Targeting protein-protein interactions through directed evolution of lanthipeptid

描述(由申请人提供)：据估计，人类基因组中10%至30%的基因可能是疾病治疗的靶点(1)。然而，其他估计表明，只有不到10%的人类蛋白质组可以被传统的小分子药物下药，这两种情况的交集只剩下2%到5%的人类蛋白质与疾病有关和可下药(2)。这些因素突出表明，需要新类别的药物分子，能够扩大目前难以治疗的疾病的可用药靶标的范围。一类通常被认为具有挑战性的靶点是蛋白质-蛋白质相互作用(3)。直觉上，多肽将是破坏特定蛋白质-蛋白质相互作用的理想分子。事实上，筛选多肽库的高通量方法，如噬菌体、核糖体和酵母展示，已经成功地识别了能够在体外扰乱蛋白质-蛋白质相互作用的多肽(4-6)。然而，作为药物分子，多肽具有易蛋白分解和构象灵活性等缺点(7)。不过， 自然界已经进化出了以多肽为基础生产具有生物活性的小分子的机制。核糖体合成和翻译后修饰多肽(RIPP)是一大类天然产品，目前正在研究用于治疗从细菌和寄生虫感染到癌症的各种疾病(8)。研究最多的一类RIPP是含有羊毛硫氨酸或甲基羊毛硫醚交联物的羊硫肽(9)。这些交联物可以提供对蛋白质水解性切割的抵抗力，并赋予羊毛肽(10-13)的构象稳定性。安装这些交联物的一条途径涉及一种一般称为LAMM的双功能酶，它可以使丝氨酸或苏氨酸残基脱水，并催化半胱氨酸残基在这些脱水残基上的迈克尔式加成，从而产生硫醚交联物(9)。研究已经在浮游海洋氰杆菌MIT9313中发现了一种特殊的兰姆酶，称为ProcM，它能够处理29种内源序列不同的羊硫肽前体(14)以及来自不同细菌属的羊硫肽前体(15)，这表明它对其底物具有相当的耐受性。由于羊硫肽是遗传编码的，它们有助于通过组合DNA文库合成来方便地合成大型文库。这项建议将集中在开发羊毛肽的酵母展示系统，设计能够被ProcM环化的多肽库，以及利用荧光激活细胞分类(FACS)破坏羊毛肽的进化，以破坏与人类疾病有关的特定蛋白质-蛋白质相互作用。

DESCRIPTION (provided by applicant): It has been estimated that between 10 and 30 percent of genes in the human genome may be targets for the treatment of disease (1). However, other estimates suggest that less than 10 percent of the human proteome is druggable by conventional small molecule drugs and that the intersection of these two conditions leaves as little as 2 to 5 percent of human proteins both involved in disease and druggable (2). These factors highlight the need for new classes of drug molecules capable of expanding the scope of druggable targets for the treatment of currently intractable diseases. One class of targets that is generally considered challenging to perturb is protein- protein interactions (3). Intuitively, peptides would be ideal molecules for disrupting specific protein- protein interactions. Indeed, high throughput methods of screening peptide libraries, such as phage, ribosome, and yeast display, have been successful at identifying peptides that can disrupt protein- protein interactions in vitro (4-6). However, as drug molecules peptides have drawbacks including susceptibility to proteolysis and conformational flexibility (7). Nevertheless, nature has evolved mechanisms for the production of biologically active small molecules based on peptides. Ribosomally synthesized and post-translationally modified peptides (RiPPs) are a large class of natural products that are currently being investigated for the treatment of conditions ranging from bacterial and parasite infections to cancer (8). The most studied class of RiPPs is the lanthipeptides, which contain lanthionine or methyllanthionine thioether crosslinks (9). These crosslinks can provide resistance to proteolytic cleavage and confer conformational stability on lanthipeptides (10-13). One route for the installation of these crosslinks involves a bifunctional enzyme generically named LanM, which can dehydrate serine or threonine residues and catalyze the Michael-type addition of a cysteine residue onto these dehydrated residues to produce the thioether crosslink (9). Studies have identified a particular LanM enzyme in the planktonic marine cynaobacterium Prochlorococcus MIT9313, called ProcM, which is capable of processing 29 endogenous sequence-diverse lanthipeptide precursors (14) as well as a lanthipeptide precursor from a different genera of bacteria (15), suggesting that it is quite tolerant with respect to its substrate. As lanthipeptides are genetically encoded, they lend themselves to the facile synthesis of large libraries through combinatorial DNA library synthesis. This proposal will focus on the development of a yeast display system for lanthipeptides, the design of peptide libraries that are capable of being cyclized by ProcM, and the evolution of lanthipeptides towards the disruption of specific protein-protein interactions that have been implicated in human disease using fluorescent-activated cell sorting (FACS).