Chapter 1: Crystallisation

          Hattori and Gouaux faced their most difficult challenge: the notoriously tricky crystallisation of a membrane protein.

They needed crystals of ATP bound P2X4 receptors in their functional state of good enough quality to diffract to atomic resolution. They also needed as high a resolution structure as possible of the apo closed state protein for the best comparisons.

They first tried to produce a crystal from the same zebrafish protein sequence construct used in 2009 this time bound to ATP. However these crystals only diffracted to 7 angstroms resolution.

We have summarised below the techniques they employed to find a solution.

Getting a crystallisation suitable protein:

·    They created a gene library of many different constructs of the P2X4 gene and called them ‘receptor candidates’

·    This library contained P2X4 genes from other organisms and zebra fish constructs that had different C-terminal residue deletions all fused to a GFP gene.

·    Construct genes were expressed via a baculovirus vector/sf9 expression system

·    The functional constructs were identified by their successful insertion into insect cell membranes.

·    All purified constructs underwent Fluorescence detection size exclusion chromatography (FDSEC)

What’s this FDSEC all about?

 A gel filtration chromatography method that

 measures the fluorescence signal of a protein

 as it elutes to produce a fluorescence elution profile

     A large single symmetrical peak it tells us expression is good, 

    proteins have homogenous MWs and folds and the

    protein is stable-all essential to produce a good crystal. 

    We can quickly select the best protein for crystallisation.

Pre-crystallisation screening of membrane proteins by FDSEC as described by 'Fluorescence-Detection Size Exclusion Chromatography for Precrystallisation Screening of Integral Membrane Proteins'-Kawate and Gouaux 2006  

·    From FDSEC they identified the best construct for crystallisation and called it ∆P2X4-C.

·    They also expressed proteins from the original 2009 construct called ∆P2X-B

 A bit more about the ∆P2X4-C construct

The final construct, compared to the WT protein,

 had 28 N terminal and 24 C terminal residues missing.

 Asn78 had been mutated to a lys residue and asn187

 to an arg residue. These mutations minimised glycosylation

 of protein constructs which may have introduced

 heterogeneity into the proteins to be crystallised. 

Voltage clamp experiments showed this construct 

had similar gating to the WT protein hence the

 mutations hadn’t altered any important structural features.

Crystallisation:

·    ATP was added to purified ∆P2X4-C but not to ∆P2X4-B

·    Crystals were obtained for both constructs by the vapour diffusion method.

·    Crystals were harvested and prepared for cryogenic preservation by addition of glycerol

·    Crystals were flash frozen in liquid nitrogen for X-ray diffraction experiments

X-ray experiments:

·    They obtained X-ray diffraction data sets for ∆P2X4-C and ∆P2X4-B as described in the video above. This gives a very simple explanation of the X-ray crystallography process.

·    They used the original ∆P2X4-B 2009 structure as a starting point for interpreting their X-ray data.

·    Successive rounds of refinement/computer modelling were performed to improve the structure.

·    They successfully produced a crystal structure of the ATP-bound receptor to 2.8 angstroms resolution!

·    They also produced an improved apo ∆P2X4-B structure to a higher resolution in which a mistake in the old apo structure had been corrected.

All in all, Hattori and Gouaux were now very well placed to make some important P2X4 related discoveries.