The MAP4 Projection Domain Accelerates Hypoxia-Induced Mitophagy Disruption through LIR Motif in Cardiomyocytes

Previously, we and other investigators have demonstrated that phosphorylated microtubule-associated protein 4 (p-MAP4) impacts myocardial hypertrophy and ischemic heart failure. However, the detailed mechanism behind this remains under elucidated. Published studies have suggested that impaired mitophagy contributes to hypoxia-induced myocardial damage, hence the involvement of p-MAP4 in mitophagy in cardiomyocytes was investigated. The results herein revealed that there was increased degradation of mitochondria, accumulated mitophagosomes and disrupted autophagic ux in both neonatal and adult ones of MAP4-knockin (KI) mice. This indicated that p-MAP4 persistently degraded mitochondria through activating mitophagy. Next, Tom70 was found as the importer of p-MAP4 in the context of mitochondrial translocation. And, the LC3-interacting region (LIR) motif (47–50aa) caused p-MAP4-induced mitochondrial engulfment, and the ubiquitin-interacting motif (UIM) domain determined the characteristics of p-MAP4-induced mitophagosomes, which were structure and membrane potential-independent. Moreover, p-MAP4 enhanced hypoxia-induced mitophagic ux impairment, and p-MAP4 LIR (47–50aa) mutation decreased hypoxia-induced autophagy both in MAP4 knockout and wildtype cardiomyocytes. Overall, this study identied that p-MAP4 as a novel mediator and cargo receptor in mitophagy, and that the degradation of the MAP4 PJ domain as a promising therapeutic target for improving the cardiac function of hypoxia-related heart failure or cardiac remodelling.


Introduction
Mitophagy, an essential part of selective autophagy, is a highly regulated process involved in the degradation of damaged or redundant mitochondria 1 . To date, two mitophagy mechanisms have been proposed. One depends on PTEN-inducible putative kinase 1 (Pink1)-Parkin pathway for the ubiquitination of mitochondrial proteins followed by interactions with the adaptor protein p62 (SQSTM1), which is responsible for connecting ubiquitin with LC3 (a marker of phagophores) 2 . The other mechanism relies upon either proteins or lipids on the outer mitochondrial membrane as receptors for LC3. NIP3-like protein X (Nix/Bcl-2L), Bcl-2, and FUN14 domain containing 1 (FUNDC1) have been shown to act as LC3 receptors and are known as LC3-binding proteins [3][4][5] . The LIR and the UIM have been regarded as the functional domains in the LC3-binding proteins, which underly binding to the LC3 LIR/AIM docking site (LDS) or UIM/AIM docking site (UDS) 6-7 . LIR is a short linear motif that consists of up to 13 amino acids with a W/F/Y-XX-L/I/V core sequence (where X indicates sites that can be replaced by any amino acid) 7 . The UIM domain consists of 20 amino acids with a principal sequence of XXXXX-L-XX-A-XXX-S-XXXXXXX, which is associated with the induction of ubiquitination and binding to ubiquitinated proteins 8 .
In mitophagy, the interaction between LIR or UIM-containing proteins and LC3 is the core step, which promotes the engulfment of mitochondria by autophagosomes. Importantly, Huang et al 9 proposed that mitophagy played a role in the regulation of heart function in a myocardial ischemia-reperfusion model. Therefore, a complete understanding of mitophagy in the context of cardiac hypoxia, as well as the associated regulatory mechanisms, is essential for the development of new therapeutic interventions to treat hypoxia-induced heart dysfunction.
Microtubules (MTs) are thought to participate in autophagy processes, including the formation of autophagosomes and their fusion with lysosomes 10,11 . Microtubule-associated proteins (MAPs) are cytosolic skeleton proteins that are important for the polymerisation of MTs 12 . MAP4 is among the most extensively studied MAPs, and it is ubiquitously expressed in non-neural cells and becomes functional upon phosphorylation (p-MAP4) [12][13][14] . MAP4 consists of a N-terminal projection (PJ) domain and a Cterminal MT-binding (MTB) domain 15 . The PJ domain is further subdivided into three distinct regions, the Na-region (the N-terminal 247 amino acid residues), the KDM-region (344 amino acid residues from 248 to 592) and the b-region (94 amino acid residues from 593 to 687) 15 . The MTB domain is also divided into the proline-rich region (Pro-rich), the assembly promoting-repeated (AP) sequence region and the hydrophobic tail region 16 . To date, studies on MAP4 have mainly focused on the MTB domain, which can be phosphorylated (at S667, S737, and S760) and leads to MT depolymerisation, myocardial apoptosis and myocardial hypertrophy 13,17,18 . It is noteworthy that p-MAP4 (at the MTB domain) was shown to translocate into the outer mitochondrial membrane leading to mitochondrial injury and the initiation of apoptosis 18 . However, its effects on mitophagy and the function of the PJ domain remain obscure.
Heart failure is associated with a high mortality rate (50% for 5 years) no matter whether it is induced by pathological cardiac remodelling, hypoxia, ischemic-reperfusion injury or other diseases 17 . Previous investigations, by both others and ourselves, have demonstrated that p-MAP4 caused myocardial hypertrophy and ischemic heart failure 17,19 . However, the mechanisms underlying the incidence of hypoxia induced heart failure requires further investigation. Recent data has highlighted mitophagy as one of the principal mediators of heart function under several pathological conditions, including hypoxia and myocardial ischemia-reperfusion injury 9 . Moreover, p-MAP4 was recently shown to translocate into mitochondria in cardiomyocytes under hypoxic conditions 18 . Thus, to determine the possible mechanisms of p-MAP4 induced myocardial damage, the effects of p-MAP4 on mitophagy in cardiomyocytes were investigated herein.

Results
Increased p-MAP4 degrades mitochondria through impacting mitophagy in the whole life of mice MAP4-knock in (KI) mice were used to determine the effects of MAP4 phosphorylation on mitophagy in the context of normoxia, the protocol for the generation of these mice has previously been described 17 . As shown in Fig. 1A, increased phosphorylation of MAP4 was observed at residues S737 and S760 in the myocardium of MAP4-KI mice (P=<0.05). Next, the mitochondrial proteins (TOM20, VDAC1 and TIM23) were signi cantly decreased in the myocardium of MAP4 KI mice, but the endoplasmic reticulum (calnexin) and Golgi marker (GM130) were not changed (Fig. 1B, P=<0.05). Meanwhile, elevated expressions of LC3II, Atg3, Atg7, Beclin1, Bcl-2, and P62 were observed in the myocardium of MAP4-KI mice, which correlated with an increased number of mitophagosomes ( Fig. 1C and 1D, p<0.05). These results demonstrated that p-MAP4 degraded mitochondria and activated mitophagy in adult mice.
Next, to investigate the effects of p-MAP4 on mitophagy in neonatal mice, the primary cardiomyocytes of neonatal MAP4 KI mice were isolated. Importantly, similar signi cant expressions increases were observed for autophagy-related proteins in the isolated primary cardiomyocytes of MAP4 KI mice (Fig. 1E, P=<0.05). Furthermore, increased mitophagy activation (with GFP-LC3 and Dsred-Mito signals) and more mitophagosomes using TEM (Transmission Electron Microscopy) were observed. Slightly disrupted autophagic ux (based upon mRFP-GFP-LC3 signals) was also observed ( Fig. 1F-J, P=<0.01). Together, these results indicated that p-MAP4 degraded mitochondria by impacting mitophagy both in neonatal and adult mice under normoxia, hinting the in uences of p-MAP4 on mitophagy in cardiomyocytes were persistent.

Tom70 Mediates The Mitochondrial Translocation Of P-map4
The mitochondrial translocation of p-MAP4 has previously been reported to be the rst step of mitophagy initiation 18 . To determine the mechanism behind the mitochondrial translocation of p-MAP4, a qualitative (shotgun) protein analysis was used to detect MAP4-interacting proteins in primary cardiomyocytes. This analysis revealed that Tomm70/Tom70 and Hspd1/HSP60 were the potential importers of MAP4 into the mitochondria ( Fig. 2A). As demonstrated in Fig. 2B, Ad-MAP4 (Glu) cardiomyocytes effectively increased the levels of p-MAP4 (at both the S737 and S760 residues; P=<0.05). Importantly, the expressions of HSP60 and Tom70 were increased in cardiomyocytes overexpressing MAP4 (Glu) or those subjected to hypoxia (Fig. 2C). Moreover, speci c siRNAs targeting Tom70 and HSP60 were constructed that e ciently knocked down their expressions (Fig. 2D). As shown in Fig. 2E-F, Tom70 siRNA signi cantly decreased Ad-MAP4 (Glu) and hypoxia-induced high LC3II expression in extracted mitochondrial proteins. Importantly, knocking down Tom70 decreased the co-localisation of p-MAP4 with the mitochondria and mitophagosomes ( Fig. 2G-J, P=<0.05). Together, these results demonstrated that p-MAP4 translocation to the mitochondria (and thereby mitophagy activation) was Tom70 dependent.
The UIM domain maintains the speci c characteristics of p-MAP4 induced mitophagy Firstly, LC3 was found to co-localised in linear mitochondria in Ad-MAP4 (Glu)-transduced or hypoxia treated cardiomyocytes, but not in Carbonyl cyanide 4-(tri uoromethoxy)phenylhydrazone (FCCP)-treated ones (Fig. 4A). Notably, the number of such mitophagosomes was highest in cardiomyocytes treated by Ad-MAP4 (Glu) (Fig. 4C). Moreover, the engulfed mitochondria showed high membrane potential in Ad-MAP4 (Glu)-or hypoxia-treated cardiomyocytes, but not in FCCP-treated ones (Fig. 4B). Again, the proportion of engulfed mitochondria with high membrane potential was highest in Ad-MAP4 (Glu) treated cardiomyocytes when compared to the hypoxia or FCCP treated ones (Fig. 4D). These results indicated that p-MAP4-induced mitophagy was neither structurally nor potential-dependent.
Further studies were conducted focusing upon the MAP4 UIM domain because ubiquitination has emerged as a universal cargo recognition signal, although it appears irrelevant to mitochondrial function 22 . It's noteworthy that the level of ubiquitination was increased under normoxia in the myocardium, primary cardiomyocytes and mitochondrial membrane of MAP4 KI mice ( Fig. 4E and S1F). To verify the function of the MAP4 UIM domain, the mutual effects of MAP4 and ubiquitin were explained, and determined to be increased in Ad-MAP4 (Glu)-treated cardiomyocytes (Fig. 4F).
Then, the effects of p-MAP4 on hypoxia induced mitophagy in vivo in adult MAP4-KI mice were further investigated. As shown in Fig. 5H, there were higher expressions (P=<0.05) of LC3-II, Atg3, Atg7, Atg16L1, Beclin1, Bcl-2 and p62 in the myocardium of MAP4-KI mice after expose to hypoxia (when compared to wild-type mice), which was also associated with more mitophagosomes (Fig. 5I). These results were in line with the in vitro data.
Next, the in uences of p-MAP4 inhibition on hypoxia induced autophagy were investigated. As a complementary approach, MARK4 was silenced in cardiomyocytes using speci c siRNAs, because it has been identi ed as an upstream regulator of p-MAP4 19 . As shown in Fig. 5K, MARK4 was effectively silenced and the reduction of its expression led to a decline in the overexpressed LC3-II in cardiomyocytes under hypoxia in vitro. These data indicated p-MAP4 inhibition decreased hypoxia activated autophagy.

Discussion
Mitochondrial dysfunction is a crucial contributor to myocardial injury. Previous research revealed that p-MAP4 caused myocardial hypertrophy and ischemic heart failure, which was dependent upon mitochondrial dysfunction 17,19 . Autophagy, including mitophagy and neutrophil autophagy, is essential for the degradation of disrupted mitochondria and is consequently important for the regulation of myocardial function [23][24][25] . We have previously reported that p-MAP4 translocated to mitochondria in hypoxic cardiomyocytes 18 ; however, the role of p-MAP4 on mitophagy in cardiomyocyte had not previously been explored. Herein, p-MAP4 was demonstrated to mediate mitochondrial engulfment through its LIR (47-50aa), and maintain structure-and voltage-independent characteristics through its UIM domain. This study highlights MAP4 not only as a new LIR and UIM domain-containing protein, but also as a novel autophagic cargo acceptor. It is noteworthy that these domains predominantly locate within the MAP4 PJ domain and are activated by phosphorylation of the MTB domain. This discovery indicates that the PJ domain is responsible for the core functionality of MAP4 in mitophagy. Moreover, p-MAP4 suppresses the disruption of mitophagic ux induced by hypoxia. Overall, this article identi ed MAP4 as another novel cargo receptor and that the degradation of MAP4 PJ domain is a promising therapeutic target to improve the cardiac dysfunction of hypoxia-related cardiac remodelling or heart failure.
The N-terminus has been reported to be the functional region for the vast majority of molecules 26 . Therefore, the role that the PJ domain (which is in the N-terminal region of MAP4) was determined to play in this study is not unprecedented. Moreover, three new domains within the MAP4 PJ domain were also described, namely: the BH3, LIR and UIM domains. The importance of the PJ domain is also supported by the ndings by Junko et al 15 , which revealed that the PJ domain represses the microtubule-bundling activity of the MTB domain. In the present study, the PJ domain was shown to strongly interact with both LC3 and Bcl-2 (Fig. 3). Together, this data strongly supports the hypothesis that the PJ domain is arguably the most biologically important MAP4 domain. It is notable that the phosphorylation of the MTB domain underlies the MAP4/LC3 and MAP4/Bcl-2 interactions ( Fig. 3 and 4). This study also demonstrates that MAP4 is a cargo receptor of mitophagy in cardiomyocytes through its characteristic LIR domain and speci cally interacts with LC3. Under both physiological and hypoxic conditions, MAP4 impacts mitophagy after its phosphorylation at S737 and S760, resulting in increased co-localisation and interaction of MAP4/LC3. This leads to the selective mitochondrial incorporation into LC3-bound isolation membranes for the subsequent removal of mitochondria. It is noteworthy that mitophagic cargo receptors use LIR or UIM motif that associates with the LDS or UDS site on LC3, respectively, which culminates in the formation of mitophagosomes 6, 8, 28 . Importantly, canonical LIR and UIM motifs were found within the MAP4 PJ domain (Fig. 3C and 3H). However, detailed interaction experiments con rmes that only the MAP4 LIR motif (47-50aa) is responsible for LC3/MAP4 interactions, mitochondrial engulfment and mitochondrial degradation (Fig. 3). This underlies the structural foundation for p-MAP4-induced mitophagy. Our ndings thus provide insights into how MAP4 acts as a mitophagy receptor to couple with the core autophagic machinery and how MAP4-dependent mitophagy is regulated by phosphorylation.
In this article, MAP4-induced mitophagy was shown to be structure and potential-independent. Phagophore formation and mitochondrial recognition are therefore the key aspects of these mechanisms, of which MAP4 was shown to stimulate the former, resulting in phagophore over-accumulation. In turn, the over-accumulation of phagophores make the cells unable to accurately distinguish healthy from impaired mitochondria. In fact, a UIM domain is found in MAP4, which is notable since UIM domain leads to the ubiquitination of mitochondrial membrane proteins and functions as the bridge between them and proteins in phagophores 6, 8 . Therefore, proteins with canonical UIM domains, other than MAP4, may also induce mitophagy in a structure-and potential-independent manner.
Based upon its distribution and structural status, and the results from this study, combined with the fact that MAP4 is conserved in higher eukaryotic organisms and ubiquitously expressed in non-neural cells [12][13] , MAP4 appears to be a quali ed cargo receptor in mitophagy. However, some differences exist between MAP4 and other known mitophagy cargo receptors. Traditionally, NIX and FUNDC1 are involved in hypoxia-induced autophagy and are indispensable for the programmed elimination of mitochondria during reticulocyte maturation 29 . Yet, based upon our ndings the mechanism of MAP4-induced mitophagy is distinct from that of NIX or FUNDC1. Unlike NIX or FUNDC1, MAP4 contains BH3, LIR, and UIM domains, indicating it can function in the entire process of autophagy (including autophagy activation and cargo recognition) by itself. Additionally, whilst all these proteins are involved in hypoxia induced mitophagy, the expression of NIX or FUNDC1 is altered, whereas MAP4 remains unchanged. However, unlike FUNDC1, both Nix and MAP4 are distributed in other organelles under physiological conditions and are involved in regulation of mitochondrial apoptosis under hypoxia conditions 18, 30 . Therefore, whilst the roles of MAP4, NIX, and FUNDC1 in the induction of mitophagy are different to some extent, they probably cooperate with each other during this process.
Based upon the results contained herein, the increased and decreased MAP4 phosphorylation, accelerates or mitigates (respectively) hypoxia-induced mitophagy impairment (Fig. 6). This is precedented given that the activity of MAPT/Tau (MAP4 a family member) is also known to disrupt autophagic ux 31 . Abnormal mitophagy has been closely associated with heart failure, ischemiareperfusion injury and cardiac hypertrophy 30 . Considering that the mechanism of MAP4-related cardiac hypertrophy and ischemic heart failure remains obscure, this study provides powerful evidence to support this notion. This article de nes MAP4 as a potential mediator in cardiac dysfunction, including hypertrophy and heart failure.
We et al 17 and Yu et al 19 demonstrated that MAP4 in uences cardiac pathology, including cardiac hypertrophy and ischemic heart failure. Therefore, it is possible that the MAP4 PJ domain might be the primary region responsible for these effects. Moreover, the results reported herein could explain the mechanism of MAP4-induced myocardial damage and ischemic heart failure, which has profound biological and clinical implications. Additionally, the current study indicates that the MAP4 PJ domain is the core functional region of MAP4 in mitophagy, and it can be successfully synthesised and used in vitro. However, further research is required to determine the precise in uence of the MAP4 PJ domain on heart failure or cardiac remodelling and to fully clarify the relevance of the MAP4 PJ domain in other diseases. Nevertheless, the MAP4 PJ domain appears to be a promising therapeutic target for a degradation strategy, which could have considerable potential for the treatment/prevention of hypoxiarelated myocardial damage and heart failure.

Animal studies
The MAP4 (S667A, S737E, and S760E)-KI mice were generated and bred as previously described 17 ; wildtype (C57BL6/J) mice were purchased from the Animal Centre, Army Medical University (Third Military Medical University, CHN). Male mice, 8-10-weeks-old and weighing 18-22 g, were used for the experiments. The mice were fed a standard rodent chow diet, allowed access to water ad libitum, and housed under 12-hour light/dark cycles. Prior to the experiments, all animals were allowed to acclimate to the facility for one week. The animals were randomly divided into four groups, control, MAP4-KI, hypoxia, and hypoxia+MAP4 KI. Hypoxia was induced with 7.5% O 2 , 0.1% CO 2 and 92.5% N 2 for seven days. The desired temperature (22℃) was maintained using a Forma Series II Water Jacket CO 2 incubator (model 3131; Thermo Fischer Scienti c, Waltham, MA, USA) and the O 2 level was controlled through a continuous ow of nitrogen. Afterward, the mice were immediately euthanised, and their heart tissues were collected for transmission electron microscopy or western blot analysis.

Mitochondria fractions preparation
Mitochondrial fractions were prepared and validated from primary cardiomyocytes according to the manual instruction of Cell Fractionation Kit (Abcam, ab109719). Brie y, primary cardiomyocytes were collected, counted and diluted to 6.6*10^6 cell/mL. Then, after centrifuged at 500g at 4℃ for 1min, the supernatant was collected and again centrifuged at 10000g at 4℃ for 1min. Then, the pellets were collected and resuspend, and after same centrifugation process, the supernatants containing mitochondrial proteins were collected.

Western blot (WB) analysis
The proteins were extracted as previously reported 11 . Brie y, samples were harvested using lysis buffer containing a protease inhibitor, phosphatase inhibitor and PMSF (Beyotime, China). After quantitated with Quick Start TM Bradford 1×Dye Reagent (Bio-Rad, USA), the protein extracts were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Bio-Rad, USA), and then transferred to polyvinylidene uoride (PVDF) membranes (Bio-Rad, USA). Furthermore, the dissected PVDF membranes were incubated overnight at 4℃ with the following primary antibodies: LC3 (Sigma

Immuno uorescence Staining
The immuno uorescence was performed as previously reported 17 . Cardiomyocytes were xed with 1% paraformaldehyde and then permeabilized in 1% Triton X-100 (Sigma, T9284). The nonspeci c binding sites were blocked with 10% normal goat serum for 60 min. Cells were stained with p-MAP4 antibody at 1:100, and uorescence conjugated secondary antibodies: Alexa Fluor 488-conjugated goat anti-rabbit IgG antibody (Invitrogen) at 1:100 and Alexa Fluor 561-conjugated goat anti-mouse IgG antibody at 1:100. To detect mitochondrial membrane potential, the living cells are stained with TMRE (200nmol/L) for 20min. The cells were viewed and analyzed through a TCS SP5 laser confocal microscope (CANT; Leica, Wetzlar, Germany) or Leica LAS AF 2.3.0 software (Leica Microsystems, Germany), respectively.
RFP-LC3, GFP-LC3, mRFP-GFP-LC3 and Dsred-Mito adenoviruses The mRFP-GFP-LC3, mRFP-LC3, GFP-LC3 and Dsred-Mito adenovirus were purchased from Hanbio Biotechnology (Shanghai, China) to respectively detect autophagy ux and mitophagosome. The cardiomyocytes grown on glass coverslips in 24-well plates were transferred with these adenoviruses overnight before corresponding treatment, and then the slides were visualized using a confocal microscope (CANT; Leica, Wetzlar, Germany). precipitates, after ve washing steps with PBS at 0 °C, were analysed via label-free shotgun proteomics. In brief, sample preparation, protein digestion, in-gel digestion, liquid chromatography-electrospray ionisation tandem MS analysis (Q Exactive) and sequence database searching and data analysis were performed. The MS data was searched against the Uniport database. Intensity-based absolute quanti cation (iBAQ) with MaxQuant was performed on the identi ed peptides to quantify the protein abundance.

Statistical analysis
All data are presented was presented using the mean ± SEM. The statistical analysis was performed using SPSS, v. 22.0. The independent-sample t-test and the one-way analysis of variance were applied for comparisons between two groups or more than two groups, respectively. Pearson's correlation coe cient was calculated using Image J (Fiji) software. P=<0.05 (two tailed) was used as the threshold to de ne statistical signi cance.     The MAP4 LIR (47-50aa) determinates p-MAP4 increased autophagy under hypoxic conditions. (A)