Harris S L, Waters M G
Department of Molecular Biology, Princeton University, New Jersey 08544, USA.
J Cell Biol. 1996 Mar;132(6):985-98. doi: 10.1083/jcb.132.6.985.
To analyze the mechanism of integral membrane protein localization in the early Golgi apparatus of Saccharomyces cerevisiae, we have used Och1p, a cis-Golgi mannosyltransferase. A series of influenza virus hemagglutinin (HA) epitope-tagged fusion proteins was constructed in which invertase is appended to the Golgi-luminal carboxy terminus of full-length Och1p. Several constructs included a Kex2p cleavage site between the Och1p and invertase moieties to monitor transit to the Kex2p-containing TGN. Cells expressing an Och1p-invertase fusion do not secrete invertase, but those expressing an Och1p-Kex2p site-invertase fusion protein secrete high levels of invertase in a Kex2p-dependent manner. The Och1p-Kex2p site-invertase fusion protein is cleaved with a half-time of 5 min, and the process proceeds to completion. Before cleavage the protein receives glycosyl modifications indicative of passage through the medial- and trans-Golgi, therefore cleavage occurs after ordered anterograde transport through the Golgi to the TGN. Transit to distal compartments is not induced by the invertase moiety, since noninvertase fusion constructs encounter the same glycosyltransferases and Kex2p as well. The Och1p-HA moiety, irrespective of whether it is generated by cleavage of the fusion protein in the TGN or synthesized de novo, is degraded with a half-time of about 60 min. Thus, the half-time of degradation is 12-fold longer than the time required to reach the TGN. At steady state, de novo-synthesized and TGN-generated HA epitope-tagged Och1p reside in a compartment with a buoyant density identical to that of wild-type Och1p and distinct from that of the vacuole or the TGN. Finally, och1 null cells that express an Ochlp fusion construct known to rapidly encounter the TGN glycosylate invertase to the same extent as wild-type cells, indicating that they have phenotypically wild-type Och1p activity. These results lead us to propose a model for Och1p-HA localization that involves movement to distal compartments, at least as far as the TGN, followed by retrieval to the cis compartment, presumably by vesicular transport.
为了分析酿酒酵母早期高尔基体中整合膜蛋白定位的机制,我们使用了顺式高尔基体甘露糖基转移酶Och1p。构建了一系列带有流感病毒血凝素(HA)表位标签的融合蛋白,其中将转化酶附加到全长Och1p的高尔基体腔内羧基末端。几个构建体在Och1p和转化酶部分之间包含一个Kex2p切割位点,以监测向含有Kex2p的反式高尔基体网络(TGN)的转运。表达Och1p-转化酶融合蛋白的细胞不分泌转化酶,但表达Och1p-Kex2p位点-转化酶融合蛋白的细胞以Kex2p依赖性方式分泌高水平的转化酶。Och1p-Kex2p位点-转化酶融合蛋白以5分钟的半衰期被切割,并且该过程会进行到完成。在切割之前,该蛋白接受了表明其通过中间高尔基体和反式高尔基体的糖基修饰,因此切割发生在通过高尔基体向TGN的有序顺向转运之后。向远端区室的转运不是由转化酶部分诱导的,因为非转化酶融合构建体也会遇到相同的糖基转移酶和Kex2p。Och1p-HA部分,无论它是由融合蛋白在TGN中的切割产生还是从头合成,都以约60分钟的半衰期被降解。因此,降解的半衰期比到达TGN所需时间长12倍。在稳态下,从头合成的和TGN产生的带有HA表位标签的Och1p存在于一个浮力密度与野生型Och1p相同且与液泡或TGN不同的区室中。最后,表达已知会快速到达TGN的Och1p融合构建体的och1缺失细胞将转化酶糖基化的程度与野生型细胞相同,这表明它们具有表型上野生型的Och1p活性。这些结果使我们提出了一个Och1p-HA定位的模型,该模型涉及向远端区室移动,至少到达TGN,随后可能通过囊泡运输被检索回顺式区室。