Ronald, the only time that I have seen a baby that is K+ typing as K- at birth (unless serial IUTs have been given of course), is a case where the mother had such a high titre anti-K that all of the baby's K antigen sites were blocked by the maternal anti-K, so I don't think that is the answer.
It could be that the baby is Ser193, which is a K antigen mutation that means the K antigen does not react with certain monoclonal anti-K reagents, and so is giving a false negative result. This mutation is incredibly rare, and so I doubt if that is the answer.
It could be that the mother has been carrying a bacteria that expresses a sort of "K antigen" on it's surface (Escherichia coli, Enterococcus faecalis, Morganella morganii and mycobacterium sp have eebm implicated) and the mum has made anti-K as a result, but this is another incredibly rare scenario, and so I doubt that is the answer.
"Naturally occurring" is very, very rare, so I doubt if that is the answer.
I'm with shily on this one. I think that it is a contaminant of the RhIG.