Changes in environmental and engineered conditions alter the plasma membrane lipidome of fractured shale bacteria

Some microorganisms that are inadvertently introduced into the deep biosphere during hydraulic fracturing to extract natural gas and oil from shale formations survive numerous stressors and persist for long periods of time (1 – 4). The persisting microbial communities in engineered shale reservoirs are mostly comprised of anaerobic halophilic and halotolerant taxa (5, 6). Halanaerobium has been found to be ubiquitous and dominate many geologically distinct formations (3, 6). The composition of the shale microbiome is, however, variable and dependent on several spatiotemporal factors including intrinsic geochemistry, engineering parameters, and production duration. These bacteria become key drivers of subsurface biogeochemistry through secretion of metabolic by-products such as the well foulant, hydrogen sulfide (7), and formation of biofilms on fracture surfaces (8). Hydrogen sulfide is produced through sulfate reduction by sulfate-reducing bacteria including Halanaerobium (7). This gas alters the composition of the reservoir and also facilitates corrosion of engineering infrastructure (9). On the one hand, cellular aggregation on shale matrices further reduces the permeability of the rock and clogs its microfractures, which are gas flow channels. These decrease the efficiency of hydrocarbon extraction. On the other hand, some microorganisms in fractured shale wells enhance secondary natural gas recovery through methanogenesis. This usually involves metabolic synergy between bacterial fermenters and archaeal methanogens. In one well-described instance, glycine betaine (GB) fermentation by Halanaerobium yields trimethylamine, which is consumed by Methanohalophilus to produce methane (3, 10).

A comprehensive understanding of the complex microbiology of fractured shale reservoirs is very important to the global energy sector. In the US, shale accounted for over 79% of dry natural gas production in 2020 and is projected to continue to meet most of the dry natural gas needs of the country through 2050 (11, 12). Unfortunately, knowledge of in situ growth kinetics, metabolisms, and physiological adaptations of shale bacteria is still limited despite recent breakthroughs (1, 3, 10), which constrains effective biocontrol and efforts to harness their beneficial potential. The plasma membrane (PM) protects the cell, mediates many of its critical functions (e.g., materials exchange, energy metabolism, and surface attachment) (13), and responds to intracellular cues and ecological perturbations through physicochemical modifications (14, 15). As such, it holds valuable insights into the behavior and physiological adaptations of microorganisms in disturbed environmental systems such as engineered shale, which is largely untapped.

Fractured shale is an extreme environment for microbial growth and, by virtue of drilling/hydraulic fracturing, a highly disturbed ecosystem. It is characterized by elevated temperatures, nutrient limitation, anoxia, elevated pressures, and brine-level salinities (1, 5, 16). For instance, the salinity of flowback and produced water, which is co-collected with natural gas, could increase more than fourfold up to 6,000 ppm in just about 30 days post-fracturing (17, 18). This is due to the mixing of produced fluids with formation of brine as well as dissolution of minerals and salts on fracture surfaces (19 – 21). In addition, hydraulic retention time (HRT), which is the length of time that injected fluids reside belowground before recovery at the surface, broadly varies and influences carbon flux and availability. These injected fluids inadvertently carry fresh-water bacteria that must adapt to the deep shale environment by making regular fitting morphological and physiological adjustments to cope with these dynamic stresses. This involves remodeling the membrane lipidome to maintain or acquire optimal physicochemical states such as those needed for rewiring intracellular metabolism or forming biofilms. Intact polar lipid (IPL) chemistry is a key determinant of plasma membrane biophysics and functions and is, thus, especially adjusted accordingly (14, 15). Membrane IPLs differ based on headgroup net charge; hence, there are neutral, anionic (negatively charged), and zwitterionic lipids, which have a balance of positive and negative charges.

Bacterial membranes are most commonly comprised of phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and cardiolipin (CL). Other less common lipids are phosphatidylcholine (PC), phosphatidylserine, phosphatidylinositol, ornithine lipids, diacylglycerol (DG), triacylglycerol (TG), glycolipids, and sphingolipids. The IPL composition largely depends on taxa and cell envelope structure. PE comprises about 75% of all phospholipids (PLs) in Escherichia coli and other Gram-negative bacteria (22). Variations in IPL chemical and physical properties, due to differences in headgroup and acyl chain compositions, greatly influence the biophysical attributes and functions of the bacterial plasma membrane.

Evidence suggests that some bacteria increase the amount of negatively charged intact polar lipids in their plasma membrane in response to increasing salinities (23, 24). This is due to their bilayer-stabilizing properties and implicated role in osmoprotection (24 – 27). Bacterial plasma membrane lipidomic changes in response to changing conditions are, however, very diverse and still being unraveled across taxa and environments (14). To the best of our knowledge, there is no existing study on how changes in salinity and hydraulic retention time affect membrane IPL chemistry in shale reservoir bacteria, and the implications of these changes on their activities and system biogeochemistry are still elusive. In addition, there is little information on the cellular membrane lipidome adjustments underlying biofilm formation on fracture surfaces. Biofilms are a major concern for engineering subsurface hydrocarbon reservoirs—they exacerbate biocontrol failure, clog fractures, and diminish the operational performance of engineering infrastructure (7, 8, 28). Therefore, understanding how changing subsurface conditions influence biofilm biology, which is mediated by the plasma membrane, is crucial for more efficient natural resource extraction.

Here, by coupling laboratory experiments with field investigations, we unravel how changes in salinity (environmental) and hydraulic retention time (engineered) influence the plasma membrane IPL chemistry of model bacterium, Halanaerobium congolense WG10, and mixed microbial consortia enriched from shale-produced fluids. We also elucidate adjustments in membrane IPL chemistry during biofilm growth relative to planktonic cells. We link these nanoscale events to ecosystem-scale reservoir biogeochemistries and discuss implications for the efficiency of energy recovery.

Anionic PGs and CLs and zwitterionic PEs (Fig. 3 and 4) predominated lipids that significantly increased in H. congolense WG10 planktonic and biofilm cultures under high salinity (20% NaCl) relative to the optimum (13% NaCl). The H. congolense genome encodes phosphatidylglycerophosphatase A (pgpA; accession ID: SAMN04515653_10658) and cardiolipin synthase (cls, accession ID: SAMN04515653_11647), which catalyze PG and CL synthesis, respectively. Hypersalinity exerts osmotic stress and ionic toxicity on biological cells (32, 33), causing loss of intracellular water via exosmosis and direct damage to cellular structures including the plasma membrane. Monovalent cations, such as Na⁺, induce hydrophobic interactions among membrane lipid headgroups, which result in biophysical transition from lamellar (normal bilayer configuration) to deleterious non-lamellar (hexagonal-II or cubic) phases (34, 35). An increase in levels of anionic PGs and CLs, which we observed in H. congolense WG10 growing under high salinity, is necessary to stabilize the membrane in the normal bilayer (lamellar) orientation. This is because unlike their zwitterionic counterpart, anionic headgroups are not very susceptible to such cation-induced interactions (27, 34, 36, 37). Moreover, anionic PLs are thought to play essential roles in osmosensing, accumulation of ions at the cell–environment interface by electrostatic forces, and channel activation for osmolyte acquisition (23, 24, 38). They are, therefore, key to bacterial survival in high-salinity shale reservoirs, through stabilizing the membrane’s normal bilayer phase and enabling “salt-in” osmoadaptive functions. Prior studies support the role of anionic phospholipids in bacterial hypersalinity tolerance (23, 39, 40). We are not sure why there was a net increase in membrane PE composition in H. congolense WG10 planktonic cells under high salinity as prior studies have established an opposite trend (24, 41). However, being the major phospholipid subclass in most bacterial membranes (42), they might be key to maintaining the default molecular order of the matrix and modulating its adaptive biophysical modifications and functions. It is important to take into consideration the difficulty of directly comparing the magnitude of lipidomic changes induced by each growth parameter.

Phosphatidylcholines were dramatically suppressed in H. congolense WG10 biofilm cells growing under high salinity relative to the optimum (Fig. 4). This suggests subsequent PC degradation to choline residues. We believe that like several bacteria (43 – 45), H. congolense WG10 expresses functional phospholipases that hydrolyze redundant PCs to choline (see Supplementary Appendix Fig. S1 and text). Phospholipase is, in fact, encoded in the genome of a H. congolense strain (accession ID: SAMN04515653_13217). Therefore, it is possible that this bacterium accumulates choline as an osmoprotectant (10) in a controlled laboratory reactor with limited alternatives. Proteomics and metabolomics are needed to verify this. However, within the fractured shale ecosystem, H. congolense might utilize other osmoprotecting compounds such as sugars and amino acids (10), therefore dispose choline (if accumulated), which it lacks the ability to catabolize. Other shale-dominating taxa including Halomonadaceae and Marinobacter can scavenge and oxidize choline to glycine betaine (3), which serves as a nutrient/energy source, as well as an osmoprotectant for several species including Halanaerobium. Viral infection and lysis of GB producers most likely account for the supply of GB to non-producers (3). In a prior study, GB levels in shale-produced fluids were positively correlated with well salinity (10). We argue that co-metabolic PC hydrolysis in osmotically stressed Halanaerobium and choline oxidation in other taxa are the missing intermediate pieces of this puzzle. Our finding adds to growing evidence that H. congolense utilizes the “compatible solutes” (choline and possibly GB) strategy for high salinity stress adaptation, in addition to the more established salt-in mechanism (10, 25). This, however, needs to be further substantiated by proteomics and metabolomics analyses.

While there are strong indications that in engineered shale, Halanaerobium accumulates GB as an osmoprotectant, some strains might ferment it to trimethylamine (3), which is readily consumed by methanogens especially Methanohalophilius to produce methane (3, 10). The enzyme responsible for fermentative GB reduction via the Stickland reaction, GB reductase (grdHI), is encoded in some shale Halanaerobium genome (3). Also see the accession ID: SAMN04515653_10521. Biogenic methanogenesis enhances secondary natural gas recovery from reservoirs (46). By implication, H. congolense adaptation to high salinity through PC suppression and subsequent recycling could be intricately linked to increased biogenic methane production. The fact that this phenomenon was only observed in the biofilm cells suggests that execution of the choline-based “compatible solutes” strategy has unfavorable energetics and is largely incumbent on PC overabundance and constraints to extracellular ion acquisition, which might be imposed by the biofilm architecture. In fact, H. congolense planktonic cultures showed relatively marginal PC abundance compared to their biofilm counterpart (Fig. 1).

As fractured shale wells mature, reservoir salinity increases (18), and so does fluid retention time. The rate of flowback and produced water recovery fluctuates—it naturally declines over time due to constant production and is deliberately controlled in line with seasonal demand (47). We simulated this phenomenon in the lab using the hydraulic retention time concept in a continuous culture setup. Moreover, studying cells at a steady state is the only way to obtain meaningful and reproducible results (48). The longer the medium/fluids spend in the shale reservoir or at the lab scale, in the chemostat bioreactors, the more resident bacterial cells are metabolically stressed due to nutrient (carbon) depletion and accumulation of toxic metabolic by-products. Anionic phosphatidylglycerols predominated lipid species that significantly increased in both H. congolense WG10 cells and mixed microbial consortia, grown and enriched, respectively, under the highest HRT (48 and 72 h, respectively) (Fig. 7) relative to the lowest (19.2 h). The fact that no zwitterionic lipid was significantly increased under the highest HRTs in both the single and mixed cultures suggests that the upsurge in PGs might have been geared in large measure toward increasing the net negative charge of the membrane and cell surface. This perhaps helps repel toxic anionic metabolites such as sulfide and organic acids (49) from the cell envelope. In addition, PGs facilitate protein folding and binding (50) in bacteria, processes that are key to cellular survival under metabolic stress.

Biofilms in engineered shale reduce rock permeability and clog fracture networks that constitute hydrocarbon flow channels (8, 51). They also confer increased ecological persistence and biocide resistance on the inhabitant microbiome. The plasma membrane directly mediates cellular surface attachment and biofilm development, and this depends on the bilayer’s physicochemical attributes, largely determined by its intact polar lipid headgroup chemistry (52). In this study, we found that zwitterionic phosphatidylcholines and phosphatidylethanolamines overwhelmingly predominated lipids elevated in H. congolense WG10 and mixed microbial consortia enriched from shale-produced fluids during biofilm growth relative to planktonic (Fig. 5). Moreover, in uncultivated microbial consortia obtained from field-filtered shale-produced fluids sampled across four time points, the relative abundance of zwitterionic PCs and PEs spiked in December 2020 (Fig. 8) following a steep drop in well flow rate 5 months prior, corresponding to an increase in HRT. This is offset by a decrease in levels of neutral lipids. Anionic lipids are maintained at fairly high proportions due to their ever-needed role in high salt tolerance. Low shale well flow rates would typically facilitate cellular attachment and biofilm formation in the reservoir.

Hence, our hypothesis is that PCs and PEs modulate the community (biofilm) physiology of the fractured shale microbiome. Their roles might include decreasing energy-intensive membrane fluidity (53) [an unnecessary burden to sessile biofilm cells with attenuated metabolic activity (54)], reducing the net negative charge of the cell envelope to prevent cell–cell (or cell–surface) electrostatic repulsion, and forming specialized lipid rafts to present protein receptors for quorum sensing and other intercellular exchanges. A prior study reported that levels of PCs and PEs were higher in the early and mature phases of biofilm formation in Candida albicans compared to the planktonic cells (52). This opens up a new frontier of discussion about the use of choline-containing additives for shale well development, such as choline chloride currently preferred as clay stabilizers to the more toxic tetramethylammonium chloride (55). Free choline residues could be acquired by shale microbes such as Marinobacter (3) and assimilated into membrane PCs, which we believe facilitate deleterious biofilm formation. Geophysical and technical constraints make it nearly impossible to sample and monitor intact biofilms in fractured shale reservoirs. Our machine learning model based on 15 discriminant lipids (Fig. 6) is a useful tool for estimating biofilm growth in non-sterile subsurface hydrocarbon systems using lipid signatures in produced fluids. A sudden increase in PE abundance could especially be indicative of aggressive biofilm development.

Besides salinity, hydraulic retention time, and biofilm formation, whose effects were evaluated in the present study, there are other strong drivers of plasma membrane lipid chemistry adjustments in fractured shale microbes, including temperature and pressure. The downhole temperature and pressure of fractured shale reservoirs vary depending on formation characteristics and operating parameters and fluctuate over time during and after well development (56, 57). This drives adaptive morphological and physiological remodeling in the inhabitant microbiome. High temperature and low pressure exert a similar effect on the biophysical properties of the membrane, which is a disruption of acyl chain packing order, which results in an increase in fluidity (58, 59). In response, bacteria primarily increase the proportion of saturated fatty acids in the bilayer. Conversely, low temperature and high pressure gelatinize the membrane, necessitating an increase in levels of unsaturated and branched chain fatty acids to restore optimal fluidity. This phenomenon is commonly known as homeoviscous adaptation (60). In addition, temperature and pressure changes affect headgroup chemistry. In general, high temperature and pressure usually induce an overall increase in relative abundance of polar lipids in the membrane (59, 61). Also, zwitterionic lipids, which possess balanced positive and negative charges, pack more densely than their anionic counterparts, hence are elevated in some bacteria under high temperature and low pressure (59). It is, thus, important to keep in mind that it is an interplay of these factors—including temperature, pressure, salinity, and hydraulic retention time—that will determine the bulk plasma membrane biophysical properties of the fractured shale microbiome.

Microbial plasma membrane intact polar lipid chemistry adjustments in response to environmental and engineered changes in fractured shale reservoirs have significant implications for biocontrol efficacy. Non-oxidizing glutaraldehyde (GTD) and quaternary ammonium compounds (QACs) are biocides very commonly used to control microbial activities in engineered wells (28). QACs either directly target the plasma membrane or use it as a conduit to access the intracellular space. On the other hand, GTD interacts with unprotonated amines on the cell surface, which disrupts membrane and cell wall integrity (62, 63). QACs, which possess a cationic headgroup, electrostatically bind the cell envelope, inserting their hydrocarbon tails into the matrix in ways that cause lethal structural disruptions (64).

Our results indicate that even though anionic phospholipids are elevated under increasing salinities (Fig. 3 and 4), levels of zwitterionic species increase dramatically during controlled biofilm formation (Fig. 5) and peak in shale-produced fluids immediately following a steep drop in flow rates (Fig. 8). Zwitterionic lipid headgroups have a perfect balance of positive and negative charges, therefore reducing the zeta potential of the membrane. This might limit the efficacy of cationic biocides such as QACs—used in nearly every shale formation in the US (28)—whose action relies on electrostatic binding to the cell surface. However, biocidal efficacy non-linearly depends on various other factors such as the salinity of the formation. At high salinity, the reactive intermediate of GA is stabilized, potentially making it less reactive and effective (65). However, high salinity enhances the hydrophobicity of the membrane bilayer (66), which might facilitate the intercalation of the carbon tails of QACs into the matrix, therefore increasing biocidal efficacy. Further experiments under controlled conditions are needed to investigate how induced plasma membrane lipidomic changes in fractured shale taxa affect biocidal efficacy. Overall, bacterial communities may better thrive in shale reservoirs under lower (within the optimal range) salinity and flow rates due to increased plasma membrane zwitterionic phospholipid composition, which facilitates biofilm formation and might increase resistance to cationic biocides.

This work was funded by the US Department of Energy, Office of Science, Office of Biological and Environmental Research (BER), and Established Program to Stimulate Competitive Research (EPSCoR) (award number DESC0019444).

We also acknowledge Northeast Natural Energy and West Virginia University for access to the MSEEL-2 wells and use of their facilities throughout this research project. Finally, we thank the reviewers who provided very insightful comments that improved the quality and readability of the manuscript.

Conceptualization was performed by P.J.M.; data curation was performed by K.B. and C.J.U.; formal analysis was performed by C.J.U.; funding acquisition was performed by P.J.M.; investigation was performed by F.C., J.A., K.B., S.A.W., and J.E.; methodology was performed by P.J.M., F.C., and J.A.; project administration was performed by P.J.M.; visualization was performed by C.J.U. and P.J.M.; writing was performed by C.J.U.; and writing (review and editing) was performed by C.J.U., F.C., J.A., K.B., S.A.W., J.E., and P.J.M.