Development of a synthetic transcription factor-based S-adenosylmethionine biosensor in Saccharomyces cerevisiae

S-Adenosylmethionine (SAM) is a crucial small-molecule metabolite widely used in food and medicine. The development of high-throughput biosensors for SAM biosynthesis can significantly improve the titer of SAM. This paper constructed a synthetic transcription factor (TF)-based biosensor for SAM detecting in Saccharomyces cerevisiae. The synthetic TF, named MetJ-hER-VP16, consists of an Escherichia coli-derived DNA-binding domain MetJ, GS linker, the human estrogen receptor binding domain hER, and the viral activation domain VP16. The synthetic biosensor is capable of sensing SAM in a dose-dependent manner with fluorescence as the output. Additionally, it is tightly regulated by the inducer SAM and β-estradiol, which means that the fluorescence output is only available when both are present together. The synthetic SAM biosensor could potentially be applied for high-throughput metabolic engineering and is expected to improve SAM production.


Introduction
S-Adenosylmethionine (SAM) is a crucial small-molecule metabolite that serves as a methyl group donor to modify biomacromolecules (Gregoire et al. 2016). It is widely used for treating various diseases, such as affective disorders, liver disease, and osteoarthritis (De Berardis et al. 2016;Guo et al. 2015;Silveri et al. 2003). Recently, SAM alone or in combination with other drugs is emerging as a potentially effective strategy for cancer treatment and chemoprophylaxis. In addition, SAM as a dietary supplement aroused great interest and has enormous commercial value (Silveri et al. 2003;Yan et al. 2021). In the past few years, great efforts have been made to enhance the yield of SAM (Chen and Tan 2018;Kanai et al. 2017;Qin et al. 2020). However, for industrial production of SAM, titers and yields of SAM production by engineered yeast must be further improved Ruan et al. 2019).
Recent advances in the iterative "design-buildtest-learn" cycle framework of synthetic biology have enabled rapid engineering of microbial cell factories, leading to pronounced improvements in library design and construction. Libraries that generate genotypic diversity can be achieved by 1 3 Vol:. (1234567890) genome engineering techniques, such as automated multiplex genome engineering (MAGE) (Menegon et al. 2022;Si et al. 2017), global transcription machinery engineering (gTME) (Deng et al. 2022;Ke et al. 2020), multiscale analysis of library enrichment (SCALEs) (Feng et al. 2021), and random mutations of chemical mutagens or ultraviolet radiation. High performance liquid chromatography (HPLC) technique can be used to confirm and quantify SAM concentration. However, this analytical method is time-consuming and labor-intensive if lack of a convenient high-throughput screening technique. Thus, it is imperative to develop a convenient method that can easily distinguish microbial phenotypes.
To help circumvent this hurdle, synthetic biology offers several valuable approaches. Among them, biosensors have been widely applied in metabolic engineering with success in various microorganisms (Zadran et al. 2013). A broad spectrum of metabolites or other chemicals can be detected by sensing devices, such as riboswitch, transcription factor (TF), specific enzymes and other sensors (Chen et al. 2022;Feng et al. 2018). However, the TF-based biosensors remain most convenient to be designed and have been successfully applied to detect fatty acids, amino acids, and other metabolites in living cells Sun et al. 2020;Wang et al. 2020Wang et al. , 2019. The TFbased biosensors could precisely translate intracellular concentrations of target compounds into a graded fluorescence output by driving expression of fluorescent proteins (Kaczmarek and Prather 2021). When combined with FACS, biosensors have the potential to change the way microbial production strains are engineered in the future (Castano-Cerezo et al. 2020;Han et al. 2022).
In this work, we established a novel synthetic TF-based biosensor for monitoring SAM levels in Saccharomyces cerevisiae. The synthetic TF was composed of a bacterial DNA-binding domain MetJ, GS linker, the human estrogen receptor binding domain hER and the activation domain VP16. Accordingly, the recombinant promoter consists of promoter Cyc1 core region and the metO region that could be recognized by the MetJ. Then, the effect of β-estradiol and SAM on cell fluorescence intensity was evaluated to verify the performance of the constructed biosensor.

Materials and methods
Strains, media, and culture conditions All strains used in this study are described in Supplementary Table S1. E. coli strain Top 10 bought from Biomed was used for cloning host. The transformed E. coli were screened on Luria-Bertani (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl) medium supplemented with 100 μg/mL ampicillin at 37℃ under 200 rpm.
Saccharomyces cerevisiae BY4741 was collected and used as the parent strain for genetic manipulation. For selection of yeast transformants, the yeast was cultivated in 50 mL synthetic complete medium (6.7 g/L yeast nitrogen base with ammonium sulfate and without amino acids, 1.4 g/L amino acids dropout mixture lacking histidine) supplemented with amino acids as required, and 20 g/L glucose. The recombinant yeast cells were incubated in 50 mL YPD medium at 30 ℃ and 220 rpm for 24 h. The calculated amounts of pre-cultures were transferred to 100 mL YPD medium to ensure that the initial OD 600 was 0.1. Then the inducers of L-methionine and β-estradiol were added to the medium after 6 h fermentation, and the cultures were incubated at 30 ℃ and 220 rpm for analysis.

Plasmid construction
All primers used were designed using the NEBuilder assembly tool of the NEB website (http:// nebui lder. neb. com/#!/). The primers are listed in Supplementary Table S2. The constructed synthetic TF consists of metJ, GS linker, her, and VP16 (the exact coding sequence of the genes and the recombinant plasmid profile can be found in the Supplementary Data). The gene metJ containing GS linker was amplified by PCR using the genomic DNA of E. coli as the template. The gene her was amplified by PCR using the cDNA template presented by professor JiaHuai Huang. The gene VP16 was synthesized by Sangon Biotech Company. The promoter TEF2 and the terminator TPI were amplified by PCR using the genomic DNA of S. cerevisiae as the template. The fragments of the metJ-GS linker, her and VP16, were firstly assembled using the Gibson Assembly method (Gibson et al. 2009). Next, the fragments of metJ-GS-her-VP16, promoter TEF2, and terminator TPI were overlapped by the Gibson Assembly. At last, the obtained fragments TEF2-metJ-GS-her-VP16-TPIt and plasmid pRS403 were double digested by SalI and EcoRV to generate the plasmid pRS403-ATF.
The recombinant promoter based on the Cyc1 core promoter sequence from S. cerevisiae and four natural metbox conserved sequences from metA in E. coli was synthesized by Wuhan Gene Create Biological Engineering Company (the exact coding sequence of the genes can be found in the Supplementary Data). The egfp-Cyc1t fragment was amplified by PCR using plasmid of pETduet-eGFP (lab collection) as the template. The two fragments were assembled using the Gibson Assembly method. Then the overlapped fragments were double digested by SacII and SpeI, and subcloned into the plasmid pRS403-ATF to yield pRS403-ATF-PG. The recombinant plasmid was transformed into S. cerevisiae BY4741 according to the LiAc/SS carrier DNA/PEG method (Gietz and Schiestl 2007). Finally, the recombinant strain BSS was obtained using His3 in pRS403 as an integrated homologous arm and a selection marker.

Analytical procedures
Cell density was estimated by measurement of the turbidity of the culture samples at 600 nm using a spectrophotometer (Thermo Scientific, Waltham, MA). The cells were lysed with perchloric acid and centrifuged for subsequent SAM detection (Chen et al. 2015). The supernatant was used for HPLC determination (Shimadzu, Japan) according to the previous study (Chen et al. 2015). The fluorescence intensity was measured by a flow cytometer (BD, USA) after 16 h fermentation (Chen et al. 2017).

Results and discussion
Design of the SAM biosensor Development of biosensors for different metabolites is crucial for construction of high-yield strains. The biosensors could link target metabolite concentration to an output, enabling high-throughput screening via powerful selection tools such as flow cytometer (Qiu et al. 2019). In general, construction of TF-based biosensors requires two primary modules: detection module and effect module (Wan et al. 2019). The detection module is mainly composed of a promoter and a DNA-binding TF. The effect module consists of an output signal expression element, such as the complex of green fluorescent protein and recombinant promoter with specific TF binding sequences.
The architecture chosen for SAM biosensor was inspired by Umeyama, in which the synthetic TF named MetJ-B42 was developed (Umeyama et al. 2013). There are some problems with their system. Firstly, addition of the antibiotic doxycycline may interfere with the normal physiological state of the cell. Secondly, the nuclear localization signal peptide sequence needs to be added in front of the promoter of MetJ-B42 to guide the TF into the nucleus. This approach will undoubtedly affect the efficiency of gene regulation. Therefore, we design and construct a novel synthetic TF that comprises the bacterial DNAbinding domain MetJ, GS linker, the human estrogen receptor binding domain hER and an activation domain VP16.
MetJ is a transcription regulator that regulates the expression of genes involved in methionine biosynthesis and transport in E. coli. MetJ can be recognized by the specified sequences from the promoters, which contain tandem repeats of 2 to 5 MetJ binding sites (also called metboxes). The MetJ binding sites usually contain a consensus sequence AGA CGT CT (Augustus et al. 2010). When MetJ binds to SAM, the formed complex binds to the metboxes, thus shutting down the transcription of downstream genes (Augustus et al. 2009;Marti-Arbona et al. 2012). The GS linker was introduced between MetJ and hER to provide flexible conformation for the fusion protein.
VP16 is the 367-490 aa fragment of the herpes simplex virus type I trans-activator. It can activate transcriptional initiation of many eukaryotic genes and thus is widely used in eukaryotic biosensor models (McIsaac et al. 2013). The human estrogen receptor binding domain hER is an active allosteric switch. In the absence of inducer β-estradiol, hER interacts with the Hsp90 chaperone complex, sequestering the MetJ-GS-hER-VP16 to the cytoplasm. The introduction of β-estradiol displaces Hsp90, revealing a nuclear localization signal, and the MetJ-GS-hER-VP16 translocates to the nucleus to initiate gene regulation (McIsaac et al. 2013). In general, the antibiotic tetracycline or its derivative doxycycline is an often-used input. However, β-estradiol is an exciting novel input for regulating heterologous TF, as its addition ensures precise regulation of hER binding. In addition, β-estradiol did not disturb the growth of yeast cells as reported in the literature (McIsaac et al. 2013).
As illustrated in Fig. 1, the synthetic biosensor also contains an output element based on the fluorescence. The enhanced green fluorescent protein (eGFP) was inserted into downstream of the recombinant promoter as the output. The recombinant promoter consists of a Cyc1 core promoter and four natural metboxes containing MetJ binding sites in the metA operator (The full length of the recombinant promoter can be seen in Supplementary data), and pairs with the synthetic TF (MetJ-GS-hER-VP16). β-estradiol and SAM as the dual-input elements can strictly regulate the activity of chimeric MetJ-GS-hER-VP16 in the biosensor. When both SAM and β-estradiol are present in the medium, MetJ-GS-hER-VP16 combined with SAM to form a holo complex. The complex can then bind to specific sequences in the recombinant promoter, thereby activating the eGFP expression and generating the fluorescent signal (Fig. 1).
To confirm the synthetic biosensor biochemically, we measured the intracellular fluorescence intensity after 16 h fermentation of the recombinant strain BSS. This dual-input biosensor was expected to act as an "AND gate", outputting fluorescence only when both SAM and β-estradiol are present in the medium. It is well known that addition of L-methionine, an essential precursor of SAM, can significantly improve the biosynthesis of intracellular SAM. Therefore, 6 g/L L-methionine and 10 nM β-estradiol were added to the medium to induce the expression of the reporter gene egfp. As shown in Fig. 2, a very low fluorescence intensity of the cells equipped with the synthetic biosensor was observed in a medium lacking β-estradiol. Interestingly, the cells also exhibited higher intensity of fluorescence in the medium containing β-estradiol but without L-methionine addition. This is because that yeast cells could produce a certain concentration of SAM without addition of L-methionine in the medium. As expected, the cells showed the highest fluorescence intensity in the presence of both L-methionine and β-estradiol. The above results can preliminarily prove that the output of fluorescence intensity is affected by SAM and β-estradiol.
We also expected to modulate the output strength by changing concentrations of SAM and β-estradiol. To test this idea, we firstly added 6 g/L L-methionine and different concentrations of β-estradiol to the medium after 6 h fermentation. As expected, the cells failed to fluoresce without the addition of β-estradiol. The fluorescence intensity increased with increasing β-estradiol concentration. However, when the β-estradiol concentration increased from 100 nM to 1 μM, the fluorescence intensity did not increase significantly (Fig. 3).
We next examined whether the changes in β-estradiol concentrations will affect cell growth. We compared cell growth between the control and the biosensor containing strains (BSS strain) in the presence of different amounts of β-estradiol. During the cell growth cycle, the growth of BSS strain was slightly lower than that of control strain ( Fig. 4A and B). The cell growth was significantly inhibited when the concentration of β-estradiol reached 1 μM. However, β-estradiol at 10 nM or 100 nM has no negative impact on the growth of either the control strain or the biosensor containing strain (Fig. 4B).
Next, different concentrations of L-methionine (0 g/L, 3 g/L, 6 g/L, 9 g/L) were added to the medium to adjust the concentrations of SAM. After 6 h fermentation, 100 nM β-estradiol and various concentrations of L-methionine were added to the medium. Then the fluorescence intensity and the intracellular SAM concentration were detected after 16 h fermentation. As shown in Fig. 5, the addition of L-methionine had a significant effect on the fluorescence intensity. The fluorescence intensity and the SAM titer were both increased with increasing L-methionine concentration. When the L-methionine concentration reached 9 g/L, the fluorescence intensity and SAM concentration were the highest. Furthermore, the fluorescence intensity of the strain containing SAM biosensor is positively correlated with the concentration of SAM. In conclusion, the synthetic SAM biosensor could be used to monitor SAM levels in living S. cerevisiae. In the future, cells with high SAM level can be sorted by flow cytometry, which will facilitate the construction of high-yield SAM strains.

Conclusions
Biosensors have shown enormous potential in metabolic engineering applications for fast highthroughput screening of high-yield strains (Kaczmarek and Prather 2021; Qiu et al. 2020;Zhang and Shi 2021). Here we design and construct a novel genetically encoded biosensor for monitoring SAM level in living S. cerevisiae. The constructed biosensor comprises bacterial DNA-binding protein MetJ, GS linker, human estrogen receptor binding domain hER and activation domain VP16. The synthetic biosensor is integrated into the chromosomes of S. cerevisiae without affecting the normal cell growth. Unlike other SAM biosensors, the regulation of the biosensor we constructed is a dual-input mode, which means that the output of fluorescence intensity is controlled by adjusting the concentrations of two inputs, L-methionine and β-estradiol. In addition, the fluorescence intensity of the biosensor is SAM dose-dependent under certain conditions. Overall, we believe that the developed synthetic biosensor can be further applied as a high-throughput metabolic engineering tool to optimize the SAM synthesis pathway and ultimately obtain the high-yield SAM strains.