the nucleoside diphosphate kinase of mycobacterium smegmatis: identification of proteins that...

The nucleoside diphosphate kinase of Mycobacteriumsmegmatis : identification of proteins that modulatespecificity of nucleoside triphosphate synthesis bythe enzyme

Sandeep Shankar, C. Douglas Hershberger and A. M.Chakrabarty *

Department of Microbiology and Immunology, Universityof Illinois College of Medicine, 835 South Wolcott Avenue,Chicago, Illinois 60612, USA.

Summary

We report the purification and characterization of theenzyme nucleoside diphosphate kinase (Ndk) fromMycobacterium smegmatis . The N-terminus of theenzyme was blocked but an internal sequence showedapprox. 70% homology with the same enzymes fromPseudomonas aeruginosa and Escherichia coli . Immo-bilization of the mycobacterial nucleoside diphosphatekinase on a Sepharose 4B matrix and passing the totalcell extract through it revealed four proteins (P 70, P65,P60, and P50, respectively) of Mr 70 kDa, 65 kDa, 60 kDaand 50 kDa that were retained by the column. Whilethe proteins of Mr 70 kDa and 50 kDa modulated theactivity of Ndk directing it towards GTP synthesis,the 60 kDa protein channelled the specificity of Ndkentirely towards CTP synthesis. The 65 kDa proteinmodulated the specificity of Ndk directing it entirelytowards UTP synthesis. The specificity for such myco-bacterial proteins towards NTP synthesis is retainedwhen they are complexed with P. aeruginosa Ndk.We further demonstrate that the P 70 protein is pyru-vate kinase and that each of the four proteins formsa complex with Ndk and alters its substrate specifi-city. Given the ubiquitous nature of Ndk in the livingcell and its role in maintaining correct ratios of intra-cellular nucleoside triphosphates, the implicationsof the occurrence of these complexes have been dis-cussed in relation to the precursor pool for cell wallbiosynthesis as well as RNA/DNA synthesis.

Introduction

Maintenance of the correct levels of nucleoside triphos-phates (NTPs) as well as their deoxy derivatives (dNTPs)

is important to all developmental processes. The enzymenucleoside diphosphate kinase (ATP:nucleoside diphos-phate phospho-transferase; E.C. 2.7.4.6; Ndk) utilizes anautophosphorylated enzyme intermediate to catalyse thetransfer of 58 terminal phosphate from NTPs to nucleosidediphosphates (NDPs) via a reversible mechanism whichis considered to be substrate non-specific (Ray and Math-ews, 1992). This enzyme is known to play a major role inmaintaining the correct levels of NTPs as well as dNTPsin the cell in not only prokaryotes but also eukaryotes. InDictyostelium discoideum, ndk gene expression appearsto be developmentally regulated and is strongly decreasedwhen a developmental cycle is induced by starvation (Wal-let et al., 1990). In Drosophila, defects in the cells of thebrain as well as the ovary result when a null mutation isintroduced in the awd gene in addition to a suicidal defectin the metaphase stage of mitosis (Biggs et al., 1990). Wehave previously reported that in Pseudomonas aeruginosathe 16 kDa form of Ndk is truncated to a 12 kDa form bya protease (Shankar et al., 1996). The truncated formis membrane associated and complexes with pyruvatekinase to specifically produce GTP as the preferred nuc-leotide (Sundin et al., 1996a). GTP plays a crucial role inregulating cellular events such as signal transduction, elon-gation steps in protein biosynthesis, tubulin formation andmalignant transformation (Pall, 1985) including as a majorcomponent in the pathway of alginate biosynthesis by pro-viding the precursor GDP-mannose (May and Chakra-barty, 1994).

With the cell wall lipids making up about 10% of the dryweight of mycobacterium (Dhariwal et al., 1976), wall andenvelope biosynthesis represents a massive effort on thepart of Mycobacterium tuberculosis and other Mycobac-teria spp. Occupying a central position in nucleic acidmetabolism as well as cell wall biosynthesis is the avail-ability of a large, albeit carefully regulated, pool of NTPsand dNTPs. However, little is known about the regulationof the maintenance of NTP pools in the case of Mycobac-teria. Because cell wall polysaccharides and glycolipidsare synthesized via nucleotide-linked sugar intermediates,and because lipid-associated polysaccharides are essen-tial components of the mycobacterial envelope (Besra andChatterjee, 1994), an important question is the mode ofsynthesis of specific NTPs during cell envelope biosynthesis

Molecular Microbiology (1997) 24(3), 477–487

Q 1997 Blackwell Science Ltd

Received 13 November, 1996; revised 10 February, 1997; accepted19 February, 1997. *For correspondence. E-mail [email protected]; Tel. (312) 9964586; Fax (312) 9966415.

m

in Mycobacteria. While the lipoarabinomannans and lipo-mannans are known to be derived from GDP-mannose(Wheeler and Ratledge, 1994), which is known to besynthesized from mannose-1-phosphate and GTP (Mayet al., 1994), much less is known about the role of specificNTPs in mycobacterial cell-wall lipid or polysaccharidesynthesis. In this report, we demonstrate the presence ofspecific proteins that modulate Ndk activity by complexformation and consequent attenuation of its specificitytowards synthesis of specific NTPs.

Results

Ammonium sulphate fractionation of the M. smegmatiscell-free supernatant reveals non-specific NTP-synthesizing activities

We have previously observed (S. Shankar, unpublished)that to be able to assay a system for its ability to synthe-size NTPs, fractionation to remove any putative phospha-tases and/or NTPases is usually helpful. There have beenno previous reports on purification and characterization ofNDP-kinase like activities in Mycobacteria. However, suchan activity has been acknowledged to be vital for the sur-vival of the mycobacterial cell (Winder and Coughlan,1969). When using an NTP-synthesizing assay for thepurpose of evaluating any NDP kinase activities of themycobacterial cell extracts, we were unable to detect sig-nificant NDP to NTP conversion. The radioactive terminalphosphate of ATP was, however, shown to be convertedto inorganic phosphate, indicating the presence of ATPasesand/or Ndk phosphatases (Shankar et al., 1995). We there-fore decided to fractionate M. smegmatis cell-extract super-natant as well as the membrane fraction after separatingthem from each other. Ammonium sulphate fractionationof the soluble supernatant from M. smegmatis demon-strated NTP production by a 0–30% (NH4)2SO4 fractionand CTP/UTP production by a 50–70% fraction, both ofwhich could be inhibited by anti-Ndk antibodies madeagainst the P. aeruginosa enzyme (data not shown). Incontrast, fractionation of the membrane suspension showednon-specific NTP-synthesizing activity (0–30% ammo-nium sulphate fraction) which could be abolished by anti-Ndk but not by anti-pyruvate kinase (Pk) antibodies. The30–50% fraction as well as the 50–70% ammonium sul-phate fraction showed specific GTP synthesis that couldbe abolished by anti-Ndk antibodies (data not shown).The supernatant of the 70% ammonium sulphate fractionallowed specific synthesis of UTP as well as GTP, sug-gesting that various fractions may give rise to specificNTPs because of an alteration of Ndk substrate specificity.

Because modulation of NTP synthesis has previouslybeen demonstrated in P. aeruginosa as a result of complexformation between Ndk and other proteins such as Pk orthe Ras-like protein Pra (Sundin et al., 1996a; Chopade

et al., 1997), it was necessary to dissect the M. smegmatissystem in terms of multiple protein species that might beinvolved in modulating various NTP synthesis as a resultof complex formation with Ndk. To delineate the role ofputative proteins that might be associated with Ndk, wepurified the Ndk protein from M. smegmatis cell extractsand determined the affinity of other proteins to be retainedin a column where Ndk was cross-linked to Sepharosebeads.

Purification and internal sequence determination ofthe NdK of M. smegmatis

We have previously described the procedure for the puri-fication of Ndk from P. aeruginosa (Kavanaugh-Black etal., 1994). To purify M. smegmatis Ndk, we began theinitial fractionation of the 50–70% ammonium sulphate frac-tion of the crude cell extract containing the bulk of Ndkactivity on a TSK–phenyl hydrophobic interaction column(Bio-Rad). Contrary to what is observed in the case ofP. aeruginosa (Kavanaugh-Black et al., 1994), we foundthat in the case of the M. smegmatis extracts, the fractionscontaining Ndk (as confirmed by NTP-synthesizing activ-ities as well as Western blot analysis; data not shown)also contained two other strongly autophosphorylatingspecies (fractions 42–48, Fig. 1A) which the hydrophobicinteraction matrix was unable to dissociate. To dissociatevarious complexing proteins, we concentrated the Ndkpool obtained from the hydrophobic interaction columnand brought it to a final concentration of 80% ammoniumsulphate. The resulting pellet was centrifuged and resus-pended in a small volume of buffer A (see the Experimen-tal procedures) containing NDPs at a final concentration of1 mM. The suspension was immediately applied to a gel fil-tration (Sephacryl S-300) matrix and the eluted fractionsassayed for both NTP synthesis (data not shown) aswell as autophosphorylation (Fig. 1B). As is evident fromlanes 27–32, the 18 kDa NdK activity was now well sepa-rated from the 33 kDa autophosphorylating species. The85 kDa phosphorylatable species was no longer detect-able through the fraction range tested. The putative Ndkcontaining fractions were analysed for homogeneity (Fig.2A) and simultaneously analysed for cross-reactivity withP. aeruginosa anti-Ndk antibodies (Fig. 2B). The N-termi-nus of the protein could not be determined because itappeared to be blocked. However, a 10-amino-acid inter-nal sequence corresponded very well with a region foundin Ndks from P. aeruginosa, Escherichia coli, and Myxo-coccus xanthus (Fig. 2C).

Immobilization of Ndk and its complex formation withother proteins

In order to study the interaction of Ndk with other cellular

Q 1997 Blackwell Science Ltd, Molecular Microbiology, 24, 477–487

478 S. Shankar, C. D. Hershberger and A. M. Chakrabarty

proteins, we decided to immobilize Ndk to a Sepharose 4Bmatrix and pass the total cell extract through it. The objec-tive was to identify proteins from the extract that bind toNdk and modify the NTP-synthesizing activity of the M.smegmatis Ndk. The matrix was prepared as describedin the Experimental procedures. The matrix was sequen-tially treated with increasing concentrations of NH4Cl. AnNTP-synthesizing activity assay of the fractions revealedthat the fractions by themselves did not possess anysuch activity on their own under the conditions tested(data not shown).

In order to analyse if any of the eluted fractions,although not having NTP-synthesizing activities on theirown, modulated the activity of Ndk in any manner, wepre-incubated the mycobacterial Ndk with aliquots of theeluted fractions and analysed the products formed asdescribed in the Experimental procedures. As shown inFig. 3, fractions eluting at 25 mM, 150 mM, 200 mM and350 mM NH4Cl had defined effects on the Ndk activity.The fraction eluting at 25 mM as well as the fraction elutingat 150 mM NH4Cl modulated the activity of Ndk towardsGTP synthesis. The fraction eluting at 200 mM NH4Claltered the activity of Ndk towards CTP synthesis, whilethe 350 mM NH4Cl fraction shifted the activity towardsUTP synthesis. We have previously shown in the case of

P. aeruginosa that a 12 kDa truncated form of Ndk com-plexes with Pk and specifically synthesizes GTP (Sundinet al., 1996a). We were interested in determining whetherany of the proteins that we had recovered from the affinitymatrix was indeed Pk.

The fraction eluted at 25 mM NH4Cl is pyruvatekinase. To analyse if any of the fractions we obtainedwas Pk, we carried out an NTP-synthesizing reaction inthe absence and in the presence of phosphoenolpyruvate(PEP). As the results in Fig. 4 (inset) indicate, the fractioneluted at 25 mM NH4Cl (termed MsP70) showed the synth-esis of NTPs only in the presence of added PEP (lane 4)and demonstrated inhibition of this activity in the presenceof anti-Pk antibodies (lane 6). The Pk activity is also con-firmed from biochemical assays that allow production ofthe pyruvoyl phenylhydrazones (Fig. 4) and the loss ofthis activity in the presence of anti-Pk antibodies. The P.aeruginosa Pk (termed P.a Pk) was used as a positivecontrol, while non-specific proteins such as BSA wereused as negative controls (Fig. 4). While we have shownin the case of P. aeruginosa that a truncation of the16 kDa form of Ndk to a 12 kDa form is required for itscomplex formation with Pk (Sundin et al., 1996a), wehad detected no such modification in the case of the M.


Fig. 1. Purification of the NdK from M.smegmatis.A. Elution of autophosphorylating activitiesfrom the hydrophobic interaction column(TSK–phenyl). Two mg of every alternatefraction eluted from the column was incubatedwith 10 mCi [g-32P]-ATP in a 20 ml reactionreconstituted in buffer A. The reactions werestopped by the addition of 4× SDS–PAGEloading buffer. Fractions have been labelledaccording to their elution order. The sizes ofthe major autophosphorylating species havebeen indicated.B. Sephacryl S-300 chromatography of theTSK–phenyl eluted pool of Ndk activity. Thepool obtained from the TSK–phenylhydrophobic interaction column wasammonium sulphate precipitated and theprecipitate resuspended in buffer A containingeach NDP at a final concentration of 1 mM.The sample (1 ml) was applied directly to aSephacryl S-300 2 cm ×50 cm column andeluted fractions were assayed forautophosphorylation as well as NTP-synthesizing activity (data not shown).Fractions have been labelled according totheir elution order.

Mycobacterium smegmatis nucleoside diphosphate kinase 479

smegmatis Ndk. It is not known if the Ndk we have puri-fied is already modified or if mycobacteria have a differentstrategy for regulating complex formation between Ndkand other proteins. It was therefore of interest to deter-mine if the Ndk from P. aeruginosa could be modulatedby the mycobacterial proteins and if the mycobacterialNdk activity could be modulated by the P. aeruginosaNdk-complexing proteins such as Pk.

In order to see what types of proteins are present in thevarious NH4Cl fractions, the fractions were electrophor-esed on an SDS–PAGE. Basically, four proteins (P70,P65, P60 and P50, respectively) having a Mr of 70, 65, 60and 50 kDa were detected in the eluates (Fig. 5). TheP70 and P50 proteins directed the specificity of Ndk toGTP production, the P60 protein directed the specificityto CTP production, while the P65 protein directed the speci-ficity to UTP production (Fig. 3).

Mycobacterial proteins confer specificity onPseudomonas Ndk

We incubated P. aeruginosa 16 kDa as well as the 12 kDaNdk with the Ndk-complexing proteins (P70, P65, P60 andP50) from M. smegmatis and also with the P. aeruginosaPk. As indicated in Fig. 6, we found that while the P. aeru-ginosa Pk modulated the specificity of NTP synthesis toGTP by the 12 kDa form of P. aeruginosa Ndk, as pre-viously demonstrated (Sundin et al., 1996a), the mycobac-terial proteins did not distinguish between the 16 kDa (Fig.6, lanes 1–5) or the 12 kDa form (Fig. 6, lane 7–11) of theenzyme. Additionally, the P. aeruginosa Pk modulated theactivity of the mycobacterial Ndk to orientate its specificityto GTP synthesis (Fig. 6, lane 18). It is unclear whether ornot this is because the M. smegmatis Ndk has alreadyundergone some modification.

We have previously reported the use of glycerol gradi-ents to analyse complex formation between the 12 kDa


Fig. 2. A. SDS–PAGE analysis of the purified M. smegmatis NdK.Lane 1, molecular mass markers; lane 2, 1 mg purified protein; andlane 3, 5 mg of purified protein.B. Western blot analysis of the M. smegmatis Ndk for cross-reactivity against the anti-Ndk antibodies directed against theP. aeruginosa enzyme. The samples were probed using a 1:2500dilution of the primary antibodies and visualized using the ECLWestern Blot Kit (Amersham). Lane 1, 20 ng P. aeruginosa Ndk;lane 2, 20 ng M. smegmatis Ndk; lane 3, 40 ng M. smegmatisNdk; lane 4, 100 ng M. smegmatis Ndk; and lane 5, 200 ngM. smegmatis Ndk.C. N-terminus homology search and alignment of sequences fromother systems using the TBLASTN search of the National Centre forBiotechnology Information (NCBI).

Fig. 3. Reconstitution experiment using eluates from the Ndk–Sepharose affinity matrix for their ability to modulate the activityof Ndk. Two ml of the various fractions eluted from the Ndk–Sepharose column were added individually to 20 ml reactionscontaining the M. smegmatis Ndk and NDPs, and the reaction wasset up as described in the Experimental procedures. Lane 1,M. smegmatis Ndk; lane 2, wash fraction; lane 3, 10 mM NH4Cl-eluted fraction; lane 4, 25 mM NH4Cl-eluted fraction; lane 5, 50 mMNH4Cl-eluted fraction; lane 6, 100 mM NH4Cl-eluted fraction; lane 7,150 mM NH4Cl-eluted fraction; lane 8, 200 mM NH4Cl-elutedfraction; lane 9, 250 mM NH4Cl-eluted fraction; lane 10, 300 mMNH4Cl-eluted fraction; lane 11, 350 mM NH4Cl-eluted fraction; andlane 12, 400 mM NH4Cl-eluted fraction.


form of P. aeruginosa Ndk and Pk (Sundin et al., 1996a).We decided to employ this strategy to analyse the com-plex-forming ability of the four proteins that modulate theNTP-synthesizing specificity of Ndk.

Complex formation and modulation of NTP synthesisby P50, P60, P65, and P70

The principle of the glycerol-gradient centrifugation is thata protein–protein complex of higher molecular weightmigrates into fractions of higher glycerol concentrationand can be detected by Western blotting using antibodiesagainst one or more of the complexing proteins (Sundin etal., 1996a; Xie et al., 1996; Chopade et al., 1997). Asshown in Fig. 7A, P50, P60, P65 and P70 (M. smegmatisPk) form complexes individually with Ndk, altering itsmobility from a lower (5%) glycerol concentration to ahigher (60%) glycerol concentration. The NTP-synthesizingassays of these higher-glycerol-sedimenting complexesclearly demonstrated that P50 and P70 (Pk) complexes ofNdk generate predominantly GTP, while the P60–Ndkcomplex generates CTP and the P65–Ndk complex gener-ates UTP (Fig. 7B). Co-immunoprecipitation of the proteins

from the crude extracts of M. smegmatis, and subsequentwashing of the immunoprecipitate with a gradient of 0–1.0 M MgCl2 demonstrated the presence of all the fourproteins in the immunoprecipitate (data not shown), con-firming that these proteins form complexes with Ndk,with consequent changes in its substrate specificity.

N-terminal sequence determination of P70 (Pk), P50

and P60 proteins

In order to determine the nature of the four M. smegmatisproteins that modulate Ndk-substrate specificity to allowsynthesis of specific NTPs, we determined the sequenceof 10 amino acids from the N-terminal end of these pro-teins. The sequence for P65 could not be obtained as theN-terminal amino acid, and even internal amino acids,seemed to be blocked. The N-terminal sequence of P70

demonstrated its homology with Pk enzymes from othermicroorganisms, confirming its identity as Pk (Fig. 8A).The P60 protein showed homology in its N-terminus with acell-wall antigen from other mycobacteria (Fig. 8B). Inter-estingly, the P50 protein showed extensive sequenceidentity with elongation factor Tu (EF-Tu) from other


Fig. 4. Biochemical analysis of the P70 fraction shows that the protein is pyruvate kinase (Pk). The assay for Pk was performed according toLeloir and Goldemberg (1960). Five mg and 10 mg of the protein samples to be analysed were mixed with 25 ml of 10 mM PEP made up in400 mM KCl. The volume was adjusted to 100 ml after the addition of 1 mM (final conc.) UDP and 100 mM (final conc.) MgSO4. The sampleswere incubated for 15 min at 378C. Next, 150 ml of 0.1% 2,4-dinitrophenyl hydrazine (2,4-DNP) made up in 2 N HCl was added to eachreaction, followed by the addition of 200 ml of 10 N NaOH after a 5 min incubation. The product formed was measured at A 420. Antibodieswere used where indicated at a dilution of 1:500. The figure inset shows the biochemical activity of Pk with regard to its NdK activity, asdescribed in the Experimental procedures. PEP was added to a final concentration of 0.25 mM along with 0.01 mM (final conc.) MnCl2 inthe assay prior to the addition of NDPs. Known P. aeruginosa Pk was used as a positive control. The NTP-synthesizing activity of Pk isdiscernible only in the presence of PEP (Sundin et al., 1996a). Lane 1, P. aeruginosa Pk; lane 2, P. aeruginosa Pk þ PEP; lane 3, M.smegmatis P70; lane 4, M. smegmatis P70 þ PEP; lane 5, P. aeruginosa Pk þ anti-Pk antibodies; and lane 6, M. smegmatis P70 þ anti-Pkantibodies.


microorganisms (Fig. 8C), suggesting that it is likely to beEF-Tu of M. smegmatis. We have recently reported (Muk-hopadhyay et al., 1997) that P. aeruginosa EF-Tu forms acomplex with either the 16 kDa or the truncated 12 kDaNdk, modulating its substrate specificity primarily to GDP.Indeed, we have demonstrated that Ndk and EF-Tutogether are responsible for generating GTP in the P. aeru-gjnosa ribosomes that allows peptide-chain elongation(Mukhopadhyay et al., 1997). It would be of interest to

determine whether Ndk and EF-Tu (P50) of M. smegmatiscarry out a similar function in protein synthesis and whethersuch a complex may be a target for screening inhibitors ofprotein synthesis.

Discussion

It is estimated that about 1700 million people are infectedwith M. tuberculosis (Koch, 1991) worldwide. The impactof human immunodeficiency virus (HIV) infection is great-est in those populations where the prevalence of tubercu-losis (TB) infection in young adults is high. More than fivemillion people worldwide have had dual HIV and TB infec-tion. Under the most favourable conditions yet developed,M. tuberculosis divides every 18 h on average, M. smeg-matis requires 3 h for division, while E. coli requires only20 min (Wayne, 1976). Thus, a study on non-virulent M.smegmatis may provide important clues to the study ofM. tuberculosis. The ability of the mycobacterial proteinsP70, P50, P60 and P65 to form complexes with P. aerugi-nosa Ndk and the ability of P. aeruginosa Pk to form acomplex with M. smegmatis Ndk suggests that such pro-teins may be functionally conserved. Indeed, preliminaryevidence from studies in our laboratory has demonstratedthat a P. aeruginosa or M. smegmatis protein can formsimilar complexes with Ndk from an eukaryote such asyeast or human, directing the specificity of NTP synthesisto GTP or UTP. It is likely that such proteins play an impor-tant role in the regulation of energy metabolism and growthof all cells and are thus evolutionarily conserved. In vitro


Fig. 5. Homogeneity of the fractions modulating NdK activity. Thefractions that demonstrated a capability to modulate the activity ofthe mycobacterial Ndk were loaded in 5 mg amounts onto a 15%SDS–polyacrylamide gel, electrophoresed and then visualized bystaining with Coomassie brilliant blue R250. The gel was stainedfor 10 h and destained for 4 h prior to visualization. Lane 1, Broad-range molecular mass markers (Bio-Rad); lane 2, 25 mM NH4Cl-eluted fraction; lane 3, 150 mM NH4Cl-eluted fraction; lane 4,200 mM NH4Cl-eluted fraction; and lane 5, 350 mM NH4Cl-elutedfraction.

Fig. 6. Effect of the four mycobacterial proteins P70, P50, P60 and P65 that modulate mycobacterial Ndk on the specificity of NTP synthesis byP. aeruginosa Ndk. The 16 kDa as well as the truncated 12 kDa forms of P. aeruginosa Ndk (Shankar et al., 1996) were analysed for theirspecificity for NTP synthesis by the mycobacterial proteins. The NTP-synthesizing assay was carried out as described in the legend to Fig. 3.Protein concentrations were 30 pmoles in each assay. Lane 1, P. aeruginosa 16 kDa Ndk; lane 2, P. aeruginosa 16 kDa Ndk þ P70; lane 3, P.aeruginosa 16 kDa Ndk þ P50; lane 4, P. aeruginosa 16 kDa Ndk þ P60; lane 5, P. aeruginosa 16 kDa Ndk þ P65; lane 6, P. aeruginosa 16 kDaNdk þ P. aeruginosa Pk; lane 7, P. aeruginosa 12 kDa Ndk; lane 8, P. aeruginosa 12 kDa Ndk þ P70; lane 9, P. aeruginosa 12 kDa Ndk þ P50;lane 10, P. aeruginosa 12 kDa Ndk þ P60; lane 11, P. aeruginosa 12 kDa Ndk þ P65; lane 12, P. aeruginosa 12 kDa Ndk þ P. aeruginosa Pk;lane 13, M. smegmatis Ndk; lane 14, M. smegmatis Ndk þ P70; lane 15, M. smegmatis Ndk þ P50; lane 16, M. smegmatis Ndk þ P60; lane 17,M. smegmatis Ndk þ P65; lane 18, M. smegmatis Ndk þ P. aeruginosa Pk; lane 19, P. aeruginosa Pk; and lane 20, P. aeruginosa Pk þ PEP.


formation of such complexes also provides a way todesign rational inhibitors for preventing GTP synthesisand therefore growth of mycobacteria, similar to therational design of protease inhibitor of HIV.

Mycobacteria such as M. tuberculosis are known tohave a well-honed capability to protect themselves in thestressful environment of the macrophage and surviveand propagate under a diverse array of hostile factors,including a continuous low pH challenge as well as highhydrogen peroxide concentrations (Rastogi, 1991). It islikely that the mycobcterial cell envelope plays a majorpart in such survival. The biosynthesis of the complexlipids and polysaccharides present in the cell envelopeand wall of various mycobacterial species, including M.smegmatis or M. tuberculosis, involve formation of nuc-leotide-linked sugars such as UDP-galactose, GDP-man-nose, etc., which are derived from UTP or GTP. CTP isneeded for the formation of diphosphatidyl glycerolbecause CDP-diacyl glycerol is an intermediate in thisconversion (Mathur et al., 1976).

An interesting aspect of mycobacterial growth and infec-tion is the presence of a complex cell envelope (Besra andChatterjee, 1994; Wheeler and Ratledge, 1994) that isessential for cell viability. Important constituents of thiscell-wall envelope are lipoarabinomannans, lipomannans,phosphatidyl-myo-inositol mannosides, etc. The lipoara-binomannans exhibit a wide spectrum of immunoregula-tory functions including abrogation of T-cell activation, orcytokine production, thereby facilitating the survival ofmycobacteria such as M. tuberculosis within host macro-phages (Besra and Chatterjee, 1994). The formation ofmonoacyl or diacyl-PIMx (phosphatidyl inositol manno-sides), methylmannosides, mannosyl-phosphoryl-poly-prenols or mannans and arabinomannans requires GDP-mannose as an intermediate (Wheeler and Ratledge,1994). In P. aeruginosa, GDP-mannose is generated frommannose-1-phosphate and GTP by the enzyme GDP-mannose pyrophosphorylase, which is often a bifunctionalenzyme that also harbours phosphomannose isomeraseactivity (May et al., 1994). It is likely that such an enzyme


Fig. 7. A. Glycerol-gradient analysis of the complex formation between proteins P70, P65, P60 and P50 and the mycobacterial NdK. Theglycerol gradient was prepared as described previously (Sundin et al., 1996a; Xie et al., 1996) and briefly in the Experimental procedures.Lanes have been marked according to the samples they represent; anti-Ndk antibodies were used at a dilution of 1:500 in the dot blotWestern analysis. P60 formed an incomplete complex showing retention of part of Ndk in the 5% glycerol fraction.B. Activity analysis of the glycerol-gradient fractions containing the various complexes as shown in A. Alternate fractions from the glycerolgradient were used as the source of enzyme (or enzyme complex) in an NdK assay as described in the Experimental procedures. Anti-Ndkantibodies were used at a dilution of 1:1000 where indicated.Lanes 1–5 represent the Ndk–P70 complex analysis. Lane 1, Ndk/P70, 5% glycerol; lane 2, Ndk/P70, 20% glycerol; lane 3, Ndk/P70, 40%glycerol; lane 4, Ndk/P70, 60% glycerol; and lane 5, Ndk/P70, 60% glycerol fraction þ anti-Ndk antibodies.Lanes 6–10 represent the Ndk–P50 complex analysis. Lane 6, Ndk/P50, 5% glycerol; lane 7, Ndk/P50, 20% glycerol; lane 8, Ndk/P50, 40%glycerol; lane 9, Ndk/P50, 60% glycerol; and lane 10, Ndk/P50, 60% glycerol fraction þ anti-Ndk antibodies.Lanes 11–15 represent the Ndk–P60 complex analysis. Lane 11, Ndk/P60, 5% glycerol; lane 12, Ndk/P60, 20% glycerol; lane 13, Ndk/P60,40% glycerol; lane 14, Ndk/P60, 60% glycerol; and lane 15, Ndk/P60, 60% glycerol fraction þ anti-Ndk antibodies.Lanes 16–20 represent the Ndk–P65 complex analysis. Lane 16, Ndk/P65, 5% glycerol; lane 17, Ndk/P65, 20% glycerol; lane 18, Ndk/P65,40% glycerol; lane 19, Ndk/P65, 60% glycerol; and lane 20, Ndk/P65, 60% glycerol fraction þ anti-Ndk antibodies.Lanes 21–25 represent the mobility of native mycobacterial Ndk by itself. Lane 21, Ndk, 5% glycerol; lane 22, Ndk, 20% glycerol; lane 23,Ndk, 40% glycerol; lane 24, Ndk, 60% glycerol; and lane 25, Ndk, 5% glycerol fraction þ anti-Ndk antibodies.Lanes 26–30 represent the mobility of the P. aeruginosa Ndk when complexed with pyruvate kinase. Lane 26, Ndk/Pk, 5% glycerol; lane 27,Ndk/Pk, 20% glycerol; lane 28, Ndk/Pk, 40% glycerol; lane 29, Ndk/Pk, 60% glycerol; and lane 30, Ndk/Pk, 40% glycerol fraction þ anti-Ndkantibodies. Note that unlike the M. smegmatis Ndk/Pk complex (Ndk/P70) which sediments at 60% glycerol (lane 4), the P. aeruginosaNdk/Pk complex sediments at 40% glycerol (lane 28).


is involved in GDP-mannose synthesis in Mycobacterium;this would necessitate synthesis of large amounts of GTPduring cell-envelope formation. It is interesting to note thattwo specific proteins, P70 and P50, are involved in directingthe specificity of M. smegmatis Ndk from all NTPs to pre-dominantly GTP synthesis. Very little is known about thedetails of the biosynthesis of arabinomannans or otherpolysaccharides that might use CDP- or UDP-linkedsugars as intermediates and may therefore require alarge pool of CTP or UTP. Indeed, the incorporation ofUDP-[14C]-GlcNAc and UDP-[14C]-Gal in various myco-bacterial glycolipid components has recently been demon-strated (Mikusova et al., 1996). The availability of theantibodies against the M. smegmatis proteins will allowus to look for the presence of similar proteins in the virulent

strains of M. tuberculosis or even Mycobacterium lepraeby Western blotting and determine their intracellular levelsat various growth phases. The N-terminal sequence ofthese proteins will similarly allow us to isolate the corre-sponding genes from various mycobacterial species; thesecan then be inactivated and the effect of inactivation of spe-cific genes on the specificity of NTP synthesis, cell envelopecomposition and growth of mycobacteria can be evalu-ated. Similarly, it is unknown at present whether proteinssuch as P60 or P65, similar to G proteins, may bind CDPor UDP strongly and thereby make them preferentiallyavailable to the complex. Further investigation into the Km

of Ndk and its complexes with the P50, P60, P65 and P70 pro-teins with regard tospecific NDP asa substrate isunder way.

GTP is a signalling molecule and, in P. aeruginosa,large quantities of GTP are synthesized by the membraneas a result of the association of Ndk with two P. aeruginosaproteins, Pk and Pra (Chopade et al., 1997), primarily inthe stationary phase. GTP is known to be converted toguanosine tetraphosphate (ppGpp), which allows activa-tion of the stationary-phase sigma factor RpoS in E. coli(Gentry et al., 1993), facilitating its entry into the station-ary phase. An intriguing possibility is that P70 and P50,by allowing large amounts of GTP formation for cell-envel-ope development at early growth phase, also promote aslow rate of growth of Mycobacteria resembling early sta-tionary-phase growth. Mutational analyses of the genesencoding Ndk, P70, P50, P60, P65, etc., and complex-form-ing characteristics of these proteins will provide importantinsights to the mode of growth and specific NTP synthesisby various mycobacterial species, similar to the way inwhich the role of specific amino acids of Ndk has been deli-neated for alginate synthesis by P. aeruginosa (Sundin etal., 1996b). Such studies may also provide interestingclues to the biology of Mycobacteria and will probablyexplain their slow rate of growth because of the burdenof GTP and other specific nucleotide synthesis whichoccur at unusually high levels at the early log phase.

Finally, Ndk is known to be a suppressor of cancermetastasis (De la Rosa et al., 1995). Hyperexpression ofthe human ndk (nm23 H1) gene in tumour cells of highmetastatic potential has been shown to reduce the meta-static potential of such cells. How Ndk supresses thehigh invasiveness of some breast or other tumours isunknown at present. Human Ndk (of which there are twoforms, Nm23 H1 and Nm23 H2) has been studied exten-sively in isolation, but its ability to form complexes inhuman cells to modulate the specificity of NTP synthesis,particularly towards GTP, has not been reported. We haverecently observed that at least one P. aeruginosa proteinand one M. smegmatis protein, that can form complexeswith either P. aeruginosa or M. smegmatis Ndk, can alsoform complexes with yeast or human Ndk, modulatingtheir specificity from all NTP synthesis to that of UTP or


Fig. 8. Results of the N-terminus sequence analysis of themycobacterial proteins that modulate NdK activity. Proteinsequence analysis was performed at the Protein ResearchLaboratory, University of Illinois at Chicago by Dr Bob Lee usinga 477A Applied Biosystems Sequence Analyser. The homologysearch was performed using the BLAST servers of the NCBI.A. Homology results on P70.B. Homology results on P60.C. Homology results on P50.The protein P65 could not be sequenced because the N-terminusas well as certain internal sites appeared to be blocked.


GTP. GTP is a signalling molecule known to modulatetumour metastasis (Kohn and Liotta, 1995). Indeed,N-terminal-sequence determination of one of these pro-teins that direct Ndk specificity towards GTP synthesisdemonstrates its similarity to a human protein involvedin oncogenesis. Thus, studies on the characterization ofprokaryotic proteins capable of complexing with Ndksmay provide interesting clues to the identification ofhuman proteins that may modulate suppression of cancermetastasis via complex formation with human Ndk.

Experimental procedures

Separation and ammonium sulphate precipitation ofthe supernatant and the membrane fraction ofM. smegmatis

Cultures of M. smegmatis grown for 24 h in 1 l batches at 378Cwere harvested by centrifugation at 6500 × g at 48C for 15 min.The pellet from the 1 l culture was routinely suspended in10 ml of buffer A (50 mM Tris-HCl, pH 7.6, 10 mM MgCl2,1 mM dithiothreitol) and sonicated through 25 pulses of 20 sduration with a 20 s gap between pulses and a constantpulse amplitude of 25 mm. The sonicated suspension wascentrifuged at 10 000 × g for 30 min and the pellet discarded.The supernatant thus obtained is referred to as ‘superna-tant-1’. For preparation of the membrane fraction, the super-natant obtained above was centrifuged again at 42 000 × gfor 90min at 48C. The supernatant thus obtained was classifiedas ‘supernatant-2’. The pellet obtained from supernatant-2 wasresuspended in the original volume of the suspension. For thepurification as well as complexing-protein isolation proce-dures, we used the ‘supernatant-1’ as our starting material.This was necessary as we were not sure as to the localization(and/or) distribution of Ndk in the cell. Protein concentrationswere estimated using the Bio-Rad reagent and 10 mg of totalprotein was used per NTP-synthesizing assay. For the ammo-nium sulphate precipitation, we routinely started with 100 ml ofsupernatant-2 or 100 ml of the membrane suspension. Ammo-nium sulphate was added sequentially to 30%, 50% and 70%,and the precipitates were recovered as described previously(Kavanaugh-Black et al., 1994). The precipitates were resus-pended in 10 ml of buffer A and dialysed against 1000 vols ofbuffer A for 10 h prior to the assays.

NTP-synthesizing assays

The NTP-synthesizing activity of the M. smegmatis fractionsas well as the purified proteins was assayed by incubating1 mg of the material with 0.1 mM (final conc.) NDPs (CDP,UDP, GDP) and 10 mCi of [g-32P]-ATP (3000 Ci mmol¹1; Amer-sham) along with 0.1 mM ATP, in a final reaction volume of20 ml. The reaction was initiated by the addition of ATP andcontinued for 1 min at room temperature, at the end of which4 ml of 4× SDS stop buffer was added (Shankar et al., 1996).The reaction vial was vortexed briefly then 1 ml was spottedonto a polyethyleneimine–thin-layer chromatography (PEI–TLC) plate. Reaction products were chromatographed in

0.75 M KH2PO4 and visualized by autoradiography and/orquantified by densitometry, as described previously (Shan-kar et al., 1996; Sundin et al., 1996a).

Purification of NdK from M. smegmatis

The procedure followed for the purification of the mycobacter-ial Ndk was essentially that described for P. aeruginosa(Kavanaugh-Black et al., 1994) with some modifications tothe final step. Four 24 h cultures (1 l) of M. smegmatis werepelleted and lysed as described above. Supernatant-1 wasused as the starting material. To 40 ml of the supernatant,4.0 M ammonium sulphate was added with stirring to a finalconcentration of 50% and the stirring continued at 48C for1 h. The precipitate formed was centrifuged for 15 min andthe pellet discarded. Ammonium sulphate (4.0 M) was addedto the supernatant to increase the ammonium sulphate to70%. The precipitate formed was centrifuged at 10 000 ×g,48C for 15 min and the pellet obtained was resuspended in10 ml of buffer A. The sample was diluted to yield a finalammonium sulphate concentration of 1.0 M. Simultaneously,a TSK–phenyl hydrophobic interaction column (2 cm×15 cm)was equilibrated with buffer A containing 1.0 M ammoniumsulphate at a flow rate of 1.0 ml min¹1. The sample suspensionwas loaded onto the matrix and eluted with a linear gradient of1.0 M to 0 M ammonium sulphate in buffer A. Alternate frac-tions were assayed for autophosphorylation and the activefractions were pooled. Ammonium sulphate was added to afinal concentration of 80% and the pellet recovered by centri-fugation at 10 000 ×g for 15 min The pellet was resuspendedin 1 ml of buffer A containing 1 mM of each NDP, and applieddirectly to a 2.5 cm ×50 cm Sephacryl S-300 column equili-brated with buffer A at a flow rate of 18.0 ml h¹1. Then, 2 mlfractions were collected and alternate fractions were assayedfor activity as described above. Active fractions were pooledand assayed for recovery and homogeneity of NdK.

Immobilization of NdK to a Sepharose 4B matrix andanalysis of proteins retained on the matrix

Five-hundred microlitres of a 1 mg ml¹1 solution of the myco-bacterial Ndk was dialysed against 100 vols of 0.2 M NaHCO3,pH 8.5 for 10 h and kept aside. Next, 3 g of activated Sephar-ose (Sigma) was suspended in 15 ml of chilled 0.2 M NaHCO3,pH 8.5 and left stirring in the cold (48C for 15 min). The dialysedprotein sample was added to the Sepharose suspension andstirring continued gently overnight. Powdered glycine wasadded to the suspension at a final concentration of 1 M. Thematrix was sequentially treated with 100 vols each of 0.5 MNaHCO3, 0.5 M sodium acetate and 2.0 M urea. The matrixwas finally washed with 250 vols of buffer A and stored inbuffer A until use. One ml of the matrix was packed into a0.5 cm × 2.0 cm column and equilibrated with buffer A at aflow rate of 15 ml h¹1. Supernatant-1 (25 mg of total protein)was loaded onto the matrix and sequentially washed with buf-fer A, followed by NH4Cl solutions in buffer A of the followingstrengths: 10 mM, 25 mM, 50 mM, 100 mM, 150 mM, 200 mM,250 mM, 300 mM, 350 mM and 400 mM. The eluates wereimmediately dialysed individually against 1000 vols of bufferA without salt and stored at ¹208C.



Complex-formation analysis with glycerol-gradientcentrifugation

Glycerol-gradient centrifugation was carried out by a modifi-cation of the method developed by Bowman et al. (1981) asdescribed previously (Sundin et al., 1996a; Xie et al., 1996).A batch gradient of glycerol was prepared by successivelayering of 60%, 50%, 40%, 30%, 20%, 10% and 5% solutionsof glycerol in volumes of 1 ml in Beckman 10 ml ultracentrifugetubes. The protein samples to be analysed were mixed with a5% glycerol solution or layered on top of the glycerol gradientwithout loss of resolution or separation. The gradient tubeswere centrifuged at 60 000 ×g for 1 h at 48C. Samples werewithdrawn in the small-volume aliquots as originally layered.Samples were analysed by Western blotting using the P.aeruginosa anti-Ndk antibodies as the primary antibodies.

Acknowledgements

This work was supported by Public Health Service grants AI31546-03 and AI-16790-16 from the National Institutes ofHealth.

References

Besra, G.S., and Chatterjee, D. (1994) Lipids and carbo-hydrates of Mycobacterium tuberculosis. In Tuberculosis:Pathogenesis, Protection and Control. Bloom, B.R. (ed.).Washington, DC: American Society for Microbiology, pp.285–306.

Biggs, J., Hersperger, E., Steeg, P.S., Liotta, L.A., andShearn, A. (1990) A Drosophila gene that is homologousto a mammalian gene associated with tumor metastasiscodes for a nucleoside diphosphate kinase. Cell 63:933–940.

Bowman, B.J., Blasco, B.F., and Slayman, S.W. (1981) Pur-ification and characterization of the plasma membraneATPase of Neurospora crassa. J Biol Chem 256: 12343–12349.

Chopade, B.A., Shankar, S., Sundin, G.W., Mukhopadhyay,S., and Chakrabarty, A.M. (1997) Characterization of amembrane-associated Pseudomonas aeruginosa Ras-likeprotein Pra: a GTP-binding protein that forms complexeswith truncated nucleoside diphosphate kinase and pyru-vate kinase to modulate GTP synthesis. J Bacteriol 179:2181–2188.

De la Rosa, A., Williams, R.L., and Steeg, P.S. (1995) Nm23/nucleoside diphosphate kinase: toward a structural andbiochemical understanding of its biological functions.Bioessays 17: 53–62.

Dhariwal, K.R., Chander, A., and Venkitasubramanian, T.A.(1976) Alterations in the lipid constituents during growthof Mycobacterium smegmatis CDC46 and Mycobacteriumphlei. Microbios 16: 169–182.

Gentry, D.R., Hernandez, V.J., Nguyen, L.H., Jensen, D.B.,and Cashel, M. (1993) Synthesis of the stationary phasesigma factor ss is positively regulated by ppGpp. J Bac-teriol 175: 7982–7989.

Kavanaugh-Black, A., Connolly, D.M., Chughani, S.A., andChakrabarty, A.M. (1994) Characterization of nucleoside

diphosphate kinase from Pseudomonas aeruginosa: com-plex formation with succinyl CoA synthetase. Proc NatlAcad Sci USA 91: 5883–5887.

Koch, A. (1991) The global tuberculosis situation and the newcontrol strategy of the World Health Organization. Tubercle72: 1–6.

Kohn, E.C., and Liotta, L.A. (1995) Molecular insights intocancer invasion: strategies for prevention and interven-tion. Cancer Res 55: 1856–1862.

Leloir, C.F., and Goldemberg, S.H. (1960) Synthesis of glyco-gen form uridine diphosphate glucose in liver. J Biol Chem235: 919–923.

Mathur, A.K., Suryanarayana Murthy, P., Sharia, G.S., andVenkitasubramanian, T.A. (1976) Studies on cardiolipinbiosynthesis in Mycobacterium smegmatis. Can J Micro-biol 22: 354–358.

May, T.B., and Chakrabarty, A.M. (1994) Pseudomonasaeruginosa: genes and enzymes of alginate biosynthesis.Trends Microbiol 2: 151–157.

May, T.B., Shinabarger, D., Boyd, A., and Chakrabarty, A.M.(1994) Identification of amino acid residues involved inthe activity of phosphomannose isomerase-guanosine 58-diphospho-D-mannose pyrophosphorylase. J Biol Chem269: 4872–4877.

Mikusova, K., Mikus, M., Besra, G.S., Hancock, I., andBrennan, P.J. (1996) Biosynthesis of the linkage regionof the mycobacterial cell wall. J Biol Chem 271: 7820–7828.

Mukhopadhyay, S., Shankar, S., Walden, W., and Chakra-barty, A.M. (1997) The elongation factor Tu from P. aeru-ginosa: complex formation with nucleoside diphosphatekinase modulates ribosomal GTP synthesis and peptidechain elongation. J Biol Chem, in press.

Pall, M.L. (1985) GTP: a central regulator of cellullaranabolism. Curr Top Cell Regul 25: 1–20.

Rastogi, N. (1991) Recent observations concerning structureand function relationships in the mycobacterial cell envel-ope: elaboration of a model in terms of mycobacterialpathogenicity, virulence and drug resistance. Res Micro-biol 142: 464–476.

Ray, N.B., and Mathews, C.K. (1992) Nucleoside diphos-phate kinase: a functional link between intermediate meta-bolism and nucleic acid syntheis. Curr Top Cell Regul 33:343–357.

Shankar, S., Kavanaugh-Black, A., Kamath, S., and Chakra-barty, A.M. (1995) Characterization of a phosphoproteinphosphatase for the phosphorylated form of nucleosidediphosphate kinase from Pseudomonas aeruginosa. JBiol Chem 270: 28246–28250.

Shankar, S., Kamath, S., and Chakrabarty, A.M. (1996) Twoforms of the nucleoside diphosphate kinase of Pseudomo-nas aeruginosa 8830: altered specificity of NTP synthesisby the cell-membrane associated form of the truncatedenzyme. J Bacteriol 178: 1777–1781.

Sundin, G.W., Shankar, S., Chughani, S.A., Chopade, B.A.,Kavanaugh-Black, A., and Chakrabarty, A.M. (1996a)Nucleoside diphosphate kinase from Pseudomonas aeru-ginosa: characterization of the gene and its role in cellulargrowth and exopolysaccharide alginate synthesis. MolMicrobiol 20: 965–979.

Sundin, G.W., Shankar, S., and Chakrabarty, A.M. (1996b)



Mutational analysis of nucleoside diphosphate kinase fromPseudomonas aeruginosa: characterization of criticalamino acid residues involved in exopolysaccharide algi-nate biosynthesis. J Bacteriol 178: 7120–7128.

Wallet, V., Mutzel, R., Troll, H., Barzu, O., Wurster, B., Veron,M., and Lacombe, M.-L. (1990) Dictyostelium nucleosidediphosphate kinase highly homologous to nm23 and Awdproteins involved in mammalian tumor metastasis andDrosophila development. J Natl Cancer Inst 82: 1199–1202.

Wayne, L.G. (1976) Dynamics of submerged growth of Myco-bacterium tuberculosis under aerobic and microaerophilicconditions. Am Rev Resp Dis 114: 807–811.

Wheeler, P.R., and Ratledge, C. (1994) Metabolism ofMycobacterium tuberculosis. In Tuberculosis: Pathogen-esis, Protection and Control. Bloom, B.R. (ed.). Washing-ton DC: American Society for Microbiology, pp. 353–385.

Winder, F.G., and Coughlan, M.P. (1969) An NTP-dependentdeoxyribonucleic acid breakdown system and the effect ofiron limitation on the activity of this system. Biochem J 111:679–687.

Xie, Z.-D., Hershberger, C.D., Shankar, S., Ye, R.W., andChakrabarty, A.M. (1996) Sigma factor–anti-sigma factorinteraction in alginate synthesis: inhibition of AlgT byMucA. J Bacteriol 178: 4990–4996.



the nucleoside diphosphate kinase of mycobacterium smegmatis: identification of proteins that...

Documents