Based on the structure of the cyclosporin A / cyclophilin complex, a series of 2-substituted cyclosporins have been synthesised as cyclophilin ligands. Their in vitro evaluation demonstrates that although there is a large lipophilic pocket offering potential for affinity improvement, minor steric interactions or conformer distribution in aqueous solution play a dramatic role. Cyclosporin A (CsA) 1 is a natural, highly lipophilic, cyclic undecapeptide whose discovery has allowed the enormous rate of success that organ transplantation enjoys today. Its mechanism of action involves inhibition of T cell activation by blocking the transcription of a family of early activation genes, including those of the principal T cell growth factors IL-2 and IL-4. It is now clearly established that the inhibitory properties of CsA are mediated through binding to a soluble 16.5 kDa amino acid protein receptor and rotamase enzyme, cyclophilin (CyP). In the past two years, the structure of the CsA/CyP complex has been thoroughly investigated by NMR, X-ray and modelling studies. However, it was only very recently that important structural details were revealed independently by two groups, based either on crystallographic studies of a decameric CsA/CyP crystal or heteronuclear three dimensional NMR spectroscopy. It has thus been unambiguously demonstrated that CsA binds to CyP with the side chains of MeBmt-1, Abu-2, Sar-3, MeLeu-9, MeLeu-10 and MeVal-11 and that there are four direct hydrogen bonds involving MeBmt-1(CO), Abu-2(NH), MeLeu-9(CO), and MeLeu-10(CO) (Fig.1). In search for CsA derivatives having improved affinity for CyP, we were attracted by the fact that the side chain of Abu-2 points into a large hydrophobic pocket containing Gln-111 whose δ-CO is potentially available for a hydrogen bond (the amide NH₂ is most probably involved in a H-bond with Glu-63, 3.2Å away) (Fig.1). Moreover, two naturally occurring immunosuppressive cyclosporins substituted at position 2, CsC (Thr-2-CsA) 2 and CsG (Nva-2-CsA) 3 bind well to CyP indicating that there is available space within the lipophilic pocket (Table). Therefore, the synthesis of (4R)-OH-Nva-2-CsA 4 was undertaken in order to test the H-bond hypothesis (Gln-111(CO)…O= 3.2Å), that of (3R)-OH-Nva-2-CsA 5 as a CsC/CsG hybrid and that of (γ,γ-dichloro Abu)-2-CsA 6 and (cyclopropyl Ala)-2-CsA 7 as sterically demanding cyclosporins. The synthetic strategy in all four cases consisted of the preparation of the appropriate, N-protected amino acid and its coupling with the decapeptide, H-Sar-MeLeu-Val-MeLeu-Ala-(D)-Ala-MeLeu-MeLeu-MeVal-MeBmt-OMe 8, followed by deprotection and ring closure to give the desired cyclosporin derivative. Characterisation of the in vitro biological activity of all the compounds included both their affinity to CyP as well as their immunosuppressive potential (Table). The relative binding activity of derivatives was readily determined under competitive ELISA systems using protein-conjugated ligand CsA bound to a solid phase, and biotinylated CyP as specific recognition structure. The immunosuppressive activity was measured with the help of a reporter gene assay (IL2 RGA) and a mixed lymphocyte reaction (MLR). From the results obtained, it can be concluded that disubstitution at the γ-carbon of Abu-2 in native CsA 1 is detrimental for the affinity of the compounds to CyP. Indeed, molecular graphics investigations indicate that there might be some unfavorable steric interaction between the receptor's backbone and the introduced groups depending on the N-Cα-Cβ torsional angle (for Abu -49°). It can also be that the modifications influence the distribution of the various conformers in aqueous solution leading to a decreased CyP-bound population. Hydrophilicity might govern these equilibria and therefore the differences observed may not necessarily reflect effects of binding.