An intrinsically disordered protein (IDP) is defined as a folded protein that is completely or partially devoid of a unique fold in isolation in solution, and exhibits limited secondary structure (Trivedi and Nagarajaram, 2022). Viral IDPs (vIDPs) induce protein condensation via LLPS. Examples include rabies virus (RABV) or vesicular stomatitis virus (VSV) phosphoprotein P (Nikolic et al., 2017; Heinrich et al., 2018) and herpesviral proteins such as ICP4 (Seyffert et al., 2021) that are linked to animal infectious diseases.

DNA viruses create nuclear viral replication compartments (VRCs), which are the phase-separated products and function as the platform for generating progeny virions. Through the properties attributed to them, these compartments promote the spatial organization of viral processes and control virus-host interactions (Charman and Weitzman, 2020). ICP4, an immediate-early protein (IEP) family member and viral intrinsically disordered protein (IDP), plays a crucial role in viral replication and has been implicated in driving LLPS within herpes simplex virus type 1 (HSV-1) VRCs (Seyffert et al., 2021). ICP4 and its homologs are broadly conserved in alphaherpesviruses with the ICP4 [the IE180 of pseudorabies virus (PRV), IE of equine herpesvirus type 1 (EHV-1) and homologs in infectious laryngotracheitis virus (ILTV) exhibiting extensive sequence homology (Wu and Wilcox, 1991; Anderson et al., 1992; Johnson et al., 1995)]. It is likely that these proteins have similar functions, as PRV mutants deficient in the IE180 gene failed to generate new viral genomes, synthesize immediate early, early or late viral proteins and assemble infectious virions (Wu et al., 2014), and EHV mutants deficient in their IE protein only replicate in IE13.1 cells (Garko-Buczynski et al., 1998). In HSV-1, the nuclear form of VP22 specifically co-localizes with ICP4-enriched viral replication compartments (Pomeranz and Blaho, 1999), whereas EHV VP22 increases ICP4 synthesis after transcriptional activation (Okada et al., 2018). Antisense oligodeoxynucleotides against the putative start codon region of MDV ICP4 mRNA inhibit MSB-1 cell proliferation (Xie et al., 1996). Collectively, these findings indicate that the mechanism of IE protein-driven LLPS in NFAT-activated nuclear VRC formation is widely conserved among animal alphaherpesviruses. The proteins pE1, pM25, pM48.2, and pM57 of murine cytomegalovirus (MCMV) are simultaneously involved in VRCs. This process occurs in both the cytoplasm and the nucleus, functioning as a compartmentalization mechanism during the nuclear phase of replication in the viral replication cycle (Mahmutefendić Lučin et al., 2023).

The N protein of SARS-CoV-2 can undergo phase separation with viral RNA and the membrane-associated protein M, forming structures resembling viral factories (VFs). In the virion, the N protein forms vRNPs with RNA; these vRNPs are arranged in a specific shell-like configuration within the virus membrane. The interaction between the N protein and the M protein may facilitate the recruitment of vRNPs into developing virus particles, thereby promoting viral assembly (Lu et al., 2021). It is noteworthy that the vRNP of influenza A virus (IAV) and the host protein Rab11 form membraneless liquid organelles in the cytoplasm. These structures can assemble without requiring RNA interactions between different vRNP segments. While not involved in immune evasion, they support viral assembly (Alenquer et al., 2019), exhibiting features of VFs. Although IAV is not typically classified as a veterinary virus, its vRNP core proteins containing NP and RNA are evolutionarily conserved (Hu et al., 2017; Li et al., 2021). This suggests that the core vRNP or similar structures in species-specific IAVs, including avian influenza virus (AIV), swine influenza virus (SIV), and bat influenza A virus (BAT IAV), may utilize a consensus LLPS mechanism to form VFs or assembly centers. These dynamic and efficient viral replication sites not only concentrate the components required for viral genome replication and transcription, but may also spatially isolate or shield viral nucleic acids. NP inhibitors (such as F66, naproxen, RK424, etc.) may exhibit good inhibitory effects against these viruses, although challenges related to the poor bioavailability of some drugs need to be addressed (Hu et al., 2017). Additionally, porcine epidemic diarrhea virus (PEDV) uses the host factor YIPF5, whose knockdown impairs viral replication and inhibits PEDV infection. YIPF5 interacts with viral nonstructural proteins (NSP) 3, 4, and 6 to facilitate double-membrane vesicle (DMV) formation, similar to VFs. YIPF5 knockdown disrupts the interaction between NSP3 and NSP4, impairing DMV formation (Guo et al., 2025). These findings suggest YIPF5 as a key host factor in PEDV infection and highlight its potential as an antiviral target.

Inclusion bodies (IBs) are membraneless organelles that arise via viral LLPS and localize to a number of cellular compartments to support viral replication. The viral IBs also reorganize cellular architecture and cytoplasmic contents, including the cytoskeleton, a process that may disrupt immune response pathways, then facilitating viral assembly and budding (Wu et al., 2022).

The capsid of the zoonotic arthropod-borne Getah virus (GETV) undergoes LLPS droplets with the viral genomic RNA (gRNA) in vitro and forms cytoplasmic puncta to promote viral assembly. Two regions of the GETV gRNA (1–4,000 and 5,000–8,000 nucleotides) have been identified to enhance the in vitro formation of capsid protein (CP) droplets. The lysine-rich linker region within CP is crucial for phase separation. The ability of CP to undergo LLPS is nearly abrogated, when the lysine residues are substituted with the glutamic acids in the (K/E) mutant (Sun et al., 2024).

NDV can also form cytoplasmic IBs by LLPS, with the NP and P proteins serving as core drivers of condensate formation. In this process, the N terminal domain, N core region of NP and the C terminus of P play important roles (Sun et al., 2017). Considering NDV has oncolytic potential that specifically kills malignant cells (Liao et al., 2024), we speculate that their oncolytic specificity may correlate with strict LLPS regulation. In this regard, tumor cells are frequently characterized by the overexpression of IDR-containing proteins (e.g., oncogenic fusion proteins) that could act as scaffolds for NDV IBs and would therefore stabilize NP-P condensates and promote viral replication exclusively in malignant cells.

Rotavirus is a double-stranded RNA virus, and during replication forms cytoplasmic electron-dense inclusions that are sites for dsRNA synthesis and virion assembly (Geiger et al., 2021; Papa et al., 2021). Phosphorylation of NSP2 is essential for triggering LLPS-driven viral factory assembly. CK1α kinase phosphorylates NSP2 at serine 313 (S313), causing its transition from the dispersed cytoplasmic state (dNSP2) to the virus factory-localized state (vNSP2). Phosphorylated vNSP2 co-localizes with lipid droplet marker P-PLIN1 as early as 4 hpi, indicating that it targets the lipid droplet interface and may recruit host lipid metabolism components through phase separation. The hydrophobic lipid droplet surface promotes vNSP2 polymerization and phase separation. Virus-induced lipid droplet biogenesis (e.g., PLIN1 phosphorylation) optimizes the LLPS microenvironment, forming condensates favorable for viral genome replication. NSP2 phosphorylation acts as a “molecular switch” that coordinates phase separation of rotavirus proteins and host factor recruitment to optimize viral replication (Criglar et al., 2020).

RABV induces membraneless liquid organelles (Negri bodies) in the cytoplasm of the host for replication (Zhang et al., 2014), and its phosphoprotein (P) acts as a scaffold driven by LLPS to promote the assembly of Negri bodies (Nevers et al., 2022). Negri bodies contain the viral genome, antisense genome, and mRNA. Detection with BrUTP assay demonstrates the occurrence of viral transcription and replication in this compartment (Lahaye et al., 2009). This LLPS is a thermo-responsive process. In aqueous solutions at pH 7.5, RABV P undergoes autonomous phase separation above the lower critical solution temperature (LCST), modulated by salt concentration (Bouchama et al., 2025). Since RABV spreads through Negri bodies, understanding the dynamics of LLPS under specific temperature and salt conditions might help develop novel treatments for this still incurable disease.

Taken together, as viruses hijack host immune-driven LLPS to promote their replication, precisely targeted therapeutic strategies, such as disrupting virus-specific LLPS-driving domains or employing proteolytic degradation systems could be developed; meanwhile, combinatorial approaches that integrate LLPS modulators with immune agonists, such as STING activators VAX014 (Nelson et al., 2025), may synergistically enhance antiviral efficacy through multi-mechanistic interventions.

The DNA-specific sensor cGAS is a prime target for viral evasion. African swine fever virus (ASFV) utilizes virulence-associated proteins DP71L and QP383R to antagonize STING-dependent type I IFN signaling (Qi et al., 2023). DP71L contains a conserved PP1-binding motif that competes for the C-terminal tail (CTT) of STING and interferes with the STING-TBK1 interaction. Therefore, the presence of DP71L suppresses downstream phosphorylation events (e.g., STING, TBK1, and IRF3), ultimately destabilizing the LLPS of IRF3 required for antiviral IFN responses (Ranathunga et al., 2025). Concurrently, another viral factor QP383R impedes cGAS activity by inducing its palmitoylation, which blocks DNA binding, cGAS dimerization, and LLPS formation, thereby reducing cGAMP synthesis (Hao et al., 2023). Similarly, Marek’s disease virus (MDV) exploits its oncoprotein Meq to sterically prevent recruitment of TBK1 and IRF7 to the STING complex (Li et al., 2019), probably inhibiting LLPS of IRF7. In contrast, RABV adopts a resource competition strategy by which the viral N and P proteins redirect TLR3 for the strict biogenesis of Negri bodies as replication compartments (Ménager et al., 2009).

To neutralize the dsRNA-PKR-eIF2α-SG axis, many viruses encode PKR inhibitors (Langland et al., 2006). Vaccinia virus (VACV) K3 protein exhibits host-specific suppression of PKR, strongly inhibiting bovine and murine PKR to enhance viral replication by reducing eIF2α phosphorylation, whereas these effects are absent in sheep, goats and humans (Park et al., 2019), suggesting that the co-evolutionary arms race between PKR and viral inhibitors is conserved among K3 orthologs that execute species-and virus-dependent antagonism of PKR (Park et al., 2021). In contrast to PKR targeting by DNA virus, RNA viruses directly modulate SG assembly via G3BP manipulation. Foot-and-mouth disease virus (FMDV) protease L^Pro cleaves G3BP1/2 to block SG formation (Visser et al., 2019), whereas NDV concurrently sort G3BP1 and TIA-1 into replication-promoting SGs that sequester host mRNA (Miao et al., 2023). Alternatively, PRV relocates G3BPs to the nucleus by its IE180 protein and causes the suppression of cytoplasmic SG assembly (Zhao et al., 2025), suggesting that hijacking nuclear transport might be commonly used by alphaherpesviruses for immune evasion. Similarly, the condensates formed by the SARS-CoV-2 N protein, M protein, and RNA can also recruit G3BP1 to inhibit SG formation, thereby suppressing the host innate immune response. The phosphorylation state of the N protein dictates its interaction with G3BP1 and the localization of host mRNA. Mechanistically, the condensates formed by unphosphorylated N protein or phosphorylation-resistant mutants N^14SA and N^ΔSR appear larger and denser, with stronger binding of host mRNA to these condensates (Lu et al., 2021). This suggests that post-translational modifications of the N protein are key in regulating viral replication and host immune suppression.

Viruses of various animals have been shown to disrupt the LLPS-induced assembly of PML-NBs to enhance replication. The nonstructural protein nsp1 of PRRSV inhibits the activity of PML (Corpet et al., 2020), while apoptin also known as VP3 in chicken infectious anemia virus (CIAV) is modified by SUMO conjugation and targeted to PML-NBs (Janssen et al., 2007). In alphaherpesviruses (e.g., MDV, PRV), the US3 orthologs mediate PML-NB disruption through kinase-and proteasome-dependent mechanisms; but unlike the alphaherpesviral ICP0, they do not target PML for degradation (Jung et al., 2011; Liao et al., 2021). Japanese encephalitis virus (JEV) selectively destabilizes PML isoforms during infection, impairing antiviral immunity (Yang et al., 2023). In general, antagonism of LLPS-mediated intrinsic immunity is widespread in animal viral infections. In return, these evasion strategies highlight the critical role of LLPS in host antiviral defense, therefore further studies are warranted to understand how viruses exploit and disrupt phase-separated signaling hubs. Elucidating the molecular interplay between viral effectors and host phase-separated complexes will accelerate the development of LLPS-targeted antivirals and therapies.

Highly active proteins appear to be the principal modulators of LLPS, thus pharmacological modification of target protein activity might provide an effective means of increasing resistance to disease, or alleviating pathology. For example, in the case of an alphaherpesvirus such as MDV, the treatment with the proteasome inhibitor MG132 partially restores PML bodies, which are disrupted by the viral US3 protein (Liao et al., 2021), indicating that the proteasome-dependent pathway can be therapeutically targeted and represent a viable strategy for counteracting the pathogenesis of alphaherpesviruses.

Mechanistically, cGAS boosts the efficiency of DNA uptake by the LLPS-guided formation of droplets (Yao et al., 2022). Immune modulators, such as IRTKS that reinforces the phase separation of STING, can potentiate interferon production (Xie et al., 2024). Additionally, the PI3K and MAPK/p38 pathways are engaged in promoting SG assembly through the mechanistic target of rapamycin complex 1 (mTORC1), making them potential therapeutic antiviral targets (Heberle et al., 2019). A notable feature of LLPS driven by the RABV phosphoprotein is dependent on its thermo-responsive property, wherein temperature and ionic strength regulate the LLPS state (Bouchama et al., 2025), presenting a therapeutic avenue for the treatment of localized hyperthermia. For the rapidly evolving RNA viruses, the effectiveness of LLPS-based therapeutic approaches largely depends on the structural conservation of targeted proteins and the binding affinity of sequence-specific genomic loci. It has been recognized that evolutionary mutations are unevenly distributed in the genomes of RNA viruses (Kustin and Stern, 2021). If the LLPS-based antivirals are designed to target the viral glycoproteins for cellular entry and egress, in which higher frequency of sequence mutations tend to occur, their inhibitory effects will be considerably diminished. In contrast, sequences coding for RNA-dependent RNA polymerase (RdRp) and other nonstructural proteins are relatively conserved in the species of a closely related family, and thus targeting these most conserved regions is expected to favor the benefit of LLPS-driven antiviral therapeutics.

The improvement of LLPS properties through capsid engineering provides great potential for viral vector optimization. PRV, for instance, hijacks host protein-mediated LLPS for nuclear entry (Zhao et al., 2025). A central question is whether the engineered capsids can target low-complexity domains (LCDs) or nuclear VRCs to increase the efficiency of transduction. The use of such vectors would greatly benefit from engineering viral capsids to take advantage of IDR-mediated LLPS to fine-tune the efficiency of the vector. Current siRNA studies have evaluated interactions between nucleolar proteins (B23/nucleophosmin and nucleolin) and adeno-associated virus type 2 (AAV2) capsids. Similar to inhibition of proteasome, knockdown of nucleophosmin enhanced the nuclear accumulation of AAV2 and increased its transduction efficiency by 5–15 folds (Johnson and Samulski, 2009). Given the conserved mechanisms of nuclear transport between AAV and PRV, these findings may provide transferable strategies to engineer PRV vectors. In addition, mRNA can modify T cells through chimeric antigen receptor (CAR)-T cells, and LLPS-mediated antigen aggregation can enhance T cell responses (Wiesinger et al., 2019; Lee et al., 2023). This may suggest that future mRNA-lipid nanoparticle (LNP) vaccines could target LLPS structures to exert clinical potential.