Unique autolytic cleavage of human myeloperoxidase. Implications for the involvement of active site MET409.

Myeloperoxidase (MPO) is a functionally important component of the normal human neutrophil host defense system. This enzyme possesses a dimeric structure composed of two heavy subunit (55-63 kDa)/light subunit (10-15 kDa) protomers, each of which is associated with a heme-like prosthetic group. In addition, MPO species of approximately 38 and 22 kDa have been reported by many different investigators, but their nature and mode of origin are not understood. In the present study, we demonstrate that when MPO is heated under nonreducing, denaturing conditions, these two species are produced via a novel autolytic cleavage of the heavy subunit. The 38-kDa species was isolated by fast-protein liquid chromatography and identified by sequencing as the carboxyl-terminal portion of the heavy subunit, and the cleavage was shown to occur exclusively between Met409 and Pro410. In order to further characterize this unusual cleavage reaction, the 22-kDa species was digested with endoproteinase Asp-N, and the peptide corresponding to its carboxyl terminus was isolated and analyzed by sequencing and mass spectrometry. These data indicated that during cleavage of the heavy subunit, Met409 was converted to homoserine lactone. Thus, the cleavage appeared to formally resemble the cyanogen bromide-dependent cleavage of Met-X peptide bonds. Recent x-ray crystallographic data for canine MPO have indicated that Met409 is in close proximity to the heme-like prosthetic group of MPO. Our studies suggest that interaction of Met409 with this group leads to the formation of a methionyl sulfonium derivative which undergoes intramolecular rearrangement with subsequent peptide cleavage under nonreducing conditions. This derivative may be, at least in part, responsible for the unusual spectral characteristics and enzymatic properties of the enzyme. The primary structure of the 22-kDa MPO species is also reported, and direct evidence is provided for asparagine-linked oligosaccharide moieties at two of the three predicted glycosylation sites.     

Human hemi-myeloperoxidase. Initial chlorinating activity at neutral pH, compound II and III formation, and stability towards hypochlorous acid and high temperature.

Human neutrophilic myeloperoxidase (MPO) is involved in the defence mechanism of the body against micro-organisms. The enzyme catalyses the generation of the strong oxidant hypochlorous acid (HOCl) from hydrogen peroxide and chloride ions. In normal neutrophils MPO is present in the dimeric form (140 kDa). The disulphide-linked protomers each consist of a heavy subunit and a light one. Reductive alkylation converts the dimeric enzyme into two promoters, 'hemi-myeloperoxidase'. We studied the initial activities of human dimeric MPO and hemi-MPO at the physiological pH of 7.2 and found no significant differences in chlorinating activity. These results indicate that, at least at neutral pH, the protomers of MPO function independently. The absorption spectra of MPO compounds II and III, both inactive forms concerning HOCl generation, and the rate constants of their formation were the same for dimeric MPO and hemi-MPO, but hemi-MPO required a slightly larger excess of H2O2 for complete conversion. Hemi-MPO was less stable at a high temperature (80 degrees C) as compared to the dimeric enzyme. Furthermore, the resistance of the chlorinating activity of hemi-MPO against its oxidative product hypochlorous acid was somewhat lower (IC50 = 32 microM HOCl) compared to dimeric MPO (IC50 = 50 microM HOCl). The higher stability of dimeric MPO in the presence of its oxidative product compared to that of monomeric MPO might be the reason for the occurrence of MPO as a dimer.     

The microheterogeneity of rabbit testosterone-binding globulin is due to differential glycosylation of its single protomer

Affinity-purified rabbit testosterone-binding globulin (rbTeBG) is a homodimer with a molecular weight (Mr) of about 92,000 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of the chemically cross-linked protein. When noncross-linked rbTeBG is subjected to SDS-PAGE, individual protomers (Mr approximately equal to 44,400 +/- 400 and Mr approximately equal to 42,000 +/- 1300) are resolved. The protomers are present in a ratio of approximately 2 (heavy):1 (light). Enzymatic deglycosylation of native rbTeBG or of rbTeBG that had been photoaffinity-labeled with [1,2-3H]17 beta-hydroxy-4,6-androstadien-3-one was conducted. The products were then identified on immunoblots using a monoclonal antibody that cross-reacts with rbTeBG, or by fluorography. These analyses indicated that rbTeBG contained sialic acid and asparagine (Asn)-linked oligosaccharides and provided evidence for the presence of serine/threonine (O)-linked glycans on the molecule. The presumptive removal of all oligosaccharides by enzymatic or chemical means resulted in the appearance of a single subunit (Mr approximately equal to 37,150 +/- 1200). On the basis of this monomeric molecular weight, carbohydrate would contribute 16% and 11% to the relative molecular mass of the nondeglycosylated heavy and light subunits, respectively. Therefore, the size heterogeneity of the nondeglycosylated rbTeBG subunits is a result of their differential glycosylation. In addition to size heterogeneity, the rbTeBG subunits are composed of multiple-charge variants. Although enzymatic and chemical methods of glycan removal altered the isoelectric points of the isoforms, none of the treatments yielded a single isoform. Thus, it is possible that moieties other than oligosaccharides are contributing charge to the isoelectric variants of rbTeBG.     

Isolation of mitochondrial succinate: ubiquinone reductase, cytochrome c reductase and cytochrome c oxidase from Neurospora crassa using nonionic detergent.

The electron transfer complexes, succinate: ubiquinone reductase, ubiquinone: cytochrome c reductase, and cytochrome c: O2 oxidase were isolated from the mitochondrial membranes of Neurospora crassa by the following steps. Modification of the contents of the complexes in mitochondria by growing cells on chloramphenicol; solubilisation of the complexes by Triton X-100; affinity chromatography on immobilized cytochrome c and ion exchange and gel chromatography. Ubiquinone reductase was obtained in a monomeric form (Mr approximately 130 000) consisting of a flavin subunit (Mr 72 000) an iron-sulfur subunit (Mr 28 000) and a cytochrome b subunit (Mr probably 14 000). Cytochrome c reductase was obtained in a dimeric form (Mr approximately 550 000), the monomeric unit comprising the cytochromes b (Mr each 30 000), a cytochrome c1 (Mr 31 000), the iron-sulfur subunit (Mr 25 000), and six subunits without known prosthetic groups (Mr 9000, 11 000, 14 000, 45 000, 45 000, and 52 000). Cytochrome c oxidase was also isolated in a dimeric form (Mr approximately 320 000) comprising two copies each of seven subunits (Mr 9000, 12 000, 14 000, 18 000, 21 000, 29 000, and 40 000). The complexes were essentially free of phospholipid. Each bound one micelle of Triton X-100 (Mr approximately 90 000). After isolation, the bound Triton X-100 could be replaced by other nonionic detergents such as: alkylphenyl polyoxyethylene ethers, alkyl polyoxyethylene ethers and acyl polyoxyethylene sorbitan esters.     

Methylamine dehydrogenase of Pseudomonase sp. J Isolation and properties of the subunits.

Two kinds of subunits, light subunit (Mr =1300) and heavy subunit (Mr=40 000), were isolated from a methylamine dehydrogenase (Mr=105 000) of Pseudomonas sp. J. The isolation of the subunits was carried out by gel chromatography after the enzyme had been treated with 3M guanidine-HCl. Coexistence of both of the subunit exhibited an absorption maximum only at 278 nm but in addition to the peak at 278 nm. The results indicate that the prosthetic group, assumed to be a derivative of pyridoxal, was bound to the light subunit. The spectral changes of the light subunit were observed by addition of methylamine. Various physical and biochemical parameters of the subunits are reported.     

Subunit structure of sex-steroid-binding plasma proteins from man, cattle, dog, and rabbit

Sex-steroid-binding plasma proteins (SBPs) of man, cattle, dog, and rabbit were purified to apparent homogeneity by sequential chromatography on testosterone-17 alpha-ethynylcarboxyaminoethyl Sepharose and hydroxyapatite. When subjected to polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate, all the purified SBPs were resolved into two subunits, the relative amounts of which differed considerably from species to species. Two-dimensional electrophoresis according to O'Farrell also revealed that each subunit was further separable into several charged variants. The heavy subunit had somewhat more acidic molecular variants than the light subunit. One molecule of 5 alpha-dihydrotestosterone was bound per dimer of the subunits. Dissociation constants of heavy and light homodimers of rabbit SBP were 3.3 and 4.9 nM, respectively. Polypeptide fragmentation patterns resulting from digestion of heavy and light subunits with protease V8 differed from species to species but resembled each other in each species. These results suggest that the native SBPs may exist as a homodimer of a single variant or a hybrid dimer composed of various combinations of light and heavy variants.     

Biosynthesis of rat renal gamma-glutamyl transpeptidase. Evidence for a common precursor of the two subunits

Rat kidney gamma-glutamyl transpeptidase is an amphipathic heterodimer, anchored to the lumenal surface of brush-border membranes via the NH2-terminal portion of its heavy subunit. The Mr values of the two subunits of detergent-solubilized enzyme are approximately 51,000 (heavy) and 22,000 (light), respectively. Biosynthesis of transpeptidase was studied in renal slices incubated with L-[35S]methionine. Transpeptidase-related proteins were isolated by immunoprecipitation with anti-transpeptidase antibodies. The major species seen after relatively short pulse times is a 78,000-dalton protein. Immunological characterization, kinetic, and pulse-chase studies indicate that the Mr = 78,000 species is the precursor of the two subunits of the enzyme. Like the dimeric enzyme, the Mr = 78,000 species contains both the core and the peripheral sugar, fucose, on its oligosaccharide moieties. Since, only the labeled dimeric enzyme appears in the brush-border membranes, conversion of the Mr = 78,000 species to the two subunits presumably occurs after its arrival at the Golgi but before its transport to the brush-border surface.     

Use of immuno-blot techniques to discriminate between the glutathione S-transferase Yf, Yk, Ya, Yn/Yb and Yc subunits and to study their distribution in extrahepatic tissues. Evidence for three immunochemically distinct groups of transferase in the rat.

The glutathione S-transferases are dimeric enzymes whose subunits can be defined by their mobility during sodium dodecyl sulphate/polyacrylamide-gel electrophoresis as Yf (Mr 24,500), Yk (Mr 25,000), Ya (Mr 25,500), Yn (Mr 26,500), Yb1 (Mr 27,000), Yb2 (Mr 27,000) and Yc (Mr 28,500) [Hayes (1986) Biochem. J. 233, 789-798]. Antisera were raised against each of these subunits and their specificities assessed by immuno-blotting. The transferases in extrahepatic tissues were purified by using, sequentially, S-hexylglutathione and glutathione affinity chromatography. Immune-blotting was employed to identify individual transferase polypeptides in the enzyme pools from various organs. The immuno-blots showed marked tissue-specific expression of transferase subunits. In contrast with other subunits, the Yk subunit showed poor affinity for S-hexylglutathione-Sepharose 6B in all tissues examined, and subsequent use of glutathione and glutathione affinity chromatography. Immuno-blotting was employed to identify a new cytosolic polypeptide, or polypeptides, immunochemically related to the Yk subunit but with an electrophoretic mobility similar to that of the Yc subunit; high concentrations of the new polypeptide(s) are present in colon, an organ that lacks Yc.     

Subunit compositions of trypsin inhibitors from tubers of taro, Colocasia antiquorum var. nymphaifolia?

Several trypsin inhibitors with different mobilities on polyacrylamide gel electrophoresis occur in the tubers of taro (Colocasia antiquorum), and they each have a dimeric molecular weight of 40,000. Of all the constituent subunits, molecular weight 20,000, of the taro trypsin inhibitor (TTI), three major subunit components were separated by chromatography on SP-Sephadex C-25 in 8 M urea, and they were named protomers alpha, beta, and gamma in the order of their elution from the SP-Sephadex column. After removal or dilution of the urea, the three protomers could be either reassociated individually or hybridized with each other to form dimeric inhibitors. All of the reassociated dimers were powerful inhibitors of trypsin. Among them, each dimer derived from protomers alpha and gamma was a weak inhibitor of chymotrypsin, whereas the dimer of protomer beta did not inhibit the enzyme. Therefore TTI is presumed to be a mixture of heterogeneous and homogenous dimers whose properties reflect those of their constituent protomers. It was also proved that the major three trypsin inhibitors (TTI-I, TTI-II, and TTI-III) previously isolated from taro tubers are composed of protomers alpha and gamma, i.e., TTI-II is a heterogeneous dimer of protomers alpha and gamma, and TTI-I and TTI-III are homogeneous dimers of protomers alpha and gamma, respectively. The molecular weight of a trypsin-TTI complex saturated with trypsin was found to be 79,000, suggesting the formation of a tetrameric complex.     

Enzymatic and Bactericidal Activity of Monomeric and Dimeric Forms of Myeloperoxidase

Enzymatic and bactericidal activities of mature, dimeric myeloperoxidase (MPO) and its monomeric form have been compared. Dimeric MPO was isolated from HL-60 cells. Hemi-MPO obtained from dimeric MPO by reductive cleavage of a disulfide bond between protomeric subunits was used as the monomeric form. Both peroxidase and halogenating (chlorinating) activities of MPO were assayed, each by two methods. Bactericidal activity of the MPO/Н₂О₂/Cl￣ system was tested using the Escherichia coli laboratory strain DH5α. No difference in the enzymatic and bactericidal activity between dimeric MPO and hemi- MPO was found. Both forms of the enzyme also did not differ in the resistance to HOCl, the main product of MPO. HOCl caused a dose-dependent decrease in peroxidase and chlorinating activity, and the pattern of this decrease was identical for dimeric MPO and hemi-MPO. At the equal heme concentration, the hemi- MPO/Н₂О₂/Cl￣ system demonstrated a somewhat higher bactericidal effect than the dimeric MPO/Н₂О₂/Cl￣ system. This is most likely explained by higher probability of contacts between the bacterial surface and hemi-MPO molecules, since at the same heme concentration the number of hemi-MPO molecules is 2-fold higher than that of dimeric MPO molecules. Using Western-blotting with antibodies to MPO, we have shown, for the first time, that the dimeric molecule of MPO could be cleaved into two monomeric subunits by HOCl, most probably due to oxidation of the disulfide bond between these subunits. This suggests that appearance in blood of MPO with mass corresponding to its monomer may result from the damage of dimeric MPO by reactive halogen species, especially upon their overproduction inducing oxidative/halogenative stress in inflammatory diseases.     

The biosynthesis and assembly of methanol dehydrogenase in bacterium W3A1

Bacterium W3A1, a restricted facultative methylotroph, produces a periplasmic methanol dehydrogenase composed of two identical subunits of Mr = 57,300, and two noncovalently bound methoxatin prosthetic groups. A precursor form of Mr = 1,500 larger than the mature subunit was identified among the products of an in vitro translation of total RNA isolated from bacterium W3A1. The precursor form of the protein could not be detected in cells during in vivo pulse-labeling studies, suggesting that the processing of this precursor occurs entirely co-translationally. Whereas the holoenzyme was detectable only as a dimer, removal of the prosthetic group yielded an apoenzyme that could be detected as either a dimeric or monomeric species. After readdition of the purified prosthetic group to the apoenzyme, only the dimeric form of the protein, bearing the cofactor and exhibiting an absorption spectrum similar to that of the holoenzyme, was detected. Neither the mature apoprotein nor the holoenzyme demonstrated any affinity for phospholipid membranes, as assayed by their inability to bind to liposomes. Taken together, these data suggest a scheme of co-translational processing and export of the apoprotein subunits, followed by assembly of the subunits and prosthetic groups in the periplasmic space to form the mature holoenzyme. The suitability of bacterium W3A1, and other methylotrophic bacteria, for use in studies of protein biosynthesis and export, is also discussed.     

Subunit structure of human and rat glutathione S-transferases

In rat tissues different forms of glutathione (GSH) S-transferases represent various dimeric combinations of at least four different classes of subunits categorized on the basis of their Mr values as seen on polyacrylamide gels. These subunit types represent heterogeneous populations and the actual number of subunits in rat GSH S-transferases may be far more than is known at present. Human GSH S-transferases arise from dimeric combinations of at least four immunologically and functionally distinct subunits which can be classified into three types, A (Mr 26,500), B (Mr 24,500) and C (Mr 22,500). There is evidence for considerable charge heterogeneity in each of these subunit types.     

The neural type II regulatory subunit of cAMP-dependent protein kinase is present and regulated by hormones in the rat ovary

In this study purified isoforms of rat ovarian regulatory subunit of type II cAMP-dependent protein kinase (R-II) were compared with R-II purified from rat brain. A special neural form of R-II has been previously described in bovine brain. Analysis by one-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis resolved three isoforms of rat ovarian R-II (R-II54, Mr = 54,000; R-II52, Mr = 52,000; and R-II51, Mr = 51,000) compared to two R-II isoforms in rat brain (R-II54 and R-II52). Polychromatic silver-stained peptide maps of purified R-II subunits indicated that peptides generated from both rat ovarian R-II52 and R-II51 were similar (if not identical) to the peptides of the neural form, R-II52, purified from rat brain. These peptides differed markedly from those generated from R-II54 of either rat ovary, brain, or heart. Ovarian R-II52/51 photoaffinity labeled with 8-N3-[32P]cAMP and analyzed by two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis was shown to consist of three (rather than two) isoelectric variants, which were similar to three variants resolved from rat brain R-II and clearly distinct from that of rat heart R-II54. An antibody which recognized both the R-II54 and R-II52/51 isoforms of rat ovarian extracts also recognized both forms of rat brain R-II (R-II54 and R-II52) and similar forms in extracts of rat adrenal and parotid glands. These results strongly suggest that the R-II52 isoform previously designated as a neural specific form of R-II is present in high concentrations in a nonneural tissue, the rat ovary.     

The reductive cleavage of myeloperoxidase in half, producing enzymically active hemi-myeloperoxidase

Reduction and alkylation of human myeloperoxidase under nondenaturing conditions results in the cleavage of this enzyme. Sedimentation equilibrium data is presented which shows that the molecular weight of the cleavage product (78,000 +/- 2,000) is half that of the native enzyme (153,000 +/- 4,000). We conclude that the cleavage product is the half-enzyme hemi-myeloperoxidase. Hemi-myeloperoxidase retains both heme groups and contains both subunit types (Mr = 57,500 and 14,000) in the same ratio as native myeloperoxidase. The two halves of native myeloperoxidase are apparently not dependent upon one another for peroxidatic activity, as the specific activity of the half-enzyme is the same as that of the native enzyme. Analytical ultracentrifugation studies show native myeloperoxidase has a sedimentation coefficient of 8.0 and an axial ratio of 5:1, while hemi-myeloperoxidase has a sedimentation coefficient of 4.3 and an axial ratio of 10:1. When [3H]iodoacetic acid was used to prepare hemi-myeloperoxidase, the label incorporated with a stoichiometry of 1.2 [3H]carboxymethyl groups per hemi-myeloperoxidase, with 90% of this label associated with the heavy subunit. From these observations we conclude that native myeloperoxidase contains two heavy-light protomers, which are joined along their long axes by a single disulfide bond between the heavy subunits. Selective reduction of this disulfide bond by the use of nondenaturing conditions results in the formation of hemi-myeloperoxidase, a catalytically active heavy-light protomer of native myeloperoxidase.     

Assembly of dimeric myeloperoxidase during posttranslational maturation in human leukemic HL-60 cells

Myeloperoxidase is a major protein component of the azurophilic granules (specialized lysosomes) of normal human neutrophils and serves as part of a potent bactericidal system in the host defense function of these cells. In normal, mature cells, myeloperoxidase occurs exclusively as a dimer of Mr 150,000 while in immature leukemia cells, there are both monomeric (Mr 80,000) as well as dimeric species. Like other lysosomal enzymes, myeloperoxidase is synthesized as a larger glycosylated precursor (Mr 91,000) that undergoes processing through single-chain intermediates (Mr 81,000 and 74,000) to yield mature heavy (Mr 60,000) and light (Mr 15,000) subunits. To study the assembly of dimeric myeloperoxidase, azurophilic granules were isolated from either unlabeled or pulse-labeled ([35S]methionine/cysteine) HL-60 cells, and myeloperoxidase was extracted and separated into monomeric and dimeric forms by FPLC gel filtration chromatography. Steady-state levels of dimeric and monomeric myeloperoxidase were found to account for 67% and 33%, respectively, of the total peroxidase activity and were correlated with the levels of associated heme as measured by absorption at 430 nm. Labeled myeloperoxidase polypeptides were immunoprecipitated using a monospecific rabbit antibody and were identified and quantitated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis/fluorography and liquid scintillation counting. After a 2-h pulse, labeled myeloperoxidase species of Mr 74,000 and 60,000 were found in fractions coeluting with the monomeric form of myeloperoxidase.(ABSTRACT TRUNCATED AT 250 WORDS)     

Hominoid triosephosphate isomerase: Regulation of expression of the proliferation specific isozyme

Summary Three primary isoforms of the dimeric glycolytic enzyme, triosephosphate isomerase (TPI; EC 5.3.1.1), are detected in proliferating human cells. The electrophoretically separable isoforms result from the three possible combinations of constitutive subunits and subunits expressed only in proliferating cells. Only a single primary isoform is observed in quiescent cells. The two subunits, which differ by covalent modification (s), are products of the single structural locus for this enzyme. Expression of the proliferation specific subunit (TPI-2) is detected within 6–10 hr following mitogen stimulation of quiescent human cells, requires RNA synthesis and is inhibited by agents which inhibit interleukin 2 expression or function. Only the constitutive subunit (TPI-1) is detected in proliferating cells from nonhominoid primate species. A single class of TPI mRNA, which is increased > 10 fold following stimulation of quiescent cells, is detected on northern blot analysis and S1 nuclease digestion analysis of RNA from quiescent and proliferating human cells. It is similar in size to the TPI mRNA from proliferating cells of the African green monkey, a primate species not expressing TPI-2. Comparison of the structure of the TPI gene from rhesus monkey (nonexpressing species) to the gene from expressing species does not suggest a mechanism for generating TPI-2. Thus, the regulation of the expression of the hominoid restricted, proliferation specific subunit of TPI has been further defined, although the mechanism for generating TPI-2 remains elusive.     

Purification and characterization of hormone-regulated isoforms of the regulatory subunit of type II cAMP-dependent protein kinase from rat ovaries

The regulatory subunit (R-II) of cAMP-dependent protein kinase type II is induced in rat ovarian granulosa cells by the synergistic actions of estradiol and follicle-stimulating hormone. The R-II from rat ovaries was compared with R-II from rat heart, rat brain, bovine heart, and bovine brain using immunological methods, 8-N3[32P]cAMP photoaffinity labeling and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Three isoforms of R-II were identified in rat ovarian cell extract (R-II54 Mr = 54,000, R-II52 Mr = 52,000, R-II51 Mr = 51,000), two isoforms of R-II in rat brain cell extract (Mr = 54,000, Mr = 52,000), and one isoform of R-II in rat heart cell extract (Mr = 54,000). Rat ovarian R-II54, heart R-II, and brain R-II (Mr = 54,000) were recognized by antiserum against rat heart R-II, whereas rat ovarian R-II52/R-II51 and rat brain R-II (Mr = 52,000) were not. In contrast, an antiserum raised against bovine heart R-II recognized all three isoforms of ovarian R-II as well as the lower molecular weight form of rat brain R-II. Ovarian types R-II52 and R-II51 but not R-II54 were increased selectively in granulosa cells by estradiol and follicle-stimulating hormone. In addition: 1) ovarian R-II52/51 subunits were purified to homogeneity and shown to recombine with C subunit from bovine heart to form a cAMP-dependent protein kinase; 2) pure R-II52/51 were not interconvertible to a higher molecular weight form by C subunit-dependent phosphorylation; 3) pure rat heart R-II (Mr = 54,000) and ovarian R-II52/51 exhibited distinct differences based on one- and two-dimensional peptide mapping; and 4) by two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis pure R-II52/51 were resolved as three (rather than two) isoelectric variants which were clearly different from pure rat heart R-II54. Thus, the hormone-regulated form of R-II in rat ovarian granulosa cells appears to represent a gene product distinct from R-II54 in rat heart.     

Hemocyanin from the Australian freshwater crayfish Cherax destructor. Electron microscopy of native and reassembled molecules.

Examination and measurement of electron micrographs of negatively stained hemocyanin molecules from Cherax destructor show that the predominant aggregated forms, the 16S and 24S components, are typical structures for arthropod hexamers and dodecamers, respectively. In Cherax hemocyanin the hexamers are formed from the monomeric (Mr congruent to 75,000) subunits, M1 and M2, while the dodecamers contain in addition a dimeric (Mr congruent to 150,000) subunit, M3'. Studies of the composition of solutions of the subunits M1 and m2 to which calcium ions have been added at pH 7.8 show that, under these conditions, reassembly occurs to particles indistinguishable from native hexamers. It is noteworthy that dodecamers are not seen since this confirms the previous suggestion that incorporation of the dimeric subunit in the assembly process is necessary for their formation. The results obtained from Cherax hemocyanin are related to those of previous structural studies of arthropod hemocyanins. In particular, the possible controlling role of certain specific subunits in arthropod hemocyanin oligomers containing more than one kind of subunit is illustrated with a model for the Cherax dodecamer, in which the dimeric subunit is shared between the two halves of the molecule.