Role of N-terminal amino acids in the potency of anthrax lethal factor

Abstract

Anthrax lethal factor (LF) is a Zn(+2)-dependent metalloprotease that cleaves several MAPK kinases and is responsible for the lethality of anthrax lethal toxin (LT). We observed that a recombinant LF (LF-HMA) which differs from wild type LF (LF-A) by the addition of two residues (His-Met) to the native Ala (A) terminus as a result of cloning manipulations has 3-fold lower potency toward cultured cells and experimental animals. We hypothesized that the "N-end rule", which relates the half-life of proteins in cells to the identity of their N-terminal residue, might be operative in the case of LF, so that the N-terminal residue of LF would determine the cytosolic stability and thereby the potency of LF. Mutational studies that replaced the native N-terminal residue of LF with known N-end rule stabilizing or destabilizing residues confirmed that the N-terminal residue plays a significant role in determining the potency of LT for cultured cells and experimental animals. The fact that a commercially-available LF preparation (LF-HMA) that is widely used in basic research studies and for evaluation of vaccines and therapeutics is 3-fold less potent than native LF (LF-A) should be considered when comparing published studies and in the design of future experiments.

Figure 1. Schematic representation of various LF proteins.

All proteins (except LF-A/St which is wild type LF produced from Sterne strain) were generated as secreted proteins from B. anthracis BH450. All proteins contained a signal peptide that is cleaved by signal peptidases during secretion. Constructs with a factor Xa recognition sequence alone or preceded by the Myc epitope tag are labeled “/X” or “/MyX”, respectively. These proteins were produced as precursor proteins followed by cleavage with factor Xa protease to generate the indicated N-termini. Residue (Z) circled in red indicates the residue mutated for each protein, i.e., Z = A, H, M, R, F, G in different mutated LF constructs.

Figure 2. Cytotoxicity of LF proteins to RAW264.7 cells.

RAW264.7 cells were incubated with various concentrations of LF-HMA, LF-A, and LF-A/X proteins (A) or LF proteins with mutated N-termini (B) and a fixed concentration of PA (250 ng/ml). Cell viability was assessed at 3 h. Percent viability was calculated relative to cells treated with medium (no toxin).

Figure 3. Toxicity of LF proteins to mice.

(A) LF-HMA and LF-A/St (100 µg) were injected in Balb/cJ mice in combination with 100 µg PA via the IP route and survival was monitored for 120 h. Each group contained n = 5 mice. (B) LF proteins (100 µg) were injected in Balb/cJ mice in combination with 100 µg PA via the IP route and survival was monitored for 120 h. Mouse numbers used in this experiment were as follows: LF-HMA (n = 9), LF-A (n = 9), LF-A/X (n = 5), LF-G/MyX (n = 3) and all other groups n = 4. (*) The animals in LF-G/MyX and LF-R/MyX groups exhibited substantial malaise starting at 24 h and throughout the experiment and were euthanized at 128 h to prevent suffering in accordance with approved animal protocols. (C) LF preparations (40 µg) were injected in Balb/cJ mice in combination with 40 µg PA via the IP route and survival was monitored for 120 h. Mouse numbers used in this experiment were as follows: LF-HMA (n = 14), LF-A (n = 9), LF-A/X (n = 5) and all other groups n = 4. (**) The surviving animals in these groups did not display any signs of malaise over the last 48 h of the experiment.

Figure 4. Evaluation of the apparent binding affinities of LF-HMA and LF-A for PA using Schild Plot analyses.

CHO WTP4 cells were incubated with various concentrations of FP59 plus a set concentration of PA (12 nM) and different concentrations of LF-HMA (A) or LF-A (B) for 3 h. After toxin removal, cells were incubated with the toxin-free medium containing 10 mM NH₄Cl for 48 h before assessment of cell viability. Schild Plot analyses were performed as described in “Methods” to assess K _ds of each LF protein for PA. Inserts shown in panels A and B are non-linear regression curves obtained from these analyses using GraphPad Prism.