Kainate receptor agonist efficacy is thought to be controlled by the degree of closure of the ligand-binding domain (LBD) of individual receptor subunits (1). However, we have suggested the LBD dimer interface plays a more prominent role (2). To distinguish between these two models we studied a mutant KAR (Y521C/L783C GluK2) that restricts the dimer interface by disulfide bond formation. If agonist efficacy is controlled by the dimer interface, cross-linking should affect open-channel probability and/or single channel conductance. However, if efficacy is governed by the LBD’s of individual subunits, these properties should be unchanged in Y521C/L783C. Cross-linking the dimer interface caused an attenuated peak response. Mutant GluK2 peak responses were 85 ± 18 pA (n = 13) compared to the wild-type (WT) peak of 2.3 ± 0.6 nA (n = 16). Poor surface expression of Y521C/L783C did not substantially contribute to this difference, as assessed using pH-sensitive GFP-tagged WT (n=11) and mutant (n=3) GluK2 receptors. Stationary noise analysis showed that the mean weighted unitary single channel conductance of Y521C/L783C (3.9 ± 0.3 pS, n = 8) was much lower than the reported WT main open state (~27 pS) (3), suggesting that reduced channel conductance accounted for the small macroscopic responses. Discrete channel events were recorded in response to fast application of saturating (10 mM) glutamate (WT, n=5, mutant, n=4) and also under equilibrium conditions (both, n=4). Consistent with noise analysis experiments, events recorded from Y521C/L783C were restricted to 2.4 (65 %) and 4 pS (35 %). In addition to these small conductances, WT receptors could also access larger conductances under equilibrium conditions (up to 20 pS) as well as the main open state (~25 pS) recorded in response to fast agonist application. Both WT and Y521C/L783C spent most of their time shut as the predominant shut time was the longest lived shut component (WT, τ =1.5 s, 81 %, mutant, τ =1.7 s, 54 %). Interestingly the small channel conductance of Y521C/L783C is a result of the double cysteine mutations inability to access the GluK2 main open state. Taken together these data highlight the importance of the LBD dimer interface in regulating KAR agonist efficacy.