Some time ago, a piece in this blog dealt with the complexities of the Gram stain. In brief, it said: "In a simpler world, all bacteria with a single membrane, the monoderms, would be Gram-positive, all those with two (the diderms), Gram-negative... Hah! Were it that simple! Each of these groups includes outliers that stain opposite from the expected way." ... "Most monoderms fall conveniently in the Gram-positive camp all right, but there are important exceptions. First, the Firmicutes (so called for the "firmness" of their thick peptidoglycan cell wall) encompass the wall-less mycoplasmas and their allies, bacteria that cannot be stained by the Gram stain at all, which makes the point moot here. Next, some typical Firmicutes (the Negativicutes) stain Gram-negative. To make things worse, at least one firmicute, Acetonema, appears to have the prototypical Gram-negative property, an outer membrane. Having a thin peptidoglycan layer may place whole phyla within the Gram-negatives, despite being monoderms."
The question arises: Do the phages that infect the Negativicutes come from the Gram-positives or the Gram-negatives? The answer, as presented in a recent paper, is clear. They are derived from the Gram-positive Firmicutes, as revealed both by their sequences and their genome organization (the genomic position of their lysis module). The only exception is a Mu-like prophage, which is characteristic of Proteobacteria. Conclusion: these phages and their bacteria co-evolved, despite the mobility that phages possess.