gbuA (PA1421, Q9I3S3) notes

Review of Pseudomonas aeruginosa PAO1 guanidinobutyrase for the PseudoCAP IBA-seed
project (worklist rank 33). No deep-research provider was available; these notes are
from UniProt, PubMed and the cached publications.

Biology

• GbuA is a Mn2+-dependent ureohydrolase (arginase family, agmatinase subfamily;
  PANTHER PTHR11358:SF26) that hydrolyses 4-guanidinobutanoate to 4-aminobutanoate
  (GABA) and urea [file:PSEAE/gbuA/gbuA-uniprot.txt].
• Nakada and Itoh 2002 cloned gbuA by complementing the gbuA9005 mutant "defective in
  4-GB utilization"; a PA1421 knockout "did not grow on 4-GB and failed to form GBase";
  the purified enzyme "catalyzed the specific hydrolysis of 4-GB", and "No enzyme
  activity was detected with other guanidino compounds, including arginine, agmatine,
  3-guanidinopropionate, and guanidinoacetate" PMID:12029055. Mn2+ restored activity
  after EDTA PMID:12029055. Expression is induced by 4-GB through the LysR-type
  regulator GbuR PMID:12029055.
• Lee et al. 2011 solved the GbuA crystal structure (with Mn2+) alongside GpuA, the
  paralogous guanidinopropionase, and showed active-site loop differences (Met161 of
  GbuA) behind their substrate specificity PMID:21600989. UniProt: homohexamer, two
  Mn2+ per subunit.
• Pathway. Jann et al. 1988 described the fourth arginine catabolic pathway of
  P. aeruginosa: D-arginine dehydrogenase gives 2-ketoarginine, "catabolized further
  to 4-guanidinobutyraldehyde, 4-guanidinobutyrate and 4-aminobutyrate", with
  guanidinobutyrase inducible by 2-ketoarginine and strongly by 4-guanidinobutyrate;
  "In mutants blocked in the L-arginine succinyltransferase pathway ... L-arginine was
  channelled partially into the D-arginine dehydrogenase pathway" PMID:3141581.
  Later work showed that L-arginine enters this route via the AruH arginine:pyruvate
  transaminase and AruI 2-ketoarginine decarboxylase (the arginine transaminase
  pathway), and that in an aruF (AST-deficient) background "disruption of gbuA
  abolished growth on l-arginine" PMID:17416670. So GbuA is a genuine step of
  L-arginine (and D-arginine) catabolism in P. aeruginosa, downstream of 2-ketoarginine.

PAINT node PTN004253691

From .cache/paint/IBD.gaf and the PAINT leaf GAF. IBDs at PTN004253691:
GO:0047971 guanidinobutyrase (seeds human AGMAT Q9BSE5, GbuA Q9I3S3), GO:0047972
guanidinopropionase (seeds Q9BSE5 and P. aeruginosa GpuA Q9I6K2), GO:0006527
L-arginine catabolic process (sole seed GbuA), and NOTs for agmatinase activity
(GO:0008783) and putrescine biosynthesis from arginine (GO:0009446). The 34 leaves
are vertebrate and invertebrate AGMAT orthologs ("guanidino acid hydrolase,
mitochondrial" in human, mouse and rat), P. aeruginosa GbuA and GpuA.

• The clade is a guanidino-acid hydrolase clade. Human AGMAT "efficiently hydrolyzed
  linear guanidino acids such as TC and GBA, whereas GPA and arginine were turned over
  less efficiently and guanidinoacetate or compounds lacking acidic residues such as
  agmatine were not accepted as substrates" PMID:36543883.
• Every leaf receives both guanidinobutyrase and guanidinopropionase IBAs. For GbuA the
  guanidinopropionase IBA is directly contradicted: no activity with
  3-guanidinopropionate PMID:12029055.
• L-arginine catabolic process is right for GbuA, but the node propagates it to GpuA
  (which has no activity on arginine and serves 3-guanidinopropionate utilisation) and
  to the vertebrate guanidino acid hydrolases, which metabolise arginine only poorly
  and whose physiological substrates (taurocyamine, 4-GB) are not arginine
  catabolites in the sense of a catabolic pathway. The node term over-reaches for the
  clade; it would sit better on a GbuA-only (bacterial guanidinobutyrase) subnode.

PseudoCAP IBA-seed assessment

Also: the GO:0047972 guanidinopropionase IBD at the same node (not PseudoCAP-seeded)
reaches GbuA, which has no guanidinopropionase activity; REMOVE on this gene and a
GO_CENTRAL MISSING_NEGATION defect.

Decisions summary

• ACCEPT: guanidinobutyrase activity (EXP x2, IBA, IDA, IEA), L-arginine catabolic
  process (IBA, IDA).
• MODIFY: GO:0016813 amidine hydrolase activity to GO:0047971; metal ion binding to
  manganese ion binding (GO:0030145).
• REMOVE: guanidinopropionase activity (IBA), contradicted by direct assay.
• Source defects: GO_CENTRAL MISSING_NEGATION (NOT GO:0047972 at GbuA leaf, PROBABLE);
  GO_CENTRAL MISSING_NEGATION CHECK for the L-arginine catabolic process IBD on GpuA
  and vertebrate AGMAT leaves (SUSPECTED).

PANTHER tree check (2026-11-05)

Topology of PTN004253691 fetched with projects/PSEUDOCAP/panther_tree.py node PTN004253691 --depth 3 / --leaves (PANTHER treeinfo for PTHR11358).

PTN004253691: HORIZ_TRANSFER, 34 leaves, all PTHR11358:SF26.

Conclusions:

• The two P. aeruginosa enzymes are not sisters. GpuA is the outgroup leaf of the
  whole node and GbuA is nested inside the animal clade through a second
  horizontal-transfer node, so there is no bacterial guanidinobutyrase subnode.
  The L-arginine-catabolic IBD cannot be relocated; its only correct scope is the
  GbuA leaf PTN002260452. Equivalently, NOT annotations at PTN002260450,
  PTN000144490 and PTN000888144 cover all 33 other leaves without touching GbuA.
• For the same reason the GO:0047972 guanidinopropionase IBD cannot be moved off
  GbuA: the smallest clade containing both of its seeds (GpuA PTN002260450 and human
  AGMAT inside PTN000144490) is PTN004253691 itself. A NOT at the GbuA leaf
  PTN002260452 is the only fix.

Both defects raised to CONFIRMED (SUSPECTED -> CONFIRMED for the GO:0006527 defect,
PROBABLE -> CONFIRMED for the GO:0047972 defect). The annotation actions are unchanged:
ACCEPT on the GO:0006527 IBA (correct for this gene) and REMOVE on the GO:0047972 IBA
(contradicted by direct assay, and the tree confirms no relocation can rescue it).