Plant cell suspension cultures are emerging as a useful platform for the manufacture of recombinant proteins [25
]. Production lines can be generated and scaled up more quickly and with lower costs than systems based on whole plants [29
]. Plant cells are also advantageous because recombinant proteins can be secreted into and recovered from the culture medium. They allow pharmaceuticals to be produced under current good manufacturing practice in bioreactor-based processes. Efficient and reliable inducible expression systems can improve the versatility of plant cell cultures still further by restricting protein expression to the most productive growth phase and preventing exposure to proteases which reduce both the quantity and the quality of the product. Inducible promoters allow transgene expression to be delayed until the culture has reached a suitable biomass in the late exponential growth phase, allowing production to be uncoupled from plant cell growth [15
Tetracycline-specific derepressible expression [30
] is a widely-characterized inducible system in plants that has been successfully applied in tobacco BY-2 cells [17
]. We have developed an improved version using a tandem construct incorporating a constitutive visual marker gene that allows the rapid and straightforward selection of the best-performing transgenic lines.
The first improvement was achieved by transferring the expression cassette from a pBin vector to pTRA, resulting in a substantial increase in recombinant protein yields by transient expression in tobacco. This probably reflects the influence of the Tobacco etch virus
5′-UTR, which acts as a translational enhancer in plants [23
]. The pTRA vector also contains flanking scaffold attachment regions (SARs), which are known to enhance transcriptional activity and mitigate silencing effects by defining a stable chromatin domain [22
]. Although stable transgenic lines were not generated with the pBin vector, transient expression experiments usually provide a reliable indication of the outcome of stable transformation [33
], thus it is reasonable to anticipate superior performance from the pTRA vector in stable transgenic lines.
The second improvement was achieved by adding a CaMV 35
S enhancer upstream of the tetracycline-inducible Triple-Op promoter. The presence of the enhancer sequence significantly increases the activity of the native CaMV 35
S promoter [36
]. We anticipated similar benefits in the context of the Triple-Op promoter because it differs from the native promoter only in the presence of three tet
O sequences surrounding the TATA box, and their presence should not hamper the synergistic interaction between the enhancer and minimal promoter.
The third and most innovative improvement was the tandem configuration of the inducible transgene and a constitutive visual marker gene, which greatly simplified the procedure for selecting the best-performing clones. Transgene expression is strongly influenced by the chromatin surrounding the integration site so tightly linked transgenes are generally subjected to the same global regulatory factors. This means that a gene encoding a fluorescent protein can be used as a surrogate marker to determine the activity of a linked transgene whose product can only be detected using as less convenient off-line assay. In this case, we found that the intensity of DsRed fluorescence was proportional to the amount of DsRed protein, which in turn was indicative of the amount of vIL-10. Although we observed a good correlation between the marker and the primary gene product, many factors could uncouple the yields of adjacent transgenes such as multiple or partial T-DNA insertions and transgene rearrangements. The DsRed gene was placed adjacent to the right-hand T-DNA border in our vector, and since the right-hand border is transferred first this configuration is more likely to produce false positives (strong fluorescence without corresponding vIL-10 accumulation) than false negatives (strong vIL-10 expression without fluorescence). Even so, the advantage of initial screening based solely on DsRed fluorescence is that second-stage screening following induction of the linked transgene is limited to a handful of promising clones, rather than tens or hundreds [37
]. Should it be so desired, it is also still possible within our platform to analyze antibiotic resistant callus clones that lack DsRed fluorescence to identify potential false negative lines that produce high levels of vIL-10. We used a precise fluorometric assay to determine the levels of DsRed fluorescence in order to carry out a quantitative comparison between DsRed and vIL-10 expression and demonstrate the predictive accuracy of our platform. However, for routine deployment such a precise assay is unnecessary. The human eye cannot distinguish small variations in fluorescence but is sufficient to identify the brightest callus clones on a plate using a green fluorescent light and a red filter, allowing the rapid visual selection of promising clones without sophisticated apparatus.
Finally, the constitutively expressed visual marker can also be used as an internal control to monitor the genetic stability of selected transgenic lines during prolonged cultivation with multiple subcultures. Any loss of fluorescence would be a strong indicator of transgene disruption/loss by recombination or the onset of epigenetic gene silencing.
The combination of improvements discussed above resulted in a highly reproducible increase in the accumulation of vIL-10, 3.5 times higher than those obtained by constitutive expression in the same cell line. The inducible expression cassette was highly efficient, benefiting from a low basal expression level and a high induction ratio. We did not replenish Ahtc in the culture medium so vIL-10 accumulation peaked at 4 dpi in well-established suspension cultures and decreased thereafter (Figure
). Further increases in yield might be achieved by optimizing the concentration and dosing method of the inducer and the timing of induction.
As well as its value in the field of molecular farming, the inducible expression cassette could also be useful for basic research by allowing the precisely-regulated overexpression or silencing of endogenous genes [15
]. The accurate measurement of reporter protein fluorescence would allow the selection of cell lines with different levels of transgene expression, allowing quantitative effects to be monitored rather than the binary on/off choice that conventional assays allow. In that regard, DsRed has been used as reporter gene in BY-2 cells [38
], whole plants [20
], fungi [40
] and animals [41
] without any reports of adverse effects on the host.