Effects of glycosylation of the residue at position 14 in ovine angiotensinogen on the human renin reaction

A mutant angiotensinogen, S14N, in which Ser14 of ovine angiotensinogen was replaced by Asn to form a N-glycosylation site, was produced in CHO cells. The molecular weight was about 3,000 larger than that of wild-type ovine angiotensinogen, indicating that S14N angiotensinogen was glycosylated at Asn14. In the reaction with human renin, the km of mutant angiotensinogen was 3 times increased, but the Vmax was not affected by the mutation.     

N-glycosylation is requisite for the enzyme activity and Golgi retention of N-acetylglucosaminyltransferase III

UDP-N-acetylglucosamine: beta-D-mannoside beta-1,4N-acetylglucosaminyltransferase III (GnT-III, EC 2.4.1.144) is a glycoprotein involved in the biosynthesis of N-linked oligosaccharides. Rat GnT-III contains three potential N-glycosylation sites, which have been predicted to be Asn243, Asn261, and Asn399. To study the roles of N-glycosylation in the GnT-III function, rat GnT-III was expressed in COS-1 cells under tunicamycin or castanospermine treatment. The tunicamycin-treated GnT-III, which was not N-glycosylated, had almost no activity. The castanospermine-treated GnT-III was not localized in the Golgi, but glucosylation did not affect its activity. To clarify the role of individual N-glycosylations, we obtained a series of mutant cDNAs in which some or all of the potential glycosylation sites were eliminated by site-directed mutagenesis, and expressed them in COS-1 cells. All the mutants exhibited lower enzyme activity than the wild-type, but deglycosylation at individual sites had different effects on the enzyme activity. The deglycosylation at Asn243 or Asn261 was more effective on the activity than that at Asn399. The enzyme activity decreased as the number of glycosylation sites decreased. The null glycosylation mutant had no activity, corresponding to the case of tunicamycin-treated wild-type GnT-III. Kinetic analysis revealed that the deglycosylation at Asn243 or Asn261 resulted in slightly lower affinity for the donor substrate, but the other mutation did not significantly change the K(m) value for either the donor or acceptor. None of the mutant GnT-IIIs showed perinuclear localization or Golgi retention, that was observed for the wild-type protein. This is the first demonstration that the glycosyltransferase localized in the Golgi apparatus requires N-glycosylation for its activity and retention.     

Effects of substitution of Val for Leu11 of ovine angiotensinogen on renin activity

A mutant angiotensinogen, L11V, in which Val11 was substituted for Leu11 of ovine angiotensinogen was prepared to have the same scissile peptide bond as the human one. The mutation didn't vary Km and kcat of human renin for the ovine substrate, but decreased those of rat renin to one half and one fortieth, respectively. Distances between the P1' subsite of angiotensinogens and the 224th (human renin numbering) residue in the S1' subsite of renins were estimated by molecular modelings. The marked decrease in kcat of rat renin for L11V could be attributed to the elongated distance between Val11 of L11V and Val221 of rat renin. It was also suggested that the distance is the reason why the human substrate cannot be cleaved by heterologous renins.     

Site-directed mutations of the 4Fe-ferredoxin from the hyperthermophilic archaeon Pyrococcus furiosus: role of the cluster-coordinating aspartate in physiological electron transfer reactions

Ferredoxin from the hyperthermophilic archaeon Pyrococcus furiosus is a monomeric protein (7.5 kDa) that contains a single [4Fe-4S]1+, 2+ cluster. The protein is unusual in that its cluster is coordinated by three Cys and one Asp residue, rather than by the typical four Cys residues. Site-directed mutagenesis has been used to obtain mutant forms in which the cluster-coordinating Asp was replaced by Cys (D14C) and also by Ser (D14S), together with a third mutant (A1K) which contained N-Met-Lys at the N-terminus instead of N-Ala. Analyses using UV-visible absorption, far-UV circular dichroism, and EPR spectroscopy showed that there were no gross structural differences between the native and the three mutant forms and that they each contained a [4Fe-4S] cluster. The reduction potentials, determined by direct electrochemistry (at 23 degrees C, pH 8.0), of the D14S, D14C, and A1K mutants were -490, -422, and -382 mV, respectively, which compare with values of -375 mV for native [4Fe-4S]-containing ferredoxin and -160 mV for the [3Fe-4S]-containing form. The native, D14C, and A1K proteins functioned as electron acceptors in vitroat 80 degrees C for pyruvate ferredoxin oxidoreductase (POR) and aldehyde ferredoxin oxidoreductase (AOR) from P. furiosus using pyruvate and crotonaldehyde as substrates, respectively. The calculated kcat/Km values were similar for the three proteins when ferredoxin reduction was measured either directly by visible absorption or indirectly by coupling ferredoxin reoxidation to the reduction of metronidazole. In contrast, using the D14S mutant and the 3Fe-form of the native ferredoxin as electron acceptors, the activity with AOR was virtually undetectable, and with POR the calculated kcat/Km values were at least 3-fold lower than those obtained with the native (4Fe-), D14C, and A1K proteins. The ability of this 4Fe-ferredoxin to accept electrons from two oxidoreductases of the same organism is therefore not absolutely dependent upon Asp14, as this residue can be effectively replaced by Cys. However, the efficiency of electron transfer is compromised if Asp14 is replaced by Ser, or if the 4Fe-cluster is converted to the 3Fe-form, but Asp14 does not appear to offer any kinetic advantage over the expected Cys.     

N- and O-glycosylation and phosphorylation of the bar secretion leader derived from the barrier protease of Saccharomyces cerevisiae

A secretion leader derived from a domain of the extracellular Barrier protease of the yeast Saccharomyces cerevisiae has been expressed in wild-type and in mnn1, mnn9, and mnn1 mnn9 glycosylation mutant strains of S. cerevisiae. Structural comparison of the extracellular leader by mass spectrometry, peptide mapping, and elementary analysis proved that all strains produced a heterogeneous, heavily glycosylated polypeptide of 161 amino acids with both N- and O-glycosylation and phosphorylation. All three potential Asn N-linked sites were glycosylated to some extent with the expected structures. Neither the different growth media used nor the glycosylation mutations had significant effect on O-glycosylation with respect to both site selectivity and size of the carbohydrate structures. All 33 Ser and 21 Thr residues in the polypeptide were glycosylated at least partially, with an average of more than 2 mannoses/site. Although the mnn1 mutation blocks addition of alpha 1,3-linked mannose, the bar secretion domain expressed in the mnn1 and mnn1 mnn9 transformants unexpectedly contained some O-linked structures with at least 4 mannoses/chain. These O-linked structures were as large as when the leader was expressed in the mnn9 and wild-type strains. The bar secretion domain also had a previously undocumented phosphorylated O-linked structure.     

Hydrolysis of phosphodiesters through transformation of the bacterial phosphotriesterase

The phosphotriesterase from Pseudomonas diminuta catalyzes the hydrolysis of a wide array of phosphotriesters and related phosphonates, including organophosphate pesticides and military nerve agents. It has now been shown that this enzyme can also catalyze the hydrolysis of phosphodiesters, albeit at a greatly reduced rate. However, the enzymatic hydrolysis of ethyl-4-nitrophenyl phosphate (compound I) by the wild-type enzyme was >10(8) times faster than the uncatalyzed reaction (kcat = 0.06 s-1 and Km = 38 mM). Upon the addition of various alkylamines to the reaction mixture, the kcat/Km for the phosphodiester (compound I) increased up to 200-fold. Four mutant enzymes of the phosphotriesterase were constructed in a preliminary attempt to improve phosphodiester hydrolysis activity of the native enzyme. Met-317, which is thought to reside in close proximity to the pro-S-ethoxy arm of the paraoxon substrate, was mutated to arginine, alanine, histidine, and lysine. These mutant enzymes showed slight improvements in the catalytic hydrolysis of organophosphate diesters. The M317K mutant enzyme displayed the most improvement in catalytic activity (kcat = 0.34 s-1 and Km = 30 mM). The M317A mutant enzyme catalyzed the hydrolysis of the phosphodiester (compound I) in the presence of alkylamines up to 200 times faster than the wild-type enzyme in the absence of added amines. The neutralization of the negative charge on the oxygen atom of the phosphodiester by the ammonium cation within the active site is thought to be responsible for the rate enhancement by these amines in the hydrolytic reaction. These results demonstrate that an active site optimized for the hydrolysis of organophosphate triesters can be made to catalyze the hydrolysis of organophosphate diesters.     

A direct estimation of the context effect on the efficiency of termination

An in vitro assay in which terminating Escherichia coli ribosomes with different stop signals in the A-site compete for a limited amount of a release factor (RF1 or RF2) has been used to estimate the relative termination efficiencies at stop codons with different adjacent downstream nucleotides. The assay allows direct measurements of relative kcat/Km parameters for the productive association of release factors to ribosomes. The kcat/Km parameter is larger for UAA(U) than for UAA(C) programmed ribosomes and the difference in kcat/Km is much larger for RF2 (about 80%) than for RF1 (about 30%). These differences in the kcat/Km parameter are not affected by the addition of release factor RF3. The only discernible effect of RF3 is a considerable acceleration of RF1/2 recycling.The estimated kcat/Km parameters correlate well with the affinities of release factors for ribosomes programmed with different stop signals. These affinities were estimated from the extent of inhibition of ribosomal recycling by high concentrations of release factors in the absence of release factor RF3. The affinity for RF2 depends on the immediate downstream context of the stop codon in the translated mRNA and is about three times higher for UAA(U) than for UAA(C). The corresponding difference in affinities for RF1 is twofold. For all stop signals studied, the estimated affinity of RF2 for terminating ribosomes is much lower than that of RF1. It is also striking that the affinity of ribosomes for a chromosomally expressed RF2 is at least three times higher than for RF2 isolated from an overproducing E. coli strain.     

Effect of an amino acid insertion into the omega loop region of a class C beta-lactamase on its substrate specificity

The extended-substrate specificity of Enterobacter cloacae GC1 beta-lactamase is entirely due to a three amino acid insertion after position 207. To clarify the reason for the extended-substrate specificity, Ala, Ala-Ala, Ala-Ala-Ala, and Ala-Ala-Ala-Ala were inserted after position 207 on the basis of the class C beta-lactamase from E. cloacae P99, respectively. The kcat and Km values of all the mutant enzymes for cephalothin, benzylpenicillin and ampicillin were almost the same as those of the wild-type enzyme, except for those of P99-210-4A which were decreased 4-15-fold. On the other hand, the kcat and Km values for oxyimino beta-lactams such as cefuroxime, ceftazidime, and aztreonam increased with increasing numbers of inserted alanines. The kcat values of the mutant enzymes for cefroxime increased 140-7400-fold compared with that of the wild-type. The Km values also increased with almost the same magnitude, resulting in about the same kcat/Km values as that of the wild-type. On progressive inhibition analysis of aztreonam of the mutant enzymes, two kinds of inactive acyl-enzyme with distinct stabilities were observed, and the proportion of the less stable inactive enzyme increased with increasing numbers of inserted alanines. This suggests that the extension of the substrate specificity is due to instability of the acyl-intermediate caused by an increased deacylation rate in the reaction process.     

The bifunctional enzyme leukotriene-A4 hydrolase is an arginine aminopeptidase of high efficiency and specificity

Leukotriene-A4 hydrolase (EC 3.3.2.6) cleaved the NH2-terminal amino acid from several tripeptides, typified by arginyl-glycyl-aspartic acid, arginyl-glycyl-glycine, and arginyl-histidyl-phenylalanine, with catalytic efficiencies (kcat/Km) > or = 1 x 10(6) M-1 s-1. This exceeds by 10-fold the kcat/Km for its lipid substrate leukotriene A4. Catalytic efficiency declined for dipeptides which had kcat/Km ratios 10-100-fold lower than tripeptides. Tetrapeptides and pentapeptides were even poorer substrates with catalytic efficiencies below 10(3) M-1 s-1. The enzyme preferentially hydrolyzed tripeptide substrates and single amino acid p-nitroanilides with L-arginine at the NH2 terminus. Peptides with proline at the second position were not hydrolyzed, suggesting a requirement for an N-hydrogen at the peptide bond cleaved. Peptides with a blocked NH2 terminus were not hydrolyzed. The specificity constant (kcat/Km) was optimal at pH 7.2 with pK values at 6.8 and 7.9; binding was maximal at pH 8.0. Serum albumins activated the peptidase, increasing tripeptide affinities (Km) by 3-10-fold and specificities (kcat/Km) by 4-13-fold. Two known inhibitors of arginine peptidases, arphamenine A and B, inhibited hydrolysis of L-arginine p-nitroanilide with dissociation constants = 2.0 and 2.5 microM, respectively. Although the primary role of LTA4 hydrolase is widely regarded as the conversion of the lipid substrate leukotriene A4 into the inflammatory lipid mediator leukotriene B4, our data are the first showing that tripeptides are "better" substrates. This is compatible with a biological role for the peptidase activity of the enzyme and may be relevant to the distribution of the enzyme in organs like the ileum, liver, lung, and brain. We present a model which accommodates the available data on the interaction of substrates and inhibitors with the enzyme. This model can account for overlap in the active site for hydrolysis of leukotriene A4 and peptide or p-nitroanilide substrates.     

Investigation of human cyclooxygenase-2 glycosylation heterogeneity and protein expression in insect and mammalian cell expression systems

Human cyclooxygenase-2 (hCox-2) is a key enzyme in the biosynthesis of prostaglandins and the target of nonsteroidal anti-inflammatory drugs. Recombinant hCox-2 overexpressed in a vaccinia virus (VV)-COS-7 system comprises two glycoforms. Removal of the N-glycosylation consensus sequence at Asn580 (N580Q and S582A mutants) resulted in the expression of protein comprising a single glycoform, consistent with the partial N-glycosylation at this site in the wild-type (WT) enzyme. The specific cyclooxygenase activities of the purified WT and N580Q mutant were equivalent (40 +/- 3 mumol O2/min/mg) and titrations with diclofenac showed no difference in inhibitor sensitivities of WT and both mutants. Results of the expression of WT and N580Q hCox-2 in a Drosophila S2 cell system were also consistent with the N-glycosylation at this site, but low levels of activity were obtained. High levels of N-glycosylation heterogeneity are observed in hCox-2 expressed using recombinant baculovirus (BV) in Sf9 cells. Expression of a double N-glycosylation site mutant in Sf9 cells, N580Q/N592Q, resulted in a decrease in glycosylation but no clear decrease in heterogeneity, indicating that the high degree of N-glycosylation heterogeneity observed with the BV-Sf9 system is not due to partial glycosylation of both Asn580 and Asn592. N-linked oligosaccharide profiling of purified VV and BV WT and S582A mutant hCox-2 showed the presence of high mannose structures, (Man)n (GlcNAc)2, n = 9, 8, 7, 6. The S582A mutant was the most homogeneous with (Man)9(GlcNAc)2 comprising greater than 50% of oligosaccharides present. Analysis of purified VV WT and S582A mutant hCox-2 by liquid chromatography-electrospray ionization-mass spectrometry showed an envelope of peaks separated by approximately 160 Da, corresponding to differences of a single monosaccharide. The difference between the highest mass peaks of the two envelopes, of approximately 1500 Da, is consistent with the wild-type enzyme containing an additional high mannose oligosaccharide.     

Role of cysteine residues in human angiotensinogen. Cys232 is required for angiotensinogen-pro major basic protein complex formation

The M235T polymorphism of human angiotensinogen is associated with essential and pregnancy-induced hypertension. A covalent complex is formed between angiotensinogen and the proform of the eosinophil major basic protein (proMBP) during pregnancy. The sequence of human angiotensinogen contains four cysteines. Their function was analyzed. Presence of free cysteines was demonstrated by their alkylation with iodo[14C]acetic acid. A disulfide bond between Cys18 and Cys138 using a fully N-deglycosylated mutant of human angiotensinogen was identified by tryptic digestion and mass spectrometry. We produced angiotensinogen. proMBP complex by co-transfection of COS-7 cells and by co-culturing transfected CHO-K1 cells. Experiments with 8 mutated recombinant angiotensinogen, in which one or more of the four cysteines were replaced by alanine, demonstrated that Cys232 is involved in complex formation and could interact with the M235T variant. The angiotensinogen.proMBP complex was isolated by molecular sieving. Hydrolysis of the complex by human renin was 7 times slower than hydrolysis of monomeric form, whatever the M235T genotype. The complex:monomeric angiotensinogen ratio was greater for Met235 (72%) than for Thr235 (58%) angiotensinogen. These data suggest a new pathophysiological explanation for the genetic association between M235T angiotensinogen polymorphism and pregnancy-induced hypertension.     

N-linked glycosylation of the ligand-binding domain of the human urokinase receptor contributes to the affinity for its ligand

Variations in glycosylation exist among urokinase plasminogen activator receptors (u-PARs) from different cell types. We have studied the functional role of N-linked carbohydrate within the ligand-binding domain of u-PAR. Treatment with glycosidases demonstrated that all the N-linked carbohydrates on u-PAR are complex-type oligosaccharides. Substitution of a single Asn (Asn52) to Gln by means of site-directed mutagenesis led to an active receptor mutant with a ligand-binding domain devoid of carbohydrate. The cellular distribution, the glycosyl-phosphatidylinositol anchoring, and the conformational stability after solubilization were unaffected by this single substitution. However, ligand binding analysis demonstrated a 4- 5-fold decrease in affinity as compared with the wild type receptor. Two different strategies were used in order to obtain a u-PAR type completely devoid of N-linked carbohydrates. 1) Tunicamycin treatment of wild type u-PAR-expressing cells. 2) Mutation of all glycosylation sites (Hu-PARN5-mut). In neither case, unglycosylated receptors with ligand binding activity were identified. However, immunofluorescence studies demonstrated that the Hu-PARN5-mut was retained inside the cells in the endoplasmic reticulum. The same result was found for Hu-PARN4-mut, where only the glycosylation sites outside the binding domain were mutated. These results demonstrate that some extent of glycosylation of u-PAR is necessary for cellular transport and for molecular maturation events leading to ligand binding activity. Glycosylation of the binding domain per se affects only the affinity of the receptor. The positive modulation of the Asn52 carbohydrate side chain on ligand affinity suggests that the u-PAR glycosylation variants observed in various cell types may have different functional roles.     

Identifying the glycosylation sites and site-specific carbohydrate heterogeneity of glycoproteins by matrix-assisted laser desorption/ ionization mass spectrometry

A sensitive and facile method is described to identify the glycosylation sites and site-specific heterogeneity in the carbohydrate attached to glycoproteins. In this procedure, the peptide backbone of the glycoprotein is cleaved enzymatically. The resulting peptide/glycopeptide mixture is divided into three fractions. The first is analyzed directly by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), while the other two aliquots are analyzed by MALDI-MS after enzymatic release of the N-linked chains and the N- and O-linked chains. Comparison of these MALDI mass spectra provides the molecular weight of each carbohydrate side chain and of the peptide to which it was attached. This information combined with the amino acid sequence of the protein identifies the glycosylation sites, and provides information concerning site-specific oligosaccharide heterogeneity. This approach does not require time-consuming liquid chromatographic separations and can be performed on as little as 10 pmol of glycoprotein. Thus, our approach is faster and simpler than procedures currently used for glycosylation site mapping, and may offer a slight sensitivity advantage.     

Kinetic analysis of papaya proteinase omega

Papaya proteinase omega (pp omega) has been purified from dried latex both by immunoaffinity and traditional methods. Kinetic analysis revealed that (1), the pp omega-catalysed hydrolysis of N-benzoyl-L-arginine p-nitroanilide (BApNA) has a lower specificity (kcat/Km) than the same reaction catalysed by papain; (2), the pp omega-catalysed hydrolysis of a tripeptide substrate having phenylalanine at the second position (S2-site) showed a more similar specificity to that catalysed by papain; (3), the significant difference between the two enzymes is that steady state kinetics with both L-BApNA and a tripeptide enables the identification in pp omega of other ionizations affecting binding. The active sites of papain and pp omega can therefore be distinguished by pH-dependence of kcat/Km.     

N-glycosylation of recombinant human interferon-gamma produced in different animal expression systems

Recombinant human interferon-gamma (IFN-gamma) was expressed in Chinese hamster ovary cells, baculovirus-infected Sf9 insect cells and the mammary gland of transgenic mice. The N-linked carbohydrate populations associated with both Asn25 and Asn97 glycosylation sites were characterized by matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) in combination with exoglycosidase array sequencing. A site-specific analysis of dual (2N) and single (1N) site-occupancy variants of IFN-gamma derived from Chinese hamster ovary cells showed that N-glycans were predominantly of the complex bi- and triantennary type. Although Asn25-linked glycans were substituted with a core fucose residue, Asn97 N-glycans were predominantly non-fucosylated, and truncated complex and high-mannose oligosaccharide chains were also evident. Transgenic mouse derived IFN-gamma exhibited considerable site-specific variation in N-glycan structures. Asn25-linked carbohydrates were of the complex, core fucosylated type, Asn97-linked carbohydrates were mainly of the oligomannose type, with smaller proportions of hybrid and complex N-glycans. Carbohydrates associated with both glycosylation sites of IFN-gamma from Sf9 insect cells were mainly tri-mannosyl core structures, with fucosylation confined to the Asn25 site. These data demonstrate the profound influence of host cell type and protein structure on the N-glycosylation of recombinant proteins.     

Site-directed mutagenesis of the active site glutamate in human matrilysin: investigation of its role in catalysis

Glu-198 of human matrilysin is a conserved residue in the matrix metalloproteinases and is considered to play an important role in catalysis by acting as a general base catalyst toward the zinc-bound water molecule, on the basis of mechanistic proposals for other zinc proteases. In the present study, Glu-198 is mutated into Asp, Cys, Gln, and Ala, and the zinc binding properties, kinetic parameters, and pH dependence of each mutant are determined in order to examine the role of Glu-198 in catalysis. The mutations chosen either modify (C and D) or eliminate (A and Q) the general base properties of residue-198. All the mutants bind 2 mol of zinc per mol of enzyme, indicating that Glu-198 is not crucial to the binding of the catalytic zinc to the enzyme. The value of kcat/Km for the E198D mutant is only 4-fold lower than that of wild-type enzyme at the pH optimum of 7.5, while that for the E198C mutant is decreased by 160-fold. The E198Q and E198A enzymes containing the mutations that have eliminated the nucleophilic and acid/base properties of the residue are still active, having lower kcat/Km values of 590- and 1900-fold, respectively. The decrease in activity of all the mutants is essentially due to a decrease in kcat. The kcat/Km values of the mutants as a function of pH display broad bell-shaped curves that are similar to the wild-type enzyme. The acidic pKa value is not greatly affected by the change in the chemical properties of residue-198. The similarity in the pH profiles for the mutant and wild-type enzymes indicates that the ionization of Glu-198 is not responsible for the acidic pKa. Ionization of the zinc-bound water may be responsible for this pKa since the three His ligands and the scaffolding of the matrilysin catalytic zinc site are different from that observed in carboxypeptidase A and would predict a lower pKa for the metal-bound water. If the zinc-bound water is the nucleophile in the reaction, the role of Glu-198 in catalysis may be to stabilize the transition state or act as a general acid catalyst after the rate-determining step.     

Deamidation and isoaspartate formation during in vitro aging of recombinant tissue plasminogen activator

When incubated at pH 7.3, 37 degrees C, human recombinant tissue plasminogen activator accumulated 0.77 mol of isoaspartate per mol of plasminogen activator over a 14-day period. Isoaspartate was detected by enzymatic transfer of 3H-labeled methyl groups from S-adenosyl-L-methionine in a reaction catalyzed by protein L-isoaspartyl methyltransferase. Analysis of tryptic peptides derived from aged plasminogen activator revealed that the two major sites of isoaspartate accumulation resulted from deamidation of Asn58 in the sequence -FNGG- and Asn177 in the sequence -GNSD-. Significant levels of isoaspartate also accumulated via deamidation of Asn37 in the sequence -CNSG-. All three sites occur in sequences predicted from studies with synthetic peptide to be unstable. All three sites appear to be on the surface of the protein, and all three occur in regions of the protein predicted to have higher than average chain mobility. These findings add support to the idea that sequence and flexibility play major roles in determining susceptibility to deamidation and peptide bond isomerization at Asn and Asp sites under mild conditions. These studies also illustrate the utility of enzymatic methylation for characterizing sites of deamidation in a large protein that contains numerous disulfide bonds and several sites of glycosylation.     

Engineering steroid 5 beta-reductase activity into rat liver 3 alpha-hydroxysteroid dehydrogenase

Delta 4-3-Ketosteroid-5 beta-reductase (5 beta-reductase) precedes 3 alpha-hydroxysteroid dehydrogenase (3 alpha-HSD) in steroid hormone metabolism. Both enzymes are members of the aldo-keto reductase (AKR) superfamily and possess catalytic tetrads differing by a single amino acid. In 3 alpha-HSD, the tetrad consists of Tyr55, Lys84, Asp50, and His117, but a glutamic acid replaces His117 in 5 beta-reductase. By introducing the H117E point mutation into 3 alpha-HSD, we engineered 5 beta-reductase activity into the dehydrogenase. Homogeneous H117E 3 alpha-HSD reduced the double bond in testosterone to form 5 beta-dihydrotestosterone with kcat = 0.25 min-1 and Km = 19.0 microM and reduced the double bond in progesterone to generate 5 beta-dihydroprogesterone with kcat = 0.97 min-1 and Km = 33.0 microM. These kinetic parameters were similar to those reported for homogeneous rat liver 5 beta-reductase [Okuda, A., and Okuda, R. (1984) J. Biol. Chem. 259, 7519-7524]. The H117E mutant also reduced 5beta-dihydrosteroids to 5 beta, 3 alpha-tetrahydrosteroids with a 600-1000-fold decrease in kcat/Km versus wild-type 3 alpha-HSD. The ratio of 5 beta-reductase:3 alpha-HSD activity in the H117E mutant was approximately 1:1. Although the H117A mutant reduced Delta 4-3-ketosteroids, the 3 alpha-HSD activity predominated because the 5 beta-dihydrosteroids were rapidly converted to the 5 beta,3 alpha-tetrahydrosteroids. The pH-rate profiles for carbon-carbon double-bond and ketone reduction catalyzed by the H117E mutant were superimposable, suggesting a common titratable group (pKb = 6.3) for both reactions. In wild-type 3 alpha-HSD, the titratable group responsible for 3-ketosteroid reduction has a pKb = 6.9 and is assignable to Tyr55. The pH-rate profiles for 3-ketosteroid reduction by the H117A mutant were pH-independent. Our data indicate that Tyr55 functions as a general acid for both 3 alpha-HSD and 5 beta-reductase activities. We suggest that a protonated Glu117 increases the acidity of Tyr55 to promote acid-catalyzed enolization of the Delta 4-3-ketosteroid substrate. Further, the identity of amino acid 117 determines whether an AKR can function as a 5 beta-reductase by reorienting the substrate relative to the nicotinamide cofactor. This study provides functional evidence that utilization of modified catalytic residues on an identical protein scaffold is important for evolution of enzymatic activities within the same metabolic pathway.     

Mutational analysis of the major loop of Bacillus 1,3-1,4-beta-D-glucan 4-glucanohydrolases. Effects on protein stability and substrate binding

The carbohydrate-binding cleft of Bacillus licheniformis 1,3-1, 4-beta-D-glucan 4-glucanohydrolase is partially covered by the surface loop between residues 51 and 67, which is linked to beta-strand-(87-95) of the minor beta-sheet III of the protein core by a single disulfide bond at Cys61-Cys90. An alanine scanning mutagenesis approach has been applied to analyze the role of loop residues from Asp51 to Arg64 in substrate binding and stability by means of equilibrium urea denaturation, enzyme thermotolerance, and kinetics. The DeltaDeltaGU between oxidized and reduced forms is approximately constant for all mutants, with a contribution of 5.3 +/- 0.2 kcal.mol-1 for the disulfide bridge to protein stability. A good correlation is observed between DeltaGU values by reversible unfolding and enzyme thermotolerance. The N57A mutant, however, is more thermotolerant than the wild-type enzyme, whereas it is slightly less stable to reversible urea denaturation. Mutants with a <2-fold increase in Km correspond to mutations at residues not involved in substrate binding, for which the reduction in catalytic efficiency (kcat/Km) is proportional to the loss of stability relative to the wild-type enzyme. Y53A, N55A, F59A, and W63A, on the other hand, show a pronounced effect on catalytic efficiency, with Km > 2-fold and kcat < 5% of the wild-type values. These mutated residues are directly involved in substrate binding or in hydrophobic packing of the loop. Interestingly, the mutation M58A yields an enzyme that is more active than the wild-type enzyme (7-fold increase in kcat), but it is slightly less stable.     

Mutational analysis of N-linked glycosylation of esterase 6 in Drosophila melanogaster

The primary sequence of the esterase 6 (EST6) enzyme of Drosophila melanogaster contains four potential N-linked glycosylation sites, at residues 21, 399, 435, and 485. Here we determine the extent to which EST6 is glycosylated and how the glycosylation affects the biochemistry and physiology of the enzyme. We have abolished each of the four potential glycosylation sites by replacing the required Asn residues with Gln by in vitro mutagenesis. Five mutant genes were made, four containing mutations of each site individually and the fifth site containing all four mutations. Germline transformation was used to introduce the mutant genes into a strain of D. melanogaster null for EST6. Electrophoretic and Western blot comparisons of the mutant strains and wild-type controls showed that each of the four potential N-linked glycosylation sites in the wild-type protein is glycosylated. However, the fourth site is not utilized on all EST6 molecules, resulting in two molecular forms of the enzyme. Digestion with specific endoglycosidases showed that the glycan attached at the second site is of the high-mannose type, while the other three sites carry more complex oligosaccharides. The thermostability of the enzyme is not affected by abolition of the first, third, or fourth glycosylation sites but is reduced by abolition of the second site. Anomalously, abolition of all four sites together does not reduce thermostability. Quantitative comparisons of EST6 activities showed that abolition of glycosylation does not affect the secretion of the enzyme into the male sperm ejaculatory duct, its transfer to the female vagina during mating, or its subsequent translocation into her hemolymph. However, the activity of the mutant enzymes does not persist in the female's hemolymph for as long as wild-type esterase 6. The latter effect may compromise the role of the transferred enzyme in stimulating egg-laying and delaying receptivity to remating.