Old genes and new genes: the evolution of the kallikrein locus.

The human kallikrein locus consists of KLK1, the gene of major tissue kallikrein, and 14 genes of kallikrein-related peptidases (KLKs) located in tandem on chromosome 19q13.3-13.4. In this review, based on information retrieved from the literature or extracted from genome databases, it is hypothesised that the kallikrein locus is unique to mammals. The majority of genes are highly conserved, as demonstrated by the identification of 11 KLK genes in the opossum, a metatherian species. In contrast, a sublocus, encompassing KLK1-4, has gone through major transformations that have generated new genes, which in most cases are closely related to KLK1. In the primate lineage, this process created KLK3, the gene of the prostate cancer marker, prostate-specific antigen (PSA), whereas in the murine lineage it gave rise to 13 genes unique to the mouse and nine unique to the rat. The KLK proteases are effector molecules that emerged early in mammalian evolution and their importance in skin homeostasis and male reproductive function is undisputed and there are also accumulating evidence for a role of KLK proteases in the development of the brain. It is speculated that the KLK gene family arose as part of the process that generated distinguishing mammalian features, like skin with hair and sweat glands, and specialised anatomical attributes of the brain and the reproductive tract.

Kallikrein-related peptidases.

Kallikrein 1 (KLK1), a key component of the kallikrein-kinin system, originates from a locus on the long arm of chromosome 19 that contains several related serine endopeptidases. The biological role of these kallikrein-related peptidases is not clear, but emerging evidence suggests that they might be important in several physiological systems, e.g., in male reproduction, skin homeostasis, tooth enamel formation and neural development and plasticity. The kallikrein locus has undergone some major evolutionary events. Most spectacular are relatively recent duplications of KLK1 that have created 13 and 9 functional genes that are unique to the mouse and the rat, respectively. Human paralogs are KLK2 and KLK3: the latter encoding the cancer biomarker prostate-specific antigen. In this review on kallikrein-related peptidases, the focus is on their evolution, their role in skin homeostasis and semen liquefaction, and their utility as cancer biomarkers.

The semenogelins: proteins with functions beyond reproduction?

The coagulum proteins of human semen, semenogelins I and II, are secreted in abundance by the seminal vesicles. Their function in reproduction is poorly understood as they are rapidly degraded in ejaculated semen. However, more recent results indicate that it is time to put the semenogelins in a broader physiological perspective that goes beyond reproduction and fertility.

Semenogelin II gene is replaced by a truncated line 1 repeat in the cotton-top tamarin.

The human seminal vesicles secrete two proteins, semenogelin I and semenogelin II, at very high concentrations. It has previously been shown that the cotton-top tamarin (Sanguinus oedipus), a New World monkey, is lacking the semenogelin II gene. We have now determined the nucleotide sequence of DNA located 5--13 kilobases (kb) downstream of the tamarin semenogelin I gene---a region that in man is occupied by the semenogelin II gene. Two regions with homology to the human semenogelin II gene were identified in the tamarin DNA. The first region, of 3.5 kb, is homologous to DNA upstream of the human gene, and the second region, of 0.6 kb, is mainly derived from the second intron. Between these regions, equivalent to 594 base pairs (bp) upstream of the transcription initiation site to 12 bp downstream of the stop codon in the human semenogelin II gene, the cotton-top tamarin DNA carries a truncated LINE1 repeat. In another set of experiments, the tamarin DNA hybridizing to the mouse semenoclotin gene was investigated. It was concluded that hybridization is with the second intron of the semenoclotin gene, but very likely, the material does not represent a cotton-top tamarin semenoclotin gene. Thus, a mammalian ancestor probably carried a single gene that in the rodent lineage developed into the semenoclotin gene and in the primate lineage into a progenitor of the semenogelin genes.

Production and activation of recombinant hK2 with propeptide mutations resulting in high expression levels.

Human glandular kallikrein 2 (hK2) is a serine protease expressed mainly by the prostate gland with 80% identity in primary structure to prostate specific antigen (PSA). hK2 has proven to be a useful marker of prostate cancer which can be used in combination with PSA to better discriminate between prostate cancer and benign prostate hyperplasia. The studies on hK2 have been hampered by its very low phyciological levels (6 microgram.mL-1), its close similarity to PSA, and the low expression levels obtained using recombinant procedures to produce hK2 (0.7 mg.L-1). We have now generated propeptide mutations of hK2 which can be used to isolate stable, inactive prohK2 mutants. Compared with wild-type hK2, expression of the propeptide hK2 mutants increases the expression levels up to 15-40-fold giving 10-30 mg hK2.L-1. These results indicate that the low expression levels of wild-type hK2 are related to the activation or autoactivation of the wild-type enzyme and the instability of the active protease in cell culture and possibly also in tissue. The purified mutant hK2 may be activated by either enterokinase or factor Xa to generate an enzyme for use in functional studies with the characteristics of the original wild-type protein. Further, the stable inactive mutant hK2 protein may be used for immunizations to generate novel monoclonal antibodies, used as standard material for clinical assays or in crystallization studies where large quantities of protein are required.

New world, but not Old World, monkeys carry several genes encoding beta-microseminoprotein.

It was shown by Southern hybridization that cotton-top tamarin and common marmoset, New World monkeys, carry three or more genes encoding beta-microseminoprotein, also known as PSP94. In contrast, the genomes of Old World monkeys, as represented by rhesus macaque and sacred baboon, contain a single gene. Clones containing three different genes encoding beta-microseminoprotein were isolated from a cotton-top tamarin genomic library. They carry two complete genes of four exons and a third gene lacking the first exon. The structure suggests that the three genes are functionally active and give rise to transcripts that are approximately 86% similar in sequence. By sequencing one gene in full, it was shown that the introns carry an excess of interspersed repeats, on average 29% of the introns consist of Alu repeats. A phylogenetic analysis demonstrated that the genes probably arose in New World monkeys after the separation from Old World primates.

Quantitative reverse transcription-PCR assay with an internal standard for the detection of prostate-specific antigen mRNA.

BACKGROUND: Circulating prostate cells can be detected with a reverse transcription-PCR (RT-PCR) assay for prostate-specific antigen (PSA) mRNA. We have developed a new quantitative RT-PCR method for measuring PSA mRNA. METHODS: The method uses a PSA-like internal standard (IS) mRNA that is added into the sample at the beginning of the RNA extraction and coamplified by RT-PCR with the PSA in the sample. After PCR amplification, the IS and PSA products are selectively detected by hybridization in a microtitration plate using probes labeled with fluorescent europium chelates. RESULTS: The method was validated with PSA and IS mRNAs and PSA-expressing cells to obtain a detection limit of 50 PSA mRNA copies (i.e., signal 2 times the mean of zero signal), linearity up to 10(6) copies, and detection of a single PSA-expressing cell. In preliminary evaluations, 60% (n = 10) of the prostate cancer patients with skeletal metastases gave results above the detection limit (500 PSA mRNA copies in 5 mL of blood). The total number of PSA copies ranged from 900 +/- 200 to 44 100 +/- 4900 (mean +/- SD) in the samples, corresponding to approximately 1-100 PSA-expressing cells in 5 mL of blood. In the controls (n = 34), none of the healthy females and 2 of 19 healthy males had detectable PSA mRNA [700 +/- 100 and 2000 +/- 900 (mean +/- SD) PSA mRNA copies in 5 mL of blood for the 2 males]. CONCLUSIONS: The assay provides sensitive and quantitative detection of PSA mRNA expression from blood samples and can be used to establish the clinically significant number of PSA mRNA copies in prostate cancer.

Measurement of prostate-specific antigen and human glandular kallikrein 2 in different body fluids.

It has been demonstrated that prostate-specific antigen (PSA), in spite of its name, can be detected in body fluids and tumors from a variety of organs. Investigations have shown that human glandular kallikrein 2 (hK2), a related prostate-secreted protease, can activate the zymogen form of PSA, suggesting that the two enzymes might work as a functional unit, with hK2 as the activator molecule and PSA as the effector molecule. If this is true, then hK2 should be found together with PSA in body fluids other than seminal plasma, as well. Recently, a sensitive and specific assay was devised for hK2, enabling its measurement in picogram quantities. With this assay, the concentration of hK2 was determined in samples of seminal plasma, amniotic fluid, breast milk, and saliva. Simultaneously, the samples were assayed for molecular forms of PSA. In seminal plasma, the mean PSA concentration was 0.82 mg/ml, while the hK2 level was around two orders of magnitude lower: mean value, 6.4 microg/ml. Approximately the same ratio of PSA to hK2 as in seminal plasma was found in amniotic fluid and breast milk, but in most samples, the hK2 values were too low for direct measurements and had to be concentrated prior to analysis. Measurable levels of PSA, all in the free form, were detected in amniotic fluid at the thirteenth week of gestation and then gradually increased to levels around and over 1 microg/L from the twentieth week. Significant levels of PSA were detected in amniotic fluid collected at delivery, also. Measurable levels of mammary PSA were primarily detected in colostrum, with a range from less than 0.03 microg/L to 2.1 mg/L. Around half of the molecules were in complex with protease inhibitor. Most surprisingly, determinations on saliva samples showed that none of them had detectable PSA levels but had measurable concentrations of hK2 with a mean value, 0.09 microg/L. The presence in saliva suggests that hK2 can be the human equivalent to one of the mouse salivary kallikreins with important biological function, like the epidermal growth factor-binding protein or the gamma subunit of nerve growth factor. However, this was ruled out, as a phylogenetic analysis showed that the human and mouse glandular kallikreins evolved independently from a common precursor after the separation of the primate and rodent lineages. In conclusion, the measurements show that in addition to the previously known secretion in seminal plasma, hK2 is secreted in amniotic fluid, breast milk, and saliva. Furthermore, the concerted expression of PSA and hK2 in seminal plasma, amniotic fluid, and breast milk suggests that the two proteases might form a functional unit but not always as demonstrated by the sole presence of hK2 in saliva.

The cotton-top tamarin carries an extended semenogelin I gene but no semenogelin II gene.

Previous studies have shown that the predominant proteins secreted by the seminal vesicles are transglutaminase substrates which have undergone major structural alterations during evolution. In man, they are known as semenogelin I and II; recently it was shown that, similar to man, several new world and old world monkeys carry two semenogelin genes as well, the exception being the cotton-top tamarin (Saguinus oedipus) with a single gene. This gene has now been cloned and identified as a semenogelin I gene, because of a higher number of conserved nucleotides in the human semenogelin I gene (89%) than in the human and the rhesus monkey semenogelin II genes (82%). Furthermore, the difference in sequence similarity indicates that the semenogelin II gene was deleted from the genome of a progenitor to the cotton-top tamarin after the duplication that yielded the two semenogelin genes seen in man. Like several other genes expressing seminal-vesicle-secreted transglutaminase substrates, the cotton-top tamarin semenogelin I gene consists of three exons of 97, 1816 and 146 bp. It codes for a signal peptide of 23 amino acid residues and the secreted protein of 592 amino acid residues. The molecular mass of 66 kDa is 32% larger than that of the human counterpart and, contrary to human semenogelin I, the cotton-top tamarin protein has the potential to be highly glycosylated as there are 14 sites with the consensus sequence for N-linked glycosylation. Approximately half of the primary structure consists of five nearly identical tandem repeats of 58 amino acid residues, that probably evolved relatively late.

Semenogelin I and semenogelin II, the major gel-forming proteins in human semen, are substrates for transglutaminase.

The major seminal vesicle secreted proteins in human semen, semenogelin I and semenogelin II, interact non-covalently and via disulphide bridges to instantly form a coagulum upon ejaculation. The coagulum is liquefied after a few minutes due to the action of a prostatic serine protease, prostate-specific antigen (PSA). In contrast to rat semen, which forms an insoluble plug within minutes of expulsion, no transglutaminase-mediated cross-linking has been demonstrated in ejaculated human semen. However, we here show that semenogelin I and semenogelin II, both in seminal vesicle fluid and purified from semen, are substrates for factor XIIIa, the fibrin cross-linking transglutaminase. The cross-linking of the semenogelins, which was conformation-dependent, and the incorporation of a fluorescence-labelled amine, were visualised by SDS/PAGE and Western blot. Purified semenogelin I and semenogelin II could be cross-linked separately into complexes. Moreover, digestion of semenogelin with PSA produced fragments, some of which were cross-linked into complexes by factor XIIIa. We also found that PSA was unable to release any semenogelin fragments during exposure of the high molecular-mass complexes of cross-linked semenogelin to active PSA.

Chemical characterization of the predominant proteins secreted by mouse seminal vesicles.

Mouse seminal vesicles secrete four major protein components with estimated molecular masses of 95, 38, 17, and 16 kDa. Amino acid sequencing revealed that the 95-kDa component represents a protein with an unknown structure, while the 38-kDa component was identified as semenoclotin, the 17-kDa component as seminal-vesicle-secreted protein IV, and the 16-kDa component as seminal-vesicle-secreted protein V. Semenoclotin and the 95-kDa component were readily cross-linked by transglutaminase, suggesting that the two proteins are involved in the formation of the mouse copulatory plug. Treatment of mouse seminal vesicle fluid with human prostate-specific antigen rapidly degraded semenoclotin, indicating a structural resemblance of this protein to human semenogelins, despite the vast difference in primary structure. As previously reported for other seminal-vesicle-secreted proteins, the semenoclotin transcripts are shown to be under androgen control.

Cloning of the semenogelin II gene of the rhesus monkey. Duplications of 360 bp extend the coding region in man, rhesus monkey and baboon.

The semenogelin II gene from the rhesus monkey has been cloned and characterized. The transcription unit is split into three exons of 97, 2086 and 124 bp, with two intervening introns of 241 bp and 862 bp. The first exon codes for a 23-amino-acid signal peptide and the two amino-terminal residues of the secreted protein. The second exon codes for the rest of the mature protein, and the third exon contains non-coding nucleotides only. Secreted rhesus monkey semenogelin II consists of 683 amino acid residues, has a calculated M(r) of 77362, is devoid of Cys and Met, and displays a highly repetitive structure composed of ten 60-amino-acid repeats. Hybridization with genomic DNA showed that the semenogelin II gene of man, rhesus monkey and baboon has evolved through extension of the coding region with 360-bp segments. In contrast, the length of the semenogelin I gene of these species appears to be conserved. The two genes are also present in some New World monkeys, as was revealed by hybridization with genomic DNA from the marmoset. However, another New World monkey, the cotton-top tamarin, carries only one semenogelin gene, but also has a gene that is similar to the mouse semenoclotin gene.

Immunofluorometric assay for sensitive and specific measurement of human prostatic glandular kallikrein (hK2) in serum.

Prostate-specific antigen (PSA) and human prostatic glandular kallikrein (hK2) have 79% identity with the primary structure. When we used recombinant hK2 protein, only 7 of 23 monoclonal anti-PSA IgGs (monoclonal antibodies, MAbs) cross-reacted with hK2, which enabled us to design a novel immunofluorometric MAb-MAb assay for the specific detection of hK2. In the first incubation, an excess of MAb 2H11, which does not cross-react with hK2, is added to prevent both free and complexed PSA from reacting in subsequent immunoreactions. In the second incubation, biotinylated MAb H50, which cross-reacts with hK2 by an epitope overlapping with MAb 2H11, served to bind only hK2 to the microtitration wells coated with streptavidin. In the third step, Eu-labeled MAb H117, which cross-reacts with hK2, detected the immobilized hK2. The hK2 assay was calibrated with recombinant hK2. The detection limit of the assay was 0.1 microgram/L, and the cross-reactivity with recombinant PSA was < or = 0.7%. The concentration of hK2 was measured in serum samples from 334 males with total PSA concentrations ranging from 1 to 3400 microgram/L. Most of the samples (57%) had hK2 concentrations below the detection limit. The proportions of hK2 relative to total PSA were 0-2% in 79%, 2-5% in 14%, 5-10% in 4%, and >10% in 3% of the samples. Gel filtration of 10 serum samples with increased hK2 concentrations showed a single peak of hK2 immunoreactivity with an apparent molecular size of approximately 30 kDa, corresponding to that of recombinant hK2 and free PSA.

The gene of the protease inhibitor SKALP/elafin is a member of the REST gene family.

Members of the REST gene family characteristically have a transcription unit consisting of three exons. The first and the last exon are conserved among members, while the second exon--encoding almost all of the mature protein--differs considerably. The so far known REST genes are highly, and almost exclusively, expressed in the seminal vesicles. By sequence analysis we have now identified the gene for the protease inhibitor SKALP/elafin as a new member of the REST gene family. The protein is expressed in the epidermis and serves, like the product of several REST genes, as substrate for transglutaminase. We have also found what seems to be a locus encompassing both transglutaminases and REST genes centered around the region q12 on the human chromosome 20, raising the question whether the enzymes and substrates have evolved in parallel.

The structure of the semenogelin gene locus--nucleotide sequence of the intergenic and the flanking DNA.

The sequence of 15.7 kb from the human semenogelin gene locus has been determined. Together with previously published sequences, this gives the structure of a 28-kb region encompassing the two semenogelin genes. The two transcription units are separated by 11616 bp intergenic DNA comprising more than 40% repetitive DNA sequences, predominantly located to a 4-kb block of L1 and Alu repeats. Two more blocks of L1 sequences are present in the DNA flanking the genes, so that approximately 20% of the completed sequence consists of long interspersed repeated sequences, so called LINES. A comparison of the SgI gene and the SgII gene suggests that they evolved by the duplication of an approximately 8 kb DNA segment about 61 million years ago, probably by a mechanism involving recombination between L1 elements.

Distribution and tissue expression of semenogelin I and II in man as demonstrated by in situ hybridization and immunocytochemistry.

Semenogelin I and II (Sgl, Sgll) are two separate gene products of chromosome 20 with extensive (80%) identity in primary structure. They are mainly responsible for immediate gel formation of freshly ejaculated semen. Degradation of Sgl and Sgll is due to the proteolytic action of prostate-specific antigen (PSA); it results within 5-15 minutes in liquefaction of semen and release of progressively motile spermatozoa. By means of cDNA cloning and Northern blots, Sgl and Sgll transcripts have previously been shown to be abundant in human seminal vesicles, but Sgll alone is suggested to be expressed at low levels in the epididymis. To characterize the expression and tissue distribution of Sgl and Sgll in greater detail, we produced monoclonal immunoglobulin Gs (lgGs for immunocytochemistry (lCC) and specific [35S]-, digoxigenin-, or alkaline phosphatase-labeled 30-mer antisense probes to Sgl and Sgll for in situ hybridization (lSH). Immunocytochemical staining for both Sgl and Sgll, and lSH detection of both Sgl and Sgll transcripts, were demonstrated in the cytoplasm of seminal vesicle epithelium. lSH showed Sgll alone to be expressed in the epithelium of the epididymal cauda. Neither lCC nor lSH yielded any evidence of Sgl or Sgll expression in caput or corpus epithelium or in any stromal cells of the epididymis. Consistent with our previous findings using polyclonal lgG, monoclonal anti-Sgll Sgll lgGs identified epitopes on the posterior head, midpiece, and tail of ejaculated spermatozoa. Spermatozoa in the epididymal cauda were also immunoreactive, but those in the caput or corpus region of the epididymis as well as those in the testis were negative. As shown by lCC, neither Sgl nor Sgll were expressed in the testis, the prostate, the female genital tract, or other normal human tissue specimens. Although the significance of Sg attachment to epididymal and ejaculated spermatozoa remains to be established, monoclonal anti-Sg lgG might prove useful in establishing the origin of seminal vesicle tissue components in prostate core biopsies or other biopsy specimens.

The cloning of a rapidly evolving seminal-vesicle-transcribed gene encoding the major clot-forming protein of mouse semen.

Approximately 30 kb of the mouse genome, containing the gene for a major seminal vesicle transcript, has been cloned. The gene was identified by the similarity to members of a family with rapidly evolving genes that includes the gene encoding the major clot protein in rat semen, SVS II, and the human semenogelin genes. The nucleotide sequence of 16.9 kb was determined; this sequence encompasses the gene of 2215 bp plus 9-kb and 5.6-kb regions flanking the 5' and 3' ends of the gene. The transcription unit is divided into three exons, of which the first encodes the signal peptide, the second the secreted protein, while the third exon contains 3'-nontranslated nucleotides only. The transcript encodes a protein of 375 amino acid residues, including a signal peptide of 22 residues. The secreted polypeptide is a protein of Mr 38442 and is similar in sequence but smaller than the major clot-forming protein of rat semen, SVS II. It is highly charged at pH 7 and it has an isoelectric point of 10.68. The central part of the protein consists of tandem repeats that might serve as a substrate for transglutaminase.

A novel gene family encoding proteins with highly differing structure because of a rapidly evolving exon.

Despite vast differences in primary structure, it is here shown that several predominant semen proteins are encoded by genes that belongs to a common family. Members have their transcription unit split into three exons: the first encoding the signal peptide, the second the secreted protein, while the third exon solely consists of 3' non-translated nucleotides. The first and the third exon are conserved between members, but the second exon is not. The genes for human semenogelins I and II, rat SVSII, SVSIV, SVSV and guinea pig GP1 and GP2 belong to this gene family.

Production of recombinant PSA and HK2 and analysis of their immunologic cross-reactivity.

Measurements of prostate-specific antigen (PSA) in serum are widely used to monitor patients with prostate cancer, but the attenuation of the assay response by PSA complexed to protease inhibitors has been shown to affect the results in certain assay designs. Moreover, the human glandular kallikrein-2 (hK2), a kallikrein-like serine protease that is 80% similar to PSA, might interfere with the specific detection of PSA by immunological cross-reactivity. We have expressed hK2 and PSA in eucaryotic cells using the Semliki Forest Virus expression system and studied the reactivity of 18 monoclonal anti-PSA IgGs. Five of them cross-reacted with identical affinities to recombinant hK2 whereas 13 recognized PSA alone. The antibodies that recognized both PSA and hK2 bind to a region of the protein that is exposed when PSA is complexed to alpha-1-antichymotrypsin.