Science in China: Series C Life Sciences 2006 Vol.49 No.2 141—148
141
DOI: 10.1007/s11427-006-0141-3
Comparison of the virulence plasmid genomes of two strains of Shigella which lost the ability to bind Congo red XIONG Zhaohui1*, TANG Xudong1,2*, YANG Fan1, ZHANG Xiaobing1, YANG Jian1, CHEN Lihong1, NIE Huan1, YAN Yongliang1, JIANG Yan1, WANG Jing1, XUE Ying1, XU Xingye1, ZHU Yafang1, DONG Jie1, AN Lizhe2, WANG Xunling2 & JIN Qi1 1. State Key Laboratory for Molecular Virology and Genetic Engineering, Beijing 100052, China; 2. School of Life Science, Lanzhou University, Lanzhou, 743000) Correspondence should be addressed to Jin Qi (email:
[email protected])
Received January 19, 2005; accepted May 20, 2005
Abstract We determined and analyzed the Shigella flexneri serotype 5 (pSF5) and S. dysenteriae serotype 1 (pSD1) virulence plasmid genomes. The total length of pSF5 is 136513 bp, including 165 open reading frames (ORFs). Of these ORFs, 133 were identified and 32 of those had no significant homology to proteins with known functions. The length of pSD1 is 182545 bp, including 224 ORFs, of which we identified 181. The remaining 43 ORFs were not significantly homologous to proteins with known functions. The insertion sequence (IS) elements are 53787 bp in pSF5, and 49616 bp in pSD1, which represents 39.4% and 27.1% of the genome, respectively. There are 22 IS element types in pSF5 and pSD1, among which we report ISEc8 and ISSbo6 for the first time in the Shigella virulence plasmid. Compared to pCP301, there are a large number of deleted genes and gene inversions in both pSF5 and pSD1. The ipa-mxi-spa locus in pSF5 is completely absent, and the genes related to the O-antigen biosynthesis are partially missing. In contrast, the above genes in pSD1 are integral, with the exception of virF. The whole genome analysis of the two plasmids shows that the loss of genes related to gene invasion or regulation also obliterates the ability of pPF5 and pSD1 to bind Congo red (Crb). Whether these genes determine the Crb function requires continued investigation. Keywords: Shigella flexneir serotype 5, Shigella dysenteriae serotype 1, virulence plasmid, Congo red, genome comparison.
As an enteric pathogen and Gram negative bacterium, Shigella possesses high infectivity and leads to serious illness. Since its discovery in 1898 by Shiga, Shigella species have been studied widely. These studies have elucidated the Shigella pathogenicity mechanism, heredity and evolution characteristics and ― host relationship[1 4]. The pathogenicity of Shigella is * These authors contributed equally to this work.
www.scichina.com
www.springerlink.com
conferred by loci products encoded on a 220-kb virulence plasmid, and within the 30 kb region which encodes ipa-mxi-spa and the type III secretion apparatus as the core region of virulence in shigellosis. The genes encoded by the virulence plasmid 100 in the S. flexneri serotype 5, pMYSH6000 in the S. flexneri serotype 2a, pSS120 in S. sonneri and pCP301 in the
142
Science in China: Series C Life Sciences
S. flexneri serotype 2a have been well studied[5]. Also, Shigella has the ability to invade and bind Congo red dye, the loss of which induces a loss of virulence[6]. Generally, spontaneous mutation from Crb+ to Crb− occurs at a frequency of approximately 10−4[7] under natural conditions. Some experiments indicate that Shigella can regain the Congo red binding capacity upon culture with the lysate of the Crb+ bacteria or with an iron-containing medium. However, the capacity of cell invasion cannot be restored. Moreover, not all of the bacterium which can bind Congo red have the ability of invasion, and this kind of bacterium can bind Congo red both at 30 and 37℃, which was different from the wild bacteria which can only bind Congo red at 37℃[8]. Also, some experiments show that the glucose enzyme related to cell wall biosynthesis affects the Crb+ phenotype. For example, the O-antigen of S. flexnneri is related to the Ser test in guinea pigs, namely, the genes related to the membrane and cell wall also affect the binding of Shigella to Congo red. From the experiments described above, we hypothesize that the mechanism of Shigella binding Congo red involves the virulence genes and other modulated mechanisms. Indeed, the relationship between Congo red binding and virulence was observed in Yersinia pestis, Vibrio cholerae and other bacterial genuses[9,10]. Almost all of the reports on Shigella plasmids focus on the invasion properties and provide little mechanistic information on how these properties are lost. Furthermore, the high numbers of insertion sequences make it difficult to determine how the Congo red binding is lost. In this paper, we attempt to elucidate this mechanism through whole genome comparison using two Shigella strains that lack the capacity to bind Congo red. 1 1.1
Materials and methods Strains and plasmids
(i) Bacterial strains. Provided by Institute of Epidemiology and Microbiology of Chinese Disease Control and Preventive Centre, Beijing, China. (ii) Plasmid isolation. S. flenexrie serotype 5 strain 8401 and S. dysenteriae serotype 1 strain 197 were cultured in the LB medium overnight. The plas-
mids were extracted and purified by QIAGEN kit (QIAGEN company), and designated as pSF5 and pSD1, respectively. 1.2 The plasmid DNA library construction and sequencing. The plasmid DNA was fragmented by sonication. The fractions ranging from 1 to 5 kb were collected and purified and ligated with vector pBluescript to construct random shotgun libraries. The ligation mixtures were transformed to DH5a and cultured on the agar plate containing IPTG and X-gal overnight at 37℃. The clones that did not express β-galactosidase activity were picked for plasmid amplification. Templates were abstracted with a large-scale preparation protocol by using Vitagene kit (Vitagene Inc, China). Templates for sequencing were amplified in a thermocycler with cycle sequencing reactions (30 cycles, 96℃ for 2 min; 50℃ for 10 s; 60℃ for 4 min). About 1500 clones were sequenced from both ends (giving 10 times coverage) using BigDye terminator chemistry on ABI3700 (Perkin Elemer) automated sequencer. 1.3
DNA sequence assembly and annotation
The sequence assembly was carried out by using Phred/Phrap software, choosing optimized parameters and quality score (≥20) in a computer work station. Sequence gaps were filled by editing the ends of sequence traces with Consed or primer walking on plasmid clones selected according to the analysis of the forward and reverse links between contigs. The predicting protein-coding regions were initially defined by searching for open reading frames (ORFs) of more than 50 amino acids in length with Glimmer2.01. The possible ORFs were decided by a combination of searching against the NR (the non-redundant NR proteins) and COG (Clusters of Orthologous Groups of proteins) database with Blastp on internet. Significant homology was defined as more than 30% identity on query sequence and the subject sequence and over 60% homology sequence. 1.4
Nucleotide sequence accession number
The sequences described here have been assigned to GenBank. Accession number was AY879342 of pSF5,
Comparison of virulence plasmid genomes of two stratins of Shigella which lost ability to Crb
and CP000035 of pSD1. 2 2.1
Results The structure of the plasmid genomes
The whole genome analysis showed that the length of pSF5 is 136694 bp. As a circular form, it contains 165 open reading frames (ORFs), of which 133 were identified, and 32 of which show no significant homology to proteins of known function. The length of pSD1 is 182726 bp, is circular and contains 224 ORFs. One hundred and eighty one of these have ORFs and 43 are not significantly homologous to proteins of
143
known function (Fig. 1). The genes can be categorized roughly into 12 types in pSF5 and pSD1, except for the loss of virulence genes in pSF5 (Table 1).The insertion sequence (IS) elements represent 53787 bp in pSF5, which covers 39.3% of the whole genome, and 49616 bp in pSD1, which covers 27.1% of the whole genome. There are 22 types of IS elements in pSF5 and pSD1, among them ISEc8, ISSbo6 we report here for the first time in the virulence plasmid of Shigella (Table 2). The number of IS elements indicates the structural instability of the plasmid. Compared to pCP301, there are a large number of gene deletions and inversions in both pSF5 and pSD1 (Fig. 2).
Fig. 1. The circular map of whole genomes of pSF5 and pSD1. The meanings of the rings from outer to inner are as follows: 1, the genes encoded by the positive DNA chain; 2, the genes encoded by the negative DNA chain; 3, the kind of different IS sequence; 4 and 5, G+C content and GC skew. Table 1 The genes and gene clusters and their functions in pSF5 and pSD1 Gene function Gene name in pSF5 Gene name in pSD1 Genes relative to plasmid replication, repA, repB, ccdA, ccdB, parA, parB repA, repB, ccdA, ccdB, partA, partB, staA, staB, mvpT segregation Plasmid invasive genes ipaH1.4, ipaH2.5, ipaH4.5, ipaH7.8, ipaH9.8; ipaH1.4, ipaH4.5, ipaH7.8, ipaH9.8, ipgB2, ipgB2, virK, sepA mxiACDEGHIJKLMN, spa 9, virAK, sopA (icsP) ospC1-ospC4; ospD1-ospD3; ospE2; ospF, ospG Mix-spa secreted genes ospC1-ospC3; ospD1-ospD3; ospE1-ospE2; phoN2/apy, phoN1 ospF, ospG Energy metabolism phoN2/apy traD, MsbB2 DNA, lipid transport genes traD, MsbB2 traX F pilin acetylation protein traX virB, Regulation genes virF 73 total IS relative ORF 78 total trbH, yacA, yigA, yihA, hmo, total 44 Function unknown trbH, yacB, yigA, yigB, hmo, total 33 finO Fertility inhibition relative finO shf, ushA, rfbU, Acp Sugar metabolism −a) −a) Mobilization mob9 a) “−” indicates no genes and relative products.
144
Science in China: Series C Life Sciences Table 2 The number and types of IS in pSF5 and pSD1 compared with pCP301 No. of ORFs IS1 2 iso-IS1(IS1N) 2 IS2 2 IS3 2 IS4 1 iso-IS10R 1 IS21 2 IS91 1 IS100 2 IS150 3 IS186 1 IS600 2 IS629 2 IS630 1 IS911 2 IS1294 1 ISSfl1 2 ISSfl2 1 ISSfl3 1 ISSfl4 3 ISEc8 3 ISSbo6 3 Total a) Only those with IS fragments ≥100 bp are listed.
No. of intact elements pCP301 pSD1 2 3 0 8 1 2 0 1 1 1 2 0 0 0 0 0 0 1 0 0 0 0 3 2 8 4 1 0 1 0 1 0 1 3 2 1 1 0 2 0 0 0 0 0 26 26
pSF5 3 0 2 0 1 0 0 0 0 0 0 1 3 1 1 2 0 1 1 1 0 0 17
No. of partial elementsa) pCP301 pSD1 pSF5 1 0 1 5 5 5 2 4 5 7 6 5 1 1 1 0 0 0 3 3 1 6 4 4 7 3 5 2 0 2 0 1 0 10 8 6 3 5 8 2 2 2 0 1 1 7 2 3 2 0 3 1 0 0 1 0 1 2 5 4 0 1 1 0 1 3 62 52 61
Fig. 2. Schematic map of whole genomes comparison among pSF5, pSD1 and pCP301.
2.2 Genes relative to plasmid replication and stabilization Many genes, including ccdA, ccdB, parA and parB, play roles in plasmid stabilization and distribution in pSF5. In pSD1, regulatory genes include staA, staB and mvpT. In pCP301, the key regulatory genes are mvpT and mvpA, the latter of which is deleted in pSD1. The genes mvpT and mvpA are suicide factors, whose
functions are to stabilize the plasmid, and play key roles in maintaining consistent plasmid copy numbers during cell generation. The mvpT gene encodes the toxin; the mvpA encodes the antitoxin, while ccdA and ccdB have similar functions. The bacteria maintain the plasmid genome because of the existence of this compensation system. The mvp gene is closely related to the plasmid pathogenesis, since losing the plasmid equates to virulence loss. The virulence characteristic
Comparison of virulence plasmid genomes of two stratins of Shigella which lost ability to Crb
imposed by the plasmid influences the evolution of pathogenic bacteria generation. We have yet to determine which certain regions of mvpT and mvpA are responsible for conferring the virulence capacity[11]. 2.3
Genes relative to virulence
(i) The ipa-mxi-spa region. In pSD1, the sequence between 126261―157773 bp is the ipa-mxi-spa region, spanning 31512 kb (Fig. 3) and encoding approximately 34―36 genes. Among them, many contain ORFs, which exhibit high homology to known proteins. This region is almost the same as that in pCP301[12]. The guanine/cytosine (GC) composition represents 34.25%, which is lower than that of the whole plasmid (44.8%) (Fig. 4). This indicates that the origin and the content of these genes are relatively consistent from an evolutionary standpoint; between the two ends of this region, there are two partial insertion sequences of IS600 and IS100, which differ from that of pCP301, which include two copies of IS600 in the ends of this region. This structure is also different from the yscM-yopD region in the Yersinia plasmid, which contains insertion elements IS100 and IS825[13]. These results also reveal that the bacteria acquired these genes from different sources, and that the rear-
145
rangement occurred extraneously. In pSF5, the invasion genes are all lost, which represents 39.3% of the IS genome, reflecting extremely complex and frequent sequence changes, leading consequently to the large number of gene deletions.
Fig. 3. pSD1.
The Ipa-mxi-spa (124000―159000) and flank regions in
(ii) The ShET gene. The ShET2 gene occurs in serotypes of S. flexneri, S. sonnei and all S. dysenteriae[14]. In pCP301, the genes encoding enterotoxin are ShET2-1 and ShET2-2; in pWR501, the enterotoxin is encoded by ShET2-2; but in pSF5 and pSD1, the enterotoxin is encoded by OspD3(SenA). In pSF5 and pSD1, the OspD3 are 100% and 97% homologous to the ShET2-1 in pCP301, respectively, and are not homologous to the ShET2-2 gene in pCP301 at nucleic acid level. The GC composition of the ShET2 gene in pSF5, pCP301 is 54.7%, and in pSD1 is 36.42%. These results indicate that the origins of pSF5 and pSD1 are obviously different.
Fig. 4. The map of G+C content and GC skew of pSF5(A), pSD1(B) whole genomes.
146
Science in China: Series C Life Sciences
(iii) The ipaH genes. There are five ipaH alleles in pSF5 and pCP301, namely ipaH1.4, ipaH2.5, ipaH4.5, ipaH7.8 and ipaH9.8. These genes are also present in pSD1, with the exception of ipaH2.5. The sequence analyses of pWR501 and pCP301 showed that all five copies of ipaH have a 600―700 bp variable region at the 5′ end, and a 839 bp constant region at the 3′ end. Some studies showed that ipaH is activated instantaneously at the transcription level when Congo red is present in the medium or after bacterial invasion of the epithelial cell. Recent studies showed that the ipaH7.8 products can facilitate the escape of Shigella from phagocytic vacuoles of mouse macrophages and human monocytes[15]. (iv) The vir genes. The virF, virB, virA and virG genes help regulate virulence expression in the expressing plasmid. Also, the regulation is temperature dependent. At 37℃, the transcription factor H-NS (His binding protein) is released, which leads to the binding of virF, virB and icsA (virG), and then leads to the expression of virB. As positive regulartory factor, virB promotes the expression of ipa, mxi and spa[16]. Compared to pCP301, in the ipa-mxi-spa virulence genes in pSF5 are lost except virF, while pSD1 lacks virF. Whether virF exists or not is closely related to the binding of Congo red. Some experiments showed that roughly 97.7% of the Crb− phenotype is related to the virF deletion in S. flexneri strains (2457T)[17]. 2.4
The shf-rfb -msbB genes
The rfbU (capU), virK and msbB loci are regulated under the same operator, and are associated with the biosynthesis of the O-antigen, RfbU, which is a UDP-sugar hydrolase and has been described in Vibrio cholerae as an accessory protein required for O-antigen biosynthesis. The MsbB protein is an acyltransferase involved in fatty acyl modification of the O-antigen. The msbB genes are also present in Shigella and E. coli genomes. Mutations in the msbB genes result in lipopolysaccharide biosynthesis, which reduces cellular toxicity. The virK mutants have less virG mRNA molecules than the wild type, indicating the involvement of virK in posttranscriptional regulation of virG expression. It is interesting to consider that virK, being encoded within the shf-capU-msbB operator, may play a role in O-antigen biosynthesis[18].
Shigella continuously obtains the “weapon”, which allows it to escape the host through horizontal transfer, while gaining the O-antigen, which determines the serotypes. In pSF5, the genes related to O-antigen biosynthesis are lost, with the exception of msbB, which is due mainly to deletions mediated by IS elements. This is reflected by the relic IS regions in the two sides of the related genes. 2.5
The IS genes
To date, the proportion of IS elements in Shigella is the highest in all known genomes, reaching 50%[19]. The ORFs IS elements cover 47% and 32% of pSF5 and pSD1, respectively, and cover 39.3% and 27.2% of the genomes. There is a higher ratio of IS elements in pSF5 than in pCP301 and pSD1, which leads to gene density. This also indicates that IS participates in multiple gene rearrangements, as well as a large number of deletions in genes encoding regulatory and invasion factors. There are 22 kinds of IS in the two plasmids, among them the ISEc8 and ISSbo6 have never been reported in pWR100, pWR501, pMYSH6000, pSS120 and pCP301 before. The IS types in these plasmids are the same except that there is no IS186 in pSF5, and no IS150 and ISSF13 in pSD1. The complete and partial IS elements are 17 and 61 in pSF5, and 26 and 52 in pSD1, respectively (Table 2). 2.6
The unknown gene functions
There are 33 genes with unknown functions (covering 20% of all ORFs) in pSF5, including trbH, YacB, YigA and YigB. There are 44 genes of unknown functions (cover 19.6% of all ORFs) in pSD1, including trbH, YacA, YigA, YihA and spa-orf10. To date, the function of these genes is still unknown. 3
Discussion
S. dysenteriae serotype 1 and S. flexneri serotype 5 belong to Shigella subgroups I and III, respectively. These categories reflect divergent evolutionary patterns. The clustering of Shigella strains by plasmid forms is almost consistent with chromosomal gene sequence clustering[20]. The results of the whole genome comparison showed that many genes in pSF5 and pSD1 have obvious origin differences, but both
Comparison of virulence plasmid genomes of two stratins of Shigella which lost ability to Crb
exhibit the Crb− phenotype. Congo red dye is analogous to hemin in structure. Hemin, as an important source of iron, plays a key role in bacteria growth in the host. The competition experiment indicated that Congo red and hemin bind to the same site on the bacterial cell surface[21]. Both Crb+ and Crb− cells can utilize hemin as their sole source of iron. The ability of Crb+-containing organisms to invade HeLa cells significantly increases when the bacteria pre-bind hemin or when hemin is added to the HeLa cell monolayer, which indicates that the products of the Congo red genes are involved in bacterial attachment onto the intestinal epithelial cells. During the invasion stage IpaB, IpaC and IpaD play the key functions. Furthermore, Congo red can induce IpaA and IpgD secretion, and the inactivation of the above two genes attenuates the capability of Shigella to invade HeLa cells. IpaA and IpgD only exist outside the membrane and do not bind closely to the membrane, and their injection into the host also relies on the existence of the IpaB and IpaC. The expression of the Ipa genes is closely related to the activity of the regulator virF. Whole genome analysis shows that there are a number of gene deletions in pSF5 and pSD1, and the ipa-mxi-spa island is lost in pSF5 except virF, which indicates that it can no longer bind Congo red. In the pSD1, the loss of virF also causes the Congo red binding mutation, even if the ipa-mxi-spa island is complete. All the above deletions are mediated by the IS elements, which are different from the previous reported cases, in which the Congo red binding mutation by an IS1-like element inserts into the virF locus. In normal subculture conditions, the loss of plasmid invasion capability relieves the energy consumption caused by gene replication. Thus the loss is beneficial for Shigella survival outside of the host. This phenomenon is verified by the observation that bacteria that lost the invasion capability grew faster[18]. Under laboratory conditions, about 97% of the lost invasion cases were related to the absence of virF and virB. In addition, this phenomenon is temperature-dependent. At 37 ℃ , we found about 50% of S. flexneri 2457T Crb− mutations after three sequential passages, while at 30℃ the Crb− mutation occurred in only 0.09%. These mutation rates remained constant over seven days of cultivation. The results were due to the repression of virulence gene
147
expression as virF, virB cannot function properly under this temperature. But the recombinant strains capable of producing either VirF or VirB at high levels during growth at 30℃ are also induced by the Crb− emergence. Thus, these results further support the expression of virulence genes as the main cause of Crb− emergence. Once a spontaneously rearranged virulence plasmid arises in a population of S. flexneri IS1-like elements in the virulence plasmid at the virF locus, it is stably maintained, and gene rearrangements and deletion will seldom occur. This may be one of the strategies that the bacteria utilize to survive under laboratory conditions. Moreover, it is common to find a virF and/or virB deletion, but it is rare to find both lost simultaneously. The same situation holds for pSF5 and pSD1, though the mechanism is still unknown. Furthermore, part of Crb− can convert back to Crb+ if it is cultured with the Crb+ lysate. The phenomenon, in our opinion, is mainly due to regaining the function of Crb, rather than restoring the ipa-mxi-spa island. This is because it is difficult to obtain large fragments through gene rearrangement, although the number of IS elements and its product transposases favor such an event. Thus, the Crb− mutant results from inversion or deletion. Additionally, there is evidence that invasive bacteria can bind Congo red, and non-invasive bacteria can do so[8]. This result suggests that virF does not regulate the binding of Congo red. In S. flexnerri 2a, the gaIU::Tn10 (glucose-1-phosphate uridy1 transferase) mutation also can lead to Crb−. The main function of gaIU is to add to the lipopolysaccharide content of the cell wall by side chain catalysis, an activity that is similar to cell wall biosynthesis. Thus, it appears that the genes related to cell wall and outer membrane biosynthesis are also involved in binding Congo red. The genes shf- and rfbU, which together with msbB are involved in O-antigen biosynthesis, are lost in pSF5. All three of these genes, however are present in pSD1. In conclusion, the comparison of the pSF5 and pSD1 genomes showed that the loss of genes relative to invasion and regulation result in the concomitant loss of the Congo red binding property. However, whether the virF and virB regulators cause this mutation remains unclear. The completion of pSF5 and pSD1 genomes provides a good foundation for elucidating this phenomenon.
148
Acknowledgement This work was supported by the State Key Basic Research Program (Grant No. 2005CB522904), the High Technology Project (Grant No. 2001AA223011), International Science and Technology Co-operation Project (Grant No. 2001AA223116) and Beijing Science and Technology Project (Contract No. H010210360119).
Science in China: Series C Life Sciences Bacteriol., 1985, 164(3): 1233―1237. 11. Sayeed, S., Reaves, L., Radnedge, L. et al., The Stability region of the large virulence plasmid of Shigella flexneri encodes an efficient postsegregational killing system, J. Bacteriol., 2000, 182(9): 2416―2421. 12. Zhang, J. Y., Liu, H., Zhang, X. B. et al., Complete DNA sequence and gene analysis of the virulence plasmid pCP301 of
References 1. O’Loughlin, E. V., Robins-Browne, R. M., Effect of Shiga toxin and Shiga like toxins on eukaryotic cells, Microbes Infection, 2001, 3: 493―507. 2. Jin, Q., Yuan, Z. H., Xu, J. G. et al., Genome sequence of Shigella flexneri 2a: Insight into pathogenicity through comparison with genomes of Escherichia coli K12 and O157, Nucleic Acid Res., 2002, 30(20): 4432―4441. 3. Reid, S. D., Herbelin, A. C., Bumbaugh, A. C. et al., Parallel evolution of virulence in pathogenic Escherichia coli, Nature, 2000, 406(6791): 64―67. 4. Liu, H., Peng, J. P., Yang, J. et al., Analysis of components of conserved “backbone sequences” among genomes of Shigella spp. Strains, Chinese Sci. Bull., 2004, 49(2): 152―160.
Shigella flexneri 2a, Science in China, Ser. C, 2003, 46(5): 513― 521. 13. Perry, R. D., Straley, S. C., Fetherston, J. D. et al., DNA sequence and analysis of the low-Ca2+-response plasmid pCD1 of Yersinia pestis KIM5, Infect. Immun., 1998, 66(10): 4611―4623. 14. Martha, V., Joaquim, G., Maria, T. et al., Prevalence of Shigella Enterotoxins 1 and 2 among Shigella strains isolated from patients with traveler’s diarrhea, J. Clin. Microbiol., 1999, 37(11): 3608― 3611. 15. Fernandez-Prada, C. M., Hoover, D. L., Tall, B. D. et al., Shigella flexneri IpaH7.8 facilitates escape of virulent bacteria from the endocytic vacuoles of mouse and human macrophages, Infect. Immun., 2000, 68(6): 3608―3619. 16. Jost, B. H., Adler, B., Site of transcriptional activation of virB on
5. Ruiting, L., Gordon, S., Peter, R. et al., Comparison of two major
the large plasmid of Shigella flexneri 2a by VirF, a member of the
forms of the Shigella virulence plasmid pINV: Positive selection
AraC family of transcriptional activators, Microb. Pathog., 1993,
is a major force driving the divergence, Infect. Immun., 2003, 71(11): 6298―6306. 6. Payne, S. M., Finkelstein, R. A., Detection and differentiation of iron-responsive avirulent mutants on Congo red agar, Infect. Immun., 1977, 18(1): 94―98. 7. Daskaleros, P. A., Payne, S. M., Cloning the gene for Congo red binding in shigella flexneri, Infect. Immun., 1985, 48(1): 165― 168. 8. Maurelli, A. T., Blackmon, B., Curtiss, R. 3rd, Loss of pigmentation in Shigella flexneri 2a is correlated with loss of virulence and virulence-associated Plasmid, Infect. Immun., 1984, 43(1): 397― 401. 9. Surgalla, M. J., Beesley, E. D., Congo red agar plating medium for detecting pigmentation of Pasteurella pestis, Appl. Microbiol., 1969, 18(5): 834―837.
14(6): 481―488. 17. Schuch, R., Maurelli, A. T., Virulence plasmid instability in Shigella flexneri 2a is induced by virulence gene expression, Infect. Immun., 1997, 65(9): 3686―3692. 18. Somerville, J. E., Cassiano, J. L., Bainbridge, B. et al., A novel Escherichia coli lipid A mutant that produces antiinflammatory lipopolysaccharide, J. Clin. Investig., 1996, 97(2): 359―365. 19. Venkatesan, M. M., Goldberg, M. B., Rose, D. J. et al., Complete DNA sequence and analysis of the large virulence plasmid of Shigella flexneri, Infect. Immun., 2001, 9(5): 3271―3285. 20. Lan, R., Lumb, B., Ryan, D. et al., Molecular evolution of large virulence plasmid in Shigella clones and Enteroinvasive Escherichia coli, Infect. Immun., 2001, 69(10): 6303―6309. 21. Daskaleros, P. A., Payne, S. M., Congo red binding phenotype is
10. Ishiguro, E. E., Ainsworth, T., Trust, T. J. et al., Congo red agar, a
associated with hemin binding and increased infectivity of Shig-
differential medium for Aeromonas salmonicida, detects the
ella flexneri in the HeLa cell model, Infect. Immun., 1987, 55(6):
presence of the cell surface protein array involved in virulence, J.
1393―1398.