Don't know if I understand the question right, don't know if you want to get taxonids or protein sequences or both.

To get all the taxonid's of a certain group you can look at this file:

taxidlineage.dmp (ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/new_taxdump/) With something like the following command you can extract the taxonid's, for example all bacterial.

grep -w [taxonid of group] taxidlineage.dmp | awk '{print $1}' > taxonids_group
#get bacterial ids
grep -w 2 taxidlineage.dmp | awk '{print $1}' > taxonids_group

To get the protein sequences you can download the nr database, that update script comes with the blast program in the bin folder:

update_blastdb.pl nr --passive

To get the protein sequences you can do (not sure if this is the right full command):

blastdbcmd -db nr -entry_batch taxonids_group

EDIT:

Reading your question again, this is an option to get only taxonids which have proteins sequences. First download the nr database:

update_blastdb.pl nr --passive

Extract all taxonids:

blastdbcmd -db nr -entry all -outfmt '%T' > nr_taxonids

Get all taxonids of your interest:

wget ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/new_taxdump/new_taxdump.zip
unzip new_taxdump.zip 
grep -w [taxonid of group] taxidlineage.dmp | awk '{print $1}' > taxonids_group
#get bacterial ids
grep -w 2 taxidlineage.dmp | awk '{print $1}' > taxonids_group

Last step is to find a way to compare taxonids_group and nr_taxonids. I would do it with python but for sure you can find a command on this forum to better do it.