Background

Conserved across herpesviruses, the protein UL37 has multiple functions in the viral lifecycle and is required for efficient viral replication. As a component of viral tegument–a layer sandwiched between the genome-containing capsid and the lipid envelope–UL37 is not only necessary for viral assembly (Desai et al., 2001; Jambunathan et al., 2014), but is also required for efficient capsid trafficking (Leege et al., 2009), neuroinvasion (Richards et al., 2017), and counteraction of the host innate immune response (Liu et al., 2008; Zhao et al., 2016). A better understanding of the nature of its multifunctionality requires detailed knowledge of its structure and biochemical properties, which, in turn requires well-behaved purified protein.

Initial preparations of different UL37 constructs varied in solubility, monodispersity, and yield of purified protein, making protein preparations inconsistent and further characterizations irreproducible. Since buffer optimization is a straightforward approach to addressing these concerns, we developed a 24-condition screen that varies buffering agents, pH, and concentrations of NaCl and glycerol and used it in combination with a Thermofluor assay (Phillips et al., 2011) to identify buffer conditions that maximize thermostability. The N-terminal and C-terminal halves of UL37 had variable average thermal stabilities and optimal buffer conditions (Figure 1, Table 1). Yet in all cases, conditions that increased protein stability also improved protein solubility, ultimately increasing the yields of purified protein, which facilitated downstream biochemical characterization.

The Thermofluor assay has many advantages over circular dichroism (CD), which is also used to monitor protein stability (Greenfield, 2006). First, the CD can only test one condition at a time whereas the Thermofluor assay is high throughout, screening up to 384 conditions simultaneously. Second, CD only estimates the secondary structure, whereas the Thermofluor assay provides an assessment of the tertiary structure. Third, CD signal is sensitive to changes in buffer components whereas essentially any buffer can be used in the Thermofluor assay. The buffer screen described here is a good starting point for characterizing a protein of unknown biochemical properties (stability, solubility) but can be customized or expanded for individual proteins based on available biochemical information. Finally, the Thermofluor assay can also be used to assess the domain organization of larger proteins by determining whether multiple domains of a large protein represent independently folded units, thereby increasing understanding of protein structure and function.