INFECTION AND IMMUNITY, Oct. 1994, p. 4686-4689
Vol. 62, No. 10
0019-9567/94/$04.00+0 Copyright X 1994, American Society for Microbiology
Analysis of a Naturally Occurring K99+ Enterotoxigenic Escherichia coli Strain That Fails To Produce K99 RICHARD E. ISAACSON* AND SHEILA PATITERSON Department of Veterinary Pathobiology, University of Illinois, Urbana, Illinois 61801 Received 31 March 1994/Returned for modification 2 June 1994/Accepted 21 July 1994
A spontaneously occurring field isolate of enterotoxigenic Escherichia coli that was genotypically K99' but phenotypically K99- was analyzed for the reason that it did not express K99. The defect, which was cis active, was of a
located within an area 5' to the first gene required for K99 biogenesis and was the result of the deletion single base pair.
The K99 pilus adhesin of enterotoxigenic Escherichia coli (ETEC) is an important virulence factor produced by many ETEC strains that cause neonatal diarrhea in pigs, lambs, and calves (10). K99 facilitates the colonization of small intestines by ETEC by attaching to unique receptors on the epithelial cell surface (8, 20, 22). K99 is transmissible, being genetically encoded on an 87.8-kb plasmid (9). A 7.15-kb BamHI DNA fragment specifying production of K99 has been cloned and shown to contain eight genes that must be expressed for the biogenesis of K99 (Fig. 1) (1, 21). The eight required genes have been designated fanA to fanH. FanA and FanB probably function as positive regulators (1, 15), FanC is the major pilin subunit (17), and FanD is presumed to be an usher (11, 16). On the basis of DNA sequence homology, FanE is presumed to be a chaperone (2) and FanF to FanH are believed to be involved in pilus assembly and elongation (18, 19). Expression of K99 has been shown to be dependent upon cyclic AMP (cAMP) (6), cAMP receptor protein (CRP) (5, 7), and leucine-responsive protein (LRP) (3). While all eight genes are clustered contiguously, analysis of Northern (RNA) blots using the native K99 DNA sequence and several deletion mutants demonstrated that the K99 genes were divided into three independently regulated regions (Fig. 1) (5). Region 1 encodes fanA to fanD and is dependent on cAMP, CRP, and LRP; region 2 encodes fanE to fanF and is dependent on cAMP and CRP but is independent of LRP; and region 3 encodesfanG to fanH and is independent of cAMP, CRP, and LRP (5) (unpublished data). A series of field isolates of ETEC from pigs that do not produce any of the expected pilus adhesins associated with pig pathogens (K88, K99, or 987P) have been identified. These strains, which have been designated 3P-, produce the pilus adhesin F41 and the 0 antigen 0101 (13). Results obtained from Southern hybridization demonstrated that three of three 3P- strains analyzed contained K99 plasmids (9, 14). Furthermore, restriction endonuclease digestion of these plasmids showed that the 3P- strains retained a 7.15-kb BamHI fragment that hybridized with the cloned K99 genes. These results suggested that large deletions or insertions within the K99 coding region were not responsible for the loss of K99 produc-
assay have failed (13) (unpublished data). Therefore, it was concluded that the 3P- strains, while being genotypically K99+, were, indeed, phenotypically K99-. The objective of the experiments described in this article was to determine the genetic basis for why one of the 3P- strains, VAC-1676, failed to produce K99. It was hypothesized that K99 production might be suppressed because of a specific defect in one (or more) of the K99-specific genes encoded on the K99 plasmid or that there was an alteration in a required host chromosomal gene(s) (e.g., crp or lrp). E. coli VAC-1676, a 3P- strain, was obtained from Harley Moon, National Animal Disease Center (13). This strain is serotype O101:K?,F41, and produces heat-stable enterotoxin. To determine whether the defect causing the suppression of K99 production by E. coli VAC-1676 was due to the host or the K99 plasmid, the 7.15-kb BamHI fragment that encodes K99 in VAC-1676 was cloned into the vector pBR322. The new recombinant plasmid, pIX34, was introduced into E. coli K-12 strain RH202 (thi lacY tonA21 supE44 hssl; obtained from Robert Helling, University of Michigan) by electroporation (Electroporator; Invitrogen, San Diego, Calif.), and recipients were tested for K99 production by a serum agglutination assay that employs rabbit antibodies prepared against purified K99. Recipients did not produce K99, suggesting that one (or more) of the K99 genes on the 7.15-kb BamHI fragment from VAC-1676 was defective. Although expression of K99 by recombinant plasmids is not dependent upon its orientation in the vector (20), to ensure that orientation effects were not affecting these results pIX34 was digested with BamHI, T4 DNA ligase was added, and the ligated DNA was used to transform E. coli RH202. Recombinant plasmids were assessed for the orientation of the 7.15-kb BamHI fragment in the vector by digestion with EcoRI and BglII. Strains containing inserts in each direction were selected, and K99 production was measured by the serum agglutination assay. K99 was not produced, regardless of the orientation of the fragment. To confirm that the host strain VAC-1676 was inherently capable of producing K99, the recombinant plasmid pIX12, which contains the native K99 sequence, was introduced into VAC1676 by electroporation. Recipient cells containing pIX12 did produce K99. Therefore, it was concluded that the reason that VAC-1676 did not produce K99 was that it contained a defective K99 plasmid and was not due to the host strain. The 7.15-kb BamHI fragment contains a single KpnI site, and this site divides the K99 genes into a left-hand fragment of 4.15-kb encoding region 1 genes (fanA to fanD) and a righthand fragment containing region 2 and 3 genes (fanE and -F and fanG and -H, respectively) (Fig. 1). We previously pre-
tion. All attempts to express or detect K99 with the 3P- strains including successive passaging and detection of intracellular subunits by electrophoresis or enzyme-linked immunosorbent Corresponding author. Mailing address: Department of VeteriPathobiology, University of Illinois, 2001 South Lincoln Ave., Urbana, IL 61801. Phone: (217) 333-7825. Fax: (217) 244-7421. Electronic mail address:
[email protected]. *
nary
4686
NOTES
VOL. 62, 1994 Hind
III
pIX12
BamHl Hpal
=,=
kp1pal
HinclI
HI HpalIl
PvuIl B~I
FanA FanB FanC
FanD
FanE
FanF Fanf>
Region 2
Reglon 1
BamHl
4687
in pIX17 and, therefore, was not available to complement pIX31. To determine if the defect was cis active, portions of the 7.15-kb BamHI fragment were selectively replaced with wildtype sequences. pIX34 was digested with Kpnl and Sall, and the 8.15-kb fragment containing the vector and fanA to fanD was isolated by electrophoresis and electroelution. pIX12 also was digested with KpnI and Sall, and the 3.3-kb fragment containing fanE to fanH was isolated. The 8.15-kb fragment from pIX34 was ligated with the 3.3-kb fragment from pIX12 and introduced into E. coli RH202. Similarly, the 4.5-kb HindIII-KpnI fragment encodingfanA tofanD from pIX34 was deleted and replaced with the 4.5-kb HindIII-KpnI fragment from pIX12. Replacement of the fanA-to-fanD region with the wild-type sequence restored the production of K99. To more precisely map the location of the defect, the 1.55-kb HindIII-BglII fragment containingfanA,fanB, and part offanC from pIX34 was replaced with the same fragment from pIX12. This fragment also restored K99 production. Part of this fragment contains the vector sequence and is not important in K99 expression. Therefore, the cis-active defect suppressing K99 production in VAC-1676 was localized to a 1.2-kb sequence (BamHI-BglII) in the K99 genes. To determine the exact map location of the defect and to help determine the nature of the cis-acting defect, the DNA sequences of the 1.55-kb HindIII-BglIl fragment from pIX34 and pIX12 were compared. DNA sequences were determined by a dideoxy chain termination protocol using Sequenase version 2 (United States Biochemicals, Cleveland, Ohio). DNA to be sequenced was prepared by using Promega (Madison, Wis.) Magic columns, and oligonucleotide primers for DNA sequencing were synthesized in the Biotechnology Center, University of Illinois. No DNA sequence deviations were observed in the region encoding FanA to FanC (data not shown). However, in a highly AT-rich region 5' to fanA a single nucleotide change was detected (Fig. 2). The change resulted from the deletion of a single T at nucleotide 345, which is within a stretch of 10 Ts in the wild-type sequence.
Sal
FanH
Region 3
plX17 plX16
FIG. 1. Restriction endonuclease map and organization of the genes within the K99 coding region. Part of the vector pBR322 is shown with relevant restriction sites. The deletion mutants (regions 1 to 3) are indicated. Regions 1 to 3 represent unique transcriptional units.
pared two mutants by deleting the 4.15- or the 3.0-kb KpnIBglII fragment in pIX12 (yielding pIX16 and pIX17 [Fig. 1]). Using Northern blot hybridization analysis, we have previously shown that transcription of the genes encoded by these deletion mutants remains unaltered even though assembled K99 is not produced (5). To locate the general site of the defect, complementation of the defect in trans was attempted by using plasmids containing the left- and right-hand fragments. To accomplish this goal, the 7.15-kb BamHI fragment from VAC1676 was excised from pIX34 and inserted in the vector pACYC184, which is compatible with the vector pBR322 (yielding pIX31). This new plasmid was introduced into E. coli RH202. Separately, pIX17 (left-hand side) or pIX16 (righthand side) was introduced into cells containing pIX31. Clones containing pIX31 and pIX16 or pIX17 failed to produce assembled K99. Therefore, it was concluded that the defect could not be complemented in trans with the left- or the right-hand region of K99. This might indicate that there was a defect in a transcriptional regulatory element that was cis active. There is a small deletion of the C terminus of fanD in the left-hand plasmid pIX17. While this deletion does not affect transcription, it may affect K99 biogenesis in complementation experiments. Thus, an alternative hypothesis is that the defect resided in the C terminus offanD, which was deleted 50
ACGTCTAACGATAACATCCCTGCAATCGTCTLCTGGAGCTACGAAGT
Wild type
ACGTCTAACGATAACATCCCTGCAATCGTCTGCTGGAGCTACGAAGT
M utant
100
AGCATATGATGCTCTTGATAAAATTAGATTTGATTAAAACAGTGACTGTTGATTGCT AGCATATGATGCTCTTGATAAAATTAGATTTGATTAAAACAGTGACTGTTGATTGCT 150
GAAAAGTTAGCCTGATTGGGGTTITTATA1T 1IGTCTTTGTAACAMGAUGG GAAAAGTTAGCCTGATTGGGG I
TI T
IATAITTTTTTTGTCTTTGTAACAMGATTGG
200
TGT 1TGTATCTAAAGCCATACAAAAAACTAACAAAAAACTAACAATTAAAACAAT TGTTTTTTI IGTATCTAAAGCCATACAAAAAACTAACAAAAAACTAACAATTAAAACAAT T
250
TTGAAATTGAAGTAATGGTTGAATTTGTTGTTTTTT^WGTGGCGTTTTTT I GTTGTGAA TTGAAATTTGAAGTAATGGTTTGAATTTGTTGTT 300
350
TTTGATTTTTTTTAGCTTAAAACTCTGGTTCTTCTTGGCTGTTTA TTTGATTTTTTTTAGCTTAAAACTCTGGTTCTTCTTGGCTGTTTA--TT
ITTTTTCTATA IIITCTATA
400
T GTTC AGTG TG TTATTTATACTCTTCCCT-TTATTTTTG T1TTTT-TTATGCCATATAATT
TGTTCAGTGTGTTATTTATACTCTTCCCTTTATTTTTGTTTTTTTTATGCCATATAATT 450
W FanA C AA T C AGC A G AG A TG ATTGGG ATCATAAAAATG TCACTTG AGGGTATATGCG ATCT Q./ f.
ArI CAGC
AG A
TG ATTCG GATCATAAAAATG TCAC TTG AG GGT AT ATGC G ATC T
FIG. 2. DNA sequence of the region 5' tofanA from the wild-type (pIX12) and mutant (pIX34) plasmids. Nucleotide 345 is the site of the single deletion. The portion of the sequencing gel where the mutation was detected is shown on the right. (A) pIX12; (B) pIX34. Arrow in panel A, extra T in the wild-type sequence.
4688
NOTES
Because the identified mutation in pIX34 is proximal to fanA, it was assumed that it affected the expression of region 1 genes. To determine whether the region 1 genes (fanA to fanD) of pIX34 were transcribed, RNA was extracted from E. coli RH202 containing pIX12 or pIX34. Northern blots were subsequently prepared and probed with a radiolabeled, PCRderived fanA probe (5). While the probe hybridized to RNA extracted from the strain containing pIX12 (wild-type K99 genes), the probe failed to hybridize with RNA extracted from the strain containing pIX34 (data not shown). Therefore, it was concluded that the lack of K99 expression in the 3P- strain was a result of inhibiting transcription of region 1 genes. Previously, a group of field isolates of ETEC were recovered from swine that failed to produce one of the three known (at the time) pilus adhesins: K88, K99, or 987P (13). Some of these strains, which were shown to be virulent, were subsequently shown to produce a new pilus adhesin, F41. However, when analyzed by Southern hybridization, some of the 3P- strains were shown to contain K99 plasmids (9, 14). The experiments described in this article were performed to determine the genetic basis of why one 3P- strain (E. coli VAC-1676) did not produce K99. The deletion of a single thymidine in a stretch of 10 thymidines was the only alteration detected in the 3Pstrain compared with the wild-type K99 sequence. This mutation exerts a cis-active effect. The reason that this mutation leads to the loss of K99 production is not known. The fact that this mutation is cis active and that transcription of K99 region 1 genes is inhibited suggests that it is important in transcriptional regulation, perhaps by altering a protein-binding site. Two global regulators, LRP and the cAMP-CRP complex are known to regulate the fanA-to-fanD region of K99 (5). These regulators are believed to function by binding to DNA in regions adjacent to the promoter regulating fanA. This site corresponds to the region where the mutation in VAC-1676 has been mapped. While the mutation in VAC-1676 may affect the binding of LRP or cAMP-CRP to K99-specific DNA, the mutation may affect a new regulatory element. FanA and FanB, for example, are presumed to be positive regulators (1, 15). One or both of these gene products may be involved in self-regulation of the region 1 genes, and the mutated site may be important in the binding of these proteins. The significance of this mutation in disease and disease prevention could be very important. The fact that the loss of K99 production by VAC-1676 did not render it avirulent because a secondary adhesin, F41, was capable of replacing K99 suggests that this may be an important mechanism to evade the host's immune response. Many K99+ ETEC strains also produce F41, and the emergence of F41 and the concurrent suppression of K99 may become common, particularly because vaccination with K99 is a widely employed practice in pig production. In protection studies, K99 has been shown to be dominant over F41. Therefore, protection from infection by a K99+ F41+ ETEC is mediated by antibodies against K99 and not by antibodies against F41. Protection against infection with a K99- F41+ ETEC, however, is not mediated by anti-K99 but is mediated by anti-F41 antibodies. This leads to concerns of whether extensive vaccination with K99 (or other pili) will lead to the emergence of new pilus adhesins. The emergence of 3Pstrains in the field fuels this concern by demonstrating that shifts in pilus adhesins can occur through specific mutations. In addressing the question of the emergence of new pilus adhesins, it should be pointed out that such shifts are not unique to E. coli. For example, Neisseria gonorrhoeae is known to produce a vast array of different pilus adhesins that shift in response to environmental pressures such as immunity (4, 12). Whether the loss of K99 production by 3P- strains results from
INFECT. IMMUN.
ecological pressures and is mediated in all cases by the same mutational event described in this article is unknown. However, experiments similar to the ones described in this article have now been completed with a second 3P- strain (strain 1751), and the same mutation was identified, suggesting that this could be a commonly occurring defect in K99 genes. The region where the mutation in VAC-1676 occurred is within what we believe to be an important regulatory region. This region also is a very AT-rich region (75%). Because there are several stretches of thymidines in this region, a general mechanism of DNA strand slippage during DNA replication that would result in the addition or deletion of a single base pair may be responsible for the loss of K99 expression. Assuming that our hypothesis that this region contains sites that regulate transcription of fanA to fanD is correct, mutations that affect these elements should affect K99 production. Therefore, a variety of different mutations in the AT-rich region could affect the expression of K99. This work was supported by grants from the Competitive Research Grants Office (90-01678) and NRI Competitive Grants program (93-03308), USDA.
REFERENCES 1. Baecker, P. A., E. R. Shelton, H. Bursztyn-Pettegrew, R. Salazar, E. G. Osen, S. D. Stoufer, S. W. Lee, and H. W. Chan. 1988. Expression of K99 adhesion antigen controlled by the Escherichia coli tryptophan operon promoter. Infect. Immun. 56:2317-2323. 2. Bakker, D., C. E. M. Vader, B. Roosendaal, F. R. Mooi, B. Oudega, and F. K. de Graaf. 1991. Structure and function of periplasmic chaperone-like proteins involved in the biosynthesis of K88 and K99 fimbriae in enterotoxigenic Escherichia coli. Mol. Microbiol. 5:875-886. 3. Braaten, B. A., J. V. Platko, M. W. Vanderwoude, B. H. Simons, F. K. de Graaf, J. M. Calvo, and D. A. Low. 1992. Leucineresponsive regulatory protein controls the expression of both the pap and fan pili operons in Escherichia coli. Proc. Natl. Acad. Sci. USA 89:4250-4254. 4. Haas, R., and T. F. Meyer. 1986. The repertoire of silent pilus genes in Neissenia gonorrhoeae: evidence for gene conversion. Cell 44:107-115. 5. Inoue, 0. J., J. H. Lee, and R. E. Isaacson. 1993. Transcriptional organization of the Escherichia coli pilus adhesin-K99. Mol. Microbiol. 10:607-613. 6. Isaacson, R. E. 1980. Factors affecting expression of the Escherichia coli pilus K99. Infect. Immun. 28:190-194. 7. Isaacson, R. E. 1983. Regulation of expression of Eschenchia coli pilus K99. Infect. Immun. 40:633-639. 8. Isaacson, R. E., P. C. Fusco, C. C. Brinton, and H. W. Moon. 1978. In vitro adhesion of Escherichia coli to porcine small intestinal epithelial cells: pili as adhesive factors. Infect. Immun. 21:392-397. 9. Isaacson, R. E., and G. L. Start. 1992. Analysis of K99 plasmids from enterotoxigenic Escherichia coli. FEMS Microbiol. Lett. 90:141-146. 10. Jones, G. W., and R. E. Isaacson. 1983. Proteinaceous bacterial adhesins and their receptors. Crit. Rev. Microbiol. 10:229-260. 11. Kuehn, M. J., D. J. Ogg, J. Kihlberg, L. N. Slonim, K. Flemmer, T. Bergfors, and S. J. Hultgren. 1993. Structural basis of pilus subunit recognition by the PapD chaperone. Science 262:1234-1241. 12. Lambden, P. R. 1982. Biochemical comparison of pili from variants of Neisseria gonorrhoeae P9. J. Gen. Microbiol. 128:21052111. 13. Moon, H. W., E. M. Kohler, R. A. Schneider, and S. C. Whipp. 1980. Prevalence of pilus antigens, enterotoxin types, and enteropathogenicity among K88-negative Escherichia coli from neonatal pigs. Infect. Immun. 27:222-230. 14. Moseley, S. L., G. Dougan, R. A. Schnieder, and H. W. Moon. 1986. Cloning of chromosomal DNA encoding the F41 adhesin of enterotoxigenic Escherichia coli and genetic homology between adhesins F41 and K88. J. Bacteriol. 167:799-804. 15. Roosendaal, E., M. Boots, and F. K. de Graaf. 1987. Two novel
VOL. 62, 1994 fanA and fanB, involved in the biogenesis of K99 fimbriae. Nucleic Acids Res. 15:5973-5984. 16. Roosendaal, B., and F. K. de Graaf. 1989. The nucleotide sequence of the fanD gene encoding the large outer membrane protein involved in the biosynthesis of K99 fimbriae. Nucleic Acids Res. 17:1263. 17. Roosendaal, E., W. Gaastra, and F. K. de Graaf. 1984. The nucleotide sequence of the gene encoding the K99 subunit of enterotoxigenic Eschenichia coli. FEMS Microbiol. Lett. 22:253258. 18. Simons, B. L., P. Rathman, C. R. Malij, B. Oudega, and F. K. de Graaf. 1990. The penultimate tyrosine residue of the K99 fibrillar subunit is essential for stability of the protein and its interaction with the periplasmic carrier protein. FEMS Microbiol. Lett. 76:107-112. genes,
NOTES
4689
19. Simons, B. L., P. T. J. Willemsen, D. Bakker, B. Roosendaal, F. K. de Graaf, and B. Oudega. 1990. Structure, localization and function of FanF, a minor component of K99 fibrillae of enterotoxigenic Escherichia coli. Mol. Microbiol. 4:2041-2050. 20. Teneberg, S., P. T. J. Willemsen, F. K. de Graaf, and K. A. Karlsson. 1993. Calf small intestine receptors for K99 fimbriated enterotoxigenic Escherichia coli. FEMS Microbiol. Lett. 109:107112. 21. van Embden, J. D. A., F. K. de Graaf, L. M. Schouls, and J. S. Teppema. 1980. Cloning and expression of a deoxyribonucleic acid fragment that encodes for the adhesive antigen K99. Infect. Immun. 29:1125-1135. 22. Willemsen, P. T. J., and F. K. de Graaf. 1993. Multivalent binding of K99 fimbriae to the N-glycolyl-GM3 ganglioside receptor. Infect. Immun. 61:4518-4522.
