Description

The mitochondrial unfolded protein response (mitoUPR) is a stress response pathway that promotes cell survival and restores mitochondrial function when mitochondrial health is compromised (Haynes et al. 2013; Jovaisaite et al. 2014; Shpilka and Haynes 2018). While a mitoUPR was first reported in mammalian cells (Zhao et al. 2002), the initial work on the mitoUPR in C. elegans was performed by Yoneda et al. who found that treatment with ethidium bromide, which affects the replication and expression of mitochondrial DNA, increased the expression of the mitochondrial chaperone gene hsp-6 (Yoneda et al. 2004). Based on this observation, they generated hsp-6p::gfp and hsp-60p::gfp reporter strains to further study the mitoUPR. They found that either RNA interference (RNAi) targeting spg-7, the worm homolog of paraplegin, or RNAi targeting other genes encoding mitochondrial proteins resulted in activation of the hsp-6p::gfp and hsp-60p::gfp reporter strains (Yoneda et al. 2004).

The hsp-6p::gfp and hsp-60p::gfp reporter strains were used to identify other components of the mitoUPR including the transcriptional regulator UBL-5, the transcription factor DVE-1, the protease ClpP, the protein import channel HAF-1, and the transcription factor ATFS-1 (Benedetti et al. 2006; Haynes et al. 2007; Haynes et al. 2010; Nargund et al. 2012). Under normal conditions, the mitoUPR transcription factor, ATFS-1, is imported into the mitochondria through the HAF-1 import channel and degraded by the protease ClpP. However, when mitochondria are impaired, ATFS-1 import into the mitochondria is blocked. Instead, the nuclear localization signal (NLS) of ATFS-1 drives ATFS-1 into the nucleus. In the nucleus, ATFS-1 acts with DVE-1 and UBL-5 to upregulate genes that restore mitochondrial homeostasis, including genes involved in mitochondrial protein folding and metabolism. In addition to responding to disruptions in mitochondrial function and integrity, the mitoUPR also plays important roles in resistance to stress (Pellegrino et al. 2014; Campos et al. 2021; Soo et al. 2021) and longevity (Durieux et al. 2011; Houtkooper et al. 2013; Bennett et al. 2014; Wu et al. 2018).

Previous studies from our laboratory and others have used RNA sequencing (RNA-seq) or microarrays to examine gene expression under conditions that induce ATFS-1 activation. Nargund et al. identified 366 genes that are upregulated by spg-7 RNAi in an ATFS-1-dependent manner (Nargund et al. 2012). We previously identified 1704 genes that are upregulated in nuo-6 mutants in an ATFS-1-dependent manner (Senchuk et al. 2018; Wu et al. 2018) and 529 genes that are upregulated in both atfs-1(et15) and atfs-1(et17) constitutively active atfs-1 mutants (Rauthan et al. 2013; Wu et al. 2018).

In order to establish a list of mitoUPR target genes that are consistently upregulated across all three ATFS-1-activating conditions, we examined the overlap between these datasets. We identified a total of 61 genes that are upregulated in nuo-6 worms in an ATFS-1-dependent manner, upregulated by spg-7 RNAi in an ATFS-1-dependent manner, and upregulated in both atfs-1(et15) and atfs-1(et17) constitutively active mutants (Extended data, Column A,B). Surprisingly, neither hsp-6 nor hsp-60 were among the 61 genes on this list. In fact, hsp-60 was not identified in any of the three datasets. It is uncertain why spg-7 RNAi increases fluorescence in the hsp-60p::GFP reporter strain (Yoneda et al. 2004) but hsp-60 was not identified as one of the genes upregulated by spg-7 RNAi (Nargund et al. 2012). One possibility may be that hsp-60 expression is primarily upregulated during development in response to spg-7 RNAi and other mitochondrial insults. Then the half-life of the HSP-60::GFP protein might allow the increase in HSP-60::GFP to still be observed at adulthood even after the hsp-60 mRNA has returned to baseline.

In order to rank the different mitoUPR target genes, each gene was scored by adding together their expression levels in nuo-6, atfs-1(et15) and atfs-1(et17) mutants and subtracting their expression level in nuo-6;atfs-1 mutants. By this metric, 42 of the 61 mitoUPR target genes exhibited a higher score than hsp-6 (Extended data, Column J). The top genes of C07G1.7, F22B3.7, cyp-14A4, and clec-17 had scores ranging from 17084-21882, compared to a score of 350 for hsp-6. The expression of these genes compared to hsp-6 and hsp-60 is shown in Fig 1A-F.

In order to control for variability, we divided this score by the average standard deviation from WT, atfs-1(gk3094), nuo-6, nuo-6;atfs-1(gk3094), afts-1(et15) and atfs-1(et17). By this metric, 27 of the 61 mitoUPR target genes showed a higher score than hsp-6 (Extended data, Column K). The top genes of C07G1.7, hrg-9, K09E9.1, cdr-2 and F15B9.10 had variability-corrected scores of 26-38 compared to 15 for hsp-6. The expression of these genes is shown in Fig 1G-J.

To determine if these mitoUPR target genes are being activated directly by binding of ATFS-1, we examined the data from a previous chromatin immunoprecipitation sequencing (ChIP-seq) experiment that identified genes bound by ATFS-1 after treatment with spg-7 RNAi (Nargund et al. 2015). We found that 22 of the 61 mitoUPR target genes exhibited binding of ATFS-1 after spg-7 RNAi (Extended data, Column L). This suggests that the expression of these genes is directly modulated by ATFS-1, while the expression of the other mitoUPR target genes may be regulated indirectly, or perhaps bound by ATFS-1 under different conditions than were utilized in the ChIP-seq study.

Our results suggest that the genes identified here may be better for monitoring the activation of the mitoUPR by quantitative RT-PCR or RNA-seq than hsp-6 and hsp-60. To that end, we have designed and validated primers for quantitative RT-PCR to measure the levels of two of the highest-ranked direct targets, cdr-2 and F15B9.10, and two of the highest ranked indirect targets, C07G1.7 and hrg-9. We confirmed that all four sets of primers can efficiently measure activation of the mitoUPR in nuo-6 worms compared to wild-type worms (Fig 1K-N).

Finally, to validate this list of 61 genes for monitoring the activation of the mitoUPR, we examined the long-lived mitochondrial mutant isp-1, which we have previously shown to have increased activation of a hsp-6p:::gfp reporter strain and significantly overlapping gene expression changes with constitutively active atfs-1 mutants (Wu et al. 2018). 57 of the 61 genes were found to be significantly upregulated in isp-1 mutants (Extended data, Fig 1O).

Overall, this work has identified a gene expression signature that can be used to monitor the activation of the mitoUPR in RNA-seq, microarray or qPCR studies. These genes can be used to complement the use of hsp-6p::gfp and hsp-60p::gfp reporter strains, and may provide a more robust measure of mitoUPR activation in studies measuring mRNA.