Tight junction pores are physiological gatekeepers of paracellular transport in epithelial tissues. Conventionally, tight junction permeability is determined via in vitro electrophysiology measurements; however, the macroscopic readout does not provide molecular-level understanding into the mechanism of ion permeation. Insight into the factors governing selectivity across the paracellular space is just emerging. In this study, we investigated tight junction pores comprising of claudin-2 and claudin-5 proteins that are structurally similar to subnanometer radii but have measurably different in vitro ion permeabilities. To evaluate the mechanistic differences in ion transport across the pores, we computed the free-energy profiles and relative rate constants for the transport of monovalent (Na+, K+, Cl-) and divalent (Mg2+ and Ca2+) ions through the pores using replica exchange metadynamics. In claudin-2, we demonstrate how a single residue dictates selective permeability of Na+ and K+ ions. In claudin-5, we found no clear preference for anion or cation selectivity; thus, pores formed by claudin-5 are indeed barriers to ion permeation. Mutations to claudin-5 that widen the pore's steric radius did not significantly impact pore selectivity, indicating that electrostatics dominate pore selectivity. The key takeaways from this work are as follows: (a) two pores that are similar in diameter and length can have dissimilar ion conductance, (b) existence of a physical pore does not guarantee ion permeability, and (c) the electrostatic environment created by the pore-lining residues dictates the ion conductivity. These mechanistic understandings of the tight junction pores are critical for the interpretation of tight junction physiology.