Stop-codon variations in non-structural protein NS1 of avian influenza viruses

KEYWORDS:

• avian
• influenza virus A
• non-structural protein 1
• NS1
• truncations

This article refers to:

Prevalence of the C-terminal truncations of NS1 in avian influenza A viruses and effect on virulence and replication of a highly pathogenic H7N1 virus in chickens

To what extent are mutations and truncations tolerated in influenza virus proteins? One approach to address this issue relies on the identification of viable mutants isolated in vitro after genome-wide mutagenesis.¹ Another approach consists in analyzing the thousands of sequences that were deposited at the Influenza Virus Resource (http://www.ncbi.nlm.nih.gov/genomes/FLU/FLU.html). In a recently published article in Virulence, Abdelwhab et al.² carried out a systematic study of C-terminal truncations of the Non-structural protein NS1 in avian influenza A viruses (IAVs) since the beginning of the 20th century. Influenza A viruses are enveloped viruses of the Orthomyxoviridae family, whose genome consists of 8 negative-strand RNAs which collectively encode more than 10 viral proteins. The large diversity of IAVs is maintained as a viral pool that circulates among a large variety of wild aquatic and shore bird species that act as their reservoir. Viruses frequently spill over from this wild reservoir to infect domesticated birds, including gallinaceous poultry. Once established in domesticated birds, IAVs can transmit to mammalian species, notably swine, horses and humans.³ Adaptation to new hosts, notably gallinaceous poultry and mammals, can be facilitated by a number of mutations in the viral genes, including those encoding the polymerase subunits and the viral glycoproteins haemagglutinin (HA) and neuraminidase (NA).⁴ NS1 is encoded by the unspliced mRNA transcribed from segment 8 of the viral genome, while a spliced species of this mRNA encodes Nuclear Export Protein (NEP, formerly named NS2). This 230-residue protein is composed of 2 structured domains that are linked by a flexible linker: the RNA-binding domain (RBD, residues 1-73) and the effector domain (ED, residues 80-230). Through its interactions with RNAs and several protein partners, NS1 counteracts the antiviral response of the host cell, blocks the maturation and nucleo-cytoplasmic export of cellular mRNAs, and enhances the translation of viral mRNAs.^5-7

How do these length variations occur?

NS1 most often occurs as a 230 residue-protein. However, premature stop codons, or alternatively, suppression of the genuine stop codon (codon 231) result in length variations at NS1's C-terminus. C-terminal truncations of NS1 are not uncommon, since about 18% of the ~13.000 sequences analyzed by Abdelwhab et al. harbored various truncations. Although the viral polymerase is prone to errors, the sequence space sampled by these mutated variants is limited, which probably explains the fact that only a limited set of premature stop codons is observed. On the one hand, mutations are tolerated to the extent that they are not detrimental to the biological functions of the viral proteins. NS1, in spite of the constraint originating from the overlapping reading frame of NEP after NS1 codon 168, is rather permissive to mutations,¹ and the sequence variability of the NS segment stands in third position after those of the HA and NA segments.⁸ On the other hand, the spectrum of mutations introduced by the viral RNA-dependent RNA Polymerase (RdRP) is biased. A mutational analysis by Cheung et al., based on transiently expressed polymerase complexes from 2 distinct viruses in 293T cells, identified AAAG as a highly mutable sequence motif, and revealed that the viral polymerase favors transitions over transversions, with a predominance of A>G and U>C among the 4 possible transitions.⁹ All these facts concur to a biased spectrum of mutations along the viral genome. Abdelwhab et al. observed that by far the most prevalent C-terminal truncation occurs through conversion of codon 218 to a stop codon, thus resulting in the NS217 variant. This truncation is especially prevalent in NS1s of avian IAVs belonging to the H6N1, H6N2, H6N6, H7N3, H7N7, H7N9 and H9N2 subtypes. Alignment of these avian-origin sequences reveal that the mutated codon 218 (most frequently CAG>UAG) lies within the purine-rich sequence AAAGCAGAAA (with the mutated C underlined). Similar truncations are also observed in non-avian influenza viruses: more than 75% of swine influenza viruses harbor a C-terminally truncated NS1, which is also found in human H1N1 viruses that have been circulating since the 2009 pandemic. These resulted from mutations at codon 220 (CGG>UGA) for the majority, and at codon 218 (CAG>UAG) for a minority of them, likewise within purine-rich motifs GAAACGGAAAA and RRRRCAGAAA, respectively.

Do truncations alter the viral phenotype?

C-terminal truncations partially or completely remove NS1's C-terminal tail (CTT). The CTT is likely an intrinsically disordered region^10,11 and contains features that could be involved in NS1's biological activities. These include the PDZ-binding motif, which is made up of NS1's 4 C-terminal residues and generally consists of the motif ²²⁷ESEV²³⁰ in NS1 of avian influenza viruses,^8,12 as well as lysine residues 219 and 221, the SUMOylation of which was shown to enhance the stability of NS1.¹³ The viral phenotype of NS1-truncation variants could therefore provide clues as to NS1 activities involving this C-terminal tail. In order to address this issue, Abdelwhab et al. focused on a well-documented case, i.e. the avian H7N1 viruses that spread through italian poultry flocks in several waves of Low-Pathogenicity (LPAI) and High-Pathogenicity Avian Influenza (HPAI) between 1999 and 2001 and led to the death or culling of millions of birds.^14,15 These viruses harbored the B allele of NS1, including 2 C-terminally truncated variants (NS224 and NS220), with premature stop at codon 225 (CGA>UGA) or 221 (UAC>UAA), respectively.¹⁶ One could speculate that a first truncation occurred (NS224), followed by a further truncation to NS220. However, close examination of the sequences and of the dates of virus isolation does not support this view. In fact, NS220 viruses were isolated from September 1999, a few months before the emergence of HPAI viruses, up until February 2001, well after the end of the HPAI epizootic. The HPAI isolates all harbored the NS224 variant, hinting to a possible link with the HPAI phenotype. In order to experimentally address this hypothesis, Abdelwhab et al. genetically engineered 3 variants of the H7N1 HPAI virus. Two of these variants harbored the NS segment of the High-Pathogenicity (HP) virus, either unmodified (HP-NS224), or with a reversion of the premature stop codon to the initial Arginine codon (HP-NS230). The third recombinant virus (HP-NSLp) harbored the NS segment of the Low-Pathogenicity (LP) virus, which encodes a NS230 that also differs from NS1 of the HP virus by 3 substitutions (A117V, I136V and N139D). In experimentally infected chickens, the moderate differences observed between the 3 viruses do not support the hypothesis that the C-terminal truncation increases the pathogenicity. Abdelwhab et al. did not engineer a «HP-NSLp-224», i.e., with the 3 LP-specific substitutions and the C-terminal truncation; however, a previous study by Soubies et al. had shown that the C-terminal truncation of NS1 from the LP virus did not significantly alter its pathogenicity.¹⁷ Altogether, these data do not show an increased pathogenicity associated with NS1 C-terminal truncation, although it cannot be formally excluded that this truncation, along with the other substitutions observed between NS1s from the HP and LP viruses concur to marginally alter the pathogenicity. For instance, Abdelwhab et al. showed a reduced virus excretion in oropharyngeal and cloacal swabs from HP-NSLp-infected animals, as well as a decreased replication in the brain, which, as the authors suggest, may result from a decreased tropism for endothelial cells. The minor role, if any, of NS1 modifications on the viral phenotype is in line with the findings of a previous article by the same group of authors, who showed that the HA protein is the major virulence determinant of the H7N1 virus in chickens: the multibasic cleavage site of the HA, along with 3 mutations in the HA2 domain, suffice to confer the HP phenotype in chickens.¹⁸

What is the fate of viruses harboring Ns1-Truncations?

A rigorous answer would require a deep analysis of nucleotide sequences from comprehensively sampled viruses. However, Abdelwhab et al. showed that NS1 truncations are markedly prevalent in viruses isolated in non-aquatic birds (notably chickens, turkeys and quails), in which they amount to ~30% or more of the total sequences per host. Furthermore, a rapid analysis of a few NS nucleotide sequences through the Influenza Virus Resource seems to indicate that most of the sequences encoding truncated NS1s are not subsequently found in wild birds. This may suggest either that NS1 truncation somehow contributes to adaptation to non-aquatic birds, or alternatively that viruses once established in these species are more tolerant to modifications. With about 20 billion chickens on Earth, the high level of replication of AIVs in poultry probably results in a broad expansion of the viral quasispecies. Do these poultry-origin viruses add up to the wildlife virus pool? Apparently not: IAVs once adapted to gallinaceous poultry rarely go back to the wild bird IAV viral pool, probably because they are no longer adapted to wild birds.¹⁹ This de-adaptation, together with eradication by the stamping out of infected poultry flocks probably contribute to the extinction of poultry-adapted IAVs and account for the low prevalence of NS1-truncated variants in ducks and in the wild avifauna reservoir.

The low prevalence of NS1-truncated variants in the wildlife virus pool is in contrast with another length variation of NS1, resulting from a deletion of codons 80-84 in H5N1 viruses that occurred in 2000. NS segments with this deletion, which are easier to track than the C-terminal truncations, are predominant in currently circulating H5N1 viruses and have spread to several non-H5N1 viruses.^20,21

The work by Abdelwhab et al. is therefore one of the studies that could enhance our understanding of the flow of influenza virus genomic segments within and between the viral pools, in the wild avifauna reservoir, in gallinaceous poultry and in mammals.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

This work was funded in part by Institut Carnot Santé Animale (Project Flumli).