Protein acetylation microarray reveals that NuA4 controls key metabolic target regulating gluconeogenesis

Abstract

Histone acetyltransferases (HATs) and histone deacetylases (HDACs) conduct many critical functions through nonhistone substrates in metazoans, but only chromatin-associated nonhistone substrates are known in Saccharomyces cerevisiae. Using yeast proteome microarrays, we identified and validated many nonchromatin substrates of the essential nucleosome acetyltransferase of H4 (NuA4) complex. Among these, acetylation sites (Lys19 and 514) of phosphoenolpyruvate carboxykinase (Pck1p) were determined by tandem mass spectrometry. Acetylation at Lys514 was crucial for enzymatic activity and the ability of yeast cells to grow on nonfermentable carbon sources. Furthermore, Sir2p deacetylated Pck1p both in vitro and in vivo. Loss of Pck1p activity blocked the extension of yeast chronological life span caused by water starvation. In human hepatocellular carcinoma (HepG2) cells, human Pck1 acetylation and glucose production were dependent on TIP60, the human homolog of ESA1. Our findings demonstrate a regulatory function for the NuA4 complex in glucose metabolism and life span by acetylating a critical metabolic enzyme.

Figure 1. A protein acetylation microarray dentifies yeast Nua4 substrates

(A) Acetylation reactions on a yeast proteome microarray fabricated on the FAST surface. A total of 91 proteins were acetylated by the NuA4 complex. (B) In vitro validation of seven representative candidate substrates. GST fusion proteins were purified from the esa1-531 mutant grown at non-permissive temperature and incubated with/without (+/−) the NuA4 complex in vitro. The reaction complex was immunoprecipitated with anti-pan-acetyl-lysine (α-Ac-K) followed by immunoblotting with anti-glutathione-S-transferase (α-GST). (C) Identification of five in vivo substrates (Pck1p, Cdc34p, Prp19p, Gph1p and Tap42p) of NuA4 complex. Overexpressed GST fusion proteins were purified from WT and esa1-531 mutant grown at non-permissive temperature, and immunoprecipitated with α-Ac-K followed by immunoblotting with α-GST.

Figure 2. In vitro and in vivo detection of acetylated K514 of Pck1p

(A and B) Substitution of K514 with arginine (K514R) diminished in vitro acetylation of Pck1p by the NuA4 complex. A conventional HAT activity assay revealed that Pck1p, but not Pck1p-K514R (both purified from overexpression strains), was acetylated in vitro by the NuA4 complex. The results are presented in a bar graph with error bars indicating ± 1 s.e.m from three biological replicates. Double asterisk P < 0.01 and triple asterisk P < 0.001 (Student’s t-test). (B) Prominent in vitro acetylation of purified Pck1p purified from the esa1-531 mutant grown at non-permissive temperature, but not Pck1p-K514R, by the NuA4 complex. See Figure 1B legend for assay protocols. (C) Overexpressed Pck1p was acetylated at K514 in vivo by the NuA4 complex. See Figure 1C legend for assay protocols. (D) In vivo acetylation of endogenous Pck1p in WT, esa1-531, pck1-K514R and sir2Δ mutants was assessed using the protocols described in Figure 1C. (E) In vivo acetylation of overexpressed Pck1p in wild type and pck1-K19R mutant. Acetylation at K19 contributes minimally to the entire acetylation fraction of Pck1p. (F) The NuA4 complex efficiently acetylated purified Pck1p overexpressed in pck1-K19R mutant in vitro.

Figure 3. Sir2p deacetylates Pck1p K514

(A and B) Pck1p is deacetylated in vivo and in vitro by Sir2p. Overexpressed Pck1p was purified from WT, hda1Δ, rpd3Δ, or sir2Δ mutants and immunoprecipitated using anti-pan-acetyl-lysine monoclonal antibody as described in Figure 1E. Acetylated Pck1p increased significantly in sir2Δ strain when compared to WT, hda1Δ, and rpd3Δ. Purified Sir2-TAP, but not Hda1-TAP, efficiently deacetylated Pck1p in vitro in an NAD⁺-dependent manner. (C) Endogenous protein levels of Pck1p were detected by immunoblot analysis of the whole cell extracts derived from cells expressing C-terminally Myc-tagged Pck1p. (D) Levels of PCK1 mRNA were detected by RT-PCR in WT, esa1-531, sir2Δ, gcn5Δ, and hda1Δ strains. The results are presented in a bar graph with error bars indicating ± 1 s.e.m from three biological replicates. Single and triple asterisk indicates P < 0.05 and P < 0.001 (Student’s t-test), respectively. Primers used in the RT-PCR reactions are provided in Table S3.

Figure 4. Pck1p acetylation controls activity

(A) Growth of WT, pck1Δ, esa1-531, pck1-K514R (K514R), pck1-K514Q (K514Q) and esa1-531 pck1-K514Q (esa1-531 K514Q) double mutant strains was examined on SC medium plates containing 2% glucose, 3% ethanol or 2% glycerol + 2% ethanol (GE). (B) Growth of WT, esa1-531, gcn5Δ, sir2Δ, hda1Δ, and pck1Δ strains was monitored in the same conditions as described above. (C) Growth of WT, pck1Δ, pck1-K514Q (K514Q) and sir2Δ was compared on SC plates containing 2% glucose, 6% ethanol, or 15% ethanol. (D) Pck1p activity was assayed in a coupled reaction, in which oxaloacetate (OAA) formed from phosphoenol pyruvate (PEP) was reduced to malate by NADH in the presence of malate dehydrogenase (MDH). Rate of NADH oxidation was measured at 340 nm. (E) Enzymatic activity of Pck1p purified from WT, esa1-531, pck1-K514R, and esa1-531 pck1-K514Q mutants was measured by consumption of NADH as a function of time. The curves of esa1-531 and pck1-K514R overlapped with each other. (F) K_m and V_max values of Pck1p purified from WT and esa1-531 were determined based on Michaelis-Menten Kinetics.

Figure 5. PCK1 activity extends chronological life span under water starvation

(A, B) Chronological survival of wild type (WT, BY4742), esa1-531, pck1Δ, sir2Δ, pck1-K514R, pck1-K514Q, sir2Δ pck1Δ, sir2Δ pck1-K514R and sir2Δ pck1-K514Q in liquid synthetic complete medium containing either 2% glucose or water at 30°C was assessed. Error bars indicate ± 1 s.e.m from three biological replicates. (D) Ethanol concentration in cell-free media from day 1, 3, 5 and 7 cultures of wild type (WT, BY4742), esa1-531, pck1Δ, sir2Δ, pck1-K514R, pck1-K514Q, sir2Δ pck1Δ, sir2Δ pck1-K514R and sir2Δ pck1-K514Q at 30°C was measured. Error bars indicate ± 1 s.e.m from three biological replicates.

Figure 6. TIP60-dependent acetylation of human Pck1 and glucose production

(A) Human PCK1 (hPCK1) but not PCK2 (hPCK2) rescued the growth defect of yeast pck1Δ mutant on ethanol and GE. Growth was monitored as described in Figure 4A. (B) Endogenous human Pck1 level decreased in HepG2 cells when SIRT1 was silenced. (C) Knockdown of TIP60 diminished the in vivo acetylation of human Pck1 in HepG2 cells. (D) Treatment of histone deacetylase inhibitors (HDACi) such as trichostatin A (TSA) or nicotinamide (NAM) increased the in vivo acetylation of human Pck1 in HepG2 cells. (E) Glucose output of HepG2 cells was dependent on TIP60. The results are presented in a bar graph with error bars indicating ± 1 s.e.m from three biological replicates. Double asterisk indicates P < 0.01 (Student’s t-test).

Figure 7. Schematic model for regulation of Pck1p activity in yeast and humans

A “triple-switch” (transcription of PCK1, acetylation of Pck1p and allosteric modification of PC) in the gluconeogenesis pathway controls prompt adaptation of a metabolic flux to energy status in yeast (A) and humans (B). The change of metabolic flux also mediates extension of chronological longevity under water starvation in yeast cells. ACS, acetyl-coenzyme A synthetase. AceCS1, acetyl-coenzyme A synthetase 1. PC, pyruvate carboxylase. Solid lines indicate the paths that have been proved by previous and the present study; dashed lines indicate those paths that were hypothesized as part of the present study.