Diversity and abundance of bacteria in an underground oil-storage cavity

Underground cavities have been used for long-term storage of crude oil in several countries, and one of such facilities is situated at Kuji in Iwate, Japan. These cavities have been constructed in groundwater-rich rocky strata, where high groundwater pressure confines the stored oil in the cavities [1]. Consequently, groundwater migrates into and accumulates at the bottom of a cavity (cavity groundwater), and this cavity groundwater is discharged to maintain the oil storage capacity of the cavity (this system has been detailed in our previous study [1]). Our previous study [1] has also shown active growth of microorganisms in groundwater accumulating at the bottom of the cavities; the total count of microorganisms in the cavity groundwater was constantly more than 10⁶ cells per ml (densities 100 times higher than those in groundwater around the cavities). This habitat can be characterized by (i) immediate contact with a large quantity of crude oil and (ii) an excess of electron donors for microbial growth (i.e., hydrocarbons) but a shortage of electron acceptors [1]. These characteristics may be similar to those of microbial habitats associated with subterranean oil reservoirs, which have recently attracted much attention in microbiology [2–6]. Since groundwater can easily be obtained from the bottom of the oil-storage cavities using standing sampling facilities without its contamination by surface water [1], studies on the oil-storage cavity are considered to provide valuable information to understand the microbial ecology of subterranean oil fields.

Our previous study applied rRNA approaches, namely cloning and sequencing of 16S rRNA gene fragments, denaturing gradient gel electrophoresis and fluorescent in situ hybridization (FISH), to analyze bacterial populations that occurred in the cavity groundwater obtained at Kuji [1]. As a result, a group of bacteria (called cluster-1 bacteria) affiliated with the Thiovulum subgroup in the ε subclass of the class Proteobacteria was consistently detected as a major population. Quantitative comparison of the results of these approaches, however, revealed a large bias associated with the cloning and sequencing approach; it was thus considered that the bacterial biodiversity has not sufficiently been assessed yet.

The present study was conducted to obtain a more reliable view on the bacterial biodiversity in the Kuji cavity groundwater. For this purpose, this study employed widely used universal primers [7] and a recently-modified Bacteria-specific primer [8] for PCR amplification of 16S rRNA gene (16S rDNA) fragments from the cavity groundwater, followed by cloning and sequencing. In addition, in order to quantitatively assess results of the cloning approach, competitive PCR (cPCR) assays were developed for some of abundantly obtained 16S rDNA types that were considered important for anaerobic petroleum degradation.

The present study used the two combinations of PCR primers for amplifying 16S rDNA fragments from the cavity groundwater. The sequence analysis of the amplified fragments suggests that the bacterial diversity in the cavity groundwater is much larger than the diversity found in our previous study [1]. The present study abundantly detected the sequence types related to Desulfotomaculum, Acetobacterium, Desulfovibrio, Desulfobacula, Zoogloea and Thiomicrospira denitificans, while major 16S rDNA sequences obtained in our previous study were solely affiliated with the ε-Proteobacteria. This comparison clearly indicates that methods employed in the molecular ecological approaches largely affect the results obtained. Differences in the methods employed in these two studies included primers used for PCR amplification of 16S rDNA fragments from the groundwater DNA and selection of E. coli clones harboring groundwater 16S rDNA fragments. The PCR primers used in our previous study were two Bacteria-specific primers (called 8f and 1546r [1]). We assume that organisms undetected in our previous study have 16S rDNA sequences that include mismatches to these primers, since this study did not determine sequences of the 5'- and 3'-end regions of 16S rDNA where 8f and 1546r exist. In addition to the primer selection, different from our previous study [1], the present study did not employ the β-galactosidase colony screening when selecting E. coli clones, because it has been reported that the vector containing a DNA fragment up to 2 kb in-frame produces a blue colony. These two modifications may have circumvented the bias toward the ε-Proteobacteria, which occurred in our previous study [1].

For estimating the abundance of bacterial populations represented by the major sequence types, this study employed cPCR assays. We did not use FISH, because a possible bias that causes the underestimation of slow-growing bacteria has been suggested [1]. Besides, in the present study, FISH with the Bacteria- and Archaea-specific probes detected only 60% in total of the DAPI-stained cells; in addition, a large portion of labeled cells exhibited weak signals. One possible explanation for this result would be that the probes used (e.g., EUB338) failed to detect some phylogenetic groups of microorganisms. The utility of cPCR for this purpose has been demonstrated in several studies [16, 17], while attention should be paid to the variation in the number of 16S rDNA copies in a bacterial cell. For example, some Clostridium strains (affiliated with the low G+C gram positive bacteria) have been known to harbor more than 10 rDNA copies per cell [18, 19], while the mean copy number in whole-genome-analyzed bacteria is 3.6 [20]. We assume that cPCR may have overestimated the Acetobacterium rDNA copies, since Acetobacterium belongs to the Clostridiaceae. In contrast, rDNA copy numbers per cell of bacteria related to the other major sequence types were not so different from the mean value for all bacteria [20]. We therefore consider that the results of cPCR reflected the abundance in the cavity groundwater of the cluster-1 bacteria and those related to Desulfotomaculum, Desulfovibrio and Zoogloea. Comparison of the cloning and cPCR results indicated that relative abundances determined by these two methods were inconsistent with each other in several cases (Table 2 and Figure 2). Furthermore, clones obtained in abundance in one library were minorities in another library in many cases (Table 2). These results suggest that the rDNA cloning approach employed in this study also included some biases.

Environmental rDNA sequencees have been used to infer some properties of the organisms that they represent [1, 7, 21], although this type of analysis is reliable only if some common properties can be recognized for phylogenetically related organisms. In the present study, we assume that the organism represented by sequence types KU21 and KU25 are acetogens [22], and those represented by KB1, KB14, KB33 and KB47 are likely sulfate-reducing bacteria [23]. KU34 is closely related to the genus Zoogloea that has been known to grow by denitrification under anaerobic conditions [24]. The Zoogloea belong to the Rhodocyclus group in the β subclass of the class Proteobacteria, which also includes the genera Rhodocyclus and Azoarcus[25]. Since organisms in these genera can commonly perform denitrification, organisms corresponding to clone KU34 are assumed to be capable of denitirfication. Sequence types KU26 and KB13 are related to the genus Desulfotomaculum that belongs to the Syntrophomonas/Thermoanaerobacter group in the low G+C gram-positive bacteria [25]. Typical members of this group are sulfate-reducing bacteria represented by Desulfotomaculum spp. [26], fermentative bacteria represented by Sporotomaculum hydroxybenzoicum[27] and acid-oxidizing syntrophs represented by Syntrophomonas spp. [28]. It is likely that organisms represented by KU26 and KB13 were engaged in some steps in organic compound decomposition. Similar types of bacteria, including δ-Proteobacteria, ε-Proteobacteria and low G+C gram-positive bacteria, have been detected by rDNA cloning analyses of mesophilic oil reservoirs in Canada [3], while β-Proteobacteria were not detected from there. We thus assume that understanding physiology of organisms represented by clone KU34 would provide valuable information regarding differences of ecosystem functions in these two subsurface oil fields.

We have recently isolated bacterial strains that belonged to the cluster-1 bacteria [Kodama, Y. and Watanabe, K. Unpublished results]. It has been found that these strains are capable of anaerobic chemolithotrophic growth by oxidizing reduced sulfur species (sulfide, elemental sulfur and thiosulfate) coupled to nitrate reduction. This finding suggests that this group of bacteria mainly consumed nitrate that was supplied by groundwater flowing into the oil-storage cavities.

The present study unveiled the large diversity of bacteria in the cavity groundwater. The phylogenetic affiliation of cloned 16S rDNA sequences suggests that bacteria exhibiting different types of energy metabolisms coexist as major proportions. This finding is considered interesting, because the chemical analyses of the groundwater have suggested that nitrate reduction is the major electron accepting process [1]. In particular, although sulfate was not significantly consumed [1] and sulfide was always undetectable (below 0.01 mg per liter [unpublished data]) in the cavity groundwater pool, the results of the present study indicate that sulfate-reducing bacteria were present at significant levels. One possible explanation for these findings would be the existence of a sulfur cycle in the cavity groundwater that involves sulfate reducers and anaerobic sulfur oxidizers that regenerated sulfate from sulfide. We assume that similar sulfur cycles may operate in subterranean oil fields, since sulfate-reducing bacteria have been isolated from subterranean oil fields whose formation water contained very low concentrations of sulfate [29, 30]. In order to elucidate ecological niches and interactions among these bacteria, their isolation and physiological analyses in axenic cultures are needed. These studies will draw novel insights into the ecological interactions among microorganisms in subterranean oil fields.