Intrasteric control of AMPK via the gamma1 subunit AMP allosteric regulatory site

Abstract

AMP-activated protein kinase (AMPK) is a alphabetagamma heterotrimer that is activated in response to both hormones and intracellular metabolic stress signals. AMPK is regulated by phosphorylation on the alpha subunit and by AMP allosteric control previously thought to be mediated by both alpha and gamma subunits. Here we present evidence that adjacent gamma subunit pairs of CBS repeat sequences (after Cystathionine Beta Synthase) form an AMP binding site related to, but distinct from the classical AMP binding site in phosphorylase, that can also bind ATP. The AMP binding site of the gamma(1) CBS1/CBS2 pair, modeled on the structures of the CBS sequences present in the inosine monophosphate dehydrogenase crystal structure, contains three arginine residues 70, 152, and 171 and His151. The yeast gamma homolog, snf4 contains a His151Gly substitution, and when this is introduced into gamma(1), AMP allosteric control is substantially lost and explains why the yeast snf1p/snf4p complex is insensitive to AMP. Arg70 in gamma(1) corresponds to the site of mutation in human gamma(2) and pig gamma(3) genes previously identified to cause an unusual cardiac phenotype and glycogen storage disease, respectively. Mutation of any of AMP binding site Arg residues to Gln substantially abolishes AMP allosteric control in expressed AMPK holoenzyme. The Arg/Gln mutations also suppress the previously described inhibitory properties of ATP and render the enzyme constitutively active. We propose that ATP acts as an intrasteric inhibitor by bridging the alpha and gamma subunits and that AMP functions to derepress AMPK activity.

Figure 1.

Stereoribbon representation of γ₁ model. Showing CBS sequence 1 in blue and CBS sequence 2 in red, the gray loops show the unmodeled connecting loops.

Figure 2.

Comparison of phosphorylase and AMPK γ subunit binding pockets for AMP. (A) Stereofigure of the glycogen phosphorylase a AMP binding site; note helices 2 and 10 (Rath et al. 2000). (B) Surface representation of the glycogen phosphorylase a AMP binding site (Rath et al. 2000). Showing the large positively charged blue patch for AMP binding. (C) Surface representation of the γ subunit AMP binding domain showing the positive blue patch of Arg and His residues. (D) Stereo figure of the γ subunit AMP binding domain (note the cluster of basic residues [R70, H151, R152, R171] in the surface cleft).

Figure 3.

AMPK γ sequence alignment. Amino acid alignment of γ subunit sequences with putative AMP binding site residues contributed from CBS1/CBS2 pair marked by an asterisk. Residue numbers refer to the rat γ₁ isoform.

Figure 4.

AMPK γ₁ AMP binding site. (A) Stereofigure of the γ₁ AMP binding site with AMP docked into the site. Residues in close proximity to the bound AMP are shown in ball and stick. (B) Surface representation of the γ subunit binding site colored by residue type (aliphatic [white], positive [blue], negative [red], and polar [green]), with docked structures of ATP (purple) and AMP (turquoise).

Figure 5.

Effect of Arg mutations on AMP activation. Wild-type γ₁ or mutant γ₁ was cotransfected with α₁ and β₁ in COS-7 cells and AMPK αβγ holoenzyme was purified by Glutathione Sepharose chromatograph. AMPK activity was determined using the SAMS peptide (see Materials and Methods). The activity measured in the absence of AMP was subtracted from plus AMP values of the wild-type and Phe mutants and fitted to the one-site binding Michaelis-Menten Equation curve. The activity of the Arg mutants was AMP-independent.

Figure 6.

Effect of ATP concentration on AMP dependence. AMPK γ₁ wild-type (A, C) and R152Q (B, D) were cotransfected with α₁ and β₁ in COS-7 cells. The AMPK activities were measured in the varied AMP and ATP concentrations as indicated (A, B). Changes in AMPK activity, relative to AMPK activity in the absence of AMP, are shown (C, D).