I've never done mySQL replication or clustering, so this post is just a "for what it's worth" post. If I were setting this up, and have clustering AND replication, then I would made the "cluster" master-master replicate and then the second "cluster" would be the slaves setup as master-master. So basically cluster CL1 has 2 mySQL servers, setup using master-master replication (so they both have the same data). Then for cluster #2, CL2, I'd setup as the "slave" servers that get replicated from CL1, but again, they would be setup as master-master replication so that all of THOSE in THAT cluster all have the same data, since it is a cluster.
However, this might not be how mySQL replication and clustering works. Hopefully I don't get a downvote, since I'm just giving you my 2 cents about how I would do it. Again, I've never done this and I've thought about it, but no needed to, and most of our companies databases are Oracle.