Mutants with a single Cys residue in place of Phe27, Pro28, Phe29, Phe30, or Pro31 at the periplasmic end of putative transmembrane helix I were used to study the interaction of lactose permease with ligand by site-directed chemical modification or fluorescence spectroscopy. With permease embedded in the native membrane, mutant Phe27 --> Cys or Phe28 --> Cys is readily labeled with [C-14]-N-ethylmaleimide (NEM), while mutant Phe29 --> Cys, Phe30 --> Cys, or Phe31 --> Cys reacts less effectively. beta,D-Galactopyranosyl 1-thio-beta,D-galactopyranoside (TDG) has little or no effect on the reactivity of Phe27 --> Cys, Phe29 --> Cys, or Phe30 --> Cys permease. Remarkably, however, Pro31 --> Cys permease which is essentially unreactive in the absence of ligand becomes highly reactive in the presence of TDG. Ligand also enhances the NEM reactivity of the mutant with Cys in place of Pro28 which is presumably on the same face of helix I as position 31. The five single-Cys mutants which also contain a biotin acceptor domain in the middle cytoplasmic loop were purified by monomeric avidin-affinity chromatography in dodecyl beta,D-maltoside and subjected to site-directed fluorescence spectroscopy. Mutants Phe27 --> Cys, Phe29 --> Cys, and Phe30 --> Cys react rapidly with 2-(4'-maleimidylanilino)naphthalene-6-sulfonic acid (MIANS), and reactivity is not altered in the presence of TDG. In striking contrast, mutants Pro28 Cys and Pro31 --> Cys react extremely slowly with MIANS in the absent of ligand, and TDG dramatically enhances reactivity. Finally, when each mutant is labeled with N-(1-pyrenyl)maleimide and reconstituted into phospholipid vesicles, 5-doxylstearic acid is shown to quench fluorescence more effectively with pyrene-labeled Phe30 --> Cys permease than with pyrene-labeled Pro28 --> Cys or Pro31 --> Cys permease. Taken together, the results indicate that the face of helix I with Pro28 and Pro31 is buried within the permease in the absence of ligand and becomes more reactive as the result of a ligand-induced conformational change. On the other hand, the opposing face of helix I with Phe30 is likely to be in contact with the lipid phase of the membrane and unaffected by ligand binding.