This particular topology has a couple of different Cisco CallManager clusters. It is best to use a centralized H.323 gateway as a bridge between the two clusters.
H.323 is a peer-to-peer protocol. As long as there is enough memory and CPU resources to support two clusters, there are no problems assigning multiple clusters to the same H.323 gateway.
You cannot register multiple route groups, route plans or translation patterns to the same gateway. Aside from those restrictions, everything should work fine.
When two clusters have the same gateway addresses and the same route pattern, install intercluster trunking. This uses the H.323 protocol instead of a common H.323 gateway.
For more information on setting up intercluster trunking, refer to these documents: