Light-mediated control of gene expression in the anoxygenic phototrophic bacterium Rhodobacter capsulatus using photocaged inducers

Photocaged inducer molecules, especially photocaged isopropyl-β-d-1-thiogalactopyranoside (cIPTG), are well-established optochemical tools for light-regulated gene expression and have been intensively applied in Escherichia coli and other bacteria including Corynebacterium glutamicum, Pseudomonas putida or Bacillus subtilis. In this study, we aimed to implement a light-mediated on-switch for target gene expression in the facultative anoxygenic phototroph Rhodobacter capsulatus by using different cIPTG variants under both phototrophic and non-phototrophic cultivation conditions. We could demonstrate that especially 6-nitropiperonyl-(NP)-cIPTG can be applied for light-mediated induction of target gene expression in this facultative phototrophic bacterium. Furthermore, we successfully applied the optochemical approach to induce the intrinsic carotenoid biosynthesis to showcase engineering of a cellular function. Photocaged IPTG thus represents a light-responsive tool, which offers various promising properties suitable for future applications in biology and biotechnology including automated multi-factorial control of cellular functions as well as optimization of production processes.


This work
All recombinant DNA techniques were carried out using E. coli DH5α as described by Sambrook et al. [6]. Construction of expression vectors were carried out using restriction and ligation cloning. The PCR fragment containing the gene for lacI q and the Ptac promoter was amplified from a synthetic gene construct on the shuttle vector pEKEx-2 (Eurofins Genomics, Ebersberg, Germany) with oligos 1 and 2 generating appropriate SmaI/XhoI restriction sites at

5) pRholHi_fw
Binds at the 3' end of the plasmid pRholHi-2-eYFP downstream of the lacI gene variant.

6) pRholHi_rev
Binds at the 5' end of the plasmid pRholHi-2-eYFP upstream of the lacI gene variant.

11) CrtEF_genome_fw
Binds at the 5' end of the crtE gene in the R. capsulatus SB1003 genome, inserts homologous regions for InFusion cloning.

16) pRholtHi-crtEF-QC_rev
Binds at the RBS of plasmid pRholtHi-crtE-crtF and inserts optimized RBS/spacer 5'-TATGAAACCTCCTTGTGAAA TTGTTATCCG -3' This work its 5'-and 3'-ends. This fragment was inserted into the likewise hydrolyzed plasmid pRhofHi-2-eYFP to build the plasmid pRholHi-2-eYFP. For construction of the optimized pRholHi-2_Δlac-eYFP plasmid, the redundant 1.5 kb lacI fragment upstream of the original lacI gene was deleted and the lacI gene variant was replaced by the original lacI gene from E. coli K12. Both changes were conducted in one cloning step. For this purpose, three PCR fragments were generated: the first one encompasses the region between the deleted redundant lacI fragment and the original lacI gene of the plasmid backbone (oligos 3 and 6), the second one encompasses a native version of the lacI gene from E. coli (oligos 7 and 8) and the third one encompasses the plasmid backbone after the lacI gene until the redundant lacI region (oligos 4 and 5). Subsequently, the three fragments were assembled via InFusion Cloning (Takara Bio Europe, St Germain en Laye, France). The plasmid pRholtHi-eYFP was constructed using the pRholHi_Δlac-eYFP by changing the RBS consisting of the Shine-Dalgarno and spacer sequence upstream of the eyfp reporter gene. An in silico optimized RBS and RBS-spacer including an NdeI site prior to the start codon of eyfp (for sequences see chapter "DNA sequences of Ptac promoter regions including their RBS, RBS-spacers and/or MCS") was calculated with the Salis Lab RBS calculator (https://salislab.net/software/predict_rbs_calculator) [7]. The plasmid backbone was amplified via PCR using olios 9 and 10, which featured homologous regions containing the optimized RBS and reassembled via InFusion Cloning (Takara Bio Europe, St Germain en Laye, France). For the plasmid pRholtHi-crtE-crtF, the genes crtE and crFE were amplified from the genome of R. capsulatus SB1003 with oligos 11 and 12 generating appropriate homologous regions for InFusion Cloning (Takara Bio Europe, St Germain en Laye, France) at their 5'-and 3'-ends. The plasmid backbone was amplified via PCR from the pRholtHi-eYFP using oligos 13 and 14. The PCR fragment was inserted into the amplified backbone to build the plasmid pRholtHi-crtE-crtF. Additionally, an in silico optimized RBS and RBS-spacer including an NdeI site prior to the start codon of crtE (for sequences see chapter "DNA sequence of CrtE and CrtF from R. capsulatus SB1003 for cIPTG-mediated expression in R. capsulatus cultures") was calculated with the Salis Lab RBS calculator (https://salislab.net/software/predict_rbs_calculator) [7]. The adaption of the RBS spacer was performed via QuikChange using oligos 15 and 16.

Determination of suitable cultivation parameters for aerobic and microaerobic growth of R. capsulatus
Figure S1: (A) Cell growth of R. capsulatus SB1003/pRholHi-2-eYFP expression cultures under varying filling volumes and shaking frequencies. The bacteria were grown in RCV medium for 48 h in Round Well Plates in the dark at 30 °C. To identify filling volumes and shaking frequencies that are appropriate for aerobic and microaerobic growth, the following cultivation conditions were applied: (i) 800 rpm and with 800 µL RCV medium (blue), (ii) 400 rpm and 1000 µL RCV medium (green) and (iii) 400 rpm and 1500 µL RCV medium (grey). Cell growth was analyzed by determining the scattered light intensity using a BioLector system. Values are means of individual biological triplicates. Error bars indicate the respective standard deviations. (B) Dissolved oxygen tension (DOT) of the same R. capsulatus cultures as described in A). The DOT was determined in plates with oxygen sensitive optodes during cultivation of Rhodobacter in the BioLector system (ex = 520 nm, em = 600 nm). Values are means of individual biological triplicates. Error bars indicate the respective standard deviations.
To control the oxygen level during non-phototrophic growth, the filling volume of the Round Well plate as well as the shaking frequencies were appropriately adapted so that a maximal aeration (around 100%) and a minimal aeration (< 25%) were maintained throughout the cultivation. For this purpose, bacterial growth ( Figure S1A) and the dissolved oxygen tension (DOT; Figure S1B) were online-monitored using the scattered light intensity and DO optodes (m2p-labs, Germany), respectively. A filling volume of 800 µL and a shaking frequency of 800 rpm resulted in a constant oxygen tension of 100% ( Figure S1B, blue line). To reduce the DOT during R. capsulatus cultivation, a filling volume of 1000 µL and a shaking frequency of 400 rpm were applied. However, these conditions were not sufficient for constant microaerobic growth, as the oxygen tension increased up to 100% during stationary growth phase ( Figure  S1B, green line). Therefore, a lager filling volume (1500 µL) at the same shaking frequency were used, which led to a strongly decreased oxygen tension of under 25% during both logarithmic and stationary growth phase ( Figure S1B, grey line).  cultures (E,F). For light-induction of eYFP gene expression, growth media were supplemented with 1 mM NP-cIPTG prior cultivation and cultures were subsequently illuminated with UV-A light (365 nm, 2 mW/cm 2 ) for 30 min at the given time points (4h = lag phase, 20 h = early logarithmic growth phase, and 40h = late logarithmic growth phase) or kept in the dark (-). Corresponding control cultures were supplemented with 1mM IPTG at the same time points. While cultures that were induced after both 4h and 20h have 44h for eYFP production (fluorescence measurement after 48h or 64h of the total cultivation time, respectively), the cultivation time could not be expanded accordingly for cultures induced after 40h, because R. capsulatus cultures start to die after approximately 72h under the here applied cultivation conditions. For those cultures, the fluorescence measurement was conducted after 72h of the total cultivation time, corresponding to an eYFP production time of 32h. In vivo fluorescence intensities (eYFP: ex = 508 nm, em = 532 nm) and biomass of all cultures (scattered light intensity at 620 nm or absorption at 660 nm, respectively) were determined at the above stated time points by using a BioLector system or a Tecan Microplate reader and fluorescence values were normalized to cell densities. Values are means of individual biological triplicates. Error bars indicate the respective standard deviations.
The results demonstrate that the induction and uncaging works well even under higher cell densities. The slightly lower fluorescence intensity after 40h in Fig. S5 E is most probably due to a shorter time span from induction to the final eYFP measurement. The elevated scattered light values at the beginning of the cultivation (i.e., before UV-A light exposure, Figure S4 A and B) can be attributed to the poorer water solubility of NP-cIPTG at concentrations of 1 mM, as a certain amount of these compounds initially form emulsions in the cultivation medium. Consequently, they contribute significantly more to the scattered light value than the bacterial cells that are initially still present in low numbers. However, exposure to UV-A light dissolves these emulsions, which is reflected by the rapid decrease in the scattered light intensity after 9 h.

Toxicity and stability of photocaged IPTG variants in different cultivation media
This control experiment with the well-established expression strain E. coli Tuner(DE3) [5] indicated that cIPTG instability was not detectable in sole LB or RCV medium. This gives a first hint that the instability of BC-cIPTG and BEC-cIPTG might be caused by using R. capsulatus as expression host, probably due to host specific enzymes or metabolism products. Evaluation of NIR-light intensities for optimal phototrophic growth of R. capsulatus SB1003 expression cultures Figure S8: (A) Detected NIR light intensities [mW/cm 2 ] at λmax= 850 nm of NIR panels from Vossloh-Schwabe [9] for increasing intensity settings, which can be set via a rotary knob with a continuous adjustment are shown in comparison to the NIR light amount of bulb light (BL) at λmax= 850 nm. Light intensity quantifications were conducted using a Thermal Power Sensor (S302C, Thorlabs Inc, USA). (B) Medium temperature at suitable NIR light intensities after 48 h of cultivation and two pictures of corresponding R. capsulatus SB1003 cultures in the cultivation vessel as well as the respective optical density at 660 nm. The cultivation temperature of 33 °C should not be exceeded to avoid adverse effects on cell growth. The culture without any growth impairment and with the highest cell density is marked with a green frame and can be compared to cultures exposed with bulb light (BL).

Effect of UV-
For the determination of suitable NIR light intensities for efficient phototrophic growth of R. capsulatus, cells were cultivated with NIR light of increasing intensities ranging from 0.5 mW/cm 2 up to 5.1 mW/cm 2 ( Figure S5 A) and analyzed with respect to their growth behavior (Figure S5 B). Screw neck vials, which were used as cultivation vessels, were placed in a distance of approx. 10 cm from each NIR panel. Adequately grown cultures without any sunken cells could only be detected for NIR light intensities of 1.7 mW/cm 2 and 3.1 mW/cm 2 represented by cell numbers corresponding to an optical density at 660 nm of 2.9 and 2.7, which are comparable to the optical density of cultures grown in bulb light (OD660nm= 3.1). Lower or higher NIR light intensities led to decreased cell densities and unequally distributed cultures with sunken cells indicating a hampered cell viability. Presumably, this was on the one hand due to the insufficient exposure intensity and on the other hand due to the excessively high temperature of over 33 °C in the cultivation medium for the highest NIR light intensity. For all following experiments, an NIR light intensity of 1.7 mW/cm 2 was chosen as this condition offers an appropriate medium temperature and the highest cell density after 48 h.