Results and Discussion

Phage Biopanning-based Discovery of Peptide Affinity Ligands for AdP

The commercially available NEB PhD-12 phage display library was chosen for phage biopanning to identify peptide candidates that bind to AdP. As described in the methods section 2.2, three rounds of phage biopanning (Figure 1) were performed against AdP. Once the biopanning rounds were completed, 20 distinct phage clones from the final eluate were randomly picked and sequenced.
The sequencing yielded 10 unique peptides with multiple repeats obtained for two of the peptides (Table 1).

Fluorescence Polarization Screening for Peptide Ligands

The 10 lead candidates were then synthesized with a 5(6)-FAM fluorescence label and fluorescence polarization (FP) was carried out to measure the dissociation constants (Kd) of their interactions with AdP. The experiments yielded binding curves (Figure 2), which were fitted to a 4-parameter logistic equation. Table 2 lists the Kd values obtained after fitting for 6 out of the 10 peptides, which showed an appreciable increase in observable signal with increase in AdP concentration. Peptides P3, P5, P6 and P8 were either too impure (as determined by C18-RP-UPLC) or showed a very low change in signal with increase in AdP concentration. From the binding curves shown in Figure 2, out of the 6 peptides, P10 exhibited the strongest binding to AdP.
Appropriate control experiments performed with free 5(6)-FAM dye in solution incubated with AdP indicated no non-specific binding between the dye and AdP. The P10 peptide was also tested against AdI (a different Adnectin provided by Bristol-Myers Squibb) and minimal binding was observed (Kd > 100 µM) indicating that P10 was indeed specific to AdP.
Since P10 exhibited the lowest Kd, and hence the strongest binding (Kd = 19.5 μM), it was carried forward for the development of a peptide-ELP based affinity precipitation process.

Development of Peptide-ELP-based Purification of AdP

Binding and Capture of AdP using P10-ELP

Initial binding experiments between P10-ELP and AdP were performed over a range of molar ratios of P10-ELP to AdP at pH 7.4. It can be observed in Figure 3 that at lower ratios, the capture of AdP increased linearly with the ratio, but the amount of AdP captured saturated at higher ratios, plateauing at a maximum of 85%. This could be due to the moderate affinity of the peptide-AdP interaction, which precludes 100% binding.
In order to improve the binding of P10-ELP to AdP, the effect of pH was examined, using pH conditions of 3.0, 4.0 and 5.0 (Figure 4a). P10-ELP was incubated with AdP (at a ratio of ~2:1 of P10-ELP to AdP) and precipitated at these conditions. The results were compared with those obtained at pH 7.4 (Figure 3) demonstrating that the capture improved by nearly a factor of 2 with pH 4.0 resulting in the highest capture of AdP in the precipitate. Since pH 4.0 and 5.0 are close to the pI of AdP, these conditions could result in precipitation of AdP. However, controls performed with 0.33 M sodium sulfate at pH 4.0 and 5.0 with 40 μM AdP (~1 mg/ml) resulted in no product precipitation. Binding and capture experiments were repeated at pH 4.0 to examine the impact of molar ratios of P10-ELP to AdP. As can be seen in Figure 4b, while a similar trend was observed for the effect of molar ratio on product capture, a higher capture was obtained at pH 4.0 for all ratios, plateauing at approximately 94% at 8:1.

Elution of AdP from P10-ELP

A range of pH conditions were evaluated for their efficacy at eluting AdP from the P10-ELP affinity reagent. The results indicated that while low pH did not result in any elution, higher pH values were able to elute up to 40% of the AdP. In order to improve the efficacy of the elution step, mobile phase modifiers were included in the eluent. Sodium chloride, ethylene glycol and arginine were selected due to their different mechanisms of disrupting protein-ligand interactions. Sodium chloride is widely employed in bioseparation techniques such as ion exchange and affinity chromatography (Lee & Chen, 2001) and operates via charge shielding and ion binding to reduce electrostatic interactions (Tsumoto et al., 2007). Ethylene glycol has been shown to weaken hydrophobic interactions while also increasing electrostatic interactions (Hou and Cramer, 2011). In fact, Kelley and co-workers have used ethylene glycol as an eluent in peptide-based affinity purification of Factor VIII (Kelley et al., 2004). Arginine is widely used as an elution modifier for several chromatographic systems. Although the mechanism of interaction of arginine is not fully understood, molecular dynamics simulations (Shukla et al., 2011) and experiments (Hirano et al., 2014) have suggested that arginine interacts with both aromatic as well as positively and negatively charged moieties leading to the mediation of both electrostatic and hydrophobic interactions. Applications have included both multimodal (Holstein et al., 2011) as well as protein A systems (Ejima et al., 2005).
The use of elution buffers with modifiers in our affinity precipitation system requires that the P10-ELP-AdP complex remain soluble during elution while also enabling subsequent precipitation of the P10-ELP upon addition of sodium sulfate. The three modifiers employed in this study were examined under elution conditions and were shown to not have any impact on these two critical constraints. Further, since the AdP product was known to be susceptible to dimerization at higher pH, these elution studies were carried out in the presence of 5 mM TCEP (a mild reducing agent to prevent cysteine dimerization).
Figure 5a presents the effects of the three modifier solutions tested independently over a pH range of 3-9. The general trend observed was that high pH resulted in increased elution, irrespective of which additive was used. The order of modifier effectiveness was arginine > ethylene glycol > sodium chloride. In addition, distinct trends were seen with each modifier as a function of pH. While recovery with sodium chloride exhibited a linear increase in recovery with pH, elution with ethylene glycol was only moderately impacted by pH. In contrast, arginine was very effective as an eluent modifier; with the recovery showing an initial increase followed by saturation as a function of pH.
Once arginine was identified as the most effective modifier, the effect of concentration was evaluated in the pH range of 6.0-9.0 where good recoveries were observed at 1 M. The effect of concentration was less pronounced at higher pH (Figure 5b). As can be seen, recoveries of greater than 80% were attained for conditions with pH ≥ 8.0 and arginine concentrations ≥ 300 mM. Based on this data, a pH of 8.5 and an arginine concentration of 500 mM were selected as elution conditions for the complex feed study described below.

Purification of AdP from Crude Mixtures

After establishing the operating conditions for the affinity precipitation process using pure AdP, we then evaluated the performance of this process with a crude feed, now incorporating three intermediate precipitate wash steps. Importantly, this process worked extremely well with the crude feed, resulting in a final AdP purity of 88% and an overall recovery of 81%. Figures 6 and 7 show the efficacy of this process using SDS PAGE and RPLC analysis, respectively. As can be seen in the gel, while the feed containing the AdP was complex (lane 5), the final supernatant from the process (lane 8) was quite pure. The RPLC analysis also clearly demonstrates that the final product is significantly purified during the process.