School of Biological Sciences, University of Sussex, Brighton BN1 9QG, UK1
Trafford Centre for Medical Research, University of Sussex, Brighton BN1 9RY, UK2
Wellcome Trust Centre for the Epidemiology of Infectious Disease, Department of Zoology, University of Oxford, Oxford OX1 3PS, UK 3
Author for correspondence: Lucas D. Bowler. Tel: +44 1273 678309. Fax: +44 1273 678433. e-mail: bafn7{at}sussex.ac.uk
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Keywords: cDNA representational difference analysis, differential gene expression, RNA isolation, iron regulation, Neisseria meningitidis
Abbreviations: cDNA RDA, representational difference analysis of cDNA; PAS, p-aminosalicylic acid; TNS, triisopropylnaphthalene sulphonic acid
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Identifying such differential gene expression is a frequent goal in modern biomedical research, and as a consequence, a variety of differential screening methods have been developed over the last few years with varying degrees of success (Soares, 1997 ; Sagerström et al., 1997
). Many of these techniques require large quantities of mRNA and are relatively insensitive, although a few take advantage of PCR to amplify messages to detectable levels. These include differential display PCR (DD-PCR), the conceptually similar technique, RNA fingerprinting by arbitrarily primed PCR (AP-PCR), and representational difference analysis of cDNA (cDNA RDA). cDNA RDA is a cDNA-specific modification of the PCR-based subtractive hybridization technique of representational difference analysis first described for application to whole genomes by Lisitsyn et al. (1993)
. cDNA RDA employs a positive selection approach in which target cDNA fragments are sequentially enriched by favourable hybridization kinetics and subsequently amplified by PCR. In contrast to DD- or AP-PCR, cDNA RDA has the major advantage that sequences common to both groups are eliminated, which greatly simplifies the interpretation of results and identification of the differentially expressed genes. In addition, the exponential degree of enrichment achieved by the use of PCR in cDNA RDA enables the detection of very rare transcripts (Hubank & Schatz, 1999
).
A number of microarray-based methods for the detection of differentially expressed genes have been described recently (De Saizieu et al., 1998 ; Ramsay, 1998
). These involve the gridding of PCR products or oligonucleotides corresponding to all of the genes of the organism under study onto a solid support so that the levels of expression of each gene under any growth condition can be monitored by hybridization with labelled RNA preparations. Microarray technology has the potential to revolutionize the study of differential gene expression but there are currently a number of drawbacks to this approach. Firstly, application of the methodology is limited to organisms whose genomes have been sequenced. Secondly, although the method has been recently used successfully for bacterial transcript imaging (De Saizieu et al., 1998
), it does not at present have the sensitivity of PCR-based techniques such as cDNA RDA.
Given these considerations, it is apparent that cDNA RDA, with its wide applicability and high sensitivity, is potentially one of the most powerful methods for the study of mechanisms of microbial pathogenesis currently available. The cDNA RDA technique is particularly useful as, in principle, it can in the same experiment analyse differences in gene expression in both the pathogen and host during different stages of their interaction. Furthermore, the sensitivity and specificity of the method should allow total (host plus pathogen) RNA preparations to be used, facilitating the rapid processing of samples without the need for the pathogen and host cells to be separated.
However, whilst the cDNA RDA method is proven with eukaryotic material (Hubank & Schatz, 1999 ), the use of cDNA RDA has yet to be reported with bacteria. This is primarily because cDNA RDA requires the isolation of high-quality representative mRNA on a reproducible basis from the organisms under study. This is particularly difficult with prokaryotes since bacterial messages can be very unstable, with mean half-lives of between 1·5 and 2·5 min at 37 °C (Alifano et al., 1994
), far less than that of eukaryotic mRNAs. Standard purification techniques using oligo(dT) affinity matrices are inapplicable to prokaryotes since bacterial mRNA effectively lack 3'-poly(A) tails, necessitating the use of total RNA preparations, in which most of the RNA will be ribosomal (rRNA). Bacterial cells are also rich in endogenous highly stable ribonucleases which are unaffected by the more commonly used proprietary RNase inhibitors. In this paper, we report development of the cDNA RDA system using total RNA preparations for the identification of differentially expressed genes in bacteria, and its successful application in a model system designed to identify iron-regulated genes in Neisseria meningitidis.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
RNA preparation.
Total RNA was isolated by a procedure adapted from the triisopropylnaphthalene sulphonic acid (TNS)/p-aminosalicylic acid (PAS) method of Felipe et al. (1993) . Work was carried out on ice where possible. Water was RNase-free, UHP quality (ELGA) and was monitored for RNase contamination using the RNaseAlert kit (Ambion). Bacterial cultures were harvested by centrifugation of 10 ml samples in Falcon tubes at 6000 r.p.m. for 5 min at 1 °C. Cell pellets were quickly resuspended in 1·5 ml cold fresh TNS/PAS buffer containing 2% TNS (Phase Separation), 6% PAS (Sigma), 1% SDS, 50 mM EDTA, 250 mM NaCl, 200 mM PIPES pH 6·0, 6% n-butanol, and 100 mM 2-mercaptoethanol. Acid phenol/chloroform (5:1, v/v, acid equilibrated to pH 4·7, 1·5 ml) and chloroform/isoamyl alcohol (49:1, v/v, 300 µl) were added, and the mixture was vortex-mixed for 3x10 s. Samples were spun at 16000 g for 5 min at 1 °C and the aqueous phase carefully transferred to new tubes. The aqueous phase was re-extracted with 1 vol. acid phenol/chloroform plus 0·2 vol. chloroform/isoamyl alcohol, and again with 1 vol. chloroform. RNA was precipitated by incubation with 2·5 M LiCl, 25 mM EDTA, pH 8·0 for at least 2 h at -20 °C. Pellets were washed twice with 70% ethanol and resuspended to 1 mg ml-1 in 3 mM CaCl2, 6 mM MgCl2, 50 mM TES pH 7·5, plus 500 U RNase block ml-1 (Stratagene). Contaminating DNA was removed by incubation with RNase-free DNase I (Ambion) for 45 min at 37 °C. The reaction was terminated by extraction with 1 vol. acid phenol followed by a further extraction with 1 vol. chloroform. RNA was then precipitated with 0·2 vol. 10 M ammonium acetate and 1 vol. 2-propanol for at least 2 h at -20 °C. Pellets were washed twice with 70% ethanol and resuspended to 1 mg ml -1 in water containing 0·1 mM EDTA plus 600 U RNase block ml-1. Quantification and crude quality assessment were done by measuring optical density of the preparation at 260 nm and 280 nm, and by examination on non- denaturing agarose gels (1%) in Tris/borate/EDTA (TBE) buffer pH 7·8 (this was found to give better, more consistent results with RNA than the usual TBE pH 8·4) stained with ethidium bromide. Aliquots of the total RNA preparations were stored at -80 °C until use.
cDNA synthesis.
cDNA synthesis was performed by random-priming of total RNA using the Pharmacia TimeSaver cDNA synthesis kit. The protocol was according to the manufacturers instructions except, after 20 min incubation of the first-strand reaction at 37 °C, 1 µl Superscript II reverse transcriptase (Stratagene) was added to the reaction mix and incubation was continued for a further 1 h. The second-strand incubation was at 12 °C for 1 h, and then 22 °C for a further 2 h.
Isolation of N. meningitidis rRNA genes and generation of rRNA representations.
Synthetic oligonucleotide primers, based on published meningococcal rRNA sequences, were used to amplify the 16S and 23S rRNA genes from N. meningitidis M986 chromosomal DNA isolated using the Genomic-Tip 100/G system (Qiagen). The primer pairs used were as follows: 16S rRNA, 16Sup (5'-CATAAGAGTTTGATCCTGGCT-3') and 16Sdn (5'-GTCATGAAGCATACCGTGGT-3'); 23S rRNA, 23Sup (5'-CAGGT/CGGATGCCTTGGCGA-3') and 23Sdn (5'- AGAGTCAAGCCTCACGA/GGCA-3'). The rRNA gene products were digested with DpnII and the R-Bgl-12/24 adapters ligated as described by Hubank & Schatz (1999) for the generation of cDNA RDA representations from cDNA. Eighteen cycles of PCR were used to produce the neisserial rRNA representations.
cDNA RDA.
The basic method used was as described by Hubank & Schatz (1994) , but with the following modifications. HPLC-purified oligonucleotides were used throughout. Driver and tester components were generated by PCR using 18 cycles of amplification of the R-Bgl- 12/24 adapter-ligated, DpnII-restricted, cDNA (R-Bgl-12, 5'-GATCTGCGGTGA-3'; R-Bgl-24, 5'- AGCACTCTCCAGCCTCTCACCGCA-3'). Digested/unligated oligonucleotide linkers (-Bgl-12/24 series) were separated from the tester when necessary using QIAquick Spin columns (Qiagen). The driver used comprised 50% N. meningitidis CE1402-derived representation and 25% each of representations derived from the 16S and 23S rRNA genes of N. meningitidis M986. Four rounds of subtractive hybridization and selective hybridization were used. The tester:driver ratios were 1:100 [difference product (DP) 1]; 1:800 (DP2); 1:5000 (DP3). A full protocol can be supplied on request.
Cloning and sequencing of difference products.
Final difference products were excised from agarose gels, cloned into the pGEM-T vector (Promega) and sequenced in both directions using M13 (-21) and M13 reverse primers and dRhodamine Terminator chemistry on an ABI 377 DNA sequencer. Resulting sequences were compared to the GenBank database and to the N. meningitidis (Sanger Centre, Cambridge, UK, and The Institute for Genome Research, Rockville, MD, USA) and Neisseria gonorrhoeae (University of Oklahoma, OK, USA) databases using the BLAST program (Altschul et al., 1990 ).
Southern blotting.
Representations were electrophoresed through 1% non-denaturing agarose gels in TBE buffer and transferred to nylon membrane (Hybond-N, Amersham), using the downward alkaline blotting method of Chomczynski (1992) . DNA probes were produced by PCR amplification using pairs of synthetic oligonucleotide primers directed to the 5' and 3' ends of the identified difference products. Non-radioactive labelling and detection were carried out using the Gene Images (Amersham) labelling and detection kit according to the manufacturers instructions.
RT-PCR.
All RNA preparations were pre-screened for residual DNA contamination by PCR using the primer pairs synthesized for the Southern blotting analysis. To confirm difference products as genuine, RT-PCR was performed using the Pharmacia Ready-To-Go RT-PCR bead system. One-step reactions (0·5 ml) were carried out using 200 ng total RNA for each reaction and the same primer pairs. To compare relative levels of expression, RT-PCR was carried out using an internal control to normalize the samples. The control used was the house-keeping gene glnA, whose expression is not iron- regulated. The primer pair for this purpose was GlnAup1165 (5'- CCCGGATGGATTTTGTTTTG-3') and GlnAdn600 (5'- CACAAAACCCGTTACGAAATCA-3'), which amplify a PCR product of approximately 750 bp. First-strand synthesis was carried out at 46 °C for 30 min. The semi-quantitative PCR amplification of the cDNA first strand was performed for 17 cycles.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Methods based on the homogenization of cellular material in guanidine.HCl or hot phenol have often proven to be unreliable (Groppe & Morse, 1993 ), and this was also found to be the case in this study. These techniques were therefore not considered further. However, relatively rapid and efficient techniques involving the use of more powerful chaotropic agents such as guanidine isothiocyanate (GITC) have been developed (Chirgwin et al., 1979
; Chomczynski & Sacchi, 1987
). Novel approaches using detergents such as Catrimox-14 (Dahle & Macfarlane, 1993
), TNS (Felipe et al., 1993
) and cetyltrimethylammonium bromide (CTAB) (Cheung et al., 1994
) have also been described, and apparently used for the production of high-quality, high-purity, intact mRNA. All of these newer methods were assessed for their ability to provide high-quality RNA from N. meningitidis. Published methods based on the use of GITC differ widely in the pH at which the lysis/extraction solutions are buffered (in the range pH 5·07·5). As it has been reported that bacterial RNA may be more stable at lower pH (Noonberg, 1995
), all isolation protocols were modified by the substitution of a variety of alternative buffering systems.
RNA derived using these methods was initially assessed qualitatively by non-denaturing agarose gel electrophoresis. Good-quality preparations were characterized by the presence of the dominant 16S and 23S rRNA species appearing as fairly sharp intense bands with apparent sizes of about 750 bp and just over 1000 bp, respectively, relative to double-stranded DNA molecular size markers (Fig. 1). In addition, a smaller more diffuse band comprising low-molecular-mass RNA should also be visible. Some diffuse smearing may be visible between 16S and 23S bands and a little below the 16S band, but there should be a clear region between this and the low-molecular-mass RNA band towards the bottom of the gel. Using these criteria, RNA isolation at pH 4·0, by whatever method, resulted in considerable degradation of the nucleic acid and a pH
8·0 also led to some degradation. Most of the various methods of isolation we used appeared to have pH optima of pH 5·56·5 and the yields from each were broadly comparable, although RNA preparations using Catrimox-14 were less efficient and those using GITC were slightly more efficient (data not shown).
|
This approach indicated that the RNA isolation method of Chomczynski & Sacchi (1987) and the TNS/ PAS protocol (Felipe et al., 1993
) carried out at pH 6·0 or pH 7·0 were the most promising. However, given the low sensitivity of Northern blot analysis, RT-PCR was used as another assay for the presence of mRNAs from a variety of genes. The transcripts targeted were of the iron-regulated lactoferrin-binding protein A (lbpA) (Pettersson et al., 1994
) and periplasmic iron-binding protein (fbpA) (Berish et al., 1990
) genes, both expressed at relatively high levels under iron limitation; the shikimate dehydrogenase (aroE ) and glutamine synthetase (glnA) genes (Zhou et al., 1997
), expressed at medium levels; and the penicillin- binding protein 2 (penA) (Spratt et al., 1989
) and rho transcription terminator (Opperman & Richardson, 1994
) genes, both expressed at relatively low levels.
Multiple total RNA preparations derived using each isolation method were tested by RT-PCR to check reproducibility and the integrity of the RNA transcripts. RNA prepared from N. meningitidis M986 using the TNS/PAS method at pH 6·0, with the addition of n- butanol to prevent PAS precipitation at low temperatures, was selected as the most suitable for use as the template for cDNA synthesis for cDNA RDA. Fig. 2 shows the results from RT-PCR screening of a total RNA preparation derived using the modified TNS/PAS method.
|
The cDNA RDA procedure
Examination of differential gene expression using cDNA RDA requires the sampling of a population of cells grown under the condition(s) of interest (in this study, under iron limitation the so-called tester component), and a population grown in exactly the same way but differing in the condition(s) of interest (in this case, with iron-sufficiency the driver component). mRNA is extracted from both cell populations and used as a template for cDNA synthesis. The derived cDNAs are digested with a frequently cutting restriction enzyme. Synthetic oligonucleotides are ligated to the ends of these restriction fragments. PCR with primers specific to these oligonucleotides is then used to amplify each of the DNA fragments in the different populations (Fig. 3a ). After amplification, the oligonucleotides are removed with the same restriction enzyme and the amplified cDNA fragments are purified (these preparations are referred to as representations). For the subsequent hybridization and amplification steps, different oligonucleotides are ligated only to the tester DNA fragments. Tester and driver DNAs (the latter in excess) are mixed, denatured and allowed to hybridize. PCR using primers specific to the new oligonucleotide extensions results in exponential amplification of sequences unique to the tester DNA population, whereas sequences present in both populations, or only in the driver population, will either be amplified linearly or not at all (Fig. 3b
). By repeating these steps, the degree of enrichment of sequences unique to the tester population is increased, and after several cycles of this procedure a number of discrete PCR products are obtained (the difference products) that represent the mRNAs unique to the tester population. These products can then be cloned and characterized.
|
Using these modifications (shown schematically in Fig. 3), two independent cDNA RDA experiments were performed. In both, the same pattern of five major distinct PCR products was obtained from the cDNA RDA of N. meningitidis, which ideally would represent mRNAs derived from the lbpA gene and from other meningococcal genes that are iron-regulated. The PCR products were cloned and the inserts in a number of the clones sequenced. Of 28 sequences from the first analysis (Fig. 4
, lane A), six (21%) were found to correspond to four different amplifiable fragments of the lbpA gene, and one (4%) to a region of the N. meningitidis RTX-like toxin gene ( frpC), which is also known to be iron-regulated. Fourteen of the remainder comprised unidentified meningococcal ORFs (50%) and seven (25%) were rRNA-derived sequences. The five amplified products from the second independent cDNA RDA experiment (Fig. 4
, lane B) also included the same four fragments from the lbpA gene and fragment of frpC.
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In this study we have, as expected, identified several difference products corresponding to regions of N. meningitidis lbpA, and in addition detected a fragment of the meningococcal RTX-like toxin gene, frpC, whose expression is known to be up-regulated under conditions of iron limitation (Thompson et al., 1993 ). Additional iron-regulated genes might have been identified by sequencing more inserts from the cloned amplified fragments since individual difference product bands produced by cDNA RDA often comprise many different fragments (M. Hubank, unpublished observation). A second cDNA RDA experiment (Fig. 4
, lane B) identified the same parts of the lbpA gene and region of frpC as genuine difference products, demonstrating the reproducibility of the technique.
It should be emphasized that this paper describes the successful development of the cDNA RDA technique for bacteria, and is not intended to be a thorough analysis of the iron-regulated genes of N. meningitidis. Indeed, most of the previously characterized neisserial iron-regulated genes were not detected (e.g. lbpB, frpB, fbpA and tbpA etc.). However, this was expected since the kinetics of the PCR-coupled subtractive hybridization in cDNA RDA preferentially enrich for DNA fragments in the 200600 bp range (as can be seen in Fig. 4 ). Thus, only those genes whose digestion with the chosen restriction enzyme results in fragments within this size range will amplify efficiently and be detected as difference products. For example, from published sequence data, we know that complete DpnII digestion of the lbpB gene from N. meningitidis M986 will not result in fragments of a suitable size to give amplifiable products using cDNA RDA. Thus, the failure to detect fragments of the lbpB gene, which was deleted together with the lbpA gene in strain CE1402 (Pettersson et. al., 1998
), was unsurprising. It is likely that this also explains the failure to find gene fragments from the periplasmic ferric-binding protein gene, fbpA, which, although not sequenced from M986, is highly conserved within both meningococci and N. gonorrhoeae (Zhou & Spratt, 1992
). Analysis of the available fbpA sequences indicates that DpnII digestion of this gene will also not result in amplifiable products under cDNA RDA. None of the other iron-regulated genes, which tend to be highly variable, have been sequenced from strain M986, which makes it difficult to assess whether they should have been recovered using DpnII. However, examination of the published meningococcal genome sequence shows that DpnII would cut on average every 600 bp or so, the upper limit for cDNA RDA, suggesting that this enzyme is not the most appropriate for optimal recovery of difference products with N. meningitidis. As a consequence of the size constraints on the DNA fragments that can be amplified by cDNA RDA, a thorough analysis to detect all of the genes that are differentially expressed under given conditions requires that cDNA RDA be carried out with a number of different frequently cutting restriction endonucleases.
The percentage of genuine difference products was relatively high (about 33%), but by using different proportions of rRNA representation in the driver, we believe that the level of rRNA contamination could be reduced further. The driver to tester ratios and/or number of subtractive hybridization/amplification cycles used could also be increased if the number of false positives is high, although it should be noted that this could lead to a failure to identify the more rare differences. In this study, no 5S-rRNA-derived sequences were detected. Nevertheless, given their abundance, it may also be useful to include a 5S rRNA component in driver 2 since in other circumstances restriction of 5S-rRNA-derived cDNA may result in amplifiable fragments. However, the use of an initial additional subtractive step using only rRNA-gene- derived driver was not successful and resulted in the identification of no genuine differences in gene expression. This may be because, although it resulted in increased selection against amplification of rRNA-derived sequences, all other cDNAs were amplified, heavily biasing the resulting cDNA population towards other abundant mRNA sequences common to each population. Subsequent rounds of PCR-coupled subtractive hybridization were then insufficient to isolate genuine difference products. Improved detection of up- or down-regulated gene expression by cDNA RDA can also be obtained using representations which have been depleted of low-copy sequences by a process known as melt depletion (Hubank & Schatz, 1994 ).
The cDNA RDA technique has great potential for studying differential gene expression during hostpathogen interactions during the course of infection, as total RNA prepared from the infected host tissue can be compared to that prepared from host and pathogen grown separately, but under the same environmental conditions. The very high sensitivity and specificity of the cDNA RDA approach should allow the genes switched on, or up- or down-regulated, in both a bacterial pathogen and its host, to be analysed in the same experiment. The demonstration that cDNA RDA can be successfully used with total RNA preparations is therefore important as infected cells can be rapidly processed without the need to separate bacterial cells from those of the host, greatly increasing the chances of identifying differentially expressed genes whose transcripts are short-lived. cDNA RDA should also be of great value in facilitating the application of microarray technology to the study of differential gene expression in bacteria since it could be used to produce probes of higher specificity, enriched for low-abundance signals.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Altschul, S. F. , Gish, W. , Miller, W. , Myers, E. W. & Lipman, D. J. (1990). Basic local alignment search tool. J Mol Biol 215, 403-410.[Medline]
Berish, S. A. , Mietzner, T. A. , Mayer, L. W. , Genco, C. A. , Holloway, B. P. & Morse, S. A. (1990). Molecular cloning and characterisation of the structural gene for the major iron-regulated protein expressed by Neisseria gonorrhoeae. J Exp Med 171, 1535-1546 .[Abstract]
Cheung, A. L. , Eberhardt, K. J. & Fischetti, V. A. (1994). A method to isolate RNA from gram-positive bacteria and mycobacteria. Anal Biochem 222, 511-514.[Medline]
Chirgwin, J. M. , Przybyla, A. E. , MacDonald, R. J. & Rutter, W. J. (1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18, 5294-5299 .[Medline]
Chomczynski, P. (1992). One-hour downward alkaline capillary transfer for blotting of DNA and RNA. Anal Biochem 201, 134-139.[Medline]
Chomczynski, P. & Sacchi, N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162, 156-159.[Medline]
Dahle, C. E. & Macfarlane, D. E. (1993). Isolation of RNA from cells in culture using Catrimox-14 cationic surfactant. Biotechniques 15, 1102-1105 .[Medline]
De Saizieu, A. , Certa, U. , Warrington, J. , Gray, C. , Keck, W. & Mous, J. (1998). Bacterial transcript imaging by hybridisation of total RNA to oligonucleotide arrays. Nature Biotechnol 16, 45-48.[Medline]
Duguid, J. R. , Rohwer, R. G. & Seed, B. (1988). Isolation of cDNAs of scrapie- modulated RNAs by subtractive hybridisation of a cDNA library. Proc Natl Acad Sci USA 85, 5738-5742 .[Abstract]
Farrell, R. E., Jr (1993). RNA Methodologies a Laboratory Guide for Isolation and Characterisation. San Diego: Academic Press.
Felipe, M. S. S. , Rogelin, R. , Azevedo, M. O. & Astolfi-Filho, S. (1993). Extended application of an efficient method for RNA isolation from different organisms. Biotechnol Tech 7, 639-644.
Groppe, J. C. & Morse, D. E. (1993). Isolation of full- length RNA templates for reverse transcription from tissue rich in RNase and proteoglycans. Anal Biochem 210, 337-343.[Medline]
Hubank, M. & Schatz, D. G. (1994). Identifying differences in mRNA expression by representational difference analysis of cDNA. Nucleic Acids Res 22, 5640-5648 .[Abstract]
Hubank, M & Shatz, D. G. (1999). cDNA representational difference analysis: a sensitive and flexible method for the identification of differentially expressed genes. Methods Enzymol 303, 325-349.[Medline]
Lisitsyn, N. , Lisitsyn, N. & Wigler, M. (1993). Cloning the differences between two complex genomes. Science 259, 946-951.[Medline]
Litwin, C. M. & Calderwood, S. B. (1993). Role of iron in regulation of virulence genes. Clin Microbiol Rev 6, 137-149.[Abstract]
Manley, J. L. (1995). Messenger RNA polyadenylation: a universal modification. Proc Natl Acad Sci USA 92, 1800-1801 .
Mekalanos, J. J. (1992). Environmental signals controlling expression of virulence determinants in bacteria. J Bacteriol 174, 1-7.[Medline]
Miller, J. F. , Mekalanos, J. J. & Falkow, S. (1989). Coordinate regulation and sensory transduction in the control of bacterial virulence. Science 243, 916-922.[Medline]
Noonberg, S. B. (1995). Effect of pH on RNA degradation during guanidinium extraction. BioTechniques 19, 731-733.[Medline]
ONeill, N. J. & Sinclair, A. H. (1997). Isolation of rare transcripts by representational difference analysis. Nucleic Acids Res 25, 2681-2682 .
Opperman, T. & Richardson, J. P. (1994). Phylogenetic analysis of sequences from diverse bacteria with homology to the Escherichia coli rho gene. J Bacteriol 176, 5033-5043 .[Abstract]
Pettersson, A. , Klarenbeek, V. , Van Deurzen, J. , Poolma, J. T. & Tommassen, J. (1994). Molecular characterization of the structural gene for the lactoferrin receptor of the meningococcal strain H44/76. Microb Pathog 17, 395-408.[Medline]
Pettersson, A. , Prinz, T. , Umar, A. , van der Biezen, J. & Tommassen, J. (1998). Molecular characterisation of LbpB, the second lactoferrin-binding protein of Neisseria meningitidis. Mol Microbiol 27, 599-610.[Medline]
Plum, G. & Clark-Curtiss, J. E. (1994). Induction of Mycobacterium avium gene expression following phagocytosis by human macrophages. Infect Immun 62, 476-483.[Abstract]
Ramsay, G. (1998). DNA chips: state of the art. Nature Biotechnol 16, 40-44.[Medline]
Sagerström, C. G. , Sun, B. I. & Sive, H. L. (1997). Subtractive cloning: past, present, and future. Annu Rev Biochem 66, 751-783.[Medline]
Shepard, A. R. & Rae, J. L. (1997). Magnetic bead capture of cDNAs from double-stranded plasmid cDNA libraries. Nucleic Acids Res 25, 3183-3185 .
Soares, M. B. (1997). Identification and cloning of differentially expressed genes. Curr Opin Biotechnol 8, 542-546.[Medline]
Spratt, B. G. , Zhang, Q.-Y. , Jones, D. M. , Hutchison, A. , Brannigan, J. A. & Dowson, C. G. (1989). Recruitment of a penicillin-binding protein gene from Neisseria flavescens during the emergence of penicillin resistance in Neisseria meningitidis. Proc Natl Acad Sci USA 86, 8988-8992 .[Abstract]
Swaroop, A. , Xu, J. , Agarwal, N. & Weissman, S. M. (1991). A simple and efficient cDNA library subtraction procedure: isolation of human retina- specific cDNA clones. Nucleic Acids Res 19, 1954.[Medline]
Thompson, S. A. , Wang, L. & Sparling, P. F. (1993). Cloning and nucleotide sequence of frpC, a second gene from Neisseria meningitidis encoding a protein similar to RTX cytotoxins. Mol Microbiol 9, 85-96.[Medline]
Zhou, J. & Spratt, B. G. (1992). Sequence diversity within the argF, fbp and recA genes of natural isolates of Neisseria meningitidis: interspecies recombination within the argF gene. Mol Microbiol 6, 2135-2146.[Medline]
Zhou, J. , Bowler, L. D. & Spratt, B. G. (1997). Interspecies recombination and phylogenetic distortion within the glutamine synthetase and shikimate dehydrogenase genes of Neisseria meningitidis and commensal Neisseria species. Mol Microbiol 23, 799-812.[Medline]
Received 7 June 1999;
revised 7 September 1999;
accepted 9 September 1999.