LPXC1 (At1g24793, F4IAT8) notes, with shared notes for the LPXC1-5 cluster

Deep research (falcon) fails in this environment (authentication), so no
-deep-research-*.md file exists. The literature summary below was compiled manually
from cached publications, the PMC full text of Li et al. 2011, and PubMed searches.
LPXC2-5 notes files point here.

The cluster

• Five LpxC genes sit in tandem in a 68-kb region of chromosome 1
  PMID:21709257,
  probably from duplication
  PMID:21709257.
• Locus-to-name mapping (Li et al. 2011 Table 1, read from PMC3136250; the table is not
  in the cached markdown): AtLpxC1 = AT1G24793 (variants AtLpxC1.1, 326 aa; AtLpxC1.2,
  266 aa), AtLpxC2 = AT1G24880, AtLpxC3 = AT1G25054, AtLpxC4 = AT1G25145, AtLpxC5 =
  AT1G25210. UniProt uses the same numbering (LPXC1 F4IAT8 ... LPXC5 F4IAW1).
• EST/cDNA support exists for the family
  PMID:21709257.
  Two variants lack the catalytic C-terminus
  PMID:21709257.
• Bioinformatics (file:ARATH/LPXC1/LPXC1-bioinformatics/RESULTS.md): the five canonical
  UniProt sequences (326 aa) are 100% identical; each is 36.2% identical to E. coli LpxC
  and retains the Zn(2+) ligands (H109, H281, D285 = E. coli H79, H238, D242), D289 (D246)
  and the proton donor H313 (H265). F4IAT8 isoform 2 (266 aa = AtLpxC1.2) loses all
  C-terminal catalytic residues. Locus-level expression cannot be resolved because
  reads from identical coding sequences multi-map.
• UniProt gene-model notes: each locus was split from a larger erroneous model
  (e.g. At1g24793/At1g24800; At1g24880/At1g24881; At1g25054/At1g25055;
  At1g25145/At1g25150; At1g25210/At1g25211).

Which locus was studied (Li et al. 2011, PMID:21709257, full text)

• Gene function was tested only collectively, by RNAi against all five genes
  PMID:21709257,
  monitored with primers from shared sequence
  PMID:21709257.
• Lipid X (2,3-diacylglucosamine 1-phosphate, the product two steps downstream of LpxC)
  fell to 37%
  PMID:21709257,
  and RNAi suppressed lipid X accumulation in atlpxb-1
  PMID:21709257.
• The authors explicitly do not know which copy is functional
  PMID:21709257.
• The only locus-specific construct is the GFP fusion of AtLpxC1.1 (AT1G24793), which is
  mitochondrial
  PMID:21709257;
  not chloroplastic
  PMID:21709257.
  Because the AtLpxC1.1 protein is identical to the canonical products of the other four
  loci, the localisation result applies to the protein each locus encodes; the 35S
  promoter construct says nothing about native expression of a given locus.
• No in vitro deacetylase assay or E. coli lpxC complementation of an Arabidopsis LpxC
  has been reported (contrast AtLpxA, PMID:22545860). The MF is inferred from orthology
  with full active-site conservation plus the in planta lipid X decrease.
• RNAi plants have no visible phenotype
  PMID:21709257.
• Pathway conclusion
  PMID:21709257;
  end product unknown
  PMID:21709257.

Other evidence

• Carrie et al. 2015 (PMID:25732537) mitochondrial N-terminomics: UniProt cites it for a
  1-21 transit peptide (ECO:0000269) and mitochondrial location (ECO:0000305) on all five
  entries. The LpxC peptide is in supplementary data, not the cached text; an N-terminal
  peptide from the mature protein cannot be assigned to one locus since sequences are
  identical. Supports mitochondrial import of the shared product.

GOA observations and decisions (applied identically to all five)

• GO:2001289 lipid X metabolic process is obsolete (QuickGO: isObsolete true, no
  replacement). The IMP rows from Li et al. capture the lipid X decrease after LpxC RNAi;
  MODIFY to GO:0009245 lipid A biosynthetic process (verified current in OLS/QuickGO).
  No lipid IVA-specific GO process term exists (OLS search).
• GO:0103117 MF: ACCEPT (IBA/IEA/ISS/IMP). The IMP is family-level RNAi; it is
  consistent with the orthology inference and is not attributable to one locus.
• GO:0005739 mitochondrion: ACCEPT. IDA rows rest on AtLpxC1.1 GFP (identical protein).
• GO:0009507 chloroplast (AtSubP ISM): REMOVE; GFP was not chloroplastic and all Lpx
  enzymes tested were mitochondrial. Lipid X is found in chloroplast fractions, but that
  is the lipid, not the enzyme.
• GO:0009245 lipid A biosynthetic process (IBA/IEA): ACCEPT; refutes the Wei et al.
  2020 RAR flag (PMID:32470107; projects/RAR_MISANNOTATION.md).
• No NEW annotations: every experimentally supported term is already present in GOA for
  each locus (any evidence code); the experimental evidence is recorded in those rows.
• LPXC3 has no IBA rows in GOA (absent from the PAINT WITH/FROM list) and LPXC2 has no
  ISM chloroplast row; the gene product is identical, so the absence is a pipeline
  artefact, not biology.

Searches

• PubMed "LpxC AND (Arabidopsis OR plant OR mitochondria)" (26 hits, bacterial LpxC
  inhibitor papers dominate) and "(lipid A OR lipid X OR UDP-3-O-acyl-N-acetylglucosamine
  deacetylase) AND Arabidopsis[Title/Abstract]" (20 hits; besides PMID:21709257,
  PMID:20124190, PMID:22545860, PMID:28202596 these are LPS perception papers, e.g.
  LORE, PMID:25729922). No later study of an individual AtLpxC locus was found.