henhaobio
Sign in

Scientific ReportsEvolutionary BiologyPeer-reviewed journal article

Evolution of intrinsic disorder in the structural domains of viral and cellular proteomes

Fizza MughaliD, Gustavo Caetano‐AnollésiD

PDFDOIPreprintData
In-depth full-text digest available · Jump to digest ↓

Abstract

Intrinsically disordered regions are flexible regions that complement the typical structured regions of proteins. Little is known however about their evolution. Here we leverage a comparative and evolutionary genomics approach to analyze intrinsic disorder in the structural domains of thousands of proteomes. Our analysis revealed that viral and cellular proteomes employ similar strategies to increase disorder but achieve different goals. Viral proteomes evolve disorder for economy of genomic material and multifunctionality. On the other hand, cellular proteomes evolve disorder to advance functionality with increasing genomic complexity. Remarkably, phylogenomic analysis of intrinsic disorder showed that ancient domains were ordered and that disorder evolved as a benefit acquired later in evolution. Evolutionary chronologies of domains indexed with disorder levels and distributions across Archaea, Bacteria, Eukarya and viruses revealed six evolutionary phases, the oldest two harboring only ordered and moderate disorder domains and delimiting two universal common ancestors of life. A biphasic spectrum of disorder versus proteome makeup captured the dichotomy in the evolutionary trajectories of viral and cellular ancestors, one following reductive evolution driven by viral spread of molecular wealth and the other following expansive evolutionary trends to advance functionality through massive domain-forming co-option of disordered loop regions. TeaserComparative and evolutionary genomic analyses of the intrinsic disorder of structural domains in thousands of proteomes reveal proteomes increase disorder as a benefit acquired later in evolution to support viral economy and multifunctionality or advanced cellular functionality.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (3 Nov 2024) · CC BY 4.0

Across 8,127 cellular and viral proteomes, ordered SCOP domains originated first in a phylogenomic chronology; highly disordered domains emerged late, and viruses dominate highly disordered proteomes.

Earlier work on disorder origins relied on sequence data, which high mutation rates, horizontal gene transfer and genetic mosaicism confound, particularly in viruses, and proteome-level surveys rarely examined evolution across organizational levels. Tracing predicted disorder of conserved SCOP structural domains along a domain chronology from 8,127 proteomes lets disorder history be compared across domains, proteomes, viruses and cells.

Key results

  1. ShownIn a phylogenomic chronology of SCOP fold families, ordered domains appeared first, then moderately disordered and finally highly disordered domains. Ordered domains also accumulated the most, with a burst near the middle of the timeline.

    Ordered domains appear first in all 15 Venn taxonomic groups; oldest family c.37.1.12 (ABC transporter ATPase domain-like) is ordered, oldest moderate one is b.40.4.5; first high-disorder family b.4.1.1 (HSP40/DnaJ) arose in Phase II, after the BEV group at nd = 0.214; the earliest high-disorder domains (nd 0.2-0.3) had mean disorder of only 0.317-0.441. · Evolution of intrinsic disorder analyzed at…; Intrinsic disorder of structural domains is a…; Figure legends

  2. ShownCellular and viral proteomes follow opposite disorder trends. In cells, mean disorder rises with mean protein length, protein count and domain use and reuse; in viruses it falls. Each taxonomic group has its own disorder profile.

    Disorder distributions differ among Archaea, Bacteria, Eukarya and viruses (Kruskal-Wallis, Holm-Bonferroni corrected p = 1.162e-224); most archaea (except Halobacteria) and bacteria are ordered, most eukaryotes moderate; viruses make up most high-disorder proteomes, ssDNA viruses being most disordered. · Intrinsic disorder of structural domains…; Viral and cellular proteomes increase disorder…; Figure legends

  3. SuggestedHigh disorder appears evolutionarily derived. The universal ABEV group holds the largest share of ordered and moderately disordered domains but only about a tenth of highly disordered ones, over a third of which fall instead in the cellular ABE and BE groups.

    ABEV holds 27.5% of ordered and 22.8% of moderate disorder domains but only 9.9% of high disorder domains; the ABE and BE groups hold 19.8% and 18.3% of high disorder domains, respectively. · Evolution of intrinsic disorder analyzed at…

  4. SuggestedOrdered domains shared between cells and viruses are more widespread across cellular proteomes than those shared only among cells, which the authors read as viruses spreading innovations to hosts. For moderately and highly disordered domains the difference is not significant.

    Two-tailed Wilcoxon tests on f-values, ordered domains: Archaea p = 3.88 x 10^-5, Bacteria p = 2.12 x 10^-11, Eukarya p = 5.81×10^-17; moderate disorder p = 0.756/0.064/0.251 and high disorder p = 0.156/0.0455/0.567, reported as non-significant. · Evolution of intrinsic disorder analyzed at…; Intrinsic disorder tailors proteome makeup; Figure legends

  5. ShownGrouping viral domains by host superkingdom largely mirrors the global pattern. The core shared by archaeo-, bacterio- and eukaryoviruses is oldest and ordered or moderately disordered, while eukaryovirus-specific domains span the full disorder range and contain most capsid domains.

    Venn analysis of 1,526 viral domain families; the abe core holds the most ancient domains (ordered and moderate disorder); the e group holds 29 of 37 known capsid domains, clustered at nd = 0.468-0.524, with five in b, 2 in be and 1 in abe. · Evolution of intrinsic disorder analyzed at…; Intrinsic disorder of structural domains is a…; Figure legends

Methods, data and reproducibility

IUPred (long mode, 0.5 residue cutoff) disorder fractions were averaged per proteome and per SCOP fold family (ASTRAL 1.75 sequences), binned as ordered (0-0.1), moderate (0.1-0.3) or high (>0.3), and mapped onto a phylogenomic domain chronology (node distance), Venn sharing groups, f-value spread and function categories.

  • Representative and reference cellular proteomes sampled across major RefSeq taxonomic groups · Archaea, Bacteria and Eukarya · n = 139 archaeal, 1,734 bacterial, 210 eukaryal proteomes
  • Nonredundant viral proteomes from the NCBI viral genomes project (as of January 2019), grouped by Baltimore class and host superkingdom · viruses · n = 6,044 of 6,540 with domain assignments; host data for 5,959
  • SUPERFAMILY HMM fold-family assignments (e-value < 0.0001) in completely sequenced genomes, used for the domain chronology and proteome analyses · cellular organisms and viruses · n = 8,127 proteomes
  • ASTRAL SCOP domain sequences (version 1.75) annotated with fold family, time of origin and molecular function (7 broad, 49 detailed categories) · n = 103,000 sequences from 110,800 structures
  • Viral domain families grouped by host superkingdom (archaeo-, bacterio-, eukaryoviruses); counts from prior Mughal et al. data · viruses · n = 1,526 domain families
  • High-disorder domains with GO biological process terms from SDADB (probability score 1) · n = 131 domains
  • Known viral capsid protein domains ordered by host-centric Venn group and time of origin · viruses · n = 37 domains
Methods & tools
IUPred (standalone, 'long' option), SUPERFAMILY database hidden Markov models, HMMER, SCOP fold-family classification (concise classification strings), ASTRAL compendium 1.75, SCOPe, NCBI RefSeq and NCBI Viral Genomes Project, Baltimore classification, Kruskal-Wallis H-test with Holm-Bonferroni correction, Wilcoxon rank sum test (two-tailed), Gene Ontology via Structural Domains Annotation Database (SDADB), Boxenplots (letter-value plots)
Reagents & environment
Standalone IUPred, 'long' disorder option (version not stated); ASTRAL compendium version 1.75; SUPERFAMILY HMM assignments with e-value < 0.0001; NCBI viral genomes as of January 2019.
Reproducibility gaps
No code repository or analysis scripts are provided; the data statement points to SCOP/SCOPe plus the article and supplementary files, and some f-value and domain-count data are reused from earlier Mughal et al. work.
Limitations (5)
  • Elevated disorder in halobacterial proteomes may be a predictor artifact: disorder estimators were developed on non-halophilic proteins in ordinary salt conditions, whereas halophile proteins keep ordered folds only at high salt. · stated by the authors
  • Because f-value distributions of moderately and highly disordered domains did not differ significantly between cell-only and cell-virus shared sets, the direction of gene transfer for these derived domains cannot yet be inferred. · stated by the authors
  • Membership in the abe viral Venn group only counts domains found across archaeo-, bacterio- and eukaryovirus proteomes; it does not imply that any single virus infects hosts in all three superkingdoms. · stated by the authors
  • Disorder comes from one predictor (IUPred long, 0.5 cutoff) applied to ASTRAL sequences of structurally solved domains, with no second predictor or experimental check; such sequences may understate disorder relative to full-length proteins. · noted in this summary
  • The bacterial high-disorder f-value test gives p = 0.0455 yet is reported as non-significant, and the significance threshold or multiple-testing correction behind these Wilcoxon calls is not stated. · noted in this summary

Read this if you study protein intrinsic disorder, virus-host evolution or phylogenomic domain chronologies and want a structure-based account of when disorder arose in cells and viruses.

Citations
16
FWCI
4.14
Field percentile
top 5%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISZarin '19Forcelloni '19Bustos '24Lafforgue '22Jiang '26Basile '19
58% semantic matchGenomics
Taraneh Zarin, Bob Strome, Alex N. Nguyen Ba+3 more·eLife·Jul 2, 2019· 214 citations