Bacteriocin Typing of Serratia marcescens A Simplified System DAVID SPROAT, M.S., AND ARNOLD BROWN, M.D.
EPIDEMIOLOGIC SURVEYS frequently necessitate typing of microbial isolates to establish transmission patterns. Of the methods available, bacteriophage typing, serotyping, and bacteriocin typing have proven most reliable.3-4-510,15 Major problems with these methods include lack of availability of antisera, bacteriophage or indicator strains, the need for specialized equipment, and the time-consuming nature of the procedures." Thus, their use is often limited to a research laboratory. We describe a simplified method for bacteriocin typing that should prove useful in the community hospital setting. Our method, applied specifically to Serratia marcescens, takes advantage of the property that marcescins (bacteriocins produced by Serratia) may inhibit the growth of species other than S. marcescens. In designing our typing method, we took advantage of the fact that S. marcescens cannot Received September 21, 1977; received revised manuscript and accepted for publication January 13, 1978. Supported by the Medical Research Service of the Veterans Administration. Address reprint requests to Dr. Brown: Infectious Diseases, Veterans Administration Hospital, University Drive C, Pittsburgh, Pennsylvania 15240.
Infectious Diseases and Microbiology Section, Veterans Administration Hospital, and Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
use L-arabinose as a carbon source but other enteric bacteria can. 3 This allowed testing for bacteriocin activity with a minimum of preparation. Materials and Methods Bacterial strains We tested 116 clinical isolates of 5. marcescens from 71 patients. Eighty-eight potential indicator strains of Klebsiella pneumoniae, Enterobacter aerogenes, Enterobacter cloacae, Citrobacter diversus, Citrobacter freundii, Escherichia coli, Serratia rubidaea, anAPseudomonas aeruginosa were also obtained from clinical specimens at the Veterans Administration Hospital, Pittsburgh. The serotypes of all S. marcescens isolates were determined by the method of Edwards and Ewing 3 with the use of 15 anti-0 and 13 anti-H sera obtained from Lee Laboratories, Grayson, Georgia. Media Cultures were maintained on nutrient agar slants and cloned by picking single colonies from nutrient agar plates. For prolonged storage, the cultures were maintained at - 7 0 C in nutrient broth with 15% glycerol. For the assay, freshly prepared Vogel and Bonner 16 minimal medium plates without citrate were supplemented with 1% L-arabinose as the sole carbon source. Bacteriocin
Assay
Pure cultures of the S. marcescens to be typed were inoculated into 5-ml tryptic soy broth* and were incubated in 16 x 25-mm screw capped tubes at 37 C overnight. The culture was then mixed with an orbital mixer, and 0.3 ml was withdrawn into a tuberculin * Difco, Detroit, Michigan.
0002-9173/79/0200/0172 $00.75 © American Society of Clinical Pathologists
172
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Sproat, David, and Brown, Arnold: Bacteriocin typing of Serratia marcescens. A simplified system. Am J Clin Pathol 71: 172-176, 1979. The authors describe a simplified system for the detection of bacteriocin production by Serratia marcescens with the use of six indicator strains, which include Escherichia coli, Klebsiella pneumoniae, Citrobacter diversus, Enterobacter aerogenes (two strains), and Serratia rubidaea grown on arabinose minimal medium plates. Of the 64 possible bacteriocin types, 11 were observed; 66% of the isolates tested were found to be one of three types. Occasionally more than one bacteriocin type was observed in an individual specimen; however, serotyping or antibiograms, or both, also indicated that this was a different strain. The marcescin types were stable markers. With the use of this technic, different endemic strains of Serratia were shown to predominate in various areas of the hospital. In addition, when urinary tract isolates were compared with respiratory tract isolates, significant differences were found in the predominate types. The typing of these isolates by bacteriocin production was supported by serotype and antibiotypefindings.The results suggest that this simple system may be a useful tool in a general hospital. (Key words: Serratia marcescens; Bacteriocin typing.)
Vol. 71 • No. 2
BACTERIOCIN TYPING OF S.
Results All strains of S. marcescens tested produced a substance that inhibited the growth of the indicators, rubidaea. This indicator was for this reason included as a positive control. Ps. aeruginosa was not inhibited by our strains. A number of potential indicator strains were tested, and six were finally selected for our typing assay (1) because of their discriminatory ability with the use of a test panel of 24 randomly selected strains of S. marcescens, and (2) because of the clarity of reaction. By dilution of suspensions of several Serratia "bacteriocin producers," bacteriophagelike plaques were not seen. In addition, while the use of mitomycin C induction increased the amount of bacteriocin produced, strains that were negative on a particular indicator could not be converted to positive. We routinely tested 25 isolates on a single indicator lawn in a 100-mm Petri dish. Figure 1 shows such a plate. Against a dark background the spotted area appears slightly opaque in the photograph, but is was free of viable bacterial growth on subculture. Most of the positive spots were surrounded by halos, suggesting the
Table I. Indicator Strains and Typical Reactions A. Indicator Strains and Code for Designation of Patterns Indicator Strain
Presence of Zone of Lysis
E. coli 682 B C. diversus 260B E. aerogenes 311A 0
Value E, aerogenes 506D K. pneumoniae 20219 5. rubidaea 302B Value
0
+ + + + - - -
+ - + +
+ +
+ + +
1 2
3
4
6
7
+ +
— + - + +
— + +
+ + +
3
4
6
7
+
— +
-
-
-
1 2
5
5
B. Typical Reactions and Their Designation E. coli 682B C. diversus 260B E. aerogenes 311A
-
0
+ }3
E. aerogenes 506D K. pneumoniae 20219 5. rubidaea 302B
- J4
4
Composite designation
:i. + + j
04
34
36
presence of a diffusible, not cell-bound, bacteriocin. Table 1 shows the panel of indicator strains used in our system, typical reactions, and their numeric designations as patterns. Table 2 shows the bacteriocin types associated with various isolates from different sites. Only nonredundant isolates are included (a type was counted only once for each site of each patient). Eleven types were identified; however, 66% of the isolates were one of three predominate types and 84% were represented by five bacteriocin types. In addition, 65% of isolates of urine were type 04, type 34, or type 33; while 64% of respiratory-tract isolates were type 36 or type 06. All blood isolates were of the same type; however, three of the five were associated with a point-source outbreak, due to a contaminated arteriography injection pump. Table 3 shows the bacteriocin type of isolates obtained from the intensive care units, the medical wards, and the surgical wards. Different types predominated for isolated from various sites in different hospital areas. Sixty-nine per cent of the isolates from the intensive care units were bacteriocin type 34, 04, or 36; 63% from the medical wards were type 36 or 04; 68% from the surgical wards were type 04, 34, 47, or 36. Table 4 demonstrates the correlation between bacteriocin type, serotype, and site of isolation from a single patient from whom two or more isolates were obtained. Significant differences, involving bacteriocin types, serotypes, and antibiotic susceptibilities were observed between urinary-tract isolates and respira-
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
syringe (26-gauge needle), air was expelled, and one freely falling drop (about 0.006 ml) was placed on a marked area of each plate containing the bacterial indicator lawns. For some experiments the strains of 5. marcescens were induced with 5 fig mitomycin C/ml.4 The indicator strains were grown from single colonies inoculated into tryptic soy broth and incubated for 2-3 hours at 37 C. The final turbidity was adjusted to equal a 0.5 MacFarland turbidity standard. The indicator strain was then streaked with a cotton-tipped swab onto the arabinose minimal-medium plate. Immediately after streaking the indicator strain the S. marcescens isolates were spotted as indicated above and incubation was begun. After incubation overnight at 37 C, the area was scored as follows: clear spot (print can be read through area) with distinct margins was considered positive (+); when the plaque was hazy or margin indistinct, it was considered weakly positive (±); when there was no evidence of clearing, it was considered negative. Weakly positive reactions were relatively uncommon (less than 5%) and, when retested several times, results usually fell into the positive or negative group. The clarity of the reaction may be improved by growing the indicator strains in minimal medium or washing with saline solution to minimize the carry-over of nutrients that may be utilized by the test strains. Six indicator strains (Table 1) were selected from 88 studied. Table 1 shows how the bacteriocin production patterns were converted into a two-digit code, designated the bacteriocin type.
173
MARCESCENS
174
SPROAT AND BROWN
A.J.C.P. • February 1979
tory-tract isolates of the endemic S. marcescens associated with nosocomial infections at our hospital. In Patient 1, colonization began in the respiratory tract with bacteriocin type 26, serotype 014:hl2 (a common respiratory-tract serotype in our hospital). When urinary-tract colonization subsequently occurred, it was with marcescin type 04, serotype 0 indeterminate:H indeterminate or 0 indeterminate:H2, which is a typical urinary-tract type at our hospital. The urine isolate with the bacteriocin type 77, while still serotype 0 indeterminate:H2 had an antibiotic sensitivity pattern different from those of the other 0 indeterminate :H2 isolates. Patient 2 had isolates with similar differences. The marcescin data support other observations we have made regarding the differences between urinary-
tract and respiratory-tract isolates of S. marcescens.la These differences of bacteriocin types, serotypes, and antibiotypes are stable after many nonselective in vitro passages. Discussion We have demonstrated the utility of a bacteriocintyping system using six indicator strains, which include K. pneumoniae, E. aerogenes, C. diversus, E. colt, and 5. rubidaea (Table 1). The indicator strains of K. pneumoniae and E. aerogenes were more likely to be susceptible to the bacteriocin activity associated with respiratory-tract S. marcescens isolates than with those associated with isolates from other sites. Sensitivity may be increased by incorporating other indicator
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
FIG. 1. Petri dish containing arabinose minimal medium with a lawn of K. pneumoniae on which 25 isolates of S. marcescens have been spotted.
Vol. 71 • No. 2
175
BACTERIOCIN TYPING OF S. MARCESCENS
Table 2. Serratia marcescens Bacteriocin Types vs. Sources of Strains Bacteriocin Type
Urine
Sputum
Blood
Other
Total
04 34 36 77 06 47 26 57 14 45 27
15 10 4 7 4 5 0 1 1 1 1
2 3 12 1 4 0 2 0 1 0 0
0 5 0 0 0 0 0 0 0 0 0
3 2 1 0 0 0 1 1 0 0 0
20 20 17 8 8 5 3 2 2 1 1
TOTAL
49
25
5
8
87
Table 3. Marcescins Types of Isolates from Three Major Hospital Areas Bacteriocin Type Intensive care units
Number
34 04 36 26 06 77
5 3 3 2 2 1 16
TOTAL
Medical service
36 04 06 34 47 14
8 4 3 2 1 1
One hundred sixteen isolates were tested. Only 'nonredundant' isolates are included.
19
TOTAL
While increasing the number of strains would not make the system overly cumbersome, we present what we feel is the simplest system with enough sensitivity for community hospitals. We cannot exclude the possibility that bacteriophage might be responsible for the reactions of some of our isolates. However, this could not be demonstrated in the several isolates that we examined for this possibility, and we feel that this should not detract from the method's usefulness in epidemiologic typing. The study of S. marcescens isolates obtained from different sites of the same patient shows that a patient
04 34 47 36 45 77 57 06
6 3 3 3 2 2 2 I 22
TOTAL
may harbor several types with quite different properties. The data also suggest that even in these patients certain types are more often present at specific sites. This may be due to different reservoirs and modes of
Table 4. Correlation between Marcescin Type, Serotype, Antibiogram and Site for Specimens from Two Patients Isolate
Date of Isolation
Source
Marcescin Type
Serotype*
Resistancet Pattern
Patient 1 1 2 3 4 5 6 7 8 9 10 11 12
4/24 4/26 5/1 5/5 5/13 5/24 5/25 6/7 6/10 6/10 6/16 6/16
Sputum Sputum Foley tip Urine Urine Sputum Sputum Urine Urine Sputum Urine Sputum
26 26 26 04 ND 04 26 04 77 26 04 ND
014:H12 014:H12 014:H12 0-:H0-:H2 02: HI 03,14:H12 0-:H2 0-:H2 03,14:H12 0-:H2 014:H12
cb c gm ka str nnt cb c gm ka str nnt cb c ka str nn$ c ka nnt cb c gm ka str nnt cb gm ka str nnt cb c gm ka nnt cb gm ka str nnj ka str nnt cb c ka nn$ cb gm ka str nn$ cb c ka str nnt
Patient 2 1 2 3 4 5 6
2/26 3/1 3/9 3/10 3/13 3/13
Urine Urine Sputum Sputum Sputum Sputum
04 04 ND 36 06 36
04: H2 04: H2 014:H12 014:H12 014:H0-:H12
gm str te§ gm str te§ c te§ c§ c§ c te§
* 0- or H- indicates that the strain failed to be agglutinated by any of the anti-0 or anti-H sera. 03, 14 indicates reaction with both antisera, a reaction that was not due to mixed culture. ND indicates not done (isolate not available). t Antibiotics tested include ampicillin, carbenicillin (cb), cephalothin, chloramphenicol (c), gentamicin (gm), kanamycin (ka), streptomycin (str), tobramycin (nn), and tetra-
cycline (te). t Isolates were also resistant to ampicillin, cephalothin. and tetracycline. § Isolates were also resistant to ampicillin, carbenicillin, cephalothin, kanamycin, and tobramycin.
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Surgical service
176
SPROAT AND BROWN
Acknowledgment. Ms. Louise Davis performed the serotyping (as part of her master's thesis).
2.
3.
4.
5.
6.
7.
8.
9.
10. 11.
12.
13. 14.
References 15. la. Brown A, Davis L, Yee RB, et al: Endemic Serratia marcescens in the Veterans Administration Hospital in Pittsburgh, Pa. 1971-1976. Health Lab Sci 15:159-167, 1978 lb. Cooksey RC, Thorne GM, Farrar WE Jr: R-factor mediated
16.
antibiotic resistance in Serratia marcescens. Antimicrob Agents Chemother 10:123-127, 1976 Dajani AS, Tom MC, Law DJ: Viridins, bacteriocins of alphahemolytic Streptococci: Isolation, characterization, and partial purification. Antimicrob Agents Chemother 9:81-88, 1976 Edwards PR, Ewing WH: Identification of Enterobacteriacease. Third edition. Minneapolis, Burgess Publishing Company, 1972 Farmer JJ III: Epidemiological differentiation of Serratia marcescens: Typing by bacteriocin production. Appl Microbiol 23:218-255, 1972 Farmer JJ III: Epidemiological differentiation of Serratia marcescens: Typing by bacteriocin sensitivity. Appl Microbiol 23:226-231, 1972 Hale EM, Hindsdill RD: Biological activity of staphylococcin 462: Bacteriocin from Staphylococcus aureus. Antimicrob Agents Chemother 7:74-81, 1975 Kawai X: Purification and characterization of pertucin produced by Pseudomonas pertucinogena. Antimicrob Agents Chemother 6:347-359, 1974 Krammes J, Brandis H: Mode of action of two Streptococcus faecium bacteriocins. Antimicrob Agents Chemother 7:117120, 1975 Morse SA, Vaughan P, Johnson D, et al: Inhibition of Neisseria gonorrhoeae by a bacteriocin from Pseudomonas aeruginosa. Antimicrob Agents Chemother 10:354-362, 1976 Pillich J, Hradecna Z, Kocur M: An attempt at phage typing in the genus Serratia. J Appl Bacterid 27:65-68, 1964. Roemisch BA, Kocka FE: Comparison of methods for differentiating among Serratia marcescens isolated from clinical specimens. Am J Clin Pathol 66:96-100, 1976 Rogers AH: Bacteriocin production and susceptibility among strains of Streptococcus mutans grown in the presence of sucrose. Antimicrob Agents Chemother 5:547-550, 1974 Sproat D, Brown A: Abstract # D 8 , Annual Meeting of the American Society for Microbiology, 1976 TaggJR, Dajani AS, Wannamaker LW: Bacteriocin of a group B Streptococcus: Partial purification and characterization. Antimicrob Agents Chemother 7:764-772, 1975 Traub WH, Raymond EA, Startsman TS: Bacteriocin (marcescen) typing of clinical isolates of Serratia marcescens. Appl Microbiol 21:837-840, 1971 Vogel HJ, Bonner DM: Acetylornithinase of Escherichia coli: Partial purification and some properties. J Biol Chem 218: 97-106, 1956
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
transmission unique to each site, to modification of the infective strain after acquisition at a site, or to the acquisition of strains with a specific site predilection, which may even be capable of replacing other colonizing strains. The production of these bacteriocins may interfere with the ability to demonstrate resistance plasmid (Rfactor)-mediated transfer of antibiotic resistance to a susceptible strain, as has been previously reported. lbl3 The unique aspect of the bacteriocin-typing method described is the use of indicator organisms of a genus or species other than the strains to be typed. This makes it possible to use selective media to prevent growth of the strains examined for bacteriocin production, thereby circumventing the necessity to remove viable cells by centrifugation or treatment with chloroform. An expanded host range has previously been described not only for the bacteriocins oi Serratia,4A but for those of Pseudomonas,1,9 streptococci ,2,SA2A4 and staphylococci6 as well. This property has not been systematically examined for epidemiologic usefulness, and may allow similar typing systems to be developed with other organisms. Indicator strains will be sent to interested investigators on request.
A.J.C.P. • February 1979
EPIDEMIOLOGIC SURVEYS frequently necessitate typing of microbial isolates to establish transmission patterns. Of the methods available, bacteriophage typing, serotyping, and bacteriocin typing have proven most reliable.3-4-510,15 Major problems with these methods include lack of availability of antisera, bacteriophage or indicator strains, the need for specialized equipment, and the time-consuming nature of the procedures." Thus, their use is often limited to a research laboratory. We describe a simplified method for bacteriocin typing that should prove useful in the community hospital setting. Our method, applied specifically to Serratia marcescens, takes advantage of the property that marcescins (bacteriocins produced by Serratia) may inhibit the growth of species other than S. marcescens. In designing our typing method, we took advantage of the fact that S. marcescens cannot Received September 21, 1977; received revised manuscript and accepted for publication January 13, 1978. Supported by the Medical Research Service of the Veterans Administration. Address reprint requests to Dr. Brown: Infectious Diseases, Veterans Administration Hospital, University Drive C, Pittsburgh, Pennsylvania 15240.
Infectious Diseases and Microbiology Section, Veterans Administration Hospital, and Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
use L-arabinose as a carbon source but other enteric bacteria can. 3 This allowed testing for bacteriocin activity with a minimum of preparation. Materials and Methods Bacterial strains We tested 116 clinical isolates of 5. marcescens from 71 patients. Eighty-eight potential indicator strains of Klebsiella pneumoniae, Enterobacter aerogenes, Enterobacter cloacae, Citrobacter diversus, Citrobacter freundii, Escherichia coli, Serratia rubidaea, anAPseudomonas aeruginosa were also obtained from clinical specimens at the Veterans Administration Hospital, Pittsburgh. The serotypes of all S. marcescens isolates were determined by the method of Edwards and Ewing 3 with the use of 15 anti-0 and 13 anti-H sera obtained from Lee Laboratories, Grayson, Georgia. Media Cultures were maintained on nutrient agar slants and cloned by picking single colonies from nutrient agar plates. For prolonged storage, the cultures were maintained at - 7 0 C in nutrient broth with 15% glycerol. For the assay, freshly prepared Vogel and Bonner 16 minimal medium plates without citrate were supplemented with 1% L-arabinose as the sole carbon source. Bacteriocin
Assay
Pure cultures of the S. marcescens to be typed were inoculated into 5-ml tryptic soy broth* and were incubated in 16 x 25-mm screw capped tubes at 37 C overnight. The culture was then mixed with an orbital mixer, and 0.3 ml was withdrawn into a tuberculin * Difco, Detroit, Michigan.
0002-9173/79/0200/0172 $00.75 © American Society of Clinical Pathologists
172
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Sproat, David, and Brown, Arnold: Bacteriocin typing of Serratia marcescens. A simplified system. Am J Clin Pathol 71: 172-176, 1979. The authors describe a simplified system for the detection of bacteriocin production by Serratia marcescens with the use of six indicator strains, which include Escherichia coli, Klebsiella pneumoniae, Citrobacter diversus, Enterobacter aerogenes (two strains), and Serratia rubidaea grown on arabinose minimal medium plates. Of the 64 possible bacteriocin types, 11 were observed; 66% of the isolates tested were found to be one of three types. Occasionally more than one bacteriocin type was observed in an individual specimen; however, serotyping or antibiograms, or both, also indicated that this was a different strain. The marcescin types were stable markers. With the use of this technic, different endemic strains of Serratia were shown to predominate in various areas of the hospital. In addition, when urinary tract isolates were compared with respiratory tract isolates, significant differences were found in the predominate types. The typing of these isolates by bacteriocin production was supported by serotype and antibiotypefindings.The results suggest that this simple system may be a useful tool in a general hospital. (Key words: Serratia marcescens; Bacteriocin typing.)
Vol. 71 • No. 2
BACTERIOCIN TYPING OF S.
Results All strains of S. marcescens tested produced a substance that inhibited the growth of the indicators, rubidaea. This indicator was for this reason included as a positive control. Ps. aeruginosa was not inhibited by our strains. A number of potential indicator strains were tested, and six were finally selected for our typing assay (1) because of their discriminatory ability with the use of a test panel of 24 randomly selected strains of S. marcescens, and (2) because of the clarity of reaction. By dilution of suspensions of several Serratia "bacteriocin producers," bacteriophagelike plaques were not seen. In addition, while the use of mitomycin C induction increased the amount of bacteriocin produced, strains that were negative on a particular indicator could not be converted to positive. We routinely tested 25 isolates on a single indicator lawn in a 100-mm Petri dish. Figure 1 shows such a plate. Against a dark background the spotted area appears slightly opaque in the photograph, but is was free of viable bacterial growth on subculture. Most of the positive spots were surrounded by halos, suggesting the
Table I. Indicator Strains and Typical Reactions A. Indicator Strains and Code for Designation of Patterns Indicator Strain
Presence of Zone of Lysis
E. coli 682 B C. diversus 260B E. aerogenes 311A 0
Value E, aerogenes 506D K. pneumoniae 20219 5. rubidaea 302B Value
0
+ + + + - - -
+ - + +
+ +
+ + +
1 2
3
4
6
7
+ +
— + - + +
— + +
+ + +
3
4
6
7
+
— +
-
-
-
1 2
5
5
B. Typical Reactions and Their Designation E. coli 682B C. diversus 260B E. aerogenes 311A
-
0
+ }3
E. aerogenes 506D K. pneumoniae 20219 5. rubidaea 302B
- J4
4
Composite designation
:i. + + j
04
34
36
presence of a diffusible, not cell-bound, bacteriocin. Table 1 shows the panel of indicator strains used in our system, typical reactions, and their numeric designations as patterns. Table 2 shows the bacteriocin types associated with various isolates from different sites. Only nonredundant isolates are included (a type was counted only once for each site of each patient). Eleven types were identified; however, 66% of the isolates were one of three predominate types and 84% were represented by five bacteriocin types. In addition, 65% of isolates of urine were type 04, type 34, or type 33; while 64% of respiratory-tract isolates were type 36 or type 06. All blood isolates were of the same type; however, three of the five were associated with a point-source outbreak, due to a contaminated arteriography injection pump. Table 3 shows the bacteriocin type of isolates obtained from the intensive care units, the medical wards, and the surgical wards. Different types predominated for isolated from various sites in different hospital areas. Sixty-nine per cent of the isolates from the intensive care units were bacteriocin type 34, 04, or 36; 63% from the medical wards were type 36 or 04; 68% from the surgical wards were type 04, 34, 47, or 36. Table 4 demonstrates the correlation between bacteriocin type, serotype, and site of isolation from a single patient from whom two or more isolates were obtained. Significant differences, involving bacteriocin types, serotypes, and antibiotic susceptibilities were observed between urinary-tract isolates and respira-
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
syringe (26-gauge needle), air was expelled, and one freely falling drop (about 0.006 ml) was placed on a marked area of each plate containing the bacterial indicator lawns. For some experiments the strains of 5. marcescens were induced with 5 fig mitomycin C/ml.4 The indicator strains were grown from single colonies inoculated into tryptic soy broth and incubated for 2-3 hours at 37 C. The final turbidity was adjusted to equal a 0.5 MacFarland turbidity standard. The indicator strain was then streaked with a cotton-tipped swab onto the arabinose minimal-medium plate. Immediately after streaking the indicator strain the S. marcescens isolates were spotted as indicated above and incubation was begun. After incubation overnight at 37 C, the area was scored as follows: clear spot (print can be read through area) with distinct margins was considered positive (+); when the plaque was hazy or margin indistinct, it was considered weakly positive (±); when there was no evidence of clearing, it was considered negative. Weakly positive reactions were relatively uncommon (less than 5%) and, when retested several times, results usually fell into the positive or negative group. The clarity of the reaction may be improved by growing the indicator strains in minimal medium or washing with saline solution to minimize the carry-over of nutrients that may be utilized by the test strains. Six indicator strains (Table 1) were selected from 88 studied. Table 1 shows how the bacteriocin production patterns were converted into a two-digit code, designated the bacteriocin type.
173
MARCESCENS
174
SPROAT AND BROWN
A.J.C.P. • February 1979
tory-tract isolates of the endemic S. marcescens associated with nosocomial infections at our hospital. In Patient 1, colonization began in the respiratory tract with bacteriocin type 26, serotype 014:hl2 (a common respiratory-tract serotype in our hospital). When urinary-tract colonization subsequently occurred, it was with marcescin type 04, serotype 0 indeterminate:H indeterminate or 0 indeterminate:H2, which is a typical urinary-tract type at our hospital. The urine isolate with the bacteriocin type 77, while still serotype 0 indeterminate:H2 had an antibiotic sensitivity pattern different from those of the other 0 indeterminate :H2 isolates. Patient 2 had isolates with similar differences. The marcescin data support other observations we have made regarding the differences between urinary-
tract and respiratory-tract isolates of S. marcescens.la These differences of bacteriocin types, serotypes, and antibiotypes are stable after many nonselective in vitro passages. Discussion We have demonstrated the utility of a bacteriocintyping system using six indicator strains, which include K. pneumoniae, E. aerogenes, C. diversus, E. colt, and 5. rubidaea (Table 1). The indicator strains of K. pneumoniae and E. aerogenes were more likely to be susceptible to the bacteriocin activity associated with respiratory-tract S. marcescens isolates than with those associated with isolates from other sites. Sensitivity may be increased by incorporating other indicator
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
FIG. 1. Petri dish containing arabinose minimal medium with a lawn of K. pneumoniae on which 25 isolates of S. marcescens have been spotted.
Vol. 71 • No. 2
175
BACTERIOCIN TYPING OF S. MARCESCENS
Table 2. Serratia marcescens Bacteriocin Types vs. Sources of Strains Bacteriocin Type
Urine
Sputum
Blood
Other
Total
04 34 36 77 06 47 26 57 14 45 27
15 10 4 7 4 5 0 1 1 1 1
2 3 12 1 4 0 2 0 1 0 0
0 5 0 0 0 0 0 0 0 0 0
3 2 1 0 0 0 1 1 0 0 0
20 20 17 8 8 5 3 2 2 1 1
TOTAL
49
25
5
8
87
Table 3. Marcescins Types of Isolates from Three Major Hospital Areas Bacteriocin Type Intensive care units
Number
34 04 36 26 06 77
5 3 3 2 2 1 16
TOTAL
Medical service
36 04 06 34 47 14
8 4 3 2 1 1
One hundred sixteen isolates were tested. Only 'nonredundant' isolates are included.
19
TOTAL
While increasing the number of strains would not make the system overly cumbersome, we present what we feel is the simplest system with enough sensitivity for community hospitals. We cannot exclude the possibility that bacteriophage might be responsible for the reactions of some of our isolates. However, this could not be demonstrated in the several isolates that we examined for this possibility, and we feel that this should not detract from the method's usefulness in epidemiologic typing. The study of S. marcescens isolates obtained from different sites of the same patient shows that a patient
04 34 47 36 45 77 57 06
6 3 3 3 2 2 2 I 22
TOTAL
may harbor several types with quite different properties. The data also suggest that even in these patients certain types are more often present at specific sites. This may be due to different reservoirs and modes of
Table 4. Correlation between Marcescin Type, Serotype, Antibiogram and Site for Specimens from Two Patients Isolate
Date of Isolation
Source
Marcescin Type
Serotype*
Resistancet Pattern
Patient 1 1 2 3 4 5 6 7 8 9 10 11 12
4/24 4/26 5/1 5/5 5/13 5/24 5/25 6/7 6/10 6/10 6/16 6/16
Sputum Sputum Foley tip Urine Urine Sputum Sputum Urine Urine Sputum Urine Sputum
26 26 26 04 ND 04 26 04 77 26 04 ND
014:H12 014:H12 014:H12 0-:H0-:H2 02: HI 03,14:H12 0-:H2 0-:H2 03,14:H12 0-:H2 014:H12
cb c gm ka str nnt cb c gm ka str nnt cb c ka str nn$ c ka nnt cb c gm ka str nnt cb gm ka str nnt cb c gm ka nnt cb gm ka str nnj ka str nnt cb c ka nn$ cb gm ka str nn$ cb c ka str nnt
Patient 2 1 2 3 4 5 6
2/26 3/1 3/9 3/10 3/13 3/13
Urine Urine Sputum Sputum Sputum Sputum
04 04 ND 36 06 36
04: H2 04: H2 014:H12 014:H12 014:H0-:H12
gm str te§ gm str te§ c te§ c§ c§ c te§
* 0- or H- indicates that the strain failed to be agglutinated by any of the anti-0 or anti-H sera. 03, 14 indicates reaction with both antisera, a reaction that was not due to mixed culture. ND indicates not done (isolate not available). t Antibiotics tested include ampicillin, carbenicillin (cb), cephalothin, chloramphenicol (c), gentamicin (gm), kanamycin (ka), streptomycin (str), tobramycin (nn), and tetra-
cycline (te). t Isolates were also resistant to ampicillin, cephalothin. and tetracycline. § Isolates were also resistant to ampicillin, carbenicillin, cephalothin, kanamycin, and tobramycin.
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Surgical service
176
SPROAT AND BROWN
Acknowledgment. Ms. Louise Davis performed the serotyping (as part of her master's thesis).
2.
3.
4.
5.
6.
7.
8.
9.
10. 11.
12.
13. 14.
References 15. la. Brown A, Davis L, Yee RB, et al: Endemic Serratia marcescens in the Veterans Administration Hospital in Pittsburgh, Pa. 1971-1976. Health Lab Sci 15:159-167, 1978 lb. Cooksey RC, Thorne GM, Farrar WE Jr: R-factor mediated
16.
antibiotic resistance in Serratia marcescens. Antimicrob Agents Chemother 10:123-127, 1976 Dajani AS, Tom MC, Law DJ: Viridins, bacteriocins of alphahemolytic Streptococci: Isolation, characterization, and partial purification. Antimicrob Agents Chemother 9:81-88, 1976 Edwards PR, Ewing WH: Identification of Enterobacteriacease. Third edition. Minneapolis, Burgess Publishing Company, 1972 Farmer JJ III: Epidemiological differentiation of Serratia marcescens: Typing by bacteriocin production. Appl Microbiol 23:218-255, 1972 Farmer JJ III: Epidemiological differentiation of Serratia marcescens: Typing by bacteriocin sensitivity. Appl Microbiol 23:226-231, 1972 Hale EM, Hindsdill RD: Biological activity of staphylococcin 462: Bacteriocin from Staphylococcus aureus. Antimicrob Agents Chemother 7:74-81, 1975 Kawai X: Purification and characterization of pertucin produced by Pseudomonas pertucinogena. Antimicrob Agents Chemother 6:347-359, 1974 Krammes J, Brandis H: Mode of action of two Streptococcus faecium bacteriocins. Antimicrob Agents Chemother 7:117120, 1975 Morse SA, Vaughan P, Johnson D, et al: Inhibition of Neisseria gonorrhoeae by a bacteriocin from Pseudomonas aeruginosa. Antimicrob Agents Chemother 10:354-362, 1976 Pillich J, Hradecna Z, Kocur M: An attempt at phage typing in the genus Serratia. J Appl Bacterid 27:65-68, 1964. Roemisch BA, Kocka FE: Comparison of methods for differentiating among Serratia marcescens isolated from clinical specimens. Am J Clin Pathol 66:96-100, 1976 Rogers AH: Bacteriocin production and susceptibility among strains of Streptococcus mutans grown in the presence of sucrose. Antimicrob Agents Chemother 5:547-550, 1974 Sproat D, Brown A: Abstract # D 8 , Annual Meeting of the American Society for Microbiology, 1976 TaggJR, Dajani AS, Wannamaker LW: Bacteriocin of a group B Streptococcus: Partial purification and characterization. Antimicrob Agents Chemother 7:764-772, 1975 Traub WH, Raymond EA, Startsman TS: Bacteriocin (marcescen) typing of clinical isolates of Serratia marcescens. Appl Microbiol 21:837-840, 1971 Vogel HJ, Bonner DM: Acetylornithinase of Escherichia coli: Partial purification and some properties. J Biol Chem 218: 97-106, 1956
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
transmission unique to each site, to modification of the infective strain after acquisition at a site, or to the acquisition of strains with a specific site predilection, which may even be capable of replacing other colonizing strains. The production of these bacteriocins may interfere with the ability to demonstrate resistance plasmid (Rfactor)-mediated transfer of antibiotic resistance to a susceptible strain, as has been previously reported. lbl3 The unique aspect of the bacteriocin-typing method described is the use of indicator organisms of a genus or species other than the strains to be typed. This makes it possible to use selective media to prevent growth of the strains examined for bacteriocin production, thereby circumventing the necessity to remove viable cells by centrifugation or treatment with chloroform. An expanded host range has previously been described not only for the bacteriocins oi Serratia,4A but for those of Pseudomonas,1,9 streptococci ,2,SA2A4 and staphylococci6 as well. This property has not been systematically examined for epidemiologic usefulness, and may allow similar typing systems to be developed with other organisms. Indicator strains will be sent to interested investigators on request.
A.J.C.P. • February 1979
