Neuropeptide Y NPY is the endogenous agonist of the human Y2 receptor, a Gi-coupled GPCR that regulates appetite and a recognized target in obesity research. A 2018 collaboration between Uppsala University and Novo Nordisk combined computational modeling, peptide synthesis, and in vitro…
Neuropeptide Y NPY binds the human Y2 receptor as an agonist, and the map of that interaction assembled in 2018 is the clearest evidence drug designers currently have for building Y2 modulators aimed at obesity. The work, led by Bo Xu at Uppsala University in collaboration with Novo Nordisk, combined computational modeling, synthetic peptide chemistry, and in vitro pharmacology to describe how the 36-residue peptide engages the receptor. The team's stated conclusion was direct: understanding how NPY binds Y2 can guide the design of novel modulators of appetite regulation to fight obesity.
The binding arrangement follows the two-domain model that applies across the NPY receptor family. The amidated C-terminal tail of NPY, carrying the essential Tyr36 amide, reaches into the orthosteric pocket of the receptor, where it is required for activation. The central alpha-helix of the peptide contacts the extracellular surface of the receptor, and this second contact, rather than the conserved tail, largely determines which of the human NPY receptors a ligand engages. A designer who wants a Y2-selective modulator has to manage both domains at once.
The clinical rationale is large. World Health Organization data cited in the reporting of this work place the proportion of overweight adults in the European Union between 30 and 70 percent, and the proportion of obese adults between 10 and 30 percent, with wide variation across member states. In the United States, the Centers for Disease Control and Prevention finds that more than one third of adults are obese. Appetite regulation is one of the few control points upstream of both figures, and Y2 occupies a distinctive position in that circuitry: it is a brake on the orexigenic NPY neurons rather than a driver. That is why a receptor-level binding map of Y2 matters for obesity drug design.
This article explains the NPY molecule, the Y2 receptor, the three-pillar method used to map the interface, and what such a map does and does not entitle a team to conclude. The short answer to the design question is a qualified yes: a validated interaction map is the right foundation for structure-guided modulator design, but it is a starting point, not a drug.
NPY is a 36-residue amidated peptide, one member of a family that also includes peptide YY PYY and pancreatic polypeptide PP . In solution NPY adopts a compact fold called the PP-fold : a left-handed polyproline II helix spanning roughly the first eight residues, a tight turn, and an amphipathic alpha-helix running from about residue 14 to residue 31, followed by a flexible C-terminal tail that carries the conserved sequence Arg-Gln-Arg-Tyr-NH2. The polyproline helix packs against the hydrophobic face of the alpha-helix, stabilizing the fold and leaving the polar face of the helix available for receptor contact. NPY contains no cysteine, methionine, or tryptophan, so its synthetic chemistry is governed instead by the two adjacent tyrosines at positions 20 and 21 and by aspartate residues in the N-terminal half of the sequence.
The human genome encodes four functional NPY receptors: Y1, Y2, Y4, and Y5. Older literature also refers to Y3 and Y6, but no human Y3 gene exists and Y6 is a pseudogene in humans. All four are class A G protein-coupled receptors that couple to Gi/o proteins and inhibit adenylyl cyclase, yet their pharmacology and distribution differ sharply. Across the family, the consensus model of peptide recognition is a two-domain interaction. The C-terminal tail of the peptide inserts into the transmembrane pocket, providing most of the binding energy and all of the activation signal. The alpha-helix contacts the extracellular loops and the receptor N-terminus, and the sequence differences in those regions are what allow a 36-residue peptide to distinguish among four similar receptors. Y2 notably tolerates N-terminally truncated ligands such as PYY3-36 and NPY13-36 far better than Y1 does.
| Receptor | G protein | Endogenous ligands | Canonical role in energy balance |
|---|---|---|---|
| Y1 | Gi/o | NPY, PYY | Postsynaptic stimulation of feeding |
| Y2 | Gi/o | NPY, PYY, PYY3-36 | Presynaptic inhibition of NPY release; Y2 agonism linked to reduced food intake |
| Y4 | Gi/o | PP | Satiety signaling; less well defined |
| Y5 | Gi/o | NPY, PYY | Stimulation of feeding, overlapping with Y1 |
Y2 is the most abundant NPY receptor in the brain and the one most closely tied to the arcuate nucleus feeding circuit. The NPY/AgRP neurons of the arcuate nucleus are the classic hunger population: when they fire, food intake rises. NPY released from these neurons promotes feeding through postsynaptic Y1 and Y5 receptors. Many of the same neurons also carry Y2 as a presynaptic autoreceptor, so released NPY feeds back onto Y2 and inhibits further NPY release. Y2 therefore operates as an endogenous brake on the orexigenic circuit. PYY3-36 , released by intestinal L cells after a meal, is the best-characterized endogenous Y2-selective agonist, and its appetite-suppressing action has been mapped to Y2 in the arcuate nucleus. None of this in vivo circuitry was tested in the 2018 study; the team's contribution was the receptor-level binding map. The context clarifies the therapeutic logic: a Y2-selective agonist is a way to apply the brake without simultaneously engaging the Y1 and Y5 accelerator.
Activation itself is a conformational event. When the NPY tail settles in the orthosteric pocket, the receptor's transmembrane helices rearrange, opening a cytoplasmic interface for the Gi/o protein. The map a drug designer needs is therefore not a single static pose but a description of which contacts stabilize the agonist-bound, signaling-competent conformation. That distinction explains why binding affinity and efficacy are different readouts: a peptide can occupy the receptor without triggering the helical rearrangement, and an antagonist is exactly such an occupant. The 2018 workflow was designed to make that distinction experimentally, which is why it included functional pharmacology alongside binding measurements.
A peptide-receptor interface resists easy study. NPY is flexible in solution, and Y2, like all GPCRs, samples multiple conformations. Neither a picture of the free peptide nor a model of the empty receptor is sufficient. The Uppsala and Novo Nordisk team approached the problem with three coordinated techniques, now a standard pipeline for peptide GPCR projects: computational modeling to generate a hypothesis, synthetic peptide chemistry to build reagents that test it, and in vitro pharmacology to measure the result.
| Pillar | Contribution | Weakest point |
|---|---|---|
| Computational modeling | Predicted binding pose and ranked contact list | Predictions are hypotheses until tested |
| Peptide chemistry | Purified analogs that isolate each predicted contact | Long, aggregation-prone peptides are hard to make cleanly |
| In vitro pharmacology | Affinity and efficacy data that confirm or reject contacts | Cell-based assays do not capture whole-body physiology |
The computational pillar produces the hypothesis. Because the published account describes a modeled complex rather than an experimentally determined structure, the work began with in silico construction of the receptor. The routine approach is homology modeling: align the Y2 sequence to a class A GPCR of known structure, build the transmembrane bundle and extracellular loops, dock NPY with its C-terminal tail guided into the orthosteric site, and refine the complex with molecular dynamics simulation. The simulation repeatedly samples the complex and reports contact frequencies and interaction energies, and the most stable contacts rise to the top of the list. The output is a ranked list of predicted contacts: which NPY side chains sit opposite which Y2 residues, and which contacts plausibly pay for the binding energy. Before any synthesis is committed, those predictions should be checked against the known structure-activity relationships of NPY, above all the non-negotiable role of the C-terminal amide.
The peptide chemistry pillar turns each prediction into a testable reagent. The standard panel starts with the native 36-mer as a reference, then removes or alters one feature at a time: an alanine substitution at a predicted contact, a truncation from either terminus, a D-amino acid replacement, or a change to the C-terminal amide. The resulting analogs are made by solid-phase peptide synthesis , and NPY is a demanding sequence. At 36 residues it is long enough to aggregate on the resin, and the amphipathic helix is the usual cause: growing chains pack through their helical faces and stall coupling. Pseudoproline dipeptides and elevated-temperature or microwave-assisted coupling are the standard workarounds. The sequence also carries aspartate residues, particularly Asp11 in an alanine-proline context, that are prone to aspartimide and isoaspartate byproducts under the basic conditions of coupling and deprotection. The C-terminal Tyr36 amide, essential for activity, requires a Rink amide resin and a carefully chosen cleavage cocktail, and the two tyrosines at positions 20 and 21 need oxygen-free handling to avoid oxidation. Every analog must be purified by reversed-phase HPLC and verified by mass spectrometry, because a single deleted or rearranged residue in a 36-mer can change receptor pharmacology without producing an obvious impurity peak.
| Analog design | What it tests |
|---|---|
| Full-length NPY | Reference agonist; baseline affinity and efficacy |
| N-terminal truncation, e.g. NPY13-36 | Whether helix and tail preserve Y2 binding; selectivity versus Y1 |
| C-terminal truncation, e.g. NPY1-32 | Role of the tail and Tyr36 amide in activation |
| Alanine substitution at a predicted contact | Whether that side chain contributes to affinity |
| D-amino acid replacement | Backbone conformation and helix stability |
| C-terminal free acid instead of amide | Contribution of the amide to receptor recognition |
The in vitro pharmacology pillar measures what each analog does. Competitive binding against a labeled NPY-family ligand on cells or membranes expressing recombinant human Y2 reports affinity. A functional readout that captures Gi signaling, classically the inhibition of forskolin-stimulated cAMP accumulation, reports efficacy and separates agonists from neutral antagonists. An analog that loses affinity when a predicted contact is deleted confirms that the contact matters. An analog that loses efficacy while retaining affinity identifies residues involved in activation rather than recognition. The iteration between prediction and assay is the point: the final map is the set of contacts that survived experimental testing, not the original docking pose.
A laboratory planning this pipeline should commit to all three arms before the first synthesis run, because the project moves at the pace of the slowest one. The modeling should be finished first so the analog panel is designed against the hypothesis. The synthesis arm should be built for volume: many small batches, which favors an automated synthesizer over manual coupling. The pharmacology arm needs a stable cell line expressing the canonical human Y2 sequence, a validated labeled ligand, and functional controls: the native peptide as the standard agonist and a C-terminally truncated peptide as a non-activating control. Projects fail when any single arm is treated as an afterthought, because an untested model is speculation and an untested peptide is an expense.
With a validated contact map, design shifts from screening to engineering. The C-terminal pharmacophore, the amidated Tyr36 and the arginine residues that anchor the tail in the pocket, is the element to preserve or mimic, whether in a stabilized peptide, a peptidomimetic, or a small molecule. The alpha-helix contacts are the element to modify for selectivity and pharmacokinetics, through substitutions that tighten Y2 binding or chemical changes such as lipid conjugation, PEGylation, or D-amino acid replacement that resist proteolysis and extend half-life. Read backwards, the same map supports antagonist design: a compound that occupies the pocket and blocks the tail would prevent NPY from activating Y2. Whether the therapeutic goal is agonism or antagonism is a physiological decision, not a structural one.
The direction of effect deserves emphasis because the 2018 report did not settle it. The NPY system is orexigenic as a whole, but Y2 sits inside it as an inhibitory autoreceptor, and Y2-selective agonism has historically been the anti-obesity direction of interest because of PYY3-36. Yet Y2 also modulates gastrointestinal motility, vagal signaling, and adipose tissue, so a receptor-level map does not predict whole-body outcomes. A modulator that is perfect at the receptor can still fail through poor distribution, rapid metabolism, or an inability to reach the relevant neurons. The map answers the recognition question; it does not answer the pharmacology question.
The evidence base for this particular map is thin in specific ways. The account of the work is a short, commercially sponsored summary rather than a peer-reviewed methods paper, and it presents conclusions without primary data. No specific binding residues are named in it, no binding affinities are given, no compound structures appear, and no in vivo results are described. The map is therefore best treated as evidence that a predictive model was built and partially validated inside the collaborating laboratories, not as a structural blueprint that an outside group could reconstruct from the published text alone.
A researcher reading a report of this kind should ask three questions before building on it. First, were the predictions and the measurements made on the same receptor construct, and was that construct the canonical human Y2 sequence? Second, did the analog panel include proper controls, and were affinity and efficacy measured separately? Third, are the claimed contacts consistent with sequence conservation across the NPY family? A contact that sits in a region identical across Y1, Y4, and Y5 cannot explain Y2 selectivity, regardless of how well it scores in a model. These checks cost little and separate a credible interaction map from a decorated docking pose.
For an industrial buyer evaluating a Y2-focused program, the interaction map is useful mainly as evidence of a working model. The de-risking events that actually move a program forward are the same as for any peptide GPCR target: confirmed affinity in a human receptor assay, clean selectivity against the other three family members, functional efficacy in a Gi readout, and ultimately target engagement in an animal model of food intake. A report that offers a map but no quantitative pharmacology should be read as a milestone, not a proof point.
The questions left open by the 2018 map are as instructive as the map itself. In rough order of how quickly they could be answered:
The obesity drug environment has moved since 2018, and that changes the context for any Y2 program. Incretin-based therapies now dominate clinical practice, so a Y2 modulator would have to justify itself against a high bar in efficacy, tolerability, and delivery. The structural toolkit has also advanced: experimentally determined structures of peptide GPCR complexes are now routine, and a structure of NPY bound to Y2 would resolve which modeled contacts are real and give medicinal chemists a much firmer foundation than a validated homology model.
The honest summary is a qualified yes. A binding map of the NPY-Y2 interaction is exactly the kind of evidence that enables structure-guided modulator design, and the 2018 Uppsala and Novo Nordisk collaboration demonstrated a credible workflow for producing one. But the public record of that work is too thin to specify residues, affinities, or drug candidates, and the appetite circuitry leaves genuine uncertainty about which direction of modulation will help patients. The map is the right starting point. The drug remains the work.
Peptides referenced: Neuropeptide Y, Peptide YY (PYY), GLP-1.
Related reading: FGFR2-Targeting Peptide for Early Esophageal Adenocarcinoma Detection, Can a CXCR4 Peptide Reverse Ovarian Cancer Drug Resistance?, Stapled Antimicrobial Peptides: How They Work and What Is Proven, The GABRD-P2X2 Complex Behind Progesterone's Calcium Signal in Sperm.