Plasmids were cloned in chemically competent Escherichia coli 10-beta (New England Biolabs) or TOP10 (Thermo Scientific). E. coli cultures were maintained at 37 °C in enriched Luria-Bertani (LB) medium (24 g/L yeast extract, 10 g/L tryptone, 5 g/L NaCl) or LB agar medium (5 g/L yeast extract, 10 g/L tryptone, 5 g/L NaCl, 15 g/L agar) plates with 50 μg/mL apramycin (Thermo Scientific). Unless described as otherwise, A. bescii strains were cultured in 50 mL of CM516 medium in 125 mL serum bottles sealed with 20 mm butyl rubber stoppers at 70 °C without shaking. CM516 medium follows the recipe for CC516 described previously by Lipscomb et al. (2016) with the only modification being a change in carbon source from 5 g/L cellobiose to 5 g/L maltose substrate. The CM516 medium was supplemented with 50 μg/mL kanamycin (IBI Scientific) as appropriate, and is referred to as CM516K medium. Sealed serum bottles containing sterile medium were made anaerobic through vacuum and gas cycling, with the headspace being replaced with 80% (v/v) N₂ and 20% (v/v) CO₂ gas. As is standard, A. bescii cell density was measured as the optical density at 680 nm (OD680) using a cuvette in a Nanodrop One C spectrophotometer (Thermo Scientific) with 1× DSM 516 salt solution used as the blanking solution (Lipscomb et al., 2016; Rodionov et al., 2021; Tjo et al., 2025).

Tables of oligonucleotide primers and synthesized DNA used to construct the plasmids in this study can be found in the Supplementary Tables S1,S2, respectively. The two promoter sequences used to express the reporter genes in this study consisted of the 200 bp sequences immediately upstream of the start codon of their associated gene (Table 1). These promoters were P_slp associated with the S-layer protein gene (Athe_2303), and P_bh associated with a bifurcating-hydrogenase gene (Athe_1295) (Table 1). Maps of plasmids constructed and utilized in this study are shown in Figure 1. pSBS4 (empty vector) was obtained from the lab of Dr. Robert Kelly (North Carolina State University) (Lipscomb et al., 2016). This vector consists of a native A. bescii replicating plasmid (pAthe02), the htk gene expressed by promoter P_S30 associated the S30 ribosomal protein (Athe_2105), as well as elements that enable cloning in E. coli including: an apramycin resistance marker (Apr), replication initiation protein A (repA), and the pSC101 origin (Figure 1) (Chung et al., 2013; Lipscomb et al., 2016). Vectors pJLG091 and pJLG093 express the ⍺-galactosidase from Pyroccous furiosus (Pf⍺gal) and the β-galactosidase from Caldivirga maquilingensis (Cmβgal) respectively with P_slp (Figure 1). This expression site is based on the protein expression construct used previously in pJMC046 with the P_slp promoter and Calkro_0402 terminator, but is relocated on the pSBS4 backbone between Apr and pAthe02 (Conway et al., 2018). The backbone DNA for these vectors was PCR amplified from a sequenced plasmid that had been constructed previously via the insertion of a different P_slp driven gene into the pSBS4 backbone at this same site (Supplementary Table S1; Primers JLG021-22). Codon optimized genes flanked by appropriate overlapping regions were purchased (Twist Biosciences) for Pf⍺gal and Cmβgal (Supplementary Table S2) and assembled into plasmids via Gibson Assembly using the NEBuilder HiFi DNA Assembly kit (New England Biosciences). Vectors pSBS4 (empty vector), pJLG091 (P_slp–Pf⍺gal), and pJLG093 (P_slp–Cmβgal) were then cloned into chemically competent E. coli 10-beta, isolated using ZymoPURE miniprep kits (Zymo Research), and sequence confirmed (Azenta Genewiz). pJLG161 is identical to pJLG093 except that expression of Cmβgal is driven instead by P_bh (Figure 1). pJLG161 was constructed from pJLG093 in partnership with the Department of Energy Joint Genome Institute (JGI) at Lawrence Berkely National Lab (Berkely, CA) as described below. pJLG093 was first modified to create unique PmeI sites, aiding subsequent promoter insertion. The vector was linearized by PCR amplification (Supplementary Table S1; B431.093.VM.F & VM.R), and re-circularized via Gibson assembly together with an ultramer (Supplementary Table S1; JGI.UM1) purchased from Integrated DNA Technologies, using the NEBuilder HiFi DNA assembly kit. After validation of the modified vector, the sequence corresponding to P_bh was flanked by linkers designed for assembly into pJLG093_PmeI linearized by PmeI digest (Supplementary Table S2), purchased (Twist Biosciences) and assembled using the NEBuilder HiFi kit. These assemblies were subsequently transformed into chemically competent E. coli Top10 of which candidate colonies were picked, sequence verified on the Pacific Biosciences Revio platform (Pacific Biosciences), and analyzed using custom pipelines at the Joint Genome Institute. pJLG161 was subsequently isolated using ZymoPURE miniprep kits (Zymo Research), and sequence confirmed (Azenta Genewiz).

For transformation in A. bescii, larger quantities of plasmid DNA were extracted from E. coli using the ZymoPURE maxiprep kit (Zymo Research). Extracted plasmids were then methylated in vitro using the M.CbeI methyltransferase and purified via phenol-chloroform extraction as previously described (Chung et al., 2012; Lipscomb et al., 2016). Wild type A. bescii DSM 6725 was obtained from the lab of Dr. Robert Kelly (North Carolina State University). Competent A. bescii were grown on CM516 media containing amino acids (CM516-AA) to an optical density at 680 nm (OD680) of 0.04–0.08 and prepared for transformation as described previously (Lipscomb et al., 2016). Fifty microliters of competent cells were transformed with 1–2 μg of plasmid in a 1 mm gap electroporation cuvette using a Bio-Rad gene pulser at 1800 V, 400 Ω, and 25 μF. Electroporated cells were immediately resuspended in 1 mL of CM516 media and transferred to tubes containing 10 mL of the same media pre-heated to 70 °C. Cells were allowed to recover for 90 min before being transferred to pre-heated bottles containing 50 mL of selective CM516K media. After 24–36 h of growth, cells were passaged into 10 mL of fresh CM516K media and allowed to grow overnight. Passaged cells were then plated and grown for 48 h in solid selective CM516K media with 1.5% (w/v) agar at 70 °C under a 95% (v/v) N₂ and 5% (v/v) H₂ atmosphere in an anaerobic chamber. Single colonies were picked and screened via colony PCR (Supplementary Figure S1) using primers JLG181 and JLG224 (Supplementary Table S1). Presence of the correct promoter-reporter sequences were confirmed by long-read sequencing of colony PCR products (Azenta Genewiz, PCR-EZ).

To detect galactosidase activity in A. bescii cells, para-nitrophenol-α-D-galactopyranoside (pNPαGal) and para-nitrophenol-β-D-galactopyranoside (pNPβGal) obtained from TCI chemicals, were used as colorimetric substrates. Substrate solutions contained 5 mM pNPαGal or pNPβGal dissolved in 100 mM sodium acetate pH 5.5 buffer which was chosen based on the previously determined optimal pH ranges of Pf⍺gal and Cmβgal (van Lieshout et al., 2003; Letsididi et al., 2017). A. bescii cells were prepared for galactosidase assays by first pelleting 5–15 mL of freshly grown cells at the maximum rotor speed (7,000 × g for 15 mL pellets or 21,000 × g for 5 mL pellets) for 10 min, followed by removal of the supernatant and storage at −80 °C for later use. Immediately prior to testing, cell pellets were resuspended and concentrated 3–5× in 1–3 mL of 100 mM pH 5.5 sodium acetate buffer to a final OD680 of 0.35–0.5, measured on a Nanodrop One C spectrophotometer with 100 mM sodium acetate buffer as the blanking solution. For assays involving heat-treatments, 50–100 μL of cells or blank buffer were aliquoted into PCR strip tubes and incubated in a thermocycler at 90 or 98 °C for 10 min unless described otherwise. To begin the reaction, 10–30 μL of cells or blank buffer were added to 60–80 μL of substrate solution or blank buffer to a total volume of 90 μL. Assays that involved wild type or the P_slp–Pf⍺gal strain of A. bescii required 30 μL of cells, while testing of Cmβgal expressing A. bescii only required 10 μL of cells per reaction. Reactions were incubated in a thermocycler at the appropriate temperature for the experimentally prescribed time after which all reactions were immediately quenched with the addition of 180 μL of 1 M sodium carbonate. The absorbance at 405 nm (A405) of 100 μL of each reaction was then measured in a flat-bottomed clear 96 well plate using a BioTek SynergyH1 microplate reader (Agilent). For all reaction conditions the following controls were included: a substrate only (no cell) condition to account for the thermal background degradation of substrate, a no substrate condition for each cell type to account for background scattering from cellular debris, and a buffer only condition to isolate the absorbance due to debris in the prior control from the buffer itself. All reaction conditions were performed in technical triplicate.

Normalized galactosidase activity was calculated as defined in Equation 1 based on the equations in “Experiments in Molecular Genetics” for measuring β-galactosidase activity in E. coli using o-nitrophenyl-β-D-galactopyranoside (Miller, 1972). The most notable modifications to the quantification formula used by Miller (1972) are cellular debris background is explicitly accounted for with a series of control reactions rather than estimated with the absorbance at 550 nm, and normalization is done with the optical density at 680 nm (OD680) rather than that at 600 nm (OD600). In Equation 1, the A405 of the no cell control (A405_NC) is subtracted from the A405 of the experimental condition (A405_exp) to remove thermal background degradation of substrate. Separately, the A405 of the buffer only control (A405_BO) is subtracted from that of the no substrate control (A405_NS). This is then subtracted from the A405_exp − A405_NC difference to account for debris scattering. This final value is then divided by the previously measured OD680 of the resuspended A. bescii input to the assay to normalize for differences in the amount of cells added.

To test for background activity from endogenous galactosidases, wild type A. bescii DSM 6725 were grown on CM516 media to an OD680 of 0.15 (late exponential), pelleted, and frozen. Cells were resuspended and heat-treated at 90 or 98 °C for 0, 10, 30, or 60 min prior to adding pNPαGal or pNPβGal solutions in the enzyme assay described above, with incubation for 2 h at 75 °C.

Prior to initial testing of the hyperthermophilic galactosidase reporters, A. bescii containing the empty vector (pSBS4), P_slp–Pf⍺gal (pJLG091), and P_slp–Cmβgal (pJLG093), were grown on selective CM516K media to an OD680 of 0.07–0.12 (mid-late exponential) and harvested as described above. To test the effects of various heat-treatment conditions on the reporters, resuspended empty vector, P_slp–Pf⍺gal, and P_slp–Cmβgal cells were heat-treated at 90 or 98 °C for 10 min which were subsequently added alongside un-heat-treated cells to both the pNP substrate solutions. For assays testing P_slp–Pf⍺gal against the empty vector, cells were incubated for 2 h, while for assays testing P_slp–Cmβgal, cells were incubated for 20 min. Signal to noise ratio was calculated as the normalized activity of reporter expressing cells on their respective preferred pNP substrates divided by that of the empty vector control at the corresponding conditions (i.e., P_slp–Pf⍺gal/empty vector activity on pNP⍺Gal, or P_slp–Cmβgal/empty vector activity on pNPβGal).

For time course experiments, cells that contained the empty vector or P_slp–Pf⍺gal were prepared and heat-treated at 98 °C. Cells were then added to pNP⍺Gal solution and incubated for 0, 1, 2, or 3 h at 98 °C. Separately this was repeated for empty vector and P_slp–Cmβgal cells except heat-treatments were carried out at 90 °C, cells were instead added to pNPβGal solution, and incubations were carried out at 90 °C for 0, 10, 20, or 30 min. For temperature optimization testing, cells that contained empty vector, P_slp–Pf⍺gal, and P_slp–Cmβgal were prepared and heat-treated at 90 °C which were subsequently added alongside un-heat-treated cells to their corresponding pNP substrate solutions (i.e., P_slp–Pf⍺gal & empty vector on pNP⍺Gal; P_slp–Cmβgal & empty vector on pNPβGal). These reactions were then incubated at the following temperatures: 75, 80, 85, 90, 95, and 98 °C. For this test, assays on pNP⍺Gal substrate were incubated for 2 h, while assays on pNPβGal were incubated for 10 min.

To start the growth curves of A. bescii, strains containing empty vector (pSBS4), P_slp–Cmβgal (pJLG093), and P_bh–Cmβgal (pJLG161) were inoculated at a target OD680 of 0.002 in 50 mL of selective CM516K media. Cultures were grown for 29 h in biological triplicate at 70 °C, with each culture’s OD680 measured at intervals of roughly 3–5 h. At 12, 18, 24, and 29 h, 4–5 mL of each culture was removed, after which cells were pelleted, frozen, and assayed as described above. To test for the activity of Cmβgal, thawed pellets were prepared as described above with heat treatment at 90 °C. Cells were then added to pNPβGal solution and incubated for 10 min at 90 °C.

Similar to previous studies, RNA was isolated from A. bescii containing pSBS4, pJLG093, and pJLG161 that were grown on CM516K media for 18 h to OD680 values of 0.11–0.16 (mid-late exponential phase), with three biological replicates for each strain (Williams-Rhaesa et al., 2018; Straub et al., 2020; Tanwee et al., 2023; Bing et al., 2024). After growth, 30–40 mL of cells were immediately pelleted at 6,000 × g for 10 min and, after removal of the supernatant, frozen at −80 °C. Prior to purification, thawed cell pellets were lysed as previously described with the addition of 240 μL of cold PBS, 75 μL of lysozyme (20 mg/mL), and 300 μL of the Monarch^® gDNA Tissue Lysis Buffer (New England Biosciences), followed by incubation at 37 °C for 15 min (Bing et al., 2024). Three hundred microliters of lystate from each pellet was then added to two volumes (600 μL) of Monarch^® StabiLyse DNA/RNA Buffer (New England Biosciences) (Bing et al., 2024). From this, RNA was purified using the Monarch^® Spin RNA Isolation Kit (New England Biosciences) as per the manufacturer’s instructions with the on-column DNase I treatment step. RNA concentrations were quantified using a Nanodrop One spectrophotometer (Thermo Scientific). qRT-PCR assays were carried out on a Viia7^™ Real-Time PCR System (Thermo Scientific). qRT-PCR on extracted RNA samples was performed using the Luna^® Universal One-Step RT-qPCR Kit (New England Biosciences) according to the manufacturer protocol, with 50 ng of total RNA added to 10 μL reactions in a 384 well plate. A no-RT control condition was included for each experimental condition to check for DNA contamination. All reaction conditions, including for each biological replicate, were performed in technical triplicate. Expression of the Cmβgal gene (Supplementary Table S1; Primers JLG219-220) was calculated relative to that of the A. bescii gapdh (Athe_1406) using primers (Supplementary Table S1; Primers JLG211_CTS480-JLG212_CTS481) utilized in a previous A. bescii study (Straub et al., 2020).

The genome of wild type A. bescii contains at least one characterized ⍺-galactosidase, along with several putative ⍺- and β-galactosidases (Lee et al., 2017; Drula et al., 2022). To assess the level of heat treatment needed to eliminate background activity from these enzymes on colorimetric pNP-glycoside substrates, prepared wild type A. bescii cells were heat treated at 90 or 98 °C for 0–60 min. Cells were then added to solutions of pNPαGal or pNPβGal and incubated for 2 h at 75 °C to test for endogenous ⍺- or β-galactosidase activity, respectively. Significant background activity was detected on both substrates with A. bescii cells that were not heat-treated (Figures 2a,b). However, this background activity was eliminated by heat treatment for as short as 10 min at either 90 or 98 °C (Figures 2a,b), indicating that native A. bescii ⍺- and β-galactosidases were inactivated with this relatively short incubation at temperatures above the organism’s optimal growth temperature.

To determine whether the hyperthermophilic galactosidases Pf⍺gal and Cmβgal could serve as effective enzymatic reporters under these conditions, wild type A. bescii DSM 6725 was transformed with plasmids pJLG091 (P_slp–Pf⍺gal) and pJLG093 (P_slp–Cmβgal), which drive strong constitutive expression of each galactosidase reporter with P_slp (Figure 1). A strain containing empty vector pSBS4 (Lipscomb et al., 2016) was also constructed as a control. Pf⍺gal and Cmβgal were selected because, among previously characterized hyperthermophilic galactosidases, they had the highest reported optimal temperatures, 115 °C and 110 °C, respectively, as determined by short in vitro assays with purified enzyme (van Lieshout et al., 2003; Letsididi et al., 2017). However, while Pf⍺gal is reported to be extremely thermostable (half-life of 15 h at 100 °C), Cmβgal is reported to lose all activity within 120 min at 95 °C and within 50 min at 100 °C (van Lieshout et al., 2003; Letsididi et al., 2017). Given the need to heat-treat A. bescii cells to eliminate native galactosidase activity and the differences between assay conditions, we evaluated the performance of both enzymes in our system.

Empty vector, P_slp–Pf⍺gal, and P_slp–Cmβgal cells were heat-treated at 90 or 98 °C for 10 min and tested alongside un-heat-treated cells at the same temperatures on both pNPαGal and pNPβGal (Figures 2c–h). In un-heat-treated samples, both reporters showed detectable activity above the empty vector control on their preferred substrates at 90 °C or 98 °C (left of Figures 2c,d,f,g). After a 10 min heat treatment at the assay temperature, background activity was reduced while preserving reporter activity in all cases except for Cmβgal at 98 °C, where the enzyme was fully inactivated during the 98 °C heat treatment (right of Figures 2c,d,f,g). For cases where the enzyme remains active, heat treatment improved signal to noise ratios (Figures 2e,h). Notably, Cmβgal at 90 °C showed a marked improvement in signal to noise ratio, increasing from 32 times to 170 times background with heat treatment (Figure 2h). Signal to noise for Pf⍺gal assays are also improved with heat treatment (Figure 2e), but much more modestly due to its weaker activity in general, though this enzyme does retain function under our assay conditions at both 90 and 98 °C (Figures 2c–e). Finally, the two reporters act orthogonally, with no detectable activity above that of the empty vector control detected on their non-preferred substrates (Figures 2c,d,f,g). Because heat treatment at 90 °C for 10 min achieved improvements in signal to noise ratio while maintaining activity of both enzymes, this heat treatment was chosen as the standard in subsequent assays.

Next, the optimal reporter assay conditions were evaluated. To ensure a sufficiently strong signal while maintaining approximately linear behavior with respect to incubation time, we measured enzymatic activity as a function of assay duration on heat-treated cells (Figures 3a,b). These assays showed that a 2–3 h incubation was appropriate for Pslp–Pf⍺gal on pNPαGal (Figure 3a), whereas a 10–20 min incubation was appropriate for Pslp–Cmβgal on pNPβGal (Figure 3b). To optimize assay temperature, assays were conducted on the reporter strains versus the empty vector control on their respective substrates from 75–98 °C with and without heat-treatment at 90 °C for 10 min (Figures 3c–f). Pf⍺gal produced the strongest signal at 98 °C (the highest reaction temperature we could reliably achieve in our thermocycler), and heat treatment was essential for eliminating the significant native ⍺-galactosidase activity at incubation temperatures below 85 °C (Figures 3c,d, without and with heat treatment, respectively). Cmβgal showed optimal activity at 90 °C, with decreasing activity at higher temperatures (Figures 3e,f). While background activity without heat treatment was much lower for pNPβGal (Figure 3e) than pNPαGal (Figure 3c), likely due to the shorter assay incubation time, heat treatment still eliminated nearly all native β-galactosidase activity (Figures 3e,f). While Cmβgal produced stronger signals than Pf⍺gal, its activity decreased at temperatures above 90 °C in the conditions of our assays (Figures 2f,g, 3e,f). This thermostability profile is in line with prior characterization of Cmβgal, which showed in vitro activity at its T_opt of 110 °C in a 2 min assay began to diminish immediately when the enzyme was pre-incubated at 100 °C, while no loss in activity was observed after incubation at 90 °C for 2 h (Letsididi et al., 2017). Considering its superior signal to noise ratio and stronger normalized activity in shorter incubation times, Cmβgal was chosen as the reporter for subsequent tests of expression in A. bescii.

Next, reporter expression was tested over the growth of A. bescii strains containing the Cmβgal reporter under the control of two previously utilized promoters, P_slp and P_bh (Table 1). Based on previous studies, the P_bh promoter should drive somewhat lower expression than P_slp (Williams-Rhaesa et al., 2018). A. bescii strains containing pSBS4 (empty vector), pJLG093 (P_slp–Cmβgal), and pJLG161 (P_bh–Cmβgal) were grown and monitored over the course of 29 h in biological triplicate (Figure 4a). At time points of 12, 18, 24, and 29 h, corresponding roughly to exponential, late exponential, early stationary, and stationary growth phases, respectively, cells were harvested for enzyme reporter measurement on pNPβGal with heat-treatment and incubation at 90 °C (Figure 4b). As expected, no significant activity was detected from the empty vector strain at any stage of growth (Figure 4b). The relative activity of the two promoters vary over the course of cell growth phase with P_bh–Cmβgal having 37, 56, 72, and 73% the activity of P_slp–Cmβgal at 12, 18, 24, and 29 h of growth, respectively (Figure 4b). In general, activity from both Cmβgal expressing strains appears to increase as A. bescii enters stationary phase (24- and 29-h timepoints), though variability between biological replicates also increases in stationary phase (Figure 4b).

To assess how the Cmβgal reporter activity levels mirror transcript levels, qRT-PCR was performed on the Cmβgal gene. RNA was extracted from empty vector, P_slp–Cmβgal, and P_bh–Cmβgal A. bescii strains in late exponential phase (18-h timepoint) grown in biological triplicate. Levels of Cmβgal transcription in each strain were calculated relative to that of the endogenous A. bescii glyceraldehyde-3-phosphate dehydrogenase gapdh (Athe_1406) housekeeping gene as is standard in the literature (Williams-Rhaesa et al., 2018; Straub et al., 2020; Tanwee et al., 2023). Results show that both P_slp and P_bh drive strong levels of transcription, with expression of 15.8× and 4.8× that of gapdh, respectively (Figure 5). P_slp is the stronger promoter with an average level of transcription 3.3× that of P_bh (Figure 5). This mirrors a smaller difference in enzyme activity, where Cmβgal expressed by P_slp produced an average enzyme activity 1.8× that driven by P_bh (Figure 4b).