Introduction

The main components of dietary phytoestrogenic isoflavones are daidzein, genistein, and their glycosides, which are distributed mainly in leguminous plants, such as soybeans (Eldridge and Kwolek 1983; Wang and Murphy 1994). Isoflavones, which undergo extensive metabolism in the intestinal tract prior to absorption, are of great beneficial effects on the treatment of many diseases (Hussain et al. 2003; Lydeking-Olsen et al. 2004). The metabolism of genistein in humans has not been well characterized, but a lot of work has shown that daidzein metabolites, including dihydrodaidzein (DHD), equol, O-desmethylangolensin, etc., are more biologically active than their precursor daidzein (Chin-Dusting et al. 2001; Coldham et al. 1999; Jiang et al. 2003; Joannou et al. 1995). Intestinal bacteria play a pivotal role in daidzein metabolism. Particular bacteria responsible for daidzein metabolism have been isolated from the intestine of both humans and animals (Hur et al. 2002; Matthies et al. 2008; Minamida et al. 2006; Wang et al. 2005a). However, all of the reported daidzein metabolizing bacteria are obligate anaerobes, which are lack of the ability to grow and convert daidzein to different metabolites under aerobic conditions.

Obligate anaerobes cannot grow even in the presence of very small amount of oxygen (less than 0.5%) in the medium (Fredette et al. 1967; Loesche 1969; Rolfe et al. 1977). It is very difficult and expensive to keep obligate anaerobes because a strictly anaerobic workspace and storage facility are needed. Therefore, researchers began to search for some methods which can make obligate anaerobes grow in aerobic conditions. Some anaerobic campylobacters were reported to be able to grow in the presence of air after the provision of a humid environment (Syed et al. 1993). However, studies on improving oxygen tolerance activity of anaerobic bacteria are still quite limited. To date, oxygen tolerance study on anaerobic isoflavone biotransforming bacteria has not been performed.

In our previous study, from bovine rumen content, we have isolated an obligate anaerobic bacterium named Niu-O16 (AY263505), which is capable of bioconverting isoflavones daidzein and genistein to DHD and dihydrogenistein (DHG), respectively, under anaerobic conditions (Wang et al. 2005b). After a long-term domestication, we obtained from the original obligate anaerobic bacterium Niu-O16, an oxygen-tolerant bacterium, which we named Aeroto-Niu-O16. In the present study, we investigated the ability of strain Aeroto-Niu-O16 to grow under aerobic conditions and studied the bioconversion activity of both daidzein and genistein in the presence of atmospheric oxygen. The adaptation to aerobic conditions of strain Aeroto-Niu-O16 was accompanied by a series of changes in morphology, growth property, biochemical traits as well as point mutations in the 16S rRNA gene.

Materials and methods

Chemicals

Daidzein and genistein were purchased from Indofine (Somerville, NJ, USA). Brain heart infusion (BHI) powder from Difco Co. (Sparks, MD, USA). Acetonitrile, ethyl acetate and methanol were of high-performance liquid chromatography (HPLC) grade. Biosynthesized DHD and DHG by previously isolated anaerobic bacterium in our lab were used as standard compounds (Liang et al. 2010).

The original obligate anaerobic bacterium and cultural conditions

The original obligate anaerobic bacterium Niu-O16 (CGMCC 1.3710) was cultured in BHI liquid medium in an anaerobic chamber (Concept 400, Ruskinn, UK) containing 5% CO2, 10% H2, and 85% N2, at 37 °C.

Oxygen-tolerant domestication process

The original obligate anaerobic bacterium Niu-O16 was inoculated and incubated in an anaerobic chamber for 24 h before being used. Eight hundred microliters of the preprepared culture broth of strain Niu-O16 was inoculated in a 10 ml cap-tube with 4 ml (approximately 4 cm in depth) fresh semisolid BHI medium, this medium was composed of 3.7% (m/v) of BHI powder, 0.6% (m/v) of agar and 0.15% (m/v) of l-cysteine, and incubated in a biochemical incubator at 37 °C. To ensure that strain Niu-O16 could grow and be transferred to the next passage stably, we repeated the process five times. We named the whole process “the first-round domestication”. The strain Niu-O16 obtained from the first-round domestication was inoculated in the conditions described in the first-round domestication except with a slightly decreased concentration of l-cysteine (from 0.15% to 0.14%). When the amount of l-cysteine added to the cultural medium was decreased to zero, we began to drop down the concentration of agar little by little (0.02% decrease of agar in each round of domestication) in a similar method until the amount of agar was reduced to zero. Finally, we obtained an oxygen-tolerant bacterium, capable of being kept and transferred to the next passage stably in the presence of atmospheric oxygen. The bioconversion activity of the gradually domesticated strain was detected by HPLC method as described in our previous study (Wang et al. 2005b) with a slight improvement. The oxygen-tolerant domestication process has been summarized in Table 1.

Table 1 Anti-oxidant domestication procedures for bacterial strain Niu-O16
Full size table

Cell morphology, biochemical traits, and 16S rRNA gene sequencing

Cell morphology of both the wild-type Niu-O16 and the oxygen-tolerant strain Aeroto-Niu-O16 was done by a light microscope equipped with a digital camera (CN15-T31, KONKYO, China) as well as a transmission electron microscope (TEM; JEOL-1010 Electron Microscope, Japan). Biochemical tests were carried out by using API 20A kits (bioMérieux, Lyon, France). Tests of the characteristics which do not exist in API 20A kits followed the protocols described in the Wadsworth-KTL Anaerobic Bacteriology Manual (Jousimies-Somer et al. 2002). Bacterial chromosomal DNA preparation, PCR amplification and purification, and sequencing of 16S rRNA genes were performed as described previously (Wang et al. 2005b). The 16S rRNA gene sequence alignment was carried out by using DNAMAN sequence analysis software (Lynnon BioSoft, Canada).

Influence of the cultural conditions on the growth of strain Aeroto-Niu-O16

To investigate the influence of the inoculation concentration on bacterial cell growth, we inoculated the different amounts (2.5%, 5.0%, and 10%; v/v) of the culture broth of strain Aeroto-Niu-O16 in a 10 ml culture tube with a 4 ml (approximately 4 cm in depth) of fresh BHI liquid medium and incubated in a biochemical incubator at 37 °C for 24 h. In order to study the influence of the depth of medium on cell growth, we inoculated 10% (v/v) of the culture broth of the strain Aeroto-Niu-O16 in a 10 ml culture tube with the different amounts (1, 2, and 4 ml) of fresh BHI liquid medium (approximately 1, 2, and 4 cm in depth). To detect the optical density (OD) at 600 nm, we sampled the culture broth after 24 h incubation.

Production and identification of isoflavone metabolites by strain Aeroto-Niu-O16

The metabolites of daidzein and genistein by the oxygen-tolerant strain Aeroto-Niu-O16 were determined by both the UV spectra and the retention time in HPLC profiles. The procedures were as follows: 200 μl of the culture broth was extracted with 1 ml ethyl acetate and evaporated to dryness using a DK-VC010 Concentrator (Daiki, Seoul, South Korea). The residue was redissolved in 200 μl of 100% of methanol and 10 μl of the solution was used for HPLC detection. The samples were analyzed on a Waters HPLC instrument equipped with a photodiode array detector and a Kromasil C18 reversed-phase column (4.6 × 250 mm). The eluting solution consisted of 10% acetonitrile in 0.1% acetic acid (A) and 90% acetonitrile in 0.1% acetic acid (B). Elution was carried out as follows: A/B at 70:30 (v/v) for 10 min; linearly to 50:50 (v/v) for 10 min; linearly to 70:30 for 5 min. The flow rate was 1 ml/min.

Bioconversion activity of daidzein and genistein by strain Aeroto-Niu-O16

We inoculated 200 μl of the preprepared culture broth of the oxygen-tolerant strain Aeroto-Niu-O16 in a 10 ml culture tube with 1.8 ml BHI liquid medium (approximately 2 cm in depth) and added 20 μl of the stock solution of daidzein (10 mM) or that of genistein (10 mM). We then incubated the culture tubes in a biochemical incubator for 48 h to detect the bioconversion activity. The bioconversion activity of daidzein and genistein was expressed as bioconversion rate, which was calculated as follows: bioconversion rate = the concentration of the metabolite (mM)/[the concentration of the metabolite produced (mM) + the concentration of the substrate left in the medium (mM)]. The concentration of the metabolite and that of the substrate left in the medium was determined based on the peak area in HPLC profiles and the standard curves of both the metabolite and the substrate. In order to investigate the maximum bioconversion capacity of strain Aeroto-Niu-O16, the organism was incubated with different amounts of the substrate (0.4, 0.8, 1.2, and 1.6 mM). We sampled the culture broth after 48 h of incubation and detected the bioconversion activity by the HPLC method mentioned previously. Each test was performed at least in triplicate. Statistical analyses were carried out by One-way ANOVA of variance (with LSD post hoc analysis) using SPSS version 13 for Windows. Differences were considered significant if p < 0.05.

Influence of different reductives on bioconversion activity of strain Aeroto-Niu-O16

In order to increase the bioconversion activity of daidzein and genistein by the oxygen-tolerant strain Aeroto-Niu-O16, we added different amounts of the reductives in the cultural medium, including l-cysteine, ascorbic acid, and sodium thiosulfate, respectively. The initial concentration of substrate (daidzein or genistein) was 0.1 mM. We then sampled the culture broth and analyzed it by the HPLC method.

The complete 16S rRNA gene sequence of the bovine rumen bacterial strain Niu-O16 has been deposited in NCBI GenBank under the accession number of AY263505.

Results

Comparison of bacterial shape, chemical traits, and 16S rRNA gene sequences

An obvious difference in cell morphology between oxygen-tolerant strain Aeroto-Niu-O16 and the wild-type strain Niu-O16 was observed through both light microscope and TEM. The typical cell morphology of the wild-type strain Niu-O16 grown anaerobically is rod-shaped under a light microscope (Fig. 1a). However, the cells of the derived strain Aeroto-Niu-O16 grown aerobically were shown to be rods occurring in filament (Fig. 1b). In order to get the ultrastructural information, the cells of both the wild-type strain Niu-O16 grown anaerobically and the derived strain Aeroto-Niu-O16 grown both aerobically and anaerobically were examined by TEM. Unexpectedly, we observed a significant difference in cell structure and propagation. The wild-type strain Niu-O16 is rod-shaped with a clear outline of cell body (Fig. 2a), and mainly relies on the common binary fission for propagation (data not shown). The outline of the bacterial cell of derived strain Aeroto-Niu-O16 was obvious only when it is grown anaerobically (Fig. 2c). In addition, we observed that the derived strain Aeroto-Niu-O16 used an asymmetric multiple-fission for reproduction (Fig. 2b).

Fig. 1
figure 1

Light micrographs showing rod-shaped cells of the original obligate anaerobic bovine rumen bacterial strain Niu-O16 grown anaerobically (a) and the rods occurring in filament of the derived oxygen-tolerant strain Aeroto-Niu-O16 grown aerobically (b). The bacterial cells were sampled in late exponential phase

Full size image
Fig. 2
figure 2

Transmission electron micrography of the wild-type strain Niu-O16 grown anaerobically (a) and the derived oxygen-tolerant strain Aeroto-Niu-O16 grown both aerobically (b) and anaerobically (c). The bacterial cells were sampled in middle exponential phase

Full size image

Study on biochemical tests showed that the oxygen-tolerant strain Aeroto-Niu-O16 and the original obligate anaerobic strain Niu-O16 had most but three biochemical traits in common. One exception was that the original strain Niu-O16 was indole-negative, while strain Aeroto-Niu-O16 was indole-positive. The second exception was that strain Aeroto-Niu-O16 had totally lost the amylohydrolysis activity which was originally possessed by strain Niu-O16. The last exception was that the capacity to utilize some carbon sources for strains Aeroto-Niu-O16, including mannitol, xylose, glycerol, and sorbitol, was slightly improved.

In order to compare the base sequence of the 16S rRNA gene, total cellular DNA was extracted from the oxygen-tolerant strain Aeroto-Niu-O16. The 16S rRNA gene was amplified by PCR and the product was sequenced. Sequence alignment was done between the 16S rRNA gene of strain Aeroto-Niu-O16 and the original obligate anaerobic strain Niu-O16. The result showed that the 16S rRNA genes from strain Aeroto-Niu-O16 and the original strain Niu-O16 were identical along their entire length with some exceptions. Two point mutations in 16S rRNA gene, including G398A and G438A, were found. The numbering of the 16S rRNA nucleotides is based on that of Escherichia coli and is used throughout. Bacterial operon and gene prediction were done to the 16S rRNA gene sequences of the wild strain Niu-O16, the derived strain Aeroto-Niu-O16 and E. coli (ATCC 11775T) by both ORF Finder and FGENESB. We found a coding sequence (CDS) from 341 to 574 bp of strain Niu-O16 and strain Aeroto-Niu-O16. The 398 bp is the first position of the 20th code representing Glu for strain Niu-O16 and Lys for strain Aeroto-Niu-O16. Similarly, the 438 bp is the second position of the 33rd codon representing Gly and Asp for strain Niu-O16 and strain Aeroto-Niu-O16, respectively. As for E. coli (ATCC 11775T), we found a CDS from 333 to 470 bp in its complementary chain of the 16S rRNA gene sequence. The corresponding amino acid is Lue and Gly, respectively.

Growth property

The culture broth of the original obligate anaerobic strain Niu-O16 was even turbid when it was grown in an anaerobic chamber. However, the cells of the oxygen-tolerant strain Aeroto-Niu-O16 agglutinated on the bottom of the culture tube when it was grown in a biochemical incubator (with atmospheric oxygen).

In order to know the growth characteristics of the oxygen-tolerant strain Aeroto-Niu-O16 in the presence of atmospheric oxygen, both the growth curve and the change of pH were detected with time. The lag phase of strain Aeroto-Niu-O16 in biochemical incubator was about 12 h when 10% (v/v) of the culture broth was inoculated. The exponential phase of strain Aeroto-Niu-O16 started after 12 h incubation and lasted for about 6 h. Meanwhile, the pH value of the culture broth sharply dropped after 12-h incubation and stably remained around 5.4 after 21-h incubation (Fig. 3).

Fig. 3
figure 3

Time course of bacterial cell growth and pH change of the oxygen-tolerant strain Aeroto-Niu-O16 in the presence of atmospheric oxygen. The inoculum concentration of strain Aeroto-Niu-O16 was 10% (v/v). The culture broth of the oxygen-tolerant strain Aeroto-Niu-O16 was sampled every 3 h for detection of both the pH and the OD600

Full size image

Influence of cultural conditions on growth of strain Aeroto-Niu-O16

Our study revealed that the growth of the oxygen-tolerant strain Aeroto-Niu-O16 in a biochemical incubator was influenced by both the inoculation amount of the culture broth and the depth of the medium in the culture tube. The strain Aeroto-Niu-O16 failed to grow when the inoculation amount was or was lower than 2.5% (OD600 ≤ 0.025). In most cases, we could transfer the organism for only one passage when the inoculation amount was increased to 5% (OD600 ≈ 0.05). However, when we increased the inoculation amount to10% (OD600 ≥ 0.1), we could transfer strain Aeroto-Niu-O16 continuously and stably for at least five passages (Fig. 4a). On the other hand, our derived strain Aeroto-Niu-O16 was also influenced by the depth of the medium in the culture tube when we inoculate 10% of the culture broth. Strain Aeroto-Niu-O16 failed to grow when the depth of the medium was or was lower than 1 cm in the culture tube. However, we were able to continuously and stably transfer strain Aeroto-Niu-O16 for at least five passages when the medium was or was more than 2 cm in depth (Fig. 4b).

Fig. 4
figure 4

Effect of inoculum concentration (a) and the cultural depth (b) on cell growth of the oxygen-tolerant strain Aeroto-Niu-O16 in the presence of atmospheric oxygen. Cultures at a final OD600 of 1.0 were used for inoculation. The growth yield was determined every 24 h followed by continuous inoculation for five passages

Full size image

Production and identification of the metabolites by strain Aeroto-Niu-O16

On the basis of the UV spectra and the retention time in HPLC profiles, peak 1 in Fig. 5a was the substrate daidzein (the maximum absorbance was 248 and 306 nm) and peak 2 was DHD, the metabolite of daidzein (the maximum absorbance was 275 and 326 nm). However, peak 2 in Fig. 5b was the substrate genistein (the maximum absorbance was 263 nm), and peak 1 was DHG, the metabolite of genistein (the maximum absorbance was 292 nm). The bioconversion activity of daidzein was 60.3% and that of genistein was 74.1%.

Fig. 5
figure 5

HPLC elution profiles of metabolism of daidzein (a) and genistein (b) by the oxygen-tolerant strain Aeroto-Niu-O16 in BHI liquid medium in the presence of atmospheric oxygen. The insets show UV spetra of the substrate daidzein (a) and genistein (b), metabolites of daidzein (peak2) and metabolite of genistein (peak1), and the authentic dihydrodaidzein (DHD) and dihydrogenistein (DHG). AU, absorbance units. The initial concentration of the substrate was 0.1 mM

Full size image

Maximum bioconversion capacity of strain Aeroto-Niu-O16

The oxygen-tolerant strain Aeroto-Niu-O16 was tested in its capacity to convert different amount of the substrate added in the cultural medium. The results showed that the average bioconversion activity of the triplicate in each performance of daidzein was 74.8%, 73.4%, and 71.6% when the initial concentration of daidzein was 0.4, 0.8, and 1.2 mM, respectively. However, when we increased the initial concentration of daidzein to 1.6 mM, the bioconversion activity sharply dropped to 56.3% (Fig. 6a). In the case of genistein, the average bioconversion rate was 88.3%, 84.8%, 78.4%, 74.3%, and 72.3% when the initial concentration of genistein was 0.4, 0.8, 1.2, 1.6, and 2.0 mM, respectively. The bioconversion activity sharply dropped to 61.1% when we increased the initial concentration of genistein to 2.4 mM (Fig. 6b).

Fig. 6
figure 6

Bioconversion capacity of the oxygen-tolerant strain Aeroto-Niu-O16 to different concentration of the substrate daidzein (a) and genistein (b) in the presence of atmospheric oxygen. Dihydrodaidzein (DHD) is the metabolite of daidzein and dihydrogenistein (DHG) is the metabolite of genistein

Full size image

Enhanced bioconversion activity of strain Aeroto-Niu-O16 by addition of ascorbic acid

In order to increase the bioconversion activity, we added different kinds of reductives, including l-cysteine, ascorbic acid, and sodium thiosulfate, respectively, in the cultural medium. We observed that neither l-cysteine nor sodium thiosulfate showed any improved effect on the bioconversion activity of strain Aeroto-Niu-O16 when it was grown in the presence of atmospheric oxygen (data not shown). However, the bioconversion activity of daidzein and genistein was increased significantly when we added a proper amount of ascorbic acid (0.15%, m/v) to the cultural medium, under which circumstance, the bioconversion rate of DHD was increased from 60.3% to 71.7% and that of DHG was increased from 74.1% to 89.2%.

Discussion

Oxygen-tolerant capacity of strain Aeroto-Niu-O16

The original bovine rumen bacterial strain Niu-O16 was an obligate anaerobic bacterium even though the oxygen sensitivity of the original strain Niu-O16 was not mentioned in our previous study (Wang et al. 2005b). After a long-term domestication, an oxygen-tolerant bacterium which we named Aeroto-Niu-O16 was obtained. To compare the growth property, strain Niu-O16 and the oxygen-tolerant strain Aeroto-Niu-O16 were each cultured in the presence and absence of atmospheric oxygen. However, strain Niu-O16 was not able to grow in the presence of atmospheric oxygen when we inoculated 10% (v/v) of the culture broth in the culture tube with 4 ml of fresh BHI liquid medium (approximately 4 cm in depth). On the contrary, the oxygen-tolerant strain Aeroto-Niu-O16 could grow well in the absence of oxygen. In the anaerobic chamber, the cells of strain Aeroto-Niu-O16 did not clump on the bottom of the culture tubes; instead, we observed loosely gathered floccules in the culture medium. However, the original strain Niu-O16 was even turbid. In addition, we also detected the oxidation-reduction potential (ORP) of fresh BHI liquid medium with and without l-cysteine. The ORP of fresh BHI liquid medium was −90 mV; however, when we added 0.15% l-cysteine to the BHI liquid medium, the ORP was dropped to −390 mV. Therefore, we concluded that the oxygen-tolerant capacity of the domesticated strain Aeroto-Niu-O16 has been much improved compared with that of the original strain Niu-O16.

All living organisms can sense change in the external environment and adjust themselves to adapt accordingly. The adaptation to aerobic conditions of the oxygen-tolerant strain Aeroto-Niu-O16 was accompanied by a series of changes, such as change in morphology, growth property, biochemical traits (including decreased final pH from 5.7 to 5.4) as well as point mutations in 16S rRNA gene (G398A and G438A). One visible change observed in our study is the reproductive strategy used by the oxygen-tolerant strain Aeroto-Niu-O16. The wild-type strain Niu-O16 reproduced by binary fission leads to a doubling of the bacterial cell number (data not shown). By TEM image, the derived oxygen-tolerant strain Aeroto-Niu-O16 was observed to use an asymmetric multiple-fission way for reproduction (Fig. 2b), which no doubt much improved the opportunity to survival. Multiple asymmetric initiation points of septum formation were also observed by electron microscopy when MRSA strains were treated with antibiotics like macrolactin A (Magally et al. 2006). Probably this is a required capacity for some bacteria to survive under adverse circumstances. Change in grow property (from even turbid to being agglutinated) might be due to the change in bacterial shape (from rod-shaped to filament). It was reported that bacteria would change their morphology to adapt to the changed new environment (Young 2006). The 16S rRNA gene is generally used as a specific marker for bacterial identification as it is highly conservative between different species of bacteria (Woese 1987). However, point mutations within 16S rRNA genes are becoming increasingly recognized. A G791A mutation in 16S rRNA gene in the E. coli affected processes involved in the initiation of protein synthesis (Tapprich et al. 1989). A C1066U mutation in 16S rRNA gene was found to be of increased cell sensitivity to fusidic acid (Johanson and Hughes 1995). In the 16S rDNA sequence of selected Brachyspira hyodysenteriae isolates, a G1058C mutation was found for all isolates with increased doxycycline MICs whereas all susceptible isolates had the wild-type sequence (Pringle et al. 2007). In our study, two point mutations (G398A and G438A) were found in the oxygen-tolerant strain Aeroto-Niu-O16. However, whether these two point mutations are of decreased cell sensitivity to oxygen is not clear. A further study is needed for figuring out the mechanisms responsible for the improved oxygen-tolerant capacity of strain Aeroto-Niu-O16.

The bioconversion activity of strain Aeroto-Niu-O16

Our study provided the first evidence that oxygen-tolerant domesticated isoflavone biotransforming bacterium was able to grow and execute the reductive activity in the presence of atmospheric oxygen. Even though the bioconversion activity of the oxygen-tolerant strain Aeroto-Niu-O16 in the biochemical incubator (averagely 60.3% for daidzein, and 74.1% for genistein) was decreased compared with that of the original strain Niu-O16 in anaerobic chamber (averagely 85.5% for daidzein and 91.3% for genistein), the bioconversion activity of daidzein and genistein by strain Aeroto-Niu-O16 was almost the same as that of the original strain Niu-O16 when we inoculated both of the two strains and incubated them in the anaerobic chamber in parallel (data not shown). In addition, we observed two strange phenomena in our study. One was that the bioconversion rate of daidzein by strain Aeroto-Niu-O16 was 60.3% when the initial concentration of daidzein was 0.1 mM (data not shown); however, it was increased to more than 70% when the initial concentration of daidzein was or was more than 0.4 mM (Fig. 6a). This is quite abnormal in microbial biotransformation process. Besides the cell number of a microorganism, biotransformation is also controlled by biochemical activity of the microorganism and the mass transfer of a chemical to the microorganism. Under a certain range of substrate concentration, the biotransformation rate is relative stable. However, when the high concentration of substrate carries toxic effects on the microorganism, the bioconversion rate will get decreased (Chang et al. 2008; Wang et al. 2005b).

The second strange phenomenon was that the total amount of the product (DHD) and the substrate (daidzein) that we extracted from the culture broth of strain Aeroto-Niu-O16 was much lower than the initial amount of the substrate we added in the cultural medium (lower than 50%; Fig. 6a). Since the recovery rate of the substrate daidzein added in fresh BHI liquid medium was about 85% in our study (data not shown), we concluded that almost half of daidzein had been lost when a higher concentration was added to the cultural medium. However, in the case of genistein, such kind of phenomenon was not so obvious: the average recovery rate of genistein was 76.5% when the concentration we added was or was more than 0.4 mM. In order to find the reason for the difference, one more experiment was carried out. We inoculated 10% of the preprepared culture of the oxygen-tolerant strain Aeroto-Niu-O16 in a 4 ml of BHI liquid medium followed by addition of 1.2 mM of the substrate daidzein. After being incubated in a biochemical incubator for 48 h, 1 ml of the culture broth was sampled and centrifuged. Both the supernatant and the cell pellet were extracted with 1 ml ethyl acetate. The total amount of the DHD we extracted from the supernatant was 0.41 mM and that of daidzein was 0.13 mM (Fig. 7a, the solid line). However, the total amount of the DHD we extracted from the cell pellet directly was only 0.01 mM and that of daidzein was 0.04 mM (data not shown). Since the total amount of the product (DHD) and the substrate (daidzein) we extracted from both the supernatant and the cell pellet (0.59 mM) was still much lower than that of the initial amount of the substrate (1.2 mM) we added in the cultural medium, we sampled another 1 ml of the culture broth from the same culture tube mentioned above and centrifuged it. We washed the cell pellet with double distilled water three times, crushed in a mortar, and extracted with 1 ml of ethyl acetate.

Fig. 7
figure 7

HPLC elution profiles of the substrates daidzein and its metabolite dihydrodaidzein (DHD) (a); substrate genistein and its metabolite dihydrogenistein (DHG) (b) extracted from the supernatant (solid line) and the crushed cell pellet (dotted line), respectively

Full size image

To our surprise, a large amount of daidzein (0.39 mM) but a small amount of DHD (0.07 mM) was extracted from the crushed cell pellet (Fig. 7a, the dotted line). In the case of genistein, the total amount of the DHG we extracted from the supernatant was 0.76 mM and that of genistein was 0.22 mM (Fig. 7b, the solid line). However, the total amount of the DHG we extracted from the cell pellet directly (data not shown) or from the crushed cell pellet (the amount of DHG was 0.01 mM, and the amount of genistein was 0.05 mM) was only a small amount (Fig. 7b, the dotted line). Furthermore, when we performed a similar process to the crushed cells of the wild-type strain Niu-O16 or the oxygen-tolerant strain Aeroto-Niu-O16 grown anaerobically, only a trace amount of the product and the substrate was extracted. Therefore, we surmised that a “protective coat” was formed with the stimulation of the atmospheric oxygen when strain Aeroto-Niu-O16 was cultured aerobically. The newly formed “protective coat” might prevent the normal transportation of oxygen through the membrane of the cell, the process of which would be beneficial for the aerobic growth of the oxygen-tolerant strain Aeroto-Niu-O16.

Furthermore, in order to justify our hypothesis, we used Congo Red to stain the bacterial cells of both the wild type grown anaerobically and the derived strain grown both aerobically and anaerobically. An obvious blue circle was seen along the cell wall of the derived strain grown aerobically (Fig. 8b), however, no similar result was seen to that of the wild strain (Fig. 8a) and the derived strain grown anaerobically (data not shown). Therefore, we surmised that exopolysaccharide like substances were produced and accumulated on the surface of the derived strain when grown aerobically, which formed the proposed “protective coat”.

Fig. 8
figure 8

Light micrographs showing stained bacterial cells of the original obligate anaerobic bovine rumen bacterial strain Niu-O16 (a) and the derived oxygen-tolerant strain Aeroto-Niu-O16 (b) with Congo Red. The bacterial cells were sampled in middle exponential phase

Full size image

On the other hand, since the isoflavone biotransforming enzyme was intracellular (Shimada et al. 2010), the bioconversion process happened inside the bacterial cells. Therefore, the “protective coat” occurring on the surface of the bacterial cell of strain Aeroto-Niu-O16 would prevent normal transportation of both the substrate and the metabolite. It is not easy for daidzein to pass through the “protective coat” because of the very low water solubility, which might explain the large remaining amount in the “protective coat”. Since the water solubility of DHD is much higher than that of daidzein, it would be much easier for DHD to move out of the cell and go into the culture broth by passing through both the membrane and the “protective coat”. This structure might provide a good explanation for the phenomenon that the total amount of daidzein we extracted from the crushed cell pellet was much higher than that of DHD (Fig. 7a, the dotted line).

In the case of genistein, the situation was different since both the substrate genistein and the metabolite DHG have relatively high water solubility in comparison with that of daidzein. Therefore, it would be difficult for either genistein or DHG to stably remain in the “protective coat” for a long period of time. It is worth noticing here that this oxygen-tolerant strain Aeroto-Niu-O16 will be more suitable for the production of DHG instead of DHD under aerobic conditions. One reason is that the productivity of DHG (the amount of DHG produced/the total amount of genistein initially added to the medium) is much higher than that of DHD when the same amount of substrate was added to the same amount of medium.

The other reason is that the maximum bioconversion capacity of genistein was much improved due to the newly formed “protective coat”. To the best of our knowledge, for a specific isoflavone biotransforming bacterium, the maximum bioconversion capacity of genistein is usually lower than that of daidzein since genistein precursor, biochanin A, had greater antimicrobial efficiency than the daidzein precursor, formononetin (Flythe and Kagan 2010). However, the newly formed “protective coat” might greatly decrease the toxicity of genistein to the cells of the oxygen-tolerant strain Aeroto-Niu-O16. The “protective coat”, just like a biofilm, which may play a role of permeability barriers to slow down the transporting speed of the substrate genistein from the medium to the inside of the bacterial cells. A biofilm is a population of cells growing on a surface and enclosed in an exopolysaccharide (EPS) matrix. This EPS matrix plays a role of permeability barriers to prevent antimicrobial substance (like antibiotics) to bacterial targets (Nikaido 1994). We cultured both the oxygen-tolerant strain Aeroto-Niu-O16 and the wild-type strain Niu-O16 in parallel in an anaerobic chamber and compared their maximum bioconversion capacity. The result showed that the maximum bioconversion capacity of genistein by the wild-type strain Niu-O16 was 0.8 mM (the bioconversion rate of which was 84.0%, data not shown); however, when we increased the concentration of genistein to 1.2 mM, the bioconversion rate sharply dropped to 64.6%. In the case of the oxygen-tolerant strain Aeroto-Niu-O16, the bioconversion rate of genistein was 78.4% when the initial concentration of genistein was 1.2 mM. The bioconversion rate sharply dropped to 61.1% when the concentration of daidzein was 2.4 mM (Fig. 6b).

Taken together, the oxygen-tolerant strain Aeroto-Niu-O16 differed from the wild-type strain Niu-O16 by various characteristics, such as cell morphology, indole formation, utilization of carbon sources, and point mutations of 16S rRNA genes. Meanwhile, strain Aeroto-Niu-O16 also keeps its isoflavone biotransformation activity in the presence of atmospheric oxygen. Because oxygen is ubiquitous in the air, successful oxygen-tolerant domestication is of great importance for the industrial application of isolated obligate anaerobes.