PCSK1N (ProSAAS / Q9UHG2) — review notes

This review is in response to upstream curator question
geneontology/go-annotation#6407
("term suitable to describe PCSK1N (ProSAAS)") asking which GO term best captures
ProSAAS's emerging "chaperone activity".

Gene identity

HGNC symbol: PCSK1N (HGNC:17301)
UniProt: Q9UHG2, human ProSAAS, 260 aa precursor; signal peptide 1–33; chain
ProSAAS 34–260
Gene names / aliases: ProSAAS, pro-SAAS, Proprotein convertase subtilisin/kexin type 1
inhibitor
Family: PANTHER PTHR15531 (PROSAAS), Pfam PF07259, InterPro IPR010832
Closest functional/structural homolog in mammals: SCG5/7B2 (similar small
neuroendocrine secretory protein with both convertase-regulatory and anti-aggregant
chaperone activities) PMID:24102330.
Tissue specificity: brain and pancreas; neurons (not glia); pituitary, adrenal,
pancreas; widespread in CNS [UniProt; PMID:10632593; PMID:24102330].

Curated proteolytic processing

ProSAAS is processed at paired-basic sites in the regulated secretory pathway
(Golgi/TGN, secretory granules) into multiple bioactive peptides
[UniProt; PMID:10632593; PMID:11435430; PMID:12914799]:

KEP (34–40)
Big SAAS (34–59) → Little SAAS (42–59)
ProSAAS(1–180) [≈21 kDa intermediate; major secreted form]
Big PEN-LEN (221–260) — contains the C-terminal convertase-inhibitory domain
PEN (221–242) — endogenous GPR83 ligand (food intake) [UniProt: by similarity]
Big LEN (245–260) → Little LEN (245–254) — endogenous GPR171 ligand
(feeding) [UniProt: by similarity]

The motif L-L-R-V-K-R (residues 239–244) within the C-terminal inhibitory
domain is necessary and sufficient for PCSK1 inhibition; alanine substitutions of
K243 or R244 abolish inhibition PMID:11435430.

Function 1 — Endogenous PCSK1 (PC1/3) inhibitor

The original characterization showed that:

"Purified proSAAS inhibits prohormone convertase 1 activity with an IC50 of 590 nM
but does not inhibit prohormone convertase 2"
[PMID:10632593, "Purified proSAAS inhibits prohormone convertase 1 activity with
an IC(50) of 590 n m but does not inhibit prohormone convertase 2"]
"Overexpression of proSAAS in the AtT-20 cells substantially reduces the rate of
processing of the endogenous prohormone proopiomelanocortin"
[PMID:10632593, "Overexpression of proSAAS in the AtT-20 cells substantially
reduces the rate of processing of the endogenous prohormone proopiomelanocortin"]
Mutagenesis maps the inhibitory determinants to the LLRVKR motif:
"K->A: Abolishes inhibition of PCSK1" / "R->A: Abolishes inhibition of PCSK1"
for K243 and R244 [PMID:11435430, UniProt MUTAGEN annotations]
ProSAAS and Big PEN-LEN, but not the further-processed peptides, reduce PCSK1
activity in the ER and Golgi [UniProt FUNCTION: by similarity, citing Q9QXV0]

PCSK1/PC1/3 is a serine endopeptidase (subtilisin-like, MEROPS S8). Therefore the
appropriate molecular function for the C-terminal inhibitory peptide / Big PEN-LEN
is GO:0004867 serine-type endopeptidase inhibitor activity (already on the
record, IEA), with the parent GO:0004866 endopeptidase inhibitor activity
(already there as IBA/TAS) acceptable as a less specific synonym. The biological
process is GO:0010955 negative regulation of protein processing and/or
GO:0060570 negative regulation of peptide hormone processing.

Function 2 — Anti-aggregant / extracellular chaperone activity (the topic of issue #6407)

A coherent body of work, primarily from the Lindberg laboratory, establishes that
full-length ProSAAS or its 21-kDa N-terminal fragment (residues ~1–180) acts as a
secreted brain chaperone that prevents aggregation of multiple amyloidogenic
client proteins, but does not refold proteins and does not disaggregate
preformed fibrils.

β-amyloid (Aβ1–42)

Hoshino et al. 2014 PMID:24102330:

"We here describe a novel anti-aggregant chaperone function for the
neuroendocrine protein proSAAS"
[PMID:24102330, "We here describe a novel anti-aggregant chaperone function for
the neuroendocrine protein proSAAS"]
"In vitro, proSAAS efficiently prevented the fibrillation of Aβ(1-42) at molar
ratios of 1 : 10, and this anti-aggregation effect was dose dependent. Structure-
function studies showed that residues 97-180 were sufficient for the anti-
aggregation function against Aβ"
[PMID:24102330, "In vitro, proSAAS efficiently prevented the fibrillation of
Aβ(1-42) at molar ratios of 1 : 10..."]
"proSAAS immunoreactivity was highly colocalized with amyloid pathology…
Immunoreactive proSAAS co-immunoprecipitated with Aβ immunoreactivity in lysates
from APdE9 mouse brains"
[PMID:24102330, "proSAAS immunoreactivity was highly colocalized with amyloid
pathology"]
"inclusion of recombinant proSAAS in the medium of Neuro2a cells, as well as
lentiviral-mediated proSAAS over-expression, blocked the neurocytotoxic effect
of Aβ(1-42) in Neuro2a cells"
[PMID:24102330, "blocked the neurocytotoxic effect of Aβ(1-42) in Neuro2a cells"]

This evidence supports MF GO:0001540 amyloid-beta binding and
GO:0051787 misfolded protein binding, plus BP GO:1905907 negative
regulation of amyloid fibril formation (and more specifically
GO:1902430 negative regulation of amyloid-beta formation is not appropriate
because GO:1902430's parent describes the secretase-driven generation of Aβ
peptide from APP, not aggregation; Hoshino et al. measured fibrillation, not APP
processing). MF GO:0044183 protein folding chaperone is not appropriate
because the curators (Hoshino et al., Jarvela et al.) explicitly state it does
not refold; the analogous CLU (clusterin) review in this repo also annotates
GO:0051787 misfolded protein binding rather than GO:0044183 for this kind of
"holdase-like" activity.

α-synuclein (aSyn)

Jarvela et al. 2016 PMID:27457957:

"proSAAS, widely expressed in neurons throughout the brain, is associated with
aggregated synuclein deposits in the substantia nigra of patients with Parkinson's
disease"
[PMID:27457957, "associated with aggregated synuclein deposits in the substantia
nigra of patients with Parkinson's disease"]
"Recombinant proSAAS potently inhibits the fibrillation of α-synuclein in an in
vitro assay; residues 158-180, containing a largely conserved element, are
critical to this bioactivity"
[PMID:27457957, "Recombinant proSAAS potently inhibits the fibrillation of
α-synuclein in an in vitro assay"]
"proSAAS-encoding lentivirus blocks α-synuclein-induced cytotoxicity in primary
cultures of nigral dopaminergic neurons, and recombinant proSAAS blocks
α-synuclein-induced cytotoxicity in SH-SY5Y cells"
[PMID:27457957, "proSAAS-encoding lentivirus blocks α-synuclein-induced
cytotoxicity in primary cultures of nigral dopaminergic neurons"]

Lindberg et al. 2022, in vivo extension PMID:35527562:

"Coinjection of proSAAS-encoding lentivirus profoundly reduced the motor
asymmetry caused by unilateral nigral AAV-mediated human aSyn overexpression.
This was accompanied by significant amelioration of the human aSyn-induced loss
of both nigral TH-positive cells and striatal TH-positive terminals"
[PMID:35527562, "Coinjection of proSAAS-encoding lentivirus profoundly reduced
the motor asymmetry…"]
"Following vagal administration of human aSyn-encoding AAV, the number of human
aSyn-positive neurites in the pons and caudal midbrain was considerably reduced
in mice coinjected with proSAAS-, but not GFP-encoding AAV, supporting proSAAS-
mediated blockade of transsynaptic aSyn transmission"
[PMID:35527562, "supporting proSAAS-mediated blockade of transsynaptic aSyn
transmission"]

This supports MF GO:0051787 misfolded protein binding (binding aSyn
aggregates) and BP GO:1905907 negative regulation of amyloid fibril formation
in vivo, plus the neuroprotective phenotype (process: protection against
neurodegeneration, captured by parent terms below).

TDP-43 — caveats about cytoplasmic activity

Peinado et al. 2022 PMID:35549000:

"We previously demonstrated that proSAAS, a small secreted neuronal protein,
exhibits potent chaperone activity against protein aggregation in vitro and
blocks the cytotoxic effects of amyloid and synuclein oligomers in cell culture
systems"
[PMID:35549000, "exhibits potent chaperone activity against protein aggregation
in vitro"]
"expression of proSAAS within the cytoplasm generates dense, membraneless 2 μm
proSAAS spheres which progressively fuse to form larger spheres, suggesting
liquid droplet-like properties"
[PMID:35549000, "expression of proSAAS within the cytoplasm generates dense,
membraneless 2 μm proSAAS spheres"]
"we demonstrate that proSAAS expression results in cytoprotection against
full-length TDP-43 toxicity in yeast. We conclude that proSAAS can act as a
functional holdase for TDP-43 via this phase-separation property"
[PMID:35549000, "proSAAS can act as a functional holdase for TDP-43 via this
phase-separation property"]

Important caveat: the TDP-43-encapsulating sphere phenotype was obtained by
forcing non-physiological cytoplasmic expression of proSAAS (signal peptide
removed). ProSAAS is normally a secreted protein that resides in the secretory
pathway (TGN, secretory granules) and the extracellular space. Therefore the
TDP-43 sphere data should not be used to support a cytoplasmic localization
annotation, but it does corroborate the protein's intrinsic capacity to bind
prion-like / aggregation-prone clients. The "holdase" characterization in this
paper is the strongest single piece of textual support for using
GO:0140309 unfolded protein holdase activity as a chaperone MF; however,
GO:0140309's definition emphasizes a carrier function ("escorts it to an
acceptor molecule or to a specific location"), which is not strictly demonstrated
for proSAAS in vivo.

Disaggregation: what proSAAS does NOT do

The upstream issue notes "Cannot disaggregate preformed fibrils (27457957)" and
this is consistent with the broader literature: proSAAS prevents initial
aggregation but does not disassemble preformed fibrils, and it does not refold
denatured proteins. Therefore terms in the disaggregase branch (e.g. ATPase
disaggregase activity) and protein folding chaperone GO:0044183 are not
appropriate.

Recommended additions / new annotations for the chaperone activity:

Aspect	GO term	Rationale
MF	GO:0051787 misfolded protein binding	Binds Aβ, aSyn, TDP-43 aggregates
MF	GO:0001540 amyloid-beta binding	Direct Aβ co-IP and binding PMID:24102330
MF	GO:0140309 unfolded protein holdase activity	Used by Lindberg 2022 for TDP-43; partial fit
BP	GO:1905907 negative regulation of amyloid fibril formation	In vitro and in vivo Aβ + aSyn fibrillation block
BP	GO:0050821 protein stabilization	Holdase-like maintenance of soluble client state

GO:0044183 (protein folding chaperone) is not recommended — proSAAS does not
refold clients. GO:1902430 (negative regulation of amyloid-beta formation) is
not recommended — that term is for APP cleavage, not Aβ aggregation.

Function 3 — Bioactive peptide ligand activity (after processing)

Big LEN and PEN, generated by paired-basic processing of proSAAS, are
GPCR ligands [UniProt FUNCTION: by similarity, Q9QXV0]:

Big LEN → endogenous ligand for GPR171, regulates feeding
PEN → endogenous ligand for GPR83, regulates feeding

This may support future processed-peptide receptor-binding annotations, but not
the existing precursor-level GO:0005102 signaling receptor binding annotation
(TAS, PMID:10632593). The literature actually supporting GPR171/GPR83 binding is
more recent (e.g. Wardman et al., Gomes et al.) than the cited PMID:10632593,
which itself does not demonstrate receptor binding. Processed peptide ligands
should be re-evidenced separately if annotated. The peptide-receptor interaction
can also support GO:0007218 neuropeptide signaling pathway (already
keyword-derived on record).

Function 4 — Disease association markers (not function per se)

Big SAAS / Little SAAS N-terminal fragments deposit with cytoplasmic tau
inclusions in Pick's disease, Alzheimer's disease, and amyotrophic lateral
sclerosis-parkinsonism/dementia complex of Guam [UniProt; PMID:12914799;
PMID:14746899]. This is a localization context rather than a discrete cellular
function, and should NOT be turned into a localization annotation (the deposit
is pathological).
ProSAAS is a recurrent CSF biomarker in dementia [reviewed in PMID:35549000].

Existing annotations — proposed actions

GOA term	Evidence	Proposed action	Rationale
GO:0004866 endopeptidase inhibitor activity	IBA, IEA, TAS	KEEP_AS_NON_CORE / MODIFY → GO:0004867	Parent of more-specific GO:0004867 (which is also already on record as IEA). PCSK1 is a serine endopeptidase (subtilisin) so the specific term is preferable.
GO:0004867 serine-type endopeptidase inhibitor activity	IEA	ACCEPT	Mutagenesis maps inhibitory determinants to LLRVKR (PMID:11435430); appropriate for the C-terminal inhibitory peptide.
GO:0005102 signaling receptor binding	TAS (PMID:10632593)	REMOVE	The cited PMID does not demonstrate receptor binding; future processed-peptide ligand annotations should use direct GPR83/GPR171 evidence if appropriate.
GO:0005576 extracellular region	IEA, ISS, TAS	ACCEPT	Secreted; well-established.
GO:0005794 Golgi apparatus	IEA	ACCEPT	True (parent of TGN).
GO:0005802 trans-Golgi network	IEA	ACCEPT	More specific Golgi sublocation; consistent with regulated secretion / processing site.
GO:0030141 secretory granule	IEA	ACCEPT	Stored in regulated secretory granules in neurons / endocrine cells [PMID:10632593, PMID:24102330].

Open questions for upstream curators (issue #6407)

Should "anti-aggregant" chaperones that prevent fibrillation but do not refold
be annotated to GO:0051787 misfolded protein binding + BP GO:1905907, or
to GO:0140309 unfolded protein holdase activity when there is no documented
carrier-to-acceptor handoff? A precedent in this repository is CLU/clusterin,
which uses the GO:0051787 + GO:1905907 + GO:0050821 combination for a closely
analogous extracellular anti-aggregant chaperone activity.
Is there an MF term capturing "holdase-like activity that prevents amyloid
fibrillation" that does not require either folding (GO:0044183) or carrier
delivery (GO:0140309)? If not, would a new term be warranted? (See similar
discussion in the SCG5/7B2 literature — Lindberg lab uses "anti-aggregant
chaperone" language for both 7B2 and ProSAAS.)
The TDP-43 sphere data PMID:35549000 is from forced cytoplasmic expression
and may not reflect physiological ProSAAS function. Should this evidence be
limited to "demonstrates capacity for prion-like client binding" rather than
used to support a normal cellular MF?

Relevant PMIDs

PMID:10632593 — Identification and characterization of proSAAS (PCSK1 inhibitor; tissue distribution; secretion)
PMID:11435430 — Mutagenesis of LLRVKR motif sufficient for PCSK1 inhibition
PMID:12914799 — N-terminal proSAAS fragment in tau deposits (Pick's disease)
PMID:14746899 — proSAAS in tau inclusions of AD and Guam
PMID:24102330 — Anti-aggregant chaperone for Aβ
PMID:27457957 — Anti-aggregant chaperone for α-synuclein (in vitro + neuronal cytoprotection)
PMID:35527562 — In vivo neuroprotection in PD models; blocks transsynaptic aSyn spread
PMID:35549000 — Holdase for TDP-43 via cytoplasmic phase separation; sphere formation

Description cleanup note

The YAML description field was revised to keep it as a standalone biological summary. Project-specific curation framing moved here instead.

Moved out of the YAML description: this review addresses upstream go-annotation issue #6407, where a curator asked which GO terms suitably capture ProSAAS chaperone activity. The N-terminal SAAS-derived peptide accumulation in tau inclusions is pathological and is not evidence for normal cytoplasmic localization.