The Wnt-specific astacin proteinase HAS-7 restricts head organizer formation in Hydra.

BACKGROUND: The Hydra head organizer acts as a signaling center that initiates and maintains the primary body axis in steady state polyps and during budding or regeneration. Wnt/beta-Catenin signaling functions as a primary cue controlling this process, but how Wnt ligand activity is locally restricted at the protein level is poorly understood. Here we report a proteomic analysis of Hydra head tissue leading to the identification of an astacin family proteinase as a Wnt processing factor. RESULTS: Hydra astacin-7 (HAS-7) is expressed from gland cells as an apical-distal gradient in the body column, peaking close beneath the tentacle zone. HAS-7 siRNA knockdown abrogates HyWnt3 proteolysis in the head tissue and induces a robust double axis phenotype, which is rescued by simultaneous HyWnt3 knockdown. Accordingly, double axes are also observed in conditions of increased Wnt activity as in transgenic actin::HyWnt3 and HyDkk1/2/4 siRNA treated animals. HyWnt3-induced double axes in Xenopus embryos could be rescued by coinjection of HAS-7 mRNA. Mathematical modelling combined with experimental promotor analysis indicate an indirect regulation of HAS-7 by beta-Catenin, expanding the classical Turing-type activator-inhibitor model. CONCLUSIONS: We show the astacin family protease HAS-7 maintains a single head organizer through proteolysis of HyWnt3. Our data suggest a negative regulatory function of Wnt processing astacin proteinases in the global patterning of the oral-aboral axis in Hydra.

Limited proteolysis by acrosin affects sperm-binding and mechanical resilience of the mouse zona pellucida.

The encounter of oocyte and sperm is the key event initiating embryonic development in mammals. Crucial functions of this existential interaction are determined by proteolytic enzymes, such as acrosin, carried in the sperm head acrosome, and ovastacin, stored in the oocyte cortical granules. Ovastacin is released upon fertilisation to cleave the zona pellucida, a glycoprotein matrix surrounding the oocyte. This limited proteolysis hardens the oocyte envelope, and thereby provides a definitive block against polyspermy and protects the developing embryo. On the other hand, acrosin, the renowned and most abundant acrosomal protease, has been thought to enable sperm to penetrate the oocyte envelope. Depending on the species, proteolytic cleavage of the zona pellucida by acrosin is either essential or conducive for fertilisation. However, the specific target cleavage sites and the resulting physiological consequences of this proteolysis remained obscure. Here, we treated native mouse zonae pellucidae with active acrosin and identified two cleavage sites in zona pellucida protein 1 (ZP1), five in ZP2 and one in ZP3 by mass spectrometry. Several of these sites are highly conserved in mammals. Remarkably, limited proteolysis by acrosin leads to zona pellucida remodelling rather than degradation. Thus, acrosin affects both sperm binding and mechanical resilience of the zona pellucida, as assessed by microscopy and nanoindentation measurements, respectively. Furthermore, we ascertained potential regulatory effects of acrosin, via activation of latent pro-ovastacin and inactivation of fetuin-B, a tight binding inhibitor of ovastacin. These results offer novel insights into the complex proteolytic network modifying the extracellular matrix of the mouse oocyte, which might apply also to other species.

Intracellular activation of ovastacin mediates pre-fertilization hardening of the zona pellucida.

STUDY QUESTION: How and where is pro-ovastacin activated and how does active ovastacin regulate zona pellucida hardening (ZPH) and successful fertilization? STUDY FINDING: Ovastacin is partially active before exocytosis and pre-hardens the zona pellucida (ZP) before fertilization. WHAT IS KNOWN ALREADY: The metalloproteinase ovastacin is stored in cortical granules, it cleaves zona pellucida protein 2 (ZP2) upon fertilization and thereby destroys the ZP sperm ligand and triggers ZPH. Female mice deficient in the extracellular circulating ovastacin-inhibitor fetuin-B are infertile due to pre-mature ZPH. STUDY DESIGN, SAMPLES/MATERIALS, METHODS: We isolated oocytes from wild-type and ovastacin-deficient (Astlnull) FVB mice before and after fertilization (in vitro and in vivo) and quantified ovastacin activity and cleavage of ZP2 by immunoblot. We assessed ZPH by measuring ZP digestion time using α-chymotrypsin and by determining ZP2 cleavage. We determined cellular distribution of ovastacin by immunofluorescence using domain-specific ovastacin antibodies. Experiments were performed at least in triplicate with a minimum of 20 oocytes. Data were pre-analyzed using Shapiro-Wilk test. In case of normal distribution, significance was determined via two-sided Student's t-test, whereas in case of non-normal distribution via Mann-Whitney U-test. MAIN RESULTS AND THE ROLE OF CHANCE: Metaphase II (MII) oocytes contained both inactive pro-ovastacin and activated ovastacin. Immunoblot and ZP digestion assays revealed a partial cleavage of ZP2 even before fertilization in wild-type mice. Partial cleavage coincided with germinal-vesicle breakdown and MII, despite the presence of fetuin-B protein, an endogenous ovastacin inhibitor, in the follicular and oviductal fluid. Upon exocytosis, part of the C-terminal domain of ovastacin remained attached to the plasmalemma, while the N-terminal active ovastacin domain was secreted. This finding may resolve previously conflicting data showing that ovastacin acts both as an oolemmal receptor termed SAS1B (sperm acrosomal SLLP1 binding protein; SLLP, sperm lysozyme like protein) and a secreted protease mediating ZP2 cleavage. LIMITATIONS, REASONS FOR CAUTION: For this study, only oocytes isolated from wild-type and ovastacin-deficient FVB mice were investigated. Some experiments involved oocyte activation by the Ca2+ ionophore A23187 to trigger ZPH. WIDER IMPLICATIONS OF THE FINDINGS: This study provides a detailed spatial and temporal view of pre-mature cleavage of ZP2 by ovastacin, which is known to adversely affect IVF rate in mice and humans. LARGE SCALE DATA: None. STUDY FUNDING AND COMPETING INTEREST(S): This work was supported by the Center of Natural Sciences and Medicine and by a start-up grant of the Johannes Gutenberg University Mainz to W.S., and by a grant from Deutsche Forschungsgemeinschaft and by the START program of the Medical Faculty of RWTH Aachen University to J.F. and W.J.D. There are no competing interests to declare.

Mammalian gamete fusion depends on the inhibition of ovastacin by fetuin-B.

The zona pellucida, a glycoprotein matrix surrounding the mammalian oocyte, hardens after intrusion of the first spermatozoon, thus protecting the embryo until implantation and preventing multiple fertilizations (polyspermy). Definitive zona hardening is mediated by the metalloprotease ovastacin, which is released from cortical granules of the oocyte upon sperm penetration. However, traces of ovastacin seep from unfertilized eggs to cause zona hardening even in the absence of sperm. These small amounts of protease are inactivated by the plasma protein fetuin-B, thus keeping eggs fertilizable. Once a sperm has penetrated the egg, ovastacin from cortical vesicles overrides fetuin-B and initiates zona hardening.

Sizzled is unique among secreted frizzled-related proteins for its ability to specifically inhibit bone morphogenetic protein-1 (BMP-1)/tolloid-like proteinases.

BMP-1/tolloid-like proteinases (BTPs) are major enzymes involved in extracellular matrix assembly and activation of bioactive molecules, both growth factors and anti-angiogenic molecules. Although the control of BTP activity by several enhancing molecules is well established, the possibility that regulation also occurs through endogenous inhibitors is still debated. Secreted frizzled-related proteins (sFRPs) have been studied as possible candidates, with highly contradictory results, after the demonstration that sizzled, a sFRP found in Xenopus and zebrafish, was a potent inhibitor of Xenopus and zebrafish tolloid-like proteases. In this study, we demonstrate that mammalian sFRP-1, -2, and -4 do not modify human BMP-1 activity on several of its known substrates including procollagen I, procollagen III, pN-collagen V, and prolysyl oxidase. In contrast, Xenopus sizzled appears as a tight binding inhibitor of human BMP-1, with a K(i) of 1.5 ± 0.5 nM, and is shown to strongly inhibit other human tolloid isoforms mTLD and mTLL-1. Because sizzled is the most potent inhibitor of human tolloid-like proteinases known to date, we have studied its mechanism of action in detail and shown that the frizzled domain of sizzled is both necessary and sufficient for inhibitory activity and that it acts directly on the catalytic domain of BMP-1. Residues in sizzled required for inhibition include Asp-92, which is shared by sFRP-1 and -2, and also Phe-94, Ser-43, and Glu-44, which are specific to sizzled, thereby providing a rational basis for the absence of inhibitory activity of human sFRPs.

Proenzyme structure and activation of astacin metallopeptidase.

Proteolysis is regulated by inactive (latent) zymogens, with a prosegment preventing access of substrates to the active-site cleft of the enzyme. How latency is maintained often depends on the catalytic mechanism of the protease. For example, in several families of the metzincin metallopeptidases, a "cysteine switch" mechanism involves a conserved prosegment motif with a cysteine residue that coordinates the catalytic zinc ion. Another family of metzincins, the astacins, do not possess a cysteine switch, so latency is maintained by other means. We have solved the high resolution crystal structure of proastacin from the European crayfish, Astacus astacus. Its prosegment is the shortest structurally reported for a metallopeptidase, and it has a unique structure. It runs through the active-site cleft in reverse orientation to a genuine substrate. Moreover, a conserved aspartate, projected by a wide loop of the prosegment, coordinates the zinc ion instead of the catalytic solvent molecule found in the mature enzyme. Activation occurs through two-step limited proteolysis and entails major rearrangement of a flexible activation domain, which becomes rigid and creates the base of the substrate-binding cleft. Maturation also requires the newly formed N terminus to be precisely trimmed so that it can participate in a buried solvent-mediated hydrogen-bonding network, which includes an invariant active-site residue. We describe a novel mechanism for latency and activation, which shares some common features both with other metallopeptidases and with serine peptidases.

Fetuin-A and cystatin C are endogenous inhibitors of human meprin metalloproteases.

Meprin α and ß, zinc metalloproteinases, play significant roles in inflammation, including inflammatory bowel disease (IBD), possibly by activating cytokines, like interleukin 1ß, interleukin 18, or tumor growth factor α. Although a number of potential activators for meprins are known, no endogenous inhibitors have been identified. In this work, we analyzed the inhibitory potential of human plasma and identified bovine fetuin-A as an endogenous meprin inhibitor with a K(i) (inhibition constant) of 4.2 × 10(-5) M for meprin α and a K(i) of 1.1 × 10(-6) M meprin ß. This correlated with data obtained for a fetuin-A homologue from carp (nephrosin inhibitor) that revealed a potent meprin α and ß inhibition (residual activities of 27 and 22%, respectively) at a carp fetuin concentration of 1.5 × 10(-6) M. Human fetuin-A is a negative acute phase protein involved in inflammatory diseases, thus being a potential physiological regulator of meprin activity. We report kinetic studies of fetuin-A with the proteolytic enzymes astacin, LAST, LAST_MAM, trypsin, and chymotrypsin, indeed demonstrating that fetuin-A is a broad-range protease inhibitor. Fetuin-A inhibition of meprin α activity was 40 times weaker than that of meprin ß activity. Therefore, we tested cystatin C, a protein structurally closely related to fetuin-A. Indeed, cystatin C was an inhibitor for human meprin α (K(i) = 8.5 × 10(-6) M) but, interestingly, not for meprin ß. Thus, the identification of fetuin-A and cystatin C as endogenous proteolytic regulators of meprin activity broadens our understanding of the proteolytic network in plasma.

The C-terminal region of human plasma fetuin-B is dispensable for the raised-elephant-trunk mechanism of inhibition of astacin metallopeptidases.

Human fetuin-B plays a key physiological role in human fertility through its inhibitory action on ovastacin, a member of the astacin family of metallopeptidases. The inhibitor consists of tandem cystatin-like domains (CY1 and CY2), which are connected by a linker containing a "CPDCP-trunk" and followed by a C-terminal region (CTR) void of regular secondary structure. Here, we solved the crystal structure of the complex of the inhibitor with archetypal astacin from crayfish, which is a useful model of human ovastacin. Two hairpins from CY2, the linker, and the tip of the "legumain-binding loop" of CY1 inhibit crayfish astacin following the "raised-elephant-trunk mechanism" recently reported for mouse fetuin-B. This inhibition is exerted by blocking active-site cleft sub-sites upstream and downstream of the catalytic zinc ion, but not those flanking the scissile bond. However, contrary to the mouse complex, which was obtained with fetuin-B nicked at a single site but otherwise intact, most of the CTR was proteolytically removed during crystallization of the human complex. Moreover, the two complexes present in the crystallographic asymmetric unit diverged in the relative arrangement of CY1 and CY2, while the two complexes found for the mouse complex crystal structure were equivalent. Biochemical studies in vitro confirmed the differential cleavage susceptibility of human and mouse fetuin-B in front of crayfish astacin and revealed that the cleaved human inhibitor blocks crayfish astacin and human meprin α and ß only slightly less potently than the intact variant. Therefore, the CTR of animal fetuin-B orthologs may have a function in maintaining a particular relative orientation of CY1 and CY2 that nonetheless is dispensable for peptidase inhibition.

Mammalian plasma fetuin-B is a selective inhibitor of ovastacin and meprin metalloproteinases.

Vertebrate fetuins are multi-domain plasma-proteins of the cystatin-superfamily. Human fetuin-A is also known as AHSG, α2-Heremans-Schmid-glycoprotein. Gene-knockout in mice identified fetuin-A as essential for calcified-matrix-metabolism and bone-mineralization. Fetuin-B deficient mice, on the other hand, are female infertile due to zona pellucida 'hardening' caused by the metalloproteinase ovastacin in unfertilized oocytes. In wildtype mice fetuin-B inhibits the activity of ovastacin thus maintaining oocytes fertilizable. Here we asked, if fetuins affect further proteases as might be expected from their evolutionary relation to single-domain-cystatins, known as proteinase-inhibitors. We show that fetuin-A is not an inhibitor of any tested protease. In stark contrast, the closely related fetuin-B selectively inhibits astacin-metalloproteinases such as meprins and ovastacin, but not astacins of the tolloid-subfamily, nor any other proteinase. The analysis of fetuin-B expressed in various mammalian cell types, insect cells, and truncated fish-fetuin expressed in bacteria, showed that the cystatin-like domains alone are necessary and sufficient for inhibition. This report highlights fetuin-B as a specific antagonist of ovastacin and meprin-metalloproteinases. Control of ovastacin was shown to be indispensable for female fertility. Meprin inhibition, on the other hand, renders fetuin-B a potential key-player in proteolytic networks controlling angiogenesis, immune-defense, extracellular-matrix-assembly and general cell-signaling, with implications for inflammation, fibrosis, neurodegenerative disorders and cancer.

Structure of mammalian plasma fetuin-B and its mechanism of selective metallopeptidase inhibition.

Mammalian fetuin-A and fetuin-B are abundant serum proteins with pleiotropic functions. Fetuin-B is a highly selective and potent inhibitor of metallo-peptidases (MPs) of the astacin family, which includes ovastacin in mammals. By inhibiting ovastacin, fetuin-B is essential for female fertility. The crystal structure of fetuin-B was determined unbound and in complex with archetypal astacin, and it was found that the inhibitor has tandem cystatin-type modules (CY1 and CY2). They are connected by an exposed linker with a rigid, disulfide-linked 'CPDCP-trunk', and are followed by a C-terminal region (CTR) with little regular secondary structure. The CPDCP-trunk and a hairpin of CY2 form a bipartite wedge, which slots into the active-site cleft of the MP. These elements occupy the nonprimed and primed sides of the cleft, respectively, but spare the specificity pocket so that the inhibitor is not cleaved. The aspartate in the trunk blocks the catalytic zinc of astacin, while the CY2 hairpin binds through a QWVXGP motif. The CY1 module assists in structural integrity and the CTR is not involved in inhibition, as verified by in vitro studies using a cohort of mutants and variants. Overall, the inhibition conforms to a novel 'raised-elephant-trunk' mechanism for MPs, which is reminiscent of single-domain cystatins that target cysteine peptidases. Over 200 sequences from vertebrates have been annotated as fetuin-B, underpinning its ubiquity and physiological relevance; accordingly, sequences with conserved CPDCP- and QWVXGP-derived motifs have been found from mammals to cartilaginous fishes. Thus, the raised-elephant-trunk mechanism is likely to be generally valid for the inhibition of astacins by orthologs of fetuin-B.

Activation mechanism of pro-astacin: role of the pro-peptide, tryptic and autoproteolytic cleavage and importance of precise amino-terminal processing.

Astacin (EC 3.4.24.21) is a prototype for the astacin family and for the metzincin superfamily of zinc peptidases, which comprise membrane-bound and secreted enzymes involved in extracellular proteolysis during tissue development and remodelling. Generally, metzincins are translated as pro-enzymes (zymogens), which are activated by removal of an N-terminal pro-peptide. In astacin, however, the mode of zymogen activation has been obscured, since the pro-form does not accumulate in vivo. Here we report the detection of pro-astacin in midgut glands of brefeldin A-treated crayfish (Astacus astacus) by immunoprecipitation and mass spectrometry. We demonstrate that the pro-peptide is able to shield the active site of mature astacin as a transient inhibitor, which is degraded slowly. In vitro studies with recombinant pro-astacin in the absence of another protease reveal a potential of auto-proteolytic activation. The initial cleavage in this autoactivation appears to be an intramolecular event. This is supported by the fact that the mutant E93A-pro-astacin is incapable of autoactivation, and completely resistant to cleavage by mature astacin. However, this mutant is cleaved by Astacus trypsin within the pro-peptide. This probably reflects the in vivo situation, where Astacus trypsin and astacin work together during pro-astacin activation. In a first step, trypsin produces amino-terminally truncated pro-astacin derivatives. These are trimmed subsequently by each other and by astacin to yield the mature amino terminus, which forms a salt-bridge with Glu103 in the active site. The disruption of this salt-bridge in the mutants E103A and E103Q results in extremely heat labile proteins, whose catalytic activities are not altered drastically, however. This supports a concept according to which the linkage of Glu103 to the precisely trimmed amino terminus is a crucial structural prerequisite throughout the astacin family.

Human meprin alpha and beta homo-oligomers: cleavage of basement membrane proteins and sensitivity to metalloprotease inhibitors.

Meprin is a zinc endopeptidase of the astacin family, which is expressed as a membrane-bound or secreted protein in mammalian epithelial cells, in intestinal leucocytes and in certain cancer cells. There are two types of meprin subunits, alpha and beta, which form disulphide-bonded homo- and hetero-oligomers. Here we report on the cleavage of matrix proteins by hmeprin (human meprin) alpha and beta homo-oligomers, and on the interactions of these enzymes with inhibitors. Despite their completely different cleavage specificities, both hmeprin alpha and beta are able to hydrolyse basement membrane components such as collagen IV, nidogen-1 and fibronectin. However, they are inactive against intact collagen I. Hence the matrix-cleaving activity of hmeprin resembles that of gelatinases rather than collagenases. Hmeprin is inhibited by hydroxamic acid derivatives such as batimastat, galardin and Pro-Leu-Gly-hydroxamate, by TAPI-0 (tumour necrosis factor alpha protease inhibitor-0) and TAPI-2, and by thiol-based compounds such as captopril. Therapeutic targets for these inhibitors are MMPs (matrix metalloproteases), TACE (tumour necrosis factor alpha-converting enzyme) and angiotensin-converting enzyme respectively. The most effective inhibitor of hmeprin alpha in the present study was the naturally occurring hydroxamate actinonin ( K(i)=20 nM). The marked variance in the cleavage specificities of hmeprin alpha and beta is reflected by their interaction with the TACE inhibitor Ro 32-7315, whose affinity for the beta subunit (IC50=1.6 mM) is weaker by three orders of magnitude than that for the alpha subunit ( K(i)=1.6 microM). MMP inhibitors such as the pyrimidine-2,4,6-trione derivative Ro 28-2653 that are more specific for gelatinases do not bind to hmeprin, presumably due to the subtle differences in the mode of zinc binding and active-site structure between the astacins and the MMPs.

Fetuin-B, a liver-derived plasma protein is essential for fertilization.

The zona pellucida (ZP) is a glycoprotein matrix surrounding mammalian oocytes. Upon fertilization, ZP hardening prevents sperm from binding to and penetrating the ZP. Here, we report that targeted gene deletion of the liver-derived plasma protein fetuin-B causes premature ZP hardening and, consequently, female infertility. Transplanting fetuin-B-deficient ovaries into wild-type recipients restores fertility, indicating that plasma fetuin-B is necessary and sufficient for fertilization. In vitro fertilization of oocytes from fetuin-B-deficient mice only worked after rendering the ZP penetrable by laser perforation. Mechanistically, fetuin-B sustains fertility by inhibiting ovastacin, a cortical granula protease known to trigger ZP hardening. Thus, plasma fetuin-B is necessary to restrain protease activity and thereby maintain ZP permeability until after gamete fusion. These results also show that premature ZP hardening can cause infertility in mice.

The protease domain of procollagen C-proteinase (BMP1) lacks substrate selectivity, which is conferred by non-proteolytic domains.

Procollagen C-proteinase (PCP) removes the C-terminal pro-peptides of procollagens and also processes other matrix proteins. The major splice form of the PCP is termed BMP1 (bone morphogenetic protein 1). Active BMP1 is composed of an astacin-like protease domain, three CUB (complement, sea urchin Uegf, BMP1) domains and one EGF-like domain. Here we compare the recombinant human full-length BMP1 with its isolated proteolytic domain to further unravel the functional influence of the CUB and EGF domains. We show that the protease domain alone cleaves truncated procollagen VII within the short telopeptide region into fragments of similar size as the full-length enzyme does. However, unlike full-length BMP1, the protease domain does not stop at this point, but degrades its substrate completely. Moreover, the protease domain cleaves other matrix proteins such as fibronectin, collagen I and collagen IV, which are left intact by the full-length enzyme. In addition, we show for the first time that thrombospondin-1 is differently cleaved by both BMP1 and its catalytic domain. In summary, our data support the concept that the C-terminal domains of BMP1 are important for substrate recognition and for controlling and restricting its proteolytic activity via exosite binding.

The interaction of recombinant subdomains of the procollagen C-proteinase with procollagen I provides a quantitative explanation for functional differences between the two splice variants, mammalian tolloid and bone morphogenetic protein 1.

The procollagen C-proteinase (PCP) is a zinc peptidase of the astacin family and the metzincin superfamily. The enzyme removes the C-terminal propeptides of fibrillar procollagens and activates other matrix proteins. Besides its catalytic protease domain, the procollagen C-proteinase contains several C-terminal CUB modules (named after complement factors C1r and C1s, the sea urchin UEGF protein, and BMP-1) and EGF-like domains. The two major splice forms of the C-proteinase differ in their overall domain composition. The longer variant, termed mammalian tolloid (mTld, i.e., PCP-2), has the protease-CUB1-CUB2-EGF1-CUB3-EGF2-CUB4-CUB5 composition, whereas the shorter form termed bone morphogenetic protein 1 (BMP-1, i.e., PCP-1) ends after the CUB3 domain. Two related genes encode proteases similar to mTld in humans and have been termed mammalian tolloid like-1 and -2 (mTll-1 and mTll-2, respectively). For mTll-1, it has been shown that it has C-proteinase activity. We demonstrate that recombinant EGF1-CUB3, CUB3, CUB3-EGF2, EGF2-CUB4, and CUB4-CUB5 modules of the procollagen C-proteinase can be expressed in bacteria and adopt a functional antiparallel beta-sheet conformation. As shown by surface plasmon resonance analysis, the modules bind to procollagen I in a 1:1 stoichiometry with dissociation constants (K(D)) ranging from 622.0 to 1.0 nM. Their binding to mature collagen I is weaker by at least 1 order of magnitude. Constructs containing EGF domains bind more strongly than those consisting of CUB domains only. This suggests that a combination of CUB and EGF domains serves as the minimal functional unit. The binding affinities of the EGF-containing modules for procollagen increase in the order EGF1-CUB3 < CUB3-EGF2 < EGF2-CUB4. In the context of the full length PCP, this implies that a given module has an affinity that continues to increase the more C-terminally the module is located within the PCP. The tightest binding module, EGF2-CUB4 (K(D) = 1.0 nM), is only present in mTld, which might provide a quantitative explanation for the different efficiencies of BMP-1 and mTld in procollagen C-proteinase activity.

Macromolecular isoforms of Daphnia magna haemoglobin.

The haemoglobin (Hb) of Daphnia magna acclimated to different oxygen conditions was sampled, and in its natively assembled state it was separated by chromatofocusing. The Hb isoforms were analysed for their subunit composition under denaturating conditions by two-dimensional gel electrophoresis. The Hb system is suggested to consist of three predominant Hb aggregates, which are characterised by a specific subunit composition and synthesised in response to different ambient oxygen conditions. In normoxia, a dominant Hb aggregate (DmHbI) with a pI of 4.4-4.6 was composed of subunits B, C, E, F and G. In severe hypoxia, a different dominant Hb isoform (DmHbIII) with a pI of 5.7-5.9 was composed of subunits A, B, C, D, E and F. Further analyses in moderate hypoxia provided evidence for a third Hb isoform (DmHbII) composed of subunits B, C, D, E and F. Sequence alignment and homology modelling of the tertiary structure of the D. magna Hb domains 1 and 2 revealed functionally relevant substitutions of amino acid residues at positions B10, E7 and E11, which determine the functional properties of D. magna haemoglobin in terms of haem contact, oxygen binding and affinity. Both domains are predicted to possess the common haemoglobin fold, but helices C and D are not properly formed, and helix G is interrupted by a short coil.