This's an interesting question. You can do this using different methodologies.

1- Download all approved PDB records for ligand-protein interactions. In the ACID database, you can find them, though this online tool is not as fast in performing docking analysis as you expected. Next, you can use VINApso to perform virtual and inverse docking.

2- You can use Raccoon 1 or 2 for docking a ligand against multiple receptors. In this case, you should provide the exact coordination of your grid box before docking analysis. If you want to know how your ligand blindly interacts with your target, you should increase the size and dimension of your docking box to cover all receptor surfaces.

3- You can also perform a text-mining search using R packages to identify relevant targets in your bacterial proteome based on previous studies. Then you can download or remodel the target receptors and use them for docking analysis. This step is a time-consuming phase but it will return high-quality results. During performing this step, please check the content of PDBbind and COvPDB databases to find relevant data.

4- Please see this post, too: https://www.researchgate.net/post/Is_there_any_free_software_available_for_inverse_docking_or_docking_where_we_are_searching_for_a_receptor_or_protein_for_a_known_ligand2

Regards

Check ChEMBL for activity data. Consider also compounds similar to yours.