Novel Polyanionic Tracts Improve Expression of a TPP1:K16ApoE Chimera

Abstract

BackgroundLow expression of highly‐charged recombinant proteins remains a critical challenge in developing protein‐based therapeutics. It has been shown that inclusion of a polyanionic tract (PAT) may improve expression through inter‐ and intramolecular interactions. We previously developed a chimeric construct consisting of tripeptidyl‐peptidase 1 (TPP1), a glycine‐serine linker ([GS]5), a glycine‐aspartate PAT ([GD3]3), and a basic cell‐penetrating peptide K16ApoE as a potential therapeutic agent for late‐infantile neuronal ceroid lipofuscinosis. While TPP1‐(GS)5‐(GD3)3‐(GS)5‐K16ApoE was expressed efficiently, it was cleaved at the PAT. We therefore sought to design novel PATs that improve expression of the full‐length chimera.MethodsTPP1:K16ApoE chimeric proteins were synthesized to contain tri‐glutamate sequences instead of tri‐aspartate sequences in the original (GD3)3PAT. All seven possible combinations of substitutions were engineered: (GE3)3, (GE3)‐(GD3)2, (GD3)‐(GE3)‐(GD3), (GD3)2‐(GE3), (GE3)2‐(GD3), (GE3)‐(GD3)‐(GE3), or (GD3)‐(GE3)2. Chimeric proteins were expressed in Chinese hamster ovary (CHO) cells and media samples were assayed for TPP1 activity. Activity was normalized to a wild‐type (unmodified) TPP1 positive control expressed in CHO cells. The identity of each construct was verified using SDS‐PAGE and mass spectrometry. Welch’s ANOVA test was used to determine expressional differences between TPP1‐(GS)5‐K16ApoE and PAT‐modified chimeras and the resulting p‐values were adjusted for multiple comparisons using Dunnett’s T3 multiple comparison test.ResultsSeven constructs were synthesized, corresponding to each of the aforementioned tri‐glutamate‐substituted PATs, following the general format of TPP1‐(GS)5‐(PAT)‐(GS)5‐K16ApoE. All chimeras expressed successfully as full‐length constructs (Figure 1). No TPP1 activity was detected in the media of untransfected CHO cells. TPP1‐(GS)5‐K16ApoE was previously determined to express at 1.4±1.0% of unmodified TPP1. Of these chimeras with novel PATs, TPP1‐(GS)5‐(GE3)‐(GD3)2‐(GS)5‐K16ApoE [mean ± standard error: 7.3±1.7%, p=0.0040] and TPP1‐(GS)5‐(GE3)2‐(GD3)‐(GS)5‐K16ApoE [5.8±3.1%, p=0.0397] exhibited significantly increased expression relative to TPP1‐(GS)5‐K16ApoE at the 0.05 level of significance.ConclusionAlteration of the a priori (GD3)3 PAT sequence by varying the aspartate and glutamate content was sufficient to prevent cleavage of the resulting TPP1:K16ApoE chimeras. However, only the (GE3)‐(GD3)2 and (GE3)2‐(GD3) PATs yielded a 5‐fold and a 4‐fold, respectively, improvement in expression over the original TPP1‐(GS)5‐K16ApoE construct. Further investigation is required to understand the electrostatic and/or structural basis of these findings.