Protein quality control at the interface of endoplasmic reticulum and mitochondria by Lon protease

Abbreviations

ER

endoplasmic reticulum

LonP1

Lon peptidase 1, mitochondrial

EFV

efavirenz

MAM

mitochondria‐associated membranes

UPR^ER

unfolded protein response in ER

UPR^MT

unfolded protein response in mitochondria

HSP

heat shock protein

We have read with great interest the recent article ‘Lon protease: a novel mitochondrial matrix protein in the interconnection between drug‐induced mitochondrial dysfunction and ER stress’ published in the British Journal of Pharmacology by Polo et al. (2017). This study adds further evidence that LonP1 is a stress protein. Previous studies have shown that LonP1 is up‐regulated under various cellular stress conditions such as hypoxic, oxidative and endoplasmic reticulum (ER) stress to adapt mitochondria to the changes caused by such stresses (Hori et al., 2002; Ngo et al., 2013). LonP1 functions by degrading oxidized and aggregated proteins in the mitochondria, thereby optimizing bio‐energetics and preserving the cell viability during various stress conditions (Bota et al., 2005; Fukuda et al., 2007; Bota and Davies, 2016). The current study by Polo et al. (2017) opened up a new area where LonP1 may also function at the ER‐mitochondria interface to preserve the mitochondrial homeostasis from being disrupted by both ER and mitochondrial stress pathways. Moreover, any effect that inhibits protein synthesis or protein folding in the ER may also affect mitochondrial proteostasis, as the majority of the mitochondrial proteins are synthesized and folded in the ER and imported into mitochondria. Previously, Hori et al. (2002) showed that HeLa cells treated with various ER stressors such as tunicamycin, brefeldin A or thapsigargin cause an accumulation of unfolded proteins in ER and an up‐regulation of LonP1 (Hori et al., 2002). In addition, hypoxia also induces ER stress and up‐regulates LonP1 (Hori et al., 2002). However, previous studies failed to look at the localization of LonP1 outside mitochondria, as Polo et al. (2017) pointed out in their discussion; hence, it is time to revisit LonP1's extramitochondrial localization, specifically at mitochondria‐associated membranes (MAMs), which is an ER‐mitochondrial interface through which the majority of cellular processes are believed to be regulated (Rowland and Voeltz, 2012; van Vliet et al., 2014).

It is interesting that the antiretroviral non‐nucleoside inhibitor efavirenz (EFV), which is used to treat HIV, induces stress in both mitochondria and ER, causing the depletion of LonP1 in mitochondria and the accumulation of LonP1 outside at MAMs (Polo et al., 2017). In this study, Polo et al. (2017) showed that EFV induces both ER stress and mitochondrial dysfunction, but the specific targets of EFV remain unknown (Apostolova et al., 2013; Tan et al., 2017). A similar class of non‐nucleoside inhibitor rilpivirine does not cause an identical effect suggesting that these drugs may have different molecular targets in cells due to their dissimilar structures. But it is also unclear whether EFV directly causes stress in both mitochondria and ER or whether stress in one organelle induces a stress response in the other. However, it is surprising to see that even in untreated cells, the LonP1 signal is observed in MAMs, cytosol and ER (Figure 8B), which is in contrast to the Western blots depicted in Figure 8A, which is also an untreated group that shows virtually all LonP1 in the mitochondrial fraction (Figure 8A; Polo et al. 2017). This makes it unclear whether LonP1 is present in the MAM at basal conditions, which is unlikely as LonP1 is specifically targeted to the mitochondrial matrix by its mitochondrial targeting sequence (Wang et al., 1993). Figure 8A,B differ in their methodology, and whether DMSO or methanol, which were used as a vehicle control, affect the LonP1 localization is not clear. However, if LonP1 is present at a basal level in MAMs, then this suggests that LonP1 has a basal functional role in MAMs as suggested by Polo et al. (2017). Further experiments are warranted to confirm and conclude that LonP1 is localized at MAMs under basal conditions. Despite this difference observed between two similar experiments (Figure 8A,B), the presence of LonP1 at MAMs after EFV treatment is intriguing. This raises many questions: is the LonP1 observed at the MAM a full‐length precursor or the matrix soluble matured protein that is translocated onto MAM upon EFV‐induced stress? Moreover, the localization of LonP1 at the MAM was observed only at a higher concentration of EFV, 50 μM (Figure 4A). At this concentration, EFV may induce the opening of the permeability transition pore with loss of mitochondrial membrane potential, leading to the release of certain soluble proteins (Pilon et al., 2002), and LonP1 may be one of them; Polo et al. (2017) failed to test for more soluble proteins.

As regards the mechanism of EFV‐induced LonP1 up‐regulation: Polo et al. (2017) observed only a partial effect of PERK (protein kinase RNA‐like ERK), but mainly, the up‐regulation was shown to be mediated through NF‐κB activation and not through CHOP (CCAAT‐enhancer‐binding protein homologous protein). Previously, it has been shown that the unfolded protein response specifically at mitochondria (UPR^MT) is initiated by ClpP, leading to the up‐regulation of certain mitochondrial chaperones, such as heat shock protein (HSP)60, HSP70, HSP10, mt. DNAJ and ClpP itself but not LonP1 suggesting that there is an alternative stress protein response pathway, which is specific for LonP1, evoked either through oxidative stress (Ngo and Davies, 2009; Pinti et al., 2011) or through the UPR^ER (Hori et al., 2002). Therefore, it may be possible that EFV causes mitochondrial stress mainly through inducing ER and oxidative stress, as observed in this study. However, Polo et al. (2017) failed to investigate ClpP, which is the protease domain of ClpXP, another matrix soluble hetero‐oligomeric protease and mitochondrial marker for UPR^MT. But instead, Polo et al. (2017) analysed the level of ClpX, which is an ATPase domain of ClpXP and misstated ClpX as a protease. We also want to point out another possible misstatement, ‘Drp1 is recruited to mitochondria in order to promote fusion’, but Drp1 is mainly recruited to promote mitochondrial fission (Imoto et al., 1998).

LonP1 protease is highly conserved from bacteria to mammals and one of the main mitochondrial protein quality control proteases in the mitochondria (Venkatesh et al., 2012). The importance of LonP1 in mammals is starting to emerge; in the recent study by Quiros et al. (2014) LonP1 was shown to be important for embryonic viability (Quiros et al., 2014). In humans, pathological mutations in LonP1 are associated with CODAS (cerebral, ocular, dental, auricular and skeletal anomalies), a severe developmental disorder, the first‐ever human disease to be linked to LonP1 (Dikoglu et al., 2015; Strauss et al., 2015; Inui et al., 2017). We believe that LonP1 is crucial for maintaining mitochondrial function not only during developmental stress but also during various pathological stresses that affect mitochondrial proteostasis. One of these mechanisms could also be by regulating the communication between the ER and mitochondria, which is largely unexplored, and we hope this article provides a starting point for exploring a new opening for LonP1.

Conflict of interest

The authors declare no conflicts of interest.

Pandey, A. K. , and Venkatesh, S. (2019) Protein quality control at the interface of endoplasmic reticulum and mitochondria by Lon protease. British Journal of Pharmacology, 176: 505–507. 10.1111/bph.14543.