This work was carried out using the Proteworks protein studio platform for protein design, https://studio.proteworks.com/.

First the 2LVQ structure of the target was manually inspected in the molecular viewer. The C-terminal beta strand was identified as a potential target for binding through beta sheet expansion, as well as the nearby cluster of hydrophobic residues (PHE 81, PHE103, and TYR 106). Model 13 of the NMR structure was selected due to the highest strand formation on the c-terminus.

We then ran RFdiffusion using the (pyrosetta free) implementation of interface conditioning in protein studio. It was run with the following settings:

{ "workflow_name": "rf-diffusion", "execution_id": "rf-diffusion-b70f5275", "submitted_at": "2026-03-26T00:58:07.084653+00:00", "files": { "input_pdb": "projects/RBX1/cleaned_2LGV.pdb" }, "parameters": { "config-name": "base", "num-designs": 1200, "contigs": "[A1-72/0 120-120]", "hotspot-res": "A6,A8,A13,A31,A32,A33,A34,A35,A36,A37,A38,A39,A40,A41,A42,A43,A44,A45,A46,A47,A48,A49,A50,A51,A52,A53,A54,A55,A56,A57,A58,A59,A63,A64,A65,A66,A67,A68,A69,A70,A71,A72", "gpu-type": "nvidia-tesla-t4", "deterministic": "false", "write-trajectory": "false", "recenter": "true", "align-motif": "true", "inference-cyclic": "false", "interface-tensor-target-adj": "A67-72", "interface-tensor-binderlen": "120", "interface-tensor-binder-ss-len": "6", "interface-tensor-binder-ss": "E" }, "output_folder": "projects/RBX1/Round 3" }

1200 backbones were generated in total which were then manually evaluated for the intended beta strand pairing and model quality.

2 sequences were then created for each backbone using solubleMPNN.