ADAMTSL1 is a large secreted glycoprotein of the extracellular matrix and the prototype of the ADAMTS-like family. It is built entirely from the ancillary modules of the ADAMTS proteases - thrombospondin type-1 repeats, four immunoglobulin-like C2-type domains, an ADAMTS spacer region and a C-terminal PLAC domain - and lacks the prodomain, the metalloprotease domain and the disintegrin-like domain, so it is not itself a protease. Its biosynthesis is unusually dependent on glycosylation of the thrombospondin repeats: O-fucosylation by POFUT2 with subsequent glucose addition by B3GLCT, together with C-mannosylation of tryptophans in the first repeat, act as a quality-control gate on export from the endoplasmic reticulum, and disrupting either modification traps the protein inside the cell. Once secreted, the protein is deposited into the pericellular matrix in a punctate pattern that gave it the name punctin, and it is cleaved by MMP10 in fibroblast secretomes. The gene is expressed most strongly in adult skeletal muscle, with additional expression in ocular anterior-segment tissues, the mandibular condyle and a myelinating Schwann cell subtype. A heterozygous substitution at a C-mannosylation site blocks secretion and acts dominant-negatively in a pedigree with congenital glaucoma together with craniofacial, dental, auditory, renal, brain-vascular and limb anomalies, and other missense variants segregate with mandibular prognathism. Its molecular function and binding partners are undetermined; by analogy with other ADAMTS-like proteins it is expected to act in the assembly or turnover of the extracellular matrix, but this has not been demonstrated for ADAMTSL1 itself.
Curated functional classes representing distinct biological activities. These may be splice variants, cleavage products, or other forms with different functions.
ADAMTSL1_PUNCTIN1_SHORTADAMTSL1_SIGNAL_PEPTIDE_LESS| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0030198 extracellular matrix organization | IEA GO_REF:0000002 | KEEP AS NON CORE | Summary: Family-signature inference that is correctly constructed and carries independent positive support at family level, but no ADAMTSL1 evidence. Kept, and explicitly not asserted as this gene's demonstrated core function. Reason: Three things give this IEA positive support rather than merely leaving it unrefuted. First, the mapping is careful: IPR013273 (ADAMTS/ADAMTS-like) is the only one of ADAMTSL1's fourteen InterPro entries with any GO mapping, and it maps to GO:0030198 alone - no peptidase, protease or hydrolase term - even though the signature also matches the catalytic ADAMTS metalloproteases. Second, GO:0030198 does not depend on the domain ADAMTSL1 lacks: catalytic family members reach the term through proteolysis, but ADAMTSL2, ADAMTSL4 and THSD4 reach it non-catalytically through fibrillin microfibril assembly, so the term sits at the level of generality at which a mechanistically heterogeneous family genuinely agrees, and a non-catalytic route to it is available to this architecture in principle. This ground has to be qualified for ADAMTSL1 in particular, and the qualification is a measured negative rather than an absence: in the one experiment that tested it, recombinant human ADAMTSL-1 did not bind the N-terminal half of fibrillin-1 in the same SPR panel in which ADAMTSL-2, ADAMTSL-3 and papilin did - and the same experiments report the C-terminal half negative for every ADAMTSL tested, albeit as data not shown, so the exclusion covers fibrillin-1 entire - and it did not bind the non-catalytic region of ADAMTS-10 either. Only the fibrillin-1 half of that is discriminating, and the two must not be double-counted: ADAMTSL-2 is equally negative against ADAMTS-10 and still reaches GO:0030198, ADAMTS-10 binding was shown only for ADAMTSL-3, and THSD4 was never tested against it. The load-bearing result is the fibrillin-1 negative alone, and it is sufficient: the specific route by which the relatives reach this term is excluded for ADAMTSL1. That does not refute GO:0030198, which is far broader than fibrillin-1 binding, but it removes the strongest mechanistic analogy and it is a further reason the term cannot be treated as core here. Third, PAINT arrived at the same term independently: GO:0030198 is an IBD annotation at PTN000347317, seeded by fourteen gene sources (eleven mouse, two fly, one nematode), and that node sits above ADAMTSL1. That last step is what makes this ground bear on the gene rather than merely on its family, so it is worth stating how it is established: mouse Adamtsl1 is in PANTHER subfamily SF157, the same subfamily as human ADAMTSL1, and it receives the GO:0031012 IBA from PTN000347317, so that node is an ancestor of SF157 and therefore of ADAMTSL1. All three are family-level, and that is the limit of the case. What the annotation does not have is any ADAMTSL1 evidence. The laboratory that discovered the protein states that its tissue function is unknown and that matrix assembly and turnover is an expectation drawn from relatives; being resident in the matrix and being an MMP10 substrate places ADAMTSL1 within matrix remodelling but does not show that it organises the matrix. Hence KEEP_AS_NON_CORE rather than ACCEPT, since ACCEPT would assert this as the gene's core function and the evidence does not reach that. Not MARK_AS_OVER_ANNOTATED either, on the three grounds above. The concurrent ADAMTSL5 review marks its identical InterPro IEA MARK_AS_OVER_ANNOTATED; the two genes are in the same IBA position for this term, because neither receives the GO:0030198 IBA from PTN000347317 that ADAMTSL2, ADAMTSL4 and THSD4 do (computed in the bioinformatics census, section 1). They are no longer in the same evidentiary position overall, and the difference now runs against this review rather than for it: ADAMTSL5 has its own IDA to GO:0031012 plus microfibril and heparin binding, while ADAMTSL1 has a measured negative against the very route by which the family reaches GO:0030198, so the better-supported gene is the one taking the harsher action. That is an argument for settling the line family-wide rather than a reason to change one gene in isolation, so it is carried into suggested_questions rather than resolved here. Supporting Evidence: PMID:28722276 The function of punctin-1 in tissues is currently unknown, however, by analogy with other family members (Apte 2009; Dubail and Apte 2015; Hubmacher and Apte 2015), there is a strong possibility that it mediates the assembly and turnover of extracellular matrix at the affected sites. PMID:24281761 we identified the extracellular matrix protein ADAMTS-like protein 1 (ADAMTSL1) as a direct MMP10 substrate file:human/ADAMTSL1/ADAMTSL1-bioinformatics/RESULTS.md Entries with a CATALYTIC ACTIVITY comment: none. file:human/ADAMTSL1/ADAMTSL1-deep-research-affinage.md Direct biochemical demonstration of an enzymatic activity for the mammalian protein has not been established in the available corpus. file:interpro/panther/PTHR13723/PTHR13723-paint.tsv PTHR13723 PTN000347317 GO:0030198 P IBD false PMID:22242013 ADAMTSL-2, -3, and papilin polypeptides interacted with the N-terminal half of fibrillin-1, while ADAMTSL-1 did not. Knowledge gap: Whether ADAMTSL1 itself participates in organising the extracellular matrix, as opposed to merely residing in it and being turned over within it, has never been tested. OPEN BIOLOGY BP_DARK Resolve: A microfibril or matrix-assembly assay using recombinant full-length ADAMTSL1, together with matrix phenotyping of an Adamtsl1-null mouse, would settle it. Both have been blocked in practice by the absence of a full-length expression construct. "The function of punctin-1 in tissues is currently unknown, however, by analogy with other family members (Apte 2009; Dubail and Apte 2015; Hubmacher and Apte 2015), there is a strong possibility that it mediates the assembly and turnover of extracellular matrix at the affected sites." — PMID:28722276 "A plasmid construct for full-length ADAMTSL1 is presently unavailable." — PMID:28722276 "ADAMTSL-2, -3, and papilin polypeptides interacted with the N-terminal half of fibrillin-1, while ADAMTSL-1 did not." — PMID:22242013 |
| GO:0005788 endoplasmic reticulum lumen | TAS Reactome:R-HSA-5173005 | KEEP AS NON CORE | Summary: True but non-core. ADAMTSL1's thrombospondin repeats really are glucosylated by B3GLCT in the ER lumen, so the protein occupies that compartment during biosynthesis; it acts in the extracellular matrix. Reason: The underlying biology is experimentally supported for this gene rather than inferred from pathway membership alone: mass spectrometry of punctin-1 detected the fucose-glucose disaccharide that B3GLCT produces on TSR2, TSR3 and TSR4, and mutating the modified residues reduces secretion. The compartment assignment is therefore correct. It is marked non-core because the endoplasmic reticulum lumen is a biosynthetic transit compartment for a secreted protein, not its site of action, and because the Reactome pipeline projects the compartment of a reaction's participant set: this reference annotates about nine gene products with the same three terms, so it is one statement about set membership rather than an observation about ADAMTSL1. Supporting Evidence: PMID:17395588 Mutation of the putative modified Ser/Thr residues in TSR2, TSR3, and TSR4 led to significantly decreased levels of secreted punctin-1. |
| GO:0005788 endoplasmic reticulum lumen | TAS Reactome:R-HSA-5173192 | KEEP AS NON CORE | Summary: Same statement as the B3GLCT row, from the upstream reaction of the same pathway. POFUT2 acts on ADAMTSL1's thrombospondin repeats in the ER lumen; kept as a transit location rather than a functional one. Reason: POFUT2 is the ER-resident enzyme that installs the O-fucose on which B3GLCT builds, and ADAMTSL1 is a demonstrated substrate: the O-fucose was mapped to Thr48, Thr312, Ser391 and Thr451 by mass spectrometry, and C-mannosylation of the adjacent tryptophans in the same repeat was mapped in the companion study, with both modifications acting together as a gate on export. The compartment is correct and better supported than the other two rows, but it remains a biosynthetic transit compartment rather than the site where the protein functions. This is one of three GOA rows asserting the identical term for the same gene from three reactions of one Reactome pathway. Supporting Evidence: PMID:19671700 Together, these modifications appear to provide a quality control mechanism for punctin-1 secretion. PMID:17395588 the data define a critical role for N-glycosylation and O-fucosylation in the biosynthesis of punctin-1 |
| GO:0005788 endoplasmic reticulum lumen | TAS Reactome:R-HSA-6785565 | KEEP AS NON CORE | Summary: The weakest of the three identical rows: it derives from a Reactome FailedReaction modelling the Peters-plus disease state, in which the B3GALTL G393E variant does not glucosylate its substrates. The location it asserts is still correct, but it carries no information the two working reactions do not already supply. Reason: Verified against the Reactome content service: R-HSA-6785565 has schemaClass FailedReaction, its catalyst is the B3GALTL G393E disease variant, and it has inputs but no outputs. The GO pipeline emits the compartment of the input participant set regardless of whether the reaction proceeds, so ADAMTSL1 receives the same endoplasmic reticulum lumen assignment it already has from R-HSA-5173005. Kept because the compartment claim is not wrong and ADAMTSL1 is a genuine B3GLCT substrate, but marked non-core on the same transit-compartment grounds as its two siblings, and flagged as redundant. Whether a FailedReaction should emit a location annotation identical to that of the corresponding working reaction is raised as a question for the Reactome-GO pipeline rather than treated as a defect in this gene's record. Supporting Evidence: PMID:17395588 Mutation of the putative modified Ser/Thr residues in TSR2, TSR3, and TSR4 led to significantly decreased levels of secreted punctin-1. |
| GO:0031012 extracellular matrix | IDA Q8N6G6-1 PMID:11805097 Punctin, a novel ADAMTS-like molecule, ADAMTSL-1, in extrace... | NEW | Summary: Proposed new annotation. ADAMTSL1 is the only one of the 26 human members of PANTHER family PTHR13723 with no extracellular matrix annotation of any kind, even though its matrix deposition is directly observed in human cells, endogenous untagged protein is recovered from human fibroblast secretomes (PMID:24281761), and its mouse orthologue receives the term by IBA from the same PANTHER node. Reason: The gap was found by census rather than by inspection. PAINT holds GO:0031012 as an IBD annotation at node PTN000347317; 24 of the 26 human members of PTHR13723 receive it by IBA from that node, and the one further exception, PAPLN, holds the term from its own evidence, which is the expected behaviour since PAINT does not lay an IBA over a direct annotation. ADAMTSL1 has neither. The gap is species-specific rather than subfamily-specific: mouse Adamtsl1 (Q8BLI0), the same PANTHER subfamily SF157, does receive the IBA from PTN000347317 and additionally carries three HDA rows from matrisome proteomics. Human evidence is not lacking. UniProt records the location as supported by three experimental references, the founding paper shows recombinant punctin deposited punctately into the cell substratum and excluded from focal contacts, and endogenous untagged ADAMTSL1 was recovered from human fibroblast secretomes and named as a direct MMP10 substrate. The caveat on the primary evidence is that the 2002 assay used epitope-tagged protein in a heterologous cell line, which is why the secretome result is cited alongside it: it is native protein in a primary human cell type, and it is independent of the tag. The isoform field records that the assayed material was the 525-residue punctin-1 form, which is what was purified and expressed; it is not a claim that the localisation is restricted to that isoform, and it should not be read as one. Isoforms 1 to 4 all retain the signal peptide, so all are expected to be secreted, while isoforms 5 and 6 delete it and would not be. Supporting Evidence: PMID:11805097 In transfected COS-1 cells, punctin is deposited in the cell substratum in a punctate fashion and is excluded from focal contacts. PMID:24281761 we identified the extracellular matrix protein ADAMTS-like protein 1 (ADAMTSL1) as a direct MMP10 substrate file:human/ADAMTSL1/ADAMTSL1-bioinformatics/RESULTS.md **Members with no `GO:0031012` annotation of any kind: ADAMTSL1.** file:interpro/panther/PTHR13723/PTHR13723-paint.tsv PTHR13723 PTN000347317 GO:0031012 C IBD false |
Loading supporting content…
Download this section (compressed HTML)Q: Where should the family draw the line between KEEP_AS_NON_CORE and MARK_AS_OVER_ANNOTATED for the InterPro GO:0030198 IEA that every ADAMTS-like protein carries? ADAMTSL1 and ADAMTSL5 are in the same IBA position for this term: both hold the IPR013273 IEA, and neither receives the GO:0030198 IBA from PTN000347317 that ADAMTSL2, ADAMTSL4 and THSD4 do. Their wider evidence now differs, and in the direction that cuts against this review: ADAMTSL5 has its own IDA to GO:0031012 plus microfibril and heparin binding, while ADAMTSL1 has a measured SPR negative against fibrillin-1, the route by which the family reaches GO:0030198. The better-supported gene is thus the one taking the harsher action, which is an argument for settling the line family-wide rather than a tie to be broken per gene. The concurrent ADAMTSL5 review marks the row MARK_AS_OVER_ANNOTATED; this review marks it KEEP_AS_NON_CORE, on the grounds that the term is the level of generality at which the heterogeneous family agrees, that PAINT independently placed it as an IBD at a node covering both genes, and that a non-catalytic route to it is demonstrably available to this architecture. The difference between the genes is therefore real but points the wrong way to explain the divergence: it would justify the harsher action on ADAMTSL1, not the milder one, so it cannot be what separates the two verdicts. The line should be settled once for the family rather than gene by gene.
Suggested experts: GO Central / PAN-GO curators
Q: Should the molecular-function keyword Hydrolase (KW-0378) be removed from Q8N6G6? It generates the GO:0016787 hydrolase activity cross-reference in the entry's own GO list, while the same entry's CAUTION comment states that the metalloprotease and disintegrin-like domains are absent, and there is no CATALYTIC ACTIVITY comment, no EC number and no active-site or metal-binding feature. Of the six human ADAMTS-like proteins, only ADAMTSL1 and THSD4 carry the keyword; ADAMTSL2, ADAMTSL3, ADAMTSL4 and ADAMTSL5 carry the identical CAUTION and no molecular-function keyword. Should THSD4 (Q6ZMP0), which carries the keyword but has no CAUTION comment at all, be reviewed at the same time? The annotation no longer reaches GOA, since keyword-derived annotations were withdrawn for cellular organisms, so this is a defect that is live in UniProt and invisible in GO.
Suggested experts: UniProt curators
Q: Why does human ADAMTSL1 carry no GO:0031012 extracellular matrix annotation at all, when the term is an IBD at node PTN000347317, 24 of the 26 human members of PTHR13723 receive it by IBA from that node, and mouse Adamtsl1 in the same subfamily SF157 does receive it from the same node? Two human members miss the IBA. PAPLN is the other, and its case is not explained by PAINT declining to overlay an existing direct annotation, because its only GO:0031012 rows are IEA and TAS, neither experimental; the concurrent ADAMTSL5 review sharpens that further by showing three PAPLN orthologs are themselves seeds at the node. ADAMTSL1 is the one that ends up with no annotation at all. Both look like propagation coverage gaps rather than curatorial decisions, and in ADAMTSL1's case the gap has landed on the family member whose matrix deposition is directly observed in human cells.
Suggested experts: GO Central / PAINT curators
Q: Should PAINT's loss calls at node PTN002673039 propagate to the other ADAMTS-like members? The node carries an inferred-from-key-residues loss of GO:0004222 metalloendopeptidase activity and an inferred-from-rapid-divergence loss of GO:0006508 proteolysis, correctly recording that catalysis was lost on this branch, but in GOA that judgement surfaces as exactly one annotation, ADAMTSL2's negated GO:0004222 row. For ADAMTSL1 the negated annotation would be the machine-readable form of a statement UniProt already makes in prose, and it would inoculate the gene against the recurring reappearance of proteolytic activity in family-level summaries.
Suggested experts: GO Central / PAINT curators
Q: Should the PANTHER family GO-slim and protein-class view be made aware of PAINT's own loss calls? The geneinfo service returns GO_SLIM_MF GO:0004222 metalloendopeptidase activity, GO_SLIM_BP GO:0006508 proteolysis and protein class PC00153 metalloprotease for every member of PTHR13723 queried, including all six ADAMTS-like proteins and including ADAMTSL2, whose GOA record explicitly negates GO:0004222, so a non-catalytic member is summarised as a metalloprotease.
Suggested experts: PANTHER curators
Q: What action should a review record for a biosynthetic-transit compartment row? ADAMTSL1 and ADAMTSL4 carry the identical three GO:0005788 TAS rows from the same three Reactome reactions; the merged ADAMTSL4 review records them as ACCEPT while its reason describes them as reflecting secretory pathway transit rather than a functional localisation, and this review records the same judgement as KEEP_AS_NON_CORE. A convention would remove the discrepancy in either direction.
Q: Should Reactome FailedReactions contribute location annotations to GO? R-HSA-6785565 models a disease state in which glucosylation does not occur, yet it emits a GO:0005788 location annotation identical to the one from the corresponding working reaction R-HSA-5173005, adding a third identical row here without adding information.
Suggested experts: Reactome curators
Q: What does ADAMTSL1 bind, given that the obvious family candidates are already excluded? Unlike ADAMTSL2, ADAMTSL3, THSD4 and papilin, recombinant human ADAMTSL-1 did not bind the N-terminal half of fibrillin-1, and it did not bind the non-catalytic region of ADAMTS-10 either, in the same study. Two questions follow. Does the negative extend to the 1762-residue isoform, since the paper describes its reagent only as full length and the group that made the protein stated in 2017 that no long-form construct existed? And if ADAMTSL1 is genuinely outside the fibrillin-1 pathway that unites the rest of the family, what does the prototype of the ADAMTS-like family actually do? FBN2 appears among its affinity-purification hits and fibrillin-2 has not been tested, which is the one family- shaped hypothesis still standing.
Suggested experts: Suneel S. Apte, Dirk Hubmacher
Experiment: Build a full-length ADAMTSL1 expression construct and purify the 1762-residue protein. This is the rate-limiting step for everything else: the 2017 study states that no such construct existed, so every published biochemical result is confined to the 525-residue punctin-1 splice form.
Hypothesis: The nine thrombospondin repeats, four immunoglobulin-like domains and PLAC domain that are absent from punctin-1 carry functions that no published experiment has been able to test.
Type: recombinant protein expression
Experiment: An unbiased partner search rather than another candidate test: pull-down or proximity- labelling from a matrix-producing human cell line using recombinant full-length ADAMTSL1 as bait, with mass-spectrometric identification. Candidate SPR should be reserved for fibrillin-2 and heparin, which are not reported as tested anywhere, and for LTBP1 and LTBP4, whose ADAMTSL1 results are absent from that study's main text but may sit in a supplementary table not present in the cached record and are worth checking before assaying. Use ADAMTSL2 or ADAMTSL3 as the positive control. Fibrillin-1 and the non- catalytic region of ADAMTS-10 should be included only as controls, since recombinant human ADAMTSL-1 has already been shown not to bind either, and repeating them with the 1762-residue isoform would additionally test whether that published negative extends to the long form.
Hypothesis: ADAMTSL1 has a matrix binding partner outside the fibrillin-1 / ADAMTS-10 axis that unites the rest of the ADAMTS-like family.
Type: protein-protein interaction assay
Experiment: Assay fibrillin microfibril deposition in fibroblast cultures with ADAMTSL1 knocked down or supplemented with recombinant protein, the standard test used to establish the matrix-organising role of the other ADAMTS-like proteins.
Hypothesis: ADAMTSL1 promotes or restrains fibrillin microfibril deposition, which would make GO:0030198 a gene-specific rather than a family-level annotation.
Type: cell-based matrix assembly assay
Experiment: Phenotype an Adamtsl1-null mouse for the tissues where the human phenotype localises - ocular anterior segment and intraocular pressure, mandibular condylar cartilage, dentition, and skeletal muscle - and compare the peripheral-nerve phenotype directly with that of Pmp2-positive Schwann cell ablation.
Hypothesis: The peripheral-nerve phenotype attributed to Pmp2-positive Schwann cells is a property of that cell population rather than of Adamtsl1, which is only one of its markers.
Type: mouse knockout phenotyping
Experiment: Map the MMP10 cleavage site in ADAMTSL1 and test whether the released fragment has activity of its own, since a matrix protein whose only reported enzymatic relationship is being a substrate may be functioning through a processed form.
Hypothesis: ADAMTSL1 acts through an MMP10-generated fragment rather than as an intact protein.
Type: degradomics and fragment activity assay
Experiment: Localise endogenous ADAMTSL1 protein, not transcript, in adult human skeletal muscle and in the ocular anterior segment. All existing localisation data come from recombinant protein in transfected cells or from secretome proteomics, so an antibody-based localisation in the affected tissues would put the extracellular matrix annotation on native protein in its own tissue context.
Hypothesis: Native ADAMTSL1 occupies the extracellular matrix of the tissues where its loss produces disease, not only that of transfected cells in culture.
Type: immunohistochemistry
What is not known — curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: No molecular function is known for ADAMTSL1: no ligand, no binding partner, no activity and no structural role has been demonstrated, and the gene carries no molecular-function annotation of any evidence code in GOA.
OPEN BIOLOGY MF_DARK
What is known: Well established: the protein's domain content (thrombospondin type-1 repeats, four Ig- like C2-type domains, ADAMTS spacer, PLAC domain, no catalytic domains), its secretion and matrix deposition, the glycosylation chemistry that gates its export, its cleavage by MMP10, and a dominant-negative human allele. Also now bounded on the other side by measured negatives rather than by silence: recombinant human ADAMTSL-1 did not bind the N-terminal half of fibrillin-1 in an SPR panel where ADAMTSL-2, ADAMTSL-3 and papilin did, and the same experiments report the C-terminal half negative for every ADAMTSL tested, which is what licenses stating the exclusion as fibrillin-1 binding rather than N-terminal fibrillin-1 binding, though that second half is given as data not shown. ADAMTSL-1 also did not bind the non-catalytic region of ADAMTS-10. Only the fibrillin-1 result discriminates ADAMTSL1 from its relatives - ADAMTSL-2 is equally negative against ADAMTS-10 and ADAMTSL-6/THSD4 was never tested against it - so what is excluded is fibrillin-1 binding, with the ADAMTS-10 result corroborating rather than adding to it, and both subject to the caveat that the construct's isoform is not stated. Entirely open: what the protein does bind, and what it does once it is in the matrix.
Significance: ADAMTSL1 is the prototype of the ADAMTS-like family and is implicated in a multisystem developmental disorder, yet has no molecular handle at all. Until a partner is identified the disease mechanism cannot be modelled beyond loss of secretion, and the family-level extracellular matrix organization inference cannot be made specific.
What would resolve it: Identification of a physical binding partner for recombinant full-length ADAMTSL1 would supply the first molecular-function term. The obvious family candidates are now poorer candidates than they look: fibrillin-1 and ADAMTS-10 were both tested by SPR in PMID:22242013 and both were negative, so an unbiased search is more appropriate than another fibrillin-1 experiment. Fibrillin-2 is not reported as tested anywhere, and FBN2 appears among the affinity-purification hits, so it is the one family-shaped hypothesis still standing. No LTBP result for ADAMTSL1 appears in the main text of PMID:22242013 either, but LTBP-1 and LTBP-4 were coupled to chips in the same experiments and the LTBP data sit in a supplementary table absent from the cached record, so untested is not a safe reading for those. Producing a construct for the 1762-residue form is the prerequisite for all of it, and would also determine whether the published negatives extend to the long isoform.
Provenance (the field's own admissions):
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)