A

gg2a2omreotam©

Ki

fpTm

tT

f

0d

Molecular & Biochemical Parasitology 153 (2007) 115–124

C-terminal proteolysis of prenylated proteins in trypanosomatids and RNAinterference of enzymes required for the post-translational processing

pathway of farnesylated proteins

John R. Gillespie a, Kohei Yokoyama b,∗, Karen Lu a, Richard T. Eastman c, James G. Bollinger b,d,Wesley C. Van Voorhis a,c, Michael H. Gelb b,d, Frederick S. Buckner a,∗∗

a Department of Medicine, University of Washington, 1959 N.E. Pacific St., Seattle, WA 98195, USAb Department of Chemistry, University of Washington, Seattle, WA 98195, USA

c Department of Pathobiology, University of Washington, Seattle, Washington 98195, USAd Department of Biochemistry, University of Washington, Seattle, Washington 98195, USA

Received 22 December 2006; received in revised form 17 February 2007; accepted 26 February 2007Available online 1 March 2007

bstract

The C-terminal “CaaX”-motif-containing proteins usually undergo three sequential post-translational processing steps: (1) attachment of a prenylroup to the cysteine residue; (2) proteolytic removal of the last three amino acids “aaX”; (3) methyl esterification of the exposed �-carboxylroup of the prenyl-cysteine residue. The Trypanosoma brucei and Leishmania major Ras converting enzyme 1 (RCE1) orthologs of 302 and85 amino acids-proteins, respectively, have only 13–20% sequence identity to those from other species but contain the critical residues for thectivity found in other orthologs. The Trypanosoma brucei a-factor converting enzyme 1 (AFC1) ortholog consists of 427 amino acids with9–33% sequence identity to those of other species and contains the consensus HExxH zinc-binding motif. The trypanosomatid RCE1 and AFC1rthologs contain predicted transmembrane regions like other species. Membranes from Sf9 cells expressing the RCE1 ortholog of T. brucei or L.ajor showed proteolytic activity against farnesylated RAS-CVIM, whereas membranes containing T. brucei AFC1 ortholog were inactive. The

esults suggest that RCE1 is responsible for proteolytic removal of the C-terminal aaX from prenyl-CaaX proteins in these parasites. All the threenzymatic post-translational processes are thought to be required for proper cellular functioning of CaaX-proteins in eukaryotic cells. We carriedut RNA interference experiments in Trypanosoma brucei of the enzymes involved in farnesyl protein post-translational modification to evaluateheir importance in cell proliferation. Knockdown of T. brucei PFT � subunit and RCE1 mRNAs resulted in >20-fold suppression of cell growth

nd dramatic morphologic changes. Knockdown of PPMT mRNA caused less dramatic effects on growth but induced noticeable changes in cellorphology.2007 Published by Elsevier B.V.

eywords: Trypanosoma brucei; Leishmania major; Protein prenylation; Protein farnesyltransferase; Ras converting enzyme; A-factor converting enzyme; RNAnterference

Abbreviations: RCE1, ras converting enzyme 1 ortholog; AFC1, a-actor converting enzyme 1 ortholog; PFT, protein farnesyltransferase; PPMT,renyl protein carboxyl methyltransferase; RNAi, RNAinterference; T. brucei,rypanosoma brucei; L. major, Leishmania major; AdoMet, S-adenosyl-l-ethionine∗ Corresponding author at: Department of Chemistry, University of Washing-

on, Campus Box 351700, Seattle, WA 98195-7185, USA.el.: +1 206 616 9214; fax: +1 206 685 8681.

∗∗ Corresponding author. Tel.: +1 206 616 4264; fax: +1 206 685 8665.E-mail addresses: koheiy@u.washington.edu (K. Yokoyama),

buckner@u.washington.edu (F.S. Buckner).

1

stiFceCg

166-6851/$ – see front matter © 2007 Published by Elsevier B.V.oi:10.1016/j.molbiopara.2007.02.009

. Introduction

Prenylation (farnesylation and geranylgeranylation) andubsequent modifications are required for proper membraneargeting and cellular functioning of a number of proteinsn eukaryotic cells such as Ras superfamily GTPases [1,2].or proteins with the C-terminal CaaX motif (where C is

ysteine, a is usually an aliphatic amino acid, and X is a vari-ty of residues), farnesylation or geranylgeranylation at theys residue by protein farnesyltransferase (PFT) or proteineranylgeranyltransferase-I is a prerequisite for further enzy-

1 chemi

mrg

iv(ttnailRphraA“lmishAofdAdpisopbeo

RlRimrsmtb

[poccf

mtbtR

oLstAbefWspiptppac�

2

2

duTb[PH(A(D

2a

ocgsSe

16 J.R. Gillespie et al. / Molecular & Bio

atic processing, i.e. the proteolytic removal of the last threeesidues “aaX” followed by methylation of exposed �-carboxylroup of the prenylcysteine [3–5].

Two different enzymes have been discovered to be involvedn removal of the tripeptide “aaX” in Saccharomyces cere-isiae [6], namely, RCE1 (Ras converting enzyme 1) and AFC1a-factor converting enzyme 1, also called Zmpste24). Thesewo enzymes show overlapping but distinct specificity towardhe CaaX sequence [7,8]. RCE1 seems to process most of theaturally occurring CaaX sequences of both farnesyl and ger-nylgeranyl proteins [7,9–11]. Disruption of the RCE1 genen mice resulted in death at the embryonic stage, and fibrob-asts from the mouse embryo failed to proteolyze prenylatedas, Rap, and transducin �, suggesting that RCE1 is indis-ensable for maturation of these GTPases [11]. On the otherand, AFC1 cleaves the a-factor precursor (at a site that is 26esidues upstream of farnesylcysteine) after removal of “aaX”nd carboxyl methylation [11,12]. Human and S. cerevisiaeFC1 orthologs have also been shown to have ability to remove

aaX” from the farnesylated a-factor precursor in the yeast cellsacking the RCE1 gene [7,8], but RCE1 is thought to be the

ajor enzyme responsible for removal of “aaX” from proteinsn yeast and mammalian cells. RCE1 and AFC1 orthologs fromeveral species have been reported [4,9,13–15]. A substrate ofuman AFC1 has been identified to be farnesylated prelamin, which undergoes two cleavages like yeast a-factor; removalf “aaX” and then cleavage at a site 15 residues upstream ofarnesylcysteine [12,16]. Failure of the proteolytic processingue to mutations of the prelamin A cleavage site [17] or theFC1 enzyme [18,19], which have been found in the geneticiseases Hutchinson–Gilford progeria and mandibuloacral dys-lasia, results in aberrant accumulation of farnesyl-prelamin An the nucleus and abnormal nucleus morphology [20–22]. Aimilar phenotype and defect in prelamin A processing werebserved in AFC1 deficient mice, indicating that AFC1 is indis-ensable for the processing of prelamin A [23]. Lamin B1 haseen shown to be processed only by RCE1 based on studies usingmbryonic fibroblasts from mice with homozygous deficiencyf either RCE1 or AFC1 [24].

Recent studies with mouse embryonic fibroblasts lackingCE1 suggest that removal of “aaX” is required for proper

ocalization of farnesylated Ras but not geranylgeranylatedho GTPases [25]. Farnesylated K-Ras is also mislocalized

n mammalian cells lacking prenyl-protein specific carboxylethyltransferase (PPMT, also called as ICMT) [26]. These

eports and studies with farnesylated and geranylgeranylatedhort peptides [27,28] suggest the importance of �-carboxylethylation, particularly for farnesylated proteins as an addi-

ional hydrophobic element to cause tight and proper membraneinding.

RCE1 and PPMT localize to the endoplasmic reticulum9,29–31] and act on both farnesylated and geranylgeranylatedroteins [32]. Gene disruption of PPMT in mice results in embry-

nic lethality [33] as seen with RCE1 disruption. Since thearboxyl methylation is thought to be reversible, although a spe-ific esterase has not been found, PPMT has been implicated inunctional regulation of certain prenylated proteins. Carboxyl

blgw

cal Parasitology 153 (2007) 115–124

ethylation of several farnesylated proteins such as K-Ras [34],ransducin � subunit [35,36], and lamin B [24] has been shown toe critical for activating downstream effectors. It is also notablehat disruption of PPMT causes a decrease in the half-life ofhoA but stabilizes Ras in mammalian cells [37,38].

We report here the cloning and characterization of RCE1rthologs from the protozoa Trypanosoma brucei (T. brucei) andeishmania major (L. major), the caustative agents of Africanleeping sickness and certain forms of leishmaniasis, respec-ively. We also report the cloning and characterization of theFC1 ortholog from T. brucei. Heterologous expression of T.rucei and L. major RCE1 and T. brucei AFC1 and in vitronzymatic studies suggest that RCE1 is the enzyme responsibleor removal of “aaX” from prenylated proteins in these parasites.

e have previously reported the cloning of T. brucei PFT � and �ubunits [39] and PPMT [40]. Thus, the full set of enzymes in theost-translational processing of farnesylated proteins is presentn this protozoan. In our previous studies, T. brucei and malariaarasites showed high sensitivity to PFT inhibitors comparedo mammalian cells, suggesting that PFT as well as subsequentrocessing enzymes may be excellent drug targets against thesearasites [39,41–44]. To evaluate the potential of these enzymess targets for the development of anti-parasite drugs, we havearried out RNA interference experiments with T. brucei PFT-subunit, RCE1 and PPMT.

. Materials and methods

.1. Materials

RAS-CVIM was expressed in E. coli and purified asescribed [44], and T. brucei PFT was produced in the bac-lovirus/Sf9 insect cell system and purified as described [45].. brucei PPMT was expressed in Sf9 cells, and the mem-ranes containing the enzyme were prepared as described40]. Sf9 cell membranes containing recombinant S. cerevisiaePMT (ICMT) were obtained as a generous gift from Dr. C.rycyna (Purdue Univ.). S-Adenosyl-l-[methyl-3H]methionine

[3H]AdoMet) was purchased from NEN/Parkin-Elmer. N-cetyl-S-farnesyl-l-cysteine (AFC) and farnesyl pyrophosphate

FPP) were from BIOMOL. Cysmethynil is a generous gift fromr. P. Casey (Duke Univ.).

.2. Cloning of RCE1 orthologs of T. brucei and L. major,nd an AFC1 ortholog of T. brucei

The genes described herein were cloned prior to completionf the genome projects for T. brucei and L. major. All genes wereloned by the following methods: first, fragments of the targetedenes were identified in GenBank by TBLASTN search usingequences of RCE1 (GI:6323930) and AFC1 (GI:1352918) ofaccharomyces cerevisiae; second, the remaining 5′- and 3′-nds of the genes were amplified off cDNA extracted from T.

rucei BF427 or L. major Friedlin promastigotes using spliced-eader RACE and oligo-dT RACE [46]; third, the full lengthenes were PCR amplified from genomic DNA, and 2 clonesere fully sequenced for each target.

chemi

2s

BRAwcFttTTteffsu1Bs

2

1wtMrompm(nMror(aT

witdwt

2a

rcmptpu[

2

P(iostvwFMopcgcihw

TO

T

T

T

T

J.R. Gillespie et al. / Molecular & Bio

.3. Expression of the proteins in the Sf9/baculovirusystem

According to the manufacturer’s instructions for the Bac-to-ac Baculovirus expression system (Invitrogen), the genes forCE1 orthologs of T. brucei (isoform 2) and L. major and anFC1 ortholog of T. brucei were transferred to pFastBac, whichere then transposed into a bacmid in DH10Bac competent E.

oli cells. Baculoviruses were plaque-purified and amplified.or expression of the proteins, Sf9 cells were infected with

he baculovirus carrying either RCE1 or AFC1 gene at mul-iplicity of infections of 1–3 and were cultured for 2–3 days.he cells were disrupted by a Dounce homogenizer in 50 mMris–HCl buffer pH 7.5 containing freshly added 1 mM dithio-

hreitol (DTT) and protease inhibitors (1 mM PMSF, 10 �g/mlach of aprotinin, leupeptin, and pepstatin A). The membraneraction was obtained by centrifugation at 10,000 × g for 10 minollowed by 120,000 × g for 80 min at 4 ◦C, and suspended in amall volume of buffer (50 mM Tris–HCl pH 7.5, 1 mM DTT)sing a Dounce homogenizer to give a protein concentration of0–20 mg/ml. Protein concentration was determined using theradford dye reagent from Bio-Rad. Aliquots of the membrane

uspension were frozen in liquid nitrogen and stored at −80 ◦C.

.4. Prenyl protein endoprotease assay

A substrate farnesyl-RAS-CVIM was prepared by incubating0 �M (22 �g) RAS-CVIM and 20 �M farnesyl pyrophosphateith 2 �M (28 �g) purified T. brucei PFT in buffer (100 �l) con-

aining 30 mM potassium phosphate pH 7.7, 5 mM DTT, 1 mMgCl2, 20 �M ZnCl2. After incubation at 30 ◦C for 2 h, the

eaction mixture was frozen at −80 ◦C and used as a sourcef the substrate. The prenyl protein endoprotease activity waseasured by 2 step-incubation as follows. Standard mixtures for

renyl protein endoprotease reaction contain 5 �l of the aboveixture as a source of farnesyl-RAS-CVIM, Sf9 cell membranes

1 �g protein) containing either recombinant RCE1 or AFC1 orone, in 20 �l buffer containing 50 mM Tris–HCl pH 7.5, 5 mMgCl2, and 5 mM DTT. After incubation at 30 ◦C for 30 min, the

eaction was stopped by adding 200 mM EDTA (5 �l), 60 mMrtho-phenanthroline (1 �l), and 6 mM TPCK (1 �l). To the

eaction mixture was added a solution (3 �l) containing 0.4 �Ci2 �M, final concentration) [3H]AdoMet (NEN/Parkin-Elmer)nd recombinant yeast PPMT/Sf9 membranes (1 �g protein).he mixture was incubated at 30 ◦C for 1 h, and the reaction

wdE(

able 1ligonucleotide primers for RNAi constructs

arget RNAi nucleotide position Primers

bPFT -� 338–783 Sense: GCTGAntisense: CC

bRCE1 51–591 Sense: GGCAAntisense: CC

bPPMT 58–529 Sense: TTTATAntisense: TT

cal Parasitology 153 (2007) 115–124 117

as stopped with 200 �l of 10% concentrated aqueous HCln ethanol. Incorporation of [3H]methyl group into the pro-ein product was measured by the glass fiber filter method asescribed for PFT assay [44] using 1 mM HCl in ethanol asashing solvent. Error values of this assay are typically lower

han 10% for the same series of measurement.

.5. Prenyl protein carboxyl methyltransferase (PPMT)ssay

PPMT activity was measured as described [40]. The standardeaction mixture (20 �l) contains 1 �M N-acetyl-farnesyl-ysteine (AFC), 0.4 �Ci (0.3 �M) [3H]AdoMet, and Sf9 cellembranes containing T. brucei PPMT or yeast PPMT (1 �g

rotein). After incubation at 30 ◦C for 20 min, the reaction mix-ure was subjected to hexane-extraction, and amounts of theroduct [3H]methyl ester of AFC were quantified by the methodsing 1 N NaOH-treatment/[3H]methanol vapor diffusion47].

.6. RNA interference

Regions of the T. brucei PFT-� subunit gene, RCE1, andPMT genes for RNAi were selected using the program RNAitTable 1) [48]. The amplicons were ligated using TA cloningnto the vector p2T7TABlue (gift of D. Horn, London Schoolf Hygiene and Tropical Medicine) [49]. T. brucei blood-tream form parasites expressing the T7 RNA polymerase andhe Tet repressor under a single selection marker were pro-ided by G. Cross (Rockefeller University) [50]. The cellsere grown at 37 ◦C in HMI-9 containing 10% heat-inactivatedCS and 2.5 �g/ml G418 under 5% CO2 atmosphere [51].id-log phase T. brucei (2.5 × 107) were suspended in 500 �l

f cytomix (120 mM KCl, 0.15 mM CaCl2, 10 mM potassiumhosphate pH 7.6, 25 mM Hepes, 2 mM EDTA, 5 mM MgCl2)ontaining 10 �g of NotI linearized p2T7TABluewith a targetene DNA. The mixture was electroporated in a 4 mm gapuvette with 1.6 kV and 24� resistance. Cells were resuspendedn HMI-9 medium under selection pressure with 2.5 �g/mlygromycin and 2.5 �g/ml G418. The expression of dsRNAas induced by addition of 1 �g/ml tetracycline to the cultures

ith 1 × 105 cells/ml. Cultures were passed at a 1:10 dilutionaily, and cell concentrations were monitored using a Perkin-lmer ATPLite Luminescence ATP detection Assay System

No. 6016941).

Northern probe position

AGATGCTCGGAATCACAG 824–1597TGGTAATAACCGCGACACAA

TCTTTTTATGGCGGAAG 607–970ACGATATTCCCAAGCCACAG

GTTTTACGCTGGGTGTTTC 547–824CGGGATGACGCAATATTGAGTA

1 chemi

BrReqb

wp1

3

3R

b

(T92srdnoGsPts

F(aciwp

18 J.R. Gillespie et al. / Molecular & Bio

Northern blots were performed using parasite total RNA.lots were hybridized with 32P-labeled DNA probes made to

egions of the gene separate from the regions targeted by theNAi constructs (Table 1). The nitrocellulose membranes werexposed to X-ray films, and amounts of the target RNAs wereuantified by densitometry and normalized against the �-tubulinand.

Metabolic labeling of T. brucei (107 bloodstream form cells)as done with 100 �Ci [3H]mevalonolactone (1.6 �М) per sam-le [52]. Proteins extracted from cell pellets were separated on2.5% SDS-PAGE and imaged by fluorography [53].

. Results and discussion

.1. Cloning and analyses of amino acid sequences of

CE1 orthologs from T. brucei and L. major

Two isoforms of the T. brucei RCE1 gene were identifiedy sequencing multiple clones amplified from genomic DNA

om2t

ig. 1. Multiple sequence alignments. (A) RCE1 orthologs: Human (GenBank acNP 014001), T. brucei (isoform 2), and L. major. The residues highlighted in black bonalyses [8,54], and those indicated by boxes are required for the full activity [8]. (erevisiae (NP 012651), and T. brucei. The zinc binding HExxH motif is indicatedn human ortholog) residues are indicated by boxes, which have been found to be mhich are associated with accumulation of farnesylated prelamin A [18,19]. Transmemrogram (http://www.cbs.dtu.dk/services/TMHMM-2.0/).

cal Parasitology 153 (2007) 115–124

GenBank accession numbers DQ831038 and DQ831039).he isoforms are of identical length (909 bp), but differ atnucleotides, and encode predicted proteins that differ atresidues. The isoforms identified in these studies are the

ame length as, but differ slightly in sequence from the cor-esponding gene deposited by the T. brucei Genome Project,esignated “CAAX prenyl protease 2, putative” (systematicame: Tb927.3.1540). A single isoform of the Leishmania RCE1rtholog was cloned from Leishmania genomic DNA (858 bp,enBank accession number DQ831041), based on the matching

equence (LmjF26.2690) annotated by the Leishmania Genomeroject (675 bp), which lacks a part of the 5′-end. The addi-

ional sequence on the 5′-end of the gene was determined by thepliced-leader RACE method on cDNA.

Sequence similarity between the orthologs from T. brucei

r L. major and other species (human, S. cerevisiae, and D.elanogaster are shown) is only 13–20%, and that between thesetrypanosomatids is 30%. T. brucei and L. major sequences con-

ain several predicted transmembrane helices (shown as shaded

cession number: NP 005124), D. melanogaster (CAB64383), S. cerevisiaexes are known to be crucial for enzyme activity based on site-directed mutationB) AFC1 orthologs: Human (AAH37283), D. melanogaster (AAF57922), S.with the black-boxed residues. W340 next to this motif and N265 (numbersutated in human patients with mandibuloacral dysplasia or progeria deseases,brane helical regions (shaded residues) were predicted with the TMHMM-2.0

J.R. Gillespie et al. / Molecular & Biochemical Parasitology 153 (2007) 115–124 119

Conti

symerwFbSirsi

3o

(gpio

Fig. 1. (

equences) as do other species orthologs (Fig. 1A). Human andeast enzymes have been shown to localize to the endoplas-ic reticulum [9,31]. Mutational analyses of the S. cerevisiae

nzyme and other orthologs [8,54] revealed several criticalesidues for enzyme activity and CaaX sequence specificity,hich are indicated by residues highlighted in black boxes inig. 1A. These residues, E156, H194, and H248 (residue num-ers in S. cerevisiae), are all conserved in both trypanosomatids.everal additional residues, E157, Y160, F190, H197, and N252,

ndicated in boxes in Fig. 1A, have also been shown to beequired for the full activity [8]. All of these residues are con-erved, except that Y160 is changed to Phe in several speciesncluding T. brucei (Fig. 1A).

(cai

nued ).

.2. Cloning and amino acid sequence analysis of an AFC1rtholog from T. brucei

The gene encoding the T. brucei AFC1 ortholog is 1281 bpGenBank accession number DQ831040) and corresponds to theene Tb09.211.0680 (designated as “CAAX prenyl protease 1,utative”). The T. brucei AFC1 ortholog shows 46% sequencedentity to the L. major ortholog and 29–33% identity to threether species and contains five predicted transmembrane helices

Fig. 1B). AFC1 is a zinc-metalloprotease containing the HExxHonsensus motif (indicated by black boxes in Fig. 1B), which islso found in the T. brucei ortholog. Mutation of the Trp residuemmediately adjacent to this motif (W340 of the human protein)

1 chemical Parasitology 153 (2007) 115–124

waIbiNratv[a

3mS

cwmtTt3b

fi

FamAwr

Fig. 3. Endoprotease activity on farnesyl-RAS-CVIM by T. brucei RCE1 andL. major RCE1 orthologs but not by the T. brucei AFC1 ortholog. The indicatedamount of Sf9 cell membranes containing recombinant L. major RCE1 (�), T.brucei RCE1 (�), T. brucei AFC1 (�), or no recombinant protein (©) was incu-bated at 30 ◦C for 30 min with farnesyl-RAS-CVIM. The proteolyzed productwas measured after conversion to [3H]methyl-esterified protein by incubationwaY

ft

20 J.R. Gillespie et al. / Molecular & Bio

as found in a patient with mandibuloacral dysplasia in whichccumulation of unprocessed farnesylated prelamin A occurs.mportance of this residue in the enzyme activity was showny mutational analysis with the yeast ortholog [18]. This Trps conserved in the T. brucei ortholog (Fig. 1B). Mutation of265 was also found in similar progeroid disease [19], and this

esidue is also conserved in the parasite enzyme. Both AFC1nd RCE1 of S. cerevisiae and human show proteolytic activityo remove the aaX tripeptide from farnesylated a-factor CaaXariants although AFC1 acts on more restricted CaaX sequences7,8], but there is no apparent homology seen between the RCE1nd AFC1 sequences.

.3. Expression of RCE1 orthologs of T. brucei and L.ajor, and the AFC1 ortholog of T. brucei in thef9/baculovirus system

The recombinant proteins were produced using the Sf9ell/baculovirus expression system, and the membrane fractionsere subjected to SDS-PAGE analysis (Fig. 2). The observedolecular masses of the expressed proteins were ∼38 kDa for

he T. brucei and L. major RCE1 orthologs and ∼40 kDa for the. brucei AFC1 ortholog, while calculated molecular masses ofhe amino acid sequences deduced from the cDNA were 33.9,

1.6, and 48.9 kDa for T. brucei RCE1, L. major RCE1, and T.rucei AFC1 orthologs, respectively.

RAS-CVIM protein was efficiently farnesylated by puri-ed recombinant T. brucei PFT as described [45]. The product

ig. 2. SDS-PAGE analysis of recombinant T. brucei RCE1, L. major RCE1,nd T. brucei AFC1. Membranes (7.5 �g protein) from Sf9 cells expressing L.ajor RCE1 (lane 1), none (lane 2), T. brucei RCE1 (lane 3), and T. bruceiFC1 (lane 4) were subjected to SDS-PAGE analysis on a 12.5% gel. Proteinsere stained by Coomassie blue. Arrow heads indicate putative bands of the

ecombinant proteins.

act[foCacRstT[ttt

htotnAte

3

tgr

ith yeast PPMT as described in “Materials and Methods”. The scale of enzymectivity shown on the right Y-axis is only for L. major RCE1, and that of the left-axis is for others.

arnesyl-RAS-CVIM was used as a substrate to measure the pro-eolytic activity removing the last three amino acids by a coupledssay utilizing PPMT since removal of “aaX” is required for thearboxyl methylation of the C-terminal prenylcysteine. The pro-eolyzed product was radiolabeled by a second incubation with3H]AdoMet and yeast PPMT. As shown in Fig. 3, membranesrom Sf9 cells overexpressing the recombinant RCE1 orthologf both T. brucei and L. major proteolyzed farnesylated RAS-VIM. The enzyme activity of T. brucei and L. major RCE1ssociated with the membranes and was not detectable in theytosolic fraction from Sf9 cells. Both T. brucei and L. majorCE1were unable to utilize non-prenylated RAS-CVIM as a

ubstrate, thus showing specificity for prenylated “CaaX” pro-eins. Incubation of membranes from Sf9 cells co-expressing. brucei RCE1 and PPMT with farnesyl-RAS-CVIM and3H]AdoMet resulted in incorporation of the radioactivity intohe protein (Fig. 4), showing that these two enzymes catalyzehe sequential modifications on prenylated “CaaX-proteins” inhe membranes.

Membranes containing the T. brucei AFC1 ortholog did notave significant activity on farnesylated RAS-CVIM comparedo control Sf9 cell membranes. Since orthologs to yeast a-factorr mammalian prelamin A (the natural substrates for AFC1 inhese organisms) have not been found in trypanosomatids, we doot have putative natural substrate(s) with which to test T. bruceiFC1. Therefore, the specific role of AFC1 in T. brucei has yet

o be elucidated, but this protein appears to have undetectablendoprotease activity for a farnesylated CaaX protein or peptide.

.4. RNA interference of T. brucei protein prenylation genes

The T. brucei PFT-� subunit gene and the two prenyl pro-ein processing genes (RCE1 and PPMT) were subjected toene silencing by overexpression of dsRNA from tetracycline-egulated constructs. mRNA levels at 72 h post-induction were

J.R. Gillespie et al. / Molecular & Biochemical Parasitology 153 (2007) 115–124 121

Fig. 4. [3H]Methylation of farnesyl-RAS-CVIM by membranes from Sf9 cellsco-expressing T. brucei RCE1 and T. brucei PPMT. Membranes (2 �g pro-tein) containing recombinant T. brucei RCE1 (�), T. brucei PPMT (�), orbC3

dgRfKim

jfltcm�(

smslbmotsoRsfSpcpr4t�o

Fig. 5. Effects of gene silencing of protein prenylation enzymes on T. bruceicell growth. Bloodstream forms of T. brucei were induced to express dsRNAagainst the target genes: PFT-� subunit, RCE1, or PPMT. dsRNA expressionwas induced by the addition (+) of tetracycline (TCN) on day 1, and growth wascompared to control cultures without (−) TCN. Growth was also compared tothe background “single marker” (SM) strain (absence of the RNAi expressionconstruct). Cumulative cell densities are shown on a log-scale as the product

oth (�) were incubated at 30 ◦C for the indicated time with farnesyl-RAS-VIM and [3H]S-AdoMet. The amount of product was measured by countingH-incorporation into protein.

ecreased by 74, 63, and 63% for the PFT-�, RCE1, and PPMTenes, respectively (Fig. 5, insets). Knockdown of the PFT-� andCE1 mRNA transcripts led, respectively, to ∼35-fold and 23-

old inhibition of growth at day 5 compared to controls (Fig. 5).nockdown of the PPMT mRNA transcripts led to 2–3-fold

nhibition of growth at day 5. Similar results were observed forultiple clones of each construct (data not shown).The morphology of cells was similar for all three targets sub-

ected to knockdown with enlargement and formation of multipleagellae (Fig. 6). The morphological changes were very similar

o those observed with treatment of bloodstream-form T. bru-ei with PFT inhibitors [52]. The number of cells with alteredorphology from RNAi was most pronounced against PFT-(∼85%) followed by RCE1 (∼70%), and least with PPMT

∼50%). This correlates to the degree of growth inhibition.To examine effects on in vivo prenylation events, cells

ubjected to RNAi were metabolically labeleled with 3H-evalonolactone which is incorporated into isoprenoids and

ubsequently into prenylated proteins [52]. RNAi against PFT-�,ed to reduced intensity of at least five protein bands as observedy fluorography (see arrows in Fig. 7). This demonstrates that theRNA knockdown led to impaired protein prenylation for each

f these gene products. The affected bands correspond closelyo those bands inhibited by PFT inhibitors shown in previoustudies [52], affirming that the changes are specific to inhibitionf protein farnesylation. There appears to be little effect of theNAi on the intense low MW proteins that are believed to be

ubstrates of geranylgeranyltransferase type II [52]. For proteinsrom the cells treated with RNAi against RCE1 or PPMT (Fig. 7),DS-PAGE migration shifts of incompletely processed prenylroteins were not detectable presumably because the structuralhanges of these proteins compared to fully processed prenylroteins are small. However, significant reduction of severaladiolabeled proteins with apparent molecular weights of 63, 48,

0, 31, and a doublet at 26 kDa was observed (Fig. 7). Most ofhese appear to correspond to protein bands reduced by the PFT-

RNAi, which may represent reduction of protein expressionr altered turnover of incompletely processed prenyl proteins as

of the cell number and the total dilution. Insets show Northern blots of RNAtaken from parasites 72 h after initiation of the experiment. The Northern blotswere probed for the target genes and �-tubulin for normalization. PFT-� RNAlevels were reduced to 23.7% of controls; RCE1 RNA was 37.5% of controls;and PPMT RNA was 36.9% of controls.

122 J.R. Gillespie et al. / Molecular & Biochemical Parasitology 153 (2007) 115–124

F l morpa agnifim singleT

hTrtci

kst

vcdigo

ig. 6. Effects of gene silencing of protein prenylation enzymes on T. brucei celbnormal morphologies in the cultures with gene silencing are shown in high-morphology of bloodstream form cultures is demonstrated in the photos of the “7 RNA polymerase and Tet repressor, but no RNAi construct.

as been described for some mammalian prenyl proteins [37,38].here may be two bands (∼55 and ∼75 kDa) with increased

adioactivity associated with RNAi (Fig. 7). This could be dueo increased expression of the proteins as reported in human can-er cells in which PFT and protein geranylgeranyltransferase-Inhibitors upregulate some GTPases [55,56].

The failure to completely arrest T. brucei cell growth withnockdown of PFT-� or RCE1 may be due to the incompleteilencing of the gene targets. Nonetheless, the observed magni-ude of growth inhibition supports the notion that these genes are

mwtT

hology. Photomicrographs are shown of the cultured cells at 72 h. Examples ofcation on the far right next to the respective low magnification photos. Normalmarker” (SM) parasites (bottom). The latter parasites contain the heterologous

ery important for growth of bloodstream form T. brucei. In thease of the PPMT RNAi experiment where cell growth was notramatically affected with 63% knockdown of the target RNA,t seems likely that this gene is more dispensable for normal cellrowth than are the PFT or RCE1 genes. Combining the findingf the PFT RNAi experiments with our observations that small

olecular weight inhibitors of PFT cause T. brucei cell deathithin ∼48 h [52] supports the contention that PFT is an essen-

ial T. brucei enzyme. It seems likely that RCE1 is essential in. brucei, although specific T. brucei RCE1 inhibitors are not

J.R. Gillespie et al. / Molecular & Biochemi

Fig. 7. Effects of gene silencing on in vivo protein prenylation in T. bru-cei. Bloodstream form parasites were subjected to RNAi against PFT-�subunit, RCE1, or PPMT mRNA transcripts for 72 h, then labeled with 3H-mevalonolactone for an 24 h. Arrows indicate prenylated protein bands that arest

ao

itbCgtcPo5

eiddhcmewioted

A

t(

R

[

[

[

[

[

[

[

[

[

[

[

ignificantly reduced in intensity in the RNAi-induced (+) cultures compared tohe RNAi-uninduced cultures (−).

vailable to provide chemical validation of the essential naturef RCE1.

Prenylcysteine analogs have been studied to developnhibitors against mammalian and yeast PPMT, although theseypes of compounds may have limited use in cellular studiesecause of poor cell permeability and poor specificity [3,4].ysmethynil, a compound without farnesyl and carboxylateroups, has been described as a useful tool to investigate selec-ive inhibition of PPMT in mammalian cells [57]. However, thisompound produced only modest inhibition against T. bruceiPMT (KI = 0.9 �M after pre-incubation) and growth inhibitionf T. brucei bloodstream form cells (a concentration causing0% growth inhibition of ∼22 �M).

In this paper, we have characterized the enzymatic machin-ry responsible for final processing of farnesylated proteinsn T. brucei. The T. brucei RCE1 ortholog appears to be theominant prenyl-CaaX protease, whereas the AFC1 orthologoes not appear to have this activity. In addition, this protozoanas an enzymatically functional prenylprotein methyltransferaseapable of performing the final step in the post-translationalodification of farnesylated proteins [47]. RNA interference

xperiments showed pronounced growth impairment associatedith knockdown of T. brucei PFT or RCE1, and modest growth

mpairment with PPMT knockdown. Together with previousbservations that inhibition of PFT severely affects growth ofrypanosomatid parasites [41–44], these studies indicate thatffective inhibitors of PFT or RCE1 have high potential to beeveloped as drugs against T. brucei infections.

cknowledgements

This work was supported by a grant from the National Insti-utes of Health (AI054384) and by Royalty Research FundUniversity of Washington).

[

[

cal Parasitology 153 (2007) 115–124 123

eferences

[1] Glomset JA, Gelb MH, Farnsworth CC. Prenyl proteins in eukaryotic cells:a new type of membrane anchor. Trends Biochem Sci 1990;15:139–42.

[2] Casey PJ, Seabra MC. Protein prenyltransferases. J Biol Chem1996;271:5289–92.

[3] Winter-Vann AM, Casey PJ. Post-prenylation-processing enzymes as newtargets in oncogenesis. Nat Rev Cancer 2005;5:405–12.

[4] Young SG, Ambroziak P, Kim E, Clarke S. Postisoprenylation proteinprocessing: CXXX (CaaX) endoproteases and isoprenylcysteine carboxylmethyltransferase. In: Tamanoi F, Sigman DS, editors. The enzymes, vol.XXI. San Diego: Academic Press; 2001. p. 155–213.

[5] Glomset JA, Farnsworth CC. Role of protein modification reactions in pro-gramming interactions between ras-related GTPases and cell membranes.Annu Rev Cell Biol 1994;10:181–205.

[6] Boyartchuk VL, Ashby MN, Rine J. Modulation of Ras and a-factor func-tion by carboxyl-terminal proteolysis. Science 1997;275:1796–800.

[7] Trueblood CE, Boyartchuk VL, Picologlou EA, Rozema D, Poulter CD,Rine J. The CaaX proteases, Afc1p and Rce1p, have overlapping but distinctsubstrate specificities. Mol Cell Biol 2000;20:4381–92.

[8] Plummer LJ, Hildebrandt ER, Porter SB, Rogers VA, McCrackenJ, Schmidt WK. Mutational analysis of the ras converting enzymereveals a requirement for glutamate and histidine residues. J Biol Chem2006;281:4596–605.

[9] Otto JC, Kim E, Young SG, Casey PJ. Cloning and characteriza-tion of a mammalian prenyl protein-specific protease. J Biol Chem1999;274:8379–82.

10] Jang GF, Gelb MH. Substrate specificity of mammalian prenyl protein-specific endoprotease activity. Biochemistry 1998;37:4473–81.

11] Kim E, Ambroziak P, Otto JC, et al. Disruption of the mouse Rce1 generesults in defective Ras processing and mislocalization of Ras within cells.J Biol Chem 1999;274:8383–90.

12] Corrigan DP, Kuszczak D, Rusinol AE, et al. Prelamin A endopro-teolytic processing in vitro by recombinant Zmpste24. Biochem J2005;387:129–38.

13] Kumagai H, Kawamura Y, Yanagisawa K, Komano H. Identification of ahuman cDNA encoding a novel protein structurally related to the yeastmembrane-associated metalloprotease, Ste24p. Biochim Biophys Acta1999;1426:469–74.

14] Cadinanos J, Varela I, Mandel DA, et al. AtFACE-2, a functional preny-lated protein protease from Arabidopsis thaliana related to mammalianRas-converting enzymes. J Biol Chem 2003;278:42091–7.

15] Cadinanos J, Schmidt WK, Fueyo A, Varela I, Lopez-Otin C, FreijeJM. Identification, functional expression and enzymic analysis oftwo distinct CaaX proteases from Caenorhabditis elegans. Biochem J2003;370:1047–54.

16] Young SG, Fong LG, Michaelis S. Prelamin A, Zmpste24, misshapencell nuclei, and progeria–new evidence suggesting that protein farnesy-lation could be important for disease pathogenesis. J Lipid Res 2005;46:2531–58.

17] Eriksson M, Brown WT, Gordon LB, et al. Recurrent de novo point muta-tions in lamin A cause Hutchinson–Gilford progeria syndrome. Nature2003;423:293–8.

18] Agarwal AK, Fryns JP, Auchus RJ, Garg A. Zinc metalloproteinase,ZMPSTE24, is mutated in mandibuloacral dysplasia. Hum Mol Genet2003;12:1995–2001.

19] Shackleton S, Smallwood DT, Clayton P, et al. Compound heterozygousZMPSTE24 mutations reduce prelamin A processing and result in a severeprogeroid phenotype. J Med Genet 2005;42:e36.

20] Yang SH, Bergo MO, Toth JI, et al. Blocking protein farnesyltrans-ferase improves nuclear blebbing in mouse fibroblasts with a targetedHutchinson–Gilford progeria syndrome mutation. Proc Natl Acad Sci USA2005;102:10291–6.

21] Toth JI, Yang SH, Qiao X, et al. Blocking protein farnesyltransferaseimproves nuclear shape in fibroblasts from humans with progeroid syn-dromes. Proc Natl Acad Sci USA 2005;102:12873–8.

22] Mallampalli MP, Huyer G, Bendale P, Gelb MH, Michaelis S. Inhibit-ing farnesylation reverses the nuclear morphology defect in a HeLa cell

1 chemi

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

24 J.R. Gillespie et al. / Molecular & Bio

model for Hutchinson–Gilford progeria syndrome. Proc Natl Acad Sci USA2005;102:14416–21.

23] Bergo MO, Gavino B, Ross J, et al. Zmpste24 deficiency in mice causesspontaneous bone fractures, muscle weakness, and a prelamin A processingdefect. Proc Natl Acad Sci USA 2002;99:13049–54.

24] Maske CP, Hollinshead MS, Higbee NC, Bergo MO, Young SG, Vaux DJ. Acarboxyl-terminal interaction of lamin B1 is dependent on the CAAX endo-protease Rce1 and carboxymethylation. J Cell Biol 2003;162:1223–32.

25] Michaelson D, Ali W, Chiu VK, et al. Postprenylation CAAX processingis required for proper localization of Ras but not Rho GTPases. Mol BiolCell 2005;16:1606–16.

26] Chen Z, Otto JC, Bergo MO, Young SG, Casey PJ. The C-terminal polyly-sine region and methylation of K-Ras are critical for the interaction betweenK-Ras and microtubules. J Biol Chem 2000;275:41251–7.

27] Ghomashchi F, Zhang X, Liu L, Gelb MH. Binding of prenylated andpolybasic peptides to membranes: affinities and intervesicle exchange.Biochemistry 1995;34:11910–8.

28] Silvius JR, l’Heureux F. Fluorimetric evaluation of the affinities of iso-prenylated peptides for lipid bilayers. Biochemistry 1994;33:3014–22.

29] Dai Q, Choy E, Chiu V, et al. Mammalian prenylcysteine carboxyl methyl-transferase is in the endoplasmic reticulum. J Biol Chem 1998;273:15030–4.

30] Romano JD, Schmidt WK, Michaelis S. The Saccharomyces cerevisiaeprenylcysteine carboxyl methyltransferase Ste14p is in the endoplasmicreticulum membrane. Mol Biol Cell 1998;9:2231–47.

31] Schmidt WK, Tam A, Fujimura-Kamada K, Michaelis S. Endoplasmicreticulum membrane localization of Rce1p and Ste24p, yeast proteasesinvolved in carboxyl-terminal CAAX protein processing and amino-terminal a-factor cleavage. Proc Natl Acad Sci USA 1998;95:11175–80.

32] Perez-Sala D, Gilbert BA, Tan EW, Rando RR. Prenylated proteinmethyltransferases do not distinguish between farnesylated and geranyl-geranylated substrates. Biochem J 1992;284:835–40.

33] Bergo MO, Leung GK, Ambroziak P, et al. Isoprenylcysteine carboxylmethyltransferase deficiency in mice. J Biol Chem 2001;276:5841–5.

34] Bergo MO, Leung GK, Ambroziak P, Otto JC, Casey PJ, Young SG.Targeted inactivation of the isoprenylcysteine carboxyl methyltransferasegene causes mislocalization of K-Ras in mammalian cells. J Biol Chem2000;275:17605–10.

35] Fukada Y, Matsuda T, Kokame K, et al. Effects of carboxyl methylationof photoreceptor G protein gamma-subunit in visual transduction. J BiolChem 1994;269:5163–70.

36] Parish CA, Smrcka AV, Rando RR. Functional significance of beta gamma-subunit carboxymethylation for the activation of phospholipase C andphosphoinositide 3-kinase. Biochemistry 1995;34:7722–7.

37] Bergo MO, Gavino BJ, Hong C, et al. Inactivation of Icmt inhibits trans-formation by oncogenic K-Ras and B-Raf. J Clin Invest 2004;113:539–50.

38] Backlund Jr PS. Post-translational processing of RhoA. Carboxyl methyla-tion of the carboxyl-terminal prenylcysteine increases the half-life of Rhoa.J Biol Chem 1997;272:33175–80.

39] Buckner FS, Yokoyama K, Nguyen LN, et al. Cloning, heterologous expres-

sion, and distinct substrate specificity of protein farnesyltransferase fromTrypanosoma brucei. J Biol Chem 2000;29:21870–6.

40] Buckner FS, Kateete DP, Lubega GW, Van Voorhis WC, Yokoyama K.Trypanosoma brucei prenylated-protein carboxyl methyltransferase prefersfarnesylated substrates. Biochem J 2002;367:809–16.

[

cal Parasitology 153 (2007) 115–124

41] Buckner FS, Eastman RT, Yokoyama K, Gelb MH, Van Voorhis WC. Pro-tein farnesyl transferase inhibitors for the treatment of malaria and Africantrypanosomiasis. Curr Opin Investig Drugs 2005;6:791–7.

42] Gelb MH, Van Voorhis WC, Buckner FS, et al. Protein farnesyl and N-myristoyl transferases: piggy-back medicinal chemistry targets for thedevelopment of antitrypanosomatid and antimalarial therapeutics. MolBiochem Parasitol 2003;126:155–63.

43] Nallan L, Bauer KD, Bendale P, et al. Protein farnesyltransferase inhibitorsexhibit potent antimalarial activity. J Med Chem 2005;48:3704–13.

44] Yokoyama K, Trobridge P, Buckner FS, Van Voorhis WC, Stuart KD, GelbMH. Protein farnesyltransferase from Trypanosoma brucei. A heterodimerof 61- and 65-kDa subunits as a new target for antiparasite therapeutics. JBiol Chem 1998;273:26497–505.

45] Buckner FS, Eastman R, Speelmon E, et al. Cloning, heterologousexpression, and substrate specificity of protein farnesyltransferases fromTrypanosoma cruzi and Leishmania major. Mol Biochem Parasitol2002;122:181–8.

46] Weston D, Patel B, Van Voorhis WC. Virulence in Trypanosoma cruziinfection correlates with the expression of a distinct family of sialidasesuperfamily genes. Mol Biochem Parasitol 1999;98:105–16.

47] Stephenson RC, Clarke S. Identification of a C-terminal protein carboxylmethyltransferase in rat liver membranes utilizing a synthetic farnesylcysteine-containing peptide substrate. J Biol Chem 1990;265:16248–54.

48] Redmond S, Vadivelu J, Field MC. RNAit: an automated web-based toolfor the selection of RNAi targets in Trypanosoma brucei. Mol BiochemParasitol 2003;128:115–8.

49] Alibu VP, Storm L, Haile S, Clayton C, Horn D. A doubly inducible systemfor RNA interference and rapid RNAi plasmid construction in Trypanosomabrucei. Mol Biochem Parasitol 2005;139:75–82.

50] Wirtz E, Leal S, Ochatt C, Cross GA. A tightly regulated inducibleexpression system for conditional gene knock-outs and dominant-negativegenetics in Trypanosoma brucei. Mol Biochem Parasitol 1999;99:89–101.

51] Hirumi H, Hirumi K. Continuous cultivation of Trypanosoma brucei bloodstream forms in a medium containing a low concentration of serum proteinwithout feeder cell layers. J Parasitol 1989;75:985–9.

52] Yokoyama K, Trobridge P, Buckner FS, et al. The effects of proteinfarnesyltransferase inhibitors on trypanosomatids: inhibition of proteinfarnesylation and cell growth. Mol Biochem Parasitol 1998;98:87–97.

53] McGeady P, Kuroda S, Shimizu K, Takai Y, Gelb MH. The farnesyl group ofH-Ras facilitates the activation of a soluble upstream activator of mitogen-activated protein kinase. J Biol Chem 1995;270:26347–51.

54] Dolence JM, Steward LE, Dolence EK, Wong DH, Poulter CD. Studieswith recombinant Saccharomyces cerevisiae CaaX prenyl protease Rce1p.Biochemistry 2000;39:4096–104.

55] Holstein SA, Wohlford-Lenane CL, Hohl RJ. Isoprenoids influence expres-sion of Ras and Ras-related proteins. Biochemistry 2002;41:13698–704.

56] Delarue FL, Adnane J, Joshi B, Blaskovich MA, Wang DA, Hawker J,Bizouarn F, Ohkanda J, Zhu K, Hamilton AD, Chellappan S, Sebti SM.Farnesyltransferase and geranylgeranyltransferase I inhibitors upregulate

RhoB expression by HDAC1 dissociation, HAT association and histoneacetylation of the RhoB promoter. Oncogene 2007;26:633–40.

57] Winter-Vann AM, Baron RA, Wong W, et al. A small-molecule inhibitorof isoprenylcysteine carboxyl methyltransferase with antitumor activity incancer cells. Proc Natl Acad Sci USA 2005;102:4336–41.