What about creating a phylogeny of the protein sequences and color coding the branches by phenotype?

Something like this:

Draw two graphs, one of top of the other, at the same width:

1. A sequence logo showing the amino acid sequence; substitutions will be visualized as lower information or higher entropy residues.
2. A bar or violin plot showing measured phenotypic change at each residue, aligned to the residues in the sequence logo.

Repeat for each amino acid sequence; tile the figures if they can be put into a group.

You may also be able to use the "genotype network" formalism, creating a separate network for each phenotype. In a genotype network, a vertex represents a sequence and edges connect vertices if their corresponding sequences are separated by a single, small mutation. For more information on this formalism, see Wagner (2008) Nature Reviews Genetics, 9:965-974. For the foundations of this formalism, see Schuster et al. (1994) Proc. R. Soc. Lond. B 255:279-284.