ULK1 seen at the single-molecule level during autophagy initiation

ABSTRACT

Macroautophagy/autophagy research often involves overexpressing proteins to investigate their localization, function and activity. However, this approach can disturb the inherent balance of cellular components, potentially affecting the integrity of the autophagy process. With the advent of genome-editing techniques like CRISPR-Cas9, it is now possible to tag endogenous proteins with fluorescent markers, enabling the study of their behaviors under more physiologically relevant conditions. Nevertheless, conventional microscopy methods have limitations in characterizing the behaviors of proteins expressed at endogenous levels. This challenge can be overcome by single-molecule localization microscopy (SMLM) methods, which provide single-molecule sensitivity and super-resolution imaging capabilities. In our recent study, we used SMLM in combination with genome editing to explore the behavior of endogenous ULK1 during autophagy initiation, yielding unprecedented insights into the autophagy initiation process.

Abbreviation: ATG13: autophagy related 13; ATG14: autophagy related 14; ATG16L1: autophagy related 16 like 1; BECN1: beclin 1; ER: endoplasmic reticulum; GABARAPL1: GABA type A receptor associated protein like 1; MAP1LC3B: microtubule associated protein 1 light chain 3 beta; MTORC1: mechanistic target of rapamycin kinase complex 1; PALM: photo-activated localization microscopy; PIK3C3/VPS34: phosphatidylinositol 3-kinase catalytic subunit type 3; PIK3R4/VPS15: phosphoinositide-3-kinase regulatory subunit 4; PtdIns3P: phosphatidylinositol-3-phosphate; SMLM: single-molecule localization microscopy; ULK1: unc-51 like autophagy activating kinase 1; WIPI2: WD repeat domain, phosphoinositide interacting 2.

Autophagy relies on the assembly of various proteins into clustered structures, characterized by the accumulation of key components involved in the process. This clustering of proteins is essential for effective regulation as it increases local concentrations of molecules within the same pathway, thus facilitating interactions and enzymatic reactions. In autophagy, this phenomenon is prominent, with diverse regulators engaging in a spatiotemporal manner to initiate autophagosome formation. Specifically, ULK1 (unc-51 like autophagy activating kinase 1), a central kinase in autophagy initiation, forms clustered structures visible through confocal microscopy during starvation, preceding autophagosome formation. Unraveling the molecular basis of these structures is an important question for understanding the mechanisms underlying autophagy initiation. While conventional fluorescence or confocal microscopy can identify ULK1 clusters, these methods have limitations in providing detailed structural information. These limitations include challenges in precisely quantifying the number of molecules within the structures, distinguishing between clusters located closer than the diffraction limit, and discerning faint clusters against a fluorescent background. Quantitative single-molecule localization microscopy (qSMLM) addresses these limitations by providing single-molecule sensitivity and a spatial resolution of approximately 20 nm.

Studying protein clusters poses a challenge associated with the transient or stable overexpression of fluorescently labeled proteins, a common approach in the field. These methods can potentially influence protein oligomerization and spatial distribution, disturbing the natural stoichiometry within protein clusters and compromising the integrity of the autophagy pathway. Recent advancements of genome-editing techniques have established a standard practice for expressing proteins with a fluorescence tag at endogenous levels. In our recent study, we used CRISPR-Cas9 to endogenously tag ULK1 with the fluorescent protein mEos2, which possesses photoswitchable characteristics suitable for quantitative photo-activated localization microscopy (PALM), one of the SMLM techniques. This approach allowed us to quantitatively analyze endogenous ULK1 clusters with single-molecule sensitivity during autophagy initiation [1]. Through PALM, we were able to discern the size and shape of starvation-induced clusters, including arc and spherical structures, providing a level of detail unattainable with conventional fluorescence microscopy. Furthermore, PALM revealed substantially different clustering patterns when applied to overexpressed ULK1, highlighting the importance of analyzing endogenous molecules.

The PALM analysis of endogenous ULK1 clusters has unveiled several previously unrecognized characteristics of ULK1 during starvation. We observed that the majority of ULK1 exists in oligomeric states, forming clusters of various sizes distributed throughout the cytoplasm, even in fed cells. Notably, clusters containing more than 30 ULK1 molecules are exclusively found in starved cells (Figure 1). The number of these clusters increases during starvation, reaching a peak before autophagy downregulation, and subsequently declines. These starvation-induced clusters display diverse yet distinct shapes, including small, dense clusters, arc-shaped structures, and spherical formations, which are all absent in fed cells. These structures likely represent different stages of autophagy, with ULK1 initially forming dense clusters that trigger morphological membrane transitions, which then result in spherical autophagosomes. Our study implies that ULK1 molecules transition from low-order oligomeric states and accumulate to form a densely clustered state competent for autophagy initiation upon starvation. To further validate this finding, future investigations should involve a time-course analysis of these structures through quantitative SMLM experiments conducted in live cells.

Figure 1.

ULK1 forms clusters consisting of more than 30 molecules, preceding the formation of the phagophore and autophagosome. In this model, the ULK1 complex constitutively interacts with the ATG14-containing BECN1-PIK3R4/VPS15-PIK3C3/VPS34 complex, which is supported by our prior study. Under fed conditions, MTORC1 phosphorylates ULK1 and ATG14, suppressing the activities of these complexes. This role of MTORC1 prevents ULK1 clustering. In starved cells, MTORC1 inhibition leads to ULK1 clustering that occurs independently of ULK1 activity. ULK1 becomes activated, subsequently phosphorylating ATG14 and BECN1 to stimulate the lipid kinase activity of PIK3C3/VPS34, catalyzing the production of phosphatidylinositol-3-phosphate (PtdIns3p). PtdIns3P may facilitate clustering of the ULK1 complex and recruit other autophagy regulators, such as WIPI2, ultimately inducing structural changes in the ER membrane.

An advantage of using SMLM approaches is the ability to quantify intracellular protein interactions. Our two-color SMLM imaging revealed that all ULK1 clusters containing over 30 ULK1 molecules colocalize with ATG13, ATG14, ATG16L1, WIPI2, GABARAPL1, and MAP1LC3B. Even in fed cells, ULK1 exhibits some degree of colocalization with ATG13, ATG14, GABARAPL1, and WIPI2, albeit to a lesser extent. This suggests the presence of pre-assembled initiation complexes involving these proteins. Additionally, our investigation unveiled that the majority of ULK1 molecules, including those within starvation-induced structures containing more than 30 molecules, are predominantly located on or near the endoplasmic reticulum (ER). This finding aligns with the established concept that mammalian autophagosomes originate from the ER. Our results also offer compelling evidence for the presence of ULK1 in various low-order multimeric states at the ER in fed cells. Upon starvation, ULK1 undergoes assembly into autophagy-competent ULK1 clusters consisting of more than 30 ULK1 molecules and other autophagy regulators. These clusters form larger structures with arc-shaped and spherical morphologies, representing the intermediate structures in the process of autophagosome formation.

The established approach of quantitatively analyzing endogenous ULK1 molecules within cells has provided a valuable tool for studying the mechanisms that initiate ULK1 clustering, a process that precedes phagophore formation and drives autophagosome formation. By using chemical inhibitors for ULK1 and PIK3C3/VPS34, we obtained key clues of how these molecules are involved in the process. Specifically, we found that inhibiting PIK3C3/VPS34 activity prevents the formation of autophagy-competent ULK1 clusters during starvation. In contrast, ULK1 inhibition leads to about a two-fold increase in the number of clusters with more than 30 ULK1 proteins in starved cells. However, these clusters are notably smaller and contain substantially fewer other autophagy regulators. Importantly, these clusters are stalled and are unable to progress into autophagosomes. This suggests that the formation of clusters with more than 30 ULK1 molecules in starved cells does not require ULK1 activity. Instead, the expansion of these small, stalled initiation clusters into autophagosomes does require ULK1 activity. Elucidating the mechanisms behind the formation of these intermediate structures may provide valuable insights into the process of autophagy initiation.

Funding Statement

This work was supported by the NIH (R21GM127965 to E.M.P. and R35GM130353 to D.H.K).

Disclosure statement

No potential conflict of interest was reported by the author(s).