Advanced uracil DNA glycosylase-supplemented real-time reverse transcription loop-mediated isothermal amplification (UDG-rRT-LAMP) method for universal and specific detection of Tembusu virus

Tembusu virus (TMUV) is a mosquito-borne flavivirus which threatens both poultry production and public health. In this study we developed a complete open reading frame alignment-based rRT-LAMP method for the universal detection of TUMV. To prevent false-positive results, the reaction was supplemented with uracil DNA glycosylase (UDG) to eliminate carryover contamination. The detection limit of the newly developed UDG-rRT-LAMP for TMUV was as low as 100 copies/reaction of viral RNA and 1 × 100.89 − 1 × 101.55 tissue culture infectious dose/100 μL of viruses. There were no cross-reactions with other viruses, and the reproducibility of the assay was confirmed by intra- and inter-assay tests with variability ranging from 0.22–3.33%. The new UDG-rRT-LAMP method for TMUV produced the same results as viral isolation combined with RT-PCR as the “gold standard” in 96.88% of cases for 81 clinical samples from subjects with suspected TMUV infection. The addition of UDG can eliminate as much as 1 × 10−16 g/reaction of contaminants, which can significantly reduce the likelihood of false-positive results during the rRT-LAMP reaction. Our result indicated that our UDG-rRT-LAMP is a rapid, sensitive, specific, and reliable method that can effectively prevent carryover contamination in the detection of TMUV.

Scientific RepoRts | 6:27605 | DOI: 10.1038/srep27605 increased blood glucose levels in infected chickens 11 . Since 2010, a highly contagious disease has spread throughout most breeder, layer, and meat-type duck farms in China, causing significant economic losses in the Chinese duck industry 2 . In addition to ducks, TMUVs are commonly detected in other domestic and wild avian species including geese 12 , layer chicken 13 , pigeons 14 , and house sparrows 3 . TMUV-infected birds exhibit various disease conditions as documented in the literature; in particular, high fever, loss of appetite, severe reduction in egg production, and/or severe neurologic disorders were commonly observed in TMUV-affected flocks 1 . Our previous study indicated that TMUV-specific antibodies and viral RNA could be detected in duck farm workers in China during the disease outbreak in ducks 15 . Although TMUV has not been reported to cause illness in humans, TMUV still has a high potential to evolve into a zoonotic pathogen that threatens public health as a mosquito-borne flavivirus. Therefore, a rapid and accurate method for diagnosing TMUV infections is important for disease control measures.
In general, TMUV is isolated using embryonated duck/chicken eggs 4,16 , primary duck/chicken embryo cell cultures 2,17 , mammalian cell lines, and mosquito cell lines 18 for initial virology diagnosis purposes. Serological methods of virus neutralization 19 , ELISA [20][21][22] , and fluorescent antibody 23 tests are also commonly used for the detection of TMUV antigens or specific antibodies. These TMUV routine virology and serology tests are useful, but they are time-consuming and have low-sensitivity issues. The nucleic acid amplification tests of conventional RT-PCR (cRT-PCR) 1,11 , nested RT-PCR, and semi-nested RT-PCR 24 for specific TMUV genes have been effectively used for the rapid diagnosis of TMUV infection since their development during the TMUV-related disease outbreak in the late 1990s and in recent years. Furthermore, SYBR ® Green and TaqMan ® real-time RT-PCR (rRT-PCR) 25,26 and reverse-transcription loop-mediated isothermal amplification (RT-LAMP) 27,28 targeting specific TMUV genes were reported for more rapid and sensitive detection of TMUV infections. The advantages of RT-LAMP assays for TMUV detection in field situations or poorly equipped laboratories were previously reported by our group and other groups 24,29 . However, there were two main limitations of the application of RT-LAMP for TMUV detection: i) most of these developed RT-LAMP assays were based on the sequences of historical TMUV strains, and thus, they were limited to TMUV reference strains of similar origin (duck, goose, or chicken), making them unsuitable for detecting divergent TMUV strains from various hosts or genotyping groups; and ii) the high sensitivity of RT-LAMP assays can easily lead to false-positive results due to carryover contamination caused particularly by aerosol droplets generated during the assay process 30 . Fortunately, carryover contamination in the nucleotide amplification reaction could be eliminated using uracil DNA glycosylase (UDG) to process the dUTP-incorporated amplicons [31][32][33] . However, there is no report describing the integration of RNA template-based RT-LAMP amplification with UDG digestion in a one-pot reaction to eliminate carryover contaminants. Therefore, the objective of this study was to develop a carryover contamination-free rRT-LAMP assay for the rapid and universal detection of various TMUV strains.

Results
The rRT-LAMP primer set for the universal detection of TMUVs. The current UDG-rRT-LAMP assay was designed to detect all TUMV genotype groups/sub-groups. We identified a conserved region among all representatives from different TMUV genotypes corresponding to the 3′ end of the NS5 gene using the mVISTA method (Fig. 1A). This conserved region was located in the RdRP domain from genome position 9,793 to 10,003 (numbering corresponds to the complete genome of strain YY5, GenBank accession No. JF270480) (Fig. 1A). The 210-bp conserved region alignment of different TMUV strains was conducted using the ClustalW method (Fig. 1B), resulting in an overall alignment score exceeding 99%. Based on the complete ORF and NS5 conserved region alignment results, four primers targeting six different regions in the NS5 gene RdRP domain (Table 1) were designed, and the six primer-targeted regions were 100% conserved among different TMUV reference strains (Fig. 1B).

UDG-rRT-LAMP reaction.
The developed UDG-rRT-LAMP is a one-pot, two-step assay for the specific detection of TMUV and elimination of carryover contamination. Compared with the RT-LAMP assay, our developed assay was supplemented with dUTP, which can incorporate dUTP into all RT-LAMP amplicons after amplifying the target RNA sequence under certain conditions (Table 2). Such processing was considered the first step for carryover contamination elimination ( Fig. 2A). To digest the dUTP incorporated RT-LAMP products, UDG treatment was performed before isothermal amplification in step 2 (Fig. 2B), whereas the template RNA and carryover contamination of non-dUTP incorporated RT-LAMP amplicons could not be digested under this step (Fig. 2B).
Sensitivity and specificity. The sensitivity of the UDG-rRT-LAMP assay was studied by testing each of the 10-fold serial dilutions of the transcribed NS5 RdRP domain RNA of the TMUV-SD2010 strain and extracted RNA from 10-fold serial dilutions of titrated TMUV viruses. As a result, the UDG-rRT-LAMP assay for the    Table 3). The detection limit of the cRT-PCR for TMUV was determined on the same viral RNA samples in UDG-rRT-LAMP tests for comparison between the two assays. The detection limits of cRT-PCR were 1 × 10 2.17 − 1 × 10 4.55 TCID50, indicating that the assay was 10-1000-fold less sensitive than UDG-rRT-LAMP. In addition, triplicate sensitivity tests of UDG-rRT-LAMP were performed, and the average Tt values of different dilutions were used to generate the standard curve plots (Fig. 3C). An extremely close linear correlation was observed between the Log of the NS5 gene copy numbers and the Tt values (R 2 = 0.998), with a regression line revealing an average intercept of 50.46 ± 0.32 and an average slope of − 4.23 ± 0.06.
To determine the specificity of UDG-rRT-LAMP in detecting TMUV, six other avian pathogens and negative control samples (allantoic fluid from SPF duck embryo) were included in specificity tests. No amplification signal was observed with avian influenza virus (AIV) subtypes H5N1 and H9N2, Newcastle disease virus (NDV), avian orthoreovirus (ARV), duck circovirus (DuCV), duck hepatitis A virus (DHAV), or the negative control (Fig. 3B, upper), or the negative control (Fig. 3B, upper). To further confirm the developed assay have no cross-reactions with non-TMUV flaviviruses, the synthesized last 1500 bp NS5 gene of Bagaza virus (BAGV), Israel turkey meningoencephalomyelitis virus (ITV), Japanese encephalitis virus (JEV), West Nile virus (WNV), tick-borne encephalitis virus (TBEV), and yellow fever virus (YFV) in pUC57 vector were also included in specificity test and no amplification signal was observed with these simulated viral nucleic acid samples (Fig. 3B, lower). Above results indicated that the developed UDG-rRT-LAMP assay has high specificity.

Reproducibility.
To evaluate the reproducibility of the developed UDG-rRT-LAMP method, samples containing concentrations of transcribed NS5 gene RNA of the TMUV-SD2010 strain ranging from 0.5 × 10 2 to 0.5 × 10 7 copies/μ L were tested in replicates of three in the same run (intra-assay) and individually in three The second step of UDG-based elimination of carryover contaminants by specifically cutting the LAMP amplicon DNA at the 5′ side of the dUTP-incorporated templates from previous LAMP reactions while having no effect on non-dUTP-incorporated DNA and RNA templates. During the RT-LAMP reaction, the digested contaminants are degraded into small fragments, and UDG is inactivated at approximately 63 °C, ensuring that only the RNA template is amplified.  independent runs (inter-assay) ( Table 4). In both intra-and inter-assay analyses, all of the samples yielded a positive result, and the proportion of each dilution yielding a positive result was 1.00. All triplicate amplifications of each dilution resulted in minor differences in Tt values but the standard deviation did not exceed 0.80 for any replicate. The coefficients of variation for the Tt values determined for each viral RNA concentration varied from 0.22 to 1.08% in the intra-assay analysis and from 1.06 to 3.33% in the inter-assay analysis (Table 4).
rRT-LAMP and UDG-rRT-LAMP detect simulated carryover contamination. To confirm that simulated carryover contamination from dUTP-incorporated RT-LAMP amplicons is sufficient to contaminate new reactions, we compared the detection limits of the rRT-LAMP and UDG-rRT-LAMP methods using serial diluted rRT-LAMP amplicons with concentrations ranging from 0.5 × 10 −19 to 0.5 × 10 −14 g/μ L. In the absence of UDG treatment, rRT-LAMP can detect as little as 1 × 10 −18 g of simulated carryover contamination per reaction (Fig. 4A), which is equivalent to an aerosol droplet diameter size of approximately 0.124 μ m. This result indicated that even at extremely low concentrations or small particle sizes, carryover contamination can cause false positive results. With UDG treatment, the UDG-rRT-LAMP assay could only detect 1 × 10 −15 g or more of simulated carryover contamination per reaction (Fig. 4B), equivalent to an aerosol droplet diameter size of approximately 1.241 μ m. These results demonstrated that UDG prevented the amplification of up to 1000-fold higher concentrations of carryover contaminant DNA and 1000-fold smaller sizes of aerosol droplets generated by UDG-rRT-LAMP products which significantly reduce the likelihood of false-positive results of TMUV diagnosis. A UDG concentration of 0.01 U/μ L in the UDG-rRT-LAMP reaction mixture is sufficient to eliminate as much as 1 × 10 −16 g of dUTP-incorporated carryover contamination per reaction.

UDG-rRT-LAMP can avoid false-positive results.
To confirm that UDG-rRT-LAMP can reduce the likelihood of false-positive results due to carryover contaminants, we performed sensitivity testing of rRT-LAMP and UDG-rRT-LAMP reactions with 10-fold serial dilutions of the transcribed NS5 gene and added 1 × 10 −18 g of simulated carryover contamination to each reaction tube. Without UDG treatment, we could not determine the detection limit of the assay, and all tested samples displayed positive results (Tt value of 30.8-37.4), including samples with undetectable levels of viral RNA (less than 0.5 × 10 2 copies/μ L) (Fig. 4C). These results were considered false positive. With UDG treatment, the detection limit of the rRT-LAMP assay (0.5 × 10 2 copies/μ L) (Fig. 4D) was consistent with the aforementioned sensitivity study. Such results were considered positive. From these data, the developed UDG-rRT-LAMP assay could successfully eliminate carryover contamination.

UDG-rRT-LAMP detects TMUV in clinical samples.
To assess whether the developed TMUV UDG-rRT-LAMP assay allows the detection of viral RNA in clinical samples, a total of 81 samples were analyzed by UDG-rRT-LAMP and VI combined with NS5 gene-based cRT-PCR. Of these, 62 samples were positive according to both UDG-rRT-LAMP and VI, 17 samples were negative according to both assays, 2 samples were      positive according to UDG-rRT-LAMP and negative according to VI, and 0 samples were negative according to UDG-rRT-LAMP and positive according to VI. The results of the two methods are summarized in Table 5 for comparison. Statistical analysis illustrated that the two methods have 96.88% agreement and the difference between the two methods was not significant (P > 0.05) in detecting clinical samples. Samples from non-inoculated SPF duck embryos tested negative using the TMUV UDG-rRT-LAMP assay in the same run described previously.

Discussion
TMUV was originally identified as a tropical pathogen in Malaysia during the 1970s, and the main hosts of the virus were confirmed to be different bird species, which were responsible for the maintenance and transmission of TMUV in nature 34 . Severe diseases with mortality caused by the infection of birds with TMUV were indicated for the first time in chickens during the 1990s 11 and subsequently studied in ducks in the early 2010s 1,2 . Our recent study reported a potential case of human infection with TMUV 15 , highlighting the urgent need for effective TMUV infection control and prevention methods, particularly detection methods. Several studies reported molecular methods for the detection of TMUV 1,11,24-26 . Nevertheless, those PCR-based assays require significant time, multiple steps, expensive instruments, and well-trained personnel to perform the experiments 24 . The use of LAMP technology is increasing in popularity for detection of TMUV and other viruses 27,28 because of its rapid, simple, sensitive, specific, and isothermal reaction capacity without the need for a specialized thermal cycler. In this study, we applied an advanced UDG-rRT-LAMP method to the detection of all recognized TMUV-related viruses. Unlike RT-PCR or rRT-PCR, which requires one pair of primers or one pair of primers and a probe, RT-LAMP requires at least four primers that recognize six regions of the target RNA sequence, and thus, its specificity is extremely high 30 . On the contrary, to achieve high specificity, the LAMP primers must have a strict match with their target sequences, or failure of amplification will occur 35 . Thus, to develop a universal TMUV detection assay, identifying conserved regions in the TMUV genome sequence is a critical step in RT-LAMP primer design. Recently, Dai et al. 36 provided evidence of natural recombination in TMUV and genetic divergence between various TMUV strains isolated in different years. Therefore, in the present study, we selected the complete ORFs of 13 representative TMUV reference strains from genotyping groups 1 and 2 (including subgroups 1-3) to screen highly conserved regions among these strains for UDG-rRT-LAMP assay development. Based on the visible alignment result, we found few regions in the TMUV ORFs that were sufficiently conserved for designing the LAMP primers for the universal detection of different TMUV strains. A TMUV variant, CQW1 (KM233707), was identified to be extremely different from other reference strains concerning its E, NS3, and NS5 genes, which were commonly used in detection assay development in previous studies 1,29,37 . Fortunately, we identified a 210-nt candidate region with relatively high sequence conservation in the 3′ -terminus of the NS5 gene, which corresponds to the RdRP domain with > 99% nucleotide identity. The size of this region was sufficient to meet the requirement of LAMP primer set designing. The highly conserved features of this region distinguished our NS5 RdRP domain-based UDG-rRT-LAMP assay from other published assays.
Although LAMP has been revealed to be a highly sensitive and specific assay in the clinical diagnosis of many pathogens, it was also found to be highly susceptible to carryover contamination due to its high sensitivity 32 . The LAMP result is commonly confirmed by DNA electrophoresis and direct observation after adding a fluorescent dye. In these processes, both observation methods require the opening of reaction tubes, easily generating aerosol droplets of different sizes that contain high concentrations of amplicons 38 . Our present study revealed that an extremely small amount of amplification products (1 × 10 −18 g/reaction) from previous LAMP reactions could serve as carryover contamination templates for re-amplification, resulting in false-positive results. When converting the minimum contaminants dose that can cause false-positive results to the size of aerosol droplets, the particle diameter was found to be extremely small (approximately 0.124 μ m). This could not be filtered by most routine processing methods in the laboratory. Importantly, due to the current lack of an effective strategy for eliminating LAMP carryover contamination, preventing carryover contamination relies solely on careful operation of the LAMP experiment. Once contamination occurs in the laboratory, the procedure for eliminating or reducing contamination is usually complicated, costly, time-consuming, and sometimes requires moving the experiment to a new laboratory room or designing a set of primes targeting different genomic regions 32 .
Although there were some pre-and post-amplification strategies for preventing or eliminating carryover contamination in other nucleic acid amplification assays, including mechanical/chemical barriers 38 , UV light irradiation 39 , primer hydrolysis 40 , and hydroxylamine treatment 41 , these methods may not be able to completely eliminate or prevent carryover contamination. Therefore, they are not ideal approaches for controlling carryover    Table 1.
contamination in LAMP reaction 33 . Compared with the aforementioned methods, the UDG-mediated digestion of dUTP-incorporated contaminant DNA is simple, inexpensive, and does not require substantial modification of existing protocols 42 . It specifically removes dUTP bases in uracil-containing DNA but has no effect on natural DNA 43 and RNA templates 44 . The prevention of carryover contamination using dUTP and UDG has been well documented in previously reported PCR and LAMP assays using both DNA and RNA as the template [31][32][33]45 . For RNA samples used in UDG-LAMP reactions, reverse transcription, UDG treatment, and the LAMP reaction are usually performed in separate steps 33 , which necessitates opening the reaction tube and increasing the risk of exposing the reaction mixture to carryover contaminants in the environment. Such a disadvantage limits the practical application of the UDG-RT-LAMP assay in the clinical detection of RNA viruses. Therefore, we performed rRT-LAMP amplification with UDG digestion in a one-pot reaction to eliminate carryover contaminants for the first time. The developed UDG-rRT-LAMP assay could remove as much as 1 × 10 −16 g/reaction of simulated contaminants with a UDG concentration of only 0.01 U/μ L. Although the simulated contaminants cannot be removed at high concentrations (1 × 10 −15 g/reaction or more), adding more UDG and physically separating large aerosol droplet particles could serve as backup methods. To avoid opening the reaction tube during result detection, EvaGreen ® fluorescent dye was integrated into the RT-LAMP reaction for quantitative monitoring of the amplification and observing the endpoint result visually. Moreover, Bst 2.0 warmStart ® DNA polymerase was used in our strategy to enhance dUTP incorporation, amplification speed, amplicon yield, salt tolerance, thermostability, and specificity in the assay, which are critical to the success of the UDG-rRT-LAMP assay for TMUV detection.
In conclusion, based on the complete ORF alignment of various TMUV strains from different genotyping groups, we developed a universal TMUV detection assay using rRT-LAMP combined with UDG treatment to prevent/eliminate carryover contamination. This UDG-rRT-LAMP method is a rapid, sensitive, specific, and reliable assay for diagnosing TMUV infection, and it provides a potentially powerful strategy for avoiding false-positive results in detecting other flaviviruses.

Materials and Methods
Ethics statement. Viral propagation and testing were performed according to the approved the Institutional Biosafety Committee, Shandong Agricultural University. All tissue samples analyzed in this study were collected for diagnostic purposes and approved by Animal Welfare Committee, Shandong Agricultural University (License number: 2015-Vet-012).The animal care and research were conducted according to Guide for the Care and Use of Laboratory Animals, National Research Council, 2011 (https://grants.nih.gov/grants/olaw/Guide-for-the-Careand-use-of-laboratory-animals.pdf).

Primer design.
To identify the most conserved region among 10 genes from different TMUV strains, 13 complete ORFs of TMUV strains (isolated from 2010 to 2014) were used in this study, including 11 duck, 1 layer, and 1 goose strain. These strains represent all different TMUV genotyping groups/subgroups described in a recent TMUV evolutionary analysis study 36 . The alignment of complete ORFs was conducted using the mVISTA online program (http://genome.lbl.gov/vista/mvista/submit.shtml) and the pairwise comparison of the selected conserved region was performed using the ClustalW method in Genomic Workbench V7.5 software (QIAGEN, Boston, MA, USA). Since the most highly conserved regions were located at the 3′ end of the NS5 gene among different TMUV strains, the last 210 bp of the NS5 gene were chosen as the optimal target region for RT-LAMP primer sequence basis. Primers were designed and optimized for sequence conservation using an online program (PrimerExplorer V4, http://primerexplorer.jp/elamp4.0.0/index.html). The designed primer set contains four primers: F3, Forward Inner Primer (FIP), Backward Inner Primer (BIP), and B3. FIP (BIP) consists of the sequence of the F1c (B1c) and F2 (B2) regions. The specificity of the primers was further confirmed using a BLAST search in the NCBI nucleotide database.
Virus strains, cell lines, and RNA extraction. For the UDG-rRT-LAMP evaluation test, six TUMV field variant strains isolated from ducks, ducklings, geese, house sparrows, and mosquitoes in Shandong Province were propagated in a African green monkey kidney cell line (Vero) (CCL-81, ATCC, Manassas, VA, USA), and titrated in tissue culture infectious doses (TCID50). AIV-H5N1, AIV-H9N2, NDV, ARV, DuCV, DHAV were also included in this study for specificity testing. Viral RNA was directly extracted from a 20% suspension (w/v) of viral transport medium diluted clinical samples, cell culture supernatant, and allantoic fluid using an RNeasy Mini Kit (QIAGEN, Valencia, CA, USA) following the manufacturer's instructions. Briefly, 250 μ L of sample was mixed with same volume of lysis buffer and then applied to the RNA binding column. After three washes, the contaminants were efficiently washed away, and total RNA eluted in 50 μ L of RNase-free water was used for further testing.
In vitro transcription of the partial NS5 gene. The in vitro-transcribed NS5 RdRP domain RNA of a TMUV field strain isolated from Peking duck (TMUV-SD2010) 4 was used to determine the detection limit of the assay. The NS5 RdRP domain (1084 bp) of TMUV-SD2010 was amplified using published procedures 46 and cloned into the pGEM-T easy Vector system (Promega, Madison, WI, USA) as described in our previous study 4 . The recombinant plasmid pGEM-NS5 was linearized by EcoRI (Promega), extracted using phenol-chloroformisoamyl alcohol (25:24:1), precipitated using ethanol, and suspended in water before using the DNA for in vitro transcription reactions. The production of "run-off " transcripts derived from the inserted M1 segment used the Riboprobe ® in vitro Transcription System with T7 RNA Polymerase (Promega) following the manufacturer's instructions. The transcribed RNA concentration was quantitated using a NanoDrop ™ 1000 spectrophotometer (Thermo Scientific, Waltham, MA, USA) and copy numbers were calculated as previously described 47 before making a 10-fold serial dilution for sensitivity testing and generating the standard curve.
Scientific RepoRts | 6:27605 | DOI: 10.1038/srep27605 UDG-rRT-LAMP conditions. The UDG-rRT-LAMP for the NS5 RdRP domain was conducted in a 20-μ L reaction system. The total reaction master mix volume was 18 μ L consisting of 1.6 μ M each of FIP and BIP primers, 0.2 μ M each of F3 and B3 primers, 1.6 mM dNTPs (Invitrogen, Waltham, MA, USA), 1 M betaine (Sigma-Aldrich, St. Louis, MO, USA), 4 mM MgSO 4 (New England Biolabs, MA, USA), 5 U of AMV reverse transcriptase (Promega), 100 mM dUTP (New England Biolabs), 0.2 U UDG (New England Biolabs), 1× ThermoPol reaction buffer, 8 U of Bst 2.0 warmStart ® DNA polymerase (New England Biolabs), and 1× EvaGreen ® fluorescent dye (Biotium, Hayward, CA, USA). After adding 2 μ L of viral RNA, the reaction mixture was incubated at room temperature for 5 min to eliminate carryover contamination. Following incubation, the reaction was performed using the 7500 Real-time PCR System (Applied Biosystems, Foster City, CA, USA) for rRT-LAMP or with a water bath for RT-LAMP at 63 °C for 1 h. The thermal cycling profile for rRT-LAMP proceeded as follows: 60 cycles of 62.5 °C for 5 s and 63.5 °C for 55 s. Fluorescence signals for each sample were collected at the end of the 63.5 °C step. The time threshold (Tt) values and standard curve were analyzed using the SDS software program (version 1.4) and visualized using OriginPro 8.5 software (OriginLab, Northampton, MA, USA). The rRT-LAMP results were also detected by fluorescence observation under UV light and 2% (w/v) agarose gel electrophoresis. Sensitivity, specificity, and reproducibility of UDG-rRT-LAMP. The sensitivity of the UDG-rRT-LAMP assay was determined in two separate systems. One system was transcribed NS5 RdRP domain RNA of the TMUV-SD2010 strain, which was diluted from 0.5 × 10 7 to 0.5 × 10 0 copies/μ L. The other system was extracted RNA from 10-fold serial dilutions of titrated TMUV field variant strains isolated from different hosts in Shandong province. Three replicates for each dilution of different samples were tested using the developed UDG-rRT-LAMP assay. The detection limit of the assay was determined as the last dilution at which all three replicates from each dilution gave a positive result. Samples of serial dilutions of six TUMV field variant strains were also tested by NS5 gene-based cRT-PCR 46 to compare its sensitivity with that of UDG-rRT-LAMP.
The specificity of UDG-rRT-LAMP was first examined using viral RNA from different avian pathogens of AIV-H5N1, AIV-H9N2, NDV, ARV, DuCV, and DHAV to confirm that no cross-reactions with non-targeted viral RNA occurred. The allantoic fluid from SPF duck embryos was also included in the test as a negative control. To further confirm the developed assay have no cross-reactions with non-TMUV flaviviruses, the last 1500 bp NS5 gene sequences of BAGV (GenBank accession no. EU684972), ITV (GenBank accession no. KC734549), JEV (GenBank accession no. AF080251), WNV (GenBank accession no. AY532665), TBEV (GenBank accession no. FJ402886) and YFV (GenBank accession no. U54798) were synthesized and cloned into pUC57 vector (Genscript, Nanjing, Jiangsu, China) for specificity test purpose and the empty pUC57 vector was working as a negative control.
The reproducibility of the developed assay was assessed using transcribed NS5 RdRP domain RNA from the TMUV-SD2010 strain. The RNA was diluted from 0.5 × 10 7 to 0.5 × 10 2 copies/μ L to prepare the intra-and inter-assay samples. Replicate samples of each dilution were stored at − 80 °C immediately after preparation. For the intra-assay test, three replicate samples from each dilution were tested in the same run. For the inter-assay test, samples from different dilutions were tested in three independent runs. Reproducibility was measured by calculating the proportion of positive samples and determining Tt values of each sample. The coefficient of variation was determined to find statistical correlations.

Elimination of carryover contamination by UDG-rRT-LAMP.
To evaluate the ability of the developed UDG-rRT-LAMP method to reduce false-positive results due to carryover contaminants in detecting TMUV, we conducted both rRT-LAMP and UDG-rRT-LAMP reactions by adding 2 μ L of diluted viral RNA templates (0.5 × 10 −1 − 0.5 × 10 4 copies/μ L) and 2 μ L of simulated carryover contamination of 0.5 × 10 −18 g/μ L in the same reaction tube. The total mass of the simulated carryover contamination for each reaction was approximately equivalent to a 0.124 μ m-diameter aerosol droplet, which cannot be efficiently blocked by either high efficiency particulate air (HEPA) filters 48 in the biosafety cabinets or fibrous pipette tip filters 49 . The detection limits of the rRT-LAMP reactions with and without UDG treatment before amplification were compared to confirm whether the studied UDG-rRT-LAMP assay can eliminate false positives.
UDG-rRT-LAMP detect TMUV in clinical sample. Eighty-one original tissue specimens, including theca folliculi, oviduct, heart, and liver from ducks or other avian species that exhibited clinical signs and lesions consistent with those of TMUV infections during necropsies, were used to evaluate UDG-rRT-LAMP in detection TMUV from clinical poultry specimens. The collected tissue samples were minced using scissors in viral transport medium 50 to produce a 20% suspension (w/v). The tissue suspension was further processed in a stomacher blender (Model 80, Seward Ltd., Davie, FL, USA) and centrifuged at 1200 rpm for 10 min at 4 °C. Thereafter, the processing tissue supernatant was filtered through a 0.45-nm syringe filter for UDG-rRT-LAMP detection and virus isolation (VI) in Vero cells. TMUV-positive isolates were confirmed by NS5 gene-based cRT-PCR 46 to detect TMUV-infected cytopathic effect cells from the specimen-inoculated Vero cell cultures. Lastly, the agreement between UDG-rRT-LAMP and VI results was summarized to investigate the clinical performance of our