FIELD OF THE INVENTION

The invention relates to the secretion of compounds from the roots of a plant.

BACKGROUND OF THE INVENTION

Genetically modified plants have numerous advantages over microbial and animal systems for the production of recombinant pharmaceuticals. Examples include the potential for large-scale, low cost agricultural production, the ability of plant cells to correctly fold, assemble and process complex proteins, the low risk of contamination with mammalian viruses and the avoidance of ethical issues raised by the use of transgenic animals.

The extraction of recombinant proteins from plant tissues can be an expensive and time-consuming process involving plant harvesting, tissue maceration and subsequent protein purification. Protease release during sample homogenisation and protein extraction may also result in degradation of recombinant protein (Ma et al., 1994) and the process of extraction and purification does not allow continuous production of recombinant protein over the lifetime of the plant. Furthermore, it allows the release of protein at all stages of production, including partially folded or processed intermediates (Cabanes-Macheteau et al., 1999), which may further complicate a purification process.

In order to bypass these problems, secretion-based systems have been investigated. For example, single chain Fv and monoclonal antibody heavy chain have been recovered from the surrounding growth medium of genetically- modified tobacco cell suspensions (Firek et al., 1994; Magnuson et al., 1996) and Agrobacterium rhizogenes-deήved hairy roots of tobacco were used to secrete assembled full-length IgGl antibody (Wongsamuth and Doran, 1997). In general, protein yields from cell suspensions have been low, averaging l-5mg/L cell suspension culture (Boehm, 2007). Cell suspension cultures are also frequently genetically unstable, whilst hairy root culture requires expensive bioreactors and cannot exploit the autotrophic capacities of the whole plant.

An alternative to these methods is to use the natural rhizosecretion mechanism of the plant which in nature has a role in the processes of nodulation, mycorrhizal colonisation, growth inhibition of neighbouring plants, acquisition of nutrients from soil and defence against toxic metals (Gleba et al, 1999).

The possibility of obtaining secondary metabolites by rhizosecretion from plants was demonstrated in yellow lupin (Lupinus luteus L) secreting the isoflavonoid genistein (Kneer et al., 1999). Genistein serves in plants as a signaling compound to initiate the legume-rhizobia symbiosis in nodule formation in nitrogen fixing plants, and also has several desirable anti-cancer pharmacological properties. Roots of hydroponically cultivated yellow lupin transferred to water secreted genistein at about 5 μg/g/fresh weight although was increased dramatically by use ofelicitors.

In 1999, green fluorescent protein (GFP), bacterial xylanase and human placental alkaline phosphatase (SEAP) were the first recombinant proteins to be produced by rhizosecretion in transgenic tobacco. The genes for GFP, xylanase and SEAP were fused to a signal peptide that ensured translated proteins were directed to the endoplasmic reticulum and the default secretory pathway. All three secreted proteins retained their biological activity and accumulated in higher amounts in the hydroponic medium than in root tissue. When a a 35S promoter was employed, the yields obtained for GFP and SEAP were 0.923 and 5.8 μg/g.root dry weight/24h respectively.

In a later study, functional, murine full-length IgGl was harvested from around the base of tobacco roots following rhizosecretion at a rate of 11.7 μg/g.root dry weight/24h (Drake et al., 2003). Sexton et al. also reported rhizosecretion of functional anti-HIV microbicide cyanovirin-N (CV-N) at a rate expressed as 0.64 μg/ml hydroponic medium after 24 days (Sexton et al., 2005).

Whilst rhizosecretion remains a promising production method, the principal disadvantage of the technology to date has been the relatively low yields of recombinant plant-made pharmaceuticals produced. Various strategies have been employed to try to increase the yield of recombinant protein obtained by rhizosecretion. The use of a root-specific promoter mas2' increased the rhizosecretion of SEAP from 5.8 μg/g.root dry weight/24h obtained with the CaMV 35S to 20 μg/g.root dry weight/24h (Borisjuk et al., 1999). The authors calculated that such a rhizosecretion rate would yield 2.4mg of recombinant protein/g root dry weight over the 4 month lifespan of a hydroponically grown plant.

Gaume et al. used Agrobacterium rhizogenes to induce hairy roots on transgenic tobacco plants expressing SEAP, and compared the rhizosecretion rate of the recombinant protein to that produced by adventitious roots in a non-sterile environment (Gaume et al., 2003). The mean rate of SEAP rhizosecretion in the former was 36 μg/g.root dry weight/24h compared to 7 μg/g. root dry weight/24h in adventious roots. The accumulation of SEAP RNA was comparable in both root types and consequently the authors postulated that the difference in secretion levels between the root types was not based on transcription, but due to enhanced protein synthesis and or secretion in hairy roots.

In a later study, it was demonstrated that replacing the original signal peptides of monoclonal antibodies with plant-derived calreticulin signal increased the levels of rhizosecreted antibody yield two-fold (Komarnytsky et al., 2006). The authors also studied the effect of Bowman-Birk Ser protease inhibitor (BBI) on antibody degradation. Initally, the effect was studied by exogenously adding BBI in vitro under various physiological conditions. BBI provided some stabilisation effect to human IgGl antibody kept on a rotatory shaker in the dark and the level of antibody protection was markedly increased by exposure to light. The protease inhibitor was co-secreted with antibody from transgenic plants by co-expression of the BBI gene fused to a plant-derived signal peptide. This increased antibody production to 36.4 μg/g.root dry weight/24h for single chain IgGl and 21.8 μg/g. root dry weight/24h for full-size IgG4 antibodies. Furthermore, SDS-PAGE indicated that BBI reduced proteolysis compared to previous studies in which antibodies have been secreted from roots (Drake et al., 2003; Sharp and Doran, 2001).

SUMMARY OF THE INVENTION

The inventors have found that rhizosecretion is much increased by applying a plant hormone to the roots of the plant. In its broadest aspect, the invention provides a method for producing a molecule from roots of a plant, comprising applying a plant hormone to the roots that increases rhizosecretion of the molecule from the roots, and recovering the molecule. The invention also provides said method further comprising applying an additional cell wall permeabilisation compound to the roots of said plant simultaneously or sequentially with respect to the plant hormone. The invention also provides the use of a plant hormone to increase rhizosecretion of a molecule from a plant. The invention further provides a solution for growing a plant hydroponically comprising a plant hormone; a dry powder which, when dissolved, provides such a solution; and a hydroponic plant culture comprising a plant and such a solution. The invention also provides a kit comprising a powder which, when dissolved, provides a solution for growing a plant hydroponically, and a plant growth hormone. The invention also provides said solution, said powder, and said kit further comprising an additional cell wall permeabilisation compound.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 shows a two-fold dilution series in an enzyme-linked immunosorbent assay (ELISA) of hydroponic medium containing Guy's 13 antibody and CV-N taken at 24h and 7d after addition of fresh medium. Figures Ia and Ib show Guy's 13 after 24h and 7d culture. Figures Ic and Id show CV-N after 24h and 7d culture. Samples in the ELISA consisted of hydroponic medium combined in equal volumes from a minimum of nine plants. Both ELISAs were functional as plates were coated with antigen: a recombinant fragment of Streptococcal antigen I/II for Guy's 13 and HIV gpl20 for CV-N. Dilutions were performed in MS medium (without NAA Group 1, with NAA Groups 2 and 3, with gelatin for Guy's 13). WT controls were cultured in medium under the same conditions as Group 3 plants, with medium combined from two plants.

Figure 2 shows a non-reducing enhanced chemiluminescence (ECL) western blot of hydroponic samples from plants expressing Guy's 13. Samples consisted of hydroponic medium combined in equal volumes from a minimum of nine plants. Lanes 1-3: samples from Groups 1-3, 24 hours after addition of fresh medium. Lane 4: 24h control sample combined from two WT plants. Lane 5: Guy's 13 positive control from hybridoma supernatant. Lane 6: molecular markers. Lanes 7-9: samples from Groups 1-3, Id after addition of fresh medium. Lane 10: 7d control sample combined from two WT plants. Detection was with horseradish peroxidase labelled anti-mouse kappa antibody.

Figure 3 shows a reducing ECL western blot of hydroponic samples from plants expressing Guy's 13. Samples consisted of hydroponic medium combined in equal volumes from a minimum of nine plants. Lanes 1-3: samples from Groups 1-3, 24 hours after addition of fresh medium. Lane 4: 24h control sample combined from two WT plants. Lane 5: Guy's 13 positive control from hybridoma supernatant. Lanes 6-8: samples from Groups 1-3, 7d after addition of fresh medium. Lane 9: 7d control sample combined from two WT plants. Lane 10: molecular markers. Detection was with a mixture of horseradish peroxidase labelled anti-mouse kappa and anti-mouse gamma chain antibody.

Figure 4 shows the effect of plant growth regulators (plant hormones) on rhizosecretion of Guy's 13 IgG. Young tobacco plants were cultured in MS medium with either no plant growth regulator (PGR), or supplements of lmg/L of IBA, BAP, KIN or NAA added on Days 0, 7 and 14. In the NAA (1) group, the addition of NAA on Days 7 and 14 was omitted. After 26 days, the hydroponic medium was replaced with fresh medium containing the same PGR, the plants were cultured for a further 7 d and medium removed for analysis. Fresh medium was again added for a further 24 h and removed for analysis.

A: Recombinant IgG yields at 7 days and 24 hours were calculated by an antigen specific ELISA using Guy's 13 standards of known concentration. Results are shown for a pooled sample from 9 individual plants. B: Rhizosecretion rates were calculated by antigen specific ELISA and measurement of root dry weights after 24 h incubation.

Figure 5 shows the effect of NAA on rhizosecretion of Cyanovirin-N. Young tobacco plants were cultured in MS medium with either no plant growth regulator (PGR), or supplements of lmg/L of NAA added either on Days 0,7 and 14, or Day 0 only (NAA (1) group). After 26 days, the hydroponic medium was replaced with fresh medium containing lmg/L of NAA, the plants were cultured for a further 7 d and medium removed for analysis. Fresh medium was again added for a further 24 h and removed for analysis. A: Recombinant CV-N yields at 7 days and 24 hours were calculated by an antigen specific ELISA using CV-N standards of known concentration. B:

Rhizosecretion rates were calculated by antigen specific ELISA and measurement of root dry weights after 24 h incubation.

Figure 6 shows western blot analysis of rhizosecreted proteins. A) Anti-kappa chain detection of intact Guy's 13 IgG antibody and antibody fragments by non-denaturing SDS-PAGE. Samples are 1 : 7 day hydroponic fluid;

2: 7 day hydroponic fluid diluted 1 :4; 3: 24hr hydroponic fluid; 4: intercellular fluid from leaf tissue. The arrow indicates intact IgG.

B) Immunodetection of recombinant CV-N with rabbit anti-CV-N antiserum in non-denaturing SDS-PAGE. Lanes 1 - 3, 24 hr samples from no PGR, NAA and

NAA(I) plants respectively; Lane 4, 24 h control sample WT plant; Lane 5, recombinant CV-N positive control from E.coli; Lane 6, molecular markers; Lanes 7-9 7 day samples from no PGR, NAA and NAA(I) plants respectively; Lane 10, 7 day control sample WT plant. The arrow indicates intact CV-N. (C) Anti-kappa chain detection of intact MAb 4E10 antibody and antibody fragments by non-denaturing SDS-PAGE. Lanes 1 and 2, 24 hr medium samples with and without NAA respectively; Lanes 3 and 4, 7 day medium with and without NAA respectively; Lane 5 MS medium only; Lane 6 positive control, purified 4E10 from hybridoma. The arrow indicates intact IgG.

Figure 7 shows gelatin zymogram of N. tabacum extracellular peptidases from leaves and roots. Leaf intercellular fluid (IF) and root hydroponic medium from Guy's 13 IgG (A) and CV-N (B) plants were subjected to SDS-PAGE and peptidase zymogram using gelatin as substrate.

(A) Lane 1 , WT leaf IF, 5μl; lane 2, WT leaf IF, 25 μl; lane 3, Guy'sl 3 leaf IF, 5 μl (1.4 ng of antibody); lane 4 Guy'sl3 leaf IF, 25 μl (7.0 ng of antibody); lane 5, Guy'sl3 root hydroponic medium, day 1, 5 μl (1.1 ng of antibody); lane 6, Guy'sl3 root hydroponic medium, day 1, 25 μl (5.5 ng of antibody); lane 7, Guy'sl 3 root hydroponic medium, day 7, 5 μl (4.8 ng of antibody); lane 8, Guy'sl 3 root hydroponic medium, day 7, 25 μl (24.0 ng of antibody).

(B) Lane 9, WT leaf IF, 5μl; lane 10, WT leaf IF, 25 μl; lane 11 , CV-N leaf IF, 5 μl (2.5ng of CV-N); lane 12, CV-N leaf IF, 25 μl (12.5ng of CV-N); lane 13,

CV-N root hydroponic medium, day 1, 5 μl (22ng of CV-N); lane 14, CV-N root hydroponic medium, day 7, 5 μl (3 Ing of CV-N); lane 15, CV-N root hydroponic medium, day 1, 25 μl (1 IOng of CV-N); lane 16, CV-N root hydroponic medium, day 7, 25 μl (155ng of CV-N).

Figure 8 shows glycosylation analysis of leaf and rhizosecreted Guy's 13 IgG. N- glycosylation profiles of the tryptic glyco-peptides d _{erived from rMAb}

Guy's 13 extracted from leaves or hydroponic medium were determined by LC- ESI-MS. Major glycoforms and their masses are shown; their relative abundance is expressed as a percentage of total glycoforms. For the Fc-located glycopeptide, the 5 most abundant glycoforms are indicated in bold. Figure 9 shows affinity purification of rhizosecreted Guy's 13 IgG. Analysis of samples by (A) silver stain of 4-15% SDS-polyacrylamide gel and (B) western Blot of 10% SDS-polyacrylamide gel, both run under non-denaturing conditions. Samples are 1 : hydroponic medium removed after 7d culture; 2: affinity column flow-through; and 3 : elution fraction.

Figure 10 and Figure 11 show the effect of plant growth regulators (plant hormones) on rhizosecretion of Cyanonvirin-N (CV-N) in tobacco. Plants were cultured in MS medium with either no plant growth regulator (PGR), or supplements of lmg/L of IBA, BAP, KIN, IAA or NAA added on Days 0, 7 and 14. After 26 days, the hydroponic medium was replaced with fresh medium containing the same PGR, and the plants were cultured for a further 24hr (Figure 10) or 7 d (Figure 11) before medium was removed for analysis of CV-N concentration (using spectrophotometric absorbance). A series of two-fold dilutions were used and plotted against a standard concentration of CV-N of 150ng/ml.

Figure 12 and Figure 13 show the effect of α-naphthalene acetic acid (NAA) on rhizosecretion of Cyanonvirin-N (CV-N) in Nicotiana benthamiana. Plants were cultured in hydroponic medium for 26 days in the presence NAA (added at lmg/L on days 0, 7 and 14). The medium was then replaced with fresh medium supplemented with NAA. After a further 24hr (Figure 12) and 7 days (Figure 13) samples were taken to measure Cyanovirin-N (CV-N) concentration (using an ELISA) via a series of two-fold dilutions. Figure 13 also shows CV-N yield without NAA treatment after 7d for N.tabacum.

Figure 14 shows transgenic CV-N tobacco plants grown in a single-tray hydroponic system Figure 15 shows the increase in rhizosecretion of CV-N by transgenic tobacco cultivated using Nutrient Film Technique (NFT) after the addition of NAA to the single-tray hydroponic system shown in Figure 14.

Figure 16 shows the synergistic effect of cell-wall degrading agents on NAA- induced rhizosecretion of CV-N. The "NAA" group was treated with lmg/L NAA for 12 days, whilst the "no NAA" group was not treated with NAA. Sub-groups: 'a' had no cell-wall degrading agents added, 'b' was supplemented with 0.16% cellulase and 0.02% pectolyase and 'c' was supplemented with ImM L-ascorbate, ImM hydrogen peroxide and 20μM CuSO ₄ on day 12. Pre samples were taken on day 12 before addition of agents and 2-day samples were taken on day 14. Values are mean of 6 plants, error bars are standard deviations.

DETAILED DESCRIPTION OF THE INVENTION

As described above, the inventors have found that plant hormones much increase rhizosecretion of molecules from plant roots. Accordingly, the invention provides a method for producing a molecule from roots of a plant, comprising applying a plant hormone to the roots that increases rhizosecretion of the molecule from the roots, and recovering the molecule.

Rhizosecretion is the secretion of a molecule from the roots of a plant. The molecule that is rhizosecreted in the method of the invention is typically a peptide, a protein, a glycoprotein, or a secondary metabolite. The rhizosecreted molecule may be produced naturally by the plant.

The plant may be a dicotyledon plant (e.g. tobacco) or a monocotyledon plant (e.g. rice, wheat, maize). Other examples of the plant include N. benthamiana, S.tuberosum, P. hybrida, Arabidopsis (e.g. A. thaliana), Ethiopian mustard, oilseed, alfalfa, apple, asparagus, banana, barley, cabbage, canola, cantaloupe, carrot, cauliflower, cranberry, cucumber, eggplant, flax, grape, kiwi, lettuce, lupins (e.g yellow lupin), melon, papaya, pea, peanut, pepper, plum, potato, raspberry, service berry, soybean, squash, strawberry, sugar beet, sugarcane, sunflower, sweet potato, tomato, turnip and walnut. Further examples include aquatic plants with roots - for example Duckweed (Family Lemnaceae).

In some examples, expression by the plant of the molecule rhizosecreted in the method of the invention may be 'induced' by the application of an elicitor or inducer. For example, US 2002/0132021 discloses elicitors or inducers that induce production of chemical compounds including those collected from root exudates.

The molecule rhizosecreted in the method of the invention may also be produced by a plant that has been genetically engineered to secrete said molecule (a transgenic plant). For example, the plant may be engineered to express a recombinant protein or glycoprotein that is itself rhizosecreted or, alternatively, that facilitates the production and/or rhizosecretion of a downstream product (such as a secondary metabolite). Recombinant proteins of interest that may be rhizosecreted in the method of the invention include therapeutic proteins and enzymes useful in industrial process, such as antibodies (and antigen-binding fragments thereof), blood product substitutes (e.g. haemoglobin, interferons, interleukins, enkephalins, coagulation factors, albumin, anticoagulants (e.g. human protein C and hirudin)), hormones (e.g insulin, somatotropin, erythropoietin, human growth hormone, human epidemal growth factor, trout growth factor, and human granulocyte-macrophage colony-stimulating factor), cytokines, vaccines, and other proteins (e.g alpha- 1 -antitrypsin, glucocerebrosidase, human homotrimeric collagen, human aprotinin, angiotensin- 1 -converting enzyme, alpha-trichosanthin from TMV-Ul subgenomic coat protein, human muscarinic cholinergic receptor, human placental alkaline phosphatase, defensins, elastin-like polypeptides (ELPs) and fusions of ELPs with other proteins).

The rhizosecreted vaccine may be hepatitis virus B surface antigen, malaria parasite antigen, rabies virus protein or epitope (e.g. rabies virus glycoprotein), E, coli heat-labile enterotoxin, human rhinovirus 14 (HRV- 14) epitope, HIV protein or epitope, norwalk virus capsid protein, diabetes-associated autoantigen, mink enteritis virus epitope, foot and mouth disease virus VPl structural protein, cholera toxin B epitope, human insulin-cholera toxin B subunit fusion protein, human cytomegalovirus glycoprotein B, or antigen of S. mutans or respiratory syncytial virus.

The rhizosecreted antibody may be monoclonal, from any species and produced by any method. The antibody may be humanised. The antibody may be of the IgG, IgM, IgA, IgD or IgE isotype. The antibody may be full-size, a Fab fragment, an Fc fragment, a single-chain antibody (scFv) fragment, a bi-specific scFv fragment, a camelid antibody or fragment thereof, a shark antibody or fragment thereof, or a chimeric antibody comprising a combination of such entities. Examples include Guy's 13 (a mouse monoclonal antibody that recognizes streptococcal antigen I/II, a major cell surface glycoprotein of

Streptococcus mutans) and 4E10 (a human monoclonal antibody that recognises the HIV gp41 antigen).

Secondary metabolites are typically organic compounds. The secondary metabolites that may be rhizosecreted in the method of the invention include alkaloids, such as nicotine, nornicotine, anabasine, anatabine, anatalline, N- formyl-nornicotine, myosmine, nicotyrine, cotinine, hyoscyamine, atropine, cocaine, codeine, morphine and tetrodotoxin; terpenoids such as azadirachtin, artemisinin, tetrahydrocannabinol, steroids and saponins; isoflavonoids such as genistein; glycosides such as nojirimycin and glucosinolates; phenols such as resveratrol; phenazines such as pyocyanin and phenazine-1-carboxylic acid (and derivatives); polyketides such as erythromycin and discodermolide; fatty acid synthase products such as FR-900848, U-106305 and phloroglucinols; non- ribosomally produced peptides such as vancomycin, thiostrepton, ramoplanin, teicoplanin, gramicidin and bacitracin; and epothilone. The method of the invention can be carried out with a plant that has been genetically engineered to increase the rhizosecretion of a molecule, including (but not limited to) expression of a protein, overexpression of a protein, root-specific expression of a protein, and/or expression of a protein fused to a signalling peptide that ensures that the protein is directed to the endoplasmic reticulum and the default secretory pathway.

Plant hormones, also known as plant growth regulators (PGRs) or phytohormon.es, are molecules that regulate plant growth. They are signal molecules that act alone or in concert with other plant hormones at sites local or distant to sites of production. Plant hormones typically affect the growth, development and differentiation of cells and tissues.

The plant hormone used in the invention can be an auxin, a cytokinin, abscisic acid (ABA), a gibberellin, a brassinolide, salicylic acid, a jasmonate, a signalling peptide, systemin, a polyamine or nitric oxide.

Cytokinins typically influence cell division and shoot and root morphogenesis. The cytokinin used in the invention may be an adenine-type cytokinin, for example kinetin, zeatin or 6-benzylaminopurine, or it may be a phenylurea-type cytokinin, such as diphenylurea or thidiazuron (TDZ).

The plant hormone is preferably an auxin. Typically, auxins play an essential role in coordination of growth and behavioural processes in the plant life cycle. Auxins can promote the formation of lateral and adventitious roots by promoting cell division and cell elongation, the latter through e.g. increasing cell wall extensibility by breaking hydrogen bonds between polysaccharide components of the cell wall. Auxins typically have an aromatic ring and a carboxylic acid group. The auxin used in the invention can be a naturally-occurring auxin, such as indole-3 -acetic acid (IAA), 4-chloro-indoleacetic acid (4-Cl-IAA), indole-3- butyric acid (IBA) and 2-phenylacetic acid (PAA), or a synthetic auxin, such as 2,4-dichlorophenoxyacetic acid (2,4-D), α-naphthalene acetic acid (α-NAA, or NAA), 2-methoxy-3,6-dichlorobenzoic acid (dicamba), 4-amino-3,5,6- trichloropicolinic acid (tordon or picloram), 2,4,5-trichlorophenoxyacetic acid (2,4,5,-T) and α-(p-chlorophenoxy)-isobutyric acid (PCIB). The auxin used in the invention is preferably α-NAA.

In one embodiment of the invention, the plant hormone is applied to the exterior of the roots of the plant, for example by adding the plant hormone to the soil surrounding the roots of the plant, or by adding the plant hormone to a hydroponic solution that is in contact with the roots of a plant grown hydroponically. Hydroponically grown plants are grown in solution, and the plant may be supported in this solution by inert mediums such as perlite, gravel or mineral wool. Typically, the hydroponic solution comprises inorganic ions that are essential for the plant to survive, including those that provide one or more of the following elements: nitrogen, phosphorus, potassium, calcium, magnesium, sulfur, boron, chlorine, iron, manganese, zinc, copper and nickel. The solution may comprise sources of any combination of these elements (e.g. all 13 elements, or 2 to 13, 4 to 13, 7 to 13, or 10 to 13 of the elements). Preferred solutions comprise sources of at least nitrogen and phosphorous. One advantage of using the method of the invention with hydroponically grown plants is that the rhizosecreted molecule can be readily collected from the hydroponic solution. For example, WO 98/49267 discloses recovery of chemical compounds from hydroponic culture. WO 99/38990 describes recovery of a recombinant protein from a transgenic plant or plant part (including roots) in contact with an aqueous medium.

The plant hormone may be applied as a single dose, as multiple doses given at different times, or as a continuous infusion. Where multiple doses are given, the frequency of doses may, for example, be from once a day to once a month, once a day to once every two weeks or once a day to once a week. The dosage of the plant hormone to be applied will depend on the selected plant hormone, and the skilled person will readily be able to identify a dosage of a particular plant hormone that is effective to increase rhizosecretion of a particular molecule. For example, the plant hormone may be applied in doses giving a final concentration of between 0.01 and 10 mg/L, preferably between 0.1 and 10 mg/L, more preferably between 0.5 and 5 mg/L, most preferably at 1 mg/L.

In an alternative embodiment, the plant hormone is applied to the interior of the roots of the plant, for example by producing the plant hormone in the roots of said plant. This may be carried out, for example, by overexpressing within the plant a protein or proteins that contribute to the biosynthesis of the plant hormone. For example, WO 90/13658 discloses the expression of an auxin in a plant by introducing a fragment of Agrobacterium T-DNA that encodes biosynthetic genes of that auxin. Preferably, said protein or proteins are preferentially expressed in the roots, by means of a root-specific promoter. Alternatively, or in addition, said protein or proteins are expressed using a controllable promoter. Such a controllable promoter may be repressed or activated by application of an exogenous compound.

Typically, rhizosecretion of a molecule is measured by its yield or rate of production. Yield may be defined in terms of mass per volume of hydroponic solution (e.g. ng/lOOμl solution). Rate of production maybe defined in terms of mass of molecule per unit of root mass per unit of time (e.g. μg/g root dry weight/24h). In the present invention, an increase in rhizosecretion is defined by any increase in the secretion of a molecule from the roots of a plant as measured, for example, by an increase in yield or, preferably, rate of production. Preferably, the increase is at least 2-fold, more preferably at least 3 -fold, more preferably at least 5-fold compared to rhizosecretion in the absence of a plant hormone. The maximum increase achieved by the invention has yet to be determined, but we suggest an increase of 10-fold or 20-fold as an upper limit.

The molecule rhizosecreted in the method of the invention can be recovered by removing the medium surrounding the roots of the plant, which will typically be the hydroponic solution in which the plant is grown and into which the molecule is rhizosecreted. Alternatively, the rhizosecreted molecule can be recovered by immobilising the molecule within the medium surrounding the roots of the plant and then removing the immobilised molecule from said medium. For example, the solution may be passed over a matrix comprising a ligand that binds to the rhizosecreted molecule (affinity chromatography), e.g. protein G binds to some classes of antibody. The rhizosecreted molecule can then be removed from the ligand.

Once the molecule is recovered, appropriate standard methods in the art can be used to concentrate and/or purify the rhizosecreted molecule (e.g. standard protein purification methods for the isolation of a protein, such as affinity chromatography). Preferably, the purified molecule is at least 50%, by weight, free from other organic molecules such as those contained within the recovered medium, more preferably at least 60%, more preferably at least 70%, more preferably at least 80%, more preferably at least 90%, more preferably at least 95%, and most preferably at least 99%. The recovered and/or purified molecule can be formulated into a composition, typically a pharmaceutical composition comprising the molecule and a pharmaceutically acceptable carrier or diluent.

The method of the invention may further comprise the application of an additional plant hormone to the roots of the plant, and this application may occur simultaneously or sequentially with respect to the first plant hormone. The additional plant hormone may be a different type of plant hormone from the first, or it may be of the same type. For example, if the first plant hormone is an auxin the alternative plant hormone may be a cytokinin or it may be a different auxin.

The method of the invention may further comprise the application of an additional cell wall permeabilisation compound to the roots of the plant, and this application may occur simultaneously or sequentially with respect to the plant hormone(s). The cell wall permeabilisation compound maybe selected from, for example, an enzyme that degrades a plant cell wall component, such as, cellulase (which degrades cellulose) or pectolyase (which degrades pectin). Preferably the cell wall permeabilisation compound is cellulase or pectolyase, even more preferably a combination of cellulase and pectolyase.

The cell wall permeabilisation compound may also be selected from a non- enzymatic compound that is known to increase the permeability of the cell wall. Such compounds include those involved in the hydroxyl radical ( ^» OH) generation system of the plant apoplast (the "apoplast ^» OH generation system"), and any combination of these compounds. The apoplast ^» OH generation system compound may be selected from a reducing agent, such as ascorbate or NADH (preferably ascorbate), a transition metal ion, such as Cu ²⁺ (or its reduced form, Cu ⁺ ) or Fe ³⁺ (or its reduced form, Fe ²⁺ ) (preferably Cu ²⁺ ), and hydrogen peroxide. Preferably a combination of any two such types of compound is used, more preferably a combination of all three types of compound, even more preferably a combination of ascorbate, Cu ²⁺ and hydrogen peroxide.

An hydroxyl radical ('OH) can be generated from a reaction of hydrogen peroxide (H ₂ O ₂ ) with a reduced transition metal ion (Fenton reaction), Cu ⁺ being 60 times more effective than Fe2 ⁺ (Halliwell and Gutteridge, 1999). The plant apoplast usually contains oxygen, ascorbate (AH ₂ ) and Cu ²⁺ , a combination that readily generates ^» OH. Fry (1998) proposed that ascorbate reduces oxygen to hydrogen peroxide and Cu ²⁺ to Cu ⁺ (Le. producing the reactants of a Fenton reaction):

AH ₂ + O ₂ ^■ » A + 2H ₂ O ₂

AH ₂ + 2Cu ^{2+ •} * A + 2H ⁺ + 2Cu ⁺

Fry (1998) showed that H ₂ O ₂ alone causes slow scission of cell wall xyloglucans, whereas ascorbate alone induces faster scission. The most dramatic effect was produced by 5mM ascorbate combined with 5 mM H ₂ O ₂ and addition of Cu ²⁺ (0.6-20 μM).

The results of Example 11 demonstrate that neither i) enzymes that degrade plant cell walls nor ii) components of the apoplast ^» OH generation system, when used alone, increase rhizosecretion in tobacco. In contrast, however, both entities increase the rhizosecretion induced by plant hormone. Indeed, the results demonstrate a synergistic effect between plant hormone and these entities for rhizosecretion.

The skilled person will readily be able to identify a dosage of a particular cell wall permeabilisation compound that is effective to increase rhizosecretion of a particular molecule produced using a particular plant hormone, more preferably to increase rhizosecretion to a level that is equal to the sum of the individual rhizosecretion rates seen with the plant hormone and the cell wall permeabilisation compound when used separately, and more preferably to increase rhizosecretion to a level that is greater than the sum of the individual rhizosecretion rates seen with the plant hormone and the cell wall permeabilisation compound when used separately (i.e. a synergistic level).

The invention further provides a solution as described above for growing a plant hydroponically comprising a plant hormone. Preferably, as described above, the solution comprises a source of one or more of the following elements: nitrogen, phosphorus, potassium, calcium, magnesium, zinc, copper, nickel, sulphur, boron, chlorine, iron and manganese.

Preferably, the hormone is selected from auxins, cytokinins, abscisic acid (ABA), gibberellins, brassinolides, salicylic acid, jasmonates, signalling peptides, systemin, polyamines and nitric oxide. Preferably, the auxin is selected from indole-3-acetic acid (IAA), 4-chloro-indoleacetic acid (4-Cl-IAA), indole-3- butyric acid (IBA), 2-phenylacetic acid (PAA), 2,4-dichlorophenoxyacetic acid (2,4-D), α-naphthalene acetic acid (α-NAA or NAA), 2-methoxy-3,6- dichlorobenzoic acid (dicamba), 4-amino-3,5,6-trichloropicolinic acid (tordon or picloram), 2,4,5-trichlorophenoxyacetic acid (2,4,5,-T) and α-(p-chlorophenoxy)- isobutyric acid (PCIB). Preferably the plant hormone is present at a concentration of between 0.01 and 10 mg/L, preferably between 0.1 and 10 mg/L, more preferably between 0.5 and 5 mg/L, most preferably at 1 mg/L. In a further embodiment of the invention, the solution further comprises a cell wall permeabilisation compound. The cell wall permeabilisation compound may be selected from, for example, an enzyme that degrades a cell wall component, such as, cellulase or pectolyase (preferably both cellulase and pectolyase), or a non-enzymatic compound that is known to increase the permeability of the cell wall, such as an apoplast OH generation system compound (selected as described above).

The invention further provides a dry powder which, when dissolved (e.g. in water), provides a solution for growing a plant hydroponically as described above.

The invention further provides a hydroponic plant culture comprising a plant and a solution as described above.

The invention further provides a kit comprising a powder which, when dissolved (e.g. in water), provides a solution for growing a plant hydroponically, and a plant hormone. Preferably, the powder comprises a source of one or more of the following elements: nitrogen, phosphorus, potassium, calcium, magnesium, zinc, copper, nickel, sulphur, boron, chlorine, iron and manganese. Preferably, the powder comprises Murashige and Skoog Media (Murashige and Skoog, 1962, Drake et al., 2003).

The hormone may be selected from auxins, cytokinins, abscisic acid (ABA), gibberellins, brassinolides, salicylic acid, jasmonates, signalling peptides, systemin, polyamines and nitric oxide. Preferably, the auxin is selected from indole-3-acetic acid (IAA), 4-chloro-indoleacetic acid (4-Cl-IAA), indole-3- butyric acid (IBA), 2-phenylacetic acid (PAA), 2,4-dichlorophenoxyacetic acid (2,4-D), α-naphthalene acetic acid (α-NAA or NAA), 2-methoxy-3,6- dichlorobenzoic acid (dicamba), 4-amino-3,5,6-trichloropicolinic acid (tordon or picloram), 2,4,5-trichlorophenoxyacetic acid (2,4,5,-T) and α-(p-chlorophenoxy)- isobutyric acid (PCIB). The hormone of the kit may be provided in solid form or in solution (e.g. in water or ethanol). Plant hormones in solid form may need to be dissolved in an appropriate solvent (e.g. ethanol) before it can be solubilised into a hydroponic solution.

In a further embodiment of the invention, the kit further comprises a cell wall permeabilisation compound. The cell wall permeabilisation compound may be selected from, for example, an enzyme that degrades a cell wall component, such as, cellulase or pectolyase (preferably both cellulase and pectolyase), or a non- enzymatic compound that is known to increase the permeability of the cell wall, such as an apoplast *OH generation system compound (selected as described above).

EXPERIMENTAL METHODS - EXAMPLE 1

Establishment of hydroponic cultures

Seeds from CV-N and F52 lines were surface sterilised and seedlings established in hydroponic culture in Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) as previously described (Drake et al., 2003). When seedlings had reached a height of approximately 5cm, all medium was removed and replaced with 20ml of fresh medium. Hydroponic cultures were randomly assigned to three groups (Table 1); plants in Group 1 were given MS medium, those in Group 2 and 3 MS medium containing lmg/L NAA. Medium for plants expressing Guy's 13 also contained 8g/L gelatin which had previously been demonstrated to stabilise antibody in MS medium (Drake et al., 2003). In all subsequent culture, gelatin was included in the medium for plants expressing Guy's 13. Hydroponic cultures were left for a period of 26d. During this 26d period, plants in Group 3 were given two further applications of NAA at 7 and 14d to a final concentration of lmg/L. In these additions, NAA was added at a concentration of 0.1mg/ml in MS medium i.e. 200μl of MS containing 0,02mg NAA. hi Groups 1 and 2, 200μl of MS lacking NAA was added to hydroponic cultures at the 7 and 14d time points. WT negative control cultures were grown under the same conditions as Group 3. Table 1. Culture medium for plants expressing Guy's 13 or CV-N in Groups 1-3.

Collection ofhydroponic samples and measurement of root dry weight

After 26d, medium was again removed and 40ml of fresh medium added. MS medium lacking NAA was added to Group 1 and medium containing lmg/L NAA added to Groups 2 and 3. After 7d, medium was removed and stored at -20 ⁰ C for future analysis. A further 40ml of fresh MS medium (with or without NAA as described previously for the 3 groups) was added to the plant cultures, collected after 24h and stored at -2O ⁰ C. Roots were immediately removed from the plants, dried at room temperature for 4 days and weighed. Hydroponic samples were defrosted and centrifuged at 20,000g/10 mins/4°C prior to analysis.

Analysis of culture samples ELISA

For Guys 13, functional ELISAs were based on binding of the antibody to a recombinant fragment of Streptococcal antigen SA VIl that contained the epitope for Guy's 13 (method modified from Drake et al., 2003). ELISA plates (Nunc Immobilon, UK) were coated with a predetermined optimal concentration of SA I/II in PBS. Plates were incubated for 2h at 37 ⁰ C, washed five times with 0.1%v/v Tween 20 in water (H-T20), dried and then blocked with 5% w/v non-fat dried milk (NFDM) in PBS for 2h at 37°C. Plates were washed and dried as previously described, pooled samples from a minimum of 9 hydroponic plant cultures were added, 100 μl/well, and a two-fold dilution series undertaken. Samples were incubated for 16h at 4°C. Plates were washed and dried and secondary antibody, affinity purified horseradish peroxidase labelled anti-mouse IgGl (The Binding Site, UK) ₅ added. Incubation was for 2h at 37 ⁰ C. After washing and drying, tetramethylbenzidine dihydrochloride peroxidase substrate was added (Sigma), colour development allowed to proceed for 10 min, 2M H ₂ SO ₄ added to each well and optical density read at 450nm on a plate reader (Tecan Sunrise).

For CV-N detection, plates were coated with lOOng/well of antigen gpl20 (strain MN, subtype B, EVA646, MRC Centralised Facility for AIDS Reagents, Potters bar UK), 50μl/well. Plates were blocked with 2.5% w/v bovine serum albumin. Detection was with rabbit anti-CV-N polyclonal rabbit serum (2h incubation,

37 ⁰ C) followed by peroxidase-conjugated anti-rabbit antibody (The Binding Site, UK) (2h incubation, 37°C). All other stages were as described previously for the Guy's 13 ELISA. Concentrations of Guy's 13 and CV-N were determined by comparison with known concentrations of Guy's 13 from hybridoma culture supernatant and recombinant E.coli derived CV-N, which were serially diluted in MS medium (with or without gelatin) lacking NAA (as this had previously been shown not to affect the ELISA reading-data not shown), in the same ELISA plates as the samples from hydroponic plants.

Western Blotting

Hydroponic medium (20μl) pooled from a minimum of 9 plant cultures was mixed with 5 μl of loading buffer (0.05M tris HCl pH 6.8, 8 % v/v glycerol, 1.6 % w/v sodium dodecyl sulphate 0.08 % w/v bromophenol blue). The loading buffer for the Guy's 13 reducing gel also contained 0.7M β-mercaptoethanol. Samples were boiled for 3 mins, centriraged 20,000 g/4°C/10 min and run on an 8% v/v polyacrylamide gel at 150V:20mA. A 10% v/v gel was used in the Guy's 13 reducing PAGE. Separated protein bands were blotted onto nitrocellulose using semi-dry transfer (Amersham), running for Ih at 50V and 4OmA. Nitrocellulose blots were blocked with 5% w/v NFDM in TBS at 4 ⁰ C for 16h. For the Guy's 13 non-reducing western blot, the detecting antibody was horseradish peroxidase- labelled anti-mouse kappa antibody (Jackson Immunoresearch, USA). In the reducing blot, a mixture of peroxidase labelled anti-mouse kappa and anti-mouse gamma chain antibodies (Jackson Immunoresearch) were used. For CV-N detection was with rabbit anti-CV-N followed by horseradish peroxidase-labelled donkey anti-rabbit antibody. For the Guy's 13 blots, washing was with TBS containing 0.1% v/v Tween 20 whilst CV-N blots were washed with TBS containing 0.2% v/v Tween 20. Blots were developed using the ECL Plus Western Blotting Detection System (Amersham).

Experimental Design and Statistical Analysis

Plants were randomly assigned to each experimental group. ELISAs and western blots were performed on samples pooled from a minimum of 9 hydroponic plant cultures. Samples were pooled by taking equal volumes from the hydroponic plant cultures and mixing. Root dry weights were compared using Student's t- test.

EXPERIMENTAL METHODS - EXAMPLES 2 TO 6

Plant material

The Guy's 13, 4E10 and CV-N transgenic plant lines were produced by transformation of tobacco (Nicotiana tabacum cv. Xanihiϊ) with Agrobacterium LBA4404 using a derivative of vector pMON530 (Horsch et al., 1985). The binary vector contains a cauliflower mosaic virus 35S promoter, a 60bp murine Ig K leader sequence to direct the recombinant protein to the plant endomembrane system and the nopaline synthase 3 'terminator. The vector also contains a bacterial selectable marker conferring resistance to spectinomycin and the npt-ll gene for selection of transformed plant tissue on medium containing kanamycin (Sexton et al., 2005).

All plants used in these studies were from a single transgenic line. The CV-N plants were from the Tl generation and were homozygous for the CV-N gene. Guy's 13 and 4E10 plants were both produced by initially expressing the heavy and light antibody chain genes in separate plants followed by sexual crossing to generate transgenic plants expressing assembled antibody. The Guy's 13 and 4E10 plants used in this study were from the T2 generation and were both homozygous for the heavy and light antibody chain genes.

Establishment ofhydroponic cultures and collection of medium samples Seeds from the transgenic tobacco plant lines expressing either Guy's 13, 4E10 or cyanovirin-N were surface sterilised and seedlings established in hydroponic culture in Murashige and Skoog medium (Murashige and Skoog, 1962) as previously described (Drake et al., 2003). When seedlings had reached a height of approximately 5cm, all medium was removed and replaced with 20ml of fresh medium. The medium for plants expressing Guy's 13 and MAb 4E10 contained 8 g/L gelatin which had previously been demonstrated to stabilise antibody in MS medium (Drake et al., 2003). The medium was supplemented with the lmg/L α- naphthalene acetic acid (NAA; Sigma, Poole, Dorset, UK). For the Guy's 13 plants, additional experiments were undertaken in which medium was supplemented with either lmg/L indole-butyric acid (IBA; Sigma), 6- benzylaminopurine (BAP; Sigma) or kinetin (KIN; Sigma). In control plants, plant growth regulator was omitted. Hydroponic cultures were left for 26 days. During this period, some plants were given two further applications of the relevant plant growth regulators at 7 and 14 days to a final concentration of 1 mg/L.

After 26 days, the medium was again removed and 40ml of fresh medium (with or without appropriate supplements) added. After 7 day culture, medium was removed and stored at -2O ⁰ C for future analysis. A further 40ml of fresh MS medium (with or without appropriate supplements) was added to the plant cultures, collected after 24h and stored at -2O ⁰ C. Roots were immediately removed from the plants, dried at room temperature for 4 days and weighed.

Analysis of medium samples Hydroponic samples were defrosted and centrifuged at 20,000 g for 10 mins at 4 ⁰ C prior to analysis. ELISA

For Guy's 13, functional ELISAs were based on binding of the antibody to a recombinant fragment of streptococcal antigen SA I/II (van Dolleweerd et al., 2003) that contains the epitope for Guy's 13. ELISA plates (Thermofisher, Roskilde, Denmark) were coated with a predetermined optimal concentration of SA I/II in PBS. Plates were incubated for 2h at 37 ⁰ C, washed five times with 0.1% v/v Tween 20 in water (H-T20), dried and then blocked with 5% w/v non-fat dried milk (NFDM) in PBS for 2h at 37 ⁰ C. Plates were washed and dried as above, pooled samples from hydroponic plant cultures were added, 100 μl/well, and a two-fold dilution series undertaken. Samples were incubated for 16h at 4 ⁰ C. Plates were washed and dried and the secondary antibody, an affinity purified horseradish peroxidase labelled anti-mouse IgGl (The Binding Site, Birmingham, West Midlands, UK), added. Incubation was for 2h at 37 ⁰ C. After washing and drying, tetramethylbenzidine dihydrochloride peroxidase substrate was added (Sigma), and colour development allowed to proceed for 10 min. The reaction was stopped with 2M H ₂ SO ₄ and the optical density read at 450 nm on a plate reader (Tecan Sunrise, Tecan, Reading, Berkshire, UK).

For CV-N detection, plates were coated with lOOng/well of antigen gpl20 (strain MN, subtype B, EVA646, MRC Centralised Facility for AIDS Reagents, Potter's Bar, Hertfordshire, UK), 50 μl/well. Plates were blocked with 2.5% w/v bovine serum albumin. Detection was with rabbit anti-CV-N polyclonal rabbit serum (2h incubation, 37 ⁰ C) followed by peroxidase-conjugated anti-rabbit antibody (The Binding Site) (2h incubation, 37 ⁰ C). All other stages were as described previously for the Guy's 13 ELISA.

For the detection of the 4E10 antibody, plates were coated with a peptide from gp41 containing the 4E10 epitope (100 ng/well). Detection was with horseradish peroxidase labelled anti-human IgG (The Binding Site). Blocking, washing and detection were as previously described for Guy's 13. Concentrations of Guy's 13, 4E10 and CV-N in hydroponic medium were determined by comparison with standards of known concentration in the ELISA.

Western Blotting Hydroponic medium (20ml) from pooled medium samples was mixed with 5 μl of loading buffer (0.05M Tris HCl pH 6.8, 8 % v/v glycerol, 1.6 % w/v sodium dodecyl sulphate, 0.08 % w/v bromophenol blue). Samples were boiled for 3 mins, centrifuged 20,000 g at 4 ⁰ C for 10 mins and run on a polyacrylamide gel. For Guy's 13 and 4E10 antibody detection, either a 4-13% gradient or an 8% v/v gel was used: separated protein bands were blotted onto nitrocellulose using semi- dry transfer (GE Healthcare, Amersham, Buckinghamshire, UK). Nitrocellulose blots were blocked with 5% w/v NFDM in TBS at 4 ⁰ C for 16 h. Blots were washed with TBS containing 0.1% v/v Tween 20. For the Guy's 13 non-reducing western blot, the detecting antibody was horseradish peroxidase-labelled anti- mouse kappa chain antiserum (Jackson Immunoresearch, West Grove, PA, USA). MAb 4E10 was detected with peroxidase-labelled anti-human kappa chain antiserum (Sigma). For CV-N, a 12% v/v PAGE gel was used to separate protein bands, blots were washed with TBS containing 0.2% v/v Tween 20 and detection was with a rabbit anti-CV-N antiserum, followed by horseradish peroxidase- labelled donkey anti-rabbit antiserum. AU blots were developed using the ECL Plus Western Blotting Detection System (GE Healthcare).

Experimental Design and Statistical Analysis

Plants were randomly assigned to each experimental group. ELISAs and western blots were performed on samples pooled from a minimum of 9 hydroponic plant cultures. Samples were pooled by taking equal volumes from the hydroponic plant cultures and mixing. Root dry weights were expressed as mean + standard deviation and were compared using one way ANOVA followed by Tukey's HSD test. Zymogram analysis of peptidases from N. tabacum leaf extracts and hydroponic medium

Leaf intercellular fluid was recovered from 4-week old plants by an infiltration- centrifugation method according to Dellanoy et al. (2008). Proteins were solubilised in Laemmli buffer (2% SDS, 0.125M Tris pH6.8, 10% glycerol,

0.002% bromophenol blue, 6OmM dithiothreitol) for 30 mins at 2O ⁰ C. SDS-PAGE was performed in 8% polyacrylamide gels containing 0.1% gelatin. After migration, the gel was washed three times in 2.5% Triton X-100, incubated overnight at 37 ⁰ C and then at room temperature with gentle agitation in 1% Triton X-100, 5mM CaCl ₂ , 1 OmM MES, pH 5.5 and finally stained with Coomassie Brilliant Blue G250.

Glycan analysis

N-Glycosylation was analysed by LC-ESI-MS of tryptic glyco-peptides, as was described previously (Stadlmann et al., 2008). Briefly, coomassie-stained SDS- PAGE bands of the respective MAb Guy's 13 heavy-chains were excised, de- stained, S-carbamidomethylated and subjected to tryptic in-gel digestion. The extracted tryptic (glyco-)peptides were then subjected to LC-ESI-MS analysis.

Affinity purification ofrhizosecreted Guy's 13 IgG

Approximately 200ml of hydroponic medium, taken from Guy's 13 plants after a 7d incubation, was thawed and gently warmed to 24 ⁰ C in a water bath. This reduced the viscosity of the solution (due to gelatin) sufficiently to allow for direct filtering over a 1.2μm membrane. The sample was then passed over a ImI antibody affinity column with a goat anti-mouse IgGl (heavy chain specific antibody as the ligand (Sigma) - at a flow rate of 1 ml/min. The column was subsequently washed with 40ml PBS before elution with 10x2ml fractions of 0.1M glycine pH2.5, which were immediately pH shifted to 7.5 with 2M TrisHCl - pH 7.5.

For silver staining of SDS-polyacrylamide gels, the PlusOne Silver Staining kit (GE Healthcare) was used. Briefly, acrylamide gels were fixed in 40% ethanol, 10% glacial acetic acid. They were sensitised in 30% ethanol, 0.2% w/v (weight/volume) sodium thiosulphate, 6.8% w/v sodium acetate and 0.125% w/v glutaraldehyde and washed with water. Staining was with 0.25 w/v silver nitrate solution, supplemented with 0.1ml formaldehyde (37% w/v) for every 250ml of solution. Gels were washed in water and developed with 2.5% w/v sodium carbonate, supplemented with 0.2ml formaldehyde (37% w/v) for every 250ml of solution. Upon sufficient development the reaction was stopped by decanting the developing solution and replacing with 3.65g of EDTA in 250ml water.

Western Blotting of purified IgG was as described above, using a 10% acrylamide gel under non-denaturing conditions and an HRPO labeled anti-Mouse IgG antiserum (The Binding Site) for detection.

EXAMPLES

Example 1 - increase of rhizosecretion of Guv's 13 antibody and Cyanovirin-N in tobacco using NAA

Summary In the present study, we included the plant growth regulator (a synthetic auxin), α- napthalene acetic acid (NAA), in the hydroponic culture medium and compared rhizosecretion rates of recombinant proteins to cultures in which NAA was omitted. Auxins promote cell elongation primarily by increasing cell wall extensibility via the breaking of hydrogen bonds between the polysaccharide components of the cell wall (Taiz, 1998).

The present study was conducted on two Nicotiana tabacum transgenic plant lines, one expressing the murine IgGIk antibody Guy's 13, and the other the anti- HIV microbicide cyanovirin-N (CV-N). Plants from each type were grown in hydroponic culture medium in the presence or absence of the plant growth regulator α-naphthalene acetic acid (NAA) for 26d. For each plant line there were three experimental groups: plants cultured in hydroponic medium lacking NAA, plants cultured in medium with a single initial addition of NAA and finally plants given multiple additions of NAA (3 in total). Samples were assayed 7d after addition of fresh medium to calculate yield of recombinant protein per volume of hydroponic medium. After the 7d culture period fresh medium was again added, samples were harvested after a further 24h and root dry weight measured to allow calculation of rhizosecretion rate in μg/g.root dry weight/24h.

For Guy's 13 plants, the antibody yield after 7d of culture in medium lacking NAA was 2.3ng/100μl compared to 40 ng/lOOμl for plants exposed to medium with a single addition of NAA. Plants given multiple additions of NAA produced an antibody yield of 95 ng/lOOμl. The corresponding figures for cyanovirin-N were 126, 448 and 625 ng/lOOμl. The maximum rhizosecretion rates were calculated as 58 μg/g. root dry weight/24h for Guys 13 and 766 μg/g.root dry weight/24h for cyanovirin-N, the highest figures so far reported for a full-length antibody and a recombinant protein respectively. Addition of NAA approximately doubled root dry weight, although this could not account for all of the increase in rhizosecretion of recombinant protein. Addition of NAA on several occasions did not significantly increase root mass compared to a single initial application.

Results

The yields of Guy's 13 and CV-N expressed in ng/lOOμl of hydroponic medium collected 24h hours and 7d after addition of fresh medium are presented in Table 2. Yields were calculated by comparison to known standard concentrations of Guy's 13 from hybridoma culture supernatant and recombinant CV-N in ELISA. Addition of NAA increased the rhizosecretion of recombinant Guy's 13 and CV- N (Group 3 > Group 2 > Group 1) and yields were higher after 7d than 24h, with the exception of Guy's 13 plants not given NAA (Group 1). Table 2. Plants were grown in hydroponic culture in MS medium until they reached a suitable size. Medium was replaced and all plants were cultured for a further 26d. After the 26d period, medium was removed and replaced with fresh medium which was removed for analysis after 7d. Fresh medium was then again added for a further 24h and then removed for analysis. Recombinant protein concentrations were calculated by ELISA using Guy's 13 and CV-N standards of known concentration.

Recombinant protein yield (ng/lOOμl hydroponic medium)

Time Guy's 13 plants CV-N plants Group 1 Group 2 Group 3 Group 1 Group 2 Group 3

24h 4.2 22 25.6 47 140 320

7d 2.3 40 95 126 448 625

A two-fold dilution series of the two recombinant proteins at the two time points are demonstrated in Figures la-Id. The dilution series corroborates the results presented in Table 2.

Table 3 contains average root dry weights of plants cultured in the 3 groups. Addition of NAA approximately doubled root dry weight in plants expressing Guy's 13 and CV-N. Root dry weight was virtually identical in plants given 1 (Group 2) or 3 (Group 3) additions of NAA.

Table 3. After the final 24h culture period, roots were removed from the plants, dried and weighed. Results are expressed as mean + standard deviation. There is a significant difference in the weights of roots from plants cultured with or without NAA (t=6.5, p<0.001, Student's t-test) but no significant difference in the weights of roots from plants given 1 (Group 2) or 3 (Group 3) applications of NAA. Root dry weight (g)

Guy's 13 plants (g) CV-N plants (g)

Groupl Group 2 Group 3 Groupl Group 2 Group 3

0.09±0.04 0.19±0.07 0.18±0.06 0.08+0.04 0.18+0. 05 0.17±0.05

Table 4 depicts the rhizo secretion rate of recombinant proteins in the 3 groups expressed as μg/g.root dry weight/24h. The rhizosecretion rate of both recombinant proteins increases with addition of NAA (Group 3 > Group2 > Groupl).

Table 4. Plants were given fresh medium and this was removed after a period of 24h. Roots were removed and dry weight measured. Recombinant protein concentration was calculated by ELISA using Guy's 13 and CV-N standards of known concentration and expressed as μg/g. root dry weight/24h.

Rhizosecretion rate (μg/g- root dry weight/24h)

Guy's 13 plants CV-N plants

Group 1 Group 2 Group 3 Group 1 Group 2 Group 3

19 46 58 235 311 766

Western blots of hydroponic medium samples from plants expressing Guy's 13 are shown in Figures 2 (non-reducing) and 3 (reducing conditions). Figure 4 is a non-reducing western blot of hydroponic samples from CV-N plants. Samples for all blots consisted of pooled medium aliquots from a minimum of 9 hydroponic plant cultures. In Figure 2, immunoreactive bands corresponding to fully-assembled antibody were detected in each of the transgenic plant medium samples tested (lanes 1,2,3,7,8,9), which matched that found in the positive control (Guy's 13 IgGIk hybridoma culture supernatant, lane 5). No immunoreactive bands were detected in the samples from non-transformed WT plants (lanes 4 and 10). Four other bands were also observed which correspond to assembly intermediates or degradation products.

In Figure 3, immunoreactive bands corresponding to antibody heavy chain (50 kDa) and light chain (25 kDa) were detected in transgenic plant hydroponic samples (lanes 1,2,3,7,8,9) which matched those observed in the positive control, although bands produced by samples from plants not supplied with NAA (lanes 1 and 6) are weak and almost below the level of detection of the blot. The positive control had two bands at 50 kDa which represent different glyco forms of the antibody heavy chain (Cabanes-Macheteau et al., 1999). A third band at 35 kDa corresponding to a degradation product was also observed in hydroponic samples. Immunoreactive bands were not observed in samples from WT control plants (lanes 4 and 9).

In Figure 6B, immunoreactive bands at approx.l 1 kDa were observed in samples from all hydroponic cultures (lanes 1,2,3,7,8,9) which corresponded to the positive control (recombinant CV-N, lane 5). Bands were not observed in samples from WT plants (lanes 4 and 10).

Example 1 - Discussion

In the present study, the effect of the auxin plant growth regulator NAA on rhizosecretion of the recombinant proteins Guy's 13, a mouse monoclonal antibody, and CV-N from transgenic hydroponic plant cultures was investigated. Both of these molecules had been shown to be rhizosecreted from the roots of transgenic tobacco plants in previous studies in our research group (Drake et al., 2003; Sexton et al., 2005). Cell wall pore size is thought to exclude the rhizosecretion of proteins larger than 30-50 kDa (Carpita et al., 1979), so the secretion of CV-N (11 kDa) is expected, but the passage of Guy's 13 (150 kDa) through the cell wall is presumably due to the highly flexible nature of antibody molecules (Drake et al., 2003).

Transgenic plants expressing these proteins were allocated to three groups (Table 1). In Group 1, the culture medium was MS lacking NAA, whilst in Group 2 and 3, lmg/ml NAA was added to the medium at the start of the culture period. Plants were cultured for 26d and during this period a further two applications of NAA were given to plants in Group 3. After the 26 d period, medium was replaced and the plants were cultured for a further 7d after which medium was harvested and stored for analysis. Plants were then cultured in fresh medium for 24h, medium was again harvested and the roots were removed from the plants and their dry weight measured. This study design permitted measurement of recombinant protein yield at 24h and 7d after addition of fresh medium in mass/volume of medium, and also allowed an accurate assessment of rhizosecretion rate in μg/g.root dry weight/24h. The former measurement is of practical value as yields can be readily calculated, whereas the latter has been most commonly employed previously (Borisjuk et al., 1999; Drake et al., 2003) and consequently the calculation of this figure allows formal comparisons to be made between our results and past studies.

For both transgenic plant lines, yields were lowest at 24 h and 7d when culture was in medium lacking NAA (Group 1) and highest in Group 3 in which plant received multiple additions of NAA (Table 2, Figure la-d). Group 2 plants which had received a single initial application of NAA demonstrated intermediate rhizosecretion levels. For the 24h medium harvest, plants in Group 2 expressed approximately 5 fold levels of Guy's 13 and 3 fold CV-N compared to plants on Group 1. For Guy's 13, at the 24 h harvest, rhizosecretion yields were slightly higher in Group 3 (25.6ng/100μl hydroponic medium) compared to Group 2 (22ng/l OOμl) but for CV-N a 2.3 fold increase was observed. For medium harvested after 7d, the same pattern of results was observed (see Table 2) although for Guy's 13 the difference between Groups 2 and 3 was more pronounced than the result obtained after 24h of culture. The yield of recombinant protein increased between 24h and 7d of culture with the exception of plants in Group 1 expressing Guy's 13. Previous studies have demonstrated that recombinant proteins in culture are subject to proteolysis by secreted proteases (Komarnytsky et al., 2006) and presumably in this group this effect was greater than the accumulation of Guy's 13 by rhizosecretion. Addition of NAA however allowed a net increase of Guy's 13 between the 24 and 7d period. In all cases, the yield of recombinant protein harvested after 7d was less than 7 times greater than the 24h figure (Table 2), suggesting that some proteolytic breakdown may be occurring.

From these figures we can calculate that a single transgenic hydroponic culture containing 40ml of medium could yield a maximum of 10.24 μg of Guy's 13 antibody and 128 μg of CV-N in 24h. The latter result is particularly impressive and probably reflects not only the ability of CV-N to passage easily through the cell wall, but also the exceptional stability of this protein. CV-N is resistant to boiling, multiple freeze-thaw cycles, dissolution in organic solvents and treatment with high salt, detergent or hydrogen peroxide (Boyd et al., 1997).

A non-reducing western blot of Guy's 13 had an immunoreactive band corresponding to fully-assembled antibody, but also several other bands of different sizes. This pattern of multiple bands is characteristic of rhizosecreted antibody (Drake et al., 2003; Wongsamuth and Doran, 1997) and they represent assembly intermediates or breakdown products. The intensity of the immunoreactive bands in different Groups collected at the two time-points corresponded to previously calculated antibody concentrations (Table 2). In the reducing western blot of Guy's 13 (Figure 3), immunoreactive bands at 50 kDa and 25 kDa represent antibody heavy and light chain respectively and an immunoreactive band representing a breakdown product is also visible at approximately 35 kDa

Figure 6B is a non-reducing western blot of CV-N. A single band was observed at approximately 11 kDa. As already stated, some proteolytic breakdown of CV- N may be occurring, but any products would probably be too small to be visualised by this polyacrylamide gel electrophoresis. The intensity of the bands reflected the concentrations of CV-N secreted in the different Groups as previously calculated (Table 2).

The rhizosecretion rate calculated in μg/g.root dry weight/24h is shown in Table 4. The differences in Groups are not as pronounced as in Table 2 due to the larger root weight observed in plants treated with NAA (Table 3). Although, much of the increase in rhizosecretion of recombinant proteins after application of NAA is presumably due to this increase in root mass (and surface area), this cannot entirely explain the higher levels observed. For example, in both Guy's 13 and CV-N plants an approximate two-fold increase in root mass was observed following a single application of NAA, but after the 24h culture period a 5-fold increase in rhizosecretion of Guy's 13 and a 3 -fold increase in CV-N was observed. After 7d these differentials rose to 17 fold for Guy's 13 and 3.5 fold for CV-N. Furthermore, there was no significant difference in root masses between plants given either a single application (Group 2) or multiple applications (Group 3) of NAA (Table 3) and yet in each instance rhizosecretion of recombinant protein was higher in Group 3 (albeit the difference was minimal in Guy's 13 plants cultured for 24h). The optimal rhizosecretion rate of 766 μg/g.root dry weight/24h for CV-N (Table 4) represents, to the best of our knowledge, the highest reported figures for rhizosecretion of a recombinant protein, whilst the corresponding figure for Guy's 13 (58 μg/g.root dry weight/24h) is the highest reported for a full-length antibody.

Example 2 - increase of rhizosecretion of Guv's 13 antibody in tobacco using a number of different plant hormones

Hydroponic tobacco plants were cultured for 26 days in the presence of one of the plant growth regulators (plant hormones) - α-naphthalene acetic acid (NAA), indole-butyric acid (IBA), 6-benzylaminopurine (BAP) or kinetin (KIN). The plant growth regulators (PGRs) were added at lmg/L on days 0, 7 and 14. Fresh medium supplemented with the same plant growth regulator was used for the next 7 days of culture. The medium was changed again for the final 24hrs of culture. The yields of Guy's 13 Mab that accumulated in the hydroponic medium during the 24 hr and 7 day cultures were calculated by comparison to known standard concentrations of Guy's 13 (from hybridoma culture supernatant) in a functional antigen specific ELISA and are shown in Figure 4A. Note that the data for NAA is derived from the experiments of Example 1.

All of the PGRs improved antibody yield. The least effective was kinetin, and NAA had the greatest effect. Here, the yields were 0.26 and 0.95μg/ml at 24hrs and 7 days respectively, compared with the yields of 0.03 and 0.02μg/ml from plants grown in culture medium alone. One group of plants were treated only once with NAA during the initial 26 day culture period (labelled NAA(I)). In comparison with the group that was treated 3 times with NAA there was a reduction in MAb yield in the first 7 day culture, but in the final 24hr culture, the yields were very similar.

The average root dry weights from each of the groups of plants are shown in Figure 4B. Inclusion of NAA and IBA significantly increased dry root mass compared to cultures in which PGR had been omitted (p<0.05, ANOVA, Tukey's HSD). BAP and KIN, despite increasing rhizosecretion of antibody, did not increase root mass compared to plants cultured in the absence of PGR (p>0.05, ANOVA, Tukey's HSD).

Also shown in Figure 4B is the rhizosecretion rate calculated for each of the culture groups and expressed as μg MAb/g.root dry weight/24 hr. In all cases, the rhizosecretion rate was higher in the presence of a plant growth regulator, and highest with NAA (58.0 μg/g.root dry weight/24 hr). As only some of the growth regulators affected total root mass, the results suggest that the increased MAb rhizosecretion may be due to a mechanism that involves higher secretion rates, as well as an overall increase in root mass. Example 3 — increase of rhizosecretion of Cyanovirin-N and 4E10 antibody in tobacco using NAA

In order to demonstrate the generic nature of the enhanced recombinant protein rhizosecretion that is mediated by NAA, two further transgenic tobacco lines were examined. Note that the data for Cyanovirin-N is derived from the experiments of Example 1.

Cyanovirin-N transgenic plants were prepared and cultivated as described for Guy's 13 MAb. Yield analysis was performed by a functional HIV gpl20 binding ELISA, using an E. coli recombinant CV-N as a standard. The baseline rhizosecretion yield (without growth regulator) for CV-N plants is higher (1.26μg/ml) at 7 days than for Guy's 13 MAb (Figure 5A). However, this could still be boosted to 6.25μg/ml in the group receiving three NAA supplements in the first 26 day cultivation period, and 4.48μg/ml in the group receiving only one.

The effect of the NAA supplementation on root dry weight was very similar to that observed with the Guy's 13 MAb plants, but the overall rhizosecretion rate for CV-N was much higher (766μg/g.root dry weight/24 hrs) in the best plant cultures.

The effect of NAA on the rhizosecretion of an HIV neutralising human IgG monoclonal antibody (4E10) was also evaluated. With this transgenic line, the MAb 4E10 was undetectable in hydroponic medium samples from transgenic plants cultured without NAA at all time points, hi the presence of NAA administered three times, the MAb was readily detectable by ELISA, with yields of 200ng/ml after 7 days culture, and 30ng/ml after 24hrs culture. The rhizosecretion rate was calculated as 10.43μg/g.root dry weight/24 h. Example 4 - rhizosecreted IgG is predominantly intact and exposed to less proteolytic enzymes than IgG extracted from transgenic leaves

We have previously shown that the majority of rhizosecreted IgG is intact full- length antibody, which is only slowly subjected to proteolytic degradation over time (Drake et al., 2003). This was confirmed here and a comparison between IgG in hydroponic medium and IgG extracted from leaf tissue intercellular fluid is shown in Figure 6A. In this western blot, labelled with anti-kappa light chain antiserum, 6 major protein bands are evident in the leaf sample (lane 4). The intact IgG (approx. M _r 160K) is arrowed and there are 5 other Ig products at ~ M _r 150, 110, 75, 45 and 25K. In lanes 1-3, which represent different hydroponic medium samples, intact IgG (M _r 160K) is the predominant band. There are identical bands at M _r 150, 110, 75 and 45, but the 25K band is absent. Densitometric analysis of this western (supported by two replicate western blots) demonstrate the intact IgG band in the leaf sample to represent 12% of the total immunoreactive bands, compared with 49% in the hydroponic fluid.

Figure 6B shows the western blot and immunodetection of CV-N in medium samples from CV-N transgenic plants, as previously described in Example 1. Immunoreactive bands of the expected size (approx.11 kDa) were observed in samples from all hydroponic cultures (arrowed in lanes 1,2,3,7,8,9) which corresponded to the positive control (recombinant CV-N, lane 5). No degradation bands were evident. No immunoreactive bands were observed in samples from WT plants (lanes 4 and 10).

Also evident in this experiment, is the increase in CV-N yield produced by the inclusion of NAA in the hydroponic medium (lanes 1 and 7 - no NAA; lanes 2 and 8 - single addition of NAA; lanes 3 and 9 three additions of NAA).

Intact 4E10 MAb was evident in hydroponic samples only when NAA was added (arrowed in Figure 6C, lanes 1 and 3). Without NAA, the yield of 4E10 was below the limit of detection (lanes 2 and 4). This MAb demonstrated greater degradation in hydroponic culture, compared with MAb Guy's 13, with major degradation products at approx. M _r 105K and 75K.

The reduction in exposure to proteolytic degradative activity observed in Guy's 13 samples, could be an advantage in hydroponic systems. To compare the activity of peptidases in hydroponic medium and intercellular fluid taken from leaves, an analysis was performed of samples by 8% SDS-PAGE containing 0.1% gelatin, followed by an in-gel assay (zymogram) at pH 5.8. As shown in Figure 7A, a broad zone of digested gelatin was observed between 170 and 100 kDa in the intercellular fluid from wild-type (lanes 1-2-9-10), Guy' s 13 plants (lanes 3 -4) and CV-N plants (lanes 11-12). In contrast, in the hydroponic medium from Guy's 13 plants, only a limited zone of protease activity can be detected around 100 kDa, while a discrete band of protease activity is detected at 70 kDa (lanes 5- 8). In these experiments, the wild-type and transgenic plant intercellular samples were matched by volume, but it is difficult to standardise these with the hydroponic samples. In order to overcome this, we compared the content of IgG in the hydroponic or intercellular fluid samples from transgenic plants, and ensured that the hydroponic samples contained equivalent or higher levels of antibody than the intercellular fluid samples. Samples were also taken from CV- N plants (Figure 7B), in which gelatin was not added to the hydroponic medium. A similar peptidase activity profile was observed (lanes 13-16).

Example 5 - there is no significant difference in N-glycosylation heterogeneity between rhizosecreted IgG and IgG extracted from transgenic leaves.

A comparison of the glyco forms associated with Guy's 13 IgG extracted either from hydroponic fluid or transgenic leaves was performed, and their relative abundances are shown in Figure 8. Guy's 13 heavy chain is N-glycoslylated at two sites, Ν ⁹³ in the variable region and Ν ³⁰⁶ in the CH2 region (Cabanes- Macheteau et al., 1999). The N- glycosylation profiles from both sites are shown. The N- glycosylation site in the Fab region at position Ν ⁹³ (located to the tryptic peptide A ⁸⁷ TLTVDNS STS AYM ^# ELR ¹⁰³ ) exhibited exclusively complex-type plant iV-glycans in both samples. In both cases, MMXF was predominant. In the leaf-derived but not the hydroponic IgG, GnMXF and GnGnXF were also major glycoforms.

The iV-glycosylation site in the Fc region at position N ³⁰⁶ (located to the tryptic peptide E ³⁰² EQLNSTFR ³¹⁰ ) was found to bear predominantly oligomannosidic- type iV-glycans in both samples (60% in leaf-derived IgG and 86% in hydroponic- derived IgG). Subtle differences between N- glycosylation patterns of the two samples were detected at this N- glycosylation site: the sample derived from the hydroponic culture almost exclusively exhibited oligomannosidic-type JV-glycans, Man4-Man8 were the most abundant structures, and only trace amounts of complex-type plan iV-glycans were detected. In contrast, the sample purified from tobacco leaves showed a more heterogeneous iV-glycosylation profile, including Man5-Man8 oligomannosidic-type iV-glycans but also significant amounts of complex-type plant iV-glycans, GnGnF and GnGnXF.

Example 6 - Rhizosecreted Guv's 13 IgG in hydroponic medium is readily purified by affinity chromatography.

Rhizosecretion of recombinant proteins into hydroponic medium should allow relatively simple processing before affinity purification. Here, 7d after its addition to the plants, the hydroponic medium was removed stored at -2O ⁰ C, clarified by centrifugation, warmed to 24 ⁰ C, filtered and then passed directly onto an affinity column. The results for Guy's 13 IgG purification are shown in Figure 9. 200 column volumes of hydroponic medium (lane 1) were passed through an affinity column. There was virtually complete depletion of IgG by the affinity resin, as evidenced by the absence of an IgG band in the flow-through fraction (lane 2) detected either by protein silver stain or specific IgG immunodetection. The fraction eluted from the affinity column (lane 3) was highly purified IgG, with no evidence of any significant degradation. At least 90% of the recombinant IgG was recovered. The functionality of the purified IgG was confirmed by an antigen specific ELISA (data not shown). Examples 2 to 6 - Discussion

Conclusions

-the maximum rhizosecretion rates achieved were 58 μg/g. root dry weight/24 h for Guys 13, 10.43 μg/g.root dry weight/24 h for 4E10 and 766 μg/g.root dry weight/24 h for cyanovirin-N, the highest figures so far reported for a full-length antibody and a recombinant protein respectively.

-the plant growth regulators indole-butyric acid, 6-benzylaminopurine and kinetin were also demonstrated to increase rhizosecretion of Guy's 13. The effect of the growth regulators differed, as α-naphthalene acetic acid and indole-butyric acid increased the root dry weight of hydroponic plants, whereas the cytokinins benzylaminopurine and kinetin increased rhizosecretion without affecting root mass, -a comparative glycosylation analysis between MAb Guy's 13 purified from either hydroponic culture medium or from leaf extracts demonstrated a similar number of glycoforms in each sample, ranging from high mannose to complex glycoforms.

-analysis of the hydroponic culture medium at harvest revealed significantly lower and less complex levels of proteolytic enzymes, in comparison with leaf extracts, which translated to a higher proportion of intact Guy' s 13 IgG in relation to other IgG products.

-hydroponic medium could be added directly to a chromatography column for affinity purification allowing simple and rapid production of high purity Guy's 13 antibody

Producing recombinant pharmaceuticals by transgenic plants is an attractive prospect for many well documented reasons (Ma et al., 2005). However, concerns over product quality and uniformity have led to doubts and delays in commercial development of the technology. Moreover, a general move away from the use of food plants has left crops such as tobacco as the main candidates for commercial production (Sparrow et al., 2007), even though there are many challenges to address in terms of processing large quantities of wet leaf tissue. Production of plant-derived recombinant pharmaceuticals in containment helps to allay regulatory fears over gene flow and may not affect production costs significantly (http://www.biodesign.asu.edu/centers/idv/projects/provacs). In addition, the application of hydroponic cultivation techniques, already well established in the horticultural industry, could provide many potential advantages, as well as allowing the use of rhizosecretion as a production tool. Overall, we speculate that the following benefits could accrue:

1. Contained cultivation allows control over environmental factors (eg light, day length, temperature and CO ₂ ) which affect plant growth and recombinant protein yield and quality. It also facilitates plant disease management and reduces the risk of contamination from external sources, as well as the risk of accidental release of transgenic material into the environment. 2. The avoidance of soil and compost and the use of defined hydroponic medium provides even greater control over the plant cultivation process. 3. Rhizosecretion of recombinant proteins into the hydroponic medium allows continuous collection of product over the lifetime of the plant, and not just at the point of harvest. 4. Rhizosecreted recombinant proteins represent fully processed and secreted proteins and are likely to be more consistent than the mix of recombinant protein extracted from processed leaf tissue.

5. Avoidance of plant tissue processing and direct purification from hydroponic medium is likely to minimise the release of proteolytic enzymes and reduce recombinant protein degradation. It would also simplify the downstream processing procedure by eliminating a costly extraction step.

Rhizosecretion of recombinant proteins from transgenic plants was first proposed in 1999, but further advance was hampered by the low yields achieved. In the present study, the enhancing effect of plant growth regulators on rhizosecretion of three recombinant proteins from transgenic hydroponic plant cultures was demonstrated. The study design permitted measurement of recombinant protein yield at 24 h and 7 d culture in relation to volume of medium, and also allowed an accurate assessment of rhizosecretion rate in μg/g.root dry weight/24 h. The former measurement is of practical value as yields can be readily calculated, whereas the latter has been most commonly employed previously (Borisjuk et al., 1999, Drake et al., 2003) and consequently the calculation of this figure allows formal comparisons to be made between our results and past studies.

In a preliminary study on plants expressing Guy's 13, four plant growth regulators were studied, and α-naphthalene acetic acid (NAA) provided the greatest enhancement in yield. Repeated supplements of NAA appeared to be optimal.

Application of NAA also enhanced rhizosecretion of 4E10 and cyanovirin-N. Cell wall pore size has previously been thought to exclude the rhizosecretion of proteins larger than 30-50 kDa (Carpita et al., 1979), so the secretion of CV-N (11 kDa) might be expected, but the passage of a monoclonal antibody (approx.150 kDa) through the cell wall is unexpected, and may be associated with the highly flexible nature of antibody molecules (Drake et al., 2003). The studies on the impact of different PGRs on rhizosecretion of Guy' s 13 lends support to this view. Although addition of NAA led to an increase in root mass this could not entirely explain the increase in recombinant protein yield. For example, in both Guy's 13 and CV-N plants an approximate two-fold increase in root mass was observed following a single application of NAA, but after the 24h culture period a 7-fold increase in rhizosecretion of Guy's 13 and a 3-fold increase in CV-N was observed. After 7d these differentials rose to 20 fold for Guy's 13 and 3.5 fold for CV-N. Moreover, whereas NAA and IBA both caused an increase in root mass in Guy's 13 plants, the cytokinins BAP and KIN also increased rhizosecretion without affecting root mass. Auxins stimulate lateral and adventitious root formation. They promote cell elongation initially by stimulating the transport of hydrogen ions into the cell wall. The resulting decrease in pH activates the cell wall proteins expansins which in turn alter the hydrogen bonding between polysaccharides increasing cell wall plasticity. Thus we suggest that rhizosecretion of large proteins such as monoclonal antibodies is enhanced by NAA, both as a result of increased root mass and due to increased permeabilisation of the cell wall. The response in Guy's 13 plants exposed to NAA is greater than in CV-N and this may be due to the change in the structure of the cell wall allowing more of the large antibody molecule to pass through the wall, an effect that maybe less important for the smaller CV-N protein.

It is interesting to note that the relative yield of Guy's 13 obtained after 7d and 24h differs between plants given either 1 or 3 applications of NAA and that the 24h figure is almost identical for both of these treatments (Figure 4A). If there was a positive correlation between amount of NAA added and rhizosecretion rate, and that secretion of Guy's 13 after addition of fresh medium occurred in a linear fashion, we would expect the relative proportions of Guy's 13 at 7d and 24h to be the same for both treatments and for the yield of Guy's 13 in the 3 NAA application treatment to be greater than the 1 application after 24h. A priori, a possible hypothesis therefore is that rhizosecretion does not occur in a linear fashion over the 7d after addition of fresh medium, but rather proceeds exponentially after an initial lag phase, with the NAA increasing rhizosecretion rate in the former. Alternatively, rhizosecretion may proceed in approximately a linear fashion for the first 24 h in both treatments, but then decline more rapidly in the 1 NAA application treatment thereafter. Clearly, further experiments involving collections of medium at time-points between 24h and 7d would be required to determine rhizosecretion over the period. These rhizosecretion kinetics may be associated with more than one mechanism of action of NAA, although root mass appears not to be important as this is virtually identical in plants given either 1 or 3 applications of NAA (Figure 4B). The effect of secreted proteases on the accumulation of Guy's 13 in the medium may also be an important factor. This pattern of rhizosecretion was not observed with CV-N (Figure 5A), and this may reflect differences in the kinetics of rhizosecretion, or susceptibility to proteases. The integrity of the rhizosecreted recombinant proteins was confirmed by western blotting of hydroponic culture medium. In blots run under non-reducing conditions, there was a similar pattern of degradation of Guy's 13 in hydroponic fluid to that found in leaf intercellular fluid, except for two major differences. Firstly, there was an extra degradation band in the leaf sample which was entirely absent from the hydroponic sample; and secondly, whereas intact IgG represented a small percentage of the total murine IgG reactivity in the leaf sample, it constituted almost half of the immunoreactivity in the hydroponic sample. There was no evidence for CV-N degradation in hydroponic samples taken at 7 days. Indeed, CV-N has been reported to be remarkably resistant to proteolytic degradation (Boyd et al., 1997), and this is probably reflected in the higher overall yields achieved for this protein. The results suggested that the potential for proteolytic degradation is reduced for rhizosecreted IgG, which may be due to greater MAb stability in hydroponic medium, as a result of minimal plant tissue wounding during extraction and/or the dilution factor in the hydroponic fluid. This is supported by the observation that the protease activity and content in hydroponic medium is less complex, as compared with extracts from whole leaves. Indeed, our data suggests that a single protease may be dominant in hydroponic medium, and this alone gives cause for optimism in the future design of anti-proteolysis strategies.

When recombinant antibodies are extracted from leaf tissue, a mixture of glycoforms have been identified (Cabanes-Macheteau et al., 1999), which was thought to be due to the release of extracellular secreted antibody as well as intracellular antibody that was still in transit. In this study, we were surprised to identify a similar heterogeneity of glycoforms associated with rhizosecreted Guy's 13 IgG, as we had expected complete processing of the majority of secreted antibody, and a predominance of complex type glycans. This suggests that the high mannose glycans associated with plant-derived Guy's 13 IgG might be the final processed form in some cases, a finding that merits further study. In the final part of this study, we have investigated the first stage of IgG purification from hydroponic fluid. The demonstration that the hydroponic fluid requires minimal processing prior to affinity chromatography is significant, as it paves the way for a continuous downstream purification process linked to recycling of the hydroponic medium. A simple first antibody extraction step would be a great advantage, in comparison to preparation of an aqueous extract from fresh transgenic leaf tissue, followed by antibody extraction.

For the best conditions identified here, a single transgenic hydroponic culture containing 40ml of medium could yield up to 10.24 μg of full length, functional Guy's 13 antibody and 128 μg of CV-N in 24 h. The rhizosecretion rate of 766 μg/g.root dry weight/24 h for CV-N represents, to our knowledge, the highest reported figures for rhizosecretion of a recombinant protein, whilst the corresponding figure for Guy's 13 (58 μg/g.root dry weight/24h) is the highest reported for a full-length antibody. In both cases, the yields have now been increased to a point where pilot scale production can be envisaged.

Example 7 - increase of rhizosecretion of Cvanovirin-N in tobacco using a number of plant hormones

Hydroponic tobacco plants were cultured for 26 days in the presence of a plant growth regulator (PGR) - α-naphthalene acetic acid (NAA), indole-3 -acetic acid (IAA), indole-butyric acid (IBA), 6-benzylaminopurine (BAP) or kinetin (KIN) - or in the absence of a PGR (control). The PGRs were added at lmg/L on days 0, 7 and 14. The medium was then replaced with fresh medium supplemented with the same PGR. After a further 24hr (Figure 10) and 7 days (Figure 11) samples were taken to measure Cyanovirin-N (CV-N) concentration (using spectrophotometric absorbance). A series of two-fold dilutions were used and plotted against a standard concentration of CV-N of 150ng/ml._All of the PGRs improved CV-N yield. The least effective was kinetin, and NAA had the greatest effect. Example 8 - increase of rhizosecretion of Cyanovirin-N in Nicotiana benthamiana using NAA

The present study was conducted on a Nicotiana benthamiana transgenic plant line expressing Cyanovirin-N (CV-N). These plants were cultured in hydroponic medium for 26 days in the presence of α-naphthalene acetic acid (NAA). NAA was added at lmg/L on days 0, 7 and 14. The medium was then replaced with fresh medium supplemented with NAA. After a further 24hr (Figure 12) and 7 days (Figure 13) samples were taken to measure Cyanovirin-N (CV-N) concentration (using an ELISA) via a series of two-fold dilutions. NAA improved CV-N yield in N. benthamiana. Note that the level of CV-N yield without NAA treatment after 7d from N. benthamiana was similar to that seen for N.tabacum (Figure 13).

Example 9 - increase of rhizosecretion of secondary metabolites in tobacco using

NAA

Hydroponic tobacco plants were cultured for 26 days in the presence (4 plants) or absence (3 plants) of α-naphthalene acetic acid (NAA). NAA was added at lmg/L on days 0, 7 and 14. The medium was then replaced with fresh medium (with or without NAA added at lmg/L). After a further 24hr samples were taken and pooled for each group before measuring tobacco alkaloid concentration using gas chromatography mass spectrometry (GC-MS) as described in Hakkinen et al., 2004. No alkaloids were detected in the pooled sample from plants grown without NAA. However, alkaloids were present following NAA treatment:

1 lμg/ml nicotine, 5μg/ml nornicotine, 0.3μg/ml anabasine and 4μg/ml anatabine.

The final average root mass of plants grown with NAA (0.13g; average of 0.176g, 0.1 Ig, 0.124g and 0.098g) was less than double that of the final average root mass of plants grown without NAA (0.08g; average of 0.0816g, 0.063g, and 0.08Ig). Example 10 - increase of rhizosecretion of Cyanovirin-N in tobacco using NAA in an ex-vitro hydroponic system

Transgenic CV-N tobacco plants were established in a single tray and cultivated using Nutrient Film Technique (NFT), a hydroponic technique whereby a very shallow stream of water containing all the dissolved nutrients required for plant growth is recirculated past the bare roots of plants in a watertight gully (Figure 14). Usually the depth of the recirculating stream should be very shallow, little more than a film of water, hence the name 'nutrient film ¹ .

Samples of the circulating medium were taken before and after the addition of NAA and the concentration of CV-N measured. Figure 15 shows that the rhizosecretion of CV-N by transgenic tobacco increases sharply and quickly after the addition of NAA.

Example 11 - synergistic effect of cell- wall degrading agents on NAA-induced rhizosecretion of CV-N in tobacco

Seeds from the transgenic tobacco plant line expressing CV-N were surface- sterilized by immersion in 20% v/v bleach (Domestos, Lever Brothers, UK), washed in sterile distilled water and sown onto MS medium solidified with 0.7% w/v agar. Seedlings were allowed to reach a length of ca. 1 cm and were then transferred to liquid culture under sterile conditions so that the root was immersed in liquid medium and the shoot was above the platform. Plants were maintained at 25 ⁰ C with a 16h photoperiod for 2 months. Old medium was removed and the roots washed with 30 mL of liquid medium and then replaced by 30 mL of new medium. At this stage plants were divided in two groups. One group had the medium supplemented with 1 mg/L α-naphthalene acetic acid (NAA; Sigma, Poole, UK) whereas the other was left without NAA. After 12 days, a sample of each plant medium was taken (pre samples) and the two groups were divided in three further groups each: group a) remained untreated, group b) medium was supplemented with 0.16% (w/v) cellulase R-10 (Apollo Scientific) and 0.02% (w/v) pectolyase (Sigma) and group c) medium was supplemented with ImM L- ascorbate, ImM hydrogen peroxide and 20μM CuSO ₄ . Plants were left for two days and a sample of the medium was collected and CV-N concentration was assessed by Enzyme-linked immunosorbent assay (ELISA) (Figure 16). Both a combination of cellulase and pectolyase and a combination of L-ascorbate, hydrogen peroxide and CuSO ₄ showed a marked synergistic effect with NAA on the rhizosecretion of CV-N. I n contrast, such combinations showed little or no effect on the rhizosecretion of CV-N in the absence of NAA treatment.

REFERENCES

Boehm,R. (2007) Bioproduction of therapeutic proteins in the 21st century and the role of plants and plant cells as production platforms. Biology of Emerging Viruses: Sars, Avian and Human Influenza, Metapneumovirus, Nipah, West Nile, and Ross River Virus 1102: 121-134.

Borisjuk,N.V., BorisjukjL.G., Logendra,S., Petersen,F., Gleba,Y., and Raskin,! (1999) Production of recombinant proteins in plant root exudates. Nature Biotechnology 17: 466-469.

Boyd,M.R., Gustafson,K.R., McMahon,J.B., Shoemaker,R.H., Okeefe,B.R., Mori,T. et al. (1997) Discovery of cyanovirin-N, a novel human immunodeficiency virus-inactivating protein that binds viral surface envelope glycoprotein gpl20: Potential applications to microbicide development. Antimicrobial Agents and Chemotherapy 41 : 1521-1530.

Cabanes-Macheteau,M., Fitchette-Laine,A.C, Loutelier-Bourhis,C, Lange,C, Vine,N.D., Ma,J.K.C. et al. (1999) N-glycosylation of a mouse IgG expressed in transgenic tobacco plants. Glycobiology 9: 365-372.

Carpita,N.C, Montezinos,D., Sabularse,D., and Delmer,D.P. (1979) Determination of the Pore-Size of Cell-Walls of Living Plant-Cells. Plant Physiology 63: 52.

Delannoy, M., Alves, G., Vertommen, D., Ma, J., Boutry, M., and Navarre, C. (2008) Identification of peptidases in Nicotiana tabacum leaf intercellular fluid. Proteomics 8, 2285-2298.

Drake,P.M.W., Chargelegue,D.M., Vine,N.D., van Dolleweerd,C.J., Obregon,P., and Ma,J.K.C. (2003) Rhizosecretion of a monoclonal antibody protein complex from transgenic tobacco roots. Plant Molecular Biology 52: 233-241. Firek,S., Draper,!, Owen,M.R.L., Gandecha,A., Cockburn,B., and Whitelam,G.C. (1994) Secretion of A Functional Single-Chain Fv Protein in Transgenic Tobacco Plants and Cell-Suspension Cultures (VoI 23, Pg 861, 1993). Plant Molecular Biology 24: 833.

Fry SC (1998) Oxidative scission of plant cell wall polysaccharides by ascorbate- induced hydroxyl radicals. Biochem. J. 332: 507-515,

Gaume,A., Komarnytsky,S., Borisjuk,N., and Raskin,! (2003) Rhizosecretion of recombinant proteins from plant hairy roots. Plant Cell Reports 21 : 1188-1193.

Gleba,D., Borisjuk,N.V., Borisjuk,L.G., Kneer,R., Poulev,A., Sarzhinskaya,M. et al. (1999) Use of plant roots for phytoremediation and molecular farming. Proceedings of the National Academy of Sciences of the United States of America 96: 5973-5977.

Hakkinen, S. T., Rischer, H., Laakso, L, Maaheimo, H., Seppanen-Laasko, T. and Oksman-Caldentey, K. -M. (2004) Anatalline and Other Methyl Jasmonate- Inducible Nicotine Alkaloids from Nicotiana tabacum cv. BY-2 Cell Cultures

Halliwell B, and Gutteridge JMC (1999) Free Radicals in Biology and Medicine, Ed 3. Oxford University Press, Oxford.

Horsch, R. B., Fry, J. E., Hoffmann, N. L., Eichholtz, D., Rogers, S.

G., Fraley, R. T. (1985) A Simple and General-Method for Transferring Genes Into Plants. Science 227, 1229-1231.

Kneer,R., Poulev,A.A., Olesinski,A., and Raskin,! (1999) Characterization of the elicitor-induced biosynthesis and secretion of genistein from roots of Lupinus luteus L. Journal of Experimental Botany 50: 1553-1559. Komarnytsky,S., Borisjuk,N., Yakoby,N., Garvey,A., and Raskin,I. (2006) Cosecretion of protease inhibitor stabilizes antibodies produced by plant roots. Plant Physiology 141: 1185-1193.

Ma, J.K.C., Lehner,T., Stabila,P., Pux,C.L, and Hiatt,A. (1994) Assembly of Monoclonal-Antibodies with Iggl and Iga Heavy-Chain Domains in Transgenic Tobacco Plants. European Journal of Immunology 24: 131-138.

Ma, J. K. C, Barros, E., Bock, R., Christou, P., Dale, P. J., Dix, P. J., Fischer, R., Irwin, J., Mahoney, R., Pezzotti, M., Schillberg, S., Sparrow, P., Stoger, E., and

Magnuson,N.S., Linzmaier,P.M., Gao,J.W., Reeves,R., An,G.H., and Lee,J.M.

(1996) Enhanced recovery of a secreted mammalian protein from suspension culture of genetically modified tobacco cells. Protein Expression and Purification

7: 220-228.

Murashige,T. and Skoog,F. A revised medium for rapid growth bioassays with tobacco tissue cultures. Plant Physiology 15, 473-497. 1962.

Sexton,A., Drake,P.M., Mahmood,N., Harman,SJ., Shattock,R.J., and Ma,J.K.C. (2005) Transgenic plant production of Cyanovirin-N, an HIV microbicide. Faseb Journal 19.

Sharp _j J.M. and Doran,P.M. (2001) Characterization of monoclonal antibody fragments produced by plant cells. Biotechnology and Bioengineering 73: 338- 346.

Sparrow, P. A. C, Irwin, J. A., Dale, P. J., Twyman, R. M., and Ma, J. K. C. (2007) Pharma-Planta: Road testing the developing regulatory guidelines for plant-made pharmaceuticals. Transgenic Res. 16, 147-161. Stadlmann, J., Pabst, M., Kolarich, D., Kunert, R., and Altmann, F. (2008) Analysis of immunoglobulin glycosylation by LC-ESI-MS of glycopeptides and oligosaccharides. Proteomics 8, 2858-2871.

Taiz,L.Z.E. (1998) Plant Physiology. Sunderland: Sinauer Associates, Inc.

WongsamuthjR. and Doran,P.M. (1997) Production of monoclonal antibodies by tobacco hairy roots. Biotechnology and Bioengineering 54: 401-415.

Twyman, R. M. (2005) Molecular farming for new drugs and vaccines - Current perspectives on the production of pharmaceuticals in transgenic plants. EMBO Rep. 6, 593-599.

van Dolleweerd, C. J., Chargelegue, D., and Ma, J. K. (2003) Characterization of the conformational epitope of Guy's 13, a monoclonal antibody that prevents Streptococcus mutatis colonization in humans. Infect.hnmun. 71 , 754-765.