document

CHINESE JOURNAL OF BIOTECHNOLOGY Volume 24, Issue 12, December 2008 Online English edition of the Chinese language journal

Cite this article as: Chin J Biotech, 2008, 24(12), 2130 2130.

Received: October 17, 2008; Accepted: November 25, 2008 Corresponding author: Jianfeng Xu. Tel: +10-870-680-4812; Fax: +10-870-972-2026; E-mail: [email protected]

Copyright © 2008, Institute of Microbiology, Chinese Academy of Sciences and Chinese Society for Microbiology. Published by Elsevier BV. All rights reserved.

RESEARCH PAPER

High-yield Expression of Therapeutic Proteins with Extended Serum Half-life in Tobacco Cells

Jianfeng Xu

Arkansans Biosciences Institute, Arkansans State University, State University, AR 72467, USA

Abstract: Small therapeutic proteins such as interferon 2b (IFN 2) and human growth hormone (hGH) generally possess short serum half-lives due to their susceptibility to serum proteases and small size, hence rapid renal clearance. Chemical derivatization overcomes these problems but at the expense of decreased bioactivity. We develop a new method that yields biologically IFN 2 and hGH in high yields and with increased serum half-lives when expressed as arabinogalactan-protein (AGP) chimeras in cultured tobacco cells. The chimeric glycoproteins typically gave 500 1800 fold greater secreted yields than their non-glycosylated controls. Importantly, the chimeric glycoproteins showed an increased in vivo serum half-life of up to 13 fold whereas their biological activities remain similar to native IFN 2 and hGH. The AGP domain was not immunogenic when injected into mice.

Keywords: plant cell culture; recombinant protein; hydroxyproline-O-glycosylation; half-life

Introduction

General approaches aimed at improving the clinical

effectiveness of small therapeutic proteins (<40 kD) involve increasing molecular size either by chemical derivatization, such as addition of PEG groups (Pegylation)[1] or expression as human serum albumin (HSA) protein chimeras[2]. However, these derivatives have significant drawbacks such as low yields of the pegylated products, extensive heterogeneity, and greatly decreased receptor binding. Therefore we explored an alternative approach that produces small therapeutic proteins as chimeric glycoproteins secreted in high yield and with enhanced stability. Our approach exploits earlier use of synthetic gene technology to design and express hydroxyproline (Hyp)-rich glycoproteins (HRGPs) where peptide sequence directs extensive Hyp O-glycosylation of proteins secreted to the cell surface[3]. Unique to plants, this post-translational modification involves an O-glycosylation code based on Hyp contiguity and yields two types of Hyp-glycomodule:

small glycomodules result from addition of short arabinooligosaccharides to contiguous Hyp residues, as typified by the Ser-Hyp4 pentapeptide motif of extensins. In contrast, much larger glycomodules result from addition of arabinogalactan polysaccharide to clustered non-contiguous Hyp residues, typically X-Hyp-X-Hyp repeats where X is most often Ser or Ala, and defines the hyperglycosylated arabinogalactan proteins (AGP)[3]. Here we report the expression of two small therapeutic proteins, IFN 2 and hGH as AGP chimeric glycoproteins, their facile secretion by tobacco cell suspension cultures and their pharmacokinetic properties.

1 Materials and methods

1.1 Plant transformation, cell culture and protein isolation

The genes encoding eight protein chimeras were constructed in the plant transformation vector pBI121 (Clontech, CA) (Fig. 1a), and then transformed into tobacco BY2 cells mediated by Agrobacterium tumefaciens strain

RETRACTED

Jianfeng Xu et al. / Chinese Journal of Biotechnology, 2008, 24(12): 2130–2130

LBA4404. Transformed tobacco cells were cultured in Shenck-Hildebrand (SH) medium. The concentration of IFN 2 and hGH in the culture medium or cytosol was determined by ELISA. Total soluble protein (TSP) was measured using a Coomassie Plus® protein assay kit (Pierce, Rockford, IL). The target proteins were isolated from culture medium with hydrophobic interaction chromatography (HIC) and reverse phased HPLC (RP-HPLC)[4]. 1.2 In vitro bioactivity and in vivo clearance of chimeric glycoproteins

Anti-viral activity of IFN 2 was assayed by PBL Biomedical Laboratories (Piscataway, NJ) using Madin-Darby Bovine Kidney (MDBK) cells challenged by Vesicular stomatitis virus (VSV)[2]. The receptor binding activity of hGH was assayed by challenging with [125I]hGH bound to hGH receptors expressed at mouse L cell surface. For in vivo clearance analysis, three female BALB/c mice (7 8 week-old, Harlan) were injected through the tail vein with a single dose of 100 g (IFN 2 equivalents)/kg of IFN 2 chimeric glycoproteins, or 20 g (hGH equivalent) of hGH chimeric glycoproteins. Blood samples were collected from the tail at a certain period of time and the protein bioactivity of the sera was determined.

2 Results and discussion We designed eight protein chimeras (Fig. 1a). Based on

the Hyp O-glycosylation codes described above, we predicted all the Pro residues in the Ser-Pro repeats would undergo hydroxylation and subsequent glycosylation with arabinogalactan (Fig. 1b). The chimeric proteins were designated IFN 2-(SO)n (n=2,10,20), hGH-(SO)10 and

hGH-(SO)10-EGFP, where O denotes Hyp. IFN 2 and hGH yields from transformed tobacco cells were presented as IFN 2 or hGH equivalents in the culture medium and cytosol. As shown in Tab., the yields of secreted IFN 2-(SO)n, hGH-(SO)10, and hGH-(SO)10-EGFP were 500–1800 fold greater than their non-glycosylated controls, with up to 95% of the total cellular IFN 2 and hGH equivalents located in the culture medium. In contrast, less than 2% of the expressed IFN 2, hGH or hGH-EGFP were secreted into medium. Secreted recombinant proteins with yields greatly enhanced by glycosylation are of practical significance to the biotechnology industry where the low yield of secreted transgenic proteins is a major obstacle to molecular pharming in plants.

Fig. 1 Design and characterization of the IFN 2 and hGH chimeric glycoproteins (a) Gene expression cassettes. SS: signal sequence; (SP)n: Ser-Pro repeats; CaMV35S: 35S cauliflower mosaic virus promoter; 3’NOS: nopaline synthase terminator. EGFP: enhanced green fluorescent protein. (b) Schematic illustration of a chimeric glycoprotein with ten Ser-Hyp repeats at the C-terminus of the therapeutic protein and a typical Hyp-arabinogalactan polysaccharide consisting of galactose, arabinose, rhamnose and glucuronic acid. (c) SDS-PAGE of IFN 2 (Lane 2), IFN 2-(SO)2 (Lane 3), IFN 2-(SO)10 (Lane 4), IFN 2-(SO)20 (Lane 5) and IFN 2 standard (Lane 6). (d) SDS-PAGE of hGH-(SO)10 (Lane 2), hGH-(SO)10-EGFP (Lane 3), hGH (Lane 4) and hGH standard (Lane 5)

RETRACTED


Tab. IFN 2 and hGH yields from tobacco BY2 cells

Constructs Protein in medium (mg/L) Protein in cytosol (mg/g FW) TSPP

(%) Secreted protein (%)

IFN 2 0.002–0.02 0.02–0.04 0.2–0.4 0.05–1.00

IFN 2-(SO)2 7–11 0.02–0.03 0.9–1.5 76–84

IFN 2-(SO)10 17–28 0.01–0.03 2.0–3.5 85–93

IFN 2-(SO)20 23–27 0.01–0.02 2.4–3.0 88–95

hGH 0.001–0.070 0.03–0.05 0.3–0.6 0.04–2.00

hGH-EGFP 0.00–0.02 0.03–0.04 0.3–0.4 0.0–0.4

hGH-(SO)10 16–35 0.02–0.03 2.2–4.0 86–91

hGH-(SO)10-EGFP 15–32 0.02–0.03 1.9–3.5 84–90

The yields were measured from 5–10 cell lines per construct after 9–10 days in culture. The cell biomass at the end of the culture was 95–110 g fresh weight (FW) per liter

Fig. 2 Fluorescence micrographs of BY2 cells expressing

hGH-EGFP and hGH-(SO)10-EGFP The cells were inspected for EGFP fluorescence using a Zeiss LSM 510 laser-scanning confocal microscope. (a) BY2 cells expressing hGH-EGFP. The fluorescence occurred at the cell wall/plasma membrane interface; (b) BY2 cells expressing hGH-EGFP after plasmolysis in 750 mM mannitol. Most of the fluorescence was retained intracellularly; (c) BY2 cells expressing hGH-(SO)10-EGFP. The fluorescence occurred mainly in the culture medium and at the wall/plasma membrane interface; (d) Plasmolysed BY2 cells expressing hGH-(SO)10-EGFP. The fluorescence is mainly in the periplasmic space between cell wall and plasma membrane, although some intracellular EGFP is apparent, probably in transit through the ER/Golgi before secretion to the wall. PM: plasma membrane; CW: cell wall.

The green fluorescence tag contained in hGH-EGFP and hGH-(SO)10-EGFP allowed us to visually compare the effects of glycans on the intra- and extra- cellular

distribution of the target proteins. As shown in Fig. 2, hGH-EGFP was largely retained intracellularly despite the presence of a secretion signal peptide. In contrast, cells expressing hGH-(SO)10-EGFP showed intense green fluorescence between the plasma membrane and wall, chimeric glycoprotein likely in transit to the medium, which suggests that the glycans greatly facilitated protein secretion in plan cells.

The molecular weights of the chimeric glycoproteins were significantly increased and migrated as broad bands on SDS-PAGE (Fig. 1c and 1d), probably resulting from micro-heterogeneity in the glycans and inefficient SDS binding by glycoproteins. The Hyp-glycans of these glycoproteins were comprised of four monosaccharide residues: high percentages of galactose and arabinose with lesser amounts of glucuronic acid and rhamnose[3,4]. The size of the Hyp-glycans ranged from 19 to 26 residues with a 22-residue glycan comprising the major species (Fig. 1b)[4].

It was interesting to see that the IFN 2 and hGH chimeric glycoproteins retained almost the full bioactivity (>85% of standards) (Fig. 3a and 3b). This was in contrast to chemically derivatization[1] or expression as a human serum albumin chimera[2] that usually greatly decreased specific bioactivities. The circulating half-life increases for IFN 2 and hGH chimeric glycoprotein were up to 13 folds (Fig. 3c and 3d), comparable to those brought about by Pegylation[1]. We further assayed the immunogenicity of the chimeric glycoproteins (data not shown) and found the purified AGP domain itself was not immunogenic when injected into mice and only mildly so when injected as a chimeric glycoprotein[4].

RETRACTED


Fig. 3 In vitro bioactivity and in vivo clearance analyses of IFN 2 and hGH chimeric glycoproteins (a) The dose response curves of IFN 2 chimeric glycoproteins. IC50: 50% cytopathic effect of the assay, the point where half the cells are dead. (b) Radioreceptor binding assays of hGH chimeric glycoproteins. IC50: binding constant. (c) Clearance of IFN 2 and IFN 2-(SO)10 & 20 from mouse plasma after intravenous injection. (d) Clearance of hGH and hGH-(SO)10 from mouse plasma after intravenous injection. The t1/2 value presented in the (c) and (d) panels is the circulating plasma half-life of each protein. Each value represents the mean of three replicates ± standard deviation.

3 Conclusions

We developed a new approach that allows the plant

cells to secrete therapeutic proteins in high yields. The AGP domain attached to the proteins has the attendant advantage of increasing the molecular size and extending the serum half-life. It also overcomes the problem of low product yields that plague the production of recombinant proteins by plant cells while secretion into the medium significantly simplifies the purification process. Thus, molecular pharming in plants may become economically feasible.

REFERENCES

[1] Harris JM, Chess RB. Effect of pegylation on pharmaceuticals. Nat Rev Drug Discov, 2003, 2: 214 221.

[2] Osborn BL, Sung C. Pharmacokinetic and pharmacodynamic studies of a human serum albumin-interferon- fusion protein in cynomolgus monkeys. J Pharmacol Exp Ther, 2002, 303: 540 548.

[3] Kieliszewski MJ, Shpak E. Synthetic genes for the elucidation of glycosylation codes for arabinogalactan-proteins. Cell Mol Life Sci, 2000, 58: 1386 1398.

[4] Xu J, Kieliszewski MJ. High-yields and extended serum half-life of human interferon 2b expressed in tobacco cells as arabinogalactan-protein fusions. Biotechnol Bioeng, 2007, 97: 997 1008.

RETRACTED

document

Documents