You grow the cultures of the 4 strains and prepare the lysate then centrifuge to clear off the cell debris. The supernatant contain all the soluble fraction of the proteome. If you need also insoluble fraction then, you dissolve the pellet either in 8M Urea or Guanidine-HCl solution.

All the best!

If you already have the Fasta files (proteins or complete genome), you can identify the core using some bioinformatics tools.

http://orthomcl.org/common/downloads/software/v2.0/

http://omabrowser.org/oma/current/

http://www.bioinfocabd.upo.es/of

In these cases, the core would be the groups with homologous sequences (with the 4 bacterial strains).

However, you could use a software to analyze pangenome ...

http://bioinformatics.oxfordjournals.org/content/early/2014/01/10/bioinformatics.btu017.full.pdf

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3268234/pdf/btr655.pdf