Selection of novel ligands from a whole-molecule randomly mutated C5a library

Development of Synthetic Human and Mouse C5a: Application to Binding and Functional Assays In Vitro and In Vivo

The complement activation peptide C5a is a key mediator of inflammation that is associated with numerous immune disorders. C5a binds and activates two seven-transmembrane receptors, C5aR1 and C5aR2. Experimentally, C5a is utilized to investigate C5a receptor biology and to screen for potential C5aR1/C5aR2 therapeutics. Currently, laboratory sources of C5a stem from either isolation of endogenous C5a from human serum or most predominantly via recombinant expression. An alternative approach to C5a production is chemical synthesis, which has several advantages, including the ability to introduce non-natural amino acids and site-specific modifications whilst also maintaining a lower probability of C5a being contaminated with microbial molecules or other endogenous proteins. Here, we describe the efficient synthesis of both human (hC5a) and mouse C5a (mC5a) without the need for ligation chemistry. We validate the synthetic peptides by comparing pERK1/2 signaling in CHO-hC5aR1 cells and primary human macrophages (for hC5a) and in RAW264.7 cells (for mC5a). C5aR2 activation was confirmed by measuring β-arrestin recruitment in C5aR2-transfected HEK293 cells. We also demonstrate the functionalization of synthetic C5a through the introduction of a lanthanide chelating cage to facilitate a screen for the binding of ligands to C5aR1. Finally, we verify that the synthetic ligands are functionally similar to recombinant or native C5a by assessing hC5a-induced neutrophil chemotaxis in vitro and mC5a-mediated neutrophil mobilization in vivo. We propose that the synthetic hC5a and mC5a described herein are valuable alternatives to recombinant or purified C5a for in vitro and in vivo applications and add to the growing complement reagent toolbox.

A system for the directed evolution of the insecticidal protein from Bacillus thuringiensis

Theoretically, the activity of AB-type toxin molecules such as the insecticidal toxin (Cry toxin) from B. thuringiensis, which have one active site and two binding site, is improved in parallel with the binding affinity to its receptor. In this experiment, we tried to devise a method for the directed evolution of Cry toxins to increase the binding affinity to the insect receptor. Using a commercial T7 phage-display system, we expressed Cry1Aa toxin on the phage surface as fusions with the capsid protein 10B. These recombinant phages bound to a cadherin-like protein that is one of the Cry1Aa toxin receptors in the model target insect Bombyx mori. The apparent affinity of Cry1Aa-expressing phage for the receptor was higher than that of Cry1Ab-expressing phage. Phages expressing Cry1Aa were isolated from a mixed suspension of phages expressing Cry1Ab and concentrated by up to 130,000-fold. Finally, random mutations were made in amino acid residues 369-375 in domain 2 of Cry1Aa toxin, the mutant toxins were expressed on phages, and the resulting phage library was screened with cadherin-like protein-coated beads. As a result, phages expressing abnormal or low-affinity mutant toxins were excluded, and phages with high-affinity mutant toxins were selected. These results indicate that a method combining T7 phage display with selection using cadherin-like protein-coated magnetic beads can be used to increase the activity of easily obtained, low-activity Cry toxins from bacteria.

Comparative agonist/antagonist responses in mutant human C5a receptors define the ligand binding site

The C terminus is responsible for all of the agonist activity of C5a at human C5a receptors (C5aRs). In this report we have mapped the ligand binding site on the C5aR using a series of agonist and antagonist peptide mimics of the C terminus of C5a as well as receptors mutated at putative interaction sites (Ile(116), Arg(175,) Arg(206), Glu(199), Asp(282), and Val(286)). Agonist peptide 1 (Phe-Lys-Pro-d-cyclohexylalanine-cyclohexylalanine-d-Arg) can be converted to an antagonist by substituting the bulkier Trp for cyclohexylalanine at position 5 (peptide 2). Conversely, mutation of C5aR transmembrane residue Ile(116) to the smaller Ala (I116A) makes the receptor respond to peptide 2 as an agonist (Gerber, B. O., Meng, E. C., Dotsch, V., Baranski, T. J., and Bourne, H. R. (2001) J. Biol. Chem. 276, 3394-3400). However, a potent cyclic hexapeptide antagonist, Phe-cyclo-[Orn-Pro-d-cyclohexylalanine-Trp-Arg] (peptide 3), derived from peptide 2 and which binds to the same receptor site, remains a full antagonist at I116AC5aR. This suggests that although the residue at position 5 might bind near to Ile(116), the latter is not essential for either activation or antagonism. Arg(206) and Arg(175) both appear to interact with the C-terminal carboxylate of C5a agonist peptides, suggesting a dynamic binding mechanism that may be a part of a receptor activation switch. Asp(282) has been previously shown to interact with the side chain of the C-terminal Arg residue, and Glu(199) may also interact with this side chain in both C5a and peptide mimics. Using these interactions to orient NMR-derived ligand structures in the binding site of C5aR, a new model of the interaction between peptide antagonists and the C5aR is presented.

Characterisation of C5a receptor agonists from phage display libraries

C5a des-Arg(74) has a 10- to 100-fold lower receptor binding affinity than intact C5a and is only a partial agonist. We have used phage display selection from randomly mutated C5a des-Arg(74) libraries to isolate variant proteins that can activate C5a receptors with similar potency to C5a. Here we explore the interactions of three variants (V1-3) with C5aR mutated at residues involved in the differential response. The mutant Asp(282)Arg-C5aR is preferentially activated by C5a des-Arg(74), probably due to repulsion between Arg(74) of C5a and the substituent Arg(282). In accordance with this hypothesis, V2 (with a polar C-terminus which has no Arg residue) but not V1 (with a C-terminal Arg residue at position 73) could activate Asp(282)Arg-C5aR. V3, with a very hydrophobic C-terminus, was the most potent agonist at Asp(282)Arg-C5aR. Arg(175) is a potential counterion for the C-terminal carboxylate of C5a. C5aR mutated to either Ala or Asp at this position lost nearly all responsiveness to both C5a and C5a des-Arg(74), suggesting that mutation of Arg(175) caused a non-specific loss of receptor conformation and a loss of signalling capacity. However, V3 could still activate Arg(175)Asp/Ala-C5aR with the same potency as wild-type C5aR, demonstrating that the mutant receptors retained high signalling capability and showed a specific loss of responsiveness. Thus C5a des-Arg(74) variants produced by phage display are potentially useful tools for the dissection of ligand-receptor interactions.

Tc-99m-labeled C5a and C5a des Arg74 for infection imaging

The complement anaphylatoxin C5a and its natural metabolite C5a des Arg(74) (C5adR) are involved in several stages of the inflammatory process. Both act on a common receptor expressed on different cell types, including neutrophils and monocytes. The receptor binding affinity of C5a is in the nanomolar range and exceeds that of C5adR by 1-2 orders of magnitude. The biologic potency of C5a is considerably higher than that of C5adR. Here we tested both proteins labeled with (99m)Tc for imaging of infection.

METHODS

The proteins were labeled with (99m)Tc via the hydrazinonicotinamide (HYNIC) chelator. The preparations were tested for imaging of infection in a rabbit model of intramuscular infection. Biodistribution of the radiolabel was determined by gamma-camera imaging and by counting dissected tissues at 5 h p.i.

RESULTS

C5a and C5adR showed in vivo abscess uptakes of 0.12 and 0.025%ID/g, abscess/muscle ratios of 76 and 14, abscess/blood ratios of 9.1 and 2.6, and ROI derived target-to-background ratios of 5.9 and 2.1, respectively at 5 h p.i.

CONCLUSION

For infection imaging (99m)Tc-labeled C5a showed excellent in vivo characteristics. However, C5a is a very bioactive protein, impeding its clinical use as an infection imaging agent. The naturally occurring partial agonist C5adR has less biological effect but showed suboptimal imaging characteristics. The present study showed that for adequate localization of a receptor binding ligand affinities for the receptor in the nanomolar range are required.