Background

The presence of double stranded DNA within the cytosol of a cell is a potential sign of infection by a DNA or retrovirus. The nucleotidyl transferase cGAS functions as a pattern recognition receptor that senses cytosolic dsDNA. cGAS is allosterically activated by dsDNA and catalyzes the conversion of ATP and GTP into the cyclic dinucleotide 2’3’-cGAMP (or simply cGAMP) (Ablasser et al., 2013; Civril et al., 2013; Diner et al., 2013; Gao et al., 2013; Kranzusch et al., 2013; Sun et al., 2013), which subsequently acts as a secondary messenger that induces an antiviral program in the infected cell. The active site of cGAS contains three acidic residues coordinating two magnesium ions. The role of these ions is to coordinate the triphosphate group of the donor nucleotide and the attacking hydroxyl group of the acceptor nucleotide. cGAS catalyzes the formation of cGAMP in two sequential steps. First, the triphosphate group of ATP is coordinated by the magnesium ions and the 2’-hydroxyl group of GTP makes a nucleophilic attack on the α-phosphate of ATP, which releases the β- and γ-phosphate as pyrophosphate. This leads to the formation of a noncanonical 2’,5’-phosphodiester linkage. The intermediate is then flipped around in the active site and now the triphosphate group of GTP is coordinated by the magnesium ions. This time the 3’-hydroxyl group of the AMP moiety makes the nucleophilic attack on the α-phosphate of GTP forming a 3’,5’-phosphodiester linkage (Civril et al., 2013; Gao et al., 2013; Hornung et al., 2014). Thus, the final product contains both a canonical and noncanonical phosphodiester linkage.

Not all dsDNA is equally efficient at activating cGAS. The minimum DNA length reported to activate cGAS in cells is 12 bp with guanosine overhangs (Herzner et al., 2015). However, the DNA’s ability to activate cGAS is strongly related to the length of the DNA. Increasing the DNA length leads to an increase in its ability to activate cGAS (Andreeva et al., 2017; Luecke et al., 2017). This effect is observed even when increasing the DNA length from 2 kb to 4 kb (Luecke et al., 2017). Furthermore, certain Y-form DNA generated during the reverse transcription of the HIV-1 genome is more potent at activating cGAS compared to conventional dsDNA of similar length (Herzner et al., 2015).

cGAMP acts as a secondary messenger that binds to the adaptor protein STING, and this leads to the induction of antiviral genes (Ablasser et al., 2013; Diner et al., 2013; Li et al., 2013; Sun et al., 2013; Zhang et al., 2013). STING is a transmembrane protein located in the endoplasmic reticulum (ER) membrane with a large C-terminal domain protruding into the cytosol (Ishikawa and Barber, 2008). When STING binds cGAMP, the complex moves to the Golgi apparatus and from there it moves to punctuated foci in the cytoplasm (Saitoh et al., 2009). The STING:cGAMP complex recruits TBK-1, and this leads to the phosphorylation of both STING and TBK-1. This phosphorylated complex can then phosphorylate and thereby activate IRF3, which then translocates to the nucleus where it induces the transcription of antiviral genes including type I interferons (Ishikawa et al., 2009; Tanaka and Chen, 2012). The STING:cGAMP complex will also activate nuclear factor kappa B (NFκB) transcription factors (Abe and Barber, 2014).

The method described in this protocol was used to show that the in vitro activation of recombinant human cGAS truncated to amino acids 155-522 (cGAS [155-522]) is dependent on DNA length. The tested interval of DNA lengths varied from 100 bp to 4,000 bp (Luecke et al., 2017). This method offers an alternative to thin layer chromatography (TLC)-based assays with radiolabeled ATP. Due to poor sensitivity, TLC-based assays normally use concentrations of both dsDNA and cGAS well above physiologically realistic concentrations. The advantage of using the protocol presented here is that no radioactivity or labeling of the substrates are needed and that the high sensitivity of this method makes it possible to use very low concentrations of both cGAS and dsDNA. In this protocol, the concentration of cGAS is ten-fold lower compared to classical TLC assays and we have avoided oversaturating the reaction with DNA. We use 1 ng/μl of dsDNA corresponding to 1.646 x 10^-6 M bp. Assuming that one cGAS molecule covers approx. 20 bp (Andreeva et al., 2017), then 1.646 μM bp corresponds to 82.32 nM cGAS binding sites. Under this assumption, there is enough DNA to occupy about 82% of the cGAS used in this protocol. The use of low and approx. equimolar concentrations of cGAS and DNA (measured in cGAS binding sites) is important if you test DNA with small differences in affinity for cGAS. The impact of different affinities might be diminished if for example the DNA concentration is substantial above the saturation point.

This protocol allows for easy and robust quantifications of the cGAS product and compare reaction conditions (such as different buffers, DNA structures, DNA lengths, and cGAS preparations) but it is more time consuming than TLC when running multiple samples. The method described in this protocol was developed from a method designed to measure the activity of the oligoadenylate synthetase (OAS) proteins (Turpaev et al., 1997).