A 60-bp Core Promoter Sequence of Murine Lactate Dehydrogenase ...

BIOLOGY OF REPRODUCTION 51, 425-432 (1994)

A 60-bp Core Promoter Sequence of Murine Lactate Dehydrogenase C Is Sufficient to Direct Testis-Specific Transcription In Vitro' WENTONG ZHOU, JIANHUA XU, and ERWIN GOLDBERG 2 Department of Biochemistry, Molecular Biology and Cell Biology, Northwestern University Evanston, Illinois 60208 ABSTRACT A clone containing the 5' flanking region of the testis-specific murine lactate dehydrogenase C (Ldbc) gene was isolated from a mouse genomic library. Promoter activity was demonstrated within a 720-bp fragment in testis nuclear extract (TN). Interestingly, the addition of liver nuclear extract (LN) significantly repressed Ldhc promoter activity in the transcription assay system. Sequence analysis of this promoter region revealed several ubiquitous cis-regulatory elements, including one TATA box, one GC box, and two putative CCAAT elements. Analysis of a series of deletion mutants revealed that a 60-bp core promoter sequence was sufficient to direct basal, testis-specific transcription in an in vitro transcription system. A 103-kDa protein in TN and a 65kDa protein in LN bind to the same palindromic sequences within the 60-bp core promoter region.

INTRODUCTION

ficiently. A 60-bp fragment is able to direct transcription in testis nuclear extract (TN) but not in liver nuclear extract (LN). This core promoter sequence contains both a TATA box and a 30-bp palindromic sequence as part of the initiation site for transcription. Gel-retardation and Southwestern assays have detected different protein-binding activity to this palindromic sequence in TN and LN.

Spermatogenesis is a complex program that consists of a variety of distinct cellular events such as meiosis and differentiation of haploid male germ cells to spermatozoa. While the developmental sequence of these events has been well characterized morphologically, the molecular mechanism that coordinates gene expression during spermatogenesis is mostly unknown, partly because of the lack of a differentiated germ cell line in culture. Among the genes expressed during spermatogenesis are some that encode testis-specific isozymes or isoforms of somatic proteins. The Ldhc gene is one example. The lactate dehydrogenases in mammals are encoded by three distinct loci designated Ldha, Ldhb, and Ldhc. The abundance and distribution of mRNAs produced from this gene family vary in different tissues. While the expression products of the Ldha and Ldhb genes are found in almost all tissues as homotetramers or heterotetramers (LDH-A 4, -A3B1, -A2B2, -AB3, -B4), the Ldbc gene product (LDH-C 4) ap-

MATERIALS AND METHODS Amplification of Intron 1 of the Mouse Ldhc Gene A sense oligonucleotide, MC5prime (5'-CATGAATTCTTACCTGTGCTGCGG-3'), and an antisense oligonucleotide, MC45-24 (5'-CATGGTACCGGAACTAGGTFCTGAATCAG-3'), were designed for polymerase chain reaction (PCR) amplification according to the mouse Ldhc cDNA sequence [5,6]. These oligonucleotides flank a putative consensus splice site in the 5' untranslated region. A synthetic EcoRI or Kpn I site was attached to the 5' end of oligonucleotide MC5prime or MC45-24, respectively, to facilitate cloning. Mouse genomic DNA was restricted by EcoRI and used as PCR template. In each PCR mixture, 1 pug of digested genomic DNA was amplified with 50 pmol each of the primers MC5prime and MC45-24 under the standard conditions of a denaturing step at 94°C for 4.5 min followed by 35 cycles of 1-min denaturing at 940C, 2-min annealing at 450C, and 3-min elongation at 720C.

pears only in sperm [1] and testis [2]. Expression of this gene is restricted to the germinal epithelium and is developmentally regulated. The murine Ldhc mRNA is first detected in the preleptotene spermatocyte and persists postmeiotically to the round spermatid stage [3]. The promoter of human Ldbc has been cloned and studied in a cell coculture system [4] in which a germinal cell-specific expression pattern of the reporter gene was demonstrated when it was driven by a 180-bp 5' sequence. To compare the regulation of human and mouse Ldhc genes, we isolated the 5' flanking region of the mouse Ldhc gene and studied its promoter activity in vitro. In this system, the mouse Ldhc promoter can drive transcription ef-

Mouse Genomic DNA Library Screening The PCR-amplified intron 1 of the mouse Ldbc gene was cloned into pBluescript KS(+) vector (Stratagene, La Jolla, CA), sequenced, and 32 P-labeled with a random-primed labeling kit (U.S. Biochemicals, Cleveland, OH). A XEMBL 3 mouse liver genomic library (kindly provided by Dr. Jose Luis Millan) was screened with this probe by standard procedure [7]. Two positive clones, 19-1 and X12-1, were

Accepted April 18, 1994. Received December 21, 1993. 'This research was supported by NICHD grant HD05863. 2Correspondence. FAX: (708) 467-1380.

425

426

ZHOU ET AL.

found and characterized. They were found to be identical by partial restriction mapping.

Sst I/Kpn I, and SstI/Hinc II, respectively, treated by mung bean nuclease, and re-ligated to create plasmid p421(C 2AT), p83(C 2AT), and p46(C 2AT).

Sequence Analysis The recombinant phage clone 19-1 was cleaved by restriction enzymes and subcloned into the pBluescript KS(+) vector. Sequencing of both strands was performed by the dideoxy-chain termination method with a Sequenase kit (U.S. Biochemicals) using universal or synthetic oligonucleotides as primers. Primer Extension A primer extension experiment [8] was performed with a synthetic 22-bp oligonucleotide Exonl-3' (5'-TACTGCTGACTCCGCAGCACAG-3') complementary to the 3' sequence of exon 1. The primer was end-labeled using y-32PATP, and 105 cpm primer was hybridized to 20 Rpg of total mouse testis RNA. The primer extension was performed by AMV-reverse transcriptase (Promega, Madison, WI) in the presence of unlabeled dNTP at 42°C for 90 min. The products were separated by electrophoresis on an 8% polyacrylamide gel along with the sequencing reaction product of M13 single-stranded DNA as size markers. Preparationof Nuclear Extracts Testis and liver nuclear extracts were prepared from 50 or 3 adult mice, respectively, as previously described [9, 10], and were designated TN and LN. The concentration of the nuclear extracts determined by the Bio-Rad (Richmond, CA) protein assay system was consistently between 5 and 10 mg/ ml. Plasmids The plasmids p(C2AT) and pML(C2AT) were generous gifts from Dr. R.G. Roeder and contained a "G-less cassette" as transcription template [11]; pML(C2AT) includes the adenovirus major late promoter and serves as a positive control for transcription, and p(C 2AT) is a promoter-less plasmid. The latter was modified slightly by double digestion with Pst I and Xba I to eliminate part of the polylinker of the original PUC13 plasmid and then re-ligated to itself. This modified p(C2AT) was linearized at a unique Sst I site immediately upstream of the G-less cassette, blunted with mung bean nuclease (GIBCO BRL, Gaithersburg, MD), and used for cloning. A 720-bp Ldc 5' flanking region fragment (-710 to +12) was amplified by PCR using a 5' sense oligonucleotide, Sstsens (5'-GTCCCAGAGAGCTCTGGG-3') (corresponding to -710 to -693 and containing an Sst I site), and a 3' antisense oligonucleotide, HA (5'-GGATAACTGTTGGGTCCAGGAGCCAACAGTTA-3') (corresponding to +12 to -20). This PCR fragment was cloned into the blunted Sst I site of the modified p(C2AT) plasmid to create plasmid p710(C 2AT). The p710(C2 AT) was then double-digested with Sst I/Acc I,

In Vitro Transcription In vitro transcription was performed as previously described [11, 12]. Appropriate amounts of mouse LN or TN were incubated in reaction buffer to a total volume of 50 ,pl [11, 12] with 600 p.M ATP and UTP, 25 p.M ca32P-CTP (Amersham Corp., Arlington Heights, IL), and 1 ug of DNA template. The mixtures were incubated for 1 h at 30°C (for TN) or 37°C (for LN). After phenol/chloroform extractions and ethanol precipitation, the RNA sample was dissolved in 98% formamide and electrophoresed on a 4% polyacrylamide gel containing 7 M urea in single-strength TBE (0.089 M Tris-borate, 0.089 boric acid, 0.002 M EDTA) running buffer. The gel was dried onto 3 M chromatography paper and subjected to autoradiography using Kodak XAR5 film (Eastman Kodak, Rochester, NY). Efficiencies of specific transcriptions were quantitated by phosphoimager measurement of total synthesized RNA. Transcription efficiencies of different constructs were normalized to percentages of positive control template (pML(C 2AT)) efficiency. Gel Retardation Equal molar amounts of oligonucleotide HS (5'-ATAACTGTTGGCTCCTGGACCCAACAGTTAT-3') and HA (5'were GGATAACTGTTGGGTCCAGGAGCCAACAGTTA-3') annealed and labeled by 32P-dCTP with Klenow fragment (Promega, Madison, WI). The probe was purified through a G-50 Sephadex spin column. Binding reactions were carried out as described elsewhere [10]. Southwestern Analysis The Southwestern assay was performed according to the procedure of Miskimins et al. [13], except that 10 ,ug/ml salmon sperm DNA was added to the hybridization solution as nonspecific competitor. RESULTS Isolation of the First Intron and the 5' FlankingRegion of the Murine Ldhc Gene To avoid nonspecific hybridization between Ldhc cDNA and Ldh pseudogenes, which have high homology to the functional Ldhc gene itself, we developed a specific probe to screen a mouse genomic library. As a member of the Ldh gene family, Ldhc was expected to be conserved with intron 1 in the 5' untranslated region [4, 14-16]. We synthesized a sense and an antisense oligonucleotide flanking the consensus splicing sequence in the 5' untranslated region of the mouse Ldhc cDNA sequence [5,6] and amplified mouse genomic DNA by PCR. One 540-bp band was visu-

TESTIS-SPECIFIC Ldbc GENE EXPRESSION

427

-730

Sst I TTTCTTTGGT TGGTGGTTTA GTCCCAGAGA GCTCTGGGGG TGCTGGTTAG

-680

TTTCTGTTGT TTCTCCTATA GAGCTACAGA CCCCTTTGGC TCTTTGGATA

-630

CATTCTTTAG CTCCTTCATT GGGGACACTG TGTTCCATCC AATGAATGAC

-580

TGTGAGCATC CACTTCTGTA TTTGTCAGGC ACTGGCAGAG CCTCCCAGGA

-530

GACACCTATA TTAGGCTCCT GTCAGCAAAA TTTTGTATCT GCAATAGTGT

-480 -430

CTGGGTTTGG TTGGTTGTTC ATGAGACGGA TCTCTTGAGT TGGGAGACCA A I GCCTGGTCTA CAGAGTTCCA GGACGGCCAG GGCTACACAG AGAAACCCTG

-380

GCTCGAAAGC AAAACAAAAC AAAACAAAAC AAAAGCAAGC AAACAGACAG

-330

ATGCCCCCAA CCCTCAATGC CTACACACAG ATGTAAGGGC AAAGGGTCCA

-280

TAGGTATGCT CCTCAAACTC AAAGTTCTGC ATTTCCTCCC AGCAGAAAAG

-230

TGGCCTTTAA AGGGATGCTG GCCAAACTGG CTGCCATCTT CTGTGTGGTC

-180

-80

CCAGGGCTTC TACATCTTAC ACTAGAGGCC TTGGAAACCG TCTGGAGTCG KpnI TCTCCTAAAA CCTGAGGTAG GCCATGCAGG TGCACGCACC AAGGTACCTG Hinc II AGCAGGCAGT gSqGGGCT TGCGTGCTGA CGTTGACTTT GTGACGTTCC

-30

TTTTCCGTTA TAACTGTTGG CTCCTGGACC CAACAGTTAT CCTGGTTTCT

+21

TACCTGTGCT GCGGAGTCAG CAGTAAGgtg tgtgtttt

-130

-

FIG. 1. Sequence of the 5' flanking region of the murine Ldhc gene. The transcription start site is numbered +1 and indicated with a dot underneath. Consensus sequence features are shown as follows: TATA box (underlined), GC box (double-underlined), and CCAAT box (half arrows). A 30-bp palindromic sequence is indicated by a pair of inverted arrows. The sequences of the 5' flanking region and exon 1 are in uppercase letters. The intron 1 sequence is in lowercase letters. The unique restriction sites for making deletion constructs are also indicated.

alized, cloned into pBluescript KS(+), and sequenced. This fragment, which contains all of intron 1 flanked by part of the first (noncoding) exon and the second (coding) exon sequence, was used as a probe for screening the mouse XEMBL 3 library. We identified two positive clones, designated 19-1 and X20-1, which appeared to be identical to each other by restriction mapping. Therefore, only 19-1 was subcloned

and sequenced. This clone contains at least 10 kb of the 5' flanking region of the mouse Ldhc gene. The sequence of the promoter region is shown in Figure 1. By primer extension assay, a single transcription start site was mapped to the C residue 18 bp downstream of the consensus TATA elements (Fig. 2). This result was confirmed by S1 nuclease protection assay (data not shown). Therefore untranslated exon 1 is 46 bp long.

428

ZHOU ET AL.

A

Exon 2

Exon 1

4

*

B

FIG. 2. Transcription initiation site mapped by primer extension assay. Synthetic oligonucleotide exon 1-3' was labeled and used to prime total mouse testis RNA. A) The gene structure is represented as follows: coding sequences in exon 2 as a hatched box; 5' untranslated sequences as open boxes; intron 1 as a solid line. The primer is represented by an arrow line with the asterisk showing the labeled end. B) Both the primer and the extension product are indicated with arrows. Their size is determined by comparison to an M13 sequencing ladder.

A computer-assisted search (GCG program, University of Wisconsin, Madison, WI) did not reveal any significant homology in the 5' flanking region of the murine and the human Ldhc genes. We were surprised to find that their putative cis-regulatory elements differed drastically in the promoter regions because the two genes show the same cell specificity of expression and are also 73% similar to each other in protein coding sequences [5, 6,17]. While the mouse Ldhc gene has a single transcription initiation site, a consensus TATA box (-18), one GC box (-62), and two inverted CCAAT box sequences (-608, -586) (with respect to the transcription start site as +1), the human gene has only a stretch of GC-rich region and multiple transcription start sites ([4]). Also of interest in the 5' flanking region of the murine Ldhc gene is a 31-bp sequence (-21 to +10) containing two 15-bp inverted repeats with a single mismatch (Fig. 1). This palindromic sequence overlaps part of the TATA box and includes the transcription initiation site. Testis-Specific Transcriptionfrom a 710-bp Promoter Fragment The in vitro transcription system described by Sawadogo and Roeder [11] was used to assay functional activity of the Ldhc gene promoter sequence. TN and LN were prepared

FIG. 3. Testis-specific transcriptional activity of the Ldhc promoter. In vitro transcription assay in TN (solid column) and LN (hatched column) using constructs described under Materials and Methods: -, negative control p(C2AT); +, positive control pML(C2AT); Ldhc, p710(C 2AT) containing 720bp (-710 to +13) mouse Ldhc genomic fragment. The vertical axis is the relative ratio of phosphoimager readings shown as the percentage of the positive control pML(C2AT) in TN (mean + SEM; N = 3). A representative autoradiogram is shown below the graph.

from adult mice as described by Bunick et al. [9] and Gebara et al. [10]. In comparison with the promoter-less negative control p(C2AT), the positive control pML(C 2AT), which contains the adenovirus major late promoter, was transcribed efficiently by both LN and TN (Fig. 3). It is interesting but not surprising that pML(C 2AT) has 1.7-3-fold higher activity in LN than in TN. This result is consistent with the previous observation that the adenovirus major late promoter has a higher activity in somatic cells than in testis in vitro [18]. To study Ldhc promoter activity, we constructed a plasmid, p710(C 2AT), with a PCR-amplified 720-bp Ldhc genomic fragment (-710 to +13) (see Materials and Methods). The p710(C 2 AT) had 20- to 40-fold higher transcription activity in TN than in LN (Fig. 3). The transcription level of p710(C 2AT) in LN was not significantly different from that of the negative control p(C2AT) (Fig. 3). Similar results were obtained when nuclear extracts from kidney and two somatic cell lines were used in this in vitro transcription system (data not shown). These results indicate that the 720bp Ldhc genomic fragment contains essential cis-regulatory elements required for testis-specific expression of the Ldhc gene. The 720-bp Ldhc fragment is not transcriptionally active in LN (Fig. 3). Either the presence of an effective repres-

429

TESTIS-SPECIFIC Ldhc GENE EXPRESSION

sor(s) or the absence of an essential transcription factor(s) in liver could be responsible. To distinguish between these two possibilities, we assayed transcription of p710(C2AT) in a mixture of TN and LN (Fig. 4). Again, p710(C2AT) in TN alone (Fig. 4, lane 3) was transcribed efficiently in comparison with the positive control pML(C 2AT) (lane 1). When 40 Cig of TN protein was premixed with 20, 40, 60, or 80 ,Ig of LN protein (Fig. 4, lanes 4-7), the transcriptional activity of p710(CzAT) decreased 2.2- to 18-fold, respectively. In a control experiment, transcription driven by the positive control pML(C 2AT) in the mixture of LN and TN (Fig. 4, lane 2) was slightly increased above that of TN alone (Fig. 4, lane 1). This result suggests that there is an Ldhc promoter-specific repression of transcription in LN. A repressor present in LN but not in TN would effectively inhibit transcription controlled by the Ldhc promoter. Most important of all, this repressor activity is targeted to the 720bp genomic fragment.

800 .o

C

= 600

U

CO

F 400 0 X

Delineation of the Ldhc Promoter

To locate the cis-regulatory elements in the 720-bp fragment responsible for testis-specific expression of the Ldhc gene, we constructed a series of deletions from the 5' end. These constructs, schematically illustrated in Figure 5A, were all transcribed in TN or LN. Transcription efficiency data were compared (Fig. 5, B and C). Deletion from the Sst I to the Acc I site (-710 to -421) led to a 3-fold loss of promoter activity (Fig. 5B, lanes 2 and 3). Deletion of the sequence between Acc I and Kpn I (-421 to -83) caused an additional 3-fold loss of activity (Fig. 5B, lanes 3 and 4), suggesting the presence of positive cis-acting elements in these two regions. There are two inverted repeats of consensus CCAAT boxes (-608, -586), which may be responsible for the decreased promoter activity when the region -710 to -421 is deleted. The construct p83(C2AT) is 2- to 3-fold more efficient than p46(C2AT), the latter of which does not have a GC box (Fig. 5B, lane 5). All of these constructs show high to low levels of transcriptional activity in TN, but not in LN (Fig. 5C, lanes 2-5). Differential Protein-BindingActivity in LN and TN

The same TN and LN used for the in vitro transcription assay were used also to study DNA-protein interactions. Gel retardation by an end-labeled palindromic oligonucleotide incubated with LN (Fig. 6A, lanes 1 and 2) and TN (lanes 3 and 4) was assayed. One mobility shift band was observed with LN (Fig. 6A, lane 1). A very faint band with the same mobility was also detected in TN (Fig. 6A, lane 3), probably due to the contribution of the somatic cells in the testis. A predominant band with higher mobility was seen in testis (Fig. 6A, lane 3). Both of the shift bands were competed with 100-fold excess palindromic oligonucleotide (Fig. 6A, lanes 2 and 4). In a Southwestern assay, a single polypeptide (103 kDa) in TN bound to the palindromic oligonu-

I

200

16

Y/

n

n_ I b

u

I ·· 1·

1 '2

TN 40 LN 0

40 40

. a

_

S

FV-,, I

I

.ll

X

I11 II

3

4

5

40 0

40

40

40

40

20

40

60

80

6

7

FIG. 4. Repression of transcriptional activity by LN. Positive control pML(C 2AT) is transcribed in 40 fig of TN (lane 1) or in the mixture of 40 jig of TN and 40 fig of LN (lane 2). Mouse Ldhc promoter construct p710(C 2AT) (-710 to +13) is transcribed in 40 pig of TN premixed with 0 (lane 3), 20 (lane 4), 40 (lane 5), 60 (lane 6), or 80 (lane 7) BIg of LN to a final reaction volume of 50 I. The vertical axis is the relative ratio of phosphoimager readings shown as the percentage of the positive control pML(C2AT) (mean SEM; N = 3). Phosphoimager counts of the radioactive transcripts of expected sizes reflect the relative transcriptional activity of different constructs or the activity under different conditions. A representative autoradiogram is shown above the graph.

cleotide probe (Fig. 6B). This band was barely detectable in LN. A second polypeptide (65 kDa) was found only in LN and was not detectable in TN even after longer exposure. Presumably, the lower sensitivity of the Southwestern assay precludes detection of the small amount of protein contributed by testis somatic cells. When a nonspecific double-stranded oligonucleotide hcex2 is used as a probe, no protein-DNA hybridization is detectable in either TN or LN (Fig. 6B). DISCUSSION In the present study, we have isolated the 5' flanking region of the murine Ldhc gene. Analysis of the Ldhc promoter sequence reveals several common regulatory motifs, e.g., a TATA box, a GC box, and a CCAAT box, all of which are located at appropriate distances from the cap site. The presence of a single transcription initiation site suggests that the TATA box is functional. We also demonstrated that a 720-bp promoter fragment has potent in vitro transcrip-

430

ZHOU ET AL.

A1. pML(C2AT)

Template

Ad ML Promoter Accl

Setl 2. p710(C2AT)

3. p421 (C2AT)

I

CAAT

b,

Ac

/l

I

C

Kpnl

Template

I \\\

Hincll

palindrome

ATA tAtt

-421

lpnI

3 -83

Hincll palindrome Hinc l TAIA GC Hinell

5. p46(C2AT)

palindrome

TAA

-710

Kpnl 4. p83(C2AT)

Hinlil

KpnI

I

\\\\\\\

Template \\\\\

Template palindrome \\\\\\\\\\\ TA'A Template

e. p(C2AT)

TESTIS-SPECIFIC Ldhc GENE EXPRESSION

FIG. 6. Protein binding to the palindromic sequence in LN (L} and TN (T). A) Gel-retardation assay: labeled palindromic oligonucleotide was incubated with LN (lanes 1 and 2) and TN (lanes 3 and 4). A 100-fold excess of specific competitor was added in lanes 2 and 4. Specific mobility shift bands are indicated by arrows. B) Southwestern assay using the palindromic oligonucleotide (palind) or a nonspecific oligonucleotide (hcex2) as probe. Sizes of molecular mass markers are indicated.

tional activity in TN but not in LN. This result is consistent with the observation that the mouse Ldhc message is abundant in testis but undetectable in somatic tissues [19]. Several other spermatogenic-specific promoters (e.g., angiotensin-converting enzyme, RT7, mouse protamine 1, mouse protamine 2, and mouse Pgk-2) were studied previously by in vitro transcription assay [9,18, 20-22]. It has been shown that repression of some of the testis-specific genes in nonexpressing tissues is caused by cis-negative regulatory elements [9, 23] rather than by other mechanisms like chromatin structure and/or DNA methylation (reviewed in [24]). Our data indicate also that the mouse Ldhc promoter is inhibited effectively by liver extract. It is possible that somatic cell repression may be targeted to the core promoter in the mouse Ldhc gene whereas other testis-specific genes may be inhibited through negative regulatory elements further upstream or downstream; this would require bending of DNA to juxtapose a repressor protein and the transcriptional machinery, an event not perfectly reproducible in an in vitro system. In vitro studies of other testis-specific promoters [9,18,2022] identified several interesting regions that enhance transcription in TN but not LN. Gel-retardation or footprinting assays detected testis-specific DNA-protein interactions with

FIG. 5. Transcriptional activity of the 5' deletion constructs of the Ldhc promoter. A) Schematic representation of positive control, negative control, and 5' deletion constructs of the Ldhc promoter; (B) transcription in TN; (C) transcription in LN. Construct numbers in (A) correspond to the construct numbers in (B) and (C). Different preparations of the plasmid DNA were used. The relative transcriptional activity of each construct is shown as the percentage of the positive control pML(C 2AT) (mean SEM; N = 3). Two representative autoradiograms are shown above the corresponding graph. CAAT, CCAAT box; GC, GC box; TATA, TATA box.

431

putative promoter sequences of protamine-2 [25], Pgk-2 [10], and the rat histone Hit gene [26]. However, there was no significant homology between those putative testis-specific activating elements and the 5' flanking region of mouse Ldbc. This finding also implies a different regulatory mechanism for mouse Ldhc gene transcription. Lim and Chae [27] identified a repressor binding site between the TATA box and the transcription initiation site in the rat testis-specific TH2B histone gene. The cognate repressor protein was detected in pre-meiotic spermatogenic cells but not in adult testis and liver, suggesting a role in developmental regulation of the gene. However, this element apparently is not involved in repression of the TH2B gene in somatic tissue or in the mature testis. Again, there is no sequence in the Ldhc promoter homologous to this repressor sequence. Overall, the picture is still far from clear. It seems unlikely that only one or a few testis-specific factors could govern the regulation of all the testis-specific genes, especially since different genes are turned on at different stages during spermatogenesis. Transcriptional activity of the 720-bp Ldhc promoter decreases with the deletion of 5' end sequences, probably because of the loss of the common cis-activating elements such as CCAAT boxes or GC boxes. Usually, multiple cisregulatory elements are needed to mediate tissue-specific expression in eukaryotes [28, 29]. In this respect, the simplicity of the mouse Ldhbc promoter is remarkable in that while the upstream regions are required for a high level of transcription, short promoter sequences (60 bp) are sufficient for directing testis-specific transcription in vitro. In vivo, however, Ldhc gene regulation may turn out to be much more complicated. It would not be too surprising if multiple cis elements are needed to impose the testis-specific expression pattern of this gene. Within the 60-bp core promoter region, there are two recognizable sequence motifs: the TATA box and an unusual 30-bp palindromic sequence overlapping the TATA box and transcription start site. The TATA box, of course, is a common regulatory element that should be functional in both LN and TN. Therefore, the palindromic sequence may be very important, because of its positioning, in determination of testis-specific expression. Binding of transregulators to this palindromic sequence could either facilitate or inhibit the assembly of the transcription initiation complex. The latter could also cause effective inhibition of transcription in vitro. It is tempting to speculate that tissuespecific transcription of the Ldhc gene is a direct result of the differential binding to the palindromic sequence in expressing and nonexpressing tissues that we detected by both Southwestern and gel-retardation assays. In order to understand the potential functions of the palindromic sequence in gene regulation and the significance of protein interactions in TN and LN, extensive mutagenesis of this 30bp inverted repeat is needed to identify the essential nucleotides correlated with both protein binding and effects

432

ZHOU ET AL.

on transcriptional activity. Ultimately, in vivo experiments (e.g., a transgenic study) are needed to verify the in vitro results. ACKNOWLEDGMENTS We thank Dr. R.G. Roeder for providing the plasmid p(C2 AT) and pML(C 2AT) and Dr. Jose Luis Millan for providing the mouse genomic library. We also thank Dr. D. Engel, Dr. S. Ness, and Dr. C. Bonny for advice and critical reading of this paper.

REFERENCES 1. Goldberg E. Lactic and malic dehydrogenase in human spermatozoa. Science 1963; 139:602-603. 2. Blanco A, Zinkham WH. Lactate dehydrogenases in human testes. Science 1963; 139:601-602. 3. Thomas K, Del Mazo J, Eversole P, Bellv A, Hiraoka Y, Li SS-L, Simon M. Developmental regulation of expression of the lactate dehydrogenase (LDH) multigene family during mouse spermatogenesis. Development 1990; 109:483-493. 4. Cooker LA, Brooke CD, Kumari M, Hofmann M-C, Millan JL, Goldberg E. Genomic structure and promoter activity of the human testis lactate dehydrogenase. Biol Reprod 1993; 48:1309-1319. 5. Sakai I, Sharief FS, Li SS-L. Molecular cloning and nucleotide sequence of the cDNA for sperm-specific lactate dehydrogenase-C from mouse. Biochem J 1987; 242:619-622. 6. Wu KC, Chan K, Lee C-YG, Lau Y-F. Molecular isolation and sequence determination of the cDNA for the mouse sperm-specific lactate dehydrogenase-X gene. Biochem Biophy Res Commun 1987; 146:964-970. 7. SambrookJ, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual, 2nd edition. New York: Cold Spring Harbor Laboratory Press; 1989. 8. Ausubel FM. Current Protocols in Molecular Biology. New York: Greene Publishing Associates and Wiley Interscience; 1987. 9. Bunick D, Johnson PA, Johnson TR, Hecht NB. Transcription of the testis-specific mouse protamine 2 gene in a homologous in vitro transcription system. Proc Natl Acad Sci USA 1990; 87:891-895. 10. Gebara MM, McCarrey JR. Protein-DNA interactions associated with the onset of testis-specific expression of the mammalian Pgk-2 gene. Mol Cell Biol 1992; 12:1422-1431. 11. Sawadogo M, Roeder RG. Factors involved in specific transcription by human RNA polymerase II: analysis by a rapid and quantitative in vitro assay. Proc Natl Acad Sci USA 1985; 82:4394-4398. 12. Dignam JD, Lebowitz RM, Roeder RG. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res 1983; 11:1475-1489.

13. Miskimins WK, Roberts MP, McClelland A, Ruddle FH. Use of a protein-blotting procedure and a specific DNA probe to identify nuclear proteins that recognize the promoter region of the transferrin receptor gene. Proc Nad Acad Sci USA 1985; 82:6741-6744. 14. Chung F-Z, Tsujibo H, Bhattacharyya U, Sharief FS, Li SS-L. Genomic organization of human lactate dehydrogenase-A gene. Biochem J 1985; 231:537-541. 15. Fukasawa KM, Li SS-L. Nucleotide sequence of the putative regulatory region of mouse lactate dehydrogenase-A gene. Biochem J 1986; 235:435-439. 16. Takano T, Li SS-L Structure of the human lactate dehydrogenase B gene. Biochem J 1989; 237:921-924. 17. Millan JL, Driscoll CE, LeVan KM, Goldberg E. Epitopes of a human testis-specific lactate dehydrogenase deduced from a cDNA sequence. Proc Natl Acad Sci USA 1987; 84:5311-5315. 18. Tamura T, Makino Y, Mikoshiba K, Muramatsu M. Demonstration of a testis-specific trans-acting factor Tet- in vitro that binds to the promoter of the mouse protamine 1 gene. J Biol Chem 1992; 267:4327-4332. 19. Salehi-Ashtiani K, Goldberg E. Differences in regulation. of testis-specific lactate dehydrogenase in rat and mouse occur at multiple levels. Mol Reprod Dev 1993; 35:1-7. 20. Goto M, Masamune Y, Nakanishi Y. A factor stimulation transcription of the testisspecific Pgk-2 gene recognizes a sequence similar to the binding site for a transcription inhibitor of the somatic-type Pgk-1 gene. Nucleic Acids Res 1993; 21:209214. 21. Howard T, Balogh R, Overbeek P, Bernstein KE. Sperm-specific expression of angiotensin-converting enzyme (ACE) is mediated by a 91-base-pair promoter containing a CRE-like element. Mol Cell Biol 1993; 13:18-27. 22. van der Hoorn FA, Tarnasky HA. Factors involved in regulation of the RT7 promoter in a male germ cell-derived in vitro transcription system. Proc Natl Acad Sci USA 1992; 89:703-707. 23. Mizuno K, Goto M, Masamune Y, Nakanishi Y. A silencer-like cis element for the testis-specific phosphoglycerate-kinase-2-encoding gene. Gene 1992; 119:293-297. 24. Cedar H, Razin A. DNA methylation and development. Biochem Biophys Acta 1990; 1049:1-8. 25. Johnson PA, Bunick D, Hecht NB. Protein binding regions in the mouse and rat protamine-2 genes. Biol Reprod 1991; 44:127-134. 26. Grimes SR, Wolfe SA, Koppel DA. Tissue-specific binding of testis nuclear proteins to a sequence element within the promoter of the testis-specific histone Hlt gene. Arch Biochem Biophys 1992; 296:402-409. 27. Lim K, Chae C-B. Presence of a repressor protein for testis-specific H2B (TH2B) histone gene in early stages of spermatogenesis. J Biol Chem 1992; 267:1527115273. 28. Grosschedl R, Baltimore D. Cell-type specificity of immunoglobulin gene expression is regulated by at least three DNA sequence elements. Cell 1985; 41:885897. 29. Maniatis T, Goodbourn S, Fischer JA Regulation of inducible and tissue-specific gene expression. Science 1987; 236:1237-1244.