**1. Introduction**

Members of the genus *Chloroflexus* are thermophilic, filamentous anoxygenic phototrophs (FAPs) in the phylum *Chloroflexota* (formerly *Chloroflexi*). They are well known to have the ability to grow photoheterotrophically under anaerobic conditions and chemoheterotrophically under aerobic conditions in the laboratory [1–3]. While photoautotrophic

**Citation:** Kawai, S.; Martinez, J.N.; Lichtenberg, M.; Trampe, E.; Kühl, M.; Tank, M.; Haruta, S.; Nishihara, A.; Hanada, S.; Thiel, V. In-Situ Metatranscriptomic Analyses Reveal the Metabolic Flexibility of the Thermophilic Anoxygenic Photosynthetic Bacterium *Chloroflexus aggregans* in a Hot Spring Cyanobacteria-Dominated Microbial Mat. *Microorg* **2021**, *9*, 652. https://doi.org/10.3390/ microorganisms9030652

Academic Editor: Johannes F. Imhoff

Received: 28 January 2021 Accepted: 17 March 2021 Published: 21 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

growth in the laboratory has been observed only in a small number of isolated strains (e.g., *Chloroflexus aurantiacus* strain OK-70-fl [4–6], *Chloroflexus* sp. strain MS-G [7], and *Chloroflexus aggregans* strains NA9-6 [8,9] and ACA-12 [10]), the genes necessary for the 3-hydroxypropionate (3-OHP) bi-cycle, which is a carbon fixation pathway found only in members of the order *Chloroflexales* among bacteria, are present in all of the available *Chloroflexus* spp. genomes [11]. *C. aggregans* strains NA9-6 and ACA-12, which were isolated from Nakabusa Hot Springs (Nagano Prefecture, Japan), can grow photoautotrophically with hydrogen gas (H2) [8] and sulfide [10] as the electron donors in pure culture, respectively. In addition to the long known phototrophic and chemoheterotrophic metabolism in *Chloroflexus* spp., chemoautotrophic growth has recently been shown in lab studies of *C. aggregans* strain NA9-6 [8]. In addition, fermentative growth has been shown in two isolates of *C. aurantiacus*, strains B3 and UZ [12].

Microbial mats in the slightly alkaline, sulfidic Nakabusa Hot Springs have been intensively studied with regard to their microbial diversity and functions [13–21]. At water temperatures of 63–70 ◦C, olive-green microbial mats ("*Chloroflexus* mats") are dominated by *C. aggregans* [14,15], and oxygenic cyanobacteria are not found. At lower temperatures of 45–62◦C, *Chloroflexus* spp. co-exist with cyanobacteria in dark blue-green microbial mats ("cyanobacterial mats"). These blue-green mats are stratified with a green upper layer dominated by the thermophilic cyanobacteria on top of an orange-colored layer that is frequently inhabited by *C. aggregans* [13].

In the anoxygenic, cyanobacteria-free phototrophic mats, *C. aggregans* is considered to be the main primary producer, using sulfide as the major electron source [9,10,14–16]. The metabolic repertoire of *C. aggregans* in the blue-green cyanobacterial mats has remained unstudied. In situ isotopic studies of similar cyanobacterial mats colonizing the effluent channels of Mushroom Spring and Octopus Spring in Yellowstone National Park (YNP; WY, U.S.) suggested that filamentous phototrophic *Chloroflexota* vary their carbon metabolisms over a diel cycle [22]. Based on transcriptomic data, Klatt et al. (2013) suggested photomixotrophic growth of a member of FAPs—i.e., *Roseiflexus* spp.—in Mushroom Spring during daytime and fermentative growth during the night [23]. Compared to the microbial mats in YNP, Nakabusa Hot Spring cyanobacterial mats are rich in *C. aggregans*, at a relative abundance of approximately 21–22% compared to only 1% in Mushroom Spring cyanobacterial mats [13,24]. This suggests an important ecological role and potential function of *C. aggregans* as a primary producer in the Nakabusa mats.

In this study, the in situ metabolic lifestyle of *C. aggregans* in the blue-green microbial mats of Nakabusa Hot Springs was analyzed by using a metatranscriptomic approach. Light is the main energy source during daytime, supporting photoautotrophic, photomixotrophic and photoheterotrophic growth of *C. aggregans*, while chemotrophic growth is prevalent during the afternoon and night. During the afternoon, under microaerobic low-light conditions chemoheterotrophic growth is based on O2 respiration, while at night fermentation is conducted under anaerobic conditions. Unexpectedly, chemoautotrophic growth using O2 as the terminal electron acceptor appeared to take place during early morning hours before sunrise, which suggests a vertical migration of *C. aggregans* cells to the microaerobic surface layers of the mats.

#### **2. Materials and Methods**

#### *2.1. Field Site and Sample Collection*

Blue-green cyanobacterial mat samples were collected from a small pool at 56 ◦C with slightly alkaline (pH 8.5–8.9) and sulfidic (46–138 μM) hot spring water [18,25–27] at Nakabusa Hot Springs, Nagano Prefecture, Japan (36◦2333 N, 137◦4452 E) [20]. Microbial mat samples of approximately 3 mm thickness with two distinct vertical layers, a green top layer and an orange-colored bottom layer (Figure 1), were randomly collected in triplicate using a size 4 cork borer (8 mm diameter) as previously described [13,14]. Samples were placed in 2 mL screw-cap tubes and snap-frozen in a dry-ice cooled, 70% (v/v) ethanol bath on site. Samples were taken at 12 different time points over a diel

cycle on 3 to 4 November, 2016 (19:00 and 23:00 on 3 November; 02:10, 05:00, 06:00, 07:00, 11:00, 15:00, 16:00, 17:00, 18:00, and 19:00 on 4 November) and were brought back to the laboratory on dry ice and stored in a −80 ◦C freezer until further processing for metatranscriptomic analyses.

**Figure 1.** Photographs of the sampling site and cyanobacteria-dominated microbial mat. (**a**) The dark blue-green microbial mats developed at the "Stream Site" of Nakabusa Hot Springs, Japan [19]. (**b**) The microbial mat core samples collected at each time point were 8 mm in diameter and approx. 15 mm thick. The upper 3 mm of the core samples was used in this study.

#### *2.2. RNA Extraction*

RNA extraction from microbial mat samples was performed as previously described [18]. Briefly, 0.10–0.21 g wet weight samples were used for RNA extraction with an RNeasy PowerBiofilm Kit (Qiagen, Valencia, CA, USA) following the manufacturer's protocol. The RNA was treated with DNase I and eluted with RNase-free water. Purity and concentration of the RNA were determined using an RNA High Sensitivity (HS) assay with a Qubit 3.0 fluorometer (Life Technologies, Grand Island, NY, USA).

## *2.3. RNA Sequencing*

Library preparation and sequencing of the RNA samples were conducted at DNALink Inc. (Seoul, Korea) as described previously [18]. RNA purity was determined by assaying 1 μL of total RNA extract on a NanoDrop8000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Total RNA integrity was assessed by the RNA integrity number (RIN) using a 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). Total RNA sequencing libraries were prepared using a Truseq Stranded Total RNA Library prep kit and Ribo-Zero bacteria kit (both from Illumina, San Diego, CA, USA) according to the manufacturer's instructions.

First, 0.5 μg of total RNA was subjected to ribosomal RNA depletion with Ribo-Zero bacteria reagen<sup>t</sup> using biotinylated probes that selectively bind rRNA species. Following purification, the rRNA-depleted total RNA was fragmented into small pieces using divalent cations under elevated temperature. The cleaved RNA fragments were copied into firststrand cDNA using random primers and reverse transcriptase, followed by second-strand cDNA synthesis using DNA polymerase I and RNase H. A single 'A' base was then added to these cDNA fragments, and the adapter was ligated. The products were purified and enriched by polymerase chain reaction (PCR) to create the final cDNA library.

The quality of the amplified libraries was verified by capillary electrophoresis using the 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). After a quantitative (q)PCR using SYBR Green PCR Master Mix (Applied Biosystems, Carlsbad, CA, USA), index-tagged libraries were combined in equimolar amounts. RNA sequencing was

performed using an Illumina NextSeq 500 system following the provided protocols for 2 × 150 sequencing.

#### *2.4. Sequence Data Analyses*

Raw RNA reads were pre-processed using FastQC [28]. Adapter sequence and lowquality reads were trimmed by Cutadapt ver. 1.12 [29]. Quality-checked reads were mapped against the complete genome of *C. aggregans* DSM 9485<sup>T</sup> (RefSeq acc. No. NC\_011831.1) [2] with bowtie2 ver. 2.3.0 [30] with default settings allowing no mismatches. The reads were then aligned using the EDGE-pro algorithm [31] with the rRNA depletion option.

Transcriptomic analyses were conducted as described previously [18]. In short, read counts were normalized for each time point by the total number of reads retrieved for the target organism. The relative transcription of each gene during the cycle was then calculated and normalized against the mean of all of the reads at each time point for that particular gene over the diel cycle. This method allows comparison of the relative transcription abundance levels (rather than the absolute values) for each gene across the diel cycle.

## *2.5. Statistical Analyses*

As described above, diel transcriptomic data in this study lacked replication. In the following Results and Discussion sections, the authors carefully interpreted and described and intentionally averaged the normalized transcriptional patterns of several genes related in a single pathway to recognize those gene transcription patterns as the pathway-level metabolic dynamics. However, some important genes function in an important enzyme reaction solely, the statistical analyses of each gene were performed to discuss the transcriptional changes over a diel cycle. For each gene in dual datasets, and for every possible pair-wise comparison of the 11 sets of adjacent samples (November 3 19:00–23:00; 3 November 23:00–4 November 2:10; 4 November 02:10–05:00, 05:00–06:00, 06:00–07:00, 07:00–11:00, 11:00–15:00, 15:00–16:00, 16:00–17:00, 17:00–18:00, 18:00–19:00), "exactTest" program in edgeR with dispersion set at 0.1 was used to determine the probability that the gene was differentially transcribed in a statistically significant manner [32,33].

## *2.6. Microsensor Analyses*

The profiles of the O2 concentration as a function of depth in the microbial mat were measured in situ by using a Clark-type O2 microsensor (OX25; Unisense, Aarhus, Denmark) with a tip diameter of <25 μm, low stirring sensitivity (<1–2%) and fast response time (t90 < 0.5 s). The O2 microsensor was mounted on a motorized micromanipulator (Unisense, Aarhus, Denmark) and connected to a PC-interfaced pA-meter (Unisense, Aarhus, Denmark), both of which were controlled by dedicated data acquisition, profiling, and positioning software (SensorTrace Pro, Unisense, Aarhus, Denmark). The micromanipulator was mounted on a metal stand placed next to the hot spring, allowing for vertical insertion of the microsensor tip into the microbial mat under natural flow, temperature and light conditions. The microsensor tip was carefully positioned at the mat surface (defined as 0 μm) by manual operation of the micromanipulator. Subsequently, O2 microprofiles were recorded automatically every 15 min for 24 h starting at 18:00 on 3 November 2016. In each profile, O2 measurements were made in 100 μm increments from the water-phase and into the mat. One measurement was taken per depth and, for each measurement a 10 s wait period was applied, to ensure steady O2 signal, and the O2 signal was then recorded averaged over a 1 s period.

## *2.7. Irradiance Measurements*

Downwelling solar photon irradiance (400–700 nm) at the water surface next to the mat was logged every 5 min throughout the 24 h diel sampling cycle with a calibrated light meter connected to a cosine-corrected photon irradiance sensor (ULM-500, MQS-B; Walz, Effeltrich, Germany).
