Introduction

Ubiquitin, a 76-amino acid protein, has a very slow evolutionary rate during the eukaryote evolution (Sharp and Li 1987; Baker and Board 1991). Besides its evolutionary conservation, ubiquitin is unique in that it can covalently attach to proteins that are destined for the protein degradation pathway. Proteins marked by ubiquitin are captured by proteasomes and disassembled into amino acids. The human genome contains four ubiquitin genes, UbA 52 (Baker and Board 1991), UbA 80 (Lund et al. 1985), UbB (Baker and Board 1987), and UbC (Wiborg et al. 1985; Nenoi et al. 1996). UbA 52 and UbA 80 encode monomeric ubiquitin genes fused with ribosomal proteins, whereas UbB and UbC encode polyubiquitin genes. The polyubiquitin genes comprise multiple tandem ubiquitin molecules without a spacer sequence or intron, and the translated products are cleaved into monomeric ubiquitin molecules by carboxyl-terminal-specific hydroxylase (Redman and Rechsteiner 1989). Human UbB and UbC are located on 17p11-12 (Webb et al. 1990) and 12q24.3 (Board et al. 1992), respectively.

Tandem duplicated genes have a tendency to be homogenized through multiple events of interlocus recombination and gene conversion (e.g., Kawamura et al. 1992; Kitano and Saitou 1999; Liao 1999; Gonzalez and Sylvester 2001). Coding units within polyubiquitin genes in a number of lineages are more similar to each other than to their orthologues in other species, suggesting that polyubiquitin genes are subjected to concerted evolution (Sharp and Li 1987; Tan et al. 1993; Keeling and Doolittle 1995; Vrana and Wheeler 1996; Nenoi et al. 2000). Concerted evolution is a phenomenon for multigene families when the members of a repeated sequence family are highly similar, whereas members from closely related species may differ greatly (Li 1997). Unequal crossover is considered to be the main evolutionary mechanism for creating concerted evolution, but gene conversion may also play a role (Nei 1987).

Recently, Nei et al. (2000) observed that the coding units within polyubiquitin genes are diverged in silent substitution sites, and they proposed that the majority of polyubiquitin genes do not evolve in concert but, rather, through a birth-and-death process, in which the monomeric coding units are assumed to evolve almost independently, and protein homogeneity is attained by strong purifying selection. Comparison of ubiquitin genes for closely related species can provide a great insight to the evolution of this gene. We thus determined UbB and UbC nucleotide sequences of human, chimpanzee, gorilla, and orangutan. Nucleotide sequence similarities of ubiquitin coding units were compared among closely related species to determine whether the observed homogenization occurred by concerted evolution or birth-and-death evolution.

Materials and Methods

Source of Genomic DNA

Genomic DNAs of Homo sapiens (95 Japanese, 80 European-Americans, and 95 African-Americans), Pan troglodytes (12 individuals), Gorilla gorilla (1 individual), and Pongo pygmaeus (1 individual) were obtained from whole blood or cell lines. DNAs of European- and African-Americans were supplied by the Coriell Institute for Medical Research (Camden, NJ), while those of Japanese were obtained through informed consent.

Amplification of Ubiquitin Units and Detection of Ubiquitin Repeat Number Polymorphisms

DNA fragments containing entire polyubiquitin genes were amplified by polymerase chain reaction (PCR) using LA Taq polymerase (Takara). The primer set, oligo 3 and oligo 6, designed for human UbC amplification, as described in Nenoi et al. (1996), was used for UbC amplification in great apes. For the amplification of UbB, a primer set was designed based on human genomic contig clone NT_02718 (forward, 5′-AGGAAGGTTTCTTCAACTCAAATTC-3′; reverse, 5′-TAATGACTTAAACAGCAAAGAAGGC-3′ and 5′-TGACTTAACTTTGAGCACATTACCA-3′). The product sizes of amplified UbB and UbC were measured with agarose gel electrophoresis, and the numbers of ubiquitin repeats were estimated from the product size.

Sequencing of Polyubiquitin Genes

The UbB sequence was determined directly using BigDye Terminator cycle sequencing on an ABI PRISM 377 DNA analyzer (Applied Biosystems). Parts of the sequences of gorilla and orangutan UbB could not be determined by direct sequencing, therefore TA-cloning was performed to obtain complete sequencing. The sequencing strategy for UbC was as follows: PCR-amplified fragments containing the entire UbC of various species were purified using a QIAquick PCR Purification Kit (QIAGEN) and subcloned into pGEM-T Easy TA-cloning vector (Promega). The clones containing a full-length coding sequence of UbC were subjected to SalI and SacI digestion and exonuclease III treatment using the Erase-a-Base System (Promega). A series of truncated clones were then constructed. The clones were sequenced from a universal primer on the vector sequence with DNA sequencer model 377 (Applied Biosystems). Sequences of the deletion clones were aligned using Sequencher (Gene Codes) and the sequence diversity was analyzed.

Phylogenetic Tree Construction and Examination of Ubiquitin-Repeat Homogenization

Each ubiquitin coding unit of UbB and UbC from human and great apes was separated and nucleotide diversities were calculated for both within and between species. To examine the extent of sequence divergence, the number of synonymous differences per synonymous sites (p s ) was calculated using a modified Nei–Gojobori method (Kumar et al. 2001; Nei et al. 2000) with MEGA2 software (http://www.megasoftware.net/ ). The p s values for all pairs of ubiquitin units were compared within and between polyubiquitin genes in all species. We obtained neighbor-joining trees (Saitou and Nei 1987) with Kimura’s (1980) two-parameter model using Clustal W (Thompson et al. 1994) and visualized it using Dendromaker (http://www.cib.nig.ac.jp/dda/timanish/dendromaker/home.html ).

Results

Repeat Number Polymorphisms of UbB and UbC

Two polyubiquitin genes, UbB and UbC, of chimpanzee, gorilla, and orangutan were amplified by PCR using primer sets designed for human UbB and UbC. By detecting the nucleotide similarities of the 5′-UTR regions between PCR products from various species, we confirmed that UbB and UbC were specifically amplified.

The number of ubiquitin coding units of UbC was estimated to be 6 to 11 for human (n = 270), 10 to 12 for chimpanzee (n = 12), 8 for gorilla (n = 1), and 10 for orangutan (n = 1) (Fig. 1 and Table 1). The repeat number 9 was predominant among three distinct human populations, African-Americans, European-Americans, and Japanese. We identified novel repeat number polymorphisms, 6 in Japanese and 10 and 11 in African-Americans. The proportion of minor alleles of UbC repeat number polymorphisms, i.e., alleles other than repeat 9, in human were 0.053 for Japanese, 0.062 for European-Americans, and 0.117 for African-Americans, while that for Australian Caucasians is reported to be 0.127 (Baker and Board 1989). The repeat number polymorphisms of chimpanzee UbC showed a higher heterozygosity than those of human UbC (Table 1). These variations in coding unit number most likely resulted from unequal crossover of two UbC alleles, presumably promoted by the repetitive nature of the polyubiquitin gene. To compare sequence variations between species, we defined the representative repeat number in each species that is the most common allele: 9 for human, 10 for chimpanzee, 8 for gorilla, and 10 for orangutan.

Table 1 Frequencies of repeat number polymorphisms of UbC among human and great apes
Full size table
Figure 1
figure 1

Analysis of ubiquitin repeat number of UbC among human and great apes. UbC of human and great apes was amplified by PCR and the PCR products were separated on 0.9%agarose gels with size standards. Six individuals for human, four for chimpanzee, one for gorilla, and one for orangutan are shown.

Full size image

The number of repeat of ubiquitin coding units in UbB was three in human and great apes (data not shown). Because of the small number of ubiquitin repeats in UbB, the unequal crossover might not have occurred in the great ape lineage.

Nucleotide Sequences Determined

The consensus sequence was defined from multiple individuals (13 humans homozygous for 9 UbC repeats and 5 chimpanzees homozygous for 10 UbC repeats) or from multiple clones (6 clones for gorilla and 5 for orangutan) to avoid PCR-induced error. Polyubiquitin gene sequences of human and great apes were deposited in the DDBJ/EMBJ/GenBank International Nucleotide Sequence Database under accession numbers AB089613–AB089616 for UbC and AB089617–AB089620 for UbB. Multiple alignments of these sequences only for variants sites are shown in Fig. 2. Because of repeat structure, each repeat was considered as a unit of comparison.

Figure 2
figure 2

Multiple alignments of the representative UbB and UbC nucleotide sequences for human and great apes. A UbB. B UbC. Each repeat was separated for comparison. Only variant nucleotide sites are presented, and site positions are shown above sequences. Periods designate same nucleotide as the first sequence, while hyphen means gaps. Sites with numbers 1 and 3 shown at the bottom of alignment A are those having repeat 1- and 3-specific nucleotides, respectively.

Full size image

Phylogenetic Trees of Monomeric Ubiquitin Units

Each monomeric ubiquitin coding unit of UbB and UbC from human, chimpanzee, gorilla, and orangutan was used as unit of comparison. Because UbB and UbC are located in different chromosomes, they probably diverged long time ago. We therefore conducted a BLAST homology search (Altschul et al. 1997) using human ubiquitin B repeat unit 1 as query and obtained seven homologous nonhominoid mammalian sequences (accession numbers are AB071067 for crab-eating macaque, AF038129 for sheep, Z18245 for bovine, BC019850 for mouse, D16554 for rat, AB003731 for Chinese hamster, and AF506969 for horse). Phylogenetic trees for those sequences as well as UbC repeat sequences of human, chimpanzee, gorilla, and orangutans were constructed as shown in Fig. 3. Human UbC repeat 1 was used as outgroup. We found clustering of different mammalian orders; rodents (mouse, rat, and Chinese hamster), ungulates (sheep and bovine), primates (hominoids and crab-eating macaque). Four horse repeat sequences are so diverged that they did not form a monophyletic cluster. Because relatively short nucleotides (228 bp) were compared, bootstrap probabilities were not so high. In any case, it is clear that homogenization of repeats did not occur frequently, as pointed out by Nei et al. (2000).

Figure 3
figure 3

A phylogenetic tree for monomeric ubiquitin coding units of UbB among various mammalian species. Human ubiquitin C repeat 1 was used as outgroup. Bootstrap probabilities are given only for branches leading to major mammalian orders, except for horse. B-Hu, human UbB; B-Ch, chimpanzee UbB; B-Go; gorilla UbB; Ub-Or, orangutan UbB; chamster, Chinese hamster; cemacaque, crab-eating macaque. Numbers after hyphen designate repeat number for each species.

Full size image

The phylogenetic relationship of the ubiquitin B repeat 2 in primates is somewhat unusual. The human and crab-eating macaque repeat 2 are identical, while chimpanzee and gorilla repeat 2 are identical. These two pairs are clustered and orangutan repeat 2 is outside of this cluster (Fig. 3). This topology is clearly different from the species phylogeny of primates, and some kind of homogenization must have occurred during the primate evolution. It is at least very difficult to explain the sequence identity of human and macaque repeat 2 by assuming a birth-and-death process. In contrast, the phylogenetic relationship of repeats 1 and 3 is more or less concordant with the primate phylogeny, though human and orangutan were clustered in repeat 1.

We also conducted the BLAST homology search by using human ubiquitin C repeat unit 1 as query and obtained five homologous nonhominoid mammalian sequences (accession numbers are M18159 for pig, BC025894 and BC021837 for mouse, D17296 for rat, D63782 and AB003732 for Chinese hamster, and D83208 for guinea pig). As in the case of UbB, all the hominoid repeat sequences became monophyletic (tree not shown). We therefore present the phylogenetic tree of hominoid ubiquitin C repeats in Fig. 4. Human UbB repeat 1 was used as outgroup. Ubiquitin coding units of UbC of the four hominoid species have complicated patterns. The last ubiquitin coding units formed a distinct cluster with bootstrap probability of 97%. The first repeat also formed a monophyletic cluster, though its bootstrap support (35%; not shown in Fig. 4) was low. In contrast, there was no monophyletic clustering for intermediate repeats. In gorilla, a high degree of homogenization of ubiquitin coding units was observed, and repeats 2–7 formed a monophyletic cluster. In chimpanzee, the homogenization of ubiquitin repeats was weaker than gorilla, however, eight intermediate repeats 2–9 belong to either one of two clusters, suggesting a certain homogenization. In human, partial homogenization was observed; repeats 5–8 formed one cluster, while repeats 2 and 4 are very similar. On the other hand, the homogenization of ubiquitin repeats was not extensive in orangutan; only repeats 4 and 6 made a cluster (see Fig. 4).

Figure 4
figure 4

A phylogenetic tree for monomeric ubiquitin coding units of UbC among the four hominoid species. Human ubiquitin B repeat 1 was used as outgroup. Bootstrap probabilities (%) are given only when they were higher than 80%. C-Hu, human UbC; C-Ch, chimpanzee UbC; C-Go, gorilla UbC; C-Or, orangutan UbC. Numbers after hyphen designate repeat number for each species.

Full size image

Synonymous Differences of UbC

We evaluated the UbC homogenization in gorilla and others using the following statistical analysis as described by Nei et al. (2000). The mode of evolution was tested by comparison of the proportion of nonsynonymous differences (p N ) and the proportion of synonymous differences (p S ). The p N of ubiquitin was zero among human and great apes, therefore only the p S was compared. Table 2 shows intraspecific comparison of UbC repeats. The model of concerted evolution predicts small p S values. Because the first and last ubiquitin coding units of UbC are more conserved between species and less conserved within species, p S was also estimated after removing the first and last ubiquitin repeats (data not shown). The homogenization was remarkable when ubiquitin coding units were compared between internal units (mean p S = 0.05 in gorilla).

Table 2 Synonymous differences (p S ) per site among human and great apes (UbC) using the Nei–Gojobori method
Full size table

Discussion

Two polyubiquitin genes, UbB and UbC, were examined and each ubiquitin repeat was compared among the four hominoid species (human, chimpanzee, gorilla, and orangutan). Because the chromosomal locations of UbB and UbC are assigned to the same regions as HOXB (17q21.3) and HOXC (12q13.3), respectively (Apiou et al. 1996), the ubiquitin gene family might have arisen through genome duplication events as already shown for the HOX gene family (Ohno 1970; Sidow 1996; Kasahara et al. 1996; Bailey et al. 1997; Skrabanek and Wolfe 1998; Pennisi 2001). A large amount of ubiquitin appears to be advantageous, especially under stressful conditions, therefore the number of coding units per allele should gradually increase with time until a maximum is reached, which might be balanced between increased selective advantage and reduced allele stability in a species-specific manner. As shown in Table 1 and Fig. 1, the number of ubiquitin repeats in UbC varies even among closely related species; 9 for human, 10 for chimpanzee, 8 for gorilla, and 10 for orangutan. The repeat number varies within species, e.g., 6 to 11 for human, indicating that the repeat number itself might not be under strong selection. On the other hand, hominoid and macaques have maintained 3 ubiquitin repeats in UbB (see Fig. 3).

Concerted evolution is a phenomenon of intraspecific homogenization that is observed in many repeat sequence families (Liao 1999). The high degree of intrasequence homogenization of UbC typically observed in gorilla (Table 2) could be due to concerted evolution, as previously reported for the Chinese hamster polyubiquitin gene (Nenoi et al. 1994, 1998). A major mechanism of concerted evolution is an unequal crossing-over of two alleles, which is a reciprocal recombination process. Another possible mechanism is gene conversion, which is a nonreciprocal transfer of genetic information between similar sequences. When we compared each ubiquitin unit of UbC in different species, there were high similarities in the first ubiquitin unit and the last ubiquitin unit between species (Figs. 2B and 4) and the homogenization within species occurred inside the repeat. This might be regarded as a signature of an interlocus crossing-over event between alleles. A similar tendency was also observed for the UbB repeats. Repeat specific nucleotides were observed only for repeats 1 and 3, not for the intermediate repeat 2 (see Fig. 2A). Therefore, we believe that unequal crossing-over has a key role in the homogenization of UbC among human and great apes. However, gene conversion may also play some role. Homogenization seems to occur even among the UbC last repeats. If we examine variant sites for UbC shown in Fig. 2B, the last nine consecutive sites (205–228) are mostly specific to last repeats (C-Hu-9, C-Ch-10, C-Go-8, and C-Or-10). This pattern suggests that the very end of the last repeat escaped from homogenization, while the remaining part of the last repeat experienced some kind of homogenization. If so, this homogenization may have been caused by gene conversion, since repeat was not the unit of homogenization.

Recently, Nei et al. (2000) proposed the concept of birth-and-death evolution to explain the homogenization of multigene families, where a selective pressure has a major role in homogenization. Birth-and-death evolution is a form of evolution that assumes that new genes are created by repeated gene duplication and that some duplicate genes stay in the genome for a long time, whereas others are deleted or become nonfunctional. To distinguish concerted evolution from the birth-and-death evolution under strong purifying selection, the extent of synonymous differences within and between species was examined in this study. Intraspecific synonymous differences that are nearly as high as interspecific differences supports birth-and-death evolution. As described by Nei et al. (2000), concerted evolution can be distinguished from birth-and-death evolution by examining ubiquitin gene sequences from many pairs of closely related species. In our sequence comparison of UbC among closely related species, marked homogenization was observed in gorilla and the mean p S value (0.05) was considerably low, suggesting that birth-and-death evolution is less likely. When three ubiquitin repeat pairs, [2, 3], [4, 5], and [6, 7], of chimpanzee UbC were combined, the mean p S became 0.05. On the other hand, relatively high mean p S values were observed for orangutan UbC, suggesting that the concerted evolution is not so strong. Therefore, concerted evolution was more responsible for homogenization of UbC in gorilla and chimpanzee but not so in orangutan. It suggests that lineage-specific homogenization occurred in closely related species of hominoids. An interlocus crossing-over event could conceivably accelerate further crossing-over events so that the lineage specific interlocus crossing-over occurred at a relatively rapid pace. The lineage-specific concerted evolution as shown for UbC of closely related species executes a role in the evolution of polyubiquitin genes.