Optimising Biotherapeutic Protein Yields
The primary determinants for successful recombinant protein expression are high expression levels of the recombinant genes, translation efficacy and post-translational modifications. Transcriptomic and proteomic tools have helped in our understanding of the molecular basis of protein expression. However, while RNA-Seq provides insight into transcription it fails to take into consideration that translational control of mRNA is a highly regulated step in determining levels of an expressed protein. Ribosome profiling (Ribo-Seq) on the other hand offers insight into the translational kinetics of proteins at single-codon resolution, thereby offering a method to exactly determine which mRNAs are being translated and at what rates.
Other translatomic approaches, such as polysome profiling, can also be used to investigate the translational efficiency of individual mRNAs, with transcripts that are located in polysome fractions deemed to be more translationally active than those in sub‐polysomes. This information can be used to identify ribosome signatures of cell stress and reprogramming of translation within cells. Additionally, biotherapeutic functionality depends heavily on its structure. Therefore, changes to the structure can result in a loss of therapeutic effect and/or an immunogenic reaction. When optimising yields of these biologics, it is imperative to ensure that this doesn’t impact the drug quality negatively.
Translatomics can aid in determining potential alterations in protein conformation, thus preventing any negative impact that may occur upon increasing protein yields. Additionally, translatomics can help us optimise cell feeding strategies and identify targets for cell engineering strategies to prolong or enhance recombinant gene synthesis. The publications outlined below exemplify the potential for using translatomics for this purpose.
Translatome analysis of CHO cells to identify key growth genes
Journal of Biotechnology, 2013; 167(3):215-224
Courtes FC, Lin J, Lim HL, Ng SW, Wong NS, Koh G, Vardy L, Yap MG, Loo B, Lee DY
Chinese hamster ovarian (CHO) cells are the most commonly used mammalian cell lines for the production of recombinant proteins in industry. A better understanding of the intracellular mechanisms that govern production capacity in CHO cells is required to fully optimise cell culture processes. While individually transcriptomics and proteomics studies have had success in addressing this limitation, very few researchers had attempted to combine both levels of information and those that have generally fail to show a strong correlation between mRNA and protein expression levels in CHO cells. Thus, for a more accurate insight into the translational landscape of CHO cells, a deeper understanding of translational regulation is required. This is where the field of translatomics comes into play. The authors of this paper conducted the first investigation of translational efficiency on a global scale of a CHO DG44 cell line producing monoclonal antibodies (mAb). They used polysome profiling along with Nimblegen microarrays to probe the translation of more than 13 thousand annotated CHO-specific genes. The aim of this study was to identify and prioritise high potential candidate-genes for improving cellular growth using translatomics as a novel strategy to do so.
Key Findings
- The distribution of ribosome loading during the exponential growth phase allowed the identification of stably and highly translated genes as potential key players for cellular growth.
- Highly translated mRNAs at the exponential phase were growth related.
- Key growth genes such as Hnmpc, Utp6, Pcna, Vcp and Mcm5 were found to be lowly expressed in terms of mRNA abundance. Despite this, these genes exhibited high levels of translational efficiency. The overexpression of such genes, where translation is unlikely to be a rate limiting step in their expression, could offer an appropriate strategy for an increase in protein synthesis.
Implications
This paper was one of the first to generate data that helped bridge the gap between the transcriptome and proteome. The closing of this gap has provided invaluable information to researchers in the field of biotherapeutics by enabling researchers to identify potential cell engineering targets that promote growth of host cells while also identifying potential bottlenecks in the gene expression pathway.
Ribosome profiling-guided depletion of an mRNA increases cell growth rate and protein secretion
Scientific Reports, 2017; 7:40388
Kallehauge TB, Li S, Pedersen LE, Ha TK, Ley D, Andersen MR, Kildegaard HF, Lee GM, Lewis NE
Unlike polysome profiling, ribosome profiling provides information regarding exact ribosome positioning on a single transcript. It also uniquely provides insight into how a recombinant mRNA integrates into and affects the translated endogenous mRNA pool. Kallehague et al. present the first comprehensive genome-wide view of translation in response to the introduction of a recombinant mRNA, in this case mAb, in the CS13-1.0 CHO cell line which uses the dihydrofolate reductase (DHFR) selection system to amplify the antibody. This cell line has approximately 90 gene copies per cell of the heavy chain encoding gene and 280 gene copies per cell of the light chain encoding gene. The cells were grown in a batch culture bioreactor, and ribosome profiling and RNA-Seq were carried out at early and late growth phases to study the translatome during IgG-production. The authors then analysed how the translational power of the cell was distributed across all genes and using a platform called Proteomaps they visually grouped genes into their corresponding cellular processes, providing a global view of total ribosomal occupancy.
Key Findings
- The recombinant mRNAs (antibody heavy and light chain mRNA) were found to be the most abundant transcripts.
- The recombinant mRNAs were translated as efficiently as the endogenous mRNAs and were accountable for up to 15% of the total ribosome occupancy.
- The authors identified that the resistance marker NeoR was transcribed and translated at a high rate unnecessarily. siRNA depletion of NeoR resulted in an increase in cellular growth and an 18% increase in antibody titers.
Implications
Ribosome profiling offers a better understanding of how a mammalian expression host utilises and distributes its translational power across its mRNA and how recombinant genes integrate into the steady-state mRNA pool. In the CS13-1.0 cell line, NeoR was found to be the dominant mRNA species in the cell, but with NeoR being a non-essential gene during exponential growth, depletion of NeoR resulted in an increase in cellular growth and protein titre, indicating the possibility to increase protein titre by removing non-essential mRNAs.
Polysome profiling of mAb producing CHO cell lines links translational control of cell proliferation and recombinant mRNA loading onto ribosomes with global and recombinant protein synthesis
Biotechnology Journal, 2017; 12(8):1700177
Godfrey CL, Mead EJ, Daramola O, Dunn S, Hatton D, Field R, Pettman G, Smales CM
During protein manufacture at industrial scale, CHO cells can experience stresses that lead to reprogramming of translation and decreased protein synthesis. Translational profiling such as polysome profiling and ribosome profiling can be used to investigate the translational efficiency of individual mRNAs, with mRNAs located in polysomes deemed to be more translationally active than those in sub‐polysomes. An exception to this is when elongation is slowed or blocked due to stress or limitation in translational efficiency. Comparing the polysome profiling data to protein yields will help determine which event is occurring. The authors of this paper used polysome profiling in recombinant monoclonal antibody‐producing (mAb) CHO cell lines during batch culture to investigate this reprogramming of translation by profiling key mRNA markers across the polysome profiles. They also investigated differences in translational activity between cell lines. This is a key study in demonstrating the application of polysome profiling to help optimise protein yields in CHO cell systems.
Key Findings
- A large proportion of the transcript copies for all recombinant genes were found in the mid and heavy polysome fractions across the cell lines investigated.
- The highest titre cell line was that which sustained recombinant protein synthesis and had the highest copy numbers in the heavy polysomes fraction for all recombinant gene transcripts including recombinant heavy chain mAb transcripts.
- A shift in the distribution of ribosomes from polysomes to monosomes is indicative of translational reprogramming.
- Elevated RagC mRNA transcripts in the heavy polysome fraction were found in the highest antibody-producing cell line. RagC is involved in the mTOR signalling pathway which is a key regulator of ribosome biogenesis, cell growth and protein synthesis.
Implications
Transcripts found preferentially in polysome fractions at various stages of batch culture highlights these as potential targets for either over‐expression or down‐regulation, depending on the role they play in the cell and their effects on the recombinant protein of interest. The 5′ untranslated regions (UTRs) of these transcripts could also potentially be used to ‘load’ recombinant transcripts onto ribosomes to improve recombinant protein expression. Additionally, the RagC polysome fingerprint found in the highly productive cell line could be a potential target for the generation of host cell lines with enhanced growth and productivity characteristics.
Effects of codon optimization on coagulation factor IX translation and structure: Implications for protein and gene therapies
Scientific Reports, 2019; 9(1):15449
Alexaki A, Hettiarachchi GK, Athey JC, Katneni UK, Simhadri V, Hamasaki-Katagiri N, Nanavaty P, Lin B, Takeda K, Freedberg D, Monroe D
Codon usage bias is where synonymous codons occur at different rates in various species. Codon optimisation is the replacement of multiple synonymous codons within a gene sequence and is often done in the biopharmaceuticals industry to increase protein yields. Recently there has been some evidence that codon optimisation does affect protein conformation. Very little is understood about how proteins synthesised using codon-optimisation compare to those obtained from the WT gene. In this paper, the authors investigated the impact of synonymous codon substitutions on protein expression and function in coagulation factor IX (FIX). They expressed coagulation factor IX (FIX) in HEK293T cells and compared the optimised (CO) FIX variant to the wild-type (WT) FIX. They used ribosome profiling to study the translation kinetics of both variants at single-codon resolution. This allowed identification of regions on the gene where synonymous substitutions are most likely to alter protein conformation.
Key Findings
- The data of this study strongly indicated that the WT and CO FIX have different conformations.
- Codon optimisation of the F9 gene altered the ribosomal distribution pattern when compared to the WT transcript which suggests a significant change in local translational kinetics.
- The ribosome profiling data showed overall comparable transcript and ribosome protected fragment (RPF) occupancy, indicating similar translational efficiencies (TE) between WT and CO F9.
Implications
The findings of this paper offer novel insights into the effects of codon optimisation when used for the production of recombinant proteins and gene therapies. The outcome of this study is a better understanding of how codon optimisation strategies can be done in a way that contributes to the development of safer and more efficient FIX biotherapeutics.
Accurate design of translational output by a neural network model of ribosome distribution
Natural Structure and Molecular Biology, 2019; 25(7):577-582
Tunney R, McGlincy NJ, Graham ME, Naddaf N, Pachter L, Lareau LF
As a ribosome moves along a transcript, it encounters various codons and tRNAs which affects translation elongation and thus gene expression. It is becoming increasingly more well-known that synonymous coding mutations and mutations within tRNAs can have dramatic effects on the folding and stability of proteins which may be undesirable in the field of biotherapeutics, as significantly altered biotherapeutic structure can render the therapeutic ineffective or may illicit an immune reaction in the patient. The authors of this paper developed a neural network model of translation elongation and trained it using a high-quality ribosome profiling dataset collected in yeast. The approach was integrated into a software package, named iXnos, which uses the trained model to design coding sequences spanning a range of predicted translation-elongation speeds for any given gene. To test the translation predictive ability of iXnos, the authors expressed synonymous variants of the fluorescent protein eCitrine in yeast and were able to demonstrate that their model could capture information regarding translation dynamics in vivo.
Key Findings
- Initiation is usually thought to be rate limiting step for protein synthesis of most endogenous genes, although the findings of this paper amongst others is that elongation has a significant impact on protein synthesis. One contribution could come from pileups behind stalled or slow-moving ribosomes, which would diminish the maximum throughput of protein production.
- Optimised elongation rates do result in more protein per mRNA, and this does not depend entirely on mRNA stability.
- A linear relationship between predicted elongation rate and translation efficiency was observed.
- Slow translation of CCG (proline), CGA (arginine), and CGG (arginine) codons at the A site was observed.
- Sequences in the P site contribute to elongation speed with the CGA codon showing a strong inhibitory effect in the P site.
Implications
iXnos captures information regarding translation dynamics in vivo which can be utilised to design coding sequences. Codon preferences vary across species, tissues, and even conditions and so codon optimisation can be complex. Utilising this neural network model will aid in capturing information about codon preferences in any system that can be profiled using ribosome profiling. This will be extremely helpful in the field of biotherapeutics for optimisation purposes and thus protein yields.