Optimizing Expression Systems

July 1, 2012
Amy Ritter, PhD

Scientific Editor, BioPharm International

BioPharm International, BioPharm International-07-01-2012, Volume 25, Issue 7

Industry experts discuss methods for optimizing protein expression in bacterial and mammalian cell lines.

Most biopharmaceutials are produced in either Chinese Hamster Ovary (CHO) cells or in Escherichia Coli. Here, biopharmaceutical manufacturers provide examples of how they have optimized protein expression in these commonly used cell types to increase product titers and effieciency.


Jonathan Pointon, PhD, principal scientist, at Fujifilm Diosynth Biotechnologies

One of the major bottlenecks in the production of biopharmaceuticals is the efficient expression of therapeutic proteins in microbial or mammalian cells. Fujifilm Diosynth Biotechnologies' Escherichia coli pAVEway expression system described here has been developed to ensure high product titers and efficient scale up to GMP manufacture while minimizing issues such as leaky expression (i.e., expression of recombinant protein in the absence of inducer), that can have detrimental effects on host and plasmid stability. Moreover, gene to fermentation can take as little as five weeks using the generic platform pAVEway fermentation processes outlined below.

How it works

Several key components of the pAVEway expression vectors contribute to the functionality of the system. The use of a number of powerful E. coli RNA polymerase promoters, such as T7A3, λpL and tac, opens up a large host range in comparison with the popular T7 system that is limited to hosts carrying the λDE3 prophage. Although the λDE3 prophage lacks the elements required to excise the prophage, there is evidence suggesting some λDE3 strains release phage particles under certain conditions, which is highly undesirable in fermentation plants (1).

Figure 1: Schematic representation of pAVEway repression and subsequent induction. The lac repressor tetramer binds to each pPOP, positioned either side of the promoter. This causes a DNA loop to form and in combination with the increased affinity of the lac repressor for pPOP compared to native lac operator sequences, extremely tight repression is observed. Addition of the inducer (IPTG, in yellow) displaces the lac repressor tetramer allowing transcription of gene of interest mRNA to begin. (FIGURES 1– 3 COURTESY OF FUJIFILM DIOSYNTH TECHNOLOGIES)

Control over basal expression from the strong pAVEway promoters is provided by the use of perfectly palindromic lac operator (pPOP) sequences. pPOP sequences facilitate much tighter binding of the lac repressor tetramer than native lac operator sequences. Moreover, the use of two optimally spaced pPOP sequences (91– 92 base pairs apart), one upstream and one downstream of the promoter, leads to DNA looping and, in combination with tighter lac repressor binding, ensures extremely tight control over basal expression (see Figure 1). This level of control is extremely important in large scale (i.e., ≥5 L) fermentations as it allows high biomass accumulation prior to induction.

Figure 2: A generic pAVEway fermentation. Before induction, tight repression of basal expression ensures all cells are capable of protein expression. After induction, recombinant cells continue to grow (blue graph) and express protein (red graph) leading to high titers.

In addition, such tight repression can be helpful when expressing proteins that are potentially toxic to E. coli, because induction is performed at a higher biomass than is achievable in a leaky expression system. Because leaky expression can have a seriously negative impact on cell growth and plasmid stability, the ability to effectively switch off protein expression until induction ensures that all cells are capable of protein production. For an expression system used in large-scale manufacture this is a highly desirable characteristic and demonstrates control over the process (see Figure 2). This feature also enables a generic high cell-density fermentation protocol to be used for any protein, with no specific optimization required (see Figure 2). Additional backbone components that confer further stability on pAVEway vectors led to the development of an antibiotic-free production system without compromising product yield. The presence of antibiotic in fermentation media is increasingly becoming a regulatory concern so the ability to run antibiotic-free pAVEway fermentations highlights the versatility of the system.

Figure 3: Control of gene transcription with varying IPTG inducer concentration. The ability to tune protein expression allows protein translation to match the folding/secretion capacity of the cell.

The rate of target gene transcription in the pAVEway system can be controlled by varying the concentration of inducer (IPTG), because there is a linear relationship between IPTG concentration and gene expression from these vectors (see Figure 3). The ability to control the rate of expression can be useful in situations where the maximum rate of expression may not be needed for optimal accumulation of the appropriate form of the protein. For example, it may be beneficial to slowly express a recombinant protein targeted for secretion, so that the host secretion machinery is not overloaded as this can greatly reduce the growth and productivity of recombinant cells. Similarly, if producing a soluble protein intracellularly, for which E coli has a limited folding capacity, especially for complex mammalian proteins, a slower induction phase may aid in the folding process. In addition to the tuning effects exerted by varying IPTG concentration, different combinations of the two promoter region components lead to a range of pAVEway vectors with expression kinetics that can be tailored to the requirements of a specific protein and its production route, whether it be intracellular insoluble/soluble or secreted soluble. When combined with generic high cell-density fermentation protocols (possible due to the tight repression of pAVEway vectors), high titers of a diverse range of biopharmaceuticals, from viral and bacterial proteins through to complex mammalian proteins such as growth factors, cytokines, and antibody fragments have been produced. These titers can be boosted further with subsequent fermentation optimization (see Table I).

Table I: Examples of proteins produced using the pAVEway expression system. Titers shown are from the first fermentations using the generic fermentation process for each accumulation method (intracellular soluble/insoluble or secreted soluble). Subsequent fermentation optimization has lead to titers in excess of 20 g/L for erythropoietin, tumour necrosis factor-α and granulocyte colony stimulating factor (G-CSF), in addition to titers of 1.1 g/L for the secreted D1.3 Fab fragment.


1. S. Shuman, Proc. Natl. Acad. Sci. USA, 86, 3489–3493, (1989).


Anne B. Tolstrup, Barbara Enenkel, Stefan Schlatter, Jochen Schaub, Harald Bradl, Anja Puklowski, Patrick Schulz, Anurag Khetan and Hitto Kaufmann, Mammalian Cell Line and Process Development in Process Science, Boehringer Ingelheim

Manufacturing of new biological entities (NBEs) such as monoclonal antibodies (mAbs) and bispecific derivatives typically depends on mammalian expression because of the necessary post-translational modification of these types of molecules. To meet today´s demands of short and cost-efficient development, the expression system employed should be robust, high-titered, and fast. The BI-HEX mammalian expression platform developed by Boehringer Ingelheim meets these requirements, and is comprised of four interlinked elements (see Figure 1):

  • A proprietary vector system

  • A Chinese hamster ovary (CHO) host cell line lacking the dhfr gene

  • A proprietary medium and

  • A complete upstream and downstream manufacturing process.

Figure 1A: Elements of Boehringer Ingelheirm’s BI-Hex expression system. (FIGURES 1A–3A COURTESY OF BOEHRINGER INGELHEIM)

Behind each element, there is a panel of choices that can be exploited for further optimization of the individual NBE. For example, a panel of BI-HEX expression vectors are at hand comprising different IgG isotypes, different genetic elements, and different frameworks. Choice of the right vector constellation is important for optimization of the desired effector function as well as for the manufacturability of the NBE.

The glycosylation structures on the mAb influence the effector function of the molecule. For example, antibody-dependent cell cytotoxicity is enhanced if the fucose content of the N-linked carbohydrate structure attached to Asn297 in the constant region of the molecule is reduced. The BI-HEX platform comprises two substrains of the host cell line, HEX1 and HEX2. Both cell lines are derivatives of the CHO DG44 cell line originally established in 1980, but they differ in their glycoprofiles (1). HEX1 has a higher content of A2FG0 and also a higher level of defucosylated carbohydrates compared with HEX2 which has a higher A2FG1 content (see Figure 2A). This allows preselection of a more optimal glycoprofile dependent on the desired function of a given NBE. Furthermore, for development of biosimilars, where it is critically important to match the originator molecule with respect to all clinically relevant product quality attributes, the platform provides the option to preselect a profile being most similar to the originator molecule.

Figure 2A: Relative distributions of different glycoforms produced in HEX1 and HEX2 Chinese hamster ovary (CHO) cell lines.

Having generated the stable BI-HEX production cell line based on the optimal vector constellation and host cell substrain, the next important element is upstream process development. Based on this platform process, a fast and robust design-of-experiments-driven approach for media and process development can be used. This part of the BI-HEX expression platform involves the ability to manipulate or influence the product quality, such as the fine-tuned level of antibody-dependent cell-mediated cytotoxicity (ADCC), is highest—provided that the right vector and host cell line were selected from the start. Bioreactor parameters such as pH settings, stirring, and gassing as well as media adjustments, composition, and timing of the feed are all elements that have a high impact on product quality as well as on product titers. A significant increase in titer is typically seen during this part of the optimization and high titers of 5–8 g/L have been reached with BI-HEX (see Figure 3A).

Figure 3A: Comparison of titers resulting from different feed strategies.

A recent addition to the BI-HEX platform is the generation of a BI-HEX cell line modified to produce antibodies with very low levels of fucose. This is achieved by means of the GlymaxX technology developed by ProBioGen with whom BI entered into a collaboration agreement in 2011. The BI-HEX GlymaxX antibodies exhibit a significant increase in ADCC activity which is desirable, for example, in cancer indications (see Figure 4A).

Figure 4A: Comparison of antibody-dependent cell-mediated cytotoxicity by antibodies produced in an unmodified BI Hex expression system, and antibodies produced using GlyMaxX technology.

Looking into the future, further automation of all parts of the BI-HEX platform, including miniaturization of clone and process screening as well as media and feed optimization is in the pipeline. Also, use of disposable systems wherever possible is being pursued to leverage the fast and flexible turnaround of such systems.


1. G. Urlaub and L.A. Chasin, Proc. Natl. Acad. Sci. USA 77 (7), 4216–20 (1980).