Calcium binding of transglutaminases: a 43Ca NMR study combined with surface polarity analysis.

Transglutaminases (TGases) form cross-links between glutamine and lysine side-chains of polypeptides in a Ca2+-dependent reaction. The structural basis of the Ca2+-effect is poorly defined. 43Ca NMR, surface polarity analysis combined with multiple sequence alignment and the construction of a new homology model of human tissue transglutaminase (tTGase) were used to obtain structural information about Ca2+ binding properties of factor XIII-A2, tTGase and TGase 3 (each of human origin). 43Ca NMR provided higher average dissociation constants titrating on a wide Ca2+-concentration scale than previous studies with equilibrium dialysis performed in shorter ranges. These results suggest the existence of low affinity Ca2+ binding sites on both FXIII-A and tTGase in addition to high affinity ones in accordance with our surface polarity analysis identifying high numbers of negatively charged clusters. Upon increasing the salt concentration or activating with thrombin, FXIII-A2 partially lost its original Ca2+ affinity; the NMR data suggested different mechanisms for the two activation processes. The NMR provided structural evidence of GTP-induced conformational changes on the tTGase molecule diminishing all of its Ca2+ binding sites. NMR data on the Ca2+ binding properties of the TGase 3 are presented here; it binds Ca2+ the most tightly, which is weakened after its proteolytic activation. The investigated TGases seem to have very symmetric Ca2+ binding sites and no EF-hand motifs.

Identification of essential active-site residues in the cyanogenic beta-glucosidase (linamarase) from cassava (Manihot esculenta Crantz) by site-directed mutagenesis.

The coding sequence of the mature cyanogenic beta-glucosidase (beta-glucoside glucohydrolase, EC 3.2.1.21; linamarase) was cloned into the vector pYX243 modified to contain the SUC2 yeast secretion signal sequence and expressed in Saccharomyces cerevisiae. The recombinant enzyme is active, glycosylated and showed similar stability to the plant protein. Michaelis constants for hydrolysis of the natural substrate, linamarin (K(m)=1.06 mM) and the synthetic p-nitrophenyl beta-D-glucopyranoside (PNP-Glc; K(m)=0.36 mM), as well as apparent pK(a) values of the free enzyme and the enzyme-substrate complexes (pK(E)(1)=4.4-4.8, pK(E)(2)=6.7-7.2, pK(ES)(1)=3.9-4.4, pK(ES)(2)=8.3) were very similar to those of the plant enzyme. Site-directed mutagenesis was carried out to study the function of active-site residues based on a homology model generated for the enzyme using the MODELLER program. Changing Glu-413 to Gly destroyed enzyme activity, consistent with it being the catalytic nucleophile. The Gln-339Glu mutation also abolished activity, confirming a function in positioning the catalytic diad. The Ala-201Val mutation shifted the pK(a) of the acid/base catalyst Glu-198 from 7.22 to 7.44, reflecting a change in its hydrophobic environment. A Phe-269Asn change increased K(m) for linamarin hydrolysis 16-fold (16.1 mM) and that for PNP-Glc only 2.5-fold (0.84 mM), demonstrating that Phe-269 contributes to the cyanogenic specificity of the cassava beta-glucosidase.

Carbon source regulation of beta-galactosidase biosynthesis in Penicillium chrysogenum.

Growth and beta-galactosidase activity of the penicillin producer industrial Penicillium chrysogenum NCAIM 00237 strain were examined using different carbon sources. Good growth was observed using glucose, sucrose, glycerol and galactose, while growth on lactose was substantially slower. beta-Galactosidase activity was high on lactose and very low on all the other carbon sources tested. In glucose grown cultures after exhaustion of glucose as repressing carbon source a derepressed low level of the enzyme was observed. cAMP concentration in lactose grown cultures was relatively high, in glucose grown cultures was low. Caffeine substantially decreased glucose consumption and growth but did not increase beta-galactosidase activity and did not prevent glucose repression which rules out the involvement of cAMP in the regulation of beta-galactosidase biosynthesis in Penicillium chrysogenum.

Genomic organization and structure of alpha-hydroxynitrile lyase in cassava (Manihot esculenta Crantz).

Two clones with homology to the alpha-hydroxynitrile lyase (HNL) cDNA clone, MeHNL10, were isolatedfrom a lambdaEMBL3 cassava (Manihot esculenta Crantz) genomic library. Analysis of the sequences showed that both genomic clones contain HNL genes (MeHNL4, MeHNL24) which are interrupted by two introns. RT-PCR analysis of MeHNL4 shows that it is expressed at high levels in seedling roots and at lower levels in cotyledons and young leaves. The deduced amino acid sequences of MeHNL4, MeHNL10, and MeHNL24 show high sequence identity and homology to the HNL from Hevea brasiliensis whose tertiary structure has been solved at 1.9-A resolution by X-ray crystallography. This high homology allowed the construction of model structures for all of the cassava proteins using the MODELLER program. Homology modeling indicates that the short variable exon 2 encodes the "cap" region which is thought to influence the substrate specificity of the protein. Two hybrid proteins were modeled using the core alpha/beta domain of MeHNL10 and the cap region of either the Hevea HNL or a structurally related Zea protein of unknown function. This analysis suggests that changes in the active site can be engineered by swapping exons.

Homology modelling and molecular dynamics aided analysis of ligand complexes demonstrates functional properties of lipid-transfer proteins encoded by the barley low-temperature-inducible gene family, blt4.

The homology modelling technique was used to predict the tertiary structures of three members of the low-temperature-inducible barley vegetative shoot epidermal lipid-transfer protein (LTP) family, BLT4, on the basis of the X-ray crystallographically determined three-dimensional structure of a maize seedling LTP. Differences between the maize LTP and the BLT4 family include amino acid substitutions around the entrance and inside the predicted hydrophobic binding tunnels of these proteins. Because of the deletion of the loop region corresponding to Val60-Gly62 of the maize LTP from all three BLT4 LTPs, their internal hydrophobic tunnels are longer. Molecular dynamics modelling shows that BLT4.9 can accommodate hexadecanoic acid in its binding tunnel in similar conformation to the maize LTP. However, modelled cis,cis-9, 12-octadecandienoic acid had a more favourable interaction with the BLT4.9 LTP than with the maize protein. Di-cis,cis-9, 12-octadecandienoyl phosphatidylglycerol and di-cis,cis-9, 12-octadecandienoyl phosphatidylcholine were modelled in the BLT4.9 structure with the fatty acyl group at position 1 embedded in the binding tunnel and the group at position 2 located on the solvent accessible surface of the protein. The results of the modelling suggest that the phospholipid headgroup can form hydrogen and salt bridges with polar and charged residues outside the binding tunnel and the exposed hydrocarbon chain interacts with hydrophobic amino acids on the surface. These results are consistent with the proposal that BLT4 LTPs have a lipid-transfer function associated with frost acclimation in barley.

Co-purification from Escherichia coli of a plant beta-glucosidase-glutathione S-transferase fusion protein and the bacterial chaperonin GroEL.

The coding sequence of the mature cyanogenic beta-D-glucosidase (beta-D-glucoside glucohydrolase, EC 3.2.1.21) (linamarase) of Manihot esculenta Crantz (cassava) was cloned into the vector pGEX-2T and expressed in Escherichia coli. The bacterial chaperonin GroEL [Braig, Otwinowski, Hedge, Boisvert, Joachimiak, Horwich and Sigler (1994) Nature (London) 371, 578-586] was found to be tightly associated with the fusion protein and co-purified with it. In the presence of excess MgATP, release and folding of the fusion beta-glucosidase were demonstrated by a fast increase in both linamarase and p-nitrophenyl-beta-D-glucopyranosidase activity at a low protein concentration. A slow endogenous folding process was also detected by activity measurements. Michaelis constants (Km) and the ratio between the maximal velocities and efficiency constants (Vmax., Vmax./Km) for the hydrolysis of the natural substrate, linamarin, and p-nitrophenyl beta-D-glucopyranoside (PNP-Glc) by the recombinant protein were found to be almost identical with those of the native glycosylated plant enzyme [Keresztessy, Kiss and Hughes (1994) Arch. Biochem. Biophys. 314, 142-152]. Molecular dissociation constants for the free enzyme (pK(E)1, pK(E)2) obtained with linamarin and PNP-Glc, and the enzyme substrate complexes (pK(ES)1, pK(ES)2) were also in accordance with that of the original protein. The reactive substrate analogue N-bromoacetyl beta-D-glucosylamine inactivated the fusion enzyme according to pseudo-first-order kinetics with first-order rate constant (k1=0.007 min-1) and apparent inhibition constants (k1=20 mM) comparable with those of the plant protein [Keresztessy, Kiss and Hughes (1994) Arch. Biochem. Biophys. 315, 323-330]. In comparison with the native glycosylated plant protein, the recombinant protein was, however, found to be extremely sensitive to proteolysis and misfolding.

Investigation of the active site of the cyanogenic beta-D-glucosidase (linamarase) from Manihot esculenta Crantz (cassava). II. Identification of Glu-198 as an active site carboxylate group with acid catalytic function.

The broad-specificity cyanogenic beta-D-glucosidase (beta-D-glucoside glucohydrolase, EC 3.2.1.21) (linamarase) from Manihot esculenta Crantz (cassava) was irreversibly inactivated by N-bromoacetyl-beta-D-glucopyranosylamine according to pseudo-first-order kinetics with a second-order efficiency constant (ki/Ki = 0.1 min-1 M-1) identical for p-nitrophenyl-beta-D-glucopyranosidase, p-nitrophenyl-beta-D-galactopyranosidase, and linamarase activities of the enzyme. The competitive inhibitor p-nitrothiophenyl-beta-D-glucopyranoside protected the enzyme from inactivation. pH dependence of the pseudo-first-order rate constant of inactivation revealed the involvement of an amino acid side chain in the inactivation process with pKa 7.0, which is very similar to that of the acid catalyst group of the enzyme (pKE2 = 7.2). The involved amino acid, which has to be ionized for the inactivation, was identified as Glu-198 using 14C-labeled inactivator to label the enzyme, cleaving the labeled protein into peptides and then purifying and sequencing the labeled peptide. This residue is highly conserved in the homologous family A beta-glucosidases and family A1-A5 cellulases and lies in a consensus Asn-Glu-Pro motif occurring in all of these enzymes.

Investigation of the active site of the cyanogenic beta-D-glucosidase (linamarase) from Manihot esculenta Crantz (cassava). I. Evidence for an essential carboxylate and a reactive histidine residue in a single catalytic center.

The broad-specificity cyanogenic beta-D-glucosidase (beta-D-glucoside glucohydrolase, EC 3.2.1.21) (linamarase) from Manihot esculenta Crantz (cassava) was kinetically characterized in mixed substrate systems and with the transition-state analogue glucono(1-5)lactone and a series of 1-thio substrate analogues. The results indicate a common catalytic and a common sugar binding site in the enzyme for all of the investigated substrates. Kinetic parameters of the hydrolysis of linamarin and p-nitrophenyl beta-D-glucopyranoside were determined over the pH range 3.5-9.0. The pH-dependence curves gave apparent pK values of 4.5 (4.6) and 7.1 (7.3) for the free enzyme, while values of 4.5 (3.7) and 9.3 were obtained for the enzyme-substrate complexes, using either linamarin or p-nitrophenyl beta-D-glucopyranoside as the substrate. Kinetic analysis of the modification indicated that one molecule of water-soluble carbodiimide or Woodward's reagent K is required to bind to the enzyme for inactivation. The enzyme was protected against inactivation by the competitive inhibitors p-nitrothiophenyl beta-D-glucopyranoside, beta-D-glucopyranosylamine, and glucono(1-5)lactone. Spectrophotometric analysis at 340 nm showed that from the three carboxylate groups modified by Woodward's reagent K essentially one was protected by p-nitrothiophenyl beta-D-glucopyranoside. During modification Vmax decreased to 30% of that of the unmodified enzyme and Km remained unchanged. The pH dependence of inactivation showed the involvement of a group with a pK value of 4.6, indicating the modification of a carboxyl residue essential for activity. Treatment of the enzyme with the histidine-group-specific reagent diethylpyrocarbonate resulted in 80% loss of enzyme activity, in biphasic kinetics. A treatment with 0.5 M hydroxylamine at pH 7.0 regenerated 92% of the original enzyme activity. The presence of the competitive inhibitor beta-D-glucopyranosylamine protected the enzyme against inactivation, preventing the modification of one histidine residue. Statistical analysis of the residual fractional activity against the number of modified residues indicated that the modification of one histidine is responsible for 40-50% of the inactivation. The pH dependence of the inactivation gave a pK value of 7.0 for the histidine group upon which the activity depends. During modification, Vmax decreased to 30% and Km decreased to 50% of the original values.