key: cord-269340-o9jdt86j
authors: Callison, Scott Andrew; Jackwood, Mark W.; Hilt, Deborah Ann
title: Infectious Bronchitis Virus S2 Gene Sequence Variability May Affect S1 Subunit Specific Antibody Binding
date: 1999
journal: Virus Genes
DOI: 10.1023/a:1008179208217
sha: 
doc_id: 269340
cord_uid: o9jdt86j

The S2 gene of several strains of infectious bronchitis virus (IBV) belonging to the Arkansas, Connecticut, and Florida serotypes was sequenced. Phylogenetic analysis of the S2 gene nucleotide and deduced amino acid sequence data resulted in groups of strains that were the same as groupings observed when S1 sequence data was used. Thus, it appears that S2 subunits are conserved within a serotype but not between serotypes. Although the sequence differences were small, we found that only a few amino acid differences were responsible for different secondary structure predictions for the S2 subunit. It is likely that these changes create different interactions between the S1 and S2 subunits, which could affect the conformation of the S1 subunit where serotype specific epitopes are located. Based on this sequence data, we hypothesize that the S2 subunit can affect specific antibody binding to the S1 subunit of the IBV spike glycoprotein.

Infectious bronchitis (IB) is an acute, highly transmissible, upper respiratory tract disease in chickens. Clinical signs include tracheal rales, nasal exudate, coughing, and sneezing. Infectious bronchitis affects both sexes and the disease may spread to the reproductive and renal systems (1) . It is of economic importance because it can cause poor weight gain and reduced feed ef®ciency in broilers and a decline in egg production and egg quality in layers (2) .

Infectious bronchitis virus (IBV), the causal agent of IB, is a member of the Coronaviridae family. The virion is pleomorphic (diameter 90±200 nm) and enveloped with club-shaped surface projections (spikes) on the surface of the virion. It contains a single stranded, positive-sense RNA genome approximately 27.5 kb in length (3) . The virion contains four major structural proteins: a nucleocapsid (N) protein associated with the viral RNA, the integral membrane (M) glycoprotein, a small membrane (sM) protein, and the spike (S) glycoprotein. The S glycoprotein is a polypeptide of approximately 1200 amino acids. It is proteolytically cleaved after translation into two subunits, S1 and S2 (4) . Both subunits are glycosylated with high mannose, N-linked oligosaccharides (5) .

The virion spike is thought to be an oligomeric protein composed of two polypeptides each of the S1 and S2 subunits. The two subunits associate by noncovalent forces and retain their three-dimensional shape by way of intrapeptide, but not interpeptide, disul®de bridges (5) . The S2 subunits, which form the stalk portion of the spike, anchor it in the membrane, whereas the S1 subunits form the globular head of the spike glycoprotein (5) . The S1 subunit encodes amino acids involved in the induction of neutralizing, serotype speci®c, and hemagglutination inhibiting antibodies (6, 7) .

Although the S1 subunit of IBV has been examined extensively, the S2 subunit remains enigmatic. Based on the highly conservative nature of the S2 subunit among different members of the Coronavirus genus *Corresponding author. E-mail: mjackwoo@arches.uga.edu and different strains of IBV, it would appear that it plays little or no role in the induction of a host immune response (8) . However, it has been shown for IBV that an immunodominant region localized in the N-terminal half of the S2 subunit can induce neutralizing, but not serotype speci®c antibodies (9) . A DNA-binding protein region or leucine zipper motif has also been identi®ed in the S2 subunit of other coronaviruses (10) . Leucine zipper motifs are thought to be involved in transcriptional activation.

Furthermore, site-directed mutagenesis of the S2 subunit of another Coronavirus, mouse hepatitis virus (MHV), inhibited the binding of a virus neutralizing monoclonal antibody to the unchanged S1 subunit (11) . Last, a monoclonal antibody neutralization resistant mutant was reported to have an S1 gene sequence identical to the parental virus, suggesting that the mutant escapes neutralization due to changes in the S2 gene sequence (12) .

Thus, we are interested in examining the S2 gene and its deduced amino acid sequence of IBV strains in an attempt to determine if it plays a role in the binding of S1 subunit speci®c antibodies to the virus. We selected four strains belonging to the Arkansas serotype, Ark 99, Ark DPI, 3668-4, and GAV 92 because their S1 deduced amino acid sequences were very similar, 4 90%. Strains 3668-4 and GAV 92 were determined to be Ark-``like'' strains by restriction fragment length polymorphism (RFLP) analysis and later con®rmed by serology studies. (13) . We also selected Connecticut 46 and Florida 18288 for S2 gene sequencing because these strains are known to share 96.6% deduced amino acid identity for their S1 subunits, yet remain serologically distinct (14, 15) .

Infectious bronchitis virus strains used in this study are listed in Table 1 . Viruses were inoculated into embryonating eggs for propagation (16) . The allantoic uid was harvested and stored at À 70 C until needed.

The Boehringer Mannheim (BM) High Pure PCR Template Preparation Kit (Indianapolis, IN) was used to extract viral RNA from allantoic¯uid per the manufacturer's directions.

The S2 gene of the IBV strains was ampli®ed using primers that¯anked both sides of the entire S2 gene. The 3 H PCR primer (5 H -TTGAATCATTAAACAGAC-3 H ) was designated S2-3 H Ark, and the 5 H PCR primer (5 H -GTAGGTATTCTTACTTCACGTA-3 H ) was designated S2-5 H Ark. The relative primer positions using the ATG start site for the Beaudette strain S1 gene (M95169) as 1, were 1516 to 1537 for S2-5 H Ark and 3480 to 3497 for S2-3 H Ark. The reverse transcriptase (RT) and polymerase chain reaction (PCR) were conducted as previously described (17) . The amplicon was puri®ed and concentrated using GenElute TM spin columns (Supelco, Bellefonte, PA 16823-0048) and Microcon TM 30 columns (Amicon, Beverly, MA 01915), respectively. (20) ) and hydrophobicity plots (Hopp and Woods (21) and Kyte and Doolittle (22)) using S2 deduced amino acid sequence data were done with computer algorithms using MacDNASIS Pro V3.5 and Lasergene V 3.12.

There was a high nucleotide similarity for the S2 genes from the IBV strains used in this study ( Table  2 ). The S2 gene sequence for the related Arkansas serotype strains Ark 99 and Ark DPI were identical, while 3668-4 and GAV 92 were respectively 98.9% and 98.6% similar to both Ark 99 and Ark DPI. The S2 gene nucleotide sequences of the Florida 18288 and Connecticut 46 strains were 99.8% similar.

The deduced amino acid sequence of the S2 subunit was also compared ( There were few amino acid differences among all the IBV strains (Fig. 1) . The strains 3668-4 and GAV 92 had 7 and 11 amino acid substitution differences, respectively, when compared with the Ark 99 and Ark DPI strains. The Florida 18288 and Connecticut 46 strains had only two differences between themselves, and both were nonconservative.

Sequence data for the S2 genes of other IBV strains was used to construct a phylogenetic tree for the deduced amino acid sequence of the S2 subunit (Fig.  2) . In the alignment, members of the U.S. serotypes Arkansas, Mass, Connecticut, Florida, and foreign S2 Gene Sequence Variability serogroups B and C, fall into the same groupings as observed when deduced amino acid sequence data for the S1 subunit is used for phylogenetic analysis (Fig.  3) . However, the range of percent similarities was much less for the S2 subunit sequence than that observed for the S1 subunit sequence data.

There were no amino acid differences within the immunodominant region of the S2 subunit for the Arkansas serotype strains (approximately the ®rst 30 residues). There were also no differences between the Connecticut 46 and Florida 18288 strains in the immunodominant region, however, there were differ- Hydrophobicity plots using the Hopp and Woods (21) algorithm gave identical values of À 0.24 + 0.01 for each strain. However, there were differences in the predicted secondary structures using the method of Chou and Fasman (19) . The predicted secondary structure of the S2 subunit of the Ark 99 and Ark DPI strains were identical due to their identical protein sequence. The predicted secondary structure of the S2 subunit of the 3668-4 strain differed from that of the 

Ark 99 and Ark DPI strains due to amino acid substitutions at position 50 (E to G) and 70 (H to N) that resulted in the addition of two turns (Fig. 4) . The GAV 92 strain differed tremendously due to an amino acid substitution at position 50 (E to G), resulting in an odd number of turns between amino acids 40 and 75. The odd number of turns resulted in a 180 ¯i p in the middle of the predicted secondary structure.

The predicted secondary structure for the S2 subunit of the Florida 18288 and Connecticut 46 strains was remarkably different (Fig. 5 ). There were two nonconservative amino acid changes at positions 227 and 274. The alanine residue at position 227 for the Connecticut 46 strain was changed to a threonine residue for the Florida 18288 strain. This resulted in the changing of some amino acid residues from helix to sheet and the addition of a turn in the predicted secondary structure of the S2 subunit for the Florida 18288 strain. The histidine residue at position 274 for the Connecticut 46 strains was changed to a tyrosine residue for the Florida 18288 strain. This resulted in the changing of some amino acid residues from helix to sheet and the addition of a coil in the secondary structure of the S2 subunit for the Florida 18288 strain.

We analyzed six strains of IBV in the Arkansas, Connecticut, and Florida serotypes. Although S2 sequence data are more conserved among different strains of IBV than S1 sequence data, it appears that strains can be grouped into serotypes based on S2 gene nucleotide sequence data, as well as deduced amino acid sequence for the S2 subunit. This agrees with S1 gene phylogenetic trees for U.S. and international viruses. The only exception for grouping is between the Connecticut and Florida serotypes, which cannot be grouped into different serotypes using S1 gene or deduced amino acid sequence data, but can be separated serologically (14, 15) .

Based on the secondary structure predictions using the Chou and Fasman (19) algorithm it appears that only a few amino acid changes in the correct location can alter the shape of the S2 subunit. One change in the GAV 92 S2 deduced amino acid sequence ( position 50 E?G) led to a 180 ¯i p in the secondary structure prediction of the S2 subunit. The two nonconservative amino acid changes between the Florida 18288 and Connecticut 46 strains led to radically different secondary structure predictions. It is plausible that these S2 subunit secondary structure changes could affect the tertiary structure of the S2 subunit. Therefore, creating different interactions between the S1 and S2 glycoproteins that could change the quaternary structure of the spike glycoprotein. Such changes would affect antibody binding and therefore account for serologic differences between GAV 92, 3668-4, and Arkansas viruses as well as the serotype differences between the Connecticut and Florida strains.

The S1 and S2 subunits are known to interact by noncovalent attractive forces (5) . Other research on a different Coronavirus, mouse hepatitis virus, by Grosse et al., showed that a single amino acid change in the S2 subunit could create a S1 subunit speci®c monoclonal antibody resistant mutant (11) . This suggests that the interaction between S1 and S2 subunits may determine the shape or availability of S1 subunit speci®c epitopes. Whether the S2 subunit is actually involved in S1 subunit speci®c antibody recognition, sterically hinders antibody from binding to the S1 subunit, or effects the presentation of S1 subunit epitopes is not known. However, from our sequence data we hypothesize that the S2 subunit can affect binding of S1 subunit speci®c antibody due to S2 gene variability and subsequent secondary structure differences.

The nucleotide sequence data reported in this paper have been submitted to the GenBank nucleotide sequence data base and have been assigned the following accession numbers: Arkansas 99, AF094814; Arkansas DPI, AF094815; 3668-4, AF094816; GAV 92, AF094817; Connecticut 46, AF094818; Florida 18288, AF094819.

The Coronaviridae

Diseases of Poultry. Iowa State University Press

Fundamental Virology

The Coronaviridae

Arch of Virol 142, 2249±2256

A Laboratory Manual for the Isolation and Identi®cation of Avian Pathogens

Secondary structure prediction of the S2 glycoprotein for the Connecticut 46 and Florida 18288 strains of IBV using the Chou and Fasman

Proceedings from the National Academy of Science