A critical starting point in drug discovery or development is often cDNA cloning and expression of the corresponding protein. Availability of large amounts of the purified recombinant protein of interest is often key to optimization of lead compounds using high-throughput screening and structure-based drug design.

Advances in recombinant protein expression technology were recently discussed November 2-5, 1997, at the Third Annual Current Topics in Gene Expression Systems conference. The meeting, held in San Diego, California, was sponsored by Invitrogen and Research Corporation Technologies. In attendance were more than 300 scientists, representing a mixed balance between pharmaceutical and biotech companies and academia.

Escherichia coli is often the host of choice for expressing recombinant proteins. This bacterial strain is well characterized, and numerous genetic tools are available for its manipulation. However, E. coli and other bacterial strains are ill suited for certain applications. For example, it is not an appropriate host for expressing large multidomain or modular proteins, secreted proteins with extensive disulfide bonds, or proteins requiring specific post-translational modifications like glycosylation and phosphorylation where eukaryotic systems are appropriate.

The choice of both the host cell and the genetic vector is critical for the successful expression of the gene of interest. One needs to keep in mind problems of gene maintenance, control of expression, and the cellular destination of the heterologous protein in order to select constitutive versus inducible promoters or secreted versus intracellular expression. Also, compromises between quantity and authenticity of heterogeneously produced proteins are usually necessary.

Bacteria

Escherichia coli

For many purposes, E. coli still offers several advantages: it is well defined genetically, it grows rapidly (with a doubling time of approximately 20 minutes); DNA manipulations are easy, fermentation technology is well developed, and, most importantly, it is a proven host for many heterologous proteins. In many cases, recombinant proteins can be expressed easily in E. coli when optimal expression elements are selected both to regulate gene expression and to direct the nascent product to an appropriate cellular compartment.

Often there is the need for a strong promoter system that can be regulated and fully repressed in the absence of inducer while assuring high-level expression upon induction. This is especially critical when the protein of interest is cytotoxic or aggregates as inclusion bodies.

The pBAD vectors provide both modulated induction and tight repression of protein production in E. coli. These vectors carry the P_BAD promoter, which can be regulated by common sugars. Arabinose induces expression, whereas glucose acts as a catabolite repressor. Manipulating the amounts of arabinose, glucose, and glycerol allow precise control of expression levels. Therefore, with pBAD vectors one can optimize expression and downstream purification for maximal yield of their functional, soluble product. In some cases, heterologous proteins accumulated at a level of 1 g/L, as reported by XOMA Corporation scientist Marc Better.

Another exciting development was the use of L-form strains, that is, bacteria without cell walls, of E. coli and Proteus mirabilis by Johannes Gumpert (Institute of Molecular Biotechnology, Jena, Germany). These L-forms have been cultivated as genetically stable strains for more than 30 years. They can grow in complex media under shaking conditions and in semitechnical fermentation regimes. The main advantage of this system is the synthesis of recombinant proteins as extracellular, soluble, and functionally active products that remain cell-bound. Secreted proteins are not degraded and remain stable during 20-50 hours of fermentation because there is no extracellular proteolytic activity in the L-form cultures. Examples of foreign proteins expressed in these systems include staphylokinase, prochymosin, penicillin G acylase, green fluorescent protein, and antibodies (scFv). The yields of the recombinant proteins tested were in the range of 10-800 mg/L. These systems without cell walls are useful for elucidating topological problems of secretion, folding, and modification of recombinant proteins.

Streptomyces lividans

Streptomyces, a gram-positive bacterium lacking an outer membrane, is a simple alternative host when secretion is desired. Streptomyces are becoming an increasingly employed host system primarily because their secreted products typically remain soluble, can be purified from fermentation media, and in many cases are correctly folded and biologically active. Additionally, they do not produce the endotoxins associated with gram-negative bacteria.

Useful vectors for gene expression in Streptomyces have been developed. In addition, a variety of replicons for episomal replication and integrative vectors have been engineered. A new bifunctional E. coli-Streptomyces vector for high-level protein expression in actinomycetes is being developed by John Verberg and colleagues at MDS-Panlabs, Inc. This new plasmid, pMCExpress, exploits the mitomycin C inducible MCRA (mitomycin resistance gene product) promoter to overexpress target genes in a mitomycin C concentration-dependent manner.

Codon preferences in Streptomyces are quite different from other systems because its genome is GC-rich. Understanding both codon usage (GC-bias) and codon context (what codon is next to what codon) are key elements that may impact on the design of DNA constructs for protein expression of foreign genes.

Yeast

Pichia pastoris

The methylotrophic yeast Pichia pastoris is a useful host system for production of foreign proteins of commercial and academic interest. Heterologous gene transcription is controlled by P_AOX1, an efficient promoter obtained from the methanol-regulated alcohol oxidase 1 gene of Pichia. For expression of genes whose products are not deleterious to the cells, the P_GAP promoter of glyceraldehyde-3-phosphate dehydrogenase can be used for constitutive expression (James Cregg, Oregon Graduate Institute of Science and Technology, Portland).

Systems for generating secreted (pPICZa) or intracellular (pPICZ) protein products are available. Fermentation methodology is well developed, allowing high-level recombinant protein production exceeding 500 mg/L. This method allows for production of ¹⁵N-labeled protein at reasonable cost, less than $3,000 per liter for NMR structural analyses of proteins, as reported by Elizabeth Komives, University of California at San Diego.

Interesting exploitations of this system include metabolic pathway engineering and expression of multimeric enzymes (Central Research and Development, DuPont & Company). These advances should have great impact on the fields of biocatalysis and biotransformation of commercially important processes or industrial applications of specialty chemicals. The best current example of commercial utility is nitrile hydrolysis (production of acrylamide from acrylonitrile).

Cell Culture

Mammalian

While bacterial and yeast hosts have the advantages of ease of culturing and high yield of heterologous protein production, they are limited by the absence of the pathways needed for authentic mammalian glycosylation. To this end, the Ecdysone-Inducible Mammalian Expression System (EIMES) has literally revolutionized inducible gene expression in mammalian cells. Developed by Salk Institute investigators in Ronald Evans's laboratory, the EIMES two-vector system, available from Invitrogen, is designed to allow regulated expression of the gene of interest in mammalian cells.

The system is distinguished by its tightly regulated expression mechanism, which permits very little basal expression and a more than 200-fold inducibility. The expression system uses the heterodimeric ecdysone receptor of Drosophila. The regulator vector encodes both monomers of the heterodimeric receptor and allows for their constitutive expression. The induction vector contains an ecdysone-responsive promoter that drives expression of the gene of interest. Once the receptor binds ecdysone (an insect hormone) or the ecdysone analog muristerone A, the receptor activates the ecdysone-responsive promoter to yield high-level expression of the gene. Muristerone A has neither toxic nor pleiotropic effects on mammalian cells. Moreover, the regulator plasmid responds in a ligand dose-dependent manner.

Insect

(1) Drosophila Expression System (DES). DES is a revolutionary advance in recombinant protein expression in eukaryotic cells. Initially developed by SmithKline Beecham researchers (Martin Rosenberg and colleagues), DES combines the best features of mammalian and insect expression systems for recombinant protein production. DES is nonviral/nonlytic and uses simple plasmid transfection for transient or stable protein expression in Drosophila S2 cells.

The system has many convenient properties: Cells grow at ambient temperature to high densities in suspension without CO₂, and they readily adapt to serum-free media. Plasmid vectors with strong promoters are available for regulated expression (e.g., the Drosophila metallothionein promoter, for high-level, inducible expression with CuSO₄ or CdCl₂) or continuous production (the Drosophila Ac5 actin gene promoter). The system can support expression of more than one transfected gene product. The cells recognize eukaryotic signal sequences for specific endoproteolytic cleavage and secretion. They can stably incorporate foreign DNA at high or low copy number in a single transfection event. Multicopy integration of expression plasmids allows for isolation of high-producing cell lines, which continue to express recombinant proteins at elevated cell densities.

The production of recombinant proteins in the DES system is less expensive and more efficient than expression in mammalian cells and is faster than expression using the baculovirus system described below. A distinct feature of DES is that endogenous Drosophila proteins generally do not interact with mammalian proteins, so, S2 cells provide a "null background" for functional studies of proteins and cell based assays or screens. This system is now available commercially from Invitrogen.

(2) Baculovirus Expression System (BvES). The recombinant Bv vectors are very useful for heterologous expression in insect cell culture. One major advantage is the abundant expression of recombinant proteins and the production of biologically active products. Therapeutic agents, vaccines, and reagents are among the pharmaceuticals being made by this system.

Under certain circumstances, two issues may be of concern. The first is the different glycosylation patterns of insect and mammalian cells. The second concern, since this is a lytic system, is that lysed cells release proteases that can potentially degrade the recombinant protein.

Advances in understanding the biology of insect cells may lead to resolution of the first issue (Donald Jarvis, Texas A&M University at College Station). Engineering the (mammalian) N-glycosylation pathway in insect cells is an active area of research with demonstrated success. These enhanced features, no doubt, will add a dimension to this system as an excellent alternative to bacterial and mammalian expression systems.

Transgenics

Plants

Biosource Technologies of Vacaville, California, has developed autonomously replicating RNA viral vectors for the production of heterologous proteins in plants. Infectious RNA transcripts from cDNA clones can be passaged through packaging hosts to generate a viral inoculum for systemic transfection of field cultivars. Large-scale production of recombinant proteins can be easily obtained by simply increasing the size and number of inoculated plants. Examples of high-level, functional expression of recombinant proteins are antimicrobial peptides, a blood product (hemoglobin), an anticancer scFv antibody, and an AIDS therapy drug (a-trichosanthin), as reported by Monto Kumagai.

Animals

Transgenic animals offer a more efficient and a more cost-effective alternative for the production of large quantities of therapeutic proteins than conventional manufacturing protocols. Production of recombinant proteins from the milk of transgenics involve four basic components:

(1) targeting expression of the gene of interest to the mammary gland of the host animal
(2) generating transgenic animals that express the targeted gene, verifying stable integration of the recombinant DNA, and evaluating the levels of recombinant protein expression in the milk of lactating transgenic animal
(3) maintaining and expanding the transgenic population
(4) purifying the recombinant protein from milk.

Using transgenic animals, Genzyme Corporation (Li-How Chen and Susan Schiavi) has produced a variety of therapeutic recombinant proteins, including enzymes (human chitinase in mice), enzyme inhibitors (a1-proteinase inhibitor in goats and mice at 20-35 g/L), soluble receptors (srCD4 and human transferrin in mice at 8 and 2 g/L, respectively), monoclonal antibodies (anticancer MAb in goats and mice at 10 g/L) as well as a host of growth factors and cytokines. The transgenically produced antithrombin III (expression levels in mice and goats at 10-14 g/L) is currently in phase III clinical trials, and more products are scheduled to enter the clinic this year. Other such proteins have also been expressed in rabbits, sheep, and cows. These developments announce the approaching maturity of an industry capable of low-cost production of biopharmaceutical proteins.

Conclusions

Biotechnology and Bioengineering - a journal including articles on various aspects of production and purification of recombinant proteins. Beginning with the 1996 issues, tables of contents, abstracts, and a search function are available online without cost.

Scale-up of Mammalian and Insect-Cell Cultures for the Production of Recombinant Proteins and Viral Particles - a course to be hosted May 25-29, 1998, in Montreal by the Canadian National Research Council.