- Email: [email protected]
Thin aggregative fimbriae enhance Salmonella enteritidis biofilm formation
FEMS Microbiology Letters 162 (1998) 295^301
Thin aggregative ¢mbriae enhance Salmonella enteritidis bio¢lm formation John W. Austin a , Greg Sanders a , William W. Kay b; *, S. Karen Collinson a b
b
Bureau of Microbial Hazards, Food Directorate, Health Protection Branch, Health Canada, Ottawa, Ont. K1A 0L2, Canada Department of Biochemistry and Microbiology and the Canadian Bacterial Diseases Network, P.O. Box 3055, Petch Building, University of Victoria, Victoria, B.C. V8W 3P6, Canada Received 5 January 1998; revised 18 March 1998; accepted 22 March 1998
Abstract Salmonella enteritidis enteropathogens produce a variety of potentially adherent fimbrial types including SEF14, SEF17, SEF18 and SEF21 (type I). In a simplified, pure culture, biofilm generating system the virulent isolate, S. enteritidis 3b, readily adhered to Teflon (polytetrafluoroethylene) and stainless steel forming thick cell aggregates. The inability of an isogenic SEF17-deficient mutant to form thick biofilms suggested a role for SEF17 in stabilizing cell-cell interactions during biofilm formation. Epifluorescent detection of SEF17 in biofilms confirmed the association of these fimbriae with aggregated cells but not with adherent mutants unable to produce SEF17. The reduced adherence observed with an isogenic SEF14/SEF21deficient strain implicated the involvement of additional cell surface adherence factors, possibly including SEF21 (type I) fimbriae in the adherence of S. enteritidis to stainless steel or Teflon. The role of SEF17 fimbriae in biofilm formation and the contributions of SEF17 to the persistence of Salmonellae on surfaces and in food are discussed. z 1998 Published by Elsevier Science B.V. Keywords : Salmonella ; Thin aggregative ¢mbriae ; SEF17; SEF14; Type I ¢mbriae (SEF21) ; Bio¢lm ; Food industry
1. Introduction Salmonella enteritidis (S.e.) is an important foodborne enteric pathogen [1] and, in Canada, is the second most commonly isolated serovar after S. typhimurium [2]. Like some other bacterial food-borne pathogens, Salmonella are able to colonize inert food contact surfaces to form bio¢lms [3^5]. Understand-
* Corresponding author. Tel.: +1 (250) 721-7078; Fax: +1 (250) 721-8882; E-mail: [email protected]
ing the factors and biological processes involved in the establishment and development of surface adherent microbial communities that constitute bio¢lms [6] is of fundamental and practical signi¢cance since bio¢lms containing pathogens can pose signi¢cant health risks and economic problems in the food industry [7,8]. Little is known of the molecular mechanisms of Salmonella adherence and bio¢lm formation on inert surfaces, but the actual Salmonella cell surface components involved might be expected to include ¢mbriae or pili, £agella, lipopolysaccharide, and/or capsules or slime layers [9]. Previously we observed ¢mbriated strains of S.e.
0378-1097 / 98 / $19.00 ß 1998 Published by Elsevier Science B.V. All rights reserved. PII S 0 3 7 8 - 1 0 9 7 ( 9 8 ) 0 0 1 3 7 - 2
FEMSLE 8147 7-5-98
296
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
to be unusually autoaggregative, forming abundant pellicles and adhering to a variety of culture surfaces [10,11]. This suggested a possible role for these peripheral cellular organelles in bio¢lm formation and as factors important in the persistence of infectious cells in various environments. To determine whether certain ¢mbriae from S.e. play a role in bio¢lm formation on the commercially important materials, Te£on and stainless steel, we compared bio¢lm formation of two ¢mbrial mutants to the parent strain in a simple, pure culture, bio¢lm generating model system. One mutant was de¢cient in production of thin aggregative ¢mbriae (SEF17) [10], whereas the other was unable to produce both the O-serogroup D1 speci¢c SEF14 ¢mbriae and the type I ¢mbriae, SEF21 [12]. The results establish for the ¢rst time an important environmental role for SEF17 ¢mbriae. Preliminary results of this study were presented at the Canadian Society of Microbiologists Annual General Meeting, 11^15 June 1996, Charlottetown, P.E.I.
2. Materials and methods 2.1. Bacterial strains and growth conditions The phenotypes of the S.e. 3b human isolate [11] and isogenic ¢mbrial mutants, 3b 2-2a [10] and 3b122 [12] are presented in Table 1. Stock cultures were stored at 380³C on ceramic beads (Prolab Diagnostics, Ontario). Strains were routinely grown from stock cultures on trypticase soy agar (TSA) at 37³C overnight. Cultures and bio¢lms were grown in CFA broth [10] at 22³C, 50 rpm for 20 h unless stated otherwise. 2.2. Bio¢lm generation Liquid seed cultures inoculated from freshly grown TSA plates were grown in £asks containing CFA broth (50 ml). Plastic petri dishes (14 cm diameter) containing 54 ml CFA broth were inoculated with the seed culture (10% v/v). Stainless steel coupons (1 cm2 , type 304, #4 ¢nish) or polytetra£uoroethylene surfaces (1 cm2 , Te£on1) were submerged in the inoculated media and then the petri plate cultures were incubated at 50 rpm as above.
2.3. Quantitation of S. enteritidis adherence in bio¢lms To measure bacterial adherence on the stainless steel and Te£on surfaces, a modi¢ed dye binding method was used [13]. Brie£y, sample surfaces incubated with bacteria for the speci¢ed time were gently rinsed in 20 ml of phosphate bu¡ered saline (PBS) for several minutes. The surfaces, with adherent cells, were stained for 5 min with 0.25% (w/v) Coomassie blue G-250 (Bio-Rad) dissolved in 50% ethanol and 7% acetic acid. The stained surfaces were rinsed for 5 min in 20 ml of PBS twice and then cell-associated dye was released from the adherent cells with 2% (w/ v) sodium deoxycholate in distilled water. The amount of cell-associated Coomassie blue released was measured spectrophotometrically at 610 nm. ANOVA and 95% con¢dence intervals were calculated using Prism version 2 (GraphPad Software Inc., San Diego, CA). 2.4. Scanning electron microscopy (SEM) To visualize S.e. bio¢lms generated on Te£on and stainless steel surfaces, the bacteria-coated surfaces were removed from the broth cultures and ¢xed in 0.2 M cacodylate bu¡er pH 7.4 containing 2.5% (w/ v) glutaraldehyde. Sample preparation for SEM was as described by Austin and Bergeron [8]. Specimens were sputter-coated with 30 nm of platinum and viewed in a Vickers Nanolab LE2100 SEM operated at an accelerating voltage of 5 or 15 kV. 2.5. SDS-PAGE and Western immunoblot analysis of S. enteritidis ¢mbriae S.e. bio¢lm cells were scraped from the Te£on and stainless steel surfaces using glass slides. Whole cell lysates of approximately 4 OD units at 650 nm (approx. 6U109 cells) were used to prepare ¢mbriae for SDS-PAGE and Western immunoblotting as previously described [10,14]. 2.6. Immuno£uorescence light microscopy S.e. bio¢lms adherent to stainless steel and Te£on samples were removed from the cultures, rinsed in PBS and then blocked in PBS supplemented with
FEMSLE 8147 7-5-98
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
297
0.1% (w/v) bovine serum albumin. Bio¢lm-covered surfaces were incubated for 1 h with speci¢c rabbit antiserum to label SEF17. After three washes with PBS, samples were incubated with goat anti-rabbit IgG conjugated to £uorescein isothiocyanate (FITC) for 30 min, then ¢nally washed three times with PBS. Adherent cells were visualized using a Zeiss light microscope (Zeiss Axiophot) equipped with epi£uorescent imaging (546 nm excitation wavelength, 450^ 490 nm excitation wavelength).
3. Results 3.1. Quantitation of S. enteritidis in bio¢lms The relative binding of S.e. 3b and the ¢mbrial mutants grown on stainless steel and Te£on surfaces was determined using a dye binding assay [13]. S.e. 3b adherence on both surfaces was signi¢cantly higher than either the SEF17-de¢cient 3b 2-2a or the SEF14/SEF21 double mutant, 3b-122 (P 6 0.001) (Fig. 1). Adherence of 3b-122 cells was signi¢cantly higher on stainless steel than that of 3b 2-2a (P 6 0.01). However, 3b-122 adherence to Te£on was variable and not always signi¢cantly higher than 3b 2-2a (P s 0.05). 3.2. SEM of bio¢lm formation by S. enteritidis Bio¢lms produced on stainless steel and Te£on surfaces by S.e. 3b and the two ¢mbrial mutants were observed by SEM to determine the microscopic
Fig. 1. Adherence of S.e. 3b, 3b-122 and 3b 2-2a to stainless steel and Te£on at 22³C. Each value represents an average of 13 replicate samples minus the average negative control OD610 values of 0.061 or 0.051 determined for cell-free stainless steel or Te£on surfaces, respectively. Error bars indicate the 95% con¢dence intervals.
cellular arrangements on these surfaces. By 24 h S.e. 3b had colonized the stainless steel surfaces forming extensive aggregates of cells (Fig. 2A). The mutant 3b-122, unable to produce SEF14 and type 1 (SEF21) ¢mbriae, also adhered to stainless steel but was limited in its ability to form aggregates of cells (Fig. 2B). The SEF17-de¢cient strain 3b 2-2a
Table 1 Western immunoblot detection of ¢mbriae on S. enteritidis bio¢lm cellsa S. enteritidis strain
3b 3b-122 3b 2-2a
Fimbrins producedb
Genotype
parent S.e. 27655-3b [11] ¢mU promoter Tn10 insertion [12] agfA: :TnphoA [10]
a
AgfA
SefAc
FimAd
SefD
+ + 3
3 3 3
+ 3 +
+ + +
Western immunoblots were performed on samples collected from three separate experiments. The major ¢mbrial subunits AgfA, SefA, FimA and SefD comprise SEF17, SEF14, SEF21 and SEF18 ¢mbriae, respectively [11,12,14]. Fimbrin subunits separated by SDS-PAGE were detected with speci¢c rabbit antisera on Western immunoblots for samples prepared from adherent cells as described in Section 2. +, detected ; 3, not detected. c SefA (SEF14) are not synthesized when cells are grown at temperatures less than 28³C [14]. d FimA typically runs on SDS-PAGE as a doublet due to incomplete acid hydrolysis of the ¢rst two N-terminal amino acids during sample preparation. b
FEMSLE 8147 7-5-98
298
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
Fig. 2. Scanning electron micrographs of S.e. bio¢lms on stainless steel. Bio¢lms of S.e. 3b (A), S.e. 3b-122 (B) and S.e. 3b 22a (C) were grown for 24 h as outlined in Section 2. The bars in A and B represent 2 Wm; the bar in C represents 4 Wm.
Fig. 3. Scanning electron micrographs showing an enlargement of S.e. cells attached to stainless steel. A: S.e. 3b. B: S.e. 3b122. C: S.e. 3b 2-2a. The bars represent 1 Wm.
FEMSLE 8147 7-5-98
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
299
Fig. 4. Western immunoblot analysis of FimA (A) or AgfA (B) produced by S.e. bio¢lms formed on stainless steel (ss) or Te£on (t) by S.e. 3b (3b), S.e. 3b-122 (122), or S.e. 3b 2-2a (2-2a). Of the prestained protein standards (BRL, Bethesda, MD) loaded in lane labeled Mr, lysozyme is visible at 14.3 kDa.
also colonized the stainless steel but only di¡usely. Single cells were distributed over the surface without the formation of cell aggregates (Fig. 2C). Similar results were seen for bio¢lms of the three S.e. strains grown on Te£on (data not shown). SEM observation of the bio¢lms at higher magni¢cation showed a meshwork of ¢bers associated with S.e. 3b and 3b-122 cells grown on stainless steel (Fig. 3A,B) but not with 3b 2-2a cells unable to produce SEF17 (Fig. 3C). Some ¢bers appeared to form an interconnecting mesh between cells, where as others appeared to attach cells to the substratum. S.e. 3b122 demonstrated occasional ¢laments interconnecting cells and also produced ¢bers which held the cells to the substratum (Fig. 3B). S.e. 3b 2-2a did not produce a ¢ber meshwork interconnecting cells but the occasional ¢ber apparently attaching cells to the substratum was present (Fig. 3C). Similar results were seen for the three strains grown on Te£on (data not shown). 3.3. Fimbrial production by S. enteritidis bio¢lm cells The status of the S.e. strains with respect to ¢mbriae produced was determined by detection of ¢mbrial proteins on Western immunoblots prepared from adherent cells. SEF17 was produced by adherent S.e. 3b and 3b-122 but not by the SEF17-de¢cient 3b 2-2a mutant (Fig. 4, Table 1). All strains produced SEF18 but they did not produce SEF14 (Table 1). SEF21 (type 1) ¢mbriae were weakly de-
Fig. 5. Epi£uorescent microscopy of S.e. 3b (A), S.e. 3b-122 (B) and S.e. 3b 2-2a (C) labeled with antiserum speci¢c for SEF17 as outlined in Section 2. The bars represent approximately 10 Wm.
FEMSLE 8147 7-5-98
300
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
tected as a doublet band in S.e. 3b and 3b 2-2a but were not detected in 3b-122 (Fig. 4, Table 1). 3.4. Epi£uorescent detection of SEF17 in bio¢lm cells Adherent cells of S.e. were labeled with antibody to SEF17 to determine the presence of SEF17 in bio¢lms. When bio¢lms of S.e. 3b and 3b-122 cells adherent to stainless steel were probed with antiSEF17 polyclonal sera, material associated with the cell aggregates was labeled with FITC-conjugated second antibody (Fig. 5A,B). Conversely, similar FITC-labeled material was not associated with the SEF17-de¢cient 3b 2-2a mutant adherent on stainless steel (Fig. 5C). Thus, SEF17 was associated with cell aggregates in the S.e. bio¢lms grown on stainless steel.
4. Discussion This study demonstrates for the ¢rst time that thin aggregative ¢mbriae (SEF17) of S.e. are important in bio¢lm formation on inert surfaces. Since SEF17producing S.e. 3b and 3b-122 produced substantially thicker layers of cells than the SEF17-de¢cient mutant 3b 2-2a, it appears that SEF17 functions to stabilize intracellular contacts allowing aggregation to occur and thereby increasing the bio¢lm thickness. This role for SEF17 in bio¢lm formation is consistent with previous studies. In S.e., these ¢mbriae are responsible for the characteristic sti¡ colony phenotype, cell clumping, and pellicle formation on liquid medium surfaces [10,11] and cause convoluted colony morphology in Salmonella typhimurium [15]. Various pathogenic and non-pathogenic bacteria similarly use ¢mbriae to reinforce intra-bacterial cell contact including: type 4 pili of Neisseria [16,17], type 2 ¢mbriae of dental plaque-forming Actinomyces-Streptococcus coaggregates [18] and conjugative pili between bacterial mating pairs. Like thin aggregative ¢mbriae of Salmonella, E. coli curli likely play a similar role in bio¢lm formation since curliproducing E. coli also form sti¡ colony phenotypes and cell clumping [19] or convoluted colonies [15]. In addition to the intracellular ¢ber networks, S.e. 3b appear in some cases to be bound to the substratum. It is not known if SEF17 bind directly to stain-
less steel and Te£on surfaces however in vitro aggregates of puri¢ed ¢mbriae also readily adhere to polypropylene, glass and metal (Collinson, unpublished observations). Other S.e. cell surface component(s), apart from SEF17, are required for some aspect of bio¢lm formation since SEF17-de¢cient 3b 2-2a adhered somewhat to surfaces. Reduced adherence by 3b-122 compared to the parent S.e. strain 3b on stainless steel and Te£on may have been due to the absence of SEF21 or possibly an as yet uncharacterized cell surface component also a¡ected by the ¢mU promoter Tn10 insertion [12]. Further studies with SEF18 isogenic mutants are required to establish whether these ¢mbriae a¡ect adherence, however such mutants are currently unavailable. S.e. bio¢lm formation on inert surfaces is likely complex, not only involving the interplay of various bacterial cell surface components expressed by cells in a given environment but also the nature of surface adsorbed solvent molecules and physicochemical properties of the inert surface itself [6,9]. S.e. not only adheres to physically very di¡erent hydrophobic (Te£on) and hydrophilic (stainless steel) surfaces common in the food industry, but SEF17 ¢mbriae actually promote the buildup of thick bio¢lms in the pure culture bio¢lm system studied herein. Several important implications arise from this study which are consistent with the current understanding of bio¢lm developmental biology and the character of adherent microbial populations [6,9]. Clumps of aggregated S.e. were readily sloughed o¡ the stainless steel and Te£on which suggest the potential contamination of food by Salmonella aggregates during food processing as an obvious health hazard. In addition, the recalcitrance of SEF17 to certain detergents, chaotropic agents, acids and heat [11] raises the question as to whether these ¢mbriae also protect Salmonella aggregates in bio¢lms during equipment sanitization. Conceivably, residual SEF17-producing S.e., whether alive or killed, remaining on sanitized surfaces would also facilitate S.e. recolonization. Indeed, a recent Salmonella outbreak was linked to residual bacteria on blender blades apparently una¡ected by routine sanitizing procedures [20]. Salmonella aggregates in food would also result in under estimation of bacterial load if standard cell counting procedures did not disperse the cell aggregates. Further studies will be needed
FEMSLE 8147 7-5-98
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
to establish whether these thick Salmonella bio¢lms are resistant to sanitizers and other antimicrobial agents, as is often observed with various bio¢lms [4,6,7,9]. If so, knowledge of the Salmonella surface components contributing to bio¢lm development and resistance will lead to the development of more effective sanitizers.
Acknowledgments This work was supported by the Natural Sciences and Engineering Research Council of Canada and the Canadian Bacterial Diseases Network. We thank Dr. S.C. Clouthier for helpful discussions and assisting with the Western immunoblots and Dr. J. Farber for proofreading the manuscript.
References [1] D'Aoust, J.-Y. (1997) Salmonella species. In: Food Microbiology. Fundamentals and Frontiers (Doyle, M.P., Beuchat, L.R. and Montville, T.J., Eds.), pp. 129^158. ASM Press, Washington, DC. [2] Poppe, C. (1994) Salmonella enteritidis in Canada. Int. J. Food Microbiol. 21, 1^5. [3] Jones, K. and Bradshaw, S.B. (1996) Bio¢lm formation by the enterobacteriaceae : A comparison between Salmonella enteritidis, Escherichia coli and a nitrogen-¢xing strain of Klebsiella pneumoniae. J. Appl. Bacteriol. 80, 458^464. [4] Dhir, V.K. and Dodd, C.E. (1995) Susceptibility of suspended and surface-attached Salmonella enteritidis to biocides and elevated temperatures. Appl. Environ. Microbiol. 61, 1731^ 1738. [5] Helke, D.M. and Wong, A.C.L. (1994) Survival and growth characteristics of Listeria monocytogenes and Salmonella typhimurium on stainless steel and buna-N rubber. J. Food Protect. 57, 963. [6] Palmer, Jr., R.J. and White, D.C. (1997) Developmental biology of bio¢lms: implications for treatment and control. Trends Microbiol. 5, 435^440. [7] Carpentier, B. and Cerf, O. (1993) Bio¢lms and their consequences, with particular reference to hygiene in the food industry. J. Appl. Bacteriol. 75, 499^511.
301
[8] Austin, J.W. and Bergeron, G. (1995) Development of bacterial bio¢lms in dairy processing lines. J. Dairy Res. 62, 509^ 519. [9] Fletcher, M. (1996) Bacterial attachment in aquatic environments: A diversity of surfaces and adhesion strategies. In: Bacterial Adhesion (Fletcher, M., Ed.), pp. 1^24. Wiley-Liss, New York. [10] Collinson, S.K., Doig, P.C., Doran, J.L., Clouthier, S.C., Trust, T.J. and Kay, W.W. (1993) Thin, aggregative ¢mbriae mediate binding of Salmonella enteritidis to ¢bronectin. J. Bacteriol. 175, 12^18. [11] Collinson, S.K., Emoëdy, L., Muëller, K.H., Trust, T.J. and Kay, W.W. (1991) Puri¢cation and characterization of thin, aggregative ¢mbriae from Salmonella enteritidis. J. Bacteriol. 173, 4773^4781. [12] Clouthier, S.C., Collinson, S.K., White, A.P., Banser, P.A. and Kay, W.W. (1998) tRNAarg (¢mU) and expression of SEF14 and SEF21 in Salmonella enteritidis. J. Bacteriol. 180, 840^845. [13] Sonak, S. and Bhosle, N.B. (1995) A simple method to assess bacterial attachment to surfaces. Biofouling 9, 31^38. [14] Clouthier, S.C., Collinson, S.K. and Kay, W.W. (1994) Unique ¢mbriae-like structures encoded by sefD of the SEF14 ¢mbrial gene cluster of Salmonella enteritidis. Mol. Microbiol. 12, 893^901. [15] Sukupolvi, S., Lorenz, R.G., Gordon, J.I., Bian, Z., Pfeifer, J.D., Normark, S.J. and Rhen, M. (1997) Expression of thin aggregative ¢mbriae promotes interaction of Salmonella typhimurium SR-11 with mouse small intestinal epithelial cells. Infect. Immun. 65, 5320^5325. [16] Dekker, N.P., Lammel, C.J. and Brooks, G.F. (1991) Scanning electron microscopy of piliated Neisseria gonorrhoeae processed with hexamethyldisilazane. J. Electron Microsc. Tech. 19, 461^467. [17] Marceau, M., Beretti, J.L. and Nassif, X. (1995) High adhesiveness of encapsulated Neisseria meningitidis to epithelial cells is associated with the formation of bundles of pili. Mol. Microbiol. 17, 855^863. [18] Whittaker, C.J., Klier, C.M. and Kolenbrander, P.E. (1996) Mechanisms of adhesion by oral bacteria. Annu. Rev. Microbiol. 50, 513^552. [19] Collinson, S.K., Emoëdy, L., Trust, T.J. and Kay, W.W. (1992) Thin aggregative ¢mbriae from diarrheagenic Escherichia coli. J. Bacteriol. 174, 4490^4495. [20] Anonymous (1992) Hospital outbreak of Salmonella enteritidis infection ^ Ontario. Canada Communicable Disease Report 18-8, pp. 57^60.
FEMSLE 8147 7-5-98
Thin aggregative ¢mbriae enhance Salmonella enteritidis bio¢lm formation John W. Austin a , Greg Sanders a , William W. Kay b; *, S. Karen Collinson a b
b
Bureau of Microbial Hazards, Food Directorate, Health Protection Branch, Health Canada, Ottawa, Ont. K1A 0L2, Canada Department of Biochemistry and Microbiology and the Canadian Bacterial Diseases Network, P.O. Box 3055, Petch Building, University of Victoria, Victoria, B.C. V8W 3P6, Canada Received 5 January 1998; revised 18 March 1998; accepted 22 March 1998
Abstract Salmonella enteritidis enteropathogens produce a variety of potentially adherent fimbrial types including SEF14, SEF17, SEF18 and SEF21 (type I). In a simplified, pure culture, biofilm generating system the virulent isolate, S. enteritidis 3b, readily adhered to Teflon (polytetrafluoroethylene) and stainless steel forming thick cell aggregates. The inability of an isogenic SEF17-deficient mutant to form thick biofilms suggested a role for SEF17 in stabilizing cell-cell interactions during biofilm formation. Epifluorescent detection of SEF17 in biofilms confirmed the association of these fimbriae with aggregated cells but not with adherent mutants unable to produce SEF17. The reduced adherence observed with an isogenic SEF14/SEF21deficient strain implicated the involvement of additional cell surface adherence factors, possibly including SEF21 (type I) fimbriae in the adherence of S. enteritidis to stainless steel or Teflon. The role of SEF17 fimbriae in biofilm formation and the contributions of SEF17 to the persistence of Salmonellae on surfaces and in food are discussed. z 1998 Published by Elsevier Science B.V. Keywords : Salmonella ; Thin aggregative ¢mbriae ; SEF17; SEF14; Type I ¢mbriae (SEF21) ; Bio¢lm ; Food industry
1. Introduction Salmonella enteritidis (S.e.) is an important foodborne enteric pathogen [1] and, in Canada, is the second most commonly isolated serovar after S. typhimurium [2]. Like some other bacterial food-borne pathogens, Salmonella are able to colonize inert food contact surfaces to form bio¢lms [3^5]. Understand-
* Corresponding author. Tel.: +1 (250) 721-7078; Fax: +1 (250) 721-8882; E-mail: [email protected]
ing the factors and biological processes involved in the establishment and development of surface adherent microbial communities that constitute bio¢lms [6] is of fundamental and practical signi¢cance since bio¢lms containing pathogens can pose signi¢cant health risks and economic problems in the food industry [7,8]. Little is known of the molecular mechanisms of Salmonella adherence and bio¢lm formation on inert surfaces, but the actual Salmonella cell surface components involved might be expected to include ¢mbriae or pili, £agella, lipopolysaccharide, and/or capsules or slime layers [9]. Previously we observed ¢mbriated strains of S.e.
0378-1097 / 98 / $19.00 ß 1998 Published by Elsevier Science B.V. All rights reserved. PII S 0 3 7 8 - 1 0 9 7 ( 9 8 ) 0 0 1 3 7 - 2
FEMSLE 8147 7-5-98
296
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
to be unusually autoaggregative, forming abundant pellicles and adhering to a variety of culture surfaces [10,11]. This suggested a possible role for these peripheral cellular organelles in bio¢lm formation and as factors important in the persistence of infectious cells in various environments. To determine whether certain ¢mbriae from S.e. play a role in bio¢lm formation on the commercially important materials, Te£on and stainless steel, we compared bio¢lm formation of two ¢mbrial mutants to the parent strain in a simple, pure culture, bio¢lm generating model system. One mutant was de¢cient in production of thin aggregative ¢mbriae (SEF17) [10], whereas the other was unable to produce both the O-serogroup D1 speci¢c SEF14 ¢mbriae and the type I ¢mbriae, SEF21 [12]. The results establish for the ¢rst time an important environmental role for SEF17 ¢mbriae. Preliminary results of this study were presented at the Canadian Society of Microbiologists Annual General Meeting, 11^15 June 1996, Charlottetown, P.E.I.
2. Materials and methods 2.1. Bacterial strains and growth conditions The phenotypes of the S.e. 3b human isolate [11] and isogenic ¢mbrial mutants, 3b 2-2a [10] and 3b122 [12] are presented in Table 1. Stock cultures were stored at 380³C on ceramic beads (Prolab Diagnostics, Ontario). Strains were routinely grown from stock cultures on trypticase soy agar (TSA) at 37³C overnight. Cultures and bio¢lms were grown in CFA broth [10] at 22³C, 50 rpm for 20 h unless stated otherwise. 2.2. Bio¢lm generation Liquid seed cultures inoculated from freshly grown TSA plates were grown in £asks containing CFA broth (50 ml). Plastic petri dishes (14 cm diameter) containing 54 ml CFA broth were inoculated with the seed culture (10% v/v). Stainless steel coupons (1 cm2 , type 304, #4 ¢nish) or polytetra£uoroethylene surfaces (1 cm2 , Te£on1) were submerged in the inoculated media and then the petri plate cultures were incubated at 50 rpm as above.
2.3. Quantitation of S. enteritidis adherence in bio¢lms To measure bacterial adherence on the stainless steel and Te£on surfaces, a modi¢ed dye binding method was used [13]. Brie£y, sample surfaces incubated with bacteria for the speci¢ed time were gently rinsed in 20 ml of phosphate bu¡ered saline (PBS) for several minutes. The surfaces, with adherent cells, were stained for 5 min with 0.25% (w/v) Coomassie blue G-250 (Bio-Rad) dissolved in 50% ethanol and 7% acetic acid. The stained surfaces were rinsed for 5 min in 20 ml of PBS twice and then cell-associated dye was released from the adherent cells with 2% (w/ v) sodium deoxycholate in distilled water. The amount of cell-associated Coomassie blue released was measured spectrophotometrically at 610 nm. ANOVA and 95% con¢dence intervals were calculated using Prism version 2 (GraphPad Software Inc., San Diego, CA). 2.4. Scanning electron microscopy (SEM) To visualize S.e. bio¢lms generated on Te£on and stainless steel surfaces, the bacteria-coated surfaces were removed from the broth cultures and ¢xed in 0.2 M cacodylate bu¡er pH 7.4 containing 2.5% (w/ v) glutaraldehyde. Sample preparation for SEM was as described by Austin and Bergeron [8]. Specimens were sputter-coated with 30 nm of platinum and viewed in a Vickers Nanolab LE2100 SEM operated at an accelerating voltage of 5 or 15 kV. 2.5. SDS-PAGE and Western immunoblot analysis of S. enteritidis ¢mbriae S.e. bio¢lm cells were scraped from the Te£on and stainless steel surfaces using glass slides. Whole cell lysates of approximately 4 OD units at 650 nm (approx. 6U109 cells) were used to prepare ¢mbriae for SDS-PAGE and Western immunoblotting as previously described [10,14]. 2.6. Immuno£uorescence light microscopy S.e. bio¢lms adherent to stainless steel and Te£on samples were removed from the cultures, rinsed in PBS and then blocked in PBS supplemented with
FEMSLE 8147 7-5-98
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
297
0.1% (w/v) bovine serum albumin. Bio¢lm-covered surfaces were incubated for 1 h with speci¢c rabbit antiserum to label SEF17. After three washes with PBS, samples were incubated with goat anti-rabbit IgG conjugated to £uorescein isothiocyanate (FITC) for 30 min, then ¢nally washed three times with PBS. Adherent cells were visualized using a Zeiss light microscope (Zeiss Axiophot) equipped with epi£uorescent imaging (546 nm excitation wavelength, 450^ 490 nm excitation wavelength).
3. Results 3.1. Quantitation of S. enteritidis in bio¢lms The relative binding of S.e. 3b and the ¢mbrial mutants grown on stainless steel and Te£on surfaces was determined using a dye binding assay [13]. S.e. 3b adherence on both surfaces was signi¢cantly higher than either the SEF17-de¢cient 3b 2-2a or the SEF14/SEF21 double mutant, 3b-122 (P 6 0.001) (Fig. 1). Adherence of 3b-122 cells was signi¢cantly higher on stainless steel than that of 3b 2-2a (P 6 0.01). However, 3b-122 adherence to Te£on was variable and not always signi¢cantly higher than 3b 2-2a (P s 0.05). 3.2. SEM of bio¢lm formation by S. enteritidis Bio¢lms produced on stainless steel and Te£on surfaces by S.e. 3b and the two ¢mbrial mutants were observed by SEM to determine the microscopic
Fig. 1. Adherence of S.e. 3b, 3b-122 and 3b 2-2a to stainless steel and Te£on at 22³C. Each value represents an average of 13 replicate samples minus the average negative control OD610 values of 0.061 or 0.051 determined for cell-free stainless steel or Te£on surfaces, respectively. Error bars indicate the 95% con¢dence intervals.
cellular arrangements on these surfaces. By 24 h S.e. 3b had colonized the stainless steel surfaces forming extensive aggregates of cells (Fig. 2A). The mutant 3b-122, unable to produce SEF14 and type 1 (SEF21) ¢mbriae, also adhered to stainless steel but was limited in its ability to form aggregates of cells (Fig. 2B). The SEF17-de¢cient strain 3b 2-2a
Table 1 Western immunoblot detection of ¢mbriae on S. enteritidis bio¢lm cellsa S. enteritidis strain
3b 3b-122 3b 2-2a
Fimbrins producedb
Genotype
parent S.e. 27655-3b [11] ¢mU promoter Tn10 insertion [12] agfA: :TnphoA [10]
a
AgfA
SefAc
FimAd
SefD
+ + 3
3 3 3
+ 3 +
+ + +
Western immunoblots were performed on samples collected from three separate experiments. The major ¢mbrial subunits AgfA, SefA, FimA and SefD comprise SEF17, SEF14, SEF21 and SEF18 ¢mbriae, respectively [11,12,14]. Fimbrin subunits separated by SDS-PAGE were detected with speci¢c rabbit antisera on Western immunoblots for samples prepared from adherent cells as described in Section 2. +, detected ; 3, not detected. c SefA (SEF14) are not synthesized when cells are grown at temperatures less than 28³C [14]. d FimA typically runs on SDS-PAGE as a doublet due to incomplete acid hydrolysis of the ¢rst two N-terminal amino acids during sample preparation. b
FEMSLE 8147 7-5-98
298
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
Fig. 2. Scanning electron micrographs of S.e. bio¢lms on stainless steel. Bio¢lms of S.e. 3b (A), S.e. 3b-122 (B) and S.e. 3b 22a (C) were grown for 24 h as outlined in Section 2. The bars in A and B represent 2 Wm; the bar in C represents 4 Wm.
Fig. 3. Scanning electron micrographs showing an enlargement of S.e. cells attached to stainless steel. A: S.e. 3b. B: S.e. 3b122. C: S.e. 3b 2-2a. The bars represent 1 Wm.
FEMSLE 8147 7-5-98
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
299
Fig. 4. Western immunoblot analysis of FimA (A) or AgfA (B) produced by S.e. bio¢lms formed on stainless steel (ss) or Te£on (t) by S.e. 3b (3b), S.e. 3b-122 (122), or S.e. 3b 2-2a (2-2a). Of the prestained protein standards (BRL, Bethesda, MD) loaded in lane labeled Mr, lysozyme is visible at 14.3 kDa.
also colonized the stainless steel but only di¡usely. Single cells were distributed over the surface without the formation of cell aggregates (Fig. 2C). Similar results were seen for bio¢lms of the three S.e. strains grown on Te£on (data not shown). SEM observation of the bio¢lms at higher magni¢cation showed a meshwork of ¢bers associated with S.e. 3b and 3b-122 cells grown on stainless steel (Fig. 3A,B) but not with 3b 2-2a cells unable to produce SEF17 (Fig. 3C). Some ¢bers appeared to form an interconnecting mesh between cells, where as others appeared to attach cells to the substratum. S.e. 3b122 demonstrated occasional ¢laments interconnecting cells and also produced ¢bers which held the cells to the substratum (Fig. 3B). S.e. 3b 2-2a did not produce a ¢ber meshwork interconnecting cells but the occasional ¢ber apparently attaching cells to the substratum was present (Fig. 3C). Similar results were seen for the three strains grown on Te£on (data not shown). 3.3. Fimbrial production by S. enteritidis bio¢lm cells The status of the S.e. strains with respect to ¢mbriae produced was determined by detection of ¢mbrial proteins on Western immunoblots prepared from adherent cells. SEF17 was produced by adherent S.e. 3b and 3b-122 but not by the SEF17-de¢cient 3b 2-2a mutant (Fig. 4, Table 1). All strains produced SEF18 but they did not produce SEF14 (Table 1). SEF21 (type 1) ¢mbriae were weakly de-
Fig. 5. Epi£uorescent microscopy of S.e. 3b (A), S.e. 3b-122 (B) and S.e. 3b 2-2a (C) labeled with antiserum speci¢c for SEF17 as outlined in Section 2. The bars represent approximately 10 Wm.
FEMSLE 8147 7-5-98
300
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
tected as a doublet band in S.e. 3b and 3b 2-2a but were not detected in 3b-122 (Fig. 4, Table 1). 3.4. Epi£uorescent detection of SEF17 in bio¢lm cells Adherent cells of S.e. were labeled with antibody to SEF17 to determine the presence of SEF17 in bio¢lms. When bio¢lms of S.e. 3b and 3b-122 cells adherent to stainless steel were probed with antiSEF17 polyclonal sera, material associated with the cell aggregates was labeled with FITC-conjugated second antibody (Fig. 5A,B). Conversely, similar FITC-labeled material was not associated with the SEF17-de¢cient 3b 2-2a mutant adherent on stainless steel (Fig. 5C). Thus, SEF17 was associated with cell aggregates in the S.e. bio¢lms grown on stainless steel.
4. Discussion This study demonstrates for the ¢rst time that thin aggregative ¢mbriae (SEF17) of S.e. are important in bio¢lm formation on inert surfaces. Since SEF17producing S.e. 3b and 3b-122 produced substantially thicker layers of cells than the SEF17-de¢cient mutant 3b 2-2a, it appears that SEF17 functions to stabilize intracellular contacts allowing aggregation to occur and thereby increasing the bio¢lm thickness. This role for SEF17 in bio¢lm formation is consistent with previous studies. In S.e., these ¢mbriae are responsible for the characteristic sti¡ colony phenotype, cell clumping, and pellicle formation on liquid medium surfaces [10,11] and cause convoluted colony morphology in Salmonella typhimurium [15]. Various pathogenic and non-pathogenic bacteria similarly use ¢mbriae to reinforce intra-bacterial cell contact including: type 4 pili of Neisseria [16,17], type 2 ¢mbriae of dental plaque-forming Actinomyces-Streptococcus coaggregates [18] and conjugative pili between bacterial mating pairs. Like thin aggregative ¢mbriae of Salmonella, E. coli curli likely play a similar role in bio¢lm formation since curliproducing E. coli also form sti¡ colony phenotypes and cell clumping [19] or convoluted colonies [15]. In addition to the intracellular ¢ber networks, S.e. 3b appear in some cases to be bound to the substratum. It is not known if SEF17 bind directly to stain-
less steel and Te£on surfaces however in vitro aggregates of puri¢ed ¢mbriae also readily adhere to polypropylene, glass and metal (Collinson, unpublished observations). Other S.e. cell surface component(s), apart from SEF17, are required for some aspect of bio¢lm formation since SEF17-de¢cient 3b 2-2a adhered somewhat to surfaces. Reduced adherence by 3b-122 compared to the parent S.e. strain 3b on stainless steel and Te£on may have been due to the absence of SEF21 or possibly an as yet uncharacterized cell surface component also a¡ected by the ¢mU promoter Tn10 insertion [12]. Further studies with SEF18 isogenic mutants are required to establish whether these ¢mbriae a¡ect adherence, however such mutants are currently unavailable. S.e. bio¢lm formation on inert surfaces is likely complex, not only involving the interplay of various bacterial cell surface components expressed by cells in a given environment but also the nature of surface adsorbed solvent molecules and physicochemical properties of the inert surface itself [6,9]. S.e. not only adheres to physically very di¡erent hydrophobic (Te£on) and hydrophilic (stainless steel) surfaces common in the food industry, but SEF17 ¢mbriae actually promote the buildup of thick bio¢lms in the pure culture bio¢lm system studied herein. Several important implications arise from this study which are consistent with the current understanding of bio¢lm developmental biology and the character of adherent microbial populations [6,9]. Clumps of aggregated S.e. were readily sloughed o¡ the stainless steel and Te£on which suggest the potential contamination of food by Salmonella aggregates during food processing as an obvious health hazard. In addition, the recalcitrance of SEF17 to certain detergents, chaotropic agents, acids and heat [11] raises the question as to whether these ¢mbriae also protect Salmonella aggregates in bio¢lms during equipment sanitization. Conceivably, residual SEF17-producing S.e., whether alive or killed, remaining on sanitized surfaces would also facilitate S.e. recolonization. Indeed, a recent Salmonella outbreak was linked to residual bacteria on blender blades apparently una¡ected by routine sanitizing procedures [20]. Salmonella aggregates in food would also result in under estimation of bacterial load if standard cell counting procedures did not disperse the cell aggregates. Further studies will be needed
FEMSLE 8147 7-5-98
J.W. Austin et al. / FEMS Microbiology Letters 162 (1998) 295^301
to establish whether these thick Salmonella bio¢lms are resistant to sanitizers and other antimicrobial agents, as is often observed with various bio¢lms [4,6,7,9]. If so, knowledge of the Salmonella surface components contributing to bio¢lm development and resistance will lead to the development of more effective sanitizers.
Acknowledgments This work was supported by the Natural Sciences and Engineering Research Council of Canada and the Canadian Bacterial Diseases Network. We thank Dr. S.C. Clouthier for helpful discussions and assisting with the Western immunoblots and Dr. J. Farber for proofreading the manuscript.
References [1] D'Aoust, J.-Y. (1997) Salmonella species. In: Food Microbiology. Fundamentals and Frontiers (Doyle, M.P., Beuchat, L.R. and Montville, T.J., Eds.), pp. 129^158. ASM Press, Washington, DC. [2] Poppe, C. (1994) Salmonella enteritidis in Canada. Int. J. Food Microbiol. 21, 1^5. [3] Jones, K. and Bradshaw, S.B. (1996) Bio¢lm formation by the enterobacteriaceae : A comparison between Salmonella enteritidis, Escherichia coli and a nitrogen-¢xing strain of Klebsiella pneumoniae. J. Appl. Bacteriol. 80, 458^464. [4] Dhir, V.K. and Dodd, C.E. (1995) Susceptibility of suspended and surface-attached Salmonella enteritidis to biocides and elevated temperatures. Appl. Environ. Microbiol. 61, 1731^ 1738. [5] Helke, D.M. and Wong, A.C.L. (1994) Survival and growth characteristics of Listeria monocytogenes and Salmonella typhimurium on stainless steel and buna-N rubber. J. Food Protect. 57, 963. [6] Palmer, Jr., R.J. and White, D.C. (1997) Developmental biology of bio¢lms: implications for treatment and control. Trends Microbiol. 5, 435^440. [7] Carpentier, B. and Cerf, O. (1993) Bio¢lms and their consequences, with particular reference to hygiene in the food industry. J. Appl. Bacteriol. 75, 499^511.
301
[8] Austin, J.W. and Bergeron, G. (1995) Development of bacterial bio¢lms in dairy processing lines. J. Dairy Res. 62, 509^ 519. [9] Fletcher, M. (1996) Bacterial attachment in aquatic environments: A diversity of surfaces and adhesion strategies. In: Bacterial Adhesion (Fletcher, M., Ed.), pp. 1^24. Wiley-Liss, New York. [10] Collinson, S.K., Doig, P.C., Doran, J.L., Clouthier, S.C., Trust, T.J. and Kay, W.W. (1993) Thin, aggregative ¢mbriae mediate binding of Salmonella enteritidis to ¢bronectin. J. Bacteriol. 175, 12^18. [11] Collinson, S.K., Emoëdy, L., Muëller, K.H., Trust, T.J. and Kay, W.W. (1991) Puri¢cation and characterization of thin, aggregative ¢mbriae from Salmonella enteritidis. J. Bacteriol. 173, 4773^4781. [12] Clouthier, S.C., Collinson, S.K., White, A.P., Banser, P.A. and Kay, W.W. (1998) tRNAarg (¢mU) and expression of SEF14 and SEF21 in Salmonella enteritidis. J. Bacteriol. 180, 840^845. [13] Sonak, S. and Bhosle, N.B. (1995) A simple method to assess bacterial attachment to surfaces. Biofouling 9, 31^38. [14] Clouthier, S.C., Collinson, S.K. and Kay, W.W. (1994) Unique ¢mbriae-like structures encoded by sefD of the SEF14 ¢mbrial gene cluster of Salmonella enteritidis. Mol. Microbiol. 12, 893^901. [15] Sukupolvi, S., Lorenz, R.G., Gordon, J.I., Bian, Z., Pfeifer, J.D., Normark, S.J. and Rhen, M. (1997) Expression of thin aggregative ¢mbriae promotes interaction of Salmonella typhimurium SR-11 with mouse small intestinal epithelial cells. Infect. Immun. 65, 5320^5325. [16] Dekker, N.P., Lammel, C.J. and Brooks, G.F. (1991) Scanning electron microscopy of piliated Neisseria gonorrhoeae processed with hexamethyldisilazane. J. Electron Microsc. Tech. 19, 461^467. [17] Marceau, M., Beretti, J.L. and Nassif, X. (1995) High adhesiveness of encapsulated Neisseria meningitidis to epithelial cells is associated with the formation of bundles of pili. Mol. Microbiol. 17, 855^863. [18] Whittaker, C.J., Klier, C.M. and Kolenbrander, P.E. (1996) Mechanisms of adhesion by oral bacteria. Annu. Rev. Microbiol. 50, 513^552. [19] Collinson, S.K., Emoëdy, L., Trust, T.J. and Kay, W.W. (1992) Thin aggregative ¢mbriae from diarrheagenic Escherichia coli. J. Bacteriol. 174, 4490^4495. [20] Anonymous (1992) Hospital outbreak of Salmonella enteritidis infection ^ Ontario. Canada Communicable Disease Report 18-8, pp. 57^60.
FEMSLE 8147 7-5-98