CRYSTAL STRUCTURE AND THE SPECIFICITY DETERMINING REGIONS OF Cry1Ac, A d-ENDOTOXIN FROM Bacillus thuringiensis
Dean Derbyshire and Jade Li
MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH, U.K.
Keywords: X-Ray Crystal Structure, d-Endotoxin, Bacillus thuringiensis, Specificity
Cry1Ac is a member of the Cry family of insecticidal, membrane pore -forming d-endotoxins produced by the Gram-positive bacterium Bacillus thuringiensis (Bt)[1]. These toxins are expressed as inactive protoxins and become proteolytically activated in the midgut of the target insect. Cry1Ac displays specific toxicities towards the larvae of lepidopteran insects. Its membrane receptor, present on the larval midgut epithelium, has been identified as the glycoprotein aminopeptidase N (APN)[2]. N-acetylgalactosamine (GalNAc) has been shown to be part of the receptor and in in vitro experiments GalNAc competitively inhibits the binding of Cry1Ac to APN[3].
We have crystallised trypsin-activated Cry1Ac from Bt ssp. HD-73 in spacegroup P1 with the unit cell dimensions a=51.73A, b=113.06A, c=123.34A, a=113.11o, b=91.49o, and g=110.46o, and four molecules in the unit cell. Self-Patterson and self-rotation searches indicate that the four molecules pack as two pairs: the two pairs are related by the translation vectors x=30.1A, y=1.1A, z=63.8A, and within each pair the molecules are related by an approximate, non-crystallographic two-fold rotation about an axis almost parallel to the a axis. The structure has been solved to 2.36A resolution by molecular replacement from the coordinates of Cry1Aa (PDB code 1CIY) followed by four-fold molecular averaging and solvent flattening. The initial model, built into the averaged map, had an R-factor of 35.7%, and 81.2% of its main chain dihedral angles are in the most favourable regions of the Ramachandran plot. The atomic model has been refined to an R-factor of 22.4% and a R-free of 28.2% with excellent geometry. The current model comprises a continuous polypeptide chain from residues 33 to 611 for each of the four molecules and a total of 389 bound solvent molecules.
Cry1Ac displays the same three domain topology of the Cry toxin family first seen in Cry3A[4]. The Cry1Ac structure is very similar to Cry1Aa in the alpha helical bundle (domain I), where their sequences differ at only two positions. The rms distance between Ca positions of Cry1Aa and Cry1Ac after superposition is only 0.36A for all 221 residues in this domain. In the beta prism domain (domain II) Cry1Ac is largely similar to Cry1Aa over the sheet regions but two of its three apical loops (loops 2 and 3) display distinct conformations. The rms distance between superimposed Ca's, excluding these loop residues is 0.61A. These localised structural variances correlate well with sequence exchange and mutagenesis data which implicate residues in loops 2 and 3 in determining the specificity, particularly in the initial toxin binding to the midgut membranes[5].
The most significant structural differences between Cry1Aa and Cry1Ac are found in the beta sandwich domain (Domain III) where only 91 out of 151 Caís can be superimposed and these have an rms distance of 0.82A. In particular, Cry1Ac shows an insertion relative to Cry1Aa of six residues which form part of a protruding loop. This sequence insertion is unique to Cry1Ac among the Cry toxins. Sequence exchange of a short segment which included the observed insertion loop was reported to affect the specific toxicity of Cry1Ac and abolish the GalNAc-sensitive binding of Cry1Ac to APN in ligand blots; moreover alanine substitutions which map within the observed insertion were found to reduce specific toxicity and binding to the insect membrane[6].
In conclusion, the structurally distinct regions, both at the
apex of domain II and around the protruding insertion of domain
III, are the most likely regions of Cry1Ac to interact directly
and specifically with its membrane receptor.
1. H. Hofte and H. R. Whiteley, Microbiol
Rev. 53 (1989) 242 - 255
2. P. Knight, N. Crickmore and D. Ellar, Mol. Microbiol.
11 (1994) 429 - 436
3. B. Knowles, W. Thomas and D. Ellar, FEBS letts. 168
(1984) 197 - 202
4. J. Li, J. Carroll and D. Ellar, Nature 353
(1991) 815 - 821;
P. Grochulski et al, J. Mol. Biol. 254
(1995) 447 - 46
5. D. Smedley and D. Ellar, Microbiol. 142 (1996)
1617 - 1624; H. E. Schnepf, K. Tomczak, J. P. Ortega and H. R.
Whiteley, J. Biol. Chem. 265 (1990) 20923 - 20930
6. R. DeMaagd et al., Proceedings of the 30th Annual
Meeting for the Society for Invertebrate Pathology, Banff,
Canada, 1997; M. K.Lee, F. Gould and D. H. Dean ibid