The Proteolipid Code
A project to restore empirical legitimacy to membrane biology, currently held intellectually hostage by the unfalsifiable, evidence-averse, and spectacularly self-congratulatory lipid raft delusion.
Membrane Biology Is Broken
Membrane biology is in a state of theoretical emergency. The lipid raft hypothesis, the idea that cholesterol and sphingolipids spontaneously form submembrane platforms that recruit proteins, got the execution order backwards. Lipids do not self-organize into submembrane platforms that summon proteins. Proteins and lipids co-organize from the outset. This is not a subtle distinction. It is the difference between a theory and a fantasy, and the field built on the fantasy for three decades.
The Proteolipid Code Is a Fix
The proteolipid code restores the correct execution order. Every protein has a characteristic lipid fingerprint. Lipids serve as glue, antagonists, codons, and more. Together, they define distinct membrane zones. The framework incorporates the ubiquitin, glycan, and phosphoinositide codes and applies to eukaryotic, prokaryotic, and viral membranes alike.
The proteolipid code framework. Not shown: the tears shed developing it.
Figure 2d // Kervin TA, Overduin M. BMC Biol. 2024. doi:10.1186/s12915-024-01849-6
The proteolipid code at a glance. Each component earns its keep.
Figure 1a // Overduin M et al. Curr Opin Struct Biol. 2025. doi:10.1016/j.sbi.2025.103061
In Memoriam
Obituaries for Failed Models
They served the field. Sort of. For a while. We're being generous.
The Lipid Raft Hypothesis
1997 – Ongoing DenialCause of death: Reality
The lipid raft hypothesis proposed that cholesterol and sphingolipids spontaneously coalesce into submembrane zones which then recruit proteins. Lipids first, proteins second. This is backwards. Proteins and lipids co-organize from the outset. No one has ever demonstrated a lipid-only raft forming first and recruiting a protein to it in a living cell. The field treated this absence of evidence not as a problem but as an invitation to keep looking, which is the scientific equivalent of insisting your imaginary friend is just very shy. Survived by an embarrassing volume of citations and the sunk-cost fallacy.
doi:10.1038/42408Modified Lipid Raft Theories
2006 – Still twitchingCause of death: Patch upon patch upon patch
When the original raft hypothesis started taking on water, the field did what any reasonable community would do: they added epicycles. "Rafts are real, they're just smaller!" "Rafts are real, they're just transient!" "Rafts are real, you just can't see them!" At some point, an invisible, transient, nanoscale entity that can't be directly observed is less a hypothesis and more a shared hallucination.
doi:10.1101/cshperspect.a041395Picket-and-Fence Model
2005 – Quietly retiredCause of death: Overengineered irrelevance
A baroque contraption proposing that cytoskeletal "fences" and anchored protein "pickets" corral lipids into compartments. Think Rube Goldberg, but for membranes. The so-called "hop diffusion" it invoked turns out to be a general feature of crowded membranes, not a picket fence required. A needlessly complex model that somehow managed to explain membrane organization without meaningfully accounting for the membrane.
doi:10.1146/annurev.biophys.34.040204.144637Tiered Mesoscale Model
2012 – Dead on deliveryCause of death: Doubled down on a losing hand
When the picket-and-fence model didn't quite work, the solution was apparently to make it bigger. The tiered mesoscale model layered additional complexity onto an already shaky foundation (ad hoc modifications stacked on analogical reasoning), achieving the rare feat of being both more complicated and less explanatory than what it replaced.
doi:10.1016/j.tibs.2011.08.001Shelby-Veatch Model
2023 – Prognosis terminalCause of death: Correlation mistaken for causation
Shelby, Veatch, and colleagues proposed that eukaryotic membranes are poised near a critical point of phase separation, making this the organizing principle of the cell surface. The model rejects lipid fingerprints (the observation that each protein is surrounded by a characteristic lipid environment) and instead assumes proteins sit in homogeneous regions that interpolate between two artificial phases. The evidence derives from model membranes and infers causation from correlation, without the control groups needed to rule out simpler explanations. A framework built on binary phases in synthetic systems is not a theory of the living membrane.
doi:10.5281/zenodo.17201252The "Functional Paralipidome"
2023 – Dead on arrivalCause of death: Internally inconsistent; forgot about solvent
Lipid fingerprints were established by Corradi V et al. The functional paralipidome is a renaming by investigators sympathetic to the raft framework who could not accept that a raft skeptic was advancing the field. Even its own authors concede the overlap. The definition is inconsistent: the paralipidome is called "the local lipid environment" but its energetic content is equated with ΔGP–L alone, excluding the ΔGL–L term in the same framework. Elsewhere it is treated as a component within the fingerprint rather than equivalent to it. The free energy decomposition omits solvent contributions entirely, and the ΔGL–L term is undefined. The model claims to be predictive but offers no falsification criteria.
doi:10.1038/s41580-022-00524-4
Zone architecture of the proteolipid code. Each zone is a distinct entity. Unlike rafts, these actually exist.
Figure 1c // Overduin M et al. Curr Opin Struct Biol. 2025. doi:10.1016/j.sbi.2025.103061
Lipids: Not Just "Structural" and "Signaling"
For decades, membrane biology operated with a taxonomy of lipid function that would embarrass a toddler's crayon box: structural and signaling. Two categories. For thousands of lipid species. Two.
The proteolipid code calls for a reassessment. Lipids play roles that are far more varied, specific, and, dare we say, interesting:
The components in action. Every lipid has a job. Some have several.
Figure 2b // Kervin TA, Overduin M. BMC Biol. 2024. doi:10.1186/s12915-024-01849-6
Consolidation
Codes Within the Code
The proteolipid code maintains a respectful, collegial relationship with the genetic code. It has not extended the same courtesy to the others. The ubiquitin code, the glycan code, and the phosphoinositide code each governed their own small territories of membrane biology — admirably, in some cases — but the era of independent operation is over. The proteolipid code has absorbed them all into a single unified jurisdiction. Resistance was not a factor. There was nothing to resist. A superior organizational framework simply arrived, and the former codes now operate under its authority, retaining their local functions but reporting to a higher structure. This is not a merger of equals.
| Code | Former Purpose | Status | Terms of Absorption |
|---|---|---|---|
| Ubiquitin Code | Tags proteins for degradation, trafficking, signaling | ASSIMILATED | Cooperated immediately. Retained operational autonomy. Now serves as a regional governor under the new order. |
| Glycan Code | Cell recognition, immune modulation, protein folding | ASSIMILATED | Bewilderingly complex, but complexity is not sovereignty. Integrated without incident. Glycobiologists have been offered counseling. |
| Phosphoinositide Code | Membrane identity, vesicle trafficking, signaling | ASSIMILATED | Seven lipids governing an outsized bureaucracy. Folded neatly into the regime once the org chart was made clear. |
| Others | Various | PENDING | Annexation is ongoing. Compliance is expected. Timetable at our discretion. |
Why We're Allowed to Do This
Proposing a unifying theory and absorbing other codes into it is obnoxious. We know. But the proteolipid code makes two falsifiable predictions, which is the minimum price of admission for calling something a theory.
Falsifiable Prediction #1: Each Protein Has a Lipid Fingerprint
Every membrane protein is surrounded by a characteristic set of lipids. The Shelby-Veatch model says the opposite: proteins sit in homogeneous lipid regions determined by bulk phase behavior. Show that membrane proteins lack distinct lipid fingerprints and the proteolipid code is dead.
Corradi V et al. ACS Cent. Sci. 2018. doi:10.1021/acscentsci.8b00143
Falsifiable Prediction #2: Proteins Are Not Recruited to Lipid-Only Platforms
Lipids do not spontaneously form submembrane platforms that then recruit proteins. Proteins and lipids co-organize from the start. This contradicts the original lipid raft hypothesis. Show that a lipid-only platform forms first and recruits proteins to it in a living cell and the proteolipid code is dead.
