Inhibition of Bacterial RNase P RNA by Phenothiazine Derivatives

There is a need to identify novel scaffolds and targets to develop new antibiotics. Methylene blue is a phenothiazine derivative, and it has been shown to possess anti-malarial and anti-trypanosomal activities. Here, we show that different phenothiazine derivatives and pyronine G inhibited the activities of three structurally different bacterial RNase P RNAs (RPRs), including that from Mycobacterium tuberculosis, with Ki values in the lower μM range. Interestingly, three antipsychotic phenothiazines (chlorpromazine, thioridazine, and trifluoperazine), which are known to have antibacterial activities, also inhibited the activity of bacterial RPRs, albeit with higher Ki values than methylene blue. Phenothiazines also affected lead(II)-induced cleavage of bacterial RPR and inhibited yeast tRNAPhe, indicating binding of these drugs to functionally important regions. Collectively, our findings provide the first experimental data showing that long, noncoding RNAs could be targeted by different phenothiazine derivatives.


Introduction
Ribonuclease P (RNase P) is an essential and ubiquitous endoribonuclease responsible for generating tRNAs with matured 5 ends. In Bacteria, RNase P is composed of an approximately 400 nucleotides-long RNA (RPR) and one small basic protein, while in Archaea and Eukarya several proteins associate with the sole RNA. Irrespective of origin, the catalytic moiety of RNase P is the RPR, as evident from its ability to mediate cleavage of pre-tRNA as well as various other natural RNAs including mRNA and artificial substrates without the protein cofactor(s) [1][2][3][4]. Comparison of the bacterial and human RPR reveals structural differences in regions of functional importance. For example, the human RPR lacks the region in bacterial RPR that interacts with the 3 end of pre-tRNAs; differences also exist in the region that constitutes the binding site for the pre-tRNA T-stem/loop in bacterial RPR [5]. In addition, 10 proteins bind to human RPR compared to 1 in bacteria [6]. Together, these findings suggest that bacterial RNase P is a promising target for new anti-infective agents [7][8][9][10][11][12][13].
The activity of bacterial RNase P is inhibited by various aminoglycosides, aminoglycoside-arginine conjugates, and puromycin [10,[14][15][16][17]. Biochemical and structural data suggest that aminoglycosides can compete with binding of functionally important Me(II)-ion(s) to RNA [15,17,18]. In addition, acitretin and aminoglycosides inhibit the activity of the Dictyostelium discoideum RNase P holoenzyme [19,20]. Thus, we were motivated to search for other small molecules that interfere with bacterial RNase P function. In this study, we focused on bacterial RPR activity without protein since we were interested in identifying inhibitors of RNA function.
The phenothiazine methylene blue was first synthesized and described in the late 1800s. Given its properties, it was used as both a stain and a drug for the treatment of bacterial infections and malaria. Its biochemical properties have also been instrumental in developing protocols for microscopic detection and identification of microorganisms such as Mycobacterium tuberculosis, the causative agent of tuberculosis [21]. Moreover, staining procedures involving methylene blue have been used to stain RNAs in tissues and for visualization of RNA following separation by gel electrophoresis and membrane blotting [22,23].
Here, we provide data showing that various phenothiazine derivatives such as toluidine blue O and chlorpromazine, an antipsychotic drug with antibacterial activity, inhibit the activity of three structurally different bacterial RPRs, including the M. tuberculosis variant, as well as lead(II)-induced cleavage of yeast tRNA Phe .

Results
To investigate whether phenothiazine derivatives inhibit the activity of bacterial RPR we followed the protocol described by Mikkelsen et al. ([15], see Materials and Methods) and monitored cleavage of Escherichia coli (Eco) pre-tRNA Tyr Su3 (pSu3) and pre-tRNA Ser Su1 (pSu1; Figure 1). Bacterial RPRs have been classified into type A (ancestral) and type B (Bacillus type) based on their secondary structures ( [24]; Figure 2). Hence, we decided to use Eco RPR wt (M1 RNA; type A) and also Eco RPR G235 , which carries a change that alters the structure in the vicinity of the binding site for the pre-tRNA T-stem/loop region ( [25]; see below), and Mycoplasma hyopneumoniae (Hyo) RPR (type B; Figure 2; [26]). In addition, we included the type A M. tuberculosis (Myc) RPR because of its medical importance and the alarming increase in the appearance of multidrug-resistant strains ( Figure 2; [27,28]). The structures of the different phenothiazine derivatives and similar compounds that we tested are depicted in Figure 3. The phenothiazine methylene blue was first synthesized and described in the late 1800s. Given its properties, it was used as both a stain and a drug for the treatment of bacterial infections and malaria. Its biochemical properties have also been instrumental in developing protocols for microscopic detection and identification of microorganisms such as Mycobacterium tuberculosis, the causative agent of tuberculosis [21]. Moreover, staining procedures involving methylene blue have been used to stain RNAs in tissues and for visualization of RNA following separation by gel electrophoresis and membrane blotting [22,23].
Here, we provide data showing that various phenothiazine derivatives such as toluidine blue O and chlorpromazine, an antipsychotic drug with antibacterial activity, inhibit the activity of three structurally different bacterial RPRs, including the M. tuberculosis variant, as well as lead(II)-induced cleavage of yeast tRNA Phe .

Results
To investigate whether phenothiazine derivatives inhibit the activity of bacterial RPR we followed the protocol described by Mikkelsen et al. ([15], see Materials and Methods) and monitored cleavage of Escherichia coli (Eco) pre-tRNA Tyr Su3 (pSu3) and pre-tRNA Ser Su1 (pSu1; Figure 1). Bacterial RPRs have been classified into type A (ancestral) and type B (Bacillus type) based on their secondary structures ( [24]; Figure 2). Hence, we decided to use Eco RPRwt (M1 RNA; type A) and also Eco RPRG235, which carries a change that alters the structure in the vicinity of the binding site for the pre-tRNA T-stem/loop region ( [25]; see below), and Mycoplasma hyopneumoniae (Hyo) RPR (type B; Figure 2; [26]). In addition, we included the type A M. tuberculosis (Myc) RPR because of its medical importance and the alarming increase in the appearance of multidrug-resistant strains ( Figure 2; [27,28]). The structures of the different phenothiazine derivatives and similar compounds that we tested are depicted in Figure 3.    [29]. (A) Eco RPR (type A), the dotted line demarcates the S-and C-domains. Roman numerals and grey arrows mark the Pb 2+ -induced cleavage sites I to V. Residue C235 highlighted in black was mutated to G to generate Eco RPRG235. The Eco CP RPRwt lacks the S-domain and constitutes only the C-domain, with A79 to G88 forming a loop, marked with a grey line/arrow; for details see Kikovska et al. [30] and Wu et al. [31]; (B) M. hyopneumoniae RPR (type B) [26]. The Ia Pb 2+ -induced cleavage site is marked with a grey arrow. As in A, the dotted line demarcates the two major RPR domains; (C) P15 RNA. The grey arrow marks the Pb 2+ -induced cleavage at sites III to V.  [29]. (A) Eco RPR (type A), the dotted line demarcates the S-and C-domains. Roman numerals and grey arrows mark the Pb 2+ -induced cleavage sites I to V. Residue C 235 highlighted in black was mutated to G to generate Eco RPR G235 . The Eco CP RPR wt lacks the S-domain and constitutes only the C-domain, with A 79 to G 88 forming a loop, marked with a grey line/arrow; for details see Kikovska et al. [30] and Wu et al. [31]; (B) M. hyopneumoniae RPR (type B) [26]. The Ia Pb 2+ -induced cleavage site is marked with a grey arrow. As in A, the dotted line demarcates the two major RPR domains; (C) P15 RNA. The grey arrow marks the Pb 2+ -induced cleavage at sites III to V. Figure 3. Chemical structures of the small molecules tested in this study. The structure marked with a box represents the phenothiazine scaffold structure, with the ring atoms numbered.

Inhibition of RNase P RNA Cleavage by Phenothiazine Derivatives
Increasing concentrations of the selected phenothiazine derivatives inhibited the activity of all three bacterial RPRs. A representative inhibition curve is shown in Figure 4 and the Ki values were determined to be in the low μM range irrespective of the RPR we tested (Table 1). Comparing the Ki values for Eco RPRwt, Myc RPR and Hyo RPR revealed little variation, at most twofold. Replacing pSu3 with pSu1 in our assays resulted in a twofold increase in Ki for inhibition of Eco RPRwt by toluidine blue O (cf. Tables 1 and 2). These data demonstrated that the addition of phenothiazine compounds inhibits the activity of structurally different bacterial RPRs.

Inhibition of RNase P RNA Cleavage by Phenothiazine Derivatives
Increasing concentrations of the selected phenothiazine derivatives inhibited the activity of all three bacterial RPRs. A representative inhibition curve is shown in Figure 4 and the K i values were determined to be in the low µM range irrespective of the RPR we tested (Table 1). Comparing the K i values for Eco RPR wt , Myc RPR and Hyo RPR revealed little variation, at most twofold. Replacing pSu3 with pSu1 in our assays resulted in a twofold increase in K i for inhibition of Eco RPR wt by toluidine blue O (cf. Tables 1 and 2). These data demonstrated that the addition of phenothiazine compounds inhibits the activity of structurally different bacterial RPRs. K i corresponds to the concentration that led to 50% inhibition of cleavage activities. Each value was calculated from plots like the one in Figure 4 and represents the average of three independent experiments ± the maximum deviation from the average value.

Eco RPR Variant Toluidine Blue O (µM) Chlorpromazine (µM)
Eco RPR wt 20 ± 4 450 ± 40 Eco RPR G235 17 ± 3 200 ± 30 Eco CP RPR 30 ± 5 ND K i corresponds to the concentration that led to 50% inhibition of cleavage activities by the indicated Eco RPR variants. Each value was calculated from plots like the one in Figure 4 and represents the average of three independent experiments ± the maximum deviation from the average value. The concentration of pSu1 was ≤0.02 µM while the concentrations of Eco RPR wt and Eco RPR G235 were 0.082 µM and Eco CP RPR was 5.4 µM (for details see Materials and Methods). ND, not determined.
The intercalator RNA dye pyronine G, which has a similar structure compared to phenothiazine ( Figure 3), also inhibited the activity of all three RPRs in the low µM range, with a threefold variation in K i (Table 1). Pyronine G has an oxygen and a carbon at positions 5 and 10, respectively, while phenothiazine has sulfur and nitrogen at those positions ( Figure 3). Thus, having sulfur and nitrogen at positions 5 and 10 is not essential for inhibiting the activity of bacterial RPRs.
The inhibition of Eco RPR activity by aminoglycosides is pH-dependent [15]. We therefore inquired whether pH also influences the inhibition by phenothiazine. For toluidine blue O, we did not detect any change in K i when the reaction was performed at pH 6, pH 7.9 (our standard pH), or pH 9 (Table 3). We also inquired whether higher salt concentrations influenced inhibition by toluidine blue O and thionine acid. The K i values increased two-to almost fourfold in the presence of 100 mM Mg 2+ and 100 mM NH 4 + , i.e., 10 to 100 times higher concentrations, respectively, compared to 10 mM Mg 2+ and 1 mM NH 4 + (see Materials and Methods). Based on this data alone, however, we cannot determine whether the increase in K i is due to the higher concentration of NH 4 + and/or the higher concentration of Mg 2+ .
Next, we asked whether the order of addition of Eco RPR vs. substrate (S; pSu3) and inhibitor (I; toluidine blue O or thionine acid), had any effect on K i . The data showed that pre-incubating pSu3 together with the I resulted in at least a fourfold lower K i than when the RPR and I were pre-incubated together or when RPR, S, and I were pre-incubated separately and then mixed (Table 3). These results indicated that toluidine blue O or thionine acid might bind to pre-tRNA. RPR + I then S: RPR + I were pre-incubated for 10 min at 37 • C followed by the addition of S; S + I then RPR: S + I were pre-incubated for 10 min at 37 • C followed by the addition of pre-incubated Eco RPR wt ; RPR, I and S: RPR, I and S were pre-incubated at 37 • C separately for 10 min and then mixed. K i corresponds to the concentration that led to 50% inhibition of cleavage activities as indicated. Each value was calculated from plots like the one in Figure 4 and represents the average of three independent experiments ± the maximum deviation from the average value. The concentration of pSu3 was ≤0.02 µM while the concentration of Eco RPR was 0.082 µM.

Lead(II)-Induced Cleavage of Yeast tRNA Phe , P15 RNA, and Eco RPR
To test whether phenothiazine indeed binds to tRNA, we studied lead(II)-cleavage of yeast tRNA Phe based on the precedent that aminoglycosides interfere with Pb 2+ binding to RNA ( [15,18]; see Materials and Methods). Our data indicate that all phenothiazine derivatives inhibit lead(II)-cleavage of yeast tRNA Phe ( Table 1). The K i values for inhibition of Pb 2+ -cleavage were higher compared to values determined for inhibiting the bacterial RPRs. In addition, we tested pyronine G and it also inhibited lead(II)-cleavage of yeast tRNA Phe . However, K i was significantly higher than for the phenothiazine derivatives. Together, these data suggested that phenothiazine and pyronine G bind to tRNA, albeit with different affinities.
The functionally important P15 domain in Eco RPR wt (type A; Figure 2; [1]) is cleaved at specific sites in the presence of Pb 2+ in the context of both full-length Eco RPR wt and as a separate module [32]. Hence, we asked whether lead(II)-induced cleavage of P15 RNA, a 31-nt-long RNA [33] representing the P15-domain, is also inhibited by phenothiazine and pyronine G. K i values ranging from 14 to 60 µM indicate that these inhibitors also bind to the P15 RNA. Collectively, these data suggested that phenothiazines bind to RNA and interfere with binding and/or positioning of Pb 2+ .

Inhibition of Bacterial RPR Activities by Antipsychotic Phenothiazine Variants
The antipsychotic drugs thioridazine, trifluoperazine and chlorpromazine are phenothiazine variants with different functional (R) groups, R 1 and R 2 , at positions 2 and 10 ( Figure 3). As was observed for the other phenothiazine derivatives, addition of any of these molecules inhibited the activity of all three RPRs. However, the K i values, which ranged from 67 to 437 µM, were higher than the values for the other phenothiazine derivatives (Table 1; see also Table 2 where pSu1 was used as substrate). This increased K i might be due to the presence of bulky R 2 groups. Moreover, thioridazine, trifluoperazine, and chlorpromazine also inhibited lead(II)-induced cleavage of P15 RNA, with K i values similar to the values for inhibition of RPR activity. We also tested if the "three ring" antidepressants imipramine and amitriptyline ( Figure 3) showed any inhibitory effect on bacterial RPR activity, but we did not detect any inhibition under our conditions (data not shown).

Binding of Phenothiazine Affects the Structure of Eco RPR
To elucidate where phenothiazine binds RPR wt we used Pb 2+ , which induces cleavage at specific and well-defined positions in the Eco RPR wt ( Figure 2). As shown in Figure 5, Pb 2+ -induced cleavage of Eco RPR wt in the presence of toluidine blue O and chlorpromazine resulted in changes in the vicinities of sites Ib, IIa, and IIb as well as near III, IV, and V in the case of chlorpromazine ( Figure 5, marked with •). The change in cleavage at III and V is consistent with phenothiazine inhibition of lead(II)-induced cleavage of P15 RNA (Table 1). Sites IIa and IIb are located in the S domain, which constitutes the binding site for the pre-tRNA T-stem/loop in bacterial RPR [1,34]. Hence, these data suggested that toluidine blue O and chlorpromazine bind to Eco RPR, altering its structure and interfering with the binding of Pb 2+ ions in several Eco RPR regions.

Phenothiazine Inhibition of Structurally Altered Eco RPR
To further test the influence of the Eco RPR wt S domain with respect to phenothiazine inhibition, we determined the K i values for inhibition of two variants, Eco RPR G235 and Eco CP RPR, by toluidine blue O and chlorpromazine (the latter was not tested for inhibition of Eco CP RPR). Eco RPR G235 carries a mutation in the region near lead(II)-induced cleavage sites Ia, Ib, IIa and IIb (Figure 2; see also Figure 6) and it influences cleavage at sites Ia and IIb ( Figure 5, cf. lane 9 vs. 4; [25]). The other variant Eco CP RPR lacks the S domain [30,31]. With pSu1 as substrate, toluidine blue O inhibited the activity of these two Eco RPR variants with similar K i values as observed for Eco RPR wt . For Eco RPR G235 and chlorpromazine, the K i was roughly twofold lower compared to that determined for Eco RPR wt (Table 2). Moreover, the Pb 2+ -induced cleavage pattern of Eco RPR G235 in the presence of these two phenothiazine derivatives was similar compared to that observed for Eco RPR wt . Of note, while cleavage at site Ia for Eco RPR G235 was reduced in the absence of inhibitor, the addition of either toluidine blue O or chlorpromazine led to an even greater reduction in cleavage at that site ( Figure 5, cf. lane 9 (without inhibitor) and lanes 7 and 10 (with inhibitor)). Together, these data suggest that both the S and C domains are targeted by phenothiazine derivatives (see also above).

Phenothiazine Inhibition of Structurally Altered Eco RPR
To further test the influence of the Eco RPRwt S domain with respect to phenothiazine inhibition, we determined the Ki values for inhibition of two variants, Eco RPRG235 and Eco CP RPR, by toluidine blue O and chlorpromazine (the latter was not tested for inhibition of Eco CP RPR). Eco RPRG235 carries a mutation in the region near lead(II)-induced cleavage sites Ia, Ib, IIa and IIb (Figure 2; see also Figure 6) and it influences cleavage at sites Ia and IIb ( Figure 5, cf. lane 9 vs. 4; [25]). The other variant Eco CP RPR lacks the S domain [30,31]. With pSu1 as substrate, toluidine blue O inhibited the activity of these two Eco RPR variants with similar Ki values as observed for Eco RPRwt. For Eco RPRG235 and chlorpromazine, the Ki was roughly twofold lower compared to that determined for Eco RPRwt (Table 2). Moreover, the Pb 2+ -induced cleavage pattern of Eco RPRG235 in the presence of these two phenothiazine derivatives was similar compared to that observed for Eco RPRwt. Of note, while cleavage at site Ia for Eco RPRG235 was reduced in the absence of inhibitor, the addition of either toluidine blue O or chlorpromazine led to an even greater reduction in cleavage at that site ( Figure 5, cf. lane 9 (without inhibitor) and lanes 7 and 10 (with inhibitor)). Together, these data suggest that both the S and C domains are targeted by phenothiazine derivatives (see also above). (B) Three-dimensional structure of the P7, P10 and P11 region in the S domain that is suggested to constitute one binding site for phenothiazine derivatives. Residues marked with black arrows 1-3 and in blue (C122, C123 and A124), magenta (A232 and A233) and green (C235, changed to G235, see Figure 2); for color code see panel A; (C) The nuclear magnetic resonance structure of human immunodeficiency virus type 1 (HIV1) trans-activation response element (TAR) RNA in complex (A) Secondary structures of various RNAs shown to bind to phenothiazine; (B) Three-dimensional structure of the P7, P10 and P11 region in the S domain that is suggested to constitute one binding site for phenothiazine derivatives. Residues marked with black arrows 1-3 and in blue (C 122 , C 123 and A 124 ), magenta (A 232 and A 233 ) and green (C 235 , changed to G 235 , see Figure 2); for color code see panel A; (C) The nuclear magnetic resonance structure of human immunodeficiency virus type 1 (HIV1) trans-activation response element (TAR) RNA in complex with acetopromazine. Black arrows mark acetopromazine (in green) and the nucleotides (in magenta) forming the UCU-bulge (see panel A).

Discussion
Various phenothiazine derivatives have been demonstrated to bind to RNAs ( Figure 6). For example, the human immunodeficiency virus (HIV) trans-activation response element (TAR) RNA that interacts with the trans-activator of transcription (Tat) protein, forming a complex that is essential for activating the transcription of the HIV genome [35][36][37]. Moreover, nuclear magnetic resonance (NMR) data revealed binding of the phenothiazine acetopromazine [38] to small RNAs representing the E. coli ribosomal A site, the coxsackievirus (CV) B3 loop D RNA (CV loop D RNA), and the loop B of the poliovirus (PV) internal ribosome entry site (IRES) (PV loop B RNA). The K D values for HIV-TAR RNA, ribosomal A site, and CV loop D RNA were reported to be ≈300 µM whereas the K D for the PV loop B RNA, at 1.8 mM, was much higher [38]. Here, we provide data showing that different phenothiazines and pyronine G inhibited the activities of three structurally different bacterial RPRs, including M. tuberculosis RPR, while the three-ring compounds imipramine and amitriptyline did not. Interestingly, three antipsychotic phenothiazines known to have antibacterial activities [39] also inhibited the activity of all three bacterial RPRs; while their K i values were higher than those of the other phenothiazines tested, they were in the same range as the K D value for acetopromazine binding to, e.g., HIV-TAR RNA (Table 1; [38]). Collectively, our findings are, to the best of our knowledge, the first to show that long functional non-coding RNAs (≈400 nt) are targeted by different phenothiazine derivatives.

Comparison with Aminoglycoside Interaction with RNA
The K i values for the phenothiazines are similar to those determined for inhibition of Eco RPR wt with and without the C5 protein by aminoglycosides such as neomycin B [12,15]. However, in contrast to aminoglycosides (and other RNA-based activities [40,41]) inhibition of bacterial RPR activity by phenothiazines does not appear to be pH-dependent. We also did not detect as pronounced an effect on inhibition in response to increasing Mg 2+ and NH 4 + concentrations as was reported for aminoglycoside derivatives [15,17]. At most, a fourfold increase in K i was observed (Table 3). Electrostatics is a major factor that accounts for aminoglycoside binding to RNA, and for E. coli and human rRNA A sites, the electrostatic contribution is about one-half of the total binding energy [42]. Contributions also come from non-ionic and pseudo-base-pair interactions, water-mediated contacts, shape complementarity, and conformational adaptation [15,17,40,41]. Moreover, NMR studies revealed that the interaction between TAR RNA and the aminoglycoside neomycin B results in distortion of the RNA structure, whereas binding of acetopromazine to TAT RNA induces only minor changes [43]. Hence, these differences suggest that phenothiazine binding to RNA is governed by features other than aminoglycoside binding [41]. However, as discussed below, structural information on phenothiazine RNA interaction is limited.

Phenothiazines Interact Preferentially with Internal RNA Bulges
Although, we do not know the mode of interaction of phenothiazines with bacterial RPRs, the NMR structure of the acetopromazine-TAR RNA complex and the phenothiazine's interaction with other RNA molecules such as the E. coli A site RNA provide several insights ( Figure 6; [36,38,43]). First, the K D values for acetopromazine binding to HIV-TAR RNA, E. coli A site RNA, and CV D loop RNA are reported to be approximately 300 µM. This is within the same range as the K i values for RPR inhibition by chlorpromazine, thioridazine, and trifluoperazine (Table 1), while the values for the methylene derivatives were significantly lower (Table 1). Second, phenothiazines have a nonplanar structure, and NMR data suggest that the tricyclic ring system is in closest contact with the RNA and forms most of the contacts, whereas weaker contacts are contributed by the aliphatic side chain. Third, acetopromazine has a preference for internal bulges and binding to smaller (two to three nucleotides in size) bulges is slightly stronger. Moreover, NMR data suggest only minor differences in the RNA structure with and without the inhibitor. Fourth, acetopromazine binds poorly to RNA loop structures and no detectable binding to a UUCG-tetraloop has been reported [38]. Fifth, there is no preference for internal bulges composed of purines or pyrimidines. Sixth, the identity of the 2-position substituent affects binding affinity. Finally, acetopromazine has a preference for internal bulges and hence it is likely that phenothiazines bind in the vicinity of bulges and internal loop regions in RPRs.

Phenothiazine Interaction with the S Domain and the Pre-tRNA D/T Loop
In the Eco RPR S domain, there are two internal bulges ( Figure 2). Lead(II)-induced cleavage in the presence of toluidine blue O (and chlorpromazine) was affected at and near sites Ib, IIa, IIb and IIb in the S domain, which are proximal to these two internal bulges. The crystal structure of the type A RPR from Thermus thermophilus showed that sites Ia, Ib, IIa, IIb and IIb are in close proximity to each other ( [44,45]; see also [34]). Moreover, the C to G change at position 235 in Eco RPR, which is located in the vicinity of these sites in the S domain (Figure 2), resulted in a twofold lower K i value for chlorpromazine while the K i of toluidine blue O was unaffected ( Table 2). In addition, the lead(II)-induced cleavage patterns for Eco RPR wt and Eco RPR G235 with and without phenothiazines were similar except for site Ia, in particular, which is known to be affected in response by having G at position 235 [25]. Taken together and based on the preference of acetopromazine for internal bulges, it is likely that phenothiazines bind at or in the vicinity of the bulges near site Ia, Ib, IIa, IIb, and IIb in the S domain. The S domains of type A and type B RPRs are similar [44] and type B Hyo RPR is cleaved by Pb 2+ at a site in the S domain that corresponds to site Ia in Eco RPR ( [26]; see also [46,47]). It is therefore reasonable to suggest that phenothiazines also bind the corresponding region(s) (i.e., near site Ia) in Hyo RPR. This putative phenothiazine binding site in the S domain would be close to the site that interacts the pre-tRNA T stem/loop (TSL) region, referred to as the TSL-binding site (TBS) [25]. Thus, binding of phenothiazine at (or near) this location might interfere with the interaction between TSL and TBS.
Our analysis further revealed that phenothiazines interact with yeast-tRNA Phe and inhibit lead(II)-induced cleavage in the D loop between residues 16 and 17. Residues in the TSL region are involved in coordinating the Pb 2+ ion that activates the 2 OH at position 16 that results in hydrolysis of the RNA [48,49]. The phenothiazines may also bind in the TSL region. However, neomycin B inhibition of lead(II)-induced cleavage of yeast tRNA Phe is the result of binding distant to the cleavage site [18]. Thus, we cannot exclude that the phenothiazine binds elsewhere and indirectly interferes with the structural topology of the T/D loop such that binding of the catalytic Pb 2+ is affected. Interestingly, pre-incubating phenothiazine and pre-tRNA prior to adding Eco RPR resulted in a four-to eightfold reduction in K i for toluidine blue O and thionine acid, respectively (Table 3). This indicates that inhibitor binding to the pSu3 substrate is more inhibitory than inhibitor binding to the RPR, which may be due to a structural change in the precursor tRNA that results in enhanced affinity for the inhibitor. However, inhibition of lead(II)-induced cleavage of yeast tRNA Phe required higher phenothiazine concentrations compared to inhibition of RPR (Table 1). This might reflect differences in the pSu3 and yeast tRNA Phe structures, which affect the binding affinities for the phenothiazines. Nevertheless, these observations raise the question that binding of some phenothiazine derivatives like chlorpromazine might interfere with the interaction between the TSL and TBS [25,50]. Additional studies are required to address these possibilities.

Phenothiazine Interaction with the C Domain and the Pre-tRNA D/T Loop
Eco CP RPR, which lacks the S domain, is still catalytic [30,31] and addition of toluidine blue O inhibited the activity (Table 2). This finding suggested that phenothiazines also bind to one or more other sites in Eco RPR in addition to the binding site in the S domain. Moreover, binding of Pb 2+ ions in the P15-P17 region promotes hydrolysis at sites III, IVa, and V in full-length Eco RPR and the addition of chlorpromazine in particular influenced cleavage at these sites (Figures 2 and 5; [51,52]). Lead(II)-induced cleavage at sites III and V was also detected using P15 RNA, a model representing the L15 internal loop (Figure 2; [32,33]); addition of phenothiazines inhibited lead(II)-induced cleavage of this RNA (Table 1). This suggests that phenothiazines bind the L15-loop of bacterial type A RPR, the region that interacts with the 3 RCC end of precursor tRNA [1,2] Thus, we conclude that phenothiazines interact with functionally important regions in both the S and C domains of bacterial RPRs. In this context it is of interest that both the L15 and TBS regions have been exploited as part of antisense approaches to inhibit the activity of bacterial RNase P [13,[53][54][55].

Phenothiazine, a Scaffold with Potential
Phenothiazines are promiscuous and they bind and interfere with targets such as nitric oxide synthase, acetylcholinesterases, disulfide reductases, DNA, Ca 2+ -dependent enzymes, enzymes involved in lipid metabolism, ATP transporters, and various bacterial efflux pumps, and it has been suggested that trifluoperazine acts as a calmodulin antagonist [21,39,56]. Our present data adds RNA, particularly bacterial RPRs, to the list of phenothiazine targets. Comparing chlorpromazine inhibition of bacterial RPR (Table 1) and chlorpromazine inhibition of previously reported targets such as trypanothione reductase (K i = 10.8 µM [57]), Na + /K + -ATPase from rat brain (IC 50 = 29 µM [58]) and mitochondrial malate dehydrogenase (inhibition observed at >140 µM [59]) reveals that higher concentrations are required to inhibit bacterial RPR. It should also be noted that since the discovery of methylene blue's antimalarial activity, other phenothiazines have been synthesized and shown to be useful drugs in various settings. The best-known examples are the antipsychotic drugs chlorpromazine and thioridazine. These drugs also have antibacterial activity; in particular, thioridazine has apparently been instrumental in treating patients infected with extreme drug-resistant M. tuberculosis strains [60][61][62]. Given the structural differences between human and bacterial RNase P and the inhibition of bacterial RPRs, including M. tuberculosis RPR, by selected phenothiazine derivatives, development of new phenothiazine derivatives that are more specific in targeting bacterial RNase P warrants consideration. Additionally, it has been discussed that methylene blue might slow down the development of Alzheimer's disease since it can inhibit the aggregation of tau protein ( [21] and references therein). In this context, aminoglycosides have been reported to interact with the tau exon 10 splicing regulatory RNA element [63]. Therefore, this raises the possibility that phenothiazines also interfere with the expression of the tau protein and RNA splicing in general.

Preparation of Compounds
The different phenothiazine derivatives, pyronine G, imipramine, and amitriptyline used in the present study were purchased from Sigma-Aldrich (Stockholm, Sweden) and VWR International (Stockholm, Sweden). All compounds were dissolved in double-distilled water prior to use.

Determination of the Inhibition Constant K i
The K i value was defined as the inhibitor concentration that results in 50% inhibition of cleavage. Conditions were adjusted so that measurements were carried out when product formation was in the linear part of the curve of kinetics of the cleavage reaction. Each value in µM was calculated from plots as the one in Figure 4 and the values are averages of three independent experiments ± the maximum deviation of the average value.