Comparative kinetics of cofactor association and dissociation for the human and trypanosomal S-adenosylhomocysteine hydrolases. 3. Role of lysyl and tyrosyl residues of the C-terminal extension.

On the basis of the available X-ray structures of S-adenosylhomocysteine hydrolases (SAHHs), free energy simulations employing the MM-GBSA approach were applied to predict residues important to the differential cofactor binding properties of human and trypanosomal SAHHs (Hs-SAHH and Tc-SAHH), within 5 Šof the cofactor NAD(+)/NADH binding site. Among the 38 residues in this region, only four are different between the two enzymes. Surprisingly, the four nonidentical residues make no major contribution to differential cofactor binding between Hs-SAHH and Tc-SAHH. On the other hand, four pairs of identical residues are shown by free energy simulations to differentiate cofactor binding between Hs-SAHH and Tc-SAHH. Experimental mutagenesis was performed to test these predictions for a lysine residue and a tyrosine residue of the C-terminal extension that penetrates a partner subunit to form part of the cofactor binding site. The K431A mutant of Tc-SAHH (TcK431A) loses its cofactor binding affinity but retains the wild type's tetrameric structure, while the corresponding mutant of Hs-SAHH (HsK426A) loses both cofactor affinity and tetrameric structure [Ault-Riche, D. B., et al. (1994) J. Biol. Chem. 269, 31472-31478]. The tyrosine mutants HsY430A and TcY435A alter the NAD(+) association and dissociation kinetics, with HsY430A increasing the cofactor equilibrium dissociation constant from approximately 10 nM (Hs-SAHH) to ∼800 nM and TcY435A increasing the cofactor equilibrium dissociation constant from approximately 100 nM (Tc-SAHH) to ∼1 mM. Both changes result from larger increases in the off rate combined with smaller decreases in the on rate. These investigations demonstrate that computational free energy decomposition may be used to guide experimental studies by suggesting sensitive sites for mutagenesis. Our finding that identical residues in two orthologous proteins may give significantly different binding free energy contributions strongly suggests that comparative studies of homologous proteins should investigate not only different residues but also identical residues in these proteins.

Evaluation of NAD(H) analogues as selective inhibitors for Trypanosoma cruzi S-adenosylhomocysteine hydrolase.

S-Adenosylhomocysteine (AdoHcy) hydrolases (SAHHs) from human sources (Hs-SAHHs) bind the cofactor NAD(+) more tightly than several parasitic SAHHs by around 1000-fold. This property suggests the cofactor binding site of this essential enzyme as a potential anti-parasitic drug target, e.g., against SAHH from Trypansoma cruzi (Tc-SAHH). The on-rate and off-rate constants and the equilibrium dissociation constants were determined for NAD(+)/NADH analogues and suggested that NADH analogues were the most promising for selective inhibition of Tc-SAHH. None significantly inhibited Hs-SAHH while S-NADH and H-NADH (see Figure 1) reduced the catalytic activity of Tc-SAHH to < 10% in six minutes of exposure.

The rationale for targeting the NAD/NADH cofactor binding site of parasitic S-adenosyl-L-homocysteine hydrolase for the design of anti-parasitic drugs.

Trypanosomal S-adenoyl-L-homocysteine hydrolase (Tc-SAHH), considered as a target for treatment of Chagas disease, has the same catalytic mechanism as human SAHH (Hs-SAHH) and both enzymes have very similar x-ray structures. Efforts toward the design of selective inhibitors against Tc-SAHH targeting the substrate binding site have not to date shown any significant promise. Systematic kinetic and thermodynamic studies on association and dissociation of cofactor NAD/H for Tc-SAHH and Hs-SAHH provide a rationale for the design of anti-parasitic drugs directed toward cofactor-binding sites. Analogues of NAD and their reduced forms show significant selective inactivation of Tc-SAHH, confirming that this design approach is rational.

Synthesis of 5'-functionalized nucleosides: S-Adenosylhomocysteine analogues with the carbon-5' and sulfur atoms replaced by a vinyl or halovinyl unit.

Adenosine and uridine analogues functionalized with alkenyl or fluoroalkenyl chain at C5' were prepared employing cross-metathesis, Negishi couplings, and Wittig reactions. Metathesis of the protected 5'-deoxy-5'-methyleneadenosine or uridine analogues with six-carbon amino acids (homoallylglycines) in the presence of Grubbs catalysts gave nucleoside analogues with the C5'-C6' double bond. Alternatively, the Pd-catalyzed cross-coupling between the protected 5'-deoxy-5'-(iodomethylene) nucleosides and suitable alkylzinc bromides also provided analogues with alkenyl unit. Stereoselective Pd-catalyzed monoalkylation of 5'-(bromofluoromethylene)-5'-deoxyadenosine with alkylzinc bromides afforded adenosylhomocysteine analogues with a 6'-(fluoro)vinyl motif. The vinylic adenine nucleosides produced time-dependent inactivation of the S-adenosyl-l-homocysteine hydrolases.

Comparative kinetics of cofactor association and dissociation for the human and trypanosomal S-adenosylhomocysteine hydrolases. 2. The role of helix 18 stability.

The S-adenosyl- l-homocysteine (AdoHcy) hydrolases (SAHH) from Homo sapiens (Hs-SAHH) and from the parasite Trypanosoma cruzi (Tc-SAHH) are very similar in structure and catalytic properties but differ in the kinetics and thermodynamics of association and dissociation of the cofactor NAD (+). The binding of NAD (+) and NADH in SAHH appears structurally to be mediated by helix 18, formed by seven residues near the C-terminus of the adjacent subunit. Helix-propensity estimates indicate decreasing stability of helix 18 in the order Hs-SAHH > Tc-SAHH > Ld-SAHH (from Leishmania donovani) > Pf-SAHH (from Plasmodium falciparum), which would be consistent with the previous observations. Here we report the properties of Hs-18Pf-SAHH, the human enzyme with plasmodial helix 18, and Tc-18Hs-SAHH, the trypanosomal enzyme with human helix 18. Hs-18Tc-SAHH, the human enzyme with trypanosomal helix 18, was also prepared but differed insignificantly from Hs-SAHH. Association of NAD (+) with Hs-SAHH, Hs-18Pf-SAHH, Tc-18Hs-SAHH, and Tc-SAHH exhibited biphasic kinetics for all enzymes. A thermal maximum in rate, attributed to the onset of local structural alterations in or near the binding site, occurred at 35, 33, 30, and 15 degrees C, respectively. This order is consistent with some reversible changes within helix 18 but does require influence of other properties of the "host enzyme". Dissociation of NAD (+) from the same series of enzymes also exhibited biphasic kinetics with a transition to faster rates (a larger entropy of activation more than compensates for a larger enthalpy of activation) at temperatures of 41, 38, 36, and 29 degrees C, respectively. This order is also consistent with changes in helix 18 but again requiring influence of other properties of the "host enzyme". Global unfolding of all fully reconstituted holoenzymes occurred around 63 degrees C, confirming that the kinetic transition temperatures did not arise from a major disruption of the protein structure.

Molecular dynamics simulations of domain motions of substrate-free S-adenosyl- L-homocysteine hydrolase in solution.

S-Adenosyl-L-homocysteine hydrolase (SAHH) is an enzyme regulating intracellular methylation reactions. The homotetrameric SAHH exists in an open conformation in absence of substrate, while enzyme:inhibitor complexes crystallize in the closed conformation, in which the ligands are engulfed by the protein due to an 18 degrees domain reorientation within each of the four subunits. We present a microscopic description of the structure and dynamics of the substrate-free, NAD(+)-bound SAHH in solution, based on a 15-ns molecular dynamics simulation in explicit solvent. In the trajectory, the four cofactor-binding domains formed a relatively rigid core with structure very similar to the crystal conformation. The four substrate-binding domains, located at the protein exterior, also retained internal structures similar to the crystal, while undergoing large amplitude rigid-body reorientations. The trajectory domain motions exhibited two interesting properties. First, within each subunit the domains fluctuated between open and closed conformations, while at the tetramer level 80% of the domain motions were perpendicular to the direction of the open-to-closed structural transition. Second, the domain reorientations in solution could be represented as a sum of two components, faster, with 20-50 ps correlation time and 3-4 degrees amplitude, and slower, with 8-23 ns correlation time and amplitude of 14-22 degrees . The faster motion is similar to the 1.5 cm(-1) frequency hinge-bending vibrations found in our recent normal mode analysis (Wang et al., Biochemistry 2005;44:7228-7239). The slower motion agrees with fluorescence anisotropy decay measurements, which detected a 10-20 ns domain reorientation of ca. 26 degrees amplitude in the substrate-free enzyme (Wang et al., Biochemistry 2006;45:7778-7786). Our simulations are thus in excellent agreement with experimental data. The simulations allow us to assign the observed nanosecond fluorescence anisotropy signal to fluctuations in domain orientations, and indicate that the microscopic mechanism of the motion involves rotational diffusion within a cone of 10-20 degrees . Overall, our simulation results complement the existing experimental data and provide important new insights into SAHH domain motions in solution, which play a crucial role in the catalytic mechanism of SAHH.

The antiviral drug ribavirin is a selective inhibitor of S-adenosyl-L-homocysteine hydrolase from Trypanosoma cruzi.

Ribavirin (1,2,4-triazole-3-carboxamide riboside) is a well-known antiviral drug. Ribavirin has also been reported to inhibit human S-adenosyl-L-homocysteine hydrolase (Hs-SAHH), which catalyzes the conversion of S-adenosyl-L-homocysteine to adenosine and homocysteine. We now report that ribavirin, which is structurally similar to adenosine, produces time-dependent inactivation of Hs-SAHH and Trypanosoma cruzi SAHH (Tc-SAHH). Ribavirin binds to the adenosine-binding site of the two SAHHs and reduces the NAD(+) cofactor to NADH. The reversible binding step of ribavirin to Hs-SAHH and Tc-SAHH has similar K(I) values (266 and 194 microM), but the slow inactivation step is 5-fold faster with Tc-SAHH. Ribavirin may provide a structural lead for design of more selective inhibitors of Tc-SAHH as potential anti-parasitic drugs.

Comparative kinetics of cofactor association and dissociation for the human and trypanosomal S-adenosylhomocysteine hydrolases. 1. Basic features of the association and dissociation processes.

The S-adenosyl-l-homocysteine (AdoHcy) hydrolases catalyze the reversible conversion of AdoHcy to adenosine and homocysteine, making use of a catalytic cycle in which a tightly bound NAD+ oxidizes the 3-hydroxyl group of the substrate at the beginning of the cycle, activating the 4-CH bond for elimination of homocysteine, followed by Michael addition of water to the resulting intermediate and a final reduction by the tightly bound NADH to give adenosine. The equilibrium and kinetic properties of the association and dissociation of the cofactor NAD+ from the enzymes of Homo sapiens (Hs-SAHH) and Trypanosoma cruzi (Tc-SAHH) are qualitatively similar but quantitatively distinct. Both enzymes bind NAD+ in a complex scheme. The four active sites of the homotetrameric apoenzyme appear to divide into two numerically equal classes of active sites. One class of sites binds cofactor weakly and generates full activity very rapidly (in less than 1 min). The other class binds cofactor more strongly but generates activity only slowly (>30 min). In the case of Tc-SAHH, the final affinity for NAD+ is roughly micromolar and this affinity persists as the equilibrium affinity. In the case of Hs-SAHH, the slow-binding phase terminates in micromolar affinity also, but over a period of hours, the dissociation rate constant decreases until the final equilibrium affinity is in the nanomolar range. The slow binding of NAD+ by both enzymes exhibits saturation kinetics with respect to the cofactor concentration; however, binding to Hs-SAHH has a maximum rate constant around 0.06 s-1, while the rate constant for binding to Tc-SAHH levels out at 0.006 s-1. In contrast to the complex kinetics of association, both enzymes undergo dissociation of NAD+ from all four sites in a single first-order reaction. The equilibrium affinities of both Hs-SAHH and Tc-SAHH for NADH are in the nanomolar range. The dissociation rate constants and the slow-binding association rate constants for NAD+ show a complex temperature dependence with both enzymes; however, the cofactor always dissociates more rapidly from Tc-SAHH than from Hs-SAHH, the ratio being around 80-fold at 37 degrees C, and the cofactor binds more rapidly to Hs-SAHH than to Tc-SAHH above approximately 16 degrees C. These features present an opening for selective inhibition of Tc-SAHH over Hs-SAHH, demonstrated with the thioamide analogues of NAD+ and NADH. Both analogues bind to Hs-SAHH with approximately 40 nM affinities but much more weakly to Tc-SAHH (0.6-15 microM). Nevertheless, both analogues inactivated Tc-SAHH 60% (NAD+ analogue) or 100% (NADH analogue) within 30 min, while the degree of inhibition of Hs-SAHH approached 30% only after 12 h. The rate of loss of activity is equal to the rate of dissociation of the cofactor and thus 80-fold faster at 37 degrees C for Tc-SAHH.

Effects of ligand binding and oxidation on hinge-bending motions in S-adenosyl-L-homocysteine hydrolase.

Domain motions of S-adenosyl-l-homocysteine (AdoHcy) hydrolase have been detected by time-resolved fluorescence anisotropy measurements. Time constants for reorientational motions in the native enzyme were compared with those for enzymes where key residues were altered by site-directed mutation. Mutations M351P, H353A, and P354A were selected in a hinge region for motion between the open and closed forms of the enzyme, as identified in a previous normal-mode study [Wang et al. (2005) Domain motions and the open-to-closed conformational transition of an enzyme: A normal-mode analysis of S-adenosyl-l-homocysteine hydrolase, Biochemistry 44, 7228-7239]. In wild-type, substrate-free AdoHcy hydrolase (NAD(+) cofactor in each subunit), reorientational motions were detected on time scales of 10-20 and 80-90 ns. The faster motion is attributed to the domain motion, and the slower motion is attributed to the tumbling of the enzyme. The domain motion was also detected for the enzyme complexes E(NADH/3'-keto-adenosine) and E(NAD(+)/3'-deoxyadenosine) but was absent for the complex E(NADH/3'-keto-neplanocin A). The results indicate that AdoHcy hydrolase exists in equilibrium of open and closed structures, with the equilibrium shifted toward the more mobile open form for the substrate-free enzyme, E(NAD(+)), and for intermediates formed early in the catalytic cycle after substrate binding or formed late prior to product release, E(NAD(+)/ligand). However, the strong inhibitor neplanocin A upon binding undergoes oxidation, forming the complex E(NADH/3'-keto-neplanocin). For this complex, which is analogous to the enzyme complex with the central catalytic intermediate, the equilibrium was shifted toward the more rigid closed form. A similar pattern was observed for M351P and P354A mutants. In contrast, the domain motion could not be detected, either in the absence or presence of ligands or with the cofactor in either the oxidized or reduced state, for the H353A protein, suggesting that this mutation changes the hinge-bending dynamics of the enzyme.

Effect of N-1 and N-2 residues on peptide deamidation rate in solution and solid state.

The deamidation kinetics of 7 model peptides (VYPNGA, VYGNGA, VFGNGA, VIGNGA, VGGNGA, VGPNGA, and VGYNGA) were studied at 70 degrees C in pH 10 buffer solutions and at 70 degrees C and 50% relative humidity in lyophilized solid formulations containing polyvinyl pyrrolidone (PVP). The disappearance of the model peptides from solution and solid-state formulations followed apparent first-order kinetics, proceeding to completion in solution. In the solid state, the reactions showed plateaus with approximately 10% to 30% of the model peptides remaining; this was thought to be due to reversible complexation of the peptides and the PVP followed by slow dissociation of the complexes. The residues immediately N-terminal to asparagine (N-1, N-2) influenced the rate of deamidation significantly in the solid state but had minimal effect in solution. Increases in the volume and hydrophobicity of the N-1 and N-2 residues decreased the rate of deamidation in the solid state, but neither parameter alone adequately accounted for the observed effects. An empirical model using a linear combination of volume and hydrophobicity was developed; it showed that the influences of the volume and the hydrophobicity of the residues in the N-1 and N-2 positions are approximately equally important for the N-1 and N-2 residues.

Deamidation of model beta-turn cyclic peptides in the solid state.

To investigate the importance of secondary structure on peptide deamidation in the solid state, two cyclic beta-turn peptides and their linear analogs were used as models of Asn residues in structured and unstructured domains, and incorporated into poly(vinyl pyrrolidone) (PVP)-based lyophilized solids. The secondary structure of the model peptides was determined in solution and the solid state using a combination of nuclear magnetic resonance (NMR) spectroscopy, circular dichroism (CD), and Fourier transform infrared (FTIR) spectroscopy. The model beta-turn cyclic peptides were found to be type II beta-turns while the linear analogs were determined to be predominantly unstructured. Quantitatively, the cyclic peptides consisted of approximately 80% beta-turn while the linear analogs contained only 30%-35% beta-turn. To characterize the solid environment, T(g), and moisture content of the solid-state formulations were determined. Accelerated stability studies were conducted in the solid state at 37 degrees C using formulations lyophilized from solutions at pH 8.8 (0.1 M borate buffer). The effect of matrix mobility on solid-state deamidation was investigated by altering the moisture content through variation of relative humidity or the addition of a plasticizer. Cyclic peptides degraded 1.2-8 times slower than the linear analogs under all of the conditions studied. The observed rate constants, however, for all of the peptides decreased dramatically (four orders of magnitude) in the glassy solids. This suggests the greater importance of matrix mobility in solid-state degradation. Molecular dynamics (MD) simulations were also performed to explore the low energy, preferred state of the peptides, and determine the structure around the beta-turn.

Effects of sucrose and mannitol on asparagine deamidation rates of model peptides in solution and in the solid state.

Asparagine (Asn) degradation kinetics in two model peptides, Gly-Gln-Asn-Gly-Gly (GQNGG) and Val-Tyr-Pro-Asn-Gly-Ala (VYPNGA), were studied at 50 degrees C in pH 7 buffer solutions in the presence and absence of 5% (w/v) sucrose or mannitol and at 50 degrees C and 30% relative humidity in solid samples lyophilized from these solutions. Solid formulations were characterized using Karl Fischer coulometric titration, thermal gravimetric analysis (TGA), differential scanning calorimetry (DSC), Fourier-transform infrared spectrometry (FTIR), and solid-state nuclear magnetic resonance (NMR) spectroscopy. GQNGG and VYPNGA showed similar pseudo first-order deamidation rates in solution in the absence of sucrose and mannitol. Adding 5% sucrose or mannitol decreased the rates by no more than 17%. The model peptides degraded 2- to 80-fold more slowly in the solid formulations of sucrose and mannitol than in 5% solutions of these carbohydrates. Ratios of deamidation rates of the model peptides depended upon the solid matrix. In the mannitol solid, the ratio of deamidation rates of GQNGG and VYPNGA (GQNGG:VYPNGA) was approximately 8, while in the sucrose solid, the model peptides deamidated at similar rates (GQNGG:VYPNGA congruent with 1). DSC showed the mannitol formulations to be largely amorphous immediately after lyophilization with some ordered, crystalline-like structure; the extent of ordered structure increased during storage as shown by FTIR and ssNMR. In contrast, the sucrose formulation was largely amorphous after lyophilization and remained so during storage. Together, the results showed that 5% sucrose or mannitol in solution does not significantly change the rates of Asn deamidation of the model peptides, while sucrose stabilizes the model peptides against deamidation more than mannitol in the solid state.

Are L-adenosine and its derivatives substrates for S-adenosyl-L-homocysteine hydrolase?

Moffatt oxidation of 2',3'-O-isopropylidene-L-adenosine and treatment of the resulting crude 5'-aldehyde with hydroxylamine followed by deprotection gave L-adenosine 5'-carboxaldehyde oximes, whose enantiomers are known to be potent inhibitors of S-adenosyl-L-homocysteine (AdoHcy) hydrolase. The L-adenosine and its 5'-aldehyde oxime derivatives were found to be inactive as inhibitors of AdoHcy hydrolase. Docking calculations showed that binding of L-adenosine to AdoHcy hydrolase is weaker (higher energy) and less specific (larger number of clusters) compared to D-Ado.

Domain motions and the open-to-closed conformational transition of an enzyme: a normal mode analysis of S-adenosyl-L-homocysteine hydrolase.

The structure and fluctuations of the enzyme S-adenosyl-L-homocysteine hydrolase (SAHH) are analyzed in an effort to explain its biological function. Besides the previously identified open structure, characteristic of the substrate-free enzyme, we find two distinct structures in enzyme-inhibitor complexes, the closed and closed-twisted conformers. Both closed conformers differ from the open form by a hinge bending motion of two large domains within each subunit, which isolate the inhibitor bound in the active site from the bulk solvent. The closed-twisted form further differs from the closed form by a rigid body twist of the two-subunit dimers. The local structural fluctuations of SAHH are analyzed by performing block normal mode analysis of the tetrameric enzyme in its three forms. For the open form, we find that the four lowest-frequency normal modes, corresponding to the collective motions of the protein with the largest amplitudes, are essentially combinations of the hinge bending deformations of the individual subunits. Thus, the mechanical properties of the open structure of SAHH lead to the presence of structural fluctuations in the direction of the open-to-closed conformational transition. A candidate for such a motion has been observed in previous fluorescence depolarization studies of the enzyme. Both structural and normal mode analyses indicate that residues 180-190 and 350-356 form hinge regions, connecting large domains which tend to move as rigid bodies in response to interactions with substrate, intermediates, and the product of the enzymatic reactions. We propose that these hinge regions play a crucial role in the enzymatic mechanism of SAHH. In contrast to the open form, normal mode calculations for the closed conformations show strong coupling of the hinge bending motions of the individual subunits to each other and to other low-frequency vibrations. Thus, information about structural changes related to reaction progress in one active site may be mechanically transmitted to other subunits of the protein, explaining the cooperativity found in the enzyme kinetics.

Stereoselective synthesis of sugar-modified enyne analogues of adenosine and uridine. Interaction with S-adenosyl-L-homocysteine hydrolase and antiviral and cytotoxic effects.

Sonogashira coupling of (E)-6'-iodohomovinyl nucleosides 1 with (trimethylsilyl)acetylene gave the conjugated 8'-(trimethylsilyl)enyne derivatives of the adenosine 2a and uridine 2b with expected E-stereochemistry. Desilylation of 2a,b with tetrabutylammonium fluoride followed by treatment with N-iodosuccinimide/AgNO(3) afforded 8'-iodoenynes 4a,b. Analogous coupling of (Z)-6'-iodohomovinyl nucleosides 7a,b produced (Z)-8'-(trimethylsilyl)enynes 8a,b, which were deprotected with aqueous trifluoroacetic acid to give the Z-enynes 9a,b. Stereoselective coupling of the adenosine 4'-acetylenic 11 and ethyl (Z)-3-bromoacrylate followed by deprotection gave the conjugated enyne system attached in the reverse orientation at C4' 13. Because of their diverse stereochemical attributes, deprotected enyne analogues 5a, 6a, 9a, and 13 derived from adenosine require a different vicinity for binding with S-adenosyl-l-homocysteine (AdoHcy) hydrolase and/or addition of enzyme-bound water across the conjugated enyne system. Enyne 5a and 8'-iodoenyne 6a produced time-dependent and concentration-dependent inhibition of AdoHcy hydrolase (K(i), 0.55 and 118.5 microM, respectively). No reduction in NAD(+) content of the enzyme and no iodide ion released were observed upon incubation of 6a with the enzyme, while incubation of 5a produced 30% reduction in the NAD(+) content of the enzyme. No specific antiviral activity was noted for 5a,b, 6a,b, 9a,b, and 13 against any of the viruses tested; the E-iodoenynes 6a and 6b inhibited HIV-1 virus (IC(50), 1.1 and 1.8 microM; selectivity index, 7 and 3, respectively). The E-enyne 5a showed activity against cytomegalovirus at a concentration (EC(50), 30 microM) that was 3- to 10-fold lower than the cytotoxic concentration.

Stereocontrolled synthesis of diene and enyne sugar-modified nucleosides and their interaction with S-adenosyl-L-homocysteine hydrolase.

Conjugated diene 5-7 and enyne 8 analogs derived from adenosine and uridine were synthesized employing Pd-catalyzed cross-coupling reactions.

Asparagine deamidation in recombinant human lymphotoxin: hindrance by three-dimensional structures.

The chemical stability of recombinant human lymphotoxin (rhLT) was evaluated at pH 7, 9, and 11 and 40 degrees C using quantitative tryptic map and urea-IEF methods. Degradation products were characterized by mass spectrometry. The stability of denatured rhLT protein was also evaluated to elucidate the effects of three-dimensional structures on Asn deamidation in rhLT. Two sites that underwent Asn deamidation were identified in rhLT, Asn(19) and Asn(40)-Asn(41). At pH 11 and 40 degrees C, deamidation at Asn(19) and Asn(40)-Asn(41) had half-lives of 14 +/- 4 and 80 +/- 24 days, respectively. Upon denaturation, 31- and ninefold acceleration in the degradation rates was observed at the Asn(19) and Asn(40)-Asn(41) sites, respectively. The rate of Asn(19) degradation in denatured rhLT was comparable to that of the model peptide possessing the same primary sequence as the Asn(19)-containing region in rhLT. Analysis of the rhLT crystal structure revealed that both Asn deamidation sites were located in beta-turn structures with extensive hydrogen-bonding networks created with nearby residues in the tertiary structures. The results suggested that these tertiary and secondary structures, if held true in solution, were probably responsible for the stabilization of Asn in the native rhLT protein by reducing flexibility, thus preventing adoption of the favorable conformation required for cyclic-imide formation.

Reaction of a peptide with polyvinylpyrrolidone in the solid state.

During stability studies at high temperature (70 degrees C) and low relative humidity ( approximately 0%), the recovery of an asparagine containing hexapeptide (VYPNGA) and its known deamidation products from solid polyvinylpyrrolidone (PVP) matrices was incomplete. To determine the causes of this mass loss, formulations were prepared by lyophilizing solutions containing PVP, glycerol, and the Asn-hexapeptide in pH 7.5 phosphate buffer, followed by storage at 70 degrees C and 0% relative humidity. Asn-hexapeptide loss was mono-exponential and reached a plateau at about 30% remaining. Total recovery of the peptide and its known deamidation products was approximately 30% of peptide load. Size exclusion chromatography with fluorescence detection indicated the formation of a PVP-peptide adduct that was stable in the presence of 6 M guanidine hydrochloride. Similar stability studies using N-acetyl phenylalanine, phenylalanine ethyl ester, and N-acetyl tyrosine ethyl ester demonstrated that the reaction involves the peptide N-terminus. The adduct was disrupted in the presence of carboxypeptidase-A, suggesting the formation of an amide bond between the peptide and PVP. (15)N solid-state nuclear magnetic resonance spectroscopy using (15)N-labeled valine as a model of the peptide N-terminus showed different populations of (15)N, suggesting that noncovalent peptide-polymer interactions precede amide bond formation.