Molecular Cloning of Rat Prostate Transglutaminase Complementary DNA

Complementary DNA (cDNA) that codes for a major androgen-dependent secretory protein of rat coagulat- ing gland and dorsal prostate, dorsal protein 1 (DPl), was isolated by molecular cloning. Recombinant DP1 cDNA clones were identified from a bacteriophage X gt 11 rat coagulating gland expression library using an affinity purified polyclonal antibody. Amino acid sequence deduced from DNA contained sequences iden- tical with several DPl cyanogen bromide cleavage fragments. Northern blot hybridization of poly(A) RNA isolated from intact rat dorsal prostate and coagulating gland revealed a predominant messenger RNA (mRNA) species of approximately 3200 nucleo- tides. Tissue-specific expression of DP1 mRNA was indicated by the absence of DP1 mRNA in ventral prostate and other tissues of the rat. Expression of DP1 mRNA was androgen-dependent, decreasing approxi- mately 80% 7 days after castration and increasing rapidly following androgen replacement. Southern blot analysis of restriction enzyme-digested rat DNA indicated that DP1 is encoded by a single gene and that no major genomic rearrangements accounted for its lack of expression in the dorsal prostate-derived rat Dunning tumor. Sequence comparisons revealed that rat prostate DP1 shares sequence identity with Factor XIIIa and tissue transglutaminase, including the active center, GQCWVF, indicating that DP1 is a member of the transglutaminase gene family.

The multiple lobes of the rat prostate synthesize in response to androgen several abundant proteins destined for secretion into the seminal fluid. Although the functions of most of these proteins remain unknown, they have served as useful indicators for studies on androgen regulation of prostate gene expression (1)(2)(3)(4)(5)(6)(7)(8). Two anatomically distinct regions of the rat prostate complex are the anterior lobes, known as the coagu-*This work was supported by Grants HD16910, HD04466, and P30-HD18968 (DNA core) from the National Institute of Child Health and Human Development Center for Population Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) M903 IO.
$ To whom correspondence should be addressed Laboratories for Reproductive Biology, CB#7500 MacNider Bldg., University of North Carolina, Chapel Hill, NC 27599. Tel.: 919-966-5159;Fax: 919-966-2203. lating glands, and dorsal lobes. Each coagulating gland is situated on the concave side of the seminal vesicle, whereas the dorsal prostate forms a portion of the prostate complex that includes the lateral prostate and surrounds the urethra at the neck of the bladder (9). Despite their anatomical differences, the coagulating gland and dorsal prostate synthesize a common abundant protein referred to as dorsal protein 1 (DP1)l (9). Western blot analysis and immunoblotting revealed DP1 in the glands and ducts of rat type 1, but not type 2, lateral prostate (10). DP1 is a 150-kDa dimeric protein (62 kDa on SDS-polyacrylamide gels) that constitutes about 25% of total intracellular protein (9). In uitro translation studies indicated another major protein that is highly glycosylated, DP2, and secreted in abundance by the dorsal prostate and coagulating glands (11).
In this report, an affinity-purified DP1 antibody was used to isolate DP1 cDNA clones from a library of coagulating gland cDNA recombinants in the bacteriophage X gtll expression vector. The polymerase chain reaction (PCR) and rapid amplification of cDNA ends method were used to complete the 5"terminal end of the coding and noncoding regions. Amino acid sequence deduced from the cDNA matched the sequence of cyanogen bromide fragments of the purified protein, indicating that the cDNA is authentic. Expression of the 3.2-kb DP1 mRNA species was tissue-specific and regulated by androgen. DNA sequence analysis indicates that rat prostate DP1 is a member of the transglutaminase family.

EXPERIMENTAL PROCEDURES
Antibody Production and Protein Sequence-DP1 was purified and used as immunogen to raise rabbit polyclonal antibodies as previously described (9). DP1 antibodies were affinity-purified by coupling DP1 (5 mg/ml gel) to  according to the manufacturer's specification (Pharmacia LKB Biotechnology Inc.). Anti-DP1 antibodies were eluted with 0.2 M glycine HCl, pH 2.5, containing 0.5 M NaCl and neutralized immediately using 1 M Tris.
To determine protein sequence, purified DP1 protein was reduced and carboxymethylated (12) and cleaved with cyanogen bromide (13). The cleavage products were separated using high performance liquid chromatography and sequenced by automated Edman degradation as previously described (14).
galactoside. The recombination frequency was 24%. The library was amplified in E. coli Y1088 to 9.5 X 10" pfu on agar plates, and phage were eluted in SM (0.1 M NaC1, 10 mM MgSO,, 0.1% gelatin, and 50 mM Tris, pH 7.5). The library was screened for recombinants as described by Young and Davis (19) using affinity-purified anti-DP1 antibody (0.6 mg/ml) a t a 200-fold dilution and 0.85% gelatin rather than fetal calf serum to block nonspecific binding sites. X phage were plated on lawns of E. coli Y1090 and incubated a t 42 "C for 1 h. Nitrocellulose filters saturated with 10 mM isopropyl P-D-thiogalactopyranoside were overlaid, and incubations were continued at 37 "C for 3 h. Filters were incubated with a 1/200 dilution of affinitypurified DP1 antibody, washed, and incubated with '"1-protein A. Positive clones were detected by autoradiography. 10; plaques on a 150-mm plate were screened, and positive plaques were selected and incubated overnight in SM. Phage from positive plaques were rescreened a t 300 plaques on a 90-mm plate. The third screen of selected plaques were all positive.
h DNA was prepared as described previously (17,20). Rat prostate DNA was isolated as described by Blin and Stafford (21). Northern and Southern blot analysis were performed as described previously (22-24) a t 42 "C in the presence of 50% formamide and 0.1% SDS. Filters were washed a t high stringency in 0.1 X SSC (1 X SSC: 15 mM NaCl, 1.5 mM sodium citrate, containing 0.1% SDS at 60 "C). T h e DNA sequences of both strands were fully determined using the chemical cleavage (25) and Sanger dideoxy (26) methods. Assignments of nucleotide numbers for DP1 cDNA clones and oligonucleotides were based on the designation of nucleotide 1 as the first residue at the 5' end of DP1 cDNA, determined by PCR, and verified by genomic DNA sequence analysis. Oligonucleotide primer sequences are listed in ascending (5'-3') or descending (3'-5') order. PCR products were cloned into M13 mp19, and both strands of a t least three clones were sequenced using the dideoxy method.

5'
End of DP1 cDNA-The method of rapid amplification of cDNA ends described by Ohara et al. (27) and Frohman (28) with some modifications was used to generate the 5' region of DP1 cDNA. First strand cDNA was prepared from coagulating gland poly(A) RNA, and a stretch of dG residues was added to the 3' end using terminal deoxynucleotidyltransferase. The first cycle of the 100 p1 PCR reaction containing oligonucleotide (oligo) A (40 bp) and B (433-395) (see Table I) was as follows: denaturation at 95 "C, 5 min; annealing at 55 "C, 1 min; extension a t 72 "C, 1 min. Oligo A contains 20 repeated deoxycytosine residues at its 3' end and three 5' restriction enzyme sites: HindIII, EcoRI, and BamHI (Table I). The PCR reaction was continued for 40 cycles: denaturation at 95 "C, 40 s; annealing a t 55 "c, 1 min; extension at 72 "C, 1 min plus 3 s/cycle. One-tenth of the reaction product (10 pl) was reamplified with oligonucleotide A and the nested oligo C (153-119) (see Table I) using the above PCR conditions.
Primer Extension Analysis-Primer extension was performed as described previously (29), with some modifications. Briefly, the 5' end "P-labeled oligonucleotides C, E, and F (Table I) were incubated with increasing amounts (0.01-1 pg) of coagulating gland poly (A) RNA in a solution containing 50% formamide, 0.4 M NaC1, 1 mM EDTA, and 40 mM PIPES, pH 6.4, at 85 "C for 10 min and placed a t 42.5 "C for 4-6 h. After ethanol precipitation, the oligonucleotide was extended with avian myeloblastosis virus reverse transcriptase (Promega Biotec, GIBCO-BRL) using DP1 mRNA as template. DNA products were subjected to electrophoresis on a 5% sequencing gel in parallel with dideoxy sequencing ladders of M13 mp19 size markers.

RESULTS
Identification of D P I cDNA-A cDNA expression library of 3.6 X lo5 pfu was prepared from rat coagulating gland poly(A) RNA using the X gtll cloning system and amplified. Screening of approximately 3 X IO6 pfu with affinity-purified anti-DP1 antibody yielded 27 positive clones. The clones were digested with EcoRI and analyzed on agarose gels; 15 were selected for further analysis based on their different sized inserts (not shown). The smallest insert from clone 15 was sequenced by the Maxam/Gilbert chemical cleavage procedure (25) and found to contain a 221-bp open reading frame that included the predicted amino acid sequence of DP1 peptide 7 (see Fig. 1). Identity between protein and predicted sequence suggested that the isolated DNA represented a partial sequence of DP1. Southern blot analysis of EcoRI digested DNA from each of the positive clones probed with a '"P- To further establish the identity of clone 11, the deduced amino acid sequence was compared with amino acid sequence determined directly from purified DP1.
The presence of a blocked NH, terminus necessitated chemical cleavage of DP1 with CNBr prior to sequence analysis. The sequences of nine CNBr fragments were obtained, five of which matched the deduced sequence of cDNA clone 11 (peptide fragments 4-8 shown underlined in Fig. 1). In each instance, the peptide sequence was preceded by a methionine residue in the predicted amino acid sequence of clone 11, in agreement with TABLE I Sequence of oligonucleotide primers used in PCR amplification of the 5' end of the DPI cDNA All sequences are listed in the 5' to 3' direction. Numbers at the right indicate the sequence position, where 1 is the mRNA initiation site. Numbers increasing are 5' to 3' and numbers decreasing are 5' to 3' of the opposite strand shown in Fig. 1. Oligonucleotides F and G were determined not to he DP1 sequence and are derived from the 5' end of clone 10. The sequence of oligonucleotide H was based on sequence obtained by PCR.       TO our surprise, one CNBr peptide sequence was encoded by the opposite strand at the 5' end of clone 11, indicating the possibility of a cloning inversion (shown correctly as peptide 1 in Fig. 1). The putative inversion contains sequences 135-17 (shown correctly as 17-135 in Fig. 1) contiguous with residue 378 of clone 11. Two additional cDNA clones, detected by rescreening the coagulating gland cDNA library with a 5' portion of clone 11 cDNA, terminated in the region of residue 378. These results suggested that secondary structure near residue 378 interfered with cDNA extension during library preparation. This hypothesis was later supported by the identification of a GC-rich sequence with complementarity between residues 12-18 and 378-384 (ouerlined in Fig. l). Furthermore, PCR analysis confirmed the inversion as discussed below. DNA inversions of this type at the 5' end of cDNA clones were reported previously (30-34). Determination of the 5' End of DPl cDNA-Since several peptide sequences were not identified in the deduced amino acid sequence and there was evidence for a 5' inversion, an additional cDNA clone 10 was isolated (Fig. 2) by rescreening the rat coagulating gland cDNA library using a 5' 270-bp HpaII fragment from clone 11. Clone 10 contained 249 bp within which two regions of amino acid sequence matched two as yet unidentified DP1 peptide fragment sequences, which are shown as peptides 1 and 2 (underlined in Fig. 1; diagrammed in Fig. 2). However, no overlapping sequence was detected that linked clones 10 and 11. Furthermore and as discussed below, the 5' 99 bp of clone 10 were not DP1 sequence.

C M C C T T G T M C T G T G T T A G C A C T G G G G T C T G G G T C T G C T T 2106 TGCCTGCTGAGCCATCTCCCCMCCTTCATTTTGATTTTTCAGGTCTGGGGATTGCMCTCATGGC 2112
PCR was used to determine the missing sequence between cDNA clones 10 and 11. Two oligonucleotide primers, oligo B (433-395) and oligo D (86-107), were prepared with sequence derived from clones 11 and 10, respectively (Fig. 2, Table I). An amplified fragment of 348 bp (shown schematically in Fig.  2) represented nucleotide residues 86-433 and had an open reading frame continuous with the flanking sequence of cDNA clones 10 and 11. Furthermore, the 139-bp gap sequence between clones 10 and 11 obtained by PCR extended from residues 239-378 and contained the sequence encoding peptide fragments 3 and 4 shown in Figs. 1 and 2.
The 5' terminal sequence of DP1 cDNA was obtained by the method of rapid amplification of cDNA ends. First strand cDNA template was extended with dG residues using terminal deoxynucleotidyltransferase to provide a substrate for priming with 5' oligo A (see "Experimental Procedures," Fig. 2, and Table I). In an initial PCR reaction using oligos A and B (433-395), no prominent band was detected by ethidium bromide staining of an agarose gel. The PCR reaction mixture was further amplified using oligos A and C (153-119, Fig. 2). A 193-bp fragment was obtained (shown schematically in Fig.  2) that contained an open reading frame but coded for no additional CNBr peptide sequences of DP1. The 193-bp fragment hybridized to the 3.2-kb DP1 mRNA, suggesting that it represented authentic DP1 sequence (data not shown). However, sequence comparisons revealed that the 5' 106 bp of the 193-bp PCR fragment differed from clone 10, yet 87 bp of 3' sequence matched the sequence in clone 10; the boundary of the two was a HinfI site in clone 10 (see Fig. 2).
To establish the correct 5' end of DP1, primer extension of the cDNA was performed using, as primers, oligos C (153-119), E (51-24), and F (54-34) (see Table I  7 and a), and with oligo E, a 51-bp fragment (Fig. 3, lanes 5 and 6). These results were in agreement with the 193-bp PCR amplification product reflecting authentic 5' sequence (153 residues + the 40-residue oligo A). Oligo F failed to yield an extended product (Fig. 3, lunes 1-3). Since the sequence of oligo F derives from the 5' region of clone 10, a cloning artifact must have arisen in this region perhaps by insertion of an EcoRI fragment at the HinfI site.
Further proof for the authenticity of DP1 5' sequence utilized an alternative PCR strategy. Oligo G (1-26 of clone 10, see Table I and Fig. 2) sequence was derived from the 5' end of clone 10 that was suspected of being a cloning insertion artifact. The sequence of oligo H-(1-23) was derived from the 5' end of the amplified 193-bp fragment thought to be the correct 5' DP1 sequence (Fig. 2). Each of these oligos was used in PCR paired with oligo C-(153-119) or B-(433-395, see Fig. 2). Only the oligo pairs H/C and H/B yielded the expected products of 153 and 433 bp, respectively. As predicted, the G/C and G/B oligo pairs failed to yield PCR amplification products, indicating that the 5' 99 bp of the 249-bp clone 10 was not authentic DP1 sequence and represented a second cloning artifact. Further evidence that the 5' region of clone 10 was not derived from DP1 came from Northern blot analysis using the 5' HinfI fragment as probe; it failed to hybridize with DP1 poly(A) RNA. The 3' HinfI fragment of clone 10 showed strong hybridization with DP1 mRNA (data not shown).
Preliminary analysis of genomic DNA indicates an intron of undetermined size positioned between nucleotide residues 84 and 85. A 5' genomic fragment was obtained, however, that allowed us to confirm the 5' cDNA sequence. Several additional large introns within the genomic sequence interrupted the coding sequence at residues 389,466, and 585. An additional 722 bp of 3"noncoding cDNA sequence (Fig. 1) were determined from the 2-kb insert of clone 25 (shown schematically in Fig. 2) making a 3000-bp cDNA.
Regulation of DPl mRNA-A 32P-labeled 270-bp HpaII 5' fragment of DP1 cDNA clone 11 was used as a hybridization probe to determine mRNA levels in the intact rat and under conditions of androgen withdrawal and restimulation. Northern blot analysis of poly(A) RNA isolated from dorsal prostate and coagulating gland revealed a prominent band of approximately 3.2 kb (Fig. 4A, lanes D and C), which was absent in poly(A)-depleted RNA of coagulating gland (Fig. 4.4, lane 2'). The similarity in band intensity of DP1 mRNA in coagulating gland and dorsal prostate is in agreement with earlier studies indicating that DP1 is a major product of these prostate regions. Poly(A) RNA of ventral prostate (Fig. 4A (lanes 2,5, and 8), 1 pg (lanes 3, 6, and 7). An M13 mp19 sequencing ladder (GACT) determined by the dideoxy method was used as a size marker. Arrows indicate the positions of the longest DNA products.

A.
B.

A, hybridizing bands are shown with poly(A) RNA from dorsal prostate (D) and coagulating gland (C) but not with poly(A)-deprived coagulating gland RNA (T), poly(A) RNA from ventral prostate ( V ) , or poly(A) RNA from the Dunning tumor (H). RNA samples (20 pg)
were glyoxylated and electrophoresed on a 1% agarose gel and transferred to a Biodyne nylon membrane. The membrane was hybridized to a '"P-labeled cDNA from the 5' end of clone 11, washed at high stringency, and exposed to x-ray film. B, total RNA was isolated from ventral ( V ) , dorsal (D), and lateral (L) lobes of prostate, coagulating gland (C), seminal vesicle (S), and epididymis (E). RNA samples were processed as described in A. lacked a signal for DP1 mRNA, as did poly(A) RNA isolated from the Dunning R3327H tumor (Fig. 4A, lane H ) . The Dunning tumor is thought to originate from the rat dorsal or vtaminase from Rat Prostate lateral prostate (35). Sublines of the Dunning tumor including the G, AT1, AT3, ML, MLL, HI, and HIF (36) also lacked DP1 mRNA (data not shown). Using total RNA, no DP1 mRNA signal was observed in ventral prostate or epididymis (Fig. 4B, lanes V and E ) , and weak signals were detected in lateral prostate and seminal vesicle RNA (Fig. 4B, lanes L and S, respectively). These latter signals likely resulted from a failure to completely dissect these glands free from dorsal prostate and coagulating glands.
Androgen regulation of DP1 mRNA was examined by isolating RNA from dorsal prostate of rats castrated and sacrificed at intervals up to 7 days. As shown in Fig. 5A, the high steady state level of the 3.2-kb DP1 mRNA in intact dorsal prostate decreased by 20% within 1 day and by 80% within 7 days after castration. Administration of testosterone propionate (2 mg intramuscularly) to animals 4 days after castration rapidly increased the level of DP1 mRNA (Fig. 5B). Within 4 h, the 3.2-kb mRNA species increased 2-fold, and within 8 h of androgen treatment, by 4-fold. After a slight decline between 12-16 h, DP1 mRNA continued to increase up to 72 h of androgen stimulation. The RNA band at approximately 2.1 kb is attributed to cross-hybridization with 18 S ribosomal RNA, based on its size and reactivity with other DNA probes.
The decrease in DP1 mRNA after castration could result from a change in mRNA turnover or from a decreased rate of transcription. To investigate this further, cycloheximide, a reagent known to stabilize rapidly turning over mRNAs (37), was administered to 350-g rats. Four days after castration, rats were either untreated or treated with testosterone propionate 30 h prior to sacrifice. Cycloheximide (20 mg intraperitoneally) injected 2 h before sacrifice increased the level of DP1 mRNA by 2.3-fold in untreated castrated rats but did not cause an increase after 30 h of androgen stimulation (Fig.  6). In a previous study under identical conditions, cycloheximide caused a 7-fold increase in the level of c-myc mRNA of ventral prostate in untreated castrated rats but only a 2-fold increase 30 h after androgen replacement (38). These results suggest that in the absence of androgen, DP1 mRNA degradation is more rapid, and like c-myc mRNA, it can be stabilized by cycloheximide. Androgen stimulation itself appears to stabilize DP1 mRNA and decrease its rate of degradation since cycloheximide did not increase DP1 mRNA after androgen treatment (Fig. 6).

One Gene Locus for DPI-In an attempt to establish if DP1
mRNA is encoded by a single gene, total genomic DNA was isolated from rat dorsal and ventral prostate and from the R3327H Dunning tumor. DNA was digested with several restriction enzymes known not to cleave within DP1 cDNA clone 11 (BamHI,BglII,PstI,PuuII,EcoRI,and SstI). Using the entire 2021-bp cDNA insert of clone 11 as a radiolabeled probe, a smear on the autoradiogram was observed, probably resulting from cross-hybridization of an Alu-like repetitive element a t nucleotides 2175-2251 within the 3"noncoding region of clone 11. Genomic DNA was then analyzed using the 5' HpaII fragment of clone 11. Digestion with each of several restriction enzymes (BamHI, BglII, and EcoRI) yielded a single band using DNA isolated from dorsal and ventral prostate and Dunning tumor (Fig. 7, A-C), suggesting the presence of one gene locus for DP1. Digestion with PstI, PuuII, or SstI yielded two or three bands, which likely resulted from enzyme cleavage within intron regions. One gene locus for DP1 is consistent with findings using an EcoRI partially digested rat genomic DNA library constructed in X Charon 4A and provided by T. Sargent 5. Effect of castration and androgen replacement on DP1 mRNA in dorsal prostate. A, dorsal prostate was removed from groups of rats that were intact (int) or castrated and sacrificed 1, 2, 4, and 7 days later. B, total RNA was isolated from dorsal prostate of intact rats and from rats that were administered testosterone propionate (2 mg intramuscularly) 4 days after castration and killed 0, 4,8,12,16,20,24,30,48, and 72 h later. In A and B, total RNA (20 pg) was analyzed by Northern blot hybridization as described in the legend to  BglII (lanes 2 ) , PstI (lanes 3 ) , PuuII (lanes 4 ) , EcoRI (lanes 5), and SstI (lanes 6). DNA was transferred to Biodyne nylon membranes for hybridization with a 32P-labeled DNA probe from the 5' portion of clone 11. terns when hybridized with radiolabeled DP1 cDNA (data not shown). Furthermore, the identical banding patterns observed with restriction enzyme digests of prostate and tumor DNA suggested that a gross rearrangement of the DP1 gene did not account for the lack of expression of DP1 in the Dunning R3327H tumor.
Sequence Comparisons with DP1-The DP1 coding sequence was compared with three types of transglutaminases (TG) (Fig. 8). Striking sequence similarity was observed be-tween DP1 and guinea pig TG, rat keratinocyte TG, and human clotting factor XIIIa (human placental factor XIII). A completely homologous active site required for TG activity was common to all four. These and additional properties, including similarity in PI (PI 6.0-6.12), molecular weight (71-77 kDa), potential calcium binding, and glycosylation sites, suggest that DP1 is rat prostate transglutaminase.

DISCUSSION
The full coding sequence for dorsal protein 1 (DPl), a major androgen-regulated protein of rat dorsal prostate and coagulating gland, was determined. The transcription initiation site was identified by primer extension and PCR methodology, and 3000 bp of cDNA were sequenced. The 2004-bp open reading frame encodes 668 amino acids and predicts a molecular weight of 75,479 for DP1. The amino acid sequence aligned with the sequence of eight cyanogen bromide fragments of purified DP1, authenticating the sequence. The DP1 mRNA, approximately 3.2 kb by Northern blot analysis in rat coagulating gland and dorsal prostate, was similar in size to the cDNA sequenced. Androgen regulation of DP1 was demonstrated by a decrease in mRNA subsequent to androgen withdrawal by castration and an increase following androgen administration. Amino acid sequence comparisons revealed that DP1 is a member of the transglutaminase family and most likely accounts for the TG activity previously described in rat prostate (40).
Our earlier studies indicated that DP1 is a dimer of two 60-70-kDa subunits (9). The homology among all cDNA clones selected with the DP1 polyclonal antibody suggested that DP1 is a homodimer. However, one of the nine cyanogen bromide cleavage fragments of DP1 protein did not match the amino acid sequence derived from DP1 cDNA. The unidentified 9-amino acid peptide may represent a contaminant of the DP1 protein preparation. Its sequence likewise showed no homology with other members of the TG family. Alternatively, DP1 may be heteromeric, with the DP1 antibody recognizing only one subunit. Examination of purified DP1 on a low percentage gradient SDS-polyacrylamide gel reveals two closely migrating bands spaced about 2 kDa apart with a smear in between. Since DP1 contains 12 potential glycosylation sites (double underlined in Fig. 1) and may become glycosylated (11), this disparity in banding pattern may simply reflect differences in glycosylation. Rat coagulating gland cytosol and secretory fluid are reported to contain at least two forms of TG (41). Thus, although the authenticity of the DP1 cDNA is established, the origin of the unidentified cyanogen bromide cleavage peptide remains unresolved.
The rate of decrease in DP1 mRNA after androgen withdrawal by castration was somewhat slower than that of the

F13a E I P -W --~V S E L O S G K -W G A K I~R E D R S~L S I O S~~K~I~G K F . -R~ 160
act to prolong DP1 mRNA half-life.

F C C G P C S V E S I K N G L V -Y n K~D T~F I F~E~s D K V Y -Y~G T L 5 3 4
(46) revealed a sequence similarity exceeding 80%. The human sequence also resembled guinea pig TG at this level (47).

GP I S T K S V G R D E R E D I T H T Y K Y P E G S E E E R E A F V R A N H L N K L A T . K~T G V M R I R V G Q N M 481
In contrast, the primary structure of human endothelial TG

DP1 I S T K U V G E N R R O D I T L H Y K F P E G S P E E R K M E K A S G~K R P D D -K L N S R~~T -L H I S V -L O N S V 465
shows only 35% similarity to human Factor XI11 and 32%

F13a IVTKOIGGDGHMDITDTYKFOEGOEEERLALETA G A K K P M K S R S N V D M D F EVENA 528
similarity to the rabbit tracheal epithelial cell TG described  TG) (46), and unofficially, a "type 111" (Factor XIIIa) (51).  . .

DP1 E -L G H P I N L T I V L K R K T A T P P -N V N l~-C S L D L O T Y T G N K K T~L G V~O K T V~I O G~E -S E V S L 524
type TG, and type TG represent three distinct

GP O N R K L I A E -V S L K N P L P -W L L G C I F T V E G A G L T K D O K S~V P D P V E -A C E O A K~V D L L P T E 662
TGs) (52, 53). Since DP1 resembles other TGs at the 30% DP1 VNOPLTIT-CNFKNTLP-IPLTNIKFSVESLGLNN-MKSWE-QETVP-PGKTINFOIECTP~ 6 3 9 level, it appears that it represents a fourth, hithedo not

824
suggests that DP1 is the prostate TG originally referred to as vesiculase (40).

FIG. 8. Amino acid sequence comparisons between DP1 (rat
The rat coagulating gland was, in fact, notable for its keratinocyte TG (RK) (49). Double dots, identical amino acid lent cross-linking Of seminal Vesicle proteins through the residues relative to DP1; single dots, homologous amino acid residues formation of t-(y-g1utamyl)lysine bonds between polypeptide relative to DP1. Gaps in the sequence (indicated by a blank for more chains (41,54,55). This prostate enzyme, unique to coagulatthan 2 amino acids or a hyphen for a single amino acid) were inserted ing gland and dorsal prostate, has an important role in forto achieve maximal sequence alignment. Amino acids at the active are underlined.
One of the major rat seminal vesicle secretory proteins, SVIV, was shown to act as an acyl donor and acceptor substrate for prostatein C3 subunit, the major androgen-regulated secretory TGs, both from rat coagulating gland and guinea Pig liver* protein of rat ventral prostate (1)(2)(3)(4)(5)(6). prostatein c 3 m~~~ The formation of SVIV products of several different molecular decreases to low levels 4-7 days after androgen withdrawal, weights suggested that TGs catalyzed both intra-and interwhereas substantial amounts of DP1 mRNA (4-5-fold higher) molecular cross-linking (57). Other modified forms of SVIV remain after this withdrawal period (38, 42-44). Administra-were produced by TG in the presence of spermidine, an 47), the a subunit of human factor XI11 (3'13a) (51), and rat secretion Of a Ca2+-dependent TG, which the cavasite with cysteine at its center are boxed. Potential Ca2+-binding sites mation Of the postejaculatory vagina1 Plug in rodents (56)' abundant polyamine of seminal plasma. Rodents are unique among mammals by forming a copulatory plug. Since the TG activity is required to stabilize the plug, it is conceivable that the prostate-specific TG evolved in rodents for this purpose.