A sixth paper systematically measured every human CXXC domain's DNA binding, deposited the MLL co-crystal, and failed all five retrieval routes
also called Xu 2018 Structure, unreadable · 4NW3's paper
One paper measured DNA binding for all the human proteins that carry this module and solved the crystal structure of the fusion's version bound to DNA. It is behind a publisher's wall and could not be read, so whatever numbers it holds are not in this analysis.
The same thing, for a scientist
PMID 29276034 (Xu C, Liu K, Lei M, et al., Structure 2018, 10.1016/j.str.2017.11.022) reports systematic measurement and comparison of the DNA-binding selectivities of all known human CXXC domains by several binding assays, plus a series of crystal structures including 4NW3, the MLL CXXC domain with CpG DNA at 2.82 A. It is the only source that could place the KMT2A CXXC affinity in a family-wide, single-laboratory, single-assay context - exactly the controlled comparison ana-cxxc-dna-affinity-ladder is assembled from cross-paper fragments to approximate. All five documented retrieval routes failed on 2026-09-21. This entry records the identifier and what the abstract asserts, marks the confidence weak, names the routes tried, and deliberately does not state what the paper probably says.
The five routes, and how each failed
Run 2026-09-21, in the order the brief prescribes 1:
| route | result |
|---|---|
1. Europe PMC core, EXT_ID:29276034&resultType=core | pmcid: None, isOpenAccess: N |
1b. cross-check efetch db=pubmed for <ArticleId IdType="pmc"> (the rule that route 1 can lie) | only pubmed, doi and pii - no pmc id. Route 1's negative is genuine here. |
2. {PMCID}/fullTextXML | cannot be attempted - no PMCID |
3. efetch db=pmc | cannot be attempted - no PMCID |
| 4. PMC HTML page | cannot be attempted - no PMCID |
5. curl -sL https://doi.org/10.1016/j.str.2017.11.022 with a browser User-Agent | HTTP 200, 11 characters after tag-stripping - the documented Elsevier redirect-stub failure |
5b. ScienceDirect directly by PII S0969-2126(17)30396-9 | HTTP 403 Forbidden |
So the failure is the rule-60 shape twice over: route 5 returns a well-formed 200 with a stub, and the direct publisher URL returns a clean 403. Both are loud failures and neither is ambiguous, which is the good case - a 13-character stub and a 27,000-character article both return 200, and measuring the stripped length is what distinguishes them.
Not attempted and worth attempting by someone with access: an institutional Elsevier subscription, an author-hosted copy, or the PDB's own 4NW3 primary-citation record for structure-level detail. Do not read the RCSB title as the PubMed title (the PDB trap already documented in this brief).
What can and cannot be said about it
Can be said, from the abstract fetched from PubMed: the authors systematically measured and compared the DNA-binding selectivities of all known human CXXC domains by different binding assays; they complemented the existing function-based classification with one based on DNA selectivity; and they solved a series of CXXC-DNA crystal structures including single CXXC domains and tandem CXXC-PHD domains.
Cannot be said, and must not be guessed: what Kd the MLL CXXC domain has in their hands, at what salt, with what construct, by which assay, or where it ranks among the twelve. That is precisely the information this beat most wants, and it is behind the wall.
What it would change if read. A single-laboratory, single-assay family-wide comparison would either corroborate Risner's rank order - MLL tightest, CGBP second, DNMT1 34-fold weaker - or contradict it. Risner's rank order is load-bearing for ana-affinity-rank-does-not-predict-cxxc-function, which in turn is the replacement for this project's withdrawn affinity argument. So one unreadable paper sits directly under the argument this beat substituted in, and that should be said out loud rather than left implicit.
Following the brief: identifier recorded, title asserted, confidence weak, routes named, content not invented. See also concept-known-but-unreadable-literature.
| pmid | 29276034 |
|---|---|
| doi | 10.1016/j.str.2017.11.022 |
| pii | S0969-2126(17)30396-9 |
| journal | Structure |
| year | 2018 |
| what the abstract asserts | systematic measurement and comparison of the DNA-binding selectivities of all known human CXXC domains by different binding assays; a DNA-selectivity-based classification complementing the existing function-based one; a series of crystal structures of CXXC domains with DNA ligands, including single CXXC domains and tandem CXXC-PHD domains |
| what it would settle | whether a single-laboratory single-assay family-wide comparison corroborates or contradicts Risner 2013's rank order (MLL tightest, CGBP second, DNMT1 about 34-fold weaker), which is load-bearing for ana-affinity-rank-does-not-predict-cxxc-function and therefore for the argument this beat substituted for the withdrawn affinity argument |
| deliberately not stated | any dissociation constant, construct, buffer, salt or rank order from this paper. None of it has been read. |
retrieval routes — the numbers behind the chart
[
{
"route": 1,
"endpoint": "Europe PMC search?query=EXT_ID:29276034&resultType=core",
"result": "pmcid None, isOpenAccess N",
"date": "2026-09-21"
},
{
"route": "1b",
"endpoint": "efetch db=pubmed, PubmedData/ArticleIdList/ArticleId",
"result": "only pubmed, doi and pii ids present; no pmc id, so route 1's negative is genuine",
"date": "2026-09-21"
},
{
"route": 2,
"endpoint": "{PMCID}/fullTextXML",
"result": "not attemptable - no PMCID"
},
{
"route": 3,
"endpoint": "efetch db=pmc",
"result": "not attemptable - no PMCID"
},
{
"route": 4,
"endpoint": "pmc.ncbi.nlm.nih.gov/articles/{PMCID}/",
"result": "not attemptable - no PMCID"
},
{
"route": 5,
"endpoint": "https://doi.org/10.1016/j.str.2017.11.022 with a browser User-Agent",
"result": "HTTP 200, 2,980 raw bytes, 11 characters after tag-stripping - the documented Elsevier redirect-stub failure",
"date": "2026-09-21"
},
{
"route": "5b",
"endpoint": "https://www.sciencedirect.com/science/article/pii/S0969212617303969",
"result": "HTTP 403 Forbidden",
"date": "2026-09-21"
}
]routes not attempted
[ "institutional Elsevier subscription", "author-hosted copy or interlibrary request", "extracting the construct boundaries from the 4NW3 deposited entity sequence via data.rcsb.org, which would give the construct span without the paper" ]
- Building a correction on five papers while a sixth on exactly the same question is unread is a real limitation of this beat, and the correction should be presented as provisional on that paper rather than as complete.
- What Kd does this paper report for the MLL CXXC domain, and under what conditions? It is the single most valuable unread number in this beat.
- Does its family-wide ranking agree with Risner 2013's? If it does not, the replacement argument for the withdrawn affinity claim needs re-examining.
- Can the 4NW3 deposited entity sequence be aligned to Q03164 to recover the construct boundaries without the paper? That is a one-call PDB query and would answer part of the question for free.
- E1weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper
PMID 29276034 has no PMC identifier in either Europe PMC or PubMed efetch, its DOI resolves to an 11-character stub after tag-stripping, and its ScienceDirect PII URL returns HTTP 403, so all five documented retrieval routes fail; its abstract states that the authors systematically measured and compared the DNA-binding selectivities of all known human CXXC domains by different binding assays and solved a series of CXXC-DNA crystal structures.
DNA Sequence Recognition of Human CXXC Domains and Their Structural Determinants. · Xu C, Liu K, Lei M, et al. · Structure 2018What this evidence cannot show
Everything recorded here comes from the PubMed record and the abstract fetched via Europe PMC core on 2026-09-21. No sentence of the paper's body has been read and nothing is inferred about its numbers. Routes tried and their exact failure modes are in data.retrieval_routes. Institutional access was not attempted and would very likely succeed - this is a paywall, not an absence.
Evidence that argues against this (1)
- This graph already records 4NW3's entry-level metadata from RCSB - 2.82 A, Q03164 CXXC with a His-TEV tagged construct, 12-mer duplex GCCATCGATGGC - fetched 2026-09-20 by the structural-interface beat. So the structural half of the paper is partially reachable through the PDB without reading the paper, and a beat with instrument time could extract the construct boundaries from the deposited entity sequence, which is the one thing this beat most needs from it. RCSB Protein Data Bank 2018
A sixth paper measured all human CXXC domains in one laboratory and could not be read; it would either corroborate or contradict the ladder's ordering.
A sixth on-topic paper found, on topic, and unreadable through all five documented routes.
If read, it would either corroborate or contradict the rank order this replacement argument rests on.
The same paper supplies 4NW3, which that entry already records from RCSB metadata.