Peptidyl tRNA Hydrolase Is Required for Robust Prolyl-tRNA Turnover in Mycobacterium tuberculosis

Abstract

Enzymes involved in rescuing stalled ribosomes and recycling translation machinery are ubiquitous in bacteria and required for growth. Peptidyl tRNA drop-off is a type of abortive translation that results in the release of a truncated peptide that is still bound to tRNA (peptidyl tRNA) into the cytoplasm. Peptidyl tRNA hydrolase (Pth) recycles the released tRNA by cleaving off the unfinished peptide and is essential in most bacteria. We developed a sequencing-based strategy called copper sulfate-based tRNA sequencing (Cu-tRNAseq) to study the physiological role of Pth in Mycobacterium tuberculosis (Mtb). While most peptidyl tRNA species accumulated in a strain with impaired Pth expression, peptidyl prolyl-tRNA was particularly enriched, suggesting that Pth is required for robust peptidyl prolyl-tRNA turnover. Reducing Pth levels increased Mtb's susceptibility to tRNA synthetase inhibitors that are in development to treat tuberculosis (TB) and rendered this pathogen highly susceptible to macrolides, drugs that are ordinarily ineffective against Mtb. Collectively, our findings reveal the potency of Cu-tRNAseq for profiling peptidyl tRNAs and suggest that targeting Pth would open new therapeutic approaches for TB. IMPORTANCE Peptidyl tRNA hydrolase (Pth) is an enzyme that cuts unfinished peptides off tRNA that has been prematurely released from a stalled ribosome. Pth is essential in nearly all bacteria, including the pathogen Mycobacterium tuberculosis (Mtb), but it has not been clear why. We have used genetic and novel biochemical approaches to show that when Pth levels decline in Mtb, peptidyl tRNA accumulates to such an extent that usable tRNA pools drop. Thus, Pth is needed to maintain normal tRNA levels, most strikingly for prolyl-tRNAs. Many antibiotics act on protein synthesis and could be affected by altering the availability of tRNA. This is certainly true for tRNA synthetase inhibitors, several of which are drug candidates for tuberculosis. We find that their action is potentiated by Pth depletion. Furthermore, Pth depletion results in hypersensitivity to macrolides, drugs that are not active enough under ordinary circumstances to be useful for tuberculosis.

Keywords: Mycobacterium tuberculosis; antibiotic resistance; bacterial genetics; ribosomes; tRNA; tRNA sequencing; translation.

FIG 1

Complementary techniques show that Pth is required for normal growth in Mtb. Growth curves are shown for different Pth knockdown (KD) constructs: proteolytic degradation (left) and transcriptional repression (right). In both panels, “(−)” indicates uninduced and “(+)” indicates induced for Pth knockdown. Growth was measured spectrophotometrically for each strain with OD₆₀₀ measurements over the course of 10 days in the presence or absence of an inducer. Prior to taking measurements, strains were diluted to the starting OD₆₀₀ indicated at day 0. Numbers next to the plasmid names indicate the promoter strengths for SspB expression (6 is lower and 10 is higher, corresponding to the expected relative levels of protein depletion). “pre-depletion” refers to growing cells in the presence of aTC for 3 to 4 days prior to diluting strains to the indicated starting OD₆₀₀s. Predepletion enables a higher level of pth knockdown to be achieved before conducting experiments. The results for both graphs are from two biological replicates.

FIG 2

Application of tRNA sequencing to measure pools of peptidyl and N-acetylated tRNAs. (A) Overview of the copper sulfate-based tRNA sequencing (Cu-tRNAseq) protocol. Strains are grown and harvested at a normalized cell density. Total RNA is subjected to chemical treatments (boxed) to distinguish between aminoacyl or uncharged tRNA and peptidyl or N-blocked (e.g., N-acetylated) aminoacyl tRNA (top and middle, respectively). A control treatment is performed for each replicate without CuSO₄ to measure total charged tRNA levels (bottom). tRNA fractions are extracted from 1 to 2 μg of total RNA for sequencing library preparation as described in Materials and Methods (26). (B) The ratio of 3′ A+C-ending reads to the total 3′ C+C- and 3′ A+C-ending reads for each tRNA is computed as described previously (27) to estimate the fraction of peptidyl (or N-acetylated) tRNA of each tRNA species. A higher fraction of A+C-ending reads following Cu-tRNAseq suggests a higher proportion of peptidyl or N-acetylated aminoacyl tRNA. (Figure created with BioRender.com.)

FIG 3

CuSO₄-tRNA sequencing (Cu-tRNAseq) distinguishes N-acetylated aminoacyl tRNA from aminoacyl or uncharged tRNA. Cu-tRNAseq was carried out on a previously described M. smegmatis (Msmeg) strain with inducible Mtb TacT overexpression (22). The fraction of peptidyl tRNA is plotted as a ratio of the number of 3′ CCA-ending reads to the sum of the 3′ CCA- and 3′ CC-ending reads for each tRNA species (Data Set S1). Cu-tRNAseq detected N-acetylated glycyl-tRNA under induced (TacT overexpression) conditions (dashed box). The results from one biological replicate are shown.

FIG 4

Depletion of Pth leads to the accumulation of peptidyl tRNAs in Mtb and a reduction in usable aminoacyl tRNA pools. Cu-tRNAseq was carried out in an Mtb CRISPRi Pth knockdown construct. (A) Strains were grown to an OD₆₀₀ of 0.3 to 0.4 in the presence or absence of the inducer (50 ng/mL aTC), and total RNA was harvested. The fraction of peptidyl tRNA is calculated as the ratio of the number of 3′ CCA-ending reads to the sum of the 3′ CCA- and 3′ CC-ending reads for each tRNA species. The bar plot at the top right displays the ratios of peptidyl tRNA fractions between induced and uninduced strains. Values from two biological replicates are shown. (B) Mtb RNA samples were treated with periodate and β-elimination only, without CuSO₄ (see Data Set S2 in the supplemental material), to measure the fraction of the tRNAs protected by either a single amino acid or a peptide. The aminoacyl tRNA fraction for each species is calculated by subtracting the protected fraction with CuSO₄ treatment from the total protected fraction. Dashed lines connect protected tRNA fractions in uninduced (no Pth knockdown) with those in induced (Pth knockdown) samples to show the overall decrease in usable charged tRNA pools. Values from three biological replicates are shown. Serine and threonine tRNAs are typically acylated at lower levels than other tRNAs in bacteria (30).

FIG 5

Pth depletion increases Mtb sensitivity to a candidate lysine tRNA synthetase inhibitor. Mtb growth in the presence of a candidate lysine tRNA synthetase inhibitor (compound 7 [31]) is plotted across drug concentrations as determined by fluorescence in an alamarBlue assay. Briefly, antibiotic-containing plates were incubated with Mtb cells for 6 days, at which point resazurin was added, followed by 48 h of additional agitation at 37°C. Fluorescence was normalized to the OD₆₀₀ and to the positive control for each strain (no antibiotic). The fractions of bacteria surviving relative to the no-drug control (“Normalized bacterial fraction surviving”) are plotted against the drug concentrations, along with a least-squares fit of the dose response. Pth CRISPRi strains were grown to mid-log phase, diluted to an OD₆₀₀ of 0.001, and plated with serial dilutions of antibiotics in 96-well plates. The fraction of Mtb cells surviving is plotted by normalizing the fluorescence values to the values for the control wells with no drug, and a least-squares fit of the dose-response data is plotted. Means with standard errors from two biological replicates are shown.

FIG 6

Pth depletion sensitizes Mtb to macrolides. (A) Mtb growth in the presence of two macrolides, erythromycin and clarithromycin, is plotted across drug concentrations as determined by fluorescence in an alamarBlue assay. Briefly, antibiotic-containing plates were incubated with Mtb cells for 6 days, at which point resazurin was added, followed by 48 h of additional agitation at 37°C. Fluorescence was normalized to the OD₆₀₀ and to the positive control for each strain (no antibiotic). The fractions of bacteria surviving relative to the no-drug control (“Normalized fraction surviving”) are plotted against drug concentrations, along with a least-squares fit of the dose response (39). The Pth CRISPRi strain (22) was grown to mid-log phase, diluted to an OD₆₀₀ of 0.001, and plated with serial dilutions of antibiotics in 96-well plates. The fraction of Mtb cells surviving is plotted by normalizing the fluorescence values to the values for the control wells with no drug. Means with standard errors from three biological replicates are shown. (B) Cu-tRNAseq was performed on Pth CRISPRi strains in the presence and absence of erythromycin. Strains were grown in the presence or absence of the inducer as described in the legend of Fig. 3. Uninduced strains were then treated with 25 μg/mL of erythromycin overnight, and Pth knockdown-induced strains were treated with 2 μg/mL for the same duration. Means with standard errors from three biological replicates are shown.