RNA-Binding Protein Hfq: A Role in Cellulose Decomposition

Objective To study the function of the RNA-binding protein Hfq in Bacillus subtilis cellulose decomposition. Results In the medium with sodium carboxymethylcellulose (Na-CMC) as the sole carbon source, the knockout of Hfq resulted in a 38.0% ± 2.1% and 76.6% ± 7.1% decrease in cellulose hydrolysis ability and cellulase activity, respectively. The results of real-time quantitative PCR revealed that several cellulase genes (eglS, bglA, and bglC) were signicantly downregulated in the Hfq knockout strain. The isogenic ΔHfq complemented strain recovered the cellulose hydrolysis ability, cellulase activity, and expression level of cellulase genes. In addition, the survival of Hfq mutant in stationary phase was signicantly affected.


Introduction
The development of new eco-friendly bioethanol fuels is crucial to reducing the consumption of fossil fuels, but the bioconversion of cellulose into bioethanol remains a challenge (Zhao et al., 2020). Cellulose cleavage is catalyzed by a group of enzymes referred to as cellulases, including endoglucanase (EC  Sadhu and Maiti, 2013). All three enzymes have been used to produce bioethanol; however, the e ciency of cellulase in the existing industrial Bacillus subtilis strains remains at a low level (Gu et al., 2018;Zhao et al., 2020). Therefore, it is of great importance to deepen our understanding of the underlying regulatory mechanisms of B. subtilis cellulase and to develop new B. subtilis strains with more e cient cellulase for industrial applications.
Endoglucanase encoded by the eglS gene is the main cellulase in B. subtilis, and can be activated and transcribed by a σ A -dependent promoter (Robson and Chambliss, 1987). σ A (encoded by sigA), as well as σ B (encoded by sigB), are found in B. subtilis and are σ factors, which are components of the multiplesubunit bacterial RNA polymerase (RNAP) and play important roles in response to environmental stresses (Haldenwang, 1995). 6S RNA, an antagonist of σ A , can compete with a σ A -dependent promoter in binding with σ A and then produce the 6S RNA-σ A -RNAP complex, resulting in transcription inhibition (Trotochaud and Wassarman, 2005).
As the rst identi ed small non-coding RNA (sRNA) in bacteria, 6S RNA can regulate a variety of metabolic processes and other sRNAs (Bobrovskyy and Vanderpool, 2013;Hindley, 1967). Previous studies revealed that sRNAs often need the incorporation of the RNA-binding protein Hfq to function well (Updegrove et al., 2016;Vogel and Luisi, 2011). 6S RNA encoded by gene ssrS has been demonstrated to be a Hfq-associated sRNA in Escherichia coli, but the underlying mechanism and functional roles of this interaction remain unclear (Zhang et al., 2003). Considering that the functional and structural properties of Hfq and 6S RNA are highly conserved in most bacteria (Barrick et al., 2005;Sun et al., 2002;Vogel and Luisi, 2011), B. subtilis Hfq (Hfq BS ) may interact with 6S RNA. As 6S RNA and σ A are involved in the activation of cellulase genes such as eglS (Robson and Chambliss, 1987;Trotochaud and Wassarman, 2005), it is reasonable to hypothesize that Hfq may participate in the regulation of cellulose metabolism via 6S RNA.
To analyze the cellular roles of Hfq BS in more detail, we constructed an Hfq mutant (B. subtilis168ΔHfq) and its complemented strain (B. subtilis168ΔHfq-C). We studied the effects of Hfq BS on bacterial growth, cellulose hydrolysis ability, cellulase activity, and cellulase gene expression, which were used as the theoretical basis to determine whether Hfq BS acted as a key factor in cellulose stress adaptation. In addition, the potential relationship between Hfq BS , 6S RNA, sigA, and eglS was also explored.

Materials And Methods
Bacterial strains, plasmids, and growth conditions The bacterial strains and plasmids used in this study are listed in Table S1. Bacillus subtilis strains were grown at 37°C in LB medium, Na-CMC medium (70 mmol K 2 HPO 4 , 5 g Na-CMC, 30 mmol KH 2 PO 4 , 25 mmol (NH 4 ) 2 SO 4 , 0.5 mmol MgSO 4 , 0.5 mmol CaCl 2 , 10 μmol MnSO 4 , 5 µmol MnCl 2 , and 1000 mL of ddH 2 O at pH = 7.2), or on Congo red plates (2 g (NH 4 ) 2 SO 4 , 2 g Na-CMC,1 g KH 2 PO 4 , 0.5 g MgSO 4 ·7H 2 O, 0.5 g NaCl, 20 g agar, and 1000 mL of ddH 2 O at pH natural). Escherichia coli strains were cultured at 37°C in LB medium or on a Luria agar (LA) plate (LB with 1.5 % Agar). When required, the antibiotics tetracycline (10 μg/ml), ampicillin (100 μg/ml), and erythromycin (200 μg/mL) were added to the growth media. Hfq mutant and complemented strain construction The Hfq mutant was generated from B. subtilis 168 by allelic homologous recombination. Brie y, Hfq anking regions and erythromycin resistance genes (emr) were ampli ed by PCR using the primer pairs UP F/R, DOWN F/R, and EMR F/R (Table S2). The upstream, downstream, and emr PCR products were digested with SacI/SmaI, XbaI/SbfI, and SmaI/XbaI, respectively. The digested fragments emr, down, and up were then ordinal ligated into the clone vector pUC18 to obtain the recombinant plasmid pUC18-emr, pUC18-emr-down, and pUC18-up-emr-down, respectively. The up-emr-down fragment of 2556 bp was ampli ed using UP F and DOWN R as primers. The up-emr-down fragment was transformed into B. subtilis 168 by Spizizen transformation (Spizizen, 1958). The Hfq mutant, B. subtilisΔHfq, was obtained after homologous recombination and con rmed by PCR.
To complement the hfq mutant, a DNA fragment from B. subtilis 168 containing the entire Hfq gene was ampli ed. The ampli ed fragment was digested with SalI/SmaI enzyme and cloned into SalI/SmaIdigested pHYP43, which contains the P43 strong promoter and results in the pHYP43+Hfq plasmid. The complementing plasmid was transformed and one complement mutant strain, named B. subtilisΔHfq-C, was selected on the LA plate with tetracycline resistance. Determination of growth curve Pellets of B. subtilis, B. subtilisΔHfq, and B. subtilisΔHfq-C were grown overnight in LB medium. Then, they were sub-cultured on fresh LB and Na-CMC medium for 60 h independently according to the inoculation amount of initial OD 600 = 0. was carried out using the SYBR green RT-qPCR kit (TB Green ® Premix Ex Taq ™ II, Takara, Japan). In order to study the potential relationship between Hfq and cellulose stress adaptation of B. subtilis, 11 genes (eglS, bglP, bglA, bglC, licB, licC, licA, ccpA, sigA, bsrA, bsrB) closely related to cellulose decomposition were chosen for RT-qPCR (Table S3). The 16S rRNA gene was used as an endogenous control (Jagtap et al., 2016). The relative fold change in target gene expression was calculated using the 2 -ΔΔCT method (Livak and Schmittgen, 2001).

Statistical analysis
The growth curve, cellulose hydrolysis ability, cellulase activity, and the expression of cellulase-related genes were evaluated three times in each group to reduce error. Signi cant differences were determined by an alpha level = 0.05 using one-way analysis of variance (ANOVA) in IBM SPSS statistics software (version 20.0).

Results
Hfq knockout attenuates B. subtilis growth A mutant strain, B. subtilisΔHfq, was generated by homologous recombination and con rmed by PCR ( Figure S1). Wild type B. subtilis, mutant B. subtilisΔHfq and its isogenic complemented strain B. subtilisΔHfq-C were cultured in Luria-Bertani (LB) and Na-CMC medium separately to examine their growth curve. In LB medium, no signi cant difference was observed from 0 to 6 h; from 8 h to 60 h, the optical density at 600 nm (OD 600 ) of B. subtilisΔHfq decreased by 10.9% ± 3.2% on average compared with that of B. subtilis ( Figure 1A). In contrast, in Na-CMC medium, no signi cant difference was observed from 0 to 12 h; from 24 h to 60 h, the OD 600 of B. subtilisΔHfq decreased by 26.5% ± 5.2% on average compared with that of B. subtilis ( Figure 1B). Therefore, our experiments showed that the presence of Hfq gives B. subtilis a survival advantage in the stationary phase, whether in LB medium or Na-CMC medium.
Importance of Hfq for cellulose hydrolysis in B. subtilis All strains were cultured on Congo red plates for 48 h and then stained with Congo red. As shown in Figure 2 and Table S3, the knockout of Hfq reduced the cellulose hydrolysis circle of B. subtilis by an average of 38.0% ± 2.1%, while the isogenic ΔHfq complemented strain cured the lost phenotype. Therefore, our results showed that Hfq is necessary for B. subtilis to maintain e cient cellulose hydrolysis ability, and the existence of the protein contributes to the positive regulation of cellulose hydrolysis.
Positive regulatory effect of Hfq on cellulase activity of B. subtilis Bacillus subtilis lacks the gene to synthesize exoglucanase, so the cellulase activity measured in this experiment mainly refers to the total enzyme activity of endoglucanase and β-glucosidase (Accession number: NC_000964). As shown in Figure 3, regardless of whether the observations were noted at 12, 24, 36, 48, or 60 h, the cellulase activity of the Hfq mutant was signi cantly lower than that of the control group (B. subtilis), while the isogenic complemented strain of the Hfq mutant recovered part of the phenotype; the cellulase activity of the Hfq mutant decreased by an average of 76.6% ± 7.1%. In comparison, the complemented strain of the Hfq mutant recovered an additional 40% ± 12% of the cellulase enzyme activity. These results showed that Hfq has a positive regulatory effect on the cellulase activity of B. subtilis.

Hfq regulation of cellulase gene expression in B. subtilis
Tests of cellulose hydrolysis ability and cellulase activity have con rmed that the knockout of Hfq BS seriously hinders the decomposition and utilization of cellulose nutrients such as Na-CMC. To obtain more details of the protein's function in the regulatory network of cellulose catabolism, 11 genes related to cellulose decomposition were used for real-time quantitative PCR (RT-qPCR) analysis (Table S2). All 11 genes measured in our study showed signi cant differences, among which 6 genes (eglS, bglA, bglC, sigA, ccpA and bsrB) were downregulated and 5 genes (bglP, licB, licC, licA and bsrA) were upregulated in B. subtilisΔHfq (Figure 4). The main cellulase gene eglS encoded by B. subtilis was downregulated by 42.7% ± 10.9% in the Hfq mutant ( Figure 4A). In addition, after the knockout of Hfq, the β-glucanase genes bglA and bglC were downregulated by 40.4% ± 7.8% and 63.1% ± 7.7%, respectively ( Figure 4A). The isogenic ΔHfq complemented strain recovered the lost phenotype. Hence, our results con rmed that Hfq is a core regulatory factor that can positively regulate the expression of cellulase genes in B. subtilis. . Compared with those in LB medium, the growth defects caused by Hfq knockout are more intense in Na-CMC medium (Figure 1). One possible explanation is that Hfq is more important for the survival of B. subtilis in a nutrient-poor environment than in a nutrient-rich environment. In addition, the decrease in survival rate after 48 h may be related to the lack of nutrients.

Aid of Hfq in B. subtilis adaptation to cellulose stress
The positive regulation of cellulase activity by Hfq further explains the advantages given by this protein in cellulose hydrolysis ability and stationary phase survival under cellulose stress, since the cellulase activity of B. subtilis is closely related to the rate of obtaining nutrients regardless of being in Na-CMC medium or on Congo red plates (Figure 1-Figure 3 and Table S3). High-ber conditions constitute a stress environment for B. subtilis (Ziyao et al., 2015). The possible role of Hfq BS in stress adaptation has been emphasized by this study; nevertheless, it remains to be con rmed.
In B. subtilis, the two promoters that initiate Hfq transcription are regulated by σ B and σ H (encoded by . Therefore, some interactions may also occur between Hfq BS and 6S RNA. Previous studies have shown that 6S-1RNA is not expressed in the early exponential phase, and the expression of 6S-1 RNA is four-fold higher than that of 6S-2RNA from the late exponential to stationary phase (Beckmann et al., 2011). Similarly, our results also show that the knockout of Hfq BS mainly leads to an increase in the overall transcriptional level of 6S RNA in the stationary phase (among them, the transcriptional level of 6S-1 RNA is also nearly four-fold higher than that of 6S-2RNA) (Figure 4). The negative regulation of Hfq on the stability of 6S RNA may be due to the formation of the Hfq-6S RNA complex, which induces the degradation of 6S RNA by ribonuclease E (RNase E). It is well known that the binding of Hfq to some Hfq-associated sRNA causes the sRNA to be degraded (degradation by RNase E recruited by Hfq-sRNA complex) or protected, while the Hfq itself is unaffected and re-released into the cell (Vogel and Luisi, 2011;Zhang et al., 2003). Compared with 6S-2 RNA, 6S-1 RNA was more likely to be degraded by Hfq. In B. subtilisΔHfq, the level of 6S-1 RNA was nearly three-fold higher than that of wild strains, while 6S-2 RNA was downregulated by 84.9% ± 3.1% ( Figure 4). On the other hand, 6S-1 RNA is the ortholog of E. coli 6S RNA (Barrick et al., 2005). In addition, 6S-2 RNA may be protected by Hfq during the stationary phase, but the signi cance of this protection needs to be further studied.
Increase in the compensatory activity of the phosphotransferase system from Hfq knockout In B. subtilis, the combination of the complex formed by ccpA (catabolite control protein A) and P-Ser-HPr or P-Ser-Crh with the catabolite responsive elements (Cre) of the target operons can lead to carbon catabolite activation (CCA) or carbon catabolite repression (CCR) (Fujjta, 2009). The ccpA transcript levels also requires the assistance of a σ A -dependent promoter (Bingyao and Jörg, 2018). After the knockout of Hfq, the ccpA transcript levels decreased by 45.5. ± 3.1% ( Figure 4B). The downregulation of ccpA may be related to the decrease of functional σ A (Bingyao and Jörg, 2018). When the preferred carbon source is limited (such as Na-CMC medium), CCR usually does not exist (Fujjta, 2009). However, four phosphotransferase system (PTS) related genes (such as bglP, licA, licB, and licC) that are inhibited by ccpA (only in the presence of CCR) are upregulated ( Figure 4B) (Fujjta, 2009). In Hfq mutants, the transcript levels of the four genes were more than two-fold higher than those in the wild strains ( Figure  4B). The PTS is generally composed of non-speci c enzyme I (EI), histidine-containing phosphocarrier protein (HPr), and sugar-speci c enzyme II (EII), of which, the latter is mainly responsible for the transmembrane transport and phosphorylation of PTS-sugar (Fujjta, 2009). Therefore, the upregulation of these four PTS-EII genes might partly explains the residual cellulose decomposition and utilization ability after Hfq knockout because secondary metabolites produced by extracellular cellulase decomposition cellulose need to be transported through the PTS to enter the cell (Fujjta, 2009). Another important reason is that the expression of the cellulase gene itself can be directly induced by some small molecular soluble sugars produced by secondary metabolites, although the amount of cellulase produced by this induction effect is relatively small (Aro et al., 2001;Fujjta, 2009).
Despite these ndings, the regulatory network related to the downregulation of two β-glucosidase genes (bglA and bglC) in Hfq mutants is not clear. One possible hypothesis is that the decrease in the transcription level of β-glucosidase may be due to the weakening of CCA mediated by ccpA. For example, several enzymes (such as degS, degU, and serA) involved in carbohydrate metabolism have been con rmed to be positively regulated by CCA (Bingyao and Jörg, 2018). Identi cation of Cre sequences of bglA and bglC will help to further understand the regulatory network of these two genes.
The positive regulatory effect of Hfq BS on cellulose hydrolysis ability, cellulase activity, cellulase gene expression, and stationary phase survival was con rmed. Therefore, Hfq gives B. subtilis a survival advantage under cellulose stress. Additional targets and functional identi cation could further consolidate the important position of Hfq as a core regulator of multiple metabolic pathways in bacteria, and contribute to a better understanding of the overall function of the protein in bacteria.