Qualitative and Quantitative Real-Time PCR Methods for Assessing False-Positive Rates in Genetically Modified Organisms Based on the Microbial-Infection-Linked HPT Gene

The hygromycin phosphotransferase (HPT) gene as a selective marker is normally used in screening tests as a first step in detecting and quantifying genetically modified organisms (GMOs) in seeds, food, and feed materials. Nevertheless, if researchers only focus on the HPT gene, it is difficult to distinguish genetically modified (GM) crops from microbial infection, leading to miscalculation of the rate of GM materials in a given sample set. Here, we cloned the 7259 bp sequence carrying the HPT gene from soybean sprouts using the genome walking strategy. BLAST analysis revealed that this sequence was derived from plasmids naturally occurring in microorganisms, such as Escherichia coli, Klebsiella pneumoniae or Salmonella sp. Using the reconstructed plasmid pFP-hpt, qualitative PCR and quantitative real-time PCR (qPCR) methods were established, and 261 bp and 156 bp products were produced. The specificity of these assays was assessed against related pFP-hpt plasmids, plant species with important agronomic traits, and GM crops containing the HPT gene. No unexpected results were observed between samples using these qualitative PCR and qPCR methods. The sensitivity of this qualitative PCR assay was determined at 20 copies, while the limit of detection (LOD) and limit of quantification (LOQ) of qPCR were both 5 copies per reaction. Our in-house validation indicated that the amplification efficiency, linearity, and repeatability of this qPCR assay were in line with performance requirements. Furthermore, a qualitative and quantitative duplex PCR showed high reliability for the simultaneous detection of the HPT gene in a plant sample and environmental micro-organisms harboring the HPT gene in one PCR reaction. These qualitative PCR and qPCR assays were able to differentiate between plants infected with E. coli harboring the HPT gene from GM plants, indicating that these two methods are broadly applicable for routine GMO testing.


Introduction
Due to the annual increase in the variety and number of genetically modified (GM) plants, monitoring and management systems for seeds, food, and feed products for the presence of GMOs are enforced in many countries worldwide, which require reliable and cost-effective detection methods. DNA analysis based on polymerase chain reaction (PCR) is routinely performed to assess the presence or absence of GMOs. Several common genetic elements present in many GMOs, such as the CaMV35S promoter derived from the cauliflower mosaic virus and the NOS terminator derived from Agrobacterium tumefaciens, Int. J. Mol. Sci. 2022, 23, 10000 3 of 15 by environmental microorganisms harboring the HPT gene often found on processed vegetables.

Isolation and Characterization of the HPT Gene Sequence
To obtain the complete HPT gene and flanking sequences, a genome walking PCR strategy was adopted to determine novel and unknown sequences adjacent to HPT. A 7259 bp fragment was obtained by overlapping two derived fragments with the HPT gene. BLAST analysis revealed that 7259 bp of the isolated sequence was highly homologous to plasmid DNA found in Escherichia coli, Klebsiella pneumoniae, and Salmonella sp., with sequence identity of about 98.1%. Ten plasmid DNA sequences were then collected from Genebank: six Klebsiella pneumoniae strains MH15-269 (

Design of Primers and Probes
Based on this sequence alignment, a conserved fragment was found to be present in all collected sequences from microorganisms in this study. Therefore, multiple candidate primer/probe sets were designed based on these conserved segments without any SNPs, targeting the junction site between the 5 -plasmid DNA sequence and the HPT gene. This segment was then constructed into a pFP-hpt plasmid. The forward primer bound to 5 -plasmid DNA, while the reverse primer binding site was in the HPT gene adjacent to the plasmid DNA. The probe bound to the junction between the 5 -plasmid DNA sequence and the HPT gene fragment. The locations of primers and probes for qualitative and qPCR methods are shown in Figure 1, and details of primers and probes can be found in Table S1. To select the best primer/probe set, all possible primer and probe combinations were tested for application size and specificity using 500 copies of pFP-hpt as a template. The most efficient, reliable, and stable qualitative/quantitative PCR primer/probe sets were Ghpt-F3/Ghpt-R3 and Ghpt-QF3/Ghpt-QR3/Ghpt-QP, yielding 261 bp and 156 bp size products, respectively (Table S3). cessed vegetables.

Isolation and Characterization of the HPT Gene Sequence
To obtain the complete HPT gene and flanking sequences, a genome walking P strategy was adopted to determine novel and unknown sequences adjacent to HPT 7259 bp fragment was obtained by overlapping two derived fragments with the HPT ge BLAST analysis revealed that 7259 bp of the isolated sequence was highly homologous plasmid DNA found in Escherichia coli, Klebsiella pneumoniae, and Salmonella sp., with quence identity of about 98.1%. Ten plasmid DNA sequences were then collected fro Genebank: six Klebsiella pneumoniae strains MH15-269 (Accession No. AP023338 AR0125 (Accession No. CP021858.1), AR0113 (Accession No. CP021755.1), AR0112 (A cession No. CP021547.1), AR0115 (Accession No. CP021547.1), and CAV1596 (Access No. CP011644.1); three Escherichia coli strains TD-33 (Accession No. MN915013.1), GD (Accession No. MN915011.1), TJ-33 (Accession No. MN915010.1); and one Salmonella (Accession No. CP071694.1). Homologous sequence alignment indicated that these co served regions exhibited some single nucleotide polymorphisms (SNP) (Supplementa Figure S1). Although HPT genes from the plasmid DNA of microorganisms and the i lated sequence had some SNPs (Supplementary Figure S2), they simultaneously all s probably encoded hygromycin B phosphotransferase (Supplementary Figure S3).

Design of Primers and Probes
Based on this sequence alignment, a conserved fragment was found to be present all collected sequences from microorganisms in this study. Therefore, multiple candid primer/probe sets were designed based on these conserved segments without any SN targeting the junction site between the 5′-plasmid DNA sequence and the HPT gene. T segment was then constructed into a pFP-hpt plasmid. The forward primer bound to plasmid DNA, while the reverse primer binding site was in the HPT gene adjacent to plasmid DNA. The probe bound to the junction between the 5′-plasmid DNA sequen and the HPT gene fragment. The locations of primers and probes for qualitative and qP methods are shown in Figure 1, and details of primers and probes can be found in Ta S1. To select the best primer/probe set, all possible primer and probe combinations w tested for application size and specificity using 500 copies of pFP-hpt as a template. T most efficient, reliable, and stable qualitative/quantitative PCR primer/probe sets w Ghpt-F3/Ghpt-R3 and Ghpt-QF3/Ghpt-QR3/Ghpt-QP, yielding 261 bp and 156 bp s products, respectively (Table S3).

Specificity and Sensitivity Test of Singleplex Qualitative
The sequence of the primer set of Ghpt-F3/Ghpt-R3, generating a 261 bp amplicon product, was next analyzed using BLAST. The sequence was mainly homologous to the plasmid DNA found in E. coli, K. pneumoniae, Salmonella sp., and the genomic DNA of Proteus mirabilis, with 100% identity (Supplementary Figure S4). No sequence homology was observed with related non-target organisms, such as plant viruses or crop plants. The amplification specificity of the primer set Ghpt-F3/Ghpt-R3 was tested using genomic DNA from non-transgenic plants (Arabidopsis thaliana, soybean, maize, peanuts, rice, rapeseed, cotton, wheat, sugar beet, and tomato), GM rice harboring HPT gene (Kefeng 6, Kemingdao and SD rice), GM rice (TT51-1 and G6H1), GM soybean mixed sample, GM maize mixed sample, GM rapeseed mixed sample, GM cotton mixed sample, and plasmid DNA (pFPhpt). The results are presented in Table 1. Plant endogenous genes from 18S rRNA [17] have been observed in specific amplification bands. However, the 261 bp target fragment and 472 bp of HPT gene from GM crops have not been observed from plant species, GM rice (TT51-1 and G6H1), GM soybean mixed sample, GM maize mixed sample, GM rapeseed mixed sample, and GM cotton mixed sample in similar assays. Electrophoresis showed that a unique 261 bp fragment was amplified from pFP-hpt plasmid, but there was no amplification observed from Kefeng 6, Kemingdao, and SD rice (Figure 2a). Specificity tests indicated qualitative PCR amplification using the primer set Ghpt-F3/Ghpt-R3 could reliably and specifically detect micro-organisms carrying the HPT gene.  ple, GM rapeseed mixed sample, and GM cotton mixed sample in similar assays. Ele phoresis showed that a unique 261 bp fragment was amplified from pFP-hpt plasmid there was no amplification observed from Kefeng 6, Kemingdao, and SD rice (Figure Specificity tests indicated qualitative PCR amplification using the primer set G F3/Ghpt-R3 could reliably and specifically detect micro-organisms carrying the HPT g In practice, DNA extracted from GM food or feed tends to be commonly infecte contaminated with trace levels of microorganisms. To evaluate the sensitivity of our qualitative assay system, pFP-hpt plasmid DNA was serially diluted from 500, 100, 50 and 10 copies per microliter and used as a template for PCR analysis. Specific produc 261 bp could be observed in reactions with as few as 20 copies of pFP-hpt as a temp ( Figure 2b). These results indicate that the detection sensitivity of the qualitative me established in this research was as low as 20 copies.

Target Species Plant
18S rRNA Qualitative PCR qPCR Ghpt-F3/R3 HPT Gene Ghpt-QF3/QR3/QP HPT G Arabidopsis thaliana In practice, DNA extracted from GM food or feed tends to be commonly infected or contaminated with trace levels of microorganisms. To evaluate the sensitivity of our new qualitative assay system, pFP-hpt plasmid DNA was serially diluted from 500, 100, 50, 20, and 10 copies per microliter and used as a template for PCR analysis. Specific products at 261 bp could be observed in reactions with as few as 20 copies of pFP-hpt as a template ( Figure 2b). These results indicate that the detection sensitivity of the qualitative method established in this research was as low as 20 copies.

Determination of the Applicability of Qualitative PCR
To verify the applicability of this new qualitative PCR method, a routine testing sample was numbered 2519, set as a blind sample, and investigated in detail. pFP-hpt plasmid and SD rice were used as positive and negative controls, respectively. The material consisting of extracted rice seed was determined to be positive for the presence of the HPT gene, yielding a 472 bp product ( Figure 3a). However, it could not be assigned to a GM plant event due to the absence of other screening elements or GM-specific events. Therefore, the material was analyzed with conventional PCR using the primer set Ghpt-F3/Ghpt-R3. Specific 261 bp products from rice sample 2519 and pFP-hpt were observed by electrophoresis (Figure 3b). To confirm the specificity of fragment, the obtained PCR product was sequenced using both forward and reverse primers from this qualitative assay. The sequence alignment of the 261 bp amplicon showed 100% homology with major microbes from E. coli, K. pneumoniae, Salmonella sp., and Proteus mirabilis, which carried the complete HPT gene in a plasmid or genomic DNA by further alignment analysis (Supplementary Figure S4). These results reveal that the qualitative PCR assay established in this research could be applied to exclude false positives for the HPT gene during routine GM analysis.
microbes from E. coli, K. pneumoniae, Salmonella sp., and Proteus mirabilis, which carried complete HPT gene in a plasmid or genomic DNA by further alignment analysis (Sup mentary Figure S4). These results reveal that the qualitative PCR assay established in research could be applied to exclude false positives for the HPT gene during routine analysis.

Verification of qPCR Specificity
The sequence of primer/probe set Ghpt-QF3/Ghpt-QR3/Ghpt-QP was next evaluated for use in qPCR. Based on the above analysis, this 156 bp fragment was specific. The experimental specificity of qPCR assay was tested in the same way as that of our qualitative PCR method. There were no positive signals observed in the plant species, GM rice (TT51-1 and G6H1), GM soybean mixed sample, GM maize mixed sample, GM rapeseed mixed sample, and GM cotton mixed sample using qPCR detection for Ghpt-QF3/Ghpt-QR3/Ghpt-QP and the HPT gene (Table 1). Only pFP-hpt plasmid DNA was amplified with the primer/probe set Ghpt-QF3/Ghpt-QR3/Ghpt-QP, and no positive signals were observed in GM rice harboring HPT gene (Kefeng 6, Kemingdao, and SD rice). These results indicate that qPCR could be used to specifically detect microorganisms carrying the HPT gene.

Standard Curve and Repeatability of qPCR
To evaluate the amplification efficiency and linearity of our qPCR assay, a 10-fold series of dilutions of plasmid-pFP-hpt was prepared, corresponding to 500,000, 50,000, 5000, 500, and 50 copies/µL. The serial dilutions were used to establish a standard curve, and each dilution was assayed in triplicate with three parallel reactions. Standard curves were created by plotting Ct values against the logarithm of pFP-hpt plasmid copy numbers. There was a good agreement observed between the quantity of the template and the Ct values, with R 2 values between 0.999 and 1.000. The slope of the standard curve was calculated between −3.27 to −3.369, with corresponding amplification efficiencies of 98.10% to 99.70% (Figure 4, Table 2). These data met the minimum performance requirements for analytical methods defined by the European Network of GMO Laboratories [18].
Ct values, with R 2 values between 0.999 and 1.000. The slope of the standard calculated between −3.27 to −3.369, with corresponding amplification effi 98.10% to 99.70% (Figure 4, Table 2). These data met the minimum performan ments for analytical methods defined by the European Network of GMO L [18].   The repeatability of the qPCR system was analyzed using these standard curves as described above. With the reduction in plasmid template from 500,000 to 50 copies, the average Ct values of the real-time assay increased from 21.41 to 34.77, with RSD values ranging from 0.24% to 0.99%, and the relative repeatability standard deviation (RSD r ) ranging from 0.03% to 0.30% (Table 3). These results also comply with the minimum performance requirements for analytical methods defined by the European Network of GMO Laboratories [18] and reveal that the quantitative real-time PCR assay had good stability and reliability.

Limits of Detection and Quantification with qPCR
To determine the sensitivity of this qPCR method, the LOD and LOQ were next determined. The pFP-hpt plasmid was diluted to 50, 20, 10, 5, and 1 copies for analysis of 9 replicate real-time PCR reactions. The detection capacity of our quantitative assay decreased with decreasing template copy numbers (Table 4). All nine PCR replicates had typical amplification curves when the template copy number was 5, whereas only five reactions were positive when the template copy number was 1. Therefore, the LOD of this qPCR method was estimated to be five copies of plasmid. Relative repeatability standard deviation (RSD r ) of Ct values from the nine replicates were all below 25% with the reduction in the template copy number. The LOQ of our method was approximately five copies of plasmid based on the implementation of the guidance document of the Joint Research Centre, Institute for Reference Materials and Measurements (JRC-IRMM). These results indicate that singleplex qPCR could be used to detect micro-organism infection carrying the HPT gene with high sensitivity but should not be used to accurately quantify the amount of infection.

Analysis of HPT-Positive Seed Material Using qPCR
The rice sample 2519 was also used to verify the applicability of our qPCR method. The pFP-hpt plasmid, non-transgenic rice, and sterile water were used as a positive control, negative control, and no template control, respectively. The results show that rice 2519 had a typical amplification curve with Ct values of less than 30, which correspond to approximately 500 copies of plasmid from an infected bacterium ( Figure 5). The results of the 156 bp PCR product alignment are consistent with those of the 261 bp sequence produced using our qualitative method, as indicated by BLAST analysis. These results reveal that the qPCR assay developed in this paper could also be applied to exclude false positives for the HPT gene during routine GM analysis.

Analysis of HPT-Positive Seed Material Using qPCR
The rice sample 2519 was also used to verify the applicability of our qPCR The pFP-hpt plasmid, non-transgenic rice, and sterile water were used as a posit trol, negative control, and no template control, respectively. The results show t 2519 had a typical amplification curve with Ct values of less than 30, which corres approximately 500 copies of plasmid from an infected bacterium ( Figure 5). The r the 156 bp PCR product alignment are consistent with those of the 261 bp seque duced using our qualitative method, as indicated by BLAST analysis. These result that the qPCR assay developed in this paper could also be applied to exclude fa tives for the HPT gene during routine GM analysis.

Development of a Qualitative and Quantitative Duplex PCR
The specificity of the designed primers/probes above were assessed using a singleplex PCR assay, as the conventional and real-time PCR methods that have been used in the field to detect the HPT gene were singleplex and tested for specificity [19]. Considering the time and cost of analysis, we developed a qualitative and quantitative duplex PCR test for simultaneous detection of the HPT gene from possible GM introduction and environmental micro-organisms harboring the HPT gene. Figure 6 shows the results of a qualitative duplex PCR with rice sample 2519 infected with environmental microorganisms. We have also developed a quantitative duplex test utilizing the two specific sets of primers and probes, one for the HPT gene and Ghpt-QF3/Ghpt-QR3/Ghpt-QP for the detection of environmental micro-organisms harboring the HPT gene (Table 5). Consequently, the combination of both targets, the HPT gene and environmental micro-organisms harboring the HPT gene, was viable in one PCR reaction system. qualitative duplex PCR with rice sample 2519 infected with environmental microorganisms. We have also developed a quantitative duplex test utilizing the two specific sets of primers and probes, one for the HPT gene and Ghpt-QF3/Ghpt-QR3/Ghpt-QP for the detection of environmental micro-organisms harboring the HPT gene (Table 5). Consequently, the combination of both targets, the HPT gene and environmental micro-organisms harboring the HPT gene, was viable in one PCR reaction system.

Discussion
Due to its extensive use as selectable marker gene in the first generation of GMOs, the HPT gene derived from E. coli has become the generally used PCR target for screening tests in routine GMO analysis [3]. In this study, only the HPT gene was detected from soybean sprouts in a routine test in our laboratory. For further analysis of whether this sample was a non-authorized GMO, the genome walking strategy was used to identify unknown sequences based on the HPT gene. Eventually, a 7259 bp sequence containing the HPT gene was retrieved after multiple rounds of PCR amplification. BLAST analysis revealed that this 7259 bp plasmid fragment might have derived from strains of E. coli, K. pneumoniae or Salmonella sp. The findings suggest that the presence of environmental micro-organisms in soybean sprouts resulted in the presence of the HPT gene in the samples. Determining whether the HPT gene is derived from genetically modified crops or nearby micro-organisms will require reliable and stable detection methods.
In recent years, transgenic testing laboratories have paid increasing attention to infection and cross-contamination during GMO detection from food, feed, and seeds. Formerly, conventional PCR methods based on ORF VI and ORF V were used to detect CaMV to distinguish P35S derived from CaMV or transgenic crops [4]. Real-time PCR and multiplex PCR methods targeting different conserved regions were designed to enhance the accuracy and coverage of CaMV detection [5][6][7][8]. Real-time PCR methods for the detection of figwort mosaic virus (FMV) were developed to complement the FMV 34S promoter-specific PCR assay used for the screening of GMOs in food and feed [9]. In addition, a QRT-PCR assay targeting the TMS1 region was successfully applied to distinguish elements such as T-nos, P-nos, and T-3 g7, which have been introduced by Agrobacterium sp. into GMOs plant events from natural Agrobacterium sp. [20].
Here, through integrating a 3207 bp sequence into a plasmid vector, we developed a novel plasmid as a positive sample, pFP-hpt, carrying 285 bp of the incomplete HPT gene and 2922 bp of the 5 flanking sequence. A specific qualitative and qPCR assay targeting 261 bp and 156 bp between the HPT gene and its 5 flanking adjacent sequence, which were conserved sections without SNPs by limited alignment analysis, were established to complement HPT gene assays used for screening first generation GMOs. These qualitative and qPCR tests showed high specificity and sensitivity, even when a trace amount of the HPT gene from micro-organisms was present. Combining this with our previously validated HPT gene method [16], these assays can be applied to routine testing in the laboratory to avoid false positives of the HPT selectable marker. The LOD and LOQ of our qPCR assay were estimated to be five copies of plasmid DNA. In this paper, we demonstrated that the pFP-hpt plasmid prepared in this study can be used as an alternative standard for the quantification of the HPT gene from natural occurrences.
Generally, GMO detection is divided into two steps: element-screening PCR is often the initial step in GMO testing procedures. When results indicate the presence of specific genetic elements, identification or quantification of the GM plant is followed by eventspecific assays [21]. This two-step detection procedure requires a large amount of time and reagents. The developed qualitative and quantitative duplex PCR could detect two targets in one PCR reaction at the same time, thereby reducing the reagent cost, shortening the detection time, improving the detection efficiency, and providing an effective method for the rapid detection of the HPT gene and micro-organisms harboring the HPT gene. This is the first study to attempt to avoid HPT gene false-positive tests derived from bacterial plasmids, and future work should be conducted on bacterial genomes carrying the HPT gene to improve false-positive exclusive detection systems. We will also expand the application of this method for daily detection in transgenic testing laboratories.

DNA Extraction and Purification
For GMO screening of all samples and genome walking, genomic DNA was isolated and purified using the DNeasy Plant Maxi Kit (QIAGEN, Hilden, Germany) according to the manufacturer's protocol. For qualitative and quantitative method in-house validation, genomic DNA was extracted and purified from young leaves or seeds of species plants using a cetyltrimethylammonium bromide (CTAB)-based protocol [23]. Plasmid DNA was prepared using the TIANprep mini plasmid kit (TIANGEN, Beijing, China) following the manufacturer's recommendations. DNA concentrations were analyzed using a NanoDrop ONE c (Thermo Scientific, Waltham, MA, USA) spectrophotometer.

Genome Walking PCR
The complete HPT gene and its flanking sequence were isolated utilizing the Genome-Walker Universal Kit (Clontech, Mountain View, CA, USA) according to the manufacturer's instructions. Gene-specific primers (GSPs) were designed according to the sequence of the HPT gene (Accession No. AF234296) and these isolated sequences. The primers designed for this study are listed in Table S3.

Construction of a Reference Plasmid
After characterizing the isolated sequence, a 3207 bp sequence (Supplementary Table S2) consisting of 285 bp of the HPT gene sequence and its 2922 bp 5 flanking DNA sequence were synthesized by Sangon Biotech and sub-cloned into plasmid vector pUC57, giving rise to the reference plasmid pFP-hpt. pFP-hpt was used for detecting false-positive GM plants due to micro-organism infection or sample contamination.

Primer and Probe Design
Qualitative PCR oligonucleotide primers were designed using the software Primer Premier Version 5.00 (PREMIER Biosoft International, Palo Alto, CA, USA). qPCR primers and TaqMan fluorescent probes were designed using the software Beacon Designer 8.0 (PREMIER Biosoft, USA). All the primers and probes were synthesized by Sangon Biotech (Table S3).

Qualitative PCR Conditions
Qualitative PCR amplifications were performed using a Bio-rad C1000 thermal cycler using an optimized PCR mixture including: 10× Taq Buffer, 2.5 mM MgCl 2 , 200 µM of each dNTP, 0.4 µM of each primer, 0.7 unit Taq DNA Polymerase (TIANGEN, Beijing, China), and 50 ng of genomic DNA or plasmid DNA. Then, 500 copies, 100 copies, 50 copies, 20 copies, or 10 copies were used as templates in a final volume of 25 µL.

Specificity Tests
To evaluate the specificity of our qualitative and qPCR assays, non-GM plants, such as A. thaliana, soybean, maize, peanuts, rice, rapeseed, cotton, wheat, sugar beet, tomato, GM soybean mixed sample, GM maize mixed sample, GM rapeseed mixed sample, GM cotton mixed sample, GM rice including TT51-1, G6H1, Kefeng 6, Kemingdao, and SD rice, as well as a reconstructed plasmid for pFP-hpt, were all tested. The assay was considered specific when the anticipated amplification was detected for each sample.

Sensitivity Tests
To determine sensitivity of our qualitative PCR assay, the serial dilutions of the pFPhpt plasmid from 500 copies to 10 copies were tested for each target. Two parallel reactions were performed.
To evaluate the LOD and LOQ for our qPCR assay, a plasmid DNA series with copy numbers from 50, 20, 10, 5, and 1 was used. Three parallel reactions were performed on nine replicates of PCR runs. The LOD was analyzed using the cycle threshold (Ct) values of each reaction and all replicates could be reliably detected. The LOQ was estimated as the lowest copy number with the relative standard deviation (RSD) of all replicates below 25% for the measured copy number in this assay.

Conclusions
In summary, the quantitative PCR and qPCR system demonstrated here was more sensitive and specific for the HPT of a bacterial donor during routine GM testing of materials such as seeds. Even though the frequency of HPT false-positive test results is not currently known, the methods here allow GMO testing laboratories to distinguish between the HPT gene from GM plants and naturally occurring bacteria.