While NMR has proven to be an efficient technology for molecular imaging, it is also known for being expensive. This is related to the expensive labeled media required during the fermentation process. The study highlighted in this article shows that high cell density fermentation (HCDF) may provide an efficient alternative to traditional protein labeling methods.
2H-, 13C-, 15N-labeled outer-membrane protein G (DCN-OmpG) over-expressed in E. coli was needed for structural studies by solid-state (ss) magic-angle spinning (MAS) NMR. The protocol for production, purification, refolding, and reconstitution for NMR purposes and method development has been established at the FMP lab (www.fmp-berlin.de). Typically a 20-mg preparation starting from 2-L E. coli shaking culture grown to an OD600nm of 2 is used to produce one sample for ssNMR. This implies the consumption of 2-L D2O as well as the 15N and 13C isotopes for the preparation of one triple (DCN) labeled ssNMR-sample. To generate higher yields of labeled protein, Dasgip’s (www.dasgip.com) fermentation system fed batch pro® was adapted to produce recombinant protein for NMR analysis and other studies.
Dasgip bioreactor systems allow a fully automated HCDF process—growing the cells in the batch phase, starting the fed-batch after the substrates have depleted, switching on the labeled feed 1 h before protein expression is induced, maintaining the feed during protein expression, then cooling down for harvest. Each step can be controlled with one system and can be integrated into one process.
The OmpG gene was cloned into the pET-26b vector (kanamycin resistance). The cloning procedure and all traditional protein expression and purification protocols were published previously.
For HCDF, Dasgip’s bioreactor system was used with four parallel fermentation vessels, each contained 200-mL medium. A modified M9 minimal medium was used for the cultivation.
Enhancing Cell Density
The expression experiments for OmpG using HCDF were performed on nondeuterated, unlabeled media for optimization. The batch medium contained sufficient substrates (8 g/L glucose, 2 g/L NH4Cl) for growing the E. coli to OD600nm of about 8. Before induction, a biomass accumulating phase was introduced. First the effects of 0, 2, 4, and 6 h additional unlabeled feeds were studied. Six hours of additional cell growth from unlabeled substrates resulted in an OD600nm of about 13 (Figure 1).
The effects of enhancing cell density on the expression level were analyzed by SDS-PAGE. The specific protein production rate was kept constant over the time tested, and the generation of higher biomass leads to more recombinant protein yields.
The four deuterated HCDF set-ups for production of DCN-OmpG were programmed with a biomass accumulating feed that lasted 6 hours (Figure 2). After a prolonged lag phase the cells grew at an easy rate, finally attaining ODs comparable to the nondeuterated experiments. The automatic switch from batch to fed-batch was initiated by the rising dissolved oxygen (DO) once the substrates had depleted.
The signal triggered the start of DO-based feeding (deuterated medium with unlabeled N- and C-sources) for further biomass enrichment. The DO allowed adjustment to the requirements of the cells. After six hours the feed automatically switched to labeled feed. The feeding rate was kept constant for sufficient protein production. After one hour the IPTG as inducer for protein production was automatically added. The duration of expression was another six hours, after which the feeding was stopped by the process control. The reactors were cooled for later cell harvest.
The specific protein production rate was even better than that of the shaking culture, producing an 11-fold higher yield of protein from the same volume of culture.
Two different NMR-spectra are shown in Figure 3. Spectrum A reflects 15N labeling as 1H-15N correlations. The original deuterated nitrogens were protonated due to the purification procedure with water-containing buffers. Spectrum B shows missing 1H-13C-proton correlations, as deuterated carbon atoms keep the deuteron bound, even in water. The spots to be seen in spectrum B originated from the natural abundant 1H-13C-proton correlation of the nondeuterated detergent used for refolding.
These results show that the HCDF labeling protocol is a useful tool for 2H13C15N -labeled protein production under the requirements of NMR spectroscopy.
Due to the toxic effects of D2O, deuteration of proteins is a challenge. Data from the literature often underline a required adaptation of cells to deuterated conditions via stepwise enhancement of the D2O content in the medium. This is time consuming and expensive.
We inoculated 99% deuterated medium relatively thickly (starting at OD600nm 0.25–0.3) with cells collected from preculture grown on nondeuterated rich medium. That led to good and reproducible results for cell growth and protein production. Following this strategy the HCDF was initialized at OD600nm 0.5 growing with about one-half of the speed of the water-based preliminary experiments. Nevertheless the same yield of recombinant protein could be achieved.
The depletion phase and the head start of the labeled feed with respect to the induction is essential to ensure high labeling of the protein of interest. The NMR-spectra demonstrate the successful labeling of OmpG, producing a product suitable for NMR-supported structural biology, while the protocol shows the reduction of costs down by 25%.
The HCDF was used for a relatively difficult application of isotopic labeling of recombinant proteins. The system is applicable to all variants of unlabeled expression as well.