Identification and characterization of small proteins in Staphylococcus aureus

小分子蛋白在细菌生理和毒力中扮演着不同的和重要的角色。在金黄色葡萄球菌中，只含有26个氨基酸(AA)的β-溶血素和20-40个氨基酸的苯酚可溶调素(20-40个氨基酸)可引起各种人体血细胞的膜功能紊乱，直至溶解。由于它们固有的有限的序列信息和它们具有物种特异性的倾向，小蛋白极难预测和实验检测。这就解释了为什么在目前的研究中，这组蛋白质的代表性仍然明显偏低。在最近的一项研究中，我们优化了我们的分析策略，用于质谱分析检测不超过100个氨基酸的蛋白质(SP100)，并能够在金黄色葡萄球菌纽曼体内识别185个SP100。其中69个未被使用的注释所涵盖(NC_009641.1；2013年7月6日)。利用核糖体图谱，已经成功地证明了139个相应的小阅读框(SORF)的翻译活性。在这个项目中，我们的目标是(I)使用监督学习模型和优化的质谱学方法来改进金黄色葡萄球菌中sORF及其产物的全局检测，以及(Ii)我们先前工作中已经确定的选定小蛋白的功能表征。为了改进金黄色葡萄球菌中sORFs的经验和基于证据的注释，我们将同时应用机器学习(基于随机森林)和实验方法(基于MS的标记N端肽的检测)。此外，我们的蛋白质基因组分析流水线Pepper(开源；https://gitlab.com/s.fuchs/pepper))将进一步开发，以处理和集成多个组学结果(MS、RNASeq、RiboSeq)。特别关注的是小蛋白在膜蛋白质组中的比例。通过研究(互补)缺失突变体在不同膜胁迫诱导条件下的基因表达，对我们目前工作中已证实存在的SP100候选基因进行功能鉴定。

Small proteins play diverse and essential roles in bacterial physiology and virulence. In Staphylococcus aureus, delta-hemolysin with only 26 amino acids (aa) and phenol-soluble modulins (20 - 40 aa) induce membrane disorders up to lysis of various human blood cells. Due to their intrinsically limited sequence information and their tendency to be species-specific, small proteins are extremely difficult to predict and experimentally detect. This explains why this group of proteins is still clearly underrepresented in current studies. In a recent study, we optimized our analytical strategies for the mass spectrometric detection of proteins with not more than 100 aa (SP100) and were able to identify 185 SP100 in S. aureus Newman. Of these 69 were not covered by the annotation used (NC_009641.1; 07/06/2013). Using Ribosome profiling, translational activity for 139 of the corresponding small reading frames (sORFs) has been successfully proven. In this project we aim at (i) an improved global detection of sORFs and their products in S. aureus using a supervised learning model and optimized mass spectrometric methods and (ii) a functional characterization of selected small proteins already identified in our previous work. For an improved empirical and evidence-based annotation of sORFs in S. aureus we will apply both machine learning (based on random forest) and experimental methods (MS-based detection of labeled N-terminal peptides). Furthermore, our proteogenomics analysis pipeline Pepper (open-source; https://gitlab.com/s.fuchs/pepper) will be further developed to process and integrate multi-omics results (MS, RNASeq, RiboSeq). A special focus will be on the proportion of small proteins in the membrane proteome. Selected SP100 candidates, whose existence has been demonstrated in our current work, will be functionally characterized by studying gene expression of (complemented) deletion mutants under different membrane stress inducing conditions.