Four plasmids, pBS-Ikan, pDTW202, pBS-Mkan and pKD-Cre, were constructed to facilitate the chromosomal gene integration and/or gene deletion in E. coli (Figure 2a). Plasmid pBS-Ikan contains the fragment lacIU-loxp-kan-loxp-lacID, which was designed to facilitate the deletion of the lacI gene and the subsequent removal of the inserted kan gene from the chromosome. Plasmid pDTW202 contains the fragment loxpLE-kan-loxpRE, which was designed for the easy removal of the kan gene by recombination of loxpLE and loxpRE. Plasmid pBS-Mkan contains the fragment lpxMU-loxpLE-kan-loxpRE-lpxMD, which was designed to facilitate the deletion of the lpxM gene and the subsequent removal of the kan gene from the chromosome. Plasmid pKD-Cre was designed to remove the inserted kan gene from the chromosome at the later stage of deletion.

Three E. coli mutants, HW000, HW001 and HW002, were constructed by using the above plasmids (Figure 2b). HW000 was constructed by deleting the lacI gene to avoid the lactose regulation. HW001 was constructed by integrating the F. novicida lpxE into the site of lacZ gene in the lac operon of E. coli HW000. HW002 was constructed by deleting the lpxM gene in the chromosome of HW001. The correct insertion or deletion of these genes and the removal of kan in these three E. coli mutants were confirmed by PCR analysis and DNA sequencing.

The growth rates of all these E. coli mutants were similar to that of wild-type W3110 (Figure 3).

Lipid A was isolated from strains of W3110, HW000, HW001 and HW002 and analyzed by TLC (Figure 4). The typical lipid A band was shown on the TLC for samples isolated from W3110 and HW000. Considering the hydrophilic property of phosphate and hydrophobic property of fatty acid chain, lipid A containing less phosphate groups and more and longer fatty acid chains should migrate faster on TLC. A more rapidly migrating substance was observed on the TLC in the sample from HW001; likely MPLA. For the sample from HW002, a substance migrating faster than lipid A, but slower than MPLA, was observed; likely P-MPLA.

Lipid A isolated from W3110, HW000, HW001 and HW002 were further analyzed by ESI/MS (Figure 5). The lipid A isolated from W3110 yielded a major peak at m/z 1796.2 in the spectrum, suggesting that the peak is created by the molecular ion [M − H]⁻ of lipid A. The minor peak at m/z 1716.3 might arise by loss of a phosphate, indicating that the hydrolysis conditions for the cleavage of the lipid A from the core-oligosaccharide was too harsh and resulted in a partial loss of the α-anomeric C-1-phosphate at GlcN(I). Lipid A isolated from HW000 showed a similar pattern as that from W3110. In the spectrum of lipid A isolated from HW001, there is only one major peak at m/z 1716.3, suggesting that the lipid A in HW001 is completely dephosphorylated at the 1-position. In the spectrum of lipid A isolated from HW002, there is only one major peak at m/z 1506.1, suggesting that lipid A isolated from HW002 is P-MPLA. Both TLC and ESI/MS data indicate that HW001 can convert all lipid A into MPLA and HW002 can convert all lipid A into P-MPLA. This indicates that the gene lpxE inserted in the chromosome of HW001 and HW002 was well expressed and LpxE can modify the structure of lipid A in vivo and that lpxM encodes the only late acyltransferase in E. coli that adds a secondary acyl chain to the 3′-position of lipid A.

LPS was purified from W3110, HW001 and HW002 and used to stimulate macrophage cell line RAW264.7. Levels of TNF-α released by the cells after stimulation were examined (Figure 6). E. coli O111:B4 LPS was used as a positive control, and PBS (phosphate-buffered saline) was used as a negative control.

The levels of TNF-α released by RAW264.7 cells were proportional to the concentration of LPS up to 100 ng/mL, but there was little difference between LPS concentrations of 100 ng/mL and 1000 ng/mL. At all concentrations, lower levels of TNF-α were released when RAW264.7 cells were stimulated by HW001 LPS and HW002 LPS than by the controls O111:B4 LPS or W3110 LPS (Figure 6). The lowest levels of TNF-α were induced by HW002 LPS from 17,472 pg/mL at 1 ng/mL to 37,578 pg/mL at 1000 ng/mL, which were 39% and 55% of the levels induced by W3110 LPS, respectively.

These results demonstrated that LPS from HW001 and HW002 induced lower secretion of TNF-α than that of wild-type E. coli LPS, suggesting they are good candidates for vaccine adjuvant development. The changes in the number of the acyl chains and the phosphate groups of lipid A might decrease the binding ability by TLR4/MD-2 and lead to the lower release of TNF-α.

Restriction enzymes, shrimp alkaline phosphatase, T4 DNA ligase and DNA ladder were purchased from Sangon (Shanghai, China). Plasmid DNA was prepared by using the EZ-10 spin column plasmid mini-preps kit from Bio Basic Inc. (Markham, Canada). PCR reaction mixtures (50 μL) used in this study contain 5 μL 10× Ex Taq buffer, 4 μL dNTP mixture (2.5 mM each), 1 μL plasmid template (100 ng/μL), 1 μL forward primer (20 μM), 1 μL reverse primer (20 μM) and 0.5 μL TaKaRa Ex Taq DNA polymerase. The PCR reaction was first heated to 94 °C and maintained for 10 min, followed by 35 cycles of denaturation (94 °C for 30 s), annealing (30 s at 55 °C or 58 °C) and extension (72 °C for 3 min). At the end, an additional 10 min incubation at 72 °C was used. PCR products were purified by using the TIANgel midi purification kit from Tiangen (Beijing, China). Primer synthesis and DNA sequencing were performed by Sangon. The sequences of primers used in this study are listed in Table 1.

All bacterial strains and plasmids used in this study are listed in Table 2. Bacteria were usually grown at 37 °C in LB media (10 g/L trypton, 5 g/L yeast extract, 10 g/L NaCl). When necessary, antibiotics were added with a final concentration of 100 μg/mL for ampicillin, 30 μg/mL for kanamycin and 30 μg/mL for chloramphenicol.

Four plasmids pBS-Ikan, pDTW202, pBS-Mkan and pKD-Cre were constructed in this study (Figure 2a). Plasmid pBS-Ikan was constructed by inserting a DNA fragment, lacIU-loxp-kan-loxp-lacID, in the vector pBlueScript II SK+. Primers lacIU-F and lacIU-R were used to amplify the upstream fragment of lacI; primers lacID-F and lacID-R were used to amplify the downstream fragment of lacI; primers kan-loxP-F and kan-loxP-R were used to amplify the fragment loxp-kan-loxp from pMG2-kan^r [20]. Plasmid pDTW202 was constructed by inserting a fragment, loxpLE-kan-loxpRE, in the vector pBlueScript II SK+. Fragments loxpLE and loxpRE were obtained by the annealing of the primers loxpLE-F, loxpLE-R and loxpRE-F, loxpRE-R, respectively, according to the method of Arakawa [21], and the kanamycin resistance gene kan was PCR amplified from pK18mobsacB [22] with primers kan-F and kan-R. Plasmid pBS-Mkan was constructed by inserting a fragment, lpxMU-loxpLE-kan-loxpRE-lpxMD, in the vector pBlueScript II SK+. Primers lpxMU-F and lpxMU-R were used to amplify the upstream fragment of lpxM; primers lpxMD-F and lpxMD-R were used to amplify the downstream fragment of lpxM, and primers kan-loxpLE-F and kan-loxpRE-R were used to amplify the fragment loxpLE-kan-loxpRE from pDTW202. Plasmid pKD-Cre was constructed by replacing the λ Red genes with the gene cre in the vector pKD46 [23]. The gene cre was amplified from the vector pSH47 [24], using primers cre-F and cre-R, digested with SacI and NcoI and ligated with the vector pKD46, which is similarly digested. The gene cre was put under the arabinose-inducible ParaB promoter. The plasmid pKD-Cre has a temperature sensitive replicon.

Three E. coli mutants, HW000, HW001 and HW002, were constructed by deleting and/or integrating genes in the chromosome of E. coli (Figure 2b). Firstly, the DNA fragment lacIU-loxp-kan-loxp-lacID was amplified from pBS-Ikan using primers lacIU-F and lacID-R and transformed into E. coli W3110/pKD46. Under the help of Red enzymes expressed by pKD46, the lacI gene in the chromosome should be replaced by DNA fragment loxp-kan-loxp; the transformants were selected on LB agar plates containing kanamycin. The kan gene inserted in the chromosome was further eliminated by transforming pKD-Cre. Plasmid pKD-Cre expresses the Cre recombinase, which directly acts on the repeated loxp (Cre recognition target) sites flanking the kan gene and excises the kan gene, resulting in the lacI mutant HW000. Secondly, the DNA fragment lacZU-lpxE-FRT-kan-FRT-lacZD was amplified from pWSK29-FnlpxE-Fkan [19] using primers lacZ-F and lacZ-R and transformed into HW000/pKD46. Under the help of Red enzymes expressed by pKD46, the lacZ gene was broken by the inserted DNA fragment lpxE-FRT-kan-FRT. The kan gene was then eliminated by transforming pCP20. Plasmid pCP20 expresses the FLP recombinase, which acts on the directly repeated FRT (FLP recognition target) sites flanking the kan gene and excises the kan gene [23], resulting in the double mutant HW001. Thirdly, the DNA fragment lpxMU-loxpLE-kan-loxpRE-lpxMD was amplified from pBS-Mkan using primers lpxMU-F and lpxMD-R and transformed into HW001. Under the help of Red enzymes expressed by pKD46, the lpxM gene in the chromosome was replaced by the DNA fragment loxpLE-kan-loxpRE. The kan gene was then eliminated by transforming pKD-Cre, resulting in the triple mutant HW002. Since the short sequences loxp left by deleting lacI, loxpLR left by deleting lpxM and FRT left by inserting lpxE are different, they could not recombine with each other in the double or triple mutants.

Typically, 200 mL cultures were inoculated from overnight cultures and grown to an OD₆₀₀ of 1.0. The cells were harvested by centrifugation at 8000 rpm for 20 min and washed with PBS. The final cell pellet was resuspended in a single phase Bligh-Dyer mixture [25] (76 mL) consisting of chloroform/methanol/water (1:2:0.8 v/v) and agitated by stirring for 1 h at room temperature. The cell debris was collected by centrifugation at 2000 rpm for 20 min, and the insoluble material was washed twice with the same solvent. The pellets were resuspended in 27 mL 12.5 mM sodium acetate (pH 4.5) by sonication and heated at 100 °C for 30 min to release lipid A from LPS. After cooling to room temperature, the suspensions were converted to two-phase Bligh-Dyer mixtures consisting of chloroform/methanol/water (2:2:1.8 v/v) by adding 30 mL chloroform and 30 mL methanol. The mixture was vigorously shaken to extract lipid A and centrifuged at 2000 rpm for 20 min. The lower phase containing lipid A was dried with a rotary evaporator and stored at −20 °C.

The dried lipid A was dissolved in chloroform/methanol (4:1 v/v) and spotted onto a silica gel 60 TLC plate. The plate was developed in the solvent of chloroform/methanol/water/ammonia (40:25:4:2 v/v/v/v). After drying and spraying with 10% sulfuric acid in ethanol, the separation of lipid A on the plate could be visualized by charring at 175 °C.

For ESI/MS analysis, dried lipid A was dissolved in chloroform/methanol (4:1 v/v) and subjected to ESI/MS in the negative ion mode. The mass spectra were acquired on a Waters SYNAPT Q-TOF mass spectrometer equipped with an ESI source. Data acquisition and analysis were performed using MassLynx V4.1 software [26].

LPS were isolated from E. coli strains using the phenol-water procedure [27]. Wet E. coli cells (3 g) were suspended in water (10 mL) at 68 °C, and 90% aqueous phenol (10 mL) was added. The mixture was stirred for 1 h at 68 °C, then cooled to 4 °C and centrifuged at 4000 rpm for 20 min to separate the water and phenol phases. The water phase was collected, dialyzed against water and lyophilized. The crude LPS (0.5 g) was resuspended in 100 mM Tris-HCl (pH 7.5, 10 mL) and then by treating with DNase I, RNase A at the concentration of 100 and 50 μg/mL at 37 °C for 2 h, respectively. Then 100 μg/mL proteinase K at 37 °C for 2 h was performed. For further purification, 5 mL water-saturated phenol was added to the crude LPS sample. The mixture was centrifuged at 4000 rpm for 20 min. The water phase was collected, dialyzed and lyophilized. Phospholipid contaminants in the sample were removed by extraction with chloroform/methanol (2:1 v/v) [28]. Purified LPS product was lyophilized and stored at 4 °C [29].

E. coli O111:B4 LPS was purchased from Sigma-Aldrich. The murine macrophage cell line RAW264.7 was obtained from the cell bank of the Chinese Academy of Science (Shanghai, China). It was maintained at 37 °C with 5% CO₂ in DMEM (Gibco BRL, USA) containing 10% FBS (Hyclone Logan, UT, USA), supplemented with penicillin (100 U/mL) and streptomycin (100 μg/mL). Cells were cultured at 10⁵/well in 96-well plates with 200 μL medium each well. After 18 h, the RAW264.7 cells were stimulated with increasing concentration of LPS (1, 10, 100 and 1000 ng/mL). 24 h later, culture supernatants were collected, centrifuged to remove the contaminates and stored at −80 °C for cytokine analysis. Concentrations TNF-α were determined using the enzyme-linked immunosorbent assay kit (eBioscience), according to the manufacturer’s instruction.

One-way analysis of variance was performed to determine the statistical significance of the differences between mean values for various experimental and control groups. Data were expressed as means ± standard deviation, and the experiments were performed in triplicate. The means were compared using the least significant difference test. p < 0.05 was considered a significant difference and p < 0.01 was considered extremely significant difference. All data were analyzed with SPSS Statistics 17.0.