International Journal of Food Microbiology, 10 (1990) 303-316 303 Elsevier

FOOD 00314

The bacteriology of fresh and spoiling Lake Victorian Nile perch ( Lates niloticus)

Lone Gram, Christina Wedell-Neergaard and Hans Henrik Huss Technological Laboratory, Ministry of Fisheries, Technical University, Lyngby, Denmark

(Received 11 April 1989; accepted 30 December 1989)

A total of 177 bacterial cultures isolated from Lake Victorian Nile Perch (Lates niloticus) were investigated. The flora on newly caught Nile perch consisted of organisms belonging to the genera Moraxella, Alcaligenes, Acinetobacter, Pseudomonas, Aeromonas, Micrococcus and other Gram-positive organisms. 39% were identified as Gram-positive species and 61% were negative in the Gram-reaction. Three cultures out of 53 investigated caused weak rotten off-odours in sterile fish broth and one culture, an Aeromonas spp. produced strong rotten, fishy, hydrogen sulphide off-odours. From Nile perch spoiled at ambient temperature, 15 of the 42 strains isolated caused rotten, fishy, hydrogen sulphide off-odours. These specific spoilage bacteria were all identified as Aeromonas and all reduced trimethylamine oxide to trimethylamine and produced hydrogen sulphide. From spoiled iced Nile perch, 74 out of 82 (90%) of the bacteria isolated were identified as Pseudomonas. A small proportion of these (13 out of 74) produced off-odours in sterile fish broth resembling the spoiling fish. These specific spoilers could not be separated from the non-spoilers based' on biochemical activities used in classical taxonomy. While the Pseudomonas spp. isolated did not produce trimethylamine or H2S, a few of the remaining isolates (two Shewanella putrefaeiens and five Aeromonas spp.) did produce these compounds. The role of Shewanella putrefaciens in the iced spoilage of Nile perch was, however, insignificant, since they only very late in the storage reached numbers where their spoilage could be detected.

Key words: Nile perch; Aeromonas; Pseudomonas; Specific spoilers; Fish spoilage

Introduction

T h e qua l i t y c h a n g e s o c c u r r i n g d u r i n g i ced s t o r ag e o f f i sh h a v e b e e n i n v e s t i g a t e d

in m a n y s tud ies , a n d d u r i n g the p a s t 30 yea r s p r ec i s e k n o w l e d g e has b e e n g a i n e d o n

th is sub jec t . T h e c o n c l u s i o n s d r a w n in n u m e r o u s s tud ies , r e v i e w e d by L i s t o n (1980)

a n d H o b b s a n d H o d g k i s s (1982), a re t ha t b a c t e r i a l d e c o m p o s i t i o n o f the n o n - p r o -

Correspondence address." L. Gram, Technological Laboratory, Ministry of Fisheries, Technical University, bldg. 221, DK-2800 Lyngby, Denmark.

0168-1605/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

304

tern-ni t rogen fract ion of tile flesh gives rise to the of f -odours associa ted with spoilage. The autoly t ic changes preceding bacter ial growth arc general ly believed to be insignif icant in the true spoilage. The microfh)ra fol.nld on newl,¢ caught fish is dependen t on the water in which the fish b, caught (Shewan, 1977) and is on fish from cold marine waters domina ted bv ( ; r am-nega t ive species (Pseudomomls, Moraxella, Flarobacterium and Shewanella). Often Gram-pos i t i ve genera as Micro- coccus, coryneforms and Bacillus can be identif ied as well, and thesc may e',en on fish from toM waters accounl for 5(}¢~ of the n-licroflora (Shewan ct al. 1960). ()f these bacter ia , only a few genera are par t ic ipa t ing in the spoilage. The spccific spoi lage bacteria, especially' 5;hewanella [)lttrQlil('iPlt& have been associa ted with fish spoi lage due to their ab i lhv to reduce t r imethy lamine oxide and p roduce volat i le su lphides from sulphur con ta in ing amino acids (Herber t and Shewan, 1975l. A number of Pseudomonas spp. p roduce of f -odours recognizable in the odour profi le of spoi l ing fish (Castell and (h"eenough, 1957).

Whils t the above picture is well cc>nfirnled for fish l'ronl toM marine \~,aters. much less is kno,an about spoi lage of fish from warm. tropical v, aters. Early da ta on t ropical fish species were included in the review by Shewan (1977), and he con- c luded that new'Iv caught fish from warm waters carr ied a heavier bacter ia l load than fish from cold waters, and that the former had a higher count of Gram-pos i t i ve species. This was, however, ques t ioned by Lima dos Santos (1978). who observed only a tendency to a lower count of t'.veu~&m(ma.~ spp. on warm water fish as c o m p a r e d to cold water fish. Many studies have dur ing the last decade de te rmined the s torage life of warm water fish s tored in ice, but the b iochemis t ry of spoi lage has not been s tudied as thorough as for t o m watcr fish. Rarely has the role of par t i cu la r bacter ia l species been documcnted . Even less da ta :.ire avai lable on spoi lage pa t te rns of warm water fish at ambicrl t temperatures , t towever . Tasman ian studies have ident i f ied Aeromonas lo'drophila :.is specific spoi lagc bacteria of trout s tored at 3 7 o ( ` (Gorczyca et al., 1985: Gorczyca and Pek Poh Len, 1985).

The qual i ty changes dur ing s torage in ice and at ambien t t empera tu re of Nile perch (l.ates mlot~cus) from I.akc Victoria ,~serc in ' ,es t igaled dur ing a rcccnl t::AO project in Kenya ( ( ; r a m ct al.. 1989). It ~as concluded thal Nile perch stored at ambien t t empera tu re (20 300(") spoi led rapidly and was unaccep tab le for human consumpt ion aftcr 11 17 h, whereas icing ensured a su~rage lifc of 4 weeks. [:ish caught in t empera te waters rarclx, keep more than 2 3 v~ccks in ice and the s torage trials with Nile perch confirnned that fish C:.lught ill ,aal-ill waters often have ex tended iced s torage lives as comparcd to tempera te species (Poul ter et al., 1981). Several p roposa ls have been made to explain the long iced s torage lives of warm water fish such as suggesting a lowcr initial nulnbcr of p,, ,ychrotrophic bacter ia (Shewan, 1977), a different spoi lage flora (Nai r et al.. 1974) c,r a non-bacter ia l spoi lage (Bremner et al., 1988). None of the hypotheses can. however, be conf i rmed or rejected due to the lack of exper imenta l data. The present study' is a con t iuua t ion of the s torage trials with Nile perch ( G r a m ct ill.. 1989), /lnd the purpose has been to ident i fy and charactcr izc the t-,actcrial flora Iound on fresh and spoil ing Nile perch and to de te rmine the role of the various bacterial species in the spoi lage process in o rde r to address the observed long iced sewage stabi l i ty of this fish species.

305

Materials and Methods

Media Iron agar was used for counting total number of bacteria as well as hydrogen

sulphide producing organisms. Iron agar was also used for isolation and pure culturing of bacterial strains and TMAO-medium was used when testing for ability to produce trimethylamine and hydrogen sulphide. Both were prepared and used according to Gram et al. (1987). Veal Infusion Broth (Difco) was used when growing pure cultures. Sterile fish broth was prepared from Nile perch and cod according to Gram et al. (1987).

Isolation of bacteria The quality changes during storage at ambient temperature and in ice of Nile

perch ( Lates niloticus) from Lake Victoria were studied using sensory, chemical and microbiological techniques as described by Gram et al. (1989). Two fish were taken on each day of sampling and tenfold dilutions were prepared from skin segments from the two fish. Psychotrophic organisms were enumerated by incubating iron agar plates at 5 °C for 14 days and colony forming units (cfu) were counted on iron agar plates incubated at room temperature (25-30 o C) for 3 days.

The bacterial flora from fresh and spoiling fish was investigated. Colonies were picked from iron agar plates sampled from the fresh fish and from the spoiled fish (i.e. poured on the day the fish passed the limit of acceptability).All colonies from a sector of the plate or all colonies from a whole plate were isolated and purified by subcultivation in veal infusion broth and on iron agar.

Identification and characterization The strains were characterized using a number of tests shown in Table I and

tentatively identified based on the seven first-mentioned reactions following Dainty et al. (1979) and Bergey's Manual (Krieg and Holt, 1984).

Spoilage potential The bacterial cultures were all inoculated in sterile Nile perch and sterile cod

broth (Gram et al., 1987) in order to determine spoilage potential. Mesophilic bacteria were incubated at 25 °C for 2 days and psychrotrophic strains at 5 °C for 7 days. A panel of six people, all trained in quality assessment of fish, evaluated the developing odours. They were asked to indicate the strength of the odour and describe it using terms as "fishy", "sulphidy", "cabbage-like", "sour", "sweet" etc.

Amino acid analyses Mesophilic and psychrotrophic bacteria were grown in sterile Nile perch broth

for 4 days at 25 ° C and 3 weeks at 5 o C, respectively, and the free amino acid profile determined with a Kontron Liquimat III Automatic Amino Acid Analyzer as described by Klausen and Lund (1986).

306

TABLE I

Biochemical reactions used when identifying bacteria isolated from fresh and spoiling Nile perch

Reaction Medium References

Gram reaction 3% KOH Gregersen (1978) Shape/moti l i ty Phase contrast microscopy Cytochrome oxidase Tetramethyl p-phenylen

diamin-dihydrochloride Kovacs (1956) Catalase 3% H202 Glucose metabolism Hugh and Leifson 1 Hugh & Leifson (1953) TMAO-reduct ion TMAO-med ium Gram et al. (1987) H2S-production TMAO-med ium Gram et al. (1987) Nitrate reduction API 20NE 4 Indole production API 20NE Arginine dihydrolase API 20NE Urease API 20NE Aesculin hydrolysis API 20NE Gelatine hydrolysis API 20NE fl-Galactosidase API 20NE Glucose 2 API 20NE Arabinose 2 API 20NE Mannose 2 API 20NE Mannitol 2 API 20NE N-acetylglucosamin 2 API 20NE Maltose 2 API 20NE Gluconate 2 API 20NE Caprate 2 API 20NE Adipate 2 API 20NE Malate 2 API 20NE Citrate 2 API 20NE Phenyl-acetate 2 API 20NE Growth at 4 ° C /41 o C Veal Infusion Broth Acetyl methyl carabinol

production Peptone-glucose broth Clark & Lubs (1915) Mannitol 3 Hugh & Leifson Hugh & Leifson (1953) Arabinose 3 Hugh & Leifson Starch 3 Hugh & Leifson Aesculin 3 Aesculin broth Cowan (1974) Haemolyses Blood agar Resistance to vibriostaticum 0 /129 (2,4-diamino-6,7-

di-isopropyl pteridin phosphate) Bain & Shewan (1968)

Fluorescent pigment King's Agar B King et al. (1954)

1 5 g of peptone was substituted by 1 g of (NH4)2SO 4. 2 Assimilation of the carbon sources. 3 Hydrolyses of the carbon sources. 4 API Systems S.A.

Results

307

F l o r a o n f r e s h f i s h

The count on i ron agar incuba ted at 2 5 - 3 0 ° C for 3 days was 104-106 c f u / c m 2 on newly caught Nile perch and the n u m b e r of psychrotrophic bacter ia was 104-105 c f u / c m 2, indica t ing that a significant fract ion of the microflora on Nile perch is to

be characterized as psychrotrophic organisms. Hydrogen sulphide-producing organisms occurred only in low number s and never exceeded 1% of the total flora on newly caught fish. The flora consisted of a variety of organisms, of which 39% were Gram-posi t ive and 61% were Gram-negat ive . The biochemical reactions and the subsequent ident if icat ions of the 53 strains isolated from 25 ° C Iron Agar plates are shown in Table II.

The majori ty (52 of 53) were unable to reduce t r imethylamine oxide (TMAO) and did not produce hydrogen sulphide. All these were classified as non-spoi lers since no offensive off-odours were produced in sterile fish broth. They belonged to different genera as M o r a x e l l a , A l c a l i g e n e s , A c i n e t o b a c t e r , P s e u d o m o n a s , M i c r o c o c c u s ,

and other Gram-pos i t ive organisms (Table II). Only one strain, isolated from a black colony on iron agar, was capable of

reducing T M A O and produced H2S. This was the only strain which produced fishy, hydrogen sulphide off-odours in sterile fish broth. This Gram-negat ive , motile bacter ia was fermentat ive and showed positive oxidase and catalase reactions. In further tests it was resistant to vibriostat icum, did no t produce gas from glucose, and hydrolysed starch, manni to l , arabinose and aesculin. It was classified as a 'mo t i l e ae romonad ' bu t the pa t te rn of reactions did not fit any of the described

TABLE 11

Composition of the bacterial flora on newly caught Nile perch

Number Gram Shape Mo- Oxi- Cata- H&L TMAO H2S Off Identification of til- dase lase reduc- odour (species) strains ity tion

3 -- r + + + Ox - _ _ 1 P s e u d o m o n a s

19 - c/r _ / + + + . . . . 1 M o r a x e l l a

5 - r + - + . . . . A I c a l i g e n e s

4 - c/r - - + . . . . A c i n e t o b a c t e r

1 - r + + + F + + + A e r o m o n a s 2

11 + c + - ~ + . . . . M i c r o c o c c u s

5 + c + _ 4 + F - - - S t a p h y l o c o c c u s

5 + r + - / + + - / F - - - B a c i l l u s / C o r y n e f o r m

53 Total number

a Strain produced weak off odours. 2 Resistant to vibriostaticum. 3 Two strains were oxidase positive. 4 One strain was oxidase positive.

308

species (Krieg and Holt, 1984). One Moraxella and one Pseudomonas produced weak sour off-odours in sterile fish broths.

Flora on fish stored at ambient temperature For fish stored at ambient temperature, the bacterial count on iron agar was

5 .10 -108 c fu / cm 2 when the fish were rejected. The proportion of H2S-producers increased during storage and they constituted 20-50% of the bacterial flora on the unacceptable fish. Fifteen of the 42 cultures isolated from Nile perch spoiled at ambient temperature were Aeromonas spp. (Table II1). The remaining bacterial strains belonged to almost the same genera as found on the fresh fish with the majority being Gram-negative, i.e. Acinetobacter, Moraxella, and Alcaligenes. Seven strains were Gram-positive cocci. One of the non-fermentative Gram-negative strains produced sour, rotten off-odours in sterile fish broth, but otherwise none which appeared as white colonies on iron agar showed any spoilage potential. Correspondingly, none reduced TMAO or produced H2S.

The above mentioned 15 Aeromonas isolates produced off-odours characterized as "fishy", "rot ten" and "sulphidy" when inoculated in sterile fish broth. All caused a black precipitate to be formed on iron agar and all were positive in the TMAO-medium. They also showed haemolytic activity with a green clearing zone on blood agar plates. Four of these aeromonads were grown in sterile Nile perch broth at 25 °C for 4 days and the free amino acid profiles determined to investigate the spoilage pattern in more detail (Table V). The growth of all aeromonads resulted in a large increase in NH 3 and in more than a doubling of the amount of free amino acids (from 13 nmol to 33 nmol amino acid per ml). The composition of free amino acid in fresh Nile perch was similar to that of cod except for a very high content of taurine, which constituted approximately 75% of the free amino acids. In cod, this amino acid accounts for 40% (E. Lund, personal communication).

TABLE I l l

Composit ion of the bacterial flora on Nile perch spoiled at ambient temperature

Number Gram Shape Mo- Oxi- Cata- H&L TMAO H2S Off Identification of til- dase lase reduc- odour (species) strains ity tion

1 5 - r + + + F + + + A e r o m o n a s 1

6 - r + + - 2 A l c a l i g e n e , ~

7 c / r - - - + A c i n e t o b a c t e r

6 r - / + + + M o r a x e l l a

4 + c - - - + M i c r o c o c ~ ' u s

3 + c + F - S t a p h y [ o c o c { ' u . v

1 + r - + . . . . ?

42 Total number

All strains resistant to vibriostaticum. 2 One strain produced sour, rotten off odours in cod broth.

309

Flora on fish stored in ice Icing of Nile perch caused a slight decrease in the bacterial count and a lag phase

of 5-7 days was seen before proliferation started (data not included). The counts on plates incubated at 5 °C and at ambient temperature were alike in the last 3 weeks of the storage period and the count had reached 10 s 10 ~° c fu / cm z when the fish were rejected after 4 weeks, Only a small proportion of these were HeS-producers, and occurring in numbers of 106-10 v c fu / cm z on rejectable fish. The bacterial flora found on Nile perch after 33 days in ice was dominated by Pseudomonas spp. (Table IV). Of 82 isolates 74 were identified as Pseudomonas spp. Seven strains were positive in the TMAO medium and these strains also produced fishy, hydrogen sulphide off-odours when inoculated in sterile fish broth. They were capable of producing TMA and HeS at 0°C, but positive reactions were not evident until after 3-6 weeks. Two were identified as Shewanella putrefaciens due to their inability to ferment glucose, whereas the remaining five were fermentative in Hugh and Leifson's medium (H&L) and were identified as Aeromonas (resistant to vibriostaticum). Twelve of the 74 strains identified as Pseudomrmas spp. produced strong fruity/sulfhydryl odours (characterized as "mango", "'fruit", "'sickening sweet", "'sulphur") when growing in cod broth and Nile perch broth, whereas one strain developed similar weak off-odours. These odours were very similar to those noticed in spoiling Nile perch. Five strains produced weak rotten, sour, muslv off-odours in cod broth whereas no off-odours were detected in Nile perch broth. These isolates were further characterized on API 20NE and a number of supplementary tests and it was not possible to differentiate between the five cod-spoilers and the 13 Nile perch spoilers on the basis of their biochemical reactions (data not included).

Five Pseudomonas spp. causing spoilage and five Pseudomonas spp. which did not produce off-odours were inoculated in sterile Nile perch broth, and the free amino acid profiles were determined after incubation at 5 °C for 3 weeks. Varying

T A B t . E IV

Compos i t ion of the bacterial flora on Nile perch spoiled in ice

N u m b e r G r a m Shape Mo- Oxi- Cata- H & L T M A O H~S Off Ident i f icat ion

of til- dase lase reduc- odour (species)

strains ity tion

56 r + + + Ox - - P ~ e u d o m c m a s

13 - - r + I @ q_ ( } x - Jr- Ps'eudomonas

5 r + + + Ox + 2 P ~ e u d o m o n a s

5 - r + + + F + + + . , l e r o m o n a x

2 r + ~ + - - / O x + + + . S ' h e w a n e l l a

1 + r - + O x "

82 [ o t a l n u m b e r

1 T w o strains were non-moti le .

2 Off odours were only p roduced in cod extract.

3 Resis tant to vibr ios ta t icum.

310

TABLE V

Amino acid profiles in Nile perch broth during growth of bacteria isolated from spoiling Nile perch. The analysis was carried out after incubation at 5 ° C for 3 weeks for the Pseudomonas spp. and at 25" C for 4 days for the Aeromonas spp.

Sample: NP 1 Pseudomonas spp. Aeromonas spp.

Amino acid 2 Non-spoilers

(5 isolates Spoilers (5 isolates)

Spoilers (4 isolates)

Aspargic acid ND 3 ND-0.12 ND 0.46 0.19-0.24 Threonine 0.20 0.06-0.12 0.02-0.22 0.00-0.52 Serine 0.18 0.04-0.08 0.04-0.17 0.32-0.49 Glutamic acid 0.1 ~ 0.08-0.50 0.00-0.46 2.44-2.92 Proline 0.13 ND ND 0.56-0.91 Glycine 1.82 0.45-1.31 0.86-1.71 3.79-4.91 Alanine 0.83 0.23-0.43 0.05-1.08 4.87-5.49 Valine 0.15 0.04-0.15 0.02 -0.47 1.17 - 1.42 Methionine 0.05 0.00-0.14 0.00- 0.43 1.03 - 1.57 Isoleucine 0.08 0.05-0.07 0.00-0.23 0.92-1.11 Leucine 0.13 0.10-0.15 0.00 0.43 1.73-2.20 Tyrosine 0.05 ND ND 0.09-0.19 Phenylalanine 0.06 ND ND 0.43-0.50 Lysine 0.16 0.16-0.19 0.03-0.38 3.18-4.40 Histidine 0.07 0.06-0.08 0.01-0.14 0.11-0.16 Arginine 0.05 0.00-0.05 0.00-0.09 0.00 Taurine 8.96 8.38-9.09 8.40-8.86 8.49-9.65 Ammonia 4 2.45 3.61 8.46 3.43-5.37 28.2 -35.3

Total amino acid 13.1 9.9 -11.6 9.8 -14.2 32.8 -34

1 Sterile Nile perch broth. 2 nmol amino acid per ml of Nile perch broth (range for the isolates 3 Not detected. 4 Including nitrogen formed during hydrolysis.

investigated).

(small) amounts of ammonia were produced by all 10 strains and no significant increase was seen in any of the amino acids (Table V). It appears that the five spoilage bacteria may have caused a decrease in the content of some of the amino acids, e.g. threonine, serine, alanine and valine. The non-spoilers were variable. No consistent pattern emerged from these experiments, i.e. no single (group of) amino acid(s) could be identified as substrate for the off-odours. However, the differences between these species and the Aeromonas spp. identified as spoilage organisms are clearly demonstrated in Table V.

Discussion

The natural bacterial flora found on newly caught Nile perch was composed of a variety of Gram-negative and Gram-posit ive organisms, and the genera identified resemble the flora on temperate water fish (Hobbs and Hodgkiss, 1982). These

311

findings are not in agreement with Shewan (1977) who concluded that the micro- flora on tropical fish differed from the microorganisms on temperate fish by having a much larger percentage of Gram-posit ive species. However, the data agree with the conclusion reached by Colwell and Liston (1962) and by Lima dos Santos (1978) who found that apart from a slightly higher incidence of Gram-posit ive bacterial types, the bacterial flora on tropical fish resembles the microflora on temperate fish species. The hypothesis that tropical fish carry a much lower number of psychrotrophic bacteria than do temperate fish (Shewan, 1977) is not supported by the data from Nile perch. The psychrotrophic count at 5 °C amounted to 10% or more of the total number of bacteria. It is therefore not likely that a the prolonged iced storage stability of tropical fish is caused exclusively by an especially low number of psychrotrophs, Similar to the findings of Gillespie and MacRae (1975) none of the Gram-posit ive strains isolated in this work were classified as spoilage bacteria. Only one of the strains isolated from the newly caught fish was classified as spoiler. This strain was identified as an Aeromonas spp. and it was identified as the dominant spoilage organism of Nile perch stored at ambient temperature. The results are thus partly in agreement with the experiments by Gorczyca et al. (1985) and Gorczyca and Pek Poh Len (1985), in which Aeromonas hydrophila was identified as the active spoilage organism on Australian fish stored without cooling. In the present study a variety of motile aeromonads were found, and no single subspecies could be pin-pointed as spoiler. All bacteria producing fishy and rotten off-odours appeared as black colonies on iron agar and were positive in the TMAO-medium, whereas none of the strains isolated as 'white colonies' were classified as spoilage organisms. This is not in agreement with the Australian findings (Gorczyca et al. 1985, Gorczyca and Pek Poh Len, 1985) where only a minority (17%) of the spoilage aeromonads produced hydrogen sulphide. They concluded that the number of H2S-producers did not indicate the number of spoilage bacteria. In contrast, we have demonstrated that the iron agar gives a very good indication of the number of ambient spoilage bacteria on Nile perch. The analyses of free amino acids in extracts inoculated with aeromonads showed that these were strongly proteolytic resulting in a significant production of ammonia which probably adds to the odours profile of the spoiling fish.

All Aeromonas spp. were hemolytic, leading to a green discolouration. This reaction probably explains the greening of the gills observed when Nile perch spoiled at ambient temperature (Gram et al., 1989). The green pigment may be caused by reaction between hydrogen sulphide and myoglobin resulting in forma- tion of sulfmyoglobin (Taylor and Shaw, 1977).

The qualitative changes in the microflora observed during storage in ice were very similar to the changes seen on iced fish from temperate waters (Liston, 1980). The psychrotrophic flora isolated from spoiled iced Nile perch was dominated by Pseudomonas spp. which did not produce TMA and H2S. The analysis of the newly caught fish indicated that these Pseudomonas were part of the natural flora. The off-odours produced by some of the psychrotrophic Pseudomonas spp. in sterile Nile perch broth were very pronounced and resembled the off-odours and off-flavours which characterized the spoiling Nile perch.

312

Biochemical testing of the psychotrophic bacteria showed a rather complex pattern of reactions and the 'spoilers' did not differ significantly from the non- spoilers. The same conclusion was reached by Gillespie (1981) who investigated psychrotrophic Pseudomonas isolated from spoiled Australian fish. He defined a group of Pseudomonas fragi and this group included both spoilage and non-spoilage organisms inseparable by biochemical reactions.

Recent investigations of food-associated pseudomonads have classified these into clusters based on numerical analyses of a large number of biochemical features (Gillespie, 1981; Shaw and Latty, 1982; Molin and Ternstr/Sm, 1986: Shelley et al., 1987). It is evident from these studies that quite a variation can be found within a cluster; thus Molin and Ternst~Sm (1986) reported that 4% of the strains belonging to the (non-fluorescent) Pseudomonas fragi cluster actually produced fluorescent pigment.

The off-odours produced by the spoilage bacteria in the present study are very like those described for Pseudomonas fragi (Castell et al., 1959), and many of the biochemical reactions resemble this non-fluorescent Pseudomonas spp., although some of the isolates showed positive urease- and nitrate-reactions and produced fluorescent pigments which are normally not found for Pseudomonas fragi (Castell et al., 1957). These reactions are, however, found for Pseudomonas putida, which also has been implicated in fish spoilage (Herbert et al., 1971).

According to our experience fluorescence is an unstable characteristic, and hence unsuitable in the identification procedure. The same conclusion was reached by Shaw and Latty (1984) and they suggested that identification of Pseudomonas could be based on assimilation of 18 carbon sources. Whether such an approach would show any taxonomic difference between spoilers and non-spoilers of Nile perch is not known.

For the Pseudomonas spp. the amino acid profile did not lead to any firm conclusions. However, some amino acids, e.g. valine and serine appeared to be decomposed and degradation of these amino acids has been shown to give rise to the fruity and onion like off-odours produced by Pseudomonas fragi (Castell et al., 1959). No significant changes were seen in any of the sulphur containing amino acids and thus the substrate for the volatile sulphides detected by the sensory analyses could not be defined. In agreement with results of Herbert and Shewan (1975), the major free amino acid, the sulphur containing taurine, was not decom- posed.

Analysis of the odour profile, e.g. by Gas Chromatography, may reveal if any specific compounds are responsible for the off-odours as described by Miller et al. (1973abc) and Edwards et al. (1987) who showed that a variety of aldehydes, ketones and volatile sulphides were produced by Pseudomonas spp. But they did not indicate the compound(s) acting as substrate for the bacteria.

Five strains were characterized as 'cod spoilers' since off-odours were detectable in cod broth but not in Nile perch broth. These offoodours were described as sour, musty and rotten rather than the fruity, sweet, sulphydryl terms used for the other isolates. These five strains may be related to Pseudomonas perolens, which as described by Castell et al. (1957), produces musty, potato-like odours.

313

Psychrotrophic strains forming black colonies on iron agar and cultivated from fish stored 4 weeks in ice were identified as either Shewanella putrefaciens or Aeromonas spp. Normally, Aeromonas is associated with mesophilic environments, but it is apparently able to adapt to chill temperatures (Barile et al., 1985: Ravn-Jorgensen, 1986). These strains reduced TMAO and produced hydrogen sulphide at 0 ° C, but positive reactions took approximately 4 weeks to develop in the medium. Ravn Jargensen et al. (1988) found that TMAO reduction in iced fish started when the bacterial count reached 108 cfu/g , and Huss et al. (1988) reported that the reduction of TMAO in a fish slurry started when cell concentrations were above 10 7 cfu /ml . Thus the time taken for the present isolates to reach numbers where TMA and H2S can be detected appears to be longer than e.g. in iced cod where the number of black colonies reach these levels in about 2 weeks (Ravn Jorgensen et al. 1988). This may be caused by a period needed for adaptation to chill temperatures of the psychrotrophic organisms on Nile perch and probably explains why only low amounts of TMA were found in the iced fish during the storage trials, and only in fish stored for more than 3 weeks where the number of black colonies exceeded 108 c f u / c m 2 (Gram et al., 1989).

Castell and Greenough (1957) found that the shelf life of iced fish was prolonged when TMAO-reducers were inhibited, and similar observations were published by Veerappa and Karunasagar (1985). The absence of TMA from the odour profile of Nile perch and a possible different spoilage potential of the Pseudomonas spp. may be factors explaining the observed long shelf life of iced Nile perch, i.e. 28-30 days.

It is intriguing why the Shewanella putrefaciens did not play a more important role in the spoilage of iced Nile perch, since this bacteria has in many studies been identified as the most important specific spoilage bacteria of iced fish (Gram et al., 1987, Ravn-Jorgensen and Huss, 1989). The bacteria were present and capable of growing and reducing TMAO at 0 ° C. The increase in hydrogen sulphide producing organisms, which was seen during iced storage was parallel to the increase in total count. This means that an initial lower number of these organisms as compared to the Pseudomonas spp. causing spoilage but not producing H2S is the most apparent reason for the insignificant role of H2S-producing organisms in the iced spoilage. The generation times at 0 ° C for a Shewanella putrefaciens and a spoilage Psue- domonas spp. were determined and found to be 20 and 14 h, respectively (Gram, 1989). Different growth rates may explain different spoilage potentials. Shelley et al. (1986) found Pseudomonas fragi to be more important in the spoilage of milk at 4 ° C than Pseudomonas fluorescens, due to a higher growth rate of the former.

In conclusion, the composition of the microflora on fresh Nile perch resembles the microflora on cold water marine fish, apart from a slightly higher incidence of Gram-posit ive species. Motile aeromonads were causing the spoilage at ambient temperature, partly due to their ability to produce TMA and HzS. During iced storage, Gram-negative rods (primarily Pseudomonas spp.) are favoured. Despite the content of TMAO in Nile perch production of TMA and H2S did not characterize the spoilage.

314

Acknowledgements

T h e a s s i s t a n c e o f K e n y a M a r i n e a n d F i s h e r y R e s e a r c h S t a t i o n a n d o f J o h a n n e s

B o n f r o m T N O , H o l l a n d , d u r i n g t he w o r k in K i s u m u is a c k n o w l e d g e d . T h e a m i n o

a c i d a n a l y s i s was c a r r i e d o u t b y E r i k L u n d a n d the m e s o p h i l i c a e r o m o n a d s w e r e

i d e n t i f i e d a c c o r d i n g to p r o c e d u r e s d e s c r i b e d b y S u s a n n e K n o c h e l . A g r a n t f r o m t h e

R o y a l V e t e r i n a r y a n d A g r i c u l t u r a l U n i v e r s i t y , C o p e n h a g e n , f i n a n c e d t he w o r k of

o n e o f t he a u t h o r s .

References

Bain, N. and Shewan, J.M. (1968) Identification of Aeromonas, Vibrio and related organisms. In: B.M. Gibbs and D.A. Shapton (Eds.), Identification Methods for Microbiologists part B, Academic Press, New York, NY, pp. 79-84.

Baffle, L.E., Estrada, M.H., Milla, A.D., Reilley,A. and Villadsen, A. (1985) Spoilage patterns of mackerel (Rastrelliger faughni Matsui). 2. Mesophilic versus psychrophilic fish spoilage of tropical fish. ASEAN Food J. 1 (3).

Bremner, A., Olley, J. Statham, J.A. and Vail, A.M.A. (1988) Nucleotide catabolism: Influence on the storage of tropical species from the North West Shelf of Australia. J. Food Sci. 53, 6-11.

Castell, C.H. and Greenough, M.F. (1957) The action of Pseudomonas on fish muscle. 1. Organisms responsible for odours produced during incipient spoilage of chilled fish muscle. J. Fish Res. Board Can. 14, 617-625.

Castell, C.H., Greenough, M.F. and Jenkins, N.L. (1957) The action of Pseudomonas on fish muscle: 2. Musty and potato-like odours. J. Fish. Res. Board Can. 14 (5), 775-782.

Castell, C.H., Greenough, M.F. and Dale, J. (1959) The action of Pseudomonas on fish muscle. 3. Identification of organisms producing fruity and oniony odours. J. Fish. Res. Bd. Can. 16, 13-19.

Clark, W.M. and Lubs, H.A. (1915) The differentiation of bacteria of the colon-aerogenes family by the use of indicators. J. Infect. Dis. 17, 160.

Colwell, R.R. and Liston, J. (1962) Bacterial flora on seven species of fish collected at Rongelap and Eninetok Atols. Pacific Sci. 16, 264-270.

Cowan, S.T. (1974) Cowan and Steel's Manual for The Identification of Medical Bacteria, 2nd edn, Cambridge University Press, Cambridge.

Dainty, R.H., Shaw, B.G., Hardinger, C.D. and Michanie, S. (1979) The spoilage of vacuum packaged beef by cold tolerant bacteria. In: A.D. Russell and R. Fuller, (Eds.), Cold Tolerant Bacteria in Spoilage and the Environment, Academic Press, New York, NY, pp. 83-110.

Edwards, R.A., Dainty, R.H. and Hibbard, C.M. (1987) Volatile compounds produced by meat pseudomonads and related reference strains during growth on beef stored in air at chill temperatures. J. Appl. Bacteriol. 62, 403-412.

Gillespie, N.C. (1981) A numerical taxonomic study of Pseudomonas-like bacteria isolated from fish in Southeastern Queensland and their association with spoilage. J. Appl. Bacteriol. 50, 29-44.

Gillespie, N.C. and MacRae, I.C. (1975) The bacterial flora of some Queensland fish and its ability to cause spoilage. J. Appl. Bacteffol. 39, 91-100.

Gorczyca, E. and Pek Poh Len (1985) Mesophilic spoilage of bay trout (Arripis trutta), bream (Acanthropagrus butcheri) and mullet (Aldrichetta forsteri). Reilly, A. (Ed.), FAO Fish Rep. (317) Suppl. 474 p., 123-132.

Gorczyca, E., Sumner, J.L. Cohen, D. and Brady, P. (1985) Mesophilic fish spoilage. Food Technol. Austral. 37, 24-26.

Gram, L. (1989) Identification, characterization and inhibition of bacteria isolated from tropical fish. Ph.D. Thesis, Technological Laboratory, Ministry of Fisheries, Lyngby.

315

Gram, L., Trolle, G. and Huss, H.H. (1987) Detection o[ specific spoilage bacteria from fish stored at low (0 o C) and high (20 ° C) temperatures. Int. J. Food Microbiol. 4, 65-72.

Gram, L., Oundo, J. and Bon, J. (1989) Storage life of Nile perch (Lates niloticus) dependent on storage temperature and initial bacterial load. Trop. Sci. 29, 221-236.

Gregersen, T. (1978) Rapid method for distinction of Gram-negative and Gram-positive bacteria. Eur. J. Appl. Microbiol. 5, 123-127.

Herbert, R.A. and Shewan, J.M. (1975) Precursors of the volatile sulphides in spoiling North Sea cod (Gadus rnorhua). J. Sci. Food Agricol. 26, 1195-1202.

Herbert, R.A., Hendrie, M.S., Gibson, D.M. and Shewan, J.M. (1971) Bacteria active in the spoilage of certain sea foods. J. Appl. Bacteriol. 34, 41-50.

Hobbs, G. and Hodgkiss, W. (1982) The bacteriology of fish handling and processing. In: R. Davis, (Ed.): Developments in Food Microbiology 1, Applied Science Publishers, London and New Jersey.

Hugh, R. and Leifson, E. (1953) The taxonomic significance of fermentative versus oxidative Gram-nega- tive bacteria. J. Bacteriol. 66, 24-26.

Huss, H.H., Trolle, G. and Gram, L. (1988) New rapid methods in microbial evaluation of fish quality. In: D.E. Kramer and J. Liston (Eds.): Seafood Quality Determination, Elsevier, Amsterdam, pp. 299-308.

King, E.D., Ward, M.K. and Roney, D.E. (1954) Two simple media for the demonstration of pyocianin and fluorescin. J. Lab. Clin. Med. 44, 301-307.

Klausen, N.K. and Lund, E. (1986) Formation of biogenic amines in herring and mackerel. Zeitschrift fiir Lebensmittel-Untersuchung und -Forschung. 182, 459-463.

Kovacs, N. (1956) Identification of Pseudomonas pyocyaneae by the oxidase reaction. Nature 178, 703. Krieg, N.R. and Holt, J.G. (1984) Bergey's Manual of Systematic Bacteriology, Williams & Wilkins,

Baltimore & London. Lima dos Santos, C.A.M. (1978) Bacteriological spoilage of iced Amazonian freshwater catfish

(Brachyplatistorna vaillanti Valenciennes). Master's Thesis, Loughborough University of Technology, Loughborough.

Liston, J. (1980) Microbiology in fishery science. In: J.J. Connell (Ed.): Advances in Fishery Science and Technology, Fishing News Books, Farnham, Surrey, pp. 138-157.

Miller Ill, A., Scanlan, R.A., Lee, J.S., Libbey, L.M. and Morgan, M.E. (1973a) Volatile compounds produced in sterile fish muscle (Sebastes melanops) by Pseudomonas perolens. Appl. Microbiol. 25 (2), 257-261.

Miller III, A., Scanlan, R.A., Lee, J.S, and Libbey, L.M, (1973b) Identification of the volatile compounds produced in sterile muscle (Sebastes rnelanops) by Pseudomonas fragi. Appl. Microbiol. 25 (6), 952-955.

Miller III, A., Scanlan, R.A., Lee, J.S. and Libbey, LM. (1973c) Identification of the volatile compounds produced in sterile muscle (Sebastes melanops ) by Pseudomonas putrefaciens, Pseudomonas fluorescens, and an Achromobaeter species. Appl. Microbiol. 26, 18-21.

Molin, G. and TernstriSm, A. (1986) Phenotypically based taxonomy of psychrotrophic Pseudomonas isolated from spoiled meat, water and soil. Int. J. Syst. Bacteriol. 36, 257 274.

Nair, R.B., Tharamani, P.K. and Lahiry, N.L. (1974) Studies on the chilled storage of fresh water fish. II. Factors affecting quality. J. Food Sci. Technol. 11,118-122.

Poulter, R.G., Curran, C.A. and Disney, J.G. (1981) Chill storage of tropical and temperate water fish-differences and similarities. Proceedings of the IIR Conference on Advances in the Refrigerated Treatment of Fish, 111-124. International Institute of Refrigeration, Boston, 1981-4.

Ravn Jorgensen, B. (1986) Sorbic Acid - Effect on Fish - The Influence of pH on the Effect of Sorbic Acid. Master's Thesis (in Danish), Technological Laboratory, Ministry of Fisheries, Lyngby.

Ravn Jorgensen, B, and Huss, H.H. (1989) Growth and activity of ShewaneUa putrefaciens isolated from spoiling fish. Int. J. Food Microbiol. 9, 51-62.

Ravn Jorgensen, B., Gibson, D.M. and Huss, H.H. (1988) Microbiological quality and shelf life prediction of chilled fish. Int. J. Food Microbiol. 6, 295 307.

Shaw, B.G. and Latty, J.B. (1982) A numerical taxonomic study of Pseudomonas strains from spoiled meat. J. Appl. Bacteriol. 52, 219-228.

316

Shaw, B.G. and Latty, J.B. (1984) A study of the relative incidence of different Pseudomonas groups on meat using computer-assisted identification technique employing only carbon sources tests. J. Appl. Bacteriol. 57, 59-67.

Shelley, A.W., Deeth, H.C. and MacRae, I.C. (1986) Growth of lipolytic psychrotrophic pseudomonads in raw and ultra-heat treated milk. J. Appl. Bacteriol. 61, 395-400.

Shelley, A.W., Deeth, H.C. and MacRae, I.C. (1987) A numerical taxonomic study of psychrotrophic bacteria associated with lipolytic spoilage of raw milk. J. Appl. Bacteriol. 62, 197 207.

Shewan, J.M. (1977) The bacteriology of fresh and spoiling fish and the biochemical changes induced by bacterial action. Proceedings of the Conference on Handling, Processing and Marketing of Tropical Fish, Tropical Products Institute (Now: Tropical Development and Research Institute), London, pp. 51 66.

Shewan, J.M., Hobbs, G. and Hodgkiss, W. (1960) The Pseudomonas and Achrornobacter groups of bacteria in the spoilage of marine white fish. J. Appl. Bacteriol. 23, 463 468.

Taylor, A.A, and Shaw, B.G. (1977) The effect of meat pH and package permeability on putrefaction and greening in vacuum packed beef. J. Food Technol. 12, 515 521.

Veerappa, H.S. and Karunasagar, I. (1985) Effect of protease inhibitors from Adenanthera pa~,om on the hi{x:hemical and microbiological quality of fishes. J. Sci. Food Agric. 36, 1113-1119.