Genotypic and phenotypic characterization of garlic-fermenting lactic acid bacteria isolated from som-fak, a Thai low-salt fermented fish product.

AIMS: To evaluate the importance of garlic for fermentation of a Thai fish product, and to differentiate among garlic-/inulin-fermenting lactic acid bacteria (LAB) at strain level. METHODS AND RESULTS: Som-fak was prepared by fermentation of a mixture of fish, salt, rice, sucrose and garlic. pH decreased to 4.5 in 2 days, but omitting garlic resulted in a lack of acidification. LAB were predominant and approximately one third of 234 isolated strains fermented garlic and inulin (the carbohydrate reserve in garlic). These strains were identified as Lactobacillus pentosus and Lact. plantarum. Randomly Amplified Polymorphic DNA (RAPD) analysis revealed one major RAPD type (29 strains) isolated from all stages of fermentation. CONCLUSION: Garlic was essential for acidification of som-fak and garlic-fermenting strains constituted a significant, homogeneous part of the LAB flora. SIGNIFICANCE AND IMPACT OF THE STUDY: The present study indicates the role of fructans (garlic/inulin) as carbohydrate sources for LAB. Fructan fermenters may have several biotechnological applications, for example, as probiotics.

Significance of volatile compounds produced by spoilage bacteria in vacuum-packed cold-smoked salmon (Salmo salar) analyzed by GC-MS and multivariate regression.

Changes were studied in the concentration of 38 volatile compounds during chilled storage at 5 degrees C of six lots of commercially produced vacuum-packed cold-smoked salmon and sterile cold-smoked salmon. The majority of volatile compounds produced during spoilage of cold-smoked salmon were alcohols, which were produced by microbial activity. Partial least-squares regression of volatile compounds and sensory results allowed for a multiple compound quality index to be developed. This index was based on volatile bacterial metabolites, 1-propanol and 2-butanone, and 2-furancarboxaldehyde produced by autolytic activity. Only a few of the volatile compounds produced during spoilage of cold-smoked salmon had an aroma value high enough to indicate contribution to the spoilage off-flavor of cold-smoked salmon. These were trimethylamine, 3-methylbutanal, 2-methyl-1-butanol, 3-methyl-1-butanol, 1-penten-3-ol, and 1-propanol. The potency and importance of these compounds was confirmed by gas chromatography-olfactometry. The present study provides valuable information on the bacterial reactions responsible for spoilage off-flavors of cold-smoked salmon, which can be used to develop biosensors for on-pack shelf-life determinations.

Elucidation of Listeria monocytogenes contamination routes in cold-smoked salmon processing plants detected by DNA-based typing methods.

The contamination routes of Listeria monocytogenes in cold-smoked salmon processing plants were investigated by analyzing 3,585 samples from products (produced in 1995, 1996, 1998, and 1999) and processing environments (samples obtained in 1998 and 1999) of two Danish smokehouses. The level of product contamination in plant I varied from 31 to 85%, and no L. monocytogenes was found on raw fish (30 fish were sampled). In plant II, the levels of both raw fish and product contamination varied from 0 to 25% (16 of 185 raw fish samples and 59 of 1,000 product samples were positive for L. monocytogenes). A total of 429 strains of L. monocytogenes were subsequently compared by random amplified polymorphic DNA (RAPD) profiling, and 55 different RAPD types were found. The RAPD types detected on the products were identical to types found on the processing equipment and in the processing environment, suggesting that contamination of the final product (cold-smoked salmon) in both plants (but primarily in plant I) was due to contamination during processing rather than to contamination from raw fish. However, the possibility that raw fish was an important source of contamination of the processing equipment and environment could not be excluded. Contamination of the product occurred in specific areas (the brining and slicing areas). In plant I, the same RAPD type (RAPD type 12) was found over a 4-year period, indicating that an established in-house flora persisted and was not eliminated by routine hygienic procedures. In plant II, where the prevalence of L. monocytogenes was much lower, no RAPD type persisted over long periods of time, and several different L. monocytogenes RAPD types were isolated. This indicates that persistent strains may be avoided by rigorous cleaning and sanitation; however, due to the ubiquitous nature of the organism, sporadic contamination occurred. A subset of strains was also typed by using pulsed-field gel electrophoresis and amplified fragment length polymorphism profiling, and these methods confirmed the type division obtained by RAPD profiling.

Diversity of Listeria monocytogenes isolates from cold-smoked salmon produced in different smokehouses as assessed by Random Amplified Polymorphic DNA analyses.

One hundred and forty-eight Listeria monocytogenes isolates originating from vacuum packed cold-smoked salmon produced in 10 different Danish smokehouses were compared by Random Amplified Polymorphic DNA (RAPD) profiling. A total of 16 different reproducible RAPD profiles were obtained using a standardised RAPD analysis by four primers separately. The grouping of the 148 strains was exactly the same for the four primers used. For a sub-set of 20 strains typed by Pulsed Field Gel Electrophoresis (PFGE), only one strain was allocated into a different group as compared to the grouping by RAPD typing. Different RAPD types dominated in products from different smokehouses. Some identical RAPD types were isolated in several smokehouses. In each of four smokehouses, one particular RAPD type could be repeatedly isolated from products. Each smokehouse/product carried its own specific RAPD type and this may indicate a possible persistence of closely related strains of L. monocytogenes in smokehouses.

The effect of biogenic amine production by single bacterial cultures and metabiosis on cold-smoked salmon.

AIM: Biogenic amines are important indicators of spoilage in vacuum-packed cold-smoked salmon. It is the aim of this study to identify bacteria responsible for biogenic amine production in cold-smoked salmon. METHODS AND RESULTS: The present study identified spoilage microflora from cold-smoked salmon and determined biogenic amine production of single and co-cultures growing in cold-smoked salmon. Photobacterium phosphoreum was the only species that produced histamine when inoculated on sterile cold-smoked salmon. Production of putrescine was enhanced 10-15 times when cultures of Serratia liquefaciens or Hafnia alvei were grown with Carnobacterium divergens or Lactobacillus sakei subsp. carnosus. This phenomenon was explained by interspecies microbial metabolism of arginine, i.e., metabiosis. CONCLUSIONS: The amounts of biogenic amines produced by single and co-cultures corresponded to those observed during spoilage of naturally-contaminated cold-smoked salmon. Photobacterium phosphoreum and Lact. curvatus were identified as the specific spoilage organisms in cold-smoked salmon. SIGNIFICANCE AND IMPACT OF THE STUDY: Determination of the specific spoilage organism is needed before a model can be developed for shelf-life predictions of cold-smoked salmon.

Multiple compound quality index for cold-smoked salmon (Salmo salar) developed by multivariate regression of biogenic amines and pH.

Production of biogenic amines during chill storage of 12 lots of cold-smoked salmon was studied. These data allowed for a multiple compound quality index to be developed by multivariate regression (partial least square regression). The quality index was based on concentrations of cadaverine, histamine, putrescine, and tyramine and pH and showed good correlation with sensory assessments. Biogenic amines were indicators of spoilage rather than casual agents of spoilage off-flavors. Four different biogenic amine profiles were found at the time of spoilage in cold-smoked salmon. These were the results of differences in the spoilage microflora. Histamine was detected above regulatory limits but below toxic levels. Measurements of salt and dry matter for calculation of water phase salt could be substituted by rapid water activity measurements.

Carbon dioxide and nisin act synergistically on Listeria monocytogenes.

This paper examines the synergistic action of carbon dioxide and nisin on Listeria monocytogenes Scott A wild-type and nisin-resistant (Nis(r)) cells grown in broth at 4 degrees C. Carbon dioxide extended the lag phase and decreased the specific growth rate of both strains, but to a greater degree in the Nis(r) cells. Wild-type cells grown in 100% CO(2) were two to five times longer than cells grown in air. Nisin (2.5 microg/ml) did not decrease the viability of Nis(r) cells but for wild-type cells caused an immediate 2-log reduction of viability when they were grown in air and a 4-log reduction when they were grown in 100% CO(2). There was a quantifiable synergistic action between nisin and CO(2) in the wild-type strain. The MIC of nisin for the wild-type strain grown in the presence of 2.5 microg of nisin per ml increased from 3.1 to 12.5 microg/ml over 35 days, but this increase was markedly delayed for cultures in CO(2). This synergism between nisin and CO(2) was examined mechanistically by following the leakage of carboxyfluorescein (CF) from listerial liposomes. Carbon dioxide enhanced nisin-induced CF leakage, indicating that the synergistic action of CO(2) and nisin occurs at the cytoplasmic membrane. Liposomes made from cells grown in a CO(2) atmosphere were even more sensitive to nisin action. Liposomes made from cells grown at 4 degrees C were dramatically more nisin sensitive than were liposomes derived from cells grown at 30 degrees C. Cells grown in the presence of 100% CO(2) and those grown at 4 degrees C had a greater proportion of short-chain fatty acids. The synergistic action of nisin and CO(2) is consistent with a model where membrane fluidity plays a role in the efficiency of nisin action.

Control options for Listeria monocytogenes in seafoods.

At least three outbreaks of listeriosis associated with seafood have been reported. Listeria monocytogenes is widely distributed in the general environment including fresh water, coastal water and live fish from these areas. Contamination or recontamination of seafood may also take place during processing and low levels (< 100 cfu/g) of L. monocytogenes are frequently found on seafood including ready-to-eat (RTE) products. Apart from heat treatment, which is very effective, there are few options for eliminating L. monocytogenes from foods and equipment. It is essential therefore, that growth of L. monocytogenes in the final product be inhibited. The preventive measures include the formulation of a cleaning and sanitising program specifically designed at reducing the presence of L. monocytogenes in the factory environment, the safe elimination of L. monocytogenes from heat treated products and prevention of growth in RTE products within the normal shelf life and conditions stated on the label. If any sampling is required, the sampling plans suggested by the International Commission on Microbiological Specifications for Foods [Int. J. Food Microbiol., 22 (1994) 89-96] are useful.

Growth control of Listeria monocytogenes on cold-smoked salmon using a competitive lactic acid bacteria flora.

A Lactobacillus sake strain LKE5 and four strains of Carnobacterium piscicola were evaluated as biopreservation cultures to control the growth of Listeria monocytogenes on vacuum-packed, cold-smoked salmon stored at 5 degrees C. All five strains were antilisterial as live cultures in an agar diffusion assay. Cell-free supernatants of two strains of C. piscicola and L. sake LKE5 were also antilisterial because of the production of bacteriocins. The presence of high cell numbers of strains of C. piscicola had no influence on the sensory quality of cold-smoked salmon stored at 5 degrees C, but L. sake LKE5 caused strong sulfurous off-flavors and was rejected as a culture for biopreservation of cold-smoked salmon. A bacteriocin-producing strain of C. piscicola (A9b) initially caused a 7-day lag phase of L. monocytogenes, followed by a reduction in numbers of L. monocytogenes from 10(3) CFU/ml to below 10 CFU/ml after 32 days of incubation, coinciding with the detection of antilisterial compounds. The presence of a nonbacteriocin-producing strain of C. piscicola (A10a) prevented the growth of L. monocytogenes during the 32-day incubation. The growth of L. monocytogenes was strongly repressed on cold-smoked salmon in the presence of C. piscicola A9b and A 10a, respectively. The initial cell numbers of L. monocytogenes that were found on Oxford plates incubated at 25 degrees C reached low maximum cell counts of 10(4) and 2 x 10(3) after 14 and 20 days of storage in mixed culture with C. piscicola A9b and A10a.

Characterization of lactic acid bacteria isolated from a Thai low-salt fermented fish product and the role of garlic as substrate for fermentation.

Lactic acid bacteria (LAB) isolated from raw materials (fish, rice, garlic and banana leaves) and processed som-fak (a Thai low-salt fermented fish product) were characterized by API 50-CH and other phenotypic criteria. Lactococcus lactis subsp. lactis and Leuconostoc citreum were specifically associated with fish fillet and minced fish, Lactobacillus paracasei subsp. paracasei with boiled rice and Weisella confusa with garlic mix and banana leaves. In addition, Lactobacillus plantarum, Lactobacillus pentosus and Pediococcus pentosaceus were isolated from raw materials. A succession of aciduric, homofermentative lactobacillus species, dominated by Lb. plantarum/pentosus, was found during fermentation. In total, 9% of the strains fermented starch and 19% fermented garlic, the two main carbohydrate components in som-fak. The ability to ferment garlic was paralleled by a capacity to ferment inulin. An increased percentage of garlic fermenting strains was found during fermentation of som-fak, from 8% at day 1 to 40% at day 5. No starch fermenting strains were isolated during fermentation. Three mixed LAB cultures, composed of either starch fermenting Lc. lactis subsp. lactis and Lb. paracasei subsp. paracasei, or garlic fermenting Lb. plantarum and Pd. pentosaceus, or a combination of these strains were inoculated into laboratory prepared som-fak with or without garlic. In som-fak without garlic, pH was above 4.8 after three days, irrespective of addition of mixed LAB cultures. The starch fermenting LAB were unable to ferment som-fak and sensory spoilage occurred after three days. Fermentation with the combined mix of starch and garlic fermenting strains led to production of 2.5% acid and a decrease in pH to 4.5 in two days. The fermentation was slightly slower with the garlic fermenting strains alone. This is the first report describing the role of garlic as carbohydrate source for LAB in fermented fish products.

Prevalence and growth of Listeria monocytogenes in naturally contaminated seafood.

Listeria monocytogenes contamination of seafood varies with product category. The highest prevalence was found in cold-smoked fish (34-60%), while the lowest was found in heat-treated and cured seafood (4-12%). The prevalence of L. monocytogenes differed greatly in cold-smoked salmon between production sites, ranging from < 1.4% (nil out of 70 samples) to 100%. The prevalence at the individual production sites was reproducible at repeated sampling. The results indicate that it is possible to produce cold-smoked salmon with a low prevalence of L. monocytogenes. The organism showed moderate growth in naturally contaminated cold-smoked, and 'gravad', fish while the growth appeared faster in hot smoked fish. Thus L. monocytogenes is not under control in these products. Finally, the prevalence and growth of L. monocytogenes in naturally contaminated cold-smoked salmon are discussed in relation to controlling this risk.

Evaluation of the role of Carnobacterium piscicola in spoilage of vacuum- and modified-atmosphere-packed cold-smoked salmon stored at 5 degrees C.

The microflora on spoiled cold-smoked salmon often consists of a mixture of lactic acid bacteria (LAB) and Gram-negative bacteria. To elucidate the role of the different groups, a storage trial was carried out in which nisin and CO2 were used for the selective inhibition of the two bacterial groups. The shelf-life of vacuum-packed cold-smoked salmon, recorded by sensory evaluation, was four weeks at 5 degrees C and the microflora was composed of LAB (10(6)-10(7) cfu/g) with an associate Gram-negative flora in varying levels (10(5)-10(7) cfu/g). The addition of nisin and/or a CO2-atmosphere increased the shelf-life to five or six weeks and limited the level of LAB to about 10(4)-10(6), 10(3)-10(6) and 10(2)-10(4) cfu/g, respectively. CO2-atmosphere +/- nisin inhibited the growth of Gram-negative bacteria, whereas nisin had no effect on these in vacuum packages. The Gram-negative flora on vacuum-packed salmon was dominated by a Vibrio sp., resembling V. marinus, Enterobacteriaceae (Enterobacter agglomerans, Serratia liquefaciens and Rahnella aquatilis) and occasionally Aeromonas hydrophila. Irrespective of the addition of nisin and/or CO2-atmosphere, the LAB microflora was dominated by Carnobacterium piscicola, which was found to account for 87% of the 255 LAB isolates characterized. Whole-cell-protein patterns analysed by SDS-PAGE confirmed the Carnobacterium species identification. The spoilage potential of C. piscicola isolates was further studied by inoculation of approx. 10(6) cfu/g in cold-smoked salmon stored at 5 degrees C. The salmon did not spoil within 4 weeks of storage in vacuum- or CO2-atmosphere, and it is concluded that despite high levels (> 10(7) cfu/g) of C. piscicola, sensory rejection was caused by autolytic changes. This was supported by the development of soft texture and sour, rancid and bitter off-flavours at the point of spoilage, irrespective of the length of shelf-life and low or high total counts of LAB and Gram-negative bacteria.

Application of an iterative approach for development of a microbial model predicting the shelf-life of packed fish.

An iterative approach was used to develop a microbial model for shelf-life prediction of cod fillets packed in modified atmospheres. The effect of temperature (0-15 degrees C) and CO2 (0-100%) on growth of the specific spoilage organism, Photobacterium phosphoreum, was studied in packed cod and in liquid media. P. phosphoreum was a dominant part of the spoilage microflora of packed cod stored at the extremes of the range of conditions studied. The organism is therefore likely to be important for spoilage and the development of a microbial model within this domain seems relevant. A liquid medium was developed to provide growth kinetics of P. phosphoreum similar to those observed in packed cod. Using this medium, the effect of temperature and CO2 on the maximum specific growth rate of Photobacterium phosphoreum was determined by absorbance measurements and modelled by a square root equation and by a polynomial equation. Product validation studies were carried out during summer and winter using naturally contaminated packed cod fillets which were stored at constant and at changing temperatures. The shelf-life of the packed fillets was predicted on the basis of the initial numbers of P. phosphoreum, product temperature profiles and the level of CO2 in the modified atmosphere. The average deviations between shelf-life determined by sensory evaluation and shelf-life predicted by the square root equation and by the polynomial equation were 17% and 9%, respectively.

Inhibition of Listeria monocytogenes on cold-smoked salmon by nisin and carbon dioxide atmosphere.

The bacteriostatic and bacteriocidal effect of nisin in combination with carbon dioxide, NaCl and low temperature on the survival of Listeria monocytogenes was investigated in in vitro model studies and in trials with cold-smoked salmon. Addition of nisin caused various degrees of inhibition and sometimes death of L. monocytogenes in model experiments performed at 10 degrees C. The antilisterial effect of nisin was improved in the presence of 100% CO2 and increasing NaCl concentrations (0.5 to 5.0% w/v). Minimal bactericidal concentrations (MBC) of nisin varied from 30 to more than 500 IU/ml. The most pronounced effect of nisin was found when 10(2) cfu/ml was grown in media with 5.0% NaCl and incubated in CO2 atmosphere (MBC = 30 IU/ml). The bactericidal effect of nisin was reduced in air and vacuum, and did not increase systematically with increasing NaCl concentrations. In general, nisin concentration < or = 50 IU/ml resulted in the survival and growth of L. monocytogenes in all combinations with other preservatives (NaCl, CO2). Addition of nisin (500 or 1000 IU/g) to cold-smoked salmon inoculated with L. monocytogenes and stored at 5 degrees C delayed, but did not prevent growth of L. monocytogenes in vacuum-packs. Numbers of L. monocytogenes increased to 10(8) cfu/g in vacuum packed cold-smoked salmon in 8 days, whereas CO2 packing of cold-smoked salmon resulted in an 8-day lag phase of L. monocytogenes, with numbers eventually reaching 10(6) cfu/g in 27 days. Addition of nisin to CO2 packed cold-smoked salmon resulted in a 1 to 2 log reduction of L. monocytogenes followed by a lag phase of 8 and 20 days in salmon with 500 and 1000 IU nisin/g, respectively. The levels of L. monocytogenes remained below 10(3) cfu/g during 27 days of storage at both concentrations of nisin.

Microbiological spoilage of fish and fish products.

Spoilage of fresh and lightly preserved fish products is caused by microbial action. This paper reviews the current knowledge in terms of the microbiology of fish and fish products with particular emphasis on identification of specific spoilage bacteria and the qualitative and quantitative biochemical indicators of spoilage. Shewanella putrefaciens and Pseudomonas spp. are the specific spoilage bacteria of iced fresh fish regardless of the origin of the fish. Modified atmosphere stored marine fish from temperate waters are spoiled by the CO2 resistant Photobacterium phosphoreum whereas Gram-positive bacteria are likely spoilers of CO2 packed fish from fresh or tropical waters. Fish products with high salt contents may spoil due to growth of halophilic bacteria (salted fish) or growth of anaerobic bacteria and yeasts (barrel salted fish). Whilst the spoilage of fresh and highly salted fish is well understood, much less is known about spoilage of lightly preserved fish products. It is concluded that the spoilage is probably caused by lactic acid bacteria, certain psychotrophic Enterobacteriaceae and/or Photobacterium phosphoreum. However, more work is needed in this area.

Prevalence of vibrio cholerae and salmonella in a major shrimp production area in Thailand.

In 1992 and 1993, a 7 months study carried out in a major shrimp-producing area in Southern Thailand to study the prevalence of Vibrio cholerae and Salmonella. A total of 158 samples were examined including water, sediment, shrimp, pelleted feed, shrimp gut, and chicken manure. Salmonella was not recovered from any sample type studied. V. cholerae O1 was isolated from 2 (2%) and V. cholerae non-O1 was isolated from 35 (33%) of 107 samples examined. The occurrence of V. cholerae was not significantly influenced by water salinity, temperature, dissolved oxygen or pH. There was no correlation between fecal coliform counts and the prevalence of V. cholerae. The results indicate that V. cholerae non-O1 is ubiquitous in aquatic environments where shrimp culture is practised under a variety of environmental conditions. The public health significance of non-O1 V. cholerae in shrimp culture remains to be determined. V. cholerae O1 and Salmonella do not appear to constitute a hygienic problem even if chicken manure was used as fertilizer.

Application of ribotyping for differentiating Vibrio cholerae non-O1 isolated from shrimp farms in Thailand.

A collection of 143 Vibrio cholerae non-O1 strains isolated from shrimp farms in Thailand were characterized and grouped by ribotyping. Sixty-four ribotypes were distinguished following digestion of chromosomal DNA with the restriction enzyme BglI, and the reproducibility of the method was 100%. There was no correlation between specific ribotype distributions and the locations of the shrimp farms. Ribotype similarity was examined by cluster analysis, and two main groups with 10 and 54 ribotypes, respectively, were found. Correlation between ribotype and O-antigen expression was shown to exist among those isolates tested. Ribotyping appears to be a suitable method for differentiating environmental V. cholerae non-O1 strains, and comparison of ribotype patterns showed a high degree of genetic divergence within V. cholerae non-O1.

Production of histamine and tyramine by lactic acid bacteria isolated from vacuum-packed sugar-salted fish.

The incidence of histamine- or tyramine-producing lactic acid bacteria was examined in several products of vacuum-packed sugar-salted fish (salmon, halibut, mackerel). No histamine-producing isolates were observed, whereas the majority of tyramine-producing isolates were identified as Carnobacterium spp. These organisms were shown to be important members of the microbial flora during storage of vacuum-packed sugar-salted salmon at 5 degrees C. The amount of tyramine produced was reduced by lowering the temperature from 9 degrees C to 4 degrees C for all of five strains of carnobacteria or lactobacilli. The majority of tyramine was produced during the exponential growth phase for Carnobacterium piscicola N 5 and Lactobacillus viridescens N 69. The ability of these bacteria to produce tyramine may be used as an index of microbial quality/acceptability of stored vacuum-packed sugar-salted fish.

Heat resistance of Listeria monocytogenes in vacuum packaged pasteurized fish fillets.

The heat resistance of two strains of Listeria monocytogenes in sous-vide cooked fillets of cod and salmon was investigated. Fish sticks of 5 g were inoculated, vacuum-packed and heated at different combinations of time and temperature (58-80 degrees C). Time-temperature combinations allowing survival and time-temperature combinations at which the bacteria were destroyed, were used to determine D- and z-values. D-values were in the range of what has been published for other food products. D60-values were between 1.95 and 4.48 min depending on the strain and the fish. Both strains were one-four-times more heat resistant in salmon than in cod, showing the importance of the heating menstruum. This difference may be due to the higher fat content in salmon as compared to cod. Z-values were calculated to be 5.65 and 6.4 degrees C, respectively, for the two strains. The suitability of methods for heat resistance experiments and the survival of L. monocytogenes in sous-vide cooked fish fillets are discussed.

Spoilage and shelf-life of cod fillets packed in vacuum or modified atmospheres.

Microbial growth, sensory and chemical changes and composition of gas atmosphere were studied in vacuum packed (VP) and modified atmosphere packed (MAP) cod fillets stored at 0 degree C. Contrary to previous studies, coccobacilli and pleomorphic Gram-negative microorganisms (2-4 by 2-5 microns) and not Shewanella putrefaciens were found most likely to be the main spoilage organisms. These microorganisms, which may be Photobacterium phosphoreum, can explain the short shelf-life extension of VP and MAP fish products compared to meat products. It is suggested that they may inhibit the typical H2S-producing fish spoilage bacteria, S. putrefaciens, as the maximum concentration of H2S-producing bacteria found in MAP fish products is very low. Compared to VP, a shelf-life extension of 6-7 days was obtained with 48% CO2 in MAP. However, with pure CO2 the shelf life was only extended by 2-3 days. Poor texture and high drip loss indicated that the shelf life of these fillets was limited by chemical reactions and not only by microbial activity.