Steady-state Kinetics and Inhibitory Action of Antitubercular Phenothiazines on Mycobacterium tuberculosis Type-II NADH-Menaquinone Oxidoreductase (NDH-2)Takahiro Yano, Linsheng Li, Edward A. Weinstein et al.|Journal of Biological Chemistry|2006 Type-II NADH-menaquinone oxidoreductase (NDH-2) is an essential respiratory enzyme of the pathogenic bacterium Mycobacterium tuberculosis (Mtb) that plays a pivotal role in its growth. In the present study, we expressed and purified highly active Mtb NDH-2 using a Mycobacterium smegmatis expression system, and the steady-state kinetics and inhibitory actions of phenothiazines were characterized. Purified NDH-2 contains a non-covalently bound flavin adenine dinucleotide cofactor and oxidizes NADH with quinones but does not react with either NADPH or oxygen. Ubiquinone-2 (Q2) and decylubiquinone showed high electron-accepting activity, and the steady-state kinetics and the NADH-Q2 oxidoreductase reaction were found to operate by a ping-pong reaction mechanism. Phenothiazine analogues, trifluoperazine, Compound 1, and Compound 2 inhibit the NADH-Q2 reductase activity with IC50 = 12, 11, and 13 μm, respectively. Trifluoperazine inhibition is non-competitive for NADH, whereas the inhibition kinetics is found to be uncompetitive in terms of Q2. Type-II NADH-menaquinone oxidoreductase (NDH-2) is an essential respiratory enzyme of the pathogenic bacterium Mycobacterium tuberculosis (Mtb) that plays a pivotal role in its growth. In the present study, we expressed and purified highly active Mtb NDH-2 using a Mycobacterium smegmatis expression system, and the steady-state kinetics and inhibitory actions of phenothiazines were characterized. Purified NDH-2 contains a non-covalently bound flavin adenine dinucleotide cofactor and oxidizes NADH with quinones but does not react with either NADPH or oxygen. Ubiquinone-2 (Q2) and decylubiquinone showed high electron-accepting activity, and the steady-state kinetics and the NADH-Q2 oxidoreductase reaction were found to operate by a ping-pong reaction mechanism. Phenothiazine analogues, trifluoperazine, Compound 1, and Compound 2 inhibit the NADH-Q2 reductase activity with IC50 = 12, 11, and 13 μm, respectively. Trifluoperazine inhibition is non-competitive for NADH, whereas the inhibition kinetics is found to be uncompetitive in terms of Q2. The Gram-positive bacterium Mycobacterium tuberculosis (Mtb) 3The abbreviations used are: Mtb, Mycobacterium tuberculosis; NDH-2, type-II NADH-dehydrogenase; TPZ, trifluoperazine; FAD, flavin adenine dinucleotide; INH, isoniazid; Q, ubiquinone; Q2, ubiquinone-2. causes tuberculosis, one of the leading causes of morbidity and mortality in the world. Each year nine million active cases of the disease are diagnosed, accounting for three million deaths. Multidrug-resistant tuberculosis and the existence of “persistent” organisms that are tolerant to antibiotics exacerbate the problem, for which more effective and efficient treatments need to be urgently developed. Mtb is traditionally considered an obligate aerobe, yet during the normal course of events in the infectious cycle, the bacillus is able to survive in conditions of low oxygen and nutrient concentrations, such as those postulated to exist within granulomas. Mtb adapts its metabolic activity, cellular transcription, and protein expression accordingly (1Wayne L.G. Hayes L.G. Infect. Immun. 1996; 64: 2062-2069Crossref PubMed Google Scholar). It is therefore of great importance to understand how Mtb generates ATP under a variety of environmental conditions. Type-II NADH-dehydrogenase (NDH-2) is a critical enzyme in the life cycle of Mtb. The enzyme has been purified from Saccharomyces cerevisiae (2de Vries S. Grivell L.A. Eur. J. Biochem. 1988; 176: 377-384Crossref PubMed Scopus (164) Google Scholar), Escherichia coli (3Young I.G. Poulis M.I. Gene. 1978; 4: 175-179Crossref PubMed Scopus (20) Google Scholar, 4Bjorklof K. Zickermann V. Finel M. FEBS Lett. 2000; 467: 105-110Crossref PubMed Scopus (55) Google Scholar), Bacillus subtilis (5Bergsma J. Van Dongen M.B. Konings W.N. Eur. J. Biochem. 1982; 128: 151-157Crossref PubMed Scopus (54) Google Scholar), Methyloccocus capsulatus (6Cook S.A. Shiemke A.K. Arch. Biochem. Biophys. 2002; 398: 32-40Crossref PubMed Scopus (45) Google Scholar), Corynebacterium glutamicum (7Matsushita K. Otofuji A. Iwahashi M. Toyama H. Adachi O. FEMS Microbiol. Lett. 2001; 204: 271-276Crossref PubMed Google Scholar, 8Nantapong N. Kugimiya Y. Toyama H. Adachi O. Matsushita K. Appl. Microbiol. Biotechnol. 2004; 66: 187-193Crossref PubMed Scopus (26) Google Scholar), Acidianus ambicalens (9Gomes C.M. Teixeira M. Biochem. Biophys. Res. Commun. 1998; 243: 412-415Crossref PubMed Scopus (18) Google Scholar, 10Bandeiras T.M. Salgueiro C. Kletzin A. Gomes C.M. Teixeira M. FEBS Lett. 2002; 531: 273-277Crossref PubMed Scopus (26) Google Scholar), and Sulfolobus metallicus (11Bandeiras T.M. Salgueiro C.A. Huber H. Gomes C.M. Teixeira M. Biochim. Biophys. Acta. 2003; 1557: 13-19Crossref PubMed Scopus (18) Google Scholar) and is, in general, composed of a single polypeptide chain, which contains a flavin as a sole cofactor. It is noteworthy that this enzyme is not found in mitochondria. The essential role of NDH-2 in Mtb is supported by extensive evidence from biochemical (12Weinstein E.A. Yano T. Li L.S. Avarbock D. Avarbock A. Helm D. McColm A.A. Duncan K. Lonsdale J.T. Rubin H. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 4548-4553Crossref PubMed Scopus (230) Google Scholar) and transcriptional studies (13Boshoff H.I. Myers T.G. Copp B.R. McNeil M.R. Wilson M.A. Barry III, C.E. J. Biol. Chem. 2004; 279: 40174-40184Abstract Full Text Full Text PDF PubMed Scopus (499) Google Scholar), gene deletion analysis, investigation of bacterial growth in various media and under various culture conditions, and animal experiments (12Weinstein E.A. Yano T. Li L.S. Avarbock D. Avarbock A. Helm D. McColm A.A. Duncan K. Lonsdale J.T. Rubin H. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 4548-4553Crossref PubMed Scopus (230) Google Scholar). Mtb contains two copies of ndh genes (ndh and ndhA). The Mtb NDH-2 and NDH-2A share 67% sequence identity, and the genes are separated by 17 kb. Mtb NDH-2 is highly homologous to those of Mycobacterium leprae and Mycobacterium smegmatis with 91 and 81% amino acid sequence identity, respectively. A strain of Mtb in which ndh has been disrupted by transposon mutagenesis is nonviable (14Sassetti C.M. Rubin E.J. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 12989-12994Crossref PubMed Scopus (1057) Google Scholar); however, a ndhA deletion mutant of Mtb can be easily isolated (15McAdam R.A. Quan S. Smith D.A. Bardarov S. Betts J.C. Cook F.C. Hooker E.U. Lewis A.P. Woollard P. Everett M.J. Lukey P.T. Bancroft G.J. Jacobs Jr., W.R. Duncan K. Microbiology. 2002; 148: 2975-2986Crossref PubMed Scopus (112) Google Scholar). We previously demonstrated that purified NDH-2A is a competent oxidoreductase (12Weinstein E.A. Yano T. Li L.S. Avarbock D. Avarbock A. Helm D. McColm A.A. Duncan K. Lonsdale J.T. Rubin H. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 4548-4553Crossref PubMed Scopus (230) Google Scholar). Therefore, we suggest that ndhA, although present in Mtb, is probably not expressed and cannot rescue mutations in Mtb ndh. NDH-2 is likely to be the sole NADH-dehydrogenase enzyme in the Mtb respiratory chain utilized for growth in an aerobic environment. Isoniazid (INH) and ethambutol are two of the standard anti-tuberculosis medications used throughout the world. Increasing resistance to these medications is recognized as a serious global public health threat. The discovery that a mechanism of INH drug resistance in Mtb is linked to mutations in Mtb NDH-2 that decrease its activity is profoundly important (16Miesel L. Weisbrod T.R. Marcinkeviciene J.A. Bittman R. Jacobs Jr., W.R. J. Bacteriol. 1998; 180: 2459-2467Crossref PubMed Google Scholar, 17Lee A.S. Teo A.S. Wong S.Y. Antimicrob. Agents Chemother. 2001; 45: 2157-2159Crossref PubMed Scopus (139) Google Scholar, 18Vilcheze C. Weisbrod T.R. Chen B. Kremer L. Hazbon M.H. Wang F. Alland D. Sacchettini J.C. Jacobs Jr., W.R. Antimicrob. Agents Chemother. 2005; 49: 708-720Crossref PubMed Scopus (199) Google Scholar). Although measured indirectly, INH-resistant mutant NADH oxidase activity is decreased 10–50% compared with the wild-type level. It has been hypothesized that reduced NDH-2 activity in the mutants leads to an increase in the intracellular NADH/NAD+ balance and accounts for the mechanism of INH resistance (16Miesel L. Weisbrod T.R. Marcinkeviciene J.A. Bittman R. Jacobs Jr., W.R. J. Bacteriol. 1998; 180: 2459-2467Crossref PubMed Google Scholar, 18Vilcheze C. Weisbrod T.R. Chen B. Kremer L. Hazbon M.H. Wang F. Alland D. Sacchettini J.C. Jacobs Jr., W.R. Antimicrob. Agents Chemother. 2005; 49: 708-720Crossref PubMed Scopus (199) Google Scholar). The present understanding of this effect is that increased concentrations of NADH decrease binding of activated INH adduct to InhA, which is an NADH-dependent enoyl-ACP reductase necessary for mycolic acid synthesis in Mtb. The anti-mycobacterial activity of phenothiazines has been reported for a number of years (19Ratnakar P. Murthy P.S. FEMS Microbiol. Lett. 1992; 76: 73-76Crossref PubMed Google Scholar, 20Molnar J. Beladi I. Foldes I. Zentralbl Bakteriol [Orig. A]. 1977; 239: 521-526PubMed Google Scholar, 21Kristiansen J.E. Vergmann B. Acta Pathol. Microbiol. Immunol. Scand. B. 1986; 94: 393-398PubMed Google Scholar, 22Gadre D.V. Talwar V. Gupta H.C. Murthy P.S. Int. Clin. Psychopharmacol. 1998; 13: 129-131Crossref PubMed Scopus (27) Google Scholar, 23Amaral L. Kristiansen J.E. Abebe L.S. Millett W. J. Antimicrob. Chemother. 1996; 38: 1049-1053Crossref PubMed Scopus (139) Google Scholar, 24Ordway D. Viveiros M. Leandro C. Bettencourt R. Almeida J. Martins M. Kristiansen J.E. Molnar J. Amaral L. Antimicrob. Agents Chemother. 2003; 47: 917-922Crossref PubMed Scopus (196) Google Scholar). Trifluoperazine (TPZ) (Fig. 1), for example, reduced in vitro ATP synthesis in M. leprae, suggesting that one of the target sites might be the electron transport pathway itself (25Katoch V.M. Saxena N. Shivannavar C.T. Sharma V.D. Katoch K. Sharma R.K. Murthy P.S. FEMS Immunol. Med. Microbiol. 1998; 20: 99-102Crossref PubMed Google Scholar). TPZ is effective against a virulent Mtb strain H37Rv in a macrophage model of infection, and it is synergistic with both INH and rifampicin (26Crowle A.J. Douvas G.S. May M.H. Chemotherapy. 1992; 38: 410-419Crossref PubMed Scopus (129) Google Scholar, 27Reddy M.V. Nadadhur G. Gangadharam P.R. J. Antimicrob. Chemother. 1996; 37: 196-197Crossref PubMed Scopus (23) Google Scholar). In our previous investigation (12Weinstein E.A. Yano T. Li L.S. Avarbock D. Avarbock A. Helm D. McColm A.A. Duncan K. Lonsdale J.T. Rubin H. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 4548-4553Crossref PubMed Scopus (230) Google Scholar), we have shown that phenothiazines block NADH-dependent oxygen consumption by Mtb membranes. Furthermore, we have demonstrated that phenothiazines inhibit purified recombinant ndh and ndhA and hinder growth of Mtb both in culture and in a mouse model of tuberculosis. Although NDH-2 is essential for growth in Mtb and plays an important role in drug resistance, little is known about the catalytic reaction mechanism of NDH-2 nor is the mode of action of inhibitors of the enzyme well characterized. Therefore, in-depth understanding of the reaction catalyzed by NDH-2 is crucial to predicting and controlling the behavior of the organism. In the present study, the kinetic properties of the enzyme and the mode of action of phenothiazines were investigated in detail using highly active NDH-2 enzyme purified in an M. smegmatis expression system. Bacterial Culture and Conditions—E. coli strains and their plasmid-harboring derivatives were grown in Luria broth medium containing appropriate antibiotics (100 μg/ml ampicillin, 50 μg/ml hygromycin) at 37 °C. M. smegmatis and its transconjugants were aerobically grown in 7H9 medium containing 50 μg/ml hygromycin at 37 °C. Molecular Cloning of Mtb ndh and ndhA Genes and Construction of Expression Plasmids—All DNA manipulations were carried out according to standard protocols as described in Ref. 28Sambrook J. Fritsch E. Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd Ed. Cold spring Harbor Press, New York1989Google Scholar. Mtb ndh was amplified by PCR with genomic DNA from M. tuberculosis as a template using the following PCR ndh ndh PCR was carried out by of for for and for 2 The PCR of the ndh gene was by and and was under an of the from V. Laboratory The was by enzyme and DNA Expression of Mtb in M. expression was M. smegmatis by A well isolated was and in 2 of 7H9 medium containing 50 μg/ml protein was and were grown to at 37 for were in containing and at °C. of Mtb from M. were by with a in a mode with for The was at for to and The was at for The was in and and was to the at a of for and the was and for with The was by at for was to the and the was at for The was to a and with of containing by of containing and The bound were with containing and and active were was to be and the enzyme was in and at standard reaction 50 and oxidoreductase and were The reaction was by and was by following the at = at °C. The were using the the enzyme was with various concentrations of inhibitors in the reaction at for of flavin was from the purified protein in the of acid for by for The purified enzyme was to in the of at acid was to be and the was for by of the was measured with an at and an at The of flavin was by according to Ref. E. Biochem. PubMed Scopus Google Scholar. was by J. Biol. Chem. Full Text PDF PubMed Google Scholar) with a A. D. Biochem. PubMed Scopus Google Scholar) using as a protein PubMed Scopus Google Scholar) and C. Y. J. Biol. Chem. Full Text PDF PubMed Google Scholar, A. Smith J. Bacteriol. 176: PubMed Google Scholar) were carried out according to the respectively. was a from V. used were of the from Expression and of Mtb our previous investigation (12Weinstein E.A. Yano T. Li L.S. Avarbock D. Avarbock A. Helm D. McColm A.A. Duncan K. Lonsdale J.T. Rubin H. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 4548-4553Crossref PubMed Scopus (230) Google Scholar), we have expressed Mtb NDH-2 (ndh and ndhA in E. coli and have active The NADH with and were by the purified ndh and ndhA low and were not In our to highly active and we expressed Mtb ndh in M. smegmatis as described under Mtb NDH-2 was expressed and to the and be with The Mtb NDH-2 expressed with a at its was purified by a single to an highly (Fig. The purified Mtb NDH-2 was as from the DNA and recognized by the (Fig. The purified NDH-2 activity as high as which is that of the previous 2 of a highly NDH-2 enzyme was from of and of the Purified Mtb Mtb NDH-2 noteworthy The of the Mtb NDH-2 is shown in The as at and of A of was recognized at The protein was an at a was at not the of the increased that the protein was and the flavin was to and of the purified Mtb NDH-2 in the of The of the flavin two at and whereas one at was The bound flavin be with we the bound flavin to be flavin adenine dinucleotide Mtb NDH-2 expressed in M. smegmatis its catalytic activity at and was for the purified Mtb NDH-2 The NDH-2 was highly to a number of of was able to Mtb NDH-2 from the Mtb NDH-2 activity the of and which have been used for of NDH-2 from such as E. coli (3Young I.G. Poulis M.I. Gene. 1978; 4: 175-179Crossref PubMed Scopus (20) Google Scholar, 4Bjorklof K. Zickermann V. Finel M. FEBS Lett. 2000; 467: 105-110Crossref PubMed Scopus (55) Google Scholar) and C. glutamicum (7Matsushita K. Otofuji A. Iwahashi M. Toyama H. Adachi O. FEMS Microbiol. Lett. 2001; 204: 271-276Crossref PubMed Google Scholar, N. Otofuji A. C.T. Adachi O. Toyama H. Matsushita K. Biotechnol. Biochem. 2005; PubMed Scopus Google Scholar). of Mtb with these or in in of the activity of the E. and H. The purified NDH-2 its high catalytic activity in the of and be at for to one of the of the Purified Mtb the properties of the purified Mtb NDH-2, we the and Mtb NDH-2 oxidoreductase reaction with a of (Fig. The activity at and to of the We NADPH as an electron however, Mtb NDH-2 not oxidoreductase activity within a of In we not electron activity to oxygen oxidase NDH-2 from S. cerevisiae and C. as reported N. Otofuji A. C.T. Adachi O. Toyama H. Matsushita K. Biotechnol. Biochem. 2005; PubMed Scopus Google Scholar). We measured the electron using various the was found to the activity of and was by 50 TPZ to and respectively. NADH-Q2 reductase activity was to by the NADH oxidase activity of the Mtb respiratory chain is by TPZ, as demonstrated (12Weinstein E.A. Yano T. Li L.S. Avarbock D. Avarbock A. Helm D. McColm A.A. Duncan K. Lonsdale J.T. Rubin H. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 4548-4553Crossref PubMed Scopus (230) Google Scholar), the electron of the purified NDH-2 with and to the reaction with in the the the reductase activity was to TPZ and not be by high concentrations of analogues, and showed electron compared with those for The for is The electron with these are to analogues, and were reduced by Mtb NDH-2 with and respectively. electron were not to TPZ, that of these not the electron and of the reductase of the purified Mtb NDH-2 and inhibition by is as of NADH by 50 inhibition of the reductase by 50 TPZ was using NADH and was used as an electron at a of was used as an electron at a of was used as an electron at a of was used as an electron at a of not not not was used as an electron at a of were for these electron were measured using NADH and these low electron reductase were measured as described under were for these electron were measured using NADH and was used as an electron at a of these low electron reductase were measured as described under were for these electron were measured using NADH and is as of NADH inhibition of the reductase by 50 TPZ was using NADH and NADH was used as an electron at a of these low electron reductase were measured as described under not were for these electron were measured using NADH and in a by Mtb NDH-2 a we investigated the kinetic mode of the electron of Mtb NDH-2 with as an electron a of the NADH-Q2 oxidoreductase reaction with concentrations and a of The for the reaction conditions the In a (Fig. of the are not by NADH that the NADH-Q2 oxidoreductase reaction by a ping-pong reaction the NDH-2 with NADH and a We the kinetic using the following for the ping-pong reaction mechanism We the kinetic and = of TPZ Mtb (12Weinstein E.A. Yano T. Li L.S. Avarbock D. Avarbock A. Helm D. McColm A.A. Duncan K. Lonsdale J.T. Rubin H. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 4548-4553Crossref PubMed Scopus (230) Google Scholar), we have demonstrated that phenothiazines the growth of Mtb in vitro as well as in a mouse model and have biochemical evidence that the inhibit the electron activity of the Mtb NDH-2 in as well as the purified recombinant understand the mode of action of phenothiazines Mtb NDH-2, we the of phenothiazines the steady-state kinetics of the highly active purified the of the NADH-Q2 oxidoreductase activity with TPZ, Compound 1, and Compound 2 inhibit NADH-Q2 oxidoreductase activity with IC50 of 12, 11, and 13 μm, under the conditions The the of NADH with at and suggesting that of at the binding as in the the NADH-Q2 oxidoreductase activity of Mtb NDH-2 with IC50 = μm, and the activity not be by concentrations of not a of TPZ inhibition kinetics of the NADH-Q2 oxidoreductase activity in terms of The and the suggesting that the inhibition kinetics is non-competitive in terms of The inhibition kinetics with to a inhibition In a shown in the of the are at concentrations of TPZ, suggesting that the inhibition kinetics an uncompetitive of TPZ inhibition kinetics of the purified Mtb NADH-Q2 oxidoreductase were measured in the of various concentrations of TPZ concentrations used and was at and NADH was NADH was at μm, and was In this study, we highly active and recombinant enzyme by using M. smegmatis as a of the Mtb NDH-2 of the is that the Mtb NDH-2 contains a non-covalently bound as a cofactor. NDH-2 has been three and in the sequence T.M. Teixeira M. Microbiol. Biol. 2004; PubMed Scopus Google Scholar). of A have two binding and flavin are A NDH-2 such as those from E. coli and C. a bound The sequence of Mtb NDH-2 that this enzyme is as A. The biochemical properties of the Mtb NDH-2 in the present are in with those for A. The activity of the Mtb NDH-2 expressed in M. smegmatis is the previous that were expressed in E. The expression and of the highly active enzyme to the properties of Mtb NDH-2, a number of properties of the Mtb NDH-2 the of NADH with quinones and does not either NADPH or oxygen as under conditions in this NDH-2 from E. coli J.A. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar), C. glutamicum N. Otofuji A. C.T. Adachi O. Toyama H. Matsushita K. Biotechnol. Biochem. 2005; PubMed Scopus Google Scholar), and J. Biol. Med. 2003; PubMed Scopus Google Scholar) have been reported to oxygen by a single electron to oxygen Although it oxygen by electron electron to as in Mtb NDH-2, efficient respiratory activity and the purified Mtb NDH-2 is able to various analogues, and the catalytic The electron activity is to TPZ the the reductase activity with is by suggest that the Mtb NDH-2 with these quinones It is that the chain at the plays an important role in the of with the binding of the The Mtb NDH-2 can analogues, and in a however, the are those with Mtb contains in the J. Appl. Bacteriol. PubMed Scopus Google Scholar), the low catalytic of the Mtb NDH-2 with are A low of for for to the electron It be that Mtb NDH-2 in Mtb or M. smegmatis electron with E. and H. it is that the low catalytic activity with is to the of Mtb The present steady-state kinetics that the NADH-Q2 oxidoreductase reaction of the purified NDH-2 a ping-pong reaction NDH-2 with NADH and in two reaction a The mechanism that following of with NADH to and of to its binding and from an of The ping-pong reaction mechanism has been for NDH-2 from A. A. W. U. S. J. Biol. Chem. 2005; Full Text Full Text PDF PubMed Scopus Google Scholar, I. Arch. Biochem. Biophys. 2001; PubMed Scopus Google Scholar). and analysis, the kinetic cannot about the mechanism of the of Mtb NDH-2 with the two It is that NADH in the and in the the of to out the electron NADH to the NADH binding containing the Although not in the Mtb NDH-2, and one or two the and that an in amino acid and and amino acid The present that the plays an important role in binding to the Therefore, it is to that the an that a of to the of flavin from the the catalytic sites is essential for understanding of the reaction mechanism of the The expression and of the highly active NDH-2 demonstrated in this a to of Mtb NDH-2 by mutagenesis and biochemical are in our to understand the kinetic and mechanism of Mtb three in the present inhibit the Mtb NDH-2 activity with the of IC50 = μm, the that of these as shown in are composed of two in general, a and a chain to the of the not either NADH or Therefore, it is highly that phenothiazines to the NADH or binding and inhibit the TPZ to with and inhibit an during such as or It is that the binding of TPZ with of or binding of The present that a these three by the in 1, is essential for their inhibitory In with the in our inhibition with Mtb NDH-2 the of be for of more in the We of for