THE BARRIER FUNCTION OF THE GRAM‐NEGATIVE ENVELOPE

The envelope of gram-negative bacteria contains three structures that can be visualized in the electron microscope (FIGURE 1 ) and can now also be separated, which permits their analysis: an inner, or cytoplasmic, membrane that contains lipid and protein; a thin peptidoglycan layer; and an outer membrane that contains lipopolysaccharide (LPS), in addition to protein and lipid. Although gram-positive organisms also have a cytoplasmic membrane and peptidoglycan layer, the outer membrane is unique to gram-negative organisms and confers special properties upon them. This paper will summarize the evidence that the outer membrane provides a permeability barrier that makes gramnegative bacteria less permeable than gram-positive bacteria to a large variety of molecules, which includes many drugs and antimetabolites. This evidence is twofold: chemical, because it has been found that ethylenediaminetetraacetate (EDTA) both increases cell permeability and specifically alters the outer membrane by releasing LPS; and genetic, because some mutants altered in LPS structure exhibit permeability changes and others with increased permeability display outer membrane alteration. A tentative model of the structure of the outer membrane and the aspects of this structure that provide the permeability barrier will be presented. The barrier function of the outer membrane first became accessible to study when it was found that EDTA disrupts this function. In 1958, MacGregor and Elliker discovered that mutants of Pseudornonns nerziginosa resistant to quartenary ammonium compounds become sensitive when exposed to EDTA, and these workers correctly concluded that EDTA increased permeability to these compounds; this observation appears to have been ignored, probably because, as will be discussed below, EDTA alone is much more toxic to this organism than to other gram-negative bacteria. Soon thereafter, Repaske found that EDTA permits lysozyme action on several gram-negative bacteria.g Spheroplasts formed by this method are sensitive to actinomycin, a drug that normally cannot affect intact Escherichia coli:', 1 While pursuing this observation in my laboratory, I found that EDTA alone, without lysozyme, conferred subsequent sensitivity to actinomycin upon the cells,5 and I concluded that EDTA permits entry of the drug. This conclusion was verified by the finding that EDTA-treated, but not control, cells, could bind radioactive actinomycin and that other molecules normally excluded could also enter.'; The permeability increase is very general, as is expected for partial breakdown of a passive permeability barrier. Many substances normally excluded can enter many species of gram-negative bacteria after they are treated with EDTA (TABLES 1 & 2 ) . Of special interest to this conference is the wide range of drugs, unrelated either in their structure or their action, that can attack such