Identification and Oxidation Products of Gluconobacter Strains Isolated from Fruits and Flowers in Thailand

The intent of the study is to identify and screen the oxidative products of acetic acid bacteria (AAB) from fruits and flowers in Thailand. Twenty-four isolates of AAB isolated from 22 fruit samples and two flower samples were grouped and identified at the species level in the genus Gluconobacter on the basis of their phenotypic and chemotaxonomic characteristics including molecular techniques. They were divided into four groups and identified as G. frateurii (Group 1, four isolates), G. japonicus (Group 2, two isolates), G. thailandicus (Group 3, 6 isolates), and G. oxydans (Group 4, 12 isolates) using 16S-23S rRNA gene internal transcribed spacer regions (ITS) restriction analysis and 16S-23S rRNA gene ITS phylogenetic analysis. All isolates were screened for dihydroxyacetone (DHA) and L-sorbose production. The test isolates produced DHA ranged from 25.24 to 42.52 g/l at 30°C. G. oxydans isolate PHD-27 produced a large amount of DHA. The test isolates produced L-sorbose ranged from 15.77 to 39.68 g/l at 30°C. G. frateurii isolate PHD-30 produced a largest amount of L-sorbose.

DHA is the oxidative product of glycerol while L-sorbose is from D-sorbitol that is industrially produced by G. oxydans (Deppenmeier et al., 2002;Salusjärvi et al., 2004 ).The DHA is commonly used as a tanning agent in the cosmetics industry and constitutes a building block for several chemical compounds, e.g., methotrexate, which is used in the chemotherapeutical treatment of cancer patients (Claret et al., 1994;Mishra et al., 2008).L-Sorbose has been used as an intermediate for synthesis of vitamin C by the Reichstein process (Hancock and Viola, 2002).L-Sorbose is produced from D-sorbitol by D-sorbitol dehydrogenase, which is bound to the cell membrane (Kim et al., 1999;Shinagawa et al., 1982).
The aim of this study was to identify Gluconobacter strains isolated from fruits and flowers in Thailand based on 16S-23S rRNA gene ITS restriction analysis and 16S-23S rRNA gene ITS sequence along with the phenotypic and chemotaxonomic characterization including the screening of their oxidative products.

Isolation of acetic acid bacteria
AAB were isolated from fruits (22 samples) and flowers (two samples) collected in Thailand by an enrichment culture approach using glucose/ethanol/yeast extract (GEY) medium (Kommanee et al., 2008) (Table 1).An isolation source was incubated at pH 4.5 and 30 o C for 3-5 days in a liquid medium.When microbial growth was found, the culture was streaked onto a GEY-agar plate containing 0.3% CaCO 3 (w/v) (Yamada et al., 1976).AAB were selected as an acid-producing bacterial strain that formed a clear zone around the colony on the agar plate.

Phenotypic and chemotaxonomic characterization of AAB
Phenotypic characterization was carried out by incubating test strains at 30 o C and pH 6.8 for two days on glucose/yeast extract/peptone/glycerol (GYPG) agar (Kommanee et al., 2008).For Gram stain of bacterial cells was performed as described by Hucker & Conn, 1923.Physiological and biochemical characterizations were determined as previously reported (Asai et al., 1964;Gosselé et al., 1980).The cells of isolates grown in GYPG broths on a rotary shaker (150-200 rpm) at 30 o C for 24 h were collected for the chemotaxonomic characterization.Ubiquinone was analyzed by the method of Yamada et al. (1968) andTamaoka et al. (1983).
Multiple sequence alignments were performed with a program CLUSTAL X (version 1.83) (Thompson et al., 1997).Gaps and ambiguous bases were eliminated from calculation.Distance matrices for the aligned sequences were calculated by the two-parameter method of Kimura (1980).A phylogenetic tree based on 16S-23S rRNA gene ITS sequences was constructed by the neighbor-joining method (Saitou & Nei, 1987) with the program MEGA 4 (Tamura et al., 2007).The confidence values of individual branches in the phylogenetic tree were determined by using the bootstrap analysis (Felsenstein, 1985) based on 1,000 replications.

Screening of strains producing DHA from glycerol
Twenty-four isolates of Gluconobacter were cultivated in potato medium, which contained 10.0% potato extract, 0.5% glucose, 1.0% glycerol, 1.0% yeast extract and 1.0% peptone (all by w/v), at 30°C on a rotary shaker (200 rpm) for 24 h and transferred to a DHA production medium containing 5% glycerol and 1% yeast extract at pH 5.0 (all by w/v) and cultivated at 30°C on a rotary shaker (200 rpm) for four days.The quantitative analysis of DHA was made by diphenylamine reaction (Karklinya et al., 1982).The potential selected strain was cultivated in 10 ml potato medium for 24 h.One ml of the cultures (0.5 optical density at 600 nm) was transferred to 200 ml DHA production medium and incubated at 30°C on a rotary shaker (200 rpm) for 96 h.An aliquot was taken every 12 h for biomass evolution and quantitative DHA production was analyzed by diphenylamine reaction.

Screening of strains producing L-sorbose from D-sorbitol
Twenty-four isolates of Gluconobacter were cultivated in potato medium at 30°C on a rotary shaker (200 rpm) for 24 h and were transferred to L-sorbose production medium (Moonmangmee et al., 2000) incubated at 30°C on a rotary shaker (200 rpm) for 48 h.The quantitative analysis of L-sorbose was made by resorcinol reaction (Adachi et al., 2001).This assay was rapid and simple in determining the presence of ketoses as observed by the cherry-red color development.The sugar quantity could be estimated based on the standard graph of a series of known concentration of L-sorbose.

Identification of isolates
Twenty-four acetic acid bacteria were isolated from fruits and flowers collected in Thailand.They were Gram-negative, aerobic and rod-shaped.They produced catalase and grew on mannitol agar.They did not oxidize acetate and lactate and not grew on glutamate agar.They produced D-gluconate, 2-keto-D-gluconate and 5-keto-D-gluconate from D-glucose, but not 2,5-diketo-D-gluconate.Major ubiquinone was Q-10 (Table 2).They were assigned to the genus Gluconobacter (Kersters et al., 2006;Skerman et al., 1980;Tanasupawat et al., 2004) and were divided into four groups; Group 1 that contained four isolates of G. frateurii, Group 2 that contained two isolates of G. japonicus, Group 3 that contained six isolates of G. thailandicus and Group 4 that contained twelve isolates of G. oxydans.
Group 1 was composed of four isolates, PHD-30, PHD-31, PHD-66 and PHD-67 (Table 1).All the isolates produced dihydroxyacetone from glycerol and grew at 30 c C. Some strains did not grow at 37 o C.They produced acids from L-arabinose, D-galactose, D-glucose, meso-erythritol (weak), and D-xylose (weak) glycerol, maltose (weak), and sucrose but not from D-arabinose, dulcitol, D-fructose, lactose, D-mannose, D-mannitol, D-melibiose, L-rhamnose, raffinose, L-sorbose, and D-sorbitol.The isolates grown on D-arabitol, L-arabitol and meso-ribitol (weak) but not on meso-erythritol, being different from strains of G. oxydans and G. cerinus.Isolate PHD-30, a representative strain had 55.2 mol% of DNA G+C content.The isolates showed almost the same phenotypic characteristics as G. frateurii NBRC 3264 T (Table 2) (Yukphan et al., 2004;Kommanee et al., 2008).The isolates were located within the cluster of G. frateurii in the phylogenetic tree based on 16S-23S rRNA gene ITS sequences (Figure 1) and had 99.9% pair-wise 16S-23S rRNA gene ITS sequence similarity to the type strain of G. frateurii.Isolate PHD-30 gave the same restriction patterns as the type strain of G. frateurii when digested with BstNI, MboII and MboI (Figure 2).From the data obtained above, all the isolates grouped into Group 1 were identified as G. frateurii.
Group 2 was composed of two isolates, PHD-28 and PHD-29 (Table 1).They produced acids from D-glucose and meso-erythritol (weak), and in some cases, from L-arabinose, D-galactose, glycerol, D-mannose, D-melibiose, sucrose and D-xylose.Some isolates produced acids weakly from D-arabinose, dulcitol, D-fructose, lactose and D-mannitol, but none produced acids from maltose, raffinose, L-rhamnose, L-sorbose or D-sorbitol.They grew on meso-erythritol, D-arabitol, L-arabitol (weak) and meso-ribitol.The isolates produced DHA weakly from glycerol.All the isolates were distinguished from G. frateurii by the ability to grow on meso-erythritol and produced acid weakly from raffinose.Isolates PHD-28, a representative isolate had 56.2 mol% of DNA G+C content.They showed almost the same phenotypic characteristics as G. japonicus NBRC 3271 T (Table 2) (Malimas et al., 2009).All isolates were located within the cluster of G. japonicus in the phylogenetic tree based on 16S-23S rRNA gene ITS sequences (Figure .1) and had 99.8% pair-wise 16S-23S rRNA gene ITS sequence similarity to the type strain of G. japonicus.Isolate PHD-28 gave the same restriction patterns as the type strain of G. japonicus when digested with BstNI, MboII and MboI (Figure 2).From the data obtained above, all the isolates grouped into Group 2 were identified as G. japonicus.
Group 3 was composed of six isolates, PHD-11, PHD-21, PHD-22, PHD-36, PHD-39 and PHD-40 (Table 1).All isolates produced DHA from glycerol.They grew on D-arabitol, L-arabitol, meso-ribitol (weakly) and meso-erythritol, but not on dulcitol.They produced acids from L-arabinose, D-fructose, D-galactose, D-glucose, glycerol, D-mannitol (variable and weak), D-ribose, L-sorbose (weak), D-xylose and ethanol.In contrast, no acid production was found from maltose, D-mannose, melibiose, raffinose, L-rhamnose, D-sorbitol and sucrose.They grew on mannitol agar but not on glutamate agar.The isolates grow at 37ºC.Isolate PHD-21, a representative strain had 57.2 mol% DNA G+C content.They showed almost the same phenotypic characteristics as the type strain of G. thailandicus (Table 2) (Tanasupawat et al., 2004;Kommanee et al., 2008).All isolates were located within the cluster of G. thailandicus in the phylogenetic tree based on 16S-23S rRNA gene ITS sequences (Figure 1) and had 99.9% pair-wise 16S-23S rRNA gene ITS sequence similarity to the type strain of G. thailandicus.Isolate PHD-11 gave the same restriction patterns as the type strain of G. thailandicus when digested with BstNI (Figure 2).From the data obtained above, all the isolates grouped into Group 3 were identified as G. thailandicus.
In this study, four isolates were identified as G. frateurii, two isolates were as G. japonicus, six isolates were as G. thailandicus and 12 isolates were as G. oxydans.There were distributed in fruits and flowers in Thailand (Table 1).Yamada et al. (1999) reported the isolates of G. oxydans and G. frateurii, and there were no the isolates of G. cerinus from the Indonesian sources.Tanaka et al. (1999) reported three species, G. cerinus, G. asaii, and G. frateurii as the low G+C contents isolates and G. oxydans as the high G+C contents isolates from Japanese sources.Huong et al. (2007) identified 44 Thai Gluconobacter isolates and grouped into seven groups, but there were no G. japonicus and G. thailandicus strains.In Gluconobacter isolates from fruits, flowers and other materials collected in Thailand, Kommanee et al. (2008) identified 17 isolates as G. oxydans, 12 isolates as G. cerinus, nine isolates as G. frateurii, six isolates as G. thailandicus and one as an unidentified isolate.
From, this result the Gluconobacter isolates was successfully identified at the species level by the phenotypic and chemotaxonomic characterization including the 16S-23S rRNA gene ITS restriction analyses using BstNI, MboII and MboI and 16S-23S rRNA gene sequence were useful.

Dihydroxyacetone production
Twenty-four isolates of Gluconobacter were tested for the ability to produce DHA.Initial screening was tested based on the diphenylamine method.Production of DHA was observed by blue color development.The sugar quantity can be estimated based on the standard graph of a series of known concentration of DHA.It was found that the twenty-four isolates produced a large amount of DHA ranged from 25.24 to 42.52 g/l at 30°C for four days (Table 1).Isolate PHD-27 identified as G. oxydans showed the highest DHA production reached to 42.52 g/l and was therefore selected for DHA production.
In DHA production, it is clear that the DHA concentration increased with the decrease of glycerol.The DHA concentration reached the maximum of 44.1 g/l at 30°C by conversion time of 84 h and generated DHA at a rate of 0.52 g/l/hr (Figure 3).Interestingly, the isolate PHD-27 showed good growth without any lag phase and produced DHA rapidly after the late log phase.The data suggested that G. oxydans isolate PHD-27 could produce DHA higher than the strain of G. oxydans improved by genetic engineering that produced DHA reached 30 g/l, while the wild type strains produced 18-25 g/l of DHA from 50 g/l glycerol (Kim et al., 1999).These data suggested that G. oxydans isolate PHD-27 suitable for using DHA production.

L-sorbose production
The twenty-four isolates of Gluconobacter were preliminary screened for their ability to produce L-sorbose using the resorcinol assay method.They produced a large amount of L-sorbose ranged from 15.77 to 39.68 g/l at 30°C for 48 h (Table 1).The pH of the culture broth came down from 6.3 to 4.7 (data not shown).In the part of L-sorbose production, the fermentation efficiency and the fermentation rate by G. frateurii isolate PHD-30 was quite high, and the isolate rapidly oxidized D-sorbitol to L-sorbose at almost 100% within 24 h at 30°C.However, the amount of produced L-sorbose was less than that (50.0 g/l by G. frateurii isolate CHM54 for 48 h) reported by Moonmangmee et al. (2000).

Conclusion
The present study revealed that 24 isolates from flowers and fruits collected in Thailand belonged to the genus Gluconobacter and were identified as G. frateurii, G. japonicus, G. thailandicus and G. oxydans routinely by phenotypic and chemotaxonomic characterizations, and finally by the 16S-23S rRNA gene ITS restriction analysis and the 16S-23S rRNA gene ITS phylogenetic analysis.All isolated Gluconobacter species produced DHA from glycerol and L-sorbose from D-sorbitol.Of the isolates, G. oxydans isolate PHD-27 and G. frateurii isolate PHD-30 produced a large amount of DHA and L-sorbose, respectively.

Figure 1 .
Figure 1.Phylogenetic relationships of isolates of Groups 1 to 4 based on 16S-23S rRNA gene ITS sequences.The phylogenetic tree was constructed by the neighbor-joining method.Numbers at nodes indicate bootstrap percentages derived from 1,000 replications