Since their first detection in South Africa in the late 1970s, avian metapneumoviruses (AMPV), family Pneumoviridae, genus Metapneumovirus (1), have now been detected in most regions of the world (2). AMPV fundamentally cause respiratory and reproductive diseases in poultry and have had serious economic impact on the poultry industry. To date, based on genetic and antigenic differences, four subgroups are described (AMPV-A, B, C, and D) (3–8) and two genetic lineages of AMPV-C (8) have been identified. Recently, two divergent AMPVs from a monk parakeet (9) and a gull (10) were described and do not cluster genetically with any of the four previously described subgroups. The principal host species for AMPVs are turkeys, chickens and ducks, however, they have also been isolated from pheasants (11), and been detected in guinea fowls (12), or wild water fowls (13) by serological and molecular methods, respectively.

Concerning field detection of AMPVs in the three principal poultry species (turkeys, chickens and ducks), all subgroups have been detected in turkeys (14–17), although to date, the subgroup C viruses detected in this species have been of a lineage unique to North America (US AMPV-C). In domestic Muscovy and Pekin ducks, only AMPV-Cs of a Eurasian lineage have been reported (18, 19). In chickens, AMPV A, B and an AMPV-C have been described (20–24), however the lineage of the latter is yet to be defined. Recently, details on the extent of infection of the four recognized subgroups and lineages of AMPV in turkeys, chickens and Muscovy ducks under controlled experimental conditions have been published (25). Interestingly, apart from confirming the subgroup / host relationships described above, this study showed that AMPV-C viruses of both US and Eurasian lineage could be re-isolated from chickens, and AMPV-C viruses of Eurasian lineage from turkeys despite a lack of viral RNA detection using a well-characterized real-time RT-PCR (26) targeting the hypervariable SH gene. These results demonstrate that replication of AMPVs in a “non-conventional host” may modify the viral consensus sequence, especially in hypervariable regions of the genome, and thus go undetected by molecular tests, such as those currently available, that target these regions.

To date, both classical and real-time RT-PCRs have been developed for either broad range (3, 27) or subgroup specific detection. Concerning real-time RT-PCR tests that are subgroup specific, tests exist in duplex for subgroups A and B (26, 28, 29), in triplex for subgroup C (both lineages) and the new identified parakeet and gull AMPVs (30) and in individual tests for subgroups C and D (26). All these current subgroup specific molecular tests are based in the hypervariable G or SH genes except for that detecting the AMPV gull virus which is in the polymerase L gene.

In this context, the development of a multiplex reaction PCR fully based on a non-hypervariable region of the genome that is capable of detecting and differentiating the four recognized subgroups found in poultry would be beneficial. Theoretically, this could be addressed using recent real-time PCR machines that now have the capability of reading up to four different fluorescent wavelengths. In reality however, real-time PCRs that differentiate more than three targets are rare due to the level of competition between primers, probes and target molecules within the reaction mixture. The technology of droplet digital PCR (ddPCR), based on the amplification of target molecules of the same mixture but distributed into thousands of reaction droplets, reduces this competition and results in an improved capacity for multiplexing (31). ddPCR is based on end-point PCR reactions and uses droplet reader machines capable of reading several fluorescent wave lengths. Different positive targets can be separated into specific clusters using probes labeled with different fluorescent dyes but also by using probes at different concentrations that are labeled with the same dye. The method of generating specific clusters by varying probe concentrations is based on the fact that the fluorescent amplitude value is directly linked to the number of probe molecules in the PCR reaction mixture. Combining these two techniques means that many specific clusters can be obtained even with a two-color reader.

The current study describes the development of a five-plex RT-ddPCR for the identification of AMPV-A, B, C or D and the housekeeping gene Glyceraldehyde-3-phosphate dehydrogenase (GAPDH). For the amplification of AMPV molecules, two common primers and four subgroup specific probes were used which targeted a small region of the conserved viral polymerase gene. Distinction between subgroups was achieved on the principle of varying probe concentrations.

Multiplex PCRs are difficult to develop mainly because of the competition between the molecules in the same reaction mixture, however, when one PCR reaction mixture containing many different molecules can be separated into thousands of individual reactions (droplets in the case of ddPCR), then competition is drastically reduced without losing the multiplexing capabilities of the reaction as a whole. Here, a five-plex RT-ddPCR for the detection and differentiation of any one of the four recognized AMPV subgroups has been achieved using an amplitude-based multiplexing approach with double-quenched probes at different final concentrations coupled with the same dye (31) to achieve sufficient cluster separation for clear identification of each subgroup. Amplitude reads on CH1 corresponding to FAM labeled probes were more extensive on the value scale than those of HEX, and for this reason, three targets (AMPV-A, B and D) were separated using this fluorochrome. Only two targets were used with HEX (AMPV-C and GAPDH). Ten clearly defined clusters were obtained: one for each AMPV subgroup (four in total), one for each AMPV subgroup in association with GAPDH (four in total), one for GAPDH only and a final cluster for negative droplets absent of target molecules (Figure 1).

Prior to the current study, all but one subgroup specific AMPV PCR have targeted open reading frames (ORFs) of high nucleotide variability between subgroups such as the SH or G ORFs (26, 28–30, 36) because sequences specific to each subgroup in these regions are in abundance. However, the problem with this approach is that hypervariable sequences are subject to frequent nucleotide change implicating potential mismatch with primers and/or probes, as has been previously evoked (25). The current RT-ddPCR has been therefore developed using a low-variability region in the L ORF (4, 34) located between nucleotide positions 5′ 1,980 and 5′ 2,158 to overcome the above-mentioned limitations. Other conserved ORFs in the AMPV genome, such as that coding for the nucleocapsid protein (N), which are more abundant as viral RNA molecules in an infected cell due to their proximity to the 3′ leader end of the genome (37, 38), were initially considered as targets for the design. However, no suitable regions could be found.

As described, the main focus of the current study was on the multiplexing aspect of the RT-ddPCR and not on its quantification capabilities. However, one advantage of developing ddPCRs is that absolute quantities of target molecules are calculated automatically without the need for the development of reference standards. Quantities in ddPCR are calculated using the “Poisson distribution model” based on the total number of droplets in a given reaction volume and the number of positive droplets within this population. This type of quantity estimation is not established by referring to a concentration curve generated using RNA transcripts that may give imprecise measures due to the initial method of dosage. Thus, quantities of target molecules in the initial sample as calculated by ddPCR could be interpreted to be more accurate; however, the efficiency of the RT-PCR also needs to be taken into account. The LoDs of the five-plex RT-ddPCR developed in the current study were in the same order as those described previously for subgroup specific AMPV real-time RT-PCR tests (26, 28). However, as described above, the differences in the methods of quantification by RT-ddPCR and real-time RT-PCR means that comparisons should be treated with caution.

In conclusion, an end-point RT-ddPCR has been developed that is capable of identifying AMPV-A, B, C and D in a single reaction. This multiplex RT-ddPCR is based on a highly conserved region of the genome and thus should have less risk of being affected by viral genome evolution and the impact it has on the specificity of such tests. A diagnostic test of such type with greater durability in terms of specificity will also help improve the accuracy when assessing the distribution of AMPV-A, B-C and D in the field. Further developments of this method could include the ability to detect co-infections with different AMPV wild type viruses or of wild type virus and vaccine strains, both unreported to date. Finally, because the primers used for this test also match the L gene sequences of the two newly identified gull and parakeet AMPVs, it may be possible to extend this five-plex RT-ddPCR to a seven-plex using machines offering more numerous fluorescent color options.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

EL: study design, experiments, analyses, and writing of the manuscript. SB: statistical analysis of data. CA: experiments. CC and NE: critical analysis and contribution to manuscript text. PB: supervision of experiments and manuscript text. All authors reviewed and approved the manuscript.

This work was supported by the “Appel à Manifestation d'Intérêt Transversalité Inter-Laboratoires et autres Directions” of Anses, the Département des Côtes d'Armor, Saint-Brieuc Armor Agglomération, the Conseil Général des Côtes d'Armor (CG22), and the Conseil Régional de Bretagne.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.