The determination of ubiquinone profiles by reversed-phase highperformance thin-layer chromatography as an aid to the speciation of Legionellaceae

Transcription

1 Journal of General Microbiology (990), 6, 050. Printed in Great ritain 05 The determination of ubiquinone profiles by reversedphase highperformance thinlayer chromatography as an aid to the speciation of Legionellaceae K. MITCHLL* and R. J. FLLON epartment of Laboratory Medicine, Ruchill Hospital, ilsland rive, Glasgow G0 9N, UK (Received January 990; revised 8 June 990; accepted 5 July 990) Ubiquinones exacted from sains of LegioneZZa pneumophiza (including the type sains of serogroups and one proposed new serogroup) and from sains of other LegwneZh species (including 8 type sains and sains of five proposed new species) were analysed by reversedphase thinlayer chromatography. Ubiquinone profiles as determined by this method were reproducible, both qualitatively and semiquantitatively, and provided information to aid in the identification of species of Legionella. Results of ubiquinone profiles determined by different laboratories were compared. Some quantitative differences in results between laboratories were observed, which may be due to different analytical procedures. For this reason laboratories should establish their own library of ubiquinone profiles. New information is presented on the ubiquinone profiles of seven LegioneZh species: L. birmhghamensis, L. brunensis, L. c ~~&nsis, L. moravica, L. quidd, L. tucsonensis and proposed new species no. 7, L. worsleiensis. Inoduction Ubiquinones and menaquinones are constituents of bacterial plasma membranes, where they play an important role in respiratory elecon ansport. Suctural differences in these respiratory quinones have been used as a basis for bacterial classification (Collins et al., 977). The Legionellaceae possess an unusual pattern of respiratory ubiquinones where the number of isoprene units (Q) on the benzoquinone ring ranges between 9 and 5. In other families of bacteria, the chain length ranges from 7 to 0 units (Gilbart & Collins, 985). These differences in hydrocarbon chain length allow a mixture of ubiquinones to be separated (Collins & Jones, 98). Using reversedphase thinlayer chromatography (RPTLC) and mass specomey, Karr et al. (98) identified Q9 to in a range of Legionella species. These observations were extended by Moss et al. (98), who detected, as well as Q, in L. feeleii WOC by both RPTLC and reversedphase highperformance liquid chromatography ; these ubiquinones were confirmed by mass specomey. Collins & Gilbart (98) published quantitative ubiquinone profiles, which included and Q5, of twelve Legionella test sains bbreviations : RPTLC, reversedphase thinlayer chromatography; SG, serogroup. determined by highperformance liquid chromatography (HPLC). Comprehensive information on ubiquinones found in Legionella species analysed by HPLC has been published by Lambert & Moss (989). In our laboratory HPLC is not available and a library of ubiquinone profiles of all known Legionella species has been established by RPTLC. This method is convenient in that it does not require sophisticated and expensive equipment, is experimentally undemanding and gives highly reproducible results. It is therefore a simple method which could be employed by clinical laboratories. Here we describe the method and report ubiquinone profiles for seven species of Legionella which have not been investigated previously in this respect. Methods acterial sains and cultural conditions. Sains of Legionella (Table ) were grown on CSbuffered charcoal yeast exact agar (CY) (delstein, 98 ) for 8 h at 7 & C in air in sealed jars. Two sains were also grown on Oxoid CY (code CM 655 with Legionella growth supplement SR 0) and incubated as described for CY. irect exaction of ubiquinones. modification of the procedure of Krivankova & adak (980) was used, as described by Karr et al. (98). riefly, fresh whole cells scraped from the surface of two Pei dishes (00 x 5 mm), taking care to avoid picking up medium fragments, were suspended in ml methanol/nhexane ( :, v/v) in a glass container. This suspension was vortexmixed for min and then O 990 SGM ownloaded from by IP: On: Sun, ec 07 09::07

2 06 K. Mitchell and R. J. Fallon Table, Legionella sains investigated Species Serogrou p Sain no. Souref L. anisa L. birminghamensis L. bozemanii L. bozemanii L. bnmensis L. cherrii L. cincinnatiensis L. dumofli L. erythra L. feeleii L. feeleii L. geestiae * L. gormanii L. gratiana L. hackeliae L. hackeliae L. israelensis L. jamestowniensis L. jordanis L. londoniensis * L. Iongbeachae L. longbeachae L. macheachernii L. micdadei L. moravica L. nautarum * L. oakridgensis L. parisiensis (Camperdown) (Knoxville ) (Olda la) (Olda lc) (Philadelphia ) (Pontiac ) (Hastings sain) CH7C 07LH WIG Toronto ORW 70OHH TXKL SC8 w w c 69 W H 08 LS TCC 9 Lansing 798PH ercovier J6Gl L50 Longbeach Tucker PX G TTLOCK 66 7 OR0 PF09CC NCTC NCTC 9 Togus loomington Lyon Nottingham PHL 8ml nhexane was added and the mixing procedure repeated. Cenifugation at 0g for 0 min assisted partition of the solvents, and the top layer of nhexane was removed without disturbing any cellular debris at the methanol/nhexane interface. The aqueous phase was reexacted with a further 8 ml nhexane. The combined hexane exacts were dried in vacuo in foilcovered vials to avoid photodegradation. RPTLC analysis of ubiquinones. Cellular exacts (050 pl) were spotted onto 0 x 0 cm octadecylsilanebonded reversedphase KCl8 F TLC plates (Whatman), ensuring similar quantities of exacted material between samples, together with 0 p of ubiquinone standards Q6 and Q0 (Sigma), each at concenations of 0.5 mg mll. Plates were developed in acetone/water (9 :, v/v) until the solvent had risen to within cm of the top. Then, after air drying, plates were sprayed with 0% (w/v) phosphomolybdic acid (Sigma) from a distance of 0 5 cm followed by heating at 60 C for 050 s, when bluegreen spots developed on a yellow background. In addition to ubiquinones, other hexanesoluble materials, e.g. phospholipids, were observed on the RPTLC plates but these were well separated from ubiquinones and did not interfere with their interpretation. Relative mobility (RF values) were calculated by dividing the distance avelled by each ubiquinone individually by the distance avelled by the solvent. This was done for the two commercial ubiquinone standards, as well as ubiquinones exacted from L. pneumophila SG (Knoxville), used as a conol for the exaction method. These RF values were plotted against the number of isoprenoid units (Q) (see Fig. ). This allowed tentative identification of ubiquinones exacted from other legionellae. Saponification of ubiquinones. modification of the procedure of be et al. (978), described by Karr et al. (98), was employed. Fresh whole cells (0. g wet weight) were scraped from two CY agar plates; a suspension was made in ml % (w/v) pyrogallol in methanol and this was ansferred to a 0 ml Quickfit flask. To this, 0. mlso% aqueous KOH was added. The flask was attached to a reflux condenser (60 mm) and the mixture refluxed for 0 min at 00 C followed by immediate cooling by placing the flask under running tap water. ml quantity of distilled water and 5ml nhexane were added and the mixture shaken vigorously for 5 min and then cenifuged at 0 g for 0 min. The hexane layer was removed and the aqueous phase was reexacted with an additional 5 ml nhexane. The hexane layers were combined and evaporated to dryness in v am and the resulting residue dissolved in ethyl acetate for RPTLC. Mass specomey. Some of the exacts used in this work were examined at the PHLS Cene for pplied Microbiological Research, Porton own, Salisbury, UK, and the presence of ubiquinones in these exacts was confirmed by mass specomey using a Kratos MS 80 RF mass specometer fitted with an IonTech NF fastatom gun ownloaded from by IP: On: Sun, ec 07 09::07

3 Ubiquinone procfiles of 5 Legionella species 07 Table (continued) Species Serogroup Sain no. Source7 Subsp. Losngeles 5 fraseri allas MICU Columbia Subsp. u7w Columbia pascullei U8W Columbia Oxford C hicago8 Concord INGlC Leiden 797PH 570COH Seattle 69MNH * Subsp. fraseri L. quateriensis * L, quinlivanii 5 US 8US 9US L. quinlivanii 50US L. rubrilucens L. sainthelensi L. santicrucis L. spiritensis L. steigerwaltii L. tucsonensis L. wadsworthii L. worsleiensis * 5US 5US 59US 67 Mt St Helens SC6C7 Mt St Helens9 SC8C 087H * Proposed new species or unnumbered new serogroup. t Cultures were obtained from : the culture collection of the Centers for isease Conol (), tlanta, Georgia, US; PHLS, Porton own, Salisbury, UK; r N. ornstein, Cene National de Reference des Legionelloses, Lyon, France; r. rown, Wm Jennings ryon orn Veterans Hospital, Columbia, US; and r.. Macrae, PHL Nottingham, UK. and a Kratos fastatom bombardment (F) source. This method of negativeion fastatom bombardment has not previously been applied to the identification of ubiquinones but provides valuable data (R. Wait, personal communication). Results and iscussion Fig. shows the relationship between the number of isoprenoid units and the relative mobility (RF) of ubiquinones on RPTLC plates. Some plates were examined under UV light (5 nm) and areas of UV absorption representing ubiquinones were observed. lthough absorption patterns were seen, it was difficult to estimate in what quantities these ubiquinones were present. The phosphomolybdate spray reagent, although not specific for ubiquinones, did yield more detail in the areas where ubiquinones were expected to be resolved. This greater sensitivity and the resultant semipermanent RPTLC plate made results easier to read and record and this outweighed the No. of isoprene units (Q) Fig.. Plot of the RF values on RPTLC (log scale) of ubiquinones exacted from SG (Knoxville)(0) and of purified standards, 6 and Q0 (Sigma)(0) against the number of isoprenoid units (Q). disadvantage of other unsaturated lipids being shown up by the phosphomolybdate on the plate. The method used for saponification of the exacts removed many of the other lipids observed on the ownloaded from by IP: On: Sun, ec 07 09::07

4 08 K. Mitchell and R. J. Fallon Table. Ubiquinone projles of 5 Legionella species as determined semiquantitatively by RPTLC and grouped according to the ubiquinone observed to be of the highest relative concenation The numbers refer to visual estimates of the relative amounts of ubiquinones, with the major components designated as, (half the amount of ), (> but < ), and (half the amount of )., ace amounts;, not detected. Predominant ubiquinone: group, ; group, ; group C, Q ; group, 0; group, species with two or more ubiquinones present in equally high concenation. Ubiquinone content Species L. anisa L. birminghamensis L. bozemanii SGl and L. brunensis L. cherrii L. cincinnatiensis L. dumofii L. erythra L. feeleii SG and L. geestiae ( 08) L. gormanii L. gratiana L. hackeliae SGl and L. israelensis L. jamestowniensis L. jordanis L. londoniensis ( ) L. longbeachae SGI and L. maceachernii L. micdadei L. moravica L. nautarum (7) L. oakridgensis L. parisiensis SG L. quuteriensis ( 5) L. rubrilucens L. sainthelensi L. santicrucis L. spiritensis L. steigerwaltii L. tucsonensis L. wadsworthii L. worsleiensis ( 7) 9 0 Group C RPTLC plate beyond the area where ubiquinones were resolved. Results obtained by this method were compared, for 0 sains, with direct exaction. Only one sain (L. maceachernii) showed any difference, less Q0 being found in the saponified exact than in the direct exaction. The direct exaction procedure took less time and thus was chosen to compile the library of ubiquinone profiles of the Legionellaceae. Table was compiled from the results of ubiquinone analyses by RPTLC of 5 Legionella species. semiquantitative assessment of the ubiquinones was achieved subjectively, the density of spots being assigned a numerical value on the basis of intensity observed on the RPTLC plate. ll analyses by RPTLC were repeated at least twice to check ubiquinone profile stability, with identical results being obtained for each LegwneZla sain tested. ll sains analysed contained ubiquinones with 0 isoprenoid units in the side chains. Twelve species (fourteen sains) had ace to small amounts of Q, but L. feeleii SG and, and proposed new species no. 08, L. geestiae, had substantial amounts of this ubiquinone. Q9 was found in 9 of the species (7 sains) analysed, including L. longbeachae SG and, and the seven sains of (Table ). However, Q9 was more difficult to detect on a RPTLC plate since it ran close to Q0 and other lipid components. Visually, the other ubiquinones were more easily distinguished. No qualitative or quantitative differences were detec ownloaded from by IP: On: Sun, ec 07 09::07

5 Ubiquinone projles of 5 Legionella species 09 Fig.. RPTLC plate of ubiquinones (, Q0;, Q; C, ;, Q) exacted from fresh whole cells of species: lane, L. pn. SG (Knoxville); lane, L.pn. SG ; lane, L.pn. SG ; lane, L.pn. SG ; lane 5, L.pn. SG 5 (allas ); lane 6, L.pn. SG 6; lane 7, L.pn. SG 7; lane 8, L.pn. SG 8; lane 9, L.pn. SG 9; lane 0, L.pn. SG 0; lane, L.pn. SG ; lane, L.pn. SG ; lane, L.pn. SG ; lane, L.pn. SG ; lane 5, 6 standard (Sigma); lane 6, Q0 standard (Sigma). table in the ubiquinone profiles of the sains analysed (Fig. ), but many differences, both qualitative and quantitative, were evident when profiles of some of the other Legionella species were examined (Fig. ). The ubiquinone profiles of the 68 Legionella sains allowed each species to be placed in one of five groups, designated (Table ). ll the sains of L. pneumophila, including subsp. pneumophila (N group I), subsp. fiaseri (N group ) (Fig. ) and subsp. pascullei (renner et al., 988), gave ubiquinone profiles which were indistinguishable. These were assigned to group (Table ) as was present in the highest relative quantity. Groups, C, and contained species for which the ubiquinone with the highest relative concenation was, Q and Qlo, respectively. Group contained species where no one ubiquinone predominated, i.e. two or more ubiquinones were observed in the highest quantity. This system of grouping legionellae according to their predominant ubiquinone is simple both in terms of definition of a group and of allocation of sains to a group. [It should not be confused with that of Lambert & Moss ( 989), where Legionella species were placed in groups according to their ubiquinone profiles as analysed by HPLC.] The five groups dividing the Legionellaceae are disibuted as follows: group, eight species (%); group, eight species (%)); group C, one species (%); group, eight species (%); and group, ten species (8 %). Legionella species assigned to any one group may be further differentiated if necessary on the relative amounts of ubiquinones other than that which is dominant. Ubiquinone profiles analysed by RPTLC may also be used to discriminate between the blue/white autofluorescent and the red autofluorescent species where these may be difficult to identify serologically. Of the eight currently described blue/white autofluorescent species, L. dumofii, L. gormanii, L. steigerwaltii and L. tucsonensis were designated members of group by their ubiquinone profiles and could be differentiated from L. bozemanii, L. cherrii and L. parisiensis of group and L. anisa of group. The two red autofluorescent species L. erythra and L. rubrilucens could also be differentiated, as they were assigned to groups and respectively. The reproducibility of ubiquinone profiles was investigated in two legionellae, a human pathogenic sain, L. pneumophila SG (Knoxville sain), and an environmental sain, 'L. geestiae'. Profiles were constant whether cultures were incubated for 8,96, 9 or 0 h at 7 "C and also whether grown on Oxoid or inhouse CY. RPTLC has been used to analyse sains other than the stock sains shown in Table. Six legionellae isolated from the environment reacted with unabsorbed L. ownloaded from by IP: On: Sun, ec 07 09::07

6 00 K. Mitchell and R. J. Fallon Fig.. RPTLC plate showing some of the similarities and differences of ubiquinone profiles (summarized in Table ) amongst twelve sains of legionellae. Lane, L. dumofii; lane, L. erythra; lane, L. feeleii SG ; lane, L. feeleii SG ; lane 5, L. geestiae ; lane 6, L. gormanii; lane 7, L. hackeliae SG ; lane 8, L. hackeliae SG ; lane 9, L. isruelensis; lane 0, L. jamestowniemis; lane, L. jordanis; lane, L.pneumophila SG (Knoxville) conol; lane,q6 standard (Sigma); lane, Q0 standard (Sigma). Ubiquinones are:, Q9;, Q0; C, Q;, ;, ; F,. Fig.. RPTLC plate showing ubiquinone profiles from six legionellae isolated from environmental samples (lanes 6) and two type sains, L. hackeliae SG (lane 7) and SG (Knoxville) (lane 8). Lane 9, 6 standard (Sigma); lane 0, Q0 standard (Sigma). Ubiquinones are:, Q0,, Q; C, ;,. hackeliae SG guineapig antiserum (IF tie = 5, positive conol with the homologous organism, 0). Their ubiquinone profiles (Fig. ), matched closely those of and not those of L. hackeliae. Further studies of fatty acids by GLC supported these findings and the organisms were subsequently identified using serogroupspecific rabbit antisera to be SG (three sains), SG6 (two sains) and SG (one sain). nother sain of Legionella (ML76) was analysed for its ubiquinone profile and found to be indistinguishable from L. spiritensis, agreeing with the findings of Harrison et al. (988). lso our findings with L. gratiana, in which Q0 is the predominant ubiquinone, agree with those of ornstein et al. (989), which were derived using HPLC. These results suggest that ubiquinone profiles of the Legwnellaceae as determined by RPTLC are a stable and constant feature which can be used to aid in species identification. Our results for 0 Legionella species are compared with those of two other groups, who used HPLC, in Table. One group (Collins & Gilbart 98; Gilbart & Collins, 985) expressed their HPLC peaks of ubiquinones as percentages of the total ubiquinone content. The other group (Lambert & Moss, 989) reported their ubiquinone data based on a visually estimated scale of dependent on the relative concenations of ubiquinones observed. s the method of presentation between these workers differed, comparison of results was achieved by applying the simple method of grouping as described in this paper. Where only the predominant ubiquinones were used to group the legionellae, the method of RPTLC compared ownloaded from by IP: On: Sun, ec 07 09::07

7 Ubiquinone profiles of 5 Legionella species 0 Table. summary of the groups into which 0 Legionella species fall when analysed by three groups of workers using HPLC or RPTLC to determine ubiquinone profiles Group, predominates; group, predominates; group C, Q predominates; group, Q0 predominates; group, more than one ubiquinone predominates in the overall profile of a species. Groups obtained with results of: Collins & Gilbart Lambert & Moss (98, 985) ( 989) This paper Species (HPLC) (HPLC) (RPTLC) L. bozemanii L. dumfii L. gormanii L. jordanis L. longbeachae L. micdadei L. oakridgensis L. sainthelensi L. wadsworthii well with the results produced by HPLC. However, some differences were observed between the quantities of ubiquinones of some legionellae both between RPTLC and HPLC results and also between HPLC results where these had been carried out in different laboratories. We observed greater amounts of Q0 in three of the 0 species in Table L. jordanis, L. longbeachae and L. rnicdadei) than reported by other workers. The reason for this is unclear but the possibility that it could be due to a lipid component of RF value similar to Q0 warrants further investigation, including mass specomey. ifferences between analytical methods may account for other minor differences noted between laboratories. RPTLC, like HPLC, when employed to analyse the composition of ubiquinones cannot be used alone to give a definitive identification of a Legionella to species level, but is most useful when used in conjunction with serology, analysis of fatty acids by GLC, autofluorescence, N resiction fragment length polymorphisms (Harrison et al., 988), and most definitively with N homology (renner et al., 988). We wish to express our sincere thanks to Mr R. Wait, PHLS, Porton own, Salisbury, for mass specomey analyses and to Professor J. H. Freer, epartment of Microbiology, University of Glasgow, for helpful discussion and for critical appraisal of this manuscript. This work was supported by the Scottish Home and Health epartment under a grant for the New evelopments in Health Care. References, K., ISHISHI, M., OHM, M., KW, K. & KTSUI, G. (978). etermination of ubiquinone in serum and liver by high speed liquid chromatography. Journal of Nuitional Science and Vitaminology, ORNSTIN, N., MRMT,., SURGOT, M., NOWICKI, M., MUGNIR, H., FLURIT, J., GRON,., GRIMONT, F., GRIMONT, P..., TUCKR, W. L., NSON, R. F. & RNNR,. J. (989). Legwnella gratiana sp. nov. isolated from French spa water. Research in Microbiology 0, 555. RNNR,. J., STIGRWLT,. G., PPL, P., I, W. F., MCKINNY, R. M., STRNS, R. W., COLVILL, J. M., SLNR, R. K., LSTIN, P. H. & Moss, C. W. (988). Legwnella pneumophila serogroup Lansing isolated from a patient with fatal pneumonia, and descriptions of L. pmmphila subsp. pneumphila subsp. nov., subsp. fraseri subsp. nov., and L. pneumophila subsp. pascullei subsp. nov. Journal of Clinical Microbiology 6, COLLINS, M.. & GILRT, J. (98). New members of the coenzyme Q series from the Legionellaceae. FMS Microbiology Letters 6, 555. COLLINS, M.. &JONS,. (98). note on the separation of natural mixtures of bacterial ubiquinones using reversephase partition thin layer chromatography and high performance chromatography. Journal of pplied acteriology 5, 9. COLLINS, M.., PIROUZ, T., GOOFLLOW, M. & MINNIKIN,.. (977). isibution of menaquinones in actinomycetes and corynebacteria. Journal of General Microbiology 00, 0. LSTIN, P. H. (98). Improved semiselective medium for isolation of Legionella pneumophila from contaminated clinical and environmental specimens. Journal of Clinical Microbiology, 980. GILRT, J. & COLLINS, M.. (985). High performance liquid chromatographic analysis of ubiquinones from new Legionella species. FMS Microbiology Letters 6, 778. HRRISON, T. G., SUNRS, N.., OSHI, N. & WIT, R. (988). Serological diversity within the species Legionella spiritensis. Journal of pplied acteriology 65, 5. KRR,.., I, W. F. & Moss, C. W. (98). Isoprenoid quinones of the genus Legwnella. Journal of Clinical Microbiology, 008. KRIVNKOV, L. & K, V. (980). Semimicro exaction of ubiquinones and menaquinones from bacteria. Methods in nzymology 67,. LMRT, M.. & Moss, W. C.(989). Cellular fatty acid compositions and isoprenoid quinone contents of Legionella species. Journal of Clinical Microbiology 7, 657. Moss, C. W., I, W. F., KRR,.., GURRNT, G. 0. &LMRT, M.. (98). Cellular fatty acid composition and ubiquinone content of Legionella feeleii sp. nov. Journal of Clinical Microbiology 8, ownloaded from by IP: On: Sun, ec 07 09::07