Truncated, branched, and/or cyclic neuropeptide Y (NPY) analogues were tested for their ability to bind to the neuroblastoma cells, SK-N-MC (Y1 receptor) and SK-N-BE(2) (Y2 receptor). The design of such analogues was inspired by models of NPY based on the crystal structure of avian pancreatic polypeptide. The minimum length of the backbone was investigated using the following truncated analogues [binding affinity (nM) for Y1 and Y2 receptor subtypes respectively are given in parentheses]: des-AA10-17[D-Ala9]NPY (100, 0.9), des-AA7-23[D-Ala6]NPY (> 1000, 1.2), des-AA4-26[D-Ala3]NPY (> 1000, 120), cyclo(7,20)-des-AA10-17[Glu7,D- Ala9,D-Dpr20]NPY (100, nd), cyclo(2,27)-des-AA7-23[Glu2,D-Ala6,D-Dpr27]NPY (> 1000, 3.6), cyclo(2,30)- des-AA7-23[Glu2,D-Ala6,-D-Dpr30]NPY (> 1000, nd), cyclo(1,30)-des-AA4-26[Glu1,D-Ala3,D-Dpr30]NPY (> 1000, > 1000). A new family of branched NPY analogues corresponding to the partial deletion of the polyproline helix with conservation of the N-terminus was also examined: des-AA7-23[(Ac-NPY14-22)-epsilon-D-Lys6]NPY (> 1000, 2.1), des-AA7-23[(Ac-NPY7-22)-epsilon-D-Lys6]NPY (> 1000, 5.1), des-AA7-23-[(Ac-LEALEG-NPY14-22)-epsilon-D-Lys6]NPY (> 1000, 4.8). Finally, the role played by the flexible tail (residues 32-36) was studied with the following cyclic analogues: cyclo(30,34)-[Lys30,Glu34]NPY18-36 (> 1000, 360), cyclo(30,34)-[Orn30,Gly34]NPY18-36 (> 1000, 950), cyclo(30,34)-[Dpr30,Glu34]NPY18-36 (> 1000, 590), cyclo(33,36)-[Lys33,Glu36]NPY (> 1000, > 1000), cyclo(33,36)-[Lys33,Glu36]NPY18-36 (> 1000, > 1000). These results suggest that the Y1 receptor is highly discriminatory since deletion of residues 10-17, shown to have little effect on Y2 binding affinity, reduces Y1 affinity 50-fold. Bridging sites and constructs have been identified that may serve as useful leads in the design of more potent and selective analogues. We have identified two positions (9 and 6) where the introduction of a D amino acid is not detrimental to binding affinity. Whether this modification leads to the stabilization of a yet unidentified turn compatible with high Y2 receptor affinity will have to be determined by spectroscopic methods. Finally, stabilizing a putative alpha-helical conformation of the C-terminal heptapeptide of NPY18-36 has a deleterious effect on the Y1 and Y2 receptors.