(419g) Tailoring the Thermo-Responsiveness of Elastin-like Polypeptides with Short Charged Sequences

Elastin-like polypeptides (ELPs) are widely used as functional domain in recombinant protein design for their entropy-driven thermo-responsiveness. When below the lower critical solution temperature (LCST), ELPs are soluble, and the solution stays in one homogeneous phase. However, when the temperature is above the LCST, ELPs phase-separate into a protein-rich phase, or coacervate. This coacervation occurs with a sharp transition (within 2 °C) and is reversible upon lowering the temperature below the LCST. Because of the LCST behavior, ELPs have been used as purification tags for recombinant proteins and to construct temperature responsive protein-based materials or drug delivery vehicles.

The LCST of an ELP is determined by several factors including protein concentration, salt identity and concentration in the solution, and pH of the solution, but is predominantly determined by the amino acid sequence of the ELP. Although the effect of amino acid composition of ELP sequences on the LCST is well studied, it has not been well established how other protein domains in an ELP-fusion protein can influence the LCST behavior. Here we report that the LCST behavior of an ELP-fusion protein can be significantly changed by short recombinant protein domains outside of the ELP sequence. We constructed two ELP-fusion proteins, c-ELP and v-ELP, with the same ELP sequence but with different N-terminal leading sequences consisting of 5-23 amino acids with different charge densities.

Despite the fact that the two N-terminal sequences comprise less than 10% of the total length of the ELP-fusion proteins, c-ELP and v-ELP showed a difference greater than 3 °C in their LCSTs in phosphate buffered saline (PBS) at the same protein concentration. We further quantified LCSTs of both ELPs at varying pH values using a tri-buffer system. c-ELP maintained a constant LCST of 45 °C from pH 5.7 to 9.0. In contrast, the LCST of v-ELP decreased from 65 °C at pH 5.7 to 35 °C at pH 9.0. To explain these large differences, we estimated the protein chain charge and found that both the ELP sequence and the leading sequence on c-ELP remained positively charged and had no significant change in charge over the tested pH range. On the other hand, the leading sequence on v-ELP changed from positively charged to negatively charged at pH 6.6. Based on the charge analysis, we hypothesize that the LCST behavior of ELP-fusion proteins can be modulated not only by the ELP sequence itself, but also by the charge sign and density on sequences outside of the ELP region. Moreover, when both ELP and the leading sequence have the same charge, the LCST will be higher to overcome the energy penalty from higher charge repulsion in the coacervate. When the ELP sequence is oppositely charged from the leading sequence, the charge repulsion in the coacervate will be lower, and the LCST will be lower.

Thus, our results provide insights into the design of ELP-fusion proteins with specific LCST behavior: the charge behavior of leading sequences should be taken into consideration and can also be harnessed as a new parameter in ELP-fusion protein design to fine tune the LCST. This allows us to have more flexibility in designing ELP sequences. A more precise control over the LCST behavior will allow us to build sophisticated ELP-based biomaterials for biological applications such as temperature-triggered hydrogel and pH-responsive drug delivery vehicles.