Site-directed mutagenesis of acyl carrier protein (ACP) reveals amino acid residues involved in ACP structure and acyl-ACP synthetase activity.

Acyl carrier protein (ACP) interacts with many different enzymes during the synthesis of fatty acids, phospholipids, and other specialized products in bacteria. To examine the structural and functional roles of amino acids previously implicated in interactions between the ACP polypeptide and fatty acids attached to the phosphopantetheine prosthetic group, recombinant Vibrio harveyi ACP and mutant derivatives of conserved residues Phe-50, Ile-54, Ala-59, and Tyr-71 were prepared from glutathione S-transferase fusion proteins. Circular dichroism revealed that, unlike Escherichia coli ACP, V. harveyi-derived ACPs are unfolded at neutral pH in the absence of divalent cations; all except F50A and I54A recovered native conformation upon addition of MgCl(2). Mutant I54A was not processed to the holo form by ACP synthase. Some mutations significantly decreased catalytic efficiency of ACP fatty acylation by V. harveyi acyl-ACP synthetase relative to recombinant ACP, e.g. F50A (4%), I54L (20%), and I54V (31%), whereas others (V12G, Y71A, and A59G) had less effect. By contrast, all myristoylated ACPs examined were effective substrates for the luminescence-specific V. harveyi myristoyl-ACP thioesterase. Conformationally sensitive gel electrophoresis at pH 9 indicated that fatty acid attachment stabilizes mutant ACPs in a chain length-dependent manner, although stabilization was decreased for mutants F50A and A59G. Our results indicate that (i) residues Ile-54 and Phe-50 are important in maintaining native ACP conformation, (ii) residue Ala-59 may be directly involved in stabilization of ACP structure by acyl chain binding, and (iii) acyl-ACP synthetase requires native ACP conformation and involves interaction with fatty acid binding pocket residues, whereas myristoyl-ACP thioesterase is insensitive to acyl donor structure.     

Cytidine deaminase from Escherichia coli B. Purification and enzymatic and molecular properties

Cytidine deaminase (cytidine aminohydrolase, EC 3.5.4.5) from Escherichia coli has been purified to homogeneity through a rapid and efficient two-step procedure consisting of anion-exchange chromatography followed by preparative electrophoresis. The final preparation is homogeneous, as judged by a single band obtained by disc gel electrophoresis performed in the absence and presence of denaturing agents. The native protein molecular weight determined by gel filtration is 56 000. Sodium dodecyl sulfate disc gel electrophoresis experiments conducted upon previous incubation of the enzyme with dimethyl suberimidate suggest an oligomeric structure of two identical subunits of 33 000 molecular weight. The absorption spectrum of the protein reveals a maximum at 277 nm and a minimum at 255 nm. The isoelectric point is at pH 4.35. Amino acid analysis indicates an excess of acidic amino acid residues as well as six half-cystine residues. No interchain disulfide groups have been evidenced. According to Cleland's nomenclature, kinetic analysis shows a rapid-equilibrium random Uni-Bi mechanism. Cytidine deaminase is competitively inhibited by various nucleosides. Km values for cytidine, deoxycytidine, and 5-methylcytidine are 1.8 X 10(-4), 0.9 X 10(-4), and 12.5 X 10(-4) M, respectively.     

Purification and characterization of the chaperonin 10 and chaperonin 60 proteins from Rhodobacter sphaeroides.

Two heat-shock proteins that show high identity with the Escherichia coli chaperonin 60 (groEL) and chaperonin 10 (groES) chaperonin proteins were purified and characterized from photolithoautotrophically grown Rhodobacter sphaeroides. The proteins were purified by using sucrose density gradient centrifugation and Mono-Q anion-exchange chromatography. In the presence of 1 mM ATP, the chaperonin 10 and chaperonin 60 proteins bound to each other and comigrated as a large complex during sucrose density gradient centrifugation. The native molecular weights of each protein as determined by gel filtration chromatography were 889,200 for chaperonin 60 and 60,000 for chaperonin 10. Chaperonin 60 is comprised of monomers with a molecular weight of 61,000 and chaperonin 10 is comprised of monomers with a molecular weight of 12,700 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Chaperonin 60 was 9.3% of the total soluble cell protein during photolithoautotrophic growth which increased to 28.5% following heat-shock treatment. When cells were grown photoheterotrophically or chemoheterotrophically, chaperonin 60 was reduced to 6.7% and 3.5%, respectively, of the total soluble protein. The N-terminal amino acid sequence of each protein was determined; chaperonin 60 of R. sphaeroides showed 72% identity to E. coli chaperonin 60 protein, and R. sphaeroides chaperonin 10 showed 45% identity with E. coli chaperonin 10. R. sphaeroides chaperonin 60 catalyzed ATP hydrolysis with a specific activity of 134 nmol min-1 mg-1 (kcat = 0.13 s-1) and was inhibited by R. sphaeroides chaperonin 10, but not E. coli chaperonin 10. The E. coli chaperonin 60 ATPase activity was inhibited by chaperonin 10 from both R. sphaeroides and E. coli.     

Repair of alkylated DNA in Escherichia coli. Physical properties of O6-methylguanine-DNA methyltransferase

An inducible methyltransferase of Escherichia coli acts on O6-methylguanine in DNA by conveying the methyl group to one of its own cysteine residues. The protein has now been purified to apparent homogeneity from a constitutively expressing strain. The homogeneous methyltransferase exhibits no DNA glycosylase or endonuclease activity on alkylated DNA. Further, the methyltransferase activity is strikingly resistant to heat inactivation under reducing conditions. The protein has Mr = 18,000 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, while the sedimentation coefficient and Stokes radius of the native enzyme yield Mr = 18,400. The amino acid composition of the purified protein shows 4 to 5 cysteine residues/transferase molecule. The methylated, inactive form of the transferase has an unaltered molecular weight.     

Purification and characterization of a low molecular weight zinc binding protein from human placenta

A low molecular weight, native zinc binding, cytosolic protein (LMZP) has been isolated, purified and characterized from human normal term placenta. Gel filtration of heat treated placental cytosol after sequential acetone precipitation (80% ppt) revealed a major zinc binding protein in the range of low molecular weight. This partially purified zinc binding fraction was further fractionated on DEAE-Sephadex A-25. The zinc was eluted in one of the three peak fractions. Further, the purity of zinc binding protein was confirmed on fast protein liquid chromatography (FPLC). The purified placental LMZP was homogenous on SDS-polyacrylamide gel electrophoresis with a single band. Ultraviolet (UV) spectrum of LMZP showed an absorption maximum at 257 nm which disappeared at pH 2. Molecular weight of LMZP as determined by gel chromatography, SDS-polyacrylamide gel electrophoresis and amino acid analysis was 6 kDa. It was calculated that 1 g atom of zinc was bound to 1 mole of the LMZP. Unlike in classical metallothionein, the amino acid composition of placental LMZP revealed the presence of aromatic amino acids, lower content of cysteine and higher content of histidine, glutamic acid and aspartic acid (10, 9 and 5 residues/mole, respectively).     

R plasmid dihydrofolate reductase with a dimeric subunit structure

Dihydrofolate reductase specified by plasmid R483 from a trimethoprim-resistant strain of Escherichia coli has been purified 2,000-fold to homogeneity using dye-ligand chromatography, gel filtration, and polyacrylamide gel electrophoresis. The protein migrated as a single band on nondenaturing polyacrylamide gel electrophoresis and had a specific activity of 250 mumol/mg min(-1). The molecular weight was estimated to be 32,000 by gel filtration and 39,000 by Ferguson analysis of polyacrylamide gel electrophoresis. When subjected to electrophoresis in the presence of sodium dodecyl sulfate, the protein migrated as a single 19,000-molecular weight species, a fact that suggests that the native enzyme is a dimer of similar or identical subunits. Antibody specific for R483-encoded dihydrofolate reductase did not cross-react with dihydrofolate reductase encoded by plasmid R67, T4 phage, E. coli RT500, or mouse L1210 leukemia cells. The amino acid sequence of the first 34 NH2-terminal residues suggests that the R483 plasmid dihydrofolate reductase is more closely related to the chromosomal dihydrofolate reductase than is the enzyme coded by plasmid R67.     

Purification and characterization of an agglutinin from mucus of the snail Achatina fulica

The mucus of the snail Achatina fulica shows the presence of an agglutinin that nonspecifically agglutinates human erythrocytes. The agglutinin has been purified by affinity chromatography using Sepharose 4B-hog gastric mucin as the affinity matrix. Homogeneity was checked by polyacrylamide gel electrophoresis, immunodiffusion, immunoelectrophoresis, and gel filtration. The agglutinin is a glycoprotein of native molecular weight 70,000. The isoelectric point of the protein was found to be 8.0. The predominant amino acids are aspartic acid and glutamic acid (or amides) and serine, which account for 32% of the total amino acid residues. The agglutinin has 10% carbohydrate (wt/wt) and the most abundant sugar is N-acetylglucosamine. The cd spectra of the agglutinin show the presence of random coil conformation. The inhibition of hemagglutination data indicates that the agglutinin is specific for beta glycosides of D-Gal and D-GalNAc.     

Isolation, purification, and partial characterization of type V-A hemagglutinin from Escherichia coli GV-12, O1:H-

A hemagglutinin which specifically agglutinates human type A erythrocytes (mannose resistant) was isolated from the growth medium of cultures of Escherichia coli GV-12, serotype O1:H-, and purified by chromatography on Bio-Gel A-1.5 and DEAE-Sephadex A-25. The purity of the hemagglutinin was established by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoelectrophoresis. N-terminus analysis indicated that only asparagine resides on the amino terminus. The native hemagglutinin is an aggregate exhibiting a sedimentation coefficient of 9.25, which corresponds to a molecular weight of approximately 200,000. The monomeric molecular weight was found to be approximately 16,300. Amino acid analysis indicated that the hemagglutinin consists of 131 residues, corresponding to a molecular weight of 13,400.     

A clottable protein (coagulogen) from amoebocyte lysate of Japanese horseshoe crab (Tachypleus tridentatus). Its isolation and biochemical properties.

A clottable protein, named coagulogen, was highly purified from the amoebocyte lysate of Japanese horseshoe crab (Tachypleus tridentatus) by a method similar to that used for the lysate of Limulus polyphemus amoebocytes. The isolated material gave a single protein band on analytical gel electrophoresis at pH 3.2, gel electrofocusing, and sodium dodecyl sulfate (SDS) gel electrophoresis with or without 2-mercaptoethanol. It was 90 percent coagulable, and the total yield from 10 ml of the amoebocyte lysate was about 40 mg. The sedimentation coefficient of purified coagulogen was 2.6 S and its molecular weight was estimated to be about 15,300 by sedimentation equilibrium analysis. The molecular weight estimated by SDS-gel electrophoretic analysis was 19,500 +/- 1,000. This discrepancy was apparently due to abnormal mobility arising from the basic nature of this protein on electrophoresis. The protein had a high isoelectric point of pH 10.0 +/- 0.2, as measured by the isoelectric focusing technique. It consisted of a total of 132 to 135 amino acid residues and contained high levels of basic amino acids, which accounted for more than 16 per cent of the total amino acid residues. No methionine was detected. High contents of valine, half-cystine, glutamic acid (glutamine), and phenylalanine were found. The N-terminal sequence of the first three residues of the coagulogen was Ala-Asx-Thr, and its C-terminal residues was identified as phenylalanine, indicating that it consists of a single polypeptide chain. It is of interest that the first three N-terminal residues are homologous with those of the Aalpha-chain of non-human primate fibrinogen.     

Purification and Characterization of Colicin D

Colicin D-CA23, obtained by sonic treatment of mitomycin C-induced cells of Escherichia coli K-12 W1485 (colD), was purified by ammonium sulfate precipitation, gel filtration on Sephadex G200, ion-exchange chromatography on diethylaminoethyl cellulose, and isoelectrofocusing. Polyacrylamide-gel electrophoresis, sedimentation velocity analysis, and antigenic analysis indicated that the preparation was homogeneous. Colicin D is composed entirely of amino acids and hence is a simple protein uncomplexed with lipid or lipopolysaccharide. It contains six residues of cysteine per molecule. The molecular weight of colicin D is approximately 92,000, as determined by sodium dodecyl sulfate-polyacrylamide-gel electrophoresis and gel filtration on Sephadex G200. Its sedimentation coefficient is 4.41S. The behavior of colicin D in solutions of sodium dodecyl sulfate and 2-mercaptoethanol indicates that it does not consist of subunits and exists as a single polypeptide chain. Its high molecular weight and presence of six cysteine residues per molecule distinguish colicin D from all colicins previously described. Although colicins D and E3 have similar modes of action, their gross molecular properties are entirely different.     

Purification and characterization of 210,000-dalton inhibitor of calcium-activated neutral protease from rabbit skeletal muscle and its relation to 50,000-dalton inhibitor

An endogenous inhibitor of calcium-activated neutral protease (CANP), which was isolated from rabbit skeletal muscle with chemically drastic pretreatments, comprised major (high-molecular-weight form, HMW-inhibitor) and minor (low-molecular-weight form, LMW-inhibitor) components. HMW-inhibitor was purified to homogeneity using FPLC and preparative electrophoresis. The purified inhibitor appeared as a single protein with a molecular weight of 110,000 on SDS-polyacrylamide gel electrophoresis, and a molecular weight of 210,000 on gel filtration. It was therefore presumed that the inhibitor is a dimer protein under native conditions. It contained large amounts of glutamic acid, alanine, and proline, and small amounts of aromatic amino acids, showing an amino acid composition similar to that of LMW-inhibitor. HMW-inhibitor inhibited CANPs with both low (m-type) and high (mu-type) Ca2+-sensitivity but had no effect on any other proteases examined. It was demonstrated that the inhibition was due to the formation of a stoichiometric complex between rabbit mCANP and inhibitor subunit in the ratio of five to one. These results suggest that HMW-inhibitor might have several reactive sites per molecule and that LMW-inhibitor subunit might be a proteolytic fragment of HMW-inhibitor containing an active site.     

Purification and properties of a binding protein for branched-chain amino acids in Pseudomonas aeruginosa.

A binding protein for branched-chain amino acids was purified to a homogeneous state from shock fluid of Pseudomonas aeruginosa PML14. It was a monomeric protein with an apparent molecular weight of 4.3 x 10(4) or 4.0 x 10(4) by sodium dodecyl sulfate-polyacrylamide gel electrophoresis or gel filtration, respectively. The isoelectric point was determined to be pH 4.1 by electrofocusing. Amino acid analysis of the protein showed that aspartic acid, glutamic acid, glycine, and alanine were major components and that the protein contained only one residue each of tryptophan and cysteine per molecule. The binding protein contained no sugar. The binding activity of the protein was specific for the branched-chain amino acids. The protein also bound alanine and threonine with lower affinity. The dissociation constants of this protein for leucine, isoleucine, and valine were found to be 0.4, 0.3, and 0.5 microM, respectively. Mutants defective in the production of the binding protein were identified among the mutants deficient in a transport system for branched-chain amino acids (LIV-I). The revertants from these mutants to LIV-I-positive phenotype simultaneously recovered normal levels of the binding protein. These findings suggest strongly the association of the binding protein with the LIV-I transport system.     

Glutamate-41 of Vibrio harveyi acyl carrier protein is essential for fatty acid synthase but not acyl-ACP synthetase activity

Bacterial acyl carrier protein (ACP) is a small, acidic, and highly conserved protein that supplies acyl groups for biosynthesis of a variety of lipid products. Recent modelling studies predict that residues primarily in helix II of Escherichia coli ACP (Glu-41, Ala-45) are involved in its interaction with the condensing enzyme FabH of fatty acid synthase. Using recombinant Vibrio harveyi ACP as a template for site-directed mutagenesis, we have shown that an acidic residue at position 41 is essential for V. harveyi fatty acid synthase (but not acyl-ACP synthetase) activity. In contrast, various replacements of Ala-45 were tolerated by both enzymes. None of the mutations introduced dramatic structural changes based on circular dichroism and native gel electrophoresis. These results confirm that Glu-41 of ACP is a critical residue for fatty acid synthase, but not for all enzymes that utilize ACP as a substrate.     

Purification and structural characterization of herpes simplex virus glycoprotein C

Purification of herpes simplex virus glycoprotein C (gC) in microgram amounts yielded sufficient material for an analysis of its secondary structure. Purification was facilitated by using the mutant virus gC-3, which bears a point mutation that interrupts the putative hydrophobic membrane anchor sequence, causing the secretion of gC-3 protein into the cell culture medium. gC-3 protein was purified by size fractionation of concentrated culture medium from infected cells on a gel filtration column of Sephacryl S-200, followed by immunoaffinity chromatography on a column constructed of gC-specific monoclonal antibodies cross-linked to a protein A-Sepharose CL-4B matrix. Purified gC-3 had a molecular weight of 130,000 as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, the size expected for gC, was reactive with gC-specific monoclonal antibodies in protein immunoblots, and contained amino acid sequences characteristic of gC as determined by radiochemical amino acid microsequence analyses. Polyclonal antisera obtained from a rabbit immunized with gC-3 reacted with wild-type gC in immunoprecipitation, enzyme immunoassay, and immunoelectroblot (western blot) assays. Deglycosylation by treatment with trifluoromethanesulfonic acid reduced the molecular weight of gC-3 by approximately 35%. Analyses of both native and deglycosylated gC-3 by Raman spectroscopy showed that the native molecule consists of about 17% alpha-helix, 24% beta-sheet, and 60% disordered secondary structures, whereas deglycosylated gC-3 consists of about 8% alpha-helix, 10% beta-sheet, and 81% disordered structures. These data were in good agreement with the 11% alpha-helix, 18% beta-sheet, 61% beta-turn, and 9% disordered structures calculated from Chou-Fasman analysis of the primary sequence of gC-3.     

Altered molecular form of acyl carrier protein associated with beta-ketoacyl-acyl carrier protein synthase II (fabF) mutants.

Acyl carrier protein (ACP) is a required cofactor for fatty acid synthesis in Escherichia coli. Mutants lacking beta-ketoacyl-ACP synthase II activity (fabF1 or fabF3) possessed a different molecular species of ACP (F-ACP) that was separated from the normal form of the protein by conformationally sensitive gel electrophoresis. Synthase I mutants contained the normal protein. Complementation of fabF1 mutants with an F' factor harboring the wild-type synthase II allele resulted in the appearance of normal ACP, whereas complementation with an F' possessing the fabF2 allele (a mutation that produces a synthase II enzyme with altered catalytic activity) resulted in the production of both forms of ACP. The structural difference between F-ACP and ACP persisted after the removal of the 4'-phosphopantetheine prosthetic group, and both forms of the protein had identical properties in an in vitro fatty acid synthase assay. Both ACP and F-ACP were purified to homogeneity, and their primary amino acid sequences were determined. The two ACP species were identical but differed from the sequence reported for E. coli E-15 ACP in that an Asn instead of an Asp was at position 24 and an Ile instead of a Val was at position 43. Therefore, F-ACP appears to be a modification of ACP that is detected when beta-ketoacyl-ACP synthase II activity is impaired.     

Partial characterization of oat and rye phytochrome

Purified oat and rye phytochrome were examined by analytical gel chromatography, polyacrylamide gel electrophoresis, N-terminal, and amino acid analysis. Purified oat phytochrome had a partition coefficient on Sephadex G-200 (sigma(200)) of 0.350 with an estimated molecular weight of 62,000; sodium dodecyl sulfate polyacrylamide electrophoresis gave an equivalent weight estimate. Purified rye phytochrome had a sigma(200) value of 0.085 with an estimated molecular weight of 375,000; sodium dodecyl sulfate electrophoresis gave a weight estimate of 120,000, indicating a multimer structure for the nondenatured protein. Comparative sodium dodecyl sulfate electrophoresis with purified phycocyanin and allophycocyanin gave a molecular weight estimate of 15,000 for allophycocyanin, and two constituent classes of subunits for phycocyanin with molecular weights of 17,000 and 15,000. Amino acid analysis of oat phytochrome confirmed a previous report; amino acid analysis of rye phytochrome differs markedly from a previous report. Oat phytochome has four detectable N-terminal residues (glutamic acid, serine, lysine, and leucine, or isoleucine); rye phytochrome has two detectable groups (aspartic and glutamic acids). Model experiments subjecting purified rye phytochrome to proteinolysis generate a product with the characteristic spectral and weight properties of oat phytochrome, as it has been described in the literature. It is concluded that the structural characteristics of purified rye phytochrome are likely those of the native protein.     

Sequence of the D-aspartyl/L-isoaspartyl protein methyltransferase from human erythrocytes. Common sequence motifs for protein, DNA, RNA, and small molecule S-adenosylmethionine-dependent methyltransferases

A widely distributed protein methyltransferase catalyzes the transfer of a methyl group from S-adenosyl-methionine to the free carboxyl groups of D-aspartyl and/or L-isoaspartyl derivatives of L-aspartyl and L-asparaginyl residues. This enzyme has been postulated to function in the repair or the catabolism of age-damaged proteins. We present here the complete amino acid sequence of the more basic isozyme I of this enzyme from human erythrocytes. The sequence was determined by Edman degradation and mass spectral analysis of overlapping trypsin, Staphylococcus aureus V8 protease, Pseudomonas fragi endoproteinase Asp-N, cyanogen bromide, and hydroxylamine-generated fragments. The NH2-terminus is modified by acetylation and the protein contains 226 amino acids for a calculated molecular weight of 24,575. This value is in good agreement with the molecular weight determined for the purified protein by polyacrylamide gel electrophoresis in the presence of dodecyl sulfate and by gel filtration chromatography under nondenaturing conditions. The identification of 2 different amino acid residues at both positions 22 and 119 may indicate the presence of allelic variants or of two or more closely related structural genes. Finally, comparison of this sequence with those of methyltransferases for RNA, DNA, and small molecules, as well as other S-adenosylmethionine-utilizing enzymes, shows that many of these proteins share elements of three regions of sequence similarity and may be structurally or evolutionarily related.     

Purification and biochemical characterization of a soluble human interferon gamma receptor expressed in Escherichia coli

We purified and characterized a soluble human interferon gamma receptor expressed in Escherichia coli. The soluble receptor comprises the amino acids 15-246 of the encoded protein (Aguet, M., Dembic, Z., and Merlin, G. (1988) Cell 55, 273-280) and was purified from large scale fermentations through four chromatographic steps with an overall recovery of 28%. The refolded soluble receptor shows some heterogeneity on nonreducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis, where it appears as the major band of 27 kDa molecular mass, accompanied by a few minor bands with molecular masses between 26 and 30 kDa. On reducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis it appears as a homogeneous protein of 32 kDa molecular mass. The soluble interferon gamma receptor is an active and stable protein and is recognized by specific antibodies raised against the native receptor. When nonreduced it has the capacity to specifically bind interferon gamma and to compete for the binding of interferon gamma to the cell surface receptor. The observed heterogeneity of the soluble interferon gamma receptor under nonreducing electrophoretic conditions is probably due to different conformational forms resulting from the formation of non-native intramolecular disulfide bonds among the 8 cysteine residues present in the soluble interferon gamma receptor molecule.     

Reconstitution of rapeseed high molecular weight protein from isolated subunits

The high molecular weight protein was isolated from rapeseed and characterised. Six subunits were isolated in SDS (0.01%) solution on polyacrylamide-gel electrophoresis and by gel filtration on Sephadex G-100. Reassociation by removing SDS by co-dialysis, against 10 mM sodium phosphate buffer (pH 7.9) was done and the yield was about 90%. The reconstituted protein was indistinguishable from the intact protein in all respects. The subunits isolated from the native protein and the reconstituted protein were found to have identical molecular weights and N-terminal amino acids. No disulphide bonds were observed in the subunit association. Amino acid analysis of the proteins and the six subunits was performed and the number of each amino acid residue calculated.     

Histidinoalanine, a naturally occurring cross-link derived from phosphoserine and histidine residues in mineral-binding phosphoproteins

Native mineral-containing phosphoprotein particles were isolated from the Heterodont bivalve Macrocallista nimbosa. The native particles are discrete structures about 40 nm in diameter which migrate as a single band during electrophoresis in agarose gels. Removal of the mineral component with ethylenediaminetetraacetic acid dissociates the native protein into nonidentical subunits. The lower molecular weight subunits, representing 8% of the total protein, were obtained by differential centrifugation. The native protein is characterized by a high content of aspartic acid, phosphoserine, phosphothreonine, histidine, and the bifunctional cross-linking residue histidinoalanine. The low molecular weight subunits have the same amino acid composition except for a reduction in histidinoalanine and a corresponding increase in phosphoserine and histidine residues, demonstrating that the alanine portion of the cross-link is derived from phosphoserine residues. Ion-exchange chromatography and molecular sieve chromatography show that the low molecular weight subunits have a similar charge density but differ in molecular weight, and the relative mobilities of the subunits on agarose gels indicate that they are polymers of a single phosphoprotein molecule. The minimum molecular weight of the monomer is about 140 000 on the basis of the amino acid composition. The high molecular weight subunits are rich in histidinoalanine and too large to be resolved by either molecular sieve chromatography or gel electrophoresis. On the basis of the ultrastructural, electrophoretic, chromatographic, and compositional evidence, native phosphoprotein particles are composed of subunits ionically cross-linked via divalent cations. These subunits are variable molecular weight aggregates of a single phosphoprotein molecule covalently cross-linked via histidinoalanine residues. Evidence for a nonenzymatic cross-linking mechanism is discussed.