The molecular basis of polyspecificity of Mdr1p, a major drug/H+ antiporter of Candida albicans, is not elucidated. We have probed the nature of the drug-binding pocket by performing systematic mutagenesis of the 12 transmembrane segments. Replacement of the 252 amino acid residues with alanine or glycine yielded 2/3 neutral mutations while 1/3 led to the complete or selective lost of resistance to drugs or substrates transported by the pump. Using the GlpT-based 3D-model of Mdr1p, we roughly categorized these critical residues depending on their type and localization 1°/ main structural impact ("S" group), 2°/ exposure to the lipid interface ("L" group), 3°/ buried but not facing the main central pocket, inferred as critical for the overall H+/drug antiport mechanism ("M" group) and finally 4°/ buried and facing the main central pocket ("B" group). Among "B" category, 13 residues were essential for the large majority of drugs/substrates, while 5 residues were much substrate-specific, suggesting a role in governing polyspecificity (P group). 3D superposition of the substrate-specific MFS Glut1 and XylE with the MDR substrate-polyspecific MdfA and Mdr1p revealed that the B group forms a common substrate interaction core while the P group is only found in the 2 MDR MFS, distributed into 3 areas around the B core. This specific pattern has let us to propose that the structural basis for polyspecificity of MDR MFS is the extended capacity brought by residues located at the periphery of a binding core to accommodate compounds differing in size and type.