Retinal proteins are transmembrane proteins with potential applications in optical memory and optogenetics for example. Proteorhodopsin (PR) is a recently discovered microbial retinal protein that acts as a proton pump. Activation of PR occurs when the covalently bound chromophore, retinal, absorbs a photon and undergoes all-trans → 13-cis isomerization. Two variants of PR are tuned to absorb blue (490 nm) and green (520 nm) light. This difference is modulated by a single residue at position 105 that acts as a light-tuning switch—in blue PR it is glutamine whereas in green PR it is leucine. Despite this basic knowledge of the color-tuning mechanism, the electronic environment in the retinal binding pocket that determines the green and blue variants of PR remains poorly understood. Using quantum mechanical/molecular mechanical calculations, we have characterized the color-tuning mechanism of PR at the B3LYP level of theory. We discovered that the difference between blue and green PR depends on a complex interplay between the water-mediated hydrogen-bonded network within the binding-pocket and the twist of the retinal polyene chain. Geometry optimization of the native form of blue PR produced two retinal structures, one with a planar and another with a slightly twisted polyene chain. In both structures a weak hydrogen bond exists between the color-tuning switch, Q105, and the polyene chain. However, the twisted structure is higher in energy, with an elongated N-H bond in the protonated Schiff base (PSB) and shorter hydrogen bond with a coordinated water. Mutation of the color-tuning switch (Q105L) abolishes the hydrogen bond with retinal, leading to a shortened N-H bond in the PSB and a lengthened hydrogen bond with water. Our results provide a detailed understanding of the color-tuning mechanism in PR and can be applied to other retinal proteins as well.