Overexpression of mqsR in Xylella fastidiosa Leads to a Priming Effect of Cells to Copper Stress Tolerance

Copper-based compounds are widely used in agriculture as a chemical strategy to limit the spread of multiple plant diseases; however, the continuous use of this heavy metal has caused environmental damage as well as the development of copper-resistant strains. Thus, it is important to understand how the bacterial phytopathogens evolve to manage with this metal in the field. The MqsRA Toxin–Antitoxin system has been recently described for its function in biofilm formation and copper tolerance in Xylella fastidiosa, a plant-pathogen bacterium responsible for economic damage in several crops worldwide. Here we identified differentially regulated genes by X. fastidiosa MqsRA by assessing changes in global gene expression with and without copper. Results show that mqsR overexpression led to changes in the pattern of cell aggregation, culminating in a global phenotypic heterogeneity, indicative of persister cell formation. This phenotype was also observed in wild-type cells but only in the presence of copper. This suggests that MqsR regulates genes that alter cell behavior in order to prime them to respond to copper stress, which is supported by RNA-Seq analysis. To increase cellular tolerance, proteolysis and efflux pumps and regulator related to multidrug resistance are induced in the presence of copper, in an MqsR-independent response. In this study we show a network of genes modulated by MqsR that is associated with induction of persistence in X. fastidiosa. Persistence in plant-pathogenic bacteria is an important genetic tolerance mechanism still neglected for management of phytopathogens in agriculture, for which this work expands the current knowledge and opens new perspectives for studies aiming for a more efficient control in the field.


INTRODUCTION
Xylella fastidiosa is a phytopathogen with a broad host range that affects plants worldwide (Almeida et al., 2019). Plant diseases caused by this bacterium include citrus variegated chlorosis (CVC), Pierce's disease (PD) in grapevines, and the olive quick decline syndrome (OQDS), which constitute important threats for these crops (Almeida et al., 2019;Saponari et al., 2019;Coletta-Filho et al., 2020). Copper-based compounds are widely used in agriculture as a chemical strategy to limit the spread of multiple plant diseases (Lamichhane et al., 2018). Although X. fastidiosa is not itself controlled by copper spraying, biocomplexes containing copper, zinc, and citric acid have been used to control X. fastidiosa in olive groves (Girelli et al., 2019). Copper has an important contribution in crop protection; however, there are many issues related to the use of this heavy metal such as phytotoxicity, soil accumulation, negative effects on soil biota, and development of copper-resistant strains (Lamichhane et al., 2018). Thus, regarding plant-pathogen interaction, it is important to understand how the bacterial phytopathogens evolve to deal with this metal in the field.
In X. fastidiosa, the mqsRA toxin-antitoxin (TA) system type II is a genetic mechanism that has been associated with tolerance to copper stress (Muranaka et al., 2012). There are six types of TA system, which are distinct according to the action, nature, and mechanisms used by the antitoxins to neutralize the activities of the toxins (Page and Peti, 2016). Typically, in these systems, the toxin gene product is a protein and the antitoxin gene is a non-coding RNA (in types I and III) or a protein (in types II, IV, V, and VI) (Page and Peti, 2016;Harms et al., 2018). Bacterial toxin-antitoxin (TA) systems encode a stable toxin that disrupts cellular function and its labile cognate antitoxin in the same operon. The antitoxin neutralizes toxin activity under normal conditions, while proteases degrade the antitoxin under stress, allowing the toxin activity (Wang and Wood, 2011;Fisher et al., 2017). Moreover, the antitoxin usually regulates the expression of its own TA operon by binding to a palindromic sequence in the promoter region and repressing its transcription . TA systems have been shown to play a role in persistence, biofilm formation, cell movement, pathogenicity, DNA maintenance, and phage-defense Wen et al., 2014;Shidore and Triplett, 2017). In addition, they are highly expressed in persister cells and, thus, are generally responsible for the persistence phenotype (Wang and Wood, 2011;Fisher et al., 2017). A persister cell constitutes a tolerant cell (Lewis, 2010) originating from a population that displays antibiotic persistence, being a subpopulation phenomenon (sometimes referred to as heterotolerance) (Balaban et al., 2019), while, a tolerant cell is the capacity of an entire population of bacteria to survive a bactericidal antibiotic exposure (Balaban et al., 2019). Multidrug resistance in bacteria can occur by distinct ways like the accumulation of the resistance factors like plasmids or genes, each one encoding for resistance to a particular agent, and can or cannot occur along with the activity of multidrug efflux pumps (Nikaido, 2009). The mqsRA TA system was originally described in Escherichia coli and shown to be involved in biofilm and persister cell formation (Wang and Wood, 2011). The toxin mqsR was the most induced gene in E. coli persisters and the first TA system to reduce persister formation upon deletion, while increasing this phenotype after overexpression (Kim and Wood, 2010). It has been demonstrated that X. fastidiosa may form persister cells under copper stress (Muranaka et al., 2012;Merfa et al., 2016), representing an important survival strategy still unexplored in plant pathogenic bacteria (Martins et al., 2018). The MqsRA TA system is composed of the MqsR toxin, which is an endoribonuclease that degrades messenger RNA (mRNA) with GCU motifs and the MqsA antitoxin that binds and inactivates the toxin via its N-terminal domain (Brown et al., 2009;Yamaguchi et al., 2009;Lee et al., 2014). Due to its ability to selectively degrade mRNA, MqsR also acts as a global regulator . Thus, aiming to identify genes modulated by MqsR in X. fastidiosa, we overexpressed this toxin under the control of its native promoter and performed RNA-Seq when growing cells under normal and upon copper stress conditions.
Our results show that MqsR is a key gene regulator in the pathway tolerance of X. fastidiosa to copper stress, mediating several genes that prompt the cells to enter in a state that suggests the formation of persisters. In addition, copper induces MqsRindependent responses related to proteolysis and multidrug resistance through transcriptional regulator, transporters, and efflux pumps in order to increase the bacterial tolerance to this metal. This study presents unexplored mechanisms in phytopathogens that could have important impacts on how they can deal with agrochemicals and highlight the persistence phenomenon that could be occurring in the field.

Copper Sensitivity Assay
To evaluate the effects of copper on X. fastidiosa growth and formation of persisters, Xf -EV and Xf -mqsR cells were grown in PW broth (PWG without Gelzan) (Davis et al., 1981) and treated with 3 mM CuSO 4 .5H 2 O (Sigma). Control samples of both strains were grown in non-copper PW broth. Cells grown on solid PW were harvested from plates, resuspended in PBS buffer, and the optical density (OD 600 nm ) was adjusted to 0.3 and inoculated into PW broth to grow for another 7 days. The cells were then collected and the OD 600 nm was adjusted to 0.1. From each of these bacterial suspensions, 10-mL aliquots were inoculated into 90 mL fresh PW broth and incubated at 28 • C for 14 days at 150 rpm. Subsequently, Xf -EV and Xf -mqsR cells were exposed to 0 ("C-0" for Xf -EV, and "M-0" for Xf -mqsR) and 3 mM copper ("C-3" for Xf -EV and "M-3" for Xf -mqsR) for 24 h . Each treatment was performed in duplicates for each strain. The cells of each culture were collected, rinsed with DEPC water, and resuspended in 11 mL of PBS buffer. An aliquot of 1 mL from each suspension was used to determine colony formation units (CFU/mL), and to perform electron microscopy analysis, as described below. The cells of the remaining 10 mL were collected under the same conditions and stored at −80 • C for RNA extraction. Three independent biological replicates were performed.

Bacterial Growth Under Copper Stress
Aliquots of the entire experimental condition described above (Supplementary Figure 2) were collected to determine the CFU/mL of each biological experiment at the following time course: inoculation time (t0), 14 days after growth (stationary phase, Campanharo et al., 2003) in fresh PW broth when copper was added (t1) and 24 h after copper treatment (t2), completing 15 days of growth. From each sample, a 10fold serial dilution was performed and plated in PWG to estimate CFU. Four replicates were used for each sample, which were grown at 28 • C for 30 days. The measurements were performed in triplicates, and results were scored as the means ± standard deviation and compared using the Student's t-test (p ≤ 0.05).

Scanning Electron Microscopy
Scanning electron microscopy was performed under the experimental conditions described above (Supplementary Figure 2). Briefly, an aliquot of the planktonic and biofilm cells was sampled 24 h after copper addition for each X. fastidiosa strain. Controls without copper were also collected for both Xf -EV and Xf-mqsR. Samples were centrifuged and resuspended in RNA Later solution (Thermo Fisher Scientific, Waltham, MA, United States), frozen in liquid nitrogen and stored at −80 • C. For microscopy analysis, cells were thawed, centrifuged, fixed in 2.5% glutaraldehyde in 0.2 M sodium cacodylate buffer (v/v) and kept at 4 • C until use. Preparation of samples for visualization was done according to Kozlowska et al. (2014). Electron micrographs were captured with a magnification of 4000 × using a Hitachi TM 3000 scanning electron microscope (Hitachi, Tokyo, Japan). The Xf -EV and Xf-mqsR cells were measured using the ImageJ software (ImageJ, 2018) to determine the length and proportion of elongated and small cells at 100 cells per treatment. Only cells longer than 4.0 µm were considered elongated (Liu et al., 2014;Merfa et al., 2016), while only cells with a length smaller than 2.0 µm were considered small. The length of each cell in each treatment was analyzed through comparison of means by one-way analysis of variance (ANOVA) followed by Holm-Sidak multiple comparison test or Tukey's HSD test (p ≤ 0.05).

RNA Isolation and RNA-Seq
RNA-Seq reads were produced from 12 RNA samples: three from non-treated Xf -EV cells, three from non-treated Xf -mqsR, three from copper-treated cells of Xf -EV, and three from copper-treated Xf -mqsR cells. Total RNA was extracted using the hot phenol method (Khodursky et al., 2003), treated with DNase I RNase free (Qiagen, Hilden, Germany), purified using the RNeasy Plus Kit (Qiagen, Hilden, Germany) and eluted in 30 µL of RNase-free water. Concentrations were determined by spectrophotometry (NanoDrop 8000, Thermo Fisher Scientific). Ribo-Zero rRNA TM Removal Kit (Illumina, San Diego, CA, United States) was used for rRNA removal. The depleted RNA was precipitated using ethanol according to the manufacturer's instructions and resuspended in 10 µL of RNase-free water. Samples were quantified for the presence of rRNA using the 2100 Bioanalyzer system (Agilent Technologies, Santa Clara, CA, United States) at the Life Sciences Core Facility (LaCTAD). cDNA libraries were prepared using the Illumina TruSeq Stranded mRNA Library Prep Kit (Illumina). Sequencing was performed using the HiSeq High Output kit (Illumina) on a HiSeq 2500 system (Illumina), run with 2 × 100 bp pairedend reads.

RNA-Seq Data Analysis
The sequencing reads were analyzed in the FastQC program (Wingett and Andrews, 2018) and processed using Trimmomatic (Bolger et al., 2014) to remove adapters and extremities with poor quality. The reads were mapped to the genome of X. fastidiosa 9a5c (NCBI BioProject accession PRJNA271) using the STAR program (Dobin et al., 2013). From the mapped data, the gene-mapped reads were counted using the Subread package (Liao et al., 2019). Standardization and analysis of differential gene expression (p < 0.05) was performed using the EdgeR package (Robinson et al., 2010), computed using data from all three biological replicates. Differentially expressed genes obtained from EdgeR analyses were used for functional categorization by Blast2GO (Götz et al., 2008). Venny 2.1.0 (Oliveros, 2007) was used to show exclusive genes regulated in Xf-mqsR under copper stress.

Data Validation by Quantitative Real Time-PCR (RT-qPCR)
RNA samples were obtained from three other experiments using the same experimental condition as the RNA-Seq. A total of 250 ng of purified RNA from each condition was used as input for cDNA synthesis with the Reverse Transcription System kit (Promega, Madison, WI, United States). RT-qPCR was performed using the GoTaq qPCR Master Mix (Promega) in an ABI PRISM 7500 Sequence Detection System (Applied Biosystems, Foster City, CA, United States). Relative expression values were normalized to the X. fastidiosa 16S ribosomal RNA endogenous control (Merfa et al., 2016). Cycling parameters were performed according to the manufacturer's protocol. The relative expression quantification (RQ) was calculated as previously described (Livak and Schmittgen, 2001). The selected genes are based on the RNA-Seq analysis of M-0 and M-3 (Supplementary Table 2). Three independent biological replicates were used for data validation.

Palindrome Search
The motif 5'-ACC (N)7 GTT-3' , used as target sequence for DNA binding by the antitoxin MqsA, was searched in the genome of X. fastidiosa strain 9a5c using PATLOC (Mrazek and Xie, 2006), and also at the differentially expressed genes (DEG) data set herein generated.

MqsR Overexpression Changes X. fastidiosa Phenotype
To evaluate the effects of copper on wild-type X. fastidiosa and the mqsR-overexpressing strain, bacterial growth with and without copper was evaluated. In a previous work we verified that overexpression of MqsR increased the formation of persister cells under 3 mM of copper stress . Here, to access the phenotypic and genetic regulation mediated by MqsR, we used the same condition, where copper was added after 15 days of bacterial growth.
At the time of the inoculation (t0), and after 15 days of growth in fresh PW broth (t1), no significant difference in bacterial growth was observed between C-0 and M-0 (Figure 1). However, 24 h after addition of copper (t2), there was a significant reduction in population size of approximately 100fold between copper-treated samples (C-3 and M-3) and their respective untreated controls (C-0 and M-0) (Figure 1). We observed an approximately 10% increase in cell survival after copper treatment in populations overexpressing mqsR (M-3) in comparison to the control (C-3). However, the difference in CFU counts between M-3 and C-3 was not significant (F = 0.06, p = 0.11) (Figure 1).
To verify possible phenotypic changes in X. fastidiosa cells potentially caused by the overexpression of mqsR and copper treatment, samples from each experimental condition were used for scanning electron microscopy. Under normal growth condition, biofilm and planktonic cells in C-0 did not show any significant morphological change (Figures 2A,B). However, when copper was added (C-3) a reduction was observed in biofilm size (Figure 2C), and curiously, copper induced aggregation and elongated cells in the planktonic condition ( Figure 2D, red arrows).
On the other hand, X. fastidiosa overexpressing mqsR (M-0) presented more elongated cells even without copper stress (Figures 2E,F, red arrows), and at an even greater extent than C-3 ( Figure 2D, red arrows). In addition, X. fastidiosa overexpressing mqsR (M-0) displayed a phenotypic heterogeneity that can be demonstrated by the presence of a higher population of shorter cells when compared to the other treatments ( Figure 2F, blue arrows). Elongated cells were also observed in X. fastidiosa overexpressing mqsR in presence of copper (M-3) in both biofilm and planktonic conditions ( Figures 2G,H, red arrows).
Microscopy images for each condition (n = 100) were used for counting elongated and short cells in the planktonic fraction (Figures 2I,J). The results showed a good agreement with the visual observation, with a higher population of elongated cells in X. fastidiosa overexpressing mqsR (M-0) compared to C-0. Copper induced an increase of elongated cells in both populations (C-3 and M-3). Interestingly the number of elongated cells in M-0 is naturally even higher than C-3 ( Figure 2I). Similarly, higher percentages of short cells were observed in presence of copper (C-3 and M-3) or X. fastidiosa overexpressing mqsR (M-0) ( Figure 2J).
Overall, these results show that besides copper treatment, mqsR overexpression also led to changes in X. fastidiosa morphology and pattern of aggregation, culminating in a global phenotypic heterogeneity. Interestingly, heterogeneous phenotypes in single bacterial populations have been described as indicative of persister cells (Michiels et al., 2016;Fisher et al., 2017).

RNA-Seq Data
RNA-Seq reads were produced for C-0, C-3, M-0, and M-3 (Supplementary Figure 3). Raw sequencing reads were deposited under the NCBI Bio-Project ID PRJNA718853. Average posttrim read length ranged from 36 to 105 bp, the reads aligned to the genome of X. fastidiosa 9a5c. Variable rRNA and small RNA depletion efficiencies between samples resulted in 0-7.3% in library preparation. Mapped reads were used to determine transcript boundaries and normalized expression for all proteincoding genes by EdgeR (Supplementary Material 1, Data Sets 1-5). Pearson's correlation coefficient for protein-coding gene expression between experimental replicates ranged from 0.89 to 0.93. Highlighted DEGs of libraries were characterized according mqsR Differentially Modulates Global Gene Expression of X. fastidiosa To investigate global expression changes likely to be associated with the phenotypes described above, we performed RNA sequencing analysis. To identify which genes were modulated by MqsR under normal growth conditions, we assessed the pairwise comparison between M-0/C-0 libraries (without copper treatment). Amongst the DEGs, 189 genes showed upregulation by the overexpression of mqsR alone, while 164 genes were downregulated (Supplementary Material 1, Data Set 1; p < 0.05). RNA-Seq expression values (log 2 foldchange) were confirmed by RT-qPCR for 10 selected genes based on Table 1, with a Pearson correlation coefficient of 0.89 (Supplementary Figure 4). According to the data obtained through RNA-Seq (Supplementary Material 1, Data Set 1), the selected genes that are possibly modulated by MqsR are listed in Table 1.
Functional categorization of these 353 DEGs comprised genes associated with peptide metabolic process, transport, proteolysis, transcriptional regulation, and RNA metabolic processes (Figure  3, Supplementary Material 1, and Data Sets 6, 7). Genes associated with proteolysis were exclusively downregulated, including the proteases clpA and clpP. On the other hand, genes related to peptide metabolism were exclusively upregulated, including those related to ribosomal subunit scaffolding of RNA polymerase (RNAP), such as rpoA and rpoZ. These genes are also listed in the regulatory function category, together with mqsR, lysR, and the post-transcriptional regulator hfq, which were induced. The regulators genes mqsR, rpoZ, and lysR are related to bacterial survival, stress responses, and pathogenicity (Maddocks and Oyston, 2008;Santiago et al., 2015;Merfa et al., 2016;Weiss et al., 2017). rpoZ mutants of Mycobacterium smegmatis were deficient in motility and biofilm formation, consequently affecting the formation of extracellular matrix (Mathew and Chatterji, 2006). Besides, the overexpression of transcriptional regulator type LysR from X. fastidiosa in E. coli was described to a play role in maturation of biofilm during its development (Santiago et al., 2015). In X. fastidiosa, it is important to emphasize that the formation of biofilm is characterized as the main pathogenicity mechanism (Coletta-Filho et al., 2020). The transport category included the upregulation of tolC and acrB, both related to efflux pumps (Weston et al., 2018) and bacterial persistence (Pu et al., 2016). Efflux pumps are important for broad cellular homeostasis during stress responses. They export a wide variety of compounds, such as signaling molecules and antimicrobial compounds (Langevin and Dunlop, 2018). rlpA, a gene involved in cell division (Jorgenson et al., 2014;Berezuk et al., 2018), was downregulated, and fimD, which is involved with type I fimbrial adhesin (Meng et al., 2005), was upregulated. These genes are involved in bacterial physiology and biofilm formation, respectively. Furthermore, our results showed TA-related genes. The relE (XF_RS12805) toxin was downregulated by MqsR; this gene is associated with inhibition translation by cleavage of mRNA in the ribosome (Fiebig et al., 2010). Another toxin, parE gene, was upregulated and it is responsible for inhibiting gyrase and thereby blocks chromosome replication (Jiang et al., 2002). Modulation of relE and parE suggests that these bacterial cells maintain basal activities with reduced metabolism as shown in persister cells (Lewis, 2007). In addition, the repression of rlpA, inhibiting cell division and induction of the toxin encoding parE (Yuan et al., 2011), which inhibits bacterial division, could contribute with the observed elongated phenotype. Taken together, these observations suggest that overexpression of mqsR contributes to bacterial survival during stress response by activating pathogenicity regulators and inhibiting proteolysis and cell division.

Overexpression of mqsR Modulates Translation in X. fastidiosa Under Copper Stress
To identify the influence of copper on the gene expression, the following pairwise comparisons of the sequencing libraries were performed: i. M-3/M-0, and ii. C-3/C-0. Each pairwise comparison generated 417 and 662 DEGs, respectively. The M-3/M-0 comparison resulted in 238 upregulated and 179 downregulated genes (Supplementary Material 1, Data Set 2), while the C-3/C-0 analysis resulted in 335 upregulated and 327 downregulated genes (Supplementary Material 1, Data Set 3).
To verify X. fastidiosa genes modulated by MqsR in response to copper stress, a Venn diagram was used to compare the upand downregulated genes in the M-3/M-0 and C-3/C-0 libraries (Figure 4). This comparison provided genes modulated only by mqsR-overexpressing cells under copper stress (M-3/M-0), resulting exclusively in 111 upregulated and 84 downregulated genes (Figure 4 and Supplementary Material 1, Data Set 4).
In the search for genes oppositely modulated between Xf -EV (C-3/C-0) and Xf -mqsR (M-3/M-0), a set of nine genes were found to be downregulated in C-3/C-0 and upregulated in M-3/M-0. Of those, there is the yeiP elongation factor (XF_RS09585), three ribosomal subunits (XF_RS00715, XF_RS00285, XF_RS09575), the msrB (XF_RS03590) and yuxK (XF_RS04035), the aminotransferase astC (XF_RS06015, also known as argM or cstC), and two hypothetical proteins (XF_RS03860, XF_RS05645). Interestingly though, the ribosomal protein (XF_RS12125) was the only one showing an opposite behavior, being induced in C-3/C-0 but suppressed in M-3/M-0. This protein is involved in translation, and accordingly, this category was downregulated in M-3/M-0. In E. coli, it is known that persister cells have very low metabolism, with nongrowing cells as a result of a depletion in translation and, thus, in protein production capacity, cessation of transcription and reduction in ATP production Kim et al., 2018). Moreover, the ability to wake up from this persister state was related to ribosome content (Kim et al., 2018). Our results show categories such as peptide metabolic processes and translation downregulated, suggesting low-metabolism and depletion of protein production in X. fastidiosa in such condition.   Finally, 94 genes remained downregulated, and 118 genes were upregulated in both conditions. We believe that these sets of genes are modulated due to treatment with copper itself and are likely to be independent of MqsR functions (Supplementary  Material 1, Data Set 5).
Next, the functional characterization of the M-3/M-0 data set was performed to identify which genes are differentially modulated by the overexpression of MqsR under copper stress. RNA-Seq log 2 fold-change values were confirmed by RT-qPCR for 10 selected genes selected from Table 2 with a Pearson correlation coefficient of 0.93 (Supplementary Figure 4). Genes related to translation and peptide metabolic processes were exclusively repressed in the M-3/M-0 libraries, whereas proteolysis and drug metabolic processes were induced ( Figure 5A and Supplementary Material 1, Data Sets 8,  9). Other categories identified in this analysis included transport, regulatory functions, and RNA metabolic processes. Therefore, besides lowering the metabolism, the cells activate these specific salvage mechanisms allowing copper tolerance.
The proteolysis category included another peptidase S4 clpP (XF_RS02140) and tldD (XF_RS04775) metalloprotease, important regulators of bacterial metabolism. These genes are related to protein degradation. Although yet unclear, tldD was described as a putative regulator of chromosome-encoded TA system activities (Hu et al., 2012). Among the upregulated genes involved in transport, there were genes that encode ion transporters and sulfate transporters belonging to the ABC transporter family. ABC transporters are known to be involved in the influx or efflux of a wide diversity of molecules, and also with antimicrobial peptide resistance (Orelle et al., 2019). The categories associated with translation and peptide metabolic process showed downregulated genes encoding ribosomal subunits and the elongation factors EF-Tu and EF-G. Interestingly EF-Tu is described as the most enriched protein in X. fastidiosa outer membrane vesicles (OMVs) important for pathogen systemic dissemination throughout the host xylem vessels (Feitosa-Junior et al., 2019). The category linked to regulatory functions showed various downregulated genes, such as the global regulator fis, which is involved in virulence and pathogenicity. The rice pathogen Dickeya zeae showed remarkably decreased virulence capacity after fis deletion (Lv et al., 2018). This global virulence regulator is involved in exopolysaccharide production, motility, biofilm formation, and cellular aggregation in Dickeya zeae. All these processes are of utmost importance for X. fastidiosa pathogenicity, being associated with host colonization. The rpoA and a DNA-binding regulator hypothetical protein (XF_RS07050) were also repressed in the overexpressing strain under copper treatment.
To identify genes exclusively modulated by MqsR under copper stress, we analyzed the gene ontology of the 111 upregulated genes and 84 downregulated genes presented in Figures 4, 5B (Supplementary Material 1, Data Sets 4,  10, 11). The exclusively downregulated categories included translation and the peptide metabolic process (Figure 5B and Supplementary Material 1, Data Sets 10, 11). Other categories identified were transport, regulatory functions, and RNA metabolic process. The genes dksA and rpoZ from the regulatory functions group are transcriptional regulators associated with stress responses (Mathew and Chatterji, 2006;Wang et al., 2018) and were induced under copper stress. The highlighted genes that are modulated by MqsR under copper stress are listed in Table 2.
Considering all the above-mentioned results, we built a hypothetical model for the mqsR overexpression and its influence on the X. fastidiosa regulatory mechanisms under normal and copper-induced stress conditions (Figure 6).

DISCUSSION
Copper-containing compounds are among the most used chemicals in agriculture (Lamichhane et al., 2018). The antimicrobial effects of copper were previously attributed to stress-induced responses in many bacterial plant pathogens (Wright et al., 2019;Martins et al., 2021), including X. fastidiosa (Rodrigues et al., 2008;Muranaka et al., 2012;Merfa et al., 2016;Ge et al., 2021). In this bacterium, the MqsRA TA system was reported to play a key role when the pathogen is under copper stress. MqsRA is likely to function as an indicator for exogenous stressors through the induction of cell elongation, formation of structured biofilm aggregations, and reduction in cell movement . To better understand the roles the toxin MqsR may be playing over stress-induced responses in X. fastidiosa, we assessed the major phenotypic outcomes and the global transcriptional profile of the mqsR-overexpressing strain under copper-stress conditions through microscopy and RNA-Seq analysis.
The mqsR overexpression triggers genetic response where cells activate genes associated to stress adaptation (Figure 6A), from which many are conserved in the presence of copper ( Figure 6B). These characteristics suggest that increasing the amount of MqsR leads to a priming effect of cells to stresses that normally induce expression of mqsR, like copper. We observed an approximately 10% increase in cell survival after copper treatment in the population overexpressing mqsR and considering that, in stationary phase, only up to ∼1% of cells are persisters (Keren et al., 2004;Lewis, 2007), we can infer that a higher number of persisters were present under this condition. Therefore, our results demonstrate that the presence of the stressor is not needed to induce the genes and consequent cell morphology changes when mqsR is overexpressed. These morphologies include the elongated cell formation and population heterogeneity, indicative of persister cell activation (Michiels et al., 2016;Fisher et al., 2017), which is supported by tolC induction (Figure 6A), that was associated with E. coli persistence (Pu et al., 2016). Indeed, the presented results fit perfectly in the mathematical model in which systems that do not present bistability produce the hysteretic switch to the persistent state (Fasani and Savageau, 2013), represented in our condition by the overexpression of mqsR.
Besides, the overexpression of mqsR in E. coli exhibited cellular toxicity, resulting in increased persister cell formation . Taken together and with previous results (Muranaka et al., 2012;Merfa et al., 2016), the role of mqsR in X. fastidiosa seems to be similar to E. coli which involves the induction of persister cells.
It has been demonstrated that the MqsRA TA system in X. fastidiosa likely autoregulates its own expression to balance the toxin and antitoxin in the most beneficial ratio for the cells to oppose the stress . The mqsR overexpression itself presents a stress condition to the cell, thus to inactivate the toxin, the antitoxin MqsA should be produced to reach a T:A balance (Brown et al., 2013). Indeed, we observed an induction of mqsA under both conditions (Figures 6A,B). An upregulation of hfq in Xf -mqsR was observed in both conditions with and without copper stress. The hfq gene encodes an RNA chaperone that, among other regulatory functions, is related to the downregulation of proteases (Kim and Wood, 2010). It suggests that hfq is a key gene in the autoregulation of the MqsRA TA system, and we propose it could be one of the factors responsible for keeping the ideal T:A ratio in the cell by controlling the expression of proteases and consequently the cell morphologies observed in this work.
The genes modulated clpP, hfq, and clpB by MqsR in X. fastidiosa resemble those modulated by the same regulon in E. coli . These genes are involved in stress responses and contribute to toxicity and, consequently, to persister cell formation in E. coli (Kim and Wood, 2010;. Differences in the global transcriptional profile were also observed, suggesting a potential X. fastidiosa-exclusive mechanism. Among the exclusive genes modulated by MqsR only in X. fastidiosa are two gene regulators (XF_RS07050 and XF_RS07310). The regulator XF_RS07310 has the same type of HTH domain as the MqsA antitoxin, suggesting that it could also bind to promoter regions of target genes and modulate their expression. Some regulators related to bacterial survival and stress responses previously described in several bacteria were also modulated (Maddocks and Oyston, 2008;Santiago et al., 2015;Merfa et al., 2016;Weiss et al., 2017).
The MqsA antitoxin regulates the expression of mqsRA and other genes in E. coli by binding to palindromic sequences in their promoter regions and repressing their expression (Brown et al., 2009;Wang and Wood, 2011;Soo and Wood, 2013). The MqsA antitoxin encoded by X. fastidiosa has the same amino acid residues in its HTH domain responsible for DNA binding . Therefore, we searched for the MqsAlike palindromic sequence 5'-AAC (N)7 GTT in the genome of X. fastidiosa (Supplementary Table 3), seeking to identify those genes that were specifically differentially expressed in our RNA-Seq analyses. We investigated gene regulations in conditions where mqsRA expression is increased, such as under copper stress and mqsR overexpression. We verified 526 palindromic regions throughout the X. fastidiosa genome, with 77 corresponding to intergenic regions (Supplementary Table 3). Among the DEGs, a few showed the searched palindromic sequence in their intergenic regions (Supplementary Table 4). These genes included clpP, htpX, clpB, and rpfB, besides mqsR itself. According to data RNA-Seq, clpP and clpB genes remained downregulated, while mqsA expression remained upregulated, suggesting that MqsA may be regulating proteolysis under stress conditions in mqsR overexpression.
In our model, copper stress induces responses independent of MqsR involving protein degradation and multidrug resistance. When mqsR is overexpressed under copper stress, other clpP (XF_RS02140) and tldD encoding proteases were induced ( Figure 6B). Thus, the observed upregulation of proteases could contribute to the consequent upregulation of mqsRA. The regulator dksA, which plays an important role in the multidrug resistance in E. coli (Wang et al., 2018), shifted from downregulation in normal growth conditions to upregulation under copper stress, supporting its role of multidrug resistance. Copper also induces the expression of transporter genes associated with multidrug efflux pumps including cutC, which is specific for copper efflux (Rodrigues et al., 2008;Li et al., 2009). It has been shown that multidrug efflux pumps induce persistence, and persister cells combine active efflux with passive numbness to survive antibiotic attacks (Pu et al., 2016). This demonstrates the interplay between resistance and tolerance mechanisms, which are complementary and redundant bacterial strategies to survive under stress conditions (Lewis, 2007).
Overall, with the results herein presented, we were able to expand the knowledge on the genes and mechanisms associated with MqsR, as well as the function of the MqsRA TA system in X. fastidiosa. MqsR regulates genes that alter cell behavior in order to prime them to respond to environmental stress, which is related to induction of persistence. The persistence in plant-pathogenic bacteria is an important tolerance mechanism to this agrochemical which is still neglected in the management of agricultural diseases.

DATA AVAILABILITY STATEMENT
The datasets generated and analyzed for this study can be found under the NCBI Bio-Project ID PRJNA718853. Other data used in this study are available on request from the corresponding author.

AUTHOR CONTRIBUTIONS
AS and MT conceived and designed this research, provided reagents, analytical tools, and revised the manuscript. IC and PM conducted the experiments and analyzed the data. IC, MM, NT-S, MT, and AS wrote the manuscript. IC, MM, PM, MT, and AS contributed to the interpretation of the data and provided intellectual input. All authors read and approved the final manuscript.

ACKNOWLEDGMENTS
We thank the staff of the Life Sciences Core Facility (LaCTAD) from the State University of Campinas (UNICAMP) for the RNA-Seq (Genomics) analysis.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb.