Critical domain interactions for type A RNase P RNA catalysis with and without the specificity domain

The natural trans-acting ribozyme RNase P RNA (RPR) is composed of two domains in which the catalytic (C-) domain mediates cleavage of various substrates. The C-domain alone, after removal of the second specificity (S-) domain, catalyzes this reaction as well, albeit with reduced efficiency. Here we provide experimental evidence indicating that efficient cleavage mediated by the Escherichia coli C-domain (Eco CP RPR) with and without the C5 protein likely depends on an interaction referred to as the "P6-mimic". Moreover, the P18 helix connects the C- and S-domains between its loop and the P8 helix in the S-domain (the P8/ P18-interaction). In contrast to the "P6-mimic", the presence of P18 does not contribute to the catalytic performance by the C-domain lacking the S-domain in cleavage of an all ribo model hairpin loop substrate while deletion or disruption of the P8/ P18-interaction in full-size RPR lowers the catalytic efficiency in cleavage of the same model hairpin loop substrate in keeping with previously reported data using precursor tRNAs. Consistent with that P18 is not required for cleavage mediated by the C-domain we show that the archaeal Pyrococcus furiosus RPR C-domain, which lacks the P18 helix, is catalytically active in trans without the S-domain and any protein. Our data also suggest that the S-domain has a larger impact on catalysis for E. coli RPR compared to P. furiosus RPR. Finally, we provide data indicating that the absence of the S-domain and P18, or the P8/ P18-interaction in full-length RPR influences the charge distribution near the cleavage site in the RPR-substrate complex to a small but reproducible extent.


Introduction
Almost all tRNAs carry a phosphate at their 5' ends due to the action of the endoribonuclease RNase P. Bacterial RNase P consists of one protein (C5), and one RNA subunit [1]. The composition of archaeal and eukarayal RNase P is more complex where the sole RNA subunit binds several proteins [2,3]. Available data suggest that the catalytic activity resides in the RNA irrespective of origin, and the RNA alone can cleave various substrates in the absence of a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 substrates when the C5 protein is absent, which is not the case for full-size type A RPR. However, in the presence of the C5 protein a modest reduction in cleavage activity for Eco CP RPR wt was detected upon deleting P18. Our data also show that deletion of the S-domain of Pfu RPR resulted in an RPR that is catalytically active in the absence of proteins. We also found that deletion as well as disruption of the P8/ P18-interaction in full-size Eco RPR lowers the cleavage efficiency of a model substrate. On the basis of our data combined with the fact that the P8/ P18-interaction (or P6) is not in direct vicinity of where substrate cleavage occurs we raise the possibility that the P8/ P18-interaction acts as a structural mediator between the TSL/ TBS-interaction site and the active center leading to positioning of chemical groups and Mg 2+ that ensures correct and efficient cleavage.

Preparation of substrates and RPR
The substrates were purchased from Dharmacon, USA and were purified on a 15% (w/v) denaturing PAGE gel followed by an overnight Bio-Trap extraction (Schleicher and Schuell, GmbH, Germany; Elutrap in USA and Canada) and phenol-chloroform extraction. γ-ATP 5' end-labeled substrates were generated and gel-purified using standard protocols.
The genes encoding full-size Eco RPR wt (M1 RNA), Eco CP RPR wt and Pfu RPR wt have previously been described [22,31,33]. The genes encoding the variants Eco RPR P18CUUG , Eco RPR G235 , Eco CP RPR C83C84 , Eco CP RPR G278G279 , Eco CP RPR C83C84/G278G279 , Eco CP RPR delP18 , Eco CP RPR delP18P3Mini , Eco CP RPR 31 , Eco CP RPR 31delP18 and Pfu CP RPR wt behind the T7 promoter were generated following the same procedure as outlined elsewhere [22,31,33] using the Eco CP RPR wt and Pfu CP RPR wt genes as template and appropriate oligonucleotides. Eco RPR delP18 was generated by replacing the 3' half of Eco RPR wt with the 3' half of Eco CP RPR delP18 using appropriate restriction enzymes. The different RPRs were generated as run-off transcripts using T7 DNA-dependent RNA polymerase and PCR-amplified templates [34,35]. The C5 protein was purified as described in [34,36].

Assay conditions
The cleavage reactions without the C5 protein were conducted in buffer C [50 mM 4-morpholineethanesulfonic acid (MES) and 0.8 M NH 4 Cl (pH 6.1)] at 37˚C and 800 mM Mg(OAc) 2 or as otherwise indicated (see Supporting information S1 Fig). The RPRs were pre-incubated at 37˚C in buffer C and 800 mM Mg(OAc) 2 for at least 10 min to allow proper folding before mixing with pre-heated (37˚C) substrate. In all the experiments the concentrations of substrates were 0.02 μM while the concentrations of the different RPR variants were as indicated in Table and Figure legends. Reactions with the C5 protein were done in buffer A [50 mM Tris-HCl (final pH 7.2), 5% (w/v) PEG 6000, 100 mM NH 4 Cl] and 10 mM Mg(OAc) 2 as described in [34].
In the RPR alone reactions the k app values were determined from experiments done under single turnover conditions where we measured the percentage of cleavage as a function of time (with C5 the RPR concentration varied between 0.004 and 0.009 μM). For the calculations we used the 5' cleavage fragment. To be able to compare with our previously reported data we refer to the so obtained rates as k app values. The concentration of substrate was 0.02 μM while the RPR concentration varied dependent on RPR variant (see Table 1).
Cleavage of pATSerU am G at 37˚C was performed in buffer C, 0.8 M NH 4 Cl and 800 mM Mg(OAc) 2 at pH 5.2, pH 6.1 and pH 7.2 [37,38].
The cleavage reactions were terminated by adding double volumes of stop solution (10 M urea, 100 mM EDTA) and the products were separated on 25% (w/v) denaturing polyacrylamide gels.

Structural probing
Structural probing of the Eco RPR variants, labeled at the 3'-end with [ 32 P]pCp, was conducted using Pb 2+ -induced cleavage and limited RNase T1 digestion under native conditions as described elsewhere [34,39,40,41]. Approximately 2 pmols of labeled RPR in 10 μl was preincubated for 10 min at 37˚C in 50 mM Tris-HCl (pH 7.5), 100 mM NH 4 Cl and 10 mM MgCl 2 together with 4 μM unlabeled tRNA. Cleavage was initiated by adding freshly prepared Pb(OAc) 2 to a final concentration of 0.5 mM (or as indicated in Fig 4 legend) and the reaction was stopped after 10 min. In the digestion with RNase T1, the RPR was pre-incubated as described above. One unit RNase T1 was added followed by incubation on ice for 10 min. The reactions were stopped after 10 min by adding two volumes of stop solution (see above) and the products were analyzed on an 8% (w/v) denaturing polyacrylamide gel.

RNase H cleavage
Approximately one μg of 3'-[ 32 P]pCp labeled RPR was re-suspended in H 2 O and incubated for 3 min at 95˚C. Following this the RPR was re-natured prior to the reaction at 55˚C for 5 min in the buffer supplied by the company (20 mM Tris-HCl, 20 mM KCl, 10 mM MgCl 2 , 0.1 mM EDTA, 0.1 mM DTT and final pH 7.5; ThermoFisher Scientific) followed by incubation at room temperature. The RPR was mixed with 120 pmols of DNA oligonucleotides 1 (5'TGCCCT) or 2 (5'TGGGCT) and incubated in reaction buffer (see above) for 15 min at 28˚C (similar results were obtained using 12 pmols of DNA oligonucleotides 1 or 2). The reaction was initiated by adding one unit RNase H (ThermoFisher Scientific) and the reaction was terminated after 30 min by adding double volumes of stop solution (see above). The reaction products were separated on 10% (w/v) denaturing polyacrylamide gels (see also Fig 3 legend and Ref [39]).

Determination of k app and the kinetic constants k obs , k obs /K sto and K sto
The rate constants k obs and k obs /K sto were determined under saturating single-turnover conditions at pH 6.1 (where cleavage is suggested to be rate limiting) and 800 mM Mg 2+ as described elsewhere [19,23,34]. On the basis of the simplified scheme k obs reflects the rate of cleavage (Fig 1). We have argued elsewhere that K sto % K d in the Eco RPR-alone reaction [19,23,31,34,42,43]. The final concentrations of the different RPR variants were between 0.4 and 47 μM (depending on the combinations of substrate and RPRs); the concentration of the pATSerUG substrate was 0.02 μM. To ensure that the experiments were done under the single-turnover conditions the lowest concentration of RPR was >10 times higher than the concentration of the substrate. For the calculations we used the 5' cleavage fragment and the time of cleavage was adjusted to ensure that the velocity measurements were in the linear range (i.e., 40% of the substrate had been consumed). To be able to compare with our previously published data k obs and k obs /k sto were obtained by linear regression from Eadie-Hofstee plots as described elsewhere [19,23,31,34,44,45]. Each value was an average of at least three independent experiments and is given as a mean ± the deviation of this value.

Structural probing of Eco CP RPR and full-size Eco RPR variants
Presence of a "P6-mimic" in Eco CP RPR wt . The "P6-mimic" might form in Eco CP RPR wt since residues in the P17-loop that constitute one part of P6 are open to pair with other residues in Eco CP RPR as a result of deleting the S-domain (marked in green in Fig 2A). To test for the presence of the "P6-mimic" we generated the following Eco CP RPR variants (Fig 2A and Table 1): Eco CP RPR C83C84 ("P6-mimic" disrupted), Eco CP RPR G277G278 ("P6-mimic" disrupted), Eco CP RPR C83C84/G277G278 ("P6-mimic" restored), Eco CP RPR 31 ("P6-mimic" disrupted due to replacement of the P15-17 domain with P15 RNA [22,23]). These variants (except Eco CP RPR 31 ) were probed with respect to the accessibility of residues 5'A 81 GGGCA 86 (underlined residues altered in the respective CP RPR variant; Fig 2A) to RNase H in the presence of DNA oligonucleotides (i) 5'TGCCCT (oligo 1), complimentary to residues (underlined) 5'A 81 GGGCA 86 in Eco CP RPR wt and Eco CP RPR G277G278 and ii) 5'TGGGCT (oligo 2), complimentary to residues 5'A 81 GCCCA 86 in Eco CP RPR C83C84 and Eco CP RPR C83C84/G277G278 ; Fig 2A; see also Ref [39]). We expected that disruption of the "P6-mimic" should result in RNase H cleavage of Eco CP RPR in the presence of oligo 1 and oligo 2 in a predictable manner. Subjection to cleavage with RNase H in the presence of either of the two DNA oligonucleotides indeed revealed strong cleavage for Eco CP RPR G277G278 (oligo 1), and Eco CP RPR C83C84 (oligo 2) as expected if the "P6-mimic" does not form. By contrast, for Eco CP RPR wt and Eco CP RPR C83C84/G277G278 we did observe significant lower  [6,46,47], residues in green refer to residues involved in formation of the "P6-mimic" and residues highlighted (black circles) were changed as indicated: i) Eco CP RPR C83C84 , Eco CP RPR G277G278 and Eco CP RPR C83C84/G277G278 ,. The G and C residues marked in red were added during the construction of the original Eco CP RPR wt construct (see Refs [22,23]). Replacements marked with grey boxes and red arrows indicate: i) the 31-mer "P15 RNA" module replacement of P15-17 in Eco CP RPR 31 and Eco CP RPR 31delP18 ), ii) P3Mini module replacement of the native P3 in Eco CP RPR delP18P3Mini and iii) deletion of P18 in Eco CP RPR delP18 , Eco CP RPR 31delP18 and Eco CP RPR delP18P3Mini . The roman numerals and arrows in blue mark the Pb 2+ -induced cleavage sites IIc to VI (in Fig 2B, I to VI). For convenience the "332-region" is also indicated in blue and a vertical blue line. For details see the text. B. Type A Eco RPR, the residues constituting the Sdomain are shown in light gray letters and the dotted line demarcates the border separating the S-and the C-domain. Residues highlighted (black circles) were changed as indicated to generate Eco RPR P18CUUG and Eco RPR G235 while the red arrow indicate deletion of P18 generating Eco RPR delP18 . C. Type A Pfu RPR derived from Pyrococcus furiosus. On the basis of the results using Eco CP RPR residues in green refer to residues involved in formation of the "P6-mimic". cleavage (Fig 3). We interpreted these data as an indication that the "P6-mimic" is likely to form in both Eco CP RPR wt and Eco CP RPR C83C84/G277G278 . Structural impact of the P8/ P18-interaction and P18 on full-size Eco RPR and Eco CP RPR. To investigate the influence of P18 on the Eco CP RPR structure and its contribution to catalysis (see below) we generated the following variants (Fig 2A) Eco CP RPR delP18 (P18 deleted), Eco CP RPR delP18P3Mini (P18 deleted, P3 size reduced) and Eco CP RPR 31delP18 ("P6 mimic" and P18 deleted). For comparison we also generated two full-size variants, Eco RPR delP18 and Eco RPR P18CUUG , which both disrupt the P8/ P18-interaction albeit in different ways, in the former P18 is deleted while in the latter the P8/ P18-interaction is disrupted ( Fig  2B). Comparing Eco CP RPR wt and Eco RPR wt allowed us also to assess whether removal of the S-domain (and the P8/ P18-interaction) affected the structure of the C-domain. First we studied the impact of the P8/ P18-interaction on full-size Eco RPR.
Structural probing of Eco RPR wt and Eco RPR P18CUUG with Pb 2+ and RNase T1 (Fig 4A) suggested that disruption of the P8/ P18-interaction affected the P18 structure and the region near Pb 2+ -induced cleavage sites IIb and possibly also IIb' (marked with a dot and absent in Eco RPR P18CUUG at 10 mM Pb 2+ , cf. lanes 3 and 4; see also Fig 4A legend; Pb 2+ cleavage sites are marked in Fig 2). We also noted that a weak RNase T1 cleavage product (marked with a dot; Fig 4A, cf. lanes 10 and 11) was absent just upstream of the Pb 2+ -induced cleavage site IIc at A248 in Eco RPR P18CUUG (Fig 2B), where A248 is close to the tRNA 5' end in the RNase P-tRNA complex [18]. Of note, a higher concentration of Pb 2+ was needed for Eco RPR P18CUUG (Eco RPR delP18 ) also resulted in some changes in the Pb 2+ cleavage pattern with the appearance of an extra band in the IIa/IIb region (marked with a dot, Fig 4, cf. lanes 14 and 15). Also, Pb 2+ mediated cleavage at IIb in Eco RPR delP18 , which is in contrast compared to Eco RPR P18CUUG (Fig 4A, cf. lanes 4 and 15). Moreover, apart from the P18 region the RNase T1 cleavage patterns were similar comparing Eco RPR wt and Eco RPR delP18 (Fig 4A, cf. lanes 16 and 17). Together these data indicate some influence on the overall RPR structure, apart from the P18 region, when the P8/ P18-interaction is absent (or disrupted), in particular in the region near the Pb 2+ cleavage site IIb.  [41,48] and the vertical line marks the cleavage sites in the 332-region [49]. Roman numerals in grey (IIa Ã , IIb Ã , IIc Ã , III Ã and V Ã ; site IV Ã is questionable) mark the Pb 2+ cleavage sites (see bands marked with grey Ã ) and the grey lines the RNase T1 cleavage sites in Eco RPR delP18 . The • mark differences comparing Eco RPR wt and Eco RPR P18CUUG or Eco RPR delP18 . Incubations of RPRs without the addition of Pb 2+ or RNase T1 lanes 1, 2, 5, 6, 12 and 13, and OH refer to the alkaline ladder (lane 7). The reactions were performed as outlined in Materials and Methods using 0.5 mM Pb (OAc) 2 for Eco RPR wt and Eco RPR delP18 while for Eco RPR P18CUUG we used 10 mM Pb(OAc) 2 . B. Pb 2+ -induced cleavage of Eco RPR wt (lane 6), Eco CP RPR wt (lane 7), Eco CP RPR delP18 (lane 8) and Eco CP RPR delP18P3Mini (lane 9). Roman numerals marked in black refers to the Pb 2+ -induced cleavage sites that are present in Eco RPR wt and Eco CP RPR wt while those marked in grey (Eco CP RPR delP18 and Eco CP RPR delP18P3Mini , IIc Ã , III Ã and V Ã likely correspond to the sites IIc, III and V present in Eco RPR wt and Eco CP RPR wt on the basis of the migration of the bands). Band marked with a grey Ã present in Eco CP RPR delP18 and Eco CP RPR delP18P3Mini (X Ã ) likely correspond to cleavage in the vicinity of residues that are part of the "P6-mimic". Bands marked with • refer to the appearance of new cleavage sites in Eco CP RPR wt that are not detected in the full-size Eco RPR wt . Lanes 1-4 incubations of RPRs in the absence of Pb 2+ and OH = alkaline ladder (lane 5). For experimental details see Materials and methods. C. RNase T1 cleavage of Eco RPR wt (lane 6), Eco CP RPR wt (lane 7), Eco CP RPR delP18 (lane 8) and Eco CP RPR delP18P3Mini (lane 9) as indicated. Lanes 1-4 incubations of the RPRs without the addition of RNase T1 while OH = alkaline ladder (lane 5). The grey vertical line mark the region that constitute P18 in Eco RPR wt and Eco CP RPR wt whereas the numbers given in grey mark the RNase T1 cleavage sites in Eco CP RPR delP18 and Eco CP RPR delP18P3Mini and these sites likely correspond to the sites detected in Eco RPR wt and Eco CP RPR wt , e.g. the "276" (marked in grey) cleavage site correspond to the cleavage site at 276 (marked in black). The vertical black lines mark the "248-region" in the CP RPRs. https://doi.org/10.1371/journal.pone.0192873.g004 Next we investigated the influence of P18 on the structure of Eco CP RPR wt , which lacks the S-domain (and the P8/ P18-interaction). The data are shown in Fig 4B (cleavage with Pb 2+ ) and Fig 4C (cleavage with RNase T1). As expected, gel mobility of equivalent cleavage products (relative to full-size Eco RPR wt ) of Eco CP RPRs lacking P18, Eco CP RPR delP18 and Eco CP RPR delP18P3Mini , shifted (cf. shift of bands comparing the patterns for Eco RPR wt and Eco CP RPR wt , e.g. the band that corresponds to Pb 2+ -induced cleavage site V; Fig 4B, compare lanes 6 and 7 with lanes 8 and 9). However, the Pb 2+ -induced cleavage sites IIc, III, and V as well as those near residue 332 are most likely still present in the Eco CP RPR variants (Fig 4B, cf. lanes 6 and 7 vs. 8 and 9; the grey Ã marks the shift of sites IIc, III and IV, and referred to as IIc Ã , III Ã and IV Ã in Eco CP RPR delP18 and Eco CP RPR delP18P3Mini ). This suggested that the metal(II)ion binding sites in the vicinity of these sites likely remain intact despite the absence of the Sdomain nor do they depend on the presence of P18 or the length of P3. The Pb 2+ -induced cleavage at sites IIc, III and V, and in the 326-335-region are most likely also present in the Eco CP RPR constructs that lack P18 (Fig 4B, cf. lanes 8 and 9). Site VI that is present in both Eco RPR wt and Eco CP RPR wt is absent in the Eco CP RPR variants that lack P18 (see also Ref [49]). The bands upstream of IIc (marked with X; Fig 4B) might possibly be the result of Pb 2+induced cleavage near residues constituting the "P6-mimic". RNase T1 cleavage of the Eco CP RPR variants revealed that most of the cleavage sites detected using full-size Eco RPR wt were present (Fig 4C, cf. lanes 6-9). However, we noted one apparent difference in the region referred to as the "248-region" (marked with a short vertical grey line, Fig 4C, cf. lanes 8 and 9) where we detected new and stronger cleavage.
Taken together, the P8/ P18-interaction appears to affect the overall structure of full-size RPR to a certain degree (e.g., cf. Fig 4A lanes 3 and 4) while deletion of the S-domain and P18 does not affect the overall structure of the C-domain to any significant extent except for changes, in particular in the P18-region (as expected) and the region referred to as the "248-region". From our data it also appears that the overall structure of the C-domain is not much affected by deleting the S-domain (compare Eco RPR wt and Eco CP RPR wt ).

Impact of the "P6-mimic", P18 and P8/ P18-interaction on the catalytic performance of Eco CP RPR and full-size Eco RPR
As substrates we used different well-characterized model hairpin-loop substrates, which are derived from the E. coli tRNA Ser Su1 precursor (Fig 5; [34,39,50], and references therein). The longer substrates, pATSerUG and pATSerCG, can interact with the TBS-site in the RPR while the shorter, pMini3bpUG and pMini3bpCG, cannot. As we reported elsewhere optimal cleavage of these substrates by Eco RPR wt or Eco CP RPR wt requires higher Mg 2+ -concentrations [16,19,23,37]. This is also the case for cleavage of pATSerUG with Eco CP RPR delP18 and Eco CP RPR delP18P3Mini (variants described below) where optimal cleavage of pATSerUG was reached at approximately 800 mM Mg 2+ (Supporting Information S1 Fig, we assume this to be the case irrespective of which RPR substrate combination used in this study). On the basis of these data and our earlier studies, the experiments presented here were done at 800 mM Mg 2+ . The choice of this Mg 2+ concentration also allowed us to directly compare the results with our previously published data. Cleavage in the presence of the C5 protein was done at 10 mM Mg 2+ ( [4,33] and Materials and methods). Finally, Eco RPR wt and Eco CP RPR wt can cleave model hairpin loop substrates at the correct (or canonical) site between residue -1 and +1 (referred to as the +1 site) and at the alternative site between positions -1 and -2 (referred to as miscleavage or the -1 site; Fig 5; see e.g., Refs [16,19,34,50]).
The "P6-mimic" affects the catalytic performance of Eco CP RPR wt . As shown in Fig  6A both Eco CP RPR C83C84 and Eco CP RPR G277G278 cleaved pATSerUG with reduced The arrows mark the cleavage sites as indicated: black arrows mark the canonical cleavage sites between residues -1 and +1, and gray arrows mark the alternative site between residues -2 and -1. The differences with respect to the identity of residue -1 are indicated, as is the replacement of the 2'OH to 2'NH 2 at the -1 position in pATSerUG. The numbering in the vicinity of the cleavage sites corresponds to that used for tRNA and precursor tRNA [51]. The precursor tRNA (pre-tRNA) is included to efficiency compared to Eco CP RPR wt and Eco CP RPR C83C84/G277G278 . This was the case in particular in the RNA alone reaction. Determination of the rate of cleavage (k app ; Table 1) for these Eco CP RPR variants without C5 corroborated these findings with 8-and 28-fold lower k app values for Eco CP RPR C83C84 and Eco CP RPR G277G278 , respectively, compared to Eco CP RPR wt . For Eco CP RPR 31 , which cannot form the "P6-mimic" (Fig 2A; substitution of P15-17 with P15 RNA) we detected an almost 50-fold reduction in k app (Table 1; see also Fig 6B and  below). The higher impact in response to replacing the P15-17 domain with P15 RNA might reflect a structural effect on establishing the pairing between the 3'ACC in the substrate and the RPR [51].
Together with the structural probing data discussed above suggested that the "P6-mimic" is likely to be present in Eco CP RPR wt and that it contributes to its catalytic performance.
The P8/ P18-interaction influences cleavage of pATSerUG by full-size Eco RPR. Before analyzing the impact of P18 on catalysis in the Eco CP RPR context we first inquired whether the P8/ P18-interaction influences cleavage of the model hairpin loop substrate pATSerUG in the full-size Eco RPR context. Hence, we used the Eco RPR delP18 and Eco RPR P18CUUG variants (see above), which allowed us also to assess the response upon deleting P18 (Eco RPR delP18 ) and "disruption" of the P8/ P18-interaction (Eco RPR P18CUUG ). Both these variants cleaved pATSerUG mainly at the +1 site with reduced efficiency (Fig 6B, cf. lane 9; and not shown). Compared to Eco RPR wt , the cleavage rates of pATSerUG (k app ; Table 1) for Eco RPR P18CUUG and Eco RPR delP18 were reduced %20-fold and almost 700-fold, respectively. Determination of the kinetic constants under single turnover conditions revealed that "disruption" of the P8/ P18-interaction resulted in a %20-and 160-fold decrease in k obs and k obs /K sto , respectively, while deleting P18 lowered both k obs (and k obs /K sto ) >3000-fold (Table 2; cf. values for Eco RPR wt , Eco RPR P18CUUG and Eco RPR delP18 ). The K sto values correspond to % K d values (see Materials and methods) and no difference was detected comparing Eco RPR wt and Eco RPR delP18 while for Eco RPR P18CUUG K sto was %10-fold higher. These data suggest that the P8/ P18-interaction influence cleavage of pATSerUG, which is consistent with previous findings using pre-tRNAs [28][29][30]. Also, while deleting P18 (Eco RPR delP18 ) affected k obs "disruption" of the P8/ P18-interaction resulted in changes in both k obs and K sto (see Discussion).
P18 does not influence the catalytic performance for Eco CP RPR. On the basis of the data discussed above one prediction was that this might also be the case for Eco CP RPR (see above). However, given that P18 helps to connect the S-and C-domains [18,46] another possibility was that P18 does not affect catalysis since in Eco CP RPR the S-domain is missing (Fig  2A). To test this, and get insight into the contribution of P18 to catalysis, we studied cleavage of pATSerUG, pATSerCG, pMini3bpUG and pMini3bpCG (Fig 5) followed by determinations of the rate constant k app (for pATSerUG) without the C5 protein and for a selected few in its presence (Table 1). We also determined the kinetic constants k obs and K sto in the absence of C5 using pATSerUG as substrate (see Materials and methods; Table 2) and Eco CP RPR wt and Eco CP RPR delP18 (P18 deleted). For Eco CP RPR delP18P3Mini (P18 deleted, P3 size reduced) and Eco CP RPR 31delP18 ("P6-mimic" and P18 deleted) we only determined k app values ( Table 1).
The different Eco CP RPR variants with and without P18 cleaved the four model hairpin loop substrates preferentially at the correct position +1 (see Figs 6B and 7, cleavage with Eco CP RPR wt , Eco CP RPR delP18 and Eco CP RPR delP18P3Mini ; for Eco CP RPR 31 and Eco CP RPR 31delP18 we only tested cleavage of pATSerUG, Fig 6B, cf. lanes 5 and 6; see above). Consistent with our previous data [16,23,34] the cleavage efficiencies for pATSerCG and illustrate the design of the model hairpin loop substrates and their structural differences relative to full-length pre-tRNA substrates (the grey circles correspond to the D-loop, anticodon and variable-loop).
https://doi.org/10.1371/journal.pone.0192873.g005 A. Cleavage with different Eco CP RPR carrying changes that affect the "P6-mimic" (Fig 2A). The experiments were done with and without the C5 protein as indicated. Reactions without the C5 protein were performed in buffer C and 800 mM Mg 2+ (cf. lanes 1 to 5) while those with the protein were done in buffer A and 10 mM Mg 2+ (cf. lanes 6 to 10). All the reactions were done at 37˚C and black and open circles as defined above. The concentrations of the RPRs were 0.7 nM with C5 and 2.7 μM without. The pMini3bpCG were lower than using pATSerUG and pMini3bpUG (Fig 7; note that longer reaction times were needed to cleave pATSerCG and pMini3bpCG). Moreover, the cleavage efficiency of pATSerUG by Eco CP RPR 31delP18 (and Eco CP RPR 31 ; see above) was reduced compared to Eco CP RPR wt (Fig 6B; Table 1, see below).
The substrates with C -1 (pATSerCG and pMini3bpCG; Fig 7, cf. lanes 14-17 and 18-21) were cleaved at the alternative site -1 irrespective of Eco CP RPR variant while the U -1 substrates were cleaved only with a low frequency at -1. A comparison of cleavage of pATSerCG and pMini3bpCG suggested that the latter is cleaved more frequently at -1 by the Eco CP RPR variants. The reason to this is at present unclear, however, it might be related to that these two substrates interact differently with Eco CP RPRs and/ or that the positioning of Mg 2+ in the vicinity of the respective cleavage site differs. For example it has been suggested that residues near the conserved U69 in Eco RPR ( Fig 2B) interact with the residue positioned five bases 3' of the cleavage site [53] and pMini3bp only has a stem of three base pairs (Fig 5).
Consistent with our previous data [23] Eco CP RPR wt cleaved pATSerUG with a reduced rate (k app decreased %100-fold) both with and without the C5 protein compared to full-size Eco RPR wt (Table 1). In contrast to full-size Eco RPR, where "disruption" (Eco RPR P18CUUG ) of the P8/ P18-interaction resulted in a 20-fold (or almost 700-fold upon deleting P18; see above) reduction in k app , deleting P18 in Eco CP RPR did not affect k app (if anything, there was a modest %two-fold increase for Eco CP RPR delP18 ). In the presence of the C5 protein, there was a modest three-to four-fold decrease in k app for the Eco CP RPR variant lacking P18 (Table 1; cf. values for Eco CP RPR wt , Eco CP RPR delP18 and Eco CP RPR delP18P3Mini ).
Determination of the kinetic constants for cleavage of pATSerUG without C5 corroborated the data presented in Table 1 and revealed no change in either k obs or K sto when P18 in Eco CP RPR was deleted ( Table 2; cf. values for Eco CP RPR wt and Eco CP RPR delP18 ). These data are in contrast to full-size Eco RPR were deleting and "disrupting" the P8/ P18-interaction affected k obs and k obs /K sto , respectively (see above; Table 2). Of note, the K sto values (%K d , see above) for Eco RPR P18CUUG and Eco CP RPR wt (or Eco CP RPR delP18 ) only differed by a factor of two ( Table 2).
Taken together, these data emphasized the importance of the S-domain and the P8/ P18-interaction for catalysis and substrate binding for full-size Eco RPR while P18 does not contribute to the catalytic performance of Eco CP RPR to any significant extent. However, the presence of P18 in full-size Eco RPR that cannot properly interact with P8 does affect pAT-SerUG binding whereas its absence does not (see Discussion). Also, comparing k app values (Table 1) for Eco CP RPR delP18 and Eco CP RPR delP18P3Mini suggested that the length of P3 does not appear to influence the catalytic performance in an Eco CP RPR context.

The C-domain derived from Pyrococcus furiosus (Pfu) is catalytically active in the absence of the S-domain and protein.
The type A Pfu RPR lacks P18 (Fig 2C and S2 Fig) and it is catalytic also in the absence of the S-domain but only in the presence of proteins ( [6];  (lane 9). The reaction times were 20 min for Eco CP RPR wt and Eco CP RPR delP18 , 60 min for Eco CP RPR 31 , Eco CP RPR 31delP18 and Eco RPR delP18 . Controls, incubation of pATSerUG alone without RPR (lane 1), cleavage of pATSerUG (lane 2) and pATSerCG GAAA (known to cleave at +1 and -1, see 18 Wu et al. 2011;lane 8) with Eco RPR wt . S = substrate, 5'-L = 5' cleavage fragments and +1 and -1 marks cleavage sites (see text for details). https://doi.org/10.1371/journal.pone.0192873.g006 RNase P RNA mediated cleavage and RNA processing see Discussion). Full-size Pfu RPR wt alone cleaves the model hairpin substrates used above at high Mg 2+ concentration [31]. To test whether Pfu RPR wt is catalytically active also without the S-domain and protein we generated Pfu CP RPR wt (Fig 2C). Indeed, Pfu CP RPR wt cleaved pATSerUG, pATSerCG, pMini3bpUG and pMini3bpCG mainly at the correct position +1 (Fig 7, cf. lanes 9, 13, 17 and 21). In addition, Pfu CP RPR wt cleaved pMini3bpCG at the alternative site -1 (Fig 7, lane 21) while we could not detect any cleavage of pATSerCG at -1 ( Fig  7B, lane 17). However, this could be because Pfu CP RPR wt cleaved pATSerCG with a very low efficiency such that cleavage at -1 could not be detected and quantified.  The rate of cleavage (k app ; Table 1) for Pfu CP RPR wt (without protein) was % 7-fold lower than for Pfu RPR wt while determination of k obs and K sto revealed that both were affected %four-to five-fold, respectively, resulting in a 17-fold reduction in k obs /K sto (Table 2). This is a significant lower reduction compared to the 500-fold drop in k obs /K sto in response to deleting the S-domain in the Eco RPR system (Table 2; see also Refs [20,23]).
We conclude that the S-domain is not essential for cleavage in the Pfu RPR alone reaction and as such supporting that the C-domain is responsible for catalysis also in the case of type A archaeal RNase P (see also Refs [11,24,32,42]). However, the S-domain boosts the catalytic performance but to a lesser extent than for Eco RPR (see Discussion).
The absence of the S-domain or disruption of the P8/ P18-interaction affects the charge distribution at and in the vicinity of the cleavage site. A correct TSL/ TBS-interaction leads to efficient and correct cleavage [16,19,23]. Moreover, cleavage of pATSer derivatives in which the 2'OH at position -1 in the substrate had been replaced with 2'NH 2 showed that the frequency of cleavage at -1 is reduced with increasing pH. Most likely this is because the 2'NH 2 becomes protonated and positively charged with decreasing pH, thereby reducing cleavage at the canonical site +1 [38,54]. The shift of the cleavage site is also dependent on the structural topography of the +1/+72 base pair in the substrate. We have argued that this is due to a change in the charge distribution at the cleavage site in the RPR-substrate complex ( [37]; however, see Ref [55] for an alternative model). Hence, to test whether the absence of the S-domain and disruption of the P8/ P18-interaction influence the charge distribution/ protonation near the cleavage site in the RPR-substrate complex we studied the cleavage pattern of the pAT-SerUG variant pATSerU am G, in which the 2'OH was replaced with 2'NH 2 at -1, at different pH (Fig 5; see Materials and methods).
The cleavage frequency of pATSerU am G at +1 increased with increasing pH for all the RPRs variants as expected from our previous data (Fig 8 and S3 Fig). However, compared to Eco RPR wt and Eco CP RPR wt the trend was that higher pH was required to reach 50% cleavage at +1 for the other RPR variants (including the Pfu RPR variants). For the all-ribo substrate pATSerUG the site of cleavage did not change with pH irrespective of RPR variant (S3 Fig). We also inquired if a structural change in the TBS region (in the vicinity of where P18 contact P8; Fig 2B) in the S-domain affects cleavage of pATSerU am G as a function of pH differently compared to Eco RPR wt . Hence, we examined the cleavage pattern for Eco RPR G235 at different pH. This change in the RPR is known to influence cleavage site recognition (Fig 2B; [16,19,23]). Again a higher pH was needed to give 50% cleavage at the +1 position compared to Eco RPR wt (Fig 8).
To conclude, we interpret these data to suggest that the S-domain, the P8/ P18-interaction and the structural topology of the TBS region influence the pKa of the 2'NH 2 and/or the charge distribution at the cleavage site.

Importance of the P6-and P8/ P18-interactions
The P6-and P8/ P18-interactions play important structural roles in folding the RPR where P6 is an intra C-domain interaction while P8/ P18 is involved in connecting the S-and Cdomains ( Fig 2B). However, information of their impact on the structure and function of RPRs lacking the S-domain, e.g. Eco CP RPR wt , is scarce. In this context, type T archaeal RPRs are equipped with a degenerated S-domain and secondary structure modeling suggests the presence of P6 but its contribution to catalysis has not been studied. As for Pfu RPR, P18 is also absent in type T RPRs [6,12,13]. Studying RNase H accessibility and cleavage of the model hairpin loop substrates pATSerUG we provide data suggesting that residues 5'G 276 C 277 C 278 C279 are likely engaged in pairing with residues 5'G 82 G 83 G 84 C 85 in the absence of the S-domain (Fig 2A). We refer to this interaction as the "P6-mimic" and our data indicate that its presence contributes to the catalytic performance by Eco CP RPR. As such, our findings also provide support for the existence and functional importance of P6 in type T RPR. Moreover, deleting the S-domain in the type B B. subtilis RPR decreased the cleavage rate %25000-fold [21]. This is in contrast to the 120-fold reduction in the rate observed for the type A Eco RPR (Table 1; see also Refs [22,23]). Type B lacks P6, however, recent data indicate that disruption of the intra-domain interaction between L5.1 and L15.1 in the C-domain affects both folding and the catalytic activity in a full-size RPR context [27]. Hence, it will be of interest to understand whether the L5.1/ L15.1 interaction (or a mimic) is present in the absence of the S-domain (see also Ref [21]). If this is the case we predict that it contributes to catalysis by the type B RPR lacking the S-domain. In contrast to disruption of the "P6-mimic", removal of P18 did not result in any significant change in the catalytic performance for Eco CP RPR (Table 2). However, disruption of the P8/ P18-interaction, or deletion of P18, in full-size Eco RPR reduced cleavage of pAT-SerUG and affected the overall structure to a certain degree. While disrupting the P8/ P18 (P18 still present) influenced both the kinetic constants (k obs and K sto ; 24-and %10-fold change, respectively, compared to Eco RPR wt ; Table 2) deleting P18 resulted in a dramatic decrease in k obs (>3000-fold) in cleavage of the model substrate pATSerUG while no change in K sto was detected. Given that K sto % K d this might indicate that P18 interferes with binding of pATSerUG when its interaction with P8 is disrupted while this is not the case in its absence. Nevertheless, previous multiple turnover kinetic studies using pre-tRNA and pre-4.5S RNA reported that disruption of P8/ P18 or deletion of P18 affects binding affinity (K m ) and k cat in cleavage of pre-tRNAs with type A RPR with and without protein. The levels of change differ comparing our single turnover data and previously reported results ( [28][29][30]; see below). This is likely due to different reaction conditions, choice of substrate and RPR (Eco RPR and Thermus thermophilus, Tth, RPR). In this context we also note that earlier data suggested that the impact of deleting P18 is suppressed by raising the ammonium concentration to 3 M [29]. We conclude that irrespective of substrate presence of P18 and the P8/ P18-interaction have an impact on the catalytic performance by bacterial type A RPR while in the absence of the S-domain (and the P8/ P18 interaction), as in Eco CP RPR, P18 has no significant impact on catalysis.
Calculating the ΔΔG using k obs /K sto values [52] revealed that the contribution of the P8/ P18-interaction is between 3.1 and 5 kcal/mol for full-size Eco RPR while the contribution of the S-domain is approximately 3.8 kcal/mol ( Table 2). Extracting and using the k cat /K m (= k obs /K sto ) values (for type A Eco and Tth RPRs; Table 2) from previous reports [28][29][30] to calculate the ΔΔG values suggest that the contribution of P18 and the P8/ P18-interaction varies between 1 and 1.8 kcal/mol. In the Eco RPR pre-tRNA-system without the C5 protein disruption of the P8/ P18-interaction resulted in a loss of 1.5 kcal/mol [30], which is two-fold lower than the value obtained using pATSerUG ( Table 2). The reason for this discrepancy could be due to the difference in reaction conditions (e.g., here we used higher [Mg 2+ ]) and/or the way the two substrates interact with the RPR where pre-tRNA has a structurally intact TSL-region. The importance of P18 in the full-size RPR context can be rationalized by its structural role in connecting the S-and C-domains and structurally orient these domains in a productive/ correct manner as discussed by Li et al. ([24]; Figs 2 and 3). This together with that a productive TSL/ TBS-interaction in the S-domain affects catalysis [16,19,23,31,34] opens for the possibility that the P8/ P18-interaction acts as a structural mediator in the "communication" between TSL/ TBS-interaction and the cleavage site leading to positioning of chemical groups and Mg 2+ that result in correct and efficient cleavage. Consistent with this is that disruption of the P8/ P18-interaction, removal of the S-domain (and P18) as well as alteration in the vicinity of the structure were P18 connects (as in the Eco RPR G235 variant; Fig 2B; Ref [19]) seems to influence the charge distribution in the vicinity of the cleavage site (Fig 8). That the P8/ P18-interaction influences events at the cleavage site is also supported by data using a derivative of pATSerCG in which the loop had been replaced with a GAAA-tetra loop (pAT-SerCG GAAA ; Fig 5). Eco RPR wt cleaves this substrate preferentially at -1 (81±2%) and as we reported elsewhere this is most likely due to the absence of a productive/ correct TSL/ TBSinteraction [16,19]. For Eco RPR P18CUUG (disrupted P8/ P18-interaction) we observed a lower but reproducible cleavage (68±2%) at the alternative site -1 while absence of the S-domain (and the P8/ P18-interaction) results in cleavage preferentially at +1 [23]. Taken together, the type A Eco RPR S-domain and P8/ P18-interaction play important roles for the catalytic performance and site selection. Moreover, since removing P18 in Eco CP RPR did not affect cleavage (or the structure to any significant extent) it is likely that P18 itself does not influence catalysis but the P8/ P18-interaction does, consistent with that Eco RPR delP18 and Eco RPR P18CUUG are poor catalysts in cleaving pATSerUG (Table 2). But noteworthy, the presence of P18 that cannot interact with P8 does affect pATSerUG binding and the reason to this is unclear (however see above). In this context, our unpublished structural probing data of fullsize Eco RPR variants suggest that substitution of A248, which is positioned close to the cleavage site in the RNase P-tRNA structure [18], influence the structure of P18.

Comparing the type A Eco and Pfu RPRs
Our data show that the Pfu RPR retained its catalytic activity upon removing the S-domain. Tsai et al. [6] provided data where they showed that a Pfu CP RPR construct is indeed catalytic however only in the presence of proteins. A rational for that they did not detect any cleavage activity without proteins might be differences in reaction conditions. We used a significantly higher Mg 2+ -concentration, which has been reported to increase low or unnoticed cleavage efficiency by Eco RPR variants [22,34]. In addition, the choice of substrate differs in these two studies, pre-tRNA vs. pATSer, which might also be a factor. It should also be noted that our data are in accordance with that the activity of the archaeal type A M. thermoautotrophicus RPR substantially increases when its C-domain is linked to the Eco RPR S-domain [24].
Comparison of our current data, where we removed the S-domain of Pfu RPR wt , with our previous data [31] shows an effect on both the kinetic constant k obs and K sto in cleaving pAT-SerUG. This is similar to the situation for Eco RPR but the magnitude of change in k obs for Eco RPR was higher (cf. 35-fold vs. 3.6-fold in the case of Pfu RPR; Table 2). Using the k obs /K sto values for Pfu RPR wt and Pfu CP RPR and calculating ΔΔG gives a loss of 1.7 kcal/mol as a result of removing the S-domain (Table 2). This should be compared to the 3.8 kcal/mol loss seen for Eco RPR wt (see above). Full-size Eco RPR wt can form a productive/ correct interaction with the substrate TSL-region while Pfu RPR wt in the absence of proteins interacts differently with TSL [31]. Together this suggests that in the RPR alone reaction the Pfu S-domain plays a less important role in cleavage of the model hairpin loop substrate pATSerUG than in the Eco RPR case. Similar reduction in the rate (12-fold) was also observed upon removing the S-domain in a cis RPR-pre-tRNA construct based on the archaeal M. jannaschii type M system [42]. In this context, as reported previously the S-domain of type A archaeal RPR appears to hamper the activity in the RNA alone reaction [24,32]. These authors also provided data where structural changes improved the activity of the type A archaeal Methanothermobacter thermoautotrophicus RPR and their findings are likely to be applicable to rationalize the difference in activity comparing Eco and Pfu RPRs.
Taken together, for Pfu RPR the influence of the S-domain is perhaps reflected by the fact that two of the five RNase P proteins, Rpp21 and Rpp29, bind to the S-domain and influence the interaction with the T-loop region of pATSer model substrates [31]. Moreover, considering the absence of P18 in archaeal type A RPRs it has been discussed that its role has been taken over by the Pop5 and Rpp30 proteins [24]. In this context we emphasize that removal of P18 in full-size Eco RPR reduced k obs to a level %10-fold lower compared to k obs as determined for Pfu RPR wt and a loss of 5 kcal/mol relative to Eco RPR wt , one kcal/mol lower compared to Pfu RPR wt Table 2; both Eco RPR delP18 and Pfu RPR wt lack P18). To conclude, combined these data raise the question whether the evolution of a more complex RNase P in terms of the number of protein subunits is linked to a decrease in the contribution of the S-domain to catalysis see also Refs [24,32].