A lysine-rich protein from spermatozoa of the mollusc Mytilus edulis.

Abstract A small basic protein φ3II has been purified from spermatozoa of Mytilis edulis. It was extracted from chromatin by ethanol-HCl and further purified by chromatography on Biogel P-10 and Sephadex G-100. The protein has a molecular weight of about 5000. Half its residues are lysine and its is also rich in alanine, proline, and serine. It shows some compositional similarities to protamines if lysines are considered homologous with arginines. The composition of φ3II is also somewhat similar to that of the carboxyl region of lysine-rich histone.

It was extracted from chromatin by ethanol-HCl and further purified by chromatography on Biogel P-10 and Sephadex G-100. The protein has a molecular weight of about 5000. Half its residues are lysine and its is also rich in alanine, proline, and serine.
It shows some compositional similarities to protamines if lysines are considered homologous with arginines. The composition of 4311 is also somewhat similar to that of the carboxyl region of lysine-rich histone.
A search is being made to determine how restricted variations are in amino acid sequence in lysine-rich histones compared t.o the severe restrictions in argininerich histonc (1). Structural variations have been found among the lysine-rich histoncs of a single tissue (2,3) and between the lysine-rich histones of different species (4) and raise the possibility that a broad spectrum of structural variations will be found in lysinc-rich histones, perhaps cvcn eliminating the clear distinction now apparent between the lysine-rich class (Fl) and the moderately lysine-rich class (F2b).
Indeed the F2c histone from avian erythrocytes (5,6), histone T (7), and a new family of lysine-rich histones in IIolothuria tubulosa (3) could all be looked upon as intermediates in a series of structures with characteristics ranging from those of lysine-rich histones, such as calf thgmus Fl, to those of moderately lysine-rich histones, such as F2b, from the same tissue. In somewhat the same vein a spectrum of structural variation might be looked for bridging protamines and histones, and Stellwagcn and Cole (8) pointed out an intriguing compositional analogy between protamine and the carboxyl half of lysine-rich histone.
More generally, Bloch (9) included in his classification Further comparison to protamine is encouraged by the observation that this very lysinc-rich component occurs only in ripe gonads. Since its properties are intermediate between those of histones and the previously studied protamines from fish (13, 14) and bull (15), we felt led to purify and characterize it further as described in the present paper.

MATERIALS AND METHODS
Electrophoresis was conducted on I5 y0 polyacrylamide gel at pH 4.5 according to Panyim and Chalkley (16).
Amino acid analyses were performed according to Moore and Stein (17) on a Beckmarl-Spinco amino acid analyzer after hydrolysis in 6 N HCl for 22 hours at 110".
NHz-terminal residues were determined by the S-dimethylaminonapthalene-l-sulfonyl (dansyl) method according to Gray (18) and the hydrolysates were chromatographed on polyamide layers according to Woods and Wang (19). The phenylisothiocyanate method was also used as described by Konigsberg and Hill (20).
COOH-terminal residues were determined with diisopropylphosphorofluoridate-treated carboxypeptidase A and B obtained from Worthington.
The digests were prepared at 37" in 0.2 M NaCl in 0.05 M sodium barbital, pH 8.0. The enzyme to substrate ratio was 1: 100 by weight.
The samples were applied directly to a Beckman-Spinco analyzer in citrate buffer, pH 2.2.
Peptide maps were prepared by tryptic hydrolysis of approximately 2 mg of histone in 250 ~1 of 0.02 M ammonium bicarbonate using 1 y0 trypsin by weight.
The mixture was frozen after 16 hours and freeze-dried.
The freeze-drying was usually repeatea after redissolving in 250 ~1 of distilled water.
The peptides were dissolved in 25 ~1 of 5y0 formic acid and the entire sample was applied to Whatman No. 3MM filter paper sheets for electrophoresis at pH 3.5 in pyridine-acetic acid-water (1: lo:69 by volume) for 50 min at 5000 volts. Chromatography was used 1100 to separate the peptides in the second dimension using butanolacetic acid-pyridine-water according to Waley and Watson (21). The solvent was run to the bottom of the paper.
In general peptides were located by dipping in ninhydrincadmium reagent (22). Arginine peptides were detected on identical but separate peptide maps by dipping the paper in a mixture of equal volumes of 0.027, phenanthrenequinone in anhydrous ethanol and 10% sodium hydroxide in 60% ethanol and allowing the paper to dry at room temperature in a hood (23).

AND DISCUSSION
A general description of the basic proteins in the spermatozoa of mussels has been presented elsewhere (12). Of the total basic protein 72 '% is a protamine-like protein containing 30 y0 arginine and 24.3% lysine. Ripe spermatozoa also contain somatic histones comprising 15% of the basic proteins.
The remaining 13% of the basic proteins was prepared by Method 2 of Johns (24) which yields histone F3 from most tissues; for M. e&&s the fraction was called 43. This fraction was shown by gel electrophoresis to contain some somatic histones and a component with a very high mobility resembling that of protamines.
When 43 was submitted to chromatography on Amberlite IRC-50 under the conditions used for histones (25), the great majority of the protein was irreversibly adsorbed. This observation was not surprising since salmine requires 40% guanidinium chloride for elution (26). Columns of Bio-Gel P-10 (27) were then used for chromatography of 43 with the results presented in Fig. 1 and Table I. Three peaks were evident: amino acid analysis demonstrated that little peptide material was present in the third peak and that the second peak represented the lysine-rich component.
The first peak probably represents the somatic histone F3, as shown by the presence of cystine, the high amounts of glutamic acid and alanine, and the low content of glycine.
The material represented by the second peak was termed 4311 and was next chromatographed on Sephadex G-100 yielding a single slightly skewed peak (Fig. 2). Amino acid composition was constant (Table I) across the peak, suggesting that the material represented by the peak was essentially homogeneous.
The material showed a single band when tested by polyacrylamide gel electrophoresis. The truly remarkable feature of the composition of 4311 is that 50y0 of the amino acid residues are lysine, compared to 25 to 30% in the case of lysine-rich histones and compared to about 67% arginine in the case of the protamines (e.g., clupeine). High alanine, serine, and proline contents, along with low levels of hydrophobic amino acids, are reminiscent of both protamine and lysine-rich histone.
The composition is also intriguingly similar to the carboxyl half of a lysine-rich histone (28). This might be taken to suggest that $311 is an artifact derived from proteolysis of lysine-rich histone during isolation, especially since proteases are known to be present in chromatin (29). Proteolysis is not a likely explanation, however, since the numerous fast-running electrophoretic bands which apparently (30) result from this sort of degradation were not detected in the histone preparations from which 43 was isolated. If proteolytic degradation did occur, it must have been far more specific than observed previously, in which case it might well be an important step in the maturation of the sperm cells in the mussel. It was clearly important to estimate the molecular weight of 4311, and this was first studied using the method of sodium   dodecyl sulfate gel electrophoresis but preliminary studies showed the method unreliable for salmine and iridine which are very small and highly charged.
This was so even though the calibration standards used were histones.
Therefore the molecular weight was estimated by quantitative analysis of the NHzterminal residue using the subtractive Edman procedure. Such an analysis depends on prior determination of the amino acid composition and so the results shown in Table II were used to calculate a molecular weight of 5015, assuming the integral figures shown for each amino acid. On the basis of this molecular weight, the subtractive Edman procedure indicated 0.88 residue of NH&erminal alanine per mole of protamine. This is a reasonable yield since a subtractive Edman degradation carried out in parallel on insulin indicated 0.75 mole of glycine and 0.90 mole of phenylalanine per mole of insulin. That alanine was the only NHz-terminal residue in the protamine was confirmed by 5-dimethylaminonaphthalene-l-sulfonyl analysis; this analysis was used because the subtractive Edman procedure would not be reliable for NHz-terminal lysine or arginine.
The molecular weight was confirmed by COOH-terminal analysis using carboxypeptidase A + B digestions. As shown in Fig. 3 The amount of alanine released (about 1 mole for a molecular weight of 5000) confirms the assumed molecular weight because the next integral value would be most difficult to reconcile with the NHz-terminal result.
The molecular weight was further confirmed by the tryptic peptide map shown in Fig. 4. The distribution of tryptic peptides by paper electrophoresis and paper chromatography revealed about 10 spots (compare to clupeines (31)  Carboxypeptidase B (enzyme to substrate, 1:lOO) was added at 0 min; carboxypeptidase A (enzyme to substrate, 1:lOO) was added at 50 min (see arrow).
A molecular weight of 5015 was assumed for +311 in calculating moles of amino acid released per mole of protein.
FIG. 4. Peptide map of tryptic digest of 4311. The digest (2 mg) was spotted at the origin and electrophoresis was carried out at pH 3.5 from left (+) to right (-); chromatography was run vertically for the length of the paper. The paper was dipped in cadmium ninhydrin solution.
The arrows point to spots containing arginine as revealed by dipping in phenanthrenequinone.
be expected to yield about 7 tryptic peptides from their 30 to 33 residues). Two of these spots were found to contain arginine (apparently in equal amounts) when analyzed with phenanthrenequinone (23). Since the number of arginine residues was calculated at 2 per molecule on the basis of the amino acid composition and an assumed molecular weight of about 5000, the latter assumption is confirmed.
Further consideration of the peptide map leads to the speculation that (6311 may have some homology with a portion of lysinerich histone.
A comparison of Fig. 4 with similar maps of the