Sulphate-reducing bacterial community structure from produced water of the Periquito and Galo de Campina onshore oilfields in Brazil

Sulphate-reducing bacteria (SRB) cause fouling, souring, corrosion and produce H2S during oil and gas production. Produced water obtained from Periquito (PQO) and Galo de Campina (GC) onshore oilfields in Brazil was investigated for SRB. Produced water with Postgate B, Postgate C and Baars media was incubated anaerobically for 20 days. DNA was extracted, 16S rDNA PCR amplified and fragments were sequenced using Illumina TruSeq. 4.2 million sequence reads were analysed and deposited at NCBI SAR accession number SRP149784. No significant differences in microbial community composition could be attributed to the different media but significant differences in the SRB were observed between the two oil fields. The dominant bacterial orders detected from both oilfields were Desulfovibrionales, Pseudomonadales and Enterobacteriales. The genus Pseudomonas was found predominantly in the GC oilfield and Pleomorphominas and Shewanella were features of the PQO oilfield. 11% and 7.6% of the sequences at GC and PQO were not classified at the genus level but could be partially identified at the order level. Relative abundances changed for Desulfovibrio from 29.8% at PQO to 16.1% at GC. Clostridium varied from 2.8% at PQO and 2.4% at GC. These data provide the first description of SRB from onshore produced water in Brazil and reinforce the importance of Desulfovibrionales, Pseudomonadales, and Enterobacteriales in produced water globally. Identifying potentially harmful microbes is an important first step in developing microbial solutions that prevent their proliferation.


Materials and methods
Sample collection. Produced  Using a sterile syringe, 10 ml of media plus 1 ml of produced water samples was injected into a vial and purged with nitrogen free oxygen gas for 2-min, before being clamped with a rubber and aluminium cap. The sealed tubes were stored for 20 days at 30° C under anaerobiosis, using anaerobic atmosphere generation bags and Gaspak jars (Merck Ltd).The anaerobic system contains a palladium catalyst and an indicator of anaerobiosis (methylene blue). The culture media contains the oxidation-reduction indicator resazurin, and an envelope containing sodium borohydride and sodium bicarbonate (Merck Ltd). For each oil field nine sub samples were investigated, 3 of each type of medium for each oilfield. After 20 days, bacteria from the 18 samples were collected by centrifugation for DNA extraction. Tubes under the same conditions containing just the medium without produced water provided the negative control.
DNA extraction and molecular identification of bacteria. Genomic   www.nature.com/scientificreports/ dowsize = 50, maxambig = 0, minlength = 410, maxlength = 500. The sequences were then aligned using the Silva database (v132) as a reference 40 and the resultant alignments were submitted to screen.seqs and filter.seqs to remove sequences with bad alignments and to remove uninformative columns of the alignment. The sequences where then pre-clustered using the command pre.cluster with parameter diffs = 3. Chimeras were detected with the chimera.uchime command, which uses the Uchime software 41 , chimeric sequences were removed from further analyses. The remaining sequences were then classified using classify.seqs command, with RDP database as reference and a bootstrap cut-off of 80. Sequences classified as chloroplasts, mitochondria, and those not assigned to any kingdom were removed, No eukaria or archaea sequences were seen in the samples. The resultant sequences, were used as input for the construction of a distance matrix and for clustering the sequences into operational taxonomic units (OTUs), with a cut-off of 3% of dissimilarity. The samples were then randomly normalized to the same number of sequences (n = 46,736) for biodiversity analyses. Then the taxonomy summary was used to identify the bacterial compositions of each sample, and the OTU distributions were used to calculate the diversity indices (Chao and Shannon index).

Statistical analysis.
Regarding the nature of data distribution, all indices showed a normal distribution.
The final files generated by Mothur were imported into R (v4.0) 42 where the microbiome data was processed through the Phyloseq 43 and Microbiome R 44 packages. The Non-Metric Dimensional Scaling (NMDS) analysis was used to visualize the (dis)similarities between the samples, with NMDS ordinations in two dimensions calculated from a Bray-Curtis dissimilarity matrix. The quality of the NMDS representation was evaluated through the generated stress value; considering values < 0.2 as good representation of the data 45 . To verify the separation of the groups observed in the NMDS, a Permutational Multivariate Analysis of Variance (PERMANOVA) was performed. Finally, to ensure that the result of PERMANOVA was not affected by dispersions in the groups, an analysis of multivariate homogeneity of group dispersion (MHGD) was carried out. Differential OTU abundance between oil fields (Galo de Campina and Periquito) was analyzed using the DESeq2 46 . The outputs were adjusted using the Benjamini-Hochberg (BH) method and the values of differential abundances were measured in log2-Fold-Change. A taxon was considered significant when it met the criteria of adjusted p-value < 0.05 and absolute value of log2-Fold-Change > 2.0. Finally, the selected taxa were assessed separately using the t-test with adjustment of the p-value using the BH method 47 . All statistical analyses were performed using R (v4.0) 42 with the support of the vegan package 48 . The BCM, PERMANOVA and MHGD analyses were performed using the functions vegdist(), adonis() and betadisper() from the vegan package.

Results and discussion
Bacterial biodiversity of produced water. The sequence data were deposited in the NCBI Sequence Read Archive (SRA) and are available under accession number SRP149784. In this study, 4.206.240 sequencing reads were generated, all the sequences studied were bacterial.. From 4,206,240 reads a total of 841,248 sequences were analyzed the remainder were trimmed. The bacterial diversity indices for produced water enriched for SRB for the Galo de Campina (GC) and Periquito (PQO) oil fields are shown in Table 1.
Based on a 3% dissimilarity cut off, a total of 992 OTU´s were defined and compared. The number of OTUs (species richness/alpha diversity) varied between 95 and 150 at Campo de Galo and between 93 to 144 at Comparisons of the Chao (P = 0.69) and Shannon (P = 0.75) indices were made and there were no significant differences in biodiversity between the oil fields. At the 3% sequence dissimilarity cut off, there was no significant difference between the different culture media analyzed. NMDS (Fig. 2) and Bray-Curtis produced a stress = 0.06. It was possible to observe a good clustering and a separation by oil field. The GC samples clustered in the positive NMDS1 portion, while GC samples were found in the negative NMDS1. NMDS results were further supported by a PERMANOVA analysis (P = 9.99 e-05, R 2 = 0.94). demonstrating that there is a significant difference between the groups analyzed (Galo de Campina and Periquito oil fields). This result was supported by the homogeneity of variances (P = 0.09) which revealed that there is no inter-individual variation. Overall, these results suggest that there is a real difference in SRB between the oil fields. The results of the BCM analysis indicate that differences seen between samples is explained by differences between the two oilfields.
Taxonomic distributions. No differences in bacterial diversity were observed as a result of using three different enrichment media. An anlysis of media composition reveals that they are similar. The dominant phyla, orders, classes were dominant regardless of which of the three medias was used. The three SRB enrichment media were used to recover a diverse collection of bacteria with the desired metabolic trait, with hindsight, one would have sufficed 49 .
Starting at the phylum level, the distribution of phyla seen in the produced water enriched to detect SRB at both the Galo de Campina and Periquito oilfields was dominated by the Proteobacteria, 93.8%, and Firmicutes with 6.2%. While this study purposefully enriched and selected for SRB it is interesting to report that in other studies of total bacterial diversity from produced water and not SRB enriched, show that phylum level results are similar and global. Belgini et al. 50 identified Proteobacteria (Brevundimonas sp, Sphingobium sp., Sphingopyxis sp., Sphingomonas xenophaga, Achromobacter sp., Delftia tsuruhatensis, Aeromonas sp., Pseudomonas sp., Rheinheimera sp. and Serratia sp.), Firmicutes (Bacillus sp. and Lysinibacillus sp.) and Bacteroidetes (Chryseobacterium sp., Flavobacterium sp.) as the dominant phyla in samples from the Gabriel Passos Refinery (REGAP) in the city of Betim (MG, Brazil). In a study of produced water by de Sousa Pires 51 , from an offshore oil field in southeast Brazil, the bacterial community was dominated by sequences associated with SRB and within the Proteobacteria (specifically Desulfobacterales), Firmicutes and Bacteroidetes. A similar pattern was found in produced water from the Diyarbakır oil Fields in Turkey 52 . In Turkey, the Proteobacteria were the predominant phylum in all samples. In a study from the Mae Soon Luang Field, Fang Basin, Thailand 53 , Proteobacteria, Firmicutes and Actinobacteria dominated the bacterial communities in most samples. Similar reports are seen from Texas, USA by Santillan et al. 13 , Li et al. 54 and from China by Lan et al. 55 . Silva et al. 31 studied the microbial communities in petroleum samples from Brazilian offshore oil fields and found that the most abundant phylum was the Proteobacteria, followed by the Firmicutes (Bacilli and Clostridia). Despite the differences in geographical locations of oilfields and methods used to investigate them, it is interesting to note that these two phyla always dominate. www.nature.com/scientificreports/ This is perhaps not surprising if we consider that conditions within an oil reservoir will strongly select for specific groups of bacteria that can survive and evolve in such environments. The main orders detected in the produced water from Galo de Campina and Periquito were Desulfovibrionales, Pseudomonadales, Enterobacteriales, Clostridiales, Alteromonadales, Rhizobiales and Bacillalles and their distributions are shown in (Fig. 3A). Note at the order level it was possible to identify and classify all sequences. Li et al. 54 , in their work on biofilms in water injection systems in the Daqing oil field (China), also identified Pseudomonadales, Enterobacteriales, Clostridiales and Desulfovibrionales and associated them with hydrogen sulphide production, biofilm formation and biocorrosion 12,13,27 . Desulfovibrionales strains were present in all samples and dominant in both oil fields accounting for 46% of all sequences. The majority of bacteria in this order are SRB and are responsible for biocorrosion. Their presence in these samples might explain problems with hydrogen sulphide at the Periquito oilfield. Pseudomonadales was the dominant order in samples from the Galo de Campina oilfield and Enterobacteiriales strains were present in both oil fields. The Enterobacteriales and Alteromonadales reported here have been previously associated with oil environments 56,57 .
Genus level diversity of the SRB is shown in Fig. 3B and the main genera are summarized in Table 2. Desulfovibrio is the main and dominant genus in produced water from both the Galo de Campina (16,08%) and Periquito (29,7%) oilfields. Next Pseudomonas in Galo de Campina (19,88%) and Shewanella in Periquito (4,69%). Some bacterial sequences were too divergent for classification at the genus level but their order level identity is shown in Fig. 3A. It is possible that the 300 bp sequence length explains this lack of taxonomic resolutionç but also one can infer, but not confirm, that novel SRB genera are present in both oilfields. Clostridium were detected in low numbers from both oil fields. The Postgate C media, favoured the detection of Desulfovibrio over the other media though this trend was not statistically significant at P 0.95. Here we used three media with lactic acid as a carbon source but there are other SRB metabolic types with different types of electron donorsç 18 as such, our study might under represent the SRB diversity. A wide range of organic acids including acetate, propionate, butyrate, pentanoate, and hexanoate have been found in oil reservoirs 58 and are alternative electron donors. Bacteria of the genera Desulfovibrio and Clostridium are well described producers of hydrogen sulphide and involved in the corrosion of metallic structures 28,59 . According Christman et al. 60 Clostridiales' sequences are often reported in produced water, however, their role is uncertain because they might not be actively producing H 2 S and simply present in a dormant sporulated form.
Pseudomonas sequences were detected in much larger numbers at the Galo de Campina oilfield in comparison with Periquito and this could be a possible biomarker for that site. Pleomorphomona were all but exclusively   61 . Strains from the genus Desulfovibrio are established sulphide producers and thus represent a major concern for the oil and gas industry 62,63 . What stands out immediately from the results is the importance of the Desulfovibrio 64,65 . Desulfovibrio are known to cause corrosion, hydrogen sulphide production, souring and biofouling within the operational plants 71,72 . More than 220 species from 60 genera of SRB have been described; of which, the most commonly isolated mesophilic SRB from produced water are from the Desulfovibrio genus 64 . Also of note at the genus level were the Pseudomonads at Galo de Campina, and the large number of (unidentified at genus level) Enterobacteriales. The frequent detection of Pseudomonas and unclassified Enterobacterales in the enrichments from the Galo de Campina are also of high research interest. To further characterize these taxa it will be necessary to isolate and properly describe these produced water taxa. The genera Pleomorphomonas and Shewanella were more abundant in samples from the Periquito oil field. Pleomorphomonas was first described in 2005 by Xie and Yokota 66 and is genus composed of gram-negative, pleomorphic, nitrogen-fixing, non-spore-forming, non-motile rods and first isolated from rice fields 67,68 . At the time of revision, there is no data correlating the presence of strains of this genus with problems in oil industry. Shewanella are facultative anaerobic, gram-negative, motile and rod-shaped bacteria, most of which have been isolated from marine environments, such as seawater, marine sediments or sand, tidal flats or marine invertebrates. Some species have, however, been isolated from clinical samples, oilfield fluids, activated sludge and coal-mine sludge 69,70 . Shewanella strains were described as a potential hazard to the oil industry causing souring of crude oil 71 .
A deeper analysis at the genus level was undertaken to evaluate if there were significant differences in the distribution of genera between the oilfields, results are shown in Fig. 4.
In Fig. 4A, the DESeq2 bar plot shows on the x-axis the log2-fold-change of abundance of the orders that showed the greatest variation between oil fields. The DESeq2 method selected the 8 most relevant genera to differentiate the two oil fields. After applying the adjusted p-value < 0.05 and log2FoldChange > 2 criteria, 3 orders were selected, namely Clostridiales, Pseudomonadales and Rhizobiales. Figure 4B illustrates the variation of these OTUs in each oil field, where it can be seen that there was a significant variation (BH adjusted p-value < 0.05) between the two fields analyzed. It is interesting to note that Clostridiales and Rhizobiales showed greater abundance in the Periquito field, while Pseudomonadales was more abundant in the Galo de Campina field.
The limited differences between the oil fields is in agreement with the enrichment methodology; and, work carried out by Kim et al. 72 who analyzed the microbial community in produced water collected from four geographically distant oil reservoirs and reported highly similar microbial communities in oil reservoirs in geographically distant locations. One can infer that geographical location is a less important driver of diversity in produced water than are oil reservoir conditions (temperature, pressure, pH) and that similar conditions in oil reservoirs naturally select similar groups of organisms. The order Rhizobiales is constantly found in microbial communities related to oil [73][74][75][76][77] and according to Angelova et al. 73 , strains from this genus are involved in polymer degradation, biofilm formation, and sulphur oxidation.
The data obtained here from SRB enriched produced water and with large scale NGS sequencing is consistent with the results reported in literature based on isolation methods and on nucleic acid methods where organisms were neither enriched nor isolated 29,31,78 . Other groups of microorganisms such as the extremophile halophiles and halotolerant microorganisms could be present in produced water after hydraulic fracturing. Although these microorganisms were not the focus of our study, it is interesting to note their presence in   76,[79][80][81][82] . It is only at the genus level that differences between the oil fields become evident. Bacterial strains of the genus Pseudomonas are producers of extracellular polymeric substances that form biofilms within oil facilities 25 . The presence of mesophilic Pseudomonas strains that are sensitive to high temperatures is not believed to originate from pristine oil reservoirs. Their presence in systems with high temperature oil wells is likely to follow flooding with cooler produced water, and contamination with Pseudomonas strains that are very versatile heterotrophs with the competitive capability to survive and form biofilms in pipes with oil/water mixtures 29 . Zdanowski et al. 83 analyzed the anaerobic microbiome of subglacial samples and compared the phylogenetic diversity of native samples with enriched ones. Enrichment was done by incubating native sediments in Postgate C medium for 8 weeks using airtight bottles to emulate subglacial conditions. They found that Psychrosinus, Clostridium, Paludibacter, Acetobacterium, Pseudomonas, Carnobacterium, and Desulfosporosinus were more abundant in the enriched medium: Pseudomonas and Carnobacterium were found only in the enriched medium. Thus, the authors suggest that it may be that bacteria of the genus Pseudomonas have proliferated under enrichment conditions, depleting available oxygen in the early stages of the process, and then helping anaerobes to develop. In a study by Guo et al. 84 , the authors demonstrated that Pseudomonas strain sp. C27 has an enzymatic system to perform sulfide removal. The authors cultivated the C27 strain in an anaerobic environment and demonstrated that the sulfide metabolism occurred through the expression of succinate dehydrogenase, iron-sulfur protein oxidoreductase, serine hydroxymethyl transferase, and iron superoxide dismutase. Brahmacharimayum and Ghosh 85 analyzed the removal of sulphate in an anaerobic environment and reported that a P. aeruginosa strain was dominant in the consortium and that it was involved in reducing Sulphate. Tüccar et al. 52 , found Pseudomonas as the dominant genus in produced water from Diyarbakir oil fields in Turkey.The authors suggested that these strains may have been inoculated into the oil reservoirs through the injection of produced water and that strains adapt to the conditions of the reservoir to survive. Several studies have found Pseudomonas in oil reservoirs [86][87][88][89][90][91][92] , and according to Cui et al. 93 , Pseudomonas and Acinetobacter are genera that can effectively use crude oil as a carbon source, being able to survive and reproduce at the oil-water interface. Species of the genus Pseudomonas are facultative anaerobes capable of performing nitrification and nitrate reduction using various carbon substrates 94,95 . In a study by Braun and Gibson 96 , the authors reported on two bacterial strains of the genus Pseudomonas that were able to degrade, under anaerobic conditions, 2-aminobenzoate (anthranilic acid) to CO2 and NH 4 +. According to Cai et al. 97 , in the petroleum area, despite the fact that the oil is considered a hostile and toxic environment for these microorganisms, there is now a lot of evidence that demonstrates the presence and activity of these microbes in the crude oil. Cai et al. 97 analyzed the microorganisms found in oil and water samples and showed that the genus Pseudomonas dominated the oil samples 98 . The Pseudomonas may well soon be considered at problematic as the Desulfovibrio 99 .

Conclusion
DNA sequencing analysis of over 800 000sequences from the Galo de Campina and Periquito oil fields detected a dominant phylum, the Proteobacteria (93.8%) which was followed by the Firmicutes (6.2%). These phyla are typically described in all global studies of reservoir and produced water bacterial biodiversity. The rDNA data in this study therefore supports previous findings based on isolation and direct DNA extraction studies. Seven major orders were detected, Clostridiales and Bacilliales belonging to the phylum Firmicutes, and Desulfovibrionales, Pseudomonadales, Enterobacteriales, Alteromonadales and Rhizobiales that belong to the phylum Proteobacteria. Among the genera detected, we noted that Pseudomonas, Desulfovibrio and Clostridium were the dominant genera, and that bacteria from these genera are routinely isolated from samples of produced water. They are known to be involved in biofilm formation processes, production of hydrogen sulphide and corrosion of metallic structures. For these two onshore oil fields we now know which bacteria in the produced water responded to the experimental conditions and that could possibly respond to attempts to manage those bacteria. To address the challenge of SRB at these oilfields and elsewhere requires monitoring SRB communities and biocides that target specific SRB groups. Microbial communities are both the problem and a solution to corrosion and souring; to decreasing maintenance and operating costs as well as to minimising negative environmental impacts in Brazil and elsewhere. The next step is to target specific problematic strains and employ targeted biocides that are more efficient at eradicating them and are less toxic to the wider environment than chemical alternatives.