BCR — domain architecture and residue coordinates

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

• Targeting the oligomerisation interface removes an interface that wild-type BCR also uses; BCR is expressed ubiquitously, so selectivity for the fusion cannot come from the interface itself, only from the dependency.
• Because the C2 domain is left as a truncated fragment rather than cleanly removed, p210 carries a partly unfolded segment that could act as a degron or a chaperone sink; that is a route by which a stabilising drug could inadvertently raise fusion levels.

1. Does the BCR oligomerisation element form the same swapped-helix tetramer inside full-length p210 as it does as an isolated 72-residue peptide, or does the rest of BCR change the stoichiometry?
2. p210 keeps BCR's RhoGEF module and discards its Rho-GAP domain. Does that one-sided retention contribute anything to CML biology, or is the GEF module a passenger that happens to lie 5' of the breakpoint?
3. Where exactly does the coiled coil end? A structure of residues 1-120 rather than 1-72 would settle whether UniProt's 28-55 or the domain databases' ~1-75 is the functionally relevant unit.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   BCR is 1,271 residues with a coiled coil at 28-55, an ABL-SH2-binding region at 197-385, DH 498-691, PH 708-866, C2 893-1020, Rho-GAP 1054-1248, and an annotated BCR::ABL translocation site between residues 426 and 427.

   UniProtKB entry P11274 (BCR_HUMAN): reviewed sequence, length and positional feature annotations · UniProt Consortium · UniProtKB/Swiss-Prot 2026
   ▸What this evidence cannot show

   UniProt positional features are curated from heterogeneous primary sources; the 'Kinase' region 1-426 reflects an older functional assignment of BCR as a Ser/Thr kinase that has never been supported by a structure of a BCR kinase fold.

   ▸Evidence that argues against this (1)

   • Two thirds of the region UniProt calls a kinase is annotated as disordered by the same record, and no structural database assigns a kinase fold anywhere in BCR. The union of UniProt's Disordered regions inside 1-426 (67-173, 185-247, 286-392, 416-476) covers 288 of 426 residues, 67.6%. The 41 InterPro matches for P11274 assign only four folds - DH 498-691, PH 708-868, C2 893-1020, Rho-GAP 1054-1249 - none a protein kinase, and inside 1-426 the sole structural assignment is the BCR-ABL oligomerisation domain (PF09036 / IPR015123, 3-75). The one SMART hit whose name contains 'protein kinase' is SM00239, 'Protein kinase C conserved region 2 (CalB)', at 912-1017, which is the C2 domain. BOUNDED ABSENCE: the wet-lab kinase claim has never been refuted in print - PubMed `(BCR[tiab] OR "breakpoint cluster region"[tiab]) AND ("kinase activity"[tiab] OR "serine/threonine kinase"[tiab] OR phosphotransferase[tiab]) AND (artifact[tiab] OR artefact[tiab] OR contaminant[tiab] OR contaminating[tiab] OR "failed to reproduce"[tiab] OR irreproducible[tiab] OR "not reproducible"[tiab])` returns 0 records, and the founding paper is not retracted. WHAT WOULD FALSIFY IT: recombinant BCR 1-426 expressed in E. coli, free of mammalian co-purifying kinases, showing no autophosphorylation - or a structure of that region showing a catalytic fold. UniProt Consortium 2026
2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   The BCR oligomerisation domain crystallised as residues 1-72 forms a tetramer: two monomers swap N-terminal helices and pair their C-terminal helices as an antiparallel coiled coil, then two dimers stack.

   Structure of the Bcr-Abl oncoprotein oligomerization domain · Zhao X, Ghaffari S, Lodish H, Malashkevich VN, Kim PS · Nature Structural Biology 2002
   PMID 11780146DOI
   ▸What this evidence cannot show

   An isolated 72-residue fragment; the oligomeric state of the same element inside full-length p210 on a membrane has not been measured directly. The PDB entry title for 1K1F was retrieved and confirmed to read 'Structure of the Bcr-Abl Oncoprotein Oligomerization domain'.

   ▸Evidence that argues against this (1)

   • UniProt annotates the coiled coil as only 28-55, which is narrower than the 1-72 crystallised unit and would omit the swapped N-terminal helix entirely. UniProt Consortium 2026
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   Independent domain databases place the BCR oligomerisation domain at 3-75 (Pfam PF09036 / InterPro IPR015123), 1-67 (SUPERFAMILY) and 1-72 (CATH-Gene3D), and reproduce the UniProt DH, PH, C2 and Rho-GAP boundaries to within a few residues.

   InterPro/Pfam/CATH-Gene3D/SUPERFAMILY domain matches for UniProt P11274 · InterPro Consortium; EMBL-EBI · InterPro 2026
   ▸What this evidence cannot show

   Database boundaries are model-match extents, not experimental measurements; Pfam and CATH disagree with each other by several residues at both ends.

   ▸Evidence that argues against this (1)

   • They do not - two of the four differ by tens of residues, and one is not even modelled as a single segment. Against UniProt's Rho-GAP at 1054-1248, Pfam PF00620 gives 1068-1215 (14 residues in at the N-terminus, 33 in at the C-terminus) and SUPERFAMILY SSF48350 gives 1050-1218; against UniProt's C2 at 893-1020, Pfam PF00168 gives 913-1007 (20 and 13). The PH domain is worse than a boundary disagreement: CATH-Gene3D G3DSA:2.30.29.30 and SUPERFAMILY SSF50729 both split it into two discontinuous fragments (692-775 plus 833-894, and 706-775 plus 833-868), so the structural classifications treat residues 776-832 as an insertion while UniProt draws 708-866 as one continuous domain. Only DH is close. InterPro Consortium; EMBL-EBI 2026
4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review

   In chronic myeloid leukaemia the p210 BCR::ABL1 fusion arises from breaks in the BCR major breakpoint cluster region between exons 13 and 15, giving b2a2/e13a2 and b3a2/e14a2 transcripts, while p190 arises from a break in BCR intron 1.

   The molecular biology of chronic myeloid leukemia · Deininger MW, Goldman JM, Melo JV · Blood 2000
   PMID 11071626
   ▸What this evidence cannot show

   A 2000 review; transcript nomenclature has since shifted from b2a2/b3a2 to e13a2/e14a2 but the underlying exon assignments are unchanged. This citation has no DOI in PubMed and is cited by PMID.

   ▸Evidence that argues against this (1)

   • Screening at scale turns up BCR breaks in neither region, at a small but measurable rate. Multiplex PCR screening of 4,750 CML patients identified 19 distinct uncommon BCR-ABL1 transcript types in 83 patients (1.7%), including e6a2, e8a2, e12a2 and e19a2 - all requiring a BCR break outside M-bcr and outside intron 1, with e19a2 sitting in the micro-bcr between exons 19 and 20. e19a2 was the single most frequent uncommon type (n=16) and was not clinically inert: against 571 newly diagnosed chronic-phase e13a2/e14a2 patients on frontline imatinib, the e19a2 group had significantly lower 1-year complete cytogenetic response (P=0.0004) and major molecular response (P=0.0018) and lower 2-year event-free (P=0.0004) and progression-free survival (P=0.0067). Qin YZ et al. 2018
5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   Converting BCR coding exon ends against the Ensembl BCR-201 (ENST00000305877) coding sequence places the end of exon 1 at codon 426, exon 13 at codon 902 and exon 14 at codon 927.

   Ensembl transcript ENST00000305877 (BCR): exon structure and coding-sequence coordinates, used to convert exon breakpoints into residue numbers · Ensembl; EMBL-EBI · Ensembl REST API 2026
   ▸What this evidence cannot show

   Computed by us from Ensembl exon coordinates; each of these exons ends mid-codon (+1 nt), so the junction residue itself is a chimeric codon and the '1-927' description is a convention that assigns the shared codon to BCR. The exon-1 result agreeing with UniProt's independently curated 426/427 site is the check that the method is sound.

   ▸Evidence that argues against this (1)

   • Those codon numbers are properties of one of seventeen Ensembl BCR transcripts, and the others shift them by up to 23 residues. The fetched gene record for ENSG00000186716 lists 17 protein-coding BCR transcripts with translation lengths spanning 1,267-1,294 aa, including two at 1,294, one at 1,278, one at 1,277 and one at 1,269, alongside ENST00000305877 at 1,271. Any BCR coordinate derived by exon-to-codon arithmetic therefore needs the transcript named with it - and a 23-residue spread is larger than the 25-codon difference between b2a2 and b3a2 that the same arithmetic is used to resolve. Ensembl / EMBL-EBI 2026
6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   An AlphaFold model of full-length BCR exists under accession AF-P11274-F1 covering residues 1-1271.

   AlphaFold Protein Structure Database entry for P11274 (model AF-P11274-F1) · Jumper J, Varadi M, et al.; EMBL-EBI/DeepMind · AlphaFold DB 2026
   ▸What this evidence cannot show

   A prediction, not an experiment. Confidence is low across the long disordered linkers (UniProt annotates disorder at 67-173, 185-247, 286-392 and 416-476), so inter-domain geometry in the model should not be trusted.

   ▸Evidence that argues against this (1)

   • Two fifths of that model is in the band AlphaFold itself says not to interpret. The fetched record for AF-P11274-F1 (v6) gives a global pLDDT of 64.81 with 38.4% of residues at pLDDT below 50 (very low), 7.4% low, 32.7% confident and 21.5% very high - so 488 of 1,271 residues are very low, close to the 288 residues UniProt annotates as disordered inside 1-426 plus the remaining linkers. 'Covers residues 1-1271' is a statement about the file's extent and says nothing about whether any part of it is usable. AlphaFold DB; EMBL-EBI / Google DeepMind 2026

• refutes →BCR::ABL1 (Philadelphia chromosome), t(9;22)(q34;q11)

  Counter-evidence added cycle 41: two breakpoint regions do not suffice. Multiplex PCR across 4,750 CML patients found 19 uncommon BCR-ABL1 transcript types in 83 patients (1.7%), including e6a2, e8a2, e12a2 and e19a2, all requiring a BCR break outside M-bcr and outside intron 1; e19a2 had significantly worse 1-year CCyR and MMR and 2-year EFS and PFS on frontline imatinib (Qin 2018, PMID 29974949).

• part of →BCR::ABL1 (Philadelphia chromosome), t(9;22)(q34;q11)

  5-prime partner domain architecture; the coiled-coil oligomerisation domain is the retained element that activates the kinase

• fuses with →ABL1 — domain architecture and residue coordinates (with the 1a/1b numbering trap)

  BCR::ABL1, t(9;22)

• supports →Hypothesis — forced-oligomerisation interfaces are the mutation-resistant drug targets the fusion field has been ignoring

  the BCR coiled coil is the canonical fusion oligomerisation interface outside the KMT2A family

• ← analogous toAFF1 / AF4 — domain architecture and residue coordinates

  cross-family comparison: both fusions retain a partner-side oligomerisation element, drawn on the same residue-scaled axis

• supports →Mechanism class 3 — constitutive kinase activation (BCR::ABL1, NPM1::ALK, ETV6::NTRK3)

  N-terminal oligomerisation domain (1-72) tetramerises and forces ABL1 trans-autophosphorylation

• supports →Breakpoint anatomy, and an audit of the prognosis claim

  e1a2 (BCR 1-426), e13a2 (1-902) and e14a2 (1-927) retain different amounts of the same protein

• ← fuses withABL1 — domain architecture and residue coordinates (with the 1a/1b numbering trap)

  ABL1 a2 exon joins the BCR major or minor breakpoint cluster

• related to →KMT2A / MLL1 — the gene that breaks

  comparator: unlike KMT2A, BCR contributes a compact crystallised oligomerisation domain with exact coordinates