ATG5-HSP90.2-mediated micromitophagy as cytological basis for maternal inheritance of plant mitochondria

ATG5-HSP90.2-mediated micromitophagy as cytological basis for maternal inheritance of plant mitochondria | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article ATG5-HSP90.2-mediated micromitophagy as cytological basis for maternal inheritance of plant mitochondria Meng-Xiang Sun, Xiaorong Huang, Linlin Zhao, Zonglin Liu, Ni Long, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6690586/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 21 Jan, 2026 Read the published version in Nature Plants → Version 1 posted You are reading this latest preprint version Abstract Mitochondria are inherited maternally in most plants as a classical paradigm of non-Mendelian inheritance, but the mechanism underlying paternal mitochondrial elimination (PME) remains almost unknown. We report here that angiosperms have evolved a micromitophagy-mediated PME, in which vacuoles directly engulf paternal mitochondria via tonoplast invagination. We show that the micromitophagy occurs specifically in male germline (MG) cells. To gain mechanistic insights, we used a vegetative to germline cell fate transition system to establish that micromitophagy is triggered by MG-cell-fate determination. We evidence that ATG5 is translocated to vacuole upon MG-cell-fate determination and interacts with mitochondrion-located HSP90.2 during mitochondrial engulfment by vacuoles, elucidating a cell-type specific ATG neofunctionalization to mediate micromitophagy. The ingenious mechanism not only contributes to maternal inheritance of plant mitochondria, but also supports zygote-to-embryo transition. We further determine that micromitophagy is conserved in angiosperms but was continually optimized during evolution to ensure a highly efficient PME strategy in MG cells with different properties. These findings bridge a long-standing gap in understanding plant PME with emerging mechanistic knowledge. Biological sciences/Plant sciences/Plant reproduction Biological sciences/Plant sciences/Plant development Biological sciences/Plant sciences/Plant cell biology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Mitochondria and plastids of the plant embryo are provided by the egg cell but not by the sperm cell (SC) via fertilization, which has been known as cytoplasm maternal inheritance, a classic example of non-Mendelian inheritance 1 , 2 . The first event of non-Mendelian inheritance was discovered in 1909, Carl Correns reported the plastid maternal inheritance of leaf color in Mirabilis jalapa 3 , 4 . Subsequent research showed that the mitochondria of major angiosperms are strictly maternally inherited 5 . However, after many years of the discovery of this phenomenon, what mechanisms regulate PME in plants remain obscure. Although no clear mechanism has emerged, various hypotheses have been advanced for what might underlie the maternal inheritance of mitochondria in plants. The finding that male gametic mtDNA was reduced in angiosperms led to the suggestion that it might contribute to maternal mitochondrial inheritance 6 . However, the mitochondrial endonuclease responsible for this process is still unknown. Potential DNA nucleases, the pollen-expressed Mg 2+ -Dependent exonuclease DEFECTIVE IN POLLEN ORGANELLE DNA DEGRADATION1 (DPD1) and Mitochondrial Endonuclease M20, have been reported to degrade mtDNA in VC but not in SCs 7 – 9 . These nucleases are therefore unlikely to be relevant to PME. In addition, in the dpd1 mutant, the paternal mtDNA carried by the SC was never transmitted to offspring 7 , 10 , and thus, there must be some unknown pathways specifically in MG cells for ensuring mitochondrial maternal inheritance in angiosperms. It was also proposed that paternal mitochondria might remain outside of the egg cell during gamete fusion 11 or could be excluded physically from the GC during microspore asymmetric division 12 . However, no follow-up works have been reported to confirm these proposals. Recently, a breakthrough has been made in the mechanism of plastid maternal inheritance in tobacco. During male gametogenesis, an environmental temperature-regulated organelle exclusion mechanism and a DPD1-mediated genome degradation mechanism act in concert to prevent paternal transmission of plastid genes 13 . During sperm development in Caenorhabditis elegans , male mitochondria are exported out of the spermatids through mitopherogenesis 14 . In fact, plant scientists have also been curious about whether PME occurs in the MG cells to ensure maternal mitochondrion heredity for decades. Autophagy is an evolutionarily conserved mechanism and provides cells with a pathway for the continuous turnover of undesired organelles. It is usually divided into three main types: macroautophagy, microautophagy and chaperone-mediated autophagy 15 . We found that the vacuoles needed for plant autophagy existed in the MG cells 16 and the number of paternal mitochondria continue to decrease during SC development, providing an intriguing clue that mitochondrial depletion by autophagy might be the means for PME in plants. Unfortunately, none of the three conventional autophagy mechanisms have a confirmed involvement in PME in angiosperms. Here, we report that micromitophagy exists in plants and is triggered by MG cell fate determination. We have uncovered a novel role for the core autophagy gene ATG5 to engage micromitophagy specifically for PME in the MG cells. We discovered that ATG5, which typically is located in the isolation membrane of developing autophagosome, is transferred to vacuoles where it binds to HSP90.2 of paternal mitochondria to mediate vacuole-mitochondria interaction and micromitophagy. We also demonstrate that superfluous paternal mitochondria entering the egg cell via fertilization will affect zygote development and subsequent embryogenesis, underscoring the biological importance of PME and revealing the deleterious effect if PME is hampered. Finally, our results reveal that this cell-type specific micromitophagy is a conserved mechanism contributing to mitochondrial maternal inheritance in angiosperms. Results Persistent mitochondrial depletion during MG cell development We first carefully observed the mitochondrial behavior in the MG cells. After the asymmetric division in the microspore of A. thaliana that produces the GC and VC of pollen, the mitochondria inherited from microspore divided into smaller mitochondria in MG cells, instead of fusion to form larger mitochondria (Extended Data Fig. 1 a-i). Similar mitochondrial fission events were also observed during the development of O. sativa (rice) MG cells, suggesting that this is a conserved phenomenon in both dicotyledons and monocotyledons (Extended Data Fig. 1 j-n). However, although the mitochondrial fission looked normal, the mitochondrial number did not increase in MG cells during male gametogenesis. Instead, the number of mitochondria decreased from GC (average 19.7 ± 4.0) to SCs (average 14.5 ± 4.5 per SC pair in mature pollen grain), the product of another mitotic division of the GC, in A. thaliana (Fig. 1 a,c). Importantly, by the in vivo–in vitro pollen growth assay 16 , we also found that the mitochondria in the SCs continued to reduce during 8 hours after pollination (HAP), the time needed for them to move into pollen tube (average 11.7 ± 2.6 per SC pair in pollen tube) and before they are released to the embryo sac and fusing with the egg cell. This clearly suggests that the reduction of mitochondria in the SCs is not due to simple dilution by GC division but depends on an active elimination process (Fig. 1 a,c). The function of the pollen is to deliver the SCs to fertilize female gametes located at some distance from where it lands on a receptive pistil. The decline of mitochondria observed here therefore indicates that the PME process was activated before fertilization and still spatially distant from male-female gamete fusion. ATG5 governs PME In most eukaryotic organisms, autophagy (often refers to macroautophagy) is a major intracellular degradation system that removes damaged and/or useless organelles and unwanted cytoplasmic components and is regulated by conserved autophagy-related ( ATG ) genes 17 – 19 . To understand if an autophagy-dependent pathway is the mechanism underlying the PME process in plants, we analyzed the transcriptomes of wild-type (WT) VC and WT SC 16 , and found that the transcripts of ATG5 , one of the core ATG genes, were more abundant and preferentially expressed in WT SC, implying that ATG5 might play a potential key role in PME in A. thaliana (Extended Data Fig. 3 a). Detailed characterization of atg5 mutant 20 revealed that at the early bicellular stage, the mitochondrial number of GC was not significantly different between WT and atg5 (Fig. 1 a-c). These observations imply that ATG5 dysfunction did not affect the mitochondrial dynamics in the microspore. However, in contrast to the significant reduction of mitochondria in WT SCs during male gametogenesis, the mitochondrial numbers increased dramatically and continuously in atg5 SCs (Fig. 1 a-c), which indicates that ATG5 regulates PME. Taken together, these results reflect that PME was continuing in SCs prior to fertilization under normal conditions, and this process during active mitochondrial fission was blocked by ATG5 dysfunction, leading to mitochondrial accumulation in atg5 SCs. 3-Methyladenine (3-MA), a phosphatidylinositol 3-kinase (PtdIns3K) inhibitor, can disrupt autophagy by blocking autophagosome formation. To further examine the role of autophagy in PME, we treated the pollen grains of the transgenic lines, in which the male mitochondria were labeled by GFP with different concentrations of 3-MA, and then counted the mitochondrial number in the SCs within the pollen tube after 5 hours. Surprisingly, 3-MA treatment could not induce the increase of mitochondrial number in the SCs and did not show a defective phenotype similar to atg5 (Extended Data Fig. 2 ). This result suggests that the ATG5-regulated PME may be involved in a non-autophagosome pathway. PME by a MG-specific micromitophagy pathway During macroautophagy, the double membrane-bound autophagosome sequesters organelles and then fuses with vacuoles (in yeast and plant) or lysosomes (in mammal) to release the internalized vesicle as autophagic body for the subsequent degradation. Therefore, this process involves complex membrane biogenesis 15 . Whereas, during microautophagy, it was suggested that the organelles were directly engulfed by the vacuoles without autophagosome mediation 21 . Previous studies have suggested that ATG5 may play an as-yet-undefined function in microautophagy in yeast 22 . As transmission electron microscopy (TEM) is the classical approach essential to describe microautophagy that lacks other effective methods 23 , we used TEM to analyze the male gametogenesis process. Firstly, we carefully followed the behavior of the vacuole (Extended Data Fig. 3 ), an indispensable organelle for macroautophagy and microautophagy in plant 15 , 21 . In the microspores, a large central vacuole formed from the fusion of pre-existing small vacuoles predominated. After the microspore asymmetric division, the large vacuole was re-divided into small vacuoles, which were assigned to both GC and VC. Then, while the vacuoles in the VC gradually disappeared, those in the GC were retained. Eventually, at the mature pollen stage, many spherical vacuoles existed uniquely in the SCs, while vacuoles were no longer clearly visible in the VC (Extended Data Fig. 3 ). Since the vacuole is indispensable for autophagy, these results imply that the autophagy responsible for plant PME could only occur in the MG cells. Meanwhile, our live-cell imaging showed that the mitochondria could indeed enter the vacuole during male gametogenesis, although it remains very difficult to catch the transient moments of mitochondrial contact and entrance (Fig. 1 d; Supplementary Video1). Furthermore, the exhaustive TEM analysis of a huge number of samples revealed clearly that the mitochondrial degradation in the MG cells of A. thaliana was actually via microautophagy rather than macroautophagy, and documented in detail the microautophagic process stepwise (Fig. 1 e-p). In MG cells, the vacuoles and mitochondria came into contact, the vacuole membrane invaginated (Fig. 1 e,f) and engulfed the mitochondria (Fig. 1 g,h). The trapped mitochondria entered micromitophagic compartments, and they could be observed in progressively degraded states in the highly hydrolytic environment of the compartments (Fig. 1 i-p). We also found that consistent with the aforementioned results (Fig. 1 a-c), as development of MG cells, the paternal mitochondria were continually reduced, but a few mitochondria always remained in SCs (Fig. 1 g,o), probably because necessary energy supply is still required for the later maturation of SCs and their function during fertilization, after which the remaining mitochondria would finally be removed. During plant male gametogenesis, the cytoplasmic contents, mainly the number of organelles, in the SCs is markedly reduced 16 . Thus, it would seem more practical and efficient by using the existing vacuoles to directly eliminate paternal mitochondria than to synthesizing new membrane structures and assembling autophagosomes for macromitophagy. In fact, we observed at least 1,000 A. thaliana pollen grains by TEM, and found that autophagosomes, characterized by their double-membrane structure and depredating cytoplasmic contents, occurred exclusively in the VC (Extended Data Fig. 4 b), but never in the SCs. These observations support a model that macroautophagy does not usually occur in the SCs, and PME is executed by MG-specific microautophagy rather than macroautophagy. ATG5 interacts with HSP90.2 to regulate PME So far, the molecular mechanisms regulating microautophagy remains poorly understood and there is no technique available yet to purposefully disrupt the microautophagy of any targeted organelles 21 , 24 . It was reported that ATG5 dysfunction affected micromitophagy in yeast, not in mammalian cells 22 , 24 , but the mechanism in yeast is not known. The above-mentioned results (Fig. 1 ; Extended Data Fig. 2 ) support that ATG5 might participates in micromitophagy in plant MG cells, offering a unique opportunity for exploring this novel function of ATG5 in multicellular organisms. Normally, ATG5 typically localizes in the isolation membrane, which is extending to envelop the autophagosome during its formation and involved in macroautophagy 15 , 25 . However, this only known location of ATG5 cannot explain its participation in microautophagy. We proposed that the neofunctionalization of ATGs should be accompanied by their translocation from the isolation membrane to the vacuole membrane. Therefore, we made the high-resolution subcellular localization analysis of GFP-tagged ATG5 and ATG5 immuno-labeling. We found that upon MG cell fate determination ATG5 can indeed translocate onto the vacuoles in the MG cells (Fig. 2 a,b), and this enables ATG5 to play a non-canonical role in micromitophagy. To further reveal the molecular mechanism of ATG5-mediated micromitophagy, we obtained an ATG5 interacting protein, HEAT SHOCK PROTEIN 90.2 (HSP90.2), by selection of candidate genes from SC transcriptome data together with yeast two-hybridization (Y2H) verification (Fig. 2 c). Moreover, Bimolecular fluorescence complementation (BiFC) assay and Co-immunoprecipitation (Co-IP) analysis further supported the interaction between ATG5 and HSP90.2 (Fig. 2 d,e). In animals, a chaperone network comprising Hsp90 and its related molecules specifically target mitochondria in tumor cells, but not in most normal tissues, and thus antagonizing mitochondrial permeability transition and regulating mitochondrial homeostasis 26 . This previous work in animal provides an important clue, namely, Hsp90 can be localized specifically to mitochondria in some specific cells to regulate mitochondrial fate. Thus, we firstly analyzed the HSP90.2 expression pattern during pollen development and male gametogenesis. In microspores, HSP90.2-RFP driven by native HSP90.2 promoter presented no clear fluorescence signals (Fig. 2 f). However, right after microspore asymmetric division, the HSP90.2-RFP exhibited MG-specific signals in GC and SCs (Fig. 2 f). Subsequently, we introduced MG-specific mitochondrial marker (Mito-GFP) into the cells with HSP90.2-RFP fusion protein, and found that HSP90.2 was indeed able to locate onto SC’s mitochondria (Fig. 2 g). Therefore, a logical explanation is that the vacuole-localized ATG5 interacts with the mitochondria-localized HSP90.2 to mediate paternal micromitophagy in MG cells. We sought genetic evidence to further confirm the conclusion from molecular interaction and localization results discussed above. We analyzed the hsp90.2-2 27 phenotype. Indeed, like what we observed in atg5 , the mutation of HSP90.2 led to a significant increase of mitochondrial number in SC pair (Fig. 2 h,i). This suggests that either ATG5 or HSP90.2 deletion can disrupt the interaction between vacuoles and mitochondria in MG cells, thereby blocking micromitophagy and PME, and leading to abnormal mitochondrial accumulation. This evidence supports that ATG5-HSP90.2 interaction is required for paternal micromitophagy. MG cell fate determination triggers micromitophagy Since distinct factors or developmental events can trigger different types of autophagy, clarifying the induction condition is a crucial part of studying an autophagic process and revealing its underlying regulatory mechanism. Since the micromitophagy reported here can only be detected in the MG cells but not in the VC, we contended that this process might be triggered by MG cell fate determination and evolved specifically for PME. H3K27me3 reprogramming is vital for the plant MG cell fate determination 16 , 28 . Previously, we have successfully induced VC fate switch towards SC by the VC-targeted H3K27me3 erasure, resulting in a sperm-like VC cell in pollen 16 . These results aligned with increased expression level of ATG5 and its relocalization on the vacuoles in the sperm-like cell (Extended Data Fig. 4 a,d). The finding not only implies an underlying involvement of ATG5 in MG cell fate determination-triggered micromitophagy but also provides a unique material with which the detailed micromitophagy events that accompany the SC differentiation process can be monitored starting from the switch to the MG cell fate. Therefore, taking advantage of this unique material, we monitored the autophagy events in both WT VC and the sperm-like VC. Astoundingly, the shift of VC fate into sperm-like not only successfully activated micromitophagy but also the process was brought to completion as in SCs (Fig. 3 , Extended Data Fig. 4 c and Supplementary Videos2-5) and could be conspicuously observed in these sperm-like cells (Fig. 3 a,b). The micromitophagic process has thus far not been well characterized cytologically. With the much larger volume of its cytoplasm relative to the GC and SC, this sperm-like cell model provides an exceptional opportunity to monitor the process, allowing us an unprecedented view of a detailed step-by-step process of microautophagy. When a mitochondrion approached a vacuole, it was first captured by the vacuole membraneinvagination, then totally enveloped and finally released into the vacuole (Fig. 3 c-i). Incipiently, a mitochondrion was boxed in a complete monolayer vacuous, and then this micromitophagic compartment flowed within the vacuole as its morphology became a “comet-like” form with a trailing tail (Fig. 3 j,k). Later, the membrane of this micromitophagic compartment gradually disintegrated, resulting in the mitochondrion being exposed to the vacuolar matrix (Fig. 3 l). Then, the mitochondrion would soon start to degrade (Fig. 3 m) and finally become totally digested (Fig. 3 b). Furthermore, the electron tomography (ET) analysis exhibited the micromitophagic process in sperm-like cell in a three-dimensional manner (Fig. 3 n, Extended Data Fig. 4 c and Supplementary Videos2-5), which more accurately exhibits the spatial position of the vacuole membrane, micromitophagic compartment and mitochondria. These 3D reconstructions further confirm that mitochondria shown by ultrastructural sections are indeed inside the vacuoles and micromitophagy plays a critical role in the process of PME. Clearly, it is the MG cell fate determination that triggers the microautophagy-dependent degradation of paternal mitochondria. This interesting phenomenon also explained why micromitophagy was not detected in the VC. During this cell fate transition triggered-microautophagy, three obvious features are particularly noteworthy. Firstly, a lot of vacuoles were generated simultaneously and they eliminated mitochondria synchronously (Fig. 3 a). Secondly, a vacuole was able to form multiple membrane invaginations at different sites to devour mitochondria at the same time (Fig. 3 g). Thirdly, a vacuole membrane invagination was routinely capable of swallowing not only one but at least two mitochondria at the same time (Fig. 3 i; also see Fig. 1 n). These features suggest remarkable efficiency can be achieved in the microautophagy-dependent PME in plant MG cells. Micromitophagy-mediated PME is a conserved strategy in angiosperms It is well known that the two main branches of angiosperms, dicots and monocots, both include three-celled pollen and two-celled pollen species. In three-celled pollen type the GC divides in pollen grain and each mature pollen contains a VC and two SCs. Whereas, in two-celled pollen type mature pollen contains two cells, a VC and a GC, and the GC divides in pollen tubes after pollen germination 29 . Use of microautophagy to continuously eliminate paternal mitochondria in the immobile SC of A. thaliana clearly has diverged from the macroautophagy process employed by the mobile SC of non-seed plants Physcomitrella patens (moss) and Marchantia Polymorpha (liverwort) for cytoplasmic reduction and mitochondrial reorganization 30 , 31 . It is thus interesting to know whether the micromitophagy-mediated PME had become a universal strategy in MG cells of all type of angiosperms during plant evolution or that the processes have diverged between different branches of angiosperm. With A. thaliana being the model dicotyledonous three-cell species, we selected three other model plants, Nicotiana tabacum (tobacco), O. sativa (rice) and Lily longiflorum (lily) as the representatives of dicotyledonous two-celled, monocotyledonous three-celled and monocotyledonous two-celled pollen, respectively, and observed mitochondrial dynamics in their MG cells. Indeed, a micromitophagy process as in A. thaliana was observed in all three plants (Fig. 4 a-l and Extended Data Fig. 5 ), establishing that micromitophagy is indeed a conserved approach for PME for angiosperms. Further examination of O. sativa showed that similar to A. thaliana , micromitophagy only occurred in the small, dispersed vacuoles of the MG cells (Extended Data Fig. 5 a-i), providing key evidence to further support that the vacuoles in the MG cells indeed had differentiated for the specialized PME process. However, there is also clear evidence for adaptive evolution of the micromitophagy process in PME during speciation. We observed, interestingly, that except for the vacuole membrane invagination-mediated paternal mitochondria engulfment (Fig. 4 i,j and Extended Data Fig. 5 j,k), the L. longiflorum GC, which is larger and contains more mitochondria than its counterparts in N. tabacum , O. sativa and A. thaliana , tended to have evolved an even more straightforward and efficient strategy to eliminate mitochondria, namely, the direct fusion between the paternal mitochondria and the vacuoles (Fig. 4 k,l and Extended Data Fig. 5 l). TEM revealed vacuole membrane extending a protrusion to capture mitochondria (Fig. 4 k, upper right), and then fusing with the mitochondrial membrane to produce a vacuole- mitochondrion contact (Fig. 4 k, lower left, and Extended Data Fig. 5 l). The formation of this contact structure led to the fusion of the membranes and mixing of the two organelles with each other, resulting in the mitochondrial matrix internalized into the vacuole (Fig. 4 l) and was digested. Moreover, several projections could occur simultaneously from a single vacuole and directed towards mitochondria. Thus, a vacuole was able to synchronously capture multiple mitochondria and quickly degrade them (Fig. 4 k,l). The L. longiflorum GC, probably due to its large size, more cytoplasm contents, and paternal mitochondria, has therefore evolved an even more efficient micromitophagy process than that observed in A. thaliana and O. sativa to achieve timely removal of mitochondria from the cell. These observations indicate that although the micromitophagy-dependent PME is a conserved strategy in angiosperms, the modes of micromitophagy may not be invariant in the diversified groups of angiosperms. The optimal mode of microautophagy for efficient PME to meet species-dependent requirement may have been evolved or selected in different MG cells. Mechanistically, we further investigated whether ATG5, which regulates PME in A. thaliana that belongs to tricellular pollen species, also governs PME in plants of bicellular pollen species. The GC undergoes mitosis in the growing pollen tube to form SCs in N. tabacum . Thus, to gain clearer images, we isolated the late GC from the pollen grains and the SC pair from the pollen tubes of WT and Ntatg5 20 N. tabacum respectively and then mitochondrial dynamics were monitored by the Mitotracker Green staining (Fig. 4 n). Our result showed that, like A. thaliana , active paternal mitochondrial reduction also occurred during N. tabacum spermatogenesis, and the NtATG5 absence disturbed PME, which led to the abnormal accumulation of a large number of mitochondria in the MG cells (Fig. 4 n,o), suggesting that the ATG5- regulated PME is also a conserved mechanism in different groups of angiosperms. Excess paternal mitochondria entering egg cell affects zygote development While maternal organelle inheritance is the norm, the benefit from the exclusion of paternal organelles is unclear. To test whether the excess paternal mitochondria may influence male-female gametes fusion or postfertilization development if PME is blocked before fertilization, we crossed the atg5 as paternal parent with egg cell pDD45:GFP marker line as maternal parent. In the hybrid ovules at 20 HAP, all fertilized egg cells developed into zygotes, indicating that atg5 SC was able to fuse with egg cell for fertilization and showed no difference from WT SCs. However, the zygotic development in atg5 pollen-pollinated ovaries was significantly delayed compared to that in WT pollen-pollinated ovaries (Fig. 5 a), demonstrating the importance of the deletion of the paternal mitochondria from the zygote for post fertilization development. We next monitored the paternal mitochondrial dynamic after fertilization in A. thaliana zygotes at different stages to establish the importance of timely PME to zygote development. Firstly, we crossed the vacuole (RFP) and mitochondria (GFP) double markers (two-markers line) as the paternal parent into a WT maternal parent. With only a few mitochondria remaining in the SCs (Fig. 1 a,c), these residual paternal mitochondria entered the egg cell together with the SC-delivered vacuoles via fertilization (Extended Data Fig. 6). Interestingly, the paternal mitochondria were found also co-localized with the SC-delivered vacuole in the fertilized egg cell (Extended Data Fig. 6). The continued co-compartmentation of the engulfed paternal mitochondria within the SC-delivered vacuole suggests that the SC-initiated microautophagy clearance system of paternal mitochondria may continue to function after fertilization. To test this, we crossed the two-markers line as paternal parent and monitored the fate of the paternal mitochondria in developing zygotes (Fig. 5 b). About 6 (6.1 ± 1.8, n = 18) paternal mitochondria could be detected in the zygotes at 8 HAP, but by 14 HAP almost all of paternal mitochondria were already eliminated, and at 20 HAP paternal mitochondria could no longer be observed (Fig. 5 b). Therefore, mitochondrial clearance continues to function after fertilization. We then introduced the double markers line into atg5 [referred to as atg5 (two-markers line)]. When atg5 (two-markers line) was used as the paternal parent, more than 19 (19.5 ± 3.2, n = 15) paternal mitochondria could be detected in zygotes at 8 HAP, and the paternal mitochondria persistent even in the 20HAP zygotes along with the delayed zygote elongation (Fig. 5 c). This result suggests that though the seed fertility was not significantly disturbed, the excess paternal mitochondria delivered from atg5 SC to the normal egg cell have postponed the process of zygote-to-embryo transition in A. thaliana . To further confirm the conclusion, we used WT or Ntatg5 pollen to pollinate WT N. tabacum ovaries, with saturating amount of pollen grains. By isolating the embryo sacs from 108 HAP hybrid ovaries to analyze the embryogenesis process, we found that the hybrid embryos from the ovules pollinated with WT pollen had developed overwhelmingly to the developmental stages of 2-celled (64.3 ± 1.7%) or 4-cell (35.4 ± 1.8%) proembryos. On the other hand, in the ovules pollinated with Ntatg5 pollen, the proportion of zygote increased significantly (15.2 ± 0.5%), while 4-celled proembryos decreased notably (7.7 ± 1.4%) (Fig. 5 d). Next, we observed the ovules from 144 HAP hybrid ovaries using an ovule whole-mount clearing technique. In contrast to all zygotes having developed into embryos in the hybrid ovules pollinated with WT pollen, at least ~ 3% of zygotes stopped developing in the hybrid ovules pollinated with Ntatg5 pollen (Fig. 5 e), and then exhibited as seed abortion at 12 days after pollination (Fig. 5 f). Compared to WT SC, loss of ATG5 resulted in about 3-fold increase in the number of paternal mitochondria in A. thaliana SC and about 5-fold increase in N. tabacum SC (Fig. 1 c, Fig. 4 o). The hybrid zygotes in N. tabacum displayed a more severe phenotype than that in A. thaliana , probably due to the introduction of more additional paternal mitochondria into the fertilized egg cells, which was more than the zygotes can tolerate. These results not only elucidate a direct role of paternal mitochondria in hindering zygotic development, but also indicate that the negative effect of abnormal PME in MG cells on zygote-to-embryo transition is conserved in different groups of angiosperms. Discussion Compared to macroautophagy, microautophagy is much less investigated, especially in plants 21 , 32 where to-date just a few reports have dealt with microautophagic activities during its development. The pigment aggregates in A. thaliana cotyledons and E. grandiorum purple petals were directly engulfed by a single-layer vacuole membrane and became an autophagic body-like structure in vacuoles 33 . Similarly, the aleurone cells of maize endosperm delivered the prolamins to vacuoles via a microautophagic manner 34 . In addition, the degradation of whole photo-damaged swollen chloroplasts in A. thaliana was shown to be microautophagy, termed as “microchlorophagy” 35 , 36 . However, the spatial and temporal events during plant microautophagic processes are far from precisely described, prohibiting advances in understanding the molecular mechanism that regulates microautophagy in plants. Herein, we evidenced for the first time that micromitophagy exists in angiosperms with detailed documentation and emerging mechanistic insights. This micromitophagy specifically occurred in MG cells but not in VC during pollen development. By triggering cell fate transition of VC towards SC, we experimentally demonstrated that the micromitophagy is triggered by MG cell fate determination. Mechanistic understanding of micromitophagy is in fact still insufficient in animal and remains very poor in yeast after several decades of studies, while entirely lacking in plants. In mammalian cells, micromitophagy, defined as lysosomal degradation of mitochondrial-derived vesicles, is independent of autophagosomes and key autophagy regulators, such as ATG5 24,37 . Whereas, the micromitophagy induced by nitrogen starvation in the lactate-grown S. cerevisiae and the microchlorophagy induced by UV-B damage in A. thaliana suggested that the microauthophagy process of semi-autonomous organelles might involve core ATG proteins 24 , 35 , 36 . As an example, the photo-damaged chloroplasts accumulated in A. thaliana atg5 and atg7 mutants due to aberrant microchlorophagy 35 , 36 . Herein, we found that ATG5 was expressed preferentially in SC and was distributed on the vacuoles of MG cells (Fig. 2 a,b and Extended Data Fig. 4 a). Meanwhile, HSP90.2 is rapidly transferred to the paternal mitochondria in GC right after MG cell fate determination (Fig. 2 f,g). The ATG5-HSP90.2 interaction mediates the vacuole-mitochondrion recognition and thus promotes PME in the MG cells. In A. thaliana , HSP90s are widely involved in different processes, e.g., for regulatory protein translocations in developmental and immunity signaling pathways 37 , 38 . The voluminous work in HSP90.2 will provide useful clues for future mechanistic investigation of whether HSP90.2 have evolved a unique function in micromitophagy during plant spermatogenesis. At present, it is unclear how the MG cell fate determination drives the assembly of ATG5 onto the vacuoles and HSP90.2 onto the mitochondria to regulate micromitophagy. It is important to note that even in yeast and mammal cells that have a considerably longer history of investigations, the recognition mechanism of micromitophagy is still unknown. The discovery of ATG5-HSP90.2 interaction bridges this knowledge gap for understanding the regulating mechanism of micromitophagy, providing a novel approach for research in future. The microspore undergoes just one asymmetric cell division to generate a VC and a GC (Extended Data Fig. 1 a), so achieving paternal micromitophagy must be a rapid and unique developmental event in MG cells. The core ATG proteins, including ATG5, are wildly known as key regulators of macroautophagy and was critical for VC development 20 , 39 . Although ATGs are universally expressed in all plant organs, we have previously discovered that macroautophagy can act compartmentally at a specific position in the VC for a specific developmental event, namely at the pollen germinal aperture for pollen tube extrusion, thus to support pollen germination 20 . Together with establishing microautophagy for PME during spermatogenesis here, our works evidence that macroautophagy and microautophagy have co-evolved in two different cell types of pollen, the VC and MG cells, to carry out the distinct functions of pollen germination and PME, respectively. Together, our studies in pollen germination 20 and MG cells here established that ATG5 plays distinct roles in two sister cell types in different developmental events. They also suggest that ATG5 regulating the broadly utilized macroautophagy in the VC to permit pollen germination 20 has undergone neo-functionalization during evolution to function in the MG cell-specific microautophagy-mediated PME (Fig. 2 ). In illustrating the occurrence of ATG5 neofunctionalization in a specific cell type for a specific developmental issue, like PME, our findings establish cell type specific neofunctionalization of ATG proteins, previously unknown not only in plants but also in animals or yeast. The discovery of ATG5 neofunctionalization here also reveals an adaptive evolution of the core ATG genes or different autophagy pathways for specific developmental events. In fact, neofunctionalization of existing proteins for new functions should be considerably more efficient and resource-economical relative to establishing a totally new group of proteins within a very narrow developmental window, such as after the GC formation to SC maturation. Given the importance of autophagy in developmental processes, the discovery of cell type-specific and functional diversification of core ATGs opens an unexplored territory of autophagy investigation, which obviously could only be addressed in multicellular organisms like angiosperms as we have pioneered here. These foundational discoveries should invite in-depth investigations of the mechanism of the cell type-dependent ATG translocation and functions in multicellular organisms, as well as the identification of triggers for ATGs’ functional transition in different cell types. The fact that angiosperms have evolved micromitophagy suggests it must be a more fitting strategy to contribute to maternal inheritance while ensuring successful fertilization. The SCs of flowering plants have lost their motility and are passively delivered to the embryo sac for fertilization by a pollen tube produced by the VC. Accordingly, large numbers of mitochondria are no longer required to power autonomous movement when the immotile angiosperm SCs have fully developed. Obviously, eliminating the unnecessary paternal mitochondria before fertilization is the simplest and ideal option to reduce reliance on complex mechanisms of distinguishing paternal or maternal mitochondria in zygotes 40 . However, after SCs have been released from the pollen tube in embryo sac, they still need energy to maintain an active metabolism, not only to stay alive but also to trigger several crucial pre-fertilization events and during gamete fusion 41 – 43 . Therefore, consistent with this, the paternal mitochondria being greatly deleted but never eliminated completely in the SC before fertilization (Fig. 1 a-c) is clearly an adaptation to fuel the SCs in their last journey to fertilization. In fact, in the long evolution of angiosperms, maintaining minimum mitochondria in SCs and avoiding the damaging impact of excess paternal mitochondria on zygote and early embryo development (Fig. 5 a,d-f) might be optimum for balancing the needs to contribute mitochondrial maternal inheritance and to fuel gamete communication, fusion and the zygote-to-embryo transition. Online content Any methods, additional references, Nature Portfolio reporting summaries, extended data, supplementary information, acknowledgements; details of author contributions and competing interests are available at______. Methods Plant material and growth conditions All A. thaliana lines including the atg5-1 20 and hsp90.2-2 27 mutants were in Columbia background, and were grown in a greenhouse under long-day conditions (16h light/8h dark) at 22°C. O. sativa plants used in this study were of the Zhonghua 11 (ZH11) cultivar of the japonica group, and were grown in a glasshouse under the diurnal cycle of photoperiod and temperature (30°C : 22°C, 12 h : 12 h, light : dark). N. tabacum L. cv. Petite Havana SR1 and Ntatg5 20 plants were grown in a glasshouse under 16h light/8h dark at 25°C. L. longiflorum plants were grown in the field of Wuhan University, in Wuhan, China. Constructs and plant transformation The 2435bp fragment including TIP5;1 promoter and coding sequence was amplified by PCR to introduce SacI and KpnI flanking restriction sites. The SacI-KpnI-digested fragment was then cloned into the SacI-KpnI-digested P182 vector (modified from the PART27 vector, carrying a RFP gene and a kanamycin-resistance gene). The DUO1 promoter was amplified as described previously 44 , and then cloned into the SacI-KpnI-digested Q108 vector (modified from the PART27 vector, carrying a GFP gene and a Hygromycin-resistance gene). Subsequently, a 129-bp DNA fragment containing the mitochondria-targeted pre-sequence of the located F1-ATPase gene At5g13450 was amplified and cloned to generate the Pro DUO1 :Mito-GFP reporter. The constructs were transferred into Agrobacterium tumefaciens strain GV3101, which was used to transform A. thaliana plants by the floral dip method 45 . For selection, the medium was supplemented with 50 μg/ml hygromycin or kanamycin. Mitochondrial dynamic in male germline cells For A. thaliana , the pollen grains from ProTIP5;1: TIP5;1-RFP ProDUO1:Mito-GFP line (two markers line) and atg5-1 ProTIP5;1: TIP5;1-RFP ProDUO1:Mito-GFP line at different developmental stage were dissected rapidly and then examined using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems). To visualize SC within pollen tube, the in vivo–in vitro pollen growth assay was performed according to the previous method 16, 46 . The styles pollinated by the mature pollen from the aforesaid two lines were incubated on pollen germination medium (PGM: 1mM CaCl 2 ·2H 2 O, 1mM Ca(NO 3 ) 2 ·4H 2 O, 1mM MgSO 4 ·7H 2 O, 1% H 3 BO 3 , 18% sucrose, pH7.0) at 22°C for 8 hours. The samples were placed in the same culture dish to guarantee the consistency of experimental conditions. For autophagy inhibition experiments, the pollen grains from two markers line germinated in the PGM containing 0mM, 1mM, 2.5mM and 5mM 3-MA (Sigma, M9281). With pollen germination, SC pair migrated into the pollen tube. After 5 hours of treatment, the mitochondrial dynamics of SCs treated with different concentrations of 3-MA were observed by Leica SP8 CLSM. The inverted confocal microscope system (Zeiss LSM 980) was used for live imaging of the micromitophagy process in SCs. The mature pollen grains from the two markers line were incubated on PGM. The microscope was mounted on a Zeiss Axio Observer Z1 basic stand equipped with an incubator (XLmulti S1) to maintain the temperature at 22 °C. Time-lapse images were acquired every 30 seconds for 5 to 10 mins, and image processing was performed with ZEN software. For N. tabacum , the late GC and SC were isolated from the growing pollen tubes 47 , and they were stained with 100nM Mitotracker Green (Invitrogen, M7512), and then the mitochondria were observed with a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems). For every experiment, unless otherwise indicated in the figure legends, the images are representative of at least three independent biological experiments. Transmission electron microscopy (TEM) and electron tomography The anthers for A. thaliana and the pollen grains for O. sativa , N. tabacum and L. longiflorum at different developmental stage were dissected carefully in high-pressure freezer specimen holder. Then, they were embedded with the external cryo-protectant (2% low melting point agarose, Sigma, A9539), and any air bubbles must be removed to avoid disrupting the tissues. Samples were rapidly freeze fixed using the Lene and stoica EM ICE and stored in liquid nitrogen. Fixative, freeze substitution and Spurr resin embedding of samples and Ultra-thin section preparation were executed from a procedure described previously 16 . The Ultra-thin sections (70 nm) were examined with a JEM-1400 plus transmission electron microscope (JEOL) at 100 kV. For electron tomography analysis, the sections (100 nm) were stained with 2% (w/v) uranyl acetate followed by 2.6% (w/v) lead citrate. Tilting series were collected from +60° to -60° (1° intervals) using a Talos L120C G2 electron microscope (Thermo Fisher). Tomograms were reconstructed and 3D models were generated using the Inspect3D and Amira software package. Immuno-electron microscopy The anther of A. thaliana was fixed with 4% (w/v) paraformaldehyde for 16 h at 4°C. After dehydration through an ethanol series (30%→50%→60%→70%→80%→90%→100% three times; 4 h per step), the samples were embedded in K4M resin. The Ultra-thin sections (70 nm) were prepared and then the blocking (PBS, 0.05% Triton X-100, 0.05% Tween-20, 10g/L BSA, RT, 5min) were performed. Following the pretreatment, the sections were incubated with primary antibody (anti-ATG5, 1:100, Agrisera, AS153060) for 2h at RT and 24h at 4°C. After washed six times in PBS for 2min each time, the sections were incubated with secondary antibody (Anti-Rabbit IgG (whole molecule)–Gold antibody, 10nm, 1:200, Sigma, G3779) for 1h at RT. After washed six times in PBS and twice in water for 2min each time, the sections were stained with 2% (w/v) uranyl acetate and 2.6% (w/v) lead citrate aqueous solution. The samples were examined with a JEM-1400 plus transmission electron microscope (JEOL) at 100 kV. For double immunolabelling, the sections were incubated with primary antibody (anti-ATG5, 1:100, Agrisera, AS153060) and secondary antibody (Anti-Rabbit IgG (whole molecule)–Gold antibody, 5nm, 1:200, Sigma, G7277) following the steps mentioned above. Then, the sections were fixed with 2% (w/v) paraformaldehyde and 0.01% (v/v) glutaraldehyde for 10 min. After washed two times in PBS for 2min each time, the sections were blocked again. Continuedly, the sections were incubated with primary antibody (anti-RFP, 1:100, Invitrogen, MA5-15257) and secondary antibody (Anti-mouse IgG (whole molecule)–Gold antibody, 10nm, 1:200, Sigma, G7652) following the steps mentioned above. Finally, the sections were stained with 2% (w/v) uranyl acetate and 2.6% (w/v) lead citrate aqueous solution, and then examined with a JEM-1400 plus transmission electron microscope (JEOL) at 100 kV. Yeast two-hybrid (Y2H) assay The Y2H assay was performed following the manufacturer’s instructions (The Yeast Two-Hybrid System Kit (Clontech), http://www.clontech.com/) and used to test the protein–protein interaction. The full-length coding sequences of the genes for ATG5 and HSP90.2 were cloned into pGBKT7 vector or pGADT7 vector, and fused with the GAL4 DNA binding domain (DBD) and the activating domain, respectively. The different plasmid combinations were transformed into the yeast strain AH109. The transformants were selected on SD/-Leu/-Trp culture medium, and then interactions were tested on SD/-Leu/-Trp/-His and SD/-Leu/-Trp/-His/-Ade culture medium. Bimolecular fluorescence complementation (BiFC) assay For BiFC experiment, the full-length coding sequences of the genes for ATG5 and HSP90.2 were inserted into the pUb-nYFP and pUb-cYFP vectors, respectively. Then, the different plasmids were transformed into Agrobacterium tumefaciens strain GV3101. We infiltrated fully expanded leaves of Nicotiana benthamiana with transformed A. tumefaciens strains containing constructs at the density OD 600 = 0.2. After 48 hours, the epidermal cells of infiltrated leaves were examined and images were collected using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems). Co-immunoprecipitation (Co-IP) assay For the Co-IP assay, the constructs Pro35S: ATG5-mCherry , Pro35S: HSP90.2-GFP and Pro35S: GFP were transformed into A. tumefaciens strain GV3101, and then two groups of mixed bacterial solution ( Pro35S: ATG5-mCherry/Pro35S: HSP90.2-GFP and Pro35S: ATG5-mCherry / Pro35S: GFP ) infected N. benthamiana leaves and caused the co-expression of ATG5-mCherry with HSP90.2-GFP or GFP, respectively . After infiltration for 48 hours, the fluorescence signal intensity of mCherry and GFP and the expression levels of fusion proteins in two groups of tobacco leaves were evaluated using Leica SP8 CLSM. Subsequently, the leaves were fully ground in liquid nitrogen and the proteins were extracted using the extraction buffer (50mM Tris–HCl, pH 7.5, 150mM NaCl, 1mM EDTA, 1% (v/v) Triton X-100, 1mM PMSF and 1× protease inhibitor cocktail). Immunoprecipitation was performed using the ChromoTek GFP-Trap® Magnetic Agarose for overnight with gentle rotation at 4 °C. The eluted protein samples together with Input were separated by SDS–PAGE (ACE Biotechnology, FuturePAGE TM 4-20%), and then detected with anti-GFP antibody (Abcam, ab13970) at 1/4,000 dilution or anti-mCherry antibody (Abcam, ab125096) at 1/2,000 dilution. The western blots were acquired using a Amersham Imager (GE, 680RGB). Zygote phenotypic analysis The healthy pistils of ProDD45:GFP marker line at same developmental state were hand-pollinated by the pollen grains from WT and atg5-1 , respectively. It's important to note that pollination should be done in 10 minutes or less. 20 hours after pollination, all pistils are picked synchronously and placed in a wet box. Immediately, the ovules in the anterior half of the pistils were dissected and the developmental stage of the zygote was observed using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems). Tobacco embryo sac isolation For tobacco embryo sac isolation to observe zygote and pro-embyro development, the 108 HAP hybrid seeds were firstly treated in the enzyme solution (1% cellulase, 0.8% macerozyme, and 0.058% MES in 11% mannitol, pH 5.8) for 30 min in dark. Then, the enzyme solution was removed, and seeds were washed twice with 11% mannitol solution (including 0.058% MES, pH 5.8). The suspension of the enzymolytic seeds was gently grinded by a flat-headed glass rod on slide. The releasing embryo sacs with the living zygote or pro-embryo were observed using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems) 48,49 . Ovule whole-mount clearing The ovule clearing was performed from a procedure described previously 49 . Briefly, the whole ovules located in the middle of the 144 HAP hybrid ovaries were collected in a 2ml-centrifuge tube with fixative (50% methanol and 10% acetic acid) and then fixed at 4°C for 12 hours. Subsequently, the fixed ovules was treated by 1% SDS and 0.2M NaOH at room temperature (RT) for overnight. Thereafter, the samples were incubated successively as follows: 2.5% NaClO for 1 hour at RT, 1% periodic acid for 1 hour at RT, 80% ethanol for 10min at 80°C, and the same fixative for 1 hour at RT, taking care to rinse the ovules with water in between each step. After washing, the samples were treated in 0.1mg/ml propidium iodide (PI) (P4170; Sigma) for 3 hours and then transferred in a chloral hydrate solution (including 4 g chloral hydrate, 1 ml glycerol, and 2 ml water) and kept overnight at RT. Finally, the ovules were stored in Hoyer’s solution (200 g chloral hydrate, 30 g gum arabic, 20 g glycerol, and 50 ml water) for about 7 days and further observation using a confocal microscope (Leica SP8 CLSM, Leica Microsystems). Statistical analysis Statistical analyses were performed using Microsoft Excel 2003 and GraphPad Prism software (version 8.0), and significance is determined by t-test, one-way ANOVA or two-way ANOVA, respectively (see figure legends for details). Declarations Data availability Sequence data from this article can be found in TAIR database (www.arabidopsis.org). High-throughput sequencing data that support this study are available in the NCBI Gene Expression Omnibus (GSE162640). All unique/stable reagents and plasmids with transgene constructs generated in this study are available from the corresponding authors on reasonable request with a completed materials transfer agreement. Acknowledgments We thank Prof. Shi Xiao (Sun Yat-sen University) for providing A. thaliana atg5-1 (SAIL_129_B07) mutant seed; Prof. Shuhua Yang (China Agricultural University) for providing A. thaliana hsp90.2-2 mutant seed; Prof. Gary N. Drews (University of Utah) for offering pDD45:GFP marker line; Dr. Danyang Li (the Core Facility, Wuhan University) for electron tomography analysis. This work was supported by the National Natural Science Foundation of China (31991201, 31800265 and 32130031). Author contributions X.H. designed and performed the research, analyzed data, and wrote the draft. L.Z., Z.L., W.Z., F.G., T.C., C.S., X.Z., W.W., and H.C. performed the research. A.C. analyzed the data and finalized the manuscript. M.-X.S. designed the research, analyzed data, and wrote and finalized the manuscript. Competing interests The authors declare no competing interests. References Birky, CW. Jr. The inheritance of genes in mitochondria and chloroplasts: laws, mechanisms, and models. Annu Rev Genet. 35, 125–48 (2001). Mittelsten Scheid, O. Mendelian and non-Mendelian genetics in model plants. Plant Cell 34(7), 2455–2461 (2022). Correns, C. Vererbungsversuche mit blass (gelb) grünen und buntblättrigen Sippen bei Mirabilis jalapa, Urtica pilulifera und Lunaria annua. Z Vererbungsl 1, 291–329 (1909). Birky, CW. Jr. Uniparental inheritance of organelle genes. Curr Biol. 18(16), R692-5 (2008). Nagata, N., Saito, C., Sakai, A., Kuroiwa, H. & Kuroiwa, T. The selective increase or decrease of organellar DNA in generative cells just after pollen mitosis one controls cytoplasmic inheritance. Planta 209(1), 53–65(1999). Wang, D.Y., Zhang, Q., Liu, Y., Lin, Z.F., Zhang, S.X., Sun, M-X. & Sodmergen. The levels of male gametic mitochondrial DNA are highly regulated in angiosperms with regard to mitochondrial inheritance. Plant Cell 22(7), 2402–16 (2010). Matsushima, R., Tang, L.Y., Zhang, L., Yamada, H., Twell, D. & Sakamoto, W. A conserved, Mg²+-dependent exonuclease degrades organelle DNA during Arabidopsis pollen development. Plant Cell 23(4), 1608–24 (2011). Ma, F., Qi, H., Hu, Y.F., Jiang, Q.R., Zhang, L.G., Xue, P., Yang, F.Q., Wang, R., Ju, Y., Uchida, H., Zhang, Q. & Sodmergen. The Mitochondrial Endonuclease M20 Participates in the Down-Regulation of Mitochondrial DNA in Pollen Cells. Plant Physiol. 178(4), 1537–1550 (2018). Takami, T., Ohnishi, N., Kurita, Y., Iwamura, S., Ohnishi, M., Kusaba, M., Mimura, T. & Sakamoto, W. Organelle DNA degradation contributes to the efficient use of phosphate in seed plants. Nat Plants 4(12), 1044–1055 (2018). Tang, L.Y., Matsushima, R. & Sakamoto, W. Mutations defective in ribonucleotide reductase activity interfere with pollen plastid DNA degradation mediated by DPD1 exonuclease. Plant J. 70(4), 637–49 (2012). Mogensen, H.L. Exclusion of male mitochondria and plastids during syngamy in barley as a basis for maternal inheritance. Proc Natl Acad Sci U S A. 85(8), 2594–7 (1988). Nagata, N. Mechanisms for independent cytoplasmic inheritance of mitochondria and plastids in angiosperms. J Plant Res. 123(2), 193–9 (2010). Chung, K.P., Gonzalez-Duran, E., Ruf, S., Endries, P. & Bock, R. Control of plastid inheritance by environmental and genetic factors. Nat Plants 9(1), 68–80 (2023). Liu, P., Shi, J., Sheng, D., Lu, W., Guo, J., Gao, L., Wang, X., Wu, S., Feng, Y., Dong, D., Huang, X. & Tang, H. Mitopherogenesis, a form of mitochondria- specific ectocytosis, regulates sperm mitochondrial quantity and fertility. Nat Cell Biol. 25(11), 1625–1636 (2023). Nakatogawa, H., Suzuki, K., Kamada, Y. & Ohsumi, Y. Dynamics and diversity in autophagy mechanisms: lessons from yeast. Nat Rev Mol Cell Biol. 10(7), 458–67 (2009). Huang, X. & Sun, M-X. H3K27 methylation regulates the fate of two cell lineages in male gametophytes. Plant Cell 34(8), 2989–3005 (2022). Mizushima, N. & Komatsu, M. Autophagy: renovation of cells and tissues. Cell 147(4), 728–41 (2011). Gatica, D., Lahiri, V. & Klionsky, D.J. Cargo recognition and degradation by selective autophagy. Nat Cell Biol. 20(3), 233–242 (2018). Yoshimoto, K.& Ohsumi, Y. Unveiling the Molecular Mechanisms of Plant Autophagy-From Autophagosomes to Vacuoles in Plants. Plant Cell Physiol. 59(7), 1337–1344 (2018). Zhao, P., Zhou, X.M., Zhao, L.L., Cheung, A.Y. & Sun, M-X. Autophagy- mediated compartmental cytoplasmic deletion is essential for tobacco pollen germination and male fertility. Autophagy 16(12), 2180–2192 (2020). Schuck, S. Microautophagy-distinct molecular mechanisms handle cargoes of many sizes. J Cell Sci. 133(17), jcs246322 (2020). Kiššová, I., Salin, B., Schaeffer, J., Bhatia, S., Manon, S. & Camougrand, N. Selective and non-selective autophagic degradation of mitochondria in yeast. Autophagy 3(4), 329–36 (2007). Mijaljica, D., Prescott, M. & Devenish, R.J. Microautophagy in mammalian cells: revisiting a 40-year-old conundrum. Autophagy 7(7), 673–82 (2011). Wang, L., Klionsky, D.J. & Shen, H.M. The emerging mechanisms and functions of microautophagy. Nat Rev Mol Cell Biol. 24(3),186–203 (2023). Mizushima, N., Yamamoto, A., Hatano, M., Kobayashi, Y., Kabeya, Y., Suzuki, K., Tokuhisa, T., Ohsumi, Y. & Yoshimori, T. Dissection of autophagosome formation using Apg5-deficient mouse embryonic stem cells. J Cell Biol. 152(4), 657–68 (2001). Kang, B.H., Plescia, J., Dohi, T., Rosa, J., Doxsey, S.J. & Altieri, D.C. Regulation of tumor cell mitochondrial homeostasis by an organelle-specific Hsp90 chaperone network. Cell 131(2), 257–70 (2007). Bao, F., Huang, X., Zhu, C., Zhang, X., Li, X. & Yang, S. Arabidopsis HSP90 protein modulates RPP4-mediated temperature-dependent cell death and defense responses. New Phytol. 202(4), 1320–1334 (2014). Borg, M., Jacob, Y., Susaki, D., LeBlanc, C., Buendía, D., Axelsson, E., Kawashima, T., Voigt, P., Boavida, L., Becker, J., Higashiyama, T., Martienssen, R. & Berger, F. Targeted reprogramming of H3K27me3 resets epigenetic memory in plant paternal chromatin. Nat Cell Biol. 22(6), 621–629 (2020). Williams, J.H., Taylor, M.L. & O'Meara, B.C. Repeated evolution of tricellular (and bicellular) pollen. Am J Bot. 101(4), 559–71 (2014). Sanchez-Vera, V., Kenchappa, C.S., Landberg, K., Bressendorff, S., Schwarzbach, S., Martin, T., Mundy, J., Petersen, M., Thelander, M. & Sundberg, E. Autophagy is required for gamete differentiation in the moss Physcomitrella patens . Autophagy 13(11), 1939–1951 (2017). Norizuki, T., Minamino, N., Sato, M., Tsukaya, H. & Ueda, T. Dynamic rearrangement and autophagic degradation of mitochondria during spermiogenesis in the liverwort Marchantia polymorpha . Cell Rep. 39(11), 110975 (2022). Zhou, X., Zhao, P. & Sun. M-X. Autophagy in sexual plant reproduction: new insights. J Exp Bot. 72(22), 7658–7667 (2021). Chanoca, A., Kovinich, N., Burkel, B., Stecha, S., Bohorquez-Restrepo A., Ueda T., Eliceiri K.W., Grotewold E. & Otegui M.S. Anthocyanin Vacuolar Inclusions Form by a Microautophagy Mechanism. Plant Cell 27(9), 2545–2559 (2015). Ding, X., Zhang, X., Paez-Valencia, J., McLoughlin, F., Reyes, F.C., Morohashi, K., Grotewold, E., Vierstra R.D. & Otegui M.S. Microautophagy Mediates Vacuolar Delivery of Storage Proteins in Maize Aleurone Cells. Front Plant Sci. 13, 833612 (2022). Izumi, M., Ishida, H., Nakamura, S., Hidema, J. Entire Photodamaged Chloroplasts Are Transported to the Central Vacuole by Autophagy. Plant Cell 29(2), 377–394 (2017). Nakamura, S., Hidema, J., Sakamoto, W., Ishida, H. & Izumi, M. Selective Elimination of Membrane-Damaged Chloroplasts via Microautophagy. Plant Physiol. 177(3), 1007–1026 (2018). Samakovli, D., Margaritopoulou, T., Prassinos, C., Milioni, D. & Hatzopoulos, P. Brassinosteroid nuclear signaling recruits HSP90 activity. New Phytol. 203(3), 743–57 (2014). Huang, S., Monaghan, J., Zhong, X., Lin, L., Sun, T., Dong, O.X. & Li, X. HSP90s are required for NLR immune receptor accumulation in Arabidopsis. Plant J. 79(3), 427–39 (2014). Yamamoto, H., Zhang, S. & Mizushima, N. Autophagy genes in biology and disease. Nat Rev Genet. 12, 1–19 (2023). Sato, M., Sato, K., Tomura, K., Kosako, H. & Sato, K. The autophagy receptor ALLO-1 and the IKKE-1 kinase control clearance of paternal mitochondria in Caenorhabditis elegans . Nat Cell Biol. 20(1), 81–91 (2018). Mori, T., Kuroiwa, H., Higashiyama, T. & Kuroiwa, T. GENERATIVE CELL SPECIFIC 1 is essential for angiosperm fertilization. Nat Cell Biol. 8(1), 64–71 (2006). Sprunck, S., Rademacher, S., Vogler, F., Gheyselinck, J., Grossniklaus, U. & Dresselhaus, T. Egg cell-secreted EC1 triggers sperm cell activation during double fertilization. Science 338(6110), 1093–7 (2012). Wang, W., Xiong, H., Cyprys, P., Malka, R., Flores-Tornero, M., Zhao, P., Peng, X., Sprunck, S. & Sun, M-X. DMP8 and 9 regulate HAP2/GCS1 trafficking for the timely acquisition of sperm fusion competence. Proc Natl Acad Sci U S A. 119(45), e2207608119 (2022). Rotman, N., Durbarry, A., Wardle, A., Yang, W.C., Chaboud, A., Faure, J.E., Berger, F., Twell, D. (2005). A novel class of MYB factors controls sperm cell formation in plants. Curr Biol. 15, 244-248. Clough, S.J., Bent, A.F. (1998). Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana . Plant J. 16(6), 735-43. Li, W., Huang, X., Zou, J., Wu, J., Jiao, H., Peng, X., Sun, M-X. (2020). Three STIGMA AND STYLE STYLISTs Pattern the Fine Architectures of Apical Gynoecium and Are Critical for Male Gametophyte-Pistil Interaction. Curr Biol. 30(23) , 4780-4788. Xin, H.P., Peng, X.B., Ning, J., Yan, T.T., Ma, L.G., Sun M-X. (2011). Expressed sequence-tag analysis of tobacco sperm cells reveals a unique transcriptional profile and selective persistence of paternal transcripts after fertilization. Sex Plant Reprod. 24(1) , 37-46. Zhao, P., Zhou, X.M., Zhang, L.Y., Wang, W., Ma, L.G., Yang, L.B., Peng, X.B., Bozhkov, P.V., Sun, M-X. (2013). A bipartite molecular module controls cell death activation in the Basal cell lineage of plant embryos. PLoS Biol. 11(9) , e1001655. Shi, C., Luo, P., Du, Y.T., Chen, H., Huang, X., Cheng, T.H., Luo, A., Li, H.J., Yang, W.C., Zhao, P., Sun, M-X. (2019). Maternal control of suspensor programmed cell death via gibberellin signaling. Nat Commun. 10(1) , 3484. Additional Declarations There is NO Competing Interest. Supplementary Files HuangxrSI.pdf SUPPLEMENTARY FIGURES VideoS1.avi Video S1 VideoS4.avi Video S4 VideoS2.avi Video S2 VideoS5.avi Video S5 VideoS3.avi Video S3 Cite Share Download PDF Status: Published Journal Publication published 21 Jan, 2026 Read the published version in Nature Plants → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6690586","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":465842248,"identity":"54756ec4-b5d4-463e-a51f-682a8cc8fa7c","order_by":0,"name":"Meng-Xiang Sun","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAArklEQVRIiWNgGAWjYNCCCoYEECVBgpYzJGthbCNFi+6M9GfShfMO5xkcYD54m4fBLo+gFrMzZ8ykZ247XGxwgC3ZmochuZiwluM9bNK8224nbjjAYybNw3AgsYGglsPsz6R554C08H8jUsvxBjNp3gawLWxEajlzxtia59j/xJmH2Ywt5xgkE6HlRvrD2zw1aYl9x5sf3nhTYUdYCwIwgwgD4tWPglEwCkbBKMADAGIwOzy2sWj6AAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-6959-8405","institution":"Wuhan University","correspondingAuthor":true,"prefix":"","firstName":"Meng-Xiang","middleName":"","lastName":"Sun","suffix":""},{"id":465842249,"identity":"a06506fa-78a5-4aaa-86a6-084113b0a298","order_by":1,"name":"Xiaorong Huang","email":"","orcid":"","institution":"Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Xiaorong","middleName":"","lastName":"Huang","suffix":""},{"id":465842250,"identity":"0ceefe87-fcd5-4a0f-8f6e-a8fb9b8aa89b","order_by":2,"name":"Linlin Zhao","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Linlin","middleName":"","lastName":"Zhao","suffix":""},{"id":465842251,"identity":"cb246ce2-1f0c-41f8-988d-992ab88e1e54","order_by":3,"name":"Zonglin Liu","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Zonglin","middleName":"","lastName":"Liu","suffix":""},{"id":465842252,"identity":"08288ee4-b42c-453b-b536-64496e734ead","order_by":4,"name":"Ni Long","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Ni","middleName":"","lastName":"Long","suffix":""},{"id":465842253,"identity":"cf5d05b3-a8f0-475f-acf0-3da469016583","order_by":5,"name":"Wenxuan Zou","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Wenxuan","middleName":"","lastName":"Zou","suffix":""},{"id":465842254,"identity":"2d679389-98ef-40f0-9107-375a320bd8e7","order_by":6,"name":"Feng Gong","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Gong","suffix":""},{"id":465842255,"identity":"eb493316-e5a8-46e1-8f6c-a350b220834a","order_by":7,"name":"tianhcheng Cheng","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"tianhcheng","middleName":"","lastName":"Cheng","suffix":""},{"id":465842256,"identity":"e313a351-463b-47ae-a6d7-9201d3a601e8","order_by":8,"name":"Ce Shi","email":"","orcid":"https://orcid.org/0000-0002-8920-8750","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Ce","middleName":"","lastName":"Shi","suffix":""},{"id":465842257,"identity":"14917c11-34af-4443-8236-e3dc3db40918","order_by":9,"name":"Xuecheng Zhang","email":"","orcid":"https://orcid.org/0000-0001-8667-4740","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Xuecheng","middleName":"","lastName":"Zhang","suffix":""},{"id":465842258,"identity":"a2453353-2c95-4803-9f50-f26af911e703","order_by":10,"name":"Wei Wang","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Wang","suffix":""},{"id":465842259,"identity":"1cd8e168-d2d5-4755-899a-5b51a41465c7","order_by":11,"name":"Hong Chen","email":"","orcid":"","institution":"Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Hong","middleName":"","lastName":"Chen","suffix":""},{"id":465842260,"identity":"2649e113-09e4-45fb-8375-2c169f251b45","order_by":12,"name":"Alice Cheung","email":"","orcid":"https://orcid.org/0000-0002-7973-022X","institution":"University of Massachusetts","correspondingAuthor":false,"prefix":"","firstName":"Alice","middleName":"","lastName":"Cheung","suffix":""}],"badges":[],"createdAt":"2025-05-18 08:30:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6690586/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6690586/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41477-025-02216-1","type":"published","date":"2026-01-21T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":83919749,"identity":"f357eb96-2401-4ad2-a53b-10b463202f6d","added_by":"auto","created_at":"2025-06-04 13:24:03","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2057471,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMicromitophagic elimination of paternal mitochondria during male gametogenesis.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e,\u003cstrong\u003eb\u003c/strong\u003e, Mitochondria in GC, SC pair (Mature Pollen stage, MP) and SC pair (Pollen Tube stage, PT) of wild-type (WT) (\u003cstrong\u003ea\u003c/strong\u003e) and \u003cem\u003eatg5-1 \u003c/em\u003e(\u003cstrong\u003eb\u003c/strong\u003e). The \u003cem\u003eProTIP5;1: TIP5;1-RFP ProDUO1:Mito-GFP \u003c/em\u003edouble-labeled transgenic line (Two-markers line) was used as WT and crossed with \u003cem\u003eatg5-1\u003c/em\u003eto display vacuole and mitochondria respectively. GCs and SC pairs were scanned using Z-stack of confocal microscope, and all sections were merged to exhibit cell images by Maximum Intensity Projection, and then the number of mitochondria (GFP signal dots) were counted (See Methods for additional details). \u003cstrong\u003ec\u003c/strong\u003e, The comparison of mitochondrial (Mt) numbers among WT GC, WT SC pair (MP), WT SC pair (PT), \u003cem\u003eatg5-1\u003c/em\u003e GC,\u003cem\u003e atg5-1\u003c/em\u003e SC pair (MP) and \u003cem\u003eatg5-1\u003c/em\u003e SC pair (PT). Results are from three independent experiments. Statistical significance is determined by two-way ANOVA (P \u0026lt; 0.001, ***; n.s., no significance) and the error bars show standard deviation (SD). \u003cstrong\u003ed\u003c/strong\u003e, The dynamic of paternal mitochondria in the two-marker line. The GFP-labeled mitochondria were co-localized inside the RFP-labeled vacuole, showing that paternal mitochondria had entered the vacuoles in male germline (MG) cells. The white dotted lines define the cell morphology. White arrows indicate the paternal mitochondria in vacuole; magenta arrows indicate vacuoles. \u003cstrong\u003ee-p\u003c/strong\u003e, The TEM analysis of mitochondrial microautophagy in GC (\u003cstrong\u003ee-l\u003c/strong\u003e) and SC (\u003cstrong\u003em-p\u003c/strong\u003e) of \u003cem\u003eA. thaliana\u003c/em\u003e. (\u003cstrong\u003ef\u003c/strong\u003e, \u003cstrong\u003eh\u003c/strong\u003e,\u003cstrong\u003ej\u003c/strong\u003e, \u003cstrong\u003el\u003c/strong\u003e, \u003cstrong\u003en\u003c/strong\u003e, \u003cstrong\u003ep\u003c/strong\u003e) showed the higher magnification view of the dashed box regions in (\u003cstrong\u003ee\u003c/strong\u003e, \u003cstrong\u003eg\u003c/strong\u003e, \u003cstrong\u003ei\u003c/strong\u003e, \u003cstrong\u003ek\u003c/strong\u003e, \u003cstrong\u003em\u003c/strong\u003e,\u003cstrong\u003e o\u003c/strong\u003e). The vacuole membrane invaginated and then engulfed paternal mitochondria (\u003cstrong\u003ee-h\u003c/strong\u003e). The trapped paternal mitochondria entered into vacuole and formed the micromitophagic vesicle. Several mitochondria could be sequestered simultaneously in the same micromitophagic vesicle (\u003cstrong\u003ei\u003c/strong\u003e, \u003cstrong\u003ej\u003c/strong\u003e, \u003cstrong\u003em\u003c/strong\u003e, \u003cstrong\u003en\u003c/strong\u003e). The micromitophagic vesicle with paternal mitochondria was degraded in the vacuole (\u003cstrong\u003ek\u003c/strong\u003e, \u003cstrong\u003el\u003c/strong\u003e, \u003cstrong\u003eo\u003c/strong\u003e, \u003cstrong\u003ep\u003c/strong\u003e). The magenta arrows indicate vacuoles; the blue arrows indicate the membrane of micromitophagic vesicles; the asterisks (*) indicate mitochondria, and the yellow asterisks indicate specifically degraded mitochondria. The observation is based on over 1000 samples of pollen at different development stages. Scale bars, 5μm (\u003cstrong\u003ea\u003c/strong\u003e, \u003cstrong\u003eb\u003c/strong\u003e, \u003cstrong\u003ed\u003c/strong\u003e), 1μm (\u003cstrong\u003ee\u003c/strong\u003e, \u003cstrong\u003eg\u003c/strong\u003e, \u003cstrong\u003ei\u003c/strong\u003e, \u003cstrong\u003ek\u003c/strong\u003e, \u003cstrong\u003em\u003c/strong\u003e,\u003cstrong\u003e o\u003c/strong\u003e), 200nm (\u003cstrong\u003ef\u003c/strong\u003e, \u003cstrong\u003eh\u003c/strong\u003e,\u003cstrong\u003ej\u003c/strong\u003e, \u003cstrong\u003el\u003c/strong\u003e, \u003cstrong\u003en\u003c/strong\u003e, \u003cstrong\u003ep\u003c/strong\u003e). GC, generative cell; SC, sperm cell; GN, generative cell nucleus; SN, sperm cell nucleus.\u003c/p\u003e","description":"","filename":"HuangxrFigures1.png","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/89cf3c60c283c7b3b0cb52ff.png"},{"id":83919755,"identity":"811e446d-9dac-47a5-95d1-35f1826f41fa","added_by":"auto","created_at":"2025-06-04 13:24:03","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1208193,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVacuole-localized ATG5 interacts with mitochondria-localized HSP90.2 to regulate micromitophagy.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, The subcellular localization analysis of fusion protein exhibits that ATG5 migrates to the vacuole, revealing as a non-classical localization. The GFP-tagged ATG5 is driven by MG-specific DUO1 promoter. TIP5;1-RFP is the marker labeling the vacuoles in the SCs. \u003cstrong\u003eb\u003c/strong\u003e, Double immunolabelling with ATG5 and RFP antibodies detected on the GC (left panel) and SC (right panel) from the two-markers line show the ATG5 targeting to vacuoles. In the two-markers line, the vacuoles in GC and SC are labeled by TIP5;1-RFP, and the magenta and black arrows indicate gold particles for anti-ATG5 (5 nm) and anti-RFP (10 nm) respectively. \u003cstrong\u003ec\u003c/strong\u003e-\u003cstrong\u003ee\u003c/strong\u003e, Yeast two-hybridization (Y2H) (\u003cstrong\u003ec\u003c/strong\u003e), Bimolecular fluorescence complementation assay (BiFC) (\u003cstrong\u003ed\u003c/strong\u003e) and Co-immunoprecipitation (Co-IP) analysis (\u003cstrong\u003ee\u003c/strong\u003e) reflect directly the interaction (\u003cstrong\u003ec\u003c/strong\u003e, \u003cstrong\u003ed\u003c/strong\u003e) and molecular complex formation (\u003cstrong\u003ee\u003c/strong\u003e) between ATG5 and HSP90.2. \u003cstrong\u003ef\u003c/strong\u003e, The expression pattern of HSP90.2 during pollen development and male gametogenesis. The RFP-tagged HSP90.2 is driven by native promoter. White circles (middle) and white squares (bottom) indicate GC and SCs (insets show magnified), respectively. \u003cstrong\u003eg\u003c/strong\u003e, The subcellular localization analysis showed that HSP90.2 is able to localize in the mitochondria in SC. The paternal mitochondria are indicated by MG-specific Mito-GFP signals. \u003cstrong\u003eh\u003c/strong\u003e,\u003cstrong\u003ei\u003c/strong\u003e, Deletion of HSP90.2 results in abnormal accumulation of paternal mitochondria in the SC pair. The \u003cem\u003eProTIP5;1: TIP5;1-RFP ProDUO1:Mito-GFP \u003c/em\u003edouble-labeled transgenic line (Two-markers line) was used as WT and crossed with \u003cem\u003ehsp90.2-2 \u003c/em\u003eto display mitochondria (\u003cstrong\u003eh\u003c/strong\u003e). The SC pairs were scanned using Z-stack of confocal microscope, and then all sections were merged to exhibit cell images by Maximum Intensity Projection. The number comparison of paternal mitochondria between WT SC pair and \u003cem\u003ehsp90.2-2\u003c/em\u003e SC pair in mature pollen grains (\u003cstrong\u003ei\u003c/strong\u003e). Results are from three independent experiments; in each of the experiments, the mitochondrial number from about 20 SC pair of WT and \u003cem\u003ehsp90.2-2 \u003c/em\u003ewere counted. Statistical significance is determined by Student's t-test (P \u0026lt; 0.001, ***) and the error bars show standard deviation (SD). Scale bars, 2μm (\u003cstrong\u003ea\u003c/strong\u003e, \u003cstrong\u003eg\u003c/strong\u003e), 100nm (\u003cstrong\u003eb\u003c/strong\u003e), 25μm (\u003cstrong\u003ed\u003c/strong\u003e), 5μm (\u003cstrong\u003ef\u003c/strong\u003e, \u003cstrong\u003eh\u003c/strong\u003e). GC, generative cell; SC, sperm cell; Va, vacuole.\u003c/p\u003e","description":"","filename":"HuangxrFigures3.png","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/3f168ddd0fe13f25dd10b712.png"},{"id":83920166,"identity":"b2e00f12-6765-4bd1-b9fa-46768c05d42f","added_by":"auto","created_at":"2025-06-04 13:32:03","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2389467,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMale germline cell fate determination triggers micromitophagy.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea-m\u003c/strong\u003e, The TEM analysis of micromitophagy in the \u003cem\u003eA. thaliana\u003c/em\u003e sperm-like cell. The asterisk (*) indicate mitochondria. \u003cstrong\u003ea\u003c/strong\u003e, Micromitophagy occurred within various-sized vacuoles in the sperm-like cell established by Huang and Sun (2022), which were indicated by white boxes. \u003cstrong\u003eb\u003c/strong\u003e, The vestige of degradative mitochondrion (white asterisk (*)) in a vacuole. The degradation product is indicated by green arrows. \u003cstrong\u003ec-f\u003c/strong\u003e, The mitochondria were captured by the vacuole membrane invagination and the micromitophagic vesicles were formed. \u003cstrong\u003eg\u003c/strong\u003e, A vacuole was able to generate multiple invaginations in different sites to devour mitochondria at the same time. Black boxes indicate that the vacuole membrane was swallowing up the mitochondria, and black arrows indicate the micromitophagic vesicles. \u003cstrong\u003eh\u003c/strong\u003e, One mitochondrion was engulfed by the vacuole along with its surrounding cytoplasm, while a second mitochondrion appeared to be in waiting. \u003cstrong\u003ei\u003c/strong\u003e, Two mitochondria were sequestered simultaneously in the same micromitophagic vesicle. The blue arrows mark the constricted region of the engulfing vacuolar membrance; the magenta arrows indicate the membrane of micromitophagic vesicles. \u003cstrong\u003ej-l\u003c/strong\u003e, The membrane of micromitophagic vesicle was gradually disintegrated, which is indicated by yellow arrows. \u003cstrong\u003em\u003c/strong\u003e, Mitochondrion (yellow asterisk (*)) was digested from outside to inside, and the degradation products dissolved into the vacuolar matrix, which is indicated by green arrows. (\u003cstrong\u003ed\u003c/strong\u003e, \u003cstrong\u003ef\u003c/strong\u003e, \u003cstrong\u003ek\u003c/strong\u003e) showed the higher magnification view of the regions enclosed in the dashed boxes in (\u003cstrong\u003ec\u003c/strong\u003e, \u003cstrong\u003ee\u003c/strong\u003e, \u003cstrong\u003ej\u003c/strong\u003e). \u003cstrong\u003en\u003c/strong\u003e, 3D models from electron tomography (ET) analysis of micromitophagy. Magenta, green, yellow and purple indicated vacuole membrane, mitochondrial membrane, the degrading vacuole membrane and the degrading mitochondrial membrane, respectively. The observation is based on over 500 samples of pollen. Tomographic slice images and movies are shown in Extended Data Fig. 4c and Movie S2-S5. Scale bars, 5μm (\u003cstrong\u003ea\u003c/strong\u003e), 1μm (\u003cstrong\u003ec\u003c/strong\u003e, \u003cstrong\u003ee\u003c/strong\u003e, \u003cstrong\u003ej\u003c/strong\u003e), 200nm (\u003cstrong\u003eb\u003c/strong\u003e, \u003cstrong\u003ed\u003c/strong\u003e, \u003cstrong\u003ef-i\u003c/strong\u003e, \u003cstrong\u003ek-m\u003c/strong\u003e). Va, vacuole.\u003c/p\u003e","description":"","filename":"HuangxrFigures5.png","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/a4863c508b1012c33270a42d.png"},{"id":83919752,"identity":"e0419565-4f9e-4501-ad70-bdc1203931bf","added_by":"auto","created_at":"2025-06-04 13:24:03","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2025596,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePre-fertilization paternal micromitophagy is conserved in angiosperms.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, Panoramic image of an \u003cem\u003eO. sativa\u003c/em\u003e GC displaying various-sized vacuoles. \u003cstrong\u003eb\u003c/strong\u003e, \u003cstrong\u003ec\u003c/strong\u003e, The micromitophagy occurred in \u003cem\u003eO. sativa\u003c/em\u003e GC. (\u003cstrong\u003eb\u003c/strong\u003e) and (\u003cstrong\u003ec\u003c/strong\u003e) were derived from different sections of the same pollen, indicating that a vacuole was able to generate multiple invaginations in different sites to devour paternal mitochondrion (PM) simultaneously. \u003cstrong\u003ed\u003c/strong\u003e, Panoramic image of a \u003cem\u003eN. tabacum\u003c/em\u003e GC displaying various-sized vacuoles. \u003cstrong\u003ee-g\u003c/strong\u003e, The micromitophagy occurred in \u003cem\u003eN. tabacum\u003c/em\u003eGC. (\u003cstrong\u003ee\u003c/strong\u003e) The higher magnification view of the regions enclosed in the dashed boxes in (\u003cstrong\u003ed\u003c/strong\u003e). \u003cstrong\u003eh\u003c/strong\u003e, Panoramic image of a \u003cem\u003eL. longiflorum\u003c/em\u003e (lily) GC displaying various-sized vacuoles. \u003cstrong\u003ei\u003c/strong\u003e,\u003cstrong\u003ej\u003c/strong\u003e, Type 1: the vacuole membrane invagination-mediated PM microautophagy was exhibited in \u003cem\u003eL. longiflorum\u003c/em\u003e GC. \u003cstrong\u003ek\u003c/strong\u003e, \u003cstrong\u003el\u003c/strong\u003e, Type 2: the micromitophagy mediated direct fusion of vacuole and PM in \u003cem\u003eL. longiflorum\u003c/em\u003e GC. The vacuole membrane protruded a projection to capture the PM, and then fused with the PM membrane to produce a vacuole-mitochondrion contact (\u003cstrong\u003ek\u003c/strong\u003e). The formation of this connective structure led to the mitochondrial matrix entering the vacuole and being digested (\u003cstrong\u003el\u003c/strong\u003e). (\u003cstrong\u003ei\u003c/strong\u003e, \u003cstrong\u003ek\u003c/strong\u003e, \u003cstrong\u003el\u003c/strong\u003e) indicate that a vacuole could generate multiple invaginations or protrusions to devour PM at the same time. For (\u003cstrong\u003ea\u003c/strong\u003e-\u003cstrong\u003el\u003c/strong\u003e), the black asterisk (*) and Va indicate vacuoles, the green asterisk (*) indicate intact mitochondria, and the yellow asterisk (*) indicate degraded mitochondria. \u003cstrong\u003em\u003c/strong\u003e, Statistical analysis of the PM number in GCs of \u003cem\u003eN. tabacum\u003c/em\u003e, \u003cem\u003eO. sativa\u003c/em\u003e and \u003cem\u003eL. longiflorum\u003c/em\u003e. Statistical significance is determined by one-way ANOVA (P \u0026lt; 0.001, ***) and the error bars show standard deviation (SD). \u003cstrong\u003en\u003c/strong\u003e, Mitochondria in the late GC and SC from wild-type (WT) and \u003cem\u003eNtatg5\u003c/em\u003etobacco. The late GC and SC were isolated from pollen tubes at 1.5 hours and 24 hours after germination and subsequently stained by Mitotracker Green to display mitochondria. GCs and SC pairs were scanned using Z-stack of confocal microscope, and then all sections were merged to exhibit cell images by Maximum Intensity Projection. \u003cstrong\u003eo\u003c/strong\u003e, The comparison of paternal mitochondrial numbers among WT GC, WT SC pair, \u003cem\u003eNtatg5 \u003c/em\u003eGC and \u003cem\u003eNtatg5 \u003c/em\u003eSC pair. Results are from three independent experiments; in each of the experiments, about 10-20 samples were counted. Statistical significance is determined by two-way ANOVA (P \u0026lt; 0.001, ***) and the error bars show standard deviation (SD). Scale bars, 500nm (\u003cstrong\u003eb\u003c/strong\u003e, \u003cstrong\u003ec\u003c/strong\u003e, \u003cstrong\u003ee-g\u003c/strong\u003e, \u003cstrong\u003ei-l\u003c/strong\u003e), 2μm (\u003cstrong\u003ea\u003c/strong\u003e, \u003cstrong\u003ed\u003c/strong\u003e), 10μm (\u003cstrong\u003eh\u003c/strong\u003e, \u003cstrong\u003en\u003c/strong\u003e).\u003c/p\u003e","description":"","filename":"HuangxrFigures7.png","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/a01f00bc72a811f29f1a7870.png"},{"id":83919754,"identity":"6b2b012b-63da-459b-ace7-af601a41c1ad","added_by":"auto","created_at":"2025-06-04 13:24:03","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2045825,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDisruption of pre-fertilization paternal micromitophagy delays zygote development.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e, Statistics of \u003cem\u003eA. thaliana\u003c/em\u003e hybrid zygotes at different stages 20 hours after pollination (HAP). Wild-type (WT) and \u003cem\u003eatg5-1\u003c/em\u003e were used as the paternal parent (♂), respectively; the \u003cem\u003epDD45:GFP\u003c/em\u003e marker line (EC line) was used as the maternal parent (♀). Magenta dotted lines indicate nucleus. The five stages were divided according to their morphological characters: the egg cell exhibited an obvious polarity with a nucleus at the apical pole and a large vacuole (Va, indicated by an arrow) at the basal pole. After fertilization, the large vacuole disappeared and the fused nucleus migrated toward the center in the fertilized egg cell, and then giving rise to a spherical zygote. The elongating zygote is undergoing polar growth and the nucleus has not yet moved to the top of the cell; in the elongated zygote, the zygote elongation has completed and the nucleus is at its apex and the polarity is established. Results are from three independent experiments; in-each of the experiments, about 200 ovules were counted. The number of ovule counted: WT (♂)×EC line (♀), n=739; \u003cem\u003eatg5 \u003c/em\u003e(♂)×EC line (♀), n=465. Tabulated data average from three experiments, ± SD. Statistical significance is determined by two-way ANOVA (P \u0026lt; 0.001, ***). \u003cem\u003eP\u003c/em\u003e value is for \u003cem\u003eatg5 \u003c/em\u003e(♂)×EC line (♀) vs. WT (♂)×EC line (♀). NA indicates not applicable. \u003cstrong\u003eb\u003c/strong\u003e,\u003cstrong\u003ec\u003c/strong\u003e, The paternal mitochondria were shown in the hybrid zygotes at different developmental stages (8HAP, 14HAP and 20HAP). TIP5;1-RFP displays the synergid cell (SY)-specific expression. After fertilization, the RFP fluorescence (white signals) from the degrading synergid cell can serve as a foil to clear outline the zygotes and early embryos for convenient observation. thus, two-marker line was used as the maternal parent (♀). (\u003cstrong\u003eb\u003c/strong\u003e) The results of two-marker line (♂)×two-markers line (♀); (\u003cstrong\u003ec\u003c/strong\u003e) The results of \u003cem\u003eatg5-1\u003c/em\u003e two-marker line (♂)×two-marker line (♀). \u003cstrong\u003ed\u003c/strong\u003e, Statistics of \u003cem\u003eN. tabacum\u003c/em\u003e hybrid zygotes or proembryos at different developmental stages in ovaries at 108HAP. WT and\u003cem\u003e Ntatg5\u003c/em\u003e were used as the paternal parent (♂); WT was used as the maternal parent (♀). Magenta arrows indicate cells derived from the zygote. Results are from three independent experiments; in each of the experiments, about 60 ovules were counted. The number of ovules counted: WT (♂)×WT (♀), n=190; WT (♀)×\u003cem\u003eNtatg5\u003c/em\u003e (♂), n=199. Tabulated data average from three experiments, ± SD. Statistical significance is determined by two-way ANOVA (P \u0026lt; 0.001, ***). \u003cem\u003eP\u003c/em\u003e value is for \u003cem\u003eNtatg5 \u003c/em\u003e(♂)×WT (♀) vs. WT (♂)×WT (♀). \u003cstrong\u003ee\u003c/strong\u003e, The stagnant zygotes can be observed in the hybrid ovules with \u003cem\u003eNtatg5 \u003c/em\u003eas the paternal parent at 144HAP, while all zygotes in hybrid ovules with WT as the paternal parent have been transformed into embryos. Results are from three independent experiments. The number of ovules counted: WT (♂)×WT (♀), n=515; WT (♀)×\u003cem\u003eNtatg5\u003c/em\u003e (♂), n=922. \u003cstrong\u003ef\u003c/strong\u003e, The \u003cem\u003eNtatg5 \u003c/em\u003eSC as the paternal parent resulted in seed abortion.The representative images of seed set from the the hybrid ovaries at 12 days after pollination (WT (♂)×WT (♀) is left panel, WT (♀)×\u003cem\u003eNtatg5\u003c/em\u003e (♂) is right panel) showed the aborted seeds indicated by arrows. Scale bars, 10μm (\u003cstrong\u003ea-e\u003c/strong\u003e), 1mm (\u003cstrong\u003ef\u003c/strong\u003e). PMt, paternal mitochondrion; DSY, degraded synergid cell.\u003c/p\u003e","description":"","filename":"HuangxrFigures9.png","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/de92e99dd3b84e39991409e3.png"},{"id":100863291,"identity":"29bf8d0c-1c81-4f13-accd-9c9221eb5c32","added_by":"auto","created_at":"2026-01-22 08:07:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":12911110,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/7de86869-7e49-43ab-ae00-445c45b11431.pdf"},{"id":83920167,"identity":"32245670-5a96-4fe9-b09b-586d37239525","added_by":"auto","created_at":"2025-06-04 13:32:03","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":11219761,"visible":true,"origin":"","legend":"SUPPLEMENTARY FIGURES","description":"","filename":"HuangxrSI.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/f9357bc659ee274173bd12c2.pdf"},{"id":83919757,"identity":"606ed186-7d31-4ff5-85b3-9742ec7532d8","added_by":"auto","created_at":"2025-06-04 13:24:03","extension":"avi","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":14905856,"visible":true,"origin":"","legend":"Video S1","description":"","filename":"VideoS1.avi","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/fe8a544f8159f810c8ef0aad.avi"},{"id":83919758,"identity":"e5b8bb6c-b399-42ba-b881-13e80e704c06","added_by":"auto","created_at":"2025-06-04 13:24:03","extension":"avi","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":42177024,"visible":true,"origin":"","legend":"Video S4","description":"","filename":"VideoS4.avi","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/60496ce283d7b899b7bd14d6.avi"},{"id":83919761,"identity":"77ae9eb4-1e0e-4dcb-b227-fd993b90e9cc","added_by":"auto","created_at":"2025-06-04 13:24:04","extension":"avi","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":42177024,"visible":true,"origin":"","legend":"Video S2","description":"","filename":"VideoS2.avi","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/1aea181021ff30048b6aacd9.avi"},{"id":83919760,"identity":"86ee22ef-054d-48e1-b070-68333b071b29","added_by":"auto","created_at":"2025-06-04 13:24:03","extension":"avi","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":42177024,"visible":true,"origin":"","legend":"Video S5","description":"","filename":"VideoS5.avi","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/52cb4f240875517ea9e238d0.avi"},{"id":83920168,"identity":"12b86203-06f6-4c5a-a132-97568feee49b","added_by":"auto","created_at":"2025-06-04 13:32:03","extension":"avi","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":42177024,"visible":true,"origin":"","legend":"Video S3","description":"","filename":"VideoS3.avi","url":"https://assets-eu.researchsquare.com/files/rs-6690586/v1/3d4096be7ea5da30e6a0ec90.avi"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"ATG5-HSP90.2-mediated micromitophagy as cytological basis for maternal inheritance of plant mitochondria","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMitochondria and plastids of the plant embryo are provided by the egg cell but not by the sperm cell (SC) via fertilization, which has been known as cytoplasm maternal inheritance, a classic example of non-Mendelian inheritance\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. The first event of non-Mendelian inheritance was discovered in 1909, Carl Correns reported the plastid maternal inheritance of leaf color in \u003cem\u003eMirabilis jalapa\u003c/em\u003e\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Subsequent research showed that the mitochondria of major angiosperms are strictly maternally inherited\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. However, after many years of the discovery of this phenomenon, what mechanisms regulate PME in plants remain obscure.\u003c/p\u003e \u003cp\u003eAlthough no clear mechanism has emerged, various hypotheses have been advanced for what might underlie the maternal inheritance of mitochondria in plants. The finding that male gametic mtDNA was reduced in angiosperms led to the suggestion that it might contribute to maternal mitochondrial inheritance\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. However, the mitochondrial endonuclease responsible for this process is still unknown. Potential DNA nucleases, the pollen-expressed Mg\u003csup\u003e2+\u003c/sup\u003e-Dependent exonuclease DEFECTIVE IN POLLEN ORGANELLE DNA DEGRADATION1 (DPD1) and Mitochondrial Endonuclease M20, have been reported to degrade mtDNA in VC but not in SCs\u003csup\u003e\u003cb\u003e\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e–\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. These nucleases are therefore unlikely to be relevant to PME. In addition, in the \u003cem\u003edpd1\u003c/em\u003e mutant, the paternal mtDNA carried by the SC was never transmitted to offspring\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e, and thus, there must be some unknown pathways specifically in MG cells for ensuring mitochondrial maternal inheritance in angiosperms. It was also proposed that paternal mitochondria might remain outside of the egg cell during gamete fusion\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e or could be excluded physically from the GC during microspore asymmetric division\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. However, no follow-up works have been reported to confirm these proposals. Recently, a breakthrough has been made in the mechanism of plastid maternal inheritance in tobacco. During male gametogenesis, an environmental temperature-regulated organelle exclusion mechanism and a DPD1-mediated genome degradation mechanism act in concert to prevent paternal transmission of plastid genes\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. During sperm development in \u003cem\u003eCaenorhabditis elegans\u003c/em\u003e, male mitochondria are exported out of the spermatids through mitopherogenesis\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. In fact, plant scientists have also been curious about whether PME occurs in the MG cells to ensure maternal mitochondrion heredity for decades.\u003c/p\u003e \u003cp\u003eAutophagy is an evolutionarily conserved mechanism and provides cells with a pathway for the continuous turnover of undesired organelles. It is usually divided into three main types: macroautophagy, microautophagy and chaperone-mediated autophagy\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. We found that the vacuoles needed for plant autophagy existed in the MG cells\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e and the number of paternal mitochondria continue to decrease during SC development, providing an intriguing clue that mitochondrial depletion by autophagy might be the means for PME in plants. Unfortunately, none of the three conventional autophagy mechanisms have a confirmed involvement in PME in angiosperms.\u003c/p\u003e \u003cp\u003eHere, we report that micromitophagy exists in plants and is triggered by MG cell fate determination. We have uncovered a novel role for the core autophagy gene \u003cem\u003eATG5\u003c/em\u003e to engage micromitophagy specifically for PME in the MG cells. We discovered that ATG5, which typically is located in the isolation membrane of developing autophagosome, is transferred to vacuoles where it binds to HSP90.2 of paternal mitochondria to mediate vacuole-mitochondria interaction and micromitophagy. We also demonstrate that superfluous paternal mitochondria entering the egg cell via fertilization will affect zygote development and subsequent embryogenesis, underscoring the biological importance of PME and revealing the deleterious effect if PME is hampered. Finally, our results reveal that this cell-type specific micromitophagy is a conserved mechanism contributing to mitochondrial maternal inheritance in angiosperms.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePersistent mitochondrial depletion during MG cell development\u003c/h2\u003e \u003cp\u003eWe first carefully observed the mitochondrial behavior in the MG cells. After the asymmetric division in the microspore of \u003cem\u003eA. thaliana\u003c/em\u003e that produces the GC and VC of pollen, the mitochondria inherited from microspore divided into smaller mitochondria in MG cells, instead of fusion to form larger mitochondria (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea-i). Similar mitochondrial fission events were also observed during the development of \u003cem\u003eO. sativa\u003c/em\u003e (rice) MG cells, suggesting that this is a conserved phenomenon in both dicotyledons and monocotyledons (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ej-n). However, although the mitochondrial fission looked normal, the mitochondrial number did not increase in MG cells during male gametogenesis. Instead, the number of mitochondria decreased from GC (average 19.7\u0026thinsp;\u0026plusmn;\u0026thinsp;4.0) to SCs (average 14.5\u0026thinsp;\u0026plusmn;\u0026thinsp;4.5 per SC pair in mature pollen grain), the product of another mitotic division of the GC, in \u003cem\u003eA. thaliana\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea,c). Importantly, by the \u003cem\u003ein vivo\u0026ndash;in vitro\u003c/em\u003e pollen growth assay\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e, we also found that the mitochondria in the SCs continued to reduce during 8 hours after pollination (HAP), the time needed for them to move into pollen tube (average 11.7\u0026thinsp;\u0026plusmn;\u0026thinsp;2.6 per SC pair in pollen tube) and before they are released to the embryo sac and fusing with the egg cell. This clearly suggests that the reduction of mitochondria in the SCs is not due to simple dilution by GC division but depends on an active elimination process (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea,c). The function of the pollen is to deliver the SCs to fertilize female gametes located at some distance from where it lands on a receptive pistil. The decline of mitochondria observed here therefore indicates that the PME process was activated before fertilization and still spatially distant from male-female gamete fusion.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eATG5 governs PME\u003c/h3\u003e\n\u003cp\u003eIn most eukaryotic organisms, autophagy (often refers to macroautophagy) is a major intracellular degradation system that removes damaged and/or useless organelles and unwanted cytoplasmic components and is regulated by conserved autophagy-related (\u003cem\u003eATG\u003c/em\u003e) genes\u003csup\u003e\u003cb\u003e\u003cspan additionalcitationids=\"CR18\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. To understand if an autophagy-dependent pathway is the mechanism underlying the PME process in plants, we analyzed the transcriptomes of wild-type (WT) VC and WT SC\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e, and found that the transcripts of \u003cem\u003eATG5\u003c/em\u003e, one of the core ATG genes, were more abundant and preferentially expressed in WT SC, implying that \u003cem\u003eATG5\u003c/em\u003e might play a potential key role in PME in \u003cem\u003eA. thaliana\u003c/em\u003e (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Detailed characterization of \u003cem\u003eatg5\u003c/em\u003e mutant\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e revealed that at the early bicellular stage, the mitochondrial number of GC was not significantly different between WT and \u003cem\u003eatg5\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea-c). These observations imply that ATG5 dysfunction did not affect the mitochondrial dynamics in the microspore. However, in contrast to the significant reduction of mitochondria in WT SCs during male gametogenesis, the mitochondrial numbers increased dramatically and continuously in \u003cem\u003eatg5\u003c/em\u003e SCs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea-c), which indicates that ATG5 regulates PME. Taken together, these results reflect that PME was continuing in SCs prior to fertilization under normal conditions, and this process during active mitochondrial fission was blocked by ATG5 dysfunction, leading to mitochondrial accumulation in \u003cem\u003eatg5\u003c/em\u003e SCs.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e3-Methyladenine (3-MA), a phosphatidylinositol 3-kinase (PtdIns3K) inhibitor, can disrupt autophagy by blocking autophagosome formation. To further examine the role of autophagy in PME, we treated the pollen grains of the transgenic lines, in which the male mitochondria were labeled by GFP with different concentrations of 3-MA, and then counted the mitochondrial number in the SCs within the pollen tube after 5 hours. Surprisingly, 3-MA treatment could not induce the increase of mitochondrial number in the SCs and did not show a defective phenotype similar to \u003cem\u003eatg5\u003c/em\u003e (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003e). This result suggests that the ATG5-regulated PME may be involved in a non-autophagosome pathway.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003ePME by a MG-specific micromitophagy pathway\u003c/h3\u003e\n\u003cp\u003eDuring macroautophagy, the double membrane-bound autophagosome sequesters organelles and then fuses with vacuoles (in yeast and plant) or lysosomes (in mammal) to release the internalized vesicle as autophagic body for the subsequent degradation. Therefore, this process involves complex membrane biogenesis\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Whereas, during microautophagy, it was suggested that the organelles were directly engulfed by the vacuoles without autophagosome mediation\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Previous studies have suggested that ATG5 may play an as-yet-undefined function in microautophagy in yeast\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. As transmission electron microscopy (TEM) is the classical approach essential to describe microautophagy that lacks other effective methods\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e, we used TEM to analyze the male gametogenesis process. Firstly, we carefully followed the behavior of the vacuole (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e), an indispensable organelle for macroautophagy and microautophagy in plant\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. In the microspores, a large central vacuole formed from the fusion of pre-existing small vacuoles predominated. After the microspore asymmetric division, the large vacuole was re-divided into small vacuoles, which were assigned to both GC and VC. Then, while the vacuoles in the VC gradually disappeared, those in the GC were retained. Eventually, at the mature pollen stage, many spherical vacuoles existed uniquely in the SCs, while vacuoles were no longer clearly visible in the VC (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Since the vacuole is indispensable for autophagy, these results imply that the autophagy responsible for plant PME could only occur in the MG cells. Meanwhile, our live-cell imaging showed that the mitochondria could indeed enter the vacuole during male gametogenesis, although it remains very difficult to catch the transient moments of mitochondrial contact and entrance (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed; Supplementary Video1).\u003c/p\u003e \u003cp\u003eFurthermore, the exhaustive TEM analysis of a huge number of samples revealed clearly that the mitochondrial degradation in the MG cells of \u003cem\u003eA. thaliana\u003c/em\u003e was actually via microautophagy rather than macroautophagy, and documented in detail the microautophagic process stepwise (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee-p). In MG cells, the vacuoles and mitochondria came into contact, the vacuole membrane invaginated (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee,f) and engulfed the mitochondria (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eg,h). The trapped mitochondria entered micromitophagic compartments, and they could be observed in progressively degraded states in the highly hydrolytic environment of the compartments (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ei-p). We also found that consistent with the aforementioned results (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea-c), as development of MG cells, the paternal mitochondria were continually reduced, but a few mitochondria always remained in SCs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eg,o), probably because necessary energy supply is still required for the later maturation of SCs and their function during fertilization, after which the remaining mitochondria would finally be removed. During plant male gametogenesis, the cytoplasmic contents, mainly the number of organelles, in the SCs is markedly reduced\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Thus, it would seem more practical and efficient by using the existing vacuoles to directly eliminate paternal mitochondria than to synthesizing new membrane structures and assembling autophagosomes for macromitophagy. In fact, we observed at least 1,000 \u003cem\u003eA. thaliana\u003c/em\u003e pollen grains by TEM, and found that autophagosomes, characterized by their double-membrane structure and depredating cytoplasmic contents, occurred exclusively in the VC (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), but never in the SCs. These observations support a model that macroautophagy does not usually occur in the SCs, and PME is executed by MG-specific microautophagy rather than macroautophagy.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eATG5 interacts with HSP90.2 to regulate PME\u003c/h3\u003e\n\u003cp\u003eSo far, the molecular mechanisms regulating microautophagy remains poorly understood and there is no technique available yet to purposefully disrupt the microautophagy of any targeted organelles\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. It was reported that \u003cem\u003eATG5\u003c/em\u003e dysfunction affected micromitophagy in yeast, not in mammalian cells\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e, but the mechanism in yeast is not known. The above-mentioned results (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003e) support that ATG5 might participates in micromitophagy in plant MG cells, offering a unique opportunity for exploring this novel function of ATG5 in multicellular organisms. Normally, ATG5 typically localizes in the isolation membrane, which is extending to envelop the autophagosome during its formation and involved in macroautophagy\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. However, this only known location of ATG5 cannot explain its participation in microautophagy. We proposed that the neofunctionalization of ATGs should be accompanied by their translocation from the isolation membrane to the vacuole membrane. Therefore, we made the high-resolution subcellular localization analysis of GFP-tagged ATG5 and ATG5 immuno-labeling. We found that upon MG cell fate determination ATG5 can indeed translocate onto the vacuoles in the MG cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003ea,b), and this enables ATG5 to play a non-canonical role in micromitophagy.\u003c/p\u003e \u003cp\u003eTo further reveal the molecular mechanism of ATG5-mediated micromitophagy, we obtained an ATG5 interacting protein, HEAT SHOCK PROTEIN 90.2 (HSP90.2), by selection of candidate genes from SC transcriptome data together with yeast two-hybridization (Y2H) verification (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003ec). Moreover, Bimolecular fluorescence complementation (BiFC) assay and Co-immunoprecipitation (Co-IP) analysis further supported the interaction between ATG5 and HSP90.2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003ed,e). In animals, a chaperone network comprising Hsp90 and its related molecules specifically target mitochondria in tumor cells, but not in most normal tissues, and thus antagonizing mitochondrial permeability transition and regulating mitochondrial homeostasis\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. This previous work in animal provides an important clue, namely, Hsp90 can be localized specifically to mitochondria in some specific cells to regulate mitochondrial fate. Thus, we firstly analyzed the HSP90.2 expression pattern during pollen development and male gametogenesis. In microspores, \u003cem\u003eHSP90.2-RFP\u003c/em\u003e driven by native \u003cem\u003eHSP90.2\u003c/em\u003e promoter presented no clear fluorescence signals (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003ef). However, right after microspore asymmetric division, the HSP90.2-RFP exhibited MG-specific signals in GC and SCs (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003ef). Subsequently, we introduced MG-specific mitochondrial marker (Mito-GFP) into the cells with HSP90.2-RFP fusion protein, and found that HSP90.2 was indeed able to locate onto SC\u0026rsquo;s mitochondria (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eg). Therefore, a logical explanation is that the vacuole-localized ATG5 interacts with the mitochondria-localized HSP90.2 to mediate paternal micromitophagy in MG cells.\u003c/p\u003e \u003cp\u003eWe sought genetic evidence to further confirm the conclusion from molecular interaction and localization results discussed above. We analyzed the \u003cem\u003ehsp90.2-2\u003c/em\u003e\u003csup\u003e\u003cb\u003e27\u003c/b\u003e\u003c/sup\u003e phenotype. Indeed, like what we observed in \u003cem\u003eatg5\u003c/em\u003e, the mutation of \u003cem\u003eHSP90.2\u003c/em\u003e led to a significant increase of mitochondrial number in SC pair (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003eh,i). This suggests that either ATG5 or HSP90.2 deletion can disrupt the interaction between vacuoles and mitochondria in MG cells, thereby blocking micromitophagy and PME, and leading to abnormal mitochondrial accumulation. This evidence supports that ATG5-HSP90.2 interaction is required for paternal micromitophagy.\u003c/p\u003e\n\u003ch3\u003eMG cell fate determination triggers micromitophagy\u003c/h3\u003e\n\u003cp\u003eSince distinct factors or developmental events can trigger different types of autophagy, clarifying the induction condition is a crucial part of studying an autophagic process and revealing its underlying regulatory mechanism. Since the micromitophagy reported here can only be detected in the MG cells but not in the VC, we contended that this process might be triggered by MG cell fate determination and evolved specifically for PME. H3K27me3 reprogramming is vital for the plant MG cell fate determination\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Previously, we have successfully induced VC fate switch towards SC by the VC-targeted H3K27me3 erasure, resulting in a sperm-like VC cell in pollen\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. These results aligned with increased expression level of \u003cem\u003eATG5\u003c/em\u003e and its relocalization on the vacuoles in the sperm-like cell (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea,d). The finding not only implies an underlying involvement of ATG5 in MG cell fate determination-triggered micromitophagy but also provides a unique material with which the detailed micromitophagy events that accompany the SC differentiation process can be monitored starting from the switch to the MG cell fate.\u003c/p\u003e \u003cp\u003eTherefore, taking advantage of this unique material, we monitored the autophagy events in both WT VC and the sperm-like VC. Astoundingly, the shift of VC fate into sperm-like not only successfully activated micromitophagy but also the process was brought to completion as in SCs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec and Supplementary Videos2-5) and could be conspicuously observed in these sperm-like cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ea,b). The micromitophagic process has thus far not been well characterized cytologically. With the much larger volume of its cytoplasm relative to the GC and SC, this sperm-like cell model provides an exceptional opportunity to monitor the process, allowing us an unprecedented view of a detailed step-by-step process of microautophagy. When a mitochondrion approached a vacuole, it was first captured by the vacuole membraneinvagination, then totally enveloped and finally released into the vacuole (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ec-i). Incipiently, a mitochondrion was boxed in a complete monolayer vacuous, and then this micromitophagic compartment flowed within the vacuole as its morphology became a \u0026ldquo;comet-like\u0026rdquo; form with a trailing tail (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ej,k). Later, the membrane of this micromitophagic compartment gradually disintegrated, resulting in the mitochondrion being exposed to the vacuolar matrix (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003el). Then, the mitochondrion would soon start to degrade (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003em) and finally become totally digested (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). Furthermore, the electron tomography (ET) analysis exhibited the micromitophagic process in sperm-like cell in a three-dimensional manner (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003en, Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec and Supplementary Videos2-5), which more accurately exhibits the spatial position of the vacuole membrane, micromitophagic compartment and mitochondria. These 3D reconstructions further confirm that mitochondria shown by ultrastructural sections are indeed inside the vacuoles and micromitophagy plays a critical role in the process of PME. Clearly, it is the MG cell fate determination that triggers the microautophagy-dependent degradation of paternal mitochondria. This interesting phenomenon also explained why micromitophagy was not detected in the VC.\u003c/p\u003e \u003cp\u003eDuring this cell fate transition triggered-microautophagy, three obvious features are particularly noteworthy. Firstly, a lot of vacuoles were generated simultaneously and they eliminated mitochondria synchronously (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Secondly, a vacuole was able to form multiple membrane invaginations at different sites to devour mitochondria at the same time (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003eg). Thirdly, a vacuole membrane invagination was routinely capable of swallowing not only one but at least two mitochondria at the same time (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ei; also see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003en). These features suggest remarkable efficiency can be achieved in the microautophagy-dependent PME in plant MG cells.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eMicromitophagy-mediated PME is a conserved strategy in angiosperms\u003c/h2\u003e \u003cp\u003eIt is well known that the two main branches of angiosperms, dicots and monocots, both include three-celled pollen and two-celled pollen species. In three-celled pollen type the GC divides in pollen grain and each mature pollen contains a VC and two SCs. Whereas, in two-celled pollen type mature pollen contains two cells, a VC and a GC, and the GC divides in pollen tubes after pollen germination\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Use of microautophagy to continuously eliminate paternal mitochondria in the immobile SC of \u003cem\u003eA. thaliana\u003c/em\u003e clearly has diverged from the macroautophagy process employed by the mobile SC of non-seed plants \u003cem\u003ePhyscomitrella patens\u003c/em\u003e (moss) and \u003cem\u003eMarchantia Polymorpha\u003c/em\u003e (liverwort) for cytoplasmic reduction and mitochondrial reorganization \u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. It is thus interesting to know whether the micromitophagy-mediated PME had become a universal strategy in MG cells of all type of angiosperms during plant evolution or that the processes have diverged between different branches of angiosperm.\u003c/p\u003e \u003cp\u003eWith \u003cem\u003eA. thaliana\u003c/em\u003e being the model dicotyledonous three-cell species, we selected three other model plants, \u003cem\u003eNicotiana tabacum\u003c/em\u003e (tobacco), \u003cem\u003eO. sativa\u003c/em\u003e (rice) and \u003cem\u003eLily longiflorum\u003c/em\u003e (lily) as the representatives of dicotyledonous two-celled, monocotyledonous three-celled and monocotyledonous two-celled pollen, respectively, and observed mitochondrial dynamics in their MG cells. Indeed, a micromitophagy process as in \u003cem\u003eA. thaliana\u003c/em\u003e was observed in all three plants (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea-l and Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), establishing that micromitophagy is indeed a conserved approach for PME for angiosperms. Further examination of \u003cem\u003eO. sativa\u003c/em\u003e showed that similar to \u003cem\u003eA. thaliana\u003c/em\u003e, micromitophagy only occurred in the small, dispersed vacuoles of the MG cells (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea-i), providing key evidence to further support that the vacuoles in the MG cells indeed had differentiated for the specialized PME process.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eHowever, there is also clear evidence for adaptive evolution of the micromitophagy process in PME during speciation. We observed, interestingly, that except for the vacuole membrane invagination-mediated paternal mitochondria engulfment (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ei,j and Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ej,k), the \u003cem\u003eL. longiflorum\u003c/em\u003e GC, which is larger and contains more mitochondria than its counterparts in \u003cem\u003eN. tabacum\u003c/em\u003e, \u003cem\u003eO. sativa\u003c/em\u003e and \u003cem\u003eA. thaliana\u003c/em\u003e, tended to have evolved an even more straightforward and efficient strategy to eliminate mitochondria, namely, the direct fusion between the paternal mitochondria and the vacuoles (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ek,l and Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003el). TEM revealed vacuole membrane extending a protrusion to capture mitochondria (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ek, upper right), and then fusing with the mitochondrial membrane to produce a vacuole- mitochondrion contact (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ek, lower left, and Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003el). The formation of this contact structure led to the fusion of the membranes and mixing of the two organelles with each other, resulting in the mitochondrial matrix internalized into the vacuole (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003el) and was digested. Moreover, several projections could occur simultaneously from a single vacuole and directed towards mitochondria. Thus, a vacuole was able to synchronously capture multiple mitochondria and quickly degrade them (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ek,l). The \u003cem\u003eL. longiflorum\u003c/em\u003e GC, probably due to its large size, more cytoplasm contents, and paternal mitochondria, has therefore evolved an even more efficient micromitophagy process than that observed in \u003cem\u003eA. thaliana\u003c/em\u003e and \u003cem\u003eO. sativa\u003c/em\u003e to achieve timely removal of mitochondria from the cell. These observations indicate that although the micromitophagy-dependent PME is a conserved strategy in angiosperms, the modes of micromitophagy may not be invariant in the diversified groups of angiosperms. The optimal mode of microautophagy for efficient PME to meet species-dependent requirement may have been evolved or selected in different MG cells.\u003c/p\u003e \u003cp\u003eMechanistically, we further investigated whether ATG5, which regulates PME in \u003cem\u003eA. thaliana\u003c/em\u003e that belongs to tricellular pollen species, also governs PME in plants of bicellular pollen species. The GC undergoes mitosis in the growing pollen tube to form SCs in \u003cem\u003eN. tabacum\u003c/em\u003e. Thus, to gain clearer images, we isolated the late GC from the pollen grains and the SC pair from the pollen tubes of WT and \u003cem\u003eNtatg5\u003c/em\u003e\u003csup\u003e\u003cb\u003e20\u003c/b\u003e\u003c/sup\u003e \u003cem\u003eN. tabacum\u003c/em\u003e respectively and then mitochondrial dynamics were monitored by the Mitotracker Green staining (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003en). Our result showed that, like \u003cem\u003eA. thaliana\u003c/em\u003e, active paternal mitochondrial reduction also occurred during \u003cem\u003eN. tabacum\u003c/em\u003e spermatogenesis, and the \u003cem\u003eNtATG5\u003c/em\u003e absence disturbed PME, which led to the abnormal accumulation of a large number of mitochondria in the MG cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003en,o), suggesting that the ATG5- regulated PME is also a conserved mechanism in different groups of angiosperms.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eExcess paternal mitochondria entering egg cell affects zygote development\u003c/h3\u003e\n\u003cp\u003eWhile maternal organelle inheritance is the norm, the benefit from the exclusion of paternal organelles is unclear. To test whether the excess paternal mitochondria may influence male-female gametes fusion or postfertilization development if PME is blocked before fertilization, we crossed the \u003cem\u003eatg5\u003c/em\u003e as paternal parent with egg cell \u003cem\u003epDD45:GFP\u003c/em\u003e marker line as maternal parent. In the hybrid ovules at 20 HAP, all fertilized egg cells developed into zygotes, indicating that \u003cem\u003eatg5\u003c/em\u003e SC was able to fuse with egg cell for fertilization and showed no difference from WT SCs. However, the zygotic development in \u003cem\u003eatg5\u003c/em\u003e pollen-pollinated ovaries was significantly delayed compared to that in WT pollen-pollinated ovaries (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea), demonstrating the importance of the deletion of the paternal mitochondria from the zygote for post fertilization development.\u003c/p\u003e \u003cp\u003eWe next monitored the paternal mitochondrial dynamic after fertilization in \u003cem\u003eA. thaliana\u003c/em\u003e zygotes at different stages to establish the importance of timely PME to zygote development. Firstly, we crossed the vacuole (RFP) and mitochondria (GFP) double markers (two-markers line) as the paternal parent into a WT maternal parent. With only a few mitochondria remaining in the SCs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea,c), these residual paternal mitochondria entered the egg cell together with the SC-delivered vacuoles via fertilization (Extended Data Fig.\u0026nbsp;6). Interestingly, the paternal mitochondria were found also co-localized with the SC-delivered vacuole in the fertilized egg cell (Extended Data Fig.\u0026nbsp;6). The continued co-compartmentation of the engulfed paternal mitochondria within the SC-delivered vacuole suggests that the SC-initiated microautophagy clearance system of paternal mitochondria may continue to function after fertilization. To test this, we crossed the two-markers line as paternal parent and monitored the fate of the paternal mitochondria in developing zygotes (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb). About 6 (6.1\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8, n\u0026thinsp;=\u0026thinsp;18) paternal mitochondria could be detected in the zygotes at 8 HAP, but by 14 HAP almost all of paternal mitochondria were already eliminated, and at 20 HAP paternal mitochondria could no longer be observed (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb). Therefore, mitochondrial clearance continues to function after fertilization. We then introduced the double markers line into \u003cem\u003eatg5\u003c/em\u003e [referred to as \u003cem\u003eatg5\u003c/em\u003e (two-markers line)]. When \u003cem\u003eatg5\u003c/em\u003e (two-markers line) was used as the paternal parent, more than 19 (19.5\u0026thinsp;\u0026plusmn;\u0026thinsp;3.2, n\u0026thinsp;=\u0026thinsp;15) paternal mitochondria could be detected in zygotes at 8 HAP, and the paternal mitochondria persistent even in the 20HAP zygotes along with the delayed zygote elongation (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec). This result suggests that though the seed fertility was not significantly disturbed, the excess paternal mitochondria delivered from \u003cem\u003eatg5\u003c/em\u003e SC to the normal egg cell have postponed the process of zygote-to-embryo transition in \u003cem\u003eA. thaliana\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eTo further confirm the conclusion, we used WT or \u003cem\u003eNtatg5\u003c/em\u003e pollen to pollinate WT \u003cem\u003eN. tabacum\u003c/em\u003e ovaries, with saturating amount of pollen grains. By isolating the embryo sacs from 108 HAP hybrid ovaries to analyze the embryogenesis process, we found that the hybrid embryos from the ovules pollinated with WT pollen had developed overwhelmingly to the developmental stages of 2-celled (64.3\u0026thinsp;\u0026plusmn;\u0026thinsp;1.7%) or 4-cell (35.4\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8%) proembryos. On the other hand, in the ovules pollinated with \u003cem\u003eNtatg5\u003c/em\u003e pollen, the proportion of zygote increased significantly (15.2\u0026thinsp;\u0026plusmn;\u0026thinsp;0.5%), while 4-celled proembryos decreased notably (7.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.4%) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed). Next, we observed the ovules from 144 HAP hybrid ovaries using an ovule whole-mount clearing technique. In contrast to all zygotes having developed into embryos in the hybrid ovules pollinated with WT pollen, at least\u0026thinsp;~\u0026thinsp;3% of zygotes stopped developing in the hybrid ovules pollinated with \u003cem\u003eNtatg5\u003c/em\u003e pollen (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ee), and then exhibited as seed abortion at 12 days after pollination (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef). Compared to WT SC, loss of ATG5 resulted in about 3-fold increase in the number of paternal mitochondria in \u003cem\u003eA. thaliana\u003c/em\u003e SC and about 5-fold increase in \u003cem\u003eN. tabacum\u003c/em\u003e SC (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eo). The hybrid zygotes in \u003cem\u003eN. tabacum\u003c/em\u003e displayed a more severe phenotype than that in \u003cem\u003eA. thaliana\u003c/em\u003e, probably due to the introduction of more additional paternal mitochondria into the fertilized egg cells, which was more than the zygotes can tolerate. These results not only elucidate a direct role of paternal mitochondria in hindering zygotic development, but also indicate that the negative effect of abnormal PME in MG cells on zygote-to-embryo transition is conserved in different groups of angiosperms.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eCompared to macroautophagy, microautophagy is much less investigated, especially in plants\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e where to-date just a few reports have dealt with microautophagic activities during its development. The pigment aggregates in \u003cem\u003eA. thaliana\u003c/em\u003e cotyledons and \u003cem\u003eE. grandiorum\u003c/em\u003e purple petals were directly engulfed by a single-layer vacuole membrane and became an autophagic body-like structure in vacuoles\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Similarly, the aleurone cells of maize endosperm delivered the prolamins to vacuoles via a microautophagic manner\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. In addition, the degradation of whole photo-damaged swollen chloroplasts in \u003cem\u003eA. thaliana\u003c/em\u003e was shown to be microautophagy, termed as “microchlorophagy”\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. However, the spatial and temporal events during plant microautophagic processes are far from precisely described, prohibiting advances in understanding the molecular mechanism that regulates microautophagy in plants. Herein, we evidenced for the first time that micromitophagy exists in angiosperms with detailed documentation and emerging mechanistic insights. This micromitophagy specifically occurred in MG cells but not in VC during pollen development. By triggering cell fate transition of VC towards SC, we experimentally demonstrated that the micromitophagy is triggered by MG cell fate determination.\u003c/p\u003e \u003cp\u003eMechanistic understanding of micromitophagy is in fact still insufficient in animal and remains very poor in yeast after several decades of studies, while entirely lacking in plants. In mammalian cells, micromitophagy, defined as lysosomal degradation of mitochondrial-derived vesicles, is independent of autophagosomes and key autophagy regulators, such as ATG5\u003csup\u003e\u003cb\u003e24,37\u003c/b\u003e\u003c/sup\u003e. Whereas, the micromitophagy induced by nitrogen starvation in the lactate-grown \u003cem\u003eS. cerevisiae\u003c/em\u003e and the microchlorophagy induced by UV-B damage in \u003cem\u003eA. thaliana\u003c/em\u003e suggested that the microauthophagy process of semi-autonomous organelles might involve core ATG proteins\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. As an example, the photo-damaged chloroplasts accumulated in \u003cem\u003eA. thaliana atg5\u003c/em\u003e and \u003cem\u003eatg7\u003c/em\u003e mutants due to aberrant microchlorophagy\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Herein, we found that ATG5 was expressed preferentially in SC and was distributed on the vacuoles of MG cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003ea,b and Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Meanwhile, HSP90.2 is rapidly transferred to the paternal mitochondria in GC right after MG cell fate determination (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003ef,g). The ATG5-HSP90.2 interaction mediates the vacuole-mitochondrion recognition and thus promotes PME in the MG cells. In \u003cem\u003eA. thaliana\u003c/em\u003e, HSP90s are widely involved in different processes, e.g., for regulatory protein translocations in developmental and immunity signaling pathways\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e,\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. The voluminous work in HSP90.2 will provide useful clues for future mechanistic investigation of whether HSP90.2 have evolved a unique function in micromitophagy during plant spermatogenesis. At present, it is unclear how the MG cell fate determination drives the assembly of ATG5 onto the vacuoles and HSP90.2 onto the mitochondria to regulate micromitophagy. It is important to note that even in yeast and mammal cells that have a considerably longer history of investigations, the recognition mechanism of micromitophagy is still unknown. The discovery of ATG5-HSP90.2 interaction bridges this knowledge gap for understanding the regulating mechanism of micromitophagy, providing a novel approach for research in future.\u003c/p\u003e \u003cp\u003eThe microspore undergoes just one asymmetric cell division to generate a VC and a GC (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea), so achieving paternal micromitophagy must be a rapid and unique developmental event in MG cells. The core ATG proteins, including ATG5, are wildly known as key regulators of macroautophagy and was critical for VC development\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Although ATGs are universally expressed in all plant organs, we have previously discovered that macroautophagy can act compartmentally at a specific position in the VC for a specific developmental event, namely at the pollen germinal aperture for pollen tube extrusion, thus to support pollen germination\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Together with establishing microautophagy for PME during spermatogenesis here, our works evidence that macroautophagy and microautophagy have co-evolved in two different cell types of pollen, the VC and MG cells, to carry out the distinct functions of pollen germination and PME, respectively. Together, our studies in pollen germination\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e and MG cells here established that ATG5 plays distinct roles in two sister cell types in different developmental events. They also suggest that ATG5 regulating the broadly utilized macroautophagy in the VC to permit pollen germination\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e has undergone neo-functionalization during evolution to function in the MG cell-specific microautophagy-mediated PME (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In illustrating the occurrence of ATG5 neofunctionalization in a specific cell type for a specific developmental issue, like PME, our findings establish cell type specific neofunctionalization of ATG proteins, previously unknown not only in plants but also in animals or yeast. The discovery of ATG5 neofunctionalization here also reveals an adaptive evolution of the core ATG genes or different autophagy pathways for specific developmental events. In fact, neofunctionalization of existing proteins for new functions should be considerably more efficient and resource-economical relative to establishing a totally new group of proteins within a very narrow developmental window, such as after the GC formation to SC maturation. Given the importance of autophagy in developmental processes, the discovery of cell type-specific and functional diversification of core ATGs opens an unexplored territory of autophagy investigation, which obviously could only be addressed in multicellular organisms like angiosperms as we have pioneered here. These foundational discoveries should invite in-depth investigations of the mechanism of the cell type-dependent ATG translocation and functions in multicellular organisms, as well as the identification of triggers for ATGs’ functional transition in different cell types.\u003c/p\u003e \u003cp\u003eThe fact that angiosperms have evolved micromitophagy suggests it must be a more fitting strategy to contribute to maternal inheritance while ensuring successful fertilization. The SCs of flowering plants have lost their motility and are passively delivered to the embryo sac for fertilization by a pollen tube produced by the VC. Accordingly, large numbers of mitochondria are no longer required to power autonomous movement when the immotile angiosperm SCs have fully developed. Obviously, eliminating the unnecessary paternal mitochondria before fertilization is the simplest and ideal option to reduce reliance on complex mechanisms of distinguishing paternal or maternal mitochondria in zygotes\u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. However, after SCs have been released from the pollen tube in embryo sac, they still need energy to maintain an active metabolism, not only to stay alive but also to trigger several crucial pre-fertilization events and during gamete fusion\u003csup\u003e\u003cb\u003e\u003cspan additionalcitationids=\"CR42\" citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e–\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Therefore, consistent with this, the paternal mitochondria being greatly deleted but never eliminated completely in the SC before fertilization (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea-c) is clearly an adaptation to fuel the SCs in their last journey to fertilization. In fact, in the long evolution of angiosperms, maintaining minimum mitochondria in SCs and avoiding the damaging impact of excess paternal mitochondria on zygote and early embryo development (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea,d-f) might be optimum for balancing the needs to contribute mitochondrial maternal inheritance and to fuel gamete communication, fusion and the zygote-to-embryo transition.\u003c/p\u003e \n\u003ch3\u003eOnline content\u003c/h3\u003e\n\u003cp\u003eAny methods, additional references, Nature Portfolio reporting summaries, extended data, supplementary information, acknowledgements; details of author contributions and competing interests are available at______.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003ePlant material and growth conditions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll \u003cem\u003eA. thaliana\u0026nbsp;\u003c/em\u003elines including the \u003cem\u003eatg5-1\u003c/em\u003e\u003cstrong\u003e\u003csup\u003e20\u003c/sup\u003e\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eand \u003cem\u003ehsp90.2-2\u003c/em\u003e\u003cstrong\u003e\u003csup\u003e27\u003c/sup\u003e\u003c/strong\u003e mutants were in Columbia background, and\u003cem\u003e\u0026nbsp;\u003c/em\u003ewere grown in a greenhouse under long-day conditions (16h light/8h dark) at 22\u0026deg;C.\u0026nbsp;\u003cem\u003eO. sativa\u003c/em\u003e plants used in this study were of the Zhonghua 11 (ZH11) cultivar of the \u003cem\u003ejaponica\u003c/em\u003e group, and were grown in a glasshouse under the diurnal cycle of photoperiod and temperature (30\u0026deg;C : 22\u0026deg;C, 12 h : 12 h, light : dark). \u003cem\u003eN. tabacum\u003c/em\u003e L. cv. Petite Havana SR1 and \u003cem\u003eNtatg5\u003c/em\u003e\u003cstrong\u003e\u003csup\u003e20\u003c/sup\u003e\u003c/strong\u003e plants were grown in a glasshouse under 16h light/8h dark at 25\u0026deg;C.\u003cem\u003e\u0026nbsp;L. longiflorum\u003c/em\u003e plants were grown in the field of Wuhan University, in Wuhan, China.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConstructs and plant transformation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe 2435bp fragment including \u003cem\u003eTIP5;1\u003c/em\u003e promoter and coding sequence was amplified by PCR to introduce SacI and KpnI flanking restriction sites. The SacI-KpnI-digested\u003cem\u003e\u0026nbsp;\u003c/em\u003efragment was then cloned into the SacI-KpnI-digested P182 vector (modified from the PART27 vector, carrying a RFP gene and a kanamycin-resistance gene). The \u003cem\u003eDUO1\u003c/em\u003e promoter was amplified as described previously\u003cstrong\u003e\u003csup\u003e44\u003c/sup\u003e\u003c/strong\u003e, and then cloned into the SacI-KpnI-digested Q108 vector (modified from the PART27 vector, carrying a GFP gene and a Hygromycin-resistance gene). Subsequently, a 129-bp DNA fragment containing the mitochondria-targeted pre-sequence of the located F1-ATPase gene \u003cem\u003eAt5g13450\u003c/em\u003e was amplified and cloned to generate the\u0026nbsp;\u003cem\u003ePro\u003c/em\u003e\u003cem\u003eDUO1\u003c/em\u003e\u003cem\u003e:Mito-GFP\u003c/em\u003e reporter.\u0026nbsp;The constructs were transferred into \u003cem\u003eAgrobacterium tumefaciens\u003c/em\u003e strain GV3101, which was used to transform\u0026nbsp;\u003cem\u003eA. thaliana\u003c/em\u003e plants by the floral dip method\u003cstrong\u003e\u003csup\u003e45\u003c/sup\u003e\u003c/strong\u003e. For selection, the medium was supplemented with 50 \u0026mu;g/ml\u0026nbsp;hygromycin\u0026nbsp;or\u0026nbsp;kanamycin.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMitochondrial dynamic in male germline cells\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor \u003cem\u003eA. thaliana\u003c/em\u003e, the pollen grains from \u003cem\u003eProTIP5;1: TIP5;1-RFP ProDUO1:Mito-GFP\u003c/em\u003e line (two markers line) and \u003cem\u003eatg5-1\u003c/em\u003e \u003cem\u003eProTIP5;1: TIP5;1-RFP ProDUO1:Mito-GFP\u003c/em\u003e line at different developmental stage were dissected rapidly and then examined using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems). To visualize SC within pollen tube, the \u003cem\u003ein vivo\u0026ndash;in vitro\u003c/em\u003e pollen growth assay was performed according to the previous method\u003cstrong\u003e\u003csup\u003e16,\u003c/sup\u003e\u003c/strong\u003e\u003cstrong\u003e\u003csup\u003e46\u003c/sup\u003e\u003c/strong\u003e.\u0026nbsp;The styles pollinated by\u0026nbsp;the mature pollen from the aforesaid two lines were incubated on pollen germination medium (PGM:\u0026nbsp;1mM CaCl\u003csub\u003e2\u003c/sub\u003e\u0026middot;2H\u003csub\u003e2\u003c/sub\u003eO,\u0026nbsp;1mM Ca(NO\u003csub\u003e3\u003c/sub\u003e)\u003csub\u003e2\u003c/sub\u003e\u0026middot;4H\u003csub\u003e2\u003c/sub\u003eO,\u0026nbsp;1mM MgSO\u003csub\u003e4\u003c/sub\u003e\u0026middot;7H\u003csub\u003e2\u003c/sub\u003eO,\u0026nbsp;1% H\u003csub\u003e3\u003c/sub\u003eBO\u003csub\u003e3\u003c/sub\u003e,\u0026nbsp;18% sucrose, pH7.0) at 22\u0026deg;C for 8 hours. The samples were placed in the same culture dish to\u0026nbsp;guarantee the consistency of experimental conditions. For autophagy inhibition experiments, the pollen grains from\u0026nbsp;two markers line\u0026nbsp;germinated in the PGM containing 0mM, 1mM, 2.5mM and 5mM 3-MA (Sigma, M9281). With pollen germination, SC pair migrated into the pollen tube. After 5 hours of treatment, the mitochondrial dynamics of SCs treated with different concentrations of 3-MA were observed by\u0026nbsp;Leica SP8 CLSM.\u003c/p\u003e\n\u003cp\u003eThe inverted confocal microscope system (Zeiss LSM 980) was used for live imaging of the micromitophagy process in SCs. The mature pollen grains from the two markers line were incubated on PGM. The microscope was mounted on a Zeiss Axio Observer Z1 basic stand equipped with an incubator (XLmulti S1) to maintain the temperature at 22 \u0026deg;C. Time-lapse images were acquired every 30 seconds for 5 to 10 mins, and image processing was performed with ZEN software.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor \u003cem\u003eN. tabacum\u003c/em\u003e, the late GC and SC were isolated from the growing pollen tubes\u003cstrong\u003e\u003csup\u003e47\u003c/sup\u003e\u003c/strong\u003e, and they were stained with 100nM Mitotracker Green (Invitrogen, M7512), and then the mitochondria were observed with a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems). For every experiment, unless otherwise indicated in the figure legends, the images are representative of at least three independent biological experiments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTransmission electron microscopy (TEM) and electron tomography\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe anthers for\u0026nbsp;\u003cem\u003eA. thaliana\u003c/em\u003e and the pollen grains for\u0026nbsp;\u003cem\u003eO. sativa\u003c/em\u003e, \u003cem\u003eN. tabacum\u003c/em\u003e and\u003cem\u003e\u0026nbsp;L. longiflorum\u003c/em\u003e at different developmental stage were dissected carefully in high-pressure freezer specimen holder. Then, they were embedded with the external cryo-protectant (2% low melting point agarose, Sigma, A9539), and any air bubbles must be removed to avoid disrupting the tissues. Samples were rapidly freeze fixed using the Lene and stoica EM ICE and stored in liquid nitrogen. Fixative,\u0026nbsp;freeze\u0026nbsp;substitution and Spurr\u0026nbsp;resin\u0026nbsp;embedding of samples and Ultra-thin section preparation were executed from a procedure described previously\u003cstrong\u003e\u003csup\u003e16\u003c/sup\u003e\u003c/strong\u003e. The\u0026nbsp;Ultra-thin sections (70 nm) were\u0026nbsp;examined with a JEM-1400 plus transmission electron microscope (JEOL) at 100 kV.\u003c/p\u003e\n\u003cp\u003eFor electron\u0026nbsp;tomography analysis, the sections (100 nm) were stained with 2%\u0026nbsp;(w/v)\u0026nbsp;uranyl acetate followed by\u0026nbsp;2.6% (w/v)\u0026nbsp;lead citrate.\u0026nbsp;Tilting\u0026nbsp;series were collected from +60\u0026deg; to -60\u0026deg; (1\u0026deg; intervals)\u0026nbsp;using a\u0026nbsp;Talos L120C G2\u0026nbsp;electron microscope (Thermo Fisher). Tomograms were reconstructed\u0026nbsp;and\u0026nbsp;3D models were generated using the Inspect3D and Amira software package.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImmuno-electron microscopy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe anther of\u0026nbsp;\u003cem\u003eA. thaliana\u0026nbsp;\u003c/em\u003ewas fixed with 4% (w/v) paraformaldehyde for 16 h at 4\u0026deg;C. After dehydration through an ethanol series (30%\u0026rarr;50%\u0026rarr;60%\u0026rarr;70%\u0026rarr;80%\u0026rarr;90%\u0026rarr;100% three times; 4 h per step), the samples were embedded in K4M resin. The Ultra-thin sections\u0026nbsp;(70 nm)\u0026nbsp;were prepared and then the blocking (PBS, 0.05% Triton X-100, 0.05% Tween-20, 10g/L BSA, RT, 5min) were performed. Following the pretreatment, the sections were incubated with primary antibody (anti-ATG5, 1:100, Agrisera, AS153060) for 2h at RT and 24h at 4\u0026deg;C. After washed six times in PBS for 2min each time, the sections were incubated with secondary antibody (Anti-Rabbit IgG (whole molecule)\u0026ndash;Gold antibody, 10nm, 1:200, Sigma, G3779) for 1h at RT. After washed six times in PBS and twice in water for 2min each time, the\u0026nbsp;sections\u0026nbsp;were stained with 2% (w/v) uranyl acetate and 2.6% (w/v) lead citrate aqueous solution. The samples were\u0026nbsp;examined with a JEM-1400 plus transmission electron microscope (JEOL) at 100 kV.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; For double immunolabelling,\u0026nbsp;the sections were incubated with primary antibody (anti-ATG5, 1:100, Agrisera, AS153060) and secondary antibody (Anti-Rabbit IgG (whole molecule)\u0026ndash;Gold antibody, 5nm, 1:200, Sigma, G7277) following the steps mentioned above. Then, the sections were fixed with 2% (w/v) paraformaldehyde and 0.01% (v/v) glutaraldehyde for 10 min. After washed two times in PBS for 2min each time, the sections were blocked again. Continuedly, the sections were incubated with primary antibody (anti-RFP, 1:100, Invitrogen, MA5-15257) and secondary antibody (Anti-mouse IgG (whole molecule)\u0026ndash;Gold antibody, 10nm, 1:200, Sigma, G7652) following the steps mentioned above. Finally, the\u0026nbsp;sections\u0026nbsp;were stained with 2% (w/v) uranyl acetate and 2.6% (w/v) lead citrate aqueous solution, and then\u0026nbsp;examined with a JEM-1400 plus transmission electron microscope (JEOL) at 100 kV.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eYeast two-hybrid (Y2H) assay\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Y2H assay was performed following the manufacturer\u0026rsquo;s instructions (The Yeast Two-Hybrid System Kit (Clontech), http://www.clontech.com/) and used to test the protein\u0026ndash;protein interaction. The full-length coding sequences of the genes for ATG5 and HSP90.2 were cloned into pGBKT7 vector or pGADT7 vector, and fused with the GAL4 DNA binding domain (DBD) and the activating domain, respectively. The different plasmid combinations were transformed into the yeast strain AH109. The transformants were selected on SD/-Leu/-Trp culture medium, and then interactions were tested on SD/-Leu/-Trp/-His and SD/-Leu/-Trp/-His/-Ade culture medium.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBimolecular fluorescence complementation\u0026nbsp;(BiFC) assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor BiFC experiment, the full-length coding sequences of the genes for ATG5 and HSP90.2 were inserted into the pUb-nYFP and pUb-cYFP vectors, respectively. Then, the different plasmids were transformed into \u003cem\u003eAgrobacterium tumefaciens\u003c/em\u003e strain GV3101. We infiltrated fully expanded leaves of \u003cem\u003eNicotiana benthamiana\u003c/em\u003e with transformed \u003cem\u003eA.\u0026nbsp;tumefaciens\u003c/em\u003e strains containing constructs at the density OD\u003csub\u003e600\u003c/sub\u003e = 0.2. After 48 hours, the epidermal cells of infiltrated leaves were examined and images were collected using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCo-immunoprecipitation (Co-IP) assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor the Co-IP assay, the constructs\u0026nbsp;\u003cem\u003ePro35S: ATG5-mCherry\u003c/em\u003e,\u0026nbsp;\u003cem\u003ePro35S: HSP90.2-GFP\u0026nbsp;\u003c/em\u003eand\u003cem\u003e\u0026nbsp;Pro35S: GFP\u0026nbsp;\u003c/em\u003ewere transformed into \u003cem\u003eA. tumefaciens\u003c/em\u003e strain GV3101, and then two groups of mixed bacterial solution (\u003cem\u003ePro35S: ATG5-mCherry/Pro35S: HSP90.2-GFP\u0026nbsp;\u003c/em\u003eand\u003cem\u003e\u0026nbsp;Pro35S: ATG5-mCherry\u003c/em\u003e/\u003cem\u003ePro35S: GFP\u003c/em\u003e) infected \u003cem\u003eN. benthamiana\u003c/em\u003e leaves and caused the co-expression of ATG5-mCherry with HSP90.2-GFP or GFP, respectively\u003cem\u003e.\u0026nbsp;\u003c/em\u003eAfter\u0026nbsp;infiltration for\u0026nbsp;48 hours, the fluorescence signal intensity of mCherry and GFP and the expression levels of fusion proteins in two groups of tobacco leaves were evaluated using\u0026nbsp;Leica SP8 CLSM. Subsequently, the leaves were fully ground in liquid nitrogen and the proteins were extracted using the extraction\u0026nbsp;buffer (50mM Tris\u0026ndash;HCl, pH 7.5, 150mM NaCl, 1mM EDTA, 1% (v/v) Triton X-100, 1mM PMSF and 1\u0026times; protease inhibitor cocktail).\u0026nbsp;Immunoprecipitation was performed using the ChromoTek GFP-Trap\u0026reg; Magnetic Agarose for overnight with gentle rotation at 4\u0026thinsp;\u0026deg;C. The eluted protein samples together with Input were separated by SDS\u0026ndash;PAGE (ACE Biotechnology, FuturePAGE\u003csup\u003eTM\u003c/sup\u003e 4-20%), and then detected with anti-GFP antibody (Abcam, ab13970) at 1/4,000 dilution or anti-mCherry antibody (Abcam, ab125096) at 1/2,000 dilution. The western blots were acquired using a Amersham Imager (GE, 680RGB).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eZygote phenotypic analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe healthy pistils of \u003cem\u003eProDD45:GFP\u003c/em\u003e marker line at same developmental state were hand-pollinated by the pollen grains from WT and \u003cem\u003eatg5-1\u003c/em\u003e, respectively. It\u0026apos;s important to note that pollination should be done in 10 minutes or less. 20 hours after pollination, all pistils are picked synchronously and placed in a wet box. Immediately, the ovules in the anterior half of the pistils were dissected and the developmental stage of the zygote was observed using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTobacco embryo sac isolation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor tobacco embryo sac isolation to observe zygote and pro-embyro development, the 108 HAP hybrid seeds were firstly treated in the enzyme solution (1% cellulase, 0.8% macerozyme, and 0.058% MES in 11% mannitol, pH 5.8) for 30\u0026thinsp;min in dark. Then, the enzyme solution was removed, and seeds were washed twice with 11% mannitol solution (including 0.058% MES, pH 5.8). The suspension of the enzymolytic seeds was gently grinded by a flat-headed glass rod on slide. The releasing embryo sacs with the living zygote or pro-embryo were observed using a Leica SP8 confocal laser scanning microscope (CLSM, Leica Microsystems)\u003cstrong\u003e\u003csup\u003e48,49\u003c/sup\u003e\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOvule whole-mount clearing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe ovule clearing was performed from a procedure described previously\u003cstrong\u003e\u003csup\u003e49\u003c/sup\u003e\u003c/strong\u003e. Briefly, the whole ovules located in the middle of the 144 HAP hybrid ovaries were collected in a 2ml-centrifuge tube with fixative (50% methanol and 10% acetic acid) and then fixed at 4\u0026deg;C for 12 hours. Subsequently, the fixed ovules was treated by 1% SDS and 0.2M NaOH at room temperature (RT) for overnight. Thereafter, the samples were incubated successively as follows: 2.5% NaClO for 1 hour at RT, 1% periodic acid for 1 hour at RT, 80% ethanol for 10min at 80\u0026deg;C, and the same fixative for 1 hour at RT, taking care to rinse the ovules with water in between each step. After washing, the samples were treated in 0.1mg/ml propidium iodide (PI) (P4170; Sigma) for 3 hours and then transferred in a chloral hydrate solution (including 4 g chloral hydrate, 1 ml glycerol, and 2 ml water) and kept overnight at RT. Finally, the ovules were stored in Hoyer\u0026rsquo;s solution (200 g chloral hydrate, 30 g gum arabic, 20 g glycerol, and 50 ml water) for about 7 days and further observation using a confocal microscope (Leica SP8 CLSM, Leica Microsystems).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStatistical analyses were performed using Microsoft Excel 2003 and GraphPad Prism software (version 8.0), and significance is determined by t-test, one-way ANOVA or two-way ANOVA, respectively (see figure legends for details).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSequence data from this article can be found in TAIR database (www.arabidopsis.org). High-throughput sequencing data that support this study are available in the NCBI Gene Expression Omnibus (GSE162640). All unique/stable reagents and plasmids with transgene constructs generated in this study are available from the corresponding authors on reasonable request with a completed materials transfer agreement.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Prof. Shi Xiao (Sun Yat-sen University) for providing\u003cem\u003e\u0026nbsp;A. thaliana\u003c/em\u003e \u003cem\u003eatg5-1\u003c/em\u003e (SAIL_129_B07) mutant seed; Prof. Shuhua Yang (China Agricultural University) for providing\u003cem\u003e\u0026nbsp;A. thaliana\u003c/em\u003e \u003cem\u003ehsp90.2-2\u003c/em\u003e mutant seed; Prof. Gary N. Drews (University of Utah) for offering \u003cem\u003epDD45:GFP\u003c/em\u003e marker line; Dr. Danyang Li (the Core Facility, Wuhan University) for electron\u0026nbsp;tomography analysis. This work was supported by the National Natural Science Foundation of China (31991201, 31800265 and\u0026nbsp;32130031).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eX.H. designed and performed the research, analyzed data, and wrote the draft. L.Z., Z.L., W.Z., F.G., T.C., C.S., X.Z., W.W., and H.C. performed the research. A.C. analyzed the data and finalized the manuscript. M.-X.S. designed the research, analyzed data, and wrote and finalized the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003e\u003cspan\u003eBirky, CW. Jr. The inheritance of genes in mitochondria and chloroplasts: laws, mechanisms, and models. \u003cem\u003eAnnu Rev Genet.\u003c/em\u003e 35, 125\u0026ndash;48 (2001).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMittelsten Scheid, O. Mendelian and non-Mendelian genetics in model plants. \u003cem\u003ePlant Cell\u003c/em\u003e 34(7), 2455\u0026ndash;2461 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eCorrens, C. Vererbungsversuche mit blass (gelb) gr\u0026uuml;nen und buntbl\u0026auml;ttrigen Sippen bei Mirabilis jalapa, Urtica pilulifera und Lunaria annua. \u003cem\u003eZ Vererbungsl\u003c/em\u003e 1, 291\u0026ndash;329 (1909).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBirky, CW. Jr. Uniparental inheritance of organelle genes. \u003cem\u003eCurr Biol.\u003c/em\u003e 18(16), R692-5 (2008).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNagata, N., Saito, C., Sakai, A., Kuroiwa, H. \u0026amp; Kuroiwa, T. The selective increase or decrease of organellar DNA in generative cells just after pollen mitosis one controls cytoplasmic inheritance. \u003cem\u003ePlanta\u003c/em\u003e 209(1), 53\u0026ndash;65(1999).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eWang, D.Y., Zhang, Q., Liu, Y., Lin, Z.F., Zhang, S.X., Sun, M-X. \u0026amp; Sodmergen. The levels of male gametic mitochondrial DNA are highly regulated in angiosperms with regard to mitochondrial inheritance. \u003cem\u003ePlant Cell\u003c/em\u003e 22(7), 2402\u0026ndash;16 (2010).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMatsushima, R., Tang, L.Y., Zhang, L., Yamada, H., Twell, D. \u0026amp; Sakamoto, W. A conserved, Mg\u0026sup2;+-dependent exonuclease degrades organelle DNA during Arabidopsis pollen development. \u003cem\u003ePlant Cell\u003c/em\u003e 23(4), 1608\u0026ndash;24 (2011).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMa, F., Qi, H., Hu, Y.F., Jiang, Q.R., Zhang, L.G., Xue, P., Yang, F.Q., Wang, R., Ju, Y., Uchida, H., Zhang, Q. \u0026amp; Sodmergen. The Mitochondrial Endonuclease M20 Participates in the Down-Regulation of Mitochondrial DNA in Pollen Cells. \u003cem\u003ePlant Physiol.\u003c/em\u003e 178(4), 1537\u0026ndash;1550 (2018).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eTakami, T., Ohnishi, N., Kurita, Y., Iwamura, S., Ohnishi, M., Kusaba, M., Mimura, T. \u0026amp; Sakamoto, W. Organelle DNA degradation contributes to the efficient use of phosphate in seed plants. \u003cem\u003eNat Plants\u003c/em\u003e 4(12), 1044\u0026ndash;1055 (2018).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eTang, L.Y., Matsushima, R. \u0026amp; Sakamoto, W. Mutations defective in ribonucleotide reductase activity interfere with pollen plastid DNA degradation mediated by DPD1 exonuclease. \u003cem\u003ePlant J.\u003c/em\u003e 70(4), 637\u0026ndash;49 (2012).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMogensen, H.L. Exclusion of male mitochondria and plastids during syngamy in barley as a basis for maternal inheritance. \u003cem\u003eProc Natl Acad Sci U S A.\u003c/em\u003e 85(8), 2594\u0026ndash;7 (1988).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNagata, N. Mechanisms for independent cytoplasmic inheritance of mitochondria and plastids in angiosperms. \u003cem\u003eJ Plant Res.\u003c/em\u003e 123(2), 193\u0026ndash;9 (2010).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eChung, K.P., Gonzalez-Duran, E., Ruf, S., Endries, P. \u0026amp; Bock, R. Control of plastid inheritance by environmental and genetic factors. \u003cem\u003eNat Plants\u003c/em\u003e 9(1), 68\u0026ndash;80 (2023).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLiu, P., Shi, J., Sheng, D., Lu, W., Guo, J., Gao, L., Wang, X., Wu, S., Feng, Y., Dong, D., Huang, X. \u0026amp; Tang, H. Mitopherogenesis, a form of mitochondria- specific ectocytosis, regulates sperm mitochondrial quantity and fertility. \u003cem\u003eNat Cell Biol.\u003c/em\u003e 25(11), 1625\u0026ndash;1636 (2023).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNakatogawa, H., Suzuki, K., Kamada, Y. \u0026amp; Ohsumi, Y. Dynamics and diversity in autophagy mechanisms: lessons from yeast. \u003cem\u003eNat Rev Mol Cell Biol.\u003c/em\u003e 10(7), 458\u0026ndash;67 (2009).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHuang, X. \u0026amp; Sun, M-X. H3K27 methylation regulates the fate of two cell lineages in male gametophytes. \u003cem\u003ePlant Cell\u003c/em\u003e 34(8), 2989\u0026ndash;3005 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMizushima, N. \u0026amp; Komatsu, M. Autophagy: renovation of cells and tissues. \u003cem\u003eCell\u003c/em\u003e 147(4), 728\u0026ndash;41 (2011).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGatica, D., Lahiri, V. \u0026amp; Klionsky, D.J. Cargo recognition and degradation by selective autophagy. \u003cem\u003eNat Cell Biol.\u003c/em\u003e 20(3), 233\u0026ndash;242 (2018).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eYoshimoto, K.\u0026amp; Ohsumi, Y. Unveiling the Molecular Mechanisms of Plant Autophagy-From Autophagosomes to Vacuoles in Plants. \u003cem\u003ePlant Cell Physiol.\u003c/em\u003e 59(7), 1337\u0026ndash;1344 (2018).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eZhao, P., Zhou, X.M., Zhao, L.L., Cheung, A.Y. \u0026amp; Sun, M-X. Autophagy- mediated compartmental cytoplasmic deletion is essential for tobacco pollen germination and male fertility. \u003cem\u003eAutophagy\u003c/em\u003e 16(12), 2180\u0026ndash;2192 (2020).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSchuck, S. Microautophagy-distinct molecular mechanisms handle cargoes of many sizes. \u003cem\u003eJ Cell Sci.\u003c/em\u003e 133(17), jcs246322 (2020).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKi\u0026scaron;\u0026scaron;ov\u0026aacute;, I., Salin, B., Schaeffer, J., Bhatia, S., Manon, S. \u0026amp; Camougrand, N. Selective and non-selective autophagic degradation of mitochondria in yeast. \u003cem\u003eAutophagy\u003c/em\u003e 3(4), 329\u0026ndash;36 (2007).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMijaljica, D., Prescott, M. \u0026amp; Devenish, R.J. Microautophagy in mammalian cells: revisiting a 40-year-old conundrum. \u003cem\u003eAutophagy\u003c/em\u003e 7(7), 673\u0026ndash;82 (2011).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eWang, L., Klionsky, D.J. \u0026amp; Shen, H.M. The emerging mechanisms and functions of microautophagy. \u003cem\u003eNat Rev Mol Cell Biol.\u003c/em\u003e 24(3),186\u0026ndash;203 (2023).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMizushima, N., Yamamoto, A., Hatano, M., Kobayashi, Y., Kabeya, Y., Suzuki, K., Tokuhisa, T., Ohsumi, Y. \u0026amp; Yoshimori, T. Dissection of autophagosome formation using Apg5-deficient mouse embryonic stem cells. \u003cem\u003eJ Cell Biol.\u003c/em\u003e 152(4), 657\u0026ndash;68 (2001).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKang, B.H., Plescia, J., Dohi, T., Rosa, J., Doxsey, S.J. \u0026amp; Altieri, D.C. Regulation of tumor cell mitochondrial homeostasis by an organelle-specific Hsp90 chaperone network. \u003cem\u003eCell\u003c/em\u003e 131(2), 257\u0026ndash;70 (2007).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBao, F., Huang, X., Zhu, C., Zhang, X., Li, X. \u0026amp; Yang, S. Arabidopsis HSP90 protein modulates RPP4-mediated temperature-dependent cell death and defense responses. \u003cem\u003eNew Phytol.\u003c/em\u003e 202(4), 1320\u0026ndash;1334 (2014).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBorg, M., Jacob, Y., Susaki, D., LeBlanc, C., Buend\u0026iacute;a, D., Axelsson, E., Kawashima, T., Voigt, P., Boavida, L., Becker, J., Higashiyama, T., Martienssen, R. \u0026amp; Berger, F. Targeted reprogramming of H3K27me3 resets epigenetic memory in plant paternal chromatin. \u003cem\u003eNat Cell Biol.\u003c/em\u003e 22(6), 621\u0026ndash;629 (2020).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eWilliams, J.H., Taylor, M.L. \u0026amp; O\u0026apos;Meara, B.C. Repeated evolution of tricellular (and bicellular) pollen. \u003cem\u003eAm J Bot.\u003c/em\u003e 101(4), 559\u0026ndash;71 (2014).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSanchez-Vera, V., Kenchappa, C.S., Landberg, K., Bressendorff, S., Schwarzbach, S., Martin, T., Mundy, J., Petersen, M., Thelander, M. \u0026amp; Sundberg, E. Autophagy is required for gamete differentiation in the moss \u003cem\u003ePhyscomitrella patens\u003c/em\u003e. \u003cem\u003eAutophagy\u003c/em\u003e 13(11), 1939\u0026ndash;1951 (2017).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNorizuki, T., Minamino, N., Sato, M., Tsukaya, H. \u0026amp; Ueda, T. Dynamic rearrangement and autophagic degradation of mitochondria during spermiogenesis in the liverwort \u003cem\u003eMarchantia polymorpha\u003c/em\u003e. \u003cem\u003eCell Rep.\u003c/em\u003e 39(11), 110975 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eZhou, X., Zhao, P. \u0026amp; Sun. M-X. Autophagy in sexual plant reproduction: new insights. \u003cem\u003eJ Exp Bot.\u003c/em\u003e 72(22), 7658\u0026ndash;7667 (2021).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eChanoca, A., Kovinich, N., Burkel, B., Stecha, S., Bohorquez-Restrepo A., Ueda T., Eliceiri K.W., Grotewold E. \u0026amp; Otegui M.S. Anthocyanin Vacuolar Inclusions Form by a Microautophagy Mechanism. \u003cem\u003ePlant Cell\u003c/em\u003e 27(9), 2545\u0026ndash;2559 (2015).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eDing, X., Zhang, X., Paez-Valencia, J., McLoughlin, F., Reyes, F.C., Morohashi, K., Grotewold, E., Vierstra R.D. \u0026amp; Otegui M.S. Microautophagy Mediates Vacuolar Delivery of Storage Proteins in Maize Aleurone Cells. \u003cem\u003eFront Plant Sci.\u003c/em\u003e 13, 833612 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eIzumi, M., Ishida, H., Nakamura, S., Hidema, J. Entire Photodamaged Chloroplasts Are Transported to the Central Vacuole by Autophagy. \u003cem\u003ePlant Cell\u003c/em\u003e 29(2), 377\u0026ndash;394 (2017).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNakamura, S., Hidema, J., Sakamoto, W., Ishida, H. \u0026amp; Izumi, M. Selective Elimination of Membrane-Damaged Chloroplasts via Microautophagy. \u003cem\u003ePlant Physiol.\u003c/em\u003e 177(3), 1007\u0026ndash;1026 (2018).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSamakovli, D., Margaritopoulou, T., Prassinos, C., Milioni, D. \u0026amp; Hatzopoulos, P. Brassinosteroid nuclear signaling recruits HSP90 activity. \u003cem\u003eNew Phytol.\u003c/em\u003e 203(3), 743\u0026ndash;57 (2014).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHuang, S., Monaghan, J., Zhong, X., Lin, L., Sun, T., Dong, O.X. \u0026amp; Li, X. HSP90s are required for NLR immune receptor accumulation in Arabidopsis. \u003cem\u003ePlant J.\u003c/em\u003e 79(3), 427\u0026ndash;39 (2014).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eYamamoto, H., Zhang, S. \u0026amp; Mizushima, N. Autophagy genes in biology and disease. \u003cem\u003eNat Rev Genet.\u003c/em\u003e 12, 1\u0026ndash;19 (2023).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSato, M., Sato, K., Tomura, K., Kosako, H. \u0026amp; Sato, K. The autophagy receptor ALLO-1 and the IKKE-1 kinase control clearance of paternal mitochondria in \u003cem\u003eCaenorhabditis elegans\u003c/em\u003e. \u003cem\u003eNat Cell Biol.\u003c/em\u003e 20(1), 81\u0026ndash;91 (2018).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMori, T., Kuroiwa, H., Higashiyama, T. \u0026amp; Kuroiwa, T. GENERATIVE CELL SPECIFIC 1 is essential for angiosperm fertilization. \u003cem\u003eNat Cell Biol.\u003c/em\u003e 8(1), 64\u0026ndash;71 (2006).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSprunck, S., Rademacher, S., Vogler, F., Gheyselinck, J., Grossniklaus, U. \u0026amp; Dresselhaus, T. Egg cell-secreted EC1 triggers sperm cell activation during double fertilization. \u003cem\u003eScience\u003c/em\u003e 338(6110), 1093\u0026ndash;7 (2012).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eWang, W., Xiong, H., Cyprys, P., Malka, R., Flores-Tornero, M., Zhao, P., Peng, X., Sprunck, S. \u0026amp; Sun, M-X. DMP8 and 9 regulate HAP2/GCS1 trafficking for the timely acquisition of sperm fusion competence. \u003cem\u003eProc Natl Acad Sci U S A.\u003c/em\u003e 119(45), e2207608119 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003eRotman, N., Durbarry, A., Wardle, A., Yang, W.C., Chaboud, A., Faure, J.E., Berger, F., Twell, D. (2005). A novel class of MYB factors controls sperm cell formation in plants. Curr Biol. \u003cem\u003e15,\u0026nbsp;\u003c/em\u003e244-248.\u003c/li\u003e\n \u003cli\u003eClough, S.J., Bent, A.F. (1998). Floral dip: a simplified method for Agrobacterium-mediated transformation of \u003cem\u003eArabidopsis thaliana\u003c/em\u003e. Plant J.\u003cem\u003e\u0026nbsp;16(6),\u0026nbsp;\u003c/em\u003e735-43.\u003c/li\u003e\n \u003cli\u003eLi, W., Huang, X., Zou, J., Wu, J., Jiao, H., Peng, X., Sun, M-X. (2020). Three STIGMA AND STYLE STYLISTs Pattern the Fine Architectures of Apical Gynoecium and Are Critical for Male Gametophyte-Pistil Interaction. Curr Biol. \u003cem\u003e30(23)\u003c/em\u003e, 4780-4788.\u003c/li\u003e\n \u003cli\u003eXin, H.P., Peng, X.B., Ning, J., Yan, T.T., Ma, L.G., Sun M-X. (2011). Expressed sequence-tag analysis of tobacco sperm cells reveals a unique transcriptional profile and selective persistence of paternal transcripts after fertilization. Sex Plant Reprod. \u003cem\u003e24(1)\u003c/em\u003e, 37-46.\u003c/li\u003e\n \u003cli\u003eZhao, P., Zhou, X.M., Zhang, L.Y., Wang, W., Ma, L.G., Yang, L.B., Peng, X.B., Bozhkov, P.V., Sun, M-X. (2013). A bipartite molecular module controls cell death activation in the Basal cell lineage of plant embryos. PLoS Biol. \u003cem\u003e11(9)\u003c/em\u003e, e1001655.\u003c/li\u003e\n \u003cli\u003eShi, C., Luo, P., Du, Y.T., Chen, H., Huang, X., Cheng, T.H., Luo, A., Li, H.J., Yang, W.C., Zhao, P., Sun, M-X. (2019). Maternal control of suspensor programmed cell death via gibberellin signaling. Nat Commun. \u003cem\u003e10(1)\u003c/em\u003e, 3484.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6690586/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6690586/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMitochondria are inherited maternally in most plants as a classical paradigm of non-Mendelian inheritance, but the mechanism underlying paternal mitochondrial elimination (PME) remains almost unknown. We report here that angiosperms have evolved a micromitophagy-mediated PME, in which vacuoles directly engulf paternal mitochondria via tonoplast invagination. We show that the micromitophagy occurs specifically in male germline (MG) cells. To gain mechanistic insights, we used a vegetative to germline cell fate transition system to establish that micromitophagy is triggered by MG-cell-fate determination. We evidence that ATG5 is translocated to vacuole upon MG-cell-fate determination and interacts with mitochondrion-located HSP90.2 during mitochondrial engulfment by vacuoles, elucidating a cell-type specific ATG neofunctionalization to mediate micromitophagy. The ingenious mechanism not only contributes to maternal inheritance of plant mitochondria, but also supports zygote-to-embryo transition. We further determine that micromitophagy is conserved in angiosperms but was continually optimized during evolution to ensure a highly efficient PME strategy in MG cells with different properties. These findings bridge a long-standing gap in understanding plant PME with emerging mechanistic knowledge.\u003c/p\u003e","manuscriptTitle":"ATG5-HSP90.2-mediated micromitophagy as cytological basis for maternal inheritance of plant mitochondria","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-04 13:23:58","doi":"10.21203/rs.3.rs-6690586/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-plants","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nplants","sideBox":"Learn more about [Nature Plants](http://www.nature.com/nplants/)","snPcode":"","submissionUrl":"","title":"Nature Plants","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"496616f4-2711-4c14-b888-4573a39d6545","owner":[],"postedDate":"June 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":49454866,"name":"Biological sciences/Plant sciences/Plant reproduction"},{"id":49454867,"name":"Biological sciences/Plant sciences/Plant development"},{"id":49454868,"name":"Biological sciences/Plant sciences/Plant cell biology"}],"tags":[],"updatedAt":"2026-01-22T08:07:02+00:00","versionOfRecord":{"articleIdentity":"rs-6690586","link":"https://doi.org/10.1038/s41477-025-02216-1","journal":{"identity":"nature-plants","isVorOnly":false,"title":"Nature Plants"},"publishedOn":"2026-01-21 05:00:00","publishedOnDateReadable":"January 21st, 2026"},"versionCreatedAt":"2025-06-04 13:23:58","video":"","vorDoi":"10.1038/s41477-025-02216-1","vorDoiUrl":"https://doi.org/10.1038/s41477-025-02216-1","workflowStages":[]},"version":"v1","identity":"rs-6690586","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6690586","identity":"rs-6690586","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}