At3g08840 (A0A1I9LPE3) review notes

Deep research via just deep-research-falcon was not available in this session (auth
failure), so the literature survey below was done manually with PubMed searches
(2026-10-19). No -deep-research-*.md file was created.

Context: projects/RAR_MISANNOTATION.md (re-examination of Wei et al. 2020,
PMID:32470107). Wei et al.
flagged the IBA GO:0009252 peptidoglycan biosynthetic process on this protein; that row
is no longer in GOA. The MF IBA GO:0008716 D-alanine-D-alanine ligase activity
(PTN000566166; donors E. coli DdlB P07862, DdlA P0A6J8, M. tuberculosis Ddl P9WP31) remains.

Identity

• 949 aa TrEMBL entry, "D-alanine-D-alanine ligase family"; two tandem Ddl modules
  (Pfam Dala_Dala_lig_N PF01820 x2, Dala_Dala_lig_C PF07478 x2; two ATP-grasp domains
  223-452 and 672-925); PANTHER PTHR23132:SF0. PROSITE PS00843 and PS00844 (D-Ala-D-Ala
  ligase signatures) both match (file:ARATH/At3g08840/At3g08840-uniprot.txt).
• Called At-DDL in the moss literature. The fused two-module architecture is shared with
  moss and Selaginella DDL PMID:27325639.
• Note: Machida et al. 2006 abstract says Arabidopsis has "two genes for D-Ala-D-Ala ligase
  (Ddl)" PMID:16618924. Later work treats At3g08840 as the single fused
  DDL gene (the two "genes" were probably the two halves/ two models); Hirano et al. 2016
  analysed one locus.

Localization

• Chloroplast stroma by MS proteomics (Zybailov 2008, HDA row). Hirano et al. restate this:
  PMID:27325639.
• Moss DDL transit peptide-GFP is plastidial PMID:27325639.
• The ISM cytoplasm (AtSubP) row is a prediction; chloroplast is technically part of
  cytoplasm in GO but the specific supported location is chloroplast stroma.

Function: what is and is not known

• Moss ortholog: genuine D-Ala:D-Ala ligase in plastid peptidoglycan synthesis needed for
  plastid division. Knockout -> macrochloroplasts; D-Ala accumulates; rescued by
  D-Ala-D-Ala but not LA-LA, DA-LA, LA-DA or D-Ala PMID:27325639; EDA-DA labelling shows a
  peptidoglycan layer around moss chloroplasts.
• Arabidopsis: two T-DNA lines (SALK_092419 At-ddl-1, SAIL_906_E06 At-ddl-2) lack
  transcript but have no chloroplast morphology defect PMID:27325639.
• Activity not tested PMID:27325639. Authors suggest possible change of function, as for AtMurE
  PMID:27325639.
• No Pbp in Arabidopsis PMID:16618924; angiosperms generally retain only MurE, Ddl, MraY, MurG
  PMID:41351887.
  Sugita et al. 2025 speculate AtDdl acts in D-amino acid metabolism PMID:41351887.
• Precedent for repurposing: AtMurE is required for chloroplast biogenesis/PEP-dependent
  transcription, not division, and does not complement moss murE PMID:18036201. Interesting counterpoint: gymnosperm MurE (Larix) rescues the Arabidopsis
  murE albino phenotype PMID:28158764, and spruce MurE has bona fide MurE enzymatic activity and also rescues
  the Arabidopsis mutant PMID:41351887.
  That is, an enzymatically active MurE can fulfil the Arabidopsis "repurposed" role -
  so loss of the division role does not imply loss of the ancestral catalytic activity.
  By analogy, At-DDL may retain ligase activity while the downstream pathway is gone.
• Review of plastid PG evolution: loss happened independently several times
  PMID:19647785.
• D-Ala-D-Ala dipeptide occurs in seed plants (tobacco) PMID:27325639.

PubMed searches with no additional Arabidopsis DDL-specific primary literature:
"Arabidopsis D-alanine D-alanine ligase" (0 hits); "(D-Ala-D-Ala OR D-alanyl-D-alanine) AND
(plant OR Arabidopsis OR chloroplast)" (only moss/bacterial/irrelevant hits);
"plastid peptidoglycan Arabidopsis MurE ..." (7 hits, all cited above).

Bioinformatics

See At3g08840-bioinformatics/RESULTS.md. Summary: C-terminal module carries both PROSITE
Ddl signatures, active-site Glu and all three Mg2+ ligands of E. coli DdlB; the N-terminal
module has a diverged Mg2+-binding loop. The same pattern (including a Ser150->Cys in
module 2) is present in the moss DDL, whose ligase function is genetically supported, so
there is no Arabidopsis-specific loss of catalytic residues.

Verdicts

• Wei et al. BP flag (peptidoglycan biosynthetic process): conclusion holds for
  Arabidopsis (no Pbp; no plastid phenotype; pathway incomplete), though their stated
  reason (plant walls are cellulose) is wrong - the relevant compartment is the plastid,
  and the IBA node is correct for bryophytes/lycophytes/gymnosperms.
• MF IBA D-Ala-D-Ala ligase: node placement defensible (target inside the plant DDL clade;
  moss member functionally supported); no target-specific evidence of catalytic loss.
  Accept, flagging that it is biochemically untested in Arabidopsis.