Reevaluating the role of human mitochondrial uL18m in the cytosolic stress response

A variety of stresses, including heat shock, lead to changes in gene expression patterns in mammalian cells caused by a general shutdown and reprogramming of cytosolic translation. This translational reprograming allows stress-resistant synthesis of heat-shock and other cytoprotective proteins. In their article, Zhang et al. describe a moonlighting role of MRPL18 (uL18m hereafter), a canonical protein of the mitochondrial ribosome, in cytosolic translation reprogramming following heat-shock¹. Specifically, the authors propose that heat-shock induces de novo translation of an uL18m isoform from a downstream open reading frame (dORF) that lacks the mitochondrial targeting sequence. This isoform associates directly with the cytosolic ribosome and causes it to preferentially translate stress-responsive mRNAs, such as Hsp70. Several years after the original publication, these findings still draw attention because of their implications for basic mitochondrial processes and translational control mechanisms, but confirmatory observations have yet to be published. Despite our genuine effort to replicate and build upon one of the key findings of Zhang et al., our data indicate that uL18m is not necessary for the translational upregulation of HSP70 following heat-shock in human cells.

Central to the model of Zhang et al. is the observation that silencing of uL18m attenuates the induction of HSP70 upon heat shock. From these data, Zhang et al. conclude that uL18m is necessary for a full-strength heat shock response (HSR). We attempted to replicate these findings in HeLa cells, a cell line also used by Zhang et al., and U87 cells, a human glioblastoma cell line commonly used in heat-shock studies. U87 cells have lower baseline levels of HSP70 protein and a more robust HSP70 induction in response to heat-shock compared to HeLa cells (Supplementary Fig 1A & B). siRNA-mediated silencing of uL18m using two distinct RNAi constructs (an siRNA pool from Dharmacon used by Zhang et al. or a single siRNA from Invitrogen targeting exon 2 of uL18m) was highly efficient (Supplementary Fig 1C & D). Validation of our anti-uL18m antibody alongside that used by Zhang et al. is provided in Supplementary Fig 1D. When we combined uL18m silencing for 72 hours using the Dharmacon siRNA pool with subsequent heat-shock, we found that depletion of uL18m protein did not impede HSP70 protein induction in either HeLa or U87 cell lines (Fig 1A & B).

Figure 1. Depletion of uL18m does not affect HSP70 induction upon heat shock.

HeLa cells (a, c) or U87 (b, d) were treated with uL18m targeting siRNAs for 72 hours (a, b) or 14 hours (c, d), followed by a one-hour heat-shock and up to 8 hours of recovery as indicated. C indicates control cells that did not receive heat shock. (Left) Immunoblot of whole-cell lysates probed with the indicated antibodies. β-ACTIN is used as loading control. (Right) Quantification of relative HSP70 abundance is presented as interleaved scatter plot. In all graphs, data are presented as mean values +/− S.D., n=3 independent experiments. Statistical analyses were performed using the two-sided Student’s t-test in the Prism-6 software. No significant difference (p<0.05) in HSP70 abundance between si-uL18m and control conditions was detected for any time point. Additional replicates and data behind the graphs are available as source data online. Uncropped images of the presented blots are available at https://doi.org/10.6084/m9.figshare.14378510.v1.

We repeated the experiment after silencing uL18m for only 14 hours. At this time point uL18m mRNA is reduced by more than 90% in both cell lines (Supplementary Fig 1C) while the steady-state levels of uL18m protein are not yet affected (Fig 1C & D). Thus, de novo synthesis of uL18m protein, including the alternative dORF isoform, is inhibited, while uL18m protein generated prior to silencing is still present and can carry out its canonical function in the mitochondrial ribosome^2,3. This experimental set up allows us to test the assertion by Zhang et al. that the de novo synthesized uL18m protein isoform is necessary for induction of the HSR without confounding effects due to loss of mitochondrial uL18m function. As at the 72h time point, we see no decrease in HSP70 induction upon heat shock in either HeLa or U87 (FIG 1C & D).

In addition, we tested an alternative siRNA targeting uL18m and in this experimental set-up, we compared cells transfected with the siRNA to untreated cells as a control. As before, we did not observe a change in the HSP70 induction upon uL18m silencing in HeLa or U87 cells (Extended Data Figure 1). Taken together, our data suggest that uL18m is not necessary for the induction of HSP70 in human cells.

Methods

A detailed description of the methods is available in Supplementary Note 1.

Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Extended Data

Extended Data Fig. 1 |. Depletion of uL18m using an alternative siRNA construct does not affect HSP70 induction upon heat shock.

a, HeLa cells were treated for 72-hours with siRNA targeting uL18m (Invitrogen), or left untreated, followed by heat-shock and recovery for 0–6 hours as indicated. ‘C’ indicates control cells that did not receive heat shock. (Left) Immunoblot of whole-cell lysates with the indicated antibodies. β-ACTIN is used as loading control. (Right) Quantification of relative HSP70 abundance presented as interleaved scatter plot. n = 5 independent experiments. b, As a), but in U87 cells, with recovery after heat shock for 0–8 hours as indicated. n = 3 independent experiments. c, As a), after 14 hours of treatment with the siRNA. n = 3 independent experiments. d, As b) after 14 hours of treatment with the siRNA. n = 4 independent experiments. In all the graphs in this figure, data are presented as mean values ± S.D. Statistical analyses were performed using the two-sided Student’s t-test in the Prism-6 software. No significant difference in HSP70 abundance between si-uL18m and control conditions was detected for any time point set; p<0.05. White space between the western blots indicates cropping of the blots to remove an irrelevant sample, but all samples were run on same gel. Additional replicates and data behind the graphs are available as source data online. Uncropped images of the presented blots are available at https://doi.org/10.6084/m9.figshare.14377337.v1.

Data availability

Source Data are available with this paper online.