Characterising the interaction between Caenopore-5 and model membranes by NMR spectroscopy and molecular dynamics simulations

Abstract

Caenopore-5 (Cp5) is a pore-forming antimicrobial protein constitutively and exclusively expressed in the intestine of C. elegans, which protects the nematode from pathogens. The interaction between Cp5 and model membranes has been shown to be dependent on pH, salt concentration and lipid composition of the model membranes. Using site-directed mutagenesis, solution-state NMR spectroscopy and molecular dynamics simulations, this project provided residue-specific characterisations of the membrane binding activity of Cp5. Production of native and mutant Cp5 (both 15N-labeled and 15N-13C-labeled) was achieved using bacterial recombinant expression systems. The side chain pKa values of all acidic residues and three histidines were determined both in the absence and presence of 200 mM NaCl by solution-state NMR spectroscopy to gain insights into the pH-dependency of the membrane-binding activity of Cp5. Interestingly, only the pKa of E17 (5.34 and 5.21 in the absence and presence of 200 mM NaCl, respectively) was found to coincide with pH 5.2 where Cp5 was observed to possess high antimicrobial activity against bacteria [1]. This observation suggests that E17 modulates the pH-dependent binding of Cp5 to model membranes as a pH trigger. Among all acidic residues, only three residues (E17, E10 and E65) have elevated pKa values than their intrinsic values, suggesting that the protonation state of these three residues probably constitutes the first (E17), second (E65) and third (E10) switches of the pH dependent binding of Cp5 to model membranes. Both Cp5 and saposin C are SAPLIPs. While Cp5 acts as a membrane-disrupting antimicrobial protein, saposin C acts as an essential activator for glucosylceramidase for the hydrolysis of glucosylceramide and interacts with model membrane in a pH-dependent manner [2], similar to that of Cp5. Both sequence and structural alignments of Cp5 and saposin C revealed that K22 of Cp5 and K23 of saposin C were conserved. K23 of saposin C was reported previously to be indispensable in both the membrane fusion activity and the enzymatic activation of saposin C. Therefore, it was hypothesized that K22 is also important for the function of Cp5. To identify key residues involved in the pH-dependent activity of Cp5, a set of 1D 1H NMR experiments were run to measure the pKa for the binding of Cp5 to model membranes. With site-directed mutagenesis, the results confirmed the indispensable involvement of K22 in membrane-binding activity of Cp5, and identified a link between the protonation state of the E17 side chain and the pH-dependent binding of Cp5 to model membranes, as evidenced by an increase of 0.93 pH units for the apparent binding pKa value for the E17Q mutation. To further characterise the side chains of Cp5, ten 100 ns molecular dynamics simulations were performed for five different pH points in the absence and presence of 200 mM NaCl. In addition to the salt bridge and hydrogen bond analysis and accounting for the unusual pKa values for acidic residues, the analysis on the MD trajectories suggested that no salt bridge was formed between E17 and K22, an indication that the two residues act independently. In addition, this observation suggests that a change in the protonation state of E17 functions by shifting the surface charge balance of Cp5 to favor membrane interaction. Moreover, stable salt bridges formed between R1, E10 and E48 involved a dramatic movement for the R1 residue side chain. This has not been reported previously in structural studies. This result suggests the potential importance of the positively charged side chain of R1 in the electrostatic interaction and initial contact between Cp5 and model membranes. Additionally, the salt bridge analysis suggested a possible lock-key mechanism of H62 and E65. In this mechanism, the positively charged side chain of H62 is locked by the negatively charged side chain of E65 at pH values between the H62 pKa and E65 pKa, but released when the pH is lower than the E65 pKa where the side chain of E65 becomes protonated, and then the the positively charged side chain of H62 can function by interacintg with membrane head group moieties. Among the four saposins derived from the saposin precursor protein prosaposin, saposin C is the only one whose structure has been determined by NMR spectroscopy both in the absence and presence of SDS micelles. Nonetheless, the coordinates of the SDS micelle is absent in the PDB file for the saposin C-SDS micelle complex structure. Therefore, using a model SDS micelle and unambiguously assigned experimental NMR data describing the interaction interface between the micelle and sapsonin C, an saposin C-SDS micelle complex structure was built through a manual rigid-body docking process with the help of the Kabsch algorithm and criteria of minimum solvent accessible surface areas. With the complex structure of saposin C and SDS, it is possible to visualize the binding interface of saposin C and SDS. Moreover, K23 and K38 were found to act as two clips on the SDS micelle in the electrostatic interaction between their positively charged side chains and the negatively charged headgroups of the SDS molecules in the micelle. Given the structural homology and the functional similarity of Cp5 and saposin C (pHdependent membrane-interaction activity), a Cp5-SDS micelle complex structure was also built based on the saposin C-SDS micelle complex structure. In the subsequent pKa analysis by PROPKA, Y16 turned out to be the most outstanding residue among the eleven titrateable residues which experienced a pKa larger than 3 pH units. Among the eleven titrateable residues, Y16 was the only residue with hydrophobic side chain, and also the one with the largest pKa. More interestingly, Y16 sits right beside E17 in Cp5 structure, whose role has been experimentally confirmed in the membrane-binding activity of Cp5.