Abstract
CTXφ is a filamentous bacteriophage that encodes cholera toxin, the principal virulence factor of Vibrio cholerae. CTXφ is unusual among filamentous phages because it encodes a repressor and forms lysogens. CTXφ can infect the existing live-attenuated V. cholerae vaccine strains derived from either the El Tor or classical V. cholerae biotypes and result in vaccine reversion to toxinogenicity. Intraintestinal CTXφ transduction assays were used to demonstrate that El Tor biotype strains of V. cholerae are immune to infection with the El Tor-derived CTXφ, whereas classical strains are not. The El Tor CTXφ repressor, RstR, was sufficient to render classical strains immune to infection with the El Tor CTXφ. The DNA sequences of the classical and El Tor CTXφ repressors and their presumed cognate operators are highly diverged, whereas the sequences that surround this “immunity” region are nearly identical. Transcriptional fusion studies revealed that the El Tor RstR mediated repression of an El Tor rstA-lacZ fusion but did not repress a classical rstA-lacZ fusion. Likewise, the classical RstR only repressed a classical rstA-lacZ fusion. Thus, similar to the mechanistic basis for heteroimmunity among lambdoid phages, the specificity of CTXφ immunity is based on the divergence of the sequences of repressors and their operators. Expression of the El Tor rstR in either El Tor or classical live-attenuated V. cholerae vaccine strains effectively protected these vaccines from CTXφ infection. Introduction of rstR into V. cholerae vaccine strains should enhance their biosafety.
Vibrio cholerae is the cause of the severe diarrheal disease cholera. After oral ingestion of contaminated food or water, this Gram-negative rod colonizes the human small intestine. In the small intestine, V. cholerae secretes cholera toxin (CT), an A-B-type toxin that binds to GM1 ganglioside on host intestinal epithelial cells (1). The activity of this ADP-ribosylating exotoxin largely accounts for the secretory diarrhea, which is characteristic of cholera (2). The classical biotype of Vibrio cholerae O1 gave rise to the fifth (1881–1896) and sixth (1899–1923) cholera pandemics (3). The ongoing seventh pandemic of cholera, which began in 1961 in Indonesia, is caused by the El Tor biotype of V. cholerae O1. The observation that cholera seems to engender long-lived immunity to repeat V. cholerae infection has led to efforts to develop an oral live-attenuated V. cholerae vaccine (4). In the past decade, both classical and El Tor V. cholerae strains have been used to construct several candidate live-attenuated V. cholerae vaccine strains that currently are undergoing clinical trials (4–6).
The genes encoding cholera toxin (ctxAB) are part of the genome of CTXφ, a 6.9-kb single-stranded DNA filamentous bacteriophage (7). Whereas the DNA encoding most filamentous phages remains extrachromosomal as plasmids (8), the CTXφ genome encodes a site-specific recombination system that catalyzes the integration of the phage DNA into the attRS site on the El Tor V. cholerae chromosome to form lysogens (7, 9). After infection of El Tor-derived live-attenuated vaccine strains that are deleted for attRS and all CTXφ sequences (10) or classical-derived vaccine strains that are deleted for ctxA (which encodes the enzymatically active A subunit of cholera toxin) (4), CTXφ remains extrachromosomal and replicates as a plasmid (7, 9). Thus, the discovery that ctxAB is transmissible as part of the CTXφ genome suggests that CTXφ infection could mediate the reversion of live-attenuated V. cholerae vaccine strains. In fact, we have found that V. cholerae vaccine strains derived from either biotype can be transduced efficiently by CTXφ within the intestinal tract (7).
The CTXφ genome is divided into a 4.6-kb core region, which encodes cholera toxin as well as functions that are required for virion morphogenesis (7), and a 2.4-kb RS2 region, which encodes functions required for regulation, replication, and integration of CTXφ (9). CTXφ, similar to lysogen-forming (integrating), double-stranded DNA bacteriophages, and different from other filamentous phages, encodes a repressor (9). This repressor, RstR, has been shown to repress expression of a rstA, a gene that is divergently transcribed from rstR and is required for CTXφ replication (9). In lambdoid phages, repressors provide immunity to secondary infection by an identical phage (11). Because lysogeny is unusual among filamentous phages, in the current study we explored whether V. cholerae CTXφ lysogens exhibit immunity to infection by CTXφ and the role of RstR in CTXφ immunity. El Tor lysogens were found to be immune to infection by El Tor-derived CTXφ whereas classical strains were not. RstR was sufficient to render classical strains immune to CTXφ infection. The sequences of the classical and El Tor CTXφ encoded repressors (rstR genes) and their cognate operators were found to be highly diverged and repression mediated by these rstR alleles was biotype-specific. The use of rstR as a means to protect classical and El Tor live-attenuated V. cholerae vaccine strains from reversion to toxinogenicity mediated by CTXφ infection is described.
MATERIALS AND METHODS
Plasmid Constructions.
pHK1 contains the El Tor rstR gene cloned into the arabinose-inducible promoter vector pBAD33 (12) and was constructed as follows: oligonucleotide primers rstR-3 (GGGAGCTCAAAGGGGATTGTTAT) and rstR-2 (CCTCTAGATAGTATTACGGGGGT) were used to amplify rstR from CTXφ replicative form (RF) DNA with PCR. PCR products were purified by using QIAquick PCR purification kit (Qiagen, Chatsworth, CA), digested with SacI and XbaI (New England Biolabs), and ligated to SacI–XbaI-digested pBAD33 plasmid DNA. pHK2, containing the rstRCL gene, was constructed similarly by using oligonucleotides rstR-5 (GGGAGCTCGTTCAAAAATAAGCACAA) and rstR-6 (CCTCTAGAGATTACCTACCTAAATTTC). Chromosomal DNA from the classical strain 0395 was used as the template DNA for PCR. To construct pHK3 and pHK4, which contain the El Tor rstR gene cloned into pBR322 (13) and pACYC184 (14), respectively, primers rstR-8 (AACGGCCGCTAAGCACCATGATTT) and rstR-9 (GGGGATCCTTCGACATCAAATGGCA) were used to amplify rstR. The purified PCR product then was digested with EagI and BamHI and ligated into EagI- and BamHI-digested pBR322 and pACYC184.
The classical rstACL-lacZ reporter plasmid pHK101 contains the ig-2CL region and the 5′ end of rstACL cloned into the β-galactosidase (lacZ) reporter plasmid pCB192 (15) and was constructed by PCR cloning by using oligonucleotide primers crstR-1 (GGAAGCTTGTTTAGATCTCTCTCAAC), crstR-2 (GGTCTAGACCAGATAAGCGAGGACAA), and classical strain 0395 DNA as template. PCR products were purified, digested with HindIII and XbaI, and ligated to HindIII- and XbaI-digested pCB192 plasmid DNA. The El Tor rstA-lacZ reporter plasmid, pHK102, contains the El Tor ig-2 region and the 5′ end of rstA and was constructed by digesting CTXφ RF DNA with NheI and KasI. DNA fragments were blunt-ended with T4 DNA polymerase, and the 290-bp fragment containing ig-2 and the 5′ end of rstA was gel-purified and ligated to EcoRV-digested pBluescript II-KS(+) (Stratagene), generating pHK100. Finally, pHK100 was digested with HindIII and XbaI, and the 300-bp fragment containing ig-2 was cloned into HindIII- and XbaI-digested pCB192 plasmid DNA yielding pHK102.
Intraintestinal CTX-Knφ Transduction Assay.
The suckling mouse V. cholerae colonization model (16) was used to detect intraintestinal CTX-Knφ transduction of recipient strains. Five-day-old CD-1 mice were intragastrically inoculated with 2 × 105 cells of a CTX-Knφ donor strain along with 1 × 105 cells of different potential recipient strains. The lacZ− classical strain LAC-1 (pCTX-Kn) (16), which harbors the replicative form of CTX-Knφ, was used as the CTX-Knφ donor strain. All the recipient strains were lacZ+. Thus, transductants of the different recipient strains were identified as kanamycin-resistant (Knr) LacZ+ colonies. The different inocula mixtures were plated to verify that there were no Knr LacZ+ cells in any of the inocula. After 20 hr of intraintestinal growth, homogenates of the small intestines were plated to determine the total number of recipient cells as well as the number of recipient cells that were transduced to Kn resistance with CTX-Knφ within the intestine. There were at least six mice in each group.
Sequencing the Classical CTXφ RS2 Region.
Plasmid subclones containing the two chromosomal copies of the classical strain 569B RS2 region, pGP2 and pGP19 (17), were used as templates for dye terminator cycle sequencing of the classical RS2 region. DNA sequences were determined with an Applied Biosystems 373A DNA sequencer. The blastp (18) program was used to detect similarities of RstRCL to other bacteriophage repressors, and geneworks (Oxford Molecular Group, Oxford) was used to align the classical and El Tor RS2 nucleotide and protein sequences. The classical RS2 region has been assigned GenBank accession number AF055890.
RESULTS AND DISCUSSION
Immunity of El Tor Strains to CTXφ Infection.
An essential V. cholerae intestinal colonization factor, the toxin coregulated pilus (TCP) (19), is the receptor for CTXφ (7). In El Tor strains, TCP is not expressed efficiently during in vitro growth; therefore, we used an intraintestinal transduction assay to study whether El Tor CTXφ lysogens exhibit immunity to CTXφ infection. In this assay, a donor strain harboring the RF of CTX-Knφ [the El Tor CTXφ RF, which contains a Kn resistance gene replacing ctxAB (7)], was coinoculated with either an El Tor lysogen (E7946) or its CTXφ− attRS− vaccine derivative (Bah-2) (10) into the gastrointestinal tracts of suckling mice. These two potential CTX-Knφ recipient strains colonized the suckling mouse small intestine approximately equally well (Fig. 1). Comparison of the numbers of CTX-Knφ transductants of E7946 or Bah-2 in intestinal homogenates revealed that although nearly 1 in 10 Bah-2 cells were transduced to KnR with CTX-Knφ in the intestine, only 1 in 104 E7946 cells were transduced within the intestine (Fig. 1). This three-order-of-magnitude difference in the frequency of recovery of transductants between these strains indicates that El Tor CTXφ lysogens exhibit immunity to further CTXφ infection.
RstR Is Sufficient for CTXφ Immunity.
Strains of the classical biotype of V. cholerae are not immune to infection with the El Tor-derived CTXφ. Classical strains readily express TCP either in vitro, by using appropriate growth conditions, or within the intestine and can be efficiently transduced with the El Tor-derived CTX-Knφ both in vitro and in vivo (7). In these transductants, CTXφ does not integrate but replicates as a plasmid (7, 9). In lambdoid phages repressors are known to mediate phage immunity (11). Although CTXφ is unrelated to lambdoid phages, we investigated whether El Tor CTXφ RstR could render a classical strain immune to transduction with the El Tor CTXφ. To test this possibility, a plasmid (pHK1) that contains the El Tor rstR under the transcriptional control of PBAD, an arabinose-inducible promoter (12), was introduced into classical strain O395 (20). After growth of O395(pHK1) in the presence or absence of inducer, the cells were infected with the El Tor CTX-Knφ. Induction of rstR expression with arabinose resulted in an approximately 900-fold reduction in the number of transductants of O395(pHK1) with CTX-Knφ compared with the number of transductants of O395(pHK1) after growth in the absence of arabinose. Arabinose did not alter the transducibility of O395 harboring the pBAD33 vector without an insert. These findings indicate that El Tor RstR is sufficient to render classical V. cholerae immune from transduction with El Tor-derived CTXφ. The converse experiment, testing whether the classical RstR can render an El Tor strain immune to infection with classical CTXφ, is not possible at this time because to date we have not been able to induce CTXφ virion production from classical lysogens. This may suggest that the classical CTX prophage is defective.
Divergence of the Classical and El Tor CTXφ Repressors.
To begin to address the molecular basis for the lack of immunity of classical lysogens to El Tor CTXφ infection, we sequenced the RS2 region of the classical CTXφ by using subclones (17) of the chromosomal copies of the classical CTX prophage. Previous studies of the RS2 region of the El Tor CTXφ genome have revealed that this region encodes products required for phage DNA replication and integration (rstAB) as well as a repressor, rstR, which is transcribed divergently from all the other CTXφ genes and which represses transcription of rstA (9) (Fig. 2). DNA sequence analysis revealed that like the El Tor CTXφ RS2 region, classical CTXφ RS2 contains three ORFs, designated rstRCL, rstACL, and rstBCL, and two intergenic regions (ig-1CL and ig-2CL) (Fig. 2). Also, like the El Tor RS2 region, an intergenic region (ig-2CL) separates the divergently transcribed rstRCL and rstACL. The nucleotide sequences of rstACL and rstBCL were 94% identical to their El Tor homologues, and the predicted amino acid sequences of these proteins are 99% identical. In striking contrast to this, the nucleotide sequences of the rstR and ig-2 sequences were highly divergent in the two biotypes (Fig. 2). Despite the nucleotide and predicted amino acid sequence divergence of the classical and El Tor rstR genes (RstR and RstRCL are 24% identical and 32% similar), the classical RstRCL, like the El Tor RstR (9), is similar to other bacteriophage repressor proteins.
Repression Mediated by the RstR Alleles Is Biotype-Specific.
To address the specificity of repression mediated by the classical and El Tor rstR genes, lacZ transcriptional fusions to the classical and El Tor rstA alleles (pHK101 and pHK102, respectively) were constructed. Also, plasmids containing the classical and El Tor rstR genes under the transcriptional control of an arabinose-inducible promoter (12) were constructed. Combinations of plasmids containing either of the two biotype rstA∷lacZ reporters and either of the two biotype-inducible repressors were introduced into an E. coli Δ lacZ strain [CC118 (21)]. After growth in the presence of the inducer arabinose, El Tor RstR repressed expression of the El Tor rstA-lacZ fusion nearly 200-fold but had no repressive effects on expression of the classical rstA-lacZ fusion (Table 1). Similarly, RstRCL repressed expression of the classical rstA-lacZ fusion nearly 80-fold but had no repressive effect on expression of the El Tor rstA-lacZ fusion. That is, RstR-mediated repression of rstA expression is biotype-specific. Expression of the rstA-lacZ reporters also was repressed in a biotype-specific manner after these reporters were introduced into El Tor (E7946) or classical (O395) lysogens (Table 1). This indicates that the repressors are expressed and active in lysogens of their respective biotypes. Our data suggest that classical lysogens lack immunity to El Tor CTXφ infection because the classical RstRCL is unable to repress El Tor rstA expression.
Table 1.
Reporter | RstR repressor*
|
|||||
---|---|---|---|---|---|---|
Classical (pHK2)
|
El Tor (pHK1)
|
V. cholerae CTXφ lysogen†
|
||||
−Arabinose | +Arabinose | −Arabinose | +Arabinose | Classical | El Tor | |
Classical rstA-lacZ (pHK 101) | 527 | 7 | 688 | 933 | 5 | 74 |
El Tor rstA-lacZ (pHK102) | 157 | 167 | 255 | 1.2 | 260 | 1 |
β-Galactosidase units are reported as nmol of o-nitrophenyl β-d-galactoside hydrolyzed per min per OD600. Assays were performed in triplicate, and the average value is presented.
The arabinose-inducible classical (pHK2) or El Tor (pHK1) rstR plasmids were introduced along with an rstA-lacZ reporter plasmid into Escherichia coli strain CC118 (21). Cultures were grown in L broth (25) for 16 hr in the presence or absence of 0.05% arabinose, and the β-galactosidase activity was determined (25).
The classical and El Tor rstA-lacZ reporters were introduced into lacZ− derivatives of classical (0395) and El Tor (E7946) V. cholerae CTXφ lysogens, and the activity of β-galactosidase was determined (25).
The molecular bases of phage immunity have been well studied in the temperate, double-stranded lambdoid phages. Among this large group of bacteriophages it has been demonstrated that divergence of the sequences of the repressor, cI, and its operators establishes the molecular basis for the finding that λ lysogens are not immune to lytic infection by closely related lambdoid phages such as 434 (11). This lack of immunity among very closely related lambdoid phages is referred to as “heteroimmunity” (11). Our observations strongly suggest that the divergence of the classical and El Tor rstR genes and their operators in ig-2 establishes a heteroimmunity-like phenomenon among CTXφ. Because filamentous CTXφs and lambdoid phages are distinct classes of viruses, it is remarkable that in both cases repressors mediate phage immunity and divergence of repressors and their cognate operators accounts for heteroimmunity. This may suggest that the evolutionary history of CTXφ included acquisition of the rstR-ig-2 immunity region by horizontal gene transfer.
Use of rstR to Protect Live-Attenuated V. cholerae Vaccine Strains from CTXφ Infection.
The live-attenuated V. cholerae vaccine strains derived from both biotypes are transducible with the El Tor-derived CTXφ (7) and therefore are capable of reversion to toxinogenicity. The finding that rstR and ig-2 function as an immunity region for CTXφ suggested that El Tor RstR could be used to protect El Tor-derived live vaccine strains, which contain a deletion of the entire CTXφ (5, 6), as well as classical-derived vaccine strains, which contain a deletion of ctxA (4), from infection with the El Tor CTXφ. To test this possibility, plasmids containing El Tor rstR (pHK3 or pHK4) or the same plasmids without inserts, pBR322 and pACYC184, respectively, were introduced into the El Tor vaccine strain Bah-2 (10), a CTXφ− attRS− strain, and into classical vaccine strain O395-N1 (22), a ctxA− strain. Then, the numbers of intestinally derived CTX-Knφ transductants of these vaccine strains were determined by using the in vivo transduction assay as described above.
No detectable intestinally derived KnR transductants of Bah-2 (pHK3) or O395-N1 (pHK4), which contain plasmid-expressed El Tor rstR, were recovered in intestinal homogenates (Table 2). In contrast, nearly 1 in 10 Bah-2 (pBR322) and 1 in 20 O395-N1 (pACYC184) were transduced with CTX-Knφ within the intestine (Table 2). This intestinal transduction assay can detect transduction frequencies as low as 1 in 100,000. Thus, introduction of the El Tor rstR into live-attenuated V. cholerae vaccine strains provides a means to ensure that these vaccines will be immune to CTXφ infection and thereby should significantly lower the possibility of vaccine reversion to toxinogenicity.
Table 2.
Recipient strain* | CFU in intestinal homogenates†
|
|
---|---|---|
Total no. of recipients | Transductants | |
BAH-2(pHK3) | 3.7 × 105 | 0 |
BAH-2(pBR322) | 5.6 × 105 | 5.3 × 104 |
O395-N1(pHK4) | 2.2 × 106 | 0 |
O395-N1(pACYC184) | 1.9 × 106 | 8.1 × 104 |
V. cholerae CTXφ donor strain LAC-1(pCTX-Kn) was mixed with each of the different recipient strains and gastrointestinally inoculated into suckling mice.
Intestinal homogenates were plated on selective media to enumerate the total numbers of colony-forming units (CFU) of each recipient strain and the numbers of kanamycin-resistant CFU (tranductants) of each recipient strain after 20 hr of intraintestinal growth. There were at least eight mice in each group, and the median number of CFU in each group is presented.
Other strategies to prevent CTXφ-mediated reversion of live vaccine strains can be envisioned. For example, deletion of the CTXφ receptor, TCP, is another approach. However, strains harboring tcpA deletions do not colonize the intestine and do not induce a protective immune response (23). Therefore, unless TCP pili, which are not functional CTXφ receptors but which enable V. cholerae colonization, can be constructed, this approach may not be successful. Expression of rstR in V. cholerae vaccine strains has potential limitations as well. For example, analogous to other temperate phages, vir mutants of CTXφ may arise. Similarly, if there are many ctxAB+ CTX phages with different immunity regions present in the world, then this strategy will not be useful. To begin to address this possibility, we compared the nucleotide sequences of rstR-ig-2 derived from El Tor clinical isolates from different continents over the past 25 years. The nucleotide sequences were identical in all five strains studied. This finding supports the clonality of the seventh pandemic of cholera (24). In addition, the nucleotide sequence of rstR-ig-2 in MO45, a V. cholerae O139 serogroup strain, was identical to the El Tor sequence. The conservation of the sequence of the immunity region of CTXφ in these El Tor strains suggests that the existence of many different CTXφ immunity regions is improbable. Thus, introduction of a stably maintained rstR into the existing V. cholerae vaccine candidates should improve their biosafety. Heterologous expression of viral regulatory genes may have applications in the development of other live bacterial vaccines and potentially in the design of human antiviral gene therapy as well.
Acknowledgments
We are most grateful to Dr. John Mekalanos for his continued advice and support. We thank Greg Pearson and John Mekalanos for providing the plasmids used for sequencing classical RS2, and Eric Rubin and Dan Steiger for assistance with DNA analysis and sequencing. We thank Luzma Guzman and Jon Beckwith for pBAD33. We are grateful to our colleagues Drs. Andrew Wright, Kevin Killeen, Linc Sonenshein, Sara Lazar, and Andrew Camilli for their helpful comments on this manuscript. This work was supported by National Institute of Allergy and Infectious Diseases Grant AI42347, the Tupper Research Fund, and the New England Medical Center GRASP Digestive Disease Center.
ABBREVIATIONS
- Kn
kanamycin
- CFU
colony-forming units
- RF
replicative form
Footnotes
Data deposition: The sequence reported in this paper has been deposited in the GenBank database (accession no. AF055890).
References
- 1.Sears C L, Kaper J L. Microbiol Rev. 1996;60:167–215. doi: 10.1128/mr.60.1.167-215.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Levine M M, Kaper J B, Black R E, Clements M L. Microbiol Rev. 1983;47:510–550. doi: 10.1128/mr.47.4.510-550.1983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Barua B. In: Cholera. Barua D, Greenough W B III, editors. New York: Plenum; 1992. pp. 1–36. [Google Scholar]
- 4.Levine M M, Pierce N F. In: Cholera. Barua D, Greenough W B, editors. New York: Plenum; 1992. pp. 285–328. [Google Scholar]
- 5.Mekalanos J J, Sadoff J C. Science. 1994;265:1387–1389. doi: 10.1126/science.8073279. [DOI] [PubMed] [Google Scholar]
- 6.Waldor M K, Mekalanos J J. In: Mucosal Vaccines. Kiyono H, Ogra P L, McGhee J R, editors. San Diego: Academic; 1996. pp. 229–240. [Google Scholar]
- 7.Waldor M K, Mekalanos J J. Science. 1996;272:1910–1914. doi: 10.1126/science.272.5270.1910. [DOI] [PubMed] [Google Scholar]
- 8.Model P, Russel M. In: The Bacteriophages. Calendar R, editor. Vol. 2. London: Plenum; 1988. pp. 375–456. [Google Scholar]
- 9.Waldor M K, Rubin E J, Pearson G D N, Kimsey H, Mekalanos J J. Mol Microbiol. 1997;24:917–926. doi: 10.1046/j.1365-2958.1997.3911758.x. [DOI] [PubMed] [Google Scholar]
- 10.Pearson G D N, Woods A, Chiang S L, Mekalanos J J. Proc Natl Acad Sci USA. 1993;90:3750–3754. doi: 10.1073/pnas.90.8.3750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ptashne M. A Genetic Switch: Phage and Higher Organisms. Cambridge, MA: Cell Press; 1992. [Google Scholar]
- 12.Guzman L-M, Belin D, Carson M J, Beckwith J. J Bacteriol. 1995;177:4121–4130. doi: 10.1128/jb.177.14.4121-4130.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bolivar F, Rodriguez R L, Greene P J, Betlach M C, Heynecker H W, Boyer H W, Crosa J W, Falkow S. Gene. 1977;2:95–113. [PubMed] [Google Scholar]
- 14.Chang A C Y, Cohen S N. J Bacteriol. 1978;134:1141–1156. doi: 10.1128/jb.134.3.1141-1156.1978. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Schneider K, Beck C F. Gene. 1986;42:37–48. doi: 10.1016/0378-1119(86)90148-4. [DOI] [PubMed] [Google Scholar]
- 16.Waldor M K, Mekalanos J J. Infect Immunol. 1994;62:72–78. doi: 10.1128/iai.62.1.72-78.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Pearson G D N. Ph.D. dissertation. Cambridge, MA: Harvard Univ.; 1989. [Google Scholar]
- 18.Altshul S F, Gish W, Miller W, Myers E W, Lipman D J. J Mol Biol. 1990;215:403–410. doi: 10.1016/S0022-2836(05)80360-2. [DOI] [PubMed] [Google Scholar]
- 19.Taylor R K, Miller V L, Furlong D B, Mekalanos J J. Proc Natl Acad Sci USA. 1987;84:2833–2837. doi: 10.1073/pnas.84.9.2833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Mekalanos J J. Cell. 1983;35:253–263. doi: 10.1016/0092-8674(83)90228-3. [DOI] [PubMed] [Google Scholar]
- 21.Manoil C, Beckwith J. Proc Natl Acad Sci USA. 1985;82:8129–8133. doi: 10.1073/pnas.82.23.8129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Mekalanos J J, Swartz D J, Pearson G D, Harford N, Groyne F, deWilde M. Nature (London) 1983;306:551–557. doi: 10.1038/306551a0. [DOI] [PubMed] [Google Scholar]
- 23.Herrington D A, Hall R H, Losonsky G, Mekalanos J J, Taylor R K, Levine M M. J Exp Med. 1988;168:1487–1492. doi: 10.1084/jem.168.4.1487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Karaolis D K, Lan R, Reeves P R. J Bacteriol. 1995;177:3191–3198. doi: 10.1128/jb.177.11.3191-3198.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Miller J H. A Short Course in Bacterial Genetics. Plainview, NY: Cold Spring Harbor Lab. Press; 1992. [Google Scholar]