Aquifer thermal energy storage (ATES), leveraging on seasonal heat storage and extraction, promises a more sustainable alternative to reduce fossil fuels dependence for energy supply, especially in the heating and cooling sector (Elsland et al., 2017). For example, this sector serves residential and industrial consumers by providing solutions to energy demands for space or process heating and cooling. ATES allows for the storage of excess thermal energy during the warm summer to meet higher thermal energy demands in the cooler seasons. ATES operating temperatures vary from one installation to another (Fleuchaus et al., 2018), although they typically function under an upper temperature limit as regulatory threshold, e.g., 25°C, adopted by many Low-Temperature ATES (LT-ATES) operations (Drijver et al., 2012). Operating above this temperature threshold, e.g., up to 60°C and higher, as in the case of High Temperature ATES (HT-ATES), has since garnered interest as authorities sought to diversify their sources of energy supply. Sites where ATES systems are situated are usually urban or industry areas which are frequently loaded with contaminants released by former anthropogenic activities and thus limited as resources for drinking water. Contaminants commonly found in urban aquifers consist of petroleum hydrocarbons introduced via spillage or improper disposal (Rooney-Varga et al., 1999; Da Silva and Alvarez, 2004; Vogt et al., 2007), where their persistence remains a big hurdle to remediation up to today (NARA, 2018). In the subsurface, microbial biodegradation is crucial since contaminants can be mineralized to harmless products (e.g., carbon dioxide and water), whereas remediation techniques based on abiotic processes are usually limited in their ability to mineralize contaminants (EPA, 1999). Most hydrocarbons are biodegradable at oxic and anoxic conditions, in aquifers, however, anaerobic biodegradation is the more important attenuation process since conditions turn rapidly anoxic once oxygen is consumed. Anaerobic hydrocarbon degradation pathways are well-known and are usually carried out by microbial consortia (Mancini et al., 2002; McKelvie et al., 2005; Gieg et al., 2010; Winderl et al., 2010; Fuchs et al., 2011; Meckenstock et al., 2015; Lueders, 2017). Acetate is a key intermediate especially during syntrophic hydrocarbon biodegradation. It serves a dual role as an electron donor and a soluble carbon source which can be easily transported and utilized by microorganisms (Wu et al., 1997; Ho et al., 2002; Heuer et al., 2010; Head et al., 2014), both aerobes (Wu et al., 1997) and anaerobes (Sansone and Martens, 1982; Schink and Stams, 2006; Heuer et al., 2010). At anoxic conditions, sulfate is a main electron acceptor in marine but also freshwater habitats making dissimilatory sulfate reduction a commonplace in contaminant degradation (National Research Council, 2000; Miao et al., 2012).

Generally, biodegradation processes accelerate at higher temperature leading to the hypothesis that organic contaminants may be preferentially degraded at elevated temperatures during ATES operation, enabling the possibility to combine in situ bioremediation and ATES. However, unlike Arrhenius-like models where biochemical rates tend to increase with temperature (Mohr and Krawiec, 1980; Huang et al., 2011), the temperature-microbial growth paradigm is somewhat different, i.e., species are categorized according to their maxima, minima, and optimal growth temperatures (Dettmer, 2002). In general, growth ranges of psychrophiles, mesophiles, thermophiles and hyperthermophiles are: 0–20, 8–48, 40–70, and 65–90°C, respectively (Madigan et al., 2017). Optimal temperatures allow maximal growth rates; rates will be significantly lower and finally cease when temperatures reach sub-optimal and boundary conditions. This is paramount since ATES are likely to drive seasonal temperature fluctuations in the subsurface. As such, temperature fluctuations from ATES could influence the structure of the indigenous microbial community, its biodegradation potential, as well as the physicochemical characteristics of the contaminants such as adsorption and solubility (Koproch et al., 2019).

In ATES, it is postulated that while temperatures < 40°C may enhance biodegradation by increasing microbial activities of mesophiles, temperatures as high as 80°C or above may considerably affect the structure and function of subsurface microbial communities and adversely impact biodegradation. Nonetheless, the mineralization capacity of the indigenous microbial community is imperative in a comprehensive assessment of combined ATES and remediation potential. While the effects of temperature on the degradation of hydrocarbons and its key intermediate acetate have been widely reported, temperatures are not specifically geared toward ATES-related and microbial growth temperature targets. Moreover, not much has been reported on microbial community shifts during the transition from an oxic to an anoxic environment. Henceforth, in this study, the effects of temperature on (i) extent and rates of acetate mineralization by a hydrocarbon-adapted microbial community at ATES-relevant temperatures and (ii) temperature-induced shifts in the microbial community structure are addressed.

The investigated microbial community was originally enriched from a sulfidogenic, hydrocarbon-contaminated aquifer, and shown to be diverse and capable of degrading several aromatic hydrocarbons (Kleinsteuber et al., 2008; Herrmann et al., 2010; Taubert et al., 2012). Acetate was shown to be a metabolite upon sulfidogenic benzene mineralization, putatively metabolized by various phylogenetically different organisms (Rakoczy et al., 2011; Starke et al., 2016). We used this community as a model to determine the temperature range of anaerobic acetate mineralization and study the potential acetate-utilizing sulfate reducing degraders at different temperatures.

We postulate that mineralization occurred in two phases, a first aerobic phase and a second sulfidogenic phase. Aerobic mineralization is indicated by (i) the lack of sulfide (Supplementary Figure S2) or methane production from the reduction of sulfate and bicarbonate, the two principal electron acceptors present during the experimental set-up, and (ii) the bloom of typical aerobes affiliated to Alphaproteobacteria belonging to the genus Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium (ANPR) (at 12–38°C) or Bacilli (at 45 and 60°C) (Figure 4). We rule out that nitrate or ferric iron might have been used as electron acceptors in this study, as both compounds were not amended. The coarse sand contains iron, but reduction of ferric iron has never been observed in former studies using this material (Vogt et al., 2007; Herrmann et al., 2008, 2010; Kleinsteuber et al., 2008; Taubert et al., 2012). Our data show a succession from oxic to anoxic conditions, illustrating conditions which usually occur in shallow subsurface systems and which could be promoted by ATES by introducing oxic heated or cooled water, thereby allowing us to compare mineralization rates of acetate under oxic and sulfate-reducing conditions in an altering microbial community. Given that acetate-degraders are ubiquitous since acetate occurs as a key intermediate in both aerobic/anoxic systems (Thauer et al., 1989), studying temperature or electron acceptor related shifts are therefore essential. Other studies focusing on communities’ behavior and successions at permanent anoxic conditions under different heat-change regimes are in progress (data not shown) and will be published separately. Approximately 0.2 mM acetate (equivalent to ∼20% of the amended amount) was consumed in the first phase, implying that up to 0.4 mM oxygen was supposedly available (Table 1). Under oxic conditions, acetate oxidation to carbon dioxide with oxygen (ΔGr∘ = −908 kJ mol^–1) is far more exergonic, outcompeting other electron acceptor such as sulfate (ΔGr∘ = −63 kJ mol^–1) (Thauer et al., 1989; Jesußek et al., 2013). However, once oxygen was depleted, the dominant aerobic phylotypes were replaced by communities consisting mainly of anaerobic phylotypes such as Deltaproteobacteria, Ignavibacteria and Anaerolineae.

The ANPR taxon encompasses a wide range of species (Berrada and Fikri-Benbrahim, 2014), aerobes or facultative anaerobes (mainly utilizing nitrates). Most phylotypes are mesophilic (Berrada and Fikri-Benbrahim, 2014), such as the recently isolated R. terrae sp. nov, exhibiting growth between 0–42°C (Ruan et al., 2020); accordingly, the immediate acetate-dependent growth of these putative aerobic phylotypes at 12, 25, and 38°C indicates a broad mesophilic temperature range of members of this genus. Besides ANPR, Thiobacillus-like members were also thriving in the initial aerobic phase at 12°C, 25°C, and especially at 38°C, likely from the aerobic oxidation of iron sulfides which have been previously identified in the used coarse sand (Vogt et al., 2007). These form black precipitates visible to the eye, and may be supported by acetate which is used as a carbon source by some Thiobacilli species (Robertson and Kuenen, 2006).

The Bacilli which became enriched in phase 1 at 45 and 60°C, belonging to the genera Bacillus, Aeribacillus, Brevibacillus, and Planifilum are affiliated to typical thermophiles. Some bacilli are capable of growth at 50°C (B. subtilis) and 65°C (B. stearothermophilus, range: 30–75°C), as well as on anaerobic agars at 50°C (e.g., B. licheniformis, B. laterosporus); and many can form resistant endospores (Warth, 1978; Slepecky and Hemphill, 2006). Aeribacillus, represented by only two species recently, A. pallidus and A. composti, has been described as thermophilic, aerobic, endospore-forming bacteria which can utilize acetate (Yasawong et al., 2011; Filippidou et al., 2015; Finore et al., 2017). Brevibacillus isolates (e.g., B. borstelensis) were reported to grow in the range of 40–70°C (Panda et al., 2014; Khalil et al., 2018). Also, Planifilum species were described as thermophiles (Hatayama et al., 2005; Zhang et al., 2007). Generally, formation of endospores allows survival at lower temperatures in the absence of oxygen; hence, the immediate endospore reactivation at higher temperatures explains the quick bloom of aerobic thermophiles at 45 and 60°C (Keynan et al., 1964; Berg and Sandine, 1970). The low observed mineralization of acetate at 80°C in one replicate is possibly due to weak activity of thermophiles thriving at maximal but lethal incubation temperature; indeed, the data indicate that hyperthermophilic phylotypes were absent or below detection limit in the microbial community investigated. These findings demonstrate a resilience of acetate-mineralizing aerobes over a temperature range of 12–60°C stemming from an inoculum which was held at anoxic, sulfidic conditions for several years and presumed to be dominated by obligate anaerobic phylotypes at the onset of the experiment (Figures 3, 4). Furthermore, the established aerobic communities are distinct between 12–38°C and 45–60°C.

Once oxygen was depleted, the second mineralization phase commenced following a period of lag phase and was more apparent (54–94% mineralized). Sulfide production peaked and stabilized between 0.8 and 1 mM for temperatures of 12, 25, and 38°C. Sulfide produced in Phase 2 was comparable to the theoretical sulfide production stoichiometry with respect to acetate mineralization, verifying that these two processes are connected (Table 2) and that sulfate was utilized as the terminal electron acceptor. The complete oxidation of 1 mM of acetate by dissimilatory sulfate reduction would yield 1 mM of sulfide and 2 mM of carbon dioxide (Eqn. 2). We calculated the ratio of sulfide produced to acetate mineralized for the second phase, discounting mineralization arising from the first aerobic phase (Table 2). Ratios of sulfide produced to acetate mineralized were < 1 at 12, 38, and 45°C-3, likely due to precipitation of FeS as previously described (Vogt et al., 2007; Herrmann et al., 2008; Kleinsteuber et al., 2008). Still, this relationship is within theoretical expectations (Kleikemper et al., 2002; Ozuolmez et al., 2015), including the value of 1.01 at 25°C (this study).

In the 12–38°C setups, prolonged incubation saw putative sulfate reducing taxa becoming more evident including the genera Desulfotomaculum and Desulfoprunum (Figure 5), with the former seemingly more prevalent at 12 and 25°C, and not being represented at temperatures 38°C and above. Desulfotomaculum form endospores, and some species have been described for acetate oxidation (Widdel and Bak, 2006); psychrophilic Desulfotomaculum strains have been also previously reported (Klemps et al., 1985; Bahr et al., 2005; Aullo et al., 2013). Notably, mesophilic and moderately thermophilic Desulfotomaculum species were seemingly absent or below detection limit in our study although often described (Klemps et al., 1985; Scholten and Stams, 2000; Widdel and Bak, 2006; Aullo et al., 2013; Liu et al., 2018). For the genus Desulfoprunum, only a single species has been described so far, D. benzolyticum, a mesophilic freshwater isolate capable of acetate utilization (Junghare and Schink, 2015). While their study described an optimal growth of D. benzolyticum of 30°C with only little growth observed at 20 or 37°C, our results suggest that the Desulfoprunum-related phylotypes grown at 12°C are adapted to lower temperature ranges.

At 38°C, typical sulfate reducers also significantly increased in abundance in the sulfidic phase. For example, at day 36, these phylotypes comprised of Desulfurispora (11.6%) and Desulfoprunum (2.9%). Meanwhile, at day 122, Desulfatiglans (∼3.4%), Desulfoprunum (∼1.1%), Desulforivibrio (2.2%), and Desulfurispora (1.3%) were amongst the abundant sulfate reducers. It is worth to note that at day 36 acetate was still available, while these was completely utilized prior to sampling day 122, hence acetate unavailability could be a reason for the lower abundance, which could be attributed to other factors such as cell lysis or spore formation. The fact that Desulfurispora was highly abundant at day 36 hints that this organism metabolized acetate at 38°C. Notably, to our knowledge only one such species was described yet, growing at 40-67°C and unable to utilize acetate (Kaksonen et al., 2007). In contrast, the Desulfurispora phylotype seen in our study is abundant under mesophilic conditions, and as we did not detect any sulfate-reducing activity at 60°C. Lastly, Desulfatiglans, also detected in communities at 12 and 25°C have been reported for acetate oxidation under sulfate reducing conditions elsewhere (Jochum et al., 2018).

Besides the sulfate reducers, several other phylotypes also saw an uptick in abundance at 38°C. At day 36, phylotypes included an uncultured member of the Ignavibacteria (5.2%), Sulfurovum (6.5%), and uncultured member of the Anaerolineae (3.7%). Similarly, increased abundances were also reported at day 122, amongst them, an uncultured Anaerolinea member (5.1%), Sulfurovum (9.1%), three uncultured members of the Ignavibacteria (2.3–3.9%), which hinted at a possible role in acetate oxidation under mesophilic conditions. Ignavibacteria, with several members abundant in both 25 and 38°C setups, are not known for sulfate reduction, but were detected in various hydrocarbon-degrading microbial communities under various electron acceptor conditions including sulfate (Kummel et al., 2015; Westphal et al., 2017; Keller et al., 2018). However, their function in acetate metabolization at sulfate reducing conditions in our study is unclear. Sulfurovum, belonging to the Campylobacterales (Epsilonproteobacteria), was detected in several replicates at 38°C. Interestingly, Campylobacterales, have been reported as potential acetate consumers at sulfate-reducing conditions detected by protein-stable isotope probing studies using uniformly ¹³C-labeled acetate as substrate (Starke et al., 2016). In addition, Epsilonproteobacteria or specifically Sulfurovum-like phylotypes were enriched in hydrocarbon-degrading consortia (Kleinsteuber et al., 2008; Muller et al., 2009; Herrmann et al., 2010) which suggest that the Sulfurovum enriched in our 38°C setup could have consumed the acetate. While Sulfurovum was detected over a range of temperatures (22–66°C) in water-flooded oil reservoirs, they were not particularly enriched (Tian et al., 2017). However, in our study, Sulfurovum was particularly enriched at 38°C (up to 9%) and 45°C (up to ∼11%) while abundances were lower at 12°C (up to ∼4%) and 25°C (∼2%). In contrast, Sulfurovum abundance was negligible in almost all replicates at 60°C (<1%) which did not mineralize acetate at sulfate-reducing conditions. Hence, this Sulfurovum candidate might be active within the higher end of the mesophilic range (38–45°C).

One key result of our study is the observation that temperatures above 38°C (45°C, 60°C) strongly inhibited acetate mineralization at anoxic conditions, as only one replicate at 45°C could do so. Notably, none of the typical sulfate reducers that became enriched at 12–38°C seem to be abundant in this replicate, which raises the question of which phylotypes is actually responsible for sulfate reduction. The Spirochaetes, which increased to up to 32%, are often observed at anoxic hydrocarbon-contaminated sites; they are suggested to recycle necromass by fermentation, and have not been described to be directly involved in hydrocarbon degradation at anoxic conditions (Dong et al., 2018). Our data indicate that Spirochaetes are especially adapted to temperatures above the typical mesophilic range, which seem to be true as well for members of the Anaerolineaceae whose abundance increased to up to 27%. Anaerolineae were shown to comprise thermophilic members and were reported to reach high abundances during anaerobic digestion under methanogenic or sulfate reducing conditions, but due to a general lack of isolates and genomes their physiological capabilities are less well understood (Xia et al., 2016; Roy et al., 2018). Hence, it might be that these phylotypes become enriched by scavenging on dead biomass similar to the Spirochaetes. Although not yet described for dissimilatory sulfate reduction, we cannot rule out the possibility that some of them are able to do so. For instance, in a down-flow fluidized bed reactor operated at 25°C and mainly fed with acetate and sulfate, a community developed which comprised of 50% of Anaerolineaea and to a lesser extent of typical acetate-oxidizing sulfate reducers (Montoya et al., 2013). Recently, Anaerolineae phylotypes were shown to assimilate ¹³C-labeled acetate upon syntrophic acetate oxidation in a chemostat under methanogenic conditions (Yi et al., 2020); thus, it is also possible that acetate was syntrophically oxidized in our culture by Anaerolineae phylotypes, leaving the question open which organisms were actually responsible for sulfate reduction. Also 45°C lies within the overlap of the maxima end of the mesophilic range and the minima end of the thermophilic range, in the so called mesophilic-thermophilic boundary. Not much is known of bacteria tolerant in this boundary. Notably, in other studies, sulfide production was observed at higher temperatures compared to our study (Elsgaard et al., 1994; Bonte et al., 2013; Jesußek et al., 2013; Westphal et al., 2017).

Three distinct clusters in the Bray–Curtis confirmed the described shift in the microbial communities at 12, 25, and 38°C in phase 1 (cluster 1) to phase 2 (cluster 2), which are both markedly different from the inoculum (cluster 3). Community shifts are usually observed in environments shifting from oxic to anoxic conditions (Broman et al., 2017; Bush et al., 2017). Additionally, samples from replicates at 45 and 60°C in phase 1 contained sequences belonging to cluster 3 and were thus similar to the inoculum; the main difference between them is the emergence of putative aerobic species such as the Bacilli (Figure 5). The sulfidogenic replicate 45°C-3 also grouped in Cluster 3, although having a lower abundance of Bacilli-like phylotypes which were replaced by higher Anaerolineae and Spirochaetia abundances. The composition of cluster 3 of the Bray–Curtis analysis indicates that only a few members of the inoculum were actually active at temperatures ≥ 45°C. Notably, the results of our study are in contrast to those reported by Westphal et al. (2017) who observed acetate mineralization under both oxic and sulfate-reducing conditions in aquifer sediment columns experiments at 10, 25, 40, and 70°C.

The temperature growth/biochemical reaction correlation (Davey, 1993) would imply that mineralization rates increase with temperature (Meier et al., 2005; EPA, 2017). Similarly, increased rates in organic matter mineralization between 25 and 50°C were also reported (Hossain et al., 2017). However, in our study, this was not necessarily the case. In the aerobic phase, two temperature maxima were observed, first at 25°C (75 ± 34.4 μM d^–1) and second at 60°C (71.4 ± 24.3 μM d^–1). Mineralization rates increased from 12 to 25°C, in agreement with increased mineralization from 6 to 28°C observed before (Sansone and Martens, 1982). However, rates at 38, 45, and 60°C did not fit with this relationship, given that rates at 25 and 38°C (74.1 ± 20.2 μM d^–1) were similar; and a rate minimum was observed at 45°C, being about six times lower than the values at 25 and 38°C. In the anaerobic phase, considering only replicates at 12–25–38°C, acetate mineralization rates (9.8 ± 4.3, 18.8 ± 1.5, 22.9 ± 5.3 μM d^–1) matched the Arrhenius relationship, increasing with temperature (Table 1). The increased mineralization activity at 25 and 38°C compared to 12°C corroborates with prior studies on the enhanced biodegradation with increased temperatures (Jesußek et al., 2012; Griebler et al., 2016; Westphal et al., 2017). In contrast, the single sulfidogenic replicate 45°C-3, does not fit with the linearity observation, indicating non-optimal conditions for the acetate consumers at 45°C as microbial growth rates outside optimal conditions decrease with temperature until inhibition (Mohr and Krawiec, 1980; Okabe and Characklis, 1992; Huang et al., 2011). Inhibition of degradation processes at temperatures above 45°C has also been shown in other studies, e.g., for dechlorination (Deeb and Alvarez-Cohen, 1999; Friis et al., 2007b; Fletcher et al., 2011). Clearly, more studies are needed using different types of sediments to derive general assumptions on the physiological capacities of indigenous microbial communities at different, especially high (T ≥ 45°C) temperatures; inhibition temperatures might be experiment- or site-specific.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ebi.ac.uk/ena, PRJEB40338.

MB planned and performed the experiments, analyzed the data, and wrote the manuscript. CV planned the experiments and edited the manuscript. HR edited the manuscript. All the authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.