Selective carboxymethylation of cysteine-174 of the beta 2 beta 2 and beta 1 beta 1 human liver alcohol dehydrogenase isoenzymes by iodoacetate

The beta 1 beta 1 and beta 2 beta 2 human liver alcohol dehydrogenase isoenzymes differ by only one residue at the coenzyme-binding site; Arg-47 in beta 1 is replaced by His in the beta 2 subunit. Since Arg-47 is thought to facilitate the carboxymethylation of Cys-46 in horse liver alcohol dehydrogenase by binding halo acids in a Michaelis-Menten complex prior to inactivation, the specificity and kinetics of modification of the two human liver beta beta isoenzymes with iodoacetate were compared. Both of the beta beta isoenzymes were inactivated by treatment with iodo[14C]acetate, and one Cys per subunit was carboxymethylated. Cys-174, which is a ligand to the active-site zinc atom in horse liver alcohol dehydrogenase, was selectively carboxymethylated in each of the human beta beta isoenzymes; less than 15% of the iodo[14C]acetate incorporated into the enzyme appeared in Cys-46. Therefore, the three-dimensional structure of the basic amino acids in the anion-binding site of the human beta beta isoenzymes appears to be different from that of horse liver alcohol dehydrogenase. The kinetics of alkylation are consistent with the formation of a Michaelis-Menten complex before inactivation of the isoenzymes. The average Ki values for iodoacetate were 10 and 16 mM for beta 1 beta 1 and beta 2 beta 2, respectively, and maximal rate constants for inactivation were 0.22 and 0.17 min-1, respectively. From these data, it can be concluded that there is a relatively minor effect of the substitution of His for Arg at position 47 on the kinetics of inactivation.     

Localization of the 5-phospho-alpha-D-ribosyl-1-pyrophosphate binding site of human hypoxanthine-guanine phosphoribosyltransferase.

Human erythrocyte hypoxanthine-guanine phosphoribosyltransferase (HPRT) is inactivated by iodoacetate in the absence, but not in the presence, of the substrate, 5-phospho-alpha-D-ribosyl-1-pyrophosphate (PRib-PP). Treatment of HPRT with [14C]iodoacetate followed by tryptic digestion, peptide separation and sequencing has shown that Cys-22 reacts with iodoacetate only in the absence of PRib-PP; this strongly suggests that Cys-22 is in or near the PRib-PP binding site. In contrast, Cys-105 reacts with [14C]iodoacetate both in the presence and absence of PRib-PP. Carboxymethylation of Cys-22 resulted in an increase in the Km for PRib-PP, but no change in Vmax. Storage of HPRT also resulted in an increase in the Km for PRib-PP and a decrease in its susceptibility to inactivation by iodoacetate. Dialysis of stored enzyme against 1 mM dithiothreitol resulted in a marked decrease in Km for PRib-PP. The stoichiometry of the reaction of [14C]iodoacetate with Cys-22 in HPRT leading to inactivation (approx. 1 residue modified per tetramer) showed that, in this preparation of HPRT purified from erythrocytes, only about 25% of the Cys-22 side chains were present as free and accessible thiols. Titration of thiol groups [with 5,5'-dithiobis(2-nitrobenzoic acid)] and the effect of dithiothreitol on Km for PRib-PP indicate that oxidation of thiol groups occurs on storage of HPRT, even in the presence of 1 mM beta-mercaptoethanol.     

Examination of subunit interaction in human ADH: carboxymethylation of the heterodimer beta 2 gamma 1.

Conflicting results have been obtained on whether the subunits of the human dimeric enzyme alcohol dehydrogenase interact kinetically. To examine this question, chemical modification by iodoacetate was used to selectively inactivate the beta 2 subunit of the heterodimeric isozyme beta 2 gamma 1. Subsequent studies of the modified beta 2 gamma 1 which, presumably, has only one functional active site per dimeric molecule, indicate that it is still active. Moreover, the properties of this hybrid are similar to those of the unmodified subunit. From these results, it is fair to conclude that the individual subunits of alcohol dehydrogenase contribute to the activity of the dimeric isozymes in an independent manner.     

Human liver alcohol dehydrogenase. 1. The primary structure of the beta 1 beta 1 isoenzyme

Determination of the amino acid sequence of the beta 1 subunit from the class I (pyrazole-sensitive) human liver alcohol dehydrogenase isoenzyme beta 1 beta 1 revealed a 373-residue structure differing at 48 positions (including a gap) from that of the subunit of the well studied horse liver alcohol dehydrogenase EE isoenzyme. The structure deduced is compatible with known differences in composition, ultraviolet absorbance, electrophoretic mobility and catalytic properties between the horse and human enzymes. All zinc-liganding residues of the horse E subunit are strictly conserved in the human beta 1 subunit, despite an earlier report of a mutation involving Cys-46. This residue therefore remains conserved in all known alcohol dehydrogenase structures. However, the total cysteine content of the beta 1 structure is raised from 14 in the subunit of the horse enzyme to 15 by a Tyr----Cys exchange. Most exchanges are on the surface of the molecule and of a well conserved nature. Substitutions close to the catalytic centre are of interest to explain the altered substrate specificity and different catalytic activity of the beta 1 homodimer. Functionally, a Ser----Thr exchange at position 48 appears to be of special importance, since Thr-48 in beta 1 instead of Ser-48 in the horse enzyme can restrict available space. Four other substitutions also line the active-site pocket, and appear to constitute partly compensated exchanges.     

Elongation factor 1 beta gamma from Artemia. Purification and properties of its subunits

The guanine nucleotide exchange factor, elongation factor 1 beta gamma (EF-1 beta gamma) has been purified from Artemia cysts using an improved method. The protein consists of two distinct polypeptides with relative molecular masses of 26,000 (EF-1 beta) and 46,000 (EF-1 gamma). A nucleoside diphosphate phosphotransferase activity often found in EF-1 beta gamma preparations has been completely separated from the actual guanine nucleotide exchange stimulatory activity of EF-1 beta gamma, thus indicating that nucleotide diphosphate phosphotransferase is not an intrinsic property of EF-1 beta. Both EF-1 beta gamma and EF-1 beta have been shown to stimulate the following three reactions to a comparable degree: (a) exchange of GDP bound to EF-1 alpha with exogenous GDP; (b) EF-1 alpha-dependent binding of Phe-tRNA to ribosomes; (c) poly(U)-dependent poly(phenylalanine) synthesis. However, a significantly higher nucleotide exchange rate was observed in the presence of EF-1 beta gamma compared to EF-1 beta alone. Concerning elongation factor 1 gamma (EF-1 gamma) the following observations were made. In contrast to EF-1 beta, pure EF-1 gamma is rather insoluble in aqueous buffers, but the tendency to precipitate can be partially suppressed by the addition of detergents. In particular, EF-1 gamma partitions solely into the detergent phase of Triton X-114 solutions. EF-1 gamma is also more susceptible to spontaneous, specific fragmentation. It is remarkably that about 5% of the cellular pool of EF-1 beta gamma was found to be present in membrane fractions, under conditions where no EF-1 alpha was detectable in these fractions. Furthermore it was noted that EF-1 beta gamma copurified strongly with tubulin on DEAE-cellulose. Moreover, it was observed that from a mixture of EF-1 beta gamma and tubulin, EF-1 gamma coprecipitates with tubulin using a non-denaturating immunoprecipitation technique. These findings suggest that EF-1 gamma has a hydrophobic domain and interacts with membrane and cytoskeleton structures in the cell.     

G-protein beta gamma forms: identity of beta and diversity of gamma subunits

Signal-transducing G-proteins are heterotrimers composed of GTP-binding alpha subunits in association with a tightly bound complex of beta and gamma subunits. While the alpha subunits are recognized as a family of diverse structures, beta and gamma subunits have also been found as heterogeneous isoforms. To investigate the diversity and tissue specificity of the beta gamma complexes, we have examined homogeneous oligomeric G-proteins from a variety of sources. The beta and gamma subunits isolated from the major-abundance G-proteins from bovine brain, bovine retina, rabbit liver, human placenta, and human platelets were purified and subjected to biochemical and immunological analysis. Protease mapping and immune recognition revealed an identical profile for each of the two distinctly migrating beta isoforms (beta 36 and beta 35) regardless of tissue or G-protein origin. Digestion with V8 protease revealed four distinct, clearly resolved terminal fragments for beta 36 and two for beta 35. Trypsin and chymotrypsin digestion yielded numerous bands, but again each form had a unique profile with no tissue specificity. Tryptic digestion was found to be conformationally specific with the most resistant structure being the native beta gamma complex. With increasing trypsin, the complex was digested but in a pattern distinct from that for denatured beta. In contrast to the two highly homologous beta structures, examination of this set of proteins revealed at least six distinct gamma peptides. Two unique gamma peptides were found in bovine retinal Gt and three gamma peptides in samples of bovine brain derived Go/Gi. Human placental and platelet Gi samples each contained a unique gamma.(ABSTRACT TRUNCATED AT 250 WORDS)     

Calcium channel beta subunit promotes voltage-dependent modulation of alpha 1 B by G beta gamma

Voltage-dependent calcium channels (VDCCs) are heteromultimers composed of a pore-forming alpha1 subunit and auxiliary subunits, including the intracellular beta subunit, which has a strong influence on the channel properties. Voltage-dependent inhibitory modulation of neuronal VDCCs occurs primarily by activation of G-proteins and elevation of the free G beta gamma dimer concentration. Here we have examined the interaction between the regulation of N-type (alpha 1 B) channels by their beta subunits and by G beta gamma dimers, heterologously expressed in COS-7 cells. In contrast to previous studies suggesting antagonism of G protein inhibition by the VDCC beta subunit, we found a significantly larger G beta gamma-dependent inhibition of alpha 1 B channel activation when the VDCC alpha 1 B and beta subunits were coexpressed. In the absence of coexpressed VDCC beta subunit, the G beta gamma dimers, either expressed tonically or elevated via receptor activation, did not produce the expected features of voltage-dependent G protein modulation of N-type channels, including slowed activation and prepulse facilitation, while VDCC beta subunit coexpression restored all of the hallmarks of G beta gamma modulation. These results suggest that the VDCC beta subunit must be present for G beta gamma to induce voltage-dependent modulation of N-type calcium channels.     

Function and reactivity of sulfhydryl groups of rat liver glycine methyltransferase

Rat liver glycine methyltransferase is completely inactivated by 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB). Treatment of the inactivated enzyme with KCN results in a reactivated enzyme having values of Vmax and S0.5 for S-adenosyl-L-methionine comparable to those of the native enzyme and about a 4-fold greater Km value for glycine. Kinetics of inactivation and reactivation show that one cysteine residue is involved in this process. Reaction of the methyltransferase with iodoacetate leads to partial inactivation of the enzyme; about 22% of the initial activity is retained in the modified enzyme. The relationship between the loss of enzyme activity and the number of iodoacetate molecules incorporated and the sequence analysis of peptides containing the modified residues indicate that carboxymethylation of Cys-282 is responsible for loss of activity. The observations that the activity of the cyanylated glycine methyltransferase shows no decrease upon incubation with iodoacetate and, conversely, the residual activity associated with the iodoacetate-modified enzyme is not abolished by DTNB suggest that Cys-282 is also involved in the inactivation by DTNB. Besides this residue, Cys-185, Cys-246, and Cys-262 are modified upon prolonged incubation with iodoacetate. 5'-[p-(Fluorosulfonyl)benzoyl]adenosine (FSBA) inactivates glycine methyltransferase by forming 1 disulfide/subunit [Fujioka, M., & Ishiguro, Y. (1986) J. Biol. Chem. 261, 6346-6351]. Despite this stoichiometry, treatment of the FSBA-inactivated enzyme with unlabeled iodoacetate and then with iodo[14C]acetate after reduction with 2-mercaptoethanol and subsequent peptide analysis show that the incorporated radioactivity is distributed equally among Cys-185, Cys-246, Cys-262, and Cys-282.(ABSTRACT TRUNCATED AT 250 WORDS)     

Mapping the functional domains of the eukaryotic elongation factor 1 beta gamma

The functional domains of the eukaryotic elongation factor (EF) 1 beta gamma have been delineated with the use of limited proteolysis, protein microsequencing, gel electrophoresis under non-denaturing conditions and antibodies against EF-1 beta and EF-1 gamma. By means of limited proteolysis, it was possible to obtain large fragments of EF-1 beta. In contrast to amino-terminal fragments, those derived from the carboxy-terminal part of EF-1 beta were still active in enhancing the guanine nucleotide exchange of GDP bound to EF-1 alpha. With the same technique of limited proteolysis, it was possible to isolate a trypsin-resistant core from EF-1 beta gamma containing polypeptide chain fragments derived from both subunits. A polyvalent antiserum against EF-1 beta and two monoclonal antibodies against EF-1 gamma were used to identify the protein fragments in this core. The monoclonal antibodies were shown to recognize different epitopes, one localized on the amino-terminal and another on the carboxy-terminal half of EF-1 gamma. The antiserum against EF-1 beta and one of the monoclonal antibodies (mAb 36E5), which recognized the amino-terminal half of EF-1 gamma, reacted with this trypsin-resistant core. We conclude that the amino-terminal halves of both EF-1 beta and EF-1 gamma are firmly attached to each other, and that the carboxy-terminal part of EF-1 beta interacts with EF-1 alpha.     

Human liver alcohol dehydrogenase. 2. The primary structure of the gamma 1 protein chain

The primary structure of the gamma 1 subunit of human liver alcohol dehydrogenase isoenzyme gamma 1 gamma 1 was deduced by characterization of 36 tryptic and 2 CNBr peptides. The polypeptide chain is composed of 373 amino acid residues. gamma 1 differs from the beta 1 subunit of human liver alcohol dehydrogenase at 21 positions, and from the E subunit of horse liver alcohol dehydrogenase at 43 positions including a gap at position 128 as in the beta 1 subunit. All zinc-liganding residues from the E subunit of the horse protein and the beta 1 subunit of the human enzyme are conserved, but like beta 1, gamma 1 also has an additional cysteine residue at position 286 (in the positional numbering system of the horse enzyme) due to a Tyr----Cys exchange. Most amino acid exchanges preserve the properties of the residues affected and are largely located on the surface of the molecules, away from the active site and the coenzyme binding region. However, eight positions with charge differences in relation to the E subunit of the horse enzyme are noticed. These result in a net positive charge increase of one in gamma 1 versus E, explaining the electrophoretic mobilities on starch gels. Of functional significance is the conservation of Ser-48 in gamma 1 relative to E. The residue is close to the active site but different (Thr-48) in the beta 1 subunit of the human enzyme. Thus, the closer structural relationship between human gamma 1 and horse E enzyme subunit than between beta 1 and E is also reflected in functionally important residues, explaining a greater similarity between gamma 1 gamma 1 and EE than between beta 1 beta 1 and EE.     

AMP-activated protein kinase subunit interactions: beta1:gamma1 association requires beta1 Thr-263 and Tyr-267

AMP-activated protein kinase (AMPK) plays multiple roles in the body's overall metabolic balance and response to exercise, nutritional stress, hormonal stimulation, and the glucose-lowering drugs metformin and rosiglitazone. AMPK consists of a catalytic alpha subunit and two non-catalytic subunits, beta and gamma, each with multiple isoforms that form active 1:1:1 heterotrimers. Here we show that recombinant human AMPK alpha1beta1gamma1 expressed in insect cells is monomeric and displays specific activity and AMP responsiveness similar to rat liver AMPK. The previously determined crystal structure of the core of mammalian alphabetagamma complex shows that beta binds alpha and gamma. Here we show that a beta1(186-270)gamma1 complex can form in the absence of detectable alpha subunit. Moreover, using alanine mutagenesis we show that beta1 Thr-263 and Tyr-267 are required for betagamma association but not alphabeta association.     

Characterization of commercial laminin preparations from human placenta in comparison to recombinant laminins 2 (alpha2beta1gamma1), 8 (alpha4beta1gamma1), 10 (alpha5beta1gamma1).

Laminins, a family of large heterotrimeric (alphabetagamma) proteins, are major components of basement membranes implicated in a variety of cellular functions. Different commercial laminin preparations isolated from human placenta have been widely used in functional studies but their molecular properties are poorly known. In the present study, we characterized several of these preparations by ELISA, silver staining and Western blotting, in comparison to mouse laminin 1 (alpha1beta1gamma1), and recombinant human laminins 2 (alpha2beta1gamma1), 8 (alpha4beta1gamma1) and 10 (alpha5beta1gamma1). The cell migration-promoting activity of different batches was also tested. The placenta laminin preparations differed from one another and consisted of highly fragmented proteins, a mixture of laminin isoforms, and/or contaminating fibronectin. Major functional differences between batches were also observed, reflecting molecular heterogeneity. Previous data obtained in functional studies using these preparations need to be interpreted with caution and may require revision, and future functional studies demand prior molecular characterization of the laminins, particularly their alpha-chain.     

Unique assignment of inter-subunit association in GABA(A) alpha 1 beta 3 gamma 2 receptors determined by molecular modeling

Recent publications defined requirements for inter-subunit contacts in a benzodiazepine-sensitive GABA(A) receptor (GABA(A)R alpha 1 beta 3 gamma 2). There is strong evidence that the heteropentameric receptor contains two alpha 1, two beta 3, and one gamma 2 subunit. However, the available data do not distinguish two possibilities: When viewed clockwise from an extracellular viewpoint the subunits could be arranged in either gamma 2 beta 3 alpha 1 beta 3 alpha 1 or gamma 2 alpha 1 beta 3 alpha 1 beta 3 configurations. Here we use molecular modeling to thread the relevant GABA(A)R subunit sequences onto a template of homopentameric subunits in the crystal structure of the acetylcholine binding protein (AChBP). The GABA(A) sequences are known to have 15-18% identity with the acetylcholine binding protein and nearly all residues that are conserved within the nAChR family are present in AChBP. The correctly aligned GABA(A) sequences were threaded onto the AChBP template in the gamma 2 beta 3 alpha 1 beta 3 alpha 1 or gamma 2 alpha 1 beta 3 alpha 1 beta 3 arrangements. Only the gamma 2 alpha 1 beta 3 alpha 1 beta 3 arrangement satisfied three known criteria: (1) alpha 1 His(102) binds at the gamma 2 subunit interface in proximity to gamma 2 residues Thr(142), Phe(77), and Met(130); (2) alpha 1 residues 80-100 bind near gamma 2 residues 91-104; and (3) alpha 1 residues 58-67 bind near the beta 3 subunit interface. In addition to predicting the most likely inter-subunit arrangement, the model predicts which residues form the GABA and benzodiazepine binding sites.     

Chain specificity assignment of monoclonal antibodies to human laminins by using recombinant laminin beta1 and gamma1 chains.

In the present study, the chain specificity of 16 commonly used monoclonal antibodies to human laminin(s) was analysed by using recombinant laminin beta1 and gamma1 chains. By ELISA, all antibodies reacted with purified placenta laminin, and most antibodies recognised either recombinant beta1 or gamma1 chains. Reactivity and chain specificity was confirmed against the recombinant chains in Western blotting under non-reducing conditions, and only a few antibodies were reactive under reducing conditions. Most antibodies were able to immunoprecipitate associated laminin beta1/gamma1 chains from platelet lysates. Based on these results and data from the literature, a tentative epitope map is presented.     

Interaction between elongation factors 1beta and 1gamma from Bombyx mori silk gland

Elongation factor 1 (EF-1) from the silk gland of Bombyx mori consists of four subunits: alpha (51 kDa), beta (26 kDa), gamma (49 kDa), and delta (33 kDa). The EF-1alpha subunit catalyzes the binding of aminoacyl-tRNA to the ribosome concomitant with the hydrolysis of GTP. The EF-1alpha-bound GDP is then exchanged for GTP by the EF-1betagammadelta complex. To facilitate analysis of the roles of the individual EF-1beta, gamma, and delta subunits in GDP/GTP exchange on EF-1alpha, we cloned the cDNAs for these subunits and expressed them in Escherichia coli. EF-1beta, EF-1gamma, and the carboxyl-terminal half of EF-1delta were expressed, purified, and examined for protein:protein interactions by gel filtration chromatography and by a quartz-crystal microbalance method. An 80-kDa species containing EF-1beta and gamma subunits in a 1:1 molar ratio was detected by gel filtration. A higher molecular weight species containing an excess of EF-1gamma relative to EF-1beta was also detected. The amino-terminal region of EF-1beta (amino acid residues 1-129) was sufficient for binding to EF-1gamma. The carboxyl-terminal half of EF-1delta did not appear to form a complex with EF-1gamma.     

Effects of conditioning depolarization and repetitive stimulation on Q beta and Q gamma charge components in frog cut twitch fibers

Charge movement was measured in frog cut twitch fibers with the double Vaseline-gap technique. Steady-state inactivation of charge movement was studied by changing the holding potential from -90 mV to a level ranging from -70 to -30 mV. Q beta and Q gamma at each holding potential were separated by fitting the Q-V plot with a sum of two Boltzmann distribution functions. At -70 mV Q beta and Q gamma were inactivated to 54.0% (SEM 2.2) and 82.7% (SEM 3.0) of the amounts at -90 mV. At holding potentials greater than or equal to -60 mV, more Q gamma was inactivated than Q beta, and at -30 mV Q gamma was completely inactivated but Q beta was not. There was no holding potential at which Q beta was unaffected and Q gamma was completely inactivated. The differences between the residual fractions of Q beta and Q gamma are significant at all holding potentials (P less than 0.001-0.05). The plot of the residual fraction of Q beta or Q gamma versus holding potential can be fitted well by an inverted sigmoidal curve that is a mirror image of the activation curve of the respective charge component. The pair of curves for Q gamma correlates well with those for tension generation or Ca release obtained by other investigators. The time courses of the inactivation of Q beta and Q gamma were studied by obtaining several Q-V plots with conditioning depolarizations lasting 1-20 s and separating each Q-V plot into Q beta and Q gamma components by fitting with a sum of two Boltzmann distribution functions. The inactivation time constant of Q beta was found to be 5-10 times as large as that of Q gamma. During repetitive stimulation, prominent I gamma humps could be observed in TEST-minus-CONTROL current traces and normal Q gamma components could be separated from the Q-V plots, whether 20 or 50 mM EGTA was present in the internal solution, whether 2 or 10 stimulations were used, and whether the stimuli were separated by 400 ms or 6 s. Repetitive stimulation slowed the kinetics of the I gamma hump and could shift the Q-V curve slightly in the depolarizing direction in some cases, resulting in an apparent suppression of charge at the potentials that fall on the steep part of the Q-V curve.     

Phosphoinositide-dependent kinase l (PDK1) haplo-insufficiency inhibits production of alpha/beta (alpha/beta) but not gamma delta (gamma/delta) T lymphocytes

In the present study, we have explored the impact of deleting a single allele of PDK1 in T cell progenitors on alpha/beta and gamma/delta T cell development. The data show that deleting a single allele of PDK1 allows differentiation of alpha/beta T cells but prevents their proliferative expansion in the thymus. Accordingly, mice with T cells that are haplo-insufficient for PDK1 have reduced numbers of thymocytes and alpha/beta peripheral T cells. T cell progenitors also give rise to gamma/delta T cells but in contrast to the loss of alpha/beta T cells in T-PDK1 null and haplo-insufficient mice, there were increased numbers of gamma/delta T cells. The production of alpha/beta T cells is dependent on the proliferative expansion of thymocytes and is determined by a balance between the frequency with which cells enter the proliferative phase of the cell cycle and rates of cell death. Herein, we show that PDK1 haplo-insufficient thymocytes have no defects in their ability to enter the cell cycle but show increased apoptosis. PDK1 thus plays a determining role in the development of alpha/beta T lymphocytes but does not limit gamma/delta T cell development.     

Recombinant rat liver guanidinoacetate methyltransferase: reactivity and function of sulfhydryl groups

Rat liver guanidinoacetate methyltransferase, produced in Escherichia coli by recombinant DNA technique, possesses five cysteine residues per molecule. No disulfide bond is present. Analysis of the chymotryptic peptides derived from the iodo[14C]acetate-modified enzyme shows that Cys-90, Cys-15, Cys-219, and Cys-207 are alkylated by the reagent in order of decreasing reactivity. Incubation of the enzyme with excess 5,5'-dithiobis(2-nitrobenzoate) (DTNB) in the absence and presence of cystamine [2,2'-dithiobis(ethylamine)] causes the appearance of 4 and 5 mol of 2-nitro-5-mercaptobenzoate/mol of enzyme, respectively. Reaction of the methyltransferase with an equimolar amount of DTNB results in an almost quantitative disulfide cross-linking of Cys-15 and Cys-90 with loss of a large portion of the activity. The methyltransferase is completely inactivated by iodoacetate following nonlinear kinetics. Comparison of the extent of inactivation with that of modification of cysteine residues and the experiment with the enzyme whose Cys-15 and Cys-90 are cross-linked suggest that alkylation of Cys-15 and Cys-90 results in a partially active enzyme and that carboxymethylation of Cys-219 completely eliminates enzyme activity. The inactivation of guanidinoacetate methyltransferase by iodoacetate or DTNB is not protected by substrates. Furthermore, disulfide cross-linking of Cys-15 and Cys-90 or carboxymethylation of Cys-219 does not impair the enzyme's capacity to bind S-adenosylmethionine. Thus, these cysteine residues appear to occur outside the active-site region, but their integrity is crucial for the expression of enzyme activity.