Vol. 3, 495-505, August 1992 Cell Growth & Differentiation 495

Identification and Characterization of the RegulatedPattern of Expression of a Novel Mouse Gene,megi, during the Meiotic Cell Cycle’

Jeremy Don and Debra J. Wolgemuth2Departments of Genetics and Development [J. 0., D. J. W.] andObstetrics and Gynecology [D. J. W.] and The center for ReproductiveSciences [J. D., D. I. W.], Columbia University college of Physiciansand Surgeons, New York, New York 10032

AbstractThe gene designated megl (meiosis expressed gene) is

a new mouse gene identified during a search formammalian genes potentially involved in meioticprocesses. Two classes of complementary DNAs wereisolated from an adult mouse testis complementaryDNA library, which shared the same 3’ end includingthe entire putative coding region but differed in their5’ ends. Only one of these complementary DNA classesappeared to correspond to the very abundant 0.75-kilobase testicular transcript of megl. Sequenceanalysis predicts a 10.8-kilodalton protein which ishighly charged and lysine rich. It is also relatively richin potential phosphoacceptor amino acids (‘�-1 7%),several of which are located in phosphorylationconsensus sequences. The paftern of expression ofmegl was studied utilizing a combined Northern blotand in situ hybridization analysis. Of the adult tissuesexamined, megl transcripts were detected exclusivelyin testis. Analysis of mRNA from testes of two germ celldeficient mutant strains did not reveal significant levelsof megl transcripts. Analysis of RNA from enrichedpopulations of spermatogenic cells from adult testesand localization by in situ hybridization revealed thatmegl transcripts are most abundant in pachytenespermatocytes. These results suggest a role for meglduring germ cell differentiation, possibly duringmeiotic prophase.

IntroductionThe existence of sexually reproducing species is depend-ent upon a process called meiosis, in which the numberof chromosomes in the gametes is reduced by half,without any loss of essential genetic information. At thechnomosomal level, cells undergoing meiosis perform avery defined and evolutionarily conserved sequence ofevents including chromosome duplication, pairing andrecombination of homologous chromosomes, and reduc-tional and equational divisions. Our understanding of the

Received 3/9/92.1 This work was supported in part by a Wolfson postdoctoral fellowshipfrom the Israeli Academy of Sciences and Humanities (J. D.), a grantfrom the Lucille Markey Trust (D. J. W.), and NIH Grant P50 HD05077(D. J. W.).2 To whom requests for reprints should be addressed, at Department ofGenetics and Development, columbia University College of Physiciansand Surgeons, 630 West 168th Street, New York, NY 10032.

meiotic process in general, and of that process in mam-mals in particular, is mainly at the descriptive level. Thesequence of events has been detailed with microscopicobservations, and classical genetic analysis had beenused to document the process of recombination. Little isknown about the regulation of the various stages ofmeiosis and of the factors that participate in thisregulation.

Insight into some regulatory mechanisms, operating inboth the mitotic and the meiotic cell cycles, has beenobtained from the recent identification of key cell cycleregulatory genes such as the cdc2 gene in yeast and itshomologues in higher organisms, and the various mem-bers of the cyclin family (1, 2). With respect to theregulation of earlier meiotic stages, such as the transitionfrom mitotically dividing cells to entry into meiosis andprogression through the first meiotic prophase, the Iarg-est body of information has accumulated from studieson the yeast Saccharomyces cerevisiae (3-1 5).

In higher organisms, and mammals in particular, muchless is known. A good model system for studying mam-malian meiosis is mouse spermatogenesis. A detailed,morphologically characterized developmental scheduleof the germ cell differentiation has been established (16,17), mutant strains of mice exist in which the cells arearrested at specific stages of spermatogenesis (18-20),and enriched populations of spermatogenic cells at dif-ferent stages of differentiation are accessible via cellseparation techniques (21). Many mouse genes whichdisplay a unique pattern of expression during spermato-genesis have been described (22). It has been shown thatan intensive transcriptional activity takes place in meioticspenmatocytes, especially in mid and late pachytenestages, but a considerable proportion of the RNAs (andproteins) synthesized during this stage are stored in thecell for subsequent use (23-25). Several genes exhibitexpression only in postmeiotic stages, such as the pro-tamine 1 and 2 and the transition protein (TP) genes (26),and as such are more likely to be associated with thedramatic changes that occur during spermiogenesisrather than in meiotic events. To date, none of the geneshas been correlated with a specific meiotic event.

In this paper, we describe the identification and char-actenize the pattern of expression of a new mouse gene,meg!,3 that was identified during a search for mammaliangenes potentially involved in meiotic processes.

ResultsIdentification of megl. Civen the conservation of boththe meiotic process itself and of several key moleculesthus fan identified, such as p34�2 and cyclins, we hy-

3 The sequence of megl was submitted to the EMBL database and wasassigned the accession number X64455.

496 Novel MouseMeiosis Expressed Gene

pothesized that other genes important in regulating entryinto meiosis in yeast might have evolutionanily conservedcounterparts in mammals. The initial goal at the outsetof this study was to search for munine genes structurallyhomologous to the yeast meiosis regulating gene, RME1(7, 8). Since the deduced amino acid composition of theRME1 protein contains three zinc fingers in its carboxyterminus (7), an RME1 clone (Sp�-l 17) was hybridized,under low stringency conditions, to a Southern blot ofeight uncharactenized putative zinc finger containing tes-ticular cDNA4 clones.5 One of these clones, 3.2 kb insize, gave a positive signal (data not shown). This clonecontained an internal EcoRl site, yielding two fragmentsof “-‘1.5 kb and 1.7 kb [designated 1.5 and 1.7, respec-tively (Fig. la)]. To determine whether this positive signalwas due to sequence similarities to RME1 other thanwithin the conserved zinc finger regions, a Southern blotofthe 1 .5 and 1 .7 fragments was hybridized with differentregions of the RME1 clone, Spi�-i 17. Both the 1 .5 and1 .7 fragments hybridized with the complete Spa-i 17probe at relatively high stringency conditions [final washat o.5x SSC-0.l% SDS at 42#{176}C(Fig. lb)]. When a 350-bp Bg!ll-EcoRl fragment of Sp�-1 17, which does notcontain the zinc finger domain, was used as a probe,only the 1.5-kb band could be detected (Fig. 1 b). Con-versely, only the 1 .7-kb band was detected by the 750-bp EcoRl-EcoRl fragment of Spz�-1 1 7, which contains thezinc finger domain (Fig. 1 b). These results suggested thatthe 1.7 fragment contained the zinc finger domain andthat the hybridization by Sp�-l 1 7 with the 1.5-kb frag-ment reflected homology to RME1 outside the conservedzinc finger domain. However, as discussed below, onlya portion of this prediction was borne out.

A preliminary characterization of the transcripts con-responding to the cDNAs was performed by Northernblot hybridization analysis of total testis RNA, with the1 .5 and 1.7 inserts as probes. The 1 .5 probe detected avery abundant transcript of 0.75 kb and another relativelyweakly hybridizing transcript of 3.2 kb (Fig. ic). The 1.7probe detected only the 3.2-kb transcript (Fig. ic). Tofurther characterize these clones and their relationshipto one another, direct DNA sequence analysis was un-dertaken, initially of the 1 .5-kb fragment. This analysisrevealed that the 1 .5-kb fragment contained a stretch of(C)15A(C)9 0.55 kb from the 3’ end (Fig. la). A 264-bp-long ORF, initiating with an ATC within a canonicaltranslation initiation consensus sequence (27, 28), wasobserved in this 0.55-kb fragment beginning 102 bp 3’to the CC region on one strand. However, a secondundisrupted ORF (at least 700 bp long), again initiatingwith an ATC, was located 220 bp 5’ to the CC region onthe complementary strand (Fig. la). These results sug-gested the possibility that the 1.5 fragment represented,in fact, two independent cDNAs that were artifactuallylinked at their ends during the generation of the testiscDNA library.

4The abbreviations used are: cDNA, complementary DNA; bp, basepair(s); cAMP-PK, cyclic AMP dependent protein kinase; cGMP-PK, cyclicGMP dependent protein kinase; PKC, protein kinase C; CaM kinase II,calcium/calmodulin protein kinase II; kb, kilobase(s); ORF, open readingframe; PBS, phosphate buffered seline; poly(A)�, polyadenylated; SDS,sodium dodecyl sulfate; SSC, standard saline citrate.S p, 5, Burke and D. J. Wolgemuth, in press.

To further examine this possibility, the 1.5 fragmentwas digested with BamHI to generate two fragments ofabout 0.75 kb each, which were used as probes forNorthern blot hybridization of total testis RNA. The 3’0.75-kb fragment, which contains the CC region, wasdesignated 0.71, and the 5’ fragment was designated 0.711(Fig. la). Probe 0.71 detected only the abundant 0.75-kbtranscript, whereas the 0.711 probe detected only the 3.2-kb transcript (Fig. 1d). These results further supportedthe notion that the 1 .5 fragment represented cDNAs oftwo independent genes. Further sequence and expres-sion analysis has revealed that the 5’ portion of the 1.5fragment (0.711) contains a portion of a cDNA that con-responds to a 3.2-kb transcript that contains a zinc fingerdomain. These studies will be presented in a subsequentreport. The focus of this study is the characterization ofthe gene, represented by the 3’ fragment (referred to asclone 0.71), that gives rise to the very abundant 0.75-kbtranscript in testis. For reasons that will be more apparentfrom the results presented below, this gene has beendesignated meg! (meiosis expressed gene).

To determine whether megl represents a single geneor is a member of a family of several related genes,Southern blot hybridization analysis of EcoRl digestedmouse genomic DNA was performed. Cenomic DNAfrom male and female mice from two strains was exam-ined, using clone 0.71 as a probe. Two bands of ‘-9.5 kband 4.4 kb were observed, independent of the sex orstrain of the source of DNA (data not shown). We con-dude that megl is not linked to the Y chromosome andis not likely to be a member of a multigene family. Inaddition, clone 0.71 was hybridized under low stringencyconditions to a blot that consisted of EcoRl digestedgenomic DNA from human, rat, rattlesnake, chicken, frog(Xenopus), fish (Oryzias !atipe), sea urchin, Drosophila,and yeast (S. cerevisiae), to obtain preliminary informa-tion regarding its evolutionary conservation. Weakly hy-bnidizing bands of about 7.0 kb and 4.0 kb and of 7.0 kband 2.3 kb could be detected in human and rat, respec-tively (data not shown). No such apparent bands weredetected under these conditions in the other organismstested (data not shown). These results, although suggest-ing the existence of a closely related gene in othermammals, do not rule out the possibility that geneshomologous to meg! exist in lower organisms as well.

Characterization of the megl cDNA(s). To confirmthat clone 0.71 corresponds to a bona fide cDNA repre-sentative of the meg! transcript, it was used as a probeto screen a second, independently generated adultmouse testis cDNA library. Four clones were isolated anddesignated 1a2, 2a2, 1 1a2, and 2c2 (Fig. 2a). DNA Se-quence analysis revealed that the longest clone (1 1a2,684 bp) shared 100% identity over 447 bp with the 3’end ofclone 0.71 (not including the poly(A) tail), includingall of the ORF. However, the sequence diverged in the219 bp corresponding to the 5’ region upstream of theputative ATG. Clone 2c2 (372 bp long) was similar to11a2, including its 5’ end. Clone 2a2 (532 bp long) wassimilar to 0.71, including its entire 5’ end, except for 19bp flanking the CC region. Finally, clone 1a2 (297 bplong) was similar to the 3’ end of clones 0.71, 2a2, and1 1a2 (Fig. 2a). These five independent clones thereforeshare the same 3’ end, including the whole ORF, buthave two different 5’ ends (Fig. 2a). The possibility ofcloning artifacts accounting for these differences was

a

EcoRl

bC d

Cell Growth & Differentiation 497

___________ 1.7 -1.5�-- I 10.711 I 0.71

1\\��,/ EcoRI�xccOkF

EcoRI BamHI �

sp�-1 17

B/X

�1.7 1.5

B/I

1.7

E

1.5

E/ES .

1.7 1.5

1.7kb

1.5 kb -

1.5 1.7

- 3.2 kb -

-0.75 kb -

0.71 0.711

Fig. 1. a, schematic representation of the 3.2-kb putative zinc finger containing clone that hybridized with the yeast gene RME1 (clone SP.�.117). An

internal EcoRl site divides the clone into two fragments designated 1.7 and 1.5. A BamHI site divides the 1.5 clone into two fragments designated 0.71

and 0.711. A GC stretch located in the 5’ portion of 0.71 is indicated, as well as two ORFs located on complementary strands and oriented opposite toone another. b, Southern blot hybridization of clones 1.7 and 1.5 with three different restriction fragments ofthe RME1 clone, SP�-117. A BglII/Xhot (B/X) fragment corresponds to the entire SPA-i 1 7 clone. The BgllI/EcoRI (B/E) and EcoRl/EcoRl (E/E) fragments correspond to the non-zinc finger containingdomain and to the zinc finger containing domain of RME1, respectively. Exposure time, 24 h. c, Northern blot hybridization of mouse total testis RNA (20;�g) to clones 1.5 and 1.7 as probes (exposure time, 16 h). d, Northern blot hybridization of mouse total testis RNA (20 �g) to clones 0.71 and 0.711 as

probes (exposure time, 16 h).

considered unlikely since it was improbable that totallyindependent clones from different cDNA libraries wouldshow the same artifact.

To address the question of whether both of theseclasses of cDNAs corresponded to the previously de-tected 0.75-kb transcript (Fig. ld), the different 5’ endsas well as the common 3’ end of the cDNA clones wereused as probes to hybridize to blots of total testis RNA.The prediction was that the 5’ end of the correspondingcDNA would detect the 0.75-kb transcript. An EcoRI-SacIfragment from clone 1 1a2 (209 bp) was used as a proberepresenting the entire 5’ end of this clone (Figs. 2a and3a), whereas a BamHI-Nael fragment from clone 0.71(231 bp) was used as a probe representing the second 5’end (Figs. 2a and 3b). The 473-bp SacI-XhoI fragment ofclone 1 1a2 was used as a probe for the common 3’ end(Fig. 2a). As can be seen in Fig. 2b, both the common 3’probe and the unique 5’ probe from clone 0.71 detectedthe 0.75-kb transcript. This transcript was not detectedby the unique 5’ probe of clone 1 1a2, even after longerexposure and reduced washing stringency conditions.However, a very faint band of about 3.0 kb was observedafter exposing the film for 3 days. When the entire clone1 1a2 was used as a probe, a similar 3.0-kb band as wellas faint 1 .87-kb, 4.7-kb, and 6.5-kb bands were detectedafter prolonged exposure in total testis RNA, in additionto the broad 0.75-kb band Fig. 2c). None of the bands

other then the 0.75-kb transcript was observed in testispoly(A)�’ RNA (Fig. 2c). The 1.87-kb and 4.7-kb bands inFig. 2c might be attributable to nonspecific probe bindingto the nibosomal bands. The 3.0-kb as well as the 6.4-kbtranscripts could represent an unprocessed precursor ofmeg! mRNA since they were detected by both cDNAs(2a2 and 1 1a2) in total RNA after prolonged exposure(data not shown) but not in poIy(A)� RNA.

Sequence analysis of clone 1 1a2 provided additionaldata, suggesting that it represents a nontranslated RNA.First, clone 1 1a2 has in its 5’ end an ORF of 1 14 bp,initiating with an ATG at position 101 (Fig. 3a). Suchupstream ORFs are believed to reduce significantly theefficiency of translation of the main initiation site (27,29). Secondly, 1 1a2 contains two complementarystretches of 9 bp each, upstream of the ATG yielding thelongest ORF of the cDNA [positions 1 19-127 (Fig. 3a)and 85-93 (Fig. 3b)}, which could possibly form second-any structures. If such structures do occur, an inhibitionof translation of the main ORF could result, especiallysince one ofthese stretches is located only 5 bp upstreamof the major ATG and includes pant of the translationinitiation consensus sequence which is believed to bepart of the nibosomal recognition site (28). Therefore, thestructural and expression data together suggest that clone2a2, rather then 1 1a2, represents the cDNA correspond-ing to the translatable mRNA of this gene.

a

GGGCC AT1-�’

CIoneO.71 \�7

Clone 2a2

Clone 1a2

F�oE

FLORI

Xh�

EcoRl

Ecoll Xhol

L

X�ol

s�I

EcoRl Xbol

CIoneZc2 I r

sat,

b C

TT d7� TT�

3.0kb -

0.75 kb - #{149}#{149}

3.0 kb -

0.75 kb -

498 Novel Mouse Meiosis Expressed Gene

Clone 11a2

B.mHI N�I

11a2 0.71 11a2

SIX B/N

The 3’ untranslated region common to all ofthe cloneshas two unique characteristics: it is relatively AT rich, and

it lacks the consensus AATAAA polyadenylation signal.Instead, a closely related sequence, CATAAA, which hasbeen observed as a polyadenylation signal in 1% ofpolyadenylated RNAs of vertebrates (30), is located 20bp upstream the poly(A) tail and substitutes for the

canonical sequence.A search through the GeneBank/EMBL database did

not reveal any DNA sequences with homology to meg!.A search through the SwissProt database for any relatedknown proteins revealed no similarity between the pu-

tative megl protein and the proteins in the database.The database was also researched for similarity betweenthe 38-amino acid potential product of the short ORF inthe 5’ end of clone 1 1a2, but no homologous proteinswere identified. Thus, meg! is likely to be a new genewhich is not significantly related to any of the knowngenes represented in the databases searched.

Sequence Characterization of the Putative megi Pro-tein. The meg! cDNA has a 264-bp ORF that putativelyencodes a basic protein of 88 amino acids with an esti-mated p1 of 9.16 and a calculated molecular weight of10,800. The deduced amino acid composition (Fig. 3b)revealed several interesting characteristics. First, meglprotein is highly charged, with 39.7% charged aminoacids, including 13.64% lysine, 7.9% anginine, 1.1% his-

Fig. 2. a, five independent cDNAclones obtained from two independ-ently generated adult testis cDNA librar-es (see text), sharing the same 3’ end,

includingthe whole ORF (thick bars), butwith two different 5’ untranslated ends(thinner and thinnest bars). Clones 0.71

and 2a2 represent one 5’ end (thinner

bar), and clones 1 1a2 and 2c2 representthe second 5’ end (thinnest bar). TheNaeI and Sad restriction sites character-istic to the two 5’ ends are indicated. b,

Northern blot hybridization oftotal testisRNA (20 �zg) to the two different 5’ ends,and to the common 3’ end. Sacl/Xhol (5/X) fragment, from clone 1 1a2, representsthe common 3’ end of all clones (expo-sure time, 16 h), BamHI/Nael (B/N) frag-ment, from clone 0.71 (exposure time,16 h), and the EcoRI/SacI (E/S) fragments,from clone 1 1a2 (exposure time, 72 h),represent the two different 5’ ends. c,Northern blot hybridization oftotal testisRNA (TT; 20 gg), and of poly(AY� RNAfrom day 7 postnatal testis (d7+; 4 ��g)and adult testis (TT+; 4 gg), to clone1 1a2 as a probe (exposure time, 48 h).

tidine, 1 1 .4% glutamic acid, and 5.7% aspartic acid. Thelysine residues are predominantly distributed within theprotein in pains (amino acids 7,9: 15,16; 57,58; 82,84). Inaddition, there are pains of arginines (amino acids 61,62)and lysine/anginine (amino acids 71,73). Secondly, thepredicted megl protein contains a high proportion ofpotential phosphoaccepton amino acids (>1 7% of thetotal amino acid complement), including 8 tyrosine nesi-dues (representing >9% of the total amino acids), 4senines (all of which are within the first 18 amino acids atthe amino terminus), and 3 threonine residues. Third,meg! contains 9 valine residues, 4 ofwhich are clusteredbetween amino acids 41-56, with a second cluster of 4residues at the canboxy terminus between amino acids80-87. In a search for functional consensus sequences,no calcium binding sites (31), ATP binding sites (32), zincbinding sites (33), on glycosylation sites (34) could bedetected. However, as discussed later, several consensussequences for phosphorylation of proteins (35) wereobserved.

Tissue Specificity of megl Expression. To determinethe tissue distribution of meg! expression in the mouse,total RNA was isolated from 10 adult tissues, includingtestis, epididymis, ovary, spleen, lung, liver, kidney,heart, brain, intestine, and from placenta (from day 12conceptuses), and analyzed by Northern blot hybnidiza-tion analysis, using clone 2a2 as probe. The very abun-

b �“#{176}‘ N.�i

3 CAA1’TCGGCACGAGCCAGTAGAGAACACGAACACCGCTGCGCCCTCCTGCGAGGCCGGCT �#{216}

e � A T S D V K P 801 GAACc�G�AAATAATTGC1AATAACGCCTCTCTAGCCA1GCCTACTTCTGACCTGAAACC 120

9 K S I S P A K K W S E E I K N L V P F q 28

12 1 AAAATCAATAAGTCCTCCCAAGAAATGCTCAGACCAAATAGAAAATCTCI’ACAGATTTCA 180

29 6 A G V P D 6 I E S K Q V K Q V A � V 0 481 8 1 ACAAGCAGGATATC000AIGAAAT TGAATATAAACAAGTGAAACAAGTTGCCATGGTCGA 240

49 P 8 P E I 0 V V K K L Q P P 0 N I F F V 08

241 CCCATGGCCACACACAGCCTACGTCAAGAAACTTCAGCGGA000ACAATACTTTCTTCTA 300

09 5 N K E R E C K 0 K E V H K V K V V V V 88

301 C IACAACAAAGACACCGAGTGCGAGGACAAGGACC1’CCACAAAGTGAAGCTTTACCICTA 300

89

301 CTGACCTTTTCCTTTCTTCGGCTTGGCAATGCTCCTITAAGAATTGGTTGTITACAITCT 420

421 TCCATCGTGTAAATCTCATTTTACAAAACAATTCACAA1’TCTGTCTTTAATTCATGGTGT 480

481 CTTACACAACATAAACACCCACCT1GAAACCTTTAAA... 617

Fig. 3. a, nucleotide sequencesofthe 5’ end ofclone 11a2; b, nucleotide

sequences of the entire clone 2a2. The deduced amino acid sequencesof the short ORF in the 5’ end of 1 1a2 and of the longest ORF in clone2a2 are shown. The EcoRl, NaeI, and Sad restriction sites are indicated.The AGG sequence in clone 2a2 at the junction point between clone2a2 and 1 1a2 is boxed, with an arrow indicating the exact junction point(the next nucleotide of clone 1 1a2 would be the G on the right side ofthe arrow). Complementary stretches of 9 bp each within clone 1 1a2 aremarked with dashed lines. The putative CATAAA polyadenylation signalis underlined.

285 - � -

185- � 4

rT,V

� VrSe � (.n�) �.(n�= � V fl

0�P -� (D � 1- (D <� �) � < C (D ��-...-, mm � (D r� �3S � ,., ..< -, � � �) “C U)

..-. .1 �‘ I I I t I I I I

0.75 kb -

2.3 kb - -----��- -:=: ‘.�

Cell Growth & Differentiation 499

a

I CCAATTCCGCACCA000AACTAGAGGCAACTCTTCTCCATAIACACACCAACTTATACAT a,

hi V F P P L P 7

61 TCCACACAGTAGIACCCGTAAAGCCTTGGTITTCTATAAGAIGGTCITTCCIAGA(IACC 120

a E I I A V 0 0 L S L 5 D N V F C 7 P S C 27121 CGAGACAATTCCTCTCCAT000TTCTCTCTACACCACAATACAT1CTCCACTGAGTC ‘CC 180

20 P P 5 N F S E V C A P181 TCGTCGACAGAATTT( TCAGAAGTTGGAGCTC(TIGAGA 219

S..’

dant 0.75-kb transcript was detected exclusively in testisRNA (Fig. 4). No signal was detected in any other tissue,even after longer exposures (data not shown). The filterwas washed and neprobed with fl-actin to confirm RNAintegrity (Fig. 4). Since liver RNA showed a very faintsignal with fl-actin, this blot was neprobed once againwith glycenaldehyde-3-phosphate dehydnogenase as aprobe. Liven RNA integrity was verified by the appear-ance of the expected 1 .4-kb band (data not shown). Inaddition, no expression of meg! transcripts was detectedin total RNA from embryos day 12.5 and day 15.5postcoitum (data not shown) or in poly(A)� RNA fromtestes of day 7 postnatal animals (Fig. 2c). Thus, amongthe tissues examined, the meg! transcript was expressedexclusively in the samples of mature testis.

Cellular Localization of megl Expression in the Testis.To determine the cellular specificity of meg! expressionin the testis, total RNA was isolated from two mutantstrains of mice: white spotted (W) and atrichosis (at).Gonads of mice homozygous for the mutant alleles (W/�,N on at/at) are virtually devoid of germ cells, whereasthe somatic complement, including the interstitial Leydigcells and the Sertoli cells in the seminiferous tubules,appears normal (20, 36-38). Hetenozygous animals (+/w,+/wv, on +/at) are phenotypically indistinguishable fromwild-type animals. Testis RNA from homozygous andhetenozygous mice as well as from normal Swiss Websteradult mice was analyzed by Northern blot analysis, usingclone 2a2 as a probe. The 0.75-kb transcript was readilyobserved in RNA from testes of the normal and heteno-zygous mice but not in the RNA from homozygous

1.65kb -

Fig. 4. Northern blot hybridization analysis of total RNA (20 ag/sample)from 1 1 adult tissues, using clone 2a2 as a probe (exposure time, 16 h).The ethidium bromide staining of the gel suggests that samples wererelatively equally loaded. RNA integrity was verified by washing the filterand reprobing it with $-actin as a probe (bottom panel; exposure time,24 h).

animals. The 0.75-kb transcript was therefore most abun-dant in germ cells. The trace levels of 0.75-kb transcriptsdetected in the mutant testes after prolonged exposure(Fig. 5a) may reflect transcripts from the few partiallydifferentiated germ cells that can occasionally be foundin the seminiferous tubules of mutant mice (37) on mayresult from very low levels of transcripts in somatic cells.

To determine which spermatogenic cells express the0.75-kb transcript, enriched populations of meiotic pro-phase spermatocytes (predominantly in the pachytenestage of meiosis), postmeiotic early spermatids, and amixture of residual bodies and cytoplasmic fragmentsfrom elongating spenmatids were obtained (21). TotalRNA was isolated from the various cell types and ana-

lyzed by Northern blot hybridization with the 2a2 cloneas a probe (Fig. 5b). The 0.75-kb transcript was veryabundant in RNA from the meiotic prophase cells. Den-sitometnic analysis revealed that the levels of the 0.75-kbtranscript decreased slightly (--20%) in the early sper-matid fraction. The presence of meg! transcripts in theresidual bodies fraction suggests that this mRNA is notdegraded rapidly, since transcriptional activity ceases

a b

0., �

- 285

- 185

Fig. 5. Northern blot hybridization analysis of total RNA from: (a) testesof two mutant strains of mice, atrichosis (at), and white spotted (w). +/atand +/w are samples from phenotypically normal heterozygous animals,whereas at/at and w/w are samples from germ cell deficient homozygousanimals (30 pg/sample). Total testis RNA from normal adult mice (iT; 20

SAg)was used as a control (exposure time, 48 h); (b) enriched populationsof meiotic prophase spermatocytes (mainly pachytene, P), postmeioticearly spermatids (ES), and a mixture of residual bodies and cytoplasmicfragments (RB; 20 ,�g/sampIe). Total testis RNA from adult mice (iT; 18�Lg) was used as a control (exposure time, 8 h). The ethidium bromidestaining of the gels suggests that samples were relatively equally loaded,unless otherwise stated. In all blots, 2a2 was used as a probe, and either$-actin or glyceraldehyde-3-phosphate dehydrogenase was used as aprobe to verify RNA integrity (data not shown).

500 Novel Mouse Meiosis Expressed Gene

185 $0 $1 N � *4

� �; � 3

� �;T�k .� .1 ‘� I � � I ..�. ES � Pb

�15 kb � iLl 1Pl�’��� #{149}-C 75 kb

prior to stage 9 of spenmiogenesis (39). This pattern ofdecreasing abundance of the 0.75-kb transcript frommeiotic prophase through early spenmatids to elongatingspenmatids suggests that meg! is predominantly tran-scnibed in cells at meiotic prophase and is slowly turnedoven once the meiotic divisions are completed. However,the possibility that this gene is actively transcribed inhaploid early spenmatids as well cannot be ruled out.

In Situ Hybridization Analysis. To further define thecellular expression of meg!, we performed a series of insitu hybridization experiments on testis sections fromnormal animals and atrichosis mutants (homozygous aswell as heterozygous). Antisense and sense RNA probesdnived from clone 2a2 were used for experimental andcontrol hybridizations, respectively. An intense hybnidi-zation signal was observed in all tubules from normaladult testes, independent of the stage of their seminif-erous epithelium cycle [as described by Oakbeng (40);Fig. 6, a and b]. Examination of the cellular distributionrevealed that this signal was exclusive to the meioticspermatocytes and early spenmatids (Fig. 6, c and d). Nosignal above background was noted in the spenmatogon-al cells in the periphery of the tubules, non in the

elongated spermatids which are located just adjacent tothe lumen (Fig. 6, a and b; c and d). No signal abovebackground was detected in the interstitial regions of theadult testis (Fig. 6, a and b), nor when sense orientedprobes were used (data not shown). A similar pattern ofmeg! localization was observed in hetenozygous mutants(+/at) testes (data not shown). In sections from homo-

zygous mutant (at/at) testes, which are deficient in germcells, no signal above background could be detected (Fig.6, e and 1), in agreement with the Northern blot hybrid-ization analysis (Fig. 5a).

Discussion

In this study, we describe the identification of a novelmouse gene, meg!, which exhibits a unique pattern ofexpression during germ cell differentiation. A very abun-dant meg! transcript of 0.75 kb was detected in totaltestis RNA from adult mice, but not in poIy(A)� RNA fromtestes of day 7 postnatal mice. Transcripts of meg! werenot observed in RNA from 10 other adult tissues on intotal RNA from midgestation embryos. Moreover, thistranscript was not detected in testes of two germ celldeficient mutant strains, except for trace amounts thatwere detected by Northern blot analysis under condi-tions of increased amounts of RNA and prolonged ex-posune. These results indicate that meg! not only isexpressed in a testis specific pattern but is also primarilygerm cell specific as well. Analysis of RNA from theenriched populations of spermatogenic cells revealedthat meg! was most abundant in the meiotic cells, apattern that was confirmed by in situ analysis.

The meg! mRNA levels appeared to decrease in thehaploid cells. This expression pattern of meg! is distinctfrom that observed for other genes also expressed inspenmatocytes, such as phosphoglycenate kinase 2 (41)or some acrosomal specific proteins (25), whose mRNAsappear to be stoned for postmeiotic translation and func-tion. The expression pattern of meg! would be morecharacteristic of genes associated with meiotic events.Moreover, analysis of the temporal synthesis pattern ofproteins constituting the spermatozoon (25) did not re-veal labeled proteins of a size corresponding to that ofthe predicted megl protein. The simplest interpretationwould be that the megi protein is not present inspermatozoa.

Although speculative, comparison of megi with othergerm cell specific, size related proteins, with severalcommon structural features, might give a clue as to itspossible function. The predicted megi protein is a small,basic (lysine rich) protein, which is also rich in potentialphosphoaccepton amino acids (‘-‘17%), mainly tyrosine(9%). It is reminiscent of the transition proteins TP1 andTP2, which are germ cell specific, small (54 and 1 17amino acids, respectively), basic (lysine and angininerich), DNA associated proteins that are believed to func-tion in DNA condensation (26). TP1 and TP2 containmany phosphoacceptor residues, mainly senine and thre-onine (42, 43). The tynosine residues present within thetransition proteins have been shown to intencalate be-tween the DNA bases, possibly inducing conformationalchanges as well as nicks in native DNA (44). Such inter-actions might be of significance during meiosis as well asduring the chromatin remodeling of postmeiotic spen-matids. The pattern of expression of meg! versus theexclusively haploid expression of TP1 and TP2 (26, 45)might suggest roles during meiotic versus postmeioticevents, respectively. Of other size related, testis specificgenes that share some similarities with meg!, the rat RT7

(46) deduced product is 90 amino acids long and is senineand threonine rich (10% each). However, this gene is

,�, �-

Fig. 6. In situ hybridizationanalysis of testis sections from:(a) normal adult animals, brightfield and (b) epiluminescenceoptics, exposure time, 6 days;

(c) normal testis, at higher mag-nification, bright field, and (d)epiluminescence; exposuretime, 6 days; (e) at/at germ celldeficient mutants, bright field,and (1) epiluminescence; 12

days. All sections were hybrid-

ized with an antisense RNAprobe drived from clone 2a2.Bars, 50 �m.

,�.. � IL.1� �

5’

Cell Growth & Differentiation 501

1�1.s.�”, �

p.

�‘ ,i.

also a haploid expressed gene, and its function is notknown.

As mentioned above, p1 1megl is relatively rich in po-tential phosphoacceptor amino acids (‘-l 7%). We ex-amined meg! for consensus phosphorylation sequencesmost frequently recognized by various senine/threonineprotein kinases (35). Four protein kinases were noted tohave the potential to phosphorylate megi protein (Table

1). Furthermore, all 4 senines in the megi protein, as wellas thneonine at amino acid 65, lie within such consensussequences (Table 1). This would argue in favor of megibeing phosphonylated, especially at residues S� and/orT65, which could potentially be phosphonylated by threedifferent kinases: cAMP-PK, cGMP-PK, and PKC for S,8and cAMP-PK, PKC, and CaM kinase II for T65 (Table 1).Interestingly, two isozymes of cAMP-PK have been re-

502 Novel Mouse Meiosis Expressed Gene

Table 1 Consensus sequences for serine and threonine phosphorylation sites located within the predicted megi protein

Modified from Kennelly and Krebs (35). Numbers to the left of the consensus sequences represent the relative frequency and efficiency with whichthe specific kinase phosphorylates at this sequence. X, any amino acid; #{176},the phosphoacceptor serine or threonine; /, interchangeable amino acids;?,relative frequency and efficiency not known.

Potential phosphoacceptor residues within megiKinase Consensus sequence

54 S,� 512 S,� T65

cAMP-PK 1 RR/KXS#{176}/T#{176}2 Rx25”/T” RDNT#{176}- RXS#{176}/T#{176}? (R/K)2xS#{176}/T” KKWS#{176}

cGMP-PK 1 (R/K)23x5”/T” KKWS#{176}

PKC 1 (R/K),3x�2S#{176}/-T#{176}X,2(R/K),3 KSIS#{176}RAK

2 S*fl*x�2(R/K)13 S#{176}DVK S#{176}ISR3 (R/K),3x�,2S#{176}/T” KS” KKWS” RRDNT#{176}

CaM kinase II 1 Rx2S”/T” RDNT”

ported in mouse spermatogenic cells: type I is present inpachytene spermatocytes and in round spenmatids,whereas type II is present in round and elongating sper-matids (47, 48). The relatively high percentage of tyrosineresidues in megi suggests that meg! might be phosphor-ylated on one or more tyrosine residues. Interestingly,the tyrosine positioned at amino acid 38 lies within anECF receptor tyrosine kinase consensus phosphorylationsite [XE/DY*X (49)]. None of the other tyrosine residuesin megi are located within sequences recognized byspecific tyrosine kinases, as listed by Pearson and Kemp(49). With respect to potential tyrosine kinases, the pu-tative tyrosine kinase ferT might be of special interest. Ithas been shown that ferT has a testis specific transcriptand that it is predominantly expressed in spenmatocytesat the mid and late pachytene stage (50, 51). The phos-phorylation site of ferT, however, has not been deter-mined. Other potential tynosine kinases shown to beexpressed during spermatogenesis include the pnotoon-cogenes c-ab! (52) and pim-1 (53).

An additional structural characteristic of the predictedmegi protein is that the basic residues, mainly lysine, arepredominantly distributed along the protein in pairs. Painsof basic amino acids have been shown to play an impon-tant role in the recognition of specific sequences byvarious serine/threonine protein kinases (35). Moreover,pairs of basic amino acids seem to be important indesignating specific domains of the protein. For example,pairs of lysine and anginine residues separate the com-ponent peptides of the bovine corticotropin-fl-lipotropinprecursor (54) and bracket all enkephalin sequences inmouse proenkephalin [opioid precursor (55)]. This sug-gests that creating small basic zones along the proteinmight be important in recognition sites for enzymes orother cellular components.

Two cDNA classes were isolated in this study. The twocDNA classes, represented by clones 2a2 and 1 1a2,shared the same 3’ end, including the entire putativeORF, but differed in their 5’ end. Clone 2a2 was consid-ered to represent the translatable 0.75-kb mRNA, sinceits 5’ end region detected the 0.75-kb mRNA, whereasthe 5’ end of clone 1 1a2 did not. The fact that clones2a2 and 1 1a2 represent different 5’ ends does not favorthe possibility that 1 1a2 is an unprocessed precursor of

meg!, in which case both 5’ ends were expected to beincluded in the precursor RNA. Although we cannotexclude the possibility that the 5’ end of clone 1 la2represents an intron and that this clone, as well as clone2c2, resulted from an incomplete reverse transcription ofa precursor RNA during the generation of the mousecDNA library, we suggest an alternative promoter modelas the most likely explanation for the alternative 5’ exons(56). Alternative promoters resulting in somatic specifictranscript and germ cell specific transcript have beenobserved in other genes, such as the proenkephalin gene(55) and the c-mos gene (57). An ACC sequence, whichwas found in clone 2a2 at a site corresponding to thejunction point between clone 2a2 and 1 1a2 [positions66-68 (Fig. 3b)}, is consistent with the sequence ex-pected at a joining point of two exons (58, 59). Thissuggests that 2a2 is a postsplicing product and thereforepotentially transcribed from a more upstream promoter.This is further supported by the presence of the sequenceCTCCTTCAGA at the most 3’ region of the 11a2 5’ end(Fig. 3a) that shows good homology to the consensusacceptor sequence (T/C)�(C/A)(C/T)AC (58, 59). A reser-vation to this model would be the extra nucleotide A atthe 3’ end of the proposed acceptor sequence [position219 (Fig. 3a)]. Although it is possible that this A wasadded after splicing via RNA editing processes (60), thispoint requires further investigation. Confirmation of themode of generation of the different RNA species can beachieved by analyzing the 5’ region of meg! in genomicclones, which is currently under investigation.

In conclusion, meg! is a novel mouse gene with aunique pattern of expression in male germ cells thatmakes it a candidate for being involved in meiotic proc-esses. It will be interesting to examine the expression ofmeg! during specific developmental stages of postnataltestes at both the RNA and protein levels as well as toexamine the comparable stages of meiosis of female germcells to determine the generality of its potential function.

Materials and Methods

Source of Tissues and Cells. Swiss Webster mice (CharlesRiver, Wilmington, DE), >60 days of age, were used as asource of normal adult mouse tissues for all experiments,except those using mouse mutant strains. For studies on

Cell Growth & Differentiation 503

germ cell deficient testes, hetenozygous and homozygouslittermates from two mouse mutant strains, white spotted(W) and atrichosis (at), were purchased from the JacksonLaboratory (Bar Harbor, ME). Enriched populations ofgerm cells in specific stages of spermatogenesis wereseparated by sedimentation at unit gravity as describedby Wolgemuth et a!. (21). Tissues for DNA or RNAisolation were frozen in liquid nitrogen immediately upondissection and stored at -70#{176}Cuntil used. All animalswere sacrificed by cervical dislocation.

Probes. The following probes were used: (a) Spi�-1 17,a 1 .3-kb Bg!ll-XhoI fragment of the yeast Saccharomycescerevisiae RME! gene (a gift from Dr. A. P. Mitchell,Columbia University, New York, NY), which consists of700 bp from the 3’ end of the coding region of the geneand an additional 600 bp from the 3’ untranslated endof the gene (7); (b) pAl, a 2.0-kb chicken fl-actin cDNA(61); (c) pRCAPDH-i, a 1.4-kb Pstl-PstI fragment fromthe rat glyceraldehyde-3-phosphate dehydrogenasegene (62) (obtained from Drs. R. K. Assoian and L. Scotto,Columbia University); and (d) clones that were isolatedduring this study and fragments of these clones were alsoused as probes as specified in the text. All DNA probeswere labeled with [a-32P]dCTP (DuPont, Wilmington, DE)by the random priming labeling technique, to a specificactivity of greater than iO� cpm/�zg, using the AmershamMultiprimer DNA labeling kit. Sense and antisense RNAprobes for in situ hybridization experiments were gen-erated from linearized clones constructed in pBluescnipt5K (Stratagene, La Jolla, CA), by using either T7 or T3RNA polymerase (Promega Biotech, Madison, WI). RNAprobes were labeled with [a-35S]UTP (DuPont), to aspecific activity of greater than i0� cpm/�g, and hydro-lyzed to an approximate size of 0.1 kb, according to Coxet a!. (63).

Southern Blot Hybridization Analysis. Mouse genomicDNA or cloned DNA was digested with appropriaterestriction enzymes, electrophoresed on 0.8% agarosegels, blotted onto nitrocellulose filters (Schleicher &Schuell, Keene, NH) according to standard procedures(64), and cross-linked to the filters by baking for 2 h at80#{176}Cin a vacuum oven. Prehybnidization (3-4 h) andhybridization (15-20 h) took place at 37#{176}Cor 42#{176}Cforlow or high stringency hybridization conditions, respec-tively, in 43% formamide, sx SSC, 1 x Denhardt’s, 50mM or 20 mM NaPO4 (for low or high stringencies, re-spectively), 0.1% SDS, and 100 �g/ml sonicated anddenatured salmon sperm DNA. Dextran sulfate wasadded to the hybridization solution to a final concentra-tion of 10%, as well as labeled probe to a final concen-tration of 1 -2 x 106 cpm/mI. For low or high stringencywashing conditions, final washes of filters, 20 mm each,consisted of: ix SSC-0.1% SDS at 37#{176}C,or 0.Sx SSC-0.1% SDS at 65#{176}C,respectively. Filters were then airdried and exposed to Kodak XAR film with an intensifyingscreen at -70#{176}Cfor the times indicated in the text.

RNA Isolation and Northern Blot Analysis. RNA wasisolated using the LiCI precipitation method describedby Cathala et a!. (65). Poly(A)� RNA was selected byoligo-dT cellulose chromatography according to Aviv andLeder (66). RNA samples were electrophoresed on de-naturing 1% agarose- 2.2 M formaldehyde gels, blottedonto Nytran membranes (Schleicher & Schuell) for 8-12h according to standard procedures (64), and UV cross-linked to the membranes (1 20 mi/cm2) with a Stratalinker

UV cross-linker (Stratagene). Pnehybnidization was pen-formed at 42#{176}Cfor 3-8 h in 50% formamide, 5x SSC,5X Denhardt’s solution, 50 mr�i NaPO4, 0.1% SDS, and300 �tg/mI sonicated and denatured salmon sperm DNA.Hybridization, at 42#{176}Cfor 15-20 h, was performed in50% formamide, sx SSC, 1x Denhardt’s solution, 20 m�iNaPO4, 10% dextran sulfate, and 100 pg/mI sonicatedand denatured salmon sperm DNA. Final washes, 20 mmeach, consisted of: o.ix SSC- 0.1% SDS at 65#{176}C,ando.ix SSC at 65#{176}C.Filters were exposed to autoradi-ographic film at -70#{176}Cfor the indicated times. Densi-tometry analysis of autoradiograms utilized the JAVAsoftware (Jandel Scientific, Corte Madera, CA), after non-malizing RNA levels from the ethidium bromide image.

Library Screening. Approximately 4 x 10� plaquesfrom an amplified adult mouse testis cDNA library, con-structed using the Stratagene ZAP-cDNA synthesis kit(provided by D. L. Chapman, Columbia University) werescreened with the clone 0.71 as a probe (see “Results”),according to standard procedures (64). The filters (dupli-cates for the first two plaque purification rounds) wereprewashed in 50 mM Tnis (pH 8.0), 1 M NaCI, 1 m�i EDTA,and 0.1% SDS at 42#{176}Cfor 1 h. Prehybnidization, hybrid-ization, and posthybnidization washes were identical tothose described for low stringency Southern hybnidiza-tions except that they were performed at 42#{176}C.Positiveclones were purified by three rounds of plaque selection.Inserts from the tertiary screen were excised in vivoaccording to the protocol of Stratagene. Resulting pha-gemids contained the cDNA of interest at the EcoRl andXhol site of pBluescnipt SK(-).

DNA Sequencing. The chain termination method (67)was utilized for sequencing. For each clone, both strandswere fully sequenced using the T3 and T7 promoterprimers (Promega Biotech) and the Perkin Elmer CetusAmplilaq sequencing kit, according to the manufactun-en’s protocol. Sequences were analyzed using the lBlPustell Sequence Analysis Software (68) and the se-quence analysis programs for the VAX (69).

In Situ Hybridization. For in situ hybridization analysisof testicular tissues, frozen sections were used. Sampleswere fixed immediately upon dissection for 24 h in 4%parafonmaldehyde (in PBS, pH 7.4) at 4#{176}C,washed inPBS, and thereafter kept in 30% sucrose at 4#{176}Cuntilembedding. Samples were frozen in Tissue-Tek medium(Miles, Elkart, IN) with liquid nitrogen, and 6-8-sm sec-tions were cut on a cryostat and placed on Tespa-treatedslides, which were then heated to 37#{176}Cfor 2 h andstored at -70#{176}Cwith desiccant until further use. Prior tohybridization, the slides were allowed to defrost for 30mm at room temperature. The sections were then treatedas follows: postfixation in 4% paraformaldehyde (20 mm),PBS washes (2x 5 mm each), proteinase K treatment for8 mm (20 zg/ml in 5 mt�i EDTA-50 mt�i Tnis-HCI, pH 7.5),PBS wash (5 mm), refixation in 4% paraformaldehyde (5mm), incubation in 0.1 M tniethanolamine, pH 8.0, withacetic anhydnide added to 0.25%, PBS and 0.85% salinewashes (5 mm each), and dehydration in increasingethanol concentrations (50-95%, 2 mm each and 2x100%, 2 mm each).

Prehybnidization and hybridization were performedusing the procedure described by Jaffe et a!. (70). Theslides were viewed on a Leitz photomicroscope underbright field, dank field, and epiluminescence optics. Pho-tographs were taken using Kodak Technical Pan film.

504 Novel Mouse Meiosis Expressed Gene

Acknowledgments

We would like to thank Dr. Aaron P. Mitchell for providing the yeastRME1 clone, Deborah L. Chapman for providingthe cDNA library, PatrickBurke for providing the testicular zinc finger containing cDNA clones,and Cynthia Montanez and Chris Marshal for technical assistance. Weare grateful to Dr. Luigi Scotto, Dr. Martin Winer, Deborah L. Chapman,and Patrick Burke for fruitful discussions and critical comments.

References

1 . Nurse, P. Universal control mechanism regulating onset of M-phase.Nature (Lond.), 344: 503-508, 1990.

2. Pines, J. Cyclins: wheels within wheels. Cell Growth & Differ., 2: 305-

310, 1991.

3. Kassir, Y., Granot, D., and Simchen, G. IME1, a positive regulator geneof meiosis in S. cerevisiae. Cell, 52: 853-862, 1988.

4. Mitchell, A. P., Driscoll, S. E., and Smith, H. E. Positive control ofsporulation-specific genes by the IME1 and IME2 products in Saccharo-

myces cerevisiae. Mol. Cell. Biol., 10: 2104-2110, 1990.

5. Neigeborn, L., and Mitchell, A. P. The yeast MCK1 gene encodes aprotein kinase homolog that activates early meiotic gene expression.Genes & Dev., 5: 533-548, 1991.

6. Smith, H. E., and Mitchell, A. P. A transcriptional cascade governs

entry into meiosis in yeast. Mol. Cell. Biol., 9: 2142-2152, 1989.

7. Covitz, P. A., Herskowitz, I., and Mitchell, A. P. The yeast RME1 geneencodes a putative zinc finger protein that is directly repressed by ai-a2.Genes & Dev., 5: 1982-1989, 1991.

8. Mitchell, A. P., and Herskowitz, I. Activation of meiosis and sporulationby repression of the RME1 product in yeast. Nature (Lond.), 319: 738-742, 1986.

9. Alani, E., Padmore, R., and Kleckner, N. Analysis of wild-type andradSO mutants of yeast suggests an intimate relationship between meioticchromosome synapsis and recombination. Cell, 61: 419-436, 1990.

10. Engebrecht, j., and Roeder, G. S. MER1, a yeast gene required forchromosome pairing and genetic recombination, is induced in meiosis.Mol. Cell. Biol., 10: 2379-2389, 1990.

1 1 . Hollingsworth, N. M., Goetsch, L., and Byers, B. The HOP1 geneencodes a meiosis specific component of yeast chromosomes. Cell, 61:

73-84, 1990.

12. Klapholz, S., Waddell, C. S., and Esposito, R. E. The role ofthe spoilgene in meiotic recombination in yeast. Genetics, 1 10: 187-216, 1985.

13. Klapholz, S., and Esposito, R. E. Recombination and chromosomesegregation during the single division meiosis in spol2-l and spoi3-l

diploids. Genetics, 96: 589-61 1, 1980.

14. Thompson, E. A., and Roeder, C. S. Expression and DNA sequenceof RED1, a gene required for meiosis I chromosome segregation in yeast.Mol. Gen. Genet., 218: 293-301, 1989.

15. Wang, H. T., Frackman, S., Kowalisyn, J., Esposito, R. E., and Elder,R. Developmental regulation of Spol3, a gene required for the separationof homologous chromosomes at meiosis I. Mol. Cell. Biol., 7: 1425-1435,1987.

16. BeIlve, A. R., Cavecchia, J. C., Millette, C. F., O’Brien, D. A., Bhat-nagar, Y. M., and Dym, M. Spermatogenic cells of the prepuberal mouse:isolation and morphological characterization. J. Cell Biol., 74: 68-85,1977.

17. BeIlve, A. R., Millette, C. F., Bhatnagar, Y. M., and O’Brien, D. A.Dissociation of the mouse testis and characterization of isolated sper-

matogenic cells. I. Histochem. Cytochem., 25: 480-494, 1977.

18. Bennett, W. I., Gall, A. M., Southard, J. 1., and Sidman, R. L. Abnormalspermiogenesis in quaking, a myelin deficient mutant mouse. BioI. Re-prod., 5: 30-58, 1971.

19. Bryan, I. H. D. Spermatogenesis revisited: spermiogenesis in micehomozygous for two different male-sterile mutations (ps and hpy). CellTissue Res., 221: 169-180, 1981.

20. Hummel, K. I. Mouse News Lett. 34: 31, 1966.

21. Wolgemuth, D. I., Gizang-Ginsberg, E., Englemyer, E., Gavin, B. J.,and Ponzetto, C. Separation of mouse testis cells on a CeIsepTM apparatusand their usefulness as a source of high molecular weight DNA or RNA.Gamete Res., 12: 1-10, 1985.

22. Wolgemuth, D. j., and Watrin, F. List of cloned mouse genes withunique expression patterns during spermatogenesis. Mamm. Genome, 1:

283-288, 1991.

23. Geremia, R., Boitani, C., Conti, M., and Monesi, V. RNA synthesis inspermatocytes and spermatids and preservation of meiotic RNA duringspermiogenesis in the mouse. Cell Differ., 5: 343-335, 1977.

24. Monesi, V. Synthetic activities during spermatogenesis in the mouse:RNA and protein. Exp. Cell Res., 39: 197-224, 1965.

25. O’Brien, D. A., and BeIlve, A. R. Protein constituents of the mousespermatozoon: temporal synthesis during spermatogenesis. Dev. Biol.,75: 405-418, 1980.

26. Hecht, N. B. Regulation of “haploid expressed genes” in male germcells. J. Reprod. Fertil., 88: 679-693, 1990.

27. Kozak, M. Selection of initiation sites by eucaryotic ribosomes: effectof inserting AUG triplets upstream from the coding sequence for pre-proinsulin. Nucleic Acids Res., 12: 857-872, 1984.

28. Kozak, M. At least six nucleotides preceding the AUG initiator codonenhance translation in mammalian cells. J.Mol. Biol., 196: 947-950, 1987.

29. Bradley, A. A., Lee, A. L., Grendell, R. L., and Derynck, R. Inhibitionof translation of transforming growth factor-fi3 mRNA by its 5’ untrans-lated region. Mol. Cell. Biol., 1 1: 4306-4313, 1991.

30. Wickens, M., and Stephenson, P. Role of the conserved AAUAAAsequence: four AAUAAA point mutants prevent mRNA 3’ end formation.Science (Washington DC), 226: 1045-1051, 1984.

31. Baum, P., Furlong, C., and Byers, B. Yeast gene required for spindlepole body duplication: homology of its product with Ca2 binding protein.Proc. NatI. Acad. Sci. USA, 83: 5512-5516, 1986.

32. Walker, J. E., Saraste, M., Runswick, M. J., and Gay, N. J. Distantlyrelated sequences in the a- and j�-subunits of ATP synthase, myosin,kinase and other ATP-requiring enzymes and a common nucleotidebinding fold. EMBO J., 1: 945-951, 1982.

33. Berg, J. M. Potential metal-binding domain in nucleic acid bindingproteins. Science (Washington DC), 232: 485-487, 1986.

34. Hunt, L. T., and Dayhoff, M. 0. The occurrence in proteins of thetripeptides Asn-X-Ser and Asn-X-Thr and of bound carbohydrate.Biochem. Biophys. Res. Commun., 39: 757-765, 1970.

35. Kennelly, P. 1., and Krebs, E. G. Consensus sequences as substratespecificity determinants for protein kinases and protein phosphatases. J.Biol. Chem., 266: 15555-15558, 1991.

36. Coulombre, J. L., and Russel, E. S. Analysis of the pleiotropism at thew-Iocus in the mouse. J. Exp. Zool., 126: 277-291, 1954.

37. Handel, M. A., and Eppig, J. J. Sertoli cell differentiation in the testesof mice genetically deficient in germ cells. Biol. Reprod., 20: 1031-1038,1979.

38. Mintz, B., and Russel, E. S. Gene-induced embryological modifica-tions of primordial germ cells in the mouse. J. Exp. Zool., 134: 207-230,1957.

39. Kierszenbaum, A. L., and Tres, L. L. Structural and transcriptionalfeatures ofthe mouse spermatid genome. J. Cell Biol., 65: 258-270, 1975.

40. Oakberg, E. F. A description of spermiogenesis in the mouse and itsuse in analysis of the cycle of the seminiferous epithelium and germ cellrenewal. Am. J. Anat., 99: 391-413, 1956.

41. Gold, B., Fujimoto, H., Kramer, J. M., Erickson, R. P., and Hecht, N.B. Haploid accumulation and translational control of phosphoglyceratekinase-2 messenger RNA during mouse spermatogenesis. Dev. Biol., 98:392-399, 1983.

42. Kleene, K. C., and Flynn, J. F. Characterization of a cDNA cloneencoding a basic protein, TP2, involved in chromatin condensation duringspermiogenesis in the mouse. J. Biol. Chem., 262: 17272-17277, 1987.

43. Kleene, K. C., Borzorgzadeh, A., Flynn, J. F., Yelick, P. C., and Hecht,N. B. Nucleotide sequence of a cDNA clone encoding mouse transitionprotein 1. Biochim. Biophys. Acta, 950: 215-220, 1988.

44. Sing, J., and Satyanarayana Rao, M. R. Interaction of rat testis protein,TP, with nucleic acids in vitro. J. Biol. Chem., 262: 734-740, 1987.

45. Yelick, P. C., Kwon, Y. K., Flynn, J. F., Borzorgzadeh, A., Kleene, K.C., and Hecht, N. B. Mouse transition protein 1 is translationally regulatedduring the postmeiotic stages of spermatogenesis. Mol. Reprod. Dev., 1:

193-200, 1989.

46. Van Der Hoorn, F. A., Tarnasky, H. A., and Nordeen, S. K. A new ratgene RT7 is specifically expressed during spermatogenesis. Dev. Biol.,142: 147-154, 1990.

47. Conti, M., Geremia, R., and Monesi, V. Adenosine 3’,5’-cyclic mon-ophosphate dependent protein kinase activity in differentiating germ cellof mouse testis. Mol. Cell. Endocrinol., 13: 137-148, 1979.

48. Conti, M., Adamo, S., Geremia, R., and Monesi, V. Developmentalchanges of cyclic adenosine monophosphate-dependent protein kinaseactivity during spermatogenesis in the mouse. Biol. Reprod., 28: 260-

269, 1983.

Cell Growth & Differentiation 505

49. Pearson, R. B., and Kemp, B. E. Protein kinase phosphorylation sitesequences and consensus specificity motifs: tabulations. Methods En-zymol., 200: 62-81, 1991.

50. Fischman, K., Edman, J. C., Shackleford, G. M., Turner, J. A., Rutter,W. J., and Nir, U. A murine fer testis-specific transcript ( ferT) encodes atruncated Fer protein. Mol. Cell. Biol., 10: 146-153, 1990.

51. Keshet, E., Itin, A., Fischman, K., and Nir, U. The testis-specifictranscript ( ferT) of the tyrosine kinase FER is expressed during spermato-genesis in a stage-specific manner. Mol. Cell. Biol., 10: 5021-5025, 1990.

52. Ponzetto, C., and Wolgemuth, D. J. Haploid expression of a uniquec-abl transcript in the mouse male germ line. Mol. Cell. Biol., 5: 1791-1794, 1985.

53. Sorrentino, V., McKinney, M. D., Giorgi, M., Geremia, R., and Fleis-sner, E. Expression of cellular protooncogenes in the mouse male germline: a distinctive 2.4-kilobase pim-l transcript is expressed in haploidpostmeiotic cells. Proc. NatI. Acad. Sci. USA, 85: 2191-2195, 1988.

54. Nakanishi, S., lnoue, A., Kita, T., Nakamura, M., Chang, A. C. Y.,Cohen, S. N., and Numa, S. Nucleotide sequence of cloned cDNA forbovine corticotropin-fi-lipotropin precursor. Nature (Lond.), 278: 423-427, 1979.

55. Kilpatrick, D. L., Zinn, S. A., Fitzgerald, M., Higuchi, H., Sabol, S. L.,and Meyerhardt, J. Transcription of the rat and mouse proenkephalingenes is initiated at distinct sites in spermatogenic and somatic cells. Mol.Cell. Biol., 10: 3717-3726, 1990.

56. Breitbart, R. E., Andreadis, A., and Nadal-Ginard, B. Alternativesplicing: a ubiquitous mechanism for the generation of multiple proteinisoforms from single genes. Annu. Rev. Biochem., 56: 467-495, 1987.

57. Propst, F., Rosenberg, M. P., Iyer, A., Kaul, K., and Vande Woude,G. F. c-mos proto-oncogene RNA transcripts in the mouse tissues: struc-tural features, developmental regulation, and localization in specific celltypes. Mol. Cell. Biol., 7: 1629-1637, 1987.

58. Breathnach, R., and Chambon, P. Organization and expression ofeucaryotic split genes coding for proteins. Annu. Rev. Biochem., SO: 349-383, 1981.

59. Mount, S. M. Catalogue of splice junction sequences. Nucleic AcidsRes., 10: 459-472, 1982.

60. Cech, T. R. RNA editing: world’s smallest introns. Cell, 64: 667-669,1991.

61. Cleveland, D. W., Lopata, M. A., McDonald, R. I., Cowan, R. J.,Rutter, N. J., and Kirschner, M. W. Number and evolutionary conservationof a- and fi-tubulin and cytoplasmic �- and ‘y-actin genes using specificcloned cDNA probes. Cell, 20: 95-105, 1980.

62. Piechaczyk, M., Blanchard, J. M., Marty, L., Dani, C., Panabieres, F.,Sabouty, S. E. I., Fort, P., and Jeanteur, C. Post-transcriptional regulationof glyceraldehyde-3-phosphate dehydrogenase gene expression in rattissues. Nucleic Acids Res., 12: 6951-6962, 1984.

63. Cox, K. H., DeLeon, D. V., Angerer, L. M., and Angerer, R. C.

Detection of mRNAs in sea urchin embryos by in situ hybridization usingasymmetric RNA probes. Dev. Biol., 101: 485-502, 1984.

64. Sambrook, J., Fritsch, E. F., and Maniatis, T. Molecular Cloning: ALaboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Labo-ratory Press, 1989.

65. Cathala, G., Savouret, J. F., Mendez, B., West, B. L., Karin, M.,Martial, J. A., and Baxter, J. D. Laboratory methods: a method for isolationof intact translationally active ribonucleic acid. DNA, 2: 329-335, 1983.

66. Aviv, H., and Leder, P. Purification of biologically active globinmessenger RNA by chromatography on oligothymidylic acid-cellulose.Proc. Natl. Acad. Sci. USA, 69: 1408-1412, 1972.

67. Sanger, F., Nicklen, S., and Coulson, A. R. DNA sequencing withchain-termination inhibitors. Proc. NatI. Acad. Sci. USA, 74: 5463-5467,1977.

68. Pustell, J. M. Interactive molecular biology computing. Nucleic AcidsRes., 16: 1813-1820, 1988.

69. Devereux, J., Haeberli, P., and Smithies, 0. A comprehensive set ofsequence analysis programs for the VAX. Nucleic Acids Res., 12: 387-395, 1984.

70. Jaffe, L., Jeannotte, L., Bikoff, E. K., and Robertson, E. I. Analysis of

�92-microglobulin gene expression in the developing mouse embryo andplacenta. I. Immunol., 10: 3474-3482, 1990.