Isolation of Legionella longbeachae Serogroup 1 from Potting Mixes

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Jn. 1990, p. 49-53 0099-2240/90/010049-05$02.00/0 Copyright C) 1990, Americn Society for Microbiology Vol. 56, No. 1 Isoltion of Legionell longbeche Serogroup 1 from Potting Mixes TREVOR W. STEELE,* JANICE LANSER, AND NORMA SANGSTER Division of Clinicl Microbiology, Institute of Medicl nd Veterinry Science, Adelide 5000, South Austrli Received 30 My 1989/Accepted 2 October 1989 Following sttewide outbrek of legionellosis due to Legionell longbeche serogroup 1 in South Austrli in 1988 nd 1989, studies were performed to find source of the orgnism. A number of wter nd soil smples with nd without cid decontmintion were exmined for L. longbeche by using selective medium contining vncomycin, ztreonm, nd pimfucin. There were no isoltions of L. longbeche from wter smples. Orgnisms resembling L. longbeche were isolted from number of smples of potting mixes nd from soil surrounding plnts in pots collected from the homes of four ptients. The orgnisms were found to persist for 7 months in two potting mixes stored t room temperture. Legionelle were isolted with difficulty from potting mixes which were llowed to dry out. Identifiction of isoltes s L. longbeche serogroup 1 ws confirmed by quntittive DNA hybridiztion nd serologicl tests. Restriction-frgment-length-polymorphism studies showed minor differences between ptient nd environmentl isoltes but differentited these redily from L. longbeche serogroup 2 nd other ntigeniclly relted legionelle. The isoltion of L. longbeche from some potting mixes nd the prolonged survivl of the orgnisms in this medium suggest tht soil rther thn wter is the nturl hbitt of this species nd my be the source of humn infections. Legionell longbeche ws first described s new species in 1981 fter it ws isolted from ptient with pneumoni who resided in Long Bech, Clif. (8). Within short time, second serogroup of this orgnism ws described (1). Despite reports over the next 8 yers of spordic infections due to these orgnisms, n environmentl source for infections by L. longbeche hs not been found. In My 1987, L. longbeche serogroup 1 ws isolted for the first time from ptient in South Austrli, nd two more cses occurred in October 1987 nd June 1988. Two of these cses were reported by Lim et l. (5). In October 1988, sttewide outbrek of infection with L. longbeche serogroup 1 begn in South Austrli nd involved 23 ptients in 3 months. Spordic single cses were dignosed in Februry, April, nd June 1989. Between My 1987 nd June 1989, the dignosis of infection with this serogroup ws mde in 30 ptients. The dignosis ws confirmed by isolting the orgnism from 11 ptients, 8 of whom showed seroconversion, nd by demonstrting seroconversion to L. Iongbeche lone in nother 8 ptients. A presumptive dignosis of infection ws mde on the bsis of n elevted level of ntibody (1 in 256 or more) to L. longbeche in the remining 11 ptients. Epidemiologicl investigtions crried out by the South Austrlin Communicble Diseses Control Unit suggested tht grdening ws mjor risk fctor in cquiring L. longbeche infection. Between October nd December 1988, lbortory investigtions concentrted on the exmintion of wter smples collected from the homes of confirmed cses; prticulr ttention ws pid to wter collected from polypropylene wtering systems, since these were thought to be possible source of infection. When these investigtions proved negtive, it ws decided to extend the serch to soil smples. The Communicble Diseses Control Unit stff collected soil smples from the grdens of four ptients whose infection occurred in or fter December 1988 nd ws confirmed by culture or seroconversion. All ptients from whose homes soil smples were collected were ctive * Corresponding uthor. 49 grdeners nd hd used commercilly vilble potting mixes shortly before becoming ill. A limited number of wter smples from the homes of two ptients were lso exmined for L. longbeche. This pper describes the investigtion of these environmentl smples, the methods used to isolte L. longbeche from soils, nd the chrcteristics of the strins of L. Iongbeche found in these smples. MATERIALS AND METHODS Environment. Smples were collected from the homes of four ptients (Tble 1). VAP medium. An ntibiotic medium contining Legionell chrcol yest extrct gr bse (25 g/liter; CM 655; Oxoid Ltd., Bsingstoke, United Kingdom), ACES [N-(2-cetmido)-2-minoethnesulfonic cid; 10 g/liter]; ferric PPi (0.25 g/liter), L-cysteine hydrochloride (0.4 g/liter), -ketoglutrte (1 g/liter), vncomycin (2 mg/liter), ztreonm (8 mg/ liter), nd pimfucin (250 mg/liter; Gist-Brocdes, Delft, Netherlnds) (VAP medium) ws prepred. This followed tests which showed tht seven isoltes of L. longbeche serogroup 1 from ptients nd the type strin ATCC 33462 were reltively resistnt to ztreonm (MIC, 16 to 32 mg/ liter) nd were not significntly inhibited by this medium nd tht colonies of these orgnisms could be detected fter 3 dys of incubtion t 35 C. In qulity control tests, VAP medium selectively inhibited most common grm-negtive nd grm-positive bcteri found in clinicl bcteriologicl specimens, some soil bcteri, nd wide rnge of fungi. This medium ws prepred by utoclving chrcol yest extrct bse gr contining ACES nd -ketoglutrte. The ph ws djusted to 6.9 with KOH. Ferric PP1 nd L-cysteine were filter sterilized nd dded seprtely to the molten gr. Antibiotics were dded fter the medium ws cooled to 50 C. Processing of smples. Wter smples from the homes of ptients 1 nd 3 were concentrted by centrifuging 30 ml t 6,000 x g for 15 min. The deposit ws suspended in 1 ml of originl wter. After tretment with cid for 5 nd 10 min, it ws inoculted in duplicte onto VAP medium with 5-mm nichrome wire loop. Acid-treted smples were not neutrlized. Pltes were incubted t 35 C for 7 dys nd were

50 STEELE ET AL. TABLE 1. Cse no. Grden smples from four ptients with confirmed L. longbeche infection Smple(s) (n) 1 Wter from grden nd house (12) Home-mde compost (1) Potting mixes A nd B (2) Grden soil (3) Potted-plnt soil (4) 2 Potting mix C nd D (2) Homemde compost (1) Composted cow mnure (1) Blood nd bone fertilizer mix (1) 3 Potting mix E (1) Potted plnt-soil (brnd E) (1) Grden soil (2) Rin wter smple from tnk (1) 4 Potting mix F (1) Homemde potting mix (2) exmined dily with n Olympus plte microscope fter 72 h of incubtion. Soil smples, including smples of potting mixes, were prepred for culture by mking 1-in-5 (wt/vol) dilution of soil in 70 ml of sterile tp wter. The suspensions were held t room temperture for 2 to 4 h nd were shken two or three times before being filtered through guze. A portion of unconcentrted smple ws retined for testing. A concentrted smple ws prepred by centrifuging 30 ml of soil suspension t 6,000 x g for 15 min. Most of the superntnt ws discrded, nd the deposit ws suspended in the remining 1 to 2 ml of wter. Concentrted nd unconcentrted smples of soil suspensions were decontminted by mking 1-in-20 (wt/vol) dilution of the smple in 0.2 M HCI-KCI buffer. A loopful (5-mm nichrome wire loop) of ech smple ws plted in duplicte on VAP medium fter 5 nd 10 min of cid tretment. Suspensions of soil smples, including potting mix smples from the homes of ptients 1 nd 2, were cultured with nd without cid tretment nd were plted to Legionell-selective gr (Oxoid) contining cefmndole, polymyxin, nd nisomycin; to modified Wdowsky nd Yee (MWY) medium (Oxoid) contining vncomycin, polymyxin, nd nisomycin s selective gents; nd to chrcol yest extrct gr without ntibiotics. Pltes were incubted t 35 C in humidified incubtor nd were exmined dily for 7 dys with n Olympus plte microscope fter 72 h of incubtion. A smple of potting mix brnd A plced in plstic flower pot (50 by 75 mm) ws left outdoors for 18 weeks nd wtered weekly with legionell-free rinwter when the wether ws wrm nd dry but not during utumn nd winter (My through August 1989). Smples of wter were collected from the bottom of this flower pot by wtering the soil t 4, 6, 8, nd 18 weeks nd were tested for legionelle by the methods described bove. Severl soil smples found to be positive for L. longbeche were suspended in wter. These were treted with cid for vrious times nd heted t 50 C for 15 nd 30 min. Heted smples were plted in duplicte with wire loop directly without cid tretment nd fter cid tretment. Resistnce to drying. The effect of drying on two potting mixes found to contin L. longbeche ws exmined. Two mixes were dried in uncovered petri dishes, one t 36 C for 7 dys nd one t 36 C for 2 h nd t room temperture for APPL. ENVIRON. MICROBIOL. TABLE 2. Pttern of gglutintion rections shown by ltex regents coted with ntibodies to L. longbeche serogroup 1 nd L. cincinntiensis Ltex gglutintion rections' Test orgnism (ATCC no.) L. Iongbeche L. cincinnserogroup 1 tiensis L. longbeche serogroup 1 (33462) 3+ L. Iongbeche serogroup 2 (33484) - L. cincinntiensis (43753) 2+ 3+ L. sinthelensi (35248) 1+ 1+ L. snticrucis (35301) 1+ 1+ Grding of gglutintion rections: 3+, 75% to 100% of cells gglutinted; 2+, pproximtely 50% of cells gglutinted; 1+, pproximtely 25% of cells gglutinted; -, no gglutintion. 5 dys. Undried smples of both mixes in seled continers were held t room temperture nd 36 C to serve s controls. All smples were tested in duplicte fter cid tretment. Chrcteriztion of isoltes. Colonies of legionelllike orgnisms ppering on VAP pltes nd other ntibioticcontining pltes fter 3 or more dys of incubtion were tested for gglutintion with ltex regents prepred with L. longbeche serogroup 1 ntiserum nd L. cincinntiensis ntiserum obtined by immunizing rbbits with suspensions of Americn Type Culture Collection type strins. These unbsorbed ntiser were diluted by dding different mounts to 1% ltex suspensions which were then tested ginst pnel of 26 Legionell species. Ltex regents showing resonbly good specificity nd rpid gglutintion with the lowest concentrtion of ntiserum were used throughout this study. The pttern of rections found with these ltex regents when tests were performed on ntigeniclly relted Legionell spp. is shown in Tble 2. Orgnisms were identified s L. longbeche serogroup 1 if they hd typicl colonil morphology on VAP medium, were dependent on L-cysteine, did not show blue-white utofluorescence, nd gglutinted rpidly (within 30 s) with the ltex ntiserum to L. Iongbeche serogroup 1 but not with the ntiserum to L. cincinntiensis. In this study, ll isoltes which showed gglutintion on slide tests were checked by direct fluorescent-ntibody (DFA) stining with fluorescein isothiocynte-conjugted ntiser prepred ginst L. longbeche serogroups 1 nd 2 (Centers for Disese Control, Atlnt). The pttern of rections shown by these DFA regents is shown in Tble 3. DNA homology nd RFLP studies. The identity of the isoltes ws confirmed by DNA homology studies using the quntittive method of Brenner et l. (2), which ws modified for use with smll volumes (5). The stndrd strin used in the homologous rection ws L. longbeche serogroup 1 ATCC 33462. The ntigeniclly relted L. Iongbeche serogroup 2 ATCC 33484, L. cincinntiensis ATCC 43753, L. snticrucis ATCC 35301, nd L. sinthelensi ATCC 35248 were lso included. Whole chromosoml DNA for restriction-frgment-length-polymorphism (RFLP) studies ws purified from ech isolte by the method of Mnning et l. (7). Smples of the preprtions were electrophoresed on 0.8% grose gels to confirm tht miniml frgmenttion of the chromosoml DNA hd occurred. The DNA ws fully digested with restriction enzymes HindlIl nd BmHI, nd the frgments were seprted by overnight electrophoresis on 0.8% grose gels in TAE buffer (6). The frgments were then trnsferred by the procedure of Southern (11) to nylon filter membrnes (Biotrce RP; Gelmn Sciences, Inc., Ann Arbor, Mich.)

VOL. 56, 1990 ISOLATION OF L. LONGBEACHAE FROM POTTING MIXES 51 TABLE 3. Pttemn of rections shown by DFA regents tested on serologiclly relted Americn Type Culture Collection cultures of Legionell spp. Species nd group DFA rection to L. longbeche serogroup: 1 2 L. Iongbeche serogroup 1 serogroup 2 4+ 1+ 1-2+ 4+ L. cincinntiensis 3+ 1-2+ L. sinthelensi 1+ 2+ L. snticrucis 4+ 1-2+ L. bozemnii serogroup 1 - - serogroup 2-2+ L. nis - 3+ Grding of rections: 4+, brillint yellow-green stining of bcteril cells; 3+, bright yellow-green stining; 2+, definite but dim stining; 1+, brely visible stining; -, no stining. which were then bked t 80 C for 2 h nd stored t room temperture. Derivtion of DNA probe. A genoml bnk for L. longbeche serogroup 1 ATCC 33462 ws constructed in LmbdGEM-11 (Promeg Biotec, Mdison, Wis.) ccording to the instructions of the mnufcturer. Amplified stocks of severl clones were prepred nd used to infect lwns of Escherichi coli LE392 t concentrtions which would yield confluent lysis. Bcteriophge ws hrvested from the plques by the method detiled by Mnitis et l. (6). The phge suspension ws then clrified t 17,500 x g for 20 min t 4 C. The pellet ws discrded. The phge ws pelleted from the superntnt by centrifugtion t 40,000 rpm for 2 h t 4 C in n L8, 70 Ti rotor (Beckmn Instruments, Inc., Fullerton, Clif.). The superntnt ws discrded, nd the phge pellet ws suspended in phosphtebuffered sline. The vector with the L. longbeche insert ws purified from the phge prticles by the method of Mnitis et l. (6). The vector with insert ws lbeled with [-32P]dCTP by nick trnsltion (6). Hybridiztion. The filters were hybridized overnight t 42 C in hybridiztion fluid (50% formmide, 1% skim milk TABLE 4. Five grden smples found to contin legionelle resembling L. longbeche Smple Lbortory Ptient illness Dte of designtion Type purchse of isoltes 1 Februry 1989 Potting mix A Jnury 1989 El Soil from one Unknown E2 potted plnt 2 December 1988 Potting mix Db Februry 1989 E3 3 April 1989 Soil from April 1989 E4 potted plnt 4 June 1989 Potting mix F April 1989 E5 L. longbeche serogroup 1 strins isolted from ptients were designted Hi, H2, nd H3. Legionelle were not isolted from ptient 4, who showed seroconversion to L. longbeche. b Brnd normlly used by the ptient but purchsed fter his illness. Mix used before illness ws unvilble for testing. powder, 7% sodium dodecyl sulfte, 5 x SSPE (1x SSPE is 0.18 M NCl, 10 mm sodium phosphte, nd 1 mm EDTA), 200,ug of slmon sperm DNA per ml, rdiolbeled probe DNA [5 x 105 cpm/ml]). The next dy, filters were wshed twice for 15 min t room temperture in 2 x SSC (lx SSC is 0.15 M NCl plus 0.015 M sodium citrte [ph 7.0])-0.1% sodium dodecyl sulfte nd then once t 68 C in lx SSC- 0.1% sodium dodecyl sulfte for 45 min nd utordiogrphed t -70 C. RESULTS Orgnisms resembling L. longbeche were isolted from the soil of two potted plnts nd three potting mixes on VAP medium (Tble 4) but not from ny grden soil or wter smples. All medi, including VAP pltes inoculted with smples nd not treted with cid, were overgrown with soil bcteri, nd legionelle were not detected on these pltes. In ddition, chrcol yest extrct nd both Oxoid selective medi inoculted with smples treted with cid were overgrown with soil bcteri nd soil fungi fter 3 dys of incubtion nd hd to be discrded. L. Iongbeche could still be isolted from soil suspensions treted with cid for 5 to 20 min nd from those heted t 50 C for 30 min, lthough cid tretment of the heted smples ws necessry. Acid tretment of soil suspensions for 10 min proved to be the most effective wy of reducing the numbers of soil orgnisms present in these smples. When present in moderte to lrge numbers (>104 CFU/g of soil), legionelle were generlly isolted from both cidtreted smples. When lower numbers were present, they were more esily detected in smples treted for 10 min, becuse of better suppression of other soil bcteri. The numbers found in the potted plnt soil nd the potting mix smples were clculted to rnge from 102 to 105 CFU/g of soil. Potting mix brnds A nd D remined positive on testing for 7 months, nd L. longbeche ws isolted from concentrted smple of lechte fter 18 weeks of exposure of the flower pot to nturl climtic conditions. Legionelle were not isolted from known positive soil smples which were dried for 7 dys, but they remined detectble in undried smples held t room temperture nd t 36 C. The L. longbeche strins isolted from environmentl smples nd ptients showed similr colonil morphologies on VAP medium. Incubtion of pltes for longer thn 4 dys resulted in the development of colonies with concentric rings, which gve them chrcteristic trget ppernce. All three humn nd five environmentl Legionell isoltes required L-cysteine for growth. All isoltes rected strongly nd rpidly with the L. longbeche serogroup 1 ltex regent nd showed bright brillint stining with the DFA regent for this serogroup. They did not rect with the L. cincinntiensis ltex regent but did rect wekly with the L. longbeche serogroup 2 DFA regent. The identity of the environmentl isoltes ws confirmed by DNA homology studies with L. longbeche ATCC 33462. The DNA reltive binding rtios of the three humn nd five environmentl isoltes were bove 80% t the stringent temperture of 75 C. L. cincinntiensis ATCC 43753, L. snticrucis ATCC 35301, nd L. sinthelensi ATCC 35248 gve reltive binding rtios of 13, 13, nd 11%, respectively. Figure 1 shows the RFLP ptterns obtined by Southern hybridiztion of isoltes digested with HindIll nd BmHI using probe mde from cloned frgment of the L. longbeche genome. The type strin of L. longbeche

52 STEELE ET AL. 1 2 3 4 5 6 7 8 9 FIG. 1. Southern hybridiztion of Hindlll nd BmHI digests of genomic DNA. Lnes: 1, L. longbeche serogroup 1 ATCC 33462; 2, L. longbeche serogroup 2 ATCC 33484; 3, L. cincinntiensis ATCC 43753; 4, L. snticrucis ATCC 35301; 5, L. sinthelensi ATCC 35248; 6 to 13, isoltes Hi, El, E2, H2, E3, H3, E4, nd E5, respectively. serogroup 1 differed from the Americn Type Culture Collection strins of L. longbeche serogroup 2, L. cincinntiensis, L. snticrucis, nd L. sinthelensi. The humn nd environmentl isoltes closely resembled ech other nd L. longbeche ATCC 33462. Among the L. longbeche serogroup 1 strins tested, only minor differences in profiles were observed. Two humn isoltes, Hi nd H2, hd n extr bnd below the hevily lbeled double bnd not seen in the third humn strin or in the environmentl strins, while E5 lcked smller bnd found in ll other strins. The Americn Type Culture Collection strin of L. Iongbeche serogroup 1 hd single bnd of high moleculr weight, wheres Austrlin strins ll hd finter double bnd in tht region. DISCUSSION Orgnisms identified s L. longbeche serogroup 1 were isolted from three potting mixes nd from the soil of two potted plnts obtined from the homes of four ptients who hd hd pneumoni due to this species. The orgnisms were shown to persist in some of these smples for 7 months. They were not isolted from ny wter or nturl soil smples exmined in this study. Severl reports tht L. longbeche hs been found in qutic environments hve been published (3, 9, 10). In these studies, identifiction relied minly on serologicl tests. Definitive DNA homology tests were not performed to confirm the identity of the L. longbeche serogroup 2 isoltes in Europe (3) or of the isolte of L. longbeche (serogroup not stted) from hotel in the Cribben (10). L. longbeche ws reported to be present in sewter in Puerto Rico when DFA tests were used, but isoltion of these orgnisms ws not reported (9). Serologicl crossrections between L. longbeche serogroup 2 nd two species which show blue-white utofluorescence, L. nis nd L. bozemnii serogroup 2 (13), re known to occur. Cross-rections lso occur between L. longbeche serogroup 1, L. cincinntiensis, L. snticrucis, nd L. sinthelensi (12, 13) nd were demonstrted in this study with DFA tests. Therefore, DFA cnnot be used to differentite mong these species relibly. All these Legionell spp. hve common ubiquinone profiles nd belong to group B of Lmbert nd Moss (4). The slide gglutinting ltex regents used in APPL. ENVIRON. MICROBIOL. this investigtion distinguished L. Iongbeche serogroup 1 from the other serologiclly relted nonutofluorescing species. Our soil isoltes rected in the sme wy s the type strin of L. longbeche serogroup 1 when tested with DFA nd slide gglutintion. The identifiction of these environmentl isoltes s L. longbeche ws confirmed by DNA homology studies. RFLP nlysis provided more precise comprison of L. longbeche isoltes from ptients nd the environment. Minor differences in profiles were noted between humn isoltes Hi nd H2 nd isolte H3 nd between environmentl isoltes El through E4 nd isolte E5. RFLP nlysis clerly differentited L. Iongbeche serogroup 1 from the serogroup 2 strin nd from other legionelle which hve very similr ubiquinone ptterns nd show some DNA nd serologicl reltedness. This result is comptible with the view tht these strins re closely relted nd tht L. longbeche in soil could be the source of some humn infections. The potting mixes nd potted plnt soils from which the environmentl isoltes El, E3, nd E4 of L. longbeche were grown were mde by one mnufcturer but distributed under seprte brnd nmes. Isolte E5 ws obtined from potting mix (brnd F) mde by different compny. It is generlly ccepted tht pneumoni results from inhltion of erosols contining Legionell spp. It is not known if this is the route for L. Iongbeche infections. Our investigtions hve shown tht legionelle re leched from the soil during wtering of potted plnts. It is possible tht erosols contining L. Iongbeche re formed by vigorous wtering of plnts in pots or when wter flls onto hrd surfces from potted plnts plced bove ground level. Cse-controlled studies re needed to determine if potting mixes re responsible for infection nd how such infections re cquired. Now tht n environmentl source of L. longbeche hs been found, it should be possible to investigte the environmentl, host, nd orgnism fctors which led to infection. ACKNOWLEDGMENTS We thnk Crolyn Wlker nd Scott Cmeron of the South Austrlin Helth Commission for their ssistnce, Sen McDermott for ntibiotic testing of legionelle, Philip Mugg nd Shron Blckett for prepring medi nd ltex regents, Robyn Doyle for ssisting with the RFLP studies, nd M. Priede for prepring the mnuscript. LITERATURE CITED 1. Bibb, W. F., R. J. Sorg, B. M. Thomson, M. D. Hicklin, A. G. Steigerwlt, D. J. Brenner, nd M. R. Wulf. 1981. Recognition of second serogroup of Legionell longbeche. J. Clin. Microbiol. 14:674-677. 2. Brenner, D. J., G. R. Fnning, A. V. Rke, nd K. E. Johnson. 1969. Btch procedure for therml elution of DNA from hydroxyptite. Anl. Biochem. 28:447-459. 3. Desplces, N., N. Nhepetin, A. Bure, nd E. Dournon. 1982. Legionell longbeche serogroup 2 in Europe. Lncet ii:1400. 4. Lmbert, M. A., nd C. W. Moss. 1989. Cellulr ftty cid compositions nd isoprenoid quinone contents of 23 Legionell species. J. Clin. Microbiol. 27:465-473. 5. Lim, I. L., N. Sngster, D. P. Reid, nd J. A. Lnser. 1989. Legionell longbeche pneumoni: report of two cses. Med. J. Aust. 150:599-601. 6. Mnitis, T., E. F. Fritsch, nd J. Smbrook. 1982. Moleculr cloning: lbortory mnul. Cold Spring Hrbor Lbortory, Cold Spring Hrbor, N.Y. 7. Mnning, P. A., M. W. Heuzenroeder, J. Yedon, D. I. Levesley, P. R. Reeves, nd D. Rowley. 1986. Moleculr cloning nd

VOL. 56, 1990 ISOLATION OF L. LONGBEACHAE FROM POTTING MIXES 53 expression in Escherichi coli K-12 of the 0 ntigens of the Inb nd Ogw serotypes of the Vibrio cholere 01 lipopolyscchrides nd their potentil for vccine development. Infect. Immun. 53:272-277. 8. McKinney, R. M., R. K. Porschen, P. H. Edelstein, M. J. Bissett, P. P. Hrris, S. P. Bondell, A. G. Steigerwlt, R. E. Wever, M. E. Ein, D. S. Lindquist, R. S. Kops, nd D. J. Brenner. 1981. Legionell longbeche species nov, nother etiologic gent of humn pneumoni. Ann. Int. Med. 94:739-743. 9. Ortiz-Roque, C. M., nd T. C. Hzen. 1987. Abundnce nd distribution of Legionellcee in Puerto Ricn wters. Appl. Environ. Microbiol. 53:2231-2236. 10. Schlech, W. F., III, G. W. Gormn, M. C. Pyne, nd C. V. Broome. 1985. Legionnires' disese in the Cribben: n outbrek ssocited with resort hotel. Arch. Intern. Med. 145: 2076-2079. 11. Southern, E. M. 1975. Detection of specific sequences mong DNA frgments seprted by gel electrophoresis. J. Mol. Biol. 98:503-517. 12. Thcker, W. L., R. F. Benson, J. L. Stneck, S. R. Vincent, W. R. Myberry, D. J. Brenner, nd H. W. Wilkinson. 1988. Legionell cincinntiensis sp. nov. isolted from ptient with pneumoni. J. Clin. Microbiol. 26:418-420. 13. Thcker, W. L., B. B. Plikytis, nd H. W. Wilkinson. 1985. Identifiction of 22 Legionell species nd 33 serogroups with the slide gglutintion test. J. Clin. Microbiol. 21:779-782.