Molecular cloning and phylogenetic analysis of the smallcytoplasmic RNA from Listeria monocytogenes

Tom Barry a *, Mary Kelly a, Barry Glynn a, John Peden b

a Molecular Microbiology Laboratory, Department of Microbiology, National University of Ireland, Galway, Irelandb Department of Genetics, Queens Medical Centre, University of Nottingham, Nottingham, UK

Received 11 January 1999; accepted 27 January 1999

Abstract

A molecular cloning strategy has been designed to isolate the gene that encodes the small cytoplasmic RNA (scRNA)component of bacterial signal recognition particles. Using this strategy a putative Listeria monocytogenes scRNA Vgt11recombinant clone was isolated. A previously described complementation assay developed to genetically select functionalhomologues of 4.5S RNA and scRNA of bacteria confirmed that the Vgt11 recombinant clone isolated encoded for the scRNAfrom L. monocytogenes. A secondary structure for this scRNA is proposed and a phylogenetic comparison of the 276 baseL. monocytogenes scRNA with previously characterised Gram-positive bacterial scRNAs is also presented. z 1999 Feder-ation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords: Small cytoplasmic RNA; Cloning strategy; Listeria monocytogenes ; Phylogenetic analysis

1. Introduction

Signal recognition particles (SRP) are ribonucleo-proteins that have been identi¢ed in Eukaryotae,Archaea and Eubacteria cells. In bacteria, SRP ribo-nucleoproteins may act as chaperones which are spe-ci¢c for signal sequences in nascent membranebound pre-proteins and as such are employed inmaintaining the pre-protein in a translocation-com-petent conformation prior to cellular secretion [1].Analysis of bacterial SRP associated RNAs has dem-onstrated that they all share a highly homologouscentral structural motif, helix 8, there is, however,

considerable sequence and structural heterogeneityassociated throughout the remainder of these RNAspecies [2]. SRPs associated RNAs from the Gram-negative bacteria are termed 4.5S RNAs and havebeen characterised from Escherichia coli [3], Thermusthermophilus [4], Legionella pneumophilia and Pseu-domonas aeruginosa [5]. Characterisation of SRP-as-sociated RNAs from Gram-positive bacteria hasbeen con¢ned to three Mycoplasma species [6^8] Mi-crococcus luteus [5], Clostridium perfringens [9] and13 Paenibacillus, Brevibacillus and Bacillus species[10]. The Mycoplasma species 4.5S RNAs, in com-mon with the Gram-negative species examined, com-prise of helix 8 and a partial helix 5, while the Micro-coccus SRP associated RNA comprise of helix 8 andan extended helix 5. To date, the only bacterial SRPassociated RNAs that contain, in common with the

0378-1097 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 3 7 8 - 1 0 9 7 ( 9 9 ) 0 0 0 5 5 - 5

FEMSLE 8654 17-3-99

* Corresponding author. Tel. : +353 (91) 524411;Fax: +353 (91) 525700; E-mail: thomas.barry@nuigalway.ie

FEMS Microbiology Letters 173 (1999) 47^53

Archaea and Eukaryotae SRP associated RNAs,helices 1, 2, 3, 4, 5 and 8, are the scRNAs from C.perfringens and the endospore-forming genus Bacil-lus. We sought to expand the range of scRNAs char-acterised from Gram-positive bacteria, by developinga general molecular cloning based strategy for thispurpose. The species of choice for this analysis wasListeria monocytogenes a non-spore forming Gram-positive rod shaped bacterium. It was selected pri-marily because of the phylogenetic relatedness to theBacillus species [11]. Characterisation of the SRP as-sociated RNA from L. monocytogenes could resolvewhether the additional structural features of Bacillusspecies and C. perfringens scRNAs are wide spreadamongst Gram-positive bacteria.

2. Materials and methods

2.1. Bacterial strains, plasmids, DNA and RNAmanipulations

L. monocytogenes serotype 1/2b was obtained fromNCIMB. E. coli S1610 and plasmids pTUBE809 andpTUBE822 were used for scRNA complementationanalysis [5,12]. L. monocytogenes was cultured in Lis-teria broth (3% Tryptone soya broth, 0.6% yeast ex-tract) at 30³C. Other E. coli cloning hosts were usedas recommended by the manufacturers of the Vgt11cDNA cloning kit (Amersham). Chromosomal DNAand total RNA preparations were carried out as pre-viously described [13,14].

2.2. Molecular cloning strategy

The initial aim of this cloning strategy was to iso-late and characterise a partial nucleotide sequence ofthe L. monocytogenes scRNA gene for use as a spe-ci¢c oligonucleotide DNA probe to screen a L.monocytogenes EcoRI Vgt11 genomic library con-structed in E. coli Y1090 (data not presented). Align-ment of known scRNA sequences using the CLUS-TAL W alignment program [15], indicated thatpolymerase chain reaction (PCR) oligonucleotideprimers could be designed which would amplify ashort variable nucleotide sequence of scRNA genes,approximating to the nucleotide sequence spanninghelices 5 and 8. Subsequently, sequence information

derived from this short ampli¢ed variable regionwould then be used to design a speci¢c oligonucleo-tide probe for use in screening the L. monocytogenesVgt11 library to identify putative scRNA gene re-combinants.

A 17-mer 5P degenerate oligonucleotide PCR pri-mer A (5P-T/G G/T TTGG T/G TC T/C C/T C/GCGCAA-3P), corresponding to helix 5 of 4.5S/scRNA genes, and an 18-mer 3P homologous oligo-nucleotide PCR primer B (5P-TCGAAGGAAGGC-CTGGAC-3P), corresponding to the conserved se-quence motif of helix 8 were synthesised andpurchased from Genosys. PCR ampli¢cation wascarried out using 250 ng of L. monocytogenes ge-nomic DNA, with ampli¢cation conditions set at94³C for 30 s, 50³C for 30 s and 72³C for 30 s inthe presence of 2.5 mM MgCl2, 250 pmol of eachPCR primer, 1 mM dNTPs, 1UPCR reaction bu¡erand 2.5 U of Taq DNA polymerase (Boehringer) toa ¢nal volume of 100 Wl in a Perkin-Elmer thermo-cycler. The 75-bp ampli¢ed PCR product was puri-¢ed from a 4% NuSieve (FMC) agarose gel and se-quenced directly using the homologous 3P PCRprimer B (data not shown). Sequence data indicatedthat a speci¢c DNA oligonucleotide probe could begenerated for use in screening for scRNA recombi-nants from the L. monocytogenes Vgt11 library andthat the possibility of non-speci¢c cross hybridisa-tion with E. coli, (the library host used), 4.5S RNAsequences was eliminated. The 18-mer L. monocyto-genes speci¢c scRNA oligonucleotide C, correspond-ing to the short variable region spanning helices 5and 8, 5P-TGGGAACCTGTGAACCAT-3P and itscomplement D, were synthesised. Oligonucleotide Cwas then [Q-32P]ATP (Amersham) 5P-end-labelledwith T4 polynucleotide kinase (Promega) and usedas an oligonucleotide probe to screen 3000 recombi-nants of the L. monocytogenes Vgt11 library. Hybrid-isation conditions for library screening was as fol-lows. Nytran ¢lters were prehybridised for 2 h in100 ml of 6USSC, 10UDenhardts and 0.1% SDS.The prehybridisation solution was removed and hy-bridisation was carried out in 5 ml of 6USSPE, 0.1%SDS with 10 ng ml31 of end labelled oligonucleotideprobe (107 cpm) at 50³C for 2 h. Three washes (100ml each) were then carried out in 6USSC and 0.1%SDS at room temperature for 5 min, with a ¢nalwash (100 ml) in the same solution at the hybrid-

FEMSLE 8654 17-3-99

T. Barry et al. / FEMS Microbiology Letters 173 (1999) 47^5348

isation temperature for 2 min. Autoradiography wascarried out for 48 h at 370³C.

2.3. Nucleotide sequence characterisation of theputative L. monocytogenes scRNA Vgt11recombinant clone

Small-scale V preparations were carried out on iso-lated recombinant clones [13] and EcoRI restrictionanalysis determined the molecular weight of thecloned inserts from these recombinants. Recombi-nant inserts were subsequently gel puri¢ed from2.5% NuSieve agarose gels and direct sequencing ofthe putative L. monocytogenes scRNA Vgt11 cloneswere then carried out using oligonucleotide C forsequencing in the 5P^3P direction and D in the 3P^5P direction.

Mapping of the 5P-end of the mature scRNA wascarried out using oligonucleotide B as a reverse tran-scriptase (BRL) extension primer [16]. The cDNAreaction products were analysed on a 6% DNA se-quencing gel. The length of the extended DNA frag-ment was estimated by comparison with sequencingladders from an M13 control transcript. 3P-end map-ping was determined by comparative sequence anal-ysis of CLUSTAL W multiple-sequence alignmentsof Bacillus, Brevibacillus and Paenibacillus speciesscRNAs.

2.4. Complementation and genetic selection analysis

The putative L. monocytogenes scRNA nucleotidesequence was PCR ampli¢ed from the Vgt11 re-combinant clone using PCR primers sc276.1 (5P-GT-TGATGAGCGTGAAGCC-3P), corresponding tothe 5P-end of the gene and sc276.2, (5P-TTAGTGT-CGCGCACCTCA-3P), corresponding to the 3P-endof the gene. The ampli¢ed product was then sub-cloned into the SmaI site of the E. coli^B. subtilisshuttle vector, pTUBE809, and the resultant con-struct, pTUBE922 was subjected to the in vivo com-plementation assay system used to genetically select4.5S/scRNA homologues of bacteria [5]. Essentially,this procedure consists of an in vivo complementa-tion assay system dependent on the presence of aplasmid borne 4.5S/scRNA to complement a 4.5Schromosomal defect in the strain E. coli S1610. Inthis strain, the sole intact copy of the gene for 4.5S

RNA is present on a thermoinducible prophage, ren-dering the bacterium temperature sensitive forgrowth. Since 4.5S RNA is essential for growth,the cured progeny of this strain are non-viable unlesstransformed with a plasmid encoding a 4.5S homo-logue.

2.5. Computer analysis and secondary structureprediction

Secondary structure prediction of the L. monocy-togenes scRNA was achieved by comparative analy-sis of multiple-sequence alignments with the secon-dary structure predictions of previously characterised4.5S RNA and scRNA sequences. All multiple align-ments were calculated by CLUSTAL W, the secon-dary structure for scRNA was drawn using the sec-ondary structure drawing program Loop-D-loop(written by D. Gilbert, available from ftp:\\ftp.bio.indiana.edu\molbio\loopdloop). Phylogenetic treeswere constructed using the Neighbor-Joining method[17], as implemented by CLUSTAL W, and Maxi-mum parsimony as implemented by the GCG 9.1PAUP program, the reliability of di¡erent phyloge-netic trees was estimated using bootstrapping [18].

3. Results and discussion

3.1. Characterisation of the L. monocytogenesscRNA gene

A putative L. monocytogenes scRNA Vgt11 clonewas identi¢ed and isolated after three rounds of se-lective hybridisation screenings with oligonucleotideC. Small-scale V preparation and restriction analysisrevealed a recombinant insert of 2.2 kb. Direct se-quencing with oligonucleotide C generated 200 bp ofnucleotide sequence data, while direct sequencingwith oligonucleotide D generated nucleotide se-quence data of 175 bp (data not shown). 5P-Primerextension analysis identi¢ed the initial 5P-nucleotideof the mature putative scRNA, with a major singlecDNA product observed from RNA isolated fromtwo points on the L. monocytogenes growth curve(Fig. 1). This partial sequence data was aligned usingCLUSTAL W with scRNAs previously characterisedfrom 13 Bacillus, Brevibacillus and Paenibacillus spe-

FEMSLE 8654 17-3-99

T. Barry et al. / FEMS Microbiology Letters 173 (1999) 47^53 49

cies. Comparative sequence analysis of this align-ment revealed that the putative L. monocytogenesscRNA was probably 276 bases in length, with anoverall sequence identity with the other scRNAs se-quences that ranged between 54 and 65%. This in-creased to 86% over the region homologous to 4.5SRNAs (i.e. bases 110^230).

The secondary structure of the putative L. mono-cytogenes scRNA demonstrates the presence of heli-ces 1^4. Compared with the scRNA of B. subtilis,helices 1, 2, 4 have fewer base pairing and the sec-ondary interaction between 3 and 4 is weaker (Fig.2).

Con¢rmation that recombinant gene isolated en-coded for the scRNA of L. monocytogenes was car-ried out by subjecting the putative scRNA gene se-quence to the genetic selection procedure. PlasmidspTUBE809 a negative control, pTUBE822 a positivecontrol, a derivative of pTUBE809 containing the

FEMSLE 8654 17-3-99

CFig. 2. Proposed secondary structure of the L. monocytogenesscRNA as deduced from comparative CLUSTAL W alignments.Uppercase bases are those conserved with B. subtilis scRNA,lower case bases are those unique to the L. monocytogenesscRNA. Helices are labelled with Arabic numerals (1, 2, 3, 4, 5and 8), and dashed lines infer possible tertiary interactions be-tween helices 3 and 4.

Fig. 1. Determination of the 5P-end of the L. monocytogenesscRNA by primer extension. Total RNA was prepared frommid-exponential phase, 4.5 h after inoculation (lane 1), and pre-stationary phase, 7 h after inoculation (lane 2) and subjected toprimer extension analysis. The length of the extended cDNAfragments were estimated by comparison with sequencing laddersfrom an M13 control transcript. The 177 base cDNA extensionproducts are arrowed.

T. Barry et al. / FEMS Microbiology Letters 173 (1999) 47^5350

scRNA gene of B. subtilis and pTUBE922 containingthe putative L. monocytogenes scRNA gene weretransformed into E. coli S1610 and subjected to ge-netic selection. No survivors were detected upontemperature pulse curing at 42³C of E. coli S1610transformed with pTUBE809, while approximately106 CFU ml31 heat-resistant survivors were obtainedwhen E. coli S1610 was transformed, and subse-quently pulsed cured at 42³C, with either pTUBE822or pTUBE922.

3.2. Phylogenetic analysis of the L. monocytogenesscRNA

The heterogeneous phylogenetic nature of the ge-nus Bacillus has been well documented [19]. Analysisof protein coding sequences indicates that even the

closely related species B. amyloliquefaciens and B.subtilis have 22% divergence in sequence identity[20]. A phylogenetic tree estimated from the 16SrRNA sequences of those Bacillus species whosescRNAs sequences have also been determined is pre-sented in (Fig. 3a). The topology is essentially thesame as that reported in a much more extensive phy-logenetic analysis of the genus Bacillus [21], with theexception that the 16S rRNA sequence of L. mono-cytogenes is included and that since the original phy-logenetic analysis of the genus Bacillus [21], the spe-cies Bacillus polymyxa, Bacillus macerans andBacillus brevis have been reclassi¢ed as Paenibacilluspolymyxa, Paenibacillus macerans and Brevibacillusbrevis, respectively. Analysis of available 23S rRNAsequences data supports this clustering pro¢le (datanot shown).

FEMSLE 8654 17-3-99

Fig. 3. Phylogenetic trees based on (a) 16S rRNA and (b) scRNA of L. monocytogenes and selected Bacillus species. Phylogenetic treeswere constructed using the Neighbor-Joining method as implemented by the CLUSTAL W program, gaps in alignments were excluded be-cause of the high divergences, corrections were not made for multiple substitutions and con¢dence values for individual branches were ob-tained by bootstrap analysis, in which 1000 bootstrap trees were generated from resampled data. The trees are unrooted and the distancebetween two species is obtained by summing connecting branch lengths, using the relevant distance scale. The percentage of the numberof bootstraps out of 1000 replications, that support a phylogenetic group of more than 85%, is placed beside the relevant branch.

T. Barry et al. / FEMS Microbiology Letters 173 (1999) 47^53 51

It has been reported that the genus Bacillus can besubdivided into groups on the basis of comparativephylogenetic 16S rRNA sequence analysis [21]. B.subtilis, B. amyloliquefaciens and B. pumilus areplaced in group 1, with B. sphaericus in group 2,and B. stearothermophilus in group 5. The L. mono-cytogenes 16S rRNA sequence is shown to representa separate branch equidistant from the three Bacillusgroups, Brevibacillus brevis and Paenibacillus species.Surprisingly, a similar phylogenetic analysis ofscRNA sequences produced a very di¡erent topology(Fig. 3b), with species, P. polymyxa, P. macerans, B.sphaericus, B. subtilis, B. amyloliquefaciens and B.pumilus clustering together. This topology was sup-ported by a boot-strap analysis.

To determine whether the scRNA or 16S rRNAphylogenetic trees best represented the true speciestree, phylogenetic trees of sequences with homo-logues in at least four species that included eitherP. polymyxa, P. macerans, or B. sphaericus were con-structed. The sequences that met these conditionswere the hyper variable region of the 23S rRNA,spoIIA, endo-L-1,3-1,4 glucanase, serine proteases,cytosine speci¢c methylases, RNA polymerase sigmafactors, spoA, bsuRI and cdgT. None of these phy-logenetic trees supported the scRNA topology inpreference to the 16S rRNA topology.

The P. polymyxa, P. macerans, B. sphaericus andB. subtilis scRNA sequences are obviously clusteringtogether because of sequence identity. Why thescRNA from these otherwise quite distantly relatedspecies have such a high sequence identity is notclear. This sequence identity is unlikely to be the

result of convergent evolution it is far more likelythat is the result the horizontal transfer of thesescRNAs between these species.

The most divergent scRNAs are L. monocytogenesand B. brevis (Fig. 3b). As the branch lengths havenot been corrected for superimposed substitutions,this represents a conservative estimation of evolu-tionary distance. The scRNA of L. monocytogenesand these other Gram-positive species can be consid-ered as having a region which is homologous to theE. coli 4.5S RNA (HEc4.5) and those regions whichare not (nHEc4.5). Divergence values between theseregions are outlined in Table 1. These divergencesindicate that it is the nHEc4.5 regions of L. mono-cytogenes and B. brevis scRNA that account for theirhigh sequence divergence. The high sequence diver-gence observed in the nHEc4.5 regions is indicativethat the L. monocytogenes and B. brevis helices 1^4are under a lower evolutionary constraint on se-quence conservation than the other scRNAs. Thecombination of a low evolutionary constraint andthe less frequent base-pairing in the secondary struc-ture of L. monocytogenes for helices 1^4 would seemto imply that these helices may have a biologicallydiminished functional role in L. monocytogenes rela-tive to the corresponding helices of the otherscRNAs. The base-pairing in the secondary structureof the equivalent B. brevis scRNA helices is muchmore like the other Bacillus species [10]. When weconsider that L. monocytogenes is the only one ofthese species that is non-sporulating, it would sup-port a hypothesis that the scRNA helices 1, 2, 3, 4and 5 have a possible biological role in the ability ofthe other species and C. perfringens to sporulate[9,12], but it does not adequately explain the highevolutionary rate of the equivalent helices of the B.brevis scRNA.

3.3. Concluding remarks

We have demonstrated that the scRNA helices 1,2, 3, and 4 which had only previously been identi¢edwithin the spore-forming Bacillaceae and C. perfrin-gens are also present in the non-spore forming L.monocytogenes. Genetic complementation and dele-tion experimentation of L. monocytogenes and otherphylogenetic related scRNAs with the inducer de-pendent scRNA gene of B. subtilis strain SC200NA

FEMSLE 8654 17-3-99

Table 1Comparison of the divergence percentage values for scRNAs ofBacillus species, B. brevis and L. monocytogenes, (a) between theregion that is homologous to E. coli 4.5S RNA (bases 110^230L. monocytogenes ^ HEc4.5) and (b) the remaining part of theirscRNA (bases 1^109, 231^276 L. monocytogenes ^ nHEc4.5)

(a)Bacillus species 0^24%B. brevis 19^24%L. monocytogenes 14^20%

(b)Bacillus species 0^20%B. brevis 19^38%L. monocytogenes 38^41%

T. Barry et al. / FEMS Microbiology Letters 173 (1999) 47^5352

[12], will be useful in determining discrete functionalbiological sequence motifs of these scRNAs.

3.4. Nucleotide sequence accession number

The L. monocytogenes scRNA has been submittedto the signal recognition particle database and Gen-Bank (accession number U15684).

Acknowledgments

We thank P. Sharp for the use of his laboratoryand helpful comments and C. Zwieb and N. Larsenfor signal recognition particle associated RNA align-ment analysis. We also thank S. Browne for E. coliS1610 and K. Nakamura for plasmids pTUBE809and pTUBE822. J.P. is funded by a studentshipfrom the University of Nottingham, B.G. andM.K. are funded by a studentship from EnterpriseIreland.

References

[1] Luirink, J. and Dobberstein, B. (1994) Mammalian and E. colisignal recognition particles. Mol. Microbiol. 11, 9^13.

[2] Zweib, C. and Larsen, A. (1994) The signal recognition par-ticle database (SRPDB). NAR 22, 3483^3484.

[3] Hsu, L.M., Zagorski, J. and Fournier, M.J. (1984) Cloningand sequence analysis of the E. coli 4.5S RNA gene. J. Mol.Biol. 178, 509^531.

[4] Struck, J.C.R., Toschka, H.Y. and Erdmann, V.A. (1988)Nucleotide sequence of the 4.5S RNA gene from T. thermo-philus HB8. NAR 16, 9042.

[5] Brown, S., Thon, G. and Tolentino, E. (1989) Genetic selec-tion and DNA sequences of 4.5S RNA homologues. J. Bac-teriol. 171, 6517^6520.

[6] Samuelsson, T. and Guindy, Y. (1990) Nucleotide sequence ofa M. mycoides RNA which is homologous to E. coli 4.5SRNA. NAR 18, 4938.

[7] Simoneau, P. and Hu, P.C. (1992) The gene for a 4.5S homo-log from M. pneumoniae : genetic selection, sequence, andtranscription analysis. J. Bacteriol. 174, 627^629.

[8] Ushida, C. and Muto, A. (1993) Small stable RNAs in M.capricolum. Nucleic Acids Symp. Ser. 157^158.

[9] Nakamura, K., Hasizume, E., Shibata, T., Nakamura, Y.,Mala, S. and Yamane, K. (1995) Small cytoplasmic RNA(scRNA) gene from Clostridium perfringens can replace thegene for Bacillus subtilis scRNA in both growth and sporula-tion. Microbiology 141, 2965^2975.

[10] Nakamura, K., Nishiguchi, M., Minemura, M., Honda, K.and Yamane, K. (1994) Small cytoplasmic RNA of B. brevis :transcriptional and phylogenetic analysis. Microbiology 140,493^498.

[11] Olsen, G.J., Woese, C.R. and Overbeek, R. (1994) The windsof (evolutionary) change: breathing new life into microbiol-ogy. J. Bacteriol. 176, 1^6.

[12] Nishiguchi, M., Honda, K., Amikura, R., Nakamra, K. andYamane, K. (1994) Structural requirements of B. subtilis smallcytoplasmic RNA for cell growth, sporulation and extracellu-lar enzyme production. J. Bacteriol. 176, 157^165.

[13] Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) MolecularCloning, A Laboratory Manual, 2nd edn., Cold Spring Har-bor Laboratory Press, Cold Spring Habor, NY.

[14] Barry, T., Geary, S., Hannify, S., MacGearailt, C., Shalloo,M., Heery, D., Gannon, F. and Powell, R. (1992) A rapidmethod for the isolation of total cellular RNA from micro-organisms. NAR 20, 4940.

[15] Thompson, J.D., Higgins, D. and Gibson, T.J. (1995) ClustalW: improving the sensitivity of progressive multiple sequencealignment through sequencing weighting, position gap penal-ties and weight matrix choice. NAR 22, 4673^4680.

[16] McKnight, S.L. and Kingsbury, R. (1982) Transcription con-trol of a eukaryotic protein-coding gene. Science 217, 316^324.

[17] Saitou, N. and Nei, M. (1987) The neighbour joining method;a new method for reconstructing phylogenetic trees. Mol.Biol. Evol. 4, 406^425.

[18] Felsenstein, J. (1985) Con¢dence limits on phylogenies: anapproach using the bootstrap. Evolution 39, 783^791.

[19] Woese, C.R. (1987) Bacterial evolution. Microbiol. Rev. 51,221^271.

[20] Sharp, P.M., Nolan, N.C. and Devine, K.M. (1995) Evolutionof gene sequences between and within species of Bacillus. In:Population Genetics of Bacteria (Baumberg, S., Young,J.P.W., Wellington, E.M.H. and Saunders, J.R., Eds.),SGM Symposium volume, Cambridge University Press, Cam-bridge.

[21] Ash, C., Farrow, A.E., Wallbanks, S. and Collins, M.D.(1991) Phylogenetic heterogeneity of the genus Bacillus re-vealed by comparative analysis of small-subunit ribosomalRNA sequences. Lett. Appl. Microbiol. 13, 202^206.

T. Barry et al. / FEMS Microbiology Letters 173 (1999) 47^53 53