Journal of Applied Bacteriology 1992,73,317-323
Species of Pseudomonas obtained at 7OC and 3OoC during aerobic storage of lamb carcasses M. Prieto, M.R. Garcia-Armesto, M.L. Garcia-Lopez, C. Alonso and A. Otero Department of Food Hygiene and Food Technology, University of Ledn, Ledn, Spain
4133102192:accepted 27 March 1992 M . P R I E T O , M . R . G A R C ~ A - A R M E S T OM.L. , G A R C ~ A - L O P E Z ,c. ALONSO AND A . O T E R O . 1992.A
total of 268 strains of Pseudornonas isolated during storage life of lamb carcasses was identified to species level. One-hundred and thirteen strains obtained at 30°C were Ps. frugi (51), Ps. lundensis (17), Ps. Juorescens biovars I (lo), I11 (9) and VI (l), Ps. putidu biovar A (8 strains) and unidentified (17 strains). Species and biovars isolated a t 7°C (155) were Ps. frugi (101), Ps. lundensis (32), Ps.Juorescens biovar I (6), Ps. putidu biovar A (8) and unidentified (8). Numerical analysis (82% SsM , UPGMA) of ‘psychrotrophic’ and ‘mesophilic’ strains resulted in the formation of nine and eight clusters respectively. The dendrograms obtained exhibited similar structures. Most of the strains of Ps. lundensis and Ps.frugi clustered together. Strains of this latter species also joined the type strain of Ps. testosteroni and appeared included with phenons containing the Ps. putidu strains. There were clusters made u p exclusively of strains assigned to one biovar or group (Ps.Juorescens biovars I and I1 and unidentified). A high level of similarity was observed between clusters of Ps. Juorescens biovar I and those containing the Ps. frugi-Ps. lundensis complex ( > 74% SSM) and Ps. lundensis ( > 80%). T h e recovery of pseudomonads seemed to be affected by the sampling day.
INTRODUCTION
Spoilage of carcasses is a surface phenomenon caused predominantly by organisms capable of growth a t refrigeration temperatures. Although the composition of the flora will depend on the time and conditions of storage, Pseudomonas usually predominate on chilled meats stored in air. The taxonomy of pseudomonads associated with meat spoilage has received a great deal of attention over the last years. Initially, Davidson et al. (1973) found that ‘the majority of commonly occurring pseudomonads on meat did not conform with the species proposed by Stanier et al. (1966)’. The application of numerical taxonomy to strains isolated from spoiled meat (Shaw & Latty 1982; Molin & Ternstrom 1982) showed that a high percentage of them formed clusters which were clearly distinct from the known species. Later on, several reports of the phenotypic relationships among strains obtained from meat were published (Shaw & Latty 1984; Molin & Ternstrom 1986) and a new species from this origin (Ps. hdensis) was proposed (Molin et al. 1986). Correspondenceto :Professor Maria-Luisa Garcia-Ldpez, Departamento de Higiene y Tecnologia de 10s Alimentos, Univrrsidadde Ledn, 2407/-Ledn, Spam.
Information about the spoilage flora of lamb carcasses is relatively sparse compared with that of beef. There is evidence that sheep meat tends to have higher pH values and shorter shelf-life than beef. Several authors (Gill & Penney 1985; Prieto et al. 1991) have reported ultimate pH values between 5-6-54 and 6.4. Although elevated pH does not seem to affect the composition of spoilage micro-organisms (Gill & Newton 1982; Shaw & Latty 1984), comparison of floras on chicken breast and leg muscle (high pH) shows significant differences (Gill 1986). Therefore, it was considered of interest to identify the Gram-negative aerobic motile bacteria associated with lamb carcasses spoilage. Evolution of species during aerobic storage and the effect on recovery of two incubation temperatures were also investigated. MATERIALS AND M E T H O D S
Stralns A total of 268 strains of Gram-negative motile rods were studied. They were isolated from eight lamb carcasses at three sampling areas (leg, brisket and neck) during aerobic storage (O-slaughtering day-5, 10, 15 d, and when
318 M . P R I E T O E T A L .
spoiled) at chill (4 1°C) temperatures and 90 f 5% R.H. The strains were selected on the basis of their Gram reaction and motility from a total of 1200 strains. A Harrison Disk (Harrigan & McCance 1976) was used for random selection from counting plates (Plate Count Agar Oxoid) incubated at 7°C for 10 d (155 colonies) and 30°C for 2 d (1 13 colonies). After purification, stock cultures were maintained at 3°C on Tryptic Soy agar (TSA, Difco) slopes and subcultured every 4 months. Strains were also lyophilized in 20% skim milk (Oxoid) and maintained at 4°C. Inocula for tests were grown on Tryptic Soy broth (TSB, Difco) at 25°C for 24-36 h unless otherwise stated. Control panels with TSA were used to ensure viability and purity of inocula. The following reference strains were included in the numerical analysis : Acinetobacter calcoaceticus ATCC 23055 ;Alteromonas putrefaciens NCTC 10735; Enterobacter aerogenes NCIB 366; Escherichia coli ATCC 25922 ; Flavobacterium lucecoloratum ATCC 17964; (Moraxella) osloensis ATCC 19976; Pseudomonas aureofaciens ATCC 13895; Ps. Juorescens ATCC 12841; Ps. caryophylli NCPPB 349; Ps. fragi ATCC 4973; Ps. putida ATCC 12633; Ps. testosteroni ATCC 11996; Ps. cepacia ATCC 17759 and Ps. maltophilia ATCC 17450. Type strains used as biological control of tests were those recommended by Lanyi (1987). Tests on strains
Type and field strains were subjected to the following tests: (1) Colony characteristics. Colonies were examined after incubation on Nutrient Agar (NA; Oxoid) (48 h at 25”C), with a stereoscopic microscope (Wild M3) fitted with a calibrated eye-piece micrometer. The following characteristics were examined: size, edge and superficial aspect (smoothness/roughness), presence of odours, opacity and elevation. (2) Morphology. Cell shape, cell arrangement and Gram reaction were recorded according to the recommendations of Hendrie & Shewan (1979). Jensen’s Gram modification (Cruickshank 1965) was used with young cultures ( < 24 h). (3) Type of growth in broth. Cultures grown for 48 h (25°C) in 5 ml of Nutrient broth (NB) were examined for surface growth, turbidity (scanty, moderate and profuse) and sediment properties (presence and dispersal). (4) Pigment production. Pyoverdin and phenazine production was investigated on King’s A and B agar (King et al 1954). A U.V. lamp (365 nm, Hendrie & Shewan 1979; 254 nm, Palleroni 1984) was used to visualize fluorescein.
Production of other pigments was observed on NA after incubation at 25°C for 2 and 7 d. (5) Growth at dtflerent temperatures. Ability of strains to grow at 4°C (10 and 14 d), 37°C (36 h) and 41°C (48 h) was observed in NB. (6) Growth at p H 4-5. Growth in NB at pH 4.5 was recorded after incubation at 25°C for 2 and 4 d. (7) Salt tolerance. Growth was observed on NA containing 3% and 6% NaCl after incubation at 25°C for 4 d.
(8) Oxidase test. The method of Kovacs (1956) was used according to the recommendations of Hendrie 8i Shewan ( 1979). (9) Catalase production. This was tested on NA after incubation at 25°C for 48 h (Cowan 1974). (10) Motility. Motility was tested by both the hanging drop technique (Harrigan & McCance 1976) and motility medium (Cowan 1974). (1 1) Oxidative and fermentative (0-0 utilization of glucose. The miniaturized (microtitre plate) procedure described in the ‘Handbook of Manual Microtiter Procedures’ (Anon. 1978) was used with a multipoint inoculator Ridgway Watt AM80 (Dynatech, Billingshurst, England) with 96 pins. The basal medium of Hugh & Leifson (1953), with 0.3% agar, 1% glucose and 0.0025% bromcresol purple indicator (180 p1 into each well) was used. After inoculation two drops of sterile melted vaseline were added for fermentation tests but omitted for oxidation. The plates were incubated at 25°C for 14 d. (12) Penicillin sensitivity. Susceptibility to penicillin G (Sigma) was determined as described by Shaw & Latty (1988). (13) Acid production from carbohydrates. The following filter sterilized carbohydrates (1 %, final concentration) were added to the Hugh-Leifson medium (Hugh & Leifson 1953): adonitol, L-arabinose, cellobiose, D-galactose, glycerol, Pfructose, rneso-inositol, lactose, maltose, mannitol, D-mannose, melibiose, raffinose, sucrose, salicin, sorbitol, trehalose, xylitol and D-xylose. A procedure similar to that of the 0 - F test was used. Microtitre plates were incubated at 25°C for 4 d and observed daily. (14) Hydrolysis of DNA, gelatin, casein and starch. The methods of Hendrie & Shewan (1979) were used. All tests were performed on 15 cm diameter Petri dishes with the multipoint inoculator Ridgway Watt AM80 and the 48 pins-header.
PSEUDOMONAS F R O M L A M B CARCASSES 319
(15) Amino acid decarboxylase rests. Miieller medium (Difco) was used to determine L-ornithine and L-lysine decarboxylases (Hendrie & Shewan 1979). Microtitre plates were used as described by Lee & Fung (1982). Control plates without amino acids were included. (16) Arginine dihydrolase. The method of Lee & Fung (1982) was used according to the recommendations of Hendrie & Shewan (1979). (17) Levan from sucrose. The method of Lelliot et al. (1966) was used. (18) Growth on MacConkey agar. Strains were incubated at 25°C for 4 d. (19) Egg-yolk reaction. Cultures were grown at 25°C for 5 d on NA supplemented with 5% (w/v) of egg-yolk emulsion (Oxoid). (20) Carbon source utilization. The organisms were tested for the ability to grow on 20 compounds as the sole source of carbon (L-arabinose, n-butanol, L-arginine, cellobiose, cysteine, ethanol, phenol, D-galaCtOSe, glycine, &glucose, guaiacol, meso-inositol, lactose, L-lysine, maltose, rhamnose, sorbitol, L-sorbose, tryptophan and wxylose). T h e basal medium was that of Palleroni & Doudoroff (1972) and the substrate concentration was 0.1% (w/v). T h e test compounds were either filter sterilized or autoclaved (121°C, 15 min) depending on their thermostability. T h e media were dispensed into microtitre plates (200 pI per well) and inoculated with a suspension of the organisms (phosphate-buffered saline solution 0.01 mol/l containing 0.85% NaCI) with the multipoint inoculator and the 96 pin-header. Plates were incubated at 30°C for 7 d. Precautions were taken to avoid contamination of the pins (Cox & Macrae 1989). Microtitre plates without carbon sources were used as controls. Readings were done with a Test Reading Mirror (Dynatech). Taxonomic schemes
For genera identification the schemes of Hendrie & Shewan (1979), Dainty (Brown & Baird-Parker 1982) and Palleroni (1984) were used. Characteristics employed for the identification of species were those of Molin & Ternstrom (1982, 1986) and Shaw & Latty (1982, 1984). Phenotypic properties useful for the identification of Ps. lundensis (Molin et al. 1986) were also included. Numerical anaiysls
Four properties (Gram reaction, motility, catalase activity and anaerobic fermentation of glucose) were identical for all
strains (including type strains) and were therefore excluded from numerical analysis. T h e rest of the characters were coded as negative (0), positive (1) or doubtful (2). The simple matching coefficient (SSM, Sokal & Michener 1958) and the Jaccard coefficient (SJ, Sneath 1957) were used and clustering achieved by unweighted pair groups average linkage (UPGMA, Sneath & Sokal 1973). T h e analysis was performed in a Compaq Deskpro 386s personal computer (Compaq Computer Corp., T X , USA). The software used was SPSS P C + (SPSS Inc., CH,USA) for clustering and BASIC computer programs (Prieto 1990) for similarity coefficients. All tests were repeated on 27 strains. T h e average probability of error ( P ) was calculated according to Sneath & Johnson (1972). Feature frequency tables were generated by computer and used to determine the most discriminatory characters (separation index) (Sneath 1978). RESULTS Classical taxonomy
All 268 strains were Pseudomonas spp. Isolates obtained at 30°C were identified as Ps. fragi (51 strains), Ps. lundensis (17), Ps. Juorescens biovars I (lo), 111 (9) and VI (1 strain), Ps. putida biovar A (8) and unidentified (17). Similar species and biovars were found amongst strains isolated at 7°C: Ps. fragi (101 strains), Ps. lundensis (32), Ps.JPuorescens biovar I (6),Ps. purida biovar A (8) and unidentified (8). Numerical taxonomy of strains Isolated at 30°C
The average probability of error ( P ) was 0.02 which would be unlikely to result in any serious distortion of taxonomic structure (Sneath & Johnson 1972). T h e cophenetic correlation values were 0.923 (SsM) and 0.851 (SJ). Numerical analysis of the strains (Fig. 1) resulted in the formation of nine groups at the 82% similarity level (SsM) . There were 16 unclustered strains, of which 10 were reference and six isolated strains. Clusters I and VII contained all strains identified as Ps. JPuorescens biovars I and 111, respectively. Strains of Ps. fragi, grouped together with Ps. lundensis (clusters 111 and V), with Ps. putida biovar A (cluster 11) and with the reference strain of Ps. testosteroni (cluster IX). Cluster VI consisted of 12 unidentified strains. Numerical taxonomy of strains Isolated at 7°C
The probability of error for the taxonomic tests was 0.03. This level of reproducibility is within the acceptable levels proposed by Sneath 8i Johnson (1972). T h e correlation cophenetic values were 0.945 (&) and 0.860 (SJ). At the 82% similarity level (SsM)eight clusters were formed (Fig. 2). Cluster IV was entirely composed of
320 M . P R I E T O E r A L .
82.0 Group
No*
100 1
93.0 I
86.0 1
I
79.0
72.0
65.0
I
I
1
I
II
m
ITI P
d
TLI I
I
I
PII
nm
7I t
I -r
I
I
I
I
I
7
7 I Ix
Group
J
100
I
I
93.0
86.0
I
I
I
I
L
79.0
72.0
65.0
strains identified as Ps. Jluorescens. Clusters I1 (10 strains), V (2) and VII (101) consisted of cultures classified as Ps. lundensis and Ps. fragi. Strains of this latter species and the reference strain of Ps. testosteroni formed cluster I (14 strains). Cluster I11 comprised eight strains of Ps. putida and one of Ps. fragi. Strains included in cluster VI (2) were Ps. putida. Finally, five unidentified isolates were included in cluster VIII. There were 17 unclustered strains (10 reference and seven field strains). Dendrograms generated with the S,coefficient exhibited very similar structures. DISCUSSION
Pseudomonas fragi is frequently reported as the most common pseudomonad on meat (Shaw & Latty 1982, 1984; Molin & Ternstrom 1982, 1986). In this work fewer strains of Ps. fragi were isolated at 30°C (45.1%) than among those at 7°C (65.1). The overall percentage within the whole population of Pseudomonas isolated strains was 56.7% com-
Fig. 1 Simplified dendrogram (SsM and UPGMA) of strains of Gram-negative aerobic motile rods obtained at 30°C from lamb carcasses. No.*, Number of strains per group
pared with a 61-79% incidence found by the above authors. Pseudomonas lundensis is a new species proposed by Molin et al. (1986) to accommodate strains attached to one of the major clusters defined by numerical analysis in the investigations of Shaw & Latty (1982). Molin et al. (1986) reported that it was the principal spoilage organism of 20% of cold stored beef and pork in Sweden. It was not uncommon on lamb carcasses, constituting 18% of the strains examined. Pseudomonas lundensis could thus be considered as the second component of the Pseudomonas flora of lamb. The four biovars of Ps. Jluorescens are also frequently reported on meat (Molin & Ternstrom 1982, 1986). Shaw & Latty (1984), who found Ps. JZuorescens biovars I (4%) and I11 (0-90/) among isolates obtained from beef, pork and lamb, suggest that this species can be regarded as a relatively minor component of the pseudomonad population on meat. Both biovars accounted for 9.6% of the strains examined by us. It is of interest that the incidence of Ps. JZuorescens was higher at 30°C (17.6%) than at 7°C (3.8%).
PSEUDOMONAS F R O M L A M B CARCASSES
321
82.0 Group
I
II
[ [
No*
I00
92.2
84.4
I
76.6
68.8
61.0
14
10 ' I
I
I
I
h Fig. 2 Simplified dendrogram (SsM and UPGMA) of strains of Gram-negative aerobic motile rods obtained at 7°C from lamb carcasses. No.*, Number of strains per group
Group
No*
In addition, strains of Ps. Jluorescens biovars 111 were not detected at the latter temperature. Perhaps Ps. Juorescens might compete more favourably at 30°C on crowded plates. Although strains of Ps. putida have been implicated in the spoilage (sulphide odours) of fish (Herbert et al. 1971; Gillespie 1981), they are not common on red meats. Cluster 4 of Shaw & Latty (1982) and subcluster B3 of Molin & Ternstrom (1986) resembled Ps. putida but they could easily be distinguished from this species pn the basis of their ability to use many carbon sources. In any case, the low incidence of strains found by us (5.9%) suggests that Ps. putida biovar A cannot be considered as an important lamb-spoiling species. None of the strains belonging to cluster VI (12 strains, Fig. 1) and VII (5, Fig. 2) could be assigned to recognized species (Palleroni 1984; Molin et al. 1986). All members of the two clusters were very similar and they could be differentiated from strains in other clusters by the lack of pigmentation and their capacity to produce acid from salicin. It is of interest that both clusters, which were aerobic,
I00
92.2
84.4
I
76.6
68.8
61.0
joined the type strain of Enterobacter aerogenes at the 79% (SsM)similarity level. T h e phenotypic and genotypic similarities between Ps. Jluorescens, Ps. fragi and Ps. lundensis have been discussed by Molin & Ternstrom (1986). In our study most of the strains belonging to Ps. fragi and Ps. lundensis grouped together and they joined the clusters of Ps. Jluorescens biovar I above the 75% level (SsM).It is widely accepted that on meat held under aerobic conditions these three Pseudomonas spp. have a marked advantage because of their ability to compete and utilize different compounds (Molin & Ternstrom 1986). A high level of similarity ( >80%SsM) was also observed between clusters containing strains of Ps. Juorescens biovar I and those mainly formed by Ps. putida (Figs 1 and 2). The ability to grow in simple mineral media containing single carbon sources is recommended as a reliable means of differentiating Pseudomonas spp. (Shaw & Latty 1984; Molin & Ternstrom 1986). Although our results confirm this observation (Table l), the highest separation indexes
322
M . PRIETO E T A L .
Table 1 Major features distinguishing the clusters
Test Pyoverdin Acid from: D-fructose salicin sorbitol maltose Growth on sole carbon source : D-XylOSe arginine &galactose rhamnose sorbitol glucose inositol Casein hydrolysis Gelatin hydrolysis
Table 3 Distribution of Pseudomonas species according to the sampling site
Strains isolated at 30°C (9 clusters)
Strains isolated at 7°C (8 clusters)
S.P.
S.P.
Species
Leg
10
Pseudomonas fragilps. lundensis complex fluorescens putida Unidentified Total
55
78
68
4 7 4 70
17 5 7 107
5 4 14 91
20 14 14 15
18 15 15 15 14
9 10 12 10 12
14 12
S.P., Separation index; calculated after Sneath (1978).
corresponded to tests based on aerobic acid production from carbohydrates. T h e analysis of results shown in Table 2 indicates certain relationships between the storage life of carcasses and the evolution of Pseudomonas species. Thus, most of the strains identified as Ps. juorescens and Ps. putida were obtained after storage for 15 d, while those belonging to the Ps. fragi-Ps. lundensis clusters tended to decrease as the spoilTable 2 Evolution of pseudomonads species during aerobic storage of lamb carcasses
Sampling site Brisket
Neck
age progressed. The numbers of unidentified pseudomonads recovered on days 0 and 15 were considerably higher than those obtained after 5 and 10 d. In all cases, the lowest figures were obtained at the spoilage time. This observation could be due to the adverse effect on pseudomonads of desiccation. Shaw & Latty (1984) found that the extent of spoilage did not affect the incidence of detection of Ps. fragi. T h e recovery of Pseudomonas spp. was not clearly affected by the sampling site. Table 3 shows that the unidentified strains and Ps. Jluorescens were mainly located on neck and brisket, respectively. The main difference between both areas was the degree of humidity because the thinner portions of the carcasses (brisket) dried sooner than neck.
ACKNOWLEDGEMENTS
This work was supported by a grant from the Spanish Comisibn Interministerial de Ciencia y Tecnologia (Project No. ALI88-0405). M. Prieto was beneficiary of a fellowship from the Spanish Ministerio de Educacih y Ciencia.
Number of strains Sampling day Species/cluster Unidentified (m) Unidentified (p) Pseudomonasfluorescens (m) fluorescens (p) fragi/Ps. lundensis (m) fragi/Ps. lundensis (p) lundensis (m) lundensis (p) putida (m) putida (p) Total
0
5
5 3
1 1 4
1
19 36
19 20 1
1 2 67
46
m, Isolated at 30°C; p, isolated at 7°C.
REFERENCES
10
15
1 1 1 3 10 37 1 2 2 1 59
9 3 13 3 11 18
5 5 67
Spoiled 1 1 7 19
1 29
A N O N . (1978) Bacteriology. In Handbook of Manual Microtiter Procedures, 2nd edn, ed. Conrath, T.B. & Coupe, N.B. pp. 28-105. London: The Whitefriars Press Ltd. R , (1982) Microbiology B R O W NM , . H . & B A I R D - P A R K EA.C. of carcass meats. In Meat Microbiology ed. Brown, M.H. pp. 423-520. Barking: Applied Science Publishers. (1974) Cowan and Steel's Manualfor the IdentiC O W A NS.T. , fication of Medical Bacteria, 2nd edn. London : Cambridge University Press. C o x , J . M . & M A C R A EI ,. C . (1989) A numerical study of proteolytic and lipolytic psychrotrophs isolated from caprine milk. Journal of Applied Bacteriology 66, 137-152 R ., (1965) Medical Microbiology, 11th edn. CRUICKSHANK Edinburgh : Livingstone.
PSEUDOMONAS F R O M L A M B CARCASSES 323
DAVIDSON,C.M., DOWDELL,M.J. & B O A R D ,R . G . (1973) Properties of Gram-negative aerobes isolated from meats. Journal of Food Science 38, 303-305. G I L L , C . O . (1986) The control of microbial spoilage in fresh meats. In Advances in Meat Research. Meat and Poultry Microbiology ed. Pearson, A.M. & Dutson, T.R. pp. 49-88. Basingstoke : Macmillan Publishers. G I L L , G.C. & N E W T O NK , . G . (1982) Effect of lactic acid concentration on growth on meat of Gram-negative psychrotrophs from a meat works. Applied and Environmental Microbiology 43, 284-288. G I L L , C.O. & P E N N E YN. , (1985) Modification of inpack conditions to extend the storage life of vacuum packaged lamb. Meat Science 14,4340. G I L L E S P I E ,N . C. (1981) A numerical taxonomy of Pseudomonas-like bacteria isolated from fish in southeastern Queensland and their association with spoilage. Journal of Applied Bacteriology 50, 2 W 4 . H A R R I G A NW, . E . & M C C A N C E M , . E . (1976) Laboratory Methods in Food and Dairy Microbiology, 2nd edn. pp. 10 and 47-49. London : Academic Press. H E N D R IM E ,. S . & S H E W A NJ ,. M . (1979) The identification of pseudomonads. In Identijcation Methods for Microbiologists, 2nd edn, ed. Skinner, F.A. & Lovelock, D.W. pp. 1-14. London : Academic Press. H E R B E RR.A., T, H E N D R I EM, . S . & S H E W A NJ ,. M . (1971) Bacteria active in the spoilage of certain seafoods. Journal of Applied Bacteriology 34, 41-50. H U G H ,R . & L E I F S O NE, . (1953) The taxonomic significance of fermentative versus oxidative metabolism of carbohydrates by various Gram-negative bacteria. Journal of Bacteriology 66, 24-26. K I N G , E.O., W A R D ,W . K . & R A N E Y D , . E . (1954) Two simple media for the demonstration of pyocyanin and fluorescein. Journal of Laboratory of Clinical Medicine 44,301-307. K O V A C SN, . (1956) Identification of Pseudomonas pyocyanea by the oxidase reaction. Nature (London) 178,703. L A N IY, B . (1987) Classical and rapid identification methods for medically important bacteria. In Methods in Microbiology, Vol. 19, ed. Colwell, R.R. & Grigorova, R. pp. 1-67. London: Academic Press. LEE,C.-Y. & F U N G ,D.Y.C. (1982) Miniaturized methods and computer analysis in meat microbiology. In Rapid Methods and Automation in Microbiology ed. Tilton, R.C. pp. 191-193. Washington DC : American Society for Microbiology. L E L L I O TR.A., , B I L L I N GE, . & H A Y W A R D A.C. , (1966) A determinative scheme for the fluorescent plant pathogenic pseudomonads. Journal of Applied Bacteriology 29, 470-489. M O L I N ,G . & T E R N S T R O M A ., (1982) Numerical taxonomy of psychrotrophic pseudomonads. Journal of General Microbiology 128, 1249-1264.
A . (1986) Phenotypically based M O L I N , G. & TERNSTROM, taxonomy of psychrotrophic Pseudomonas isolated from spoiled meat, water and soil. International Journal of Systematic Bacteriology 36, 257-274. M O L I N ,G . , TERNSTROM, A. & U R S I N GJ, . (1986) Pseudomonas lundensis, a new bacterial species isolated from meat. International Journal of Systematic Bacteriology 36, 339-342. P A L L E R O N N.J. I, (1984) Genus I. Pseudomonas. In Bergey's Manual of Systematic Bacteriology, Vol. 1, ed. Krieg, N.R. & Holt, J.G. pp. 141-199. Baltimore: The Williams and Wilkins. & DOUDOROFF, M . (1972) Some properP A L L E R O NN.J. I, ties and taxonomic subdivisions of the genus Pseudomonas. Annual Review of Phytopathology 10,73-100. P R I E T OM , . (1990) Asociaciones bacterianas que participan en la alteraci6n de canales de ovino refrigeradas : taxonomia, evoluci6n e influencia de diversos factores. Ph.D. Thesis, L e h , Spain: Servicio de Publicaciones de la Universidad de Leon. P R I E T O ,M . , G A R C I AM , . L . , G A R C I AM , . R . , O T E R O ,A. & MORENO,B. (1991) Distribution and evolution of bacteria on lamb carcasses during aerobic storage. Journal of Food Protection 54,945-949. S H A W ,B.G. & L A T T Y ,J.B. (1982) A numerical taxonomic study of Pseudomonas strains from spoiled meat. Journal of Applied Bacteriology 52, 2 19-228. SHAW,B.G. & L A T T Y ,J.B. (1984) A study of the relative incidence of different Pseudomonas groups on meat using a computer-assisted identification technique employing only carbon source tests. Journal of Applied Bacteriology 57, 59-67. S H A WB, . G . & LATTY,J.B. (1988) A numerical taxonomic study of non-motile non-fermentative Gram-negative bacteria from foods. Journal of Applied Bacteriology 65, 7-2 1. S N E A T H, P .H .A . (1957) The application of computers to taxonomy. Journal of General Microbiology 17, 201-226. SNEATH P.H.A. , (1978) Identification of microorganisms. In Essays in Microbiology ed. Norris, J.R. & Richmond, M.H. pp. 10/1-10/32. Chichester: John Wiley. S N E A T HP, . H . A . & J O H N S O N , R. (1972) The influence on numerical taxonomic similarities of errors in microbiological tests. Journal of General Microbiology 72, 377-392. S N E A T H ,P.H.A. & S O K A L ,R . R . (1973) Numerical Taxonomy. The Principles and Practice of Numerical Classification. pp. 230-234. San Francisco: W.H. Freeman. SOKAL R.R. , 19M I C H E N E RC, . D . (1958) A statistical method for evaluating systematic relationships. Kansas University Bulletin 38, 1409-1438. S T A N I E R R.Y., , P A L L E R O N IN, . J . & D O U D O R O F FM ,. (1966) The aerobic pseudomonads : a taxonomic study. Journal of General Microbiology 43, 159-271.
Species of Pseudomonas obtained at 7OC and 3OoC during aerobic storage of lamb carcasses M. Prieto, M.R. Garcia-Armesto, M.L. Garcia-Lopez, C. Alonso and A. Otero Department of Food Hygiene and Food Technology, University of Ledn, Ledn, Spain
4133102192:accepted 27 March 1992 M . P R I E T O , M . R . G A R C ~ A - A R M E S T OM.L. , G A R C ~ A - L O P E Z ,c. ALONSO AND A . O T E R O . 1992.A
total of 268 strains of Pseudornonas isolated during storage life of lamb carcasses was identified to species level. One-hundred and thirteen strains obtained at 30°C were Ps. frugi (51), Ps. lundensis (17), Ps. Juorescens biovars I (lo), I11 (9) and VI (l), Ps. putidu biovar A (8 strains) and unidentified (17 strains). Species and biovars isolated a t 7°C (155) were Ps. frugi (101), Ps. lundensis (32), Ps.Juorescens biovar I (6), Ps. putidu biovar A (8) and unidentified (8). Numerical analysis (82% SsM , UPGMA) of ‘psychrotrophic’ and ‘mesophilic’ strains resulted in the formation of nine and eight clusters respectively. The dendrograms obtained exhibited similar structures. Most of the strains of Ps. lundensis and Ps.frugi clustered together. Strains of this latter species also joined the type strain of Ps. testosteroni and appeared included with phenons containing the Ps. putidu strains. There were clusters made u p exclusively of strains assigned to one biovar or group (Ps.Juorescens biovars I and I1 and unidentified). A high level of similarity was observed between clusters of Ps. Juorescens biovar I and those containing the Ps. frugi-Ps. lundensis complex ( > 74% SSM) and Ps. lundensis ( > 80%). T h e recovery of pseudomonads seemed to be affected by the sampling day.
INTRODUCTION
Spoilage of carcasses is a surface phenomenon caused predominantly by organisms capable of growth a t refrigeration temperatures. Although the composition of the flora will depend on the time and conditions of storage, Pseudomonas usually predominate on chilled meats stored in air. The taxonomy of pseudomonads associated with meat spoilage has received a great deal of attention over the last years. Initially, Davidson et al. (1973) found that ‘the majority of commonly occurring pseudomonads on meat did not conform with the species proposed by Stanier et al. (1966)’. The application of numerical taxonomy to strains isolated from spoiled meat (Shaw & Latty 1982; Molin & Ternstrom 1982) showed that a high percentage of them formed clusters which were clearly distinct from the known species. Later on, several reports of the phenotypic relationships among strains obtained from meat were published (Shaw & Latty 1984; Molin & Ternstrom 1986) and a new species from this origin (Ps. hdensis) was proposed (Molin et al. 1986). Correspondenceto :Professor Maria-Luisa Garcia-Ldpez, Departamento de Higiene y Tecnologia de 10s Alimentos, Univrrsidadde Ledn, 2407/-Ledn, Spam.
Information about the spoilage flora of lamb carcasses is relatively sparse compared with that of beef. There is evidence that sheep meat tends to have higher pH values and shorter shelf-life than beef. Several authors (Gill & Penney 1985; Prieto et al. 1991) have reported ultimate pH values between 5-6-54 and 6.4. Although elevated pH does not seem to affect the composition of spoilage micro-organisms (Gill & Newton 1982; Shaw & Latty 1984), comparison of floras on chicken breast and leg muscle (high pH) shows significant differences (Gill 1986). Therefore, it was considered of interest to identify the Gram-negative aerobic motile bacteria associated with lamb carcasses spoilage. Evolution of species during aerobic storage and the effect on recovery of two incubation temperatures were also investigated. MATERIALS AND M E T H O D S
Stralns A total of 268 strains of Gram-negative motile rods were studied. They were isolated from eight lamb carcasses at three sampling areas (leg, brisket and neck) during aerobic storage (O-slaughtering day-5, 10, 15 d, and when
318 M . P R I E T O E T A L .
spoiled) at chill (4 1°C) temperatures and 90 f 5% R.H. The strains were selected on the basis of their Gram reaction and motility from a total of 1200 strains. A Harrison Disk (Harrigan & McCance 1976) was used for random selection from counting plates (Plate Count Agar Oxoid) incubated at 7°C for 10 d (155 colonies) and 30°C for 2 d (1 13 colonies). After purification, stock cultures were maintained at 3°C on Tryptic Soy agar (TSA, Difco) slopes and subcultured every 4 months. Strains were also lyophilized in 20% skim milk (Oxoid) and maintained at 4°C. Inocula for tests were grown on Tryptic Soy broth (TSB, Difco) at 25°C for 24-36 h unless otherwise stated. Control panels with TSA were used to ensure viability and purity of inocula. The following reference strains were included in the numerical analysis : Acinetobacter calcoaceticus ATCC 23055 ;Alteromonas putrefaciens NCTC 10735; Enterobacter aerogenes NCIB 366; Escherichia coli ATCC 25922 ; Flavobacterium lucecoloratum ATCC 17964; (Moraxella) osloensis ATCC 19976; Pseudomonas aureofaciens ATCC 13895; Ps. Juorescens ATCC 12841; Ps. caryophylli NCPPB 349; Ps. fragi ATCC 4973; Ps. putida ATCC 12633; Ps. testosteroni ATCC 11996; Ps. cepacia ATCC 17759 and Ps. maltophilia ATCC 17450. Type strains used as biological control of tests were those recommended by Lanyi (1987). Tests on strains
Type and field strains were subjected to the following tests: (1) Colony characteristics. Colonies were examined after incubation on Nutrient Agar (NA; Oxoid) (48 h at 25”C), with a stereoscopic microscope (Wild M3) fitted with a calibrated eye-piece micrometer. The following characteristics were examined: size, edge and superficial aspect (smoothness/roughness), presence of odours, opacity and elevation. (2) Morphology. Cell shape, cell arrangement and Gram reaction were recorded according to the recommendations of Hendrie & Shewan (1979). Jensen’s Gram modification (Cruickshank 1965) was used with young cultures ( < 24 h). (3) Type of growth in broth. Cultures grown for 48 h (25°C) in 5 ml of Nutrient broth (NB) were examined for surface growth, turbidity (scanty, moderate and profuse) and sediment properties (presence and dispersal). (4) Pigment production. Pyoverdin and phenazine production was investigated on King’s A and B agar (King et al 1954). A U.V. lamp (365 nm, Hendrie & Shewan 1979; 254 nm, Palleroni 1984) was used to visualize fluorescein.
Production of other pigments was observed on NA after incubation at 25°C for 2 and 7 d. (5) Growth at dtflerent temperatures. Ability of strains to grow at 4°C (10 and 14 d), 37°C (36 h) and 41°C (48 h) was observed in NB. (6) Growth at p H 4-5. Growth in NB at pH 4.5 was recorded after incubation at 25°C for 2 and 4 d. (7) Salt tolerance. Growth was observed on NA containing 3% and 6% NaCl after incubation at 25°C for 4 d.
(8) Oxidase test. The method of Kovacs (1956) was used according to the recommendations of Hendrie 8i Shewan ( 1979). (9) Catalase production. This was tested on NA after incubation at 25°C for 48 h (Cowan 1974). (10) Motility. Motility was tested by both the hanging drop technique (Harrigan & McCance 1976) and motility medium (Cowan 1974). (1 1) Oxidative and fermentative (0-0 utilization of glucose. The miniaturized (microtitre plate) procedure described in the ‘Handbook of Manual Microtiter Procedures’ (Anon. 1978) was used with a multipoint inoculator Ridgway Watt AM80 (Dynatech, Billingshurst, England) with 96 pins. The basal medium of Hugh & Leifson (1953), with 0.3% agar, 1% glucose and 0.0025% bromcresol purple indicator (180 p1 into each well) was used. After inoculation two drops of sterile melted vaseline were added for fermentation tests but omitted for oxidation. The plates were incubated at 25°C for 14 d. (12) Penicillin sensitivity. Susceptibility to penicillin G (Sigma) was determined as described by Shaw & Latty (1988). (13) Acid production from carbohydrates. The following filter sterilized carbohydrates (1 %, final concentration) were added to the Hugh-Leifson medium (Hugh & Leifson 1953): adonitol, L-arabinose, cellobiose, D-galactose, glycerol, Pfructose, rneso-inositol, lactose, maltose, mannitol, D-mannose, melibiose, raffinose, sucrose, salicin, sorbitol, trehalose, xylitol and D-xylose. A procedure similar to that of the 0 - F test was used. Microtitre plates were incubated at 25°C for 4 d and observed daily. (14) Hydrolysis of DNA, gelatin, casein and starch. The methods of Hendrie & Shewan (1979) were used. All tests were performed on 15 cm diameter Petri dishes with the multipoint inoculator Ridgway Watt AM80 and the 48 pins-header.
PSEUDOMONAS F R O M L A M B CARCASSES 319
(15) Amino acid decarboxylase rests. Miieller medium (Difco) was used to determine L-ornithine and L-lysine decarboxylases (Hendrie & Shewan 1979). Microtitre plates were used as described by Lee & Fung (1982). Control plates without amino acids were included. (16) Arginine dihydrolase. The method of Lee & Fung (1982) was used according to the recommendations of Hendrie & Shewan (1979). (17) Levan from sucrose. The method of Lelliot et al. (1966) was used. (18) Growth on MacConkey agar. Strains were incubated at 25°C for 4 d. (19) Egg-yolk reaction. Cultures were grown at 25°C for 5 d on NA supplemented with 5% (w/v) of egg-yolk emulsion (Oxoid). (20) Carbon source utilization. The organisms were tested for the ability to grow on 20 compounds as the sole source of carbon (L-arabinose, n-butanol, L-arginine, cellobiose, cysteine, ethanol, phenol, D-galaCtOSe, glycine, &glucose, guaiacol, meso-inositol, lactose, L-lysine, maltose, rhamnose, sorbitol, L-sorbose, tryptophan and wxylose). T h e basal medium was that of Palleroni & Doudoroff (1972) and the substrate concentration was 0.1% (w/v). T h e test compounds were either filter sterilized or autoclaved (121°C, 15 min) depending on their thermostability. T h e media were dispensed into microtitre plates (200 pI per well) and inoculated with a suspension of the organisms (phosphate-buffered saline solution 0.01 mol/l containing 0.85% NaCI) with the multipoint inoculator and the 96 pin-header. Plates were incubated at 30°C for 7 d. Precautions were taken to avoid contamination of the pins (Cox & Macrae 1989). Microtitre plates without carbon sources were used as controls. Readings were done with a Test Reading Mirror (Dynatech). Taxonomic schemes
For genera identification the schemes of Hendrie & Shewan (1979), Dainty (Brown & Baird-Parker 1982) and Palleroni (1984) were used. Characteristics employed for the identification of species were those of Molin & Ternstrom (1982, 1986) and Shaw & Latty (1982, 1984). Phenotypic properties useful for the identification of Ps. lundensis (Molin et al. 1986) were also included. Numerical anaiysls
Four properties (Gram reaction, motility, catalase activity and anaerobic fermentation of glucose) were identical for all
strains (including type strains) and were therefore excluded from numerical analysis. T h e rest of the characters were coded as negative (0), positive (1) or doubtful (2). The simple matching coefficient (SSM, Sokal & Michener 1958) and the Jaccard coefficient (SJ, Sneath 1957) were used and clustering achieved by unweighted pair groups average linkage (UPGMA, Sneath & Sokal 1973). T h e analysis was performed in a Compaq Deskpro 386s personal computer (Compaq Computer Corp., T X , USA). The software used was SPSS P C + (SPSS Inc., CH,USA) for clustering and BASIC computer programs (Prieto 1990) for similarity coefficients. All tests were repeated on 27 strains. T h e average probability of error ( P ) was calculated according to Sneath & Johnson (1972). Feature frequency tables were generated by computer and used to determine the most discriminatory characters (separation index) (Sneath 1978). RESULTS Classical taxonomy
All 268 strains were Pseudomonas spp. Isolates obtained at 30°C were identified as Ps. fragi (51 strains), Ps. lundensis (17), Ps. Juorescens biovars I (lo), 111 (9) and VI (1 strain), Ps. putida biovar A (8) and unidentified (17). Similar species and biovars were found amongst strains isolated at 7°C: Ps. fragi (101 strains), Ps. lundensis (32), Ps.JPuorescens biovar I (6),Ps. purida biovar A (8) and unidentified (8). Numerical taxonomy of strains Isolated at 30°C
The average probability of error ( P ) was 0.02 which would be unlikely to result in any serious distortion of taxonomic structure (Sneath & Johnson 1972). T h e cophenetic correlation values were 0.923 (SsM) and 0.851 (SJ). Numerical analysis of the strains (Fig. 1) resulted in the formation of nine groups at the 82% similarity level (SsM) . There were 16 unclustered strains, of which 10 were reference and six isolated strains. Clusters I and VII contained all strains identified as Ps. JPuorescens biovars I and 111, respectively. Strains of Ps. fragi, grouped together with Ps. lundensis (clusters 111 and V), with Ps. putida biovar A (cluster 11) and with the reference strain of Ps. testosteroni (cluster IX). Cluster VI consisted of 12 unidentified strains. Numerical taxonomy of strains Isolated at 7°C
The probability of error for the taxonomic tests was 0.03. This level of reproducibility is within the acceptable levels proposed by Sneath 8i Johnson (1972). T h e correlation cophenetic values were 0.945 (&) and 0.860 (SJ). At the 82% similarity level (SsM)eight clusters were formed (Fig. 2). Cluster IV was entirely composed of
320 M . P R I E T O E r A L .
82.0 Group
No*
100 1
93.0 I
86.0 1
I
79.0
72.0
65.0
I
I
1
I
II
m
ITI P
d
TLI I
I
I
PII
nm
7I t
I -r
I
I
I
I
I
7
7 I Ix
Group
J
100
I
I
93.0
86.0
I
I
I
I
L
79.0
72.0
65.0
strains identified as Ps. Jluorescens. Clusters I1 (10 strains), V (2) and VII (101) consisted of cultures classified as Ps. lundensis and Ps. fragi. Strains of this latter species and the reference strain of Ps. testosteroni formed cluster I (14 strains). Cluster I11 comprised eight strains of Ps. putida and one of Ps. fragi. Strains included in cluster VI (2) were Ps. putida. Finally, five unidentified isolates were included in cluster VIII. There were 17 unclustered strains (10 reference and seven field strains). Dendrograms generated with the S,coefficient exhibited very similar structures. DISCUSSION
Pseudomonas fragi is frequently reported as the most common pseudomonad on meat (Shaw & Latty 1982, 1984; Molin & Ternstrom 1982, 1986). In this work fewer strains of Ps. fragi were isolated at 30°C (45.1%) than among those at 7°C (65.1). The overall percentage within the whole population of Pseudomonas isolated strains was 56.7% com-
Fig. 1 Simplified dendrogram (SsM and UPGMA) of strains of Gram-negative aerobic motile rods obtained at 30°C from lamb carcasses. No.*, Number of strains per group
pared with a 61-79% incidence found by the above authors. Pseudomonas lundensis is a new species proposed by Molin et al. (1986) to accommodate strains attached to one of the major clusters defined by numerical analysis in the investigations of Shaw & Latty (1982). Molin et al. (1986) reported that it was the principal spoilage organism of 20% of cold stored beef and pork in Sweden. It was not uncommon on lamb carcasses, constituting 18% of the strains examined. Pseudomonas lundensis could thus be considered as the second component of the Pseudomonas flora of lamb. The four biovars of Ps. Jluorescens are also frequently reported on meat (Molin & Ternstrom 1982, 1986). Shaw & Latty (1984), who found Ps. JZuorescens biovars I (4%) and I11 (0-90/) among isolates obtained from beef, pork and lamb, suggest that this species can be regarded as a relatively minor component of the pseudomonad population on meat. Both biovars accounted for 9.6% of the strains examined by us. It is of interest that the incidence of Ps. JZuorescens was higher at 30°C (17.6%) than at 7°C (3.8%).
PSEUDOMONAS F R O M L A M B CARCASSES
321
82.0 Group
I
II
[ [
No*
I00
92.2
84.4
I
76.6
68.8
61.0
14
10 ' I
I
I
I
h Fig. 2 Simplified dendrogram (SsM and UPGMA) of strains of Gram-negative aerobic motile rods obtained at 7°C from lamb carcasses. No.*, Number of strains per group
Group
No*
In addition, strains of Ps. Jluorescens biovars 111 were not detected at the latter temperature. Perhaps Ps. Juorescens might compete more favourably at 30°C on crowded plates. Although strains of Ps. putida have been implicated in the spoilage (sulphide odours) of fish (Herbert et al. 1971; Gillespie 1981), they are not common on red meats. Cluster 4 of Shaw & Latty (1982) and subcluster B3 of Molin & Ternstrom (1986) resembled Ps. putida but they could easily be distinguished from this species pn the basis of their ability to use many carbon sources. In any case, the low incidence of strains found by us (5.9%) suggests that Ps. putida biovar A cannot be considered as an important lamb-spoiling species. None of the strains belonging to cluster VI (12 strains, Fig. 1) and VII (5, Fig. 2) could be assigned to recognized species (Palleroni 1984; Molin et al. 1986). All members of the two clusters were very similar and they could be differentiated from strains in other clusters by the lack of pigmentation and their capacity to produce acid from salicin. It is of interest that both clusters, which were aerobic,
I00
92.2
84.4
I
76.6
68.8
61.0
joined the type strain of Enterobacter aerogenes at the 79% (SsM)similarity level. T h e phenotypic and genotypic similarities between Ps. Jluorescens, Ps. fragi and Ps. lundensis have been discussed by Molin & Ternstrom (1986). In our study most of the strains belonging to Ps. fragi and Ps. lundensis grouped together and they joined the clusters of Ps. Jluorescens biovar I above the 75% level (SsM).It is widely accepted that on meat held under aerobic conditions these three Pseudomonas spp. have a marked advantage because of their ability to compete and utilize different compounds (Molin & Ternstrom 1986). A high level of similarity ( >80%SsM) was also observed between clusters containing strains of Ps. Juorescens biovar I and those mainly formed by Ps. putida (Figs 1 and 2). The ability to grow in simple mineral media containing single carbon sources is recommended as a reliable means of differentiating Pseudomonas spp. (Shaw & Latty 1984; Molin & Ternstrom 1986). Although our results confirm this observation (Table l), the highest separation indexes
322
M . PRIETO E T A L .
Table 1 Major features distinguishing the clusters
Test Pyoverdin Acid from: D-fructose salicin sorbitol maltose Growth on sole carbon source : D-XylOSe arginine &galactose rhamnose sorbitol glucose inositol Casein hydrolysis Gelatin hydrolysis
Table 3 Distribution of Pseudomonas species according to the sampling site
Strains isolated at 30°C (9 clusters)
Strains isolated at 7°C (8 clusters)
S.P.
S.P.
Species
Leg
10
Pseudomonas fragilps. lundensis complex fluorescens putida Unidentified Total
55
78
68
4 7 4 70
17 5 7 107
5 4 14 91
20 14 14 15
18 15 15 15 14
9 10 12 10 12
14 12
S.P., Separation index; calculated after Sneath (1978).
corresponded to tests based on aerobic acid production from carbohydrates. T h e analysis of results shown in Table 2 indicates certain relationships between the storage life of carcasses and the evolution of Pseudomonas species. Thus, most of the strains identified as Ps. juorescens and Ps. putida were obtained after storage for 15 d, while those belonging to the Ps. fragi-Ps. lundensis clusters tended to decrease as the spoilTable 2 Evolution of pseudomonads species during aerobic storage of lamb carcasses
Sampling site Brisket
Neck
age progressed. The numbers of unidentified pseudomonads recovered on days 0 and 15 were considerably higher than those obtained after 5 and 10 d. In all cases, the lowest figures were obtained at the spoilage time. This observation could be due to the adverse effect on pseudomonads of desiccation. Shaw & Latty (1984) found that the extent of spoilage did not affect the incidence of detection of Ps. fragi. T h e recovery of Pseudomonas spp. was not clearly affected by the sampling site. Table 3 shows that the unidentified strains and Ps. Jluorescens were mainly located on neck and brisket, respectively. The main difference between both areas was the degree of humidity because the thinner portions of the carcasses (brisket) dried sooner than neck.
ACKNOWLEDGEMENTS
This work was supported by a grant from the Spanish Comisibn Interministerial de Ciencia y Tecnologia (Project No. ALI88-0405). M. Prieto was beneficiary of a fellowship from the Spanish Ministerio de Educacih y Ciencia.
Number of strains Sampling day Species/cluster Unidentified (m) Unidentified (p) Pseudomonasfluorescens (m) fluorescens (p) fragi/Ps. lundensis (m) fragi/Ps. lundensis (p) lundensis (m) lundensis (p) putida (m) putida (p) Total
0
5
5 3
1 1 4
1
19 36
19 20 1
1 2 67
46
m, Isolated at 30°C; p, isolated at 7°C.
REFERENCES
10
15
1 1 1 3 10 37 1 2 2 1 59
9 3 13 3 11 18
5 5 67
Spoiled 1 1 7 19
1 29
A N O N . (1978) Bacteriology. In Handbook of Manual Microtiter Procedures, 2nd edn, ed. Conrath, T.B. & Coupe, N.B. pp. 28-105. London: The Whitefriars Press Ltd. R , (1982) Microbiology B R O W NM , . H . & B A I R D - P A R K EA.C. of carcass meats. In Meat Microbiology ed. Brown, M.H. pp. 423-520. Barking: Applied Science Publishers. (1974) Cowan and Steel's Manualfor the IdentiC O W A NS.T. , fication of Medical Bacteria, 2nd edn. London : Cambridge University Press. C o x , J . M . & M A C R A EI ,. C . (1989) A numerical study of proteolytic and lipolytic psychrotrophs isolated from caprine milk. Journal of Applied Bacteriology 66, 137-152 R ., (1965) Medical Microbiology, 11th edn. CRUICKSHANK Edinburgh : Livingstone.
PSEUDOMONAS F R O M L A M B CARCASSES 323
DAVIDSON,C.M., DOWDELL,M.J. & B O A R D ,R . G . (1973) Properties of Gram-negative aerobes isolated from meats. Journal of Food Science 38, 303-305. G I L L , C . O . (1986) The control of microbial spoilage in fresh meats. In Advances in Meat Research. Meat and Poultry Microbiology ed. Pearson, A.M. & Dutson, T.R. pp. 49-88. Basingstoke : Macmillan Publishers. G I L L , G.C. & N E W T O NK , . G . (1982) Effect of lactic acid concentration on growth on meat of Gram-negative psychrotrophs from a meat works. Applied and Environmental Microbiology 43, 284-288. G I L L , C.O. & P E N N E YN. , (1985) Modification of inpack conditions to extend the storage life of vacuum packaged lamb. Meat Science 14,4340. G I L L E S P I E ,N . C. (1981) A numerical taxonomy of Pseudomonas-like bacteria isolated from fish in southeastern Queensland and their association with spoilage. Journal of Applied Bacteriology 50, 2 W 4 . H A R R I G A NW, . E . & M C C A N C E M , . E . (1976) Laboratory Methods in Food and Dairy Microbiology, 2nd edn. pp. 10 and 47-49. London : Academic Press. H E N D R IM E ,. S . & S H E W A NJ ,. M . (1979) The identification of pseudomonads. In Identijcation Methods for Microbiologists, 2nd edn, ed. Skinner, F.A. & Lovelock, D.W. pp. 1-14. London : Academic Press. H E R B E RR.A., T, H E N D R I EM, . S . & S H E W A NJ ,. M . (1971) Bacteria active in the spoilage of certain seafoods. Journal of Applied Bacteriology 34, 41-50. H U G H ,R . & L E I F S O NE, . (1953) The taxonomic significance of fermentative versus oxidative metabolism of carbohydrates by various Gram-negative bacteria. Journal of Bacteriology 66, 24-26. K I N G , E.O., W A R D ,W . K . & R A N E Y D , . E . (1954) Two simple media for the demonstration of pyocyanin and fluorescein. Journal of Laboratory of Clinical Medicine 44,301-307. K O V A C SN, . (1956) Identification of Pseudomonas pyocyanea by the oxidase reaction. Nature (London) 178,703. L A N IY, B . (1987) Classical and rapid identification methods for medically important bacteria. In Methods in Microbiology, Vol. 19, ed. Colwell, R.R. & Grigorova, R. pp. 1-67. London: Academic Press. LEE,C.-Y. & F U N G ,D.Y.C. (1982) Miniaturized methods and computer analysis in meat microbiology. In Rapid Methods and Automation in Microbiology ed. Tilton, R.C. pp. 191-193. Washington DC : American Society for Microbiology. L E L L I O TR.A., , B I L L I N GE, . & H A Y W A R D A.C. , (1966) A determinative scheme for the fluorescent plant pathogenic pseudomonads. Journal of Applied Bacteriology 29, 470-489. M O L I N ,G . & T E R N S T R O M A ., (1982) Numerical taxonomy of psychrotrophic pseudomonads. Journal of General Microbiology 128, 1249-1264.
A . (1986) Phenotypically based M O L I N , G. & TERNSTROM, taxonomy of psychrotrophic Pseudomonas isolated from spoiled meat, water and soil. International Journal of Systematic Bacteriology 36, 257-274. M O L I N ,G . , TERNSTROM, A. & U R S I N GJ, . (1986) Pseudomonas lundensis, a new bacterial species isolated from meat. International Journal of Systematic Bacteriology 36, 339-342. P A L L E R O N N.J. I, (1984) Genus I. Pseudomonas. In Bergey's Manual of Systematic Bacteriology, Vol. 1, ed. Krieg, N.R. & Holt, J.G. pp. 141-199. Baltimore: The Williams and Wilkins. & DOUDOROFF, M . (1972) Some properP A L L E R O NN.J. I, ties and taxonomic subdivisions of the genus Pseudomonas. Annual Review of Phytopathology 10,73-100. P R I E T OM , . (1990) Asociaciones bacterianas que participan en la alteraci6n de canales de ovino refrigeradas : taxonomia, evoluci6n e influencia de diversos factores. Ph.D. Thesis, L e h , Spain: Servicio de Publicaciones de la Universidad de Leon. P R I E T O ,M . , G A R C I AM , . L . , G A R C I AM , . R . , O T E R O ,A. & MORENO,B. (1991) Distribution and evolution of bacteria on lamb carcasses during aerobic storage. Journal of Food Protection 54,945-949. S H A W ,B.G. & L A T T Y ,J.B. (1982) A numerical taxonomic study of Pseudomonas strains from spoiled meat. Journal of Applied Bacteriology 52, 2 19-228. SHAW,B.G. & L A T T Y ,J.B. (1984) A study of the relative incidence of different Pseudomonas groups on meat using a computer-assisted identification technique employing only carbon source tests. Journal of Applied Bacteriology 57, 59-67. S H A WB, . G . & LATTY,J.B. (1988) A numerical taxonomic study of non-motile non-fermentative Gram-negative bacteria from foods. Journal of Applied Bacteriology 65, 7-2 1. S N E A T H, P .H .A . (1957) The application of computers to taxonomy. Journal of General Microbiology 17, 201-226. SNEATH P.H.A. , (1978) Identification of microorganisms. In Essays in Microbiology ed. Norris, J.R. & Richmond, M.H. pp. 10/1-10/32. Chichester: John Wiley. S N E A T HP, . H . A . & J O H N S O N , R. (1972) The influence on numerical taxonomic similarities of errors in microbiological tests. Journal of General Microbiology 72, 377-392. S N E A T H ,P.H.A. & S O K A L ,R . R . (1973) Numerical Taxonomy. The Principles and Practice of Numerical Classification. pp. 230-234. San Francisco: W.H. Freeman. SOKAL R.R. , 19M I C H E N E RC, . D . (1958) A statistical method for evaluating systematic relationships. Kansas University Bulletin 38, 1409-1438. S T A N I E R R.Y., , P A L L E R O N IN, . J . & D O U D O R O F FM ,. (1966) The aerobic pseudomonads : a taxonomic study. Journal of General Microbiology 43, 159-271.
