Jun 4, 1979 - B-hemolysis, motility, and fermentation of raffi- nose and sucrose were the most useful in differ- entiating among strains. The 959 strains were.
JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1979, p. 275-278 0095-1137/79/09-0275/04$02.00/0
Vol. 10, No. 3
Biotyping of Escherichia coli by a Simple MultipleInoculation Agar Plate Technique FREDERICK J. BUCKWOLD,'-3 ALLAN R. RONALD,1-3* GODFREY K. M. HARDING,'-3 THOMAS J. MARRIE,1' LINDA FOX,2 AND CLAUDETTE CATES2 Department of Clinical Microbiology, Health Sciences Centre,' and Departments of Medical Microbiology2 and Medicine,3 Faculty of Medicine, University of Manitoba, Winnipeg, Canada R3E 0W3
Received for publication 4 June 1979
A nine-test system using multiple-inoculation agar plates for biotyping of Escherichia coli is described. Testing of 959 strains resulted in 78 biotypes. On repeated testing, 96% of 182 strains had identical biotypes or differed by only one test. This system provides satisfactory differentiation among strains and is reproducible. Precise standardization of inoculum size is not required. Multiple inoculation allows time and cost-efficient testing of large numbers of strains.
Management of recurrent urinary tract infections is simplified if relapse of infection with the original pathogen can be differentiated from reinfection by a new organism (9). When a recurrence is caused by the same bacterial species as the initial infection, this distinction may be difficult to make. A simple epidemiological marker to aid in differentiation of relapse from reinfection would be useful in both clinical and research settings. Escherichia coli is the most common pathogen causing urinary tract infections. Serotyping of E. coli, based on the presence of 0, K, and H antigens, has been useful in studies of urinary tract infections (4, 5, 8, 9). This technique, however, is limited by the relatively large number of organisms causing urinary infection that are untypable (15 to 20%) and by the time and cost required to perform the test. Biotyping is simpler, and all strains can be typed (10). The Analytab (API 20E) system has been used for biotyping (2); however, the reproducibility of this system is in question (1, 6). We describe a simple biotyping system that we have found useful in differentiating relapse from reinfection in recurrent E. coli urinary tract infections. This system may also be useful in investigations of the role played by the perineal and rectal flora in the pathogenesis of recurrent urinary infections.
nose, production of lysine decarboxylase, and motility; and fermentation of lactose, raffinose, and sucrose. Points are allotted for each positive test: one point for the first test in each group of three, two for the second, and four for the third. The points allotted for each group of three tests are added to give the code number. Three code numbers are produced, each varying from 0 to 7. Commercially prepared agar plates are purchased to test for lysine decarboxylase production (KVL Laboratories, Cambridge, Ontario). Sheep blood and MacConkey agar plates, used to test for f,-hemolysis and lactose fermentation, are made in the Health Sciences Centre Clinical Microbiology Laboratory by using standard methods (11). Other carbohydrate-containing agars are made in our research laboratory. Base agar containing Andrade peptone water (with Andrade indicator incorporated) and 2% Oxoid agar is prepared and autoclaved. Each carbohydrate is prepared in a 10% solution and added to the autoclaved base agar to give a final concentration of 1%. Approtimately 20 ml of the liquid agar is poured into a standard petri dish (100 by 15 mm) and allowed to harden at room temperature. Plates are stored for up to 6 weeks at 4°C until use. Motility is tested in tubes containing semisolid agar (SIM media). Isolates identified as E. coli by using standard laboratory methods are stored at -70°C before biotyping. E. coli are batched into lots of 25 to 30 to minimize test variability and to use the entire plate. Serial urine, fecal, and periurethral isolates from a patient under investigation for urinary infection are included in a single batch whenever possible. After removal from the freezer, organisms are passed twice on blood agar plates before biotyping. MATERIALS AND METHODS Tips of three to five colonies are touched with a E. coli were identified by the standard procedures straight wire, emulsified in Steers replicator wells conof Edwards and Ewing (3). A total of 17 biochemical taining 0.5 ml of Trypticase soy broth, and incubated tests were initially selected to differentiate among E. at 37°C for 3 h. The agar test plates are then inocucoli strains. After preliminary testing, a biotyping sys- lated by using the Steers replicator (inoculum size, tem was developed consisting of nine tests, divided approximately 107 at imprint), and the agar is cut with into three groups of three: fernentation of melibiose a sterile scalpel to isolate each strain and prevent and adonitol and ,B-hemolysis; fermentation of rham- cross-diffusion of positive reactions. An organism with 275
276
J. CLIN. MICROBIOL.
BUCKWOLD ET AL.
a well-established, stable biotype is included in each run as a positive control. A plate of base agar without an added carbohydrate is included with each run as a growth control. A citrate agar plate is included to confirm the initial identification of the organism as E.
coli. Motility is tested by inoculation into SIM media with a straight wire which has sampled bacteria from the top of an isolated colony. Plates and the SIM media are incubated at 37°C and are read at 24 h. ,B-Hemolysis is said to be present if there is a clear zone of any size around the colonies. Adonitol, lactose, raffinose, and sucrose tests are interpreted as positive if an appropriate color change occurs in the colonies, whereas lysine decarboxylase, melibiose, and rhamnose tests are interpreted as positive only if the entire square of agar undergoes the appropriate color change. We tested 959 strains of E. coli isolated from urine specimens and anal canal periurethral swabs obtained from patients at the Urinary Infection Clinic of the Health Sciences Centre. Methods for collection of these specimens have been described previously (7).
RESULTS The test results for 959 strains of E. coli and the ensuing commonest code numbers for each group of tests are shown in Table 1. Tests for
,B-hemolysis, motility, and fermentation of raffinose and sucrose were the most useful in differentiating among strains. The 959 strains were represented by 78 separate biotype codes. Eleven biotypes accounted for 67% of the strains tested: 171 (16%), 177 (11.5%), 131 (8.5%), 571 (8.5%), 175 (5%), 371 (4%), 135 (3%), 173 (3%), 575 (2.5%), 137 (2.5%), and 161 (2.5%). To investigate the reproducibility of the system, 182 separate strains were retested blindly at 1 to 4 months after initial biotyping, and the initial and repeat biotype codes were compared. Identical biotypes occurred in 150 strains (82.5%), one test difference occurred in 25 strains (13.5%), and two test differences occurred in the remaining 7 strains (4%). Frequency of variance for each test is shown in Table 2. Fermentation TABLE 1. Test results for 959 strains of E. coli Test
% Posi-
Pitive
Melibiose fermentation Adonitol fermentation
89 8
B8-Hemolysis
23
Rhamnose fermentation Lysine decarboxylase produc-
88 92
Motility
70
TABLE 2. Frequency of test variance in 182 E. coli strains biotyped twice No. of times
Test
varied Melibiose fermentation .5 (3)a Adonitol fermentation .0.O (0) ,B-Hemolysis ... 2 (1) 2 (1) 6 (3) 2 (1)
Rhamnose fermentation .............. Lysine decarboxylase production .
Motility ..
Lactose fermentation .0.O (0) 12 (6.5)b Raffinose fermentation .. 10 (5.5)b .... Sucrose fermentation .... a Values in parentheses are percentages. bp < 0.02 compared with ,-hemolysis, motility, and fermentation of adonitol, rhamnose, and lactose.
Common-
est code number(s)a
TABLE 3. Patient I history Culture
1, 5
Date 14 January (ini-
7, 3
tion
94 Lactose fermentation 24 1, 7, 5 Raffinose fermentation 37 Sucrose fermentation Code number occurred in >15% of instances. a
of raffinose and sucrose varied most frequently and significantly more often than f8-hemolysis, motility, and fermentation of adonitol, rhamnose, and lactose (P < 0.02, calculated by chisquare analysis). In the course of establishing this system, we investigated other biochemical and fermentation tests. Fermentations of sorbitol, maltose, and xylose were unable to differentiate among most strains, as they were positive in 92 to 98% of all isolates. Dulcitol fermentation and production of arginine dihydrolase and ornithine decarboxylase could differentiate among strains (19 to 63% of strains were positive); however, results were inconsistent, with a variance on reproducibility testing of 9 to 12%. We were unable to include fermentation of salicin or bile esculin hydrolysis in the panel because of technical difficulties. The value of this system is illustrated in its ability to distinguish between relapse and reinfection in patients with recurrent urinary infections. Two of these patients, for whom somatic antigen serotyping and biotyping were carried out, are briefly summarized. Patient 1. (i) History. This 19-year-old fe-
tial culture)a 21 January (post-
treatment)
Midstream urine E. 105 coli serotype 06, biotype 177 No growth
Rectal
E. coli serotype
01, biotype 171
E. coli serotype 105 E. coli sero01, biotype 171 type 06, biotype 577 a From 17 to 21 January the patient was treated with nalidixic acid.
6 February (re-
currence)
BIOTYPING OF E. COLI
VOL. 10, 1979
277
TABLE 4. Patient 2 history Culture Date Midstream urine
7 January (initial cul-
ture)a 14 January (during
treatment) 15 February (posttreatment) 8 March (reinfection) During treatmentc
Periurethral
>i10 E. coli serotype undetermined, biotype 470 Negative
ND'
ND
ND
ND
Negative
ND
ND
>10" E. coli serotype 075, biotype 135 Negative
ND
ND
E. coli serotype 075, biotype 135 >10" E. coli serotype E. coli serotype 075, Posttreatment 075, biotype 135 biotype 135 From 12 to 26 January the patient was treated with nalidixic acid. 'ND, Not done. 'From 9 to 23 March the patient was treated with cephalexin.
male had a history of recurrent infections for 1 year. At the initial visit she had an asymptomatic infection with E. coli serotype 06, biotype 171 (Table 3). The rectal culture grew only E. coli serotype 01, biotype 177. (ii) Interpretation. This patient had a recurrence after a short course of nalidixic acid. E. coli serotype 06, biotype 177 was found in the rectum. After therapy the organism recurred with a change of a single test difference (hemolysis). We consider this a relapse confirmed by serotype and biotype. Patient 2. (i) History. This 54-year-old female presented with asymptomatic bacteriuria due to E. coli serotype undetermined (with the sera available), biotype 470. She was treated with a 14-day course of nalidixic acid, which eradicated the infection (Table 4). (ii) Interpretation. After successful eradication of this organism she was reinfected with a new E. coli of a different serotype and biotype. This infection (E. coli serotype 075, biotype 135) relapsed after a 2-week course of cephalexin.
DISCUSSION We have described a simple system for biotyping E. coli. We believe this system has advantages over the previously described Analytab (API 20E) system. Davies found that only three of the tests in the API 20E system contributed to differentiation among strains and resulted in 55 biotypes for 574 strains of E. coli (2). Six of these biotypes accounted for almost 70% of the strains. In the system we have described, four tests contribute to differentiation among strains. The 959 strains of E. coli were divided into 78 biotypes. Eleven biotypes accounted for 67% of the strains.
Rectal
ND
E. coli serotype 075, biotype 135
Murray found that standardization of inoculum size is necessary to obtain satisfactory reproducibility of the API 20E system (6). The system we have described has good reproducibility without the necessity for more precise standardization of the inoculum. With our system, isolates that have two or more test differences on biotyping are considered to be different strains. Use of the agar plate replicator technique offers the advantage that large numbers of organisms can be tested simultaneously. All isolates from a single patient are biotyped concurrently in order to reduce variability in results. ACKNOWLEDGMENTS This work was supported by grant MA-5973 from the Medical Research Council of Canada. F.J.B. and T.J.M. were recipients of a fellowship from the Medical Research Council of Canada. LITERATURE CITED 1. Butler, D. A., C. M. Lobregat, and T. L. Gavan. 1975. Reproducibility of the Analytab (API 20E) system. J. Clin. Microbiol. 2:322-326. 2. Davies, B. I. 1977. Biochemical typing of urinary Esche-
richia coli strains by means of the API 20E Enterobacteriaceae system. J. Med. Microbiol. 10:293-298. 3. Edwards, P. R., and W. H. Ewing. 1972. Identification of Enterobacteriaceae. Burgess Publishing Co., Minneapolis, Minn. 4. Mabeck, C. E., F. 0rskov, and I. 0rskov. 1971. Studies in urinary tract infection. Acta Med. Scand. 190:279282. 5. McGeachie, J. 1965. Serological grouping of urinary Escherichia coli. J. Clin. Pathol. 18:428-431. 6. Murray, P. R. 1978. Standardization of the Analytab Enteric (API 20E) system to increase accuracy and reproducibility of the test for biotype characterization of bacteria. J. Clin. Microbiol. 8:46-49. 7. Ronald, A. R., G. K. M. Harding, R. Mathias, C. K. Wong, and P. Muir. 1975. Prophylaxis of recurring urinary tract infection in females: a comparison of nitrofurantoin with trimethoprim-sulfamethosazole. Can.
278
BUCKWOLD ET AL.
Med. Assoc. J. 112:13S-16S. 8. Turck, M., R. G. Petersdorf, and M. R. Fournier. 1962. The epidemiology of non-enteric Escherichia coli infections: prevalence of serologic groups. J. Clin. Invest. 41: 1760-1765. 9. Turck, M., A. R. Ronald, and R. G. Petersdorf. 1968. Relapse and reinfection in chronic bacteriuria. II. The correlation between site of infection and pattern of recurrence in chronic bacteriuria. N. Engl. J. Med. 278:
J. CLIN. MICROBIOL. 422-427. 10. van der Waaij, D., T. M. Speltie, P. A. M. Guinee, and C. Agterberg. 1975. Serotyping and biotyping of 160
Escherichia coli strains: comparative study. J. Clin. Microbiol. 1:237-238. 11. Vera, H. D., and M. Dumoff. 1974. Culture media, p. 881-929. In E. H. Lennette, E. H. Spaulding, and J. P. Truant (ed.), Manual of clinical microbiology. American Society for Microbiology, Washington, D.C.
