In vitro correction of erythrocyte glucose 6-phosphate dehydrogenase (G6PD) deficiency.

The permeability characteristics of egg phosphatidylcholine (EPC) vesicles to Eu³⁺ has been examined by ³¹P-nmr, in various mixed lipid systems. It has been found that incorporation of small amounts of sodium taurocholate (NaTC) in the EPC vesicle greatly increased the vesicular permeability to Eu³⁺. Incorporation of cholesterol in the EPC vesicle significantly inhibits the ability of NaTC to induce permeability alterations in this mixed system. It has been reported that at low EPC:NaTC ratios (2:1-0.65:1), mixed micelles of the components are formed [N. A. Mazer, R. F. Kwasnick, M. C. Carey, and G. B. Benedek, 1977, in Micellization, Solubization, and Microemulsions (Mittal, K. L., ed.) Vol. 1, pp. 483-402, Plenum Press, New York]. By examination of the ³¹P-nmr linewidths of EPC, at various ratios of EPC:NaTC, it is possible to follow the decrease in size of the EPC vesicle, as it becomes incorporated into the smaller NaTC micelles. The ³¹P{¹H} nuclear Overhauser enhancement of the simple EPC vesicle is not significantly different from this same parameter measured for EPC, when it exists in mixed micellar systems with NaTC. This indicates that there must be considerable internuclear head group interactions of EPC molecules in EPC-NaTC mixed micelles. This argues for sequestering of EPC in such micelles.     

Localization of the disulfide bond in human antithrombin III required for heparin-accelerated thrombin inactivation

Heparin accelerates the rate of inhibition of thrombin by antithrombin III. Reduction of one of the three antithrombin disulfide bonds with dithiothreitol under mild conditions abolishes this rate-enhancing effect without affecting the rate of reaction in the absence of heparin. Alkylation of mildly reduced antithrombin III with [³H]iodacetic acid followed by digestion with cyanogen bromide yielded two major labeled peptides. The smaller peptide, containing Cys-422, was identified as extending from Gly-414 to the C-terminus, Lys-424. Our data are consistent with the larger labeled peptide being the one extending from Glu-104 to Met-243 and containing Cys-239. Cys-422 has been shown by others to be linked to Cys-239. These data indicate that the sensitive disulfide bond in antithrombin III extends between Cys-239 and Cys-422; the site at which thrombin cleaves the antithrombin III is between these two half-cystines.     

Purification of two forms of kanamycin acetyltransferase from Escherichia coli

Kanamycin acetyltransferase acylates aminoglycoside antibiotics using acetyl-CoA, and thereby conveys bacterial resistance to several clinically important antibiotics, notably amikacin. The enzyme was quantitatively and reproducibly released from Escherichia coli W677 harboring plasmid pMH67 by a modified osmotic shock procedure (bacterial cells are incubated overnight in sucrose and again without sucrose before onset of osmotic shock). The enzyme was purified by dye-ligand chromatography on Affi-Gel Blue in addition to antibiotic affinity chromatography on neomycin-Sepharose-4B. The activity did not increase with subsequent chromatography on ion-exchange, hydrophobic, or molecular-exclusion gels. However, both dye-ligand and molecular-exclusion chromatography, as well as disc-gel electrophoresis, separated the purified enzyme equally into two active protein fractions. Based on the more active of the two forms, the purification was 112-fold with a specific activity of 1.9 IU/mg. The less-active form has an unusual absorbance spectrum, with a maximum near 255 nm, which cannot be explained by the amino acid composition. Chromatography of this form alone regenerated both forms, suggesting that the enzyme is noncovalently conjugated to an uncharged chromophore, such as a lipid. The purified enzyme has a very sharp pH optimum at 5.5 with a plateau on the alkaline side, but is most stable between pH 8.5 and 9.5. Data from electrophoresis in the presence of sodium dodecyl sulfate and gel-filtration on Ultrogel AcA 44 are consistent with a tetrameric protein of 60-70,000 Da.     

Characteristics of the induction of aldehyde dehydrogenase in rat liver

The activity of rat liver supernatant aldehyde dehydrogenase is increased by phenobarbital treatment in one selected strain (RR) but not in another strain (rr) of animals derived from randomally bred populations (Deitrich, Collins, and Erwin (1972) J. Biol. Chem., 247, 7232). Before 14 days of age, increased enzyme activity after phenobarbital treatment is minimal but between 30 and 60 days of age there is a maximal increase in activity after phenobarbital treatment. Using animals of this age, it was shown that both cycloheximide and actinomycin D block this response to phenobarbital. Phenobarbital treatment decreases heat stability of crude preparations of the enzyme from RR rats, but increases heat stability of the enzyme from rr animals.     

Regulation of C4 photosynthesis: characterization of a protein factor mediating the activation and inactivation of NADP-malate dehydrogenase.

The leaf NADP-malate dehydrogenase of Zea mays is rapidly activated when leaves are illuminated and inactivated in the dark. The present studies show that inactive enzyme isolated from darkened leaves was activated by dithiothreitol and that the active enzyme was rapidly inactivated by oxygen in dithiothreitol-free solutions. Following the fractionation of leaf extracts, both the activation and inactivation of NADP-malate dehydrogenase in vitro were partially or totally dependent upon a separate small molecular weight protein factor. Activation and inactivation were largely or solely dependent upon this factor at pH 8.0 or less, but apparently only partially factor dependent at pH 9.0. The factor was heat stable, inactivated by incubation with trypsin, and had a molecular weight of about 10,000. It was mostly associated with the chloroplasts of mesophyll cells.     

Mechanism of reversible glycine cleavage reaction in Arthrobacter globiformis. Function of lipoic acid in the cleavage and synthesis of blycine.

The physical and chemical characterization of horse serum butyrylcholinesterase has been extended. The results show that the enzyme is a glycoprotein containing about 20% carbohydrate by weight. Mannose, glucosamine, galactose, and sialic acid are the sugar residues found. The extinction coefficient of butyrylcholinesterase, E_1cm^1% at 280 nm, was found to be 15.2 ± 0.3 by dry weight determination. The molecular weight of the protein in dilute phosphate buffer was determined to be (31.7 ± 1.2) × 10⁴ by high speed equilibrium sedimentation with a redetermined partial specific volume of 0.723 ± 0.003 ml/g. Subunit molecular weights for the dissociated protein were found to be (7.9 ± 0.4) × 10⁴ and (8.1 ± 0.1) × 10⁴, respectively, in guanidine hydrochloride and in a solution at pH 11.8. The subunit molecular weight was also estimated to be (8.8 ± 0.2) × 10⁴ by analytical sodium dodecyl sulfate-gel electrophoresis. This apparently higher subunit molecular weight from dodecyl sulfate gels is expected for glycoproteins containing significant amounts of carbohydrate. No free sulfhydryl group was detected, even though there are six half-cystines in each subunit. Therefore, it seems likely that there are three pairs of disulfide bonds per subunit. The available data indicate that native butyrylcholinesterase is a tetrameric glycoprotein consisting of subunits of equal molecular weight.     

Competitive binding assays for riboflavin and riboflavin-binding protein

A competitive binding procedure that can be used to determine either riboflavin or riboflavin-binding protein has been developed. Riboflavin-binding protein from chicken egg white binds tightly to DEAE-cellulose while free riboflavin does not. Stock [2-¹⁴C]riboflavin solutions, diluted with varying amounts of a standard unlabeled riboflavin solution or an unknown sample, are mixed with aporiboflavin-binding protein and washed through small DEAE-cellulose columns. The protein-bound riboflavin is batch eluted into scintillation vials, counted, and the unknown samples compared to a standard curve. This is a simple, rapid method for assaying riboflavin by isotope dilution. By a slight modification of the incubation conditions of this procedure, the degree of saturation and amount of riboflavin-binding protein can be determined. Data from both assays can be represented by linear plots in which slopes or intercepts correspond to unknown values. The principles presented here have been extended to the assay of biotin and avidin and should apply to other vitamins and vitamin-binding proteins.     

Isolation and characterization of a tartrate-sensitive splenic acid phosphatase in Gaucher's disease

The acid phosphatase activity that is increased in the spleens of patients with Gaucher's disease can be separated into two principal isoenzymes by chromatography on sulphopropyl-Sephadex. The acid phosphatase species that is resistant to inhibition by l-(+)-tartrate is retained by the cation-exchange resin while the tartrate-sensitive species passes through. We have isolated and characterized the tartrate-sensitive acid phosphatase (designated SP_I) from the spleen of a patient with the adult (type 1) form of Gaucher's disease. SP_I acid phosphatase, representing approximately 30 to 50% of the total acid phosphatase activity in a detergent (Triton X-100) extract of spleen tissue, has been purified approximately 400-fold to a specific activity of 48 units/mg of protein (substrate, 4-methylumbelliferyl phosphate). The final preparation of acid phosphatase contains at least two protein components—each with phosphatase activity—when analyzed by polyacrylamide gel electrophoresis at pH 8.9 or isoelectric focusing. SP_I acid phosphatase exhibits a broad substrate specificity and catalyzes the hydrolysis of a variety of artificial and natural phosphate-containing compounds including p-nitrophenyl phosphate, α-naphthyl phosphate, phosphoenolpyruvate, and CMP. The enzyme is inhibited by l-(+)-tartrate, sodium fluoride, and ammonium molybdate and has the following properties: pH optimum, 4.5; K_m on 4-methylumbelliferyl phosphate, 44 μm; pI, 3.8–4.1; M_r, 177,400; s_20,w, 6.8.     

Substrate specificity of Gaucher spleen phosphoprotein phosphatase

The spleen in Gaucher's disease contains elevated levels of two distinct acid phosphatases. One of the isoenzymes, a tartrate-resistant type 5 acid phosphatase which we have designated SP_II acid phosphatase, possesses considerable phosphoprotein phosphatase activity. The enzyme dephosphorylates phosvitin and casein at specific rates (V) of 38.6 and 45.0 units/mg, respectively. The dephosphorylation of the oligophosphoproteins as well as various fragments of phosvitin, histories, and monophosphopeptides was studied kinetically. Positive cooperativity (Hill coefficient = 1.3–2.0) was observed for the dephosphorylation of phosvitin and casein as well as for the dephosphorylation of fragments of phosvitin which contained as few as two vicinal phosphoserine residues. In contrast, the hydrolysis of phosphomonoesters such as o-phosphorylserine or various monophosphopeptides exhibited typical Michaelis-Menten kinetics. Cooperativity appears to depend upon the substrate rather than the enzyme. The cooperativity of dephosphorylation was not affected by altering the secondary structure of phosvitin from a random to β conformation or by acetylation of the protein; however, acetylated phosvitin was dephosphorylated more rapidly (V = 50.8 units/mg) than native phosvitin indicating that the very basic phosphatase enzyme (pI = 8.5) prefers more acidic phosphoproteins as substrates rather than basic proteins such as histone (V= 0.0013 unit/mg). A monophosphohexa-peptide (V = 0.47 unit/mg) and monophosphoheptapeptide (V = 0.18 unit/mg) proved to be much poorer substrates than phosvitin, and monophosphoproteins such as glycogen phosphorylase, phosphorylase kinase, and glycogen synthase were not dephosphorylated by the enzyme. Although the phosphatase is active on monophosphopeptides and the presence of flanking amino acids considerably decreases the K_m of the enzyme for the phosphoserine residue (up to 100-fold), the enzyme appears to prefer peptide or protein substrates that contain two or more phosphoserine residues in close proximity. Finally, previous results showing the spleen phosphatase to be composed of 16,000- and 20,000-dalton subunits were apparently due to proteolysis during isolation since when 1.0 mm phenylmethylsulfonyl fluoride was included in the isolation media, the enzyme appeared as a single 35,000-dalton species when subjected to polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate.     

Subcellular localization of the heat-stable glucocerebrosidase activator substance in Gaucher spleen

The spleen in Gaucher's disease contains relatively large quantities of a heat-stable activator of the glucocerebrosidase of normal human tissues (Ho, M. W., and O'Brien, J. S. (1971) Proc. Nat. Acad. Sci. USA68, 2810–2813) that has been shown to be an 11,000 molecular weight acidic glycoprotein (Peters, S. P., et al. (1977) J. Biol. Chem.252, 563–573). In an effort to determine the subcellular location of the activator, a mannitol-sucrose homogenate of fresh, unfrozen spleen obtained from a 26-year-old patient with adult, nonneuropathic (Type 1) form of Gaucher's disease was subjected to subcellular fractionation. The tissue used in these experiments exhibited a β-glucocerebrosidase deficiency (11% of control tissue characteristic of Gaucher's disease. Mitochondrial and lysosomal fractions obtained by centrifugation of the spleen homogenate at 6900 and and 20,000g, respectively, contained greater than 80% of the recovered acid phosphatase and heat-stable glucocerebrosidase activator activities. In addition, 60% of the residual glucocerebrosidase activity was recovered in the mitochondrial and lysosomal fractions. The lysosomal and mitochondrial fractions were subjected to equilibrium sucrose density gradient centrifugation. Analysis of the sucrose gradient of the crude mitochondrial fraction demonstrated the mitochondrial marker enzyme (cytochrome oxidase) banding with a specific gravity of 1.19 g/ml, whereas the heat-stable activating factor banded in an acid phosphatase-rich fraction having a specific gravity of 1.12 g/ml. Sucrose gradient analysis of the crude lysosomal fraction obtained from differential centrifugation indicated the activating factor banding with a specific gravity of 1.12 g/ml. Coincident with the activating factor was glucocerebrosidase and acid phosphatase activity. Electron microscopic examination of fractions from each of the sucrose density gradients demonstrated that the glucocerebrosidase activating factor was located in the same acid phosphatase-rich fractions that contained the characteristic Gaucher deposits. Furthermore, when Gaucher deposits were isolated and purified independently by a sucrose gradient procedure, they were found to contain high concentrations of the heat-stable glucocerebrosidase activator. The specific activity of the glucocerebrosidase activating factor was approximately 15-fold greater in the extensively purified Gaucher deposits than in the crude extract of Gaucher spleen from which the deposits were isolated. These observations indicate that the heat-stable activator is associated with the storage deposits contained in lysosomes of the Gaucher cell.     

Syntheses of 2-acetamido-2-deoxy-4-O-βD-galacto-pyranosyl-D-mannose and -D-glucose and of 2-acetamido-2-deoxy-4-O-βD-mannopyranosyl-D-mannose and -D-glucose

2-acetamido-2-deoxy-4-O-β-D-galactopyranosyl-D-mannose (6) and -D-glucose (7) were prepared by addition of nitromethane to 3-O-β-D-galactopyranosyl-D-arabinose, followed by acetylation, ammonolysis, and application of the Nef reaction. Similarly, 2-acetamido-2-deoxy-4-O-β-D-mannopyranosyl-D-mannose (14) and -D-glucose (15) were prepared by the same scheme from 3-O-β-D-mannopyranosyl-D-arabinose. In the two series of experiments, 6 and 14 were the respective major products. Epimerization of the 2-acetamido-2-deoxy-D-mannose residue in 6 and 14 yielded 7 and 15, respectively.     

Inactivation of pig heart alanine aminotransferase by beta-chloroalanine.

The substrate analog β-chloro-l-alanine rapidly inactivates the pig heart alanine aminotransferase. Inactivation is dependent upon formation of an intermediate. The substrate alanine protects the enzyme by competing with the inhibitor for the formation of this intermediate. Halide and carboxylate anions accelerated the rate of inactivation once the enzyme-inhibitor complex was formed. This acceleration appears to mimic the action with the substrate, for the rate of exchange transamination between unlabeled alanine and labeled pyruvate is similarly accelerated. When labeled inhibitor was used, the inactive enzyme became labeled. The spectral changes which occur resemble, in many respects, those which occur with aspartate aminotransferase when its active-site lysine undergoes alkylation by β-chloroalanine. We conclude that chloroalanine fulfills the criteria for a “suicide substrate.”     

Calmodulin-dependent glycogen synthase kinase: Identification in liver of normal and phosphorylase kinase-deficient rats

We have purified a calmodulin-dependent glycogen synthase kinase from livers of normal and phosphorylase kinase-deficient (gsd/gsd) rats. No differences between normal and gsd/gsd rats were apparent in either (a) the ability of liver extracts to phosphorylate exogenous glycogen synthase in a Ca²⁺- and calmodulin-dependent manner or (b) the purification of the calmodulin-dependent synthase kinase. Although extracts from rat liver, when compared to rabbit liver extracts, had a significantly reduced ability to phosphorylate exogenous synthase, the calmodulin-dependent synthase kinase could be purified from rat liver using a protocol identical to that described for rabbit liver. Moreover, the synthase kinase purified from rat liver had properties very similar to those of the rabbit liver enzyme. The enzyme was completely dependent on calmodulin for activity against glycogen synthase, was unable to phosphorylate phosphorylase b, catalyzed the rapid incorporation of 0.4 mol phosphate/mol of glycogen synthase subunit, selectively phosphorylated sites 1b and 2 in the glycogen synthase molecule, had a Stokes' radius of about 70 Å, and appeared to be composed of subunits of M_r 56,000 and 57,000. These observations led us to conclude that (1) calmodulin-dependent glycogen synthase kinase is distinct from other kinases previously described and (2) the rat liver kinase and the rabbit liver kinase are very similar enzymes.     

Coordinate regulation of adenylate cyclase, protein kinase, and specific enzyme synthesis by cholera toxin in hormonally unresponsive hepatoma cells

Since none of the hormones which activate adenylate cyclase in other tissues have been found to activate adenylate cyclase or to induce tyrosine aminotransferase in cultured Reuber hepatoma cells (H35), despite the stimulatory effects of cyclic AMP derivatives on the latter enzyme, we tested the ability of cholera toxin to influence these processes. At low concentrations cholera toxin was found to mimic the ability of cyclic AMP derivatives to selectively stimulate the synthesis of the aminotransferase. Adenylate cyclase and protein kinase activity were also enhanced, but only after a lag period as in other systems. Specific phosphorylation of endogenous H₁ histone was also shown to be increased by cholera toxin treatment. The increase in tyrosine aminotransferase activity is due to an increase in de novo synthesis as shown by radiolabeling experiments utilizing specific immunoprecipitation. The activity of another soluble enzyme induced by dibutyryl cyclic AMP, PEP carboxykinase, was also stimulated by exposure of H35 cells to cholera toxin. Combinations of cholera toxin and dexamethasone led to greater than additive increases in the activity of both the aminotransferase and carboxykinase. Close coupling of cyclic AMP production with protein kinase activation and enzyme induction was suggested by the observation that the ED₅₀ values for the stimulation of adenylate cyclase, cyclic AMP production, protein kinase, and tyrosine aminotransferase activities were found to be the same (5–7 ng/ml) within experimental error. The results indicate that the adenylate cyclase system in H35 cells is functionally responsive and they support the suggestion that activation of protein kinase is functionally linked to induction of specific enzymes.     

Dissymmetric recognition of the helical sense of deoxyribonucleic acid and evidence for binding of reporter molecules from the minor groove of DNA

Through the utilization of optically active DNP-derivatives of l- and d-proline, evidence is presented which suggests that nucleic acids exist as right-handed helices in solution. The results of ultraviolet absorption, circular dichroism, proton magnetic resonance (pmr), T_m of the helix-coil transition, viscometric, and binding studies are consistent with the above interpretation. It is shown that several types of DNA (i.e., salmon sperm, calf thymus, Micrococcus luteus, poly d(A-T)-poly d(A-T) and poly d(I-C)-poly d(I-C)) exist in a right-handed helical structure in solution. In addition, evidence is presented which strongly indicates that the 2,4-dinitroaniline ring of DNP-proline is intercalated between base-pairs of DNA and the prolyl side chain situated in the minor groove. Moreover, it is shown that the more sterically hindered DNP-derivatives exhibit a higher selectivity for A-T binding sites.     

Decrease of hepatic mono and oligo adenosine diphosphoribose content and augmentation of [14C]ribose incorporation during induction of growth by bovine growth hormone in hypophysectomized rats

Following $\text{in}\text{vivo}$ pulse labeling with [¹⁴C]ribose the specific radioactivities of mono- and polyadenosine diphosphoribose, NAD and adenine nucleotides were determined in livers of hypophysectomized Long Evans rats. These analyses were also performed after induction of growth with four successive daily injections of bovine growth hormone. As a consequence of brief treatment with growth hormone polyadenosine diphosphoribose content markedly diminished whereas its specific radioactivity increased. NAD concentration did not vary but its specific radioactivity increased similarly to that of the homopolymer. The steady state concentration of adenine nucleotides remained unchanged, except for ATP which decreased, and their specific radioactivities uniformly decreased.     

The synthesis of a labile triosephosphate isomerase isozyme in human lymphoblasts and fibroblasts

Human peripheral lymphocytes contain a single electrophoretic form of triosephosphate isomerase (pI = 5.6). However, when induced to undergo blastogenesis by mitogens such as phytohemagglutinin or convanavalin A, a second isozyme (pI = 5.2) is also produced. This new isozyme is also found in human fibroblasts, but is present only in low concentrations in most other human tissues. The two isozymes were isolated from lymphocytes, lymphoblasts, and fibroblasts by isoelectric focusing, and their properties and the requirements for their synthesis were studied. The production of a new isozyme occurs concomitantly with blastogenesis and DNA synthesis, but when DNA synthesis is delayed by hydroxyurea, the appearance of the new isozyme is unaffected. The formation of the new isozyme is inhibited by actinomycin D and puromycin, and thus, appears to be dependent on both RNA and protein synthesis. Lymphocytes grown in the presence of [³H]leucine synthesize the new isozyme which is isotopically labeled, and pulse-chase experiments show that the two isozymes are not interconvertible. Although the two isozymes exhibit essentially identical catalytic properties, they differ markedly with regard to their stability, with the more acidic isozyme being much more labile. The lability of the more acidic isozyme may account for its low levels in most other tissues.     

Comparison of cytochromes P-450 with high activity toward benzo[a]pyrene purified from liver microsomes of beta-naphthoflavone and 3-methylcholanthrene-pretreated rats

Cytochromes P-450 with high activity toward benzo[a]pyrene were isolated from liver microsomes of rats treated with either β-naphthoflavone or 3-methylcholanthrene and examined for similarity using several physical and catalytic criteria. The β-naphthofla-vone-inducible cytochrome P-446 and the 3-methylcholanthrene-inducible cytochrome P-448 have the same subunit molecular weight (56,000 ± 1000) and electrophoretic mobility. Antibodies prepared to either form cross-react with each form without spurring in Ouchterlony double-diffusion experiments suggesting immunochemical identity. After proteolytic digestion with Staphylococcus aureus SV-8 protease and electrophoresis, both Cytochromes P-450 show the presence of the same bands. Both cytochromes have the same absorption maximum (446.5 ± 0.5 nm) in the CO-reduced absolute spectrum. The catalytic activity toward benzo[a]pyrene of cytochrome P-446 is somewhat greater than that of cytochrome P-448. Thus, all the physical evidence suggests identity of the two cytochromes. The significance of the difference in catalytic activity remains to be defined.     

Defect in 3'-phosphoadenosine 5'-phosphosulfate synthesis in brachymorphic mice. II. Tissue distribution of the defect

The tissue distribution of the defective PAPS synthetic pathway in homozygous brachymorphic mice ($\text{bm}\text{bm}$) has been investigated using four different criteria: (i) incorporation of ³⁵SO₄^2? into adenosine 5′-phosphosulfate (APS), 3′-phosphoadenosine 5′-phosphosulfate (PAPS), and endogenous macromolecular acceptors, (ii) APS kinase (adenylylsulfate kinase; ATP:adenylylsulfate 3′-phosphotransferase, EC 2.7.1.25) activity, (iii) ATP sulfurylase (sulfate adenylyltransferase; ATP:sulfate adenylyltransferase, EC 2.7.7.4) activity, (iv) thermostability of ATP sulfurylase. With respect to the first three criteria, the results indicate that liver is affected as profoundly as cartilage (K. Sugahara and N. B. Schwartz, Arch. Biochem. Biophys. (1982) 214, 589–601). In contrast, skin and brain show no differences between normal and mutant. Kidney is significantly, but only moderately, affected. The results from thermostability studies demonstrate that ATP sulfurylase activity is more labile in $\text{bm}\text{bm}$ cartilage, liver, and kidney, but not in skin or brain, supporting the above-observed distribution of the defect. Therefore, the present results indicate a multiple, but not universal, tissue distribution of the defective PAPS synthetic pathway in $\text{bm}\text{bm}$ mice. Furthermore, these findings support the suggestion that ATP sulfurylase as well as APS kinase is defective in brachymorphic mice.     

Formation and stability of the complex formed between human antithrombin-III and thrombin

The inhibition of thrombin by antithrombin-III involves formation of a 1:1 covalent complex between protease and inhibitor and concomitant cleavage of the antithrombin-III peptide chain after Arg-385. The resultant fragment remains connected to the complex via a disulfide bond. This complex spontaneously breaks down into a fragment of approximately 55,000 daltons and smaller peptides. Breakdown is prevented by the presence of hydroxylamine or diisopropylflurophosphate, or by denaturation with urea. It occurs even if the purified complex is treated with diisopropylflurophosphate prior to purification, and can be greatly accelerated by the presence of small amounts of active thrombin. The initial sites of proteolytic attack on the complex are after Arg-13 of the thrombin A chain and Arg-68 of the thrombin B chain. These data indicate that active thrombin can be released from the antithrombin-thrombin complex, and that thrombin becomes more susceptible to proteolytic attack when complexed with antithrombin.