The Protein Data Bank (PDB) is a large and rapidly growing standard repository of complex biological macromolecules. Existing software tools for the PDB, including storage management, are quite inflexible and work in isolation. We are using object-oriented design and object-oriented database techniques in the construction of a toolbox for the PDB. The main components of the object-oriented toolbox include: (1) a uniform graphical user interface for fast, intuitive, and interactive access to the PDB; (2) several interoperable storage managers, compatible with present and future PDB formats; (3) a new set of efficient, modular, and flexible software tools. The toolbox is being implemented in C++. We describe our experience in the the implementation and use of the first prototype, PDBTool, to display, manipulate, and verify the structure of proteins. This project is a successful collaboration of a Biophysics and a Computer Science research and development team.