The evolutionary origins of Middle East respiratory syndrome (MERS) coronavirus (MERS-CoV) are unknown. Current evidence suggests that insectivorous bats are likely to be the original source, as several 2c CoVs have been described from various species in the family Vespertilionidae Here, we describe a MERS-like CoV identified from a Pipistrellus cf. hesperidus bat sampled in Uganda (strain PREDICT/PDF-2180), further supporting the hypothesis that bats are the evolutionary source of MERS-CoV. Phylogenetic analysis showed that PREDICT/PDF-2180 is closely related to MERS-CoV across much of its genome, consistent with a common ancestry; however, the spike protein was highly divergent (46% amino acid identity), suggesting that the two viruses may have different receptor binding properties. Indeed, several amino acid substitutions were identified in key binding residues that were predicted to block PREDICT/PDF-2180 from attaching to the MERS-CoV DPP4 receptor. To experimentally test this hypothesis, an infectious MERS-CoV clone expressing the PREDICT/PDF-2180 spike protein was generated. Recombinant viruses derived from the clone were replication competent but unable to spread and establish new infections in Vero cells or primary human airway epithelial cells. Our findings suggest that PREDICT/PDF-2180 is unlikely to pose a zoonotic threat. Recombination in the S1 subunit of the spike gene was identified as the primary mechanism driving variation in the spike phenotype and was likely one of the critical steps in the evolution and emergence of MERS-CoV in humans.IMPORTANCE Global surveillance efforts for undiscovered viruses are an important component of pandemic prevention initiatives. These surveys can be useful for finding novel viruses and for gaining insights into the ecological and evolutionary factors driving viral diversity; however, finding a viral sequence is not sufficient to determine whether it can infect people (i.e., poses a zoonotic threat). Here, we investigated the specific zoonotic risk of a MERS-like coronavirus (PREDICT/PDF-2180) identified in a bat from Uganda and showed that, despite being closely related to MERS-CoV, it is unlikely to pose a threat to humans. We suggest that this approach constitutes an appropriate strategy for beginning to determine the zoonotic potential of wildlife viruses. By showing that PREDICT/PDF-2180 does not infect cells that express the functional receptor for MERS-CoV, we further show that recombination was likely to be the critical step that allowed MERS to emerge in humans.