A comprehensive protein-protein interactome for yeast PAS kinase 1 reveals direct inhibition of respiration through the phosphorylation of Cbf1

Abstract

Per-Arnt-Sim (PAS) kinase is a sensory protein kinase required for glucose homeostasis in yeast, mice, and humans, yet little is known about the molecular mechanisms of its function. Using both yeast two-hybrid and copurification approaches, we identified the protein-protein interactome for yeast PAS kinase 1 (Psk1), revealing 93 novel putative protein binding partners. Several of the Psk1 binding partners expand the role of PAS kinase in glucose homeostasis, including new pathways involved in mitochondrial metabolism. In addition, the interactome suggests novel roles for PAS kinase in cell growth (gene/protein expression, replication/cell division, and protein modification and degradation), vacuole function, and stress tolerance. In vitro kinase studies using a subset of 25 of these binding partners identified Mot3, Zds1, Utr1, and Cbf1 as substrates. Further evidence is provided for the in vivo phosphorylation of Cbf1 at T211/T212 and for the subsequent inhibition of respiration. This respiratory role of PAS kinase is consistent with the reported hypermetabolism of PAS kinase-deficient mice, identifying a possible molecular mechanism and solidifying the evolutionary importance of PAS kinase in the regulation of glucose homeostasis.

FIGURE 1:

A Psk1 Y2H construct with increased protein–protein interaction proficiency. (A) A diagram of previously isolated hyperactive truncations of Psk1 (Grose et al., 2009) that were screened for their ability to bind protein partners in a Y2H screen. (B) The relative Y2H interaction strength of each Psk1 truncation is shown using Pbp1, a Psk1 binding partner identified in this study by both Y2H and copurification approaches. Y2HGold cells (Clontech) containing the Pbp1 prey (pJG1001) were cotransformed with bait plasmids harboring full-length Psk1 (pJG441), truncated Psk1 (ΔN477Psk1 [pJG1005], pΔN692Psk1 [pJG598], pΔN694Psk1 [pJG1006], pΔN931Psk1 [pJG568]) or empty vector (pJG425). Overnights were grown in SD-Leu-Trp for plasmid maintenance, diluted fivefold serially, and plated on Y2H selective media (SD-Leu-Trp-His-Ade + X-α-Gal), as well as a control plate (SD-Leu-Trp). Plates were incubated at 30°C for 3–4 d until colonies were apparent. The Y2H Gold strain contains four reporters under the control of three different Gal4-responsive promoters (Clontech), allowing selection for histidine (His) or adenine (Ade) biosynthesis, as well as blue colony formation on media containing X-α-Gal.

FIGURE 2:

Evidence for in vitro phosphorylation of Psk1 binding partners. A subset of 25 Psk1 binding partners was expressed and purified from E. coli BL21 cells and then assayed for in vitro phosphorylation by Psk1. (A) Utr1, Osh7, Cbf1, and Zds1 were shown to be phosphorylated by Psk1 in vitro. (B) Mot3 displayed faint bands with full-length Psk1 and was confirmed as an in vitro Psk1 substrate using a hyperactive Psk1 truncation (∆N931Psk1). (C) Cbf1, Utr1, Zds1, and Mot3 were verified as Psk1 substrates using a kinase-dead mutant (Psk1-D1230A). For kinase assays, purified proteins were incubated with radiolabeled ATP (³²P) in the presence or absence of purified Psk1. Kinase reactions were visualized on 8 or 12% SDS–PAGE gels, stained with Coomassie brilliant blue (CB), and exposed on x-ray film.

FIGURE 3:

Enrichment of GO processes and cellular localization for the putative Psk1 binding partners. (A) GO processes enriched in the Psk1 interactome with corresponding p-values from FunSpec (Robinson et al., 2002) analysis. Putative Psk1 binding partners belonging to each process are provided above the bar. GO process names are simplified due to space constraints, and abbreviations include signal transduction (sig. trans.) and negative regulation (neg. reg.). (B) FunSpec analysis reveals enrichment of proteins localized to the cytoplasm and mitochondria. Percentage of proteins from the Psk1 interactome and corresponding p-values are reported. Of the 93 putative binding partners, eight were not enriched in either the cytoplasm or mitochondria and were placed in an “other” category.

FIGURE 4:

In vitro and in vivo evidence for the Psk1-dependent phosphorylation of Cbf1 at T211/T212. (A) Mass spectrometry of Cbf1 phosphorylated in vitro revealed five Psk1-dependent phosphorylation sites (T138, T154, S156, T211, and T212) that are indicated on a diagram, along with the conserved helix-loop-helix (HLH) domain. (B) In vitro kinase assays using purified wild type and mutant Cbf1 (T211A/T212A, T138A, and T154A/S156A) reveal T211/T212 as necessary sites for Psk1-dependent phosphorylation. Kinase assays were run with radiolabeled ATP (³²P), visualized on 12% SDS–PAGE gels, stained with Coomassie brilliant blue (CB; bottom), and exposed on x-ray film (top). (C–F) In vivo evidence for Psk1-dependent respiratory inhibition through the phosphorylation and inactivation of Cbf1. (C) PSK-deficient yeast (JGY1244) display increased respiration rates, whereas CBF1-deficient (JGY1227) or CBF1PSK-deficient (JGY1261) yeast display decreased rates. (D) Overexpression of PSK1 (JGY1241) causes a significant decrease in respiration rates. (E) Yeast harboring Cbf1-T211A/T212A (JGY1263) display dramatically increased respiration rates compared with yeast harboring wild type (JGY1265) or Cbf1-T154A/S156A (JGY1264). Routine respiration (R) was determined by measuring O₂ consumption using an Oroboros O₂K Oxygraph in the absence of substrate and then with glucose (20 mM), ethanol (2%), or in the presence of the uncoupler carbonyl cyanide p-(trifluoro-methoxy) phenylhydrazine (FCCP; 70 μM). OD₆₀₀ was taken to ensure equal growth among strains. (F) Phosphosite mutants were expressed in yeast and analyzed by Western blot to ensure equal expression.

FIGURE 5:

A model for the role of PAS kinase in regulating glucose allocation. The PAS kinase–dependent phosphorylation of Ugp1 directs UDP-glucose utilization away from glycogen biosynthesis and toward the biosynthesis of structural carbohydrates such as β-glucans (Rutter et al., 2002; Smith and Rutter, 2007; Grose et al., 2007, 2009). The inhibitory phosphorylation of Psk1 on Cbf1 decreases respiration, which may lead to a further diversion of glucose toward structural carbohydrates necessary for growth. Dashed lines indicate inhibition of a pathway, and thick lines indicate activation.