Fluorescent AMP-Kinase Biosensors

Abstract

The present invention deals with heterotrimeric AMP-activated protein kinase (AMPK) comprising a fluorophore pair wherein the conformational change can be measured by FRET. It represents an advanced tool to screen and identify AMPK interactors in vitro and in cells in vivo. Such invention can also be considered as a reporter of the cellular energy status as it allows the spatiotemporal monitoring, in situ, of fluctuations in the ratio of AMP and ADP versus ATP.

Claims

1-29. (canceled)

30: A heterotrimeric AMP-activated protein kinase (AMPK) construct comprising: an AMPK -subunit or a mutant or fragment thereof, an AMPK -subunit or a mutant or fragment thereof, an AMPK -subunit or a mutant or fragment thereof, and a fluorescent dye pair tagging at least one of said -subunit, -subunit, -subunit, mutants or fragments, said fluorescent dye pair being placed to allow detection of conformational changes within the AMPK construct.

31: The AMPK construct of claim 30, wherein the AMPK construct retains the activity of a kinase and the capacity of being allosterically activated.

32: The AMPK construct of claim 30, wherein at least two of the subunits selected from the group consisting of the -subunit, the -subunit and the -subunit are tagged with the fluorescent dye pair.

33: The AMPK construct of claim 30, wherein the -subunit and either the -subunit or the -subunit are tagged with the fluorescent dye pair.

34: The AMPK construct of claim 30, wherein at least one of said -subunit, the -subunit, the -subunit, mutants or fragments is tagged with the fluorescent dye pair at the C-terminus.

35: The AMPK construct of claim 30, wherein the subunit is the 2 isoform, the subunit is 2 isoform, and the subunit is 1 isoform.

36: The AMPK construct of claim 35, wherein the 2 isoform is tagged with the fluorescent dye pair at the C-terminus, and either the 2 or the 1 isoform is tagged at the C-terminus with the fluorescent dye pair.

37: The AMPK construct of claim 30, wherein the fluorescent dye pair is a Frster Resonance Energy Transfer (FRET) pair.

38: The AMPK construct of claim 30, wherein the fluorescent dye pair is a pair of genetically encoded fluorescent proteins.

39: The AMPK construct of claim 30, wherein the fluorescent dye pair is a FRET pair, which comprises a first fluorophore and a second fluorophore, wherein the first fluorophore and the second fluorophore are each independently selected from the group consisting of Green Fluorescent Proteins (GFP), Cyan Fluorescent Proteins CFP, Yellow Fluorescent Proteins (YFP), Orange Fluorescent Proteins and Red Fluorescent Proteins.

40: The AMPK construct of claim 39, wherein the first fluorophore and the second fluorophore are each independently selected from the group consisting of GFP, CFP and YFP.

41: The AMPK construct of claim 39, wherein the first fluorophore is CFP and the second fluorophore is YFP; or wherein the first fluorophore is mseCFP11 (SEQ ID NO: 67) and the second fluorophore is cpVenus (SEQ ID NO: 55).

42: The AMPK construct of claim 35, wherein the 2 isoform is tagged with mseCFP11 (SEQ ID NO: 67) at the C-terminus, and either the 2 or the 1 isoform is tagged with cpVenus (SEQ ID NO: 55) at the C-terminus; or wherein the 2 isoform is tagged with CFP at the C-terminus, and either the 2 or the 1 isoform is tagged with YFP at the C-terminus.

43: The AMPK construct of claim 30, wherein at least one of the -subunit, -subunit and -subunit is a metazoan AMPK subunit.

44: The AMPK construct of claim 30, wherein the -subunit, -subunit, and -subunit are mammalian.

45: The AMPK construct of claim 44, wherein the a subunit is murine, rat, human, bovine or ovine; the subunit is murine, rat, human, bovine or ovine; and the -subunit is murine, rat, human, bovine or ovine.

46: The AMPK construct of claim 30, wherein the AMPK construct is comprised within a host cell.

47: The AMPK construct of claim 30, which is comprised in a kit for identifying the presence of an allosteric interactor of AMPK in a sample.

48: A method for identifying an AMPK allosteric interactor in a sample, the method comprising the steps of: (a) contacting the sample with a heterotrimeric AMPK construct comprising: an AMPK -subunit or a mutant or fragment thereof, an AMPK -subunit or a mutant or fragment thereof, an AMPK -subunit or a mutant or fragment thereof, and a fluorescent dye pair tagging at least one of said -subunit, -subunit, -subunit, mutants or fragments, said fluorescent dye pair being placed to allow detection of conformational changes within the AMPK construct; and (b) detecting, in the sample, a modification of the FRET fluorescence by fluorescence techniques; wherein the detecting of the modification of the FRET fluorescence in the sample indicates the presence of an AMPK allosteric interactor in the sample.

49: An ex vivo method of screening for an AMPK allosteric interactor, the method comprising the steps of: (a) providing a cell culture comprising cells expressing a heterotrimeric AMPK construct comprising: an AMPK -subunit or a mutant or fragment thereof, an AMPK -subunit or a mutant or fragment thereof, an AMPK -subunit or a mutant or fragment thereof, and a fluorescent dye pair tagging at least one of said -subunit, -subunit, -subunit, mutants or fragments, said fluorescent dye pair being placed to allow detection of conformational changes within the AMPK construct; (b) providing candidate AMPK allosteric interactor; (c) contacting the cells in the cell culture with said candidate AMPK allosteric interactor; and (d) detecting a modification of the FRET fluorescence in the cells by fluorescent techniques; wherein detecting of the modification of the FRET fluorescence in the cells indicates that the candidate AMPK allosteric interactor is an AMPK allosteric interactor.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0205] FIG. 1: Conformational change model showing operating mode of AMPfret sensors.

[0206] AMPfret sensors are constructed from an AMPK heterotrimer (consisting of -, -, and -subunits) with two additional GFP-derived fluorescent proteins (CFP, YFP) fused to different N- and C-termini of AMPK subunits. Binding of AMP or ADP to two CBS domains in the AMPK -subunit induces a conformational change which reduces the distance between the fluorophore couple. This increases fluorescence (or Foerster) resonance energy transfer (FRET) between the two fluorophores. Experimentally, when CFP is excited at 439 nm, FRET reduces direct CFP fluorescence emission at 476 nm, while energy transferred to YFP increases YFP fluorescence emission at 527 nm.

[0207] FIG. 2: Initial AMPfret constructs.

[0208] AMPfret A and C exhibit variation of FRET ratio upon AMP binding. Top: Schema showing structural organization of the sensors. CFP and YFP are respectively represented as hatched- and dotted-circles. (a) Fluorescence emission spectra of AMPfret constructs excited at 430 nm. Spectra show fluorescence peaks of CFP (476 nm) and YFP (527 nm), and their variation upon AMP binding (dotted line: 3 mM ATP, continuous line: 20 M AMP). (b) FRET variation of AMPfret constructs calculated from data above (hatched column: 3 mM ATP, dotted column: 20 M AMP) and autoradiograms of in vitro kinase activity assays with these constructs using acetyl-CoA carboxylase (ACC) as a substrate. Data correspond to meanSEM (AMPfret A: n=7; AMPfret C: n=10). Note: AMPfret constructs exhibit similar activity as native AMPK.

[0209] FIG. 3: Optimized AMPfret constructs.

[0210] Second generation of AMPfret constructs 1.1 and 2.1. based on constructs AMPfret C and A, respectively. Top: Schema showing structural organization of the sensors. Both optimized constructs contain mseCFP.sub.11/cpVenus as GFP-derived fluorescent couple instead of CFP/YFP. mseCFP.sub.11 and cpVenus are respectively represented as hatched- and checkered circles. AMPfret 2.1 - and -subunits also contain small deletions in their protein sequence to shorten C-terminal non-folded linker sequences (A.sub.551 and R.sub.552 in and K.sub.270, P.sub.271 and I.sub.272 in ). In addition, a putatively rigid helix (7 amino acids) was inserted between the -subunit C-terminus and CFP (see small box with curled lines). (a) Fluorescence emission spectra of AMPfret constructs excited at 430 nm. Spectra show fluorescence peaks of mseCFP.sub.11 (476 nm) and cpVenus (527 nm), and their variation upon AMP binding (dotted line: no AMP, continuous line: 20 M AMP). (b) FRET variation of AMPfret constructs calculated from data above (hatched column: no AMP, dotted column: 20 M AMP) and autoradiograms of in vitro kinase activity assays with these constructs using acetyl-CoA carboxylase (ACC) as a substrate. Data correspond to meanSEM (AMPfret 1.0: n=10; AMPfret 1.1: n=7); *=p<0.001 (significance assessed by a Student-Newman-Keuls test). Note: All AMPfret constructs exhibit similar activity as native AMPK. AMPfret 2.1 reveals improved FRET variation range as compared to AMPfret 1.1, providing proof of principle that optimization of FRET is possible.

[0211] FIG. 4: FRET response of AMPfret sensors correlates with the concentration of AMPK activator AMP.

[0212] AMP concentration dependence of the normalized FRET ratio of AMPfret sensors (a) AMPfret 1.1 and (b) AMPfret 2.1. The FRET ratio was calculated from fluorescence emission spectra excited at 430 nm. Data points correspond to meanSEM (n>3). Data were fitted with Sigma Plot 1.1 software to single site binding kinetics, yielding affinities of 1.8 M (AMPfret 1.1) and 1.5 M (AMPfret 2.1.). Note: AMPfret sensors are sensitive to AMP concentrations in a physiological range (0-20 M)

[0213] FIG. 5: FRET response of AMPfret sensors correlates with the concentration of AMPK activator ADP.

[0214] ADP concentration dependence of the normalized FRET ratio of AMPfret sensors (a) AMPfret 1.1 and (b) AMPfret 2.1. The FRET ratio was calculated from fluorescence emission spectra excited at 430 nm. Data points correspond to meanSEM (n>3). Data were fitted with Sigma Plot 1.1 software to single site binding kinetics, yielding affinities of 5 M (AMPfret 1.1) and 7.4 M (AMPfret 2.1). Note: AMPfret sensors are sensitive to ADP concentrations in a physiological range (0-50 M for free ADP).

[0215] FIG. 6: AMPfret sensors as in vitro tools to identify AMPK allosteric activators.

[0216] AMPfret 2.1 is incubated in absence (grey mesh bars) or in presence (black bars) of (a) 20 M AMP, (b) 50 M A-769662 or (c) 500 M Metformin. Structure and names of the molecules are given below the bars. Data correspond to meanSEM (AMP: n=7; A-769662: n=4; Metformin: n=4); *=p<0.001 (significance assessed by paired T-test).

[0217] FIG. 7: AMPfret sensors as cellular in vivo tools to identify AMPK allosteric activatorsHeLa cells

[0218] HeLa cells transfected with AMPFret 2.1 were exposed for 60 min to 1 mM AICAR. (a) Fluorescence emission spectra showing the increase of cpVenus peak (527 nm) over time; dotted black line: 0 min; dashed black line: 15 min and solid black line: 30 min. (b) Time course of the FRET signal. Normalized FRET ratio determined each 15 minutes (meanSEM; n=45; *=p<0.001 according to the performed Mann-Whitney Rank Sum Test). (c) AMPK activation at t=0 min and t=60 min. Phosphorylation of the AMPK substrate ACC as determined by immunoblotting (lower panel) and quantification of the resulting P-ACC/total ACC ratio (upper panel). P-ACC/total ACC ratios at t=0 and t=60 min are respectively represented as a white dotted bar and black dotted bar. Data correspond to meanSEM (n=3).

[0219] FIG. 8: AMPfret sensors as cellular in vivo tools to identify AMPK allosteric activators3T3-L1 cells

[0220] HeLa cells transfected with AMPFret 2.1 were exposed for 60 min to 1 mM AICAR. (a) Time course of the FRET signal. Normalized FRET ratio determined each 15 minutes (meanSEM; n=9). (b) Time course of AMPK activation. Phosphorylation of the AMPK substrate ACC as determined by immunoblotting (lower panel) and quantification of the resulting P-ACC/total ACC ratio (upper panel). Data correspond to meanSEM (n=3).

[0221] FIG. 9: Effect of 1 hour ischemia followed by 1 hour reperfusion on HepG2 cell followed by AMPfret.

[0222] AMPfret 2.1 normalized FRET ratio evolution during 1 h ischemia (light grey bar) and 1 h reperfusion (dark grey bar). Transfected HepG2 cells were cultured on a glass slide mountable onto the incubation flow-through chamber of our Leica TCS SP2 AOBS confocal microscope. At t=0, the cell was placed under ischemia-like conditions: hypoxic conditions (2% O2) and glucose-free medium at 37 C. Deprived medium was previously bubbled with N2 for at least 10 minutes before its addition onto the cells. After 1 hour of deprivation, started the 1 hour-reperfusion period with glucose-rich medium and 02 (21%). FRET values were record every minute from a single isolated cell using the Leica confocal software. The FRET ratio was followed by recording simultaneously mseCFP11 (476 nm) and cpVenus (527 nm) fluorescence emitted within 4 nm windows using two independent channels, under excitation set at 458 nm. FRET ratio was normalized to 1 at t=0.

[0223] FIG. 10: Strategy for optimizing the AMPK sensors according to the present invention. Starting from the most promising original constructs, FRET signal was optimized by mutations, deletions and addition of sequences.

[0224] The compounds and processes of the present invention will be better understood in connection with the following examples, which are intended as an illustration only and not limiting of the scope of the invention

EXAMPLE 1: AMPK CONSTRUCTS

[0225] AMPK Constructs and Protein Preparation

[0226] The .sub.2, .sub.2 and .sub.1 AMPK subunits tagged or not with fluorescent protein, were respectively cloned in the pACE, pDC and pDS vectors of the ACEMBL expression system (Bieniossek et al., Nat Meth 6:447) using SLIC (Li et al., Methods Mol Biol 852:51, 2012) and conventional cloning. Created vectors, containing a single subunit fluorescently tagged or not, were fused via their Lox-P site using the CRE-recombinase (EMBL Heidelberg): a single expression vector coding for a chimeric AMPK that contains two of its three subunits flanked with the mseCFP.sub.11/cpVenus fluorescent proteins pair (respectively variant of cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) at their termini was obtained. A decaHis-tag, cleavable by the TEV protease, was inserted at the N-terminus of the .sub.2 subunit in order to purify easily the heterotrimer.

[0227] BL21 (DE3) Star cells were transformed by electroporation and protein expression was carried overnight at 18 C. in autoinducing medium. Cells were collected by centrifugation at 6000 rpm for 20 min using a Beckman Coulter centrifuge (rotor JLA-8.1000) and wash with PBS. Cells were then suspended in lysis buffer (0.5 M sucrose, 30% glycerol, 50 mM Tris pH8, 100 mM NaCl, 2 mM MgCl2, 2 mM -mercaptoethanol, lysosyme 1 mg/mL, 20 mM imidazole, Complete EDTA free tablet (Roche), leupeptin, pepstatin). 200 U Benzonase were added to the suspension and it was gently stirred for 1 h in the cold room. Cells were then lysed by sonication using a MisonixSonicator 4000 (5 min total at 80%20 s ON/1 min OFF).

[0228] Cell-free extract, obtained by centrifugation at 20000 rpm for 80 min (rotor JS 25.50) was applied on Ni-NTA Superflow resin (Qiagen) pre-equilibrated with lysis buffer. Resin was washed using washing buffer (50 mM Tris pH 8, 100 mM NaCl, 20 mM imidazole, 2 mM MgCl2, 2 mM -mercaptoethanol) and high salt buffer (wash buffer+1 M NaCl). Proteins were eluted by applying elution buffer (wash buffer+400 mM Imidazole). Imidazole was removed through a overnight dialysis in buffer A (50 mM Tris pH8, 100 mM NaCl, 2 mM MgCl2, 2 mM -mercaptoethanol). Eluted proteins were passed over a 5 mL QXL column (GE Healthcare) in order to remove proteins bound to nucleic acids and non-stoichiometric AMPK complexes. Proteins were eluted using a gradient of buffer B (50 mM Tris pH8, 1 M NaCl, 2 mM MgCl2, 2 mM -mercaptoethanol). Finally chimeric AMPK heterotrimers were applied to a Superose 6 gel filtration column (GE Healthcare) pre-equilibrated with SEC buffer (50 mM Tris pH8, 200 mM NaCl, 2 mM MgCl2, 2 mM -mercaptoethanol, 5 mM spermidine). Spermidine diminished concentration dependent AMPK oligomer formation. After adding glycerol to a final concentration of 50%, the purified AMPK (untagged or AMPK 221WT) and AMPK heterotrimers of the invention (AMPK tagged hereafter AMPFret or AMPFret sensors) were stored at 20 C. for further experiments.

[0229] Finally, combinations of AMPK tagged with mseCFP.sub.11 and cpVenus on two of the termini of its 3 subunits were created in order to identify constructs that show FRET signal variation upon AMP binding (hereafter termed AMPFret).

TABLE-US-00002 TABLE 1 Overview of the AMPfret constructs containing two fluorescent permutated at the N- and C-termini of the three AMPK subunits. protein tags AM Pfret construct Vector name and composition AMPK 221 pACEMBL .sub.2-.sub.2-.sub.1 AMPfret A pACEMBL .sub.2-CFP.sub.-.sub.2-YFP.sub.-.sub.1 AMPfret C pACEMBL .sub.2-CFP.sub.-.sub.2-.sub.1-YFP Abbreviations: pACEMBL, plasmid resulting from the Cre-LoxP fusion of vectors pACE, pDC and pDS of the MutliColi expression system; CFP, Cyan Fluorescent Protein; YFP, Yellow Fluorescent Protein; 2, 2, 1, AMPK subunits.

[0230] Characterization of AMPFret Sensors In Vitro

[0231] ATP containing buffers were always freshly prepared to limit AMP contamination. Aqueous solutions of nucleotides (adenine nucleotides, NAD) were analyzed by HPLC (stationary phase: Polaris C18/mobile phase: 60% CH.sub.3CN 40% H.sub.20) to evaluate spontaneous ATP and ADP hydrolysis and contaminations.

[0232] Enzymatic assay: AMPK 221WT and AMPfret constructs (3 pmol) were activated by incubation with purified CamKK (1 pmol) for 20 min at 30 C. in kinase buffer (200 M ATP, 40 M AMP, 5 mM MgCl2, 1 mM DTT and 10 mM Hepes pH 7.4). Purified ACC fragment targeted by AMPK (200 pmol) was then incubated for 20 min at 37 C. in presence or absence of pre-activated AMPK 221WT or AMPFret sensor in kinase buffer containing [-.sup.32P]ATP. Reaction mixtures were then load on SDS-PAGE gel, P-ACC signals were revealed using a Typhoon and activities were evaluated with ImageJ.

[0233] FRET assay: FRET signal variation in presence of different compounds (nucleotides, chemicals, ions) was measured using a fluorimeter (Photon Technology International). AMPfret constructs (20 pmol) were incubated in a quartz cuvette in a final volume of 150 L (spectro buffer: 50 mM Tris pH8, 200 mM NaCl, 5 mM MgCl.sub.2, 2 mM -mercaptoethanol). Effects of nucleotides and others compounds (previously prepared in the spectro buffer) on the FRET ratio given by AMPfret sensor was determined by comparing FRET ratio (peak value at 527 nm/peak value at 476 nm) in presence or absence of the compounds. Excitation wavelength was set to 430 nm, and emission spectra were recorded from 450 to 600 nm with an integration time of 0.2 s. Mg.sup.2+ effect on FRET was investigated in spectro buffer without Mg.sup.2+.

[0234] The two constructs, AMPFret A (.sub.2-CFP-.sub.2-YFP-.sub.1; CFP tagged at the 2 C-terminus, and YFP at the 1 C-terminus) and AMPFret C (.sub.2-CFP-.sub.2-.sub.1-YFP; CFP tagged at the 2 C-terminus, and YFP at the 1 C-terminus) showed both a significant difference in their FRET signal (.sup.10%) depending on the presence of AMP or ADP (FIG. 2). It appears that, during allosteric activation, the -subunit C-terminus approaches the C-termini of the 3- and -subunits.

EXAMPLE 2: OPTIMIZED AMPK CONSTRUCTS

[0235] Constructs were optimized to achieve a superior FRET signal amplitude. The construct AMPfret 1.1 is based on AMPfret C, containing full length AMPK subunits 2, 2 and 1. The -subunit is tagged with mseCFP11 at its C-terminus and the -subunit is tagged at its C-terminus with cpVenus; the -subunit remains untagged. The construct AMPfret 2.1 is based on AMPfret A, where CFP/YFP were exchanged for the same different fluorophore pair as AMPfret 1.1. In addition, the sequence of the construct was modified. First, small truncations based on the crystal structure (PDB 2Y94) and secondary structure prediction (nps@consensus (ucbl)) were inserted via PCR and self SLIC between the N-terminus of fluorescent protein tags and the C-terminus of the tagged AMPK subunits. Such shortening of the sequence between AMPK core and tag may remove flexibility other than the conformational change induced by AMP. Second, a short insert supposed to fold into a rigid -helix (Sivaramakrishnan et al., PNAS 105:13356, 2008 and 108:20467, 2011) was inserted between the 2 C-terminus and the CFP N-terminus to rigidify the AMPK backbone of the invention and to stabilize the CFP tag in a given position relative to AMPK.

[0236] This engineering comprised the following mutations. The last 2 C-terminal amino acids (AR) of 2 and the first 3 N-terminal amino acids (MSK) of mseCFP11 were removed and the new termini linked via 8 amino acids insert supposed to fold into an -helix (EEEEKKKK, SEQ ID No.1). Further, the last C-terminal (non-folded) 3 amino acids (KPI) of 2 were also removed, and directly fused to the N-terminus of YFP. Since 2 amino acids resulting from the restriction site previously used were also removed by the SLIC technique, this yielded a construct lacking in total 5 amino acids between the 2-subunit and YFP. The optimized AMPfret sensor showed an almost 100% increased FRET ratio (FIG. 3).

[0237] The optimized AMPfret sensors allow titration of the allosteric AMPK-activators AMP and ADP, confirming that both induce conformational changes in the AMPK heterotrimer. The affinity (Kd) for AMP and ADP could be determined as 1.5 M and 7.4 M, respectively.

[0238] The AMPfret sensors thus represent a pioneering powerful and easy-to-use tool to decipher the activation mechanisms of AMPK. They contain full length AMPK heterotrimer that behaves the same way as native AMPK WT as judged by (i) its kinase activity after phosphorylation via CamKK and allosteric activation by AMP and (ii) its affinities for adenine nucleotides. The AMPfret FRET signal is directly dependent of the AMP concentration; in a physiological range (1-10 M) it shows almost linear relationship (FIGS. 4 and 5).

EXAMPLE 3: IN VITRO INTERACTION OF OPTIMIZED AMPFRET WITH ALLOSTERIC ACTIVATORS

[0239] The optimized AMPfret sensors not only translate the adenylate-dependent movements of the AMPK heterotrimer into a FRET signal, which are triggered by adenylate binding to specific sites at the -subunit. Their readout also reports conformational changes of other, pharmacological direct AMPK activators such as the compound A-769662 (FIG. 6). This molecule interacts with the -subunit, but clearly induces a FRET signal comparable to AMP, even if the triggering conformational change may be of different nature according the different binding mode.

[0240] Metformin, a widely used anti-diabetes drug, which was postulated to directly interact with -subunit (Zhang et al., Mol Cell Biochem 368:69, 2012) did not induce any FRET variation emission of AMPfret (FIG. 6). This absence of conformational changes in vitro supports the generally accepted indirect mode of action, whereby this drug inhibits mitochondrial respiration and increases the AMP/ATP ratio, thus activating AMPK by the canonical AMP binding mechanism at the -subunit.

[0241] Taken together, AMPfret appears as a valuable and accurate tool for in vitro applications, notably screening for AMPK interactors.

EXAMPLE 4: EX VIVO EXPERIMENTS WITH THE OPTIMIZED AMPFRET CONSTRUCTS

[0242] For cellular ex vivo experiments, subunits of the optimized construct AMPfret 2.1 were cloned in the vectors (pACEMam2, pMDS and pMDK) of the MultiMam expression system. Created plasmids were fused via their Lox-P site to yield to a single mammalian expression vector coding for the sensor AMPFret 2.1 according to well-known techniques to the skilled man in the art.

[0243] 3T3-L1 and HeLa cells were cultivated in glucose containing DMEM (4.5 g/L) supplemented with SVF, glutamine, penicillin and streptomycin. Once cells reached around 80% confluence, medium was replaced by OptiMEM (Lifetechnologies) and AMPfret 2.1 coding plasmid was transfected using Lipofectamine2000 (Lifetechnologies). After 5 h, OptiMEM was exchanged by complete DMEM and cells grew for >36 h until their observations under the confocal microscope.

[0244] 3T3-L1 or HeLa cells, cultivated in 8 wells LabTek cover glass plates (Nunc), were observed with a Leica TCS SP2 AOBS confocal microscope. LabTek plates were placed in an incubation chamber in which the temperature and O.sub.2 concentration were maintained at 37 C. and 21%, respectively. Without moving the Labtek, 200 L medium was replaced by the same volume of complete medium containing 2 mM AICAR (1 mM final). Excitation wavelength was set to 458 nm and emission spectra showing FRET signal were monitored through A scans from 463 nm to 600 nm every 15 min. ROI (region of interest) were drawn in order to cover entire cells. FRET ratio variations were calculated from those measured emission spectra.

[0245] Under the microscope, cells were treated with 1 mM AICAR (AMPK allosteric activator) to visualize the allosteric activation of AMPK through the FRET signal of AMPfret 2.1 and hence validate its use for ex vivo applications.

[0246] AMPfret 2.1 was excited using a 458 nm laser and emission spectra showing FRET signal were monitored through A scans from 463 nm to 600 nm every 15 min. The AMPfret 2.1 FRET signal increased with time upon AICAR addition, suggesting that AMPfret 2.1 can monitor allosteric activation of AMPK in cells (FIGS. 7 and 8).

[0247] More than half-maximal response was already reached after 15 min of treatment, and the maximal effect reached after about 30 minutes.

[0248] The activation kinetics of AMPK upon AICAR addition was independently verified by Western blotting for the AMPK-specific phosphosite in acetyl-CoA carboxylase (ACC; widely used as reporter for AMPK activity) in 3T3-L1 cells.

[0249] All the results presented above show that the AMPfret sensor can be used as a suitable tool for cellular in vivo applications.

EXAMPLE 5: AMPFRET 2.1 DURING ISCHEMIA-REPERFUSION IN A HEPG2 SINGLE CELL

[0250] Using an incubation flow-through chamber fitted to the confocal microscope which permits to control temperature as well as 02 concentration, HepG2 cells were placed under ischemia-like conditions, comprising hypoxic conditions (2% 02) and glucose free medium at 37 C. ATP pools may not be affected when hypoxia is applied in a high nutrient containing medium since cells can adapt to hypoxia by switching their energy metabolism through anaerobic pathways to compensate for aerobic ATP production. The deprivation period was followed by 1 hour of reperfusion with complete medium and O.sub.2 (21%). During the 2 hours of the ischemia-reperfusion protocol, the FRET ratio was monitored every minute by recording simultaneously mseCFP11 and cpVenus fluorescence emitted within 4 nm windows (corresponding to fluorescence emission) using two independent channels. Images were collected and processed using ImageJ in order to i) remove eventual background fluorescence and ii) isolate individual cells from acquired pictures. Then, the fluorescence intensities were extracted from single cell images using Volocity. Thus, the effect of ischemia-reperfusion on the AMPfret 2.1 signal in single cells was analyzed (FIG. 9).

[0251] During ischemia in HepG2 cells, the FRET signal did not vary. Changes in AMP/ATP ratio under such conditions were proposed to happen in the liver and AMPK becomes activated, but a recent study suggested that AMPK was activated during ischemia through adenylate-independent pathways. FIG. 9 shows results of a single cell.

[0252] During reperfusion of HepG2 cells, the FRET signal increases over the first 30 minutes indicating increased AMP and ADP concentrations. Subsequently, the FRET ratio remained at unchanged high values, suggesting that elevated AMP and ADP concentrations were maintained. In fact, AMPfret should revert the FRET ratio as soon as AMP and ADP levels drop again. These results suggest that in HepG2 cells, reperfusion represented a more drastic energy stress than ischemia regarding adenylate concentrations and AMPK allosteric activation.

[0253] Through these experiments, using AMPfret 2.1 in HepG2 cells, we did not detect any FRET signal changes during ischemia, suggesting that AMP and ADP concentrations remained unchanged. However, we showed an increase of AMPfret FRET signal during reperfusion, suggesting an elevation of intracellular AMP and ADP and allosteric activation of AMPK.

[0254] These experiments achieved in living cells using AMPfret 2.1 show that AMPfret 2.1 was properly transfected and its fluorescence monitored over time. These results show that AMPfret 2.1 provides a readout of AMP/ZMP concentrations and AMPK allosteric activation by reporting the related conformational changes. Experiments involving ischemia-reperfusion showed that AMPfret 2.1 can monitor endogenous changes of adenylates and AMPK allosteric activation over time.

[0255] Monitoring of transient events related to AMPK allosteric activation is promising to decipher or unravel new aspects of its regulation.