NUCKS1 (Q9H1E3) - a RAD51AP1 paralogue that binds the same DNA with weaker affinity, and whose measured phenotype is mitomycin-C sensitivity rather than the D-loop binding this graph reached for
also called NUCKS1 · NUCKS · Q9H1E3 · NUCKS_HUMAN · Nuclear ubiquitous casein and cyclin-dependent kinase substrate 1 · RAD51AP1 paralogue
NUCKS1 is a near-copy of a known DNA-repair protein, binding the same kinds of DNA but less tightly. Removing it makes cells break their chromosomes and become very sensitive to a crosslinking drug, which is the same weakness this project already tracks in the Fanconi pathway.
The same thing, for a scientist
The second of three entries the six-member read-queue adjudication committed to, and the only member of that set with **fully coded nuclear and chromosomal localisation** (ECO:0000269). **Identity.** Q9H1E3, `NUCKS_HUMAN`, **243 aa**, *Nuclear ubiquitous casein and cyclin-dependent kinase substrate 1*; described in the primary literature as a 27 kD chromosomal, **vertebrate-specific** protein for which limited functional data existed. Locations: Nucleus (ECO:0000269, twice) and Chromosome (ECO:0000269). Tissue: widely expressed, highest in thyroid, prostate and uterus and in fetal liver, thymus and lung. **It is a RAD51AP1 paralogue, and the weaker one.** Parplys et al. showed NUCKS1 shares extensive sequence homology with RAD51AP1, and that recombinant NUCKS1 has **the same DNA binding preference as RAD51AP1 but binds with reduced affinity**. Depleting NUCKS1 in human cells impairs homologous recombination and chromosome stability, and functional loss slows replication-fork progression with a concomitant **increase in the use of new replication origins**. **The measured phenotype is crosslink sensitivity, and that is where it touches this graph.** NUCKS1 depletion causes *"greatly increased cellular sensitivity to mitomycin C"* and increased levels of **spontaneous and MMC-induced chromatid breaks**. MMC is the canonical interstrand-crosslinking agent, and this project already holds an ICL/Fanconi beat - `cell-state-replication-stress-icl-repair`, `ana-fa-aldehyde-vs-replication` and the FA pan-essentiality risk entries. **A CORRECTION TO MY OWN ADJUDICATION, MADE THE SAME CYCLE.** The analysis that committed to writing this entry justified it partly on NUCKS1 binding *"secondary DNA structures, such as D-loop structures"* and said that D-loop binding *"sits between the break-repair beat and the R-loop beat"*. That phrase is UniProt's, and it has **no title-or-abstract literature behind it**: `(NUCKS1 OR NUCKS) AND "D-loop"` returns **0**, a genuine zero with a parsed querytranslation and warninglist, and `AND "R-loop"` returns **0** the same way. The primary paper's abstract does not mention D-loops. So the D-loop detail is a curated phrase, the R-loop link I speculated about has an empty literature, and the entry is earned by the HR and MMC phenotypes instead. **Never studied in this disease.** `(NUCKS1 OR NUCKS) AND (MLL OR KMT2A)` returns **0**, a genuine zero; `AND leukemia` returns 2. **And a near-disjoint alias split with a tiny literature.** `NUCKS1[tiab]` returns 96 records and `NUCKS[tiab]` 47, intersecting in **one**. This is a thinly studied protein by any measure, and its dependency is the flattest of the six queue members at −0.0091.
| uniprot | Q9H1E3 |
|---|---|
| uniprot entry name | NUCKS_HUMAN |
| gene symbol | NUCKS1 |
| protein name | Nuclear ubiquitous casein and cyclin-dependent kinase substrate 1 |
| length AA | 243 |
| molecular weight as described in the literature | 27 kD; chromosomal; vertebrate-specific |
| domains note | UniProt Q9H1E3 annotates no DOMAIN features. The functional description is by homology to RAD51AP1 rather than by domain architecture, so no coordinates are asserted. Empty array rather than an omitted field, because the viewer requires these to be arrays. |
| localization | UniProt Q9H1E3: Nucleus [ECO:0000269 x2]; Chromosome [ECO:0000269]. The only member of the six adjudicated read-queue candidates with fully CODED nuclear and chromosomal localisation. |
| subnuclear localization | CHROMATIN-ASSOCIATED, described as such in the primary paper and consistent with the coded Chromosome assignment. No compartment finer than that - no nucleolar, speckle or heterochromatin assignment exists. NOT measured in any KMT2A-rearranged cell. |
| localization shown for fusion | false |
| cell type | No KMT2A-rearranged cell and no haematopoietic cell. The HR and MMC work is in unspecified human cells; the tissue-expression record is bulk tissue, highest in thyroid, prostate and uterus and in fetal liver, thymus and lung. |
| cell state | REPLICATION STRESS and DNA DAMAGE are the states in which it has a measured phenotype: loss slows fork progression and increases new-origin firing, and sensitises cells to mitomycin C with more spontaneous and MMC-induced chromatid breaks. Nothing measured under a menin or DOT1L inhibitor, and nothing across the cell cycle despite the protein being named for being a CDK substrate. |
| epigenetic mark named | None. It is chromatin-associated but writes, reads and erases no histone mark. |
| epigenetic readout | Not applicable. Its route to this graph's chromatin layer is structural and repair-based, not mark-based. |
| shown for fusion | false |
| absence class | never-measured |
domains
[]
the paralogue pair
{
"partner": "RAD51AP1",
"relationship": "extensive sequence homology; the primary paper calls them paralogues",
"the_functional_asymmetry": "recombinant NUCKS1 shares the SAME DNA binding preference as RAD51AP1 but binds DNA with REDUCED affinity",
"phenotypic_similarity": "NUCKS1 depletion phenocopies RAD51AP1 knockdown for impaired HR and chromosome instability",
"why_this_matters_for_this_project": "a paralogue pair where one member is the weaker version of the other is exactly the shape this project's registry warns about - a swap between them would not produce an obviously wrong entry, it would produce a quantitatively wrong one. gene-rad51ap1 does not exist in this graph."
}the measured phenotypes
{
"homologous_recombination": "depletion impairs DNA repair by HR and chromosome stability; NUCKS1 is described as critical to maintaining wild-type HR capacity",
"crosslink_sensitivity": "greatly increased cellular sensitivity to mitomycin C, with increased spontaneous AND MMC-induced chromatid breaks",
"replication": "functional loss slows replication fork progression with a concomitant increase in the utilisation of new replication origins",
"which_connects_where": "MMC is the canonical interstrand-crosslink agent, so the crosslink phenotype connects to this graph's ICL/Fanconi beat rather than to its R-loop beat"
}correction to my own adjudication same cycle
{
"what_i_wrote": "that NUCKS1 binds 'secondary DNA structures such as D-loop structures' and that D-loop binding 'sits between the break-repair beat and the R-loop beat'",
"the_problem": "the D-loop phrase is UniProt's curation and the primary paper's abstract does not mention D-loops",
"measured": {
"(NUCKS1 OR NUCKS) AND \"D-loop\"[tiab]": "0 - GENUINE zero, parsed querytranslation with warninglist 'No items found.'",
"(NUCKS1 OR NUCKS) AND \"R-loop\"[tiab]": "0 - GENUINE zero, same form",
"(NUCKS1 OR NUCKS) AND \"homologous recombination\"[tiab]": 4
},
"so": "the R-loop connection I reached for has an empty literature under either token, and the entry stands on the HR and MMC phenotypes instead. The adjudication entry has been amended."
}never studied in this disease
{
"(NUCKS1 OR NUCKS) AND (MLL OR KMT2A)": "0 - GENUINE zero",
"(NUCKS1 OR NUCKS) AND leukemia": 2
}alias split measured
{
"NUCKS1[tiab]": 96,
"NUCKS[tiab]": 47,
"\"nuclear ubiquitous casein\"[tiab]": 22,
"intersection NUCKS1 and NUCKS": 1,
"reading": "near-disjoint on a tiny total literature. Both tokens were run per the registry rule that the alias question has no default direction; here the approved symbol is the larger set but the overlap is a single record, so neither token alone describes the literature.",
"a_token_tested_and_discarded": "P1[tiab] returns 42,355 records and is unusable as an operand for this gene. It was tested speculatively and is NOT asserted here to be a UniProt-listed synonym - the UniProt record returned no gene synonyms for NUCKS1."
}depmap
{
"mean_gene_effect": -0.0091,
"frac_below_minus_0_5": 0.0135,
"n_lines": 1258,
"reading": "the flattest of the six adjudicated queue members. Recorded because the adjudication's whole point was that a flat gene effect is not a rejection ground, not because flatness supports anything.",
"depmap_release": null,
"rule_83": "the Open Targets depMapEssentiality payload carries no release identifier; figure measured in cycle 58 with all tissue blocks flattened and a payload guard at the panel maximum"
}- Nothing here is a target nomination. NUCKS1's dependency is the flattest of the six candidates adjudicated this cycle (-0.0091, 1.4% of 1,258 lines below -0.5), it has never been studied in this disease - a genuine zero against MLL and KMT2A - and its interaction evidence in this project's census is a single atlas-weight observation.
- The one clinically-adjacent fact points away from intervention rather than towards it. NUCKS1 loss sensitises cells to mitomycin C and raises chromatid breaks, so inhibiting it would be a genotoxic strategy in a child, and this graph's own Fanconi entries already record that crosslink-repair pathways are pan-essential enough to make that window narrow.
- Does NUCKS1 bind D-loops, and is that in the body of PMID 26323318? UniProt states it with a citation and the abstract does not mention it, so the full text would settle whether the claim is supported or is a curation extrapolation from 'secondary DNA structures'.
- Why does `gene-rad51ap1` not exist in this graph? NUCKS1 is written here as its weaker paralogue and the stronger member of the pair is unwritten - and a paralogue pair where one is quantitatively weaker is precisely the swap-risk shape this project's registry warns about.
- Does the mitomycin-C sensitivity connect NUCKS1 to this graph's ICL/Fanconi beat mechanistically, or only phenotypically? MMC sensitivity is shared by many HR factors and is not specific to the FA pathway.
- Is NUCKS1 a CDK substrate in a way that matters? It is named for being a casein-kinase and CDK substrate and nothing in this entry's evidence touches its phosphorylation, in any cell state.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
NUCKS1 is a RAD51AP1 paralogue with the same DNA-binding preference but reduced affinity, required for normal homologous recombination, and its loss confers strong mitomycin-C sensitivity. NUCKS1 is a 27 kD chromosomal, vertebrate-specific protein that shares extensive sequence homology with RAD51AP1, and depletion of NUCKS1 in human cells impairs DNA repair by homologous recombination and chromosome stability, phenocopying RAD51AP1 knockdown. Depletion also results in greatly increased cellular sensitivity to mitomycin C and in increased levels of spontaneous and MMC-induced chromatid breaks. Functional loss of NUCKS1 slows DNA replication fork progression with a concomitant increase in the utilisation of new replication origins. Recombinant NUCKS1 shares the same DNA binding preference as RAD51AP1 but binds DNA with reduced affinity.
NUCKS1 is a novel RAD51AP1 paralog important for homologous recombination and genome stability. · Parplys AC, Zhao W, et al. · Nucleic Acids Research 2015What this evidence cannot show
Abstract-level; no figure or methods section was read, so the depletion method, the magnitude of the MMC sensitisation and the fork-speed measurement are not established here. The cell systems are not named in the abstract, so nothing is established in a haematopoietic or KMT2A-rearranged cell. 'Paralogue' is the authors' inference from sequence homology and shared phenotype, not from a structure or a phylogeny presented in the abstract. This is a 2015 single-paper basis for the protein's principal function, and the total indexed literature for this gene is fewer than 150 records under either token.
Evidence that argues against this (2)
- Reduced affinity for the same substrate as an existing paralogue is an argument that NUCKS1 is the dispensable member of the pair rather than an independent factor - consistent with its Chronos gene effect of -0.0091 and 1.4% of 1,258 lines below -0.5. A protein that does what RAD51AP1 does, less well, and whose loss almost no cell line minds, is a weak basis for anything mechanistic in this disease.
- The replication phenotype points the opposite way to this graph's usual framing. Loss of NUCKS1 SLOWS forks and increases new-origin firing; this project's break model is about transcription-replication conflict generating breaks, and a slower fork with more origins is not straightforwardly more or less conflict-prone. Nothing here licenses a directional prediction at the KMT2A locus.
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
The D-loop and R-loop connections this graph reached for have no title-or-abstract literature, while the crosslink connection does. Measured on ID sets with both tokens: (NUCKS1[tiab] OR NUCKS[tiab]) AND "D-loop"[tiab] returns 0 and AND "R-loop"[tiab] returns 0, both genuine zeros with non-empty querytranslations and warninglist entries; AND "homologous recombination"[tiab] returns 4; AND (MLL[tiab] OR KMT2A[tiab]) returns 0, a genuine zero; AND leukemia[tiab] returns 2. The alias sets are near-disjoint: NUCKS1[tiab] 96 records, NUCKS[tiab] 47, intersecting in one.
What this evidence cannot show
Title-and-abstract scope only, so a D-loop binding assay reported in a figure or a supplementary file would not be seen - and UniProt's D-loop phrase presumably rests on exactly such data in the body of PMID 26323318, which was not read. The correct reading is therefore that the D-loop claim is not RETRIEVABLE at title-abstract level, not that it is unsupported. The zeros are small-set zeros on a gene with under 150 total records, where absence is weak evidence either way.
Evidence that argues against this (1)
- This measurement was used to correct my own adjudication and it may be too harsh on it. UniProt curators state the D-loop binding with a citation, and a curated statement from a reviewed record is ordinarily good enough for this graph - the standard applied here, that a claim must also be reachable by a title-abstract query, is stricter than the one used elsewhere in the corpus and is not applied consistently.
same adjudication, opposite ends of its dependency range - HMGA1 required in 90.9% of 1,258 lines, NUCKS1 in 1.4% - and both kept on function rather than on number
the entry that analysis commissioned corrected it: the D-loop detail it leaned on is a UniProt phrase with a genuine zero at title-abstract level, and the R-loop link was speculation with an empty literature
the phenotype that actually connects: NUCKS1 loss greatly increases mitomycin-C sensitivity with more spontaneous and MMC-induced chromatid breaks, and MMC is the canonical interstrand-crosslink agent
- related to →Neither rival mechanism leaves its fingerprint, and the best correlate of Fanconi dependency is menin
same crosslink-repair axis; NUCKS1 is an HR factor whose loss phenocopies crosslink hypersensitivity without being an FA gene
two of the three read-queue members kept on function rather than on dependency, from the same adjudication - NUCKS1 the flattest of the six at -0.0091, ASB7 the second flattest at -0.0457