M. Kariuki Njenga
Humphrey Lwamba
Richard Bennett,
David Halvorson
Department of Veterinary Pathobiology,
University of Minnesota,
St. Paul, MN 55108,
U.S.A.
Several years after the first outbreak of Avian pneumovirus (APV) in commercial turkeys the US, the disease continues to ravage turkey flocks primarily in the state of Minnesota. Genetic analysis of more than seventeen US viruses isolated between 1996 and 2000 demonstrated that the US viruses were relatively homogenous, even viruses isolated from wild birds. However, the US viruses were different from the European viruses. Most interestingly, the US viruses were shown to be closely related an isolated human pathogen, Human metapneumovirus, which has been detected in Netherlands, Canada, Australia, and United States.
Introduction
Avian pneumovirus (APV) induces a highly contagious acute infection of the upper respiratory tract in turkeys sometimes referred to as turkey rhinotracheitis disease. The US was considered free from APV until an outbreak of upper respiratory tract infection among turkeys occurred in the state of Colorado in 1996. At about the same time, APV outbreaks were reported in the north central state of Minnesota, where six years later (1997 – 2002), APV has emerged as a major economic problem for turkey farmers. For example, over 37% of commercial turkey flocks in the state have experienced annual APV outbreaks since 1999, resulting in economic losses estimated at 15 million US dollars per year.
In Minnesota, most outbreaks occurred within counties located at the central region of the state, which are also the counties with the highest population of commercial turkeys. Turkey sera from the neighboring states of North Dakota, South Dakota, Iowa, and Wisconsin were negative for APV antibodies in 1998. However between 1999 and 2001, a small number of commercial flocks from North Dakota, South Dakota, and Canada, tested positive for APV antibodies.
Seasonal outbreaks
Analysis of the annual prevalence of the outbreaks in Minnesota revealed increased incidence of APV outbreaks during the spring (April through May) and fall (October through December) among turkeys. This seasonal trend of APV suggested that environmental factors such as migratory birds were involved in the spread of the virus. Consistent with this hypothesis, RNA belonging to APV genes was isolated from nasal turbinates of wild sparrows, geese, blue-winged teal and starlings.
The nucleotide and predicted amino acid sequences of M and F gene RNA isolated from these wild birds had 90% to 95% nucleotide and 97% to 99% amino acid sequence identity with viruses isolated from turkeys. Later, infectious APV was isolated from sentinel ducks placed in a pond in close proximity with an APV infected turkey farm and this isolate (APV/MN-12/C) had high nucleotide and predicted amino acid sequence identity with four virus strains isolated from a neighboring turkey farm. Recently, infectious APV was isolated from Canada geese captured in Minnesota. Whereas these findings clearly demonstrate that APV can replicate in wild birds, the epidemiologic picture of US APV outbreaks has made it difficult to ascribe a major role to migratory bird in disseminating the virus. This is because Canada, which neighbors Minnesota to the north; and states neighboring Minnesota to the east, west and south (Iowa, Illinois, South Dakota, North Dakota, and Wisconsin) have not reported serious APV outbreaks.
Sequence comparison of the first 5 genes
Comparison of the nucleotide and amino acid sequences of nucleoprotein (N), phosphoprotein (P), matrix (M), fusion (F), and second matrix (M2) genes of 15 U.S. APV strains isolated between 1996 and 1999 revealed between 89% and 94% nucleotide sequence identity, and 81% to 95% amino acid sequence identity. In contrast, genes from U.S. viruses had 41% to 77% nucleotide sequence identity and 52% to 78% predicted amino acid sequence identity with European subgroup A or B viruses, confirming that U.S. viruses belonged to a separate subgroup. Of the five proteins analyzed in U.S. viruses, P was the most variable (81% amino acid sequence identity) and N the most conserved (95% amino acid sequence identity). Phylogenetic comparison of subgroups A, B, and C viruses indicated that A and B viruses were more closely related to each other than either A or B viruses were to C viruses.
Sequence comparison of the highly variable G gene
In a major breakthrough towards completing the genome sequence of the US viruses, we determined gene nucleotide and predicted amino acid sequence the highly variable cell attachment (G) glycoprotein of US APV viruses (APV/Co and 8 Minnesota isolates) and showed that the nucleotide sequence comprised 1321 nucleotides with only one predicted open reading frame encoding a protein of 435 amino acids, with a predicted Mr 48,840. Structurally, the predicted G protein of US APV viruses maintained similar characteristics to the human respiratory syncytial virus (hRSV) attachment G protein. Comparison of the deduced G protein sequence of US APV with that of European APVs (subtypes A, B, and D) revealed an overall predicted amino acid sequence identities ranging from 4 to 16.5%, suggesting a distant relationship. However, G protein sequence identities among the US viruses ranged from 72% to 97% and 21% when compared with the hMPV human virus.
Isolation of Human metapneumovirus
Human metapneumovirus (hMPV) was first isolated in the Netherlands in 2001 from nasopharyngeal aspirates collected over 20 years from young children suffering mild to severe respiratory tract illnesses characterized by cough, bronchiolitis, and pneumonia. Characterization of the 28 hMPV strains, including cytopathic effect on tertiary monkey kidney cells, absence of hemagglutinating activity, morphology by electron microscopy, and high sequence homology with APV indicated that the viruses were different from human respiratory syncytial virus, resulting in their classification into the Metapneumovirus genus. Retrospective studies showed high seroprevalence of hMPV antibodies among humans in 1958 in the Netherlands, indicating that the virus had been circulating in human populations for at least 45 years.
Following this report, clinicians worldwide began to aggressively test for the virus in cases often suspected of human respiratory syncytial virus, whose clinical symptoms are indistinguishable from those associated with hMPV infection. The hMPV virus was reported in Australia where serological screening showed a high incidence of the viruses among children between 2 and 4 years old, and in Canada from patients ranging from 2 months to 87 years of age. Some reports suggest that hMPV is widely distributed and more isolation is expected now that molecular and diagnostic tools are available. Complete genome sequencing of hMPV has confirmed the genomic organization similar to APV consisting of the eight genes and nine open reading frames.
Sequence comparison between hMPV and APV
Early sequencing data from the first hMPV isolates from Netherlands indicated high levels of nucleotide identity and a genomic organization similar to APV, particularly the absence of the NS genes at the 3' end of the genome. Comparison of partial sequences of N, M, F, and L genes from nine hMPV isolates (from samples collected between 1993 and 2000) in Netherlands revealed clustering suggestive of two subgroups within hMPV, with nucleotide sequence identity between 90 and 100% for viruses within a cluster, and between 81 and 88% for viruses from different clusters. The complete genome sequencing of the hMPV 00-1 strain enabled detailed comparisons between the four APV subgroups (APV/A, AV/B, APV/C, APV/D) and hMPV. The results revealed between 56 and 88% amino acid sequence identity between APV and hMPV within the nine open reading frames, and some sequence identity within the non-coding regions.
Most importantly, comparison of N, P, M, F, and M2 proteins demonstrated that the highest amino acid identity (overall 80%) was between APV/C and hMPV, significantly higher than amino acid identity between hMPV and either APV/A or APV/B. In fact, the overall percentage of sequence identity between the N, P, M, F, and M2-1 genes of hMPV with APV/C was sometimes higher than that observed within intra-subgroup comparison of APVs. Because only the F gene sequence of APV/D is available, we compared the amino acid sequence alignments for the F protein and performed phylogenetic analysis to further demonstrate hMPV closeness to APV/C strains.
The close relationship between hMPV and APV/C, lead to the hypothesis that either of the two viruses can cross infect, i.e hMPV can infect turkeys and/or APV/C infect humans. Preliminary studies in which juvenile turkeys, chickens and cynomologous macaques were inoculated with hMPV resulted in virus replication and mild respiratory clinical signs in monkeys, but not in turkeys or chickens. The investigators suggested that the restricted host range of hMPV might be because the attachment G gene has low amino acid sequence identity with APV/C. However, the APV/C G gene sequence is not yet available for comparison. It is worth noting that APV had been detected in Netherlands in 1992 and unpublished reports indicate that several turkey flocks in Canada are positive for APV. However, the detection of hMPV in Australia is intriguing because there is no previous report of APV on that continent.
From Proceedings of the "Midwest Poultry Federation Convention", St. Paul, Minnesota, U.S.A.
