The Potential Application of Microorganisms for Sustainable Petroleum Recovery from Heavy Oil Reservoirs

: A microbial enhanced oil recovery (MEOR) technique was tested at low-temperature heavy oil reservoirs (Russia). The bioaugmentation approach used is based on the introduction of hydrocarbon-oxidizing bacteria into the oilﬁeld in combination with an injection of oxygen as a H 2 O 2 solution in order to initiate the ﬁrst stage of hydrocarbon oxidation and of (NH 4 ) 2 HPO 4 as a source of biogenic elements. Before the pilot trials, the microorganisms of petroleum reservoirs were investigated by high-throughput sequencing, as well as by culture-base and radioisotope techniques. Molecular studies revealed the di ﬀ erences in microbial composition of the carbonate and terrigenous oil reservoirs and the communities of injection and formation water. Aerobic bacteria Rhodococcus erythropolis HO-KS22 and Gordonia amicalis 6-1 isolated from oilﬁelds oxidized oil and produced biosurfactants. Fermentative enrichment and pure cultures produced considerable amounts of low fatty acids and alcohols from sacchariferous substrates. In core-ﬂooding tests, 43.0–53.5% of additional heavy oil was displaced by aerobic bacteria, producing biosurfactants, and 13.4–45.5% of oil was displaced by fermentative bacteria, producing low fatty acids, alcohols, and gas. A total of 1250 t additional oil was recovered as a result of the application of an MEOR technique at the Cheremukhovskoe heavy oil reservoir and Vostochno-Anzirskoe reservoir with light conventional oil.


Introduction
The estimated reserves of heavy oil and bitumen in Russia are 6-7 billion tons, exceeding the remaining reserves of conventional oil, and may be considered the main basis for the development of oil production in the next few years. The most promising methods for heavy oil recovery are thermal methods, which decrease the oil viscosity (fire flooding, steam flooding, and cyclic or joint injection of steam, solvents, and surfactants). However, the application of these methods requires a large energy

Characteristics of the Studied Heavy oil Reservoirs and Sampling Procedures
The microorganisms of injection and production water from petroleum reservoirs in Volgograd oblast and Tatarstan (Russia) were studied. Oil strata were located at the depth of 830-1650 m and had a temperature of 20-40 • C (Table A1). According to analyses of surface samples, oil density varied from 0.813 g/cm 3 to 0.924 g/cm 3 (at 20 • C). This made it possible to classify the oil from two out of four oilfields as heavy oil. The studied areas of the Arkhangelskoe and Romashkinskoe oilfields (Tatneft, Tatarstan) and Cheremukhovskoe and Vostochno-Anzirskoe oilfields (Ritek, Volgograd oblast) are exploited with water-flooding [23].
Water samples were collected directly at the head of injection wells and production wells into sterile bottles, hermetically sealed without air bubbles, and used for radioisotope and microbiological analyses on the day of sampling. Within 4-6 h after sampling, inoculations were carried out in order to determine cell numbers of cultivated microorganisms. To estimate the rates of sulfate reduction and methanogenesis, aliquots of the water samples were supplemented with labeled compounds (Na 2 35 SO 4 , NaH 14 CO 3 , and 14 CH 3 COONa) and analyzed as described previously [24]. The samples for chemical analyses were stored at 6 • C. For molecular biological research, water samples (1 L each) were fixed with ethanol (1:1 vol/vol) at the day of sampling, filtered through 0.22 µm membranes (Millipore, United States), and stored at −20 • C prior to analysis.

Analysis of Microbial Diversity in Heavy oil Reservoirs
The composition of microbial communities from injection and formation water was determined by high-throughput sequencing of the 16S rRNA genes on the Illumina platform. Cell biomass obtained by the filtration of ethanol-fixed water samples was washed off with the lysing solution containing 0.15 M NaCl and 0.1 M Na 2 EDTA (pH 8.0) and used for DNA isolation. Total DNA was isolated using the PowerSoil DNA Isolation Kit (MoBio, USA) according to the manufacturer's recommendations and was stored at −20 • C. The purified DNA preparation was used as a template for polymerase chain reaction (PCR) with a pair of universal primers to the V3-V4 regions of the 16S rRNA gene: 319F (5 -ACTCCTACGGGAGGCAGCAG-3 ) and 806R (5 -GGACTACHVGGGTWTCTAAT-3). The primers were supplemented with oligonucleotide identifiers for sequencing on MiSeq (Illumina, USA). Sample preparation and sequencing were carried out according to the manufacturer's recommendations. Quality control of obtained reads was carried out using UPARSE [25]. After qualification, the reads were grouped into operational taxonomic units (OTUs) with 97% similarity using USEARCH (https: //www.drive5.com/usearch/) [26]. The taxonomic position of the representative sequence for each OTU was determined using the RDP (https://rdp.cme.msu.edu/tutorials/classifier/classifer_cover_page.html) classifier [27]. Analysis of community composition using heatmaps was carried out using ClustVis [28]. Statistical analysis was carried out using Microsoft Excel, Rstudio (the vegan package) [29].

Media Composition and Cultures Used in the Study
Cell numbers of the microorganisms of the main metabolic groups were determined by inoculation of ten-fold dilutions of the water samples into liquid media in triplicate, as described earlier [23]. Salinity of the studied water samples was taken into consideration when preparing the media. Samples of injection water were inoculated into the media containing 10 g NaCl per liter. Samples of high-salinity formation water were inoculated into the media with 45 g NaCl per liter. Aerobic organotrophs were enumerated in trypton-extract-glucose (TEG) medium containing the following (g·L −1 ): Bacto tryptone, 5.0; yeast extract, 2.5; glucose, 1.0; and NaCl, as required; pH 7.0-7.2. The numbers of hydrocarbon-oxidizing bacteria were determined using a mineral medium supplemented with a mixture of C 10 -C 22 n-alkanes (2% vol/vol). Anaerobic fermentative bacteria were enumerated in the medium supplemented with peptone (4 g·L −1 ) and glucose (10 g·L −1 ). The numbers of sulfate-reducing bacteria were determined by an increase in sulfide concentration in dilution series of formation water in Postgate B medium with sodium lactate (3.5 g·L −1 ), supplemented with microelements and reduced with Na 2 S·9H 2 O (200 mg·L −1 ). The numbers of methanogens were assayed by methane production in dilution series in the Zeikus medium with acetate (2 g·L −1 ), methanol (1 mL·L −1 ), and H 2 /CO 2 (4:1), supplemented with microelements and yeast extract (1 g·L −1 ). The composition of the media used is cited in the publication by Bonch-Osmolovskaya et al. [24]. Inoculated media were incubated for 30 days at 22-24 • C and then examined using an Olympus microscope with a phase contrast device.
Aerobic heavy oil degrading bacteria Gordonia amicalis strain 6-1 (=VKM Ac-2795D) and Rhodococcus erythropolis strain HO-KS22 (=VKM Ac-2807D) were isolated earlier from the Cheremukhovskoe and Vostochno-Anzirskoe oilfields (Russia), respectively [23]. Fermentative enrichment 7SA was obtained by inoculation of the medium with peptone and glucose with water injected into well 1010 of the Vostochno-Anzirskoe oilfield. Pure culture of a fermentative bacterium Oscillibacter ruminantium strain AIK was obtained from the sludge of a biogas reactor processing oil-containing and municipal waste [30]. For core flooding experiments, O. ruminantium strain AIK was grown in the medium for fermentative bacteria (FM) containing the following (g·L −1 distilled water): KH 2 [31]. Argon was used as the gas phase. Fermentative enrichment 7SA was grown in liquid mineral FM medium supplemented with glucose (2.0 g·L −1 ) and sucrose (2.0 g·L −1 ) as substrates, with 10.0 g·L −1 NaCl; without addition of the reducing agent. Strain AIK and fermentative enrichment 7SA were grown for 20 days under stationary conditions at 25 • C. Grown liquid cultures with microbial biomass and metabolites were used in core flooding experiments on the enhancement of oil recovery.
Oil biodegradation by aerobic hydrocarbon-oxidizing bacteria was studied in the medium of the following composition (g·L − 30.1% resins, and 4.1% asphaltenes. The cultures and uninoculated medium with oil (control) were incubated for seven days on a shaker (60 rpm) at 28 • C. After incubation, oil fractions were separated as described previously [32]. Residual oil was extracted with chloroform; the extract was dried with Na 2 SO 4 , filtered, and the solvent was removed on a rotary evaporator. Oil was then dispensed into vials, dried for 24 h under draught, and analyzed by gas-liquid chromatography (GLC).

Analytical Methods
The contents of methane, hydrogen, and carbon dioxide accumulated in the gas phase of the primary enrichment cultures were determined by gas chromatography. Sulfide was determined by the colorimetric method of Pachmayr with dimethyl-p-phenylenediamine [33]. The chemical composition of formation water was determined as described by Bonch-Osmolovskaya et al. [24]. Volatile fatty acids and alcohols were analyzed with a Shimadzu GC 2010 Plus chromatograph (Japan) in a column (30 m × 0.32 mm) with a Zebron ZB-FFAP phase thickness of 0.25 µm, as described previously [34]. The surface tension was measured at the interface between formation water and air. The interfacial tension was determined at the interface between the studied liquid and the n-hexadecane by the ring-tearing-off method using a Kruss K10 ST tensiometer (Kruss GmbH, Germany) at 22 • C.
The aliphatic fraction of oil was analyzed on a Crystal 5000.1 gas-liquid chromatograph (Chromatec, Russia) equipped with a flame ionization detector and a 15 m ZB-FFAP capillary column, with helium as the carrier gas. The column temperature changed from 100 • C to 320 • C in the course of analysis at the rate of 5 • C/min.

Core Flooding Experiments
The oilfield models were constructed using several cores of carbonate rocks with intact structure; each core was~30 mm in diameter and~40 mm long. The cores were cleaned from hydrocarbons by using the Soxhlet extraction method, using benzene and methanol in the proportion of 75:25 vol/vol. After cleaning, the cores were dried at 100-110 • C for 24 h, and their absolute permeability was determined. The columns packed with two cores each were 8.0 cm in length and 3.02 cm in diameter and were made of stainless steel. The cores were saturated with filtered formation brine with 195 g·L −1 salinity using vacuum desiccators for 24 h, and pore volume (PV) was determined using the dry and wet weights of each core, as described previously [16]. The cores were then flooded with heavy crude oil with a density of 0.93 g/cm 3 and viscosity of 190 mPa·sec at 1 cm 3 /min until no more water was removed. This filtration rate was chosen since microbial processes in the near-bottom zone of injection wells were simulated, which is characterized by a high rate of injected water flow. The oil initially in place (OIIP) was determined, which was indicated by the volume of water displaced. The core was subjected to water flooding at 1 cm 3 /min until no further oil was removed. The residual oil was calculated by measuring the amount of oil recovered by water-flooding. After preliminary operations, the model was incubated for one to seven days at the experimental values of temperature and pressure. Then, 10 PV of the grown liquid culture was injected as a tertiary recovery stage and extra oil recovery was determined. All core floods were conducted at 7-8 MPa and 22-24 • C to mimic the pressure and temperature of the oilfield, respectively.

Nucleotide Sequence Accession Numbers
The 16S rRNA gene sequences of strains Gordonia amicalis 6-1 (= VKM Ac-2795D), Rhodococcus erythropolis HO-KS22 (= VKM Ac-2807D), and Oscillibacter ruminantium AIK were deposited in the NCBI database under accession numbers MN101280, MN622878, and MN623416, respectively. The libraries of the 16S rRNA gene fragments, obtained as a result of high-throughput sequencing, were deposited in the NCBI Sequence Read Archive (SRA) under accession number PRJNA589080.

Physicochemical and Microbiological Characteristics of Injection and Production Water of the Studies Oilfields
The petroleum reservoirs studied are exploited with water-flooding with fresh water from the surface or with its mixture with formation water, separated from oil. The injection of fresh water results in decreased salinity in the area of water mixing. Formation water from the Arkhangelskoe heavy-oil reservoir with carbonate collectors had salinity from 36 g·L −1 to 244 g·L −1 , belonged to the chlorine-calcium type, and contained high levels of bicarbonate (0.12 g·L −1 to 0.70 g·L −1 ) and sulfates (0.44 g·L −1 to 1.58 g·L −1 ) ( Table 1). In the terrigenous Cheremukhovskoe heavy oil reservoir, production water with salinity from 12 g·L −1 to 186.9 g·L −1 belonged to the chlorine-calcium type; the content of bicarbonate in studied water samples varied within a narrow range from 0.29 g·L −1 to 0.32 g·L −1 , and that of sulfate from 0.08 g·L −1 to 0.22 g·L −1 . At the terrigenous Romashkinskoe oilfield (Almetievskaya bed) with light conditioned oil, which was used for comparison, the salinity of production water samples varied within a wide range exceeding 190 g·L −1 . Fresh river water injected into the Almetievskaya bed through injection wells 20990* and 10055* had a density of~1.000 kg dm −3 . A tendency to higher abundance of culturable microorganisms in the desalinated zone compared to the highly mineralized formation water was observed. In the Arkhangelskoe heavy oil reservoir with carbonate collectors, the highest microbial numbers were found in injection fresh water (well 7528) and in the production water sample with salinity up 73 g·L −1 (well 4634) ( Figure A1). In production water with higher salinity, the numbers of microorganisms of the major physiological groups did not exceed tens of cells per mL. Sulfate reduction rate was registered by the radioisotope method in water samples with salinity below 105 g·L −1 , but methanogenesis rate was detected up to 244 g·L −1 (Table A2).
In the terrigenous Cheremukhovskoe heavy oil reservoir, the numbers of microorganisms were considerably lower in production water samples with salinity exceeding 72 g·L −1 , although fermentative microorganisms occurred at up to 186.9 g·L −1 (well 5464). In spite of the broad salinity range of formation water at the terrigenous Romashkinskoe oilfield (Almetievskaya bed) with light oil, the numbers of aerobic organotrophs, as well as of fermentative and methanogenic prokaryotes, were higher than in the studied heavy oil reservoirs. While sulfide formation was detected in fermentative enrichments after long-term incubation, cultured sulfate-reducing bacteria were not retrieved when medium with lactate and 45 g·L −1 NaCl was inoculated with formation water of the Romashkinskoe oilfield.

Phylogenetic Diversity of Archaea and Bacteria in Injection and Production Water
High-throughput sequencing of the V3-V4 region of the 16S rRNA genes was used to determine the composition of microbial communities in the samples of injection and formation water from the Arkhangelskoe and Romashkinskoe petroleum reservoirs. A total of 107,846 reads were obtained and operational taxonomic units (OTUs) were defined based on 97% identity. The 16S rRNA gene The sequences of Archaea were revealed in the libraries of production water from wells 4610 and 7806 of the Arkhangelskoe heavy oil reservoir, as well as in library 20963 from the Romashkinskoe oilfield, where they constituted 18.0%, 10.1%, and 2.1% of the total number of genes in the library, respectively. Microbial communities of these reservoirs with high salinity of formation water are characterized by a noticeable occurrence of methylotrophic prokaryotes (aerobic methylotrophic bacteria and anaerobic methylotrophic methanogenic archaea).
The libraries of production water from wells 4610 and 7806 of the Arkhangelskoe heavy oil reservoir revealed the sequences of aerobic organotrophic bacteria of the genera Bacillus, Pseudomonas, Methylobacterium, of anaerobic fermentative Halanaerobium, and of syntrophic bacteria Smithella and Syntrophus, typical for water-flooded petroleum reservoirs. Sequences of less frequently occurring bacteria of the genera Herminiimonas, Acidovorax, and Sphingomonas were also detected ( Figure 2).
The minor sequences belonged to fermentative bacteria of the genera Geotoga and Mesotoga and to sulfate-reducing bacteria of the genera Desulfocurvus, Desulfotignum, Desulfonatronovibrio, Desulfovibrio, Desulfosalsimonas, Desulfovermiculus, and Desulfoglaeba (PRJNA589080). Euryarchaeota sequences belonged mainly to organotrophic halophilic archaea of the genus Halomicrobium (class Halobacteria), which are capable of both aerobic and anaerobic (by nitrate reduction) growth, and to halophilic methanogenic archaea of the genus Methanohalophilus (class Methanomicrobia). The coccoid methylotrophic methanogen Methanococcoides euhalobius, capable of growth in the presence of 14% NaCl (wt/vol), was originally isolated from the Bodyuzhskoe oilfield of Tatarstan [35,36].
Among bacteria revealed in the libraries of production water 20963, 47B, and 2356D from the Romashkinskoe oilfield were halophilic sulfate-reducing bacteria of the genera Desulfosalsimonas and Desulfonatronovibrio, aerobic organotrophic bacteria of the genera Ochrobactrum, Pseudomonas, and Brevundimonas, as well as members of the genera Exiguobacterium, Blastomonas, Nitrosospira, which seldom occur in oilfields, and magnetotactic bacteria of the genus Magnetococcus. Members of the genus Exiguobacterium are facultatively anaerobic organotrophs [37]. Magnetococcus can grow chemolithoautotrophically with thiosulfate or sulfide as electron donors, and The sequences of Archaea were revealed in the libraries of production water from wells 4610 and 7806 of the Arkhangelskoe heavy oil reservoir, as well as in library 20963 from the Romashkinskoe oilfield, where they constituted 18.0%, 10.1%, and 2.1% of the total number of genes in the library, respectively. Microbial communities of these reservoirs with high salinity of formation water are characterized by a noticeable occurrence of methylotrophic prokaryotes (aerobic methylotrophic bacteria and anaerobic methylotrophic methanogenic archaea).
The libraries of production water from wells 4610 and 7806 of the Arkhangelskoe heavy oil reservoir revealed the sequences of aerobic organotrophic bacteria of the genera Bacillus, Pseudomonas, Methylobacterium, of anaerobic fermentative Halanaerobium, and of syntrophic bacteria Smithella and Syntrophus, typical for water-flooded petroleum reservoirs. Sequences of less frequently occurring bacteria of the genera Herminiimonas, Acidovorax, and Sphingomonas were also detected ( Figure 2).
The minor sequences belonged to fermentative bacteria of the genera Geotoga and Mesotoga and to sulfate-reducing bacteria of the genera Desulfocurvus, Desulfotignum, Desulfonatronovibrio, Desulfovibrio, Desulfosalsimonas, Desulfovermiculus, and Desulfoglaeba (PRJNA589080). Euryarchaeota sequences belonged mainly to organotrophic halophilic archaea of the genus Halomicrobium (class Halobacteria), which are capable of both aerobic and anaerobic (by nitrate reduction) growth, and to halophilic methanogenic archaea of the genus Methanohalophilus (class Methanomicrobia). The coccoid methylotrophic methanogen Methanococcoides euhalobius, capable of growth in the presence of 14% NaCl (wt/vol), was originally isolated from the Bodyuzhskoe oilfield of Tatarstan [35,36].
of the genus Polynucleobacter, which utilize a narrow range of substrates (acetate, pyruvate, malate, etc.) under aerobic conditions or grow as facultative anaerobes, are known to be common in freshwater environments [39]. The chemoheterotrophic marine bacteria of the SAR11 clade Candidatus Pelagibacter, which were revealed in both samples of injection water, are "the Earth's most abundant organisms" [40]. The reason for the occurrence of these bacteria far from marine habitats remains unclear.  Among bacteria revealed in the libraries of production water 20963, 47B, and 2356D from the Romashkinskoe oilfield were halophilic sulfate-reducing bacteria of the genera Desulfosalsimonas and Desulfonatronovibrio, aerobic organotrophic bacteria of the genera Ochrobactrum, Pseudomonas, and Brevundimonas, as well as members of the genera Exiguobacterium, Blastomonas, Nitrosospira, which seldom occur in oilfields, and magnetotactic bacteria of the genus Magnetococcus. Members of the genus Exiguobacterium are facultatively anaerobic organotrophs [37]. Magnetococcus can grow chemolithoautotrophically with thiosulfate or sulfide as electron donors, and chemoorganoheterotrophically on acetate; thus, the conditions in the water-flooded oilfield are in agreement with their growth requirements [38].
Application of the principle component analysis (PCA) method revealed considerable similarity in the composition of microbial communities in production water samples from the same petroleum reservoirs (Figure 3). The 4610 and 7806 communities from Arkhangelskoe oilfield, where formation water contained high concentrations of sulfate, a diverse archaeal population, and sulfate-reducing bacteria as minor phylotypes, formed a separate group. Another separate group was represented by communities of production water from the Romashkinskoe oilfield (47B, 2356D, and 20963), where sulfate concentration in the water was low and various halotolerant aerobic organotrophic bacteria and halophilic sulfate reducers were present. Microorganisms of injection water communities formed two groups, which were different from formation water communities. sequences to the total number of sequences from each sample. Injection wells are marked with *, the rest are production wells.
Application of the principle component analysis (PCA) method revealed considerable similarity in the composition of microbial communities in production water samples from the same petroleum reservoirs (Figure 3). The 4610 and 7806 communities from Arkhangelskoe oilfield, where formation water contained high concentrations of sulfate, a diverse archaeal population, and sulfate-reducing bacteria as minor phylotypes, formed a separate group. Another separate group was represented by communities of production water from the Romashkinskoe oilfield (47B, 2356D, and 20963), where sulfate concentration in the water was low and various halotolerant aerobic organotrophic bacteria and halophilic sulfate reducers were present. Microorganisms of injection water communities formed two groups, which were different from formation water communities. Earlier analysis of the 16S rRNA genes of the microbial community from injection water 5600* of the Cheremukhovskoe oilfield revealed bacteria of the genera Dechloromonas (15%), Phenylobacterium (12%), hydrocarbon-oxidizing bacteria of the genera Acinetobacter (5%) and Rhodococcus (1.2%), and of sulfate-and sulfur-reducing bacteria of the genera Desulfovibrio, Desulfomicrobium, and Desulfuromonas [23]. Denaturing gradient gel electrophoresis (DGGE) analysis of the mcrA genes in injection water 5600* revealed methanogens of the genera Methanospirillum, Methanobacterium, and Methanoregula; production water from sample 5464 revealed methanogens of the genera Methanosarcina, Methanoculleus, and unidentified Thermoplasmata. These results indicate that both microorganisms from the surface freshwater basins and the reservoir microbiota adapted to oilfield conditions are delivered into the oil reservoir with injection water (a mixture of surface fresh water and formation water after oil removal). Earlier analysis of the 16S rRNA genes of the microbial community from injection water 5600* of the Cheremukhovskoe oilfield revealed bacteria of the genera Dechloromonas (15%), Phenylobacterium (12%), hydrocarbon-oxidizing bacteria of the genera Acinetobacter (5%) and Rhodococcus (1.2%), and of sulfateand sulfur-reducing bacteria of the genera Desulfovibrio, Desulfomicrobium, and Desulfuromonas [23]. Denaturing gradient gel electrophoresis (DGGE) analysis of the mcrA genes in injection water 5600* revealed methanogens of the genera Methanospirillum, Methanobacterium, and Methanoregula; production water from sample 5464 revealed methanogens of the genera Methanosarcina, Methanoculleus, and unidentified Thermoplasmata. These results indicate that both microorganisms from the surface freshwater basins and the reservoir microbiota adapted to oilfield conditions are delivered into the oil reservoir with injection water (a mixture of surface fresh water and formation water after oil removal).

Aerobic Bacteria Producing Biosurfactants
A number of aerobic organotrophic bacteria from the genera Rhodococcus, Gordonia, Nocardia, Pseudomonas, Bacillus, Salinicola, Chromohalobacter, and Cellulomonas were previously isolated from Russian heavy oil reservoirs [23]. These bacteria produced biosurfactants in the media with a variety of organic substrates (acetate, glucose, sucrose, ethanol, peptone, and glycerol), as well as in the medium with crude oil.
The strains Gordonia amicalis 6-1 and Rhodococcus erythropolis HO-KS22, which exhibited the most pronounced decrease in the surface tension, were studied in more detail. R. erythropolis HO-KS22 grew aerobically within broad ranges of salinity (0 g·L −1 to 80 g·L −1 NaCl, optimum at 10 g·L −1 ), pH (5.5-9.5, optimum at pH 7.0-7.5), and temperature (5-37 • C, optimum at 28 • C). G. amicalis 6-1 grew aerobically at 0 g·L −1 to 100 g·L −1 NaCl (optimum at 15 g·L −1 ), pH from 6.5 to 8.2 (optimum at 7.4), and at 5 • C to 45 • C (optimum at 28 • C). Biodegradation of heavy oil by G. amicalis 6-1 resulted in the utilization of C 11 , C 12 , C 13 , C 15-C21 , and C 23 n-alkanes ( Figure A2a,b). After seven days of incubation, R. erythropolis HO-KS22 consumed most of the n-alkanes from heavy oil ( Figure A2c). In the medium with crude oil, the strains G. amicalis 6-1 and R. erythropolis HO-KS22 decreased the interfacial tension from 21 mN/m to 0.2-1.0 mN/m. In the medium with glucose, G. amicalis 6-1 synthesized biosurfactants during the highest biomass accumulation (late logarithmic to early stationary growth phases), as was evident from the decrease in surface tension from 72 mN/m to 20 mN/m and of the interfacial tension against n-hexadecane from 47 mN/m to 1-2.1 mN/m ( Figure A3a-c). The most pronounced decrease in surface tension was observed at the glucose/nitrate ratio of 5/0.5 (g/g per 1 L); the highest decrease in interfacial tension occurred at the 20/1.0 ratio (g/g). Salinity and temperature optima of G. amicalis 6-1 and R. erythropolis HO-KS22 and the ability to consume crude oil indicate their adaptation to the conditions of the near bottom zone of injection wells, where dissolved oxygen is supplied by injected water; these physiological characteristics may be used in the application of strains 6-1 and HO-KS22 for injection into oilfields in order to increase oil recovery or to remove asphaltene-resin-paraffin deposits.
The microorganisms capable of decreasing surface tension by more than 10 mN/m are considered potential biosurfactant producers [41]. The isolation of biosurfactants from the media and investigation of their structure were beyond the scope of the present work. The surface tension decrease of the media caused by the isolates G. amicalis 6-1 and R. erythropolis HO-KS22 was compared to that caused by sulfonol, an anionic synthetic surfactant used in the oil industry for oil recovery enhancement. The biosurfactants present in liquid cultures of the strains decreased the surface tension to the same degree as a sulfonol solution at 1.3 g L −1 . Thus, G. amicalis 6-1 and R. erythropolis HO-KS22 are promising for application in the biotechnologies for enhanced oil recovery and for oil removal from contaminated environments. G. amicalis 6-1, which is also able to use dibenzothiophene as a sulfur source in the presence of other organic substrates, was patented [42].

Fermentative Bacteria Producing Low Fatty Acids and Alcohols
While the population of anaerobic fermentative bacteria was the most numerous one in formation water samples ( Figure A1), their growth on heavy oil was very poor (data not shown). The growth of fermentative enrichments in media with peptone and glucose did not result in surface tension changes. The studied fermentative enrichment 7SA and pure culture Oscillibacter ruminantium strain AIK fermented organic substrates with the production of low fatty acids, alcohols, and gases (H 2 and CO 2 ). The 7SA fermentative enrichment produced low fatty acids (acetic, 50.0 mg·L −1 ; propionic, 190.0 mg·L −1 ; and n-butyric, 3590.0 mg·L −1 ), the pH of the medium decreased from 7.0 to 5.5. O. ruminantium AIK produced low fatty acids (LFA) (acetic, 1895.0 mg·L −1 ; propionic, 390.0 mg·L −1 ; isobutyric, 370.0 mg·L −1 ; n-butyric, 3830.0 mg·L −1 ; and other LFA, 790.0 g·L −1 ), isopropanol, 110.0 g·L −1 ; and gases (H 2 and CO 2 ); pH decreased to 4.5. Volatile fatty acids are known to promote the dissolution of carbonate oil-bearing rocks and to increase their porosity, permeability, and oil recovery.
Methanogens were detected in a number of samples by molecular, cultural, and radiotracer techniques. It is, however, hardly probable that fermentative bacteria may provide enough substrates for methanogens for methane to have a significant effect on oil displacement, even in the presence of sugar-containing substrates. The cultures of aerobic bacteria and anaerobic fermentative bacteria were therefore used in core flooding experiments.

Core Flooding for Heavy Oil Recovery
The cultures of biosurfactant-producing aerobic bacteria and of fermentative bacteria-producing volatile acids and alcohols were grown in liquid media with sugar-containing substrates and then injected into the models. In these experiments only the effect of introduced liquid cultures of aerobic and anaerobic bacteria on heavy oil recovery was estimated. The distribution of bacteria in the pore space of the models and bacterial growth in time were not analyzed.
The oilfield models were prepared using the natural carbonate cores from oilfields and were saturated with heavy oil (density 0.93 g/cm 3 ). The cores were then flooded with NaCl solution (30 g·L −1 ), which was similar to formation water salinity, to 100% water content in the outflowing liquid. After primary oil displacement with the NaCl solution, the residual oil saturation of the cores varied from 67.3% to 80%, indicating that most of the water present in the column was replaced by oil. The results obtained with the carbonate cores are shown in Table 2.
Secondary core flooding with 10 PV of the liquid cultures of aerobic and anaerobic bacteria resulted in decreased residual oil saturation in the cores and additional oil displacement after MEOR. Injection of the liquid culture O. ruminantium AIK, which contained low fatty acids (7.27 g·L −1 ), isopropanol, and the biomass, was accompanied by additional displacement of 30.5% and 60.4% of oil (45.5% on average) from cores #1 and #2. Additional oil recovery from cores #3 and #4 flooded with fermentative enrichment 7SA was 11.6% and 15.3% (13.4% on average), which was probably the result of lower concentrations of the fermentation products in the culture 7SA (3.83 g LFA L −1 ).
The flooding of cores #5 and #6 with the aerobic culture R. erythropolis HO-KS22 resulted in the highest additional oil displacement of 60.6% and 46.4% (53.5% on average) (Figure 4). substrates for methanogens for methane to have a significant effect on oil displacement, even in the presence of sugar-containing substrates. The cultures of aerobic bacteria and anaerobic fermentative bacteria were therefore used in core flooding experiments.

Core Flooding for Heavy Oil Recovery
The cultures of biosurfactant-producing aerobic bacteria and of fermentative bacteria-producing volatile acids and alcohols were grown in liquid media with sugar-containing substrates and then injected into the models. In these experiments only the effect of introduced liquid cultures of aerobic and anaerobic bacteria on heavy oil recovery was estimated. The distribution of bacteria in the pore space of the models and bacterial growth in time were not analyzed.
The oilfield models were prepared using the natural carbonate cores from oilfields and were saturated with heavy oil (density 0.93 g/cm 3 ). The cores were then flooded with NaCl solution (30 g·L −1 ), which was similar to formation water salinity, to 100% water content in the outflowing liquid. After primary oil displacement with the NaCl solution, the residual oil saturation of the cores varied from 67.3% to 80%, indicating that most of the water present in the column was replaced by oil. The results obtained with the carbonate cores are shown in Table 2.
Secondary core flooding with 10 PV of the liquid cultures of aerobic and anaerobic bacteria resulted in decreased residual oil saturation in the cores and additional oil displacement after MEOR. Injection of the liquid culture O. ruminantium AIK, which contained low fatty acids (7.27 g·L −1 ), isopropanol, and the biomass, was accompanied by additional displacement of 30.5% and 60.4% of oil (45.5% on average) from cores #1 and #2. Additional oil recovery from cores #3 and #4 flooded with fermentative enrichment 7SA was 11.6% and 15.3% (13.4% on average), which was probably the result of lower concentrations of the fermentation products in the culture 7SA (3.83 g LFA L −1 ).
The flooding of cores #5 and #6 with the aerobic culture R. erythropolis HO-KS22 resulted in the highest additional oil displacement of 60.6% and 46.4% (53.5% on average) (Figure 4).   The flooding of cores #7 and #8 with liquid cultures of strains HO-KS22 and 6-1 in the 1:1 ratio (vol/vol) caused additional oil replacement of 54.2% and 31.8% (43.0% on average). The increase in oil recovery observed in core-flooding experiments #5-#8 was probably due to the ability of the aerobic Rhodococcus and Gordonia isolates to produce biosurfactants, both associated with their cell walls and released into the medium. Biosurfactants are known to be bipolar molecules concentrating at the oil-water interface and causing oil emulsification and a decrease in the interfacial tension between oil and water.
Additional oil displacement in experiments with fermentative cultures was probably caused by the dissolution of carbonate rocks and by their porosity increasing due to the action of the fermentation products (primarily acetic and butyric acids and CO 2 ). Short-chain fatty acids interacted with the carbonate matrix according to the following equations: As a result, soluble salts were produced and rock permeability in the contact zone increased, while calcium and magnesium salts of fatty acids with over seven atoms in their carbon chain are insoluble, so their accumulation resulted in decreased permeability. In the collectors consisting of aragonite and calcite, increased carbonate concentration in the medium may initiate the process of dolomitization, with magnesium incorporating into the structure of carbonates. The dolomite thus formed has lower volume than calcite or aragonite, and this process should in theory increase the rock permeability [43]. Inflow of the biomass of aerobic and anaerobic bacteria was accompanied by sealing of the highly permeable interlayers, which resulted in the changed phase permeability of the models (data not shown). The interaction of microbial metabolites with the rock, oil, and water resulted in more widespread water flooding and increased oil recovery. The results of model experiments were confirmed in the course of pilot trials at heavy oil reservoirs.

Pilot Trials of MEOR Biotechnology at Petroleum Reservoirs
As a substrate for microbial growth, heavy oil, being by its nature a product of biodegradation, is therefore a poor source of the easily utilized fractions most available to microorganisms. It may be expected, however, that in the presence of oxidants, the oilfield microbial community will oxidize oil with the production of a spectrum of oil-displacing metabolites. A field pilot experiment for enhancing oil recovery using the microbiological method was carried out at the Cheremukhovskoe oilfield, in the zone of the injection well 5600. The Cheremukhovskoe oilfield, containing heavy oil with a density of 0.932 kg/m 3 , has been exploited since 1997. Water-flooding with fresh river water mixed with production water separated from oil was initiated in 2001. The initial recoverable reserves were 802 × 10 3 t. The average oil saturation was 0.810 U. The average water cut of oil production is presently 51%. The same experiment was carried out in the zone of the injection well 1010 at the Vostochno-Anzirskoe oilfield with oil with a lower density of 0.856 kg/m 3 . Since 2003, the Vostochno-Anzirskoe oilfield has been exploited with focal water injection in order to preserve the seam pressure. The initial recoverable reserves were 582 × 10 3 t. The average oil saturation was 0.713 U.
The biotechnological treatment at the Cheremukhovskoe and Vostochno-Anzirskoe oilfields included the injection of oil-oxidizing bacteria adapted to oilfield conditions into the reservoir through injection wells. For the sustainable growth of oil-oxidizing bacteria, oxygen (as a H 2 O 2 solution) and (NH 4 ) 2 HPO 4 solution as a source of biogenic elements were injected in order to initiate the first stage of the oxidation of oil hydrocarbons. After the treatment of injection wells 5600 and 1010, the cultured microorganisms and the rates of microbial processes in the zone of production wells were studied and the effect of microorganisms on oil recovery was assessed. Unfortunately, we had no possibility of monitoring the changes in the microbial community during the pilot trials by molecular techniques.
The water in wells 5454 and 5463 of the Cheremukhovskoe oilfield showed the highest oil displacement, the highest numbers of aerobic organotrophic bacteria (including hydrocarbon oxidizers), and anaerobic fermentative bacteria were found to increase. The numbers of hydrogenotrophic and acetate-utilizing methanogens did not change in the course of the biotechnological treatment, although radiotracer analysis revealed a 1.5-4.5-fold increase in methanogenesis rates in a number of production wells (Table 3). In sulfate-containing formation water, sulfate reduction is unavoidable, unless special measures for its suppression are taken. Sulfate reduction rate in the zone of responding wells increased from 4-98 ng S 2− ·L −1 ·day −1 (during the baseline period) to 13-213 ng S 2− ·L −1 ·day −1 in the course of the trial. Heavy oil biodegradation was accompanied by increased concentrations of acetate (from 1.3-3.0 mg·L −1 to 20-30 mg·L −1 ) and bicarbonate (from 244-305 mg·L −1 to 476-671 mg·L −1 ) in the formation water of responding wells. In the course of the experiment, microbial metabolites spread from the near bottom zone of the injection well and affected oil displacement 5-7 months after the introduction of bacteria, the oxidant, and the nitrogen and phosphorus salt in the reservoir ( Figure 5). As a result of the pilot trial at the Cheremukhovskoe oilfield, 1141 t of oil was additionally recovered in the course of a year after the onset of the treatment. Although the abundance of hydrocarbon-oxidizing and fermentative bacteria, as well as bicarbonate concentration, increased the zone of production well 5464, this did not result in enhanced oil recovery, probably due to the high salinity of formation water in this area (over 186 g·L −1 ), which prevented microbial growth and the formation of oil-releasing metabolites.
At the Vostochno-Anzirskoe oilfield, the biotechnological treatment resulted in a very low increase in oil recovery. A total of 1250 t of oil was obtained in the course of two pilot trials. The efficiency of the biotechnological treatment was confirmed by the water content in the production of the responding wells decreasing by 18-35% ( Figure 5).
In the MEOR pilot trials at the heavy oil reservoirs, both at the high-temperature Dagang oilfield (China) [18][19][20] and at the low-temperature Cheremukhovskoe and Romashkinskoe oilfields [44,45] (Russia), oil recovery became more sustainable after the biotechnological treatment, which had a positive effect of total oil recovery. The Cheremukhovskoe oilfield is characterized by a low reservoir temperature (20.2-21.3 • C). The efficiency of carbonate dissolution by acetate at a low temperature was estimated in laboratory experiments [46]. It was shown that the dissolution of calcium carbonate in aqueous solutions of acetic acid occurs even at 10 • C. At the low-temperature (17-23 • C) bed 302 with carbonate collector and heavy oil at the Romashkinskoe oilfield, an MEOR pilot trial was performed based on the introduction of Clostridium tyrobutyricum and molasses into the oil stratum [44,45]. Monitoring the concentrations of fatty acids, calcium, and carbonates dissolved in the formation water of bed 302 provided direct evidence of carbonate dissolution by acidic products of microbial metabolism. According to the van't Hoff equation, the rate of this reaction would have certainly been higher at higher temperatures. The dissolution rate of carbonate rock is probably less important than the fact of formation of low fatty acids directly in the pore water, where their effect on the oil-bearing rocks is maximal.
Due to the multistage and interrelated nature of microbial processes occurring in oilfields, it is difficult to attribute enhanced oil recovery to a single mechanism. Field experiments at heavy oil reservoirs indicated that all groups of microbial metabolites produced during molasses fermentation [44,45] and oil oxidation [18][19][20] affected oil displacement. Volatile fatty acids dissolved the carbonate rocks, increasing their porosity and permeability. The gases produced in the course of transformation of the fermentation products in the microbial trophic chain (CO 2 , hydrogen, and methane) caused local restoration of the seam pressure, increased gas factor, oil swelling, and decreased oil viscosity, as well as increased collector permeability due to carbonate dissolution by CO 2 . The surfactants caused a decrease in the interfacial tension at the oil-water and oil-rock boundaries and to oil emulsification. Both microbial exometabolites and the biomass itself affected the motility of formation fluids and the direction of hydrodynamic flows in the oilfield, also increasing the waterflooding coverage.  Due to the multistage and interrelated nature of microbial processes occurring in oilfields, it is difficult to attribute enhanced oil recovery to a single mechanism. Field experiments at heavy oil reservoirs indicated that all groups of microbial metabolites produced during molasses fermentation [44,45] and oil oxidation [18][19][20] affected oil displacement. Volatile fatty acids dissolved the carbonate rocks, increasing their porosity and permeability. The gases produced in the course of transformation of the fermentation products in the microbial trophic chain (CO2, hydrogen, and methane) caused local restoration of the seam pressure, increased gas factor, oil swelling, and decreased oil viscosity, as well as increased collector permeability due to carbonate dissolution by CO2. The surfactants caused a decrease in the interfacial tension at the oil-water and oil-rock boundaries and to oil emulsification. Both microbial exometabolites and the biomass itself affected the motility of formation fluids and the direction of hydrodynamic flows in the oilfield, also increasing the waterflooding coverage.

Conclusions
Our results indicate that the studied low-temperature water-flooded heavy oil reservoirs (Russia) contain microbial populations capable of producing oil-displacing metabolites.
Aerobic bacteria Rhodococcus erythropolis HO-KS22 and Gordonia amicalis 6-1 isolated from oil reservoirs were capable of oxidizing heavy crude oil, and produce biosurfactants which significantly decrease the surface and interfacial tension of the media. G. amicalis 6-1 and R. erythropolis HO-KS22 have potential for applications in paraffin control, microbial enhanced oil recovery (MEOR), and the bioremediation of hydrocarbon-polluted environments.
The aerobic microbial community of heavy oil reservoirs was poor. Thus, its enrichment (bioaugmentation) with hydrocarbon-oxidizing bacteria adapted to oilfield conditions is a prerequisite for application of the MEOR technologies based on oil oxidation.
Fermentative bacteria from heavy oil reservoirs did not grow on oil; however, they produced volatile fatty acids, alcohols, and gases (H2 and CO2) when growing on sugar-containing substrates.
In core-flooding tests, liquid cultures of aerobic bacteria-producing biosurfactants were more effective in heavy oil displacement than fermentative bacteria-producing low fatty acids, alcohols, and gas, which resulted in a recovery of 43.0-53.5% and 13.4-45.5% of additional oil.

Conclusions
Our results indicate that the studied low-temperature water-flooded heavy oil reservoirs (Russia) contain microbial populations capable of producing oil-displacing metabolites.
Aerobic bacteria Rhodococcus erythropolis HO-KS22 and Gordonia amicalis 6-1 isolated from oil reservoirs were capable of oxidizing heavy crude oil, and produce biosurfactants which significantly decrease the surface and interfacial tension of the media. G. amicalis 6-1 and R. erythropolis HO-KS22 have potential for applications in paraffin control, microbial enhanced oil recovery (MEOR), and the bioremediation of hydrocarbon-polluted environments.
The aerobic microbial community of heavy oil reservoirs was poor. Thus, its enrichment (bioaugmentation) with hydrocarbon-oxidizing bacteria adapted to oilfield conditions is a prerequisite for application of the MEOR technologies based on oil oxidation.
Fermentative bacteria from heavy oil reservoirs did not grow on oil; however, they produced volatile fatty acids, alcohols, and gases (H 2 and CO 2 ) when growing on sugar-containing substrates.
In core-flooding tests, liquid cultures of aerobic bacteria-producing biosurfactants were more effective in heavy oil displacement than fermentative bacteria-producing low fatty acids, alcohols, and gas, which resulted in a recovery of 43.0-53.5% and 13.4-45.5% of additional oil.
The pilot trial at the Cheremukhovskoe oilfield confirmed the efficiency of the MEOR technology based on the introduction of hydrocarbon-oxidizing bacteria together with injection of the oxidant as an H 2 O 2 solution and of (NH 4 ) 2 HPO 4 as a nitrogen and phosphorous source for low-temperature heavy oil reservoirs.     Figure A1. Numbers of aerobic and anaerobic microorganisms in injection and production water samples from the Arkhangelskoe, Romashkinskoe and Cheremukhovskoe oilfields. Designations: Ferm-fermentative bacteria, AOB-aerobic organotrophic bacteria, SRB-sulfate-reducing bacteria, Met-methanogenic archaea. Injection well marked with *, the rest are production wells.