Proc. Natl. Acad. Sci. USAVol. 89, pp. 6521-6525, July 1992Medical Sciences

Modulation of adenylate cyclase toxin production as Bordetellapertussis enters human macrophages

(bacterial invasion/gene expression)

H. ROBERT MASURELaboratory of Molecular Infectious Diseases, The Rockefeller University, 1230 York Avenue, New York, NY 10021

Communicated by Maclyn McCarty, April 20, 1992

ABSTRACT During the course of human infection, Bor-detella pertussis colonizes sequential niches in the respiratorytract that include intracellular and extracellular environments.In vitro the expression of virulence factors such as the adenylatecyclase toxin is coordinately regulated by the bvg locus, whichis an example ofa two-component sensory transduction system.With this toxin as a reporter, enzyme activities were comparedbetween a wild-type and an altered strain to determine whetherbacterial entry into human macrophages affected gene expres-sion. BPRU140, a strain containing an inducible expressionvector, produced enzyme activity independent of bvg. Samplesof the parent, the induced, and the uninduced BPRU140 wereincubated individually with macrophages for 30 min. Extra-cellular bacteria were then killed by gentamicin. The numberof viable intracellular bacteria and the internalized bacterialenzyme activity were measured over time. By 2.5 hr all samplesreached a steady-state concentration of 105 bacteria per 106macrophages. Following an initial peak of enzyme activity,adenylate cyclase values for the parent and the uninducedBPRU140 decreased to a basal level, while the values for theinduced strain remained at least 3-fold greater. Therefore,compared with the persistence of enzyme in the induced strainBPRU140, the decrease in enzyme production by the parentand the uninduced BPRU140 upon entry into macrophagesindicates in vivo down-modulation of gene expression. Theseobservations support the hypothesis that sensory transductioncontributes to adaptations for bacterial survival in the infectedhost.

Several genera of pathogenic bacteria, such as the gastroin-testinal pathogens Salmonella, Yersinia, Shigella, and List-eria, invade and survive within host cells. Recent reportsdescribing the entry and intracellular survival of Bordetellaefrom rabbit alveolar macrophages (1) and cultured mamma-lian cells (2,3) suggest at least two different niches in the host:one extracellular on the ciliated epithelium (4) and anotherwithin leukocytes (1,5). The recovery ofbacteria from withinrabbit alveolar macrophages in vivo indicates an intracellularstate during infection (1). It is assumed that bacteria mountan adaptive response to a changing environment by thecoordinate regulation ofthose genes which convey a selectiveadvantage for survival.The virulence factors of Bordetella pertussis, such as

pertussis toxin (Ptx), filamentous hemaglutinin (Fha), fim-briae (Fim), and the adenylate cyclase (AC) toxin, areorganized into a regulon. Activation of expression of thegenes in this regulon is controlled by the genetic locus bvg,which encodes two proteins, BvgA and BvgS, that aremembers of the two-component family of bacterial sensorytransduction proteins (6, 7). In vitro coordinate regulation bybvg, a process termed phenotypic modulation, produces two

distinct phenotypes (Vir+ and Vir-) in response to definedmedium components. An increased concentration of MgSO4or nicotinic acid or a decreased temperature produces a Vir-phenotype (8-11). Genes expressed in the Vir+ state havebeen termed virulence-activated genes (i.e., those encodingPtx, Fha, and AC) and those expressed in the Vir- state havebeen termed virulence-repressed genes (9). Little is knownabout the in vivo role of phenotypic modulation because ofthe inability to measure the expression of the regulated genessuch as Ptx, Fha, and Fim with sufficient sensitivity andspecificity in vivo in response to a changing environment. Incontrast, the bvg-regulated AC, with its high specific activity,was chosen as a reporter for the bacterial response to thenatural transition of B. pertussis into human macrophages.The cya operon, responsible for the AC (Cya) and hemo-

lytic (Hly) phenotypes, contains five genes, cyaA, -B, -C, -D,and -E (12, 13) with the genetic organization depicted in Fig.1. Transcription of the structural gene, cyaA, is driven by asingle promoter that is dependent on bvg for expression. Anadditional promoter is present between cyaA and cyaB thatis constitutive for the transcription of cyaB, -D, and -E (15).The gene product of cyaA is a calmodulin-stimulated, extra-cellular AC that also confers the hemolytic phenotype and theability to increase cAMP in mammalian cells (16-18). Theproducts of the other contiguous genes cyaB, -D, and -E arerequired for secretion ofAC (12). In the opposite orientationto this polycistron and separated by a short, 260-base-pairintergenic region is another gene, cyaC, whose productconveys the hemolytic and invasive properties to the toxin,presumably through covalent modification of AC (19-21).

MATERIALS AND METHODSBacterial Strains and Media. Escherichia coli strains used

were DH5a, which is F-080dlacZA(lacZYA-argF)U169recAl endAl hsdRJ7 (rK-mK+) supE44 A- thy-i gyrA relAl(Bethesda Research Laboratories), and SM10, which is thi,thr, leu, su111, RP4-2-Tc::Mu (22). B. pertussis strains usedwere BP348, a cyaA::TnS derivative of BP338 (12, 14);BP536, a spontaneous naladixic acid-resistant derivative ofBP338 (23); BP537, a spontaneous bvg- derivative of BP536(23); and BPRU44, a cya::TnStacl derivative of BP338 (20).BPRU176, -178, and -182 are derivatives of BP348 thatcontain plasmids pAMN2, -10, and -13, respectively (Fig. 1).The conjugal transfer of plasmids from E. coli to B. pertussiswas performed as described (23, 24).

E. coli were grown in Luria-Bertani (LB) liquid or solidmedium (25). B. pertussis were grown in Stainer Scholte (SS)medium (26) or on Bordet-Gengeou (BG) blood agar plates.To phenotypically modulate B. pertussis in vitro, bacteriawere grown in medium supplemented with 40mM MgSO4. Tocultivate B. pertussis under growth-limiting conditions, bac-terial suspensions were incubated in SS medium lacking a

Abbreviations: AC, adenylate cyclase; IPTG, isopropyl .6-D-thiogalactopyranoside; cfu, colony-forming unit(s).

6521

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Janu

ary

16, 2

020

Proc. Natl. Acad. Sci. USA 89 (1992)

cyaoperon

A I B D IPlasnid consbcons

cyaA

E I

K

K

K

cyaA

cyaA

FIG. 1. Structural organization of the cya locus and plasmidconstructions. The cya operon consists ofa polycistron that containscyaA, -B, -D, and -E. cyaC is in the opposite orientation and isseparated by a 260-base-pair intergenic region. The TnS chromoso-mal insertion in the mutant strain BP348 (14) resides in cyaA at theposition indicated (12). Restriction sites are indicated by verticallines (B, BamHI; K, Kpn I) and triangles represent promoters. Ptac,tac promoter.

supplement ofessential components required for growth (26).Antibiotics when required were used at the following con-centrations: ampicillin, 100 ,ug/ml; chloramphenicol, 34,ug/ml for E. coli or 10 pAg/ml for B. pertussis; and strepto-mycin, 300 ,g/ml. When indicated, isopropyl f3-D-thiogalactopyranoside (IPTG) was added at 250 ,ug/ml.Plasmid Construction and DNA Manipulation. To identify

the bvg-dependent regulatory region for cyaA and create tacpromoter fusions, a 6.1-kilobase (kb) BamHI-Kpn I fragmentthat contains the gene and its promoter but lacks any up-stream sequence was subcloned from pRM003 (27) andinserted into the corresponding sites of the broad-host-rangecloning vectors pMMB207 and pMMB208 (28) to generatepAMN10 and pAMN2, respectively (Fig. 1). These plasmidsconstitute a pair in which cyaA is in opposite orientations tothe tac promoter of the expression vector, creating inducible(pAMN2) and noninducible (pAMN10) vectors. PlasmidpAMN13, which contains cyaA, its promoter, and 2.4 kb ofupstream sequence spanning the intergenic region and cyaC,was created by subcloning a 2.4-kb BamHI fragment (20) intopAMN10. Plasmids derived from pUC derivatives wereisolated by the alkaline hydrolysis method (29), whereasplasmids derived from pMMB207 and pMMB208 were iso-lated by a boiling method (30). Recombinant DNA manipu-lations were performed by standard protocols. DNA-modifying enzymes were obtained from United States Bio-chemical.AC Assay. Extracellular cell-associated AC was isolated

from whole bacteria as a detergent extract (31). Production ofcAMP was measured as described (32). Units of specificenzyme activity are presented as nmol ofcAMP produced permin per mg of total protein. Samples were assayed in thepresence of 2.5 ,uM calmodulin, and protein concentrationswere determined by the Coomassie blue G250 method (33).

Bacteria--Macrophage Invasion Assay. Human monocyteswere isolated from buffy coats and harvested from Teflonbeakers after 5-10 days and suspended at 2 x 106 per ml inRPMI-1640 medium (GIBCO) (34). Frozen stocks of isogenicstrains of B. pertussis were grown for 1-2 days on selectiveBG medium. Bacterial cells were harvested from the platesand resuspended in RPMI-1640 and adjusted to an OD620 of0.5 [109 colony-forming units (cfu)/ml]. At zero time, 1 ml ofthe bacterial cell suspension was mixed with 1 ml of macro-phages (2 x 106 per ml) and placed on a tumbler at 3TC for30 min. Macrophages and bound bacteria were separatedfrom unbound bacteria by centrifugation for 10 min at 800rpm in a Beckman Accuspin centrifuge. The bacteria/macrophage mixtures were washed three times with RPMI-

1640 and resuspended in a fresh 2 ml of RPMI-1640 plusgentamicin (100 ,g/ml) to kill extracellular bacteria. Inter-nalized bacteria were determined at timed intervals by platingserial dilutions of a macrophage lysate on BG medium. Thelysate was prepared by pelleting and resuspension ofa 100-,lIsample ofthe bacterial/macrophage mixture with phosphate-buffered saline containing 0.1% Triton X-100.

Determination of the Internalized Bacterial AC Activity.Trypsin treatment has been shown to completely destroy anybacteria-associated AC activity (31). Thus to distinguish ACactivity derived from internalized bacteria from that of ex-tracellular bacteria, samples of the bacteria/macrophagesuspension were treated with trypsin to destroy any enzymeactivity outside the macrophage. Samples from the bacteria/macrophage mixture were removed at timed intervals,adjusted to a final trypsin concentration of 20 ,ug/ml, andincubated for 10 min at 37°C. After a 5-min incubation witha 20 molar excess of soybean trypsin inhibitor, the macro-phages were lysed as described above, and the samples werefrozen and subsequently assayed for AC activity. AC activitymeasured in the macrophage lysates was derived from thebacteria and not the endogenous eukaryotic enzyme sinceenzyme activity in a lysate of macrophages challenged witha cya- strain, BPRU44, was three log units less than with theCya+ strains (data not shown).

RESULTS AND DISCUSSIONIdentification of a Regulatory Region for the Expression of

AC and Construction of an Inducible Gene (cyaA). In vitroexpression of the AC in a derivative of a wild-type strain,BP536, is mediated by the bvg locus and can be down-modulated by the addition of MgSO4 (40 mM) to the culturemedium of growing B. pertussis (Table 1). BP536 expresses30-fold less enzyme activity in the presence of MgSO4. Incontrast, the bvg- strain BP537 expresses only a low levelenzyme activity regardless of the presence or absence ofMgSO4. B. pertussis may down-modulate the expression ofthe AC as the bacteria make the transition from outside toinside the macrophage in a manner similar to that observedin vitro with MgSO4. To test this hypothesis it was necessaryto create a strain in which expression of the structural genecyaA was inducible and independent of bvg regulation. Theexpression of enzyme activity in this strain would serve as abasis for comparison (positive control) to the predicteddecreased expression in the parent strain. Plasmid-bornecopies of the cyaA gene were constructed first to identify thebvg-dependent regulatory region and second to place thecyaA gene under the control of the inducible tac promoter.(Plasmid constructions are depicted in Fig. 1.) To test plas-

Table 1. The bvg-dependent and IPTG-induced expression ofAC from B. pertussis

Specific activity, nmol of cAMPper min per mg of protein

No MgSO4Strain Description addition MgSO4 IPTG + IPTG

BP536 bvg+ 6.15 0.22 6.44 0.12BP537 bvg- 0.15 0.19 ND NDBP348 cyaA::Tn5 0 0 ND NDBPRU182 BP348/pAMN13 4.2 0.66 ND NDBPRU178 BP348/pAMN10 0.61 0.51 0.17 0.61BPRU176 BP348/pAMN2 0.43 0.17 6.08 5.19

Bacteria were grown for 48 hr on selective BG agar in the presenceor absence of 40 mM MgSO4 and 250 AuM IPTG as indicated.Extracellular AC was solubilized from whole bacteria in 0.1% TritonX-100 and assayed for the production ofcAMP (25, 32). Mean valuesof two independent determinations are presented. SD for any valuedid not exceed 8% of that value. ND, not determined.

Th5 oi BP348

B B

PAMN13B

pAMN2O

6522 Medical Sciences: Masure

Dow

nloa

ded

by g

uest

on

Janu

ary

16, 2

020

Medical Sciences: Masure

mid-encoded AC activity, these plasmids were introducedinto a mutant strain, BP348, that contains a TnS insertion inthe chromosomal copy of cyaA. An intact plasmid-bornecopy of cyaA, with 2.4 kb of upstream sequence that includesthe intergenic region, and cyaC (pAMN13) expressed en-zyme activity in a bvg-dependent manner in response toMgSO4 (Table 1). In contrast, a plasmid (pAMN10) lackingthis upstream sequence but retaining the promoter for cyaAproduced only low levels of enzyme activity. This suggestedthat the deleted region contained cis-acting elements requiredfor the bvg-mediated expression of cyaA. This is in goodagreement with the results of lacZ fusion studies character-izing bvg-dependent expression of cyaA (35) and gel retar-dation analysis of the intergenic region (36).Next the cyaA gene lacking the regulatory region was

positioned just downstream from the inducible tac promoterof an expression vector (pAMN2). In this case, significant ACactivity was detected only in the presence of the exogenousinducer IPTG, regardless of the presence or absence of themodulator MgSO4 (Table 1). Two other features of thisexpression system define other members of the cya operon.Detection of extracellular pAMN2-encoded enzyme con-firms bvg-independent transcription of the chromosomalcopies of cyaB, -D, and -E, which are required for secretion(15). Second, even though the mutant strain BP348 containsan intact chromosomal copy of cyaC and the intergenicregion, the hemolytic phenotype was eliminated when theplasmid-encoded, IPTG-induced enzyme was expressed inthe presence of MgSO4 (BPRU176; data not shown). Thissuggests that cyaC, oriented in the opposite direction to therest of the operon, is regulated in a bvg-dependent manner.Thus, the plasmid-derived inducible form of the AC enzymecould be used as a reporter for the expression of cyaAuncoupled from control by the bvg locus.

In Vitro Modulation of cyaA Expression. The inducibleexpression vector pAMN2 containing cyaA was introducedinto a wild-type strain of B. pertussis containing a functionalchromosomal copy of the cya operon. This strain (BPRU140)expressed AC activity in a bvg-dependent manner, but down-modulation could be overcome by the addition of IPTG (Fig.2). One hour following addition of40mM MgSO4, AC activityof BPRU140 decreased to a basal level that remained con-stant over 4 hr (i.e., one generation time). This time courseof down-modulation was identical to that of the parentalstrain BP536 and is consistent with bvg-dependent down-modulation of AC expression measured at the level of tran-scription (37). In contrast, AC activity remained 4-fold higherin BPRU140 when induced with IPTG, despite the presenceof 40 mM MgSO4. Thus, this strain expressed AC activity dueto the chromosomal copy of the gene, in a bvg-dependentmanner leading to a 4-fold reduction in enzyme activity withthe addition of MgSO4, but expression of AC activity due tothe vector pAMN2 was independent of bihg but rather undercontrol of the tac promoter.

In Vitro Modulation of cyaA Expression Under Slow GrowthConditions. Little is known about the in situ state of bacterialgrowth upon colonization of the host or the effect of growthrate on the ability to phenotypically modulate gene expres-sion. Bacteria may have limited access to nutrients withinintracellular compartments. Therefore, a decrease in ACexpression by the bacteria with the transition to inside themacrophage could be attributable to either down-modulationvia the bvg locus or a decrease in expression in response tolimiting growth conditions. To assure that reduced growthrate per se would not affect bvg-dependent AC expression,bacteria were subjected to essential nutrient deprivation andevaluated for the ability to phenotypically modulate theexpression of the AC. The presence or absence of themodulator MgSO4 altered the expression of AC in normal orslow-growth conditions (Fig. 3). Therefore, the bvg-

Proc. Natl. Acad. Sci. USA 89 (1992) 6523

12

x=X

8EI

4

Time, hr

FiG. 2. In vitro modulation of AC expression in B. pertussis. (A)Growth curve of the parent strain BP536 (-) and two samples of thederivative BPRU140 (BP536/pAMN2) with (e) and without (5) 250/iM IPTG. Cultures were grown in liquid medium (26) and at timezero were adjusted to a final concentration of 40 mM MgSO4 tomodulate the expression ofAC. Results (cfu/ml) were determined byserial dilution of liquid cultures onto solid medium. (B) ExtracellularAC activity. Samples from liquid cultures described in A (300 pl)were removed at timed intervals, centrifuged, and suspended in 0.1%Triton X-100 to solubilize the extracellular associated enzyme.Samples were centrifuged again to remove whole cells and thesupernatants were assayed for enzyme activity (26, 32) and proteincontent. Results are presented as units of specific activity (nmol ofcAMP produced per min per mg of protein). Error bars represent theSEM of duplicate samples. Symbols are the same as in A.

dependent regulatory circuit controlling the expression ofACis functional under limiting growth conditions.In Vivo Modulation of cyaA Expression with the Entry of B.

pertussis into Human Macrophages. To determine whether B.

12

xa 8

4

.2

V:

¢a)

2

Time, hr

FIG. 3. In vitro modulation of AC expression under growth-limiting conditions. (A) Growth curve of BPRU140 grown in com-plete (26) (in, ) or growth limiting (E, a) medium in the absence (-,a) or presence (a, e) of 40mM MgSO4. To phenotypically modulatethe bacteria, cultures initially grown in the absence of MgSO4 were

switched to medium containing the divalent cation or, alternatively,cultures grown in the presence of MgSO4 were switched to mediumdevoid of this component. (B) AC activities of cultures described inA (determined as described in Fig. 2).

.

A

A,

Dow

nloa

ded

by g

uest

on

Janu

ary

16, 2

020

Proc. Nat!. Acad. Sci. USA 89 (1992)

pertussis modulates expression ofAC upon entry into humanmacrophages, bacterial AC enzyme activity was monitored inparallel with bacterial invasion and survival in macrophages.Three samples were studied. (i) The parent strain, BP536,which contains only a bvg-regulated chromosomal copy ofcyaA, was compared with two controls. (ii) BPRU140, whichcontains an IPTG-inducible plasmid-borne copy of cyaA in awild-type background, was assayed as a positive control toassess induced AC production independent ofbvg. (iii) KilledBP536 was assayed as a negative control to establish basalAC activity associated with internalized bacteria not activelyproducing the enzyme. Bacterial AC activity within themacrophage was assessed by destroying extracellular en-zyme with trypsin and liberation of the internalized bacterialenzyme by lysis of the macrophages that contained ingestedbacteria. Intracellular bacteria invasion and survival weredefined as colony-forming units obtained following gentami-cin treatment of a bacteria/macrophage mixture.The parental strain (BP536) and BPRU140 bound macro-

phages in a ratio of 10:1 (107 bacteria per 106 macrophages perml; Fig. 4A). After the addition of gentamicin (100 ,ug/ml),bacterial killing progressed until a steady-state concentrationof 105 bacteria per ml was reached by 2.5 hr. Bacteriarefractory to the bactericidal activity of gentamicin after 2.5hr were taken to represent the intracellular pool of bacteria.Similar results were obtained with polymixin B (100 &g/ml)(data not shown). Ninety-five percent of the macrophagesremained viable during the course of the experiment asdetermined by visual examination for trypan blue exclusion.Induction of cyaA with IPTG in the mutant strain BPRU140did not affect the survival of these bacteria. To kill thebacteria prior to incubation with the macrophages, a sampleof BP536 was pretreated for 2 hr with gentamicin at 100,ug/ml. This led to a four-log reduction of surviving bacteriacompared with bacteria that were not pretreated. The bvg-strain BP537 does not express any known virulence deter-minant. This strain bound less avidly to macrophages thanthe bvg+ strains (2 x 105 cfu/ml vs. 1 x 107 cfu/ml). Thosethat were bound did not survive treatment with gentamicinand macrophage challenge (103 cfu/ml). This is consistentwith other studies describing the inability of avirulent strainsto attach to human macrophages (38) or invade other cells (2,3). Bacterial multiplication in any strain was not observedover the 4-hr time course of the experiment. This is notunexpected given the 4- to 6-hr doubling time observed forthe growth of B. pertussis in liquid medium (Fig. 3). There-fore, virulent B. pertussis is able to survive challenge byhuman macrophages in this invasion assay.AC Activity Associated with the Entry of B. pertussis Into

Human Macrophages. There were clearly two components tothe internalized AC activity identified from the bacteria/macrophage mixtures. Within 30 min of the attachment ofwild-type BP536 or BPRU140, comparable peaks ofbacterialAC activity were measured in the macrophage lysates (Fig.4B). Since this peak could be reproduced by killed BP536(pretreated with gentamicin) the initial peak ofAC activity forall toxin-containing strains was attributed to internalizationof preexisting toxin regardless of the viability of the bacteria.By 1.5 hr, this peak of initial AC activity decreased to a basalvalue for the nonviable BP536, which is not actively produc-ing new enzyme. It seems unlikely that the basal AC valuesobtained with the nonviable bacteria represent activity pro-duced by the small number of surviving cells (y102 bacteria),and these basal values are probably due to the residualactivity from the initial internalization of bacteria that deliv-ers AC to the macrophage. This interpretation is consistentwith numerous studies using purified toxin, which is capableof increasing cAMP in a wide variety of eukaryotic cells(16-18). During the same time period, the positive control,BPRU140 induced with IPTG to express the plasmid derived

8-

1 2 3 4Time, hr

Fio. 4. Survival and AC activity of B. pertussis within humanmacrophages. (A) Survival of bacteria following uptake into macro-phages. Bacteria were mixed with human macrophages and, 30 milater, exposed to gentamicin (100 Iug/ml, arrow) to kill extracellularbacteria. Samples were removed at timed intervals, macrophageswere lysed, and the solution was plated to quantitate viable intra-cellular bacteria. Values are representative of experiments con-ducted in triplicate. The parent BP536 (n) was compared with killedBP536 (treated with gentamicin for 2 hr prior to mixture withmacrophages) (0), isogenic BPRU140 (BP536 carrying the plasmid-borne, IPTG-inducible cyaA) treated (e) or untreated (o) with 250,uM IPTG (beginning 10 min prior to admixture with macrophages),and BP537 (bvg-) (x). (B) AC activity associated with internalizedbacteria. Symbols are as described in A. AC activity associated withinternalized bacteria was determined by trypsin treatment of thebacteria/macrophage mixture to destroy enzyme activity outside themacrophage, followed by quenching of the trypsin with soybeantrypsin inhibitor, lysis of the macrophage, and quantitation ofenzyme activity. Values are means ofduplicate samples presented asunits of specific activity (nmol ofcAMP produced per min per mg ofprotein). SD for any value did not exceed 7%. Results are represen-tative of experiments conducted in triplicate.

form ofthe enzyme, produced AC activity 5-fold greater thanthe basal values of the killed parental strain. Thus, after 1.5hrAC values were taken to be indicative ofenzyme producedby bacteria within the macrophage.Between 2 and 4 hr AC activity for all samples reached a

steady-state value. The wild-type parent and the uninducedmutant responded to internalization with low AC valuescomparable to those of the nonviable bacteria (Fig. 4B).When the AC values were plotted as a function of activity percolony-forming unit over time (Fig. 5), the wild-type parent,BP536, produced only basal AC activity that was consistentlyonly one-third that obtained with the induced strainBPRU140. These observations demonstrate down-modula-tion of AC expression in the parent strain with entry intomacrophages and suggest a setting in which phenotypicmodulation, so well studied in vitro, also occurs in vivo.Failure to down-modulate would have produced AC valuescomparable to those obtained with the induced strain. Theability of the parent bacteria to alter AC expression with the

6524 Medical Sciences: Masure

Dow

nloa

ded

by g

uest

on

Janu

ary

16, 2

020

Proc. Natl. Acad. Sci. USA 89 (1992) 6525

1.00

i 0 0.75C Xxu 0.50u Era 0.25U

2 3Time, hr

4

FIG. 5. In vivo modulation of cyaA expression with the entry ofB. pertussis into human macrophages. AC activity associated withviable bacteria internalized within human macrophages was mea-sured as described in Fig. 4. Units of AC specific activity (nmol ofcAMP per min per mg of protein) from the experiment depicted inFig. 4 are plotted as a function of cfu over time. The parent BP536(-) was compared with isogenic BPRU140 (BP536 carrying theplasmid-borne, IPTG-inducible cyaA) treated (e) or untreated (o)with 250 .uM IPTG.

transition to inside the macrophage is presumably mediatedby the products of the bvg locus despite potential growth-limiting conditions. This argument is strengthened by theresults in Fig. 3, which demonstrate modulation of ACexpression in vitro by the bacteria even under limiting growthconditions. It is certainly possible that the stimulus for thisdown-modulation is contained within the milieu of the en-docytic vesicles of the macrophage encountered by thebacteria with internalization.

Survival of B. pertussis upon challenge with human mac-rophages has been attributed to expression of adhesins andvirulence factors in the Vir+ state (39). If the regulatorycircuit between the bvg and the cya loci serves as a paradigmfor gene regulation of other members of the vir regulon, thenthe results reported here suggest that the natural progressionof events in the disease may require that B. pertussis phe-notypically changes to a Vir- state upon entry into themacrophage. This hypothesis is supported by studies withanimal models for the disease. A mutant with a transposoninsertion in vrg6 (a vir-repressed gene) was less able tocolonize the lung and trachea in a mouse model for respira-tory infection (40). Similar studies also demonstrated that theAC (the product ofa vir-activated gene) is absolutely requiredfor colonization (41). Therefore, B. pertussis has developeda survival strategy based on the coordinate regulation ofgenes including the cya operon that results in two distinctphenotypes. In the Vir+ state the bacteria adhere and colo-nize the host in several niches. Thereafter access to anotherenvironment, the alveolar macrophage, leads to the Vir-state, cessation of toxin production, and establishment ofintracellular bacteria that remain sequestered from host de-fenses. The ability of the bacteria to sense its environmentand switch between two phenotypes confers a selectiveadvantage for B. pertussis in promoting successful and per-sistent survival within the host.

I thank Drs. E. Tuomanen and C. A. Butler for many thoughtfuldiscussions, support, and critical reading of the manuscript. I thankAnne Naughton for her expert technical assistance and Drs. JeffWeiser and Emil Gotschlich for their encouragement. This work wassupported by National Institutes of Health Grants R29 A132024-01and RO1 A123459 and by a Career Scientist Award from the I. T.Hirshl Trust.

1. Saukkonen, K., Cabellos, C., Burroughs, M., Prasad, S. &Tuomanen, E. (1991) J. Exp. Med. 173, 1143-1149.

2. Ewanowich, C. A., Melton, A. R., Weiss, A. A., Sherburne,R. K. & Peppler, M. S. (1989) Infect. Immun. 57, 2698-2704.

3. Lee, C. K., Roberts, A. L., Finn, T. M., Knapp, S. & Mekal-anos, J. J. (1990) Infect. Immun. 58, 2516-2522.

4. Tuomanen, E. & Weiss, A. (1985) J. Infect. Dis. 152, 118-125.5. Steed, L. L., Setareh, M. & Friedman, R. L. (1991) J. Leuk.

Biol. 50, 321-330.6. Scarlato, V., Prugnola, A., Arico, B. & Rappuoli, R. (1990)

Proc. Nati. Acad. Sci. USA 87, 6753-6757.7. Stibitz, S. & Yang, M. S. (1991) J. Bacteriol. 173, 4288-4296.8. Idigbe, E. O., Parton, R. & Wardlaw, A. C. (1981) J. Med.

Microbiol. 14, 409-418.9. Knapp, S. & Mekalanos, J. J. (1988) J. Bacteriol. 170, 5059-

5066.10. Lacey, B. W. (1960) J. Hyg. 58, 57-93.11. Melton, A. R. & Weiss, A. A. (1989) J. Bacteriol. 171, 6206-

6212.12. Glaser, P., Sakamoto, H., Bellalou, J., Ullmann, A. & Danchin,

A. (1988) EMBO J. 7, 3997-4004.13. Glaser, P., Ladant, D., Sezer, O., Pichot, F., Ullmann, A. &

Danchin, A. (1988) Mol. Microbiol. 2, 19-30.14. Weiss, A. A., Hewlett, E. L., Myers, G. A. & Falkow, S.

(1983) Infect. Immun. 42, 33-41.15. Laoide, B. M. & Ullmann, A. (1990) EMBO J. 9, 999-1005.16. Confer, D. L. & Eaton, J. W. (1982) Science 217, 948-950.17. Friedman, E., Farfel, Z. & Hanski, E. (1987) Biochem. J. 243,

145-151.18. Shattuck, R. L. & Storm, D. R. (1985) Biochemistry 24, 6323-

6328.19. Berry, E. L., Weiss, A. A., Ehrmann, I. E., Grey, M., Hew-

lett, E. L. & Goodwin, M. (1991) J. Bacteriol. 173, 720-726.20. Saukkonen, K., Sande, S. & Masure, H. R. (1991) in Proceed-

ings of the Sixth International Symposium on Pertussis, ed.Manclark, C. R. (Dept. Health and Human Services, Bethesda,MD), DHHS Publ. No. (FDA) 90-1164, pp. 267-274.

21. Issartel, J. P., Koronakis, V. & Hughes, C. (1991) Nature(London) 351, 759-761.

22. Simon, R., Priefer, U. & Piihler, A. (1983) BiolTechnol. 1,784-791.

23. Relman, D. A., Domenighini, M., Tuomanen, E., Rappuoli, R.& Falkow, S. (1989) Proc. Nati. Acad. Sci. USA 86,2637-2641.

24. Stibitz, S., Black, W. & Falkow, S. (1986) Gene 50, 133-140.25. Miller, J. F. (1972) Experiments in Molecular Genetics (Cold

Spring Harbor Lab., Cold Spring Harbor, NY).26. Masure, H. R., Donovan, M. G. & Storm, D. R. (1991) Meth-

ods Enzymol. 195, 137-152.27. Masure, H. R., Au, D. C., Gross, M. K., Donovan, M. G. &

Storm, D. R. (1990) Biochemistry 29, 140-145.28. Morales, V. M., Backman, A. & Bagdasarian, M. (1991) Gene

97, 39-47.29. Birnboim, H. C. & Doly, J. (1979) Nucleic Acids Res. 7,

1513-1518.30. Holmes, D. S. & Quigley, M. (1981) Anal. Biochem. 114,

193-199.31. Masure, H. R. & Storm, D. R. (1989) Biochemistry 28, 438-

442.32. Salomon, Y., Londos, C. & Rodbell, A. M. (1974) Anal.

Biochem. 58, 541-548.33. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254.34. Wright, S. & Silverstein, S. C. (1982) J. Exp. Med. 156,

1149-1164.35. Goyard, S. & Ullmann, A. (1991) FEMS Microbiol. Lett. 77,

251-256.36. Huh, Y. J. & Weiss, A. A. (1991) Infect. Immun. 59, 2389-

2395.37. Scarlato, V. & Rappuoli, R. (1991) J. Bacteriol. 173, 7401-7404.38. Relman, D., Tuomanen, E., Falkow, S., Golenbock, D. T.,

Saukkonen, K. & Wright, S. D. (1990) Cell 61, 1375-1382.39. Saukkonen, K., Cabellos, C., Burroughs, M., Prasad, S. &

Tuomanen, E. (1991) J. Exp. Med. 173, 1143-1149.40. Beattie, D. T., Shahin, R. & Mekalanos, J. J. (1992) Infect.

Immun. 60, 571-577.41. Goodwin, M. S. & Weiss, A. A. (1990) Infect. Immun. 58,

3445-3447.

Medical Sciences: Masure

Dow

nloa

ded

by g

uest

on

Janu

ary

16, 2

020