Human cytomegalovirus interactome analysis identifies degradation hubs, domain associations and viral protein functions

(A) Relative abundance of all canonical and non-canonical viral proteins quantified in experiment whole cell lysate 3 (WCL3) from Weekes et al. (2014) and whole cell lysate series three from Fielding et al. (2017). Further details of the calculations employed are given in Figure 1—figure supplement 1A and the Materials and methods section. (B) Details of all 172 baits. Bait expression was verified by IB, MS or RT-qPCR (Figure 1—figure supplement 1B). (C) Relative abundance of all human proteins expressed in HFFFs, calculated as described in (A). The ‘rank’ column indicates the ranked average IBAQ abundance. The most abundant protein calculated by this method was ranked 1, and least abundant ranked 8129. (D) Coding sequences of all viral genes used in this study. A six base-pair linker region, a V5 tag then a stop codon directly followed each sequence (Key Resources Table). Codon usage was optimised for expression for US14, US17 and UL74. (E) Oligonucleotides and templates employed in the generation and RT-qPCR of each viral vector. (F) Oligonucleotides and templates employed in the generation and RT-qPCR of each human overexpression vector.

(A) Numbers of HCIPs per bait, excluding bait-bait interactions. (B) HCIPs for each bait (see Figure 1 and the Materials and methods section for details of the filtering employed, and the scores shown in this table). For baits solubilised in NP40, VHCIPs are shown in green. The ‘Prey IBAQ rank’ column shows the ranked IBAQ abundance from Supplementary file 1C, and gives an indication of how abundant each prey protein was in infected HFFFs. A range of ranks is shown where more than one isoform of a protein could be detected, in order to reflect data for all isoforms of that protein. Abundantly expressed prey may be more easily validated using IB with antibodies against an endogenous protein; less abundant proteins may require overexpression to enable detection. (C) All detected interacting proteins for each bait, without filtering.

Columns give details of the database(s) that included each interaction, the method used, and cell type employed. Interactome scores from the present study are shown in columns H-K. Column L shows whether a given interaction was validated in this interactome. A value of 1 indicates validation; 0 indicates detection of the interaction but failure to pass stringent scoring thresholds; ‘ND’ indicates the interaction was not detected by the interactome. Column M shows the ranked abundance of each human prey protein from Supplementary file 1C. Interactions that were not detected in this study included a number of prey proteins that could not be detected in HFFFs. Further details are given in the Materials and methods section.

(A) All enriched functional pathways amongst all human HCIPs (p<0.05, after Benjamini-Hochberg adjustment). Column D shows the bait(s) interacting with each pathway component. (B) Further details of viral baits interacting with components of each pathway. Two values are shown: ‘% interaction’, the percentage of human interactors of each bait that belonged to the pathway (relates to Figure 2, where viral baits are included if >33% of interactors belonged to a given pathway). ‘% function’ illustrates the percentage of proteins from the pathway that interacted with the bait (relates to Figure 2—figure supplement 1, where viral baits are included if >33% of the pathway components identified interacted with a given viral bait). Values of >33% are coloured in this table. The ‘count’ column shows the total number of interacting pathway members; Figure 2 and Figure 2—figure supplement 1 included data with counts ≥ 2. (C) All enriched functional pathways amongst human HCIPs from each temporal class (p<0.05, after Benjamini-Hochberg adjustment). Column E shows the bait(s) interacting with each pathway component. This data underlies Figure 2—figure supplement 2A. (D) Temporal interactions of viral bait and viral prey proteins. This data underlies Figure 2—figure supplement 2B.

(A) HCMV proteins that contain each described Pfam domain. Links are given to additional information on each domain on the Pfam website. Overall 96 domains have been identified in HCMV proteins by Pfam, however only 10 domains were identified in two or more baits. Only this subset was examined to increase confidence in domain association predictions. (B) Subset of Supplementary file 2B illustrating individual protein-protein interactions that underpin data shown in Figure 4A.

Interactome data identified viral baits for 31 of these degraded proteins.