The monomer subunits of giant extracellular hemoglobins from earthworms Pontodrilus matsushimensis and Pheretima communissima that belong to the family Megascolecidae, Oligochaeta, were purified by a reversed-phase column, Resource RPC, and sequenced. The complete amino acid sequences of the two monomeric globin chains were determined: 141 amino acid residues with a molecular weight of 16,366 Da for Pontodrilus matsushimensis and 140 amino acid residues with a molecular weight of 16,000 for Pheretima communissima, respectively. The Pontodrilus matsushimensis monomer globin has three cysteine residues, and the two located at positions 2 and 131 are conserved as those observed in all annelids and contribute to form a disulfide-bonded interchain. The third cysteine residue at position 73 is the first evidence for the annelid monomer globin subunits. The physiological functions of the third cysteine residue, however, are still unknown. The monomer sequences of the two species were aligned with those of five known sequences from annelids, including a polychaete, Tylorrhynchus heterochaetus, and four oligochaetes, Pheretima hilgendorfi, Pheretima sieboldi, Lumbricus terrestris and Tubifex tubifex. Using computer analysis, a 87.9% identity of the amino acid sequences between two monomeric subunits of Pheretima communissima and Pheretima hilgendorfi hemoglobins showed the highest degree of sequence similarity. A molecular phylogenetic tree of seven species of annelids has constructed, suggesting that the divergence times among the three species of Pheretima and between Pheretima and Pontodrilus were 50 to 100 and about 209 million years ago, respectively.
INTRODUCTION
Annelid hemoglobin is a giant extracellular protein comprised of two main heme-containing subunits, a monomer and a disulfide-bonded trimer, and a linker including two or three chains (Vinogradov, 1985). The four constituent chains of two main heme-containing subunits that are classified into two strains of globins, strain A and strain B, by Gotoh et al. (1987) are akin to vertebrate α- and β-globin. To date, five monomeric globins that belong to the strain A have been purified from annelid hemoglobins whose amino acid sequences have been established: chain I of Tylorrhynchus heterochaetus by Suzuki et al. (1982), chain I of Lumbricus terrestris by Shishikura et al. (1987), chain I of Pheretima sieboldi by Suzuki (1989), chain I of Tubifex tubifex by Stern et al. (1990) and the monomer chain of Pheretima hilgendorfi by Shishikura (1996). Compared with all known sequences of annelid hemoglobins described elsewhere (Kapp et al., 1995), hemoglobin molecules sharing an orthologous relationship, such as a lineage of monomer globin chains in strain A, are good materials for analyzing the molecular evolution of proteins and genetic divergence patterns among annelids (Shishikura, 1996).
The Pheretima group of genera, family Megascolecidae, contains over 750 nominal species from southern and eastern mainland of Asia and is a dominant member of the earthworm fauna (Easton, 1984). In Japan, there are as many as 155 species of the genus of Pheretima (Kobayashi, 1941) including such abundant species as Pheretima hilgendorfi and Pheretima communissima in the central and northern regions. A previous study (Shishikura, 1996) showed that a 81.6% identity of the amino acid sequence was found between the two monomeric globin chains from the Pheretima species, Pheretima hilgendorfi and Pheretima sieboldi. The value, as expected, was the highest degree of sequence similarity among known annelid sequences (Goodman et al., 1988; Vinogradov et al., 1993; Kapp et al., 1995). Hence, here we aimed to sequence another monomer globin chains of the member of Pheretima. Consequently, only 17 amino acid residues out of 140 in total were different between Pheretima hilgendorfi and Pheretima communissima, which accounts for a 87.9% identity.
On the contrary, Pontodrilus matsushimensis also belongs to the family of Megascolecidae (Jamieson, 1971,1988). The genus Pontodrilus is a terrestrial earthworm, normally found burrowing in sand around the high tide mark of sea shores (lizuka, 1897). These worms have a wide tolerance to the salinity and may be different from common terrestrial earthworms in their osmoregulatory functions (Takeuchi, 1980a, b). Such internal conditions may cause adaptation of protein functions involving amino acid substitutions of primary structure (Perutz, 1984; Clementi et al., 1994). Therefore, we also aimed to establish the amino acid sequence of the monomer subunit of Pontodrilus matsushimensis hemoglobin and compared its primary structure with known annelid structures. Here we provide significant evidence concerning the Pontodrilus monomer globin sequence: The monomer globin chain has three cysteine residues, and the two located at positions 2 and 131 might be responsible for linking together by a disulfide bond, while the other at position 73 is noteworthy in annelid monomer globin subunits but its biological function is still unknown.
In this study we also describe a convenient method for separating individual chains from the whole hemoglobin. A reversed-phase column, Resource RPC, recently introduced by Pharmacia Biotech., showed very powerful globin chain resolution with high yield, in particular for relatively larger molecules such as monomer and trimer subunits of hemoglobin constituents. The two monomer globins of Megascolecidae were thus obtained and sequenced.
A phylogenetic tree was constructed for seven known monomer sequences of annelid hemoglobins, supporting the traditional relationship among these annelids.
MATERIALS AND METHODS
Materials
Live earthworms Pheretima communissima were collected in the field of the Nihon University School of Medicine and Pontodrilus matsushimensis collected from high tide of Matsushima Bay were kindly donated by Dr. N. Takeuchi, Miyagi College of Education, Sendai, Japan. Acetonitrile, p-amidinophenylmethanesulfonyl fluorideHCI, tri-n-butyl phosphine, and 4-vinylpyridine were purchased from Nacalai Tesque (Kyoto, Japan). Methylethylketone, trifluoroacetic acid (TFA), and 1-L-tosylamide-2-phenylethyl chloromethyl ketone treated trypsin from bovine pancreases (type XIII; 10,000-13,000 BAEE units per mg protein) were obtained from Sigma Chemical Co. (St. Louis, MO, USA). Lysyl endopeptidase (Achromobactor protease I) was purchased from Wako Pure Chemicals Co. (Tokyo, Japan). A Bradford's protein assay kit was purchased from Nippon Bio-Rad Laboratories KK (Tokyo, Japan). Separation columns, Resource RPC (3 ml packed with Source 15 RPC gel matrix) and PepRPC HR 5/5 placed in a Fast Protein Liquid Chromatography (FPLC), were purchased from Pharmacia Biotech. (Upsala, Sweden).
Preparation of hemoglobin
Samples of hemoglobin were prepared as previously described (Shlom and Vinogradov, 1973; Shishikura, 1996). The hemoglobin was converted to the cyanomet form by adding 2% K3Fe(CN)6, 0.5% KCN, and 0.1% NaHC03 (Moss and Ingram, 1968). Heme was removed from the hemoglobin using methylethylketone as described (Teale, 1959).
Purification of monomer globin subunits
Each of the monomer globin subunits of Pontodrilus matsushimensis and Pheretima communissima was separated and purified by reversed-phase chromatography on Resource RPC in a 0.1% TFA buffered gradient to 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).
SDS-polyacrylamide gel electrophoresis (PAGE)
SDS-PAGE was conducted on a mini-slab gel apparatus, Mighty Small II (Hoefer Scientific Instruments, San Francisco, CA, USA) according to the instructions of the manufacturer with slight modification of Laemmli's (Laemmli, 1970) procedure. Molecular weight markers were: 66 kDa (bovine serum albumin), 45 kDa (ovalbumin), 29 kDa (carbonic anhydrase) and 18.4 kDa (β-lactoglobulin).
Protein concentration
Bradford's (1976) protein assay was performed according to the micro-assay procedure described in the Bio-Rad technical bulletin attached to the kit, using γ-globulin as the standard. Protein fractions eluted from reversed-phase columns were lyophilized and then redissolved in 50 mM ammonium bicarbonate buffer, pH 8.9. A 10 μl of redissolved sample was measured by a spectrophotometer Uvidec 340Q (Jasco Co., Tokyo, Japan).
Protein modification
Reduction and S-pyridylethylation of globins were performed by methods described elsewhere (Friedman et al., 1970; Shishikura, 1996). Finally, the remaining reagents in the reaction mixture were completely removed from the sample by reversed-phase chromatography on a Resource RPC column.
Enzymatic cleavages
Lysyl endopeptidase digestion was performed essentially as described (Jekel et al., 1983), and was detailed in the previous study (Shishikura, 1996). Large peptides derived from the intact globin by the digestion of lysyl endopeptidase were digested further with trypsin at an enzyme/substrate ratio of 1/200 (by mass) at 37°C for 8 hr in 0.1 M Tris-HCI, pH 8.0, containing 2 M urea.
Peptide separation
All peptides derived from monomer globin as well as peptides from the lysyl endopeptidase digestion were separated using a PepRPC HR 5/5 column in a 0.1% TFA buffered gradient to 60% acetonitrile in 0.08% TFA. Flow rates were maintained at 0.3 ml/min. All fractions were monitored at 214 nm and 280 nm. Rechromatographies of selected peptides were performed as previously described (Shishikura et al., 1987).
Amino acid sequence determination
Sequence analysis was performed using a gas phase protein sequencer, PPSQ-10 (Shimadzu Co., Kyoto, Japan), equipped with a PTH-10 amino acid analyzer. PTH-derivatives from the sequencer were separated and quantitated.
Computer analysis of aligned sequences
Sequences were aligned by a multiple alignment program, Clustal W (Thompson et al., 1994), which can be found on the home page of the Baylor College of Medicine (curator; Dr. Smith RF, Human Genome Center, Baylor College of Medicine). Pairwise distances and a rooted tree with branch lengths among seven monomer sequences were analyzed using the computer program Protdist under the Dayhoff PAM matrix option, Neighbor under the unweighted pair group method with arithmetic (UPGMA) means and Neighbor-Joining option of the PHYLIP package (Felsenstein, 1993). Other computations were performed using the DNASIS programs (NBRF-Protein Sequence Database, version 23, 1995) of Hitachi Software Engineering Co., Ltd. (Yokohama, Japan).
RESULTS
Chain separation
Figure 1A shows a typical chain separation pattern of the whole hemoglobin of Pontodrilus matsushimensis on a reversed-phase column (Resource RPC), whose SDS-PAGE patterns are shown in Fig. 2A, B. Similarly, Fig. 1B and Fig. 3A, B show a typical separation pattern of whole hemoglobin of Pheretima communissima and SDS-PAGE patterns of fractions eluted from the column, respectively. In both species, the subunit constituents are: monomer (Fig. 2A4, B4 for Pontodrilus, Fig. 3A5, B5 for Pheretima), trimer (Fig. 2A5, B5 for Pontodrilus, Fig. 3A7, B7 for Pheretima) and several protein bands including linker chains (Fig. 2A1, 2, 3, B1, 2, 3 for Pontodrilus, Fig. 3A1, 2, 3, 4, B1, 2, 3, 4 for Pheretima). Both FPLC chromatograms and SDS-PAGE analyses also show two monomer components from both Pontodrilus matsushimensis and Pheretima communissima. The components in the peaks indicated by arrows are identified to be monomer subunits by SDS-PAGE analyses under reduced and non-reducued conditions. One of the minor components of Pheretima communissima monomer globins is shown in Figs. 3A6 and 3B6 but others are not shown here. The recoveries of the first eluted monomer globins (indicated by bars in Fig. 1) were as much as 520 μg (20.8%) for Pontodrilus matsushimensis and 480 μg (19.2%) for Pheretima communissima, when 2.5 mg of each hemoglobin sample was applied.
Peptide mapping of the two monomeric globins
The major monomer globins from Pontodrilus matsushimensis and Pheretima communissima were reduced and S-pyridylethylated and were then digested, separately, by lysyl endopeptidase. Figures 4A and 4B show their FPLC patterns on a PepRPC HR 5/5 column. These lysyl endopeptidase digested (K−) peptides of Pontodrilus (Fig. 4A) and Pheretima (Fig. 4B) were designated by considering the order of elution, position and homology in the sequences of the two species; for example, Pontodrilus matsushimensis (Pm) K-6 peptide and Pheretima communissima (Pc) K-6 peptide indicate the last eluted peaks sharing sequence homology and the peptide positions of the two species. Definitive differences between Pontodrilus matsushimensis and Pheretima communissima on the chromatograms were observed: One peak with a retention time of 60.9 min (PmK-5 peptide) in Pontodrilus disappeared in Pheretima. In contrast, two different peaks with retention times of 18.8 min (PcK-5a peptide) and 36.2 min (PcK-5b peptide) emerged in Pheretima. The elution order of K-2 and K-3 peptides in Pontodrilus appeared in reverse order in Pheretima.
Sequencing and alignment
The amino acid sequences of monomer globin chains of Pontodrilus matsushimensis and Pheretima communissima were determined as follows; their reduced and S-pyridylethylated heme deprived proteins were digested with lysyl endopeptidase and separated by reversed-phase chromatography whose elution pattern was depicted above (Fig. 4). The lysyl endopeptidase digested peptides were sequenced by a Shimadzu amino acid sequencer PPSQ-10. All the K-peptides except for PmK-6 and PcK-6 were sequenced completely. The rests were digested further by trypsin and separated by reversed-phase chromatography on a Pep RPC column (data not shown), followed by sequencing. Sequence similarities to the monomer globin of Pheretima hilgendorfi previously determined (Shishikura, 1996) as well as to those of annelid hemoglobins were used to determine the position of sequenced peptides. Only one small peptide derived from the monomer globin of Pheretima communissima could not be found in the elution of lysyl endopeptidase digestion peptides, however, its amino acid sequences, Val10-Lys11, were determined by the analysis of the first 20 amino-terminal amino acid sequences. The peptide fragment containing the third cysteine residue in Pontodrilus monomer subunit was analyzed in triplicate. The results are shown in Fig. 5 with the five known sequenced monomers of annelids, including Tylorrhynchus heterochaetus (Suzuki et al., 1982), Lumbricus terrestris (Shishikura et al., 1987), Pheretima sieboldi (Suzuki, 1989), Tubifex tubifex (Stern et al., 1990) and Pheretima hilgendorfi (Shishikura, 1996). The Pontodrilus matsushimensis monomer subunit constitutes of 141 amino acid residues, containing three cysteine residues, and has a calculated molecular mass of 16,366 Da. The Pheretima communissima monomer subunit has 140 amino acid residues, containing two cysteine residues, and has a calculated molecular mass of 16,000 Da. There are 82 positions of identity among these four monomeric globins of the family Megascolecidae. There are also 26 invariant positions in this grouping of seven globins of annelids. When these globins are compared with the Pheretima communissima, there are 123 identities (87.9%) in Pheretima hilgendorfi, 121 identities (85.8%) in Pheretima sieboldi, 103 identities in Pontodrilus (66.5%), 82 identities (56.2%) in Lumbricus, 78 identities (51.3%) in Tubifex and 70 identities (45.8%) in Tylorrhynchus. A matrix of amino acid identities among the rest was also developed (Table 1). Figure 6 shows a phylogenetic tree of seven species of annelids based on the amino acid sequence analyses of their monomeric globins.
Table 1.
Pair wise distances between the complete amino acid sequences of monomer subunits of Pontodrilus, three species of Pheretima, Lumbricus, Tubifex and Tylorrhynchus hemoglobins
DISCUSSION
Many attempts have been carried out to separate and purify individual chains of annelid hemoglobins: gel filtration (Shlom and Vinogradov, 1973; Suzuki et al., 1983b), DEAEion exchange column chromatography (Suzuki et al., 1983a), preparative gel electrophoresis (Ochiai and Enoki, 1979), chromatofocusing (Gotoh, 1982; Shishikura et al., 1986), reversed-phase column chromatography (Fushitani et al., 1988; Suzuki et al., 1990a; Ownby et al., 1993) and others (Bonaventura and Bonaventura, 1981). Our separation method described here had a similar efficiency as previously obtained in separation of Lumbricus hemoglobin by Ownby et al. (1993) and Lamellibrachia hemoglobin by Suzuki et al. (1990b), who used reversed-phase columns, Synchropak RP-P C18 and Cosmosil 5C18-300, respectively. However, several advantages still exist: Resource RPC has a large quantity of sample capacity with high chemical stability, gives excellent resolution in separation at high flow rates with low back pressure (versus flow rates), and can be run in pH ranging from 1-12. In this study, amounts as large as 520 μg (20.8%) and 480 μg (19.2%) for monomer subunits and 1.16 mg (46.4%) and 1.56 mg (62.4%) for trimer subunits of Pontodrilus matsushimensis and Pheretima communissima, respectively, were obtained from 2.5 mg of each whole hemoglobin sample.
According to the SDS-PAGE analyses under reduced and unreduced conditions, the Pontodrilus monomer subunit which migrated slower than any of the constituent polypeptide chains of the trimer subunit (Fig. 2B, lane 5) was unique feature because corresponding subunits of the other annelid hemoglobins, in general, migrated faster than the three chains of trimer subunit (Vinogradov, 1985). Furthermore, an apparent molecular weight of the Pontodrilus monomer subunit was estimated to be about 18,400 Da on SDS-PAGE. After sequenced, however, the molecular mass of the monomer subunit was determined to be 16,366 Da. This value was in good agreement with those reported on monomer chains of Megascolecidae (Suzuki, 1989; Shishikura, 1996), including that of Pheretima communissima. The value obtained by SDS-PAGE analysis is, therefore, misunderstanding and probably due to dissimilarity in molecular shape and apparent net charge of SDS-treated Pontodrilus monomer subunit. Figures 1, 2 and 3 also indicated that heterogeneities of monomer globin subunits existed in both Pontodrilus matsushimensis and Pheretima hilgendorfi hemoglobin as demonstrated previously in Lumbricus terrestris hemoglobin (Shishikura et al., 1987; Ownby et al., 1993) as well as Neanthes diversicolor hemoglobin (Suzuki et al., 1994). In this experiment, only the major fractions of both hemoglobin monomers were purified and sequenced.
One of the most striking features of sequenced monomers is that the Pontodrilus monomer globin has been shown to have three cysteine residues at positions of 2, 73 and 131. In annelids, the globin chains belonging to the member of strain A have a conserved characteristic containing two cysteine residues located in the amino- and carboxyl-termini. The two cysteine residues in both species apparently form a disulfide-bonded interchain as already reported in the known sequences of monomer subunits of annelid hemoglobins. Therefore, the third cysteine residue at position 73 may be free and is a conspicuous characteristic of monomer globin of Pontodrilus matsushimensis. This is the first observation of the presence of the third cysteine residue in the monomer subunits of annelid hemoglobins. Recently, Suzuki et al. (1990b) demonstrated the presence of three cysteine residues in the constituent polypeptide chain (AIII) of deep-sea tube-worm Lamellibrachia (phylum Vestimentifera) hemoglobin. They showed that the third cysteine residue at position 74 in Lamellibrachia hemoglobin is responsible for sulphide-binding ability because the hemoglobin may have been acquired high-molecular adaptation to transport enough sulphide to internal sulphide-oxidizing bacterial symbionts (Childress et al., 1987). It is, however, difficult to directly adjust their discussion to explain the function of the third cysteine residue in Pontodrilus hemoglobin because Pontodrilus matsushimensis is only found in sea shore environments where the animals live and spend their time 50 cm underground in winter (Dr. N. Takeuchi, personal communication). In this respect of their adaptations to hypoxic conditions, it might be said that they have developed particular mechanisms for maintaining an adequate 02 supply to tissues. Although there is a lack of data on globin structures, in particular on aquatic or marine oligochaetes with tolerance to a wide range of salinity, it is noteworthy that the Pontodrilus hemoglobin has a distinct similarity in the amino acid sequences compared with the deep-sea tube worm Lamellibrachia. Hence, it would be interesting to investigate the biochemical meanings of the third cysteine residue in the Pontodrilus hemoglobin for the molecular evolution of hemoglobins as well as physiology of invertebrate hemoglobins.
Previous works (Gotoh et al., 1987; Shishikura, 1996) suggested that homologous subunits sharing an orthologous relationship in the hemoglobin molecules are useful clues for analyzing the genetic relationship of earthworms. Table 1 lists the percentage differences and protein distances between complete amino acid sequences of Pheretima and Pontodrilus compared with the known sequences of oligochaetes and polychaetes. The identical scores (81.6%-87.9%) among three species of Pheretima showed very high degrees of similarity and might provide evidence biochemical support for the previous morphological study of Easton (1984) who grouped several species of Pheretima, including above three species into Amynthas hilgendorfi species-complex. Recently, Boore and Brown (1995) sequenced the complete mitochondrial DNA of earthworms (from Lumbricus terrestris). It is also hoped that it may be used for analyzing genetic relationships among these species as well as annelids.
The UPGMA analysis (Felsenstein, 1993) of distance matrices shown in the lower left half of Table 1 produced a rooted tree with the assumption of an evolutionary clock (Fig. 6). The same tree topology was obtained by a neighbor-joining analysis (Felsenstein, 1993). This rooted tree was coincident to the traditional phylogeny of annelids. Furthermore, the evolutionary rates of monomer subunits of annelid hemoglobins were also calculated by assuming that the divergence between the Oligochaeta and the Polychaeta occurred about 450 million years ago (Fushitani et al., 1988). This gave an estimated divergence time of Pheretima communissima and Pheretima hilgendorfi of 62.5 million years ago, that of Pheretima sieboldi and the other two species of Pheretima 83.3 million years ago, and that of Pontodrilus and three species of Pheretima 208.8 million years ago. Although very close species of Pheretima were analyzed, it was clearly shown that evolutionary divergence patterns can be studied using the amino acid sequence data from the monomer subunits of annelid hemoglobins.
ACKNOWLEDGMENTS
We gratefully acknowledge Dr. N. Takeuchi of the Biology Department, Miyagi College of Education, Sendai, Japan, for his cooperation in collecting the Pontodrilus matsushimensis and Dr. M. Esumi and the staff of the First Department of Pathology, Nihon University School of Medicine, for their generosity in allowing the use of DNASIS and helping set up the equipment for analyzing amino acid compositions and amino acid sequences. We also express our gratitude to the staff of the Department of Biology, Nihon University School of Medicine for their cooperation.