Toxic small alarmone synthetase FaRel2 inhibits translation by pyrophosphorylating tRNAGly and tRNAThr

Abstract

Translation-targeting toxic small alarmone synthetases (toxSAS) are effectors of bacterial toxin-antitoxin systems that pyrophosphorylate the 3'-CCA end of transfer RNA (tRNA) to prevent aminoacylation. toxSAS are implicated in antiphage immunity: Phage detection triggers the toxSAS activity to shut down viral production. We show that the toxSAS FaRel2 inspects the tRNA acceptor stem to specifically select tRNA^Gly and tRNA^Thr. The first, second, fourth, and fifth base pairs of the stem act as the specificity determinants. We show that the toxSASs PhRel2 and CapRel^SJ46 differ in tRNA specificity from FaRel2 and rationalize this through structural modeling: While the universal 3'-CCA end slots into a highly conserved CCA recognition groove, the acceptor stem recognition region is variable across toxSAS diversity. As phages use tRNA isoacceptors to overcome tRNA-targeting defenses, we hypothesize that highly evolvable modular tRNA recognition allows for the escape of viral countermeasures through tRNA substrate specificity switching.

Fig. 1.. Immunoprecipitated FaRel2 associates with type I tRNAs.

(A) Ten-fold dilutions of overnight cultures of E. coli strains transformed with pBAD33 vector or pBAD33 derivatives expressing either wild-type (WT) Сoprobacillus sp. D7 FaRel2 or alanine-substituted FaRel2 variants: K28A, R29A and the non-toxic double-substituted K28A R29A. (B) Purification of C-terminally FLAG₃-tagged FaRel2 using anti-FLAG–conjugated beads. Samples were separated on SDS-PAGE and visualized by blue-silver staining and by Western blotting with anti-FLAG antibodies. To counter the toxicity of wild-type FaRel2, the toxSAS was coexpressed with PaSpo^SSU5 SAH from Salmonella phage SSU5. (C) RNA coeluted with FaRel2-FLAG₃ resolved on urea-PAGE and visualized by SYBR Gold staining. (D) Immunoprecipitation of FaRel2-FLAG₃:ATfaRel2 for purification of co-IPed tRNA. (E) Comparison of the RNA samples coeluted with either FaRel2-FLAG₃ or FaRel2-FLAG₃:ATfaRel2 with a commercial preparation of E. coli small RNA fraction, tBulk_Roche. Additional replicates are shown on fig. S2. All experiments were performed at least two times; representative images are shown.

Fig. 2.. tRNA^Gly and tRNA^Thr are specifically bound by FaRel2 and FaRel2:ATfaRel2.

(A and B) tRNA species abundance in (A) FaRel2-FLAG₃- or (B) FaRel2-FLAG₃:ATfaRel2–co-IPed RNA fraction and E. coli tBulk was quantified by mim-tRNAseq (31). Fold change of the relative tRNA abundance (fraction in co-IPed pool versus tBulk) is shown as gray numbers. (C) Heatmap representation for log₂ fold change in abundance of tRNA species in co-IPed tRNA samples relative to their abundances in tBulk. The experiments were performed at least three times; the data are shown as average ± SD.

Fig. 3.. FaRel2 inhibits translation by specifically pyrophosphorylating tRNA^Gly and tRNA^Thr.

(A) Total RNA prepared from cells either expressing FaRel2 or transformed with an empty vector plasmid was resolved on acidic urea-PAGE and probed against tRNA^Thr1 or tRNA_i^fMet. (B and C) tRNA pyrophosphorylation by FaRel2 or B. subtilis la1a PhRel2 assayed using ³²P-labeled ATP. (B) FaRel2 specifically labels the type I tRNAs from E. coli total tRNA, tBulk_Roche. A total of 50 nM toxSAS-FLAG₃ was reacted with 5 μM tBulk_Roche at 37°C for increasing periods of time. tBulk_Roche stands for commercial preparation of E. coli small RNA fraction. (C) RNA^Gly and FaRel2-FLAG₃–co-IPed tRNA fractions are more efficiently modified by FaRel2 as compared to tBulk_Roche and individual E. coli tRNAs tRNA_i^fMet, tRNA^Phe, and tRNA^Val. A total of 50 nM toxSAS-FLAG₃ was reacted with 0.4 μM of tRNA preparations at 37°C for 10 min. (D) Reporter expression assays in cell-free protein synthesis system. Addition of FaRel2 abrogates production of Strep-tagged DHFR reporter proteins, which harbors full codon set or in which all Gly or Thr codons were substituted to Ala (Gly-to-Ala or Thr-to-Ala), but not the mutant variant in which all Gly and Thr codons were converted to Ala (Gly/Thr-to-Ala). Addition of B. subtilis la1a PhRel2 abrogates the expression of all reporters equally. All experiments were performed at least two times; representative gels are shown.

Fig. 4.. Conservation and diversification of tRNA recognition by toxSAS.

(A to C) AF3-generated models of tRNA^Gly1-bound Сoprobacillus sp. D7 FaRel2, B. subtilis la1a PhRel2, and CapRel^SJ46 colored by the surface charge. (D to F) Same toxSAS:tRNA^Gly1 complexes as on (A) to (C), but colored by amino acid conservation as computed using ConSurf (33).

Fig. 5.. Molecular determinants defining tRNA selection by FaRel2.

(A) Acceptor stem sequences for E. coli tRNA^Gly and tRNA^Thr isoacceptors. (B) E. coli tRNA^Gly1 secondary structure and the tested base pair swapping mutations. The minihelix part is outlined with an orange line. Base pair swapping mutations are indicated by arrowheads. Pink background shows the base pairs where swapping mutation abrogates pyrophosphorylation of mini helix by FaRel2, and green background indicates the base pair where the swapping mutation did not decrease the pyrophosphorylation. (C) Pyrophosphorylation of tRNA^Gly1-mimicking RNA minihelix by FaRel2 assayed with ³²P-labeled ATP. Base pairs at positions 1-72, 2-71, 4-69, and 5-68 but that at 3-70 are crucial for substrate recognition by FaRel2. A total of 5 nM FaRel2-FLAG₃ was reacted with 5 μM tRNA minihelix at 37°C for 5 min. The experiments were performed at least three times, and representative gels are shown. (D) Kinetic analysis of FaRel2-medited modification of wild type and A₅:U₆₈ and A₅:U₆₈ C₃:G₇₀ variants of tRNA^Gly1-mimicking RNA minihelix. The experiments were performed analogously to (C). (E) Acceptor stem sequences for E. coli tRNA_i^fMet, tRNA^Phe, tRNA^Val1, and tRNA^Val2.