That's huge, unless that's a typo. 1.5 megabases is ~1/3 the size of the E. coli genome, and about the size of some fertility plasmids (F plasmids). I understand these are transferred between appropriate E. coli strains by conjugation, but lack the expertise to advise on how to perform this protocol.
Strip the cell wall to prepare spheroplasts (N-actylmuramidase treatment), followed by electroporation.
If the above proves ineffective or lethal, condense the DNA to nanoparticles and then zap them into spheroplasts. It goes without saying that spheroplasts will have to be treated gently with care or the lysis could be too prevalent.
As Brian said, this would be a really really large plasmid. You could transfer it between strains by conjugation if it already exists. If you are proposing to make it, then this will be technically challenging because it is hard to manipulate DNA of this size and leave it intact.
However there are methods out there for the building and transformation of large DNA molecules (although not usually this large). I would review the literature for the creation of BAC libraries.