Newcastle Disease Viruses Causing Recent Outbreaks Worldwide Show Unexpectedly High Genetic Similarity to Historical Virulent Isolates from the 1940s

Abstract

Virulent strains of Newcastle disease virus (NDV) cause Newcastle disease (ND), a devastating disease of poultry and wild birds. Phylogenetic analyses clearly distinguish historical isolates (obtained prior to 1960) from currently circulating viruses of class II genotypes V, VI, VII, and XII through XVIII. Here, partial and complete genomic sequences of recent virulent isolates of genotypes II and IX from China, Egypt, and India were found to be nearly identical to those of historical viruses isolated in the 1940s. Phylogenetic analysis, nucleotide distances, and rates of change demonstrate that these recent isolates have not evolved significantly from the most closely related ancestors from the 1940s. The low rates of change for these virulent viruses (7.05 × 10⁻⁵ and 2.05 × 10⁻⁵ per year, respectively) and the minimal genetic distances existing between these and historical viruses (0.3 to 1.2%) of the same genotypes indicate an unnatural origin. As with any other RNA virus, Newcastle disease virus is expected to evolve naturally; thus, these findings suggest that some recent field isolates should be excluded from evolutionary studies. Furthermore, phylogenetic analyses show that these recent virulent isolates are more closely related to virulent strains isolated during the 1940s, which have been and continue to be used in laboratory and experimental challenge studies. Since the preservation of viable viruses in the environment for over 6 decades is highly unlikely, it is possible that the source of some of the recent virulent viruses isolated from poultry and wild birds might be laboratory viruses.

INTRODUCTION

Infections with virulent strains of Avian paramyxovirus 1 (APMV-1, synonymous with Newcastle disease virus [NDV]), a member of the family Paramyxoviridae, subfamily Paramyxovirinae, and genus Avulavirus (1), cause Newcastle disease (ND) in birds. NDV is a pathogen capable of producing a devastating disease in domestic fowl, with vast social and economic consequences (2). Chickens infected with NDV show a wide spectrum of clinical signs that vary with different virus strains (3). Hanson and Brandly (4) categorize ND viruses into three main pathological groups: lentogens are avirulent and cause mild enteric, respiratory, or subclinical disease (5); mesogens cause disease and death primarily for chickens younger than 8 weeks and produce mainly respiratory disease (6); velogens induce severe systemic infections with high mortality rates (7). According to the World Organisation for Animal Health (OIE) (8), virulent NDV strains, which include both mesogenic and velogenic strains, must meet one of the following criteria: (i) have an intracerebral pathogenicity index (ICPI) in day-old chicks (Gallus gallus) of 0.7 or greater; or (ii) have multiple basic amino acids at the C terminus of the F2 protein and phenylalanine at residue 117, which is the N terminus of the F1 protein. The term “multiple basic amino acids” refers to the presence of at least three arginine or lysine residues from positions 113 through 116. More information about NDV characteristics can be found in a comprehensive review that has been published recently (7).

NDV has a worldwide distribution (9) and is grouped into classes and genotypes based on either the nucleotide coding sequence of the full-length fusion gene or complete genome sequences (10). Currently, NDV isolates are grouped in two classes. Class I includes mainly avirulent isolates from wild waterfowl that occasionally spill over into poultry (11). However, the vast majority of NDV isolates belong to class II, which is further currently divided into 18 genotypes, some with subgenotypes (10, 12–14). Viruses from the “historical” genotypes II, III, and IV of class II were responsible for the first NDV panzootic during the 1940s and persisted through the 1960s (15). Virulent class II NDV isolates from other genotypes have been and continue to be isolated worldwide from ND outbreaks. Most of these isolates belong to genotypes V (North America and Africa), VI and VII (worldwide), XI (Madagascar), XII (Asia, South America), XIII (Asia), and XIV (Nigeria) (9, 16) and recently designated genotypes XVI (Dominican Republic) and XVII and XVIII (Africa) (12, 14, 17).

Class II genotype II comprises both lentogenic and velogenic strains (9, 15), first identified in North America. Genotype II virulent isolates Texas/GB/1948 and Beaudette C/1945 were collected in the 1940s in the United States. Some lentogenic strains of genotype II (e.g., LaSota/1946 and Hitchner/B1/47) were also collected in United States in the 1940s, and since their commercialization they have been used worldwide as live vaccines in poultry (9). Newcastle disease virus strain F48 (according to the number of passages in eggs often designated F48E8 and F48E9) was isolated in China in the 1940s and is designated a member of class II genotype IX (18). During the last 2 decades, isolations of genotype II virulent viruses in Egypt, China, and India have been reported in poultry and wild birds (19–21). Isolates of genotype II are reported occasionally, and surprisingly, it has been reported that they have lower substitution rates in their genomic sequences than do other genotypes (22). In the last decade, virulent viruses of genotype IX have also been isolated frequently from poultry and wild birds in China (23, 24).

Wild birds, particularly waterfowl, are considered a natural reservoir for NDV (9, 25, 26) and harbor primarily lentogenic strains (27). Evidence suggests the existence of epidemiological links between field isolates recovered from wild birds and those obtained from poultry (7, 28, 29). Additionally, there is evidence for spillover of both lentogenic and velogenic viruses from poultry into wild birds (11, 30). However, the possible release of highly virulent viruses into poultry or wild birds as a result of human activity has not been considered to be a risk factor for outbreaks. The aim of this study was to analyze the origin of recent virulent class II NDV field isolates belonging to genotypes II and IX and to determine their evolutionary history. As previously used by Ballagy-Pordány et al. (15), for the purposes of the present report, the term “historical isolates” refers to those viruses isolated prior to 1960.

MATERIALS AND METHODS

Origin of the sequenced “historical” viruses.

Ten historical NDV isolates were obtained from the repository of the U.S. Department of Agriculture (USDA) Southeast Poultry Research Laboratory (SEPRL) and used in this study. These samples were isolated during poultry outbreaks between 1944 and 1955 in the United States and Mexico (see Table S1 in the supplemental material). These historical viruses underwent between 3 and 10 passages in embryonated chicken eggs during 1955 to 1957 and since that time have been kept frozen at −80°C (with the exception of chicken/USA/California/CG179/1946, which was passed 13 times in 1965). The passage histories of chicken/USA/11914-FD/California/1944 and chicken/USA/California/119144/NAP/1944 were not available.

Virus propagation.

The 10 historical viruses obtained from the SEPRL repository were propagated in 9- to 11-day-old specific-pathogen-free (SPF) embryonated chicken eggs from the SEPRL SPF White Leghorn flock (31). To ensure advanced biosecurity, all the tests were carried out in the SEPRL biosafety level 3A (BSL-3A) facility.

RNA isolation, PCR amplification, and sequencing.

RNAs from each isolate were extracted from allantoic fluids using TRIzol LS (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. One-step reverse transcriptase (RT)-PCR (SuperScript III One-Step RT-PCR System with Platinum Taq DNA Polymerase; Life technologies, Carlsbad, CA, USA) was used to convert and amplify the extracted RNA samples. The primers used for the PCR and sequencing (4331F/5090R, MSF1/NDVR2, 4911F/5857R, 4961F/5772R, 5435F/6320R, 5669F/6433R) were described previously (32). The PCR products were subjected to electrophoresis in 1% agarose gels (0.5× Tris-borate-EDTA [TBE]). The DNA bands were excised and purified using the QuickClean II gel extraction kit (GenScript, Piscataway, NJ, USA). Nucleotide sequencing and assembly were performed as described previously by Miller et al. (33). To prevent cross-contamination of virus strains and RT-PCRs, established good laboratory practices for use of virulent viruses, proper record keeping, monitoring methods for storage, and use of reaction controls were followed.

Collection of sequences.

Full-length fusion gene (F-gene) and complete genome sequences of class II NDV isolates were downloaded from GenBank (available as of March 2015; http://www.ncbi.nlm.nih.gov/GenBank/index.html) and aligned using ClustalW (34). This resulted in two final data sets: 1,450 full-length F-gene sequences and 316 complete genome sequences. For the complete genome sequences, intergenic regions were removed, and the final alignment consisted of the concatenated coding region sequences of the six NDV genes.

Recombination analysis.

Recombination analysis was performed with all sequences (including the 10 virulent NDV sequences obtained in this study) using the RDP3 program (35) (data not shown). To identify putative recombinant sequences, four statistical methods were utilized (RDP, Geneconv, Maxchi, and Chimera). Sequences with recombination events identified by at least two detection methods (P < 0.001) were considered true recombinants and removed from the analyses.

Evolutionary analysis.

Preliminary phylogenetic analyses were performed utilizing the full-length F-gene and the complete genome sequences after the recombinant sequences were removed (data not shown). Historical virulent NDVs from genotypes II and IX that were closely related to the reported field viral isolates were selected, and two different data sets were generated for further analyses: (i) full-length F-gene sequences of recent (n = 37) and historical (n = 15) virulent NDV from genotypes II and IX (including the sequences of the 10 isolates obtained in this study), NDV of low virulence from genotype II (n = 12), and selected isolates from other genotypes (n = 17) (see Table S1 in the supplemental material); (ii) complete genome sequences of recent and historical NDV from genotypes II and IX and selected isolates from other genotypes (n = 72) (see Table S2 in the supplemental material).

The estimates of average evolutionary distances were inferred using MEGA6 (36). Analyses were conducted using the maximum composite likelihood model using 162 full-length fusion gene sequences (136 of genotype II [see Table S4 in the supplemental material] and 26 of genotype IX [see Table S1 in the supplemental material]) (37). In the analyses conducted by MEGA6, the percentages of trees in which the associated taxa clustered together are shown below the branches. The trees are drawn to scale, with branch lengths measured in the number of substitutions per site. For all methods, the codon positions included were 1st + 2nd + 3rd + noncoding. All positions containing gaps and missing data were eliminated. Previously described criteria (10) based on the phylogenetic topology of maximum likelihood trees and on the evolutionary distances based on maximum composite likelihood were employed to determine the classification of viruses into taxonomic groups.

To illustrate the dynamics of relative genetic changes over time, Bayesian analyses of the F-gene sequences of virulent NDV isolates from genotype II (n = 26) and genotype IX (n = 26) (see Table S1 in the supplemental material) were performed using BEAST version 1.8.1 (38). Substitution rates were estimated employing the generalized time reversible (GTR) substitution model for a relaxed lognormal molecular clock. All other conditions used were as previously reported by Miller et al. (39), except for the length of the Markov chain Monte Carlo analyses, which were run for 40 million iterations to ensure that the effective sample size was >200 for all estimated parameters in model runs.

Complete genome alignment between genotype II recently identified virulent isolates (n = 9, isolated after 1998), isolates of low virulence (n = 9), and Texas/GB/48 was performed using multiple alignment with fast Fourier transformation (MAFFT) implemented in Geneious v8.1.2 (40). To trace the origin of the recently identified virulent isolates, vaccine-like strains and historical virulent isolates were included in the genome alignment. The complete genome sequence of the vaccine strain LaSota was set as the reference sequence. For the analyses performed by MAFFT (complete genome alignment), a six-letter/digit code was used, and corresponding isolates are presented in Table S3 in the supplemental material. The first two letters represent the country of isolation and the first two Arabic numerals represent the year of isolation. The median joining (MJ) phylogenetic network of the full-length fusion gene sequences was constructed using 136 sequences of NDV genotype II viruses (including all available viruses of low virulence) (see Table S4 in the supplemental material). F-gene segments were aligned and used to construct a phylogenetic network using the median joining method implemented in Network ver. 4.613 (41).

Nucleotide sequence accession numbers.

The full-length F-gene sequences (n = 10) of virulent NDV obtained in this study were submitted to GenBank and are available under the accession numbers KP939087 to KP939096.

RESULTS

Sequencing of the historical isolates.

Analysis of the sequencing data from the fusion gene showed that the 10 viruses isolated from U.S. and Mexican poultry between 1944 and 1955 grouped together with other previously reported historical isolates (Fig. 1). The sequences of these 10 isolates encoded fusion protein cleavage sites with multiple basic amino acids between residue positions 113 and 116 and a phenylalanine at position 117 (₁₁₃RQKR↓F₁₁₇). Such a motif is typical of virulent NDV isolates (8).

FIG 1.

Phylogenetic analysis based on the full-length nucleotide sequence of the fusion protein gene of isolates representing Newcastle disease virus class II. The evolutionary history was inferred by using the maximum likelihood method based on the Tamura-3 parameter model with 500 bootstrap replicates (57). The tree with the highest log likelihood (−11871.6115) is shown. A discrete gamma distribution was used to model evolutionary rate differences among sites (4 categories [+G, parameter = 0.6992]). The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 36.3891% sites). The analysis involved 81 nucleotide sequences with a total of 1,653 positions in the final data set. Historical virulent strains of genotype II sequenced in this study are designated with a circle (•; clustering in a separate branch) or a triangle (▲) in front of the taxon name. The standard strains are designated with a square (■) and bold font. The Roman numerals presented in the taxon names in the phylogenetic trees represent the respective genotype for each isolate, followed by the GenBank identification number, host name (if available), country of isolation, strain designation, and country of isolation. The low-virulence LaSota and B1 vaccine strains are presented in bold italic font.

Phylogenetic analysis.

Based on their sequences, the 10 historical virulent NDV isolates sequenced here were classified as members of genotype II of class II. Six of them (USA/NJ/KD/1945, USA/Montana/1946, USA/Ky/50/1947, USA/Ohio Miller/1948, USA/Ontario/Berwick/1948, and Mexico/Texcoco/1950) clustered together with other virulent isolates from the 1940s (Texas/GB/48, Roakin/48, and Beaudette C/45). Interestingly, the other four isolates (chicken/USA/11914-FD/California/1944, chicken/USA/California/119144/NAP/1944, chicken/USA/California/CG179/1946, and USA/Nebraska/WW1/1955) formed a separate monophyletic branch within genotype II (Fig. 1).

Recent and historical virulent NDV isolates within genotypes II and IX were found to be phylogenetically very closely related. Within genotype II, recent isolates from poultry and geese from China, India, and Egypt obtained between 1997 and 2008 clustered together with virulent NDV isolated in the United States in the 1940s (Fig. 1). Within genotype IX, viruses from chickens and wild birds isolated between 1997 and 2000 grouped closely together in the phylogenetic tree with the historical F48/48 strain isolated in the 1940s.

The mean evolutionary distances for the full-length fusion nucleotide sequences between recent and historical virulent NDV isolates were determined using MEGA as indicated above, and the results are presented in Table 1 (subsets of viruses were conditionally divided into groups of historical and recent isolates). Table 1 demonstrates that the distance between historical and recent virulent isolates within genotype II is 1.2% (Table 1, bold), while the distances between the recent virulent isolates compared to the NDV viruses of low virulence, including vaccine strains, of genotype II are much higher (2.7 to 2.9%) (Table 1, underlined values). Additionally, the evolutionary diversity between historical and recent isolates within genotype IX (0.3%) (bold) is even lower than the one observed within genotype II. Some of the recent isolates (e.g., chicken/Egypt/2/2006, chicken/China/AQI-ND026/2005, quail/India/NDV2K17/Chennai/1998, goose/China/SD/6/2004, goose/China/JS/1/2005, chicken/China/NDV03/2008) were almost identical (0.1 to 0.3% nucleotide identity) to historical strains chicken/USA/TX/GB/1948 and chicken/USA/New Jersey-Roakin/1946, and more than 20 isolates from China showed almost no genetic difference from chicken/China/F48E9/1946-1948 for periods of 50 to 60 and more years (data not shown). To further support the observation that historical viruses were evolving already 60 years ago, it is interesting that four other historical virulent isolates of genotype II (chicken/USA/11914-FD/California/1944, chicken/USA/California/119144/NAP/1944, chicken/USA/California/CG179/1946, and USA/Nebraska/WW1/1955) formed a separate branch in the phylogenetic tree (Fig. 1) and displayed a considerably greater genetic distance (3.5 to 4.3%) (Table 1, bold and italic) than did the rest of the isolates within the genotype. The results from the analyses of complete genome sequences were consistent with those obtained from the complete fusion gene analyses, and the close phylogenetic relationships between the historical and recent virulent NDV isolates within genotypes II and IX were confirmed (see Fig. S1 in the supplemental material). The mean interpopulational genetic distance inferred from the complete genome nucleotide sequences was very close to the one obtained from the full-length fusion gene analysis. The evolutionary distances between the historical and recent virulent isolates within genotypes II and IX inferred from the complete genome coding sequences were 1.3% and 0.2%, respectively.

TABLE 1.

Estimates of evolutionary distances between historical and recent isolates within genotypes II and IX of class II Newcastle disease virus^a

Genotype/group	No. of base substitutions per site for genotype/group:
	II/historical virulent (n = 9)	II/recent virulent (n = 13)	II/historical low virulence (n = 2)	II/recent low virulence (n = 108)	II/historical virulent sb (n = 4)	IX/historical (n = 2)	IX/recent (n = 24)
II/historical virulent		(0.001)	(0.003)	(0.003)	(0.004)	(0.010)	(0.009)
II/recent virulent	0.012		(0.003)	(0.003)	(0.004)	(0.010)	(0.010)
II/historical low virulence	0.029	0.027		(0.001)	(0.005)	(0.010)	(0.010)
II/recent low virulence	0.031	0.029	0.006		(0.005)	(0.010)	(0.010)
II/historical virulent sb	0.036	0.035	0.040	0.043		(0.010)	(0.010)
IX/historical	0.117	0.117	0.121	0.122	0.114		(0.001)
IX/recent	0.116	0.117	0.120	0.121	0.113	0.003

^aThe distances (presented as no. of substitutions per site) were inferred from the full-length nucleotide fusion protein gene sequences. Standard error estimate(s) are shown above the diagonal in parentheses and were obtained by a bootstrap procedure (500 replicates). The rate variation among sites was modeled with a gamma distribution (shape parameter = 1). The analysis involved 162 nucleotide sequences. All positions containing gaps and missing data were eliminated. There were a total of 1,653 positions in the final data set. Evolutionary analyses were conducted in MEGA6 (36). The term “historical” refers to strains isolated prior to 1960; sb refers to separate branch of historical virulent strains in genotype II (Fig. 1). Values indicating low genetic distance between recent and historical virulent viruses (1.2% and 0.3%) are in boldface, and those indicating greater genetic distance (2.7% and 2.9 %) are underlined. Values higher than the criterion set by Diel et al. (10) for identifying subgenotypes are in bold and italic.

Complete genome alignment using concatenated complete coding regions of genotype II virulent viruses demonstrates that the sequence similarities observed between recent and historical isolates, based on the fusion gene coding region, also extend across the entire genome (Fig. 2). The distribution of single nucleotide polymorphisms (SNPs) among viruses of different virulence are graphically represented in this figure. Currently commercially available NDV vaccine strains and lentogenic isolates of genotype II (above rectangles) share among them the highest genetic identity. Similarly, the figure shows that recent and old virulent viruses are highly related to each other (enclosed in rectangles) and not to vaccine-like viruses.

FIG 2.

Nucleotide sequence alignment of complete coding sequences of Newcastle disease virus genotype II isolates. Alignment of the complete coding region sequences of the genotype II isolates was performed using MAFFT. The complete genome sequence of LaSota/1946 lentogenic vaccine strain was placed on the top of the alignment as a reference. Vertical lines indicate the single nucleotide polymorphisms (SNPs) compared to the reference sequence. The abbreviations on the left of the figure are described in Table S3 in the supplemental material. The sequences enclosed in the rectangles correspond to the virulent strains used in the analysis, and the sequences above the rectangles represent the strains of low virulence. US, United States; CN, China; EG, Egypt; IN, India.

Similar types of phylogenetic relationships were obtained utilizing a different type of analysis (maximum parsimony). Full-length fusion sequences of 136 isolates from genotype II were used to construct an MJ phylogenetic network, which formed two distinct clusters (see Fig. S2 in the supplemental material). In Fig. S2A in the supplemental material, viruses of low virulence are clearly separated from historical and recent virulent viruses. The diagram demonstrates that there is no separation into clades among the recently isolated virulent NDV and the historical virulent NDV isolated prior to 1960. The central cluster within the virulent branch of the network indicates that there are very small genetic distances among historical and recent isolated viruses. Notably, the largest genetic distance occurs among a group of 4 historical isolates in the upper right corner of the figure (see Fig. S2A in the supplemental material) (chicken/USA/11914-FD/California/1944, chicken/USA/California/119144/NAP/1944, chicken/USA/California/CG179/1946, and USA/Nebraska/WW1/1955) and the rest of the virulent samples. Based on the criteria described by Diel et al. (10) for determination of genotypes and subgenotypes, these four isolates probably represent a subgenotype within genotype II that disappeared after the 1940s. Figure S2B in the supplemental material is a similar representation of the virulent viruses by country of origin and demonstrates that virulent viruses nearly identical to historical samples are identified worldwide.

To illustrate the dynamics of substitution rates over time, Bayesian analyses of the F-gene sequences of recent and historical virulent NDV isolates were performed. Utilizing the sequences of 26 virulent isolates from genotype II (see Table S1 in the supplemental material), the mean substitution rate for the fusion gene was estimated to be 7.05 × 10⁻⁵ (standard error [SE], 3.48 × 10⁻⁶). Similar estimates were determined for the full-length fusion sequences of 26 genotype IX virulent isolates using a relaxed lognormal molecular clock (2.05 × 10⁻⁵, SE = 6.41 × 10⁻⁷) (Table 2).

TABLE 2.

Mean substitution rates and summary statistics of full-length fusion protein gene sequences of class II genotypes II and IX Newcastle disease virus virulent strains estimated using a relaxed lognormal molecular clock

Summary statistics	Estimated fusion gene substitution rate for virulent strains of genotype:
	II	IX
Mean	7.05 × 10−5	2.05 × 10−5
Standard error	3.48 × 10−6	6.41 × 10−7
Median	5.37 × 10−5	1.51 × 10−5
Effective sample size	361	992

DISCUSSION

We have investigated the relationship between historical virulent NDV isolates of genotypes II and IX within class II and viruses of the same genotypes causing recent outbreaks in poultry and wild birds worldwide. The data presented here establish the existence of a close phylogenetic relationship between recent and historical virulent isolates circulating before 1960. RNA viruses, specifically NDV, are known to evolve, and since the genomes of these recently isolated virulent strains did not change over time, our data suggest that they may not have been circulating over the past 60 years. The results from the phylogenetic trees and the nucleotide distance analysis (Fig. 1 and Table 1; see also Fig. S1 in the supplemental material) demonstrated a very close genetic relationship between the historical and recent virulent isolates of genotypes II and IX. Some of the recent isolates were almost identical to the historical viruses. Virulent isolates from genotype II considered “historical” or “early” were involved in the first panzootic of Newcastle disease (7, 10, 15). Virulent isolates belonging to genotype II have been thought to occur mainly in the United States before the 1960s. However, virulent viruses from genotypes II and IX have been occasionally isolated over the past 2 decades at distant geographic locations (19, 20, 23, 42, 43). In previous reports, Czeglédi et al. (44) and Herczeg et al. (45) indicated that the expected nucleotide sequence change of NDV caused by natural evolution is estimated to be approximately 1% per decade. The isolation dates of viruses used in this study span over 5 decades; however, the genetic distance between recent and historical viruses at both fusion gene and the complete genome appear to be unexpectedly and unusually small. This small nucleotide distance can be explained only by an epidemiological link between viruses and/or common ancestry (44).

Those recent virulent viruses of genotypes II and IX that have been isolated from wild birds and poultry lack recent ancestors in existing genetic databases, and there is no evidence of progressive genomic changes over time as observed in other viruses in the database that contains thousands of samples since the 1960s. Thus, the data suggest that these recent isolations might represent the reintroduction of already “lost” (not circulating naturally) viral genomes into the environment. The possible source of those field isolates appears to be the virulent strains Texas/GB and F48, both of which have been used as challenge viruses to show the efficacy of ND commercial vaccines before their commercial production and in experimental studies (9, 18, 46, 47).

The greater genetic distance between the recent virulent NDV isolates and the vaccine viruses or other recent NDV isolates of low virulence suggests that the recent virulent isolates did not revert from the vaccines or from other low virulence viruses. Although there are only two documented cases, other class II NDV isolates of low virulence have been shown to be capable of evolving naturally into a virulent phenotype (11), as only a few point mutations resulted in the emergence of a virulent form of NDV (48). As vaccine strains of genotype II have been found to spill over from poultry to wild birds (49), the comparison is necessary. However, the grouping of the viruses in the median joining phylogenetic network analysis of genotype II viruses (see Fig. S2 in the supplemental material) and the results from the complete genome coding region sequence alignments (Fig. 2) clearly exclude the possibility that the recent virulent isolates represent mutant forms that originated from viruses of low virulence or vaccines. The sporadic character of the outbreaks and the presence of these “challenge-like” isolates in highly distant locations support a possible limited circulation in poultry or in wild birds after the standard strains have been released into the environment. Since there is some genetic distance among the recent isolates and also between them and the parental strains, it is possible to assume that some minimal level of circulation among poultry and perhaps wild birds may have occurred.

Our data suggest that when NDV isolates almost identical to historical isolates are obtained, their source should be carefully evaluated, and if there is reasonable doubt as to their origin, these viruses may need to be excluded from evolutionary studies, or else the data will be skewed. Their use might lead to incorrect results and conclusions. The substitution rate for genotype II estimated in the present study is consistent with a previously reported data by Chong et al. (22); however, the authors did not consider the possibility that some of the viruses used in their analysis would have been escapes; thus, results are in disagreement with the rates of evolution estimated previously for class II NDV isolates. Previous studies using large data sets for the full fusion gene coding sequence and the complete NDV genomes have shown that the annual rate of change for Newcastle disease virus is approximately 10⁻³. Miller et al. (39) estimated a substitutional rate of 1.32 × 10⁻³ and 1.7 × 10⁻³ for class II virulent viruses based solely on the fusion gene using strict and lognormal relaxed molecular clocks, respectively. In a more recent study, Ramey et al. (50) reported similar rates for isolates from NDV class II. Our observed rates of changes for the full fusion coding region of the virulent viruses of genotype II and IX were considerably lower (>100-fold lower) (Table 2). Furthermore, such very low substitutional rates have not been observed among other RNA viruses and do not likely reflect the natural evolution of NDV. As the use of evolutionary studies is constantly increasing, the creation of proper data sets with detailed information and reliable sequences is essential for obtaining objective results.

The possibility that the recent virulent viruses may have survived dormant in the environment for such a long period is highly unlikely. NDV is an enveloped RNA virus that is unstable in the environment, and solar radiation destroys the virus within 1 h (51). In litter, soil, and carcasses, NDV remains infectious for days, and in feathers it remains so for several months (52), but these periods are significantly shorter than the decades between the collection times of recent and historical virulent isolates studied here. These facts preclude the accumulation of infectivity in natural habitats of birds (53). There is not yet evidence of the existence of a reservoir of these viruses that could maintain RNA viruses unchanged, especially in countries without environmental conditions favorable to preservation, such as extensive ice cover.

Other possibilities could explain the identification of recent virulent viruses that are genetically closely related to historical virulent viruses. If historical and field strains have been simultaneously used in laboratories, cross-contamination may have occurred. However, the likelihood of this possibility is low, as the sequences of the recent virulent strains studied here originate from different countries and laboratories, and it is unlikely that in all of them crossover contamination yielding similar sequencing results would have occurred. Inactivated NDV vaccines made with virulent seed strains are currently marketed in Asia. For other viral systems (e.g., Venezuelan equine encephalitis virus, foot-and-mouth disease virus, influenza A virus), it has been suggested that improperly inactivated vaccines, vaccine trials with challenge studies, and laboratory accidents were the source of isolation of almost identical viruses in distant years (54–56). None of those possibilities can be discarded based on present evidence; however, all of those point to human activities instead of natural evolution.

Taken together, our results suggest that the recent isolations of virulent ND viruses of genotypes II and IX, which are unnaturally closely related to NDV strains isolated prior to 1960, do not represent a natural circulation of NDV in poultry and wild bird populations. Although infrequently isolated, such viruses pose a significant economic threat to the poultry industries.