Potassium ion (K⁺) is a ubiquitous monovalent cation that is essential for all living cells. It plays important roles in a wide variety of cellular functions such as turgor pressure maintenance, osmo-regulation, acid-base homeostasis, membrane potential adjustment, enzyme activation, and gene expression regulation (Booth, 1985; Epstein, 1986, 2003; Dinnbier et al., 1988; Stumpe et al., 1996; Follmann et al., 2009). Generally, a higher intracellular concentration of K⁺ (150–500 mM) is required than that of extracellular environment (Epstein and Schultz, 1965; Dinnbier et al., 1988) to efficiently carry out these functions. Hence, cell has to precisely regulate the uptake of K⁺ for its growth and survival in various hostile conditions (Stumpe et al., 1996; Kuo et al., 2005). For this purpose, animal cells are usually equipped with Na⁺/K⁺ ATP pumps that are generally missing in bacterial cells (Heitzmann and Warth, 2008; Gumz et al., 2015), whereas several K⁺ channels and active transporters are present in the prokaryotic cells to maintain potassium homeostasis. These bacterial K⁺ uptake systems vary a lot in their structures and affinities for the cation, which ultimately help bacteria to precisely maintain intracellular K⁺ concentrations under diverse environmental conditions (Altendorf et al., 1992, 1998; Epstein, 2003; Nanatani et al., 2015). In prokaryotes, three major families of K⁺ uptake transporters, Trk/Ktr/HKT, Kup/HAK/KT, and Kdp, have been discovered (Buurman et al., 1995; Ballal et al., 2002, 2007; Berry et al., 2003; Sato et al., 2014). Besides these, some K⁺ channels have also been reported (Epstein, 2003). Among these K⁺ uptake transporters, Trk and Kup are constitutively expressed in Escherichia coli, with the uptake rate of Trk system nearly 10 times higher than that of Kup transporter (Rhoads et al., 1976). On the contrary, the Kdp-ATPase is an inducible system, exhibiting high affinity for K⁺ uptake that is specific to bacterial and archeal kingdoms (Diskowski et al., 2015). This system becomes operational under low K⁺ concentrations (~2 μM as estimated from the K_m-value; Epstein, 2016) when other K⁺ uptake systems such as Trk and Kup fail to function (Epstein, 1992). Generally, Kdp-ATPase complex is consisted of four subunits, encoded in the kdpFABC operon. Among these subunits, KdpF is required for stabilizing the complex (Altendorf et al., 1998; Gassel et al., 1999), KdpA is involved in the binding and translocation of K⁺ (Altendorf et al., 1992; Greie, 2011), while KdpB is associated with KdpC that is essential for the ATP hydrolysis (Haupt et al., 2005; Greie and Altendorf, 2007). In most cases, kdpFABC is found adjacent to kdpDE, which encodes proteins that are involved in the regulation of Kdp-ATPase system. The inducible Kdp-ATPase is regulated by histidine kinase KdpD and response regulator KdpE, which constitute a typical two-component system (TCS) (Walderhaug et al., 1992), with KdpD responding to a multitude of signals such as low K⁺ concentrations, osmotic pressure upshift, different intracellular ATP levels, and pH of the medium (Asha and Gowrishankar, 1993; Jung and Altendorf, 2002; Hamann et al., 2008; Epstein, 2016). Previously, it was suggested that turgor pressure was another stimulus for kdpDE induction (Laimins et al., 1981), however recent observations did not support this idea (Epstein, 2016). Other than its role in K⁺ homeostasis, KdpD/KdpE TCS is also well-reported for its role in the regulation of bacterial virulence in diverse bacterial species including Mycobacterium tuberculosis, Staphylococcus aureus, Salmonella enterica serovar Typhimurium and Pseudomonas aeruginosa (Parish et al., 2003; Alegado et al., 2011; Xue et al., 2011; Feinbaum et al., 2012; Freeman et al., 2013). Moreover, as a regulator, KdpE was found to regulate many essential genes involved in key metabolisms, stress tolerance and virulence (Njoroge et al., 2013; Burda et al., 2014; Samir et al., 2016; Parker et al., 2017).

The Kdp-ATPase and its cognate TCS (KdpD/KdpE) have been well-characterized in E. coli (Walderhaug et al., 1992; Gassel et al., 1998; Brandon et al., 2000; Heermann and Jung, 2010; Laermann et al., 2013; Epstein, 2016). Cytoplasmic stimuli such as K⁺ concentration, ionic strength and ATP content are perceived by histidine kinase KdpD (Jung and Altendorf, 2002; Heermann and Jung, 2010). After detecting these stimuli, KdpD experiences auto-phosphorylation, with its phosphoryl group transferred to the transcriptional regulator KdpE (Voelkner et al., 1993). The phosphorylated KdpE (KdpE~P) binds to a 23-bp T-rich sequence in the promoter of kdpFABC with stronger affinity as compared to un-phosphorylated KdpE (Nakashima et al., 1992, 1993; Sugiura et al., 1992). Moreover, some accessory proteins such as UspC (Heermann et al., 2009) and enzyme IIA^Ntr (Lüettmann et al., 2009) have also been reported for their interactions with KdpD for signal transduction and cross talk. Among pathogenic bacteria, KdpD/KdpE TCS has been well-studied in S. aureus (Xue et al., 2011). The kdpFABC genes were significantly up-regulated (up to 100-fold) under salt stress in this osmo-tolerant bacterial species (Price-Whelan et al., 2013). In S. aureus, KdpE also regulates the expression of many other genes. Among these are the capsular polysaccharide genes (cap) that play a vital role in the pathogenicity of S. aureus. It was found that KdpE specifically bound to the cap promoter region and deletion of kdpDE resulted in the decreased transcription of cap genes (Zhao et al., 2010). Expression of other virulence factors such as spa, hla, aur, geh, and hlgB was also regulated by KdpE, and direct binding of KdpE to the promoter regions of many of these genes, has been observed (Xue et al., 2011).

Based on sequence homology to E. coli, Kdp-ATPase was found to be widely distributed among different bacterial and archeal genera (Ballal et al., 2007; Heermann and Jung, 2010). In case of genus Mycobacterium, the Kdp system has been found in M. tuberculosis, M. avium, M. bovis, M. smegmatis, M. marinum, and others, except M. leprae and M. ulcerans (Cholo et al., 2008; Bretl et al., 2011). Among these species, M. smegmatis is a fast growing and comparatively safe bacterial species to work with, which is now commonly used as a surrogate for M. tuberculosis. Generally, M. smegmatis is perceived as a non-pathogenic species. However, some investigations have demonstrated its pathogenic potential (Wallace et al., 1988; Alvarado-Esquivel et al., 2009; Best and Best, 2009; Driks et al., 2011; Xu et al., 2013). Up to now, KdpD/KdpE TCS in M. smegmatis has not been well-characterized. Due to the different organizations of kdpFABC and kdpDE in various mycobacterial species, investigation of this system in M. smegmatis would be valuable. As KdpD/KdpE is involved in regulating K⁺ homeostasis, elucidation of its detailed function and regulatory mechanism for its downstream target genes would provide better options to cope with mycobacterial infections. Moreover, due to the structural differences between the K⁺ uptake systems of prokaryotes and eukaryotes (Cholo et al., 2008), the current system may present a potential target for antibiotic development against mycobacterial infection. In the current study, the promoter of kdpFABC was clearly identified and the regulation of kdpFABC expression by KdpD/KdpE TCS under different potential stimuli was investigated in detail. To the best of our knowledge, for the first time this study provided strong evidences that KdpE bound to the promoter region of kdpFABC operon (PkdpF) at a specific site in M. smegmatis. In addition, we also predicted the coding sequence of KdpF in M. smegmatis as well as in many other mycobacterial species.

K⁺ plays many physiological roles in bacteria and its sufficient supply is essential for the pathogenicity of various species including M. tuberculosis (Salina et al., 2014). Upon K⁺ deficiency, bacterial cells of M. tuberculosis entered into a dormant non-culturable state but could be revitalized to culturable state once the availability of K⁺ was restored. Moreover, those non-culturable cells acquired resistance to the cell wall-targeting antimicrobials. Hence this resistance dormancy may raise the probability of latent tuberculosis (Salina et al., 2014). Regarding the important roles of K⁺ in mycobacterial physiology and pathogenicity, elucidation about K⁺ uptake and its regulatory mechanism would enable us to control mycobacterial infections in a more efficient way.

The species of genus Mycobacterium relies mainly on the Trk and Kdp systems for K⁺ uptake (Cholo et al., 2008). The Kdp-ATPase is regulated by KdpD/KdpE TCS, which was found to be functional in M. tuberculosis (Agrawal and Saini, 2014) and the bacterial strain became hyper-virulent upon its deletion (Parish et al., 2003). However, noticeable variations in the genetic organization of kdpFABC and kdpDE operons were observed within the genus Mycobacterium (Steyn et al., 2003; Cholo et al., 2008; Agrawal and Saini, 2014) (Figure S8). Some of the pathogenic species of genus Mycobacterium such as M. leprae and M. ulcerans are even deprived of Kdp-ATPase system (Cholo et al., 2008). Since different species of the genus Mycobacterium pose serious threats to a wide variety of organisms, the mechanism of K⁺ uptake within this genus needs to be elucidated in detail in order to cope with their pathogenic potential. To date, regulation of Kdp-ATPase has not been thoroughly investigated in M. smegmatis, a model species that has been used as a surrogate for pathogenic mycobacterial species. This manuscript focused on the regulation of kdpFABC operon by KdpD/KdpE TCS and, for the first time, proposed a detailed regulatory mechanism in M. smegmatis (Figure 10). These findings shall provide a foundation to further explore the mechanism of K⁺ uptake by Kdp-ATPase in other pathogenic species of genus Mycobacterium.

In M. smegmatis, it was found that the expression of kdpFABC operon was regulated by a KdpD/KdpE TCS that was induced under K⁺ limiting condition. Similar phenomenon was also reported in other bacteria (Altendorf et al., 1992; Frymier et al., 1997; Treuner-Lange et al., 1997). The KdpE binding sequence in PkdpF was found to be well-conserved in various mycobacterial species (Figure 4). However, no such sequence similarity could be detected in M. tuberculosis. Our results support the previous study that attempted to reveal binding of KdpE~P to the 192-bp intergenic region between kdpFABC and kdpDE in M. tuberculosis, although no such binding was observed (Agrawal and Saini, 2014). Contrary to M. tuberculosis, the direct binding of KdpE in PkdpF was confidentially observed in M. smegmatis in the present study.

Initially, candidate PkdpF (upstream of MSMEG_5391) was used for EMSA and DNase-I foot printing studies, considering MSMEG_5391 as a part of the kdp operon. However, the newly identified TSS was the fourth nucleotide of MSMEG_5391 hence, we believed that the original annotation of ORF (MSMEG_5391) might be inaccurate since the first three nucleotides of this frame were not the part of mRNA. The downward sequence of MSMEG_5391 was therefore searched for another potential ORF of kdpF gene. The sequence similarity of this newly identified kdpF of M. smegmatis with those of other species in genus Mycobacterium was noteworthy, although in many mycobacterial species, kdpF is still un-annotated. Our findings provide a strong foundation to identify and characterize kdpF in other species of genus Mycobacterium. Previously, KdpF was reported as a regulatory peptide (Gannoun-Zaki et al., 2013, 2014), hence its role under stress conditions needs further investigations in M. smegmatis.

The sigF loci in M. smegmatis encoded an alternative stress-response sigma factor F (SigF) (Gebhard et al., 2008). Previously, SigF-dependent promoters were first identified in M. tuberculosis (Rodrigue et al., 2007), and also characterized in M. smegmatis (Gebhard et al., 2008) in later studies. A GTTT-N_(15−17)-GGGTA motif was predicted as SigF-dependent promoter consensus after inspecting promoter sequences of numerous SigF-regulated genes through bioinformatics analysis by the same group (Hümpel et al., 2010). In our study, we found that the −10 and −35 consensus sequences of PkdpF were highly similar to numerous SigF-regulated sequences so far identified, but with a smaller N numbers (GTTC-N₍₁₃₎-GGGCA). In the previous study, PkdpF was not identified as a SigF-regulated promoter (Hümpel et al., 2010). This difference might be due to two reasons. First, the previous conclusion was based on different growth phases, whereas extremely low K⁺ concentration or high salt stress are required for the activation of kdpFABC transcription. Since their microarray study was performed under a non-inducible condition, kdpFABC might not be expressed at all. Second, different criteria were adopted in choosing spacing between the −10 and −35 regions of the SigF-dependent promoter consensus.

Ionic strength has been reported as stimulus for the induction of Kdp system. For instance, the Kdp system was found to be stimulated at high Na⁺ and NH4+ concentrations in E. coli (Jung et al., 2001; Hamann et al., 2008; Epstein, 2016). Similar to these findings, Kdp-ATPase of M. smegmatis was also induced under high concentrations of Na⁺ and NH4+ salts. Results of our RT-qPCR data further verified the induction of Kdp system under Na⁺ and NH4+ salts stresses (Figure S9). In the case of M. tuberculosis, bacterial systems related to K⁺ uptake were induced at acidic pH (Cholo et al., 2015). Similarly, in E. coli, the expression of kdpFABC also altered due to the change in the medium pH (Asha and Gowrishankar, 1993). However, we found that the Kdp-ATPase system of M. smegmatis was not induced by medium pH change.

In our study on M. smegmatis, a single transcript of kdpFABCDE was observed under K⁺ limiting condition. In E. coli, however, it was proposed that some long transcripts were made to contain transcripts of both operons but most of the transcripts stopped before the start of kdpD (Altendorf et al., 1992). The single transcript (kdpZYABCXDE) in K⁺ limiting condition was reported in Clostridium acetobutylicum (Treuner-Lange et al., 1997) and (kdpABC-DE) in Acinetobacter baumannii (Samir et al., 2016). In case of M. smegmatis, it could be proposed that under low K⁺ concentration, both operons (kdpFABC and kdpDE) were co-transcribed at least to some extent from PkdpF, whereas under normal K⁺ condition, kdpDE was transcribed to very basal level and this transcription was initiated from PkdpD.

Kdp-ATPase and its regulatory KdpD/KdpE TCS have been reported in many bacterial species including E. coli, S. aureus, C. acetobutylicum, M. tuberculosis, S. typhimurium, Leptospira interrogans, Alicyclobacillus acidocaldarius, and A. baumannii for its role in K⁺ uptake (Ballal et al., 2007; Xue et al., 2011; Matsunaga and Coutinho, 2012; Samir et al., 2016). In bacterial species such as E. coli, L. interrogans, and A. baumannii, KdpE was highly expressed in order to positively regulate expression of kdpFABC operon under low K⁺ concentrations (Polarek et al., 1992; Matsunaga and Coutinho, 2012; Samir et al., 2016) except in S. aureus where KdpE negatively regulated kdpFABC expression (Xue et al., 2011). In M. tuberculosis, KdpD/KdpE TCS was found to be functional (Agrawal and Saini, 2014) and it was reported that KdpD interacted with two membrane lipoproteins (LprF and LprJ) to modulate expression of kdpFABC controlled by KdpE under low K⁺ concentrations (Steyn et al., 2003). At present, KdpE binding sequence in the PkdpF has only been identified in four bacterial species including E. coli (Sugiura et al., 1992), C. acetobutylicum (Behrens and Duerre, 2000), S. aureus (Xue et al., 2011), and M. smegmatis (current study). The C. acetobutylicum KdpE binding motif shared significant sequence similarity with that of E. coli (23 bp T-rich). However, M. smegmatis KdpE binding motif (22 bp A-rich) did not show noticeable sequence similarity with any other previously reported KdpE binding motif (Figure 4B).

In M. smegmatis, bioinformatics study revealed another constitutively expressed system for K⁺ uptake, namely the Trk system comprised TrkA (MSMEG_2771) and TrkB (MSMEG_2769) subunits. In addition, TrkA domain-containing protein (MSMEG_3665) and ion channel membrane protein (MSMEG_1945) were found as possible K⁺ channels. Expression of these genes was down-regulated in K⁺ limiting (0 mM K⁺) condition when compared to K⁺ non-limiting conditions (normal K⁺) in wild type and mutant strains (Figure S10). These results indicated that under K⁺ limiting condition, the Kdp-ATPase appeared to be functional to scavenge K⁺ as compared to other K⁺ systems.

It can be concluded that in M. smegmatis, Kdp-ATPase plays a vital role in the bacterial physiology during K⁺ limitation. The system is positively regulated by KdpD/KdpE TCS and triggered for osmo-regulation during low K⁺ and salt stress. The regulation of kdpFABC operon via KdpD/KdpE TCS appeared to be dependent upon SigF. Based on growth discrepancies of ΔkdpD and ΔkdpE, it can be proposed that some other histidine kinases may also activate KdpE, which in return regulates genes other than kdpFABC. However, a detailed study is required to verify this hypothesis.

For the first time, DNA sequence required for successful binding of KdpE with PkdpF was identified in genus Mycobacterium. Based on sequence similarity of the DNA motif, the binding of KdpE protein to positively regulate Kdp-ATPase can be expected in other mycobacterial species such as M. avium, M. fortuitum, M. marinum, and M. chubuense. On the other hand, a different mechanism might be speculated for the regulation of Kdp-ATPase in M. tuberculosis and M. bovis, which lack this KdpE-binding DNA motif and disclose different genomic organization for the kdpFABC and kdpDE operons.

In the present manuscript, we also observed a comparatively long 5′-UTR region in the transcript of kdpFABC for M. smegmatis. It would be worthy of further studies to see whether this 5′-UTR region exhibits some kind of regulatory role for the translation of Kdp-ATPase transporter.

MKA performed most of the experiments and made most of the data evaluation. XLi, QT, XLiu, FC, JX and MA participated in partial experiments and interpretation of the data. MKA, QT, and JH conceived the study and drafted the manuscript. JH, SC, and MA revised the manuscript. All authors read and approved the final manuscript.