Introduction
Genetic modification technology was developed to overcome the limitations of traditional breeding techniques, and it involves purpose-specific manipulation of genes using the latest molecular biology tools. Organisms developed through genetic modification are commonly referred to as genetically modified organisms (GMOs) or living modified organisms (LMOs) in legal terms (Lee and Suh, 2011). GMO technology has played a vital role in industrial development in the 21st century, and has been applied in several bioindustries including medicine, chemistry, environment, food, energy, agriculture, and fisheries (Choi, 2020).
GMO technology has been most successfully utilized in the development of GM crops. As of 2019, GM crops were grown on 190.4 million hectares in 29 countries. Furthermore, since the beginning of commercialization of these crops in 1994, their accumulated cultivation area has reached 2.7 billion hectares (ISAAA, 2019). Although GM crop cultivation is not currently allowed in Korea, GM crops are imported and used for food and feed purposes, and their domestic import totaled 11.45 million tons and amounted to approximately $3.4 billion as of 2021 (KBCH, 2022a).
Industrial GMOs are another widely used commercial application of GMO technology, and are defined as “Genetically modified organisms used in industries such as the textile industry, the machine industry, chemical industry, electronics industry, the energy industry, and the resource industry excluding genetically modified organisms for testing, research, agriculture, livestock, health care, environmental purification, or marine and fisheries.” by Article 1-2 Subparagraph 5 of the Act on the Movement of Genetically Modified Organisms (LMO Act). A research report from the Korean Society for Biotechnology and Bioengineering (KSBB, 2020) described active research underway worldwide on industrial GM microorganisms, for example, in the production of 1,3-propanediol, polylactic acid, a biodegradable polymer, succinic acid and 1,4-butanediol.
The industrialization of GM microorganisms in Korea is being developed and while it is lower than other leading nations, research and development are underway in several biotechnology companies and universities to generate commercially available host strains (KSBB, 2020). For example, an energy company is using GM microorganisms to develop a commercialized production process for 2,3-butanediol. Moreover, a food company has developed lactic acid-producing GM yeast strains and operates commercial production of lysine and glutamic acid using Corynebacterium glutamicum as host cells. Additionally, research is actively underway in several universities to establish GM microorganisms such as GM yeast, which generates 2,3-butanediol, and GM Escherichia coli which yields retinol, a wrinkle-improving bio-material. While no GM crops have been approved for commercialization, six industrial GMOs have been commercialized, including five microorganisms and one plant cell line (KBCH, 2022b). In addition, seven GM microorganisms were approved and commercialized for food products, including C. glutamicum, which produces the enzyme, D-fructose-4-isotropic.
Recently, as the worldwide production and industrial use of GMOs has gained momentum with an increase in the number of reported LMO research facilities, the demand for safety and safety management for using and developing GMOs has also increased. In Korea, the Ministry of Trade, Industry, and Energy is in charge of industrial GMO safety management according to Article 2, No. 5 of the LMO Act. The safety management is carried out by dividing GMOs into two groups: GMOs that are released into the natural environment (open system) and GMOs that are used in production processes (closed system) (KBCH, 2022b). Since GM plants, animals, and microorganisms are produced in manufacturing facilities, preventing their leakage into the external environment is one of the most critical issues in safety management. It is crucial to monitor possible external outflows that may occur during production, transportation, disposal, and wastewater treatment, as GM microorganisms are difficult to detect by visual inspection.
Qualitative polymerase chain reaction (PCR) methods are generally used to detect GM microorganisms, but real-time quantitative PCR methods are used for quantitative analysis (Ali et al., 2012; Dreo et al., 2014; Kim et al., 2016). Moreover, direct PCR can be used to directly quantify microorganisms in a high-throughput manner by acquiring template DNA directly from the organism, thus reducing the cost and time by bypassing the DNA extraction step (Park et al., 2021). Direct PCR analysis methods have been developed and reported in various species, including bacteria (Lee et al., 2009b; Ben-Amar et al., 2017), algae (Chen et al., 2019), fungi (Walch et al., 2016), and viruses (Choo and Kim, 2006). However, direct PCR assays can result in false positives due to DNA amplification from dead cells or ejected DNA in addition to living cells (Rudi et al., 2005). Noke et al. developed a method to circumvent this issue by treating samples with propidium monoazide (PMA) dyes (Nocker et al., 2006), which penetrate membrane-damaged or dead cells, but not living cells. Thus, PMA only binds to ejected DNA or DNA templates from membrane-damaged or dead cells and inhibits their PCR amplification. As a result, PMA treatment enables selective amplification of DNA templates from living cells during direct PCR (Nocker et al., 2006; Zhao et al., 2013).
Relatively few studies have reported the monitoring of environmental emissions of GM microorganisms, which have been described for crops such as corn (Lee et al., 2009a; Han et al., 2014; 2015), soybeans (Lee et al., 2009a), canola (Lee et al., 2007; Lee, 2020), and grass (Lee et al., 2014). Therefore, this study aimed to establish safety management protocols for monitoring the environmental emissions of GM microorganisms. Here, we describe a new method for the qualitative and quantitative monitoring of E. coli in environmental emissions by employing direct PCR combined with PMA treatment.
Materials and Methods
Escherichia coli culture and cell suspension preparation
A modified pGAL-HIR525 (6,593 bp) plasmid harboring an ampicillin-resistance gene (AmpR), was transformed into E. coli DH5α cells. A single transformed colony was inoculated into 10 mL Luria-Bertani (LB) broth (BD DifcoTM, Franklin Lakes, NJ, USA) supplemented with 50 µg·mL-1 of ampicillin and grown overnight (18 h) at 37℃ with 150 rpm of agitation. The optical density at 600 nm (OD600) was determined in triplicate samples, and 1 mL of O/N (18 h) cell culture was used for serial cell dilutions. A 10-1 diluted cell suspension was prepared by adding 100 µL of cell culture to 900 µL of LB broth. Further diluted cell suspensions of 10-2 to 10-8 were prepared using the same method. All dilutions were performed on ice.
Primer design for PCR and TaqMan real-time PCR (qPCR) assay
Two primer sets specific for the ampicillin resistance gene (AmpR) and the E. coli taxon-specific endogenous D-1-deoxyxylulose 5-phosphate synthase gene (dxs, GenBank accession number AF035440; Lee et al., 2006) were used. For dual-plex qPCR, dual-labeled hydrolysis probes of 5′-FAM/3′-BHQ1 for AmpR and 5′-HEX/3′-BHQ1 for dxs were designed as shown in Table 1.
Table 1. Primers and probes targeting the ampicillin resistance gene, AmpR, and the taxon-specific, single
copy gene, D-1-deoxyxylulose 5-phosphate synthase gene, dxs, in E. coli.
|
DNA standard curve production
E. coli genomic DNA and pGAL-HIR525 plasmid DNA were extracted using Wizard® Genomic DNA purification Kit (Promega, Madison, WI, USA) and Wizard® Plus Minipreps DNA purification system (Promega, Madison, WI, USA), respectively, following the kit instructions. DNA quality and concentration were confirmed through DNA electrophoresis using 1.5% agarose gels and a NanoDrop spectrophotometer (MicroDigital Co., Ltd., Seongnam, Korea), respectively.
Standard curves of AmpR and dxs were constructed using a 10-fold serial dilution series ranging from 1 × 103 to 1 × 108 copies·µL-1 of genomic (4,534,037 bp) and plasmid DNA (6,593 bp), respectively. Copy numbers·µL-1 were calculated from genomic and plasmid DNA concentrations using the ratio of (6.02 × 1023 [copy·moL-1] × DNA amount [g]) to (DNA length [bp] × 660 [g·moL-1·bp-1]) as described by Whelan et al. (2003). For the AmpR gene, single-plex qPCR was performed using serially diluted plasmid DNA to produce AmpR standard curve. However, the standard curve for dxs was produced by dual-plex qPCR using serial diluted genomic DNA. A total of two intra-runs with three repeats of each qPCR run were performed and mean quantification cycle (Cq) values of each diluted DNA were used for producing standard curves for AmpR and dxs. The Cq values were plotted against the logarithm of their initial template copy numbers. Each standard curve was generated using linear regression of the plotted points. PCR amplification efficiency (E) was calculated from the slope of each standard curve following the equation E = 10-1/slope - 1 (Rasmussen, 2001). Relative standard deviation (RSD%) and bias% were calculated to evaluate qPCR precision and trueness of both AmpR and dxs. Bias% was calculated using the following formula: bias% = (mean value - true value)/true value*100.
PCR and dual-plex qPCR analysis
PCR was performed using SapphireAmp® fast PCR Master mix (Takara Bio, San Jose, CA, USA) in a total volume of 25 µL, which contained 1 µL diluted cell suspension (on ice), 10 pmol each of the forward and reverse primers, and 12.5 µL master mix. Nuclease-free water (SDW) was used as the non-template control. The following amplification conditions were used: pre-denaturation at 95℃ for 5 min and 30 cycles of 95℃ for 15 s, 60℃ for 10 s, and 72℃ for 15 s. PCR products were confirmed on a 1.5% agarose gel.
Dual-plex qPCR was performed on a StepOneTM Real-time PCR System (Applied Biosystems, Foster City, CA, USA). Reactions were carried out in a final reaction volume of 20 µL and consisted of 10 µL of TOPreal™ qPCR 2X PreMIX (TaqMan Probe for multiplex, Enzynomics, Daejeon, Korea), 10 pmol of two sets of primers/probes, and 1 µL cell suspension (on ice) or 1 µL DNA. SDW and LB broth were used as negative controls. qPCR was performed under the following conditions: 95℃ for 10 min and 40 cycles of 95℃ for 30 s and 60℃ for 40 s. Target DNA abundance was calculated using Cq values, which correspond to the number of cycles required to reach a systematic default threshold value. A total of 10 colonies were cultured, and 10 qPCR intra-runs with three repeats of each qPCR run were used for qPCR assays. Relative standard deviation (RSD%) was calculated to evaluate qPCR precision of both AmpR and dxs.
PMA treatment for viable cell quantification
Each 450 µL of serially diluted E. coli cell suspension was prepared and treated with 1 µL of 20 mM PMA (PMAxx™ Dye, Biotium Inc., Fremont, CA, USA) and was incubated for 5 min in the dark (with shaking). The samples were exposed on PMA-Lite™ LED Photolysis Device (Biotium Inc., Fremont, CA, USA) for 15 min following manufacturer’s instructions. Each 1 µL PMA-treated cell suspension was directly used for qPCR assay. PMA-qPCR assays were performed in five intra-runs with one qPCR run consisting of two repeats of PMA-treated dilutions and one repeat of non-PMA treated control dilution. For sensitivity determinations, four PMA-qPCR runs with three repeats of each run were additionally performed using diluted cell suspensions from 10-3 to 10-6 dilution. The limit of detection (LOD95%) was determined using the Quodata web application (Grohmann et al., 2016; https://quodata.de/content/validation-qualitative-pcr-methods-single-laboratory).
Determination of colony forming unit (CFU) using qPCR and plate counting
Cell count·µL-1 of E. coli was estimated using qPCR mean Cq and PMA-qPCR mean PMA-Cq values of each dilution, based on the DNA standard curves of both targets. CFU per microliter of E. coli was determined using plate counting on AmpR LB plates. Each 100 µL of 10-6, 10-7, and 10-8 diluted cell suspension was spread individually on the plates, and CFU per microliter was calculated as follows: CFU·µL-1 = count number × 1/dilution rate × 1/inoculum /1,000. A total of 10 replications were performed (same samples were used for qPCR assays) and CFU per microliter was determined from the mean value of 10 repeats.
Statistical analysis
Statistical analysis were performed using Microsoft Excel to determine the average values (Mean) and standard deviation (SD) of Cq, PMA-Cq and plate counting for CFU·µL-1. Also, regression analysis was used to construct standard curves, and to determine dynamic ranges of diluted cell suspensions and PMA-treated diluted cell suspension.
Results
PCR analysis of serially diluted E. coli cell suspensions
Serially diluted E. coli cell suspensions were used as templates in place of DNA for direct PCR amplification of AmpR and dxs. Fig. 1 shows the amplification specificity of AmpR and dxs primer sets. We detected one specific fragment of the expected size corresponding to 181 and 160 bp for AmpR and dxs, respectively. No non-specific products were detected using SDW, indicating high primer specificity for both target genes. Serial dilution PCR revealed that the detection limits of AmpR and dxs were 10-4 and 10-3 diluted samples, respectively.
qPCR parameters of DNA
qPCR performance including efficiency, slope, and correlation coefficient were determined using standard curves obtained from genomic and plasmid DNA isolated from E. coli. As shown in Fig. 2, both AmpR and dxs assays showed good performances in the range between 1 × 103 and 1 × 108 DNA copies, corresponding to plasmid DNA of 7.23 - 723 pg and genomic DNA of 5 - 500 ng, respectively. The standard curve efficiencies were 102.0% and 96.5% with slopes of -3.2727 and -3.4094, and linear correlation coefficients of both standard curves were 0.9983 and 0.9987, for AmpR and dxs, respectively. qPCR assay precision, evaluated using relative standard deviation (RSD%) of the Cq values obtained in six qPCR assay varied from 0.24 to 2.57% for AmpR and 0.31 to 1.38% for dxs. Trueness expressed in terms of bias (bias%) ranged from 4.95 to 20.55% for AmpR and -3.52 to 20.34% for dxs (Table 2). qPCR assays for both targets showed an acceptable level of precision and trueness, which were within the limit of ±25% as detailed in European Network of GMO Laboratories (ENGL) guidelines.
qPCR and PMA-qPCR performance of E. coli cell suspensions
For cell suspension qPCR, 10 repeats of E. coli 18 h cultures were prepared and OD600 and CFU per milliliter were measured. As shown in Fig. 3A, 18 h cultures had an average OD600 of 0.70 ± 0.04 (range: 0.63 - 0.76), corresponding to an average CFU per milliliter of 1.31 × 109 (range: 1.21 × 109 - 1.42 × 109). The 10 culture replicates showed similar increases in cell density, determined by OD600 and CFU·mL-1 values, indicating that the growth phase of the 18 h cultures was not stationary and was suitable for qPCR and PMA-treated qPCR analysis. Dual-plex qPCR using culture stocks (100) revealed that the Cq values of AmpR and dxs were 16.17 ± 0.06 and 22.23 ± 0.31, respectively (Table 3). Compared to non-PMA treated samples, PMA treatment resulted in higher qPCR Cq values of 18.57 ± 0.19 and 24.18 ± 0.26 for both targets, respectively. Variations in Cq and PMA-Cq values were 0.37 and 1.00 RSD% for AmpR and 1.06 and 1.41 RSD% for dxs, indicating that the qPCR of three inter-run repeats and ten intra-run repeats had good repeatability and reproducibility.
Dual-plex qPCR analysis of serially diluted cell suspensions was performed and the mean Cq values from 10-fold (10-1) to 100,000-fold (10-5) cell dilutions of both target genes are shown in Table 3. AmpR and dxs qPCR showed good repeatability and reproducibility, corresponding to RSD% ranging from 1.76 to 3.04 and 1.75 to 3.62, respectively. qPCR performance for AmpR and dxs using five serial cell dilutions showed slopes of -3.285 and -3.134, with linear correlation coefficients of 0.9995 and 0.9988, corresponding to efficiencies of 101.6% and 108.5%, respectively (Fig. 3B and C). These results revealed that the target DNA concentration after qPCR amplification in 10-fold serially diluted cell suspension was identical with the diluted DNA sample, indicating no negative impact of using direct cell suspension on qPCR assays. Furthermore, LB broth and SDW were analyzed as negative controls under the same dual-plex qPCR conditions for both targets to evaluate possible matrix effects. The matrixes of LB broth and SDW had similar Cq values of 35.94 ± 0.98 and 36.10 ± 0.68 for AmpR and 36.28 ±0.91 and 36.42 ± 0.59 for dxs, respectively, indicating that the combination of primer/probe sets of AmpR and dxs and the cell culture matrix did not show non-specific cross-reactivity in the qPCR assay. The PMA-treated serially diluted cell suspensions showed qPCR performance slopes of -3.0893 and -3.2302, with linear correlation coefficients of 0.9979 and 0.9976, corresponding to efficiencies of 110.7% and 104.0%, for both targets, respectively (Fig. 3B and C). These results indicated that qPCR amplification using diluted viable cells was not affected by PMA addition to the sample matrix. As a negative control, PMA-treated LB broth showed qPCR Cq values of 35.66 ± 1.92 and 36.66 ± 0.53 for both targets, respectively.
Total cell counts and viable cell counts were obtained using qPCR and PMA-qPCR based on the Cq values and constructed DNA standard curves. The mean total cell counts of AmpR and dxs were calculated to be 126.00 ± 50.40 (× 106) and 8.13 ± 6.42 (× 106), respectively, and the mean viable cell counts per microliter were 19.60 ± 6.63 (× 106) and 1.41 ± 0.55 (× 106), respectively (Table 3). Considering the copy number of the dxs gene is one, the theoretical cell density (true copy number) calculated using PMA-qPCR should be 1.41 ± 0.55 (× 106), revealing a bias of 7.71% from the plate counting (mean CFU·µL-1: 1.31 × 106). This bias did not overlap with the limit of 25% as detailed in ENGL guidelines, suggesting acceptable trueness of the PMA-qPCR assay. In addition, the plasmid copy number was calculated using a ratio of 15.50 from the total cell count of AmpR to that of dxs in the qPCR assay. For PMA-qPCR assay, a ratio of 13.90 was calculated from the viable cell count of AmpR to that of dxs, implying a plasmid copy number of 14 - 16.
Table 4. Limit of detection (LOD) confirmation using replicate propidium monoazide–quantitative real-time PCR (PMAqPCR) runs on separate dilution series of suspension cells from E. coli.
|
Detection limit of viable cells using PMA-qPCR
The detection limit of dual-plex PMA-qPCR assay is the lowest viable E. coli cell numbers that can be detected with 95% confidence. The cell stock with an average CFU of 1.31 × 106·µL-1 was serially diluted from 10-fold (10-1) to 1,000,000-fold (10-6), while E. coli cell numbers were estimated to be 1.31 × 105 to 1.31 copies·µL-1, after PMA treatment. For all treated samples, qPCR consisted of a total of ten repeats containing two inter-run repeats and five intra-run repeats. To improve detection accuracy, we performed several replicates for a total of 22 repeat measurements for the 10-3 - 10-6-fold diluted samples. Both assays of AmpR and dxs detected an average of 131 viable E. coli cells from 10,000-fold (10-4) diluted cell suspension (Table 4). Statistical analysis revealed that the LOD95% on AmpR and dxs was 134 live cells with a 95% confidence interval of 83 - 216 cells detectable using dual-plex PMA-qPCR assays.
Discussion
GMOs are commonly utilized in the food and feed industries to produce additives, enzymes, and flavorings. They frequently harbor antimicrobial resistance genes as selection markers, conferring a resistance to different antibiotics including chloramphenicol, ampicillin, kanamycin, bleomycin, and tetracycline (Fraiture et al., 2020). The acquisition of potential antimicrobial resistance genes through horizontal gene transfer from pathogens and other microorganisms is of considerable public health and environmental concern, requiring fast and effective detection methods to monitor living GMOs released into the surrounding environment, such as wastewater, air currents, and facility surfaces from large-scale food and feed processing industries.
Most environmental microbial monitoring experiments use DNA-based real-time PCR detection methods, requiring the filtration, concentration, and culture of the environmental samples using selection media (Heijnen and Medema, 2006). However, these real-time PCR-based methods have limited monitoring efficiency due to technical difficulties in microbial DNA extraction and estimation of the actual number of viable microbial cells present in environmental samples. Moreover, DNA quality and real-time PCR efficiency can be influenced by DNA extraction protocols, which affect DNA recovery rate and integrity, as well as contamination by DNA inhibitors (Piskata et al., 2019). Culture-based methods are currently used to distinguish dead and living microbial cells in environmental sample and to assess the number of living cells. However, real environmental microbial cell numbers may be overestimated due to concentration and culture growth of the target microorganism. Additionally, sample pretreatment with nucleic acid intercalating dyes such as PMA has been proposed to avoid qPCR false-positive results (da Cunha et al., 2020). PMA only binds to DNA from dead or membrane-damaged cells following exposure to bright light, and qPCR amplification of PMA-bound DNA is inhibited (Fittipaldi et al., 2012). However, PMA treatment can reduce DNA yields from treated cells compared to those from untreated living cells (Joo et al., 2019), which can reduce the detection efficiency of DNA-based real-time PCR.
Here, we describe a PMA-qPCR method to detect viable genetically modified E. coli cells harboring the ampicillin resistance gene, AmpR, by using direct cell suspensions. To quantify the number of living cells, an E. coli single copy endogenous gene, dxs (Lee et al., 2006), was used as an internal control, combined with AmpR for TaqMan probe based dual-plex qPCR analysis. In addition, the use of the taxon-specific endogenous gene, dxs, allows the screening of natural AmpR-resistant microorganisms to enhance the assay specificity of the target ampicillin-resistant E. coli strains. Cell suspensions were prepared from 18 h cultures, and 10 replicates representing an average OD600 of 0.70 and an average CFU of 1.31 × 106 per µL using plate counting were used. As a result, serially diluted cell suspension direct qPCR and PMA-qPCR assays showed good performances on inter- or intra-run replicates, and the use of cell suspensions had little effect on qPCR accuracy and precision, and was suitable for qPCR analysis compared to the use of DNA templates. These results are consistent with similar studies by Sung and Hawkins (2020) on the detection of mycoplasma contamination in mammalian cell lines, and the results indicated that the accumulation of metabolic waste products in cultured cell suspensions had no effect on qPCR precision and robustness.
DNA standard curves generated by plasmid DNA often overestimate the amount of DNA in qPCR due to contamination of different conformations of plasmid DNA, such as supercoiled circular DNA and linear DNA, known for differing DNA amplification efficiencies (Lin et al., 2011). We generated the AmpR standard curve using plasmid DNA, and the single copy gene dxs standard curve using genomic DNA, with the same dual-plex qPCR assay as the cell suspension, to evaluate the actual cell counts in cell suspensions of qPCR and PMA-qPCR assays. Based on DNA standard curves, we calculated theoretical total cell counts and viable cell counts from the Cq values from qPCR and PMA-qPCR of serially diluted cell suspensions. The 7.71% bias between calculated viable cell counts and CFU by plate counting did not overlap with the baseline ±25% proposed by the ENGL (2015), indicating that monitoring viable AmpR-resistant E. coli cells using direct cell suspension PMA-qPCR assays was accurate and effective. Moreover, qPCR and PMA-qPCR showed similar cell count ratios of AmpR to dxs, indicating that the plasmid, pGAL-HIR525, had approximately 14 - 16 copies in E. coli.
Zhang and Fang (2006) suggested that reliable standard curves should exhibit slope values between -3.9 and -3.0, corresponding to qPCR efficiency between 80 and 115%, and R2 values higher than 0.95 when qPCR is applied for microbial quantification in environmental samples. ENGL establishes that the variance of qPCR analysis based on relative standard deviations (RSD%), should be less than 25% for dynamic method ranges. Here, we report optimal DNA standard curves for AmpR and dxs, and appropriate linear regression generated by Cq and PMA-Cq values of diluted cell suspensions. Thus, the number of living AmpR-resistant E. coli cells in environmental samples can be reliably quantified using PMA-qPCR standard curves as shown in Fig. 3B and C, instead of using DNA standard curves. The assay sensitivity confirmed that the detection limit for both targets was at least 134 viable E. coli cells with 95% confidence intervals using dual-plex PMA-qPCR. This work describes a novel TaqMan based dual-plex qPCR method to circumvent the technical limitations of DNA extraction, improve detection efficiency, reduce assay time and costs, and accurately detect the number of living AmpR-resistant genetically modified E. coli cells in environmental samples.
Conclusion
As industrial GM organisms are developed and commercialized, a detection method is needed to quickly and effectively monitor living GMOs emitted to the surrounding environment such as wastewater, air flow, and facility surfaces in large-scale food and feed processing industries. In this study, the PMA-treatment combining dual-plex TaqMan-based qPCR method targeting on Amp-resistant gene and taxon-specific gene dxs was developed to directly detect viable Amp-resistant E. coli cells from cell culture. qPCR and PMA-qPCR assays using diluted cell suspensions on both targets were confirmed acceptable levels of precision and trueness, which were within the ±25% limit detailed in ENGL guidelines. In addition, qPCR analysis was not affected by the cell culture matrix. The PMA-qPCR assay allows detection limits to be detected at 95% confidence of 134 viable Amp-resistant E. coli cells. The method can avoid high technical requirements for DNA extraction and purification, save time and cost, avoid false positive detection in living GM cells, and exclude natural Amp-resistant microorganisms. This work provides a feasible monitoring approach for living GMO detection in environmental samples.
