**4. Discussion**

#### *4.1. Light and O2 Dynamics Shape the Environmental Conditions for C. aggregans*

The results of microsensor analyses revealed a strong correlation between solar irradiance and the O2 concentration and penetration depth in the cyanobacterial microbial mats from Nakabusa Hot Springs (Figure 2). Since no increase in O2 concentration was observed in the bottom layer lower than 1mm in depth, oxygenic cyanobacteria was supposed to be absent in the undermat as supported by the Martinez et al. [13]. The diel changes between anoxia and hyperoxic conditions driven by the oxygenic activity of cyanobacteria lead to drastic changes in the conditions for their microbial metabolism. Consequently, matinhabiting microbes may be under optimal conditions during only part of the diel cycle, and may need to endure unfavorable conditions at other times. Under such conditions, a versatile metabolism is thought to be advantageous in terms of ensuring continuous energy production under dynamic environmental conditions. In the following sections it is discussed how *Chloroflexus aggregans* use their metabolic flexibility to thrive in the highly variable conditions in the microbial mat over a diel cycle.

#### *4.2. Low Light and Low O2 Dominated the Morning Hours (07:00)*

After sunrise, although no direct sunlight hit the mats, the irradiance from diffuse light increased and stimulated cyanobacterial oxygenic photosynthesis, as indicated by the increasing O2 concentrations at the mat surface as well as the significant increase in the relative transcription levels of genes for protection against reactive oxygen species in *C. aggregans*. However, deeper mat layers were still anoxic, which in combination with low-light conditions seems to provide suitable conditions for anoxygenic photosynthesis by FAPs such as *Chloroflexus spp.* [67]. The increasing transcription of genes encoding housekeeping enzymes suggested increasingly active metabolism (anabolism) in *C. aggregans* (Figure S2). In the morning, the transcription of the phototrophy-affiliated ACIII Cp gene increased slowly and the transcription levels of genes for ACIII Cr, *aurA* and cytochrome *c* oxidase decreased, which in combination indicate active phototrophy. Photoheterotrophy of *C. aggregans* is suggested by increases in the transcription of TCA-related genes, probably using fermentation products in the mats that accumulated during the nighttime, as reported in similar hot spring systems in Yellowstone National Park [68–72]. At the same time, photoautotrophy (and thus photomixotrophy) is indicated by the increased transcription of key 3-OHP genes, probably supported by anaplerotic carbon fixation as indicated by the increased transcription of phosphoenolpyruvate carboxylase genes (Figure 6).

Unexpectedly, neither molecular hydrogen nor sulfide seems to function as an electron donor for autotrophic growth at this time, as neither *hyd* nor *sqr* genes were highly transcribed. In contrast, a significant increase in the relative transcription levels of *cox* gene transcription was seen starting at 07:00, indicating that carbon monoxide is a potential electron source for photoautotrophic metabolism in the early morning, as hypothesized for the CO utilization of *Roseiflexus* spp. and *C. aurantiacus* based on the genomic analysis [11,37].

#### *4.3. High-Light and Super-Oxygenated Midday Hours (11:00)*

Oxygen started to accumulate in the upper mat layers from around 09:00 and with time also accumulated in the deeper mat layers. Hyperoxic O2 levels (>800 μmol O2 <sup>L</sup>−1) were observed in the uppermost mat layers and O2 penetration reached a >2 mm depth under high irradiance (approx. 1000–1500 μmol photons m<sup>−</sup><sup>2</sup> s<sup>−</sup>1; 400–700 nm) between 10:00 and 14:00. The relative transcription levels of genes encoding DNA gyrase, DNA/RNA polymerase and ATP synthase peaked at the same time (Figure 5 and Figure S2), indicating active growth and energy production of *C. aggregans* during midday high-light and oxic conditions.

In the laboratory, *C. aggregans* and other FAPs have long been known to grow chemoheterotrophically via oxic respiration under aerobic dark conditions [2]. Additionally, aerobic growth in the light has been shown for *C. aurantiacus* only very recently [73]. In the present study, despite the presence of O2 in the upper 2 mm of the mat during this period of the day, there was no indication of chemoheterotrophic growth or aerobic respiration. These results sugges<sup>t</sup> the absence of both active glycolysis (as indicated by low relative transcription of the 'bacterial' type of the 6-phosphofructokinase gene) and aerobic respiration (as inferred from the transcription data of cytochrome *c* oxidase genes and the ACIII Cr genes). In contrast, active phototrophy in the presence of O2 is suggested by the high relative transcription levels of genes involved in electron transport, including ACIII Cp, *aurB*, and the TCA cycle. Structural analyses indicated that AurA and AurB in *C. aurantiacus* are active during phototrophic growth and chemotrophic growth, respectively [74]. In contrast, the transcription of *aurB* under phototrophic conditions in this study is consistent with the data obtained in a proteomic study of *C. aurantiacus* in which AurA was more abundant during chemoheterotrophic growth, while AurB was observed under photoheterotrophic conditions [45]. While the phototrophic growth under aerobic conditions correlates with earlier laboratory studies demonstrating that O2 did not inhibit energy transfer between the chlorosomes and the reaction center in *C. aurantiacus* [75].

However, bacteriochlorophyll biosynthesis in *C. aurantiacus* was long believed to be inhibited by O2, and enzymes involved in the biosynthesis were only detected or significantly increased in cultures grown under anaerobic phototrophic conditions [45]. Accordingly, metatranscriptomic studies of alkaline hot spring microbial mats in Yellowstone National Park showed that the transcripts of most of the genes involved in the biosynthesis of

BChls in different FAPs, including *Chloroflexus* spp., were the most abundant at night [23]. Inhibitory effects on the biosynthesis of the photosynthetic apparatus under aerobic conditions in *C. aurantiacus* were also shown in the laboratory [76]. In contrast, transcription of phototrophy-related genes under aerobic light conditions were shown for *C. aurantiacus* only very recently [73].

In accordance with the expectation that oxygen represses the biosynthesis of phototrophic apparatus and pigments, nocturnal transcription patterns of the photosynthetic apparatus-related *puf* and *csm* genes were also observed in the present study and might be negatively correlated with the increasing O2 concentrations in the environment. In contrast, many of the *bch* genes were observed to be transcribed during the day, with the majority showing the highest relative transcription levels at 11:00. This correlates with a recent study of *C. aurantiacus* showing biosynthesis of BChl *a* and *c* under anaerobic as well as aerobic conditions in the light, while being suppressed in the presence of O2 in the dark [73]. This finding supports the hypothesis that *C. aggregans* in Nakabusa Hot Springs cyanobacterial mats can produce BChls for phototrophic growth under aerobic conditions during the daytime.

The high relative transcription levels for glutathione peroxidase-encoding genes during the midday not only reflect high O2 levels in the environment; they might also indicate high H2O2 levels, since the main function of this enzyme is to reduce lipid hydroperoxides to their corresponding alcohols and to reduce free hydrogen peroxide to water [77,78]. Hydrogen peroxide can be formed via photochemical reactions with dissolved organic carbon in hot spring waters [79], and H2O2 is produced as a byproduct of the oxidation of glycolate (to glyoxylate) [66], which has been shown to be present under high-light conditions in hot spring microbial mats, presumably produced by photoinhibited cyanobacteria [69]. The presence of glycolate and its oxidation at this time of day (11:00) is further suggested by the gene transcription pattern of glycolate oxidase genes, which showed patterns similar to those of the glutathione peroxidase, with highest relative transcription levels during the midday (Figure S3). Glycolate oxidase transcription activity supports the hypothesis that *Chloroflexus* species photoassimilate the glycolate supplied by the cyanobacteria in such microbial mats [80]. This indicates a photoheterotrophic metabolism for *C. aggregans*, which is further supported by the high relative transcript levels for glycoside, sugar, and amino acid transporter genes during high-light conditions (Figures S4 and S5).

Carbon monoxide (CO) might function as an electron and/or carbon source under aerobic conditions during this time of day, as aerobic carbon monoxide dehydrogenase genes are found to peak at midday [23]. The transcription profile of carbon monoxide dehydrogenase genes was related to the transcription pattern of the phosphoenolpyruvate carboxylase gene (Figure 6), which indicates that CO might be converted to CO2, which then is incorporated by the phosphoenolpyruvate carboxylase-catalyzed anaplerotic reaction (phosphoenolpyruvate to oxaloacetate). In *Thermomicrobium roseum*, an obligately aerobic chemoheterotroph in the phylum *Chloroflexota*, carbon monoxide dehydrogenase is utilized to produce ATP and NADPH under aerobic conditions [81]. It is speculated that *C. aggregans* may use CO for both anaplerotic carbon fixation and supplemental energy production in the presence of O2 due to the limitation of available CO2 caused by active cyanobacterial photosynthetic carbon fixation.

Two gene clusters encoding the "respiratory" complex I (NADH:menaquinone oxidoreductase, *nuo* genes) are present in *C. aggregans* and *C. aurantiacus* [54,82], as well as in the red FAP species *Roseiflexus castenholzii* and *Roseiflexus* sp. RS-1 (acc. nos. CP000804 and CP000686, respectively [23,37]). To our knowledge, neither of these clusters has been affiliated with phototrophic electron transport. The different relative transcription patterns obtained in the present study might indicate that the 14-gene set is used primarily in photosynthesis and the 12-gene set is used primarily in respiratory electron transport. However, two sets of *nuo* operons have also been described for other, non-phototrophic bacteria such as *Ignavibacterium album* [83] and "*Candidatus* Thermonerobacter thiotrophicus" [18], which might indicate specification to different O2 levels rather than phototrophic versus

respiratory electron transport. If true, this might indicate a potential higher O2 tolerance for the 14-gene set and a higher oxygen affinity for the 12-gene set. Because these results indicate that chemoheterotrophic, respiratory metabolism does not take place during highlight and superoxic conditions at midday, when the 14-gene set is highly transcribed, it is hypothesized that this NADH:menaquinone oxidoreductase plays a role in phototrophic electron transport, perhaps by donating electrons from NADH similar to as it has been proposed to be the case in the cyclic electron transport chain in heliobacteria [84]. However, further biochemical analyses are required to precisely determine the different functions of the two NADH: menaquinone oxidoreductases in *Chloroflexus*.

#### *4.4. Low Light and Low O2 Dominated the Afternoon Hours (15:00–16:00)*

In the afternoon, a substantial decrease in solar irradiance and O2 was observed after 14:00, as the sun set behind the surrounding mountains. Between 15:00 and 16:00, O2 was still detected only in the upper layer of the mat (Figure 2), and *C. aggregans* is hypothesized to experience microaerobic conditions enabling aerobic chemoheterotrophic metabolism, as indicated by the high relative transcription levels of TCA cycle, ACIII Cr, *aurA*, and cytochrome *c* oxidase-encoding genes (Figures 4, 5 and 8).

Simultaneously, high relative transcription levels of the gene encoding the O2-sensitive version of Mg-protoporphyrin monomethylester cyclase, *bchE*, indicated that part of the *C. aggregans* population was exposed to anoxic conditions, at least in the deeper mat layers, as supported by the microsensor data of the vertical O2 distribution in the mat (Figures 2 and 3). High relative transcription levels of genes involved in the 3-OHP bi-cycle sugges<sup>t</sup> autotrophic growth, especially under anaerobic conditions, where sulfide and/or H2 is available [85]. Sulfide concentrations are expected to rise under anaerobic conditions due to biological sulfate-reduction, as was similarly shown for the bacterial community of Mushroom Spring in Yellowstone National Park [86]. The sulfide oxidation capabilities of *C. aggregans* in the Nakabusa mats are supported by high relative transcription levels of *sqr* for utilization of sulfide as an electron donor [87]. At the same time, uptake hydrogenase genes are transcribed, pointing to the use of molecular hydrogen for autotrophic growth (Figure 7). However, although anoxic low-light conditions were prevalent in the deeper layers of the microbial mats for approx. 2 h before sunset, phototrophy does not seem to be the predominating metabolic growth mode during this time of day, as the decreasing relative transcription levels of ACIII Cp suggest. Thus, sulfide- and H2-oxidizing enzymes of *C. aggregans* may be employed for aerobic chemoautotrophic metabolism instead, or additionally for photoautotrophic metabolism at dusk, as suggested by the transcriptional peaks of ACIII Cr, *aurA* and cytochrome *c* oxidase (Figures 4 and 5).

#### *4.5. Dark and Anoxic Nighttime Hours (17:00–19:00, 23:00, 02:10)*

After 17:00, the microbial mat community experienced dark and anoxic conditions. The low relative transcription levels for housekeeping genes such as DNA gyrase and DNA/RNA polymerases indicate low metabolic activity of *C. aggregans* during this period. The low transcription values for transporter genes support this conclusion. The electron transport gene transcripts for both phototrophy and respiration as well as for the ATPsynthase genes were low. The unidirectional glycolysis enzyme 6-phosphofructokinase gene showed significant changes towards high relative transcription levels at the beginning of the night while TCA cycle-related gene transcriptions were low, suggesting fermentative metabolism and possibly the degradation of internal glycogen storage. This has been shown for *C. aurantiacus* in the laboratory [56] and has been suggested for FAPs inhabiting hot spring microbial mats in Yellowstone National Park [23]. High relative transcription levels of a bi-directional hydrogenase further support the fermentative growth mode and the potential production of H2 by *C. aggregans* during the night. This is in accordance with the recent detection of H2 production under fermentative conditions (dark, anaerobic) in *C. aggregans* strain NA9-6 (unpublished data).

#### *4.6. Early Morning Hours (05:00)*

Under dark, anoxic conditions in the early morning hours, an unexpected significant increase in the relative transcription of genes encoding the respiratory chain components (respiratory complexes I, II, and IV and ACIII Cr) as well as ATP-synthase was observed. It is suggested that the transcription of these genes is indicative of the occurrence of chemotrophic growth involving O2 respiration at that time of day. Although the microsensor measurements showed no presence of O2 in the mats until later in the morning, the transcription of genes encoding enzymes involved in O2 protection, such as superoxide dismutase and glutathione peroxidase, indicated a (micro)aerobic environment for *C. aggregans* at that time of day, and trace amounts of O2 were detected at the very surface throughout the night. As *C. aggregans* is known to have gliding motility and chemotaxis toward reduced O2 concentrations [2,88–91], it is speculated that *C. aggregans* migrates from anaerobic deeper layers to the micro-oxic surface layers in the early morning, in which a diffusive supply of O2 from the overlying water leads to microaerobic conditions during the nighttime, as has been suggested previously [23,92].

Similar to the ACIII Cr operon, the genes of the ACIII Cp operon, presumed to be involved in phototrophy, were also highly transcribed at 05:00. However, active phototrophy is ruled out due to the lack of light. Because the red filamentous anoxygenic phototrophic members of *Chloroflexota*—i.e., *Roseiflexus* spp.—contain only one copy of Cp-like ACIII genes, which are predicted to work under both phototrophic and chemotrophic conditions [23], the Cp-related genes in *C. aggregans* might also function under chemotrophic growth in the mats.

A capacity for chemoautotrophic growth was very recently observed in *Chloroflexus* spp. isolates obtained from Nakabusa Hot Springs microbial mats [8]. The high relative transcription levels of *hyd* uptake hydrogenase and Ni-transporter genes further sugges<sup>t</sup> the use of H2 as an electron donor for the aerobic chemoautotrophic growth in the microoxic surface layers of the cyanobacterial mats around 05:00. Additionally, the high transcription levels of genes for the TCA cycle and acetate/CoA ligase—the latter of which catalyzes the production of acetyl-CoA from acetate—indicate that acetate, supplied mainly from the fermentation of co-existing microbes as shown in similar mats in Yellowstone National Park, might be taken up at this time of day [22,68,93,94] (Table S5). This points to the possibility of an assimilation of acetate in addition to the purely autotrophic metabolism, suggesting the chemomixotrophic lifestyle of *C. aggregans* during predawn.
