| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
research-article |
University Hospital MAS,3 Division of Clinical Chemistry, Department of Laboratory Medicine, Lund University, S-205 02 Malmö, Sweden
Division of Evolutionary Molecular Systematics,4 Department of Cell and Organism Biology, Lund University, S-223 62 Lund, Sweden
Department of Reproductive Biology,5 German Primate Center, Göttingen D-37077, Germany
ABSTRACT
The semen coagulum proteins have undergone substantial structural changes during evolution. In primates, these seminal vesicle-secreted proteins are known as semenogelin I (SEMG1) and semenogelin II (SEMG2). Previous studies on the common marmoset (Callithrix jacchus) showed that ejaculated semen from this New World monkey contains semenogelin, but it remained unclear whether it carries both genes or only SEMG1 and no SEMG2, like the closely related cotton-top tamarin (Saguinus oedipus). In this study we show that there are two genes, both expressed in the seminal vesicles. Surprisingly, the genes show an almost perfect sequence identity in a region of 1.25 kb, encompassing nearly half of the genes and containing exon 1, intron 1, and the first 0.9 kb of exon 2. The underlying molecular mechanism is most likely gene conversion, and a phylogenetic analysis suggests that SEMG1 is the most probable donor gene. The marmoset SEMG1 in this report differs from a previously reported cDNA by a lack of nucleotides encoding one repeat of 60 amino acids, suggesting that marmoset SEMG1 displays allelic size variation. This is similar to what was recently demonstrated in humans, but in marmosets the polymorphism was generated by a repeat duplication, whereas in humans it was a deletion. Together, these studies shed new light on the evolution of semenogelins and the mechanisms that have generated the structural diversity of semen coagulum proteins.
epididymis, male reproductive tract, seminal vesicles, sperm, sperm motility and transport
Semen is composed of spermatozoa and seminal plasma originating from accessory sex glands. Around 95% of the ejaculated semen volume is contributed by the seminal vesicles and the prostate. The seminal vesicles of most mammals secrete large amounts of a few proteins that are involved in the formation of a seminal coagulum. The major protein components of the human seminal coagulum are semenogelin I (SEMG1) and semenogelin II (SEMG2), which have molecular masses of 50 kDa and 63 kDa, respectively [1, 2]. Much of these proteins' primary structure, which shows 79% conservation, consists of redundant tandem repeats of 60 amino acid residues. Based on the degree of conservation, the repeats have been divided into three groups in which the type I repeat has the highest number of conserved amino acid residues [2]. The size difference between SEMG1 and SEMG2 in humans is due to two type I repeats, more in the latter. The SEMG1 molecules carry no sugar moiety, but approximately half of the SEMG2 molecules are glycosylated at a single Asn [3].
The genes SEMG1 and SEMG2, located 11.6 kb apart on human chromosome 20q12–13.1, were formed by the duplication of an 8-kb region [4, 5]. They belong to a gene family encoding rapidly evolving substrates of transglutaminase (REST), which also includes the genes of the protease inhibitor elafin, the murine SVS proteins, and the SVPs of the guinea pig [6, 7]. REST genes typically consist of three exons, of which the first codes for the signal peptide, the second codes for most of the mature protein and also carries the stop codon and a few 3' untranslated nucleotides, and the third exon has only 3' untranslated nucleotides and contains the polyadenylation signal. The rapidly evolving second exon gives rise to a polypeptide that is a good substrate for transglutaminase and may therefore be modified by polyamine adducts or peptide chain cross-linking.
Structural studies of the semenogelin genes in higher primates show an unusual evolution that is characterized by species-specific expansion of the second exon with repeats encoding either 60 or close to 60 amino acid residues [8–10]. There also are reports on truncated molecules due to the introduction of one or more premature stop codons [8, 11]. It has been proposed that the size variation of SEMG1 among hominoids is the result of a positive selection that is due to the degree of polygamy in different species, such that a typical polygyneous species like the gorilla produces SEMG1 of a smaller size than a polyandrous species like the chimpanzee [8]. Furthermore, it also has been claimed that the rate of molecular evolution of SEMG2 correlates with the levels of female promiscuity, in that species with promiscuous females tend to have an accelerated evolution compared with those with monogamous females [12].
In addition to the interspecies size variation of semenogelins, there also are reports on size variation within species. Recent studies demonstrated a truncated form of human SEMG1 that is present at an allele frequency of 2%–3% in both European and Asian populations. It codes for a variant SEMG1 that lacks a tandem repeat of 60 amino acid residues [8, 13, 14].
Duplicated genes and longer repetitive sequences are distributed throughout the genome of more complex organisms. The similarity between related nucleotide sequences can promote recombination, including gene conversion and reciprocal exchange. Gene conversion is the nonreciprocal transfer of genetic information between two genes or DNA sequences. Multiple factors influence the recombination frequency, including degree of similarity, length of the sequence, and the length of perfect, uninterrupted sequence identity [15].
Studies of New World monkeys show that the cotton-top tamarin (Saguinus oedipus) SEMG1 codes for a secreted protein of 66 kDa, containing five very similar tandem repeats of 58 amino acid residues. Surprisingly, analysis of DNA sequences downstream of this gene revealed that the cotton-top tamarin SEMG2 has been deleted and replaced by a truncated LINE1 repeat [16]. In contrast, investigation of the closely related common marmoset by Southern blot hybridization has shown that this species probably has two semenogelin genes [10]. The partial sequence also has been reported of a presumably unspliced transcript of marmoset SEMG1 [17].
In a recent study we found that the predominating protein component in the common marmoset seminal plasma is heterogeneous, protease resistant, and weakly reactive with an antibody raised against human SEMG1 that also cross-reacts with human SEMG2 [18]. We also found that both the heterogeneity and protease resistance is due to glycosylation. However, the investigation gave no clear answer as to whether the common marmoset has two functional semenogelin genes or only one, as in the closely related cotton-top tamarin. In this study we address this issue by analyzing the structure of the marmoset semenogelin genes and their transcription in the seminal vesicles.
All procedures carried out for this study were in accordance with German Animal protection law and were reviewed and approved by the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit TVA Az: 509.42502/08–01.03.
The extraction of genomic DNA from the liver of a common marmoset has been described previously [10]. Additional DNA samples were isolated from common marmoset spermatozoa. In this case, semen samples were collected by penile vibrostimulation at the German Primate Centre (Göttingen, Germany) and then stored frozen. The spermatozoa in thawed samples were pelleted in a microcentrifuge and washed twice with 200 µl of 0.9% NaCl, with subsequent centrifugation. The spermatozoa were lysed in 500 µl of 10 mM Tris, pH 8; 10 mM EDTA; 100 mM NaCl; 1% SDS; and 0.5 mg/ml proteinase K by incubation at 56°C for for 2 h until readily dissolved. The DNA was purified using conventional phenol/chloroform extraction and ethanol precipitation [19]. An RNA sample was isolated from the seminal vesicles of a deceased common marmoset at the German Primate Centre. The glands were removed, frozen in liquid nitrogen, and stored in dry ice until processed. Frozen seminal vesicle tissue (0.15 g) was cut in smaller pieces and transferred to a test tube containing 2 ml Trizol reagent (Invitrogen AB, Stockholm, Sweden) and homogenized using an ultratron. The RNA was isolated according to the standard Trizol protocol provided by the supplier of the reagent.
Synthesis and Sequencing of Marmoset SEMG1 and SEMG2
DNA fragments flanking the semenogelin genes were generated by genome walking [20]. Four different genome walking libraries were constructed from common marmoset DNA using the GenomeWalker Universal kit (BD Biosciences Clontech, Stockholm, Sweden). The genomic DNA was restricted with the enzymes DraI, EcoRV, PvuII, and StuI, and was ligated to the GenomeWalker adaptor DNA. Gene-flanking DNA was generated by nested PCR using the Adaptor Primer 1 (AP1) and Nested Adaptor Primer 2 (AP2) provided with the kit in combination with gene-specific primers. Some of the primer sequences were based on DNA sequences in other species and did not always match the marmoset sequences, but primed successfully. The primers 1–8 (Table 1) are specific for the common marmoset. The PCR was done with Advantage 2 (BD Biosciences Clontech) in the reaction buffer recommended by the supplier. The PCR protocol consisted of an initial 1-min incubation at 94°C followed by 7 cycles of 25 sec at 94°C and 4 min at 72°C, then 32 cycles of 25 sec at 94°C and 4 min at 67°C, and finally 7 min of extension at 67°C. The reactions were run in a PTC-200 DNA Engine (MJ Research, Watertown, MA).
|
Genomic fragments encompassing whole or part of the semenogelin genes were generated by PCR using primers 9–12 (Table 1) synthesized on the basis of sequences obtained from genome walking fragments. Discrimination of allelic forms of SEMG1 was done with PCR using primers 13–14 (Table 1). The PCR was done with Advantage 2 according to the standard protocol provided by the manufacturer, with the extension time adjusted according to the expected size of the product—1 min of extension time was set aside for each 1 kb of template to copy. The PCR products were separated by electrophoresis on agarose gels, stained with ethidium bromide, and visualized in ultraviolet light. DNA fragments were purified with JetQuick (Genomed, Bad Oeyenhausen, Germany) and then sequenced. The sequencing reactions were done with the BIG Dye terminator cycle sequencing kit and then run on an Applied Biosystems 3730 automated DNA sequencer (Applied Biosystems, Stockholm, Sweden). Sequences from the ends of DNA fragments were gradually extended by reactions primed by 20-mer oligonucleotides, which were designed from the sequences generated in the previous step. The sequences were assembled and analyzed using the GCG computer program package (Genetic Computer Group, Madison, WI).
Multiple sequence alignments were made using the ClustalW (1.82) program and then manually inspected in the Se-Al v2.0a11 software [21, 22]. Phylogenetic relationships were analyzed using the PAUP* program package [23]. For maximum likelihood and neighbor joining analysis, the general time reversible model of sequence evolution was assumed [24].
Synthesis of cDNA was done with the first-strand cDNA Synthesis Kit (Amersham Biosciences, Little Chalfont, UK) and 2 µg seminal vesicle RNA, according to the protocol provided by the supplier of the kit. The subsequent PCR was run with the cDNA equivalent to 30 ng total RNA in a volume of 10 µl containing 0.2 units Advantage 2 and 0.4 µM gene-specific primers 15–18 (Table 1) in 40 mM Tricine-KOH, pH 8.7; 15 mM KOAc; 3.5 mM Mg(OAc)2; 3.75 mg/ml bovine serum albumin; 0.005% Tween 20; 0.005% Nonidet-P40; and 0.2 mM deoxynucleotide triphosphate (PCR buffer). The reactions were programmed for 1 min of denaturation at 95°C, followed by 25, 30, or 35 cycles of extension comprising 30 sec at 95°C and 1 min at 68°C. At the end of the programs there was an additional incubation for 1 min at 68°C. The primers also were used for analysis on genomic DNA with 35 cycles. The products were analyzed by electrophoresis on 2% (w/v) agarose gel, stained with ethidium bromide, and visualized in ultraviolet light.
Identification of Two Semenogelin Genes in the Common Marmoset
Gene-flanking nucleotide sequences were obtained from DNA fragments generated by genome walking. The gene-specific primers were based on published sequences of an unspliced common marmoset SEMG1 transcript (Gen Bank accession number AJ005842) and nucleotides that are conserved in human (Gen Bank accession number M81651) and rhesus monkey (Gen Bank accession number X92589) SEMG2. The 3' flanking sequences were obtained from 1.4-kb SEMG1 and 1.6-kb SEMG2 products generated by PCR of the DraI GenomeWalker library. Genome walking in the 5' direction was more problematic, as many of the PCR products yielded double or unreadable sequences when subjected to DNA sequencing. Eventually, a 1.0-kb fragment of the EcoRV library and a homologous 0.8-kb fragment of the DraI library yielded sequences that were identified as SEMG1 and SEMG2 by comparing their nucleotide sequences to those flanking the human genes (Fig. 1). Comparison to the human genes also confirmed the origin of 3' flanking sequences. The gene-flanking sequences were used in the design of PCR primers for amplification of the complete genes. The nucleotide sequences were then determined on PCR fragments of 3.6 and 3.5 kb, encompassing SEMG1 and SEMG2, respectively.
|
The common marmoset SEMG1 gene (GenBank accession number AY965733) is 2.5 kb and consists of three exons of 97, 1147, and 146 bp that are separated by introns of 243 and 870 bp. The mRNA is estimated to be 1390 nucleotides, of which 21 are 5' nontranslated, 1179 are translated, and 190 are 3' nontranslated. The transcript end is preceded by a conventional polyadenylation signal 18 nucleotides upstream of the predicted poly(A) site. A TATAAA sequence is located 23–28 bp upstream of the postulated transcription initiation site. Exon 1 and the major part of the second exon are translated into a protein precursor of 392 amino acid residues, whereas the third exon remains untranslated. At the amino terminus of the precursor there is a predicted signal peptide of 23 amino acid residues, which means that the secreted SEMG1 molecule is 369 residues and has a molecular mass of 40 808 Da, as calculated from the amino acid composition. However, the primary structure contains seven consensus sequences for N-linked glycosylation, suggesting that the secreted protein is heavier than what was calculated from the amino acid composition. The glycosylation will probably also affect the isoelectric point, which was calculated to be 8.72 from the amino acid composition.
The sequence of the SEMG1 transcript presented here was compared to that of a previously reported marmoset SEMG1 clone (GenBank accession number AJ005842) [17]. The latter is presumably derived from an unspliced primary transcript or genomic DNA, as it carries the second exon flanked by a few hundred base pairs of intron sequence. The two sequences are almost identical in the nucleotides that correspond to exon 2, but there is also a major difference, as the previously reported sequence carries 180 bp that are not present in the new sequence. The size difference translates to 60 amino acid residues, which is equal to the size of the poorly defined repeats that form large parts of the semenogelin molecules. The larger molecule reported earlier, SEMG1-47, carries one repeat more than the smaller SEMG1-41 reported here. The extra repeat in SEMG1-47 was formed presumably by a duplication at a relatively late phylogenetic stage, as there are two repeats of 60 amino acid residues in the molecule that differ by only 3 residues. To sequence the complete SEMG1-47 gene, we analyzed the repeat region in new DNA samples by PCR in order to identify a source of the gene. Unfortunately, the DNA from five animals that were available all displayed a single PCR product with a size that is indicative of SEMG1-41, suggesting that they were homozygous for this variant.
Since the size variation and approximate location of the polymorph repeat in the allelic variants of the common marmoset SEMG1 are reminiscent of the size polymorphism recently reported for human SEMG1 [8, 13, 14], a comparison was done with the allelic variants of human and common marmoset SEMG1. This showed that the location of the repeats in human and marmoset SEMG1 is slightly different (Fig. 2), which suggests the size polymorphisms were generated independently in the two species.
|
The common marmoset SEMG2 gene (GenBank accession number AY965734) is 2.7 kb and consists of three exons of 97, 1339, and 166 bp that are separated by introns of 243 and 811 bp. The first and second exons are translated into a polypeptide of 459 amino acid residues and include a signal peptide of 23 residues. The third exon is untranslated, in analogy with SEMG1. The predicted message is 1602 nucleotides, of which 21 are 5' nontranslated and 201 are 3' nontranslated. As in SEMG1, there is a canonical AATAAA polyadenylation signal l8 nucleotides upstream of the predicted poly(A) site. The mature marmoset SEMG2 molecule of 436 residues has a molecular mass of 48 200 Da, but the presence of nine sites for N-linked glycosylation suggests that posttranslational modification could substantially modify the molecular size. The isoelectric point (pI) calculated from the amino acid composition is 9.66, but this value is probably also affected by glycosylation of the peptide chain.
In the common marmoset SEMG2, the gene-preceding TATAAA sequence is replaced by a variant TACAAA sequence located 23–28 bp upstream of the predicted transcription initiation site. The modified promoter element in combination with the failure in a previous report [18] to detect two distinct molecular species of semenogelin in marmoset seminal plasma prompted an investigation into the transcription of the semenogelin genes. This was done by RT-PCR analysis of RNA isolated from a common marmoset seminal vesicle. The specificity of the PCR was verified on the SEMG1 and SEMG2 genes, and then their transcripts were amplified with a varying number of PCR cycles. As can be seen (Fig. 3), both SEMG1 and SEMG2 are transcribed in the marmoset seminal vesicle—presumably they are highly expressed, as PCR products were detected already at relatively few rounds of PCR.
|
A comparison of the common marmoset SEMG1 and SEMG2 genes gave an unexpected result: a region of 1.25 kb was almost identical in the two genes, whereas the remainder of the genes only displayed 77% sequence conservation. The conserved region extends from 5 bp upstream of the translation initiation site to the type 1 repeat region (Fig. 4). There are only 15 differences in a stretch of 1255 bp, of which 6 are in the first intron and the reminder in exon 2, where 4 of them give rise to amino acid replacement. The homogenization is presumably the result of gene conversion, and the phylogenetic analysis shows that the conserved region in both genes segregates with SEMG1 (Fig. 5A), whereas the nucleotides downstream of the conserved structure segregates with either SEMG1 or SEMG2 (Fig. 5B). All phylogenetic analyses gave the same topology. The gene trees are consistent with the species tree in most aspects, and only minor deviation in the relationships among the great apes occurs.
|
|
The mammalian semen coagulum proteins display a remarkable evolution. They probably originate from a gene that coded for a small elafinlike serine peptidase inhibitor with homology to whey acidic protein [25]. Much of their further development took place after the formation of the major mammalian taxa, something that gave rise to vast differences in the primary structures of the semen coagulum proteins from human, guinea pig, and mouse. A continued evolution in the primate lineage is reflected in species-specific size differences of SEMG1 and SEMG2. There are also examples of deleted genes, as in the cotton-top tamarin, where SEMG2 is replaced by a truncated copy of a LINE1 repeat. As this was the first species of New World monkeys to be analyzed with respect to semenogelin genes, the question arose as to whether deletion of SEMG2 is a frequent phenomenon in this primate suborder. We conclude from our results here that this is not the case, as the common marmoset—which like the cotton-top tamarin belongs to the Callithricidae family of the New World monkeys—has SEMG1 and SEMG2 that are both functional.
In a previous study we demonstrated that protein staining of ejaculated semen from the common marmoset is dominated by heterogeneous material with molecular mass of 55–70 kDa. The material showed a weak but specific reaction with an antiserum against human SEMG1 that also cross-reacts with human SEMG2. Endoglycosidase digestion reduced both the molecular mass and the heterogeneity, and therefore it was concluded that the common marmoset semenogelin is highly glycosylated. This is confirmed in this report, as the primary structures of the marmoset SEMG1 and SEMG2 have seven and nine consensus sequences for N-linked glycosylation respectively. In the earlier study, the deglycosylated material appeared as a single molecular species with a molecular mass of 50 kDa at SDS-PAGE and, as Southern blotting showed that the common marmoset has two semenogelin genes, it was concluded that either the molecular masses of SEMG1 and SEMG2 are identical in this species or one of the genes is silent. In this report it is shown that both SEMG1 and SEMG2 are transcribed in the common marmoset seminal vesicles and therefore presumably also give rise to products in seminal plasma. If it is assumed that the error of the size estimation by SDS-PAGE is 5%, then the molecular mass as calculated from the amino acid composition could agree with that of SEMG2 and SEMG1-47, but not with that of SEMG1-41. However, there must be an additional explanation, as the allele analysis showed that all the tested individuals, including the one used for the reported deglycosylation results, were homozygous for the shorter allele (SEMG1-41). Due to problems with the solubility and stability of the semenogelins, the seminal plasmas are routinely diluted in a buffer containing denaturing concentration of urea. This procedure was not applied to samples taken for studies of glycosylation, as the urea would have denatured the endoglycosidase. Instead, a physiologic buffer was used, and it could be that this buffer enriched SEMG2 over SEMG1 due to differences in solubility or stability.
The evolution of the mammalian genome has resulted in the duplication of numerous genes, gene segments, and gene clusters. This genome structure or architecture provides substrates for homologous recombination and DNA rearrangements that may be significant in the development of disease or traits that influence the survival capacity of the species or the individual. The similar evolution pattern in man and marmoset, with two allelic variants of SEMG1 differing only in the length of one repeat, is interesting. The larger marmoset molecule is most likely the result of a repeat duplication, since their sequences are almost identical—only 3 of 180 positions have mismatching nucleotides. The case with the human counterpart is more complicated and probably represents a deletion [13]. Studies suggest that approximately 5% of the human population is heterozygous for the shorter SEMG1–43. The allele frequency in the marmoset population is unknown. However, in this study all animals were homozygous for the shorter SEMG1 variant, even though the DNA was derived from two different populations of marmosets. The cDNA sequence reported for the larger SEMG1 provides information from a different marmoset colony, but as it was obtained from a single subclone it is representative for one allele only. Taken together, this suggests that the shorter SEMG1–41 allele is more frequent than the larger SEMG1–47 allele in the common marmoset. However, due to the inbreeding of marmoset colonies in captivity this conclusion is speculative, and furthermore, it might not mirror the situation of free living animals.
The rapid evolution is not only a matter of repeat number, as the finding of the 1.25-kb region of high identity between marmoset SEMG1 and SEMG2 shows. The most plausible explanation in this case is the recombination of homologous genes resulting in gene conversion. The perfect match of exon 1 and the almost identical sequences of intron 1 and the first 0.9 kb of exon 2 indicate that the conversion is a recent event. The origin of the region of almost complete sequence identity is not obvious, but phylogenetic analysis indicates that SEMG1 presumably is the donating gene. Bearing this in mind, it is interesting to discuss the recent results dealing with the relationship between sperm competition and molecular evolution. Postcopulatory sperm competition has been proposed to be one of the main factors affecting sexual selection, and its effect on reproductive functions and genes has been debated. Mating behavior has been argued to determine the intensity of sperm competition, leading to more pronounced competition in species with polyandry [26]. Greater sperm competition also has been associated with more prominent semen coagulation [27]. In a recent study the rate of molecular evolution of SEMG2 was correlated with female promiscuity [12]. Also, a survey of polymorphisms and divergence in the gene for SEMG1 has shown a stronger selection pressure in species with higher degrees of polyandry [11]. In the New World monkeys the genes for the semenogelins have evolved somewhat differently, showing a deleted SEMG2 in the cotton-top tamarin and homogenized SEMG1 and SEMG2 sequences in the common marmoset. Both SEMG1 and SEMG2 have been preserved in most primates, indicating a biologic function for both of the genes, and the scenario in the New World monkeys is therefore interesting, showing a partial or complete loss of SEMG2. This seems to indicate that there is a selection against SEMG2 in callitrichine monkeys. This could very well be due to sexual selection, in case the same phenomenon is confirmed also in other monogamous species in this family of New World monkeys. However, there could also be other explanations for the evolution of semenogelins in New World monkeys, as well as for the rapid evolution seen in other orders of primates. Perhaps gene conversion, expansion and deletion of repeats, and an increased frequency of amino acid substitutions is not due to positive selection, but rather the consequence of a relative lack of purifying selection. Recently, it was shown that both SEMG1 and SEMG2 bind Zn2+, and it was also suggested that they might serve as general extracellular regulators of Zn2+, also outside of the male genital tract. This function might provide the molecules with a relative structural freedom in which the number and relative spacing of Zn2+ ligands, such as His, Glu, and Asp, are important, but the overall three-dimensional structure is not.
In man, SEMG1 and SEMG2 are the predominating protein components in the semisolid coagulum that is formed immediately upon ejaculation. In a few minutes the coagulum liquefies due to the proteolytic cleavage of the semenogelins by prostate-specific antigen (PSA). To our knowledge there are no studies demonstrating a solid relationship between the length of the semenogelin genes and the biochemical properties of the molecules, nor is there any experimental evidence supporting the idea that a longer molecule should produce a more firm coagulum. The semifluid composition of the ejaculate of the common marmoset has previously been described [18, 27], but whether there are differences in semen viscosity and coagulum formation between the individuals carrying the longer or shorter SEMG1 variant remains to be studied. The shorter variant in humans has not been associated with aberrant liquefaction or altered susceptibility to PSA. The marmoset does not have PSA, meaning that if there are alterations in the semenogelin genes due to sexual selection, they are independent of the actions of PSA. The absence of PSA, however, could perhaps accelerate the evolution of the coagulating proteins, as there is no longer a selection pressure to preserve the proteolytic sites.
In conclusion, we have demonstrated for the first time that both SEMG1 and SEMG2 occur in the common marmoset and have shown that both are expressed in the seminal vesicles. Parts of the genes are highly similar, probably due to gene conversion, and SEMG1 probably has two allelic variants differing in the lengths of one 60-amino acid repeat, in accordance with the human counterpart. The phenomenon of both duplication of repeats and recombination of a gene segment make these genes most interesting to study in different primates. The association between these rapidly evolving, unstable genes and the biologic implications of the changing structure of the molecules is yet to be investigated.
ACKNOWLEDGMENTS
Margareta Persson is acknowledged for excellent technical assistance.
FOOTNOTES
1Supported by a grant from the Medical Faculty at Lund University and from MAS Cancer. 
Correspondence: 2Camilla Valtonen-André, Wallenberg Laboratory, University Hospital MAS, S-205 02 Malmö, Sweden. FAX: 46 4033 7043; e-mail: C.Valtonen-Andre{at}med.lu.se
Received: 26 September 2006.
First decision: 4 November 2006.
Accepted: 7 December 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
