Engineering a tunable bicistronic TetR autoregulation expression system in Gluconobacter oxydans

Bacterial strains and culture conditions

Gluconobacter oxydans 621H (DSMZ 2343) was routinely grown at 30 °C and 250 rpm in yeast mannitol (YM) or yeast glucose (YG) medium (6 g/L yeast extract, 20 g/L mannitol or glucose, 2.5 g/L MgSO₄ × 7 H₂O, 1 g/L (NH₄)₂SO₄, 1 g/L KH₂PO₄, pH 6.0) with 50 μg/ml cefoxitin. Escherichia coli strains (Table 1) were grown in lysogeny broth (LB; 5g/L yeast extract, 10 g/L tryptone, 10 g/L sodium chloride) at 37 °C and 250 rpm. Agar was added to 1.5% when making solid medium. Kanamycin or ampicillin were added to 50 or 100 μg/ml for plasmid maintenance (Table 1). G. oxydans was transformed either by conjugation with E. coli S17-1 or by electroporation (Kallnik et al., 2010; Kiefler, Bringer & Bott, 2017). Competent E. coli 10β and E. coli JM109 were prepared by the TSS method (Chung, Niemela & Miller, 1989).

Materials and molecular techniques

Standard molecular techniques were done according to manufacturer’s protocols or by standard protocols (Sambrook, Fritsch & Maniatis, 1989). The GeneJet Plasmid Miniprep kit (Thermo Fisher Scientific, Waltham, MA, USA) was used to purify plasmids and the GenElute Bacterial Genomic DNA kit (Millipore-Sigma, St. Louis, MO, USA) used to extracted genomic DNA. DNA was amplified using Phusion DNA polymerase and DreamTaq polymerase, and cloning done with FastDigest restriction enzymes and T4 ligase that were purchased from Thermo Fisher Scientific. Eurofins Genomics supplied all primers and sequencing services to confirm plasmids (Louisville, KY, USA) (Table 1).

Plasmid construction

The tetR gene was amplified from pASK-IBA3 (IBA Lifesciences GmbH, Göttingen, Germany) using Phusion DNA polymerase and cloned into the MluI or EcoRI and HindIII sites of pBBR1p264-SPpelB-streplong and pBBR1p452-ST such that there were 6 bp between tetR and its RBS, creating p452TetR and p264TetR (Table 1). P₀₁₆₉ was amplified from G. oxydans DNA and inserted into the SacI and EcoRI sites of p452TetR to create p0169TetR. The P_tet promotor (iGEM BBa_R0040), ribosomal binding site (iGEM BBa_B0034), a MCS region containing both BsaI restriction sites from pASK-IBA3 (IBA Lifesciences GmbH, Göttingen, Germany), and the iGEM BBa_B0010 transcriptional terminator was synthesized and cloned into pUC57 by GenScript (Piscataway, NJ, USA) (Table 1). This P_tet-MCS-terminator region was amplified using pTet_F and pTet_R primers and inserted into the PasI and Bst1107I sites of the p0169TetR, p264TetR, and p452TetR plasmids. Additionally, the uidA gene encoding the β-D-glucuronidase from E. coli DH5α was amplified and inserted into the BsaI sites of the p0169TetR, p264TetR, and p452TetR to create pTET0169-uidA, pTET264-uidA, pTET452-uidA, having 7 bp between uidA and its RBS. All plasmids were confirmed by sequencing.

The luxR gene (iGEM BBa_C0062) was synthesized by GenScript, amplified with Phusion DNA polymerase, and inserted into the MluI or EcoRI and BsaI of pBBR1p264-SPpelB-streplong or pBBR1p452-ST such that there were 9 bp between luxR and its RBS, creating p452LuxR and p264LuxR (Table 1). P₀₁₆₉ was amplified from G. oxydans DNA and inserted into the SacI and EcoRI sites of p452LuxR to create p0169LuxR. The P_lux promotor (iGEM BBa_R0062), ribosomal binding site (iGEM BBa_B0034), MCS region containing two BsaI restriction sites derived from pASK-IBA3 (IBA GmbH), and the iGEM BBa_B0010 transcriptional terminator was synthesized by Eurofins Genomics (Louisville, KY, USA). The synthesized fragment was cut with PasI and Bst1107I and ligated into similarly cut p0169LuxR, p264LuxR, and p452LuxR. The uidA gene encoding the β-D-glucuronidase from E. coli DH5α was also amplified and inserted into the BsaI sites of p0169LuxR, p264LuxR, and p452LuxR, having 10 bp between uidA and its RBS to create pLUX0169-uidA, pLUX264-uidA, pLUX452-uidA and confirmed by sequencing.

The tetracycline-inducible bicistronic expression system was constructed based on pZH512 plasmid that uses the iGEM BBa_R0040 P_tet promoter, the rrnB iGEM BBa_B0010 terminator, and the GFPmut2 reporter (Hensel, 2017). The sequence derived from pZH512 was synthesized and cloned into pUC57-mini by GenScript (Piscataway, NJ, USA). The tetracycline-based bicistronic system was amplified using TetBicis_F/TetBicis_R primers and cloned into the VspI/SacI sites of pBBR1p452-ST such that there were 9 bp between gfpmut2 and its RBS and 7 bp between tetR and its RBS, creating pBICISTRON.

Reporter assay

Well grown cultures of G. oxydans were inoculated 1:50 in new 50 ml baffled shake flasks containing 10 ml YG or YM. Cultures were immediately induced by addition of either anhydrotetracycline (ATc) or N-(3-Oxohexanoyl)-L-homoserine lactone (AHL; Chemodex Ltd, St. Gallen, Switzerland) to the indicated final concentrations and grown to an OD₆₀₀ of about 0.8. The β-D-glucuronidase (UidA) activity was monitored in Miller Units as previously described (Kallnik et al., 2010). Cells were permeabilized by adding 30 μl cells to a 96 well plate containing 100 μl Z-buffer (60 mM Na₂HPO₄ × 7H₂O, 40 mM NaH₂PO₄ × H₂O, 10 mM KCl, 1 mM MgSO₄ × 7H₂O, 3.5 mM 2-mercaptoethanol, 5% (v/v) chloroform, and 0.5% SDS, pH) and incubated for 10 min at 30 °C. Absorbance was monitored at 405 nm every minute on a BioTek EL808 plate reader after addition of 100 μl 4-nitrophenyl-β-D-glucuronide (4 mg/ml). All assays were done in at least three biological replicates each with three technical replicates and analyzed using BioTek Gen5 software.

Alternatively, expression was monitored by fluorescence using the GFPmut2 protein. Induced cultures were monitored for fluorescence in late log phase (OD₆₀₀~0.8). Fluorescence was measured with a λex of 485/4 nm and λem 508/4 nm using a SpectraMax M3 plate reader (Molecular Devices, San Jose, CA, USA). Fluorescence was measured in relative fluorescence units (RFU) and normalized to the optical density of the culture.

Microscopy

G. oxydans pBICISTRON and wildtype were grown on YG to approximately 0.4 OD₆₀₀. Anhydrotetracycline was added to 100 ng/ml and cultures were incubated overnight. Clean depression slides were preheated in a 37 °C incubator. Molten buffered agarose (25 mM Tris-Base pH 7.4, agarose 1.2%) was applied to the depression slide and a glass slide was placed on top to form an agar pad. The slide was removed once solidified and 5 µL of cells were placed onto the center of the pad. A Nikon Eclipse E600 fluorescence microscope (100×/1.40 numerical aperture oil immersion objective) with a Nikon DS-Qi2 camera. Fluorescence was measured using λex 460/40 nm / λem 510/50 nm filter. Digital images were acquired and analyzed with the NIS-Elements imaging software (version 5.20.01).

Statistical analysis

R Studio was used to perform statistical analyses and to generate all plots (R Core Team, 2021). Analyses can be found in Script S1.

Regulation of the activation-dependent Luxr expression system in G. oxydans

The P_lux promoter is a common system used to induce gene expression in many bacteria. The LuxR-P_lux system was successfully used to control gene expression in K. rhaeticus (Florea et al., 2016). However, species and genera differences in regulation have been reported in the AAB (Fricke et al., 2021a; Fricke et al., 2020; Fricke et al., 2021b). Based on its reported high induction and with relatively lower leakiness in comparison to TetR-based expression control, a LuxR-P_lux system was examined in G. oxydans. The luxR gene was cloned under the control of either the constitutive strong (P₀₁₆₉ and P₂₆₄) or moderate (P₄₅₂) promoters (Kallnik et al., 2010; Yuan et al., 2016). Induction by addition of AHL was monitored with the uidA reporter gene under the control of the P_lux promoter (Fig. 1). Regardless of the promoter used to express luxR, induction ratios were similar having average induction of 4.3 ± 0.31, 3.5 ± 0.17, and 4.90 ± 0.45 when P₀₁₆₉, P₂₆₄, and P₄₅₂ promoters were used respectively (Fig. 2). Stronger absolute expression was observed when the P₀₁₆₉ promoter was used to control luxR gene expression. However, induction ratios were similar to P₂₆₄, and P₄₅₂ despite the higher absolute activity. However, the amount of AHL inducer from 0.1 to 10 μM did not influence induction regardless of the strength of the promoter used to express the luxR gene. As such, this system functions as an on-off switch in the AHL concentration range tested rather than a tunable induction system in G. oxydans. Importantly, the LuxR-P_lux system has previously been reported as inducible with lower leakiness in the acetic acid bacterium K. rhaeticus (Florea et al., 2016). However, the condition-dependent induction was only 5-fold in free-living cells and 12-fold in cellulose pellicles. The moderate levels of induction are attributed to the leakiness of this system in AAB. Indeed, similar induction levels are observed in G. oxydans and are also attributed to the moderate leakiness of the induction system. These data suggests that while the LuxR-P_lux system shows good induction for heterologous gene expression relative to previously described systems in acetic acid bacteria (Florea et al., 2016), it is limited by its functionality as an on/off switch. Furthermore, tight control of gene expression is not possible due to the high background expression when not induced.

Regulation of the constitutive repression TetR expression system in G. oxydans

Another common system used to control heterologous gene express in bacteria is the TetR-P_tet system. This system is a constitutive repression system, where TetR constitutively represses transcription of the target gene by binding the P_tet promoter in the absence of a TetR effector. This system was examined for functionality in K. rhaeticus where gene expression was hindered by high leakiness of the repressor resulting in only circa 1.5-fold induction (Florea et al., 2016). However, variable expression in different AAB hosts has been previously observed, highlighting the need to examine expression in different hosts (Teh et al., 2019). To investigate its functionality in G. oxydans, a TetR-P_tet system was examined. The tetR gene was cloned under the control of the constitutive P₀₁₆₉, P₂₆₄, and P₄₅₂ promoters (Kallnik et al., 2010; Shi et al., 2014). Induction was monitored by addition of ATc with the uidA reporter gene under the control of the P_tet promoter derived from Tn10 (Fig. 3). The TetR-P_tet had strong maximal induction of 19-fold, 18.6-fold, and 4-fold when P₀₁₆₉, P₂₆₄, and P₄₅₂ promoters were used to control tetR transcription respectively (Fig. 4). While promoters P₀₁₆₉ and P₂₆₄ had similar induction ratios, absolute induction was 6-fold higher when the P₀₁₆₉ promoter was used. This disparity in absolute activity compared to induction ratio is attributed to the somewhat higher background expression when the P₀₁₆₉ promoter was used. Interestingly, the induction ratio of P₄₅₂ was the lowest at 4-fold. This is likely due to the weaker expression of the TetR repressor. The lower amount of TetR in the cytoplasm likely is not sufficient to fully repress the P_tet promoter, resulting in the increased amount of leaky expression that leads to reduced induction ratios. Indeed, the absolute levels of induction when the P₂₆₄ and P₄₅₂ promoter was used are similar, suggesting that the differences in induction ratios is driven by differences in promoter strength. However, induction of the TetR-P_tet system using the P₄₅₂ promoter to express tetR is similar to induction with the LuxR-P_lux system in G. oxydans and in K. rheaticus (Florea et al., 2016). Notably, the TetR-P_tet system showed tunable expression in all cases, showing an increased response to increasing ATc concentrations in the concentration range examined. Interestingly, the TetR-P_tet system showed stronger induction than the LuxR-P_lux system in G. oxydans, which differs from observations in K. rhaeticus (Florea et al., 2016). However, the TetR-P_tet system still had low levels of leaky expression when uninduced, limiting induction ratios.

Construction of a tunable bicistronic TetR expression system in G. oxydans

While the above described TetR-P_tet system showed strong induction and tunability, a low level of leakiness was still observed. This is similar to previously reported results of high basal level of expression of the Tet system in acetic acid bacteria in the absence of inducer (Florea et al., 2016; Fricke et al., 2021a). In E. coli, a bicistronic autoregulation P_tet system was predicted to have the lowest background noise and predicted to be tunable based on modeling of various gene configurations. These predictions were also validated experimental to show the optimal signal-to-noise ratio of a bicistronic gene arrangement (Hensel, 2017). We used this as a model to construct a bicistronic TetR system using a GFP reporter using the moderate strength RBS (aaagCCgagaaaggtaccgcATG, Hensel, 2017) in the 5′-UTR of the gfpmut2 gene (Fig. 5). Like the TetR-P_tet system, the bicistronic system showed a tunable response up to 200 ng/ml ATc addition (Fig. 6). The maximum GFP induction ratio was a 3,037-fold increase in fluorescence. Importantly, the bicistronic system was tightly controlled, having little to no fluorescence in the absence of inducer and signal only slightly above background levels of wildtype cells (Fig. 6). This is an approximately 160-fold improvement of induction compared to the TetR-P_tet constitutive repression system in G. oxydans (Figs. 3 and 4). This is attributed to the low leakiness of the bicistronic system that increases induction ratios dramatically. Consequently, the bicistronic TetR gene arrangement is a promising system for gene control in G. oxydans and other AAB.