Introduction

shubat is a special fermented product, prepared from unheated two-humped camel milk through indigenous fermentation process. Although it is more or less similar to yoghurt in appearance, there are important differences between these two products since shubat is liquid rather than creamy, sparking due to its CO2 production and has a high degree of sourness (pH of ca. 3.8). In the Jungghar baisin region of Xinjiang in China, shubat have been widely used both as beverage and folk medicine. The indigenous populations have believed that fermented camel milk (shubat) is safe and even has medicinal properties such as antidiabetic, anti-cancer and anti-tuberculosis activities.

Traditionally, shubat is homemade by using a semi-continuous or fed-batch fermentation process that was handed down throughout generations. Whenever part of the product is withdrawn for consumption, a portion of raw camel milk is added to make up volume and this process of retrieval and replacement of milk continues for months. The spontaneous fermentation of unheated milk takes advantage of natural microflora inherent in milk and environmental contaminants. However, many reports about traditional dairy products have shown that they have unique and different microflora dependent on the production technology as well as on the ecological localities where they have been produced (Zamfir et al. 2006; Dewan and Tamang 2007). It is well known that specific characteristics of dairy products depend on the present microflora (Wouters et al. 2002; Leroy and De Vuyst 2004). Therefore, this knowledge can contribute to improving the process of the shubat manufacture, as well as to the fermentation condition for obtaining the product of better quality. The present study aimed to isolate and identify predominant microflora in homemade shubat by using phenotypic and molecular methods, which can further developed starter cultures for camel milk fermentation. To the best of our knowledge, no similar investigation into microflora of shubat from two-humped camel milk in China has been carried out so far.

Materials and methods

Samples

Seven samples of shubat were collected from different nomadic families in Kanas and Brojin areas, the Jungghar baisin of Xinjiang in China. All the samples were homemade by traditional method in their gers (portable houses for nomads). Each shubat sample was aseptically transferred to a 200 ml sterile screw-capped bottle, transported immediately to the laboratories, and kept in refrigerator at 4°C before.

Enumeration and isolation of lactic acid bacteria and yeasts

After mixing 10 ml of each shubat sample was pipetted aseptically into 90 ml of sterile physiological saline (0.85% w/v). One hundred microlitre from tenfold dilutions of the samples was surface inoculated onto each of the following media: (1) MRS (Oxoid™) and tomato juice agar (TMJ, Oxoid™) incubated anaerobically for 48 h at 37°C for enumeration of LAB and (2) yeast extract-glucose-chloramphenicol-agar (YGCA, Merck™) incubated for 72 h at 25°C for enumeration of yeast. Colonies showing different appearance were randomly selected and further purified by successive streaking on MRS agar (for LAB) or PDA agar (for yeasts). For long term maintenance of isolates, stock cultures of LAB and yeasts were stored at −80°C in 20% (v/v) glycerol, with 80% (v/v) MRS and YPG broth (Oxoid™), respectively.

Identification of LAB isolates

A total of 48 LAB isolates were macro- and microscopically characterized. For Gram-positive, catalase-negative rods and cocci, growth at 10 and 45°C, and carbon dioxide production from glucose in MRS broth (with inverted Durham tubes) were further identification, according to the methods of Harrigan and McCance (1976). Lactic acid isomer was determined enzymatically using d-lactate and l-lactate dehydrogenase test kits (Roche Diagnostic, France), according to the manufacturer’s instructions.

Carbohydrate fermentation patterns of LAB isolates were determined using the API 50 CHL (bioMérieux, France), which enabled identification of the isolates to species level. Based on API identification, representative isolates of LAB were selected for sequencing of 16S rDNA as previously described by Najjari et al. (2008). In brief, genomic DNA was extracted using Taq DNA polymerase kit Wizard DNA purification Kit as described with the manufacturer (promega, Madison, WI USA). The extracted DNA (2 μl) was amplified by using the primers pA 5′ AGAGTTTGATCCTGGCTCAG 3′ and pH 5′ GGCTACCTTGTTACGACT 3′. The reaction mixture (50 μl) consists of 50 ng of bacterial DNA, 60 pmol of each primer, deoxynucleoside triphosphate at a concentration of 200 pM, 2.5 U of Taq DNA polymerase, 10 mM Tris–HCl (pH 9.0), 50 mM KCl, 1.5 mM MgCl2 and enough sterile deionized water to bring the volume to 25 μl. The reactions were carried out in PCR Thermocycler (Bio-Rad Mycycler™) programmed as follows: 94°C for 3 min; 35 cycles of 94°C for 15 s, 59°C for 60 s and 72°C for 2 min and finally, 72°C for 10 min.

Identification of yeast isolates

Fifteen isolates of yeast were tested for their ability to assimilate various compounds by using the ID 32 API kits (bioMérieux, France). Based on API groups, representatives of yeast isolates were selected for amplification of 18S-28S ITS region using the ITS1 forward primer (5′-TCC GTA GGT GAA CCT GCG G-3′) and TS4 reverse primer (5′-TCC TCC GCT TAT TGA TAT GC-3′) (White et al. 1990). The genomic DNA were extracted by using DNA kits for yeasts (Tiangen Biotech, Beijing, China). For each 1 μl of the prepared DNA, 49 μl of the reaction mixture were added. This mixture contained 5 μl of 10× PCR buffer (Pharmacia Biotech, Uppsala, Sweden), 4 μl of 2.5 mmol of each dNTP (Pharmacia Biotech), 4 μl of 25 mmol MgCl2, 3 μl of 50 pm of each of the primer, 0.5 of l% (v/v) formamide, 0.25 μl of 2.5 U Taq polymerase (Pharmacia Biotech), and 30.25 ml of Milli Q water (Milli-Q Plus, Ultra Pure Water System, Millipore, Molsheim, France). The reactions were carried out in PCR Thermocycler (Bio-Rad Mycycler™). The PCR conditions were: initial denaturation at 94°C for 5 min, 35 cycles of 98°C/15 s, 59°C/60 s and 72°C/2 min with a final extension of 72°C/10 min.

DNA sequencing and sequence analysis

DNA fragments amplified by PCR were separated using 2% agarose gel electrophoresis and purified by QIAquick PCR Purification kit/250 (QIAGEN GmbH, Hilden, Germany). The sequencing of purified products was performed by Sangon Company, Shanghai, China. The BLAST algorithm was used to determine the most related sequence relatives in the NCBI nucleotide sequence database (www.ncbi.nlm.nih.gov/blast).

Statistical analysis

All experiments were performed in triplicate and the results were expressed as the mean value ± the standard deviation. All microbiological counts were converted to the base-10 logarithm of cfus per milliliter (ml) of shubat samples (log cfu/ml), and from these, means and their standard deviations were calculated. Data were tested for statistical significance by the analysis of variance (ANOVA) using the Statistical Analysis System software (SAS version 9.00, SAS Institute, Inc. 2000), P < 0.05.

Results and discussion

Microbial counts and pH

Table 1 showed that lactic acid bacteria and yeasts constituted the predominant microflora of shubat. The four samples from Kanas areas contained LAB log 6.8 and 7.3 cfu/ml when determined on MRS and TMJ agar plates, respectively. Slightly higher LAB counts were observed in the three samples from Borjin areas (log 7.3 and 7.6 cfu/ml, enumerated on MRS and TMJ, respectively). Yeasts were detected in all shubat samples with counts ranging between log 4.3 and 4.7 cfu/ml (Table 1). The samples from Kanas areas had the higher yeast counts than those from Borjin areas. In addition, no mold and coliform bacteria were detected in final products of shubat (data not shown). The pH values of the seven investigated shubat samples collected from kanas and Brojin areas were found to be 4.1 and 3.7, respectively, indicating that shubat has a high degree of acidity.

Table 1 pH values and log counts of the shubat samples
Full size table

The microbial flora in shubat was comparable to those reported in previous studies on suusac (traditional Kenyan fermented camel milk) and Gariss (traditional Sudanese fermented milk product) (Lore et al. 2005; Abdelgadir et al. 2008). Lore et al. (2005) reported that the LAB and yeast counts in suusac were log 6.8 and 6.8 cfu/ml, respectively, while Abdelgadir et al. (2008) indicated that LAB and yeast counts in Gariss were between log 7.76 and 8.66 cfu/ml for LAB and between log 6.05 and 7.79 cfu/ml for yeasts.

Identification of LAB

A total of 48 isolates were Gram-positive and catalase-negative rod or cocci, which were presumptively considered to be LAB isolates. According to their observed features such as microscopy appearance, CO2 production from glucose, NH3 production from arginine, the ability to grow under different temperature and salt conditions, these isolates were grouped into six groups (Table 2). By use of API 50 CH system, the representative isolates from group I, II, III, IV and V were clearly identified as Lactobacillus sakei, Lb. helvelticus, Enterococcus feacium, Ec. feacalis, Lb. brevis and Leuconstoc lactis, respectively. Unfortunately two isolates from group VI (Le5-1 and Le5-2) were not identified by API 50 CH system (Table 2).

Table 2 Phenotypic and molecular identification of lactic acid bacteria groups isolated from shubat
Full size table

A set of representative isolates from different phenotypic groups were selected for 16S rDNA sequencing. The resulting16S rDNA sequences were aligned with all sequences present in the GenBank database and resulted in the final identification of the isolates from shubat. The API identification of all the isolates belonging to group I, II, III, IV, V were confirmed by 16S rDNA sequencing at 99% of sequence similarity (Table 2), respectively. The two unidentified isolates from group VI were identified as Weissella hellenica with 99% sequence similarity.

Identification of yeast

Based on Carbohydrate assimilation profiling using the API ID 32C kit, the 15 yeast isolates from seven shubat samples belonged to Kluyveromyces marxianus (perfect form of Candida kefyr), Kazachstania unispora (senior name of Saccharomyces unisporus) and Candida ethanolica, which was confirmed by sequencing of the 26S rDNA at 99% of sequence similarity (Table 3).

Table 3 Biochemical (API) and molecular identification of yeasts isolated from shubat
Full size table

Distribution of LAB and yeast microflora in shubat

As seen from Table 4, Lactobacillus clearly dominated the other genera (44% of total isolates), followed by Enterococcus (19%), Kluyveromyces (14%) and Leuconostoc (10%).

Table 4 Distribution of lactic acid bacteria and yeast species isolated from shubat samples
Full size table

Among the LAB micrfora, Lb. sakei were the most frequently isolated species (26% of total isolates, Table 4), which was found in all the seven investigated samples except for BJ1 sample and dominant in 5 of them (sample KS2, KS3, KS4, BJ2 and BJ3). Lb. helveticus was isolated from 5 out of 7 samples and Lb. brevis was isolated from 3 samples. Ec. faecium was found in five shubat samples where it was predominant species in BJ1 sample. Ec. feacalis was isolated from KS1 and KS2 as well as BJ2 samples. Ln. lactis was isolated from four samples (KS1, KS3, KS4, and BJ1), whereas two strains of Ws. hellenica was isolated from same samples (BJ1). The yeast micrfora in shubat were found to be Kluyveromyces marxianus and Kazachstania unispora as well as Candida ethanolica with Kl. marxianus being dominant yeast that was isolated from all the 7 samples (Table 4). Kazachstania unispora was detected from 5 samples, only one Cd. Ethanolica strain was isolated in BJ1 sample.

Abdelgadir et al. (2008) reported Streptococcus bovis group was predominant species in Gariss. Also Lb. fermentum were detected in high numbers, whereas Ec. faecium and Lb. helveticus were detected more occasionally in Gariss. Lore et al. (2005) reported that the frequently isolated species in suusac was Ln. mesenteroides (24% of total isolates), followed by Lb. plantarum. In our work, Lactobacillus clearly dominated the other LAB genera. However, Lb. sakei was the predominant species in shubat. To the best of our knowledge, it has never been reported on Lb. sakei present in fermented milks. Indeed, all the Lb. sakei isolates from shubat were homofermentative, fermenting lactose to produce lactic acid as sole product. This suggests their significant role in lactic fermentation of camel milk.

Enterococi are commonly isolated from dairy products, especially from Southern Europe (Giraffa 2003). Although the presence of enterococci in dairy products is controversial, many reports showed that enterococci play an important role in the ripening of cheeses due to their proteolytic and lypolitic activities and by the production of aromatic compounds from citrate (Moreno et al. 2006). This is a characteristic that would be of functional significance towards aroma development in shubat.

Weissella hellenica was isolated only from one sample meaning that these bacteria are probably present in low numbers. The species is occasionally detected in dairy products (Morea et al. 1998; Randazzo et al. 2006) and it is believed that their presence is a result of low sanitary conditions during the hand milking and handling.

Kluyveromyces marxianus and Kazachstania unispora have previously been isolated from indigenous fermented milk products such as Gariss, suusac, koumiss and kefir (Abdelgadir et al. 2001; Lore et al. 2005; Narvhus and Gadaga 2003; Shuangquan et al. 2004; Latorre-García et al. 2007). Kluyveromyces marxianus was isolated predominantly from Mongolian Airag and Saccharomyces cerevisiae, Issatchenkia orientalis and Kazachstania unispora were the predominant isolates from Mongolian Tarag (Watanabe et al. 2008). In the present study, Kl. marxianus and Kz. unisporus were frequently isolated and constitute the dominant yeast microflora of shubat. Surprisingly we did not detect Saccharomyces cerevisiae, Issatchenkia orientalis in all the seven shubat samples. Lachance indicated that all isolates of Kl. marxianus are capable of metabolizing lactose whereas Kz. unisporus isolates are lactose negative. Both species are able to metabolize lactate (Lachance 1998), Yeast growth in shubat is thus probably positively influenced by the metabolic activities of the LAB present (Narvhus and Gadaga 2003). Candida ethanolica was originally isolated from industrial fodder yeast cultivated on synthetic ethanol as the only source of carbon. This species differs from all recently accepted Candida species in not assimilating nitrate, not producing urease and not fermenting sugars. Thus, it is likely that the occurrence of Candida ethanolica in shubat does not have a functional significance in the fermentation process.

Conclusion

The present study has showed that the microflora in shubat comprises a combination of LAB and yeasts. The LAB were represented by Lactobacillus, Enterococcus, Leuconostoc and Weissella, with the most frequently isolated LAB being Lb. sakei, Ec. feacium and Lb. helveticus. Lactose-fremneting Kl. marxianus was found to be the predominant yeast species, the precise role of yeasts in shubat requires further study.