1) Lipid Presentation by CD1

Our lab is interested in the structural basis of lipid antigen presentation and recognition in cell-mediated immunity. Toward this goal, we determine the three-dimensional structure of CD1 antigen receptors in complex with different lipids by x-ray crystallography. Similar to peptide presenting MHC class I molecules, cell surface expression of CD1 antigen receptors is a crucial event in initiating the cellular immune response against glycolipid molecules from invading pathogens and in tumor suppression. The T cell receptor (TCR) repertoire can respond to a broad pool of self and foreign antigens presented by CD1 molecules and can trigger killing of the antigen presenting cell, through cytotoxic T lymphocytes (CTLs), or recruit help (T helper cells) from the humoral immune system through production of soluble antibodies.

The human CD1 family (hCD1) is composed of five non-polymorphic major histocompatibility complex (MHC) class I-like glycoproteins (CD1a to CD1e), and can be divided into two main sub groups based on sequence similarity. Group 1 contains CD1a, CD1b, CD1c and CD1e, whereas human and murine (m) CD1d constitute group 2. CD1 is expressed mainly on thymocytes and several bone marrow-derived dendritic cells. CD1a to CD1d binds and present a wide variety of different lipid antigens to T cells that include mycobacterial mycolates, diacylated sulfoglycolipids, polyisoprenoid lipids, phosphoglycerolipids and sphingolipids.

Many of these lipids do not simply bind to CD1 at the cell surface, but are loaded onto the protein in intracellular compartments. A tyrosine-containing hydrophobic motif (YXXZ; where Y is tyrosine, X is any amino acid and Z is a bulky hydrophobic residue) within the cytoplasmic tail is important for intracellular trafficking, with the exception of CD1a. Adaptor protein complexes (AP1, AP2, AP3 and AP4) control intracellular sorting by recognizing and binding this motif, leading to the packaging of CD1 in transport vesicles. While CD1b, CD1c and CD1d traffic into late endosomal compartments or lysosomes, similar to MHC class II, CD1a is mainly expressed on the cell surface and recycled through the early endosomal compartment, analogous to MHC class I. As a result, CD1 samples the lipid content of antigen presenting cells (APC) in order to initiate an immune response by activating T cells that are constantly surveilling APC’s.

A non-self glycolipid ligand, α-galactosylceramide (α-GalCer, from a marine sponge), has been identified as a ligand for both mCD1d and hCD1d and activates a subpopulation of T lymphocytes called NKT cells that express an invariant Vα14 (mouse) and Vα24 (human) TCR chain, paired with a limited number of Vβ chains (Vβ2, Vβ7 and Vβ8.2 in mice) and expressing a C-type lectin called the NK1.1 marker (NK1+ T cells). Four hours after binding to mCD1d-α-GalCer, these NKT cells rapidly secrete both IL-4 and IFN-γ. The pattern of cytokine secretion suggests a crucial role for these T cells in the modulation of the initial phase of the immune response, and a regulatory role in certain autoimmune diseases. Several structurally related antigens with either truncated fatty acid or sphingosine moieties have been synthesized and their potential in modulating the T helper immune response between TH1 and TH2 profiles has been rigorously studied.

To aid in the design of novel chemotherapeutic agents with such immuno-modulatory capacity, we previously determined the crystal structures of various CD1-glycolipid complexes (Figure 1).

Figure 1 Structure of mCD1d with bound sulfatide ligand (yellow). The carbohydrate headgroup is exposed at the cell surface for T cell recognition (top).

For group 1, especially CD1a and CD1b, the lipid backbone is anchored inside the hydrophobic binding grooves (lipid anchoring), whereas, for group 2 CD1d, a precise hydrogen-bonding network positions the polar ligand headgroups in a well-defined orientation at the T cell recognition surface (headgroup positioning, Figure 2).

Figure 2 Hydrogen-bonding network (blue dashed lines) between mCD1d (grey) and short-chain α-GalCer (yellow).

In addition, small, but important, structural changes occur on the surface of CD1d upon binding of the potent invariant NKT cell agonist α-galactosylceramide due to increased polar interaction with the α1 and α2-helices. Surprisingly, the A’ pocket of the α-GalCer (C8 fatty acid) and α-GalAGsl (C14) complex structures contain a “spacer lipid”, a C16 fatty acid that stabilize the hydrophobic binding groove in the absence of groove-filling long chain ligands (C26). The process of loading or exchanging these lipids against endogenously bound spacer lipids is not yet fully understood, however, various lysosomal lipid transfer proteins have been implicated to participate in the direct transfer by binding these lipids in hydrophobic cavities, before transfer to CD1 occurs.

Although the CD1d-glycolipid structures can explain the rapid activation of NKT cells by CD1d-α-GalCer, there is no structural rationale to explain the development of either a TH1 or TH2 phenotype. The stability of the CD1-lipid interaction and the correlated strength of the signal are likely to be one explanation for the observed phenotype, rather than structural changes between the different CD1-lipid complexes. However, without a CD1-glycolipid-TCR ternary complex, precise predictions about ligand recognition and TCR interaction cannot be made and one urgent question still remains:

How is it possible that, on one hand, some TCRs are so specific for their presented T cell epitope, such as several group1 CD1-restricted T cells, while on the other hand, several antigens with differing chemical structures all have the capacity to stimulate the same invariant Vα14 expressing T Cells? It is not yet clear how the model antigen α-GalCer, microbial α-glycuronosylceramides from Sphingomonas, the self-antigen isoglobotrihexosylceramide (iGB3) and possibly mycobacterial phosphatidylinosito-tetramannoside (PIM4) can use their markedly differing glycans to activate the same types of TCRs. Therefore, future efforts for the structural determination of the CD1-glycolipid complexes bound to their respective TCRs is the next major step in elucidating the immunological properties of lipid-recognizing T cells in human diseases.

2) Mycobacterial Polyketide Synthases
Worldwide, 3 billion people are infected with Mycobacterium tuberculosis and every year more than 2 million people die of its related disease tuberculosis (TB). These figures put TB in the unfavorable list of the top major killers, together with AIDS and malaria. In most cases the immune system can contain a mycobacterial infection to over 90% but with secondary infections such as HIV the immune system looses control over the invading pathogen and the risk of developing tuberculosis increases 30-fold. Despite efforts in recent years to eradicate this disease, incidents of TB related deaths are on the rise in industrialized countries. The situation is worsened by emerging, multiple drug resistance (MDR) strains.

The high resistance of mycobacteria to most common antibiotics and chemotherapeutic agents is correlated with the unusually low permeability of its cell wall and the resulting slow uptake of drugs. Some of the most effective anti-mycobacterial drugs are known to inhibit the biosynthesis of the cell wall components. One of these drugs, isoniazid, inhibits InhA, the 2-trans-enoyl-acyl carrier protein reductase of the type II fatty acid synthase (FAS) enzyme system.

M. tuberculosis has five times more genes that are predicted to be involved in fatty acid and lipid metabolism than E. coli. Many of these genes encode enzymes that are thought to be required for biosynthesis of the additional, unusual lipid components of the cell wall, such as mycolic acids, dimycocerosates, phtiocerol, lipoarabinomannan and the polyprenol phosphates.

Our main interest is the structural characterization of polyketide synthases (PKS) (Figure 3). These multimodular enzymes are similar in their reaction mechanism and catalytic domain architecture to eukaryotic type I fatty acid synthases (FAS I) and essential for the synthesis of complex lipids and, therefore, the assembly of the entire cell wall, which makes these enzymes perfect targets for structural studies.

Figure 3 Domain organization of PKS and Typ I FAS.

We are going to apply x-ray crystallographic and biochemical methods to characterize the function and structure of Mycobacterium tuberculosis proteins involved in mycolic acid and lipid biosynthesis. The resulting information will be used to facilitate the design of novel anti-tuberculosis drugs, which will interfere with lipid synthesis, thus impairing the integrity of the mycobacterial cell envelope.