Control of mitophagy initiation and progression by the TBK1 adaptors NAP1 and SINTBAD

Mitophagy preserves overall mitochondrial fitness by selectively targeting damaged mitochondria for degradation. The regulatory mechanisms that prevent PINK1/Parkin-dependent mitophagy and other selective autophagy pathways from overreacting while ensuring swift progression once initiated are largely elusive. Here, we demonstrate how the TBK1 adaptors NAP1 and SINTBAD restrict the initiation of OPTN-driven mitophagy by competing with OPTN for TBK1. Conversely, they promote the progression of NDP52-driven mitophagy by recruiting TBK1 to NDP52 and stabilizing its interaction with FIP200. Notably, OPTN emerges as the primary recruiter of TBK1 during mitophagy initiation, which in return boosts NDP52-mediated mitophagy. Our results thus define NAP1 and SINTBAD as cargo receptor rheostats, elevating the threshold for mitophagy initiation by OPTN while promoting the progression of the pathway once set in motion by supporting NDP52. These findings shed light on the cellular strategy to prevent pathway hyperactivity while still ensuring efficient progression.


SUMMARY
Mitophagy preserves overall mitochondrial fitness by selectively targeting damaged mitochondria for degradation. The regulatory mechanisms that prevent PINK1/Parkindependent mitophagy and other selective autophagy pathways from overreacting while ensuring swift progression once initiated are largely elusive. Here, we demonstrate how the TBK1 adaptors NAP1 and SINTBAD restrict the initiation of OPTN-driven mitophagy by competing with OPTN for TBK1. Conversely, they promote the progression of NDP52-driven mitophagy by recruiting TBK1 to NDP52 and stabilizing its interaction with FIP200. Notably, OPTN emerges as the primary recruiter of TBK1 during mitophagy initiation, which in return boosts NDP52-mediated mitophagy. Our results thus define NAP1 and SINTBAD as cargo receptor rheostats, elevating the threshold for mitophagy initiation by OPTN while promoting the progression of the pathway once set in motion by supporting NDP52. These findings shed light on the cellular strategy to prevent pathway hyperactivity while still ensuring efficient progression.
We therefore investigated the roles of NAP1/SINTBAD in PINK1/Parkin-dependent mitophagy and discovered their overall inhibitory role in this pathway. While they support NDP52-mediated mitophagy, they negatively regulate TBK1 recruitment and activation by OPTN. This competition for TBK1 binding prevents OPTN from fulfilling one of its primary functions during mitophagy initiation. Our findings highlight a multilayer regulation of mitophagy initiation by NAP1/SINTBAD, acting as cargo receptor rheostats that increase the threshold for mitophagy initiation but promote the progression of the pathway once set in motion. As such, NAP1/SINTBAD provide insight into the cellular strategy that prevents selective autophagy pathways from overreacting while ensuring swift progression once initiated.

NAP1/SINTBAD are recruited and co-degraded during mitophagy
To understand whether the TBK1 adaptors NAP1/SINTBAD have a function in PINK1/Parkin mitophagy, we investigated if NAP1/SINTBAD are recruited to mitochondria during this process. To this end, we stably expressed HA-NAP1 or HA-SINTBAD in wild-type (WT) HeLa cells that also expressed YFP-Parkin and assessed their subcellular localization.
However, upon induction of mitophagy using a combination of Oligomycin A and Antimycin A1 (O/A), agents targeting the mitochondrial ATP synthase and complex III, respectively, both NAP1 and SINTBAD notably accumulated on depolarized mitochondria (Fig. 1A). We then performed co-staining with WIPI2, a marker for early cup-shaped membrane structures known as phagophores, precursors to autophagosomes. This demonstrated colocalization between NAP1/SINTBAD and WIPI2 (Fig. 1B), indicating that both NAP1 and SINTBAD were recruited to sites of autophagosome formation.
To test if NAP1/SINTBAD are degraded along with damaged mitochondria during mitophagy, we assessed the proteins levels of NAP1/SINTBAD. This revealed a decrease in NAP1/SINTBAD levels upon mitophagy induction, which was partially mitigated when lysosomal degradation was inhibited by Bafilomycin A1 (Fig. 1C). This indicates that NAP1/SINTBAD are not only recruited to sites of autophagosome formation, but that a portion of NAP1/SINTBAD also undergo autophagy-dependent degradation alongside damaged mitochondria, implying a potential role for them in the PINK1/Parkin mitophagy pathway.

NAP1/SINTBAD are mitophagy inhibitors
To explore the involvement of NAP1/SINTBAD in PINK1/Parkin mitophagy, we generated knockout HeLa cells for both factors and assessed mitophagy flux. Depletion of either NAP1 or SINTBAD alone did not impact the mitophagy rate in a statistically significant manner, as shown by the mitochondrial-targeted mKeima (mt-mKeima) assay ( Fig. 2A-B) [48]. Recognizing their structural similarities, which might facilitate compensation for each other, we also generated NAP1/SINTBAD double knockout (DKO) cells. To our surprise, we observed an enhancement in mitophagy flux in NAP1/SINTBAD DKO cells (Fig. 2C), contrasting their supporting role in NDP52-mediated xenophagy [47]. This finding was validated by assessing mitochondrial protein COXII levels via western blotting, confirming accelerated mitochondrial degradation in NAP1/SINTBAD DKO cells (Fig. 2D).
To explore whether NAP1/SINTBAD also regulate non-selective bulk autophagy, we evaluated p62 degradation in starved cells. Our findings indicated no discernible changes in p62 degradation in single or double knockout cell lines when compared to control wild-type cells (Fig. S1). Therefore, NAP1/SINTBAD are involved in the regulation of selective forms of autophagy, such as mitophagy, but not in non-selective bulk autophagy.

NAP1/SINTBAD support NDP52-mediated mitophagy by enabling TBK1 binding and stabilizing interactions with the autophagy machinery
To investigate the mechanisms underlying NAP1/SINTBAD's inhibition of mitophagy, we first focused on their functional interaction with NDP52, as they were previously implicated in an NDP52-dependent selective autophagy pathway, albeit in a stimulatory manner [47].
To explore their interplay with NDP52, we generated CRISPR/Cas9 double knockout clones for NAP1/SINTBAD in the pentaKO background, which lacks five key cargo receptors OPTN, NDP52, TAX1BP1, p62, and NBR1 [30]. This allowed us to reintroduce NDP52 into these cells and to assess NDP52-driven mitophagy rates in the presence or absence of NAP1/SINTBAD, eliminating the confounding effects from other cargo receptors, including OPTN. Surprisingly, contrary to our previous observations (Fig. 2), deleting NAP1/SINTBAD in these cells resulted in reduced mitophagy. This was evident from reduced degradation of the mitochondrial marker COXII (Fig. 3A), decreased mt-mKeima conversion (Fig. 3B), and impaired TBK1 activation (Fig. 3C). Moreover, the deletion of NAP1/SINTBAD may have weakened the NDP52-FIP200 interaction, as in vitro reconstitution of NDP52-mediated mitophagy initiation revealed that SINTBAD enhanced the NDP52-FIP200 interaction (Fig.   3D), underscoring an important role for NAP1/SINTBAD in this critical early step of mitophagy initiation.
Despite the significant contribution of NAP1/SINTBAD to these important first steps of NDP52-mediated mitophagy initiation, the overall reduction in mitophagy flux was relatively modest. However, considering that NDP52 can drive mitophagy through either ULK1/2 or TBK1 [40], we knocked out ULK1/2 in NAP1/SINTBAD DKO/pentaKO cells to elucidate the necessity of NAP1/SINTBAD when NDP52 engages in mitophagy solely through the TBK1 pathway. In the absence of ULK1/2, NAP1/SINTBAD emerged as essential factors for NDP52mediated mitophagy, evident from significantly reduced COXII turnover (Fig. 3E) and diminished WIPI2 and ATG13 recruitment upon O/A treatment (Fig. 3F-G). The latter to the same extent as when we inhibited TBK1 with the small molecule BX795.
In summary, these findings reveal important roles for NAP1/SINTBAD in supporting NDP52-mediated mitophagy through the recruitment of TBK1 and stabilization of the NDP52-FIP200 complex.

NAP1/SINTBAD are sufficient to induce mitophagy when recruited to mitochondria
From our experiments above (Fig. 3), it becomes evident that NAP1/SINTBAD exhibit traits of cargo receptors, including their ability to bind FIP200 and TBK1, albeit lacking the ubiquitin binding capabilities of cargo receptors. However, ubiquitin chains are critical in marking damaged organelles for autophagic degradation. With this in mind, we hypothesized that bypassing this ubiquitin-dependent recruitment by artificially tethering NAP1 to the outer mitochondrial membrane might be sufficient to initiate autophagosome biogenesis.
To test this hypothesis, we employed a chemically induced dimerization (CID) assay, wherein FRB and FKBP can be dimerized upon rapalog addition [49,50]. By positioning FRB on the mitochondrial outer membrane through fusion with the transmembrane domain of FIS1 and attaching NDP52 or NAP1 to FKBP, we gained the ability to redirect NAP1 or NDP52 to the outer mitochondrial membrane upon rapalog treatment (Fig. 4A).
We first confirmed that NDP52 induced mitophagy upon rapalog addition (Fig. 4B), as previously demonstrated [30,51]. We then evaluated whether FKBP-NAP1 could similarly initiate mitophagy. Intriguingly, artificial tethering of NAP1 to the mitochondrial surface resulted in comparable levels of mitophagy induction upon rapalog treatment as compared to NDP52 (Fig. 4B). To rule out that this effect stemmed from the indirect recruitment of NDP52 by NAP1, we repeated the experiment in pentaKO cells. This confirmed that NAP1 could autonomously induce mitophagy, independently of NDP52 (Fig. 4B). Moreover, blocking autophagosome formation with a Vps34 inhibitor or impeding autophagosome degradation with Bafilomycin A1 validated that the mitochondrial turnover was mediated by autophagy (Fig. 4C).
Collectively, these findings highlight the resemblance of NAP1/SINTBAD to cargo receptors, with the exception of ubiquitin binding. By artificially tethering NAP1 to the mitochondrial surface, we demonstrated its competency as an autophagy cargo receptor in a TBK1-dependent manner. Based on these insights, we propose the term "cargo co-receptors" for NAP1/SINTBAD, emphasizing their ability to facilitate selective autophagy through interactions with cargo receptors like NDP52. NAP1/SINTBAD restrict mitophagy by competing with OPTN for TBK1 binding and activation While our findings above underscore the importance of NAP1/SINTBAD for NDP52driven selective autophagy pathways, these results do not explain our earlier observations in NAP1/SINTBAD DKO cells, where their overall effect on mitophagy was inhibitory rather than stimulatory. This suggests that the roles of NAP1/SINTBAD in mitophagy might be cargo receptor-specific, considering that NAP1/SINTBAD DKO cells express all five cargo receptors, while experiments in the pentaKO background were conducted in cells expressing only NDP52. Based on the fact that NAP1/SINTBAD bind to TBK1 at the same binding site as OPTN [44], we hypothesized that their inhibitory impact on mitophagy might arise from direct or indirect regulation of OPTN, the other major cargo receptor in PINK1/Parkin-dependent mitophagy.
To test whether NAP1/SINTBAD could inhibit mitophagy by competing with OPTN for TBK1 binding, we reconstituted the initiation of OPTN-driven mitophagy in vitro using purified components. Agarose beads coated with linear 4x ubiquitin, mimicking the surface of ubiquitinmarked damaged mitochondria, were co-incubated with mCherry-tagged OPTN, EGFPtagged TBK1, and increasing concentrations of NAP1 (Fig. 5A). This experiment revealed that OPTN was recruited to the ubiquitin-coated beads, subsequently recruiting TBK1 (Fig. 5B).
However, increasing NAP1 levels led to TBK1 displacement from the OPTN-bound beads, indicating that OPTN and NAP1 compete for the same binding site. This competition was further validated through conventional pull-down experiments (Fig. S3).
To assess whether NAP1/SINTBAD also competed with OPTN for TBK1 binding in cells, we used the NAP1/SINTBAD DKOs in the pentaKO background, where OPTN was reintroduced. This setup allowed us to distinguish the effects of NAP1/SINTBAD on OPTNmediated mitophagy from those on NDP52-mediated mitophagy. Following mitophagy induction in these cells, we observed increased TBK1 activation as indicated by higher levels of p-S172 TBK1 in the absence of NAP1/SINTBAD (Fig. 5C). This suggests that NAP1/SINTBAD suppress TBK1 activation in OPTN-mediated mitophagy, consistent with their competition with OPTN for TBK1 binding (Fig. 5B). We then examined whether increased TBK1 activation resulted in accelerated mitophagy. Indeed, measurements of mitophagy levels, indicated by COXII degradation in pentaKO cells rescued with OPTN, confirmed the acceleration of OPTN-driven mitophagy in the absence of NAP1/SINTBAD (Fig. 5D).
Next, we quantified the amount of activated TBK1 relative to total TBK1 on the surface of purified mitochondria in wild-type cells versus NAP1/SINTBAD DKO cells expressing all five cargo receptors. This revealed increased TBK1 activation upon NAP1/SINTBAD deletion (Fig.   5E), suggesting that NAP1/SINTBAD are indeed competing with OPTN for TBK1 binding in the cell.
To further validate that NAP1/SINTBAD inhibit mitophagy, at least in part, through competition for TBK1 binding, we engineered a NAP1 mutant (L226Q/L233Q) deficient in TBK1 binding (Fig. S5) and assessed its inhibitory potential. Upon overexpression of wild-type NAP1 or the TBK1-binding deficient mutant in wild-type HeLa cells, we observed that wildtype NAP1 reduced the overall COXII degradation, as observed earlier (Fig. 2E). However, this effect was nearly completely abolished for the TBK1-binding deficient mutant (Fig. 5F), further supporting the notion that NAP1/SINTBAD restrict mitophagy initiation through competition for TBK1 binding.
Taken together, these findings reveal that NAP1/SINTBAD, through competition for TBK1 binding, can restrict the initiation of mitophagy.

OPTN is the primary recruiter and activator of TBK1 during mitophagy initiation
The insights gathered above not only unveil a novel regulatory step at the onset of mitophagy, but also shed light on a critical role for TBK1 in ensuring the efficient progression of mitophagy. Our results show that NAP1/SINTBAD restrict mitophagy initiation by limiting the TBK1 recruitment by OPTN, hinting at a dominant role for OPTN in recruiting and activating TBK1. We therefore set out to dissect the underlying mechanisms of TBK1 recruitment and activation during mitophagy.
Consistent with prior research, we first confirmed that the activation of TBK1 strictly relies on the presence of cargo receptors, as their absence resulted in the absence of TBK1 activation (Fig. 6A) [30]. Furthermore, in line with the mechanism by which TBK1 is activated through local clustering on the ER surface by the cGAS-STING complex [52][53][54][55][56], we find that TBK1 is also activated locally on the mitochondrial surface during mitophagy (Fig. 6B). This aligns with the requirement of TBK1 dimers to be brought into close proximity, enabling trans autophosphorylation, as the kinase domain cannot access the activation loop in cis and the two kinase domains in the dimer face away from one another [57][58][59]. Our data thus propose an essential role for cargo receptors in locally clustering TBK1 dimers on the mitochondrial surface. This is consistent with a recently proposed model, positing that TBK1 is activated from a local platform of OPTN molecules [60].
To test whether TBK1 activation predominantly relies on OPTN, as implied by our NAP1/SINTBAD results, we compared TBK1 activation in wild-type HeLa cells to cells lacking either OPTN or NDP52. This comparison revealed a severe reduction in TBK1 activation upon OPTN deletion as evident from decreased TBK1 phosphorylation (Fig. 6C). In contrast, NDP52 deletion had a relatively minor impact on TBK1 activation (Fig. 6C). We then compared the amount of TBK1 recruitment during OPTN-versus NDP52-driven mitophagy by rescuing the pentaKO cells with either OPTN or NDP52. This revealed OPTN's pronounced ability to recruit TBK1 to the mitochondrial surface upon mitophagy induction, while NDP52 recruited TBK1 to a lesser extent (Fig. 6D).
To further corroborate this result, we employed the ALS-causing TBK1 E696K mutation. This mutant failed to bind OPTN in vitro, in line with prior research [39,44,61,62].
However, this mutation retained its binding capacity to NAP1 (Fig. 6E). In wild-type HeLa cells, expressing both OPTN and NDP52, the TBK1 E696K mutant was previously shown to be no longer recruited to damaged mitochondria [39,61]. Consistently, we show that this is accompanied by a drastic reduction of TBK1 activation (Fig. 6F), reinforcing the importance of clustering for TBK1 activation. Moreover, these findings are also consistent with OPTN playing a primary role in recruiting and clustering TBK1 on the mitochondrial surface, which cannot be sufficiently compensated for by the NDP52-NAP1/SINTBAD axis in HeLa cells. This underscores the importance of OPTN-mediated TBK1 recruitment.
Together, our results provide evidence for a crucial role of OPTN in recruiting and activating TBK1 during mitophagy, explaining how interference with this interaction by NAP1/SINTBAD can effectively restrict mitophagy initiation.

Crosstalk between the OPTN-axis and NDP52-axis stimulates mitophagy
We wondered whether the crucial role of OPTN in TBK1 activation might also influence the NDP52 axis. Previous research revealed that either cargo receptor alone is sufficient to initiate mitophagy [30]. However, several tissues express both cargo receptors. In tissues such as the brain, where NDP52 expression is low [30], the NDP52-related protein TAX1BP1 is expressed. We therefore hypothesized that a crosstalk might exist between OPTN and NDP52, allowing each receptor to leverage its strengths so that their combined presence results in robust mitophagy control and progression.
To be able to test this, we designed a system that enabled us to exploit OPTN's capacity to recruit TBK1 during mitophagy, but omitting its ability to interact with other components of the autophagy machinery [41,63,64]. To this end, we created a rapaloginduced dimerization assay, linking only the minimal sequence of OPTN (residues 2-119) essential for TBK1 binding to FKBP (Fig. 7A). Using this system in the pentaKO background, we tested whether this truncated OPTN fragment could effectively recruit and activate TBK1 at the mitochondrial surface. Indeed, purification of mitochondria from rapalog-treated HeLa cells revealed that rapalog induced the translocation of FKBP-OPTN(2-119) and TBK1 (Fig.   7B). Crucially, the recruitment of TBK1 to the mitochondrial surface was sufficient to induce TBK1 activation, as demonstrated by the increase in phosphorylated TBK1 in the mitochondrial fraction upon rapalog treatment.
We then assessed the extent to which this minimal OPTN peptide could initiate mitophagy, as most of its essential autophagy-driving protein domains had been removed.
Yet, treating cells with rapalog for 24 h led to a notable fraction of cells undergoing mitophagy, as demonstrated by the mt-mKeima conversion (Fig. S6), albeit to a lesser extent than with full-length OPTN. This observation shows that recruitment of TBK1 is sufficient for mitophagy initiation, consistent with our earlier finding that it can recruit the PI3KC3C1 complex [40].
With this minimal OPTN peptide at hand, we sought to elucidate whether TBK1 recruited through this truncated OPTN axis could enhance NDP52-driven mitophagy. To test this hypothesis, we rescued pentaKO cells with NDP52 and further transduced them with FKBP-OPTN(2-119) and Fis1-FRB. This experimental setup enabled us to measure mitophagy rates by NDP52 upon mitochondrial depolarization by O/A, both in the presence and absence of additional TBK1 recruited through rapalog treatment. While rapalog alone resulted in relatively slow mitophagy activation, displaying only minimal activation from 4 hours onwards, the combined treatment of O/A and rapalog substantially accelerated mitophagy flux ( Fig. 7C and S7). Importantly, this increase in mitophagy flux was not a merely additive effect, based on the kinetics of rapalog treatment alone, especially during the first three hours of treatment where we observed minimal mitophagy induction by rapalog alone, suggesting that the recruitment of TBK1 by OPTN synergistically enhances NDP52-driven mitophagy in cells.
This underscores the pivotal role of TBK1 recruitment by OPTN, not only for OPTN's own function but also for NDP52-mediated mitophagy, as the proximity of TBK1 recruitment by OPTN likely also augments NDP52-mediated mitophagy.
To dissect the interplay between OPTN and NDP52 further, we conducted biochemical reconstitution experiments using agarose beads coated with GST-4xUb to mimic damaged mitochondrial surfaces. We incubated these beads with OPTN, TBK1, and NAP1 in the presence or absence of NDP52. This confirmed that NAP1 negatively regulates the recruitment of TBK1 towards ubiquitin-bound OPTN in the absence of NDP52, as we showed above (Fig. 5A). However, when we added NDP52 to concentrations of NAP1 that would prevent any detectable TBK1 recruitment to the beads, we observed restoration and even a trend towards a slight enhancement of the TBK1 signal on the beads (Fig. 7D). To confirm the specificity of the NAP1 sequestration by NDP52, we replaced wild-type NAP1 with a NDP52binding mutant and observed complete disappearance of TBK1 signal on the ubiquitin-coated beads (Fig. 7D). This suggests that the interaction of NAP1 with NDP52 on the cargo allows the mitophagy machinery to overcome the inhibitory effect of NAP1 in terms of TBK1 recruitment during mitophagy initiation.
To assess whether TBK1 indeed converted from binding the OPTN-axis to the NDP52-NAP1/SINTBAD axis, we performed a reverse pull-down experiment using GFP-trap beads coated with EGFP-TBK1. This confirmed that the relative amounts of NAP1 and OPTN determined which of the two cargo receptors, OPTN versus NDP52, was predominately recruited to TBK1 (Fig. 7E). These results suggest that NDP52 can support OPTN-driven mitophagy by harnessing NAP1/SINTBAD to recruit further TBK1.
In summary, our findings propose a model in which NAP1/SINTBAD initially set a threshold for mitophagy activation by constraining TBK1 activation via the mitophagy receptor OPTN (Fig. 8). This is because OPTN fulfills a primary role in recruiting TBK1 during mitophagy. However, when mitochondrial damage is severe enough, NAP1/SINTBAD transition into a supportive role, acting as cargo co-receptors that bolster NDP52-driven mitophagy. Their sequestration by NDP52 increases TBK1 activation through increased recruitment by OPTN, and this, in return, then boosts NDP52-driven mitophagy again due to the crosstalk from OPTN-TBK1 towards the NDP52-axis, providing an effective feedforward loop once the mitophagy pathway is set in motion.

DISCUSSION
The regulatory mechanisms that prevent PINK1/Parkin-dependent mitophagy and other selective autophagy pathways from overreacting while ensuring swift progression once initiated are largely elusive. By focusing on the roles of the TBK1 adaptors NAP1/SINTBAD, we uncovered how tightly they are interwoven into this pathway by regulating key activities of the OPTN and NDP52 cargo receptors in completely different ways. In particular, we find that NAP1/SINTBAD act as rheostats, which inhibit mitophagy initiation by restricting recruitment and activation of TBK1 by OPTN, while enhancing NDP52-mediated engulfment of damaged mitochondria. NAP1/SINTBAD drew our attention due to the central role of TBK1 as a key regulator of selective autophagy pathways and their involvement in supporting NDP52-dependent xenophagy [7,8,10]. In addition, we found that NAP1/SINTBAD are recruited and codegraded with damaged mitochondria (Fig. 1). The observation that deletion of NAP1/SINTBAD in wild-type HeLa cells results in acceleration rather than a deceleration of mitophagy ( Fig. 2) was therefore unexpected. Using cellular and in vitro reconstitutions, we dissected how NAP1/SINTBAD interact with NDP52 and OPTN, the key cargo receptors in PINK1/Parkin-dependent mitophagy [30]. In NDP52-driven mitophagy, they exert a stimulatory role ( Fig. 3A-B), similar to their function in xenophagy, by bridging NDP52 with TBK1 to activate this kinase on the mitochondrial surface. Furthermore, they stabilize the interaction with the core autophagy factor FIP200 (Fig. 3D). In contrast, in OPTN-driven mitophagy, NAP1/SINTBAD counteract TBK1 recruitment and activation by directly competing for the same TBK1 binding site (Fig. 5). Overall, the inhibitory role of NAP1/SINTBAD seems to prevail, as evidenced by the increased presence of activated TBK1 on damaged mitochondria in the absence of NAP1/SINTBAD in cells expressing both OPTN and NDP52 (Fig. 5E).
Our study highlights the central role of TBK1 in coordinating the different mitophagy mechanisms and uncovers an interplay between OPTN-mediated mitophagy and NDP52mediated mitophagy. This interplay suggests a finely tuned regulation of mitophagy, which may be particularly important for specific cell types. For example, in the brain where NDP52 expression is low and OPTN is the primary mitophagy receptor, competition for TBK1 activation may prevent excessive initiation of mitochondrial degradation. Given the postmitotic nature of neurons, an excess in mitochondrial degradation could be as detrimental as insufficient activation. This is exemplified by disease-causing mutations in FBXL4, which results in excessive mitochondrial degradation through the NIX/BNIP3 pathway [65][66][67][68], and which lead to a severe mitochondrial encephalopathy [69,70]. Conversely, in cells expressing both mitophagy receptors OPTN and NDP52, NAP1/SINTBAD initially compete with OPTN for TBK1 binding until mitophagy is adequately activated. Subsequently, NAP1/SINTBAD convert into mitophagy-promoting factors by supporting NDP52.
Our results also highlight OPTN's dominant role in recruiting and activating TBK1 during mitophagy. This is in line with the recent finding that OPTN forms a platform for TBK1 activation from where it engages with TBK1 in a positive feedback loop and observations made with the ALS-causing TBK1-E696K mutant, which has lost its OPTN-binding capacity [39,60,61], but not its binding to NAP1 as we show. Although NDP52 can recruit and activate TBK1 in the absence of other cargo receptors, our findings indicate that when OPTN and NDP52 are co-expressed, mitophagy is accelerated when OPTN can more easily recruit TBK1. This hints at a two-tiered mechanism, with OPTN being the first cargo receptor to drive mitophagy at very early stages, followed by NDP52 in a second phase. While the existence of such a mechanism in mitophagy remains in part speculative at this point, previous work has indicated that OPTN and NDP52 are not kinetically interchangeable, with OPTN being more dominant for mitophagy at early time points [39]. Two-step cargo receptor recruitment has also been observed in xenophagy in which NDP52 is recruited initially to invading pathogens via recognition of exposed Galectin 8 molecules [42,71], subsequently leading to the recruitment of the E3 ubiquitin ligases, such as LUBAC and LRSAM1, which coat the bacterial surface with poly-ubiquitin chains [72][73][74]. This, in turn, triggers the recruitment of other cargo receptors like OPTN and SQSTM1/p62 [41,75]. Future research should address whether a similar two-step recruitment mechanism or other diversification mechanisms between cargo receptors underlie our findings.
Additionally, the recent identification of TNIP1 as another mitophagy inhibitor [76] suggests that the inhibitory effect of NAP1/SINTBAD may constitute a more widespread mechanism. TNIP1 was proposed to compete with autophagy receptors for FIP200 binding, which is distinct from how NAP1/SINTBAD inhibit mitophagy, as our results demonstrate that they instead strengthen the NDP52-FIP200 interaction. Nevertheless, identifying these regulatory steps during the early steps of autophagosome biogenesis could offer new therapeutic opportunities, especially in conditions where damaged mitochondria are insufficiently cleared.
In summary, our study uncovers an unexpected additional layer of regulation governing mitophagy initiation and expands our understanding of the complex interplay among various players involved in maintaining mitochondrial quality control. This additional layer may enable cells to respond better to cellular demands and may offer new opportunities for developing new therapeutic strategies aimed at modulating mitophagy in various pathological conditions associated with mitochondrial dysfunction.

Reagents
The following chemicals were used in this study: Oligomycin (A5588, ApexBio), Antimycin A

Plasmid Construction
The sequences of all cDNAs were obtained by amplifying existing plasmids, HAP1 cDNA, or through gene synthesis (Genscript). For insect cell expressions, the sequences were codon optimized and gene synthesized (Genscript). With the exception of the NAP1-6xAla mutant, which was obtained through gene synthesis (Genscript), all other plasmids were generated by Gibson cloning, For Gibson cloning, inserts and vector backbones were generated by PCR amplification or excised from agarose gels after restriction enzyme digestion at 37°C for two hours. The inserts and plasmid backbones were purified with Promega Wizard SV gel and PCR Cleanup System (Promega). Purified inserts and backbones were mixed in a molar 3:1 ratio, respectively, supplemented by a 2x NEBuilder HiFi DNA assembly enzyme mix (New England Biolabs). Gibson reactions were incubated for one hour at 50°C and then transformed into DH5-alpha competent E. coli cells. Transformed Gibson reactions were grown overnight on agar plates containing the appropriate selection marker (ampicillin, kanamycin, or chloramphenicol). Single colonies were picked, grown overnight in liquid cultures, and pelleted for DNA plasmid extraction using the GeneJet Plasmid Miniprep kit (Thermo Fisher). The purified plasmid DNA was submitted for DNA Sanger sequencing (MicroSynth AG). All insert sequences were verified by Sanger sequencing. Positive clones were further analyzed by whole plasmid sequencing (Plasmidsaurus). A detailed protocol is available (https://doi.org/10.17504/protocols.io.8epv5x11ng1b/v1).

Cell lines
All cell lines were cultured at 37°C in humidified 5% C02 atmosphere. HeLa

Generation of CRISPR/Cas9 knockout cells
All knockout cell lines were generated using CRISPR/Cas9. Candidate single-guide RNAs

Nutrient starvation experiments
To induce bulk autophagy, cells were starved by culturing them in Hank balanced salt medium

Rapalog-induced chemical dimerization experiments
The For Rapalog-induced mitophagy experiments, cells were seeded as described above and treated for 24 h with 500 nM Rapalog A/C hetero-dimerizer (Takara). Cells were collected as described above, and the mt-mKeima ratio (561/405) was quantified by an LSR Fortessa Cell Analyzer (BD Biosciences). The cells were gated for GFP/mt-mKeima double-positive cells.

Cellular fractionation and mitochondrial isolation
HeLa cells were seeded in 15 cm dishes and grown until confluence. Cells were treated with DMSO or O/A for the indicated time. Mitochondria were isolated as described previously [77].

Immunofluorescence and confocal microscopy
Cells were seeded on HistoGrip (Thermo Fisher) coated glass coverslips in 24 well plates.
Cells were allowed to adhere overnight and treated as indicated prior to fixation. Cells were fixed in 4% (w/v) paraformaldehyde (PFA), diluted in 100 mM phosphate buffer, for 10 min at room temperature. The PFA was removed, and samples were washed three times with 1x PBS. Cells were permeabilized with 0.1% (v/v) Triton X-100, diluted in 1x PBS, for 10 min.

Protein expression and purification
Linear tetra-ubiquitin fused to GST (GST-4xUb) was cloned into a pGEX-4T1 vector  To purify NAP1 or GST-NAP1, human NAP1 cDNA was synthesized and cloned in a pcDNA3.1 vector (Genscript), from where it was subcloned into a pGEX-4T1 vector with an N-terminal GST tag followed by a TEV cleavage site (RRID:Addgene_208870). For expression of unlabeled NAP1 in E. coli (which we used in Figure 5B, Figure 7E, and Figure   S5) or GST-NAP1 (which we used in Figure 6E and Figure S3)  To purify MBP-NAP1, human NAP1 cDNA was gene-synthesized (by Genscript) and

Quantification and statistical analysis
For the quantification of immunoblots, we performed a densitometric analysis using Fiji software. Graphs were plotted using Graphpad Prism version 9.

Data availability statement
Raw files associated with this work will be made available on Zenodo by time of final publication.
Code availability statement      Experimental results obtained by confocal imaging. One of three representative experiments is shown. (E) Pull-down assay of mCherry-OPTN, mCherry-NDP52, and NAP1 by GFP-TBK1. GFP-TBK1 was pre-loaded onto GFP-Trap beads and then incubated with the protein mixtures as indicated. The relative amounts of mCherry-OPTN, NAP1, and mCherry-NDP52 bound to TBK1 were quantified for the indicated lanes and plotted (right). Data are shown as mean ± s.d. from three independent experiments or as one of three representative Coomassie-stained gels for (E).

Figure 8. Working model for mitophagy initiation in cells expressing both mitophagy receptors OPTN and NDP52
(1) Cargo receptors OPTN and NDP52 are recruited to damaged mitochondria upon accumulation of ubiquitin and phospho-ubiquitin on their surface. OPTN recruits TBK1 but is restricted by NAP1, which competes with OPTN for TBK1-binding. (2) However, NDP52 recruits NAP1 to the mitochondrial surface and sequesters NAP1 away, allowing OPTN to recruit and activate more TBK1.  Results are representative of one of three replicates.

Figure S3. Pull-down of GFP-TBK1 with mCherry-OPTN and GST-NAP1
Pull-down assay of mCherry-OPTN versus GST-NAP1 by GFP-TBK1. GFP-TBK1 was preloaded onto GFP-Trap beads and then incubated with the protein mixtures as indicated. The relative amounts of mCherry-OPTN and GST-NAP1 bound to TBK1 were quantified for the indicated lanes and plotted (right).