Molecular recognition. Conformational analysis of limited proteolytic sites and serine proteinase protein inhibitors

The conformations of known tryptic limited proteolytic sites have been analysed and compared to the structures of the binding regions of serine proteinase inhibitors, as they are found when complexed to a serine proteinase. Conformational parameters studied include main-chain torsion angles, root-mean-square fits, accessibility, mobility and protrusion indices. As observed before, the inhibitors share a common main-chain conformation at the binding loop from P3-P'3 (Schechter & Berger notation), which is maintained throughout all the serine proteinase inhibitor families for which X-ray data is available, despite lack of similarity in the rest of the protein. This canonical structure is not found amongst the limited proteolytic sites (or nicksites), which differ markedly from the inhibitor binding loop conformation, and also amongst themselves. The experimentally determined nicksites are in general both accessible and protruding; as are the inhibitor binding loops, as well as being typically flexible regions of structure, as denoted by elevated temperature factors from crystallographic determinations. For cleavage by serine proteinases these loops must radically alter their local conformations and a large motion of the loop relative to the structure, in some cases, would be required to orientate these sites for cleavage.     

Differences in local conformation in human apolipoprotein B-100 of plasma low density and very low density lipoproteins as identified by cathepsin D

The conformational changes of human apolipoprotein (apo) B-100 which accompany the conversion of plasma very low density lipoproteins (VLDL) to low density lipoproteins (LDL) were investigated by studying the accessibility of apoB-100 in LDL and VLDL to limited proteolysis with cathepsin D, an aspartyl proteinase involved in intracellular protein degradation. We characterized the proteolytic products of apoB-100 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by NH2-terminal sequence analysis to locate cleavage sites. The results identified at least 10 cleavage products generated from apoB-100 and showed differential accessibility of cleavage sites for cathepsin D in apoB-100 between LDL and VLDL. We identified a specific peptide region (residues 2660-2710), which is preferentially accessible to limited proteolysis by cathepsin D but inaccessible to limited proteolysis by 12 other enzymes tested. Within this peptide region, cathepsin D cleaved apoB-100 of LDL and VLDL preferentially at different sites, separated by 33-36 amino acids (2665-2666 or 2668-2669 (LDL) and 2701-2702 (VLDL]. In addition, we identified a cleavage site, located at residues 3272-3273, specific for cathepsin D, which is contained within the COOH-terminal enzyme-accessible peptide region (residues 3180-3280), which we have demonstrated using 12 endoproteases with various specificities. The previously identified NH2-terminal region (residues 1280-1320) appears to be resistant to limited cleavage by cathepsin D. However, a new site was revealed only approximately 66 kDA from the NH2 terminus. We conclude that differential accessibility and the shift of the novel scission site for cathepsin D by 33-36 amino acids indicate significant differences in local conformation at these sites in apoB-100 as VLDL are converted to LDL.     

A serpin-induced extensive proteolytic susceptibility of urokinase-type plasminogen activator implicates distortion of the proteinase substrate-binding pocket and oxyanion hole in the serpin inhibitory mechanism.

The formation of stable complexes between serpins and their target serine proteinases indicates formation of an ester bond between the proteinase active-site serine and the serpin P1 residue [Egelund, R., Rodenburg, K.W., Andreasen, P.A., Rasmussen, M.S., Guldberg, R.E. & Petersen, T.E. (1998) Biochemistry 37, 6375-6379]. An important question concerning serpin inhibition is the contrast between the stability of the ester bond in the complex and the rapid hydrolysis of the acyl-enzyme intermediate in general serine proteinase-catalysed peptide bond hydrolysis. To answer this question, we used limited proteolysis to detect conformational differences between free urokinase-type plasminogen activator (uPA) and uPA in complex with plasminogen activator inhibitor-1 (PAI-1). Whereas the catalytic domain of free uPA, pro-uPA, uPA in complex with non-serpin inhibitors and anhydro-uPA in a non-covalent complex with PAI-1 was resistant to proteolysis, the catalytic domain of PAI-1-complexed uPA was susceptible to proteolysis. The cleavage sites for four different proteinases were localized in specific areas of the C-terminal beta-barrel of the catalytic domain of uPA, providing evidence that the serpin inhibitory mechanism involves a serpin-induced massive rearrangement of the proteinase active site, including the specificity pocket, the oxyanion hole, and main-chain binding area, rendering the proteinase unable to complete the normal hydrolysis of the acyl-enzyme intermediate. The distorted region includes the so-called activation domain, also known to change conformation on zymogen activation.     

Limited proteolysis of chicken gizzard 5'-nucleotidase.

Chicken gizzard 5'-nucleotidase represents an ectoenzyme which is linked to the plasma membrane via a phosphatidylinositol glycan. We have characterized the possible domain-like organization of 5'-nucleotidase by limited proteolysis. A hydrophobic proteolytic fragment carrying the intact C-terminus, as well as two major hydrophilic products, were identified. We developed procedures for specific radiolabelling of the active center of 5'-nucleotidase. This allowed us to locate the catalytic site within hydrophilic fragments obtained after limited proteolysis. We demonstrate that removal of N-linked carbohydrate chains increases the sensitivity of 5'-nucleotidase to proteolytic attack, indicating that sugar moieties protect against proteolysis. 5'-Nucleotidase represents a binding protein for components of the extracellular matrix. The interaction between 5'-nucleotidase and the laminin/nidogen complex unmasked proteolytic cleavage sites in the C-terminal portion of the enzyme. This resulted in the specific production of a hydrophilic form of 5'-nucleotidase. In summary, we have further characterized chicken gizzard 5'-nucleotidase: (1) the protein is organized into two domain-like structures, (2) the N-terminal domain harbors the active center; (3) N-linked carbohydrates protect the protein against proteolytic degradation; (4) interaction with components of the extracellular matrix alters the conformation of 5'-nucleotidase.     

Limited proteolysis of native proteins: the interaction between avidin and proteinase K.

Avidin is a tetramer of 16-kDa subunits that have a high affinity for biotin. Proteolysis of native apoavidin by proteinase K results in a limited attack at the loop between beta-strands 3 and 4, involving amino acids 38-43. Specifically, sites of proteolysis are at Thr 40-Ser 41 and Asn 42-Glu 43. The limited proteolysis results in an avidin product that remains otherwise intact and which has enhanced binding for 4'-hydroxyazobenzene-2-benzoic acid (HABA), a chromogenic reporter that can occupy the biotin-binding site. Saturation of the biotin-binding site with the natural ligand protects avidin from proteolysis, but saturation with HABA enhances the rate of proteolysis of the same site. Analysis of the three-dimensional structures of apoavidin and holoavidin reveals that the 3-4 loop is accessible to solvent and scores highly in an algorithm developed to identify sites of proteolytic attack. The structure of holoavidin is almost identical to the apoprotein. In particular, the 3-4 loop has the same structure in the apo and holo forms, yet there are marked differences in proteolytic susceptibility of this region. Evidence suggests that the 3-4 loop is rather mobile and flexible in the apoprotein, and that it becomes constrained upon ligand binding. In one crystal structure of the apoprotein, this loop appears constrained by contacts with symmetry-related molecules. Structural analyses suggest that the "lid" to the biotin-binding site, formed by the 3-4 loop, is displaced and made more accessible by HABA binding, thereby enhancing its proteolytic susceptibility.     

Inactivation of yeast ornithine decarboxylase by polyamines in vivo does not result from the incorporation of polyamines into enzyme protein

The peptide mixture obtained from controlled proteolytic digestion of ligandin with proteinase K or subtilisin retained 40% of glutathione-S-transferase and steroid isomerase activities, immunological reactivity and lower affinity bilirubin binding but binding at the primary site was abolished. When these limited proteolytic digests, which had no intact ligandin as determined by SDS gel electrophoresis, were subjected to Sephadex G-75 column chromatography, 40–50% of the peptide fragments were recovered in fractions where intact ligandin eluted. The results suggest that intact ligandin is not required for enzymatic activities, binding of bilirubin at the secondary site, or immunological reactivity; steroid isomerase and glutathione-S-transferase activities are modulated in a parallel manner and may be mediated by the same region of the protein, and primary and secondary binding sites for bilirubin are distinct and independent, despite nicks introduced by proteolysis in ligandin's subunits, some of the fragments remain associated under non-denaturing conditions and the susceptibility of the two subunits to the proteases is different.     

Cleavage of growth hormone by rabbit liver plasmalemma enhances binding

Several studies have shown that proteolytic cleavage can enhance the biological activity of the growth hormone (GH) molecule. It seemed possible, therefore, that proteolytic modification of GH might be a normal function of GH-target tissues. Plasmalemma-enriched fractions isolated from rabbit liver were found to contain a proteinase(s) which cleaves the large disulfide loop of human and rat GH. The proteolytic activity was specific to plasmalemma-enriched fractions in that much lower activities were observed in microsomal-enriched fractions prepared from the same livers. The plasmalemmal proteinase(s) may be a trypsin-like enzyme because proteolytic activity was decreased by two serine proteinase inhibitors. Inhibition by unlabeled human GH of 125I-GH binding to receptors did not prevent cleavage of the tracer; therefore, hormone-receptor interaction was not required for cleavage of the GH molecule. In binding studies, cleaved GH associated more readily than did intact hormone with rabbit liver receptors. These studies suggest that plasmalemma-enriched fractions prepared from rabbit liver contain a proteinase which cleaves the GH molecule in a highly specific manner. Moreover, it is unlikely that inactivation of GH is the function of this limited proteolysis because cleaved hormone is bound preferentially by at least a subset of receptors in rabbit liver.     

CleavPredict: A Platform for Reasoning about Matrix Metalloproteinases Proteolytic Events

CleavPredict (http://cleavpredict.sanfordburnham.org) is a Web server for substrate cleavage prediction for matrix metalloproteinases (MMPs). It is intended as a computational platform aiding the scientific community in reasoning about proteolytic events. CleavPredict offers in silico prediction of cleavage sites specific for 11 human MMPs. The prediction method employs the MMP specific position weight matrices (PWMs) derived from statistical analysis of high-throughput phage display experimental results. To augment the substrate cleavage prediction process, CleavPredict provides information about the structural features of potential cleavage sites that influence proteolysis. These include: secondary structure, disordered regions, transmembrane domains, and solvent accessibility. The server also provides information about subcellular location, co-localization, and co-expression of proteinase and potential substrates, along with experimentally determined positions of single nucleotide polymorphism (SNP), and posttranslational modification (PTM) sites in substrates. All this information will provide the user with perspectives in reasoning about proteolytic events. CleavPredict is freely accessible, and there is no login required.     

Limited proteolysis of smooth muscle myosin light chain kinase

Myosin light chain kinase plays a central role in the regulation of smooth muscle contraction. The activity of this enzyme is controlled by protein-protein interaction (the Ca2+-dependent binding of calmodulin) and by phosphorylation catalyzed by cAMP-dependent protein kinase. The effects of these two regulatory mechanisms on the conformation of myosin light chain kinase and the locations of the phosphorylation sites, the calmodulin-binding site, and the active site have been probed by limited proteolysis. Phosphorylated and nonphosphorylated myosin light chain kinases were subjected to limited digestion by four proteases having different peptide bond specificities (trypsin, chymotrypsin, Staphylococcus aureus V8 protease, and thrombin), both in the presence and in the absence of bound calmodulin. The digests were compared in terms of gel electrophoretic pattern, distribution of phosphorylation sites, and Ca2+ dependence of kinase activity. A 24 500-dalton chymotryptic peptide containing both sites of phosphorylation was purified and tentatively identified as the amino-terminal peptide. The following conclusions can be drawn: neither phosphorylation nor calmodulin binding induces dramatic changes in the conformation of the kinase; the kinase contains two regions that are particularly susceptible to proteolytic cleavage, one located approximately 25 000 daltons from the amino terminus and the other near the center of the molecule; the two phosphorylation sites are located within 24 500 (probably 17 500) daltons of the amino terminus; the active site is located close to the center of the molecule; the calmodulin-binding site is located in the amino-terminal half of the molecule, between the sites of phosphorylation and the active site, and this region is very susceptible to cleavage by trypsin.     

Structural domains of human apolipoprotein B-100. Differential accessibility to limited proteolysis of B-100 in low density and very low density lipoproteins

The structural domains of human apolipoprotein B-100 in low density lipoproteins (LDL) and the conformational changes of B-100 that accompany the conversion of very low density lipoproteins (VLDL) to LDL were investigated by limited proteolysis with 12 endoproteases of various specificities, and their cleavage sites were determined. In B-100 of LDL, we identified two peptide regions that are highly susceptible to proteolytic cleavage. One region encompassed about 40 amino acids (residues 1280-1320, designated as the NH2-terminal region) and the other about 100 amino acids (residues 3180-3280, designated as the COOH-terminal region). IN LDL, the cleavage sites in both susceptible regions of B-100 were readily accessible to limited proteolysis; but in VLDL, only sites in the COOH-terminal region were readily accessible. Moreover, B-100 in VLDL appeared less degraded than B-100 in LDL by all enzymes used. Reduction of disulfide bonds of B-100 in both LDL and VLDL before digestion by Staphylococcus aureus V8 protease and clostripain exposed additional cleavage sites and increased the rate of B-100 degradation, suggesting that disulfide bonds probably exert conformational constraints. These results indicate the presence of three principal structural domains in B-100 of LDL that are relatively resistant to limited proteolysis. These three domains are connected by the two susceptible peptide regions. Our results also demonstrate differential accessibility of cleavage sites in B-100 of LDL and VLDL to limited proteolysis. This differential accessibility suggests that substantial changes in the conformation or environment of B-100 accompany the conversion of VLDL to LDL.     

Structure formation in short designed peptides probed by proteolytic cleavage

The formation of local structure, in short peptides has been probed by examining cleavage patterns and rates of proteolysis of designed sequences with a high tendency to form beta-hairpin structures. Three model sequences which bear fluorescence donor and acceptor groups have been investigated: [see text]. Fluorescence resonance energy transfer (FRET) provides a convenient probe for peptide cleavage. MALDI mass spectrometry has been used to probe sites of cleavage and CD spectroscopy to access the overall backbone conformation using analog sequences, which lack strongly absorbing donor and acceptor groups. The proteases trypsin, subtilisin, collagenase, elastase, proteinase K and thermolysin were used for proteolysis and the rates of cleavage determined. Peptide 3 is the most susceptible to cleavage by all the enzymes except thermolysin, which cleaves all three peptides at comparable rates. Peptides 1 and 2 are completely resistant to the action of trypsin, suggesting that beta-turn formation acts as a deterrent to proteolytic cleavage.     

Increased proteolytic resistance of ribonuclease A by protein engineering.

Although highly stable toward unfolding, native ribonuclease A is known to be cleaved by unspecific proteases in the flexible loop region near Ala20. With the aim to create a protease-resistant ribonuclease A, Ala20 was substituted for Pro by site-directed mutagenesis. The resulting mutant enzyme was nearly identical to the wild-type enzyme in the near-UV and far-UV circular dichroism spectra, in its activity to 2',3'-cCMP and in its thermodynamic stability. However, the proteolytic resistance to proteinase K and subtilisin Carlsberg was extremely increased. Pseudo-first-order rate constants of proteolysis, determined by densitometric analysis of the bands of intact protein in SDS-PAGE, decreased by two orders of magnitude. In contrast, the rate constant of proteolysis with elastase was similar to that of the wild-type enzyme. These differences can be explained by the analysis of the fragments occurring in proteolysis with elastase. Ser21-Ser22 was identified as the main primary cleavage site in the degradation of the mutant enzyme by elastase. Obviously, this bond is not cleavable by proteinase K or subtilisin Carlsberg. The results demonstrate the high potential of a single mutation in protein stabilization to proteolytic degradation.     

Proteinase 3C-mediated processing of VP1-2A of two hepatitis A virus strains: in vivo evidence for cleavage at amino acid position 273/274 of VP1.

Two prominent features distinguish hepatitis A virus (HAV) from other members of the picornavirus family. A C-terminally prolonged precursor of the structural protein VP1 is incorporated into assembly intermediates (e.g., the provirion), and a single proteinase is contained within the HAV polyprotein. Using an in vivo expression system, we show that proteolytic liberation of VP1 from its precursors P1-2A and VP1-2A is catalyzed by the virus-encoded proteinase 3Cpro. Among the proposed cleavage sites within VP1-2A, the Glu/Ser pair found at VP1 amino acid position 273/274 of most HAV strains is efficiently processed, whereas proteolysis of the Val/Ser site of the attenuated HM175 strain is protracted. Two mutations within VP1-2A (Lys[297]Arg and Ser[330]Asn) had no effect on 3Cpro-mediated cleavage at this site. Additional sites in this region of VP1-2A can also be utilized as substrates by the proteinase, yet less efficiently, and might give rise to smaller and larger VP1 polypeptides also detected in HAV-infected cells.     

Inactivation of human plasma serine proteinase inhibitors (serpins) by limited proteolysis of the reactive site loop with snake venom and bacterial metalloproteinases

Human plasma serine proteinase inhibitors (serpins) gradually lost activity when incubated with catalytic amounts of snake venom or bacterial metalloproteinases. Electrophoretic analyses indicated that antithrombin III, C1-inhibitor, and alpha 2-antiplasmin had been converted by limited proteolysis into modified species which retained inhibitory activity. Further proteolytic attack resulted in the formation of inactivated inhibitors; alpha 1-proteinase inhibitor (alpha 1-antitrypsin) and alpha 1-antichymotrypsin were also enzymatically inactivated, but active intermediates were not detected. Sequence analyses indicated that the initial, noninactivating cleavage occurred in the amino-terminal region of the inhibitors. Inactivation resulted in all cases from the limited proteolysis of a single bond near, but not at, the reactive site bond in the carboxy-terminal region of the inhibitors. The results indicate that the serpins have two regions which are susceptible to limited proteolysis--one near the amino-terminal end and another in the exposed reactive site loop of the inhibitor.     

Herpesvirus proteinase: site-directed mutagenesis used to study maturational, release, and inactivation cleavage sites of precursor and to identify a possible catalytic site serine and histidine.

The cytomegalovirus maturational proteinase is synthesized as a precursor that undergoes at least three processing cleavages. Two of these were predicted to be at highly conserved consensus sequences--one near the carboxyl end of the precursor, called the maturational (M) site, and the other near the middle of the precursor, called the release (R) site. A third less-well-conserved cleavage site, called the inactivation (I) site, was also identified near the middle of the human cytomegalovirus 28-kDa assemblin homolog. We have used site-directed mutagenesis to verify all three predicted sequences in the simian cytomegalovirus proteinase, and have shown that the proteinase precursor is active without cleavage at these sites. We have also shown that the P4 tyrosine and the P2 lysine of the R site were more sensitive to substitution than the other R- and M-site residues tested: substitution of alanine for P4 tyrosine at the R site severely reduced cleavage at that site but not at the M site, and substitution of asparagine for lysine at P2 of the R site reduced M-site cleavage and nearly eliminated I-site cleavage but had little effect on R-site cleavage. With the exception of P1' serine, all R-site mutations hindered I-site cleavage, suggesting a role for the carboxyl end of assemblin in I-site cleavage. Pulse-chase radiolabeling and site-directed mutagenesis indicated that assemblin is metabolically unstable and is degraded by cleavage at its I site. Fourteen amino acid substitutions were also made in assemblin, the enzymatic amino half of the proteinase precursor. Among those tested, only 2 amino acids were identified as essential for activity: the single absolutely conserved serine and one of the two absolutely conserved histidines. When the highly conserved glutamic acid (Glu22) was substituted, the proteinase was able to cleave at the M and I sites but not at the R site, suggesting either a direct (e.g., substrate recognition) or indirect (e.g., protein conformation) role for this residue in determining substrate specificity.     

Different roles of the two disulfide bonds of the cysteine proteinase inhibitor, chicken cystatin, for the conformation of the active protein.

The Cys-71-Cys-81 disulfide bond of the cysteine proteinase inhibitor, chicken cystatin, was specifically reduced by thioredoxin or low concentrations of dithiothreitol. This cleavage, followed by S-carbamoylmethylation, induced a conformational change of the protein, as evidenced by changes in isoelectric point and circular dichroism spectra and by an increased susceptibility to digestion by nontarget proteinases. The proteinase binding ability and the immunological properties of the inhibitor, however, were not detectably altered, indicating that the conformational change was limited to the region around the disrupted bond. In contrast, reduction of both disulfide bonds of cystatin by higher concentrations of dithiothreitol and subsequent alkylation led to the slow conversion of the inhibitor into two forms lacking proteinase binding ability, indicative of more extensive conformational changes. Together, these results suggest that the less accessible Cys-95-Cys-115 disulfide bond of chicken cystatin, but not the more accessible Cys-71-Cys-81 bond, is of importance for maintaining the conformation of the inhibitor required for binding of target proteinases.     

Characterization of the single peptide generated from the amino-terminus end of alpha- and beta-hemoglobin chains by the Ca2+-dependent neutral proteinase

Human erythrocyte Ca2+-dependent neutral proteinase catalyzes a limited proteolysis of isolated globin chains. The rate of hydrolysis is very rapid using heme-deprived alpha- or beta-globin chains and is reduced to one-fifth with their corresponding native forms. In both cases, the proteinase specifically cleaves a single peptide bond, this resulting in the removal from the amino-terminus end of an octapeptide in beta-globin and of an undecapeptide in alpha-globin. Both peptides have been isolated, their amino acid composition has been characterized and the susceptible site of cleavage has been identified. Hemoglobin variants show a different rate of digestion as compared to that of normal chains. The alpha-Hasharon [alpha 47(CE5) Asp----His] undergoes rapid digestion, while the beta-G San Josè chain [beta 7(A4) Glu----Gly], which carries the mutation near the site of cleavage, reveals a high degree of resistance to proteolytic degradation by the neutral proteinase.     

Probing the conformational state of a truncated staphylococcal nuclease R using time of flight mass spectrometry with limited proteolysis.

The conformational state of C-terminally truncated staphylococcal nuclease R (SNR135), with and without bound ligands, has been studied by performing limited proteolysis with a specific endoproteinase Glu-C followed by electrophoresis and mass spectrometry. Comparison of the accessibility of the cleavage sites shows that the C-terminal truncation of 14 amino-acid residues causes significant unfolding of the C-terminal part of alpha helix 1 and the center of alpha helix 2, but there is little effect on other regions of the nuclease, in particular the N-terminal subdomain, which includes the active site of the nuclease. The truncation also makes the overall conformation of the nuclease more loose and flexible. Binding of ligands makes helices 1 and 2 more resistant to protease Glu-C attack and converts the partially unfolded state to a native-like state, although the conformational stability of the SNR135 complex is still much lower than that of the full-length enzyme. The results suggest that the amino-acid residues around the active site in the truncated nuclease are arranged in a similar topology to those in the full-length nuclease. The study shows that there is a clear-cut correlation between protease susceptibility and conformational stability of the protein, and the initial proteolytic events are the most critical for evaluating the conformational features of the protein. This study demonstrates how mass spectrometry can be combined with limited proteolysis to observe conformational changes induced by ligand binding.     

Expression of potyviral polyproteins in transgenic plants reveals three proteolytic activities required for complete processing.

All proteins encoded by the plant potyvirus, tobacco etch virus (TEV), arise by proteolytic processing of a single polyprotein. Two virus-encoded proteinases (NIa and HC-Pro) that catalyze most of the proteolytic events have been characterized previously. The two proteins that are derived from the N-terminal 87 kd region of the viral polyprotein are a 35 kd protein and HC-Pro (52 kd). It is demonstrated in this study that a third proteolytic activity is required to process the junction between these proteins. Proteolysis at the HC-Pro N terminus to separate these proteins occurred poorly, if at all, after in vitro synthesis of a 97 kd polyprotein, whereas cleavage of the HC-Pro C terminus occurred efficiently by an autoprocessing mechanism. Synthesis of the same polyprotein in transgenic tobacco plants, however, resulted in complete and accurate proteolysis at both termini of HC-Pro. A point mutation affecting an amino acid residue essential for the proteolytic activity of HC-Pro had no effect on N-terminal processing. Expression in transgenic plants of a construct with a large deletion in the 35 kd protein coding region resulted in partial inhibition of HC-Pro N-terminal cleavage, suggesting that the 35 kd protein may affect the proteolytic event but not in a catalytic role. We speculate that this cleavage event is catalyzed by either a cryptic potyviral proteinase that requires a host factor or subcellular environment for activation, or possibly a host proteinase.     

Mutagenesis of the yellow fever virus NS2B protein: effects on proteolytic processing, NS2B-NS3 complex formation, and viral replication.

To study the role of specific regions of the yellow fever virus NS2B protein in proteolytic processing and association with the NS3 proteinase domain, a series of mutations were created in the hydrophobic regions and in a central conserved hydrophilic region proposed as a domain important for NS2B function. The effects of these mutations on cis cleavage at the 2B/3 cleavage site and on processing at other consensus cleavage sites for the NS3 proteinase in the nonstructural region were then characterized by cell-free translation and transient expression in BHK cells. Association between NS2B and the NS3 proteinase domain and the effects of mutations on complex formation were investigated by nondenaturing immunoprecipitation of these proteins expressed in infected cells, by cell-free translation, or by recombinant vaccinia viruses. Mutations within the hydrophobic regions had subtle effects on proteolytic processing, whereas mutations within the conserved domain dramatically reduced cleavage efficiency or abolished all cleavages. The conserved domain of NS2B is also implicated in formation of an NS2B-NS3 complex on the basis of the ability of mutations in this region to eliminate both association of these two proteins and trans-cleavage activity. In addition, mutations which either eliminated proteolytic processing or had no apparent effect on processing were found to abolish recovery of infectious virus following RNA transfection. These results suggest that the conserved region of NS2B is a domain essential for the function of the NS3 proteinase. Hydrophobic regions of NS2B whose structural integrity may not be essential for proteolytic processing may have additional functions during viral replication.