Genetic and antigenic changes in porcine rubulavirus

Abstract

Blue eye disease, caused by a porcine rubulavirus (PoRV), is an emergent viral swine disease that has been endemic in Mexico since 1980. Atypical outbreaks were detected in 1990 and 2003. Growing and adult pigs presented neurological signs, mild neurological signs were observed in piglets, and severe reproductive problems were observed in adults. Amino acid sequence comparisons and phylogenetic analysis of the hemagglutinin-neuraminidase (HN) protein revealed genetically different lineages. We used cross-neutralization assays, with homologous and heterologous antisera, to determine the antigenic relatedness values for the PoRV isolates. We found antigenic changes among several strains and identified a highly divergent one, making up a new serogroup. It seems that genetically and antigenically different PoRV strains are circulating simultaneously in the swine population in the geographical region studied. The cross neutralization studies suggest that the HN is not the only antigenic determinant participating in the antigenic changes among the different PoRV strains.

Introduction

Blue eye disease (BED) is a viral swine infection that has been endemic in Mexico since 1980. Blue eye disease is initially characterized by a neurological and respiratory syndrome in suckling pigs and corneal opacity in 1% to 10% of cases (1). Biological and genetic similarities with mumps virus; Newcastle disease virus; and parainfluenza viruses 2, 4, and 5 (2) led to the classification of the etiologic agent of BED as a porcine virus of the Rubulavirus genus (PoRV) in the Paramyxoviridae family (3).

In the earliest outbreaks, BED reached 20% morbidity and 90% mortality in newborn pigs. Unusual clinical manifestations of BED, including neurological signs in 3- to 5-month-old pigs, were observed in 1988, associated with increased mortality in pigs > 1 mo old (1). Other PoRV outbreaks have resulted in high rates of infertility in sows and boars (4). In the past decade, severe neurological symptoms have been observed in adult pigs (5). These data indicate that PoRVs are undergoing changes in their clinical manifestations and virulence.

The PoRV genome is a single-stranded negative-sense RNA encoding 6 proteins: the nucleoprotein (NP), phosphoprotein (P), matrix (M), fusion (F), hemagglutinin-neuraminidase (HN), and large (L) proteins (6). The HN protein determines the tropism for the alpha 2,3-linked sialic acid present on cell receptors, which are highly expressed in susceptible porcine tissues (7). In addition, the immune response of PoRV infected pigs is predominantly directed toward the HN glycoprotein (8). These properties indicate that HN is a potential vaccine target.

Although several experimental and commercial vaccines against PoRV have been tested (9), the disease is far from being controlled. This may be the result of the existence of antigenic differences between several PoRV isolates (10). In this study, we did cross-neutralization assays in order to assess the antigenic differences present among PoRV isolates collected in Mexico from 1990 to 2003. Furthermore, to understand the patterns of genetic changes in PoRV, we carried out HN amino acid sequence comparisons and phylogenetic analysis of the available strains.

Materials and methods

Viruses and HN gene sequences

The viral strains used in this work (listed in Table I) have been previously described by our group (11). The viral strains are designated with its name/year of isolation: PAC2/1990, PAC3/1992, PAC4/1993, PAC6/2001, PAC7/2002, PAC8/2002, PAC9/2003, CI/1991, CII/1991, CIII/1999, and CIV/1999.

Table I.

Amino acid substitutions present in the hemagglutinin-neuraminidase (HN) antigen of the porcine rubulavirus (PoRV) isolates in comparison to the LPM/1984 strain. The amino acid changes are indicated

Position	18	20	37	68	121	122	156	193	223	252	264	291	298	347	370	382	407	432	436	447	450	456	462	467	469	475	484	487	500	511	512	514	526
LPM/1984	A	R	V	I	I	P	F	A	A	V	N	A	V	Q	S	Q	L	F	P	L	L	G	S	T	V	D	S	G	T	A	I	E	T
PAC4/1993							L
PAC2/1990							L		T			D							L				N				I		I		T
PAC3/1992							L		T			D							L				N				I		I		T
CI/1991						R	L												L				N				I		I
CII/1991							L												L	M			I	P			I		I				C
CIII/1999				M			L								P				L		R		N				I		I
CIV/1999	V				V		L							E			I	L	L				N		T	I	I	S	I
PAC6/2001		K	I				L	T		I	T		I						L			E							I	S		K
PAC7/2002		K	I				L	T		I	T		I						L			E							I	S		K
PAC8/2002		K	I				L	T		I	T		I						L										I	S		K
PAC9/2003			I				L	T		I	T		I			M			L										I	S		K

Phylogenetic analysis

A phylogenetic analysis was done using computer software (MEGA software, version 4) of the Center for Evolutionary Functional Genomics, The Biodesign Insitute, Arizona State University, Arizona, USA (12).

Antigenic relatedness

The antigenic relatedness values (ARV) among the PAC PoRV strains were assessed according to the method of Archetti and Horsfall (13).

Specific antibodies

Two rabbits per virus isolate were immunized intramuscularly with ultraviolet (UV) light inactivated virus using 10⁴ TCID₅₀ infectious doses. Rabbits were boosted on the 7th, 11th, and 14th d post-immunization. The rabbits were bled 1 wk after the last inoculation and the serum was inactivated at 56°C for 30 min. The sera for each virus were pooled and used in the virus neutralization tests.

Virus neutralization

Virus neutralization (VN) tests were done in Vero cells using the beta method (constant-virus, diluted-serum) (14). The antiserum was normalized to the same neutralization activity, and the amount of antiserum needed to neutralize 100 TCID₅₀ of homologous virus (1 antibody unit) was determined. Starting from 20 antibody units, 2-fold serial dilutions of the antiserum were mixed with a constant amount of virus (100 TCID₅₀%) for 30 min at room temperature. Each serum — virus combination was tested by duplicate, using two 96-well plates. The serum was double serial diluted along the 12 columns in such a way that each serum dilution had 6 replicates per plate (A to F lines). The G and H lines were used as negative (no virus) and positive controls (no serum). Three days post-infection, the integrity of the cells was examined by microscopy to evaluate cytopathic effects. The presence of PoRV in the supernatants of each well was confirmed by hemagglutination activity and amplification using reverse transcription-polymerase chain reaction (RT–PCR), as have been reported previously (11). The VN end point for each virus was determined and the neutralization titer for each antiserum with homologous and heterologous viruses was calculated (15).

Results

Hemagglutinin-neuraminidase genetic differences

Table I shows the predicted amino acid (aa) differences in the HN genes of different PoRV isolates in comparison to the HN gene sequence of the LPMV/1984 virus reported by Sundqvist (16). The aa comparison among PoRV isolates and the phylogenetic analysis revealed 5 major divergent genetic lines (Table I, Figure 1). One genetic line is made up of PAC4/1993, which is very similar to the LPMV/1984 isolate, with only 1 aa different. A 2nd distinct genetic group of sequences consisted of PAC6/2001 and PAC7/2002 (with the same aa sequence) and PAC8/2002 and PAC9/2003 (differing by 1 or 2 aa). A 3rd group of sequences included PAC2/1990 and PAC3/1992, which share the same aa sequence. Other viruses isolated in 1991 (CI/1991 and CII/1991) and 1999 (CIII/1999) belong to a 4th genetic group. The isolate CIV/1999 represent a 5th genetic line.

Figure 1.

Genetic relationships among the porcine rubulavirus (PoRV) strains based on the hemagglutinin-neuraminidase (HN) protein. Dendrogram showing the genetic distances and the phylogenetic relationships between the different PoRV isolates studied. The numbers in the nodes represent the percentage of times the branch was repeated in the bootstrap analysis after 1000 replicas. The bars represent the genetic distance in arbitrary units.

Antigenic variation

The ARV values determinate among the PAC strains are presented in Table II. Two virus strains are considered to be antigenically related if they have an ARV of 0.5 to 1.0 (17). According to this criterion, all PoRV strains were antigenically related except for the PAC3/1992 virus, which was not antigenically related (ARV = 0.35) to the PAC8/2002 of PAC9/2003 strains. All viruses isolated in the 2000s (PAC6/2001 to PAC9/2003) were closely related, with ARVs of 1.0, excluding PAC8/2002 and PAC9/2003 which had an ARV of 0.7. The viruses isolated in the 1990s (PAC2/1990, PAC3/1992, and PAC4/1993) had reduced ARVs (0.7), and the ARVs were even lower when these strains were compared to strains from the 2000s (ARVs of 0.5 and 0.3).

Table II.

Antigenic relatedness values among porcine rubulavirus (PoRV) strains

Virus strain	PAC2/1990	PAC3/1992	PAC4/1993	PAC6/2001	PAC7/2002	PAC8/2002	PAC9/2003
PAC2/1990	1	0.7	0.7	0.5	0.5	0.5	0.5
PAC3/1992		1	0.7	0.5	0.5	0.3	0.3
PAC4/1993			1	0.7	0.7	0.5	0.5
PAC6/2001				1	1	1	1
PAC7/2002					1	1	1
PAC8/2002						1	0.7
PAC9/2003							1

Discussion

Our findings suggest that PoRV strains are able to preserve the HN aa sequence for long periods of time. In fact, the HN sequence of the PAC4/1993 virus differed from the LPMV/1984 sequence at only 1 aa residue even though these strains were isolated almost 10 years apart.

Conversely, some virus strains were genetically and antigenically different from other contemporary isolates. The PAC2/1990 and PAC3/1992 viruses differed from the PAC4/1993 strain at 7 aa, with an ARV of 0.7. The existence of contemporary different genetic strains is also evident from the CI/1991 and CII/1991 strains, which differed at 4 aa, and from the CIII/1999 and CIV/1999 strains, which differed at 11 aa.

Furthermore, when we compared the strains isolated from 1990 to 1993 with those isolated from 1999 to 2003, we found that there were 6 to 11 aa changes. This result shows that amino acids in the HN of PoRV are changing over time and are diverging to form different genetic lines (Figure 1).

The evaluation of ARV and aa changes in the HA antigen among PoRV strains suggests that antigenic differences are somehow associated to the HN aa sequence divergence. In that sense PAC4/1993 in relation to PAC8/2002 and PAC 9/2003, which differ in 10 aa, have an ARV of 0.5; and the ARV among PAC3/1992 and PAC9/2003 is 0.3 with 12 aa changes.

However, it seems that HN is not the only antigenic determinant participating in the ARV among different PoRV strains. This is mainly suggested by the fact that among the PAC2/1990 and PAC3/1992 strains (with the same HN aa sequence) have an ARV of 0.7. Similarly, PAC2/1990 rendered an ARV of 0.5 to PAC8/2002 or PAC9/2003; but the ARV of these last strains to PAC3/1992 (with the same HN aa sequence of PAC2/1990) is 0.3. In the same way, PAC8/2002 in relation to PAC9/2003 has only 2 aa different, and their ARV is 0.7.

The ARVs below 0.5 indicate that there is no antigenic relation between strains. Based on this, we concluded that the PAC3/1992 strain is not antigenically related to the PAC8/2002 and PAC9/2003 strains; thus, they represent different serotypes. Unfortunately, there is very limited information about how many genetically and antigenically different strains are circulating at different times and in different geographic regions. It is also not known how long genetic variations are conserved and how often new ones arise. All of these questions are important to address in order to develop efficient diagnosis and prevention measures to control and eradicate BED.

This work shows that some variants of PoRV are able to preserve the HN sequence for long periods of time. Additionally, we found that other strains are changing, generating new variants that diverge genetically and antigenically. It is evident that genetically and antigenically different PoRV strains are circulating in the swine population imposing challenges to diagnostic and vaccination efforts. The cross neutralization studies shows that the HN is not the only antigenic determinant participating in the antigenic changes among the different PoRV strains.