No abstract

Magainins, a family of positively charged peptides, are partly if not wholly responsible for antimicrobial activity in skin extracts of Xenopus laevis. We report here that members of the magainn family-i.e., the 21-amino acid peptide PGLa and the 23-amino acid peptide magainin 2 amide (PGSa)-dissipate the electric potential across various energytransducing membranes and thus uncouple respiration from other free-energy-requiring processes. We propose that this is a likely mechanism for the antimicrobial effects of these compounds.The greater part of the transduction of free-energy from catabolism to anabolism (e.g., ATP synthesis) requires redox-linked proton pumps to generate an electric potential difference across a membrane (1, 2). As protons flow back through the H+-ATPase, the electric driving force (2) is applied to the synthesis of ATP from ADP and inorganic phosphate. Other membrane enzymes allow other thermodynamically uphill reactions (such as the uptake of substrates for metabolism) to be driven by such proton backflux.Agents that make membranes leaky towards protons or other ions have been shown to disrupt energy coupling (3). Among these are small synthetic molecules such as carbonylcyanide p-trifluoromethoxyphenylhydrazone (FCCP), which can shuttle protons across membranes, and mixed peptides such as the gramicidins, which form ion channels in biological membranes. The latter compounds are produced and excreted by prokaryotes as weapons against neighboring organisms. A third class of such weapons consists of "true" oligopeptides (in the sense that they consist only of the 20 "standard" amino acids in their L configuration). The colicins (4) are among these. It should be noted that not all of these compounds act primarily by depolarizing the membrane potential of their target cells; some seem to compromise the integrity ofthe outer rather than the inner membrane (4, 5).Lower eukaryotes also use such oligopeptide weapons. As a response to injection of bacteria, insects produce cecropins (6), also called sarcotoxins (7,8). This raised the question of whether vertebrates, and perhaps even mammals, use an oligopeptide-based defense mechanism in addition to their highly effective cellular and antibody immune responses. One such mechanism (presumably auxiliary to phagocytic and oxidative killing mechanisms), residing in rabbit and human leukocytes, consists of the so-called "defensins." These are cysteine-rich, lysine-poor oligopeptides of about 3.5 kDa in molecular mass (9,10).Recently, one of us (11, 12) purified antimicrobial activity from the skin of the amphibia Xenopus laevis as two 23-amino-acid-long positively charged peptides that lack cysteine but are rich in lysine. These peptides were called "magainins." Based on its primary structure (13) and activity (14), a similarly active 21-amino-acid-long oligopeptide called PGLa should also be considered a magainin. The primary structures of magainins (11) differ considerably from those of the peptides belonging to the other classes of amphibian...

In this paper, we present the implementation and results of new methodology based on linear algebra. The theory behind these methods is covered in detail in the Supporting Information, available electronically (Shrager and Hendler). In brief, the methods presented search through all possible forward sequential submodels in order to find candidates that can be used to construct a complete model for the BR-photocycle. The methodology is limited only to forward sequential models. If no such models are compatible with the experimental data, none will be found. The procedures apply objective tests and filters to eliminate possibilities that cannot be correct, thus cutting the total number of candidate sequences to be considered. In the current application, which uses six exponentials, the total sequences were cut from 1950 to 49. The remaining sequences were further screened using known experimental criteria. The approach led to a solution which consists of a pair of sequences, one with 5 exponentials showing BR* → Lf → Mf → N → O → BR and the other with three exponentials showing BR* → Ls → Ms → BR. The deduced complete kinetic model for the BR photocycle is thus either a single photocycle branched at the L intermediate or a pair of two parallel photocycles. Reasons for preferring the parallel photocycles are presented. Synthetic data constructed on the basis of the parallel photocycles were indistinguishable from the experimental data in a number of analytical tests that were applied.