lntemcltional Joumal

ofFoodlI4icrobiology
International Journal of
ELSEVIE:R Fwd Microbiology 26 (1995) 319-333

Qualitative and quantitative characterization of spoilage

bacteria from packed fish
Paw Dalgaard *
Technological Laboratory, Ministry of Fisheries, Technical University, Bldg. 221, DK-2800 Lyngby,
Denmark

Received 20 January 1994; revised 25 August 1994; accepted 3 October 1994

Abstract

The large cells recently suggested to be responsible for spoilage of packed cod, have
been identified as Photobacterium phosphoreum. The spoilage activity of these cells, of
Shewunefr’u putrefaciens and of other microorganisms isolated form spoiled packed cod has
been studied. Both qualitative and quantitative tests were used for characterization of the
microbial spoilage activity. The importance of the different groups of microorganisms was
evaluated1 by comparison of microbial spoilage activity determined in model substrates and
in product experiments. The yield factor for production of trimethylamine (Y-,,,I and
the cell concentration determined at the time of off-odour detection were used as quantita-
tive mea:surements of microbial spoilage activity. On average cells of P. phosphoreum
produced 30 times more TMA than cells of S. putrefuciens, YTMA,cFu of the two organisms
were 10-.8.0 mg-N TMA/cfu and 10m9.’ mg-N ThIA/cfu, respectively. With these yield
factors the level of Th4A found in spoiled packed cod (30 mg-N TMA/lOOg) corresponds to
about 10” cfu/g of P. phosphoreum and to lo*-lo9 cfu/g of S. putrefaciens. 10’ cfu/g of P.
phosphoreuum were actually found in spoiled packed cod suggesting this organism could be
responsible for spoilage.
High cell concentrations of more than 10” cfu/g of S. putrefuciens were required for
production of detectable off-odours and is was concluded that this organism is without
importance for spoilage of packed cod.

Keywords: Yield factors of trimethylamine production; Spoilage activity; Spoilage potential;

Photobaderium phosphoreum; Shewanella putrefaciens; CO,

l Corresponding author. E-mail: pad@ffl.min.dk. Fax: + 4545884774. Tel.: + 4545883322.

0168-1605,/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved

SSDI 0168-1605(94)00137-5
320 P. Dalgaard /ht. J. Food Microbiology 26 (1995) 319-333

1. Introduction

Recently, TMA and ammonia-like off-odours and sour off-flavours were found
in spoiled vacuum packed (VP) and modified atmosphere packed (MAP) chilled
cod fillets. Opposed to aerobically stored whole cod, no sulphidy off-odours were
detected and the sensory characteristics were supported by chemical analyses
showing large amounts of TMA but very little H,S and CH,SH in the spoiled
packed fish (Dalgaard et al., 1993).
Shewanella putrefaciens is a well-known fish spoilage bacteria that produce
intensive off-odours, TMA and H,S and this organism was found responsible for
spoilage of both packed and unpacked cod. Only low levels of S. putrefaciens,
however, were found at sensory rejection of packed cod but it was suggested that
CO, may increase this organisms activity or cause specially active strains to be
selected during storage of packed fish (Levin, 1968; Jensen et al., 1980; Cann et al.,
1983; Gram et al., 1987; Jorgensen et al., 1988; Jorgensen and Huss, 1989).
In a recent publication large coccobacilli and pleomorphic cells (2-4 by 3-5
pm) were found as an important part of the microflora in packed cod and these
organisms were suggested to be responsible for spoilage (Dalgaard et al., 1993).
The purpose of the present study is to identify these large cells and to
determine their importance for spoilage of packed fish compared to the well
known fish spoilage bacteria. Qualitative and quantitative techniques were used to
determine the importance of the different groups of bacteria and to quantify the
effect of CO, on the microbial spoilage activity.

2. Material and methods

2.1. Bacteria

Strains were isolated at the time of sensory rejection from VP and MAP cod
fillets (Table 1; Dalgaard et al., 1993). Black and white colonies were isolated from
pour plates of Iron Agar (IA, 25”C, 3 days) and large cells were isolated as
glass-like colonies from spread plates (IA, 5°C 14 days). 10 strains of large cells
from the intestines of cod and two strains of Photobacterium phosphoreum (ATCC
11040, ATCC 35080) have also been studied. From each of the five atmospheres

Table 1
Numbers of bacterial strains isolated from VP and MAP cod fillets stored at 0°C
No. of strains isolated from fish stored with different CO, concentrations Total
O-5% CO, lo-20% CO, 25-35% CO, 40-60% CO, 90-100% CO* z;z;;ns

Large cells 8 7 12 12 11 50
Black colonies 24 25 25 25 25 149
White colonies 14 15 11 14 16 84
 P. Dalgaard /Int. J. Food Microbiology 26 (1995) 319-333 321

indicated in Table 1, one strain of large cells (P-O, P-15, P-30, P-50 and P-100) and
one strain from a black colony (S-O, S-15, S-30, S-50 and S-100) were selected for a
more detailed characterization.

2.2. Media

Luminous Medium (LM), Luminous Medium Broth (LMB), Basal Medium

(BM) and Basal Medium Agar (BMA) with purified agar (Sigma A-70491, were
prepared according to Baumann and Baumann (1981). Growth Medium Broth
(GMB) was prepared as previously described (Dalgaard et al., 1994). Iron Agar
was obtained from Oxoid (CM867) and Veal Infusion Broth (VIB) from Difco. The
medium of Wood and Baird (1943) (WB) was used in tests for TMA and NH,
production. Bacto-peptone, yeast extract (YE) and tryptone (all from Difco) and
D-glucose (Merck 8337100) were used in several tests.
Cooked fish muscle juice (FJ). Fillets of newly killed cod (Gadus morhua) were
cut in pieces, 500 ml of tap water were added per kg of fish fillet and the mixture
was boiled for 5 min. Juice was separated form solids in a strainer and residual
juice pressed out of the solids in a mechanical press. The two parts of juice were
mixed and boiled for 5 min, left 5 min at ambient temperature and filtered through
standard coffee filters. 0.10 M phosphate buffer, 0.056 M H,KPO, (Merck
48711000) and 0.044 M HK,PO, (Merck 51041000), were added and pH was
adjusted to 6,60 with NaOH/HCl. The juice was heated to 100°C for 30 min and
stored at 5°C. To recompense dilution and heat decomposition, 1.6 g of trimeth-
ylamine-N-oxide (TMAO, Fluka 92277), 40 mg of L-cysteine (Sigma C-7755) and
40 mg of L-methionine (Sigma M-9625) were added per liter of juice as filter
sterilized solutions.
Fish Juice Agar (FJA) were prepared from FJ by addition of 1,5% agar.
Steti’le muscle blocks (MB) were prepared by the method of Herbert et al. (1971)
from newly killed cod.

2.3. Identification of bacteria

Unless otherwise indicated, all tests with isolates from black and white colonies
were incubated at 25°C for 3 days and all tests with large cells were incubated at
10°C for 7 days. Media used for the large cells were made up with half strength
artificial sea water (ASW, MacLeod, 1968).
Gram-reaction was tested by the KOH-method (Gregersen, 1978), cytochrome
oxidase by Bactident@ Oxidase strips (Merck 13300.0001) and catalase by the 3%
H,O, method. Shape and mobility were tested by phase contrast microscopy of
strains grown in VIB (24 h, 25°C) or for large cells in LMB (48 h, 10°C>.
TMAG-reduction was tested by the method of Gram et al. (1987) and TMAO
negative strains were also tested by the method of Laycock and Regier (1971).
H,S-production was determined in stack inoculated test tubes with IA. Resistance
to Vibriostaticum (150 pg, o/129 discs, A/S ROSCO,Taastrup, Denmark) was
tested on IA supplemented with 0.5% NaCl and bioluminescence was detected in
a dark room from the emission of visible light (Baumann and Baumann, 1981).
322 P. Dalgaard / ht. .I. FoodMicrobiology
26 (1995) 319-333

Glucose metabolism. The O/F-medium was used to test oxidative/fermentative

mode of glucose metabolism (Hugh and Leifson, 1953). Anaerobic production of
gas from glucose was tested in the O/F-medium with and without addition of
glucose. The method of Lee et al. (1979) was used for the Voges-Proskauer test,
the medium contained 1.0% Bacto-peptone, 0.5% YE, 1.0% glucose, l/2 strength
ASW and 0.3% agar.
Amino acid metabolism. Increased pH after metabolism of arginine, lysine and
ornithine was tested in Falkow’s medium (Barrow and Feltham, 1993). Glucose
was omitted from the medium when used for isolates of large cells. Indole
production was tested in 1.0% Bacto-peptone by addition of Kovas reagents and
NH, production from peptone was tested in WB by addition of Nessler’s reagent.
Hydrolytic activity. Solidified media were streak inoculated and incubated at
10°C for 14 days (large cells) and 25°C for 5 days (black colonies). Degradation of
fish protein was tested on FJA. After incubation, the plates were flooded with 10
ml of 5% (V/V) acetic acid and colonies surronded by a clear zone were defined
as positive. Chitin and Tween-80 degradation were determined from clearing zones
in solid media (Barrow and Feltham, 1993).
Utilization of carbon and energy sources were tested as visible growth on BMA
supplemented with 0.01% YE and 0.2% D-mannitol (Sigma M-4125) or 0.1%
P-hydroxy-butyrate (Sigma H-9256) or 0.2% D-alanine (sigma A-7377) or 0.2%
pyruvate (Sigma P-5280).
PHB. Accumulation of poly-&hydroxybutyrate was tested for large cells grown
48h at 10°C in BM supplemented with 0.2% D-glucose and 0.05% YE. Direct
phase contrast microscopy and Sudan Black B (Sigma S-2380) staining were used
for detection of PHB (Baumann and Baumann, 1981)
Flagella and %G + C. The strains P-O, P-15, P-30, P-50, P-100, S-O, S-15, S-30,
S-50 and S-100 were tested for %G + C by the method of Sandstedt et al. (19831
and the width and thereby the sheathing of their flagella was tested after growth in
LMB (lO°C, 48 h) by electron microscopy (Allen and Baumann, 1971).
Whole cell fatty acids were analyzed for P-O, P-100, S-O and S-100 grown at 10°C
in Erlenmeyer flasks with GMB. A modification of the method of Moss et al.
(1988) were used for extraction of lipids. The fatty acids were identified by gas
chromatography (GC) and GC-mass spectrometry (Nichols et al., 1992).
The effect of temperature and NaCl on growth. Precultures were grown 2 days in
LMB at 10°C and 10 ~1 were inoculated on LM plates incubated aerobically at 0,
20, 25, 30 and 37°C. Visible growth was detected regularly of up to 4 weeks. The
effect of NaCl was tested in medium with 1.0% tryptone, 0.5% YE and NaCl (0,
30,60 or 90 g/l). pH was adjusted to 7.5. Test tubes were incubated with 10 ~1 of
pre-culture and visible turbidity was detected for up to 4 weeks.
2.4. Qualitative spoilage potential and microbial production of sulfides
50 strains of large cells, 49 strains from black colonies and 71 strains from white
colonies were inoculated in lo-15 ml FJ and incubated at 0°C for 3 to 4 weeks in
100% N, in anaerobic jars (Oxoid). Spoilage potential was determined as the
ability of isolates to produce of off-odours. 1.0 ml FJ was transferred to a petri dish
 P. Dalgaard / ht. J. Food Microbiology 26 (1995) 319-333 323

and after 15 min. at ambient temperature off-odours were evaluated by five

panelists and categorized in three groups: I corresponding to no off-odours, II for
weak off-odours or III for strong off-odours. Off-odour production was deemed
positive when the average of notes from the five panelists were higher than II. H,S
and CEjl,SH production were determined for 33 large cells and 23 strains from
black colonies. Strains were grown as described for spoilage potential. 5 ml of FJ
were transferred to a 50 ml headspace flask (Mikro-Laboratoriet, Hojbjerg, Den-
mark) and H,S and CH,SH in the headspace gas were detected by gaschromatog-
raphy as previously described (Dalgaard et al., 1993). Responces were recorded on
a Perki:n-Elmer LIC-100 integrator (Perkin-Elmer, Norwalk, Connecticut, USA),
the same attenuation and range were used for all strains analysed.

2.5. Quantitative tests

The yield factor for TMA production. YTMA,CFU was calculated from cell concen-
trations (&/ml) and YrMA,oW was determined from dry weight of biomass by
Eq. 1.
Final - Initial TMA cone .
y’m/cFu Or YTMA/DW = Final - Initial cell cont. (1)
35 strains of large cells of which five were from cod intestines, 20 strains from
black colonies and 15 strains from white colonies were grown in FJ as described for
spoilage potential.
Effect Of co2 On yTMA jcFul YTMA/DW and on the spoilage activity of S.
putrefaciens. S-O and S-100 were grown in FJ and a mixture (MIX-S) of S-O, S-15,
S-30, S-50 and S-100 were grown in FJ and in sterile muscle blocks (MB). Spoilage
activity was determined as the cell concentration (cfu/ml) found when microbial
off-odours were detected. The cell concentration was calculated as the average of
the last cell concentration where off-odour was not determined and the first cell
concentration where off-odours were detected.
200 ml FJ was placed in Erlenmeyer flasks and incubated in anaerobic jars
(Oxoid). The jars were evacuated to lo4 Pa, filled with N,, evacuated again to lo4
Pa and filled with one of five gas mixture with 0, 25, 50, 75 or 100% CO,
(remaining gas being N,). The gas composition in the jars was initially tested by gas
chromatography (Nielsen et al., 1989). Purity of gases was 99.9% for CO, and
99.99% for N, (Hydrogas Danmark A/S, Fredericia, Denmark). After 2 days at
0°C pH[ of FJ was readjusted to 6.6 and the juice was inoculated with 102-lo4
cells/ml. Inoculum was pre-cultured anaerobically at 0°C in FJ. MB of about 20 g
were placed in Petri dishes and incubated in 0, 50 and 100% CO, as described for
FJ. After 2 days each MB was inoculated with 102-lo4 cfu/g of MIX-S. At regular
intervals samples of FJ and MB were removed for analyses of growth, TMA, pH
and off-odours. In these experiments off-odours were evaluated by two panelists.

2.6. Analyses
Growth was measured by viable counts (VC) and by absorbance measurements
(Abs). For determination of VC, FJ was appropriately diluted in physiological
324 P. Dalgaard /ht. J. Food Microbiology 26 (1995) 319-333

saline (0.9% NaCI) with 0.1% Bacto-peptone (PSI. 5-10 g of MB was diluted
lo-fold in PS, homogenized 30 s in a Colworth Stomacher 400 (Seward Laboratory,
London, UK). VC of bacteria from black and white colonies were determined in
IA (25”C, 3 days). Strains of large cells were ennumerated on spread plates of IA
supplemented with 0.5% NaCl (lOC, 7 days).
In experiments with FJ, absorbance was measured at 600 nm in 4 ml disposable
cuvettes (Kebo Lab, Herlev, Denmark) with a Ciba-Coming calorimeter 254
(Tillquist, Copenhagen, Denmark). Deionized water was used as blank and for
dilution of samples with Abs > 0.4. The relation between Abs and the dry weight
of cells was determined for MIX-S and for a mixture of strains of large cells grown
at 0°C in FJ in a Braun Biostat @L fermentor (Meda, Herlev, Denmark). Cells from
250 ml of FJ were harvested by centrifugation (4000 X g, 20 min, O’C) in a Sorvall
RC-SB centrifuge (Dupony company, Wilmington, Delaware, USA), the super-
natant decanted, the cell pellets washed, cells re-harvested, supernatant decanted
again and cells dried (lOS’C, 24 h) whereafter dry weight was determined.
TMA was determined in duplicate in about 5 ml of FJ and for about 5 g of MB
by a modified Conway micro-diffusion method (Conway and Byrne, 1933). pH was
measured by a Autocal pH meter (Radiometer, Copenhagen, Denmark).

2. I. Product experiments

Apparent values of YrMA,cFU for growth of S. putrefuciens in packed cod were

calculated from the previously described product experiments (Dalgaard et al.,
1993). The apparent values of YrMA,cFU were calculated by Eq. 1 from the
concentration of TMA and from the levels of large cells and from the levels of
black colonies found in spoiled packed cod stored with different CO, concentra-
tions. The TMA production of the two groups of microorganisms was evaluated by
comparison of YrMA,cFU values determined in model substrates and the apparent
values calculated from the product experiments.

3. Results

3.1. Identification of bacteria

Table 2 show the phenetic characteristics of the three groups of microorganisms

isolated from VP and MAP cod fillets. Results for the large cells isolated from
packed cod were similar to results for large cells from the intestines of cod and
also similar to results for ATCC 11040 and ATCC 35080. All strains of large cells
therefore tentatively were identified as Photobacterium phosphoreum. More than
99% of strains isolated as black colonies were tentatively identified as Shewanella
putrefuciens with only one fermentative strain belonging to Vibrionaceae (Dainty
et al., 1979). All strains selected for detailed characterization had unsheathed
flagella. The width of the flagellum was 13.8 f 1.5 nm for P. phosphoreum and
13.6 f 3.1 for 5. putrefaciens. Allen and Baumann (1971) indicate unsheathed and
 P. Dalgaard /ht. J. Food Microbiology 26 (1995) 319-333 325

Table 2
Qualitative characteristics of bacteria isolated from vacuum packed and MAP cod fillets stored at WC
Tests Number of strains positive (number of strains tested)
Large cells Black colonies White colonies
Gram-positive 0 (50)
Shape 50 coccobacilli (50) 149 rods (149) 81 rods, 3 cocci (84)
Motile 24 (50) 149 (149) 60 (84)
Qxidase 0 (50) 149 (149) 77 (84)
Catalase positive 50 (50) 149 (149) 77 (84)
Acid from glucose
fermentative 50 (50) l(149) 32 (84)
oxidative 0 (50) lU49) 29 (84)
none 0 (50) 147 (149) 23 (84)
Gas from glucose 50 (50) 0 (5) _a
O/129 50 (50) 0 (5) -
TMAO-reduction 50 (50) 149 (149) 28 (84)
Has-production 17 (50) 149 (149) 28 (84)
NH,-production 13 (50) 143 (149) 32 (84)
Proteolysis 0 (50) 142 (149) 54(84)
Arginine 50 (50) 0 (5) -
Lysine 22 (50) 0 (5)
Ornithine 5 (50) 5 (5)
Indole 0 (50) 0 (5)
Chitin 21(50) 3 (5)
Tween-80 22 (50) 5 (5) _
DNase 50 (50) 5 (5) -
VP 50 (50) 0 (5) -
NO; 50 (50) 5 (5)
Luminescence 16 (50) 0 (5)
Growth at 0°C 50 (50) 5 (5)
20°C 49 (50) 5 (5)
25°C 23 (50) 5 (5)
30°C 0 (50) 4 (5) -
37°C - 0 (5) -
Growth in 0% NaCl 0 (50) 5 (5) -
3% NaCl 50 (50) 5 (5) -
6% N.&l 37 (50) 5 (5) _
9% NaCl 0 (50) 0 (5) _
PHB 10 (10) - -
D-mannitol 0 (50) - -
P-hydroxybutyrate 0 (50) - -
D-alanine 0 (50) - -
Pyruvate. 0 (50) - -
Production of off- 10 (50) 32 (49) 24 (71)
odours in fish juice
Characteristics of TMA and am- putrid, TMA and putrid, fruity and
off-odours monia like ammonia like ammonia like

a Not determined.
326 P. Dalgaard /ht. .I. Food Microbiology 26 (1995) 319-333

Table 3
Major fatty acids in strains of isolated bacteria
Fatty acids Average percentage of fatty acids
large cells Isolates form black colonies
(P-O and P-100) (S-O and S-100)
12:o 3.6kO.6 5.8+2.3
i13:l a - 6.0 + 0.2
14:o 3.1+0.5 6.7 f 1.9
i15:O - 7.1*0.3
16:l 55.3 f 2.5 46.5 f 0.9
16:O 13.0*0.4 13.0+0.1
18:l 16.6k5.4 5.8fO.S
Polyunsaturaed acids 0.7+0.0 2.0+ 1.8
Gyclopropane acids - TRb
Hydroxy acids _ TR

a i, iso brached acid.

b TR, detected in trace levels

sheathed flagella to be 14-16 nm and 24-30 nm, respectively. The average

%G + C of the large cells was 41.6 f 0.4 and 46.3 + 0.9 for S. putrefaciens. The
three dominating fatty acids (FA) of the large cells (PO and PlOO) made up 84% of
the total fatty acids. SO and SlOO contained a large number of FA including
branched and polyunsaturated acids, for these strains the three major acids made
up only 58% of total FA (Table 3).
Based on biochemical test and bioluminescence, the 50 strains of Z? phospho-
reum isolated form packed cod were divided in six subgroups (Fig. 1).
The 84 strains isolated from white colonies were a heterogenous group, tenta-
tively identified as Pseudomonas spp. (39%), Vibrionaceae (21%), Shewanella
putrefaciens (19%), MoraxeIla spp. (ll%), Lactic acid bacteria (7%) and Psy-
chrobacter spp. (2%) (Dainty et al., 1979).

pFq F p,
luminous non-luminous luminous non-luminous luminous non-luminous
I I I I I I
22% 20% 10% 40% 2x 6%
(111 1101 151 (20) 11) 131
Fig. 1. Groups of 50 strains of Phorobacterium phosphoreum based on phenotypic characteristics.
Numbers in brackets indicate the number of strains in each group.
 P. Dalgaard / Int. J. Food Microbiology 26 (199s) 319-333 321

Table 4
Microbial1 production of volatile sulfur compounds
Sulfides Percentage of strains producing large amounts of sulfur compounds a
Shewanella putrefaciens b Photobacterium phosphoreum ’
H2S 56% 0%
CH,SH 78% 0%

’ Peak areas higher than 1 X 107.

b 23 strains tested.
’ 33 strains tested.

3.2. Spoilage potential and microbial production of sulfides

Some strains of P. phosphoreum produced small amounts of H,S and CH,SH

but most strains of S. putrefaciens produced large amounts of both sulfides (Table
4).

-7

-7.5
‘3
‘t
2 -9.0
f
f -0.5

!E -9.0
zl
s
-9.5

-10

0.4
o-5 1 O-20 25-35 40-60 90-100
CO* level of products where strains were isolated

Fig. 2. Average values and standard deviation of the yield factors for trimethylamine (TMA) production
for Photobacterium phosphoreum (0) and for Shewanella putrefaciens (0). Strains were isolated from
cod fillets stored with different levels of CO,.
328 P. Dalgaard /ht. J. Food Microbiology 26 (1995) 319-333

The percentages of off-odour producing strains of P. phosphoreum, S. putrefa-

ciens and of the strains from white colonies were 20, 65 and 34%, respectively
(Table 2).
3.3. Quantitative tests
The cooked fish juice was clear and growth of bacteria could be measured by
absorbance measurements. For both P. phosphoreum and for S. putrefaciens 1.00
Abs unit corresponded to 2.1 &-0.1 g dry weight per liter FJ (n = 7).
Yield factor for TM2 production. Fig. 2 show Log(YrMA,cFU) and YTMA,nW
determined for strains of P. phosphoreum and S. putrefaciens isolated for cod
fillets stored under different concentrations of CO,. The averages of
I-o~YTMA,CF”) were - 8.0 + 0.3 (SD) and - 9.5 _t 0.2 (SD), respectively. The large
values of YrMA,cFU for P.photobacterium, about 30 times higher than for S.
putrefaciens, were due to the large size of these cells as YrMA,nv., values were
similar for the two species (Fig. 2). No specially active strains of S. putrefaciens
were found i cod fillets stored with high CO, concentrations (Fig. 2). For TMAO
reducing strains isolated from white colonies Log(Y,,,,,,,) values were low,
about - 9.8.
Effect of CO, on LA /CFU and on the spoilage activity of S. putrefaciens.
Apparent values of YTMA,cFo calculated from storage trials with packed cod were
from x 10 to > x lo3 higher than values of YTMA,cF,, determined for strains of S.
putrefaciens grown in model system experiments. YrMA,cFu values for S-O, S-100
and MIX-S were similar and values determined for MIX-S grown in FJ or in MB
were also similar (Fig. 3). The average value of Log(Ym,,,,,) in these experi-
ments was -9.5 * 0.3.

-11 1 I I I I
0 20 40 60 60 100

% COP in experiments

Fig. 3. Effect of CO, on the TMA production for Shewanella putrefaciens grown in model systems (open
symbols) and apparent TMA production of the organism in VP and MAP cod fillets (m 1. Fish juice (FJ)
and sterile muscle blocks (MB) were used as model substrates. Individual strains (S-O and S-100) and a
mixture of five strains (MIX-S) were studied. 0, (S-O in FJ); A, 6100 in FJ); v , (MIX-S in FJ); 0,
(MIX-S in MB).
 P. Dalgaard /Int. J. Food Microbiology 26 (1995) 319-333 329

Table 5
Effect of CO, on the spoilage activity of Shewanella putrefaciens
Strains and model Cell concentrations at the time of off-odour detection
substrates (cfu/ml or cfu/g)
0% co, 25% co* 50% co, 75% co, 100% co,
S-O grown in FJ a 3x109 1x109 2x 109 6X108 1.ox1os
S-100 grown in FJ 2x109 2x109 5x10s nd ’ nd ’
MIX-S gronw in FJ 7x10s 2x109 5x10s 1 x 109 5x10s
MIX-S grown in MB b 3x109 - 7x10s - <3X109d

a FJ, cooked fish juice.

b MB, Sterile muscle blocks.
’ Off-odours not determined at the end of the experiments after 80 days with 75% CO, and after 150
days with 100% CO,.
d Cell cencentration where off-odours were detected. The last sampling with no off-odours had a much
lower cell concentration and the average cell concentration has not been calculated.

No off-odours were produced by S-O, S-100 and MIX-S if cell concentrations

were below lo8 cfu/ml or cfu/g (Table 5). The detected off-odours were TMA-like,
putrid and musty with S-100 producing less putrid odours than S-O and MIX-S.

4. Discussion

The large cells recently found in levels of 10’ cfu/g in spoiled packed cod were
identified as Photobacterium phosphoreum. The generally recognized fish spoilage
bacteria Shewunellu putrefuciens has been shown to be of no importance for
spoilage of packed cod stored in ice and P. phosphoreum seems most likely as the
specific spoilage organisms @SO), limiting the shelf life of these products.
The large cells belonged to Vibrionaceae (Gram negative, fermentative, o/129
sensitive and required NaCl for growth). The isolated strains were oxidase nega-
tive, had unsheathed flagellum, and did not utilize &hydroxybutyrate or D-man-
nitol, indicating the organisms to be Photobacterium spp. Production of gas from
glucose, growth at 0°C but not at 30°C the inability to utilize pyruvate and a
%G + C of 41.6 + 0.4 suggest the large cells to be Photobacterium phosphoreum
(Baumann and Baumann, 1981; Baumann and Baumann, 1984; Farmer III and
Hickman-Brenner, 1991). The major fatty acids of P-O and P-100 were typical for
Photobacterium but cyclopropane acids and hydroxy acids have been found in
strains of this genus (Lambert et al., 1983; Ratledge and Wilkinson, 1988).
P. phosphoreum is suggested as SSO because the characteristics of the isolated
large cells correspond qualitatively and quantitatively to the characteristics found
for packed cod (Dalgaard et al., 1993). With a YrMA,cFU value of 10s mg-N
TMA/cfu the 30 mg-N TMA/lOOg found at rejection of packed cod corresponded
to 3 X 1.0’ large cells/g and this level was actually found in the spoiled products.
P. phosphoreum produce only small amounts of H,S and CH,SH corresponding to
similar low level of these sulfides in spoiled packed cod and also the spoiled
330 P. Dalgaard /ht. J. FoodMicrobiology
26 (1995)319-333

products had no sulphidy or putrid off-odours. That 17 out of 50 strains of P.

phosphoreum produced H,S in IA may be due to reduction of the thiosulphate
found in IA. P. phosphoreum do not produce intensive off-odours and this may
explain why packed cod is not found sensorially unacceptable before high levels of
TMA has been produced. In packed cod 0.1 pmol hypoxanthine were produced
per pmol of TMA and a similar yield factor was found for P. phosphoreum grown
in FJ (results not shown). Finally, TMA and off-odour production in packed cod
are caused by microbial activity and no other microorganisms was detected in
concentration sufficient for these organisms to cause spoilage (see below).
Opposed to the present study, Shewan (1971) stated that luminous bacteria
were of little real importance to the practical fish technologist and Abgrall and
Cleret (1990) found I? phosphoreum (10’ cfu/g) to be without direct importance
for spoilage of packed whiting. P. phosphoreum has been detected in high levels in
spoiled cod (van Spreekens, 1974, van Spreekens, 1977). The importance of this
organism in relation to spoilage, however, has not previously been documented. It
is interesting to see that the percentage of non-luminous strains and the distribu-
tion of strains into subgroups found in the present study (Fig. 1.) were remarkably
similar to the results reported by van Spreekens.
As in other studies with spoiled fish (Levis, 1968; Gram et al., 1987), black
colonies isolated from pour plates Iron Agar were identified as S. putrefaciens.
Some strains isolated as white colonies, however, were also identified as S.
putrefaciens, indicating that not all H,S-producing organisms form black colonies
in IA. In most studies levels of lo’-10’ cfu/g of S. putrefaciens were found in
spoiled packed cod. More than lo8 cfu/g of this organism are required for
production of off-odours and the level of TMA found in spoiled packed cod
correspond to 108-10’ cfu/g of 5. putrefaciens. The hypothesis of Jorgensen et al.
(1988) that CO, increased the spoilage activity of S. putrefaciens has not been
confirmed (Fig. 2 and Fig. 3). S. putrefaciens has a high spoilage potential (produce
intensive off-odours) but this organism has a low spoilage activity and high cell
concentrations are required to cause spoilage. These quantitative results indicate
that S. putrefaciens can not possibly be responsible for spoilage of packed cod.
The off-odour producing bacteria from white colonies had a low spoilage
activity and taken into account their concentration in the spoiled packed fish
(105-lo6 cfu/g, Dalgaard et al., 1993) they seems to be of no importance for
spoilage.
The present study have shown that organisms with a common qualitative activity
(TMAO reduction) can have markedly different quantitative activities. Also, classi-
cal plate count methods may not detect all groups of microorganisms on a food
products. Consequently, techniques used for pointing out the group(s) of microor-
ganisms responsible for spoilage must be either highly specific or based on both
qualitative and quantitative tests. For meat and poultry, chemical spoilage profiles
were used to point out spoilage bacteria (Edwards et al., 1987; Edwards and
Dainty, 1987; Dainty et al., 1989; Viehweg et al., 1989a, Viehweg et al., 1989bI.
The technique mainly used to point out fish spoilage bacteria have been the
determination of spoilage potential, the ability of isolated strains to produce
 P. Dalgaard /ht. J. Food Microbiology 26 (1995) 319-333 331

off-odours. With this technique false positive results, however, can be obtained
when high levels of an isolate produce intensive off-odours but the organisms only
was found in low levels in the spoiled product.
Yield factors are well established for simple quantitative characterization of
microbial metabolism (Monod, 1942; Nagai, 1979) and determination of yield
factors for selected metabolite like TMA, histamine or others and possibly in
combination with chemical spoilage profiles seems a useful technique for evalua-
tion of the importance of groups of microorganisms during fish spoilage.

Acknowledgements

I should like to thank Jo-Anne E. Cavanagh, ACAM-Antarctic CRC, University

of Tasmania, Australia for carrying the lipid analyses and I thank Ingela Tjernberg,
Malmij General Hospital, Sweden who determined the %G + C. From the techno-
logical laboratory Hans Henrik Huss and Lone Gram gave valuable comments to
the manuscript and Anni Jensen have provided skillful technical assistance, I am
most grateful for their contributions.

References

Abgrall, 13. and Cleret, J.J. (1990) Evaluation of Petrifilm TM SM for the enumeration of the aerobic
flora of fish. 3. Food Prot. 3, 213-216.
Allen, R.D. and Baumann, P. (1971) Structure and arrangenent of flagella in species of the genus
Benecicea and Photobacterium fisheri. .I. Bacterial. 107, 295302.
Barrow, G.I. and Feltham, R.K.A. (1993). Cowan and Steel’s Manual for the Identification of Medical
Bacteria. Cambridge University Press.
Baumann, P. and Baumann, L. (1981) The marine gram-negative eubacteria: Genua Photobacterium,
Benecicea, Alteromonas, Pseudomonas, and Alcaligenes. In: M.P. Starr, H. Stolp, H.G. Triiper, A.
Balows and H.G. Schlegel (editors), The Prokaryotes. Springer-Verlag, Berlin, pp. 1302-1330.
Baumann, P. and Baumann, L. (1984) Genus II. Photobacterium Beijerink 1889, 402&. In: N.R. Krieg
and J.G. Holt (editors), Bergey’s Manual of Systematic Bacteriology. Williams and Wilkins, Vol 1,
Baltimore, pp. 539-545.
Cann, D.C., Smith, G.L. and Houston, N.C. (1983) Further studies on marine fish storage under
modified atmosphere packaging. Tony Research Station, Ministry of Agriculture, Fisheries and
Food, Aberdeen, Scotland.
Conway, E.J. and Byrne, A. (1933) An absorbance apparatus for the micro-determination of certain
volatile substances. I. The micro-determination of ammonia. Biochem. J. 27, 419-429.
Dalgaard, P., Gram, L. and Huss, H.H. (1993) Spoilage and shelf-life of cod fillets packed in vacuum or
modified atmospheres. Int. J. Food Microbial. 19, 283-294.
Dalgaard. P., Ross, T., Kampermann, L., Neumeyer, K. and McMeekin, T.A. (1994). Estimation of
bacterial growth rates from turbidimetric and viable count data. Int. J. Foof Microbial. 23, 391-404.
Dainty, R..H., Shaw, B.G. and Charmaigne, D.H. (1979) The spoilage of vacuum-packed beef by cold
tolerant bacteria. In: A.D. Russell and R. Fuller (editors), Cold Tolerant Microbes in Spoilage and
the Environment. Academic Press, London, pp. 83-100.
Dainty, R.H., Edwards, R.A., Hibbard, CM. and Marnewick, J.J. (1989) Volatile compounds associated
with microbial growth on normal and high pH beef stored at chill temperatures. J. Appl. Bact. 66,
281-2.39.
332 P. Dalgaard /Int. .I. Food Microbiology 26 (1995) 319-333

Edwards, R.A. and Dainty, R.H. (1987) Volatile compounds associated with the spoilage of normal and
high pH vacuum-packed pork. .I. Sci. Food Agric. 38, 57-66.
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 tempera-
tures. J. Appl. Bact. 62, 403-412.
Farmer III, J.J. and Hickman-Brenner, F.W. (1991) The genera vibrio and Photobacterium. In: A.
Balows, H.G. Triiper, M. Dworkin, W. Harder and K-H. Schleifer (editors), The Prokaryotes. A
Handbook on the Biology of Bacteria: Ecophysiology, Isolation, Identification, Application.
Springer-Verlag, New York, pp. 2952-3011.
Gram, L., Trolle, G. and Huss, H.H. (1987) Detection of specific spoilage bacteria from fish stored at
low (0°C) and high (20°C) temperatures. Int. J. Food Microbial. 4, 65-72.
Gregersen, T. (1978) Rapid method for distinction of Gram-negative from Gram-positive bacteria. Eur.
J. Appl. Microbial. 5, 123-127.
Herbert, R.A., Hendrie, MS., Gibson, D.M. and Shewan, J.M. (1971) Bacteria active in the spoilage of
certain sea foods. J. Appl. Bact. 34, 41-50.
Hugh, R. and Leifson, E. (1953) The taxonomic significance of fermentative versus oxidative Gram-
negative bacteria. J. Bacterial. 66, 24-26.
Jensen, M.H., Petersen, A., Rage, E.H., and Jepsen, A. (1980) Storage of chilled cod under vacuum and
at various concentrations of carbon dioxide. In: J.J Connell (editor), Advances in Fish Science and
Technology. Fishing News Books Ltd., Farnham, Surrey, England, pp. 294-297.
Jorgensen, B.R. and Huss, H.H. (1989) Growth and activity of Shewanella putrefaciens isolated from
spoiling fish. Int. J. Food Microbial. 9, 51-62.
Jorgensen, B.R., Gibson, D.M. and Huss, H.H. (1988) Microbiological quality and shelf life prediction
of chilled fish. Int. J. Food Microbial. 6, 295-307.
Lambert, M.A., Hickman-Brenner, F.W., Farmer, J.J. and Moss, C.W. (1983) Differentiation of
Vibrionaceae species by their cellular fatty acid composition. Int. J. Syst. Bacterial. 33, 777-792.
Laycock, R.A. and Regier, L.W. (1971) Trimethylamine-producing bacteria on Haddock (Melanogrm-
mus aeglefinus) fillets during refrigerated storage. J. Fish. Res. Board Can. 28, 305-309.
Lee, J.V., Hendrie, M.S. and Shewan, J.M. (1979) Identification of Aeromonads, vibrio and related
organisms. In: F.A. Skinner and D.W. Lovelock (editors), Identification Methods for Microbiolo-
gists. Academic Press, London, pp. 151-166.
Levin, R.E. (1968) Detection and incidence of specific species of spoilage bacteria on fish. I Methodol-
ogy. Appl. Microbial. 16, 1734-1737.
MacLeod, R.A. (1968) On the role of inorganic ions in the physiology of marine bacteria. Adv.
Microbial. Sea 1, 95-126.
Monod, J. (1942) Rescherches sur la Croissance Bacttriennes. Herman, Paris.
Moss, C.W., Wallace, P.L., Hollis, D.G. and Weaver, R.E. (19881 Cultural and chmical characteristics
of CDC EO-2, M-5 amd M-6, Moraxella (Moraxefla) species, Oligella urethralis, Acinefobacter
species, and Psychrobacter immobilis. J. Clin. Microbial. 26, 484-492.
Nagai, S. (1979) Mass and energy balances for microbial growth kinetics. Adv. Biochem. Eng. 2, 49-83.
Nichols, D.N., Nichols, P.D. and McMeekin, T.A. (1992) Anaerobic production of polyunsaturated fatty
acids by Shewanella putrefaciens strain ACAM 342. FEMS Microbial. Lett. 98, 117-122.
Nielsen, P.V., Beuchat, L.R. and Frisvad, J.C. (1989) Influence of atmospheric oxygen content on
growth and Fumitremogin production by a heat-resistant mold, Neosartorya fisheri. J. Food Sci. 54,
679-685.
Ratledge, C. and Wilkinson, S.G. (19881 Microbial Lipids. Academic Press, London.
Sandstedt, K., Ursing, J. and Walder, M. (1983) Thermotolerant Campylobacter with no or weak
catalase activity isolated from dogs. Curr. Microbial. 8, 209-213.
Shewan, J.M. (1971) The microbiology of fish products a -A progress report. J. Appl. Bact. 34,
299-315.
van Spreekens, K.J.A. (1974) The suitability of a modification of Long and Hammer’s medium for the
enumeration of more fastidious bacteria from fresh fishery products. Antonie v. Leeuw. 25,
213-219.
 P. Dalgaard /ht. J. Food Microbiology 26 (1995) 319-333 333

van Spreekens, K.J.A. (1977) Characterization of some fish and shrimp spoiling bacteria. Antonie v.
Leeuw. 43, 283-303.
Viehweg, S.H., Schmitt, R.E. and Schmidt-Lorenz, W. (1989a) Microbial spoilage of refrigerated fresh
boilers. Part VI. Identification of the volatile compounds produced during microbial spoilage of
chicken carcasses. Lebensm. Wiss. Technol 22, 346-355.
Viehweg, S.H., Schmitt, R.E. and Schmitt-Lorenz, W. (1989b) Microbial spoilage of refrigerated fresh
broilers. Part VII. Production of odours from poultry skin by bacterial isolates. Lebensm. Wiss.
Technol22, 356-367.
Wood, A.J. and Baird, A. (1943) Reduction of trimethylamine oxide by bacteria. J. Fish. Res. Board
Can. 6, 194-201.