Three-Dimensional Quantitative Structure-ActivityRelationship and Comparative Molecular Field
Analysis of Dipeptide Hydroxamic AcidHelicobacter pylori Urease Inhibitors
Hetal Mishra,1 Abby L. Parrill,1,2* and John S. Williamson1,3,4*Department of Medicinal Chemistry1, Research Institute of Pharmaceutical Sciences,3 and National Center for
Natural Products Research,4 University of Mississippi, University, Mississippi 38677, and Chemistry Department andComputational Research on Materials Institute, University of Memphis,
Memphis, Tennessee 38152-60602
Received 3 July 2001/Returned for modification 1 March 2002/Accepted 1 May 2002
A homology model of Helicobacter pylori urease was developed by using the crystal structure of urease fromKlebsiella aerogenes (EC 3.5.1.5) as a template. The acetohydroxamic acid moiety was docked into the activepocket of the enzyme model, followed by relaxation of the complex by use of molecular dynamics. The resultingconformation was used as a template to construct 24 potential dipeptide hydroxamic acid inhibitors with whichcomparative molecular field analysis (CoMFA) was performed. The resulting model provided a cross-valida-tion correlation coefficient (q2
L00) of 0.610, a conventional r2 value of 0.988, and an F (Fisher indication ofstatistical significance) value of 294.88. We were able to validate the CoMFA model by using the 50% inhibitoryconcentrations of six compounds that were not included in the construction of the model. A very good structuralcorrelation was observed between the amino acids in the model urease’s active pocket and the contour mapsderived from the CoMFA model. This correlation, accompanied by the validation supplied by use of the CoMFAdata, illustrates that the model can aid in the prediction and design of novel H. pylori urease inhibitors.
Helicobacter pylori is a gram-negative, spiral bacteriumthought to affect about 90% of the world’s population (11). Itis well accepted that H. pylori infection is etiologically associ-ated with chronic active gastritis, peptic ulcer diseases, muco-sa-associated lymphoid tissue-type gastric carcinoma, andother gastric cancers (16). Although H. pylori infection hasbeen implicated as an etiological factor in chronic gastric refluxdisease, new studies show that H. pylori infection may providea protective mechanism against such disease; however, theresults of those studies remain controversial (8, 18). Eradica-tion therapy heals gastritis and results in cure of peptic ulcerand the remission of mucosa-associated lymphoid tissue-typegastric carcinomas (22). Although most infections can be con-trolled by antibiotic therapy (17, 27), H. pylori antibiotic resis-tance is becoming somewhat commonplace (1). Antibiotic re-sistance in a microorganism as widespread as H. pylori is acause for immediate concern and warrants a dedicated searchfor the discovery of new drug therapies.
H. pylori, characterized by its strong urease activity (5), hasreceived a great deal of attention from the scientific commu-nity over the past two decades. It is now clear that for survivalthe organism requires the production of a urease enzyme tohelp produce ammonia to counteract the strong acidic envi-ronment of the stomach (19). It has been estimated that over
5% of the total protein in the cell is represented by this enzyme(12). The urease reaction not only provides an environmentwith a pH suitable for H. pylori colonization of the stomachmucosal lining but also provides the mechanism for eventualgastric wall damage that increases the overall likelihood andthe severity of gastric ulcers (20). Ureases are ubiquitous innature and are inhibited, in general, by a variety of agentsincluding fluorides (26), thiols (25), and hydroxamic acids (14).Urease-specific inhibitors are much less common. Recently,several mono-amino acid and dipeptide derivatives containinghydroxamic acid moieties were synthesized and tested for theirspecific inhibitory activities against H. pylori urease (23). Theinitial findings suggest that these derivatives are potent, spe-cific inhibitors of H. pylori urease but show little or no inhibi-tory activity against jack bean urease.
In order to explore the binding parameters associated withthese and potentially novel hydroxamic acid inhibitors targetedto the active pocket of H. pylori urease, a homology model wasdeveloped by using the urease crystal structure from Klebsiellaaerogenes (13) (EC 3.3.1.5) as a template. Acetohydroxamicacid was docked into the active pocket of the homology modeldeveloped with this urease, and the most probable configura-tion of the enzyme-inhibitor complex was assessed by molec-ular dynamics studies. Comparative molecular field analysis(CoMFA) was then carried out with a variety of dipeptidehydroxamic acid derivatives. Quantitative models obtained bythree-dimensional quantitative structure-activity relationship(QSAR) techniques like CoMFA and comparative molecularsimilarity indices analysis, in which the steric and electrostaticfields sampled at the intersections of one or more latticesspanning a specific three-dimensional region are compared,
* Corresponding authors. Mailing address for John S. Williamson:Department of Medicinal Chemistry, University of Mississippi, Uni-versity, MS 38677. Phone: (662) 915-7101. Fax: (662) 915-5638. E-mail:[email protected]. Mailing address for Abby L. Parrill: ChemistryDepartment and Computational Research on Materials Institute, Uni-versity of Memphis, Memphis, TN 38152-6060. Phone: (901) 678-2638.Fax: (901) 678-3447. E-mail: [email protected].
2613
have shown unprecedented accuracy in predicting specificstructure-activity relationships (15).
We have developed by CoMFA a model of 24 dipeptidehydroxamic acid derivatives, using the conformations of struc-tural ligands based on the acetohydroxamic acid-enzyme com-plex obtained by homology modeling, docking, and finally,molecular dynamics. The predictive value of the model wasevaluated and verified with data for compounds not included inthe set used to develop the original model. Overlapping of thecontour maps derived from the model obtained by CoMFAwith the amino acids associated with the enzyme active pocketresulted in a model that provides an initial conceptualizationand understanding of the steric and electrostatic requirementsfor ligand binding to and inhibition of H. pylori urease.
MATERIALS AND METHODS
Data set. A group of 24 dipeptide hydroxamic acid derivatives that wereassayed in one laboratory under the same assay conditions was selected for useas the primary set of compounds for which data were obtained. The 50% inhib-itory concentrations (IC50s) of the dipeptide derivatives were previously deter-mined by Odake et al. (23), and these data are reported in Table 1. The primarystructural variation among these compounds was the amino acid side chain.
Computational approaches. (i) Homology modeling. The amino acid sequencefor H. pylori urease was retrieved from SWISS-PROT data bank entryURE2_HELPY (5). The X-ray crystal structure of the urease of K. aerogenes,entry 2KAU (13), was obtained via the Protein Data Bank. The construction ofthe protein model was based on the homologous structure of the K. aerogenesurease, which was used as a template. Amino acid sequence alignment indicateda 61.4% residue identity between the primary structures of the urease enzymes.The three-dimensional model was constructed by copying aligned coordinates ofidentical residues, building loops, and structural refinement (10). The proteinmodeling tools available in the computer software package MOE (2000; Chem-ical Computing Group Inc. Montreal, Quebec, Canada) were used for protein
modeling. The partial equalization of orbital electronegativity (PEOE) (7)charges were calculated, and the force field was set to AMBER94 (28). This wasfollowed by geometric optimization of the protein structure, minimizing thestructure to a root mean square deviation (RMSD) gradient of 0.01 kcal/mol � Å.
(ii) Docking. The geometrically optimized protein structure was used as astarting point for docking experiments, and PEOE partial charges were calcu-lated. The acetohydroxamic acid was docked into the active pocket of the proteinby using the docking protocol of the MOE software package. Specific informa-tion identifying the amino acids involved in the formation of the urease activepocket was obtained from the literature, and this information was used to definethe docking box or volume to be explored during docking (6). The MOE softwarepackage uses grid-based potential fields to calculate interaction energies betweenflexible ligands and rigid targets. Twenty-five putative complex geometries weregenerated and optimized during the docking process by an annealing procedurewith Monte Carlo simulation (9); however, the final configuration was selectedon the basis of similarity to the crystal structure of acetohydroxamic acid asbound to the K. aerogenes urease (24). Energy minimization of the enzyme-inhibitor complex, which was selected from the docking results, was accom-plished by using the energy minimization protocol of MOE with chirality con-straints to a gradient of 2.0 kcal/mol � Å. The PEOE charges were calculatedprior to minimization of the structure using the AMBER94 force field, so thatthe PEOE charge calculation is independent of the force field.
(iii) Molecular dynamics. The molecular dynamics study was carried out byusing a constant temperature and a constant volume of the proposed enzyme-inhibitor complex obtained from docking studies. The enzyme-inhibitor complexwas heated from 1 to 300 K over 10 ps with a time step of 1 fs. Data werecollected during the following 100-ps equilibrium phase, while the temperatureresponse was fixed at 25. Snapshots were collected every 1 ps throughout the
FIG. 1. Atoms of dipeptide hydroxamic acid used for the structurealignment. Asterisks, biologically active portions of the molecules.
FIG. 2. Structure alignment of 24 dipeptide hydroxamic acids.RMSD, 0.077 Å for the atoms marked with asterisks in Fig. 1.
TABLE 1. IC50 of hydroxamic acid derivatives of dipeptidesa
Derivative no. Compound R X IC50 (�M)b
1 DP-1 H Ile 0.202 DP-2 H Leu 0.223 DP-3 H Phe 0.714 DP-4 H Met 0.315 DP-5 H Gly 0.786 DP-6 H Ala 0.797 DP-7 H Arg 0.978 DP-8 H Pro 1.19 DP-9 H Ser 1.310 DP-10 H Tyr 1.611 DP-11 H Lys 7.012 DP-12 H Glu 37.013 DP-14 Boc Ile 95.014 DP-15 Boc Leu 21.015 DP-16 Boc Met 20.016 DP-17 Boc Gly 41.017 DP-18 Boc Ala 32.018 DP-19 Boc Pro 87.019 DP-20 Boc Ser 36.020 DP-21 Boc Tyr 110.021 DP-27 Ac Phe 16.022 DP-28 Z Phe 29.023 DP-30 H Ser (Bzl) 0.2124 DP-31 H Tyr (Bzl) 1.225 DP-25 Boc Phe 2826 DP-26 Bz Phe 14
a The dipeptides were R-X-Gly-NHOH, where R is a protecting groups and Xis an amino acid. Abbreviations: Boc, tert-butoxycarbonyl; Bz, benzoyl; Ac, acetyl;Z, benzyloxycarbonyl; Bzl, benzyl.
b Data were obtained from reference 23.
2614 MISHRA ET AL. ANTIMICROB. AGENTS CHEMOTHER.
FIG. 3. (A) Overlay of the urease of K. aerogenes (white; entry 2KAU in the Protein Data Bank) with the homology model H. pylori urease(purple) (Z score, 61.7%); (B) comparison of active pockets (green, H. pylori urease; white, K. aerogenese urease).
VOL. 46, 2002 ANALYSIS OF H. PYLORI UREASE INHIBITORS 2615
equilibrium phase. Of the 100 conformations obtained in this manner, the 10conformations showing the lowest potential energy were selected for furtherminimization to a gradient of 0.01 kcal/mol � Ao, as described above.
Structure alignment. Each of the dipeptide hydroxamic acids was constructedby using the acetohydroxamic acid-H. pylori urease bound conformation as atemplate. As expected, at physiological pH the amino groups are ionized andwere modeled as such. All structures were subsequently imported to Sybyl soft-ware (version 6.7, 1995; Tripos Associates, St. Louis, Mo.), with which Gasteiger-Huckel charges were calculated. The biologically active portions of the struc-tures, as illustrated with asterisks in Fig. 1, were overlaid by using the fit atomsprotocol available in Sybyl (version 6.7) software (3, 4). The resulting structurealignment of the compounds (RMSD, 0.077 for the fitted atoms) is shown in Fig.2.
Generation and analysis of data by CoMFA. The CoMFA tools available in theSybyl (version 6.7) software were used to calculate steric and electrostatic fieldsat grid points by using the Lennard-Jones and the Coulomb potential functionsof the Tripos Standard field class. An sp3 carbon probe with a charge of �1 wasused for the grid calculations. The software program was prompted to automat-ically generate a single grid, overlapping all entered molecules and extendingpast them by at least 4 A along all axes. Steric and electrostatic cutoffs were setto 30 kcal/mol and proceeded with smooth transition, except that the electro-static contributions were dropped for each row where the steric cutoff wasreached. Both steric and electrostatic contributions were calculated by use of adistance-dependent dielectric constant of 4.0 with a smoothing function. The gridspacing was set to 2 A. Standard CoMFA scaling was applied to give eachindividual CoMFA field the same potential influence on the resulting QSAR.
Model validation. The partial-least-squares (PLS) method (3) was used toderive a linear relationship between the biological activities and the molecularfields. Leave-one-out cross validation with a 2-kcal/mol column filter was per-formed by omitting from the analysis those columns (lattice points) with anenergy variance less than 2.0 kcal/mol. This allowed determination of the opti-mum number of components associated with the lowest standard error of pre-diction.
Four components provided the highest cross-validated r2 (q2) with the loweststandard error of prediction. The cross validation was performed by dividing thetotal number of compounds into 10 groups. Only models with q2 values over 0.5and a fraction of variance greater than 0.85 for the optimum number of com-ponents were considered further. The model found to be the best by crossvalidation was used to perform the nonvalidation run.
RESULTS
Homology modeling. Our homology modeling yielded a Zscore of 66.7. This is an estimate of the statistical significanceof the alignment score (10). The homology obtained betweenthe sequences was 61.4%. An overlay of the K. aerogenes ure-ase and the H. pylori urease is illustrated in Fig. 3A.
Docking and molecular dynamics. Docking, minimization,and molecular dynamics produce an orientation of acetohy-droxamic acid in the H. pylori urease model that is very similarto that in the K. aerogenes urease crystal structure. The hydrox-amic acid binds to the nickel ions and is very important foractivity (6). The overlap of the modeled H. pylori urease-ace-tohydroxamic acid on the K. aerogenes urease complex ob-tained from the crystallographic data is shown in Fig. 3B.
CoMFA. The results of our CoMFAs by PLS analysis are asfollows: q2 was 0.610, r2 was 0.988, the standard error of pre-diction was 0.131, and F (the Fisher indication of statisticalsignificance) was 294.55 in tests with four components. Withthe 24 dipeptides, a q2
L00 of 0.609 was obtained with the singleoptimal principal component. The correlation coefficient ob-tained by the nonvalidation analysis was 0.988, and the q2
found was 0.610. It is widely accepted that a correlation with aq2 value greater than 0.5 to 0.6 is useful for the prediction ofnew biologically active molecules (2). The predicted and actualIC50s of the compounds in the test set are provided in Table 2.
DISCUSSION
CoMFA fields. In order to visualize the CoMFA results, theCoMFA contour maps were created by using the data fromPLS analysis. These maps help to explain the steric and elec-trostatic features of the compounds included in our analysis.The electrostatic contour maps are shown as red and bluepolyhedra in Fig. 4A, with the red indicating regions in whichthe electronegative groups in the ligands are associated withincreased biological activities. These can be envisioned as elec-tropositive groups within the active pocket of the receptor. Incontrast, the blue contours can be visualized as the electroneg-ative groups in the active pocket and indicate where positivelycharged groups in the ligand have improved biological activityand thus lower the IC50s. The steric contour maps (Fig. 4B)also illustrate the areas in which steric bulk on an inhibitor isfavored (green contours). The yellow polyhedra represent re-gions in which steric bulk is detrimental to binding and henceincreases the IC50s by decreasing the overall biological activity.The green polyhedra may be envisioned as hydrophobic cavi-ties in the receptor, which can accommodate hydrophobicgroups. In contrast, the yellow polyhedra represent areas al-ready occupied by the receptor, which would thus prohibit anyeffective binding.
After the development of this enzyme-acetohydroxamic acidcomplex of the H. pylori urease active pocket, we compared themaps prepared by use of the coefficients from PLS analysis ofdata obtained by CoMFA with the geometric and chemicalproperties of the binding site. Since CoMFA studies are basedsolely on a set of active compounds, these maps cannot providea picture of the binding pocket. Nevertheless, CoMFA doesreflect the regions in space where differences in the ligand-probe interaction energy can be correlated with the varianceassociated with biological activity. Therefore, CoMFA data canbe extremely useful in the design of new chemical moietiesbased on a prediction of their biological activities (2). Figure 4illustrates the contour maps prepared by use of the coefficientsfrom PLS analysis, in which the contour maps have been over-lapped on the active pocket of the H. pylori urease. Figures 4Aand B indicate that there exists considerable agreement be-tween the contour maps and the observed positions of theamino acids. Interestingly, a red negative field was observednear the nickel atoms; this is an indication that only electro-negative groups that can interact favorably with the positivecharge on the nickel ions will be capable of effectively binding
TABLE 2. Predicted and actual �Log IC50s of the compounds inthe test set
a General structures of amino acids (H-X-NHOH): for AA1, X � Phe; forAA2, X � Ile; for AA6, X � Met; for AA7, X � Leu.
b Standard error of mean for prediction, 0.028.
2616 MISHRA ET AL. ANTIMICROB. AGENTS CHEMOTHER.
FIG. 4. (A and B) electrostatic and steric maps derived by CoMFA overlaid on the H. pylori urease active pocket (acetohydroxamic acid isshown in orange). (A) Blue contours indicate where electropositive groups improve activity, and red contours show where electron-rich groupsincrease activity. (B) Yellow contours show the sterically disfavored areas of the pocket, and green contours are areas where steric bulk is predictedto have a favorable impact on the biological activity.
VOL. 46, 2002 ANALYSIS OF H. PYLORI UREASE INHIBITORS 2617
to the active pocket. The presence of a large positive fieldoccupying the Asp223, Glu222, and His221 residues, as indi-cated by a blue contour, demonstrates that a requirement fora positive charge is most consistent with the increased biolog-ical activities. We believe that this explains why the NH2 iscrucial for biological activity and why either blockage or re-moval of this group would result in compounds with increasedIC50s. The presence of Asp223 in the active pocket also sup-ports our decision to model the NH2 group in its ionized state.
In Fig. 4B, the presence of yellow polyhedra surrounding theNH2 group indicates an area of the pocket which is stericallyhindered and which most likely prohibits any effective bindingwhere large groups are present on the ligand. The green con-tour covering the hydrophobic pocket is created by amino acidresidues Met366, Met317, and Ala365; and the presence ofthese residues helps explain why ligands with polar groups thatfit this position correlate with improved biological activities.Indeed, the most active ligands possess hydrophobic substitu-ents that interact in this position. A similar interpretation wasobtained from previous QSAR studies with hydroxamic acidinhibitors of urease (21).
Our model, along with previously reported studies of ureaseinhibitors, indicates that only the inclusion of electronegativemoieties like hydroxamic acids, thiols, and phosphoramidesthat can efficiently complex with nickel ions likely results inbiologically active compounds (25). In addition, in illustratingthe requirement for an ionizable NH2 group, our model pro-vides essential information for the design of novel H. pyloriurease inhibitors. Finally, our model identifies a hydrophobicregion whose identification should provide a significant advan-tage in the design of inhibitors.
Overall, we have developed a robust three-dimensionalQSAR model that not only explains the variance in biologicalactivities of a set of specific dipeptide hydroxamic inhibitors ofH. pylori urease but also correlates with the homology model ofthis urease. In addition, we have incorporated active site-basedinformation in our CoMFA studies to aid in the prediction anddesign of structural properties for novel hydroxamic acid mol-ecules. We are synthesizing several hydroxamic acid derivativesto be evaluated against H. pylori urease in vitro and to helpfurther refine our model.
ACKNOWLEDGMENTS
A.L.P. and J.S.W. are grateful to the Chemical Computing Groupfor the MOE software package.
Financial support for this project was provided in part by grantsfrom the Centers for Disease Control and Prevention (grants CDCU50-CCU418839 and CDC UR3-CCU418652).
REFERENCES
1. Alarcon, T., D. Diego, and L. Manuel. 1999. Antibiotic resistance problemswith Helicobacter pylori. Int. J. Antimicrob. Agents 12:19–26.
2. Boehm, M., J. Stuerzebecher, and G. Klebe. 1999. Three-dimensional quan-titative structure-activity relationship analysis using comparative molecularfield analysis and comparative molecular similarity indices analysis to eluci-date selectivity differences of inhibitors binding to trypsin, thrombin andfactor Xa. J. Med. Chem. 42:458–477.
3. Clark, M., and R. D. Cramer. 1993. The probability of chance correlationusing partial least squares (PLS). III. Quant. Struct.-Act. Relationships 12:137–145.
4. Cramer, R. D., D. E. Patterson, and J. Bunce. 1988. Comparative molecularfield analysis (CoMFA): effect of shape on binding of steroids to carrierproteins. J. Am. Chem. Soc. 110:5959–5967.
5. Dunn, B. E., G. P. Campbell, G. I. Perez-Perez, and M. J. Blaser. 1990.Purification and characterization of urease from Helicobacter pylori. J. Biol.Chem. 265:9464–9469.
6. Faraci, W. S., V. Yang, D. O’Rourke, and R. W. Spencer. 1995. Inhibition ofHelicobacter pylori urease by phenyl phosphorodiamidates: mechanism ofaction. Bioorg. Med. Chem. 3:605–610.
7. Gasteiger, J., and M. Marsili. 1980. Iterative partial equalization of orbitalelectronegativity—a rapid access to atomic charges. Tetrahedron 36:3219–3222.
8. Gisbert, J. P., J. M. Pajares, and C. Losa. 1999. Helicobacter pylori andgastroesophageal reflux disease: friends or foes? Hepatogastroenterology46:1023–1029.
9. Hart, T. N., and R. J. Read. 1992. A multiple-start Monte Carlo dockingmethod. Proteins Struct. Funct. Genet. 13:206–222.
10. Hirosawa, M., Y. Totoki, M. Hoshida, and M. Ishikawa. 1995. Comprehen-sive study on iterative algorithms of multiple sequence alignment. Comput.Appl. Biol. Sci. 11:13–18.
11. Honda, S., T. Fujioka, M. Tokieda, R. Satoh, A. Nishizono, and M. Nasu.1998. Development of Helicobacter pylori-induced gastric carcinoma in Mon-golian gerbils. Cancer Res. 58:4255–4259.
12. Hu, L. T., and H. L. T. Mobley. 1990. Purification and N-terminal analysis ofurease from Helicobacter pylori. Infect. Immun. 58:992–998.
13. Jabri, E., M. B. Carr, R. P. Hausinger, and P. A. Karplus. 1995. The crystalstructure of urease from Klebsiella aerogenes. Science 268:998–1004.
14. Kobashi, K., S. Takebe, N. Terashima, and J. Hase. 1975. Inhibition ofurease activity by hydroxamic acid derivatives of amino acids. J. Biochem.(Tokyo) 77:837–843.
15. Kubinyi, H. 1997. QSAR & 3D-QSAR in drug design. Part 1. Methodology.Drug Discov. Today 2:457–467.
16. Langenberg, M. L., G. N. J Tytgat, and M. E. I. Schipper. 1984. Campy-lobacter-like organisms in the stomach of patients and healthy individuals.Lancet i:1348.
17. Logan, R. P., P. A. Gummett, B. T. Hegarty, M. M. Walker, J. H. Baron, andJ. J. Misiewicz. 1992. Clarithromycin and omeprazole for Helicobacter pylori.Lancet 340:239.
18. McColl, K. E. L., and D. Gillen. Helicobacter pylori and reflux disease.Lancet 358:9294.
19. Mobley, H. L. T., M. D. Island, and R. P. Hausinger. 1995. Molecular biologyof microbial ureases. Microbiol. Rev. 59:451–480.
20. Mobley, H. L. T. 1996. The role of Helicobacter pylori urease in the patho-genesis of gastritis and peptic ulceration. Aliment. Pharmacol. Ther. 10:57–64.
21. Munakata, K., K. Kobashi, and J. Hase. 1980. Quantitative structure-activ-ity-relationship between hydroxamic acids and their urease inhibitory po-tency. J. Pharmacobio.-Dyn. 3:457–462.
22. Nakajima, S., D. Y. Graham, T. Hattori, and T. Bamba. 2000. Strategy fortreatment of Helicobacter pylori infection in adults. Curr. Pharm. Design6:1503–1514.
23. Odake, S., T. Morikawa, M. Tsuchiya, L. Imamura, and K. Kobashi. 1994.Inhibition of Helicobacter pylori urease activity by hydroxamic acid deriva-tives. Biol. Pharm. Bull. 17:1329–1332.
24. Pearson, M. A., L. O. Michel, R. P. Hausinger, and P. A. Karplus. 1997.Structures of Cys319 variants and acetohydroxamic-inhibited Klebsiella aero-genes urease. Biochemisty 36:8164–8172.
25. Todd, M. J., and R. P. Hausinger. 1989. Competitive inhibitors of Klebsiellaaerogenes urease. Mechanisms of interaction with the nickel active site. Biol.Chem. 264:15835–15842.
26. Todd, M. J., and R. P. Hausinger. 2000. Fluoride inhibition of Klebsiellaaerogenes urease: mechanistic implication of a pseudo-uncompetitive, slow-binding inhibitor. Biochemistry 38:5389–5396.
27. Unge, P., A. Gas, K. Eriksson, B. Bergman, L. Carling, P. Ekstrom, H. Glise,H. Gnarpe, O. Junghard, C. Lindholmer, B. Sandzen, L. Strandberg, A.Stubberod, and L. Weywadt. 1993. Amoxicillin added to omeprazole pre-vents relapse in the treatment of duodenal ulcer patients. Eur. J. Gastroen-terol. Hepatol. 5:325–331.
28. Weiner, S. J., P. A. Kollman, D. A. Case, U. C. Singh, C. Ghio, G. Alagona,S. Profeta, and P. Weiner. 1984. A new force field for molecular mechanicalsimulation of nucleic acids and proteins. J. Am. Chem. Soc. 106:765.
