Báo cáo khoa học: Subunit sequences of the 4 · 6-mer hemocyanin from the golden orb-web spider, Nephila inaurata Intramolecular evolution of the chelicerate hemocyanin subunits pot

Subunit sequences of the 4 · 6-mer hemocyanin from the golden orb-web spider, Nephila inaurata Intramolecular evolution of the chelicerate hemocyanin subunits Anne Averdam, Ju¨ rgen Markl and Thorsten Burmester Institute of Zoology, Johannes Gutenberg University, Mainz, Germany The transport of oxygen in the hemolymph of many arth- ropod and mollusc species is mediated by large copper- proteins that are referred to as hemocyanins. Arthropod hemocyanins are composed of hexamers and oligomers of hexamers. Arachnid hemocyanins usually form 4 · 6-mers consisting of seven distinct subunit types (termed a–g), although in some spider taxa deviations from this standard scheme have been observed. Applying immunological and electrophoretic methods, six distinct hemocyanin subunits were identified in the red-legged golden orb-web spider Nephila inaurata madagascariensis (Araneae: Tetragnathi- dae). The complete cDNA sequences of six subunits were obtained that corresponded to a-, b-, d-, e-, f-andg-type subunits. No evidence for a c-type subunit was found in this species. The inclusion of the N. inaurata hemocyanins in a multiple alignment of the arthropod hemocyanins and the application of the Bayesian method of phylogenetic inference allow, for the first time, a solid reconstruction of the intra- molecular evolution of the chelicerate hemocyanin subunits. The branch leading to subunit a diverged first, followed by the common branch of the dimer-forming b and c subunits, while subunits d and f, as well as subunits e and g form common branches. Assuming a clock-like evolution of the chelicerate hemocyanins, a timescale for the evolution of the Chelicerata was obtained that agrees with the fossil record. Keywords: Arthropoda; Chelicerata; evolution; hemocyanin; subunit diversity. Hemocyanins are large, copper-containing respiratory pro- teins that serve to transport oxygen in many arthropod species [1,2]. Hemocyanins and their sequences have been identified in all arthropod subphyla, including the Onycho- phora, Chelicerata, Crustacea, Myriapoda and Hexapoda [3,4]. These proteins belong to a large superfamily that also includes functionally divergent proteins such as phenol- oxidases, as well as the crustacean pseudo-hemocyanins (cryptocyanins), insect hexamerins and hexamerin receptors [3,5–8]. Arthropod hemocyanins form hexamers or oligo-hexa- mers composed of distinct or related subunits in the 75 kDa-range. In each subunit, the binding of oxygen is mediated by a pair of Cu + ions that are coordinated by six conserved histidine residues [1,2]. Based on immunological differences, seven distinct hemocyanin subunit types have been identified in the chelicerates that are termed a–g [9–11]. Depending on the taxon, chelicerate hemocyanins assemble to quaternary structures of up to 8 · 6 subunits [1,11,12]. The 4 · 6-mer hemocyanin of the orthognath spider, Eurypelma californicum (tarantula) is the best studied hemocyanin in terms of structure, function and evolution [13–17]. The formation of the 4 · 6-mer hemocyanin requires the stoichiometric association of all seven subunit types (4 · a,2· b,2· c,4· d,4· e,4· f,4· g), with each subunit occupying a distinct position within the native oligomer [18]. The complete amino acid and cDNA sequences of all seven tarantula hemocyanin subunits have been determined [19]. By a combination of electron microscopic and immuno- logical methods, Markl and colleagues investigated the structure and subunit composition of 40 different spider species from 25 spider families [9,11,12]. These studies have demonstrated that all investigated mygalomorph spiders (Orthognatha; such as E. californicum) possess 4 · 6-mer hemocyanins, but in some Araneomorpha, deviations from that standard scheme have been observed. While the ÔclassicalÕ 4 · 6-mer hemocyanins are also present in many species of this taxon, many haplogyne and entelegyne spiders contain 1 · 6or2· 6-mer hemocyanins. In these hemocyanin-oligomers, some subunit types are absent. For example, the hemolymph of the entelegyne hunting spider, Cupiennius salei, contains a mixture of 1 · 6and2· 6-mer hemocyanins [20]. The C. salei hemocyanin consists of six distinct g-type subunits, while the subunits types a–f have been lost in evolution more than 200 MYA [21]. For further understanding of chelicerate hemocyanin structure and evolution, we have characterized the 4 · 6- mer hemocyanin of an araneomorph spider. We show that the hemocyanin of red-legged golden orb-web spider, Nephila inaurata madagascariensis, consists of six distinct polypeptides that can be assigned to the subunit-types a, b, d, e, f and g. These additional sequences, as well as the Correspondence to Ju ¨ rgen Markl, Institute of Zoology, University of Mainz, Mu ¨ llerweg 6, D-55099 Mainz, Germany. Fax: + 49 6131 3924652, Tel.: + 49 6131 392 2314, E-mail: markl@mail.uni-mainz.de Abbreviations: Hc, hemocyanin; MYA, million years ago. (Received 13 May 2003, revised 18 June 2003, accepted 26 June 2003) Eur. J. Biochem. 270, 3432–3439 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03730.x application of Bayesian methods for phylogenetic inference also allow the first time a reliable reconstruction of the intramolecular evolution of chelicerate hemocyanins. Materials and methods Animals Six specimens of the red-legged golden orb-web spider, Nephila inaurata madagascariensis (Chelicerata; Araneae; Tetragnathidae; Fig. 1), were kindly provided by the Zoological Garden ÔWilhelmaÕ in Stuttgart. The spiders were kept at 28 °C with a 12-h light : 12-h dark cycle and fed on insects. Specimens used in this study had a body length of 4–5 cm and leg span of 12–16 cm. Protein biochemistry After immobilization of the spiders for 1 h at 4 °C, the hemolymph was withdrawn by puncturing the heart with a syringe at the median-dorsal region of the opisthosoma. The hemolymph was collected in 20 lL50m M Tris/HCl, 5 m M CaCl 2 ,5m M MgCl 2 , 150 m M NaCl,pH7.4.Hemocytes and clotted material were removed by 10-min centrifugation at 10 000 g. In some analyses, the hemocyanin was dissociated by dialysis of the hemolymph over night at 4 °C in 130 m M glycine/NaOH, pH 9.6. SDS/PAGE ana- lyses were carried on a 7.5% gel under reducing conditions [22]. Native PAGE was performed on 7.5% polyacrylamide gels without SDS and b-mercaptoethanol. For Western blotting, the proteins were transferred to nitrocellulose at 0.8 mAÆcm )2 . Nonspecific binding sites were blocked by 5% nonfat dry milk in TBST (10 m M Tris/HCl, pH 7.4, 140 m M NaCl, 0.25% Tween-20) and the membranes were incubated overnight at 4 °C with various anti-hemocyanin Igs, diluted 1 : 5000 in 5% nonfat dry milk/TBST. The filters were washed three times for 10 min in TBST and subsequently incubated for 1 h with goat anti-(rabbit) Fab fragments conjugated with alkaline phosphatase (Dianova) diluted 1 : 10.000 in 5% nonfat dry milk/TBST. The membranes were washed as above and the detection was carried out using nitro blue tetrazolium and 5-bromo-4- chloro-3-indolyl phosphate. Antisera against crude hemo- cyanin preparations from Argiope aurantia and Araneus diadematus were raised in rabbits [9,11,12]. Crossed immu- noelectrophoresis of dissociated N. inaurata hemolymph proteins was performed as described by Weeke [23], applying anti-A. aurantia-hemocyanin antiserum. Cloning and sequencing of hemocyanin cDNAs Hematopoiesis was induced by bleeding about 1 week before RNA preparation. The spider was shock-frozen in liquid nitrogen and ground to a fine powder under continuous addition of nitrogen. Total RNA was extracted using the guanidine thiocyanate method [24]. Poly(A) + RNA was purified by the aid of the PolyATract kit (Promega). A directionally cloned cDNA expression library was established using the Lambda ZAP-cDNA synthesis kit from Stratagene. The library was amplified once and screened with a mixture of anti-A. diadematus- and anti- A. aurantia-hemocyanin Igs. Positive phage clones were converted to pBK-CMV plasmid vectors with the material provided by Stratagene according to the manufacturer’s instructions and sequenced on both strands by the com- mercial GENterprise (Mainz, Germany) sequencing service. Complete hemocyanin cDNA sequences were obtained by primer walking using specific oligonucleotides. Sequence analyses and molecular phylogenetic studies The web-based tools provided by the ExPASy Molecular Biology Server of the Swiss Institute of Bioinformatics (http://www.expasy.org) and the program GENEDOC 2.6 [25] were used for the analyses of DNA and amino acid sequences. The amino acid sequences of the N. inaurata hemocyanin subunits were added by hand to the previ- ously published alignments of chelicerate hemocyanin sequences [19,21]. Nine selected arthropod phenoloxidases [phenoloxidases from Penaeus monodon (accession number AF099741), Pacifastacus leniusculus (X83494), Marsupenaeus japonicus (AB065371), Tenebrio molitor (AB020738), Bombyx mori (D49370, D49371 and E12578), and Sarcophaga bullata (AF161260 and AF161261)) and four crustacean hemocyanins (Panulirus interruptus hemo- cyanin subunits a (P04254) and c (S21221), Homarus americanus hemocyanin A (AJ272095) and Pacifastacus leniusculus hemocyanin (AF522504)] were included in the alignment and used as outgroups for tree reconstruction. The final alignment is available from the authors upon request. Distances between pairs of sequences were cal- culated using the PAM [26] or the JTT [27] matrices implemented in the PHYLIP 3.6a2 package [28]. Tree constructions were performed by the neighbor-joining method and the reliability of the trees was tested by the bootstrap procedure with 100 replications [29]. Bayesian phylogenetic analyses were performed with MRBAYES 3.01 [30]. The PAM, JTT or WAG [31] amino acid substitution models with gamma distribution of rates was applied. Metropolis-coupled Markov chain Monte Carlo sampling was performed with four chains that were run for 100 000 generations. Prior probabilities for all trees were equal, starting trees were random, trees were sampled every 10th generation. Posterior probability densities were estimated on 5000 trees (burnin ¼ 5000). Molecular clock calcula- tions based PAM distances of orthologous subunits as described [19,21], assuming that the Xiphosura and Arachnida diverged about 450 MYA [7,32]. The confid- ence limits were calculated using the observed standard deviation of the protein distances. Fig. 1. The red-legged golden orb-web spider, Nephila inaurata. Ó FEBS 2003 Spider hemocyanin (Eur. J. Biochem. 270) 3433 Results Characterization of N. inaurata hemocyanin Hemolymph was collected from six specimens of N. inau- rata. Electron microscopic images show the presence of large particles that are indistinguishable in terms of size and shape from the 4 · 6-mer hemocyanin of E. californicum, indicating that N. inaurata contains a hemocyanin of similar structure (data not shown). Denaturating SDS/PAGE shows various main bands in the 65-kDa range, as expected for a typical chelicerate hemocyanin subunit (Fig. 2). As estimated from the Coomassie-stained gels, these polypep- tides represent more than 90% of the total hemolymph proteins, while other proteins form only a minor fraction. Thus, the hemolymph was considered a crude hemocyanin preparation and used as such in the following experiment. The putative hemocyanin bands were recognized by four polyclonal antibodies raised against different spider hemo- cyanins (Fig. 2). Minor cross-reactions with other proteins were observed with the anti- C. salei-, A. diadematus- and A. aurantia hemocyanin Igs, probably due to come con- tamination in the crude hemocyanin preparations used for immunization. The anti-E. californicum hemocyanin Igs were specific and only stained the hemocyanin bands. After dissociation of the hemocyanin subunits in alkaline buffer, the hemolymph samples were subjected to nondenaturing PAGE (Fig. 3). In the low molecular mass region, six distinct bands were identified that most likely correspond to the hemocyanin subunits. At least two slowly migrating bands were visible that probably represent nondissociated or partially dissociated 4 · 6-mer hemocyanin, or nonres- piratory proteins [33]. Two-dimensional, crossed immuno- electrophoresis of dissociated N. inaurata hemolymph proteins with anti-Argiope aurantia hemocyanin antiserum shows the presence of six immunologically distinct compo- nents that most likely correspond to distinct hemocyanin subunits (Fig. 4). N. inaurata hemocyanin subunit sequences A cDNA library that contains about 1.6 · 10 6 independent clones was constructed from a single adult specimen of N. inaurata. As no specific Igs against the hemocyanin of this species are available, we used a mixture of antisera Fig. 2. SDS/PAGE and Western blot analyses of Nephila inaurata hemolymph proteins. Hemolymph proteins (2–3 lg) of N. inaurata were applied per lane and stained with Coomassie Brilliant Blue (lane 1). Immunodetection was carried out using Igs raised against the hemocyanins of C. salei, A. diadematus, A. aurantia and E. californi- cum as indicated (lanes 2–5). Igs were diluted 1 : 5000. On the left side, the positions of the molecular mass marker proteins are given. Hc, hemocyanin subunits. Fig. 3. Native PAGE of N. inaurata hemolymph proteins. Total hemolymph proteins (10 lg) of N. inaurata (left lane) and 10 lgof purified hemocyanin of E. californicum (right lane) were applied. The hemocyanin subunits were dissociated into subunits against alkaline glycine/NaOH buffer before PAGE. The positions of the E. californ- icum hemocyanin subunits a to g, as well as the nondissociated 24mer and partially dissociated oligomers are indicated [33]. Fig. 4. Crossed immunoelectrophoresis of N. inaurata hemolymph pro- teins. Four micrograms of dissociated N. inaurata hemolymph proteins as antigen and anti-A. aurantia hemocyanin serum were used. 3434 A. Averdam et al. (Eur. J. Biochem. 270) Ó FEBS 2003 raised against the hemocyanins of the related web spiders, Araneus diadematus and A. aurantia. Two hundred and nineteen positive clones were identified, of which 124 clones containing inserts between 1.5 kb and 2.5 kb were partially sequenced at their 5¢-ends. Database comparisons show that 56 of these clones encode hemocyanin subunits. Based on the similarities with the E. californicum sequences, 17 clones were assigned to hemocyanin subunit a, 2 represent subunit b, 13 subunit d, 11 subunit e, 7 subunit f and 6 subunit g. No c-type subunit sequence was obtained. The complete sequences of each subunit were obtained by primer walking and have been deposited in the EMBL/GenBank databases (Table 1). Conceptual translation and comparisons with known chelicerate hemocyanin sequences show that all six cDNA sequences were complete and cover the whole coding regions. The cDNA sequences comprise of 2078–2350 bp, which include 22–73 bp of the 5¢-untranslated regions and open reading frames of 1878–1893 bp. The 3¢-untranslated regions comprise 137–419 bp, include the standard poly- adenylation signals (AATAAA) and are followed by poly(A)-tails of 19–58 bp. Multiple clones show the pres- ence of allelic sequences for the subunits a, d, e and g that are > 99.7% identical with the main cDNA sequence. Most of the nucleotide substitutions are silent, only in each a, d and g has a single amino acid substitution been observed (data not shown). The open reading frames of the hemocyanin subunits translate into distinct polypeptides of 625–630 amino acids, with calculated molecular masses in the range of 71–73 kDa (Table 1). As in the E. californicum and C. salei hemocya- nins [19,21], no signal peptides required for transmembrane transport have been found in the N. inaurata subunit sequences. Thus, the nascent proteins do not pass through the Golgi-apparatus and the putative N-glycosylation sites (NXT/S) in the primary structures are probably not used. There are four strictly conserved cysteine residues in domain 3oftheN. inaurata hemocyanins (Fig. 5) that form two disulfide bridges that stabilize the three-dimensional protein structure [34,35]. Sequence comparison and chelicerate hemocyanin evolution A multiple alignment of the chelicerate hemocyanin amino acid sequences was constructed and used for sequence comparisons and phylogenetic inference (Fig. 5). Compar- ison of the N. inaurata hemocyanin sequences with those of E. californicum allows the unambiguous assignment to distinct subunit types. The orthologous subunits of these species share 69.1–76.2% of their amino acids, with the a subunits being the most conserved and the b subunits the least conserved proteins (Table 2). The similarity scores of nonorthologous subunits is in the range of 55–64%. There are only few amino acid insertions/deletion among the chelicerate hemocyanin subunits that are located mainly in the loop regions between alpha-helices 1.1 and 1.3, alpha- helix 1.7 and beta-sheet 1B, beta-sheets 3B and 3C, and alpha-helices 3.3 and 3.4. Like with other web spiders belonging to the family of the Tetragnathidae, the hemolymph of the golden orb-web spider, N. inaurata, contains a 4 · 6-mer hemocyanin [11,12]. Using electrophoretic and immunological methods, we identified six distinct hemocyanin polypeptides. This result is in line with the identification of six distinct cDNA sequences, although we must consider that b-andc-type subunits usually occur as stable heterodimer and therefore may yield a single peak in crossed immuno-electrophoresis [11]. Sequence comparison and phylogenetic analyses show that the cDNAs correspond to subunits a, b, d, e, f and g. Despite the identification of 56 independent hemocyanin clones in the cDNA library, we found no evidence for the presence of a cDNA clone corresponding to a c-type subunit. We therefore designed various pairs of primers based on the known sequence of subunit c of E. californicum [19] (5¢fi3¢: bp 220–239, 591–608, 925–947, 3¢fi5¢: 947– 925, 1286–1265, 1474–1453) and used them for various PCR experiments with the cDNA library as template. However, we only obtained PCR products that correspond to the known subunit b. For reconstruction of chelicerate subunit evolution, either three selected crustacean hemocyanin sequences or 11 arthropod phenoloxidases were included in the alignment. Phylogenetic trees were calculated by the neighbor-joining method based on protein distances estimated by the PAM or JTT model of amino acid substitutions. While the grouping of orthologous subunits from different species is strongly supported, the relationships among the subunit types could not be resolved with sufficient confidence (data not shown). We therefore conducted Bayesian tree building methods. Applying three different evolutionary models of amino acid substitution, identical and solid reconstructions of the diversification scheme of chelicerate subunit types was obtained (Fig. 6). In all analyses, the clade leading to the a-type subunits diverged first. This clade includes subunit II of the horseshoe crab, Limulus polyphemus (LpoHc2). The next branch is formed by the b-andc-type subunits of N. inaurata and E. californicum, while there are common Table 1. Molecular properties of the N. inaurata hemocyanin subunits. Accession number, EMBL/GenBank DNA data. Subunit Accession number cDNA [bp] a Protein [Amino acids] c Molecular mass [kDa] b pI c NinHc-a AJ547807 2067 630 72.09 5.39 NinHc-b AJ547808 2060 628 72.59 6.06 NinHc-d AJ547809 2327 627 71.94 6.36 NinHc-e AJ547810 2106 625 71.00 5.49 NinHc-f AJ547811 2152 626 71.73 6.09 NinHc-g AJ547812 2059 626 71.19 5.88 a Without poly(A) tail. b Including the initiator methionine. c Without initiator methionine. Ó FEBS 2003 Spider hemocyanin (Eur. J. Biochem. 270) 3435 Fig. 5. Alignment of the amino acid sequences of N. inaurata and E. californicum hemocyanin subunits. Strictly conserved amino acids are shaded in grey, the secondary structure elements as deduced from L. polyphemus subunit II [34] are indicated at the bottom. Other features are given in the upper row: (*) copper binding histidine; (c) disulfide bridges. The abbreviations used are: NinHc-a to g, hemocyanin subunits a–g of N. inaurata (see Table 1 for accession numbers); EcaHc-a to g, E. californicum hemocyanin subunits a–g (acc. nos: X16893, AJ290429, AJ277489, AJ290430, X16894, AJ277491 and AJ277492). 3436 A. Averdam et al. (Eur. J. Biochem. 270) Ó FEBS 2003 branches of subunit types d and f,ande and g.TheC. salei hemocyanins group with the g-subunit of N. inaurata, while the hemocyanin subunit 6 of the scorpion Androctonus australis (AauHc6) and subunit A of the horseshoe crab, Tachypleus tridentatus (TtrHcA), (that form a common but not significantly supported branch) are basal to all spider g-subunits. A timescale of chelicerate hemocyanin evolution was inferred assuming that the LpoHc2 and the a-subunits of N. inaurata and E. californicum on the one hand, and TtrHcA and the arachnid g-subunits on the other hand are orthologous proteins (see above). The fossil record suggests that the Arachnida (A. australis, E. californicum, N. inau- rata and C. salei) and the Xiphosura (L. polyphemus and T. tridentatus) separated about 450 MYA [32]. Using the PAM evolution model, a mean replacement rate of 0.66 ± 0.03 · 10 )9 amino acid substitutions per site per year was calculated for the a-type subunits, and 0.60 ± 0.03 · 10 )9 for the g-type subunits. Based on these predictions, we calculated that the hemocyanins of the N. inaurata and E. californicum diverged about 279 ± 28 MYA (Fig. 7). The C. salei hemocyanins and N. inaurata subunit g separated 222 ± 9 MYA. Assuming that scor- pion hemocyanin AauHc6 is associated with the hemocy- anins of the Araneae rather than with TtrHcA, we estimated that AauHc6 split from the spider hemocyanins about 381 ± 32 MYA. The earliest divergence of chelicerate hemocyanin subunits (i.e., a-type subunits vs. all others) took place about 550 ± 45 MYA, the clade leading to the b/c subunits emerged 545 ± 24 MYA, while the other subunit types emerged about 450–470 MYA. Discussion Structure and subunit composition of N. inaurata hemocyanin The absence of subunit c in N. inaurata is surprising, as seven distinct subunits are required to build a functional active 4 · 6-mer hemocyanin in E. californicum [16,18]. While we cannot rule out the possibility that we missed subunit c in all of our analyses, it is, however, also possible that the 4 · 6-mer hemocyanin of N. inaurata is in fact built by only six subunit types. We must also consider the findings of Kuwada and Sugita [36], which suggests that subunit loss and duplication may also occur in some mygalomorph spiders that are generally assumed to contain 4 · 6-mer hemocyanins [11,12]. In E. californicum, subunit c builds a stable dimer with subunit b that is centred in the core of the 4 · 6-mer. Table 2. Sequence comparison of N. inaurata and E. californicum subunits. N. inaurata E. californicum DNA identity [%] Protein identity [%] NinHc-a EcaHc-a 72.0 76.2 NinHc-b EcaHc-b 66.4 69.1 NinHc-d EcaHc-d 70.1 74.9 NinHc-e EcaHc-e 69.6 71.3 NinHc-f EcaHc-f 70.1 76.2 NinHc-g EcaHc-g 69.6 72.8 Fig. 6. Phylogenetic tree of the chelicerate hemocyanin subunits. The numbers at the nodes represent Bayesian posterior probabilities esti- mated with the PAM model of amino acid substitution. Abbreviations: AauHc6, A. australis hemocyanin subunit 6 (acc. no. P80476); LpoHcII, subunit II of L. polyphemus (P04253); TtrHcA, T. trident- atus hemocyanin a [42]; CsaHc1–6, C. salei hemocyanin subunits 1 through 6 (AJ307903 – AJ307907, AJ307909). For other abbrevia- tions, see Fig. 5. Fig. 7. Timescale of the evolution in the chelicerate hemocyanin sub- units. The grey bars are the standard errors. Abbreviation: MYA, million years ago; for abbreviations, see Figs 5 and 6. Ó FEBS 2003 Spider hemocyanin (Eur. J. Biochem. 270) 3437 Assuming that the formation of hemocyanin quaternary structure is similar in E. californicum and N. inaurata, another subunit must have taken over the role and the position of the missing c-subunit in the N. inaurata hemo- cyanin that might be subunit b. Although we have no experimental evidence for such an assumption, it is therefore possible that the N. inaurata hemocyanin contains four copies of subunit b. Emergence of hemocyanin subunit diversity in the Chelicerata Our previous molecular phylogenetic analyses of the evolution of the arthropod hemocyanin superfamily agree that the chelicerate hemocyanin subunits are monophyletic [7,19,21], but failed to provide a solid reconstruction for the intramolecular evolution of the chelicerate hemocyanins. On one hand, this was due to the limited number of available sequences and on the other hand to the restrictions of the phylogenetic methods. With the inclusion of the N. inaurata hemocyanin sequences, the resolution increased slightly, although the grouping of the different subunit types is still inconsistent and the bootstrap values are low. With the application of a Bayesian method for phylogenetic inference [37], the resolution of the tree significantly increased and resulted in high posterior probabilities (Fig. 6). Notably, we obtained identical trees using three evolutionary models of amino acid substitutions (PAM, JTT or WAG). The phylogeny presented here shows for the first time a solid reconstruction of the diversification scheme of the chelicerate hemocyanin subunits and the pattern of intra- molecular evolution of this protein. The first gene duplica- tion probably took place already in the chelicerate stemline, close to 550 MYA, and gave rise to the a-subunits and the ancestor of all other subunit types. Further differentiation of subunit types occurred about 450–470 MYA, with the exception of the b/c-subunits that diverged around 420 MYA. These calculations agree with the observation of immunologically related subunits in L. polyphemus and E. californicum [10–12]. The Bayesian reconstructions show a common clade of the d and f subunit types on one hand, and the e and g subunit types on the other. A close relationship of subunits d and f is also supported by a conserved common deletion of two amino acids between b-sheets 3B and 3C (Fig. 5). Interestingly, the topology presented here was essentially proposed already, on the basis of antigenic determinants, by Markl and coworkers [10–12]. Nevertheless, it should be noted that in previous studies based on structural and immunological similarities, AauHc6 was homologised with EcaHc-e [38,39], while our phylogenetic inference strongly suggest an association of AauHc6 with the g-type subunits (Fig. 5). Latter assump- tion is, however, also supported by the higher similarity scores of AauHc6 with the g-type than with the e-type subunits. A hemocyanin-based timescale for the evolution of the Chelicerata The orthologous hemocyanin subunits of E. californicum and N. inaurata differ in 24–31% of their amino acids (Table 2). Under the assumption of a clock-like evolution and an arachnid–xiphosuran divergence of 450 MYA, these differences translate into a divergence time of the species of about 279 ± 28 MYA (Fig. 7). This estimate represents the time of separation of the mygalomorph and araneomorph spiders, and is in good agreement with previous calculations (285 MYA [21]); and the fossil data [40]. The ancestors of Nephila and Cupiennius separated about 222 ± 9 MYA that reflect the time of divergence of the Orbiculariae and the ÔRTA-cladeÕ within the entelegyne spiders [41]. The correc- ted replacement rate for the g-type subunits led to slightly earlier times ( 5%) of the differentiation of the C. salei hemocyanins than estimated before [21] that are, however, within the range of the expected standard deviations and are still in line with fossil data [40]. According to our calcula- tions, the differentiation of the Scorpiones from the other Arachnids took place 381 ± 32 MYA that agrees again with the fossils [32,40]. These accurate time estimates suggest that the hemocyanin subunits are excellent tools to investi- gate the evolution of the Chelicerata. Acknowledgements We thank J. Schultess for his help with the cDNA library, K. Kusche for her advice and W. Gebauer for the EM pictures. This work was supported by grants of the Deutsche Forschungsgemeinschaft (Bu956/ 5; Ma843/4) and by the Feldbauschstiftung Mainz. References 1. Markl, J. & Decker, H. (1992) Molecular structure of the arthropod hemocyanins. Adv. Comp. Environ. Physiol. 13, 325–376. 2. van Holde, K.E. & Miller, K.I. (1995) Hemocyanins. Adv. Protein Chem. 47, 1–81. 3. Burmester, T. (2002) Origin and evolution of arthropod hemo- cyanins and related proteins. J. Comp. Physiol. B. 172, 95–117. 4. Kusche, K., Ruhberg, H. & Burmester, T. (2002) A hemocyanin from the Onychophora and the emergence of respiratory proteins. Proc.NatlAcad.Sci.USA99, 10545–10548. 5. Beintema, J.J., Stam, W.T., Hazes, B. & Smidt, M.P. (1994) Evolution of arthropod hemocyanins and insect storage proteins (hexamerins). Mol. Biol. Evol. 11, 493–503. 6. Burmester, T. & Scheller, K. (1996) Common origin of arthropod tyrosinase, arthropod hemocyanin, insect hexamerin and dipteran arylphorin receptor. J. Mol. Evol. 42, 713–728. 7. Burmester, T. (2001) Molecular evolution of the arthropod hemocyanin superfamily. Mol. Biol. Evol. 18, 184–195. 8. Kusche, K. & Burmester, T. (2002) The hemocyanin superfamily of the Arthropoda. In Proceedings of the XIIth International Conference on Invertebrate Dioxygen Binding Proteins – IO2BiP 2000 (Lallier, F.H., Zal, F. & Toulmond, A., eds). Editions de la Station Biologique de Roscoff, France. 9. Markl, J., Sto ¨ cker, W., Runzler, R., Kempter, B., Bijlholt, M.M.C. & van Bruggen, E.F.J. (1983) Subunit heterogeneity, quaternary structure and immunological relationship of arthro- pod hemocyanins. Life Chem. Report Supplement 1, 39–42. 10.Kempter,B.,Markl,J.,Brenowitz,M.,Bonaventura,C.& Bonaventura, J. (1985) Immunological correspondence between arthropod hemocyanin subunits. II. Xiphosuran (Limulus)and spider (Eurypelma, Cupiennius) hemocyanin. Biol. Chem. Hoppe Seyler 366, 77–86. 11. Markl, J. (1986) Evolution and function of structurally diverse subunits in the respiratory protein hemocyanin from arthropods. Biol. Bull. (Woods Hole) 171, 90–115. 3438 A. Averdam et al. (Eur. J. Biochem. 270) Ó FEBS 2003 12. Markl, J., Sto ¨ cker, W., Runzler, R. & Precht, E. (1986) Immunological correspondences between the hemocyanin sub- units of 86 arthropods: evolution of a multigene protein family. In Invertebrate Oxygen Carriers (Linzen, B., ed.), pp. 281–292. Springer, Heidelberg. 13. Decker, H., Schmid, R., Markl, J. & Linzen, B. (1980) Hemo- cyanins in spiders, XII. Dissociation and reassociation of Eur- ypelma hemocyanin. Hoppe Seylers Z. Physiol. Chem. 361, 1707–1717. 14. Markl, J., Bonaventura, C. & Bonaventura, J. (1981a) Hemo- cyanins in spiders, XIII. Kinetics of oxygen dissociation from individual subunits of Eurypelma and Cupiennius hemocyanin. Hoppe Seylers Z. Physiol. Chem. 362, 429–437. 15. Markl, J., Savel, A. & Linzen, B. (1981b) Hemocyanins in spiders, XIV. Subunit composition of dissociation intermediates and its bearingonquaternarystructureofEurypelma hemocyanin. Hoppe-Seyler’s Z. Physiol. Chem. 362, 1255–1262. 16. Markl, J., Kempter, B., Linzen, B., Bijlholt, M.M.C. & van Bruggen, E.F. (1981c) Hemocyanins in spiders, XVI. Subunit topography and a model of the quaternary structure of Eurypelma hemocyanin. Hoppe-Seyler’s Z. Physiol. Chem. 362, 1631–1641. 17. Hartmann, H. & Decker, H. (2002) All hierarchical levels are involved in conformational transitions of the 4 x 6-meric tarantula hemocyanin upon oxygenation. Biochim. Biophys. Acta 1601, 132–137. 18. Markl, J., Decker, H., Linzen, B., Schutter, W.G. & van Bruggen, E.F. (1982) Hemocyanins in spiders, XV. The role of the indi- vidual subunits in the assembly of Eurypelma hemocyanin. Hoppe Seylers Z. Physiol. Chem. 363, 73–87. 19. Voit, R., Feldmaier-Fuchs, G., Schweikardt, T., Decker, H. & Burmester, T. (2000) Complete sequence of the 24-mer hemo- cyanin of the tarantula Eurypelma californicum. J. Biol. Chem. 275, 39339–39344. 20. Markl, J. (1980) Hemocyanins in spiders, XI. The quaternary structure of Cupiennius hemocyanin. J. Comp. Physiol. B 140, 199–207. 21. Ballweber, P., Markl, J. & Burmester, T. (2002) Complete hemocyanin-subunit sequences of the hunting spider Cupiennius salei: Recent hemocyanin-remodelling in entelegyne spiders. J. Biol. Chem. 277, 14451–14457. 22. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 277, 680–685. 23. Weeke, B. (1973) Crossed immunoelectrophoresis. Scand. J. Immunol. 2, 47–56. 24. Chirgwin, J.M., Przbyla, A.E., MacDonald, R.J. & Rutter, W.J. (1979) Isolation of biologically active ribonucleic acid from sour- ces enriched in ribonuclease. Biochemistry 18, 5294–5299. 25. Nicholas, K.B. & Nicholas, H.B. Jr (1997) GeneDoc: Analysis and Visualization of Genetic Variation, http://www.psc.edu/biomed/ genedoc/ 26. Dayhoff, M.O. (1979) Atlas of Protein Sequence and Structure,5, National Biomedical Research Foundation, Washington, DC. 27. Jones, D.T., Taylor, W.R. & Thornton, J.M. (1992) The Rapid Generation of Mutation Data Matrices from Protein Sequences. CABIOS, 8, 275–282. 28. Felsenstein, J. (2001) Phylip (Phylogeny Inference Package), Version 3.6b2. Department of Genetics, University of Washing- ton, Seattle. 29. Felsenstein, J. (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–791. 30. Huelsenbeck, J.P. & Ronquist, F. (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. 31. Whelan, S. & Goldman, N. (2001) A general empirical model of protein evolution derived from multiple protein families using a maximum-likelihood approach. Mol Biol. Evol. 18, 691–699. 32. Dunlop, J.A. & Selden, P.A. (1997) The early history and phylo- geny of the chelicerates. In Arthropod Relationships (Fortey, R.A. & Thomas, R.H., eds), pp. 221–236. Systematic Association Special Volume Series 55, Chapman & Hall, London. 33. Markl, J., Markl, A., Schartau, W. & Linzen, B. (1979) Subunit heterogeneity in arthropod hemocyanins. I. Chelicerata. J. Comp. Physiol. 130, 283–292. 34. Hazes, B., Magnus, K.A., Bonaventura, C., Bonaventura, J., Dauter, Z., Kalk, K.H. & Hol, W.G. (1993) Crystal structure of deoxygenated Limulus polyphemus subunit II hemocyanin at 2.18 A ˚ resolution: clues for a mechanism for allosteric regulation. Protein Sci. 2, 597–619. 35. Topham, R., Tesh, S., Westcott, A., Cole, G., Mercatante, D., Kaufman, G. & Bonaventura, C. (1999) Disulfide bond reduction: a powerful, chemical probe for the study or structure-function relationships in hemocyanins. Arch. Biochem. Biophys. 369, 261–266. 36. Kuwada, T. & Sugita, H. (2000) Evolution of hemocyanin sub- units in mygalomorph spiders: distribution of hemocyanin sub- units and higher classification of the Mygalomorphae. Zool. Sci. 17, 517–525. 37. Huelsenbeck, J.P., Ronquist, F., Nielsen, R. & Bollback, J.P. (2001) Bayesian inference of phylogeny and its impact on evolu- tionary biology. Science 294, 2310–2314. 38. Lamy, J., Bijlholt, M.C., Sizaret, P.Y., Lamy, J. & van Bruggen, E.F. (1981) Quaternary structure of scorpion (Androctonus australis) hemocyanin. Localization of subunits with immunological methods and electron microscopy. Biochemistry 20, 1849–1856. 39. Markl, J., Gebauer, W., Runzler, R. & Avissar, I. (1984) Immunological correspondence between arthropod hemocyanin subunits, I. Scorpion (Leiurus, Androctonus)andspider (Eurypelma, Cupiennius) hemocyanin. Hoppe-Seyler’s Z. Physiol. Chem. 365, 619–631. 40. Selden, P.A. (1993) Arthropoda (Aglaspidida, Pycnogonida and Chelicerata). In The Fossil Record 2 (Benton, M.J., ed.), pp. 297– 320. Chapman & Hall, London. 41. Coddington, J.A. & Levi, H.W. (1991) Systematics and evolution of spiders (Araneae). Annu. Rev. Ecol. Syst. 22, 565–592. 42. Linzen, B., Soeter, N.M., Riggs, A.F., Schneider, H.J., Schartau, W.,Moore,M.D.,Behrens,P.Q.,Nakashima,H.,Takagi,T., Nemoto, T., Vereijken, J.M., Bak, H.J., Beintema, J.J., Volbeda, A., Gaykema, W.P.J. & Hol, W.G.J. (1985) The structure of arthropod hemocyanins. Science 229, 519–524. Ó FEBS 2003 Spider hemocyanin (Eur. J. Biochem. 270) 3439 . Subunit sequences of the 4 · 6-mer hemocyanin from the golden orb-web spider, Nephila inaurata Intramolecular evolution of the chelicerate hemocyanin subunits Anne. seven subunit types (4 · a, 2· b, 2· c ,4 d ,4 e ,4 f ,4 g), with each subunit occupying a distinct position within the native oligomer [18]. The complete amino