Bridging Project 2: Enolase (EN) Superfamily

功能多样化的EN超家族是理解具有保守活性部位结构的同源酶如何催化不同反应的范例[1，2]。这些反应是由保守的部分反应启动的：镁辅助提取羧酸盐底物的a-质子以生成烯二酚。 通过与镁离子配位而稳定的中间体；中间体由反应特定的酸直接生成。活性部位位于1)包含催化基团的(p/a)7P桶结构域和2)包含大多数底物特异性决定因素的覆盖a+p结构域之间的界面上。 镁离子与桶状结构域的第三、第四和第五个p-链末端的三个Conen/ed配体(Asp/Glu)和底物的至少一个羧酸氧配位；扁桃酸消旋酶(MR)的活性位置如图所示。在第二、第六或第七条P链末端的碱基产生通过与镁离子配位而稳定的疏水甘油中间体。第二条、第三条、第六条或第七条P链末端的酸将中间体导向产物。我们现在识别了七个功能分配的亚基，它们具有不同的活性中心基序，即p-链末端的酸/碱催化剂和金属离子配体的身份和位置。更多的将被确定为“新的”职能被分配和结构被确定。 利用SFLD(SuperFamily/Genome Core)[3]开发的Cytoscapisualized序列相似性网络(SSN)分析，可以将超家族归类为家族。 序列按功能进行颜色编码，并用灰色标记未知功能。与NYSGXRC合作，确定了29个利用功能分配的未知成员的结构。大多数序列可以与两个功能不同的亚群联系在一起，称为变锥酸化内酯酶(MLE)和扁桃酸消旋酶(MR)亚群。 已知MLE亚基具有四种功能：环异构化(MLE)、脱水(o-琥珀酸基苯甲酸合成酶，OSBS)、异构化(L-丙氨酸-ID/L-谷氨酸异构体酶， AEE)和消旋(N-琥珀酸基氨基酸外消旋酶，NSAR)。底物上的羧酸根是镁离子的双齿配体。赖氨酸酸碱催化剂位于桶状结构域的第二和第六个p-链的末端。只有15%的成员(共1258名成员；2009年6月1日)具有未知的功能。 MR子组中有九个已知函数： 细菌碳水化合物分解代谢中七个酸性糖脱水酶家族(一个双功能家族)的外消旋和脱水作用。底物是a-OH酸，其中一个羧基氧和a-OH为镁离子提供双齿配体。约65%的成员(总计1310名成员)具有未知的功能。 许多成员具有未知的功能，随着更多微生物基因组的测序，其数量不断扩大。在P01 GM071790的支持下，我们开发和使用了计算方法来预测并实验指定1)N-琥珀酰精氨酸消旋酶(NSAR)功能[4]，2)二肽异构体酶[5]的不同底物特性[5](也未发表)，以及3)不同的半乳糖酸酯 脱水酶(未发表)。在前两个例子中，使用底物连接的AEE(PDB代码1TKK)作为模板进行同源建模，从而从序列预测功能；在最后两个例子中，由NYSGXRC(PDB代码20QY)确定的结构用于硅配体对接。

The functionally diverse EN superfamily is a paradigm for understanding how homologous enzymes with conserved active site architectures catalyze different reactions [1, 2]. The reactions are initiated by a conserved partial reaction: Mg -assisted abstraction of the a-proton of a carboxylate substrate to generate an enediolate intermediate stabilized by coordination to the Mg2+; the intermediate is directed to product by a reaction-specific acid. The abtive sites are located at the interface between 1) a (p/a)7P-barrel domain that contains the catalytic groups; and 2) a capping a+p domain that contains most ofthe substrate specificity determinants. The Mg2+ is coordinated to three consen/ed ligands (Asp/Glu) at the ends ofthe third, fourth, and fifth p-strands ofthe barrel domain and at least one carboxylate oxygen of the substrate; the active site of mandelate racemase (MR) is shown in the figure. A base at the end of the second, sixth, or seventh p-strand generates the enediolate intermediate that is stabilized by coordination to Mg2+. An acid at the end of the second, third, sixth, or seventh p-strand directs the intermediate to product. We now recognize seven functionally assigned subgroups that have different active site motifs, i.e., identities and locations of acid/base catalysts and metal ion ligands at the ends of the p-strands. More will be identified as "new" functions are assigned and structures are determined. The superfamily can be grouped into families with the Cytoscapevisualized sequence similarity network (SSN) analysis developed by the SFLD (Superfamily/Genome Core) [3]. Sequences are color-coded by function, with gray marking unknown function. In collaboration with NYSGXRC, structures were determined for 29 unknown members that leverage assignment of function. Most sequences can be associated with two functionally diverse subgroups designated the muconate lactonizing enzyme (MLE) and mandelate racemase (MR) subgroups. Four functions are known in the MLE subgroup: cycloisomerization (MLE), dehydration (o-succinylbenzoate synthase, OSBS), epimerization (L-Ala-ID/L-Glu epimerase, AEE) and racemization (N-succinylamino acid racemase, NSAR). The carboxylate oxygens of the substrate are bidentate ligands of the Mg2+ ion. Lys acid/base catalysts are located at the ends of the second and sixth p-strands of the barrel domain. Only 15% of the members (1258 total members; June 1, 2009) have unknown functions. Nine functions are known in the MR subgroup: racemization by MR and dehydration by seven families of acid sugar dehydratases in bacterial carbohydrate catabollsm (one bifunctional family). The substrates are a-OH acids, with one carboxylate oxygen and the a-OH providing bidentate ligands for the Mg2+. About 65% of the members (1310 total members) have unknown functions. Many members have unknown functions, with the number expanding as additional microbial genomes are sequenced. With the support of P01 GM071790 we developed and used computational approaches to predict and then experimentally assign 1) the N-succinyl Arg racemase (NSAR) function [4], 2) divergent substrate specificities for dipeptide epimerases [5] (also unpublished), and 3) a divergent galactarate dehydratase (unpublished). In the first two examples, a substrate-liganded AEE (pdb code 1TKK) was used as template for homology modeling, so functions were predicted from sequence; in the last example, a structure determined by NYSGXRC (pdb code 20QY) was used for in silico ligand docking.