Bacitracin is a metal-dependent peptide antibiotic from cultures of Bacillus subtilis and B. licheniformis, directed primarily against Gram-positive bacteria (1).  It is commercially produced in large quantities worldwide as a feed additive for livestock (2) and in human medicinal ointments like Neosporin^R and Polysporin^R (3).  This antibiotic is a mixture of many closely related analogues in which bacitracin A1 is the major component with the highest activity (4).  Bacitracin A1 contains a thiazoline ring formed by the condensation of the Ile-1 carboxylate with the –NH₂ and –SH groups of Cys-2, a cyclic heptapeptide structure formed via an amide linkage between Lys-6 side chain and the C-terminus, and four D-amino acids (Figure 1A).  These unusual structural features may protect this peptide from degradation by proteases (5).
Bacitracin requires a divalent metal ion such as Zn(II) for its potent antimicrobial activity (6) and can form a 1:1 complex with other divalent metal ions, including Co(II), Ni(II), and Cu(II) (7).  An early NMR study of Zn(II)-bacitracin suggested that His-10 and the thiazoline ring sulfur were coordinated to the metal (8). A later electron paramagnetic resonance (EPR) study on Cu(II)-bacitracin suggested that the thiazoline ring nitrogen, the His-10 imidazole, and the carboxylate of Glu-4 and Asp-11 were the ligands (7b).  The results from a recent extended X-ray absorption fine structure (EXAFS) study of Zn(II)-bacitracin in the solid form suggested a tetrahedral-like geometry with the thiazoline nitrogen, His-10 imidazole, Glu-4, and the N-terminal amino group as ligands (9), corroborating some previous observations.  However, a conclusive metal binding mode was not drawn from these studies, and the structure of metal-bacitracin complexes could not be determined from these studies.  For example, the tetrahedral geometry in solid may not retain in solution, the ligand Glu-4 cannot be conclusively assigned, the binding status of His-10 (through Ne or Nd) and the amino group (bound or not) cannot be concluded, the pyrophosphate binding moieties are not identified, and the configuration of the groups other than the coordinated ligands cannot be revealed.

    The metal(II)-bacitracin complex was found to bind tightly to C₅₅-isoprenyl pyrophosphate (K_f = 1.05 x 106 M^–1) (10), which presumably prevents the lipid pyrophosphate from being dephosphorylated by a membrane pyrophosphatase.  Since the mono-phosphate form of the lipid is required to bind UDP-sugars for transport during cell wall synthesis (11), the binding of metal-bacitracin to the lipid pyrophosphate thus inhibits N-glycosylation of nascent proteins in the lumen of the endoplasmic reticulum which serves as the key step in the inhibition of cell wall synthesis by bacitracin.  Although a Co(II)-bacitracin-pyrophosphate ternary complex has been proposed (12), detailed binding and structural information on this ternary complex and other metal-bacitracin complexes was not presented.  Furthermore, a structure-activity relationship for this antibiotic has not yet been conclusively defined.

    We have recently perfomed a structural analysis of the paramagnetic Co(II) complexes of the antibiotic bacitracin A₁ and analogues by means of NMR spectroscopy.  High-spin paramagnetic Co(II) has been utilized as a very sensitive “probe” for characterizing the structure and ligand interactions of metalloproteins via the assignment of the isotropically shifted signals (13) and also serves as a nearly perfect substitute for the Zn(II) in most zinc proteins (14).  The isotropically shifted ¹H NMR signals are attributable to protons on the ligands or the moieties near the metal, which contain structural information about the metal-binding site. Thus, Co(II) is chosen as the prototypical metal ion for the study of the structure and function of metal-bacitracin complexes. The ligand-binding mode and metal-binding environment of Co(II)-bacitracin have been fully established in our study, and a structural model is built.  In addition, several bacitracin analogues have also been purified or prepared and their Co(II) binding mode determined, which allowed us to suggest a structure-activity relationship of this antibiotic.

References

1. Brewer, G. A., & Florey, K.; Eds. (1980) Analytical Profiles of Drug Substances 9, 1–69.
2. (a) Hanson, D. J. (1985) Chem. Eng. News 63, 7-11.  (b) Yagasaki, T. (1986) Food Hyg. Soc. Jpn. 27, 451.
3. Arky, R. (1997) Physicians’ Desk Reference for Nonprescription Drugs 18th Ed. Medical Economics Company, Montvale, NJ.
4. (a) Morris, M. (1994) Biol. Mass Spect. 23, 61–70.  (b) Siegel, M. M., Huang, J., Lin, B., & Tsao, R. (1994) Biol. Mass Spect. 23, 196–204.
5. Pfeffer, S., Hohne, W., Branner, S., Wilson, K., & Betzel, C. (1991) FEBS Lett. 285, 115–119.
6. Storm, D. R., & Strominger, J. L. (1974) J. Biol. Chem. 249, 1823–1827.
7. (a) Scogin, D. A., Mosberg, H. I., Storm, D. R., & Gennis, R. B. (1980) Biochemistry 19, 3348–3342.  (b) Seebauer, E. G., Duliba, E. P., Scogin, D. A., Gennis, R. B., & Belford, R. L. (1983) J. Am. Chem. Soc. 105, 4926–4929.
8. Mosberg, H. I., Scogin, D. A., Storm, D. R., & Gennis, R. B. (1980) Biochemistry 19, 3343–3347.
9. Drablos, F., Nicholson, D. G. & Ronning, M. (1999)  Biochim. Biophys. Acta 1431, 433–442.
10. Storm, D. R., & Strominger, J. L. (1973) J. Biol. Chem. 248, 3940–3945.
11. Higashi, Y., Siewert, G., & Strominger, J. L. (1970) J. Biol. Chem. 245, 3683.
12. Stone, K. J., & Strominger, J. L. (1971) Proc. Nat. Acad. Sci. USA 68, 3223–3227.
13. (a) Bertini, I., & Luchinat, C. (1984) Adv. Inorg. Biochem. 6, 71–111.  (b) Bertini, I., & Luchinat, C. (1986) NMR of Paramagnetic Molecules in Biological Systems Benjamin/Cummings, Menlo Park, CA. (c) Ming, L.-J. (2000) in Physical Methods in Bioinorganic Chemistry, Spectroscopy and Magnetism, Que, L., Jr., Ed.; University Science Books, CA.
14. Bertini I., & Luchinat C. (1994) in Bioinorganic Chemistry, Bertini, I., Gray, H. B., Lippard, S. J., & Valentine, J. S.; Eds., University Science Books, CA, 1994, Chapter 2.