Molecular Microbiology (2002)

46

(1), 191–201

© 2002 Blackwell Publishing Ltd

Blackwell Science, LtdOxford, UKMMIMolecular Microbiology0950-382X Blackwell Science, 200246Original Article

The functions of OmpATbC. Raynaud et al.

Accepted 11 July, 2002. *For correspondence. E-mail jcolsto@nimr.mrc.ac.uk; Tel. (

+

44) 181 959 3666; Fax (

+

44) 181 913 8528.

†

Present address: GeneWorks Pty Ltd, PO Box 11, Rundle Mall,South Australia 5000.

The functions of OmpATb, a pore-forming protein of

Mycobacterium tuberculosis

Catherine Raynaud,

1

K. G. Papavinasasundaram,

1

Richard A. Speight,

1†

Burkhard Springer,

2

Peter Sander,

2

Erik C. Böttger,

2

M. Joseph Colston

1

* and Philip Draper

1

1

The Division of Mycobacterial Research, The National Institute for Medical Research, The Ridgeway, Mill Hill, London NW7 1AA, UK.

2

Institut für Medizinische Mikrobiologie, Universität Zurich, Gloriastrasse 30/32, 8028 Zurich, Switzerland.

Summary

The functions of OmpATb, the product of the

ompATb

gene of

Mycobacterium tuberculosis

and a putativeporin, were investigated by studying a mutant witha targeted deletion of the gene, and by observingexpression of the gene in wild-type

M. tuberculosis

H37Rv by real-time polymerase chain reaction (PCR)and immunoblotting. The loss of

ompATb

had noeffect on growth under normal conditions, but causeda major reduction in ability to grow at reduced pH. Thegene was substantially upregulated in wild-type bac-teria exposed to these conditions. The mutant wasimpaired in its ability to grow in macrophages and innormal mice, although it was as virulent as the wildtype in mice that lack T cells. Deletion of the

ompATb

gene reduced permeability to several small water-soluble substances. This was particularly evident atpH 5.5; at this pH, uptake of serine was minimal, sug-gesting that, at this pH, OmpATb might be the onlyfunctioning porin. These data indicate that OmpATbhas two functions: as a pore-forming protein withproperties of a porin, and in enabling

M. tuberculosis

to respond to reduced environmental pH. It is notknown whether this second function is related to theporin-like activity at low pH or involves a completelyseparate role for OmpATB. The involvement with pHis likely to contribute to the ability of

M. tuberculosis

to overcome host defence mechanisms and grow in amammalian host.

Introduction

Tuberculosis remains a major public health problem andis responsible for an estimated two million deaths annually(World Health Organization tuberculosis fact sheet;http://www.stoptb.org/tuberculosis/). An important pro-perty of the causative organism,

Mycobacterium tubercu-losis

, is its extreme impermeability, which is associatedwith the presence of an outer, lipidic permeability barrierin the cell envelope (Connell and Nikaido, 1994; Draper,1998). This barrier allows the organism to survive insidehost phagocytic cells such as macrophages, and alsorestricts access to hydrophilic molecules, including someimportant antibacterial agents such as

b

-lactam antibiot-ics (Chambers

et al

., 1995), which complicates treatmentof the disease. It is therefore important to understand thephysiological functioning of the outer permeability barrier,because it confers properties involved in the pathogenesisof tuberculosis. Although the mycobacterial barrier andthe outer membrane of Gram-negative bacteria are bio-chemically quite different, they perform a similar function.In both cases, the presence of such a barrier presentsa problem to the bacterium: availability of water-solublesmall molecules, which includes common nutrients, isrestricted. The solution adopted by Gram-negative organ-isms is the inclusion of porins in the outer membrane(Nikaido, 1993; 1994; Jap and Walian, 1996). These arepore-forming proteins that provide water-filled poresallowing diffusion of small (typically

<

600 kDa) hydrophilicmolecules through the barrier. There is good evidence thatmycobacteria make use of similar structures (Jarlier andNikaido, 1990; Trias

et al

., 1992; Trias and Benz, 1994;Chambers

et al

., 1995).One molecule, which could potentially play a role in

increasing permeability to small hydrophilic molecules,has been termed OmpATb (Senaratne

et al

., 1998). Thiswas identified during the sequencing of the genome of

M.tuberculosis

(Cole

et al

., 1998), because the gene thatencodes it shows some sequence homology to

ompA

of

Escherichia coli

. There is continuing debate about thefunction of

E. coli

OmpA (Sugawara and Nikaido, 1992;1994; Pautsch and Schulz, 1998). It is a major outermembrane protein that can form pores in artificial mem-branes (Sugawara and Nikaido, 1992, 1994; Saint

et al

.,1993), and it shares the characteristic

b

-barrel structure

192

C. Raynaud

et al.

© 2002 Blackwell Publishing Ltd,

Molecular Microbiology

,

46

, 191–201

of recognized porins (Ried

et al

., 1994; Pautsch andSchulz, 1998). OmpATb, similarly, contains significant

b

-sheet/

b

-barrel structure and can form pores with pro-perties typical of porins in liposomes and lipid bilayers(Mobasheri

et al

., 1998; Senaratne

et al

., 1998).Members of the OmpA family of proteins and other

outer membrane proteins of Gram-negative bacteriahave been found to play a role in pathogenicity. Thus,OmpA of

E. coli

contributes to serum resistance(Weiser and Gotschlich, 1991), invasion of endothelialcells (Prasadarao

et al

., 1996) and maturation of dendriticcells (Jeannin

et al

., 2000). OmpC of

Salmonella typhimu-rium

is thought to be involved in attachment to macro-phages (Negm and Pistole, 1999), the major outermembrane porin of

Neisseria gonorrhoeae

, PorB, is capa-ble of translocating into host cell membranes and modu-lating phagosome maturation (Mosleh

et al

., 1998), and aporin from

Pseudomonas aeruginosa

can induce apopto-sis in an epithelial cell line (Buommino

et al

., 1999). Thus,in addition to their role in permeability, these proteins mayhave other functions in host–pathogen interactions.

In the work described here, we have investigated thephysiological function of OmpATb in

M. tuberculosis

. Wefind that the protein does appear to function as a classicalporin. However, it has an additional involvement in theresponse of the bacteria to reduced pH and, at low pH,it appears to be particularly important for the uptake ofsmall, hydrophilic molecules. The creation of a mildlyacidic environment is one of the strategies used by phago-cytic cells to control phagocytosed microorganisms, and

M. tuberculosis

is exposed to such an environment,particularly after activation of the macrophage duringacquired immunity (Schaible

et al

., 1998). Thus, OmpATbmight be particularly important for an intracellular patho-gen, either as a porin that is active at low pH, or in playingsome other role for which its activity is optimized to thepH of the macrophage environment. This is supported bythe finding that, when compared with the wild-type strain,the

ompATb

knock-out mutant shows reduced growth incultured macrophages and in intravenously infected mice.

Results

Inactivation of

ompATb

by homologous recombination

To facilitate isolation of an

ompATb

knock-out, a strepto-mycin-resistant strain of

M. tuberculosis

H37Rv was trans-formed with the suicide plasmid pompA-aph-rpsL

+

, andthe transformants were selected on 7H11 media contain-ing kanamycin and streptomycin (positive and counter-selection markers respectively) and on kanamycin alone.The transformants were screened by Southern hybridisa-tion analysis to distinguish knock-outs from merodiploidsand the parental strain. Based on the restriction pattern

for

Nsi

I, it was predicted that the probe would hybridize toa 4.7 kb fragment in the wild type and to a 2.7 kb fragmentin the knock-out. Merodiploid strains carrying both thewild-type and knock-out alleles were expected to hybridizeto two fragments (2.7 kb

+

11 kb) in the case of a 3

¢

singlecross-over and to a 12 kb fragment in case of a 5

¢

singlecross-over. Southern analysis of

Nsi

I-digested DNAdemonstrated that an

ompATb

knock-out strain aroseafter a double cross-over homologous recombination, anda merodiploid strain arose after a 3

¢

single cross-overrecombination (Fig. 1A). Allelic exchange in the knock-out strain was also confirmed by Southern hybridizationanalysis of

Sty

I-digested DNA and polymerase chainreaction (PCR) amplification from the flanking regions(data not shown). Successful deletion of the

ompATb

genewas also confirmed by immunoblotting a

M. tuberculosis

cell lysate with an antibody raised against OmpATb(Fig. 1B).

OmpATb

is involved in the acid stress response of

M.

tuberculosis

A number of studies have demonstrated that the level of

Fig. 1.

Preparation of an

ompATb

deletion mutant.A. Southern blot analysis to confirm

ompATb

inactivation

.

Chromo-somal DNA was isolated from the wild type 1424 (WT), knock-out (KO) and merodiploid single cross-over (SCO) strains and digested with the restriction enzyme

Nsi

I, transferred to a Nylon N

+

membrane and hybridized to

a

-

32

P-labelled internal probe (located in the 3

¢

flanking DNA). The sizes of the hybridizing bands were determined from the migration distance of the DNA molecular marker

l

-

Hin

dIII–

Eco

RI and were consistent with the predicted pattern for DNA digested with the restriction enzyme

Nsi

I.B. Immunoblotting of lysates of wild-type

M. tuberculosis

(WT) and the

ompATb

-deleted mutant (KO). Mobility of the low-range prestained SDS-PAGE standards (Bio-Rad) as observed on the blots is indicated to the left.

The functions of OmpATb

193

© 2002 Blackwell Publishing Ltd,

Molecular Microbiology

,

46

, 191–201

expression of outer membrane proteins of Gram-negativebacteria is related to extracellular pH in addition to osmo-larity (Foster

et al

., 1994; Bang

et al

., 2000; Sato

et al

.,2000). Some proteins, such as OmpC, are upregulated,whereas others, such as OmpF, are downregulated at acidpH (Heyde and Portalier, 1987). The effect of low pH onthe transcription of

ompATb

in

M. tuberculosis

was inves-tigated using real-time PCR. For real-time PCR, a refer-ence gene is used that does not alter in expression underthe conditions being tested. In this study, we used

sigA

,which shows levels of expression in

M. tuberculosis

thatare independent of a variety of changes in growth condi-tions and also in

M. tuberculosis

grown in macrophages(Manganelli

et al

., 1999; 2001). It is clear that there is anincrease in transcription of

ompATb

in

M. tuberculosis

cultured at low pH, with expression levels correlating withincreased acidity as pH was lowered from 7.2 to 6.5, 6.0and 5.5 (Fig. 2A).

Because vacuole acidification is known to be involvedin macrophages infected with

M. tuberculosis

, we used asimilar approach to investigate the levels of

ompATb

expression in

M. tuberculosis

that had been phagocytosedby different types of macrophages (Fig. 2B). The resultsshow that expression is significantly increased in

M. tuber-culosis

growing in the human monocytic cell line THP1and in murine bone marrow macrophages.

We were also able to demonstrate that the increasedtranscription of

ompATb

in acidified medium was reflectedin an increase in levels of OmpATb protein. Total

M. tuber-culosis

cell lysates were prepared from bacteria grown at

pH 7.2 and then transferred to medium at pH 5.5 for 18 h;immunoblotting of these lysates with the OmpATb anti-body revealed a substantial increase in amounts ofOmpATb in the bacteria exposed to low pH (Fig. 3).

The demonstration that

ompATb

expression is upregu-lated in the response of

M. tuberculosis

to lowered pHsuggested that OmpATb might be involved in protectingthe organism under acidic conditions. Therefore, thegrowth of the wild type and mutant were compared atpH 7.2 and pH 5.5. The results demonstrate that, at

Fig. 2.

Expression of

ompATb

by

M. tuberculosis

exposed to low pH or isolated from macrophages.A. The effect of reduced pH on

ompATb

expression, measured by real-time quantitative PCR. The amount of

ompATb

mRNA relative to that of the normalizing gene,

sigA

, was determined by real-time quantitative RT-PCR, and the ratio at pH 7.2 was taken as 1. The values shown are the means; the error bars indicate the standard deviations.B.

ompATb

expression by

M. tuberculosis

growing inside the human monocytic cell line THP1 (white bar) or murine bone marrow-derived macrophages (black bar) measured by real-time quantitative PCR. Normalization was carried out as in (A). The values shown are means; the error bars indicate the standard deviations.

Fig. 3. Immunoblotting of cell lysates of M. tuberculosis exposed to low pH. OmpATb protein levels were markedly increased in bacteria exposed for 18 h to pH 5.5 compared with bacteria maintained throughout at pH 7.2. Each lane contains 30 mg of total protein. Mobil-ity of the low-range prestained SDS-PAGE standards (Bio-Rad) as observed on the blots is indicated on the left.

194 C. Raynaud et al.

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

pH 7.2, the growth of the mutant and wild type wereessentially identical, whereas at pH 5.5, growth of themutant was significantly impaired, although it eventuallyreached similar optical densities to those of the wild type(Fig. 4). The parental strain showed only slightly delayedgrowth (approximately 2 days) at pH 5.5 compared withpH 7.2, whereas the growth of the mutant was delayed bybetween 8 and 15 days in a series of similar experiments.

OmpATb appears to function as a porin

The role of OmpATb in increasing the permeability of M.tuberculosis to small hydrophilic molecules was investi-gated by comparing the uptake of radiolabelled serine bythe wild type and by the mutant (Fig. 5A). The mutant wassignificantly defective in the uptake of serine, indicatingthat lack of the protein reduced the permeability of thebacterial cells to hydrophilic substances. Similar resultswere obtained with glucose and glycerol, with wild-typelevels of uptake seen in the single cross-over merodiploidstrain that carries both mutant and wild-type alleles (datanot shown). With labelled glycine, however, uptake wasincreased in the mutant compared with the wild type (datanot shown), suggesting that the porin formed by OmpATbis not involved in glycine uptake, and that restrictedaccess to other hydrophilic substances may result in acompensatory increase in the uptake of glycine. None of

Fig. 4. Growth of wild-type M. tuberculosis and the ompATb mutant at pH 7.2 and pH 5.5. The wild-type (circles) and the ompATb mutant (squares) strains of M. tuberculosis were grown in Dubos broth at pH 7.2 (closed symbols, continuous lines) and pH 5.5 (open symbols, broken lines). The figure shows a representative experiment from three identical experiments.

Fig. 5. Uptake of serine by the ompATb-deleted mutant and by wild-type M. tuberculosis. In each case, the closed circles represent wild-type M. tuberculosis, and the closed squares represent the ompATb mutant. Each experi-ment was performed in quadruplicate. Experi-ment 1 was carried out over a time scale of 45 min and is shown in (A) and (B), with (A) showing uptake at pH 7.2 and (B) at pH 5.5. Experiment 2 was carried out over a time scale of 15 min and is shown in (C) and (D), with (C) showing uptake at pH 7.2 and (D) at pH 5.5. Each point represents the mean, with error bars representing the standard error of the mean. ** denotes that the results for the ompATb mutant differed significantly from those for the wild type by the paired Student t-test (P < 0.05). The results are expressed as the percentage uptake of 14C-labelled serine with the maximal uptake by wild-type M. tuberculosis at pH 7.2 (after 45 min for experiment 1 and 15 min for experi-ment 2) taken as 100%.

The functions of OmpATb 195

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

the molecules was totally excluded from the mutant, sug-gesting that, if OmpATb is involved in the transport ofsmall hydrophilic molecules across the outer permeabilitybarrier, it is unlikely to be the only such porin. Becauseequilibration across the cell wall is achieved within a mat-ter of minutes even with the slow penetration rates ofmycobacteria (Chambers et al., 1995), we next investi-gated the early kinetics of this uptake. The results(Fig. 5C) demonstrate that, even after the shortest timeinterval (4 min), there was a highly significant differencebetween uptake by the wild type and the mutant.

The finding that ompATb expression was upregulatedat low pH also prompted us to investigate serine uptakeby bacteria that had been incubated overnight at pH 5.5.The results showed that, under these conditions, therewas a highly significant difference between the wild typeand mutant (Fig. 5B and D). Although there was reduceduptake by the wild type at pH 5.5 compared with pH 7.2(compare wild-type uptake in Fig. 5A and C with that inFig. 5B and D), there was very little uptake at all by themutant at pH 5.5 in either the long term (Fig. 5B) or theshort-term assay (Fig. 5D). This was not attributable to areduction in viability in the cultures maintained at pH 5.5nor to reduced numbers of viable bacteria in the suspen-sions of the mutant compared with wild-type bacteria;preliminary experiments confirmed that the relationshipbetween optical density and colony-forming units (cfu)was the same for the mutant and the wild type, and thatmaintaining either strain at pH 5.5 did not result in areduction in cfu.

The ompATb mutant shows reduced growth in macrophages and in intravenously infected mice

The finding that ompATb expression is upregulated in M.tuberculosis growing in macrophages suggests that theprotein could contribute to intracellular survival of the bac-teria. In order to test this, cultured murine bone marrowmacrophages (Fig. 6A) and the human monocytic cell lineTHP1 (Fig. 6B), were infected with M. tuberculosis; themacrophages were lysed at various time intervals, and thenumber of viable M. tuberculosis was determined. In bothcell types, the mutant showed significantly reduced multi-plication compared with the wild type.

In order to see whether this reduced ability to grow inhost cells was also manifested during infection in vivo,mice were infected intravenously with either the mutant orthe parental strain, and growth in lungs (Fig. 7A) andspleens (Fig. 7B) was monitored. The levels of infectionachieved by the mutant were significantly reduced com-pared with those of the wild type. In a second experiment,the mutant and wild type were also compared with theompATb+ ompATb::Kmr merodiploid single cross-overstrain (Fig. 7C) at a single time point. In this experiment,a lower inoculum was used (5 ¥ 104 compared with5 ¥ 105 for the first experiment); although the infectionlevels were generally lower, again the mutant showedsignificantly reduced growth compared with the wild type,whereas the single cross-over strain, which carries bothmutant and wild type ompATb, behaved identically to thewild type. Interestingly, infection of athymic (nu/nu) mice

Fig. 6. Growth of wild-type M. tuberculosis and the ompATb mutant in macrophages. The wild-type (circles) and the ompATb mutant (squares) strains of M. tuberculosis were grown in murine bone marrow-derived macrophages (A) or the human monocytic cell line THP1 (B). ** denotes that the ompATb mutant differed significantly from the wild type by the paired Student t-test (P < 0.05).

196 C. Raynaud et al.

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

was similar for both mutant and wild type (data notshown), implying that the loss of ompATb impaired theability to resist acquired cell-mediated immunity ratherthan innate immune mechanisms.

Discussion

The presence of an outer, lipidic permeability barrier inthe cell envelope of mycobacteria restricts access tohydrophilic molecules. It is possible to quantify the per-meability of bacteria using b-lactam drugs with differinghydrophilicities. Such an approach has been used toinvestigate the permeability of Mycobacterium che-lonae (Jarlier and Nikaido, 1990) and M. tuberculosis(Chambers et al., 1995). The studies with M. chelonaedemonstrated that, although its permeability wasextremely low, about 1000 times lower than E. coli, thepermeation properties were consistent with the presenceof water-filled pores in the cell envelope. Similar studieswith M. tuberculosis suggested that it too had a lowpermeability, but that, in this case, it was approximately100 times lower than E. coli and comparable with thatof Pseudomonas aeruginosa, an organism noted for itsresistance to antibacterial agents. The presence of pro-teins able to form pores in liposomes and planar lipidbilayers has now been demonstrated clearly for the slow-growing mycobacteria M. tuberculosis (Senaratne et al.,1998; Kartmann et al., 1999) and Mycobacterium bovisBCG (Lichtinger et al., 1999). However, the quantities ofmaterial that can be isolated from bacterial cell envelopesof these slow-growing organisms are too small for theproteins to have been identified so far. A protein thatundoubtedly has the physiological functions of a porin hasbeen identified in the rapidly growing species Mycobacte-rium smegmatis (Niederweis et al., 1999; Stahl et al.,2001). The sequence of this protein and of the gene thatencodes it has been determined, but no structurallyrelated protein or homologous gene is present in M. tuber-culosis (Cole et al., 1998).

The M. tuberculosis gene ompATb was cloned andexpressed because it shared some sequence homologywith ompA of E. coli (Senaratne et al., 1998). The physi-ological function of OmpA of E. coli is not known withcertainty. It behaves as a porin in artificial membranes,but it appears to have no continuous water-filled pore, atleast in the form in which it has been successfully crys-tallized (Pautsch and Schulz, 1998). However, this formconsisted only of the N-terminal domain. This truncatedprotein has now been shown to form small ion channelsin planar bilayers, whereas the full-length protein isrequired for the formation of an additional larger channel(Arora et al., 2000). A similar situation appears to exist forthe major porin protein of P. aeruginosa, OprF (Nikaidoet al., 1991; Bellido et al., 1992), which forms both small

Fig. 7. Growth of wild-type M. tuberculosis and the ompATb mutant in mice. BALB/c mice were infected intravenously with ª 5 ¥ 105 cfu of M. tuberculosis. The numbers of cfu per tissue were determined for lungs (A) and spleens (B) at different time intervals. The wild type is shown by circles and the mutant by triangles. Each point represents the mean of four or five mice; error bars represent standard errors. In a second experiment (C), the single cross-over strain was included, and growth in the lungs at a single time point (65 days after infection) was determined; the black bar represents the wild type, the hatched bar represents the single cross-over and the white bar represents the ompATb mutant. ** denotes that the ompATb mutant differed signifi-cantly from the wild type by the paired Student t-test (P < 0.05).

The functions of OmpATb 197

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

and, more rarely, larger channels in lipid bilayers(Brinkman et al., 2000). OmpATb is in fact a closer homo-logue of OprF than E. coli OmpA, and it is possible that asimilar situation occurs in M. tuberculosis, with the smallerchannel-forming conformer being involved in stabilizingthe cell envelope and a rare, larger channel-forming con-former being involved in porin activity.

Loss of ompATb had no effect on the growth of M.tuberculosis under standard growth conditions (Fig. 4).This is similar to the findings with the mspA gene from M.smegmatis, which clearly seems to be a porin; deletion ofthat gene has no effect on growth rate, although there isa significant loss in the mutant of its capacity to transportglucose (Stahl et al., 2001). Indeed, the ompATb mutantdid appear to have reduced permeability to small water-soluble molecules, which is consistent with OmpATb func-tioning as a porin. The reduction, at least in bacteriamaintained at pH 7.2, was small compared with the reduc-tion in the mspA mutant of M. smegmatis, suggesting thatOmpATb is not the sole molecule involved in transport inM. tuberculosis (Stahl et al., 2001). This agrees with thefinding that at least two proteins with pore-forming activitycan be isolated from envelopes of mycobacteria of theM. tuberculosis group (Kartmann et al., 1999; Lichtingeret al., 1999).

Using both quantitative real-time PCR (Fig. 2A) andimmunoblotting (Fig. 3), we were able to show thatexpression of ompATb is upregulated at low pH. Althoughthe effect of pH on transcriptional regulation of outer mem-brane proteins has been reported previously (Bang et al.,2000; Sato et al., 2000), the dramatic effect of deletion ofthe corresponding genes on ability to grow at moderatelylowered pH, such as we see with ompATb (Fig. 4), hasnot been described previously. It is not clear whether theinvolvement of OmpATb in response to low pH and itsputative role in solute transport are related. The resultsshown in Fig. 5, in which uptake of serine at pH 5.5 in thewild type and mutant was compared, suggest that,although OmpATB plays a minor role at pH 7.2, at pH 5.5,it may be the major porin in M. tuberculosis. It is noticeablethat, in the wild type, uptake of serine at pH 5.5 is signif-icantly lower than at pH 7.2 in spite of the increasedamount of OmpATb in bacteria at the lower pH. Prelimi-nary data using lipid bilayer membranes suggest that theprotein has reduced porin activity at pH 5.5 (E. Lea, per-sonal communication), suggesting that the increasedexpression serves to compensate for reduced activity.

The expression of porins in Gram-negative bacteria isregulated by a sensor–regulator two-component system,EnvZ–OmpR (Aiba et al., 1989; Mizuno and Mizushima,1990). This system can operate as both positive andnegative regulation (Pratt et al., 1996). Two M. tuberculo-sis genes, Rv0902c and Rv0903c, designated as encod-ing a two-component regulatory system (Cole et al.,

1998), are separated from ompATb by two small openreading frames (ORFs) of unknown function. The putativeregulator protein, encoded by Rv0903c, binds to two sites,one within and one upstream of ompATb, and this bindingis enhanced by phosphorylation of the regulator (R. A.Speight, PhD thesis, University of London, 2000). Thus, itis possible that these proteins may be involved in theregulation of ompATb expression. OmpR of Salmonellaregulates expression of the porin OmpC in response tochanges in osmolarity and pH, the latter being by far thegreater effect (Foster et al., 1994). Thus, it would be ofinterest to investigate the involvement of this sensor–regulator pair in response to pH.

The most probable explanation for the reduced viru-lence of the ompATb mutant in macrophages and in micerelates to the response to lowered pH. M. tuberculosis isavidly phagocytosed by macrophages, and these cells areits major environment during the infection, despite theirusual bactericidal capacities. Live pathogenic mycobacte-ria suppress the normal lowering of pH inside phagocyticvacuoles (Sturgill-Koszycki et al., 1994), but this effect ispartial and is over-ridden when the macrophages are acti-vated (Schaible et al., 1998). Exposure to the intracellularenvironment of macrophages results in upregulation ofompATb expression (Fig. 2B). The reduced ability of theompATb mutant to grow at slightly lowered pH is consis-tent with its impaired growth in macrophages. Activationof macrophages is thought to be part of the mechanismby which cell-mediated immunity prevents the growth ofintracellular pathogens, and so its poor response to low-ered pH may explain why the ompATb mutant apparentlyceases growth in mice at about the time when specificcell-mediated immunity begins to operate (Fig. 7). Thiswould also explain the normal growth of the mutant inathymic mice, which are unable to mount an appropriatecell-mediated response to infection with M. tuberculosis.Whether there is a direct relationship between the porinactivity of OmpATb at low pH and reduced ability of themutant to grow in macrophages is not clear. M. tubercu-losis probably spends a very large part of its time, possiblymany years in latently infected individuals, in a low-pHenvironment; the ability to retain some access to hydro-philic nutrients, while generally reducing overall porinactivity, could be an essential survival strategy underthese circumstances.

Tuberculosis has a major impact on human health(World Health Organization tuberculosis fact sheet;http://www.stoptb.org/tuberculosis/), and much effort hasbeen made to identify ‘virulence factors’, gene productsthat are specifically required for survival and multiplicationin a human host (Shinnick et al., 1995; Collins, 1996;Gordon and Andrew, 1996; Quinn et al., 1996; Gomezet al., 1997; Pelicic et al., 1998; Smith et al., 1998). Suchfactors may represent new approaches for developing

198 C. Raynaud et al.

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

avirulent vaccines or new drug targets. It seems thatOmpAtb should now be added to the list of virulencefactors. There is a clear indication that it is involved in theability of the pathogen to respond to the potentially harm-ful lowering of pH that occurs inside the phagocytic cellwhere it multiplies. The mechanism by which it performsthis function is not understood; it is of particular interestto determine whether the involvement of the response toacidification is an aspect of the pore-forming capabilitiesof OmpATb, or whether these are unrelated properties ofthe same molecule.

Experimental procedures

Bacterial strains and growth conditions

Mycobacterium tuberculosis H37Rv (1424) was used forthese experiments. This is a streptomycin-resistant strainwith a mutation in the rpsL gene, which has been used forinsertional inactivation of genes via homologous recombina-tion (Springer et al., 2001). This strain and its derivativeswere grown in modified Dubos broth (Difco Laboratories) oron 7H11 agar with Dubos oleic albumin complex supplement(Difco Laboratories) at 37∞C. In the case of the ompATb-deleted mutant, 20 mg ml-1 kanamycin was included in themedia.

Growth of M. tuberculosis at low pH

The wild type and the ompATb deletion mutant of M. tuber-culosis were grown to exponential phase and then inoculatedinto Dubos medium adjusted to pH 5.5 with HCl to give anOD600 of 0.01. The OD600 of the cultures was determined atvarious time intervals after incubation at 37∞C.

Inactivation of ompATb by homologous recombination

The ompATb gene was inactivated by a previously publishedstrategy (Sander and Bottger, 1998; Sander et al., 2001;Springer et al., 2001). Briefly, a 4.4 kb EcoRI fragment(H37Rv genomic co-ordinates 1001133–1005566) contain-ing ompATb was cloned at the EcoRI site of vector ptrpA-1-rpsL+ (a pBluescript vector carrying the wild-type rpsL geneof M. bovis BCG as a counterselection marker). The ompATbgene was inactivated by replacing part of the codingsequence (425 bp fragment located between BsmI and HpaIsites; H37Rv co-ordinates 1002955 to1003379) with a1.26 kb Kmr cassette from pUC4K (Amersham PharmaciaBiotech) to construct the suicide vector pompA::aph-rpsL+.This plasmid was introduced into a streptomycin-resistantstrain of M. tuberculosis H37Rv (strain 1424) by electropora-tion, and transformants were selected on 7H11+Kan and7H11 Kan+Str agar plates. Chromosomal DNA was isolatedfrom the transformants, digested with NsiI and transferred toa Nylon N+ membrane by vacuum blotting and hybridized toa 450 bp probe located within the 3¢ end of the cloned flank-ing region (H37Rv co-ordinates 1004610–1005059). Theprobe was labelled with [a-32P]-dCTP using Klenow enzymeand buffer from the Oligolabelling kit (Amersham Pharmacia

Biotech). The hybridization and washing protocols were car-ried out under high-stringency conditions.

Immunoblotting of M. tuberculosis cell lysates

Bacterial cell pellets were collected by centrifugation, washedin Tris buffer, pH 9.5, and resuspended in urea/thiourea buffer[5 M urea, 2 M thiourea, 2% (w/v) CHAPS, 2% (w/v) SB 3-10, 65 mM dithiothreitol (DTT), 20 mM Tris, pH 9.5, 0.1 mMEDTA and Complete protease inhibitors (Roche)]. Bacteriawere lysed in the presence of glass beads (150–212 mm,Sigma) in a Ribolyser (Hybaid) at a speed setting of 6.5 for4 ¥ 30 s. The tubes were cooled on ice for 5 min betweeneach cycle of disruption. The supernatant was collected bycentrifugation, brought to room temperature and filteredthrough a low-binding Durapore 0.22 mM membrane filter(Ultrafree-MC; Millipore). An aliquot of the cell extract wasused to determine its protein concentration using a Coo-massie Plus protein assay reagent (Pierce).

Cell-free extracts (30 mg per lane) were resolved by SDS-PAGE (12% gel) and transferred at 60 mA for 1 h to polyvi-nylidene difluoride (PVDF) membrane (Millipore) in a semi-dry blotter (Hybaid) using Tris–glycine–SDS buffer (48 mMTris, 39 mM glycine, 0.037% SDS and 20% methanol,pH ª 8.3). Equal loading of the proteins was confirmed byCoomassie staining of an identical gel, and the efficiency oftransfer was verified by staining the blot with a solution of0.1% Ponceau S in 5% acetic acid. The membrane wasblocked with 10% non-fat milk in TTBS (20 mM Tris, pH 7.5,0.5 M NaCl buffer containing 0.1% Tween 20). The primaryanti-OmpATb antibody was raised in rabbits against recom-binant truncated OmpATb (Senaratne et al., 1998) and usedat 1 : 2000 dilution. The secondary goat anti-rabbit IgG anti-bodies conjugated to horseradish peroxidase (Dako) wereused at 1 : 1000 dilution. The blots were washed and devel-oped according to the enhanced chemiluminescence (ECL)detection protocol (Amersham Pharmacia Biotech).

Uptake of radioactive substrates

[14C]-glycerol (5.66 GBq mmol-1), L-[U-14C]-serine(5.74 GBq mmol-1), D-[U-14C]-glucose (11.5 GBq mmol-1)and [U-14C]-glycine (3.74 GBq mmol-1) were purchased fromAmersham Biosciences. Uptake experiments were per-formed using 1 ¥ 1010 bacteria in 1 ml of saline buffer atroom temperature (approximately 23∞C), containing either1.6 ¥ 10-6 M [14C]-glycerol, 1.9 ¥ 10-6 M L-[U-14C]-serine,5.1 ¥ 10-6 M D-[U-14C]-glucose or 3.9 ¥ 10-6 M [U-14C]-glycine. Aliquots of 0.1 ml were taken at different timesand added onto the top of 0.8 ml of 0.5 M sucrose con-tained in an Eppendorf centrifuge tube. Bacteria were rapidlysedimented through the sucrose gradient by centrifugation(13 000 g for 1 min). The supernatant was carefully removed,and the radioactivity associated with the bacterial pellet wasdetermined by liquid scintillation counting. Uptake determina-tions were repeated in at least three independent experi-ments. For experiments on uptake at low pH, exponentiallygrowing cultures (OD600 0.5–0.6) were centrifuged, and thebacteria were resuspended in Dubos medium at eitherpH 7.2 or pH 5.5. The cultures were incubated overnight at

The functions of OmpATb 199

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

37∞C before resuspension in appropriately buffered salineand assayed for uptake of L-[U-14C] serine as describedabove.

Extraction of M. tuberculosis RNA

The extraction of M. tuberculosis RNA was performed usingthe Hybaid RibolyserTM kit blue (Hybaid). M. tuberculosiscultures (OD600 of 0.6–0.8) were centrifuged at 2000 g for10 min. The pellet was resuspended in 0.6 ml of reagent A(chaotropic RNA stabilizing reagent) and then transferred toa Hybaid Ribolyser tube containing 300 ml of reagent B(phenol acid reagent) and 100 ml of reagent C (chloroform–isoamyl alcohol) and silica/ceramic matrix for optimal bacte-rial lysis. Bacteria were broken by shaking for 2 ¥ 20 s in aRibolyser (Hybaid) at a speed rating of 6. The tubes werecentrifuged (13 000 g for 15 min), and the supernatants weretransferred to new tubes containing 300 ml of reagent C. After10 min of centrifugation at 13 000 g, the aqueous phase wastransferred to a tube containing 500 ml of reagent D (DEPC-treated isopropanol precipitation solution). Total RNA wasthen precipitated overnight at 4∞C and washed with 1 ml of75% ethanol. RNA was resuspended in 90 ml of DEPC-treated water. Contaminating DNA was removed by digestionwith DNase I according to the manufacturer's instructions(Roche).

To extract RNA from M. tuberculosis grown in macroph-ages, THP1 cells or murine bone marrow macrophages wereinfected with M. tuberculosis at a multiplicity of infection (MOI)of 10 for 8 h (see below). The macrophages were lysed with500 ml of 2% saponin, the supernatant containing the bacte-ria was centrifuged for 10 min at 2000 g, and the pellet waswashed twice with saline buffer. M. tuberculosis RNA wasextracted as described above.

Real-time quantitative Taqman PCR assay

Real-time quantitative PCR was carried out on the ABI Prism7700 sequence detection system using the Taqman UniversalPCR master mix (PE Applied Biosystems). The primers andthe Taqman probes (carrying both a fluorophore and aquencher) were designed using the Primer Express softwareand obtained from PE Applied Biosystems. The sequenceswere as follows: ompA forward primer, ATGTGCCGACCTGCAATCA; ompA reverse primer, ATTTCATAGTCGGCTGGGATCA; sigA forward primer, TCGGTTCGCGCCTACCT;sigA reverse primer, TGGCTAGCTCGACCTCTTCCT; ompAFAM-labelled probe, TGGACCCATCGCGTTTGGCAA; sigAFAM-labelled probe, TTGAGCAGCGCTACCTTGCCG.

Reverse transcription (RT)-PCR experiments were carriedout with 1 mg of RNA and 2.5 pmol of specific reverse primersfor ompA and sigA in a volume of 8 ml. After denaturation at65∞C for 10 min, 12 ml of the mixture containing 2 ml of dNTP(25 mM), 4 ml of 4¥ buffer, 2 ml of DTT, 1 ml of RNasin(Promega) and 1.5 ml of Superscript II (Invitrogen) wasadded. Samples were incubated for 60 min at 42∞C, heatedat 75∞C for 15 min and then chilled on ice. Samples werethen diluted with 30 ml of H2O and stored at -20∞C. PCRconditions were identical for all reactions. The 25 ml reactionsconsisted of 12.5 ml of PCR master mix (Promega), 4 ml of

template, 5 pmol of each primer and 2.5 pmol of the appro-priate probe. The reactions were carried out in sealed tubes.Results were normalized to the amount of sigA mRNA, whichwas shown to be constant under the conditions used(Manganelli et al., 1999; 2001).

Preparation of macrophages

THP1 cells were cultured as described previously (Ragnoet al., 2001). After expansion, the cells were centrifuged at200 g for 5 min at room temperature in RPMI medium(Life Technologies) supplemented with 50 nM phorbol 12-myristate 13-acetate (Sigma) at a concentration of 1 ¥ 106

cells ml-1. 12-well tissue culture plates (Nunc) were inocu-lated with 1 ml of this suspension per well and incubated at37∞C in 5% CO2 for 24 h. Under these conditions, THP1 cellsdifferentiate into macrophages, stop dividing and adhere tothe bottom of the wells (Tsuchiya et al., 1982).

Murine bone marrow macrophages were flushed from thefemurs of 6- to 8-week-old Balb/C mice and suspended inDulbecco’s medium with low glucose (1 g l-1) and highcarbonate (3.7 g l-1) concentrations (Gibco BRL) and en-riched with 10% heat-inactivated fetal calf serum, 10% L-cell-conditioned medium and 2 mM glutamine. For the infectionassays, mice bone marrow macrophages were seeded into12-well tissues culture plates (1 ¥ 106 cells per well in avolume of 1 ml) and allowed to differentiate for 6–8 days.

Infection of macrophages to determine intracellular growth of M. tuberculosis

THP1 cells and murine bone marrow macrophages wereinfected by removing the medium and replacing it with 1 mlof RPMI containing 105 cfu of M. tuberculosis (equivalent to1 cfu per 10 macrophages). Cultures were incubated at 37∞Cin a 5% CO2 atmosphere for 16 h. The medium was removed,and the cells were washed twice with 1 ml of warm mediumto remove extracellular bacteria. A sample of 1 ml of freshculture medium was added to each well, and the plate wasreincubated at 37∞C in a 5% CO2 atmosphere. Medium wasreplaced every 48 h. At different time intervals, the mediumwas removed from three wells, and the intracellular bacteriawere released by lysing the macrophages with 500 ml of 2%saponin. The resulting lysate was immediately serially dilutedin sterile saline and plated onto 7H11 agar plates. Theseplates were incubated at 37∞C for 14 days and coloniescounted.

Infection of mice

The wild-type and ompA-deleted mutant strains of M.tuberculosis H37Rv were grown in Dubos 7H9 broth for14 days. Each strain was diluted in phosphate-buffered saline(PBS) to give a suspension of ª 106 cfu ml-1, and 0.2 ml ofthese suspensions was inoculated intravenously into 6- to8-week-old female Balb/c mice. In a second experiment, thewild-type, mutant and single cross-over merodiploid(ompATb+ ompATb::Kmr) strains were grown and diluted to105 cfu ml-1 before being inoculated into mice as describedabove. The infection was monitored by removing the lungs

200 C. Raynaud et al.

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

and spleens of infected mice and homogenizing them byshaking with 2-mm-diameter glass beads in chilled salinewith a Mini-Bead Beater (Biospec Products). Serial 10-folddilutions of the resultant suspensions were plated ontoMiddlebrook 7H11 agar with Dubos oleic albumen complexsupplement (Difco Laboratories). The numbers of cfu weredetermined after the plates had been incubated at 37∞C forapproximately 20 days.

Acknowledgements

This work was supported in part by the European Communityfor research, technological development and demonstrationactivities, Fifth Framework Programme (Contract EU-ClusterQLK2-2000-01761), the Deutsche Forschungsgemeinschaftand the Swiss National Science Foundation. CatherineRaynaud was supported by a grant from INSERM. We wishto thank Dr E. A. Lea, University of East Anglia, for usefuldiscussions on the possible role of OmpATb.

References

Aiba, H., Nakasai, F., Mizushima, S., and Mizuno, T. (1989)Evidence for the physiological importance of the phospho-transfer between the two regulatory components, EnvZ andOmpR, in osmoregulation in Escherichia coli. J Biol Chem264: 14090–14094.

Arora, A., Rinehart, D., Szabo, G., and Tamm, L.K. (2000)Refolded outer membrane protein A of Escherichia coliforms ion channels with two conductance states in planarlipid bilayers. J Biol Chem 275: 1594–1600.

Bang, I.S., Kim, B.H., Foster, J.W., and Park, Y.K. (2000)OmpR regulates the stationary-phase acid toleranceresponse of Salmonella enterica serovar typhimurium. JBacteriol 182: 2245–2252.

Bellido, F., Martin, N.L., Siehnel, R.J., and Hancock, R.E.W.(1992) Reevaulation, using intact cells, of the exclusionlimit and role of porin OprF in Pseudomonas aeruginosaouter membrane permeability. J Bacteriol 174: 5196–5203.

Brinkman, F.S.L., Bains, M., and Hancock, R.E.W. (2000)The amino terminus of Pseudomonas aeruginosa outermembrane protein OprF forms channels in lipid bilayermembranes: correlation with a three-dimensional model. JBacteriol 182: 5251–5255.

Buommino, E., Morelli, F., Metafora, S., Rossano, F.,Perfetto, B., Baroni, A., and Tufano, M.A. (1999) Porinfrom Pseudomonas aeruginosa induces apoptosis in anepithelial cell line derived from rat seminal vesicles. InfectImmun 67: 4794–4800.

Chambers, H.F., Moreau, D., Yajko, D., Miick, C., Wagner,C., Hackbarth, C., et al. (1995) Can penicillins and otherbeta-lactam antibiotics be used to treat tuberculosis? Anti-microb Agents Chemother 39: 2620–2624.

Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C.,Harris, D., et al. (1998) Deciphering the biology of Myco-bacterium tuberculosis from the complete genomesequence. Nature 393: 537–544.

Collins, D.M. (1996) In search of tuberculosis virulencegenes. Trends Microbiol 4: 426–430.

Connell, N., and Nikaido, H. (1994) Membrane Permeabilityand Transport in Mycobacterium tuberculosis. Washington,DC: American Society for Microbiology Press.

Draper, P. (1998) The outer parts of the mycobacterial enve-lope as permeability barriers. Front Biosci 3: D1253–D1261.

Foster, J.W., Park, Y.K., Bang, I.S., Karem, K., Betts, H., Hall,H.K., and Shaw, E. (1994) Regulatory circuits involved withpH-regulated gene expression in Salmonella typhimurium.Microbiology 140: 341–352.

Gomez, J.E., Chen, J.M., and Bishai, W.R. (1997) Sigmafactors of Mycobacterium tuberculosis. Tuber Lung Dis 78:175–183.

Gordon, S., and Andrew, P.W. (1996) Mycobacterial viru-lence factors. Soc Appl Bacteriol Symp Series 25: 10S–22S.

Heyde, M., and Portalier, R. (1987) Regulation of major outermembrane porin proteins of Escherichia coli K-12 by pH.Mol Gen Genet 208: 511–517.

Jap, B.K., and Walian, P.J. (1996) Structure and functionalmechanism of porins. Physiol Rev 76: 1073–1088.

Jarlier, V., and Nikaido, H. (1990) Permeability barrier tohydrophilic solutes in Mycobacterium chelonei. J Bacteriol172: 1418–1423.

Jeannin, P., Renno, T., Goetsch, L., Miconnet, I., Aubry, J.P.,Delneste, Y., et al. (2000) OmpA targets dendritic cells,induces their maturation and delivers antigen into the MHCclass I presentation pathway. Nature Immunol 1: 502–509.

Kartmann, B., Stenger, S., and Niederweis, M. (1999) Porinsin the cell wall of Mycobacterium tuberculosis. J Bacteriol181: 6543–6546.

Lichtinger, T., Heym, B., Maier, E., Eichner, H., Cole, S.T.,and Benz, R. (1999) Evidence for a small anion-selectivechannel in the cell wall of Mycobacterium bovis BCGbesides a wide cation-selective pore. FEBS Lett 454: 349–355.

Manganelli, R., Dubnau, E., Tyagi, S., Kramer, F.R., andSmith, I. (1999) Differential expression of 10 sigma factorgenes in Mycobacterium tuberculosis. Mol Microbiol 31:715–724.

Manganelli, R., Voskuil, M.I., Schoolnik, G.K., and Smith, I.(2001) The Mycobacterium tuberculosis ECF sigma factorsigmaE: role in global gene expression and survival inmacrophages. Mol Microbiol 41: 423–437.

Mizuno, T., and Mizushima, S. (1990) Signal transductionand gene regulation through the phosphorylation oftwo regulatory components: the molecular basis for theosmotic regulation of the porin genes. Mol Microbiol 4:1077–1082.

Mobasheri, H., Senaratne, R.H., Draper, P., and Lea, E.J.A.(1998) Single channel properties of a porin-like proteinfrom Mycobacterium tuberculosis H37Rv in planar lipidbilayers. Biophys J 74: A320.

Mosleh, I.M., Huber, L.A., Steinlein, P., Pasquali, C.,Gunther, D., and Meyer, T.F. (1998) Neisseria gonor-rhoeae porin modulates phagosome maturation. J BiolChem 273: 35332–35338.

Negm, R.S., and Pistole, T.G. (1999) The porin OmpC ofSalmonella typhimurium mediates adherence to macroph-ages. Can J Microbiol 45: 658–669.

Niederweis, M., Ehrt, S., Heinz, C., Klocker, U., Karosi, S.,

The functions of OmpATb 201

© 2002 Blackwell Publishing Ltd, Molecular Microbiology, 46, 191–201

Swiderek, K.M., et al. (1999) Cloning of the mspA geneencoding a porin from Mycobacterium smegmatis. MolMicrobiol 33: 933–945.

Nikaido, H. (1993) Transport across the bacterial outer mem-brane. J Bioenerg Biomembr 25: 581–589.

Nikaido, H. (1994) Porins and specific diffusion channelsin bacterial outer membranes. J Biol Chem 269: 3905–3908.

Nikaido, H., Nikaido, K., and Harayama, S. (1991) Identifica-tion and characterization of porins in Pseudomonas aerug-inosa. J Biol Chem 266: 770–779.

Pautsch, A., and Schulz, G.E. (1998) Structure of the outermembrane protein A transmembrane domain. NatureStruct Biol 5: 1013–1017.

Pelicic, V., Reyrat, J.M., and Gicquel, B. (1998) Geneticadvances for studying Mycobacterium tuberculosis patho-genicity. Mol Microbiol 28: 413–420.

Prasadarao, N.V., Wass, C.A., Weiser, J.N., Stins, M.F.,Huang, S.H., and Kim, K.S. (1996) Outer membrane pro-tein A of Escherichia coli contributes to invasion of brainmicrovascular endothelial cells. Infect Immun 64: 146–153.

Pratt, L.A., Hsing, W., Gibson, K.E., and Silhavy, T.J. (1996)From acids to osmZ: multiple factors influence synthesisof the OmpF and OmpC porins in Escherichia coli. MolMicrobiol 20: 911–917.

Quinn, F.D., Newman, G.W., and King, C.H. (1996) Virulencedeterminants of Mycobacterium tuberculosis. Curr TopicsMicrobiol Immunol 215: 131–156.

Ragno, S., Romano, M., Howell, S., Pappin, D.J., Jenner,P.J., and Colston, M.J. (2001) Changes in gene expressionin macrophages infected with Mycobacterium tuberculosis:a combined transcriptomic and proteomic approach.Immunology 104: 99–108.

Ried, G., Koebnik, R., Hindennach, I., Mutschler, B., andHenning, U. (1994) Membrane topology and assembly ofthe outer membrane protein OmpA of Escherichia coli K12.Mol Gen Genet 243: 127–135.

Saint, N., De, E., Julien, S., Orange, N., and Molle, G. (1993)Ionophore properties of OmpA of Escherichia coli. BiochimBiophys Acta 1145: 119–123.

Sander, P., and Bottger, E.C. (1998) Gene replacement inMycobacterium smegmatis using a dominant negativeselectable marker. Methods Mol Biol 101: 207–216.

Sander, P., Papavinasasundaram, K.G., Dick, T.,Stavropoulos, E., Ellrott, K., Springer, B., et al. (2001)Mycobacterium bovis BCG recA deletion mutant showsincreased susceptibility to DNA-damaging agents but wild-type survival in a mouse infection model. Infect Immun 69:3562–3568.

Sato, M., Machida, K., Arikado, E., Saito, H., Kakegawa, T.,and Kobayashi, H. (2000) Expression of outer membrane

proteins in Escherichia coli growing at acid pH. Appl Envi-ron Microbiol 66: 943–947.

Schaible, U.E., Sturgill-Koszycki, S., Schlesinger, P.H., andRussell, D.G. (1998) Cytokine activation leads to acidifica-tion and increases maturation of Mycobacterium avium-containing phagosomes in murine macrophages. J Immu-nol 160: 1290–1296.

Senaratne, R.H., Mobasheri, H., Papavinasasundaram, K.G.,Jenner, P., Lea, E.J., and Draper, P. (1998) Expression ofa gene for a porin-like protein of the OmpA family fromMycobacterium tuberculosis H37Rv. J Bacteriol 180:3541–3547.

Shinnick, T.M., King, C.H., and Quinn, F.D. (1995) Molecularbiology, virulence, and pathogenicity of mycobacteria. AmJ Med Sci 309: 92–98.

Smith, I., Dussurget, O., Rodriguez, G.M., Timm, J., Gomez,M., Dubnau, J., et al. (1998) Extra and intracellular expres-sion of Mycobacterium tuberculosis genes. Tuber Lung Dis79: 91–97.

Springer, B., Master, S., Sander, P., Zahrt, T., McFalone, M.,Song, J., et al. (2001) Silencing of oxidative stressresponse in Mycobacterium tuberculosis: expression pat-terns of ahpC in virulent and avirulent strains and effect ofahpC inactivation. Infect Immun 69: 5967–5973.

Stahl, C., Kubetzko, S., Kaps, I., Seeber, S., Engelhardt, H.,and Niederweis, M. (2001) MspA provides the main hydro-philic pathway through the cell wall of Mycobacteriumsmegmatis. Mol Microbiol 40: 451–464.

Sturgill-Koszycki, S., Schlesinger, P.H., Chakraborty, P.,Haddix, P.L., Collins, H.L., Fok, A.K., et al. (1994) Lack ofacidification in Mycobacterium phagosomes produced byexclusion of the vesicular proton-ATPase. Science 263:678–681.

Sugawara, E., and Nikaido, H. (1992) Pore-forming activityof OmpA protein of Escherichia coli. J Biol Chem 267:2507–2511.

Sugawara, E., and Nikaido, H. (1994) OmpA protein ofEscherichia coli outer membrane occurs in open andclosed channel forms. J Biol Chem 269: 17981–17987.

Trias, J., and Benz, R. (1994) Permeability of the cell wall ofMycobacterium smegmatis. Mol Microbiol 14: 283–290.

Trias, J., Jarlier, V., and Benz, R. (1992) Porins in the cellwall of mycobacteria. Science 258: 1479–1481.

Tsuchiya, S., Kobayashi, Y., Goto, Y., Okumura, H., Nakae,S., Konno, T., and Tada, K. (1982) Induction of maturationin cultured human monocytic leukemia cells by a phorboldiester. Cancer Res 42: 1530–1536.

Weiser, J.N., and Gotschlich, E.C. (1991) Outer membraneprotein A (OmpA) contributes to serum resistance andpathogenicity of Escherichia coli K-1. Infect Immun 59:2252–2258.