CD8B notes

2026-10-03 initial review (ADAPTIVE_IMMUNITY, TCR trunk)

CD8 beta (P10966) is the partner of CD8 alpha in the CD8 alpha-beta coreceptor of conventional
MHC class I-restricted T cells.

Key primary evidence:

• Human CD8B cloned; ~90% of CD8A+ thymocytes and blood lymphocytes carry surface CD8 beta;
  splice variants differ in the tail and one lacks the TM exon
  PMID:3145195.
• Needs CD8A for surface expression
  PMID:3145196.
• CD8 beta tail palmitoylation partitions CD8ab into rafts, where LCK associates and is activated (mouse T1.4 hybridomas)
  PMID:10925291.
• CD8 beta tail couples CD8 to TCR/CD3 and raises TCR-pMHC avidity
  PMID:11714755;
  Cys179 mutation abolishes raft association
  PMID:11714755.
  The human tail has a membrane-proximal Cys pair (sequence ...IHLCCRRRRARLRFMKQFYK).
• Mouse CD8ab/H-2Dd structure: CD8 beta contributes ~half the interface and sits T cell-membrane proximal
  PMID:19625641.
• Mouse CD8b-/- and tailless CD8b: lower CD8a-associated LCK activity, impaired CD8 SP development
  PMID:9647223.
  Necessity evidence only; no thymocyte-development process term proposed (project rule 3).
• LCK binds the CD8 alpha tail (CxC zinc clasp), not CD8 beta
  PMID:9830036.
  So no LCK-binding MF for CD8B; GO:0042610 CD8 receptor binding sits on LCK.

Annotation decisions

• protein binding (IPI, PMID:2493728, with CD8A): REMOVE, matching CD8A review.
• GO:0007169 RTK signaling (NAS): MODIFY -> GO:0050852. TCR signaling is not under GO:0007169
  in GO (checked ancestors via OLS); receptor has no kinase activity.
• GO:0042101 T cell receptor complex (NAS): MARK_AS_OVER_ANNOTATED. Definition = TCR heterodimer + CD3.
  Differs from CD8A review (ACCEPT); raised as a suggested question.
• Early endosome membrane (Reactome, Nef pathway): non-core.
• NEW: plasma membrane raft (GO:0044853), ISS from mouse palmitoylation data; CD8A has IDA raft.
• Curation gap (not a knowledge gap): no GO CC for the CD8 alpha-beta heterodimer.

Deep research integration (falcon)

Report: CD8B-deep-research-falcon.md (Edison/falcon, finished 2026-10-03, ~23 min). Used as leads only.

Adopted (each traced to a cached primary paper):
- Human primary-T-cell evidence that CD8 alpha-beta (not alpha-alpha) is needed for optimal HLA class I
coreceptor function, and that the extracellular domains suffice
PMID:23738014.
Added as support for coreceptor activity and core function 1.
- Human CD8B membrane isoforms M-1 to M-4; M-1 naive, M-4 effector memory; M-4 tail internalization motif
PMID:23533620.
Added to description only (no GO change).
- Mouse CD8 beta CDR2/CDR3 loops required for coreceptor activity
PMID:16356863.
Added as support for MHC class I protein binding.
- Confirms CxC LCK-binding motif is on CD8 alpha, not CD8 beta (agrees with my reading; cited on the GO:0007169 MODIFY row).

Not adopted:
- Nanobody radiotracer imaging (De Groof 2024), transcriptomics (Akula 2024), CD8ab gamma-delta T cells in T-ALL
(Sumaria 2024): applications/expression context, no bearing on GO terms.
- Stalk sialylation tuning MHC affinity: from the Srinivasan 2024 review, mostly mouse; not traced to a primary
human paper, so not used.
- Kd range 10-500 uM: review-level range across CD8 complexes, not a CD8B measurement.

Report errors/caveats:
- Table says the membrane-proximal cysteine of the TM "contributes to disulfide stabilization"; the
primary literature (Arcaro 2000/2001) assigns the membrane-proximal tail cysteine to palmitoylation, and the
interchain disulfide is in the stalk region. Not used.
- No other factual errors found; the report correctly flags that the bound structures are mouse and that
familial CD8 deficiency is a CD8A, not CD8B, defect.