Engineering a tunable bicistronic TetR autoregulation expression system in Gluconobacter oxydans

Acetic acid bacteria are well-known for their ability to incompletely oxidize their carbon sources. Many of the products of these oxidations find industrial uses. Metabolic engineering of acetic acid bacteria would improve production efficiency and yield by allowing controllable gene expression. However, the molecular tools necessary for regulating gene expression have only recently started being explored. To this end the ability of the activation-dependent Plux system and two constitutive repression Ptet systems were examined for their ability to modulate gene expression in Gluconobacter oxydans. The activation-dependent Plux system increased gene expression approximately 5-fold regardless of the strength of the constitutive promoter used to express the luxR transcriptional activator. The Ptet system was tunable and had a nearly 20-fold induction when the tetR gene was expressed from the strong constitutive promoters P0169 and P264, but only had a 4-fold induction when a weak constitutive promoter (P452) was used for tetR expression. However, the Ptet system was somewhat leaky when uninduced. To mitigate this background activity, a bicistronic TetR expression system was constructed. Based on molecular modeling, this system is predicted to have low background activity when not induced with anhydrotetracycline. The bicistronic system was inducible up to >3,000-fold and was highly tunable with almost no background expression when uninduced, making this bicistronic system potentially useful for engineering G. oxydans and possibly other acetic acid bacteria. These expression systems add to the newly growing repertoire of suitable regulatable promoter systems in acetic acid bacteria.


INTRODUCTION
Acetic acid bacteria (AAB) are a unique group of organisms that incompletely oxidize their carbon and energy substrates (e.g. sugars, polyols, alcohols). This incomplete oxidation is done using a set of membrane-bound dehydrogenases that are oriented toward the periplasmic space and release their products into the medium (Deppenmeier & Ehrenreich, 2009;Prust et al., 2005). This feature has been widely used in the production of vinegar, vitamin C, 6-amino-L-sorbose to produce the antidiabetic drug miglitol, tanning agents (dihydroxyactone and erythrulose), and flavor and fragrance products (Deppenmeier, Hoffmeister & Prust, 2002;Wang et al., 2016).
The ability to express target genes in acetic acid bacteria is of recent interest, especially in the industrially important genera Gluconobacter, Gluconacetobacter, Komagataeibacter, Acetobacter, and Acidiphilium. Gene expression allows basic research into the physiology and metabolism of these bacteria, as well as for genetic and metabolic engineering. To date most engineering in AAB has been done using constitutive promoters (Blank & Schweiger, 2018;Kallnik et al., 2010;Kiefler, Bringer & Bott, 2015;Kosciow et al., 2014;Merfort et al., 2006;Meyer, Schweiger & Deppenmeier, 2013;Siemen et al., 2018;Yuan et al., 2016). However, regulatable expression is often desired. Despite this, few regulatable systems have been reported and those that have are often reported to be leaky, resulting in low induction ratios (reviewed in: (Fricke et al., 2021b)). For example, the well-studied LacI system had low induction and high leakiness in the AAB Gluconobacter spp. and K. xylinus (Condon, FitzGerald & O'Gara, 1991;Liu et al., 2020). Similarly, the acyl homoserine lactone (AHL) inducible LuxR-P lux had high expression, but only 4-fold induction in free-living cells due to leakiness in uninduced K. rhaeticus (Florea et al., 2016). Recently two systems for regulatable gene expression in Gluconobacter oxydans have been reported, a L-arabinoseinducible AraC-P BAD system and an ATc-inducible Tn10-based TetR-dependent system. The L-arabinose-inducible AraC-P BAD system was reported to have low background leakiness and was inducible to 480-fold (Fricke et al., 2020). Interestingly, a similar system in K. rhaeticus was highly leaky and had low induction (Teh et al., 2019). While the Tn10based TetR-dependent system was tightly regulated and is highly tunable in G. oxydans, a similar system was leaky and had low induction in K. rhaeticus (Florea et al., 2016;Fricke et al., 2021b). This highlights differences in control within the AAB despite their high evolutionary similarity De Vuyst et al., 2008;Papalexandratou et al., 2009). Therefore, the response in each host must be confirmed prior to use.
To address the problem of low induction and leaky gene expression in AAB, we investigated the use of the LuxR-P lux system and a bicistronic TetR-dependent system for tightly regulated gene expression in G. oxydans using a pBBR1MCS-2 plasmid backbone. While both systems were regulatable, the amount of basal level expression varied between these systems. While the LuxR-P lux system showed strong induction, it lacked the ability for tunable expression of the bicistronic TetR system. These regulatory systems will add to the growing ability to control gene expression in AAB.

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). P 0169 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 DH5a 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). P 0169 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 DH5a 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 OD 600 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 ml cells to a 96 well plate containing 100 ml Z-buffer (60 mM Na 2 HPO 4 × 7H 2 O, 40 mM NaH 2 PO 4 × H 2 O, 10 mM KCl, 1 mM MgSO 4 × 7H 2 O, 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 ml 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 (OD 600~0 .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 OD 600 . 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 (P 0169 and P 264 ) or moderate (P 452 ) 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 P 0169 , P 264 , and P 452 promoters were used respectively (Fig. 2). Stronger absolute expression was observed when the P 0169 promoter was used to control luxR gene expression. However, induction ratios were similar to P 264 , and P 452 despite the higher absolute activity. However, the amount of AHL inducer from 0.1 to 10 mM 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. Variants and regulation of the pBBR1MCS-2-based LuxR-P lux expression system with a UidA reporter gene. (A) Plasmid map of pLUX0169-uidA, pLUX452-uidA, and pLUX264-uidA. LuxR is derived from A. fischeri iGEM part BBa_C0062 and is constitutively transcribed from either strong promoters P 0169 or P 264 or the moderate P 452 promoter (Kallnik et al., 2010;Shi et al., 2014). LuxR is terminated by the E. coli lpp terminator (Nishi et al., 1984). The uidA reporter gene is expressed from the Lux promoter/operator region from iGEM BBa_0062 and is terminated by the rrnB terminator from iGEM BBa_B0010. The strong ribosomal binding site from iGEM BBa_B0034 is used for translation of both LuxR and UidA (Hentschel et al., 2013). (B) Regulation of the activation-dependent LuxR-P lux expression system. Binding of AHL allows binding of LuxR-AHL to the P lux promoter. UidA hydrolyzes pNPG to produce a yellow color that can be quantified colorimetrically. Full-size  DOI: 10.7717/peerj.13639/ fig-1 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 P 0169 , P 264 , and P 452 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 P 0169 , P 264 , and P 452 promoters were used to control tetR transcription respectively (Fig. 4). While promoters P 0169 and P 264 had similar induction ratios, absolute induction was 6-fold higher when the P 0169 promoter was used. This disparity in absolute activity Figure 3 Variants and regulation of the pBBR1MCS-2-based TetR-P tet expression system with a UidA reporter gene. (A) Plasmid map of pTET0169-uidA, pTET452-uidA, and pTET264-uidA. The tetR gene is derived from the pASK-IBA3 plasmid (IBA Lifesciences GmbH, Göttingen, Germany). It is constitutively transcribed from either strong promoters P 0169 or P 264 or the moderate P 452 promoter (Kallnik et al., 2010;Shi et al., 2014). An E. coli lpp terminator was used to terminate tetR transcription (Nishi et al., 1984). The uidA reporter gene is expressed from the Tet promoter/operator region from iGEM BBa_0040 and is terminated by the rrnB terminator from iGEM BBa_B0010. The strong ribosomal binding site from iGEM BBa_B0034 is used for translation of both tetR and uidA (Hentschel et al., 2013). (B) Regulation of the constitutive repression TetR-P tet expression system. TetR binds ATc, relieving repression of the UidA reporter, which hydrolyzes pNPG, producing a yellow color that can be quantified colorimetrically. Full-size  DOI: 10.7717/peerj.13639/ fig-3 compared to induction ratio is attributed to the somewhat higher background expression when the P 0169 promoter was used. Interestingly, the induction ratio of P 452 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 P 264 and P 452 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 P 452 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 Figure 4 Activity of the TetR-P tet systems in G. oxydans. Background expression was low and induction was tunable with all three promoters.
(A) The PTET169-uidA using the strong P 0169 promoter to express the luxR gene had an induction ratio of 19-fold (B) pTET264-uidA using the strong P 264 promoter to express the luxR gene had an induction ratio of 18.6-fold (C) pTET452-uidA using the moderate P 452 promoter to express the luxR gene had an induction ratio of 4-fold. Data represent at least three biological replicates each with three technical repeats. An asterisk ( Ã ) indicates that the Kruskal-Wallis test had a p < 0.001. Full-size  DOI: 10.7717/peerj.13639/fig-4 (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.

DISCUSSION
We looked at the use of three regulatable expression systems in G. oxydans, one using the common LuxR-P lux system and two genetic arrangements of the TetR-P tet system. The absolute induction of the LuxR-P lux system was high in G. oxydans, having and average activity of 9,750, 4,600, and 5,300 Miller Units with the P 0169 , P 264 , and P 452 promoters respectively (Fig. 2). However, the induction ratio was only about 4-fold regardless of the constitutive promoter used to express the luxR gene. This is similar to the induction ratio of 5-fold reported in K. rhaeticus free-living cells (Florea et al., 2016). The low induction ratio maybe due to the presence of rare G. oxydans codons in luxR that lead to slower translation rates or inclusion body formation, which would decrease reporter activity and induction ratios. However, slower translation has also been shown to increase the ratio of folded protein relative to rapid translation using only common codons, which can increase total protein while decreasing the folding yield ratio (Rodriguez et al., 2018). The induction ratio was also limited because of the leakiness of the expression system. In Aliivibrio fischeri, AHL is produced by a synthase encoded by the luxI gene. The expression of the luxI gene, along with the other bioluminescence genes, is under the control of the activator, LuxR (Fuqua, Winans & Greenberg, 1994). Yet, basal levels of AHL are still produced. It may be that LuxR has weak binding affinity for P lux in the absence of AHL, causing leaky transcription in G. oxydans. In the absence of the luxR gene, UidA activity was not detected. Consequently, the difference in background expression is likely due to the increase in basal level of expression of LuxR from the different constitutive promoters, with the strongest P 0169 promoter enhancing basal level expression due to the increased levels of LuxR. Furthermore, the presence of a GinI/GinR quorum sensing system has been reported to be present in most AAB (Iida, Ohnishi & Horinouchi, 2008). This system is Lux-like and responds to AHLs. However, homologs of GinI/GinR were not identified by blastp. Despite this, an unknown GinI-like AHL-synthase or other undescribed AHL-dependent regulatory system may be present in G. oxydans. This would produce low levels of AHLs that may interact with LuxR, causing basal levels of expression, which may also explain the increased leakiness when the strong P 0169 promoter was used. Another drawback to this system is that it is not tunable with the AHL concentrations tested. Rather it functions as an on/off switch with equal induction ratios when using 0.1 µM AHL up to 10 mM of AHL (Fig. 2). Regardless of these shortcomings, the LuxR-P lux system is a promising candidate to control gene expression in G. oxydans, especially when high absolute levels of expression are desired or the basal-level expression is not of concern.
Alternatively, the TetR-P tet system has much lower levels of leaky expression in G. oxydans and up to 19-fold induction, which is approximately four-times higher compared to the LuxR-P lux system. Additionally, the TetR-P tet system was tunable, unlike the LuxR-P lux system. Interestingly, a similar TetR system was examined in K. rhaeticus but only had induction ratios of about 1.5-fold, despite both systems having a similar genetic arrangement (Florea et al., 2016). One difference in genetic arrangement is the presence of an intervening neo/kan cassette between the tetR gene and the P tet promoter (Fig. 3). It was suggested that terminator read-through may cause high leakiness of the P tet promoter in K. rhaeticus (Fricke et al., 2021b). Molecular simulations also showed that TetR dimerization is inhibited at low pH (Fricke et al., 2021b). K. rhaeticus is often grown in medium containing glucose or sucrose with an initial pH of 6.0, which decreases during growth (Florea et al., 2016;Hestrin & Schramm, 1954). The cytoplasmic pH of acetic acid bacteria can decrease to about 4.0 during growth (Menzel & Gottschalk, 1985), which may be responsible for the leakiness seen in high acid producing AAB like K. rhaeticus (Fricke et al., 2021b). Additionally, the separation of the tetR gene and the P tet promoter in our G. oxydans expression system likely preclude terminator read-through as the cause of the leaky expression. Instead, the leakiness may be attributed to the choice of promoter controlling the transcription of tetR. K. rhaeticus contains the P J23118 or P lacI promoters for expression of tetR rather than native host constitutive promoters (Florea et al., 2016). It may be that the choice of promoter is important for sufficient expression of TetR and the repression of the P tet promoter, causing a tighter control of gene expression. Additionally, the promoter must not produce excessive TetR to avoid inclusion bodies that could reduce the concentration of functional TetR, resulting in leakiness. Indeed, the highest amount of uninduced leakiness was observed when the weaker P 452 promoter controlled tetR expression (Fig. 4).
In an attempt to reduce the amount of leakiness and improve induction of the TetR-P tet system, we engineered a bicistronic TetR expression system. This gene arrangement has the best signal-to-noise ratio of any gene arrangement for the TetR-controlled expression in E. coli (Hensel, 2017). In G. oxydans, this system exhibited very low uninduced expression with high induction ratios of >3,000-fold. Importantly, the system was highly tunable when up to 200 ng/ml of ATc was added. Recently, another TetR-based system was reported in G. oxydans. This system uses the native Tn10 gene orientation. This Tn10based system was also tunable and had induction ratios up to 3,674-fold (Fricke et al., 2021b), which is similar to the bicistronic system reported here. Tunability is important in many instances. For example, membrane-bound dehydrogenases in AAB often require low-to-moderate gene expression for functional enzyme production (Mientus et al., 2017). Also, in some cases strong gene expression is known to produce inclusion bodies. Metabolic and genetic engineering often requires that genes are expressed at lower levels. Therefore, the ability to tune gene expression up or down as needed is essential in these applications. In these cases, the bicistronic or Tn10-based TetR-dependent systems may be preferred over the LuxR system that is not tunable and shows higher basal levels of uninduced expression.

CONCLUSIONS
It recently has become possible to control and tune gene expression in AAB and there are now multiple systems for gene control. In addition to the LuxR-P lux systems described here and in K. rhaeticus, there are two known tunable and highly inducible TetR-P tet systems (Florea et al., 2016;Fricke et al., 2021b). Another promising system for regulatable gene expression in AAB is the AraC-P BAD system. This system had low leakiness and is inducible up to 480-fold in G. oxydans. Importantly, this system was also tunable by varying the amount of arabinose added to the medium (Fricke et al., 2020). The current development of multiple tunable expression systems in AAB is advantageous for future studies into not only basic biological questions, but also for engineering of new strains for industrial productions.