ACTRT2

Q: Can the perinuclear-theca proteins re-earn GO:0005200 on evidence of their own, and if so should ACTL7A's removal be revisited at the same time? The merged ACTRT3 review raises this question and this review contributes what is, within the whole ARP-T and theca set, the strongest single instance of such evidence: ACTRT2 is not merely resident in the theca but is one of the two major acidic components of the purified, detergent- and salt-resistant cytoskeletal calyx, identified by fractionation and partial protein sequencing. That is bulk structural constituency of a structure GO places under GO:0005856 cytoskeleton, which is what GO:0005200's definition asks for, and it is evidence neither ACTRT3 nor ACTL7A has - ACTRT3 is not one of Heid's two components, and ACTL7A's own review explicitly withdrew a residence-based version of the argument. The reason this review nonetheless generalises its row rather than accepting it is that the row's own evidence chain runs from polymerising actins, so accepting it would publish a filament-constituent inference that the structure audit excludes. The question is therefore whether GO should have a route to assert GO:0005200 from theca evidence directly, independent of the actin node - and if it does, the ARP-T set and ACTL7A should be decided together rather than one gene at a time.

Q: The eight-node GO:0005200 rejection sweep in PTHR11937 is independently confirmed here, from the cached PAINT table: the term propagates at exactly one node (PTN000940351) and is negated as an IRD with negated=true at PTN000233596 (Arp2), PTN000233796 (Arp3), PTN000233752, PTN000233887, PTN000234048 and PTN001732543 (the nuclear ARPs of the Ino80, Swr1, NuA4 and SWI/SNF complexes), PTN007551901, and PTN008986528 (ACTL7A/ACTL7B) - the last of which alone received the parent GO:0005198 as a replacement, and GO:0005198 occurs nowhere else in the family. The merged ACTRT3 review asks whether the sweep should be extended to the ARP-T branch; the recommendation is not repeated here, only corroborated, with one addition. If the IRD is applied to that branch, GO:0005198 should be substituted at the same time as it was at PTN008986528, not merely rejected - otherwise ACTRT1, ACTRT2, ACTRT3, ACTL9 and ACTL10 lose their only molecular function outright, and four of the five reported members of one perinuclear-theca complex would then be annotated at one level while ACTR10, on the same un-adjudicated path but with an ortholog in the seed set, keeps the child.

Q: Have the perinuclear-theca actin-related proteins convergently lost actin's polymerisation-coupled ATPase, and does the retained nucleotide pocket still bind a nucleotide? Actin's hydrolysis is triggered when His161 flips on incorporation into the filament, and the flip is governed by the Pro-rich loop at Ala108/Pro109. All five reported members of the sperm perinuclear-theca ARP complex have lost His161 - ACTRT1 to Cys, ACTRT2 to Cys, ACTRT3 to Gln, ACTL7A to Tyr, ACTL9 to Tyr - while every panel member that extends or nucleates a filament retains it, and ACTL7B, which is testis-specific but is not a reported member of that complex, also retains it. ACTRT2 has additionally lost both Pro-rich loop residues - context rather than evidence, since the corresponding actin mutants polymerise and hydrolyse like wild type - while remaining identical to actin at all ten named phosphate-loop, cation-site and sensor positions, the adenine/ribose region carrying two non-conservative substitutions. The merged ACTL7A review raised this for that gene alone; the census generalises it to the whole complex. The testable prediction is a nucleotide-bound but non-hydrolysing fold, and no member of the set has ever been assayed for nucleotide binding.

Q: Should GO:0033150 cytoskeletal calyx be extended, in both its definition and its annotation set, beyond the three basic proteins? The term's definition states that the calyx contains calicin and basic cylicin proteins, and in human it is annotated to exactly CYLC1, CYLC2 and CCIN. The paper that named ARP-T1 and ARP-T2 reported them as two major acidic components of the purified calyx and listed two major actin-capping proteins there as well; CAPZA3 and CAPZB carry GO:0033011 but not GO:0033150. So the term has not tracked the composition of the structure it names.

Q: Why has the perinuclear-theca literature not reached ACTRT2 and ACTRT3 in either species? Of the six reported members of the theca complex, these two are the only ones with no experimental biological-process annotation in human or mouse (ACTRT2 has one BP row in human and two in mouse, all automatic). The reference that establishes the complex names ACTRT2 explicitly and has produced acrosome-assembly, spermatid-development and fertilization annotations for ACTL7A and ACTL9 but none for Actrt1 or Actrt2; the calicin paper gave mouse Actrt2 a cellular-component IDA and a protein-interaction row but no process term; and the Actrt2 knockout paper has not been curated at all. Because the human record here is fed by ortholog transfer, closing the mouse gap would close the human one.

Q: Where do ACTRT2 and ACTRT3 genuinely differ, given that they were reviewed independently and share three byte-identical annotation rows? Recorded so the contrast is not lost. The two reviews agree on every shared row and their two structural analyses agree numerically on every shared panel member - on ACTRT3's chemically-compatible metric, which counts identical plus conservative and so is not the identical-only metric used on the GO:0005200 row: Arp53D 33 of 38, the dynactin Arp1 paralogue 28, conventional actins 38, ACTRT2 20 and ACTRT3 18. Computed independently on both sides, which is a mutual validation worth noting. The differences are all in evidence, not in method. ACTRT2 is one of Heid's two major acidic calyx components and ACTRT3 is not, which is why only this review can supply the bulk-constituency evidence bearing on the GO:0005200 question. ACTRT3 has a knockout with acrosome and Golgi phenotypes and an informative binding partner in PFN3, so its new annotations are GO:0001675 acrosome assembly and GO:0005522 profilin binding; ACTRT2's knockout reports only shrunken seminiferous tubules, so its new process term is the more general GO:0007283 spermatogenesis - a difference in evidence, not a disagreement about level. ACTRT2 additionally has ten IntAct records including six CCT subunits where ACTRT3 has two, and it is the only one of the pair for which GO:0033150 cytoskeletal calyx is supported. ACTRT3 has nuclear and male-germ-cell-nucleus annotations that ACTRT2 lacks entirely.

Q: What is ACTRT2 doing in the flagellar midpiece? The only human protein-level localisation reports ACTRT2 in the post-acrosomal region and the middle piece, and the midpiece is not a perinuclear-theca compartment. No current annotation accounts for it, and it is unclear whether it reflects a genuine second pool, residual cytoplasm, or antibody cross-reactivity with another actin-family protein - the last being a real concern for a protein 48 per cent identical to beta-actin.

Q: Is ACTRT2 protein testis-restricted or merely testis-enriched? The 2001 cloning paper reports that hArpM1 and hArpM2 mRNAs are expressed in all tested human tissues, at lower levels than actin, whereas the calyx paper reports specific synthesis in the testis and the knockout paper describes testis-specific expression; HPA calls the gene tissue-enriched rather than tissue-specific. Low ubiquitous transcript with testis-restricted protein reconciles all three, but that has not been tested directly, and it matters for whether any non-germline function should be entertained.