INFECTION AND IMMUNITY, Jan. 1985, p. 157-165 Vol. 47, No. 1 0019-9567/85/010157-09$02.00/0 Copyright C) 1985, American Society for Microbiology Cosmid Cloning of prowazekii Antigens in K-12 DUNCAN C. KRAUSE, HERBERT H. WINKLER, AND DAVID 0. WOOD* Department of Microbiology and Immunology, University of South Alabama College of Medicine, Mobile, Alabama 36688 Received 24 May 1984/Accepted 4 October 1984

Rickettsia prowazekii DNA was partially digested with Sau3A or Hindlll, ligated with the cosmid pHC79, packaged in vitro, and transduced into Escherichia coli HB101. Cosmid cloning of Sau3A-digested rickettsial DNA yielded 1,288 ampicillin-resistant colonies; 798 cosmid clones resulted with HindlIl-digested rickettsial DNA. Chimeric cosmid DNA was extracted from the latter bank, digested to completion with HindIII, and compared by agarose gel electrophoresis with a HindIII digest of rickettsial genomic DNA. The two digestion profiles were quite similar in their overall banding patterns, indicating that the clone bank was significantly representative of the rickettsial . When both clone banks were screened for expression of rickettsial antigens by -linked immunosorbent assay with goat anti-R. prowazekii serum, ca. 20% of the clones reacted positively. Two clones were randomly selected for more detailed analysis. Each contained a large chimeric plasmid (40.2 and 38.1 kilobases) which apparently yielded smaller deletion derivatives (13.6 and 12.6 kilobases) when transformed into an E. coli minicell strain. Each recombinant plasmid directed the synthesis of new protein not observed in control minicells. One of the clones produced a 51,000-dalton protein in minicells, which comigrated with a protein reactive with anti-R. prowazekii serum. This protein was not present in negative controls. When to this protein were incubated with a of rickettsial total protein, they bound to a 52,000-dalton polypeptide. Hence, the cloned rickettsial gene product in E. coli corresponds to a protein of similar size in R. prowazekii. This study demonstrates the feasibility of cosmid cloning of rickettsial antigens in E. coli.

Rickettsia prowazekii is an obligate intracellular bacterium 30). Furthermore, interferon gamma levels have been dem- capable of parasitizing a number of different cell types, onstrated to increase in mice immunized with Rickettsia including both professional and nonprofessional phagocytes tsutsugamushi (19). (8, 25, 27, 33, 40). Unlike other intracellular , both Less is known about the rickettsial moieties responsible facultative and obligate, rickettsiae grow directly within the for eliciting cellular and humoral immune responses. Dasch cytoplasm of their rather than within phagosomes or and Bourgeois (6) have reported that the species-specific phagolysosomes. Rickettsiae adsorb to the surface of host protein antigens evoke strong cellular and humoral re- cells (20, 39), triggering internalization via "induced phago- sponses in the guinea pig. Burans and Dasch have subse- cytosis" (33). Once inside the cell, the rickettsiae rapidly quently identified additional rickettsial proteins which are escape from the phagosome into the host's cytoplasm, recognized by antibodies from human convalescent presumably by means of a phospholipase A activity (37, 38). sera (Abstr. Annu. Meet. Am. Soc. Microbiol. 1984, I46, p. R. prowazekii possesses a variety of means by which it can 129). Further definition of the role of specific rickettsial exploit the nutrient-rich environment of its host's cytoplasm. proteins (as well as nonproteinaceous moieties) in both the Although some of these mechanisms, such as lysine and host response to and the pathogenesis of typhus are ham- proline transport systems (23, 36), can be found in common pered by difficulties in obtaining sufficient quantities of extracellular bacteria, others, such as the ADP-ATP ex- purified proteins for expanded studies. change system (35) and an adenylate transporter (manuscript The advent of recombinant DNA technology has facili- in preparation), appear to be unique, if not to the genus tated the study of difficult to cultivate and Rickettsia, then to obligate intracellular bacteria. manipulate in the , such as Treponema pallidum It is apparent that host defense against rickettsial infec- and Chlamydia trachomatis (24, 34). Recently, this labora- tions involves both humoral and cell-mediated immunity. tory successfully cloned the R. prowazekii gene for citrate Pretreatment of R. prowazekii with opsonizing antibodies synthase and demonstrated its expression in Escherichia coli results in their destruction after phagocytosis by human (42). Here we report the formation of cosmid clone banks of macrophages; unopsonized rickettsiae survive and multiply the R. prowazekii genome and the identification of clones within macrophages (1). On the other hand, although im- expressing for rickettsial antigens. Two antigen-posi- mune serum fails to transfer resistance to intradermal chal- tive clones were selected for more detailed analysis. lenge with in guinea pigs, resistance is (Preliminary reports of this work were presented at the transferred by immune spleen cells (17). In addition, rickett- Fourth National Conference of the American Society for siae are destroyed by both macrophages (18, 30a, 41) and and Diseases, Airlie House, Va., cultured fibroblasts (28, 41) which have been treated with Rickettsiology Rickettsial lymphokines from concanavalin A-treated or rickettsial an- 27 to 30 October 1983, and the Annual Meeting of the tigen-treated mouse spleen cells. The antirickettsial activity American Society for Microbiology, St. Louis, Mo., 4 to 9 has subsequently been attributed to interferon gamma (29, March 1984. At the American Society for Rickettsiology and Rickettsial Diseases Conference, preliminary data were also presented by M. E. Dobson, G. A. Dasch, and J. P. Burans * Corresponding author. on the cloning of R. typhi antigens in E. coli.) 157 158 KRAUSE, WINKLER, AND WOOD INFECT. IMMUN.

TABLE 1. Bacterial strains Strain Description Source (reference) E. coli HB101 F- hsdS20 (rB-, mB-) recA13 ara-14 proA2 lacYl galK2 leuB6 rpsL20 xyl-5 mtl-l thi-I ginV44 D. Kopecko" (2)

X2687T F- hsdS20 (rB-, mB-) recA13 ara-14 proA2 lacYl galK2 leuB6 rpsL20 xyl-5 mtl-l thi-i ginV44 R. Curtiss 111b X c1857 BHB2688 F- recA [X imm434 clts b2 red3 Eam4 Sam7] R. Curtiss III (5) BHB2690 F-recA [X imm434 clts b2 red3 Daml5 Sam7] R. Curtiss III (5) MOB134 HB101 containing pHC79 R. Welch' (11) X1849 F- tonA53 dapD8 minAl purE41 ginV42 A(gal-uvrB)47 minB2 his-53 nalA25 metC65 oms-i T3r R. Curtiss III (9) iA(bioH-asd)29 ilv-277 cycB2 cycAl hsdR2 X- a Walter Reed Army Institute of Research, Washington, D.C. b Department of Biology, Washington University, St. Louis, Mo. ' Department of Medical Microbiology, University of Wisconsin Medical School, Madison.

MATERIALS AND METHODS bromide equilibrium density gradient centrifugations. Rick- ettsial DNA was extracted according to Meyers and Wisse- Bacterial strains and culture conditions. The bacterial man (16), with modifications. Briefly, the pelleted rickettsiae strains used in this study are described in Table 1. R. were suspended in SSC (0.15 M NaCl plus 0.015 M sodium prowazekii Madrid E was cultivated in the yolk sac of anti- citrate [pH 7.0]) to an optical density at 440 nm of 2.0. biotic-free, embryonated hen eggs and purified as described Protease type V was added to a final concentration of 1 elsewhere (35). Purification included Renograffin density gra- mg/ml and incubated at room temperature for 10 min. A dient centrifugation to remove contaminating yolk sac mito- solution of 25% SDS was added to a final concentration of chondria. Rickettsiae were stored as a pellet at -70°C before 0.2% and mixed by gentle inversion. The rickettsial lysate extraction of the DNA. E. coli strains were cultured in LB was incubated for 6.5 h at 37°C. An equal volume of phenol broth and on LB agar plates (7). Where appropriate, ampi- saturated with SSC and containing 0.1% hydroxyquinoline cillin, , and streptomycin were added at final was added and incubated for 60 min at 37°C with moderate concentrations of 50, 25, and 100 ,ug/ml, respectively. Except mixing. The lysate was centrifuged (10,000 x g at 20°C for 20 where indicated below, the E. coli strains were grown at min), and the aqueous layer was collected. The DNA was 370C. precipitated with ethanol at -20°C, suspended in 0.1 x SSC, and reagents. Ampicillin, tetracycline, strep- and dialyzed against 50 mM phosphate buffer (pH 7.0). tomycin, protease type V, RNase A, Trizma base, sucrose, RNase A was added to a final concentration of 50 ,ug/ml and spermidine hydrochloride, putrescine hydrochloride, 3-mer- incubated for 2.5 h at 37°C. The DNA was further purified by captoethanol, dimethyl sulfoxide, ATP, hydroxyquinoline, CsCl equilibrium density gradient centrifugation and dia- lysozyme, Tween 20, biotin, diaminopimelic acid, adenine, lyzed against TE buffer (10 mM Tris, 1 mM disodium EDTA tryptophan, methionine, cysteine, asparagine, glutamine, [pH 8.0]). and ethidium bromide were purchased from Sigma Chemical Chimeric cosmid DNA was isolated from individual clones Co., St. Louis, Mo. Technical grade cesium chloride was or from a pool of clones according to Hansen and Olsen (10). obtained from Kawecki Berylco Industries, New York, In the latter case, cosmid clones were replica plated from the N.Y. Sequanal grade sodium dodecyl sulfate (SDS) was 96-well microtiter dishes in which they were stored onto LB obtained from Pierce Chemical Co., Rockford, Ill. Agarose, agar plates containing ampicillin. After an overnight incuba- all restriction endonucleases, T4 DNA ligase, and dithiothre- tion at 32°C, the resulting colonies were scraped into phos- itol were purchased from Bethesda Research , phate-buffered saline, pH 7.2 (PBS) before extraction of the Rockville, Md. DNase I was obtained as indicated in the plasmid DNA. Extracted plasmid DNA was further purified text. HATF filters were obtained from Millipore Corp., by CsCI-ethidium bromide equilibrium density gradient cen- Bedford, Mass. Nitrocellulose was purchased from Schlei- trifugation. DNA was analyzed on vertical gels of 0.7 or cher & Schuell, Inc., Keene, N.H. Acrylamide, bis-acrylam- 0.9% agarose in 90 mM Tris-90 mM borate-3 mM EDTA ide, ammonium persulfate, N,N,N',N'-tetramethylethylene- (pH 8.1). diamine, and HRP color development reagent were obtained Formation of cosmid clone banks of R. prowazekii DNA. All from Bio-Rad Laboratories, Richmond, Calif. Triton X-100 recombinant DNA experiments were conducted according was obtained from Research Products International Corp., to guidelines of the National Institutes of Health (Fed. Elk Grove Village, Ill. Register, vol. 48, no. 106, 1 June 1983). Rickettsial DNA was Preparation of DNA. Cosmid vector pHC79 (11) was used digested with restriction endonuclease (either Sau3A or as the cloning vehicle. This plasmid contains the bacterio- HindIII) under conditions which optimized the yield of DNA phage lambda-derived cos site and genes conferring resist- fragments in the 30- to 45-kilobase (kb) range, as determined ance to ampicillin and tetracycline. E. coli containing pHC79 by agarose gel electrophoretic analysis of the digestion was cultured, and plasmid copy number was amplified products. The cosmid vector pHC79 was digested to com- according to Maniatis et al. (15). Plasmid DNA was ex- pletion with either BamHI (for Sau3A-digested rickettsial tracted by the cleared lysate procedure of Clewell and DNA) or HindIII, for both of which it possesses unique Helinski (4) and purified by two successive CsCl-ethidium cloning sites. Sau3A-digested rickettsial DNA was centri- VOL. 47, 1985 COSMID CLONING OF RICKETTSIAL ANTIGENS 159 fuged on a linear 10 to 40% sucrose gradient in 1 M NaCI-5 Whatman 3MM paper saturated with 0.1 M NaHCO3-1% mM EDTA-20 mM Tris (pH 8.0) for 20 h at 20,000 rpm at Triton X-100-2 mg of lysozyme per ml and incubated for 1 h 20°C in a Beckman SW27 rotor. Fractions containing 30- to in an atmosphere of CHCl3. The filters were then washed 45-kb fragments of DNA were identified, pooled, and dia- with two consecutive incubations of 30 min each on 3MM lyzed against TE buffer. HindIll-digested rickettsial DNA paper soaked with 20 mM Tris (pH 7.5) and 500 mM NaCl was not fractionated before ligation. (TBS). The filters were transferred to a fresh sheet of 3MM Endonuclease-digested DNAs were combined and ligated paper saturated with TBS containing 10 mM MgCl2 and 4 mg at a concentration of 410 ,ug of DNA per ml in 66 mM Tris of DNase I per ml. After a 45-min incubation, the filters were (pH 7.5)-6.6 mM MgCl2-10 mM dithiothreitol-0.1 mM washed once with TBS as before and overlaid with nitrocel- ATP-100 ,ug of DNase-free bovine serum albumin per ml-2 lulose. Filter sandwiches were incubated overnight in plastic U of T4 polynucleotide ligase for 68 h at 15°C. The ligated bags to maintain moisture. The nitrocellulose sheets were DNA was packaged into bacteriophage capsids by using incubated for 4 h at 37°C in TBS containing 2.5% gelatin to packaging extracts prepared from E. coli BHB2688 and BHB minimize nonspecific binding of probes to the 2690 (7). Packaging was accomplished as follows. Ligated nitrocellulose. The blots were incubated with primary anti- DNA (3 ,ug) was added to 20 ,ul of a packaging buffer serum at a 1:200 dilution in TBS plus 1% gelatin for 4 h at containing 40 mM Tris (pH 8.0), 10 mM MgSO4, 10 mM 37°C and then overnight at 4°C. The filters were washed spermidine hydrochloride, 2 mM putrescine hydrochloride, extensively with TBS containing 0.1% Triton X-100 and 0.1% P-mercaptoethanol, 7% dimethyl sulfoxide, and 4 mM reacted with rabbit anti-goat immunoglobulin G conjugated ATP. This was then mixed with 20 ,ul of frozen packaging with horseradish peroxidase (Cappel Laboratories, Coch- extract and incubated for 1 h at 37°C. An additional 25 ,u1 of ranville, Pa.) diluted 1:2,000 in TBS plus 1% gelatin for 2 h at packaging extract containing 10 mM MgCI2 and 0.3 p.g of 37°C. The blots were washed extensively and developed in DNase I (Boehringer Mannheim Biochemicals, Indianapolis, HRP color development reagent. Ind.) was mixed with the packaging reaction and incubated A simplified protocol was developed for screening for for 30 min at 37°C. One milliliter of 10 mM Tris (pH 7.5)-10 antigen expression in cosmid clones derived from X2687T, mM MgSO4 and 4 drops of CHC13 were added, and the which is E. coli HB101 containing a thermo-inducible lambda mixture was vortexed gently and centrifuged for 30 s in an lysogen. Cosmid clones were cultured in individual wells of Eppendorf microfuge. The resulting supernatant fluid was 96-well microtiter dishes containing 200 p.l of LB broth used to transduce E. coli HB101 or X2687T. plus 50 p.g of ampicillin per well overnight at 32°C. Elevation Recipient E. coli strains were grown to late-logarithmic of the temperature to 42°C effected cell lysis. Ten microliters phase at 37 (HB101) or 32°C (X2687T) in 20 ml of LB broth of 0.1 M MgCI2 containing 1 mg of DNase I (Sigma) per ml containing 0.4% maltose, centrifuged at 5,000 rpm for 5 min was added per well and incubated for 30 min at room at 4°C in a Beckman JA-20 rotor, and suspended in 4 ml of 10 temperature. The lysates were transferred to individual mM MgSO4. The cell suspension (1 ml) was combined with wells of a 96-well vacuum filtration manifold and collected 0.5 ml of the packaged DNA preparation and incubated for on nitrocellulose. Reactions with blocking buffer and pri- 30 min at 37 or 32°C (HB101 or X2687T, respectively). LB mary and secondary antibodies were as described above, broth (19 ml) was added, and the incubation was continued except that Triton X-100 was replaced with 0.001% Tween for an additional hour. The suspension was centrifuged at 20. 5,000 rpm for 5 min as before, suspended in 2 ml of PBS, and Minicell analysis of cosmid gene products. CaCI2-treated E. plated on LB agar containing 50 p.g of ampicillin per ml. coli minicell X1849 was transformed with purified chimeric Ampicillin-resistant colonies developed after overnight incu- cosmid DNA or vector DNA according to Davis et al. (7). bation at 37 or 32°C, as appropriate. These colonies were Transformants were selected on LB agar containing 50 p.g of picked to individual wells of 96-well microtiter dishes con- ampicillin per ml and 100 p.g of diaminopimelic acid per ml. taining 200 p.l of LB broth plus 50 p.g of ampicillin per well Minicells were purified according to Robeson et al. (22). and incubated overnight at the appropriate temperature. Briefly, minicell transformants were cultured in 50 ml of After the addition of sterile glycerol, the clones were stored glucose minimal medium (32) supplemented with 0.5 p.g of at -700C. biotin per ml, 100 p.g of diaminopimelic acid per ml, 40 p.g of Preparation of antiserum. One goat was injected intrader- adenine per ml, 0.5% Casamino Acids, and 100 p.g of mally with an emulsion of ca. 3.2 x 1010 R. prowazekii ampicillin per ml to an optical density at 600 nm of 0.50 to Madrid E cells in Freund complete adjuvant. This was 0.60. The cell suspension was centrifuged at 500 x g for 5 followed by six weekly subcutaneous injections of a similar min to remove a substantial amount of the normal cells. The dose in Freund incomplete adjuvant. Two months later, the resulting supernatant fluid was centrifuged for 20 min at goat received a booster dose of rickettsiae broken in a 5,700 x g. The pellet was suspended in PBS containing 0.1% French pressure cell (three times at 20,000 lb/in2) and gelatin. Minicells were purified by two successive centrifu- emulsified in Freund incomplete adjuvant. A similar booster gations on 5, 12.5, and 20% sucrose step gradients (5,000 immunization was administered 6 months later. Immune rpm in an SW27 rotor for 15 min). The minicells were finally serum was obtained 1 week after the final booster dose. The suspended in 0.6 ml of glucose minimal medium supple- blood was allowed to clot overnight, and serum was col- mented with 0.5 p.g of biotin per ml, 40 p.g of adenine per ml, lected after centrifugation at 2,500 x g for 10 min. After and S p.g each of tryptophan, methionine, cysteine, aspara- conjugation with fluorescein isothiocyanate, this serum had gine, and glutamine per ml and incubated for 20 min at 37°C. a direct fluorescent antibody titer against intact rickettisae of 14C-amino acids (2 p.Ci) (ICN, Irvine, Calif.; 1.78 mCi/mg) 1:1,000. were added, and the incubation was continued for an addi- In situ screening for R. prowazekii antigen expression. tional 20 min at 37°C. Supplementation with the unlabeled Cosmid clones derived from HB101 were replica plated from amino acids listed was necessary because these particular stocks onto sterile HATF filters situated on 150-mm-diame- amino acids were not present in the 14C-amino acid mixture. ter LB agar plates containing 50 p.g of ampicillin per ml and The labeled minicells were washed once with PBS contain- incubated overnight at 37°C. The filters were placed on ing 0.1% gelatin and analyzed by electrophoresis on a 160 KRAUSE, WINKLER, AND WOOD INFECT. IMMUN.

3%/10% or 3%/12.5% discontinuous SDS-polyacrylamide gel (13) and autoradiography. Western blot analysis of cosmid clones. Clones or control cultures were incubated overnight in LB broth containing 50 jig of ampicillin per ml. Samples of the cell suspensions (usually 0.5 ml) were centrifuged in an Eppendorf microfuge, and the resulting pellets were washed once with PBS. Cell pellets were resuspended in dissolving buffer (62.5 mM Tris [pH 6.8], 10% glycerol, 2% 3-mercaptoethanol) and solubi- lized with the addition of SDS to 2%. Bromphenol blue was added to a final concentration of 0.005%, and the samples were boiled for 4 min. SDS-polyacrylamide gel electropho- resis (SDS-PAGE) was performed as described above. Pro- teins were transferred electrophoretically to nitrocellulose according to Towbin et al. (26) at 100 mA for 15 h or 240 mA for 5 h. The nitrocellulose was then processed by enzyme- linked immunosorbent assay (ELISA) as described above for the colony screening, except that no detergent was included in the TBS washes. RESULTS Cosmid cloning of R. prowazekii DNA into E. coli K-12. A total of 1,288 ampicillin-resistant colonies were obtained when E. coli HB101 was infected with phage capsids con- taining recombinant molecules of Sau3A-digested rickettsial DNA and vector pHC79 DNA. When HindIll-digested DNA was cosmid cloned, 798 ampicillin-resistant colonies devel- oped. With the latter clone bank, a direct comparison on an agarose gel was possible between complete HindIlI diges- tions of pooled chimeric plasmid DNA and uncloned R. FIG. 1. Agarose gel profiles resulting from restriction endonucle- prowazekii DNA. In this way, the extent to which the clone ase Hindlil-digestion of pooled recombinant cosmid DNA from the bank was representative of the rickettsial genome in terms of HindIII clone bank (lane B) or Madrid E strain R. prowazekii genomic DNA (lane C). A total of 798 E. coli cosmid clones resulting the overall banding pattern could be assessed. The patterns from the insertion of HindIll-digested rickettsial DNA into the thus obtained are shown in Fig. 1. The two preparations corresponding site in the cosmid vector pHC79 were cultured exhibited very similar but not identical patterns. The major separately and pooled. Plasmid DNA was extracted according to difference, as expected, was the presence of the vector Hansen and Olsen (10) and purified by CsCl-ethidium bromide pHC79 DNA as the predominant DNA band in the pooled equilibrium centrifugation. Purified chimeric cosmid DNA and pu- chimeric cosmid preparation. In addition, there were bands rified rickettsial genomic DNA were digested to completion with present in the restricted genomic DNA which were not HindIII, electrophoresed on a 0.9% agarose gel, and visualized by visible in the digested chimeric cosmid profile. Hybridiza- ethidium bromide staining. Lane A contains bacteriophage lambda tion of R. prowazekii DNA labeled with 32P by nick transla- DNA digested with HindlIl for molecular size determinations. From tion (21) under conditions requiring >90% homology con- top to bottom the bands are 23.3, 9.3, 6.5, 4.3, 2.2, 1.9, and 0.5 kb. firmed that the DNA shown in Fig. 1, lane B, with the exception of the vector band, is R. prowazekii DNA (data Approximately 20% of all clones screened were positive by not shown). ELISA for antigen expression, as indicated by the develop- Evaluation of the anti-R. prowazekii antibody probe. Com- ment of the dark spots. Hybridization of R. prowazekii DNA parisons of reactivity in preimmune and immune goat anti-R. labeled with 32P by nick translation (21) with individual prowazekii sera to lysates of R. prowazekii or E. coli HB101 positive clones revealed that ca. 70% of the sample popula- harboring the vector pHC79 are shown in Fig. 2. The highest tion tested were carrying detectable amounts of rickettsial numbers of bacteria used were ca. 7.5 x 107 R. prowazekii DNA (data not shown). The limits of detection by this cells and 1.4 x 107 E. coli cells. Serial 10-fold dilutions of procedure were not assessed, however. Two putative anti- each suspension were also tested. A very faint purple color gen-positive clones from such screenings were selected at developed with the highest concentrations of E. coli with random for more detailed analysis. both normal and immune sera (Fig. 2, rows A and C), Physical characterization of DNA from selected chimeric indicating that some cross-reactivity with E. coli determi- cosmid clones. Cosmid clones F8F and FlOE were selected nants can be detected in the antibody probe. Normal goat for further analysis. These resulted from the cloning of serum failed to react with the rickettsial lysates, even at the Sau3A-digested rickettsial DNA in HB101. Comparison of highest concentration of antigen tested (Fig. 2, row B). In undigested chimeric cosmids with vector pHC79 and with contrast, the immune serum reacted positively with the covalently closed plasmid DNA molecular size standards is rickettsial lysates in a manner proportional to antigen con- shown in Fig. 4. Plasmids pMW218 and pMW219 (F8F and centration (Fig. 2, row D). FlOE, respectively) are ca. 40.2 and 38.1 kb in size, respec- Screening cosmid clone banks for antigen expression. Re- tively. Additional bands of lower intensity and smaller size sults from a representative screening of cosmid clones are were observed in the purified pMW218 and pMW219 prep- shown in Fig. 3. A positive control (R. prowazekii French arations (lanes A and C, respectively). Although the origin of pressed twice at 20,000 lb/in2) and a negative control (E. coli these bands is not clear, they could represent deletion X2687T carrying pHC79 and induced to lysis) are indicated. derivatives of the larger chimeric plasmids. Chimeric cosmid VOL. 47, 1985 COSMID CLONING OF RICKETTSIAL ANTIGENS 161

To test the latter possibility, plasmid DNA was extracted from the minicell strains by the rapid minilysis procedure of Holmes and Quigley (12) and analyzed by restriction en- donuclease digestion and agarose gel electrophoresis. Un- digested plasmid DNA migrated considerably further in an agarose gel than had been observed previously (data not shown). Profiles of restriction fragments of the plasmids after digestion with HincII are shown in Fig. 6. These B 40 restriction fragments were shared with the larger parent plasmids; the parent plasmids, however, possessed addi- tional restriction fragments not found here (data not shown). Thus, it appears that the original chimeric plasmids pMW218 and pMW219 underwent a deletion event resulting in the smaller plasmids isolated subsequently (Fig. 6). The result- ing plasmids were 13.6 and 12.6 kb and will be designated here as pMW218-d and pMW219-d. Western blot analysis of recombinant clones. The minicell strains containing recombinant plasmids pMW218-d or pMW219-d, or vector pHC79, were analyzed by SDS- PAGE, electrophoretic transfer to nitrocellulose, and ELISA D 4X (Fig. 7). There was some cross-reactivity in the anti-R. prowazekii probe with the E. coli minicell strain harboring FIG. 2. ELISA of normal and hyperimmune goat serum, using only the vector pHC79, particularly with a 60,000-dalton serially diluted lysates of R. prowazekii and E. coli. Tenfold protein (Fig. 7, lane A). Both clones exhibited the same bacterial dilutions were spotted onto agar plates and allowed to air cross-reactivity as that of the negative control. The clone dry. Cells were lysed by inverting the plates for 20 min in an containing pMW218-d, however, possessed a new 51,000- atmosphere saturated with chloroform vapors. Lysates were blotted dalton protein which was also recognized by the antiserum to nitrocellulose for 2 h and processed by ELISA as described in the (Fig. 7,-lane C). This protein is the same size as the new, text. The primary sera were used at a 1:75 dilution. The secondary recombinant plasmid clone-associated protein seen in Fig. 5, antibody-enzyme cornugate was used at a 1:2,500 dilution. The highest numbers of bacteria used were ca. 7.5 x 107 rickettsial cells lane D. and 1.4 x 107 E. coli cells. (A) and (B) are E. coli and R. prowazekii, Identification of cloned gene product as a rickettsial antigen. respectively, screened with normal goat serum; (C) and (D) are E. The rickettsial antigen being expressed in E. coli was iden- coli and R. prowazekii, respectively, screened with goat anti-R. tified by a two-step ELISA. Total protein from E. coli prowazekii serum. Arrows indicate the direction of increasing lysate containing pHC79 or pMW218-d was subjected to SDS- concentration. PAGE and blotted to nitrocellulose (26). The nitrocellulose strips carrying control or clone protein were processed separately from this point. Each was treated with blotting instability has been reported recently by others (31) and is buffer, followed by goat anti-R. prowazekii antiserum as suggested by data to follow in this study. before. After unbound antibodies were washed away, the Characterization of gene products from chimeric cosmid bound antibodies were eluted by treating the nitrocellulose DNA. E. coli minicells provide a convenient tool for the with elution buffer (0.2 M glycine, 0.2 M NaCl, 0.1% bovine examination of plasmid-coded proteins. Analysis of newly synthesized proteins in minicells carrying recombinant plasmids by SDS-PAGE and autoradiography resulted in the several new protein species originating from identification of X il: k n:. rickettsial genes (Fig. 5). The proteins seen in lane A + cv.:1,,, 91," represent newly synthesized proteins in minicells containing no plasmid DNA, perhaps transcribed from long-lived E. it ¢ f t W coli mRNA. In lane B, protein synthesis in E. coli minicells containing only the vector pHC79 was analyzed. Additional proteins represent the gene products of the ampicillin and tetracycline resistance genes. Lanes C and D show the f: proteins synthesized in minicells containing cosmids pMW218 and pMW219, respectively. A 55,000-dalton pro- tein not observed in control minicells was synthesized by r minicells containing plasmid pMW219; similarly, 51,000- and ( 04 H 15,000-dalton proteins were synthesized in minicells contain- ing plasmid pMW218 (arrows). These new protein species Ir t1 were apparently encoded by rickettsial genes in the chimeric plasmids. f f : f _-A -. We were surprised that more new proteins were not this the size of the observed by procedure, considering FIG. 3. ELISA of lysates of E. coli X2687T cosmid clones. chimeric cosmids and their potential coding capacity. Al- French-pressed R. prowazekii and X2687T harboring vector alone, though this could easily be explained by poor rickettsial as positive and negative controls, respectively, are indicated. The promoter recognition by E. coli polymerase, a portion of the development of a dark color is indicative of the expression of a parent chimeric cosmids might have been lost by deletion. rickettsial antigen(s) in the clone. 162 KRAUSE, WINKLER, AND WOOD INFECT. IMMUN. serum albumin [pH 2.8]) overnight at 20°C. After dialyzing A B C D the eluates, first with 50 mM citric acid-50 mM sodium phosphate (pH 5.5) and then with two changes of PBS, the eluted immunoglobulin preparations were used as the pri- mary antibodies in an ELISA with R. prowazekii total protein separated by SDS-PAGE and blotted to nitrocellu- lose (Fig. 8). Antibodies eluted from the E. coli control blot 97.4- reacted with several rickettsial proteins (Fig. 8, lane C). These protein bands were also observed with antibodies eluted from the clone blot. With the latter, however, an additional, albeit faint, new protein species of ca. 52,000 -.( m 4. daltons was observed (arrow, Fig. 8, lane D). This suggests that the 51,000-dalton protein that is produced in E. coli 46- minicells containing pMW218-d and is reactive with anti-R. prowazekii antiserum corresponds to a 52,000-dalton protein a~~4~ om .qm- in R. prowazekii. 30- !im DISCUSSION In recombinant cosmid DNA libraries, a relatively small number of clones is sufficient to represent an entire genome. This greatly facilitates the cloning of genes for nonselectable traits, in which positive clones can only be identified by screening procedures. We describe here the formation of cosmid clone banks of the R. prowazekii genome in E. coli. The extent to which one such bank represented the rickett- sial genome was evaluated. Comparison of restricted clone FIG. 5. Autoradiograph of 14C-labeled proteins produced in E. bank and R. prowazekii genomic DNA on agarose gels coli minicells and separated by SDS-PAGE. Minicells were pre- revealed that the two possess remarkably similar banding pared and pulse labeled as described in the text. Lanes: A, proteins portion of the produced in X1849 minicells carrying no plasmids; B, proteins patterns. This suggests that a substantial produced in minicells carrying the vector pHC79; C and D, proteins rickettsial genome is represented in the clone bank. Based produced in minicells harboring the chimeric cosmids pMW219-d upon the formula of Clarke and Carbon (3), the size of the and pMW218-d, respectively. The positions and molecular masses rickettsial genome (1 x 109 daltons), and a theoretical (kilodaltons) of 14C-labeled protein standards are indicated to the left. They were phosphorylase b, ovalbumin, and carbonic anhydr- ase, having molecular masses of 97.4, 46, and 30 kilodaltons, respectively, according to the manufacturer (New England Nuclear Corp., Boston, Mass.). New proteins synthesized by minicells containing chimeric plasmids are indicated by arrows. A 3% stacking- 10% separating polyacrylamide gel was employed here.

average insert size of 2.5 x 107 daltons with cosmid cloning, a clone bank of ca. 200 different cosmid clones should be sufficient to yield a 99% probability that any given gene sequence will be represented in that bank. That there are differences in the banding patterns of HindIII-digested clone bank and genomic DNA is probably a reflection of several factors. It has been our observation as well as that of others (31; D. Kopecko, personal communication) that cosmid clones are relatively unstable. This might account for the observation that only 70% of the ampicillin-resistant cosmid clones tested possessed detectable rickettsial DNA by DNA- DNA hybridization. It appears that deletions in the large chimeric plasmids can eventually result in the loss of much or all of the rickettsial DNA inserts. This seems to occur despite the fact that the recipient strains used here are recA mutants and are therefore deficient in their recombination capabilities. The instability of cosmid clones was also re- flected by the apparent deletions in the chimeric DNA from two of the clones, F8F and F1OE. It is possible that by reducing the ampicillin concentration in the growth medium and hence reducing the selective pressure for maximum FIG. 4. Agarose gel electrophoretic analysis of the size of chi- beta-lactamase production from the vector, the occurrence meric plasmids pMW219 and pMW218. The mobility of the chimeric in the chimeric could be reduced (D. plasmids was compared with that of pHC79 alone and with cova- of deletions plasmid lently closed circular DNA standards (14). Lanes: A, pMW219; B, Kopecko, personal communication). marker DNA, from top to bottom (x 106): 35.8, 4.8, 3.7, 3.4, 2.6, 2.0, A second factor which could produce discrepancies in 1.8, and 1.4 daltons; C, pMW218; and D, pHC79. agarose gel patterns of restricted clone bank and genomic VOL. 47, 1985 COSMID CLONING OF RICKETTSIAL ANTIGENS 163

data suggest that these smaller plasmids represent deletion mutants of the parent plasmids. It is not clear when the deletion that generated pMW218-d and pMW219-d occurred. Because smaller bacterial plasmids transform more effi- ciently than their larger counterparts, transformation into the E. coli minicell strain could have resulted in an enrich- ment of smaller plasmids. Alternatively, the deletions could have occurred in the minicell strain after transformation with the original parent plasmid. That rickettsial proteins were synthesized from genes on the chimeric plasmids was established by minicell analysis of the gene products. With clone FlOE this protein was subse- quently demonstrated to be reactive with anti-R. prowazekii serum and apparently corresponds with a 52,000-dalton protein antigen in R. prowazekii. Although a 55,000-dalton protein was produced in minicells transformed with plasmid pMW219-d, this protein did not appear to be recognized by the hyperimmune serum by immunoblotting. This could be explained in at least two ways. It is possible that the 55,000-dalton protein was not responsible for the original positive reaction obtained during colony screening, and the gene responsible for the positive reaction was lost as a result of the deletion which yielded the smaller plasmid pMW219- d. Alternatively, SDS treatment rendered the 55,000-dalton protein incapable of binding its respective antibodies. Ex- amination of this clone by radioimmunoprecipitation, a milder procedure than protein blotting after SDS-PAGE, may clarify this point. FIG. 6. Agarose gel electrophoresis profiles of plasmid DNA isolated from E. coli minicell strains and digested with restriction endonuclease HincII. Lanes: A, vector pHC79; B, chimeric plasmid pMW218-d; C, chimeric plasmid pMW219-d; and D, bacteriophage A C lambda DNA digested with Hindlll for linear DNA size determina- tions. The DNA standards are: 23.3, 9.3, 6.5, 4.3, 2.2, and 1.9 kb. DNA is the effect, either positive or negative, that the presence of different large chimeric plasmids might have on the growth of the E. coli host. Thus, variations in plasmid copy number in the clone bank due to a specific rickettsial 92.5- insert or failure of a particular clone to grow due to expres- sion of a rickettsial gene which is lethal in E. coli could result in differences in the clone bank profile when compared with 68- the genomic DNA pattern. Screening of the chimeric cosmid libraries for antigen expression was facilitated by the relatively low level of _+ reactivity in the hyperimmune serum to E. coli constituents. Antigen-positive clones consistently yielded a more intense purple color by ELISA than did their negative counterparts or E. coli negative controls. Our best success was obtained with the E. coli X2687T recipient grown in broth culture, induced to lysis, and treated with DNaseI before the collec- tion of lysates on nitrocellulose. This can probably be attributed to more efficient lysis, better binding of lysate 2S.7- proteins to nitrocellulose, or less interference by other cellular constituents with antigen-antibody binding. False- negative reactions were occasionally observed and were apparently due to the trapping of an air bubble just above the nitrocellulose at some stage during the ELISA. Likewise, FIG. 7. Western blot and ELISA analysis of total protein from occasional clogging of the vacuum outlet in the filter support recombinant plasmid- or vector-containing E. coli minicell strains. generally resulted in poor washing in a particular sample Proteins were separated on a 3%-10% polyacrylamide gel and well and hence yielded a false-positive reaction. Neverthe- blotted to nitrocellulose as described in the text. Goat anti-R. less, ca. 20% of the clones screened were consistently prowazekii serum at a 1:100 dilution was employed in the ELISA. Lanes A, B, and C are E. coli minicell strains containing pHC79, positive for antigen expression. To better characterize anti- pMW219-d, and pMW218-d, respectively. The locations of molecu- gen expression in our system, two putative positive clones lar mass standards (Bethesda Research Laboratories) are indicated were selected at random for a more detailed analysis. to the left; they included: phosphorylase b, 92,500 daltons; bovine Both clones examined contained large plasmids of ca. 40 serum albumin, 68,000 daltons; and cx-chymotrypsinogen, 25,700 kb and perhaps low levels of smaller plasmid species. The daltons. 164 KRAUSE, WINKLER, AND WOOD INFECT. IMMUN.

2. Boyer, H. W., and D. Roulland-Dussoix. 1969. A complementa- A B C D tion analysis of the restriction and modification of DNA in Escherichia coli. J. Mol. Biol. 41:459-472. f..... 3. Clarke, L., and J. Carbon. 1976. A colony bank containing 92.5 - __~~~~~~~~~~~~~~~~~~~~~1w synthetic ColEl hybrid plasmids representative of the entire E. 68 _ao coli genome. Cell 9:91-99. 4. Clewell, D. B., and D. R. Helinski. 1969. Supercoiled circular DNA-protein complex in Escherichia coli: purification and induced conversion to an open circular form. Proc. Natl. Acad. Sci. U.S.A. 62:1159-1166. 25.7 - 5. Collins, J. 1979. Escherichia coli plasmids packageable in vitro in bacteriophage particles. Methods Enzymol. 68:309-326. 6. Dasch, G. A., and A. L. Bourgeois. 1981. Antigens of the typhus group of rickettsiae: importance of the species-specific surface

- protein antigens in eliciting immunity, p. 61-70. In W. Burgdor- 18.4 fer and R. Anacker (ed.), Rickettsiae and rickettsial diseases. Academic Press, Inc., New York. 7. Davis, R. W., D. Botstein, and J. R. Roth. 1980. Advanced /~~~~~~ bacterial genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 8. Gambrill, M. R., and C. L. Wisseman, Jr. 1973. Mechanisms of immunity in typhus infections. I1. Multiplication of typhus rickettsiae in human macrophage cell cultures in the nonimmune FIG. 8. Western blot and ELISA analysis of R. prouazekii system: influence of of rickettsial strains and of protein. Rickettsial proteins were separated on a 3%-12.5% poly- chloramphenicol. Infect. Immun. 8:519-527. acrylamide gel and blotted to nitrocellulose as described in the text. 9. Hansen, J. B., Y. Abiko, and R. Curtiss III. 1981. Characteriza- Rickettsial protein in lanes A and B were treated with normal goat tion of the Streptococcus mutans plasmid pVA318 cloned into serum or goat anti-R. prowazekii serum, respectively, at a 1:100 Escherichia coli. Infect. Immun. 31:1034-1043. dilution. Lanes C and D were treated with antibodies which had 10. Hansen, J. B., and R. H. Olsen. 1978. Isolation of large bacterial been eluted from Western blots of E. coli containing pHC79 (C) or plasmids and characterization of the P2 incompatibility group pMW218-d (D) previously treated with anti-R. prowazekii serum. plasmids pMGI and pMG5. J. Bacteriol. 135:227-238. Thus, antibodies reactive with the cloned antigen were used to 11. Hohn, B., and J. Collins. 1980. A small cosmid for efficient identify that protein's counterpart in R. prowazekii. Note the cloning of large DNA fragments. Gene 11:291-298. 52,000-dalton protein present in lane D (arrow) but not in lane C. 12. Holmes, D. S., and M. Quigley. 1981. A rapid boiling method for Protein standards are indicated to the left as in Fig. 7. the preparation of bacterial plasmids. Anal. Biochem. 114:193-197. This study represents a preliminary analysis of the feasi- 13. Laemmli, U. K. 1970. Cleavage of structural proteins during the bility of cosmid cloning of R. prowazekii antigens in E. coli. assembly of the head of bacteriophage T4. Nature (London) The significance of the cloned antigen is under investigation. 227:680-685. In addition, subsequent analysis of other cosmid clones 14. Macrina, F. L., D. J. Kopecko, K. R. Jones, D. J. Ayers, and which reacted positively in the colony screening procedure D. J. McCowen. 1978. A multiple plasmid-containing Escheri- has resulted in the identification of other chia coli strain: a convenient source of size reference plasmid rickettsial protein molecules. Plasmid 1:417-420. antigens being expressed in E. coli. This preliminary screen- 15. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular ing procedure will permit the elimination of antigen-negative cloning. Cold Spring Harbor Laboratory, Cold Spring Harbor, clones and thus the judicious use of human typhus sera, N.Y. which are available only in limited quantities, in subsequent 16. Meyers, W. F., and C. L. Wisseman, Jr. 1980. Genetic related- screenings to identify clones expressing antigens potentially ness among the typhus group of rickettsiae. Int. J. Syst. important in the host response to rickettsial infections. This Bacteriol. 30:143-150. work should prove significant in permitting the purification 17. Murphy, J. R., C. L. Wisseman, Jr., and P. Fiset. 1979. of sufficient quantities of rickettsial antigens for the evalua- Mechanisms of immunity in typhus infection: adoptive transfer of to Rickettsia mooseri. Infect. Immun. 24:387-393. tion of their role in pathogenesis and host response to immunity 18. Nacy, C. A., E. J. Leonard, and M. S. Meltzer. 1981. Macro- infection. phages in resistance to rickettsial infections: characterization of lymphokines that induce rickettsiacidal activity in macro- phages. J. Immunol. 126:204-207. ACKNOWLEDGMENTS 19. Palmes, B. A., F. M. Hetrick, and T. R. Jerrells. 1984. Produc- tion of gamma interferon in mice immune to Rickettsia tsu- We gratefully acknowledge Roy Curtiss lII, Dennis Kopecko, and tsugamushi. Infect. Immun. 43:59-65. Rodney Welch for providing us with the indicated E. coli strains. We 20. Ramm, L. E., and H. H. Winkler. 1976. Identification of thank Cary Engleberg for helpful suggestions, Robin Daugherty, Bill cholesterol in the receptor site for rickettsiae on sheep erythro- Atkinson, and Chee T. Chu for technical assistance, and Mary cyte membranes. Infect. Immun. 13:120-126. Lazzari for typing this manuscript. 21. Rigby, P. W. J., J. Dieckmann, C. Rhodes, and P. Berg. 1977. This work was supported by Public Health Service research Labeling deoxyribonucleic acid to high specific activity in vitro grants AI-15035 and AI-20384 from the National Institute of Allergy by nick translation with DNA polymerase I. J. Mol. Biol. and Infectious Diseases. 113:237-251. 22. Robeson, J. P., R. G. Barletta, and R. Curtiss III. 1983. Expression of a Streptococcus mutans glycosyltransferase gene LITERATURE CITED in Escherichia coli. J. Bacteriol. 153:211-221. 1. Beaman, L., and C. L. Wisseman, Jr. 1976. Mechanisms of 23. Smith, D. K., and H. H. Winkler. 1977. Characterization of a immunity in typhus infections. VI. Differential opsonizing and lysine-specific active transport system in Rickettsia prowazekii. neutralizing action of human antibodies in cultures of human J. Bacteriol. 129:1349-1355. macrophages. Infect. Immun. 14:1071-1076. 24. Stamm, L. V., J. D. Folds, and P. J. Bassford, Jr. 1982. VOL. 47, 1985 COSMID CLONING OF RICKETTSIAL ANTIGENS 165

Expression of Treponerna palliduim antigens in Escherichia coli 33. Walker, T. S., and H. H. Winkler. 1978. Penetration of cultured K-12. Infect. Immun. 31:1238-1241. mouse fibroblasts (L cells) by Rickettsia prowvazekii. Infect. 25. Stork, E., and C. L. Wisseman, Jr. 1976. Growth of Rickettsia Immun. 22:200-208. prouazekii in enucleated cells. Infect. Immun. 13:1743-1748. 34. Wenman, W. M., and M. A. Lovett. 1982. Expression in E. coli 26. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic of Chlatnvdia trachomatis antigen recognized during human transfer of proteins from polyacrylamide gels to nitrocellulose infection. Nature (London) 296:68-70. sheets: procedure and some applications. Proc. Natl. Acad. Sci. 35. Winkler, H. H. 1976. Rickettsial permeability: an ADP-ATP U.S.A. 76:4350-4354. transport system. J. Biol. Chem. 251:389-396. 27. Turco, J., and H. H. Winkler. 1982. Differentiation between 36. Winkler, H. H., and R. M. Daugherty. 1984. Proline transport virulent and avirulent strains of Rickettsia prowazekii by and metabolism in Rickettsia prowazekii. J. Bacteriol. macrophage-like cell lines. Infect. Immun. 35:783-791. 158:460-463. 28. Turco, J., and H. H. Winkler. 1983. Inhibition of the growth of 37. Winkler, H. H., and E. T. Miller. 1980. Phospholipase A activity Rickettsia prowazekii in cultured fibroblasts by lymphokines. J. in the hemolysis of sheep and human erythrocytes by Rickettsia Exp. Med. 157:974-986. Infect. 29. Turco, J., and H. H. Winkler. 1983. Comparison of the proper- prowvazekii. Immun. 29:316-321. ties of antirickettsial activity and interferon in mouse lympho- 38. Winkler, H. H., and E. T. Miller. 1982. Phospholipase A and the kines. Infect. Immun. 42:27-32. interaction ofRickettsia prowtazekii and mouse fibroblasts (L-929 30. Turco, J., and H. H. Winkler. 1983. Cloned mouse interferon cells). Infect. Immun. 38:109-113. inhibits the growth of Rickettsia prowazekii in cultured mouse 39. Winkler, H. H., and L. E. Ramm. 1975. Adsorption of typhus fibroblasts. J. Exp. Med. 158:2159-2164. rickettsiae to ghosts of sheep erythrocytes. Infect. Immun. 30a.Turco, J., and H. H. Winkler. 1984. Effect of mouse lympho- 11:1244-1251. kines and cloned mouse interferon-y on the interaction of 40. Wisseman, C. L., Jr., and A. D. Waddell. 1975. In vitro studies Rickettsia prowiazekii with mouse macrophage-like RAW264.7 on rickettsia-host cell interactions: intracellular growth cycle of cells Infect. Immun. 45:303-308. virulent and attenuated Rickettsia prowiazekii in chicken embryo 31. Van Embden, J. D., H. J. van der Donk, R. V. van Eijk, H. G. cells in slide chamber cultures. Infect. Immun. 11:1391-1401. van der Heide, J. A. de Jong, M. F. van Olderen, A. D. 41. Wisseman, C. L., Jr., and A. D. Waddell. 1983. Interferonlike Osterhaus, and L. M. Schouls. 1983. Molecular cloning and factors from antigen- and mitogen-stimulated human leukocytes expression of Treponema pallidlum DNA in Escherichia coli with antirickettsial and cytolytic actions on Rickettsia K-12. Infect. Immun. 42:187-196. prowvazekii. J. Exp. Med. 157:1780-1793. 32. Vogel, H. J., and D. M. Bonner. 1956. Acetylornithase of 42. Wood, D. O., W. H. Atkinson, R. S. Sikorski, and H. H. Escherichia coli: partial purification and some properties. J. Winkler. 1983. Expression of the Rickettsia prow,azekii citrate Biol. Chem. 218:97-106. synthase gene in Escherichia coli. J. Bacteriol. 155:412-416.