Exploring the structure and function of key enzymes involved in microbial sphingolipid biosynthesis

Abstract

Sphingolipids (SLs) are a diverse class of lipid molecules derived from the amino acid L-serine and long chain fatty acids (e.g. carbon chain lengths C14-C26). When combined together these building blocks form a so-called “sphingoid”, also known as a long-chain base (LCB). Decades of research into these enigmatic molecules has revealed that SLs are essential components of eukaryotic cell membranes and control many critical cellular functions. The SL biosynthetic pathway begins in all organisms with the Claisen-like, decarboxylative condensation of L-Ser and long-chain fatty acid acyl-CoA thioester (CoASH) substrates (most commonly C16/palmitoyl) to form the intermediate 3-ketodihydrosphingosine (3-KDS). This first key irreversible reaction is catalysed by the pyridoxal 5’-phosphate (PLP)-dependent serine palmitoyltransferase (SPT). The 3-KDS product is then reduced by a NAD(P)H-dependent enzyme (KDS reductase, KDSR) to generate dihydrosphingosine (DHS). Acylation of this molecule with a long-chain fatty acid by a ceramide synthase (CerS) leads to the formation of ceramide (Cer). Downstream, phosphorylation with a phosphate group can occur on the serine-derived head group and lead to the formation of sphingosine-1-phosphate (S-1-P). The balance between Cer and S-1-P concentrations has been proposed to control the cell survival rate in the host system. Alternatively, sugars such as glucose can be added to give the family of glycosphingolipids (GSLs). In contrast to higher order species, research into SL biosynthesis in bacteria has lagged much further behind. However recent studies have revealed a number of important microbes that produce a range of SLs and Cers. Interestingly, SLs from the human microbiota including those from Bacteroides fragilis, a Gram-negative commensal bacterium from the human gut, and from the oral pathogen Porphyromonas gingivalis have recently been shown to play an essential role in host/microbe communication. Of note is that these bacterial SLs and Cers display many similar structural features to their mammalian homologues. However, there appears to be “chemical signatures” distinct to those of bacterial origin. For example, B. fragilis SLs have iso-Me branched acyl chains which are also similar to those found in the round-worm Caenorhabditis elegans. This suggests a PLP-dependent branched-chain amino acid transferase (BCAT/IlvE) transfers iso-Me chains from amino acid precursors such as L-Leu to branched-chain keto acids. To fully understand the role of bacterial SLs and their metabolism detailed investigations of the enzymes, pathways and metabolism are required. In this thesis, studies of different types of the three key enzymes (SPT, IlvE/BCAT and KDSR) involved in the core microbial SL biosynthetic pathway have been presented. Firstly, both recombinant B. fragilis SPT (BfSPT, encoded by gene BF2461) and P. gingivalis (PgSPT, gene PG1780) were expressed in E. coli, purified and studied with protein UV-vis spectrometry, enzyme kinetics, inhibition assays, mass spectrometry and protein crystallization screening. 3-KDS products were detected derived from a range of straight-chain CoA substrates (C14-C18) and amino acids (Gly, L-Ala and L-Ser) produced by BfSPT and PgSPT. Mutagenesis of a conserved loop (PAVAP) in SPT homologues was found to be associated with the catalytic efficiency of PgSPT. Also the presence of a Val353 residue in BfSPT was shown to be essential to allow the enzyme to interact with the C16-CoA substrate. Moreover, data suggested that the position (N- or C- terminus) of the 6His-affinity tag used to purify SPT influenced substrate inhibition by C16-CoA. A hypothetical 3D structural model of the PgSPT PLP:L-Ser external aldimine complex was built in order to explore the active site and residues involved in substrate binding and catalysis. In collaboration with Prof. Mary-Ellen Davey (Florida), the role on SL biosynthesis in P. gingivalis was also explored and found that SLs impact on the way this pathogen interacts with human cells. Secondly, a branched acid transaminase P. gingivalis IlvE (PgIlvE, gene PG1290) was expressed, purified and studied using UV-vis spectrometry to investigate substrate binding and enzyme activity. A multi-enzyme coupled assay for PgIlvE was developed in both the ‘forward’ and ‘reverse’ direction, studied with inhibitors such as L- and D- cycloserine (LCS/DCS), as well as x-ray crystallography. In collaboration with Dr. Jon Marles-Wright (University of Newcastle) the crystal structures of four different forms of PgIlvE; the PLP-bound, internal aldimine form, LCS ring-opened ring-closed form and the PMP form, were obtained. Two residues (F56 and Y188) were identified which played a role in substrate binding and activity. In the final chapter, recombinant KDSR from the yeast Saccharomyces cerevisiae (ScKDSR) was isolated from E. coli. Kinetic parameters for a soluble, truncated form of the enzyme were determined using the substrates KDS and NADPH. The product C18 DHS derived from C18 KDS was detected with MALDI-ToF-MS. Recent studies of human KDSR revealed that patients with point mutations in this enzyme suffered from skin disorders (erythrokeratoderma). To investigate the impact of these mutations, a series of ScKDSR mutant mimics (G176S, Y180F and G263E) were prepared. To date there has been no crystal structure of a KDSR determined but it is a member of the short-chain dehydrogenases/reductases (SDR) superfamily. A homology model of the 3D structure of ScKDSR with three possible NADPH docking positions were constructed, and residues involved in substrate binding and catalysis were suggested. The results in this thesis shed light on the key enzymes involved in the core biosynthetic pathway of microbial SLs and lay down a blueprint for future studies.