Vol. 61, No. 10INFECrION AND IMMUNITY, Oct. 1993, p. 4232-42390019-9567/93/104232-08$02.00/0Copyright X 1993, American Society for Microbiology

Acid Phosphatase Activity in Coxiella burnetii: a PossibleVirulence Factor

OSWALD G. BACA,12* MIRIAM J. ROMAN,' ROBERT H. GLEW,3 ROBERT F. CHRISTNER,lJOHN E. BUHLER,2 AND ADAM S. ARAGON2

Departments of Biology,'* Microbiology, 2 and Biochemistry,3 University ofNew Mexico, Albuquerque, New Me-xico 87131

Received 13 May 1993/Returned for modification 25 June 1993/Accepted 7 July 1993

High-speed supernatant fluids derived from sonicated CoxieUla burnetii contained considerable acid phos-phatase activity when assayed by using 4-methylumbelliferylphosphate; they also contained a factor thatblocked superoxide anion production by human neutrophils stimulated with formyl-Met-Leu-Phe. The pHoptimum of the enzyme was approximately 5.0. The level of phosphatase activity detected in several isolates ofC. burnetii implicated in acute (Nine Mile) and chronic (S Q217, PRS Q177, K Q154) Q fever was 25 to 60 timesgreater than that reported in other microorganisms, including Leishmania and LegioneUla spp. The enzyme wasfound in rickettsiae grown in different hosts (L929 cells and embryonated eggs) and, in the case of L929 cells,for both short periods (less than a month) and the long term (years). Cytochemical techniques coupled withelectron microscopy localized the phosphatase activity to the periplasmic gap in the parasite. Ion-exchangechromatography revealed a major species of the enzyme and showed that the enzyme of the parasite wasdistinct from that of the host cell (L929 fibroblasts); its apparent molecular weight was 74,000. Phosphataseinhibitors (i.e., molybdate heteropolyanions) had differential effects on the phosphatases of the parasite andhost cell. C. burnetii supernatant fluid inhibited superoxide anion production by formyl-Met-Leu-Phe-stimulated human neutrophils; molybdate inhibitors reversed the inhibition. Treatment of C. burnetii-infectedL929 cells with one of the molybdate compounds (complex B') significantly reduced the level of infection anddid not affect the viability or growth of the host cell. These data suggest that the acid phosphatase of the parasitemay be a major virulence determinant, allowing the agent to avoid being killed during uptake by phagocytesand subsequently in the phagolysosome.

The agent of Q fever in humans, Coxiella burnetii, residesand proliferates within acidic phagolysosomes of host cells,including phagocytes (2). Although this obligate intracellularrickettsial parasite contains superoxide dismutase and cata-lase (1), which probably protect it from host cell-generatedsuperoxide anion and hydrogen peroxide, respectively, it islikely that other virulence determinants play a significantrole in allowing the agent to thrive within phagolysosomes.Recently we reported that intact C. bumetii inhibits therespiratory burst of phagocytosing human neutrophils (3). Inlight of reports (13, 21, 24) that Legionella micdadei andLeishmania promastigotes possess acid phosphatases(ACPs) that inhibit the respiratory (oxidative) burst of neu-trophils, we investigated the possibility that C. burnetii alsopossesses such a virulence factor, which would account forits capacity to inhibit the oxidative burst of the neutrophil.Both of these facultative parasitic agents also proliferate inphagolysosomes of macrophages. Their ACPs apparentlymodulate the respiratory burst of neutrophils by reducing theamount of the second messengers, inositol 1,4,5-triphos-phate and sn-1,2-diacylglycerol, produced following recep-tor-mediated stimulation (12, 25).

In this report we show that C. bumetii, does, indeed,possess significant ACP activity and that it is probablyresponsible for inhibiting the metabolic burst of humanneutrophils.

(This work was presented, in part, at the 1992 GeneralMeeting of the American Society for Microbiology, NewOrleans, La. [abstr. no. B571.)

* Corresponding author.

MATERIALS AND METHODSC. burnetii propagation and purification. Several C. bur-

netii isolates implicated in acute and chronic Q fever (seeTable 1) were cultivated in L929 fibroblasts or in embryo-nated eggs and subsequently purified from host cell compo-nents. L929 fibroblasts were infected with C. burnetii byestablished procedures (6, 7, 22). Normal and infected cellswere maintained in suspension culture at 35°C.

Rickettsiae were purified from infected L929 cells (contin-uously infected and maintained in culture for the periodsindicated without freezing), and concentrations were deter-mined as described by Baca et al. (7). Infected cells sus-pended in 2 mM magnesium acetate were disrupted in a TenBroek homogenizer, and the released rickettsiae were puri-fied from host components by differential centrifugation:alternating low-speed (150 x g for 10 min) and high-speed(10,000 x g for 30 min) centrifugations to sediment hostdebris at low speed and rickettsiae at high speed. Thepurified rickettsiae were enumerated by the methods ofSilberman and Fiset (28) as modified by Williams et al. (33).

C. burnetii organisms, grown in pathogen-free embryo-nated eggs, were purified from infected yolk sacs afterhomogenization, differential centrifugation, and banding in30 to 60% (wtlwt) linear sucrose gradients (29).The preparations were monitored for microbial contami-

nants with the aid of thioglycolate broth and blood agar.Electron microscopy and ACP cytochemistry. Infected

L929 cells and purified C. burnetii organisms were subjectedto electron microscopy for localization of ACP activitywithin the parasite by the Gomori cytochemical technique(2, 15). The enzyme substrates used were 1.25% (wt/vol)P-glycerophosphate or p-nitrophenyl phosphate (disodium

4232

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ACID PHOSPHATASE ACTIVITY IN COXIELLA BURNETII 4233

salt) (Sigma Chemical Co., St. Louis, Mo.); 0.2% (wt/vol)lead (PbNO3) served as the capture agent. Cells and rickett-siae were fixed for 30 min in 0.1 M cacodylate buffer (pH 7.4)containing 2.5% (vol/vol) glutaraldehyde. After fixation, thecells were washed in three changes of ice-cold cacodylatebuffer and then incubated for 2 h at 37°C in a shaking waterbath with constant agitation (100 rpm) in reaction mediumbuffered to pH 5.0 with Tris-maleate. Control sampleslacked the substrate in the reaction medium. The cells andrickettsiae were postfixed for 1 h in 1% (wttvol) OS04buffered in cacodylate, embedded, sectioned, stained (withuranyl acetate and lead citrate), and examined by transmis-sion electron microscopy (2).

Standard ACP assay. ACP activity was determined fluo-rometrically with 4-methylumbelliferylphosphate (MUP)serving as the substrate (14). The standard assay was carriedout for 15 min at 37°C in a 0.1-ml reaction mixture containing0.2 M sodium acetate buffer (pH 5.5) and 5 mM MUP. Oneunit of enzyme activity is defined as the amount of enzymerequired to convert 1 nmol of substrate to product per hour.

Protein was determined by the Bradford (8) method withbovine serum albumin as the standard.

Preparation of C. burnetii supernatant fluid and ion-exchangechromatography. Purified rickettsiae suspended in N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES)buffer (0.01 M [pH 7.4] with 0.8% [wtlvol] NaCl) were

subjected to ultrasonication (30 to 60 s at ice bath tempera-ture) followed by centrifugation at 10,000 x g for 1 h at 4°C.The supernatant fluid was subjected to additional centrifu-gation at 100,000 x g for 1 h at 4°C. The resulting high-speedsupernatant fluid was chromatographed on a QAE-Sephadexcolumn and subjected to high-performance liquid chroma-tography (stationary-phase Superdex G-75 column). At eachstep, aliquots were assayed for ACP activity and usedimmediately in experimentation or stored at -70°C. In theinitial stages of this investigation the sonicates were alsosubjected to three cycles of freezing and thawing. Wesubsequently found that freezing and thawing did not en-hance ACP yields and therefore discontinued the practice. Itwas also found that freezing and thawing did not affect theresults obtained (e.g., chromatography profiles and enzymeactivities). Except when noted in this report, sonicated C.bumetii organisms were not also subjected to freezing andthawing.

Preparation of human neutrophils and superoxide anionassay. Neutrophil suspensions were prepared from humanperipheral blood by dextran sedimentation followed by den-sity gradient centrifugation on Ficoll-Hypaque gradients andisotonic NH4C1 lysis (20, 24). The generation of superoxideanions by formyl-Met-Leu-Phe-stimulated neutrophils was

measured as superoxide dismutase-inhibitable cytochrome c

reduction by using a continuous assay method (20).Treatment of infected L929 cells with a heteromolybdate

compound (complex B') and determination of degree of infec-tion. The heteromolybdate compound complex B' was dis-solved in sterile distilled water (10-mg/ml stock), and in-fected L929 cells (infected ca. 3,200 days earlier with phaseI Nine Mile C. burnetii) were exposed to either 20 or 40 jigof the compound per ml for up to 2 weeks. The percentagesof infected cells were determined as previously described byus (35, 36) by direct microscopic examination of Gimenez-stained cells. A minimum of 300 cells were examined in eachprepared slide to determine the percentage of the populationthat was infected (one or more rickettsiae per cell). Dupli-cate flasks containing 10 ml of infected cells (starting cellconcentration, 2.5 x 105/ml) were used for each concentra-

TABLE 1. ACP activity in different isolates of C. burnetiiassociated with acute or chronic disease, cultivated

in either L929 cells or embryonated eggs

Isolate Phase Host Sp act (U/mg ofprotein)a

Nine Mile Originally I L929 (952 days)b 1,510 ± 127Nine Mile I L929 (28 days) 1,540 ± 102Nine Mile I Egg embryoc 2,430 ± 154S Q217 Originally I L929 (952 days) 2,630 ± 159S Q217 I L929 (28 days) 2,620 ± 118S Q217 I Egg embryo 1,610 ± 104PRS Q177 I Egg embryo 3,840 ± 477K Q154 I Egg embryo 2,240 ± 224

a Results represent the means + standard deviations of three independentexperiments with purified intact rickettsiae.

L929 cells infected for the number of days indicated.c The egg-grown C. burnetii organisms, obtained from L. P. Mallavia,

Washington State University, were in their third serial egg passage; they werein phase I. The Nine Mile and S Q217 isolates (both phase I) grown in L929cells were previously cultivated in guinea pigs and subjected to three eggpassages.

tion of the compound; duplicate control flasks received onlythe solvent (water). The cell populations were divided twicea week, and the concentration of molybdate was adjusted.Viability was determined by the dye exclusion technique(35).

RESULTS

Detection of ACP C. burnetii isolates and optimum pHactivity. Several isolates of C. burnetii, representative of thethree major groups implicated in acute or chronic disease,exhibited significant levels of ACP activity when assayedfluorometrically with MUP serving as the substrate (Table1). The rickettsiae were purified from host cells as describedin Materials and Methods and included the isolates NineMile, which causes acute short-term disease, S Q217 (plas-midless) and K Q154 (QpRS plasmid) isolated from patientswith chronic endocarditis, and PRS Q177 isolated from agoat cotyledon and from the same plasmid group (QpRS) asK Q154 (25). The assays were performed on intact organismsat 37°C in 0.1-ml reaction mixtures containing 0.2 M sodiumacetate buffer (pH 5.5) and 5 mM MUP. Irrespective of thehost from which the rickettsiae were isolated-L929 fibro-blasts or embryonated eggs-the parasites exhibited signifi-cant levels of ACP activity. Also, rickettsiae derived fromshort-term (28 days)- and long-term (952 days)-infected L929cells exhibited comparable levels of ACP activity (Table 1).Disrupted C. burnetii organisms (sonicated for 30 to 60 s at0°C, in a Branson Sonifier with the microprobe) exhibitedslightly higher levels of ACP activity than did the intactrickettsiae (specific activities, 1,430 + 157 for the intactorganisms and 1,840 ± 147 for the sonicated organisms[results of four independent experiments]). The C. burnetiiNine Mile organisms were purified from L929 cells that hadbeen infected for over 2,500 days. Sonication for additionalperiods (up to 4 min, which is sufficient to completely disruptthe rickettsiae [1]) did not result in increased ACP activity.These results suggested that the enzyme was localized at ornear the parasite surface, i.e., in the periplasmic space.The ACP activity (on a per-milligram-of-protein basis)

present in C. bumetii was greater than has been reported todate, to our knowledge, for any other microorganism; de-pending on the isolate, it was approximately 25 to 60 timesthat observed in Legionella micdadei (24).

VOL. 61, 1993

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

4234 BACA ET AL.

180-

160-

>

140-

100-

Z~ 80-C,)O 60-

o 40-

e 20-

0

FIG. 1.activity. Nition) was pito 0.5 M NMaterials afractions otdetermined3 to 5.5), 0.6.5 and 8).which had I

Optimu:in 100,00(Nine Millperformecas the sub

Presencousing the (activity wNine Millrickettsiaecells wereophosphalcapture aprepared Iexaminedtron micrelectron-dperiplasmiapproxim;p-nitrophe

Ion-excinetii phosjC. bumetsuspendecThe disrul4°C for 1QAE-Sep]M NaCl.

cells were obtained after homogenization with the Ten Broekhomogenizer and centrifugation at 100,000 x g for 1 h, andthey too were subjected to chromatography. The ACPs ofL929 cells and parasites eluted at distinctly different NaClconcentrations (Fig. 3). The elution profile for C. bumetiishows at least one major peak (middle region) and indica-tions of two minor peaks.

Molecular mass of C. burnetii ACP. The apparent nativemolecular mass of the predominant C burnetii species wasdetermined by high-performance liquid chromatography tobe approximately 74 kDa (Fig. 4). When the host cell extractwas chromatographed on the same column, three distinctpeaks of ACP activity were observed; they eluted in frac-tions corresponding to molecular masses of 2.5, 10.5, and 50

I kDa. All three of these peaks eluted after the Coxiella ACP3 4 5 6 7 8 (data not shown). Since few enzymes are 10 kDa or smaller,

pH we speculate that the relatively low apparent molecularEffect of hydrogen ion concentration on C. burnetii ACP masses of two of the three host cell ACPs may be attributedine Mile supernatant fluid (from 100,000 x g centrifuga- to their interaction with the column resin. Supernatant fluidsrepared and chromatographed through QAE-Sephadex (0 from 100,000 x g centrifugation of the Nine Mile rickettsiaeTaCi, 10 mM phosphate buffer [pH 6.4]) as described in and L929 cells were analyzed in a stationary-phase SuperdexLnd Methods and in the legend to Fig. 3. The ACPs in the G-75 column eluted with phosphate-buffered saline at 120btained were pooled, and the activity at various pHs was lb/in2. Eluted fractions were assayed for phosphatase activ-Iby using the following buffers: 0.2 M sodium acetate (pH ity. Molecular mass standards were also chromatographed.2 M sodium cacodylate (pH 6), and 0.2 M Tris-HCI (pH on the column to establish the molecular mass of the ACP.The Nine Mile organisms were isolated from L929 cells Effect of a heteromolybdate complex on C. burnetii and hostbeen infected for 3,052 days.

cell phosphatase activity. It was demonstrated previously thata number of heteromolybdate compounds inhibit ACP activ-ity (23). The potential inhibitory effects of various chemicals

m C. bumetii ACP actvlty was observed at pH 5.0 on the ACP activity associated with intact Nine Mile C.ex g supernatant fluids derived from sonicated bumnetii and supernatant fluids from 10,000 x g centrifuga-e C.buvaetio organisms (Fig. 1). Assays were tion of infected and uninfected L929 cells were investigated

strate (Table 2). The compounds tested included sodium tartrate,e of ACP in the C. burnetii periplasmic space. By sodium fluoride, and two heteromolybdates synthesized

cytochemical technique of Gomori (2, 15), the ACP by Michael Pope, Georgetown University. The two het-ras detected only in the periplasmic space of the eromolybdates studied are designated as complex B'

e and S Q217 isolates of C bumetii. Purified {[C(NH2)3]2(CH3)2AsMo4O15H} and complex E2 [(NH4)6or parasites present in phagolysosomes of L929 As2Mo18062xH2O]. Complex B' significantly reduced theincubated with the phosphate substrate (13-glycer- ACP activity of intact C. bumetii (over 50% inhibition) and

te orp-nitrophenylphosphate- lead nitrate was the only slightly inhibited the ACP activity of uninfected L929tgent). Following incubation, the samples were cells. Although complex E2 seemed to have a slightly stron-for electron microscopy (2) and thin sections were ger inhibitory effect on the C. burnetii ACP (intact organ-by transmission electron microscopy. The elec- isms) than on that of the uninfected host cell, the differenceographs (Fig. 2) clearly showed the presence Of is probably not significant. The effect of the molybdatelense reaction products (lead phosphate) in the compounds on the C. burnetii ACP activity present in

lic space. With the glycerophosphate substrate 100,000 supernatant fluids was even more pronounced. Forately 10% of rickettsiae gave a positive result; with example, 25 and 50 ,uM concentrations of B' inhibited the C.atylphosphate, more than 50% did so. burnetii ACP by 59 and 70%, respectively; 25 and 50 ,uMbange chromatography clearly distinguished C. bur- concentrations of E2 inhibited the rickettsial ACP by 72 andphatase from that of the host cell. Purified Nine Mile 80%, respectively. The differential effect of these two heter-ii organisms were disrupted by sonication while omolybdate compounds is an additional indication that theI in 0.01 M HEPES buffer (pH 7.4)-0.8% NaCl. ACPs of parasite and host are distinct; it is also indirectpted organisms were centrifuged at 100,000 x g at evidence that the ACP associated with the parasite is C.h, and the supematant fluid was eluted through bumnetii specific (i.e., produced by the parasite). Otherhadex with a linear gradient ranging from 0 to 0.5 potential phosphatase inhibitors (sodium fluoride and so-Supernatant fluids from normal uninfected L929 dium tartrate) were examined; however, neither compound

FIG. 2. Detection of ACP in the periplasmic spaces of the S Q217 (A) and Nine Mile (B and C) isolates of C. burnetii present inphagolysosomes (panels A and B) of L929 fibroblasts or purified from L929 cells (panel C). The electron-dense reaction products (arrows)indicate the presence of ACP and were generated by usingp-nitrophenylphosphate (panel A) or 1-glycerophosphate (panels B and C) as thesubstrate. Organisms in panel D are control rickettsiae processed at the same time as those in panel C but without substrate. The L929 cellshad been continuously infected for 1,058 days (A) or 2,562 days (B); the purified C. bumetii organisms (panels C and D) were from L929 cellsthat had been continuously infected for 2,893 days. Bars, 2 pum (panels A and D), 0.6 ,um (panel B), and 1.1 pum (panel C).

INFECT. IMMUN.

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ACID PHOSPHATASE ACTIVITY IN COXIELLA BURNETII 4235

N_&WqL _;

ir

VOL. 61, 1993

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

4236 BACA ET AL.

900-

800-

E 700-

= 600-z

500-

> 400-

C 300-IL 200-n(ioo I

1'KR6

Ro3 5' 7 9 11 3 15 17 192 23 2527 29 31 33 35

FRACTION NUMBERFIG. 3. Chromatography of C. bumnetii and L929 ACPs on a

QAE-Sephadex column. Purified Nine Mile C. burnetii organisms (9mg or 3.3 x 1011 rickettsiae) in 3 ml of HEPES buffer (0.01 M [pH7.4] with 0.8% NaCl) were lysed by being subjected to sonication(Branson Sonifier; 30 s with the microprobe). L929 cells (7.3 x 108,in 10 ml of HEPES buffer-0.8% NaCl) were disrupted in a TenBroek homogenizer maintained at 4°C. After centrifugation at100,000 x g for 1 h at 4°C, the supernatant fluids from the parasitesand L929 cells were independently chromatographed through aQAE-Sephadex column (30 mm by 10 mm [inner diameter]) with aNaCl linear gradient ranging from 0 to 0.5 M, which was initiated atfraction 3 and terminated at fraction 35. Fractions (0.5 ml) werecollected and assayed for ACP activity. The assays were carried outat 37°C in 0.1 ml of reaction mixture containing 0.2 M sodiumacetate buffer (pH 5.5) and 5 mM MUP as substrate. O, L929 ACPactivity; *, C. bumetii ACP activity.

TABLE 2. Differential effect of two molybdate compounds onthe ACP activity of Nine Mile C. bumetii and

uninfected L929 cellsa

Sp act (U/mg of protein)bInhibitor

L929 supernatante C. bumetii

None 1,120 ± 232 1,380 + 94Complex B' 1,090 + 113 652 ± 80Complex E2 676 ± 253 533 ± 88

a The compounds were tested in the standard ACP assay with MUP as thesubstrate. The complex B' concentration was 100 p.M; the E2 concentrationwas 50 pM.

b Results represent the means ± standard deviations of four independentexperiments.

c Uninfected L929 cells were suspended in HEPES buffer (0.01 M [pH 7.4]with 0.8% NaCi), homogenized with the aid of a Ten Broek homogenizer, andcentrifuged at 10,000 x g for 1 h at 4'C. The supernatant fluid was assayed forACP activity.

rickettsiae. The supernatant fluid contained significant ACPactivity (1,488 U/mg of protein). Human neutrophils wereprepared from peripheral blood as described in Materials andMethods. Superoxide anion production by formyl-Met-Leu-Phe-stimulated neutrophils was measured by a continuouscytochrome c reduction method (20). In the presence of 10 p,lof the C. bumnetii supernatant preparation (which contained4 U of phosphatase activity), the neutrophils (106 cells in 1.0ml of buffer) did not generate significant levels of superoxide

significantly affected the activity of the enzymes of theparasite and host cell.

Inhibition of superoxide anion production by human neu-trophils coincubated with C. burnetii supernatant fluid. Super-natant fluids were prepared from sonicated Nine Mile Cburnetii organisms by subjecting the sonicates to centrifuga-tion at 10,000 x g for 1 h to remove cell debris and intact

60 v

-lE

: 50-

0 s-t40-wco

I- 30-0.

Q-.20-C

VO

1 6 8 10 12 14 16 18 20 24 26FRACTION NUMBER

FIG. 4. Apparent molecular mass of C. burnetii ACP. The mo-lecular mass of C. burnetii ACP was determined by high-perfor-mance liquid chromatography as described in the text. The ACP wasfrom rickettsiae derived from L929 cells that had been infected forapproximately 3,000 days.

0.10-

LO 0.08-

LU() 0.06-zm 0.04-m0Cl) 0.02-

0 t1LLLI[0 30' 60- 90 120 150 180 210 240 270

TIME (sec)FIG. 5. Apparent inhibition of superoxide anion production by

human neutrophils coincubated with C. burnetii supernatant fluid.Sonicated (30 s in a Branson S75 Sonifier fitted with a microprobe)Nine Mile C bumetii organisms (3 mg of purified rickettsiae per ml)were centrifuged at 10,000 x g for 1 h. The resulting supernatantfluid was assayed for ACP activity (1,488 U/mg of protein). Humanneutrophils were prepared from peripheral blood by dextran sedi-mentation followed by density gradient centrifugation on Ficoll-Hypaque gradients and isotonic NH4C1 lysis (20, 24). The generationof superoxide anions by formyl-Met-Leu-Phe-stimulated neutrophilswas measured as superoxide dismutase-inhibitable cytochrome c(0.05 mM) reduction by using a continuous assay method (20).Experimental neutrophils (0) (106) in 1.0 ml of Krebs-Ringer-Phosphate solution (pH 7.4) were preincubated at 37°C (2 to 3 min)with 10 p,l (4 U of phosphatase activity) of the C burnetii superna-tant fluid; control untreated (-) neutrophils (106) received 10 PI1 ofbuffer. Measurement of superoxide anion production began with theaddition of 10-7 M formyl-Met-Leu-Phe. Absorbance readings weremade at 30-s intervals during the 5-min incubation period. Theuntreated neutrophils produced significant amounts of superoxideanion (i.e., increased absorbance) during the course of the experi-ment, whereas the experimental neutrophils treated with C. bumetiisupernatant fluid did not (i.e., they exhibited only the innatebaseline absorbance, which was approximately 0.07).

IN ECT. IMMUN.

0 0

NM 0

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ACID PHOSPHATASE ACTIVITY IN COXIELLA BURNETII 4237

0.002

FIG. 6. Suposed to C. band absence cwere purifiedfor ca. 3,000 dLeu-Phe-stimiand after exp(of ACP actividates E2 (A)approximatel)

DAYS TREATEDFIG. 7. Treatment of C bumnetii-infected L929 cells with heter-

30 60 120 150 180 210 240 270 omolybdate complex B'. L929 cells continuously infected for ap-proximately 3,200 days with Nine Mile C. burnetii organisms were

TME (see) treated with complex B' (20 SLg/ml [0] and 40 .ig/ml [V]) for 15 days;

iperoxide anion production by human neutrophils ex- control infected cells (0) received the solvent (water) only. Aliquotsumnetii 100,000 x g supernatant fluid in the presence were removed on the indicated days, and the cells were stained and)f heteromolybdates. Nine Mile C. burnetii organisms examined microscopically to determine the percentage of infectedfrom L929 cells that had been continuously infected cells (see Materials and Methods). L929 cell viabilities were deter-lays. Superoxide anion production by the formyl-Met- mined at each sampling and were .98%. Control, uninfected L929ulated neutrophils was assayed before (control) (0) cell populations also exhibited .98% viability (data not shown).Dsure to 100,000 x g supernatant fluid containing 5 U Each point on the curves represents the average of two indepen-ty alone (@) or in combination with the heteromolyb- dently treated (or control) flasks. The infectivity counts obtained forand B' (A). The innate baseline absorbance was each point did not vary by more than 10%. These results are

y 0.07. representative of two independently conducted experiments.

anion during the course of the continuous assay whencompared with control untreated neutrophils (Fig. 5). Simi-lar results were obtained with 50 pl of the C. burnetiisupernatant preparation.When C. bumetii supernatant fluid from centrifugation at

10,000 x g was boiled, its capacity to inhibit superoxideanion production by neutrophils (data not shown) was de-stroyed; this is evidence that the inhibitor is probably aprotein (i.e., an enzyme such as ACP).Although the enzyme has not been purified to homogene-

ity, it was enriched sevenfold by subjecting the Nine Mile C.burnetii 10,000 x g supernatant fluid to 100,000 x g centrif-ugation for 60 min followed by chromatography on a hydrox-ylapatite column. These enrichment steps resulted in con-comitantly enhanced capacity to inhibit superoxide anionproduction by human neutrophils (data not shown).That ACP in the supernatant fluid is responsible for the

suppressed anion production is also strongly indicated by thefollowing results, obtained with the heteromolybdate phos-phatase inhibitors.

Inhibitory effect of C. burnetii supernatant fluid on super-oxide anion production by neutrophils blocked by heteromo-lybdate compounds. Inclusion of either heteromolybdate E2or B' blocked the C. bumnetii supernatant fluid from inhibit-ing superoxide anion production by human neutrophils (Fig.6). That these compounds have been shown to inhibit ACPsfrom other sources (23) implicates the C burnetii ACP as theinhibitory factor in the supernatant fluid. This conclusion isindirectly supported by the following results.

Heteromolybdate complex B' reduced the level of C. burnetiiinfection. Treatment of C bumetii-infected L929 cell popu-lations with the heteromolybdate complex B' for a period of15 days resulted in markedly decreased levels of infection(Fig. 7). The experiment depicted shows that the percentageof infected cells decreased in a dose-dependent manner

down to the 20 to 30% range after 11 days of treatment;untreated control populations were approximately 80% in-fected. After 15 days of continuous treatment, about 40% ofthe L929 cells were infected; in contrast, approximately 90%of the control untreated cells were infected. Also notable isthe observation that the treated populations continued toproliferate with similar viabilities (298%) as control un-treated infected (and uninfected) cells; this indicates that thechemical was relatively nontoxic to the host cell.

DISCUSSION

In this report evidence has been presented that C. bumetiipossesses innate ACP activity. The enzyme was detected inseveral isolates representative of strains implicated in acuteor chronic disease and was localized in the periplasmicspace. The level of activity found in the parasite far exceedsthat reported for other organisms, including Legionella andLeishmania spp. (21, 24). Supematant fluid derived from theparasite significantly depressed the metabolic burst of hu-man neutrophils stimulated by the formylated peptideformyl-Met-Leu-Phe; inclusion of heteromolybdate phos-phatase inhibitors relieved the inhibition. That the hetero-molybdate complex B' reduced the level of infection of hostL929 cells indicates that it may be specifically targeting theCoxiella ACP while sparing the host cell ACP. This isconsistent with the observation that the compound had agreater inhibitory effect on the parasite ACP than on that ofthe L929 fibroblasts. Collectively, these results stronglyindicate that ACP is the factor responsible for the inhibitoryeffect of the rickettsial extract.

Previously we conclusively demonstrated that C. bumetiigrows within phagolysosomes of host cells (2). Because ofthe detection of typical lysosomal markers ACP and 5'-nucleotidase activity within parasite-containing vacuoles,

0.16

?. 0.12-

. 0.10-

Ul

.4 0.06-

0.02-

VOL. 61, 1993

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

4238 BACA ET AL.

Burton et al. (10, 11) had earlier concluded that C. burnetii(Nine Mile isolate) proliferated within phagolysosomes ofhost cells. They also reported that the rickettsiae generatedACP reaction product within the periplasmic space whengrown in Vero cells but did not do so when grown in L929cells. These authors speculated that the pleomorphic appear-ance of the rickettsiae within Vero cells indicated that theywere undergoing degradation by host lysosomal enzymesand that the ACP in the parasite periplasm was of hostorigin, having penetrated into the parasites (no viabilitystudies were performed). As shown in our studies, we havediscovered ACP activity in several isolates of C. burnetiigrown in two different host systems, L929 cells and chickenembryos (Table 1). That Burton et al. (11) did not detectACP in the periplasm of C. burnetii organisms grown in L929cells may be explained by their use of P-glycerophosphate asthe substrate. As shown above, we found that use of,B-glycerophosphate resulted in only about 1 in 10 rickettsiaepositive for ACP in the periplasmic gap whereas use ofp-nitrophenylphosphate revealed about 50% ACP-positiveorganisms. These results suggest that thep-nitrophenylphos-phate penetrates into the periplasmic gap better than the-glycerophosphate does; alternatively, p-nitrophenylphos-

phate may be a significantly better substrate for ACP. Theinterpretation by Burton et al. that the ACP they detected inthe C. bumetii periplasm (grown in Vero cells) is hostderived is probably incorrect; in all probability the phos-phatase is encoded in and produced by the parasite.The importance of the oxidative metabolic burst in the

destruction of intracellular parasites by phagocytes has beendemonstrated by several investigators (4, 5, 9, 19, 27). Werecently reported that phagocytosis of opsonized or unop-sonized C. bumetii organisms failed to trigger a significantproduction of superoxide anion by human neutrophils (3).By failing to elicit an adequate metabolic burst (as indicatedby superoxide anion production) during phagocytosis byneutrophils, C burnetii resembles other intracellular patho-gens such as Mycobactenium leprae (17), Toxoplasma gondii(34), Leishmania donovani (16), and Legionella micdadei(13, 18, 31), an organism that is phylogenetically related toC. burnetii (30, 32). Successful parasitization of phagocytesby C. bumetii may be due, in part, to the inability of thephagocyte to generate adequate concentrations of microbi-cidal oxygen metabolites during and after rickettsial entry.That the C. bumnetii ACP may play a role in shutting downthe metabolic burst and concomitant superoxide anion pro-duction is indicated by these studies. One should bear inmind the likelihood that the C bumetii periplasmic ACPplays a role in dephosphorylating various compounds, e.g.,host nucleotides, that may subsequently be rephosphory-lated during or after transit through the cytoplasmic mem-brane of the parasite. We are now in the process of purifyingand further characterizing the ACP. One of our first goalswill be to investigate the substrate specificity of the purifiedenzyme.

ACKNOWLEDGMENTSWe thank Louis Mallavia for providing the purified egg-grown C.

burnetii organisms used in this study. We are also indebted toMichael Pope, Department of Chemistry, Georgetown University,Washington, D.C., for his generous gift of the heteromolybdatecompounds. We gratefully acknowledge the assistance provided byAndrzej Pastuzyn and colleagues of the Department of BiochemistryProtein Core Facility, UNM School of Medicine.

This work was supported by U.S. Public Health Service grantR01-AI-32492 from the National Institute of Allergy and Infectious

Diseases and by a grant (to R.H.G.) from the Research AllocationsCommittee of the UNM School of Medicine.

REFERENCES1. Akporiaye, E. T., and 0. G. Baca. 1983. Superoxide anion

production and superoxide dismutase and catalase activities inCoxiella bumetii. J. Bacteriol. 154:520-523.

2. Akporiaye, E. T., J. D. Rowatt, A. A. Aragon, and 0. G. Baca.1983. Lysosomal response of a murine macrophage-like cell linepersistently infected with Coaxiella bumetii. Infect. Immun.40:1155-1162.

3. Akporiaye, E. T., D. Stefanovich, V. Tsosie, and 0. Baca. 1990.Coxiella bumetii fails to stimulate human neutrophil superoxideanion production. Acta Virol. 34:64-70.

4. Allen, R. C. 1977. Evaluation of serum opsonic capacity byquantitating the initial chemiluminescent response from phago-cytizing polymorphonuclear leukocytes. Infect. Immun. 15:828-833.

5. Babior, B. M., R. S. Kipnes, and J. T. Curnutte. 1973. Biologicaldefense mechanisms: the production by leukocytes of superox-ide, a potential bactericidal agent. J. Clin. Invest. 52:741-744.

6. Baca, O., T. Scott, E. Akporiaye, R. DeBlassie, and H. Crssman.1985. Cell cycle distribution patterns and generation times ofL929 fibroblast cells persistently infected with Coxiella bumetii.Infect. Immun. 47:366-369.

7. Baca, 0. G., E. T. Akporiaye, A. S. Aragon, I. L. Martinez,M. V. Robles, and N. L. Warner. 1981. Fate of phase I and phaseII Coxiella bumetii in several macrophage-like cell lines. Infect.Immun. 33:258-266.

8. Bradford, M. M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of protein utilizing theprinciple of protein-dye binding. Anal. Biochem. 72:248-254.

9. Buchmuller, Y., and J. Mauel. 1981. Studies on the mechanismsof macrophage activation: possible involvement of oxygen me-tabolites in killing of Leishmania enrietii by activated mousemacrophages. J. Reticuloendothel. Soc. 29:181-192.

10. Burton, P. R., N. Kordova, and D. Paretsky. 1971. Electronmicroscopic studies of the rickettsia Coxiella bumetii: entry,lysosomal response, and fate of rickettsial DNA in L-cells. Can.J. Microbiol. 17:143-158.

11. Burton, P. R., J. Stueckemann, R. M. Welsh, and D. Paretsky.1978. Some ultrastructural effects of persistent infections by therickettsia Coxiella bumetii in mouse L cells and green monkeykidney (Vero) cells. Infect. Immun. 21:556-566.

12. Das, S., A. K. Saha, A. T. Remaley, R. H. Glew, J. N. Dowling,M. Kajiyoshi, and M. Gottlieb. 1986. Hydrolysis of phosphopro-teins and inositol phosphates by cell surface phosphatase ofLeishmania donovani. Mol. Biochem. Parasitol. 20:143-153.

13. Dowling, J. N., A. K. Saha, and R. H. Glew. 1992. Virulencefactors of the family Legionellaceae. Microbiol. Rev. 56:32-60.

14. Glew, R. H., M. S. Czuczman, W. F. Diven, R. L. Berens, M. T.Pope, and D. E. Katsoulis. 1982. Partial purification and charac-terization of particulate acid phosphatase of Leishmania dono-vani promastigotes. Comp. Biochem. Physiol. 72B:581-590.

15. Gomori, G. 1952. Microscopic histochemistry, principles andpractice, p. 189-194. University of Chicago Press, Chicago.

16. Haidaris, C. G., and P. F. Bonventre. 1982. A role for oxygen-dependent mechanisms in killing ofLeishmania donovani tissueforms by activated macrophages. J. Immunol. 129:850-855.

17. Holzer, T. J., K. E. Nelson, V. Schauf, R. G. Crispen, and B. R.Andersen. 1986. Mycobacterium leprae fails to stimulate phago-cytic cell superoxide anion generation. Infect. Immun. 51:514-520.

18. Horwitz, M. A., and S. C. Silverstein. 1981. Interaction of thelegionnaires' disease bacterium (Legionella pneumophila) withhuman phagocytes. II. Antibody promotes binding of L. pneu-mophila to monocytes but does not inhibit intracellular multi-plication. J. Exp. Med. 153:398-406.

19. Murray, H. W., and Z. A. Cohn. 1979. Macrophage oxygen-dependent antimicrobial activity. I. Susceptibility of Toaxo-plasma gondii to oxygen intermediates. J. Exp. Med. 150:938-949.

20. Newburger, P. E., M. E. Chovaniec, and H. J. Cohen. 1980.

INFECT. IMMUN.

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

ACID PHOSPHATASE ACTIVITY IN COXIELL4 BURNETII 4239

Activity and activation of the granulocyte superoxide-generat-ing system. Blood 55:85-92.

21. Remaley, A. T., D. B. Kuhns, R. H. Basford, R. H. Glew, andS. S. Kaplan. 1984. Leishmanial phosphatase blocks neutrophil02 production. J. Biol. Chem. 259:11173-11175.

22. Roman, M. J., P. D. Coriz, and 0. G. Baca. 1986. A proposedmodel to explain persistent infection of host cells with Coxiellabumetii. J. Gen. Microbiol. 132:1415-1422.

23. Saha, A. K., D. C. Crans, J. N. Dowling, M. T. Pope, C. M.Simone, and R. H. Glew. 1991. Inhibition of human seminal fluidand Leishmania donovani phosphatases by molybdate het-eropolyanions. J. Biol. Chem. 266:3511-3517.

24. Saha, A. K., J. N. Dowling, K. L. LaMarco, S. Das, A. T.Remaley, N. Olomu, M. Pope, and R. H. Glew. 1985. Propertiesof an acid phosphatase from Legionella micdadei which blockssuperoxide anion production by human neutrophils. Arch. Bio-chem. Biophys. 243:150-160.

25. Saha, A. K., J. N. Dowling, A. W. Pasculle, and R. H. Glew.1988. Legionella micdadei phosphatase catalyzes the hydrolysisof phosphatidylinositol 4,5-bisphosphate in human neutrophils.Arch. Biochem. Biophys. 265:94-104.

26. Samuel, J. E., M. E. Frazier, and L. Mallavia. 1985. Correlationof plasmid type and disease caused by Coxiella bumetii. Infect.Immun. 49:775-779.

27. Sasada, M., and R. B. Johnston, Jr. 1980. Macrophage micro-bicidal activity: correlation between phagocytosis-associatedoxidative metabolism and the killing of Candida by macro-phages. J. Exp. Med. 152:85-98.

28. Silberman, R., and P. Fiset. 1968. Method for counting rickett-siae and chlamydiae in purified suspensions. J. Bacteriol. 95:

259-261.29. Thompson, H. A., 0. G. Baca, and D. Paretsky. Presence of

ribosomal ribonucleic acid in the rickettsia Coxiella bumetii.Biochem. J. 125:365-366.

30. Tzianabos, T., C. W. Moss, and J. E. McDade. 1981. Fatty acidcomposition of rickettsiae. J. Clin. Microbiol. 13:603-605.

31. Weinbaum, D. L., R. R. Benner, J. N. Dowling, A. Alpern,A. W. Pasculle, and G. R. Donowitz. 1984. Interaction ofLegionella micdadei with human monocytes. Infect. Immun.46:68-73.

32. Weisburg, W., M. Dobson, J. Samuel, G. Dasch, L. Mallavia, 0.Baca, L. Mandelco, J. Sechrest, E. Weiss, and C. Woese. 1989.Phylogenetic diversity of the rickettsias. J. Bacteriol. 171:4202-4206.

33. Williams, J. C., M. G. Peacock, and T. F. McCaul. 1981.Immunological and biological characterization of Coxiella bur-netii, phases I and II, separated from host components. Infect.Immun. 32:840-851.

34. Wilson, C. B., V. Tsai, and J. S. Remington. 1980. Failure totrigger the oxidative metabolic burst by normal macrophages.Possible mechanism for survival of intracellular pathogens. J.Exp. Med. 151:328-346.

35. Yeaman, M. R., L. A. Mitscher, and 0. G. Baca. In vitrosusceptibility of Coxiella burnetii to antibiotics, including sev-eral quinolones. Antimicrob. Agents Chemother. 31:1079-1084.

36. Yeaman, M. R., M. J. Roman, and 0. G. Baca. 1989. Antibioticsusceptibilities of two Coaxiella burnetii isolates implicated indistinct clinical syndromes. Antimicrob. Agents Chemother.33:1052-1057.

VOL. 61, 1993

on Septem

ber 15, 2020 by guesthttp://iai.asm

.org/D

ownloaded from