J. Med. Microbio1.-Vol. 21 (1986). 139-144 (F 1986 The Pathological Society of Great Britain and Ireland Rare occurrence of Proteus vulgaris in faeces: a reason for its rare association with urinary tract infections

B. W. SENIOR and D. L. LESLIE

Department of Medical Microbiology, University of Dundee Medical School, Ninewells Hospital, Dundee DO1 9SY

Summary. The faecal carriage rates of different species of Proteeae were assessed in studies with 220 faecal isolates from 2 19 individuals of whom approximately one-third were well and the remainder had gastro-enteritis. As a result of the development of new media that allowed replacement of the phenylalanine deaminase test with the tryptophan deaminase test and made it possible to combine tests for indole and urease production and for hydrogen sulphide and ornithine decarboxylase formation in two single-tube tests, all strains were speciated with speed, economy and accuracy. Most (96%) isolates were either Proteus mirabilis (62%) or Morganella morgani (34%). The significance of these findings in relation to urinary tract infection is discussed. P. vulgaris was found in only one (0.45%) faecal specimen and this rarity of carriage in faeces is believed to be the main reason for its rare association with urinary tract infections. The frequent association of M. morgani, in the absence of other enteropathogenic bacteria, with severe gastroenteritis was noted with interest.

Introduction (Senior and Larsson, 1983) all indicate the close relationship between P. mirabilis and P. vulgaris. It Although infections of the urinary tract with strains of Proteus occur less frequently than those is remarkable, therefore, that there should be such a with Escherichia coli, they are particularly impor- wide difference in the frequency with which these organisms are associated with urinary tract infec- tant in young boys (Bergstrom, 1972; Hallett et al., 1976) and the elderly (Walkey et al., 1967; Senior, tions. P. 1979). Furthermore, the consequences of infection One explanation for the rare association of zwlgaris with Proteus are often more serious than those with with infections of the urinary tract might be E. coli, which is usually confined to the bladder, for the fact that analogous bacteriocin production/ (P/S) P. Proteus has a special predilection for the upper sensitivity types to those associated with mirabilis urinary tract (Fairley et al., 197 1; Svanborg Eden et strains having special affinity for the urinary tract (Senior, 1979) have not yet been found ul., 1980) where it may cause damage and death to P. zwlgaris the renal tubular epithelium (Braude et ul., 1960) in (unpublished results). Alternatively, there may be a much simpler explanation based on and the formation of renal stones (Griffith et al., the frequency of occurrence of P. vulgaris in the 1973). These events arise through the formation by faeces. The aim of this study was to investigate the Proteus species of potent urease isoenzymes (Senior latter theory through the development of new multi- et al., 1980) which cause the degradation of urea and combination biochemical tests which would permit rapidly make the urine alkaline (Senior, 1983). the identification of all species of members of the The common source of urinary tract pathogens is tribe Proteeae to be made with speed, ease, economy the bowel. Most infections of the urinary tract with and accuracy. Proteus are due to P. mirabilis (96.50/,)and very few to P. vulgaris (1.5%) or Morganella morgani (P. morgani) (I -5%) (Senior, 1979). DNA hybridisation Materials and methods studies (Brenner et al., 1978), urease isoenzyme profiles (Senior et al., 1980) and the sharing of Source and isolation of bacterial strains phage (Coetzee, 1963) and bacteriocin receptors Strains belonging to the tribe Proteeae were isolated from specimens of human faeces submitted to the labora-

. __ ~~~~~~~ . tory for routine investigations. A loopful of each speci- Received 29 Apr. 1985; revised version accepted 17 Jun. 1985. men was plated out for single colonies on Desoxycholate 139 140 B. W. SENIOR AND D. L. LESLIE

Citrate Agar (DCA; Oxoid CM35) and also inoculated by the formation of a stable brown pigment around into Tetrathionate Broth (Oxoid CM29). Both media strains growing on TDA plates. Ornithine decarboxylase/ were incubated overnight at 37°C. The tetrathionate- sulphide medium and triple sugar medium were seeded broth culture was plated out for single colonies on a DCA from broth cultures of all strains that formed tryptophan plate which was incubated overnight at 37°C. One non- deaminase. Those strains which failed to produce urease lactose-fermenting (NLF) colony of every morphological as indicated by an inability to turn the pH indicator in the type present on the DCA plates was picked and inocu- urea/indolc medium to pink were, in addition, inoculated lated into Nutrient Broth (Oxoid CM67). After overnight from their broth cultures into glucose, galactose, and incubation at 37°C a loopful of each broth culture was trehalose peptone-water sugars. All inoculated media placed on the surface of a dried tryptophan deaminase were incubated overnight at 37°C. Ehrlich’s reagent was agar plate and a loopful seeded into urea/indole medium. gently added to the cultures in the urea/indole medium. After overnight incubation at 37”C, those cultures that Indole formation was indicated by the reagent turning produced tryptophan deaminasc were saved and subse- red. Ornithine decarboxylase formation was indicated by quently speciated as detailed below. the development of alkalinity. Hydrogen sulphide pro- duction was revealed by the formation of a black deposit. If any isolate formed acid on the triple sugar agar Media medium, four tubes of single sugar peptone waters All commercially prepared media were dispensed and (lactose, maltose, inositol and adonitol) were inoculated sterilised according to the manufacturer’s instructions. from its broth culture, incubated overnight at 37 C and Each batch of every medium was tested with appropriate the reactions to the individual sugars recorded. The positivc and negative control test organisms before being species of each isolate was determined by reference to used in the study. Trjyptophun deaminuse agar medium table I. (TDA) contained bacteriological peptone (Oxoid L37) 30 g, Lab-Lemco powder (Oxoid L29) 50 g, L-tryptophan Results (Sigma) 10 g, Agar No. 3 (Oxoid L13) 12 g, bile salts (Oxoid L55) 3 g and yeast extract (Oxoid L21) 3 g During a 6-month period, 220 strains belonging dissolved in distilled water to a volume of 1 L. The to species of the tribe Proteeae were isolated from medium was sterilised by autoclaving at 12 1’ C for 15 min the faeces of 219 people of whom 120 (55%) were and, when cool, poured into petri dishes. Urealindole female, 94 (43%) male, and five (2%) of unstated sex; medium was prepared by supplementing 95 ml of sterile 70% of the females and 60% of the males were in Tryptone Water (Oxoid CM87) with 5 ml of a Seitz- filtered refrigerated stock solution of urea 40 g, phenolph- hospital. Clinical information was given about 202 thalein (1% w/v in propan-2-01) 10 ml made up to 100 ml of the individuals and of these, 148 (73%) had with distilled water. The medium was dispensed into symptoms of gastro-enteritis in the form of diarr- sterile, cotton wool-plugged tubes. Ornithine decarboxy- hoea only (72% of cases) or diarrhoea and vomiting lase/sulphide medium contained Tryptone Water (Oxoid (23% ofcases) or vomiting only (5% of cases). None CM87) 0-5 g, L-ornithine hydrochloride (Sigma) 1 g, was receiving antibiotics when faecal samples were ferrous ammonium sulphate 20 mg, sodium thiosulphate collected. The remaining 71 people (327; of the 20 mg and a little bromocresol purple dye, dissolved in total) were either healthy, or unwell with undis- distilled water to 100 ml. After autoclaving at 121°C for closed illness. Known intestinal pathogenic bacteria 15 min, the medium was allowed to cool and then supplemented with glucose to a final concentration of were found in only 20 (9%) of the specimens of 0.1% w/v from a Seitz-filtered glucose 10% w/v in water faeces and these and their association with species stock solution. The medium was dispensed aseptically of Proteeae are presented in table 11. into bijoux bottles. Tripfe sugar agar medium was pre- Speciation of the 220 strains of Proteeue isolated pared by supplementing sterile molten, cool Lab-Lemco showed that 136 (61.804) were P. mirubilis, 74 Agar (Oxoid CM 17) containing neutral red with lactose, (33.6”/,) were M. morgani, four ( 1.8‘:(,) were Proili- maltose and inositol, each to a concentration of 1% w/v dencia alcalifcrciens, three (1.36%) were Proz?. reit- from Seitz-filtered 107” w/v stock solutions of each sugar geri, two (0.9%) were Prov. stuarti and one (0.45”;) in water. After gentle mixing, the medium was poured was P. zdgaris. There were no isolations of Tofu- Peptone rtwter sugars into petri dishes. were made by rnella ptyseos (Hollis et a/., 198 1) or Pro[(.rustigiurzi supplementing Peptonc Water (Oxoid CM9) with a little et al., 1983) P. penrzeri bromocresol purple dye and the sugar to a final concen- (Hickman-Brenner or (Hick- tration of 1% w/v. The medium was dispensed into tubes man et ul., 1982). or bijoux bottles and sterilised by steaming for 30 min on An analysis of the frequency of isolation of the each of three successive days. different species of Proteeae from the faeces of patients of different age and sex is presented in table 111. P. mirabilis and M. nzorgani were isolated from Speciation of strains patients in every 10-year age group from infancy to The formation of tryptophan deaminase was indicated old age. Generally P. mirahilis was isolated a little RARE OCCURRENCE OF PROTEUS VULGARIS IN FAECES 141

Table I. The biochemical reactions of members of the tribe Proteeae and Taturnella ptyseos

Biochemical test and result*

Acid from

Proteus mirabilis + (+) - + (+) - - - - + NT (+) t Proteus penneri ++ - - V-+--++V Proteus iwlgaris + + + - (+)- - - - vNT(+) Morganella morguni +++(+I---- - V NT (+) Proridencia rettgeri +++---- + + - + (-) Proilidenciu stuarti + v + - - --[+][-I+ + - Proiidenciu alcalqaciens +-+----- +-- V $, Providenciu rustigiuni +-(+)------+v 9 Taturnella ptyseos

* + = Production; - =no production; ( )=most strains; V=different strains give different results; NT= not tested. t Data from Hickman el ul., 1982. $, Data from Hickmann-Brenner et ul., 1983. 9 Data from Hollis et al., 1981.

Table 11. The 20 enteropathogenic isolates* and their association with species of the tribe Proteeae

Number of isolations of pathogen

in association with

Sulmonellu sp. 10 5 4 I 0 Shigellu sp. 3 0 2 0 1 Cumprlohuc t er sp . 3 I 2 00 Clostridium chfjicile 3 2 0 I 0 Non-cholera vibrio I 0 I 0 0

* The 20 enteropathogenic isolates were from 219 faecal specimens examined (see text). more frequently than M. morgani regardless of the symptoms of gastro-enteritis, 44 (62%) were age or sex of the patient. From elderly women, P. infected with P. mirahilis and 27 (38%) with M. mirahilis isolates greatly outnumbered M. morgani morgani. A similar ratio between P. mirabilis and isolates, but the reverse situation was found in M.morgani was found among the 107 patients with males of 1 1-30 and 41-50 years. diarrhoea alone, ofwhom 69%were infected with P. An analysis of the data with regard to clinical mirabilis and 30% with M. morguni, but a different symptoms showed that of the 71 patients with no ratio was found for the 34 patients with both 142 B. W. SENIOR AND D. L. LESLIE

Table 111. Frequency of isolation of species of Proteeue from faeces of patients according to age and sex

Number of strains of the stated species

from male patients from female patients

2 9 Age 2 group Number 01 Number of .E "gg (years) patients 4; 2 1 pa tien ts a; %$$

~ - .. 0-10 44 27 170000 36 23 11 0 I10 11-20 2 0 20000 9 5 3 I 000 21 30 9 I80000 11 5 6 0 000 3140 6 4 20000 3 3 0 0 000 4 1-50 5 140000 9 5 2 0 020 5 1-60 4 2 10010 3 2 1 0 000 61-70 2 2 00000 91 7 2 0 001 7 1-80 7 4 20001 15 13 I 0 I00 81 + 6 5 10000 19 .I 5 4 0 000 Unknown 9 440100 6 5 1 0 000

~~ Total 94 5041 0 I1 1 I20 83 31 1 231 Percentage of total I 00 534401 I I 100 70 26 0.8 1.6 2.4 0.8

There were 5 patients of unstated age and sex of whom 3 yielded P. mirubilis and 2 yielded M. morguni. * One patient carried 2 different species. diarrhoea and vomiting among whom 50% were Strains belonging to the tribe Proteeae are easily infected with P. mirabilis and 47% with M.morgani. recognised for, with the exception of strains of Tatumella ptyseos (Hollis et af., 1981), they are unique among the Enterobacteriaceae in possessing Discussion the ability to deaminate amino acids oxidatively . This study sought the development of new media The test for this ability has usually been performed to permit the biochemical speciation of all represen- with phenylalanine, the deaminated product phenyl tatives of the tribe Proteeae with speed, economy of pyruvic acid forming a transient green colour upon media and the minimum of manipulations in order addition of ferric chloride (Henriksen and Closs, that the degree of correlation between frequency of 1938; Henriksen, 1950). Polster and Svobodova faecal carriage of the different species and their (1964) advocated the use of tryptophan instead of frequency in urinary tract infection might be phenylalanine and reported that only strains of the assessed. tribe Proteeae formed a red-brown diffusible pig- The new media that we developed proved reliable ment (presumably indole-3-pyruvic acid) when and are believed to be accurate; with very few growing on media containing tryptophan. Using exceptions, each strain gave results that were typical their medium, we found, as they did, that a small and fully in agreement with only one of the different number of M. morgani strains failed to produce species of the tribe Proteeae. The exceptions pigment. The modified medium we devised sup- included a small number of H2S-negative variants ported a better growth of M.morgani and all strains of P. mirabilis and P. zwfgaris, well within the of Proteeae produced pigment. All non-pigment- frequency range observed by others (Brenner et af, forming strains were also phenylalanine deaminase 1978), and one strain of the ornithine decarboxy- negative (data not shown). The inclusion of bile lase-negative biotype of M. morgani which others salts appeared to have no effect on pigment produc- have described (Hickman et al., 1980). tion but, by preventing swarming, permitted about RARE OCCURRENCE OF PROTEUS VULGARIS IN FAECES 143

12 strains to be tested together on one plate. The morgani are more frequently encountered than P. brown pigment, in contrast to the green pigment vulgaris and Providencia species and with this in from phenylalanine, did not fade nor did it require mind our triple sugar medium fulfilled two useful the addition of ferric chloride to show its presence. functions. If strains did not show acid formation on Because of these factors and the ease, reliability and it they were pure isolates of either P. mirabilis or M. reproducibility of the test, we recommend the use of morgani and free from lactose-fermenting contami- our modified tryptophan deaminase medium in nants. Whenever a strain formed acid on the preference to the phenylalanine test. medium this indicated that the isolate was either a The urea/indole medium proved helpful in that it lactose-fermenting contaminant or a strain of the permitted rapid recognition, with high probability, more rarely encountered P. vulgaris, P.penneri or a of the commonest species of Proteeae, P. mirabilis species of Prouidencia, and this was readily resolved (urease positive, indole negative), in a single tube to species level by subculture in the four single-sugar test. Phenolphthalein was selected as pH indicator tubes. instead of the more commonly used indicator The most commonly isolated species of Proteeae phenol red because it changed from colourless to from faeces was P. mirabilis and this would explain pink at a much higher pH than the latter, thereby its frequent association with urinary tract infections avoiding possible confusion between weak urease (Senior, 1979). M. morgani was also common and producers, e.g. strains of Enterobacter, Serratia and found in about one third of faecal specimens Pseudomonas species, and the strong urease pro- irrespective of the age and sex of the patient. Its rare ducers of the Proteeae. The tryptone water was association with urinary tract infections (Senior, sufficiently buffered and rich to permit good growth 1979) has been attributed to its relatively slow and tryptophan breakdown before high alkalinity growth rate in urine and its inability to make urine prevented this. There was never any difficulty in rapidly alkaline despite its potent urease enzyme reading both urease positive and indole positive (Senior, 1983). The results showed unequivocally reactions because one was pink and the other red that P. vulgaris, in spite of its name (vulgaris = com- and the acid of the Ehrlich’s reagent caused a yellow mon), was rarely found in human faeces and this is band to appear between each colour. therefore the most probable reason for its rare Detection of hydrogen sulphide formation is association with urinary tract infections (Senior, often not very reliable but the ornithine decarboxy- 1979). lase/sulphide medium we devised was always satis- It was noted with interest that, in the absence of factory and most useful in that the reactions in this other known enteropathogenic bacteria or anti- one medium when considered with the indole/urea biotic therapy, M. morgani was associated with result could distinguish four different species-P. diarrhoea and vomiting at a higher frequency than mirabilis, P. vulgaris, M.morgani and Prov. rettgeri. expected. We hope to investigate whether such In our experience, strains of P. mirabilis and M. strains can produce enterotoxins.

REFERENCES reaction and the urease test in the differentiation of Protrus from other enteric organisms. Journal sf Bacteriology Bergstrijm T 1972 Sex differences in childhood urinary tract 60:225-23 1. infection. Archities of Diseases of Childhood 47:227-232. Henriksen S D, Closs K 1938 The production of phenylpyruvic Braude A I, Siemienski J, Shapiro A B 1960 The role ofbacterial acid by bacteria. Acta Pathologica et Microbiologicu Scandi- urease in the pathogenesis of pyelonephritis. In: Quinn E L, navica 15101-1 13. Kass E H (eds) Biology of pyelonephritis, Churchill Liv- Hickman-Brenner F W, Farmer J J, Steigerwalt A G, Brenner D ingstone, London, pp 69-88. J 1983 Protiidencia rustigianii: a new species in the family Brenner D J et al. 1978 Deoxyribonucleic acid relatedness of Enterobacteriaceae formerly known as Proaidencia alcalifa- Proteus and Providencia species. International Journal of ciens biogroup 3. Journal of Clinical Microbiology 17: 1057- Systematic Bacteriology 28:269-282. 1060. Coetzee J N 1963 Lysogeny in Proteus rettgeriand the host-range Hickman F W, Farmer J J, Steigerwalt A G, Brenner D J 1980 of P. rettgeri and P. hauseri bacteriophages. Journal of Unusual groups of Morganella ( “Proteus ”) morganii iso- General Microbiology 31:2 19-229. lated from clinical specimens, lysine-positive and ornithine- Fairley K F ef al. 1971 Site of infection in acute urinary tract negative biogroups. Journal of Clinical Microbiology 12:88- infection in general practice. Lancet 2:615-618. 94. Griffith D P, Musher D M, Campbell J W 1973 Inhibition of Hickman F W, Steigerwalt A G, Farmer J J, Brenner D J 1982 bacterial urease. Investigative Urology 11:23&238. Identification of Proteuspenneri sp. nov. formerly known as Hallett R J, Pead L, Maskell R 1976 Urinary infection in boys. A Proteus vulgaris indole negative or as Proteus vulgaris three year prospective study. Lancer 2:1107-I 110. biogroup 1. Journal of Clinical Microbiology 151097-1 102. Henriksen SD 1950 A comparison of the phenylpyruvic acid Hollis D G, Hickman F W, Fanning G R, Farmer J J, Weaver R 144 B. W. SENIOR AND D. L. LESLIE

E, Brenner D J 198 I Tutumellu ptyseos gen. nov., sp. nov., a Protrus strains in relation to virulence for the urinary tract. member of the family Enferohacteriaceae found in clinical Journal qf Medical Microbiology 13507-5 12. specimens. Journal of Clinical Microbiology 1479 88. Senior B W, Larsson P 1983 A highly discriminatory multi- Polster M, Svobodova M 1964 Production of reddish-brown typing scheme for Proreus niirubilis and Proteus i~trlguris. pigment from DL-tryptophan by enterobacteria of the Juurnul of Medicul Microbiologv 16: 193-202. Pro t eus- Pror idenciu group . Esperien t ia 20:6 3 7 -6 3 8. Svanborg Edtn C, Larsson P, Lomberg H 1980 Attachment of Senior B W 1979 The special affinity of particular types of Proteus niirahilis to human urinary sediment epithelial cells Proteus mirahilis for the urinary tract. Journal i$ Medical in ilitro is different from that of Esclterichici cdi. In/i~ctiov~ Microbiology 121-8. and Inimunirj' 27:804-807. Senior B W 1983 Proteus morgani is less frequently associated Walkey F A, Judge T G, Thompson J, Sarkari N B 1967 with urinary tract infections than Proteus mirabilis-an Incidence of urinary infection in the elderly. Scoftlsh explanation. Journal of Medical Microbiology 16317-322. Medirul Journal 12:4 1 14 14. Senior B W, Bradford N C, Sinipson D S 1980 The ureases of