细胞壁是高度复杂的动态结构，不仅决定着植物细胞的大小和形状，而且在保护细胞、调控细胞粘附和细胞增殖等方面都发挥着重要作用。植物细胞壁由纤维素、半纤维素、果胶和结构蛋白等组成。纤维素是细胞壁的主要组成成分之一，由纤维素合酶 (Cellulose synthase，CesA) 在细胞质膜表面合成。CesA在内质网合成后转运到高尔基体中，进一步组装成为纤维素合酶复合体 (Cellulose Synthase Complex, CSC)，最终CSC转运至细胞质膜合成纤维素。目前关于CSC胞内转运过程的调控机制还有待深入研究。

STELLO蛋白 (STL1和STL2) 是被报道参与CSC组装的高尔基体定位蛋白。本研究在拟南芥不同细胞中观察到STL蛋白除了已报道的高尔基体定位外，还定位在一类称为SmaCCs/MASCs的囊泡中。使用纤维素合酶抑制剂isoxaben处理，使细胞质膜上的CSC内吞进入位于细胞周质的SmaCCs/MASCs，未观察到GFP-STL2与这类源于细胞质膜的SmaCCs/MASCs共定位，表明GFP-STL2标记的SmaCCs/MASCs不同于来源于细胞质膜的SmaCCs/MASCs。通过长时程成像观察发现，GFP-STL2标记的SmaCCs/MASCs是由高尔基体局部形变产生管状结构、管状结构断裂产生的。常用的高尔基体荧光标记SYP32-mCherry和GOT1P-RFP不能标记高尔基体形变的现象，表明这种形变只发生在STL和CesA定位的高尔基体特定区域。进一步的研究表明，反面高尔基体网 (Trans-Golgi Network, TGN) 不参与高尔基体形变产生SmaCCs/MASCs的过程。这些结果表明，STL特异标记高尔基体产生的SmaCCs/MASCs，由高尔基体形变产生SmaCCs/MASCs的过程不同于经典的、高尔基体中的大分子物质经TGN的囊泡转运过程，而是一个非经典的物质转运途径。

为了探究高尔基体产生SmaCCs/MASCs过程的调控机制，本研究构建并观察了STL2、微丝和微管三重荧光标记植株。高尔基体形变产生管状结构的过程可以分为三种类型：成功断裂型 (59%)、高尔基体回退型 (27%) 和末端缩回型 (14%)。在成功断裂型中，高尔基体沿着微丝运动，接触到微管时，高尔基体形变产生管状结构，管状结构末端定位于微管，而后高尔基体继续沿微丝运动，管状结构拉长，最终断裂产生SmaCCs/MASCs定位于微管上。在高尔基体回退型中，高尔基体管状结构拉伸到一定长度后，高尔基体反向运动，导致管状结构消失，没有产生SmaCCs/MASCs。在末端缩回型中，高尔基体能形变出管状结构，但管状结构末端不能锚定在微管上，而是缩回到高尔基体，最终管状结构消失，没有产生SmaCCs/MASCs。这些结果表明，高尔基体形变产生SmaCCs/MASCs至少需要满足两个条件：一是在管状结构产生后高尔基体需要持续向一个方向运动；二是管状结构末端需要锚定在微管上。

为了验证这些猜测，首先在高尔基体运动受到显著抑制的肌球蛋白三突变体xi3KO (xi1xi2xik) 和经肌球蛋白抑制剂PBP (Pentabromopseudilin) 处理的细胞中分析高尔基体产生管状结构的过程。高尔基体运动受到抑制时，高尔基体回退型比例相比对照显著增加，提示肌球蛋白驱动的高尔基体运动对于高尔基体管状结构断裂是必需的。已有文献报道，CSI1 (Cellulose Synthase Interactive Protein 1) 是连接微管和CesA的关键蛋白，参与胁迫条件下SmaCCs/MASCs的产生。为了探究CSI1是否参与高尔基体产生SmaCCs/MASCs的过程，构建并观察了GFP-STL2和mCherry-CSI1的双重荧光标记材料。结果表明，只要高尔基体管状结构末端有CSI1存在，该末端就能锚定住；末端没有CSI1，该末端就不能锚定。与此一致的是，在csi1-3突变体中，高尔基体能形变出管状结构，但所有的高尔基体管状结构均不能锚定。这些结果提示，CSI1可能不参与高尔基体形变产生管状结构的过程，但对于管状结构末端锚定在微管上至关重要。

为了探究高尔基体源SmaCCs/MASCs的功能，发现这些SmaCCs/MASCs会分裂产生在细胞质膜上以稳定速率运动的CSC。在stl1 stl2和csi1突变体中，高尔基体产生SmaCCs/MASCs的过程被抑制，高尔基体无法产生SmaCCs/MASCs，细胞质膜上的CSC密度显著降低。这些结果表明，高尔基体产生的SmaCCs/MASCs在转运CSC中发挥重要作用，是CSC在高尔基体和细胞质膜之间的转运站。

综上所述，本研究观察到高尔基体形变产生SmaCCs/MASCs囊泡的过程，结合遗传学和细胞生物学等方法解析了肌球蛋白和CSI1协同调控该过程的机制。这些结果不仅为深入理解CSC的胞内转运过程提供了新的实验证据，也为研究囊泡生成和物质运输机制提供了新的视角和突破口。

The cell wall is a highly complex and dynamic structure that not only determines the size and shape of plant cells but also plays crucial roles in cell protection, cell adhesion, and cell proliferation. The plant cell wall is composed of cellulose, hemicellulose, pectin, structural proteins and other components. Cellulose, one of the main components of the cell wall, is synthesized by cellulose synthase (CesA) on the plasma membrane (PM). After synthesis in the ER, CesA is transported to the Golgi and further assembled into the cellulose synthase complex (CSC). Finally, the CSC is transported to PM to synthesize cellulose. The regulation mechanism of CSC intracellular transport is still under investigation.

STELLO proteins (STL1 and STL2) have been reported to be involved in CSC assembly. In my study, the STL proteins were found to localize not only in the Golgi, as previously reported, but also in a group of small compartments containing CesA known as Small CesA compartments (SmaCCs) or microtubule-associated CesA compartments (MASCs) found in other cell types of Arabidopsis. Treatment with the cellulose synthase inhibitor isoxaben induced the internalization of CSC from the PM into the SmaCCs/MASCs. However, co-localization between GFP-STL2 and the internalized CSC was not observed, indicating that GFP-STL2-labeled SmaCCs/MASCs and the internalized CSC from PM are distinct. Time-lapse imaging revealed that the SmaCCs/MASCs labeled with GFP-STL2 were derived directly from the Golgi through a process known as “membrane tail-stretching”. No similar events were observed for the Golgi marker SYP32-mCherry, suggesting that this membrane tail-stretching process occurs specifically in the Golgi region where STL and CesA are localized. Subsequent studies revealed that the TGN marker SYP61-CFP did not co-localize with GFP-STL2 in either the Golgi tail or the SmaCCs/MASCs compartments. These findings support the presence of a heterogeneous population of SmaCCs/MASCs and highlight the STLs as markers for Golgi-derived SmaCCs/MASCs. Therefore, the Golgi membrane tail-stretching events represent a non-canonical route for the formation of SmaCCs/MASCs.

To investigate the regulatory mechanism of Golgi-derived SmaCCs/MASCs, triple fluorescent seedlings were generated to express microtubule visualization, actin filament, and STL2. The results revealed three types of Golgi membrane tail-stretching: successful groups (59%), Golgi reversal (27%), and tail-end retraction (14%). In successful groups, the Golgi moved along actin filaments, encountered nearby microtubules, and engaged with them. This engagement seemed to anchor the membrane to the microtubule. Further movement of the Golgi along the actin filament resulted in the formation of a membrane tail attached to the microtubule, generating microtubule-associated SmaCCs/MASCs. In Golgi reversal, the Golgi membrane tail stretched to a certain length and reversed its movement, leading to the absence of SmaCCs/MASCs. In tail-end retraction, after the tail-anchoring/membrane-stretching, the membrane detached from the microtubule, resulting in the absence of SmaCCs/MASCs. These results suggest that two conditions must be fulfilled for Golgi membrane tail-stretching: continuous unidirectional movement of the Golgi after the formation of the membrane tail and anchoring of the tail ends to microtubules. To investigate this hypothesis, a myosin xik xi1 xi2 triple-knockout mutant (xi3KO) was used and treated with the myosin protein inhibitor Pentabromopseudilin (PBP). The frequency of Golgi reversal significantly increased, indicating that myosin-dependent Golgi movement was necessary for the stretching and rupture of Golgi membrane tails. CSI1 (Cellulose Synthase Interactive Protein 1) is a crucial protein that links microtubules and CesA and is required for the formation of isoxaben-induced SmaCCs/MASCs. To investigate whether CSI1 also contributes to the formation of Golgi-derived SmaCCs/MASCs, dual fluorescence-labeled materials expressing GFP-STL2 and mCherry-CSI1 were constructed and observed. The results demonstrated that in the presence of CSI1, the Golgi membrane tail ends could be successfully anchored to microtubules through CesAs. Without CSI1, the anchoring of the ends could not occur. Consistently, in csi1-3 mutants, all events resulted in retraction of the membrane tail end. These findings suggest that CSI1 may not be involved in the process of Golgi deformation to generate the membrane tail but is necessary for anchoring the ends of the membrane tail to microtubules.

Time-lapse imaging was utilized to examine the role of Golgi-derived SmaCCs/MASCs by analyzing individual CSC insertion events. Successive splits were observed in one SmaCC/MASC, resulting in two components. The break-up product, lacking GFP-STL2 fluorescence, displayed slow and steady trajectories of active CSCs. In stl1 stl2 and csi1-3 epidermal cells, the Golgi inhibits the process of Golgi-derived SmaCCs/MASCs, resulting in a significant decrease in CSC density on the PM due to the absence of SmaCCs/MASCs production. These data suggest that Golgi-derived SmaCCs/MASCs serve as an intermediate station for the delivery of CSCs from the Golgi to the PM.

In summary, through the use of pharmacological, genetic, and cell biological methods, My study observed the process of Golgi-derived SmaCCs/MASCs and analyzed the co-regulation mechanism of myosin and CSI1. My study offers compelling evidence for a non-canonical delivery route of a significant enzyme complex to the plasma membrane in plant biology. Furthermore, it provides new insights into the mechanisms of vesicle generation and substance transport.

The cell wall is a highly complex and dynamic structure that not only determines the size and shape of plant cells but also plays crucial roles in cell protection, cell adhesion, and cell proliferation. The plant cell wall is composed of cellulose, hemicellulose, pectin, structural proteins and other components. Cellulose, one of the main components of the cell wall, is synthesized by cellulose synthase (CesA) on the plasma membrane (PM). After synthesis in the ER, CesA is transported to the Golgi and further assembled into the cellulose synthase complex (CSC). Finally, the CSC is transported to PM to synthesize cellulose. The regulation mechanism of CSC intracellular transport is still under investigation.

STELLO proteins (STL1 and STL2) have been reported to be involved in CSC assembly. In my study, the STL proteins were found to localize not only in the Golgi, as previously reported, but also in a group of small compartments containing CesA known as Small CesA compartments (SmaCCs) or microtubule-associated CesA compartments (MASCs) found in other cell types of Arabidopsis. Treatment with the cellulose synthase inhibitor isoxaben induced the internalization of CSC from the PM into the SmaCCs/MASCs. However, co-localization between GFP-STL2 and the internalized CSC was not observed, indicating that GFP-STL2-labeled SmaCCs/MASCs and the internalized CSC from PM are distinct. Time-lapse imaging revealed that the SmaCCs/MASCs labeled with GFP-STL2 were derived directly from the Golgi through a process known as “membrane tail-stretching”. No similar events were observed for the Golgi marker SYP32-mCherry, suggesting that this membrane tail-stretching process occurs specifically in the Golgi region where STL and CesA are localized. Subsequent studies revealed that the TGN marker SYP61-CFP did not co-localize with GFP-STL2 in either the Golgi tail or the SmaCCs/MASCs compartments. These findings support the presence of a heterogeneous population of SmaCCs/MASCs and highlight the STLs as markers for Golgi-derived SmaCCs/MASCs. Therefore, the Golgi membrane tail-stretching events represent a non-canonical route for the formation of SmaCCs/MASCs.

To investigate the regulatory mechanism of Golgi-derived SmaCCs/MASCs, triple fluorescent seedlings were generated to express microtubule visualization, actin filament, and STL2. The results revealed three types of Golgi membrane tail-stretching: successful groups (59%), Golgi reversal (27%), and tail-end retraction (14%). In successful groups, the Golgi moved along actin filaments, encountered nearby microtubules, and engaged with them. This engagement seemed to anchor the membrane to the microtubule. Further movement of the Golgi along the actin filament resulted in the formation of a membrane tail attached to the microtubule, generating microtubule-associated SmaCCs/MASCs. In Golgi reversal, the Golgi membrane tail stretched to a certain length and reversed its movement, leading to the absence of SmaCCs/MASCs. In tail-end retraction, after the tail-anchoring/membrane-stretching, the membrane detached from the microtubule, resulting in the absence of SmaCCs/MASCs. These results suggest that two conditions must be fulfilled for Golgi membrane tail-stretching: continuous unidirectional movement of the Golgi after the formation of the membrane tail and anchoring of the tail ends to microtubules. To investigate this hypothesis, a myosin xik xi1 xi2 triple-knockout mutant (xi3KO) was used and treated with the myosin protein inhibitor Pentabromopseudilin (PBP). The frequency of Golgi reversal significantly increased, indicating that myosin-dependent Golgi movement was necessary for the stretching and rupture of Golgi membrane tails. CSI1 (Cellulose Synthase Interactive Protein 1) is a crucial protein that links microtubules and CesA and is required for the formation of isoxaben-induced SmaCCs/MASCs. To investigate whether CSI1 also contributes to the formation of Golgi-derived SmaCCs/MASCs, dual fluorescence-labeled materials expressing GFP-STL2 and mCherry-CSI1 were constructed and observed. The results demonstrated that in the presence of CSI1, the Golgi membrane tail ends could be successfully anchored to microtubules through CesAs. Without CSI1, the anchoring of the ends could not occur. Consistently, in csi1-3 mutants, all events resulted in retraction of the membrane tail end. These findings suggest that CSI1 may not be involved in the process of Golgi deformation to generate the membrane tail but is necessary for anchoring the ends of the membrane tail to microtubules.

Time-lapse imaging was utilized to examine the role of Golgi-derived SmaCCs/MASCs by analyzing individual CSC insertion events. Successive splits were observed in one SmaCC/MASC, resulting in two components. The break-up product, lacking GFP-STL2 fluorescence, displayed slow and steady trajectories of active CSCs. In stl1 stl2 and csi1-3 epidermal cells, the Golgi inhibits the process of Golgi-derived SmaCCs/MASCs, resulting in a significant decrease in CSC density on the PM due to the absence of SmaCCs/MASCs production. These data suggest that Golgi-derived SmaCCs/MASCs serve as an intermediate station for the delivery of CSCs from the Golgi to the PM.

In summary, through the use of pharmacological, genetic, and cell biological methods, My study observed the process of Golgi-derived SmaCCs/MASCs and analyzed the co-regulation mechanism of myosin and CSI1. My study offers compelling evidence for a non-canonical delivery route of a significant enzyme complex to the plasma membrane in plant biology. Furthermore, it provides new insights into the mechanisms of vesicle generation and substance transport.