Ah, I think I understand.
You can pass the "search" function any geometry. The sample uses a polygon. You need a point. The JavaScript version has a point buffer and it will return a collection. Once you get the collection you need to select the one you want. Once you have the feature you want, most people create a graphics layer and add a new graphic based on the features geometry and your "Selected" symbology.
Attach to "OnClick"
Pass query function the Point created from the mouse click and a buffer.
The return will be a collection.
Select the one you want if there are multiple items under that point.
Create graphics layer
Create new graphic with selected features geometry
Add new graphic to graphic layer. This will "Cover up" the original symbology with your "selected" version. [You might just be able to change the symbology of the feature itself and save all the above.]
