SEARCH

SEARCH BY CITATION

Keywords:

  • evolution;
  • gene duplication;
  • mammals;
  • primates;
  • spermatogenesis;
  • testis-specific kinase

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

The testis-specific serine/threonine protein kinases TSSK1 and TSSK2 are known to be essential for male fertility, in mice. The enzymes are present in elongating spermatids, and targeted deletion of the two genes Tssk1 and Tssk2 results in dysregulation of spermiogenesis. The mouse genes are genetically closely linked, forming a Tssk1–Tssk2 tandem. In human, TSSK1 is present in the form of a pseudogene, TSSK1A, which is linked to an intact TSSK2 gene, and in the form of an intact gene, TSSK1B, which is not genetically linked to TSSK2. Studies on conservation of genes and gene function between mouse and human are relevant, to be able to use mouse models for studies on human infertility, and to evaluate possible targets for non-hormonal contraception targeting the male. Therefore, we have performed a detailed analysis of the evolution of genes encoding TSSK1 and TSSK2 among mammals, in particular among primates. This study includes functional analysis of replacement mutation K27R in TSSK2, which is frequently observed among humans. In primates, the kinase domains of TSSK1B and TSSK2 have evolved under negative selection, reflecting the importance to maintain their kinase activity. Positive selection was observed for the C-terminal domain of TSSK1B, which indicates that TSSK1B and TSSK2 may perform at least partly differential functions.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Spermatogenesis takes place within the spermatogenic epithelium, where the developing germ cells interact with the supporting somatic Sertoli cells. The cellular composition and structure of the spermatogenic epithelium is very dynamic, but also highly organized. In the control of spermatogenesis, hormonal and intercellular signalling pathways play important roles. In addition, many cell-autonomous events in the developing germ cells require intracellular signalling (Grootegoed et al., 2000; Ruwanpura et al., 2010; Jan et al., 2012). In view of the overwhelming evidence for the role of many different protein kinases in inter and intracellular signalling events in virtually all types of cells and tissues (Johnson, 2009b), it is not surprising that a number of protein kinases are involved in various aspects of spermatogenesis (Li et al., 2009; Lie et al., 2009; Almog & Naor, 2010; Luconi et al., 2011; Tang et al., 2012). Studying the properties, activities and evolution of testicular kinases is relevant, not only in relation to basic knowledge about spermatogenesis but also to advance our understanding of infertility. Moreover, protein kinases can be targeted by inhibitors (Johnson, 2009bb), so that testis-specific kinases might offer promising candidate targets for development of new methods for non-hormonal male contraception, in particular if such kinases are expressed late in spermatogenesis.

Studies on testicular proteins gain much relevance, when there is a high level of evolutionary conservation of structure and function between mouse and human. This allows the development of specific mouse models, in which a gene encoding a protein of interest is inactivated by gene knockout, or a comparable approach (Matzuk & Lamb, 2002). Herein, we focus on the highly conserved testis-specific serine/threonine kinases TSSK1 and TSSK2. Within the calcium/calmodulin-dependent protein kinase (CaMK) superfamily (Manning et al., 2002), these kinases form a branch with five or six TSSKs (Bielke et al., 1994; Kueng et al., 1997; Hao et al., 2004; Shang et al., 2010; Li et al., 2011). TSSKs are found in mammalian species, ranging from platypus (a monotreme), to marsupials and placental mammals (Shang et al., 2010). In mouse, Tssk14 and Tssk6 show testis-specific and post-meiotic expression (Li et al., 2011). The genes are transcribed late in spermatogenesis, and the encoded proteins belong to the last gene products in spermatogenesis. Hence, it is to be expected that interference with their activities impacts mainly on the last steps of spermatogenesis, or on sperm function. For Tssk1, Tssk2 and Tssk6, this has been investigated by generation of mouse knockout models (Spiridonov et al., 2005; Xu et al., 2008; Sosnik et al., 2009; Shang et al., 2010). From analysis of a mouse Tssk6 knockout, TSSK6 was suggested to be implicated in post-meiotic chromatin remodelling (Spiridonov et al., 2005) and sperm-egg fusion (Sosnik et al., 2009). The mouse Tssk1 and Tssk2 genes are located in tandem on chromosome 16, separated by an intergenic region of only 3.1 kb, which has prohibited knockout of these genes one by one (Xu et al., 2008; Shang et al., 2010). Targeted deletion of both Tssk1 and Tssk2 resulted in male chimeras carrying the mutant allele in spermatogenic cells, but this allele was not transmitted to offspring, indicating infertility because of haploinsufficiency (Xu et al., 2008). In an independent approach, on another mouse genetic background, we obtained a fertile Tssk1/2 heterozygous mutant mouse, which allowed us to generate the Tssk1/2 knockout (Tssk1 and Tssk2 double knockout) (Shang et al., 2010). These Tssk1/2 knockout mice have a testis-specific phenotype, with late spermatids showing developmental dysregulation of the formation of the mitochondrial sheath, resulting in male infertility (Shang et al., 2010). A testis-specific TSSK1/TSSK2 protein substrate, named testis-specific kinase substrate (TSKS), has been identified in mouse and human (Kueng et al., 1997; Scorilas et al., 2001; Hao et al., 2004). In mouse spermatids, this substrate colocalizes with TSSK1 and TSSK2 on a cytoplasmic ring-shaped structure which shows dynamic properties compatible with a role in mitochondrial sheath morphogenesis (Shang et al., 2010). Other functions and action mechanisms of TSSK1 and TSSK2 are not excluded.

Tssk1 and Tssk2, and also Tssk6, are intronless genes, which might reflect that these genes have originated from retroposition events. This retroposition, or retrotransposition, is a mechanism for gene duplication mediated by L1 retrotransposons, which can reverse-transcribe an mRNA and insert the resulting cDNA as a retrocopy elsewhere in the genome (Volff & Brosius, 2007). Most often, such an inserted cDNA lacks regulatory elements for proper control of its transcription, and the retrocopy will become a pseudogene. However, if the retrocopy is transcribed and exerts an essential function, it can be maintained as a functional retrogene. In human, around 50 retrogenes have functions in the testis (Vinckenbosch et al., 2006).

The human TSSK1A gene, located in tandem with TSSK2 within the DiGeorge Syndrome region on chromosome 22q11.21, has accumulated mutations transforming this gene into a non-functional pseudogene (Gong et al., 1996; Galili et al., 1997; Goldmuntz et al., 1997). However, TSSK1 activity might be indispensable, as indicated by the presence of another intronless gene TSSK1B, which is not a pseudogene, located on human chromosome 5q22.2 (Hao et al., 2004). Human TSSK1B and TSSK2 show 83% amino acid sequence identity in the kinase region (Hao et al., 2004). TSSK1B might be required next to TSSK2, to obtain a sufficient dose of TSSK1B/TSSK2 total kinase activity in developing spermatids. On the other hand, the actions of the two kinases might not be fully redundant. It is likely that TSSK1B and TSSK2 phosphorylate quite a number of different substrates and exert a series of important functions. Various functions might be related to different aspects of spermiogenesis and sperm maturation, such as the cytodifferentiation of spermatids (which includes marked changes in volume and cytoarchitecture of the cytoplasm and organelles), release of spermatids from Sertoli cells (spermiation) and the acquisition of sperm fertilizing capacity (Hao et al., 2004; Shang et al., 2010). The question whether TSSK1B might be required next to TSSK2 to exert some specific functions is relevant, to study a possible relationship between replacement mutations in either of these two genes and human male in fertility. In view of the technical difficulty to generate mouse models with targeted deletion of the single genes of the Tssk1–Tssk2 tandem, we have addressed this question by analysis of the evolutionary history of the genes encoding TSSK1B and TSSK2 in primates.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Retrieval of gene and protein sequences

The sequences of the genes encoding TSSK enzymes in different mammalian species were retrieved from the NCBI nucleotide collection (nr/nt) database and the Ensembl genome database, using BLAST (Basic local alignment search tool) (Altschul et al., 1990). Mouse Tssk1 (GeneID: 22114), Tssk2 (GeneID: 22115), Tssk3 (GeneID: 58864), Tssk4 (GeneID: 71099), Tssk5 (GeneID: 73542) and Tssk6 (GeneID: 83984) were used as reference sequences. The protein sequences were obtained by database search, or were translated from putative mRNA sequences. The functional domains and sites of the serine/threonine kinase region were retrieved by CDART (Conserved domain architecture retrieval tool) from Conserved domain database (CDD) (Marchler-Bauer et al., 2005).

Phylogenetic tree and dN/dS calculation

Protein (or putative protein) amino acid sequences of TSSK1 and TSSK2 from different species were used in pair alignment and Gonnet distance computation by ClustalW1.83 (Chenna et al., 2003). Statistical significance was calculated using SPSS Statistics 20 (IBM, Armonk, NY, USA), by two-sample t-test. A phylogenetic tree was constructed with the neighbour-joining method (Saitou & Nei, 1987) using MEGA4 evolutionary analysis package (Tamura et al., 2007). To calculate the evolutionary selection indicated by dN/dS for TSSK1 and TSSK2 in different species, the DNA sequences (coding sequence without stop codon) and the protein sequences were aligned using the Transalign program on the server of wEmboss (the Swiss node of EMBnet, hosted by the Swiss Institute of Bioinformatics). Output of aligned sequences was in the Phylip format. The dN/dS ratio was calculated as described (Yang & Nielsen, 2000) using the program yn00 from the Paml (Phylogenetic analysis by maximum likelihood) package, version 4 (University College London, London, UK).

Replacement mutations in human TSSK1B and TSSK2

Known replacement mutations in human TSSK1B and TSSK2 were retrieved from NCBI SNP database and HapMap Genome Browser release #28 (Smith, 2008). The secondary structure change induced by K27R in TSSK2 was detected by using the Garnier method on the server of wEmboss (Garnier et al., 1978).

In vitro phosphorylation activity of human TSSK2–27R

The replacement mutation A672G of human TSSK2, which causes the amino acid substitution K27R, was introduced by using the primers Fw1: 5′AAGAAGGGTTACATCGTAGGCATCAATCTTGGCAAGGGTTCCTACGCA AAAGTCAGATCTGCCTACTC3′, Fw2: 5′TATTTTCAGGGCGCCGACGATGCCACAGTCCTAAGGAAGAAGGGTTAC ATCGTAGGCATCAATC3′ and Rv: 5′ATGCTCTGCAGCACC-TCGG3′. Human TSSK2 and TSSK2–27R were expressed in the insect cell-line SF-21 with a baculovirus expression system, following the manufacturer's instructions (Invitrogen, Grand Island, NY, USA). A 50 kDa fragment (1–216 aa) of TSKS, which is an endogeneous substrate of TSSK1 and TSSK2, was expressed in Escherichia coli. The purified wild-type TSSK2 and mutant TSSK2–27R, together with the 50 kDa TSKS fragment, were used for the in vitro phosphorylation assay as described by Hao et al., 2004.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Replacement mutation K27R in human TSSK2

Among different mammalian species, ranging from monotremes to marsupials and placental mammals, TSSK1, TSSK1B and TSSK2 share a conserved serine/threonine kinase catalytic domain (Supplemental Fig. S1). Known replacement (non-synonymous) mutations in human TSSK1B and TSSK2 coding regions were retrieved from the NCBI SNP database and the HapMap Genome Browser. From this, we found that the mutation K27R (A672G) in TSSK2, substituting a lysine for an arginine, is found quite frequently in certain ethnic groups, in particular in Kenya and among African ancestry in Southwest USA, where the allele frequency 672G reaches more than 10% (HapMap Genome Browser release #28). The lysine residue at position 27 is located in the ATP-binding sub-domain of TSSK1, TSSK1B and TSSK2, and is conserved in all mammalian species (Supplemental Fig. S1). Lysine and arginine have a similar basic side chain, but the Emboss Garnier secondary structure prediction program (Garnier et al., 1978) suggested that the K27R substitution may induce some alpha helix disruption (data not shown). Therefore, we investigated if this mutation might affect the kinase activity of TSSK2. To this end, we performed an in vitro phosphorylation assay for full-length human wild-type and K27R mutant TSSK2, with a 50 kDa fragment (amino acid residues 1–216) of human TSKS (testis-specific kinase substrate) as the substrate. Wild-type and K27R mutant TSSK2 showed comparable activities with regard to autophosphorylation and substrate phosphorylation (Supplemental Fig. S2). This indicates that the K27R replacement mutation does not have a major impact on TSSK2 activity. Other authors have suggested that some replacement (or silent) mutations of TSSK2, TSSK4 and TSSK6 might have a meaning regarding male infertility (Su et al., 2008, 2010; Zhang et al., 2010), but this remains to be studied using a functional assay. At this point, there is no direct evidence that mutation of either TSSK1B or TSSK2 is implicated in human male infertility, which led us to an evolutionary approach, to study whether TSSK1B might be required next to TSSK2 to exert some specific functions.

Origin of TSSK1 and TSSK2 in evolution

Using the mouse Tssk1–6 genes as reference sequences, the sequences of genes encoding TSSK enzymes in different species were retrieved, as described in 'Materials and methods'. In the yeast (Saccharomyces cerevisiae), worm (Caenorhabditis elegans), fly (Drosophila melanogaster), fish (Danio rerio) and frog (Xenopus tropicalis) genomes, we did not find any sequences indicative for the presence of Tssk homologs. However, in the chicken (Gallus gallus) and lizard (Anolis carolinensis) genomes, we detected the presence of Tssk3, Tssk5 and Tssk6 homologs (Fig. 1). All six genes Tssk1–6 were found in the opossum (Monodelphis domestica) genome. Tssk3 and Tssk4 were not detected in platypus (Ornithorhynchus anatinus), but we feel that this might be caused by a relatively low coverage of the sequenced platypus genome. It is not certain if Tssk4 evolved later than Tssk1 and Tssk2, or around the same time (Fig. 1). In the radiation of placental mammals, we find Tssk1–6 in the genomes of elephant, dog, cattle, mouse, rat and primates. The Tssk5 gene of chicken, lizard, platypus, opossum and mouse shows conservation of seven introns (not shown), but mouse TSSK5 may not contain an intact kinase domain (Li et al., 2011), and the gene encoding this protein has become a pseudogene in human (Shang et al., 2010).

From the above, we suggest that the TSSK branch of CaMK enzymes originated in the ancestor of all amniotes (birds, reptiles and mammals) after diversification of amphibians and amniotes, more than 300 MYA (million years ago). In the present database analysis, the genes Tssk1 and Tssk2 appear on the stage in the mammalian genomes of platypus (Ornithorhynchus anatinus, a monotreme) and opossum (Monodelphis domestica, a marsupial).

image

Figure 1. Origin of the TSSK branch of protein kinases. From the present database analysis, we suggest that the origin of the TSSK branch is to be found in ancient amniotes, between 380 and 316 MYA. Tssk1 and Tssk2 first appeared in the common ancestor of all mammalian species, between 316 and 166 MYA. Tssk4 was not detected in platypus, here representing the monotremes, possibly caused by incomplete coverage of the sequenced platypus genome, but it is found in all other mammalian species we have investigated. The newest gene of this branch, Tssk1B, is found in new and old world monkeys, small and great apes, and human (TSSK1B in human). At least in human, and possibly in other primates and mammalian species, TSSK5 has become a pseudogene.

Download figure to PowerPoint

Genes encoding TSSK homologs in primates

Mouse Tssk1 and Tssk2 are located in very close proximity, within a region on mouse chromosome 16 that is syntenic to the human DiGeorge Syndrome region on chromosome 22q11.21 which harbours human TSSK2 (Gong et al., 1996; Galili et al., 1997). At a distance 3.5 kb upstream of TSSK2, human chromosome 22 also harbours the TSSK1A sequence, which represents a pseudogene homologous to mouse Tssk1 (Goldmuntz et al., 1997). Most likely, the original human TSSK1 was duplicated through a retroposition event, giving rise to TSSK1B located on chromosome 5q22.2. This was then followed by pseudogenization of TSSK1 yielding TSSK1A. To study the origin of TSSK1B in more detail, we have looked at its presence in other placental mammals and primates.

The present database analysis indicated that the tandem arrangement Tssk1–Tssk2 is conserved among non-primate placental mammals. We did not detect sequences homologous to human TSSK1B, other than Tssk1 in the tandem position immediately upstream of Tssk2, in any of the non-primate placental mammals. Moreover, none of the Tssk1 homologs in all non-primate placental mammals is a pseudogene. However, we found a homolog of human TSSK1B in regions syntenic to its location on human chromosome 5, in chimpanzee, gorilla, orangutan (great apes), gibbon (a small ape), macaque (an old world monkey) and marmoset (a new world monkey). Hence, Tssk1B originates in primates from before the radiation of all apes and monkeys.

The primate suborder Haplorrhini includes human, all apes and all monkeys (Simiiformes), and also the distant infraorder Tarssiformes with the family of the Tarsiidae, which is represented here by the tarsier (Tarsius syrichta). In the available sequence of the tarsier genome, we have not detected a homolog of human TSSK1B in the syntenic region, but we found Tssk1 and Tssk2 located on the same scaffold (GeneScaffold 1249) with a 6 kb intergenic region. Other than the Tssk1A pseudogene in the Simiiformes, the Tssk1 gene in tarsier contains an intact 5′-end with a start codon and an open reading frame encoding a part of TSSK1. The tarsier genome sequence is not yet complete, but the available data suggest that tarsier lacks Tssk1B and has maintained Tssk1 as a functional gene.

The suborder Haplorrhini separated from the suborder Strepsirrhini 76 MYA (Matsui et al., 2009). The Strepsirrhini includes the bushbaby (Otolemur garnettii). In the sequenced bushbaby genome, we found the Tssk1 and Tssk2 homologs located on the same scaffold (GL873737.1) and both of them contain a complete open reading frame. Gene scaffold GL873549.1 contains a region syntenic to the human TSSK1B gene locus (Fig. 2a), but we did not find a Tssk1B homolog. This indicates that the bushbaby, like tarsier, lacks Tssk1B and has maintained Tssk1 as a functional gene.

Taken together, we suggest that the origin of Tssk1B and the pseudogenization of Tssk1A have occurred in the ancestor of all Simiiformes, after the separation of Simiiformes and Tarsiiformes (Fig. 2b).

In marmoset (Callithrix jacchus), we found Tssk1A and Tssk2 in the DiGeorge Syndrome syntenic region located on chromosome 1 (not shown), and Tssk1B on chromosome 2, in a region syntenic with human chromosome 5 which harbours human TSSK1B (Fig. 2a). Interestingly, we also found a second copy of Tssk1B at a distance 2.7 Mb upstream, outside the region syntenic to the human TSSK1B locus, which we named Tssk1Bbeta (Fig. 2a). Two additional copies, Tssk1Bgamma and Tssk1Bdelta, were found on marmoset chromosome 10, also in a region that is non-syntenic to the corresponding region on human chromosome 5. Most likely, Tssk1Bbeta, -gamma and -delta originated from additional retroposition events. A start codon was found to be missing for marmoset Tssk1B, indicating that, in this new world monkey, Tssk1B has become a pseudogene, which we indicate herein as Tssk1Balpha. Likewise, Tssk1Bgamma was found to be a pseudogene. However, Tssk1Bbeta and TsskB1delta contain complete open reading frames (1125 and 1242 bp respectively) and both genes show high homology to human TSSK1B (91 and 93% coding sequence identity, and 88 and 86% protein sequence identity respectively). It appears that marmoset has lost Tssk1B (generating the pseudogene Tssk1Balpha), but gained Tssk1Bbeta and Tssk1Bdelta (and the pseudogene Tssk1Bgamma), which we have not detected in any other species. This reinforces the idea that the mammalian species require at least one functional copy of Tssk1 (or Tssk1b) in addition to Tssk2 (Fig. 2b).

image

Figure 2. Genes encoding TSSK1B in primates. (a) Human TSSK1B is located on chromosome 5. In marmoset, Tssk1B has become a pseudogene, which here is named Tssk1Balpha, located on marmoset chromosome 2 in a syntenic region. Marmoset Tssk1Bbeta, at a distance 2.7 Mb upstream from Tssk1Balpha, is a functional copy of Tssk1B. The marmoset genome contains two more Tssk1B copies, Tssk1Bgamma and Tssk1Bdelta, on chromosome 10 (not shown). In bushbaby, there is no Tssk1B copy located in the region syntenic to the region containing human TSSK1B. In tarsier, here representing the Tarssiformes, no sign of Tssk1B was found, and Tssk1 has not become a pseudogene (although further analysis awaits a higher coverage of the genome). Similarly, the bushbaby lacks Tssk1B and has maintained Tssk1 as a functional gene. (b) The data indicate that the origin of Tssk1B and the pseudogenization of Tssk1A have occurred in the ancestor of all Simiiformes, after the separation of the primate suborders Simiiformes and Tarsiiformes (*).

Download figure to PowerPoint

Sequence conservation and phylogenetic tree

The N-terminal part of the human TSSK1B and TSSK2 proteins (Supplemental Fig. S1), which harbours the kinase domain, shows 82.0% sequence identity between the two proteins, whereas this value is only 14.7% for the C-terminal region (Hao et al., 2004). Such a higher divergence of the C-terminal region, compared with the N-terminal region, can be observed for different mammalian species (Supplemental Fig. S1 and Table 1a). This might indicate that the C-terminal region is involved in controlling specific properties of the kinases, and that these properties have diverged after the two genes originated from a gene duplication event. This would be in agreement with a hypothesis that the two kinases have differential functions, in addition to some overlap of functions.

Table 1. Percentage identity of the N- and C-terminal sequences of TSSK1/1B and TSSK2: (upper panel) in different mammalian species, (lower panel) in a comparison between non-primate species (upper half) and human vs. non-primate species (lower half)
TSSK1/1B vs. TSSK2NC
Og 88.626.1
Mm 88.216.5
Bt 90.118.9
Cf 88.222.1
La 89.023.2
Hs 82.0 14.7
 TSSK1/1BTSSK2
NCNC
  1. Species represented: human (Homo sapiens: Hs), bushbaby (Otolemur garnetti: Og), mouse (Mus musculus: Mm), cattle (Bos taurus: Bt), dog (Canis familiaris: Cf) and elephant (Loxodonta africana; La).

Mm vs. Og97.860.298.987.2
Mm vs. Bt98.272.898.582.6
Mm vs. Cf 97.163.099.386.0
Mm vs. La97.464.098.282.6
Og vs. Bt98.965.397.888.4
Og vs. Cf96.765.698.994.2
Og vs. La97.461.197.491.9
Bt vs. Cf97.169.198.287.2
Bt vs. La97.870.897.884.9
Cf vs. La97.165.297.890.7
Mean ± SD97.6±0.665.7±4.198.3±0.687.6±3.8
Hs vs. Og92.661.395.291.9
Hs vs. Mm92.662.494.983.7
Hs vs. Bt93.062.494.182.6
Hs vs. Cf91.956.495.688.4
Hs vs. La92.364.494.587.2
Mean ± SD92.5 ± 0.461.4 ± 3.094.9 ± 0.686.8 ± 3.7

Next, we examined the percentage sequence identity for N- and C-terminal domains of TSSK1/1B and TSSK2, between species (Table 1b). Among non-primate mammals, the C-terminal part of TSSK1 shows 65.7% mean sequence conservation, which is significantly lower compared to 87.6% mean sequence conservation of the C-terminal domain of TSSK2 (p < 0.005). For human, compared to several non-primate mammalian species, the mean sequence conservation of the C-terminal domain of TSSK1B was even lower, 61.4%, as compared to 86.8% for the C-terminal region of TSSK2 (Table 1b). In this comparison of human to non-primates, also the N-terminal kinase domain of TSSK1B showed a somewhat lower sequence conservation (mean 92.5%) compared to the N-terminal kinase domain of TSSK2 (mean 94.9%) (Table 1b; p < 0.005). Taken together, this indicates that differential selective pressures are acting on TSSK1/1B and TSSK2, mainly on the C-terminal domain but also on the N-terminal kinase domain.

It is to be expected that the kinase domain of TSSK1/1B and TSSK2 offers less room for sequence variation and evolution, being restricted by structural and functional requirements for its kinase activity. However, by constructing a phylogenetic tree (based on the N-terminal sequences presented in Supplemental Fig. S1), its evolutionary dynamics can be visualized (Fig. 3). It appears that the origin of TSSK1B in primates has resulted in a series of mutations which caused divergence between TSSK1B in primates and TSSK1 in the non-primate mammals. Simultaneously, TSSK2 in primates is found at some distance from TSSK2 in non-primate mammalian species (Fig. 3).

image

Figure 3. Phylogenetic tree of TSSK proteins in mammalian species. The tree represents the N-terminal kinase domain of TSSK1, TSSK1B and TSSK2, in primates, other placental mammals, and a marsupial: human (Homo sapiens; Hs), chimpanzee (Pan troglodytes; Pt), gorilla (Gorilla gorilla; Gg), orangutan (Pongo pygmaeus; Pp), gibbon (Nomascus leucogenys; Nl), macaque (Macaca mulatta; Ma), marmoset (Callithrix jacchus; Cj), bushbaby (Otolemur garnetti; Og), mouse (Mus musculus; Mm), cattle (Bos taurus; Bt), dog (Canis familiaris; Cf), elephant (Loxodonta africana; La) and opossum (Monodelphis domestica; Md). The tree was constructed using the neighbour-joining method (Saitou & Nei, 1987), based on the multiple sequence alignment shown in Supplemental Fig. S1. The evolutionary distances, in units of number of amino acid substitutions per site per million years, were computed as described by Zuckerkandl & Pauling, 1965. Marmoset (Cj) TSSK2 is not included, because its sequence is not present in the current genome assembly. The red and green branches represent TSSK1B and TSSK2, respectively, in primates.

Download figure to PowerPoint

Positive and negative selection

Evolution of duplicated genes may occur at different rates (Cusack & Wolfe, 2007; Han et al., 2009; Jun et al., 2009; Wang et al., 2010) which can be investigated by calculation of the ratio dN/dS. This ratio represents the number of non-synonymous substitutions per non-synonymous site (dN) divided by the number of synonymous substitutions per synonymous site (dS). The first type of substitutions, indicated by dN, lead to changes in the amino acid sequence of the encoded protein, which can be either negatively selected if the function of the protein is impaired, or positively selected if the function of the modified protein is improved or changed towards gain of a new function. The synonymous substitutions, indicated by dS, are silent changes, based on the degeneracy of the genetic code, which do not cause a change in the encoded protein and hence are not subject to selection. Therefore, the ratio dN/dS provides information about negative selection (dN/dS <1) and positive selection (dN/dS >1), when analysed for closely related species (Nekrutenko et al., 2002).

We have made these calculations for the N- and C-terminal regions of TSSK1 and TSSK2 (Supplemental Fig. S1) in three rodents (Mus musculus, Rattus norvegicus and the ground squirrel species Spermophilus tridecemlineatus) and TSSK1B and TSSK2 in human and its two close relatives, chimpanzee (Pan troglodytes) and gorilla (Gorilla gorilla). From the results presented in Table 2, it appears that the kinase domains of TSSK1 and TSSK2 have evolved under negative selection, in agreement with the importance of maintenance of their kinase activity. Positive selection (dN/dS >1) was observed for the C-terminal domain of TSSK1 in rodents, and very clearly also for TSSK1B in primates (Table 2a and b). This positive selection points to divergence of the function of the C-terminal domain between TSSK1/1B and TSSK2, which might be associated with differential interaction with protein partners and substrates.

Table 2. Calculation of the dN/dS ratio for TSSK1 and TSSK2: (upper half) for rodent species, (lower half) for primate species
 NC
dN/dSdNdSdN/dSdNdS
  1. Species represented: mouse (Mus musculus: Mm), rat (Rattus norvegicus: Rn), squirrel (Spermophilus tridecemlineatus: St),

  2. human (Homo sapiens: Hs), chimpanzee (Pan troglodytes: Pt), gorilla (Gorilla gorilla: Gg). Positive selection (dN/dS >1) is in bold. When dS is zero, dN/dS cannot be calculated (NA).

 TSSK1
  Mm vs. Rn0.0360.0000.130 1.123 0.0580.052
  Mm vs. St0.0170.0030.460 1.870 2.1061.126
  Rn vs. St0.0110.0040.449 1.639 1.9271.176
 TSSK2
  Mm vs. Rn0.0000.0000.1120.3260.0390.119
  Mm vs. St0.0030.0020.4980.4420.4971.124
  Rn vs. St0.0040.0020.4190.3870.4641.199
 TSSK1B
  Hs vs. Pt0.1100.0030.028 3.429 0.0380.011
  Hs vs. Gg0.0740.0030.041 1.470 0.0320.022
  Pt vs. Gg0.0000.0000.040 3.515 0.0380.011
 TSSK2
  Hs vs. PtNA0.0040.000NA0.0090.000
  Hs vs. Gg0.4130.0030.007NA0.0090.000
  Pt vs. Gg0.6190.0040.007NA0.0090.000

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Gene duplication and the subsequent divergence of the duplicates is viewed as an important mechanism leading to the formation of new genes, in evolution (Long, 2001). The mechanisms of gene duplications are either DNA-based or RNA-based (Long et al., 2003; Kaessmann et al., 2009; Innan & Kondrashov, 2010). Possibly, either Tssk1 or Tssk2 originated by an RNA-based retroposition event, from an unknown parental gene. Such a mechanism would explain the absence of introns. In general, when a gene is duplicated, one of the gene copies is relieved from negative selection, when the other gene maintains the original function. Relieve from negative selection will result in mutation, which leads to pseudogenization or might sometimes provide a gain-of-function leading to neofunctionalization. A retrocopy is disconnected from its ancestral promoter and is integrated into a new chromatin environment, which increases mutational asymmetry between the retrocopy and the parental gene. It has been found that the retrogenes show relatively fast rates of evolution compared with their parental genes (Cusack & Wolfe, 2007).

In a possible scenario, retroposed Tssk1 (or Tssk2) escaped from pseudogenization, and gained a new function in spermatogenesis, associated with testis-specific expression. Such a testis-specific expression, in spermatogenesis, is observed for quite many retrogenes (Wang, 2004). Whoever was first, either Tssk1 or Tssk2, the tandem Tssk1–Tssk2 must have formed as a result of a next gene duplication, possibly DNA-based. Tandem duplicates are more conserved, compared to retrogenes which are relocated to other sites in the genome (Cusack & Wolfe, 2007; Han et al., 2009; Jun et al., 2009; Wang et al., 2010). Following the proposed DNA-based gene duplication leading to the Tssk1Tssk2 tandem arrangement, the syntenic context of these two genes was virtually identical, which may have limited mutational asymmetry. The fact that two genes, encoding TSSK1 and TSSK2, were maintained in the mammalian ancestor where the gene duplication has occurred, might be explained if a higher dose of TSSK1/2 kinase gave an immediate selective advantage. Subsequent positive selection may have provided TSSK1 and TSSK2 with the proposed differential functions. Our comparative analysis indicates that the C-terminal domain of TSSK1 has undergone more mutational changes (showing a lower percentage of sequence conservation), compared to TSSK2, in the evolution of mammalian species.

In the primate lineage, a next gene duplication gave rise to Tssk1B. We suggest that this represents a retroposition of Tssk1. For a relatively short time, the genes Tssk1, Tssk1B and Tssk2 may have coexisted in an ancestor of the Simiiformes (human, all apes and all monkeys), but only Tssk1B and Tssk2 have survived negative selection, with Tssk1 turning into the pseudogene Tssk1A. In this series of events, Tssk1B was inserted into the genome at a different location, having lost the genetic linkage to Tssk2. For the C-terminal domain of TSSK1B, we describe a strong positive selection of sequence changes among gorilla, chimpanzee and human, similar to what we observed for TSSK1 among rodents. However, a phylogenetic tree representing the N-terminal kinase domain of TSSK1 and TSSK2 revealed that TSSK1B in primates has diverged considerably from TSSK1 in non-primate mammals. This is in accordance with, and lends support to, the idea that adaptive evolution may occur faster for gene copies which have moved to a new genomic location (Cusack & Wolfe, 2007; Han et al., 2009; Jun et al., 2009; Wang et al., 2010). The present phylogenetic tree also indicates that, with the evolutionary appearance of TSSK1B, the kinase domain of TSSK2 in primates may have accumulated a few more sequence changes, compared to TSSK2 in non-primate mammals.

Evolutionary maintenance of both TSSK1/1B and TSSK2 among mammalian species indicates that there is an added value, most likely in relation to reproductive fitness, to have two different genes encoding similar proteins. This might be related to obtaining a sufficiently high level of kinase activity. However, mice heterozygous for the targeted Tssk1–Tssk2 deletion do not show significant impairment of spermatogenesis, when housed under normal animal breeding conditions (Shang et al., 2010). In these heterozygous mice, the total dose of TSSK1 and TSSK2 is about halved, but one allele of the Tssk1–Tssk2 tandem is still present and active, apparently able to maintain spermatogenesis, at least when the animals are not exposed to environmental stress. Dose-sensitivity is not completely excluded, in particular if mice would be exposed to natural selective conditions, but the added value to have TSSK1/1B in addition to TSSK2 might be related to differential functions of the encoded kinases. This suggestion is supported by the present observations on the positive evolutionary selection of the C-terminal domain of TSSK1/1B.

The Tssk5 gene of chicken, lizard and mouse contains seven introns, compared to one intron in Tssk3 and no introns in Tssk6. Possibly, the parental gene of the family was Tssk5, giving rise to Tssk3 and Tssk6 by retroposition, where Tssk3 would have kept or gained an intron. However, alternative scenarios cannot be excluded. Evidence for a retrogene would require the presence of flanking sequences showing marks of the retroposition event. For young retrogenes, such evidence can be obtained, but these marks are usually erased for older genes (Long, 2001; Long et al., 2003; Kaessmann et al., 2009). The origin of the intronless genes Tssk1 and Tssk2 166–316 MYA likely makes it impossible to obtain conclusive evidence from flanking sequences that these genes result from retroposition. Yet, we consider these genes as retrogenes. In spermatogenesis, an X-to-autosomal retrogene can act as a back-up copy of an X-chromosomal gene, when the X chromosome is transcriptionally silenced in meiotic prophase as a consequence of meiotic sex chromosome inactivation (MSCI) leading to formation of the XY body (Boer et al., 1987; McCarrey & Thomas, 1987; Baarends et al., 2005; Turner et al., 2005). None of the known Tssk genes, which may or may not include the parental gene, is X-chromosomal. Hence, it seems likely that the driving forces which have acted towards evolution of Tssk retrogenes may not be directly related to the evolution of the heterologous sex chromosomes and XY body formation.

Taken together, our evolutionary analysis indicates that TSSK1/1B is required next to TSSK2, to perform essential functions in spermatogenesis. Ancient gene duplication events in mammalian lineages may have provided a selective advantage based on an increased dosage of the encoded TSSK1/2 total kinase activity, but the conservation of at least one copy of a Tssk1 homolog next to a Tssk2 gene in various mammalian species might also be explained if a wider spectrum of biological functions is performed by the two encoded kinases. At this point, it is not possible to indicate which specific functions might be attributed to either TSSK1/1B or TSSK2. However, it is warranted to study differential functions of TSSK1/1B and TSSK2 in more detail, also in relation to human male infertility.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information
FilenameFormatSizeDescription
andr21-sup-0001-FigS1-Part-1.tifimage/tif12593KFigure S1. Alignment of amino acid sequences of TSSK1, TSSK1B and TSSK2 in mammalian species. The multiple sequence alignment was performed using ClustalW1.83 program (Chenna et al., 2003). The amino acid sequence 1–272 represents the N-terminal domain, described in the present report, and the sequence from amino acid 273-end represents the C-terminus. The serine/threonine protein kinase catalytic domain, amino acid residues 1–269, was defined by a NCBI CDD search (Marchler-Bauer et al., 2005), and is shown as a black arrow bar. Completely conserved amino acid residues are indicated by asterisks. Amino acid residues consistently found in TSSK1B but not in TSSK1 or TSSK2 are indicated with a blue mark. The conserved lysine K27 is indicated with a green mark. Species represented are: human (Homo sapiens; Hs), chimpanzee (Pan troglodytes; Pt), gorilla (Gorilla gorilla; Gg), orangutan (Pongo pygmaeus; Pp), gibbon (Nomascus leucogenys; Nl), macaque (Macaca mulatta; Ma), marmoset (Callithrix jacchus; Cj), bushbaby (Otolemur garnetti; Og), mouse (Mus musculus; Mm), cattle (Bos taurus; Bt), dog (Canis familiaris; Cf), elephant (Loxodonta africana; La) and opossum (Monodelphis domestica; Md).
andr21-sup-0002-FigS1-Part-2.tifimage/tif7575K 
andr21-sup-0003-FigS2.tifimage/tif993KFigure S2. TSSK2 and TSSK2–27R in an in vitro phosphorylation assay. The enzymatic activity of human TSSK2–27R was examined in an in vitro phosphorylation assay, in which the wild-type human TSSK2 was used as a positive control and a 50 kDa human TSKS fragment was used as the substrate. Both wild-type TSSK2 and TSSK2–27R showed autophosphorylation (open arrow heads) and phosphorylation of the TSKS fragment (closed arrowheads).

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.