The X-ray crystal structure of the Trichoderma reesei family 12 endoglucanase 3, Cel12A, at 1.9 A resolution.

We present the three-dimensional structure of Trichoderma reesei endoglucanase 3 (Cel12A), a small, 218 amino acid residue (24.5 kDa), neutral pI, glycoside hydrolase family 12 cellulase that lacks a cellulose-binding module. The structure has been determined using X-ray crystallography and refined to 1.9 A resolution. The asymmetric unit consists of six non-crystallographic symmetry-related molecules that were exploited to improve initial multiple isomorphous replacement phasing, and subsequent structure refinement. The enzyme contains one disulfide bridge and is glycosylated at Asp164 by a single N-acetyl glucosamine residue. The protein has the expected fold for a glycoside hydrolase clan-C family 12 enzyme. It contains two beta-sheets, of six and nine strands, packed on top of one another, and one alpha-helix. The concave surface of the nine-stranded beta-sheet forms a large substrate-binding groove in which the active-site residues are located. In the active site, we find a carboxylic acid trio, similar to that of glycoside hydrolase families 7 and 16. The strictly conserved Asp99 hydrogen bonds to the nucleophile, the invariant Glu116. The binding crevice is lined with both aromatic and polar amino acid side-chains which may play a role in substrate binding. The structure of the fungal family 12 enzyme presented here allows a complete structural characterization of the glycoside hydrolase-C clan.

The role of lysine-234 in beta-lactamase catalysis probed by site-directed mutagenesis.

Lys-234 has been postulated to participate in beta-lactamase catalysis by acting as an electrostatic anchor for the C3 carboxylate of penicillins [Herzberg, O., & Moult, J. (1987) Science 236, 694-701]. To test this hypothesis, site-directed mutagenesis was used to convert the Lys-234 in Bacillus licheniformis beta-lactamase into Glu-234 or Ala-234. The wild-type, Glu-234, and Ala-234 beta-lactamases have been expressed in Bacillus subtilis and purified to homogeneity. The wild-type, K234E, and K234A enzymes have virtually identical circular dichroism and fluorescence spectra, similar thermal stabilities at neutral pH, and the same susceptibilities to proteolysis, indicating the lack of significant structural perturbation caused by the mutation. At acidic and basic pH the mutant enzymes have the same native circular dichroism as the wild-type enzyme but the thermal stability is significantly different. The mutations cause perturbations of the pK values of the ionizing groups responsible for the pH dependence of the catalytic reaction in both the free enzyme and the E.S complex. As expected, conversion of Lys-234 to Ala or Glu decreased substrate binding (Km) by 1-2 orders of magnitude for several penicillin and cephalosporin substrates at neutral and higher pH. However, at low pH, Km is essentially the same for the K234E and K234A enzymes as for the wild-type enzyme. Furthermore, decreases of 2-3 orders of magnitude in kcat were also observed, indicating substantial effects on the transition-state binding, as well as on ground-state binding. Surprisingly, changing the C3 carboxylate of phenoxymethylpenicillin to a hydroxymethyl group led to little difference in kinetic properties with the K234E or K234A enzyme. The results of this investigation indicate the Lys-234 is an important active-site residue involved in both ground-state and transition-state binding.

Protein engineering of disulfide bonds in subtilisin BPN'.

Five single-disulfide mutants were studied in subtilisin BPN', a cysteine-free, secreted serine protease from Bacillus amyloliquefaciens. The disulfides were engineered between residues 26-232, 29-119, 36-210, 41-80, and 148-243. These bonds connected a variety of secondary structural elements, located in buried or exposed positions at least 10 A from the catalytic Ser-221, and linked residues that were separated by 39 up to 206 amino acids. All disulfide bonds formed in the enzyme when the expressed protein was secreted from Bacillus subtilis, and the disulfides had only minor effects on the enzyme kinetics. Although these disulfide bonds varied by over 50-fold in their equilibrium constants for reduction with dithiothreitol, there was no correlation between the strength of the disulfide bond and the stability it imparted to the enzyme to irreversible inactivation. In some cases, the disulfide-bonded protein was stabilized greatly relative to its reduced counterpart. However, no disulfide mutant was substantially more stable than wild-type subtilisin BPN'. Some of these results can be rationalized by destabilizing effects of the cysteine mutations that disrupt interactions present in the folded enzyme structure. It is also possible that the rate of irreversible inactivation depends upon the kinetics and not the thermodynamics of unfolding and so the entropically stabilizing effect expected from a disulfide bond may not apply.

The design and production of semisynthetic ribonucleases with increased thermostability by incorporation of S-peptide analogues with enhanced helical stability.

Recent work has shown that, with synthetic analogues of C-peptide (residues 1-13 of ribonuclease A), the stability of the peptide helix in H2O depends strongly on the charge on the N-terminal residue. We have asked whether, in semisynthetic ribonuclease S reconstituted from S-protein plus an analogue of S-peptide (1-15), the stability of the peptide helix is correlated with the Tm of the reconstituted ribonuclease S. Six peptides have been made, which contain Glu9----Leu, a blocked alpha-COO- group (-CONH2), and either Gln11 or Glu11. The N-terminal residue has been varied; its charge varies from +2 (Lys) to -1 (succinyl-Ala). We have measured the stability of the peptide helix, the affinity of the peptide for S-protein (by C.D. titration), and the thermal stability of the reconstituted ribonuclease S. All six peptide analogues show strongly enhanced helix formation compared to either S-peptide (1-15) or (1-19), and the helix content increases as the charge on the N-terminal residue changes from +2 to -1. All six peptides show increased affinity for S-protein compared to S-peptide (1-19), and all six reconstituted ribonucleases S show an increase in Tm compared to the protein with S-peptide (1-19). The Tm increases as the charge on residue 1 changes from +2 to -1. The largest increment in Tm is 6 degrees. The results suggest that the stability of a protein can be increased by enhancing the stability of its secondary structure.

Effects of sulphate and urea on the stability and reversible unfolding of beta-lactamase from Staphylococcus aureus. Implications for the folding pathway of beta-lactamase.

The reversible denaturation by urea of beta-lactamase from Staphylococcus aureus was followed in the presence and absence of ammonium sulphate by circular dichroism studies, difference absorption spectroscopy and measurement of enzyme activity. The multiple unfolding and refolding transitions demonstrate the existence of a thermodynamically stable state of intermediate conformation in equilibrium with the native (N) and fully unfolded (U) states. Its physical properties show that it is identical to the state H found on denaturation by guanidinium chloride. State H is 10.1 (+/-1.5) kJ mol-1 less stable than the native state and 10.1 (+/-1.6) kJ mol-1 more stable than the unfolded state. Ammonium sulphate shifts both the N in equilibrium H and H in equilibrium U transitions to concentrations of urea higher by 5.3 M per mole of sulphate. It has markedly different effects on the thermodynamic stabilities of states N and H, making delta G'N-H, O and delta G'H-U, O more negative by 41 kJ mol and 20 kJ mole, respectively, per mole of ammonium sulphate. The change in equilibrium constant for the N-H transition is reflected almost exclusively in a dramatic change of the unfolding rate constant, which is decreased by a factor of 10(11) on addition of 1.4 M-sulphate. The presence of the substrate benzyl penicillin has little effect on the equilibria or kinetics of the N-H transition. The results are discussed in terms of the nature of the N-H transition and of the ordering of intermediate states on the folding pathway.

The relative effectiveness of guanidinium and some biguanide salts as denaturants. Assessment against penicillinase.

Penicillinase (penicillin amido-beta-lactamhydrolase, EC 3.5.2.6) has been used as a model for quantitating the effectiveness of several guanidine-derived denaturants. It was chosen because the mechanism of its denaturation by urea and guanidinium chloride has been worked out in detail, because it has a low thermodynamic stability bringing it in the range of weak denaturants, and because its denaturation can readily be followed by enzyme activity as well as by spectroscopic probes. Contrary to previous reports, biguanide HCl is found to be no more effective than guanidinium chloride. The denaturant effectiveness of the various compounds studied is found to increase in the order: guanidinium chloride identical to biguanide HCl less than propylbiguanide HCl less than hexylbiguanide HCl much less than n-decylbiguanide HCl. n-Decylbiguanide HCl is a particularly powerful denaturant, unfolding penicillinase at a concentration of less than 0.015 M.

1H-NMR studies on nucleotide binding to the sarcoplasmic reticulum Ca2+ ATPase. Determination of the conformations of bound nucleotides by the measurement of proton-proton transferred nuclear Overhauser enhancements.

The glycosidic bond torsion angles and the conformations of the ribose of Mg2+ATP, Mg2+ADP and Mg2+AdoPP[NH]P (magnesium adenosine 5'-[beta, gamma-imido]triphosphate) bound to Ca2+ATPase, both native and modified with fluorescein isothiocyanate (FITC), in intact sarcoplasmic reticulum have been determined by the measurement of proton-proton transferred nuclear Overhauser enhancements by 1H-NMR spectroscopy. This method shows clearly the existence of a low-affinity ATP binding site after modification of the high-affinity site with FITC. For all three nucleotides bound to both the high-affinity (catalytic) site and the low-affinity site, we find that the conformation about the glycosidic bond is anti, the conformation of the ribose 3'-endo of the N type and the conformation about the ribose C4'-C5' bond either gauche-trans or trans-gauche. The values for the glycosidic bond torsion angles chi (O4'-C1'-N9-C4) for Mg2+ATP, Mg2+ADP and Mg2+AdoPP[NH]P bound to the low-affinity site of FITC-modified Ca2+ATPase are approximately equal to 270 degrees, approximately equal to 260 degrees and approximately equal to 240 degrees respectively. In the case of the nucleotides bound to the high-affinity (catalytic) site of native Ca2+ATPase, chi lies in the range 240-280 degrees.

Identification of a labelled peptide after stoicheiometric reaction of fluorescein isothiocyanate with the Ca2+ -dependent adenosine triphosphatase of sarcoplasmic reticulum.

Incorporation of 4.5 nmol fluorescein isothiocyanate/mg rabbit sarcoplasmic reticulum, or of 7.4 nmol/mg purified ATPase, was sufficient to inhibit the activity completely. These results are not consistent with the suggestion (Pick, U. and Karlish, S.J.D. (1980) Biochim. Biophys. Acta 626, 255-261) that 2 mol ATPase were inhibited by each mole of reagent incorporated. A single labelled peptide was purified from the inhibited ATPase and it was shown that Lys 3/190, 10 residues from the N-terminus of tryptic fragment B, was the reactive lysine residue. This site is close to a potential nucleotide-binding fold in the ATPase sequence. A similar peptide showing only 2 conservative replacements was isolated from the sarcoplasmic reticulum of the lobster.