Novel dTDP-l-Rhamnose Synthetic Enzymes (RmlABCD) From Saccharothrix syringae CGMCC 4.1716 for One-Pot Four-Enzyme Synthesis of dTDP-l-Rhamnose

Deoxythymidine diphospho-l-rhamnose (dTDP-l-rhamnose) is used by prokaryotic rhamnosyltransferases as the glycosyl donor for the synthesis of rhamnose-containing polysaccharides and compounds that have potential in pharmaceutical development, so its efficient synthesis has attracted much attention. In this study, we successfully cloned four putative dTDP-l-rhamnose synthesis genes Ss-rmlABCD from Saccharothrix syringae CGMCC 4.1716 and expressed them in Escherichia coli. The recombinant enzymes, Ss-RmlA (glucose-1-phosphate thymidylyltransferase), Ss-RmlB (dTDP-d-glucose 4,6-dehydratase), Ss-RmlC (dTDP-4-keto-6-deoxy-glucose 3,5-epimerase), and Ss-RmlD (dTDP-4-keto-rhamnose reductase), were confirmed to catalyze the sequential formation of dTDP-l-rhamnose from deoxythymidine triphosphate (dTTP) and glucose-1-phosphate (Glc-1-P). Ss-RmlA showed maximal enzyme activity at 37°C and pH 9.0 with 2.5mMMg2+, and the Km and kcat values for dTTP and Glc-1-P were 49.56μM and 5.39s−1, and 117.30μM and 3.46s−1, respectively. Ss-RmlA was promiscuous in the substrate choice and it could use three nucleoside triphosphates (dTTP, dUTP, and UTP) and three sugar-1-Ps (Glc-1-P, GlcNH2-1-P, and GlcN3-1-P) to form nine sugar nucleotides (dTDP-GlcNH2, dTDP-GlcN3, UDP-Glc, UDP-GlcNH2, UDP-GlcN3, dUDP-Glc, dUDP-GlcNH2, and dUDP-GlcN3). Ss-RmlB showed maximal enzyme activity at 50°C and pH 7.5 with 0.02mM NAD+, and the Km and kcat values for dTDP-glucose were 98.60μM and 11.2s−1, respectively. A one-pot four-enzyme reaction system was developed by simultaneously mixing all of the substrates, reagents, and four enzymes Ss-RmlABCD in one pot for the synthesis of dTDP-l-rhamnose and dUDP-l-rhamnose with the maximal yield of 65% and 46%, respectively, under the optimal conditions. dUDP-l-rhamnose was a novel nucleotide-activated rhamnose reported for the first time.


INTRODUCTION
L-rhamnose is a common 6-deoxy hexose found in the cell wall polysaccharides of many bacteria and the carbohydrate moieties of many natural products from bacteria and plants (Halimah et al., 2015;Khalil et al., 2015;Wittgens et al., 2018;Edgar et al., 2019;Garcia-Vello et al., 2020). These rhamnosecontaining biomolecules show great potential in the development of drugs and vaccines owing to their distinct bioactive properties (Wang et al., 2011;Lee et al., 2014;Mei et al., 2019;Pequegnat and Monteiro, 2019). Moreover, the anti-rhamnose antibody present in high titer in human serum can specifically bind to rhamnose or rhamnose-containing compounds, which inspires the design of antitumor vaccines conjugated with rhamnose for augmenting immunogenicity based on an antibody-dependent antigen uptake mechanism (Sarkar et al., 2010;Hossain et al., 2018). Therefore, the synthesis of rhamnose-containing biomolecules with pharmacological effects has been considered to be of great importance and attracted much attention (Xing et al., 2015;Sylla et al., 2019;Xu et al., 2019).
Over the past few decades, many Rml enzymes from different prokaryotes have been reported (Graninger et al., 2002;Han et al., 2007;Steiner et al., 2008;van der Beek et al., 2019). Most of those works focused on the biological roles of these enzymes and their potential as drug targets for the treatment of pathogenic bacteria, such as Mycobacteria tuberculosis H37Rv (Brown et al., 2017), Streptococcus pyogenes 5,448 (van der Beek et al., 2019), Bacillus anthracis str. Ames (Gokey et al., 2018), etc. However, there are a few studies on the in vitro synthesis of dTDP-l-rhamnose using these enzymes.
Several efforts have been made in the enzymatic synthesis of dTDP-l-rhamnose using Rml enzymes from different bacteria. A two-step synthesis of dTDP-l-rhamnose from dTDP-d-glucose was achieved using RmlBCD from Salmonella enterica LT2 (Marumo et al., 1992). This method first used RmlB to catalyze the conversion of dTDP-d-glucose into dTDP-4-keto-6-deoxyglucose, and then RmlC and RmlD were added to further catalyze the formation of dTDP-l-rhamnose. Another one-pot synthesis of dTDP-l-rhamnose from dTDP-d-glucose was accomplished with RmlBCD from Aneurinibacillus thermoaerophilus DSM 10155 (Pazur and Shuey, 2002). These two approaches employed the expensive substrate dTDP-dglucose, which limited the scale of production of dTDP-lrhamnose. A low-cost synthesis of dTDP-l-rhamnose from dTMP and Glc-1-P was achieved by a combined enzymatic pathway with six crude enzymes, in which three enzymes of TMP kinase, acetate kinase, and dTDP-glucose synthase (RmlA) were cloned from Escherichia coli, RmlB was cloned from Salmonella enterica LT2, and two enzymes of RmlC and RmlD were from Mesorhizobium loti (Oh et al., 2003;Kang et al., 2006). However, when using cell extracts as catalysts, the replicability of the method and the purification of product were challenging. Another two-step synthesis of dTDP-lrhamnose was conducted using five enzymes from dTMP and sucrose (Elling et al., 2005). This method first used dTMPkinase from Saccharomyces cerevisiae, sucrose synthase from potato, and RmlB from S. enterica LT2 to catalyze the synthesis of dTDP-4-keto-6-deoxy-glucose from dTMP and sucrose, and then the purified dTDP-4-keto-6-deoxy-glucose was used as substrate in the reaction catalyzed by RmlC and RmlD from S. enterica LT2 to generate dTDP-l-rhamnose. This method started synthesis from cheap substrates but the operation was complicated. Recently, a one-pot four-enzyme synthesis of dTDP-l-rhamnose with Cps23FL, Cps23FM, Cps23FN, and Cps23FO from Streptococcus pneumoniae 23F was developed by the addition of the enzymes in two portions, in which, in order to avoid the probable inhibitory effect of dTDP-l-rhamnose on the enzyme activity, Cps23FL was mixed with the substrates and reagents for the formation of dTDP-d-glucose firstly, and then the other three enzymes, Cps23FM, Cps23FN, and Cps23FO, were added for further catalytic reaction to produce dTDP-lrhamnose . From all the published work mentioned above, it could be seen that the discovery of new Rml enzymes and the development of a simple and efficient reaction system for the practical enzymatic synthesis of dTDPl-rhamnose should be of great significance.
In this work, four putative genes Ss-rmlABCD encoding for dTDP-l-rhamnose biosynthesis pathway in Saccharothrix syringae CGMCC 4.1716 were successfully cloned and expressed in E. coli BL21 (DE3), and the sequential synthesis of dTDP-l-rhamnose by these four enzymes Ss-RmlABCD was proved. The enzymatic properties of Ss-RmlA and Ss-RmlB, the first two enzymes in the dTDP-l-rhamnose biosynthesis pathway, were studied in detail. A new simple Frontiers in Microbiology | www.frontiersin.org and efficient four-enzyme reaction system for the synthesis of dTDP-l-rhamnose by simultaneously mixing all of the substrates, reagents, and four enzymes Ss-RmlABCD in one pot was then developed and optimized. This reaction system could also convert Glc-1-P and dUTP to dUDP-l-rhamnose. Our work characterized a novel set of dTDP-l-rhamnose synthesis enzymes derived from S. syringae CGMCC 4.1716 and further revealed its potential as the biocatalyst for the practical enzymatic synthesis of dTDP-l-rhamnose and dUDP-l-rhamnose.

Strains and Culture Conditions
The S. syringae CGMCC 4.1716 strain purchased from the China General Microbiological Culture Collection (CGMCC) Center was recovered and cultured in the 0234 broth (peptone 10 g/l, yeast extract 2 g/l, hydrolyzed casein 2 g/l, NaCl 6 g/l and glucose 10 g/l, pH 7.2) as described by the supplier's instruction. E. coli BL21 (DE3) used for protein expression was grown in Luria-Bertani (LB) medium at 37°C with 50 μg/ml ampicillin (Sangon Biotech, China) added when required.

Genes Cloning and Heterogeneous Expression
The four genes rmlABCD from S. syringae CGMCC 4.1716 were amplified with four pairs of specific primers (Supplementary Table S1) which were designed based on the target sequences (GenBank accession no. WP_033428542 for rmlA, WP_033429095 for rmlB, WP_033434852 for rmlC, and WP_033429096 for rmlD) using the genomic DNA of S. syringae CGMCC 4.1716 as PCR template. The purified PCR product was cloned into pET-22b plasmid with a histidine tag coding sequence fused to the C-terminal of each gene. The recombinant plasmid was transformed into E. coli BL21 (DE3) and the proper transformants were cultured in LB medium at 37°C. When the cell density reached 0.6-0.8 at 600 nm, 0.25 mM isopropyl β-d-thiogalactoside (IPTG) was added to the culture to induce the expression of the recombinant proteins. The cells were further cultured at 16°C for 16 h, harvested by centrifugation, and lysed by ultrasonic treatment. The cell lysate was centrifuged, and the supernatant was used for protein purification with nickel affinity chromatography. The concentration of the purified proteins was determined using the Bradford method with bovine serum albumin as a standard.

Functional Confirmation of Ss-RmlABCD
The glucose-1-phosphate thymidylyltransferase activity of Ss-RmlA was confirmed by detecting the generation of dTDP-d-glucose from dTTP and Glc-1-P in a 100 μl of reaction mixture containing 5.0 mM dTTP, 5.0 mM Glc-1-P, 10 mM MgCl 2 , and 100 μg/ml of Ss-RmlA in 40 mM Tris-HCl buffer (pH 8.0). The dTDP-d-glucose 4,6-dehydratase activity of Ss-RmlB was confirmed by detecting the formation of dTDP-4-keto-6-deoxy-glucose in the Ss-RmlA reaction mixture supplemented with 5 mM NAD + and 100 μg/ml of Ss-RmlB. The dTDP-4-keto-6-deoxy-glucose 3,5-epimerase activity of Ss-RmlC was confirmed by detecting the dTDP-4-keto-rhamnose formation in the Ss-RmlB reaction mixture supplemented with 100 μg/ml of Ss-RmlC. The dTDP-4-ketorhamnose reductase activity of Ss-RmlD was confirmed by detecting the dTDP-l-rhamnose formation in the Ss-RmlC reaction mixture supplemented with 5 mM NADPH and 100 μg/ml of Ss-RmlD. All the reactions were performed at 37°C for 5 min and then terminated by mixing with 100 μl chloroform. After centrifugation, the upper water phase was used for the detection of products by TLC and HPLC. The products were purified by HPLC and desalted for MS analysis.

Enzyme Assays of Ss-RmlA and Ss-RmlB
The enzyme activity of Ss-RmlA was determined by measuring the amount of released PPi using a colorimetric method (Sha et al., 2011) with minor modifications. The 100 μl reaction mixture containing 40 mM Tris-HCl buffer (pH 8.0), 5.0 mM dTTP, 5.0 mM Glc-1-P, 10 mM MgCl 2 , 100 μg/ ml of Ss-RmlA and 2 U/ml YIPP was incubated at 37°C for 5 min and terminated by mixing with 100 μl malachite green reagent containing 0.03% (w/v) of malachite green, 0.2% (w/v) of ammonium molybdate, 0.05% (v/v) of Triton X-100, and 0.7 N HCl. The absorbance of the mixture was measured at 630 nm (OD 630 ) with a microplate reader (BioTek, United States) after a 5-min incubation at 37°C. The amount of PPi released from the reaction was calculated using a standard curve relating OD 630 value to PPi concentration. One unit of Ss-RmlA activity was defined as the amount of Ss-RmlA catalyzing the generation of 1 μmol PPi per min under the assay conditions (Sha et al., 2011). The enzyme activity of Ss-RmlB was determined by measuring the amount of generated dTDP-4-keto-6-deoxyglucose using a colorimetric method (Shi et al., 2016) with minor modifications. The 100 μl reaction mixture containing 40 mM Tris-HCl buffer (pH 8.0), 5.0 mM dTDP-d-glucose and 100 μg/ml of Ss-RmlB was incubated at 37°C for 5 min and terminated by mixing with 10 μl of 1 M NaOH solution at room temperature for 10 min. Then the absorbance at 320 nm (OD 320 ) was detected with the microplate reader and the amount of dTDP-4-keto-6-deoxy-glucose was calculated using the standard curve of the OD 320 value vs. dTDP-4-keto-6-deoxy-glucose concentration. One unit of Ss-RmlB activity was defined as the amount of Ss-RmlB catalyzing the generation of 1 μmol dTDP-4-keto-6-deoxyglucose per min under the assay conditions.

Biochemical Studies of Ss-RmlA and Ss-RmlB
The optimal temperature for the reaction catalyzed by Ss-RmlA or Ss-RmlB was examined by assaying the enzyme activity at 16-80°C. Thermostability was determined by examining the residual enzyme activity after incubating the enzyme in the abovementioned temperature range for 1 h. The optimal pH for the reaction was determined by assaying the enzyme activity at pH 3.5-12.0 in 40.0 mM buffers (sodium acetate buffer at pH 3.5-6.5, Tris-HCl buffer at pH 7.5-9.0, and NaHCO 3 -NaOH buffer at pH 9.5-12.0). The pH stability was determined by incubating the enzyme in the abovementioned buffers at 4°C for 1 h and assaying the residual enzyme activity. The effects of metal ions on Ss-RmlA activity were examined by assaying the enzyme activity in the presence of 2.0 mM of EDTA, NaCl, KCl, ZnCl 2 , AgNO 3 , MnCl 2 , CaCl 2 , MgCl 2 , CuCl 2 , FeCl 2 , NiSO 4 , HgSO 4 , or CoSO 4 . The reaction without metal ions was used as the control. The optimal concentration of MgCl 2 for Ss-RmlA activity was explored by assaying the enzyme activity in the presence of 0-5.5 mM MgCl 2 . The optimal NAD + concentration for Ss-RmlB activity was examined by assaying the enzyme activity in the presence of 0-0.08 mM NAD + .
The kinetic analysis of Ss-RmlA for both substrates was conducted under the optimal reaction conditions using varied dTTP (0-0.5 mM) and varied Glc-1-P (0-0.3 mM) with the other one of these two substrates at a saturated concentration (1 mM). The kinetic analysis of Ss-RmlB was conducted under the optimal reaction conditions using varied concentrations of dTDP-d-Glc (0-1.0 mM). The K m and k cat values of the two enzymes were calculated using the software GraphPad Prism.6 1 The substrate acceptance of Ss-RmlA for different NTPs and sugar-1-Ps was examined with a 100 μl reaction mixture containing 2 U/ml Ss-RmlA, 2 U/ml YIPP, 2.5 mM MgCl 2 , 5 mM NTP, and 5 mM sugar-1-P in 40 mM NaHCO 3 -NaOH buffer (pH 9.5) at 37°C for 12 h. The substrate dTTP was used to test other sugar-1-Ps (GlcNAc-1-P, GlcA-1-P, GlcNH 2 -1-P, GlcN 3 -1-P, and Man-1-P), and Glc-1-P was used to test other NTPs (ATP, dATP, GTP, dGTP, CTP, dCTP, UTP, and dUTP). Product formation was verified by MS analysis.

One-Pot Synthesis of dTDP-l-Rhamnose and dUDP-l-Rhamnose
The synthesis was performed by simultaneously mixing all of the substrates, reagents, and four enzymes Ss-RmlABCD in one pot. To achieve the maximum yield, the reaction conditions including temperature, pH, NADPH concentration, enzyme concentration, and reaction time were evaluated. For the synthesis of dTDP-l-rhamnose, 10 mM dTTP, 10 mM Glc-1-P, 2.5 mM MgCl 2 , 0.02 mM NAD + , and 40 mM Tris-HCl buffer were used. The effects of temperature (16-80°C) were determined using 5.0 mM NADPH and 100 μg/ml of each enzyme at pH 9.0. The effects of pH (3.5-12.0) were determined using 5.0 mM NADPH and 100 μg/ml of each enzyme at 30°C. The effects of NADPH concentration (0-8.0 mM) were explored using 100 μg/ml of each enzyme at pH 8.5 and 30°C. The effects of each enzyme concentration (100-300 μg/ml) were determined using 1.5 mM NADPH and 100 μg/ml of each of the other three enzymes at pH 8.5 and 30°C. The effects of reaction time were evaluated by using 1.5 mM NADPH, 100 μg/ml of each of three enzymes (Ss-RmlA, Ss-RmlB, and Ss-RmlD) and 200 μg/ml of Ss-RmlD at pH 8.5 and 30°C with interval sampling within 3 h. For the synthesis of dUDP-l-rhamnose, except for 10 mM dUTP, the components of the reaction mixture and conditions were the same as those for the synthesis of dUDPl-rhamnose. All the reactions were performed for 20 min and then terminated by mixing with an equal volume of chloroform. After centrifugation, the upper water phase was used for the detection of the product by TLC and HPLC. The products were purified by HPLC, and the identified fractions were concentrated by lyophilization. Then, the concentrated sample was desalted by being eluted from the Sephadex G10 column (10 × 300 mm, GE Healthcare, United States) with distilled water at a flow rate of 1 ml/min and detected at 260 nm. The obtained pure product was lyophilized to dry powder and redissolved in distilled water and deuterated water for MS and nuclear magnetic resonance (NMR) analysis, respectively. The product yield was defined as the ratio of the concentration of the synthesized nucleotide-activated rhamnose (mM) to the concentration of added deoxynucleoside triphosphate (mM).

TLC and HPLC Analysis
TLC was performed by loading samples on silica gel 60 F254 plates (Merck, Germany). The loaded samples were developed by a mixture of 95% ethanol/1 M acetic acid (5:2, pH 7.5; Kaminski and Eichler, 2014) and visualized by spraying the plate with 0.5% (w/v) 3,5-dihydroxytoluene in 20% (v/v) sulfuric acid and heating it at 120°C for 5 min.
HPLC was performed on an Agilent 1,260 series HPLC system coupled with a UV detector (Agilent Technologies, Inc. United States) using the CarboPac™ PA-100 column (4 × 250 mm, 4 μm particle size, Thermo Fisher Scientific, United States). The sample was eluted with a gradient concentration of ammonium acetate, set as 0-30 mM (0-12 min), 30-60 mM (12-22 min), 100 mM (22-32 min), and 0 (32-37 min), as the mobile phase at a flow rate of 1.0 ml/min and detected at 260 nm. The corresponding fractions were combined and concentrated through lyophilization.

Mass Spectrometry and NMR
The mass spectra (MS) were recorded on a Shimadzu liquid chromatography-mass spectrometry ion trap time of flight (LCMS-IT-TOF) instrument (Kyoto, Japan) equipped with electrospray ionization (ESI) source in negative ion mode at a resolution of 10,000 full width at half-maximum. The nuclear magnetic resonance (NMR) spectra were recorded on an Agilent DD2 600 MHz spectrometer (Agilent Technologies, Inc. United sTATES) at 600 MHz for 1 H and at 150 MHz for 13 C, and at 242 MHz for 31 P at 25°C. Chemical shifts were expressed in parts per million (ppm) downfield from the internal tetramethylsilane of D 2 O. Homo-and heteronuclear correlation experiments, including 1 H− 1 H correlation spectroscopy (COSY), and heteronuclear single quantum coherence (HSQC) were run using the standard pulse sequences.

RESULTS
Sequence Analysis and Expression of Ss-RmlA, Ss-RmlB, Ss-RmlC, and Ss-RmlD Four genes Ss-rmlABCD in the genome of S. syringae CGMCC 4.1716 are annotated to encode the putative dTDP-l-rhamnose synthetic pathway according to the GenBank database. The distribution pattern of these four genes in the genome of S. syringae CGMCC 4.1716 was compared with those of the homologs reported from four gram-positive bacteria (Saccharopolyspora spinosa, Mycobacteria tuberculosis H37Rv, Streptomyces sp. MK730-62F, and Streptococcus pneumoniae 23F), two gram-negative bacteria (E. coli K12 and S. enterica LT2), and two archaea (Haloferax volcanii DS2 and Sulfurisphaera tokodaii str. 7). The rml genes in the genome of S. syringae CGMCC 4.1716 were organized in three separate regions as Ss-rmlC, Ss-rmlDB, and Ss-rmlA, being different from those of the other homologs (Figure 2). The deduced amino acid sequences of the enzymes Ss-RmlA, Ss-RmlB, Ss-RmlC, and Ss-RmlD encoded by the rml genes from S. syringae CGMCC 4.1716 shared sequence identities of 33-81%, 46-80%, 35-56%, and 32-66% with those of the abovementioned homologs, respectively ( Table 1).
The multiple alignments conducted with the predicted enzymes Ss-RmlABCD and their respective homologs from the abovementioned eight species indicated that there were several functionally critical motifs in the four Ss-Rml enzymes (Figure 3). Ss-RmlA possessed the motifs of GXGT/SRLXPXTX 4 K and LGDNX 4 for the recognition and binding of dTTP, the motifs of XEKP and SXRGEXEIT for the recognition and binding of Glc-1-P, and Mg 2+ -stabilizing motifs of DTG and GDN within the LGDNX 4 . Ss-RmlB contained GG/AAGFIG, the signature motif of the short chain dehydrogenase/reductase (SDR) superfamily, as well as H/NXAAES/TH and STDEVYG, the motifs for recognition and binding of NAD + and dTDPglucose. Ss-RmlC had the substrate-binding motifs DXRGXF/ LX 2 and Q/MXN/YXSXS/T. Ss-RmlD possessed the GX 2 GX 2 G, a signature motif of reductases/epimerases/dehydrogenase superfamily, and the substrate-binding motifs STDYVFXG and YG/AXT/SKL/RXGE (Bais et al., 2018;Dhaked et al., 2019).

Functional Confirmation of Ss-RmlABCD
The enzyme activities of Ss-RmlA, Ss-RmlB, Ss-RmlC, and Ss-RmlD were analyzed by examining their catalytic products through TLC, HPLC, and MS. With dTTP and Glc-1-P as the substrates, Ss-RmlA catalyzed the formation of the product P1 in the reaction. As shown in Figure 5 Figure S2). When Ss-RmlC was incubated in the reaction mixture of Ss-RmlA and Ss-RmlB, the spot of P3 on the TLC plate and its peak in HPLC were quite similar to those of P2 ( Figure 5). Moreover, the substrate P2 (dTDP-4-keto-6-deoxy-Glc) and P3 (dTDP-4-keto-rhamnose) shared the identical molecular mass of 546.07, so the product P3 could not be confirmed here. After NADP + and Ss-RmlD were incubated in the reaction mixture of three enzymes Ss-RmlABC, a new spot could be found on the TLC plate and a peak with the retention time of 18 min could be recognized in HPLC analysis (Figure 5). Frontiers in Microbiology | www.frontiersin.org data of dTDP-l-rhamnose . So, dTDP-l-rhamnose was successfully synthesized by the stepwise-catalysis of Ss-RmlABCD.
The optimal temperature for Ss-RmlB activity was 50°C, and the enzyme was stable below 42°C ( Figure 7A). Ss-RmlB was highly active in the pH range of 7.0-8.5 with the maximal activity obtained at pH 7.5, and the enzyme was stable at pH 7.0-9.0 ( Figure 7B). NAD + could promote Ss-RmlB activity, and the optimal concentration of NAD + was 0.02 mM. When NAD + exceeded 0.02 mM, the enzyme activity dropped considerably ( Figure 7C). The K m and k cat values of Ss-RmlB for dTDP-glucose were 98.60 μM and 11.2 s −1 , respectively.

One-Pot Synthesis of dTDP-l-Rhamnose and dUDP-l-Rhamnose
The one-pot synthesis of dTDP-l-rhamnose was performed by incubating four enzymes Ss-RmlABCD with dTTP and Glc-1-P as the starting substrates. The effects of temperature, pH, and concentrations of NADPH and enzymes on dTDP-l-rhamnose yield were investigated in detail. As shown in Figure 8A, the reaction temperature markedly affected dTDP-l-rhamnose formation. As the temperature was raised from 16 to 70°C, dTDP-l-rhamnose yield rapidly increased from 14% to the maximum of 47% at 30°C, decreased to 28% at 50°C, and dropped to 0 at 70°C. Thus, subsequent reactions were performed at 30°C. The pH values also strongly affected dTDP-l-rhamnose yield. Figure 8B showed that dTDP-l-rhamnose yield increased at pH 3.5-7.5 and then stabilized at pH 7.5-9.5 with the maximal yield of 53% obtained at pH 8.5. When pH exceeded 9.5, dTDPl-rhamnose yield sharply decreased to 0 at pH 12.0. Thus, the subsequent reactions were performed at pH 8.5. NADPH was an essential cofactor for the final step of reduction catalyzed by Ss-RmlD and the effect of its concentration on dTDP-lrhamnose yield was examined. As shown in Figure 8C, when NADPH was increased from 0 to 1.5 mM, dTDP-l-rhamnose yield increased from 0 to the maximum of 52% and then kept stable at 1.5-8.0 mM. Thus, the subsequent reactions were performed using 1.5 mM of NADPH. Figure 8D showed the effect of each enzyme concentration on dTDP-l-rhamnose yield. When the concentration of each enzyme changed from 100 to 300 μg/ml, the change of the concentration of Ss-RmlC from 100 to 200 μg/ml significantly influenced dTDP-l-rhamnose yield with a significant increase from 42 to 63%, but the change of the concentration of each of three enzymes Ss-RmlA, Ss-RmlB, and Ss-RmlD did not affect the dTDP-l-rhamnose yield. Therefore, the optimal conditions for one-pot synthesis of dTDP-l-rhamnose was 10 mM dTTP, 10 mM Glc-1-P, 0.02 mM NAD + , 1.5 mM NADPH, 100 μg/ml of each of three enzymes Ss-RmlABD, and 200 μg/ml of Ss-RmlC at pH 8.5 and 30°C. The time curves indicated that dTDP-l-rhamnose yield reached the maximum of 65% at 90 min (Figure 9). Next, the positive substrates of Ss-RmlA, including dTTP, dUTP, UTP, Glc-1-P, GlcNH 2 -1-P, and GlcN 3 -1-P were tested as the starting substrates in a one-pot reaction under the optimal synthesis conditions for dTDP-l-rhamnose. In addition to dTTP and Glc-1-P, dUTP and Glc-1-P could also be used by Ss-RmlABCD as the starting substrates to form a new product with a yield of 46% at 90 min (Figure 9). This product was purified by HPLC, and identified by MS with the peak of [M-H] − at m/z 533.0538 in the negative ion ESI mass analysis, consistent with the theoretical molecular mass of dUDP-l-rhamnose (534.0652; Supplementary Figure S13). The chemical structure of this new product was further elucidated by 1 H, 13 C, 31 P, 1 H-1 H COSY, 1 H-13 C HSQC, and 1 H-13 C HSQC without decoupling NMR analysis (Supplementary Figures S14-S19). The proton signal of the peak at δ 5.04 ppm and the carbon signal of the peak at δ 95.4 ppm in the 1 H-and 13 C-NMR spectra were, respectively, assigned to be the H-1 and C-1 of the rhamnose moiety (Supplementary Figures  S14 and S15). According to the 1 H-13 C HSQC spectrum without decoupling (Supplementary Figure S19), the coupling constant of the C-1 and H-1 (J C1, H1 ) of the rhamnose moiety was determined to be 162 Hz, revealing a β-configuration of its anomeric carbon. Therefore, the product was identified as dUDP-β-l-rhamnose.
Frontiers in Microbiology | www.frontiersin.org of the glycosyl donor dTDP-l-rhamnose which is difficult to achieve through chemical synthesis. In bacteria and archaea, dTDP-l-rhamnose is synthesized by four conserved enzymes RmlABCD. Thus far, some studies of enzymatic synthesis of dTDP-l-rhamnose in vitro with a set of bacterial RmlABCD or partial bacterial Rml enzymes combined with other enzymes from potato and Saccharomyces cerevisiae have been reported (Marumo et al., 1992;Pazur and Shuey, 2002;Oh et al., 2003;Elling et al., 2005;Kang et al., 2006;Li et al., 2016). The discovery of a novel set of RmlABCD enzymes and the development of a simple and efficient enzymatic synthesis reaction system would be very worthy of further exploration. In this work, a novel set of four novel Ss-RmlABCD enzymes from S. syringae CGMCC 4.1716 were heterologously expressed and functionally confirmed, and one-pot synthesis of dTDPl-rhamnose (yield of 65%) and dUDP-l-rhamnose (yield of 46%) by employing Ss-RmlABCD enzymes was developed. The genomic organizations of rmlABCD genes in prokaryotic microorganisms are polymorphic. The rmlABCD genes of E. coli K12, S. enterica LT2, Sulfurisphaera tokodaii str. 7, Haloferax volcanii DS2, and Streptomyces sp. MK730-62F are successively located within the biosynthetic gene clusters for cell wall glycans or the secondary metabolite. On the contrary, for some other species, the rmlABCD genes are separately located on their genomes, spaced apart by other genes. For example, the four genes were arranged in three regions in the order of rmlDB, rmlA, and rmlC in Saccharopolyspora spinosa, rmlA, rmlD, and rmlBC in Mycobacteria tuberculosis H37Rv, as well as rmlD, rmlAC, and rmlB in Streptococcus pneumoniae 23F. In this study, the four Ss-rmlABCD genes from S. syringae CGMCC 4.1716 were found to be also located in three discrete regions of the genome in the order of Ss-rmlC, Ss-rmlDB, and Ss-rmlA, presenting a different pattern compared with those of the reported homologs.
Although RmlABCD enzymes act in synergy to synthesize dTDP-l-rhamnose, some individual Rml enzymes show different optimal reaction conditions compared with other enzymes in the same pathway. The variance was observed for both bacterial and archaeal Rml homologs. For example, Cps23FL (RmlA) and Cps23FN (RmlB) from Streptococcus pneumonia serotype 23F shared the same optimal pH for enzyme activity, but the optimal temperature for Cps23FN (37°C) was 12°C higher than that of Cps23FL (25°C; Li et al., 2016). For the RmlABCD enzymes from Sulfolobus tokodaii strain 7, RmlA showed maximal activity at pH 8.5 and 95°C, whereas RmlB was most active at 80°C (Zhang et al., 2005;Teramoto et al., 2012). In our study, Ss-RmlA displayed maximal enzyme activity at pH 9.0 and 37°C, while Ss-RmlB showed maximal enzyme activity at pH 7.5 and 50°C. Thus, in order to maximize dTDP-l-rhamnose yield, we optimized the conditions for the one-pot reaction catalyzed by Ss-RmlABCD. The optimal pH and temperature Frontiers in Microbiology | www.frontiersin.org determined for the one-pot reaction were pH 8.5 and 30°C which were different from those of Ss-RmlA and Ss-RmlB, suggesting that the optimal conditions for the one-pot fourenzyme reaction could be a compromise for those of the four enzymes.
After the optimization of conditions including pH, temperature, and NADPH concentration for the one-pot reaction, we next analyzed the effect of enzyme concentration on dTDPl-rhamnose production. As shown in Figure 8D, the increase of Ss-RmlC concentration from 100 to 200 μg/ml led to a corresponding rise in dTDP-l-rhamnose yield from 42 to 63%, but the concentration change from 100 to 300 μg/ml of Ss-RmlA, Ss-RmlB or Ss-RmlC could not lead to a similar effect, which suggested that these three enzymes might be surplus for catalysis even at the concentration of 100 μg/ml.
The one-pot enzymatic synthesis of dTDP-l-rhamnose by RmlABCD involves multiple factors possibly limiting the product yield. RmlA, the first enzyme in the pathway, was reported to be significantly inhibited by the end product dTDP-l-rhamnose in vitro, which in turn limited the final yield. Li et al. reported that quite a low yield (~1%) of dTDP-l-rhamnose was obtained when Cps23FL, Cps23FN, Cps23FM, and Cps23FO were simultaneously used in one pot. On the contrary, the dTDP-lrhamnose yield could be improved to 63% when the four enzymes were added in two portions, in which the first enzyme Cps23FL was added to synthesize dTDP-d-glucose firstly, and then the enzymes Cps23FN, Cps23FM, and Cps23FO (RmlBCD) were supplemented to the reaction mixture to yield dTDP-l-rhamnose . With such an operation, they minimized the inhibition of dTDP-l-rhamnose on Cps23FL (RmlA) activity. In  Frontiers in Microbiology | www.frontiersin.org contrast, we did not observe the obvious inhibitory effect of dTDP-l-rhamnose on Ss-RmlA activity and obtained a similar yield of dTDP-l-rhamnose whether we added Ss-RmlABCD simultaneously or in two portions as Li et al. reported (data not shown), suggesting that dTDP-l-rhamnose probably had no significant inhibitory effect on the enzyme activity of Ss-RmlA. Several sets of Rml enzymes from different bacterial species have been successfully employed to synthesize dTDP-l-rhamnose in vitro. The highest yield of dTDP-l-rhamnose reported so far is 63% achieved by Li et al. using four Rml homologs Cps23FL, Cps23FN, Cps23FM, and Cps23FO from S. pneumonia serotype 23F . However, the authors here had to carry out the synthesis reaction by adding the four involved enzymes in two portions in order to avoid the inhibition on Cps23FL by dTDP-l-rhamnose. In this work, by simultaneously mixing Ss-RmlABCD enzymes, the substrates, and the necessary reagents in one pot, we developed a simple four-enzyme reaction system for the synthesis of dTDP-l-rhamnose with a comparable conversion yield (65%). Moreover, using this reaction system, we successfully synthesized a structural analog of dTDP-lrhamnose, dUDP-l-rhamnose, which was an unnatural nucleotide-activated rhamnose reported for the first time. Next, in order to develop a more cost-effective one-pot enzymatic process for the synthesis of dTDP-l-rhamnose or dUDP-lrhamnose, enzyme immobilization and cell surface display techniques could be strategies worthy of attempts to reduce the cost of enzymes (Zhang et al., 2006;Silva-Salinas et al., 2021).
In conclusion, this work identified and characterized a novel set of dTDP-l-rhamnose synthetic enzymes Ss-RmlABCD from S. syringae CGMCC 4.1716 and provided a new simple and efficient reaction system that laid a foundation for the practical enzymatic synthesis of dTDP-l-rhamnose and dUDP-l-rhamnose.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.

AUTHOR CONTRIBUTIONS
SY conceived and designed the research and wrote the manuscript. XA and SY performed the experiments and analyzed the data. SY, LX, and MX revised the manuscript. LX, XJ and MX supervised the project. All authors contributed to the article and approved the submitted version.