Improved Antiviral Activity of a Polyamide Against High-Risk Human Papillomavirus Via N-Terminal Guanidinium Substitution.

We report the synthesis of two novel pyrrole-imidazole polyamides with N-terminal guanidinium or tetramethylguanidinium groups and evaluate their antiviral activity against three cancer-causing human papillomavirus strains. Introduction of guanidinium improves antiviral activity when compared to an unsubstituted analog, especially in IC90 values. These substitutions change DNA-binding preferences, while binding affinity remains unchanged.

Differential effects of isomeric incorporation of fluorophenylalanines into PvuII endonuclease.

Incorporation of fluorine into proteins has long served as a means of probing structure and function, yet there are few studies that examine the impact of fluorine substitution, particularly at locations distant from the active sites of enzymes. The flexibility of isomeric fluorine incorporation at Phe is used to explore subtle substitution effects on enzyme activity and conformation. The unnatural amino acids o-, m-, and p-fluorophenylalanines were incorporated biosynthetically into the representative PvuII restriction endonuclease. Interestingly, m-fluoro-Phe-PvuII endonuclease exhibits very similar conformational stability to that of the native enzyme, but it exhibits a reproducible, 2-fold higher average specific activity. Given the level of incorporation and the distribution of species, the species of modified enzyme responsible for this increase in specific activity is most likely even faster. Further, moving the fluorine atom from the meta- to the para-position of Phe results in a 4-fold decrease in specific activity and a decrease in conformational stability of 1.5 kcal/mol. Since none of the Phe residues in PvuII endonuclease lies near the DNA recognition or catalytic sites, this differential behavior alludes to the impact of subtle changes in enzyme conformation on endonuclease activity and suggests novel ways to influence catalytic behavior.

The PD...(D/E)XK motif in restriction enzymes: a link between function and conformation.

The active sites of Mg(II)-dependent nucleases feature a cluster of conserved charged residues which includes both acidic (Asp and Glu) and basic (Lys) side chains. In restriction enzymes, these side chains are part of the conserved PD...(D/E)XK functional sequence motif which has been implicated as being important in metal ion binding and catalytic steps. Recent work revealing the unusual behavior of the active site variant D58A of the representative PvuII endonuclease prompted speculation that the array of charged groups in the nuclease active site may also be linked to conformational behavior [Dupureur, C. M., and Conlan, L. H. (2000) Biochemistry 39, 10921-10927]. To address this issue, we analyzed the conformational behavior of active site variants of PvuII endonuclease using both NMR spectroscopic and thermodynamic methods. NMR spectroscopic analysis via (19)F and (1)H-(15)N HSQC experiments indicates that a number of side chain and backbone amide groups are perturbed upon Ala substitution at conserved active site residues Asp58, Glu68, and Lys70. Spectral changes are particularly pronounced for the lowest-activity mutants (D58A and K70A). These changes are accompanied by perturbations in conformational stability. Ala substitution at each of these positions results in 2-5 kcal/mol of stabilization over the wild-type enzyme at pH 7.7, changes which constitute increases in DeltaG(d)(H2O) of 20-50%. The pH dependencies of mutant enzyme stabilities are distinct from those of the wild type, results which confirm that these ionizable groups strongly influence stability. Wild-type enzyme stability is correlated with the ionization of groups shown to be important to metal ion binding and orientation. Correlations between spectral changes and conformational stability indicate that the latter measurements may prove useful in the evaluation of site-directed mutant restriction enzymes. More importantly, these results indicate that structure-function relationships in restriction enzyme active sites can be complex, and that the ensemble of conserved charged residues which mediate DNA hydrolysis in Mg(II)-dependent nucleases constitutes a critical link between function and conformation.

A catalytically deficient active site variant of PvuII endonuclease binds Mg(II) ions.

In efforts to understand the mechanisms of many nucleic acid enzymes, the first site-directed mutations are made at conserved acidic active residues. Almost without exception, the low or null activities of the resulting variants are attributed to the importance of the acidic residue(s) to the ligation of required metal ions. Using (25)Mg NMR spectroscopy as a direct probe of metal ion binding and the homodimeric PvuII restriction endonuclease as a model system, this interpretation is examined and clarified. Our results indicate that Mg(II) binds wild-type PvuII endonuclease in the absence of DNA with a K(d,app) of 1.9 mM. Hill analysis yields an n(H) coefficient of 1.4, a value consistent with the binding of more than one Mg(II) ion per monomer active site. Variable pH studies indicate that two ionizable groups are responsible for Mg(II) binding by wild-type PvuII endonuclease near physiological pH. The pK(a,app) for these ionizations is 6.7, a value which is unusual for acidic residues but consistent with data obtained for critical groups in MunI endonuclease and a number of other hydrolases. To assign residues critical to ligating Mg(II), binding measurements were performed on the low activity catalytic site mutants E68A and D58A. As expected, E68A binds Mg(II) ions very weakly (K(d,app) approximately 40 mM), implicating Glu68 as critical to Mg(II) binding. Interestingly, while D58A has only residual specific activity, it retains an affinity for Mg(II) with a K(d,app) of 3.6 mM and exhibits a Hill coefficient of 0.7. Moreover, in this variant, multiple ionizable groups with pK(a,app) of 7.2 are involved in Mg(II) binding, suggesting a shuffling of Mg(II) ligands in the active site. These data indicate that Asp58 is important for the critical positioning of metal ion(s) required for catalysis.

Modulating restriction endonuclease activities and specificities using neutral detergents.

It is well known that type II restriction enzyme activities and specificities can be modulated by altering solution conditions. The addition of co-solvents such as dimethyl sulfoxide (DMSO), alcohols and polyols can promote star activity, which is the cleavage of non-cognate sequences. While neutral detergents are often used to control protein aggregation, little is known about the effect of neutral detergents on restriction enzyme activities and specificities. We report here that BamHI, BglI, BglII, EcoRI, EcoRV, HindIII, MluI, PvuII, SalI and XhoI restriction endonucleases are remarkably tolerant of high concentrations of neutral detergents Triton X-100, CHAPS and octyl glucoside. In most cases, lambda DNA cleavage rates were comparable to those observed in the absence of detergent. Indeed, the specific activities of SalI and XhoI were appreciably increased in the presence of Triton X-100. For all enzymes active in the presence of detergents, sequence specificity toward lambda DNA was not compromised. Assays of star cleavage of pUC18 by EcoRI, PvuII and BamHI endonucleases in equimolar concentrations of Triton X-100 and sucrose revealed reduced star activity in the detergent relative to the sucrose co-solvent. Interestingly, under star activity-promoting conditions, PvuII endonuclease displayed greater fidelity in Triton X-100 than in conventional buffer. Taken altogether, these results suggest that in some cases, neutral detergents can be used to manipulate restriction endonuclease reaction rates and specificities.

Effects of divalent metal ions on the activity and conformation of native and 3-fluorotyrosine-PvuII endonucleases.

The activities of restriction enzymes are important examples of Mg(II)-dependent hydrolysis of DNA. While a number of crystallographic studies of enzyme-DNA complexes have also involved metal ions, there have been no solution studies exploring the relationship between enzyme conformation and metal-ion binding in restriction enzymes. Using PvuII restriction endonuclease as a model system, we have successfully developed biosynthetic fluorination and NMR spectroscopy as a solution probe of restriction-enzyme conformation. The utility of this method is demonstrated with a study of metal-ion binding by PvuII endonuclease. Replacement of 74% (+/- 10%) of the Tyr residues in PvuII endonuclease by 3-fluorotyrosine produces an enzyme with Mg(II)-supported specific activity and sequence specificity that is indistinguishable from that of the native enzyme. Mn(II) supports residual activity of both the native and fluorinated enzymes; Ca(II) does not support activity in either enzyme, a result consistent with previous studies. 1H- and 19F-NMR spectroscopic studies reveal that while Mg(II) does not alter the enzyme conformation, the paramagnetic Mn(II) produces both short-range spectral broadening and longer range changes in chemical shift. Most interestingly, Ca(II) binding perturbs a larger number of different resonances than Mn(II). Coupled with earlier mutagenesis studies that place Ca(II) in the active site [Nastri, H. G., Evans, P.D., Walker, I.H. & Riggs, P.D. (1997) J. Biol. Chem. 272, 25761-25767], these data suggest that the enzyme makes conformational adjustments to accommodate the distinct geometric preferences of Ca(II) and may play a role in the inability of this metal ion to support activity in restriction enzymes.

Quantitative evaluation of metal ion binding to PvuII restriction endonuclease.

Restriction enzymes are important examples of phosphodiester hydrolysis activity and as such have been of increasing interest to structural biologists. Much of the architecture of endonuclease active sites has been derived from X-ray crystallographic studies. These structures implicate conserved active site acidic residues and the scissile bond of the substrate as coordination ligands of required metal ions. Central to the development of restriction enzyme mechanism is our understanding of the role of metal ion binding in the reaction, an important feature of which is identifying the energetic contributions of the enzyme and the substrate to metal ion affinity. To begin to address this issue, isothermal titration calorimetry (ITC) and 19F NMR spectroscopy have been applied to evaluate metal ion binding by the representative PvuII endonuclease in the absence of substrate. In separate experiments, ITC data demonstrate that PvuII endonuclease binds 2.16 Mn(II) ions and 2.05 Ca(II) metal ions in each monomer active site with Kd values of approximately 1 mM. While neither calorimetry nor protein NMR spectroscopy is directly sensitive to Mg(II) binding to the enzyme, Mn(II) competes with Mg(II) for common sites(s) on PvuII endonuclease. Substitution of the conserved active site carboxylate Glu68 with Ala resulted in a loss of affinity for both equivalents of both Ca(II) and Mn(II). Interestingly, the active site mutant D58A retained an affinity for Mn(II) with Kd approximately 2 mM. Mn(II) paramagnetic broadening in 19F spectra of wild-type and mutant 3-fluorotyrosine PvuII endonucleases are consistent with ITC results. Chemical shift analysis of 3-fluorotyrosine mutant enzymes is consistent with a perturbed conformation for D58A. Therefore, free PvuII endonuclease binds metal ions, and metal ion binding can precede DNA binding. Further, while Glu68 is critical to metal ion binding, Asp58 does not appear to be critical to the binding of at least one metal ion and appears to also have a role in structure. These findings provide impetus for exploring the roles of multiple metal ions in the structure and function of this representative endonuclease.

Phospholipase A2 engineering. The roles of disulfide bonds in structure, conformational stability, and catalytic function.

Site-directed mutagenesis was used to probe the contribution of each of the seven disulfide bonds of bovine pancreatic phospholipase A2 (PLA2, overexpressed in Escherichia coli) to the structure, conformational stability, and catalytic function of the enzyme. Each of the seven disulfide bonds, C11-C77, C27-C123, C29-C45, C44-C105, C51-C98, C61-C91, and C84-C96, was deleted separately by changing both cysteine (C) residues to alanine (A). The structural properties of the mutants were analyzed by 1D and 2D proton NMR, the conformational stability by guanidine hydrochloride-induced denaturation, and the catalytic property by measuring kinetic parameters toward DC8PC (1,2-dioctanoyl-sn-glycero-3-phosphocholine) micelles. The results led to the following significant findings: (i) All but one (C84A-C96A) mutants have been refolded and purified by use of the same procedure for wild-type PLA2. Thus, the disulfide bonds are generally not important to the folding pathway of PLA2. (ii) The disulfide bond C11-C77 is most important to the conformation and conformational stability of the enzyme since deletion of this disulfide bond resulted in greatly perturbed NMR properties and in a decrease of 6.2 kcal/mol in conformational stability. However, the C11A-C77A mutant displayed little change in catalytic function. (iii) The effects of deleting disulfide bonds on the catalytic function of PLA2 are small, except the disulfide bond C29-C45 which connects the calcium binding loop with the helix C. However, the conformation and conformational stability of the C29A-C45A mutant was found to decrease by a factor of 10 or greater. (ABSTRACT TRUNCATED AT 250 WORDS)

Binding proteins on synaptic membranes for crotoxin and taipoxin, two phospholipases A2 with neurotoxicity.

Crotoxin and taipoxin are both neurotoxic phospholipases A2 capable of affecting the presynaptic activity to bring about ultimate blockade of synaptic transmission. The enzymatic activity has generally been considered to be necessary but not sufficient for the blockade. Since many phospholipases A2 with comparable or even higher enzymatic activity are not toxic, it has been postulated that the difference lies in the affinity of binding to the presynaptic membrane. In confirmation of this proposition, we and others have previously shown that iodinated crotoxin and taipoxin bind specifically with high affinity to the isolated synaptic membrane fraction from guinea-pig brain, whereas specific binding is not detected with the nontoxic pancreatic phospholipase A2. Experiments based on photoaffinity labeling and simple chemical cross-linking techniques have led to the identification of three polypeptides preferentially present in neuronal membranes as (subunits of) the binding protein(s) for crotoxin and/or taipoxin. Some, but not all, other toxic phospholipases A2 also appear to be ligands for the three polypeptides. We now report studies on partial purification of these polypeptides using affinity chromatography and other techniques. In order to learn the normal physiological roles played by the toxin-binding proteins, the phospholipase-independent effects of the toxins on the synaptosomes have been sought. We have found that under Ca(2+)-free condition, taipoxin or crotoxin inhibits with IC50 of 20-1000 nM the Na(+)-dependent uptake of norepinephrine, dopamine and serotonin by the synaptosomes. In contrast, choline uptake is not affected.(ABSTRACT TRUNCATED AT 250 WORDS)

Conversion of bovine pancreatic phospholipase A2 at a single site into a competitor of neurotoxic phospholipases A2 by site-directed mutagenesis.

A 45-kDa polypeptide preferentially present in neuronal membranes was previously identified as a subunit of a binding (or receptor) protein for several phospholipase A2 variants with neurotoxicity, including crotoxin, by chemical cross-linking experiments (Yen, C.-H., and Tzeng, M.-C. (1991) Biochemistry 30, 11473-11477). The binding of crotoxin to this receptor protein was completely suppressed by sufficient F22Y, a mutated bovine pancreatic phospholipase A2 generated by site-directed mutagenesis of Phe22 of the wild-type enzyme to Tyr. The IC50 of this inhibition was estimated to be 1 microM. In sharp contrast, the wild-type enzyme gave no effect even at 50 microM. This mutation resulted in only minor and localized structural perturbations with little effect on enzymatic activity. Other phospholipase A2 molecules capable of competing with crotoxin for this binding invariably have Tyr at this position. It was concluded that this Tyr residue is an important determinant for the binding of a number of phospholipase A2 variants to the 45-kDa receptor.

Structure and function of the catalytic site mutant Asp 99 Asn of phospholipase A2: absence of the conserved structural water.

To probe the role of the Asp-99 ... His-48 pair in phospholipase A2 (PLA2) catalysis, the X-ray structure and kinetic characterization of the mutant Asp-99-->Asn-99 (D99N) of bovine pancreatic PLA2 was undertaken. Crystals of D99N belong to the trigonal space group P3(1)21 and were isomorphous to the wild type (WT) (Noel JP et al., 1991, Biochemistry 30:11801-11811). The 1.9-A X-ray structure of the mutant showed that the carbonyl group of Asn-99 side chain is hydrogen bonded to His-48 in the same way as that of Asp-99 in the WT, thus retaining the tautomeric form of His-48 and the function of the enzyme. The NH2 group of Asn-99 points away from His-48. In contrast, in the D102N mutant of the protease enzyme trypsin, the NH2 group of Asn-102 is hydrogen bonded to His-57 resulting in the inactive tautomeric form and hence the loss of enzymatic activity. Although the geometry of the catalytic triad in the PLA2 mutant remains the same as in the WT, we were surprised that the conserved structural water, linking the catalytic site with the ammonium group of Ala-1 of the interfacial site, was ejected by the proximity of the NH2 group of Asn-99. The NH2 group now forms a direct hydrogen bond with the carbonyl group of Ala-1.

Crystal structure of the Y52F/Y73F double mutant of phospholipase A2: increased hydrophobic interactions of the phenyl groups compensate for the disrupted hydrogen bonds of the tyrosines.

The enzyme phospholipase A2 (PLA2) catalyzes the hydrolysis of the sn-2 ester bond of membrane phospholipids. The highly conserved Tyr residues 52 and 73 in the enzyme form hydrogen bonds to the carboxylate group of the catalytic Asp-99. These hydrogen bonds were initially regarded as essential for the interfacial recognition and the stability of the overall catalytic network. The elimination of the hydrogen bonds involving the phenolic hydroxyl groups of the Tyr-52 and -73 by changing them to Phe lowered the stability but did not significantly affect the catalytic activity of the enzyme. The X-ray crystal structure of the double mutant Y52F/Y73F has been determined at 1.93 A resolution to study the effect of the mutation on the structure. The crystals are trigonal, space group P3(1)21, with cell parameters a = b = 46.3 A and c = 102.95 A. Intensity data were collected on a Siemens area detector, 8,024 reflections were unique with an R(sym) of 4.5% out of a total of 27,203. The structure was refined using all the unique reflections by XPLOR to a final R-factor of 18.6% for 955 protein atoms, 91 water molecules, and 1 calcium ion. The root mean square deviation for the alpha-carbon atoms between the double mutant and wild type was 0.56 A. The crystal structure revealed that four hydrogen bonds were lost in the catalytic network; three involving the tyrosines and one involving Pro-68. However, the hydrogen bonds of the catalytic triad, His-48, Asp-99, and the catalytic water, are retained. There is no additional solvent molecule at the active site to replace the missing hydroxyl groups; instead, the replacement of the phenolic OH groups by H atoms draws the Phe residues closer to the neighboring residues compared to wild type; Phe-52 moves toward His-48 and Asp-99 of the catalytic diad, and Phe-73 moves toward Met-8, both by about 0.5 A. The closing of the voids left by the OH groups increases the hydrophobic interactions compensating for the lost hydrogen bonds. The conservation of the triad hydrogen bonds and the stabilization of the active site by the increased hydrophobic interactions could explain why the double mutant has activity similar to wild type. The results indicate that the aspartyl carboxylate group of the catalytic triad can function alone without additional support from the hydrogen bonds of the two Tyr residues.

Phospholipase A2 engineering. The structural and functional roles of aromaticity and hydrophobicity in the conserved phenylalanine-22 and phenylalanine-106 aromatic sandwich.

The highly conserved phenylalanine-22 and phenylalanine-106, arranged as an aromatic sandwich, form part of an invariant hydrophobic wall that shields the active site of bovine pancreatic phospholipase A2 (PLA2) from bulk solvent [Dijkstra, B. W., Drenth, J., & Kalk, K. H. (1981) Nature 289, 604-606]. The residues have also been suggested to interact with the sn-2 acyl chain of bound phospholipid substrate [White, S. P., Scott, D. L., Otwinowski, Z., Gelb, M. H., & Sigler, P. B. (1990) Science 250, 1560-1563]. We now report the importance of these two residues in the structure and function of PLA2 in terms of aromaticity (changing to Ile) and hydrophobic (changing to Ala) and hydrophilic (changing to Tyr) character of these residues. The structural properties of the mutants were analyzed by proton NMR and by guanidine hydrochloride-induced denaturation. The functional properties were determined by measuring kinetic parameters toward various substrates in the forms of monomers, micelles, and vesicles, and by measuring equilibrium dissociation constants at the interface. The results show that (i) The conformational stability of each mutant was as good as that of wild-type PLA2; none of the mutants was significantly perturbed structurally as judged from detailed 1H NMR analysis. These results suggest that neither the Phe-22/Phe-106 face-to-face pair nor the Phe-22/Tyr-111 edge-to-face pair plays a significant structural role. (ii) Mutations to Ile at either position 22 or position 106 resulted in only minor perturbations in activity. This suggests that the aromaticity is not important to the function of these two residues.(ABSTRACT TRUNCATED AT 250 WORDS)

Phospholipase A2 engineering. Structural and functional roles of highly conserved active site residues tyrosine-52 and tyrosine-73.

Site-directed mutagenesis was used to probe the structural and functional roles of two highly conserved residues, Tyr-52 and Tyr-73, in interfacial catalysis by bovine pancreatic phospholipase A2 (PLA2, overproduced in Escherichia coli). According to crystal structures, the side chains of these two active site residues form H-bonds with the carboxylate of the catalytic residue Asp-99. Replacement of either or both Tyr residues by Phe resulted in only very small changes in catalytic rates, which suggests that the hydrogen bonds are not essential for catalysis by PLA2. Substitution of either Tyr residue by nonaromatic amino acids resulted in substantial decreases in the apparent kcat toward 1,2-dioctanoyl-sn-glycero-3-phosphocholine (DC8PC) micelles and the v(o) (turnover number at maximal substrate concentration, i.e., mole fraction = 1) toward 1,2-dimyristoyl-sn-glycero-3-phosphomethanol (DC14PM) vesicles in scooting mode kinetics [Berg, O. G., Yu, B.-Z., Rogers, J., & Jain, M. K. (1991) Biochemistry 30, 7283-7297]. The Y52V mutant was further analyzed in detail by scooting mode kinetics: the E to E* equilibrium was examined by fluorescence; the dissociation constants of E*S, E*P, and E*I (KS*, KP*, and KI*, respectively) in the presence of Ca2+ were measured by protection of histidine-48 modification and by difference UV spectroscopy; the Michaelis constant KM* was calculated from initial rates of hydrolysis in the absence and presence of competitive inhibitors; and the turnover number under saturating conditions (kcat, which is a theoretical value since the enzyme may not be saturated at the interface) was calculated from the vo and KM* values. The results indicated little perturbation in the interfacial binding step (E to E*) but ca. 10-fold increases in KS*, KP*, KI*, and KM* and a less than 10-fold decrease in kcat. Such changes in the function of Y52V are not due to global conformational changes since the proton NMR properties of Y52V closely resemble those of wild-type PLA2; instead, it is likely to be caused by perturbed enzyme-substrate interactions at the active site. Tyr-73 appears to play an important structural role. The conformational stability of all Tyr-73 mutants decreased by 4-5 kcal/mol relative to that of the wild-type PLA2. The proton NMR properties of Y73A suggested significant conformational changes and substantially increased conformational flexibility. These detailed structural and functional analyses represent a major advancement in the structure-function study of an enzyme involved in interfacial catalysis.

Phospholipase A2 engineering. X-ray structural and functional evidence for the interaction of lysine-56 with substrates.

Site-directed mutagenesis studies of bovine pancreatic phospholipase A2 (PLA2, overproduced in Escherichia coli) showed that replacement of surface residue Lys-56 by a neutral or hydrophobic amino acid residue resulted in an unexpected and significant change in the function of the enzyme. The kcat for phosphatidylcholine micelles increases 3-4-fold for K56M, K56I, and K56F and ca. 2-fold for K56N and K56T but does not change for K56R. These results suggest that the side chain of residue 56 has significant influence on the activity of PLA2. In order to probe the structural basis for the enhanced activity, the crystal structures of wild-type and K56M PLA2 were determined by X-ray crystallography to a resolution of 1.8 A. The results suggest that the mutation has not only perturbed the conformation of the side chain of Met-56 locally but also caused conformational changes in the neighboring loop (residues 60-70), resulting in the formation of a hydrophobic pocket by residues Met-56, Tyr-52, and Tyr-69. Docking of a phosphatidylcholine inhibitor analogue into the active site of K56M, according to the structure of the complex of cobra venom PLA2-phosphatidylethanolamine inhibitor analogue [White, S.P., Scott, D. L., Otwinowski, Z., Gleb, M. H., & Sigler, P. (1990) Science 250, 1560-1563], showed that the choline moiety [N(CH3)3]+ is readily accommodated into the newly formed hydrophobic pocket with a high degree of surface complementarity. This suggests a possible interaction between residue 56 and the head group of the phospholipid, explaining the enhanced activities observed when the positively charged Lys-56 is substituted by apolar residues, viz., K56M, K56I, and K56F. Further support for this interpretation comes from the 5-fold enhancement in kcat for the mutant K56E with a negatively charged side chain, where there would be an attractive electrostatic interaction between the side chain of Glu-56 and the positively charged choline moiety. Our results also refute a recent report [Tomasselli, A. G., Hui, J., Fisher, J., Zürcher-Neely, H., Reardon, I.M., Oriaku, E., Kézdy, F.J., & Heinrikson, R.L. (1989) J. Biol. Chem. 264, 10041-10047] that substrate-level acylation of Lys-56 is an obligatory step in the catalysis by PLA2.