Apprehending the NAD+-ADPr-Dependent Systems in the Virus World

Abstract

NAD⁺ and ADP-ribose (ADPr)-containing molecules are at the interface of virus-host conflicts across life encompassing RNA processing, restriction, lysogeny/dormancy and functional hijacking. We objectively defined the central components of the NAD⁺-ADPr networks involved in these conflicts and systematically surveyed 21,191 completely sequenced viral proteomes representative of all publicly available branches of the viral world to reconstruct a comprehensive picture of the viral NAD⁺-ADPr systems. These systems have been widely and repeatedly exploited by positive-strand RNA and DNA viruses, especially those with larger genomes and more intricate life-history strategies. We present evidence that ADP-ribosyltransferases (ARTs), ADPr-targeting Macro, NADAR and Nudix proteins are frequently packaged into virions, particularly in phages with contractile tails (Myoviruses), and deployed during infection to modify host macromolecules and counter NAD⁺-derived signals involved in viral restriction. Genes encoding NAD⁺-ADPr-utilizing domains were repeatedly exchanged between distantly related viruses, hosts and endo-parasites/symbionts, suggesting selection for them across the virus world. Contextual analysis indicates that the bacteriophage versions of ADPr-targeting domains are more likely to counter soluble ADPr derivatives, while the eukaryotic RNA viral versions might prefer macromolecular ADPr adducts. Finally, we also use comparative genomics to predict host systems involved in countering viral ADP ribosylation of host molecules.

Keywords: ADP-ribose; ADP-ribosyltransferase; NADase; RNA polymerase; RNA repair; anti-phage systems; cyclic ADP-ribose; nicotinamide adenine dinucleotide; nucleotides; sirtuin; virus evolution.

Figure 1

(a) Structure of NAD⁺ and (b) the substrates and products of various enzymes in the NAD⁺–ADPr network. Enzymes are color-coded based on the pathway in which they are involved. Bonds that are the target of particular enzymes are highlighted with colored circles.

Figure 2

Distribution of NAD⁺–ADPr network domains in the virus world. (a) Fraction of RNA (left) and DNA (right) viruses with a genome ≥ a given length containing NAD⁺–ADPr network domains. (b) Distribution of various NAD⁺-ADPR domains in the Myoviruses, Siphoviruses and Podoviruses. The graphs depict the number of phages per 1000 containing the given domain in that morphological category. (c) Prevalence of various NAD⁺–ADPr processing domains depicted as a percentage of the total number.

Figure 3

Phylogenetic trees of select viral NAD⁺–ADPr network components illustrating their origins as described in the text. (a) TIR, (b) Arc, (c) VIP2-like ART, (d) PART, (e) Ecto-ART, (f) Macro domain. Clades with a bootstrap support of > 75% are marked by colored circles. Several clades are collapsed in the trees for brevity. Relevant exchanges of genes are indicated. The raw data for the phylogenetic trees can be obtained from Supplementary S3.

Figure 4

(a) Topology diagram of the core of the Macro domain and the EF-Tu-like GTPases illustrating their structural relationship. Strands labeled with an ‘S’ prefix followed by their order number in the core structure are shown as yellow arrows, whereas helices which are similarly labeled with a ‘H’ prefix are shown as red cylinders. (b) The structure of the Nlig-Ia domain (cartoon rendering) that exists as a solo domain only in viruses. The figure illustrates the residues involved in binding NAD⁺ and its relative position with respect to the C-terminal ATP-grasp and RAGNYA domain of the NAD⁺-dependent ligases (rendered as a tube).

Figure 5

Representative contextual associations including domain architectures and gene neighborhoods of various domains of the NAD⁺–ADPr network. Gene neighborhoods are shown as box arrows with the arrowhead pointing to the 3′ gene. Domain architectures are shown by other shapes. The contextual associations are categorized based on their genomic contexts or their function including (a) domains associated with the Terminase-portal genes and those encoding other virion components; (b) domains involved in NAD⁺ synthesis; (c) secreted toxin domains; (d) domains that are components of T–A and related conflict systems; (e) domains in RNA virus polyproteins; (f) domains involved in a predicted RNA repair system; (g) viral TIR systems; (h) domains involved in ADPr-processing and; (i) SLOG sensor-activated systems. Gene neighborhoods are labeled with the accession number and species name of the gene marked with an asterisk.

Figure 6

Complete and pairwise co-occurrence patterns of NAD⁺–ADPr domains depicted as Euler diagrams. (a,b) Domains involved in NAD⁺ biosynthesis/salvage in viruses. (c,d) Co-occurrence of the Macro, SLOG, Nudix and NADAR domains in DNA viruses. Co-occurrences are measured as a percentage of all the proteins that are being compared in a particular Euler diagram.

Figure 7

Contextual associations including (a) domain architectures and (b) gene neighborhoods of the ARG-associated systems. (c) More gene neighborhoods of the ARPP domain. (d) Contextual network diagram and (e) co-occurrence frequencies of domains associated with the ARG domain. (f) Structural comparison of the newly identified members of the TY-chaperone superfamily found in these systems.