A Simple Truth: Pre-crystallization screening assays
that lead to increased stability also lead to higher standards.

Structural studies of membrane proteins present a number of unique challenges which take a significant investment of time and resources to overcome. At nearly every stage of the membrane protein production pipeline, numerous roadblocks exist including: low expressing constructs, solubilization efficiency, purification of a homogeneous sample, protein stability, among others(1). Due to these issues, there is often limited protein sample available for downstream assays, including crystallization. Therefore, ensuring that the protein sample is well behaved and stable prior to structural studies is important to maximize the output from these experiments.

Pre-crystallization screening assays are designed to determine the stability of a membrane protein prior to structural studies. The overall goal of pre-crystallization screening is to rule out “bad” samples early on in the pipeline, and only push forward well behaved samples.  Additionally, sample conditions (e.g. detergent, buffer, additives) that increase the stability of the protein can also be explored in a high-throughput manner. A well accepted hypothesis is that there is a link between protein stability and crystallizability of the sample, most notably seen with thermostabilized GPCRs crystallizing more readily then wild-type proteins(2).

Here, we’ll highlight some of the commonly used methods for determining the stability of membrane proteins. These pre-crystallization screening techniques are also applicable to other structure determination methods, including cryo-EM, NMR, and XFEL. 

GFP Fusions. By creating a construct where green fluorescent protein (GFP) was added to the C-terminus, membrane protein overexpression in E. coli could be observed(3). The GFP protein will only properly fold (and emit a signal) if the upstream membrane protein is properly inserted into the bilayer. This assay was further expanded to a high-throughput method to screen and optimize expression conditions in both E. coli(4) and S. cerevisiae(5)

Fluorescence Size Exclusion Chromatography (fSEC). Further taking advantage of the GFP fusion construct Kawate and Gouaux introduced fSEC in 2006(6). This method allows for the screening of membrane protein constructs for expression, solubilization efficiency in a panel of detergents, and monodispersity. Furthermore, the ability to screen directly from crude cell lysate dramatically reduces the sample and time requirements for the assay. An example of the utility of fSEC can be seen with the human voltage-gated protein channel (Hv1), where after screening a panel of detergents both Anzergent 3-12 and DDM were shown to be able to solubilize the protein with little aggregation(7). Recently, a new method called FA-SEC (Fluorophore Absorption SEC) was introduced. This modification to fSEC monitors the absorption of GFP at 485 nm instead of the fluorescence allowing for the assay to be performed with standard HPLC equipment.

Size Exclusion Chromatography – Multi-Angle Laser Light Scattering (SEC-MALLS). After purification, there will typically be a concentration of free detergent micelles, which can have a drastic effect on crystallization(8). There exist a number of colormetric methods to determine the concentration of total detergent in a sample; however, these methods typically require a lot of sample, are not applicable to all classes of detergents, and do not report on the amount of free detergent micelles(9). SEC-MALLS utilizes three detectors: ultraviolet (UV), light scattering (LS), and refractive index (RI) positioned after a SEC column(10, 11). The chromatogram from the UV detector measures the concentration of the protein in the sample, the RI detector measures the protein and detergent concentrations in the sample, and the LS detector is used to measure the absolute molecular weight of the protein sample(12). Taken together, all three measurements report on the molecular weight of the membrane protein – detergent complex, the oligomeric state, and quantity of free detergent micelles. There are many ways to reduce the amount of free detergent micelles in a sample prior to crystallization, including the DetEx MiniSpin Columns.

CPM Assay. Dye-based methods typically are unsuitable for measuring thermostability of membrane proteins due to the presence of detergent in the sample. Commonly used dyes will partition into the detergent micelle leading to high levels of background fluorescence. Alexandrov and co-workers introduced the use of a thiol-specific dye, CPM, which is compatible with detergent solubilized membrane proteins for use in a fluorescence based thermal stability assay(13). Upon heating, the protein will unfold exposing buried cysteine residues which then bind to the CPM dye. This assay provides a high-throughput way of screening multiple variables that will increase protein stability, including protein constructs, buffer conditions, additives, and detergents. The CPM assay has been modified to monitor membrane protein stability in lipidic cubic phase (LCP)(14), as well as in nanodiscs(15).

LCP-FRAP. The LCP technique for crystallizing membrane proteins has been highly successful(16). In this method, detergent solubilized membrane proteins are reconstituted into lipid bilayer of the cubic phase. In successful experiments, membrane proteins diffuse within the plane of the lipid bilayer and pack together to form type I crystals(17). The LCP-FRAP (fluorescence recovery after photobleaching) technique is a way to monitor the diffusion of membrane proteins reconstituted into LCP(18,19). When used as a pre-crystallization screening assay, LCP-FRAP allows for the selection of screening conditions (host lipids, protein constructs, etc.) for LCP crystallization experiments that increase the likelihood of crystallization.

Differential Filtration Assay (DFA).  Deposits to the Protein Data Bank for membrane proteins comprise broad range of detergent and detergent-lipid complexes as part of the protein preparations leading the diffraction-quality crystals.  In 2015 alone, 25 different detergents were used in the crystallization of 62 unique membrane protein structures.  Developed by the lab of Michael Wiener at the University of Virginia, the DFA assay is a high-throughput method that reports on the size and stability of membrane proteins after exchange into a panel of 94 different detergents(20).  The DFA Assay is the basis for the Analytic Selector Kit, which requires ~500 µg of protein and can be performed in under three hours. 


  1. Wiener, M. C. (2004) Methods 34(3), 364-372.
  2. Loll, P. J. (2014) Acta Cryst F 70(Pt 12), 1576-1583.
  3. Drew, D., et al. (2001) FEBS Letters 507(2), 220-224.
  4. Drew, D., et al. (2006) Nature Methods 3(4), 303-313.
  5. Drew, D., et al. (2008) Nature Protocols 3(5), 784-798.
  6. Kawate, T. and Gouaux, E. (2006) Structure 14(4), 673-681.
  7. Agharkar, A., et al. (2014) Protein Science 23(8), 1136-1147.
  8. Lin, S.-Y., et al. (2016) PLoS One 11(6), e0157923.
  9. Prince, C. and Jia, Z. (2015) Methods in Enzymology 557, 95-116.
  10. Wen, J., et al (1996) Analytical Biochemistry 240(2), 155-166.
  11. Strop, P. and Brunger, A. T. (2005) Protein Science 14(8), 2207 – 2211.
  12. Moraes, I. and Archer, M. (2015) Methods in Molecular Biology 1261, 211-230.
  13. Alexandrov, A. I., et al. (2008) Structure 16(3), 351-359.
  14. Liu, W., et al. (2010) Biophysical Journal 98(8), 1539-1548.
  15. Ashok, Y. and Jaakola, V.-P. (2016) MethodsX 3, 212-218.
  16. Caffrey, M. (2015) Acta Cryst F 71, 3-18.
  17. Landau, E. M. and Rosenbusch, J. P. (1996) PNAS 93(25), 14532-14535.
  18. Cherezov, V., et al. (2008) Crystal Growth and Design 8(12), 4307-4315.
  19. Fenalti, G., et al. (2015) Methods in Enzymology 557, 417-437.
  20. Vergis, J. M., et al. (2010) Analytical Biochemistry 407(1), 1-11.