Regulation of multiple paralogs of a small subunit ribosomal protein in Francisella tularensis

Abstract

Francisella tularensis is a highly infectious human pathogen that must replicate inside macrophage to cause disease. The ribosomes of F. tularensis can incorporate one of three different paralogs for the small ribosomal subunit protein bS21. One of these paralogs positively impacts translation of key virulence genes and promotes intramacrophage replication. Although ribosomal bS21 content influences F. tularensis virulence, the factors that control bS21 paralog production are not well understood. Here, we reveal that all three bS21 proteins influence the transcript abundance of the paralog important for virulence, bS21-2. In contrast, the other bS21 paralogs (bS21-1 and bS21-3) do not affect their own production. We further determined that the leader sequence of the bS21-2 mRNA is sufficient for bS21-mediated repression of mRNA abundance, suggesting that bS21-2 is autogenously regulated. Yet we determined that the increase in bS21-2-encoding mRNA is not reflected by increased protein production, suggesting that translation of this transcript is controlled by other factors. Finally, we found that bS21-2 exerts at least some of its effects on the bS21-2 transcript by decreasing its stability. Together, our findings suggest that F. tularensis integrates multiple signals into a regulatory network to control the appropriate production of each bS21 paralog, and particularly the paralog important for virulence, bS21-2. This regulatory network in turn may control ribosomal heterogeneity and virulence gene expression.

Figure 1.. All three F. tularensis bS21 paralogs negatively regulate the rpsU2 operon.

Normalized transcript abundance reads from RNA-Seq experiments in the area surrounding the rpsU2 operon. Data from cells with bS21–2 (LVS pF), cells lacking bS21–2 (LVS ΔrpsU2 pF), and cells lacking bS21–2 but ectopically expressing either bS21–2, bS21–1, or bS21–3 (LVS ΔrpsU2 pF-bS21–2-V, pF-bS21–1-V, pF-bS21–3-V, respectively). Y-axis is truncated at 1x10⁴ for clarity. Grey rectangles represent genes and those above the line indicate genes encoded on the positive strand; those below the black line represent those encoded on the negative strand. The purple box indicates an experimentally-determined promoter region (Ramsey et al., 2015).

Figure 2.. bS21–1 and bS21–3 do not regulate their own transcript abundance.

(A) Transcript abundance of the rpsU1 5´ UTR in either wild-type cells or cells lacking rpsU1, relative to tul4 (a control gene). (B) Transcript abundance of the rpsU3 5´ UTR in either wild-type cells or cells lacking rpsU3, relative to rpoA (a control gene with lower transcript abundance than tul4). (A and B) Differences did not reach statistical significance by t test. Experiments were repeated at least twice in biological triplicate and data from a representative experiment are shown.

Figure 3.. The rpsU2 5´ UTR mediates transcript abundance yet transcript and protein abundance changes are not correlated.

(A) Diagram of translational reporters. (B) Relative mRNA and protein abundance for indicated products from translational fusions in cells with (+; wild-type) or without (−; ΔrpsU2) bS21–2. Quantitative RT-PCR was used to determine the relative lacZ transcript normalized to the tul4 gene. β-galactosidase activity was used to determine the relative protein abundance. Error bars represent 1 SD. Lines above bars indicate comparisons between wild-type and ΔrpsU2 of the same translational reporter, *p < 0.05. Experiments were repeated at least twice in biological triplicate and data from a representative experiment are shown.

Figure 4.. Predicted stem-loops in the rpsU2 mRNA are dispensable for control by bS21–2.

(A) Secondary structure predictions of wild-type and modified rpsU2 5 UTRs, generated by MXfold2. (B) Relative mRNA and protein abundance for indicated products from translational fusions in cells with (+; wild-type) or without (−; ΔrpsU2) bS21–2. Quantitative RT-PCR was used to determine the relative gfp transcript normalized to the tul4 gene. Fluorescence was used to determine the relative protein abundance. Error bars represent 1 SD. Lines above bars indicate comparisons between wild-type and ΔrpsU2 of the same translational reporter, *p < 0.05. Experiments were repeated at least twice in biological triplicate and data from a representative experiment are shown.

Figure 5.. The presence of bS21–2 destabilizes transcripts with the rpsU2 5´ UTR.

Half-lives of (A) rpsU operon mRNA as measured by yqeY transcript or (B) lacZ mRNA in cells containing P_rpsU2-5´ UTR_rpsU2-lacZ translational fusions with (+; wild-type) or without (−; ΔrpsU2) bS21–2. Linear regression analysis was used to calculate and compare half-life values. Error bars indicate standard error. Experiments were repeated at least twice in biological triplicate, and data from a representative experiment are shown. ***p<0.0001