Interaction with the solvent plays a critical role in modulating the structure and dynamics of a protein. Because of the heterogeneity of the interaction strength, it is difficult to identify multi-scale structural response. Using a coarse-grained Monte Carlo approach, we study the structure and dynamics of a protein (H3.1) in effective solvent media. The structural response is examined as a function of the solvent-residue interaction strength (based on hydropathy index) in a range of temperatures (spanning low to high) involving a knowledge-based (Miyazawa-Jernigan(MJ)) residue-residue interaction. The protein relaxes rapidly from an initial random configuration into a quasi-static structure at low temperatures while it continues to diffuse at high temperatures with fluctuating conformation. The radius of gyration (Rg ) of the protein responds non-monotonically to solvent interaction, i.e., on increasing the residue-solvent interaction strength (fs ), the increase in Rg (fs ≤fsc ) is followed by decay (fs ≥fsc ) with a maximum at a characteristic value (fsc ) of the interaction. Raising the temperature leads to wider spread of the distribution of the radius of gyration with higher magnitude of fsc . The effect of solvent on the multi-scale (λ: residue to Rg ) structures of the protein is examined by analyzing the structure factor (S( q ),|q| = 2π/λ is the wave vector of wavelength, λ) in detail. Random-coil to globular transition with temperature of unsolvated protein (H3.1) is dramatically altered by the solvent at low temperature while a systematic change in structure and scale is observed on increasing the temperature. The interaction energy profile of the residues is not sufficient to predict its mobility in the solvent. Fine-grain representation of protein with two-node and three-node residue enhances the structural resolution; results of the fine-grained simulations are consistent with the finding described above of the coarse-grained description with one-node residue.