A growth selection system for sucrose synthases (SuSy): design and test

High throughput screening (HTS) methods of enzyme variants are essential for the development of robust biocatalysts suited for low impact, industrial scale, biobased synthesis of a myriad of compounds. However, for the majority of enzyme classes, current screening methods have limited throughput, or need expensive substrates in combination with sophisticated setups. Here, we present a straightforward, high throughput selection system that couples sucrose synthase activity to growth. Enabling high throughput screening of this enzyme class holds the potential to facilitate the creation of robust variants, which in turn can significantly impact the future of cost effective industrial glycosylation. Supplementary Information The online version contains supplementary material available at 10.1186/s13568-024-01727-y.


Introduction
Glycosides are widely distributed in nature and can be found in nearly every living organism (Lis and Sharon 1993).In plants, glycosides are part of the secondary metabolism and play a role in different processes like accumulation, storage, and transport of hydrophobic substances, as precursors for flower fragrances (Watanabe et al. 1993) and UV radiation protection (Hofmann 2000).Glycosylation of lipophilic low molecular weight compounds is an efficient tool to reduce chemical toxicity, enhance water solubility and improve stability (Fu et al. 2019), and hence positively affect shelf life and bioavailability of various chemicals and pharmaceuticals (Bowles et al. 2006;Seeberger et al. 2022).Additionally, glycosides themselves have industrial applications, e.g. in detergents and cosmetics (de Roode et al. 2003), as antimicrobials of sucrose from fructose and NDP-glucose, allow for the recycling of the NDP-sugar donor by pushing the thermodynamic equilibrium, using abundant and cheap sucrose.SuSys have already been applied at lab scale together with different UGTs in one pot reaction systems to produce a variety of glycosides, including steviol glycosides (Chen et al. 2018), glycyrrhetinic acid glucoside (Ali et al. 2021), polydatin (Chen et al. 2021), nothofagin (Putkaradze et al. 2023), indican (Bidart et al. 2024), and orientin and vitexin (Gu et al. 2021).
In order to achieve the high titers required for glycosylation reactions at scale, organic solvents are usually applied to help dissolve the hydrophobic substrates (Pei et al. 2019;Tao et al. 2023).In this context, SuSy stability seems to be challenged, limiting further industrial implementation (Diricks 2017;Liu et al. 2022;Orrego et al. 2017;Trobo-Maseda et al. 2020;Zhao et al. 2023).In recent years, multiple studies have focused on improvement of SuSy stability using different approaches such as rational (Diricks 2017) and semi rational (Zhao et al. 2023) enzyme engineering, immobilization (Liu et al. 2022;Orrego et al. 2017;Trobo-Maseda et al. 2020), and enzyme mining (Chen et al. 2023), to find a robust variant suitable for industrial application.So far, enzyme engineering has successfully led to variants with increased kinetic stability, providing longer half lives, however this did not translate in higher thermo or chemostability (Zhao et al. 2023).Enzyme immobilization was also successful at increasing the stability compared to free enzyme, although with a concomitant activity loss (Liu et al. 2022;Orrego et al. 2017;Trobo-Maseda et al. 2020).Lastly, with the rapid increase in DNA sequence data, enzyme mining has emerged as an attractive approach to find robust variants.This is the case of the SuSy identified in the green algae Micractinium conductrix (McSuSy).
McSuSy showed an attractive DMSO tolerance, however its kinetic stability is strongly reduced at temperatures above 50 °C (Chen et al. 2023).
In vivo growth couple selection systems are a versatile methodology for directed evolution of enzymes (Luo et al. 2019;Neuenschwander et al. 2007;Wu et al. 2022).By coupling the activity of a targeted enzyme with growth of a host organism, in vivo selection allows simple and robust identification of active variants within a large library.Although several pathways for microbial sucrose utilization are known, and despite the industrial relevance of sucrose utilizing enzymes (as SuSy), in vivo growth coupled selection systems for engineering of these industrial biocatalyst have never been demonstrated.In this work we report a generally applicable growth coupled selection system for the engineering of robust SuSy variants.

Bacterial strains, medium and growth conditions
The strains used in this study are listed in Table 1.E. coli MGcscBKA strain and derivatives were used in this study.Genome engineering of E. coli MGcscBKA was performed by CRISPR/MAD7 method as described previously (Phaneuf et al. 2021).E. coli DH5α™ subcloning efficiency™ (ThermoFisher Scientific, Germany), was used as host in cloning experiments.All strains were mantained in Luria-Bertani (LB) medium at 37 °C under agitation, and transformants were selected with Chloramphenicol (25 µg mL −1 ).Growth curves were done on M9 minimal medium following the protocol described below.

Plasmids
The plasmids used in this study are listed in Table 1.DNA fragments containing the Sucrose Synthases genes were obtained by PCR using pET28a + plasmids encoding the targeted enzymes (purchased from Genscript, United States) as templates and the corresponding oligonucleotide pairs.The PCR products were purified using a NucleoSpin Gel and PCR kit (Macherey-Nagel, Germany) and cloned into the backbone of pMTP7 (obtained by PCR using primers pMPT7-Fw and pMPT7-Rv) by USER cloning (New England Biolabs, US).Chemically competents E. coli DH5α™ were transformed by heat shock with 3 µl of the USER reactions and after 1 h recovery cells were plated on LB with Chloramphenicol.The resulting plasmids were purified using a NucleoSpin

Metabolic model based analysis
The growth coupling design for SuSy was confirmed and analyzed using iML1515 (Monk et al. 2017), a comprehensively curated genome scale metabolic model for E. coli K-12 MG1655.ADP, UDP dependent variants of SuSy, and the proton symporting sucrose permease were added to iML1515 to enable the uptake and metabolization of sucrose.To disable trehalose-6-phosphate synthase activity (UDP recycling), thus resembling the genetic background of the strain SDT365, otsA (b1896) was knocked out in the model.Note that the gene cscA and its associated metabolic reactions are not present in iML1515.Flux balance analyses were conducted assuming a maximum sucrose and oxygen uptake rate of 7.5 mmol gDW −1 h −1 and 20 mmol gDW −1 h −1 , respectively.Model handling and all simulations were performed using the COBRApy toolbox (Ebrahim et al. 2013) (version 0.26.2) and the Gurobi optimization suite (Gurobi-Optimization 2023) (version 9.0.2).A Jupyter notebook including all relevant in silico analyses is provided in the Supplementary information.

Genome engineering of parental strain MGcscBKA
Simultaneous knockout of cscA and otsA from sucrose utilizing MG1655 derived strain, EMP4, was achieved by CRISPR/MAD7 method as described in a previous study (Phaneuf et al. 2021) with a slight modification.gRNA plasmid containing both cscA and otsA targeting gRNA sequence (pSD85) was constructed by USER cloning (Cavaleiro et al. 2015) for multiplex gene editing.gRNA plasmid backbone was amplified by a pair of primers, SD_PR356 and SD_PR357 (Supplemental Table 1), to achieve the array 1 structure applied from Lin et al. (Lin et al. 2021).We followed the same procedure for genome engineering as the previous study, except that 100 pmol of two MAGE oligos (SD_PR361 and SD_PR364) for knocking out the cscA and otsA genes were mixed before being added.The knockout of genes was validated through colony PCR using flanking primers for each gene.

Design of a growth coupled selection system for SuSy variants
SuSy catalyzes the cleavage of sucrose into fructose and ADP/UDP-glucose in the presence of either NDP.To design a growth coupled selection system based on SuSy activity, we leveraged its byproduct fructose as the sole carbon source for microbial growth.Moreover, we took advantage of the previously optimized E. coli MGcscBKA strain for improved sucrose utilization (Mohamed et al. 2019), which contains genes for a sucrose permease (cscB) to allow the sucrose uptake, an invertase (cscA) to allow sucrose hydrolysis, and the fructose kinase (cscK) which allows fructose to directly enter glycolysis via fructose-6P (Fig. 1).We then knocked out cscA to avoid the utilization of sucrose as carbon source (Fig. 1).Therefore, when sucrose is supplied as the only carbon source in a minimal medium, only cells harboring an active sucrose utilizing enzyme can survive and grow on either glucose, fructose, or both, resulting from the sucrose utilizing reaction.
The design and growth rescue by SuSy activity was computationally validated using a genome scale metabolic model.The model confirms that SuSy restores growth when sucrose is provided as the sole carbon source, even without considering feeding ADP or UDP.If an extracellular ADP or UDP feed is added to the model and made solely available to SuSy, growth rates increase by 34% and 30%, respectively.These growth advantages demonstrate the benefit to the cell of using ADP/UDP recovered from ADP/UDP-glucose in metabolic processes other than SuSy activity.Furthermore, the model predicts a slightly higher growth rate (+ 6%) for an ADP specific SuSy in contrast to an UDP specific variant, since recycling of ADP from ADP-glucose by glycogen synthase (glgA) inflicts less metabolic burden on the cell than UDP recycling by the glucose-6-phosphate dependent trehalose-6-posphate synthase (otsA).
To prevent the synthesis of trehalose, a storage disaccharide that would otherwise act as a carbon sink in our selection scheme, the otsA gene was knocked out leading to strain SDT 365.

Growth rescue by SuSy activity
We then proceeded to experimental validation.Growth curves of strain SDT 365 in M9 medium supplemented with sucrose confirmed that this strain is not able to grow on sucrose without harboring a complementary sucrose utilizing enzyme (Fig. 2A).Furthermore, the strain was neither able to grow on M9 medium supplemented with both sucrose and UDP, or sucrose and ADP (Fig. 2A).To test growth rescue by SuSy activity, SDT 365 was transformed with three vectors, each carrying a different SuSy variant under a constitutive strong promoter.These were the plant SuSy from Glycine max (GmSuSy), the most frequently used SuSy so far, which has the strongest affinity towards UDP (K M = 5 µM (Diricks et al. 2016), the bacterial SuSy from Acidithiobacillus caldus (AcSuSy) which is more thermostable than plant SuSys, and has higher affinity towards ADP (K M = 0.3 mM) compared with UDP (K M = 7.8 mM) (Diricks et al. 2016); and a previously reported engineered variant of AcSuSy (L637M-T640V) with increased affinity towards UDP (K M = 0,13 mM) (Diricks et al. 2016), giving rise to the strains EESB 4, EESB 5 and EESB 6. Cultures of EESB 4 and EESB 5 showed no noticeable growth on a medium with only sucrose (Fig. 2B-C).Both strains partially restored growth when further supplemented with UDP or ADP (Fig. 2B-C), respectively, in accordance with their individual substrate preferences.In line with reported substrate affinities, strain EESB 6, which carry the engineered variant of AcSuSy with high affinity towards both ADP and UDP, exhibited significant growth when the medium was supplemented with sucrose and either of the two NDPs (Fig. 2D).Curiosly, this strain also exhibited some growth when the medium was supplemented only with sucrose.This indicates that intracellular provision of NDP is sufficient to allow for activity of this enzyme variant, which confirms model based observations.(Fig. 2D).

Discussion
Enzyme engineering for the development of industrially suitable biocatalysts entails complex improvements of activity, stability, and other desired features.Even with HTS methods available, the process can become tedious when multiple substitutions are needed to obtain the targeted properties, leading to a combinatorial landscape that can quickly exceed the experimentally possible.In vivo selection systems are generally regarded as simple and low tech methodology to screen large librarys, which size are only limited by transformation efficiency (Neuenschwander et al. 2007;Wu et al. 2022).However, developing effective growth coupled selection systems can be challenging, as the link between enzyme kinetic activity and growth must be carefully established in order to discriminate enhanced variants from the starting enzyme during several rounds of directed evolution.Different tools have been developed to increase selection pressure and allow a wider dynamic range of in vivo selection systems, such as the use of regulable promoters, a variety of ribosomal binding sites, and even protein degradation tags (Nearmnala et al. 2021;Neuenschwander et al. 2007;Wu et al. 2022).These tools remain to be further implemented in our selection platform in order to be able to enhance SuSy activity by directed evolution selecting for faster growth rates.
The challenge to use in vivo selection systems becomes especially difficult when aiming to improve enzyme stability for its use in in vitro biocatalysis, given that reaction conditions (e.g., high substrate load, presence of organic solvent, pH, and temperature) are often impossible to mimic in an intracellular environment.However, the development of computational approaches fed with big data generated from growth coupled platforms, such as global multi mutant analysis (GMMA) (Johansson et al. 2023) opens the door for such endeavours.GMMA relys on the fact that beneficial mutations, even if they do not result in a discernible phenotype, can be identified by their ability to compensate for deleterious mutations, which do have a phenotype (Norrild et al. 2022).By integrating the information of phenotype (in this case, growth/no growth) and genotype (mutations in the gene) of a large library of multiple mutated variants, it would be possible to assign the effects of individual mutations.
SuSys hold potential to become an industrial biocatalysts when coupled with UGTs for the production of valuable glycosides with diverse uses and bioactive properties.However, lack of a robust variant, ideally in the form of high kinetic stability in the prescense of organic solvents and high aglycon loads, prevent its widespread application.Moreover, lack of efficient HTS methods for this enzyme class hinders the engineering efforts towards desired properties, limiting tests to a reduced number of rational designs.Our work presents a solution to overcome these limitations.The proposed procedure for the stability engineering of any desired SuSy using our strain platform and GMMA is summarized in Supplemental Fig. 1.
TDP activated sugars are the most structurally diverse class of nucleotide sugars found in nature, and the preferred option in the biosynthesis of most of the bacterial glycosylated natural products (Thibodeaux et al. 2008).In fact, the Bacillus licheniformis (UDP) glycosyltransferase YjiC has already been engineered to increase specificity towards TDP-glucose, as it has been hypothetised that microbial production of glycosides would be favoured given that TDP-sugars are abundantly present in microbial cytosol (Cho et al. 2019).In this context, a SuSy with specificity towards TDP could play an important role for the recycling of sugar donors within microbial cell factories for in vivo production of glycosides, as currently proosed in the recombinant production of steviol glycosides (Dalgaard Mikkelsen et al. 2012).The growth coupled selection platform presented in this work is readily applicable for the engineering of SuSy NDP specificity, by providing an alternative NDP in the culture medium.
In summary, our selection design works as a flexible platform that can be used for the engineering of different features of SuSy, and other industrially relevant sucrose utilizing enzymes such as invertases (involved in the production of high fructose syrup), or β-Fructofuranosidase (for fucosyl-oligosaccharides synthesis).

Fig. 2
Fig. 2 Growth curves of E. coli SDT 365 (without plasmid) (A), EESB 4 (B), EESB 5 (C), and EESB 6 (D) on M9 minimal medium supplemented with sucrose, and with or without (control) supplemented UDP/ADP.Data presented are average values based on at least three replicates.Error bars indicate standard deviations

Table 1
Strains and plasmids used in this studyPlasmid kit (Macherey-Nagel, Germany), and its sequence verified (Eurofin Genomics).E. coli SDT 365 electrocompetents were transformed with each of these plasmids using a single electric pulse in a BioRad Micro-Pulser as indicated by the manufacturer (program Ec2; 10 mF, 600 Ω, 12,5 kV/cm).