Low PLE but total available memory not being used

I am dealing with a performance issue on a production server that hosts many databases. Two of the databases are dominant on this instance consuming about 90% of the total memory in use. These databases are both read and write intensive which is not ideal but it is what I have inherited.

We are running SQL Server 2005 SP1.

As part of my troubleshooting I have noted that PLE is chronically low anywhere between 0 and 40 seconds pretty much continuously apart from a small 2 hour let up during the night.

One of the troubleshooting steps we took was to increase the maximum memory by 4GB (This instance is part of an active/active FCI so there is more physical memory available to the node). I expected to see SQL Server gobble up the extra memory but instead it grabbed an extra 500MB or so and wouldn't go any higher but the PLE remained low and disk activity remained high along with lazy writes.

We then tried forcing SQL server to take up the extra memory by setting min memory 2GB higher than total memory used but it remained resolute and didn't budge.

I am using sys.dm_os_buffer_descriptors to identify how much data is in the buffer pool.

Has anyone come across this behaviour before and have any suggestions for troubleshooting steps to identify the route cause of the memory behaviour I am seeing?