Regulation of potassium dependent ATPase (kdp) operon of Deinococcus radiodurans

The genome of D. radiodurans harbors genes for structural and regulatory proteins of Kdp ATPase, in an operon pattern, on Mega plasmid 1. Organization of its two-component regulatory genes is unique. Here we demonstrate that both, the structural as well as regulatory components of the kdp operon of D. radiodurans are expressed quickly as the cells experience potassium limitation but are not expressed upon increase in osmolarity. The cognate DNA binding response regulator (RR) effects the expression of kdp operon during potassium deficiency through specific interaction with the kdp promoter. Deletion of the gene encoding RR protein renders the mutant D. radiodurans (ΔRR) unable to express kdp operon under potassium limitation. The ΔRR D. radiodurans displays no growth defect when grown on rich media or when exposed to oxidative or heat stress but shows reduced growth following gamma irradiation. The study elucidates the functional and regulatory aspects of the novel kdp operon of this extremophile, for the first time.


Introduction
Potassium (K + ) is essential for physiological functions such as regulation of intracellular pH, transmembrane electrical potential and turgor pressure in all living organisms [1]. K + homeostasis is critical for adaptation to several biotic or abiotic stresses in bacteria [1][2][3][4][5][6]. Bacteria accumulate 0.1-0.6 M K + intracellularly from trace amounts (0.1-10 mM) of this cation present in the environment [7] through a variety of low and high affinity K + transporters. In E. coli, low affinity K + transporters Trk and Kup are constitutively expressed and maintain physiological concentration of potassium in the cell [8] while high affinity ATP-dependent K + uptake system (K m 2 μM), is effected by an inducible KdpATPase expressed under K + limiting conditions (active below 5 mM K + ) [5]. Ktr, a fourth constitutively expressed K + uptake transporter, is absent in E. coli but encoded by several other microbes [9].
The homologs of kdp operon and its two-component signaling system are present in Grampositive as well as Gram-negative bacteria and archaeal species [10] wherein kdp operon expression is stimulated by low K + concentration and high osmolarity. In addition, its expression has also been coupled with regulation of virulence genes and pathogenesis in Staphylococcus aureus, Salmonella typhimurium, Yersinia pestis, mycobacteria etc. [11,12] Expression of kdp operon is also associated with high NaCl tolerance in S. aureus [12], phosphate limitation in E. coli [13], or drought stress in Anabaena 7120 [14]. Thus, kdp operon forms a part of basic metabolism and also contributes to survival under a variety of stressful conditions. Gram-positive D. radiodurans exhibits extreme resistance to gamma radiation as well as to desiccation [15], owing to its customized DNA damage repair system, enzymatic/non-enzymatic antioxidants and metabolites. A number of elegant studies reported differential expression/levels of transcriptome, small RNAs, proteome, antioxidants, metabolites etc. during the phase of DNA repair [16][17][18][19][20][21]. K + is necessary for the stability of replication process and maintaining genome integrity in cells [22,23], both the functions actively carried out by D. radiodurans while recovering from DNA damage. Thus, role of K + homeostasis in the stress resistances of D. radiodurans needs detailed exploration.
The genome of D. radiodurans shows presence of two K + uptake systems, namely, Ktr and K + -transporting ATPase (Kdp). The organization of kdp genes in D. radiodurans is distinct from well studied kdp operon in E. coli, especially in the two component regulatory genes ( Fig  1). In D. radiodurans, structural genes kdpBAC (DR_B0083, DR_B0086, DR_B0087, respectively) were present in one operon along with a naturally truncated regulatory component kdpD-N (representing N-terminal half of the E. coli kdpD, DR_B0088) (Fig 1). The remaining regulatory components, sensor kinase (SK, representing C-terminal half of the E. coli kdpD, DR_B0082) and the response regulator (RR, DR_B0081) were present in a separate operon, in an opposite orientation, immediately upstream of the kdpBACD-N operon (Fig 1) [24]. Homologs of such truncated kdpD genes are found in the genomes of Deinococcus-Thermus group and cyanobacteria (Fig 1), that share close evolutionary relationship [25]. In cyanobacteria, that harbor 2 separate kdp operons in yet another distinct organization of regulatory components (Fig 1), only one kdp operon responds to potassium limitation, high osmolarity or desiccation [14,[26][27][28]. However, the role and regulation of kdp operon in the normal physiology and stress responses of D. radiodurans remains completely unexplored.
In this study, we report that the kdp operon of D. radiodurans is induced following K + limitation but does not respond to high osmolarity, unlike E. coli Kdp ATPase. We also demonstrate that regulation of kdp operon is effected through the cognate response regulator (RR) protein. The knockout mutant of RR (ΔRR) displayed kdp expression null phenotype even under potassium limiting conditions. In rich media, the growth of ΔRR mutant was similar to wild type D. radiodurans. Survival response of ΔRR mutant to other relevant stresses was also evaluated.

Stress conditions
D. radiodurans cells or its mutant was grown overnight in TGY medium. The cells were washed twice in either TGY, 20 mM K + phosphate supplemented minimal medium (hereafter referred to as K20 medium) or 20 mM sodium phosphate supplemented minimal medium (hereafter referred to as K0 medium) and the washed cells were inoculated in TGY, K20 medium or K0 medium, respectively, at an initial cell density of OD 600 = 0.5/ml. The cultures were incubated at 32˚C with 150 rpm agitation. The cells resuspended in K0 medium experienced K + limitation while the cells resuspended in K20 medium or TGY served as controls. For ionic or osmotic stress, K1 medium (1:20 dilution of K20 medium in K0 medium) was supplemented with either 0.1M NaCl or 0.2M sucrose. For gamma irradiation experiments, overnight grown cells of wild type or mutant D. radiodurans cells were resuspended in fresh TGY medium (OD 600 = 3.0), 2 μl aliquots of the cell suspensions and their serial dilutions were spotted onto the TGY plate. The plate was exposed to 5 kGy gamma irradiation (Gamma Cell 5000, Bhabha Atomic Research Centre, dose rate: 1.85 kGy/hr). Following stress, the plate was incubated at 32˚C for 24 h for recovery before the result was scored. For heat and oxidative stresses, the overnight grown cells of wild type or mutant D. radiodurans cells were resuspended in fresh TGY medium (OD 600 = 0.5) and exposed to either 42˚C or 100 mM H 2 O 2 for 1h with 150 RPM agitation. Further, the cells were concentrated (OD 600 = 3.0) and 2 μl aliquots of the cell suspensions and their serial dilutions were spotted onto the TGY plate. The plate was incubated at 32˚C for 24 h for recovery before the result was scored.
This study

This study
Primers for sequencing kdpB-548-F_Seq Cloning of DR_B0083 (complete or partial) and DR_B0081 ORFs, and over-expression and purification of corresponding proteins Genomic DNA of D. radiodurans was prepared as per the protocol reported previously [34]. The DR_B0083 (kdpB), DR_B0083/826-1740bp (skdpB), DR_B0081 (RR) ORFs were PCR amplified from the genomic DNA of D. radiodurans using gene-specific primers listed in Table 1. The PCR products were restriction digested using NdeI/XhoI and cloned into pET-29b plasmid vector at the identical sites. Correct clone was ascertained by DNA sequencing. The resultant constructs pET-kdpB, pET-skdpB or pET-RR were transformed in E. coli BL21 (DE3)pLysS and expression of corresponding proteins was induced in auto-induction medium at 20˚C. The cells were harvested after 15h of growth and the over-expressed proteins (sKdpB and RR) were purified by affinity chromatography using Ni-nitrilotriacetic acid (Ni +2 -NTA)agarose resin (Qiagen, Germany), as per the manufacturer's protocol. The purified RR protein was used for promoter interaction studies. The sKdpB and RR proteins were purified by gel elution method and used to generate polyclonal antibodies in rabbit, at a commercial facility of Bangalore Genei (India).

Deletion of DR_B0081 (RR) gene and confirmation of the knockout mutant
The ΔDR_B0081 mutant was constructed as per the procedure described previously [35]. In brief, mutagenesis strategy involved complete replacement of DR_B0081 ORF with an aph cassette that conferred kanamycin resistance. For this purpose, the upstream (RR-up) and downstream (RR-dn) sequences (540 bp each) of DR_B0081 were PCR amplified using primer pairs RR-up-F/R and RR-dn-F/R, respectively ( Table 1). The PCR amplified RR-up and RR-dn DNA fragments were restriction digested with KpnI/EcoRV and EcoRV/BamHI, respectively and were ligated to the KpnI/BamHI sites of pBlueSkript vector to obtain plasmid pΔRR1. To obtain aph cassette, plasmid pUC4K was restriction digested with HincII to release the 1252 bp DNA fragment harboring aph cassette. The aph cassette was cloned into the EcoRV site of pΔRR1 plasmid, by blunt end ligation. The resultant construct pΔRR2, with aph cassette inserted between the RR-up and RR-dn sequences, was used to transform WT D. radiodurans. Kanamycin resistant transformants were selected on TGY plates containing 10μg/ml kanamycin. The ΔRR knockout mutant was confirmed by PCR and western blot analysis of RR protein.

Immuno-detection of KdpB and RR proteins
The cellular extracts were prepared by sonication (Branson Digital Sonifier, Model 250). Cell debris was removed by centrifugation (10000 rpm for 10 min). The supernatant was centrifuged at 100000g for 1h. The supernatant was collected as cytosolic fraction. The pellet was washed 2 times and then collected as membrane fraction. The cellular proteins extracts, cytosolic or membrane fractions from stressed or unstressed cells were resolved by 12% SDS-PAGE and electro-blotted onto nitrocellulose membrane (Amersham Biosciences, India). The presence of KdpB or RR proteins was probed with primary polyclonal anti-D. radiodurans-sKdpB or anti-D. radiodurans-RR antibodies (1:10,000 dilution), respectively, followed by goat anti-rabbit IgG tagged to alkaline phosphatase (Sigma, 1:5000 dilution). The western blots were developed using NBT-BCIP solution (Roche, Germany). The immuno-detected protein bands which appeared on the membranes were quantified using CLIQS 1D PRO (Total Labs, UK).

Electrophoretic mobility shift assay (EMSA)
The promoter region (200 bp, DNA sequence immediately upstream of DR_B0083 i.e. kdpB gene) of kdp operon was PCR amplified using specific primers (Table 1). Alternatively, the intergenic region (38 bp) between DR_B0082 (sensor kinase) and DR_B0083 (kdpB) was generated in vitro by two primer (P38-F/ P38-R) annealing (  [28]. The DNAprotein complexes were resolved on 10% native polyacrylamide gel at 50 V in 0.5X TBE gel running buffer. The resolved DNA-protein complexes were electro-blotted onto nylon membrane for 30 mins. The nylon membrane was cross-linked with UV, probed with anti-DIG antibody and subsequently developed using colorimetric substrate, NBT-BCIP. Each experiment was repeated three times. The bands which appeared on membranes were quantified using gel quant software (Biochem lab solutions) and the K D value for RR protein was obtained by fitting the data to Hill's equation [36]. Specificity of RR interaction was confirmed by titration of DNA-protein complexes with unlabeled target-specific or non-specific DNA. Further, to determine specific interaction of RR protein with the kdp promoter, EMSA was carried out either by replacing kdp promoter fragment with non-specific DNA or by replacing RR with BSA protein. Non-specific DNA was a DNA sequence (153) bp from DR_0906 gene amplified using primers listed in Table 1.

Bioinformatic analyses
BLASTP [37] and ClustalW [38] tools were used to analyze sequence similarities and for phylogenetic analyses of the proteins encoded by kdp operon of D. radiodurans. Presence of E. coli like gene promoter elements in the upstream sequence of kdp operon was analyzed using BPROM software [39]. Phobius software was used to predict presence of a transmembrane domain in KdpB protein [40].

Results and discussion
The kdp operon of D. radiodurans responds to K + limitation To ascertain if the kdp operon of D. radiodurans responds to K + limitation, we cloned the genes for soluble domain of KdpB (sKdpB, amino acids 277-573 of KdpB protein) and response regulator protein, and over-expressed and purified the corresponding proteins and raised polyclonal antibodies (S1 Fig). The genes for full length kdpB and sensor kinase were also cloned but expression of corresponding proteins could not be achieved in E. coli despite several attempts. We probed the induction of KdpB as well as RR proteins in response to K + limitation. The KdpB as well as RR proteins were undetectable in D. radiodurans cells grown in TGY medium or K20 medium, but the cells incubated in K0 medium showed prominent induction of both KdpB and RR proteins (Fig 2A and 2B). The result shows that the kdp operon of D. radiodurans responds to K + limitation. The induction of RR protein post K + limitation is intriguing since in Anabaena L-31, the basal level of RR remains unaffected by K + limitation [28], while in E. coli as well as in Clostridium acetobutylicum the expression of KdpE increases as a result of synthesis of a polycistronic mRNA encompassing kdpABCDE genes and a read-through effect [41,42]. Further, we checked the kinetics of gene expression and observed that both the KdpB and RR proteins were synthesized within 30 minutes of transfer of D. radiodurans cells from TGY medium to K0 medium (Fig 2C and 2D). The response to K + limitation was further assessed at different concentrations of K + . Both KdpB as well as RR proteins were induced in K0 medium but not in K1 medium (1 mM K + phosphate added to K0 medium) (Fig 2E and  2F), indicating that unlike E. coli, the kdp operon of D. radiodurans could be induced only below 1 mM K + concentration. Similar results were observed for Anabaena L-31 earlier [28]. While most of the KdpB protein localized in membrane (Fig 2G), the RR protein was entirely localized in the membranes (Fig 2H). RR is expected to be a cytosolic protein as it does not possess membrane anchoring or spanning domains. RR protein, when over-expressed in heterologous E. coli, localized entirely in the cytoplasm. In D. radiodurans, DNA is present in the nucleoid [43] and about 70% of the DNA is associated with the membranes [44]. The localization of RR protein in the membrane may be due to its possible interaction with kdp operon signaling proteins (KdpD-N and SK) localized in the membrane and its functionality can be facilitated by association of DNA with the membranes.

The kdp operon of D. radiodurans does not respond to ionic or osmotic stresses
In addition to response to K + limitation, the kdp operon is also induced by high osmolarity in bacteria [45][46][47]. In E. coli K12, the expression of kdp operon was about 10-fold higher in response to K + limitation as compared to its expression in response to increased osmolarity by 0.4M NaCl, which in turn is about 10-fold higher compared to its expression in response to 0.6M sucrose stress [48]. Expression of kdp operon was originally proposed to be in response to changes in turgor pressure, but was but later debated [49][50][51]. Here we show that the kdp operon of D. radiodurans does respond to K + limitation but does not respond at all to increased osmolarity. We changed the osmolarity of the K1 medium either by adding 0.1M NaCl or 0.2M sucrose and checked induction of KdpB and RR proteins. Neither KdpB nor RR protein were induced upon addition of NaCl or sucrose (Fig 2I and 2J). The kdp operon of Anabaena L-31, signaling components of which are similar to those of kdp operon of D. radiodurans, does respond to NaCl stress but does not respond to sucrose [27], although, kdp operon of another cyanobacterium Anabaena torulosa does respond to sucrose [26]. In contrast, the kdp operon of D. radiodurans appears to be highly specific for K + limitation alone, and free from the influences of osmolarity.

RR protein induces kdp operon under K + limitation
Comparison of the RR protein encoded by DR_B0081 gene of D. radiodurans revealed 23.9% identity with KdpE of E. coli. BLAST analysis also revealed homology of RR protein with similar two-component response regulators from Deinococcus-Thermus group and cyanobacteria. Therefore, it was pertinent to ascertain if the RR protein was indeed the response regulator of the kdp operon of D. radiodurans. We constructed a knockout mutant of RR by replacing the DR_B0081 gene with aph gene conferring kanamycin resistance, by site directed mutagenesis, as per the strategy outlined in Fig 3A. The knockout mutant was confirmed by PCR analysis. When RR-up-F/RR-dn-R primer pair was used, 2.33 kb and 1.71 kb bands were amplified radiodurans cells grown either in K20 or K0 media, or exposed to ionic (-I, 0.1M NaCl) or osmotic (-O, 0.2M sucrose) stresses in K1 medium. The cellular proteins (100 μg) were resolved by 12% SDS-PAGE, electroblotted onto nitrocellulose membrane and immuno-stained using anti-KdpB or anti-RR antibodies as detailed in materials and methods. The top most protein band (~125 kDa) in the corresponding Coomassie stained gel is shown below Fig 2A, 2B, 2C, 2D, 2E, 2F, 2I and 2J, as loading control. For membrane or cytosolic protein extracts, 63 kDa protein band or 44 kDa protein band, respectively, are shown as loading controls (See S2 Fig for details on loading controls). Red bold numbers below the KdpB or RR immuno-stained bands indicate fold increase in their levels over the lanes in which these bands were not observed (denoted by 1).
from the genomic DNA of the prospective RR knockout mutant and wild type D. radiodurans, respectively, thereby confirming that the 0.63 kb DR_B0081 gene had been replaced by 1.25 kb aph cassette (Fig 3A and 3B). The mutant thus obtained failed to induce KdpB upon K + limitation. KdpB as well as RR proteins were induced in wild type D. radiodurans under similar culture conditions (Fig 3C). The data confirmed that the RR protein encoded by DR_B0081 gene was indeed responsible for expression of Deinococcal kdp operon under potassium limitation.  Regulation of kdp operon in D. radiodurans RR protein binds to an intergenic DNA sequence between structural and signaling genes of kdp operon [T n ]-rich region present in the kdp promoter region in E. coli has been shown to bind response regulator KdpE [52,53]. A similar [T n ]-rich region is present in the intergenic region (38 bp) between the kdpB and SK genes of D. radiodurans and in the upstream sequences of kdp operons in other bacteria as well [53,54] (Fig 4A). Thus, the 38 bp intergenic region harbouring [T n ]-rich sequence is a potential site for regulation of kdp operon. This short sequence did not possess any sequence elements essential for gene expression as analysed by BPROM software. However, several of the Deinococcal promoters are devoid of the typical E. coli type sequence elements but are still expressed constitutively [55]. Here, we chose a short (38 bp) as well as longer sequence (200 bp) for our study. We determined the binding of RR protein to the promoter sequence of kdpBACD operon (PkdpB-38/ PkdpB-200). Retardation in the mobility of PkdpB-200 incubated with increasing concentrations of RR protein confirmed binding of RR to PkdpB-200 (Fig 4B). We obtained a similar profile of retardation in the mobility of PkdpB-200 incubated with phosphorylated RR protein (S3 Fig). The apparent equilibrium dissociation constant (K D ) was calculated to be 3.62 ± 0.56 μM, which signifies the amount of protein required for 50% binding to PkdpB-200. The Hill coefficient value was determined to be 1.36 ± 0.17. Thus, RR showed very low or no co-operativity in binding (Fig 4C). Similarly, the mobility of PkdpB-38 was also retarded in the presence of increasing concentrations of RR protein (S4 Fig), confirming that this small intergenic region indeed provided the site for RR binding. To probe specific interaction of RR protein with PkdpB-200, competitive EMSA was carried out by titrating the DIG-labelled PkdpB-RR complexes with unlabeled PkdpB-200. With increasing concentration of unlabelled PkdpB-200, the labelled PkdpB-200 DNA was released from the PkdpB-200-RR complexes, where the labelled free DNA was clearly visible following competition with increasing amounts of unlabeled specific DNA (Fig 4D). Non-specific interactions of RR with nsDNA or of PkdpB-200 with any protein were ruled out by either incubating RR protein with PkdpB-200 or with nsDNA, or incubating BSA with PkdpB-200 or with nsDNA. DNA-protein complexes were observed only when PkdpB-200 was incubated with RR ( Fig 4E), thereby confirming the specific interaction of RR with PkdpB-200.

Relevance of kdp operon to stress resistance of D. radiodurans
Since kdp operon is reported to give survival advantage under stressful conditions in both Gram-positive as well as Gram-negative bacteria [4][5][6][7][8][9][10][11][12], we checked if presence of inducible kdp operon offered any advantage to D. radiodurans during stress. We used a kdp expression null ΔRR mutant to evaluate survival fitness following different stresses. The ΔRR mutant did not show any growth defect, as compared to wild type D. radiodurans, when grown in TGY medium under standard growth conditions (Fig 5A), or when exposed to 100 mM H 2 O 2 ( Fig  5B) or 42˚C heat stress (Fig 5C). However, survival of ΔRR mutant was reduced, as compared to wild type D. radiodurans, following 5 kGy gamma irradiation (Fig 5D).
The study, for the first time, reports functional characterization of the kdp operon of the radioresistant extremophile D. radiodurans, having a distinct operon organization. Unique features of the kdp operon of D. radiodurans include (a) induction of both, the structural and regulatory components, of the kdp operon under extreme potassium limitation, (b) lack of stimulation by increased osmolarity either by NaCl or sucrose, (c) expression of kdp operon through specific interaction of unphosphorylated RR protein with kdp promoter, and (d) confers a minor survival advantage to the organism while recovering from gamma radiation stress.  under standard growth conditions, (B) exposed to 100 mM H 2 O 2 , (C) exposed to 42˚C or (D) exposed to 5 kGy gamma irradiation.