Chromosome-free bacterial cells are safe and programmable platforms for synthetic biology

Significance We constructed simple cells (SimCells) whose native chromosomes were removed and replaced by synthetic genetic circuits. The chromosome-free SimCells can process designed DNA and express target genes for an extended period of time. The strategy of SimCell generation is applicable to most bacteria, creating a universal platform for reprogramming bacteria. We demonstrated that SimCells can be designed as safe agents for bacterial therapy through synthesis and delivery of a potent anticancer drug against a variety of cancer cell lines. This showed that the nonreplicating and programmable property of SimCells is advantageous for applications in sensitive environments. The results of this work will both improve our understanding of natural living systems and simultaneously lay the foundations for future advances in synthetic biology.


Construction and application of broad host-range SimCell-generating plasmids
To demonstrate SimCell formation in a wider array of organisms we constructed the plasmid pRH121 (Fig. S5a) via a HiFi Assembly approach, based on the pSEVA231 backbone kindly provided to us by Victor de Lorenzo 5 . Primers for the amplification of the various parts of pRH121 are displayed in Table S2. Tight transcriptional control of I-CeuI was deemed mandatory to avoid the cellular defences of various bacteria that may inactivate the SimCellforming machinery. The Jungle Express repressor system has been shown to provide exceptionally tight transcriptional control in several different species, including P. putida 8 . This tight level of control was also witnessed in R. eutropha (Fig. S6). Therefore, in pRH121 the EilR repressor and PJExD promoter from pJC580 (JBEI Part ID: JPUB_010723, Fig. S5b) governed the transcription of an I-CeuI gene that had been codon optimised for R. eutropha H16 (synthesized by Integrated DNA Technologies).
The final pRH121 plasmid was sequenced and transferred to both R. eutropha H16 and P. putida UWC1 by conjugation via the S17-1 E. coli donor. R. eutropha and P. putida cells containing pRH121 were then selected on LB plates containing 50 µg/mL kanamycin, and either 50 µg/mL gentamicin or 50 µg/mL rifampicin, respectively. I-CeuI expression was induced in both strains with 1 µM crystal violet.

Characterization of SimCell populations
Flow cytometry analysis was done with an S3e Cell Sorter (Bio-Rad). The FL1 filter was used to detect fluorescence from GFP, which has an excitation/emission at wavelengths 488 nm/507 nm. The FL3 filter was used to detect fluorescence from mCherry, which has an excitation/emission at 587/610 nm.

Imaging
Cells were visualized with a Nikon Ti Eclipse. To visually compare SimCell versus parent cell growth, cells were fixed in agar containing 1.2% Noble Agar, 10X diluted LB media, 100 nM anhydrotetracycline (ATc), 0.2% arabinose and 200 μM erythromycin with the thickness of a coverslip on a glass slide. The slide was incubated overnight at room temperature and visualized the following day. To take the videos demonstrating SimCell synthesis of mCherry, cells were stained with 100 µM DAPI and fixed in the same agar with the same components except LB. Visualisation of SimCell formation from R. eutropha and P. putida containing pRH121 was conducted as described previously, but used 1 µM crystal violet to induce I-CeuI expression. Frames were taken every 15 min for about 24 hr.

Longevity of SimCells
Purified SimCell cultures expressing luminescence (pJKR-OmphR-ilux) or an unstable mCherry variant (pJKR-OmphR-ASV) were kept in 50 mL Falcon tubes shaking at 100 rpm at 37°C. Every two days 25 µg/mL D-cycloserine was added to the cultures to maintain SimCell purity. At days 1, 3, 5, 10, 14, and 28, aliquots (200 µL, n=4) were taken and measured for luminescence or fluorescence production over 24 hours. The maximum reading during this period was recorded. OD600 was compared at t=0 and t=24 hours to see if there was an increase and therefore contamination by parent cells.

Sample preparation for proteomics and LC-MS/MS analysis:
The E. coli DH5α pellets of the wildtype, with pLO11 and with pLO11-ICeuI, were resolubilized in a solution containing 8 mol/l urea (Merck, Darmstadt, Germany) and 2 mol/l thiourea (Merck, Darmstadt, Germany). Solubilised cells were disrupted by 5 cycles of freezing in liquid nitrogen and subsequent incubation at 30°C for 10 min. Afterwards, medium intensity ultrasonic pulses were applied for 30 seconds. Cell fragments were removed by centrifugation at 20,000 g for 1h at room temperature. The resulting supernatant was collected. AG, Schlieren, Switzerland). The generated raw data were further analysed in a Renvironment (v 3.5.1). MS2-peak-areas were median normalized on replicate level and conditions were compared on peptide level using reproducibility-optimized peptide change averaging method (ROPECA) 9 . Candidates with adjusted P-values (Benjamini and Hochberg) < 0.05 and fold changes of +/-1.5 were considered as significantly changed. Mass spectrometry settings, mass windows for DIA analysis, and R packages used for analysis are listed in a related study 10 .
The heatmaps were generated using the seaborn Python data visualization library 11 . For the dendrograms, correlation was used as the distance metric with single-linkage clustering.
Protein abundances were row scaled to show the fold change.

Modelling unstable mCherry-ASV production and degradation
The time-dependent change in the mCherry-ASV concentration is modelled as: The rate of mCherry-ASV synthesis is , which also possesses a Michaelis-Menten dependence on the energy level, with , as the kinetic constant. The degradation of mCherry-ASV is achieved by the protease . Therefore the degradation process is considered to follow the Michaelis-Menten kinetics for enzymatic reactions. The (maximum) rate of mCherry-ASV degradation is . As proteolysis of mCherry-ASV is also energy dependent, substrate level ( ) and energy level ( ) are both incorporated into the kinetic terms, with and , being the corresponding kinetic constants, respectively.
The dynamic change in the concentration of protease is modelled as Where is the rate of protease degradation. New proteases are not synthesized during the considered process.
The change in glycolytic enzyme follows a similar form to Equation (1) shown in the main text Where is the rate of glycolytic enzyme synthesis, , is the kinetic constant, and is the rate of enzyme degradation.
The change in the overall energy can be calculated by the energy production by glycolysis, energy consumption for the expression of glycolytic enzymes, energy consumption for mCherry-ASV synthesis and that for mCherry-ASV hydrolysis.
Where is the rate of energy production via glycolysis. As the substrate is shown to be irrelevant in this system, is considered as a constant. The parameters , and denote the energy demands per unit rate of glycolytic enzyme synthesis, mCherry synthesis and mCherry degradation, respectively.
Finally, the wearing-out of cell machinery is modelled by the time-dependent decay of protein synthesis rate, applied to both glycolytic enzymes and mCherry-ASV: Where is the rate of wearing-out of cell machinery.

Equations (5)-(10)
were used to describe the potential mechanisms that govern the experimental observation of the mCherry-ASV production.
The metabolite separation was achieved using a ZORBAZ Eclipse Plus C18 packed with 95 Å pore, 5 µm particle size and 4.5 × 150 mm column (Agilent, US). Elution was performed using isocratic mixture of water, methanol and acetic acid (690:280:30) as previously described by Sawyer and Kumar at 0.5 ml min-1 for 10 min 12 . The oven temperature was 30°C. The injection volume was 5 µl. The UV detector was set to a wavelength of 275 nm for catechol detection. Data were collected at an acquisition rate of 5Hz. Control, experimental samples and catechol standards were run sequentially for comparison (n=3). For culture, 300 µL purified SimCells or parent cells in LB were induced or not induced for catechol production in 1.5 mL Eppendorf tubes. Tubes were placed in a shaking incubator overnight at 37°C.
Parent cells and SimCells were spun down at 10,000 x g and the supernatant was analyzed for catechol concentration. OD600 was recorded to calculate the number of cells/mL and subsequently the moles of catechol produced per cell.
Absorbance was measured at 595 nm.         induction of glycolysis will briefly yield a small increase in mCherry concentration, which was observed in the experiment. This is because a suitable concentration of glycolytic proteins will eventually be released but the transcription and translation machinery will have been worn out before the benefits of the glycolytic proteins will be realized.