Fasting, autophagy and the mTOR switch

mTORC1 is mechanistic target of rapamycin complex 1 — a kinase that sits on the outer face of the lysosome and decides whether the cell is allowed to build. Amino acids arriving through the Rag GTPases, and insulin arriving through PI3K and Akt, turn it on. When it's on it phosphorylates S6K1 and 4E-BP1, translation rises, and it phosphorylates ULK1 at a residue that keeps autophagy off. AMPK is AMP-activated protein kinase, a heterotrimeric fuel gauge that listens to the AMP/ATP ratio. When that ratio rises, AMPK phosphorylates TSC2 and Raptor, mTORC1 turns down, and AMPK phosphorylates ULK1 at a different residue that starts autophagy. Fed: the cell builds. Fasted: the cell inventories and recycles. One switch, two states, both useful. Treating either state as a personality is how a kinase cascade becomes a lifestyle argument. Yoshinori Ohsumi's yeast genetics made the recycle side a named, assayable programme. Resistance training still needs the build side. A week that uses both on purpose is a protocol. A week that only clocks one of them is a week with a missing shift.

In short. Eat, and the cell builds. Fast, and it starts recycling worn parts. Both jobs matter; they are two states of one switch.

Yoshinori Ohsumi's 2016 Nobel Prize in Physiology or Medicine made autophagy a named, assayable programme rather than a rumour about hungry cells. Working in Saccharomyces cerevisiae in the 1990s, Tsukada and Ohsumi isolated mutants that could not recycle their own cytoplasm when nitrogen was withdrawn, and then cloned the ATG genes those mutants had broken. Mammalian biology inherited the list. Mizushima, Yoshimori and Ohsumi put LC3, the mammalian Atg8 homolog, on autophagosome membranes in 2000; you can blot LC3-II, the phosphatidylethanolamine-conjugated form, and you can count puncta with an antibody. Levine put Beclin-1, the mammalian Atg6 homolog, on tumour suppression in 1999. Klionsky's later guidelines, a paper the size of a short book, told the field how not to lie to itself with a snapshot. A mood isn't an assay, and a brand isn't a gene list. A Western blot of LC3-II, run with a lysosomal-protease trap so that flux is visible rather than inferred, is a measurement. Sixteen empty hours can be a useful timetable. They are not, by themselves, that blot.

In short. A 2016 Nobel prize made cell recycling a named programme you can measure on a blot, not a feeling or a brand.

We stock lyophilised β-NAD+, 1000 mg, ≥98% by HPLC, because fasting and caloric restriction raise the NAD+/NADH ratio and give sirtuins something to spend, and because that cofactor is the molecule those papers actually weigh into a tube. Molecular weight 663.43. Formula C21H27N7O14P2. CAS 53-84-9. The neighbouring sirtuin essay is the deacylase family; the neighbouring NAD+ essay is the pool as pool. This essay is the mTOR–AMPK–ULK1 switch those redox changes sit beside. We stock MOTS-c because it is a 16-mer a mitochondrion translated from its own 12S rRNA — Lee, Kim, Cohen, Cell Metabolism, 2015 — and it talks to AMPK, one door from this cascade. Neighbourhood isn't identity. A dinucleotide, a mitochondrial peptide and a meal schedule share an energy campus and don't share a mechanism. The 1000 mg cake isn't an intermittent fast; intact extracellular NAD+ is a charged molecule of 663 daltons and a poor passenger across a plasma membrane. Write the vial as a feeding window and you've mixed a cofactor with a timetable.

In short. The catalogue holds the cofactor fasting also moves, and a mitochondrial peptide next door. They are neighbours, not a diet.

What follows is the physiology rather than a protocol. Named complexes, named sensors, named residues, named papers, and a legal class stated once at the close. Sixteen-to-eight is a convenient eating window, not a constant of nature; how long is long enough for AMP/ATP to rise and for mTORC1 to quiet depends on the last meal, on liver glycogen, and on whether you're walking around or asleep. Panda's time-restricted eating work is the same switch, aligned to the clock, and it isn't a separate physics. Resistance training still wants mTORC1 on, in the fibres you loaded, with leucine in the blood. People who live in permanent build and people who live in permanent recycle both miss half the map. DiRECT-style remission of type 2 is, among other things, a long recycle shift applied to a fatty liver. The GH axis and IGF-1 LR3 in this catalogue are how you study the build side in a dish. NAD+ is how you study the redox budget that fasting also moves. Different doors. The rest of this piece is the biochemistry you'd want before designing the assay you actually have the controls for.

In short. This page is the switch, the papers, and why a popular sixteen-hour window is a timetable, not a law of nature.

mTORC1 is a complex, and the address is part of the decision. The kinase subunit is mTOR, a phosphatidylinositol 3-kinase-related kinase of about 289 kilodaltons. Raptor, regulatory-associated protein of mTOR, is the substrate-recruiting partner that defines the complex as complex 1 rather than complex 2. mLST8 sits on the kinase lobe. PRAS40 and Deptor are the inhibitory subunits that come off as the complex activates. Sabatini's laboratory, then others, showed that the active complex is recruited to the cytosolic face of the lysosome when amino acids are present. The lysosome isn't a bin in this picture. It's the surface on which nutrient sensing, Rag GTPases, Ragulator and Rheb are allowed to meet the kinase. Rapamycin, with FKBP12, occupies an allosteric site and was the pharmacological reason the kinase has its name; mTORC2 is largely rapamycin-resistant on a short clock, which is why writing 'mTOR' without the C1 hasn't yet chosen an experiment. The complex that builds translation and brakes ULK1 is mTORC1, on the lysosome, listening.

In short. The build kinase sits on the cell's recycling depot and listens there. Location is part of the decision.

Rag GTPases are how amino acids speak to that surface. Mammals have RagA or RagB paired with RagC or RagD, heterodimers tethered to the lysosome by the Ragulator complex, LAMTOR1 through LAMTOR5. When amino acids are plentiful, RagA/B is GTP-bound and RagC/D is GDP-bound; that configuration binds Raptor and docks mTORC1 at the lysosome, where Rheb-GTP can then activate the kinase. When amino acids fall, the nucleotide state flips and mTORC1 leaves the surface. Sancak, Sabatini, Science 2008, put the Rags on raptor. Sancak, Bar-Peled, Sabatini, Cell 2010, put Ragulator on the targeting. The v-ATPase, SLC38A9 as a putative arginine transceptor, and a set of other lysosomal proteins sit in the same neighbourhood and are still being mapped. What you need is simpler and ruder: amino-acid sensing isn't insulin sensing. The Rag axis is the amino-acid inbox. A dish from which you've withdrawn serum but left glutamine and leucine isn't an amino-acid experiment. A dish from which you've withdrawn the amino acids is.

In short. Amino acids talk to the build kinase through a pair of small G proteins on that depot's surface.

Sestrin2 is the leucine sensor the field can actually name, and it's why a whey shake is a molecular event rather than a gym habit. Wolfson, Chantranupong, Sabatini, Science 2016: Sestrin2 binds leucine; when leucine is absent, Sestrin2 inhibits GATOR2, GATOR1 (a GAP for RagA/B) keeps RagA/B GDP-bound, and mTORC1 stays off the lysosome. When leucine binds, Sestrin2 lets GATOR2 inhibit GATOR1, RagA/B go GTP-bound, and mTORC1 can dock. CASTOR1 is the arginine sensor on the same GATOR2 node. SAMTOR is the S-adenosylmethionine sensor, a methionine-status microphone. Three amino acids, three proteins, one Rag cycle. Leucine is the loudest of the three for skeletal muscle, which is why the protein-and-training essay next door will not let you skip the dose at the meal after the session. Sestrin2 isn't a brand. It's a fold that holds a leucine. Writing 'protein' as if all amino acids were equal at this node is how a whey paper becomes a brochure. Name the residue. Name the sensor.

In short. A protein called sestrin2 is how the cell tastes leucine. Without that amino acid the build kinase stays off.

The growth-factor inbox is a different cable into the same complex, and confusing the two inboxes is how a fasting essay becomes an insulin essay by accident. Insulin or IGF-1 occupies its receptor tyrosine kinase; IRS proteins recruit PI3K; PIP3 recruits Akt; Akt phosphorylates TSC2 and thereby inhibits the TSC complex, which is the GAP for Rheb. Rheb stays GTP-bound on the lysosome and can activate mTORC1 that has already been delivered by the Rags. Amino acids without insulin: Rags can dock the complex, Rheb may still be GDP-bound, activation is weak. Insulin without amino acids: Rheb is GTP-bound, the complex isn't at the lysosome, activation is weak. Both inboxes, both required for a full on-signal, is the Sabatini sentence you have to keep honest. TSC1/TSC2 is also where AMPK leans when energy falls: AMPK phosphorylates TSC2 at a different site and favours the GAP-on state, Rheb goes GDP, mTORC1 quiets. One hamartin–tuberin complex, two kinases writing on it in opposite directions. That's the switch, drawn as biochemistry rather than as a personality.

In short. Insulin is the other inbox. It tells the same kinase, through a different path, that the body is in a fed state.

Once mTORC1 is on, the jobs are named and they aren't mysterious. S6K1 is phosphorylated at Thr389, a blot a thousand papers have run; S6K1 then phosphorylates ribosomal protein S6 and several translation factors, and it also feeds back on IRS-1, which is why chronic mTORC1 tone can blunt insulin signalling — a real physiology, not a curiosity. 4E-BP1 is phosphorylated at several sites; the hypophosphorylated form binds eIF4E and holds cap-dependent translation down; phosphorylation releases eIF4E, the cap-binding protein, and 5'-TOP mRNAs for ribosomal proteins are among the messages that then run. ULK1 is phosphorylated at Ser757 (human numbering; Ser758 in some tables), a mark that prevents AMPK from activating ULK1, which is how the build state keeps autophagy off. Lipin 1, TFEB, a set of other substrates: the complex writes more than translation, including a brake on lysosomal biogenesis when the cell is already building. The point of the census is to stop writing 'mTOR is anabolic' as if that were a mechanism. Name the substrate. Name the residue. Then the recycle side has something to oppose.

In short. Once on, the kinase turns protein-making up and recycling down. Those two jobs are the same decision, written twice.

The fed state is therefore a coincidence of two signals, and a meal that supplies both is the native protocol mTORC1 already expected. Dietary protein supplies leucine, arginine, methionine; carbohydrate and, to a lesser extent, protein itself supply the insulin. A protein-rich meal after a loaded session is the textbook way to turn mTORC1 on in skeletal muscle: amino acids in the blood, insulin not at zero, mechanical tension having already primed the fibre through a set of kinases we won't pretend to own. Wolf, Phillips, Tipton, the muscle-protein-synthesis literature of the last twenty-five years, put numbers on the leucine threshold and on the dose at the meal, and those numbers are why 'just fast' is an incomplete instruction for anyone who also wants to keep the muscle they have. The liver sees the same meal differently. Hepatic mTORC1 is a lipogenic and a translational microphone; a constant grazing pattern that never lets amino acids or insulin fall is a liver that never quite leaves build. That isn't a moral. It's a kinase that wasn't given an off-shift.

In short. Amino acids and insulin together turn building on. Either alone is a weaker signal than the pair.

Translation is the first job the on-state actually pays for, and it's the job a muscle paper is usually trying to measure. eIF4E, freed from 4E-BP1, binds the 5' cap. eIF4G and eIF4A join. The 43S pre-initiation complex scans. S6K1-dependent phosphorylation of eIF4B and of PDCD4, among other substrates, leans the helicase work toward messages that were waiting. The 5'-TOP mRNAs — ribosomal proteins, translation factors, a set of growth-associated messages with a run of pyrimidines at the cap — are disproportionately mTORC1 clients, which is how a kinase on a lysosome changes the composition of the ribosome pool over hours. Puromycin incorporation, SUnSET, a flooding-dose phenylalanine method in a rodent, a primed constant infusion of a labelled amino acid in a human: those are the assays that earn the phrase muscle protein synthesis. A phospho-S6K blot is a proxy. A hypertrophy measurement on a tape is a much later invoice. Confusing the three is how a whey advertisement and a lab write-up swap clothes. The kinase turned translation up. Whether the fibre grew is a different figure.

In short. The first thing the on-switch does is let the cell make more protein from existing messages.

Resistance training still needs that build side, which is the sentence a fasting timetable keeps trying to edit out. Mechanical loading of a fibre, through a still-argued mix of phosphatidic acid, TSC2 localisation and integrin-adjacent kinases, makes mTORC1 easier to turn on in that fibre when the amino acids arrive. The session isn't the leucine. The session is why the leucine has somewhere useful to write. A week of empty hours with no loaded bars is a week in which autophagy and fat oxidation may well be running, and in which myofibrillar protein is on the menu along with the damaged mitochondria you actually wanted to recycle. Protein synthesis in muscle is a several-hour window after the session, leucine-sensitive, insulin-permissive, sleep-adjacent. Skipping the meal because a timetable said the window was shut is how a recycle protocol becomes a sarcopenia protocol in slow motion. The protein-and-training essay next door is the dose table. This essay's job is smaller: the same kinase that you quieted overnight is the kinase you need on after the lift. Both states. Same week.

In short. Muscle that you loaded still needs the build switch on, with leucine in the blood, after the session.

Permanent build is the other missing shift, and it's the more common one in the animals this literature actually worries about. Constant grazing, a late large carbohydrate load, amino acids never far from the sestrin2 pocket, insulin never far from IRS: mTORC1 spends the day on, ULK1 spends the day phosphorylated at the inhibitory residue, autophagy runs as a trickle. A hepatocyte in that state keeps writing lipid and VLDL; Taylor's twin-cycle work, in the neighbouring liver-fat essay, is among other things a picture of a liver that wasn't given an empty hour. Overnight fasting is the native off-shift most humans already had, before a fridge and a night kitchen took it away. That overnight off-shift isn't sixteen hours, and it isn't a protocol. It's the reason a breakfast that breaks a real fast and a breakfast that continues last night's grazing are different experiments at this node. People who never empty, and people who never load, fail in opposite directions. The week is the unit that can hold both failures at once, which is why the unit is the week and not the meal.

In short. Grazing all day is a week with no empty hours. The recycle shift never clocks on.

AMPK is the other half of the switch, and it's older than the brand that later borrowed it. A heterotrimeric serine/threonine kinase: a catalytic α subunit, a scaffolding β subunit that binds glycogen, a γ subunit that binds AMP, ADP and ATP at Bateman domains. Grahame Hardie's laboratory spent decades on this enzyme; the yeast Snf1 kinase is the orthologue a geneticist already knew. AMP binding to γ allosterically activates the kinase and makes the activating phosphorylation at α-Thr172 more resistant to phosphatases. LKB1 (STK11) is the principal upstream kinase in most tissues, constitutive; CaMKK2 can substitute when calcium rises, which is one reason a contracting myocyte can turn AMPK on before the adenine-nucleotide ratio has fully moved. The ratio that matters is AMP/ATP, not a mood about energy. A cell with millimolar ATP and micromolar AMP is a fed cell at this microphone. A cell whose adenylate kinase has started to mint AMP from ADP is a cell AMPK can hear. Exercise mints that AMP in muscle on a minutes-scale. A fast mints it more slowly, as glycogen falls and the liver begins to notice.

In short. A fuel-gauge kinase listens to how much AMP sits next to ATP. When energy is short, it turns on.

The two marks on ULK1 are the cleanest picture of the switch as a switch. Kim, Guan, Nature Cell Biology 2011: AMPK phosphorylates ULK1 at Ser317 and Ser777 (mouse numbering; human tables shift by a residue), and those marks activate the kinase that starts autophagosome formation. mTORC1 phosphorylates ULK1 at Ser757 and that mark blocks the AMPK–ULK1 interaction. When mTORC1 is on, AMPK can be shouting and ULK1 still will not start the programme. When mTORC1 is off, AMPK's marks can land. Egan, Shaw, Science 2011, put the same AMPK–ULK1 connection on mitophagy, the specialised recycling of mitochondria. AMPK also phosphorylates TSC2, favouring Rheb in the GDP state, and phosphorylates Raptor, both of which turn mTORC1 down from the other side. So the fuel gauge does two useful things at once: it takes the brake off ULK1, and it leans on the kinase that was applying the brake. One enzyme, two substrates, a coherent off-shift. Writing AMPK as 'the fasting kinase' without naming ULK1 and TSC2 is how a heterotrimeric enzyme becomes a poster.

In short. That fuel gauge turns the build kinase down and the recycling starter on, by marking different proteins.

AMPK has jobs that aren't autophagy, and a fasting essay that skips them hasn't yet described a fast. Acetyl-CoA carboxylase 1 and 2 are phosphorylated at the AMPK sites; malonyl-CoA falls; carnitine palmitoyltransferase 1 is released from inhibition; fatty-acid oxidation rises. That's the fat-burning sentence, and it's a blot (phospho-ACC Ser79) as much as it is a respiratory quotient. PGC-1α is phosphorylated and, in some tissues, transcriptionally induced; nuclear genes for respiratory subunits and for TFAM get written; the mitochondrial census can rise over days. TFEB and the CLEAR network, partly through the mTORC1-off state, write lysosomal and autophagic genes, so the recycle programme is both an acute ULK1 event and a slower transcriptional one. GLUT4 trafficking in muscle, a glucose-uptake invoice, is an AMPK client during contraction. None of that requires a personality theory of willpower. It requires a rising AMP/ATP ratio, an intact LKB1, and a tissue that still has the enzyme. A late-passage dish with a tired AMPK is a poor model of a morning walk.

In short. The same gauge also tells fat-burning enzymes to run and asks mitochondria to increase their census.

A fast raises AMP/ATP as liver glycogen runs down and as the periphery begins to oxidise fat, and the timescale is hours, not a ringtone at sixteen. Overnight, in a human who ate dinner at a reasonable hour, hepatic glycogen is already falling; glucagon is up; insulin is down; AMPK tone in liver and in muscle is a real, graded thing by morning. A sixteen-hour window, if the last meal was at 20:00 and breakfast is at 12:00, is that overnight plus a late morning. It isn't a different physics from the overnight. It's more of the same physics, with more time for glycogen to fall and for fatty acids to rise. A morning walk, or a training session in the empty hours, mints AMP in muscle on a minutes-scale that a sofa will not match. Treating sixteen as a constant of nature is how people skip dinner and call it a protocol. Treating any empty hour as equivalent to a loaded AMPK session is how a timetable replaces a nucleotide ratio. Measure the ratio, or admit you measured the clock.

In short. A fast raises that gauge as liver starch runs down. A hard training session can raise it faster in muscle.

Autophagy, in the cell, is a kinase cascade with names Ohsumi's yeast genetics actually gave us. Tsukada and Ohsumi, FEBS Letters 1993: mutants that accumulated un-recycled cytoplasm under nitrogen starvation, the genetic start. The ATG genes that followed — more than thirty in yeast, a mammalian set that is smaller and more tissue-specific — encode the machinery that nucleates a double membrane, grows it around cargo, closes it, and delivers it to the vacuole, which in a mammal is the lysosome. Yoshimori, Mizushima, Ohsumi, EMBO Journal 2000: LC3 localises to autophagosome membranes after processing. Levine, Nature 1999: Beclin-1, induction of autophagy, inhibition of tumorigenesis. The 2016 Nobel citation was for the mechanisms of autophagy, not for a diet. Mammalian physiology later added a meal schedule as one way to lean on the programme. You can't 'hack' ULK1 with a timetable. You can withdraw amino acids, raise AMP/ATP, let insulin fall, and lift the brake mTORC1 had parked on ULK1. How long is long enough is an empirical question in the tissue you have, not a number a window invented.

In short. Yeast genetics named the recycling genes. A 2016 Nobel made that list a mammalian programme you can assay.

ULK1 is unc-51-like kinase 1, the mammalian Atg1, and it's the starter we'll keep naming because the rest of the sentence skips it. The ULK1 complex is ULK1 (or ULK2 in some tissues), ATG13, FIP200/RB1CC1, and ATG101. When the inhibitory mTORC1 mark is off and the activating AMPK marks are on, ULK1 phosphorylates components of the class-III PI3K complex and of its own neighbourhood, and autophagosome nucleation begins. ATG13 and FIP200 aren't decoration; they're how the kinase is held at the right membrane and how the signal is scaffolded. ULK1 also phosphorylates Beclin-1, which is one of the ways the starter talks to the lipid kinase that comes next. If you report 'autophagy' and can't say whether ULK1 was phosphorylated at Ser757 or at the AMPK sites, you've reported a blur. The two marks move in opposite directions in a genuine fast. They can move together, or fail to move, in a dish that was merely starved of serum. Write the residue. The starter is a kinase. Kinases have substrates and they have phosphosites. Moods don't.

In short. A starter kinase opens the recycling programme. The build switch marks it one way; the fuel gauge marks it another.

Beclin-1 is the mammalian Atg6 homolog, and it's the partner that brings the lipid kinase to the nucleation site. Beclin-1, VPS34 (the class-III PI3K), VPS15, and ATG14L form PI3K complex I, which makes phosphatidylinositol 3-phosphate at a specialised domain of the endoplasmic reticulum sometimes called the omegasome. PI3P then recruits WIPI proteins and the rest of the machinery that grows the isolation membrane. Bcl-2 binds Beclin-1 and holds this complex quiet; JNK phosphorylation of Bcl-2, among other inputs, can release Beclin-1 when the cell is stressed. Levine's 1999 Nature paper put Beclin-1 on autophagy and on tumour suppression, a dual job that later cancer literature spent a generation arguing about, because a programme that recycles damaged mitochondria can be a tumour suppressor and a programme that helps a tumour survive a nutrient-poor core can be the opposite. For a fasting essay the useful sentence is smaller: Beclin-1 isn't a mood. It's a named subunit of a lipid-kinase complex you can immunoprecipitate. If your blot can't find it, you aren't yet doing nucleation.

In short. A partner protein then helps draw a membrane around the cargo. That partner has a name, Beclin-1.

LC3 is microtubule-associated protein 1 light chain 3, MAP1LC3, and the I-to-II conversion is the assay the field actually runs. Pro-LC3 is cleaved by ATG4 to LC3-I, a soluble form. LC3-I is then conjugated to phosphatidylethanolamine by a ubiquitin-like cascade: ATG7 as the E1, ATG3 as the E2, the ATG12–ATG5–ATG16L1 conjugate as the E3-like ligase. The product, LC3-II, is membrane-bound and runs faster on an SDS–PAGE gel, which is why a Western blot can tell the two forms apart. You can blot LC3-II. That sentence is the whole of the 'not a mood' argument, and it has to be qualified immediately: LC3-II accumulation can mean more autophagosomes being made, or fewer being consumed by lysosomes, or both. A snapshot without a lysosomal-protease inhibitor, or without bafilomycin A1 or chloroquine to trap the last step, is a number you can't interpret. Klionsky's guidelines exist because that ambiguity was costing people papers. Run the trap. Then the rise in LC3-II is flux, and flux is the programme.

In short. A small protein called LC3 is stitched onto the new membrane. You can see the stitched form on a blot.

The sealed autophagosome fuses with a lysosome, hydrolases meet cargo, and the products — amino acids, fatty acids, nucleosides — return to the cytosol. p62/SQSTM1 is a cargo adaptor that binds ubiquitinated proteins and binds LC3; it is itself degraded in the autolysosome, so a falling p62 with a rising LC3-II flux is the two-blot signature you'd want from a colleague. NBR1, NDP52, OPTN are other adaptors, more specialised, more often named in mitophagy and in xenophagy. Fusion machinery — STX17, SNAP29, VAMP8, a SNARE set — is how the two vesicles actually become one. Lysosomal acidification, which bafilomycin A1 blocks at the v-ATPase, is required for the hydrolases to work and for LC3-II to be turned over. Chloroquine raises lysosomal pH by a different route and is the tool a mouse study can still buy. Electron microscopy of double-membrane vesicles is the picture Ohsumi's yeast work began with, and it's still the picture a reviewer can ask for if the blots look too convenient. The programme ends as a lumen, not as a poster.

In short. The sealed bubble fuses with the acid depot, cargo is broken down, and a labelled adaptor protein falls on the same blot.

Selective autophagy is a set of named jobs, not a synonym for the whole programme, and mitophagy is the one a mitochondrial essay is allowed to steal. PINK1 accumulates on a depolarised outer membrane; Parkin ubiquitinates outer-membrane proteins; adaptors including OPTN and NDP52 bring LC3; the mitochondrion is wrapped and delivered. Egan and Shaw put AMPK–ULK1 on that path in 2011. MOTS-c, the 16-mer this catalogue holds, is a peptide the organelle wrote from 12S rRNA and is AMPK-adjacent; it isn't a mitophagy ligand, and sitting on AMPK isn't sitting on PINK1. NAD+ is the hydride coin Complex I wants oxidised, and a drained mitochondrial pool is one of the stresses that can drop the membrane potential PINK1 is watching. Two jobs, one organelle, a recycle programme that can take the organelle itself. Xenophagy, aggrephagy, ER-phagy, lipophagy: other cargoes, other adaptors, the same ATG chassis. A fasting essay that writes 'autophagy' and means 'the cell ate a mitochondrion' has skipped the chassis. Write mitophagy if you mean mitophagy. Write LC3 flux if you mean the programme.

In short. Worn mitochondria have their own tagged path into that bubble. That is a specialised recycling job, not the whole programme.

Assays, named, because 'we measured autophagy' isn't a methods line. A Western for LC3-I and LC3-II, with and without bafilomycin A1 or chloroquine, is flux. A Western for p62, moving in the opposite direction when flux is on, is the cargo-adaptor check. Phospho-ULK1 Ser757 falling and phospho-ULK1 Ser317 rising is the switch. Phospho-S6K Thr389 falling is the mTORC1-off check. Phospho-ACC Ser79 rising is the AMPK-on check. Immunofluorescence of LC3 puncta, counted, with the same trap, is the spatial version of the blot. Electron microscopy of autophagosomes is the ultrastructure. GFP-LC3–RFP-LC3 tandem reporters, in a dish that can take them, distinguish autophagosomes from autolysosomes by colour, because GFP quenches in the acid lumen and RFP does not. A Seahorse XF will not measure autophagy; it will measure the oxygen-consumption invoice that a recycled mitochondrion, or a mitochondrion that was not recycled, might change. Pick the assay that matches the sentence. Klionsky 2021 is the citation a reviewer will reach for if you do not.

In short. A single snapshot of the stitched protein can lie. Trap the last step, then read the blot. That is flux.

Satchidananda Panda's time-restricted eating work is the same switch, aligned to the clock, and it's the literature a sixteen-eight timetable is quietly borrowing. Hatori, Panda, Cell Metabolism 2012: mice fed a high-fat diet inside an eight-hour window, calories matched to ad-libitum controls, were protected against the obesity, insulin resistance and hepatic fat the ad-libitum animals acquired. The calories weren't the variable. The clock was. Feeding is a zeitgeber for peripheral oscillators; a hepatocyte that sees food at all twenty-four hours is a hepatocyte whose BMAL1, whose NAMPT, whose gluconeogenic enzymes, never quite know which shift they are on. Panda, Hogenesch, Kay and the microarray era had already shown that a large fraction of the protein-coding genome is circadian in at least one tissue. Time-restricted eating is a way to put food back into the hours those genes expected. It isn't a separate autophagy physics. It's the mTOR–AMPK switch, run on a 24-hour cycle rather than on a random fridge. The clock essay next door is the transcriptional loop. This one is the kinase that loop is allowed to time.

In short. Eating inside a daily window, rather than around the clock, is the same switch aligned to the body clock.

Sixteen-to-eight is a window, not a constant. The number is popular because it's easy to say, because skipping breakfast or skipping dinner fits a working day, and because eight hours is long enough to eat two or three meals without a lecture. Autophagy is graded, not a light that flips at sixteen hours elapsed. A person whose last meal was small, whose liver glycogen is already low, who walked to work, will be further into the AMPK-on, mTORC1-off state at hour twelve than a person whose last meal was large, late, and followed by a sofa. A person asleep is a different metabolic object from a person typing. Treating sixteen as a law of nature is how people skip dinner and call it a protocol, and how a late eating window that runs from 13:00 to 21:00 gets filed under the same heading as an early window that runs from 08:00 to 16:00. Those two windows aren't the same experiment at the insulin-sensitivity microphone, and they aren't obviously the same experiment at ULK1. The clock is the other variable. The nucleotide ratio is the first.

In short. Sixteen hours empty is a popular timetable. How long is long enough depends on the last meal, liver starch, and whether you moved.

Circadian insulin sensitivity is the reason an early window and a late window aren't interchangeable, even when both last eight hours. Glucose tolerance is better in the morning in most humans; insulin is more effective; the liver is more willing to stop making glucose. Van Cauter's sleep-and-metabolism work, and the later CGM-era replications, already showed that a restricted night moves the next-day curve. A late eating window asks a liver that thinks it is evening to handle a glucose load it would rather not see. Sutton, Peterson, Cell Metabolism 2018: early time-restricted feeding, a six-hour window finished by mid-afternoon, improved insulin sensitivity, blood pressure and oxidative-stress markers in men with prediabetes, even without weight loss. That's a clock result as much as a calorie result. Wilkinson, Panda, Taub, Cell Metabolism 2020: a ten-hour window in metabolic syndrome moved weight, blood pressure and atherogenic lipids. Ten hours, not sixteen; six hours, not sixteen. The popular sixteen-eight is a convenience. The papers are a set of windows, often earlier, often shorter, sitting on a liver clock that believed you would sleep.

In short. Insulin works better in the morning in most people. An early eating window uses that fact; a late one fights it.

Fasting also lifts NAD+ and sirtuin tone, which is the redox reason this essay sits next to the NAD+ catalogue item. NAMPT, the kinetic bottleneck of salvage, is a circadian output of CLOCK/BMAL1 in some tissues — Ramsey, Bass, the Northwestern papers — and restriction and exercise raise it in the tissues those papers actually measured. A rising NAD+/NADH ratio gives SIRT1 and SIRT3 something to spend: PGC-1α deacetylation, FOXO, a mitochondrial neighbourhood the sirtuin essay already mapped. AMPK can stabilise NAMPT, so the fuel gauge and the salvage enzyme talk. That's a measured physiology. It isn't a reason to write a 1000 mg cake of β-NAD+ as a feeding window. The dinucleotide is charged, 663 daltons, a laboratory reagent for a sirtuin tube and a PARP tube and a standard curve. Intact NAD+ doesn't stroll through a plasma membrane. The neighbouring essay wrote the pharmacokinetics and the three drains. This essay's job is to keep the redox ledger on the same page as ULK1 without collapsing a cofactor, a kinase and a meal schedule into one juice. Three jobs. One campus.

In short. Hours without food also raise the NAD+ pool the neighbouring essay maps. Sirtuins then have more cofactor to spend.

A feeding window, a calorie cut, and an every-other-day fast are different protocols sitting on the same two kinases, and pretending they are synonyms is how a literature becomes a brand. Time-restricted eating, in the Panda sense, can be isocaloric: the same energy, a tighter clock. Caloric restriction, since McCay, is a cut in energy, often with the clock left unspecified. Alternate-day fasting and five-two patterns are intermittent cuts, with feast days that can still be late and large. All three can raise AMP/ATP, quiet mTORC1, and lean on ULK1 in some tissues, on some clocks, in some animals. They don't do it equally, and they don't do it in every organ on the first morning. Autophagy in liver comes on faster than autophagy in muscle, which is one reason a short window can change hepatic fat while leaving a myofibril relatively un-recycled. Muscle needs the build side on the days you load it. A protocol that only ever empties is a protocol that has picked a side. The kinases don't require you to pick a side. They require you to name which state you wanted, in which tissue, on which day.

In short. A feeding window, a calorie cut, and an every-other-day fast are different protocols sitting on the same two kinases.

A week that uses both states on purpose is a protocol, and it's a more honest object than a religion of empty hours. A training day with a protein-rich meal is a build shift: mTORC1 on in the fibres you loaded, leucine in the sestrin2 pocket, S6K1 marked, 4E-BP1 released, ULK1 held at Ser757. An overnight fast plus a morning walk is a recycle shift: AMP/ATP up in the working muscle, AMPK on, mTORC1 quieter, ULK1 marked at the AMPK sites, LC3 flux a real question in the tissues you could actually biopsy. Those are shifts, not enemies. People who live in permanent build — constant grazing, never empty — miss the recycle. People who live in permanent recycle — under-eating, never loading a bar — miss the build. Both miss half the physiology. Write the training. Write the meal. Write the empty hours. Then the kinases have a timetable they can actually recognise, in British English and in residues. If you can't say which shift a given day was for, you haven't yet started. The week is the unit that can hold both.

In short. A training day with a protein-rich meal is a build shift. An overnight fast plus a walk is a recycle shift.

The lift still wants mTORC1, which is the sentence a recycle timetable will try to talk you out of and which the muscle literature will not. Mechanical tension, leucine, a permissive insulin, sleep: those are the native inputs to myofibrillar protein synthesis. A session in the empty hours is lawful, and some people prefer it; the meal that follows still has to arrive while the fibre is listening, which is a window of hours, not a rumour about an 'anabolic window' measured in minutes. A session followed by more empty hours, because the timetable hadn't yet opened, is a session whose translational invoice wasn't paid. Autophagy will take what it is given. It doesn't know that you meant only the damaged mitochondria. Protein at the meal after the session is how you tell mTORC1, through sestrin2 and the Rags, that this fibre is a build project today. The recycling can have the night. Both, in one week, is the protocol. Picking only one, and making it a personality, is how a serious kinase pair becomes a side.

In short. Lifting still needs the build side in the fibres you loaded. Skipping protein after a session is not a clever recycle.

The two failure modes are easy to name and hard to occupy at once, which is why the week is the right unit. Permanent build looks like a fridge that never shuts, a late glucose load, a liver writing VLDL at midnight, a phospho-S6K that never quite falls, a TFEB programme that never quite runs, an LC3 flux that is a trickle. Permanent recycle looks like a chronic energy deficit, a training log that is only walking, a leucine dose that never clears the sestrin2 threshold, a myofibril that is being inventoried along with the cargo you wanted, a resting metabolic rate that sags, a bone that notices. DiRECT, in the neighbouring type-2 essay, is a supervised, finite recycle shift applied to a fatty liver and a pancreas, with food reintroduced on purpose; it isn't a personality of emptiness. Resistance training plus protein, in the neighbouring training essay, is a supervised build shift applied to a fibre. A week that clocks both, on purpose, in named tissues, is the smallest protocol this switch actually justifies. A timetable that clocks only one has picked a side.

In short. People who never empty, and people who never load a bar, both miss half the physiology. The week is the unit.

DiRECT is the clinical picture of a long recycle shift, and it belongs on this page so that autophagy doesn't get to pretend it invented type-2 remission. Lean, Taylor, the Counterweight-Plus formula, a steep energy deficit, mean weight loss in the neighbourhood of ten to fifteen kilograms, remission odds that tracked the loss and, on the magnetic-resonance pictures, the emptying of hepatic and pancreatic fat. Twin-cycle: a fatty liver overproduces glucose and VLDL; that fat lands in the pancreas; first-phase insulin goes quiet; empty the liver and a large fraction of people in the first years of the disease get the glucose back. That's an energy-deficit physiology, a glycogen-and-triacylglycerol physiology, and only secondarily an autophagy physiology. ULK1 will have been on. AMPK will have been on. mTORC1 in the hepatocyte will have been quieter. Those kinases aren't the reason to write DiRECT as an autophagy protocol. They're the reason a fasting essay is allowed to sit next to a diabetes essay without being a collage. Empty the depot. Name the tool. The tool in DiRECT was a formula diet, supervised, finite.

In short. A long, supervised energy deficit that empties a fatty liver is a recycle shift applied to one organ, not a personality.

Ketones are the neighbouring fuel and signal, and they aren't autophagy. When hepatic glycogen is low and fatty-acid oxidation is high, acetyl-CoA overflows; HMGCS2 commits the overflow to ketogenesis; β-hydroxybutyrate and acetoacetate leave the liver for brain, muscle and kidney. Cahill already measured the fuel job in starving humans. Newman and Verdin later put BHB on class-I HDACs; Youm, Dixit and colleagues put it on NLRP3. A millimolar metabolite gets to be both a substrate and a ligand. A well-formulated ketogenic pattern or a serious fast will raise BHB, and AMPK and ULK1 will often be on in the same hours, which is why the two essays sit on the same desk. Correlation isn't identity. You can have ketones with a still-fed mTORC1 if protein is high enough to hold sestrin2 and the Rags in the on-state; you can have autophagy in a dish with no ketones at all, because a dish doesn't have a liver. The ketone is a readout of hepatic acetyl-CoA overflow. Autophagy is a kinase cascade. File them next door. Don't file them as each other.

In short. Ketones are what the liver writes when carbohydrate is scarce. They are a neighbouring fuel and signal, not the recycling programme itself.

The mouse TRE literature is large enough to be a field, and it isn't a cartoon. Hatori 2012, already named: isocaloric eight-hour feeding, high-fat diet, protection against the metabolic disease the ad-libitum cage-mates acquired. Subsequent Panda-laboratory papers and independent groups extended the finding across diets, sexes and shift-work-like light cycles, with the usual caveats of strain, thermoneutrality and whether the control was truly matched for energy. Mouse rooms are often 22 °C, which is cold for a clothed mouse and a chronic brown-fat stimulus; human rooms are not. Mouse feeding windows are often tighter than a human sixteen-eight. Those are reasons to read the figures rather than to throw the field away. What the mice show, repeatedly, is that the clock of feeding is a variable, that hepatic fat and insulin sensitivity notice it, and that the mTOR–AMPK switch is one of the microphones those papers can actually blot. What they don't show is that a human who skips breakfast has recapitulated an eight-hour mouse protocol. Species, dose, window, temperature. Write them.

In short. Mouse feeding-window papers, calories matched, still protected against a bad diet. Human trials are smaller and more mixed.

Human TRE trials are smaller and more mixed, which is what you should expect when you perturb a node this central in an animal that also has a job, a family and a fridge. Sutton 2018, already named: early six-hour window, prediabetic men, insulin sensitivity up, blood pressure down, weight not required. Wilkinson, Panda, Taub 2020: ten-hour window, metabolic syndrome, weight, blood pressure, lipids. Other trials have moved the feeding clock and left body weight where it was, or moved weight and left clamp insulin sensitivity where it was, or lost the effect once calories were tightly matched. Alternate-day and five-two patterns — Varady's work among the named human files — move weight more reliably because they often move energy, which is a different lever from the clock. That pattern is information. You can tighten a window and not own the phenotype the window sits under. The mouse figures are in print; the human windows are in print; the leap from an eight-hour mouse to a sixteen-eight timetable is a leap. A catalogue that sells a cofactor next door need not referee a diet argument. The tables already do.

In short. Every-other-day and five-two patterns also move weight and insulin in named studies. They are not the sixteen-eight window.

Measuring autophagy in a living human is the unglamorous gap a timetable never mentions. LC3-II on a muscle biopsy, with or without a chloroquine trap you are not about to give a volunteer, is a sparse literature. p62 on the same biopsy is sparser. Phospho-ULK1 residues in human muscle after an overnight fast plus exercise exist in a handful of papers; they don't exist as a consumer test. Plasma LC3 isn't a thing a clinic can sell honestly. Ketones, insulin, glucose, a CGM, a weight, a waist: those are the readouts a human TRE trial actually has, and they are one or two floors above ULK1. A trial that moves insulin sensitivity hasn't thereby shown LC3 flux in a hepatocyte. A trial that doesn't move weight hasn't thereby shown that autophagy failed. The mouse can give you the blot. The human can give you the clamp, the MRI-PDFF, the blood pressure, the adverse-event table. Translating one into the other is the work, and it's slower than a timetable. Design the human study as if someone who wasn't in the room will ask which floor you measured.

In short. Measuring recycling in a living person is hard. Muscle biopsies and a handful of blots are what the literature actually has.

Muscle is on the menu if the week is only empty, and that sentence has to stay on a page that is otherwise tempted to celebrate ULK1. During a steep energy deficit, amino acids from myofibrils contribute to gluconeogenesis; mTORC1 in the fibre is quiet; the loading stimulus, if it exists, is walking. Older adults, already closer to the sarcopenic threshold, notice first. Protein intake on the eating days, and actual resistance work, are how you keep the build shift on the timetable while the recycle shift does hepatic and mitochondrial jobs. A TRE protocol that accidentally becomes a protein-restriction protocol is a different experiment, closer to a leucine-sensing paper than to a clock paper. Sestrin2 doesn't know that you meant only liver fat. It knows whether leucine arrived. The protein-and-training essay is the dose. This essay is the warning: the same switch that recycles a damaged mitochondrion will recycle a myofibril if you never turn mTORC1 back on in that fibre. Both states. Named days. A bar, not only a window.

In short. If the week is only empty, muscle is on the menu. Load the fibres and feed them, on purpose, on some of the days.

A popular sixteen-eight often becomes a late window, which is the version the clock papers least support. Work, family, a restaurant, a 21:00 dinner: the eight hours slide toward the evening, insulin meets a liver that is trying to go off-shift, and the empty hours fall across a morning that already had the better glucose tolerance. Panda's human work, and Sutton's early-window study, leaned the other way: eat earlier, finish earlier, let the night be empty for a reason the SCN and the hepatocyte already agreed on. An early window is harder to live. It's also closer to the mouse papers and closer to the insulin-sensitivity microphone. None of that makes sixteen-eight a crime. It makes the clock a covariate you have to write down. If you report a feeding window and can't say whether it was 08:00–16:00 or 13:00–21:00, you haven't yet described the intervention. Light, sleep, the last meal's composition, the training day: write those too. The kinases will not mind the paperwork. The paper will.

In short. A popular sixteen-eight often becomes a late eating window. The original clock papers preferred an earlier one.

The catalogue listing next door is lyophilised β-NAD+, 1000 mg, ≥98% by HPLC, molecular weight 663.43, formula C21H27N7O14P2, CAS 53-84-9. A white to off-white cake in a vial, a certificate of analysis, a reconstitution kit. It's the same carbon skeleton every sirtuin, PARP and dehydrogenase paper names, and it's the cofactor a fasting-and-restriction physiology raises in the tissues those papers measured. It's a laboratory reagent. The form is the form you weigh into a sirtuin assay, a PARP assay, a cycling assay, a set of isolated mitochondria, a standard curve for LC-MS. Purity is a chromatogram, not a feeling. None of those sentences is a dose. None of them is a feeding window. None of them is a schedule. A fasting mimetic, if the phrase means anything, would be a molecule that recapitulates the transcriptional and metabolic programme of an empty stretch without the empty stretch. Rapamycin on mTORC1 is the example the gerontology field actually argues about. β-NAD+ wasn't asked to be a mimetic by any serious paper. It's the co-substrate. Don't write a gram of characterised dinucleotide as a meal timetable.

In short. The vial on this listing is freeze-dried NAD+, a gram, for the bench. It is not a fast in a bottle.

MOTS-c is the other catalogue object in this neighbourhood, and the neighbourhood has to stay a neighbourhood. MRWQEMGYIFYPRKLR, sixteen residues, translated from an open reading frame in mitochondrial 12S rRNA — Lee, Kim, Cohen, Cell Metabolism 2015: AMPK, the folate–methionine cycle, metabolic homoeostasis in mice. Sitting on AMPK is one door from ULK1 and from TSC2, which is why an autophagy essay is allowed to mention a 16-mer without claiming it. Retatrutide, if it is on your bench, is occupancy at GLP-1R, GIPR and GCGR, organism-level fuel demand, a different floor of the building; hepatic fat will move because energy moves, not because a triple agonist is an ULK1 ligand. IGF-1 LR3 is an 83-residue analogue at IGF1R, PI3K–Akt–mTOR in a myotube or a hepatocyte, the build side in a dish, the opposite shift. Four objects. A stack that treats them as interchangeable fasting juice hasn't named a cofactor, a reading frame, a receptor or a kinase. Neighbourhood, here, is a courtesy on a reading list. It isn't a combination claim, and it isn't a protocol.

In short. A mitochondrial sixteen-mer, a triple gut-hormone agonist, and a dish-level growth peptide sit nearby. Three other clocks.

Rapamycin is the pharmacological close of mTORC1, and it's the object a gerontology paper means when it says fasting mimetic with a straight face. Streptomyces hygroscopicus, Easter Island, a macrolide that with FKBP12 occupies an allosteric pocket on mTOR and brakes complex 1. Translation falls, ULK1 is released from the Ser757 mark, autophagy can start, TFEB can run. The mouse lifespan literature is real and argued; the immune and wound-healing costs are real and not argued. Metformin on AMPK and on Complex I is the other candidate, with a messier enzymology and a cleaner human-safety file. Neither is a meal schedule. Neither is a 1000 mg cake of β-NAD+. Use rapamycin to close mTORC1, and then show LC3 flux with a trap, and you've done the pharmacological version of the off-shift. Skip dinner in a mouse and then blot the same residues, and you've done the physiological version. Both are lawful. Collapsing them, or collapsing either with a cofactor vial, is how a journal becomes a stack. Use rapamycin as a tool. Use a fast as a fast. Weigh the cake for the assay that needs the standard.

In short. Rapamycin turns the build kinase down as a drug. That is a different object from skipping dinner.

Decide what you're measuring before you claim the recycle programme ran. LC3-II without a lysosomal blockade is a snapshot, and snapshots lie in both directions: a rise can be production or a jammed lysosome; a fall can be consumption or a failure to conjugate. Bafilomycin A1, chloroquine, or a protease-inhibitor cocktail for the last hours of the experiment is the trap. p62 should move the other way if flux is on. Phospho-S6K Thr389 and phospho-4E-BP1 are the mTORC1-off checks; if they haven't fallen, you aren't yet in the off-shift, whatever the clock says. Phospho-AMPK Thr172 and phospho-ACC Ser79 are the fuel-gauge checks. Phospho-ULK1 Ser757 falling and Ser317 rising is the switch itself. If you only have LC3-II, you have a number. If you have the trap, the cargo adaptor, the two ULK1 marks, and the mTORC1 substrates, you have a programme. Design the assay as if someone from another lab will have to believe it. That's the reason Klionsky's guidelines run to hundreds of pages.

In short. To claim recycling rose, trap the last step and then blot. A single un-trapped snapshot can mean the opposite.

The ULK1 residues aren't interchangeable, and writing 'phospho-ULK1' without the site is how a switch becomes a blur. Ser757 is the mTORC1 mark, inhibitory, the brake. Ser317 and Ser777 are the AMPK marks, activating, the starter. Commercial antibodies exist for both; they are only as good as the peptide they were raised against and the controls you run. A lambda-phosphatase-treated lane, a rapamycin lane, an AICAR or A-769662 lane, an amino-acid-withdrawal lane: those are the brackets. If Ser757 falls and the AMPK sites don't rise, you may have taken the brake off without pressing the starter; residual ULK1 activity and other inputs can still nucleate, but you haven't shown the AMPK half. If the AMPK sites rise and Ser757 stays high, mTORC1 is still sitting on the complex and the Kim-and-Guan paper predicts a weak start. Both marks, in opposite directions, is the picture a genuine fast in a competent tissue should be able to give you. One mark is a hint. Write the residue on the figure, not only in the methods.

In short. Name which mark on the starter kinase you measured. The build switch and the fuel gauge write on different amino acids.

Time since the last amino acid is the variable most dish papers under-report, and it's the variable that decides whether you have a Rag experiment or a growth-factor experiment. Amino-acid withdrawal, with insulin left on, asks the Rag–sestrin2 axis. Serum withdrawal, with amino acids left on, asks the PI3K–Akt–TSC axis and a cloud of other factors in foetal bovine serum. True starvation — amino acids, serum and glucose down — asks AMPK as well, and will mint AMP on a faster clock. A sixteen-hour serum starve in HEK293 isn't a sixteen-eight. A two-hour leucine withdrawal in a myotube is a sestrin2 experiment. Write the time. Write what was withdrawn. Write what was left. In an animal, the last meal's composition and clock are the equivalent honesty test: a high-protein dinner holds mTORC1 longer than a small carbohydrate dinner; a 22:00 meal is a different hepatic clock from a 18:00 meal. If you add β-NAD+ to a medium and harvest at twenty-four hours, you've done an ectonucleotidase-plus-salvage experiment unless you've shown otherwise. The cake is a standard. The minute you stop the chemistry is the result.

In short. Write the time since the last amino acid. Withdrawal of serum is not the same experiment as a true empty.

Cell-type choice is a control, not a convenience. A HEK293 well will tell you whether a construct and an amino-acid withdrawal move LC3-II. It won't tell you what a hepatocyte does with a sixteen-hour window, and it won't tell you what a myotube does with leucine after a stretch. Primary hepatocytes still have a serious AMPK and a serious mTORC1 at the lysosome; neurons mostly live on a different salvage and a different risk. Muscle is the tissue in which the build side is a myofibril; liver is the tissue in which the recycle side is glucose output and fat; endothelium is a SIRT1-and-nitric-oxide neighbourhood the NAD+ essay already owns. Aged mouse tissue isn't a late-passage dish, and a late-passage dish isn't ageing, it's a Hayflick neighbourhood with its own mTOR story. If the claim is fasting, the animal has to have been fasted, the last meal named, the tissue named, and the cell sort, if you can afford it, done, because a whole-liver LC3 number is a hepatocyte diluted by immune cells that may be running a different programme. That last sentence is a design demand, not a flourish.

In short. Pick the cell that has the job you are claiming. A kidney-cell line is not a muscle fibre or a liver cell.

Machines, named, because a cascade this central will not survive a vague line in the write-up. A wet tank and an LC3 antibody, with a loading control that isn't itself an autophagy substrate, is the undergraduate blot. A phospho-specific ULK1 pair, a phospho-S6K Thr389, a phospho-AMPK Thr172, a phospho-ACC Ser79, p62, Beclin-1, ATG5: the panel. Immunofluorescence and a spinning-disk or a confocal, puncta counted by a script not by hope. Electron microscopy, double membranes, a morphologist who has seen a genuine autophagosome. A plate reader will not save you. A Seahorse XF or an Oroboros O2k is the mitochondrial invoice if mitophagy is the claim; TMRM with FCCP and oligomycin as the brackets for membrane potential. qPCR for TFEB targets (CTSD, LAMP1, SQSTM1) if the transcriptional limb is the claim. LC-MS for NAD+, NADH, NMN if the redox ledger is on the same figure. None of that is glamorous. All of it is how you stop a fasting essay becoming a mood. The cake on the shelf is for the assays that need the standard. The window on the clock is for the animal you actually fasted.

In short. Name the machine: a blot, a microscope count of bubbles, an electron micrograph. Saying you measured recycling is not a method.

The node is conserved, which is the only reason a yeast ATG mutant, a mouse eight-hour window and a human clamp can sit in one essay without being a collage. Ohsumi's ATG genes still have mammalian names. ULK1 is Atg1. Beclin-1 is Atg6. LC3 is Atg8. Rag GTPases and sestrin2 are the amino-acid inbox a mammal added with a lysosome. AMPK is Snf1 with a γ subunit that binds AMP. mTORC1 is the build kinase rapamycin named. Panda put food back on the clock. Sutton put an early window on a clamp. Taylor put a long recycle shift on a fatty liver. Hardie, Sabatini, Guan, Shaw, Mizushima, Levine, Klionsky: the map is public. Conservation isn't a licence to treat a mouse figure as a human protocol. It's a licence to take the biochemistry seriously enough to measure it, in the organism you have, with the residues named, the trap run, and the last meal written down. The popular story got loud because the node is central. The work got hard for the same reason. Sixteen-eight is a timetable. The switch is a pair of kinases.

In short. From yeast to humans the same two kinases decide build versus recycle. A mouse feeding window is not automatically a human plan.

Leave with a topology, not a shopping list. Amino acids and insulin turn mTORC1 on at the lysosome, through Rag GTPases and sestrin2 on one cable and PI3K–Akt–TSC–Rheb on the other; S6K1 and 4E-BP1 then run translation, and ULK1 is held off at Ser757. AMPK, kicked by AMP/ATP, marks TSC2 and Raptor and marks ULK1 at a different residue; autophagy starts, with Beclin-1 and LC3 lipidation as the named steps you can blot. Ohsumi made that programme assayable. Panda aligned the same switch to the clock; sixteen-eight is a window, not a constant; earlier windows often sit better with insulin. Resistance training still needs the build side. A week that uses both states on purpose is a protocol. NAD+ is the redox ledger fasting also moves; the 1000 mg cake is that dinucleotide for a tube. MOTS-c is a different object on the same campus. Rapamycin is another. If your experiment needs the cofactor, weigh it, quench it, and name the drain. If it needs a fast, empty the hours, on purpose, and turn mTORC1 back on in the fibres you loaded. If it needs a medicine, this catalogue doesn't sell one.

In short. Leave with the map: two inboxes into the build kinase, a fuel gauge, a starter kinase, a blot for the stitched protein, both states in one week.

Research-use-only. Not for human consumption / not a medicine. The lyophilised β-NAD+ on the neighbouring listing is a laboratory reagent, HPLC-characterised at ≥98 percent, labelled for in-vitro work: a sirtuin tube, a PARP assay, a standard curve, a set of isolated mitochondria whose oxygen consumption you actually record. The physiology in the paragraphs above is public, cited, and older than the vial. Use it to design the experiment you have the controls for, with the residue named, the trap run, the last amino acid written down, and the tissue named. Read Ohsumi, read Sabatini, read Hardie, read Panda, read Klionsky, then weigh the cake if the assay needs the cofactor, or skip dinner if the assay needs the empty hours. We'll sell you the dinucleotide those empty hours also raise. We won't tell you it's a feeding window, a sixteen-eight, or a personality. The switch is a pair of kinases. Both states are useful. This rate you can measure, on a blot, with a chromatogram on the bench beside it if the redox ledger is part of the figure.

In short. The vial is a research chemical for experiments, not a medicine and not food. The biology is public. Weigh it, or empty the hours, and keep the claim the size of the blot.