Temperature-dependent expression of different guanine-plus-cytosine content 16S rRNA genes in Haloarcula strains of the class Halobacteria

Haloarcula strains, which are halophilic archaea, harbour two to three copies of 16S rRNA genes (rrsA, rrsB and rrsC) in their genomes. While rrsB and rrsC (rrsBC) show almost identical sequences, rrsA shows 4–6% sequence difference and 1–3% guanine-plus-cytosine content (PGC) difference compared to rrsBC. Based on the strong correlation between the PGC of 16S rRNA genes and the growth temperatures of the prokaryotes, we hypothesised that high-PGCrrsA and low-PGCrrsBC are expressed at high and low temperatures, respectively. To verify the hypothesis, we performed sequence analyses and expression surveys of each 16S rRNA gene in eight Haloarcula strains. The secondary structure prediction of the 16S rRNA via computer simulation showed that the structural stability of 16S rRNAs transcribed from rrsA was higher than that of 16S rRNAs transcribed from rrsBC. We measured expression levels of rrsA and rrsBC under various temperature conditions by reverse-transcriptase quantitative PCR. The expression ratio of high-PGCrrsA to low-PGCrrsBC increased with cultivation temperatures in seven of eight Haloarcula strains. Our results suggest that the transcription of high-PGCrrsA and low-PGCrrsBC may be regulated in response to environmental temperature, and that 16S rRNAs transcribed from high-PGCrrsA function under high temperature conditions close to the maximum growth temperature. Electronic supplementary material The online version of this article (10.1007/s10482-018-1144-3) contains supplementary material, which is available to authorized users.


Introduction
Prokaryotes harbour 16S rRNA genes encoding rRNA in the small subunit of the prokaryotic ribosome, and the sequences are highly conserved. Therefore, the 16S rRNA gene sequence is generally used for the identification of prokaryotes and for the assessment of prokaryotic diversity in various environments (Narihiro and Kamagata 2017;Yabe et al. 2017). A high percentage of prokaryotic genomes (83%) contain two or more copies of 16S rRNA genes, and the sequences are nearly identical among the copies of 16S rRNA genes (Sun et al. 2013). However, recent genome analyses revealed that some prokaryotes show intragenomic heterogeneity of 16S rRNA genes (Větrovský and Baldrian 2013;Sun et al. 2013). In particular, halophilic archaea of the genus Haloarcula have long been known to harbour significantly different sequences of 16S rRNA genes in their genomes (Mevarech et al. 1989;Acinas et al. 2004;Boucher et al. 2004;Pei et al. 2010;Sun et al. 2013).
Haloarcula strains have been isolated from solar salterns and salt lakes in deserts (Juez et al. 1986;Oren et al. 1990), where large day-night temperature variations are observed (Wieland et al. 2005;Sima et al. 2013;Andrade et al. 2015;Naghoni et al. 2017). These strains have two or three copies of rRNA operons (rrnA, rrnB and rrnC) in their genomes (Baliga et al. 2004;Liu et al. 2011;Ding et al. 2014;Yun et al. 2015;Sato et al. 2017). Each rRNA operon contains a 5S rRNA gene, a 16S rRNA gene (rrsA, rrsB or rrsC, respectively) and a 23S rRNA gene. The 5S rRNA gene sequences are almost identical among the rRNA operons. The 23S rRNA gene sequences show limited diversity (0.3-2%). For 16S rRNA gene sequences, rrsB and rrsC (rrsBC) show almost identical sequences, whereas rrsA shows different sequences (4-6%) from rrsBC (Mevarech et al. 1989;Mylvaganam and Dennis 1992;Dennis 1999;Sato et al. 2017). The different sequences of 16S rRNA genes were first reported in Haloarcula marismortui (Mevarech et al. 1989). Adjacent regions such as the putative promoter of 16S rRNA genes and the maturation process of the transcribed 16S rRNA have already been studied (Mylvaganam and Dennis 1992;Dennis et al. 1998;Dennis 1999;Amann et al. 2000). Previous studies showed that variation of the salinity in the medium did not change the expression level of each 16S rRNA gene in Haloarcula strains (López-López et al. 2007;Cui et al. 2009). However, it has been reported that variation of the cultivation temperature changed the expression level of each 16S rRNA gene in H. marismortui (López-López et al. 2007).
It has been shown that 16S rRNA gene sequences contain information regarding the thermal features of prokaryotes (Galtier and Lobry 1997;Khachane et al. 2005;Kimura et al. 2006;Wang et al. 2006;Kimura et al. 2010Kimura et al. , 2013. Those findings were based on a high correlation between the guanine-plus-cytosine contents (P GC ) of 16S rRNA genes and the growth temperatures of archaea; the P GC of 16S rRNA genes in thermophilic and hyperthermophilic archaea tend to be high, whereas the P GC of 16S rRNA genes in psychrophilic and mesophilic archaea tend to be comparatively low (Kimura et al. 2007). On the basis of the relationship between the P GC of 16S rRNA gene sequences and the growth temperatures of archaea, Kimura et al. (2013) proposed the microbial molecular thermometer (MMT), which uses linear regression equations to infer the minimum (T min ), optimal (T opt ) and maximum (T max ) growth temperatures of cultured and not-yet cultured archaea based on the P GC value of each 16S rRNA gene. Sato et al. (2017) used Haloarcula hispanica to verify a hypothesis that high-P GC rrsA (58.9%) and low-P GC rrsBC (56.4-56.5%) are expressed at high and low temperatures, respectively. They also investigated the physiological advantage of harbouring high-and low-P GC 16S rRNA genes with respect to the growth of H. hispanica under various temperature conditions. An expression survey by reverse-transcriptase quantitative PCR (RT-qPCR) showed that the expression ratio of rrsA to rrsBC increased with the incubation temperature. At the T max , i.e., 50°C, the expression level of rrsA was significantly higher than those of rrsBC. Sato et al. (2017) also constructed mutant strains with only one of three rRNA operons and cultivated the strains at various temperatures. The mutant strain harbouring only rrnA including high-P GC rrsA grew significantly slower than the wild-type strain at the T min 25°C, but it grew as fast as the wildtype at 45°C (T opt ) and 50°C (T max ). Sato et al. (2017) thus demonstrated that high-P GC 16S rRNA is important for the rapid growth of H. hispanica at high temperature.
Here, we investigated eight Haloarcula strains harbouring different sequences of 16S rRNA genes in their genomes (Sato et al. 2017). We sequenced the 16S rRNA genes in each strain and estimated the T min , T opt and T max based on the P GC of the 16S rRNA genes, using the MMT proposed by Kimura et al. (2013), in addition to measuring the growth temperatures by general cultivation methods. We also calculated the thermal stability of the secondary structure for each 16S rRNA transcribed from rrsA, rrsB or rrsC in the Haloarcula strains via computer simulation (in silico). The expression ratio of rrsA to rrsBC was determined under various temperature conditions by RT-qPCR. We used the MMT, secondary structure analysis and RT-qPCR to test our hypothesis that Haloarcula strains express and use high-P GC 16S rRNAs at high temperature and low-P GC 16S rRNAs at low temperature. Each strain was cultured using JCM Medium 307: 2 g of casamino acid (BD, Franklin Lakes, NJ, USA), 2 g of Bacto yeast extract (BD), 1 g of sodium glutamate, 3 g of trisodium citrate, 10 g of MgSO 4 Á7H 2 O, 1 g of CaCl 2 Á2H 2 O, 1 g of KCl, 200 g of NaCl, 0.36 mg of FeCl 2 Á4H 2 O and 0.36 mg of MnCl 2 Á4H 2 O per a litre of distilled water.

Materials and methods
The Haloarcula strains were incubated in the dark in 60 ml screw-capped test tubes containing 20 ml of JCM Medium 307 with shaking at 120 or 180 rpm at 20,25,30,35,40,45,50, 55 and 60°C. The culture experiment was performed in quadruplicate. The optical density at 660 nm (OD 660 ) of the cultures was monitored using a Spectronic 200 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) against sterilised medium that was incubated under the same condition as a reference. The growth rates were calculated from OD 660 values as described earlier (Sato et al. 2017). The actual T min , T opt and T max were determined based on the growth rate at each temperature.
DNA extraction, 16S rRNA gene sequencing and sequence analyses Cells of Haloarcula strains were inoculated into 60 ml screw-capped test tubes containing 20 ml of JCM Medium 307 and incubated in the dark with shaking at 120 or 180 rpm at 40°C for 100-150 h. The cells were then collected by centrifugation at 12,0009g for 3 min, and the pelleted cells were stored at -25°C until DNA extraction.
DNA extraction from the pelleted cells was performed using the described methods with modifications (Tchinda et al. 2016;Sharmin et al. 2018). Briefly, the cells of Haloarcula strains were lysed by lysozyme and proteinase K. Bulk DNAs were extracted using both phenol/chloroform/isoamyl alcohol and chloroform/isoamyl alcohol and purified with ethanol precipitation. Next, 16S rRNA gene fragments were amplified by PCR with KOD -Plus-DNA polymerase (Toyobo, Osaka, Japan) from the bulk DNAs using a primer set 8aF/1448hR (Table 1). The primer 1448hR, which is specific for 16S rRNA genes in Haloarcula strains, was designed using Genetyx-Mac ver. 19.0.1 software (Genetyx, Tokyo, Japan).
The PCR products were purified with a MicroSpin S-400 HR column (GE Healthcare, Little Chalfont, UK) or a QIAquick gel extraction kit (Qiagen, Hilden, Germany). The purified PCR products were cloned using a Zero Blunt TOPO PCR cloning kit (Life Technologies, Carlsbad, CA, USA) following the manufacturer's protocol. Insert DNA in each clone was sequenced by the dideoxy cycle-sequencing method using an Applied Biosystems 3730xl DNA analyzer (Life Technologies). We checked the sequences with the Bellerophon program (Huber et al. 2004) and no chimeric products were detected. We determined the current nearest relative of each 16S rRNA gene by using the BLAST program (Altschul et al. 1990) and the sequences of rrsA, rrsB and rrsC in each strain were identified.
We deposited the 16S rRNA gene sequences of the Haloarcula strains in the DDBJ/EMBL/GenBank database under Accession Nos. LC085248, LC085249 and LC198781 to LC198794 (Table S1).
We also performed sequence analyses of 16S rRNA genes in other halophilic archaea. We obtained the sequences derived from 233 species in 58 genera in class Halobacteria from the DDBJ database and manually searched halophilic archaea with high intragenomic heterogeneity (more than 1%) of 16S rRNA genes. The sequence differences and P GC offsets among copies of their 16S rRNA genes in the halophilic archaea were calculated using Genetyx-Mac ver. 19.0.1 (Genetyx).

Estimation of growth temperatures
The 16S rRNA genes of thermophilic and hyperthermophilic archaea have higher P GC values compared to those of psychrophilic and mesophilic archaea (Galtier and Lobry 1997;Khachane et al. 2005;Kimura et al. 2007). Based on the strong correlation between the P GC values of 16S rRNA genes and the growth temperatures of archaea, linear regression equations to estimate the T min , T opt and T max of archaea based on P GC of the 16S rRNA genes have been proposed (Kimura et al. 2013). We manually selected the internal sequences between the Archaea-specific primers 109aF and 915aR, used by Kimura et al. (2013), from the 16S rRNA gene sequences (Table 1). The P GC values of the internal sequences were calculated using Genetyx-Mac ver. 19.0.1 (Genetyx). The T min , T opt and T max of eight Haloarcula strains were estimated based on the P GC of rrsA, rrsB and rrsC using the linear regression equations as follows. We used SSU-align ver. 0.1.1 to predict secondary structures of 16S rRNAs transcribed from rrsA, rrsB and rrsC based on the sequences determined in this study (Narwocki 2009). In addition, we calculated the minimal folding free energy (MFE, DG in kcal mol -1 ) value of each 16S rRNA with mfold web server ver. 2.3 (http://unafold.rna.albany.edu/?q=mfold/RNA-Folding-Form2.3) to investigate the thermal stability of each 16S rRNA secondary structure (Zuker 2003). The nearly full-length 16S rRNA sequences (approx. 1405 bp) determined in this study were used for the  calculation. The sequence of rrsC in H. hispanica (Accession No. LC085247) contains Y (T or C) at position 392, and the nucleotide was C at the same position in rrsC obtained from the complete genome (Liu et al. 2011). We therefore calculated the MFE value of 16S rRNA transcribed from rrsC in H. hispanica by using the sequence modified from Y to C. The analyses using mfold web server were performed at the actual T min , T opt and T max of each strain. The significance of differences was determined by Student's t test between MFE values of 16S rRNAs transcribed from rrsA and rrsB or rrsC under same temperature condition. To check the selectivity of these primer sets, we performed a qPCR with diluted PCR products (1/100, 1/1000, 1/10,000 and 1/100,000) of rrsA, rrsB and rrsC (partial regions of 16S rRNA genes between the Archaea-specific primers, 109aF/915aR) amplified from the clone libraries constructed in this study. Each PCR mixture contained 2 ll of diluted PCR products (1/100, 1/1000, 1/10,000 or 1/100,000), 2 ll of each designed primer set (each 300 nM), 10 ll of PowerUP SYBR Green master mix (Life Technologies) and 4 ll of nuclease-free water. The qPCR was performed on an Applied Biosystems 7300 real time PCR system (Kataoka et al. 2018). The PCR cycling conditions included an initial step of 50°C for 2 min and 95°C for 2 min followed by 40 cycles of 95°C for 15 s, 58°C for 15 s and 72°C for 1 min. Electrophoresis of PCR products was performed to check primer selectivity and no contamination.
The quantification of rrsA and rrsBC contained in the cDNA samples was performed by the qPCR method described above. Each cDNA sample was tenfold diluted. The standard curves were prepared from diluted PCR products (1/100, 1/1000, 1/10,000 and 1/100,000) of rrsA, rrsB and rrsC (partial regions of 16S rRNA genes between the Archaea-specific primers 109aF/915aR) as described above. Each PCR reaction was performed in triplicate for technical replicates. Quantification of 16S rRNAs in H. japonica and H. argentinensis was performed in two batches for biological replicates. The significance of differences was calculated by Student's t test among the results under different temperature conditions.

Results and discussion
Actual growth temperatures of the Haloarcula strains Maximum growth rate and OD 660 value of Haloarcula strains at each temperature are shown in Figs. S1 and S2. The growth temperatures of the eight Haloarcula strains are summarised in Table 2. H. quadrata and ''H. californiae'' grew at 20-45°C with optimal growth at 40°C. This result is not consistent with an earlier study in which H. quadrata grew up to 55°C (Oren et al. 1999). H. argentinensis, H. vallismortis, H. amylolytica, H. japonica and H. hispanica grew from 20 to 50°C, with optimal growth at 45°C. The growth temperatures of Haloarcula sp. CBA1115 ranged from 20 to 55°C, and the T max is the highest growth temperature among the eight Haloarcula strains examined here.

Sequence analyses of 16S rRNA genes in the Haloarcula strains and other halophilic archaea
Almost full-length 16S rRNA genes (1405 bp) were amplified from bulk DNAs of Haloarcula strains, and clone libraries were constructed. We randomly selected and sequenced a total of 16 clones in each clone library. The BLAST analysis revealed that the 16S rRNA genes mostly matched with rrsA, rrsB and Estimated growth temperatures were calculated from P GC of rrsA, rrsB and rrsC using the microbial molecular thermometer (MMT) of Kimura et al. (2013) rrsC of the Haloarcula strains registered in the DDBJ/ EMBL/GenBank database (Table S1). While rrsB and rrsC show nearly identical sequences, the sequence of rrsA was 4.6-6.0% (65-85 bp) different from those of rrsBC in the eight Haloarcula strains. We also performed sequence analyses of 16S rRNA genes in other halophilic archaea and found 26 strains (representing 14 genera) with highly different sequences of 16S rRNA genes (more than 1.0%) (Table S2). Notably, five strains belonging to genera Halomicrobium and Halosimplex (5.8-8.7%) show higher sequence differences of 16S rRNA genes compared with Haloarcula strains (4.8-5.6%). On the other hand, P GC offsets among the 16S rRNA genes are less than 1.7% in these halophilic archaea compared with Haloarcula strains.
Estimation of the growth temperatures based on the P GC of the 16S rRNA genes We selected the internal sequences between the Archaea-specific primers 109aF and 915aR (795 bp) to estimate the growth temperatures of the Haloarcula strains, using the linear regression equations proposed previously (Kimura et al. 2013). The P GC of the internal sequences of rrsA ranged from 58.4 to 59.0% ( Table 2). The P GC of the internal sequences of rrsB and rrsC ranged from 56.1 to 57.0%. The minimum and maximum P GC offsets between rrsA and rrsB were 1.4% and 2.9% in H. vallismortis and Haloarcula sp. CBA1115, respectively.
The approximate growth temperatures were estimated based on the P GC values of the partial 16S rRNA gene sequences ( Table 2). The T opt estimated based on the P GC of rrsA in each Haloarcula strain ranged from 49.1 to 52.2°C, whereas the T opt estimated based on the P GC of rrsB or rrsC in each Haloarcula strain ranged from 37.8 to 42.2°C. Actual T opt were typically temperatures between the estimated T opt from the P GC of rrsA and rrsBC. The offsets between T opt estimated based on P GC of rrsA and rrsBC in Haloarcula strains were 6.9-14.4°C. These results indicate that harbouring high-and low-P GC 16S rRNA genes may enable Haloarcula strains to rapidly grow under a wide range of temperature conditions.

Prediction of the secondary structures of 16S rRNAs in the Haloarcula strains
We summarised different positions between secondary structures of 16S rRNAs predicted from rrsA and rrsB or rrsC in Haloarcula strains (Figs. 1, 2). The majority of differences (49-60%) were observed in helices 21, 22 and 26 of the central domain. A previous study showed that the affinity of ribosomal protein S8 to helix 21 in thermophiles and hyperthermophiles is much higher than that in mesophiles (Gruber et al. 2003). Furthermore, another study suggests that formation of helix 26 is essential for assembly of functional small subunit of ribosomes in Escherichia coli (Xu and Culver 2010). Therefore, these regions may be important in determining the functional differences among 16S rRNAs transcribed from rrsA and rrsBC.
The thermal stability of each 16S rRNA secondary structure was predicted with the use of the mfold web server (Table 3). The calculation results showed that the MFE values of the 16S rRNAs transcribed from rrsA were significantly lower than those of the 16S rRNAs transcribed from rrsB and rrsC under the same temperature conditions (p \ 0.05 by Student's t test). These results suggest that the thermal stability of the secondary structure of the 16S rRNAs transcribed from high-P GC rrsA is greater than those of the 16S rRNAs transcribed from low-P GC rrsB or rrsC.

Expression of rrsA and rrsBC under various temperature conditions
The selectivity of the primer sets used in this study was verified by qPCR using PCR product of rrsA, rrsB or rrsC amplified from the clone libraries constructed in this study (Fig. S3). Proper products were amplified from only rrsA in H. quadrata, rrsA in H. vallismortis and rrsA in the other Haloarcula strains using the primer sets, rrsAf2/rrsAr, rrsAf/rrsAr2 and rrsAf/ rrsAr, respectively. Proper products were amplified from only rrsB and rrsC in all Haloarcula strains using the primer set rrsBCf/rrsBCr. In addition, melting curve analyses showed that the melting temperature of the products from rrsA is about 1°C higher than those of products from rrsB or rrsC in all Haloarcula strains. These results suggest that all of the primer sets were well designed to precisely amplify rrsA or rrsBC. We quantified each 16S rRNA in Haloarcula strains by using RT-qPCR. No significant differences were found between biological replicates. The results of our RT-qPCR suggest that the expression levels of rrsA and rrsBC in seven of the Haloarcula strains varied depending on the incubation temperatures (Fig. 3). The expression ratio of rrsA to rrsBC (rrsA:rrsBC ratio) in particular increased with the cultivation temperature, and the maximum expression ratios were observed at each T max . We were able to divide the expression patterns in the Haloarcula strains into three types. First, the expression ratios continuously increased with temperature. This expression pattern was observed in H. amylolytica, H. japonica and H. hispanica (Fig. 3a-c). Although the expression ratio did not vary at temperatures close to T opt and T max , the ratio at T max was significantly higher than those below 40°C (p \ 0.05 by Student's t test). Second, the expression ratios increased suddenly at T max in Haloarcula sp. CBA1115, H. argentinensis, H. quadrata and H. vallismortis (Fig. 3d-g). This ratio at T max was significantly higher than those at other tested temperatures (p \ 0.05 by Student's t test). Third, the expression ratio stayed constant under all tested temperatures in ''H. californiae'' (Fig. 3h).
Although we observed differences in the expression levels of rrsA and rrsBC, all eight of the Haloarcula strains simultaneously expressed these 16S rRNA genes at all tested temperatures. The rrsA:rrsBC ratio increased with temperature in seven of the eight Haloarcula strains (Fig. 3). A previous study using H. marismortui also showed that the expression ratio of a high-P GC 16S rRNA gene compared to low-P GC 16S rRNA genes increased markedly at T max (50°C) (López-López et al. 2007). Our present findings extend those of the previous study and suggest that the expression ratios of high-P GC 16S rRNA genes to low-P GC 16S rRNA genes increase with temperature in most Haloarcula strains. The variation in expression ratios may occur due to temperature-dependent expression of 16S rRNA genes and/or thermal degradation of 16S rRNAs with low stability transcribed from rrsBC at high temperatures. Moreover, we observed that the expression ratios at 20°C and 25°C ranged from 0.28 to 0.59 in seven Haloarcula strains, which suggests that the total expression of rrsB and rrsC was 1.7 to 3.6 times as much as that of rrsA. Low-P GC 16S rRNA gene can be transcribed and replicated with low energy costs because of two reasons. First, costs for de novo production of adenine, thymine and uracil is lower than those of guanine and cytosine (Chen et al. 2016). Second, force for unwinding adenine-thymine rich region is lower than that for unwinding guanine-cytosine rich regions Essevaz-Roulet et al. 1997;Bockelmann et al. 1998;Rief et al. 1999). Therefore, low-P GC rrsB and rrsC in Haloarcula strains can help to reduce energy consumption in transcription and replication under low temperature conditions, which might explain why total expression of rrsB and rrsC increased at low temperatures close to the actual T min in seven of the eight Haloarcula strains. Although the expression mechanisms of these 16S rRNA genes are unclear, they can be verified in future studies.
High temperature adaptation of Haloarcula strains in their habitats Sato et al. (2017) constructed rRNA operon doublemutant strains of H. hispanica that harbour only rrnA including high-P GC rrsA (58.9%) or only rrnC including low-P GC rrsC (56.4%). Both mutant strains grew slower than the wild-type strain at low temperature close to the T min , whereas the mutant harbouring rrnA containing only high-P GC rrsA grew as fast as the wild-type strain at high temperature close to the T opt and T max . Their findings suggest that rRNA transcribed from the rRNA operon containing high-P GC rrsA is important for growth at high temperature in H. hispanica.
In the present study, the rrsA:rrsBC ratios increased with the incubation temperature in six Haloarcula strains as well as H. hispanica (Fig. 3). The results of the secondary structure prediction in silico indicate that 16S rRNAs transcribed from high-P GC rrsA in Haloarcula strains can form a more stable secondary structure at high and low temperatures compared to those transcribed from low-P GC rrsBC. In addition, only Haloarcula sp. CBA1115 harbouring the highest b Fig. 1   P GC (59.0%) of 16S rRNA gene grew at 55°C, the highest growth temperature among the Haloarcula strains reported previously (Bowers and Wiegel 2011;Namwong et al. 2011). These findings imply that 16S rRNAs transcribed from high-P GC rrsA are potentially important for growth at high temperatures close to actual T max (45-55°C) in the Haloarcula strains inhabiting solar salterns and salt lakes, where environmental temperature varies greatly in a day (Wieland et al. 2005;López-López et al. 2007;Sima et al. 2013;Andrade et al. 2015).

Conclusion
In this study, we performed the sequencing of 16S rRNA genes in eight Haloarcula strains, and we estimated the growth temperatures using MMT based on the P GC values of the 16S rRNA gene sequences. The estimated growth temperatures of cells carrying the high-P GC 16S rRNA gene (rrsA) were approximately 6-15°C higher than those carrying the low-P GC 16S rRNA genes (rrsB and rrsC). In addition, the prediction analysis via computer simulation revealed that the stability of the secondary structure of 16S rRNAs transcribed from high-P GC rrsA was higher than that of 16S rRNAs transcribed from low-P GC rrsBC in the Haloarcula strains examined in this study, suggesting that small ribosomal subunits containing those 16S rRNAs may function effectively under various temperature conditions. These results suggest that Haloarcula strains harbour different sequences of 16S rRNA genes to maintain rapid growth over a wide range of temperatures.
We characterised the expression of rrsA and rrsBC of H. hispanica at different growth temperatures. We found that the expression ratio of high-P GC rrsA to low-P GC rrsBC increased with the incubation temperature in seven of the eight Haloarcula strains. Our results indicate that the temperature-dependent expression of 16S rRNA genes is a common feature in most Haloarcula strains. The environmental b Fig. 2    californiae''. Error bars denote the standard deviation for triplicate measurements. The results for H. hispanica (c) are taken from Sato et al. (2017). Asterisks indicate significant differences among samples (Student's t test, p \ 0.05) temperature in salt lakes and solar salterns varies in a day, and Haloarcula strains may express and utilise different P GC of 16S rRNAs in response to variable temperature in their habitats.