ULK1 Membrane Recruitment Stops at WIPI3-ATG13 and the PVP Motif

The PI3P paradox that ULK1 membrane recruitment had to solve

For years the field accepted that ULK1C is stabilized at phagophore initiation sites by PI3P, an idea originally attributed to basic residues in ATG13 [1]. But ULK1, FIP200, ATG13, and ATG101 contain no canonical PI3P-binding domain, and structural work failed to find a PI3P site in ATG13 [1]. The ULK1C also has to be coordinated with PI3KC3-C1, which produces PI3P and whose subunits—like WIPI2 and ATG16L1—are themselves ULK1 substrates [1][6]. The anchor paper's introduction frames this as the central unresolved question: how does a complex without a lipid-binding module become PI3P-dependent [1]?

Earlier work offered partial answers. ULK1 palmitoylation by ZDHHC13 was reported to drive translocation of the complex to autophagosome formation sites and to enhance ATG14L phosphorylation [5], and TRIM27/STK38L was shown to restrain ULK1 levels and activity through ubiquitination and Ser495 phosphorylation [7]. Neither explains PI3P dependence. The 2025 Nature Cell Biology reconstitution of BNIP3/NIX mitophagy supplied the missing conceptual link: transmembrane cargo receptors can initiate autophagosome biogenesis through a WIPI–ATG13 complex rather than through FIP200/ULK1, establishing that WIPI–ATG13 engagement is a bona fide initiation route [2][8]. The anchor paper generalizes that route to canonical, PI3P-dependent ULK1C recruitment [1].

WIPI3 docks ATG13 at a DHF motif distinct from the WIPI2 site

The anchor study used DeepMSA2 alignments and AlphaFold2 models to identify a conserved Asp213-His214-Phe215 (DHF) motif in the ATG13 IDR, with a β-strand-(X)n-Φ-X-X-Φ-X-X-X-Ψ-F signature resembling the WIPI-binding region of ATG2A but divergent from the ATG16 W2IR motif [1]. In the predicted ATG13(1–220):ATG101:WIPI3 complex, the His-Phe dyad sits in the WIPI3 WIR pocket, with WIPI3 Lys42, Lys44, Thr126, and His127 contacting ATG13 and Asp217 positioned for salt bridges [1]. GST-bead binding assays confirmed a clear preference for WIPI3 over WIPI4, and the ATG13(H214D,F215D) double mutant—termed HF|DD—nearly abolished WIPI3 binding [1].

Critically, the same HF|DD mutation did not substantially affect the WIPI2:ATG13:ATG101 interaction, and a WIPI3(4D) mutant (K42D, K44D, T126D, H127D) showed profoundly reduced binding [1]. This separates two docking modes: WIPI3 uses the newly defined DHF/W3IR motif, whereas WIPI2 engages ATG13:ATG101 through a distinct, previously described site [1][2]. The 2025 mitophagy reconstitution had already shown NIX and BNIP3 recruit WIPI proteins that then engage the ULK1 complex via ATG13/101, while FUNDC1 and BCL2L13 use FIP200 instead [2]. The anchor paper now assigns a specific sequence determinant to the WIPI3 arm of that route [1].

The ATG101 WF finger and CTH insert into the bilayer

With GUVs containing 5% PI3P, 200 nM WIPI3 robustly recruited ATG13(1–230)-GFP:ATG101 to membranes; neither PI3P nor WIPI3 alone sufficed [1]. Mutating ATG101 Trp110 and Phe112 to Asp (WF|DD) substantially diminished membrane binding, CTH truncation produced a quantifiable defect, and the double mutant behaved like WF|DD alone—indicating the WF finger is the major membrane-binding contributor [1]. At physiologically relevant concentrations of 40–50 nM, the WF finger was required for efficient recruitment, whereas residual binding in its absence was primarily driven by the WIPI3–PI3P interaction [1]. In SUV assays at a superphysiological 1 μM protein concentration, high curvature exposed packing defects and allowed substantial WT binding, a permissive condition that can bypass normally required signals [1].

This finally explains why ATG101 is essential and why its WF finger matters. ATG101's only known prior role was bridging ULK1C to ATG9 via the ATG9 C-terminal IDR, but disrupting that site only partially blocks autophagy, and the WF residues are not involved in ATG9 binding [1]. The anchor paper argues the WF finger's essential function is synergistic membrane recruitment with WIPIs, sufficient to account for the decade-old phenotype [1]. The authors also note that their models correspond to the established crystal structure of human ATG13:ATG101 and that none of the findings depend on HORMA metamorphosis [1].

A PVP motif delivers the ULK1 kinase domain to the membrane

Even with the HORMA dimer anchored, the ULK1 kinase domain sits far from the membrane: the KD is separated from the EAT domain by a ~500-residue IDR, plus ~150 residues of ATG13 IDR, leaving the KD free to explore roughly 6 × 10⁴ nm³ with a maximum membrane–KD separation near 40 nm [1]. The anchor study identified a KD-proximal Pro-Val-Pro (PVP) motif in the ULK1 IDR that docks onto the ATG13:ATG101 HORMA dimer, halving the maximum separation to ~20 nm and bringing the kinase to its membrane-localized substrates [1]. The PVP motif was essential for in vitro ULK1 phosphorylation of ATG16L1 and important for autophagy and mitophagy [1].

The ATG16L1 connection matters because ATG16L1 phosphorylation at Ser278 is a marker of newly forming autophagosomes and tracks the rate of LC3B lipidation, though S278 mutation is dispensable for starvation-induced autophagy activation [4]. The anchor paper's cell-based assays show WIPI3:ATG13 engagement promotes ATG16L1 phosphorylation, autophagy, and mitophagy [1]. In ULK1-KO ARPE-19 cells, reexpressed WT ULK1 restored DFP-induced mitophagy and reduced mitochondrial OMI, whereas an ADA mutant—disrupting the HORMA interaction—did not; ADA also impaired ULK1 puncta formation and reduced ATG13 co-immunoprecipitation [1]. This positions the PVP–HORMA contact as a functional requirement, not a structural curiosity.

Where the two-step model stops and what remains open

The conclusions rest mainly on in vitro reconstitution, molecular dynamics, and cell-based assays, and the authors explicitly note these cannot fully represent dynamic events on autophagic membranes in vivo [1]. The WIPI3 literature is also not uniform: one CRISPR knockout study found no effect on starvation-induced autophagy flux in HEK293 cells, which the authors acknowledge is unexplained and requires further study, while the preponderance of published data—including WIPI3 co-IP and colocalization with FIP200 upon starvation—supports a role in ULK1C recruitment [1]. The anchor paper also notes that ULK1 EAT Cys palmitoylation is compatible with its pathway, that RAB1A binding appears dominant for PI3KC3-C1 localization, and that ATG8 scaffolding of ULK1 likely acts downstream—interplays left for future work [1][5].

Broader context reinforces the multistep framing. WIPI2 recruits ATG12–ATG5–ATG16L1 to PI3P-positive membranes, and ATG16L1 mutants that cannot bind WIPI2 fail to rescue starvation-induced autophagy despite binding FIP200 [6]. STING can directly bind WIPI2 and bypass canonical upstream machinery for LC3 lipidation, showing that WIPI proteins are convergence points for multiple initiation modes [3]. The 2025 reconstitution showed remarkable hierarchical flexibility—NIX/BNIP3 use WIPI–ATG13, FUNDC1/BCL2L13 use FIP200/ULK1, and FKBP8/TEX264 can use both [2][8]. The anchor paper's contribution is to define the molecular rules for one arm of that flexibility, but whether the two-step WIPI3–ATG13–PVP pathway operates identically across cell types, receptors, and organisms remains an open question [1][2].

About These Sources

This research page is built on 8 peer-reviewed studies — published from 2014 to 2026, 4 from 2024 or later, collectively cited 1,058 times — selected as the most relevant from 13 studies that passed quality screening, drawn from 61 papers retrieved from a database of over 500 million.

Sources used in this answer