The microalga Chlamydomonas reinhardtii as a platform for the production of human protein therapeutics

Abstract

Microalgae are a diverse group of eukaryotic photosynthetic microorganisms. While microalgae play a crucial role in global carbon fixation and oxygen evolution, these organisms have recently gained much attention for their potential role in biotechnological and industrial applications, such as the production of biofuels. We investigated the potential of the microalga Chlamydomonas reinhardtii to be a platform for the production of human therapeutic proteins. C. reinhardtii is a unicellular freshwater green alga that has served as a popular model alga for physiological, molecular, biochemical and genetic studies. As such, the molecular toolkit for this microorganism is highly developed, including well-established methods for genetic transformation and recombinant gene expression. We transformed the chloroplast genome of C. reinhardtii with seven unrelated genes encoding for current or potential human therapeutic proteins and found that four of these genes supported protein accumulation to levels that are sufficient for commercial production. Furthermore, the algal-produced proteins were bioactive. Thus, the microalga C. reinhardtii has the potential to be a robust platform for human therapeutic protein production.

The first recombinant therapeutic protein approved by the US Food and Drug Administration was insulin over 25 years ago. Since then, the class of protein-based therapeutics has grown quickly. As of 2002, there were about 140 therapeutic proteins approved in the US and Europe.¹ About 40% of recombinant therapeutic proteins on the market are non-glycosylated and are typically expressed in either bacteria (Escherichia coli) or yeasts (S. cerevisiae or Pichia pastoris). N-glycosylated proteins are usually made in mammalian tissue culture cells, such as Chinese hamster ovary (CHO) cells. Other therapeutic productions systems include insect cell culture and transgenic plants and animals.² An ideal organism for recombinant protein expression would include the following characteristics: rapid growth, high protein yield in a biologically active form, generally regarded as safe (lacking endotoxins or infectious viruses), genetically manipulatable, capable of rapid scale-up, cost effective culturing and protein production, capable of appropriate post-translation modifications, and able to produce, fold and assemble complex proteins. Because each production system mentioned above has unique advantages and disadvantages,² choice of expression system is often protein-dependent.

Interestingly, the microalga C. reinhardtii has been shown to contain many of the above characteristics. The singled-celled green alga C. reinhardtii is a popular model alga for studies of photosynthesis and flagellar function, for example. As such, C. reinhardtii is genetically well characterized and has an extensive molecular toolkit. All three genomes (the nuclear, chloroplast and mitochondrial) have been sequenced^3–5 and genetic transformation methods are well established.⁶ C. reinhardtii grows relatively quickly, doubling every 5–8 hours, and can grow to densities above 10⁷ cells/ml. Because C. reinhardtii is photosynthetic, its media requirements are minimal and thus it is inexpensive to culture and relatively easy to keep sterile.

The chloroplast of C. reinhardtii is a particularly attractive organelle for the production of some recombinant proteins. The algae has a single cup-shaped chloroplast that occupies about 40% the volume of the cell.⁷ The chloroplast genome is stably transformed using biolistics.⁸ It has been shown in both algae and in higher plants that recombinant proteins accumulate to much higher levels when expressed from the chloroplast genome in comparison to the nuclear genome.^9,10 The chloroplast of C. reinhardtii has been shown to support expression of a range of recombinant proteins including reporters such as GUS,¹¹ luciferase,^12,13 and GFP,⁹ vaccines,¹⁴ and industrial enzymes (our unpublished data). Many of the proteins that have been quantified have been shown to accumulate as high as 2–20% of total soluble protein (TSP).^14–16 Importantly, the chloroplast contains the proper machinery to form disulfide bonds and assemble large, complex proteins such as full-length antibodies.^17,18

While the chloroplast of C. reinhardtii has been shown to support expression of a variety of recombinant proteins, the number of successful reports remains low, especially in comparison to other expression systems such as Escherichia coli, S. cerevisiae and CHO cells. Thus, we sought to challenge the microalga chloroplast expression system with seven recombinant genes encoding for a diverse set of current or potential human therapeutic proteins (Table 1).¹⁶ The seven recombinant genes were codon-optimized for C. reinhardtii chloroplast codon bias and tagged with a FLAG tag to enable detection by western blot and for purification. Each was cloned into three distinct chloroplast transformation vectors: (1) pD1-KanR under the control of the psbA promoter and 5′ and 3′ untranslated regions (UTRs) with an integration site at the psbA locus resulting in replacement of the endogenous psbA gene, (2) the N-terminal tagging vector pD1-KanR-SAA, which is identical to pD1-KanR except that the recombinant gene is cloned downstream of the well expressed M-SAA gene,¹⁵ resulting in a M-SAA fusion protein, and (3) pAtpA in which expression of the human therapeutic gene is controlled by the atpA promoter and 5′ UTR and the rbcL 3′ UTR and integrates into an intergenic region of the chloroplast genome near the psbA locus (Fig. 1).¹⁶ The promoter regions are responsible for initiating transcription, while 5′ UTRs function to regulate both mRNA stability as well as translation.¹⁹ The 3′ UTRs appear to influence mRNA processing and stability,^20–23 but may also interact with 5′ UTRs to influence translation.^19,24 In general, the regulation of chloroplast gene expression is primarily controlled at the level of translation.^25–27

Table 1.

Genes encoding for human protein therapeutics used in the study¹⁶

Gene	Vector	Expression?	%TSP
	pD1-KanR	−	−
EPO	pD1-KanR-SAA	−	−
	pAtpA	−	−
	pD1-KanR	+	low*
10FN3	pD1-KanR-SAA	+++	ND
	pAtpA	−	−
	pD1-KanR	+++	3
14FN3	pD1-KanR-SAA	+++	ND
	pAtpA	++	0.15
	pD1-KanR	−	−
Interferon β	pD1-KanR-SAA	−	−
	pAtpA	−	−
	pD1-KanR	+	low*
proinsulinp	D1-KanR-SAA	−	−
	pAtpA	−	−
	pD1-KanR	+++	2
VEGF	pD1-KanR-SAA	+++	ND
	pAtpA	++	0.1
	pD1-KanR	+++	2.5
HMGB1	pD1-KanR-SAA	+++	1
	pAtpA	+++	1

Approximate protein expression levels are represented by plus (+) signs. A minus (−) sign indicates that no protein was detected by either western blot or immunoprecipitation. The asterisk (*) indicates that the protein was only detected by immunoprecipitation. ND indicates ‘not determined.’

Figure 1.

Transformation of the Chlamydomonas reinhardtii chloroplast genome with genes encoding for human therapeutic proteins. Schematic diagram of the transformation vectors used and the corresponding integration sites. pD1-KanR: Replacement of the endogenous psbA gene with the gene of interest under the control of the psbA promoter and UTR elements. The kanamycin resistance gene aphA6 (KanR) under the control of the atpA promoter and 5′ UTR is genetically linked to the gene of interest. Grey regions flanking the gene of interest and resistance gene corresponds to regions of the chloroplast genome used for homologous recombination between the insertion plasmid and the C. reinhardtii chloroplast genome. pD1-KanR-SAA: Identical to pD1-KanR except that the gene of interest fused to the C-terminus of M-SAA. pAtpA: Vector used for the insertion of the genes of interest under the control of the atpA promoter and 5′ UTR and the rbcL 3′ UTR into the BamHI silent site between the psbA gene the 5S rRNA.¹⁹ All recombinant proteins were C-terminally fused to the 1 x FLAG-tag sequence (DYKDDDDKS) for western blotting and purification.

Three of the human therapeutic proteins; domain 14 of fibronectin (14FN3), VEGF and HMGB1 expressed well in each of the three vectors (Table 1). However, the highest levels of protein expression, between 2–3% of TSP, were achieved with the psbA regulatory elements in the psbA deletion strains. This is consistent with previous published data.^14,15 Indirect immunofluorescence of transgenic C. reinhardtii stably expressing HMGB1-1xFLAG verifies the chloroplast localization of the biopharmaceutical as well as demonstrates the large volume that the single algal chloroplast occupies within the cell (Fig. 2). 14FN3, VEGF and HMGB1 also expressed well when fused to M-SAA. In addition, fusion of 10FN3 to the well-expressed M-SAA enabled it to accumulate to high levels. Thus, of the seven genes encoding for human therapeutic proteins, four of them expressed to relatively high levels in the chloroplast of C. reinhardtii. This success rate is similar to those achieved in yeast and mammalian cells,^28,29 and much better than the rate of soluble protein expression achieved in Escherichia coli.³⁰

Figure 2.

Indirect immunofluorescence of a strain expressing HMGB1-1xFLAG from the chloroplast genome. Wildtype (137c) or transgenic algae expressing aptA::HMGB1-1xflag were fixed on to poly-L-lysine treated coverslips, stained with mouse anti-FLAG primary antibody, washed and then stained with anti-mouse-texas red (TR). The DNA was stained with DAPI. The merge image demonstrates that HMGB1 does not co-localize with the nucleus, as indicated by the large DAPI-staining body, as would be expected from a chloroplast-localized protein.

Next, we tested whether the algal-expressed proteins were bioactive. Because 10FN3 and 14FN3 are potential antibody mimics whose purpose would be to bind target proteins, the most important requirement for these biopharmaceuticals is that they are soluble. And indeed, 10FN3 and 14FN3, along with VEGF and HMGB1, are soluble in the algal chloroplast.¹⁶ VEGF and HMGB1 were purified from algal lysates and tested for bioactivity. Native VEGF is active as a dimer and functions to stimulate the growth of new blood vessels. This process is initiated by the binding of VEGF to its tyrosine kinase receptor, the VEGF receptor (VEGFR). Algal-expressed VEGF assembles into its dimer form within the chloroplast, suggesting the protein is folded properly. Bioactivity of algal VEGF was determined by competitive binding to VEGFR in the presence of varying concentrations of bacterial-derived VEGF using an ELISA assay. Bacterial VEGF was able to displace algal VEGF (200 ng/ml) from VEGFR with an IC50 of ∼40 ng/ml, consistent with a shared binding site and broadly similar affinity.¹⁶ Thus, algal-expressed VEGF is indeed bioactive. HMGB1 mediates a number of important functions involved in wound healing including endothelial cell activation, stromagenesis, recruitment and activation of innate immune cells and dendritic cell maturation.³¹ Fibroblast chemotaxis assays were used to determine whether algal-expressed HMGB1 was bioactive. Indeed, algal-expressed HMGB1 displayed similar bioactivity compared to bacterial HMGB1.¹⁶ Taken together, the data demonstrates that algal chloroplasts can support high levels of recombinant protein accumulation for a variety of human protein therapeutics. Furthermore, the proteins are soluble, easily purified and bioactive.

Thus, our results support the conclusion that the microalgal chloroplast has the potential to be a commercially viable recombinant protein production platform. Indeed, microalgae are easily transformed, they grow rapidly and to high densities, many species are generally regarded as safe, photosynthetic strains are inexpensive to culture and can be scaled rapidly and the chloroplast can support expression of large, complex proteins that contain disulfide bonds. Our study demonstrated that the algal chloroplast can support the expression of a variety of human proteins with therapeutic potential and thus is a versatile system.

While we and others have shown that the algal chloroplast can support expression of a broad range of heterologous proteins,^32,33 it remains unclear why some proteins accumulate and others do not. It is possible that certain recombinant proteins are susceptible to degradation by proteases. Another possible explanation is that the heterologous mRNA folds into a secondary structure that inhibits the initiation of translation, i.e., by prohibiting mRNA processing, translation factor binding or ribosome assembly and initiation. Alternatively, the mRNA may be folded in such a way as to block translation elongation that results in premature termination of protein translation. Indeed, because the heterologous genes are codon-optimized for chloroplast expression, the native mRNA secondary structures are not maintained.

Recombinant protein production in algal chloroplast does have some limitations. Proteins expressed in the algal chloroplast cannot be secreted, which means the cells must be harvested and lysed in order to purify the desired recombinant protein. And, like bacteria, the algal chloroplast is not capable of protein glycosylation. However, both these limitations can be overcome if the recombinant gene is expressed from the nuclear genome of microalgae. Nuclear transformation of C. reinhardtii is well-established and many heterologous genes have been expressed from the nuclear genome.³⁴ Importantly we have successfully engineered C. reinhardtii to secrete nuclear-encoded GFP from the cells into the media (our unpublished data). Nuclear expression also enables recombinant protein glycosylation, although this glycoslyation is unlikely to be identical to mammalian glycosylation. However, the disadvantage of recombinant protein expression from the microalgal nuclear genome is product yield. To date, the yield achieved through nuclear expression tends to be much lower compared to chloroplast expression and therefore is not yet commercially viable, although certainly an area of active research and development.

Another limitation of our algal chloroplast expression system is that the best expression achieved to date has been with non-photosynthetic psbA-deletion strains. However, the ideal protein production strain should be photosynthetic in order to keep production costs low. We are currently taking a multi-tiered approach to improve recombinant protein accumulation in photosynthetic strains, including high-throughout screening for psbA auto-attenuation mutants and evaluating alternative promoters and UTR regulatory elements to identify those supporting robust recombinant protein production.

Because the algal expression system is still in its infancy, there are many potential improvement strategies that still need to be tested to increase recombinant protein accumulation. For example, strain improvements through genetic mutation, directed evolution or genetic engineering will likely yield positive results. Identifying protease deficient strains has proven useful for other expression systems,² and should therefore benefit the algal chloroplast expression system. Algal production will also be improved through an expanded molecular toolkit. The types of tools that are being developed take their cues from other production systems, including fluorescent or luminescent reporter systems, tags that promote easy protein purification, improved selection markers, improved expression and inducible expression. As each of these improvements are introduced we can expect production to increase and costs to decrease, potentially making algal recombinant protein production the most cost effective system for a variety of recombinant proteins.

Addendum to: Rasala BA, Muto M, Lee PA, Jager M, Cardoso RM, Behnke CA, et al. Production of therapeutic proteins in algae, analysis of expression of seven human proteins in the chloroplast of Chlamydomonas reinhardtii. Plant Biotechnol J. 2010;8:719–733. doi: 10.1111/j.1467-7652.2010.00503.x.