Production of the biocommodities butanol and acetone from methanol with fluorescent FAST-tagged proteins using metabolically engineered strains of Eubacterium limosum

Background The interest in using methanol as a substrate to cultivate acetogens increased in recent years since it can be sustainably produced from syngas and has the additional benefit of reducing greenhouse gas emissions. Eubacterium limosum is one of the few acetogens that can utilize methanol, is genetically accessible and, therefore, a promising candidate for the recombinant production of biocommodities from this C1 carbon source. Although several genetic tools are already available for certain acetogens including E. limosum, the use of brightly fluorescent reporter proteins is still limited. Results In this study, we expanded the genetic toolbox of E. limosum by implementing the fluorescence-activating and absorption shifting tag (FAST) as a fluorescent reporter protein. Recombinant E. limosum strains that expressed the gene encoding FAST in an inducible and constitutive manner were constructed. Cultivation of these recombinant strains resulted in brightly fluorescent cells even under anaerobic conditions. Moreover, we produced the biocommodities butanol and acetone from methanol with recombinant E. limosum strains. Therefore, we used E. limosum cultures that produced FAST-tagged fusion proteins of the bifunctional acetaldehyde/alcohol dehydrogenase or the acetoacetate decarboxylase, respectively, and determined the fluorescence intensity and product concentrations during growth. Conclusions The addition of FAST as an oxygen-independent fluorescent reporter protein expands the genetic toolbox of E. limosum. Moreover, our results show that FAST-tagged fusion proteins can be constructed without negatively impacting the stability, functionality, and productivity of the resulting enzyme. Finally, butanol and acetone can be produced from methanol using recombinant E. limosum strains expressing genes encoding fluorescent FAST-tagged fusion proteins. Supplementary Information The online version contains supplementary material available at 10.1186/s13068-021-01966-2.

faces technical challenges as the mass transfer of gases into the liquid state of media is quite poor and consequently limits microbial productivity [1]. An alternative to the C1 gases as a substrate is the use of methanol, since it bears the advantage of easy transport, storage, and is completely soluble in water, and does, therefore, not suffer from mass transfer issues [2]. Methanol can be produced from CO-, CO 2 -, and H 2 -containing syngas [3] and is a cheap, sustainable, and relatively pure feedstock for a variety of bacteria [4][5][6][7]. Several acetogens can utilize methanol via the Wood-Ljungdahl pathway [8], which has been elucidated in detail for Acetobacterium woodii [9]. Since the production of chemicals with methanol-utilizing acetogens seems to be auspicious, it is also of great interest to expand their molecular toolbox. Molecular tools with different levels of possibilities for metabolic engineering are available for acetogens such as A. woodii, Clostridium ljungdahlii, C. autoethanogenum, and E. limosum. That includes, e.g., genomic editing tools [10][11][12][13][14] and the expression of recombinant pathways to produce biocommodities such as butanol [15,16], acetone [17][18][19], isopropanol [20], 3-hydroxybutyrate [21], or poly(3-hydroxybutyrate) [22]. However, fluorescent reporter systems which are well-established and often used tools in molecular biology to study gene expression [23,24], promoter activities [25,26], or the dynamics in microbial populations and co-cultures [27][28][29] are still restricted for acetogens, basically due to the lack of proteins that show bright fluorescence under anaerobic conditions. Recently, the fluorescence-activating and absorption shifting tag (FAST) [30] was established in C. acetobutylicum [31] and C. ljungdahlii [32] and opened the door for application in other anaerobic bacteria, since its fluorescence is bright and independent of oxygen. FAST is a small-sized protein with a mass of 14 kDa that only shows fluorescence when it forms a non-covalent reversible complex with a fluorogenic ligand, a so-called fluorogen [30,33]. Those fluorogens are hydroxybenzylidene rhodanine derivatives that are non-fluorescent by themselves and only show fluorescence when bound to FAST [30]. Due to the small size, FAST is perfectly suited for genetic fusion to any protein of interest (POI) [30]. Such fusion proteins enable investigations of protein localization and intracellular dynamics [31,[34][35][36].
In this study, we aimed to expand the genetic toolbox of the Gram-positive, methanol-utilizing acetogen E. limosum [37], which is one of the few acetogens that is genetically accessible [13,38], by establishing FAST as a fluorescent reporter system. We used FAST to construct FAST-tagged fusion proteins. As an initial approach, the bifunctional acetaldehyde/alcohol dehydrogenase (AdhE2) from C. acetobutylicum, which mediates the reaction from butyryl-CoA to butanol [39], was tagged with FAST and the respective gene expressed in recombinant E. limosum strains (Fig. 1). Since E. limosum produces butyryl-CoA naturally [40], the heterologous expression of the gene encoding the FAST-tagged AdhE2 fusion protein is sufficient to produce butanol. As a second approach, the acetoacetate decarboxylase (Adc) originating from C. acetobutylicum was FAST-tagged and the respective gene used to assemble an artificial acetone production operon (APO) also including the genes thlA and ctfA/B (encoding thiolase and acetoacetyl-CoA: acetate/butyrate-CoA transferase) originating from C. acetobutylicum [41,42]. Thus, acetone production can be achieved by expression of the APO using recombinant E. limosum strains (Fig. 1). Evidence for heterologous production of the FAST-tagged AdhE2 and Adc fusion proteins within E. limosum mutants was provided by the determination of the fluorescence intensity of the respective bacterial populations producing the FAST-tagged fusion proteins during growth. Moreover, a comparison of production patterns of the biocommodities butanol and acetone using the C1 carbon source methanol provides first insight regarding stability, functionality, and productivity of the AdhE2 and Adc FAST-tagged fusion proteins.

Growth of E. limosum using different carbon sources
E. limosum NG-6894 was cultivated heterotrophically using glucose or methanol as well as autotrophically using H 2 + CO 2 and syngas as carbon and energy source. Heterotrophically cultivated cells using glucose reached a maximal optical density (OD 600 ) of 2.83 with a growth rate of 0.23 h −1 . Cell cultures grown on various methanol concentrations of 50, 100, and 200 mM showed maximal OD 600 values of 1.4, 2, and 2.6, respectively, while having a similar growth rate of about 0.09 h −1 (Fig. 2a and Table 1). Autotrophically grown cells using H 2 + CO 2 showed the lowest growth rate of 0.02 h −1 and a maximal OD 600 value of 0.6. Cultivation using syngas resulted in a maximal OD 600 of 2.9 and a growth rate of 0.06 h −1 (Fig. 2b and Table 1). E. limosum NG-6894 produced acetate and butyrate under heterotrophic as well as under autotrophic growth conditions, while the ratio between these products varied depending on the substrates used. Acetate and butyrate production started at the end of the exponential growth phase and reached their maximum in the stationary growth phase after the substrate was consumed completely. Heterotrophically grown cells using glucose showed a product ratio of acetate:butyrate of 6:1, which shifted to 2:1 when cells were cultivated with methanol, independent of the used amount of methanol (Table 1). E. limosum NG-6894 cells cultivated autotrophically using H 2 + CO 2 produced the highest amounts of acetate (85 mM) and only low amounts of butyrate (2.4 mM), which resulted in a product ratio of 36:1. Cells grown using syngas achieved an acetate:butyrate ratio of 9:2 ( Table 1). As a control, cells cultivated without the addition of a defined carbon source except for the standard medium components carbonate and yeast extract accomplished only one doubling and produced traces of acetate (2.2-6 mM).

Engineering of the fluorescence activation and absorption shifting tag in E. limosum
After transformation, the recombinant E. limosum [pMTL83251_P bgaL _FAST] expressed the FAST-encoding gene (feg) controlled by the lactose-inducible bgaR-P bgaL promoter. Cells of that strain showed bright fluorescence during growth only in the presence of the fluorogen TF Lime after induction of gene expression (Fig. 3). E. limosum [pMTL83251_P bgaL _FAST] was cultivated on glucose and reached a 26-fold higher maximum fluorescence intensity after 48 h of cultivation compared to the non-induced strain. Cells of the non-induced strain did not show any fluorescence except autofluorescence in the presence of TF Lime. Both, the induced and the non-induced strain reached similar maximal OD 600 values of 3 and 3.1 after 24 h of cultivation, respectively. The empty vector control E. limosum [pMTL83251] reached an OD 600 of 2.7 and showed only autofluorescence when supplemented with TF Lime (Fig. 3a). Fluorescence microscopy was used to image fluorescent cells (Fig. 3b) [9] coupled with the recombinant butanol and acetone production pathway in E. limosum. Butanol and acetone production can be achieved with FAST-tagged AdhE2 and Adc fusion proteins, respectively. ThlA, thiolase; CtfA/B, acetoacetyl-CoA:acetate/butyrate-CoA transferase; Adc, acetoacetate decarboxylase; Hbd, 3-hydroxybutyryl-CoA dehydrogenase; Crt, crotonase; Bcd, butyryl-CoA dehydrogenase; AdhE2, bifunctional acetaldehyde/alcohol dehydrogenase 62% of induced E. limosum [pMTL83251_P bgaL _FAST] cells were fluorescent, while 38% of them were non-fluorescent. The number of fluorescent cells increased in the stationary growth phase up to 72%. In the late stationary growth phase, after 72 h of cultivation, the number of fluorescent cells decreased to 59%. The population of the non-induced cells showed no shift and were non-fluorescent during all determined time points (Fig. 3d).
Furthermore, the fluorescence of FAST-producing cells cultivated with methanol as carbon and energy source was examined. In addition to E. limosum [pMTL83251_P bgaL _FAST] a strain termed E. limosum [pMTL83251_P thlA _FAST] was constructed that expressed feg under the control of the constitutive P thlA promoter. E. limosum [pMTL83251] harboring the empty vector control showed only autofluorescence when the fluorogen TF Lime was supplemented (Fig. 4a). The constitutive expression of feg by E. limosum [pMTL83251_ P thlA _FAST] caused a clear and increasing fluorescence during growth and a maximum intensity during the stationary stage (Fig. 4b). The fluorescence intensity of cells that produced FAST constitutively was 5. Also, tdFAST2 was codon optimized for E. limosum and the respective gene used to construct the plasmid pMTL83251_P bgaL _tdFAST2 harboring the gene tdFAST2-opt under the control of the bgaR-P bgal lactose-inducible promoter (Fig. 4e). The induced expression of the codon-optimized tdFAST2-opt gene with E. limosum [pMTL83251_P bgaL _tdFAST2] resulted in a 1.5-fold improved fluorescence compared to E. limosum [pMTL83251_P bgaL _FAST] expressing feg controlled by the same promoter (Fig. 4f ). The non-induced cells of E. limosum [pMTL83251_P bgaL _tdFAST2] only showed autofluorescence.
Moreover, the fluorescence intensity of recombinant E. limosum strains expressing feg or tdFAST2-opt was examined when supplemented with the red fluorescent dye TF Coral after excitation at 516 nm (Fig. 4f ). The use of the fluorogen TF Coral caused clear fluorescence of E. limosum [pMTL83251_P bgaL _FAST] after induction of feg gene expression and showed 3-and 2.6-fold higher intensities compared to non-induced cells and those harboring the empty vector control, respectively. Furthermore, induced E. limosum [pMTL83251_P bgaL _tdFAST2] cells Butanol production using FAST-tagged AdhE2 fusion proteins by recombinant E. limosum strains FAST was fused to the C-or the N-terminus of AdhE2 (CA_P0035) using a flexible glycine linker. The resulting fusion proteins were used to examine the impact of the fluorescent FAST tag on functionality and productivity of AdhE2. Both AdhE2 fusion proteins were heterologously produced using the strains E. limosum [pMTL83251_ produced the native AdhE2 from C. acetobutylicum and was constructed and served as a control for butanol production using glucose or methanol as a carbon source.
Cultivation of E. limosum [pMTL83251_P bgaL _C-FAST-AdhE2] using glucose as a carbon source and induced gene expression resulted in a maximal OD 600 of 2.5 and a growth rate of 0.16 h -1 . Non-induced cells reached an OD 600 of 3 with a growth rate of 0.17 h -1 . Clear fluorescence was observed after induction of gene expression, proving production of the C-terminal FAST-tagged AdhE2 fusion protein. Fluorescence increased during growth and reached a maximum in the late exponential growth phase. Non-induced cells showed only autofluorescence. The acetate:butyrate ratio was 10:  Table 2, and Additional file 1: Table S1). E. limosum [pMTL83251_P bgaL _C-FAST-AdhE2] cultivated using methanol as a carbon source showed a growth rate of 0.05 h -1 with induced and non-induced gene expression and reached OD 600 values of 1.6 and 1.8, respectively. Clear fluorescence was determined after induction of gene expression that caused production of the FAST-tagged AdhE2 fusion protein and resulted in a 3.7-fold higher intensity compared to autofluorescence of non-induced cells. The induced culture showed an acetate:butyrate ratio of 3:1, while the non-induced culture showed a ratio of 2:1. Only the induced culture of E. limosum [pMTL83251_P bgaL _C-FAST-AdhE2] produced traces of ethanol (0.1 mM) and up to 0.4 mM butanol from methanol at the end of the exponential growth phase. Non-induced strains did not produce any alcohols (Fig. 5c, Table 3, and Additional file 1: Table S2).  Table 2, and Additional file 1: Table S1).
E. limosum [pMTL83251_P bgaL _N-FAST-AdhE2] cultivated using methanol showed a growth rate of 0.06 h -1 , while the induced and non-induced cells reached an OD 600 of 1.7 and 2, respectively. Induced cells produced the N-terminal FAST-tagged AdhE2 fusion protein after induction of gene expression and showed distinct fluorescence during growth with a maximum at the early stationary stage. Non-induced cells only showed autofluorescence. The acetate:butyrate ratio shifted from 3:1 to 2:1 for the induced and non-induced cells, respectively. Consequently, that strain produced 0.2 mM of ethanol and 0.6 mM of butanol at the end of the exponential growth phase. Non-induced cells again did not produce any alcohols (Fig. 5d, Table 3, and Additional file 1: Table S2). Finally, the strain that produced the N-terminal tagged version of AdhE2 showed a 1.4-fold higher fluorescence intensity compared to the strain that produced the C-terminal tagged version.
The control culture E. limosum [pMTL83251_P bgaL _ AdhE2] that produced the non-tagged version of AdhE2 showed similar characteristics regarding growth rate and production patterns. Cells were grown with glucose as a carbon source and reached a maximal OD 600 of 2.5 with a growth rate of 0.17 h −1 when adhE2 gene expression was induced. Non-induced cells showed improved growth with a maximal OD 600 of 2.8 and a growth rate of 0.2 h -1 . The acetate:butyrate ratio was 6:1 in the case of the induced and 5:1 in the case of the non-induced cells. Gene expression of adhE2 in an inducible manner resulted in the production of up to 1.9 mM ethanol and 1.6 mM butanol at the end of the exponential growth phase. Non-induced cells only produced traces of ethanol and butanol (0.3 mM) (Additional file 1: Fig. S1A, Table S1, Table 2). Both, the induced as well as the noninduced cells of E. limosum [pMTL83251_P bgaL _AdhE2] were cultivated with methanol and showed a growth rate of 0.05 h -1 with maximal OD 600 values of 1.5 and 1.7, respectively. The acetate:butyrate ratio shifted to 3:1 for the induced and 2:1 for the non-induced cells. Induced  expression of adhE2 caused production of 0.6 mM butanol and traces of ethanol (0.1 mM) in the late exponential growth phase during these batch experiments.

Acetone production with FAST-tagged Adc fusion proteins using recombinant E. limosum strains
The fluorescent FAST tag had no obvious negative effect on the catalytic properties of AdhE2. Subsequently, feg was fused to adc and the respective gene was expressed constitutively, together with the remaining genes of the APO, using the promoter P thlA . Thus, the two strains designated E. limosum [pMTL83251_P thlA _C-FAST-Adc] and E. limosum [pMTL83251_P thlA _N-FAST-Adc] were constructed that produced either the C-or N-terminal FAST-tagged Adc fusion protein, respectively. Those constructed strains were used to perform growth experiments to prove the functionality and productivity of the FAST-tagged Adc fusion proteins.
E. limosum [pMTL83251_P thlA _C-FAST-Adc] and E. limosum [pMTL83251_P thlA _N-FAST-Adc] were cultivated with glucose and showed growth rates of about 0.15 h -1 while reaching maximal OD 600 values of 3.3 and 3.1, respectively ( Fig. 6a and Table 4). Both strains produced acetone, acetate, and butyrate as metabolic end products (Additional file 1: Table S1 and Table 4). E. limosum [pMTL83251_P thlA _C-FAST-Adc] cells that constitutively expressed the APO genes produced 0.8 mM of acetone, while fluorescence intensity was 2.3-fold higher compared to the autofluorescence of the empty vector control strain. E. limosum [pMTL83251_P thlA _N-FAST-Adc] cells that expressed the gene encoding the N-terminal FAST-tagged Adc fusion protein produced only traces of acetone (0.1 mM) and exhibited 1.5-fold higher fluorescence compared to the autofluorescence of the empty vector control strain ( Fig. 6b and Table 4). The control strain E. limosum [pMTL83251_P thlA _act], expressing the non-FAST-tagged native genes of the APO controlled by the P thlA promoter, reached an OD 600 of 3, had a growth rate of 0.15 h −1 , and produced acetate, butyrate as well as  Table 4). The empty vector control E. limosum [pMTL83251] had a maximal growth rate of 0.17 h -1 , reached a maximal OD 600 of 2.5, showed only autofluorescence, and did not produce any acetone.
Cultivation of E. limosum [pMTL83251_P thlA _C-FAST-Adc] and E. limosum [pMTL83251_P thlA _N-FAST-Adc] using methanol as a carbon source resulted in maximal growth rates of 0.08 and 0.07 h -1 , respectively ( Fig. 6c and Table 4). Both cultures reached a maximal OD 600 of 1.6 and produced acetate, butyrate, and acetone (Additional file 1: Table S2 and Table 4). E. limosum [pMTL83251_ P thlA _C-FAST-Adc] produced the highest amount of acetone (1.6 mM) cultivated using methanol, while the strain E. limosum [pMTL83251_P thlA _N-FAST-Adc] expressing the gene encoding N-terminal tagged Adc fusion protein only produced 0.1 mM acetone. Fluorescence intensity of E. limosum [pMTL83251_P thlA _C-FAST-Adc] was fourand three-fold higher compared to the fluorescence of E. limosum [pMTL83251] and E. limosum [pMTL83251_ P thlA _N-FAST-Adc], respectively (Fig. 6d). E. limosum [pMTL83251_P thlA _C-FAST-Adc] cultivated using methanol showed a twofold higher acetone production as well as fluorescence intensity when compared to the cells cultivated using glucose. E. limosum [pMTL83251_ P thlA _N-FAST-Adc] showed minor fluorescence and low amounts of acetone under both conditions tested. E. limosum [pMTL83251_P thlA _act] cultivated using methanol reached an OD 600 of 1.7, had a growth rate of 0.6 h -1 , and produced acetate and butyrate in a ratio of 2:1 and 0.3 mM acetone (Additional file 1: Fig S2B, Table 4). The empty vector control E. limosum [pMTL83251] reached a maximal OD 600 of 1.9 with a growth rate of 0.07 h −1 , did not produce any acetone, and only showed autofluorescence in the presence of TF Lime.

Methanol promotes butyrate production in E. limosum
Since we are aiming to produce butanol and acetone using recombinant E. limosum strains, different carbon sources were investigated as alternatives to commonly used sugars. Therefore, we focused on non-food feedstock-derived C1 carbon sources such as methanol, H 2 + CO 2 , and CO containing syngas and their impact on growth behavior and product spectrum of E. limosum. Cultivation of E. limosum using methanol resulted in higher growth rates compared to cells cultivated with the C1 gases as reported before [43,44]. The C1 carbon sources clearly influenced the product spectrum of E. limosum, since high amounts of methanol (200 mM) resulted in improved butyrate production titers. Methanol is oxidized to CO 2 via the WLP which provides three mol NAD(P)H per methyl group oxidized [45,47]. The NAD(P)H-NAD(P) + balance is regulated during butyrate production since the reactions catalyzed by 3-hydroxybutyryl-CoA dehydrogenase and crotonase are NADH dependent and, therefore, regenerate NAD(P) + . Thus, the acetate:butyrate ratio shifts in favor of butyrate when cells grow on methanol [46]. Moreover, the utilization of methanol yields 1.83 ATP/ butyrate, while H 2 + CO 2 utilization only yields 1 ATP/ butyrate [47]. As shown for C. autoethanogenum and A. woodii mutants, improved ATP supply benefits fast growth and abolishes acetate production [48,49]. In general, cultivation with methanol results in high efficiencies of bioproduction compared to the cultivation with the C1 gases [50,51]. High amounts of methanol (200 mM) had no clear negative impact on the growth of E. limosum. Thus, sustainably produced methanol could be a promising substitute for glucose and C1 gases as substrate. During our experiments, the wildtype strain E. limosum NG-6894 did not produce any butanol. However, E. callanderi KIST612 (former E. limosum KIST612) produces low amounts of butyrate from methanol [52] as well as traces of butanol from CO [53]. Natural butanol production from C1 gases is reported for the methanol-utilizing acetogen Butyribacterium methylotrophicum, which is closely related to E. limosum [54]. In B. methylotrophicum, the production of one mol butanol from methanol requires six reducing equivalents and yields 1.5 ATP/butanol [47].

Establishment of FAST as fluorescent reporter in E. limosum
We established FAST as a fluorescent reporter protein for E. limosum and determined the fluorescence of feg expressing cells using a microplate reader, fluorescence microscopy, and flow cytometry. Lactoseinduced expression of feg caused bright fluorescence of respective E. limosum cells, no matter if fluorescence analytics were carried out under aerobic or anaerobic conditions. Interestingly, the bgaR-P bgaL lactose-inducible system originating from C. perfringens was tightly regulated in E. limosum, since only cells with induced feg expression showed bright fluorescent cells. This tightly regulated expression of the bgaR-P bgaL promoter was also reported for C. perfringens [55] C. acetobutylicum [56], and C. ljungdahlii [17], while gene expression was found to be leaky in A. woodii [49]. Moreover, we also showed constitutive FAST production with the P thlA promoter from C. acetobutylicum. This constitutive expression caused improved fluorescence of cells compared to the lactose-induced feg expression. As of yet, only a very limited number of fluorescent proteins were shown to be somewhat functional in anaerobes, since the fluorescence of the most common fluorescent reporters including GFP and its derivatives is oxygen dependent [57]. Mostly, the flavin mononucleotidebased fluorescent proteins (FbFPs) [58] were used in anaerobes as fluorescent reporters as reported for Clostridium cellulolyticum [26], Clostridioides difficile, Clostridium acetobutylicum, Clostridium sordellii [35], C. ljungdahlii [59], Clostridium beijerinckii [60], and Clostridium tyrobutyricum [61]. Moreover, Shin et al. [13] established an engineered FbFP version (CreiLOV) as a reporter gene in E. limosum ATCC8486. Although FbFPs are functional under anoxic conditions, they show a fairly low brightness compared to control strains that do not express the respective fluorescent protein.
In contrast, the FAST system was recently established in C. acetobutylicum and C. ljungdahlii and showed a clear and obvious fluorescence [31,32,62]. The fluorescence of recombinant E. limosum strains expressing feg constructed in this work is up to 26-fold improved compared to the control strain. Thus, FAST seems to be perfectly suited as a fluorescent reporter for anaerobic bacteria. An alternative to FAST could be the SNAPand the Halo-Tag, which were shown to be functional in C. acetobutylicum and C. ljungdahlii [32]. Both tags show oxygen-independent bright fluorescence when covalently bound to a fluorogenic ligand. However, both tags require a quite long labeling time. Moreover, compared to the 14 kDa of FAST the SNAP-and Halo-Tag are larger proteins with a mass of 19.4 and 33 kDa respectively. In sum, FAST seems to be best suited for the construction of fusion proteins.
As dimerized fluorescent proteins can result in brighter fluorescence [63], we used tdFAST2 to improve fluorescence brightness with E. limosum. Compared to cells producing FAST, Tebo et al. [64] could show a 3.8-fold brighter fluorescence of tdFAST2-producing HEK 293 T cells. Here, we achieved a 1.5-fold improved fluorescence by expressing the codon-optimized gene encoding tdFAST2 in E. limosum confirming the findings of Tebo et al. [64] that production of tdFAST2 improves fluorescence brightness. With approx. 28 kDa, the molecular mass of tdFAST2 is comparable to the mass of EGFP or mCherry [65]. The improved brightness of tdFAST2 also improves the detection limit of tdFAST2-tagged proteins which might be important especially for protein localization experiments [66]. As we only showed a 1.5-fold improved fluorescence of tdFAST2, we chose the less bright but smaller FAST for the construction of AdhE2 and Adc fusion proteins.
Fluorescence microscopy of FAST-producing E. limosum cells revealed that plasmid-based induced expression of feg caused bright fluorescent E. limosum cells showing an inhomogeneous population. This finding was confirmed by respective flow cytometry experiments which revealed that the number of fluorescent cells varies between 72 and 59%. Such a phenotypic heterogeneity of bacterial populations with plasmid-based induced gene expression was previously reported [67][68][69]. One possible reason that might cause these heterogeneous populations is the usage of the lactose-inducible promoter system. The resulting feg gene expression may not only depend on the amount of inducer, but also on the uptake mechanism of the inducer molecule [68]. The lactose-inducible bgaR-P bgaL system was also used in C. perfringens to express the gene yfp-pilB which encodes a yellow fluorescent protein-tagged fusion protein and caused an inhomogeneous fluorescent population [55]. In this experiment, Hartman et al. [55] reported that YFP-PilB was detected in 73% of the cells. Moreover, the inhomogeneous population might be caused by plasmid instability or even the loss of the plasmid encoding the reporter gene [67]. Siebert et al. [69] reported an inhomogeneous population of Oligotropha carboxidovorans which expressed the gene encoding mCherry in a plasmid-based manner. The reported inhomogeneity was overcome by the genomic integration of the respective gene encoding mCherry into the genome of Oligotropha carboxidovorans. Genomic integration of feg into E. limosum should be possible using the CRISPR-Cas9 genome editing tool [13].
Five different fluorogens are commercially available at The Twinkle Factory which can bind FAST and its variants to form the fluorescent FAST:fluorogen complex [33,70]. Expression of feg or the gene encoding tdFAST2 Flaiz et al. Biotechnol Biofuels (2021) 14:117 resulted in green or red fluorescence when TF Lime or TF Coral was used, respectively. As different bacteria and proteins show autofluorescence upon excitation with different wavelength [34,71,72], FAST should be considered as a fluorescent reporter protein in a large variety of bacteria as the choice of the fluorogen can be adapted to the spectral conditions needed for the study. Recently, Tebo et al. [73] reported two engineered FAST versions termed greenFAST and redFAST. While greenFAST binds TF Lime, redFAST only binds TF Coral, which we showed can both be used in E. limosum. These two FAST variants augmented the FAST system and facilitate applications including two-color or even mixed culture imaging in anoxic environments.

C-and N-terminally tagged AdhE2 and Adc fusion protein construction
In this study, we designed AdhE2 and Adc FAST-tagged fusion proteins and expressed respective genes in recombinant E. limosum strains. The X-ray crystal structure of Adc (3BH2) originating from C. acetobutylicum was provided by Ho et al. [74], but to our knowledge there is none for AdhE2 from C. acetobutylicum. Based on the crystal structure, first hints can be derived if either the C-or N-terminal end of the protein should be tagged without causing negative impacts regarding the folding, activity, and productivity of the enzyme. The probability of creating a functional fusion protein was maximized as both the C-and N-terminal tagged versions of both Adc and AdhE2 were constructed [75]. This approach diminishes the risk of constructing non-functional fusion proteins since the interactions of the C-and N-terminal end with other structures within the enzyme are unpredictable, especially for AdhE2. We further used a flexible GS linker as it was shown to promote proper folding of the enzyme and improves the stability of fusion proteins [76]. We cloned the native genes and genes encoding the FAST-tagged fusion proteins without affecting non-coding regions including promoters. As a result, the native genes as well as the genes encoding the FAST-tagged fusion proteins were expressed under the control of the same promoter. We consequently assume that native and tagged genes are transcribed at comparable levels.
The increasing fluorescence of E. limosum [pMTL83251_P bgaL _C-FAST-AdhE2] and E. limosum [pMTL83251_P bgaL _N-FAST-AdhE2] during growth proves the production of the C-and N-terminal tagged AdhE2 fusion proteins after induction of gene expression. Butanol production with strains producing the C-and N-terminal tagged AdhE2 fusion proteins prove that AdhE2 is still functional even with the fluorescent tag. The functionality of the N-terminal tagged version of AdhE2 seems not to be affected since the respective strain produced similar amounts of alcohols as the strain expressing the non-tagged native adhE2 of C. acetobutylicum. However, the productivity of the C-terminal tagged version seems slightly reduced since less alcohol was produced compared to the strains just expressing adhE2. In addition to butanol, ethanol was produced, since AdhE2 converts not only butyryl-CoA to butanol but also acetyl-CoA to ethanol via the respective aldehyde [77]. Interestingly, if cells of both strains were cultivated with glucose as carbon source, fluorescence of the N-and C-terminal tagged fusion proteins was improved compared to cultivation on methanol. This also indicates an improved production of the fusion proteins. Moreover, both strains cultivated on glucose produced more butanol as well as ethanol compared to the cells of the respective strain cultivated using methanol. This finding indicates a coherence between the fluorescence of cells expressing genes encoding adhE2 fusion proteins and the amount of recombinantly produced product.
In the case of the constitutive production of FASTtagged Adc fusion proteins, only the C-terminal FAST-tagged version of Adc produced by E. limosum [pMTL83251_P thlA _C-FAST-Adc] showed clear fluorescence. The strain E. limosum [pMTL83251_P thlA _N-FAST-Adc] producing the N-terminal tagged version of Adc only showed a low fluorescence intensity and acetone productivity. Several factors might affect the fluorescence brightness of fluorescent fusion proteins including intrinsic brightness, folding efficiency, and translation efficiency [78]. Adc itself forms a homododecameric complex with a total mass of 365 kDa [74]. We assume, based on the X-ray structure, that the free C-terminal end of the non-tagged Adc subunits is not interfering with any residues of the protein and, therefore, not harming folding efficiency. However, the N-terminal end of Adc seems to form hydrogen bonds with LYS45, ARG44, VAL47, and GLU49 of the closest subunit. Therefore, the fusion of FAST with the sterically hindered N-terminal end might indeed affect the folding efficiency of the fusion protein or even prevents the formation of the homododecameric complex [79]. In contrast, the folding efficiency of the C-terminal FAST-tagged Adc does not seem to be affected as cells producing the respective fusion protein are fluorescent. Heterologous expression of adc originating from C. acetobutylicum and the FAST-tagged variants with E. limosum might result in inefficient translation regarding their different codon usage [80]. Adapting adc as well as genes encoding respective FAST-tagged Adc fusion proteins to the codon usage of E. limosum might improve translation efficiency causing enhanced protein production and improved fluorescence brightness. Nevertheless, E. limosum [pMTL83251_P thlA _C-FAST-Adc] produced up to 0.8 mM acetone when cultivated with glucose and 1.6 mM from methanol providing evidence that the C-terminal tagged Adc fusion protein is functional and forms the catalytically active homododecameric Adc complex. E. limosum [pMTL83251_P thlA _act] expressing the native genes of the APO produced 0.8 mM and 0.3 mM acetone when cultivated with glucose or methanol, respectively. This finding shows that the productivity of the C-terminal tagged Adc fusion protein is not negatively affected. To our knowledge, acetone was neither produced with E. limosum nor from the sustainable C1 carbon source methanol before. Although acetone production from methanol seems to be promising, several improvements of this recombinant pathway must be performed including codon harmonization, genomic integration, and bioreactor experiments which usually show a beneficial effect on bioproduction.

Conclusions
In this study, we implemented the fluorescent activating and absorption shifting tag as an oxygen-independent fluorescent reporter protein in E. limosum. We showed that this reporter protein is functional under various growth conditions and can be adapted to the natural spectral properties of the organism using different fluorogens. We showed that the fluorescence of feg-expressing cells can be determined using a microplate reader, fluorescence microscopy, or flow cytometry. Butanol, as well as acetone, was produced by expressing the native genes of C. acetobutylicum and genes encoding fluorescent FAST-tagged fusion proteins. The expression of the genes encoding the fluorescent FAST-tagged fusion proteins enabled to monitor protein production during growth. We showed that those fusion proteins are functional and that the productivity is not negatively affected, however, dependent on which end the protein was tagged. The use of FAST as a fluorescence reporter not only expands the molecular toolbox of E. limosum but also is a promising reporter tool for other anaerobic bacteria as well. For the first time, we could show the production of the biocommodities acetone and butanol from methanol with recombinant E. limosum strains.

Plasmid construction
All plasmids and primers used in this study are listed in Tables 6 and 7, respectively. DNA fragments for cloning purposes were amplified using the KAPA Hifi To establish FAST as a reporter protein in E. limosum, plasmids harboring feg were constructed. Therefore, feg was amplified together with the promoter P thl sup from plasmid p95thl sup FAST [31] using primers FW_Pthlsup_ FAST_NdeI and RV_Pthlsup_FAST_XhoI ( Table 7). The amplified PCR fragment was cloned via NdeI and XhoI recognition sites into vector pMTL83251 resulting in the plasmid pMTL83251_P thl sup _FAST. The promoter region of P thl sup from plasmid pMTL83251_P thl sup _FAST was further exchanged with the lactose-inducible bgaR-P bgaL system from C. perfringens [55] and P thlA from C. acetobutylicum [18]. Therefore, the bgaR-PbgaL-encoding DNA fragment was amplified from plasmid pMTL83151_ gusA_P bgaL [49] using the primers FW_PbgaL_NdeI and RV_PbgaL_BamHI. The DNA fragment containing the promoter region of P thlA was amplified from plasmid pJIR750_act [18] using primers FW_Pthl_NdeI and RV_Pthl_BamHI (Table 7). Both promoter containing DNA fragments were assembled with pMTL83251_ P thl sup _FAST (excision of P thl sup ) resulting in plasmids pMTL83251_P thlA _FAST and pMTL83251_P bgaL _FAST.
The gene feg of pMTL83251_P bgaL _FAST was exchanged with the gene encoding tdFAST2. The sequence of tdFAST2 was obtained from Tebo et al. [64] and codon optimized for E. limosum using the GENEius tool (Eurofins Genomics GmbH, Luxemburg) and synthesized by Eurofins Genomics Germany GmbH (Ebersberg, Germany). The gene tdFAST2-opt was amplified from the delivered plasmid pEX-A128_tdFAST2-opt using the primers FW_tdFAST2-opt_BamHI and RV_tdFAST2-opt_XhoI (Table 7) and ligated with BamHI and XhoI digested pMTL83251_P bgaL _FAST (excision of feg) vector DNA resulting in plasmid pMTL83251_P bgaL _tdFAST2.
For butanol production with E. limosum, the plasmid pMTL83251_P bgaL _AdhE2 was constructed. The gene adhE2 was amplified from genomic DNA of C. acetobutylicum using primers FW_C1_adhe2 and RV_ N2_adhE2 ( Table 7). The gene was assembled under the control of the lactose-inducible bgaR-P bgaL promoter using the plasmid DNA of pMTL83251_P bgaL _FAST, which was digested using BamHI and XhoI (excision of feg).
For acetone production with E. limosum, the plasmid pMTL83251_P thlA _act was constructed encoding the genes of the APO under the control of promoter P thlA . The origin of these genes as well as their locus tags are listed in Table 6. These genes as well as the DNA fragment containing the promoter region of P thlA were amplified from plasmids pJIR750_act using primers FW_act_NdeI and RV_act_XhoI ( Table 7). The amplified PCR fragments were cloned via the NdeI and XhoI restriction sites of pMTL83251 resulting in the plasmids pMTL83251_P thlA _act.
Moreover, plasmids were constructed to produce Cand N-terminal FAST-tagged fusion proteins of AdhE2 and Adc. FAST was C-and N-terminal fused to AdhE2 or Adc using a glycine linker (GGGGS) with the sequence ggtggtggtggttct. For the construction of C-terminal tagged fusion proteins, the 3' end of the gene of interest was fused to the 5' end of feg. Therefore, the gene of interest was amplified without its stop codon while feg was amplified without its start codon. To construct N-terminal FAST-tagged fusion proteins, the 3′ end of feg was fused to the 5′ end of the gene of interest. The gene of interest was amplified without its start codon, while feg was amplified without its stop codon.
The plasmid pMTL83251_P thlA _C-FAST-Adc was constructed to produce the C-terminal FAST-tagged Adc fusion protein. adc was amplified from plasmid pMTL83251_P thlA _act using primers FW_C1_adc and RV_C1_adc-FAST, and feg was amplified from plasmid pMTL83251_P bgaL _FAST using primers FW_C2_adc-FAST and RV_C2_FAST (Table 7). Primers RV_C1_adc-FAST and FW_C2_adc-FAST contain the sequence for the glycine linker. The plasmid pMTL83251_P thlA _N-FAST-Adc was constructed to produce the N-terminal FAST-tagged Adc fusion protein. The gene adc was amplified from plasmid pMTL83251_P thlA _act using primers FW_N1_adc-FAST and RV_N1_FAST-adc and feg was amplified from plasmid pMTL83251_P bgaL _FAST using primers FW_N2_FAST-adc and RV_N2_adc (Table 7). Primers RV_N1_FAST-adc and FW_N2_FASTadc contain the sequence for the glycine linker.

Transformation
Electroporation and preparation of electrocompetent E. limosum cells were performed according to the protocol of Leang et al. [84] with several modifications. All steps were carried out under anaerobic conditions in an anaerobic chamber (gas atmosphere 95% N 2 and 5% H 2 ). Plastic materials were placed 24 h before use into the anaerobic chamber to remove any traces of oxygen. For the preparation of electrocompetent cells, E. limosum was cultivated at 37 °C in 100 mL modified DSM 135 medium supplemented with 100 mM methanol and 40 mM DL-Threonine. Cells were cultivated until the early exponential growth phase (OD 600 0.3-0.5) and harvested by centrifugation at 7690×g for 10 min at 4 °C.
Subsequently, cells were washed two times with anaerobic SMP buffer (270 mM sucrose, 1 mM MgCl 2 , 7 mM NaH 2 PO 4 , pH 6) by centrifugation at 7.690×g for 10 min at 4 °C. Afterwards, the cell pellet was suspended in 648 µL SMP buffer and 72 µL DMSO, distributed into cryotubes, and stored at − 80 °C until use.

Fluorescence determination
Fluorescence of FAST and FAST-tagged fusion proteinproducing recombinant E. limosum strains, cultivated anaerobically until the stationary growth phase, were determined using a microplate reader, fluorescence microscopy, or flow cytometry. Therefore, 2 mL culture broth was taken anaerobically during growth and   14:117 harvested by centrifugation at 7711×g for 15 minutes at 4°C. The supernatant was discarded and harvested cells were washed with anaerobic PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 ) followed by centrifugation at 7711×g for 15 minutes at 4°C. Cell pellets were suspended in anaerobic PBS buffer (end OD 600 1).

Microplate reader
100 µL of suspended cells was transferred to black flatbottomed 96-well microtiter plates (Greiner Bio-One GmbH, Frickenhausen, Deutschland) and supplemented with 10 µM fluorogen of the green dye TF Lime (ex. 480/ em. 541) or the red dye TF Coral (ex. 516/em. 600) (The Twinkle Factory, France, Paris). Excitation and emission wavelengths both correspond to the respective maximum of the fluorogen. Fluorescence intensities of the whole population were determined using the SYNERGY H1 microplate reader (BioTek, Bad Friedrichshall, Germany) located in an anaerobic chamber and finally normalized to the OD 600 of PBS-washed cells.

Fluorescence microscopy
Washed cells were stained with 10 µM TF Lime and transferred onto an 1% agarose pad on a microscopy slide, covered with a glass coverslip, and sealed with nail polish. Cells were viewed with a 63x objective, using a Zeiss Axio Observer Z1 (Zeiss, Oberkochen, Germany). Green fluorescence of the FAST:fluorogen complex was detected using the Zeiss filter set 38 HE (ex. BP 470/40, em. BP 525/50). Data were analyzed using the Zeiss Zen 2.6 blue edition software.

Flow cytometry
Washed cells were stained with 10 µM TF Lime diluted in PBS buffer (end OD 600 : 0.01). Green fluorescence of the FAST:fluorogen complex was assessed using an excitation wavelength of 488 nm and a 528/46 nm emission filter. For analysis, at least 10,000 events were recorded using an Amnis ® CellStream ® flow cytometer (Luminex Corporation, Austin, TX, USA). Acquired data were analyzed using the CellStreamTM Analysis tool version 1.2.152 (Luminex Corporation, Austin, TX, USA).

Analytics
2 mL samples were withdrawn from cell cultures during the growth of E. limosum to analyze OD 600 , product spectrum as well as substrate consumption. After OD 600 of the withdrawn culture broth was determined at 600 nm using the GENESYS 10vis photometer (Thermo Scientific, Waltham, MA, USA), the remaining culture was centrifuged at 17,968×g for 20 minutes at 4 °C to remove cell debris. The supernatant was used for HPLC and GC analysis.

High-performance liquid chromatography
The concentration of glucose, acetate, and butyrate from culture supernatant was determined using the Agilent 1260 Infinity II HPLC system (Agilent Technologies, Santa Clara, CA, USA), equipped with a diode array detector and a refractive index detector. To achieve separation, 20 μL of the supernatant was injected into the organic acid resin 150 x 8 mm column (CS-chromatographie-Service GmbH, Langerwehe, Germany) packed with polystyrene divinylbenzene copolymer operating at a constant temperature of 40 °C. As mobile phase, 5 mM H 2 SO 4 was used with a flow rate of 0.7 mL min -1 . The software OpenLAB CDS ChemStation Edition A.01.03 (Agilent Technologies, Santa Clara, CA, USA) was used for data analysis.

Gas chromatography
The concentration of methanol, ethanol, butanol, and acetone in the culture broth was analyzed via gas chromatography. A PerkinElmer Clarus 680 GC system (Perkin Elmer LAS GmbH, Waltham, MA, USA) equipped with an Elite-FFAP capillary column (length 30 m x inner diameter 0.32 mm, film thickness 0.25 µm) (Perkin Elmer LAS GmbH, Waltham, MA, USA) and FID detector was used. Supernatants were acidified with 2 M HCL. H 2 was used as the carrier gas. The injector and detector were operated at 225 and 300 °C, respectively. 0.5 µL of the culture broth was injected into the gas chromatograph and analyzed using the following temperature profile: 40 °C for 2.5 minutes; 40 °C-250 °C with 40 °C min -1 ; 250 °C for 2 minutes. and acetone production. Error bars indicate standard deviations. n = 3. Table S1. Growth characteristics and product formation of recombinant E. limosum strains characterized in growth experiments using glucose as carbon source. Table S2. Growth characteristics and product formation of recombinant E. limosum strains characterized in growth experiments using methanol as carbon source.