Food Chemistry 114 (2009) 933–938

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier.com/locate/foodchem

Histamine content and histamine-forming bacteria in dried milkfish (Chanos chanos) products

Hsiu-Hua Hsu a, Tin-Chen Chuang a, Hung-Chou Lin b, Yu-Ru Huang c, Chia-Min Lin b, Hsien-Feng Kung d, Yung-Hsiang Tsai b,* a Graduate Institute of Applied Science of Living, Tainan University of Technology, Tainan, ROC b Department of Seafood Science, National Kaohsiung Marine University, No. 142, Hai-Chuan Road, Nan-Tzu, Kaohsiung City, 811, Taiwan, ROC c Department of Food Science, National Penghu University, Penghu, Taiwan, ROC d Department of Food Science and Technology, Tajen University, Pingtung, Taiwan, ROC article info abstract

Article history: Thirty-two dried milkfish products sold in five retail markets in southern Taiwan were purchased and Received 15 August 2008 tested to determine the occurrence of histamine and histamine-forming bacteria. Except for histamine Received in revised form 14 September and cadaverine, the average content of various biogenic amines in tested samples was less than 2008 8.5 mg/100 g. Most of the tested dried milkfish products (78.1%) had histamine levels greater than the Accepted 22 October 2008 FDA guideline of 5 mg/100 g for scombroid fish and/or product, while fourteen of them (43.7%) contained > 50 mg/100 g hazard action level. Thirty histamine-producing bacterial strains, capable of producing 5.4 ppm to 562 ppm of histamine in trypticase soy broth supplemented with 1.0% L-histidine (TSBH), Keywords: were identified as Enterobacter aerogenes (seven strains), Citrobacter sp. (one strain), Staphylococcus xylo- Histamine Dried milkfish sus (ten strains), S. sciuri (one strain), Bacillus thuringiensis (two strains), Citrobacter freundii (five strains), Histamine-forming bacteria Klebsiella pneumoniae (one strain) and E. cloacae (three strains), by 16S rDNA sequencing with PCR Quality amplification. Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction Biogenic amines are formed mainly through the decarboxyl- ation of specific free amino acids by exogenous decarboxylases re- Histamine is the causative agent of scombroid poisoning, a leased by the microbial species associated with the seafood (Taylor food-borne chemical hazard. Scombroid poisoning is usually a mild & Speckard, 1983). Many different bacterial species are known to illness with a variety of symptoms including rash, urticaria, nausea, possess histidine decarboxylase and have the ability to produce vomiting, diarrhea, flushing, and tingling and itching of the skin histamine. Although only Morganella (Proteus) morganii, Klebsiella (Taylor, 1986). Severity of the symptoms can vary considerably pneumoniae and Hafnia alvei have been isolated from the fish with the amount of histamine ingested and the individual’s sensi- incriminated in scombroid poisoning, a variety of other bacterial tivity to histamine. Scombroid fish such as tuna, mackerel, bonito, species capable of producing histamine have been identified in fish and saury that contain high levels of free histidine in their muscle (Taylor & Speckard, 1983). Among them are the enteric bacteria, are often implicated in scombroid poisoning incidents (Taylor, that include Proteus vulgaris, P. mirabilis, Enterobacter aerogenes, 1986). However, several species of nonscombroid fish such as Enterobacter cloacae, Serratia fonticola, Serratia liquefaciens, Raoul- mahi-mahi, bluefish, herring, and sardine have often been impli- tella (formerly Klebsiella) planticola, Raoultella ornithinolytica and cated in incidents of scombroid poisoning. In Taiwan, scombroid Citrobacter freundii (Kim et al., 2003; Tsai, Lin et al., 2005b). In addi- poisoning occurs occasionally (Chen & Malison, 1987; Tsai, Kung tion to the enteric bacteria, Clostridium spp., Vibrio alginolyticus, et al., 2005a), and the fish implicated in these outbreaks are tuna, Acinetobacter lowffi, Plesiomonas shigelloides, Pseudomonas putida, mackerel, and black marlin. Recently, due to their popularity in Pseudomonas fluorescens, Aeromonas spp., and Photobacterium spp. Taiwanese people, sailfish and marlin fillets have become the most have also been reported as histamine producers (Okuzumi, Hirai- frequently implicated fish species in scombroid outbreaks in Tai- shi, Kobayashi, & Fujii, 1994; Yatsunami & Echigo, 1991). Recently, wan (Hwang, Chang, Shiau, & Chai, 1997; Hwang, Chang, Shiau, & we demonstrated the presence of histamine-forming Proteus, Cheng, 1995). Enterobacter, Klebsiella, Rahnella and Acinetobacter in sailfish fillets in Taiwan, but failed to isolate any of the three above mentioned * Corresponding author. Tel.: +886 7 3617141 3605; fax: +886 7 3640634. major histamine-formers – the Hafnia alvei, Morganella morganii E-mail addresses: [email protected], [email protected] (Y.-H. Tsai). and K. pneumoniae (Tsai, Kung, Lee, Lin, & Hwang, 2004).

0308-8146/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodchem.2008.10.040 934 H.-H. Hsu et al. / Food Chemistry 114 (2009) 933–938

Milkfish (Chanos chanos) is an important aquacultured fish in terial numbers in the milkfish samples were expressed as log10 col- the Indo-Pacific region, particularly the Philippines, Indonesia, ony forming units (CFU)/g. and Taiwan (Chen, 1990). Histidine at approximately 441 mg/ To isolate histamine-forming bacteria, a 0.1 ml aliquot of the 100 g is the most prominent free amino acid (FAA) found in the sample dilute was spread on histamine-forming bacterium isola- white muscle of milkfish, accounting for 80% of the total FAAs in tion agar (HBI agar) fortified with L-histidine (Niven, Jeffreg, & Cor- the fish (Chiou, Shiau, & Chai, 1990). Tsai, Chang, Kung, Wei, and lett, 1981). Following incubation of the differential agar plates for Hwang (2005c) reported that milkfish was a better substrate than 4 d at 35 °C, colonies with blue or purple color on the plates were sailfish for histamine formation by bacterial histidine decarboxyl- picked and further streaked on trypticase soy agar (TSA) (Difco) to ation at elevated temperatures (>15 °C). Recently, it was first re- obtain pure cultures. Their ability to produce biogenic amines was ported that dried milkfish product could cause histamine determined by inoculating the isolates in trypticase soy broth (TSB) intoxication, which occurred in February 2006, in southern Taiwan. (Difco) supplemented with 1% L-histidine (TSBH) and incubated Raoultella ornithinolytica was the major histamine-producing bac- without shaking at 35 °C for 24 h. Two milliliters of the culture terium responsible for the high content of histamine in the impli- broth were taken for quantitation of biogenic amines. cated milkfish sample (Tsai et al., 2007). Analyses of total coliform and Escherichia coli in these dried In Taiwan, the traditional process for dried milkfish involves milkfish samples were conducted using the three-tube most prob- back-cutting, degutting, salting, and sun-drying for several days. able number (MPN) methods (FDA, 1992). Lauryl sulphate tryptose There exists no report on the occurrence of biogenic amines, broth (LST broth) and brilliant green lactose bile (2%) broth (BGLB including histamine, histamine-forming bacteria and related bac- broth) were used for presumptive and confirmed tests for total teria in dried milkfish products. Therefore, this research was coliform, respectively. E. coli was determined by using the LST undertaken by testing 32 dried milkfish products sold in retail broth and EC broth. Cultures that showed positive production of markets in southern Taiwan so that a better understanding of the gas were then confirmed by eosine methylene blue agar (EMBA) safety quality of the products could be accomplished to better pro- and IMViC tests. tect the consumers. 2.4. Identification of histamine-forming isolates 2. Materials and methods The presumptive histamine-forming isolates were identified on the basis of morphology, Gram stain, endospore stain, catalase and 2.1. Samples oxidase reaction. The identity of histamine-forming isolates was further confirmed by amplifying and sequencing approximately Thirty-two dried milkfish products purchased from five retail 1400 bp of the 16S ribosomal DNA (rDNA) for bacteria (Kuhnert, markets in southern Taiwan, including Tainan county (eight sam- Capaul, Nicolet, & Frey, 1996; Kuhnert, Heyberger-Meyer, Nicolet, ples), Tainan city-A (six samples), Tainan city-B (three samples), & Frey, 2000). Amplification of histamine-forming bacteria was Tainan city-C (five samples), and Kaohsiung county (ten samples), performed using the universal primers UNI-L (5’-AGA- were collected. All collected samples were wrapped in aseptic bags, GTTTGATCATGGCTCAG-3’) and UNI-R (5’-GTGTGACGGGCGGTGTG- placed in ice, and immediately transported to the laboratory for TAC-3’) (Kuhnert et al., 1996 and 2000). Bacterial cells were use within 8 h. cultured overnight in 2 ml of TSB at 35 °C and then centrifuged at 5,000 x g for 10 min. The cell pellet was washed and resus- 2.2. pH value, salt content, water content and water activity pended in 0.5 ml of TE-buffer (10 mM Tris–HCl, 1 mM EDTA; pH determination 8.0), and then lysed by 20% sodium dodecyl sulfate (SDS). After the solution was boiled for 20 min and the cellular debris was dis- Dried milkfish samples (10 g) were homogenised in sterile carded following centrifugation at 13,000 x g for 3 min, the total blenders with 10 ml of distilled water to make a thick slurry. The DNA in the supernatant was precipitated with 70% ethanol and pH of this slurry was then measured using a Corning 145 pH meter used as template DNA for PCR. (Corning Glass Works, Medfield, MA, USA). The salt content in each PCR amplification was performed in 20 ll reaction mixture con- sample was determined according to the AOAC (1995) procedures taining 10 mM Tris–HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2,20 by homogenizing 2 g of dried milkfish sample with 18 ml of dis- pmol of each primer, a 0.2 mM concentration for each of the four tilled water. The homogenate was titrated with 0.1 M AgNO3 using deoxynucleotide triphosphates, 0.5 U of Taq DNA polymerase (Ap- 10% w/v K2CrO4 solution as an indicator. The water content was plied Biosystems, Foster City, CA, USA), and template DNA (10 ng). conducted with the standard gravimetric method by drying 1–3 g Amplifications were carried out for 35 cycles (94 °C for 30 s, 55 °C of a test sample at 102.0 ± 2.0 °C under atmospheric pressure for for 30 s, and 72 °C for 60 s) in a GeneAmp PCR 2400 Thermal Cycler 2 h. Consistancy of mass was tested by additional drying steps of (Applied Biosystems) with an initial denaturation at 94 °C for 1 h until the difference in mass did not exceed 0.5 mg. Water activ- 4 min and a final extension at 72 °C for 7 min (Kuhnert et al., ity was determined by an Electric Hygrometer (Hygrodynamics, 1996 and 2000). Amplicons were detected by electrophoresis on Inc., Silver Spring, MD) at 27 °C. a 1.5% agarose gel staining with ethidium bromide. Amplicons were purified using a QIAquick PCR Purification Kit (Qiagen, Valen- 2.3. Microbiological analysis and isolation of histamine-forming cia, CA, USA) eluted in Tris–HCl (10 mM, pH 8.5) prior to sequenc- bacteria ing. The amplified DNA was directly sequenced with the ABI Taq Dye Deoxy Terminator Cycle sequencing kit and ABI Model 377 A 25-g portion of the dried milkfish sample was homogenised at automated DNA sequencer (Applied Biosystems). The sequences high speed for 2 min in a sterile blender with 225 ml sterile potas- were analysed with the BLAST (NCBI) for identification of hista- sium phosphate buffer (0.05 M, pH 7.0). The blender was sterilised mine-forming bacteria. by autoclaving for 15 min at 121 °C. The homogenates were seri- ally diluted with a sterile phosphate buffer, and 1.0 ml aliquots 2.5. Determination of total volatile base nitrogen (TVBN) of the dilutes were inoculated into aerobic plate count (APC) agar (Difco, Detroit, MI, USA) containing 0.5% NaCl. Bacterial colonies The TVBN content of the dried milkfish sample was measured were counted after the plates were incubated at 35 °C for 48 h. Bac- by the method of Conway’s dish (Cobb, Aoaniz, & Thompson, H.-H. Hsu et al. / Food Chemistry 114 (2009) 933–938 935

1973). The TVBN extract of the fish sample in 6% trichloroacetic levels of pH, water content, water activity, salt content, total vola- acid (TCA, Sigma, St. Louis, Mo., USA) was absorbed by boric acid tile basic nitrogen (TVBN), aerobic plate count (APC), E. coli and total and then titrated with 0.02 N HCl. The TVBN content was ex- coliform in all samples ranged from 5.4 to 6.8, 20.5–57.9%, 0.72– pressed in mg/100 g fish. 0.93, 1.2–11.6%, 12.9–146.5 mg/100 g, 1.4–9.1 log CFU/g, <3 to 2200 MPN/g and <3 to 6500 MPN/g, respectively. The average water 2.6. Biogenic amine analysis content (21.7%), water activity (0.73), and APC (2.3 log CFU/g) of Tainan city-B samples were significantly (P < 0.05) lower than those Each dried milkfish sample was ground in a Waring Blender for of the other milkfish samples, while the average TVBN level 3 min. The ground samples (5 g) were transferred to 50 ml centri- (100.9 mg/100 g) in Tainan city-B samples was the highest fuge tubes and homogenised with 20 ml of 6% trichloroacetic acid (P < 0.05) (Table 1). However, the average APC, E. coli and total coli- (TCA) for 3 min. The homogenates were centrifuged (10,000g, form values in Kaohsiung county samples (8.3 log CFU/g, 150 MPN/ 10 min, 4 °C) and filtered through Whatman No. 2 filter paper g and 1100 MPN/g, respectively) were significantly (P < 0.05) higher (Whatman, Maidstone, England). The filtrates were then placed than those in the other four retailers samples, while its average salt in volumetric flasks, and TCA was added to bring to a final volume content (1.7%) was the lowest (Table 1). The rates of unacceptable of 50 ml. Samples of standard biogenic amine solutions and 2 ml dried milkfish samples were 68.8% (22/32) for TVBN, based on the aliquots of the dried milkfish extracts were derivatised with ben- decomposition limit level of 30 mg/100 g for fish quality determi- zoyl chloride according to the previously described method nation, and 46.9% (15/32) and 12.5% (4/32), using the Taiwanese (Hwang et al., 1997). Two milliliters of each bacterial culture broth regulatory standard of 6.47 log CFU/g for APC and 50 MPN/g for were also benzoylated using the same procedures for dried milk- E. coli, respectively. The dried milkfish samples were made by fish extracts. The benzoyl derivatives were dissolved in 1 ml of sun-drying for several days, and stored at room temperature, allow- methanol, and 20 ll aliquots were used for HPLC injection. ing flies to easily contaminate and multiply. Therefore, the unhy- The contents of biogenic amines in the dried milkfish samples gienic handling or processing of dried milkfish resulted in the were determined with a Hitachi liquid chromatograph (Hitachi, poor microbiological quality observed in this study. Tokyo, Japan) consisting of a Model L-7100 pump, a Rheodyne Table 2 summarised the contents of biogenic amines in the Model 7125 syringe loading sample injector, a Model L-4000 UV– tested dried milkfish products. Except for histamine and cadaver- Vis detector (set at 254 nm), and a Model D-2500 Chromato-inte- ine, the average content of various biogenic amines in tested sam- grator. A LiChrospher 100 RP-18 reversed-phase column (5 lm, ples was less than 8.5 mg/100 g. Among them, three samples 125 Â 4.6 mm, E. Merck, Damstadt, Germany) was used for chro- collected from Tainan city-B retailer had the highest average hista- matographic separation. The gradient elution program began with mine content of 409.7 mg/100 g and cadaverine content of 50:50 (v/v) methanol:water at a flow rate of 0.8 ml/min for the first 94.9 mg/100 g. The second highest average content of histamine 0.5 min, followed by a linear increase to 85:15 methanol:water (240.9 mg/100 g) and cadaverine (40.8 mg/100 g) was obtained (0.8 ml/min) during the next 6.5 min. The methanol:water mix from samples collected from the Kaohsiung county retailer (Table was held constant at 85:15 (0.8 ml/min) for 5 min, and then de- 2). Table 3 shows the distribution of histamine contents in tested creased to 50:50 (0.8 ml/min) during the next 2 min. dried milkfish products, with 78.1% (25/32) samples containing greater than 5 mg/100 g of histamine, the allowable limit of the 2.7. Statistical analysis US Food and Drug Administration (FDA) for scombroid fish and/or product. Fourteen samples (43.7%) with greater than Pearson correlation was carried out to determine relationships 50 mg/100 g of histamine could be hazardous to the health of between pH, salt content, water content, water activity, TVBN, consumers based on the data collected from numerous outbreak APC, E. coli, TC, and histamine contents in the 32 dried milkfish reports (Taylor, 1989). However, Bartholomev, Berry, Rodhouse, samples. All statistical analyses were performed using the Statisti- and Gilhouse (1987) demonstrated that histamine at greater than cal Package for Social Sciences, SPSS Version 9.0 for windows (SPSS 100 mg/100 g in fish would be toxic and unsafe for human con- Inc., Chicago, Il. USA). Value of P < 0.05 was used to indicate signif- sumption. Based on information on the toxicological levels of his- icant deviation. tamine in various seafood products in causing health hazards, 13 samples (40.6%) could have caused disease symptoms when 3. Results and discussion consumed, mainly due to their greater than 100 mg/100 g level of histamine content (Table 3). High levels of histamine content Values of the pH, water content, water activity, salt content, APC, have been found in various types of fish implicated in scombroid TVBN, E. coli, and total coliform in the thirty-two dried milkfish poisoning. Hernandez-Herrero, Roig-Sagues, Rodriguez-Jerez, samples from the five retail markets are presented in Table 1. The and Mora-Ventura (1999) detected 68 mg/100 g of histamine in

Table 1 Values of the pH, water content, water activity, salt content, total volatile basic nitrogen (TVBN), aerobic plate count (APC), E. coli, and total coliform (TC) in dried milkfish products.

Sample No. of pH Water Water activity Salt content TVBN APC (log E. coli TC (MPN/g) sources samples content (%) (%) (mg/100 g) CFU/g) (MPN/g)

Tainan 8 5.45.6 41.350.5 0.720.74 9.311.6 12.931.4 3.14.1 <3 <3 county (5.5 ± 0.1)aC (44.8 ± 2.5)B (0.73 ± 0.01)C (10.3 ± 0.7)A (19.4 ± 6.0)D (3.6 ± 0.3)C Tainan city 6 5.75.8 32.452.9 0.730.83 2.45.5 21.535.2 3.86.8 <3 <390 (25 ± 36) -A (5.8 ± 0.1)B (39.9 ± 9.9)BC (0.79 ± 0.03)B (3.8 ± 1.0)C (29.0 ± 4.1)C (6.1 ± 1.1)B Tainan city 3 6.26.8 20.522.6 0.720.75 3.14.3 65.3146.5 1.43.3 <3 <3 -B (6.4 ± 0.3)A (21.7 ± 0.8)D (0.73 ± 0.01)C (3.8 ± 0.5)C (100.9 ± 41.3)A (2.3 ± 0.8)D Tainan city 5 5.86.0 31.940.9 0.900.93 3.26.2 35.545.9 5.06.5 <3 <3 -C (5.9 ± 0.1)B (36.0 ± 3.1)C (0.92 ± 0.01)A (4.9 ± 1.3)B (38.2 ± 4.4)B (5.6 ± 0.5)B Kaohsiung 10 6.06.8 31.157.9 0.920.93 1.22.3 38.949.8 8.09.1 <32200 1506500 county (6.2 ± 0.2)A (50.1 ± 6.9)A (0.92 ± 0.01)A (1.7 ± 0.4)D (43.1 ± 3.5)B (8.3 ± 0.3)A (150 ± 490) (1100 ± 2500)

a Mean ± S.D. values in the same column with different letters are statistically different (P < 0.05). 936 H.-H. Hsu et al. / Food Chemistry 114 (2009) 933–938

Table 2 Contents of biogenic amines in dried milkfish products.

Sample No. of Contents of biogenic amine (mg/100 g) sources samples Puta Cad Try Phe Spd Spm His Tyr Agm Tainan 8NDb -1.9 3.1-8.5 ND-5.1 ND ND ND-4.0 15.4-99.9 ND ND-3.3 county (1.0 ± 0.5)c (5.5 ± 1.7) (1.1 ± 2.0) (2.8 ± 1.6) (32.5 ± 27.1) (0.4 ± 1.1) Tainan city 6 ND-4.3 0.3-8.6 ND ND ND ND-5.0 ND-38.2 ND-1.5 ND -A (0.7 ± 1.6) (2.7 ± 2.8) (3.5 ± 1.7) (18.2 ± 14.1) (0.3 ± 0.6) Tainan city 3 ND-13.2 43.6-147.4 ND ND-2.9 ND-0.9 2.0-2.5 233.2-544.9 6.1-11.4 ND –B (4.4 ± 6.2) (94.9 ± 42.3) (1.4 ± 1.2) (0.7 ± 0.2) (2.3 ± 0.2) (409.7 ± 130.5) (8.5 ± 2.2) Tainan city 5 0.2–0.4 ND-1.6 ND ND-0.7 ND ND-2.5 0.5–1.2 (0.8 ± 0.3) ND ND –C (0.2 ± 0.1) (0.4 ± 0.6) (0.1 ± 0.3) (1.4 ± 0.8) Kaohsiung 10 ND-5.6 22.6–70.2 ND ND-0.4 ND-1.2 1.0–3.4 114.0–531.5 ND-11.3 ND-4.2 county (1.4 ± 1.9) (40.8 ± 13.5) (0.1 ± 0.1) (0.3 ± 0.4) (1.9 ± 0.7) (240.9 ± 115.8) (4.8 ± 4.4) (0.6 ± 1.3)

a Put: putrescine; Cad: cadaverine; Try: tryptamine; Phe: 2-phenylethylamine; Spd: spermidine; Spm: spermine; His: histamine; Tyr: tyramine; and Agm: agmatine. b ND: not detected (amine level less than 0.1 mg/100 g). c Means ± S.D.

Table 3 between pH and salt content (r = À0.78, p < 0.05), salt content Distribution of the histamine contents in 32 dried milkfish products. and APC (r = À0.73, p < 0.05), and Aw and salt content (r = À0.72, p < 0.05). Content of histamine (mg/100 g) Dried milkfish products Table 4 lists the identity of thirty histamine-forming bacteria No. of samples % of samples isolated from the dried milkfish products as determined by 16S < 4.9 7 21.9 rDNA sequences using NCBI database analysis. Thirty histamine- 5.0–49.9 11 34.4 producing bacterial strains were capable of producing 5.4 ppm to 50.0–99.9 1 3.1 >100 13 40.6 562 ppm of histamine in trypticase soy broth supplemented with Total 32 100 1.0% L-histidine (TSBH). They were identified as E. aerogenes (seven strains) and Citrobacter sp. (one strain) for prolific histamine form- ers (>553 ppm), and Staphylococcus xylosus (ten strains), S. sciuri semipreserved Spanish anchovies that were implicated in an inci- (one strain), B. thuringiensis (two strains), C. freundii (five strains), dent of scombroid poisoning. Marlin implicated in a poisoning inci- K. pneumoniae (one strain) and E. cloacae (three strains) for weak dent had a histamine content ranging between 93.5 and 276 mg/ histamine formers (<51.4 ppm) (Table 4). Histamine-forming bac- 100 g (Morrow, Margolies, Rowland, & Robert, 1991). In Taiwan, teria were more frequently isolated from Kaohsiung county sam- incidences of scombroid poisoning have only occurred occasion- ples (17/30, 56.7% of isolates) than from Tainan county (6/30, ally, and the fish implicated in those occasional outbreaks are tuna, 20% of isolates), Tainan city-A (3/30, 10% of isolates), Tainan city- mackerel, and black marlin (Chen & Malison, 1987; Tsai, Kung B (2/30, 6.7% of isolates), and Tainan city-C (2/30, 6.7% of isolates) et al., 2005a). Sailfish and marlin fillets have recently become the ones. Based on the finding that the higher levels of histamine, APC, most frequently implicated fish species in scombroid outbreaks E. coli, TC and histamine-forming bacteria strains occurred in Kaoh- in Taiwan (Hwang et al., 1995; Hwang et al., 1997). Recently, the siung county samples, we postulate that these samples had been high content of histamine at 61.6 mg/100 g in the suspected dried seriously contaminated during food preparation and processing. milkfish product could be the etiological factor for this fish borne However, two strains of S. xylosus isolated from Tainan city-B sam- poisoning in Taiwan (Tsai et al., 2007). Therefore, it is also very ples containing high levels of histamine produced small amounts important for people, especially those from the Indo-Pacific region of histamine in culture broth, indicating that they are not the main such as the Philippines, Indonesia and Taiwan, to be aware that contributors to histamine accumulation in Tainan city-B samples. milkfish products could become a hazardous food item in causing It was possible that the major histamine-forming bacteria that con- histamine poisoning. tributed to the high levels of histamine in Tainan city-B samples Although high content of histamine was detected in dried milk- were killed or inhibited by low water activity and content. More- fish products tested in this study, only very few cases of food-borne over, the use of poor quality fish as raw material for processing histamine intoxication were reported due to consumption of dried milkfish products. Symptoms of histamine poisoning are not par- Table 4 ticularly definitive. Therefore, histamine intoxication is frequently Identification of histamine-forming bacteria isolated from the dried milkfish products confused diagnostically with an allergic reaction. Also, histamine by 16S rDNA analysis, and their production of histamine (ppm) in culture broth. poisoning is not a reportable illness even in those countries that Source Organism identified Number Histamine keep surveillance records (Taylor, 1986). High levels of putrescine isolated (ppm) and cadaverine were also found in dried milkfish products in this Tainan county Staphylococcus xylosus 4 9.9–10.6 (10.2 ± 0.3)a study (Table 2). Putrescine and cadaverine have been shown to Bacillus thuringiensis 2 10.0–19.1 (14.6 ± 4.6) potentiate histamine toxicity when present in spoiled fish by Tainan city -A S. xylosus 3 9.8–10.0 (9.9 ± 0.1) inhibiting the intestinal histamine metabolizing enzyme (Antoine Tainan city -B S. xylosus 2 9.8–9.9 (9.9 ± 0.1) et al., 2002). Tainan city -C Citrobacter freundii 1 5.6 Pearson correlation was conducted to determine if there existed Staphylococcus sciuri 1 6.0 Kaohsiung S. xylosus 1 51.4 any relationship among the pH, salt content, water content, water county Klebsiella pneumoniae 1 5.4 activity, TVBN, APC, E. coli, TC, and histamine contents of the 32 C. freundii 4 5.5–20.8 (13.4 ± 7.0) samples tested. In general, the positive correlations existed be- Enterbacter cloacae 3 18.1–23.5 (20.8 ± 2.2) tween Aw and APC (r = 0.85, p < 0.01), pH and histamine (r = 0.76, Enterbacter aerogenes 7 553–562 (558 ± 2.8) Citrobacter sp. 1 561 p < 0.05), pH and TVBN (r = 0.74, p < 0.05), and histamine and TVBN (r = 0.73, p < 0.05). However, the negative correlations were noted a Mean ± S.D. H.-H. Hsu et al. / Food Chemistry 114 (2009) 933–938 937 results in the presence of high levels of TVBN and histamine in in these products could lead to scombroid poisoning in consumers. Tainan city-B samples. While the bacterial isolates of S. xylosus (ten strains), S. sciuri In this study, most of histamine-forming isolates belonged to (one strain), Bacillus thuringiensis (two strains), C. freundii (five Enterobacteriaceae, such as: E. aerogenes, E. cloacae, Citrobacter sp. strains), K. pneumoniae (one strain) and E. cloacae (three strains) C. freundii, K. pneumoniae, while they accounted for 56.7% (17/30) were identified to be weak histamine-formers, the E. aerogenes (se- of the total histamine-forming isolates. Meanwhile, E. aerogenes ven strains) and Citrobacter sp. (one strain) isolates were proven to (seven strains) and Citrobacter sp. (one strain) isolated from Kaoh- be prolific histamine-formers with ability to produce >500 ppm siung county samples produced 553–561 ppm of histamine in TSBH histamine in TSBH medium. (Table 4), accounting for 26.7% of histamine-forming isolates. Enter- obacter spp. and Citrobacter spp. have often been isolated from var- Acknowledgements ious species of scombroid fish. E. aerogenes and Citrobacter sp. were most frequently reported prolific histamine formers in tuna (Lopez- The study was supported by the National Science Council, R.O.C. Sabater, Rodriguez-Jerez, Hernandez-Herrero, Roig-Sagues, & (Contract No. NSC 96-2313-B-022-002). Mora-Ventura, 1996), albacore (Kim et al., 2001) and sailfish (Tsai et al., 2004). Recently, the enteric bacterium, Raoultella (formerly References Klebsiella) ornithinolytica, isolated from dried milkfish implicated in a food-borne poisoning was a potent histamine-former capable Antoine, F. R., Wei, C. I., Otwell, W. S., Sims, C. A., Littell, R. C., Hogle, A. D., et al. of producing 1243 ppm of histamine in TSBH (Tsai et al., 2007). (2002). Analysis of biogenic amines and their precursor free amino acids in mahi-mahi (Coryphaena hippurus). Journal of Food Biochemistry, 26, 131–152. Staphylococcus spp. was the most frequent reported histamine- AOAC. (1995). Official methods of analysis of AOAC international. 16th ed. Arlington, former in fermented salted fish, accounting for nearly 50% of hista- VA: AOAC International. mine-forming microorganisms. They were usually shown to have Bartholomev, B. A., Berry, P. R., Rodhouse, J. C., & Gilhouse, R. J. (1987). Scombrotoxic fish poisoning in Britain: Features of over 250 suspected incidents from 1976 to powerful histamine-forming activity (Yatsunami & Echigo, 1991; 1986. Epidemiology and Infection, 99, 775–782. Yatsunami & Echigo, 1992). For example, Staphylococcus epidermi- Chen, L. C (1990). Aquaculture in Taiwan. London: Fishing News (Books) (pp. 119– dis and Staphylococcus capitis isolated from salted Spanish ancho- 137). vies produced more than 1,000 ppm and 400 ppm of histamine, Chen, K. T., & Malison, M. D. (1987). Outbreak of scombroid fish poisoning, Taiwan. American Journal of Public Health, 77, 1335–1336. respectively, in TSBH broth (Hernandez-Herrero et al., 1999). How- Chiou, T. K., Shiau, C. Y., & Chai, T. J. (1990). Extractive nitrogenous components of ever, the recently isolated Staphylococcus pasteuri from miso prod- cultured milkfish and tilapia. Nippon Suisan Gakkaishi, 56, 1313–1317. ucts in Taiwan was a weak histamine-former, producing only Cobb, B. F., Aoaniz, I., & Thompson, C. A. (1973). Biochemical and microbial studies on shrimp: Volatile nitrogen and amino nitrogen analysis. Journal of Food 28.1 ppm histamine in TSBH broth (Kung, Tsai, & Wei, 2007). Sim- Science, 38, 431–435. ilarly, the S. xylosus and S. sciuri, which were isolated in this study, FDA. (1992). Bacteriological analytical manual. Arlington, VA: AOAC International. were also weak histamine-formers, capable of producing 6.0– Hernandez-Herrero, M. M., Roig-Sagues, A. X., Rodriguez-Jerez, J. J., & Mora- Ventura, M. T. (1999). Halotolerant and halophilic histamine-forming bacteria 51.4 ppm of histamine in TSBH (Table 4). Since staphylococci are isolated during the ripening of salted anchovies. Journal of Food Protection, 62, one of the major microbial groups that inhabit human skin, it is 509–514. reasonable to expect that they would be transferred to food prod- Hwang, D. F., Chang, S. H., Shiau, C. Y., & Cheng, C. C. (1995). Biogenic amines in the flesh of sailfish (Istiophorus platypterus) responsible for scombroid poisoning. ucts through considerable human contact during food preparation Journal of Food Science, 60, 926–928. and processing. The Bacillus spp. isolates from salted anchovies Hwang, D. F., Chang, S. H., Shiau, C. Y., & Chai, T. (1997). High-performance liquid produced low levels of histamine at 10.5–12.4 ppm (Hernandez- chromatographic determination of biogenic amines in fish implicated in food poisoning. Journal of Chromatography B, 693, 23–30. Herrero et al., 1999). The recently isolated B. coagulans and B. Kim, S. H., Field, K. G., Morrissey, M. T., Price, R. J., Wei, C. I., & An, H. (2001). Source megaterium from fermented fish products in Taiwan were also and identification of histamine-producing bacteria from fresh and temperature- identified as weak histamine-forming bacteria (Tsai et al., 2006). abused albacore. Journal of Food Protection, 64, 1035–1044. Similarly, two B. thuringiensis isolates isolated from the Tainan Kim, S. H., Barros-Velazquez, J., Ben-Gigirey, B., Eun, J. B., Jun, S. H., Wei, C. I., et al. (2003). Identification of the main bacteria contributing to histamine formation county samples were also weak histamine-formers capable of pro- in seafood to ensure product safety. Food Science and Biotechnology, 12, ducing 10.0 and 19.1 ppm of histamine. 451–460. H. alvei, M. morganii and K. pneumoniae have been implicated as Kuhnert, P., Capaul, S., Nicolet, J., & Frey, J. (1996). Phylogenetic positions of Clostridium chauvoei and Clostridium septicum based on 16S rRNA gene causative organisms in the formation of toxicologically significant sequences. International Journal of Systematic Bacteriology, 46, 1174–1176. levels of histamine in outbreaks of scombroid fish poisoning. Kuhnert, P., Heyberger-Meyer, B., Nicolet, J., & Frey, J. (2000). Characterization of Among them, M. morganii has consistently been shown to form PaxA and Its operon: A cohemolytic RTX toxin determinant from pathogenic Pasteurella aerogenes. Infection and Immunity, 68, 6–12. high levels of histamine (>1,000 ppm) in culture broth and isolated Kung, H. F., Tsai, Y. H., & Wei, C. I. (2007). Histamine and other biogenic amines and from spoiled fish stored above 15 °C (Lopez-Sabater, et al., 1996). histamine-forming bacteria in miso products. Food Chemistry, 101, 351–356. Nevertheless, the above major histamine-former (M. morganii) Lopez-Sabater, E. I., Rodriguez-Jerez, J. J., Hernandez-Herrero, M., Roig-Sagues, A. X., & Mora-Ventura, M. A. T. (1996). Sensory quality and histamine formation was not isolated from dried milkfish products in this study. Subb- during controlled decomposition of tuna (Thunnus thynnus). Journal of Food uraj, Karunasagar, and Karunasagar (1984) demonstrated that M. Protection, 59, 167–174. morganii and other histamine formers occur frequently in ice, Morrow, J. D., Margolies, G. R., Rowland, J., & Robert, L. J. (1991). Evidence that histamine is the causative toxin of scombroid-fish poisoning. New England water used for wetting fish, fish carrying baskets and market floors. Journal of Medicine, 324, 716–720. From our results, M. morganii does not seem a prevalent microor- Niven, C. F., Jeffreg, M. B., & Corlett, D. A. (1981). Differential plating medium for ganism in the processing environment of dried milkfish in Taiwan. quantitative detection of histamine-producing bacteria. Applied and Environmental Microbiology, 41, 321–322. Okuzumi, M., Hiraishi, A., Kobayashi, T., & Fujii, T. (1994). Photobacterium histaminum sp Nov., a histamine-producing marine bacterium. International 4. Conclusion Journal of Systemic Bacteriology, 44, 631–636. Subburaj, M., Karunasagar, I., & Karunasagar, I. (1984). Incidence of histidine- This study, to determine the safety of 32 dried milkfish products decarboxylating bacteria in fish and market environs. Food Microbiology, 1, 263–267. sold in Taiwan, showed that the TVBN content in 22 samples Taylor, S. L. (1986). Histamine food poisoning: Toxicology and clinical aspects. (68.8%) exceeded the 30 mg/100 g decomposition limit. The hista- Crititical Reviews in Toxicology, 17, 91–117. mine contents in most of the dried milkfish products exceed the Taylor, S. L. (1989). Histamine poisoning (scombroid fish poisoning): An allergy-like intoxication. Clinical Toxicology, 27, 225–240. 5 mg/100 g USFDA guideline value, and 40.6% (13/32) of the tested Taylor, S. L., & Speckard, M. (1983). Isolation of histamine-producing bacteria from samples contained >100 mg/100 g of histamine. Consumption of frozen tuna. Marine Fisheries Review, 45, 35–39. 938 H.-H. Hsu et al. / Food Chemistry 114 (2009) 933–938

Tsai, Y. H., Kung, H. F., Lee, T. M., Lin, G. T., & Hwang, D. F. (2004). Histamine related Tsai, Y. H., Kung, H. F., Chen, H. C., Chang, S. C., Hsu, H. H., & Wei, C. I. (2007). hygienic qualities and bacteria found in popular commercial scombroid fish Determination of histamine and histamine-forming bacteria in dried milkfish fillets in Taiwan. Journal of Food Protection, 67, 407–412. (Chanos chanos) implicated in a food-born poisoning. Food Chemistry, 105, Tsai, Y. H., Kung, H. F., Lee, T. M., Chen, H. C., Chou, S. S., Wei, C. I., et al. (2005a). 1289–1296. Determination of histamine in canned mackerel implicated in a food born Yatsunami, K., & Echigo, T. (1991). Isolation of salt tolerant histamine-forming poisoning. Food Control, 16, 579–585. bacteria from commercial rice-bran pickle sardine. Bulletin of the Japanese Tsai, Y. H., Lin, C. Y., Chang, S. C., Chen, H. C., Kung, H. F., Wei, C. I., et al. (2005b). Society of Scientific Fisheries, 57, 1723–1728. Occurrence of histamine and histamine-forming bacteria in salted mackerel in Yatsunami, K., & Echigo, T. (1992). Occurrence of halotolerant and halophili Taiwan. Food Microbiology, 22, 461–467. histamine-forming bacteria in red meat fish products. Bulletin of the Japanese Tsai, Y. H., Chang, S. C., Kung, H. F., Wei, C. I., & Hwang, D. F. (2005c). Histamine Society of Scientific Fisheries, 58, 515–520. production by Enterobacter aerogenes in sailfish and milkfish at various storage temperatures. Journal of Food Protection, 68, 1690–1695. Tsai, Y. H., Lin, C. Y., Chien, L. T., Lee, T. M., Wei, C. I., & Hwang, D. F. (2006). Histamine contents of fermented fish products in Taiwan and isolation of histamine forming bacteria. Food Chemistry, 98, 64–70.