Molecular and functional characterization of hemocyanin of the giant African millipede, Archispirostreptus gigas

The size and complexity of metazoan animals compelled the evolution of specific anatomical and molecular traits (notably circulatory systems and O₂-transporting proteins) to secure the transfer of O₂ from the respiratory surfaces to the respiring tissues to support aerobic metabolism. In contrast to the vast majority of animals, in which O₂ is transported by respiratory proteins such as hemoglobin (Hb), hemocyanin (Hc) and hemerythrin circulating in body fluids, terrestrial arthropods possess a tracheal system that permits gaseous diffusion of O₂ to individual tissue cells (Keilin and Wang, 1946). In this regard, the Myriapoda – which includes centipedes (class Chilopoda) and millipedes (class Diplopoda) – are of particular interest in having trachea as well as Hc (Rajulu, 1969; Mangum et al., 1985; Kusche and Burmester, 2001; Kusche et al., 2002). In contrast to Hbs, which are widely distributed in the animal kingdom, the copper-containing Hcs are extracellular proteins found only in arthropods (recorded in all subphyla) (Burmester, 2001) and molluscs (Mangum et al., 1985). However, the Hcs of these two phyla are not related but emerged independently (Burmester, 2001). Arthropod Hcs are highly multimeric. Myriapod Hcs are 6×6, 36-mer structures (Jaenicke et al., 1999) composed of up to four different subunit types (Markl et al., 2009). Each subunit consists of ~650 amino acid residues that are distributed among three structural domains and include six highly conserved histidines, which coordinate two copper (I) ions that reversibly bind one O₂ molecule (Terwilliger, 1998; Markl and Decker, 1992). The multimerization increases the O₂ transport capacity of the hemolymph without markedly raising its osmotic pressure.

As with Hbs, Hcs are allosteric proteins that are in equilibrium between two conformational states (Wyman, 1969): a tense (T) state with a low O₂ equilibrium association constant (K_T), and a relaxed (R) state with a high O₂ equilibrium association constant (K_R) (Loewe, 1978). When P_O2, and thus O₂ saturation, is high (at the respiratory surfaces), most Hc molecules are in the R state, which favors O₂ loading to the hemolymph. At low P_O2 (as in the working muscle), the molecules are predominantly in the T state, which enhances unloading. Hc–O₂ binding is allosterically modulated by homotropic interactions (between O₂-binding sites) and heterotropic interactions (between the sites for binding O₂ and effectors) (Bonaventura and Bonaventura, 1980; Brouwer and Serigstad, 1989; van Holde et al., 2001). Heterotropic effectors modulate O₂ affinity by stabilizing or destabilizing the T and/or the R state or by altering the T–R allosteric equilibrium constant between the two protein conformations, L.

In contrast to the intensively studied erythrocytic vertebrate Hbs, where protons, organic phosphates and chloride ions decrease Hb–O₂ affinity by lowering K_T, striking variation is encountered amongst the multimeric invertebrate O₂-binding proteins, where effectors, commonly divalent cations, may increase or decrease K_T and/or K_R (which can be graphically represented by the intercepts of the lower and upper asymptotes, respectively, of the extended Hill plot with the y-axis at logP_O2=0). The diversity in regulatory mechanisms encountered in extracellular O₂-binding proteins calls for a parallel analysis of myriapod Hcs.

The diplopod Archispirostreptus gigas is a large, slow-moving, forest floor-dwelling deposit feeder from East Africa (Pechenik, 2005), which suggests it may have markedly different metabolic requirements from those of chilopods, which are fast-moving carnivores. As with insects, myriapods may have occludable spiracles and exhibit discontinuous gas exchange whereby their tissues may intermittently become hypoxic (Schmitz and Harrison, 2004). As soil dwellers they are moreover subjected to temporary flooding, which decreases O₂ availability. With the aim of gaining insight into the evolutionary origin of millipede Hc, its functional adaptations and its allosteric regulatory mechanisms, we investigated the primary structure of A. gigas Hc and its O₂-binding properties in the native hemolymph and Hc solutions under differing physico-chemical conditions as regards pH, (l)-lactate and divalent cation concentrations.

Specimens of the giant millipede A. gigas (Peters 1855) weighing 25–82 g were obtained from a dealer (Exotera GmbH, Holzheim, Germany), and kept in a terrarium at room temperature and fed regularly.

Hemolymph for O₂-binding studies was drawn from 12 animals using a hypodermic syringe fitted with 23 gauge × 1¼ in needles to pierce the dorsal intersegmental membranes. All subsequent preparative steps were carried out at 0–4°C. The samples were pooled, and centrifuged for 10 min at 10,000 g to remove cellular debris. Samples used to record extended Hill plots (see below) were concentrated by ultrafiltration in Amicon Ultra Centrifugal filters (cut-off: 100,000 kDa; Millipore, Billerica, MA, USA). To study the effect of cations (Na⁺, Ca²⁺, Mg²⁺), the hemolymph was dialyzed against three changes of 10 mmol l⁻¹ Tris-HCl buffer, pH 7.5 at a 1:300 sample:buffer ratio and concentrated. The dialyzed hemolymph is denoted as ‘stripped Hc’. The hemolymph and Hc preparations were divided into 100 μl aliquots that were stored at −80°C and thawed individually for measurement of O₂ equilibria.

Hc concentration was derived from the absorbance at 335 nm using the ε=0.0175 l mol⁻¹ cm⁻¹ reported for crab Carcinus manaes Hc (Nickerson and Van Holde, 1971; Weber et al., 2008) and a subunit mass of 73.5 kDa as deduced from cDNA sequence analysis (described below).

A polyclonal antibody was generated against purified A. gigas Hc in guinea pigs. SDS-PAGE was performed on a 7.5% gel using standard methods as previously described (Jaenicke et al., 1999). Two-dimensional gel electrophoresis (2D-PAGE) using pH 3.5–10 ampholines was performed according to O'Farrell (O'Farrell, 1975). For western blotting, the proteins were transferred to nitrocellulose at 0.8 mA cm⁻². Non-specific binding sites were blocked by 5% non-fat dry milk in TBST (10 mmol l⁻¹ Tris-HCl, pH 7.4, 140 mmol l⁻¹ NaCl, 0.25% Tween-20). Incubation with the anti-Hc antibody in 5% non-fat dry milk/TBST was carried out for 2 h at room temperature. The filters were washed 3×10 min in TBST and subsequently incubated for 1 h with a goat anti-rabbit secondary antibody conjugated with alkaline phosphatase, diluted in 5% non-fat dry milk/TBST. The membranes were washed as above and detection was carried out using nitro-blue-tetrazolium and bromo-chloro-indolyl-phosphate.

Mass spectrometry was carried out by nanoLC-ESI-ion trap analysis after tryptic digest of the Hc bands from SDS-PAGE. Protein identification was performed with Mascot software (Perkins et al., 1999) using the NCBI nr database.

Total RNA was extracted from the whole animal after removal of the cuticle. Poly(A)⁺ RNA was purified from total RNA using a PolyATract kit (Promega, Madison, WI, USA); 5 μg poly(A)⁺ RNA was then used for the construction of a directionally cloned cDNA expression library with the Lambda ZAP-cDNA synthesis kit (Stratagene, Agilent Technologies, Santa Clara, CA, USA). The library was screened with the anti-A. gigas Hc antibodies. Positive phage clones were converted to plasmid vectors using the material provided in the kit. The cDNAs inserted in the pBK-CMV vector were sequenced on both strands by a commercial sequencing service (Genterprise, Mainz, Germany). A full-length Hc cDNA sequence was obtained and is available from the EMBL/GenBank databases under the accession number HE574799.

The web-based tools provided by the ExPASy Molecular Biology Server of the Swiss Institute of Bioinformatics (http://www.expasy.org) were used for sequence analyses. A multiple sequence alignment of the amino acid sequences of selected arthropod Hcs and phenoloxidases (supplementary material Table S1) was constructed with MAFFT 6 (Katoh et al., 2005) (http://mafft.cbrc.jp/alignment/server/). The L-INS-i routine and the BLOSUM 45 matrix were selected. The final alignment covered 93 sequences and 906 positions (supplementary material Fig. S1).

Bayesian phylogenetic analysis was performed using MrBayes 3.1.2 (Huelsenbeck and Ronquist, 2001). We assumed the WAG model with a gamma distribution of substitution rates. Metropolis-coupled Markov chain Monte Carlo (MCMCMC) sampling was performed with one cold and three heated chains. Two independent runs were performed in parallel for five-million generations. The average standard deviation of split frequencies was <0.005. Starting trees were random and the trees were sampled every 1000th generation. Posterior probabilities of the nodes were estimated on the final 4000 trees (burnin=1000).

O₂ equilibria were measured at 25°C using a modified gas-diffusion chamber (Weber, 1981) coupled to two serially linked precision Wösthoff gas mixing pumps (Bochum, Germany) for mixing pure (>99.998%) N₂ and atmospheric air. In the procedure, absorbance of 3 μl hemolymph samples was recorded at 365 nm following equilibration with pure O₂ and pure N₂ to obtain the full saturation range, and thereafter monitored during stepwise increases in the P_O2 of the equilibrating gas mixture. O₂ equilibrium curves were derived as the relationship between absorbance levels (corresponding to O₂ saturation) and calculated O₂ tensions at the individual steps. In order to assess the Bohr effect, pH values in the measured samples were varied by adding 1 mol l⁻¹ Tris-HCl (for measurements at >pH 7) or bisTris buffers (<pH 7) to obtain a final buffer concentration of 0.1 mol l⁻¹ (unless otherwise indicated). The pH values were measured at 25°C using a BMS 2 MK 2 microelectrode coupled to a PHM 64 Research pH meter (Radiometer, Copenhagen, Denmark).

For each O₂ equilibrium curve, P₅₀ and n₅₀ values (O₂ tension and Hill's cooperativity coefficients, respectively, at 50% oxygenation) were derived from the zero intercepts and slopes, respectively, of Hill plots (logS/(1–S) versus logP_O2, where S is the fractional saturation) based on at least four equilibrium steps in the 25–75% saturation range. Extended Hill plots for assessment of K_T, K_R and other allosteric parameters were obtained by recording at least 10 additional steps at low (<10%) and high (>90%) O₂ saturations.

SDS-PAGE and 2D-PAGE followed by western blotting employing specific anti-Hc antibodies identified two distinct Hc subunits in the hemolymph of A. gigas (Fig. 1). Mass spectrometry (nanoLC-ESI-ion trap) showed that the two bands corresponded to the same Hc subunit and identified several glycosylated peptides. A single Hc cDNA sequence was obtained by screening an A. gigas cDNA expression library. No additional subunit could be detected despite extensive screening and additional RT-PCR experiments with the ESTs of this species (Meusemann et al., 2010). Thus, in contrast to most arthropod Hcs (Markl and Decker, 1992; Kusche et al., 2003), the native A. gigas Hc likely consists of only a single subunit type.

The full-length Hc cDNA sequence encompasses 2115 bp (plus a polyA tail of 18 bp), including an open reading frame of 1962 bp beginning with a methionine ATG at bp 40 (supplementary material Fig. S1). The typical polyadenylation signal AATAAA is located 34 bp upstream of the start of the polyA tail. A polypeptide of 653 amino acids was deduced. The predicted molecular mass was 75.2 kDa. The N-terminal sequence harbors a typical signal peptide of 17 amino acids (supplementary material Fig. S1). Thus, the native secreted Hc subunit comprises 636 amino acids with a calculated molecular mass of 73.5 kDa. Three putative N-glycosylation sites were detected at amino acid positions 404, 435 and 630. Their usage is supported by the glycosylated peptides detected in mass spectrometry. Differential glycosylation most likely explains the two Hc bands detected in the western blot.

Archispirostreptus gigas Hc is most similar to the Hc of the diplopod Spirostreptus sp. (Kusche and Burmester, 2001), with which it shares 94.0% of the amino acids and 95.3% of the nucleotides in the coding region. The six copper-coordinating histidines that are strictly conserved in all arthropod Hcs (Burmester, 2002; Linzen et al., 1985) are present in the copper-binding sites A and B. The amino acid sequence of A. gigas Hc was included in a multiple sequence alignment of a total of 86 other Hcs from all arthropod subphyla. Six phenoloxidase sequences were added as an outgroup. Bayesian phylogenetic analyses support the monophyly of the myriapod (diplopod and chilopod) Hcs (posterior probability: 1.0; Fig. 2). The A. gigas Hc groups with the Spirostreptus sp. Hc. The two diplopod Hcs are a sister group of the HcA subunit of Scutigera coleoptrata (Kusche et al., 2003), while the other four subunits of S. coleoptrata are basal. This demonstrates an early divergence of distinct subunit types in the myriapods before the separation of Diplopoda and Chilopoda, which occurred at least ~420 million years ago in the Silurian period (Wilson and Anderson, 2004). Thus, the formation of the oligohexameric Hc (probably a 36-mer) is an ancient event. The phylogeny of the Hc subunits further supports a monophyly of the Mandibulata (Myriapoda + Crustacea; 0.96 support), but rejects Myriochelata (Myriapoda + Chelicerata).

Archispirostreptus gigas whole hemolymph contains high levels of Hc (159 mg ml⁻¹) compared with centipedes (93 mg ml⁻¹ in Scutigera coleoptrata) (Mangum et al., 1985). The whole hemolymph (Fig. 3) exhibits a high O₂ affinity [P₅₀=3.45 mmHg (0.46 kPa) at pH 8.1 and 25°C] compared with other crustacean Hcs studied under similar temperature and pH conditions (P₅₀=2.8–24 mmHg) (Morris and Bridges, 1994). The cooperativity coefficients are low (~1.5 at <pH 7.0 and >pH 8.0) but increase to ~3.0 near pH 7.4 (Fig. 3), a pattern commonly observed in other Hcs (Miller and Van Holde, 1974; Miller and Mangum, 1988). These O₂-binding properties differ sharply from those of centipedal S. coleoptrata Hc, despite similar quaternary structures and subunit sequence homology in millipede and centipede Hcs (supplementary material Fig. S1). The high O₂ affinity and low cooperativity in A. gigas Hc predictably favor O₂ binding at low O₂ tensions, as expected to prevail in myriapods with gas-impermeable exoskeletons that lack the capacity for tracheal ventilation (Rajulu, 1970). This contrasts with the low affinity and high cooperativity observed in centipede S. coleoptrata Hc, which may be adaptive in enhancing O₂ release in the highly active centipedes (Markl et al., 2009; Jaenicke et al., 1999).

The O₂ affinity of whole A. gigas hemolymph was not materially affected by pH or by the addition of lactate or Ca²⁺ at <pH 6.7 (Fig. 3B). However, at pH 7.0–8.0, a marked Bohr factor (ϕ=−0.73) was observed that was further increased by effectors [ϕ=−0.77 and −0.85, respectively, in the presence of 10 mmol l⁻¹ (l)-lactate and 20 mmol l⁻¹ Ca²⁺]. The marked Bohr effect implies that activity-induced acidification of tissues bathed by the hemolymph will enhance unloading of O₂ from the Hc. A similarly large Bohr effect (ϕ=−0.87) has been observed in the house centipede, S. coleoptrata (Mangum et al., 1985).

In the intensively studied tetrametric vertebrate Hbs, the majority of the Bohr effect is attributable to oxygenation-linked binding of protons to surface His residues that increase their pK values (Lukin and Ho, 2004; Berenbrink, 2006). The subunits of A. gigas Hc contain 34 His residues, including the six copper-binding residues and four histidines in the interface between hexamers (H152, H155, H443 and H446) (supplementary material Fig. S1), and some of the free surface residues may undergo oxygenation-linked proton binding under physiological conditions and transmit forces during allosteric interaction (Markl et al., 2009), contributing to the Bohr effect. In tarantula Eurypelma californicum Hc, salt bridges between conserved surface histidine and glutamate residues located at particular intersubunit interfaces are considered responsible for the observed Bohr effect (Sterner and Decker, 1994). However, the mechanism of O₂ binding and its modulation by allosteric effectors depends on the interplay of many factors, including the coordination geometry and redox potential of copper. Studies on crustacean (Panulirus interruptus and Carcinus aestuarii) Hc suggest that the Bohr effect of arthropod Hcs results from pH-dependent structural modulation of the geometry of copper in the deoxygenated Hc molecules, ‘leaving the question of identifying the role of individual amino acids to future studies’ (Hirota et al., 2008).

Strikingly, the addition of 10 and 50 mmol l⁻¹ (l)-lactate to whole hemolymph slightly decreased O₂ affinity in the range pH 7–8 (Fig. 3B), which contrasts sharply with the commonly encountered lactate-induced increase in O₂ affinity observed in crustacean Hcs, which would favor O₂ binding under hypoxic conditions (Truchot, 1980; Morris and Bridges, 1994; Bridges et al., 1984; Paoli et al., 2007; Hellmann et al., 2010).

The slight lactate effect observed in A. gigas supports the correlation between terrestrial life and reduced lactate effects reported in crustaceans – and extends it to diplopods. Accordingly, lactate has no effect on Hc–O₂ affinity in the tropical land hermit crab Coenobita clypeoatus (Morris and Bridges, 1986) and the terrestrial, obligate air-breathing decapod Birgus latro (Morris et al., 1988). Compared with marine crustaceans, Hc of the Christmas Island red landcrab, Gecarcoides natalis, even shows a reverse lactate effect [O₂ affinity falls with increasing (l)-lactate concentration] (Adamczewska and Morris, 1998) that predictably favors O₂ unloading in the tissues in terrestrial habitats where O₂ loading is secured.

Addition of Ca²⁺ ions to whole hemolymph drastically increased Hc–O₂ affinity in the pH range 7.0–8.0. The Ca²⁺ effect increased with pH (Fig. 3B) suggesting that the O₂-linked Ca²⁺-binding sites become increasingly accessible and/or deprotonated with rising pH, as also observed in crustacean (Penaeus monodon) Hc (Beltramini et al., 2005). In contrast, addition of Mg²⁺ to whole hemolymph does not affect Hc–O₂ affinity, indicating that the O₂-linked cation-binding sites are specific to Ca²⁺ ions or that potential O₂-linked Mg²⁺-binding sites are saturated in the whole hemolymph.

The pH of native hemolymph was 8.2 at 25°C, which falls within the range measured in freshly collected diplopod hemolymph (pH 8.0–8.5) (Xylander, 2009). This implies (see Fig. 3) that variations in Ca²⁺ concentration exert a tangible effect on O₂ affinity at physiological pH. The dose–response curve at ~pH 7.96 for the relationship between [Ca²⁺] and P₅₀ in stripped Hc (Fig. 4A), moreover, shows high sensitivity of P₅₀ to Ca²⁺ at low cation levels that characterize arthropod hemolymph (3.5–4.3 mmol l⁻¹ in arachnids) (Paul et al., 1994; Decker et al., 1980). Fitting the logP₅₀/[Ca²⁺] curve to a hyperbola (Fig. 4A) indicates that the Ca²⁺ concentration required for half the maximum increase in logP₅₀ – which reflects the apparent Ca²⁺ binding constant – is 12.3 mmol l⁻¹. The slope of the double logarithmic plot (Fig. 4B) of 0.55±0.05 (mean ± s.e.m.) reflects binding of 0.5–0.6 Ca²⁺ ions per O₂ molecule. In contrast to Ca²⁺, 100 mmol l⁻¹ Na⁺ – the major cation in extracellular fluids – had no detectable effect on O₂ affinity of the stripped Hc (Fig. 4A) – supporting the view that the cation-binding site is highly specific for Ca²⁺.

The exoskeleton of arthropods contains up to 40% mineral salts, predominantly calcium carbonate, that hardens the cuticle. Decreases in hemolymph pH associated with lactic acidoses under hypoxia may thus be expected to mobilize calcium carbonate from exoskeleton stores, raising hemolymph Ca²⁺ and stabilizing hemolymph pH. This scenario is analogous to that in turtles, whose shells are rich in calcium carbonate. In the painted turtle, Chrysemys picta, increases in lactic acid levels (up to 200 mmol l⁻¹) under anoxia are buffered largely by carbonate, resulting in the accumulation of calcium and lactate in the extracellular fluid or incorporation of calcium lactate into the shell (Jackson and Ultsch, 1982; Jackson, 2004). Our finding that Ca²⁺ increases Hc–O₂ affinity thus indicates that Ca²⁺ released under hypoxia will increase Hc–O₂ loading under these conditions.

The parameters obtained by fitting the MWC model to the data (Table 1) reveal close agreement between P_m and P₅₀, and between n_max and n₅₀ values, reflecting highly symmetrical O₂ equilibrium curves, which permit analysis of allosteric effects in terms of changes in P₅₀ values. In the absence of added effectors, the non-exclusive binding coefficient c in the hemolymph (=K_T/K_R=0.14), indicates an ~7-fold higher O₂ affinity in the relaxed than in the tense state.

The q values (5.3–10.7) suggest an interaction between at least 11 O₂-binding sites, indicating that cooperativity is not confined within the hexameric subunits but extends between the individual hexamers. Although the major structural elements are strictly conserved amongst arthropod Hcs (Markl and Decker, 1992; van Holde and Miller, 1995; Burmester, 2001), identification of the salt bridges implicated in inter-hexameric interactions is complicated by the large number of charged amino acid residues found in the five types of inter-hexamer interfaces (Markl et al., 2009). Given the large difference in cooperativity coefficients of the sluggish millipede Spirostreptus and the swift house centipede S. coleoptrata (n=1.3 and ~10, respectively), comparison of diplopod and chilopod Hcs promises insight into structure–function coupling.

Cooperativity between hexamers is accounted for by several salt bridges at inter-hexamer interfaces. Two salt bridges (Lys502 ↔ Asp575 and Asp567 ↔ Asn573) may act to connect the two 3×6mer half-molecules, and connections between adjacent hexamers may be stabilized by one salt bridge (Glu395 ↔ Lys628) (supplementary material Fig. S1) (Markl et al., 2009). The observation that q values exceeding 6 were only encountered in the hemolymph samples that had a higher Hc concentration than the stripped samples (2.43 and 1.70 mmol l⁻¹, respectively; Table 1) is congruent with dissociation of Hc subunits at low concentration (Svedberg and Heyroth, 1929). Analogously, the high q values found in the hemolymph at low pH and in the presence of lactate (q=10.7 and 8.08, respectively; Table 1) indicate that these conditions may favor the association of the hexameric subunits. The lower q value in stripped Hc, however, might also be due to the loss of other factors that are present in whole hemolymph, such as Ca²⁺, which is known to increase the stability of the Hc quaternary structure (van Holde and Miller, 1995). Indeed, as shown in Table 1, addition of 20 mmol l⁻¹ Ca²⁺ increases q of the stripped Hc from 3 to 5 and n₅₀ from 1.32 to 1.68 (Table 1). This is in agreement with structural studies on diplopod Hc that show the presence of two opposing glutamate residue pairs (E410 and E411), which are not compensated by positively charged residues and might be responsible for the bridging of Ca²⁺ between hexameric subunits (Markl et al., 2009).

The slopes of unity seen at extremely low and high O₂ saturations in extended Hill plots (Fig. 5) reflect non-cooperative binding of the first and last O₂ molecules bound by the Hc molecules. As evidenced by the more than 2-fold increase in logL upon addition of lactate (Table 1), lactate increases P₅₀ by modulating the allosteric constant L, without affecting K_T and K_R (Fig. 5A). Contrasting with this, the Bohr effect is associated with a tangible decrease in K_T as well as K_R upon acidification (by >75% when hemolymph pH falls from ~7.8 to ~6.8; Table 1, Fig. 5A). Analogously, the drastically lower Hc–O₂ affinity in stripped Hc compared with that in the whole hemolymph (P₅₀=~26 and 4 mmHg, respectively, at pH ~7.9) results from marked reductions in both K_T and K_R (Table 1, Fig. 5B). Strikingly, addition of 20 mmol l⁻¹ Ca²⁺ almost completely annuls these effects, resulting in left-shifting of the curve via increases in both K_T and K_R. The almost exact superimposition of the plot of stripped Hc in the presence of 20 mmol l⁻¹ Ca²⁺ and of the whole hemolymph indicates that Ca²⁺ is the main physiological regulator of O₂ affinity in A. gigas Hc at constant pH.

O₂ equilibrium measurements were also carried out on stripped Hc in the presence and absence of 20 mmol l⁻¹ Ca²⁺ at pH 6.4. However, it was not possible to fit the MWC model to these data because of the low level of cooperativity and almost identical K_T and K_R values, which suggests that the Hc molecules do not ‘switch’ between the T and R states and that Ca²⁺ does not exert an effect at low pH.

The mechanisms of allosteric control in A. gigas Hc differ strikingly from those characterizing the intensively studied vertebrate Hbs, where the major effectors (organic phosphates, chloride ions and protons) decrease Hb–O₂ affinity by lowering K_T, and other multimeric invertebrate O₂-binding proteins (other Hcs and extracellular Hbs). Thus in ~3000 kDa annelid (Arenicola marina, Perineries aibuhitensis and Macrobdella decora) Hbs and ~1750 kDa gastropod (Biomphalaria glabrata) Hb (Weber, 1981; Tsuneshige et al., 1989; Weber et al., 1995; Bugge and Weber, 1999) increased pH and divalent cation concentrations increase O₂ affinity predominantly by increasing K_R. The A. gigas Hc mechanism is moreover at variance with those encountered in other arthropod and in molluscan Hcs. Thus, as with multimeric Hbs, increased pH and divalent cation (Ca²⁺) concentrations raise Hc–O₂ affinity of gastropod Lymnea stagnalis Hc predominantly by increasing K_R (Dawson and Wood, 1982), whereas Mg²⁺ ions raise K_R and lower K_T of Haliotis iris Hc (Behrens et al., 2002). The A. gigas Hc mechanism also differs from those observed in crustacean Hcs, where Mg²⁺ ions and increased pH increase O₂ affinity by raising K_R in Callianassa californiensis (Miller and Van Holde, 1974) and (l)-lactate increases Hc–O₂ affinity by increasing K_T in Carcinusmaenas and Callinectes sapidus Hc (Weber et al., 2008; Johnson et al., 1988). Analyses of the allosteric interactions in terms of a nesting model that reveal hierarchies of interactions based on known hierarchy of subunit structure (Robert et al., 1987; Decker et al., 1988; Hellmann et al., 2003; Menze et al., 2005; Hellmann et al., 2008) will further elucidate the mechanisms regulating Hc–O₂ in myriapods.

Our study of A. gigas Hc shows that millipede Hcs are a sister group to centipede HcA, reflecting an early divergence of Hc subunits, which is associated with distinct differentiation in functional properties, notably in Hc–O₂ affinity and its regulation by protons (pH) and Ca²⁺ and lactate concentrations. The O₂-binding properties of the Hc are compatible with a role of the circulating hemolymph in complementing the tracheal system in supplying O₂ to the respiring tissues in myriapods. The pattern of allosteric effects observed in A. gigas Hc provides an ideal foundation for studies on the molecular basis and physiological significance of the variant patterns of allosteric control mechanisms encountered in multimeric gas-binding proteins.

We thank Anny Bang (Aarhus) and Christian Pick (Hamburg) for technical assistance and help with data handling. We thank Marcel Kwiatkowski (Hamburg) for the nanoLC-ESI-ion trap analysis.

 
• A

  absorbance

 
• c

  non-exclusive binding coefficient K_T/K_R

 
• Hb

  hemoglobin

 
• Hc

  hemocyanin

 
• K_R

  relaxed (R)-state association equilibrium constant for O₂ binding

 
• K_T

  tense (T)-state association equilibrium constant for O₂ binding

 
• L

  allosteric constant ([T]/[R]) at zero oxygenation

 
• n₅₀

  Hill's cooperativity coefficient at half-saturation

 
• n_max

  maximal cooperativity coefficient

 
• P₅₀

  oxygen tension at half saturation

 
• P_m

  median oxygen tension

 
• P_O2

  partial pressure of oxygen

 
• q

  number of interacting binding sites

 
• S

  apparent saturation

 
• ΔG

  free energy of cooperativity

 
• ϕ

  Bohr factor (ΔlogP₅₀/ΔpH)