CASP-like (CASPL) family curation

Bottom line: CASP-like (CASPL) proteins are small four-span plasma-membrane
scaffolds; the five true CASPs build the root Casparian strip, but most CASPLs
are uncharacterized. We reviewed 21 Populus trichocarpa CASPLs and five
characterized Arabidopsis orthologs (CASPL1B1, 1B2, 1D1, 1D2, 4C1), wrote
descriptions for the two PANTHER families (PTHR33573, PTHR36488), and ran
reproducible orthology and expression analyses. We did this to test whether
functions from the few characterized Arabidopsis members can be transferred to
poplar. The answer is yes at group level: for all 20 poplar CASPLs in the
analysis, the best Arabidopsis hit is in the same Roppolo group (12 are
reciprocal best hits). Expression data support the transfer: poplar CASPL1B1 is
root-specific and CASPL4C1 is the most stress-responsive member. The GO
annotations themselves are sparse. Each poplar gene has one IEA plasma-membrane
row (all accepted), and the 21 Arabidopsis rows gave 13 ACCEPT, 1
KEEP_AS_NON_CORE, 2 REMOVE (ISM extracellular region and nucleus) and 5
UNDECIDED (HDA Golgi/endosome rows on CASPL1D1). The component list below says
20 poplar genes; the repo also has a 21st, CASPL2A1, from the same batch.

Cross-cutting curation of the plant CASP / CASP-like (CASPL) membrane-scaffold family,
extending the Populus trichocarpa CASPL gene-review batch to its Arabidopsis orthologs,
the underlying PANTHER families, and a reproducible orthology analysis.

Background

CASP / CASPL proteins are small, non-enzymatic, four-transmembrane-span (tetraspan)
plasma-membrane proteins related to the MARVEL-domain superfamily. The five bona fide CASPs
(CASP1–CASP5 in Arabidopsis) build the Casparian strip by recruiting the lignin-polymerization
machinery in the root endodermis ([PMID:21593871], [PMID:24920445]). The much larger CASPL
subfamily is mostly uncharacterized, but specific clades have defined functions (below).

Components

1. Poplar CASPL reviews (20 genes)

The 20 reviewed P. trichocarpa CASPLs (genes/POPTR/CASPL*). All carry a single IEA
plasma-membrane annotation; five were enriched with characterized-ortholog context.

2. Arabidopsis ortholog curation (5 genes, real deep research)

Full reviews of the characterized Arabidopsis orthologs with falcon deep research
(genes/ARATH/CASPL{1B1,1B2,1D1,1D2,4C1}):

• CASPL1B/1D — expressed in suberized endodermis; interact with aquaporin PIP2;1;
  implicated in suberization / water-transport modulation ([PMID:30767240]).
• CASPL4C1 — cold-inducible; negative regulator of growth/biomass/flowering; vascular role
  beyond the Casparian strip ([PMID:26399665]).

3. PANTHER families (2 reviewed)

• PTHR33573 "Casparian Strip Membrane" — corrected the LLM description to separate the bona
  fide CASP1–5 clade from the dominant, uncharacterized CASPL subfamily.
• PTHR36488 (InterPro IPR044173, "CASPL") — added a description for this divergent CASPL
  clade (the 1U + 1B/1D subgroups, including the aquaporin-interacting CASPL1B/1D members).

4. Bioinformatics — orthology validation

projects/CASPL_FAMILY/bioinformatics/ — reproducible Biopython pairwise-identity / reciprocal-best-hit
analysis of all 20 poplar + 34 Arabidopsis CASPLs. Confirms that every poplar CASPL's best
Arabidopsis hit is in the same Roppolo group, validating group-level (not strict 1:1) functional
transfer. See bioinformatics/RESULTS.md.

Scope notes

• Peanut / watermelon have no reviewed (Swiss-Prot) CASPL entries — the genome-wide peanut
  set ([PMID:39124195]) and watermelon ClCASPL ([PMID:26399665]) are TrEMBL only, so they cannot be
  curated to the repository's reviewed-entry standard. Rice (29) and maize (23) do have reviewed
  CASPL entries and are the curatable cross-species extension if desired.

Slides

• Slides (Marp source: CASPL_FAMILY-slides.md) — AI generated