Morganella psychrotolerans sp. nov., a histamine-producing bacterium isolated from various seafoods

Morganella was described by Fulton (1943) and in 1978 it was established as a genus within the Enterobacteriaceae (Brenner et al., 1978). Strains previously known as Proteus morganii now belong to the genus Morganella; the two species have the same type strain, with Morganella morganii having priority. The genus Morganella comprises a single species (Morganella morganii) that includes two subspecies (Morganella morganii subsp. morganii and Morganella morganii subsp. sibonii), which can be distinguished from one another on the basis of trehalose fermentation (Jensen et al., 1992). In addition, M. morganii isolates have been divided into biogroups depending on glycerol fermentation, lysine- and ornithine-decarboxylase activities and tetracycline sensitivity (Jensen et al., 1992; Janda & Abbott, 2005). M. morganii is often found in the intestines of humans and animals. It grows between 4 and 45 °C and has the ability to produce toxic concentrations of histamine in seafood, but only at storage temperatures above 7–10 °C (Lehane & Olley, 2000; Janda & Abbott, 2005). Recently, psychrotolerant M. morganii-like bacteria have been isolated from fresh tuna, cold-smoked tuna and garfish. The psychrotolerant isolates produce toxic concentrations of histamine at 0–5 °C and appear to be important in histamine formation in chilled seafood. In fact, these strains have been isolated from seafood implicated in incidents of histamine fish poisoning (Emborg et al., 2005; Emborg & Dalgaard, 2006; Dalgaard et al., 2006). The present study includes a polyphasic characterization with multi-locus sequencing of psychrotolerant Morganella isolates.

Nine mesophilic M. morganii and 13 psychrotolerant M. morganii-like isolates were studied including M. morganii subsp. morganii LMG 7874^T and M. morganii subsp. sibonii DSM 14850^T (Table 1⇓). As shown in Table 1⇓, the isolates originated from different geographical locations, date of isolation and habitats, including different seafoods. Prior to sequence analysis all strains were cultured overnight at 25 °C on blood agar and prior to phenotypical testing on Long and Hammer agar (van Spreekens, 1974) with 1 % NaCl.

A partial sequence of the 16S rRNA gene (1356 bp) and fragments from seven protein-encoding housekeeping genes: atpD (β-subunit of ATP synthase; 509 bp), dnaN (DNA polymerase III β-subunit; 582 bp), gyrB (DNA gyrase; 443 bp), hdc (histidine decarboxylase; 576 bp), infB (translational initiation factor; 467 bp), rpoB (β-subunit of DNA-dependent RNA polymerase; 454 bp) and tuf (elongation factor; 526 bp), were selected for use in classification and identification of the isolates. Gene fragments were amplified by PCR from chromosomal DNA, with denaturation at 94 °C for 2 min, 35 cycles of 94 °C for 1 min, 1 min at the appropriate annealing temperature, 72 °C for 1.5 min and finally an extension step at 72 °C for 10 min, in a reaction volume of 50 μl. The primers used were designed for the present study, except the hdc and dnaN primers, which were used previously by Takahashi et al. (2003) and Torpdahl et al. (2005), respectively. Primer sequences and annealing temperatures are given in Supplementary Table S1 in ISJEM Online. Amplified gene fragments were purified using QIAquick spin columns (Qiagen) and sequenced on an ABI377 automatic sequencer using a Prism BigDye Terminator Cycle sequencing kit (Applied Biosystems). The 16S rRNA gene was sequenced by using standard methods, as described by L'Abee-Lund et al. (2003). DNA base composition (G+C content) and DNA–DNA hybridization studies using strains U2/3^T, LMG 7874^T, DSM 14850^T, FD24, JB-T16, 03A11 and 1F10 were performed at Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ) GmbH, Braunschweig, Germany. DNA–DNA hybridization was carried out as described by De Ley et al. (1970) and Huß et al. (1983) using a model Cary 100 Bio UV/VIS-spectrophotometer equipped with a Peltier-thermostatted 6×6 multi-cell changer and a temperature controller with an in situ temperature probe (Varian). Alignments and phylogenetic analysis were made using the software mega version 3.1 (Kumar et al., 2004). Distances and clustering were determined using parsimony maximum and neighbour-joining methods. Distance matrices were calculated by using Kimura's two-parameter distances (Gascuel, 1997). Reliability of topologies was assessed by bootstrap analysis with 1000 replicates. Reference 16S rRNA gene sequences for Proteus and Providencia species were obtained from GenBank ().

Phenotypic characterization of the strains was conducted by using commercially available kits: API 20E and API 50 CH-E (bioMérieux), according to the manufacturer's protocols. Other standard tests were performed as described previously (Emborg et al., 2005) and by Barrow & Feltham (1999), but were conducted at 25 °C instead of 30–37 °C to obtain growth of the psychrotolerant isolates. Growth at various temperatures (0, 2, 4, 25, 30, 35 and 37 °C) was determined using a nutrient-rich broth with 1.0 % NaCl (Dalgaard et al., 1994). The same broth with Tris/HCl instead of phosphate buffer was used to determine growth at various pH values (pH 4.1–9.6). Growth at pH values above 8 was determined using Hungate tubes with butyl-rubber stoppers to prevent CO₂ from air reducing the pH of the medium during the experiment. Growth with 0–10 % (w/v) NaCl was tested in broth containing 1 % tryptone and 0.5 % yeast extract. The presence of tyrosine lyase was tested at 25 and 30 °C instead of 37 °C as suggested by Sheth & Kurup (1975). Formation of a reddish-brown pigment from tryptophan was tested according to Muller (1985). The test was performed at 25 °C. The antibiotic sensitivity of the isolates was tested using methods from the Clinical and Laboratory Standards Institute with commercially prepared, dry-form broth micro-dilution panels (Sensititre; TREK Diagnostics) and Mueller–Hinton broth (National Committee on Clinical Laboratory Standards, 2002, 2003; Clinical and Laboratory Standards Institute, 2005). The whole-cell fatty acid composition was determined for strains U2/3^T, LMG 7874^T, DSM 14850^T, 1F10 and 03A11 at DSMZ, using gas chromatography according to MIDI Inc. Histamine formation by the isolates in broth at 10 °C was tested as described previously (Emborg et al., 2005).

Phylogenetic analysis of almost-complete 16S rRNA gene sequences (1356 bp) differentiated psychrotolerant M. morganii-like isolates from mesophilic M. morganii, Proteus and Providencia isolates (Fig. 1⇓). The similarity between group means of the psychrotolerant and mesophilic Morganella isolates was 98.6 %, which is above the 97 % similarity generally used to distinguish two groups of isolates as representing separate species (Stackebrandt & Goebel, 1994). However, among some taxa, such as the Enterobacteriaceae, 16S rRNA gene sequence similarity well beyond 97 % is common (Cilia et al., 1996). Protein-encoding genes frequently show more sequence variation than 16S rRNA genes and multi-locus sequencing of such housekeeping genes can be valuable for classification and identification of isolates (Enright & Spratt, 1999; Giraffa, 2001; Stackebrandt et al., 2002). Phylogenetic analysis of merged sequences of all seven housekeeping gene fragments showed 88.7 % similarity between group means of the psychrotolerant and mesophilic Morganella isolates (Fig. 2⇓). Furthermore, the gene fragments from each of the seven studied housekeeping genes clearly separated the psychrotolerant and mesophilic Morganella isolates (Supplementary Figs S1–S7 in IJSEM Online). The sequence similarity of each of the housekeeping gene fragments was less than 90.9 % and bootstrap values of 100 for all of the seven housekeeping gene fragments as well as for the merged sequences supported the separate clustering of psychrotolerant and mesophilic Morganella isolates. The degree of sequence similarity for psychrotolerant M. morganii-like isolates corresponded to the degree of similarity observed between the mesophilic Morganella isolates studied (Fig. 2⇓).

DNA–DNA hybridization showed a low relatedness between the psychrotolerant isolate U2/3^T and the type strains of both M. morganii subsp. morganii (LMG 7874^T; 41.0 %) and M. morganii subsp. sibonii (DSM 14850^T; 17.8 %). The DNA–DNA relatedness was lower than that of 51.0 % found between M. morganii subsp. morganii LMG 7874^T and M. morganii subsp. sibonii DSM 14850^T (Table 2⇓). This low level of DNA–DNA relatedness between the psychrotolerant isolate and M. morganii supports that the strains represent separate species (Wayne et al., 1987). The relatedness values between strain U2/3^T and the psychrotolerant isolates FD24, JB-T16 and 1F10 were 85.3, 97.0 and 67.7 %, respectively (Table 2⇓). The mesophilic strain 03A11 showed significantly higher DNA–DNA relatedness to strain LMG 7874^T than to strain DSM 14850^T (Table 2⇓). This is in conflict with the strain's fermentation of trehalose, indicating that strain 03A11 represents M. morganii subsp. sibonii. Strains 03A11 and 03B10 from smoked fish from New Zealand thus appear to represent a separate cluster of M. morganii. This was supported by the phylogenetic analysis of 16S rRNA and housekeeping genes (Fig. 2⇑ and Supplementary Figs S1–S7 in IJSEM Online). The other mesophilic M. morganii isolates were divided into two clusters corresponding to M. morganii subsp. morganii and M. morganii subsp. sibonii. However, the tuf gene fragment resulted in a slightly different clustering for M. morganii subsp. sibonii DSM 14850^T (Fig. 2⇑ and Supplementary Figs S1–S7 in IJSEM Online).

The DNA G+C contents for M. morganii subsp. morganii LMG 7874^T and M. morganii subsp. sibonii DSM 14850^T were identical (49.6 mol%) and similar to that of strain 03A11 (50.7 mol%), whereas slightly lower values of 46.5 and 47.1 mol% were found for the psychrotolerant isolates U2/3^T and 1F10, respectively.

Phenotypic characterization confirmed that all the isolates studied belonged to the genus Morganella (Janda & Abbott, 2005). Cells were Gram-negative, fermentative, motile rods, with negative oxidase and positive catalase reactions. The isolates were phenylalanine deaminase-positive, gelatin-negative, urease-positive (except strain 25a32), indole-positive (except strain JB-T11) and citrate-negative (except strain 1F10). They all degraded tyrosine crystals after 72 h at 25 and 30 °C (except strains 2F6 and JB-T11) and produced a reddish-brown pigment from tryptophan after 48 h at 25 °C (except strain 2F6). With regard to the whole-cell fatty acids, the Morganella strains all contained higher concentrations of 12 : 0 and lower concentrations of 18 : 1ω7c (Supplementary Table S2 in IJSEM Online) than usually observed for Proteus and Providencia (Vasyurenko & Chernyavskaya, 1990). This further supported their allocation to the genus Morganella.

As observed previously (Emborg et al., 2005; Emborg & Dalgaard, 2006; Dalgaard et al., 2006), the psychrotolerant M. morganii-like isolates differed from mesophilic M. morganii by their ability to grow at 0–2 °C, but not at 37 °C. In addition, fermentation of d-galactose and growth in 8.5 % (w/v) NaCl support this differentiation (Table 3⇓). At 10 °C, high concentrations of histamine were produced by all the psychrotolerant M. morganii-like isolates (6420–10 980 p.p.m.) as well as by the M. morganii isolates (5000–9120 p.p.m.). It has been shown previously that various mixtures of psychrotolerant M. morganii-like isolates produced 7400±1050 p.p.m. histamine in vacuum-packed tuna steaks after storage at 2 °C. At 5 °C, 3500–4000 p.p.m. histamine was formed in vacuum-packed, cold-smoked tuna and 1190–3310 p.p.m. histamine in broth (Emborg et al., 2005; Emborg & Dalgaard, 2006; Dalgaard et al., 2006). Such a pronounced histamine formation at 2–5 °C has not been reported for mesophilic M. morganii isolates.

In the present study, all M. morganii subsp. morganii isolates tested were sensitive to tetracycline, whereas M. morganii subsp. sibonii isolates were all resistant. This is in agreement with earlier findings by Janda & Abbott (2005). Jensen et al. (1992) used tetracycline resistance among other criteria to divide Morganella into biogroups. Interestingly, we observed the same pattern for the psychrotolerant M. morganii-like isolates and the mesophilic M. morganii isolates, with d-trehalose-fermenting isolates being resistant to tetracycline. Only strain 1F10 was resistant to tetracycline and unable to ferment d-trehalose. Janda & Abbott (2005) reported that adonitol and d-arabitol fermentation performed using the conventional test differed between M. morganii subsp. morganii and M. morganii subsp. sibonii, but this was not confirmed in the present study; however, in the present study, the test was performed using API 50 CH-E with results being recorded after 24 and 48 h at 25 °C (Supplementary Table S3 in IJSEM Online). Consequently, API 50 CH-E can be used for identification of M. morganii, but not for separation into subspecies, where conventional tests are needed.

In combination with the sequence data (Fig. 2⇑), the phenotypic characterization suggests that strain U2/3^T and the other psychrotolerant M. morganii-like isolates studied represent a novel Morganella species, for which the name Morganella psychrotolerans sp. nov. is proposed.

Similar to M. morganii, the psychrotolerant isolates studied varied with respect to d-trehalose fermentation, l-ornithine decarboxylase activity and tetracycline sensitivity (Table 3⇑ and Supplementary Table S3 in IJSEM Online). The nine psychrotolerant isolates from fresh and cold-smoked tuna were d-trehalose-negative and l-ornithine decarboxylase-positive and were able to grow at 0 °C. They differed from the two isolates from lumpfish roe (F39-1 and F39-3) in these phenotypic characteristics, as well as in the gene fragments studied (Fig. 2⇑ and Supplementary Figs S1–S7 in IJSEM Online). The two ‘garfish’ isolates also differed from the ‘tuna’ isolates in the gene fragments studied (Fig. 2⇑ and Supplementary Figs S1–S7 in IJSEM Online). In addition, strain 1F10 was the only psychrotolerant isolate that was able to grow in 8 % NaCl; it was also citrate-negative and showed only 67.7±2.3 % DNA–DNA relatedness with strain U2/3^T. Strain 2F6 differed from the ‘tuna’ isolates with respect to formation of pigment from tryptophan and tyrosine degradation at 30 °C. From the present study, it appears that a future division of M. psychrotolerans may be relevant, but analysis of a larger number of isolates would be required to determine whether a novel subspecies is appropriate or whether a novel species can be separated from M. morganii and M. psychrotolerans.

In this study, the usefulness of multi-locus sequencing to elucidate the genomic relatedness at inter- and intraspecific levels is illustrated. The sequence information obtained has improved the taxonomy of the genus Morganella and made data available for future studies of molecular identification and detection methods. The recognition and acknowledgement of M. psychrotolerans sp. nov. and its growth conditions will most probably be of significant importance for the investigation of future outbreaks of histamine fish poisoning.