Embed Size (px)
Citation preview
Molecular Analysis of Normal and Mutant Forms of the Androgen Receptor and Their Interactive Properties
Valerie Panet-Raymond Department of Biology
McGilI University
June 1999
A Thesis submitted to the Faculty of Graduate Studies and Research in partial fùlfillment of the requirements for the degree of Master of Science
O Vaierie Panet-Raymond, 1999
National Library Bibliothèque nationale du Canada
Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 w.yhgton Street 395, nict W ~ g t o f l OtÉawaON KlAôN4 W O N K l A W CPnada Canada
The author has granted a non- exclusive licence ailowing the National Library of Canada to reproduce, loan, distniute or sel1 copies of this thesis in microfonn, paper or electronic fonnats.
The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts fiom it may be printed or otherwise reproduced without the author's permission.
L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la foxme de microfiche/film, de reproduction sur papier ou sur format électronique.
L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de ceiie-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.
For my parents Giles and Pam
ABSTRACT
The androgen receptor (AR) is a ligand-activated transcription factor and a member
of the nuclear receptor superfatnily. Mutations in the androgen receptor are associated
with androgen insensitivity syndrome (AIS), and a neurodegenerative disease, spinal
bulbar muscula. atrophy (SBMA). Most of the mutations causing AIS are losssf-function
missense mutations whereas SBMA is caused by a gain-of-fùnction polyglutamine
expansion in the N-temiinal domain of the protein. Characterization of AR mutations has
led to a better understanding of structure-fùnction relationships of the AR and serves as a
prototype for steroid receptors mechanisms of action.
In the first paper, we examine the role of an AR mutation in causing rnild androgen
insensitivity syndrome. We found that this mutation conferred reduced transactivation by
AR through impaired interactions with the AR coactivator, TIFZ, and impaired
homodimerization.
In the second paper, we investigate the role of the AR polyGln expansion mutation
in SBMA pathogenesis. Recent evidence has implicated proteolytic degradation of
polyûln-expanded proteins and their subsequent intracellular aggregation in poiyGn-
expanded disease pathogenesis. We examined the role and composition of aggregates
using BuorescentIy-tagged AR and found that proteolysis need not be a prerequisite for
aggregation and that aggregation is not necessary for polyGln-induced cellular toxicity.
Finaliy, we characterize the novel heterodimerization of AR and ERa. We
deterrnined that this direct interaction has fùnctional implications for the transactivational
properties of both receptors.
Le récepteur des androgènes (AR) est un facteur de transcription activé par ligand
et est membre de la superfamille des récepteurs nucléaires. Plusieurs mutations dans le AR
sont associées au syndrome d'insensibilité aux androgènes (AIS), ainsi qu'à l'atrophie
musculaire spinobulbaire (SBMA), une maladie neurodégénérative. Alors que laplupart
des mutations causant AIS sont des mutations faux-sens entraînant une perte de fonction,
SBMA est causée par une expansion de I'insert polyglutaminique dans le domaine N-
terminal du récepteur, lui conférant ainsi une nouvelle fonction. La charactérisation de
dirérentes mutations du AR a pennis une meilleure compréhension des relations structure-
fonction de ce récepteur et par le fait même des méchanismes d'action des récepteurs
stéroïdiens en général.
Dans le premier article, nous examinons le rôle d'une mutation causant une faible
insensibilité androgénique. Nous avons trouvé que cette mutation diminue la propriété
transactivatnce du AR dû à un défaut d'intéraction avec son CO-activateur TIF2, ainsi qu'à
une altération de son homodimérisation.
Dans le second article, nous étudions le rôle de l'expansion de I'insert
polyglutaminique du AR dans la pathogénèse de SBMA. Nous avons examiné le rôle et
la composition d'agrégats cellulaires en utilisant un AR lié à une protéine fluorescente.
Nous avons trouvé que l'agrégation de ce récepteur ne dépend pas de la protéolyse et
n'est pas nécessairement toxique pour la cellule. La protéolyse des protéines arborant un
inseat polyglutaminique ainsi que leur aggrégation intracellulaire ont récemment été
impliquées dans la pathogénèse de SBMA.
Finalement, nous charactérisons I'hétérodimérisation du AR et du récepteur des
estrogènes de type a (ERa). Nous avons déterminé que cette interaction directe à des
implications fonctiomelles quant au pouvoir transactivateur de ces deux récepteurs.
TABLE OF CONTENTS
ABSTRACT
RÉsuMÉ
TABLE OF CONTENTS
ACKNOWLEDGEMEMENTS
ABBREVLATIONS
LIST OF FIGURES
INTRODUCTION
1. The Androgen Receptor and the Nuclear Receptor Superfamify
II. Steroid Hormone Rcceptors
m. The Androgen Receptor
ïU. 1 The Androgen Receptor Gent
m.2 The Androgen Receptor Protein Structure
III.2.1 The N-Terminal domain
m2.2 The DNA-Binding domain
UI.2.3 The Binge tegion
III.2.4 The Ligand-Binding dornain
IV. Androgen Receptor Function
IV. 1. Androgen Receptor Coregulators
N.1.1. Coactivators
N. l.2. Corep ressors
IV. 1.3. Specific AR Coregulators
V. Molecular Mechanism of Androgen Action
V. 1. Steroid Hormone Receptors Conformational
Change Associated with Ligand Binding
V.2. Dirnerization
V.3. Heterodimerization
VI. Sexual Differtntiation
W. Androgen Rceeptor Mutations
Molecular and Clinical Perspectives
W.1. Androgen Insensitivity Syndrome
ViI.2. Androgen Receptor Mutations in Prostate
and Brtast Cancers
VII.3. Spinal Bu1 bar Muscular Atrophy
W.3.1. Ctinical Features
VII.3.2. Pat hogenesis
VI1.3.2A In tracellular Aggregates
REFERENCES
OBJECTIVES
CONTRIBUTIONS OF AUTHORS
INTRODUCTION TO CHAPTER 1
CHAPTER 1: Oligospermic infertility associated with androgen receptor mutation that reduces DNA binding and disrupts interactions between domains and with the coactivator TIFZ.
INTRODUCTION TO CHAPTER LI
CHAPTER II: Cbaracterization of intracellular aggregates using fluorescen tly-tagged polyglu tamine-erpanded androgen receptor.
INTRODUCTION TO CEtAPTER III
CHAPTER III: Eeterodimerization between the androgen receptor and estrogen receptor a affects their respective transactivational properties.
CONCLUSIONS
1 would like to thank my supervisor, Dr. Leonard Pinsky, for his wondefil s u p e ~ s i o n and support and for always making tirne for me. Your discussions were invaluable.
1 would like to thank Dr. Mark Trifiro, my CO-supervisor, for his never-ending patience and support and his technical expertise.
1 am also grateful to Dr. Lenore K. Beitel for her unbeiievable technical help and for always taking the time to answer my questions and lend a helping hand. Thank you also for your fnendship and for providing the lab with your cheerful attitude combined with quiriq stones.
My thanks to Rose Lumbroso for always being cheerfùl and willing to help me. Thanks to Carlos Alvarado for his assistance and for always being quick to prevent my invariable lab disasters.
1 would like to thank Dr. Bruce Gottlieb for knowing so much about everything and sharing it with us d l .
1 would like to thank Annie for her fnendship and kindness and for translating the abstract. 1 would also like to thank Devorah Felman, Shereen Ghali, Vadim Khalil, Dr. Zhi Qiang Yuan for their wondefil attitudes in the lab.
Thank you to past students who introduced me to the Iab including Sunita de Tourreil, Youssef Elhaji and Dr. Abdullah AR. Abdullah who taught me many lab techniques.
1 would also like to acknowledge Fran Langton, Lynda McNeil, Susan Bocti and Rhona Rozensweig for their help and patience.
1 would like to thank the members of my supervisory cornmittee, Drs Yutaka Nishioka and Rima Rozen for their guidance.
1 would also like to thank al1 my fnends and loved ones, especiaily Paul, for their support.
Finally, 1 would like to thank rny parents Giles and Pam and sister Christine Panet- Raymond without whom none of this would have been possible. Thank you for your endless love and support.
aa AD AF AIS ALP Anfm AR ARA ARE Ar43 Am p-Ga1 BFP BP BrCa CA1 cDNA COUP-TF
C-terminal DBD DHT DNA DRPLA EMG ER ERE FBS FL h o l GAL GFP GH Gln G ~ Y GR GRE GSF H HAP HAT HD
amino acid Activation domain Activation fùnction Androgen insensitivity syndrome Alkaline p hosphat ase Anti-müilerian hormone Androgen recepto r Androgen receptor activation Androgen response element Arginine Adenine triphosphate Beta-galactosidase Blue fluorescent protein Binding protein Breast cancer Complete androgen insensitivity Complementary DNA Chicken ovalbumin upstream promoter- transcription factor Carboxy-terminal DNA-binding domain Dihydrotestosterone Deoxyribonucleic acid Dentatorubral pallidoluysian atrophy Electromyography Estrogen receptor Estrogen response element Fetal bovine semm Full-length femtamole Gdactosidase Green fluorescent protein Growth hormone Glutamine Glycine GIucocorticoid receptor Glucocorticoid response element Genital skin fibroblast Helix Huntingtin-associated protein Histone acetylase activity Huntington disease
HDAC HDCRG HNF4 HRE HSP kb kDa Ig LBD LTR LUC LYS M mAb MAI mAR MB MEM Met MJD CLM mI MMTV MR mRNA MT N- t enninal NH2-terminal n M NLS NSC-34 NR OD ONPG ORF PBS PCa PCR PPAR Poly(A) PolyGln PR PRE Q
Histone deacetylases Huntington disease collaborative research group Hepatocyte nuclear factor 4 Hormone response element Heat shock protein kilobase kiloDalton Immunoglobulin Ligand-binding domain .
Long temiinal repeat Luci ferase Lysine Molar monoclonal Antibody Mild androgen insensitivity mouse androgen receptor Mibolerone Modified eagle's medium Met hionine Machado Joseph disease micromolar millilitre Mouse mamrnary tumor virus Mineralocorticoid receptor messenger RNA Met hylt rienolone Amino-terminal Amino-terminal nanomolar Nuclear localization signal Neuroblastoma x spinal cord Nuclear receptor Optiral density O-nitrophenyl P-D-galactopyranoside Open reading frame Phosphate-buffered saline Prostate cancer Polymerase chah reaction Peroxisome proliferator-activated receptor Polyadenylation P olygiut amine Progesterone receptor Progesterone response element Glutamine
RAL rAR mu RAR RNA IUCR SBMA SCA SeAP SF-1 SHR SSCP SRC SV40 T TAD TF TIF-2 TIS TR UTR Val VDR W16 WT Zn
Raiosene Rat androgen receptor Relative light units Retinoid acid receptor Ribonucleic acid Retinaic X receptor Spinal bulbar mupailar atrophy Spinocerebeilar ataxia Secreted aikaline phosphatase Steroidogenic factor4 Steroid hormone receptor Single stranded conformational polymorphism Steroid receptor coactivator Sirnian virus 40 Testosterone Transcriptional activation domain Transcription factor Transcriptional int ermediary factor 2 Transcription initiation site Thyroid receptor Untranslated region Valïne Vitamin D receptor Virai protein 16 Wild-Type Zinc
LIST OF FIGURES
1. The structurai organUation of the AR gene and protein p. 14 2. Androgen Receptor coregulators p.2 t 3. Liganded and unliganded AR ligand-binding domains p.23 4. SSCP analyses of family members of two M886V mutation probands p.60 5 . Dissociation kinetics of normal and mutant genital skin fibroblasts p.60 6. Tramactivation activity of M886V and WT AR p.6 1 7. Transactivation activity of M886V and Wï AR with p.6 1
increasing amounts of DNA 8. DNA mobility gel shift assays of M886V and WT AR p.62 9. Transcription of normal and mutant AR fragments p.63 10. Effect of TIF2 on AR activity in Hela cells p.64 1 1. Schematic representation of the AR p.82 12. Aggregation in COS- 1 cells using polyGln-expanded and WT GFP p.85
and BFP tagged AR 13. Quantitation of the percentage of GFP-AR and GFP-AR-BFP transfected cells p.86
containing aggregates 14. Aggregation in NSC-34 cells transfected with GFP-AR, BFP-AR and p.87
GFP-AR-BFP 15. Aggregation in transfected cells treated with MI3 and 2-DEVD-FMK p.87 16. Western analysis of GFP-A.& BFP-AR and GR-AR-BFP transfected cells p.89 1 7. Yeast two-hybrid AR-ERa interactive assays p.117 18. Mamrnalian two-hybnd AR-ERa and AR-ERP interactive assays p. 120 19. Impaired AR-induced transactivation by ERa p. 121 20. Impaired ERa-induced trasactivation by AR p. 123
INTRODUCTION
1. THE ANDROGEN RECEPTOR AND NUCLEAR RECEPTOR
SUPERFAMILY
The androgen receptor is a member of the nuclear receptor superfamily, the largest
family of transcription factors in eukaryotes (reviewed in Tsai and O'Malley, 1994). By
1995, over 150 members of this superfamily had been identified (Mangelsdorf et al., 1995)
in a range of species from humans to drosophila (Koelle et al., 199 1). The high degree of
evolutionary conservation from species to species underlines the importance of this
superfiamily in the control of gene expression. The superfamily can be subdivided into four
discrete classes of receptors based on their DNA-binding affinities and their dimerization
(Stunnenberg, 1993, reviewed by Mangelsdorf et al., 1995). Class I receptors are the
steroid hormone receptors that are activated by ligand binding and bind DNA through
their respective hormone response elements (HIES). Class II receptors are also ligand-
dependent receptors and include retinoid X receptor (RXR) and some of its
heterodimerization partners. This class of recepton usually binds to direct repeats. Class
III and N include the orphan receptors, receptors with no known ligands. Class III
receptors bind homodimerically to direct repeats whereas class IV receptors typicaily bind
to extended core sites (reviewed in Mangelsdorf et al., 1995). Class IV recepton are the
only receptors in the superfamily known to bind as monomers.
Another comrnon subdivision of the superfâmily is based on ligand recognition and
groups the steroid receptors separately fiom the nonsteroid receptors. The nonsteroid
receptors include the thyroid (TR), vitamin D (VDR) and retinoid receptors (RAS R X t )
and are found to mostly interact as heterodimers. The steroid receptors include the
estrogen receptor (ER), androgen receptor (AR), rnineralocorticoid receptor (MR),
glucocoriicoid receptor (GR) and progesterone receptor (PR) and usually fùnction as
homodimers although recent evidence suggests that heterodimerization may also occur in
steroid signaling pathways (Lee et al., 1998).
II. STEROID HORMONE RECEPTORS
The steroid hormone receptors are a subdivision of the nuclear receptor
superfamily and bind ligands that are derived fiom cholesterol. Steroid hormones include
sex steroids such as progesterone, testosterone and estradiol and adrenal steroids such as
cortisol and aldosterone. Steroid hormone receptors (SHR) have a fairly consented
modular structure includig an amino-terminal transactivation domain, a DNA-biding
domain, a nuclear localization signal and a ligand-binding domain (Figure 1). They usually
bind as homodimen to hormone response elements (HRE) located upstream of target
genes (reviewed in Beato et al., 1995). The GR PR and AR al1 recognize a
consensus HRE sequence organized as an inverted repeat (AGAACA) (Beato et al.,
1995). They each ais0 have their own specific response elements that are determined by
nearby DNA sequences and may bind tissue-specific transcriptional coregulators. The
estrogen receptor recognizes a half-site AGGTCA that is used by other receptors in the
superfamily. When unliganded, SHRs have been shown to be complexed with a number of
chaperones that maintain the receptors in an inactive conformation (reviewed by Pratt,
1993; Beato et al., 1995). Once bound to their ligand, they shed their chaperones and
interact with their HRE to bring about transcription of their target genes.
While steroid hormone receptors are known to be crucial for a multitude of
physiological processes, transgenic animal knock-outs have dlowed a closer delineation of
the different receptors' roles in development and pathophysiology. Interestingly,
transgenic rnice lacking estrogen receptor were viable but both sexes were infertile
implicating the estrogen receptor in both fernale and male sexual development (Lubahn et
al., 1993). Dismption of the human estrogen receptor in a male patient proved not to be a
lethal mutation although it aected both bone maturation and mineralization (Smith et al.,
1994). Glucoco~icoid receptor-nul1 rnice were also viable but died shortly after birth due
to a lack of lung maturation (Cole et al., 1995). There have been human examples of SHR
gene deletions and loss-of-fùnction mutations most especially in the androgen receptor
(AR). The wide spectrum of AR mutations associated with variable clinical phenotypes
has pemiitted a better understanding of the protein structure and fùnction and, by
homology, of the other steroid recepton as weil (for a review see Sultan et al., 1993).
III, THE ANDROGEN RECEPTOR
III. 1. ANDROGEN RECEPTOR GENE
The human androgen receptor (AR) has been mapped to Xq 1 1 - 12 on the long arm
of the chromosome (Brown et al., 1989). It is encoded by a single copy gene that spans
over 90 kilobases (kb) of genomic DNA (Kuiper et al., 1989) although only -2750 bp code
for arnino acids. The coding region comprises eight exons (Brown et al., 1988) that are
separated by introns up to 26 kb in size (Kuiper er al., 1989) (Figure 1). The first exon is
the longest (over 3 kb) and codes for a long 5' untranslated region (5'UTR) and the N-
terminal portion of the protein (1586 bp). The AR varies in length due to two tnnucleotide
stretches in the first exon that are both polyrnorphic in size. The second and third exons
each encode a DNA-binding zinc finger. Each zinc finger encompasses an a-helical
domain, the first of which is responsible for direct DNA-binding, the second allows for
stabilization of the interaction through hydrophobie interactions. The 5' region of exon 4
encodes the hinge region of the receptor including the nuclear localization signal. The rest
of exon 4 and remaining exons (4-8) contain the sequences for the Iigand-binding domain
as well as the 3' untranslated region (UTR). The 3' UTR is 6.8 kb and contains two
polyadenylation sites (Faber et al., 1991). The overall genomic stnicture (eight exons) is
shared by al1 related rnembers of the steroid receptor family.
Gln tract (mnowidr) 1 58 78 538 625 91 9
hAR protein TAO . Figure 1. The structural organization of the AR gene and protein. Top: exons and introns of the AR gene are shown. Bottom: schematic representation of AR protein with main fimctional domains. The transcriptionaI activation domain (TAD). DNA-binding domain @BD) and ligand-binding domain (LBD) of the AR protein are shown.
The 5' OTR, encoded by exon 1, is approximately 1.1 kb in size and contains two
translation-initiating sites (TIS) that are 13 bp apart (Faber et al., 1991), although only the
fint (TIS 1) has been used in tissues examined (Tilley et al., 1990). The finction of the
second (TIS 2) is unknown. The AR gene does not have a TATA or CAAT box in its
promoter although it does contain a GC-rich binding site for the SPI transcription factor
at -46/-37 and other transcription factors (Faber et al., 1993). It also contains an
adenindguanine rich stretch, a CAMP response element and AP-1 binding sites (Mirokami
et al., 1994). The nucleotides -74 to +84 surrounding the TIS contain the minimal area of
the promoter necessary for AR gene transcription (Quigley et al., 1995 and refs therein).
Overall, very little is known about the control of AR gene expression.
Northern blot analysis of various tissues have detected two different AR mRNA
species, one of 8 kb and another of 1 1 kb (Faber et al., 199 1). The srnaller mRNA species
is less abundant and is the result of alternative splicing of the pnmary RNA resulting in a
3'UTR truncated RNA. Both messages contain an open-reading frame of 2757 nucleotides
(Brinkmann et al., 1989). AR mRNA and protein have been detected in a vanety of genital
and non-genital tissues including liver, testis, prostate, hair follicles and sebaceous and
prepubertal glands in hurnans (Ruizeveld de Winter, 199 1). Whole-mount rat
irnrnunohistochemistry studies have show AR expression in tissues ranging fiom breast
and neural tissues to kidney and bone.
111.2. AR PROTEIN STRUCTURE
The 2757 bp ORF encodes a 910-919 amino acid (aa) protein of 110-1 14
kilodaltons (Quigley et al., 1995). Wilson and McPhaul (1994) have found an 87
kilodalton form of the AR believed to result fiom a downstream translation initiation site
in the amino-terminus. There is variability in protein size and weight due to two
polymorphic triplet repeats in the first exon of the protein. The first is a polyglutamine
tract which varies in repeat size fiom 9 to 36 (Andrew et al., 1997) and averages 2122
glutamines (La Spada et al., 1991). It is expanded in spinal bulbar muscular atrophy
patients (discussed below). The second is a polyglycine stretch that varies nom 10 to 3 1
arnino acids (Lumbroso et al, 1997).
The AR is a single polypeptide chah that contains three pnmary fundional
domains. The first is an amino-terminal (N-terminal) transactivation domain encornpassing
amino acids 1 to 537, the second is the central DNA-binding domain including aa 538 to
627 and finally there is a carboxy-tenninal (C-terminus) androgen-binding domain
(residues 670-9 19). A nucIear localization signal exists between the LBD and DBD of the
AR and is encoded by part of the founh exon. There are also phosphoqdation sites in the
N-terminus of AR that cause a shifiing from a 110 kDa protein to a 112 kDa isoform
(Kuiper et al. 1994)
III.2.1 The N-terminal domain
The N-terminal domain is encoded by exon 1 and includes the S'UTR and amino
acids 1-539 of the AR. It is the least homologous domain among the steroid receptors
with only 16-25% conservation between AR and the receptors for glucocorticoid (GR),
progesterone (PR) and mineralocorticoid (MR). Nonetheless, each receptor has an amino-
terminus that is hydrophilic and negatively charged. The N-terminus contains two of the
AR'S transcriptional modulatory domains: AF- 1 a and AF-1 b. In the rat AR (AR), AF- 1 a
is a 14 residue stretch homologous to the human AR that forms a beta turn followed by an
alpha helix (Chamberlain et al., 1996). Mutations of either of two hydrophobic arnino
acids within the AF-la is accompanied by a 60% reduction in transactivation by AR
(Chamberlain el al., 1996). The AF-lb is 65 amino acids long and contains multiple
aspartate and glutamate residues. Its deletion decreases AR-transactivation by 55%
(Chamberlain et al., 1 996).
An interesting feature of the amino-terminal domain is its many homopolymeric
stretches. There are six diffèrent ones, two of them are polymorphic. The polymorphic Gln
tract is the most N-terminal of the hornopolymeric tracts and is encoded by (CAG),,CAA.
The other polymorphic tract, the polyglycine tract, is encoded by (GGN),,. Of the
remaining non-variable homopolymeric tracts, the longest is an octaproline (8 proünes)
stretch. There are also two invariant Gln tracts (6Gln and 5Gln) and a pentaalanine stretch.
Many of these tracts have been conserved arnong species. Both the mouse AR (mAR) and
rat AR (AR) contain polygiutarnine tracts that are encoded by a mixture of both CAG and
CAA codons (Faber et al., 1991; Chang et aL, 1988). Polyproline, polyalanine,
polyglycine and polyarginine tracts are also found in the rAR possibly indicating an
evolutionary importance to homopolymeric tracts.
The roles of the homopolymeric stretches are largely unknown. The length of the
polyrnorphic Gln repeat is known to modulate transactivational abiiity of the AR. Longer
tracts cause a reduction in AR transactivational cornpetence (Mhatre et al-, 1993; Kazemi-
Esfajani et al., 1995). It has been postulated that the tracts impact on transcriptional
regdation through interactions with other transcription factors (TF) (Gerber et al., 1994).
This postdate was supported by the finding that proteins in a SwissProt database
containhg a higher degree of glutamine and proline residues were mainiy TF (Gerber et
al., 1994).
ïïI.2.2. The DNA-binding domain
The DNA-binding domain (DBD) of AR is approximately 70 aa long and is
encoded by exons 2 and 3 of the AR. In contrast to the N-terminal domain. the DBD is
highly conserved arnong members of the steroid receptor family. Out of 65 arnino acids,
20 residues are conserved. Most of the conserved residues are cysteines that anchor the
zinc fingers. Al1 10 cysteines of the AR, PR and GR and 9 in the ER are conserved (Chang
et al., 1988 and refs therein). The zinc fingers are key stmctures in a steroid receptor's
DBD. In each finger, a zinc ion is surrounded by a tetrahedron of four cysteines and 2 a-
helixes (reviewed in Pinsky et al., 1996). The zinc fingers in AR have approximately 80%
homology with those of the GR MR and PR (Freedman et al., 1992). The DBD mediates
the interaction between the AR and the hormone response elements (HRE) in target genes
through its fint zinc finger. Crystailographic studies of the glucocorticoid receptor have
shown that the first zinc fuiger mediates contact between the receptor and the major
groove of the DNA (Luisi et al., 1991). The second zinc finger mediates
homodimerization and stabilizes the AR-DNA interaction through interaction with the
DNA phosphate backbone and through hydrophobic interactions (reviewed by Quigley et
al., 1995). The "D-box", formed by 5 residues at the base of the C-terminal zînc finger, is
necessary for AR dimerkation and contributes to the stabilization of the subsequent
interaction with DNA (Pinsky et al., 1996).
The hîgh degree of residue conservation arnong the DBDs of the GR, PR, MR and
AR contributes to their recognition of the same hormone response element (HRE). The
half-sites are recognized by several residues at the base of the first Pnc finger which confer
specificity of binding between the SHR and the HRE @anielsen et al., 1989). Although,
G& AR and PR are al1 able to tramactivate genes downstream fiom a mouse mamrnary
tumor virus (MMTV) promoter, they aiso each have their own specific hormone response
elements. Recognition of specific HREs by the SHR is thought to be mediated by
upstrearn and downstream adjoining sequences to the HEE on the target genes. Cell-
specific coregulators are also believed to be a factor in target gene recognition.
iII.2.3 The hinge region
The hinge region of the AR molecule is composed of amino acid sequences
between the DBD and ligand-binding domain (LBD) of the AR. It is encoded by the 5'
region of exon 4 and is not highly homologous among steroid receptor family members. It
contains a bipartite nuclear localization signai (Zhou et al., 1994) which mediates
transport of the AR from the cytoplasm to the nucleus following androgen binding. It is
formed by five highly basic residues preceded by a dyad of Arg-Lys. The hinge region may
also be involved in repression of the transcriptional activation domain in the AR-LBD
possibly by providing a binding region for interacting proteins. In a yeast-two hybrid
experiment, addition of the hinge region to the LBD in constructs caused a significant
attenuation of the AR-LBD transactivational domains' (AF-2) activity (Moilanen et al.,
1997).
IIL2.4. The Iigaad-binding domain
The C-tenninal portion of AR contains the ligand-binding domain (LBD). The
main finction of the LBD is to mediate the interaction between androgens and the AR. It
also suppresses AR tramactivational activity in the absence of ligand. When unbound, AR-
LBD interacts with heat shock proteins and other chaperones (HSP) such as HSP 90 that
maintain the AR in an inactive conformation while displayhg the androgen binding site
(reviewed in Sultan et al., 1993). ARS lacking 75% of the LBD display androgen-
independent constitutive activity suggesting that the LBD does behave as a repressor in
the unbound state (Rundlett et al., 1990). The LBD binds directly and specifically to both
physiological androgens, testosterone (T) and dihydrotestosterone (DHT). There is a
ligand-dependent transcriptional activation domain within the LBD that is called AF-2.
The AF-2 is known to provide an intenace for many AR-interacting coregulators such as
the transcriptional intermediary factor 2 (TIF2) and may mediate ligand-dependent AR
transcriptional activity through its interactions with coactivators and corepressors. In
accordance with this postdate, a point mutation in the AR-LBD, E888Q, was s h o w to
decrease the stimulatory effect of TIF2 on AF-2 Nnction (Berrevoets et ai., 1998). The
LBD also contributes to the process of receptor homodimerization; it is unknown whether
this function is accomplished through N-terminal or C-terminal interactions.
While each domain of the AR and related steroid receptors is ascnbed a specific
fiinction, the modularity is indistinct. The various domains may contribute differentially to
a specific task in steroid signaling; however, their interactions with each other and other
interacting proteins cause them to affect other fùnctions as well, including nuclear
localization, dimenzation and gene transactivation (Pinsky el al., 1998).
IV. AR FUNCTION
IV.1. AR coregulators
Many proteins have been isolated over the 1s t few years that interact with SHRs
and fiinction in modulating their transcriptional effects. SHRs are believed to fùnction as
part of a large complex of proteins that coordinate the transcription of target genes. The
coregulators can be subdivided into 2 distinct categories: coactivaton and corepresson.
IV. 1.1. Coactivators
The AF-2 domain in the LBD o f many SHRs is known to mediate transcription
through interactions with coactivators. Once bound to ligand, the AF-2 and the rest of the
SHR LBD assumes a conformation to recruit interacting coactivators that will potentiate
transactivational AF-2 activity (Torchia et ai., 1998; Horwitz et ai., 1998; Barettino et
al., 1 994). Three families of SHR coactivators have been isolated to date that are related
by their overall structure. SRC/NCOA- 1, TIFUGrip lNCoA-2 and p/CIP/AIB l/TRAM 1
al1 have an N-terminal helix-loop-helk structure followed by a senne threonine-rich region
and a C-terminal glutamine-rich region (Halchmi et al., 1994; Cavailles et ai., 1994; Onate
et al., 1995). They each possess two interacting domains: one allows interaction with
S m , the other with the CBP/p300 family of transcriptional coactivators (Kamei et al.,
1996; Torchia et al., 1997; Voegel et ai., 1998; Yao et aL, 1996). Many are capable of
interacting with multiple memben of the SHR family and potentiate transactivation
through their two transactivation domains @ing et al., 1998).
In addition to their relatively conserved structure, the coactivaton also possess
LXXLL motifs. These sequences have been show to be both necessary and suficient for
SHR interaction (Heery et al., 1997).
Afier binding a ligand-activated SHR, coactivators recruit other mernbers of the
activating complex, notably CBPlp300 (Yao et al., 1996). CBPfp300 and SRC proteins
both possess histone acetylase (HAT) activity, an enzyme necessary for transcnptionai
activity (Ogryrco et al, 1996; Bannister et al., 1996; Spencer et al., 1997). HAT activity
facilitates nucleosome disruption and allows the binding of TFs to the target genes.
N. 1.2. Corepressor~
When unliganded, several SHRs have been s h o w to be involved with
corepressors. NCoR and SMRT are two proteins that mediate SHR repression in the
absence of ligand-binding (Horlein et al., 1995; Kurokawa et al.; 1995; Lee et al., 1995;
Chen et al., 1995). Corepressor binding is sometimes dependent on antagonist binding by
the SHRs. Both PR and ER need to have undergone a conformational change instigated
by antagonist binding, in order to bind corepressors. These findings impiicate both
corepressors and coactivators in SHR responses to various antagonist and agonist iigands.
Corepressor-SHR binding allows fkther interactions with other proteins such as histone
deacetylases (HDAC) that are associated with transcriptionally silent DNA elements.
IV. 1.3. Specific AR coregulators
While many coregulators have been shown to bind multiple SHRs, certain
coregulators have been found that specifically interact with AR (Fig. 2). ARA-70
(Miyarnoto et al., 1998), ARA-55 (Fujimoto et al., 1999), ARA54 Wang et al., 1999) al1
bind to the AR-LBD in the AF-2 region and fùnction as coactivators. Al1 were isolated in
prostatic ce11 lines.
Figure 2. AR coregulators. AR coregulators and interacting proteins are s h o w binding to their putative binding sites on the AR protein. (Modified from Beitel et aL, 1998). The androgen receptor-activation proteins ARA-?O, ARA-55 and ARA44 are shown.
V. MOLECULAR MECHANISM OF ANDROGEN ACTION
The AR fiinctions like the other members of the SHR. Its action is dictated by
ligand binding. Androgen induces a conformational change in the AR structure (discussed
below) that is due in part to the loss of chaperone proteins like HSP90 that maintain the
AR in an inactive conformation. The associated change in AR conformation dlows
dimerization with another androgen-bound AR as well as allowing interaction with the
many coactivators that serve to potentiate AR-mediated transactivation. Before
dimerization, the androgen-AR complex is transported fiom the cytoplasm uito the
nucleus via its nuclear localization signal. Once in the nucleus, AR homodimers bind to an
androgen response element (ARE) on a target gene. Then, through interactions with
coactivators and members of the transcription initiation cornplex, increased transcription
of the target gene occurs.
V.1. SHR conformational change associated with ligand binding
X-ray crystallographic studies of certain SHRs have revealed the molecular details
of SHR structure and allowed a closer examination of the conformational change induced
upon steroid binding. The three-dimensional structure of the LBDs of RXRa, M y , TR
ERa and PR have ail been determined (Bourguet et al., 1995; Renaud et d, 1995;
Wagner et al., 1995; Brzorowski et al., 1997; Williams et al., 1998 respectively). The
SHR LBD al1 show a similar conformation reflecting the overall LBD sequence homology
in the SHR family that leads to a conservation of structure (Wurtz et al., 1996) (Fig. 3).
The LBD is composed of twelve a-helices, one smd, two-stranded, antiparallel f3-sheet
and one R loop. The a-helices form a three-layered antiparallel a-helicai sandwich in
which the central core of helices HW6, H9 and H10, are bordered by helices H1-4 and H7,
H8 and Hl1 on either side (Brzozowski et al., 1997). Hl2 and the P-sheet flank the
helical sandwich. At the narrow end of the structure, the a-helices wedge and form a
highly conserved hydrophobic pocket that fùnctions as a cavity for ligand-binding. H3, H6,
H8, Hl 1, Hl2 and the SllS2 hairpin al1 contribute to forming the ligand-binding pocket
(B~ozowski et al., 1997). Upon ligand-binding, Hl0 and Hl 1 rearrange to fom one
continuous helix and Hl2 shortens releasing it from its interaction with the fl loop. The R
loop then flips over and allows H l 2 to move over and cover the ligand-binding cavity
(Fig. 3). This is referred to as the mouse trap model where the spring (H12) is released
upon ligand-binding (reviewed in Parker and White, 1996). The sequence conservation
among LBDs of the steroid receptor family suggests that ligand-binding most likely
induces the sarne structural change in al1 SHRs. Indeed, the realignment of Hl2 over the
ligand-binding cavity has been observed in ail liganded nuclear receptor LBD structures
that have been elucidated to date (Wagner et al., 1995; Renaud et al., 1995; Bnozowski
et al., 1997; Williams et al., 1998). It is believed that this realignment is essential for
transcnptional activation of the receptors as it generates an accessible AF-2 domain
allowins it to interact with necessary coactivators thus promoting transactivation.
Brzozowski et al. (1997), also determined the structure of the ERa LBD in the presence
of the ER-antagonist, raloxifene (RAL). They detennined that RAL binding in the
hydrophobic pocket did not allow the realignment of the Hl2 over the pocket and instead
Hl2 moved into a groove formed by helix 5 and 3 thereby preventing the AF-2 fiom
becoming accessible to coregulators (Brzozowski et of., 1997). The conformational
differences of receptor LBDs in the presence of agonists and antagonists provide a
molecular basis for their different physiological effects.
Unllgsnded Uganded
Figure 3. Schematic representation of the liganded and unliganded AR ligand-binding domain (Gottiieb et al., 1998). The conformation of many AR domains are altered upon ligand-binding with respect to both position and direction within the protein.
23
V.2. Dimerization
M e r ligand-binding, PR, GR, Ek MR and AR each bind response elements on
target genes as homodimers. Most SHRs, like PR, GR and ER, have been shown to
homodimerize prior to binding HREs (Cairns et ai., 199 1 ; Rodnguez et ai-, 1990; Fawell
et ai., 1990). While steroid receptor diierization is well documented, the exact dornain-
interactions and orientations of the receptors during dimerimion remain unclear. One of the
known dunerization sequences is the D-box of the DNA-binding domain as s h o w by
crystaiiographic studies of the ER and GR (Luisi et al., 199 1; Schwabe et al, 1993). The
LBD is also beiieved to contain dirnebtion sequences. Using GST fùsion proteins in
Ekcherichia cdi, Nernoto et al., (1994) demonstrated that AR-LBD fùsion proteins could
homodimerize in vivo. Yeast-NO-hybrid studies have proven vaiuable for elucidating the
hctional in v h interactions between the various domains of nuclear receptors. This system
has proven that dimerization of the ER and AR are Liganddependent in v i w (Doesburg et
ai., 1996; Wang et ai., 1995). However, this system has yielded conflicting results regarding the
parallel or anti-parallel nature of the dimerization interaction (Doesburg et ai-. 1996; Langley,
et al.. 1998). In Doesburg et aL (1996), the AR N-terminal transcription activation domain
(AR-TAD) was shown to strongly interact with the AR-LBD in vivo in a yeast-two hybrid
system. The LBD-LBD interaction was only detectable afier a high expression vector for the
GAUDBD-ARLBD fùsion protein was introduced. Doesburg et al. (1996) proposed that the
AR homodimerizes on rnany levels and that a direct or indirect interaction between the AR-
TAD and AR-LBD as weU as interactions between the DBDs and a weak LBD-LBD
interaction occur in AR homodimen. This mode1 daers fiom one suggesting that the AR
homodiierizes in an antiparallel N-C terminai fahion. Langiey et al. (1995) detemiined that
the AR-TAI3 and AR-LBD directly interact, without any LBD-LBD interaction, and proposed
an antiparaiiel mode1 of AR dimerization.
Ln a recent paper by Ghadessy et al. (1999) (a part of this thesis), a mutant AR
affecting dimerization was studied. The mutation, a single adenine-to-guanine transition
that changes codon 886 in exon 8 fiom methionine to valine, had no effect on androgen
binding. However, the Met886Vai mutant receptor did show a consistent decrease in
transactivation of two different androgen-inducible luciferase reporter genes that were CO-
transfected into three ce11 types. Through yeast and marnmalian two-hybnd studies, the
authors demonstrated that the mutation affected both dimerization and interaction with the
coactivator TIFZ. Interestingly, both TAD-LBD and LBD-LBD interactions were shown
to be disnipted by the mutation.
Recent crystaUization studies of the ER and PR LBDs have aided in the detennination
of the dimerization interfaces in SHRs (Brzozowski et al., 1997; Williams et al., 1998). A
direct interaction between a-helices of the LBDs was demonstrated in both cases. In ER
LBDs, the H8 and Hl 1 a-helices line up and form a dimecbation interface that interacts with
the same dimer interface on another ER-LBD molecule (Brzozowski et al., 1997). Most of
the contact is made through the Hl 1 helices although sections of Hg-Hl0 are also involved. In
the PR the Hl2 has a C-terminal extension which prevents it fiom f o h g the same strong,
dunerized structure as ER (Williams et al., 1998). Given the sequence conservation arnong
members of the SHR f d y , direct LBD-LBD interactions probably occur in ali steroid
hormone recepton. The AR bean more homology to the PR and is likely to foliow its mode1 of
dimerization which indicates a weak interaction between LBDs.
V.3. Hetemdimerization
Whiie certain SHRs are known to fiinction as homodimers, others are known to exist
primarily as heterodimers. The 94s retinoic acid X receptor (RXR) is a cornmon heterodimer
partner for several nuclear receptors including TR, RA& VDR and the peroxisome
proliferator-activated receptor (PPAR) (Yu, 199 1, Kliewer, 19924 Kiiewer, 1 W2b, Marks,
1992, Issemann, 1993, Bugge, 1992, Zhang, 1992). More recently, some of the SHRs that
fùnction as homodimers have been show to intetact with other members of the nuclear
receptor superfamily. ERa heterodiierizes with various nuclear receptors includiig ERP,
TR and RXR aibeit under forced conditions (Lee et al., 1998). These interactions
provide a basis for the diversity in steroid sigMUing pathways and may indicate the existence of
novel steroid receptor tnuiscnptional effis .
VL S E X U A L DIFFERENTIATION
The SRY gene, located in the testisdetermining region of the Y-duomosome, is
believed to trigger testis developrnent fiom the bipotential gonad (Berta et al., 1990, Sinclair et
al., 1990). SRY is a DNA-bindhg protein and contains an HMG box Wce rnany other
transcription fkctors. Upon DNA-biidiig, SRY causes the activation of various tdcular
differentiation genes including aeroidogenic factor I @Fi). SF-1 is an orphan nuclear
rrccptor that is expressed in the urogenital ridge before testis differentiation and is believed to
be a key component in testis formation (Ikeda et al., 1994). SOX9 and DAX-1 are also
believed to be involveci in male gonadal development.
There are two rnasculinizing hormones necasary for development of male sex
characteristics: anti-miillerian homone (AMH) and testosterone. AMH is produced by the
Sertoli cells and tesiosterone by the Leydig cells. AMH is raponsible for MüUerian duct
degeneration and may be regulated by SF-1 (Josso et al., 1977; Shen et al., 1994). Without
AiW& vestiges of the Miillerian duct can be seen in vintized males. Testosterone is converted
by the enzyme Sa-dihydrotestosterone in the urogenital sinus into dihydrotestosterone (DHT),
a more potent androgen (Siteri and Wdson, 1974). DHT is essentid for male development as
extemal genitalia formation is under t s control. Woltnan duct dierentiation is controlied by
testosterone itself (Gilbert, 1 994).
VIL AR MUTATIONS; MOLECULAR AND CLINICAL PERSPECTIVES
To date, there are 309 entries in the androgen receptor gene mutations database
that represent 200 different AR mutations (Gottlieb et al., 1998). Most of the mutations
reported are rnissense mutations causing a form of androgen insensitivity syndrome (AIS).
There are also reported mutations that are associated with and perhaps predisposing to
certain cancer, like prostate or breast cancers, as well as trinucleotide repeat
polymorphisms associated with increased cancer-risk. Of special interest is a CAG
expansion in the amino-terminus of AR that causes Spinobulbar Muscular Atrophy, a
neurodegenerative disease aKecting spinal cord and motor neurons.
A variety of mutations causing disease have been reported in AR including
missense mutations and partial gene deletions (Gottlieb et al., 1998). The AR gene
mutations in tum affect AR mRNA and protein structure, amount or fbnction.
ml. Androgen Insensitivity Syndrome
Androgen insensitivity syndrome(AIS) comprises a wide spectrum of clinical
phenotypes. The phenotypes range in severity fiom complete androgen insensitivity,
characterized by female body habitus, to mild androgen insensitivity in which the habitus is
male usualIy in association with gynecomastia and infertility (Trifiro et al., 1991). AIS can
be caused by many different molecular defects in the androgen receptor. Typically,
mutations causing AIS affect either the DNA- or steroid-binding domains causing
androgen resistance (McPhaul et ai., 1993). Although a functional androgen receptor is
essential for genetic viability and reproduction, most mutations do not appear to affect the
mortality or morbidity of afTected individuals. Notably, mutations causing a complete
abolition of AR fiinction lead to an external female phenotype in XY individuals (Trifiro et
al., 1991, Quigiey et al., 1995). AIS affects approximately 1 in every 20 000 to 60 000
XY births (Quigley et al., 1995). This estimate may not include some undiagnosed milder
forms of disease.
Complete androgen insensitivity (CAI) results from AR mutations that severely
alter the quality or quantity of the AR and affects 2-5 per 100 000 (Pinsky et al., 1998).
Subjects are female in appearance but may have undescended testes. While their extemal
genitalia are female, they may have clitoromegaly or posterior labial fision. Development
is usually normal until puberty. They are sterile and amenorrheic since Müllerian duct
regression being androgen-independent occun nomally. CA1 individuals usually have
absent to sparse pubic and axillary hair. Occasionally Müllenan or Wolffian duct remnants
are seen (Rutgers and Scully, 199 1).
Mild androgen insensitivity subjects are phenotypically male. After puberty, MAI
subjects show a variety of charactenstics associated with disease including small genitalia,
gynecomastia, a high-pitched voice, sparse sex hair or impotence (Pinsky et al., 1998).
Fertility rnay or rnay not be Sected depending on the severity of the AR mutation. The
hquency of MAI in the population is unknown since many MAI cases with minor
syrnptoms are not diagnosed.
Partial androgen insensitivity comprises a wide spectrum of phenotypes which
includes al1 the forms of AIS in between MAI and CAL Individuals with PA1 range fiom
having a predominantly femaie phenotype to being predorninantly male with cases of
fiankly ambiguous genitalia in between. Unlike families with CA1 where most afTected
individuals in one family bear the same phenotypic expression of disease, families with PA1
rnay show a great deal of variation in disease expression: while some a f k t e d individuals
rnay have nearly male extemal genitalia, others rnay be nearly female or show ambiguous
genitalia.
The continuous spectmm of phenotypes associated with AIS mirrors the genetic
spectrum of mutations in the AR. Many of the mutations in the N-terminal portion of the
AR result in CAI by causing premature termination of translation. Two complete deletions
of AR have been reported each resulting in CAI (Trifiro et al., 1991; Quigley et al., 1992).
Some missense mutations cause a decrease in AR M A levels (Choong et al., 1996;
Marcelli et al., 199 1). Most M A I mutations are missense mutations that do not severely
impair AR tùnction.
VII.2. AR mutations in prostate and breast cancers
Prostate cancer (PCa) is the second leading cause of male cancer deaths in the
United States. Over the last 25 years, the number of men that have been diagnosed each
year with the cancer has risen 30% (reviewed in Ruijter er al., 1999). Multiple
epiderniological studies have revealed a clear genetic component to PCa. Most PCa
progress fiom an androgen-dependent phase where up to 75% of tumors are responsive to
anti-androgen treatrnent to an androgen-independent phase where the tumors become
resistent to the endocrine therapy (MacLean et al., 1995; Suaiki et al., 1993). It was
initially believed that AR rnay be involved in the progression to androgen-independence in
PCa perhaps through mutations and decrease in expression. However, studies perfonned
on PCa tumors revealed AR expression in al1 tumors regardless of disease stage (reviewed
in Ruijter et al., 1999 and refs therein).
Nonetheless mutations in AR have been reported in PCa samples. To date, 39 AR
gene mutations have been reported (Gottlieb et al., 1998). The mutations are spread dong
the length of the gene, although mutations in the LBD have generated the most interest.
Several mutations in the AR-LBD have been found to alter the ligand-specificity of the
receptor causing the AR to be activated by a wider spectrum of steroids (Veldsholte et al,
1990; Culig et al., 1993). Veldsholte et al. (1990) discovered a mutation in the AR in
LNCaP cells, a PCa ce11 line, that had an increased growth rate in response to estrogens
and progestins. It was later demonstrated that the mutation in the LBD conferred an
increased binding affinity for estrogens and progestins (Veldsholte et al., 1992).
In addition to missense AR mutations, AR gene amplification has also been
suggested as a cause for PCa progression to androgen-independence. In one study, AR
gene amplification was observed in 28% of hormone-recurrent tumors but not in any
primary tumors (Koivisto et al., 1997).
PCa risk was found to be modulated by the length of the polymorphic
homopolymenc tracts in the AR (IMne et al, 1995). Shorter polyglutamine @olyGln)
tracts and longer polyglycine (polyGly) have both been associated with a higher nsk of
PCa (Giovannucci et al., 1997; Sleddens et al., 1993). The exact mechanism by which
homopolymenc tract length may modify risk is unknown. Shorter polyGln tracts in AR
have been associated with increased transactivity (Mhatre et al., 1993; Kazemi-Esfarjani et
al., 1995). AR'S increased transcriptional capacity may result in increased androgen-
induced gene transcription and contribute to disease progression.
In primary PCa, AR mutation rates have varied between O and 41% indicating that
initial development of disease often occurs without any AR mutations (reviewed in Ruijter
et al., 1999). The role of AR in PCa is still undefined and requires fùrther investigation.
In addition to modifjing PCa risk, some AR mutations have been associated with
male breast cancer. Male breast cancer is very uncornmon, accounting for 1% of the total
number of breast cancer (BrCa) cases (Sasco et al., 1993). Two consecutive mutations in
the AR-DBD have been associated with male BrCa. AR missense mutations at positions
607 and 608 have been found in patients with PA1 and BrCa (Wooster et al., 1992;
Lobaccaro et al., 1993). The mechanism by which these mutations increase BrCa risk is
unknown ait hough loss of androgens' protective effect and altered protein-protein
interactions are believed to play a role (Lobaccaro et al., 1993; Poujol et al., 1997).
Though numerous AR mutations have been described to date, only two have been
associated with an increased risk of male BrCa indicating that mutations in the AR rnay
not be an important determinant in male BrCa.
VII.3. SBMA
Spinal and bulbar muscular atrophy (SBMA) is an X-linked, recessive form of
motor neuron disease that affects males. It is a slowly progressive syndrome that is
characterized by a loss of motor neurons in the spinal cord and brainstem (Tnfiro et al.,
1994). SBMA is caused by an expansion of a tnnucleotide (CAG) repeat in the first exon
of the androgen receptor (AR) (La Spada et al., 1991). Seven other inherited
neurodegenerative diseases have since been found to be caused by expanded CAG repeats,
including Huntington disease (HDCRG, 1993), spinocerebellar ataxia types 1, 2, 6, and 7
(Orr, et al., 1993; Pulst et al., 1996; Sanpei et al., 1996; David et al., 1997), dentatorubral
pallidoluysian atrophy (Koide, R. et al., 1994; Nagafiichi S. et al., 1994) and Machado-
Joseph disease (Kawaguchi et al., 1994). Each of these diseases shows a progressive loss
of a specific group of neurons. Other in SBMA (AR) and SCA6 (ai, calcium channel), the
fiinctions of the W T proteins encoded by the polyCAG expanded genes are completely
unknown.
A common mechanism of pathogenesis is believed ta underlie the polyglutamine
expansion diseases. The extensive analysis and characterization of the androgen receptor
makes it ideal for investigating that mechanism.
The mutation that causes SBMA is an expansion of the glutamine tract coded by
CAGs in the variable amino terminus region. in normal individuals, this tract ranges from
11-33 CAGs; in afKected individuals, the nurnber of CAG repeats is increased, ranging
fiom 40-62 (Brooks, et al., 1995). Its description is credited to Kennedy, Alter and Sung
(1968). As such, SBMA is sometimes referred to as "Kennedy Disease". La Spada et al.
(1991) were the first to descnbe the genetic expansion in the AR as being the cause of the
disease. Description of the CAG expansion provoked a stream of successN1 investigations
into other adult-onset neurodegenerative diseases and opened up an entire field of
research.
V11.3.1. Clinical features
SBMA affects less then 1 in 40 000 men (Andrew et al., 1997). Frequency is
higher in the Japanese population even though this disease is considered panethnic (Beitel
et aï., 1998). It is X-linked recessive and as such, only affects males. Female carriers show
few symptoms associated with the disease (Discussed below). The disease is adult-onset
with muscular weakness and atrophy usually beginning in the third to fifth decades of life,
although earlier and later onsets have been noted (reviewed by Beitel et al., 1998). It is
characterized by a progressive muscular weakness of the proximal upper and lower
extremities sometimes preceded by muscle cramps. Tremors and muscle twitching may be
seen. Eventually, with disease progression, patients also typically suffer from dysarthria
and dysphagia.
The muscular weakness is secondary to the neural degeneration suffered including
a characteristic loss of the lower motor neurons in the spinal cord and brainstem (Trifiro et
al., 1994). The motor neuronal loss in Kennedy disease is limited to motor nuclei that
express the androgen receptor (Sar, M. et al., 1977). However, some sensory neurons in
the dorsal root ganglia are affected as well (Sobue et al., 1989).
In addition to the atrophied musculature, patients also develop symptoms of
androgen insensitivity. Their symptoms are typical of mild androgen insensitivity although
in SBMA patients, symptoms develop only much later in life. Gynecomastia is fiequently
identified and is sometimes the first symptom noticed by Sected individuais. Patients may
also develop a decrease in libido7 impotence and testicular atrophy (reviewed in Quigley et
al., 1995; Batel et al., 1998). Although previously fertile, patients' spermatogenesis may
become impaired. The MAI symptoms in SBMA afEected males are indicative of a loss of
AR function with disease progression. ARS containing an expanded polyglutarnine tract
have been reported to have decreased ligand aninity in pubic skin fibroblasts (MacLean et
al., 1995) and decreased transactivation capacity in heterologous cell culture (Mhatre et
al., 1993; Kazerni-Esfajani et al., 1995). These observations can explain the androgen
insensitivity associated with SBMA. -0ther studies suggest, however that the partial
androgen insensitivity associated with SBMA is a result of reduced expression of the
androgen receptor with first exon CAG repeat expansion (Choong, et al., 1996).
Female carriers seldom show syrnptoms of SBMA. A few cases have however
been reported describing rnild, almost unnoticeable signs in a few female camen ranging
fiom EMG abnonnalities to muscle weakness and fatigue (Sobue et al., 1993; Belsham et
al., 1992). Low levels of androgens and random X-inactivation have both been
hypothesized as possible protective mechanisms in women (MacLean et ai., 1996; Beitel
et al., 1998).
VlL3.2. PATHOGENESIS
Complete deletion of the androgen receptor is not sufficient to cause motor neuron
degeneration (Trifiro, M. et al., 1991). Therefore, it has been hypothesized that the
expansion of the CAG repeat in the AR causes a gain of fbnction that is toxic to certain
motor neurons in the brainstem and spinal cord. Several mechanisms have been proposed
to explain the possible gain of function acquired with an expanded CAG repeat. The first
is that the polyglutarnine tract can fonn "polar rippers" mediated by the hydrogen bonds
formed between amide groups of neighboring proteins (Perutz, M.F. et al., 1994). In this
model, the expanded polyglutarnine tracts would form very stable complexes that would
accumulate and cause neuronotoxicity. The second hypothesis is that transglutaminases
catalyse the formation of covalent isopeptide bonds between glutamines in the expanded
polyGln tract and lysyl residues in other proieins. The durability of these bonds would
promote a slow but progressive aggregation leading ultimately to ce11 death (Green, H. et
ai., 1993). Provocatively, Igarashi et al. (1 998) found that truncated DRPLA proteins
containing expanded polyGln stretches form aggregates and cause apoptotic ce11 death,
and that aggregate formation is suppressible by transglutarninase inhibitors. Other theories
hold that because the polyglutamine-expanded AR has an enhanced ability to bind other
proteins it could titrate out important molecules that the cell needs to suMve (MacLean,
HE. et al. ., 1 996). Goldberg et ai. ( 1 996) demonstrated that the polyglutamlne-expanded
huntingtin protein has an increased rate of cleavage by apopain, a cysteine protease
involved in apoptosis. They hypothesized that cleavage of the polyglutamine-expanded
protein could result in a toxic breakdown product or in the "inactivation of an anti-
apoptotic tùnction" of the protein resulting in pathogenesis (Goldberg, et ai., 1996). Other
groups have since shown that polyGln-expanded proteins have caspase cleavage sites.
Huntingtin, atrophin-1, ataxin-3 and AR have dl been shown to be substrates for caspases
and are cleaved by apoptotic cell extracts (Goldberg et al., 1996; Miyashita er al., 1997;
Wellington et al., 1998). Mutation of one of the two caspase cleavage sites in the AR
prevented the formation of intracellular aggregates and reduced apoptosis (Wellington er
ai., 1998). The mechanism that confers cellular specificity in polyGln-expanded diseases is
unknown but it may involve cell-specific expression of proteases iike the proapoptotic
caspases. The extent of the involvement of caspases in the pathogenic process is unknown
and requires fùrther investigation. Caspase cleavage of polyûln-expanded proteins may be
a consequence of cells entering apoptosis rather a cause of neurodegeneration.
Al1 of the former hypotheses involve a translated protein. It is conceivable
however, that neuronotoxicity could be induced by a DNA-mediated mechanism without
the need of a translatable product. For example, a DNA-protein interaction could occur
whereby the CAG trinucleotide repeat would bind proteins, again titrating out important
molecules that the ce11 needs. Also the repetitive sequence in mRNA could be a magnet for
binding other proteins, RNA molecules or DNA molecules. Recent transgenic models,
however, strongly support the hypothesis that pathogenesis requires translation of the
C AG-expanded genes.
VII.3.2.A. In tncellular aggregatts
Cha and Dure (1994) postulated that polyGln-expanded fiagments may be
neuronotoxic by accumulating in neurons due to a reduction in the efficacy of
endopeptidases to break them down. Proteolytic cleavage of the poly-Gln containing
proteins may liberate a toxic Gln tract containing fiagment that would promote
aggregation. In accordance with this postdate, a group studying Machado-Joseph disease
induced apoptosis in cultured cells by expressing a portion of a gene, MID 1, containing
the expanded CAG repeat (Ikeda et aL, 1996). The expression of the polyGln-expanded
MID 1 protein fiagment caused aggregation in transfected cells that correlated with
cellular toxicity. Recently, there has been more evidence to suggest that tmncation of
polyGln-expanded proteins and their aggregation may play an important role in the
initiation or propagation of cellular toxicity. Initially, mice transgenic for the HD mutation
were found to develop neuronal intranuclear inclusion bodies containing huntingtin and
ubiquitin proteins (Davies et al., 1997). Aggregates are believed to represent an
accumulation of insoluble proteolytic fhgments of parent NIl-length protein. Since their
initial description, inclusions have been described in many human pathological studies,
DNA transfection studies and in transgenic animals expressing polyGln-expanded protein
fiagments (Skinner et al., 1997; Davies et al., 1997; Perez et al., 1998; Miyashita et al.,
1998; Cooper el al., 1998; Martindale et al., 1998; Kim et al., 1999). Such inclusion
bodies are typically found in the neuropathologically afKected areas and have been
characterized in HD, DRPL4 SBMA, SCAl, SCA3 and SCA7 (reviewed by Lunkes and
Mendel, 1997; Ross, 1997; Davies et al., 1998; Kim and Tanzi, 1998; Holmberg et al.,
1998). Interestingly, insertion of an expanded polyGln track in an irrelevant protein,
hypoxanthine phosphoribosyltransferase, caused a neurological phenotype and inclusion
formation in the transgenic mice expressing the mutant gene (Ordway et al., 1997). These
results indicate that polyGln expansions need not be in a specific parental protein
framework to induce neurological disease. Moulder et al. (1999) expressed polyCAG
repeats of v w n g lengths fiised to green fluorescent proteins (GFP) and found that
toxicity and aggregation increased with polyGln length in DNA transfection experiments.
Drosophila expressing a polyGln-expanded portion of the SCA3/MJD protein were also
shown to have nuclear inclusions although inclusion formation was insufficient for cellular
degeneration (Warrick et al., 1 998).
While intracellular aggregates rnay be directly toxic, they rnay also be a
consequence rather a cause of the underlying neural degeneration. Recently two reports
have undennined the importance of cellular inclusions in polyGln-expanded disease
pathogenesis. In ceUs transfected with mutant huntingtin, onset of ce11 death did not
correlate with aggregate formation (Sandou et al., 1998). In a transgenic mouse mode1
expressing a CAG expanded ataxin-1 lacking a self-association region, ataxia and Purkinje
cell pathology both developed similady to the mice expressing full-length ataxin-1 but in
the absence of cellular inclusions (Klernent et ai., 1998). While disease progression may
occur in the absence of aggregate formation in mice, it is unknown if inclusions do indeed
promote pathogenesis in humans.
Increasing evidence has also implicated proteasomal dysfunction in aggregate formation
and expanded polyGln protein pathogenesis. The proteosorne and folding chaperone HDJ-
2/HSDJ has been shown in aggregates in SCA-1 patients and in transgenic animal models
(Cummings et aL, 1998). Aggregates fonned by expanded polyGln AR were also found to
stain positively for HDJ-ZMSDJ (Stenoien et al., 1999). Interestingly, overexpression of
the chaperone protein in cells significantly repressed aggregate formation by polyGln-
expanded ataxin-1 and AR (Cumrnings er al, 1998, Stenoien et al., 1999).
References
Andrtw SE, Goldberg YP, Hayden MR: Hum Mol Gertet 6: 2005-20 1 0, 1 997.
Bannister AJ, Kouzarides T: N i r e 384: 64 1-643, 1996.
Barettino D, Vivanco Ruiz MM, Stu~enberg HG: M O J 13 : 303993049, 1994.
Beato 1M, H e d c h P, Schütz G: Cell83: 85 1-857, 1995.
Beitd ï& Trifiro M., Pinsky L: A?m&sis of TbpIet Repeaî Disorders Rubinsztein DC, Hayden MR (eds), BIOS Scientific Pubtishing Ltd., Mord, Chapter 16 1, 1998.
BeJsham DD, Yee WC, Greenberg CR Wrogemann K: J Neicrosci Sci 1 12: 133- 138 1992.
Benwoets CA, Doesburg P, Steketee K, Trapman J, Brinbnann AO: Mol Endocrimi 12: 1172-1 183, 1998.
Berta P, Hawkins Ji& Sinclair AH, Taylor A, GdEths BL, Goodfeiiow PN, Feiious M: Natire 348: 448-450, 1990.
Bourguet W, Ruff M, Chambon P, Gronemeyer H, Moras D: Nuizue 375 : 377-3 82, 1995.
Briniunann AO, Faber PW, van Rooij HCJ, Kuiper GGJM, Ris C, Kiaassen P, van der Korput JAGM, Voorhorst MM, van Laar JH, Mulder E, Trapman J: J Steroid Biochern 34: 307-3 10, 1989.
Brooks BP, Fishbeck KH: Trerd Neurosci 18: 45946 1, 1995.
Brown CJ, Goss SJ, Lubahn DB, Joseph DR Wilson EM, French FS, Wdard HF: Am J Hum Genet 44: 264-269, 1989.
Bnaowski AM, Pike ACW, Dauter 2, Hubbard RE, Bonn T, Engstrom O, ohman L, Greene GL, Gustdsson J-4 Carlquia M: Nmre 3 89: 753-758, 1997.
B u g TH, Pohl J, Lonnoy O, Stunnenberg HG: M O J 1 1 : 1409- 14 1 8, 1992.
Cairns W, Caims C, Pongratz I, Poeihger L, Okret S: JBioi C h 266: 1 122 1 - 1 1226, 1991.
Cavailies V, Dauvois S, Danielain PS, Parker MG: Prm N d A d Sci USA 91: 10009- 10013, 1994.
Chi J-a, Dure LS 4& : Life Sciences 54: 1459-1464, 1994.
Chamberlain NL, Whitacre DC, Miesfeld RL: JBiol Chem 27 1 : 26772926778, 1996.
Chang C, Kokontis J, Liao S: Proc Nat! AcadSci USA 85: 72 1 1-72 15, 1988.
Chen JD, Evans RM: Nature 377: 454-457, 1995.
Cboong CS, Kemppahen JA, Zhou 2-Y Wdson EM: Mol Endorrd 10: 1527-1 535, 1996.
Cote TJ, Blendy J 4 Monaghan AP, Krieglstein K, Schmid W, Aguzzi 4 Fanniai G, Hummler E, Unsicker K, Schutz G: Genes & Dewlopment 9: 1608-162 1, 1995.
Cooper Jy Schilling G, Peters MF, Herring WJ, Sharp AH, Kaminsky 2, Masone J, Khan F q Delanoy M, Borchelt DR Dawson VL, Dawson TM, Ross CA: Htîm Mol Genet 7: 783-790, 1998.
CuIig 2, Hobisch A, Cronauer MV, Cato AC, Hittmair 4 Radmayr C, Eberle J, Bartsch G, Klocker H: Mol EjadixrinoI7: 154 1- 1550, 1993.
Cummings CJ, Mancini Antale B, Defiancm DB, Orr HT, Zoghbi HY: Nat Genet 19: 148-154, 1998.
Danielsen HE, Fanants G, Reith A: Scan. Micros.: 3: 297-302, 1989.
David G, Abbas N, Stevanin G, Durr 4 Yvert G, Cancel G, Weber C, Imbert G, Saudou F, Antoniou E, Drabkin FI, Gemmd R, Giunti P, Benornar A, Wood N, Ruberg M, Agid Y, Mandel J-L, Bnce A: Nat Genet 17: 65-70, 1997.
Davics SW, Turmaine M, Cozens BA DiFiglia M, Sharp AH, Ross C 4 Sheninger E, Wanker EE, Mangiarini L, Bates GP: CeII 90(3): 537-548, 1997.
Ding XF, Anderson CM, Ma H, Hong H, Uht RM, Kushner PJ, Stallcup MR: Mol Eihnnol 12: 302-3 12, 1998.
Doesburg P, Kuü CW, Berrevoets CA, Steketee K, Faber PW, Mulder E, Brinbnann AO, Trapman J: Biochemisiry 36: 1052-1 064, 1996.
Fabcr PW, King 4 van Rwij HCJ, Brinkmann AO, de Both NJ, Trapman J: Biuckm 5278: 269-278, 1 99 1.
Fabu PW, van Rooij HCJ, Kuiper GGM, Ris C, van der Korput JAGM, Baarends WhA, Brinkmann AO, Grootegoed JA, Traprnan J: JBiol Chem 266: 10743-1 0749,1991.
Fabcr PW, van Rooij HCJ, Schipper Y Brinkmann AO, Trapman 1: JBiol Chem 268: 9296- 9301, 1993.
Faweü SE, White R, Hoare S, Sydenham M, Page h4, Parker MG: Proc Nad Acad Sei USA 87: 68836887, 1990.
F d m a a LP: EkbwRev 13: 129-145, 1992.
Fujimoto N, Yeh S, Kang H-Y, Inui S, Chang H-C, Mitokami A, Chang C: J Bid Chem 274: 8316-8321, 1999.
Gerber H-P, Seipel K, Georgiev O, Hofferer M, Hug M, Rusconi S, S c h a e r W: Scierce 263 : 808-81 1, 1994.
Ghadessy FJ, Lim J, Abdullah AAR Panet-Raymond V, Choo CK, Lumbroso Tut TG, Gottlieb B, Pinsky L, Trifiro MA, Yong EL: J C h Ikvest 103 : I 5 1 7- 1525, 1999.
Gübert SF: Developmental Biology (In ed): Chapter 22, 1994.
Giovannucci E, Stampfer MI, Krithivas K, Brown M, Bnifsky A, Taïma I, Hennekens CH, KantoffPW: ProcNallAcadSci UW 94: 3320-3323, 1997.
Coldbeg YP, Nicholson DW, Rasper DM, Kalchman M4 Koide HB, Graham RK, Bromm M, Kazemi-Esfajani P, Thomberry NA Vaiiiancwrt JP, Hayden MR: Nat G e m 13 : 44249, 1996.
Gottiieb B, Lehvaslaiho H, BeiteI LI& Lumbroso R, P i k y L, Trifiro M : Ntrcleic Acids Res 26: 234-238, 1998.
Green H: Cell74: 955-956, 1993.
Halchmi S, Marden E, Martin G, MacKay H, Abbondarua C, Brown M: Science 264: 1455- 1458, 1994.
Heery DM, Kalkoven E, Hoare S, Parker MG: Nahrre 387: 733-736, 1997.
Eolmberg M., Duyckaerts C, Durr A, Cancel G, Goudhkel-An ï, Damier P, Faucheux B, Trottier Y, Hirsch EC, Agid Y, Brice A: Hum MolGenet 7: 913-918, 1998.
Horiein AJ, Naar AM, Heinzel T, Torchia J, Gloss B, Kurokawa & Ryan 4 Karnei Y, Soderstrom M, Glass CK, RoSenfeld MG: Nmre 377: 397-404, 1995.
Homitz KB, Jackson TA, Bah DL, Richer IK, Takimoto GS, Tung L: Mol Ei#Iocrir~of 10: 1167-1 177, 1996.
Huntington's Diseut Coliabomtive Resarch Gmup (EDCRG): Cell 72: 97 l-983,1993.
Igarashi S, Koide R, Shohata T, Yamada M, Hayaski Y, Takano H, Date H, Oyake M, Sato T, Sato 4 Egawa S, h c h i T, Tanaka H, Nakano R, Tanaka K, Hozymi 1, huzuka T, Takahashi H, Tsuji S : Nat Gewr 1 8: 1 1 1 - 1 17, 1 998.
Ikeda H, Yamagushi M, Sugar S, Aze Y, Narurniya S, Kakinika A: Nat Gerret 13 : 196-202, 1996.
Drcda Y, Shen W-H, Ingraham Parker KL: Mol Eidocn'mi 8: 654-662, 1994.
h i n e RA, Yu MC, Ross RK, Coetzee GA: Cancer Res 55 : 193 7- 1 940, 1995.
bsemann ï, Prince Rq Tupvood JD, Green S: Biochimie 75: 25 1-256, 1993.
Jouo N, Icarf JY, Tran D: Rec Prog Horm Res 33 : 1 17- 160, 1977.
Kang H-Y, Yeh S, Fujimoto N, Chang C: JBiol C h 274: 8570-8576, 1999.
Kawaguchi Y, Okamoto T, Taniwalà M., Aizawa M, houe m, Katayama S, Kawakarni H, Nakamura S, Nishimura M, Akigushi 1 ef al. : Nat. Gemt. 8: 22 1-228, 1994.
Kazcmi-Esfa jani P, Trifiro Pinsky L: H1m Mol Genet 5: 523-527, 1995.
Kennedy WR, Alter M, Sung JH: Nei~rofogy 18: 671-680, 1968.
Kim T-W, T d RE: Neurort 221: 657-659, 1998.
Kiement IA, Skiner Pl, Kaytor MD, Yi H, Hench SM, Clark HB, Zoghbi HY, Orr HT: Cell 95: 41-53, 1998.
Kliewer SA, Umesono K, Mangelsdorf DJ, Evans EM: Nature 3 55 : 446449, 1 99ta.
Klimer SA, Umesono K, Noonan DJ, Heyman Evans RM: Nature 3 58: 77 1-774, 1992b.
K d e MR, Talbot WS, Segraves W 4 Bender hdT, Cherbas P, Hogness DS : Cell67: 59-77, 1991.
Koide R, Ikeuchi T, Onodera O, Tanaka H, Igarashi S, Endo K., Takahashi Y Kondo R, Ishikawa A, Hayashi T et al. : Nat Genet 6: 9-13, 1994.
Koivisto P, Kononen J, Paimberg C, T'ammela T, Hyytinen E, Isola J, Traprnan J, Cleutjens K, Noordzij 4 Visakorpi T, Kallioniemi OP: C'cer Res 57: 3 14-3 19, 1997.
Kuiper CGJM, Brinkmann AO: Mol Ceil Eirdocrinoi 100: 103- 107, 1994.
Kuiper CGJM, Faber PW, van Rooij HU, van der Korput JAGM, Ris-Stalpers C, Kiaassen P, Trapman J, Brinkmann AO: JMol fisdocriml 2: R1-R4, 1989.
Kumar V, Chambon P: Cell55: 145-1 56, 1988.
Kurokawa R, Soderstrom M, Horiein A, Halachmi S, Brown M., Rosdeld M, G l a s C: Nuîure 377: 45 1-454, 1995.
La Spada AR, Wdson Eh.i Lubahn DB, Harding AE, Fischbeck KH: Naîure 352: 77-79, 1991.
Langley E, Kernppainen J& Wdson EM: J Bioi Chem 273 : 92- 10 1, 1998.
Langley E, Zhou Z - x Wilson EM: JBiol C h 270: 29983-29990, 1995.
Lee JW, Choi HS, Gyuris J, Brent R, Moore DD: Mol Eitdocri1zol9: 243-254, 1995.
Lee Sy Choi HS, Song M-R, Lee M-O, Lee LW: Mol Eidocirtd 12: 1 184- 1 192, 1998.
Lobaccaro J-M, Lumbroso S, Belon C, Galtier-Dereure F, B ~ g e r J, Lesimple T, Heron J-C Pujol H, SultanC: .Vat Genet 5: 109-1 10, 1993.
Lubahn DB, Moyer IS, Golding TS, Couse IF, Korach KS, Srnithies O: Proc N i l Acad Sci USA 90: 11162-11166, 1993.
Luisi BF, Zu WX, Otwinowski Z, Freedman LP, Yamamoto I(S Sigler PB: Nahre 352: 497-505, 199 1.
Lumbroso R, Beitel & Vasihou DM, Trifiro MA, Pinsls, L: Hum Genet 10 1 : 43-46, 1997.
Lunkes A, Mandel J-L: Nat Med 3 : 120 1- 1202, 1997.
MacLean HE, Choi WI; Rekaris G, Warne GL, Zajac JD: J Clin EdàcrinolMeiab 80: 508- 516,1995.
MacLtrin HE, Warne GL, Zajac JD: Mol Cefi Endocrimi 1 12: 133- 142, 1995.
MacLean HE, Warne GL, Zajac JD: J Nertrosci Sci 135: 149-1 57, 1996.
Mangdadorf DJ, Thurne1 C, Beato M, Herrlich P, Schütz G, Umesono K., Blumberg B, Kastrier P, Mark M, Chambon P, Evans RM: Cell83 : 835-839, 1995,
Marcdi M, T i i WD, Zoppi S, Griflin JE, Wilson JD, McPhaul MJ: JCIin Endocriml Metab 73: 318-325, 1991.
Marks MS, Halenbeck PL, Nagata T, Segars JY Appela E, Nkodem VM, Ozato K: M O J Il: 1419-1435,1992.
Martindaie D, Hackam 4 Wiecmrek 4 EUerby L, Wellington C, McCutcheon Y Singaraja q Kazerni-Esfiani P, Devon R Kim SU, Bredesen D, Tufàro F, Hayden M R Nat Chet 18: 150-154, 1998.
McPhaul MJ, Marcelli M, Zoppi S, Griflin JE, Wlson JD: J Clin W'ol Metab 76: 17- 23, 1993.
Mhatre A, Trifiro Kaufinan M, Kazemi EP, Figlewicz D, Rouleau G, Pinsky L: Nat Gemr 5: 184-188, 1993.
Miyamoto H, Yeh S, Wdding G, Chang C: Proc Nat1 AcadSci US4 95: 7379-7384, 1998.
Miyashita T, Nagao Y Ohmi Y Yanagisawa Y Okarnura-Oho Y, Yamada M: Biochem Biophys Res Comm 249: 96- 102, 1998.
Miyashita T, Okamura-Oho Y, Mito Y, Nagafuchi S, Yamada M: J Biof Chem 272 : 2923 8- 29242, 1997.
Mizokami A, Yeh S-Y, Chang C: Mo~firdocrinol 8: 77-88, 1994.
Moilanen A, Rouleau N, Ikonen T, Palvimo JJ, Janne OA: FEBSLen 4 12: 355-358, 1997.
Moulder KL, Onodera O, Burke iR, Strittmatter WJ, Johnson EM Jr: JNéurosci 9: 705-71 5, 1999.
Nagafwhi S, Yanagisawa H, Sato Y Shirayama T, Ohsaki E, Bundo M, Takeda T, Tadokoro K, Kondo 1, Murayama N et al. : Nat. Genet- 6 : 14- 1 8, 1994.
Nemoto T, Ohara-Nemoto Y, Shimazaki S, Ota M: JSteroidBiuchem MolBid 50: 225-233, 1994.
Ogmyzko W, Schlitz RL, Russanova V, Howard BH, Nakatiani Y: Cell87: 953-960, 1996.
Onate SA, Tsai SY, Tsai M-J, O'Mdey BW: Science 270: 1354-1 3 57, 1995.
Ordway JM, Tallaksen-Greene S, Gutekunst CA, Bernstein EM, Cedey J 4 HW, Dure LS 4th Lindsey Et, Hersch SM, Jope RS, Aibin RL, Detloff PJ: Cell9 1: 753- 763, 1997.
O- HT, Chung MY, Banfi S, Kwiatkowski TJ, Senadio 4 Beaudet AL, McCaii AE, Duvick L q Ranum LP, Zoghbi HY : Nat Getet 4: 22 1-226, 1993.
Parktr MG, White R: Nilire Stnict Biol3 : 1 13-1 15, 19%.
Perez MK, Paulson HL, Pendse SJ, Saionz S, Bonini NM, Pittman RN: J Cell Biol 143: 1457-1470, 1998.
Ptmb MF, Johnson T, Suzuki M, Finch J ï : Proc NatIAcadSci USA 91: 5355-5358, 1994.
Pinsky L, Beitel LK, Gottlieb B, Yong EL, Trifiro MA: In: FerM& and Reprodr~ctive Medicine, Kernpers RD, Cohen J, Haney AF, Younger JB (Eds), Elsevier Science B.V., pp. 783-797, 1998.
Pins@ L, Beitel LK, Kazerni-Esfârjani P, Lumbroso R, Vasihou DM, Shkohy D, Abduilah Gottlieb B, Trifiro M . In: Sex Differentiation: Clinical and Biological Aspects
Frot~tiers in Endocritiology 20 : 95- 1 1 4, 1 996.
Pinsky L, Triiiro MA, Beitel LK K a h a n M: Molecular genetics of androgen insensitivity syndromes in humans. In: Molead . Genetics of Sex Detenninatiotz pp. 34 1 -365, 1994.
Poujol N, Lobaccaro J-A4, Chiche L, Lurnbroso S, Sultan C: Mol Cell Endmnnol 130: 43- 51, 1997.
Pratt W B : JBiol Chem 268: 21455-2 1458, 1993.
Puist SM, Nechiporuk 4 Nechiporuk T, Gispert S, Chen XN, Lopes-Cendes L, Peralrnan S, Starkman S, Orozco-Diaz G, Lunkes A, DeJong P, Rouleau GA, Auburger G, Korenberg ni Figueroa C, Sahba S: Nat Gerret 14: 269-276, 1996.
Quigiey CA, De Beiiis A, Marschke KB, El-Awady MK, WiIson EM, French FS: Eitdmrine Rev 16: 271-321, 1995.
Quigley CA, Friedman KJ, Johnson 4 Meniere RG, Silvemian LM, Lubahn DB, Brown Wilson EM, Wiuard HF, French FS: JClin Embcn'nol Me@ 74: 927-933, 1992.
Renaud J-P, Roche! N, Ruff M, Vivat V, Chambon P, Gronemeyer H, Moras D: Nohrre 378: 681-697, 1995.
Roôrigua R, WEigel NL, O'Malley BW, Schrader W: MolWmrimf 4: 1782-1790, 1990.
Ruijter E, van de Kaa C, Miller G, Ruiter D, Debruyne F, Schalken J: fidoçrine Rev 20: 22- 45, 1999.
RuPevdd de Winter JA, Trapman J, Vemey M, Mulder E, Zegers ND and van der Kwast M: J. Hisi. Cytochem.39: 927-936, 199 1.
RundIett SE, Wu XP, Miesfeld RL: Mol fid'rimI 4: 708-7 14, IWO.
Rutgers JL, Scuiiy RE: hr J Gp~ecoI PathoI 10: 126- 144, 199 1.
Sanpei Y Takano H, Igarashi S, Sato Tl Oyake M, Sasaki H, Wakisaka 4 Tashiro K, Ishida Y, Ikeuchi T, Koide Saito M, Sato A, Tanaka T, Hanyu S, Takiyama Y, Nishizawa Ad, Shimizu N, Nomura Y, Segawa hd, Iwabuchi K, Eguchi I, Tanaka H, Takahashi FI, Tsuji S: Nat Genet 14: 277-284, 1996.
Sasco AJ, Lowenfels AB, Pasker de Jong P: Ihter. J Cmcer 53(4): 538-549.
Sar M, Stumpf WE: Scierice 1 W(4298): 77-79, 1977.
Saudou F, Finkbeiier S, Devys D, Greenberg ME: Ce11 95,55066, 1998.
Schwabe JW, Chapman L, Finch JT, Rhodes D: Ceif 75: 567-578, 19%.
Shen W-H, Moore CCD, Ikeda Y, Parker KL, Ingraham HA: Cell77: 65 1-66 1, 1994.
Shikama N, Lyon J, La Thangue NB: Trends CeIf W 7 : 230-236, 1997.
Siiteri PK, Wilson JD: J Clin Eradocrinol Metab 38: 1 13- 125, 1974.
Sinclair AH, Berta P, Palmer MS, Hawkins JR, WfEths BL, Smith MJ, Foster JW, Frischauf A-M, Lovel-Badge Goodfeiiow PN: Nature 346: 240-244, 1990.
Skinner PJ, Koshy BT, Cummings CJ, Kiement IA, Heün Y Servadio 4 Zoghbi HY, Orr HT: Nature 389: 97 1-974, 1997.
Sleddens HF, Oostra BA, Brinkmann AO, Trapman J : Hum Mol Genet 2: 493-494, 1993.
Smith EP, Boyd J, Frank GR Takahashi H, Cohen RM, Specker B, Wiarns TC, Lubahn DB, Korach KS: N h g l JMed 33 1 : 1088-1089,1994.
Sobue G, Doyu M, Kachi T, Yasuda T, Mukai E, Kumagai T, Mitsuma T : JNetrrosci Sci 1 17: 74-78, 1993.
Sobue G, Hashinime Y, Mukai E, Wirayama M, Mitsurna T, TakahaShi A: Br& 112: 209- 232, 1989.
Sobue G: Nagoya JMed Sci 58: 95- 106, 1995.
Spencer TE, Jenster G, Burcin MM, Allis CD, Zhou J, Mizzen CA McKenna NJ, Onate SA, Tsai SY, Tsai MJ, On-Idey BW: Nahrre 389: 194- 198, 1997.
Stenoien DL, Cummîngs CJ, Adams HP, Mancini MG, Patei K, DeMartirno GN, Marceüi M, Weigel NL, Mancini MA: Hum Mol Genet 8: 73 1-74 1, 1999.
Stunnenberg HG: Bioessqys 15: 309-3 15, 1993.
Sultan C, Lumbroso q Poujol N, Belon C, Boudon C, Lobaccaro J-M: J Sterod Biochem Mol Bi01 46: 5 19-530, 1993.
T'iey WD, Marcelh M, McPhaul M.: J Biol Chem 265: 13 776- 1378 1, 1 990.
Torchia J, Glass C, RosenfeId MG: Czm Opin Cell Biol 1 q3): 373-3 83.
Torchia J, Rose DW, hostroza J, Kamei Y, Westin S, Glas CK, Roserifeld MG: Naiwe 387: 677-684, 1997.
Trüiro MA, Kazemi-Esfarjani P, Pinsky L: Tremk Endocrinol Metab 5: 4 16-42 1, 1994.
Triiïro MA, Gottlieb B, Pinsky L, Kaufnian M, Prior L, BeIsham DD, Wrogemann K, Brown CS, Wdard HF, Trapman J : Mol Cell Ertcoctiinol 75: 37-47, 199 1.
Tsai M-J, OMalley BW: A m Rev Biachem 63 : 45 1-486, 1986.
Vddscholte J, Berrevoets C q Brinkmann AO, Grootegoed J q Mulder E: Biuchemistry 3 1 : 239302399, 1992.
Vddscholte J, Ris-Stalpers C, Kuiper GGJM, Jenster G, Berrewoets C, Claassen E, van Rooij HCJ, Trapman J, Brinkmann AO, Mulder E: Biochem Biophys Res Commun 173: 534-540, 1990.
Voegd JJ, Heine MJS, Tini M, Vivat V, Chambon P, Gronemeyer H: -0 J 17: 507.5 19, 1998.
Wpgntr RL, Apriletti JW, McGrath ME, West BL, Baxter JD, Fletterick RI: Nature 378: 690-697, 1995.
Wang H, Peters GA Zeng X, Tang M, Ip W, Khan SA: JBiol Chem 270: 23322-23329, 1995.
Wamck JM, Paulson HL, Gray-Board GL, Bui QT, Fischbeck KH, Pittman RN, Bonini NM: Ce1193 : 939-949, 1998.
Wellington CL, EUerby LM, Hackam AS, Margoüs RL, Trifiro Singaraja q McCutcheon Y Salvesen GS, Propp SS, Bromm M, Rowland KJ, Zhang T, Rasper D, Roy S. Thronbeny N, PinsiCy L, Kakinika A, ROSS C 4 Nicholson DW, Bredesen DE, Haydn MR: J B i d C h 273: 9158-9167, 1998.
Williams SP, Sigler PB: Nature 393 : 392-396, 1998.
Wilson CM, McPhaul MJ: P m Na11 A d S c i USA 91 : 12341238, 1994.
Wooster Ft, Mangion J, Eeles Smith S, Dowsett M, Averill D, Barrett-Lee P, Easton DF, Ponder BAJ, Stratton MR: Nat Genet 2: 132-134, 1992.
Wuid J-M., Bourguet W. Renaud JP, Vivat V, Chambon P, Moras D, Gronemeyer H: Nature Stnrct B o l 3 : 87-94, 1996.
Yao T-P, Ku G, Zhou N, SwUy S Livingston DM: Proc NaIl AcadSci USA 93: 10626- 10631, 1996.
Yu VC, Delsert C, Andersen B, Hoiioway JM, Devaty OV, Near AhA, Kim SY, Boutin N Glass CK, Rosenfeld MG: Cell67: 1 25 1 - 1266, 1 99 1.
Zhang XK, Hofnmann B, Tran PB, GraupnerG, Pfahl M: Nature 3 55 : 44 1-446, 1 992.
Zhou 2, Sar M, Simental JA, Lane MV, Wdson EM: JBiol C h : 269: 13 1 15-1 3 123, 1994.
1. To define the molecular abnormalities of a novel androgen receptor missense mutation
that causes mild androgen insensitivity. Using a yeast-two hybrid system, dimerization
between mutant receptors will be tested and quantified and compared with wild-type
receptor dimerization in order to define what molecular deficiencies are associated with
this particularly novel mutation. Such analysis will contribute to the ever-growing
structure tiinction map defined by the multiple mutations underlying androgen
resistance syndromes.
2. To help define the molecular composition of the polyGln AR-induced intracellular
aggregates and to test whether parent protein proteolysis is a prerequisite for such
aggregation. This will be done by fluorescently tagging both the N-terminal and C-terminal
of the AR, independently and doubly, concurrent with immunocytochemistry using two
different AR antibodies. Western analysis of the transfected ce11 lysates using an AR
antibody, a polyûln-expanded tract antibody and a GFP/BFP antibody will also help define
the contents of these intracellular aggregates. Finally, caspase inhibitors will be added to
assess the role of caspases in aggregate formation.
3. To test a hypothesis that androgen-estrogen physiological interactions might
reflect direct AR-ERa interaction. Using both a yeast two-hybrid system and a mammalian
two hybrid system, several domains of the AR and ERa will be tested to document
possible AR-ERa interactions. As well, DNA cotransfections using AR and ER expression
vectors and a dual reporter gene construct will be done to mess possible functional
interactions between AR and ERa.
CONTRIBUTIONS OF AUTHORS
Chapter 1 The Sinna~ort laboratorv
Fand JO Ghadessy Joyce Lim Thtin G o Tut Chee Keong Choo Dr. Eu Lcong Yong
1) Screened AR gene o f subjects and controls by SSCP. 2) Cloned mutant and WT AR into mammalian 2-hybnd conamcts 3) Performed irnmunoblot analyses 4) Performed DNA mobility gel shifi assays 5) Wrote and edited manuscript
Abdullah A. R Abdullah 1) Recreated the 886 mutation and subcloned it into an expression plasmid 2) Performed the transactivational studies of the mutant and wild-type receptors
Valerie Panet-Raymond 1) Cloned mutant and full-length AR (and fragments thereof) into yeast-two hybrid vectors. 2) Performed al1 yeast-two hybrid expenments 3) Analyzed and represented yeast-two hybrid data 4) Participated in writing manuscript
Rose Lumbroso 1) Performed the Scatchard analyses and chase experiments 2) Sequenced constmcts to verify cloning accuracy
Bruce Gottlieb 1) Performed DNA binding studies
Leonard Pinsky 1) My s u p e ~ s o r 2) Participated in vniting and editing manuscript
Mark A. Trifiro 1) My co-supervisor 2) Participated in writing and editing the manuscript
Chapter 11
Valerie Panet-Rnymond 1) Performed dl transient transfections in both ce11 lines, COS-1 and NSC-34 and related ceil culture 2) Cloned aii fluorescently labeled-constructs including WT and polyGln expanded GFP-AR, BFP-AR and GR>-AR-BFP 3) Perfomed al1 ceU staining and immunocytochemiary experiments and photographing of cells 4) Penormed ail irnmunoblotting 5) Perfonned confocal microscopy experiments 6) Anaiyzed al1 resulting data and constructed figures 7) Wrote and edited manuscnpt
Dr. Lenore K. Beitel 1) Participated in editing manuscript and supewision
Dr. Hyman Schipper 1) Provided confocal microscopy
Dr. Micheal Tymianski 1) Provided confocal microscopy
Dr. Leonard Pinsky 1) My supervisor 2) Participated in editing manuscnpt
Dr. Mark A. Trifiro 1) My CO-supervisor 2) Participated in editing manuscript
Chapter III Valerie Panet-Ray mond 1) Perfomed al1 transient DNA transfections 2) Cloned al1 yeast-two hybnd constructs used 3) Performed al1 yeast-two hybrid experiments 4) Cloned al1 mammalian two-hybnd constructs used 5) Perfonned al1 mamalian two-hybrid experiments 6) Constructed pSEAP2-ERE and pSeAP2-ERE-MMTV-GH vectors 7) Performed both AR and ER transactivation studies using GH and SeAP reporter genes 8) Performed ail data analysis and figure construction 9) Wrote and edited manuscript
Dr. Leonard Pinsw 1) My supervisor 2) Participated in editing manuscript
Dr. Mark A. Trifiro 1 ) My co-supervisor 2) Participated in editing manuscnpt
Chapter 1
INTRODUCTION TO CHAPTER 1
The recent description of steroid receptor coactivators as important trancnptional
regdators has prompted several investigations into the analysis of specific steroid receptor
mutations that may iack the ability to interact with such coregulators.
Several missense mutations of the AR associated with androgen insensitivity could
potentially have abnormal coactivator interactions as a predorninant property underlying
AR dysfbnction. Characteriration of such mutants would help in developing fiirther the
structure-fünction map of the AR and contribute to the understanding of the molecular
mechanisrn of action of AR. Indeed, such information rnay be reIevant and applicable to
other steroid receptors as well.
In the following paper, we describe a missense mutation in the AR-LBD that is
novel and does not confer abnomai Iigand binding but has abnormal transactivational
properties. The missense mutation was found to intei.fere with coactivator TIF2
interactions as well as DNA binding and homodimerization.
Oiigospermic infertility associated with andmgen receptor mutation that reduces DNA-binding, and disrupts interactions between domains and with the coactivator, TIF2
Farid J. Ghadessy*, Joyce Lim*, Abdullah A.R. AbduUah*, Valerie Panet-Raymond, Chee Keong Choo, Rose Lurnbroso, Thein G. Tut, Bruce Gottlieb, Leonard Pinsky, Mark A. Trifiro, Eu Leong Yong
* These authors contnbuted equaily
Department of Obstetrics and Gynaecology (FJG, JL, CKC, TGT, ELY) National University of Singapore Republic of Singapore 1 19074
Lady Davis Institute for Medical Research, Sir Mortimer B. Davis-Jewish General Hospital (AAR4 VP, RL, MAT, BG, LP) and Departrnents of Biology (AARA, LP), Medicine (MAT, LP), Pediatrics (LP), Human Genetics (LP), McGiIl University, Montreal, Quebec, Canada H3T 1E2
Proofs and al1 correspondence to: E. L. Yong, MD, PhD Department of Obstetrics and Gynaecology National University Hospital Lower Kent Ridge Road Republic of Singapore 1 19074 Fax: (65) 779 4753 e-mail : [email protected]. sg
Abstract
Structural changes in the androgen receptor (AR) are one of the causes of defective
spematogenesis. We screened the AR gene of 173 idenile males with impaired
spermatogenesis and identified three of them, umelated, who each had a single adenine-to-
guanine transition that changed codon 886 in exon 8 fiom methionine to valine. This
mutation was significantly associated with the severely oligospennic phenotype and was
not detected in 400 control AR aileles. Despite the location of this substitution in the LBD
of the AR, neither the genital skin fibroblasts of the subjects nor transfected ceIl types
expressing the mutant receptor had any androgen-binding abnormality. However, the
mutant receptor had a consistently (-50%) reduced capacity to transactivate each of two
different androgen-inducible reporter genes in three different ce11 lines. Deficient
transactivation correlated with reduced binding of mutant androgen-receptor complexes to
androgen response elernents. Co-expression of AR domain fiagrnents in mammaiian and
yeast two-hybrid studies suggest that the mutation dismpts interactions of the LBD with
another LBD, with the N-terminal transactivation domain, and with the transcriptional
interrnediary factor, TIF2. These data, for the first time, suggest that a functional element
centered around MS86 has a role, not for ligand binding, but for inter-domain and
coactivator interactions cutminating in the formation of a normal transcription cornplex.
Introduction
Infertility afEects about 10 to 15% of al1 couples (1). Spennatogenesis is at fault in
about half of them, but its cause is offen covert. Androgens are required for normal
spermatogenesis; however, most infertile males with impaired spermatogenesis have
normal serum androgen levels. Therefore, attention has tumed to the androgen-response
apparatus, particularly the androgen receptor (AR). Mutation of the X-linked androgen
receptor gene causes a wide range of clinical androgen insensitivity: complete, when the
extemal genitalia are female; partial, when they are sufficiently ambiguous to require
corrective surgery; rnild, when they are phenotypically male. Three surveys of males with
idiopathic Uifertility have yielded widely disparate frequencies of androgen-binding
abnormalities in cultured genital skin fibroblasts (2-4). Genetic defects of the AR that
cause rnild androgen insensitivity with impaired or preserved spermatogenesis (5, 6, 7) are
of particular interest because they may illuminate the fine structure-function attributes of
the AR that permit differential regulation of androgen-inducible genes. To this end, we
screened a large group of infertile males with defective spermatogenesis for abnormalities
in the coding segments of their AR genes. We found three unrelated subjects with the
sarne missense substitution in the C-terminal portion of the ligand-binding domain (LBD)
of the AR. One shaves infrequently; another has low-grade but persistent postpubertal
gynecomastia. Unexpectedly, this novel mutation does not affect the ligand-binding
characteristics of the AR, rather, it reduces the transactivational cornpetence of the AR by
impairing interactions between the receptor domains, binding to androgen response
elements (ARE) and fbnction of the steroid receptor coactivator, TIF2.
Materials and Methods
The Study Population
Our patients presented to infertility clinics because they could not conceive for at
least two years. None was referred with the question of androgen insensitivity. Two or
more spem samples were collected at least 3 months apart atter 3- to 5-day abstinence,
and assessed according to WHO criteria (8). Patients with abnormal sperm analyses due to
obstruction of the genital tract, hypopituitarism, hyperprolactinaemia or markedly raised
FSH were excluded. Of the remaining 173 patients, 33 had azoospermia, 84 were severely
oligosperrnic (CS million/rnl), 36 were moderately oligospermic (5-20 million/ml), and 20
had only abnormal spem motility. Control male subjects (n=100) of proven fertility, no
previous infertility history or treatment, and without any genetic disease were recruited
fiom the contraceptive chic. A further 200 (1 10 males and 95 fernales) healthy subjects
were also screened to determine whether AR allelic variations exist in the general
population.
The Mutant Subjects
CML presented at 3 1 years d e r four years of Ulfertility. He shaves infrequently,
about once a week. Right-sided cryptorchidism was corrected at 7 years. Testicular
volume was 10 ml bilaterally and his sperm count about 0.5 million~rnl. KLH presented at
40 years &er 7 years of infertility. He had Tanner grade 2 persistent postpubertal
gynecomastia. Testicular volume was 6 ml bilaterally with sperm counts around 0.3
millionhnl. Secondary sexual development was othenvise normal in both subject S. The
third subject, EHS, had a sperm count of 0.7 millionhnl. Ail three patients had normal
male karyotypes and normal serurn levels of gonadotropins, androgens, sex hormone-
binding globulin, prolactin and estradiol.
Mutational screening with single strand conformation polymorphism (SSCP)
DNA was extracted from peripheral blood. Coding segments and flanking intronic
sequences of exons 2 to 8 of al1 173 patients were examined by SSCP(9). DNA fiapents
that exhibited differential mobilities were sequenced. Exons 2 to 8 of the patients (CML,
KLH, EHS) with mutant alleles were re-examined by SSCP to exclude any CO-existing
mutations that might have been missed on initial screening. Most of exon 1 of the mutant
subjects was screened by SSCP. The two highly polymorphic tinucleotide (CAG and
GGC) tracts were directly sequenced.
In vivo androgen-binding properties of AR
Fibroblast cultures were obtained fiom skin biopsies of subjects and fiom nomal
controls. The androgen-binding properties of the AR were detennined according to
standard techniques (1 1). Finastende (IO-' M) was added to the cultures to inhibit
endogenous Sa-reductase activity that might degrade androgens. Scatchard analysis was
perfiormed by plotting specifically bound hormone versus the BoundFree ratio; and &,
determined fiom the negative dope of the graph. B- was obtained fiom the intercept of
the line on the X-axis. Thennolability was examined by comparing the binding properties
at 32°C or 3PC, and 42°C. A reduction of B, of more than 40% defined thennolability.
Chase experiments: to determine &, the rate constant of dissociation, ceIl monolayers
were pre-incubated with the radiolabelled androgens and the proportion of labelled
homone still specifically bound d e r exposure to excess unlabelled hormone was
determined at defined time intervals.
Plasmids
M886V AR: The mutation in Our subjects was recreated in a cDNA fragment by
site-directed mutagenesis (6) and then substituted into the homologous section of an AR
expression vector, ;>SVhARo. TAD and LBD AR fragments were fomied by excising
ftagments bounded by the unique restriction sites KpnYEcoRI and NheYKpnI respectively
from AR cDNA Marnmalian two-hybrid (Clontech, Pa10 Alto, CA): The pGAL4DBD-
LBD was prepared by amplifjing cDNA encoding the AR LBD (exons 4 to 8) and ligating
in-fiame into the SmaVHindIII site of pM containing GAL4 DNA-binding domain. The
pVP16AD-ARTAD was made by restncting pSVhARo with EagI and HindIII to release
the fragment encoding arnino acids 14-565 of the AR. The 5' end of this fragment was
then ligated to a synthetic linker encoding the first 13 amino acids of ARTAD and the
resultant fiagment encoding the entire ARTAD was cloned in frame with the VP16
activation domain using pVP16 vector. Plasmid pVP16AD-TIF2 was constmcted by a
double digest of pSGS-TIF2 with HindIII/XbaI followed by ligating in-frame to pVP 16.
The (1 lm),-ElbTATA-Luc repofier vector was obtained by amplifjing the five GAL4
binding sites and the adenovirus El b minimal promoter of pG5CAT and ligating upstrearn
of the luciferase gene in pGL-basic (Promega, Madison, WI) vector. Yeast two-hybrid
(Clontech, Pa10 Alto, CA): WT or M886V LBD fiagments including (amino acids 502 to
919) or excluding (amino acids 659-919) the DBD were generated by double digesting the
fiill-length AR vector with KpnVBamHl or PvuYBamM, respectively. The LBD
fiagrnents were cloned into vectors pAS2-1 and pACT2, which contain the GAL4 DNA-
binding domain (amino acids 1 to 147) and the GAL4 transactivation domain (arnino acids
768-88 1) respectively, to generate GAL4-AR hybrid constructs. Full length AR constmcts
were made by excising the AR from pSVhA&-BHEX using SmaYBamHI and then
ligating the resulting fiagment to pAS2-1 and pACT2 that had been similarly digested. Ail
constructs were sequenced to confirm the fidelity of the enzymatic manipulations.
Mammalian cell culture and transient transfection
Mutant and WT plasmids were transfected into COS-7, CV-1 or HeLa cells using
lipofection technique (12). pCMV-PGal was used to assess transfection efficiency. In
some replicates, radiolabelled Mibolerone (MB) was added to the culture medium and
specific MB-binding activity determined (6). Transactivation activity was measured in
relative light units (RLU) and normalized to protein content and transfection efficiency.
Immunoblot analyses
Immunoblot analyses were used to study the effect of the mutation on AR pmtein
production. The rabbit polyclonal antibody, PG-2 1, which recognizes the first 2 1 N-
terminal amino acids of the human AR, was used to detect AR protein (1 1). A mouse
monoclonal antibody SC5 1 0 (Santa Cruz Biotech, CA) was used to locate GAL4DBD
ftsion proteins. Protein-antibody complexes were subsequently visualized by enhanced
cherniluminescence (1 2).
DNA mobility gel shift assays and quantitation of androgen-AR complexes bound to
androgen response elements
DNA mobility gel shift assays: COS-7 cells were transfected with WT or mutant
AR plasmids, harvested in extraction buffer (20 mM HEPES, pH 7.9, 20% glycerol, LOO
mM KCI, 0.2 rnM EDTA containing protease inhibitors (phenylmethylsulfonyIfluoride,
leupeptin, aprotinin), lysed by three fieezdthaw cycles to release AR. A consensus
synthetic ARE (5'-ctagaagtctggtacagggtflctt~gca-3'1, served as specific-binding DNA
(13). Two non-specific competitor oligonucleotides were used: the first incorporates a
Wï estrogen response element fiom the promoter of the Xenopus vitellogenin A2 gene,
5'-gtccaaagtcaggtcacagtgacctgatcaaagtt-3 the second a transcription factor Oct2A , 5'-
gtacggagtatccagctccgtagcatgcaaatcctctgg-3' response element . T4 polynucleotide kinase
and [ g-32P] dATP were used to label the oligonucleotides. Binding reactions containing 5
pg ce11 extract, 60 pg BSA, 10% glycerol, 2 mM DTT, 2 pg poly(d1-dC), 0.2 m M EDTA
and 20 mM HEPES, pH 7.9, in a total volume of 20 pl were preincubated on ice for 20
min. Some reactions contained a 25-fold excess of unlabelled oligonucleotide as
competitor DNA Labeled oligonucleotides ( 4 . 3 ng, 10,000 cpm) were added and
incubation carried out a further 20 min at room temperature. AR-DNA complexes were
resolved by electrophoresis on 5% PAGE gels and autoradiographed. Quantitation of
androgen-AR complexes bound to ARES (14,15). AR expressed in COS-1 cells was
exposed to 3 n M [3HJMB, harvested and diquots were measured for radioactivity. Each
assay mixture contained 10 pg of poly (dl-dC), a sample of charcoal-treated supernate
with 50,000 dpm of MB-receptor complexes, and binding buffer to a final volume of 500
pl. Then, we added 150 pmol of 3'-biotinylated double-stranded oligonucleotide sequence
of the synt hetic ARE (5'-ctagaagtctggtacagggtgttctttttgca-biotin or MMTV-ARE (5'-
tatggttacaaactgttcttaaaa-biotin) and 50 pl of streptavidin-agarose beads before continuing
incubation for 2 h. The beads were collected by centrifugation, washed thrice, and bound
radioactivity measured by liquid scintillation. Assay mixtures lacking an ARE bound
negligible amounts of radioactivity. Each data point was the mean of duplicate
experiments calculated as a percentage of binding observed with wild-type AR.
Yeast two-hybrid assay for LBD-LBD interactions
Hybrid GAL4-AR proteins were expressed in a S. Cerevisiae Y190 containing
integrated GAL4 binding sites upstream of the U A S G ~ ~ - I ~ C Z reporier gene. Interaction
between hybrid proteins results in a transcriptionally active complex inducing B- galactosidase activity. Yeast transformation was perfonned using the [ithium acetate
method according to Clontech protocols. Yeast were grown in standard YEDP medium or
an appropriate selective medium in the presence or absence of 1 ph4 Methyltrienolone
(MT). Yeast were harvested and liquid P-galactosidase assays were pefiormed using
ONPG as a substrate according to standard protocols with the following modification: the
yeast were pemeabilised by incubation in a 0.2% sodium lauryl sarcosinate Z buffer
solution (60nM Na2HP04, 40nM NaH2P04, lOmM KCI, 1mM MgzSO*, 50mM B- mercaptoethanol) instead of fieeze-thawing before the addition of ONPG.
Results
Mutations detected in the AR of three unrelated oligospemic patients
Thee umelated patients? CML? KLH and EHS, showed differential migration of
exon 8 PCR fragments when screened by SSCP. Two subjects, CML and KLH, inherited
the mutation nom their mothen who were heterozygous carriers Vig. 1A). The suspected
fiagments were sequenced and al1 3 patients had the same mutation in exon 8 involving
aminoacid 886, an A to G transition resulting in substitution of vaine for methionine (Fig.
1B). This mutation resulted in the creation of a new Bbr P 1 restriction site &ig. 1C). Ail
three subjects were near-azoospermie with repeated sperm counts of less than 1 million
per ml. CML had 23 codons in each exon 1 polymorphic trinucleotide repeat tract, while
KLH had 21 glutamine and 24 glycine codons, confinning that the two subjects were not
genetically related. No coexisting mutations were detected in AR exons 1 to 8 of the 3
patients on SSCP analyses, nor in exon 8 PCR fiagments of 400 AR alleles from healthy
controls. The presence of 3 mutations in our 173 infertile patients and none in 400 control
alleles makes it unlikely that M886V exists in the general population (Fisher's Exact Test,
p=0.027). On the other hand, since the mutations occur only in a subset of infertile males
with severe oligospermia (n=84), it is more likely @=0.005) that M886V is significantly
associated with the phenotype of severe oligospermia.
M886V i ~ t the LBD had 110 effect on androgen-bindi~~g churacteristics
The Kd for dihydrotestosterone (DHT) was 0.31 and 0.52 nM, and a B, of 36 and 37
h o 1 DHT/rng protein for CML and KLH, respectively at 37°C for fibroblast monolayers
cultured fiorn scrotal skin biopsies (Normal: &, 0.3-0.6 n . ; B,, 26-43 hoVmg
protein). Similarly there were no significant differences in affinty constants for testoterone
(T), the & for CML being 1 -28 nM (normal : - 1 -8 nM). With mibolerone (MB), & for
KLH fibroblasts was 0.1 nM, B-, 52 hoilmg protein; with methyltrienolone (MT,
Rl881), Kd was 0.12 n . B,, 52 hoVmg protein (Normal: &, 0.1-0.3 nM; B,, 15-50
hoilmg protein). When exposed to DHT at 42OC, the & and B- values of AR in CML
fibroblasts (0.27 nM and 3 5 h o 1 DHWmg protein respectively), were no different from
C T A G C T A G I f f r
Figure 2 ~ ~ s ï a c ! a ~ ~ c r r : G ~ C ; C I 3 f A x 5 t r i senical jkin n5rC=:;5:5. \ cr -21 :=e- c:rcios ana 5oiC ;;ces; Zr;= .Tr;canc !C.LiL. %let c:r=:es cqc 3 0 r ~ a i iices.
KLh, Cilec wuarcs and :.-iin !ines] ~brooiasï cccaiase-s ..vere rxposec :a 2 nX: [2ui4\la ( top, . 5 z:\l [ j?]DnÏ(middle , or 3 IL! ['LijT,.boc- - - -p - tom! ac 3 1 -L ,lek) or 22-C <nghc, s r 2 hours. . g.e rzCioiz=e;r= meci- ~ r n was ciscw5ed I P G r ~ ~ i a c t d 'aicn one c O R t z : i l r n ~ :QQ--Oic ~ X C ~ S S
ur;iaoe!es' z n c r o g c : r?s;iczce sz.~?i?s .fier? r e . ~ c v e c r : :.te .?ctcacs5 rimes 2nd assa:/ed fur f 2 + lanz ' ropn chat -.vas si:;! rec29:ar j c u O z = . Ex: &:a coiri:. :ne ?i?zr: ûf2 ? ~ I k o : ~ s . z s exsr2ss& 35 2 =zr:?r;:z -3 3F, - t x - 5 t n u n bin,-:rs 2: ::ce ? '/r-t1<3l i.\?S 3re an r h ~ 54.72 Ics;r::.-.-ic 5 ~ 2 : e .
e x c x ~ ?or 35::3m ripr, =x.ze! .
the corresponding values at 3PC, indicating the absence of thermolability. The expenment
was repeated using ME3 and MT with similar results (data not shown). Prolonged labelling
for 15 h at 37°C with 2 n M MT, or MB, followed by a swïtch to 41°C or 42'C for up to 6
h in the presence of 100 m M cycloheximide, failed to demonstrate any difference between
mutant and normal cells at the higher temperatures. Chase experiments were perfonned to
examine the dissociation kinetics of mutant androgen-receptor complexes. When chased at
37°C or 42"C, the dissociation rates for T, DHT and MI3 were similar for both normal and
mutant fibrobhsts (Fig. 2). Mer an 18 h incubation with T, the remaining androgen-
receptor complexes dissociated linearly with a k value of 5 (10-3rnin-') at 37OC. This value
is typical of DHT-receptor complexes, and indicates normal conversion of T to DHT. The
AR content of CML fibroblasts rose fiom 48 to 104 finoVmg protein when incubated with
MB for 2 and 20 h, respectively indicating normal receptor upregulation. Normally a
doubling, or greater, of specific androgen-binding activity after overnight incubation with
androgen is observed, presumably because of AR stabilization by ligand. Likewise the
aflinity and dissociation characteristics of androgens bound to mutant receptors expressed
in COS-7 cells were no different fiom the WT (data not shown). In aggregate, these
experiments independently replicated in two laboratories, indicate that the M886V
mutation of the LBD did not change any androgen-binding properties of the receptor.
Impaired transactivation capaciîy of the mutant receptor
However, the mutant receptor had only 50-70% of the transactivation capacity of
the WT receptor in COS-7 cells when exposed to various concentrations of T, DHT and
MI3 (Fig. 3). The ARS encoded by the normal and mutant plasmids dernonstrated dose-
dependent increases in transactivation capacity with testosterone, but at al1 doses fiom 10
to 300 nM the transcriptional capacity of M886V AR was only about half that of the WT
(data not shown). The transactivational defect persisted even with doses of androgens up
to 1 pM. Thus, the mutant AR had a modest, but consistent, reduction in its
transactivation capability compared to the Wï, for al1 three androgens in repeated
expenments. Using a ce11 line (CV-1) that does not express the SV40 antigen, the
20 .I DHT 76- T
Figure 3 Transaccivation accivicy o f iLISs6V AR with high cioses o f androgms.
WT (open circles and solid i i nes ) or m u t a n t ( f i l l e t i c i r c les and d o ~ é d lines) receptors were cransiencly expressed in COS-7 ce l ls ana' exposed ro increasing d o s e s ( n M ) o f T (a), DHT (bj, or M B (c) . Ï r a n s a c ~ i v a - rion activicy was expressed as fold increase in iuciferase accivity corn-
pared wich ce l l s n o c exposed co androgen. P-galactosidase acrivir;~ and prorein concent were used co normalize For cransfec:ion efficiên- cy a n d cell numbers, respeccively. Each daca point r e p r e s e n t s :ne
mean 2 SE o f 4 r epkaces .
5 0 100 2 0 0 cDNA (ng)
Figure 4 ;a) Transac~ivacion rc~ iv i t y o f MSS6V AR with increzsang doses o f - cDNA 011 cêils were corransfecced wich the indicaced zmouncs O?.'
ûr 1V896V AR cONA ana t h e reporcer plasmid ,DMALI~-LUC. Each d~ ?om. ï h e mean oicriplicaces. represencs che foid increase in luciferc i c t i v ig oicel ls cxposed CO 30 ntM M B compared wich chose wicho mdrogén. Bars are 5 SE. (b) Immunoblot oFWT (W) or rnucanc ('.
-ecepcors. CV1 sefi exrracrs ( 1 O ~g :ocal proceln sach) de9ic:ad in a we
eiêccrop*oresed on an S O S - P , G E gel, and AR procein iden ï i ficd WC'-
specihc antibodv (PC-21). Films were overexposed co enhance signai ' riie !ow-abundance .AR procein.
transactivation defect of the mutant AR was even more evident. Thus, when 50, 100 and
200 ng of AR cDNA were used in the transfections, the WT gave 3-10, 1 1.1- and 3.Cfold
higher transactivation activity, respectively, compared to the mutant (Fig. 4A).
Irnrnunoblot analyses showed that AR protein levels were equivalent for mutant and wild-
type transfections, at al1 three cDNA doses (Fig 4B). Androgen-binding activity of the
mutant and \KT AR proteins in COS cells was also similar (data not shown) confirming
that the transactivation defect of M886V was not due to changes in AR protein levels.
Impired binding of M886V AR-ligand complexes to ondrogen resporse elements.
Both WT' and mutant receptors displayed an increase in ARE binding with the
higher hormone dose (Fig. SA, comparing lanes 1 and 3; 2 and 4). However the mutant
receptor, despite being present in slightly greater quantities (Fig. SB), was unable to bind
synthetic ARE as effectively as the WT (Fig. 5 4 lanes 1 and 2; 3 and 4). Addition of
excess unlabelled synthetic ARE reduced both WT and mutant signals, with the mutant
band being less prominent than the WT (Fig. 5 4 lanes 5 and 6). The presence of non-
specific competitor DNA (estrogen response element, ERE; and Oct2a oligonucleotide,
the binding site for a ubiquitous transcription factor) did not reduce the band shifts, but
mutant signal was still less intense than WT in both instances (Fig. SA, lanes 7 and 8; 9
and 10). The specificity of the band shift was confirmed by the absence of the most
prominent band when radiolabelled ERE was used @ig. 5 4 lanes 11, 12). The ARE-
binding deficit of the mutant ligand-AR complexes was quantified. In a first series of
expenments using two different ARES, the mutant receptor had only 57.7% (SE, 1 5.17)
of the WT DNA-binding capacity (Fig. 5C, left panel). In a second series of expenments, a
slightly modified technique was used to reduce background counts and the DNA-binding
defect was even more marked with M886V displayhg only 38% of the activity compared
to normal controls (Fig. 5C, right panel). In comparison, two DNA-binding domain
mutants, APheS82 and AArg6 15, displayed less than 20% of normal binding. In aggregate,
the M886V AR had only about half the DNA-binding activity of the WT, and this defect in
'N Y 'N M 'E/ M 'CI .W ' e V M W M IO 100 ARE 3 E Cc: -E3€
I
W M W M I O roo
Figure 5 .,a) ûiuA i n o ~ r i t r r ~ c ! s n i k assav. 'NT (W; s r rnt.canc (hl) reczpcors u e r J expresscc :n Cas-7 :?ils and èxooscd CO :O n.L: *!anes 1 a n c 2 ; o r :00 niL1: !anes 3-7 2) . OHT. E q u i ~ i e n r quanciaes o k t n ï n o r e a c t i v e .AR r i o m cne ceil emracs were addea ro binding reacrions conta inmg 1:P-laoeled synrneric ARE (lanes : -1 0) o r €RE (Iants : 1 i n d 7 2) o i i p o n u d e 0 ~ i 2 t secuences. 5cess unrabeled ARE Janes 5 ana 5;. ERE (lanes 7 a n d 3 j. o r O a ( lanes 9 ana : 0) oligonuc!eocrdei ..vere acded as compecicor DNA ;O derno-sc.qte :nc s o e c r f i c i ~ ~ oi:ne Sinding -eaction. Ï n e d a r k b a n d 3: :ne b o c r o n resresezcs u n b o u n a I2P-kbeicd ONA. (b) I r r i m u n o d o c of
s r mucarrc recaotors used i n p i shik assav. F r e micro i iwrs o f rc?re- icnracive cet1 -.ma ! useti in the get shik assav c e p c e d in Figure s a j .&as cqaosed :O +.:.lier ;O s r 1 O 0 nhl arOHT ana W ~ S seoaraced o n a n SOS- PACE 3ei- A n procein ,-vas iaentifiea ~ t h a S D ~ C ~ ~ C antiboav (PC-2' ). :cl Quanci6cat:on oi l i n a i n g co ARE?;. Recezor r ivere expresse0 i n COS ce!ls. excosad :O f2+!h1B a n d cqutvaleiit quariocies [30.000 dom; o i
j h l B - A a =omute.xes .ncubacec wtii ? 5û p m o i ofcicher oeoon-labe:èd ?a:zai A 2 5 a n ARE k m .blSlT\/-iÏR; a r jwnttie:ic ARE in 2 ~ n d e p e n c - e x ser~es sfccxnrre-KS. Screccaviatn-Diocin-jounc .ARES were col lccr- e~ fv c'en:r+*Lgxicn. 3r.c ['H i X l 9 - : a ~ c . c i r e e s c o r jound co chc ARES .*as Z,ÏJPC:&C 3v 5; ~ c ! i f a c ~ o n c3uncins 'C.io\vn DNA btrc ing-Comain .nucants ; l r ' 532 . l R 5 ; 5 ) wtr: sc4ere wwcarrrneic oiDNA 51ncing were ~ s c a -kir zcccar tson . 'n :ne mgn: Janet. ~ s k 3 r o u n j ccuncs were :ow- drè5 5v Cva: cg : ic vsa:e wc;r CcrfraniOateC c-arcaal ane ~v c c n t n k - g1::or: 3: 'CO.300 J Gr f rrour. A s a r s . m n g -0cic-.yjn5f<c:eC x * i r :-J,,CC -- - - - a. =a<<;-ou-C i=r;wrd. E X - :,ri :ctnc .*as :ne -ex- ar ..
DNA binding was commensurate with the degree of impaired transactivation observed
above.
The mutation affects TAD-LBD interactions
To determine if the M886V mutation impairs interaction between the LBD and
TAD, vectors encoding the C-terminal LBD, and the N-tenninal TAD were constmcted
and expressed simultaneously in the presence of an ARE-driven reporter gene @ig. 6A).
The LBD or TAD fragments by themselves did not display any androgen-inducible
reporter gene activity. However, WT LBD fiagment when CO-expressed with the TAD,
resulted in an androgen-dependent increase in luciferase activity when exposed to
physiological doses of DHT, indicating that AR fiagments can interact in vivo to yield a
fùnctional protein (Fig. 6 4 left panel). Mutant LBD fiagment, when CO-expressed with
TAD, resulted consistently in a 22-25% lower androgen-inducible activity compared to the
WT LBD fragment in three independent experiments. Fusion proteins, comprising
ARTAD fused to VP16AD, and ARLBD to the GAL4DBD, were CO-expressed in HeLa
cells and protein-protein interactions measured with a reporter vector containing multiple
GAL4DNA binding sites in the marnmalian two-hybrid assay (Fig. 6B). Interactions
between TAD and LBD were specific and androgen-dependent. Differences between WT
and mutant were most evident when about equivalent quantities (50-100ng) of LBD and
TAD hybnd plasmids were CO-transfected (Fig.7B. left panel). Mutant LBD fusion protein
had only half the activity of the WT in this assay when exposed to subnanomolar doses of
MB (Fig.7B. right panel). Thus experiments with AR fiagments using an ARE-reporter,
and AR fusion proteins using GAL4 driven reporter indicated that M886V impaired TAD-
LBD interactions.
ïhe mutatior~ Mects LBD-LBD i11teraction.s
To examine whether the M886V mutation can affect LBD-LBD interactions, we
constmcted hybrid expression vectors wherein hll-length or truncated AR fragments were
fused to the GAL4 DNA-binding [GPBD)] or the GAL4 transactivation [G(TAD)]
1 TAO DBD LBD 919
Figure 6 !a) Trânscriptror. ~cr!vi ty o f AR fra~mencs. Delecion construcs encoa'ing :ne ÀRTAD (a;nino arids 1-50.1) or the ARLBD (amino acids 507-9: 9 ) were c r ~ n s ~ e n t i ~ ~ :rans:2cted inco CV-1 ceils. Transcriccional acwicy I mecs- ured in RLU) o f r r c or L1386V La0 iirzgmencs, aione or coaxpressea with TriD ( 1 US/'. :vas mrasured wich a pMAA-1-LUC in the presence and absence oFindic2:cd amouncs DHT (niL1) ( lef i ) . The experimencs =vere repcaceo on anc re r 1 occascns using independenc plasmia pre?arârions. and ï he resuics -vecs expressed as fold increase in luciferase activity in che presence ana soseice o f 3 n M DHT(nghr). (b! Inreraccions o iLBD and - i AD fusion crocers in the mammalian MO-nybriti sssay. The Fusion pro- ceins ' J P l oAD-.AZT.AD ana CALSDBO-AR180 were coexprcssed in HrLa cells. and rec3Frt.r TAO-LBD inceraccions were measured wich ( ! .7mls-
E l iiTATA-ioc r2pcr:er piasmie. Cells were exposed ;O increasing doses a i
pVPl6AD-ARTAD or MB as ,ntiicaced. Data points represenc mean = SE o h !easc 3 re?iic=ces and reriecc Fold increase in iuciferase activ~ry o t c ~ i l s mposed to hl8 over rnose noc exposed ro tne anàrogen.
domains in the yeast two-hybrid assay (Table 1). No androgen-inducible activity was
observed when GAL4-AR hybnd proteins were CO-expressed with the GAL4 domains
[G(DBD) or G(TAD)] alone (data not shown). Coexpressed G(TAD)-AR(LBD) and
G@BD)-AR(LBD) hybrid proteins containing the WT LBD induced a 5 to 7-fold increase
in transcriptional activity in the presence androgen. In contrast tiision proteins containing
the mutant LBD were consistently only half as transcriptionally active as the WT in three
independent expenments, each performed in tnplicate. Defective interactions were also
observed when a larger AR fiagrnent G(TAD)-AIX@BDLBD) was used (data not shown).
Notably when fiill-length AR [G(TAD)-(AR)] fùsion protein was CO-expressed,
interactions were two orders of magnitude higher than with truncated AR recombinant
proteins lacking ARTAD, indicating the importance of TAD-LBD interactions to the
dimerization process. This increase in magnitude of interaction may reflect intramolecular
LBD-TAD association (1 6) or the effect of the TAD on the LBD intermolecularly (1 7). In
al1 these expenments, MS86V nision proteins were consistently transcnptionally defective,
having only about haif the activity of the WT.
Mt~fation disrupts inferactions with the cwctivafor, TIF2
Methionine 886 of the AR is close to residues known to interact with the steroid
receptor coactivator TIF2 (18). We therefore examined the effect of M886V on
coactivator Nnction in three ways. Firstly, we measured the effect of full length TIF2 on
WT and mutant AR transcriptional activity, secondly the interaction of fusion proteins
containing TIF2 and LBD, and thirdly the effect of full length TIF2 on TAD-LBD
interactions. Full length TIF2 enhanced WT AR activity -3-fold in a dose-dependent
manner (Fig. 7 4 left panel). However M886V was defective and lowered CO-activator
function by 2241%. Significant impairment o f TIF2 ninction was most prominent at a
dose of O. 1 nM MB (Fig. 7A right panel). When fision proteins containing GAL4DBD-
m B D were CO-expressed with VP16AD-TIF2, M886V caused a mean 37% lower
receptor/coaaivator interaction with doses of ME3 fiom 0.01 to 1 n M (Fig 7B left panel).
Impaired interaction of TIF2 with M886V was not corrected by increasing the amount of
Figure 7
TIF2 cDNA (Fig 7B, right panel). Defective interactions of mutant LBD with TIF2 were
not due to dserences in LBD fision protein content as measured by immunoblots and
[3mMB binding (Fig 78 lowest panels). To determine whether N- and C-terminal
interactions of the AR were also regulated by coactivator, full length TIF2 was co-
expressed with WT or mutant GAL4DBD-ARLBD and VP 16AD-ARTAD fusion proteins
in the mamrnalian two-hybrid system. Interactions between the WT LBD and the TAD
Nsion proteins were increased up to 20-fold in the presence of TIF2. M886V disrupted
this effect of TIF2 by up to 28% consistent with the impaired TAD-LBD interactions
observed previously (Fig. 7C).
Discussion
Earlier studies (2-4, 19, 20) have provided endocrine-biochemicai evidence linking
oligo/azoospermia, with or without other signs of undervirilisation, to quantitative andor
qualitative abnomalities of the AR There have been only two case reports (6, 2 1) of AR
point mutations in men with diEerent degrees of impaired sperrnatogenesis. This study was
undertaken to discover constitutional AR mutations in a series of 173 men with varying
degrees of impaired sperrnatogenesis. Among the subset of 84 men with severe, idiopathic
oligospennia, we found three, unrelated, each of whom had the sarne novel AR gene
variant: M886V. The fact that we did not find the variant among 400 control AR alleles
indicated that this was a significant allelic variation. One of the men shaves intiequently;
the other has Tanner II persistent postpubertal gynecomastia. These two faas
strengthened the idea that the AR variant was pathogenic. To prove its pathogenicity we
undertook the studies whose results are discussed below.
The location of the Met886Val mutation in the LBD of the AR led us to expect
some degree of androgen-binding abnormality since a nearby mutation, V889M, 3 residues
downstream, has been reported to cause nearly complete androgen insensitivity due to
defective androgen-binding capacity (22). Furthemore, ail mutations so far described in
exon 8 of the AR manifest some abnormality of androgen binding (23). To our surprise, al1
androgen-binding properties of the M886V in genital skin fibroblasts of both probands
were within normal lirnits. Similady, the androgen-binding propenies of the mutant AR
were normal in two transfected mammalian ce11 lines. Thus the M886V mutation, although
residing in the hormone-binding domain of the AR, does not affect the conformation of the
ligand-binding pocket of the LBD. However the mutant AR was unable to tramactivate
normally and disptayed only 50 to 70% of the transactivation ability of the WT in three
dif5erent ce11 lines (COS, CV-1, HeLa) as measured by natural (MMTV-LTR) or
multimeric ARE-reponer genes. Al1 transactivation-defective LBD mutations reviewed to
date (23) have been associated with some form of androgen-binding abnomalities. In this
regard it is striking to note that the nearby V889M mutation caused nearly complete
androgen insensitivity with increased androgen dissociation kinetics despite norrnal
equilibrium androgen-binding affinity (24). The activated mutant receptor had an impaired
ability to bind oligonucleotides contaking two dinerent ARES: the first, an artificial ARE
based on the consensus DNA-binding site for androgens; the second, MMTV-LTR, the
naturally occurring steroid response element. Binding to either ARE was about 60% of the
WT. Thus it is vev likely that reduced ARE binding contributes to the reduced
transactivation competence of the M886V AR. However as discussed below, faulty intra-
and intermolecular interactions of the mutant AR may, directly or indirectly, also be
contributory. Mutations such as M886V are particularly interesting as they can illuminate
fiinctional subdomains that reside in the LBD.
Interactions between the N and C-terminal domains of the AR have been observed
using the marnmalian two-hybnd system (17, 25), and a transcriptionally active complex
forms when truncated proteins having only the N- and C-terminal domains of the AR are
CO-expressed in mamrnalian cells (16). Several lines of evidence indicate that defective
TAD-LBD and LBD-LBD interactions are important in the pathogenicity of M886V.
M886 in the C-terminal region, directly or indirectly, interacts with the N-terminal portion
of the AR. M886V impairs TAD-LBD interactions in both AR- and GAL4-dnven reporter
systems. M886V also consistently disrupted LBD-LBD interactions of G U - A R fusion
proteins in the yeast two-hybrid assay. These data indicate that Mg86 has a role in
interactions between the functional domains of the AR. It is important to note that the
nearby residue, V889, although a part of the ligand-binding pocket, also interacts with the
TAD (26). Similarly, mutations of the human estrogen receptor affecting C530 (27), about
5 residues N-terminal to the homologue of AR M886, lose DNA-binding activity but
retain normal estradiol atfinity. Collectively, these data suggest strongly that a functional
element centered around residue 886 of the AR has a role, not for ligand binding, but for
inter-domain interactions.
Recently several steroid receptor CO-activators (SE) have been characterized that
mediate nuclear receptor interactions with the pre-initiation complex and the chromatin
template (28). The SRC gene family has several closely related homologues and of these,
TiF2, has the greatest activity with respect to AR (29). Three highly conserved regions,
each containing the nuclear-receptor interacting motif LXXLL, are located within the
nuclear-receptor interacting domain of TIF2 (Fig. 78). The LXXLL motifs interact with
core motifs in LBDs of nuclear receptors close to residue 886 of the AR (30). M886V
significantly impairs TIF2 coactivator fiinction in an AR-dnven reporter system and also
with chimeric proteins in the mamrnalian two-hybrid assay. In both systems, MSS6V
reduces TIF2 CO-activation function by +O%, equivalent to the defects in transactivation,
DNA-binding and inter-domain interactions observed. There is evidence that TIF2
interacts with both the ?AD and LBD of steroid receptors. enhances the transcriptional
activity of TAD and LBD separately, and has an additive effect when both TAD and LBD
are expressed simultaneously (3 1). Our data shows that TIF2 increases by about 20-fold
the interactions between the TAD and LBD of the AR and this interaction is dismpted by
M886V (Fig. 7C). Thus it is not surprising that our mutation cm affect inter-domain
interactions and DNA binding since TAD-LBD interactions are thought to be critical for
both fûnctions (26). It is known that ligand-induced recruitment of a L-?UCLL motif of
SRC- 1 to the retinoic acid receptor (RAR) promotes heterodimerization to the retinoic
acid X receptor (RXR), and binding of a second LXXLL motif molecule to RXR. The two
adjacent LXXLL motifs in SRC-1 contacts homotogous residues in the LBDs of the RXR-
RAR dimers, thereby fonning a charge clamp to stabilize the dimerization complex (32).
Our findings that the M886V mutation affects many AR processes including TAD-LBD
and LBD-LBD interactions, DNA binding, and transactivation competence, are consistent
with the integrative role that coactivators like TIF2 may have in these post ligand-binding
events.
Since a nearby AR mutation (E897Q) also preserves ligand binding whilst
disrupting TAD-LBD and LBD-TIF2 interactions (12). our data suggest that M886
contributes to a tiinctional subdomain of the AR LBD that does not mediate androgen-
binding, but is important for interaction with TE2. This ftnctional dichotomy in the AR
LBD is not irnplausible considering that intermolecular interactions depend on surface
properties, whilst ligand-binding pockets are buried within the hydrophobie core of ail
steroid receptor LBDs crystallised to date (33). The identification of other transactivation-
defective LBD mutations (6,21) that do not affect ligand binding suggests the possibility
that disruption of one or another coactivator activity may be a common contributor to
pathogenesis of severe oligospermia and male infertility. Since at least one of these
abnormalities is correctable by hormonal manipulation (6, 21), understanding the
pathogenetic mechanisms may lead to effective therapeutic strategies.
Acknowledgments
We wish to thank Dr G. Jenster for his kind gifi of ARE-TATA-Luc; Dr G. Pnns,
for the AR antibody, PG-21; and Professor P. Chambon for pSG5-TIF2. This work is
supported by grants fiom the Fonds de la Recherche en Santé du Québec (Hydro-
Québec); Fonds pour la Formulation de Chercheurs et l'Aide a la Recherche; and the
Medical Research Councüs of Canada and Singapore. A.A.R. Abdullah is grateful for
personal support fiom the University of Baluain.
References
1. World Health Organization 199 1 Infertility: a tabulation of available data on the prevalence of primary and secondary infertility. WHO Geneva, pp 1-72.
2. Aiman J., Griffin J.E. 1982 The frequency of androgen receptor deficiency in infertile men. J. Clin. Endocrinol. Metab. 54:725-732.
3. Bouchard P., Wright F., Portois M C , Couzinet B., Schaison G., Mowzowicz 1. 1986 Androgen insensitivity in oligospedc men: A reappraisal. J. Clin. Endocrinol. Metab. 63: 1242-1246.
4. Morrow A.F., Gyorki S., Warne G.L., Burger H.G., Bangah M.L. Outch K.H., Mirovics A., Baker H.W.G. 1987 Variable androgen receptor levels in infertile men. J. Clin. Endocrinol. Metab. 64: 1 1 1 5- 1 1 2 1 .
5. Gnno P.B. Griffin J.E. Cushard W.G., Wilson J.D. 1988 A mutation of the androgen receptor associated with partiai androgen resistance, familial gynaecomastia, and fertility. J. Clin. Endocrinol. Metab. 66:754-76 1.
6. Wang Q., Ghadessy F. J., Trounson A. Kretser D., McLachlan R., Ng SC. , Yong E.L. 1998 Azoosperrnia associated with a mutation in the ligand-binding domain of an androgen receptor displaying normal ligand binding, but defective transactivation. J. Clin. Endocrinol. Metab. 83 A303 -43 09.
7. Tsukada T., Inoue M., Tachibana S., Nakai Y., Takebe H. 1994 An androgen receptor mutation causing androgen resistance in the unde~rilised male syndrome. J. Clin. Endocrinol. Metab.79: 1202- 1207.
8. World Health Organization 1992 Laboratory manual for the examination of human semen and sperm-cervical mucus interactions. ed.3 Aitken, RJ, et al. (eds) WHO special program of research, development and research training in human reproduction. Geneva
9. Yong E.L., Chua K.L., Yang M., Roy A., Ratnam S.S. 1994 Complete androgen insensitivity due to a splice-site mutation in the androgen receptor gene and genetic screening with single stranded conformation polymorphism (SSCP). Fertil. Steril. 6 1 : 856-862.
10. Gottlieb B., Kauhan M., Pinsky L., Leboeuf G., Sotos J.F. 1987 Extracellular correction of the androgen-receptor transformation defect in two families with complete androgen resistance. J. Steroid. Biochem. 28:279-284.
1 1 . Prins G.S. Birch L. Greene G.L. 199 1 Androgen receptor localkation in diserent cell types of the adult rat prostate. Endocrinology 1 t9:3 187-99
12. Tut TG, Ghadessy F, Trifiro MA, Pinsky L, Yong EL. Long polyglutamine tracts in the androgen receptor are associated with reduced transactivation, defective sperm production and male infertility. J Clin. Endocrinol. Metab. l997:82, 3 777-3782.
13. Roche P.J., Hoare S.A., Parker M.G. 1992 A consensus DNA binding site for the androgen receptor. Mol. Endocrinol. 6: 2229-2235
14. Yu WC., Delsert C., Andersen B. Holloway J.M., Devaly O. V., Naar A.M., Kim S.Y., Boutin J.M. Glass C.K., Rosenfeld M.G. 199 1 RXRb: A coregulator that enhances binding of retinoic acid, thyroid hormone, and vitamin D receptors to their cognate response elements. Ce11 67: 125 1 - 1266.
15. Glass C.K., Franco R., Weinberger C., Albert v.R., Evans R.M., Rosenfeld M.G. 1987. A c-erb-A site in rat growth hormone gene mediates transactivation by thyroid hormone. Nature 329: 73 8-74 1 .
16. Doesburg P., Kuil C. W., Berrevoets C.A., Steketee K., Faber P. W., Mulder E., Brinkmam A.O., Trapman J. 1997 Functional in vivo interaction between the amino- terminal, transactivation domain and the ligand binding domain of the androgen receptor. Biochemistry. 36: 1052- 1064.
17. Langley E., Zhou Z.X., Wilson E.M. 1995 Evidence for an anti-parallel orientation of the ligand-activated human androgen receptor dimer. J. Biol. Chem. 270:29983-29990.
18. Berrevoets C.A., Doesburg P., Steketee K., Trapman J., Brinhan A.O. 1998 Functional interactions of the AF-2 avtivation domain core region of the human androgen receptor with the amino-terminal domain and with the transcriptional coactivator TIF2 (Transcriptional Intermediary Factor 2) Mol. Endocrinol -12: 1 172- 1183.
19. Migeon C. J., Brown T.R., Lanes R., PaIacios A., Amrhein J. A., Schoen E. J. 1984 A clinical syndrome of mild androgen insensitivity. J. Clin. Endocrinol. Metab. 59:672- 678.
20. Smallndge R.C., Vigersky R., Glass A.R., Griffin J.E., White B. J., Eil C. 1984 Androgen receptor abnorrnalities in identical twins with oligospermia. Am. J. Med. 77: 1049- 10%.
2 1. Yong E.L., Ng SC., Roy A C , Yun G., Ratnarn S.S. 1994 Pregnancy d e r hormonal correction of severe spennatogenic d e f a due to mutation in androgen receptor gene. Lancet 344:826-827
22. De Bellis A., Quigley C.A., Marschke K.B., El-Awady M.K., Lane M.V., Smith E.P., Sar M., Wilson E.M., French F.S. 1994 Characterisation of mutant androgen receptors causing partial androgen insensitivity syndrome. J. Clin.Endocnno1. Metab. 785 13-522
23. Gottlieb B., Lehvaslaiho H., Beitel L.K., Lumbroso R., Pinsky L., Trifiro M. 1998 The Androgen Receptor Gene Mutations Database. Nucleic Acids Res. 26:234-238.
24. Zhou Z.X., Lane M.V., Kemppainen J. A., French F. S., Wilson E.M. 1995 Specificity of ligand-dependent androgen receptor stabilization: Receptor domain interactions influence ligand dissociation and receptor stability. Mol. Endocnnol. 9:208-2 18.
25. Ikonen T., Pdvimo J.J., Jame O.A. 1997 Interaction between the amino- and carboxyl-terminal regions of the rat androgen receptor modulates transcriptional activity and is iduenced by nuclear receptor coactivators. J. Biol. Chem. 272: 2982 1 - 29828.
26. Langley E. Kemppainen J.A., Wilson E.M. 1998 Intermolecular NH2-Karboxyl- terminal interactions in androgen receptor dimerization revealed by mutations that cause androgen insensitivity. J. Biol. Chem.273 : 92- 10 1
27. NeE S., Sadowski C., Miksicek R. J. 1994 Mutational analysis of cysteine residues within the hormone-binding domain of the human estrogen receptor identifies mutants that are defective in both DNA-binding and subcellular distribution. Mol. Endocrinol. 811215-1223.
28. Voegel J. J., Heine M.J. S ., Tini M., Vivat V., Chambon P., Gronemeyer H. 1998 The coactivator TIF2 contains three nuclear receptor-binding motifs and mediates transactivation through CBP binding-dependent and -independent pathways. EMBO J. 17:507-5 19.
29. Xiu F.D., Anderson C.M., Ma H., Hong H Uht R.M., Kushner P.J., Stallcup M.R. 1998 Nuclear receptor binding sites of coactivators glucocorticoid receptor interacting protein (GRIP 1) and steroid receptor coactivatorl (SRC- 1): Multiple motifs with different binding specificities. Mol. Endocrinol. 1 Z:302-3 13.
30. Nolte R.T., Wisely G.B., Westin S., Cobb J.E., Lambert M.H., Kurokawa R., Rosenfeld M.G., Willson T.M., Glass C.K., Milburn M.V. 1998 Ligand binding and co- activator assembly of the peroxisome proliferator-activated receptor-g Nature 395:137143.
3 1. Onate S.A., Boonyaratanakornkit V., Spencer T.E. Tsai S.Y., Tsai M.J., Edwards D.P., O'Malley B.W. 1998 The steroid receptor coactivator-l contains multiple receptor interacting and activation dornains that cooperatively enhance the activation
ninction 1 (AF 1) and AF2 domains of steroid receptors. J Bi01 Chem 273 : 12 10 1 - 12108.
32. Westin S., Kurokawa R., Nolte R.T., Wisely G.B., Mclnemey E.M., Rose D. W., Milbun M.V., Rosenfeld M.G., Glass C.K. 1998 Interactions controlling the assembly of nuclear-receptor heterodirnen and CO-activaton. Nature 395: 199-202.
33. Tanenbaum D.M., Wang Y., Williams S.P., Sigler P.B. 1998 Crystallographic cornparison of the estrogen and progesterone receptof s ligand binding domains. Proc Nat1 Acad Sci U S A, 95:5998-6003
34. Shiau A.K., Barstad D., Lona P.M., Cheng L., Kushner P.J., Agard D.A., Greene G.L. 1998 The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Ce11 95927-37.
Chapter II
INTRODUCTION TO CHAPTER II
Contrary to loss-of-function mutations secting classical AR functionality, CAG
expansion in exon 1 of the AR leads to a gain-of-fùnction mutation responsible for a
unique neurodegenerative disorder (SBMA) not seen with classical androgen receptor
mutations.
SBMA is one of eight neurodegenerative diseases caused by a CAG expansion
encoding an expanded polyglutarnine tract @olyGln). The pathological event leading to
neuronotoxicity in the polyGln-expanded (encoded by CAG tracts) related diseases has
not been elucidated. Severai hypotheses have been discussed including the requirement of
the parent protein to undergo proteolysis to release a "death substrate". These "death
substrates" would in tum have the propensity of forming intracellular aggregates believed
to be neurotoxic.
In this original article, we address whether the polyGln-expanded and wild-type
AR can form such aggregates. We define the composition of aggregates and whether
aggregation requires proteolytic fragmentation of the parent full-lengt h AR. We also
investigate the relationship between aggregation and increased cellular toxicity as well as
the possible involvement of caspases in the aggregation process. Using novel approaches,
we ascertained that proteolytic cleavage is not a prerequisite for aggregation to occur and
that aggregation may not be necessary for induction of cellular toxicity by the polyûln-
expanded AR.
Cha racterization of Intniceiiular Aggregates Using Fîuorescently-Tagged
Polyglubimine-Expaaded Andmgen Receptor
Valerie panet-~ayrnond'f Lenore K. ~eitel', Hyrnan schipperl"", Michael ~ i m i a n s k ~ ~ ,
Leonard ~insky'~', Mark A. ~rifiro''.'.
Lady Davis Innitute for Medical Research, Sir Mortimer B. Davis- Jewish General ~ospital'
and Departments of ~ i o l o d , ~ e d i c i n e ~ , ~ediatrics" Hurnan ~enet ics~ , Neurology and
~eurosur~e$, McGill University, Quebec, Division of Neurosurgery, Toronto Western
~ospi ta l~ , Ontario, Canada .
Proofs and correspondence to :
M.A Trifiro, M.D.
Lady Davis Institute for Medical Research
Sir Mortimer B. Davis- Jewish Gened Hospital
3 755 Cote-Ste-Catherine Rd.,
Montreai, Quebec
Canada H3T LE2
Fax: (5 14) 340-7502
E-mail:mdtm@mu~ica.mcg~~ca
Abstract
Spinal buibar m u d a r atrophy (SBMA) is a classic CAG-repeat neurodegenerative
disease. It is caused by expansion of a polyGln tract in the androgen receptor (AR). Recent
evidence has indicated a potential role for nuclear and cytoplasmïc inclusions in the
pathogmesis of polyûin-expanded diseases. We have used blue and green fluorescently-tagged
AR to show that both wild-type and polyûln-expanded MI-length AR can fom aggregata and
that aggregation is not related to cytotoxicity. Twenty to thùty-five percent of al1 transfected
cell types showed aggregation containing both amino- and carboxy-terminal fluorescent tags.
The aggregates form in COS-1 celis and in NSC-34 celis, a hybrid motor-neuron-like line. The
aggregates reacted with an anti-AR antiiody (F39.4. l), and with lC2, an expanded polyGln
track antibody. Western analysis of protein extracts revealed Iittle evïdence of proteolysis
although some cleavage of the fusion proteins was seen. The general caspase inhibitor, Z-
DEW-FMK, did not affect aggregation in either wild type (WT) or polyGIn-expanded GFP-
AR &'id celis. Surprisin& addition of a synthetic androgen was found to signïfjcantly
decrease inclusion formation in both WT and polyGh-expanded AR-tranâected ceUs. Overail,
we show that both WT and polyGln expanded hl-length AR are found in aggregates and that
proteolysis need not be a requirement for aggregation. Our rmlts also suggest that toxicity is
not related to cellular aggregation in po1yGl.n expanded AR
Introduction
Genomic expansions of translated CAG repeats have now ben recognized to be the
etiology of eïght neurodegenerative disorders. Glutamine (Gln) tract-stpanded d i m include
spinal bulbar musaila. atrophy (SBMA) (La Spada et aL, 1991), Huntington disease
(HDCRG, 1993), dentatorubral pallidoluysian atrophy (DRPLA) (Koide et d, 1994;
Nagfishi et al., 1994) and the spinocerebeilar &as 1-3,6,7 (Orr er al., 1993; Kawagucbi et
d., 1994; Imbert et al., 1996; PuIst et al., 1996; Sanpei el al., 19%; David et al., 1997; Koob
et al., 1998). Each disease is characterized by a degeneration of speafic neuronal populations
and each resuits in a characteristic neurologicai phenotype. The molecular pathological
pathways have yet to be elucidated although a toxic gain off'unction is thought to underlie the
neurodegenerative process. The toxic gain-of-finaion hypothesis is strongly suppoited by
several lines of evidence. For instance, a total deletion of the X-Linked androgen receptor gene
(AR) causes complete androgen insensitivity without any neural degeneration. Likewise, a
mouse lacking atêuin- 1 does not d e r f b m any ataxia (Matilla et al., 1998).
Several mechanisrns for this toxic gain of finction have been proposed. One theoiy is
that Gin-expandeci tracts may be substrates for transglutaminases thereby covalently
sequeste~g other st i i i undehed proteins (Kahlem et al., 1996). Alternatively, Gh-expandeci
tracts may interact with specific targeted proteins through polar zippering and potentialiy create
insoluble complexes (Pemtz, et al., 1994).
Proteolytic cleavage of the parent protein has also been postulated as an important step
leading to neuronal degeneration. Several studies have shown that fiil-length polyûin-
expanded parent proteins do not themselves confer toxicity, while specifically cleaveâ or
synthetidy truncated products promote aggregation and neuronal degeneration (Ikeda et al.,
19%; Martindale et al., 1998; Paulson et al., 1997; Igarashi et al., 1998). Moreover, several
proteins involved in potyGin-expanded diseases may be susceptible to cleavage by s p d c
caspases, proteases activated during apoptosis. The importance of caspases in neuronotoxicity
has been undemred by tramgenic mice deficient in caspase-3 that dernonsaate a lethal
dysregulation of apoptosis (Kuida et al., 1996). Fuii-length huntingin, atrophin-1 and
amin-3 are substrates for caspases and theû susceptiiity to cleavage may relate to theu
polyGLn tract length (Goldberg et al., 1996; Miyashita et a L ,1997; Wellington et al., 1998;
Miwa et a[. , 1998).
Recent daîa have been contlcting on whether formation of intrecellular aggregates is
an indispensible part of the disease process (reviewed by Lunkes and Mandel, 1997; Ross,
1997; Davies et a[., 1998; Kim and Tami, 1998). These aggregates are thought to represent an
accumulation of insoluble proteolytic byproducts of full-length parent protein These inclusions
have beem descn'bed in many hurnan pathological studies, DNA transfection studies and in
transgenic animais of polyGin-expanded diseases. Such inclusion bodies are typicaüy but not
exclusively found in the targeted areas of neuropathology and have been characterized in HD,
DRPL& SBMA, SCAl, SCA3 and SCA7 (reviewed by Lunkes and Mendel, 1997; Ross,
1997; Davies et al., 1998; Kim and Tanzi, 1998; Hoimberg et al., 19%). Inclusions have been
found both in the cytoplasm and the nudeus of cells. The pathogenic relevance of nuclear
localization is still unclear.
The mechankm that confers neurodegenerative cellular s p d c i t y in polyGin-expanded
diseases has not been dehed; it may relate to neuron-specific proteolytic enzymes. Proteolytic
cleavage of the polyGin-expandeci proteins may tiberate a toxk hgment containing the
expanded Gln tract that could promote aggregation. Many mouse models and cellular
transfection experiments have supponed this mode1 by induchg aggregation through
expression of a fiagrnent of a polyGh-expanded protein (Skinner et al., 1997; Davies et al.,
1997; Perez et al., 1998; Myashita et al., 1998; Cooper et al., 1 998; Martindale et al., 1 998).
Expanded polyglutamine disease was also recreated in drosophila by targeted expression of a
portion of ataxui-3, the disease protein in SCA3 (Warrick et al., 1998). While nuclear inclusion
formation was observed in the drosophila, it was insufficient for degeneration. Interestingiy,
expression of an expanded Gln tract alone has been found to be cytotoxic in ceus.
Recently, however, two reports have undermineci the importance of cellular inclusions
in polyGin-expanded disease pathogenesis. in Sandou et al. (1998), onset of celi death did not
correlate with aggregate formation in ceUs transfected with mutant huntingtin. in transgenic
mice expressing CAG expanded ataxin-1 lacking a seIfIfassociation region, ataxia and Puricinje
ceU pathology both developed similariy to the mice expresshg fiiU-length ataxin-1 but in the
absence of ceUdar inclusions (Klement et al., 1998). While disease progression rnay occur in
the absence of aggregate formation in mice, it is unknown if inclusions do indeed promote
pathogenesis in humans.
SBMA is a rare disease affecting motor neurons which causes muscular weakness and
atrophy beginn.ig at 30 to 50 years of age (Kemedy et al., 1968; Sobue et a[, 1989).
Expansion of the CAG trinudeotide repeat fiom 10-36 to 40-62 also confers mild partial
androgen resistance which is chanicterized clinidy by gynecomastia and am- or oligospemiia
(La Spada et al., 1991). AR and ubiquitin-positive inclusions have been found in anterior
motomeurons of SBMA patients (Li et al., 1998a). AR-positive aggregates have also been
seen În nonneural ceils of SBMA patients (Li et al., 1998b).
The exact content of Gln-expanded AR aggregates has not yet been determined. Gin-
expanded AR and other Gln-expanded protein aggregates have been shown to lack specific
antigenic epitopes. This suggests that proteolytic processing may be a requirement for
aggregate formation ( DiFigiia et al, 1997; Li et al., 1998a). It is alço possible that post-
translationai events or abnormal protein folding are part of the aggregation process and may, as
a consequence, mask certain epitopes. In this study, N-terminal, C-temiinai and w m b i i N-C
terminal fluorescently tagged WT and Ginexpandeci AR were used to help detennine if
proteolysis is a prerquisite for AR aggregation. As weU, several AR antibodies were employed
to help define AR in aggregates. The effects of androgen and caspase inhibition on cellular
aggregation were also examined. We show that full-length normal and polyûin-expanded AR
can aggregate in COS-1 ceils and in a motor neuron-üke ceU line (NSC-34). The presence of
both the N- and C-terminal fluorescent tags as weii as an anti-AR [aa(301-3 19)] and anti-
expanded Gin tract antibody epitopes in the aggngates suggest that ttClength AR is Îndeed
present in the cellular inclusions. The percent of ceils containhg aggregates was not found to
coirelate with toxicity as measured by ethidium monoazide bromide siainuig and was not
affécted by addition of the caspase inhibitor 2-DEVD-FMK. Moreover, western analysis of ce11
lysates and irnmunocytochemistry revealed litde evidence that proteolytic degradation
contributes to intraceiiular aggregation. Taken together, our findings suggest that proteolysis
conm'buting to the aggregation process may not be an essential component of SBMA
pathogenesis.
Materials and Methoàs
Cell Chlare and Tnursii T e d ~ a 3 n s
COS4 ceiis were maintained in OptiMEM (Giiibco, Grand Island, NY) supplemented
with 5% fetai bovine senun (FBS). NSC-34 cds were maintained in Ddbecco's modifieci
Eagle's medium cuntaining 100/o FBS and 0.1 mghl penicillin. Celis were placed on chamber
slides (Nalge Nunc intemationai, Napiede, IL) the day before t d ê c t i o n at a density of 2 x
lo4/100 mmZ. Transfechons were performed using LipoféctAMINE (Gibco BRL, Grand Isle,
NY). Cells were 6rst rinsed with serum-fke media and 2.5 pg of DNA were mixed with 12 pi
of LipofectAMINE in a total of 1.2 ml of senim-fiee media 200 pi of the DNA-
Lipof- mixture were then added to each chamber. Mer 6 hours of lipofaion, the
media was replaced with 5% FBS supplemented media. 10 nM of Mibolerone (MB), a
synthetic androgen, was added to ceils post-transfection or 1 hour before fixation. 100 ph4 of
the caspase inhibitor 2-DEVD-FMK was also added to celis post-transfection and lefl on the
celis until fixation.
Transfdons for western analysis were perfonned as descn i previously (Beitel et
al., 1994).
Ejqrnmi'on Çonsthwts
The pEGFP-AR wild-type (Wï) and glutamine-expanded (Gln-expanded) consmias
were made by excising the Smd-BamHI hAR cDNA âagment from pSVARBHEXE 0
containhg either a 20- or a 50-glutamine(Q) polyQ tract (Beitel et ai., 1994). The AR
hgnents were then each inserted in frame into the similady digested pS65T-Cl vector
(Clontech, Pa10 Alto, Caüfomia) that codes for the green fluorescent protein (GFP). Ligation
renilted in the pGFP-AR construct coding for GFP fised to the amho terminai of the fiill-
length mutant or wild-type AR (aa 35-919). The BFP-tagged hAR constructs were made by
f h t digesting the 204 and 5 0 4 pSVARBHEXE constnicts with NheI and h I and isolating
BFP
GFP-AR
BFP-AR
Gln tnct
1 58 78 301 319 538 625 919
1 . TAO . .--;.; LBD I
35 91 9
TAD 1 w0l LBD 1
35 91 9
GFP-AR-BFP TAD f 060 ( LBI)
Figure 1. A) Schematic representaîion of the androgen receptor (AR). The key domains of the steroid receptor are shown. The DN-4-binding domain is shaded and the glutamine tract is hatched. Nmbers represent the amino acid positions. The epitopes of the F39.4.1 and the 1C2 anthdies are shown.(B) The expresseci fiiston proteins used in the transfection experiments are show schematically. The regions e n d i n g GFP and BFP are hatched and dotted respectively .
the AR fragments. For the remaining C-terminal portion of the AR, PCR extension was used
and the resulting PCR product was digested with BamHI and M. The primers used to
ampli@ the needed hgment were: TCFA (S'CAGTGGATGGGCTGAAAAAT3 ') and 10 12B
(S'CGTGGATCCCCCTGGGTGTGGAAATAG 3'). The stop codon at the end of AR was
abolished such that translation of the ttsion protein would be unintempted. The EBFP-NI
veztor (Clontech, Palo Aho. California) encoding the blue fluorescent protein (BFP) was
digested with BamHI and NheI. A three point Ligation was penonned to fiise the NheI-PvuI
204 or 50Q hgment of AR with the M-BarnHI PCR fiagrnent and the BamHI-Mie1
double-digested pEBFP-NI vector. The resulting vectors encode a fuii length AR (aa 1-919)
C-terminally fùsed to BFP. To construct the dual GFP-BFP lakled hAR constructs, the entire
GFP coding sequence was rernoved from the pS65T-CI vector by double-digestion with NheI
and SmaI. AAer isolation, the GFP fiagment was inserted in h e into the pEBFP-AR 2ûQ
and 50Q constructs that had been sunilarly digested. The pGFP-AR-BFP constnicts code for a
near full-length expandeci ( 5 0 or WT(2OQ) hAR (aa 35-9 19) fùsed N-terminally to GFP and
C-temllnally to BFP. The constructs used are shown schematidy in Figure 1.
Seventy-two hours post-transfèction, ceiis in charnber siides were rùised twice with
PBS, k e d in 4% formalin for 10 minutes, rinsed again with PBS then overlayed with 5%
gIy~efol-PB S.
The *munocytochernistry protocol was modined from Jenster et al.. (199 1). Ceiis
were permeabilized with -20°C methanol for 4 minutes and then with -20°C acetone for 2
minutes foliowuig fornalin fixation. PBS was added for 5 minutes to rehydrate the celis. CeUs
were preincubated with a blocking solution of PBS with 10% FBS for 15 minutes. Two
anti'bodies were used for imrnunocytochemistry: the F39.4.1. monoc10nal anti'body which
recognim arnino acids 301-3 19 of the hAR (Zegers et al., 1991) and the mAblC2 which
recognizes an expandeci glutamine tract (Trottier et d . 1995, Lesaire et al., 1994). The
monoclonal antibody F39.4.1 was a gift fiom Dr. A. O. Brinkmann, Erasmus University,
Rotterdam and mAb 1 C2 was a gi f t fkom Dr.Y. Trottier, Université de Strasbourg, Strasbourg.
Incubation with primary antibody was performed for I hour at roorn temperature in PBS with
1% BSA with either a 1:1000 dilution of F39.4.1 antibody or a 1500 dilution of mAblC2.
Afier incubation with prirnary antibody and additional washes in PBS, cells were exposed to a
rhodamine-labeled seconciaty antibody TRITC goat Anti-mouse IgG (Zymed, San Francisco,
CA) used at a 1500 dilution for 1 hour. After washing, ceiis were mounted in 500h giycerol in
PBS solution and visuaüzed with a Leitz Aristoplan fluorescence microscope.
Inunrcnoblotthg
Western analysis was performed as descriied previously (Okajima et ai., 1993). Three
amibodies were used for UNnunoanalysis: F39.4.1, mAb l C2 and an anti-GFP monoclonal
amibody. The anti-GFP monoclonal antibody (Clontech, Pa10 Alto, Catifornia) was used as
recommended in the Clontech protocol in a 1500 dilution. A goat anti-mouse IgG secondary
antibody conjugated with horseradish peroxidase (Amersharn Life Sciences) was used at a 1 : 10
000 dilution in PBS with 0.5% Tween. Blots were developed using the ECL western blotting
cherniluminescence detection system (Arnersham Lie Sciences).
Results
GFP-tagged and BFP-tagged hAR causes aggtegah0on in vivo
To assess possible abnormal proteolytic processing of the polyGin-expanded AR
versus the wild-type AR in Mvo, both poIyGIn-expanded and WT AR were tagged with
blue and green fluorescent proteins (BFP and GFP) and transfected h o COS-1 cells. AR was
tagged on its amino-texminus with GFP and on its carboxy-terminus with BFP. It was
established that the W T GFP-AR and BFP-AR behaved O<e untagged AR in W. Both
remained mostly in the cytoplasm unless miilerone (MB), a synthetic androgen, was added to
the media. Upon addition of 10 to 100 nM of MB for 1 hour, the fluorescently tagged proteins
localized to the nucleus. Our results confinn those of Georget et ai., (1997) who found little
effèct of GFP fùsion on AR localization and activity.
Transfcted ceUs were then investigated for diierences in fluorescence between the
polyGln-expanded and WT AR-transfécted cds. It was hypothesized that protein processing
may lead to pattern and cellular localkation diEerences between GFP- and BFP-tagged AR and
between the Wf or polyGln-expanded venions of the AR Ceiis were also examinai for
possible aggregation, as aggregates or inclusion bodies have been suggested as a possible
source of pathogenicity in expanded-polyglutamine repeat diseases.
Interestingiy, we found aggregates in the 20Q and 50Q GFP-AR and BFP-AR
t d e c t e d celis 2-4 days p o s t - W d o n (Fig.2 A and C, only 50Q shown). Expetiments
were repeated multiple times to confjrm our findings, each time showing aggregation in both
the 2 0 4 and 50Q GFP-AR and BFP-AR transfected cds. Both the WT and polyGin-expandeci
GFP-tagged AR proteins were found to eggregate to the same extent, with 20 to 35% of ceils
showing aggregation (Fig. 3). Our hdings, therefore, indicate that the arnino-terminus and
carboxy-tenninus of both n o d and mutant GFP-AR, can be incorporated into intraceUular
aggregates in vitno.
The BFP-AR 204- and SOQ-transfected ceils were thm compareci to the GFP-AR
204 and 5ûQ transfkcted ceils to examine potential diaerences in aggregation and or protein
processing between C- and N-terminaüy labeled proteins. Aggregates were determineci to be
sirniiar in size and number in both polyGLn-expanded and WT. There were no differences seen
Figure 2. Aggregation in COS-1 cells using polyGln-expanded and 'NT GFP and
BFP tagged AR. GFP-AR 50Q (A) and BFP-AR SOQ (C) transfected cells were
immunostained with lC2 (B) and (D) respectively. The green fluorescence and blue
fluorescence of GFP-AR-BFP 50Q transfected cells are shown in (E) and (F), (H)
and 0. Immunocytochemisîry with F39.4.1 was performed on the ce11 (E, F) which
is shown in (G) and with 1C2 on the ce11 (HJ), shown in (J).
between c d s transfected with BFP-tagged AR or GFP-wed AR. in certain cds, aggregates
were srnaller and nurnerous (Fig. 2A); others had only one or two large inclusions (Fig. 2 E and
H). The distribution of aggregates was the same in ail cases: they were found both in the
cytoplasm and in the nucleus, although cytoplasrnic aggregates were predominant. GFP or
BFP, alone, spread dfisely across the celis and did not form inciusion bodies. Our results
demonstrate that the aggregation pattern seen in aansfeaed celis is independent of glutamine
tract length and that both N- and C-terminal fluorescently labeled WT or polyGin-expanded
AR display similar patterns.
To determine whether the aggregates seen in the BFP-AR and GFP-AR transfected
c d s contah intact AR protein, AR was duaiiy tagged with BFP and GFP and cornpared
intraceUular aggregation patterns to the GFP-AR and BFP-AR t d é c t e d COS-1 cells. In the
GFP-AR-BFP t d e c t e d ceiis, aggregation was again observai for both the poIyGln-
expandeci (Fig. 2 E and F, H and I) and WT forms (Fig. 5 B and C). The aggregates found
were both green and blue fluorescently labeled, indicating that both tags were in the aggregates.
Our results indicate that both ends of the carboxy and arnino, are found in aggregates in
vivo. To ensure that both fluorescent tags were indeed in the same aggregates, the ceUs were
Msuaiized using a confocal microscope. Using confocai Mcroscopy, we found that the blue and
green fluorescence colocaliteci in the ceUs in the aggregates. No dserences in the percentage
of cells with aggregates were seen between WT or polyGln-expanded AR or between double-
tagged (GFP-AR-BFP) and single-tagged AR(GFP-AR) (Fg. 3). In the absence of hormone,
aIl sets of celis showed 20 to 35% aggregation.
Duaiiy-ta& AR trandected c d s were then investigated for potential diffierences
between WT and polyGlnexpanded AR either in size, location or frequency of aggregates.
When compad with Wî or polyGin-expanded GFP-AR and BFP-AR transfected cells,
minimal diaerences were seen between aggregates. As in the GFP-AR and BFP-AR
transfected ceus, the GFP-AR-BFP cellular aggregates were seen in both the nucleus and
cytoplasm. Their size and number vsned within the cells (compare Fig. 2 E with H or F with 1).
Both WT and polyûln-expanded GFP-AR-BFP showed the same aggregational patterns.
These results are the fïrst to dernonstrate that WT proteins can agpgate in vivo to the same
AR20 ARS0 AR20 ARS0
GFP 1 GFP-BFP
Figure 3. Quantitation of percentage of GFP-AR and GFP-AR-BFP transfected cells containing aggregates. Both the 20Q and 50Q GFP-AR and GFP-AR-BFP tram fec ted cells are represented. The percentage of cells with aggregates was calculated by counting multiple fields on slides hom different transfection experiments. Greater than 200 cells were assessed per slide. Results are expressed as a percentage of total cells in the presence (black columns) or absence (white columns) of I O nM of mibolerone (MB). Results are also show in the presence of 100 pib! of 2-DEVD-FMK (hatched columns).
extent as polyGin-expandeci proteins. Interestingiy, toxicity studies previously pdorrned
demonsirated a signifiant increase in toxicity when polyGin-expanded AR was transfected
instead of WT (Sculptoreanu et aL, 1999 accepteci). PolyGin-expanded AR transfécted cells
showeâ twice the amount of ce11 death when compared with WT AR transfected ceUs. Taken
together, these findings indicate that aggregation rnay not promote toxicity in polyGh-
expanded AR transfected cells.
T d e c t i o n s were repeated in NSC-34 celis, a hybrid neuroblastoma-embryonic spinal
cord ce11 h e (Cashrnan et al., 1992). Since NSC-34 cells were developed using spinal cord
neurons, a type of newon highiy affecteci in SBMA patients, they may provide a better mode1
to investigate polyGlnexpanded AR pathogenesis. Mer successful NSC-34 Iipofedion, we
found aggregation of both the polyGh-expanded and WT versions of GFP-AR (Fig 4 4 WT not shown). BFP-AR and double-labeled GFP-AR-BFP also showed aggregation, independent
of glutamine tract length (Fig 4 B, C and D). in the GFP-AR-BFP NSC-34 transfected ceUs
(Fig. 4C and D), the aggregates contained both blue and green fluorescent labels indicating that
both the N- and C-tenninals of AR could be found in the aggregates. Again, aggregates
displayed divenity in s k 7 number and cellular locahtion within transfected cells. Aggregates
were shown in both the cytoplasrn and nucleus, independent of the type of fluorescent protein
tagging or Gln tract length. NSC-34 4 s transfected sole1y with GFP or BFP showed ditkse
fluorescent labeüng without aggregation. Our findings suggest that both WT and polyGln-
expanded full-length AR can bring about aggregation in vivo in a motor neuron-like cell line.
Our results aiso demonstrate that both the amino and carboxy tennini of AR can be
incorporated into aggregates fomed within NSC-34 cds.
Aggregates reaci wlh AR antiborljl and Gln-tmcî antihdy
nie presence of both fluorescent tags in the aggregates cwld also indicate that prior
proteolysis permitteci only the ends of the protein to be incorporated into the aggregates.
Likewise, aggregates or protein fbpents that have lost their fluorescent tagging through
processing wodd not be seen. To address whether full-length protein was indeed in the
inclusions and to determine if untagged AR was present in the d s , imrnunocytochemistry with
Figure 4. NSC-34 cells show aggregation with GFP-AR, BFP-AR and GFP-AR-
BFP. Fluorescent microscop y was used to visualize NS C-34 cells transfected with
GFP-AR 50Q(A), BFP-AR 50Q (B) and GFP-AR-BFP 50Q (C and D).
Figure 5. Aggregation in transfected cells after addition of M B and 2-DEVD-FMK.
10 pM of MB was added to a BFP-AR 50Q transfected ce11 1 hour before fixation
(A). 100 pM of 2-DEVD-FMK was added to GFP-AR-BFP 204 cells shown in
green (B) and blue (C).
anti-AR ant~hdies was performed on the fluorescently-tagged AR transfècted cells. Anti-AR
anh'bodies rnay help localize portions of the AR within ceils that can not be seen with end-
labeling. The monoclonal antibody F39.4.1 which recognizes amino acids 301-3 19 of the AR
was used in polyGin-expandeci and normal GFP-AR, BFP-AR- and GFP-AR-BFP-tfaflsfied
cells. The antibody reacted with diffusely expressed tagged-AR protein as well as with
aggregates (Fig. 2 G) fiom al types of transfkcted cells, including the 204 or 50Q GFP-AR,
BFP-AR and GFP-AR-BFP. The F39.4.1 antibody a h reveaied that some of the AR is
àiisely spread throughout the cd . Antibody staining was strongest in the aggregates
indicating that the F39.4.1 epitope can be found in the inclusions. Further studies using
confocal rnicroscopy dernonstrated clear colocalization of the GFP tag and the antibody,
reveaiing that the F39.4.1 epitope is in the sarne aggregates as the GFP tag.
Using the rnAblC2 antibody, an antibody known to react with expanded Gh tracts,
aggregates were detected in the poIyGin-expanded GFP-AR (Fig. 2B), BFP-AR (Fig. 2D) and
duaily tagged AR transfiied ceUs (Fig. 21). Staining was specific and clear colocaiization was
seen between GFP-AR and the rhodamine wnjugated secondary antibody recogniting the 1C2
antibody. hensity was strongest in the aggregates attesting to the presence of the expanded
Gln tract. However, a certain arnount of immunofluorescence was observeci using lC2 that
was not seen with the fluorescent tagging. This may be due to non-speci6c binding of the
primary antibody or it rnay indicate cleavage of the fluorescently tagged AR fùsion protein
generating untagged protein. Expectedly, the 1 C2 antibody did not react with aggregates in the
WT transfected celis. All transfécted cells were incubateâ with the hodamine-conjugated
secondary antibody done to connmi specifcity of the primary antibodies. These data indicate
that both the F39.4.1 epitope and the expanded Gh tract are indeeâ present in aggregates.
Taken together with the findings that both the carboxy and amino termini of the tagged AR are
in the aggregates, our results suggest that both WT and polyGin-expanded fùll-length AR can
be found in aggregates in vivo.
Figure 6. Westem analysis of GFP-AR, BFP-AR and GFP-AR-BFP transfected ce11
lysates. 40 pg of protein were loaded ont0 two different gels and nitrocellulose
membranes were probed with 1C2 (A), F39.4.1 (B) or anti-GFP(C) antibodies. Mots
were developed using the ECL Westem blotting chemiluminscence detection system
(Amersharn Life Sciences).
Western ana&sis of p~l~tein ewkwcts
To fùrther determine if full-length AR was beiig recruited by degradative products into
the aggregates, western anaiysis was perfomed on total ceU extracts of polyGln-expanded and
WT GFP-Aï& BFP-AR and GFP-AR-BFP-transfixted cells. The protein extracts were loaded
on diierent gels and the resulting nitrocellulose membranes were reacted with di8érent
an t i i i e s . The F39.4.1, 1C2 and an anti-GFP antibody were al used as prirnary antibodies.
The anti-GFP antibody recognizes both the green and blue variants of the fluorescent protein.
As shown in Fig. 6, al1 three antibodies were specific and did not react with mock- tdec ted
cells (first lane). The anti-GFP, F39.4.1 and lC2 antibodies show that &ion proteins of
appropriate sires are made in the cells: the single tagged-AR protein is approxhately 137 kDa;
the duaily tagged AR is -164 kDa The BFP-AR proteins are very îàintly recognized by the
ad-GFP antibody indicating that it may be more specific for the GFP variant. hunoblo t t ing
with F39.4.1 shows some secondary bands below the fuson proteins, possibly indicating some
proteolysis of the fiision proteins. ï h e bands may also represent some non-specific biiding
since the sarne bands are not apparent using the lC2 antibody. The 1C2 antibody was s h o w to
bind d e l y to the polyGln-expanded proteins. No h g n e n t s contahing the expanded glutamine
tract were found in any of the transfecteed ceIl proteins. The lC2 antibody demonstrated that
the major product containhg a polyGln tract in the polyGin-expanded AR-transfected ceUs
were the GFP and BFP-AR tùsion proteins.
There was some evidence of deavage of the fluorescent protein tags, generating
unlabeled AR in both the anti-GFP and F39.4.1. incubated blots, especidy in the double
tagged GFP-AR-BFP tfaflsfected ceils Oast lanes). This was not unexpected since fùsion
proteins are known to be more susceptible to proteases at their junaion points. Western
analysis was repeated to confirm our findiigs and ensure that potentiaily smaller proteolytic
fragments would not be rnissed. Repeat experirnents demonstrated an absence of smaller
hgments in aii c d lysates.
AHtion o/MB q # ! i e ~ ~ ~ a m a t r i n , the caspuse inhibitor, 2-DE VD-FiWK dwF not
The possible role of proteases in causing aggregation was fiirther explored by using the
general caspase inhibitor 2-DEVD-FMK. 1 ûûpM 2-DEVD-FMK was added post-tdeaion
to the GFP-AR, BFP-AR and GFP-AR-BFP-transfected c d s until the cells were fixed- It did
not alter the pattern of aggregation in any of the uansfecteed c d s (Fig.5 B and C). It neither
inhiiited nor fàcilitated aggregate formation: the nurnber of celis with aggregates rernained at
23935% (Fig. 3); the size and distriiution of the aggregates within the transfected cells was
unchangeci. These results indicate that AR cleawge caspases does not contribute to aggregate
fonnation.
Importantly, addition of 10 nM mibolerone (MB) signincantly decreased the
percentage of ceils with aggregates, both in the polyGin-expanded and WT-transfied cens.
Oniy 13 to 23% of al1 GFP-AR and GFP-AR-BFP t d e c t e d d s showed cytoplasmic
aggregation in the presence of androgen (Fig. 3). These results may indicate a protective effect
of androgen. They may, iiowever, also indicate an increase in nuclear aggregation in ceUs that
could not be resolved by our fluorescent microscope and were therefore not counted in the
tally of ceUs with aggregates. The s k e of the nucleus and low intensity of fluorescence rnay
rnake it difficult to distinguish dense nuclear aggregation from dise nuclear staining.
Interestingly, when MB was added 1 hour before 6xing the ceils instead of post-transfèction, it
did not seem to affect aggregation (Fig. 5A). Its addition caused the nuclear locaiization of
labeled AR proteins not already incorporated h o aggregates while not afiècting the presence
of intraceiiular aggregates. These results indicate that MB cannot disperse already formed
aggregates in the ceUs but may have a protective effect in decreasing or delaying aggregate
formation.
Discussion
The seledve neuronopathogenesis of polyûln-expanded proteins has yet to be
elucidated. Curent theones include the potentiai neurotoxicity of intracellular aggregates
associat with the expression of polyGln-contaking proteins, and specific proteolytic
processing of parent pdy(;in-proteins to produce "death substrates" (Miwa et al., 1998).
There has been much evidence supporthg a pathogenic role for inclusions in
neurodegenerative diseases. Aggregation of abnormal and mutant proteins are now suspecteci
of king a key elernent in many neurodegenerative diseases including Aizheimer's disease,
Parkinson's disease and prion diseases. In polyûln-arpanded disorders, inclusions were initialiy
found in animai models of Huntington's disease (IID) and in spinocerebeUar ataxia type 1
(SCAI) (Davies et al., 1997; Skinner et ai., 1997). Inclusions have now also been identified in
the affected neurons of patients with Hunthgton disease, ddetatombral-pallidoluysian atrophy
(DRPLA), SCAl, SCA3, SCA7 and spinal bulbar muscular atrophy (SBMA) (reviewed by
Lunkes and Mendel, 1997; Ross, 1997; Davies et al., 1998; Kim and Tanzi, 1998; Hohberg et
ai., 1998). They have also been found upon expression of polyGln-expanded protein hgments
in various ceil lines and in animal transgenic models (Skinner et ai., 1997; Davies et ai., 1997;
Perez et al., 1998; Myashita er al., 1998; Cooper et al., 1998; Martindaie et al., 1998; Kim et
al., 1999). Increasing evidence implicates proteasome dystiinction in aggregate formation and
polyGh-expanded protein pathogenesis. Proteosorne components and folding chaperone HDJ-
S/HSDJ have been found in aggregates in SCA-I patients and transgenic animal models
(Cummings et al., 1998). Aggregates formed by polyGln-expanded AR were also found to
stain positively for HDJ-2RISDJ (Stenoien et al., 1999). Interestingly, overatpression of the
chaperone protein in ceUs significantly decreased aggregate fomtion by po1yGln-expanded
ataxin- 1 and AR (Cummings et ai., 1998, Stenoien et ai., 1999). Recent evidence, however,
has show that inclusions may not play a direct role in toxicity. In two landmark studies, ceU
death was not found to correlate with inclusion formation. In Sandou et al. (1998),
cotfaflsfection of mutant huntingtin with a dominant-negative forrn of the ubiquitin-conjugating
enryme in striata1 neurons led to enhanceci ceU death but diminished inclusion formation. in
Klement et al. (1998), transge~c mice expressing at;ucin-1 with a deleted seIf-association
region developed ataxia and Puricinje c d neurodegeneraîion without inclusion formation.
In this study, we sought to îùrther elucidate the role of inclusions and their composition
by expressing polyGln-expanded luorescentiy-tagged AR fuson proteins and vinialling
fluorescence patterns by fluorescent miaoscopy. Our resuits showed that both 2OQ ONT) and
50Q (polyGin-expanded), GFP-AR, BFP-AR and GFP-AR-BFP proteins aggregated in
transfected celis. No o v e d ciifferences were seen between the size, location and density of
WT and polyGln--expanded AR aggregates. Toxicity studies were perfonned using the same
celi h e (Sdptoreanu et al. 1999) as in the present study and a significant increase in ce11 death
was found when polyûln-expanded AR was t d ê c t e d in lieu of WT AR. Overali, our
findings suggest that aggregation may not cause cellular toxicity.
Our observations were extended to a motor neuron like ce11 üne, NSC-34 cells. NSC-
34 ceUs are a hybrid newoblastorna and ernbryonic spina! cord ceii h e which makes them an
ideal mode1 to hvestigate SBMA pathogenesis since spinal cord neurons are highly af5ected in
patients. It was found that aggregation was not affècted by cell type indicating WT and
polyGln-expanded AR aggregation must occur by pathways present in both ceii Iines.
The percentage of ceUs containing aggregates was not affecteci by Gln tract length nor
by N- or C-tenninal tagging. AU transfected ceU types showed 20 to 35 % aggregation in the
absence of androgen. The hi& percentage and pattern of aggregation in the COS-1 cells may
be related to overexpression of proteins. In contrast, NSC-34 ceiis do not overexpress proteins.
in Miyashita et UL ( 1 998), a smd percentage of ceils expressing fiill-length mutant or normal
DRPLA were found to contain aggregates. Inclusion formation was however gmtly
potentiated by expression of an N-terminaiiy truncated construct expressing an expanded Gh
tract (Miyashita et al., 1998). In the absence of overexpression, polyGh-expanded proteins
may have an inneased susceptibiiity to aggregation. The observed aggregation in poly-Gin
expanded proteins rnay merely be an auxlerateci version of the aggregation seen with WT AR
protein.
in GFP-AR-BFP transfected cells, aggregates containexi colocaüzed blue and green
fluorescent tags. This aiggests that intact AR is present within the aggregates. These results
wae repeated in a rnotor-neuron iike celi line, NSC-34 cells. Our hdings thus demonstrate
that both te- of AR are found in aggregates fonned in both non-neuronal (COS- 1) and
neuronal (NSC-34) celis. Using irnrnunocytochemistry, we demonstrated that the F39.4.1 (aa
301-319) and rnAblC2 (expanded-Gin tract) epitopes are in the aggregates as well.
Altogeber, Our findiigs niggest that full-length AR can be found in cellular aggregates.
Stenoien et al. (1999) recently found aggregates in HeLa ceus expressing niIl-length normal
and polyGin-expanded AR Previously, aggregates have been found in t d e c t e d ceUs
expressing polyGin-expanded tracts alone or tnincated polyGin-expanded proteins but not fùll
length parent proteins (Miyashita et al, 1998; Cooper et al., 1998; Hackarn et al., 1998;
Moulder et al., 1 999; Martindale et al., 1 998; Paulson et al., 1 997). In Myashita et al ( 1 W8),
few aggregates were found in hii-length WT and polyGln-expanded DRPLA transfeaed c d s
howwer transféction of a polyGin-expanded DRPLA m e n t lachg the N-terminus
significantly increaseù intracellular aggregation. Another study found that expression of a
COOH-temiinal ataxin-3 hgment contauiing an expanded polyGln tract aggregated but fW-
length expanded ataxin-3 did not result in aggregation (Paulson et ai., 1997). In HD,
expression of a 464 containhg fÙU-length huntingtin protein Med to cause nuclear inclusions
while a lOûQ cunstruct demonstrated clear inclusions (Km et al., 1999). In HD, brain
inclusions reacted ody with an N-terminal anti-huntingtin antibody and failed to react with
other huntingtin antibodies suggesting that only portions of the protein may aggregate (DiFiglia
et al., 1997; Becher et al., 1998). It may however be evidence that certain epitopes are
rninirnaüy accessible in the aggregates and therefore cannot be detected by
imrnunocytochemist~. Indeed, in HD, some inclusions in afFiected brauis stained positively for
ubiquitin but not for huntingtin implying a àiierence in antibody strength, epitope accessibiity
or aggregate composition (Gourfinkel-An et al., 1998).
While our results suggest that full-length proteins can aggregate in viw, they may also
reflect recniitment of the parent protein by proteolytic produas. Perez et al. (1998)
demonstrated recruitment of M-length expanded ataxin-3 by a polyGinsxpanded a t h - 3
h p e n t and by both n o d and expanded fiiii-length ataxin-1 proteh, a protein unrelated to
SCA3.
To invesbgate reccuitment of fiiii-length protein by degradative hgments, we looked
for proteolytic products on Western blots and used immunocytochemiaiy on transfected cells.
Using antiidies against AR (F39.4.1) and the expanded Gin tract (1C2), we could visualize
AR tiagments whose fluorescent tags may have been cleaved. While J39.4.1 and 1C2 both
reacted strongly with aggregates, a certain level of non-specinc immunofluorescence outside
the aggregates was observeci. Western analysis of the protein extracts allowed us to hrther
examine the likelihood of protein processhg in the GFP-AR, BFP-AR and GFP-AR-BFP
transfected cds. The major products identifid using the three antibodies were the expeded
fiision proteins. The secondary bands in the F39.4.1 blot may represent non-speci6c bindhg
since the sarne bands were not apparent using the lC2 or anti-GFP antiidies. They may,
however, be indicative of a minimal arnount of proteolytic degradation seen only with F39.4.1
due to the increased strength and specificity of this antibody.
Mbolerone addition (ME3) for 72 hours, led to a significant decrease in aggregate
formation in wiid-type and polyGin-expanded GFP-AR and GFP-AR-BFP transfection
experiments. The addition of hormone alters AR conformation and this in tself may lead to a
physiochemical dierence in the ability of AR-ligand complexes vs. eee AR to undergo
aggregation. Furthemore, hormone exposure lads to a quick translocation of AR to the
nucleus. The nuclear environment may be les conducive to aggregate formation. It is also
possible that hormone addition favored the formation of nuciear aggregates that were near
invisible using fluorescent microscopy and that dense nuclear staining made the resolution of
small nuclear inclusions difficult. Stenoien et al. (1999), documented both androgen-dependent
and androgen-independent AR aggregation. The addition of hormone led to AR cytoplasmic
aggregation within 20 minutes. The length of hormone exposure may account for the diierent
effects of hormone on AR aggregation. When mibolerone was added one hour prior to fixation
instead offor 72 hours, no decrease in aggregation was seen. The addition of androgen caused
the nuclear translocation of any AR that was not in the aggregates but did not impact on
cytoplasmic aggregation (Fig SA). MB was unable to clear any of the existing aggregates and
may have even potentiated an increase in aggregate formation. While short-term hormone
Btposure may increase awegation, longer exposures rnay provide long terrn protection to the
cells and an ovedi decrease in the nurnber of aggregates.
The effea of a general caspase inhibitor, 2-DEVD-FMK, on aggregaîion was also
tested. Several polyGln-tract containing proteins have been shown to be susceptible to cleavage
by caspases. Huntingtin, atrophin-1, a t h - 3 and the AR are substrates for caspases (Goldberg
et ai., 1996; Miyashita et al, 1997; Wellington et ai., 1 998; Miwa et al., 1998) with caspase
cleavage resulting in the fomation of an N-terminal hgment containing the Gh tract. These
N-terminai 6iagments have b e n postulated to bring about aggregation andl or contribute to
cellular toxicity. In one study, inhibition of caspase-8 blocked polyûh-induced cell death
(Sanchez et al 1999). Provocatively, addition of the caspase uihibitor 2-DEVD-FMK
significantly reduced the nuckar and cytoplasmic inclusions fonned by mutant N-terminal
huntingtin wnstmcts (Kim et al., 1999). Our results, however, show that addition of the
general caspase inhibitor Z-DEVD-FMK does not influence aggregation in GFP-AR or GFT-
AR-BFP transfected c d s indicating that AR aggregate fomation may not require cleavage by
a Z-DEVD-FMK-sensitive caspase. Inherent differences in protein structure or in the ceU lines
used rnay explain why caspase inhibition &ected mutant huntingtin aggregation and not AR
aggregation. Our findings are supported by a siudy by Sandou et al.(1998) which demonstrated
that nuclear inclusion formation did not require cleavage by an Ac-DEVD-CHO sensitive
caspase.
Although caspase cleavage rnay not be required for aggregation, it rnay ail1 be required
for polyGln protein-induced cellular toxicity. in our studies, the percentage of cells containhg
aggregates was unaffected by polyGln tract length even though toxicity studies revealed a 2-
fold increase in ceU death when polyûln-expanded AR was expresseci instead of Wï AR
(Sculptoreanu et ai., 1999) . Our findings indicate that inclusion formation rnay not be required
for cellular toxicity and while proteolysis rnay play a role in the pathogenicity of polyGln-
expanded diseases, it rnay not be necessary for protein aggregation.
References
Becher, M.W., Kotzuk, J . k , Sharp, AH., Davies, S.W., Bates, G.P., Price, D.L. and Ross,
C.A. (1 998) Intranuclear neuronal inclusions in Huntington's disease and dentatorubrd
and pallidoluysian atrophy: correlation between the density of inclusions and [T l5
CAG triplet repeat length. Neurobiol. Dis. 4,387-397.
Beitei, L.K., Prior, L., Vasiiiou, D.M., Gottlieb, B., Kaufhn, M., Lumbroso, R, Alvarado, C.,
McGiUi~ay, B., Trifiro, M and Pinsky, L. (1994) Complete androgen insensitivity due
to mutations in the probable a-heiical segments of the DNA-binding domain in the
human androgen receptor. Hum. Mol. Genet. 3,21-27.
Cashman, N-R, Durham, H.D., Blusztajn, J.K., Oda, K., Tabira, T., Shaw, I.T., Dahrouge, S.
and Antel, J.P., (1992) Neuroblastoma x spinal cord (NSC) hybnd ceIl lines resemble
developing motor neurons. Devei. Dyn. 194(3), 209-22 1.
Cooper, J.K., Schilling, G., Peters, M.F., Herring, W.J., Sharp, AH., Kaminsky, Z., Masone,
J., Khan, F-A, Delanoy, M., Borchelt, D.R, Dawson, V.L., Dawson, T.M. and Ross,
C.A (1998) T~ncated N-terminai fragments of huntingtin with expanded g l u t e e
repeats fom nuclear and cytoplasrnic aggregates in ceU culture. Hum. Mol. Genet. 7,
783-790.
Curnmings, C.J., Mancini, M.A, Antaif&, B., DeFranco, D.B., Or, H.T. and Zoghbi, H.Y.
(1998) Chaperone suppression of aggregation and altered subcellular proteasorne
1orali;ration hply protein misfolding in SCAI. Nat. Genet. 19, 148- 154.
David, G., Abbas, N., Stevanin, G., Dum, A, Yvert, G., CanceI, G., Weber, C., hôert, G.,
Saudou, F., Antoniou, E. et al. (1997) Cloning of the SCA7 gene reveals a highiy
unstable CAG repeat expansion. Nat. Genet. 17,65-70.
Davies, S.W., Beardsaii, K., Turmaine, M., DiFiglia, M., Aronin, N. and Bates, G.P. (1998)
Are neuronal intranuclear indusions the common neuropathology of triplet-repeat
disorders with polyglutamine-repeat expansions? Lancet 3 5 1, 1 3 1 - 1 33.
Davies, S.W., Tumiaine, M., Cozens, B.A. DiFiglia, M., Sharp, AR, Ross, C.A,
Scherzinger, E., Wanker, E.E., Mangiarini, L. and Bates, G.P. (1997) Formation of
neuronal intranuclear inclusions underlies the neurology dysfùnction in rnice transgenic
for the HD mutation. Cell90, 537-548.
DiFiglia, M., Sapp, E., Chase, K.O., Davies, S.W., Bates, GR, Vonsattel, J.P. and Aronh, N.
(1997) Aggregation of huntingtin in neuronal htranuclear inclusions and dystrophie
neurites in brain. Science 277, 1990- 1993.
Georget, V., Lobaccaro, J.M., Terouanne, B., Mangeat, P., Nicolas, J - C . and Sultan, C.
(1 997) TrafEcking of the androgen receptor in living ceUs with Nsed green
fluorescent proteh-androgen receptor. Mol. Cel. Endocrinol. 129, 17-26.
Goldberg, Y.P., Nicholson, D. W., Rasper, D.M., Kalchman, M. A, Koide, H.B., Graham,
RK., Brornm, M., Kazemi-Esfajani, P., Thombeny, NA, Vaillancourt, J.P. and
Hayden, M.R (1 996) Cleavage of huntingtin by apopain, a proapoptotic cystehe
protease is modulated by the polyglutarnine tract. Nat. Genet. 13,442-449.
Gouinnkel-An, L, Cancei, G., Duyckaerts, C., Faucheux, B., Hauw, J.J., Trottier, Y., Brice,
A, Agid, Y. and Hirsch, E.C. (1998) Neuronal distribution of intranuclear inclusions in
Huntington's disease with adult onset. Neuroreport 9, 1823-1826.
Hackam, A., Singaraja, T., Weüington, CL., Metder, M., McCutcheon, K., Zhang, T.,
Kalchaman, M. and Hayden, M.R (1998) The influence of huntingtin protein size on
nuclear IOcalization and cellular toxicity. J. Cel Biol. 14 1, 1097- 1 105.
Hohberg, M., Duyckaerts, C., Dun; A, Cancei, G., Gourf'rnkel-An, I., Damier, P.,
Fauchery B., Trottier, Y., Hirsch, E.C., Agid, Y. and Bnce, A (1998)
Spinocerebeiiar ataxia type 7 (SCA7): a neurodegenerative disorder with neuronal
intranuclear inclusions. Hum. Mol. Genet. 7,9 13-9 1 8.
Huntington's Disease Collaborative Research Group (1 993) A novel gene containing a
trinucleotide repeat that is expanded and unstable on Huntington's disease
chromosomes. Ceil 72, 97 1-983.
Igarashi, S., Koide, R, Shimohata, T., Yarnada, M., Hayashi, Y., Takano, H-, Date, H.,
Oyake, M., Sato, T., Sato., A et al (1 998) Suppression of aggregate formation and
apoptosis by transglutarninase inhibitors in d s expressing truncated DRPLA protein
with an expanded polyglutamine stretch. Nat. Genet. 1 8, 1 1 1 - 1 1 7.
Ikeda, H., Masahiro, Y., Satoshi, S., Yoshiya, A, Shuh, N. and Kakinika, A (1996) Expanded
polyglutamine in the Machado-Joseph disease protein induces ceîi death in vitro and in
vivo. Nat. Genet. 13, 196-202.
Imbert, G., Saudou, F., Yvert, G., Devys, D., Trottier, Y., Garnier, J.M., Weber, C., Mandel,
J.L., Cancei, G., Abbas, N. et al (1 996) Cloning of the gene for spinocerebeiîar ataxia
2 r e v d s a locus with hi& sensitivity to expanded CAG/glutamine repeats. Nat.
Genet. 14,285-291.
Kahlem, P., Terre, C., Green, H. and Djian, P. (1 996) Peptides containing glutamine repeats as
substrates for transglutamuiase-catalyzed cross-ünking: relevance to diseases of the
nervous system. Proc. Natl. Acad, Sci. USA 93, 14580- 14585.
Kawaguchi, Y., Okamoto, T., T a n i d , M., Aizawa, M., houe, M., Katayarna, S.,
Kawakami, H-, Nakamura, S., Nishirnura, M., Akiguchi, 1. et al. (1994) CAG
expansions in a novel gene for Machado-Joseph disease at chromosome 14q32.1.
Nat. Genet. 8,221-228.
Kennedy, W.R, Alter, M. and Sung, J.H. (1968) Progressive proximal spinal and bulbar
muscular atrophy of late onset. A sex-iinked trait. Neurology t 8,67 1-680.
Kim, M., Lee, H.-S., LaForet, G., McIntyre, C., Martin, E.J., Chang, P., Kim, T.W., Williams,
M, Reddy, P.H., Tagle, D., Boyce, F.M., Won, L., Heller, A, Aronin, N. and
DiFigiia, M. (1999) Mutant huntingtin expression in clonai striatai d i s : dissociation
of inclusion formation and neuronal sunival by caspase inhibition. J. Neuraci. 19,
964-973.
Kim, T.-W. and Tanzi, RE. (1998) Neuronal intranuclear inclusions in polyglutamine diseases:
nuclear weapons or nuclear fdout? Neuron 2 1,657-659.
Klement, LA, Skiner, P.J., Kaytor, M.D., Y, H., Hersch, S.M., Clark, H.B., Zoghbi, H.Y. and
On, H.T. (1998) Ataxin-1 nuclear localization and aggregation: role in polyglutamine-
inducPd disease in SCAI transgenic rnice. Ce1 95,41-53.
Kobayashi, Y., Miwa, S., Merry, D.E., Kume, A, Mei, L., Doyu, M. and Sobue, G. (1998)
Caspase-3 cleaves the expanded androgen receptor protein of spinal and bulbar
muscular atrophy in a polyglutamine repeat length-dependent manner. Biochem.
Biophys. Res. Commun. 252, 145-1 50.
Koide, R, Ikeuchi, T., Onodera, O., Tanaka, H., Igarashi, S., Endo, K., Takahashi, H.,
Kondo, R, Ishikawa, A, Hayashi, T. et al. (1994) Unstable expansion of CAG repeat
in hereditaxy dentatonbral-pallidoluysian atrophy (DRPLA). Nat. Genet. 6, 9- 13.
Kwb, M.D., Benzow, KA, Bird, T.D., Day, J.W., Moseley, M.L. and Ranum, L.P. (1998)
Rapid cloning of expanded thucleotide repeat sequences fiom genomic DNA Nat.
Genet. 1 8, 72-75.
Kuida, K., Zheng T.S., Na, S., Kuan., C., Yang, D., Karasuyama, H., Rakic, P. and Haveii,
RA (1996) Derrased apoptosis in the brain and premature lethaiity in CPP32-
deficient mice. Nature 3 84(6607), 368-72.
LaSpada, AR, Wdson, E.M., Lubahn, D.B., Harding, AE. and Fisc&& K.H. (1991)
Androgen receptor gene mutations in X-linked spinal and bulbar musailar atrophy.
Nature 3 52, 77-79.
Lescure, A, Lu& Y., Eberhard, D., Jacq, X., Krol A, Gmmmt, I., Davidson, I., Chambon, P.
and Tora, L. (1994) The N-terminal domain of the human TATA-binding protein plays
a role in vansaiption f?om TATA-contahhg RNA polymerase 11 and ID promoten.
EMBO J. 13, 1 166- 1 175.
Li, M., Miwa, S., Yasushi, K., Merry, D.E., Yamamoto, M., Tanaka, F., Doyu, M.,
Hashinime, Y., Fishbeck, K.H. and Sobue, G. (1998) Nuclear inclusions of the
androgen receptor protein in spinal and bulbar musatlar atrophy. Ann. Neurol. 44(2),
249-254.
Li, M., Nakagomi, Y., Kobayashi, Y., Meny, D.E., Tanaka, F., Doyu, M., Mitsuma, T.,
Hashiaime, Y., Fischbeck, K.H. and Sobue, G. (1998b) Nomeurai nuclear hclusions
of androgen receptor protein in spinai and bulbar m u d a r atrophy. Am. J. Pathol.
1 53,695-70 1.
Lunkes, A and Mandel, J:L- (1997) Polygiutarnines, nuclear incIusions and
neurodegeneration. Nat. Med. 3, 120 1 - 1202.
Martindale, D., Hackam, A, Weczorek, A, Ellerby, L., Wellington, C., McCutcheon, K.,
Singaraja, R, Kazemi-Esfajani, P., Devon, R, Kim, SU., Bredesen, D., Tuf'o, F.
and Hayden, M.R (1998) Length of huntingtin and its polyglutamllie tract inûuences
IOcalization and eequency of htraceilular aggregates. Nat. Genet. 18, 150- 154.
Matilla, A, Roberson, E.D., B a S., Morales, J., Armstrong, D.L., Bumght, E.N., On; H.T.,
Sweatt, J.D., Zoghbii H.Y. and Matzuk, M.M. (1998) Mice lacking ataxh-1 display
leaniing deficits and decreased hippocampal paired-pulse facilitation. J. NeUrosci. 1 8,
5508-55 16.
Miyashita, T., Nagao, K., Ohmi, K., Yanagisawa, H, Okarnura-Oho, Y. and Yamada, M.
( 1998) Intracellular aggregate formation of dentatorubral-paliidoluysian atrophy
(DRPLA) protein with the extended polyglutamine. Biochem. Biophys. Res.
Commun. 249,96- 102.
Miyashita, T., Okarnura-Oho, Y., Mito, Y., Nagafùchi, S. and Yamada, M. (1997)
Dentatombral pallidoluysian atrophy (DRPLA) protein is cleaved by caspase-3 during
apoptosis. J. Biol. Chem. 272,29238-29242.
Miwa, K., Asano, M., Horai, R, Iwakura, Y., Nagata, S. and Suda, T. (1998) Caspase 1-
independent I L 1 beta release and inflammation induced by the apoptosis induced
Fas ligand. Nat. Med. 4(11): 1287-1292.
Moulder, K.L., Onodera, O., Burke, J-R, Strittmatter, W. J. and Johnson E.M. Jr. (1 999)
Generation of neuronal intranuciear inclusions by polyglutamineGFP: anaiysis of
inclusion clearance and toxicity as a tiinction of polygîutamine length. J. Newosci. 9,
705-7 15.
Nagafùchi, S., Yanagisawa, H., Sato, K., Shirayama, T., Ohsaki, E., Bundo, M., Takeda, T.,
Tadokoro, K., Kondo, L, Murayama, N. et a1.(19940 Dentatonibral and
paliidoluysian atrophy expansion of an unstable CAG trinucleotide on chromosome
12p. Nat. Genet. 6(1), 14-18.
Okajirna, T., Tanabe, T. and Yasuda, T. (1993) Nonurea sodium dodecyt sulfate-
polyacrylamide gel electrophoresis with high-molarity bders for the sepration of
proteins and peptides. Anal. Biochem. 2 1 1,293-300.
Ordway, J.M., Tailaksen-Greene, S., Gutekunst, C.A, Bernstein, E.M., Cearley, J.A, Wiener,
H.W., Dure, L.S. 4th., Lindsey, R, Hers&, S.M., Jope, RS., Albin, RL. and Detloff,
P.J. (1997) Eaopically expressed CAG repeats cause intranuclear inclusions and a
progressive late onset neurological phenotype in the mouse. Cd 91,753-763.
OIT, H.T., Chung, M X , Banfi, S., Kwiatkowski, T.J., Servadio, A, Beaudet, AL., McCall,
AE., Duvick, L.A, Ranum, L.P. and Zoghbï, H.Y. (1993) Expansion of an unstable
trinudeotide CAG repeat in spinocerebellar ataxia type 1. Nat. Genet. 4,22 1-226.
Paulson, HL., Pereq M.K., Trottier, Y., Trojanowski, J.Q., Subramony, S.H., Das, S.S., Vig,
P., Mandel, J.-L., Fischbeck, K.H. and Pittman, R.N. (1 997) Intranuclear inclusions of
expandeci polyglutamine protein in spinocerebeUar ataxia type 3. Neuron 19,333-
344.
P e r q M.K., Paulson, HL., Pendse, S.J., Saiom, S., Bonini, N.M. and Pittrnan, RN. (1998)
Recmitment and the role of nuclear 1Ocalization in polyglutamuie-mediated
aggregation. J. Cell Biol. 143, 1457-1470.
Perug M.F., Johnson, T., Suzuki, M. and Fmch, J.T. (1994) Glutamine repeats as polar
zippers: their possible role in inhented neurodegenerative diseases. Proc. Natl. Acad.
Sci. USA 91,5355-5358.
Pukt, SM., Nechiporuk A, Nechiporuk, T., Gispert, S., Chen, X.N., Lopes-Cendes, I.,
Peariman, S., Starkman, S., Orozco-Diaz, G., Lunkes, A et al. (1996) Moderate
expansion of a normaiiy biaüelic trinucleotide repeat in spinocerebehr ataxia type 2.
Nat. Genet. 14, 269-276.
Ross, C.A (1 997) Intranuclear neuronal inclusions: a common pathogenic mechanism for
glutamine-repeat neurodegenerative diseases? Neuron 19, 1 147- 1 1 50.
Sinchez, L, Xu, CJ, h o , P., Kalàzaka, A, Blenis, J. and Yuan, J. (1 999) Caspase-8 is
required for ceU death induced by expanded poiygiutamine repeats. Newon 22,623-
633.
Sanpei, K., Takano, H., Igarashi, S., Sato, T., Oyake, M., SaralO, H., Wakisaka, A, Tashko,
K., Ishida, Y., Ikeuchi, T. et al. (1 996) Identification of the spinocerebeliar ataxia type
2 gene using a direct identitication of repeat expansion and cloriing technique,
DIRECT. Nat. Genet. 14,277-284.
Saudou, F., Finkbeiner, S., Devys, D. and Greenberg, M.E. (1 998) Huntingtin acts in the
nucleus to induce apoptosis but death does not correlate with the formation of
intranuclear inclusions. CeU 95, 55-66.
Sai l ptoreanu, A, Abramovici, H., Abdullah, AAR, Bibikova, A, Panet-Raymond, V.,
Frankei, D., Schipper, H.M., Pinksy, L. and Tntiro, M A (1999) Inaeaseâ T-type
ca2' channel activity as a determinant of cellular toxicity in neuronal cell iines
expressing polyglutamine-expanded human androgen receptors. Mol. Ceii. Biochem.
Skinner, P.J., Koshy, B., Cumrnings, C.J., Klement, LA, helin, K., servadio, A, Zoghbi H.Y.
and Orr, HT. (1997) SCAl pathogenesis involves alterations in nuclear ma&
associateci stnictures. Nature 389,97 1-974.
Sobue, G., Hashiaune, Y., Mukai, E., Hirayama, M., Mitsuma, T. and Takahashi, A (1989)
X-linked recessive bulbospinal neuronopathy. A c~mpathological study. Brain 1 12,
209-232.
Stenoien, D.L.,. Curnmings, C.J., Adams, HP., Mancini, MG., PateI, K., DeMartino, G.N.,
Mardi, M., Weigel N. L. and Mancini, M. A ( 1 999) Polyglutamine-expanded
androgen receptors fom aggregates that sequester heat shock proteins, proteasome
cornponents and SRC-1, and are suppressed by the HDI-2 chaperone. Hum. Mol.
Genet. 8, 73 1-74 1.
Trottier, Y., Lutz, Y., Stevanïn, G., Imbert, G., Devys, D., Cancel, G., Saudoy F., Weber, C.,
David, G., Tora, L. et al. (1995) Polyglutamine expansion as a pathological epitope in
Huntington's disease and four dominant cerebeiiar ataxias. Nature 378,403-406.
Warrick, J.M., Paulson, H.L., Gray-Board, G.L., Bui, Q.T., Fischbeck, K.H., Pittman, RN.
and Bonini, N.M. (1998) Expanded polyglutamine protein f o m nuclear inclusions and
causes neural degeneration in Drosophda. Cell93,939-949.
Wehgton, C.L., EUerby, L.M., Hackam, AS., Margolis, RL., Trifiro, M.A, Singaraja, R,
McCutcheon, K., Salvesen, G.S., Propp, S.S., Brornm, M., Rowland, K.J., Zhangm T.,
Rasper, D., Roy, S., Thombeny, N., Pinsky, L., Kakiaika, A, Ross, C A , N~cholson,
D.W., Bredesen, D.E. and Hayden, M.R (1998) Caspase cleavage of gene products
associated with triplet expansion disorders generates truncated fhgments containing
the polyglutamine tract. J. Biol. Chem. 273,9 158-9167.
Zegers, N.D., Claassen, E., Neelen, C., Mulder, E., van Laar, J.H., Voorhorst, MM.,
Berrevoets, CA, Brinkmann, AO., van der Kwast, T.H., Ruizeveld de Wmter, J. A
et aL (1991) Epitope prediction and confinnation for the human androgen receptor:
generation of monoclonal antibodia for multi-assay perfomance foUowing the
synthetic peptide Qrategy. Biochirn. Biophys. Acta 1073, 23-32.
Chapter III
INTRODUCTION TO CHAPTER III The characterization of AR mutations has helped define the structure-fùnction
relationships of the AR and has promoted the understanding of the AR mechanism of
action. Substantial AR analysis has enabled the use of the AR as a mode1 for all steroid
receptors.
AR mutations affecthg receptor homodimerkation underline the importance of
effective dimerization for transactivation by the AR (see chapter 1). Recent evidence
suggests that another important facet of steroid receptor functionality is the
documentation of steroid receptor heterodimers which possess unique functional
properties that may expand the physiological rotes of steroids themselves.
In this paper, we entertain the possible heterodimerization between the androgen receptor
(AR) and the estrogen receptor a @Ra). The long standing physiological interactions
between androgens and estrogens have been well documented.
The characterization of AR-ERa heterodimers may explain this classical
physiological relationship between androgens and estrogens and are in concert with the
recent documentation of colocalization of androgen and estrogen receptors in select
endocrine targeted tissue.
Heterodimerization Between the Andmgen Receptor and Estrogen
Receptor a Mects Their Respective Transactivational Properties
Vaierie panet-~aynond", Lenore K. ~eitel', Leonard ~insk~'", Mark A ~rifiro'".
Lady Davis Institute for Medicai Research, Sir Mortimer B. Davis-Jewish General ~ospital'
and Departments of ~ i o l o d , ~edicine), ~ediatrics~, Human ~enetics', McGdi University,
Montreai, Quebec, Canada.
Proofs and correspondence to :
M.A Trifiro, M.D.
Lady Davis institute for Medicai Research
Sir Mortimer B. Davis- Jewish General Hospital
3755 Cote-Ste-Catherine Road
Montreai, Quebec
Canada H3T 1E2
Fax: (5 14)340-7502
E-mail: [email protected]
A bstract
The physiological interplay of androgen and estrogen action in targeted endocrine tissues
is weii recognized. The biochemicai processes responsible for this interplay have yet to be ftlty
defined.
In this paper, we show novel direct interactions between the androgen receptor (AR) and
estrogen receptor alpha @Ra) by means of the yeast and rnammalian two-hybrid systerns.
These interactions occurred behueen the C-terminai ERa ligand-binding dornain and both the
N-terminal AR transactivational domain and the ftll-length AR Estrogen receptor beta (Em)
did not interact with the AR
DNA CO-transfection studies employing AR E h and Ew expression vectors and AR-
or ER-reporter gene consmicts identifid and measured potential fiinaional effects of putative
AR-ER interactions. Coexpression of AR with ERa decreased AR transactivation by 35%;
coexpression of ERcr with AR decreased transactivation by 75%. Coexpression of AR and
Ew did not significandy modulate AR or Ew transactiation.
in summary, we have shown that speàfic domains of AR and ERa interact physically and
CO-expression transfection studies have revealed fbnctional implications of such interactions.
Introduction Physiologid interplay between androgens and estrogens in various endocrine targets is
weU established (1). ln such targeted cells and tissues, androgens and estrogens typically have
opposing actions. Such counteractive effects have been dexnonstrated in breast (2-6), ovarïes
(7,8), tesbs (9), brain (10-1 6), utenis (1 7- 19) and prostate (20-25). Androgen's and estrogen's
efféctor mechanisms oca~ at the promoter regions of selected gens where individual estrogen
receptors (ER) and androgen recepton (AR) affêct gene expression independently through
theù distinct response elernents. Androgen-estrogen physiological interplay rnay solely reflect
the combined effects of estrogen- and androgen-regulated expression of multiple genes within a
targeted celi. As weli, each respective hormone may contdateraily affect receptor expression
(26-28).
Many nuclear hormone receptors are known to interact as heterocomplexes with
distinctive physiological effects. Thyroid receptors (TR), vitamin D receptors, peroxisome
proWerator-activated receptors (PPAR), retinoid acid receptor (RAR), chicken ovalbumin
upstream promoter-transcription factor (COUP-TFs) and some orphan receptors bind retinoid
X receptor (RXR) or other receptors allowing diverse fùndonality (29-37). More specifically,
ERa recently has been shown to bind directly to hepatocyte nuclear f iaor 4 (HNF4), COUP-
TF, % - RXR and ElU3 (37,38). These interactions have been show to impart unique
propenies to the heteroàïmers, such as altered ligand responsiveness and transactivational
inhibition or activation (39,40).
Hormone-hormone interplay m y also result 60m the contribution of a steroid-receptor
coactivator to steroid-binding. For example, ARA-70 may c o n . cn the androgen receptor an
ability to bind estrogen, thereby allowing for an estrogen effect on androgen targets (4 1).
In this paper, we repon for the first time that classical androgen-estrogen interaction
may be confe~ed by direct interaction between AR and ERa. Spdcally, we have
docwnented interactions between the N-terminus of AR and the ligand-binding domain of the
ERa ushg the yeast and mammalian two-hybrid systems. Comparable N-temiinaVC-terminai
interactions have been shown to be important in detemüning the transcriptional properties of
homdienc steroid-receptor complexes (42-44).
We also used DNA transfection studies in the presence of both androgen and estrogen
reporter gene constmcts to mess the fimaional consequences of AR-ERa interaction.
0 Materials and Metbods
Expression and Reporter Constructs
Yeast two-hybrid co~tstnrcts: Construction of the GAL4DBD-AR(LBD), GAL4DBD-AR
(DBDLBD) and the GAL4AD-AR (fùll-length; FL) have previously been descnbed (45).
AR fragments were cloned into the pAS2-1 vector containing the GAL4 DBD and the
pACT2 vector containing the GAL4AQ both fiom Clontech, Pa10 Alto, California. The
pAS2- 1-ERa(LBD) was kindly provided by Dr. J. White, McGill University. The
segments used in the fiision constructs are shown in Figure 1A.
MammaIian two-hybrid constructs: Fragments of the AR, ERa and ERP cDNAs were
cloned h o the pM vector containing the GAL4DBD, amino acids 1-147 (Clontech, Palo
Alto, California), and into the pVP16 vector containing the VP16 activation domain,
amino acids 41 1-455, shown schematically in Figure 2A. The GAL4DBD-AR(LBD),
GAL4AD-AR(TAD) and the pGAL-Luciferase vectors were generously provided by Dr.
E.L. Yong, National University of Singapore (45). The pVPl6-AR(FL) was made by first
cloning the Accl-AccI fragment of AR20 (arnino acids 1 1-34 1) into the SmaI-digested,
dephosphorylated pAS2-1. The resulting construct was digested with EcoRI and BamHI
and the AR-containing fragment was isolated and ligated to pVP16 that had been similarly
digested. The pVPl6-hARAccI-AccI vector and pSVAR-BHEXE (46) were then cut with
Nad and SaII. The VPl6 vector backbone, with the AccI to Nad AR fiagrnent of AR,
and the Narl-SaII AR piece isolated from pSVARBHEXE, were then ligated to generate
a full-length AR construct in pVP16. The PM-ERa(LBD) was constructed by excision of
the ERuLBD (amino acids 270-595) from the pACT2-ERaLBD plasmid using EcoRI and
SalI and ligating in h e into the pM vector that had been similarly digested. The
GAL4DBD-ER (LBD) was made by digesting the ERQ cDNA (provided by Dr. Sylvie
Mader, Université de Montréal) with KpnI and blunt-ending before fun her digestion wit h
NheI. The pM vector was digested with Sma 1 and Xba 1 before the ERP(LBD) was
inserted.
Reporler Gene Constnrcts: The pSeAP2-basic vector encoding the secreted alkaline
phosphatase reponer gene (Clontech, Palo Alto, California) was digested with HindIiI and
0 KpnI and the ERE fiom the pS2Luc (provided by Dr. Vincent Giguère, McGill University)
that had been similarly digested was inserted in frame into the reporter construct. The
pSeAP2-ERE-MMTV-GH vector was constructed by first digesting the pMMTV-GH
with EcoRl and NdeI thereby releasing an MMTV-LRE fiagrnent containing androgen
response elements (ARE) and the human growth hormone gene (GH). The fkagment was
then blunt-ended and ligated into the pSeAP2-ERE vector which had been cut with Not 1
and blunt-ended. Vectors were sequenced to veriijr the fidelity of the manipulations.
Tranfection Experiments
CV-I cells were maintained in OptiMEM medium (Gibco BRL, Grand Island, NY)
with 5% fetal bovine serum (FBS). Before transfection, 2 million cells were placed
overnight in T25 flasks (Gibco BRL, Grand Island, NY). The following day, the cells
were rinsed with serum-fiee media and the appropriate DNA was rnixed with 35 pl
LipofectAMINE (Gibco BRL, Grand Island, NY) in a total of 2 ml of serurn-fiee media
and allowed to rest on the cells for 6 to 8 hours. M e r lipofection, the LipofectAMZNE
solution was replaced ovemight by 5 ml of phenol-red free Alpha-MEM medium (Gibco
BRL, Grand Island, NY) with 5% stripped serum. Cells tiom each flask were then
trypsinized and placed in 4 separate 3 5mm petn dishes each with 1.5 ml of Alpha MEM
containing 5% stripped serum in the presence or absence of 3 nM Mibolerone (MB)
andor 10 nM 17P-Estradiol ( 1 7P-Ez).
Yeast Two-Hybrid Assays
Hybnd GAL4-AR and GAL4-ERa proteins were expressed in the yeast strain
Saccharomyces Cerevisiae Y 1 90 which contains an integrated UASGAL 1 -Lac2 reporter
gene. Interactions between hybrid proteins results in the formation of a transcriptionally
active cornplex formed by the GAL4-DBD and GALA-AD that induces P-galactosidase
activity. The yeast were transformed using the lithium acetate method according to the
Clontech (Palo Alto, California) protocols. Yeast were grown in standard YPD medium
initially and then plated on double-dropout (-Leu, -Trp) selective medium in the presence
or absence of 1 pM Methyltrienolone (MT) post-transformation. Yeast were harvested
and liquid P-galactosidase assays were performed using ONPG as a substrate according to
standard protocols with the following modification: instead of fieeze-thawing before the
addition of ONPG, yeast were permeabiiized by incubation in a 0.2% sodium lauryl
sarcosinate Z buffer solution (60 nM Na&IP04, 40 n M Na&P04, 10 mM KCI, 1 mM
MasoI, 50 m M j3-mercaptoethanol) (47).
Mammalian Two-Hybrid Assays
1 pg of GAL4DBD-AR or ER construct and 1 pg of VP 16-AR or ER construct
were CO-transfected into cells with 4 pg of the pGAL-LUC reporter. 72 hours post-
transfection, celis were rinsed twice with PBS and then lysed for 15 min with the 1X
luciferase lysis buffer (Promega, Madison, WI). The ceUs were then scraped and ce11 debns
was centrifuged. 20 pl of the supernatant were then added to 100 pl of the Luciferase
reagent to quanti@ luciferase reporter activity (Bio Orbit Luminometer) as described in
Promega protocols. Luciferase activity values were norrnalized for cellular protein,
measured by a Lowry assay, and for transfectional efficiency using pCMV-j3Gal CO-
transfection (48).
Growth Hormone and Secreted Alkaline Phosphatase Assays
To determine the transactivational properties of ERa, ERP and 4 pg of the
pSeAp2-ERE MMTV-GH vector were cotransfected with 1 pg of pcDNA3-hAR (49).
pSGS-hERa (provided by Dr. Wilson Miller, McGill University) or pCMV-hERJ3 or a
combination of the expression vectors. 1 pg of pCMV-PGal was also CO-transfected to
assess transfection efficiency.
An MMTV-GH containhg vector was used to quantitate AR tramactivation.
Growth homone(GH) concentrations were measured according to the protocol in the
Allegro Human Growth Hormone Kit (Nichols Institute Diagnostics, San Juan Capistrano,
CA) with the following modifications: 20 MI of culture media were centrifuged to remove
ce11 debris then assayed for activity by adding them to 1 80 pl of H20 and 50 pl o f '=I-
labeled GH Antibody. To quantitate ERa and ERP transactivational activity, an ERE
fused to a secreted alkaline phosphatase (SeAP) gene was used. SeAP activity was
e measured by first removing 150 pl of culture media and heating to 65°C for 15 minutes.
The samples were then clarified by centfigation and added to an equal volume of 2X
Alkaline Phosphatase reagent (ALP) wit h 2OmM L-homoarginine (Sigma Diagnostics, S t-
Louis, MO). The samples were then lefi at 3 P C and Am5 of the reaction was read in an
automatic plate reader at 30 minute intervals. The SeAP assay was based on Sigma
Diagnostics protocol and Berger et al. (50).
Both the change in SeAP activity and the GH values were noxmalized for cellular
protein, determined through Lowry assays, and for transfection efficiency using P- gaiactosidase cotransfection.
Results
GAL4-AR and GAL4-ERa Do Nol Activate Tramscription of~gdactosidùse On Their
Own in The Yeast Two-Wybrid System
Both ERa and AR contain two independent transcnptional activation domains
(AF-1 and AF-2) that could potentially interfere with a protein interaction assay. We first
verified the inability of the fusion proteins to induce P-galactosidase formation on their
own, without interaction. The ligand-binding domain (LBD), the combined DNA-binding
and ligand-binding domains (DBDLBD), and the fidi-length (FL) human AR cDNA were
cloned in fiame into the pACT2 vector, containing the GAL4 transcnptional activation
domain. The ligand-binding domain of the human ERa was cloned into the pAS2-1 vector
containing the GAL4 DNA-binding domain. Fusion vectors were introduced into yeast
with the corresponding GAL4AD or GAL4DBD vector, and plated on selective media
before performing liquid B-galactosidase assays. Both the GAL4A.D + GAL4DBD-
ERu(LBD) (Fig. 1B; d) and the GALADBD + GAL4AD-AR (not shown) yeast were
found to have negligible P-galactosidase activity. Treatment with 1 pM Methyltnenolone
(MT), a synthetic androgen, 1 p.M 17P-Estradiol (17P-Ez) or both did not promote
GALl-lac2 expression. Since the fusion proteins alone do not exhibit substantial
transactivational properties, we concluded that our system could be used to test possible
interactions between ERa and AR.
AR ànteructs with ERa in a Hormone-lndependenf Fashion in a Yeasî Two-Hybrid
Sysîem
To test the ability of ERa and AR to interact with each other, we cotransformed
yeast with GAL4DBD-ERa(LBD) and GAL4AD-AR (LBD, DBDLBD or FL).
Cotransformations with pTD1 and pVA3 or with GAL4AD-AR and GAL4DBD-
AR(LBD) fusion proteins were used as positive controls. The pTD 1 plasmid encodes a
GUAD-SV40 large T-antigen protein known to interact with the GAL4DBD-murine
p53 fiision protein encoded by the pVA3-I vector. GALAAD-AR(FL) and GAL4DBD-
AR(LBD) are known to positively interact either through LBD-LBD interactions (43,45)
Fig. 1 .Yeast two-hybrid AR-ERa interictive assays. (A) Scheanatic representation of
ERa and AR The key domains of the steroid receptors are shown. Numbers represent the
amino acid positions. (B) The constnicts that were used in the yeast two-hybnd system are
presented. These constnicts were cotransformed into Y 190 yeast which were then
subjected to liquid P-galactosîdase asays. (C) f3-galactosidase activity values are show
for the various construct combinations and in the absence or ptesence of 1 pibf 17B-
estradiol (E) or lpM MethyItrienolone (MT). Values represent an average of 6 assays
each done in duplicate. p-gal values have been corrected for amount of yeast. TDl and
VA3-1 are positive controls for yeast two-hybrid. Statistical analysis was performed using
two-tailed unpaired t-tests. Asterisks represent values found to be statistically significantly
(p=0.05) different then negative controls.
1
hAR I 538 91 9 1 4-8 I
TAD D6D LBO
768 881
1-f GAL4A&AR(LBD)
502 979 I G A L C A ~ ~ { m- 1 LBD ] GAL4APAR(DBDlBD)
919 GAL4-AD +( -. a -- CBD GAL4WAR(FL)
1 1 47 659 919
GAL4DBD-AR(LBD)
m 270 595
E 1 t-l GAL4tX35ER(MCj
or N-terminal TAD-LBD interactions (42,43). Bot h positive controls significantly induced
f3-galactosidase production (Fig. 1B; e,f).
Yeast cotransformed with both the GAL4-ERa(LBD) construct and the GAL4-
AWBDLBD) or GAL4-AR(LBD) failed to significantly activate t ranscnption of the j3-
galactosidase gene (Fig. 1B; a,b). ~ddition of 1 pM of 17P-E2,l pM MT, or both, did not
enhance transactivation of the reporter. In comparison, cotransformed yeast with
GAL4DBD-ERa(LBD) + GAL4AD-ARW) were found to have significantly hi@ tevels
of P-galactosidase activity in triplicate liquid assays (Fig. 1B; c). Fold induction of the
reporter gene was comparable to the level attained with the pS3-SV40 large-T-antigen
cotransfonnation. The AR + AR yeast displayed significantly more P-galactosidase
induction then the AR + ERa transformed yeast in the presence of MT, indicating a
preference for homodimerization by the androgen receptor. In al1 cases, ligands were
found to have no effect on yeast growth on "double-dropout" selective medium.
The presence or absence of different ligands had no significant effect on the
interaction between ERa(LBD) and AR(FL) aithough a small increase in activity upon
addition of 1 pibl MT was noted. 1 pM 17P-Ez neither enhanced nor inhibited the putative
interaction between the tiision constnicts as measured by reporter gene activity. Results
indicate a ligand-independent interaction between full-length AR and the Ligand-binding
domain of ERa but failed to demonstrate any stable interaction between the ERa(LBD)
and the DBDLBD or LBD of the AR.
Mammaiian T w d y b r i â System Confirms Interaction Between ERa and AR
A limitation of the yeast-two hybrid system is that it allows mamrnalian protein
interactions to occur within a heterologous environment. Therefore, we examined the
interaction between ERa and AR in a mammalian cell system using a GAL4-luciferase
reporter activated by the stable interaction between the GALDBD and VP16 chimeras.
CV-1 cells were cotransfected with the reporter, the fusion protein expression vectors and
a vector for B-galactosidase to assess transfection efficiency.
We first investigated the ability of the fusion proteins to induce transcription of the
luciferase reporter gene alone. No significant transactivation was seen in cotransfection
with a parent plasmid lacking an AR or ERa sequence (Fig. 2B; 1). The presence or
absence of 3 nM Mibolerone (MB), or 10 nM L 7P-E2, prolonged hormone exposure (96
hours), or significant increases in hormone concentration failed to produce significant
reporter activity. We concluded that the GAL4-AR or ER fusion constnicts alone do not
have significant GAL4 transactivational properties in the mamrnalian two-hybnd system.
Cotransfection of VP16-AR(FL) and GAL4DBD-AR(LBD) was used as a positive control
and was found to induce significant luciferase activity in the presence of MB (Fig. 2B; 2).
The AR + AR(LBD) interaction caused - 3000-fold induction in the presence of MB and
- 2000-fold induction in the presence of both MB and 17P-E2 as compared to the negative
controls (Fig. 2B; 2). Reporter gene activity was negligible in the absence of androgen. As
expected, addition of 17p-E2 was not able to induce transactivation in the VP16-AR +
GAL4DBD-AR(LBD) cotransfected cells.
We next exarnined the ability of the full-length AR to interact with the ligand-
binding domain of the ERa. Cotransfection of VP 16-AR(FL) + GAL4-ERa(LBD)
induced significant luciferase activity in the presence of steroids. Treatment of the
cotransfected cells with both MB and 1 7P-E2 increased luciferase induction 1 Pfold (Fig.
2B; 3). Addition of 17P-E2 alone did produce significant transcriptional activation.
However, this was less then that seen with both steroids, induction was consistently 4.3-
fold over controls. Interestingly, addition of MB or a complete lack of hormones failed to
induce luciferase production in the AR-ERa transfected CV-1 cells. We also tested
interaction between the transcriptional activation domain (TAD) of AR and the
ERa(LBD) using this system. Cells cotransfected with VP 16-AR(TAD) and GAL4DBD-
ERa(LBD) were measured for luciferase activity under different hormonal conditions
(Fig. 2B; 4). Without steroids, no transcriptional activity was seen; in their presence a
modest though significant increase in luciferase activity was seen.
To examine the specificity of the ERa and AR interaction, we cloned human
ERPGBD) cDNA in fiame into the GAL4 DBD vector and then transfected cells with the
ERB and VPl6-AR@%) constmcts. Results shown in Fig. 2B;S revealed negligible
luciferase activity in the presence of the ERP and AR fusion constnicts. This lack of
transcriptional activation reflects an absence of stable interaction between ERP and AR,
Fig. 2. Mammalian two-hybrid AR-ERa and AR-ER$ interactive assays. (A)
Schematic representation of AR, ERa and ERP constructs used in the marnmalian two-
hybrid assays. Numbers represent arnino acid positions. (B) Results of the mammalian
two-hybrid assays. Luciferase values are shown as optical units. Values represent
averages of four independent experiments in the absence or presence of 3 n M rnibolerone
(MB) or 10- 1 00nM 1 7a-estradiol (E). Fold increases over negative controls are indicated
over the columns. Statisticai analysis was performed using two-tailed unpaired t-tests.
Asterisks represent values found to be statistically significantly (p-O.05) different then
negative controls.
VP16 H TAD VP 1 6-AR(FL) L .#illD'I LBD 1
659 9 19 LBD
GALIDBD-AR(LBD)
and it confirms a specific interaction between ERa and AR. Experiments were repeated
four times to confirm reproducibility. Our results indicate a consistent, specific interaction
between the full-length AR and the LBD of ERu in the presence of MB and 17P-Ez.
ERa Modrtlates AR Transcriptional Activity
To identiQ a potential physiological role to the AR-ERa interaction, we
cotransfected AR and ERu cDNAs into CV-1 celIs with the pSeAP2-ERE MMTV-GH
reporter gene. AR is able to bind the androgen response elements (ARE) in the MMTV
and induce growth hormone production upon androgen addition. A dual AR-ER reporter
constnict was used so that ER activity could be measured in the same transfected cells.
Cells were exposed to different hormonal conditions to examine transcriptional activation
with specific steroids. In the AR-transfected cells, there was substantial androgen-
responsive GH activity (Fig. 3). In the presence of 3 nM MB, there was a 500-fold
increase in responsiveness. 10- 100 nM 1 7P-E2 failed to induce any transcriptional activity
when used alone. The addition of both androgens and estrogens in the media did not alter
the quantity of GH activity seen with androgen alone. When ERa was cotransfected with
Aq a 3 5% decrease in GH activity was seen in the presence of both steroids compared to
the cells treated with MB alone. This decrease in transcriptional activity was statistically
significant and consistent in four independent expenments. These results are supporied by
Kumar (5 1) who found an inhibition of AR-induced transactivation that was ER cDNA
dose-responsive and ranged nom a 43% to an 87% decrease. The decrease in AR-
responsive reporter activity was estradiol dependent (5 1).
Mock transfections and transfections of the double reporter construct alone were
performed as controls (not show). The pSeAP2-ERE MMTV-GH reporter gene alone,
as well as the mock-transfected cells, failed to show induction of GH transactivation under
any homional condition. This suggests an absence of endogenous AR in the cells. ERa
alone also failed to activate GH transcription in the presence of estrogen or androgen
demonstrating that ERu alone cannot induce MMTV-GH.
To examine the specificity with which ERa modulates AR transactivation,
experiments were repeated using ERP. First, we demonarated that cells transfected with
Fig. 3. Impaired AR-induced transactivation by ERa Transcriptional activation of MMTV-GH gene by AR in the presence or absence of ERa and ER@ Values shewwi are the-averagesof four indepenckenx e ~ p e n ~ e n t s : Values are represented in each group as a percentage o f MB-induced GH reporter gene activity. Hormonal conditions include 3nM mibolerone (MB), 10 nM 178- estradiol (E), MB and E or no hormone. Values were conected for transfection eficiency ($-galactosidase assays) and for protein levels (Lowry assays). The asterisk represents a value that is significantly different then controls according to an unpaireci, two-tailed T-test (p-0.05).
human ERP cDNA alone do not induce MMTV-GH. When ERB was cotransfected with
a MB-induced GH activity was unaltered by the presence or absence of 17p-E~ (Fig. 3).
Moreover, AR transcriptionai activity was unaffected by the presence of ERP and there
was no decrease in GH activity when 17B-Ez was added. These results indicate that MB-
induced AR transactivation decreases only in the presence of the ERa/l7P-E2 complex
and may be a result of a direct AR-ER interaction. The inability of ERa to activate GH
transcription under any hormonal condition, combined with the lack of difference in GH
transcription between Ml3 and MB+17P-E2 treated AR-transfected celk, irnplies that the
decrease of activity is ERa rnediated.
AR Modulates ERa Transcriptional Activity
The same transfected cells used to investigate AR transactivational ability were
also used to measure ERa's ability to stimulate transcription of SeAP. As described, the
reporter used in the GH assays also contained an estrogen response element (ERE)
attached to a secreted alkaline phosphatase gene (SeAP). Similar controls were performed
and phenol-red fiee media with stnpped serum was used to ensure estrogen depletion in
the media. Both ERa- and ERB-transfected cells induced SeAP activity in the presence of
1 7f3-E2 (Fig.4). Interestingly, ERa was also minimally activated by addition of MB. When
androgen was added to the 17P-E2-treated cells, ERa showed a rnodest decrease in SeAP
activity compared to estrogen addition alone (Fig. 4). ERB showed a modest decrease
under the same conditions. The ability of AR to induce SeAP expression was also
exarnined and was found to be comparable to rnock-transfected and reponer-alone-
transfected cells. Addition of 3 nM of MB or 10 n M of 17P-EZ did not increase the basal
levels of SeAP activity seen in the controls. Even increases in 17P-Ez levels, from 10 nM
to LOO nM, failed to induce SeAP activity in the mock-, reporter- or AR-transfected cells.
The lack of SeAP activity in the reporter transfected cells demonstrates that endogenous
ER cannot substantially activate the reporter even in the presence of 100 n M of 17P-E2.
The effect of AR on ER transactivation was then exarnined by cotransfection of the two
expression plasmids. A substantial inhibition of ERdI7P-E2 induced SeAP activity was
Fig. 4. tm paired ERa-induced transactivation b y AIL Transcriptional activation of pS2-SeAP by ERa or ERf3 in the presence or absence of AR. Values show are the averages of three independent experiments. Values are represented in each group as a percentage of E-induced SeAP activity. Transcriptional activation was tested in the presence or absence of 3nM miblerone (M.) or 1 OnM 17fkstradiol (E ). Values were conected for tram fection efficiency (f3-galactosidase assay s) and for protein levels (L,owry assays). The asterisk represents a value that is significantiy different then controls according to an unpaired, two-tailed T-test (p-0.05)-
seen when AR was cotransfected and Ml3 was added: there was a 74% reduction in the
fold increase of SeAP activity compared to 17p-E~ alone treated cells. A smaller,
insignificant decrease in 17P-E2 fold induction was seen in ERP/AR transfected cells upon
addition of both hormones. Overall the results indicate an --dependent decrease in
Ez- 17PERa-induced transcriptional activity that is both significant and specific to ERa.
Discussion
Multiple examples of physiological interplay between androgens and estrogens
have led us to hypothesize that there may be a direct interaction between the androgen
receptor and the estrogen receptor. Heterodimerization between members of the steroid
receptor superfarnily is not uncornmon and provides one basis for the diversity of steroid
signaling pathways (39). For instance, the 9-cis retinoic acid X receptor (RXR) is known
to serve as a comrnon partner for several nuclear receptors including the retinoic acid
receptor (RAR), the thyroid hormone receptor (TR), the vitamin D receptor (VDR) and
the peroxisome proliferator-activated receptor (PPAR), and many other receptors once
thought to bind only as homodimen are now known to interact with other rnembers of this
diverse farnily of transcnptional regdatory proteins (29-36). In this study, we used the
yeast and marnmalian two-hybrid systems to explore direct interaction between the
androgen and estrogen receptors.
AR-ER Two-Hybrid Studies
In both the marnrnalian and yeast two-hybrid interactive protein assay systerns, the
ligand-binding domain of ER was used to demonstrate AR-ER interactions. The ligand-
binding domain (LBD) of the estrogen receptor is known to contain potential interactive
domains (38,52). For example, fiision of the ERaCBD) to the myc oncogene created a
protein whose activity was then homonally regulated (53). In another study, the isolated
ERa-LBD was show to bind ligand, dimerize and undergo conformational changes
associated with cooperative ligand binding, comparable to full-length ERa (52). In
addition, ERa(LBD) was proven to interact with a subset of nuclear receptors including
m 4 , TR, RGR, RXR, and ERP (38)-
We considered that the AR LBD may be sufficient for interaction with ERu LBD.
We also wanted, however, to examine potential amino (N) tenninalcarboxy (C) terminal
interactions between the two receptors. N-C terminal interactions have been shown to be
critical for full transcription capabilities of homodimeric AR interactions (42-44). It was
postulated that ftnctionai N-C terminal interactions may also occur between AR and ER.
As such, a fùll-length AR construct was used in both two-hybrid systems in addition to
LBD, DBDLBD and the AR transcriptional activation domain (TAD) GAL4 fiision
constructs. In the yeast two-hybrid experiments, a strong interaction was seen between the
fiII-length AR and the ERa(LBD) whereas AR(LBD) and AR@BDLBD) failed to
interact with ERa(LBD). The lack of reporter gene activity of the latter two may, indeed,
reflect a lack of direct physical interaction between the respective steroid receptor ligand-
binding domains; however, it may also indicate that the AR must be in a conformation
appropriate for a fùll-length receptor in order for ERa to properly interact with it. For
instance, the heterodimerization interface may be more readily available in fùll-length AR.
The interactions between the various AR and ER domains described were also tested
under different hormonal conditions. In the mammalian system, the interaction(s) (were)
was potentiated by addition of both androgens and estrogens. This may suggest that within
the confines of the yeast system, there may be coactivators, interfering proteins or
conformationaI differences that negate the usual ligand-dependence of the interaction.
The GAL4 fùsion control constructs were tested alone to confim their inability to
activate reporter transcription. None of the vectors was found to have activity in the
absence of interaction. The inability of the GAL4DBD-ERa(LBD) to induce transcription
of the reporter gene when coexpressed with GAL4AD reflects a fitnctional failure of the
fusion protein to activate transcription, even though it can bind the GAL4 response
elements upstrearn of the reporter gene. It also implies that the AF-2 activation domain
contained within the ligand-binding domain of ERa does not confer a transcriptional
activation potential ont0 the fusion protein. These results are supported by those of Wang
et al. who found no transactivation by full-length ERa-GAL4 fiision proteins when
transfonned singly into yeast (54). The GAL4AD-AR does not interact with the
GAL4DBD alone and therefore fails to activate transcription of P-galactosidase. The
inability of the fision proteins to activate the yeast NO-hybrid system on their own
validate AR-ERa interactions.
Potential AR-ER interactions were fùrther investigated in mammalian cells. Our
findings demonstrated an unequivocal interaction between AR hll-length and ERa(LBD)
that was dependent on steroids. Although addition of 17j3-E2 did cause some reporter
activation, the AR-ER interaction was greatly enhanced by the presence of both MI3 and
17&EZ. One interpretation of these results is that the ligand dependence reflects a need
for ligand-induced conformational changes prior to heterodimerization. A previous report
established that heterodimer formation between TR and RXR was hormone-mediated
(36). While some interfaces may be readily available for interaction when expressed in
yeast, competing chaperones, coactivators and steroids rnay mask these same interfaces in
unliganded proteins in mammalian cells. Steroids rnay also impact on the dissociation
kinetics of the fiision protein dimers in marnmalian cells. Addition of 17P-Ez was found to
significantly decrease the rate of dissociation of ER(LBD) dimers in gel filtration
chromatography (52). Differences in protein half-lives could explain the differences in
interaction ligand-dependence between the systems: the increases in AR and ER halGlife
perpetuated by steroid addition would be necessary for a stable interaction in mammalian
cells but not in yeast.
Interestingly, a weak but persistent interaction between the AR(TAD) and
ERa(LBD) fùsion constructs was also demonstrated in the mammalian two-hybrid system.
This implies a direct N-C terminal interaction between AR and ERa. The weakness of the
interaction rnay be due to a decrease in heterodimerization interface availability in the AR
TAD truncated construct as opposed to full-length. An altemate hypothesis is that ER rnay
interact with multiple sites on the AR, only one of which is found in the AR (TAD), and
these sites may only be accessible d e r hll-length protein folding.
AR(FL) is known to interact strongly with AR(LBD) (42-44); therefore it was
used as a positive control in this study. Comparatively, the interaction between ER(LBD)
and &Il- length AR was found to be approximately 100-fold weaker then between AR(FL)
and AR(LBD). This difference is not surprising and rnany factors rnay contribute.
Androgen recepton are known to interact through both C-C terminal interactions (43,45)
and through N-C terminal interactions. With multiple, specific interfaces interacting, it is
not surprising that AR homodimerization is stronger then heterodimerization with ER.
Also, there rnay be differences in either the stability or the expression of the AR and ER
fusion proteins. Several attempts at Western Blotting were unsuccessfùl due to the low
expression of proteins in CV-1 cells. Attempts to repeat experiments in COS-1 cells to
achieve higher receptor expression levels were also hitless due to the ability of the
GAL4-ER and GAL4-AR fùsion constmcts to activate the luciferase reporter
independently in control experiments. The sarne fusion constmcts did not cause
transcription of the reponer when transfected into CV-1 cells with their empty GAL4
vector counterpart, either GAL4AD or GAL4DBD. Presumably, the ER and AR fusion
proteins interact with proteins or factors containing transcriptional properties that are
found in COS- 1 cells, and not CV- 1 cells thus transactivating the GAL4 luciferase. These
findings underline the importance of the effects of different environrnents on interaction
studies.
To examine the specificity of AR-ERa interaction, we exploited a recently cloned
novel estrogen receptor, ERP (55-58). ERP is highly homologous to ERa; it is able to
bind estradiol with similar affinity and activate gene transcription through estrogen
response elements (58). It also been show to dimerize with ERa (38,60,6 1). When used
in the same experiments as ERa, the ERP(LBD) was shown to have minimal interaction
with full-length a level of reporter activation similar to that of the negative controls.
Our findings suggest a stable, distinct and specific interaction between AR and ERu that
has potential physiological ramifications. It is possible for example, that in organs and
systems where al1 three receptors are located, ERP may control estrogen signaling since it
would not be Bected by the high levels of AR expression.
AR-ER Coîmnsfectioon Erperimenîs
To fùrther examine the fùnctional implications of ERa-AR interaction, we
cotransfected AR and ERa expression vectors and examined the impact on their
respective transactivation systems. ERa modulated AR transcriptional activity, supporting
our findings of a direct interaction between AR and ERa steroid receptors. The effect on
AR'S activity was 17P-estradiol-dependent. Importantly, when transfected alone, 17P-E2
did not cause transactivation of the AR-regulated gene, nor did it alter the amount of
growth hormone produced when used in combination with MB. Only addition of MI3
caused reporter gene transactivation, proving that ttis system is specifically regulated by
androgens. Estradiol had no effea on the AR system unless ERa was cotransfected.
Taken together, these data suggest that ERa does not afFect ARE binding, but rather is
directly interacting with the AR protein. Endogenous AR was not found to induce any
substantial amount of reporter since GH levels in the reporter-transfected cells were
minimal. Overail, we demonstrated that addition of 17B-E2 and ERa caused a substantial
modulation of MB-induced AR transactivation. Our findings are supported by Kumar et
al. (5 1) who also found a dose-dependent decrease of AR transcriptional activity when ER
was coexpressed in the presence of estradiol. Transcriptional inhibition of the
progesterone and glucocorticoid receptors has also been noted when ER was
cotransfected (62).
CV-1 cells were used in the present expenments because of their low expression
(15 finoilmg protein) of transiently transfected vector proteins. With low AR and ER
expression, squelching should be minimal or nonexistant (binding studies and Western
blotting data not show). In addition, modulation of AR transactivation was not seen
when al1 experiments were repeated with ERB. Coexpression of ERP and AR in the
presence of both androgens and estrogens did not show any decrease or inhibition of AR
target gene transcription. If squelching were a major factor, it could be expected a
receptor as simiIar to ERa as ERP would have the same modulatory effect on
transactivation by AR.
In addition to investigating the effect of ER on AR activity, we aIso sought to
study the modulation of ER transcriptional activity by AR. The use of a double reponer
construct allowed us to examine the effect of AR coexpression on ER transcriptional
activity in the sarne cells as those used to investigate AR transactivational competence.
Substantial transcriptional inhibition of 1 7P-E2/ERa induced activity was demonstrated by
the MWAR cornplex. Estradiol induced ERa activity was suppressed 74% when AR was
coexpressed in the presence of androgen. AR expression alone in the presence of
hormones did not activate the reporter gene demonstrating that AR cannot bind the ERE.
The reporter alone failed to induce SeAP expression under any hormonal condition,
revealing the absence of endogenous ER activity. Dextran-charcoal stripped semm and
phenol red-free media were used to ensure that the medium was tmly estrogen-free. When
experiments were repeated with ERP, AWMB caused an insignificant decrease in
transcriptional activity; it did not approach the 74% reduction in ERa activity under the
same conditions.
Transcnptional suppression as a result of heterodimerization is not uncommon in
the nuclear receptor superfarnily. Cotransfection of TR, RAR or RXR with ER caused
either down- or up-regulation of ER transactivation (38). In Forman et al. (39), the
unliganded thyroid receptor and RAR LBDs both conferred transcriptional inhibition on
the retinoic acid X receptor LBD. In the same paper, RXR was shown to suppress Nunl
constitutive activity after heterodimerization.
Overall, Our fùnctional data correlate well with our interactive studies. The
interaction between ERa and AR is specific and has functional implications that are both
reproducible and significant. Our results showed that the ERa-AR interaction is
substantially weaker then the AR-AR interaction indicating that AR homodimerization
may have more fùnctional relevance then heterodimerization in human physiology.
However, both heterodimerization and homodimerization likely play important and
complimenting functions is steroid signalling. The ERa-AR heterodimer has unique
propetties and seems to alter the transcriptional efficacy of its modulatory units. Our
findings also demonstrate a lack of interaction between ERP(LBD) and AR. These results
may provide one explanation for the phenomena of the physiological interplay between
androgens and estogens. A direct interaction between the receptors allows an additional
level of control and adds to the increasing cornplexity of steroid signaling pathways. It
may also indicate that the novel estrogen receptor subtype P is the mediator of estradiol-
induced activation that is androgen-resistant.
Acknowledgmenb
We wish to thank Drs. E.L. Yong for his kind gift of the mamrnalian two-hybrid
vectors, J. White for the PAS 1-ER(LBD) vector, Sylvie Mader for the ERP expression
vector and Wilson Miller for the ERa expression vector. This work was supported by
grants fiom Fonds pour la Formation de Chercheurs et l'Aide a la Recherche and the
Medical Research Council of Canada.
Bibliography
Kreitmann B, Bayard F. (1979). Androgen interaction with the oestrogen receptor
in human tissues. J Steroid Biochem 1 1, 1589- 1595.
Thijssen IH, Poortman J, Schwart F. (1975). Androgens in postrnenopausal breast
cancer: excretion, production and interactions with estrogens. J Steroid Biochem
6, 729-734.
Dauvois S, Li SM, Martel C, Labrie F. (1989). Inhibitory effect of androgens on
DMBA-induced mamrnary carcinoma in the rat. Breast Cancer Res Treat 14, 299-
306.
Simard J, de Launoit Y, Haagensen DE, Labrie F. (1992). Additive stimulatory
action of glucocorticoids and androgens on basal and estrogen-repressed
apolipoprotein-D messenger ribonucleic acid levels and secretion in hurnan breast
cancer cells. Endocrinology 13 0, 1 1 1 5- 1 12 1.
Burak WE Jr., Quim AL, Farrar WB, Brueggeneier RW. (1997). Androgens
influence estrogen-induced responses in human breast carcinoma cells through
cytochrome Pd50 aromatase. Breast Cancer Res Treat 44, 57-64.
Liao DZ, Pantazis CG, Hou X, Li SA. (1998). Promotion of estrogen-induced
mammasy gland carcinogenesis by androgen in the male Noble rat: probable
mediation by steroid receptors. Carcinogenesis 19, 2 1 73-2 180.
Lee HT, Bahr JM. (1990). Inhibition of the activities of Pl50 cholesterol side-chah
cleavage and 3 beta-hydroxysteroid dehydrogenase and the amount of P ~ I O
cholesterol side-chah cleavage by testosterone and estradiol-17 beta in hen
granulosa cells. Endocrinology 126, 779-786.
Ghenevich S, Poutanen M, Tapanainen J, Vihko R (1994). Hormonal regulation
of rat 17 beta-hydroxysteroid dehydrogenase type 1 in cultured rat granuloso celis:
effects of recombinant follicle-stimulating hormone, estrogens, androgens, and
epidermal growth factor. Endocrinology 13 5, 1963- 197 1.
Moger WH. (180). Studies on the interaction of 5 alpha-androstane-3 beta, 17
beta-di01 with the testicular estrogen receptor. Cdn J Physiol Pharmacol 58, 124-
128.
Sheridan PJ, Sar M, Stumpf WE. (1974). Interaction of exogenous steroids in the
developing rat brain. Endocrinology 95, 1749- 1753.
Titova GV, Kliueva NN. (1977). On glutamate dehydrogenase interaction with sex
hormones. Biokhimiia 42,635-63 8.
Dohanich GP, Clemens LG. ( 1 983). Inhibition of estrogen-activated sema1
behavior by androgens. Hormones & Behavior 17,366-373.
Gooren L. (1989). Androgens and estrogens in their negative feedback action in
the hypothalamo-pituitary-testis axis: site of action and evidence of their
interaction. J Steroid Biochem 33, 757-76 1.
Clay CM, Keri RA, Finicle AB, Heckert LL, Hamernik DL, Marschke KM, Wilson
EM, French FS, Nilson M. (1993). Transcriptional repression of the glycoprotein
hormone alpha subunit gene by androgen may involve direct binding of androgen
receptor to the proximal promoter. J Bi01 Chem 268, 13556-1 3564.
Borski RI, Tsai W, DeMott-Friberg R, Barkan AL. (1996). Regulation of somatic
growth and the somatotropic axis by gonadal steroids: primary effect on insulin-
like growth factor 1 gene expression and secretion. Endocrinology 137, 253-259.
Roselli CE, Abdelgadir SE, Resko JA. (1997). Regulation of aromatase gene
expression in the adult rat brain. Brain Res Bull 44, 35 1-357.
Jellinck PH, Newcombe AM. (1983). Androgen receptor-mediated inhibition of
estrogen-induced uterine peroxidase. J Steroid Biochem 19, 17 1 13 - 17 1 7.
Leiva MC, Lyttle CR, Jellinck PH. (1 99 1). Complement C3 synthesis, peroxidase
activity and eosinophil chemotaxis in the rat uterus: effect of estradiol and
testosterone. Mol Ce11 Endocnnol 8 1, 105- 1 1 1.
Arao Y, Miyatake N, Yamamoto E, Usiku H, Matsuura Ninorniya Y,
Masushige S, Hasegawa T, Kato S. (1996). Steroid hormones differentially induce
transcription of the chicken ovalburnin gene, but stabilize the mRNA with the sarne
half-life. J Biochem 120, 7 10-7 1 5 .
Krieg M, Bartsch W, Thomsen M, Voigt KD. (1983). Androgens and estrogens:
their interaction with stroma and epithelium of hurnan benign prostatic hyperplasia
and normal prostate. J Steroid Biochem 19, 155-161.
21 - Martikainen PM, Makela SI, Santti RS, Harkonen PL, Suominen JJ. (1 987).
Interaction of male and female sex hormones in cultured rat prostate. Prostate 1 1,
29 1-303.
22. Hansen RJ, Jungerma~ K. (1987). Sex differences in the control of glucosed-
phosphate dehydrogenase and 6-phosp hogluconate dehydrogenase. Interaction of
estrogen, testosterone and insulin in the regdation of enzyme levels in vivo and in
cultured hepatocytes. Biol Chem Hoppe-Seyler 3 68, 95 5-962.
23. Kiel DP, Baron JA, Plyrnate SR, Chute CG. (1989). Sex hormones and
lipoproteins in men. Am J Med 87, 35-39.
24. Yu YM, Domene HM, Sztein J, Counts DR, Cassorla F. (1996). Developmental
changes and differential regulation by testosterone and estradioi of growth
hormone receptor expression in the rabbit. Eur J Endocrinol 135, 583-590.
25. Croston GE, Milan LB, Marschke KB, Reichman M, Briggs MR. (1997).
Androgen receptor-mediated antagonism of estrogen-dependent low density
lipoprotein receptor transcription in cultured hepatocytes. Endocnnology 138,
3 779-3 786.
26. Puolin R, Simard J, Labrie C, Petitclerc L, Dumont M, Lagace L, Labrie F. (1989).
Down-regdation of estrogen receptors by androgens in the ZR-75-1 human breast
cancer ce11 line. Endocrinology 125, 392-399.
27. Tetsuka h4, Hillier SG. (1996). Androgen receptor gene expression in rat
granulosa cells: the role of follicle-stimulating hormone and steroid hormones.
Endocrinology 137,43924397.
28. Adesanya-Farnuyiwa 00, Zhou J, Wu Gy Bondy C. (1999). Localization and sex
steroid regulation of androgen receptor gene expression in rhesus monkey utems.
Obstet GynecoI 93, 265-270.
29. Yu VC, Delsert C, Andersen B. Holloway JM, Devary OV, Near AM, Kim SY,
Boutin JM, Glass CK, Rosenfeld MG. (1991). RXRB: a coregulator that enhances
binding of retinoic acid, thyroid hormone, and vitamin D receptors to their cognate
response elements. Ceil 67, 125 1 - 1266.
Kliewer S A Umesono K, Mangelsdorf DJ, Evans RM. (1992a). Retinoid X
receptor interacts with nuclear receptors in retinoic acid, thyroid hormone and
vitamin Dj signaling. Nature 355, 446-449.
Kliewer SA, Umesono K, Noonan DJ, Heyman Rq Evans RM. (1992b).
Convergence of 9 4 s retinoic acid and peroxisorne proliferator signaling pathways
through heterodimer formation of t heir receptors. Nature 3 5 8, 77 1-774.
Marks MS, Hallenbeck PL, Nagata T, Segars IH, Appella E, Nikodem VM, Ozato
K. (1992). H-2RUBP (RXRD) heterodimerization provides a mechanism for
cornbinatonal divenity in the regdation of retinoic acid and thyroid hormone
responsive genes. EMBO J 1 1, 14 19- 143 5.
Bugge TH, Pohl J, Lomoy O, Stunnenberg HG. (1992). RXRa, a prorniscuous
partner of retinoic acid and thyroid hormone receptors. EMBO J 1 1, 1409- 14 18.
Zhang XK, Hoffinam B, Tran PB, Graupner G, Pfahl M. (1992). Retinoid
receptor is an auxiliary protein for thyroid hormone and retinoic acid receptors.
Nature 355,441446.
Issemann 1, Prince lU, Tugwood ID, Green S. (1993). The retinoid X receptor
enhances the function of the peroxisorne proliferator-activated receptor.
Biochimie 75,25 1-256.
Collingwood TN, Butler 4 Tone Y, Clifion-Bligh RI, Parker MG, Chatterjee
VKK. (1997). Thyroid hormone-mediated enhancement of heterodimer formation
between thyroid hormone receptor l3 and retinoid X receptor. I Bi01 Chem 272,
13060-13065.
Seo1 W, Chung M, Moore DD. (1997). Novel receptor interaction and repression
domains in the orphan receptor SHP. Mol Ce11 Biol 17, 7 126-7 13 1.
Lee S-K, Choi H-S, Song M-& Lee M-O, Lee W. (1998). Estrogen receptor, a
common interaction partner for a subset of nuclear receptors. Mol Endocrinol 12,
1184-1 192.
Forman BM, Umesono K, Chen J, Evans RM. (1995). Unique response pathways
are established by allostenc interactions among nuclear hormone receptors. Cell
81, 541-550.
Seol W, Hanstein B, Brown M, Moore DD. (1998). Inhibition of estrogen receptor
action by the orphan receptor SHP (short heterodimer partner). Mol Endocrinol
12, 1551-1557.
Yeh S, Miyamoto H, Shima H, Chang C. (1998). From estrogen to androgen
receptor: a new pathway for sex hormones in prostate. Proc Nat1 Acad Sci USA
95, 5527-5532.
Langley E, Zhou 2-X, Wilson EM. (1995). Evidence for an anti-pardiel
orientation of the ligand-activated human androgen receptor dimer. J Bi01 Chem
270, 29983-29990.
Doesburg P, Kuil CW, Berrevoets C& Steketee K, Faber PW, Mulder E,
Brinkmann AO, Trapman J. (1997). Functional in vivo interaction between the
amino-terminal, transactivation domain and the ligand-binding domain of the
androgen receptor. Biochemistry 36, 1052- 1064.
Langley E, Kemppainen Jk Wilson EM. (1998). Intemolecular NH2/carboxyl-
terminal interactions in androgen receptor dimenzation revealed by mutations that
cause androgen insensitivity. J Biol Chem 273,92- 10 1.
Ghadessy FJ, Lim J, Abdullah AAR, Panet-Raymond V, Choo CK, Lumbroso R,
Tut TG, Gottlieb B, Pinsky L, Trifiro Mq Yong EL. (1999). Oligospermic
infenility associated with androgen receptor mutation that reduces DNA-binding,
and disrupts interactions between domains and with the coactivator, TIF2. J Clin
Invest 103, 15 17- 1525 .
Beitel LK, Kazemi-Esfarjani P, Kaufman M, Lumbroso 9 DiGeorge AM,
Killinger DW, Trifiro Pinsky L. (1994). Substitution of arginine-839 by
cysteine or histidine in the androgen receptor causes different receptor phenotypes
in cultured cells and coordinate degrees of clinical androgen resistance. J Clin
Invest 94, 546-554.
Kippert F. (1995) A rapid pecmeabilization procedure for accurate quantitative
determination of beta-galactosidase activity in yeast cells. FEMS Microbiol Letts
128-20 1-206.
Sarnbrook J, Fritsch EF, Maniatis FU 1983 MolecuIar Cloning- A Laboratory
Manual, ed.2. Cold Spring Harbor, pp 16.57, 16.66-16.67.
Abdullah &Ut, Tnfiro MA, Panet-Raymond V, Alvarado C, de Tourreil S,
Frankel D, Schipper HM, Pinsky L. (1998). Spinobulbar muscular atrophy:
polyglutamine-expanded androgen receptor is proteolytically resistant in vitro and
processed abnomally in transfected cells. Hum Mol Genet 7, 379-384.
Berger J, Hauber, J, Hauber R, Geiger R, Cullen BR. 1988. Secreted placental
alkaline phosphatase: a powerful new quantitative indicator of gene expression in
eukaryotic cells. Gene 66: 1- 10.
Kumar MV, Leo ME, Tindall DJ. (1994). Modulation of androgen receptor
transcnptional activity by the estrogen receptor. J Androl 1 5, 534-542.
Brandt ME, Vickery LE. (1997). Cooperativity and dimerization of recombinant
human estrogen receptor hormone-binding domain. J Bi01 Chem 272, 48434849.
Eilers M, Picard D, Yamamoto Bishop M. (1989). C h e r a s of myc oncoprotein
and aeroid receptors cause hormone-dependent transformation of celis. Nature 340,
66-68.
Wang H, Peters GA Zeng X, Tang M, Ip W, Khan SA (1995). Yeast wo-hybnd
system demonstrates that estrogen receptor dimerization is ligand-dependent in vivo. J
Bi01 Chem 270,23322-23329.
Kuiper GGJM, Enmark E, Pelto-Huikko M., Ndsson S, Gustafsson J-A (1996).
Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc Nat1
Acad Sci USA 93, 5925-5930.
Mosselman S, Polrnan J, Dijkema (1996) ERD: identscation and characterization of
a novel human estrogen receptor. FEBS Lett. 392,4943.
TremMay, GB, Tremblay 4 Copeland, NG, Gilbert, DJ, Jenkins, NA, Labrie, F,
Giguere V. (1997) Cloning, Chromosomal L o a h t i o n and Functional Analysis of the
Murine Estrogen Receptor Mol. Endocrinol 1 1,3 53-3 65.
Petterson K, Grandien Y Kuiper GGJM, Gustat'sson J-A (1997). Mouse estrogen
receptor D forms estrogen response elernent-binding heterodimers with estrogen
receptor a. Mol Endocrinol 1 1, 1486- 1496.
59. Kuiper GGJhi Cdsson B, Grandien K, Emark E, Haggblad J, Ndsson S, Gustafsson
J-A (1997). Cornparison of the ligand bhding specificity and transcript tissue
distribution of estrogen receptors a and B. Endocrinology 138, 863-870.
60. Pace P, Taylor J, Suntharaiingam S, Coombes RC, Ah S. (1997). Human estrogen
receptor L) binds DNA in a rnanner s h d a r to and dimerizes with estrogen receptor a. J
Bi01 Chem 272,25832-2583 8.
61. Ogawa S, houe S, Watanabe T, Hiroi H, ûrimo 4 Hosoi T, Ouchi Y, Muramatsu M.
(1998). The cornplete primary structure of hurnan estrogen receptor B (h.ERB) and its
heterodiierization with ER a in vivo and in vitro. Biochem Biophys Res Commun
243, 122-126.
62. Meyer M-E, Gronemeyer H, Turcotte B, Boquel M-T, Tasset D, Chambon P. (1989).
Steroid hormone recepton compete for factors that mediate their enhancer fùnction.
CeU 5 7 , 4 3 3 4 2 .
1 have descnbed the following novel findings:
1) A missense mutation in the LBD of the AR impacted on AR dimerkation but did not
affect ligand-binding. We further showed that this mutation decreased interaction with an
AR-coactivator TIF2. The decrease in mutant receptor homodimerization and interaction
with TE2 explain the reduced transactivity of the mutant AR and contnbute to the clinical
infertility seen in carriers.
2) Results using poiyûln-expanded AR show for the first time that fiill-length AR can be
found in aggregates and that both polyGln and normal AR can aggregate in vivo in a COS-
1 ce11 Iine and in NSC-34 cells, a hybrid motor-neuron like ce11 line. Our data indicate that
proteolysis need not be a requirement for aggregation. Furthemore, we showed that
caspase cleavage of the polyG1n-expanded AR was not a prerequisite for aggregation.
3) Using a yeast and mamrnalian two-hybrid systems, we demonstrated that AR and ERa
directly interact in vivo. We also showed that this interaction had direct fiinctional
implications on AR and ERa transcriptional properties. This novel work suggests that
AR-ER heterodimers may have an important physiological role and explains androgen-
estrogen physiological interactions that have been previously documented. These results
underline the ever-increasing diversity in steroid signaling pathways.
Oligospermic infertiLity associated with an androgen receptor mutation that disrupts interdomain and coactivator (TIF2) interactions
Farid J. Ghadessy,' Joyce Lim,' Abdullah AR Abdullah.2-Waierir Panec-Rayrnond,l Cher Krong Choo,' Rose Lumbroso,' Thein G. Tut.' Bruce Go~cl ieb ,~ Leonard ~insky,'-j-~i-~ Mark A. Trifiro,'-' and Eu Leong Yonj!
Strucnimi changes in the androgen recepcor (AR) are one ofthe causes ofdefective spermato-enesis. a-
W e screened che&? gene of 173 int'srtilc men with irnpaired spermatogenesis and identified 3 ot [hem, unrelaced, who each had a single adenineiguaninc transition thac changed codon 556 in eson S €rom methionine co vsline. This mutation \vas significandy ~ s o c i a t e d wich the severel. oligospermic phe- notype and \,.as noc detected in 400 conrroI AR rillsles. Despite the location o f chis substicucion in the Iigand-binding domain (LBD) ofthe AR, neither the genitaI skin fibrobIasrs o f the subjects nor crans- fected ceil types expressing the mucanc recepcor had an? sndrogen-binding sbnonnality. However, the mutant recepcor had a consistencly (approsimateiy 50%) reduced capacicy CO cr;uisac~ivate each of2 dif- Cerenc andmgen-inducible reporter genes in 3 different ceIl iines. Deficient crmsactivsrion correisted with reduced binding of mutant AR cornpIexes CO androgen response elements. Coexpression of AR domain fragments in mammalian and y e j s c 0.0-hybrid studics suggests rhac the mutacion disruprs incer;lcuons of the LBD wirh anocher LBD, nith the XHr-cennind cransaccivation dornjin, and wich the trsnscrip&nJ intermedian. factor TiP,. These data suggest that a funccional efement cenrered around MSSG h a a role, noc for ligand binding, but for interdomain and coacci\~acor inrersctions culminacing in the f o m x i o n o f a normal transcription compIes.
j. C!:r.. fnzwr. 103:15 1 T-ijZ5 (1999).
Irrtroduction InÇerciliy d k c u appro.uimarrly 10- 15% ofdl couples ( 1). Spermxogenrsis is ac Cauic in abouc hdfofchcm. but ics cause is ohen covert. .bdrogens are required for normal spetrnacogenesis; however, mosc infcrdc men with impired spermatogenesis have normal serum androgen leve!s. Therefore, accencion has curncd to the androgen rrzsponse apparatus, particularly the androzen recepcor (AR). >lutacion of the X-iinked .-IR grne causes a widc range of c l i n id artdrogen insensitivicy cornpiete, when chç exrernai genicdia are femaie; partial, when they are suficiencly ambiguous CO require corrective surgery and mild. when chey are phenocypically male. Three sumeys of men aich idiopathic infenilicy have yieIdçd wideiy dis- parate frequencies ofandrogen-binciing abnonnaiiçies in cultured gçnicd skin fibroblasts (24). Gçnecic defeccs o f che AR chat cause miId androgen insensitivity wich impaired or preservrzd spermacogenesis (5-7) are of par- ciculsir intercsc because they may illuminrice chc fine structure-function artributes of the AR char permit di6 ferrntid reguhcion ofandrogen-inducible gcnes. To chis
end. ur screene;? a Iarge group of infertile men nith defec- cive spcrmato:;entsis for abnorrndicies in chr coding seg- mencs ofcheir AR gents. WC found 3 unrciated subjeccs tr-ith the ssrnr missense subscitucion in the COOH-ter- minai portion ofche ligand-binding domain (LBD) otthe A R One shavts infrequcnti?.; anocher has Iow-grade but persistent pospberral grmecornasria L'nexpecrediy, chis novefel mutation does noc affect the ligand-binding char- acteristics of che AR; rather, it reduces chc cransacciva- tional compccence of the AR bv impairing inreracrions benvecrn the recepcor domains. binding co androgen response elemencs (ARES), and function of the steroid receptor coacti~acor TiFZ.
Methods The~tu@poprtleon. Our patients presentca CO ir.krr:licy clinics because chev codd noc conceit.e for ac Izut 1 yetrs. Sonr \vas rderred wich the problem ofsexud smbiguicy. Two or more sperm urnples twre coIkctÈd ar Ie=r 3 months aparr a h r 3- CO
;&y abs:incnce md wcre Y X ~ K ~ sccordins CO X'orid Hedch Organkt ion critcna(8). Patiencs wvieh aonormd sperm sndy ses dur CO obstruction of the genitd crast. h~popicuit~nsm.
~ h e ~ o u r n d ofClinical Invcstig~cion 1 Junc 1999 1 Volume 103 ( Sumbcr 1 1 1517
h'-~erprolaccincrnis, or rnarkcidly raisai follicle-scimukting hor- mone wsre cxcluded. Of che remaining 173 pacicnw. 33 haJ ~oosperrriia, S4 \vere severely oligospermic (iess chan 5 mi[- lion/nL). 36 urre rnoderjcely oiigosprnnic (5-20 million/mL). and 10 hsd only abnormal sperrn motility. C o n c d male sub- jecrs !n - 100) ofproved fenilin: no prwous infertilicy hrscoq- or crexmenc. and wirhout an). genclcic d i s e ~ c w r e recruiced trorn che concrjcepti~-e clinic. .A furchcr 105 ( L 10 men and 95 rvomen) hedchy subjeccs were ais0 screrned co decrrmine wnechcr AR dlelic c;zriacions cUsc in the ycned populxion,
The mur~nrircbjeç:i CXCL presmced a t 3 1 years o f açe afcer 4 yèars of inferrilicy. He shaves infrequencly, approximace- ly once a week. Right-sided crypcorchidisrn w-ss correcced ac 7 years of age. Tescicuiar voiurne was 10 mL bilacemlly. and his sperm counc was approximactly 0.5 rnillion/rnL. RLH presenced ac 40 yctrs o t age aber 7 yesrs O€ infercilicy. He had Tanner gracie 7 persiscenc poscpubertril gyneco- mascia. TrscicuIar volume 6 mL bilacerally. wich spe rn councs t round 0.3 rniI1ion;rnL. Secondary sexual dtvelop- ment \vas ocherwis~. normal in boch subjeccs. The chird subjrcr, EHS. had a spcrm counc oiO.7 million,'mL. Al! 3 paciencs had normal male kiiryoc!.pcs and normal seru= 1evtIs of gonadotropins. androgens. ses hormone-binding globulin. prolsctin. and cscradiol.
.LI:~t.xtion~l sc+erning uirh ir~gfe-s;r=rx?'d con18nn~ri4n;! poiynor- ? h n . DKA ws excracced €rom peripheml blood. Coding ses- mena and tlanking incronic sequences of exons 2-S ofall 173 pxiencs wrre examind bu single-scranded conForrnaciona1 ~o lymor~h i sm (SSCP) (9). DS.4 fnyrnencs chsc eshibired dif- ferencial rnobilities wrre s eq t i e~~cd . Exons 2-5 of the paciencs (CLIL, KLH, EWSj wirh mucznc alIrtics tvere ret~iimined 5~ SSCP :O c~clude any cocxiscing mutscions chat miyh: have bcen missed on inicriil screening. Most ofeson 1 oiche mt;:anc sub- iccts u s scrernçd bu SSCP- The 2 hiçhly pol!morphic crinu- ckotidc (CAC and GGC) crac= uere ifirecc!y scqucnccd.
In r w o ~ircirogen-bozd111g propm:er of.4R. FibrobIas cvlcurcs obcained frorn A in biopslrs oisu~jects and f ron fiornai
cc~crols. Tne udroyen-bindin'; propértirs of the AR nxrr de:rr- rninrd according co scandard cechniqucs (10). Finj-scendr (IO-- 111 ws addrd CO che iuitures co inhibic rndo~enous 5a-redxc- case activiy., chac mighc degx!e rindroyens. Scacchard anak.sis !!-as perfomed by plocciny spec::?cdIy bound hormone versus the bound.'frre ratio. & \\= decçrrnined from che netacive slope ciche grjph. B . w z s oocained from the incerce~c of the line on rhe xasis. T h e r m o l a b i l i ~ \ \ ~ cxsmtned by cornparin'; chr bind- in- proFercies ac 32°C or3i"C. and W C . .A rrduccion o f B , ot' more chan 40% drfined cherrnolabilicy.
Chse c\tpmmrnx To decemine Ka. the race conscanc of dis- soiiacion. ce11 monolayers were preincubxed uich che ndiola- belcd androgens. and che proporcion oflabeled hormone scill s~ecificrilly bound dcer exposurc co excess unlabeled hormone rus decemined ac defincd rime incervds.
Plasmids JIS86VAR The muracion in our subjecrs \bas recreartd in a
cDNA fragment by sice-directed mucasenesis (6) and chcn sub- scicuced inco che homologous secrion ofsn ARespression WC-
cor. pSVhARo. Tmwccivacion domain (TAD) and LBD AR fragments wcre formed €rom AR cDNA by excising fragments bocndcd by the unique resirtccion sicts h'pni,'EcoM and .'J?rtI!hpnI. rrspccriwly.
.\l:rnm.zli~n nuo-h+id. The pC.AL-bDBD-LBD ICLONïECH hboncories Inc.. Pdo Aleo. CA) \r;rr prepsred by arnplifjins CDS.-\ encoding che .AR LBD (esons -1-9) and liyacing in-fijrne into the Smai/NindlI sice of ph1 containing CAL4 DNA-bind- ing domain. The ~VPI~AD-ARTAD t s s made bu rescrïccins pSC'MRo wich E J ~ I and HindIIl CO reltase the €mgrr.rnt encod- ing arnino x i & 14-565 o f chc A R The 5' end ofchis tkymenc
1
Normal CML
Figure 1
C T A G C T A G 1 #A
- (a) SSCP analyses oitarnrly memben of 2 probands Exon S r'ragmeccs were ampliried from genomic DNA 2nd electrùphoresed on a PACE 3 1 CO &play SSCP mobilip shiks. The probands. CLIL (laces Z and 6) and KLH (lanes 1 and 5). have a mucanc DNA stranci (laver arrow! chic m y x s kster chan the alieie ( u ~ s e r amrv) fiorri the unat~eced 580-
lins$ i sister of CML, lane 1; brocher aiKLH. lane 3). Mochers o i C x i l (lane 3) and KLH (lane 7) nave boch mutant and WT aileles. indicacrng chat h c h women are carners. (b) Sequcncing aïtorridioyrzrm oFa por- cron oFAR exon 8 from pacienc CML ïornparea wrh che normal Tnrèe patients had the sanc A-C subs~icucion. Icj 2esniction analyses oFFan- ily rembers or2 paoencs ;CILIL. KLH ). Gon 8 fragments wrre a m jlified Eoir: gènomic DNA and res:ncccd wicn BbrPl. T h hlSS6V rr.u:a:.on cre- ates a new BbrPT sire such chsc enzymacic diges:ion resclcs in 2 r'ragrncncs (6 and C) in the probands (lanc 3. CLIL; lane 5. KLH). wnereas normal a!leles (lane 1, scscer of CM1; lane 6. normal ferrile man; lane 7 . brocher of KLH) have oniy 1 Fmgnenc (A), measunng 327 bp. htocners aiCXIL (lane 2) and KLH (lanr 6) dispiay al1 3 fragments, indicacing chetr Sec- ero=.pus nacus. Outer lanes are 123-bp ONA IaCdeis (L).
was cnen lisareci co a synchecic linker encodin'; the first 1.3 amino acids ofART.AD.3nd che rcsuicanc k g m e n t encoding the encire ARTAD \\a doned in-Enne wich the VP 16 accivxion donain using pW16 veccor. Plssmid pVPI6AD-TIF2 \vas con- scrucced by a double digest ofpSG5-TIF1 wïch HiniiilI/.Yb~[ foi- lowrci by ligjcing in-frjme CO pVP 1G. The ( I in i ) , -El bTAT4-Luc reporcerveccor \ras o b c ~ n e d by amplie~ng the 5 CAL4 binding sires and chc adenovims El b minimal prornocer ofpG5C.T and iigacing upstrevn of che Iuciferyc gene in pGL-buic (Prornega Corp. Madison. Wisconsin. USA) vector.
Yc;xir~r*o-~bnd t1ïId-type ( K T or MS86V LBD fngmencs includiny (amino acids 503-9 19) or excluding (smino ac:ds 639-9 19) che DNA-bindiny dornain (DBD) w:e yeneraced bu docble diyesciny the full-Icnyrh ARveccor wich i\pn1/B~nzHI or PruI,'B.zrnHI. rrsycrtvcI!-. The LBD frjgmcnts rrrre cloned inco veccors pASZ-1 and p.iCT2. which contain chr GAL4 OSA-binding domain (amino acids 1- 147) and the GALA cruiwccivation durnain (smino ac ib 768-58 1). respectively. to genrrace GAL-b-AR hybrid conscrucrs. Full-length AR con- scruccs w r e made by excising rhe .AR from pSttiiARo.BHEX usiny SmtI/B~rnH1 and chen ligaring che resulang h ~ r n r n c CO
1S18 The Journd ofClinica1 Investigation 1 June 1999 1 Volume 103 1 Number I l
Figure 2 Dtssocration knecics OFARS in senical skin fibrocfasx. Nomal (C. open circics and boid lines) and rnucanc (CibIL. filfed circles and normal lines; KL.4. :?Ilcd scuares and rnin lines) Gbroblasc monolayc?r were exposed ca 7 nX1 [ 3 H i X ~ B (rop), 3 n,\t [JHjDHT(rniddle;. or 3 nXi ['HjT jboc- tom! at 37Y ,.left) or 27°C (righc) For 2 hours. Tne radioiabelcd medi- u m was cisarded and replaced with one concaintng 200-Fold excess unlace!ed anCrosen; repficace samples were removed at :5e indicaced cimes and assa:ted for ['Fi jandrogen chat was SCIII recepcor iound. Each dacr oolnc. the mean 0F4 repitcates. is cxpressed as a pcrcencage of max- imum bindins rc cime O. Vertical axes are on the sane loganchmic scale. excr?c for borrom ngnc panel.
pASZ- 1 and P A C E chac hsd bern sirntlarly digcsccd. Al1 con- sxucts u-ere sequcnccd CO confirma ch<: fidelin oichc rnqmat- ic rncinipulations.
X l ~ n t n r ~ i i ~ n ce!f cdrure and mjnstenc n ~ n r f i s i o n . Slucmc and U T plasmids were cmsfecrcd into COS-7. CV-1. or HeLs ceUs u s q Iipokccion cechnique (1 1). pCMV-PCd used co assess ctjnsfeccion et-ficienq. In some teplicaces. radiolabeled mibolerone (MB) u y added to che culcure medium and spe- ciEc h1B-bindiny activicy was dccermined (6). Transactivacïon acriviry was measured in relative lighc unics (RLL-') and nor- mdized co prorein content and cruisfeccion etTiciency
hmunablocon;r+s. Immunobloc an;rl>ses were ~ t d to study the eff~ec: olche mucation on AR protein production. The rab- bic pal!-clona1 mcibody PG-11, which rrcognizcs the first 21 SH:-cermind amino jci& of the h u m a AR. US^ CO decect AR procein (11) . ilfouse mAb SC510 (Santa Crut Biocechnolo- W Inc.. Sancs Cruz. CdiCornis, USA) W;LS uscd co Iocace +. GAL.CDBD fusion prottins. Procein-jntibody comple..a were s~b~~quent iyviewed by enhancd chernilurnin~cencc (1 1).
DNA mobiliry gel shifr assays and quan t i t a r i on of androgen-AR complexes b o u n d CO ARES
D&W m o b d i ~ g d i h i f ~ s j ~ ~ ~ . COS-7 cetis wcre trsnsfccrrd wrh \TT or m u m c AR plumi&. han-&ced in excrriccion bufkr (10 rnM HEPES [pH 7.91. 20% giycerol. IO0 rnl l KCI. O.:! m31 EDTA) conczinin~ proccuc inhibitors (phenylrncth~lsulfon~~1- tluoridc. Irupeptin. zprocinin). Iysed by 3 ~rerze:chsw qc1c.s CO
release A R A consensus synchecrc ARE (5'-CTAGh4GTCTG- GT.ACAGGGTGlTCttTiTCCi-Y). sen-cd as sprcifis-bind- in^ DNA (13). Two nonspecific cornpecicor oligonusfcocides were used. The first incorporates a KT esrroyen rrsponse clcmcnc (ERE) kom che promocer of the .Yenopus r~:reffogenrr: =\Z Srne: 5'-GTCCtMCTCAGGTC.4C.iGTGACCTCAT- C.L%AG71-3*; the secand incorporaces s cranscription hccor OcGA response clerncnc: 5*-GT.4CGGAGTATCCAGCTCCG- TACCATGCAAATCCTCTGG-3'. T-i polynucleocrde kinase m d Pf-JZPldATFj werc u c d ro label che ~Iigonucl~ocides. Bind- in^ reaccions concaining 5 p~ ce11 cïrrsct. 60 US BSA. 10-0 gis- crol, 1 mhl DTT. 2 US pal>-(dI4C). 0.1 rnhl EDTA. and 20 mXI HEPES (pH 7.9) in a coca1 volume of70 uL were preincubaced on ice for 10 minutes. Some reaccions containcd 3 13-told exccss of untabeled oiigonucIeocide sr cornpetitor DNA. hbeied oligonuclcocides ( 4 . 3 ns. 10.000 cpm) irrre added and incubation czmed ouc a furcher 10 minutes ac roorn tempera- cure. .AR-DNA cornpIexes were resolved by eiec:rophorrsrs on 5% PAGE sels anci auroradiopphed. * Qzttnrrr~tion of ~ndrogen-AR complercs borind CO ARES. AR c~pressed in COS-I cells \GIS cx~osea CO 3 n M [3HJXIB and har- vestrd, and aliquocs were rneasured for rsdioaxivicy. Etch s s a y rniscure containcd 10 uy of po[yidI-JC). a sample of cnsrcoal-crexed supernate wxh 50.000 Jpm of >SB-rrceptor complrxrs. and bindin'; buffer CO s final voiurnr of500 uL. Then we addcd 130 pmoi of 3'-btocinylated double- scrsnded ~Iigonucfeocide sequence o f che syachccic ARE ( S ' - C T . ~ G . ~ + G T C ~ G G T A U G G G T G T ~ G C + - ~ ~ O ~ ~ ) , or L I M TV-ARE (5'-TATG GTTAC.LLKTG~CTTXA..M - kitocin) and 50 KL otsrrrpravidin-asarose beads, behre con- cinuing incubacron for 7 hours. The bra& were coIIected by cencrifuyation. washed thrrtt cimes, and bound radioacrivity wss messurrd by liquid scintillation. .ks* mixtures lackins i n ARE bound ne&jble jmouncs ot ' r jdioact ivi . Esch daca point w u chc mean ofifuplicscc experïmencs calcukced ss a ;rrrcmcqe ofbindiny obsrncd wich UT AR (14, 15).
I'eancwo-bbnd J S S ~ ~ O ~ L B D - L B D tnrersctrons. Hybrid CAL-)- AR proccins n-err esprcssed in Sircchdmm~ces cerev:si~r Y 190 concaining incegraced GAL4 binding sites upscrcam of che LXcA~:-LcZ reporter gene. Incemcrion bcween hybrid pro- teins resuics in s ctjnscnptionally active cornplex inducing fi- galacrosidve accrvicy. Yesc transformation was performrd usiny the lichium scerace mechod accordiny CO prococols. Yevt were grown in scandard YEDP medium o r an appropnatr scleaive medium in the presence or absence o f 1 bf mechyl- trienoIone (MT). Yessc wcre harvesced. and liquid P-gdactosi- d s c assays were petfomed using O-nicrophenyl-P-D-galacto- side (ONPG) as a subscrace according CO scandard prococols wich die following modification: the F a s t were permeabilized by incubarion in a 0.2% sodium IauryI sarcosinate 2 buffer solucion (60 n M NI~HPOJ. 40 nhl NaH2P04. 10 rnbl KCI. 1 rn~M M ~ S O J . $0 mX1 P-merctpcoechanol) instead oifreeze- thawing beforc the addition ofONPG.
Results ,~~ururionr'detecred in the M of 3 rrnteljred oligospennic . putimrs. Three unreltced parienu - CICIL. KLH, and EHS - showed differential migration OC e x o n 8- PCR h g - menu when screened by SSCP. Two subjects, CXIL and
Figure 3 Transactivation accivtty of MSS6V AR wich nigh doses oiandrogens. b\T(operi circles and solid linès) or murme (filleci circles and docred lines) rece?cors wcre transiericly expressed in COS-; cells and exposèd ro kcreasing doses (nM) ofT (a), DHT (bj. or M B ( c ) . Transacriva- clon acrtvity was cxpressèd as M d increase in iueiterase acrrvity com- pared rvich cells noc exposea CO androgen. ~-g3lactosidase activicy and procein concenc were uscd CO normalire For eransFec;ion eficien- cy and ceil numbcrs. respectively. Each dara point represencs :he mean = SE OF 4 re?licaces.
KLH, inhrricrà che rnutacion rfom cheir mochers. *ho rvcre hecerozygous carriers (Figure la). The suspecred trsgmencs were sequenccd. and al1 2 patiencs had che sarne mucscion in exon S invoiviny rrrnino acid 886. an .A+G cransicion tesulting in substitution ofvalinr for methionine (Figure lb). This mutation resulced in rhr crexion o fa new BbrPI restriction sire (Figure Ic). -411 3 subjeccs were ntrar-luoosprrmic, wich repeaced sperrn councs Iess chan 1 million/mL CML had 23 codons in each txon 1 poIyrnorphic tnnucieocide repeac tracc. wheress KLH had 2 1 glucamine and 71 glycine codons, conftrming chat the 2 subjeccs were not gencùcdly relac- td No coexisting mucacions were detected in rlR exons 1-8 OF the 3 patiencs on SSCP analyses, nor in exon S PCR fmgmencs of 400 AR alldes from healchy concrols. The presence o f 3 mucations in our 173 infertile paciencs snd none in 400 concrol alleles m d e s ic unlikely chac MSSW exists in the g t n e d popu1acion (Fisher's exact test. P = 0.027). On che ocher hand, as the mucacions occur onIy in a subsec of infercile men wich severe oligospermia in - 84), il: is more likely (P = 0.005) chat MSSGV is significanciy sssociaced with che p h e n o q x o f severe oligospermia.
M886Vin tbe LBD had no eficr on androgen-bindntg charac- r d c s . The Kd for dihydrocescosceronr (DHT) \vas 0.3 I and 0.52 nM, and che B-x was 36 and 37 fin01 DHT/mg protein For CML and KLH, respeccively, ac 37°C for
9 520 The J o u d ot'ClinicJ Investigaaon 1 Ju
tîbroblsist m o n o l a ~ ~ n cultured tiom sirocal sk:n biopsirs (normal: A;. 0.3-0.6 nhl: B,,, 16-43 fmoL'm~ procem). Similady. chere n-ere no sigriir?cant ditTcrenccs in afiîni- y cons t a t s for cesrosterone (T). che & for CLIL bemg 1-23 nhI (normal: -1.S n\l). T'ith XIB. AG for KLH fibroil- Iasa trss 0.1 nb1 and B,, n y 52 fntol!-'rn= prsccin: trich AIT (RISSI). A:, \.as 0.11 nAL and B,, \ v a 52 ïmoi. mg procein (normal: Kd. 0.1-0.3 nh[: B,,. 15-50 h o 1 m- protein). \Shen c~posed co DWT at 41°C. the .Y< an,! B-, t~;l!ces of AR in CXlL fibrobIascs (0.27 nhI and 35 fmol DHT,'rny proce", mplrctiwly) uure no ditT'rrrnc Som the corresponding d u e s ac 37°C. indicacing the absence of chermoisbdity. Thc espcrimenc \vas tepescd using N B an2 MT. nith similx results (data not shor~m). Prolongeci labelin'; for 15 hours ac 37°C uich 1 a\[ LIT. or AlB, fol- lo~red by 3 swicch :O 4l'C or 47OC tOr up CO 6 hours in che ~resence OC 100 mhI qclohesirnitie, failcd co demon- srrace snv aiFerence benrem mutanc anci normal te!ls ac t h higher crrnperatures. Chase expenmencs uwe pcr- Corrned CO examine che dissociation kinittics o i rnu tmt .U complexes. When chascd ac 3 i 0 C or 42 'C. the disso- cmcion rares for T, DHT, and XIB U-ce sirnilar for boch norrnd and mutant fibrobiascs (Figcrc 1)- .+fier an LS- hour incubacion \\+th T, rhr rernaining .4R tonpk .~es dis- sociated Iinexly wich a k valüe o f 5 ( IO-','min) ac 27°C. This \ d u r is vpicai of DHT-recepcor cornpieses and indicares normal conversion of T co DHT. The AR csn- cecc ofCML fibroblascs rosc fiom 4S CO 104 h o 1 'rns prorein whcn incubatrd wch hIB for 3. and 70 hours. res?ecciveIy. indicarin:: normal rece? tor u pre,ouIxiun. Xormally. a Coublinj or j rexer ofspctcrfic androyen- binding accivicy is obser:eci afier overniyhc incubation wich androgen. presumably because ofAR scaf5Iizstmn by !ipnd. Liken-ise, t he arTinlcy and dissociation charac- cerisrics of androgens bound to mucant receprors expressed in COS-T cdls were no ciifferenc korn the X? (data noc shonn). In assre;xli. these esperirnencs, indr- prndcntly replicaced in 2 laboracorics. indicace chat the lLISSGV mucation of the LBD Ji9 noc change any andro- gen-bindiny ~roperries of the rccepcor.
In:?.+ed ~ ~ ~ r w n i r r ~ a r : c ~ p x r ~ qfrhr nrimnr rece?ror. The mutam receptor, hs\r.evrr, haci only 50-70% of che crans- activacion capacicy of chr \iT recepcor in COS-7 tells when esposed CO tarious concentrarions of T. DHT. and LtB. The ARS encoded by the normal and mutant plas- mids demonscrsced dose-deyendent increases in crans- activation capscicy wich T, buc ac a11 doses €rom 10 to 300 nXI. the cranscriptional capacicy o€rL[SS% AR u'as only about h d f chat of the KT (data noc shotvn). The cransactivationtl defect persisceci even wlch doses of 1 androgens up co 1 PM (Figure 3). Thus. chc murant AR ! ha9 s modest but consistent reciuccion in iu crsnsacci- vacion capabilicy compared \viril cliac ofche WT. for al1 5 androgens. in repeaced experimcnts. üsing a ce11 line (CVl ) chac does noc express the SV40 ancigen. che crans- accivacion defecc of the mursnc AR was even more evi- dent. Thus. when 50,100, and 100 ng of AR cD'i.4 uvre
i r
used in che cransfeccions, the KT gave X I - . 1 LI-, and 3.+foId higher transactivxion acnticy, respecrivcly. com-
i pared wich the mutant (Figure -!a). Imrnunobloc analy- ses s h o w d chat AR protein Ieveis were cqu idenc For rnucyic and W'T trruisfections. 3 c d i 3 cDNA doses (Fig-
I : l
inc 1999 1 volume 103 1 Number 11
Figure 4 (a) Transacivacion ac:ivicy o f MSSoll AR wtch incressing doses D C A ? cDNA. CV1 cells were cocransfeced wch the indicaced anouncs oiWT or ?.lSS6V AR cDNA and the reporter plasmid plLIAAI-LUC. Eack daca p o i ~ c . che rnean oicnplicaces. represenrs che foia increase in luctierase activicy of cells çxposed CO 30 nM M B compared wicn :hose rvichouc androgen. Bars are = SE. (b) lmmunobloc o i W T (W) or rnucanc (LI) receptors. CV1 crll èxrraccs (IO pg cota1 procein each) d e ? i c d in a were elearopnorèsed on an SOS-PACE 3el. and AR procetn idencifieci rv~ch a specific ancibodv (PC-2 1 ). Films ivere overexposed to enhance signai For chê low-abundance AR procein.
ure 4b). Andro~en-binding ricciviry of the mucrint and KT AR proceins in COS cells \vas also sirnilu (drica noc shown). confirming chat the cransaccivacion d&cc of LtSSGV \vas noc due to changes in AR procein Ieveis.
I n p i r e d binding ofM886V ligand-AR conzpkexes ro ARE,-. Boch W T and murant tecepcon displayed an increve in ARE binding wich chr higher hormone dose (Figure 5a. compare Iane 1 with 2. and lsne 3 wich 4). Hosever the mu& receptor, despite being present in slighdy greacer quanticies (Figure Ib), \vas unable CO bind synchecic ARE Y etTectively as the K T (Figure sa, compare Imr 7 with 8. and ime 9 wich 10). Addition of excess uniabeled synthec- ic ARE reduced boch W ï and mutant signais, wich the mutant bancl being Içss prominent chan che (Figure 5s. lanes 5 and 6). The presence ~Fnonspecific cornpeciror D N A (ERE; and OccZa oligonucleocide. the binding sicc for a ubiquicous cranscri prion factor) did noc reduce the band shifts, but mucanc signai wss scilf less intense chan W'T in boch instances (Figure Sa, compare lanes 7 wich S and lsne 9 wich IO. The spccificiey of che band shifc \vas confirmed by che absence of che mosc prominent band when radiotabeied ERE \s used (Figure 53, lates 1 1 and 12). The ARE-binding deficic of the mucanc ligand-AR
cornpIeses \vas qusnciried In a t k c series 0fe.perirnencs using 2 ditferenc ARES. che mutant recepcur Iiad only 57.7% (= 5-17 SE) ofciie XT DNA-binding capsci? (Fis- ure Sc. ldt). In a sscond serics oftsprrirnencs. a siighcfy rnodified cechnique was used co reduce background counts. and che OSA-binding d e k t \vas even more markec!. wich hISSG\' dispIa!ing only 33% ofthr activiq- cornpared wich n o m i concrols (Figure je, righc). In com- parison. 2 DK.4 bindins-domain mucancs. APhrjSI and J.4~615, ciisplayed less chan 10% of normal bindinÿ. fn aggre~acc, the SISSGi' AR had onIy about hdf the DXA- binding acnvïryoFchc KT. and chis defrct in DK.4 bind- in^ was cornmensurace ~ i c h chr degrer ofimpaireci cians- acrivation O bsen-rd. as discussed earlier here.
The mnr~rion .zflecr.i T U - L B D inrerzction. To decerrr.in<: wherher the 1ISSGY mutation impairs interaction benveen che LBD and TAD, \*ectors encoding the COOH-cerminal LBD and the NHi-terminal TAD wcrc conscrucccd and espressed simulc~neously in che pres- ence of an ARE-driren reporter ycne (Figure 6a). The LBD or TAD Frqrnencs by chemselves did noc dispiay an? androgen-induciolc reporter gene accivity- However, KT LBD fragment. whcn cotivpressed wich the TAD. resulce:! in an androgcn-dependenc increue in luciferase a c t i v i ~ when cxposed to physiotogical doses of DHT. indicacin; char AR fra~rnencs can incerricc to yïeld ri h n c - cional protein (Fisurs Ga. Iet't). blutant LBD fragment. whcn coe.upressed tri th TAD. resulced consiscenciy in ci 22-25% \O\\-er sndrogtn-inducible accivicy compared ivich rhe KT LBD frasrnent in 3 independenc esperi- ments. Fusion proceins. tornprising ARTAD Fuseci co VP16AD. and .ARLBD co the G-ALJDBD, were cocs- pressed in HeL3 cens. and protein-procein interactions wrre messureci n-ich a reporter vçccor concaining muici- p k CAL4 DS.1-binding sites in the rnammalian ctt-O-
hvbrid assay (F ip r e 60). Inceraccions benvecn TAD and L-BD were specific and nndropn dependent. Di5erencis benveen KT and mucanc w r e mosc evidenc when ~pprosirnsce!y equi\denc quacicies (50-100 ng) otLSD and TAD hybnd plsrnids wxe cocrsnsfecced (Figuré Gb. Mc). Stctant LBD Fusion procein had only half the acciv- icy of the \C? in chis assay whrn exposed co sub- nanomclar doses ot-hl6 (Figure 6b. righc). Thus, expen- mrncs wrch AR frasmencs using an ARE-driven reporter, and AR fusion proceins using GAL4-driven reporcer. indicated chat h(SS6V impaired TAD-LBD interactions.
Table 1 Interactions ofAR tüsion proceins in che yeasc cwo-hybrid assay
Erpenmenr WT ,MS86V Gperirncnc \KT ~ 8 8 6 V
1 5.53 = 0.72 3.03 = 0.68 1 24-81 -1 jl.4 ~ 2 . 5 2 7.08 = 5.66 3.33 = 0.29 2 3 7 9 ~ 1 9 3 197~57 3 5.45 i 2-93 3.55 = 1.26 3 154~S1.4 t t 4 = 4 3
Mean 6.03 3.30 Mean 259 124
Ycan C;rLLactivation [C(Tr\0)1 and ONA-binding [C(DBO)] damains were hscd to full-leng:, AR or G O hgmenr CO srudy merdomain inccn~cianr. Oaca (mean : SE) rcpresenc the iold incnase in P-~ala~tos~dase act~vitv. in the prcscnce and absence of1 uaM MB. c i 3 independenc urpcnmtnrs. cach In v ip l i a~c .
The ~ o u r n d oFClinicd In=?sripcion ) Junc 1999 1 Volume 103 i Number 1 I 1521
Figure 5
W M W M W M W M W M W M 10 100 ARE S E Oct 'EFIE
t W M W M
10 1 O 0
(a) DNA mobilicy gel shtk assay. W ( W ) or mutant (LI) recepcors were expresscd ln COS-7 cefls and wposed CO 7 O nM (lancs 1 and 2) or 100 nM (lanes 3-1 2) D m . Equivalenc quanticies ofimmunoreactive AR frorn the cell exvacu wcre added CO binding reacrions concaining "P-labeled synthetic ARE (lanes 1-1 0) or ER€ (lanes 11 and 12) oliponucleocide sequences. Gcess unlabelcd ARE (Ianes 5 and 6) . €RE (lanes 7 and 8 ) , or Oct (lancs 9 and 10) oligonucleocidcs wtre added as compericor DNA ro dernonsvatt the specificiq o f the binding naction. The dark band at che bocrom represenu unbound "P-labelcd DNA. (b) lrnrnunobloc o f W or mutanc recepcors used in gel shik assay. Fire microliters oirepre- senmove cell ucu;lct (used in r)ie gel shift a s a y depicced in Figure Sa) was exposcd to etcher 10 or 100 nM o f DHT and was separated on an SDS- PACE gel. AR protein was idcntified wich a spuific ancibody (PC-21). (c) Quancification of binding to ARES. Receptors were expresscd in COS cells, exposed co ['HIMB and equivalenc quantities (50,000 dpm) o f ('H],MS-AR complexes incubated wich 150 pmol of either biorin-labeled nacurai ARE (an ARE fiorn MMTV-LTR) or synthecic ARE in 2 independ- ent senes of uperimcnu. Screpcavidin-biocïn-bound ARG were collea- ed by cenrdÙgacron, and [JHIMB-labcfed receptor bound co the A R G was quancified by scintillacion counring. Known DNA binding-domain mutants (JFS82, AR61 5 ) with sevem impairment OFDNA binding were used for cornparison. In the righc panel. background cauncs were low- ered by veating the lysate with durran-coared charcoal and by cenuifir- gacion at 100.000 g for 1 hour. Assayr using mock-cransfecccd cells showcd minimai background acrivicy. &ch dasa point was the mean o f 2 experimencs. and bars indicacc cheir ringe.
The rnntxion dfftcr,- LBD-LBD znterxtions. To esamme whether the Xf856V mutauon can aiTecc LBD-LBD incer- actions. rre consc~cted hybrid expression veccors cvhere- in Ml-length or cruncaced AR fmsments were fucd co the GAL4 DNA-binciing [G(DBD)] o r the GALJ-crans- activation [G(TAD) J domains in che yeasc nro-h!.brid assay (Table 1). X o androgcn-inducibk activicy \t-as observed when G - U t A R hybrid proceins were coes- pressed wich the CAL4 domains [G(DBD) or G(TAD)] done (dota not shown). Coexpressed G(TAi3)-AR(LBD) and G(DBD)-AR(LBD) hybrid proctins concaininy che WT LBD induced a 5- co 7-fold increase in transcrip- tional acci\i;icy in the presence ofandrojcn. In concmr. fusion proceins concaining che rnucanc LBD were con- siscenrly only haif as cranscriptionally acrive as the KT in 3 independent expenmencs, each perhnned in tripli- cace- Defeccive interactions were also observed when a larger AR fragment, G(TAD)-AR(DB D LBD), was used (daca noc shown). Nocably, when full-Iength AR [G(TAD)-(AR)] fu ion procein &as coespressed, intetric- rions were 7 orders of magnitude higher chan with crun- caced AR recombinant proceins lacking ARTAD, in&- cacing the importance OTTAD-LBD inceraccions to the dimerizacion process. This increase in magnitude of interaction may rerlect intramolçcular LBD-TAD asso- ciarion (16) or che incermolrcular efiect of the TAD on the LBD (17). In dl chese esperimencs, iLlSS6V fusion proceins were consistencly transcriptiona~ly defecrive, h a n n g only about hdfthe activicy oF&e KT.
M8562' riirrrrpt- intmctionj wirh the coa~m~zlc~tor T I F 2 Slechionine 886 oÇche .LR is close CO residues known co inceract wich che sceroid receptor coaccir*acor TTFZ (1s). K'c cherefore exmincd che efEecr oflLISS6t'on coactira- cor function in 3 ways. Firsc, we rneasured the eFFect of &Il-lengch TiFZ on \TT and mutant AR cranscriptionai acriviq.; second, chè inceraction offusion proceins con- taining TIF? and LBD; and chird. the eEecc of FuU-lengh TIF2 on T-AD-LBD inceraccions. Full-Iengch TiFZ enhanceci [TT AR acciriy approsimacely 3-foId in a dosc- dcpendent rnanner (Figure Ta, lefc). However, MSSGV was defeccive and lowemd coaccivacor Funccion by 22-4 1%. Significanc impairment ofTiF2 function was mosc porninenc at a dose of0.1 n M M B (Figure Ta, right). t r i e n h i o n proceins contsining GAL4DBD- ARLBD were coexpressed wich VPL6AD-TTFZ, XlS86V caused a mean 3n lower recepcor/coaccivacor incerac- tion wich doses of MB from 0.01 to 1 nM (Figure 7b. left). Impôired inceracuon of TIF2 with iMS86V was noc correcred by increasing rhe arnounc ofTIFZ cDNA (Fig- ure 7b, nghc). Defeccive inceraccions of mutant LBD wich TIF2 were nor due co clifferences in iBD L i o n pro- cein contenc. s measured by immunoblou and [3H]M B binding (Figure 7b, boctom). To determine whether NH2- and COOH-terminal inceraccions oFche AR were also reguiared by coacrivacor, Mi-Iengch TE2 tvas cosx- pressed wich WT or mutanc GAL4DBD-ARLBD and VPL6AD-ARTAD fu ion proceins in the mammalian nvo-hybrid systern. Interacrions berneen the %'T LBD and the T.D fusion proteins were increased up co 20- fold in the presence otTiF2. MS86V disrupced chis effecc ! ofTIF2 by up CO 28%, consiscenc wich che impaired TAD- LBD interactions observed previousty (Figure 7c).
k
Figure 6 !a) Transcripcion acrwicy o f AR fiagrnencs. Oelecion conrtniccs enc0ci.g :ne ÂRTAD (amino acids 1-504) or che ARLBO (amino acids 507-9: Ç)
were cranscenrly mnsfecced inca CV-1 cells. Tranxnpoonal aciviry (mczs- ured in RLU) of the VVT or h18S6V LBO hagments. alone or coexpres~ed wrch TAD ( 1 pg), was rnrasured wtth a phIAII-LUC in the Fresence azd aosence oiindicatcd amounrs OHT (nM) (lek). The expenrnencs wè-e repesced on anocher Z occasions using independenc plasmid preparatiors, and che results were expressed as fold increase Ïn luciferase acrivicy in r5c presence and absence o f 3 n M DtiT(nghc). (b) lnceracrions ofLBO and TAO Fiision procerns in the mammalian wo-hybnd assty. The fusion pro- reins VPl6AD-ARTA0 and CALJOBD-ARLBD were coupressed in HeLa cells. and recepcor TAO-LE0 inceraccions wcre rneasured wich ( ? / m l : -
EibTATA-Luc reporcer piasmid. Cells were uposéd to ~ncreasing doses o i pVPl6AD-ARTAD or M B as indicaced. Data points represenc mean = SE of ac leasc 3 replicares and reflcct fold increase in luciferase acrivicy ofceils exposed CO M B over tnose noc uposed CO the androgen.
Oiscussion Eariier snidies (23, 19,20) have provided endocrine-bio- chcmicai ecvidence iinking oli~ospermia a n d azoosper- mi% with or wichout other signs oiunderviriliwuon. :O
quancicative a.nd/or quditacive abnormalicies ot'che AR Thlrre have been onIy 2 case reporu (6,21) o f AR point mutacions in men wich different degrees o f impaired spennatogenesis- This scudy $vas undertaken co discov- er conscinicional AR mutations in a series o f 173 men wich varying degrees of impaired spennacogencsis. Amony chc subsec of S1 men wich severe, idiopachic oligospennis, we found 3, unrelaced, cach of whom had the same novel AR variruir: btS86V. The facr chat we did noc find the variant among 400 concroi AR aileles indi-
cxed chac chis w u a pachoyenic ali&t variricion. O n e of the men shs-es infrequrncfy the ocher h u Tanner çrsde 3 persiscenr postpubercal yunecornascia These fsccs screngthened the idra chac a cause-2nd-&ect relation- ship mighc exisc benvcen the AR variant and screre oligospermiz To prove ics pachoyenicicy. we undercook the seudies. the resuIcs oFwhich are discussed Iacer hete.
The locacion otche hIecSSG\kI mucation in the LBD of che AR Ied us to espect some deyrec otandroyen-bind- iny abnorrndicy because a nearby mucation. VSS9h1.3 residues aownscre;irn. has been reported co cause nezrfy complece androgen insensiciviy due to Jefecrive andro- gen-binding capacity (11). Furchermore. al1 mutacions so Far described in exon S of che AR manifest some abnonnalicy o f androgen binding (23). To our surprise. d l androgen-binding propcrties of rhe XISS6V in genicd skin fibroblascs o f boch probands were wichin normal Lirnits. Sirnilarty, che androgcn-binding propercies 0 5 t h ~ rnucanc .a were norrnai in 2 cransfected m m m d i m ce11 Lines. Thus. chr MSSGl' mutation, althou,oh residing in the hormone-binding domain ofche a does noc attecr: chr conformacion o f the ligand-bindiny pockec o f the LBD. However, the rnucanc AR \vas unable to cransacci- vace normally and disptayed only 50-70% ofche c m s a c - citacion abiiicy of the \VT in 3 diKerent ceIl Iines [COS. CV-1, HeLs), as mesu rcd bu naturai (MhI7-V-LTR) o r mulcimeric ARE reporcer Senes. -411 cransaccivacion- defective LBD mutations revierved to dace (33) have bern ssociaced with sorne form ofandrogen-binding abnor- malicies. In chis regard. ic is srriking ro noce char the nesrby VSS9M mutacion caused neariy complete andro- @en insensicivicy wich increased androgen dissociation 3
kinecics, despice normd equilibriurn androgen-binding affinicy (24). The actkaced murant recepcor had a n impaired abi l iy to bind oIigonucleocides containing 2 ditfrrenc ARES: the first. an artifici31 ARE b u e d o n the consensus DNA-binding sice for androgens; chr second, XIXITV-LTR the nacurally occurring scrroid responsr element. Bindiny co eicher ARE wss approximacely 60% chat o f the KT. Thus, i t is very Likdy chac reduced ARE bindin; concribuces co che reduced cransaccivacion com- pecence OF the ILISSGV A R Howevcr, as discussed lacer. F'ulcy intra- and incemolecuIar interactions o f the murant AR may, direcciy or indirectly, be concributory as well. h.iutacions such as MS86V are parricuIxiy interest- ing, Y chey can illuminace funcrional subdomains thac reside in the LBD.
Interactions benween the NH2- and COOH-terminal domains otche AR have been observed using the mam- rnalian ovo-hybnd system (17,15), and a transcription- dl? active cornpleu f o m s when truncrrted proceins hav- ing only the NHr- and COOH-terminal dornains OF the AR ;ue coexpressed in rnjmrndian cells (16). Seversi lines ofevidence indicate chat defective TAD-LBD and LBD- LBD inceraccions are important in the pachogenicicy of rLISS6V. MSS6 in the COOH-terminal region, direcciy or indireccly. inceraccs with the NH2-terminal portion o f the A R hISS6V impairs T a - L B D interactions in both AR- and GhLUsiven repomr syscems. bfSS6V &O con- siscencly disrupcçd LBD-LBD interactions of GALIAR fusion proceins in the yevc WO-hybrid assay.These data indicate chat &I8S6 has s role in incetriccions benvecn the
TheJoumJl ofctinicai Invescïgacion 1 June 1999 ( Volume L03 [ Numbcr 11 t S U
(O) EZecr ofTIF2 o n AR accivity in H e i a cells. M u t a n t o r WT AR plasrnids were cocransfecrei svirh the indicated amounts OFCONA e n c o l i n g &Il- length TE, wich (+) orwichouc (-) 0.01 n M M B (lefi). Cells were :ransfec:ed wirh 5J n g TIF2 cDNA a n d cxposed co increasing doses o i M B (nghr i . AR ac t i v ig was measured w i r h a mult imeric AR reporter gene (ARE-TATA-Luc). (b) Interactions o f T i F Z and ARLBO h g m e n t s ,n the mammal ian r-vo-hybnd assay. The Fusion proreins VP16AD-TIF2 a n d CAL~DBO-ARLSD were coexpresscd in HeLa cells. and procein-protein interactions were measured wich ( 17m)~-EfbTATA-Luc reporter plasmid. A m i n o acid posioons o f T l F l and ARLSO fiagrnanu are numbered. Vertical bars wirh in the nFZ fragmenc indicare nuclear receptor-;nteracting box motifs (L)(XLL), azd numbcrs rndicate the firsc L oreach consensus moctf. In the I c k panel. cells were cxposed to increasing doses o f M B ; in the r ignt panel, cells were exposed :a increasicg doses o r t h e VP16AD-TIF2 ex~ression vecror. D a t a are expresscd as fold increase i n lucir'crase a c c i v i g \ ~ i t h o r wichout 1 n M Ma. B o a o n panels skow a n irnrnunab(oc ofWT(lanes 1-3) and m u r a n t (lanes 4-6) CAL4DBB-ARLBD fusion proteins (-51 kDa) from representative celi :ysates (1 0 ug procetn p e r lane) a n d specific [ IH jh lB-b ind ing aerivity ( h o l i m g procein) of celis uansfecred wich LVTor mueam CALJDBB-ARLBD *remor. (c! Effect o r T F 2 on TAO-L50 inre.ac:ions in the m a m - mal ian cwo-hybnd assay. W o r mutan t CALJDBD-ARLBD fusion pmcein :vas coex=rtssed w i th V P l &AD-ARTAD in H e l a cells. and TAD-LBD incer- sc:ions we-e mcasured wirh a (ffm)rE15TATA-Luc repocer plasmid (as i n bi. The ofcocransfeccing & x e a s i n ~ doses ofa fourth veccor. pSC5- TIF'>, encoding the Fiill-lengrh Tif2 procein. was measured as fold increase tn luciTe.ase activicj svirn a n d wiehouc 0.1 ntU M B . Dara were che mean = SE o f a t feasc 3 repiicaces.
hnccional domains of the A R It is imporcanc CO note chat cnt nearby residue, VSS9, alchough a parc ofthe lig- and-binding pockee, also inceraccs uich che TAD (26). Similarly, mutations of the human escrogen recepcor affeccing C53O (27), approxirnacely 5 residues KH,- 7 cer- mind CO the homologue or'= h.1686, lose DNA-binding aceiwty bue recYn normal estadiot atxnicy. Colleccively, chese daca susgesr scronyly thac a Functional çlernenc cencerrzd around residue 886 of the AR has a d e , noc for Iigand binding, but for incerdomain inceraccions.
Rccencly, severd sceroid recepcor coactivators (S RG) have biten characeerized chat mediate nuclear receptor inceraccions wich che preuiinacion cornplex and chç chro- matin remplace (28). The SRC gene fimily has several closely relaced homolopes; of these, TIF2 has che greac- est accivicy wich respect to .4R (29). Three highIy consen-ed regions. each concrtining the nuclear recepcor-ineeraccing mocir' W L L , are locaced within the nuclear rece pcor-inceracung domain of' TiF2 (Figure 7b). The LXYLL motifs incrracc wich core motif's in LBDs of nuclear recepcors close CO residue 886 of the AR (30). MSS6V significancly impairs TIF2 coaccivacor hncaon in
anXR-tiren reporcer s>xeern and dso wich chimeric pro- ceins ir. the mammdizui nro-hybrid assay. In boch sys- cerns. hlSS6V reduces TIF2 coactivacion iùnccion by approximacdy 4OQ, equivdrnc co the defem in cransacci- vation, DNA binding, and incerdomain intcraccions obsrzmed. There is evidence char TIF2 incemcts wich boch the TAD and LBD ofsceroid recepcors, enhances the cran- scripcional accivicy otTAD and LBD sepriracely and hris an additi.de e f k t when TAO and LBD are expressed simulcaneously (3 1). Our daca show chat TIF3 increases by approxirnateIy 20-f01d che interactions benveen che TAD and LBD of the AR and chac chis incersrction is dis- mpted by MS86V (Figure 7c). Thus, ic is not surprising chat our mutation c m affect interdomain inceraccions and DSA binding, as TAD-LBD inceraaions are thoughc CO be criacd for boch Fiinctions (26). Ie is known chac lig- md-induced recruiemenc ofan UCYLL motifof SRC- 1 to the retinoic acid recepcor (EUR) promaces hecerodimer- kation co the retinoic acid X receptor (RYR) and binding O fa second L W rnou€molede to RXR The 2 adjacenc UCYLL motifs in SRC-1 concacc homologous residues in che LBDs of che RYR-FUR dirners, thercby foming a
-- -
1524 TheJoumJl o f C l i n i d fnvesugacion ( Junc 1999 1 Volume 103 1 Number 11
m
1 -
n
C
: 5
1
The Jourrd of C l i n i d lnvedpüon 1 june 1999 1 Volume 103 1 Prumber I l
charge clamp to stabiiizr' the dimerizarion comp1es (32). Our findings chat che M S S 6 V mucacion aGem man? AR processes, inciuding T6AD-LBD and LBD-LBD inccrac- uons. D M bindins, and cramaCn~tion cornpetence. are consistent wich the incapcive rote chac coaccivrirors like TIF2 m. have in thtse post-ligand-binding evencs.
Because a nearby AR mutacion (E897Q) &O presemes Libmd binding whiie disru pang Tm-LBD and LB D-TTFZ inceraccîons (18). our data suggest chac MSS6 conmibuces CO a hncuonai subdomain ofche AR LBD chac d o s noc rnediace androgen binding buc is important for interaction wich TIR. This F:ncciond dichocomy in the AR LBD is noc irnplausibie considering chac incermolecuIar inarac- cions depend on surface propertiçs, wherev Iigand-bind- ing pockecs are buried wirhin che hydrophobie cores ofaii sceroid receptor LBDs ~ c ~ e d co dace (33). The idenci- fication O focher c r ~ l s o c t i ~ ~ t i o n - d e f e a i LBD mutations (6.2 1) chat do noc Liwd binding suggests che pos- sibiiicy chat disrupuon ofone or anochcr coacà\.acor a&*- iy may bir a common concribucor CO pachogeniris ofsevere oii~ospermiaand male i n f e d c y Given chat at l e s t one OF these abnomdicies is c o r r e d l e by hormonal rnanipuia- cion (6.21), understanding the padiogenetic rnechuiisms mzy Iead CO eEeccive cherapeuuc scrategies.
Acknowledgments K'e chuik G. jçnscer (.-derson Cancer Cencer. Houscon.T~sx. USA) for his kind gik ofARE-TATA-bc. G. Pins (University or' Illinois, Chicago, Iliinois. USA) For the AR ancibody PG-2 1. y i d P. Chambon (IGBMC. Strasbourg. France) for ?SGj-TIF?. This wotk is supported by gmncs t'rom rhe Fonds de 13 Recherche en Sanci du Quibrc (Hydro-QuCbcc). Fonds pour L Fomukr~on de Chercheurs ec l'Aide 3 1s Recherche. w d the hkdicd Resrarch Councils of Canada and Singsporr. -4XR AbdulId~ is ,orscehl for prnond supporr from the L'ni\-ersicyot'Bakwt.
cion oichc androgc" reccpcor assoaatcd u~ch parnai andmgen raisc- ancc. h i l i a i mmatcomuaa. and icrcility. J. Clm. Endamnai Slzub.
andmgen reccpcor muoaon arisingandrogcn raisunce in dit under- vinkd d e syndrome/. Uin. EnZormoL .W#.ib. 79:llOT-1207.
S. World Hdrh O r p n k a o n 1991. Lboramy m u t d tÔr the cxaminz- Oon o i h u m u l wmcn and sperm-cenicd mucus inreracuons. 3rd cdi- a o n RI. Arcken a ai.. d m n . WHO S p d Royrun o i k c i r d t Dcvei- opmcnï. and Rcsurch Training o n Human Rrprodumon. Cetera, Srmacriand 1-10?.
9. Yony. EL.Chui. KL. Yang, XI.. Roy. A. and Ramam.SS.1994. Corn- plrcc andmgen insemiciwcy duc CO a spiice-sire muunon in chc ~x5-0- yca cecepcor genc and àcnecic scnening wch sinyle s m J c d conior- manon poiymorpttum (SSCP). FmL S r m L 6l:S56-961
10. Gocrlr~-. B.. Luiman. hl.. R ~ k y . L. ~ b e e u c 6.. =ri Socor. J.F. 19s- ~~~~ei!ukr carrrcnon oicke mciropn-rmq?cùr crtr.s~~rnanon &fkr in rivo 5 m i I i a uxh camc!a:z a n d q c n .r j istz~ic.. J. Srrnira' Brui;rl i . 28 23-34.
1 1. Tur. T.G.. Glia2asy. F.. Tcnro. SI-1. Piruky. L. u i d \-an$. EL 19 9' Pol\-giucamtne i n tkc andmgen nrcpror arc ~ s c u c d trich redkcf t ~ v a c : o n . &<cccir-e s p m producnon and mdc rnicrnir- ey.3. & EndomwL .\lc?.& 823ïïï-3782
IL Pcns. G.S.. Bircit- L. and Grclnc, G.L 1991. -4nrf=ogcn ~ K C ~ C J ? !ocil- i ~ n o r : in ditErenc =el! of chc adulc r z c pro~cacs. S n d o ~ n n o i ~ 1293 191-3 199.
13. Rxke. P.J.. Hoxe. S.A..z?d PIrkzr. MG. 1991. .- \~7i~xr~us O S A b in t - i ny sicc s r chc andropn rsqcor. .Ud ~ d a r r d . 6 2219-2235.
LJ. tauu. V-C.. rc A. 1991. RYRS: a cont?guLcor char e n h x i a SinCina JC xcino:c xd. th>mid homoncandviumin D rccc;cars :O cheir <obgtxs rcspçinrc elernmrs. 663:iSL- 1366.
15. G b r . CL. cr ai. 1937. .A ;-&-A sicc :n n c p u * kom.oec scnc m d i -
