Loss of chloroplast ClpP elicits an autophagy-like response in Chlamydomonas

Abstract

Chloroplast genomes contain a single ClpP1 gene encoding one of the catalytic subunits of the evolutionarily conserved ATP-dependent Clp protease. Efforts to inactivate this protease in the chloroplast through targeted disruption of the clpP1 gene have failed, suggesting that it is essential for cell survival in plants. To circumvent this problem, a repressible chloroplast gene expression system was developed in the green unicellular alga Chlamydomonas reinhardtii. This system takes advantage of the nuclear Nac2 gene fused to the MetE promoter and Thi4 riboswitch, which can be repressed by adding vitamin B12 and thiamine to the growth medium. Nac2 encodes a chloroplast protein that interacts specifically with the 5′UTR of the psbD mRNA and is involved in processing/translation of this transcript. Loss of Nac2 leads to the specific degradation of psbD mRNA. Because the psbD 5′UTR is necessary and sufficient for the Nac2-dependent stability of psbD mRNA, this dependence can be transferred to any chloroplast gene by linking its coding sequence to the psbD 5 ‘UTR. In this way it was possible to repress the clpP1 gene in a reversible way with vitamins.

Using this genetic system, we have followed the time course of ClpP1 depletion through biochemical, transcriptomic, and proteomic approaches as well as light and electron microscopy. These studies showed that Atg8 and Atg3, 2 key proteins involved in autophagosome biogenesis, are highly upregulated upon ClpP1 depletion. In particular, there is a striking inverse correlation between the decrease of ClpP1 and the increase of Atg8. Furthermore, similar morphological changes such as cellular swelling and extensive cytosolic vacuolization are observed in Chlamydomonas cells upon ClpP1 depletion or treatment with rapamycin, a chemical inducer of autophagy.

A quantitative comparison of changes in gene expression upon ClpP1 depletion and rapamycin treatment allowed us to identify a group of mRNAs that respond in a similar manner to these 2 stress conditions. Among the most upregulated transcripts, we found genes involved in autophagy (Atg3, Atg6, Atg7, Atg8, and Atg12). In contrast, downregulated transcripts include genes implicated in cell cycle (e.g., CycB1, Cdc45, and Mat3/Rb-like) and DNA replication (e.g., genes encoding subunits of the Mcm, Smc, and Orc complexes). These findings are of particular interest because they highlight the possibility of uncovering novel autophagy-related proteins in the list of coregulated and unknown genes and to eventually explore their degree of conservation in other organisms. As an interesting example, while our study was in progress, SNARE proteins were biochemically demonstrated to play a role in membrane fusion during autophagy in mammalian cells. In agreement with these findings, we discovered that ClpP1 depletion and rapamycin treatment both cause a significant increase of 2 transcripts encoding SNARE proteins (Vmpl2 and Vamp74) in Chlamydomonas. Likewise, the increased abundance of a transcript encoding a putative mechanosensitive channel conserved in almost every organism deserves to be investigated further since very little is known about how cells sense the mechanical tension in the lipid bilayer associated with cellular swelling observed during autophagy. In general, our data confirm the evolutionary role of autophagy as a mechanism that eukaryotic cells employ to devote their resources to maintenance and survival at the expense of growth and division.

Importantly, given the different stress inputs used in our experiments, we have also identified a group of genes that are selectively responsive to ClpP1 depletion but not to rapamycin treatment under continuous growth in dark or light conditions. Because some of these genes encode chloroplast chaperones, proteases, and proteins involved in chloroplast membrane biogenesis and/or repair, they are indicative of an unfolded protein response. Among these proteins several have isoforms in different cellular compartments. This is the case for the Hsp22 and Hsp70 heat shock proteins, the Deg proteases, the ClpB diaggregases, and for the GroE type chaperonins. Intriguingly, the proteomic analysis revealed that mainly the chloroplast isoforms are upregulated during ClpP1 depletion, suggesting the existence of an organelle-specific stress response that we propose to name chloroplast unfolded protein response (cpUPR).

To validate our results, 2 of the early-responsive cpUPR genes, Vipp2, presumably involved in lipid trafficking and thylakoid membrane assembly, and Deg11, involved in proteolysis, were examined using reporters driven by their promoters and 5′UTRs. Interestingly, expression of the Vipp2-driven reporter was increased upon ClpP1 depletion or high light stress but not under many other stress conditions. Expression of the Deg11-driven reporter followed a similar pattern in these experiments indicating that induction of these 2 genes depends on a distinct pathway from the autophagy pathway. Deg11 could also be induced upon inhibition of glycosylation in the endoplasmic reticulum, however, thus raising questions about interactions between this compartment and the chloroplast.

Surprisingly, expression of several genes involved in RNA surveillance is also significantly affected upon ClpP1 downregulation. In particular, the abundance of the mRNAs of a putative homolog of the yeast 5′-3′ exoribonuclease Xrn1 and that of the 3′-5′ exoribonucleases Rnb1 and chloroplast Rnb2 strongly increase. Moreover, 3 of the most upregulated transcripts in this category are related to the bacterial Rsr and to the human Ro 60-kDa autoantigen. These proteins are not predicted to be targeted to the chloroplast and they can be weakly induced by rapamycin treatment, too. However they deserve attention because they act as cofactors for the association of nucleases with misfolded RNAs during stress. Therefore, the question arises whether an unfolded RNA response is integrated with an unfolded protein response under the particular stress conditions induced by ClpP depletion in Chlamydomonas.

Taken together, our results indicate that the ClpP protease plays an essential role in plastid protein homeostasis and that interference of its function through ClpP1 depletion triggers distinct responses. These include autophagy, a chloroplast protein unfolded response, and possibly an RNA unfolded response (Fig. 1). A challenge for the future will be to identify the molecular events underlying these signaling pathways between chloroplast and nucleus.

Figure 1. Scheme showing how ClpP depletion affects chloroplast protein homeostasis and triggers various cellular responses including autophagy (ATG), a chloroplast unfolded protein response (UPR), and possibly an unfolded RNA response (URR). Induction of autophagy may occur through modulation of TOR signaling upon ClpP depletion. These responses lead to a bulk degradation of macromolecules and to an increase of several chloroplast chaperones, proteases, and proteins involved in thylakoid membrane assembly/repair and proteins involved in RNA quality control.

Ramundo S, Casero D, Mühlhaus T, Hemme D, Sommer F, Crèvecoeur M, Rahire M, Schroda M, Rusch J, Goodenough U, et al. Conditional Depletion of the Chlamydomonas Chloroplast ClpP Protease Activates Nuclear Genes Involved in Autophagy and Plastid Protein Quality Control. Plant Cell. 2014;26:2201–22. doi: 10.1105/tpc.114.124842.