Research

The following are a list of current research projects in the ASU Structural Biology of Cancer Lab, which are described further below:


Figure 1. Lipidic Cubic Phase (LCP) membrane protein Figure 1. Lipidic Cubic Phase (LCP) membrane protein crystallization. A) Liquid crystals of lipids create tiny cavities that simulate a native membrane-like environment that is ideal for stabilizing and crystallizing membrane proteins. The gel-like material (B) generates high-density crystals as observed by light microscope (C) and SONICC (D).

LCP Tool Development

Membrane proteins are an important class of proteins responsible for cell signaling, transport, cell adhesion and catalyzing chemical transformation. They make up approximately a quarter of all gene products and are the targets for about half of all modern pharmaceutical drugs.  Despite their importance, the number of unique membrane protein structures solved to date is only around 500, as compared to over 110,000 for soluble proteins.  The complex interactions between membrane proteins and the surrounding lipids, as well as their bipolar nature, with both hydrophobic and hydrophilic domains, make it challenging to work with and crystallize membrane proteins for structural interrogation.  In 1996, Landau and Rosenbusch designed a three dimensional matrix made of lipids, water and proteins called lipidic cubic phase (LCP), which facilitated crystallization of bacteriorhodopsin proteins through lateral diffusion of proteins to the nucleation and crystallization pockets (Fig. 1A).  LCP is a transparent mixture with a viscous, gel-like consistency (Fig. 1B) that provides a native membrane-like environment.  LCP generates a high-density of small crystals (Fig. 1C) that leads to a high nucleation rate for growing large, high-quality crystals for analysis by traditional macromolecular crystallography methods using Synchrotron X-ray sources.  LCP also promotes growing nanocrystals for serial femtosecond crystallography (SFX), a technique pioneered by members of the Center for Applied Structural Discovery (CASD), for analysis at cutting-edge X-ray Free Electron Laser (XFEL) sources.  Nanocrystals can be detected using second-order non-linear optical imaging of chiral crystals (SONICC, Fig. 1C) and injected into the XFEL beam while suspended in LCP.

Figure 2. LCP Tool Chest. Set of tools for working with LCP (A-H) and LCP-compatible assays (I-L).

A wide variety of tools have been developed (“LCP tool chest”) for advancing membrane protein crystallization.  These include high-performance mixers for generating LCP matrix (Fig. 2A), LCP dispensers for allocating the proper amount of LCP (Fig. 2B), crystallization robots (Fig. 2C) to generate a wide range of crystallization conditions in 96-well plates (Fig. 2D); developing advanced LCPs using novel lipids (Fig. 2E), monitoring crystal growth (Fig. 2F), generating microcystals for analysis at microfocus beamlines (Fig. 2G), forming and injecting nanocrystals for SFX (Fig. 2H).  A number of common assays can also be performed in situ within LCP including fluorescence recovery after photobleaching (FRAP, Fig 2I), temperatures of dissociation ( Tm, Fig. 2J), SONICC (Fig. 2K), and small angle X-ray scattering (SAXS, Fig. 2L).
 


G-Protein-Coupled Receptors

G-protein coupled receptors (GPCRs) are an important class of signaling membrane proteins mediating our sense of sight, taste and smell, and regulating our mood, the inflammation and immune system as well as the automatic nervous system, including blood pressure.  These receptors have very specific binding pockets for a wide range of molecules within the body, such as hormones, neurotransmitters, lipids, peptides, and light.  However, over 15% of identified GPCRs have no corresponding ligand identified or function determined, known as orphan GPCRs.  All GPCRs contain seven transmembrane domains and are phylogenetically grouped into five main subfamilies: Glutamate, Rhodopsin, Adhesion, Frizzled/Taste2, and Secretin (the GRAFS classification system).  LCP has become increasingly important for crystallizing and studying GPCRs, leading to a rapid increase in the number of solved GPCR structures (Fig. 3, right).  Figure 4 shows a summary of structures solved using LCP, from GPCRs, transporters to photosynthetic proteins, which represent a growing fraction of the total (> 150) and unique structures of membrane proteins (approximately 50) in the protein data bank (PDB).

Figure 4. Membrane protein structures solved in LCP.

Our group uses traditional macromolecular crystallography (MX) to solve GPCR structures, such as the Adeonosine (A2A) receptor (Liu, Science, 2012), at high resolution to reveal intricate details regarding the function of the protein, including the sodium binding site (Fig. 5A-B) and interactions of cholesterols and oleic acid near the ligand binding pocket (Fig. 5C-D).  Macromolecular crystallography presents some limitations, such as requiring well-ordered crystals of 10’s to 1000 microns in size and cryogenic samples to limit radiation damage during X-ray analysis. To overcome these limitations, an increasing amount of our research effort is focused on growing nanocrystals of GPCRs, which can be examined using SFX with XFELs.  This opens the door to many GPCRs that would not form large crystals and allows the dynamics of these proteins to be explored at room temperature (close to physiological conditions).  Our group solved the first GPCR structure, the serotonin receptor 5-HT2B, using SFX in 2013 (Liu, Science, 2013) and has since made significant contributions to the structures of the δ-opioid receptor, as a potential therapeutic alternative to addictive alkaloied opiate analgesics (Fenalti, Nat. Struct. Mol. Biol., 2015), the angiotensin receptor together with leading drug compounds for treating high blood pressure (Zhang, Cell, 2015), and rhodopsin in complex with its main signaling partner arrestin (Kang, Nature, 2015).  GPCRs are also involved in the growth and metastasis (spreading) of tumors in cancer and are the target of 40% of all modern drugs.  Therefore, novel structural information of GPCRs will have a direct impact on human health.  Although orphan GPCRs lack the corresponding ligands, Dr. Liu is looking to identify potential ligands that can bind to these orphan GPCRs to elucidate their natural function and their significance as promising drug targets.
 


Membrane Protein Fusions

Figure 5. Adenosine Receptor (GPCR). High resolution structure revealed the Na+ binding pocket (A, B) and several important interactions near the ligand pocket (C, D).

When exogenous membrane proteins are overexpressed within non-native systems, such as E. coli, they may not be properly transported or inserted into membranes to remain soluble as the concentration increases.  As a result, they often end up within inclusion bodies and are difficult to isolate.  For those proteins that can be isolated, many are not stable outside of their native lipid bilayer during subsequent purification steps.  Membrane proteins that can be purified still face another challenge during crystallization, because of the relative dearth of hydrophilic domains for providing bridges and contacts to adjacent proteins during crystal formation.  A promising approach to improve the expressed membrane protein yields, stability and crystallization are to fuse them with soluble proteins or antibodies, or through point mutagenesis or cleavage targeting problematic domains and residues (Fig. 6).  A significant part of our research entails exploring new construct designs and methods to improve all aspects of protein preparation and crystallization to facilitate biophysical interrogation.

One strategy used in our group is to incorporate the fusion protein within the third intracellular loop to improve solubility and stability (Fig. 7).  Figure 7 shows some examples of fusion protein partners, including T4 lysozyme (T4L), flavodoxin, rubridoxin, T4L fragment, cytochrome b562RIL and xylanse (Chun, Structure, 2012).  T4L, and the corresponding fragment, have been used as protein fusion partners with GPCRs with some success due to their high propensity for forming well-ordered crystals.  Apocytochrome b562RIL is ideally suited as a fusion protein because its flexible outer surface allows it to fit within the GPCR intracellular loop region, while its thermostable core improves the overall stability of the fusion construct.

Figure 6. Protein fusion strategies. Several fusion protein strategies to improve protein expression, purification and crystallization.


Figure 7. GPCR-protein fusion partners and constructs. Illustrates the location for incorporating a number of different protein partners for fusing with the target GPCR.


GPCR Structure-to-Function Pipeline

Structural biology has a dramatic impact on society by providing knowledge of nature’s fundamental building blocks as well as a blueprint for how we can design interventions to improve human life.  Our laboratory research brings together a broad range of tools, skills and expertise to accelerate the structure to function pipeline for GPCR proteins as outlined in Figure 8.  Building on a suite of tools for generating fusion protein constructs and working with LCP, the structure-to-function pipeline utilizes computer-based ligand screening to rapidly identify potential ligands for orphan GPCRs, parallel purification and formulation of ligand-bound GPCR proteins that feeds into high throughput (HT) pre-crystallization assays to optimize the construct design, HT crystallization and imaging systems, X-ray data collection and structure determination, computer-based drug target screening against the crystal structures leading ultimately to structure based drug design, where drug designs are optimized based on the crystal structure of predicted and actual binding of the drugs to the target protein.

Figure 8. GPCR Structure-to-Function Pipeline. Illustration of how the various tools for working with membrane proteins and GPCRs can be coordinated to have a positive impact on human health.

Serial femtosecond crystallography allows us to improve the pipeline further by determining structures under more physiological conditions and on timescales where GPCR dynamics can be observed and assembled into molecular movies.  In this fashion, computer-based drug screening can sample more physiologically relevant conformations for each target protein in order to identify conformations, such as the transition state, which are completely unique to a particular protein so that drugs can be designed with little to no side effects and can be given in lower doses.  Our laboratory is exploring all aspects of this structure-to-function pipeline with the ultimate goal of providing new methods for structure-based-drug design that can have significantly improve the quality of life for patients.