Influenza A virus PB1-F2 gene.

To the Editor: Recently, Chen and co-workers described the expression of an 11th influenza A virus protein, designated PB1-F2 because this protein is encoded in the +1 open reading frame of the segment-2 RNA (1). Later, Chen et al. presented a preliminary analysis of 336 PB1 sequences from GenBank (2). We have extended the work on PB1-F2 and analyzed 1,864 partial and complete segment-2 sequences deposited in GenBank; these sequences belong to 79 influenza A virus subtypes. In summary, the following 8 observations should receive attention:

First, the size of PB1-F2 polypeptides ranges from 79 to 101 amino acids (aa); most isolates encode versions of either 87 or 90 aa. Because polypeptides of 79 aa are located within mitochondria, their truncation has no effect on the protein function. The frequency of the 79-aa PB1-F2 is [approximately equal to] 5%.

Second, a functional PB1-F2 is expressed by 92% of all segment-2 sequences, i.e., a polypeptide >78 aa. The proportion of intact PB1-F2 varies according to host (humans 90%, swine 76%, other mammals 100%, birds 95%).

Third, the H1N1 subtype comprises 3 genetic lineages. One clade has 2 branches: 1 branch includes the human viruses, with the pandemic 1918 virus at its root; the other branch includes the classic swine viruses. The third clade represents the European porcine isolates. Although all classic swine sequences have a truncated PB1-F2 (in-frame stop codons after 11, 24, and 35 codons), the early human isolates (H1N1 sequences from 1918 through 1947) have an intact PB1-F2. After 1956, however, a mutation became prevalent such that the recent sequences starting from A/Beijing/1/56 terminate after 57 codons. An exception to this rule is A/Taiwan/3355/97. Two human H1N1 isolates with an intact PB1-F2 coding sequence cluster in the H3N2 clade (A/Kiev/59/79, A/Wisconsin/10/98). The PB1 sequences of European porcine influenza A virus isolates cluster with European porcine H3N2 and H1N2.

Fourth, all H2N2 sequences are monophyletic and encode an intact PB1-F2. Fifth, the main sequence cluster of the H3N2 subtype comprises 3 branches: 1) porcine H3N2 and porcine H1N2 sequences from the United States, 2) porcine H3N2 isolates from Hong Kong and human H1N2, and 3) recent human H3N2 and some Japanese H3N2 isolates. Most of these sequences encode an intact PB1-F2.

Sixth, the cluster of European porcine influenza A virus isolates comprises the subtypes H1N1, H1N2, and H3N2. The lack of distinct clades for each subtype indicates frequent reassortment in the evolution of these viruses. Of the segment-2 sequences, 56% encode an intact PB1-F2.

Seventh, other porcine isolates of various subtypes represent trans-species infections or single reassortment events. And eighth, the segment-2 sequences of many avian influenza A virus isolates encode intact PB1-F2. Considerable proportions of truncated PB1-F2 genes were found in the H5N2, H6N6, H9N2, and H13N2 subtypes. However, because of the small number of sequences available, this observation may not be important.

In conclusion, PB1-F2 is expressed in most avian and many porcine influenza A virus isolates. This finding contrasts with those in the initial publication, which stated that PB1-F2 is not expressed in many animal isolates, particularly those of porcine origin (1). Because PB1-F2 was described as a proapoptotic protein probably counteracting the host immune response, why numerous human and porcine isolates lack this protein without selective disadvantage remains unclear.

Roland Zell, * Andi Krumbholz, * and Peter Wutzler *

* Friedrich Schiller University, Jena, Germany

References

(1.) Chen W, Calvo PA, Malide D, Gibbs J, Schubert U, Bacik I, et al. A novel influenza A virus mitochondrial protein that induces cell death. Nat Med. 2001;7:1306-12.

(2.) Chen GW, Yang CC, Ysao KC, Huang CG, Lee LA, Yang WZ, et al. Influenza A virus PB1-F2 gene in recent Taiwanese isolates. Emerg Infect Dis. 2004;10:630-6.

Address for correspondence: Roland Zell, Institute of Virology and Antiviral Therapy, Medical Centre at the Friedrich Schiller University, Hans Knoell Str 2, D-07745 Jena, Germany; email: [e-mail]

In response: Zell et al. (1) performed an extensive genetic investigation of PB1-F2, based on up-to-date GenBank sequences. Their sample size (1,864) greatly outnumbered ours (336) in a previous study (2) and thus definitely better portrays the genetic characteristics of PB1-F2. We appreciate their analyzing these samples by subdividing nonhuman strains into different species, which we did not do (2). Their analysis is especially meaningful for the global pandemic threat from avian influenza viruses, which increases the need to study interspecies adaptation and transmission.

Zell et al. found that 92% of PB1 RNA encodes a functional PB1-F2, compared with our 79% (264/334), which supports the increasingly crucial role of PB1-F2 in influenza virology. They found the proportion of intact human PB1-F2 to be 90%, a substantial boost from our 68% (67/99), which was based on data from late 2003 (2). This increase is apparently caused by the increasing number of human H3N2 sequences (mostly encoding an intact PB1-F2 compared with H1N1) deposited in the past 2 years.

Human H1N1 from 1918 through 1947 contains full-length PB1-F2, whereas human H1N1 beginning in 1956 has a truncated PB1-F2 after codon 57. As reported by Zell et al., only 3 human H1N1 strains contain full-length PB1-F2: A/Kiev/59/79, A/Taiwan/3355/97, and A/Wisconsin/ 10/98. The PB1 genes of A/Kiev/ 59/79 and A/Wisconson/10/98 were found clustered with human H3N2 as a result of natural reassortment between human H1N1 and H3N2 strains. On the other hand, the asynonymous mutation found on A/Taiwan/3355/97 enabled the translation to get past the usual stop codon at position 58, which other H1N1 strains exhibit. A/Taiwan/3355/97 (H1N1) was isolated from a patient with severe pneumonia. Animal study has demonstrated that the existence of full-length PB1-F2 contributed to pathogenesis in mice (3). We speculate that the expression of a full-length PB1-F2 may contribute to disease severity in humans.

The C-terminal domain of PB1-F2 contains the mitochondrial signal and can trigger apoptosis in specific immune-related cells. Our recent work (4) comparing avian and human influenza A viruses also found that many species-associated amino acid signatures are located on the C terminal of PB1-F2. This finding highlights the importance of further investigating the role of PB1-F2 on interspecies infection.

Guang-Wu Chen * and Shin-Ru Shih *

* Chang Gung University, Taoyuan, Taiwan

References

(1.) Zell R, Krumbholz A, Wulzler P. Influenza A virus PB1-F2 gene [letter]. Emerg Infect Dis. 2006;12:1607-8.

(2.) Chen GW, Yang CC, Tsao KC, Huang CG, Lee LA, Yang WZ, et al. Influenza A virus PB1-F2 gene in recent Taiwanese isolates. Emerg Infect Dis. 2004;10:630-6.

(3.) Zamarin D, Ortigoza MB, Palese P. Influenza A virus PB1-F2 protein contributes to viral pathogenesis in mice. J Virol. 2006;80:7976-83.

(4.) Chen GW, Chang SC, Mok CK, Lo YL, Kung YN, Huang JH, et al. Genomic signatures of human versus avian influenza A viruses. Emerg Infect Dis 2006:9:1353-60.

Address for correspondence: Guang-Wu Chen, Chang Gung University, Department of Computer Science and Information Engineering, 259 Wen-Hua 1st Rd, Kwei-shan, Taoyuan, Taiwan 333; email: gwchen@ mail.cgu.edu.tw