EIN3 → Response to Hypoxia (GO:0001666): Hypothesis Adjudication Report

Gene: EIN3 (ETHYLENE INSENSITIVE 3), Arabidopsis thaliana
UniProt: O24606 (EIN3_ARATH) · Locus: At3g20770
Focus: function_assignment — existing_annotations[16].function_hypothesis
Seed hypothesis: EIN3 has response to hypoxia (GO:0001666)
Annotation under review: GO:0001666, evidence IMP, reference PMID:25284079(https://pubmed.ncbi.nlm.nih.gov/25284079/), qualifier acts_upstream_of_or_within, assigned by TAIR

Summary

Executive verdict — source-specific: REFUTED / over-annotated. Overall biological: weakly / partially supported (indirect, redundant, upstream).

The EIN3 → response to hypoxia (GO:0001666) IMP annotation does not survive independent reading of its cited source. PMID:25284079(https://pubmed.ncbi.nlm.nih.gov/25284079/) (Xie et al., 2015, The Plant Journal) is a study of the acyl-CoA-binding protein ACBP3, not of EIN3. In that paper ein3-1 appears only as one genotype in an epistasis panel built around an ACBP3 overexpressor (OE-1). The decisive experiment (Figure 6b) shows that the ein3-1 single mutant was not significantly different from wild type under either dark submergence (DS) or light submergence (LS), and that the OE-1 ein3-1 double combination still resembled the hypoxia-sensitive OE-1 parent. In other words, removing EIN3 neither produced a hypoxia phenotype on its own nor modified the ACBP3 phenotype. The authors' own conclusion attributes the attenuated hypoxic tolerance of ACBP3 overexpressors to NPR1- and CTR1-mediated signaling, explicitly not to EIN3. An IMP (Inferred from Mutant Phenotype) annotation requires a mutant phenotype; here the mutant phenotype was null. The paper's proposed EIL1 redundancy was invoked to explain the null result but never tested — no ein3 eil1 double mutant was assayed for hypoxia in this work. This is a textbook case of a database carry-over annotation where the qualifying phenotype does not exist.

Separately from the source, the broader biology offers modest, indirect support for an EIN3/EIL1 role in low-oxygen adaptation, but only as an upstream ethylene-signaling step and only in redundancy with EIL1. Hartman et al., 2019 (PMID:31488841(https://pubmed.ncbi.nlm.nih.gov/31488841/)) show that ethylene perception on submergence pre-adapts plants to hypoxia by depleting nitric oxide (via PHYTOGLOBIN1) and thereby stabilizing group VII ERF transcription factors; this pathway runs through EIN2 and the redundant EIN3/EIL1 pair, and EIN3-GFP nuclear accumulation is used only as a readout of ethylene perception. Crucially, direct oxygen sensing and the transcriptional activation of hypoxia-responsive genes are executed by the N-degron pathway and group VII ERFs (RAP2.2/2.12/2.3, HRE1/2; PMID:26668304(https://pubmed.ncbi.nlm.nih.gov/26668304/)), not by EIN3.

Bottom line for the curator: The specific IMP annotation to PMID:25284079 should be removed or re-sourced, because its cited experiment reports a null EIN3 phenotype. If any hypoxia link is to be retained for EIN3, it should be a non-core, indirect BP annotation citing Hartman et al. 2019, with a qualifier that does not imply direct oxygen sensing, and it should acknowledge EIN3/EIL1 redundancy. No molecular-function (MF) term implying oxygen sensing is warranted. Identity was verified: O24606 is EIN3 (At3g20770), distinct from its paralog EIL1 (At2g27050), so there is no paralog misassignment — the annotation is on the correct protein but rests on inadequate evidence.

Key Findings

F001 — The cited source (PMID:25284079) reports a null ein3-1 hypoxia phenotype; it does not support the IMP annotation

PMID:25284079(https://pubmed.ncbi.nlm.nih.gov/25284079/) (Xie et al., 2015, Plant J, full text PMC4309432) is an ACBP3 paper — "Arabidopsis acyl-CoA-binding protein ACBP3 participates in plant response to hypoxia by modulating very-long-chain fatty acid metabolism." EIN3 is not the subject of the study. The ein3-1 mutant enters only as one of several genotypes in an epistasis screen designed to place the ACBP3 overexpressor (OE-1) hypoxia-sensitivity phenotype within a signaling network.

The relevant result is in the Results subsection "Hypoxia hypersensitivity in OE-1 is dependent on NPR1 and CTR1" and Figure 6b. Two verbatim observations define the evidence:

"the ein3-1 mutant was not significantly different to wild type under either DS or LS (Figure 6b), possibly due to the functional redundancy of the EIL1 gene in Arabidopsis."

"the OE-1 coi1-2 and OE-1 ein3-1 double combinations were hypoxia sensitive, resembling the OE-1 phenotype."

Both statements point the same way. First, the ein3-1 single mutant had no hypoxia phenotype (no significant difference from wild type under either dark or light submergence). Second, adding ein3-1 to the ACBP3 overexpressor did not rescue or modify the OE-1 hypoxia-sensitivity phenotype — the double still behaved like OE-1. Both are the signatures of a gene that is not required for the phenotype under assay. The abstract's own conclusion makes the exclusion explicit:

"we observed that the attenuated hypoxic tolerance in ACBP3-OEs was dependent on NPR1- and CTR1-mediated signaling pathways"

EIN3 is not named as a required node; NPR1 and CTR1 are.

Methodologically, the annotation is further weakened by the quality of the negative result reported. No survival counts, no sample size (n), and no explicit statistical test are reported for the ein3-1 comparison — the claim is a qualitative "not significantly different." An IMP (Inferred from Mutant Phenotype, ECO:0000315) annotation is meant to record a phenotype conferred by loss of the gene. Here the loss-of-function mutant conferred no detectable phenotype. Using a null result to justify a positive functional annotation inverts the logic of IMP evidence.

F002 — The proposed EIL1 redundancy was invoked, not tested; independent biology supports only an indirect, redundant, upstream role

The paper itself hedges the null ein3-1 result by "possibly due to the functional redundancy of the EIL1 gene." That is a hypothesis to explain absence of phenotype, not evidence of function. No ein3 eil1 double mutant was assayed for hypoxia tolerance anywhere in PMID:25284079. Redundancy invoked to rescue a null result cannot be used to support the very annotation the null result fails to justify.

However, redundancy of EIN3/EIL1 in hypoxia adaptation is independently supported by a different, stronger study. Hartman et al., 2019 (PMID:31488841(https://pubmed.ncbi.nlm.nih.gov/31488841/), Nat Commun) demonstrate that ethylene perception pre-adapts Arabidopsis to subsequent hypoxia:

"This ethylene-mediated NO depletion and consequent ERFVII accumulation pre-adapts plants to survive subsequent hypoxia"

Mechanistically, entrapment of ethylene during submergence is perceived through the canonical pathway (EIN2 → EIN3/EIL1), leading to PHYTOGLOBIN1-mediated depletion of nitric oxide, which stabilizes group VII ERF transcription factors (ERF-VIIs) and thereby pre-adapts the plant. In that study the ein3eil1-1 double mutant is among the ethylene-signaling mutants in which the pre-adaptation response is abolished, and EIN3-GFP nuclear accumulation is used as a readout of ethylene perception on submergence. This places EIN3/EIL1 upstream of the oxygen-sensing machinery as part of ethylene signal transduction — it is not itself an oxygen sensor or a direct activator of the core hypoxia-responsive gene battery. The requirement is genetically satisfied only by the double mutant, consistent with EIN3 and EIL1 being redundant.

Thus the overall biological picture allows a defensible but weak, indirect claim: EIN3, redundantly with EIL1, participates in ethylene-mediated pre-adaptation to hypoxia. That claim is not the one tested by PMID:25284079, and it is BP-level, indirect, and non-core.

F003 — Identity verified: O24606 = EIN3 (At3g20770); no paralog misassignment, but no oxygen-sensing molecular function

UniProt REST (O24606.json) returns uniProtKBId = EIN3_ARATH, protein name "ETHYLENE INSENSITIVE 3", gene EIN3, ordered locus At3g20770. The functional keywords are Activator, Chromatin regulator, DNA-binding, Ethylene signaling pathway, Nucleus, Transcription regulation — the profile of a nuclear DNA-binding transcriptional activator in the ethylene pathway. There is no oxygen-sensing, dioxygenase, or hypoxia molecular-function keyword. Among UniProt's GO cross-references, exactly one hypoxia/oxygen-related term appears: GO:0001666 response to hypoxia, with evidence code IMP:TAIR — i.e., the very annotation under review, and the only hypoxia link on the record.

The paralog EIL1 is a distinct gene at a distinct locus (At2g27050). Therefore the annotation sits on the correct protein — there is no confusion of EIN3 with EIL1 at the identity level. The problem is not misassigned identity; it is inadequate evidence for the assigned process.

F004 — QuickGO confirms a single, TAIR-only annotation with a loose qualifier and no orthology support

The EBI QuickGO annotation API (geneProductId=O24606, goId=GO:0001666) returns numberOfHits = 1. That single annotation has: evidence IMP (ECO:0000315), reference PMID:25284079, assignedBy = TAIR, qualifier acts_upstream_of_or_within, withFrom = null, taxon 3702. There is no second, independent annotation of GO:0001666 for EIN3, and no with/from ortholog transfer supporting it. The annotation is therefore a single-source record whose entire evidentiary weight depends on the one paper analyzed above — and that paper reports a null.

The acts_upstream_of_or_within qualifier is itself telling: it is the loosest causal relationship in the GO relations vocabulary, signaling that even the curator did not assert a direct role. That loose qualifier is consistent with — but does not remedy — the underlying null phenotype.

Mechanistic Model / Interpretation

The core question is whether EIN3 directly has the function "response to hypoxia," or whether it is a bystander/upstream node that was captured by an epistasis screen and then over-annotated. The evidence supports the latter. The two pathways involved should be kept sharply distinct:

  DIRECT OXYGEN SENSING & HYPOXIA-GENE ACTIVATION  (EIN3 NOT a member)
  ─────────────────────────────────────────────────────────────────
   O2 / NO ──► N-degron pathway (PCO oxidation of Cys) ──► ERF-VII
                                          stability
   ERF-VIIs (RAP2.2, RAP2.12, RAP2.3, HRE1, HRE2)  ──►  ~50 hypoxia-
     bind Hypoxia-Responsive Promoter Element (HRPE)     responsive
                                         genes (HRGs)
                        [PMID:26668304, 41732087]

  UPSTREAM ETHYLENE SIGNALING  (where EIN3/EIL1 acts — indirectly)
  ─────────────────────────────────────────────────────────────────
   Submergence ──► ethylene entrapment ──► ETR1/CTR1 ──► EIN2
│                                                  │
│                                                  ▼
│                              EIN3 / EIL1 (redundant TFs, nucleus)
│                                                  │
└──► PHYTOGLOBIN1 ──► NO depletion ──► ERF-VII stabilization
                    "pre-adaptation to hypoxia"
                        [PMID:31488841 (Hartman 2019)]

EIN3's authenticated molecular function (F003) is that of a nuclear DNA-binding transcriptional activator in the ethylene pathway. Its documented connection to hypoxia is upstream and indirect: as part of ethylene signal transduction that can pre-adapt the plant by feeding into ERF-VII stability. It is not a component of the oxygen-sensing N-degron machinery and not one of the ERF-VII activators that directly bind hypoxia-responsive promoters.

Mapping this onto GO evidence logic:

The single IMP annotation conflates the "upstream relay" layer (weakly, indirectly true) with a direct "response to hypoxia" claim, and sources it to a paper whose only EIN3 experiment produced a null. That is the definition of an over-annotation.

Evidence Matrix

GO Curation Implications

Lead (requires curator verification): remove or re-source the GO:0001666 IMP annotation on EIN3 (O24606).

• Term type: GO:0001666 response to hypoxia is a Biological Process term. There is no molecular-function claim in play, and none is warranted — EIN3 has no oxygen-sensing or dioxygenase activity (F003). Do not add any MF term implying direct oxygen sensing.
• Action on the existing annotation: The IMP evidence code requires a mutant phenotype attributable to the gene. The cited paper's ein3-1 mutant showed no phenotype and did not modify the ACBP3 phenotype (F001). The evidence therefore does not meet the IMP standard. Recommended action: remove the annotation, or at minimum re-source it.
• If a hypoxia link is retained for EIN3: cite Hartman et al. 2019 (PMID:31488841(https://pubmed.ncbi.nlm.nih.gov/31488841/)) instead, framed as a non-core, indirect BP role in ethylene-mediated pre-adaptation, with an acts_upstream_of relationship and explicit recognition of EIN3/EIL1 redundancy (ideally supported by the ein3eil1 double-mutant evidence, not ein3-1 alone). This is an IGI/IMP-on-double-mutant style claim, not an IMP on the single mutant.
• Qualifier: The existing loose acts_upstream_of_or_within qualifier already signals a non-direct role; it is consistent with removal or with a re-sourced, indirect BP annotation, but it does not repair the missing phenotype in the current source.
• Avoid "protein binding": Not applicable; the informative framing is transcriptional activator in the ethylene signaling pathway (already captured by other EIN3 annotations).

GO decision table

Mechanistic Scope

• Immediate molecular function of EIN3 (direct): sequence-specific nuclear DNA-binding transcriptional activator in the ethylene signaling pathway (UniProt keywords: Activator, DNA-binding, Nucleus, Ethylene signaling). This is the gene product's authenticated core activity.
• What GO:0001666 asserts: participation in the response to hypoxia biological process. For EIN3 this is at best an upstream, indirect contribution via ethylene signal transduction (EIN2 → EIN3/EIL1), feeding into ERF-VII stabilization (Hartman 2019).
• Downstream / not-direct effects that must not be conflated with EIN3 activity: the actual oxygen-sensing step (N-degron/PCO), the direct transcriptional activation of the hypoxia-responsive gene battery (ERF-VIIs binding the HRPE motif), and the whole-plant submergence-survival phenotypes assayed in PMID:25284079 (which are driven by ACBP3/very-long-chain fatty acid metabolism and NPR1/CTR1 signaling, not EIN3).
• Effects inferred only from loss of function: none usable here — the ein3-1 loss-of-function produced no hypoxia phenotype, so no function can be inferred from it.

Conflicts and Alternatives

1. The source is not an EIN3 study. PMID:25284079 is an ACBP3 paper; EIN3 enters only as an epistasis control. Database carry-over of a term from a secondary genotype is the most likely origin of this annotation.
2. Null result mislabeled as IMP. The mutant phenotype that IMP requires is absent (ein3-1 = WT under submergence). This is the central conflict with the seed hypothesis.
3. Redundancy invoked, not demonstrated. EIL1 redundancy is a rescue narrative for the null in PMID:25284079; the double mutant was not tested there. Genuine redundancy evidence exists elsewhere (Hartman 2019) but for a different, upstream process.
4. Competing direct actors. The direct transcriptional response to hypoxia is executed by group VII ERFs (RAP2.2/2.12/2.3, HRE1/2; PMID:26668304(https://pubmed.ncbi.nlm.nih.gov/26668304/)), not EIN3. Attributing "response to hypoxia" to EIN3 risks implying it is one of these direct activators, which it is not.
5. No paralog/identity artifact. Identity is clean: O24606 = EIN3 (At3g20770) ≠ EIL1 (At2g27050). The problem is evidentiary, not identity-based.

Limitations and Knowledge Gaps

• No raw data / no reanalysis possible. PMID:25284079 reports the ein3-1 comparison qualitatively ("not significantly different"). No survival counts, sample size, or test statistic are given, and no deposited raw dataset was located, so an independent reproducible reanalysis of the ein3-1 phenotype could not be performed. The verdict rests on the authors' own reported null and stated conclusions, read independently. Gap: the quantitative strength of the null (was the study powered to detect a small EIN3 effect?) is unknown; a properly powered ein3-1 and ein3 eil1 submergence assay would resolve it. Missing evidence is not refutation of the broader biology — only of the specific IMP claim.
• Direction of the Hartman 2019 requirement. The ein3eil1-1 double abolishes ethylene-mediated pre-adaptation, but the study's primary aim was the ethylene→NO→ERF-VII axis; the quantitative EIN3/EIL1 contribution versus other ethylene-pathway components is not fully partitioned. Gap: an ein3 eil1 hypoxia-tolerance survival curve with ERF-VII target readouts would quantify the contribution.
• Curation provenance vs. curator intent. QuickGO confirms a single TAIR IMP with a loose qualifier, but the internal TAIR curation note that led to term assignment was not retrievable. Gap: the original TAIR curation rationale, if it cited a figure other than 6b, should be checked before deletion.
• Single-species scope. All primary evidence is Arabidopsis; ortholog behavior (e.g., rice EIL1/2 under submergence) may differ and is not with/from-linked here.

Discriminating Tests

The following would most efficiently separate "EIN3 directly has response to hypoxia" from "EIN3 is an upstream, redundant, indirect node":

1. Powered ein3-1, eil1, and ein3 eil1 submergence survival assays (dark and light), with explicit n and statistics, plus reoxygenation recovery. Prediction: single mutants ≈ WT; only the double shows reduced tolerance if the pre-adaptation role is real.
2. ERF-VII stability / hypoxia-marker readout in ein3 eil1 (e.g., ADH1, PDC1, SUS4 induction; RAP2.12 protein) under submergence to test whether EIN3/EIL1 loss impairs the ethylene→ERF-VII pre-adaptation relay (as Hartman 2019 predicts) without EIN3 binding the HRGs directly.
3. ChIP-seq / DAP-seq of EIN3 across hypoxia-responsive promoters to test directly whether EIN3 binds the core HRG battery (expected: it does not, in contrast to ERF-VIIs and their HRPE motif).
4. Epistasis re-test of ACBP3 with ein3 eil1 double, since PMID:25284079 only used ein3-1; this would test whether redundancy actually masks a contribution to the ACBP3 phenotype (expected: still NPR1/CTR1-dependent).

Curation Leads (require curator verification)

Candidate action: Remove GO:0001666 IMP (PMID:25284079) from EIN3, or re-source it as a non-core indirect BP.

Candidate reference snippets to verify in the source (PMID:25284079):
- Results, Fig. 6b: "the ein3-1 mutant was not significantly different to wild type under either DS or LS (Figure 6b), possibly due to the functional redundancy of the EIL1 gene in Arabidopsis."
- Results: "the OE-1 coi1-2 and OE-1 ein3-1 double combinations were hypoxia sensitive, resembling the OE-1 phenotype."
- Abstract: "we observed that the attenuated hypoxic tolerance in ACBP3-OEs was dependent on NPR1- and CTR1-mediated signaling pathways" (EIN3 not named as required).

Candidate replacement reference (if a hypoxia link is retained): PMID:31488841(https://pubmed.ncbi.nlm.nih.gov/31488841/) — snippet to verify: "This ethylene-mediated NO depletion and consequent ERFVII accumulation pre-adapts plants to survive subsequent hypoxia." Frame as indirect, redundant (EIN3/EIL1), acts_upstream_of, non-core BP.

Candidate competing-context reference (for scope, not for EIN3 annotation): PMID:26668304(https://pubmed.ncbi.nlm.nih.gov/26668304/) — identifies the actual direct hypoxia transcriptional activators (ERF-VIIs), which EIN3 is not.

Suggested curator questions:
- Does the TAIR curation note point to any figure other than 6b? If not, the IMP basis is the null.
- Should the annotation be transferred to EIN3/EIL1 as a redundant pair with double-mutant evidence rather than to EIN3 alone?

Suggested experiments: the four discriminating tests above, prioritizing the powered ein3 eil1 submergence survival + ERF-VII readout.

Final Verdicts

• Source-specific GO verdict (PMID:25284079): REFUTED / over-annotated. The cited experiment reports a null ein3-1 phenotype and attributes the hypoxia phenotype to NPR1/CTR1, not EIN3; EIL1 redundancy was invoked but untested. The IMP evidence standard is not met.
• Overall biological verdict: Weakly / partially supported. An indirect, redundant, upstream EIN3/EIL1 role in ethylene-mediated pre-adaptation to hypoxia is defensible (Hartman 2019), but EIN3 is not an oxygen sensor and not a direct activator of hypoxia-responsive genes.
• Narrowest defensible function: EIN3 acts, redundantly with EIL1, as an upstream ethylene-signaling transcription factor that can contribute to pre-adaptation to hypoxia — i.e., acts_upstream_of the low-oxygen response, non-core.
• What would change the verdict: a properly powered, statistically reported ein3-1 (or ein3 eil1) submergence-survival experiment showing a reproducible loss-of-function hypoxia phenotype attributable to EIN3 — ideally with direct evidence that EIN3 binds/activates hypoxia-responsive genes. Absent that, the specific IMP annotation remains unsupported by its source.

Artifacts

• OpenScientist EIN3 hypoxia GO decision table
• OpenScientist EIN3 hypoxia evidence matrix
• OpenScientist final report
• OpenScientist final report