Engineering of glycosyltransferases to obtain glycan binding proteins

大多数人胞外蛋白是由连接到ASN上的N-连接的多糖翻译后修饰的，而O- 连接在丝氨酸/苏氨酸上的葡聚糖。这种多糖控制或微调许多生物过程，包括 细胞生长、分化、细胞黏附和信号传递。因此，糖基化的变化也是相关的。 与哺乳动物的病理生理过程，如肿瘤转移，宿主病原体识别，炎症 等。一个重要的障碍是理解蛋白质-碳水化合物相互作用在人类健康和 疾病是缺乏一种简化的技术来快速和准确地表征任意细胞/中的多糖。 组织系统。碳水化合物结合凝集素通常被用来表征细胞表面的多糖，但 天然植物和动物凝集素的结合特异性和亲和力较差。在以下方面也取得了一些成功 通过工程技术开发新的糖链结合蛋白(GBP)，例如唾液酸酶以识别 含有多糖的唾液酸，但这些试剂通常只与末端残基结合，对 多聚糖骨架。在本提案中，我们描述了从以下内容开始设计Gbps的替代方法 糖基转移酶，尤其是唾液酸基转移酶(ST)家族。我们假设 设计这类酶可能使对较大的糖链结构的特异性检测具有高度的特异性。 在这一点上，STS催化不同的葡聚糖受体的立体和区域特异性唾液酸化，表明他们的 工程可能产生同时识别唾液酸和受体的唾液酸结合蛋白(SiaBP) 底物。因此，SiaBP可能具有独特的结合特异性，这是传统凝集素或 基因工程糖苷酶。为了测试这一概念，在目标1中，我们对三个不同的人进行了蛋白质工程 唾液酸基转移酶产生三个SiaBP，以高亲和力识别特定的碳水化合物表位 专一性。这些包括：识别Neu5Ac(2-3)Gal(β1-3)GalNAcα的ST3Gal-I突变体；ST6Gal-I突变体 识别Neu5Ac(2-6)Gal(β1-4)GlcNAcβ；和ST8Sia3突变体结合多唾液酸。我们会做模特 通过计算对接配体结合酶的结构，并合理设计突变以 提高结合特异性。我们还将进行脱辅酶的分子动力学(MD)模拟和 分析模拟结构以识别低频集体运动(主成分分析)。 分析将使我们能够引入突变，使酶的构象偏向于有利于产品的构象 有约束力的。合理设计的突变体将通过定向进化得到进一步提炼。在目标2中，提纯的SiaBP 将使用带有各种唾液酸聚糖的葡聚糖阵列进行表征。我们还将测试结合 包含或存在缺失的SiaBP到等基因HEK293T克隆和CRISPR-Cas9 KO细胞库 特定的糖类药物。纤维素分析中的这些是对多糖阵列的补充，并提供了以下生物学背景 这些经过改造的蛋白质最终将被使用。该项目的成功完成还将产生一个平台 这一策略可推广到碳水化合物活性酶数据库(CAZy.org)中的其他糖类。

Most human extracellular proteins are post-translationally modified by N-linked glycans attached to Asn, and O- linked glycans attached to Ser/Thr. Such glycans control or fine-tune a number of biological processes including cell growth, differentiation, cell adhesion, and signaling. As a result, changes in glycosylation are also associated with mammalian pathophysiological processes like tumor metastasis, host-pathogen recognition, inflammation etc. An important impediment to understanding the role of protein-carbohydrate interactions in human health and disease is the lack of a streamlined technology to rapidly and accurately characterize glycans in arbitrary cell/ tissue systems. Carbohydrate binding lectins are commonly used to characterize cell-surface glycans, but the binding specificity and affinity of natural plant and animal lectins is poor. There has also been some success in developing novel glycan binding proteins (GBPs) by engineering, for example, sialidases in order to recognize sialic acid containing glycans, but these reagents typically only bind terminal residues with less specificity for the glycan backbone. In this proposal, we describe an alternative approach to engineering GBPs starting with glycosyltransferases (glycoT), particularly with a focus on the sialyltransferase (ST) family. We hypothesize that engineering this class of enzymes may enable specific detection of larger glycan structures with high specificity. In this regard, STs catalyze stereo and regiospecific sialylation of distinct glycan acceptors, suggesting that their engineering may yield sialoglycan binding proteins (SiaBP) recognizing both the sialic acid and the acceptor substrate. Thus, SiaBPs may have unique binding specificity that is not recapitulated by traditional lectins or engineered glycosidases. To test this concept, in Aim 1, we perform protein engineering on three different human sialyltransferases to generate three SiaBPs that recognize specific carbohydrate epitopes with high affinity and specificity. These include: ST3Gal-I mutants to recognize Neu5Ac(2-3)Gal(β1-3)GalNAcα; ST6Gal-I mutants to recognize Neu5Ac(2-6)Gal(β1-4)GlcNAcβ; and ST8Sia3 mutants to engage poly sialic acids. We will model the ligand-bound enzyme structures through computational docking and rationally design the mutations to improve binding specificity. We will also perform molecular dynamics (MD) simulation of apo-enzymes and analyze the simulated structures to identify low frequency, collective motions (principal component analysis). The analysis will enable us to introduce mutations to bias the enzyme conformation to one that favors product binding. The rationally designed mutants will be further refined using directed evolution. In Aim 2, purified SiaBPs will be characterized using glycan arrays that bear various sialoglycans. We will additionally assay the binding of SiaBPs to isogenic HEK293T clones and CRISPR-Cas9 KO cell libraries that either contain or have deletions of specific glycoTs. These in cellulo assays complement the glycan arrays and provide a biological context where the engineered proteins will ultimately be used. Successful completion of the project will also result in a platform strategy that may be extended to other glycoTs in the Carbohydrate-Active enzyme database (CAZy.org).