Technical Highlight - September 2009
Short description: The researchers who published the Ebola virus glycoprotein structure explain how to successfully express, purify and crystallize glycoproteins.
In July 2008, Erica Ollmann Saphire's lab at The Scripps Research Institute in La Jolla, California, published in Nature the structure of an Ebola virus glycoprotein bound to an antibody from a survivor of a 1995 Ebola outbreak 1 . This landmark paper was all the more striking because so few glycoprotein structures are available, despite the abundance and importance of these molecules and their importance in pathogenicity.
Obtaining large amounts of stable, homogeneous protein for growing crystals for structural studies can be a frustrating process. But now Ollmann Saphire's team provide details in Nature Protocols 2 of how they routinely express and screen proteins for crystallization, achieving cloning, expression and crystallization of glycoproteins within 4 weeks.
Their advice is to pick a vector that expresses glycoprotein in mammalian cells but also replicates with a high copy number in Escherichia coli because milligrams of DNA are required for transient transfection of the mammalian cell cultures used to express the glycoprotein. The gene encoding the required protein should be linked to a strong mammalian promoter, such as the human cytomegalovirus promoter.
The team also suggest that the expression vector encodes sequence signals that target the protein for secretion via the endoplasmic reticulum, the cell compartments where the protein is correctly folded and oligosaccharide side chains are first added. They use the pDISPLAY expression vector, although various other commercial vectors designed for glycoprotein expression in mammalian cells are available. They also add a tag, such as the hemagglutinin (HA) epitope, to aid protein purification and detection.
For small-scale transfection in the laboratory, Ollmann Saphire's group use the transfection facilitator FuGENE HD, but for large-scale production they use traditional cheap transfection facilitators such as calcium phosphate.
The group expresses glycoproteins in adherent human embryonic kidney (HEK) 293T cells, which are relatively easy to maintain and so can be cultured equally conveniently in individual laboratories or at large-scale structural genomics centers. Secreted protein is detected by ELISA or by western blotting using different antibodies targeted against linear epitopes and conformational epitopes in the protein. The use of both types of epitope is helpful because some conformational epitopes are disrupted by the SDS-PAGE used to separate proteins for western blotting.
Purification of the protein is by anti-HA affinity column in preference to a metal-based affinity-chromatography column, as a number of host-cell contaminants tend to co-purify with the desired protein when the metal column is used.
One difficulty in crystallizing glycoproteins is that the N-linked oligosaccharide side chains vary considerably from molecule to molecule of a given glycoprotein and such a heterogeneous protein mix is not suitable for crystal formation. Ollmann Saphire and colleagues use one of three methods to remove or prevent glycosylation. One approach is to trim off glycans using PNGasF; another is to use an N-acetylglycosaminyltransferase I (GnT I)-deficient HEK293T cell line that results in the incorporation of smaller, more homogeneous glycans; a third approach is to add The other is to add glycan anabolic inhibitors such as swainsonin or kifunensine to the cell culture. The latter two approaches give proteins bearing uniform high-mannose oligosaccharide chains, which can subsequently be removed in vitro by endoglycosidase H.
For crystal screening, Ollmann Saphire's team uses a robotic liquid-handling system that can dispense very small quantities of protein sample. They mainly use Fluidigm's Topaz protein crystallization system, with its microfluidic free-interface diffusion system, but other robotic systems are available. Their system can test 96 different crystallization conditions using only 1.5 microliters of protein sample.
In their studies of the Ebola virus glycoprotein, the team used this approach to screen 140 different constructs of the glycoprotein in complex with seven different antibodies, eventually obtaining a crystal suitable for structural studies — and the subsequent paper in Nature.
J. E. Lee et al. Structure of the Ebola virus glycoprotein bound to an antibody from a human survivor.
Nature 454, 177-182 (2008). doi:10.1038/nature07082
J. E. Lee et al. An efficient platform for screening expression and crystallization of glycoproteins produced in human cells.
Nature Protoc. 4, 592-604 (2009). doi:10.1038/nprot.2009.29