We generated SARS-CoV-2 variants resistant to three SARS-CoV-2 main protease (M) inhibitors (nirmatrelvir, TKB245, and 5h), by propagating the ancestral SARS-CoV-2 in VeroE6 cells with increasing concentrations of each inhibitor and examined their structural and virologic profiles. A predominant E166V-carrying variant (SARS-CoV-2), which emerged when passaged with nirmatrelvir and TKB245, proved to be resistant to the two inhibitors. A recombinant SARS-CoV-2 was resistant to nirmatrelvir and TKB245, but sensitive to 5h. X-ray structural study showed that the dimerization of M was severely hindered by E166V substitution due to the disruption of the presumed dimerization-initiating Ser1'-Glu166 interactions. TKB245 stayed bound to M, whereas nirmatrelvir failed. Native mass spectrometry confirmed that nirmatrelvir and TKB245 promoted the dimerization of M, and compromised the enzymatic activity; the Ki values of recombinant M for nirmatrelvir and TKB245 were 117±3 and 17.1±1.9 µM, respectively, indicating that TKB245 has a greater (by a factor of 6.8) binding affinity to M than nirmatrelvir. SARS-CoV-2 selected with 5h acquired A191T substitution in M (SARS-CoV-2) and better replicated in the presence of 5h, than SARS-CoV-2. However, no significant enzymatic or structural changes in M were observed. The replicability of SARS-CoV-2 proved to be compromised compared to SARS-CoV-2 but predominated over SARS-CoV-2 in the presence of nirmatrelvir. The replicability of SARS-CoV-2 surpassed that of SARS-CoV-2 in the absence of 5h, confirming that A191T confers enhanced viral fitness. The present data should shed light on the understanding of the mechanism of SARS-CoV-2's drug resistance acquisition and the development of resistance-repellant COVID-19 therapeutics.