Introduction

Peptidoglycan is a network of interconnected polymers comprised of N-acetylglucosamine and N-acetylmuramic acid, which is unique to eubacteria and possesses many biological activities including pyrogenicity, toxigenicity, immunogenicity and somnogenicity (Biberstine and Rosenthal 1994; Burroughs et al. 1993). The biosynthesis of peptidoglycan is a two-stage process. The monomeric building blocks N-acetylmuramyl pentapeptide and N-acetylglucosamine are formed in the cytoplasm during the first stage (Silver 2006). As the primary committed step of biosynthesis, the condensation of phosphor (enol) pyruvate (PEP) and UDP-N-acetylglucosamine is catalyzed by MurA; then the enol-pyruvate moiety is reduced by MurB to generate UDP-N-acetylmuramate (UDP-MurNAc). After that, a series of ATP-dependent amino acid ligases (MurC, MurD, MurE and MurF) catalyze the stepwise addition to form a pentapeptide side chain at the newly reduced d-lactyl group (Fig. 1). The second stage of peptidoglycan biosynthesis involves the transportation of precursor across the membrane by a lipophilic carrier, and subsequent addition to the growing cell wall polymer catalyzed by penicillin-binding proteins (PBPs) (Heijenoort 1994; Matsuhashi 1994; Raymond et al. 2003; Zhang et al. 2007b). All enzymes involved in the early cytoplasmic stage of biosynthesis are highly conserved among bacteria, thus a possible inhibitor of any Mur enzyme could present a wide antibiotic spectrum. However, although each Mur enzyme has been crystalized and characterized (Schönbrunn et al. 1996; Benson et al. 1996; Deva et al. 2006; Bertrand 2007; Gordon and Flouret 2001; Yan 2000), no available inhibitor from known antibiotics or synthetic chemicals has been found, only except MurA which could be inhibited by phosphonomycin (Heijenoort 1994; Zoeiby et al. 2001). Due to the incomprehension of pathway intermediates, the mechanistic and structural studies of Mur enzymes, as well as their inhibitor screening, have been severely hampered. Although the enzymatic synthesis of cytoplasmic intermediates via homogeneous overexpressed Mur enzymes have been attempted (Zoeiby et al. 2001; Reddy et al. 1999), only UDP- N-acetylglucosamine, which is the substrate of MurA, has been synthesized in vitro and commercially available (Zhou et al. 2010a, 2010b).

Fig. 1
figure 1

The cytoplasmic assembly of murein precursors, which arecatalyzed by the Mur enzymes (MurA through MurF). (1 UDP-N-acetylglucosamine; 2 UDP-N-acetylmuramate; 3 UDP-MurNAc-l-Ala; 4 UDP-MurNAc-l-Ala-γ-d-Glu; 5 UDP-MurNAc-l-Ala-γ-d-Glu-meso-2,6-diaminopimelate; 6 UDP-MurNAc-l-Ala-γ-d-Glu-meso-2,6-diaminopimeloyl-d-Ala-d-Ala)

Full size image

The cell-free protein synthesis (CFPS) system is newly emerged in vitro tool for efficient and rapid translation of target genes (Katzen et al. 2005; Swartz 2006; Carlson et al. 2011), which could be used as a substitution for traditional expression approaches. In addition, CFPS system is compatible with other high throughput technologies because of its direct usage of PCR-generated templates. Thus, due to the open nature and the absence of endogenous DNA templates, CFPS system is a promising platform for reconstitution of biosynthetic pathways, even metabolic networks (Jewett et al. 2008; Harris and Jewet 2012; Karig et al. 2011; Noireaux et al. 2003). The construction of artificial cells in CFPS system to explore the origin of life (Kurumaa et al. 2009; Murtasa et al. 2007) has testified the merits of this system to accelerate the synthetic biology studies. More important, CFPS system holds potential applications for the development of novel high throughput drug-screening models, which target on the entire pathway instead of a specific protein (Casteleijn et al. 2012). Thus, the adaption of CFPS system in the reconstitution of the peptidoglycan biosynthetic pathway could circumvent mentioned challenges during the peptidoglycan biosynthetic studies.

Usually, 20-nt-size antisense oligonucleotides can modulate gene expression by hybridizing with cognate RNAs at complementary target sites (Shao et al. 2006; Zamecnik and Stephenson 1978). The binding of antisense oligonucleotide to the mRNA may physically block the movement of ribosomes along the mRNA, thus the synthesis of defective proteins caused by splicing errors could be prevented and the degrading rate of mRNA within the cytosol could be increased. The fast and specific inhibition of gene expression induced by antisense oligonucleotides makes them a potential tool for functional genomics studies and drug target validation (Taylor et al. 1994; Ding and Lawrence 2001; De Backer et al. 2001). Combined with the application of CFPS system, the rational design of antisense oligonucleotides accompanied by target screening may provide novel insight into the modulation of the peptidoglycan biosynthetic pathway.

It was proposed that some purified enzymes could be used to construct the synthetic enzymatic pathways (SEPs) in vitro for small molecule production, (Zhang 2010; Roessner and Scott 1996; Wendell et al. 2010). SEPs can be achieved for rational programming of biosynthetic networks (Zhang et al. 2007a), but the process of expression and purification of the enzymes is still time consuming and laborious. In this study, we constructed the Mur biosynthetic pathway in CFPS system which was proved to successfully synthesize the final product UDP-N-acetylmuramic acid within about 5 h. Also, several antisense oligonucleotides were rationally designed and screened to inhibit the expression of Mur enzymes, thereby blocking the synthetic pathway.

Material and methods

Material and strains

Escherichia coli str. K-12 substr. MG1655 (ATCC-700926) was used as the original strain to amplify Mur genes; E. coli DH5α was used for the plasmids preparation. RTS 100 E. coli HY cell-free protein expression kit was purchased from 5 prime Co. (St. Bucksfield, MD, USA). Unless otherwise specified, all reagents used in this study were purchased from Sigma (St. Louis, MO, USA).

Plasmid construction

The original genes murA (EcoGene:EG11358), murB (EcoGene:EG11205), murC (EcoGene:EG10619), murD (EcoGene:EG10620), murE (EcoGene:EG10621), and murF (EcoGene:EG10622) were amplified from E. coli genomic DNA by PCR, respectively. The designed primers (Table 1) used for gene amplification were synthesized by Sangon (Shanghai, China) and the engineered restriction endonuclease sites were introduced to facilitate the subsequent cloning into the pIVEX2.4c (Roche, Indiana, USA) expression vectors. pIVEX-MurA was constructed by cloning the resulting PCR product into pIVEX2.4c via NcoI and BamHI restriction sites. Similarly, pIVEX-MurB (NcoI/PstI), pIVEX-MurC (NcoI/PstI), pIVEX-MurD (NcoI/BamHI), pIVEX-MurE (NcoI/BamHI) and pIVEX-MurF (XhoI/BamHI) were constructed. All resulting expression plasmids were validated by restriction mapping and DNA sequencing (Sangon).

Table 1 Nucleotide sequences of primers
Full size table

Cell-free expression of Mur enzymes

The constructed expression plasmids pIVEX-MurA, pIVEX-MurB, pIVEX-MurC, pIVEX-MurD, pIVEX-MurE and pIVEX-MurF were prepared from overnight cultures, respectively, using QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany). Then each linear CFPS template was amplified from the corresponding plasmid using two primers 5′-GGAATTCCATATGGTGGCTTTCAAAGG-3′ and 5′-GCCCGCTCGAGTACGGAAGACAAT-3′. To express a certain Mur enzyme, about 10 ng/μl linearized template was added into the RTS 100 E. coli HY Kit, and the reaction was carried out at 30 °C with a stirring speed of 400 rpm for 5 h. Then the reaction mixture was centrifuged at 13,000 rpm for 10 min and the resulting supernatant was analyzed by SDS-PAGE electrophoresis and Western blotting. For Western blotting analysis, the target gel was transferred onto a 0.45-μm nitrocellulose (NC) membrane (Pall, USA), then the membrane was incubated with primary mouse anti-His antibody (Qiagen, Germany) and HRP-conjugated secondary antibody of goat anti-mouse IgG subsequently. Finally, the membrane was visualized by chemiluminescence and exposed to a BioX-film (Kodak, USA).

Synthesis of UDP-N-acetylmuramyl pentapeptide

To initiate the reaction, 10 μl of co-expressed MurA and MurB CFPS sample was added into the reaction mixture (50 mM Tris–HCl, pH 8.0; 20 mM UDP-N-acetyl glucosamine; 20 mM PEP; 20 mM β-NADPH and 5 mM DTT) with the final volume of 100 μl in a 1.5-ml Eppendorf tube. The reaction mixture was incubated at 37 °C for 5 h and stopped by heating at 70 °C for 10 min. After centrifuging at 13,000 rpm for 10 min, the collected supernatant was used as the substrate to synthesize UDP-N-acetylmuramyl pentapeptide. The synthesis reaction mixture contained 50 mM Tris–HCl (pH 8.0), 20 mM ATP, 5 mM MgCl2, 5 mM DTT, 5 mM l-alanine, 5 mM d-glutamic acid, 5 mM meso-diaminopimelic acid, 5 mM d-alanine-d-alanine and 20 μl of co-expressed MurC–MurF CFPS sample. Following the same procedures, the resultant supernatant was analyzed by HPLC and electrospray ionization-mass spectrometry (ESI-MS).

Product analysis

To identify the product, 5 μl reaction sample was analyzed by Agilent 1100 Series HPLC equipped with a TC-C18 column (3.9 × 240 mm, 10 μm pore size) and a UV detector, using an isocratic flow of 50 mM ammonium formate (pH 3.5) containing 7.5 % menthol at 1 ml/min. The program was performed at 30 °C for 20 min, and the expected nucleotide sugars and ADP/ATP were detected at 262 nm. ESI-MS analysis was carried out in the negative ion detection mode using a Bruker LC/MSD Trap mass spectrometer (Bruker, Massachusetts, USA). The scan range was 100–2,000 m/z using five averages and 13,000 m/z s−1 resolution. The capillary voltage was +4,000 V for negative ions, with an end plate offset of 3,500 V. The collected data was processed by HP Chemstation.

Rational design of antisense oligonucleotide and in vitro inhibition of MurA and MurB expression

Sfold (http://sfold.wadsworth.org/cgi-bin/index.pl) was used to generate possible mRNA structures of Mur enzymes. To design expected antisense oligos, the local folding constraint of window width was set at 50nt, so the bases separated by the sequence longer than 50nt could not form a pair. The preferred length of antisense oligos was 20 nt.

To examine the possible inhibition of designed antisense oligos, the Mur genes were expressed in the CFPS system using 0.5 μg/50 μl linear template in presence of 40 μM unmodified antisense oligonucleotides. The reaction was performed at 30 °C for 4 h and then stopped by chilling on ice for 10 min, and finally analyzed by Western blotting. The quantity of target protein was calculated by Quantity One (Bio-Rad, USA). The inhibition percentage of tested antisense oligo was calculated as below:

$$ \mathrm{inhibition}\kern0.5em \%=\left(1-\frac{\mathrm{Protein}\kern0.5em \mathrm{band}\kern0.5em \mathrm{intensity}\kern0.5em \mathrm{with}\kern0.5em \mathrm{oligo}}{\mathrm{Protein}\kern0.5em \mathrm{band}\kern0.5em \mathrm{intensity}\kern0.5em \mathrm{with}\mathrm{out}\kern0.5em \mathrm{oligo}}\right)\times 100\% $$

Results

In vitro expression of Mur enzymes

The SDS-PAGE analysis clearly indicated that each of the six Mur enzymes was successfully expressed as a soluble form in CFPS system with a fused C-terminal poly-(His) 6 tag (Fig. 2). The linear PCR templates were equally added into the batch-mode CFPS system for co-expression of Mur enzymes and reconstitution of muramic acid biosynthetic pathway. Western blotting analysis has validated the successful co-expression of MurA and MurB. Considering about the very similar molecular weight of MurD (49.5 kDa) and MurF (50.8 kDa), we have concluded that the co-expression of four enzymes (MurC/MurD/MurE/MurF) and even all six enzymes were also accomplished (Fig. 2). However, the expression levels of different Mur enzymes were varied in a defined CFPS system, and the quantity of a certain expressed enzyme was significantly affected by the co-expression of other enzymes.

Fig. 2
figure 2

Expression of Mur enzymes in CFPS system. a Soluble protein samples were analyzed by SDS-PAGE. Lane M molecular marker; lane C negative control without plasmid addition; lane 1 MurA (47.6 kDa); lane 2 MurB (40.8 kDa); lane 3 MurC (56.4 kDa); lane 4 MurD (49.5 kDa); lane 5 MurE (56.1 kDa); lane 6 MurF (50.8 kDa). b Co-expressed Mur enzymes analyzed by Western blotting. Lane 7 Co-expression of MurA&B; lane 8 co-expression of MurC-F; lane 9 co-expression of all six Mur enzymes. Arrows indicate the position of target proteins

Full size image

Reconstitution of Mur pathway and biosynthesis of UDP-N-acetylmuramyl pentapeptide

The biosynthesis of UDP-N-acetylmuramyl pentapeptide involves the formation of UDP-N-Acetylmuramate catalyzed by MurA and MurB, and the addition of a pentapeptide side chain catalyzed by a series of ATP-dependent amino acid ligases (MurC, MurD, MurE and MurF). To reconstitute the Mur pathway, MurA and MurB were co-expressed in CFPS system to synthesize the important intermediate UDP-MurNAc. The HPLC analysis of ended reaction mixture indicated a distinct new signal of nucleotide sugar species (Fig. 3a), and the subsequent LC-MS analysis had detected the standard signal of UDP-MurNAc at m/z 678.1 (Fig. 4a). Using the reaction mix catalyzed by MurA and MurB which contained UDP-MurNAc as the substrate, a series of amino acid addition intermediates including UDP-MurNAc-l-Ala produced by MurC (Fig. 3b); UDP-MurNAc-l-Ala-γ-d-Glu produced by co-expressed MurC and MurD (Fig. 3c) and UDP-MurNAc-l-Ala-γ-d-Glu-meso-2,6-diaminopimelate produced by co-expressed MurC–MurE (Fig. 3d) were detected by corresponding HPLC analysis and validated by LC-MS (Fig. 4b–d), respectively, and the molecular weight of each intermediate was identical to the reported value (Raymond et al. 2003; Benson et al. 1996; Deva et al. 2006). When UDP-MurNAc was mixed with co-expressed MurC-MurF, a distinct peak eluted at 11.5 min of HPLC analysis (Fig. 3e) had been further identified as the final pathway product UDP-N-acetylmuramyl pentapeptide by LC-MS, giving the characteristic mass signal at m/z 1192.5 (Fig. 4e). These results confirmed the activities of the cell-free expressed Mur enzymes and successful in vitro reconstitution of the cytoplasmic biosynthetic pathway for peptidoglycan precursor.

Fig. 3
figure 3

HPLC analysis of intermediates from Mur biosynthetic pathway. a Control reaction without enzymes. UDP-N-Acetylglucosamine (1). b Reaction of UDP-MurNAc (2) synthesis with coexpressed MurA and MurB. c Reaction of UDP-MurNAc-l-Ala synthesis with MurC (3). d Reaction of UDP-MurNAc-l-Ala-γ-d-Glu (4) synthesis with coexpressed MurC and MurD. e Reaction of UDP-MurNAc-l-Ala-γ-d-Glu-meso-2,6-diaminopimelate (5) synthesis with coexpressed MurC/D/E. f Reaction of UDP-MurNAc-l-Ala-γ-d-Glu-meso-2,6-diaminopimeloyl-d-Ala-d-Ala (6) synthesis with coexpressed MurC/D/E/F

Full size image
Fig. 4
figure 4

ESI-MS analysis of intermediates from Mur biosynthetic pathway. a The spectrum of UDP-N-Acetylmuramate (m/z 678.1). b The spectrum of UDP-MurNAc-l-Ala (m/z 749.1). c The spectrum of UDP-MurNAc-l-Ala-γ-d-Glu (m/z 878.2). d The spectrum of UDP-MurNAc-l-Ala-γ-d-Glu-meso-2,6-diaminopimelate (m/z 1050.7). e The spectrum of UDP-MurNAc-l-Ala-γ-d-Glu-meso-2,6-diaminopimeloyl-d-Ala-d-Ala (m/z 1,192.5)

Full size image

Rational design and screening of antisense oligos for Mur enzymes

Four empirical rules were applied to filter the computer assisted rational designs of antisense oligos: (a) 40 % ≤ GC % ≤ 60 %; (b) no GGGG in the target sequence; (c) average unpaired probability of target site nucleotides ≥0.5; (d) the threshold probability of each peak in the accessibility profile should be above 0.5. It is possible that the accessibility of some regions in the target mRNA could be severely influenced by the tertiary folding of mRNA and the binding of other proteins. To avoid potential regional bias due to these effects, the designed oligos were distributed across the entire length of the mRNA sequence. As the most important prediction factor for antisense activity, ΔG disruption is the target disruption energy describing the free energy cost to open the secondary structure at the target site, so it also should be considered during the design and screening of antisense oligos. According to the screening result, four or five oligos targeting at different parts of mRNA were selected as the training set for MurA and MurB (Table 2).

Table 2 Antisense oligos of the training set
Full size table

Each antisense oligo was tested in CFPS system and evaluated its inhibition towards expression of target gene via quantitative Western blot analysis. As shown in Fig. 5, the addition of various oligos could weaken the expression of each target gene at different levels, and the resistance of each mur gene towards antisense oligos was also different. The strongest antisense oligo inhibitor for murA was A281 (Fig. 5a) and for murB was B139 (Fig. 5b). In other aspects, the inhibition effects of universal oligos U79 and U84 binding at the upstream of ORFs were less effective than oligos designed within the ORFs.

Fig. 5
figure 5

Oligo induced inhibition on expression of target Mur enzyme was quantified by Western blot analysis. Each result was derived from triplet samples in one experiment and every oligo was tested independently at least three times. a Oligos induced inhibition on expression of MurA. b Oligos induced inhibition on expression of MurB. c Expression inhibition of MurA and MurB induced by different concentrations (0–40 μM) of oligos, with 10 ng/μl target template in a reaction volume of 50 μl

Full size image

Antisense oligos designed for murC–murF were listed in Table S1. As shown in Fig. S1, the addition of various oligos could affect the expression of each target gene in different levels, and the resistance of different mur genes towards antisense oligos was also different. Among the Mur genes, murD was the most sensitive one; almost all selected antisense oligos could significantly repress the expression of MurD (Fig. S1b). The best antisense oligo inhibitors were C1539 for murC (Fig. S1a), E1378 for murE (Fig. S1c) and F936 for murF (Fig. S1d). Moreover, the promotion of improper protein expression caused by antisense oligos was also observed in the study of MurC (Fig. S1a).

The repression of enzymatic activity of cell-free expressed proteins by oligos

Different concentrations of A281and B139 ranging from 0 to 40 μM were respectively investigated to determine the effective concentration of target oligo for proper inhibition of mRNA translation. The Western blot results showed that the translation of murA or murB could be effectively repressed by adding 40 μM oligo A281 or 40 μM oligo B139, respectively (Fig. 5c). When co-expression of MurA and MurB was evaluated, addition of 20 μM oligo A281 and 20 μM oligo B139 into CFPS system could totally inhibit the protein expression. Moreover, the subsequent enzyme activity assay analyzed by HPLC revealed that the biosynthesis of UDP-acetylmuramate was eliminated in the presence of oligos (Fig. 6), indicating the inhibition of muramic acid biosynthetic pathway.

Fig. 6
figure 6

Enzyme activity assay of Mur enzymes derived from oligo repressed cell-free expression and related HPLC analysis. a MurA and MurB were co-expressed with 20 μM oligos and 5 ng/μl of each target template; the product UDP-MurNAc could not be detected in the reaction. b MurA and MurB were co-expressed without oligos; UDP-MurNAc was detected in the reaction mixture

Full size image

Discussion

The reconstitution of different metabolic pathways is very important to understand the synthetic biology of microorganisms and manipulate the synthesis of related microbial products. However, in vivo studies of these pathways are hampered by the challenge of multiple genes co-expression. Construction of multi-gene expression plasmids or transformation of more than two compatible plasmids into the host strains has been developed to solve the problem in vivo (Tolia and Joshua-Tor 2006). But the deficiencies such as unsynchronized and imbalanced multi-gene expression, the instability of coexisted multiple plasmids, and the heavy burden of host cells has severely limited the effective reconstitution of metabolic pathways. The open nature and easy accessibility are two important merits of CFPS system, and its synthetic potential for enzyme production has already been exhibited by many paradigms. Therefore, the CFPS system has the potential to be employed to construct some targeted biosynthetic pathways via multi-gene high-throughput expressions. It is reported that one complex channel protein (Matthies et al. 2011) can be successfully produced in E. coli cell-free system via the construction and expression of all nine subunits under one promoter, which encouraged us to express six pathway-related genes to construct the first biosynthetic pathway in the cell-free system. In our study, the biosynthetic pathway of UDP-N-acetylmuramyl pentapeptide, the cytoplasmic precursor of bacterial peptidoglycan, was targeted due to its through-understand enzymatic mechanism and potential application in drug discovery. The preliminary results indicated that the usage of three expression plasmids for co-expression of all six Mur genes was not feasible (data not shown). To solve this problem, linear PCR templates derived from expression plasmids were applied to synthesize the target Mur enzymes in CFPS system. Our results showed that the expression of single gene, or combined multiple genes, or even all six genes, had been accomplished, with preserved enzyme activity to synthesize expected pathway products. The production of important intermediate UDP-MurNAc and final pathway product UDP-N-acetylmuramyl pentapeptide suggested the successful reconstitution of Mur pathway in CFPS system. This successful first example has revealed the great potential of CFPS system as the in vitro expression platform for co-expression of multiple genes and reconstitution of important metabolic pathways.

In vivo, the gene expression could be regulated by many methods, such as using different promoters and RBS sites. But almost all the CFPS systems were constructed on the base of initiating the transcription with T7 RNA polymerase, leaving us limited ways to regulate the gene expression in vitro. Noireaux used a CFPS transcription and translation system to construct a circuit which is analogous to linear electrical amplifiers to synthesize different proteins of interest that require time-delayed or coordinated expression (Noireaux et al. 2003). Furthermore different types of Multi-step reaction cascades such as transcriptional oscillators (Kim and Winfree 2011), strand displacement cascades (Zhang and Seelig 2011), Lotka–Volterra oscillators (Soloveichik et al. 2010), limit-cycle oscillators (Soloveichik et al. 2010), chaotic systems (Soloveichik et al. 2010) and logic AND gates (Takinoue et al. 2008) have achieved to elicit desired functions in vitro. All these efforts have shown us a variety of ways to regulate gene expression in CFPS system. Nevertheless all the methods above were exquisite and complicated themselves, simple approaches were needed to settle the problem of direct and quick regulating the reconstitute pathway. In the case of Antisense oligos which prevent protein translation of certain messenger RNA strands by binding to them, can be used to target a specific, complementary (coding or non-coding) RNA. And the method of rapid screening of antisense oligonucleotides for gene modulation has been established in vitro (Shao et al. 2006). It turned out to be an appropriate way to regulate the reconstituted biosynthetic pathway in CFPS system. Using the rational designed oligos for MurA and MurB, we easily repressed the expression of the target enzymes. Also, the antisense oligos for Mur C, D, E and F have been chosen to severely inhibit the expression of the corresponding enzymes (data not shown), respectively. It is imaginable that the composition of these various oligos with different concentrations could be used to directly regulate the expression of multi-enzymatic genes thus influencing the efficiency of targeted bioproduct synthesis in this In vitro artificial-constructed pathway.

Since Mur pathway is generally existed in bacteria, the related Mur enzymes have been considered as potent antibiotic targets. However, the screening and development of chemical inhibitors for Mur enzyme were not effective, and very few commercial chemogents have been developed by targeting at Mur enzymes. The developed CFPS-oligo assay strategy in our work provides new-type inhibitors for the targeted biosynthetic pathway via rapid screening of antisense oligos, which avoided the obstacle in the in vivo expression system associated with the penetration of oligos through the cell membrane into the cytoplasm. Furthermore, besides the oligos, this system could also be used in the screening of a variety of compounds to inhibit muramic acid biosynthesis by targeting at one artificial in vitro pathway. Given its high-throughput characteristics, such CFPS assay has the potential to develop as an efficient and novel methodology to for discovery of cell wall inhibitors, which are highly imperative to confront the increased bacterial resistance threats.