Computational protocol: Analysis of the SUMO2 Proteome during HSV-1 Infection

Similar protocols

Protocol publication

[…] A method involving nickel affinity purification of His-tagged sumoylated proteins under denaturing conditions [] gave efficient recovery of both overall sumoylated proteins () and the specific example of sumoylated PML (). HA-HisSUMO2 cells were grown in isotopically normal (light; L) or heavy (H) SILAC media (in which lysine and arginine have the heavy isotopes of nitrogen (15N) and carbon (13C)). Uninfected HA-His only control cells were grown in isotopically intermediate (M) medium, as defined in the methods section. The L cells were infected with wt HSV-1 at MOI 10, then all but one plate of cells from each condition were harvested directly into guanidinium denaturing buffer 12 h later. The H, L and M lysates were mixed in equal protein amounts then used for nickel affinity purification of SUMO2-modified proteins. To allow analysis of total protein abundance changes and confirm efficient virus infection and degradation of previously characterized cellular proteins, cells from one plate of each set were harvested directly into SDS-PAGE loading buffer. Samples of the crude and affinity purified mixtures were subjected to SDS-PAGE, then the gels were stained and cut into slices for in gel tryptic peptide production. The peptides were analyzed by LC-MS/MS and the data analyzed using MaxQuant software (See for details). shows a flow diagram of the experimentation and images of the resulting SDS-PAGE stained gels. Data derived from the crude and purified samples gave information on the relative levels of total protein and putative sumoylated protein species respectively. Analysis of the H/M ratios in the purified sample allowed the separation of likely sumoylated species from non-specifically purified proteins, while the H/L ratios of these protein IDs in the same sample can be used to assess changes in relative abundance during HSV-1 infection. [...] H/L ratios can be used to study changes to either the total proteome (via ‘crude’ data), or the SUMO proteome (via ‘pure’ data) upon HSV-1 infection. Frequency distribution charts comparing ‘crude’ ratios from infected cells and uninfected cells showed few changes in total protein levels (, blue line). However, while most proteins in purified preparations also showed no change in abundance, the distribution is skewed towards larger ratios (, red line), with putative SUMO2 substrates being mostly responsible for this high ratio tail (, insert). This shows that the non-substrates are largely unchanged during HSV-1 infection, while SUMO2 conjugates have a tendency toward high ratios, indicative of a widespread loss of SUMO2 conjugation during infection. This is consistent with western blot data (). To test the reproducibility of these data a similar quantitative proteomic experiment was undertaken, this time only including Light infected and Heavy uninfected samples (). Although the total number of proteins was lower in this compared to the triple labeled experiment (), there was considerable overlap between the proteins in the purified fractions (), and the details of their ratio changes correlated substantially (; ). Because the triple labeled experiment was both larger and included the His-only control, subsequent sections mostly refer in detail only to this dataset.To determine which SUMO2 substrates changed significantly during HSV-1 infection, Significance B (SigB) values were calculated using Perseus from the MaxQuant suite of software []. SigB is calculated using both signal intensity and SILAC ratio, and is indicative that a protein abundance change significantly deviates from the bulk of the quantified proteins. Of the 877 putative sumoylated proteins, 260 changed in abundance (either up or down) in the purified fraction of the triple SILAC experiment with SigB values of less than 0.1 (, sheet 3). This number includes duplicate entries for proteins, such as PML, for which more than one isoform was detected. Removal of these duplicates and restricting the list to entries with H/L ratio increases of 2-fold or more results in a list of 185 proteins, shaded according to degree of change, and listed in order of decreased abundance (). An additional 18 proteins on the putative sumoylated substrates list were detected with H/L ratios of 2 or greater and SigB values of less than 0.1 in the purified fraction of the replicate experiment, but which were not recorded as significantly regulated substrates in the primary experiment (). As shown below, at least one of these is an authentic sumoylated substrate whose abundance decreases during HSV-1 infection. Overall therefore, up to around 200 cellular proteins identified in the purified fractions of the experiments decreased significantly in abundance during infection.The degree of change in abundance of the putative sumoylated forms of these proteins varies considerably, up to a maximum of greater than 20-fold. In broad view, these data are consistent with previous observations on the decreased stability of sumoylated cellular proteins during HSV-1 infection. They also support the idea that there is considerable specificity or selectivity to the extent of sensitivity to desumoylation, as two-thirds of sumoylated proteins remain largely unaltered during infection while only 14% and 1% decrease in abundance by over 3-fold and 7-fold respectively.We also identified a further 72 proteins that were not defined as potentially sumoylated on the basis of H/M ratios, which nonetheless had H/L ratios of greater than 2 and SigB of less than 0.1 in the purified fraction (), although only 32 of these also complied with these criteria in the double labeled experiment. These may represent proteins that are non sumoylated, but which have an affinity for the Ni-agarose beads and whose abundance decreases during HSV-1 infection.A small number of putative sumoylated proteins listed in sheet 3 also exhibited H/L ratios of greater than 2 in the crude fraction (), indicating candidate SUMO2 substrates whose total protein levels also reduced in infected cells. In the cases of IFI16, CENPB and PML, this has been reported previously [,–]. Another protein on this list (NACC1) will be considered below, and further data presented below suggests that this list does not include all proteins that behave in this manner.The levels of a number of proteins were changed significantly in the crude samples (, sheet 4). Of these, and excluding those already noted above as regulated SUMO2 substrates, 128 proteins gave H/L ratios of greater than 2 in the crude samples (). Given that HSV-1 infection causes the shut-off of host protein synthesis through induction of mRNA instability [], this list is perhaps shorter than might be expected. It is possible that low abundance proteins with short half-lives (and thus those most susceptible to decreased host transcription) have not been detected efficiently in the ‘crude’ preparations by this approach, which will naturally favor the most abundant cellular proteins. […]

Pipeline specifications

Software tools MaxQuant, Perseus
Application MS-based untargeted proteomics
Organisms Viruses, Human poliovirus 1 Mahoney
Diseases Infection, Leukemia, Promyelocytic, Acute, HIV Infections, Epstein-Barr Virus Infections