AMPK alters proteasome phosphorylation status and prevents persistent proteasome condensates

AMPK; Proteasome condensate; Phosphorylation; Glucose starvation; 29 Nucleocytoplasmic translocation; Budding yeast 30 31 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 2

32 Proteasomes are large multiprotein complexes required for selective intracellular protein 33 degradation, regulating numerous cellular processes and maintaining protein homeostasis and 34 organismal health. In the budding yeast Saccharomyces cerevisiae grown under different 35 glucose conditions, proteasomes undergo dynamic phase transitions between free and 36 condensate states concomitant with nucleocytoplasmic translocation. Low glucose-induced 37 cytoplasmic proteasome condensates are usually reversible but become persistent in the 38 absence of AMP-activated protein kinase (AMPK). AMPK is important for proteasome 39 condensate dissolution and proteasome nuclear reimport upon glucose refeeding of quiescent 40 cells. Here we found that AMPK activities and the AMPK signaling pathway affect proteasome 41 subunit phosphorylation, which correlates with the solubility and reversibility of proteasome 42 condensates. Nuclear and cytoplasmic AMPK functions redundantly in proteasome condensate 43 dissolution. AMPK interacts with the proteasome regulatory particle in an AMPK activity-44 independent manner. At least 50 kinases and phosphatases have been found to associate with 45 the AMPK complex. Therefore, the prevention of persistent proteasome condensate formation 46 by AMPK likely results from regulating the antagonistic effects of downstream kinases and 47 phosphatases on proteasome phosphorylation. A mechanistic understanding of the downstream 48 effector proteins of AMPK that directly regulate proteasome subunit phosphorylation will provide 49 insights into how proteasome phosphorylation is linked to proteasome condensate regulation. 50 51 Article summary 52 Proteasomes undergo dynamic nucleocytoplasmic translocation and phase transitions in 53 response to glucose starvation. AMP-activated protein kinase (AMPK) is important for 54 cytoplasmic proteasome condensate dissolution and proteasome nuclear reentry in budding 55 yeast cells upon glucose refeeding of quiescent cells. This study demonstrates that AMPK 56 interacts with proteasomes, and the AMPK pathway regulates proteasome phosphorylation 57 status and condensate solubility during reversible proteasome condensate formation. AMPK 58 and the PP1 phosphatase dynamically regulate phosphorylation of multiple proteasome 59 subunits. Therefore, the regulation of proteasome phosphorylation by AMPK is likely to be 60 central to proteasome biomolecular condensate formation and dissolution. 61 62 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 3

63 The ubiquitin-proteasome system (UPS) is essential for selective protein degradation and 64 regulates almost every aspect of cellular function, such as cell cycle progression and antigen 65 processing (ROCK et al. 1994; TU et al. 2012). Dysfunction of the UPS is involved in the 66 pathogenesis of many diseases, including many cancers, neurodegeneration, infection, 67 inflammation, and developmental disorders (COUX et al. 2020). At the center of the UPS is the 68 26S proteasome, a large multiprotein complex comprising a proteolytic core particle (CP) and a 69 regulatory particle (RP). The RP is assembled from lid and base subcomplexes (BARD et al. 70 2018). 71 Proteasomes undergo dynamic phase transitions and intracellular movement under 72 various stress conditions (ENENKEL AND ERNST 2025). Stress can induce proteasomes to 73 transition from a free state to a condensate-like state and can regulate proteasome 74 nucleocytoplasmic translocation. The cellular regulation of proteasome phase transitions and 75 intracellular movements is strongly conserved from yeast to plants and mammalian cells. 76 Depending on the specific stress conditions, reversible proteasome condensates have been 77 observed in the nucleus of mammalian cells and the cytoplasm or perinuclear region of yeast 78 cells (ENENKEL et al. 2022). In exponentially growing budding yeast Saccharomyces cerevisiae, 79 proteasomes are highly enriched in the nucleus (PACK et al. 2014). When yeast enter stationary 80 phase, a glucose starvation condition, most proteasomes are exported from the nucleus to form 81 reversible proteasome condensates or proteasome storage granules (PSGs) in the cytoplasm 82 (LAPORTE et al. 2008). 83 Upon glucose refeeding of such quiescent cells, proteasome condensates dissolve 84 rapidly, and proteasomes are reimported into the nucleus within minutes (LAPORTE et al. 2008; 85 BUTCHER et al. 2025). Cellular factors or conditions that modulate these movements and 86 physical changes include ubiquitin (GU et al. 2017); cytosolic pH (PETERS et al. 2013); 87 endosomal sorting complex required for transport (ESCRT) complexes (LI et al. 2019; LI AND 88 HOCHSTRASSER 2022); membrane fusion protein Pep12 (VANDERVEN et al. 2025); proteasome 89 shuttle factors Rad23, Dsk2, and Ddi1; and ubiquitin polymers (WAITE et al. 2024). By contrast, 90 cellular factors required for proteasome condensate dissolution are poorly understood. Our 91 previous studies demonstrated that AMP-activated protein kinase (AMPK) is required for 92 proteasome condensate dissolution (LI et al. 2019). However, how AMPK regulates proteasome 93 condensate dissolution is unclear. 94 AMPK is an evolutionarily conserved master regulator of cellular energy homeostasis in 95 all examined eukaryotes (GHILLEBERT 2011). AMPK activation mediates a metabolic switch from 96 an anabolic to catabolic state under stress conditions. Dysregulation of AMPK has been 97 implicated in the pathogenesis of many human diseases, such as metabolic disorders, 98 neurodegenerative diseases, and cancer (ASHRAF AND VAN NOSTRAND 2024). AMPK is a 99 heterotrimeric complex, comprising a catalytic α subunit (Snf1 in yeast), a regulatory β subunit, 100 (the three yeast paralogs–Gal83, Sip1, and Sip2–determine the subcellular location of the 101 AMPK complexes), and a regulatory γ subunit (Snf4) (COCCETTI et al. 2018). In yeast cells, Snf1 102 is activated by upstream kinases Sak1, Elm1, and Tos3 through the phosphorylation of 103 threonine-210 (T210), equivalent to T172 in mammalian cells (JEON 2016), in the Snf1 activation 104 loop (SUTHERLAND et al. 2003). The Reg1-Glc7 protein phosphatase 1 (PP1) reverses AMPK 105 activation through T210 dephosphorylation (MCCARTNEY AND SCHMIDT 2001). Snf1 can also be 106 negatively regulated by autoinhibition, which occurs through interaction of the Snf1 kinase with 107 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 4 its regulatory domains (JIANG AND CARLSON 1996). Many drugs and compounds have been 108 developed and identified that alter AMPK activity (STEINBERG AND CARLING 2019). 109 Understanding the regulation of proteasomes by AMPK may offer a means to repurpose these 110 FDA-approved AMPK-related drugs for proteasome regulation and proteasome dysfunction-111 related disease treatment. 112 Proteasome condensates are one of the hundreds of biomolecular condensates that can 113 form under various specific conditions (NARAYANASWAMY et al. 2009). Biomolecular 114 condensates play crucial roles in fine-tuning cellular processes, such as gene expression and 115 intracellular signal transduction (JEON et al. 2025). At the cellular level, it is evident that AMPK 116 plays an important role in regulating biomolecular condensate formation and dissolution. For 117 instance, under arsenite stress, the Glc7 catalytic subunit of yeast PP1 relocates from the 118 nucleus and assembles into cytoplasmic condensates, resulting in translational inhibition. The 119 regulatory subunit Reg1 of PP1 and its cellular target Snf1 are important for regulating 120 reversible Glc7 cytoplasmic condensate formation (SCHNELL et al. 2021). Notably, the RNA-121 binding protein Smaug1 forms cytosolic condensates, thereby inhibiting mRNA translation; 122 activation of AMPK induces dissolution of the Smaug1 condensates, releasing mRNA for 123 translation, whereas inhibition of AMPK locks Smaug1 in the condensed state (THOMAS et al. 124 2025). 125 Our previous studies demonstrated that persistent proteasome condensates are formed 126 in the absence of AMPK when budding yeast are grown under low glucose conditions (LI et al. 127 2019). On the other hand, the UPS can also regulate AMPK (ZUNGU et al. 2011). Here, we show 128 that the AMPK signaling pathway regulates proteasome condensate (PSG) dissolution. AMPK 129 complexes interact with proteasomes and maintain proteasome condensate solubility under low 130 glucose conditions. Snf1 kinase activity is necessary for regulating proteasome condensate 131 dissolution, which correlates with altered proteasome phosphorylation status. Our findings thus 132 illuminate a cellular pathway of dynamic proteasome phase transitions regulated by AMPK. 133 134

135 Proteasome condensates gradually lose reversibility without AMPK 136 Proteasome condensates (PSGs) are reversible (LAPORTE et al. 2008). They dissolve 137 rapidly in a repetitive “contact and release” manner at the nuclear periphery, where 138 proteasomes reimport to the nucleus within 15 minutes upon glucose recovery (BUTCHER et al. 139 2025). AMPK is required for proteasome condensate dissolution upon glucose refeeding after 140 cells are starved for four days in low-glucose medium (LI et al. 2019). To understand the role of 141 AMPK in regulating proteasome condensate dissolution, we examined and quantified the 142 percentage of cells with proteasome condensates after one and four days in low glucose and 143 after glucose refeeding following these starvation treatments. As expected, proteasome 144 condensates were reversible in wild-type (WT) cells under low glucose conditions on both day 145 one and day four (Figure 1). By contrast, in the AMPK deletion mutants snf1∆ and snf4∆, 146 proteasome condensates rapidly dissolved upon glucose addition on day one but became 147 persistent by day four of starvation (Figure 1). Parallel results were observed using GFP-tagged 148 CP (Pre10), RP base (Rpn2), and RP lid (Rpn5) subunits. These data indicate that AMPK is 149 important for maintaining the reversibility of PSGs following prolonged glucose limitation. 150 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 5 Persistent proteasome condensates are associated with abnormal proteasome 151 phosphorylation status 152 To investigate how AMPK might control proteasome condensate reversibility, we used 153 the Phos-tag gel mobility shift method to examine proteasome subunit phosphorylation status. 154 Phos-tag molecules preferentially capture phosphomonoester dianions bound to serine, 155 threonine, and tyrosine residues (KINOSHITA et al. 2006). Phosphorylation of proteasome 156 subunits could be inferred by their slowed migration in Mn2+-Phos-tag SDS-PAGE gels. By anti-157 GFP immunoblotting of GFP-tagged subunits, we observed dynamic proteasome 158 phosphorylation changes in cells under different glucose conditions; specifically, we switched 159 the glucose concentration from 2% to 0.025% for one or four days and then switched the four-160 day-starved cells back to 2% glucose for 15 minutes to examine recovery (Figure 2). 161 When isolated from exponentially growing (‘log’) WT cells in 2% glucose, all three tested 162 GFP-tagged proteasome subunits migrated more slowly in Phos-tag gels than when grown 163 under glucose-starved conditions. The CP subunit Pre10 is smaller than the lid subunit Rpn5 164 and the base subunit Rpn2, so phosphorylated Pre10 isoforms were better separated in the 165 Phos-tag gels (Figure 2a). WT proteasome subunit phosphorylation decreased dramatically on 166 day one of glucose limitation, based on the more rapid migration of the subunits in Phos-tag 167 gels (Figure 2). While their degree of phosphorylation increased slightly on day four, it did not 168 return to the levels seen in non-starved cells, and this also did not change detectably after 15 169 minutes of glucose refeeding. Therefore, proteasome phosphorylation undergoes striking 170 changes in response to glucose availability. This correlates with proteasome nucleocytoplasmic 171 translocation and proteasome transitions between free particles and proteasome condensates. 172 By contrast, proteasome subunits were underphosphorylated (faster migrating) in snf1∆ 173 and snf4∆ cells growing exponentially in 2% glucose (Figure 2). Subunit phosphorylation 174 decreased slightly by day one but increased by day four under low glucose to levels equivalent 175 to WT levels or nearly so. These levels decreased slightly again during glucose recovery. 176 Together, the results demonstrate that AMPK regulates proteasome phosphorylation status 177 under changing glucose conditions. Moreover, the persistence of proteasome condensates 178 observed in AMPK mutants correlates with abnormal proteasome subunit phosphorylation 179 status. 180 Snf1 ki nase activity regulates proteasome condensate reversibility 181 Snf1 is activated by upstream kinases through phosphorylation of the conserved T210 182 residue in the Snf1 activation loop (SUTHERLAND et al. 2003). Lysine-84 (K84) and glycine-53 183 (G53) in Snf1 are important parts of the ATP-binding site and are essential for kinase activity 184 (ESTRUCH et al. 1992). Mutations of T210 to alanine (snf1-T210A) or aspartate (snf1-T210D), or 185 mutation of K84 to arginine (snf1-K84R) disrupt Snf1 kinase activity while G53 to arginine (snf1-186 G53R) mutation enhances Snf1 kinase activity (ESTRUCH et al. 1992). To test whether Snf1 187 activation through T210 and Snf1 kinase activity are essential for proteasome condensate 188 dissolution and affect proteasome subunit phosphorylation status, we expressed WT SNF1 or 189 the aforementioned snf1 mutants from the native promoter on a low-copy CEN plasmid in snf1∆ 190 cells (Figure 3). 191 Like snf1∆ cells with an empty vector (EV), abnormally persistent proteasome 192 condensates were observed following glucose refeeding of cells that had been in low glucose 193 for four days and expressed the inactive snf1-T210A, -T210D, or -K84R mutant proteins (Figure 194 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 6 3a and 3b). Correlating with the persistence of proteasome condensates, CP and RP 195 proteasomal subunits remained more heavily phosphorylated in these mutants, comparable to 196 snf1 null cells (Figure 3c and 3d, Phos-tag SDS-PAGE gel blots). Microautophagy of 197 proteasomes was also similarly impaired based on the reduced fragmentation or loss of these 198 subunits compared to snf1∆ cells complemented with the WT SNF1 plasmid. In contrast to the 199 kinase-dead mutants, the percentage of snf1-G53R mutant cells (which have elevated kinase 200 activity) bearing proteasome condensates was decreased relative to WT under low glucose 201 conditions. Importantly, proteasome condensates were fully reversible in the snf1-G53R mutant 202 upon glucose refeeding (Figure 3b), and proteasome subunit phosphorylation and autophagic 203 degradation were unaffected (Figure 3d). These results indicate that Snf1 kinase activity 204 regulates proteasome condensate reversibility, proteasome autophagy and subunit 205 phosphorylation. 206 Proteasome condensate formation is a highly regulated cellular process that occurs 207 alongside proteasome degradation by autophagy under low glucose conditions (LI et al. 2019). 208 This helps ensure that normal proteasomes are sorted into reversible proteasome condensates 209 while aberrant proteasomes are removed through autophagy (LI AND HOCHSTRASSER 2020). 210 Snf1 has a dual role in regulating proteasome condensate dissolution and autophagic 211 degradation (LI et al. 2019). Phosphorylation of T210 activates Snf1, whereas its 212 dephosphorylation by PP1 inactivates the kinase. Reg1 is a regulatory subunit for the Glc7 213 phosphatase in the PP1 complex, and cells lacking Reg1 can no longer dephosphorylate T210 214 and exhibit constitutive phosphorylation of T210 and activation of Snf1 kinase activity 215 (MCCARTNEY AND SCHMIDT 2001). Since elevated Snf1 kinase activity in the G53R mutant 216 appeared to reduce the levels of proteasome condensates in low glucose and enhance their 217 reversibility upon glucose refeeding (Figure 3b), we tested whether constitutively active Snf1 218 could also alter reversible proteasome condensates in reg1∆ cells. As predicted, a lower 219 percentage of reg1∆ cells contained proteasome condensates compared to WT cells, and the 220 condensates were smaller than in WT cells (Figure S1). To confirm that regulation of 221 proteasome condensate formation by Reg1 is through the AMPK signaling pathway, we tested 222 proteasome condensate reversibility in snf1∆ reg1∆ and snf4∆ reg1∆ cells. As expected, 223 persistent proteasome condensates were formed in snf1∆ reg1∆ and snf4∆ reg1∆ cells (Figure 224 S2). This demonstrates that snf1∆ and snf4∆ are epistatic to reg1∆ in regulating proteasome 225 condensate dissolution. 226 These results indicate that constitutive Snf1 kinase activity caused by PP1 phosphatase 227 inactivation promotes dissolution of proteasome condensates and that Snf1 activity levels can 228 fine-tune proteasome condensate formation and dissolution. 229 230 The r egulatory β subunits Gal83 and Sip2 function redundantly in regulating proteasome 231 condensate dissolution 232 The regulatory β subunits Sip1, Sip2, and Gal83 determine the subcellular activity of 233 Snf1 in the vacuole, cytoplasm, and nucleus, respectively (JIANG AND CARLSON 1997; VINCENT 234 et al. 2001). To understand which β subunit(s) is required for regulating proteasome condensate 235 dissolution and which subcellular population(s) of AMPK regulates proteasome condensate 236 dissolution, we examined proteasome condensates in single and double deletion mutants of the 237 three β subunits. Proteasome condensates were reversible in all three single and two of the 238 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 7 three double deletion mutants (Figure S3); the lone exception was the sip2∆ gal83∆ double 239 mutant (Figure 4). Sip2 and Gal83 mediate cytoplasmic and nuclear Snf1 activity, respectively, 240 suggesting that nuclear and cytoplasmic Snf1 activity are each sufficient to mediate proteasome 241 condensate dissolution and nuclear reimport. The Sip1-Snf1-Snf4 complex is the only remaining 242 Snf1 kinase complex in sip2∆ gal83∆ cells. However, previous work revealed low in vitro kinase 243 activity due to the reduced levels of Sip1 in sip2∆ gal83∆ cells (NATH et al. 2002). Therefore, we 244 cannot fully exclude the possibility that the persistent proteasome condensates in sip2∆ gal83∆ 245 cells are due to overall reduced Snf1 kinase activity rather than loss of Snf1 activity at a 246 particular subcellular site. 247 AMPK interacts with proteasome RP in a Snf1 kinase activity-independent manner 248 To examine whether AMPK interacts with proteasomes, we performed GFP-Trap 249 coimmunoprecipitation in snf1∆ or snf4∆ cells expressing the GFP-tagged proteasome CP 250 subunit Pre10 or RP subunit Rpn5 and plasmid-based expression of HA-tagged Snf1 or Snf4. 251 Although SNF1 and SNF4 were both expressed from the GPD (TDH3) promoter, the levels of 252 Snf4 protein appeared to be much higher than that of Snf1, possibly due to downregulation of 253 Snf1 under low glucose conditions (Figure 5a, total cell lysates). Snf4 interacted with both the 254 CP (Pre10-GFP) and the RP (Rpn5-GFP), whereas Snf1 only interacted detectably with the RP 255 (Figure 5a, long exposure blot). WT Snf1, kinase activity-defective mutants T210A, T210D, and 256 K84R, and elevated kinase-activity mutant G53R interacted with Rpn5 in a Snf1 kinase activity-257 independent manner (Figure 5b). Snf1 protein accumulation and that of the snf1 mutants varied 258 and were negatively correlated with the kinase activity; specifically, Snf1 with the kinase-259 inhibiting T210A, T210D, or K84R mutations accumulated to higher levels than WT Snf1, while 260 the hyperactive snf1-G53R kinase was at lower levels compared to WT Snf1 in total cell lysates 261 (Figure 5b). No AMPK components have been detected in a previous tandem mass 262 spectrometry analysis of WT strain without GFP tag (LI AND HOCHSTRASSER 2022). These 263

suggest that Snf1 kinase activity impacts Snf1 protein levels through a post-264 transcriptional mechanism but does not modulate Snf1-proteasome (RP) interactions. 265 We next purified Snf1 complexes from yeast cells expressing a 3xFlag-tagged Snf1 from 266 the native chromosomal locus. By anti-Flag affinity purification followed by tandem mass 267 spectrometry analysis, we could determine the Snf1 complex interactome under different growth 268 conditions. Consistent with the co-immunoprecipitation results in Figure 5, Snf1 complexes 269 interacted with the proteasomal CP, RP, and proteasome assembly chaperones in exponentially 270 growing (log) cells, but only the RP complex copurified under low glucose conditions (Figure S4; 271 Table S5). The lack of CP in the anti-Flag isolates after glucose limitation may reflect separation 272 of CP and RP subcomplexes under these conditions (LI AND HOCHSTRASSER 2022) and a direct 273 interaction of Snf1 only with the RP. Additionally, 37 kinases and 13 phosphatases were 274 detected in the Snf1 interactome (Table S1). The results confirm the interactions between the 275 Snf1 complex and the proteasomal RP and suggest the downstream effectors of AMPK 276 signaling pathway regulate proteasome condensate reversibility. 277 Persistent proteasome condensates in AMPK mutants behave as solid-like condensates 278 In the absence of AMPK in long-term glucose starved cells, proteasome condensates 279 became resistant to dissolution following glucose refeeding and proteasomes were abnormally 280 phosphorylated (Figure 2). More generally, protein phosphorylation plays important roles in 281 regulating protein-protein interactions and can alter biomolecular condensate solubility and 282 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 8 reversibility (SRIDHARAN et al. 2022; RANGANATHAN et al. 2023). Liquid-like, but not solid-like, 283 biomolecular condensates can be dissolved by the solvent 1,6-hexanediol (1,6-HD), which 284 interferes with their weak hydrophobic interactions. Therefore,1,6-HD is a useful tool to 285 differentiate between liquid-like and solid-like condensates in cells (KROSCHWALD et al. 2017). 286 Using 1,6-HD treatments, we tested the physical properties of proteasome condensates 287 (PSGs) in WT and AMPK mutant cells. Proteasome condensates in WT cells were readily 288 dissolved by 1,6-HD after one day of glucose limitation (Figures 6a and 6b) but became less 289 soluble after longer-term glucose starvation stress (Figures 6a and 6c). By contrast, proteasome 290 condensates in AMPK deletion mutants were much more resistant to dissolving in 1,6-HD at 291 both early and later times under glucose limitation, suggesting they had solid-like condensate 292 properties (Figure 6). These data suggest that proteasome condensates lose solubility and 293 behave more like solids or gels in cells lacking AMPK. Hence, an increase in solid-like 294 proteasome condensates under low glucose stress correlates closely with loss of condensate 295 reversibility and abnormal proteasome phosphorylation status in cells without AMPK (Figures 2 296 and 6). 297 In an attempt to identify the phosphorylation sites of proteasome subunits that lead to 298 persistent proteasome condensates in cells, we performed gel phospho-staining and phospho-299 enrichment mass spectrometry of affinity-purified proteasomes from WT, snf1∆, and snf4∆ cells. 300 The Pro-Q® Diamond phosphoprotein gel stain allows direct, in-gel detection of phosphate 301 groups attached to tyrosine, serine, or threonine residues. The phosphorylation status of several 302 proteasome subunits, such as Rpt1, Rpn6, Rpn11, and Rpn12, was strongly impacted, with 303 most showing decreased phosphorylation in snf1∆ and snf4∆ cells relative to WT cells, 304 especially on day four under low glucose conditions (Figure 7). By contrast, proteasome subunit 305 phosphorylation status was much less affected in reg1∆ cells, which have constitutively high 306 AMPK activity (Figure S5). Therefore, loss of AMPK alters the phosphorylation status of multiple 307 specific proteasome subunits in response to glucose starvation. 308 Phospho-enrichment mass spectrometry of purified proteasomes also revealed changes 309 in phosphorylation status of individual proteasome subunits in AMPK null cells relative to WT 310 when cells were starved for glucose (Figure S6). In most cases, the changes were in overall 311 agreement with in-gel phospho-staining, such as for Rpn6 and Rpn11, but some discrepancies 312 were seen (e.g., Pre10 phosphorylation was apparently lower in snf1∆ cells based on gel 313 staining but was slightly increased based on the proteomic analysis). We investigated the 314 impact on proteasome condensate dissolution of specific proteasome phosphorylation sites of 315 several subunits (Rpn8, Rpn11, Rpn13, and Pre10) by mutating the relevant residues to alanine 316 but failed to identify specific phosphorylation sites that were necessary. Given that proteasomes 317 have 31 subunits that have been reported to be phosphorylated and hundreds of potential 318 phosphorylation sites that may function redundantly (HIRANO et al. 2015), it is perhaps not 319 surprising that we could not identify specific sites that recapitulate the phenotype of cells lacking 320 AMPK. 321 322

323 In this study, we have demonstrated that Snf1 kinase activity is important for normal 324 proteasome subunit phosphorylation patterns under low glucose growth conditions and for the 325 solubility and reversibility of proteasome condensates (Figures 3, 6, and 7). The nuclear and 326 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 9 cytoplasmic forms of Snf1 complexes together appear to be necessary for proteasome 327 condensate dissolution, and either by itself is sufficient; the vacuole-localized Snf1 complex is 328 neither necessary nor sufficient for condensate regulation (Figures 1 and 4). We observed a 329 physical interaction between Snf1 and the proteasomal RP by co-immunoprecipitation and mass 330 spectrometry; this interaction appears to be independent of Snf1 kinase activity (Figures 5 and 331 S4). 332 To date, we have not been able to isolate Snf1 mutants that disrupt Snf1-RP interaction 333 under low glucose conditions but retain AMPK activity. Therefore, it remains unclear how 334 AMPK-proteasome interactions regulate proteasome trafficking and condensate properties. 335 Snf1 complexes are unlikely to play a structural role in PSGs since glucose starvation-induced 336 condensates are mainly composed of proteasomes and monoubiquitin (GU et al. 2017). 337 Moreover, we did not detect direct binding between purified Snf1 complexes and proteasomes 338 in vitro (not shown), indicating that additional factors may be involved in the physical interactions 339 between these complexes in cells. 340 The physical interactions between Snf1 complexes and the proteasomal RP might 341 suggest that AMPK can directly affect proteasome subunit phosphorylation. In our 342 phosphorylation enrichment mass spectrometry analysis of purified proteasomes from WT and 343 snf1∆ cells, we observed most proteasome subunits with reduced phosphorylation in snf1∆ cells 344 on day four under low glucose (Figure S6). This is generally consistent with our phospho-345 staining results of purified proteasomes (Figure 7). Day four was also when persistent 346 proteasome condensates were observed in cells lacking AMPK activity (Figures 1 and 3). 347 Therefore, it is possible that specific AMPK-dependent proteasome subunit phosphorylations 348 help maintain cytoplasmic proteasome condensates in a state capable of rapidly dissipating and 349 allowing reimport of proteasomes back into the nucleus. 350 On the other hand, there were no clear differences in the phosphorylation patterns of 351 purified proteasomes between WT and snf1∆ cells during exponential growth in high glucose or 352 after one day in low-glucose conditions (Figure 7). These data using isolated proteasomes 353 suggest it is less likely that AMPK acts directly on proteasomes since we should have detected 354 more rapid responses to AMPK, and it should not take days to see changes in proteasome 355 phosphorylation status. However, we did observe clear differences in proteasome 356 phosphorylation profiles under changing glucose conditions when whole cell lysates were 357 examined using Phos-tag gels and immunoblotting, and these changes correlated closely with 358 loss of solubility of PSGs in 1,6-HD (Figures 2 and 6). As noted, the difference in samples of 359 purified proteasomes versus total cell lysates and methodologies of phospho-stain versus Phos-360 tag gels may lead to different results (Figures 2 and 7). 361 Although we do not know the reason for the discrepancy be tween analyses of purified 362 proteasomes analyzed by phospho-staining versus analysis of total cell lysates using Phos-tag 363 gels, both methodologies suggest that AMPK functions indirectly on proteasomes by regulating 364 other kinases and phosphatases that more directly impact the overall proteasome 365 phosphorylation profile under the indicated growth conditions. This is consistent with the large 366 number of phosphorylation sites found in proteasomes (HIRANO et al. 2015), and our detection 367 of 37 kinases and 13 phosphatases in the Snf1 complex interactome that could function as part 368 of the AMPK signaling pathway acting on proteasomes (Table S1). For example, Cdc14 369 phosphatase and Torc1 kinase have extensive network interactions with other kinases and 370 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 10 phosphatases (BREITKREUTZ et al. 2010), and similar antagonistic effects of kinases and 371 phosphatases on the overall proteasome phosphorylation status are also possible. 372 Phosphorylation plays a critical role in regulating formation and dissolution of 373 biomolecular condensates, which are involved in diverse cellular and biological processes 374 (SRIDHARAN et al. 2022). Phosphorylation alters protein conformations and interactions, which 375 subsequently modulate condensate properties and biological functions (NOSELLA AND FORMAN-376 KAY 2021; RANGANATHAN et al. 2023). For instance, the N-terminal intrinsically disordered 377 region of adenovirus 52K protein is phosphorylated at specific Ser residues, promoting liquid-378 like properties of viral 52K protein condensates and viral particle production (GRAMS et al. 379 2024). Previous studies have demonstrated that phosphorylation can alter proteasome function 380 and activities (WANI et al. 2016; GUO et al. 2017; KORS et al. 2019; LIU et al. 2020). Our data 381 demonstrate that phosphorylation is tightly linked to proteasome condensate solubility and 382 reversibility (Figures 1, 2, 6, and 7). This is expected to alter proteasome functions, for example, 383 by altering nuclear re-import dynamics of proteasomes upon glucose refeeding. 384 To identify the phosphorylation sites in proteasome subunits that are important for 385 proteasome condensate dissolution, we performed TiO2 phospho-enrichment mass 386 spectrometry analysis of purified proteasomes from WT and snf1∆ cells (Figure S6). By 387 comparing the number of phosphorylation sites and spectrum counts of WT and snf1∆ (Figure 388 S5), we performed mutagenesis and deletion of the potential phosphorylation sites and regions 389 in four proteasome subunits, i.e., Rpn8, Pre10, Rpn13, and Rpn11 (Table S2). None of the 390 mutated phosphorylation sites caused defects in proteasome condensate dissolution. Therefore, 391 an optimized strategy will need to tease out the direct effector proteins that regulate proteasome 392 condensate dissolution and phosphorylation status in the AMPK pathway. 393 The AMPK pathway plays a major role in regulating cellular energy homeostasis. 394 Dysfunction of AMPK has been observed in various diseases, such as cancer, metabolic 395 disorders, and aging (JEON 2016). Therefore, AMPK has attracted widespread interest as a 396 therapeutic target for disease treatment. Our results indicate that AMPK plays an important role 397 in proteasome condensate dissolution by interacting with the proteasome RP and altering 398 proteasome phosphorylation status and condensate solubility, which is likely relevant to 399 proteasome condensate regulation by AMPK in humans. Moreover, repurposing AMPK-400 modulating drugs for proteasome regulation could offer new strategies to tackle drug resistance 401 in proteasome malfunction-related disease treatments. 402 403

404 Yeast strains and cell cultures 405 Yeast strains used in this study are listed in Table S3. Yeast cells were grown overnight in 406 synthetic complete (SC) medium (LI AND HOCHSTRASSER 2022) at 30°C with vigorous agitation; 407 tryptophan and uracil single or double dropout media were used for plasmid selection. Cultures 408 were back-diluted in fresh SC medium and grown to mid-exponential phase. Cells were 409 harvested and rinsed once with sterile ultrapure water and then resuspended in SC medium 410 containing 0.025% glucose. The cells grown in low glucose cells were incubated for one and 411 four days at 30°C with vigorous agitation. For glucose recovery (BUTCHER et al. 2025), the low-412 glucose starved cells were pelleted and rinsed once with sterile ultrapure water and then 413 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 11 resuspended in fresh SC medium containing 2% glucose. The cells were incubated in this 414 medium for 15 min and subjected to different experiments as described in the figures. For 415 fluorescence microscopy, glucose-recovered cells were fixed with 2% formaldehyde at room 416 temperature (RT) for 5 min. The fixed cells were pelleted, washed once with 0.1M KPO4 pH 6.5, 417 and resuspended in 0.1M KPO4 pH 7.5 before imaging. 418 Plasmid construction 419 Plasmids used in this study are listed in Table S4. Genes of interest were amplified from yeast 420 genomic DNA and using restriction enzyme digestions and T4 DNA ligase (NEB)-mediated 421 ligation were cloned into the following plasmids: pRS316 (SIKORSKI AND HIETER 1989), 422 p424GPD, and p426GPD (MUMBERG et al. 1995). The plasmids were verified by enzymatic 423 digestion and DNA sequencing. QuikChange site-directed mutagenesis (Agilent) was performed 424 to generate snf1-T210A, snf1-T210D, snf1-K84R, and snf1-G53R mutants. 425 Fluorescence microscopy 426 Yeast cells were visualized on an Axioskop microscope (Carl Zeiss) equipped with a Plan-427 Apochromat 100×/1.40 NA oil DIC objective lens, a CCD camera (AxioCam MRm; Carl Zeiss), 428 and an HBO100W/2 light source. Images were taken using AxioVision software with an auto-429 exposure setup. Yeast cells used for Figure S2 were visualized on a Nikon Ts2R inverted 430 fluorescence microscope equipped with a CFI60 Plan Apochromat Lambda D 100x/1.45 NA oil 431 DIC objective lens, a Nikon DS-Fi3 camera, and an LED light source. Images were taken using 432 NIS-Elements D software with a 400 ms exposure setup. Images were taken under a single 433 focal plane and processed using Adobe Photoshop CC software. Percentage of cells with 434 proteasome condensates were quantified using ImageJ. ANOVA single factor significance 435 analysis was performed using Excel. 436 Protein extraction and Western blotting 437 Total yeast proteins were extracted using the post-alkaline extraction method (KUSHNIROV 438 2000), and Western blotting was performed as described previously with minor changes (LI et 439 al. 2016). One optical density unit at 600 nm (OD600) of cells were harvested at 8,000 rpm for 1 440 min and rinsed once with sterile ultrapure water at room temperature (RT). For phosphorylation 441 analyses, cells were incubated in 400 µl 0.1 M NaOH with PhosSTOP phosphatase inhibitor 442 cocktail tablets (Millipore Sigma, catalog # 4906837001) for 5 min at RT and pelleted at 10,000 443 rpm for 2 min. The alkali-treated cells were resuspended in 100 µl SDS sample buffer (10% 444 glycerol, 2% SDS, 0.1 M DTT, 62.5 mM Tris-HCl pH 6.8, 4% 2-mercaptoethanol, 0.008% 445 Bromophenol Blue, PhosSTOP phosphatase inhibitor) and heated at 100°C for 5 min. Cell 446 debris was removed by centrifugation at 15,000 rpm for 1 min. Equal volumes of the 447 supernatants were loaded onto 10% (v/v) SDS-PAGE gels with and without adding Mn2+-Phos-448 tag to the resolving gels. Additionally, the Phos-tag SDS PAGE gels were incubated with 449 transfer buffer (20% methanol, 0.3% trizma base, 1.44% glycine, 0.01% SDS) containing 10 mM 450 EDTA and then incubated in fresh transfer buffer before the transfer of proteins to PVDF 451 membranes (EMD Millipore, catalog # IPVH00010). 452 The membranes were incubated with JL-8 anti-GFP monoclonal antibody (Clontech 453 catalog # 632381) at 1:2,000 dilution or an anti-Pgk1 monoclonal antibody (Thermo Fisher 454 Scientific, catalog # 459250) at a 1:10,000 dilution. Primary antibody binding was followed by 455 anti-mouse-IgG (Cytiva, catalog # NXA931) secondary antibody conjugated to horseradish 456 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 12 peroxidase at a 1:10,000 dilution. The membranes were incubated in ECL detection reagent 457 (MRUK AND CHENG 2011). The ECL signals were detected using film (Thomas Scientific, catalog 458 #1141J52). 459 GFP-Trap immunoprecipitation of GFP-tagged proteasomes 460 GFP-Trap immunopurifications were performed as described previously (LI AND HOCHSTRASSER 461 2022). Briefly, 35 OD600 units of cells expressing GFP-tagged proteasome subunits were 462 harvested and rinsed once with ice-cold sterile ultrapure water. Cells were resuspended in lysis 463 buffer (20 mM Tris pH 7.5, 0.5 mM EDTA pH 8, 200 mM NaCl, 10% glycerol, 1 mM PMSF, 10 464 mM NEM, Roche cOmplete mini protease inhibitor catalog # 11836153001). Total cell lysates 465 were prepared by beating with acid-washed glass beads (Sigma-Aldrich, catalog # G8772). The 466 cell membranes were solubilized by adding 0.1% Triton X-100 to the lysates and incubated on 467 ice for 30 min. Crude cell lysates were cleared at 16,000 x g for 10 min at 4°C. Cleared cell 468 lysates were incubated with GFP-Trap agarose resin slurry for 1 h. The resin was washed three 469 times with the lysis buffer containing 0.02% Triton X-100, resuspended in 2xSDS sample buffer, 470 and then incubated at 42°C for 10 min to elute the bound proteins. The elutes were analyzed by 471 Western blotting as above with monoclonal anti-HA antibody (Sigma, catalog # H9658) at a 472 1:2,000 dilution as well as with JL-8 anti-GFP and anti-Pgk1 monoclonal antibodies. 473 In vivo proteasome condensate solubility test 474 Yeast cell treatment with 1,6-hexanediol (1,6-HD) was performed as described previously 475 (KROSCHWALD et al. 2017). Low glucose-grown yeast cells were harvested and treated with a 476 final concentration of 6% 1,6-hexanediol (1,6-HD) and 10 µg/ml digitonin in the media on day 477 one, and 10% 1,6-HD and 10 µg/ml digitonin on day four. The cells were incubated at RT for 20 478 min and fixed with 2% formaldehyde before imaging by fluorescence microscopy as described 479 above. 480 Proteasome affinity purification and gel staining 481 Proteasome affinity purification was performed using yeast cells expressing Rpn11-3xFlag as 482 described previously (SULTANA et al. 2023). Briefly, chromosomally tagged RPN11-3xFlag cells 483 were harvested following incubation at the indicated growth conditions and ground to fine 484 powder with liquid nitrogen. Cell powders were resuspended in buffer A (50 mM Tris-HCl pH 485 7.5, 150 mM NaCl, 10% glycerol, 5 mM MgCl2, 5 mM ATP, Roche cOmplete EDTA-free 486 protease inhibitor, catalog #1872580001, PhosSTOP phosphatase inhibitor), and incubated on 487 ice for 15 min. Cell debris were pelleted at 30,000 x g for 20 min at 4°C. Total protein 488 concentration was determined with a Pierce BCA protein assay kit (Thermo Scientific, catalog 489 #23227). The supernatants were incubated with anti-FLAG M2 affinity gel (Sigma, catalog 490 #A2220) for 2 h on a rotator at 4°C. The resin was washed twice with buffer A for 10 min and 491 then incubated with 3 resin volumes of 200 µg/ml 3xFLAG peptide (Sigma, catalog #F4799) for 492 45 min to elute proteasome complexes. Proteasomes were concentrated with 100 K MWCO 493 centrifugal filters (Merck Millipore, catalog #UFC510024) and quantified with a BSA standard 494 using a G:Box Chemi HR16 imager (Syngene). 495 Proteasome staining with Pro-Q® Diamond phosphoprotein gel stain (Invitrogen, catalog 496 # P33301) and SYPRO® Ruby protein gel stain (Invitrogen, catalog # S12001) were performed 497 according to the manufacturer’s protocols. Briefly, 10 µg of proteasomes were resolved in 12% 498 SDS-PAGE gels. The gels were fixed twice with a solution of 50% methanol and 10% acetic 499 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 13 acid with gentle agitation at RT for 30 min per fixation. The fixation step was to ensure the 500 removal of SDS from the gels. The fixed gels were washed three times with ultrapure water for 501 10 min per wash to remove methanol and acetic acid from the gels. Then the gels were stained 502 with Pro-Q® Diamond phosphoprotein gel stain with gentle agitation in the dark for 1 h and 503 destained three times with a solution of 20% acetonitrile and 50 mM sodium acetate (pH 4.0) 504 with gentle agitation at RT in the dark for 30 min each. The gels were washed twice with 505 ultrapure water at RT for 5 min per wash before imaging with a Typhoon biomolecular imager 506 (Amersham Biosciences). 507 After phosphoprotein gel stain imaging, the gels were incubated with SYPRO® Ruby 508 protein gel stain solution overnight at RT. The stained gels were washed with a solution of 10% 509 methanol and 7% acetic acid for 30 min and rinsed twice with ultrapure water for 5 min each 510 before imaging with the BioRad ChemiDoc MP Imaging System. 511 Affinity purification of Snf1 complexes 512 Yeast cells expressing Snf1-3xFlag from the chromosomally tagged SNF1 locus were used for 513 Snf1 complex affinity purification. Cells were harvested at the indicated growth conditions and 514 ground to a fine powder with liquid nitrogen. Cell powders were resuspended in modified 515 IPP150 buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Nonidet P-40, 5% Glycerol, and 516 Roche cOmplete EDTA-free protease inhibitor, catalog #1872580001) (NATH et al. 2002). The 517 lysed cells were incubated on ice for 15 min and followed the same protocol for proteasome 518 purification above. Snf1 complexes were concentrated with 10 K MWCO centrifugal filters 519 (Merck Millipore, catalog # UFC501024) and quantified with a BSA standard using a G:Box 520 Chemi HR16 imager (Syngene). The purified Snf1 complexes were resolved in 10% SDS-PAGE 521 gels. The gel slices were fixed before liquid chromatography-tandem mass spectrometry (LC-522 MS/MS) analysis (see below). 523 In-gel digestion 524 Gel bands containing the Snf1 complexes or proteasomes were cut into small pieces and rinsed 525 with 1 ml water on a tilt-table for 10 min. The bands were then washed for 20 min with 1 ml 50% 526 acetonitrile (ACN)/100 mM NH4HCO3 (ammonium bicarbonate, ABC). The samples were 527 reduced by the addition of 4.5 mM dithiothreitol (DTT) (sufficient volume to cover gel pieces) in 528 100 mM ABC with incubation at 37ºC for 20 minutes. The DTT solution was removed, and the 529 samples were cooled to room temperature. Samples were alkylated by the addition of an equal 530 volume of 10 mM iodoacetamide (IAN) in 100 mM ABC with incubation at room temperature in 531 the dark for 20 minutes. The IAN solution was removed, and the gels were washed for 20 min 532 with 1 ml 50% ACN/100 mM ABC, then washed for 20 min with 1 ml 50% ACN/25 mM ABC. 533 The gels were briefly dried in a SpeedVac, then resuspended in 100-150 µl of 25 mM ABC 534 containing 200-300 ng of sequencing grade trypsin (Promega, V5111) (volume sufficient to 535 cover gels) and incubated at 37ºC for 16 hours. The samples that were to be submitted for 536 phosphopeptide enrichment had more protein and so were digested with 500 ng of trypsin. The 537 supernatant containing the released tryptic peptides were moved to a new Eppendorf tube. 538 Residual peptides remaining in the gel were extracted by the addition of approximately 3 539 volumes of 80% ACN/0.1% trifluoroacetic acid and combined with the previous supernatant, and 540 the peptides were dried in a SpeedVac. For unenriched samples, peptides were dissolved in 23 541 µl MS loading buffer (2% ACN, 0.2% TFA in water), with 5 µl injected for LC-MS/MS analysis. 542 For select samples, after extraction, phosphorylated peptides were enriched using TiO2 543 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 14 (titanium dioxide) TopTips (TT1TIO.96; Glygen, Columbia, MD) with a slightly modified 544 manufacturer protocol. Briefly, the manufacturer protocol was utilized with the addition of 70 mM 545 L-glutamic acid in the loading buffer (65% acetonitrile, 2% trifluoroacetic acid). Bound 546 phosphopeptides on the TiO2 resin were washed with 65% acetonitrile, 2% trifluoroacetic acid, 547 and eluted with 2% ammonium hydroxide solution in water at pH 12. The enriched fraction was 548 acidified, and both enriched (EN) and flowthrough (FT) fractions were dried by SpeedVac prior 549 to dissolving in MS loading buffer for LC–MS/MS analysis. 550 LC-MS/MS on the Thermo Scientific Q Exactive Plus 551 LC-MS/MS analysis was performed as previously described (LI AND HOCHSTRASSER 2022) at the 552 Keck Mass Spectrometry & Proteomics Resource at the Yale School of Medicine on a Thermo 553 Scientific Q Exactive Plus equipped with a Waters nanoAcquity UPLC system utilizing a binary 554 solvent system (A: 100% water, 0.1% formic acid; B: 100% acetonitrile, 0.1% formic acid). 555 Trapping was performed at 5µl/min, 99.5% Buffer A for 3 min using a Waters ACQUITY UPLC 556 M-Class Symmetry C18 Trap Column (100Å, 5 µm, 180 µm x 20 mm, 2G, V/M). Peptides were 557 separated at 37°C using a Waters ACQUITY UPLC M-Class Peptide BEH C18 Column (130Å, 558 1.7 µm, 75 µm X 250 mm) and eluted at 300 nl/min with the following gradient: 3% buffer B at 559 initial conditions; 5% B at 1 minute; 25% B at 90 minutes; 50% B at 110 minutes; 90% B at 115 560 minutes; 90% B at 120 min; return to initial conditions at 125 minutes. Full MS spectra were 561 acquired in profile mode over the 300-1,700 m/z range using 1 microscan, 70,000 resolution, 562 AGC target of 3E6, and a maximum injection time of 45 ms. Data dependent MS/MS scans 563 were acquired in centroid mode on the top 20 precursors per MS scan using 1 microscan, 564 17,500 resolution, AGC target of 1E5, maximum injection time of 100 ms, and an isolation 565 window of 1.7 m/z. Precursors were fragmented by HCD activation with a collision energy of 566 28%. MS/MS spectra were collected on species with an intensity threshold of 1E4, charge 567 states 2-6, and peptide match preferred. Dynamic exclusion was set to 20 seconds. 568 Peptide identification 569 Tandem mass spectra were extracted by Proteome Discoverer software (version 2.2.0.388, 570 Thermo Scientific) and searched in-house using the Mascot algorithm (version 2.7.0, Matrix 571 Science). The data were searched against the SWISS-PROT database with taxonomy restricted 572 to Saccharomyces cerevisiae (v20211116; 7,921 sequences). Search parameters included 573 trypsin digestion with up to 2 missed cleavages, peptide mass tolerance of 10 ppm, and MS/MS 574 fragment tolerance of 0.02 Da. Cysteine carbamidomethylation was configured as a fixed 575 modification. Methionine oxidation; phosphorylation of serine, threonine and tyrosine; 576 propionamide adduct to cysteine; and GG dipeptide (ubiquitin residue) on lysine were 577 configured as variable modifications. Normal and decoy database searches were run, with the 578 confidence level was set to 95% (p<0.05). Scaffold Q+S (version 5.1.2, Proteome Software Inc.) 579 was used to validate MS/MS based peptide and protein identifications. Peptide identifications 580 were accepted if they could be established at greater than 95.0% probability by the Scaffold 581 Local FDR algorithm. Protein identifications were accepted if they could be established at 582 greater than 99.0% probability and contained at least 2 identified peptides. The phosphopeptide 583 enriched samples were further analyzed with Scaffold PTM (version 4.0.2, Proteome Software 584 Inc.). 585 586 Data availability 587 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 15 The proteomic data from this study has been deposited in Integrated Proteome Resources with 588 ID number IPX0012452000. A full list of proteins identified in the AMPK complex (Figure S4) 589 and proteasomes (Figure S6) by mass spectrometry analysis is provided in Table S5. All other 590 data are included in the manuscript. 591 592

593 We thank the Keck MS & Proteomics Resource at the Yale school of Medicine for providing the 594 necessary mass spectrometers and the accompanying biotechnology tools funded in part by the 595 YSM and NIH (S10OD02365101A1, S10OD019967, and S10OD018034). 596 597 Study Funding 598 The work was supported by the startup fund and provost seed grant from Florida Institute of 599 Technology to JL and NIH grant GM136325 to MH. 600 601 Conflicts of interest 602 The author(s) declare no conflict of interest. 603 604 Author contributions 605 JL: conceptualization, methodology, investigation, data curation, supervision, funding 606 acquisition, project administration, writing-original draft, and writing-review and editing; CB and 607 KV: investigation; MH: conceptualization, supervision, funding acquisition, project 608 administration, and writing-review and editing. 609 610 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 16

611 Ashraf, N., and J. L. Van Nostrand, 2024 Fine-tuning AMPK in physiology and disease using 612 point-mutant mouse models. Dis Model Mech 17. 613 Bard, J. A. M., E. A. Goodall, E. R. Greene, E. Jonsson, K. C. Dong et al., 2018 Structure and 614 Function of the 26S Proteasome. Annu Rev Biochem 87: 697-724. 615 Breitkreutz, A., H. Choi, J. R. Sharom, L. Boucher, V. Neduva et al., 2010 A global protein 616 kinase and phosphatase interaction network in yeast. Science 328: 1043-1046. 617 Butcher, C., K. VanderVen and J. Li, 2025 Proteasome condensates repeatedly "contact and 618 release" at the nuclear periphery during dissolution. MicroPubl Biol 2025. 619 Coccetti, P., R. Nicastro and F. Tripodi, 2018 Conventional and emerging roles of the energy 620 sensor Snf1/AMPK in Saccharomyces cerevisiae. Microb Cell 5: 482-494. 621 Coux, O., B. A. Zieba and S. Meiners, 2020 The proteasome system in health and disease. Adv 622 Exp Med Biol: 55-100. 623 Enenkel, C., and O. P. Ernst, 2025 Proteasome dynamics in response to metabolic changes. 624 Front Cell Dev Biol 13. 625 Enenkel, C., R. W. Kang, F. Wilfling and O. P. Ernst, 2022 Intracellular localization of the 626 proteasome in response to stress conditions. J Biol Chem 298: 102083. 627 Estruch, F., M. A. Treitel, X. Yang and M. Carlson, 1992 N-terminal mutations modulate yeast 628 SNF1 protein kinase function. Genetics 132: 639-650. 629 Ghillebert, R., 2011 The AMPK/SNF1/SnRK1 fuel gauge and energy regulator: structure, 630 function and regulation Regulation of AMPK/SNF1/SnRK1 kinase complexes. FEBS J 631 278: 3978-3990. 632 Grams, N., M. Charman, E. Halko, R. Lauman, B. A. Garcia et al., 2024 Phosphorylation 633 regulates viral biomolecular condensates to promote infectious progeny production. 634 EMBO J 43: 277-303. 635 Gu, Z. C., E. Wu, C. Sailer, J. Jando, E. Styles et al., 2017 Ubiquitin orchestrates proteasome 636 dynamics between proliferation and quiescence in yeast. Mol Biol Cell 28: 2479-2491. 637 Guo, X., X. Huang and M. J. Chen, 2017 Reversible phosphorylation of the 26S proteasome. 638 Protein Cell 8: 255-272. 639 Hirano, H., Y. Kimura and A. Kimura, 2015 Biological significance of co- and post-translational 640 modifications of the yeast 26S proteasome. J Proteomics 134: 37-46. 641 Jeon, S.-M., 2016 Regulation and function of AMPK in physiology and diseases. Exp Mol Med 642 48: e245-e245. 643 Jeon, S., Y. Jeon, J.-Y. Lim, Y. Kim, B. Cha et al., 2025 Emerging regulatory mechanisms and 644 functions of biomolecular condensates: implications for therapeutic targets. Signal 645 Transduct Target Ther 10: 4. 646 Jiang, R., and M. Carlson, 1996 Glucose regulates protein interactions within the yeast SNF1 647 protein kinase complex. Gene Dev 10: 3105-3115. 648 Jiang, R., and M. Carlson, 1997 The Snf1 protein kinase and its activating subunit, Snf4, 649 interact with distinct domains of the Sip1/Sip2/Gal83 component in the kinase complex. 650 Mol Cell Biol 17: 2099-2106. 651 Kinoshita, E., E. Kinoshita-Kikuta, K. Takiyama and T. Koike, 2006 Phosphate-binding tag, a 652 new tool to visualize phosphorylated proteins. Mol Cell Proteomics 5: 749-757. 653 Kors, S., K. Geijtenbeek, E. Reits and S. Schipper-Krom, 2019 Regulation of proteasome 654 activity by (post-)transcriptional mechanisms. Front Mol Biosci 6: 48. 655 Kroschwald, S., S. Maharana and A. W. Simon, 2017 Hexanediol: a chemical probe to 656 investigate the material properties of membrane-less compartments. Matters. 657 Kushnirov, V. V., 2000 Rapid and reliable protein extraction from yeast. Yeast 16: 857-860. 658 Laporte, D., B. Salin, B. Daignan-Fornier and I. Sagot, 2008 Reversible cytoplasmic localization 659 of the proteasome in quiescent yeast cells. J Cell Biol 181: 737-745. 660 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 17 Li, J., M. Breker, M. Graham, M. Schuldiner and M. Hochstrasser, 2019 AMPK regulates 661 ESCRT-dependent microautophagy of proteasomes concomitant with proteasome 662 storage granule assembly during glucose starvation. PLoS Genet 15: e1008387. 663 Li, J., S. Fuchs, J. Zhang, S. Wellford, M. Schuldiner et al., 2016 An unrecognized function for 664 COPII components in recruiting the viral replication protein BMV 1a to the perinuclear 665 ER. J Cell Sci 129: 3597-3608. 666 Li, J., and M. Hochstrasser, 2020 Microautophagy regulates proteasome homeostasis. Curr 667 Genet 66: 683-687. 668 Li, J., and M. Hochstrasser, 2022 Selective microautophagy of proteasomes is initiated by 669 ESCRT-0 and is promoted by proteasome ubiquitylation. J Cell Sci 135. 670 Liu, X., W. Xiao, Y. Zhang, S. E. Wiley, T. Zuo et al., 2020 Reversible phosphorylation of Rpn1 671 regulates 26S proteasome assembly and function. Proc Natl Acad Sci U S A 117: 328-672 336. 673 McCartney, R. R., and M. C. Schmidt, 2001 Regulation of Snf1 kinase: activation requires 674 phosphorylation of theronine 210 by an upstream kinase we well as a distinct step 675 mediated by the Snf4 subunit. J Biol Chem 276: 36460-36466. 676 Mruk, D. D., and C. Y. Cheng, 2011 Enhanced chemiluminescence (ECL) for routine 677 immunoblotting: An inexpensive alternative to commercially available kits. 678 Spermatogenesis 1: 121-122. 679 Mumberg, D., R. Müller and M. Funk, 1995 Yeast vectors for the controlled expression of 680 heterologous proteins in different genetic backgrounds. Gene 156: 119-122. 681 Narayanaswamy, R., M. Levy, M. Tsechansky, G. M. Stovall, J. D. O'Connell et al., 2009 682 Widespread reorganization of metabolic enzymes into reversible assemblies upon 683 nutrient starvation. Proc Natl Acad Sci USA 106: 10147-10152. 684 Nath, N., R. R. McCartney and M. C. Schmidt, 2002 Purification and characterization of Snf1 685 kinase complexes containing a defined Beta subunit composition. J Biol Chem 277: 686 50403-50408. 687 Nosella, M. L., and J. D. Forman-Kay, 2021 Phosphorylation-dependent regulation of 688 messenger RNA transcription, processing and translation within biomolecular 689 condensates. Curr Opin Cell Biol 69: 30-40. 690 Pack, C.-G., H. Yukii, A. Toh-e, T. Kudo, H. Tsuchiya et al., 2014 Quantitative live-cell imaging 691 reveals spatio-temporal dynamics and cytoplasmic assembly of the 26S proteasome. 692 Nat Commun 5: 3396. 693 Peters, L. Z., R. Hazan, M. Breker, M. Schuldiner and S. Ben-Aroya, 2013 Formation and 694 dissociation of proteasome storage granules are regulated by cytosolic pH. J Cell Biol 695 201: 663-671. 696 Ranganathan, S., P. Dasmeh, S. Furniss and E. Shakhnovich, 2023 Phosphorylation sites are 697 evolutionary checkpoints against liquid–solid transition in protein condensates. Proc Natl 698 Acad Sci U S A 120: e2215828120. 699 Rock, K. L., C. Gramm, L. Rothstein, K. Clark, R. Stein et al., 1994 Inhibitors of the proteasome 700 block the degradation of most cell proteins and the generation of peptides presented on 701 MHC class I molecules. Cell 78: 761-771. 702 Schnell, H. M., M. Jochem, Y. Micoogullari, C. L. Riggs, P. Ivanov et al., 2021 Reg1 and Snf1 703 regulate stress-induced relocalization of protein phosphatase-1 to cytoplasmic granules. 704 FEBS J n/a. 705 Sikorski, R. S., and P. Hieter, 1989 A system of shuttle vectors and yeast host strains designed 706 for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122: 19-27. 707 Sridharan, S., A. Hernandez-Armendariz, N. Kurzawa, C. M. Potel, D. Memon et al., 2022 708 Systematic discovery of biomolecular condensate-specific protein phosphorylation. Nat 709 Chem Biol 18: 1104-1114. 710 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 18 Steinberg, G. R., and D. Carling, 2019 AMP-activated protein kinase: the current landscape for 711 drug development. Nat Rev Drug Discov 18: 527-551. 712 Sultana, S., M. Abdullah, J. Li, M. Hochstrasser and A. H. Kachroo, 2023 Species-specific 713 protein–protein interactions govern the humanization of the 20S proteasome in yeast. 714 Genetics 225. 715 Sutherland, C. M., S. A. Hawley, R. R. McCartney, A. Leech, M. J. R. Stark et al., 2003 Elm1p is 716 one of three upstream kinases for the Saccharomyces cerevisiae SNF1 complex. Curr 717 Biol 13: 1299-1305. 718 Thomas, M. G., A. J. Fernández-Alvarez, M. Giménez, F. Corvetto Aristarain, L. H. Cozza et al., 719 2025 The non-canonical Smoothened-AMPK axis regulates Smaug1 biomolecular 720 condensates. J Cell Sci 138. 721 Tu, Y., C. Chen, J. Pan, J. Xu, Z. G. Zhou et al., 2012 The Ubiquitin Proteasome Pathway 722 (UPP) in the regulation of cell cycle control and DNA damage repair and its implication in 723 tumorigenesis. Int J Clin Exp Pathol 5: 726-738. 724 VanderVen, K., C. Butcher, R. Fokine and J. Li, 2025 Pep12 is important for proteasome 725 microautophagy under low glucose conditions. MicroPubl Biol 2025. 726 Vincent, O., R. Townley, S. Kuchin and M. Carlson, 2001 Subcellular localization of the Snf1 727 kinase is regulated by specific β subunits and a novel glucose signaling mechanism. 728 Genes Dev 15: 1104-1114. 729 Waite, K. A., G. Vontz, S. Y. Lee and J. Roelofs, 2024 Proteasome condensate formation is 730 driven by multivalent interactions with shuttle factors and ubiquitin chains. Proc Natl 731 Acad Sci USA 121: e2310756121. 732 Wani, P. S., A. Suppahia, X. Capalla, A. Ondracek and J. Roelofs, 2016 Phosphorylation of the 733 C-terminal tail of proteasome subunit α7 is required for binding of the proteasome quality 734 control factor Ecm29. Sci Rep 6: 27873. 735 Zungu, M., J. C. Schisler, M. F. Essop, C. McCudden, C. Patterson et al., 2011 Regulation of 736 AMPK by the ubiquitin proteasome system. Am J Pathol 178: 4-11. 737 738 739 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 19 Figures and figure legends 740 741 Figure 1. Proteasome condensates (proteasome storage granules or PSGs) gradually lose 742 reversibility in the absence of AMPK under low glucose conditions. (a) Fluorescence 743 microscopic images of the CP subunit Pre10-GFP, the lid subunit Rpn5-yEGFP, and the base 744 subunit Rpn2-GFP in WT and AMPK deletion mutant snf1∆ and snf4∆ strains grown under low 745 glucose conditions for one or four days and after 15 min of glucose recovery. There are only 746 one to two PSGs per cell when starved. BF: bright field (phase). Scale bar: 5 µm. Persistent 747 proteasome condensates were formed in snf1∆ and snf4∆ cells on day 4 under low glucose 748 conditions. (b, c) Quantification of proteasome condensates in WT, snf1∆, and snf4∆ cells 749 following low glucose growth and after 15 min glucose recovery on day one (b) and day four (c). 750 Bar graph results were plotted as mean±s.d. of the percentage of proteasome condensates in 751 the indicated strain. Each sample has four data points that indicate the percentages. At least 752 three repeats were conducted. More than 300 cells were counted for each sample. 753 754 755 756 757 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 20 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 Figure 2. Highly phosphorylated proteasome subunits accumulate in AMPK deletion mutants 773 after prolonged incubation in low glucose. Anti-GFP immunoblot analyses of proteasome 774 subunit phosphorylation status (Phos-tag) and total proteasome subunit accumulation levels of 775 Pre10-GFP (a), Rpn5-yEGFP (b), and Rpn2-GFP (c) in WT and AMPK mutant snf1∆ and snf4∆ 776 cells. Phosphorylation retards the migration of proteins in the Phos-tag gels. Cells were cultured 777 at the indicated glucose concentrations and harvested at the indicated times. Pgk1 778 (phosphoglycerate kinase-1) served as a loading control. 779 780 781 782 783 784 785 786 787 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 21 Figure 3. Snf1 activation and kinase activity are essential for both proteasome condensate 788 dissolution and regulation of proteasome phosphorylation during glucose starvation and 789 refeeding. (a) Percentage of cells bearing proteasome condensates tagged with Pre10-GFP, 790 Rpn5-yEGFP, or Rpn2-GFP in snf1∆ cells carrying either an empty vector (EV) or plasmids with 791 the indicated snf1 alleles expressed from the native SNF1 promoter. Mutants snf1-T210A and 792 snf1-T210D are defective for kinase activity. (b) Percentage of cells bearing proteasome 793 condensates tagged as indicated and grown under the indicated culture conditions. Mutant 794 snf1∆ cells carried either an empty vector (EV) or plasmids with the indicated snf1 alleles 795 expressed from the native SNF1 promoter. The snf1-K84R protein is kinase defective, while 796 snf1-G53R has elevated kinase activity. **P<0.01 and ***P<0.001 (one-way ANOVA analysis 797 comparing snf1-G53R to Snf1 at indicated growth conditions. (c) Anti-GFP immunoblot analyses 798 of proteasome subunit phosphorylation status (Phos-tag) and total proteasome subunit 799 accumulation levels of Pre10-GFP, Rpn5-yEGFP, and Rpn2-GFP in cells from panel a. (d) Anti-800 GFP immunoblot analyses of proteasome subunit phosphorylation status and total proteasome 801 subunit accumulation levels of Pre10-GFP, Rpn5-yEGFP, and Rpn2-GFP in cells from panel b. 802 Cells were cultured under low glucose for about one day and four days, and subjected to 803 glucose recovery for 15 min. Bar graph results were plotted as mean±s.d. of the percentage of 804 proteasome condensates in the indicated strain. Each sample has four data points that indicate 805 the percentages. At least three repeats were conducted. More than 300 cells were counted for 806 each sample in panels a and b. Pgk1 served as a loading control in panels c and d. 807 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 22 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 Figure 4. Nuclear and cytoplasmic AMPK complexes function redundantly in regulating 826 proteasome condensate dissolution. (a) Fluorescence microscopy images of Pre10-GFP, Rpn5-827 yEGFP, and Rpn2-GFP in WT and sip2∆ gal83∆ cells grown under low glucose conditions for 828 about four days and following glucose recovery for 15 min. BF: bright field. Scale bar: 5 µm. (b) 829 Quantification of cells with proteasome condensates in WT and sip2∆ gal83∆ cells. The cells 830 were from panel a. Bar graph results were plotted as mean±s.d. of the percentage of 831 proteasome condensates in the indicated strain. Each sample has four data points that indicate 832 the percentages. At least three repeats were conducted. More than 300 cells were counted for 833 each sample. 834 835 836 837 838 839 840 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 23 841 842 Figure 5. AMPK interacts with proteasomes in a Snf1 activity-independent manner. (a) Co-IP 843 analyses of Snf1-HA and Snf4-HA with proteasomes in snf1∆ and snf4∆ cells carrying plasmids 844 p424GPD-SNF1-HA and p426GPD-SNF4-HA, respectively. Snf4 interacted with both the CP 845 subunit Pre10 and the RP subunit Rpn5, while Snf1 interacted detectably only with Rpn5, as 846 shown in the long exposure blot. (b) Co-IP analyses of Snf1-HA and Snf1 kinase mutants with 847 Rpn5. Cells were grown under low glucose conditions for about one day. Pgk1 served as a 848 loading control. Relative to the amount that was loaded in immunoprecipitated samples, ∼0.83% 849 of total cell lysates were loaded for the anti-HA and 6.67% for the anti-GFP blots. 850 851 852 853 854 855 856 857 858 859 860 861 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 24 862 Figure 6. Solid-like proteasome condensates accumulate in mutant cells lacking AMPK. (a) 863 Fluorescence microscopy images of Rpn5-yEGFP in WT, snf1∆, snf4∆ cells grown under low 864 glucose conditions for about four days and following 1,6-HD treatment for 20 min. Scale bar: 5 865 µm. (b) Quantification of proteasome condensates in WT, snf1∆, and snf4∆ cells cultured under 866 low glucose conditions for about one day followed by incubation of cells in 6% 1,6-HD for 20 867 minutes at room temperature (RT). (c) Quantification of proteasome condensates in WT, snf1∆, 868 and snf4∆ cells cultured under low glucose conditions for about four days followed by 10% 1,6-869 HD treatment for 20 minutes at RT. Bar graph results were plotted as mean±s.d. of the 870 percentage of proteasome condensates in the indicated strain. Each sample has four data 871 points that indicate the percentages. At least three repeats were conducted. More than 300 cells 872 were counted for each sample. Proteasome condensate solubility in 1,6-HD was reduced over 873 time under low glucose conditions. 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint 25 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 Figure 7. AMPK affects the phosphorylation of several proteasome subunits. Proteasomes 905 were affinity purified (with an anti-Flag resin) from WT, snf1∆, and snf4∆ cells expressing 906 Rpn11-3xFlag under the indicated glucose growth conditions. Ten µg of proteasomes were 907 resolved in 12% SDS-PAGE gels, stained with protein phospho-stain (right panels) followed by 908 SYPRO Ruby stain (left panels). Pronounced changes in the phosphorylation status of several 909 proteasome subunits were observed in cells without AMPK on day four under low glucose 910 conditions. Proteasome subunits were labeled based on their molecular weights, so some labels 911 might be inaccurate. 912 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted July 2, 2025. ; https://doi.org/10.1101/2025.07.01.662652doi: bioRxiv preprint