Strategies To Improve The Expression Level Of Human Granulocyte-Macrophage Colony Stimulating Factor In Plants

When genetically engineering cells to produce pharmaceuticals, it is crucial to improve the expression of foreign genes. There is a key limit factor in the application of plant bioreactor, that is, the accumulation level of recombinant protein is so low that the cost of its production is very expensive, and therefore, no market place. In generally, the content of recombinant protein should be over 1% of total soluble protein (TSP), and it might become industrial merchandise. At present, the human granulocyte-macrophage colony stimulating factor (hGM-CSF) has been expressed in microbes, animals and plants, but it is impossible to produce recombinant hGM-CSF in large scale using plants for its low expression level.Human granulocyte-macrophage colony stimulating factor is a glycosylated protein which can stimulate proliferation, maturation and differentiation of neutrophils and macrophages. It is a hematopoietic cytokine used in the treatment of neutropenia caused by bone marrow transplantation or cancer chemotherapy by accelerating the response of immune function. In the human genome, the GM-CSF gene is approximately 2.5 kb in length, containing four exons and three introns. The hGM-CSF encodes a protein of 144 amino acids including a 17-AA signal peptide. Its mature protein has 127 AA with a predicted molecular mass of 16,293 daltons.In this dissertation, a seed-specific promoter,β-conglycininα' subunit promoter, was artificially modified to enhance the gene expression level. The 5'-untranslated region of tobacco mosaic virus, Omega sequences, was inserted in downstream ofβ-conglycininα' promoter, and the translation start site was also modified to fit for Kozak motif of plants.The hGM-CSF was further artificially modified as follows: 1) Reconstruction of hGM-CSF signal peptide. Itself signal peptide of hGM-CSF was substituted by phytohemagglutinin (PHA-E) signal peptide of common bean (phaseolus vulgaris) that biased for plant codon usage, which would increase the translation level of genes. 2) Modification of hGM-CSF terminal. The terminal of hGM-CSF was conjugated by endoplasmic reticulum (ER) retention signal (KDEL) or storage vacuolar sorting determinants (AFVY) respectively. KDEL can target the recombinant protein to ER-derived bodies, which avoids protein degradation, and therefore, the protein accumulation level increased. The vacuolar sorting determinants of common bean 7S globulin (phaseolin) can sort the recombinant protein to storage vacuoles, which elevated the protein accumulation in the seeds. 3) Reconstructed 3'-untranslated region (3'-UTR) of hGM-CSF. The domainⅣnucleotides of human 7SL RNA were inserted into 3'-untranslated region of hGM-CSF. DomainⅣnucleotides, in which the highly conserved secondary structure consists of two internal loops, might be a natural aptamer that could sense the subunit of signal recognition particle (SRP), SRP54, in vivo.The modifiedβ-conglycininα' subunit promoter and hGM-CSFs were used to construct plant expression vector pCAMBIA2300-CSF1, pCAMBIA2300-CSF-KDEL, pCAMBIA2300-CSF-AFVY and pCAMBIA2300-CSF-AP, which were tranformed into Arabidopsis thaliana.In the dissertation, the anti-sense technology was also employed to improve the accumulation of recombinant protein. The binary vectors of pCAMBIA2300-CSF-KDEL-anti2S and pCAMBIA2300-CSF-AFVY-anti2S, containing anti-at2S2 gene expression vector which was constructed with itself promoter and poly(A) signal, were transformed into Arabidopsis thaliana.The expression vectors including pCAMBIA2300-CSF1, pCAMBIA2300-CSF-KDEL, pCAMBIA2300-CSF-AFVY, pCAMBIA2300-CSF-AP, pCAMBIA2300-CSF-KDEL-anti2S and pCAMBIA2300-CSF-AFVY-anti2S were used to transform Arabidopsis thaliana, and transformants were selected on the medium added with kanamycin. In total, 50 C (from pCAMBIA2300-CSF1), 32 KD (from pCAMBIA2300-CSF-KDEL), 23 AF (from pCAMBIA2300-CSF-AFVY), 6 AP (from pCAMBIA2300-CSF-AP), 7 KDS (from pCAMBIA2300-CSF-KDEL-anti2S) and 9 AFS (from pCAMBIA2300-CSF-AFVY-anti2S) transgenic lines were acquired respectively. The part lines of AP, KDS and AFS transformants showed sterility, however, we still found that some transformed plants of AP, KDS and AFS inherited stably.PCR amplification was first used to verify transformants. Southern blotting was further employed to display that the hGM-CSF gene was integrated into the genome, although gene copies varied in AP, KDS and AFS transformed lines. The segregation of T2 generation was first calculated to define one-copy transformants in KD and AF transgenic lines, and further confirmed by Southern blots. Single-copy independent transgenic lines displaying in southern blotting results were chosen for hybrid formation. In the KD×KD and AF×KD crossed matches, all pollen donors were obtained from one-copy KD6 transgenic lines.Total RNA isolated from roots, stems, leaves and seeds of KD6 transformants respectively, with DNaseⅠdigestion, were used to confirm hGM-CSF expression. No PCR products were amplified from roots, stems or leaf tissues, but strong hGM-CSF mRNA expression was shown in seeds. We further investigated hGM-CSF mRNA levels in early and late stages of seed development using siliques at 3, 6, 9, 12 day after pollination (DAP) by semi-quantitative RT-PCR. These results suggested that the soybeanα' subunit of theβ-conglycinin promoter was activated in the early seed developmental stage, and still displayed high expression in late stages.ELISA data showed that the accumulation level of recombinant hGM-CSF was lower than 0.01% TSP in immature siliques of C transgenic lines, and heightened greatly in KD and AF transformed plants. There was no significant difference between KD and AF, which indicated that both KDEL and AFVY signal were equal to elevate the protein accumulation level. Hybrid seeds derived from KD×KD6 and AF×KD6 revealed gene additive effects, but AF×KD6 hybridizations displayed higher additive effect than that of KD×KD6. KD1 and AF18 transgenic lines displayed the same rhGM-CSF accumulation level, and hGM-CSF mRNA from the seeds of KD1, KD6, AF18, KD1×KD6, AF18×KD6 was analyzed by real-time quantitative SYBR-GreenⅠRT-PCR using gene-specific primers. The hGM-CSF transcript levels showed no difference between KD1×KD6 and AF18×KD6 hybrid matches, whereas there was highly significant difference in the accumulation level of recombinant protein, which indicated that AFVY and KDEL played an active role and independently functioned in recombinant protein sorting and transport pathways of the same transformed cells.The mean value of rhGM-CSF in AP independent transgenic lines showed far higher than that of the KD transformed lines, and there was highly significant difference between KD and AP. The quantitative PCR data displayed that the more recombinant hGM-CSF accumulation, the higher its mRNA expression level, which suggested that insertion of nucleotides of domainⅣfrom human 7SL RNA into the 3'-UTR of the hGM-CSF gene may improve mRNA stability, and therefore, enhance protein accumulation.The data of at2S2 mRNA quantitative PCR indicated that at2S2 gene was repressed in KDS and AFS transgenic lines. The results of protein SDS-PAGE displayed that the relative content of 2S, 12S-A and 12S-B was not markedly changed, which showed that at2S2 gene repression did not alter the ratio among main components of seed protein, however, the rhGM-CSF exibited significantly difference compared with KD and AF transgenic lines respectively. The repression of at2S2 gene can heighten recombinant protein accumulation, and there was not a linear correlation between the ratio of at2S2 repression and the expression level of hGM-CSF. We observed a band at 21 kD of rhGM-CSF produced by seeds in Western blots, which is similar to that recombinant hGM-CSF derived from transgenic tobacco seeds, was 5 kD bigger than that of natural human GM-CSF. This might represent a higher glycosylated form of recombinant hGM-CSF protein produced by plants. TF-1 cell proliferation data demonstrated that the recombinant hGM-CSF was biologically active.