Desulfovibrio and Pseudomonas Dominated Enriched Produced Water Reinforcing their Importance in Oil elds and Production Processes

Samyra Tiburcio Federal University of Rio de Janeiro Andrew Macrae (  amacrae@micro.ufrj.br ) Federal University of Rio de Janeiro Raquel Peixoto Federal University of Rio de Janeiro Caio Rachid Federal University of Rio de Janeiro Felipe Mansoldo Federal University of Rio de Janeiro Daniela Alviano Federal University of Rio de Janeiro Celuta Alviano Federal University of Rio de Janeiro Davis Ferreira Federal University of Rio de Janeiro Fabrício Venâncio Capital Intelectual Instituto Interdisciplinar de P&D Doneivan Ferreira Capital Intelectual Instituto Interdisciplinar de P&D Alane Vermelho Federal University of Rio de Janeiro


Introduction
Petroleum is found in reservoir rocks typically between a gas cap at the top and water at the bottom 1 . Oil extracted from a reservoir rock reaches the surface as a mixture with some sediment, gas and water. Water naturally con ned in a reservoir is denominated Formation Water; after it has been extracted with oil and gas, it is denominated Produced Water 2 . Due to technical considerations (equipment and reservoir) and environmental legislation, produced water requires special attention and demands speci c management actions 3 . Oil from a reservoir is pumped through pipelines to a primary processing unit, where density is used to separate particles, water, oil and gas and the emulsi ed zone between the oil and water. At the beginning of the operation, the liquid phase is separated from the gas phase and the water and the emulsi ed zone are separated from the oil phase 4 . Before disposing of, or injecting the produced water into back into a well, the operator must treat it to meet physical, chemical and biological quality control legislation that aims to protect the natural environment and to prevent fouling of the reservoir. If an operator decides to reinject the water into the reservoir, it is treated for: bacteria, the presence of salts, and the potential generation/deposition of fouling materials. Bacteria can create integrity problems for assets and safety problems for personnel 5 . Salts and solids can create ow assurance problems causing strangulation within the tubes and pipes. Solids will also create permeability problems within the reservoir rock 6 . Regulations covering the management and fate of produced water are increasingly stringent but not globally standardised. Treatment is required before disposing of produced water to remove concentrations of oil and grease. In Brazil, produced water is legally regarded as an "e uent" by the National Environmental Council (CONAMA). CONAMA Resolution No. 357/2005 and its amendments (393/2007 and 397/2008) establish acceptable oil limits in produced water but do not mention microbial contamination.
Qualitative and quantitative characterization of the bacteria in produced water has not been a focus of studies in Brazil, nor globally and yet bacteria can cause signi cant problems and are sometimes the solution to oil production problems (MEOR). It is interesting to identify bacteria present in produced water to understand the microbial community as well to detect harmful bacteria such a as sulphate reducing bacteria (SRB). Many papers discuss the origin of microbial populations in oil reservoirs. It is accepted that microorganisms are introduced into the wells during the drilling process, and/or by the injection of produced water or naturally through venting with ocean water offshore and ground water onshore 7 . In these instances, the microbial communities are a combination of native microorganisms in the well, and microorganisms present at the surface portion of the well added by drilling or injection 8,9 . Oil production faces four microbially attenuated SRB problems and they include: (i) microbial induced corrosion (MIC) in and on pipelines, metal structures and equipment 10 ; (ii) contamination of injection wells; (iii) impact of bioaccumulation (biofouling) on production and ow rates 11 ; and (iv) biogenic acidi cation (souring), which is caused by sulphide production in the form of Hydrogen Sulphide (H 2 S) 12 . When a high sulphate source is introduced into an oil reservoir already colonized by SRB, sulphate is reduced to sulphide thus oxidizing the organic electron donors present in crude oil, causing souring and a loss in commercial value of the oil 13 . To avoid this problem, biocides have been and are used, but without microbiological monitoring and resistance studies to assess the e cacy and impact. In this context, research that seeks new technologies and new biocidal substances must be combined with microbiological studies and not just toxicological effects for a more e cient management and maintenance of assets. Numerous studies on bio lms show that most of the biocidal substances have limited bio lm penetration, thus killing super cial cells and are more effective on planktonic cells [14][15][16] . Accurate knowledge of the microorganisms present in an oil eld and its produced water is critical for selecting and monitoring the action of antimicrobial substances. It is also critical in order to establish the required/optimum amount of biocide, its duration of use, and stage of application. There is a clear need to stop bio lms establishing because treating them subsequently is much more challenging.
A wide variety of SRB and other microorganisms have been isolated or detected in samples from oil reservoirs and produced water, including aerobic bacteria, facultative anaerobes, microaerophilic bacteria, strict anaerobes, archaea, thermophilic organisms, mesophilic and hyper thermophilic organisms 8 . Some functional groups of microorganisms such as SRB are frequently described as present. Other bacteria routinely found in reservoirs, include members of genera Clostridium, Pseudomonas and Bacillus and functional groups including the Acid Producing Bacteria (APB), Oxidizing Bacteria (OB), Sulphur Reducing Bacteria (SRB), Oxidizing Iron Bacteria (OIB), and Methanogenic Archaea (MA) [17][18][19][20][21] .
Petrobras (Brazil's State controlled International Oil Company) is a major global oil producer and exporter and of great economic and technological importance for Brazil. Petrobras proactively researches all aspects of exploration and production and very often with academic partners. Petrobras's studies on bacterial communities and especially SRB in produced water are ongoing and to date bacterial strains from the following genera have been isolated from Brazilian oil elds Curtobacterium, Brevundimonas, Brachymonas, Streptomyces, Bacillus, Pseudomonas amongst others 22,23 . Bacteria belonging to the Firmicutes are commonly isolated and within them the Bacilli and Clostridia Classes are most often reported. Speci c studies that have isolated bacteria associated with pipeline corrosion and bio lms reported strains from the genera Marinobacter, Colwellia and Pseudomonas. Strains from these genera are known to stimulate corrosion, producing large amounts of extracellular polysaccharides that protect sulphate-reducing bacteria (SRB) from biocides. Isolates from these genera have been shown to contribute to corrosion and to produce hydrogen sulphide impacting negatively on the market value of the crude oil and a threat to staff safety [22][23][24][25][26] .
In this study, we isolated cultivable bacteria from two Brazilian oil elds, Galo de Campinas and Periquito, both located in Rio Grande do Norte. Very little was known about the bacterial component of these wells. Metagenomic molecular biology methods were used to investigate samples and to establish a biodiversity benchmark as well as identifying potential biomarkers for the oil elds. 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-minutes, before being clamped with rubber and aluminium cap. The sealed tubes were incubated for 20 days at 30 ° C under anaerobiosis in Gaspak jar. For each oil eld there were nine samples, 3 of each type of medium. After 20 days bacteria from each of the 18 samples was collected by centrifugation for DNA extraction.

DNA extraction and Molecular identi cation of bacteria
Genomic DNA was extracted with the Wizard® Genomic DNA Puri cation Kit (PROMEGA) and quanti ed using QUBIT uorometer (Thermo Fisher Scienti c). DNA purity and quality were evaluated by 1% agarose gel electrophoresis containing SYBR® Safe DNA gel stain (Life Technologies™) at 90 V in 0.5X TBE buffer for 1 h and visualized by transillumination under ultraviolet light. DNA was PCR ampli cation of the 16S rDNA. The V3 and V4 variable region of the 16S rDNA was ampli ed by PCR, using the primers 338F 29 and 806R 30 with a barcode on the forward primer, in a 30 cycle PCR using the HotStarTaq Plus Master Mix Kit (Qiagen, USA) under the following conditions: 94 °C for 3 minutes, followed by 28 cycles of 94 °C for 30 seconds, 53 °C for 40 seconds and 72 °C for 1 minute, after which a nal elongation step at 72 °C for 5 minutes was performed. After ampli cation, PCR products were checked in 2% agarose gel to determine the success of ampli cation and the relative intensity of bands. Samples were puri ed using calibrated Ampure XP beads. Puri ed PCR product was used to prepare the DNA library by following the Illumina TruSeq DNA library preparation protocol. Paired-end Sequencing was performed at MR DNA (www.mrdnalab.com, Shallowater, TX, USA) on a MiSeq following the manufacturer's guidelines. After sequencing, the two reads from the paired end sequencing were joined together after q25 trimming of the ends, with the MrDNA pipeline prior to further analyses.

Sequence identi cation and bioinformatic analyses
The raw joined sequences were processed using mothur v.1.36.1 software 31 . The sequences were trimmed using trim.seqs command, with the following parameter: qwindowaverage = 30, qwindowsize = 50, maxambig = 0, minlength = 410, maxlength = 500. The sequences were then aligned using the Silva database as a reference 32 and the resultant alignments were submitted to screen.seqs and lter.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 33 , chimeric sequences were removed from further analyses. The remaining sequences were then classi ed using classify.seqs command, with RDP database as reference and a bootstrap cut-off of 80. Sequences classi ed as chloroplasts, mitochondria, eukaria, archaea and those not assigned to any Kingdom were removed. 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. 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 and to establish the relationship between samples, using a NMDS analysis.

Heat map analysis
To illustrate the heat map results, some modi cations were made to the methodology proposed by Pagé et al. 34 . A heatmap and a dendrogram with the Bray-Curtis dissimilarity index were generated from the most abundant bacterial families and the community relationships calculated using the vegdist function of the Vegan package 35 for R. Heat maps were created using the heatmap.2 of the gplots package 36 and the dendrograms generated by the hclust function from R's statistical package. Bacterial families whose relative reading abundance was less than 1% of at least one sample were removed. These analyses were performed in R software 37 .

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 sequence reads were generated, all the sequences studied were bacterial. The presence of some Archaea sequences would not have been a surprise as they are commonly found with this methodological approach in DNA derived from oil reservoirs [38][39][40] .
Crenarchaeota, Euryarchaeota were the most common archaea phyla found in produced-water 19 . The bacterial diversity indices of the produced water for the Galo de Campina (GC) and Periquito (PQO) oil elds are shown in Table 1.  (Fig. 2) indicate that differences seen between samples is explained by differences between the two oil elds and not the enrichment media. This support an inference that differences at the genus taxonomic rank between the two oil elds is real and not methodological.  46 , analyzed microbial abundance in produced water samples in the Bakken region by qPCR. the analysis was done directly from produced water, without a pre-enrichment medium. The anaerobic, fermentative Firmicutes orders (Bacillales and Halanaerobiales) and the Proteobacteria Order Pseudomonadales were the most abundant taxa across all evaluated samples, representing between 57% and 99% of the total microbial population. Despite the differences in geographical locations of oil elds, their physicochemical properties and methods used for analyses, it is interesting to note that these two phyla dominate and perhaps not surprising if we consider conditions within a reservoir will strongly select for certain groups of bacteria. No differences in bacterial diversity were observed by using three enrichment media at phylum level. The dominant phyla are dominant regardless. Use of enrichment is a deliberate strategy for recovering cultivable bacteria with desired metabolic rates 47 . The main Orders detected in the produced water from Galo de Campinas e Periquito were Desulfovibrionales, Pseudomonadales, Enterobacteriales, Clostridiales, Alteromonadales, Rhizobiales and Bacillalles and their distributions are shown in (Fig. 3). Li et al. 44 , in his work on bio lms in water injection systems in the Daqing oil eld (China), also identi ed Pseudomonadales, Enterobacteriales, Clostridiales and Desulfovibrionales orders, some organisms of these orders are already known to cause problems associated with hydrogen sulphide production, bio lm formation and biocorrosion 8,9,19 . Sequences from the order Desulfovibrionales were present in all samples and it was the dominant order found in both oil elds accounting for 46% of all sequences. The majority of bacteria in this order are SRB and are responsible for biocorrosion and possibly related to problems with hydrogen sulphide at the Periquito oil eld (personal communication). Pseudomonadales was predominant in Galo de Campinas oil eld and the, Enterobacteiriales in both oil elds. In the Galo de Campinas oil eld, no signi cative difference was observed among the media. The Firmicutes, and Proteobacteria Orders seem to have a worldwide distribution in oil eds and are abundant in produced water. Sun et al. 48 isolated 61 phylogenetic groups that belong to 32 genera in the phyla Actinobacteria, Firmicutes, and Proteobacteria in oil-production water from the Karamay Oil eld, Xinjiang, China. The Enterobacteriales and Alteromonadales orders that were detected in this study are in agreement with other reports in the literature that relate these orders to oil environments 49,50 . Liu and Liu 51 analyzed the bacterial community of oil collected from the sea surface of the northern Gulf of Mexico, as a result they found high proportions of Alteromonas, Marinobacter, Thalassospira, Bartonella, Rhodovulum and Stappia. And they point out that Marinobacter and Alteromonas, Gammaproteobacteria, are common oil-degrading microorganisms. This result is in agreement with Bacosa et al. 52 who analyzed the bacterial diversity of the same region and concluded that Alteromonas is an important class of bacteria in the fate of oil, being effective in degrading the alkanes in oil. Wang et al. 53 studied the microbial community of oil-polluted soil on agricultural land in Fushun, Liaoning province in China, where they attested that the abundance of Enterobacteriales was greater in areas with oil-contaminated soil. Figure 4 provides analysis genus level diversity of the produced water and the main genera are summarized in Table 2. Desulfovibrio is the main and dominant genus in produced water from Galo de Campinas (16,08%) and Periquito (29,7%) followed by Pseudomonas in Galo de Campinas (19,88%) and Shewanella in Periquito (4,69%). Some bacterial sequences were too divergent for classi cation and that was the case for both oil elds. Clostridia were detected in low numbers from both oil elds. Postgate C, in the majority of the samples, favoured Desulfovibrio isolation over the other media used. Pseudomonas was detected only in groups GCc and GCb, being a possible biomarker for these sites. Pleomorphomona were identi ed only at Periquito. Desulfovibrio, Pseudomonas and Clostridium are routinely found in produced water samples where strains from the Genus Desulfovibrio are noted as of major concern for the oil and gas industry. Comparing Fig. 4 with Fig. 3 we note that the gures look similar. This happens because a limited number of dominant genera are responsible for the Order level diversity. What stands out immediately is the importance of the Desulfovibrio sequences 40,54 . Desulfovibrio impact negatively causing corrosion, hydrogen production, souring and biofouling caused by bio lm formation within the operational plants. More than 220 species from 60 genera of SRB have been reported; of which, the most commonly isolated mesophilic SRB from produced water are from the Desulfovibrio genus 40 . Also of note were the Pseudomonads from two of the three Galo de Campinas sites, and, the as yet unidenti ed bacteria at genus level were from the Enterobacteriales. Pelobacter, Marinobacterium and Geotoga were detected in produced water from Petrobras Ilha Grande Bay Oil Terminal in Brazil, as well as the SRB Desulfoplanes formicivoran 55 . Bacteria of the genera Desulfovibrio and Clostridium are producers of hydrogen sulphide, it is toxic and accelerates the corrosion of metallic structures 20 and is composed of Gram-negative, pleomorphic, nitrogen-xing, non-spore-forming, non-motile rods 61,62 . At the time of writing, there is no data correlating this genus to 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 ats or marine invertebrates. Some species have, however, been isolated from clinical samples, oil eld uids, activated sludge and coal-mine sludge 63,64 . Shewanella strains were described as a potential hazard to the oil industry causing souring of crude oil 65 . * unclassi ed genera from Galo de Campinas is 11,05% and Periquito 7,62 In Fig. 5, the heat map provides a clearer representation of the genus level differences between the sites and oil elds. The heatmap illustrates the mutual occurrence of Pseudomonas and Desulfovibrio only at the GCb site. The genera Anaerosalibacter, Pleomorphominas and Shewanella are important and exclusive biomarkers of the Periquito oil eld. Non-metric multidimensional scaling analysis of Galo de Campina and Periquito oil elds (Fig. 2) was used to study the correlation between the three sampling locations with the two oil elds (PQO and GC) and culture media (Postgate B, Postgate C and Baars) and the in uence of each on the diversity detected. That analysis indicated that sampling points within each oil eld and culture media did not signi cantly in uence the microbial populations detected and the main difference in microbial distributions at the genus level was related to the two oil elds.
The metagenomic data obtained here by enriching produced water and then large scale PCR-DNA sequencing is consistent with the results reported in literature based on isolation methods and on methods where organisms were neither enriched nor isolated 21,23,66 .
It is only at the genus level that differences between individual wells and oil elds become evident. Bacterial strains of the Genus Pseudomonas are producers of extracellular polymeric substances that form bio lms within oil facilities 17 . 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 ooding with cooler produced water, and contamination with Pseudomonas strains that are very versatile heterotrophs with the competitive capability to survive including formation of bio lms in oil/water mixtures 21 .
Zdanowski et al. 67 analyzed the anaerobic microbiome of subglacial samples. In their work, the authors compared the phylogenetic diversity of native samples with enriched ones. This enrichment was done by incubating native sediments in Postgate C medium for 8 weeks using airtight bottles to emulate subglacial conditions. As a result, they found that the following genera were found more abundantly in the enriched medium: Psychrosinus, Clostridium, Paludibacter, Acetobacterium, Pseudomonas, Carnobacterium, and Desulfosporosinus. 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. Kliushnikova et al. 68 described a microorganism of the genus Pseudomonas with sulphate-reducing activity. According to the authors, this strain when grown under strictly anaerobic conditions was able to reduce sulphate more intensely than under aerobic conditions. In a study by Guo et al. 69 , the authors demonstrated that Pseudomonas strain sp. C27 has an enzymatic system to perform sul de removal. The authors cultivated the C27 strain in an anaerobic environment and demonstrated that the sul de metabolism occurred through the expression of succinate dehydrogenase, iron-sulfur protein, oxidoreductase, serine hydroxymethyltransferase, and iron superoxide dismutase. Brahmacharimayum and Ghosh 70 analyzed the removal of sulfate in an anaerobic environment by metagenomics. As a result, they show that the P. aeruginosa strain was predominant in the consortium and that it was involved in reducing Sulfate. Tüccar et al. 42 , found Pseudomonas as the dominant genus in produced water from Diyarbakir oil elds in Turkey and the authors suggested that these strains may have been inoculated into the oil reservoirs through the injection of uids, and point out that these strains may adapt to the conditions of the reservoir to survive.
Several studies have found the occurrence of the genus Pseudomonas in oil reservoirs [71][72][73][74][75][76][77] , according to Cui et al. 78 , 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 nitri cation and nitrate reduction using various carbon substrates 79,80 . In a study by Braun and Gibson [11], the authors reported two bacteria 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. 81 , in the petroleum area, despite the fact that the oil is considered a hostile and toxic environment for microorganisms, there is lot of evidence that demonstrates the presence of microbes in the crude oil. The authors analyzed the microorganisms found in oil and water samples from four oil wells. According to them, as an unexpected result, they found that the genus Pseudomonas dominated the oil samples, where several groups of functional genes were identi ed.
According to Arai (2001) 82 , Pseudomonas aeruginosa has a remarkable ability to grow in the most diverse environmental conditions, such as in soil and water and on and in animals, humans and plants.
This versatility is related to its metabolic exibility through a branched respiratory chain with multiple terminal oxidases and denitri cation enzymes. Its set of denitri cation enzymes are capable of reducing nitrate to nitrogen via nitrite, nitric oxide (NO), and nitrous oxide. Nitrogen oxides function as electron acceptors, which allows P. aeruginosa to grow in anaerobic conditions. It is interesting to note that in the absence of nitrate, P. aeruginosa is able to metabolize arginine via arginine deiminase and in the absence of both, it can ferment mixed acid pyruvate and survive for long periods in anoxic conditions 83 .

Conclusion
DNA analysis of over 4 million sequences from the Galo de Campina and Periquito oil elds detected a dominant phylum, the Proteobacteria (93.8%) 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 supports previous ndings based on isolation studies. Seven major orders were detected, Clostridiales and Bacilliales belong 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, bacteria that are routinely isolated from samples of produced water and are known to be involved in bio lm formation processes, production of hydrogen sulphide and corrosion of metallic structures were also detected. For these two oil elds 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.
Microbial dynamics and ecological studies will facilitate monitoring and guide the best use biocides in order to avoid biocorrosion and related processes. Microbial communities are both the problem and the solution to corrosion and souring; to decrease maintenance/operating costs as well as to minimise environmental preservation. These results can be used to help identify and target speci c antimicrobial agents, biotechnologies that are more e cient and less toxic to the environment resulting microbially enhanced oil recovery.

Funding
This work was supported in part by the Coordenação de Aperfeiçoamento Pessoal de Nível Superior-

Declaration of Competing interests
The authors report no declarations of interest.