Limited knowledge is currently available on the influence of fish thawing and subsequent storage conditions on bacterial growth kinetics, succession, and diversity alongside the production of biogenic amines. This study aimed to address these factors during the thawing and subsequent storage of mackerel. Thawing was either done fast in 18°C water for 2 h or slowly at 30°C overnight. Subsequent storage was at 30°C (ambient) for 36 h and 2 to 5°C (refrigerated) for 12 days. The cultivation methods used were total viable counts, hydrogen sulfide–producing bacteria, and Pseudomonas. Maximum growth rate, population density, and lag time were fitted on the counts using the Baranyi model. The bacterial diversity and succession were based on sequencing of 16S rRNA amplicons, and biogenic amines were quantified on high-pressure liquid chromatography–UV. The results show that lag time of hydrogen sulfide–producing bacteria was significantly affected by both thawing methods, and further, the interaction between thawing and storage significantly affected the maximum growth rate of these bacteria. However, the maximum growth rate of Pseudomonas was higher during refrigerated storage compared with storage at ambient temperature. Total viable counts showed longer lag time and reduced growth rate under refrigerated storage. Higher bacterial diversity was correlated to slow thawing and storage at ambient temperature compared with slow thawing and refrigerated storage. Overall, Acinetobacter and Psychrobacter genera were the dominant bacterial populations. The amine levels were low and could not be differentiated along the thawing and storage approaches, despite a clear increase in bacterial load, succession, and diversity. This corresponded well with the low abundance of biogenic amine–producing bacteria, with the exception of the genus Proteus, which was 8.6% in fast-thawed mackerel during storage at ambient temperature. This suggests that the decarboxylation potential is dependent on both microbial load and microbial community structure.

Frozen fish product is a major form in which fish is transported and distributed globally. Thawing, a process by which frozen water melts from fish, is an important intermediate step in the use of fish as human food (8). Different thawing methods based on water, air or steam, and microwave and radio frequency systems are currently used industrially and domestically. There are no existing definitive standards for thawing at present; however, some national bodies and the Codex Alimentarius Commission of the Food and Agriculture Organization of the United Nations and World Health Organization have drafted guidelines for air and water immersion thawing for frozen fish blocks. In these guidelines, air and water temperatures are recommended to be between 12 and 25°C and 16 and 22°C, respectively (1). Fish is a highly perishable product, and temperature abuse, coupled with bacterial contamination, induces proliferation of both gram-positive and gram-negative bacteria, resulting in a loss of sensory, chemical, and microbial quality (1, 14). The supply chain in the fish industry in developing countries is typically short due to low volumes and inadequate cold chain capacity. Moreover, imported frozen fish is normally thawed without control and often occurs in the imported boxes during storage at retail. The fish is then exposed to temperature abuse (26 to 32°C), resulting in large postharvest losses (16, 25).

It is well established that temperature influences the growth rates and types of bacteria that proliferate in fish. The growth of decarboxylating bacteria may result in spoilage and, consequently, the production of toxic biogenic amines in fish of the Scombridae family, such as mackerel and tuna, owing to the high concentration of biogenic amine precursor in their tissues (19, 22). Biogenic amines are low-molecular-weight organic basic nitrogenous compounds formed by microbial decarboxylation of free amino acids in fish, cheese, vegetables, and other foodstuff (18, 24, 27). The endogenous biogenic amines are physiologically important in the body, but high intake of exogenous amines may cause scombrotoxicosis (18). Only histamine from histidine and tyramine from tyrosine have been associated with high biological activities (27), whereas other amines (cadaverine and putrescine) are suggested to be histamine potentiators (14, 27, 34, 38). The decarboxylation potential of bacteria in fish is not well established, although a few studies have attempted to classify bacteria into high and low histamine producers. The species Morganella morganii, Enterobacter aerogenes, Raoultella planticola, and Photobacterium damselae have been classified as high producers (2, 19, 33); Proteus vulgaris, Erwinia sp., Citrobacter freundii, Citrobacter brakii, and Hafnia alvei are reported to be low producers (2, 33). There is limited information for the other amines; nonetheless, the bacterial families Enterobacteriaceae and Pseudomonadaceae and the genera Lactobacillus, Enterococcus, and Staphylococcus are reported to be important in their production (13).

Biogenic amine production is, however, a multifactorial process; it can be influenced by levels of free precursor amino acids, bacterial community, and bacterial growth, as determined by temperature and storage time (10, 13, 19). Although thawing is an important step, the influence of different thawing methods on microbial growth rate, diversity, and development of biogenic amines is still not known. In this study, culture and DNA sequencing techniques were used to unravel the influence of thawing and postthawing storage approaches on growth kinetics, diversity, and succession of bacteria in thawed Atlantic mackerel (Scomber scombrus). The effect on biogenic amine production was also explored.

Raw material. Atlantic mackerel weighing 290.3 ± 36.0 g was harvested in July 2013 southeast of Iceland using a purse seine. The fish was headed and gutted and was refrigerated to below −1°C until landing. Fish was frozen in 22-kg block cartons and was kept at −18°C for 6 months before it was transported frozen (−15 to −18°C) within 7 h to the laboratory.

Thawing and storage. Slow and fast thawing processes were followed by storage at ambient temperature and refrigerated (cold) storage. These conditions gave four experimental groups: slow-thawed fish subsequently stored at ambient (SH group) and refrigerated (SD group) temperatures and fast-thawed fish subsequently stored at ambient (FH group) and refrigerated (FD group) temperatures. Slow thawing was achieved by keeping the fish overnight (14 h) in closed polystyrene boxes placed in an incubator set at 30°C. Fast thawing, on the other hand, was done by immersing frozen fish without packaging in tap water with an initial temperature of 18°C for 2 h (time predetermined in a thawing trial). At each sampling point representing fast- and slow-thawed fish, four individual fish were collected and stored in polystyrene boxes under cold storage (2 to 5°C). Samples were then collected on days 4, 6, 8, 10, and 12 for analysis. At ambient (30°C) temperature storage, four pieces of both fast- and slow-thawed fish were kept in open polystyrene boxes and were sampled after 12, 24, 30, and 36 h. The temperatures during storage were recorded every 10 min inside the fish using temperature loggers (iButton type DS1922L, Maxim Integrated Products, Sunnyvale, CA).

Microbial analysis by the culture-dependent method. Bacterial counts were conducted in duplicate on fillets from two pieces of mackerel. Counts of total viable psychrotrophic bacteria (TVC) and hydrogen sulfide–producing bacteria (HSB) were enumerated on iron agar as described by Gram et al. (15). The samples were first minced in an electronic mixer (type 4262, Braun, Kronberg, Germany) for 20 s. Subsequently, 20 g of minced sample was aseptically weighed in a stomacher bag and mixed with 180 ml of cooled peptone saline (0.1% peptone, 1% NaCl; Oxoid Ltd., Basingstoke, UK). Homogenization was then done in a Waring laboratory blender (Stomacher 400, Seward, London, UK) at 230 rpm for 1 min, followed by serial 10-fold dilutions. Spread plating was then done on iron agar prior to incubation at 17°C for 5 days. Presumptive pseudomonads were enumerated on modified cephaloridine fucidin cetrimide (mCFC) agar as described by Stanbridge and Board (31). Plating was done on pseudomonads agar base (Oxoid) with CFC selective agar supplement (Oxoid), and the plates were incubated at 22°C for 3 days. Enumeration of bacteria from duplicate measurements of two pieces of fillets was expressed as mean log CFU per gram of fish.

Culture-independent method. Template genomic DNA was extracted from a 1-ml aliquot pooled from the duplicate of homogenates, as described in the culture-dependent method section, which was kept at −20°C until DNA extraction. The sample was centrifuged at 11,000 × g for 10 min to form a pellet, and the supernatant was discarded. From the recovered pellet, the DNA was isolated using MasterPure DNA purification kit according to manufacturer's instructions (cat. no. MCD85201, Epicentre, Madison, WI).

The 16S rRNA pyrosequence analysis used Visualization and Analysis of Microbial Population Structures primers to amplify the v3 to v6 region using 5′-YCTACGGRNGGCWGCAG-3′ and 5′-CGACRRCCATGCANCACCT-3′ for forward and reverse, respectively (30). Titanium adaptors A and B were attached to the forward and reverse primers, respectively, along with multiplex identifier adaptors recommended by Roche for FLX pyrosequencing. PCR was performed in a 25-μl reaction volume using the FastStart High Fidelity polymerase system (Roche, Madison, WI). Thermocycling conditions were as follows: 95°C for 3 min, 35 cycles at 95°C for 40 s, 54°C for 40 s, 72°C for 30 s, and a final extension step at 72°C for 7 min. The sequencing was done according to manufacturer's instructions for FLX amplicon sequencing (GS FLX Titanium SV emPCR kit [Lib-A], Roche Life Science, Madison, WI).

Sequencing. The 16S rRNA sequences were processed through the QIIME (Quantitative Insights Into Microbial Ecology) pipeline using the Greengenes database (version 1.2.1 (4)). Quality processing was initially performed by filtering sequences that were under 200 bp and over 1,000 bp containing incorrect primer sequences (more than one mismatch). Samples were assigned through the multiplex identifier sequences and were clustered into operational taxonomic units (OTUs) based on 97% similarity in the 16S rRNA sequences using Uclust (9). Phylogenetic taxonomy of the sequences was done through MSU's Ribosomal Database Project classifier (https://rdp.cme.msu.edu/classifier/classifier.jsp). OTU sequences were then aligned using Python Nearest Alignment Space Termination (http://biocore.github.io/pynast/). Chloroplast sequences were filtered from the data set, and taxa were then summarized through bar plots within the QIIME toolbox. The α-diversity parameters Chao1 and Shannon-Wiener index, which describe the microbial richness (number of OTUs) and evenness (homogeneity of abundance) in a sample, were calculated. The Shannon-Wiener index was computed as H = pi log pi, where H is the index, k is the total number of observed species, and pi is the proportion of k making up the ith species (29). On the other hand, the Chao1 index was computed as , where sobs is the observed species, n1 is the highest number of species captured once, and n2 is the number of species captured twice. The index describes the richness with an added correlation factor particularly useful for data sets skewed toward low abundance classes (17). A β-diversity analysis (microbial diversity among samples) was done by principal coordinate analysis using weighted phylogenetic UniFrac distances between samples and subsampling of 150 sequences from each sample as described by Lozupone et al. (23). Samples with similarity in microbial diversity tend to group together. The data were also analyzed using the SILVAngs (https://www.arb-silva.de/ngs/) pipeline, using an automatic and standardized procedure to analyze bulk 16S rRNA sequence data according to Quast et al. (28).

Extraction of biogenic amines. Four biogenic amines, namely, histamine, cadaverine, putrescine, and tyramine, were quantified using high-pressure liquid chromatography (HPLC) as described by Corbin et al. (5) on pooled samples from four pieces of mackerel per thawing and storage treatment. Minced fish sample (25 g), prepared as described in the culture-dependent method section, was weighed into a 250-ml plastic container and homogenized (Ultra Turrax T25 homogenizer, Janke and Kunkel, IKA Labortechnik, Staufen, Germany) at 9,100 rpm for 1 min in 50 ml of 10% TCA. The extract was then filtered through S-Pak 0.45-μm-pore-size filter paper (catalog no. HAWG047S6, EMD Millipore, Darmstadt, Germany) into a 100-ml volumetric flask.

Derivatization procedure and quantification. Derivatization and quantification were conducted as described earlier by Corbin et al. (5), with slight modifications. Briefly, a total of 0.25 ml of filtered sample extract or standard was added to 0.5 ml of o-phthaldialdehyde reagent in a test tube with a screw cap and then was kept in darkness for exactly 3.5 min. Two milliliters of ethyl acetate was then added and vortexed (Vortex Genie model K-550, Scientific Industries Inc., Springfield, MA) for 1 min for complete phase separation. An aliquot from the top phase was pipetted into a vial and was injected for analysis exactly 3.5 min after the addition of ethyl acetate. The derivatives (5 μl) were eluted at room temperature with 90% (vol/vol) acetonitrile-water mixture at a flow rate of 1.0 ml/min in a runtime of 40 min. The column consisted of Hypersil BDS C18 (5 μm, 250 by 4.0 mm; Thermo Scientific, Waltham, MA) with a guard column (5 μm, 10 by 4 mm; LiChrospher RP-18, Merck, Darmstadt, Germany). The Varian 9070 fluorescence detector (Varian Instruments Inc., Palo Alto, CA) was set at excitation and elution wavelengths of 336 and 440 nm, respectively. Quantification of biogenic amines was determined by using a standard curve of each amine with good linearity (R2 > 0.997).

Primary modeling. The variance in microbial counts was first standardized by transformation of the data to log counts. Dynamic Modeling Curve Fitting Software, DMFit 1.0, an Excel add-in (http://www.ifr.ac.uk/safety/DMfit/), was used to fit microbial growth data based on the Baranyi model. Parameters, including lag phase (L), maximum growth rate (μmax), and final count before maximum population density (M), were estimated. The fits with low values of mean squared error and high R2 were chosen. Ideally, models with a mean squared errors value of 0 and an R2 of 1 represent a perfect fit. Product characteristics including initial HSB counts, thawing, and storage temperature were used to estimate shelf life by fitting these parameters in the Food Spoilage and Safety Predictor (FSSP; version 4.0) (6).

Statistical analysis. Univariate analysis of variance on factorial treatment layout was conducted to test the effect of thawing and storage approaches on initial counts, L, μmax, and M using IBM SPSS version 21 (IBM Corp., Armonk, NY). QIIME and SILVAngs pipelines were used to determine diversity and succession of bacteria.

Two hours was required for the fillets to reach 7°C during fast thawing, whereas 12 h was required to reach 2°C using slow thawing. These temperatures were considered the end point of thawing. During the storage period, the mean temperatures in the fillets were 27.83 ± 5.5°C at 30°C incubation and 4.30 ± 5.0°C under refrigeration with frequent opening and closing to represent home conditions. The effect of thawing and storage approaches on microbial growth kinetics is shown in Figure 1.

FIGURE 1.

Growth curves of total viable bacteria (A and B), hydrogen sulfide–producing bacteria (C and D), and pseudomonads (E and F). The left and right panels represent growth curves for storage at ambient temperature (30°C) and cold storage (2 to 5°C), respectively. The letters F and S represent fast and slow thawing, respectively, whereas H and D denote the length of storage either in hours or in days.

FIGURE 1.

Growth curves of total viable bacteria (A and B), hydrogen sulfide–producing bacteria (C and D), and pseudomonads (E and F). The left and right panels represent growth curves for storage at ambient temperature (30°C) and cold storage (2 to 5°C), respectively. The letters F and S represent fast and slow thawing, respectively, whereas H and D denote the length of storage either in hours or in days.

HSB. There was no significant influence (P > 0.05) of thawing method on initial HSB counts. However, thawing and storage methods significantly affected the L (P < 0.05). Slow-thawed fish resulted in a longer L (46 h) than fast-thawed fish (26 h). Cold storage showed a longer L (65.0 h) than storage at ambient temperatures (4.7 h). Thawing and storage methods had a significant interactive effect (P < 0.05) on μmax of HSB bacteria. Lower μmax values for slow-thawed (0.29 ± 0.43 h−1) and fast-thawed (0.29 ± 0.04 h−1) fish stored at ambient temperature were observed in comparison with cold storage. Under cold storage, slow-thawed fish produced higher μmax (1.7 ± 0.22 h−1) than the fast-thawed fish (0.83 ± 0.06 h−1). There was no significant influence (P > 0.05) of thawing and storage, or their interactions, on final HSB counts.

Pseudomonads. Neither initial pseudomonad count nor L were significantly (P > 0.05) influenced by thawing, storage, or their interaction. The interactive effect of both thawing and storage methods did not show any significant effect on the growth rate. However, cold storage produced a higher μmax (1.03 ± 0.31 h−1) than storage at ambient temperatures (0.30 ± 0.08 h−1). Thawing and storage methods had a significant interactive effect (P < 0.05) on the M of pseudomonads. A higher M (8.2 ± 0.2 log CFU/g) was recorded for slow-thawed fish in cold storage than in storage at ambient temperatures (6.9 ± 0.1 CFU/g). After fast thawing, a high M of pseudomonads was seen regardless of the subsequent storage method.

Total viable aerobic counts. The thawing method significantly (P < 0.05) influenced the initial TVCs, with higher counts being observed in slow thawing (3.7 ± 0.1 CFU/g) than in fast thawing (3.4 ± 0.1 CFU/g). The TVC, however, was not affected by the interactive effect of the thawing and storage methods. There was no significant effect (P > 0.05) of thawing and the interaction of thawing and storage methods on L. Cold storage resulted in a longer L (49.9 h) (P = 0.05) than storage at ambient temperature (4.0 h). No effect (P > 0.05) was observed between thawing methods or their interaction with storage temperature, although storage method by itself significantly affected the μmax. Even though cold storage resulted in a higher μmax (0.78 ± 0.29 h−1) than storage at ambient temperature (0.19 ± 0.03 h−1), the M of TVCs was not affected (P > 0.05) by the method of storage or thawing or their interaction.

Shelf life prediction. Under cold storage, the fast- and slow-thawed group exhibited a predicted shelf life of 5.0 and 4.4 days, respectively, but both displayed a less than 1-day shelf life during storage at ambient temperature.

Microbial diversity during thawing and storage. All samples were analyzed by sequencing. Six samples mainly from early storage did not yield PCR products, whereas two delivered too few sequences to enable proper analysis. A total of 20,271 sequences from the remaining 12 samples were analyzed by QIIME, with an average of 1,842 sequences per sample and 586 observed OTUs, which were considered observed species. Bacterial flora of freshly slow-thawed fish from day 1 were taken to represent microbial diversity at the beginning of storage (Fig. 2) because the fast-thawed day 1 fish did not yield PCR product. The genus Rothia of family Micrococcaceae and the SAR supergroup (stramenopiles, alveolates, and Rhizaria) typically belonging to marine bacteria were found to be the most abundant at the beginning (slow-thawed day 1) at 17.9 and 14.3% abundance, respectively (Fig. 2). The Moraxellaceae family increased in abundance, from 10.7% on day 1 to a final abundance of 73.0% in both thawing and storage approaches.

FIGURE 2.

Bar graph presenting 97.7% of all taxa identified in the samples. The remaining 2.3% of the taxa contain 72 taxa averaging 0.28% in abundance. c, class; o, order; f, family; g, genus. F, fast-thawed; S, slow-thawed; H, hours stored at 30°C; D, days stored at 2 to 5°C; the numbers following these labels refer to duration of storage.

FIGURE 2.

Bar graph presenting 97.7% of all taxa identified in the samples. The remaining 2.3% of the taxa contain 72 taxa averaging 0.28% in abundance. c, class; o, order; f, family; g, genus. F, fast-thawed; S, slow-thawed; H, hours stored at 30°C; D, days stored at 2 to 5°C; the numbers following these labels refer to duration of storage.

The abundance of the Moraxellaceae family was apparent; unidentified genera belonging to the family increased from 10.7 to 50.4% by day 12 in the slowly thawed and cold-stored fish. Acinetobacter genus, also part of the Moraxellaceae family, increased in abundance and dominated in slow-thawed and cold-stored fish at 49.0% on day 6 but decreased again to 29.6% by day 12. Similarly, Acinetobacter increased in abundance in fast-thawed and cold-stored mackerel after 6 days up to 51.5%, but there was a less marked reduction by day 12 (46.1%). There was a decrease in Moraxellaceae in the fast-thawed fish stored at ambient temperature from 58.1 to 31.1% between the 24th and 36th h. Correspondingly, the Peptostreptococcaceae family increased from 15.0 to 50.8%. The slow-thawed group, however, showed a slight increase in the abundance of both Moraxellaceae (from 47.8 to 52.9%) and Peptostreptococcaceae (from 0 to 4.4%).

A higher bacterial diversity and richness (Table 1) was observed in fish stored at ambient temperature than in cold-stored fish. Fast-thawed and cold-stored fish exhibited a higher diversity than did slow-thawed fish. By contrast, during storage at ambient temperature, the slow-thawed fish had a higher and increasing diversity of bacteria, whereas a decrease in bacterial diversity was recorded during storage of fast-thawed fish. To illustrate the β-diversity between samples, in a principal coordinate analysis (Fig. 3) that subsampled 150 sequences from each sample, three clusters were resolved and were defined by storage temperature rather than by the thawing method. Common clusters had similar bacterial composition, mainly Acinetobacter and Psychrobacter genera belonging to the Moraxellaceae family. Less clustering of samples was manifested during storage at ambient temperature, even though, compared with fast-thawed, the slow-thawed group clustered relatively well together. After 36 h of storage at ambient temperature, fast-thawed fish differentiated along the first component. This sample displayed over 50% abundance of a taxa belonging to the Peptostreptococcaceae and 8.6% of the genus Proteus belonging to the Enterobacteriaceae family.

FIGURE 3.

Principal coordinate analysis of weighted UniFrac distances between samples derived from bacterial abundances in thawing and postthawing storage analysis. F, fast-thawed; S, slow-thawed; H, hours stored at 30°C; D, days stored at 2 to 5°C; the numbers following these labels refer to duration of storage. Right circle, FH samples; center circle, SH samples; left circle, FD and SD samples.

FIGURE 3.

Principal coordinate analysis of weighted UniFrac distances between samples derived from bacterial abundances in thawing and postthawing storage analysis. F, fast-thawed; S, slow-thawed; H, hours stored at 30°C; D, days stored at 2 to 5°C; the numbers following these labels refer to duration of storage. Right circle, FH samples; center circle, SH samples; left circle, FD and SD samples.

TABLE 1.

Summary of calculated diversity indicesa

Summary of calculated diversity indicesa
Summary of calculated diversity indicesa

Influence of thawing and bacterial growth on biogenic amine production. The levels of biogenic amines did not significantly differ between slow- and fast-thawed fish under both ambient and cold storage conditions, but they increased with time (Fig. 4). For cold-stored samples, biogenic amines could only be detected after 6 and 8 days of storage in slow- and fast-thawed fish, respectively. At this time, Acinetobacter was most abundant (51%), followed by Psychrobacter (48%), in the slow-thawed fish. Fast-thawed fish exhibited Psychrobacter exclusively (100%). After storage for 24 h, Psychrobacter was most abundant, occurring at 54 and 40% for the slow- and fast-thawed fish, respectively.

FIGURE 4.

Production of biogenic amines in fast (left, A, C) and slow (right, B, D) thawed mackerel during storage at ambient temperature (30°C) for 36 h (A, B) and cold storage (2 to 5°C) for 12 days (C, D).

FIGURE 4.

Production of biogenic amines in fast (left, A, C) and slow (right, B, D) thawed mackerel during storage at ambient temperature (30°C) for 36 h (A, B) and cold storage (2 to 5°C) for 12 days (C, D).

Despite the potential impact of the temperature regime used during thawing of fish on quality and microbial growth, the effect of thawing methods on microbial diversity has scarcely received attention. Thawing, an important intermediary step in frozen fish preparation, can potentially influence the quality and safety of fish products (8). At the end of the thawing period, the fast-thawed fish was, on average, at 7°C, whereas the slow-thawed fish was at 2°C, which was in accordance with the proposed thawing guidelines.

These temperature differences may have led to the significantly shorter lag phase of HSB in fish that was subsequently stored at ambient temperature than in fish kept in cold storage. However, this was not noted for TVC or pseudomonads.

The influence of thawing on shelf life was marginal, but fast-thawed fish in cold storage exhibited a 15-h extended shelf life of 5.0 days compared with 4.4 days for slow-thawed fish. Under storage at ambient temperature, the predicted shelf life was less than 1 day for both thawing methods. Estimated growth rate during simulated storage at ambient temperatures was higher using FSSP compared with DMfit; however, under cold storage, a contrary observation was made. It has been demonstrated that FSSP may predict a shorter shelf life than sensory techniques (26). Shelf life prediction is mainly an estimate, given that the shelf life is influenced by many factors, including bacterial species, diversity, growth substrates, etc., which have been estimated to influence prediction accuracy by 25% (7).

Influence of thawing and storage regime on bacterial diversity and production of biogenic amines. A low and delayed production of biogenic amines was observed during cold storage in this study. These results are in agreement with studies of mackerel inoculated with known biogenic amine producer M. morganii (19) and of Indian mackerel stored at 0 and 3°C (38). Levels of histamine, cadaverine, and putrescine lower than 70 ppm have also been previously reported for gutted Atlantic mackerel kept on ice for 12 days, but in contrast, the same author reported higher values in ungutted mackerel (22). This suggests that the mere presence of prolific biogenic amine producers cannot fully explain amine production but that it is modulated by many factors. Most importantly, the interaction of these factors, bacterial species, time, and storage temperatures, has been shown to greatly influence decarboxylation (19, 21). The importance of different species of bacteria is underscored by their genetic capacity to decarboxylate the amino acids that are precursors to biogenic amines (2, 20). Previously, it has been reported that such bacteria are gram-negative bacteria, especially of enteric and marine origin (2, 11, 32). Under optimum growth conditions, the prolific producers of histamine above 1,000 ppm include M. morganii, E. aerogenes, R. planticola, and P. damselae (2, 19, 33). On the other hand, P. vulgaris, Erwinia sp., C. freundii, C. brakii, and H. alvei are classified as low histamine producers (10 to 500 ppm) (2, 33). Under storage at ambient temperature, the Enterobacteriaceae family is reported to be responsible for the production of biogenic amines (33). Morganella spp. are reported to produce histamine to levels above 300 ppm within 24 h at an optimum temperature of 25°C (19) and within 48 h (38). In this study, however, low values of biogenic amines were revealed despite storage at a slightly higher temperature (30°C).

To determine similarities occasioned by the treatments, principal coordinate analysis was applied to the 16S rRNA sequence data set. A clustering of samples with internal similarities exhibited a higher abundance of Moraxellaceae family (genus Acinetobacter) in cold-stored samples than in samples stored at ambient temperatures. This may have influenced the ability of the bacterial microbiota to produce biogenic amines. Fast-thawed samples that were subsequently stored at ambient temperature were uniquely characterized by a higher abundance of the family Peptostreptococcaceae compared with the slow-thawed samples. The slow-thawed samples also showed patterns of bacterial dominance similar to those of the fast-thawed ones, except after 24 h of storage, when Peptostreptococcaceae was absent. This may explain why the samples that were slow-thawed, stored at ambient temperature, and collected at 24 h of storage were separated in the principal coordinate analysis plot. After 36 h of storage at ambient temperature, the fast-thawed fish was clearly separated along principal coordinate 1, while exhibiting a relatively higher abundance of Proteus (8.6%), which is a low biogenic amine producer compared with the other producers. Citrobacter, Enterobacter, and Morganella, also biogenic producers of the Enterobacteriaceae family, were present, albeit at lower abundance than Proteus. However, this did not result in higher biogenic amine production in fast-thawed fish compared with slow-thawed fish at the same temperature. Taken together, the levels of biogenic amines were below the defect action limit of 100 and 50 ppm as set by the European Union and the U.S. Food and Drug Administration, respectively (12, 35). This could be attributed to the low abundance of both the low and prolific decarboxylators, as revealed in this study. Their low load could be explained by the good postharvest handling received by the fish used in this study. Gutting and heading with immediate chilling, followed by frozen storage, reduces both the bacterial load and increases the lag time (14, 22, 36). Further investigations are, however, needed to affirm this, given that only one independent analysis of biogenic amines was done in the present study.

Our findings, nonetheless, demonstrate that Psychrobacter, Acinetobacter, and other genera of Moraxellaceae were generally the dominating bacteria in both slow- and fast-thawed mackerel stored at either refrigerated or ambient temperatures. This is in accordance with previous studies, which confirmed the dominance of these bacteria in fish from temperate and cold waters (3, 32, 37). The study also revealed that the growth of Acinetobacter was suppressed when stored at ambient temperatures as compared to cold storage. Bacteria of the Peptostreptococcaceae family were inhibited during cold storage.

In conclusion, storage and thawing methods, and their interaction, significantly affect the growth kinetics and diversity of bacteria but not of biogenic amines, which suggests a dependence on prevalent microbiota and their amino acid decarboxylation capacity. Extended shelf life (15 h) was achieved in fast- compared with slow-thawed mackerel in cold storage. Furthermore, higher bacterial diversity was observed in storage at ambient temperature than cold storage and in slow-thawed than in fast-thawed mackerel. Psychrobacter and Acinetobacter, belonging to the Moraxellaceae family, were the dominant bacteria, irrespective of the thawing and storage method. The study confirms that good postharvest practices and cold storage extend the shelf life of fresh mackerel.

The authors thank United Nations University (UNU)–Fisheries Training Programme for funding the study. Special thanks to þór Ásgeirsson, Mary Frances, and Sigríður Ingvarsdóttir at the UNU-FTP for assistance in language and proofreading. We also appreciate the technical assistance offered by Páll Steinþórsson and Ingibjörg Rósa þorvaldsdóttir at Matis. Finally, we thank Sigurjón Arason and Magnea Karlsdóttir for providing frozen mackerel samples for the study. The authors declare no conflict of interest.

1.
Archer
,
M.
,
M.
Edmonds
, and
M.
George
.
2008
.
Seafood thawing
.
Seafish Research & Development, Sea Fish Industry Authority
,
Edinburgh and Grimsby, UK
.
2.
Bjornsdottir-Butler
,
K.
,
J. L.
Jones
,
R. A.
Benner
, Jr.
, and
W.
Burkhardt
III
.
2011
.
Quantification of total and specific gram-negative histamine-producing bacteria species in fish using an MPN real-time PCR method
.
Food Microbiol
.
28
:
1284
1292
.
3.
Broekaert
,
K.
,
M.
Heyndrickx
,
L.
Herman
,
F.
Devlieghere
, and
G.
Vlaemynck
.
2011
.
Seafood quality analysis: molecular identification of dominant microbiota after ice storage on several general growth media
.
Food Microbiol
.
28
:
1162
1169
.
4.
Caporaso
,
J. G.
,
J.
Kuczynski
,
J.
Stombaugh
,
K.
Bittinger
,
F. D.
Bushman
,
E. K.
Costello
,
N.
Fierer
,
A.
Gonzalez Peña
,
J. K.
Goodrich
,
J. I.
Gordon
,
G. A.
Huttley
,
S. T.
Kelley
,
D.
Knights
,
J. E.
Koenig
,
R. E.
Ley
,
C. A.
Lozupone
,
D.
McDonald
,
B. D.
Muegge
,
M.
Pirrung
,
J.
Reeder
,
J. R.
Sevinsky
,
P. J.
Turnbaugh
,
W. A.
Walters
,
J.
Widmann
,
T.
Yatsunenko
,
J.
Zaneveld
, and
R.
Knight
.
2010
.
QIIME allows analysis of high-throughput community sequencing data
.
Nat. Methods
7
:
335
336
.
doi:10.1038/nmeth.f.303
.
5.
Corbin
,
J. L.
,
B. H.
Marsh
, and
G. A.
Peters
.
1989
.
An improved method for analysis of polyamines in plant tissue by precolumn derivatization with o-phthalaldehyde and separation by high performance liquid chromatography
.
Plant Physiol
.
90
:
434
439
.
6.
Dalgaard
,
P.
2014
.
Food spoilage and safety predictor (FSSP) version 4.0
.
Technical University of Denmark, National Institute for Aquatic Resources
,
Charlottenlund, Denmark
.
7.
Dalgaard
,
P.
,
P.
Buch
, and
S.
Silberg
.
2002
.
Seafood spoilage predictor—development and distribution of a product specific application software
.
Int. J. Food Microbiol
.
73
:
343
349
.
8.
Dincer
,
T.
,
A.
Cadun
,
S.
Cakli
, and
S.
Tolasa
.
2009
.
Effects of different thawing methods on the freshness quality of fish
.
E.U. J. Fish. Aquat. Sci
.
26
:
253
256
.
9.
Edgar
,
R. C.
2010
.
Search and clustering orders of magnitude faster than BLAST
.
Bioinformatics
26
:
2460
2461
.
doi:10.1093/bioinformatics/btq461
.
10.
Ekici
,
K.
, and
M.
Alisarli
.
2008
.
Histamine formation and microbiological changes in endemic Chalcalburnus tarichi Pallas 1811 (Inci Kefali) stored at 4°C
.
Arch. Med. Vet
.
40
:
95
98
.
11.
Emborg
,
J.
,
B. G.
Laursen
,
T.
Rathjen
, and
P.
Dalgaard
.
2002
.
Microbial spoilage and formation of biogenic amines in fresh and thawed modified atmosphere packed salmon (Salmo salar) at 2°C
.
J. Appl. Microbiol
.
92
:
790
799
.
12.
European Commission
.
2005
.
Commission Regulation (EC) No 2073/2005 of 15 November 2005 on microbiological criteria for foodstuffs
.
Off. J. Eur. Union
L 338
:
1
26
.
13.
European Food Safety Authority
.
2011
.
EFSA Panel on Biological Hazards (BIOHAZ); scientific opinion on risk based control of biogenic amine formation in fermented foods
.
EFSA J
.
9
:
2393
.
14.
Food and Agriculture Organization of the United Nations, World Health Organization
.
2013
.
Public health risks of histamine and other biogenic amines from fish and fishery products
.
Food and Agriculture Organization of the United Nations, World Health Organization
,
Rome
.
15.
Gram
,
L.
,
G.
Trolle
, and
H. H.
Huss
.
1987
.
Detection of specific spoilage bacteria from fish stored at low (0°C) and high (20°C) temperatures
.
Int. J. Food Microbiol
.
4
:
65
72
.
16.
Gram
,
L.
,
C.
Wedell-Neergaard
, and
H. H.
Huss
.
1990
.
The bacteriology of fresh and spoiling Lake Victorian Nile perch (Lates niloticus)
.
Int. J. Food Microbiol
.
10
:
303
316
.
17.
Hughes
,
J. B.
,
J. J.
Hellmann
,
T. H.
Ricketts
, and
B. J. M.
Bohannan
.
2001
.
Counting the uncountable: statistical approaches to estimating microbial diversity
.
Appl. Environ. Microbiol
.
67
:
4399
4406
.
18.
Hungerford
,
J. M.
2010
.
Scombroid poisoning: a review
.
Toxicon
56
:
231
243
.
19.
Kim
,
S. H.
,
R. J.
Price
,
M. T.
Morrissey
,
K. G.
Field
,
C. I.
Wei
, and
H.
An
.
2002
.
Histamine production by Morganella morganii in mackerel, albacore, mahi-mahi, and salmon at various storage temperatures
.
J. Food Sci
.
67
:
1522
1528
.
20.
Küley
,
E.
,
F.
Özogul
,
E.
Balikçi
,
M.
Durmus
, and
D.
Ayas
.
2013
.
The influences of fish infusion broth on the biogenic amines formation by lactic acid bacteria
.
Braz. J. Microbiol
.
44
:
407
413
.
21.
Lokuruka
,
M. N. I.
, and
J. M.
Regenstein
.
2004
.
Biogenic amines in iced and temperature-abused tropical fish
.
J. Aquat. Food Prod. Technol
.
13
:
87
99
.
22.
Lokuruka
,
M. N. I.
, and
J. M.
Regenstein
.
2007
.
Handling and storage of Atlantic mackerel (Scomber scombrus) on biogenic amine production
.
J. Aquat. Food Prod. Technol
.
15
:
17
33
.
23.
Lozupone
,
C.
,
M. E.
Lladser
,
D.
Knights
,
J.
Stombaugh
, and
R.
Knight
.
2011
.
UniFrac: an effective distance metric for microbial community comparison
.
Int. Soc. Microb. Ecol. J
.
5
:
169
172
.
24.
Naila
,
A.
,
S.
Flint
,
G.
Fletcher
,
P.
Bremer
, and
G.
Meerdink
.
2010
.
Control of biogenic amines in food—existing and emerging approaches
.
J. Food Sci
.
75
:
R139
150
.
25.
Odoli
,
C. O.
,
P. M.
Oduor-Odote
,
B.
Ohowa
, and
S. O.
Onyango
.
2013
.
Evaluation of fish handling techniques employed by artisanal fishers on quality of lethrinids and siganids fish genera at landing time along the Kenyan coast using sensory and microbiological methods
.
Afr. J. Food Agric. Nutr. Dev
.
13
:
8167
8186
.
26.
Olafsdottir
,
G.
,
H. L.
Lauzon
,
E.
Martinsdottir
, and
K.
Kristbergsson
.
2006
.
Influence of storage temperature on microbial spoilage characteristics of haddock fillets (Melanogrammus aeglefinus) evaluated by multivariate quality prediction
.
Int. J. Food Microbiol
.
111
:
112
125
.
27.
Prester
,
L.
2011
.
Biogenic amines in fish, fish products and shellfish: a review
.
Food Addit. Contam. Part A Chem. Anal. Control Expo. Risk Assess
.
28
:
1547
1560
.
28.
Quast
,
C.
,
E.
Pruesse
,
P.
Yilmaz
,
J.
Gerken
,
T.
Schweer
,
P.
Yarza
,
J.
Peplies
, and
F. O.
Glockner
.
2013
.
The SILVA ribosomal RNA gene database project: improved data processing and web-based tools
.
Nucleic Acids Res
.
41
:
D590
596
.
29.
Shannon
,
C. E.
, and
W.
Weaver
.
1963
.
The mathematical theory of communication
.
University of Illinois Press
,
Champaign
.
30.
Sogin
,
M. L.
,
H. G.
Morrison
,
J. A.
Huber
,
D. M.
Welch
,
S. M.
Huse
,
P. R.
Neal
,
J. M.
Arrieta
, and
G. J.
Herndl
.
2006
.
Microbial diversity in the deep sea and the underexplored “rare biosphere.”
Proc. Natl. Acad. Sci. USA
103
:
12115
12120
.
31.
Stanbridge
,
L. H.
, and
R. G.
Board
.
1994
.
A modification of the Pseudomonas selective medium, CFC, that allows differentiation between meat pseudomonads and Enterobacteriaceae
.
Lett. Appl. Microbiol
.
18
:
327
328
.
32.
Svanevik
,
C. S.
, and
B. T.
Lunestad
.
2011
.
Characterisation of the microbiota of Atlantic mackerel (Scomber scombrus)
.
Int. J. Food Microbiol
.
151
:
164
170
.
33.
Takahashi
,
H.
,
B.
Kimura
,
M.
Yoshikawa
, and
T.
Fujii
.
2003
.
Cloning and sequencing of the histidine decarboxylase genes of gram-negative, histamine-producing bacteria and their application in detection and identification of these organisms in fish
.
Appl. Environ. Microbiol
.
69
:
2568
2579
.
34.
Tao
,
Z.
,
M.
Sato
,
H.
Zhang
,
T.
Yamaguchi
, and
T.
Nakano
.
2011
.
A survey of histamine content in seafood sold in markets of nine countries
.
Food Control
22
:
430
432
.
35.
U.S. Food and Drug Administration
.
2011
.
Fish and fishery products hazards and controls guidance
.
U.S. Department of Health and Human Services, Food and Drug Administration, Center for Food Safety and Applied Nutrition
,
Silver Spring, MD
.
36.
Visciano
,
P.
,
M.
Schirone
,
R.
Tofalo
, and
G.
Suzzi
.
2012
.
Biogenic amines in raw and processed seafood
.
Front. Microbiol
.
3
:
188
.
37.
Wilson
,
B.
,
B. S.
Danilowicz
, and
W. G.
Meijer
.
2008
.
The diversity of bacterial communities associated with Atlantic cod Gadus morhua
.
Microb. Ecol
.
55
:
425
434
.
38.
Zare
,
D.
,
K.
Muhammad
,
M. H.
Bejo
, and
H. M.
Ghazali
.
2013
.
Changes in urocanic acid, histamine, putrescine and cadaverine levels in Indian mackerel (Rastrelliger kanagurta) during storage at different temperatures
.
Food Chem
.
139
:
320
325
.