Application of cross-priming amplification (CPA) for detection of fowl adenovirus (FAdV) strains

Fowl adenoviruses (FAdVs) are widely distributed among chickens. Detection of FAdVs is mainly accomplished by virus isolation, serological assays, various polymerase chain reaction (PCR) assays, and loop-mediated isothermal amplification (LAMP). To increase the diagnostic capacity of currently applied techniques, cross-priming amplification (CPA) for the detection of the FAdV hexon gene was developed. The single CPA assay was optimised to detect all serotypes 1-8a-8b-11 representing the species Fowl aviadenovirus A-E. The optimal temperature and incubation time were determined to be 68 °C for 2 h. Using different incubation temperatures, it was possible to differentiate some FAdV serotypes. The results were recorded after addition of SYBR Green I® dye, which produced a greenish fluorescence under UV light. The CPA products separated by gel electrophoresis showed different “ladder-like” patterns for the different serotypes. The assay was specific for all serotypes of FAdV, and no cross-reactivity was observed with members of the genus Atadenovirus, duck atadenovirus A (egg drop syndrome virus EDS-76 [EDSV]) or control samples containing Marek’s disease virus (MDV), infectious laryngotracheitis virus (ILTV) or chicken anaemia virus (CAV). The results of the newly developed FAdV-CPA were compared with those of real-time PCR. The sensitivity of CPA was equal to that of real-time PCR and reached 10−2.0 TCID50, but the CPA method was more rapid and cheaper than the PCR systems. CPA is a highly specific, sensitive, efficient, and rapid tool for detection of all FAdV serotypes. This is the first report on the application of CPA for detection of FAdV strains.

For a long time, the primary role of FAdVs in pathogenesis was not fully clarified, as FAdVs were isolated from both clinically healthy and diseased birds, or together with other immunosuppressive agents (e.g., infectious bursal disease virus [IBDV] and chicken anaemia virus [CAV] [23]). Currently, it is acknowledged that virulent FAdVs belonging to certain serotypes of different species act as a primary pathogens, inducing characteristic clinical signs alone without co-infection with any other pathogens. This knowledge came from the emergence of highly virulent strains during the late 1980s and 1990s (hydropericardium hepatitis syndrome [HHS] outbreaks in Pakistan, spreading to other countries of the world) [7,10] and successful experimental induction of characteristic signs by inoculation of SPF chickens with strains isolated from field outbreaks. Development of improved methods for the differentiation of FAdV strains has helped to elucidate the pathogenic role of FAdVs causing different clinical syndromes such as inclusion body hepatitis (IBH), HHS, and gizzard erosion and ulceration (GEU) [5,7,8,13].
The clinical picture of IBH is caused predominantly by strains from serotypes FAdV-2 and -11, representing FAdV-D, and serotypes FAdV-8a and -8b, representing FAdV-E. The current perception of the role of pathogenic FAdVs has changed, since some serotypes are able to induce specific clinical signs without co-infection with additional agents. Indeed, some FAdV strains are responsible for the clinical picture of IBH. These etiological agents are predominantly strains belonging to the species Fowl aviadenovirus D, representing serotypes FAdV-2, 3, 9, and 11, and the species Fowl aviadenovirus E, representing serotypes FAdV-6, 7, 8a, and 8b [2]. Adenoviruses have been isolated from sick birds as well as from birds without any clinical signs of infection [10,11]. FAdVs may cause disease independently, or they could be one of the factors of multi-aetiological syndromes [2]. Another example is HHS [7], which is caused by the serotype FAdV-4, which belongs to the species Fowl aviadenovirus C, and GEU, which is mainly induced by serotype FAdV-1, species Fowl aviadenovirus A [8,30,31].
FAdVs may also cause immunodeficiency or vaccination failure in chickens [26]. Due to the serious economic loss they are able to cause, FAdVs represent a major concern in the poultry industry.
Several PCR assays for the detection of members of different FAdV species and serotypes have been developed in the last 15 years, including PCR combined with restriction enzyme analysis [20]. Methods based on PCR are useful and able to identify several FAdV serotypes [1,17,39,44]. PCR was described for the first time for the detection of pIIIa and pIII genes of serotypes FAdV-8 and FAdV-1 [14], and nested-PCR assays based on the polymerase gene [16,41] and duplex-PCR assays for detection of serotypes FAdV-1 and FAdV-5 have also been developed [27]. Recently, PCR targeting the hexon gene was developed [18,20,21,37,44]. The same technique was also used for the detection of haemorrhagic enteritis virus (HEV) and egg drop syndrome virus (EDSV), belonging to the genera Siadenovirus and Atadenovirus, respectively [11,18,19]. In addition, real-time PCR has been shown to be a useful tool for detection of members of the five FAdV species (A-E) [9].
Novel techniques based on isothermal amplification, especially loop-mediated isothermal amplification (LAMP), have been developed previously for the detection of many poultry pathogens, including infectious bronchitis virus (IBV) [2], chicken anaemia virus (CAV) [12], avian influenza virus (AIV) [36], Newcastle disease (NDV) [34], infectious bursal disease (IBDV) [46] and Marek's disease virus [42,43]. Recently, the LAMP assay was shown to be a useful and sensitive method for identification of FAdV [26,28,40,45]. Detection of specific products is performed under UV illumination after addition of SYBR Green I Ò dye, resulting in greenish fluorescence in positive samples. A recent innovation of isothermal amplification methods is cross-priming amplification (CPA), which is primarily invented by Ustar Biotechnologies (Hangzhou, China) and described by Rendong et al. [38]. CPA involves five to six specific primers and leads to cross-priming amplification with an intermediate stage of hairpin-shaped products. Forward cross primer (sense) and reverse cross primer (anti-sense) are aligned with the 5 0 end of the sequence. The second defined priming sites of the CPA product are extended by inner and outer primers. After the reaction, the mixture contains several micrograms of amplified DNA products of different length, which are visualised using fluorescent dyes that bind to double-stranded nucleic acids. Bst and Bsm polymerases have been used previously because of their isothermal polymerase activity. More recently, however, the isothermal polymerase GspSSD polymerase, isolated from Geobacillus sp. has been used due to its improved activity and resistance to reaction inhibitors. So far, the CPA protocol has been adapted mainly for the detection of Mycobacterium tuberculosis. A schematic concept of CPA was presented by Rendong et al. [38,47,48]. CPA has not been developed so far for the detection of any avian pathogens.
The aim of this study was to develop and optimise CPA using five specific primers corresponding to hypervariable regions HVR-1-4 of the hexon gene, which are the most variable regions of the FAdVs genome.

Virus propagation
Chicken Embryo Fibroblasts (CEFs) were prepared from 11-day-old SPF chicken embryos (Lohmann Animal Health, Cuxhaven, Germany) according to the standard procedure. The growth medium was Eagle's medium (MEM) supplemented with 10 % bovine serum and 0.1 % antibiotic mixture (Antibiotic-Antimycotic, Gibco, Paisley, Scotland). The maintenance medium consisted of MEM with 0.1 % antibiotic mixture. A monolayer of CEFs was obtained after about 24 h of incubation at 37.5°C and 5 % CO 2 and was inoculated with reference FAdV strains and with homogenates obtained from internal organs of sick chickens. The organ samples were prepared by a triple freezing and thawing procedure, centrifugation, and filtering through filters with 450 nm pore size (Millipore, Billerica, USA). Infected cells were incubated until a cytopathic effect (CPE) was observed in 80 % of the monolayer, and collected supernatants were used for the next passage. The material from the third passage was used for DNA extraction.

DNA extraction
FAdV DNA was extracted from 200 ll of infected CEF culture using a QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. Negative control DNA for CPA and real-time PCR were extracted from uninfected CEFs. The DNA was stored at -20°C for the next step of the study as template for CPA.

PCR
The reaction was carried out in a basic gradient thermocycler (Biometra, Germany) in a final volume of 25 ll of reaction mix. The mixture contained 2.5 ll of PCR buffer, 1 ll of dNTP (10 mM), 1.5 ll of each pair of primers, 4 ll of total DNA isolated from reference strains, and 11.5 ll of sterile water. Pre-denaturation was done at 95°C for 5 min, followed by 35 cycles of denaturation at 94°C for 45 s, primers annealing at 61°C for 1 min, and chain elongation at 72°C for 2 min, and then a final elongation step was carried out at 72°C for 10 min.

Analysis of PCR products
After the amplification reaction, electrophoresis was conducted in a 2 % agarose gel with 1 lg of ethidium bromide per mL in a Mini-Sub Cell (Bio-Rad, USA). Electrophoresis was carried out at 150 V and 80 mA for 50 min in Tris-borate-EDTA buffer, pH 8.2. The sizes of amplified products were compared with a MassRuler 1,031-bp DNA marker (Fermentas). The bands were visualised using a UV transilluminator and then photographed and analysed. The results were considered positive when the products obtained with a specific pair of nucleotide primers had the predicted size.
Sequencing and phylogenetic analysis PCR products were purified using a NucleoSpin Extract II Kit (Marcherey-Nagel, France) and then commercially sequenced by GENOMED (Warsaw) using a GS FLX/Titanium sequencer (Roche, Switzerland). Phylogenetic analysis was performed by alignment of the nucleotide sequences of the amplified fragments of the hexon gene originating from the field adenovirus strain FAdV with those of the fowl adenovirus reference sequences from the GenBank database that had been identified according to the general obligatory classifications. A phylogenetic tree was generated by the neighbor-joining method by the p-distance method (on 1000 bootstrapped datasets). Four sequences of reference human adenovirus strains were also used in the dendrogram. The analyses were performed using the computer software MEGA5, Geneious 6, and BLAST. On the basis of this analysis, a phylogenetic tree was constructed and the relationships between the examined adenovirus strains were determined.

Design of FAdV-CPA primers
The loop-1 region of the hexon gene of FAdV-A reference strain CELO (GenBank accession number AF339914) was compared with other available hexon gene sequences of FAdV reference strains in order to select a conserved sequence about 151 bp long as the detection target. Five primers complementary to the hexon gene sequence of FAdV were designed manually. The locations of the CPA primer sequences in the hexon gene are shown in Fig.1.

Optimisation of FAdV-CPA
A reaction mix was made on ice in 0.2-ml optical tubes with five complementary primers. Reaction parameters were optimised using Isothermal Mastermix (OptiGene, Horsham, West Sussex, UK) with different primer concentrations: 1s (10-40 pM), 2a and 3a (5-20 pM), 4s, and 5a (2.5-10 pM). The temperature of incubation was optimised at 62°C, 64°C, and 68°C. The last parameter optimised was the incubation time, which ranged from 30 min to 3 h. After incubation, 1 ll of 10,000-fold-diluted SYBR Green Ò I dye (Invitrogen, Germany) was added to each tube. Positive reactions gave greenish fluorescence under UV light. The CPA products were analysed in agarose gels stained with 2 % GelRed and formed a ''ladder-like'' pattern with sizes ranging from 80 to 1031 bp when compared to the molecular-length marker Gen-eRuler TM 100 bp DNA Ladder Plus (Thermo Scientific, Waltham, Massachusetts, USA). Agarose gels were documented and photographed under UV light (GenoSmart Gel documentation system, VWR, Germany). During the optimization, all samples were tested in triplicate.

Specificity and sensitivity of FAdV-CPA
The specificity of the assay was evaluated using DNA extracted from uninfected CEFs as well as from the MDV Rispens/CVI988, ILTV, CAV, and EDS strains.
The sensitivity of the CPA was determined using tenfold dilutions (10 0 to 10 -5.0 ) of 2 ll DNA extracted from 200 ll of infected cell culture supernatant containing 10 3.0 TCID 50 /ml of FAdV-A reference strain CELO. The CPA products were then analysed by electrophoresis in 2 % agarose gels. The sensitivity was defined as the highest dilution showing greenish fluorescence. Thirty field strains of FAdV were examined using CPA under conditions optimised with reference strains. The assay was able to detect a 10 -5.0 dilution of DNA extracted from infected cell culture supernatant containing 10 3.0 TCID 50 FAdV-A/ CELO.
The specificity of FAdV-CPA was demonstrated using DNA extracted from Rispens/CVI988, EDS, ILTV, and CAV strains. All samples containing DNA of these viruses were negative (Fig. 4).
During the next step of this study, we applied the optimised CPA assay to detect 30 field adenovirus strains from the collection of the Department of Poultry Viral Diseases (National Veterinary Research Institute, Pulawy, Poland), representing to seven different serotypes (FAdV-1, FAdV-2/11, FAdV-4, FAdV-5, FAdV-7, FAdV-8a, and FAdV-8b). The DNA of all investigated FAdV field strains was successfully detected by the CPA. A greenish fluorescence was observed in positive samples under UV light, as was the characteristic ''ladder-like'' pattern of amplification products after gel electrophoresis (Fig. 5).

Discussion
Early diagnosis is crucial for controlling FAdV infections. Classical virological methods such as virus isolation in SPF chicken embryos or in cell culture are laborious and timeconsuming. In spite of this, they are still used as a ''gold standard'' [22].
In recent years, molecular biological methods based on PCR have been frequently used for detection and identification of adenoviruses [1,17,[19][20][21][22][26][27][28]. Our study was conducted to develop and optimise a specific, sensitive, and rapid tool for the detection and identification of FAdV strains. A variable region of the FAdV hexon loop-1 was selected as the target region for amplification, using CPA primers corresponding to a region with a high degree of conservation between all FAdV serotypes. The CPA method described here relies on the isothermal amplification of nucleic acids. CPA is catalysed by Bsm, Bst or GspSSD polymerase and does not require an initial denaturation step or other enzymes [38]. The results of this study show that CPA is a specific detection method for all FAdV serotypes and could potentially be used to differentiate between different serotypes of the virus based on the observation that each serotype yields a characteristic ladder-like pattern of bands. CPA can be performed without any special laboratory equipment. The optimised CPA method was used to detect 30 field adenovirus strains representing seven serotypes that are found on poultry farms in Poland. Using the CPA method under different reaction conditions it was possible to differentiate particular serotypes of FAdV. The CPA, which could detect 10 -5.0 TCID 50 /ml of virus with a titer of 10 3.0 TCID 50 , is more sensitive than a previously described FAdV-LAMP, which could detect 10 -4.0 TCID 50 /ml of virus with a titer of 10 3.0 TCID 50 [26]. Moreover, the CPA has the same sensitivity as a previously reported real-time PCR [26]. This is in agreement with an earlier study showing CPA and real-time PCR to have similar sensitivity for detection of thrombocytopenia syndrome virus [3]. The CPA reaction time was 2 h and allowed up to 100 copies of the virus to be detected. It therefore seems to be highly suitable for detection of FAdV DNA under field conditions and thus has great diagnostic value.
In conclusion, for the first time, we have developed and evaluated a CPA for rapid identification of FAdVs, a viral pathogen that is responsible for an economically important disease of chickens. The CPA was found to be a rapid and very sensitive tool that can be used by poorly equipped laboratories. The evaluated CPA method has the potential to be used even by mobile laboratories and small veterinary diagnostic units.