INFECTION AND IMMUNITY, May 1984, p. 222-227 Vol. 44, No. 2 0019-9567/84/050222-06$02.00/0 Copyright © 1984, American Society for Microbiology Cloning and Expression of pneumophila in N. CARY ENGLEBERG,' DAVID J. DRUTZ,1'2'3 AND BARRY I. EISENSTEIN1l2* Departments of Medicine' and Microbiology,2 The University of Texas Health Science Center at San Antonio, and the Veterans Administration Hospital,3 San Antonio, Texas 78284 Received 3 October 1983/Accepted 23 January 1984

To isolate and characterize Legionella pneumophila antigens, we constructed a genomic library of L. pneumophila serogroup 1 (strain 130b). L. pneumophila DNA fragments (2.5 to 7.5 megadaltons) obtained by partial digestion with Sau 3A endonuclease and size fractionation on a sucrose density gradient were inserted into the dephosphorylated BamHI site of vector pBR322; CaCl2-treated Escherichia coli cells of strain HB101 were transformed with hybrid plasmids. To detect expression of antigens, 2,559 ampicillin- resistant transformants were transferred to nitrocellulose paper, lysed in situ, and screened by enzyme immunoassay (EIA) with E. co/i-absorbed rabbit anti-L. pneumophila sera. A total of 77 (3%) of the colonies were reactive by EIA; 31 (1.2%) were strongly reactive, and 6 were strongly reactive by EIA without colony lysis. Analysis of 29 stable, strongly reactive clones by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electroblotting showed antigenic bands in 18 clones by EIA with E. coli-absorbed antisera. Absorption of antisera with heat- and Formalin-killed L. pneumophila eliminated or dimninished the reactivity of the antigenic bands in representative clones. These studies confirm that several L. pneumophila antigens can be cloned and expressed in E. coli.

Since the recognition of Legionella pneumophila as the others have shown that genes from a variety of bacterial causative agent of Legionnaires disease in 1977, vigorous can be transcribed and translated by Escherichia research has produced an impressive body of information coli (14, 16-18, 20, 23), we chose a simple cloning strategy about the disease, principally in the areas of clinical medi- that did not include specific engineering for gene expression. cine and public health (13). During the same period, howev- We constructed a library of the L. pneumophila genome by er, insight into the immunology and pathogenesis of legionel- using vector pBR322. We then screened the library for losis has developed slowly, and today our understanding of expression of L. pneumophila antigens by using a modified these aspects of infection remains incomplete. In general, filter binding assay (8, 14) in which colonies bound to laboratory-based research has been complicated by difficul- nitrocellulose paper are tested for antigens by in situ enzyme ties in cultivating and handling the organisms and by the immunoassay. In this article, we report the isolation of problem of maintaining their on artificial media. several E. coli clones that express a variety of L. pneumo- In early studies by Wong and co-workers, complex L. phila antigens. pneumophila antigens, prepared by ion-exchange and liquid chromatography, were isolated, and fractions having seroty- MATERIALS AND METHODS pic specificity and cross-reactivity were identified and ana- Bacterial strains. L. pneumophila serogroup 1 (130b) was lyzed (24-26). The serotypic antigen fraction is composed of used for all cloning procedures and antigen preparations. For a pronase-sensitive lipid-protein-carbohydrate complex that both purposes' L. pneumophila was isolated from infected induces active immunity to infectious challenge in mice. guinea pigs and passed only once on buffered charcoal-yeast Cross-reacting antigens, which are almost entirely composed extract agar (7). E. coli K-12 strain HB101 (mk-, rk-, recA) of protein, were found to elicit dermal hypersensitivity. was the recipient for hybrid plasmid transformations. Other workers have studied L. pneumophila antigens by the Enzymes and chemicals. Restriction endonucleases and T4 techniques of crossed immunoelectrophoresis (3, 4, 10), DNA ligase were obtained from Bethesda Research Labora- counterimmunoelectrophoresis (19), and most recently by tories, Bethesda, Md. Calf intestinal alkaline phosphatase, polyacrylamide gel electrophoresis and protein blotting lysozyme (grade I), protein A-Sepharose, 5-aminosalicylic (M. S. Hindahl and B. H. Iglewski, Abstr. Annu. Meet. Am. acid, ampicillin, and tetracycline were purchased from Sig- Soc. Microbiol. 1983, C104, p. 329; W. Ehret, G. Anding, ma Chemical Co., St. Louis, Mo. Horseradish peroxidase- and G. Ruckdeschel, 2nd International Symposium on Le- conjugated goat anti-rabbit immunoglobulin was purchased gionella, Atlanta, Ga., M-8, 1983). Although these studies from Cappell Laboratories, Cochranville, Pa. The color- have confirmed the existence of discrete serotypic and forming reagent 4-chloro-1-naphthol and nitrocellulose paper crossreacting antigens, it will be necessary to purify these for electroblotting were purchased from Bio-Rad Labora- molecules to establish their immunogenicity and to fully tories, Richmond, Calif. Nitrocellulose disks for filter bind- characterize their role in the response to infection. ing assays (type HA) were purchased from Millipore Corp., To circumvent the problems inherent in working with L. Bedford, Mass.; Whatman 3MM chromotography paper was pneumophila and the difficulties of biochemical purification obtained from American Scientific Products, McGaw Park, of antigens of interest, we chose to isolate antigens at the Il. genetic level by using recombinant DNA technology. Since Construction of the clone bank. DNA was extracted from L. pneumophila cells harvested from buffered charcoal-yeast extract agar plates in phosphate-buffered saline (PBS), pH * Corresponding author. 7.2. To assure against contaminating , samples of 222 VOL. 44, 1984 CLONING OF L. PNEUMOPHILA ANTIGENS 223 this suspension were plated on buffered charcoal-yeast ex- by an enzyme-linked immunosorbent assay (ELISA) by tract and sheep blood agar and inoculated into brain heart using whole L. pneumophila cells fixed to 96-well microtiter infusion broth. There was no growth on blood agar or brain plates (Dynatech Laboratories, Inc., Alexandria, Va.). FK heart infusion broth. Pure growth of L. pneumophila was cells corrected to optical density at 550 nm of 1.5 with PBS detected on buffered charcoal-yeast extract agar and was (pH 7.2) were further diluted 1:25 with PBS. A 0.1-ml sample confirmed by direct with fluorescein- of dilute FK cells was added to each well, and the microtiter labeled, polyvalent rabbit antisera (Centers for Disease plates were allowed to dry overnight at 42°C. Plates were Control, Atlanta, Ga.). L. pneumophila DNA was extracted washed three times with PBS containing 0.5% Tween 20, and purified by the method of Nakamura et al. (15). Purified serial dilutions of test sera (0.1-ml samples) were added, and DNA was partially restricted with Sau 3A restriction endo- plates were incubated for 2 h at room temperature on a nuclease, and the digestion fraginents were applied to a 10- Minimix agitator (Fisher Scientific Co., Silver Spring, Md.). ml 5 to 40% sucrose gradient in 1 M NaCl-20 mM Tris- After three washes with PBS-0.5% Tween 20, 0.1 ml of hydrochloride-5 mM EDTA (pH 8.0) and centrifuged at peroxidase-conjugated goat anti-rabbit immunoglobulin 100,000 x g for 21 h. Gradient fractions (0.5 ml) were (1:1,500 dilution) was added to each well, and the plates analyzed by agarose gel electrophoresis, and fractions con- were again agitated for 2 h. After three final washes with taining restriction fragments of 2.5 to 7.5 megadaltons were PBS-0.5% Tween 20, 0.1 ml of a color-forming substrate pooled. solution (0.08% 5-aminosalicylic acid and 0.006% hydrogen Vector pBR322 was prepared for cloning by complete peroxide, pH 6.0) was added to each well. After agitation for digestion with BamHI followed by 5' dephosphorylation 30 min, absorbance at 450 nm was measured in a MR580 with alkaline phosphatase (12). The latter procedure resulted MicroElisa Auto Reader (Dynatech). An absorbance of 0.5 in a 2- to 3-log reduction in recircularization and ligation of or more was considered positive. the vector as compared with untreated linear pBR322. To assess the effectiveness of serum absorption, a similar Size-fractionated L. pneumophila Sau 3A restriction frag- ELISA test was used in which live E. coli whole cells were ments were ligated to dephosphorylated pBR322 with T4 fixed to the microtiter plate wells by the methods described DNA ligase and used to transform E. coli strain HB101 above. rendered competent by treatment with 0.5 M CaCI2 (5). FB-EIA. Apr transformants were spotted onto agar plates Transformants were selected on Louria-Bertani agar con- with sterile toothpicks (ca. 325 colonies per plate), grown taining ampicillin (40,ug/mI) (12). Forty ampicillin-resistant overnight, and then blotted onto dry nitrocellulose filter (Apr) transformants were screened for tetracycline sensitiv- disks; 1 to 2 RI of FK cells was also spotted onto each filter ity (Tcs) on Louria-Bertani medium containing tetracycline as a positive control. Colonies were lysed in situ by the (15,ug/ml). method of Meyer et al. (14). Briefly, filter disks were placed Preparation of L. pneumophila cells for immunization. L. sequentially, colony side up, onto 3MM paper in a series of pneumophila cells from six buffered charcoal-yeast extract four petri dishes saturated respectively with (i) 0.1 N NaOH, agar plates were harvested, pooled, and suspended in 6 ml of (ii) 1.5 M Tris-hydrochloride (pH 7.4), (iii) 300 mM NaCI-30 PBS (pH 7.3).F,ormalin-killed (FK) cells were prepared from mM sodium citrate, and (iv) 70% ethanol for 5 min each. 3 ml of this suspension by adding Formalin to a final Each filter was then dried in a vacuum at 60°C for 2 h. concentration of 2% and holding overnight at 4°C. Heat- Antigen-bearing colonies wete detected by an enzyme im- killed (HK) cells were prepared by heating the rerhaining 3 munoassay (EIA). Dry filters were placed in Tris-buffered ml of suspended cells to 100°C for 30 min. Both preparations saline (TBS; 50 mM Tris-hydrochloride, 150 mM NaCl, pH were checked for nonviability at 24 h. 7.5) containing 3% to block nonspecific protein- Preparation of antisera. New Zealand rabbits were inject- binding sites. After gentle rotation at room temperature for 2 ed subcutaneously with 2 ml of HK (1:5 dilution) or FK (1:10 h, the filters were transferred to a solution of absorbed, dilution) cells at 0, 2, 4, and 6 weeks. Four rabbits were pooled antisera (1:800 dilution in 1.5% gelatin-TBS) and immunized; two received HK cells and two received FK incubated overnight at room temperature. The following cells. Sera were collected at 7 weeks or later and were stored morning, the filters were rinsed with distilled water, washed with preimmune sera at -70°C. four times in TBS, and incubated in a solution of peroxidase- Immune sera (0.5-ml samples) were absorbed 4 times with conjugated goat anti-rabbit immunoglobulin (1:3,000) for 2 h. E. coli HB101 (pBR322). For each absorption, cells from 175 After a final rinse and series of four washes in TBS, the ml of a stationary-phase culture were washed, mixed with filters were immersed in a color development solution con- sera, and rotated at 4°C for 1 h; sera were recovered by sisting of 0.05% 4-chloro-1-naphthol and 0.015% hydrogen centrifugation at 5,000 x g for 5 min after each absorption. peroxide in a 5:1 solution of TBS-methanol. All colonies Sera used for the filter-binding assay were also absorbed showing color development were reanalyzed by a modified with cells harvested from Louria-Bertani agar plates and filter binding (FB)-EIA in which the alkaline lysis and with sonicated E. coli. For the sonicate absorption, cells neutralization steps (i.e., on saturated 3MM paper) were from 500 ml of stationary-phase growth suspended in 10 ml eliminated, and freshly blotted colonies were placed directly of PBS were disrupted with a Branson Sonifer probe sonica- into the vacuum oven. tor and then mixed with sera for 1 h at 4°C. The serum- Protein electroblotting of positive clones. Clones that were sonicate mixture was then cleared by centrifugation at strongly reactive by FB-EIA were analyzed by sodium 100,000 x g for 1 h in a Beckman SW 40 rotor. Fifty dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis microliters of E. coli-absorbed serum was also absorbed and protein blotting. Each clone was grown to the stationary twice with mixtures of 0.1 ml of FK and 0.1 ml ofk-K L. phase in 3.0 ml of Louria-Bertani broth containing ampicillin pneumophila cells for 8 h at4°C. All sera were heated to (50 jig/ml). Cultures were washed once with PBS, trans- 56°C for 30 min before use; the coagulum that formed in the ferred to Eppendorf tubes, suspended in 0.375 ml of sample serum-sonicate mixtures was pelleted and removed by cen- buffer (75 mM Tris-hydrochloride, 5% 2-mercaptoethanol, trifugation at 10,000 x g for 5 min. 2% SDS, 10% glycerol, 0.002% bromophenol blue, pH 6.8), ELISA. Sera were tested for anti-L. pneumophila activity and heated to 100°C for 2.5 min. SDS-polyacrylamide gel 224 ENGLEBERG, DRUTZ, AND EISENSTEIN INFECT. IMMUN. electrophoresis was performed in a vertical slab gel electro- TABLE 1. Characterization of cloned L. pneumophila antigens phoresis tank (Hoefer Scientific Instruments, San Francisco, Molecular Calif.) by the method of Laemmli (11). Samples of 0.04 ml E. coli transformants masses of were run at 25 mA (constant current) through the 5% (isolate no.) cloned polyacrylamide stacking gel (pH 6.8) and at 50 mA through antigens the 15% polyacrylamide separating gel (pH 8.3). Electro- 50 68K transfer of proteins to nitrocellulose was performed by the 63 61K procedure of Towbin et al. (22) in a Trans-Blot Cell (Bio- 66K Rad) at 30 mV for 12 h with a buffer containing 0.025 M Tris 12 61K base, 0.192 M and 20% methanol Rena- 66K glycine, (pH 8.3). 68K tured proteins on nitrocellulose were visualized with an EIA 11, 13, 40, 41, 44, 70 19K to 23K' similar to that described above. Nitrocellulose sheets were 21, 31, 32, 43, 60, 71 24K preincubated in a blocking solution of TBS with 0.05% 47, 81, 82 17K Tween 20 and then transferred overnight to a similar solution containing 1:750 rabbit antisera. After four washes with a RIP resolved individual bands of 19K, 20K, 23.5K, and 28K. 0.05% Tween 20-TBS, sheets were incubated in a 1:3,000 dilution of peroxidase-conjugated goat anti-rabbit immuno- globulin in 0.05% Tween 20-TBS for 1 h. After a rinse in was transformed by hybrid plasmids at an efficiency of 103 distilled water and four additional 0.5% Tween 20-TBS per ,ug of vector DNA. A sample of 40 Apr transformants washes, the sheets were exposed to the color development was transferred to media containing tetracycline; all 40 were reagent used in the FB-EIA. The molecular weight of each Tc'. individual band was estimated by comparing its coefficient of We screened 2,559 Apr transformants by FB-EIA; 77 migration to a logarithmic plot of the migrations of protein (3.0%) of these clones produced detectable blue color. Of standards run on the same gel and visualized with Coomassie these, 31 (1.2%) were considered to be strongly reactive, and blue stain. 46 (1.8%) were considered weakly reactive. All 77 reactive Selected clones were also analyzed by protein blotting clones were transferred to a single nitrocellulose disk and under less denaturing conditions by a method of sample analyzed by a modified FB-EIA, which excluded the chemi- preparation that did not require boiling. Cultures (1 ml) of cal lysis steps. In this assay, 11 clones produced detectable each clone were centrifuged, washed once with PBS, and color; 6 were strongly positive (clones 11, 13, 40, 41, 44, and suspended by vigorous vortexing in 0.04 ml of 25 mM Tris- 70), and 5 were weakly positive (clones 33, 47, 61, 65, and hydrochloride-10 mM EDTA (pH 8.0) with lysozyme (2 mg/ 73). Of these 11 clones, 10 were also strongly reactive in the ml). After 10 min of incubation at room temperature, 0.04 ml initial FB-EIA; clone 47 was weakly reactive in both assays. of 2 x sample buffer was added, and the samples were Protein blotting of strongly reactive clones. We analyzed 29 clectrophoresed as described above. of the strongly reactive clones by SDS-polyacrylamide gel RIP. Radioimmunoprecipitation (RIP) was performed with electrophoresis and protein blotting (two clones were unsta- absorbed rabbit antisera as previously described (6), except ble and could not be maintained in antibiotic-containing that protein A-Sepharose (40 ILI) was used instead of protein media). Of the 29 clonles, 18 had protein bands detected by A-containing Staphylococcus aureus. protein blotting with E. coli-absorbed rabbit antiserum, but Agarose gel electrophoresis of recombinant plasmids. Plas- not with preimmune serum (Table 1). Clones 11, 13, 40, 41, mids were extracted from selected clones by the modified 44, and 70, which gave strong signals by FB-EIA with and rapid alkaline lysis method of Birnboim and Doly (2, 9). without chemical lysis, all expressed the same antigenic Samples were mixed (1:5) with concentrated gel loading proteins. Trhese antigens appear as a confluence of bands in buffer (0.25% bromophenol blue, 30% glycerol) and added to the 19K to 23K range (Fig. 1, lane c). Because a faintly the wells of a 0.8% agarose gel, prepared in Tris-acetate reactive 19K to 20K band was seen in immunoblots of all E. buffer (40 mM Tris-acetate, 2 mM EDTA, pH 7.85). Electro- coli strains, we analyzed a representative clone (no. 11) by phoresis was performed in a horizontal gel electrophoresis additional techniques to confirm the separate identity of the system (Bethesda Research Laboratories), with Tris-acetate cloned antigens from the background band. Protein immun- buffer at 1 V/cm for 18 h. Gels were stained in ethidium oblots probed with unabsorbed antisera confirmed that 19K bromide (0.5 p.g/ml) for 45 min at room temperature, illumi- antigens are present in both the recipient E. coli strain and L. nated on a Chromato-Vue Transilluminator (Ultra-Violet pneumophila; E. coli clone no. 11 expresses both reactivities Products, Inc., San Gabriel, Calif.), and photographed with (Fig. 2). In addition, analysis of clone no. 11 by RIP with a Polaroid MP-3 Land camera. Restriction endonuclease absorbed antisera resolved four bands of 19K, 20K, 23.5K, digestions were performed under conditions specified by the and 28K that are distinct from E. coli background antigens manufacturer to produce complete digestion. (Fig. 3). Also in this analysis, an E. coli antigen between 19K and 20K was precipitated. None of the other cloned antigen RESULTS bands that we identified by immunoblotting coincided with background antigenic bands. ELISA of rabbit antisera. Rabbits imtnunized with either To confirm that the antisera recognized L. pneumophiia FK or HK cells developed titers of >1:1280 as antigens cloned in E. coli, we absorbed antisera with L. measured by the FK (whole cell) ELISA. Anti-L. pneumo- pneumophila FK and HK cells. This additional absorption phila activity was not detected in any preimmune serum at markedly diminished the reactivity against the antigenic titers of 1:10 or greater. Absorption of antisera with E. coli bands in clone 11 and completely removed reactivity against did not reduce the anti-L. pneumophila activity by more than one of the six clones that expressed a 24K antigen (Fig. 1). In one dilution, but did reduce the anti-E. coli titer from 1:320 a similar experiment, reactivity against the higher-molecu- to 1:40 in the E. coli whole cell ELISA. lar-weight antigens in clone 12 was also reduced or removed Characterization of E. coli transformants. E. coli HB101 (data not shown); quantitatively poor expression of the 17K VOL. 44, 1984 CLONING OF L. PNEUMOPHILA ANTIGENS 225 C b c d e f g M.W. a b

92.5 K -

68K -

5. -24K 43K - - 19K

-28 K 25.7K - FIG. 1. Protein blot analysis of recombinant plasmid-bearing E. coli; samples shown here were prepared from the recipient strain -23.5 K containing pBR322 (lane a), isolate 11 (lanes b through d), and isolate 21 (lanes e through g). Blots were assayed with preimmune -20 K sera (lanes b and e), immune sera absorbed with E. coli (lanes a, c, -004,0 0 or absorbed with killed L. pneumophila (lanes d _-1- 9 K and f), immune sera 18.4K - and g). !4t_,,. . .__:. antigen has not yet permitted any firm conclusions about the postabsorptive reactivity of this band. Repeat blotting with less denaturing conditions failed to uncover antigenic bands in any of the 12 clones that were FIG. 3. RIP and SDS-polyacrylamide gel electrophoresis of solu- positive by FB-EIA and negative by protein blotting. bilized E. coli (pBR322) (lane a) and E. coli clone 11 (lane b) with Agarose gel electrophoresis of recombinant plasmids. Plas- absorbed rabbit antisera. Tritiated molecular mass standards are mids from clones 11, 13, 40, 41, 44, and 70 were isolated and indicated in left margin. analyzed by gel electrophoresis. Electrophoresis of unre- stricted plasmid extracts confirmed that each clone con- with antisera absorbed with the recipient strain prevented tained only one recombinant plasmid (data not shown). After the selection of cloned, contaminating E. coli genes. The L. digestion with HindIII, only clones 41 and 44 had plasmids pneumophila origin of the antigen bands from representative with identical restriction fragments; clones 11 and 40 con- clones was confirmed by their failure to react with either tained similar but not identical plasmids (Fig. 4). These preimmune sera or immune sera absorbed with killed L. findings were confirmed by restriction analysis with Sau 3A, pneumophila. Because reactivity against a native E. coli BamHI, AvaI, and AccI. antigen that comigrated with a cloned antigen at approxi- DISCUSSION mately 19K could not be completely deleted from our antisera after extensive absorption with E. coli, it was We have demonstrated that L. pneumophila antigens can necessary to confirm the separate identity of the cloned be cloned and expressed in E. coli. The screening FB-EIA antigen. SDS-polyacrylamide gel electrophoresis and im- munoblotting with unabsorbed antisera demonstrated the presence of this antigen in solubilized L. pneumophila and a b c showed that the L. pneumophila and E. coli antigen bands in clone 11 were not identical. Moreover, RIP of clone 11 antigens with (whole cell) E. coli-absorbed antisera resolved four distinct bands of 19K, 20K, 23.5K, and 28K. In these two studies, the native E. coli antigens migrated slightly differently relative to the smallest of the series of cloned antigens. Because of differences in antigen preparation for the two techniques and because of the difficulty in accurately assigning molecular weights within such a narrow range, we are reluctant to attempt any matching of the bands detected by RIP and immunoblotting. Although not apparent in Fig. 1, activity against a 28K component has been observed on other immunoblots of clone 11 antigens conducted under different experimental conditions (data not shown). The expression of the same series of four antigens in five individual of solubilized E. coli distinct clones suggests that some or all of the FIG. 2. Section of a protein immunoblot a clone 11 (lane a), E. coli (pBR322) (lane b), and L. pneumophila bands may result from stepwise degradation of larger (lane c) with unabsorbed rabbit antisera. Note the presence of both antigen. A similar series of antigens (61K, 66K, 68K) is the E. coli and L. pneumophila 19K bands in E. coli clone 11. represented in clone 12 (although it is not yet clear why two 226 ENGLEBERG, DRUTZ, AND EISENSTEIN INFECT. IMMUN.

ance during the conduct of these studies. We also wish to acknowl- edge the valuable advice of Philip Bassford, William Haldenwang, a bcd efgh Itzhak Kahane, and Lola V. Stamm. B.I.E. is supported in part by a Public Health Service Research Career Development Award from the National Institutes of Health. LITERATURE CITED 1. Amano, K.-I., and J. C. Williams. 1983. Peptidoglycan of Legionella pneumophila: apparent resistance to lysozyme hy- drolysis correlates with a high degree of peptide cross-linking. J. Bacteriol. 153:520-526. 2. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 7:1513-1523. 3. Collins, M. T., S.-N. Cho, N. Hoiby, F. Epersen, L. Baek, and J. S. Reif. 1983. Crossed immunoelectrophoresis analysis of Legionella pneumophila serogroup 1 antigens. Infect. Immun. 39:1428-1440. 4. Collins, M. T., F. Epersen, N. Hoiby, S.-N. Cho, A. Triismoller, and J. S. Reif. 1983. Cross-reactions between Legionella pneu- mophila (serogroup 1) and twenty-eight other bacterial species, including other members of the family Legionellaceae. Infect. Immun. 39:1441-1456. 5. Dagert, M., and S. D. Erlich. 1979. Prolonged incubation in calcium chloride improves the competence of Escherichia coli cells. Gene 6:23-28. 6. Dodd, D. C., and B. I. Eisenstein. 1982. Antigenic quantitation of type 1 fimbriae on the surface of Escherichia coli cells by an enzyme-linked immunosorbent assay. Infect. Immun. 38:764- 773. 7. Edelstein, P. H. 1982. Comparative study of selective media for FIG. 4. Agarose (0.8%) gel electrophoresis of Hindlll digests of isolation of Legionella pneumophila from potable water. J. Clin. recombinant plasmids from six strongly reactive clones that express Microbiol. 16:697-699. the same series of L. pneumophila protein antigens. Complete 8. Henning, U., H. Schwarz, and R. Chen. 1979. Radioimmunologi- digests of pBR322 and plasmids from clones 11, 13, 40, 41, 44, and cal screening method for specific membrane proteins. Anal. 70 were run in lanes a through g, respectively. pBR322 multimers Biochem. 97:153-157. were included as molecular mass references (lane h). Note that only 9. Ish-Horowicz, D., and J. F. Burke. 1981. Rapid and efficient the plasmids of clones 41 and 44 have identical restriction patterns. cosmid vector cloning. Nucleic Acids Res. 9:2989-2999. 10. Joly, J. R., and G. E. Kenny. 1982. Antigenic analysis of Legionella pneumophila and Tatlockia micdadei (Legionella other clones, 50 and 63, selectively expressed certain of micdadei) by two-dimensional (crossed) immunoelectrophore- these bands, but not others). The observation that certain sis. Infect. Immun. 35:721-729. clones were positive by FB-EIA without colony lysis sug- 11. Laemmli, U. K. 1970. Cleavage of structural proteins during the that assembly of the head of bacteriophage T4. Nature (London) gests their encoded antigens may be surface expressed; 117:680-685. peptide cleavage associated with membrane translocation 12. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular may partially explain the multiple antigenic bands. cloning: a laboratory manual, p. 133-134, 368-369. Cold Spring The cloning of individual antigens will permit us to study Harbor Laboratory, Cold Spring Harbor, N.Y. the immune response to infection in an animal model. In the 13. Meyer, R. D. 1983. Legionella infections: a review of five years nude mouse model, both cellular and humoral mechanisms of research. Rev. Infect. Dis. 5:258-278. appear to be important for protection (D. J. Drutz, P. 14. Meyer, T. F., N. Mlawer, and M. So. 1982. Pilus expression in DeMarsh, J. Richard, W. Owens, R. Rolfe, P. Edelstein, and involves chromosomal rearrangement. S. Finegold. Exp. Cell Biol. 50:325, 1982; unpublished Cell 30:45-52. Cloned can be used to and 15. Nakamura, K., R. M. Pirtle, and M. Inouye. 1979. Homology of observations). antigens quantitate the gene coding for outer membrane lipoprotein within various characterize the humoral immune response to infection and gram-negative bacteria. J. Bacteriol. 137:595-604. to investigate cross-reactivity with other L. pneumophila 16. Pearson, G. D. N., and J. J. Mekalanos. 1982. Molecular cloning strains. In an animal model in which immune serum provides of enterotoxin genes in Escherichia coli K-12. passive protection, cloned antigens can be used to absorb Proc. Natl. Acad. Sci. U.S.A. 79:2976-2980. antisera selectively and to identify potential protective anti- 17. Purcell, B. K., and S. Clegg. 1983. Construction and expression gens. In models in which protective immunity is predomi- of recombinant plasmids encoding type 1 fimbriae of a urinary nantly cell mediated, they can be used in a similar manner to isolate. Infect. Immun. 39:1122-1127. detect T-lymphocyte responsiveness selectively. In studying 18. Sako, T., S. Sawaki, T. Sakurai, Y. Yoshizawa, and I. Kondo. the immunogenic components of L. pneumophila, separation 1983. Cloning and expression of the staphylokinase gene of of Staphylococcus aureus in Escherichia coli. Mol. Gen. Genet. antigens by cloning is particularly useful since many 190:271-277. important polypeptide surface components cannot be ade- 19. Smith, R. A., T. C. DiGiorgia, J. Darner, and A. Wilhelm. 1981. quately purified by biochemical means (1). Detection of Legionella pneumophila capsular-like envelope ACKNOWLEDGMENTS antigens by counterimmunoelectrophoresis. J. Clin. Microbiol. 13:637-642. We thank John Abraham, Janice Clements, Peter DeMarsh, 20. Stamm, L. V., J. D. Folds, and P. J. Bassford. 1982. Expression Douglas C. Dodd, Suzy Engleberg, Cynthia Freitag-Hall, Eric of Treponema pallidum antigens in Escherichia coli K-12. Pearlman, Robert Thornburg, and Janine Trempy for their assist- Infect. Immun. 36:1238-1241. VOL. 44, 1984 CLONING OF L. PNEUMOPHILA ANTIGENS 227

21. Stamm, L. V., T. C. Kerner, V. A. Bankaitas, and P. J. Bassford. W. 0. Schalla. 1979. "Endotoxicity" of the Legionnaire's 1983. Identification and preliminary characterization of Trepo- disease bacterium. Ann. Intern. Med. 90:624-627. nema pallidum protein antigens expressed in Escherichia coli. 25. Wong, K. H., W. 0. Schalla, R. J. Arko, J. C. Bullard, and J. C. Infect. Immun. 41:709-721. Feeley. 1979. Immunochemical, serologic, and immunologic 22. Towbin, H., J. Straehelin, and J. Gordon. 1979. Electrophoretic properties of major antigens isolated from the Legionnaire's transfer of proteins from polyacrylamide gels to nitrocellulose disease bacterium. Ann. Intern. Med. 90:634-638. sheets: procedure and some applications. Proc. Natl. Acad. Sci. 26. Wong, K. H., W. 0. Schalla, W. C. Wong, P. R. B. McMaster, U.S.A. 76:4350-4354. J. C. Feeley, R. J. Arko. 1982. Biologic activities of antigens 23. Vodkin, M. H., and S. H. Leppla. 1983. Cloning of the protective from Legionella pneumophila, p. 434-443. In L. Weinstein and antigen gene of Bacillus anthracis. Cell 34:693-697. B. Fields (ed.), Seminars in infectious diseases, vol. 4: bacterial 24. Wong, K. H., C. W. Moss, D. H. Hochstein, R. J. Arko, and vaccines. Thieme-Stratton, Inc., New York.