The structure of arthropod and mollusc hemocyanins.

The hemocyanins from molluscs and from arthropods differ in the size of their polypeptide chains. A variety of physical techniques including sodium dodecyl sulfate polyacrylamide gel electrophoresis and column chromatography in sodium dodecyl sulfate and guanidine HCl indicate that the polypeptide chain of mollusc hemocyanin has a molecular weight of 290,000. These results were corroborated by quantitative end group analyses. Several experiments designed to count the number of tryptophan and methionine-containing peptides in the hemocyanin from the whelk Busycon canaliculatum indicate that sequence homology within the polypeptide chain of the mollusc hemocyanins accounts for their large size. Digestion of the native protein with subtilisin produces a 50,000-dalton fragment in high yield which corresponds to one binding site for oxygen. On the other hand, the polypeptide chain molecular weight of lobster hemocyanin is 76,000 to 78,000 and this seems to be a general property of all arthropod hemocyanins. The pigment from lobster consists of two very similar polypeptide chains which are not present in equal amount. Analysis of the cysteine-containing and of the tryptophan-containing tryptic peptides confirms the value of the molecular weight. However, separation of fragments which contain methionine indicates that there is sequence homology withing the polypeptide chain of this protein. It is concluded that the mollusc and arthropod hemocyanins have little structural similarity.


SUMMARY
The hemocyanins from molluscs and from arthropods differ in the size of their polypeptide chains. A variety of physical techniques including sodium dodecyl sulfate polyacrylamide gel electrophoresis and column chromatography in sodium dodecyl sulfate and guanidine HCl indicate that the polypeptide chain of mollusc hemocyanin has a molecular weight of 290,000. These results were corroborated by quantitative end group analyses. Several experiments designed to count the number of tryptophan and methionine-containing peptides in the hemocyanin from the whelk Busycon canalicufafum indicate that sequence homology within the polypeptide chain of the mollusc hemocyanins accounts for their large size. Digestion of the native protein with subtilisin produces a 50,000-dalton fragment in high yield which corresponds to one binding site for oxygen.
On the other hand, the polypeptide chain molecular weight of lobster hemocyanin is 76,000 to 78,000 and this seems to be a general property of all arthropod hemocyanins.
The pigment from lobster consists of two very similar polypeptide chains which are not present in equal amount. Analysis of the cysteine-containing and of the tryptophan-containing tryptic peptides confirms the value of the molecular weight. However, separation of fragments which contain methionine indicates that there is sequence homology within the polypeptide chain of this protein. It is concluded that the mollusc and arthropod hemocyanins have little structural similarity.
The second largest class of invertebrate respiratory pigments, the hemocyanins, are of interest because of their occurrence in two very diverse phyla: the arthropods and the molluscs. While copper-containing proteins are widespread in nature, functioning both as enzymes and as components of electron-transfer systems, it is remarkable that their role as oxygen carriers is unique to the members of these phyla.
The hemocyanins are large molecules, with masses of 1 to 9 x lo6 daltons, which contain many copper atoms and many O2 binding sites. A central question concerning these molecules is the size of the polypeptide chains which compose them. In this * This work was supported by United States Public Health Service Grant HL 08893 and bv National Science Foundation Grant GB 36827 (to Guido Guidocti).
$ Supported by National Institutes of Health Predoctoral Training Grant administered by the Committee on Higher Degrees in Biophysics, Harvard University.
paper evidence is presented to answer this question for the mollust and the arthropod pigments. The results show that arthropod hemocyanin is made of small polypeptide chains (75,000 daltons) each of which can bind one 02 molecule, while the mollust hemocyanins are composed of giant polypeptide chains (300,000 daltons) each with six 02 binding sites. MATERIALS

AND METHODS
The sources of all reagents and their preparation have been described (1). Chemicat &fodi$cations-Performic acid oxidation and reduction and alkylation with iodoacetamide were carried out as previously described (1). Cleavage with CNBr was performed essentially as described by Waxdal et al. (2), and the extent of reaction was estimated by the disappearance of methionine after acid hydrolysis as well as by the recovery of homoserine and homoserine lactone.
Periodate oxidation of the sugar moieties of the hemocyanins was performed exactly as described by Komatsu et al. (3). After stopping the reaction with polyethylene glycol, the protein was prepared for Na dodecyl-SO, polyacrylamide gel electrophoresis as described-berow.
Protein modification by succinylation was performed as described by Klotz (4), and acetylation experiments were carried out according to Bucci et al. (5). The disruptive effects of these reagents was determined by Na dodecyl-SOI polyacrylamide gel electrophoresis.
Modification of protein sulfhydryl groups with ethyleneimine after reduction was performed as described by Cole (6), except that 6 M guanidine HCl was used to denature the protein.
The methionine residues of Homarus hemocyanin were labeled with iodo[l-"Clacetamide (5 to 12 mCi/mmol) INew England NUclear) bv a modification of the method of Wilkinson (7) as discussed b"y Platt et aE. (8). Modified proteins were dialyzkd against 1% ammonium bicarbonate to remove excess reagents before digestion with trypsin.
Following two-dimensional paper electrophoresis and chromatography, autoradiographs were prepared with Kodak RP Royal X-Omat x-ray film.

Chemical
Analyses-Amino sugars were quantitated on the short column (12 cm) of a Beckman model 12OC amino acid analyzer using the pH 4.25 citrate buffer of the long column. Hydrolyses were carried out for several times between 4 and 20 hours in 3 N HCl with p-alanine as an internal standard. In this system, 8alanine elutes at 36 min, glucosamine at 44 min, and galactosamine at 49 min.
Neutral sugars were assayed by the phenol-sulfuric acid method of Dubois et al. (9) using mannose as standard. By scaling the volumes used by a factor of 5, 1 ag of sugar gives an absorbance of 0.1 at 489 nm. Sialic acid was determined bv the calorimetric assay of Warren (10).
Isolation of the neutral sugars from Busycon hemocyanin was Derformed as described bv Suiro (14) followine hvdrolssis of 10 kg of protein in 1 N HzSb, fir 8 hours at 110'~ Tie sugars were then converted to their 0-trimethylsilyl derivatives as described by Reinhold (15) and were analyzed on a Perkin-Elmer 990 chromatograph. Amino acid analyses were performed as previously described (1). Half-cystine  and methionine  were  determined  as cysteic  acid and methionine  sulfone  following  performic  acid oxidation.  The value for half-cystine  was corroborated  by determining  carboxymethylcysteine  following  reduction  and alkylation  with  pearing as two closely spaced bands after gel electrophoresis in FIG. 1 (top). Na dodecyl-SO, polyacrylamide gel electrophoresis of arthropod hemocyanins. Na dodecyl-SO, polyacrylamide gels (7.5% acrylamide, 0.135% methylene bisacrylamide) of purified arthropod hemocyanins were run in borate-acetate buffer, pH 8.5, for 5 hours. Left group, from left to right: Limulus hemocyanin; Homarus hemocyanin; Libinia hemocyanin; Cancer hemocyanin and bovine serum albumin. In the set at right, Libinia hemocyanin was subjected to co-electrophoresis with bovine serum albumin (68,900 daltons) and with transferrin (76,000 daltons) in order to show that the apparent molecular weights of these proteins on Na dodecyl-SO, gels was in fact 75,900. It can be seen that Libinia hemocyanin migrates to a position between these two standards. Na dodecyl-Sod. When the amount of protein in each peak of lobster hemocyanin was estimated by amino acid analysis of gel slices, the mass ratio of the larger to the smaller polypeptide was 1.5 to 1.0. Cross-linking experiments done on native lobster hemocyanin with dimethyl suberimidate resulted in the production of a single type of dimer. These results suggest that both types of chains are present in the same molecule. The values of the molecular weights were confirmed by the results of end group analyses. The NHz-terminal amino acid of octopus hemocyanin is aspartic acid according to the dansyl method. Quantitative end group analysis by the cyanate method showed that one mole of aspartic acid is recovered per 320,000 g of protein.
On the other hand, the NH,-terminal group of lobster hemocyanin is blocked. Accordingly, the procedure of Press et al. (29) was used to isolate the blocked NH,-terminal peptide obtained by digestion of the protein with prouase. One peptide was obtained by this procedure with the electrophoretic properties and the amino acid composition shown in Table I, which indicate that the main component is a dipeptide containing aspartic and glutamic acid.
The high anodic mobility of this dipeptide composed of aspartic acid and glutamic acid suggests that pyrrolidone carboxylic acid is not the NHpterminal residue and both amino acids are present as the acid and not as the amide. Treatment with carboxypeptidase A released no amino acids, but because this enzyme does not readily attack at acidic residues, this was not unexpected. Consequently, carboxypeptidase Y was employed and released aspartic acid which was identified on the amino acid analyzer. Tentatively, the following structure may be assigned to the dipeptide: X-Glu-Asp, where X is probably an acetyl group, although this has not been shown explicitly. To quantitate the yield of this peptide, an aliquot of a similarly digested sample was taken for amino acid analysis and the remainder was put onto Dowex as before. The flow-through was then collected, lyophilized, and acid hydrolyzed. The result of this experiment (Table I) shows that approximately one mole of dipeptide was recovered per 76,000 g of protein which is in good agreement with the molecular weight obtained by physical methods.
The COOH-terminal amino acids were released with carboxypeptidases A and B, and quantitated on the amino acid analyzer. The results obtained (not shown here) are certainly compatible with the measured molecular weights of the polypeptides, but they cannot be used as clear evidence for their sizes.

Polypeptide
Chain Structure-Additional evidence for the molecular weights of the polypeptides and information on the structure of the polypeptide chains were sought by determining the number of peptides which contain cysteine, tryptophan, and methionine residues. According to the amino acid compositions shown in Table II, the number of half-cystine, tryptophan, and methionine residues per molecule are 5, 10, and 12 for lobster hemocyanin and 42, 60, and 63 for Busycon hemocyanin.
Lobster Hemocyanin-Lobster hemocyanin contains one free -SH group and 2 cystine residues, as shown in Table III. Fractionation of the tryptic peptides obtained from lobster hemocyanin after reduction and alkylation of the -SH groups with iodo[l-14C]acetamide is shown in  contains the peptide with the free SH group, as deduced from fractionation of peptides obtained from alkylated, but not reduced, hemocyanin. Its purification was carried out independently on this material.
The considerable overlap on the Aminex column shows up quite clearly on DEAE-Sephadex chromatography (Fig. 7) where several of the radioactive peptides appear in more than one fraction. Thus, II, and III, are identical, as are IIb and ZIzb, ZZI, and IV,, and ZZId and Ivb. From the amino acid composi- Elurion Volume (ml) FIG. 4. Na dodecyl-SO1 column chromatography of Octopus vulgaris hemocyanin. Eight to ten milligrams of reduced and Na dodecyl-Sod-denatured protein were applied to a column of Sepharose 4B in 0.2yo Na dodecyl80,.
The chromatographic conditions are exactly as described in Fig. 3. The component at 390 ml contained no amino acids after acid hydrolysis, and is probably an artifact due to light scattering by the Na dodecyl-SO, micelle.
tions (Table IV) it can be seen that ZIb is the incomplete degradation product of I.
The isolation of 5 peptides confirms the molecular weight found by several other methods. Moreover, when this experiment was performed on the higher molecular weight component of lobster hemocyanin (Fig. 2), an identical elution profile on Aminex was found. This shows that both chains contain the same arrangement of cysteine residues, and that one chain probably is not contributing 2 unique peptides and the other, 3.
The question of whether the disutide bonds are i&a-or interchain was resolved by running the protein on Na dodecyl-SO4 gels without prior reduction. Because no difference was observed, it is probable that the disulfides are intrachain.
The tryptophan peptides were identified after separation of a ing in 6 M guanidine HCl, 0.1% 2-mercaptoethanol, and incubated overnight at 37". The denatured protein was applied to a Sepharose 4B column (1.5 X 45 cm) in this solvent and eluted at 2.5 g/hour; 1.2 to 1.3 g fractions were collected. The standards, which were prepared as described above, included myosin (220,000 daltons), muscle phosphorylase (92,000 daltons), bovine serum albumin (68,000 daltons), and bovine erythrocyte carbonic anhydrase (30,000 daltons). (A) denotes the calibration curve for the Na dodecyl-SO, column; (B) designates that for the guanidine column. FIG. 5. Na dodecyl80, polyacrylamide gel electrophoresis of mollusc hemocyanins. Na dodecyl-SO1 polyacrylamide gels (5% acrylamide, 0.15y0 methylene bisacrylamide) of the purified mollust hemocyanins examined in this study were run for 5 hours in borate-acetate buffer, pH 8. 5. From left to right: Busycon hemocyanin which had been incubated for 1 week in 700/o formic acid as described by Dijk et al. (28) 43.6 nmol of digested protein were applied to a column of Dowex 5OW-X2 (0.9 X 20 cm). The composition is based on a 24-hour acid hydrolysate, and analyses of controls containing only pronase or hemocyanin which had been treated in the same way as the digest gave negligible amounts of amino acids. tryptic digest of reduced and alkylated protein on DEAE-cellulose, followed by electrophoresis on Whatman No. 1 paper of the fractions that absorbed light at 280 nm (Fig. 8). Tryptophan peptides were detected by Ehrlich's reagent, and were found only in the first two ultraviolet-absorbing fractions obtained from the DEAE-cellulose column. When the material which remained at the origin in Fraction II was subjected to electrophoresis at pH 8.9 (0.1 M ammonium bicarbonate titrated with ammonia), none of the tryptophanpositive material moved. However, chromatography in l-butanol-pyridine-acetic acid-water resolved two more peptides and left no material at the origin (Fig. 8). Thus, a total of 7 tryptophan peptides could be detected of the 10 or 11 expected from the amino acid composition.
Two approaches were taken toward analyzing the methioninecontaining peptides of lobster hemocyanin. The first involved cleavage of the reduced and alkylated protein with CNBr as described by Waxdal et al. (2). Amino acid analysis shobed that greater than 95% of the methionine had been derivatized When a sample of CNBr-cleaved protein was taken up in 10% Na dodecyl-SO4 and the fragments were separated on Na dodecyl-SO, polyacrylamide gels, only four strongly staining bands were seen (Fig. 9).
Because the experiments with CNBr could account for only half of the expected number of cleavage products (unless there are several Met-Met sequences), a second approach to determine the number of unique methionine-containing peptides was tried. This entailed the specific alkylation of methionine residues at pH 3.5 to form the carboxymethylsulfonium derivative with iodo[l-14C]acetamide.
Although no attempt was made to separate the peptides, it was noted that all of them were retarded on a column of P-2, and all are therefore smaller than 10 to 15 residues.
When this material was separated by two-dimensional electrophoresis-chromatography, only 4 spots were detected by autoradiography (Fig. 10). The radioactive areas were then cut from the paper and placed in scintillation vials, and eluted into 0.5 ml of 1 M acetic acid at 37" before counting. The distribution of counts (Peptide 1, 14,722 cpm; Peptide 2, 6,850 cpm; Peptide 3, 8,200 cpm; Peptide 4,6,943 cpm) shows that one peptide contains approximately twice the number of counts as any other one. This confirms the conclusion, deduced from the isolation of free homoserine in the CNBr digest, that a Met-Met sequence exists in the protein. However, the number of methionine peptides is only half the expected number.

Busycon
Hemocyanin-Fractionation of a tryptic digest of Busycon hemocyanin on DEAE-cellulose and paper electrophoresis of the ultraviolet-absorbing peaks resolved 10 tryptophanpositive regions (Fig. 11) which could not be further resolved by chromatography.
The cleavage of reduced and alkylated protein with CNBr was carried out as described by Waxdal et al. (2). Following lyophilization, the residue was extracted with water and the soluble peptides were submitted to paper electrophoresis at pH 1.9. It was possible to identify one ninhydrin-positive spot in this fraction. A portion of the insoluble fraction was taken up in 10% Na dodecyl-SO* and when the CNBr fragments were separated on Na dodecyl-SOc polyacrylamide gels, only five strongly staining bands were seen (Fig. 12). A second aliquot was dissolved in 8 M urea and subjected to electrophoresis on polyacrylamide gels at pH 4.5 (Fig. 12). Eight to eleven bands could be identified in this system. According to these results, at best, only 36 of the expected number of peptides was obtained, indicating that the molecule is made up of repeating polypeptide fragments.
Additional evidence for this conclusion was obtained by limited proteolysis of Busycon hemocyanin. When the native protein was subjected to digestion with subtilisin, its size decreased. As shown in Fig. 13A, 40% of the material had an apparent molecular weight of 50,000, and 60% had a molecular weight of 100,000 by gel filtration on Sephadex G-100.
The material with a molecular weight of 50,000 was purified by chromatography on DEAE-cellulose (Fig. 13B). The column fractions were analyzed by Na dodecyl-SO4 gel electrophoresis. As shown in Fig. 14, the main peak in Fig. 13B contained mainly material with a molecular weight of 50,000, although there was also a significant amount of low molecular weight (25,000) protein. Gel filtration on Sephadex G-100 in 5 M urea (Fig. 13C) 6. Fractionation of the sulfhydryl-containing tryptic peptides of lobster hemocyanin. A tryptic digest of lobster hemocyanin (30 mg) in which the sulfhydryl groups had been reduced and alkylated with iodo[l-'"Clacetamide was applied to a column of Aminex AG 5OW-XY (0.9 X 15 cm) at room temperature which had been equilibrated with 0.1 M pyridine-acetate, pH 3.1. After 100 ml of starting buffer had been pumped through the column, a linear gradient was begun which consisted of 200 ml of the starting buffer and 200 ml of 2.0 M pyridine-acetate, pH 5.0. The flow rate was 18 ml/hour and 2.5-ml fractions were collected. The fractions designated I to IV were taken for further purification. The chromatogram was analyzed by the ninhydrin method; 100 ~1 were taken from alternate fractions for alkaline hydrolysis. with 75 mM ammonium bicarbonate, pH 8.0. After 50 ml of starting buffer had been pumped through the columns, a linear gradient consisting of 160 ml each of starting buffer and 1.0~ ammonium bicarbonate was begun. The flow rate was 12 to 14 ml/ hour, and 4-ml fractions were collected.
the main fraction from the DEAE-cellulose column produced a major peak of material which contained only the 50,000-dalton protein (Fig. 15).
Only tyrosine and histidine were detected as NHz-terminal residues by the dansyl technique. The amino acid composition of this component is not significantly different from intact hemo- Hydrolyses were in 6 N HCl and 0.1% merceptoacetic acid for 24 hours; no corrections have been made for the destruction of any amino acids; residues which were present in less than gc mol per mol of peptide were omitted. Peptide III, was stained for tryptophan using Ehrlich's reagent. of aliquots of Fractions Z-IV was carried out at pH 5.7, 2500 volts, for 1 hour on Whatman No. 1 paper. When the peptides were stained for tryptophan with Ehrlich's reagent, only Fractions Z and ZZ gave positive results (top). Chromatography in 1-butanol-pyridine-acetic acid-water of the material which remained at the origin after electrophoresis of Fraction ZZ resolved two more tryptophan-positive peptides (bottom).
cyanin (Table II), and the assay for neutral sugar was positive. It therefore appears that the large polypeptide of Busycon hemocyanin can be split into 50,000-dalton fragments (34 of the molecule). This result along with that on the number of tiyptophan and methionine peptides makes a strong case for the view that the 300,000-dalton polypeptide is composed of six 50,000-dalton units held together by peptide bonds. This is a case of g-fold sequence homology.

Polypeptide Chain
Carbohydrate-The small amount of carbohydrate present in lobster hemocyanin (Table II) suggested FIG. 9 (left). Na dodecyl-SO, polyacrylamide gel electrophoresis of lobster hemocyanin CNBr fragments. After treatment with CNBr and lyophilization to remove excess reagent and formic acid, a sample of reduced and alkylated hemocyanin was dissolved in 5% Na dodecyl-SO, and subjected to electrophoresis on 14-cm Na dodecyl-SO, polyacrylamide gels (15% acrylamide, 0.405% methylene bisacrylamide) for 21 hours in borate-acetate buffer, pH 8. 5. The anode is at the bottom. The 4 major bands have approximate molecular weights of 20,000, 10,000,7,000, and 4,000. FIG. 10 (right). Two-dimensional map of the methionine-containing tryptic peptides of lobster hemocyanin. After lyophilization of the radioactive fraction obtained by gel filtration, the methionine-containing tryptic peptides which had been alkylated with iodo[l-YJ]acetamide were applied to a sheet of Whatman No. 1 paper (22 X 37 cm). Electrophoresis was carried out in Varsol-cooled tanks at pH 5.7 (1 hour, 2500 volts), and was followed by descending chromatography in l-butanol-pyridineacetic acid-water for 16 hours. The radioactive peptides were detected by autoradiogrsphy for 48 hours. Reduced and alkylated hemocyanin (25 to 30 mg) was digested with trypsin and the product was applied to a column of DEAE-cellulose (25 ml) in 10 mM ammonium bicarbonate, pH 8.4. After 150 ml of starting buffer had been pumped through the column, a linear gradient was begun which consisted of 125 ml each of starting buffer and 0.5 M ammonium bicarbonate, pH 7.9. The flow rate was 12 ml/hour and 2.0-to 2.5-ml fractions were collected. Every fourth tube was lyophilised and the contents dissolved in 200 ~1 of buffer; each sample was then applied to a sheet of Whatman No. 1 paper and electrophoresis was carried out at pH 5.7 for 1 hour at 3000 volts. The maps were then stained with Ehrlich's reagent to locate the tryptophan peptides. FIG. 12 that all of it might be attached to one amino acid, and it was decided to perform a tryptic cleavage of reduced and alkylated protein, and then to purify this peptide by conventional means. This was done in 3 steps: gel filtration on G-25, in which case the peptide eluted ahead of most of the other tryptic peptides, and then on G-50; this material was then further purified on DEAEcellulose, and gave one ninhydrin-positive fluorescent spot upon paper electrophoresis.
It was clear from the amino acid composition ( Table V) that the peptide was not completely cleaved by trypsin; this is not unusual if either lysine or arginine is near the site of carbohydrate attachment.
The huge size of this peptide, however, made it difficult to resolve the question of how many sites of attachment for carbohydrate there are in lobster hemocyanin. Thus, in a similar experiment, the carbohydrate-containing fraction from a separation on G-50 of the tryptic peptides was pooled and digested with pronase for 24 hours. This material was eluted from Sephadex G-25, the appropriate fractions were pooled, and were rerun on the same column. The composition of this material, which was pure on electrophoresis, is given in Table V. Thus there is probably only one carbohydrate side chain.
Attempts to fractionate a tryptic digest of Busycon hemocyanin in order to isolate the glycopeptides were unsuccessful. After several steps of gel filtration and ion exchange chromatography two impure peptides could be resolved (data not shown). The presence, however, of mannose and glucosamine was particularly interesting because several workers (32,33) have used the intact hemocyanin molecule as an electron-dense marker for the lectin concanavalin A when the latter was bound to the membranes of animal cells. This application suggested, based on the specificity of concanavalin A for cu-n-glucosamine and a-Dmannose (34), that it should be possible to purify one or more of the glycopeptides by affinity chromatography on a column of Sepharose-bound concanavalin A. When this was done, the material that eluted with 0.1 M cY-methylmannoside gave only one highly fluorescent spot upon electrophoresis which stained with both ninhydrin and silver nitrate (Fig. (16). This does not eliminate the possibility that this spot may include more than one electrophoretically similar peptide, although the amino acid composition of this fraction is very simple ( Table V) . The yield of the glycopeptide indicates that there is one glycopeptide per 50,000 g of protein, or carbohydrate chains per polypeptide chain.
The carbohydrate composition of the glycopeptides for both FIG. 14. Na dodecyl-SO, polyacrylamide gel electrophoresis of samples taken from fractionation on DEAE-cellulose of the 50,000dalton material. Na dodecylS0, gel electrophoresis of DEAEcellulose column fractions was performed as described in Fig. 1

DISCUSSION
The experiments described here were designed to determine the subunit structure of the arthropod and the mollusc hemocyanins. The results show that the polypeptide chain molecular weights for arthropod and mollusc hemocyanins are 75,000 and 290,000, respectively. These values were obtained by gel filtration, gel electrophoresis, and by quantitation of the NH2-and COOH-terminal amino acids. Values obtained by other investigations (28,35-39) appear to be incorrect.
In the hemocyanins, 1 O2 molecule is bound by 2 copper atoms. There are 2 copper atoms per 75,000 daltons of arthropod hemocyanin and per 50,000 daltons of mollusc hemocyanin. Accordingly, one polypeptide chain of arthropod hemocyanin binds 1 02 molecule, while one polypeptide chain of mollusc hemocyanin binds six 02 molecules.
Both polypeptides appear to have some degree of sequence homology. The peptide fractionation experiments of Pickett et al. (39) and Nardi et al. (40) on lobster hemocyanin have shown that only the number of tryptic peptides which could account for 35,000 daltons of protein could be detected. This result is concordant with the isolation of the small number of cyanogen bromide peptides reported here and suggests that the 75,000-dalton polypeptide contains regions with sequence homology. On the other hand, attempts to fractionate the CNBr fragments and the tryptophan-containing tryptic peptides of Bwycon hemocyanin suggest that the minimum chemical molecular weight is 50,000. The only way to reconcile the values ob-   16. Purification of a glycopeptide of Busycon hemocyanin. In this experiment, the tryptic peptides that eluted in the void volume of a Sephadex G-25 column were applied to a 3.5-ml column of Sepharose-bound concanavalin A at 4", and were allowed to equilibrate with the column for 30 min before elution was begun with 0.1 M ammonium bicarbonate, pH 7.9. At Fraction 14 the buffer was changed to one that contained 0.1 M a-methyl mannoside. The flow rate was 36 ml/hour and 3-ml fractions were collected.
tained by chemical and physical methods is to postulate that there must be a substantial amount of sequence homology to the extent that this hemocyanin is duplicated 6 times within the polypeptide chain of 290,000 daltons.
This interpretation, in fact, is given added support by the results of limited proteolysis of the native hemocyanin with subtilisin. It is clear from column chromatography and Na dodecyl-SO4 polyacrylamide gel electrophoresis of the digestion products that a 50,000.dalton fragment is produced in large amounts from the intact 290,000-dalton polypeptide. The occurrence of polypeptides with more than one binding site is not rare (i.e. transferrin and ceruloplasmin).
More recently, however, the primary structures of several of these proteins have been shown to contain extensive regions of similar sequences (31,41), although the mechanisms which brought about these intragenic duplications are not known. Several examples include the human haptoglobin a2 chain (42)) the neurophysins (43)) the ferredoxins (44), the vertebrate immunoglobulin y-chain (45), the calcium-binding proteins (parvalbumins) from fish muscle (46,47), and rabbit skeletal muscle troponin C (48). The last example is particularly interesting.
Evidently, troponin C can be divided into four regions which are similar in sequence to each other as well as being homologous with calcium binding regions on the fish proteins. This conclusion was confirmed by the demonstration that troponin C does indeed bind four calcium ions (48). It is not known, of course, if the sequence homology in this large polypeptide is a consequence of events at the level of the genetic material or due to joining of 50,000-dalton polypeptides by a protein ligase (49).
Finally, it should be noted that the size of the polypeptide chain of mollusc hemocyanin is compatible with the model for 3805 this protein given by Mellema and Klug (50). While there is no evidence for the three kinds of proteins implied by their analysis, it is conceivable that the "collar" and "cap" at the ends of the 100 S particle are actually electron-dense regions which arise from a pinching in of the 290,000-dalton polypeptides which would be arranged along the long axis of the molecule. Moreover, the 120 structural subunits from which the wall of the particle is built are most simply interpreted to be the 50,000dalton domains which repeat along each polypeptide chain and which contain one oxygen binding site.