Molecular origin of the mosaic sequence arrangements of higher primate α-globin duplication units

Abstract

The human adult α-globin locus consists of three pairs of homology blocks (X, Y, and Z) interspersed with three nonhomology blocks (I, II, and III), and three Alu family repeats, Alu1, Alu2, and Alu3. It has been suggested that an ancient primate α-globin-containing unit was ancestral to the X, Y, and Z and the Alu1/Alu2 repeats. However, the evolutionary origin of the three nonhomologous blocks has remained obscure. We have now analyzed the sequence organization of the entire adult α-globin locus of gibbon (Hylobates lar). DNA segments homologous to human block I occur in both duplication units of the gibbon α-globin locus. Detailed interspecies sequence comparisons suggest that nonhomologous blocks I and II, as well as another sequence, IV, were all part of the ancestral α-globin-containing unit prior to its tandem duplication. However, sometime thereafter, block I was deleted from the human α1-globin-containing unit, and block II was also deleted from the α2-globin-containing unit in both human and gibbon. These were probably independent events both mediated by independent illegitimate recombination processes. Interestingly, the end points of these deletions coincide with potential insertion sites of Alu family repeats. These results suggest that the shaping of DNA segments in eukaryotic genomes involved the retroposition of repetitive DNA elements in conjunction with simple DNA recombination processes.

The adult α-globin gene locus appears to provide a model for the analysis of molecular evolution of tandem duplication units in higher primates. Sequence analyses have shown that the two human adult α-globin genes are contained within the 3′ portions of a pair of tandemly arranged duplication units which consist of three pairs of homology blocks X, Y, and Z (1–5). These homology blocks are interrupted by three nonhomology blocks, I, II, and III, with three of their boundaries precisely defined by the presence of Alu family repeats (see Fig. 1; refs. 4 and 5). It has been proposed that a tandem duplication that occurred prior to the divergence of higher primates generated two DNA segments with identical sequences (1–7). The exact homology between these duplication units was subsequently disrupted by nucleotide substitutions, as well as by DNA rearrangements resulting in a mosaic arrangement of the homology/nonhomology blocks (4, 5, 7). Segmental gene conversion, a process of nonreciprocal transfer of genetic information, may have been responsible for maintenance of the high degrees of sequence homology of the X and Z blocks (2–6).

Figure 1.

(a) Linkage maps of the human and gibbon adult α-globin locus. The range of maps spans from the pseudogene Ψα1 to immediately downstream of the α1-globin gene. Three pairs of homology blocks, X, Y, and Z, are indicated by open bars. Interspersed among the homology blocks are the nonhomology blocks I, II, III, and IV, and the different Alu family repeats indicated by the arrows. The numbers below the map represent the nucleotide positions of the ends of different blocks of the gibbon locus sequence as shown in b. X(α2), 1–1052; I(α2), 1368–1724; Y(α2), 2662–2831; Z(α2), 2832–4554; X(α1), 4555–5607; I(α1), 5608–5939 and 6239–6263; IV(α1), 6264–6444; II(α1), 6777–7103; Y(α1), 7104–7292; III(α1), 7293–7516; and Z(α1), 7517–9238. Note that the X blocks assigned in this diagram do not include the Alu family repeats, as previously defined (4, 5). (b) Nucleotide sequence alignment of the gibbon and human α-globin loci. The gibbon sequence (G) is shown in full, with the corresponding human sequence (H) aligned underneath it. The first nucleotide of the gibbon X(α2) block is denoted as number 1. Human nucleotides identical to those in the gibbon are indicated by dashes, and different nucleotides are indicated by the appropriate letters. Single-base deletions are left blank, while relative deletions longer than 1 bp are indicated by horizontal lines. The homology and nonhomology blocks are individually indicated by brackets. Alu family repeats are represented by horizontal line arrows with the short direct repeats flanking them boxed. The α2- and α1-globin genes are represented by horizontal open arrows. Not shown are the sequences of the α-globin genes (2, 3), the Alu family repeats (4, 5, 14), and the central parts of the blocks X, Y, and Z (15). The numbers of bp in the middle of the latter blocks indicate their total lengths, including the sequences shown. (c) Sequence alignment of the nonhomology blocks human I(α2), gibbon I(α2), and gibbon I(α1). The gibbon I(α1) sequence is shown in full, with those of the human and gibbon I(α2) blocks aligned underneath. The symbols used are the same as in b. (d) Sequence alignment of the human and gibbon Alu1 and Alu2 repeats. The sequence of gibbon Alu2 is shown in full, with those of the human Alu2 (4), human Alu1 (4), and gibbon Alu1 (15) aligned underneath. The pairwise homology comparisons of these four repeats are further described in Table 1.

The above evolutionary scenario is consistent with subsequent genomic mapping and DNA sequencing of the orthologous, or parallel, loci from the orangutan and two Old World monkeys, the olive baboon and the rhesus monkey (reviewed in ref. 7). Unfortunately, little information regarding the evolutionary origin of the nonhomology blocks I, II, and III is available from these analyses. One or more of these nonhomology blocks could have belonged to the ancestral DNA unit prior to its tandem duplication. Alternatively, they could be DNAs foreign to the adult α-globin locus, brought into it by various DNA recombination events. The gross organization of the α-globin duplication units of orangutan is nearly identical to that of human (8). Humans and orangutans diverged relatively recently, only 10–15 million years ago. The α2-globin unit of the Old World monkeys, whose lineage diveraged earlier from humans, is also similar to that of human (9). Unfortunately, a major portion of their α1-globin unit has been unclonable (9, 10).

More recently, we have cloned the adult α-globin locus from gibbon (Hylobates lar) (11), which started to diverge from either the great apes or the Old World monkeys 15–19 million years ago (12). As shown below in Results and Discussion, detailed sequence comparison between human and this nonhuman catarrhine has provided data regarding the evolutionary origin of the nonhomology DNA blocks, as well as several rearrangement processes that appear to have played a major role in the shaping of the adult α-globin loci in higher primates.

MATERIALS AND METHODS

Both the molecular cloning and DNA sequencing of the gibbon α-globin locus by the chain-termination method (13) have been described in detail elsewhere (14, 15).

RESULTS AND DISCUSSION

The linkage maps of the adult α-globin loci of human and gibbon are shown in Fig. 1a. The nucleotide sequences of the gibbon locus from the X(α2) block to downstream of the α1-globin gene are aligned with the orthologous human region in Fig. 1b for comparison. The gibbon locus is similar to the human locus in that it also consists of three pairs of homology blocks, X(α2)/X(α1), Y(α2)/Y(α1), and Z(α2)/Z(α1). As in human and Old World monkeys, the nonhomology block I(α2) (nucleotides 1368–1724, Fig. 1b) and an Alu family repeat (Alu3, nucleotides 1739–2661, Fig. 1b) are also present between the gibbon X(α2) and Y(α2) blocks. As shown previously, the gibbon Alu3 is a triplet Alu sequence, while doublet and singlet Alu repeats were found at the parallel positions in human and rhesus, respectively (Fig. 1; refs. 4, 9, 14). These doublet and triplet Alu sequences are believed to be the result of sequential insertion of singlet Alu family repeats, at the I(α2)/Y(α2) junctions, during primate evolution (14).

The organization of the gibbon α1-globin unit is similar to that of human in the relative arrangement of the blocks X(α1), II(α1), Y(α1), III, and Z(α1) (Fig. 1a). However, close examination of the two sequences reveals several unique features of the gibbon locus. First, an array of DNA sequences, nonhomology I(α1) (nucleotides 5608–5939 and 6239–6263, Fig. 1b), occurs immediately downstream of the gibbon X(α1). This gibbon I(α1) region is homologous in sequence to both the human and gibbon I(α2) blocks (Fig. 1c), but it is interrupted by an Alu repeat (Alu4, nucleotides 5940-6223, Fig. 1b) inserted into its 3′ portion. Second, two sequences, IV and Alu2, at nucleotide positions 6264–6444 and 6479–6776, respectively (Fig. 1b) are located between the gibbon I(α1) and II(α1) blocks. Block IV does not exist in the α2-globin unit of either human or gibbon, and it is followed by Alu2, an Alu family repeat exhibiting a 93% sequence identity to the human Alu2 repeat. This sequence identity is significantly higher compared with those of the three pairwise comparisons: gibbon Alu2/gibbon Alu1 (85%), gibbon Alu2/human Alu1 (86%), or human Alu1/human Alu2 (86%) (Table 1). Furthermore, the gibbon Alu2/human Alu2 pair has the highest number of common sites (Table 1). These sequence comparisons suggest that gibbon Alu2, which apparently has been inserted in between blocks II(α1) and IV(α1) with the sequence 5′-G_C^TCCCCACATCTCA-3′ being the site of integration, is most likely the parallel element, or orthologue, of human Alu2. Following this, human Alu1 and Alu2 repeats could not be paralogous as postulated previously (4, 5).

Table 1.

Pairwise comparison of Alu1 and Alu2 repeats of gibbon and human

Sequences compared	No. of differences	Common sites	% identity
G-Alu1/H-Alu1	24	19	92
G-Alu2/H-Alu2	21	20	93
G-Alu1/G-Alu2	42	1	85
H-Alu1/G-Alu2	40	3	86
G-Alu1/H-Alu2	38	3	87
H-Alu1/H-Alu2	39	2	86

Pairwise comparisons of the sequences among the four Alu family repeats were made to determine their evolutionary relationship. The percent identity values were calculated according to the sequence alignment of Fig. 1d. A common site is where the two compared Alu repeat sequences have the same nucleotide, but the other two Alu repeats have different base(s). G, gibbon; H, human. 

We propose the following evolutionary scenario for the α2- and α1-globin units: (i) The sequence organization of both α-globin duplication units in ancient higher primates was at one time similar to the present-day gibbon α1-globin unit except for the Alu2 repeat; (ii) subsequently, prior to the divergence of human and gibbon, Alu1 was inserted at the X(α2)/I(α2) junction; (iii) Alu2 was inserted at the IV(α1)/II(α1) junction; (iv) an ancestral singlet Alu3 was inserted at the 5′ boundary of the Y(α2) block (Fig. 2); and (v) gross sequence arrangement of this ancient α1-globin unit has since been conserved in the gibbon lineage. But during evolution of humans, DNA recombination has occurred between the A+T-rich sequence at the 3′ end of X(α1) block and the very 5′ end of the Alu2 repeat (Figs. 2 and 3). This recombination process, presumably facilitated by the homology near the recombination junctions (Fig. 3b), led to the deletion of blocks I(α1) and IV(α1), and it also brought the Alu2 repeat to a position immediately downstream from the X(α1) block (Fig. 3a). This scenario is consistent with the sequence comparison of the Alu1 and Alu2 repeats (Table 1), and it also provides an explanation for our previous finding. That is, human Alu2, despite its existence at a genomic position paralogous to the human Alu1, is flanked only at its 5′, and not the 3′, side with the sequence 5′-CTAAAATCC-3′ that has served as the insertion site for the human Alu1 repeat.

Figure 2.

Evolutionary shaping of the human α2- α1-globin duplication units by retroposition of Alu family repeats and by simple DNA deletion events. The proposed scheme of evolution has been deduced from sequence comparison of the duplication units between human and gibbon (see text for more details). The ancestral unit consisted of blocks X, I, IV, II, Y, and Z. The evolutionary origin of block III is uncertain at the present time. It could belong to the ancestral unit, but was deleted from the α2-, but not α1-, globin unit. Alternatively, it could have been inserted into the α1-globin unit after the duplication event. After tandem duplication, independent retroposition of different Alu family repeats into the adult α-globin locus occurred. Of the four depicted repeats, Alu1, Alu2, and Alu3 were inserted in the human/gibbon ancestor, with the insertion timing of Alu3 as early as prior to the separation of gibbon/human from the Old World monkeys (9, 14). Although Alu4 is present in the gibbon lineage, whether it was inserted prior to or after the human/gibbon divergence is unknown, since the DNA region containing its insertion site has been deleted from the human genome.

Figure 3.

(a) Diagram of DNA recombination process deleting blocks I(α1) and IV(α1) from the human α1-globin unit. The proposed recombination is accomplished by DNA breakage and rejoining of the X(α1)/I(α1) junction and the IV(α1)/Alu2 junction. See text for more details. (b) Sequence alignment of the above two junctions of gibbon α1-globin unit and DNA at the human X(α1)/Alu2 junction. The tandemly arranged repeats of 5′-CAAAA-3′ (horizontal arrows) are likely generated by a series of amplifications of the short sequence CA and/or CAA, presumably only in the gibbon lineage and not in the human (4). The short direct repeats, or insertion sites, of human Alu1 and gibbon Alu2 are boxed. The open triangles denote the boundaries of adjacent blocks around which the hypothesized breakage and rejoining processes have occurred. The nucleotide positions, as used in Fig. 1b, of the first and last base of each sequence are indicated by the numbers.

Alternatively, one could propose that an ancient Alu1 repeat was inserted at the 3′ end of the X block prior to duplication of the α-globin unit. Subsequently in the human lineage, DNA deletion occurred by homologous recombination between the Alu repeat at 3′ end of the X(α1) block and an Alu repeat at the IV/II junction. This would also produce the arrangement now seen in the human α1-globin unit (Fig. 1). However, neither Old World Monkeys (9) nor gibbons (Fig. 1) have an Alu repeat at the 3′ ends of their X(α1) blocks. For this latter model to be viable then, one has to hypothesize that tandem duplication of the α-globin unit and/or deletion of certain Alu repeats occurred independently in different lineages during evolution of higher primates.

In the α2-globin unit, on the other hand, it is most likely that a DNA recombination event occurred between the junction of blocks I(α2) and IV(α2), and the junction of blocks II(α2) and Y(α2) (Figs. 2 and 4a) prior to the separation of the gibbon/human lineage and the Old World monkeys. Interestingly, the DNA substrates near the breakage–rejoining points of this illegitimate recombination process also exhibit quite extensive sequence homology, 53% for perfect base-pairing and 75% if purine–purine and pyrimidine–pyrimidine homologies are accounted (Fig. 4b). It should be noted that the timing of this deletion event relative to the insertion of the Alu3 repeat at the very 5′ end of the Y(α2) block is unknown. However, it remains a viable possibility that the two events occurred at the same time during evolution.

Figure 4.

(a) Diagram of the DNA recombination process deleting blocks IV and II of the human and gibbon α2-globin units. The breakage and rejoining occurred at the I(α2)/IV(α2) and II(α2)/Y(α2) junctions. (b) Sequence alignment of the above two recombination end points of gibbon, and DNA sequences of the I(α2)/Y(α2) junctions of both gibbon and human. For simplicity of comparison, the sequences of the gibbon and human Alu3 repeat (4, 14) are omitted, but their sites of insertion are still shown (boxed). Again, the open triangles denote the boundaries of adjacent blocks around which breakage and rejoining have occurred.

These proposed DNA deletion events and the insertions of various Alu family repeats during evolution of the adult α2-/α1-globin duplication units in higher primates are summarized in Fig. 2. Despite the still unknown origin of the nonhomology block III, the model of Fig. 2 appears to provide a satisfactory explanation for the generation of the present-day mosaic organization of this locus in the human and gibbon lineages. Several interesting genetic implications emerge from this model. It seems reasonable to suggest that all eukaryotic genomes arose from series of tandem duplication events (refs. 16 and 17, and references therein), such as the one generating the ancestor of the adult α-globin gene duplication units in higher primates. Sequences of originally identical tandem duplication units then diverged due to nucleotide substitutions (reviewed in ref. 18), which could be counterbalanced by the processes of gene correction (see ref. 19 for references). The eukaryotic genomes could have been further shaped by the expansion through insertions of various DNA sequences including the repetitive DNA elements, and by down-sizing events such as DNA deletions (20). In the scenario of Fig. 2, we see no evidence for insertion processes that would bring unique DNA sequence(s) into the α-globin locus from other parts of the primate genomes. Instead, deletion of unique DNA sequences belonging to the original duplication units and retroposition (21, 22) of repetitive DNA elements—e.g., the Alu family repeats (23)—appear to be the two major mechanisms for shaping of the adult α-globin duplication units during higher primate evolution. Whether this is generally true for the evolution of most eukaryotic genomes is an intriguing question. Furthermore, the above two DNA rearrangement processes are probably tightly linked to each other in the evolution of eukaryotic genomes, a situation similar to some transposon-induced DNA rearrangements in bacteria, yeast, and plant cells (for examples, see refs. 24–26).

The scenario of Fig. 2 also reinforces the previous observation that potential insertion sites of repetitive DNA elements such as the Alu family repeats are also hot spots for other DNA recombination processes (refs. 4 and 27–32 and references therein). Three of the four recombination junctions depicted in Figs. 3 and 4 apparently functioned as the insertion sites for Alu family repeats at one time or another during genomic evolution of the higher primates. The fact that some sequences could be utilized both as insertion sites for Alu family repeats (Fig. 1b; ref. 4), and independently as the substrates for illegitimate recombination (Fig. 3) further suggests that it is an intrinsic property of these genomic sequences, not the presence of repetitive DNA elements per se, that renders them hot spots of various DNA recombination processes. Among the possibilities is the recognition of these sequences by DNA recombinases and/or recombination complexes. Interestingly, our evolutionary analysis of the α-globin loci of higher primates is similar to the recent findings in yeast cells that retrotransposons preferentially insert at, and heal, chromosomal breaks (33–35).