Table 1. Troubleshooting.

	Problem	Possible reason	Solutions
9	No colonies on plate	PCR did not work	Check whether DNA is present by running the PCR reactions on a gel. Try using a different polymerase such as pfu. Change annealing temperature and/or change the primers.
		Transformation did not work	Check competency of cells using positive control. Try different amounts of PCR mix in the transformation.
16	No fluorescence in the plasma membrane of the wild type membrane protein	Cells growing poorly	Re-thaw a new batch of cells and ensure they are doubling every 24 hours.
	Very poor expression of the membrane protein	Complex membrane protein	Engineering membrane proteins to express better in mammalian cells is not easy. Try different host cells, different N-terminal fusions or homologues from different species. Co-expression in the presence of an inhibitor/antagonist may also be helpful.
27	Very high counts are seen in every sample	Free radioligand is not being retained on the column, probably due to association with the detergent micelle	Try different gel filtration media. Use a different radioligand.
	Very low counts in every sample	The radioligand has not bound to the detergent-solubilised membrane protein	Ensure that the binding assay works for the membrane protein in membranes and optimise buffer conditions. Use the mildest possible detergent e.g. digitonin. Use the super [+] format. Add further stabilising agents e.g. 30% vol/vol glycerol during solubilisation. Use a higher affinity radioligand. Use a homologous membrane protein that is more stable.
	Measurements are extremely variable (> ± 10% of the mean)	Sample is not passing through the gel filtration column properly	Ensure samples are loaded slowly onto the centre of the columns. Make sure columns are uniform and do not contain cracks. Ensure pipetted volumes are all identical. Ensure centrifuge is properly balanced and runs without vibration.
33	All the values are the same at every temperature 20-70˚C	Detergent incorrect	If all the values are high, then use a harsher detergent. If all the values are near background, then use a milder detergent.
38	None of the mutations are additive with the most thermostable single mutation	The single mutation is stabilising a different conformation of the membrane protein compared to all the other mutations	Ignore the most thermostabilising mutation and combine the other mutations. Alternatively, perform scanning mutagenesis of the membrane protein containing the single mutation to thermostabilise that particular conformation.
42	The final thermostabilised mutant is not stable in NG or OG	Insufficient number of thermostabilising mutations	Repeat Steps 1-42, using the thermostabilised mutant as the starting construct rather than the wild type receptor.