The complete primary structures of αD-2- and β-globin of hemoglobin D (Hb D) from the Aldabra giant tortoise, Geochelone gigantea, have been constructed by amino acid sequencing analysis in assistance with nucleotide sequencing analysis of PCR fragments amplified using degenerate oligo-nucleotide primers. Using computer-assisted sequence comparisons, the αD-2-globin shared a 92.0% sequence identity versus αD-globin of Geochelone carbonaria, a 75.2% versus αD-globin of Aves (Rhea americana) and a 62.4% versus αA-globin of Hb A expressed in adult red blood cells of Geochelone gigantea. Additionally, judging from their primary structures, an identical β-globin was common to the two hemoglobin components, Hb A and Hb D. The αD-2- and β-globin genes contained the three-exon and twointron configurations and showed the characteristic of all functional vertebrate hemoglobin genes except an abnormal GC dinucleotide instead of the invariant GT at the 5′ end of the second intron sequence. The introns of αD-2-globin gene were both small (224-bp/first intron, 227-bp/second intron) such that they were quite similar to those of adult α-type globins; the β-globin gene has one small intron (approximately 130-bp) and one large intron (approximately 1590-bp).
A phylogenetic tree constructed on primary structures of 7 αD-globins from Reptilia (4 species of turtles, 2 species of squamates, and 1 species of sphenodontids) and two embryonic α-like globins from Aves (Gullus gullus) and Mammals (Homo sapiens) showed the following results: (1) αD-globins except those of squamates were clustered, in which Sphenodon punctatus was a closer species to birds than turtles; (2) separation of the αA- and αD-globin genes occurred approximately 250 million years ago after the embryonic α-type globin-genes (π' and ζ) first split off from the ancestor of α-type globin gene family.
Amniota (reptiles, birds and mammals), in general, have two or more hemoglobin components (Brown and Ingram, 1974; Moss and Hamilton, 1974; Lawn et al., 1978; Efstratiadis et al., 1980; Bunn and Forget, 1986; Fushitani et al., 1996; Gorr et al, 1998) that are expressed according to the demands of different physiological conditions. Among them, hemoglobin D (Hb D) was first found in birds as a minor component of the embryonic and adult definitive erythrocytes (Hagopian and Ingram, 1971; Brown and Ingram, 1974). Based on functional studies of Hb D, the presence of αD-globin raises the oxygen affinity and might be one such adaptation of insufficient oxygen supply as observed in the embryonic stages (Dodgson et al., 1981; Chapman et al., 1982) or extreme hypoxic and even anoxic conditions (Rücknagel and Braunitzer, 1988). On the other hand, the primary structure of αD-globin of Hb D shows closely resemblance with embryonic hemoglobins (Chapman et al., 1982) and thus, the Hb D is of interest for the study of the molecular evolution of Amniota globins because the distribution of the αD-globin, to date, has been restricted in Aves and Reptilia (Rücknagel et al., 1984; Abbasi et al., 1988; Rücknagel et al., 1988; Matsuura et al., 1989; Fushitani et al., 1996; Gorr et al., 1998; Accession No. AF304335 in GenBank; Shishikura and Takami, 2001), except for Crocodilia (Leclercq et al., 1981; Leclercq et al., 1982). Most of the studies on globin gene structures have been carried out on birds and mammals (Bunn and Forget, 1986; Kleinschmidt and Sgouros, 1987), however, only one study has been conducted on reptilian αD-globin cDNA structure from globin mRNA isolated from the red blood cells present in the adult Geochelone carbonaria (Accession No. AF304335 in GenBank). In addition to adult αD-type globins, there are many genes related to α-globins such as embryonic α-like globins termed π'-globin (Chapman et al., 1980) for birds and ζ-globin (Aschauer et al., 1981) for mammals, all of which are important clues for understanding the molecular evolution of α- and α-related globins.
This study describes the primary structures of αD-2- and βD-globin of G. gigantea Hb D (hereafter the author uses β instead of βD because the primary structure of βD-globin prepared from Hb D was definitively shown to be identical when compared with that of β-globin prepared from the G. gigantea Hb A) in assistance with nucleotide sequences of the two globin genes of G. gigantea, and constructs a phylogenetic tree concerning the molecular evolution of αD-type globins. The tree also shows the relationships of α- and embryonic α-related globins, π'- and ζ-globin, as well as a few representatives of αA-type globins from vertebrates. This study first describes the genomic structures of globins amplified by PCR with degenerate primers, and then, the nucleotide sequences, to ascertain the amino acid sequences of αD-2- and β-globin. During the course of this study, it was also demonstrated that an identical β-globin was shared in both Hb A and Hb D as predicted in the previous study (Shishikura and Takami, 2001).
MATERIALS AND METHODS
Hb D from the Aldabra giant tortoise, G. gigantea, was prepared as described in the previous study (Shishikura and Takami, 2001).
Acetonitrile, ammonium sulfate, ammonium bicarbonate, tri-n-butyl phosphine, 4-vinyl pyridine and V8 protease (from Staphylococcus aureus strain V8) were purchased from Nakalai Tesque, Inc. (Kyoto, Japan). Separation columns, Alkyl Superose column HR5/5 and Resource column (prepackaged with 3 ml source 15 RPC gel matrix), were purchased and placed in a fast protein liquid chromatography (FPLC) system (Amersham Pharmacia Biotech, Upsala, Sweden). Lysyl endopeptidase (Achromobactor protease I) was purchased from Wako Pure Chemical Industries, Ltd. (Tokyo, Japan).
Taq DNA polymerase and GenElute Agarose Spin Columns were obtained from Sigma-Aldrich Co. (St. Louis, MO, USA). DNA molecular standard markers, pHY Marker (Takara Shuzo Co., Ltd., Shiga, Japan) and 100-bp DNA Ladder (New England Biolabs Inc., MA, USA) were used. Sequencing primers, M13 forward 17-mer (5′-GTA AAA CGA CGG CCA GT-3′) and PUC/M13 reverse 17-mer (5′-CAG GAA ACA GCT ATG AC-3′), were obtained from Sigma-Aldorich Co. and Promega Co. (Madison, WI, USA), respectively. A BigDye Terminator Cycle Sequencing Ready Reaction Kit was purchased from Perkin-Elmer Japan Co. Ltd (Tokyo, Japan).
All other chemicals and solvents used were the most purified grade commercially available.
The Hb D was modified by reduction and S-pyridylethylation (Friedman et al., 1970) and then directly applied on a reversed-phase column (Resource column), which had been equilibrated with a 0.1% TFA solution. Removal of unincorporated reagents bound on the Resource column could be achieved by washing with an excess amount of 0.1% TFA solution until the base line was below 0.05 at 280 nm. The globin-peptides were, then, eluted from the column by a linear gradient with 60% acetonitrile in 0.08% TFA. Flow rates were maintained at 0.5 ml/min. All fractions were monitored at 214 nm and 280 nm by a spectrophotometer (Model 116, Gilson).
Enzymatic digestion and peptide separation
Lysyl endopeptidase digestion was performed essentially with modifications of Jekel et al (1983), the details of which were previously described (Shishikura and Takami, 2001). To obtain overlapping peptides, the globin (about 10 nmoles) was digested with the V8 protease at a ratio of 1:100 (w/w, enzyme/substrate) for 48 hr at 37°C in a 0.1M Tris-HCl solution, pH 8.5 containing 1 M urea.
All peptides derived from the parent molecules were separated using a reversed-phased column, Resource, in a linear gradient with 60% acetonitrile in 0.08% TFA. Flow rates were maintained at 0.5 ml/min. All fractions were monitored at 214 nm and 280 nm by a spectrophotometer (Model 116, Gilson). When necessary, re-chromatography of selected peptides was performed as previously described (Shishikura et al., 1987).
Amino acid sequencing
Sequence analysis was performed using a Shimadzu gas phase protein sequencer, PPSQ-10, equipped with a PTH-10 amino acid analyzer (Shimadzu Co., Kyoto, Japan). Phenylthiohydantoin (PTH)-derivatives from the sequencer were separated and quantified. PTH-cysteine was detected as pyridylethylated-PTHcysteine, the elution point of which was determined as described in the manufacturer's manual.
Isolation of genomic DNA
Prior to DNA extraction, fixed-tissue samples (80–120 mg) in absolute alcohol were dissolved in 600 μl DNA extraction buffer (10 mM Tris, 10 mM EDTA, 150 mM NaCl, pH 8.0) in a micro-centrifuge tube to obtain wet forms. Samples were treated with SDS (final concentration: 0.4%) and proteinase K (final concentration: 20 mg/ml), mixed well, and incubated for 60 min at 55°C, followed by overnight incubation at 37°C. The extraction of DNA was performed by the procedure described by Sambrook et al. (1989) with minor alterations: two rounds of precipitation with ethanol and spooling the precipitate purified DNA. DNA was then resuspended in 1 ml of TE buffer (10 mM Tris/HCl buffer containing 1 mM EDTA, pH 8.0) and stored at 4°C: about 0.1 mg/ml of high-molecular-weight genomic DNA was obtained, as evaluated by the absorption spectrum and by 0.8% agarose gel electrophoresis.
Degenerate primers were designed based on the amino acid sequences of lysyl endopeptidase digested fragments of parent molecules. In order to sequence the PCR amplified fragments with a BigDye Terminator Cycle Sequencing Ready Reaction Kit, the degenerate oligo-nucleotide primers were tailed with the M13 forward or M 13 reverse sequencing primer tail (for the tail sequences shown above). A list of degenerate primers used in PCR amplifications is shown in Table 1.
Oligo-nucelotide primers used in this study
The PCR amplifications were performed in a 25-μl volume containing about 100 ng of genomic DNA template, 3 to 30 pmoles of each degenerate primer, deoxynucleotide triphosphates (400 μM) and 1.25 U of Taq DNA polymerase in the buffer conditions recommended by the manufacturer, 2.5 mM MgCl2. The reactions started with denaturation at 95°C for 3 min, followed by 45 cycles and ended with 7 min of extension at 72°C on a DNA Thermal Cycler 9700 (Perkin-Elmer, Norwalk, CT, USA). The first five cycle profile began with denaturation for 1 min at 95°C, 5-stepwise different annealing temperatures (65°C, 62.5°C, 60°C, 57.5°C and 55°C) for 10 sec each, and ended with elongation for 1min every cycle at 72°C. The thermal profile including denaturation of the first 5 cycles modified the procedures described by Sachadyn et al. (1998), Skantar and Carta (2000), and Don et al. (1991). The remaining cycles were programmed according to the method recommended by the manufacturer.
Agarose gel electrophoresis
A 1.5% agarose gel was used to examine the purity and the size range of the PCR products amplified from the Geochelone genomic DNA. In each lane, except lanes of DNA-markers, 10 μl of each of the amplified DNA samples were loaded. The two DNA molecular weight standard markers were used. The gel was run in TBE (Tris-Borate-EDTA) buffer at 110V for 50 min. The results were then recorded using a KODAK Electrophoresis Documentation and Analysis System 290 (EDAS 290), and analyzed by a 1D Image Analysis Software (v. 3.5.4; Eastman Kodak Co., Rochester, NY, USA).
Extraction of PCR products and nucleotide sequencing analysis
After trimming away excess agarose, the gel slices (<500 μg) containing the PCR products were placed into the GenElute Agar-ose Spin column and centrifuged for 10 min at 14,000 x g. The filtrate was concentrated by Microcon-100 and sequenced using the ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction Kit with the following modifications to the manufacturer's recommended protocol: 3.6 picomoles of M13 sequencing primer (forward or reverse) were annealed with about 32 ng of PCR product by mixing primer and template with 8 μl of Terminator Ready Reaction Mix in a final volume of 20 μl. This mixture was placed in a GeneAmp PCR system 9700 and subjected to cycle sequencing depending on the manufacturer's recommended protocol: start with heating for 10 sec at 96°C, and then 25 cycles of 96°C for 10 sec, 50°C for 5 sec, and 60°C for 4 min and reactions allowed to end with rapid thermal ramp at 4°C. Purifying extension products and the removal of unincorporated dye terminators in sequencing reactions were subjected to Centri-Sep spin columns (Princeton Separations P/N CS-901). Sequences of the PCR fragments were determined for both strands with the BigDye Primer Cycle Sequencing Ready Reaction Kit and the samples were on an ABI PRISM 310 Genetic Analyzer (Perkin-Elmer Japan Co. Ltd. Tokyo).
A multiple alignment program, Clustal W (Thompson et al., 1994), was used in the alignment of reptilian and other vertebrate's globin primary structures. Pair-wise distances among the globin sequences were analyzed using a computer program PROTDIST stored in the PHYLIP package (v. 3.51C; Felsenstein, 1993) under the Kimura-formula option. Based on the pair-wise distances, Neighbor-joining/UPGMA in NEIGHBOR (Felsenstein, 1993) was used to construct the phylogenetic tree of globins. Pair-wise alignments of DNA sequences were carried out using softwares of DNASIS as well as DNA Strider (V. 1.0.1).
RESULTS AND DISCUSSION
In a preceding paper (Shishikura and Takami, 2001) we have described the isolation of the two hemoglobin components of the Aldabra giant tortoise G. gitantea, in which the two were designated as Hb A and Hb D. The nomenclature of Hb A and Hb D was adopted in Ingram's laboratory (Hagopian and Ingram, 1971; Brown and Ingram, 1974) where the various domestic fowl hemoglobins were defined. Among them, the adult definitive erythrocytes contained the major adult hemoglobin (Hb A) and the minor definitive hemoglobin (Hb D). After establishing the complete amino acid sequences of globins as described below, the presence of Hb D in the Aldabra giant tortoise, G. gigantea, was completely confirmed when compared with the known primary structures of αD-globins (Kleinshmidt and Sgouros, 1987) specific to the Hb D. The advantage of modifying the protein by reduction and S-pyridylethylation also applied for separation of globin-constituents from the Hb D. As the results, three major fractions, α-1, α-2 and β in the order of elution, were separated as shown in Fig. 1: the two peaks, α-1 and α-2, were identical to each other having characteristics of αD-type globins so far sequenced until the first 20 N-terminal amino acid residues, but in contrast their chromatograms on reversed-phase column were shown as distinctly different. There might be sequence microheterogeneity of their primary structures as found in those of Hb A (Shishikura and Takami, 2001). Hence, the author first sequenced the α-2-globin from the two α-types of globins.
For establishing complete primary structures of α-2-and β-globin, two sequencing methodologies, protein and DNA sequencing, were carried out. First, the parent molecules and their peptide fragments were sequenced and aligned tentatively with the assistance of sequence similarities toward the known primary structures of reptilian α- and β-globins, in particular, those obtained from the G. gigantea Hb A (Shishikura and Takami, 2001). Fig. 2 shows the results of amino acid sequence analyses of αD-2- and β-globin. Appendix provides the data supporting the amino acid sequences in Fig. 2. The αD-2-globin chain was composed of 141 amino acid residues and the β-globin chain was composed of 146 residues. Two lysine-lysine residues appeared in positions 60-61 of αD-2-globin chain and 82–83 in β-globin chain were difficult to determine by analyzing the peptide fragments derived from digestions with lysyl endopeptidase. To complete the primary structure, peptide fragments containing the lysine-lysine residues generated by another enzymatic digestion such as V8 protease are required to be sequenced. This was done in the construction of αD-2-globin structure (Fig. 2, top) but required time-consuming work. To cope with time-consuming problems in determining primary structures, the following methods were used: (1) based on sequencing information of both intact globins and digested fragments, degenerate oligo-nucleotide primers were synthesized with a M13 forward or M13 reverse sequencing primer tail; (2) using these primers (forward and reverse), a target gene was amplified by PCR from genomic DNA as a template; (3) the PCR fragment was purified and sequenced by cycle sequencing with the M13 forward or reverse sequencing primer. Fig. 3 shows amplified fragments on agarose gel electrophoresis: An 870-bp fragment was generated from the PCR-amplification of genomic DNA using primers M13a-1 and M13a-2, assuming amplified complete coding regions (three exons) and interviewing regions (two introns), and the remaining, 480-bp fragment and 1.75-kbp fragment, were amplified using primer-sets of M13b-1/M13b-2 and M13b-3/M13b-4, respectively; (4) nucleotide sequences of the three PCR fragments were determined with manufactured M13 sequencing primers, the BigDye Terminator Cycle Sequencing Ready Reaction Kit and the ABI PRISM 310 Genetic Analyzer; (5) finally, both protein and DNA sequencing data were complementary combined to establish complete structures of the α-2-and β-globin chains of Geochelone Hb D. As shown in Fig. 2 and Table 2 (2A and 2B), the two primary structures reinforced each other by the two different methods.
Nucleotide sequences of three exons and exon-intron boundaries of aD-2-globin gene
Nucleotide sequences of three exons and exon-intron boundaries of b-globin gene
In comparison with the structural data of β-globin of Hb A (Shishikura and Takami, 2001), the primary structure of β-globin derived from the Hb D was completely identical, indicating that the β-globin was common in the construction of the two adult hemoglobin components, Hb A and Hb D. This finding supports the studies of Rücknagel and Braunitzer (1988) who described that the red blood cells shared the same β-globin chains in Hb A and Hb D. The sharing of identical β-globin chains has also been demonstrated in crocodiles (Leclercq et al., 1981; Leclercq et al., 1982), while lizards and snakes express two adult β-types of globins (Rücknagel et al., 1988; Matsuura et al., 1989; Abbasi and Braunitzer, 1991; Naqvi et al., 1994; Fushitani et al., 1996; Gorr et al., 1998). In this context, adult mammals (Braunitzer et al., 1961; Leclercq et al., 1981) and birds (Rücknagel et al., 1984; Oberthür et al, 1983; Oberthür et al., 1986) have been reported to have one kind of β-globin, but adult frogs (Knöchel et al., 1983; Patient et al., 1983) contained two subtypes of β-globin chains. Due to an inconsistency in the number of subtypes of adult β-type globin-chains among amphibians, reptiles, birds and mammals, reinvestigations are needed, especially, in regards to the evolution of Tetrapoda (Benton, 1990; Hardison, 1998).
Comparison of the primary structure of α-type globins within G. gigantea, αD-2-globin of G. gigantea differs from αA-globin in 53 amino acid residues (62.4% identity), but when compared with homologous globin chains found in adult Geochelone carbonaria (a different species of tortoises) and adult Rhea Americana (a species of birds), only 7 (95.0% idnetity) and 35 (75.2% identity) amino acid residues were substituted, respectively.
PCR amplification of globin gene by degenerate primers
Two degenerate oligo-nucleotides (M13a-1 and M13a-2 in Table 1) which were designed from the regions of N-terminal (8 amino acid residues in length) and C-terminal (8 amino acid residues in length) of αD-2-globin successively amplified a PCR-product with 870-bp estimated by migration distance on agarose gel electrophoresis (Fig. 3, lane 1). On the contrary, in the case of amplification of β-globin using M13b-1 and M13b-4 primers no product was observed on agarose gel electrophoresis, indicating that the whole coding region of β-globin gene was impossible to amplify at once using two degenerate primers designed by its N-terminal and C-terminal amino acid sequences. It seems to be difficult to amplify extremely long nucleotides such the case over 1.75-kbp PCR-fragment. Hence, several sets of sense and anti-sense degenerate primers were synthesized and used for amplification of β-globin gene in total with the genomic DNA: the two sets of sense and anti-sense primers (M13b-1/M13b-2 and M13b-3/M13b-4 shown in Table 1) produced a single fragment in each PCR, in which nucleotide-sized fragments were determined to be a 480-bp fragment and a 1.75-kbp fragment, respectively (Fig. 3, lane 2 and 3). Both products and the 870-bp fragment of αD-2-globin gene were sequenced from both sides and aligned by computer-assisted programs. Table 2A and 2B show alignments of nucleotide sequences in encompassing whole exon regions of the αD-2- and the β-globin gene and exonintron boundaries of the two genes. Breathnach and Chambon (1981) stated that there was no exception to the GT-AG rule according to which all intron sequences start with GT and end with AG. However, Table 2A shows that a unique structural feature of the αD-2-globin gene is a GC instead of a GT dinucleotide at the 5′ end of the second intron sequence. This finding is the first exception found in reptilian hemoglobin gene and supported the previous studies on gene structures of bird's hemoglobin (Erbil and Niessing, 1983; Dodgson and Engel, 1983). Erbil and Niessing (1983) found the T to C transition at the second intron position 2 of αD-globin gene from a duck, Cairina moschata. This evidence together with the unique structure of αD-2-globin gene found in the tortoise strongly indicates that the two animals, tortoises and birds, are the closest living relatives to each other.
When compared with the intron lengths among the four α-types of globin genes (Table 3), it was clearly determined that the Geochelone αD-2-globin gene structure corresponded to that of the adult chicken αD-globin gene and not to the embryonic chicken π'-globin gene (Engel et al., 1983) nor the embryonic human ζ-globin gene (Proudfoot et al., 1982). On the contrary, the G. gigantea β-globin gene was hard to classify since the second intron length (about 1.59-kbp) was large compared with those of the adult β-globin gene (Lawn et al., 1980; Dolan et al. 1983), including embryonic β-like globins (Efstratiadis et al., 1980; Chapman et al., 1981). In addition, the evolutionary relatedness of the intron sizes of the Geochelone globin genes to the other amniotes globin genes was defined for the first time.
Comparison of exson and intron sizes (in bp) of D -2- and -globin genes
Reptilian phylogeny and diversity based on αD-Globin structures
Shishikura and Takami (2001) have constructed a phylogenetic tree based on α- and β-globins of 28 reptilian Hb As, by which the molecular phylogeny of Reptilia is highly correlated at the level of orders with the traditional phylogeny established mainly upon their morphological characteristics (Carroll, 1969; Benton, 1990). To date, there have been four different types of α-globins in amniotes reported: αA, αD, π' and ζ. The former two are adult α-type globins and the remaining are embryonic α-like globins. Fig. 4. shows a molecular tree of reptilian evolution constructed mainly by αD-globins of 7 reptiles as well as relatedness among representatives of adult and embryonic α-type globins. The tree also strongly supports the previous molecular studies (Goodman et al., 1975; Fushitani et al., 1996; Gorr et al., 1998; Shishikura and Takami, 2001), however, it is reasonable to note the following two points: (1) the two kinds of embryonic globins, π' and ζ, first split off from the ancestor of the α-type of globins and formed a cluster; (2) the ancestor of squamates (snakes; L. miliaris, lizards; Varanus komodoensis) occupied unusual positions since αD-globins of squamates began to diverge approximately 335 million years ago, much earlier than the separation of the three other clusters of α-type globin families.
The author thanks the staff of veterinary surgeons of Animal Hospital of Osaka Municipal Tennoji Zoo for their cooperation in collecting blood from the Aldabra giant tortoise, G. gigantea. The author also thanks Dr. M. Esumi and Dr Y-H. Zhou, Department of Pathology, Nihon University School of Medicine, for her generosity (Dr. Esumi) in helping to prepare the equipment for analyzing amino acid sequences of globins and his instruction (Dr. Zhou) for analyses of nucleotide sequences under the ABI PRISM 310 Genetic Analyzer.