Bacterial/archaeal protein-only RNase P: complementation in Escherichia coli uncovers coevolution of protein-only RNase P and precursor tRNA structures in Aquifex aeolicus

Complementation of the E. coli RNase P RNA depletion strain by bacterial/archaeal protein-only RNase P

We analyzed one archaeal and two bacterial HARPs for their ability to restore growth of E. coli BW cells under nonpermissive conditions where the levels of endogenous RNase P are depleted to an extent that impedes cell growth. All three HARPs from A. aeolicus (Aae), T. indicus (Tin), and M. thermoautotrophicus (Mth) basically showed complementation activity in plating assays (Nickel et al. 2017). In liquid cultures in the presence of glucose, where the P_BAD promoter of the chromosomal rnpB gene is silenced, the HARP complementation strains grew markedly slower than the BW strain transformed with the plasmid expressing the E. coli rnpB gene. However, this phenotype differed from that of BW cells transformed with the empty pDG148 vector for which growth completely stagnated after 240 min (Fig. 1). At an OD₆₀₀ of 0.5–0.75 in glucose medium, cells were harvested to prepare total RNA for RNA-seq, northern blot, and primer extension analyses. The early exponential phase was chosen to focus on primary effects in the (tRNA) transcriptome upon depletion of endogenous RNase P.

RNA-seq analysis of E. coli cells

The tRNA-seq data were evaluated according to cleavage precision (Fig. 2) and cleavage efficiency (Fig. 3), as detailed under Materials and Methods. In brief, reads starting at the native mature tRNA 5′-end +1 (at position −1 for tRNA^His and tRNA^Sec) were classified as “canonically cleaved” and reads beginning at −1 (at +1 for tRNA^His and tRNA^Sec) as “miscleaved.” The fractions of canonically and miscleaved tRNAs were plotted in Figure 2. For each tRNA covered by sufficient reads, cleavage efficiency was defined as the fraction of 5′-ends at −1 and +1 (assigned to canonical or aberrant RNase P processing) relative to the fraction of 5′-ends at position −2 and further upstream (representing pre-tRNAs not processed by RNase P). We did not find any evidence for RNase P cleavage between nt −3 and −2 to yield a 5′-end at position −2, which might have been expected for substrates such as pre-tRNA-Asp GUC 2/2 (aspV), which could form an acceptor stem that is extended by two additional G:C base pairs (Supplemental Fig. S1A). Pre-tRNA reads had to fulfill the condition that their 3′-end included at least nt +1 of the tRNA body. In Figures 2 and 3, only tRNAs covered by at least 100 reads per one million reads, summed up over all eight libraries, were considered. The classification modes for tRNA isoacceptors and their gene names in E. coli strain MG1655 (genome NCBI NZ_CP012868.1) are summarized in Supplemental Table S1. The sequences of E. coli tRNAs including 25 nt of 5′ precursor sequence, isoacceptor alignments, and potential acceptor stem extensions are shown in Supplemental Figure S2, and the organization/context of tRNA genes in E. coli is illustrated in Supplemental Figure S3.

For the E. coli BW strain expressing the plasmid-borne rnpB gene (cyan squares in Fig. 2), 5′-ends corresponding to canonical cleavage made up >95% of the read 5′-ends for almost all tRNAs, except for tRNA-Val (UAC 2/3, valU), where the read profile indicated ∼20% noncanonical cleavage at the c ₋₁ site (between nt −2 and −1; see bottom of the diagram in Fig. 2 as well as Supplemental Fig. S4). This may be related to the stretch of five G residues from position −2 to +3 enabling extension of acceptor stem pairing by two G:C bp in the precursor (with the discriminator A residue bulged out), combined with the presence of two consecutive CCA motifs (Supplemental Fig. S1B).

For the BW strains complemented with harp genes, the fractions of 5′-ends assignable to miscleavage were in many cases slightly or moderately increased (Fig. 2, red, green, and purple symbols) relative to the rnpB complementation strain (cyan symbols), with considerable variation between E. coli cells expressing Aae (red), Tin (purple), or Mth (green) harp. However, a conspicuous feature was the particularly imprecise cleavage of tRNAs Glu(UUC) 1/2, Arg(UCU), and Ser(CGA) by the strains expressing Aae and Tin harp. This trend was attenuated for the Mth harp complementation strain (Fig. 2). A common feature of the three aforementioned tRNA precursors is their capacity to form an acceptor stem extended by a C:G bp and the presence of an A residue at position −2 (Supplemental Fig. S5).

Cleavage efficiency was evaluated by the ratio of 5′-mature to 5′-precursor reads. In the BW strain complemented with E. coli rnpB, previously shown to grow as efficiently as its parental wild-type strain MG1655 under glucose conditions (Wegscheid and Hartmann 2006; Gößringer et al. 2017), the fractions of 5′-precursor reads were already substantial for many tRNA species (Fig. 3, cyan squares). Yet, these ratios do not necessarily mirror the precise intracellular ratio of 5′-mature to 5′-precursor tRNAs. Owing to less efficient 5′-adapter ligation, the base-paired acceptor stems of 5′-mature tRNAs may have entered the RNA-seq libraries less efficiently than 5′-precursor tRNAs. Nevertheless, this issue is not critical here as long as cleavage efficiencies for the harp complementation strains are related to that of the strain complemented with E. coli rnpB. Cleavage efficiencies in the harp-expressing E. coli cells were impaired for many tRNAs, in several cases most pronounced for the strain expressing Mth Harp (Fig. 3). Overall, these findings are consistent with the retarded growth of the harp complementation strains upon endogenous RNase P depletion (Fig. 1). Evidently, inefficient and/or slow RNase P processing of a variety of E. coli tRNA precursors limited translation efficiency and thus growth of these strains under nonpermissive conditions.

Analysis of 5′-end processing of selected RNase P substrates by northern blot and primer extension analysis

We next validated processing defects by northern blot and primer extension analyses of a subset of RNase P substrates (Supplemental Fig. S1A,C–F) for which the RNA-seq data (averaged on the results for the three harp complementation strains) indicated reduced cleavage precision [Glu(UUC) 1/2, Ser(CGA), Thr(UGU)] (Fig. 2), and/or reduced cleavage efficiency [Asp(GUC) 1/2, Glu(UUC) 2/2, His(GUG), Asp(GUC) 2/2, Ser(CGA), Thr(UGU)] (Fig. 3). Acceptor stems of E. coli tRNA-Glu(UUC) 1/2 precursors encoded by the genes gltU and gltV as well as those of tRNAs Ser(CGA) and His(GUG) are extended by 1 C:G bp (Supplemental Fig. S1C–E). Acceptor stems of tRNA-Glu(UUC) 2/2 precursors encoded in genes gltT and gltW are extended by 2 bp (U:G and G:C; Supplemental Fig. S1C). Similarly, acceptor stems of tRNA-Asp(GUC) precursors encoded by the gene aspV are extended by 2 G:C bp and those transcribed from genes aspT and aspU by 3 bp (U:G and 2 G:C bp; Supplemental Fig. S1A). An additional U:A bp can form in the acceptor stem of the Thr(UGU) precursor (Supplemental Fig. S1F). In addition, we included tRNA Sec(UCA) (Supplemental Fig. S1G), because this substrate is naturally cleaved aberrantly between nt −1 and −2 to leave an 8 bp long acceptor stem as in the case of tRNA^His (Burkard and Söll 1988; Gößringer et al. 2017). Moreover, we analyzed the noncoding 4.5S RNA (Supplemental Fig. S1H) that is a substrate for E. coli RNase P (Bothwell et al. 1976; Peck-Miller and Altman 1991). 4.5S RNA (also termed SRP RNA) is a component of the signal recognition particle that targets ribosomes to the plasma membrane (Ulbrandt et al. 1997; Herskovits et al. 2000). This substrate was of interest to us, as it lacks a T-loop element. Unfortunately, no 4.5S RNA reads were present in our RNA-seq libraries, although (pre-)4.5S RNA was detectable in the primer extension assay (see below) using the same total RNA preparations as for the RNA-seq experiments. We assume that (pre-)4.5S RNA molecules were depleted at some step during library preparation, for example, in the course of small RNA (<200 nt) isolation from total RNA samples using the RNeasy Mini Kit (Qiagen) or because of inefficient adapter ligation.

In vitro processing analyses

E. coli pre-tRNAs Glu(UUC) 1/2, Arg(UCU) and Ser(CGA), miscleaved by the strains expressing Aae and Tin harp (Fig. 2), can form an extra C₋₁:G₇₃ bp and carry an A residue at position −2 (Supplemental Fig. S5). Likewise, E. coli pre-tRNA Thr(UGU) able to form a weaker U₋₁:A₇₃ bp but carrying an A at position −2 (Supplemental Fig. S5) was miscleaved to ∼50% at the c₋₁ site by Aae HARP (Figs. 2, 8). In view of these observations, we examined the base identity of residue −1 by in vitro processing analyses. Beforehand, we analyzed processing of a bacterial pre-tRNA^Gly with a 14 nt 5′-flank (5′-GGAUUUUCCCUUUC; Fig. 11A, structure on the left) next to a derived hairpin substrate (pAaT_9B) (Fig. 11A, central structure). The latter substrate was cleaved at 500 nM Aae HARP in single turnover reactions with a k_obs only ∼1.4-fold lower than that obtained for the parental pre-tRNA^Gly, showing its efficient utilization by Aae HARP and its value as a diagnostic HARP substrate. We then tested chemically synthesized pAaT variants with a 9 nt leader (5′-UUUUUUUUU; Fig. 11A, structure on the right), which differed in the presence and size of a bulge separating acceptor and T-stem on the 5′ side to confer different extents of flexibility to the acceptor/T-stem junction. This included variants that only differed by a U to A exchange at position −1. All substrates with a U₋₁ and 0, 3, 6, or 9 nt bulges (variants pAaT_−1U_0B/3B/6B/9B) were basically cleaved by Aae and Tin HARP, although at different rates (Fig. 11B,C; Supplemental Table S2). When we compared the 3- and 9-bulge substrates with U₋₁ versus A₋₁, we observed enhancements in the rate constant k_obs for the substrates with A₋₁: in the case of Aae HARP, the overall k_obs increased fourfold from 0.46 to 1.89 min⁻¹ (substrates with 3 nt bulge) and 1.9-fold from 0.88 to 1.67 min⁻¹ (substrates with 9 nt bulge); for Tin HARP, the corresponding increases were sixfold from 0.17 to 1.01 min⁻¹ and 1.4-fold from 0.28 to 0.38 min⁻¹ (Fig. 11B,C; Supplemental Table S2, see exemplary gels in Supplemental Fig. S12). Moreover, analysis of cleavage site selection revealed an accuracy switch toward preferred cleavage at the canonical c₀ site in the presence of A₋₁ versus U₋₁: from 78% to 93% (Aae HARP) and from 82% to 98% (Tin HARP) for substrates with the 3 nt bulge and from 74% to 92% (Aae HARP) and 82% to 97% (Tin HARP) for the substrates with the 9 nt bulge (Fig. 11B,C; Supplemental Table S3). We then analyzed for the substrates with the 9 nt bulge if the presence of an A at −1 is also favorable in the reaction catalyzed by the archaeal Mth HARP. Indeed, cleavage efficiency and selection of the canonical cleavage site also improved for substrate pAaT_−1A_9B relative to pAaT_−1U_9B (Fig. 11D; Supplemental Tables S2, S3).

Conclusions

In summary, it is remarkable that HARPs can restore at least slow growth of RNase P-depleted E. coli bacteria, considering that all three tested HARPs are from thermophilic to hyperthermophilic bacteria and have to assemble into tetra- to dodecameric homo-oligomers for activity in the fast-growing mesophilic, nonnatural E. coli host. Our complementation analysis revealed that HARPs have difficulties to access the native cleavage site when acceptor stems are extended by G:C base pairs and lack an A residue immediately upstream of this site. As PRORPs, HARPs are essentially inactive on substrates lacking a T-loop element. This can be explained by their less comprehensive set of interactions in the vicinity of the cleavage site compared with RNA-based bacterial RNase P enzymes. Our study has further revealed that A. aeolicus as a representative of the bacterial family Aquificaceae underwent coevolution of its pre-tRNA ensemble and novel type of protein-only RNase P. Two evolutionary scenarios are conceivable. Either, for establishing the protein-only enzyme as the sole and efficient RNase P activity in the Aquificaceae, tRNA genes evolved toward reduction of stable acceptor stem extensions in pre-tRNA transcripts and enrichment of A residues upstream of 5′-maturation sites; these changes made docking of the scissile phosphodiester of pre-tRNA substrates in the metallonuclease active site more efficient. Not excludable is an alternative evolutionary scenario according to which an A upstream of the cleavage site became prevalent in the cellular pre-tRNA pool of the Aquificaceae progenitor for unknown reasons, which then favored HARP substitution of the RNA-based enzyme that acts inefficiently on pre-tRNAs with Y₋₂ A₋₁ as inferred from the E. coli system (Fig. 11E; Zhu et al. 2022).

Construction of expression plasmids

The harp gene (MTH_RS01110) of M. thermautotrophicus strain Delta H (NCBI Reference Sequence: NC_000916.1) was PCR-amplified from chromosomal DNA isolated with the QIAamp DNA Mini Kit (Qiagen) using the extraction protocol for Gram-positive bacteria and the primer pair Mth(a)-for and Mth(a)-rev (Supplemental Table S4). The harp gene (THEIN_RS10515) of T. indicus DSM 15286 (NCBI RefSeq: NC_015681.1) was PCR-amplified from chromosomal DNA (DSM 15286 from DSMZ, Germany) using the primer pair Tin(a)-for and Tin(a)-rev (Supplemental Table S4). Both genes were inserted into vector pET28a(+) via the restriction enzymes NcoI and XhoI. For complementation studies in E. coli strain BW, the MTH_RS01110 gene was PCR-amplified from the pET28a(+)-based vector using the primer pair Mth(b)-for and Mth(b)-rev (Supplemental Table S4) and inserted into the shuttle vector pDG148(S/X) (Gößringer and Hartmann 2007; Gößringer et al. 2017) using the SmaI and XbaI sites. Likewise, the THEIN_RS10515 gene was amplified via primer pair Tin(b)-for and Tin(b)-rev (Supplemental Table S4) from the pET28a(+)-based vector and inserted into shuttle vector pDG148(S/X) using the SmaI and SphI sites. For complementation studies, we further used the expression vector pACYC177_E. coli_rnpB_wt (Wegscheid and Hartmann 2006) encoding the E. coli RNase P RNA and pDG_aq880 (Nickel et al. 2017) encoding A. aeolicus RNase P.

Bacterial strains and cell growth

Complementation studies were performed in E. coli strain BW that expresses the chromosomal rnpB gene under control of an arabinose-dependent promoter and repressor system (Wegscheid and Hartmann 2006). In the absence of arabinose and presence of glucose as carbon source, growth of the E. coli BW strain is arrested owing to endogenous RNase P depletion. Only plasmid-encoded expression of a functional rnpB gene or protein-only RNase P restores growth in the absence of arabinose. For this purpose, either a pACYC177 derivative expressing the E. coli rnpB gene or derivatives of plasmid pDG148(S/X) expressing the individual harp genes were introduced into E. coli BW by heat shock of chemically competent cells using a transformation and storage solution (TSS) (Chung et al. 1989). After transformation, selected colonies were verified and analyzed by PCR amplification of the plasmid-encoded RNase P gene (using flanking, vector-specific primers) and plating of cells on LB agar plates containing 100 µg/mL ampicillin, 34 µg/mL chloramphenicol, and either 10 mM arabinose or 10 mM glucose. For RNA-seq analysis, cells were grown as described in the following. Single colonies of strains BW[pACYC177_E. coli_rnpBwt], BW[pDG_aq880], BW[pDG_MTH_RS01110], and BW[pDG_THEIN_RS10515], expressing plasmid-encoded E. coli RNase P RNA, A. aeolicus (Aae) HARP, T. indicus (Tin) HARP or M. thermautotrophicus (Mth) HARP, were used to inoculate precultures of 5 mL LB medium supplemented with 10 mM arabinose, 100 µg/mL ampicillin, and 34 µg/mL chloramphenicol. After growth for 5–6 h at 37°C and 200 rpm, an aliquot (≤0.5 mL) of each preculture was used to inoculate 200 mL fresh LB medium supplemented with 10 mM glucose, 100 µg/mL ampicillin, and 34 µg/mL chloramphenicol to the starting OD₆₀₀ = 0.01–0.03. Cells were grown at 200 rpm and 37°C to an OD₆₀₀ of ∼0.5–0.6 (0.75 for replicate 2 of strain BW[pACYC177_E. coli_rnpB_wt]; see Fig. 1), corresponding to ∼ 5 ± 1 cell divisions. Aliquots of OD₆₀₀ = 10 were then withdrawn and centrifuged for 30 min at 4800g and 4°C. The resulting cell pellets were frozen in liquid nitrogen and stored at −80°C until RNA preparation.

RNA preparation

Preparation of total RNA was performed basically by the hot phenol method as described in Damm et al. (2015). In detail, each cell pellet derived from 10 OD₆₀₀ was resuspended in 1.6 mL ice-cold extraction buffer (10 mM NaOAc, 150 mM sucrose, pH 4.8) and equally distributed to two 2 mL tubes. To each 800 µL sample, 75 µL lysozyme solution (20 mg/mL in 1× TE buffer) was added, followed by incubation for 10 min at room temperature (21°C–23°C). Then, 40 µL of 20% (w/v) SDS solution was added, and components mixed by pipetting up and down. Cells were completely lysed by incubation for 90 sec in a 65°C water bath before adding 800 µL acidic phenol (preheated to 65°C). After thorough vortexing for 15 sec, repeated inversion of reaction tubes and vortexing for another 15 sec, samples were successively incubated for another 4 min at 65°C, immediately shock-frozen shortly in liquid nitrogen, thawed for 2 min at 65°C in the water bath, vortexed for 20 sec, and centrifuged for 10 min at 15,700g. The upper aqueous phase was carefully transferred to a new 2 mL tube, and the entire phenol extraction procedure was repeated for another two rounds. To each of the aqueous phases obtained from the three phenol extractions, 800 µL chloroform was added, and samples were vortexed for 30 sec and centrifuged for 5 min at 15,700g. Each aqueous phase was then carefully transferred to a 1.5 mL tube, and the RNA was precipitated by addition of 0.8 volumes of isopropanol and 0.1 volume of 3 m NaOAc, pH 5.0, followed by storage for 1 h at −20°C. After centrifugation for 30–40 min at 15,500g and 4°C, the RNA pellet was washed with 75% (v/v) EtOH, air-dried and redissolved in 20–30 µL ddH₂O.

Northern blot analysis

Northern blots were basically performed as described in Gößringer et al. (2017). Briefly, total RNA samples (8–20 μg per lane) were mixed with 1 volume 2× denaturing loading buffer (0.02% [w/v] bromophenol blue, 0.02% [w/v] xylene cyanol blue, 8 m urea, 50% [v/v] formamide, 2× TBE [see below]) and heated for 3 min at 95°C before separation on 10% (w/v) polyacrylamide gels containing 8 m urea. After separation, the RNA was electro-blotted onto positively charged nylon membranes (Roche Diagnostics) for 140 min at 0.7 mA/cm² in 0.5× TBE (44.5 mM Tris; 44.5 mM boric acid; 1 mM EDTA) transfer buffer. The hybridization and detection procedures were carried out according to the DIG RNA Labeling Mix protocol (Roche Diagnostics). For the detection of tRNA^Asp, tRNA^His, tRNA^Sec, and 5S RNA (rrf), internally digoxigenin-labeled antisense transcripts complementary to the full-length RNAs were prepared. First, the genes were PCR-amplified from chromosomal DNA of E. coli K12 using forward primers containing a T7 promoter sequence to enable the synthesis of antisense transcripts (see Supplemental Table S4 for oligonucleotide and primer sequences). Next, the gel-purified (QIAEX II, Qiagen) DNA fragments were used as templates for in vitro transcription in the presence of digoxigenin-11-UTP (Roche Diagnostics) using T7 RNAP. For the detection of tRNA^Glu, tRNA^Ser, and tRNA^Thr, isoacceptor-specific DNA probes containing a digoxigenin label at the 5′-end (metabion) were used (see Supplemental Table S4 for sequences). After hybridization with the tRNA-specific probes at 50°C–72°C, immunodetection, and documentation, the nylon membranes were washed successively with stringency buffer I and II (Roche Diagnostics) for 20 min at 40°C and 30 min at 80°C. After stripping of the tRNA-specific probes, the membranes were additionally hybridized at 68°C with the 5S rRNA-specific probe to control for transfer efficiency of RNA onto the nylon membrane and fluctuations in RNA amounts loaded onto individual gel lanes.

Primer extension analysis

Primer extension analysis was performed with 4–20 µg total cellular RNA prepared by the hot phenol method as described (see above; Gößringer et al. 2017). The 5′-³²P-end-labeled primers were specific for mature E. coli 4.5S RNA (annealing to nt 20 to 3 of mature 4.5S RNA), tRNA^Asp(aspT/U/V) (annealing to nt 22 to 4 of mature tRNA^Asp), tRNA^Glu(gltUV) (annealing to nt 21 to 4 of mature tRNA^Glu), tRNA^His(hisR) (annealing to nt 27 to 5 of mature tRNA^His), tRNA^Sec(selC) (annealing to nt 23 to 5 of mature tRNA^Sec); tRNA^Ser(serU) (annealing to nt 22 to 4 of mature tRNA^Ser), or tRNA^Thr(thrU) (annealing to nt 28 to 4 of mature tRNA^Thr). The primers were annealed to total RNA (3–20 µg) in 10 μL 1× cDNA synthesis buffer applying a linear temperature gradient (cooling rate: 0.6°C per min) from 90°C to 55°C. Reverse transcription was performed by adding 5 µL RT-mix (1× cDNA synthesis buffer [50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgCl₂], 100 mM DTT, 10 mM dNTPs and ThermoScript or SuperScript III polymerase, Thermo Fisher Scientific Invitrogen) and incubation for 50 min at 55°C (detection of tRNA^Asp, tRNA^Glu, tRNA^His, tRNA^Sec, tRNA^Thr, and 4.5S RNA) or 62°C (detection of tRNA^Ser). The primer extension products were analyzed by 20% (w/v) PAGE in the presence of 8 m urea. As size markers, we used 5′-³²P-labeled DNA primers mimicking the extension product on the respective tRNA (or 4.5S RNA) cleaved at the c₀ site (cleavage between nt −1 and +1), the c₋₁ site (cleavage between nt −2 and −1), or c₋₂ site in the case of tRNA^Asp (see Supplemental Table S4 for primer sequences).

RNA-seq–Illumina sequencing and data evaluation

RNA-seq experiments were performed in two biological replicates each. Small RNA species (<200 nt) were isolated from total RNA samples using the RNeasy Mini Kit (Qiagen). Ribosomal RNA molecules were depleted using an in-house developed protocol (Vertis Biotechnologie AG). Oligonucleotide adapters were ligated to the 5′- and 3′-ends.

First-strand cDNA synthesis was performed using M-MLV reverse transcriptase and the 3′-adapter as primer. The resulting cDNA was PCR-amplified using a high-fidelity DNA polymerase (14 cycles). Purification was done using the Agencourt AMPure XP kit (Beckman Coulter Genomics). Samples were pooled in equimolar amounts and sequenced using an Illumina NextSeq 500 system with 75 bp read length. The raw sequencing files from this study are available at the NCBI Sequence Read Archive (SRA) under BioProject PRJNA1196355, BioSample accession SAMN45428597.

RNA-seq reads were trimmed using cutadapt (v1.18) with enabled quality trimming of reads using two-color chemistry (threshold Q28). Adapter sequences were 5′-AAT GAT ACG GCG ACC ACC GAG ATC TAC AC (5′-adapter) and 5′-ACA CTC TTT CCC TAC ACG ACG CTC TTC CGA TCT (3′-adapter). Reads with a length smaller than 16 nt or a fraction of undefined (N) bases above 10% were discarded. Infernal (v1.0) was used to annotate tRNAs in the genome of E. coli substr. K-12 substr. MG1655 (CP012868.1) based on covariance models provided by RFAM (release 14.1). The annotated tRNA sequences including 60 nt upstream were extracted and used to map RNA-seq data with segemehl (v0.3.4, parameters: -minsize 16 -E 0.001). Only perfect matches were kept (CIGAR string regex: /^[0-9]+=$/) whose 3'-ends at least overlapped the 5′-terminal nucleotide of the mature tRNA body.

Processing of tRNAs was evaluated according to Gößringer et al. (2017). However, the adjustment of ambiguously mapping reads did not include experimentally determined molar ratios of tRNA isoacceptor transcripts expressed from multiple gene loci as done previously (Gößringer et al. 2017), but such reads were equally distributed among the mapping targets. For example, reads mapping to three tRNAs with identical sequences (in the target region) are counted as the number of reads divided by three for each of them. This modification vastly simplified the analysis (relative to the protocol applied in Gößringer et al. 2017) with only minor changes to the results (data not shown). To focus on relevant tRNAs, we further limited the analysis to molecules with a total number of reads of at least 100 cpm (counts per million) summed up over all eight libraries (corresponding to two replicates each of libraries generated from the four E. coli BW complementation strains). Reads starting at the 5′-end position +1 (at position –1 for tRNA^His and tRNA^Sec) were classified as “canonically cleaved,” and reads beginning at –1 (at +1 for tRNA^His and tRNA^Sec) were classified as “miscleaved.” In addition, cleavage efficiency was measured based on the fraction of 5′-ends derived from (RNase P) processing at the c₀ or c₋₁ site (5′-end at +1 or –1) compared to the fraction of reads with 5′-ends at position −2 and further upstream (assumed not to be cleaved by RNase P). Figures were implemented using the R package ggplot2. The tools and results generated for this study can be found in the uni marburg GitLab repository (https://gitlab.uni-marburg.de/synmikro/ag-lechner/e-coli-k12-complementation-harp).

RNA substrates and 5′-[³²P]-end-labeling

The pAaT substrate RNAs were chemically synthesized by Axolabs GmbH. The substrates were designed based on Thermus thermophilus tRNA^Gly and its minimized AaT derivative (Brillante et al. 2016). The homouridine sequence of the 9 nt 5′-leader was designed for the purpose of RNA:protein cross-linking studies, and the discriminator U was changed to C to avoid base-pairing with an A at position −1. For 5′-[³²P]-end-labeling, 30 pmol of RNA (carrying a 5′-OH end) were incubated with 10 U T4 PNK (Thermo Fisher Scientific) and 2 μL [γ-³²P]-ATP (10 mCi/mL; 3000 Ci/mmol) in T4 polynucleotide kinase buffer A (50 mM Tris-HCl, 10 mM MgCl₂, 5 mM DTT, 0.1 mM spermidine, pH 7.6) for 60 min at 37°C. The reaction was stopped by addition of 2× RNA sample buffer (2× TBE buffer, 2.6 m urea, 66% deionized formamide, 0.02% [w/v] bromophenol blue, and 0.02% [w/v] xylene cyanol blue), and the sample was loaded onto an 8% (w/v) PAA gel containing 8 m urea for gel purification. After electrophoresis, the radioactively labeled RNA was localized by autoradiography and excised for elution in 500 µL 1 m NaOAc (pH 4.9) and, 0.05% (w/v) SDS at 8°C overnight under shaking (750 rpm, ThermoMixer). After addition of 2 µL GlycoBlue (15 µg/µL, Thermo Fisher Scientific Invitrogen) as coprecipitant, the eluate was concentrated by ethanol precipitation. The RNA pellet was air-dried, dissolved in 10 µL 1× TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5), shock-frozen in liquid nitrogen, and stored at −20°C.

Preparation of recombinant Aq880 and T. indicus HARP enzymes

Recombinant proteins (with C-terminal His tag) were prepared by Ni-NTA affinity chromatography (HisTrap HP, GE Healthcare) as described (Gößringer et al. 2023).

Preparation of E. coli RNase P RNA and E. coli and B. subtilis RnpA proteins

Preparation of E. coli RNase P protein with an N-terminal His-tag (MRGSHHHHHHGS, encoded in plasmid pQE-30 in E. coli JM109) was based on the previously described procedure (Rivera-León et al. 1995; Wegscheid and Hartmann 2006).

For preparation of B. subtilis RnpA, E. coli BL21(DE3)pLysS cells were transformed with plasmid pPWT1 (Niranjanakumari et al. 1998a) for protein expression. The RnpA protein was purified as described (Niranjanakumari et al. 1998a, Method 1 therein), with slight modification: the first chromatographic step using the weak cation exchange (HiTrap CM Sepharose FF, 5 mL) column was conducted in the presence of 10 m urea, followed by the second run on the same column in the absence of urea.

RNase P activity assays

Before activity assays, HARP working solutions were prepared from enzyme stock solutions by dilution in buffer F (50 mM Tris-HCl, pH 7.0, 20 mM NaCl, 4.5 mM MgCl₂, 20 µg/mL ultrapure BSA, and 5 mM DTT added immediately before use) and kept on ice before use. Activity assays with HARPs (carrying C-terminal His tags; final concentration 500 nM) were conducted as follows: The substrate solution (12 µL) containing 30,000 Cherenkov cpm of RNA substrate (∼1 pmol) in buffer F was preincubated for 5 min at 55°C und 20 min 37°C. The enzyme solution (16 µL) in the same buffer containing 625 nM recombinant Aae, Tin, or Mth HARP was preincubated for 5 min at 37°C. Then, 4 µL of substrate mix were added to 16 µL enzyme mix to start the reaction at 37°C (final concentrations: ∼3 nM substrate, 500 nM enzyme). Aliquots of 4 µL were withdrawn at different time points, mixed with 4 µL of ice-cold 2× denaturing loading buffer (0.02% [w/v] bromophenol blue, 0.02% [w/v] xylene cyanol blue, 2.6 m urea, 66% [v/v] formamide, and 2× TBE), and loaded onto a 20% (w/v) denaturing polyacrylamide gel (length: 41.9 cm; width: 34 cm; thickness: 1 mm) followed by electrophoresis for 4 h at 35 mA. For the “0 min” control lane, substrate solution only was incubated at 37°C for the same time period as the last aliquot withdrawn from the enzyme reaction tube. Then, 1 μL of this substrate solution was mixed with 3 μL ddH₂O and 4 μL of ice-cold 2× denaturing loading buffer for gel electrophoresis (see above). .

Assays with E. coli RNase P holoenzymes reconstituted with the homologous and the B. subtilis RnpA protein were conducted accordingly, with the following differences: Instead of using buffer F, substrate and enzyme solutions were prepared in buffer KN [20 mM HEPES-KOH, pH 7.4, 150 mM NH₄OAc, 2 mM spermidine, 50 μM spermine, 4 mM β-mercaptoethanol, and 4.5 mM Mg(OAc)₂ added immediately before use]. Substrate and RNase P RNA were preincubated separately for 5 min at 55°C and 15 min at 37°C; after the first 10 min at 37°C, RnpA protein was added to the RNase P RNA solution to form the holoenzyme complex during the last 5 min at 37°C. After combining substrate and enzyme solutions, final concentrations of the holoenzyme subunits were 100 nM RNase P RNA and 400 nM E. coli RnpA or 200 nM B. subtilis RnpA, respectively.

Substrate and 5′-leader product bands were visualized with a Bio-Imaging Analyzer FLA3000-2R (Fujifilm) and quantified by use of the AIDA software (Raytest). First-order rate constants of cleavage (k_obs) were calculated with Grafit 5.0.13 (Erithacus Software) by nonlinear regression analysis. In cases where substrates were cleaved at the c₀ and c₋₁ site, product formation was treated as the result of two independent parallel cleavage pathways (Cuzic-Feltens et al. 2009), deriving values for $f_{\mathrm{cleaved}{c}_0}$ by calculating $In{t}_{\mathrm{product}{c}_0}/(In{t}_{\mathrm{product}{c}_0}+In{t}_{\mathrm{product}{c}_{-1}}+In{t}_{\mathrm{substrate}})$, with f indicating the cleaved fraction and Int the respective band intensity at the respective kinetic time point; $f_{\mathrm{cleaved}{c}_{-1}}$ was calculated correspondingly.