Research Challenge 01 · Experiment / Mechanism · Core · 30 min
Is Rheb an ON/OFF switch?
Is Rheb-mediated regulation of mTORC1 determined mainly by nucleotide state, by localisation, or by both?
This is a decision environment, not a quiz. Several of the choices below are scientifically defensible, and the feedback tells you what each one buys and what it costs rather than marking it right or wrong.
What you should be able to do
By the end of this challenge you should be able to
- Distinguish a claim about a protein's activation state from a claim about where that protein is
- Rank candidate experiments by how well they separate two competing explanations
- Identify what an overexpression result can and cannot support
- Propose the control that would make a spatial claim interpretable
Research skills Causal reasoning · Experimental design · Control selection · Evidence evaluation · Spatial reasoning · Scientific uncertainty
Recommended preparation
- What is mTOR?What the kinase integrates, and what its readouts are.
- mTORC1 vs mTORC2Why a claim about ‘mTOR’ is usually a claim about one of two complexes.
- Rheb — The Critical RegulatorThe GTP/GDP cycle, the TSC complex as a GAP, and the uncertainty this challenge starts from.
Step 1
What we know
Every line here is carried by a study in the Atlas, linked beside it. Read the second list as carefully as the first: it is where your own decisions start.
What the Atlas already supports
- TSC2 acts directly on Rheb's nucleotide cycle: it accelerates the conversion of Rheb-GTP to Rheb-GDP, and cells without a working TSC complex hold more Rheb-GTP and more mTOR signalling. INOK2003
- The loading step is under physiological control. Insulin raises Rheb-GTP in a PI3K-dependent way, and TSC1/TSC2 blocks that increase. GAR2003
- Raising Rheb is sufficient to raise mTORC1 output and cell growth in the systems where it was tested. SAU2003 INOK2003
- The amino-acid input works on position rather than on loading: the Rag GTPases bind raptor and move mTOR to a compartment that also holds Rheb, and a GTP-locked Rag mutant makes signalling resistant to amino-acid withdrawal. SAN2008
- The brake moves too. TSC2 accumulates on lysosomes whenever amino acids or growth factors are missing, so ‘state’ and ‘place’ are not two separate compartments of the problem. DEM2016
Where that stops
- Several of the foundational nucleotide-state experiments used overexpressed protein. They report what Rheb can do when there is more of it, which is a different statement from what endogenous Rheb does.
- The two arms were largely measured in separate experiments, often in different cell types. Their relative contribution in one cell at one moment is not something the studies in this Atlas resolve.
- No study in this corpus moves Rheb itself between compartments at endogenous levels while leaving its nucleotide cycle intact — which is the experiment the question is really asking for.
Step 2
Build and break the model
Switch the controls and watch which combinations produce output. Then break it on purpose: each prediction below asks you to commit before it answers. These are predictions about a simplified teaching model, not measurements.
Two lanes converge on the same kinase. The upper lane is the growth-factor arm that sets Rheb's nucleotide state; the lower lane is the nutrient arm that decides whether mTORC1 is recruited to the lysosomal surface at all. The location control stands for that lower lane being engaged or not. Nodes and arrows are taken from the Atlas pathway model, so this teaching diagram cannot drift away from the scientific one.
p-S6K1 high · 4E-BP1 phosphorylated
Both conditions are met at once: the activator is loaded and the kinase is where the activator is. This is the state every hypothesis on your list predicts, which is exactly why it cannot tell them apart.
- Akt/PKB
- AGC kinase requiring PIP3 recruitment plus two phosphorylations to be fully active.The main 'grow' relay from growth factors.T308 by PDK1 and S473 by mTORC2. Recruitment and activation are separate events — a distinction the older Atlas diagram blurred.
- v-ATPase
- Lysosomal proton pump physically and functionally coupled to Ragulator.The pump the sensing machinery is built on.Required for amino-acid signalling; inhibitor experiments cannot fully separate the signalling role from loss of lysosomal acidification.
- TSC1/TSC2
- TSC1–TSC2–TBC1D7 complex; a GAP that switches Rheb off.The pathway's master brake.Integrates Akt, AMPK, ERK/RSK, GSK3 and REDD1 inputs. Regulation is substantially about lysosomal recruitment, not only phosphorylation-driven activity change.
- Ragulator
- LAMTOR1–5 complex tethering the Rags to the lysosomal surface.The bolt holding the taxi to the membrane.Also reported as a RagA/B GEF, though the GEF assignment is less secure than the tethering role; LAMTOR1 lipidation anchors the whole assembly.
- Rheb
- Small GTPase; GTP-loaded Rheb allosterically activates mTORC1.The one thing that actually switches mTORC1 on.Realigns the mTOR active site; the convergence point of the entire growth-factor arm and the reason localisation alone is not activation. Note a declared simplification in this map: Rheb is farnesylated and distributes across the endomembrane system, with a substantial ER and Golgi pool, and which pool supplies the activating Rheb is still debated. It is drawn on the lysosomal band because that is where it meets mTORC1.
- Rag GTPases
- RagA/B–RagC/D heterodimers; nucleotide state determines mTORC1 recruitment.The taxi that brings mTORC1 to the lysosome.Note the inversion: RagA/B is active when GTP-loaded, RagC/D when GDP-loaded. The Rags control mTORC1 *location*, not its catalytic activity.
- mTORC1
- mTOR–Raptor–mLST8 (+PRAS40, DEPTOR); switched on only at the lysosome.The growth decision itself.Coincidence detector: nutrients supply location via the Rags, growth factors supply activation via Rheb. Neither alone is sufficient — the single most important idea in the pathway.
- S6K1
- Ribosomal protein S6 kinase; the standard readout of mTORC1 activity.mTORC1's best-known output kinase.T389 phosphorylation by mTORC1 is rapamycin-sensitive, which is why S6K1 became the field's default assay — and why the field long over-read rapamycin as a complete mTORC1 inhibitor.
- 4E-BP1
- Translational repressor released from eIF4E upon multi-site phosphorylation.A cap on protein-making that mTORC1 removes.Only partially rapamycin-sensitive. This single fact explains the rapalog/Torin discrepancy and drove the whole ATP-competitive inhibitor programme.
This is a simplified teaching model with a fixed set of states, not a simulation. It shows the direction each control pushes the pathway, not how much, how fast, or what any particular cell would do.
Every state of this model
| Growth factors | Rheb nucleotide state | mTORC1 location | Readout | What it means |
|---|---|---|---|---|
| ON | GTP | lysosome | p-S6K1 high · 4E-BP1 phosphorylated | Both conditions are met at once: the activator is loaded and the kinase is where the activator is. This is the state every hypothesis on your list predicts, which is exactly why it cannot tell them apart. |
| ON | GTP | cytosol | p-S6K1 low · 4E-BP1 unphosphorylated | Sit with this one. The hormone is present, the brake is released, Rheb is loaded — and the output is still low, because the kinase is not at the place where the activating step happens. If you hold that nucleotide state alone determines the output, this state is the one you have to explain away. |
| ON | GDP | lysosome | p-S6K1 low · 4E-BP1 unphosphorylated | The mirror image, and the reason recruitment is not the same thing as activation: mTORC1 arrives at the surface, finds an unloaded activator, and does nothing. Docking is a precondition, not the switch. |
| ON | GDP | cytosol | p-S6K1 low · 4E-BP1 unphosphorylated | Neither condition is met. Note that the readout is identical to the two states above it — one blot cannot tell you which of the two requirements was the one that failed. |
| OFF | GTP | lysosome | p-S6K1 high · 4E-BP1 phosphorylated | Loaded Rheb with the brake engaged and no hormone. In the model this is a held state; in a cell it is what a TSC-null background looks like — mTORC1 active without the permission signal that normally licenses it. |
| OFF | GTP | cytosol | p-S6K1 low · 4E-BP1 unphosphorylated | The same loaded Rheb as the state above, and no output, because the kinase is elsewhere. Read the two together: they are the cleanest statement this model can make that position is doing work of its own. |
| OFF | GDP | lysosome | p-S6K1 low · 4E-BP1 unphosphorylated | Nutrients are there, the hormone is not, the brake is engaged and Rheb has unloaded. mTORC1 sits on the surface waiting for an input that is not coming. |
| OFF | GDP | cytosol | p-S6K1 low · 4E-BP1 unphosphorylated | The resting or starved state: no permission signal, brake engaged, activator unloaded, kinase not recruited, autophagy no longer restrained. |
Prediction 1
Break the brake. Take away the growth-factor signal and, at the same time, disable the TSC complex so Rheb stays GTP-loaded. What does the model do?
Removing the brake keeps Rheb loaded, so the growth-factor input becomes dispensable — but only in the states where mTORC1 is at the surface. Switch the location control to cytosol and the same loaded Rheb produces nothing. This is a prediction about the teaching model, not a measurement; the experiments below are where measurements come in.
Show the expected answer
B. Output is high, as long as mTORC1 is still recruited to the lysosome
Removing the brake keeps Rheb loaded, so the growth-factor input becomes dispensable — but only in the states where mTORC1 is at the surface. Switch the location control to cytosol and the same loaded Rheb produces nothing. This is a prediction about the teaching model, not a measurement; the experiments below are where measurements come in.
Prediction 2
Now remove the spatial component instead: leave growth factors on and Rheb GTP-loaded, and stop mTORC1 being recruited. What does that state tell you about the claim ‘Rheb-GTP determines mTORC1 activity’?
The claim as written says loading determines output. The model produces a state with loading and without output, so the claim needs a qualifier: Rheb-GTP determines the output given that mTORC1 is where Rheb is. Adding that qualifier is not a retreat — it is the more testable version of the same idea.
Show the expected answer
B. It weakens the claim as stated, because loading is present and output is not
The claim as written says loading determines output. The model produces a state with loading and without output, so the claim needs a qualifier: Rheb-GTP determines the output given that mTORC1 is where Rheb is. Adding that qualifier is not a retreat — it is the more testable version of the same idea.
Prediction 3
Three different states in this model give an identical p-S6K1 readout. What follows for an experiment that measures only p-S6K1?
This is the whole idea of experiment informativeness in one picture. A readout is useful in proportion to how differently the competing explanations predict it. p-S6K1 is a fine readout of output; it is a poor discriminator here, because three quite different states of the cell produce the same value. What separates them is measuring the nucleotide state or the localisation as well.
Show the expected answer
B. A low p-S6K1 alone does not tell you which requirement failed, so the design needs a second measurement that differs between the candidates
This is the whole idea of experiment informativeness in one picture. A readout is useful in proportion to how differently the competing explanations predict it. p-S6K1 is a fine readout of output; it is a poor discriminator here, because three quite different states of the cell produce the same value. What separates them is measuring the nucleotide state or the localisation as well.
Step 3
Commit to a working hypothesis
Commit to a working hypothesis before you spend anything. It is meant to be revised later — that is what the last step of this challenge is for.
What each choice commits you to
Nucleotide state is dominant. The strong version of the textbook arrow: what the kinase reads is whether Rheb carries GTP, and everything upstream matters only through that. It predicts that any manipulation raising Rheb-GTP raises output, and that position is a housekeeping detail. The state to watch is growth factors ON, Rheb-GTP, mTORC1 cytosolic — if that gives no output, this hypothesis needs a qualifier.
Localisation is dominant. The spatial version: the regulated step is whether mTORC1 is brought to the surface where its activator waits, and the nucleotide cycle is largely permissive. It predicts that forcing the location makes the nutrient input dispensable — which is close to what one of the experiments below reports. Its weak point is that forcing the location in a growth-factor-replete cell leaves loading untested.
Both contribute, and they are not independent. Two requirements that have to be met together, with the added twist that the brake itself relocates. It is the most defensible reading of the current evidence and the hardest to test, because ‘both matter’ only becomes a scientific claim once you say what would change if one of them did not.
The current evidence does not separate them. A legitimate position rather than a refusal to answer. It commits you to naming the missing experiment, which is what the last step of this challenge asks for. Hold it seriously and you should be reluctant to spend your budget on experiments that every hypothesis predicts identically.
Step 4
Run the investigation
This is a research path, not a shopping list. Each step costs what its equipment costs, and each one opens up different next steps — so the order you choose decides which questions you can still afford to answer. You can go back to any earlier step and take another branch; what you have already spent stays spent. The panel above tracks which parts of the research question your results can actually answer.
What you can answer so far 0 of 6
- Is Rheb's nucleotide state actually controlled, and by what?
- Does a physiological input move that control?
- Can changing Rheb alone change the output?Two findings, and that is the point: an overexpression result only answers this once you have shown the input arm did not move with it.
- Is where mTORC1 sits a control point in its own right?
- Are state and place two separate variables at all?
- Which arm contributes more, in one cell, at endogenous levels?Nothing in this corpus yields this. The only step that would address it returns a prediction rather than a result, so this sub-question stays open at any price — which is the honest ending of the challenge, not a gap in the menu.
Research budget
100 of 100 research units remaining
Each step costs what its equipment costs. Running everything would take 160 units, and answering every sub-question that can be answered would take 120 — so this budget buys four of the six answers at most. Which four is the whole exercise.
100 of 100 research units left
Everything else currently open to you
The research path, without JavaScript
With scripting on, this page walks you through the steps below one decision at a time and keeps the budget. Without it, every step is on the page in full and here is the map that would otherwise live in the script.
Which step answers which sub-question
| Is Rheb's nucleotide state actually controlled, and by what? | In vitro GAP assay: does TSC2 unload Rheb directly? |
|---|---|
| Does a physiological input move that control? | Does a physiological input change Rheb loading in cells? |
| Can changing Rheb alone change the output? | Raise Rheb and see whether output follows + Measure p-AKT in the same lysates |
| Is where mTORC1 sits a control point in its own right? | Force mTORC1 to the lysosomal surface and remove the nutrient signal |
| Are state and place two separate variables at all? | Watch where the brake goes when the inputs are withdrawn |
| Which arm contributes more, in one cell, at endogenous levels? | nothing in this corpus answers this |
What each step opens up
- In vitro GAP assay: does TSC2 unload Rheb directly? opens Does a physiological input change Rheb loading in cells?, Raise Rheb and see whether output follows
- Does a physiological input change Rheb loading in cells? opens Raise Rheb and see whether output follows, Force mTORC1 to the lysosomal surface and remove the nutrient signal
- Raise Rheb and see whether output follows opens Measure p-AKT in the same lysates, Watch where the brake goes when the inputs are withdrawn
- Measure p-AKT in the same lysates opens Force mTORC1 to the lysosomal surface and remove the nutrient signal
- Force mTORC1 to the lysosomal surface and remove the nutrient signal opens Watch where the brake goes when the inputs are withdrawn, Move Rheb itself, at endogenous levels, both ways
- Watch where the brake goes when the inputs are withdrawn opens Move Rheb itself, at endogenous levels, both ways
- Move Rheb itself, at endogenous levels, both ways opens nothing further
You can start with
In vitro GAP assay: does TSC2 unload Rheb directly?, Does a physiological input change Rheb loading in cells?, Force mTORC1 to the lysosomal surface and remove the nutrient signal
The cheapest routes
- 10 research units → 1 of 6 sub-questions
- 30 research units → 2 of 6 sub-questions
- 60 research units → 3 of 6 sub-questions
- 85 research units → 4 of 6 sub-questions
The best this budget allows is 4 of 6 sub-questions for 85 research units.
10 research units
In vitro GAP assay: does TSC2 unload Rheb directly?
Addresses The nucleotide arm, and only its biochemistry
Needs Recombinant protein and a nucleotide-loading assay. No cells, no imaging — the cheapest thing on this menu, and it shows.
The design
| Model | Purified proteins, in vitro |
|---|---|
| Perturbation | Add the TSC2 GAP domain to Rheb preloaded with GTP |
| Readout | Which nucleotide is bound to Rheb |
| Control | Rheb without TSC2; a GAP-dead TSC2 mutant |
Run · 10 research units spent
Not enough budget left for this one. What you have already spent stays spent.
What it opened up Having the assay working is what makes the two cell-based steps worth doing at all.
Derived from D Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling 2003
Rheb-GTP remaining
Direction and rank order as reported in the source study. The bars are a schematic of that direction, not measured values.
Interpret — what does this let you say?
What each choice commits you to
It supports my hypothesis. Be careful about what it supports. A result that every candidate explanation predicts does not favour any of them — it confirms shared background, which feels like progress and is not.
It weakens my hypothesis. It is hard to see how. No hypothesis on your list denies that TSC2 unloads Rheb; they differ about whether loading is what decides the output in a cell.
It is inconclusive for my question. That is the reading this result deserves. It is a clean, interpretable biochemical experiment about a step none of your hypotheses disputed.
It introduces an alternative explanation. Not really an alternative — more a reminder that the brake acts on the activator rather than on the kinase, which is worth keeping straight when you read the spatial experiments.
What this evidence supports
- TSC2 acts on Rheb's nucleotide cycle itself, rather than somewhere further downstream.
- The brake in the pathway diagram has a biochemical meaning: a GAP that speeds up unloading.
What it does not establish
- Anything about where either protein sits — this is a tube, and it has no compartments.
- Whether this rate is the one that matters at endogenous protein levels inside a cell.
Information value High for the mechanism of the brake, and zero for the question you actually asked: all four hypotheses predict this same result, so it cannot separate them. Spending 10 units here buys confidence in a step you were not doubting — though it is also what opens the two cell-based steps, so it is not money thrown away.
20 research units
Does a physiological input change Rheb loading in cells?
Addresses Whether the nucleotide arm is regulated at all
Needs Cell culture, serum starvation and stimulation, a GTP-loading pull-down and a western blot.
The design
| Model | Cultured mammalian cells and Drosophila |
|---|---|
| Perturbation | Stimulate serum-starved cells with insulin, with and without TSC1/TSC2 |
| Readout | Fraction of Rheb in the GTP state, alongside p-S6K1 |
| Control | Serum-starved, unstimulated cells; PI3K inhibition |
Run · 20 research units spent
Not enough budget left for this one. What you have already spent stays spent.
What it opened up A regulated loading step is the reason to ask whether loading is what decides the output — and the reason to go looking at position.
Derived from D Insulin activation of Rheb, a mediator of mTOR/S6K/4E-BP signaling, is inhibited by TSC1 and 2 2003
Rheb-GTP loading
Direction and rank order as reported in the source study, which also places the insulin effect downstream of PI3K. Schematic, not measured values.
Interpret — what does this let you say?
What each choice commits you to
It supports my hypothesis. If your hypothesis is nucleotide-dominant, this is real support for the premise — the arm is regulated. It is not yet support for the claim that this arm is the one that decides the output.
It weakens my hypothesis. Only for the strongest form of the localisation hypothesis, the one that treats loading as fixed background. The moderate form survives this result untouched.
It is inconclusive for my question. Defensible. Two things move together here, and the design does not break that pairing apart.
It introduces an alternative explanation. It does raise one: if insulin changes loading and output at once, some of the output change may come from elsewhere in the insulin arm rather than through Rheb.
What this evidence supports
- Rheb loading is a regulated variable in living cells, not a constant the cell sets once.
- TSC1/TSC2 sits between the hormone and the loading step, which is what makes it the point where growth-factor signalling enters.
What it does not establish
- That the loading change is what caused the p-S6K1 change — the two are measured together, and moving one without the other is a separate experiment.
- Anything about position: nothing here reports where Rheb or mTORC1 were.
Information value Moderate. It separates ‘loading is regulated’ from ‘loading is constitutive’, which matters if you hold the localisation hypothesis in its strongest form. It is silent on the location arm.
20 research units
Raise Rheb and see whether output follows
Addresses Whether the nucleotide arm can drive the output on its own
Needs Cell culture and transfection or stable lines, western blot, plus Drosophila tissue for the growth readout.
The design
| Model | Cultured mammalian cells; Drosophila tissues |
|---|---|
| Perturbation | Overexpress Rheb, or a form biased toward the GTP state |
| Readout | p-S6K1 and p-4E-BP1; cell and tissue growth |
| Control | Empty vector; rapamycin to test whether the effect runs through mTOR |
Run · 20 research units spent
Not enough budget left for this one. What you have already spent stays spent.
What it opened up A result with a complication in it opens exactly two moves: measure the complication, or go and look somewhere else entirely.
Derived from D Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling 2003 · D Rheb promotes cell growth as a component of the insulin/TOR signalling network 2003
mTORC1 output (p-S6K1 / p-4E-BP1)
Direction and rank order as reported across the two source studies, which also report increased growth in Drosophila tissue. Schematic, not measured values.
Interpret — what does this let you say?
What each choice commits you to
It supports my hypothesis. It supports the nucleotide arm being able to drive output. Notice what you have to add silently to get from there to ‘nucleotide state is what decides’: that the extra protein went where it was needed.
It weakens my hypothesis. For a strong localisation hypothesis, an output rise without any spatial manipulation is uncomfortable — unless the extra Rheb is also reaching the compartment, which nothing here measures.
It is inconclusive for my question. A fair reading, and the reason is worth naming: an overexpression experiment changes abundance, and abundance is a third candidate explanation you have not been tracking.
It introduces an alternative explanation. It does, and the next panel makes it concrete. Raising Rheb does not leave the rest of the network where it was.
What this evidence supports
- More Rheb is sufficient to raise mTORC1 output in these systems, and the effect runs through mTOR rather than around it.
- Rheb sits downstream of the TSC complex and upstream of the kinase, which is what put it on the pathway diagram in the first place.
What it does not establish
- What endogenous Rheb does. Sufficiency at a raised level and necessity at a normal level are different claims, and this design supports the first.
- Which variable did the work: the amount of protein changed, and for a state-biased mutant the state changed as well.
- Anything about location, which was not measured.
Information value Historically decisive for placing Rheb on the map, and weak for the question in front of you: two variables move together, and neither of them is position.
New information
Raising Rheb does not only raise mTORC1 output. In the same work that placed Rheb downstream of TSC1/TSC2, Rheb overexpression also lowered PKB/AKT phosphorylation — the arm that was supposed to be the input. GAR2003
Does this affect your conclusion?
What each choice commits you to
Yes, substantially. This is the reading the design supports. The experiment did not change one arrow; it changed a network that contains a feedback loop, and the readout you measured sits downstream of both the change and the loop.
Possibly. Reasonable if you think the feedback is slow relative to the measurement. Say what would make you sure — a time course, or measuring p-AKT in the same lysate, turns ‘possibly’ into something checkable.
No. Hard to hold once the feedback is on the table: the input arm moved, so ‘more Rheb-GTP acting on mTORC1’ is one of at least two routes from the manipulation to the readout.
Not enough information. Also defensible, and it points at the missing measurement rather than at the missing knowledge: the experiment as run does not report the state of the insulin arm.
mTORC1 output feeds back on upstream insulin signalling, so a cell with more Rheb is not the same cell with one arrow strengthened. The rise in p-S6K1 is a real observation either way. What needs care is the interpretation: attributing it specifically to more Rheb-GTP acting on mTORC1 requires the feedback arm to be held constant, or at least measured.
The control that would help Measure p-AKT in the same lysate; use an acute, inducible increase rather than a stable line, so the network has less time to re-equilibrate; or use an endogenous knock-in so that abundance is not the variable at all.
15 research units
Measure p-AKT in the same lysates
Addresses Whether the overexpression result came through Rheb at all
Needs The lysates you already have and one more antibody. The cheapest follow-up here, and the one that decides what the step above meant.
The design
| Model | The same cultured cells and the same lysates as the step above |
|---|---|
| Perturbation | None new — this reads a second output from material you already have |
| Readout | p-AKT alongside p-S6K1 |
| Control | Empty-vector lysates from the same experiment |
Run · 15 research units spent
Not enough budget left for this one. What you have already spent stays spent.
What it opened up With the feedback on the record you can go at the spatial arm without carrying an unresolved objection.
Derived from D Insulin activation of Rheb, a mediator of mTOR/S6K/4E-BP signaling, is inhibited by TSC1 and 2 2003
p-AKT in the same cells
Direction as reported in the source study, which notes that Rheb overexpression lowers PKB/AKT phosphorylation. Schematic, not measured values.
Interpret — what does this let you say?
What each choice commits you to
It supports my hypothesis. It supports being able to state your hypothesis carefully. A measured confounder is not a confirmed one — it is a known one.
It weakens my hypothesis. If you read the overexpression result as clean evidence that loading drives the output, then yes: part of that effect has a second route to the readout.
It is inconclusive for my question. It is clear about one thing: the design above is not a single-variable experiment. That is worth 15 units even though it answers nothing new about Rheb.
It introduces an alternative explanation. It confirms one that was already there. Feedback from mTORC1 onto the insulin arm is standard biology; this puts a number-free measurement next to it.
What this evidence supports
- The manipulation moved the input arm as well as the output, so what changed was a network with a feedback loop in it, not one arrow.
- With the feedback measured rather than assumed, the size of the objection is on the record and the overexpression result can be stated with its scope attached.
What it does not establish
- How much of the p-S6K1 rise travelled through Rheb and how much through the feedback. Separating those needs an acute or endogenous manipulation, not another blot.
- Anything about location — nothing here reports where any of these proteins were.
Information value Low on its own and high in combination. Fifteen units turn a result you could not interpret into one you can, which is the cheapest thing you will do in this challenge.
30 research units
Force mTORC1 to the lysosomal surface and remove the nutrient signal
Addresses Position and nucleotide state against each other
Needs Stable lines expressing Rag mutants, amino-acid withdrawal, western blot. Weeks of line-making before the first data point.
The design
| Model | Human cell lines |
|---|---|
| Perturbation | Express a GTP-locked Rag mutant that keeps mTORC1 at the lysosomal surface regardless of amino acids |
| Readout | p-S6K1 with and without amino acids; mTOR localisation by imaging |
| Control | Wild-type Rag at a matched expression level; a GDP-bound Rag mutant |
Run · 30 research units spent
Not enough budget left for this one. What you have already spent stays spent.
What it opened up Once position is a control point, the questions become how far it goes and whether it is really separable.
Derived from D The Rag GTPases bind raptor and mediate amino acid signaling to mTORC1 2008
mTORC1 output (p-S6K1)
Direction and rank order as reported in the source study: locking the Rag mutant in its GTP-bound form made mTORC1 signalling resistant to amino-acid withdrawal. Schematic, not measured values.
Interpret — what does this let you say?
What each choice commits you to
It supports my hypothesis. If you committed to localisation, this is the strongest support available in the corpus. Keep the qualifier in view: it shows the position requirement can be satisfied artificially, not that the loading requirement went away.
It weakens my hypothesis. For a strict nucleotide-dominant view it is awkward: withdrawing amino acids should not have mattered much, and in wild-type cells it did.
It is inconclusive for my question. Too cautious, but for a good reason: the experiment separates the amino-acid input from the output, and leaves the two requirements bundled together in the growth-factor-replete condition.
It introduces an alternative explanation. It reframes the question more than it answers it: position may be a gate that has to be open rather than a dial that sets the level.
What this evidence supports
- Where mTORC1 sits is a control point in its own right: fix the position and the cell no longer needs the amino-acid signal to keep signalling.
- The amino-acid arm and the growth-factor arm are doing different jobs — one moves the kinase, the other loads its activator.
What it does not establish
- That position on its own is enough. These cells still had growth factors, and therefore an activator that was still loaded.
- Anything about Rheb's nucleotide state in these cells, which was not measured in the same experiment.
- That the mutant Rag behaves like an endogenous protein at an endogenous level — it is expressed, so its amount is a variable too.
Information value The most discriminating experiment on this menu, because the hypotheses predict different things here. A nucleotide-dominant view expects the amino-acid withdrawal to matter little either way; a localisation-dominant view expects exactly the pattern reported.
25 research units
Watch where the brake goes when the inputs are withdrawn
Addresses Whether the two arms are separate variables at all
Needs Immunofluorescence microscopy and a panel of stress conditions — cheap reagents, expensive microscope time.
The design
| Model | Cultured mammalian cells |
|---|---|
| Perturbation | Withdraw amino acids, or growth factors, or apply other stresses |
| Readout | TSC2 localisation by imaging, alongside mTORC1 output |
| Control | Full medium; cells with the inputs restored |
Run · 25 research units spent
Not enough budget left for this one. What you have already spent stays spent.
What it opened up There is one step left after this, and it is the one nobody has taken.
Derived from D Lysosomal recruitment of TSC2 is a universal response to cellular stress 2016
TSC2 on lysosomes
Direction and rank order as reported in the source study, which finds that both inputs have to be present for TSC2 to stay cytoplasmic. Schematic, not measured values.
Interpret — what does this let you say?
What each choice commits you to
It supports my hypothesis. It supports the ‘both, and not independent’ reading more than either single-arm hypothesis, because it shows the arms sharing a mechanism.
It weakens my hypothesis. It weakens both strong forms at once: neither arm is the clean, isolable variable that a dominance claim needs.
It is inconclusive for my question. In the narrow sense, yes — it measures neither loading nor mTORC1 position. In the wider sense it tells you the question may be badly posed, which is worth knowing.
It introduces an alternative explanation. It does, and it is the most interesting one available: what looks like two competing mechanisms may be one mechanism with a spatial and a biochemical face.
What this evidence supports
- The brake is spatially regulated too: TSC2 moves to the compartment where Rheb and mTORC1 meet whenever an input is missing.
- ‘Nucleotide state’ and ‘location’ are not two independent variables, because the enzyme that sets the nucleotide state is itself positioned.
What it does not establish
- That TSC2 is acting on Rheb at that location — imaging shows where a protein is, not what it is doing there.
- Anything quantitative about Rheb loading, which this design does not measure.
Information value It reframes rather than answers. Run late, it turns a two-way question into a better-posed one; run first, it spends a quarter of the budget on a complication you cannot yet use.
40 research units
Move Rheb itself, at endogenous levels, both ways
Returns a prediction, not a result
Addresses Position alone at endogenous abundance — an experiment nobody has published
Needs A knock-in line, reciprocal targeting constructs and live imaging. The most equipment on this menu, and part of why nobody in this corpus has done it.
The design
| Model | Cultured cells with a knock-in, so abundance is held at the endogenous level |
|---|---|
| Perturbation | Tether Rheb to a defined membrane, and separately remove it from that membrane, leaving the nucleotide cycle intact |
| Readout | p-S6K1 and p-4E-BP1, Rheb-GTP loading, and the localisation of both proteins in the same cells |
| Control | Untethered knock-in carrying the same tag and linker |
Run · 40 research units spent
Not enough budget left for this one. What you have already spent stays spent.
Expected, not observed No study in this Atlas corpus reports this experiment. What follows is the outcome the model above predicts, written as a prediction and labelled as one. Treat it as a hypothesis to be tested, not as a result to be cited.
mTORC1 output (predicted)
Expected, not observed. These bars are what the teaching model predicts if position and loading are two requirements that both have to be met. Nothing in the Atlas corpus reports this measurement.
Interpret — what does this let you say?
What each choice commits you to
It supports my hypothesis. A prediction cannot support a hypothesis; it is generated by one. If this felt like support while you were reading it, that is the reflex the labelling is here to interrupt.
It weakens my hypothesis. Also not something a prediction can do. What it can do is tell you which hypothesis you would abandon if the experiment came out the other way.
It is inconclusive for my question. Correct, and the useful follow-up is: what would you have to see here to change your mind? Naming that in advance is most of what makes an experiment worth running.
It introduces an alternative explanation. It introduces a design problem rather than an explanation: the tether adds variables, so even the real version of this experiment would need its own controls.
What this evidence supports
- Only the reciprocal pair — adding the protein to a compartment and removing it from the same compartment — would let a spatial claim stand on its own.
- Holding abundance at the endogenous level is what would separate this from the overexpression experiment above.
What it does not establish
- Anything at all, until it is done. Reading a predicted pattern as evidence is the specific error this challenge is built to make visible.
- A tether is itself a perturbation: the tag, the linker and the local concentration are three changes riding along with the one you meant to make.
Information value The highest in principle and the most expensive, and it is the gap in the literature rather than a line in it. Choosing it costs 40 units and returns a prediction — which is a fair picture of what committing to an unpublished experiment actually feels like.
Step 5
Revise the hypothesis
Your working hypothesis, after the evidence you bought
Changing your mind here is the expected outcome, not a failure state. A first hypothesis that survives every experiment untouched usually means the experiments were not discriminating.
What each choice commits you to
Nucleotide state is dominant. Holding the nucleotide arm as dominant after these results means committing to a qualifier: loading decides the output given that the kinase is at the surface. That is a narrower and better claim than the one you started with.
Localisation is dominant. The forced-location experiment is the strongest support in the corpus for this reading. What it does not give you is the reciprocal: those cells still had growth factors, so the loading requirement was quietly satisfied throughout.
Both contribute, and they are not independent. The most defensible reading of what you have seen, and the one that inherits the hardest job: making ‘both’ specific enough to test. The brake-localisation result is the reason it is not simply a hedge.
The current evidence does not separate them. A position with a duty attached: name the experiment that would separate them. The one this challenge could not buy you a result for is the reciprocal targeting of Rheb at endogenous levels.
Step 6
Compare with published research
Now, and not before, here is what someone actually did.
What the researchers actually tested
Whether the Rag GTPases are the missing link between amino acids and mTORC1 — measured as binding to raptor, as mTOR's position inside the cell, and as whether a GTP-locked Rag makes signalling resistant to amino-acid withdrawal.
What it answered
- Amino acids act on mTORC1 largely by moving it to a compartment that also holds its activator.
- Fixing that position artificially removes the cell's need for the amino-acid signal.
- The Rag proteins do not themselves switch on the kinase activity — the recruitment step and the activation step are separable.
What it did not answer
- It did not measure Rheb's nucleotide state in those cells, so the two arms were not separated within one experiment.
- The Rag mutants are expressed rather than endogenous, so their abundance is a variable alongside their state.
- It shows that reaching the compartment is required, not that being in the compartment is by itself enough.
The published experiment is not the perfect version of your design, and it was never meant to be. It is a real design with its own controls, its own scope and its own limits — and reading it that way, rather than as a verdict, is the skill this step is for. Compare it with what you chose: which of your unspent experiments would you now run alongside it?
Study page →Step 7
Where the question stands
You have committed, spent, revised and compared. Here is what the evidence in this Atlas actually supports today — separated into what was measured, how it is read, and what is still nobody's answer.
The short answer
On the evidence in this Atlas: no, not on its own. Rheb behaves as one of two requirements that have to be met at the same time — the kinase has to be at the compartment, and the activator it meets there has to be loaded.
Observation — what has actually been measured
- TSC2 accelerates the conversion of Rheb-GTP to Rheb-GDP, and cells without a working TSC complex hold more Rheb-GTP and more mTOR signalling. INOK2003
- Insulin raises Rheb-GTP loading in cells in a PI3K-dependent way, and TSC1/TSC2 blocks that rise. GAR2003
- Raising Rheb raises S6K1 and 4E-BP1 phosphorylation and drives growth in the systems where it was tested. SAU2003 INOK2003
- A GTP-locked Rag mutant holds mTORC1 at the lysosomal surface and makes its signalling resistant to amino-acid withdrawal. SAN2008
- TSC2 itself accumulates on lysosomes whenever amino acids or growth factors are missing. DEM2016
Interpretation — how the field reads it
Read together, these support a two-requirement model rather than a single switch. Nutrients decide where the kinase is; growth factors decide whether the activator waiting there is loaded. Neither arm produces output without the other, which is why the model in step 2 has two states with an identical readout and completely different causes.
Rheb is still the step the kinase reads directly, which is why the ON/OFF language stuck to it. What the evidence does not support is the stronger version: that the nucleotide state by itself sets the output regardless of where the kinase is.
The two arms are not independent either, because the enzyme that sets the nucleotide state is itself positioned — TSC2 moves to the same compartment when an input is withdrawn. So ‘state versus place’ is a useful way to design an experiment and a poor way to describe the mechanism.
Still open — what nobody has done
- No study in this corpus moves Rheb itself between compartments at endogenous levels while leaving its nucleotide cycle intact. That is the reciprocal experiment the question actually asks for, and it is the 40-unit item on your menu that returns a prediction instead of a result.
- The relative contribution of the two arms in one cell at one moment is measured nowhere: the arms were established in separate experiments, often in different cell types, and comparing across them is comparing designs as much as biology.
- Several of the foundational nucleotide-state results used overexpressed protein, so how much of the effect survives at endogenous abundance is an open question rather than a settled detail.
How each hypothesis comes out
- Nucleotide state is dominant
- Survives only with a qualifier: Rheb-GTP determines the output given that mTORC1 is at the compartment. The forced-location experiment is what forces the qualifier, and the qualified version is the more testable claim — so this is a hypothesis that got sharper rather than one that failed.
- Localisation is dominant
- Half-right in a way worth naming. Position is a genuine control point, but the experiment that shows it was run in growth-factor-replete cells, so the loading requirement was quietly satisfied the whole time. Nothing here shows position is sufficient on its own.
- Both contribute, and they are not independent
- The reading the current evidence best supports, and the one that inherits the hardest job: making ‘both’ specific enough to test. The brake-localisation result is what saves it from being a hedge — it names a shared mechanism rather than just declining to choose.
- The current evidence does not separate them
- Defensible for the narrow question of relative contribution, which really is unmeasured. Harder to hold for the broad question, because the forced-location result does decide something: it rules out the version where position is a housekeeping detail.
Step 8
What would you do next?
What changed your mind, if anything?
Which evidence mattered most to your final position?
The list above is where a next experiment would go. Pick the one you would actually defend in front of someone holding the budget — not the most interesting one, the one that would change what the field can claim.
You have one more opportunity and a limited budget. What would you do next?
What each choice commits you to
Build the reciprocal targeting experiment for Rheb at endogenous levels. The direct attack on the gap. Expensive, slow, and the only option that turns the spatial claim into something with its own controls. Budget the reciprocal arm from the start — a one-directional tether leaves the result ambiguous.
Measure loading and localisation in the same cells, under one set of conditions. Cheaper and often underrated. Much of the difficulty here comes from arms measured in different cells on different days; a single matched dataset would not settle the question but would tell you how far apart the arms really are.
Repeat the forced-location experiment with growth factors withdrawn. The missing cell of the existing design. If fixed position still gives output without a growth-factor signal, position is doing more work than the current reading allows; if it does not, the two-requirement model gains real support.
Test whether the brake's relocation is required for the output change. Follows the most interesting complication rather than the original question. Blocking TSC2's move while leaving its activity intact is hard to do cleanly, which is worth knowing before you commit the budget.
Step 9
Challenge complete
No score, and nothing unlocked. What you take away is a shorter list of things you would need to find out, and a clearer sense of which experiment would tell you.
Research skills this challenge practised
- Causal reasoning
- Experimental design
- Control selection
- Evidence evaluation
- Spatial reasoning
- Scientific uncertainty
Where to go next
Academy lessons
Guided routes
- How does a cell learn that it is allowed to grow?The growth-factor route, which is the nucleotide arm drawn end to end.
- How does a cell know it has enough raw material to grow?The amino-acid route, which is the spatial arm drawn end to end.
In the Atlas