Ampicillin (Amp), streptomycin (Stm), spectinomycin (Spm), and kanamycin (Km)-resistant (^R) transformants were selected on a Luria Bertani (LB) agar medium (Kemin Bio Tech., Shanghai, China), containing the antibiotics at 100, 100, 50, and 25 µg/ml, respectively. All antibiotics were purchased from Sigma-Aldrich (St. Louis, MO, USA). DAP (Sigma-Aldrich) was used at 100 µg/ml, and L-arabinose (Sigma-Aldrich) was used at 0.4%. The oligonucleotides were purchased from Beijing Sunbiotech Co., Ltd. (Beijing, China). The restriction enzymes were obtained from New England Biolabs Inc. (Ipswich, MA, USA), and the Taq Plus DNA polymerase was purchased from Takara Biotechnology Co., Ltd (Dalian, China).

The DH10B (F-mcrA Δ(mrr-hsdRMS-mcrBC) Φ80lacZΔM15ΔlacX74 recA1endA1 araD139 Δ(ara leu) 7,697 galU galK rpsL nupG λ-) E. coli strain was purchased from Invitrogen Life Technologies (Carlsbad, CA, USA), and was maintained in LB medium. The pKD46 and pKan-kil plasmids were provided from Dr. Shanhu Li (China Academy of Military Medical Science), and were previously described (16), and the pDV-cmv-gene plasmid was constructed in our laboratory (unpublished data; China Institute for Food and Drug control).

A Red recombinase-based homologous recombination technique was used to knockout the chromosomal asd in the DH10B cells, as reported previously (13,20). The details of the asd knockout are presented in Fig. 1. Briefly, transformants carrying a pKD46 Red helper plasmid were induced in 5 ml super optimal broth (Hope Bio Tech., Qingdao, China) cultures with Amp and L-arabinose at 30°C. The plasmid was subsequently transformed with linear DNA fragments carrying kan-kil and asd-homologous extensions using a Gene Pulser Xcell Electroporation system with 0.1 cm chambers according to the manufacturer's instructions (Bio-Rad Laboratories Inc., Hercules, CA, USA). The fragment was obtained by polymerase chain reaction (PCR), using Taq Plus DNA polymerases with the following primer sequences: Forward, 5′-ATT TAT ACA GCA CAC ATC TTT GCA GGA AAA AAA CGC TTA TTA GTG AAT GCT TTT GCT TG-3′; and reverse, 5′-CGC ACT AAC AGG GGC GGC ATC GCG CCC CAG ATT TAA TGA ACA ACC TCC TTA GTA CAT GC-3′, and pkan-kil as a template. The electroporated cells (1×10⁹−2×10⁹ CFU/ml) were plated out to select for the Km^R and DAP-dependent transformants, and subsequently colony-purified at 37°C, prior to being tested for Amp sensitivity to isolate the cells that had lost the helper plasmid. The helper plasmid cells were then made competent again and received another round of electroporation with the linear fragments carrying the homologous extensions without the kan-kil region. The fragment was obtained by PCR using Taq Plus DNA polymerases with the following primer (Sun Bio Tech, Beijing, China) sequences: Forward, 5′-CGG CAC ATT TAT ACA GCA CAC ATC TTT GCA GGA AAA AAA CGC TTT CAT TAA ATC TGG GG-3′; and reverse, 5′-CTG CGC TTA CTC CTG TAT TAC GCA CTA ACA GGG GCG GCA TCG CGC CCC AGA TTT AAT GA-3′. The PCR mixture was comprised of 12.5 µl Q5 High-Fidelity 2X Master mix (Takara Biotechnology Co., Ltd), 10 µM forward primer (1.25 µl), 10 µM reverse primer (1.25 µl), 1 µl template DNA and made up to 25 µl with nuclease-free water. The transformants were subsequently induced at 37°C for kil expression and homologous recombination, in order to remove the kan-kil fragment that was previously integrated into the E. coli genome (Fig. 1).

PCR was used to confirm that all mutants had the correct structure. A freshly isolated colony was suspended in 50 µl water, from which 3 µl aliquots were used in separate 25 µl PCR reaction mixtures following a 2 min pre-incubation 'hot start' at 95°C. The primers (Sun Bio Tech) used were as follows: Forward, 5′-ATT TAT ACA GCA CAC ATC TTT GCA GGA AAAAAA CGC TTA TTA GTG AAT GCT TTT GCT TG-3′; and reverse, 5′-CGC ACT AAC AGG GGC GGC ATC GCG CCC CAG ATT TAA TGA ACA ACC TCC TTA GTA CAT GC-3′. Control colonies were tested side-by-side, and the amplified products were analyzed by agarose gel electrophoresis.

Detailed steps for the construction of the plasmid are presented in Fig. 2. Briefly, the DNA sequence of the entire expression cassette, including the nirB15 promoter (21), the Shine-Dalgarno sequence, the asd coding region (Salmonella enterica subsp. enterica serovar typhi str. Ty2, GenBank: AAO71451.1), and the terminator was chemically synthesized to incorporate the appropriate restriction sites (Genewiz, Inc., South Plainfield, NJ, USA). The sequence was ligated into the pMD 18-T-simple plasmid (Sun Bio Tech) and confirmed by sequencing at the junction regions, prior to double digestion with KpnI and BamHI, in order to release the expression cassette. The released fragment was ligated to the similarly digested pDV-cmv-gene plasmid in order to generate pDV-cmv-gene-nirb-asd Km⁺. The Km resistance gene (Km⁺) in the resultant plasmid was then removed by digestion with HindIII and XhoI, and the purified plasmid was ligated by annealing a dsDNA fragment carrying complementary restriction sites to the linearized plasmid prior to ligation. The sequences of the dsDNA are as follows: Sense, 5′-TCG AGT AGA TAA CTG ACT AGC ATA ACC CCT TGG GGC CTC TAA ACG GGT CTT GAG GGG TTT TTT A-3′; and antisense, 5′-AGC TTA AAA AAC CCC TCA AGA CCC GTT TAG AGG CCC CAA GGG GTT ATG CTA GTC AGT TAT CTA C-3′. The ligation products were subsequently transformed into the DH10BΔasd cells and spread onto LB agar plates to recover DAP-independent colonies, prior to being further colony tested on LB agar plates for Km-sensitivity, in order to identify any potential antibiotic resistant-marker-free vectors.

DH10BΔasd cells were streaked onto agar plates containing LB, DAP⁺LB, Stm⁺DAP⁺LB, and Spm⁺DAP⁺LB, and their growth was subsequently recorded after overnight culture at 37°C.

To test the autolysis of DH10BΔasd cells the cells were grown on DAP⁺LB agar plates, prior to being washed three times with sterile distilled water, and were then resuspended in sterile distilled water. Three types of media containing: Water only, LB, and LB + Amp, were inoculated with 100 µl aliquots prior to culture at 37°C. Subsequently, 10 µl aliquots were taken from each culture at various times, and plated onto LB agar medium, prior to overnight incubation at 37°C. For each assessment three colony plates were counted to determine the number of colonies.

To verify the structure of the expected plasmid, purified pDV-cmv-gene-nirb-asd Km⁻ DNA was digested in three double enzyme reactions: SalI + BglII, HindIII + BglII, and HindIII + SalI. The digested products were subsequently analyzed by agarose gel electrophoresis. The selected clones were fully sequenced on both strands at the Beijing Genomics Institute (Beijing, China).

The DH10BΔasd cells containing pDV-cmv-gene-nirb-asd Km⁺ and pDV-cmv-gene-nirb-asd Km⁻ were cultured overnight at 37°C with agitation at 150 rpm, in liquid LB medium with or without Km (100 µg/ml). The cells were then harvested by centrifugation and diluted with 0.9% NaCl solution with the same OD₆₀₀ values for both cultures (OD=2.0). Plasmid DNA was prepared from 2 ml aliquots of the cells (1×10⁹−2×10⁹ CFU/ml) using the Qiagen Miniprep kit (Qiagen, Inc., Valencia, CA, USA) and quantified using a UV spectrometer (DU800; Beckman Coulter, Inc., Fullerton, CA, USA) at a wavelength of 260 nm. The plasmids were subsequently separated on 1% agarose gel, and the resulting images were scanned and analyzed to determine the amount of DNA in the supercoiled conformation using MultiGauge (Version 3.1, Fujifilm, Tokyo, Japan).

After 10, 20 and 30 passages, cells from the rapidly growing cultures were taken and inoculated in LB medium with inoculant-medium ratios of 1:1,000, 1:100 and 1:10. The cells were subsequently cultured for 5 h at 37°C with 150 rpm agitation. LB agar and DAP⁺ LB agar plates were then spread with 100 µl aliquots from each of the cultures prior to further overnight culture at 37°C. The number of colonies on both plates were counted and used to calculate the percentage of the bacterial cells that had retained the plasmid and could grow normally in LB medium.

Aliquots of 10 µl of cells were removed following 0, 1.5, 3 and 4.5 h cultures in LB broth with 200 rpm agitation at 37°C. The cells were subsequently diluted 100 times, spread onto LB agar plates, and cultured at 37°C. Following overnight growth, the colonies were counted and used to calculate the cell doubling time using the Gompert function (F = K·a^b·t) (22).

DH10BΔ(asd pDV-cmv-gene-nirb-asd km⁻) cells were stored as glycerol stock at 25, 4, −20 and −70°C for 12 months and their survival rate was subsequently tested by plating 100 µl of the thawed cells onto LB agar plates, prior to counting the number of colonies formed after overnight culture.

The egfp gene was ligated into the pDV-cmv-gene-nirb-asd Km⁻ and pDV-cmv-gene-nirb-asdKm-plasmids (Fig. 3), and the resulting transformants containing the egfp gene were selected by PCR, and sequenced using egfp gene-specific primers (Sun Bio Tech). Based on the concentration and molecular mass, the copies of the two types of plasmid were adjusted to the same order of magnitude. The lipofection of the HeLa cells with the two plasmids was carried out using Lipofectamine^® reagent (Invitrogen Life Technologies), according to the manufacturer's instructions. ddH₂O was used as a control. The egfp gene was also ligated into the pDV-cmv-gene-nirb-asd and pDV-cmv-gene plasmids, which were then transformed into the DH10BAΔasd E. coli strain. DH10BΔasd cells were streaked onto agar plates containing LB medium, and their growth was recorded after overnight culture at 37°C.

The data were analyzed using SPSS version 19.0 (IBM SPSS, Armonk, NY, USA), and P<0.05 was considered to indicate a statistically significant difference. Curve fitting was conducted using OriginPro v7.0 (OriginPro, Northampton, MA, USA).

Following several preliminary gene disruption experiments, namely the use of the pKD46 plasmid to provide Red recombinase and FRT-flanked Km resistance gene fragments carrying the homologous extensions to asd, hundreds of colonies were obtained that could grow on LB medium supplemented with Km and DAP. Thus indicating that the Km gene was successfully inserted into the DH10B genome. The Km-resistant cells were subjected to a second round of recombination by electroporating a linear DNA fragment containing asd-homologous extensions without the kan-kil region, in order to remove the kan-kil fragment that had previously been inserted into the bacterial genome. Following several experiments used to comparatively screen the transformants on LB agar plates containing Km⁺ DAP⁺, DAP⁺, and LB only, a few colonies were obtained that could grow on DAP⁺ LB, but not on Km⁺ DAP⁺ LB, or LB. These results indicate that the cells in these colonies had lost resistance to Km as a result of the secondary recombination, and were dependent on the exogenous supply of DAP. This is due to the expected disruption of asd in the strain. The colonies were therefore selected as candidates for asd-deficient lines for further confirmation at the DNA level.

asd locus-specific PCRs were performed using genomic DNA extracted from several candidate template colonies. As shown in Fig. 4, a band corresponding to the expected 77 bp product was observed in the asd deletion candidate colony, whereas a 1,184 bp PCR product was amplified with DNA from the wildtype DH10B cells, indicating that asd-specific disruption had occurred in the cell line. Therefore, the cell line was named DH10BΔasd.

Deletion or disruption of functional asd in E. coli cells is expected to result in the failure of the bacterium to synthesize a functional cell wall, due to the disruption of DAP synthesis, thus resulting in cell death. To acquire a better understanding regarding how cell death occurred in the asd-disrupted cells, cell survival duration was determined in various growth conditions by exposing them to water and LB media, with or without Amp, for various durations. As demonstrated in Fig. 5, the cells gradually lost their viability following exposure to these media, but at different rates. The half-life of the cells was ~2–3 h and ≤1 day in the LB medium with and without Amp, respectively. In distilled water, the half-life of the cells was 3–4 days.

The pDV cmv gene was used as a starter plasmid for construction of an antibiotic resistance-marker-free vector. Following several steps of DNA manipulation, as described in the Materials and Methods, the DH10BΔasd cell colonies that contained the target plasmid pDV-cmv-gene-nirb-asd Km⁻ were recovered, since they could grow in the LB medium without requiring extra supplementation of DAP. The nirB-asd cassette was expected to provide ASD activity in the DH10BΔasd cells. To verify the structure of the plasmid, DNA was prepared from a number of colonies and was restricted using restriction enzymes in three reactions: SalI + BglII, HindIII + BglII, and HindIII + SalI. The digested products were analyzed and plasmids with the expected bands and sizes were identified (Fig. 6). Sequencing of the plasmids further confirmed that the entire sequences were correct (data not shown).

To investigate whether the elimination of antibiotic pressure during the plasmid amplification process would have an impact on the plasmid yields in this symbiotic system, the plasmid yield of the resultant constructs were compared prior to and following the removal of Km⁺ in the backbone, in LB medium with or without Km in the DH10BΔasd cells. The results demonstrated that the plasmid yields were similar in the two culture conditions (0.527, vs. 0.552 µg/ml of culture in LB with or without Km, respectively), and did not statistically significantly differ from each other (P>0.05). The examination of the conformation of the plasmids extracted from the cultures under the two conditions demonstrated a similar percentage (~60%) of supercoiled plasmid DNA (data not shown).

Differing from the typical plasmid amplification protocols, where antibiotics are added to the culture media in order to selectively proliferate the bacterial cells that contain plasmids bearing antibiotic resistant genes, maintenance of the plasmids in this symbiotic system relies solely on DAP production from the host plasmid containing an asd expression cassette located in the asd-disrupted cells. It is therefore important to investigate how well the plasmid can be retained under the symbiotic, but non-selective condition. The cells containing pDV-cmv-gene-nirb-asd Km⁻ were grown to various passages in LB medium, and were then plated out onto LB agar plates with or without DAP after diluting the culture up to 1,000 times. The colonies that appeared in LB medium were considered to contain the plasmid, whereas those that appeared on the LB + DAP medium were considered to be either with or without plasmid. The results shown in Table I indicate that the plasmid retention rates (ratio of colony counts on medium without and with DAP) were statistically the same (100, vs. 99%) between the 10th and the 30th passages when they were inoculated at a 1:1,000 dilution. However, the plasmid retention rates were reduced to ~80% when inoculated at a 1:10 dilution.

As a cell line deficient in cell wall composition, it is important to analyze bacterial storability at various temperatures for its further use in a plasmid production system. The results summarized in Table II demonstrated that the cells rapidly lost their viability when stored at higher temperatures (25° and 4°C), but maintained stable colony forming ability when stored at −70°C. The half-life of the cells was 2 and 15 days when the cells were stored at 25°C and 4°C, respectively, and increased to 2 months when stored at −20°C, and to 10 years when stored at −70°C.

To examine the growth speed that would impact overall plasmid productivity, the present study determined the number of cells in the culture by counting their colony forming units at various time points. The doubling time computed based on the compertz function, was 29.9 min when the cells were grown in LB medium.

The expression of enhanced green fluorescence protein (EGFP) is shown in Fig. 7. The cells transfected with the pDV-cmv-egfp-nirb-asd Km⁻ plasmid exhibited green fluorescence under blue light, indicating a good expression of EGFP protein in the cells. As assessed with the naked eye, the efficiency of transfection with the two types of plasmid appeared similar. The DH10BΔasd cells with the pDV-cmv-egfp-nirb-asd gene grew well on LB medium, whereas the cells transformed with the pDV-cmv-egfp gene did not, which is indicative of a normally functioning asd gene following the insertion of egfp.