Sunflower/Corn $8.10

Canola $8.95

Bioreactor Crop

Figure 1 Relative cost of protein production in various agricultural crops. (Adapted from Kusnadi et al., 1997.)

detectable, serum immune response (Chargelegue, personal communication, 1999). Thus, in the mouse at least, a "self" primary protein structure, decorated with plant N-linked glycans, can be nonimmunogenic. This planti-body has been applied topically in the mouth of humans with no detection of human antimouse antibodies (Ma et al., 1998).

Humanization of plantibody primary structures may go a long way to obviate the immunogenic potential of plant glycans, and further genetic engineering may, if necessary, alter the glycan structures themselves. Most dramatically, aglycosyl antibodies can be created by altering the peptide recognition sequence for N-linked glycosylation (asn-X-Ser/Thr). High-mannose glycosylation, which does not contain the core-linked xylose and fucose residues, may be favored by the addition of a C-terminal KDEL sequence (Wandelt et al., 1992) and the subsequent targeting of plantibodies to the proximal endoplasmic reticulum. Alternatively, the specific fucosyl (Leiter et al., 1999) or xylosyl transferases, which operate in the trans-Golgi, may be targeted for silencing.

C. Gene Silencing: A Potential Problem of Plant Expression

Levels of transgenic protein accumulation reported in the literature (Table 1) generally represent the highest levels found in primary transformants. It is becoming increasingly clear that stability of expression is just as important as absolute levels of expression. Many laboratories have noted the common occurrence of transgene silencing in plants, in which transgenes can become inactivated either during development or in subsequent generations (Baul-combe, 1999; Depicker and Montagu, 1997; Matzke and Matzke, 1998; Meyer and Saedler, 1996; Stam et al., 1997; Vaucheret et al., 1998). Two types of gene silencing have been observed in plants: transcriptional gene silencing (TGS), in which promoters become silenced and transcription is reduced, and posttranscriptional gene silencing (PTGS), in which transcription continues unabated but transcripts are rapidly degraded. The exact mechanisms of both types of gene silencing are still being discovered, but there are common features to both. For instance, the presence of homologous sequences is frequently involved. If repeats are present within transgenes, or if transgene sequences are homologous to endogenous genes, silencing is more frequent. Methylation of DNA may be involved. In TGS, the promoters of silenced genes are hypermethylated. In PTGS, it is the coding regions that tend to be methylated, de Neve et al. (1999) found both types of silencing in a careful examination of five different homozygous transgenic Arabidopsis lines expressing either IgG or Fab.

Silencing seems to be positively correlated with level of expression and with copy number of the transgene. That is, transgenes present in multiple copies or expressed from strong constitutive promoters are more likely to be silenced (Elmayan and Vaucheret, 1996; Jorgensen et al., 1996). This suggests some possibilities for reducing the occurrence of silencing. Silencing may be less of a problem with tissue-specific or weak constitutive promoters, although this may be antithetical to the desire for high levels of expression. The introduction of multiple copies can be reduced by the use of Agrobacterium, which tends to result in fewer transgene copies than bio-listic transformation. Flanking transgenes with matrix attachment region (MAR) sequences has been shown to reduce the occurrence of gene silencing (Uelker et al., 1999; Vain et al., 1999). This probably works by preventing formation of antisense transcripts by transcription from flanking endogenous genes into the transgene. Viral supppressors of silencing have been used quite effectively to prevent or reverse PTGS (Anandalakshmi et al., 1998; Kasschau and Carrington, 1998). When plants that are transgenic for a viral protein known as helper component protease (HC-Pro) are crossed with plants in which a transgene is silenced, expression is restored in progeny containing both transgenes. It should also be possible to use HC-Pro-expressing plants as the starting material for transformation with a gene of interest.

Finally, a number of mutants have been isolated in Arabidopsis that lack either TGS or PTGS (Elmayan et al., 1998; Furner et al., 1998; Scheid et al., 1998). Some of these mutants appear to be normal in every respect except the inability to silence transgenes. It may eventually be possible to isolate similar mutations in plants that are more practical for the production of transgenic proteins.

D. Purification and Process Development

The potential for cost reduction when using genetically engineered green plants as bioreactors, instead of conventional pharmaceutical factories engineered with concrete and steel, is a powerful argument for the use of transgenic plants to produce recombinant proteins. The commercial-scale production of proteins from transgenic plants generally requires one to envisage the growth and processing of tons of biomass to achieve the economy of scale that would fully exploit the inputs of sunlight, soil, water, and fertilizer. The processing of large amounts of biomass also anticipates large numbers of patients or consumers, large amounts of purified protein needed for each patient or consumer, lower than anticipated levels of expression, and losses during processing.

The realities of commercial-scale production benefit from a demand for simplicity early on in the development of a large-scale process. Bench-scale or laboratory-scale procedures often employ the pampering conditions that are necessary for a proof of concept but are too complicated and expensive for large-scale efforts. For instance, because of issues regarding toxicity and expense, it is preferred not to use protease inhibitors beyond the bench scale. Reagents such as ammonium sulfate are frowned upon because of the disposal-associated costs, as are organic solvents because of their toxicity and fiammability. Efforts to avoid the purchase and maintenance of centrifuges will eventually be gratefully acknowledged by maintenance personnel. In addition, and more fundamentally, it is important to realize that a commercial process will not be executed by rocket scientists, each with extensive postdoctoral training, but by more ordinary and, while no less dedicated, almost certainly less well-trained and educated people (i.e., they cost less). Thus, simplicity in the number, as well as the type, of components in a process is paramount to reduce not only costs but also errors.

Minimalist approaches to large-scale process development will often be rewarded even though they may appear simplistic. We found that grinding transgenic tobacco in water alone gave up to 70% of the expected immunoglobulin compared with tobacco ground in a buffer, poised at a specified pH, containing six components in addition to water; each component had been chosen for a particular, biochemically sound, reason. Yet this was evidence which suggested that perhaps not all of the components were vitally important, and we now use a two-component buffer at a specified pH.

Commercial-scale production of recombinant proteins from plants will also benefit from the technology and equipment commonly used in the food and beverage industry. Grain mills and coleslaw slicers will doubtless be useful off-the-shelf machinery for the initial processing of seeds and leafy tissue. The beverage industry has always been concerned with the clarification of juices and their phenolic content (van Sumere et al. 1975) and is a source of knowledge and equipment, both new and used. The chef's trick of delaying the browning of cut fruit by the addition of lemon-acidulated water (Bombauer and Becker, 1975) has led us to formulate buffers to inhibit the "tanning" of proteins in tobacco extracts.

Two constituents of stem and leaf extracts that require special consideration are membranes and cell walls. The green color of stem and leaf extracts indicates the presence of chlorophyll and the suspension of thyla-koid membranes. These and other membranes must be removed to allow filterability at or below a pore size of 0.45 /xm. This level of filterability helps to ensure, but does not guarantee, the good behavior of the extract during subsequent chromatography and serves to remove bacterial sources of contamination. The cellulose-containing debris, which forms a large part of the insoluble portion of extracts, can be used as an endogenous "filter aid" during the initial clarification to remove these membranous components.

Phenolics are a major concern when extracting most stems and leaves, and one's efforts are rewarded by their early removal. They may interact with proteins and other extract components via hydrophobic interactions, salt bridges, hydrogen bonding, and by additional reactions to nucleophilic centers (Gegenheimer, 1990; van Sumere, et al. 1975; Loomis, 1974). These interactions can dramatically and irreversibly alter the properties of proteins. Fortunately, the majority of released phenolics are generally small in size, as well as water soluble, and may be removed by tangential-flow ultrafiltration/diafiltration, which also serves to concentrate the considerably larger proteins of interest. In addition, other incompatible, water-soluble secondary metabolites, such as neonicotine (anabasine) and nicotine from tobacco, may be removed by ultrafiltration/diafiltration.

After such clarification and concentration steps, recombinant proteins can be assumed to behave independently and with regard to the peculiarities of their own biochemistry. In other words, they are now ready for chromatography. Residual phenolics may still dictate the degree of cleanup necessary before chromatography on an expensive protein A or protein G affinity column is allowed; these are economic considerations common to any purification process.

One additional note of caution: because the processed plant or seed was probably grown in dirt, or some similarly unclean stratum, one may anticipate the presence of a diverse bioburden. Contamination of product with endotoxins and mycotoxins can be minimized by rapid processing and early filtration, but process development must also always recognize the necessity to eliminate compounds as well as the necessity to purify, concentrate, and stabilize a product of interest.


Although antibodies were first expressed in plants in the mid-1980s (Steiger, Düring) by two German graduate students (Düring et al., 1990), the first report was published in 1989 (Hiatt et al., 1989). Since then, a diverse group of "plantibody" types and forms have been prepared (see Table 2). Originally, foreign antibody genes were introduced into plant cells by nonpathogenic strains of the natural plant pathogen Agrobacterium tumefaciens (Horsch et al., 1985) and regeneration in tissue culture resulted in the recovery of stable transgenic plants. Although this initial work to generate multichain proteins required crossing of plants expressing each chain, more recent studies have shown that multiple chains can be introduced via a single biolistic transformation event (Sanford, 1988; Wycoff et al., unpublished data), greatly reducing the time to final assembled plantibody.


This laboratory has focused on the production of secretory IgA (SlgA) plan-tibodies (Ma et al., 1995) (see Fig. 2). At the present time, plants offer the only large-scale, commercially viable system for production of this unique form of antibody. SlgA is the most abundant antibody class produced by the body (>60% of total immunoglobulin). SlgA is secreted onto mucosal surfaces to provide local protection from toxins and pathogens (see Fig. 3). The SlgA is composed of four different protein chains: heavy and light immunoglobulin chains that form the antigen-binding hypervariable region, a J chain that dimerizes two IgA molecules (SlgA has four antigen binding sites), and a secretory component that is derived from the mucosal epithelial

Table 2


Mab form (no. of chains)

Single domain (dAb) (1) Single chain Fv (1) Single chain Fv (1) Single chain Fv (1)

Single chain Fv (1) Single chain Fv (1) scFv


Substance P (neuropeptide) Phytochrome

Artichoke mottled crinkle virus coat Abscisic acid

Root-knot nematode

Beet necrotic yellow vein virus


Human creatine kinase Transition-state analogue

Fungal cutinase Streptococcus mutans adhesin

Plant species—comment

Nicotiana Nicotiana

Nicotiana—viral protection

Nicotiana—wilty phenotype


Nicotiana benthamiana Nicotiana, KDEL augments expression

Nicotiana Arabidopsis Nicotiana



S. mutans adhesin Nicotiana NP (4-hydroxy 3-nitrophenylacetyl) Nicotiana (Hapten) ___


Benvenuto et al.

Owen et al. (1992) Tavladoraki et al.

(1993) Artsaenko et al.

(1995); Phillips et al. (1997) Rosso et al. (1996) Fecker et al. (1996) Fiedler et al. (1997); Schouten et al. (1996, 1997); Bruyns et al. (1996) de Neve et al. (1993) Hiattet al. (1989);

Hein et al. (1991) van Engelen et al.

Figure 2 The structure of secretory IgA.

cells. Dimeric IgA containing J chain derived from submucosal B cells binds to the epithelial cell polyimmunoglobulin receptor (PIGR) that transports the IgA to the mucosal surface. Binding triggers transcytosis to the mucosal surface, where a protease releases a portion of the PIGR called secretory component conveniently used to bind the SIgA. The secretory component protects the dimeric IgA from proteases and denaturation on the mucosal surface. Previously it was not possible to obtain therapeutic quantities of this class of immunoglobulin. The recent availability of large amounts of secretory IgA plantibodies opens up a number of novel therapeutic opportunities for disorders of the mucosal immune system. These include therapies for intestinal pathogens such as hepatitis viruses, Helicobacter pylori, and enterotoxigenic E. coli, and cholera; respiratory pathogens such as rhinovirus and influenza; and genitourinary sexually transmitted diseases (e.g., herpes simplex virus) and contraception.

To date, three immunotherapeutic products produced in plants have entered the clinic. These products, listed in Table 3, include two antibodies and an oral vaccine. Clinical studies of the anti-EPCAM plantibody (co-developed by NeoRx and Monsanto) were discontinued due to significant gastrointestinal side effects. The anti—Streptococcus mutans antibody is currently in phase II trials.


The most clinically advanced SIgA plantibody, called CaroRx, recognizes and inhibits the binding of the major oral pathogen, Streptococcus mutans, to teeth. In preliminary work, a series of in vivo passive immunization ex

SlgA Transcytosis

Figure 3 SlgA is produced by two cell types. Submucosal lymphoid tissue produces dimeric, IgA, which diffuses to the serosal side of the mucosal epithelial cells, where it binds to the polyimmunoglobulin receptor. Secretory component (SC) is proteolytically derived from the polyimmunoglobulin receptor.

Figure 3 SlgA is produced by two cell types. Submucosal lymphoid tissue produces dimeric, IgA, which diffuses to the serosal side of the mucosal epithelial cells, where it binds to the polyimmunoglobulin receptor. Secretory component (SC) is proteolytically derived from the polyimmunoglobulin receptor.

periments was carried out in 84 human subjects using murine anti-S. mutans antibodies (Ma et al„ 1987, 1989, 1990; Lehner et al., 1975, 1985). Topical application of anti-S. mutans antigen SA I/II monoclonal antibodies (MAbs) prevented colonization of artificially implanted exogenous strains of S. mutans as well as natural recolonization by indigenous S. mutans. In these studies the pathogenic S. mutans was replaced by endogenous flora (see Fig. 4).

The presence of the complement-activating and phagocyte-binding sites on the Fc fragment of the MAb was not essential for activity because the F(ab')2 portion of the MAb was as protective as the intact IgG; however, the Fab fragment failed to prevent recolonization of S. mutans. Prevention of recolonization was specifically restricted to S. mutans, as the proportion of other organisms, such as S. sanguis, did not change significantly. The surprising feature of these experiments was that protection from recolonization by S. mutans lasted up to 2 years (J. Ma, personal communication), although MAb was applied for only 3 weeks and functional MAb was detected on the teeth for only 3 days following the final application of MAb. All studies indicated that this form of immunotherapy appears to be safe

Table 3

Plant-Produced Human Therapeutics (—1998)

Was this article helpful?

0 0
How To Bolster Your Immune System

How To Bolster Your Immune System

All Natural Immune Boosters Proven To Fight Infection, Disease And More. Discover A Natural, Safe Effective Way To Boost Your Immune System Using Ingredients From Your Kitchen Cupboard. The only common sense, no holds barred guide to hit the market today no gimmicks, no pills, just old fashioned common sense remedies to cure colds, influenza, viral infections and more.

Get My Free Audio Book

Post a comment