Abstract Protein synthesis is essential to cells and requires a constant supply of nutrients. Amino acid starvation leads to accumulation of uncharged tRNAs that promote ribosomal stalling, which is sensed by the protein kinase Gcn2, together with its effector proteins, Gcn1 and Gcn20. Activation of Gcn2 phosphorylates eIF2, leading to a global repression of translation. Fine-tuning of this adaptive response is performed by the Rbg2/Gir2 complex, which is a negative regulator of Gcn2. Despite the wealth of biochemical data, structures of Gcn proteins on the ribosome have remained elusive. Here we present a cryo-electron microscopy structure of the yeast Gcn1 protein in complex with stalled and colliding 80S ribosomes. Gcn1 interacts with both 80S ribosomes within the disome, such that the Gcn1 HEAT repeats span from the P-stalk region on the colliding ribosome to the A-site region of the lead ribosome. The lead ribosome is stalled in a non-rotated state with peptidyl-tRNA in the A-site, uncharged tRNA in the P-site, eIF5A in the E-site, as well as Rbg2/Gir2 located in the A-site factor binding region. By contrast, the colliding ribosome adopts a rotated state with peptidyl-tRNA in a hybrid A/P-site, uncharged-tRNA in the P/E-site and Mbf1 bound adjacent to the mRNA entry channel on the 40S subunit. Collectively, our findings provide a structural basis for Rbg2/Gir2 repression of Gcn2, and also reveal that colliding disomes are the substrate for Gcn1 binding, which has important implications not only for Gcn2-activated stress responses, but also for general ribosome quality control (RQC) pathways.