If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported by grants from the Natural Sciences and Engineering Research Council of Canada, the Alberta Heritage Foundation for Medical Research, the Canada Foundation for Innovation, and the Canada Research Chair in Bioorganic and Medicinal Chemistry (to J. C. V.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ¶ To whom correspondence may be addressed. Tel.: 315-787-2279; Fax: 315-787-2284 ‖ To whom correspondence may be addressed. Tel.: 780-492-5475; Fax: 780-492-8231
Mattacin is a nonribosomally synthesized, decapeptide antibiotic produced by Paenibacillus kobensisM. The producing strain was isolated from a soil/manure sample and identified using 16 S rRNA sequence homology along with chemical and morphological characterization. An efficient production and isolation procedure was developed to afford pure mattacin. Structure elucidation using a combination of chemical degradation, multidimensional NMR studies (COSY, HMBC, HMQC, ROESY), and mass spectrometric (MALDI MS/MS) analyses showed that mattacin is identical to polymyxin M, an uncommon antibiotic reported previously in certain Bacillus species by Russian investigators. Mattacin (polymyxin M) is cyclic and possesses an amide linkage between the C-terminal threonine and the side chain amino group of the diaminobutyric acid residue at position 4. It contains an (S)-6-methyloctanoic acid moiety attached as an amide at the N-terminal amino group, one d-leucine, six l-α,γ-diaminobutyric acid, and threel-threonine residues. Transfer NOE experiments on the conformational preferences of mattacin when bound to lipid A and microcalorimetry studies on binding to lipopolysaccharide showed that its behavior was very similar to that observed in previous studies of polymyxin B (a commercial antibiotic), suggesting an identical mechanism of action. It was capable of inhibiting the growth of a wide variety of Gram-positive and Gram-negative bacteria, including several human and plant pathogens with activity comparable with purified polymyxin B. The biosynthesis of mattacin was also examined briefly using transpositional mutagenesis by which 10 production mutants were obtained, revealing a set of genes involved in production.
nuclear Overhauser effect
double quantum filtered correlation spectrometry
proton-proton correlation spectrometry
high performance liquid chromatography
matrix-assisted laser desorption ionization
tandem mass spectrometry
rotational nuclear Overhauser effect spectrometry
total correlation spectrometry
two-dimensional transferred NOE
tryptic soy agar
tryptic soy broth
To survive in the natural environment and compete with other microorganisms for resources, many bacteria produce antimicrobial compounds to inhibit or kill other competing strains, including human and animal pathogens. One subclass of these antimicrobial compounds is the antibacterial peptides, which can be divided into two categories based on the biosynthetic pathways by which they are generated. One group consists of gene-encoded, ribosomally synthesized peptides (bacteriocins) that typically have 30–60 residues, may be either unmodified or extensively post-translationally altered (i.e.lantibiotics), and are active against closely related bacteria (
). Peptides in the second class are nonribosomal in origin and are produced by a series of condensations catalyzed by specific nonribosomal peptide synthetases using a templated multienzyme mechanism (
). These synthetases are large, multifunctional proteins composed of different modules, each of which has different domains capable of performing one step in the condensation of an amino acid onto a growing peptide chain (
), for novel peptidic antimicrobial agents. During a screening program we found that a Paenibacillus kobensis strain isolated from a soil/manure sample produced an active modified peptide with broad activity against both Gram-positive and Gram-negative organisms, including a number of human and animal pathogens. We now report that the structure of the principal antimicrobial compound formed by this strain, mattacin, is identical to polymyxin M, an uncommon antibiotic reported previously in the Russian literature (13–16; all reports in Russian) (Fig. 2). Although polymyxins represent one of the earliest classes of commercially important antibiotics to be identified (
), only polymyxin B is currently widely used and studied. In addition to efficient production and purification of mattacin, the present study describes its NMR solution structure and transfer NOE1 determination of conformational changes that occur upon binding to lipid A and compares these with previous results reported by others (
) with polymyxin B. Isothermal titration calorimetry was also employed to compare the binding of mattacin and polymyxin B to lipopolysaccharide (LPS), the major antigen of the outer membrane of Gram-negative bacteria. Finally, the biological potency of mattacin was assessed compared with that of polymyxin B, and the biosynthesis of mattacin was briefly examined.
In recent years a large number of antimicrobial peptides from Gram-positive bacteria have been discovered, including ribosomally produced bacteriocins (
). Paenibacillus spp. are Gram-positive, spore-forming bacteria from which polymyxins have been isolated, and their in vitro biosynthesis by cell-free enzyme systems has been successfully demonstrated (
), the genetic control of their biosynthesis has not been described. We utilized the Tn917 transposon to study the mattacin (polymyxin M) biosynthesis genes and screened ∼7,000 colonies for production mutants.
The majority of the production mutants exhibited decreased levels of production as opposed to a complete loss of production. This is not consistent with other studies using Tn917 for iturin and fengycin biosynthesis genes (
). In those studies, the transposon disrupted the nonribosomal peptide synthetase genes and blocked the antimicrobial peptide synthesis completely. The identified genes in our study show low homology to other known genes in the NCBI data base. The biosynthesis genes of mattacin in Paenibacillus spp. may be quite different from other peptide synthases in DNA sequence, which may be the result of different evolutionary gene origin. Despite all of the differences, our study did show some similarities between mattacin biosynthesis and the synthetic processes of other peptide antibiotics, including the use of nonribosomal peptide synthetases for peptide growth and ABC transporters for secretion.
Structural analysis of mattacin utilized a combination of chemical degradations, mass spectrometry, and multidimensional NMR analyses to obtain primary sequence, identity of the lipid side chain, connectivity, stereochemistry, and conformational preferences. As is common for small peptides, mattacin shows considerable conformational flexibility when pure in solution, as shown by the absence of extensive long range NOE interactions. Recent NMR investigations by Pristovsek and Kidric (
) have shown that in solution, two of the ring-amide protons in polymyxin B participate in intramolecular hydrogen bonding. This is detected by the temperature dependences of the amide proton chemical shifts; specifically, amide protons that are least affected by temperature change are likely shielded from solvent by participation in H bonding. In mattacin, it was observed that the amide protons of A2bu-8 and the side chain of A2bu-4 are both shielded, as in polymyxin B. These results suggest that mattacin and polymyxin B, though containing structural differences, are both highly flexible in solution and can adopt similar conformations.
It has long been believed that polymyxins elicit their bactericidal effects by binding to and disrupting the action of LPS, the major antigen of the outer membrane of Gram-negative bacteria (
). LPS contains three major structural components: lipid A, a core oligosaccharide, and an outer polysaccharide composed of repeating hetero-oligosaccharide subunits. Lipid A is a hydrophobic, lipid-rich moiety that harbors the endotoxic principle of LPS and is the most highly conserved part of the structure, containing two glucosamines, two phosphate esters, and six fatty acid chains (
). The proposed binding model for the polymyxin-LPS conjugate involves ionic interactions between the side chain amino groups of the polymyxin peptide cycle (positively charged in acidic medium) and the negatively charged phosphate groups of the lipid A disaccharide (
) recently investigated the preferred conformation(s) induced in polymyxin B when in solution with LPS from E. coli. Their results indicate that although the peptide is highly flexible, the side chain of A2bu-8 and the γ-amide NH of A2bu-4 as well as the amide NH of A2bu-8 with the side chain of A2bu-4 are in close proximity while bound to LPS. In performing these experiments with mattacin and the same LPS preparation, however, we did not detect the same correlations. Our conformational model in fact is somewhat different from that of polymyxin B (Fig. 8). Although the side chains of all three A2bu residues contained in the heptacycle were on the opposite side of the molecule from the hydrophobic Phe and Leu side chains in the polymyxin B structure, the bend in the mattacin heptacycle was in the opposite direction. This resulted in a less dramatic separation of hydrophobic and hydrophilic side chains. Whether this was because of our experiments not being able to detect the critical NOEs seen for polymyxin B or whether the conformational differences have an explanation resulting from the structural differences is not clear. Mattacin was significantly different at residues 6 and 7 (d-Leu, l-Thr) compared with the corresponding residues in polymyxin B (d-Phe,l-Leu). It was in this region of the heptacycle which the greatest conformational variation in the two models was observed, and it seems reasonable to suggest that the significantly diverse hydrophobic/hydrophilic residues present in the two peptides contribute to these differences.
To ascertain whether mattacin behaved in a manner similar to polymyxin B in its interaction with LPS, isothermal titration calorimetry was employed. The thermodynamics of polymyxin B binding to LPS have been investigated previously using isothermal titration calorimetry by the group of Surolia (
). In these studies, a highly processed LPS preparation from E. coli was used (extensive treatments with proteases, chelating agents, and purifications via dialysis and size exclusion chromatography). Using these conditions, which lead to smaller fragments of LPS, Surolia and co-workers obtained results supporting a simple 1:1 binding between polymyxin B and LPS. We chose to use an LPS from the same E. coli strain, but without extensive processing because our primary interest was the comparison of mattacin with polymyxin B and their respective binding to LPS. The same stock LPS solution and concentrations were used for isothermal titration calorimetry experiments with both mattacin and polymyxin B. As seen in Fig. 9, the binding isotherms for both peptide titrations were almost indistinguishable. The data do not support a simple binding model, and the ORIGIN analysis software could only fit the data using a complex sequential binding model. Initial interaction of the polymyxins with unprocessed LPS appears to lead to disruption of tertiary structural arrangements, thereby exposing additional lipid A binding sites. The simple 1:1 binding of polymyxin B seen in previous experiments (
) with highly processed lipid A may be caused by the predominance of smaller unconglomerated units. Whether this form provides a more accurate model of what occurs with living bacterial cells remains uncertain. However, our results clearly suggest that mattacin binds to LPS in a manner similar to that of polymyxin B.
Results from the activity assays for both mattacin and polymyxin B showed that the two peptides have virtually indistinguishable spectra of activity. Mattacin did appear to be slightly more active in most cases, but these differences are not very significant (less than 1 order of magnitude). An interesting observation in the activity assays was that all strains of Listeria and Bacilluswere inhibited by the live cells of P. kobensis M, but the purified polymyxins B and M had no effect. This suggests that another compound(s) is produced by this organism which is either lethal toListeria and Bacillus on its own or acts in synergy with another compound(s), possibly the polymyxin, to elicit its killing effects. Many multiple component bacteriocin systems are now known to be produced by Gram-positive organisms and have been reviewed recently (
). Future investigations aimed at determining whether P. kobensis M produces other novel antimicrobial compounds and at elucidating the details of polymyxin biosynthesis are in progress.
We thank Michael Carpenter (Department of Biochemistry, University of Alberta) for peptide sequencing and amino acid analysis; Bernd Keller and Liang Li (Department of Chemistry, University of Alberta) for use of the QSTAR MALDI MS instrument; Albin Otter (Department of Chemistry, University of Alberta) for extensive assistance with NMR studies; and Tara Sprules (Department of Chemistry, University of Alberta) for aid in the NOE modeling.