- Email: [email protected]
New Fimbrial Adhesins of Serratia Marcescens Isolated from Urinary Tract Infections: Description and Properties
Vol 157, 694-698, February 1997 Printed i n U S.A
NEW FIMBRIAL ADHESINS OF SERRATIA MARCESCENS ISOLATED FROM URINARY TRACT INFECTIONS: DESCRIPTION AND PROPERTIES S. LERANOZ, P. ORUS, M. BERLANGA, F. DALET From the Department
AND
M. VINAS*
Microbiology, Institut Uniuersitari de Salut Publica de Catalunya, Health Sciences Division, Unioersity of Barcelona, and the Department of Microbiology, Fundacio Puiguert, Barcelona, Spain
of
ABSTRACT
Fimbriation, hemagglutination and adherence properties were studied in two strains of S. marcescens (ATCC 43820 and 43821) isolated from the urine of two hospitalized patients in two different hospitals. Studies were performed using electron microscopy (EM), fimbrial purification, recombinant DNA and hemagglutination techniques, hydrophobicity and tests of adherence to uroepithelial cells, catheters and glass. In EM, fimbriae of these two strains showed an inner channel and were 11 nm. thick and 0.76-1.08 pm. long. Original strains and the clone GH42-pSF192 (recombinant DNA prepared using E. coli GH42 a s recipient and the cosmid SuperCos 1 a s a vector) versus negative control (E.coli GH42-SuperCos 1)showed mannose-resistant hemagglutination of tanned erythrocytes and yeast, high hydrophobicity (55.4 and 49.6% at 37C versus 22.8%) and high adherence to borosilicate glass (313,000 and 168,000 CFU/cm.2 versus 17,000 CFU/cm.2),catheters (4.7 x lo6 and 1.0 X lo6 CFU/cm.z versus 3.9 X lo4 CFU/cm.') and uroepithelial cells (adherence indexes of 3.82 and 3.29 versus 1.25). The properties of the fimbriae studied were different from those previously described in the genus Serratia, and they were designated as MFV". KEYWORDS: Serratia marcescens, fimbriae, adhesive properties, UTI
Serratia marcescens was once largely considered as a nonpathogenic microorganism. However, in the last few years the role of the bacterium as an opportunistic pathogen has gradually been recognized. It produces a large variety of human infections, mainly nosocomial. Among them, complicated urinary tract infections (UTI) are prominent in patients with or without indwelling catheters.' Its virulence is related to the production of fimbriae, cell wall antigens, serum bactericidal resistance, and siderophore and proteases synthesis. Fimbriation in the genus Serratia has been studied" and five types (Types 1 and 3, FGH MFUF', MR/P thick and MFUP thin) have been described. To date, the species S. marcescens was able to produce only three different types of these fimbriae (Types 1 and 3 and MFW), all of them with adherence to catheters and eucaryotic cells.' Some S. rnarcescens strains are able to produce a diffusible red pigment, called prodieosin. This pigment is common in strains isolated from soil and water, but very rare in clinical samples. The strains studied herein were selected due to their ability t o produce a yellow diffusible pigment (2hydroxy-5-carboxymetylmuconic acid derivative) which is obtained during the tyrosine catabolic pathway.3 The aim of this work was to determine and characterize the fimbriae produced by two yellow pigment-producing clinical strains of S. marcescens. MATERIALS A N D METHODS
Bacterial strains. S. rnarcescens strain ATCC 43821 was isolated in our department from a patient's urine affected by a complicated UTL3 Some time later, another strain of S. marcescens with similar biochemical characteristics (ATCC 43820) was isolated from the urine of another patient with Accepted for publication May 31, 1996. * Requests for reprints: De artment of Microbiology, Institut Universitari de Salut Publica, l a m us de Bellvitge, Pavello Central, University of Barcelona 08907, Ifospitalet de Llobregat. Barcelona, Spain. Supported b grant PM90-0073 from the Direccion General de Investigacinn Zientifica y Technica.
complicated UTI at a different hospital in Barcelona. Both strains were studied because of their special characteristics in aromatic catabolism, and subsequently included in the ATCC collection.4.5 Media and growth conditions. Bacteria were maintained on Tryptone Soy Agar (TSA). For experiments, bacteria were incubated in Tryptone Soy Broth (TSB). When necessaly, media were supplemented with ampicillin (Sigma) a t 50 mg./l. Synthetic urine medium was prepared according to Minuth e t a1.6 and TSB 2% was added to improve growth. Electron microscopy. In order to visualize the fimbriae in the electron microscope, bacteria were obtained from six serial static cultures in TSB, a t 37C, a t 2-day intervals. Cells were harvested by gentle centrifugation, washed three times in Phosphate Buffer Saline (PBS) and finally suspended in PBS. Bacterial suspension was placed onto fonnvar-carbon coated gold grids and staining was achieved by 2% uranyl acetate. Preparations were observed in a Hitachi MT 800 electron microscope at 75 kV. Fimbrial purification. Purification of fimbrial material was carried out following the method of Dodd and E i ~ e n s t e i n . ~ Control of purity of preparations was performed by electrophoresis in polyacrylamide gels,8 as modified by Ames et aL9 Recombinant DNA techniques. S. rnarcescens chromosomal DNA was obtained following the method of Leranoz et al.1° to prevent nuclease action. Chromosomal DNA was partially digested with Sau3Al restriction enzyme (Boehringer) and subsequently analyzed by electrophoresis in agarose. Restriction fragments were ligated with the cloning vectors using T4-DNA ligase (Boehringer). Transformation was performed by electroporation using an Invitrogen apparatus a t 150 F and 1,700 volts. Lambda transducing particles were prepared using the kit Gigapack I1 packaging extract (Stratagene). To clone the genes involved in fimbriae biosynthesis, recombinant DNA techniques were used to prepare a genonk library from S. marcescens ATCC 43821 chromosome. The cloning experiments were carried out using the nonfimbri-
694
NEW FIMBRIAE IN GENUS SERRATIA IMPLICATED IN UTI
ated E. coli GH42 strain as the recipient11.12 and the cosmid SuperCos 1 as a vector. Experimental procedures in the study of fimbriae. The ability of fimbriated bacteria to agglutinate was assayed as follows: bacteria cultured in TSB were harvested by centrifugation at 2,000 X g for 10 min. and washed twice in PBS. Finally, cells were suspended in PBS at an OD,,, of 0.4. Erythrocytes from different species (horse, chicken, rabbit, guinea pig and human 0 Rh+) were separated by centrifugation of blood at 1,500 x g, washed three times in PBS and suspended in PBS at a final concentration of 3%(v/v).Equal volumes of bacterial and fresh erythrocyte suspensions or erythrocytes treated with tannic acid (0.003%)were mixed on a slide at room temperature. When necessary, D-mannose was added to the bacterial suspension. Yeast agglutination activity was controlled by a n identical method using yeast suspension instead erythrocytes. Hydrophobicity. Hydrophobicity was assayed by the BATH method following Iimura et al.13 and Rosemberg et al.14 Briefly, 10 ml. of overnight culture were centrifuged and suspended in PBS a t an OD,,, of 0.6 and pH 6.5. Aliquots of 4 ml. were mixed with 1 ml. of toluene, vortexed for 30 sec. and left at room temperature for 1 hr. This allowed the separation of organic and inorganic phases. The presence of bacteria in the aqueous phase was measured spectrophotometrically. Results were calculated by the formula (H = (A, &)/A,), where A, and Af are initial and final absorbance respectively. Adherence to glass. Bacterial suspensions in PBS were allowed to adhere to borosilicate surfaces for 1 hr. Surfaces were then washed in PBS, fixed with glutaraldehyde at 0.5% in 0.1 M cacodylate buffer pH 7 for 1 hr., stained with acridine orange (0.01%)for 2 min., dried and examined under a Zeiss epifluorescence microscope. Ten fields were counted for every sample. Results were expressed as CFU/cm.2 Adherence to urinary catheters. To study the adherence to urinary catheters we used Foley catheters from Baxter. A system was constructed to allow the adhesion of the baderia to the inner surface of the catheters (fig. 1). The reservoir contained one litter of synthetic urine medium inoculated with the microorganism. Flux was regulated at 60 m1.h.; after 24 hours the catheter was washed in 30 ml. of PBS 3 times and then cut into 1 cm. length and opened. Adhered bacteria were recovered by an ultrasonic treatment of 30 sec. in a ultrasonic bath, followed by vortexing for 1 min. Cells were counted by enumeration of the viable count on TSA plates. Adherence to uroepithelial cells. This was tested as de-
FIG.1. Dispositive schema of adherence to Foley catheters
695
scribed previously.15.16 Experiments were performed in triplicate. Count values were divided by the count of their respective control. Several carbohydrates (D-arabinose, D-fructose, D-galactose, D-glucose, D-mannose and D-xylose) were added to determine the ability to inhibit adhesion; the concentrations tested were 1, 2 and 3% (w/v). RESULTS
The first step in this study was the observation of S. marcescens ATCC 43821 under the electron microscope. When bacteria were incubated in liquid medium and statically, a great proportion of the bacteria appeared to be fimbriated (fig. 2). The fimbriae were about 11 nm. thick and between 0.76 and 1.08 pm. long, and an inner channel was observed. S. marcescens was unable to agglutinate fresh erythrocytes of the different species tested, although it agglutinated tanned elythrocytes and yeasts. This ability of agglutination was not inhibited by D-mannose. Agglutinating ability was removed by heat treatment (65C for 30 min.) and also by treatment with formaldehyde (0.5% w/v) at 37C for 4-hr. When the PF-test (Orion)kit17 was used no reaction was observed, which suggests that our fimbriae were unrelated to those known as P-type fimbriae. We designated the fimbriae of S. marcescens ATCC 43820 and 43821 as MFUT (Mannose-resistant-thick). Purification of fimbriae produced highly purified protein preparations. Figure 3 shows a PAGE in which two main polypeptides can be observed. In order to clone genes involved in MFU" fimbriae biosynthesis, recombinant DNA techniques were used to prepare a genomic library from S. marcescens ATCC 43821 chromosome. The first attempts
FIG. 2. Electron micrograph of S. marcescens ATCC 43821 (Msgnification X33,OOO).
696
NEW FIMBRIAE IN GENUS SERRATZA IMPLICATED IN UTI
FIG. 3. SDS-PAGE of purified fimbrial preparations. Lanes: a, molecular size standards; b, S. marcescens ATCC 43821; c, S. marcescens ATCC 43820.
were made using E. coli HBIOlls as recipient strain and the ~ screening of clones was plasmid pBR328 as a ~ e c t o r . 1The carried out by selecting recombinants by their ability to agglutinate erythrocytes. The strain HBlOl gave puzzling results, so a new cloning strategy was assayed. Strain GH42 was used as recipient (a gift from Dr. Hull, Baylor College of Medicine, Houston, Texas). It is a well-known unfimbriated minicell-producer strain. Cloning was finally achieved by using SuperCos 1, which yielded several clones expressing agglutination. One of these clones (GH42-pSF192) was selected for further experiments. The clone exhibited a high ability of hemagglutination. Furthermore, ability to produce fimbriae by minicells obtained from the clone was determined by electron microscopy (fig. 4). In t h e experiments on hydrophobicity, adhesion to inert surfaces, and adhesion to human cells, three strains were used simultaneously. S. marcescens ATCC 43821 was used a s positive control; E. coli GH42 carrying the cosmid SuperCos 1 as negative control; and GH42-pSF192 as problem. The clone GH42-pSF192 exhibited t h e same standard as original strains and high ability to agglutinate mannose-resistant tanned erythrocytes. Hydrophobicity was markedly modified by temperature, since high hydrophobicity was detected when cultures of strains S. marcescens ATCC 43821 and E . coli GH42-pSF192 were incubated at 30C (experimental temperature). Hydrophobicity decreased when temperature increased, but in both strains it was still about 50%;a t 37C ( h u m a n physiological temperature) (table 1). A statistically significant difference ( p <0.001) was observed in glass adherence between ATCC 43821 (313,000 CFU/crn.') and E. coli GH42-pSF192 (168,000 CFU/cm.') with respect to E. coli GH42-SuperCos 1 (17,000 CFU/cm.') a t 24 hr. incubation. The exposure time increased the adherence averages, while the negative control did not exhibit significant variation (fig. 5 ) . The efficiency in adherence to the surface of Foley catheters showed t h a t strains ATCC 43821 (4.7 X 10' CFU/cm.') and clone GH42-pSF192 (0.99 x 10" CFU/cm.") were significantly different ( p (0.01) from negative control (3.9 x lo4 CFU/cm.',. The results of adherence to human uroepithelial cells (table 2) showed: ( i ) high ability to adhere with ATCC 43821 and the clone GH42-pSF192 strains, ( i i ) significant differences with respect to negative control, and (iiir none of carbohydrates tested inhibited or modified the adherence.
FIG. 4. Electron micrograph of E. coli GH42-pSF192 (Magnification X40.000). TABLE1. Hydrophobicity values and serniquarititatii~eeiialuation
of
agglutination ability of the three strains tested Temperature Incubation
30C
34C
37C
39C
S.marcescens ATCC 43821 QH 65.0 t 2.8 61.4 ? 3.5 55.4 z 2 . 9 47.5 I2.2 I Agglutination ++IE. coli GH42 pSF192 56.3 t 1.4 55.2 i 2 4 49.6 z '2.8 40.1 2 3.1 %H Agglutination ++ + tlE. coli GH42 SuperCos 1 QH 19.0 I 5.1 13.6 z 3.4 22.8 : 3.6 15.7 f 2.1 Agglutination % H: Percentage of hydrophobicity. i
-
~
-
DISCUSSION
All types of fimbriae described in genus Serratia have been defined from their dimensions, by t h e presence or absence of a central channel, and finally by their agglutination properties. MFUT cannot belong to any of the previously described types: (i) type 3 are thin (3-4 nm. thick) and without a n inner channel, whereas in MWT a channel 5 nm. thick was clearly observed, (ii) FGWMRP and some F/MRP a r e also thin, (5) the dimensions of MWT and those of F M W of S. fonticoh a n d type 1 fimbriae are coincident, but F MR/P only agglutinate chicken erythrocytes and type 1 is inhibited by mannose. MWT can be differentiated from these fimbriae by their agglutination properties. Additionally, thel-nio- and formaldehyde-sensitivity a r e clearly different in t.vpe 1 and MW.
NEW FIMBRIAE IN GENUS SERRATIA IMPLICATED IN UTI
FIG.5. Adherence averages to borosihcate glass of three strains tested. TABLE2. Adherence capacity to uroepithelial cells of the three strains tested a n d inhibition by addition of several concentrations of different carbohydrates Adherence Indexes Carbohydrate
c.
S. marcescens
ATCC43821
No D-Arabinose
0 3.82 ? 0.13 1 3.35 t 0.10 2 3.28 t 0.005 3 3.41 t 0.09 D-Fructose 1 3.86 f 0.02 2 3.88 f 0.003 3 3.31 f 0.10 D-Galactose 1 3.96 f 0.11 2 3.91 0.08 3 3.78 ? 0.15 D-Glucose 1 3.67 ? 0.05 2 3.85 ? 0.22 3 3.50 ? 0.04 1 3.75 ? 0.09 D-Mannose 2 3.72 ? 0.04 3 3.58 It_ 0.15 D-Xylose 1 3.47 ? 0.004 2 2.78 ? 0.005 3 3.19 ? 0.17 * Carbohydrate concentration in 70. f_
E. coli GH42-pSF192
E. coli GH42-SuperCos 1
3.29 f 0.11 3.02 f 0.0008 2.91 2 0.01 2.94 ? 0.05 2.91 ? 0.13 2.89 f 0.16 2.78 t 0.11 3.11 f 0.09 2.94 f 0.09 3.30 f 0.13 2.97 0.08 3.16 ? 0.18 3.13 ? 0.02 3.18 ? 0.09 2.85 t 0.01 3.30 It_ 0.017 3.13 f 0.17 3.33 f 0.06 3.31 f 0.01
1.25 ? 0.05 1.26 ? 0.01 1.24 f 0.01 1.24 Z 0.007 1.23 f 0.03 1.31 f 0.03 1.25 f 0.01 1.31 f 0.01 1.24 f 0.03 1.30 f 0.01 1.21 f 0.007 1.23 f 0.04 1.20 t 0.03 1.28 2 0.006 1.21 f 0.05 1.23 f 0.04 1.21 2 0.03 1.15 f 0.04 1.14 ? 0.009
f_
On the other hand, none of the sugars tested acted as a specific receptor of MWT fimbriae. It is generally accepted that agglutination of yeast is a good test for type 1 fimbriae and that mannose of the surface is the receptor for these fimbriae. However, this agglutination is inhibited by the addition of mannose. In our case, although MlW agglutinate yeast, no inhibition by mannose was detected. Then, some other receptor should be responsible for yeast agglutination. TO test whether fucose or sialic acid could act as receptors, since they form part of the yeast wall (J.P. Martinez, University of Valencia, personal communication), several lectins were tested. Since negative results were obtained, the nature of MR/T receptor remains unknown, These findings suggest that MFUT constitutes a new type of fimbria produced by uropathogenic S. marcescens, different from all previously described fimbriae in Serratia. Fimbrial purification and subsequent study using PAGE demonstrate that the material obtained was a constituent of the fimbriae. The first attempts using non-fimbriated strain E. coli HI3101 as recipient and the plasmid pBR328 as vector gave puzzling results, until a new cloning strategy was used. The isolation of the clone GH42-pSF192, able to produce fimbriae, whose properties were identical to those produced by S. marcescens ATCC 43821, demonstrated that all characteristics described (agglutination, hydrophobicity, adherence activity) could be attributed to MILT fimbriae. The modifications of hydrophobicity by temperature suggested that some effect on Serratia virulence could be modi-
697
fied in this way, since high hydrophobicity could be detected when bacteria grew at low temperatures (outside the human body), which should favour adhesion, whereas this hydrophobicity decreased (about 20%) when bacteria grew at 37C (in the human body). This property could be related to a reduction in the number of fimbriae expressed on the bacterial surface. From a practical point of view, the hydrophobicity at 37C (human infections) shown by ATCC 43821 could be used as a protection against this strain itself. Catheter-associated bacteriuria, the main cause of nosocomial infections and septic pictures, could easily be prevented using hydrophilic-coated catheters. The strong electronegative fields created by the catheter hydrophilic-layer and the bacterial surface MWT fimbriated will probably avoid the adhesion due to the repulsion forces. According to the results, MR/T fimbriae expressed by S. marcescens can be classified as able t o adhere to human uroepithelial cells. In theory, MILT-fimbriated strains are perfectly adapted for urinary tract invasion, especially when the host-defense mechanisms are diminished due to underlying urologic or systemic diseases. Acknowledgments. The technical support of 'Servei de Microscopia electronica" of the University of Barcelona is gratefully acknowledged, and we thank Elisenda Viiias for her collaboration and Robin Rycroft for his help in the preparation of the manuscript. REFERENCES
1. Maki, D. G., Hennekens, C. G., Phillips, C. W., Shaw, W. V. and Bennet, J. V.: Nosocomial urinary tract infection with Serratia marcescens: a n epidemiological study. J. Infect. Dis., 128: 579, 1973. 2. Grimont, F. and Grimont, P. A. D.: The genus Serratia. In: The Prokaryotes. Edited by A. Ballows, H. G. Triipper, M. Dworkin, W. H. Tno and K. H. Schleifer. New York: SpringerVerlag, vol. 2, chapt. 150, pp. 2822-2448, 1992. 3. Trias, J., Vifias, M., Guinea, J. and Loren, J. G.: Isolation from urine of two Serratia marcescens strains excreting a diffusible yellow pigment. J. Gen. Microbiol., 133 773, 1987. 4. Trias, J., Viiias, M., Guinea, J. and Loren, J. G.: Induction of yellow pigmentation in Serratia marcescens. Appl. Environ. Microbiol., 54: 3138, 1988. 5. Trias, J., Vifias, M., Guinea, J. and Loren, J. G.: Brown pigmentation in Serratia marcescens cultures associated with tyrosine metabolism. Can. J. Microbiol., 35: 1037, 1989. 6. Minuth, J. N., Musher, D. M. and Thorsteinsson, S. B.: Inhibition of antibacterial activity of gentamicin by urine. J. Infect. Dis., 133: 14, 1976. 7. Dodd, D. C. and Eisenstein, B. I.: Antigenic quantitation of type 1 fimbriae on the surface of Escherichia colt cells by an enzyme-linked immunosorbent inhibition assay. Infect. Immun., 38 764, 1982. 8. Laemmli, U. K.: Cleavage of structural proteins during the assembly ofthe head ofbacteriophage T4. Nature (London),227: 680,1970. 9. Ames, G. F. L., Suphch, E. N. and Nikaido, H.: Protein composition of the outer membrane of Salmonella typhimurium: effect of lipopolysaccharide mutations. J. Bacteriol., 117: 406, 1974. 10. Leranoz, A. M., Fuste, M. C., Hull, R. A. and Williams, R. P.: Cloning and expression in Escherichia coli of a gene encoding proline oxidase of Serratia marcescens. Microbios, 67: 87. 1991. 11. Karr, J. F., Nowick, B., Truong, L. D., Hull, R. and Hull, S.: Purified P fimbriae from two cloned gene clusters of a single pyelonephritogenic strain adhere to unique structures in the human ludney. Infect. Immun., 57: 3594, 1989. 12. Haraguchi, G. E., Hull, R. A., Krallmamm-Wenzel, U. and Hull, S. I.: Molecular cloning and expression of the 0 4 polysaccharide gene cluster from Escherichia coli. Microb. Pathogen., 6 123, 1989. 13. Iimura, Y., Har, S. and Otsuka, K.: Cell surface hydrophobicity as a pellicle formation factor in film strain Saccharomyces. Agric. Biol. Chem., 44: 1215, 1980.
698
NEW FIMBRIAE IN GENUS SERRATIA IMPLICATED IN UTI
14. Rosenberg, M., Gutnick, D. and Rosenberg, E.: Adherence of
bacteria to hydrocarbons: a simple method for measuring cellsurface hydrophobicity. FEMS Microbiol. Lett., 9 29, 1980. 15. Schaeffer, A. J., Amundsen, S. K. and Schmidt, L. N.: Adherence of Escherichia coli to human urinary tract epithelial cells. Infect. Immun., 24: 753, 1979. 16. Palomar, J., Leranoz, A. M. and Viaas, M.: Serratia marcescens adherence: the effect of 0-antigen presence. Microbios, 81: 107, 1995.
17. Karkktiinen, U., Ikaheimo, R., Katila, M. and Mantyjarvi, R.: p fimbriation of Escherichia coli strains from patients with UTOsepsis demonstrated by a commercial agglutination test (PF TEST). J. Clin. Microbiol., 2 9 221, 1991. 18. Soberon, X., Covarmbias, L. and Bolivar, F.: Construction and characterization of new cloning vehicles, IV. Delection denvates of pBR322 and pBR325. Gene, 9 287, 1980. 19. Sambrook, J.,Fritsch, E. F. and Maniatis, T.: Molecular Cloning. New York: Cold Spring Harbor Laboratory, vol. 3, p. 1, 1989.
NEW FIMBRIAL ADHESINS OF SERRATIA MARCESCENS ISOLATED FROM URINARY TRACT INFECTIONS: DESCRIPTION AND PROPERTIES S. LERANOZ, P. ORUS, M. BERLANGA, F. DALET From the Department
AND
M. VINAS*
Microbiology, Institut Uniuersitari de Salut Publica de Catalunya, Health Sciences Division, Unioersity of Barcelona, and the Department of Microbiology, Fundacio Puiguert, Barcelona, Spain
of
ABSTRACT
Fimbriation, hemagglutination and adherence properties were studied in two strains of S. marcescens (ATCC 43820 and 43821) isolated from the urine of two hospitalized patients in two different hospitals. Studies were performed using electron microscopy (EM), fimbrial purification, recombinant DNA and hemagglutination techniques, hydrophobicity and tests of adherence to uroepithelial cells, catheters and glass. In EM, fimbriae of these two strains showed an inner channel and were 11 nm. thick and 0.76-1.08 pm. long. Original strains and the clone GH42-pSF192 (recombinant DNA prepared using E. coli GH42 a s recipient and the cosmid SuperCos 1 a s a vector) versus negative control (E.coli GH42-SuperCos 1)showed mannose-resistant hemagglutination of tanned erythrocytes and yeast, high hydrophobicity (55.4 and 49.6% at 37C versus 22.8%) and high adherence to borosilicate glass (313,000 and 168,000 CFU/cm.2 versus 17,000 CFU/cm.2),catheters (4.7 x lo6 and 1.0 X lo6 CFU/cm.z versus 3.9 X lo4 CFU/cm.') and uroepithelial cells (adherence indexes of 3.82 and 3.29 versus 1.25). The properties of the fimbriae studied were different from those previously described in the genus Serratia, and they were designated as MFV". KEYWORDS: Serratia marcescens, fimbriae, adhesive properties, UTI
Serratia marcescens was once largely considered as a nonpathogenic microorganism. However, in the last few years the role of the bacterium as an opportunistic pathogen has gradually been recognized. It produces a large variety of human infections, mainly nosocomial. Among them, complicated urinary tract infections (UTI) are prominent in patients with or without indwelling catheters.' Its virulence is related to the production of fimbriae, cell wall antigens, serum bactericidal resistance, and siderophore and proteases synthesis. Fimbriation in the genus Serratia has been studied" and five types (Types 1 and 3, FGH MFUF', MR/P thick and MFUP thin) have been described. To date, the species S. marcescens was able to produce only three different types of these fimbriae (Types 1 and 3 and MFW), all of them with adherence to catheters and eucaryotic cells.' Some S. rnarcescens strains are able to produce a diffusible red pigment, called prodieosin. This pigment is common in strains isolated from soil and water, but very rare in clinical samples. The strains studied herein were selected due to their ability t o produce a yellow diffusible pigment (2hydroxy-5-carboxymetylmuconic acid derivative) which is obtained during the tyrosine catabolic pathway.3 The aim of this work was to determine and characterize the fimbriae produced by two yellow pigment-producing clinical strains of S. marcescens. MATERIALS A N D METHODS
Bacterial strains. S. rnarcescens strain ATCC 43821 was isolated in our department from a patient's urine affected by a complicated UTL3 Some time later, another strain of S. marcescens with similar biochemical characteristics (ATCC 43820) was isolated from the urine of another patient with Accepted for publication May 31, 1996. * Requests for reprints: De artment of Microbiology, Institut Universitari de Salut Publica, l a m us de Bellvitge, Pavello Central, University of Barcelona 08907, Ifospitalet de Llobregat. Barcelona, Spain. Supported b grant PM90-0073 from the Direccion General de Investigacinn Zientifica y Technica.
complicated UTI at a different hospital in Barcelona. Both strains were studied because of their special characteristics in aromatic catabolism, and subsequently included in the ATCC collection.4.5 Media and growth conditions. Bacteria were maintained on Tryptone Soy Agar (TSA). For experiments, bacteria were incubated in Tryptone Soy Broth (TSB). When necessaly, media were supplemented with ampicillin (Sigma) a t 50 mg./l. Synthetic urine medium was prepared according to Minuth e t a1.6 and TSB 2% was added to improve growth. Electron microscopy. In order to visualize the fimbriae in the electron microscope, bacteria were obtained from six serial static cultures in TSB, a t 37C, a t 2-day intervals. Cells were harvested by gentle centrifugation, washed three times in Phosphate Buffer Saline (PBS) and finally suspended in PBS. Bacterial suspension was placed onto fonnvar-carbon coated gold grids and staining was achieved by 2% uranyl acetate. Preparations were observed in a Hitachi MT 800 electron microscope at 75 kV. Fimbrial purification. Purification of fimbrial material was carried out following the method of Dodd and E i ~ e n s t e i n . ~ Control of purity of preparations was performed by electrophoresis in polyacrylamide gels,8 as modified by Ames et aL9 Recombinant DNA techniques. S. rnarcescens chromosomal DNA was obtained following the method of Leranoz et al.1° to prevent nuclease action. Chromosomal DNA was partially digested with Sau3Al restriction enzyme (Boehringer) and subsequently analyzed by electrophoresis in agarose. Restriction fragments were ligated with the cloning vectors using T4-DNA ligase (Boehringer). Transformation was performed by electroporation using an Invitrogen apparatus a t 150 F and 1,700 volts. Lambda transducing particles were prepared using the kit Gigapack I1 packaging extract (Stratagene). To clone the genes involved in fimbriae biosynthesis, recombinant DNA techniques were used to prepare a genonk library from S. marcescens ATCC 43821 chromosome. The cloning experiments were carried out using the nonfimbri-
694
NEW FIMBRIAE IN GENUS SERRATIA IMPLICATED IN UTI
ated E. coli GH42 strain as the recipient11.12 and the cosmid SuperCos 1 as a vector. Experimental procedures in the study of fimbriae. The ability of fimbriated bacteria to agglutinate was assayed as follows: bacteria cultured in TSB were harvested by centrifugation at 2,000 X g for 10 min. and washed twice in PBS. Finally, cells were suspended in PBS at an OD,,, of 0.4. Erythrocytes from different species (horse, chicken, rabbit, guinea pig and human 0 Rh+) were separated by centrifugation of blood at 1,500 x g, washed three times in PBS and suspended in PBS at a final concentration of 3%(v/v).Equal volumes of bacterial and fresh erythrocyte suspensions or erythrocytes treated with tannic acid (0.003%)were mixed on a slide at room temperature. When necessary, D-mannose was added to the bacterial suspension. Yeast agglutination activity was controlled by a n identical method using yeast suspension instead erythrocytes. Hydrophobicity. Hydrophobicity was assayed by the BATH method following Iimura et al.13 and Rosemberg et al.14 Briefly, 10 ml. of overnight culture were centrifuged and suspended in PBS a t an OD,,, of 0.6 and pH 6.5. Aliquots of 4 ml. were mixed with 1 ml. of toluene, vortexed for 30 sec. and left at room temperature for 1 hr. This allowed the separation of organic and inorganic phases. The presence of bacteria in the aqueous phase was measured spectrophotometrically. Results were calculated by the formula (H = (A, &)/A,), where A, and Af are initial and final absorbance respectively. Adherence to glass. Bacterial suspensions in PBS were allowed to adhere to borosilicate surfaces for 1 hr. Surfaces were then washed in PBS, fixed with glutaraldehyde at 0.5% in 0.1 M cacodylate buffer pH 7 for 1 hr., stained with acridine orange (0.01%)for 2 min., dried and examined under a Zeiss epifluorescence microscope. Ten fields were counted for every sample. Results were expressed as CFU/cm.2 Adherence to urinary catheters. To study the adherence to urinary catheters we used Foley catheters from Baxter. A system was constructed to allow the adhesion of the baderia to the inner surface of the catheters (fig. 1). The reservoir contained one litter of synthetic urine medium inoculated with the microorganism. Flux was regulated at 60 m1.h.; after 24 hours the catheter was washed in 30 ml. of PBS 3 times and then cut into 1 cm. length and opened. Adhered bacteria were recovered by an ultrasonic treatment of 30 sec. in a ultrasonic bath, followed by vortexing for 1 min. Cells were counted by enumeration of the viable count on TSA plates. Adherence to uroepithelial cells. This was tested as de-
FIG.1. Dispositive schema of adherence to Foley catheters
695
scribed previously.15.16 Experiments were performed in triplicate. Count values were divided by the count of their respective control. Several carbohydrates (D-arabinose, D-fructose, D-galactose, D-glucose, D-mannose and D-xylose) were added to determine the ability to inhibit adhesion; the concentrations tested were 1, 2 and 3% (w/v). RESULTS
The first step in this study was the observation of S. marcescens ATCC 43821 under the electron microscope. When bacteria were incubated in liquid medium and statically, a great proportion of the bacteria appeared to be fimbriated (fig. 2). The fimbriae were about 11 nm. thick and between 0.76 and 1.08 pm. long, and an inner channel was observed. S. marcescens was unable to agglutinate fresh erythrocytes of the different species tested, although it agglutinated tanned elythrocytes and yeasts. This ability of agglutination was not inhibited by D-mannose. Agglutinating ability was removed by heat treatment (65C for 30 min.) and also by treatment with formaldehyde (0.5% w/v) at 37C for 4-hr. When the PF-test (Orion)kit17 was used no reaction was observed, which suggests that our fimbriae were unrelated to those known as P-type fimbriae. We designated the fimbriae of S. marcescens ATCC 43820 and 43821 as MFUT (Mannose-resistant-thick). Purification of fimbriae produced highly purified protein preparations. Figure 3 shows a PAGE in which two main polypeptides can be observed. In order to clone genes involved in MFU" fimbriae biosynthesis, recombinant DNA techniques were used to prepare a genomic library from S. marcescens ATCC 43821 chromosome. The first attempts
FIG. 2. Electron micrograph of S. marcescens ATCC 43821 (Msgnification X33,OOO).
696
NEW FIMBRIAE IN GENUS SERRATZA IMPLICATED IN UTI
FIG. 3. SDS-PAGE of purified fimbrial preparations. Lanes: a, molecular size standards; b, S. marcescens ATCC 43821; c, S. marcescens ATCC 43820.
were made using E. coli HBIOlls as recipient strain and the ~ screening of clones was plasmid pBR328 as a ~ e c t o r . 1The carried out by selecting recombinants by their ability to agglutinate erythrocytes. The strain HBlOl gave puzzling results, so a new cloning strategy was assayed. Strain GH42 was used as recipient (a gift from Dr. Hull, Baylor College of Medicine, Houston, Texas). It is a well-known unfimbriated minicell-producer strain. Cloning was finally achieved by using SuperCos 1, which yielded several clones expressing agglutination. One of these clones (GH42-pSF192) was selected for further experiments. The clone exhibited a high ability of hemagglutination. Furthermore, ability to produce fimbriae by minicells obtained from the clone was determined by electron microscopy (fig. 4). In t h e experiments on hydrophobicity, adhesion to inert surfaces, and adhesion to human cells, three strains were used simultaneously. S. marcescens ATCC 43821 was used a s positive control; E. coli GH42 carrying the cosmid SuperCos 1 as negative control; and GH42-pSF192 as problem. The clone GH42-pSF192 exhibited t h e same standard as original strains and high ability to agglutinate mannose-resistant tanned erythrocytes. Hydrophobicity was markedly modified by temperature, since high hydrophobicity was detected when cultures of strains S. marcescens ATCC 43821 and E . coli GH42-pSF192 were incubated at 30C (experimental temperature). Hydrophobicity decreased when temperature increased, but in both strains it was still about 50%;a t 37C ( h u m a n physiological temperature) (table 1). A statistically significant difference ( p <0.001) was observed in glass adherence between ATCC 43821 (313,000 CFU/crn.') and E. coli GH42-pSF192 (168,000 CFU/cm.') with respect to E. coli GH42-SuperCos 1 (17,000 CFU/cm.') a t 24 hr. incubation. The exposure time increased the adherence averages, while the negative control did not exhibit significant variation (fig. 5 ) . The efficiency in adherence to the surface of Foley catheters showed t h a t strains ATCC 43821 (4.7 X 10' CFU/cm.') and clone GH42-pSF192 (0.99 x 10" CFU/cm.") were significantly different ( p (0.01) from negative control (3.9 x lo4 CFU/cm.',. The results of adherence to human uroepithelial cells (table 2) showed: ( i ) high ability to adhere with ATCC 43821 and the clone GH42-pSF192 strains, ( i i ) significant differences with respect to negative control, and (iiir none of carbohydrates tested inhibited or modified the adherence.
FIG. 4. Electron micrograph of E. coli GH42-pSF192 (Magnification X40.000). TABLE1. Hydrophobicity values and serniquarititatii~eeiialuation
of
agglutination ability of the three strains tested Temperature Incubation
30C
34C
37C
39C
S.marcescens ATCC 43821 QH 65.0 t 2.8 61.4 ? 3.5 55.4 z 2 . 9 47.5 I2.2 I Agglutination ++IE. coli GH42 pSF192 56.3 t 1.4 55.2 i 2 4 49.6 z '2.8 40.1 2 3.1 %H Agglutination ++ + tlE. coli GH42 SuperCos 1 QH 19.0 I 5.1 13.6 z 3.4 22.8 : 3.6 15.7 f 2.1 Agglutination % H: Percentage of hydrophobicity. i
-
~
-
DISCUSSION
All types of fimbriae described in genus Serratia have been defined from their dimensions, by t h e presence or absence of a central channel, and finally by their agglutination properties. MFUT cannot belong to any of the previously described types: (i) type 3 are thin (3-4 nm. thick) and without a n inner channel, whereas in MWT a channel 5 nm. thick was clearly observed, (ii) FGWMRP and some F/MRP a r e also thin, (5) the dimensions of MWT and those of F M W of S. fonticoh a n d type 1 fimbriae are coincident, but F MR/P only agglutinate chicken erythrocytes and type 1 is inhibited by mannose. MWT can be differentiated from these fimbriae by their agglutination properties. Additionally, thel-nio- and formaldehyde-sensitivity a r e clearly different in t.vpe 1 and MW.
NEW FIMBRIAE IN GENUS SERRATIA IMPLICATED IN UTI
FIG.5. Adherence averages to borosihcate glass of three strains tested. TABLE2. Adherence capacity to uroepithelial cells of the three strains tested a n d inhibition by addition of several concentrations of different carbohydrates Adherence Indexes Carbohydrate
c.
S. marcescens
ATCC43821
No D-Arabinose
0 3.82 ? 0.13 1 3.35 t 0.10 2 3.28 t 0.005 3 3.41 t 0.09 D-Fructose 1 3.86 f 0.02 2 3.88 f 0.003 3 3.31 f 0.10 D-Galactose 1 3.96 f 0.11 2 3.91 0.08 3 3.78 ? 0.15 D-Glucose 1 3.67 ? 0.05 2 3.85 ? 0.22 3 3.50 ? 0.04 1 3.75 ? 0.09 D-Mannose 2 3.72 ? 0.04 3 3.58 It_ 0.15 D-Xylose 1 3.47 ? 0.004 2 2.78 ? 0.005 3 3.19 ? 0.17 * Carbohydrate concentration in 70. f_
E. coli GH42-pSF192
E. coli GH42-SuperCos 1
3.29 f 0.11 3.02 f 0.0008 2.91 2 0.01 2.94 ? 0.05 2.91 ? 0.13 2.89 f 0.16 2.78 t 0.11 3.11 f 0.09 2.94 f 0.09 3.30 f 0.13 2.97 0.08 3.16 ? 0.18 3.13 ? 0.02 3.18 ? 0.09 2.85 t 0.01 3.30 It_ 0.017 3.13 f 0.17 3.33 f 0.06 3.31 f 0.01
1.25 ? 0.05 1.26 ? 0.01 1.24 f 0.01 1.24 Z 0.007 1.23 f 0.03 1.31 f 0.03 1.25 f 0.01 1.31 f 0.01 1.24 f 0.03 1.30 f 0.01 1.21 f 0.007 1.23 f 0.04 1.20 t 0.03 1.28 2 0.006 1.21 f 0.05 1.23 f 0.04 1.21 2 0.03 1.15 f 0.04 1.14 ? 0.009
f_
On the other hand, none of the sugars tested acted as a specific receptor of MWT fimbriae. It is generally accepted that agglutination of yeast is a good test for type 1 fimbriae and that mannose of the surface is the receptor for these fimbriae. However, this agglutination is inhibited by the addition of mannose. In our case, although MlW agglutinate yeast, no inhibition by mannose was detected. Then, some other receptor should be responsible for yeast agglutination. TO test whether fucose or sialic acid could act as receptors, since they form part of the yeast wall (J.P. Martinez, University of Valencia, personal communication), several lectins were tested. Since negative results were obtained, the nature of MR/T receptor remains unknown, These findings suggest that MFUT constitutes a new type of fimbria produced by uropathogenic S. marcescens, different from all previously described fimbriae in Serratia. Fimbrial purification and subsequent study using PAGE demonstrate that the material obtained was a constituent of the fimbriae. The first attempts using non-fimbriated strain E. coli HI3101 as recipient and the plasmid pBR328 as vector gave puzzling results, until a new cloning strategy was used. The isolation of the clone GH42-pSF192, able to produce fimbriae, whose properties were identical to those produced by S. marcescens ATCC 43821, demonstrated that all characteristics described (agglutination, hydrophobicity, adherence activity) could be attributed to MILT fimbriae. The modifications of hydrophobicity by temperature suggested that some effect on Serratia virulence could be modi-
697
fied in this way, since high hydrophobicity could be detected when bacteria grew at low temperatures (outside the human body), which should favour adhesion, whereas this hydrophobicity decreased (about 20%) when bacteria grew at 37C (in the human body). This property could be related to a reduction in the number of fimbriae expressed on the bacterial surface. From a practical point of view, the hydrophobicity at 37C (human infections) shown by ATCC 43821 could be used as a protection against this strain itself. Catheter-associated bacteriuria, the main cause of nosocomial infections and septic pictures, could easily be prevented using hydrophilic-coated catheters. The strong electronegative fields created by the catheter hydrophilic-layer and the bacterial surface MWT fimbriated will probably avoid the adhesion due to the repulsion forces. According to the results, MR/T fimbriae expressed by S. marcescens can be classified as able t o adhere to human uroepithelial cells. In theory, MILT-fimbriated strains are perfectly adapted for urinary tract invasion, especially when the host-defense mechanisms are diminished due to underlying urologic or systemic diseases. Acknowledgments. The technical support of 'Servei de Microscopia electronica" of the University of Barcelona is gratefully acknowledged, and we thank Elisenda Viiias for her collaboration and Robin Rycroft for his help in the preparation of the manuscript. REFERENCES
1. Maki, D. G., Hennekens, C. G., Phillips, C. W., Shaw, W. V. and Bennet, J. V.: Nosocomial urinary tract infection with Serratia marcescens: a n epidemiological study. J. Infect. Dis., 128: 579, 1973. 2. Grimont, F. and Grimont, P. A. D.: The genus Serratia. In: The Prokaryotes. Edited by A. Ballows, H. G. Triipper, M. Dworkin, W. H. Tno and K. H. Schleifer. New York: SpringerVerlag, vol. 2, chapt. 150, pp. 2822-2448, 1992. 3. Trias, J., Vifias, M., Guinea, J. and Loren, J. G.: Isolation from urine of two Serratia marcescens strains excreting a diffusible yellow pigment. J. Gen. Microbiol., 133 773, 1987. 4. Trias, J., Viiias, M., Guinea, J. and Loren, J. G.: Induction of yellow pigmentation in Serratia marcescens. Appl. Environ. Microbiol., 54: 3138, 1988. 5. Trias, J., Vifias, M., Guinea, J. and Loren, J. G.: Brown pigmentation in Serratia marcescens cultures associated with tyrosine metabolism. Can. J. Microbiol., 35: 1037, 1989. 6. Minuth, J. N., Musher, D. M. and Thorsteinsson, S. B.: Inhibition of antibacterial activity of gentamicin by urine. J. Infect. Dis., 133: 14, 1976. 7. Dodd, D. C. and Eisenstein, B. I.: Antigenic quantitation of type 1 fimbriae on the surface of Escherichia colt cells by an enzyme-linked immunosorbent inhibition assay. Infect. Immun., 38 764, 1982. 8. Laemmli, U. K.: Cleavage of structural proteins during the assembly ofthe head ofbacteriophage T4. Nature (London),227: 680,1970. 9. Ames, G. F. L., Suphch, E. N. and Nikaido, H.: Protein composition of the outer membrane of Salmonella typhimurium: effect of lipopolysaccharide mutations. J. Bacteriol., 117: 406, 1974. 10. Leranoz, A. M., Fuste, M. C., Hull, R. A. and Williams, R. P.: Cloning and expression in Escherichia coli of a gene encoding proline oxidase of Serratia marcescens. Microbios, 67: 87. 1991. 11. Karr, J. F., Nowick, B., Truong, L. D., Hull, R. and Hull, S.: Purified P fimbriae from two cloned gene clusters of a single pyelonephritogenic strain adhere to unique structures in the human ludney. Infect. Immun., 57: 3594, 1989. 12. Haraguchi, G. E., Hull, R. A., Krallmamm-Wenzel, U. and Hull, S. I.: Molecular cloning and expression of the 0 4 polysaccharide gene cluster from Escherichia coli. Microb. Pathogen., 6 123, 1989. 13. Iimura, Y., Har, S. and Otsuka, K.: Cell surface hydrophobicity as a pellicle formation factor in film strain Saccharomyces. Agric. Biol. Chem., 44: 1215, 1980.
698
NEW FIMBRIAE IN GENUS SERRATIA IMPLICATED IN UTI
14. Rosenberg, M., Gutnick, D. and Rosenberg, E.: Adherence of
bacteria to hydrocarbons: a simple method for measuring cellsurface hydrophobicity. FEMS Microbiol. Lett., 9 29, 1980. 15. Schaeffer, A. J., Amundsen, S. K. and Schmidt, L. N.: Adherence of Escherichia coli to human urinary tract epithelial cells. Infect. Immun., 24: 753, 1979. 16. Palomar, J., Leranoz, A. M. and Viaas, M.: Serratia marcescens adherence: the effect of 0-antigen presence. Microbios, 81: 107, 1995.
17. Karkktiinen, U., Ikaheimo, R., Katila, M. and Mantyjarvi, R.: p fimbriation of Escherichia coli strains from patients with UTOsepsis demonstrated by a commercial agglutination test (PF TEST). J. Clin. Microbiol., 2 9 221, 1991. 18. Soberon, X., Covarmbias, L. and Bolivar, F.: Construction and characterization of new cloning vehicles, IV. Delection denvates of pBR322 and pBR325. Gene, 9 287, 1980. 19. Sambrook, J.,Fritsch, E. F. and Maniatis, T.: Molecular Cloning. New York: Cold Spring Harbor Laboratory, vol. 3, p. 1, 1989.