Enzyme activity is based on the geometry/architecture of its active site (this is essential for the binding of the substrate and the reaction to occur between substrate and side chains involved in the catalysis) and this geometry is dependent on the 3D structure of the protein. This is the concept of structure/function Relationship. If you perturb the structure of the enzyme you affect its capacity of catalysis.
As rightly said by Dominique, the structure of the protein is on what depends the activity. The amino acid sequence determines how the protein would fold and what structure it'll assume in 3D. That will align specific amino acid residues in a way to form clefts complementary to their interacting partners (in case of enzymes, the substrate). When the binding partner interacts to those amino acids, a chain of chemical reactions occur ultimately giving rise to the product. In case of non enzymatic reaction simple binding can bring about a conformation change and initiate a signalling pathway. But majorly proteins perform their functions depending on their structures. I'm not sure if the structure-function relation can be studied computationally but as Dominique suggested you can denature a protein (disrupt its structure, simply by heating it may be) and allow it to slowly renature and measure the gradual gain of its enzymatic activity.
Enzyme or for that matter any biomolecule function is directly related to structure. This can be demonstrated or observed by using several ways. Like site directed mutagenesis where it is possible to establish role of each group in determinig the function of given enzyme. Simulation or docking studies avoiding time consuming experiments are currently available.
You can carry out computational simulations using freeware like Autodock. By comparing population of clusters, binding energy, etc. you will be able to find a correlation between structure and reactivity.
When you make a simulation you have to set some parameters, amongst them is the number of "runs" the software will do. In each run the computer will find a conformation of enzyme-substrate that minimize the potential energy of the system. If, for example, you tell the program to make 100 runs you might see that 90 of those runs have the same energy, whereas the other ten have another (probably you will have more than one or two relative minimums). That, which indicates the probability of a certain conformation happening, is the cluster population.