Highly sensitive detection of mutations in CHO cell recombinant DNA using multi‐parallel single molecule real‐time DNA sequencing

Abstract

High-fidelity replication of biologic-encoding recombinant DNA sequences by engineered mammalian cell cultures is an essential pre-requisite for the development of stable cell lines for the production of biotherapeutics. However, immortalized mammalian cells characteristically exhibit an increased point mutation frequency compared to mammalian cells in vivo, both across their genomes and at specific loci (hotspots). Thus unforeseen mutations in recombinant DNA sequences can arise and be maintained within producer cell populations. These may affect both the stability of recombinant gene expression and give rise to protein sequence variants with variable bioactivity and immunogenicity. Rigorous quantitative assessment of recombinant DNA integrity should therefore form part of the cell line development process and be an essential quality assurance metric for instances where synthetic/multi-component assemblies are utilized to engineer mammalian cells, such as the assessment of recombinant DNA fidelity or the mutability of single-site integration target loci. Based on Pacific Biosciences (Menlo Park, CA) single molecule real-time (SMRT™) circular consensus sequencing (CCS) technology we developed a rDNA sequence analysis tool to process the multi-parallel sequencing of ∼40,000 single recombinant DNA molecules. After statistical filtering of raw sequencing data, we show that this analytical method is capable of detecting single point mutations in rDNA to a minimum single mutation frequency of 0.0042% (<1/24,000 bases). Using a stable CHO transfectant pool harboring a randomly integrated 5 kB plasmid construct encoding GFP we found that 28% of recombinant plasmid copies contained at least one low frequency (<0.3%) point mutation. These mutations were predominantly found in GC base pairs (85%) and that there was no positional bias in mutation across the plasmid sequence. There was no discernable difference between the mutation frequencies of coding and non-coding DNA. The putative ratio of non-synonymous and synonymous changes within the open reading frames (ORFs) in the plasmid sequence indicates that natural selection does not impact upon the prevalence of these mutations. Here we have demonstrated the abundance of mutations that fall outside of the reported range of detection of next generation sequencing (NGS) and second generation sequencing (SGS) platforms, providing a methodology capable of being utilized in cell line development platforms to identify the fidelity of recombinant genes throughout the production process.

Enhanced CHO Clone Screening: Application of Targeted Locus Amplification and Next‐Generation Sequencing Technologies for Cell Line Development

“…These two additional readouts were not covered by the earlier reported workflows. [20,27] Comparing the point mutation frequencies, determined by TLA/NGS, genomic DNA, mRNA, and LC-MS ( Figure 1C), confirms that the identified variant could also be verified by techniques other than TLA/NGS. In general, different frequencies for the clones A and B seem to correlate quite well across different methods, except for clone C, where the frequency identified by the TLA/NGS seem to be significantly lower.…”

“…Another benefit is that the data enables the identification of sibling clones. These two additional readouts were not covered by the earlier reported workflows …”

“…While cell culture conditions can be changed and optimized to reduce mistranslations later in development, it is critical to screen and differentiate for DNA mutations early, during CHO cell line development to minimize the negative impacts on development timelines and costs. Our NGS approach with CHO clones can address these needs in one single platform making it an optimal analytical tool to obtain a variety of analytical screening parameters …”

“…Peptide mapping with liquid chromatography coupled with tandem mass spectrometry (LC–MS/MS) and protein subunit LC–MS methods are most commonly used to receive information on location, identity, and relative quantification of sequence variants at the amino acid level with high sensitivity . These techniques, however, are still labor‐intense and low throughput, especially when detection of low‐level sequence variants is required . Novel next‐generation sequencing (NGS) techniques have revolutionized the gene and genome analysis lately in terms of speed, costs, throughput, and sensitivity and are increasingly applied in biopharmaceutical research and development .…”

“…[25] These techniques, however, are still labor-intense and low throughput, especially when detection of low-level sequence variants is required. [20,21,26,27] Novel next-generation sequencing (NGS) techniques have revolutionized the gene and genome analysis lately in terms of speed, costs, throughput, and sensitivity and are increasingly applied in biopharmaceutical research and development. [28] Several recent publications describe the application of NGS for CHO cell line development workflows and clone screening.…”