It is well known that the MST boundary switch replicates the IST BPDU (instance 0 BPDU) on all VLANs to send them to a PVST+ neighbor. The MST switch does this because PVST+ switch _expects_ to receive a BPDU on every VLAN.
The question is: Does this mean that replicated BPDUs contain exactly the same information? What about extended system id in the RootID field? Does the MST boundary switch set extended sys-id in the PVST+ BPDU to 1, 0 or a VLAN number?
This is important thing to know in case the root is located outside of the MST region (i.e. the PVST+ switch is the root).
For example, suppose we have a triangle - 1 MST switch and 2 PVST+ switches.
One of the PVST+ switches has priority 0+1 for VLAN 1 and 0+10 for VLAN 10 (0 + ext. sys-id) and is the root for both VLAN 1 and 10.
This (first) root PVST+ switch sends VLAN 10 BPDU to the MST boundary switch with priority 0+10.
If the MST switch _always_ set priority to 0+1 in the PVST+ BPDUs replicated to the second PVST+ switch that switch will think that the root is accessible thru the MST region.
Then the second PVST+ switch will send the BPDU to the first PVST+ switch and this switch will trust it. It will think it is no longer the root for VLAN 10 (although it will see its own MAC in the root BID field!).
This will lead to a conflict: something like "the root for VLAN 1 is outside of the MST region and root for VLAN 10 is inside the MST region". The MST boundary port will be blocked as the "root inconsistent port".