Why BAX doesn't always finish what it starts at the membrane

What BAX is

BAX (BCL-2-associated X protein) is the pro-apoptotic effector of the BCL-2 family — the protein a cell uses to commit to programmed cell death. Where BCL-2 and BCL-xL guard the mitochondrial outer membrane, BAX and its close relative BAK are the effectors that breach it under cellular stress.

In this article

1. What BAX does and why it matters
2. VDAC2 catches BAX mid-activation
3. What the structural model reveals
4. Bench notes for researchers in this space
5. Frequently asked questions
6. What the community gets wrong

The name is historical: BAX was identified in a complex with BCL-2 and turned out to do the opposite job. Since then it has been one of the most widely studied proteins in cell-death research, which is why a first look at a previously unseen intermediate state is worth pausing on.

BAX at a glance

The detail that matters for this article is that BAX does not flip straight from inactive to porated. It occupies membrane-bound intermediate states along the way, and the structural study covered below caught it in one of them: primed, membrane-associated, and held there by VDAC2 before any pore forms.

What BAX is not

• Not an external agent, a toxin or a drug — it is an endogenous protein that cells already carry.
• Not interchangeable with every other BCL-2 family member: the family splits into protectors, activators and pore-forming effectors, and BAX sits in the last group.
• Not a simple two-state switch. Cytosolic inactive BAX and membrane-inserted pore-forming BAX are the two ends of a graded process, with stable intermediate states in between — the finding this article is about.

What BAX does and why it matters

In one line: BAX is a BCL-2 family protein that punches pores in the mitochondrial outer membrane, committing a stressed cell to self-destruction.

It works in four stages:

• Idle: it drifts inactive in the cytosol, held in check by BCL-2 survival proteins.
• Trigger: cellular stress changes its shape and sends it to the mitochondria.
• Commitment: it assembles into pores in the outer membrane — mitochondrial outer membrane permeabilization (MOMP).
• Aftermath: pro-death signals flood the cytoplasm and the death program runs.

Why it matters: MOMP is the point of no return. Once the outer membrane is breached, that decision cannot be reversed, so the steps leading up to the pore carry as much weight as the pore itself. That is what makes BAX a control point rather than a simple switch: it does not jump from idle to pore in one move, it passes through intermediate activation states, and whatever governs those intermediates decides the outcome. Those half-activated states, and what holds BAX in them, are what this study set out to answer.

VDAC2 catches BAX mid-activation

Cells lacking VDAC2 are partially resistant to BAX-driven apoptosis (programmed cell death). Without a structural picture of the two proteins together, the mechanism stayed unresolved: whether VDAC2 keeps BAX off the membrane, or stops it partway. This study, a bioRxiv preprint not yet peer reviewed, built that picture, and the answer is the second one — VDAC2 holds BAX at one defined step of activation.

The researchers reconstituted a stable VDAC2-BAX complex — both proteins produced, combined under controlled conditions, and confirmed to physically bind — then characterized it with biochemical and biophysical methods plus AlphaFold3 (Google DeepMind's AI protein-structure modeling tool), constrained by the experimental data.

The Primed Intermediate State

"Caught mid-activation" means held at one defined step — not simply kept off the membrane. At that step:

• The anchor helix (a segment called the α9 helix) is already inserted into the mitochondrial outer membrane.
• The BH3 domain — a short loop other death-promoting proteins recognize and engage — is exposed.
• A region near the N-terminal end (one end of the protein chain) is more accessible.
• BAX looks activated, but no pore has been built.

The complex is stable — which is what let the study map which BAX surfaces stay available while VDAC2 holds it.

Isoform Specificity: VDAC1 vs. VDAC2

VDAC1 — closely related, in the same protein family, structurally similar — did not form this complex. That isoform specificity (two nearly identical proteins doing different jobs) matters when designing experiments that target one but not the other: an assay built around VDAC2 binding needs VDAC2, and the isoforms are not interchangeable.

What the structural model reveals

The model answers one question directly: where is BAX, and what is it doing, at this stage? Combining AlphaFold3 predictions with experimental constraints, it places BAX in contact with VDAC2 — a channel protein that normally lets small molecules cross the outer mitochondrial membrane — before any pore forms.

What the model shows

• BAX is anchored, not assembled. Its α9 helix holds it at the membrane, while the water-soluble domain remains separate.
• The soluble domain drapes over the VDAC2 pore. The protein is held in place rather than released.
• VDAC2 contacts a defined intermediate. The model specifies where along BAX that contact occurs.

Key point: VDAC2 acts as an external checkpoint rather than a pore component, holding BAX in an activation-competent intermediate until additional signals trigger pore assembly.

What changed from earlier models

Earlier models put VDAC2 inside the death pore itself. This study argues otherwise: VDAC2 gates access to the final activation step from outside the structure. What the model captures is the checkpoint state, not a completed pore. "Activation-competent intermediate" means BAX is primed and waiting; whether it proceeds depends on signals arriving later. The kill switch sits downstream, and the signals that release it are not part of this model.

Bench notes for researchers in this space

If your assays use BH3-domain peptides or other compounds that engage BCL-2-family proteins, these handling variables shift your readout before any compound reaches the plate.

• Store cold and reconstitute fresh. Synthetic BH3 peptides can adopt unwanted secondary structure at room temperature, reducing engagement with the target groove; freeze-thaw cycles erode activity over time.
• Know what is in your diluent. Benzyl alcohol is a preservative, not an inert carrier, and oxidizes to benzaldehyde over time. Endotoxin in water or diluent can independently trigger apoptotic signaling in cell-based work, adding background signal. Use a sterile, unpreserved diluent with published endotoxin specs.
• Treat a color shift as data. GHK-Cu's blue is its copper(II)-peptide complex. A solution fading to clear typically means that complex is breaking down — copper displaced by chelators or reducing agents, or a pH drift. Check the lot before running a series.
• Purity grade matters for short peptides. BH3 peptides are short sequences with strong folding tendencies; lower-purity lots carry truncated or oxidized variants that may fail to engage the BH3-recognition groove at expected concentrations.
• Use glass for low-concentration work. Peptides at nanomolar concentrations adsorb to plastic, shifting your effective working concentration with no visible sign. Glass cartridges give your stock a more stable baseline.

The VDAC2-BAX finding turns a long-standing genetic observation into a physical picture with a mechanism attached.

What the research community gets wrong about BAX

The short answer: BAX is usually scored as a finished pore, so the intermediate steps - where most compounds act - stay invisible, and VDAC1 and VDAC2 get treated as interchangeable when the reconstituted data separate them. Five assumptions carry most of the error, and only some of them are about BAX itself.

Five assumptions worth re-testing

• Treating BAX as an on/off switch. It passes through several shape changes and can sit primed at the membrane without ever forming a pore, so a readout that scores only the final pore cannot show where a compound is acting.
• Assuming VDAC1 and VDAC2 are interchangeable. They are close relatives and look alike structurally, but they do not behave the same way here: in the reconstituted work, VDAC2 captured BAX and VDAC1 did not. Check which isoform a reagent or knockout is actually hitting.
• Thinking VDAC2 is part of the pore. The newer structural picture places it outside the pore, holding BAX at an intermediate and gating the final step - a checkpoint, not a building block.
• Treating BH3 peptides as stable stock. They fold on their own, oxidize, and lose activity across freeze-thaw cycles; at low concentration they also stick to plastic, so working concentration drifts below the label.
• Ignoring the diluent. Endotoxin in bacteriostatic water or another diluent can trigger apoptotic signaling on its own, and that background reads as compound activity.

What benzyl alcohol does to a peptide in solution

Benzyl alcohol is not inert in solution. As a preservative it works by perturbing membranes, which makes it an active small molecule in a vial of protein - and a protein whose job is membrane interaction is a poor candidate for preservative-neutral. Hold the preservative constant before attributing a stability shift to the sequence; switching diluents mid-study changes the vehicle and the peptide at once, and no later assay separates them.

Color follows the same rule: a copper-peptide that turns from blue to clear is reporting on its metal center, not on its backbone, so treat the change as an observation rather than a degradation verdict.

From our bench: split one BH3-domain peptide stock between a plastic tube and a glass vessel, then measure working concentration from each after a set hold time at your usual bench temperature. Send the paired numbers, peptide length, purity grade and hold time, and we will add real observations - not estimates - to this page.

Sources

Reminder: research and educational reference only. PreppinPeppers sells hardware and materials, not peptides. Not medical advice, not evaluated by the FDA, and not intended for human or animal use.