Molecular mechanism of the ribosome and functions of translational regulation

为了研究翻译是如何调控来控制基因表达的，实验室主要采用了高通量测序方法，例如mRNA-Seq和核糖体足迹分析与计算分析。我们还在开发使用荧光显微镜和由阵列GFP分子组成的报告(Suntag)对活的酵母和哺乳动物细胞中的单个多聚体进行成像的工具。我们还实现了特殊形式的核糖体图谱，通过足迹40年代，而不是80年代，核糖体或两个80年代核糖体(二体)之间的碰撞。我们已经证明，通常在起始密码子和终止密码子上发现40s足迹，在起始密码子中，40s亚基等待与60S亚基连接，在翻译完成和60S亚基被移除后，在终止密码子中发现40s足迹。我们现在可以通过酵母中的突变来增强这些信号，这些突变减缓了60S在起始密码子上加入或在终止密码子上解离40S亚基的速度。然后，我们使用这种方法来检测非规范开放阅读框架上的隐秘翻译。相比之下，二体足迹检测到两个80年代的核糖体在翻译过程中发生碰撞，导致足迹是单个核糖体保护的两倍。我们的工作揭示了二体形成是常见的，并且发生在随机的和程序化的核糖体停滞事件中。有趣的是，我们还发现二体在终止密码子处形成，当我们抑制翻译完成后移除核糖体的循环过程时，二体可以移动到3‘非编码区。对于清除停滞的核糖体来说，这种移动的二体可能更广泛地发挥着关键作用。 核糖体(二体)之间的碰撞通常被认为是应激的信号，并已知招募核糖体质量控制(RQC)因子，如E3泛素连接酶Hel2/ZNF598(酵母/哺乳动物)，促进它们的分解，触发mRNA衰退，并更广泛地激活应激反应通路。这一途径的缺失会导致有毒多肽的产生和神经变性。然而，核糖体在翻译过程中自然发生碰撞，这引发了一个问题，即细胞能否区分有害的二体和执行某种功能的二体。为了解决这个问题，我们使用二体图谱在全球范围内确定了核糖体碰撞的位置。有趣的是，我们发现所有的二体都被RQC途径识别，并表明丢失Hel2会减少二体的形成。因此，这一途径被广泛用于检测多种类型的双体。我们的工作还揭示了Hel2并不是唯一一个碰撞核糖体的传感器。我们发现，当Hel2丢失时，eIF2α激酶GCN2被激活。此外，我们发现GCN2的传统触发器，即氨基酸饥饿引起的核糖体停滞，也被Hel2检测到。因此，我们已经确定，核糖体碰撞是通过多条压力感应通路来检测的，这些信号的整合可能有助于细胞对有害的核糖体碰撞和功能性的核糖体碰撞做出差异反应。 就像二体形成事件触发RQC途径一样，过早的翻译终止事件也会触发一种被称为无义介导的衰退(NMD)的应激途径，从而导致mRNA的降解。这一途径是由编码突变的异常转录本触发的，突变编码编码序列中的过早终止密码子。然而，众所周知，NMD途径针对许多表面上正常的转录本，因此被认为在基因调控中扮演着额外的角色。为了寻找可能导致正常细胞提前终止翻译的隐蔽翻译事件，我们采用了40S核糖体图谱、RNA-Seq和其他方法。我们在编码序列内部发现了这样的事件，表明泄漏扫描，其中40S核糖体无法找到主要的Aug起始密码子，而是启动了下游的翻译。我们还发现了长的未解码转录异构体(Lutis)翻译的证据。在这些情况下，远上游启动子在正常情况下被用来产生编码上游开放阅读框架(UORF)的长转录本。这些uORF翻译后的过早终止会导致NMD，这表明NMD对于完全沉默这些转录本是重要的。我们现在正在探索环境条件的变化如何改变启动子的使用，使其有利于uORF不被转录的异构体，允许主要ORF被翻译，并防止NMD。 我们的工作还研究了核糖核酸酶L的激活如何影响翻译，核糖核酸酶是先天性免疫反应的一个分支。核糖核酸酶L的激活被认为是通过对转录本的内切导致广泛的基因衰退。由此产生的转录组是如何翻译的尚不清楚。我们对核糖核酸酶L激活的细胞进行了核糖体图谱分析，发现核糖体在3‘非编码区的占有率显著增加。虽然有人认为核糖核酸酶L可以调节翻译终止和再循环过程，从而产生这些效应，但我们的数据与这一假说不一致。相反，我们通常观察到非编码区的翻译增加，包括编码序列中的5‘UTRs和框外ORF。此外，这种作用依赖于RNaseL的切割活性。解释这些结果的一个模型是，核糖体可以启动对mRNA衰变片段的翻译，并翻译编码的ORF。我们计划进一步研究在这个过程中产生的多肽是如何与先天性免疫反应相关的，以及这些翻译事件是否会影响这些片段的稳定性或拟议的功能，例如激活细胞中的dsRNA传感器。 最后，我们还与益宏业实验室合作，揭示了ufm化是一种重要的核糖体修饰，用于ER特异的RQC途径。

To investigate how translation is regulated to control gene expression, the lab has primarily employed high-throughput sequencing methods, such as mRNA-Seq and ribosome footprint profiling with computational analysis. We are also developing tools for imaging single polysomes in living yeast and mammalian cells by using fluorescence microscopy and reporters consisting of arrayed GFP molecules (SunTag). We also have implemented specialized forms of ribosome profiling by footprinting the 40S, as opposed to 80S, ribosome or collisions between two 80S ribosomes (disomes). We have shown that 40S footprints are typically found at start codons, where the 40S subunit waits to join with the 60S subunit, and at stop codons, after translation is complete and the 60S subunit has been removed. We can now enhance these signals with mutations in yeast that slow down 60S joining at start codons or dissociation of 40S subunits at stop codons. We then use this approach this to detect cryptic translation on non-canonical open reading frames. In contrast, disome footprinting detects cases where two 80S ribosomes collide during translation, leading to a footprint that is twice as long as that protected by a single ribosome. Our work has revealed that disome formation is common and occurs both stochastically and at programmed ribosome stalling events. Intriguingly, we also found that disomes form at stop codons and can move into the 3'UTR when we inhibit the recycling process that removes ribosomes after translation is complete. Such moving disomes may be more broadly critical for clearing stalled ribosomes. Collisions between ribosomes (disomes) are generally thought to be a signal of stress and are known to recruit ribosome quality control (RQC) factors, such as the E3 ubiquitin ligase Hel2/ZNF598 (yeast/mammals), that promote their resolution, trigger mRNA decay, and more broadly activate stress response pathways. Loss of this pathway leads to the production of toxic peptides and neurodegeneration. However, ribosomes naturally collide during the course of translation, raising the question of whether the cell can distinguish disomes that are harmful from those that perform a function. To address this, we globally identified sites where ribosomes collide using disome profiling. Intriguingly, we found that all disomes are recognized by the RQC pathway and showed that loss of Hel2 reduces disome formation. This pathway therefore is widely used to detect many classes of disomes. Our work also revealed that Hel2 is not the only sensor of collided ribosomes. We found that the eIF2alpha kinase Gcn2 is activated when Hel2 is lost. Moreover, we found that the traditional trigger of Gcn2, ribosome stalling induced by starvation of amino acids, was also detected by Hel2. We therefore have established that ribosome collisions are detected by multiple stress-sensing pathways and the integration of these signals may help the cell to differentially respond to harmful vs functional ribosome collisions. Much as disome formation events trigger the RQC pathway, premature translation termination events also trigger a stress pathway called nonsense-mediated decay (NMD) that leads to degradation of the mRNA. This pathway is triggered by aberrant transcripts that encode mutations that encode a premature stop codon within a coding sequence. However, the NMD pathway is known to target many apparently normal transcripts and is therefore thought to play additional roles in gene regulation. To search for cryptic translation events that would cause premature translation termination in normal cells, we employed 40S ribosome profiling, RNA-Seq, and other approaches. We found such events internal to coding sequences, indicative of leaky scanning, where the 40S ribosome fails to find the main AUG start codon and instead initiates translation downstream. We also found evidence for translation of long undecoded transcript isoforms (LUTIs). In these cases, far upstream promoters are used under normal conditions to make long transcripts that encode upstream open reading frames (uORFs). Premature termination after translation of these uORFs results in NMD, suggesting that NMD is important for fully silencing these transcripts. We are now exploring how changes in environmental conditions change promoter usage to favor of isoforms where the uORFs are not transcribed, allowing the main ORF be translated and preventing NMD. Our work has also examined how translation is influenced by activation of RNase L, a branch of the innate immune response. RNase L activation is known to cause widespread mRNA decay via endonucleolytic cleavage of transcripts. How the resultant transcriptome is translated is unknown. We used ribosome profiling on RNase L activated cells and found a strong increase in ribosome occupancy in 3'UTR regions. While it was thought that RNase L can modulate the translation termination and recycling processes and therefore give rise to these effects, our data are inconsistent with this hypothesis. Instead, we observe increased translation in non-coding regions generally, including 5'UTRs and out-of-frame ORFs in coding sequences. Moreover, the effects are dependent on the cleavage activity of RNase L. One model to explain these results is that ribosomes can initiate translation on mRNA decay fragments and translate the encoded ORFs. We plan to further investigate how peptides produced from this process are relevant to the innate immune response and whether these translation events affect the stability or proposed functions of these fragments, such as activation of dsRNA sensors in the cell. Finally, we also collaborated with lab of Yihong Ye to reveal that ufmylation is an important ribosome modification that is used for a ER-specific RQC pathway.