Abstract

Introduction

Histone H1 is a main component of eukaryotic chromatin. It binds to nucleosomes and linker-DNA in the chromatin fiber, playing a key role in the folding of the nucleofilament. H1 may contribute to transcriptional regulation through several mechanisms, including nucleosome positioning (Pennings et al. 1994), binding to scaffold-associated regions (Izaurralde et al. 1989; Roque et al. 2004), effects on DNA methylation (Fan et al. 2005), modulation of chromatin higher-order structure (Thomas 1999), and binding to nuclear proteins (Halle et al. 2006).

Histone H1 is the most divergent and heterogeneous group of histones. H1 is encoded by a multigene family that has evolved faster than have the core histone families. H1 has multiple isoforms (Parseghian et al. 1994; Talbert et al. 2012). In mammalians, the somatic H1s include H1.1 to H1.5, H1x, and H1.0. Germ-line-specific H1s have been designated as H1t, H1T2, Hils1 (all testis-specific), and H1oo (oocyte-specific).

Histone H1 has three structural domains: a short amino-terminal domain (NTD) (20–35 amino acids), a central globular domain (GD) (~80 amino acids) and a long carboxy-terminal domain (CTD) (~100 amino acids) (Hartman et al. 1977). The central domain is globular and stably folded in solution, and its composition is dominated by hydrophobic amino acids. The NTD contains two distinct sub-regions. The distal half is devoid of basic residues, whereas the half immediately adjacent to the globular domain is highly basic (47%) (Böhm and Mitchell 1985). The CTD is also highly basic (~40%), with basic residues, mostly Lys, uniformly distributed (Subirana 1990). Both the basic subdomain of the NTD and the CTD are intrinsically disordered, but they become extensively folded upon interaction with DNA (Vila, Ponte, Collado, Arrondo, Jiménez, et al. 2001; Vila et al. 2002; Roque et al. 2005).

The H1.1–H1.5 subtypes have evolved mainly in their N- and C-terminal domains. The composition of the C-terminal domain and, to a lesser extent, that of the N-terminal domain, is dominated by the amino acids Lys, Ala, and Pro, residues well-known for promoting protein disorder (Lu et al. 2009). Lysine is positively charged and hydrophilic, alanine has a very small side-chain and proline is a breaker of α-helix. These residues are often arranged in simple repeats, such as SPKK, PKK, PKKA, AAKK, etc. (Churchill and Travers 1991). The terminal domains are thus of low-sequence complexity. The DNA coding for the N- and C-terminal domains is also of low-sequence complexity. Low complexity DNA sequences are prone to slippage events during replication that can cause short insertions/deletions (indels) (Tautz et al. 1986; Ponte et al. 2003). In contrast, the globular domain has a sequence complexity equivalent to that of common globular proteins. It is very similar in all subtypes (93–100% sequence identity) in both interspecies and intraspecies comparisons, and it has been suggested that it may constitute a footprint of this class of subtypes (Eirín-López et al. 2004).

The origin of histone H1 can be traced back to eubacteria, where H1-related lysine-rich DNA-binding proteins have been found, long before the addition of the globular domain (Kasinsky et al. 2001). These proteins have an amino acid composition similar to the CTD and to the basic sub-region of the NTD. H1-like proteins, completely lacking the GD are present in some unicellular eukaryotes, such as Euglenozoa and Alveolata. The appearance of the “winged-helix” motif, present in the globular domain of metazoan H1s, occurred much later in protists, independently of the appearance of the H1 basic-rich regions and core histones.

The analysis of histone H1 sequences is difficult because of the lack of conserved orthologs across even moderately distant classes or phyla. Mammals present some exceptions to the lack of detectable orthology. The somatic H1 subtypes H1.1–H1.5 form a clade, and the individual subtypes have orthologs in mammalian species, which can be clearly identified by their gene organization as well as by their sequences. They are known as replication-dependent subtypes because their synthesis rates increase during the S phase. The H1.1–H1.5 genes and the H1t gene are located together with core histone genes in two large clusters on the short arm of chromosome 6 in humans (Albig, Kioschis, Poutska et al. 1997) or on chromosome 13 in mouse (Drabent et al. 1995; Wang et al. 1997). All H1 genes outside these clusters are solitary (orphon) genes, located on other chromosomes.

It has been shown that H1 subtypes can be knocked out singly and in pairs without noticeable effects on large-scale genome structure and organization, and that the other subtypes compensate the knocked out subtypes to maintain total H1 stoichiometry. However, when three subtypes are inactivated, mice are not able to complete gestation (Fan et al. 2003, 2005).

Biochemical, cytological, and developmental observations suggest that the H1.1–H1.5 subtypes are functionally differentiated (Happel and Doenecke 2009; Parseghian 2015; Millán-Ariño et al. 2016). Evolutionary evidence also supports the functional differentiation of the subtypes of the H1.1–H1.5 gene family. In vertebrates, the rates of nonsynonymous nucleotide substitution differ significantly among subtypes. Furthermore, the synonymous substitution rates greatly exceed the nonsynonymous rates, indicating the presence of strong purifying (negative) selection (Ponte et al. 1998; Eirín-López et al. 2004). In addition, the divergence of H1.1–H1.5 subtypes occurred much earlier than did the mammalian radiation (Ponte et al. 1998; Graur and Li 2005). A phylogenetic analysis of a large group of H1 subtypes showed that they cluster by type in the topologies (Eirín-López et al. 2004), confirming that they are more closely related between than within species (Ponte et al. 1998; Albig, Meergans, Doenecke 1997). Taken together, these results support the view that H1 histones have not been subject to concerted evolution. The generation and diversification of H1 isoforms is better explained by the evolutionary process of birth-and-death with strong purifying selection at the protein level, as first proposed by Nei and Hughes (1992). In this model, new genes are created by gene duplication. Some duplicated genes stay in the genome for a long time, while others are inactivated or deleted from the genome. Protein homogeneity is maintained by the effect of strong purifying selection. This model has been proposed as the primary mode of evolution for numerous multigene families (Nei and Rooney 2005), including histone H1 (Eirín-López et al. 2004).

In the present work, we have examined the evolution of the members of the mammalian histone H1.1–H1.5 gene family. Our results have revealed a high degree of preferential conservation of the basic amino acids and of some subtype-specific post-translational modification (PTM) sites. A conserved pattern of indels in the N- and C-terminal domains, ancestral to the splitting of mammalian orders, was also observed. We have found evidences of positive selection events in certain residues of H1.1. Positive selection episodes appeared to have taken place, both previous and after mammalian radiation. Previous episodes may have contributed to the differentiation of H1.1, while later episodes may reflect the adaptive potential of this subtype. Putative recombination points were found flanking the GD, suggesting that a process of genetic exchange was involved in the acquisition of the tripartite structure, typical of H1.

Material and Methods

Results

Conserved Features in H1 Subtypes

The alignment of 130 sequences (also referred as global H1.1–H1.5 alignment) belonging to the H1.1–H1.5 gene family from species representing different mammalian orders showed several highly conserved features (fig. 1; supplementary fig. S1, Supplementary Material online). It can be observed that while both terminal domains have large sequence variability, the sequence of the GD is remarkably conserved, with more than 70% of identical residues in the 130 sequences. It is also striking that more than 90% of the basic residues are conserved throughout all the protein in all subtypes, including the residues of the GD involved in DNA-binding (Goytisolo et al. 1996). The conserved pattern of basic residues is especially apparent in the CTD, where low sequence identity in the non-basic residues can be observed among subtypes (fig. 1, supplementary fig. S1, Supplementary Material online). The overall conservation of the amount and position of basic residues is presumably associated with their role in the binding of H1 to DNA in chromatin.

Open in a separate window

Fig. 1

Multiple sequence alignment of histone H1.1–H1.5 subtypes. Selected sequences from the complete alignment (supplementary fig. S2, Supplementary Material online) representing different mammalian orders clustered by species. The limits of the domains are indicated by vertical lines. Insertions/deletions present in paralogous comparisons are highlighted in red; insertions/deletions present in orthologous comparisons are highlighted in green. The conserved positions of basic residues among paralogous sequences are highlighted in light-blue. In white, lysines post-translationally modified; in yellow, phosphorylated residues; in orange, the ARKS peptide, containing the methyl-phos switch that controls H1.4 interaction with the heterochromatin protein (HP1); asterisks indicate positively selected sites in H1.1.

Close examination of the alignment showed conservation of residues that have been found post-translationally modified in mammalian species (fig. 1, supplementary fig. S1, Supplementary Material online) (Sarg et al. 2015; Izzo and Schneider, 2016). This subset comprises nine basic residues: two in the NTD, one in the CTD and six in the GD. Some subtype-specific phosphorylation sites, including those of cyclin-dependent kinases, are conserved as well. In addition, the HP1-binding motif (ARKS) is also conserved in H1.4 of primates and carnivora.

H1 Subtypes Are under Different Degrees of Purifying Selection

As expected, the phylogenetic tree with the 130 sequences shows that H1.1–H1.5 subtypes are clustered together in different branches of the tree (fig. 2). Estimation of omega (ω), defined as the ratio between nonsynonymous and synonymous substitution rates (ω=dN/dS), is a measure of selective pressure. Maximum-likelihood estimation of ω for the tree branches would allow the detection of positive selection for the functional divergence, following a gene duplication event and also of a long-term shift in the selective pressure among the different branches. The null hypothesis, H0, of a general average ω value in the entire tree was contrasted with different alternative hypotheses: H1 and H2, propose that subtypes H1.1 and H1.5, respectively, have different ω values than do the rest of the tree. H3 proposes that subtypes H1.2, H1.3, and H1.4 have significantly different ω values among each other and also from the rest of the subtypes (fig. 3). The Likelihood Ratio Test (LRT) allowed to reject the null hypothesis in all cases, suggesting that H1.1–H1.5 subtypes are under different degrees of purifying selection (table 2).

Open in a separate window

Fig. 2

Phylogenetic tree of the 130 sequences of the mammalian H1.1–H1.5 subtypes. The maximum likelihood phylogenetic tree was reconstructed with the PhyML 3.0 package, based on the best fitted nucleotide substitution model obtained with jModelTest (T92+G+I), as described in the Materials and Methods section. For each sequence, the subtype and species are indicated. Scale bar shows nucleotide substitutions per codon.

Open in a separate window

Fig. 3

Analysis of the selective pressure in the different branches of the phylogenetic tree. (A) Schematic representation of the topology of the tree and the branch-specific ω ratios. H0, null hypothesis, all the subtypes are under the same selective pressure; H1–H3 alternative hypothesis for different selective pressures in the individual subtypes.

Table 2

Averaged ω in Branches of Phylogenetic Tree of Mammalian H1.1–H1.5 Gene Family.

Hypothesis	lnL	Branches	Omega (ω)	LRT	P-value
H0	−22708.4	All the branches	0.14116
H1	−22692.1	H1.1	0.18982	32.592414	1.137E−08
		Rest of the branches	0.12314
H2	−22690.59	H1.5	0.08853	35.622658	2.395E−09
		Rest of the branches	0.15575
H3	−22694.48	H1.2	0.1097	27.834734	0.00001347
		H1.3	0.1741
		H1.4	0.0952
		Rest of the branches	0.1497

Open in a separate window

Note.—H0, null hypothesis, which considers that all the subtypes are under the same selective pressure; H1–H3 alternative hypothesis considering different selective pressures for the individual subtypes. In italics, p-values lower than 0.05, which allowed to reject the null hypothesis. LRT, likelihood ratio test, calculated as 2(lnL_{alternative hypothesis} − lnL _{null hypothesis}).

No positive selection was detected acting on any of the five subtypes. They had ω values lower than 0.2, indicating the presence of a global strong purifying selection in all the family (table 2). H1.5 is the subtype under the strongest negative selection, followed closely by H1.4, while H1.1 is the subtype with higher average ω value. It should be noted that if functional divergence of H1 subtypes had evolved by positive selection of certain amino acid would not affect significantly the ω ratios among branches (Beliawski and Yang 2005).

Possible Biological Implications of Positive Selection in H1.1

Taking into account the results of the analysis of positive selection, only sites that were identified by more than one method were considered reliable. Using this criteria, seven sites appeared to be under positive selection in H1.1 (table 3). Of the seven sites, two are located in the NTD, one in the GD, and four in the CTD.

The positions located in the NTD were detected by both branch-site methods and represent substitutions in H1.1 of conserved amino acids in the rest of the subtypes that could affect H1.1 affinity for DNA (fig. 1). The first site, at position 20 (referred to the sequence alignment in fig. 1), is a substitution of a lysine in H1.2-H1.5 by an uncharged residue. In the second site (position 39), an arginine is substituted by a lysine, which is positively charged like arginine, but has lower affinity for DNA.

The PSS located at the beginning of the GD (position 51) belongs to the first α-helix of the winged-helix motif (helix-I). In this position, H1.2–H1.5 have a threonine residue that has been substituted by valine in H1.1 (fig. 1). Two 3D-structural models were generated using the consensus sequence of H1.1 and that of H1.2–H1.5 (fig. 4A). The overlap of both models showed great structural similarity, except for helix-I that was shorter in H1.1 model (fig. 4A and B). Stability calculations considering the mutations in helix-I in H1.1 consensus sequence revealed that the T→V mutation is unfavorable for structure stability as the change in ΔG was above 0 [ddG(change)=dG(mutant) − dG(wild-type)] (fig. 4C).

Open in a separate window

Fig. 4

Potential structural changes in the GD of H1.1. (A) Consensus sequences for H1.1 and H1.2–H1.5 subtypes including the secondary structure prediction used in the 3D-model generated with I-TASSER. In red, the positively selected sites (position 51); underlined, residues predicted to form α-helix; in italics, residues predicted in strand conformation; asterisks, changes in helix-I of H1.1. (B) Overlap of the structural models of H1.1, in green and H1.2–H1.5, in yellow, using PyMOL. Arrows point to the limits of helix-I in the H1.1 model. (C) Stability calculations for mutations in helix-I of H1.1 performed by the INPS-MD prediction server.

Four PSS were detected in the CTD by more than one method, positions 124, 125, 136, and 178. Despite the similarity of the percentages of basic residues (~38%) in all H1.1 sequences, the percentages of other amino acids types are variable among the different species. In particular, percentages of non-polar residues varied from 39% to 54%, basically due to changes in the amount of alanine and valine, while the percentage of polar amino acids varied from 8% to 22%, depending on the amounts of serine and threonine present in the sequences (supplementary table S10, Supplementary Material online). Three of the positively-selected sites, 124, 125, and 136 (fig. 5A) are located in the first 25 residues of the CTD, proximal to the GD. This region may adopt both α-helix and extended strand conformations, depending on the specific amino acid composition (fig. 5B). In addition, disorder predictions in the region with the three positively selected sites (PSS) revealed changes in protein disorder. In particular, the region analyzed appeared to be more disorder-prone in primates, but more order-prone in rodentia, artiodactyla, and cetacea (fig. 5C). It is worth noting that basic residues occupy positions 124 and 136 in H1.2–H1.5. The last PSS of the CTD, position 178, is located between two CDK consensus sequences, within a region predicted to be in random coil conformation (supplementary fig. S4, Supplementary Material online). In this position, the valine present in H1.2–H1.5 subtypes was substituted by less hydrophobic or polar residues (S, T, A, and P) in H1.1, as detected by both branch-site methods (table 3).

Open in a separate window

Fig. 5

Putative structural effects of positive selection in the CTD of H1.1. (A) Multiple sequence alignment of the 27 H1.1 sequences in the region proximal to the GD, clustering positively selected sites. In red, positively selected sites; in green, indels present in H1.1 sequences. The numbers correspond to the positions in the sequence alignment in figure 1. (B) Consensus secondary structure prediction for each sequence. The analysis was performed at the Network Protein sequence analysis server available at http://npsa-pbil.ibcp.fr/. The different secondary structure motifs are labeled as follows: h, α-helix; e, extended strand; c, random coil;?, ambiguous states; -, indel in the sequence alignment. (C) Heat map of disorder tendency by residue as predicted by IUPred.

Discussion

In mammals, histone H1 comprises eleven subtypes. Subtypes H1.1–H1.5 are evolutionarily close, forming a clade (Talbert et al. 2012). H1.1–H1.5 genes are clustered with core histones, and they are expressed in a replication-dependent manner through the cell-cycle. We have examined the evolution of the mammalian H1.1–H1.5 gene family using 130 sequences, belonging to 27 different species, finding some conserved features and more importantly, evidences of positive selection in certain residues of H1.1.

Overall, basic residues are strongly conserved throughout the protein, even in the terminal domains, which have poor sequence identity between subtypes. Basic amino acids may be preferentially conserved in the terminal domains because of their major role in the induction and stabilization of chromatin higher-order structure. Therefore, a degree of functional equivalence of H1 subtypes could be achieved, helping to the interpretation of knockout experiments with the H1.1–H1.5 group of subtypes (Fan et al. 2003, 2005). The apparent functional overlap of H1 subtypes can be made compatible with their functional differentiation, assuming that specific functions optimally performed by particular subtypes can, to some extent, be fulfilled sub-optimally by other subtypes without compromising survival. Experimental evidences showing differences between H1 subtypes in their genomic distribution, expression patterns, chromatin binding affinities, PMTs and protein–protein interactions are in favor of subtype specificity (Millán-Ariño et al. 2016).

Examination of the multiple sequence alignment showed the conservation of several subtype-specific post-translational modification sites. The position and number of cyclin-dependent kinase (CDK) phosphorylation sites among subtypes are conserved both in the NTD and the CTD, indicating the importance of this PTM in the regulation of H1 function throughout the cell-cycle (Liao and Mizzen 2016). Even the specific phosphorylated residue, serine or threonine is conserved, as it is supposed to be related to the phase of the cell-cycle in which the phosphorylation occurs (Sarg et al. 2006). Some of the mitotic specific phosphorylation sites are conserved as well, including T10 in H1.5 and S35 in H1.4 (Happel et al. 2009; Chu et al. 2011). Interestingly, although S35 phosphorylation has been characterized in H1.4, consensus sequences for phosphorylation by Aurora B kinase, responsible for this modification, are detected at equivalent positions in subtypes H1.2, H1.3 and H1.5. In addition, the tetrapeptide ARKS responsible for the interaction with HP1, which is regulated by a methyl-phos switch, is conserved in primates and carnivora, but only appears sporadically in the rest of species (Daujat et al. 2005; Hergeth et al. 2011).

We also identified the presence of an indel pattern in the terminal domains conserved among paralogs, and thus, ancestral to the radiation of mammalian orders. Indels in protein coding genes are mostly small, spanning 1–5 amino acids. They occur almost exclusively in loops linking structural elements at the solvent-exposed surfaces of protein structures and are likely to be involved in intermolecular interactions and species-specific adaptations (Ajawatanawong and Baldauf 2013). Moreover, indel surveys have also been used to identify regions of the human genome under positive selection (Chen et al. 2009). Terminal domains of H1 are coded by low-complexity DNA, which is prone to slippage events during replication, originating short insertions/deletions, favoring the rapid divergence of the subtypes (Tautz et al. 1986; Ponte et al. 2003). The conservation of the indel pattern in H1 subtypes, suggest that indels might be the footprints of positive selection events associated with subtype differentiation, previous to the radiation of mammalian orders. The virtual absence of indels in ortholog sequences confirms that strong purifying selection has maintained the pattern of indels among paralogs because it is associated with function.

One the most compelling evidence of subtype functional specificity comes from the fact that sequences of gene members are more closely related between than within species, suggesting a birth-and-death mode of evolution (Eirín-López et al. 2004). Under this model of evolution new genes are created by repeated gene duplication and protein homogeneity is maintained by a strong purifying selection (Nei and Hughes 1992). In our analysis, mammalian orthologs also clustered together, indicating the higher resemblance of individual subtypes from different species in comparison with the subtypes present in a single species. In addition, the analysis of the averaged ω values in the tree branches revealed that even though all the subtypes are under strong purifying selection, the ω values have significant differences among branches, indicating that H1 subtypes are under different degrees of purifying selection. However, substitutions fixed by positive selection on a background of purifying selection were detected in H1.1.

Signatures of positive selection, ancestral to mammalian radiation and related to subtype differentiation were provided by the branch-site analyses carried out with all the sequences under study. PAML and BUSTED found strong signs of episodic positive selection in the branch clustering H1.1 sequences. This branch is longer than the counterparts to other genes, supporting the larger divergence of this subtype. In addition, H1.1 presents the highest number of ortholog indels, resulting in a variable number of amino acids. H1.1 is present in actively proliferating cells and is the predominant H1 variant in prepachytene spermatocytes, comprising approximately 70% of the total H1 (Pan and Fan 2016). H1.1 has relatively low chromatin binding affinity and is a weak chromatin condenser (Th'ng et al. 2005; Orrego et al. 2007; Clausell et al. 2009), which presumably contributes to the adoption of loosely compacted chromatin states necessary for replication and in spermatogenesis. In the latter, open chromatin is thought to facilitate genetic recombination in pachytene spermatocytes and/or the replacement of histones with transition proteins in early spermatids. Genomic mapping of H1.1 in human fibroblasts showed a distinct binding profile, suggesting a special role of this subtype in chromatin function in those cells. In contrast with subtypes H1.2–H1.5, H1.1 is present at promoters and CpG islands, enriched in intergenic regions and associated with polycomb-type chromatin (Izzo et al. 2013; Millán-Ariño et al. 2016).

The analysis at residue level revealed several positively selected sites, which may have a role in the phenotypic features of H1.1. In particular, the loss of basic residues in three of the positive selected sites located in the terminal domains (20, 125, and 136) coupled to the substitution of arginine by lysine in position 39, contribute to the decrease in the overall basicity of this subtype. In addition, the substitution of a threonine residue by valine in the helix-I of the globular domain is unfavorable for the stability of this domain. These changes, especially those in the CTD, may explain the lower affinity of H1.1 for chromatin, and therefore underscore the role of positive selection in subtype differentiation. The role of charged amino acids of the CTD in subtype affinity is evidenced in recent FRAP studies that suggest that the increased proximity of an acidic residue to the GD may be associated with the lower affinity for chromatin of mouse H1.1 (Flanagan et al. 2016).

Evidences of positive selection after mammalian radiation has been found in the CTD of H1.1 using site-specific methods. It is important to highlight that in this case the 27 sequences of H1.1 were analyzed independently and divided in three partitions due to the presence of putative recombination points. Two of the identified positions of the CTD, 125 and 136, were previously mentioned in the branch-site analysis due to the loss of a positive charge in H1.1 CTD when compared with H1.2–H1.5 subtypes. These positions were also detected to be under positive selection in H1.1 sequences, highlighting their relevance for this subtype specific functions. Interestingly, positions 125 and 136, plus another PSS (position 124) are found in a region of 25 residues of the CTD, adjacent to the GD. The equivalent region of mouse H1.0 has been shown to fold in α-helix and is associated with altering linker DNA conformation, stabilizing the folding and facilitating self-association of the chromatin fiber (Vila, Ponte, Collado, Arrondo, Suau 2001; Lu and Hansen 2004). Changes in secondary structure and intrinsic disorder propensities among H1.1 orthologs suggest that positive selection may be associated with the structural flexibility of the CTD. In fact, it has been shown that the CTD has great conformational flexibility, depending on the ionic conditions, post-translational modifications and the linker DNA conformation (Roque et al. 2005; Roque et al. 2008; Fang et al. 2012; Lopez et al., 2015; Fang et al. 2016). It seems possible that positive selection may somehow promote the folding of the CTD in a conformation or conformations optimal for H1.1 physiological role in different species.

The results show that evolutionary footprints in histone H1 subtypes are associated with both intrinsically disordered domains (NTD and CTD), indicating the importance of intrinsic disorder in the production and maintenance of genetic variation with adaptive potential (Brown et al. 2011). The more abundant amino acid in both terminal domains is lysine, which is associated with highly conserved low-complexity regions (Radó-Trilla and Albá 2012). This fact is consistent with our results, where more than half of the negatively selected sites found by REL and ADAPTSITE in the CTD are lysine residues. The conservation of the number and position of basic amino acids, as well as, the indel pattern in paralog sequences also provides a solid argument in favor of the accuracy of our results.

Low complexity regions are also proposed to promote genetic diversity by favoring recombination (DePristo et al. 2006; Lenz et al. 2014). As previously mentioned, recombination breakpoints, located close to the boundaries of the GD were detected in H1 subtypes. This finding suggests the possibility that recombination, or any other event of genetic exchange, may have been involved in the origin of H1 tripartite structure (fig. 6). A plausible model for metazoan H1 origin is a recombination event between genes encoding lysine-rich DNA binding proteins, evolutionary related to those found in eubacteria and genes encoding the “winged-helix” motif present in the GD, which appeared much later in protists, and is also present in some transcription factors (Kasinsky et al. 2001). This hypothesis would also explain that the NTD is found after the acquisition of the GD (Kasinsky et al. 2001). Later on, the combined action of gene duplication, slippage and positive selection in a birth-and-death mode of evolution may have originated the actual mammalian subtypes, which are conserved by strong purifying selection.

Open in a separate window

Fig. 6

Hypothetical model of evolution of mammalian subtypes. Lysine-rich DNA binding proteins, related to H1, have been detected in eubacteria. The acquisition of the globular domain in protist may have involved a recombination event, originating the typical tripartite structure of metazoan H1. From protists to mammals, H1 has evolved under a birth-and-death mode of evolution. Subtype differentiation may have been determined by slippage and positive selection after gene duplication events. The mammalian subtypes are represented by the different lengths of the terminal domains. In dark-gray, highly basic regions; in light-gray, hydrophobic regions of the NTD; in black, the globular domain; in black line pattern, the sites under positive selection.

In summary, we have found conservation and variation between the members of mammalian H1.1–H1.5 gene family. Conserved features include overall basicity and some PTM sites that may explain to some extent the apparent functional overlap between H1 subtypes. Two major distinct features were found, ancestral to the radiation of mammalian orders, which may be associated with subtype differentiation. In the first place, a conserved pattern of indels was identified in paralogs. This pattern may constitute a footprint of slippage events after gene duplication, associated with subtype specificity. In the second place, strong evidence of positive selection was found in the branch of the phylogenetic tree clustering H1.1 sequences. Positive selection seems to has acted on specific positions of H1.1 causing a reduction in the basicity of the terminal domains and affecting the stability of the GD, and therefore, favoring the lower affinity and chromatin condensing capacity characteristic of this subtype. Apparently, more recent events of positive selection have occurred in the CTD of H1.1 sequences. Most of the positions under positive selection are clustered in the region of the CTD adjacent to the GD, where they may modulate the folding of this domain when bound to chromatin. Additionally, the presence of putative recombination points at both sides of the GD in most of the analyzed subtypes suggests the importance of this process in the origin of metazoan H1.

Supplementary Material

Supplementary data are available at Molecular Biology and Evolution online.

Supplementary Material

Acknowledgments

Footnotes

References