It is often difficult to reliably determine charges on larger model compounds. For this reason, we perform the actual charge optimization on small model compounds (like those methylphosphates) and transfer them to the larger compounds. There are rules for doing so; for instance, if we delete a hydrogen to form a bond with another atom, we most commonly sum the charge on the hydrogen into its parent atom. I haven't looked at it very closely, but from a distance, it appears that all the phosphate charges in the nucleic acid force field were derived from the methylphosphate charges using this rule. If you want to be sure, you could have a look at the nucleic acid force field paper. If you construct topologies for your desired compounds using the same rules, there's no reason why they should be less valid than our topologies...
If this is sufficient (and apparently quite easy to do),
I think it is, but double-check the paper to be sure.
is there a reason why protonation states of these nucleotides have not been added or at least patches included in the topology sets?
No idea, but my best guess is that at some time, the decision was taken to publish the force field - sitting in front of the computer and parameterizing more model compounds without end won't get any bills paid. The published force field contains the most common compounds and protonation states, as testified by the fact that it's widely used.