NSs (Rift Valley fever virus, strain ZH-548 M12; UniProt P21698) — curation notes

Deep research was unavailable for this review. No deep-research API keys are
configured in this container, so just deep-research was not run and no
-deep-research-<provider>.md file exists for this gene. The literature below was
assembled by hand from the cached publications/PMID_*.md records, the UniProt flat
file (NSS-uniprot.txt), PubMed Central (for figure-level detail where the cached
record is abstract-only), and QuickGO/OLS for term definitions and comparator
annotations. Where a cited paper's cached record is abstract-only, that is flagged.

Summary of the gene product

NSs is the 265-aa nonstructural protein encoded in the genomic-sense ORF of the
ambisense S segment of RVFV, and is the virus's principal virulence factor. It is
dispensable for genome replication in interferon-incompetent cells but required for
infection of interferon-competent cells, where it dismantles two arms of the host
antiviral response by acting as a substrate-recognition adaptor that redirects host
SCF ubiquitin ligases onto host proteins:

Nuclear pool → TFIIH. NSs binds the p62/GTF2H1 subunit of the general
transcription factor TFIIH through a short ΩXaV motif at its C-terminus, recruits the
F-box protein FBXO3 (with SKP1), and so drives proteasomal destruction of p62. It
also competes with XPD for p44 and sequesters p44 and XPB into nuclear filaments.
TFIIH depletion shuts down host RNA polymerase II transcription initiation globally,
which is how NSs prevents induction of the IFN-β gene.
Cytoplasmic pool → PKR. NSs binds EIF2AK2/PKR and recruits FBXW11, degrading PKR
and so preventing eIF2α phosphorylation and preserving translation.
Downstream cytopathology. By blocking host transcription NSs also prevents the
antiviral upregulation of Abl2, perturbing the actin cytoskeleton, lamellipodia and
adherens junctions.

Localization

NSs is the only RVFV protein found in the nucleus, which is unexpected for a virus
that replicates entirely in the cytoplasm.

A 34-kDa nonstructural protein was identified in nuclei of RVFV-infected Vero cells
PMID:7108491 (cached record is abstract-only).
NSs expressed alone forms the same nuclear filaments seen in infected cells, and the
acidic C-terminal domain is needed for filaments but not for nuclear import
PMID:10233964 (full text
cached). The same paper maps self-association to the last 10–17 C-terminal residues by
two-hybrid analysis.
UniProt records both compartments with experimental evidence:
Host nucleus {ECO:0000269|PubMed:10233964, ECO:0000269|PubMed:25918396, ECO:0000269|PubMed:7108491}. Host cytoplasm {ECO:0000269|PubMed:25918396}
[file:RVFVZ/NSS/NSS-uniprot.txt "SUBCELLULAR LOCATION: Host nucleus"].

Mechanism 1 — TFIIH destruction and host transcription shutoff

RVFV infection causes a rapid collapse of host RNA synthesis tracking the loss of
TFIIH, and NSs binds p44, forming nuclear filaments that also contain XPB
[PMID:14980221 "infection by RVFV leads to a rapid and drastic suppression of host
cellular RNA synthesis that parallels a decrease of the TFIIH transcription factor
cellular concentration"; "By competing with XPD, the natural partner of p44 within
TFIIH, and sequestering p44 and XPB subunits, NSs prevents the assembly of TFIIH
subunits"] (abstract-only in cache).
The p62 subunit is destroyed via the ubiquitin/proteasome system, and NSs recruits the
F-box protein FBXO3 to do it; FBXO3 or SKP1 knockdown rescues p62 and raises IFN
transcription [PMID:24403578 "Small interfering RNA (siRNA)-mediated depletion of
FBXO3 rescued p62 protein levels in RVFV-infected cells and elevated IFN transcription
by 1 order of magnitude"; "NSs recruits FBXO3 to destroy the general host cell
transcription factor TFIIH-p62, resulting in suppression of the transcriptional
upregulation of innate immunity"] (abstract-only in cache). Cullin 1, cullin 7 and
Rbx1 knockdowns did not rescue p62, so the ligase composition is unusual.
A ΩXaV motif (residues 261–265) mediates p62 binding; mutating Phe261 abolishes both
filament formation and p62 degradation and attenuates virulence PMID:25918396
(abstract-only in cache). The motif mimics host NER/transcription factors that dock on
p62.
UniProt's curated summary ties the two pools together: Cytoplasmic NSs interacts with host FBXW11 to degrade PKR whereas ... nuclear pool binds to host FBXO3 to target TFIIH subunit GTF2H1 for proteasomal degradation with the help of the linker protein SKP1 [file:RVFVZ/NSS/NSS-uniprot.txt].

Mechanism 2 — interferon antagonism is at the level of IFN induction

This distinction matters for the GO review. NSs does not interfere with signalling
downstream of the type I IFN receptor; it prevents the IFN-β gene from being transcribed
in the first place.

NSs alone blocks IFN-β at the transcriptional level, while leaving IRF-3, NF-κB and
AP-1 activation intact [PMID:15331713 "we demonstrate that NSs is sufficient to block
IFN-beta gene expression at the transcriptional level"; "In the presence of NSs, IRF-3
was still activated and moved to the nucleus. Likewise, NF-kappaB and AP-1 were
activated normally, as shown in electrophoretic mobility shift assays"]. The paper's
own conclusion is that the block is general, not IFN-specific: PMID:15331713 and PMID:15331713.
Figure-level detail (from PMC514997, outside the cached abstract, consulted for
judgment only): IFN-β promoter–luciferase reporters in 293 cells (Fig. 1D, Fig. 3A/C,
dose titration to >95% inhibition), RT-PCR and Northern blots for IFN-β mRNA after
ZH548 vs clone 13 infection (Fig. 1A/B), cytoplasmic NSs mutants losing inhibition
(Fig. 5A), suppression of a CMV-driven β-galactosidase reporter and of the SV40 early
promoter (Figs. 3B, 5B, 7D), and EMSAs showing normal NF-κB/AP-1 activation (Fig.
6D/E). So the assays are reporter/transcript assays of IFN production and of
constitutive promoters — none of them an IFN-receptor signalling assay.
Reverse genetics confirms the same at the level of the virus: NSs-deletion mutants of
MP-12 induce IFN-β mRNA and lose host shutoff [PMID:16501102 "accumulation of beta
interferon mRNA occurred after infection by these NSs deletion mutants, but not after
infection by MP-12"; "failed to shut off host protein synthesis in Vero cells"].

Consequence for GO. GO:0039502 is defined as interference with "a type I
interferon-mediated signaling in the host organism", and its parent GO:0140886 as
reducing "interferon-mediated signaling in the host organism" — i.e. the branch for
antagonists of the IFN receptor → JAK/STAT → ISG axis. What NSs is shown to do is
suppress type I IFN production. GO currently has no symbiont-mediated counterpart of
GO:0032480 negative regulation of type I interferon production (checked by QuickGO
search: the only symbiont-mediated interferon terms are GO:0039502, GO:0140884,
GO:0140885 and GO:0140886, all signalling-pathway terms), so the production-blocking
viruses are all parked on the signalling term. Comparator check confirms this is a
consortium-wide convention rather than a student error: influenza A NS1 (P03496), the
textbook IFN-induction antagonist, carries GO:0039502 only by the same
UniProtKB-KW:KW-1114 IEA mapping, and Ebola VP35 (Q05127) carries GO:0039502 by a
CACAO IMP. Hence this review records a proposed_new_terms entry and treats the four
RVFV rows on this axis as MODIFY-to-NTR rather than as errors of the annotators.

Mechanism 3 — PKR degradation

NSs induces specific degradation of PKR [PMID:19211744 title: "NSs protein of rift
valley fever virus induces the specific degradation of the double-stranded RNA-dependent
protein kinase"] and thereby blocks eIF2α phosphorylation PMID:19197350 (titles taken from the UniProt reference
block; these papers are not separately cached).
The responsible ligase is SCF^FBXW11 [PMID:26837067 title "Protein Kinase R degradation
is essential for Rift valley fever Virus infection and is regulated by SKP1-CUL1-F-box
(SCF)FBXW11-NSs E3 ligase"], and transcription inhibition is separable from PKR
degradation [PMID:23063407 title "Rift Valley fever virus NSs inhibits host
transcription independently of the degradation of dsRNA-dependent protein kinase PKR"].
Note on the molecular function: the mechanism is degradation, not occupancy-based
inhibition of the kinase active site. The UniProt keyword KW-1102 Inhibition of host PKR by virus maps to GO:0030291 protein serine/threonine kinase inhibitor activity
("Binds to and stops, prevents or reduces the activity of a protein serine/threonine
kinase"), which mis-describes the activity. GO:1990756 ubiquitin-like ligase-substrate adaptor activity ("The binding activity of a molecule that brings together a
ubiquitin-like ligase ... and its substrate. Usually mediated by F-box BTB/POZ domain
proteins") is the activity NSs actually has, for both PKR/FBXW11 and p62/FBXO3, and it
is a descendant of the GO:0030674 term UniProt already annotates by IDA (QuickGO
ancestors of GO:1990756: GO:0140767, GO:0030674, GO:0060090, GO:0003674).

Mechanism 4 — nuclear filaments

NSs self-associates and builds filaments without any other viral protein
PMID:10233964.
UniProt adds that the filaments are amyloid-like and may sequester NSs partners,
causing cell cycle arrest (PubMed:28915104, PubMed:32612175), and reports the crystal
structure of the core domain (PDB 5OOO) and the NMR structure of the ΩXaV peptide bound
to p62 (PDB 2N0Y).

Mechanism 5 — cytoskeleton and junctions (SigSci rows)

NSs prevents the antiviral upregulation of Abl2 and reshapes the actin cytoskeleton
PMID:33087469, with "reduction of lamellipodia,
cell spreading, and dissolution of adherens junctions" in hepatocytes (full text
cached). These effects are real but are downstream readouts of the transcriptional
block, so they are kept as non-core.

The CACAO rows

Six GOA rows carry ASSIGNED BY = CACAO, all dated 2014-04-02/05, from three papers.
The GONUTS wiki could not be consulted: https://gowiki.tamu.edu/wiki/index.php/PMID:15331713
and .../PMID:16501102 both return "This site is temporarily offline while we perform
upgrades", and the Internet Archive has no snapshot of either page
(archived_snapshots: {}). All three cited papers are abstract-only in the cache, so
PMC was used for figure-level reading of PMID:15331713 and the curator-deference rule was
applied throughout.

Term	Evid	PMID	Verdict
GO:0039602 suppression of host transcription initiation from Pol II promoter	IDA	14980221	Matches. The paper's mechanism is TFIIH disassembly, and TFIIH is a Pol II general transcription factor required for preinitiation-complex function. Right term, right granularity.
GO:0039653 suppression of host transcription	IMP	15331713	Matches. The paper shows inhibition of a constitutive promoter in addition to IFN-β, i.e. general transcription.
GO:0039657 suppression of host gene expression	IMP	16501102	Matches. ΔNSs viruses "failed to shut off host protein synthesis", so NSs is responsible for host shutoff; the gene-expression term is the right non-committal level for a protein-synthesis readout.
GO:0039502 suppression of host type I IFN-mediated signalling	IMP	15331713	Wrong branch, not wrong biology: the assays score IFN-β production (promoter reporters, IFN-β transcripts), not IFN-receptor signalling. Shared with the UniProt IEA and with influenza NS1/Ebola VP35, so it is a convention, not a student slip.
GO:0039502 suppression of host type I IFN-mediated signalling	IMP	16501102	Same wrong branch; the datum is IFN-β mRNA accumulation in ΔNSs infections.
GO:0039689 negative stranded viral RNA replication	IMP	16501102	Not supported, and contradicted by the cited abstract. The paper's point is that NSs is dispensable: "Mutants of MP-12 partially or completely lacking the NSs open reading frame were viable. These NSs deletion mutants replicated efficiently in Vero and 293 cells", and NSs expression from a plasmid actually "abolished virus rescue". Genome replication is performed by the L polymerase and N; NSs contributes none of the work of RNA synthesis.

The GO:0039689 row looks like a new failure shape for the CACAO taxonomy, distinct from
"phenotype read as process": a viral fitness/growth requirement read as participation
in genome replication. Here even the fitness premise fails in the cited cell lines —
the attenuation is IFN-dependent, and the paper reports it only in MRC-5 cells. It is
the same category error the repository's NEW-action guidance warns about (necessity
for an outcome mistaken for performing a step of it), applied to a term for an activity
that NSs demonstrably does not carry out.

The third case for the proposed production term, and a correction

The proposed symbiont-mediated interferon-production term (a counterpart of GO:0032480,
filed in proposed_new_terms) originally rested on two cases, RVFV NSs and Bunyamwera
NSs. Both shut down host RNA polymerase II globally, which is a narrow basis for a term
meant to be mechanism-agnostic, and an earlier draft of the justification recorded that
as an unresolved limitation.

The third case was in the same pull request all along. Sendai virus V (P69282) was
reviewed in the same batch and proposed the identical term independently. It is
mechanistically unlike the two NSs proteins: it does not shut down host transcription,
and every measured readout for it is an interferon-production readout. Its case is also
the cleanest demonstration of the gap, because no curator mis-annotated anything — the
absence of a production term shows up as nothing being recorded, rather than as a wrong
row to correct. The request now names all three cases and the duplicate proposal has been
removed from the Sendai review, so the term is counted and justified once.

A correction to the earlier reasoning. That draft declined to offer KSHV vIRF1 and
HSV-1 ICP0 as candidate third cases on the ground that both act on activated IRF3 and its
CBP/p300 coactivators, and so plausibly belong to GO:0039548 instead. That test does not
survive contact with the Sendai case: Sendai V carries GO:0039540 and GO:0039548 by direct
assay, both accepted in this project and both in its core functions, and it still needs a
term for its production phenotype. Applying the IRF3 test to vIRF1 and ICP0 while not
applying it to Sendai V was inconsistent. The honest reason for not citing those two is
simply that neither paper was read in full for this review.

What that change exposes is a scoping question rather than a weakness: if the production
term is warranted wherever interferon output is separately measured, it will layer over
much of the induction branch, coexisting with whichever sensor term describes the
mechanism. The request puts that question to the ontology editors explicitly, with the
criterion this review would apply — output measured in its own right, not inferred from a
mechanism term the gene already carries.

References consulted but not cached separately

PMID:19197350, PMID:19211744, PMID:19751406, PMID:21543505, PMID:23063407,
PMID:26837067, PMID:28915104, PMID:32612175, PMID:39366381 — all appear in the UniProt
reference block with ECO:0000269 evidence attached to the curated FUNCTION, SUBUNIT and
DOMAIN statements, and are used here only through those curated UniProt statements (cited
as file:RVFVZ/NSS/NSS-uniprot.txt) rather than as independent quotations.