| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Institute of Virology and Antiviral Therapy, Medical Center at the Friedrich Schiller University, Hans-Knoell-Str. 2, D-07745 Jena, Germany
2 Institute of Anatomy II, Medical Center at the Friedrich Schiller University, Teichgraben 7, D-07743 Jena, Germany
Correspondence
Roland Zell
Roland.Zell{at}med.uni-jena.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The GenBank/EMBL/DDBJ accession numbers for the nucleotide sequences reported in this paper are DQ836168–DQ836179.
Tables of influenza A virus subtypes expressing full-length and truncated PB1-F2, and of McDonald–Kreitman tests for PB1 and PB1-F2 are available as supplementary material in JGV Online.
These authors contributed equally to this work. 
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Eight segments encode the subunits of the transcriptase complex (PB1, PB2, PA, nucleoprotein NP), the matrix protein (M1), the pro-apoptotic NS1, the nuclear export protein (NS2/NEP) and three integral membrane proteins (haemagglutinin HA, neuraminidase NA, proton channel M2). Recently, the existence of an eleventh protein encoded by an alternative ORF of segment 2 was reported (Chen et al., 2001). This protein, designated PB1-F2, shows variable expression levels in individual infected cells, rapid proteosome-dependent degradation and a predominantly mitochondrial localization in approximately 50 % of PB1-F2-expressing cells. The latter is achieved by a mitochondrial targeting sequence (MTS) which has been predicted to form a positively charged amphipathic helix (Gibbs et al., 2003; Yamada et al., 2004). It has been proposed that PB1-F2 is a pro-apoptotic protein interacting with the mitochondrial proteins ANT3 (adenine nucleotide translocator 3) and VDAC1 (voltage-dependent anion channel 1) which are located in the mitochondrial envelope (Zamarin et al., 2005). Cell-type-dependent induction of cell death was observed in infected host immune cells, but not in epithelial cells. Originally described as an 87-residue protein of influenza virus A/PR/8/34, a 90-residue PB1-F2 has also been observed (Chen et al., 2001, 2004). Moreover, it has been reported that PB1-F2 is present in nearly all human FLUAV isolates, but absent in many animal isolates, particularly of porcine origin (Chen et al., 2001). The aim of this study was a comprehensive evaluation of the FLUAV segment 2 sequences deposited in the GenBank nucleotide sequence database to gain a detailed view on the prevalence of PB1-F2. As this systematic evaluation revealed unexpected results, the expression of PB1-F2 in cells infected with a porcine FLUAV isolate and its mitochondrial location were demonstrated. To describe the selective forces driving segment 2 evolution, a molecular genetic analysis was conducted.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2004) were obtained from the Federal Institute for Risk Assessment (Berlin, Germany). RNA preparation, amplification of segment 2-specific cDNAs and nucleotide sequencing was conducted as described previously (Schmidtke et al., 2006).
Genetic and phylogenetic analysis.
Nucleotide and amino acid sequences were aligned manually. Transition/transversion ratios (R=ts/tv) were computed with MEGA version 3.1 (Kumar et al., 2004). The transition/transversion rate ratio 
was estimated according to =2R. Segregation (frequency of sites with two or more nucleotides at a site) and diversity (mean number of differences per site between two sequences) were estimated with DnaSP4.10.7 (Rozas et al., 2003). Non-synonymous/synonymous nucleotide substitution rate ratios (
=dN/dS) (Nei & Gojobo, 1986) were computed with DnaSP4.10.7 from the Jukes–Cantor corrected synonymous and non-synonymous nucleotide divergences (dS, number of synonymous nucleotide substitutions per synonymous site; dN, number of non-synonymous nucleotide substitutions per non-synonymous site). The Jukes–Cantor correction accounts for multiple substitutions at the same site. Mean dS–dN differences and their variances were computed with MEGA 3.1. Neutrality tests [Tajima's D (Tajima, 1989), Fu & Li's D*, Fu & Li's F* (Fu & Li, 1993) and McDonald–Kreitman's G (McDonald & Kreitman, 1991)] were conducted with DnaSP4.10.7 and the Z-test was performed with MEGA 3.1. Unrooted tree topology from multiple alignments of the PB1-encoding nucleotide sequences were obtained by using the neighbour-joining method calculated with MEGA 3.1. Consistency of branching was tested with a bootstrap analysis with 10,000 resamplings of the data using MEGA 3.1.
Construction of plasmids for bacterial PB1-F2 expression and transient eukaryotic expression of PB1-F2-enhanced green fluorescent protein (EGFP) fusion proteins.
PB1-F2-encoding sequences of isolates A/swine/Potsdam/15/81 (90 aa) and A/swine/Bakum/1832/00 (79 aa) were amplified using the upstream oligonucleotides 8-PB1F2-Xho (5'-GATCGACTCGAGAATGGAACAGGGACAGGATACACC-3') and 12-PB1F2-Xho (5'-GATCGACTCGAGAATGGAACAGGGACAGGATACACC-3'), and the respective downstream oligonucleotides 8-PB1F2-Bam (5'-GATCGAGGATCCTATTTGTCCACTCTTGTTTGCTG-3') and 12-PB1F2-Bam (5'-GATCGAGGATCCTTCGTTTCAAGACACGAGTTTTC-3'). An artificial construct encoding the N-terminal 57 aa was generated by use of oligonucleotides 8-PB1F2-Xho and 8*-PB1F2-Bam (5'-GATCGAGGATCCTGGACACAATCTGTTTGTGCATATC-3'). The amplification reaction results in PB1-F2 fragments containing a 5'-terminal XhoI restriction enzyme recognition site followed by an ATG translation initiation codon and the respective coding triplets. The 3' terminus comprises a BamHI recognition sequence. PCR products were digested with BamHI and XhoI, ligated into the 5' multiple cloning site of the pHygEGFP vector (Clontech) and used to transform competent Escherichia coli cells. Accuracy of the plasmid constructs was determined by sequence analysis.
Transfection of HeLa cells, mitochondrial labelling and confocal microscopy.
HeLa cells were grown on slides up to 75 % confluence. Then, the cells were transfected with the respective plasmid constructs using the Effectene transfection reagent (Qiagen), according to the manufacturer's instructions. Twenty-four hours post-transfection, cells were examined by confocal laser scanning microscopy. To assess patterns of subcellular localization of various PB1-F2-EGFP constructs, transfected cells were labelled in vivo with the mitochondria-specific fluorescence markers JenMitoStain 1 {JMS-1, (E)-N-n-propyl-2-[2-(2-pyrollyl)ethenyl]benzothiazolium iodide} and JenMitoStain 28j {JMS-28j, (E)-N-ethoxyethyl-4-[2-(4-N,N-dimethylaminophenyl)ethenyl]chinolinium bromide} (Krieg, 2006; R. Krieg, A. Eitner, W. Günther, C. Schürer, J. Lindenau, H.-J. Halbhuber, unpublished). The selective stains used for mitochondria are newly developed dyes belonging to the class of cyanine dyes. So far, JMS-1 and JMS-28j are developed only for research purposes and are not available commercially. Weak basic and electrochromic properties of the positively charged dye provide the main preconditions for its specific incorporation into mitochondria associated with a pronounced change of spectral properties (Ross et al., 2005; Krieg, 2006). The fluorescence intensity of accumulated JMS-1 and JMS-28j are dependent on mitochondrial respiration as well as ATP synthesis. An impairment of mitochondrial bioenergetics results immediately in a decrease in fluorescence intensity.
JMS-1 and JMS-28j were dissolved in culture medium to a final concentration of 1 µM. HeLa cells were incubated at 37 °C for 1 h under growth conditions. Subsequently, EGFP as well as mitochondrial stain fluorescence in living transfected cells was examined with a confocal laser scanning microscope (LSM 310; Zeiss) equipped with an Ar laser and an HeNe laser. EGFP was excited with the Ar laser at an excitation wavelength of 488 nm. The emission signal of EGFP was detected using a bandpass optical filter of 505–545 nm and visualized in the green channel. The mitochondria markers were excited with the 543 nm laser line from the HeNe laser and recorded in the red channel using a longpass optical filter of 560 nm. The laser power was kept at very low intensity values to limit cytotoxicity.
Generation of polyclonal PB1-F2 antibodies.
Oligonucleotides PB1F2-Nde (5'-GATCGACATATGGAACAGGGACAGGATACACCATGG-3') and PB1F2-Bam-stop (5'-GATCGAGGATCCTCAATTTGTCCACTCTTGTTTGCT-3') were used to amplify the PB1-F2 sequence of A/swine/Potsdam/15/81. The amplification results in a PB1-F2 fragment which is flanked by a 5'-terminal NdeI restriction site and a 3'-terminal stop codon followed by a BamHI recognition sequence. The digested amplicon was ligated into a pET-15b expression vector (Novagen) which carries an N-terminal His-tag sequence followed by a thrombin cleavage site. PB1-F2 expression was induced in transformed Escherichia coli strain BL21 by IPTG. PB1-F2-6xHis fusion protein was purified using Ni-NTA agarose (Qiagen). One milligram PB1-F2-6xHis eluate ml–1 was emulsified with an equal volume of Freund's adjuvant (Sigma–Aldrich) dissolved in 0.9 % sodium chloride and used to immunize rabbits twice in intervals of 4 weeks. One month after the booster, the rabbits were bled. The obtained antiserum was tested against bacterial protein lysate by Western blotting.
Immunohistochemistry.
MDCK cells were grown on 16-well slides and infected at an m.o.i. of 0.1 with the respective virus isolate. After an incubation period of 12–24 h, cells were fixed with methanol and washed with PBS. The endogenous peroxidase activity was inhibited with hydrogen peroxide, and cellular biotin activity was blocked by addition of avidine and biotin. The PB1-F2 antiserum was diluted up to a factor of 1 : 9.600 and co-incubated with pre-treated cells for 12 h at 4 °C. Then, cell–antibody complexes were washed with PBS and co-incubated with biotinylated secondary antibodies for 30 min. Subsequently, streptavidine-peroxidase complex was added for an additional 30 min. A chromogenic substrate (Litaris) was administered and the development of a chromogenic reaction was monitored by microscopy. After disinfection with 4 % formalin, slides were counterstained with Mayer's haemalum and prepared for light microscopy.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
described 64 of 75 FLUAV segment 2 sequences accessible in the GenBank nucleotide sequence database to have PB1-F2 ORFs of 87 and 90 aa residues, respectively. Eleven sequences, six of which were porcine isolates, had in-frame stop codons leading to truncated PB1-F2 polypeptides. This preliminary survey led to the conclusion that PB1-F2 is present in nearly all human FLUAV isolates, but absent in many animal isolates, particularly of porcine origin (Chen et al., 2001). However, sequencing of segment 2 RNA of 12 porcine H1N1, H1N2 and H3N2 influenza viruses isolated in Germany between 1981 and 2002 revealed both intact and truncated PB1-F2 ORFs. To analyse the prevalence of intact PB1-F2 of FLUAV in more detail, GenBank was scanned for complete and partial FLUAV segment 2 sequences and 2226 entries were found to be suitable for further analysis of PB1-F2 (last update: 15 June 2006). These sequences represent 80 FLUAV subtypes: 1228 human, 137 non-human mammalian (119 porcine, 10 equine, 3 canine, 1 feline, 3 tiger, 1 leopard) and 861 avian isolates (Table 1). Regardless of their origin, 87 % of these sequences encoded a PB1-F2 polypeptide greater than 78 aa which is supposed to be functional (Yamada et al., 2004). Among these, 29 sequences contained an ORF encoding 101 aa, 27 of which belonged to the H3N2 subtype (human isolates) and one each to H5N1 and H9N2 (both avian isolates). A detailed analysis of the 80 FLUAV subtypes (supplementary Table S1) revealed the following subtypes with a strikingly low number of intact PB1-F2 ORFs: H1N1, H6N6 and H13N2. For the remaining subtypes, the fraction of intact PB1-F2 ORFs ranged from 82 to 100 %.
|
I exchange in PB1, another at position 28 yielded a T58I exchange. These mutations were observed in A/chicken/Hong Kong/739/94 (H9N2) and A/New York/541/1998 (H3N2). Another M40I substitution found in A/swine/Alberta/56626/03 (H1N1) did not result in a stop codon within the PB1-F2 ORF. Twenty-nine sequences revealed a mutation of the stop codon at position 90, yielding leucine or tryptophan. These versions of PB1-F2 have a length of 101 aa. Five sequences were observed without an AUG start codon in the PB1-F2 ORF [A/swine/Belzig/2/01 (H1N1), A/swine/Bakum/8602/99 (H3N2), A/swine/Spain/33601/2001 (H3N2), A/chicken/Chis/15224/1997 (H5N2), A/quail/Hong Kong/AF157/92 (H9N2)], but with a threonine codon instead, and in-frame stop codons at positions 8, 11 or 35. The PB1 sequences were unaffected. For molecular genetic analyses, 247 segment 2 sequences (PB1 ORF, 2271 nt), which were considered to be representative for human, porcine and avian FLUAV isolates, were subgrouped into six datasets: set 1 representing European porcine influenza virus isolates (H1N1, H1N2, H3N2; n=31 sequences); set 2 representing classic swine influenza virus isolates (H1N1; n=11); set 3 representing human H1N1 isolates (n=23); set 4 representing human H2N2 isolates (n=21); set 5 representing human H3N2 and H1N2 isolates as well as porcine H3N2 and H1N2 isolates from non-European countries and reassortants from those viruses (n=80); and set 6 representing other FLUAV subtypes, mainly avian isolates (n=81). Set 7 included the sequences of sets 1–6 and three equine/canine FLUAV sequences (250 sequences in total). To identify significant differences of the PB1- and PB1-F2-encoding gene regions, the evolutionary change in nucleotide sequences was analysed. For the molecular genetic analyses, transition/transversion (ts/tv) rate ratios, segregation, diversity and the rates of synonymous and non-synonymous substitutions of the seven datasets were determined.
The ratio of the ts/tv rates (=2R) indicates the occurrence of non-random distribution of nucleotide substitutions. The mean value of the 250 sequences (set 7) was 6.8 for the PB1 ORF and 9.2 for the PB1-F2 ORF, showing a high proportion of transitions (Table 2). Datasets 1–6 yielded a similar excess of transitions. However, the values of PB1-F2 differ from the extremely high estimate of 33.3 previously found for avian FLUAV PB1-F2 (Obenauer et al., 2006).
|
). A segregating site is defined as any nucleotide/amino acid site that shows substitutions among a compared set of sequences while diversity is the mean number of nucleotide/amino acid differences per site between two sequences. The mean value of obtained after pairwise comparisons of PB1 sequences was estimated to be 0.1174 (Table 2
). Using the sliding window option of DnaSP4.10.7 software (window length 100, step size 25), no striking differences in across the segment were observed, indicating that the complete sequence is underlying the same selective force. At the nucleotide level, the mean values of and pS are similar for the PB1 and PB1-F2 ORFs in each dataset (Table 2). For amino acid sequences, however, is 6- to 13-times and pS is 2- to 7-times greater for PB1-F2. This observation is the consequence of mostly synonymous third base substitutions in the PB1 ORF which correspond to non-synonymous exchanges at the second codon positions of the +1 shifted PB1-F2 ORF (Table 2). Such non-synonymous substitutions result in amino acid exchanges. Analysis of the base substitutions revealed 707 segregating sites at the third nucleotide position of the PB1 codons (n=250 sequences). The corresponding pS value for the third codon position of the PB1 ORF was 0.94, while the mean pS value is 0.51 (Table 2). Accordingly, 88 of 90 nucleotide positions of the PB1-F2 sequences were segregated (pS=0.98); the mean pS value of the PB1-F2 ORF is 0.51.
The rates of synonymous and non-synonymous substitutions per site (dS, dN) should provide an indication of the selective forces driving the evolution of the segment 2 sequences. For this analysis, dS and dN of each dataset were estimated und used to calculate the difference between dS and dN as well as values. The mean dS–dN differences of the PB1 gene were positive and significant (mean SEM 0.01). The mean values ranged from 0.015 to 0.062. Both estimations indicate negative (purifying) Darwinian selection (dS>dN, Table 3). In contrast, negative dS–dN values of the PB1-F2-encoding sequences and mean values greater than one may suggest positive (diversifying) selection (dS<dN). To prove the significance of these results, several tests for detecting selection were conducted. The Tajima test (Tajima, 1989) and two approaches of Fu & Li (1993) are based on DNA polymorphism. These assays examine the relationship between the number of segregating sites and nucleotide diversity. The results for Tajima's D, Fu & Li's D* and Fu & Li's F* assays, though negative, were not significant for each test (Table 3). Neutral selection (null hypothesis H0: dN=dS) was also investigated employing the Z-test first used by Miyata & Yasunaga (1980). This test is based on a parsimony analysis. It computes the variances of dS–dN by the bootstrap method. If dN is significantly greater than dS, positive Darwinian selection can be assumed. For the PB1 datasets, negative selection (alternative hypothesis H1: dS>dN) was tested using the MEGA 3.1 software [analysis settings: standard error computation by bootstrap (1000 replicates), modified Nei–Gojobori method, p-distance]. The mean probability for the rejection of the alternative hypothesis of all PB1 datasets was P<0.001. The results significantly (at the 5 % level) favour the alternative hypothesis. For PB1-F2, the Z-test favoured positive selection (dN>dS). Another neutrality test developed by McDonald & Kreitman (1991) compares closely related species or datasets. In a first step, it determines variable and invariable nucleotide sites of the tested sequences and estimates the number of fixed and polymorphic nucleotides of the variable nucleotide sites. Then, synonymous and non-synonymous nucleotides of fixed and polymorphic nucleotides are determined. The four resulting values are applied to a statistical G-test. Positive selection is demonstrated when the ratio of non-synonymous (n) and synonymous (s) fixed (F) mutations is significantly greater than the ratio of non-synonymous and synonymous polymorphic (P) nucleotides (nF/sF>nP/sP). For the sequences of several, but not all PB1 datasets, the McDonald–Kreitman test revealed a significant excess of synonymous exchanges (supplementary Table S2a). However, the pairwise comparison of closely related PB1 datasets yielded non-significant G values, and comparison of the PB2 datasets did not allow the test to be performed due to the low number of fixed substitutions (supplementary Table S2a).
|
). The H1N1 cluster divides into two main lineages (Fig. 1b), one representing isolates of classic swine influenza viruses, the other consisting of human isolates. Those human viruses isolated between 1918 and 1947 have a full-length PB1-F2. Starting from A/Fort Worth/50 (isolated 1950), all human H1N1 isolates have a truncated PB1-F2 polypeptide with 57 aa. The classic swine isolates have truncated PB1-F2 ORFs with various in-frame stops after codons 11, 25 and 34. The H1N1 clade of human and classic swine FLUAV is genetically unrelated to a collection of avian H1N1 viruses isolated between 1977 and 2002 in Alberta, Canada (not indicated in Fig. 1). Distinct genetic lineages can be demonstrated for all segments (e.g. Guan et al., 1996; Brown et al., 1998).
|
, 1994; Brown et al., 1998; Marozin et al., 2002). As demonstrated in Fig. 1(c), segment 2 of European porcine H1N1, H1N2 and H3N2 viruses has a monophyletic origin with the exception of A/swine/Finisterre/127/99 (H3N2) which is a reassortant of a porcine and a human H3N2 (Fig. 1d). Eleven of 31 full-length sequences and 7 of 12 partial sequences (in total 18 of 43 segment 2 sequences) encode a truncated PB1-F2 (42 %; compare Table 4, subgroup 1). The clade of European porcine FLUAVs also includes one human and some porcine viruses isolated in Hong Kong between 1999 and 2002 (Gregory et al., 2001).
|
) includes human and numerous porcine FLUAV isolates from America and Asia, as well as H1N2 and a few H1N1 isolates. Most of these sequences encode an intact PB1-F2. Within the H3N2 cluster, genetic drift has led to a continuous evolution of the segment 2 sequences with distinct branches representing lineages of human and porcine H3N2 viruses and the emergence of human and porcine H1N2 viruses as a result of independent reassortment events. Human H3N2 isolates from the state of New York collected in 1996 and 1997 are an example of the co-circulation of two distinct segment 2 lineages in the same area: while segment 2 of one lineage encodes an intact PB1-F2, the other lineage is characterized by a unique truncation after codon 63 (Fig. 1d). Interestingly, the PB1 sequences of five H1N1 isolates cluster within the H3N2 branch: (i) A/Kiev/59/79, an H1N1-H3N2 reassortant, (ii) A/Wisconsin/10/98, a reassortant of a classic swine H1N1 virus and porcine H3N2, isolated from a man (data not shown), and (iii) A/swine/Ontario/53518/03 (H1N1), A/swine/Ontario/11112/04 (H1N1) and A/swine/Ontario/23866/04 (H1N1) which are H1N1/H1N2 reassortants (Karasin et al., 2006).
Other porcine isolates characterized as subtypes H1N1 (A/swine/Saskatchewan/18789/02), H3N2 (A/swine/Hong Kong/126/82), H3N3 (A/swine/Ontario/K01477/01, A/swine/Ontario/42729A/01), H4N6 (A/swine/Ontario/01911-1/99), H5N1 (A/swine/Fujian/F1/2001, A/swine/Fujian/1/2003, A/swine/Shandong/2/03) and H9N2 (A/swine/Hong Kong/9/98, A/swine/Hong Kong/10/98, A/swine/Korea/S452/2004) represent either single reassortment events or trans-species infections. These viruses (not indicated in Fig. 1) encode a full-length avian-like PB1-F2 protein.
The equine and canine FLUAV PB1 sequences form a distinct phylogenetic clade. The respective viruses belong to the H3N8 and H7N7 subtypes and were collected from 1973 to 2004 in different regions (UK, USA).
Detection of PB1-F2 expression of porcine influenza A viruses
A previous study described PB1-F2 as a putative protein absent from many animal isolates, particularly in those of porcine origin (Chen et al., 2001). However, approximately 60 % of the segment 2 sequences of European porcine isolates, all sequences of other mammalian isolates and 96 % of the avian isolates exhibit a PB1-F2 expressing a functional MTS (Tables 1 and 4). To elucidate the significance of this observation, experiments were performed to demonstrate the presence of the PB1-F2 polypeptide of porcine FLUAV in virus-infected MDCK cells. For this purpose, RNA of A/swine/Potsdam/15/81-infected cells was reverse-transcribed and the cDNA was used to clone the PB1-F2 ORF into the pET15b expression vector. Recombinant PB1-F2 was purified and used to immunize rabbits. The antiserum was employed to detect the PB1-F2 polypeptide in MDCK cells, infected at an m.o.i. of 0.1 with porcine FLUAV isolates, by immunohistochemistry. Twelve hours post-infection, the 90 aa protein of A/swine/Potsdam/15/81 was readily detectable in virus-infected cells, but not in non-infected cells (Fig. 2).
|
; Gibbs et al., 2003; Yamada et al., 2004; Zamarin et al., 2005). To investigate the subcellular localization of PB1-F2 of porcine FLUAV, HeLa cells were transfected with pHygEGFP derivatives, allowing the transient expression of the Aequorea victoria EGFP fused to various PB1-F2 ORFs. The ORFs encoded PB1-F2 polypeptides of 57, 79 and 90 aa, respectively. Such ORFs are frequently observed in FLUAV isolates. Upon transfection with the pHygEGFP derivatives, cells were incubated at 37 °C for 24 h. Then, cells were treated with JMS-28j and JMS-1, respectively, which stains intact mitochondria. The fluorescent signals were analysed using confocal microscopy.
Due to the low transfection efficiency of HeLa cells, the fraction of EGFP-tagged PB1-F2-expressing cells was below 10 %. However, excitation at 488 nm revealed a mitochondrial localization of the EGFP fusion protein with full-length (90 aa) PB1-F2 of A/swine/Potsdam/15/81 (Fig. 3A–C) and a C-terminally truncated 79 aa protein of A/swine/Bakum/1832/00 (Fig. 3D–F). As a control, excitation at 543 nm was performed to demonstrate staining of intact mitochondria with JMS-28j and JMS-1, respectively. In accordance with previous findings (Chen et al., 2001), all HeLa cells expressing EGFP-tagged porcine full-length PB1-F2 showed accumulation of this polypeptide in mitochondria (Fig. 3A, B). In the overlay, co-localization of EGFP-tagged PB1-F2 and dyes JMS-28j or JMS-1 is indicated by yellow spots. In those cells, however, numerous mitochondria do not contain the fusion protein. In general, a change in morphology is observed for mitochondria affected by functional PB1-F2 constructs. While normal mitochondria are thin and thread-like, PB1-F2-expressing mitochondria have a round shape, appear to be swollen and many lose the ability to accumulate dye JMS-1. This indicates a loss of normal mitochondrial function. Fig. 3(B) shows cells with a low level of full-length PB1-F2 expression. Fig. 3(C, E, F) demonstrates cells with either high expression levels of the respective constructs or at a later stage. In these cells, the number of mitochondria showing co-localization of PB1-F2 and JMS-1 is reduced. A further C-terminal truncation of the PB1-F2 ORF (57 aa) resulted in diffuse cytoplasmic distribution of the fusion protein (Fig. 3G, H) which is indistinguishable from the control (unfused EGFP; Fig. 3I, J). However, the 57 aa polypeptide fused to GFP appears to be stable enough to accumulate in the cytoplasm. Occasionally, mitochondrial accumulation of the fusion protein was observed despite truncation of the complete MTS (Fig. 3H).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). However, a detailed analysis resulted in two surprising findings: (i) since 1950 all circulating human H1N1 isolates have a truncated PB1-F2 (57 aa), and (ii) many porcine isolates and most avian isolates (96 %) express an intact PB1-F2.
Eighty-nine of 119 sequenced porcine FLUAV isolates (75 %) express a PB1-F2 polypeptide with a size ranging from 79 to 90 residues (Table 1). While PB1-F2 of classic swine H1N1 viruses is truncated, many European porcine FLUAV (H1N1, H1N2, H2N3) and all American and Asian porcine H1N2 and H3N2 isolates express the full-length polypeptide. To prove the presence of the postulated polypeptide in virus-infected cells, PB1-F2 of a German porcine isolate was demonstrated by immunohistochemistry (Fig. 2). In previous experiments, the MTS of PB1-F2 was located to amino acids 46–75 (Yamada et al., 2004), while Gibbs et al. (2003) defined the minimal MTS ranging from amino acid 65 to 82 and the optimal MTS from 65 to 87. In transfected HeLa cells, mitochondrial co-localization of EGFP-tagged PB1-F2 of porcine FLUAVs (90 and 79 aa) was demonstrated by staining with dyes JMS-28j and JMS-1, indicating that PB1-F2 polypeptides with a C-terminal truncation up to amino acid 79 are functional. This finding fits well with the previous observation that the positively charged residues K73 and R75 are essential for mitochondrial targeting (Yamada et al., 2004). A further truncation (57 aa) usually resulted in a diffuse distribution of the fusion protein in the cytoplasm and the nucleus (Fig. 3G, H). In general, PB1-F2 expression results in a loss of normal mitochondrial function, since not all intact mitochondria (as indicated by uptake of 
m-sensitive dyes) simultaneously show PB1-F2-EGFP staining (Fig. 3A, B, D). This observation is in accordance with previous reports describing mitochondrial dysfunction in PB1-F2-expressing cells (Gibbs et al., 2003; Yamada et al., 2004; Zamarin et al., 2005). Apparently, the PB1-F2 properties of porcine FLUAV isolates do not differ from those of human FLUAV isolates. Although the details of PB1-F2 function are unclear, it is thought to bind to the mitochondrial proteins ANT3 and VDAC1, resulting in mitochondrial damage and subsequent apoptosis (Gibbs et al., 2003; Yamada et al., 2004; Zamarin et al., 2005; see also Fig. 3).
PB1-F2 was found in 80 FLUAV subtypes. Analysis of human H1N1 and European porcine FLUAV isolates showed numerous sequences with truncated PB1-F2 (Table 4). Since 216 human H1N1 sequences and 43 European porcine FLUAV sequences were analysed, this observation may be considered as significant. Other subtypes like H6N6 and H13N2 also showed large fractions of sequences with a truncated PB1-F2. However, in those cases only a rather small number of sequences are available. While recent human H1N1 isolates have a PB1-F2 truncation which first appeared between 1947 and 1950 (57 aa), the classic swine PB1-F2 polypeptides have accumulated various stop codons at amino acid positions 12, 26 and 35. This observation is based on 10 sequenced isolates. Although European porcine FLUAVs differ genetically from classic swine FLUAVs, their PB1-F2 sequences also terminate at codons 12, 26 and 35, which are highly conserved among FLUAV sequences.
Analysis of the PB1-F2-encoding nucleotide sequence revealed approximately 40 sequence positions that would allow the introduction of a stop codon upon a single nucleotide exchange. Only 14 sites were observed in available sequences and only two nonsense mutations of the PB1-F2 gene led to amino acid substitutions in PB1 (M40I, T58I). Hence, this observation and other molecular genetic sequence analyses (Tables 2 and 3) suggest that PB1-F2 is under constraint of the PB1 ORF. According to Kimura's neutral theory of molecular evolution, values less than 1 are indicative of negative selection (Kimura, 1983). The mean value of the PB1 gene is far below the conservative cut-off of 0.5, and the Z-test for neutrality supports purifying selection but rejects positive selection for all pairwise comparisons. For PB1-F2, the Z-test favours positive selection of most pairwise comparisons and rejects negative selection. To appraise the Z-tests appropriately, one has to consider that in overlapping gene regions, synonymous third base substitutions in the PB1 ORF correspond to non-synonymous exchanges at the second codon positions in the PB1-F2 ORF. The non-synonymous substitutions provoke diversification of PB1-F2 and truncation. The sliding window method which allowed calculation of diversity across the whole of the segment 2 sequence revealed no striking differences in . Therefore, it is concluded that negative selection of the PB1 ORF rather than positive selection of the +1 ORF is the reason for the observed variation in PB1-F2. The other neutrality tests did not yield clear results and have to be interpreted with caution. The McDonald–Kreitman test reveals an excess of synonymous nucleotide exchanges for the comparison of some, but not all PB1 datasets, and fails to detect positive selection for PB1-F2. Also, the neutrality tests based on DNA polymorphism (Tajima test, Fu & Li tests) did not yield significant results.
A striking difference with respect to the value of PB1-F2 estimated in this study and in a previous publication was observed. While Obenauer et al. (2006) determined to be 33.3, the values of the seven datasets in the present study ranged from 7.8 to 18.2. Two reasons may account for such a discrepancy. (i) Obenauer and colleagues used PAML software (Yang, 1997) which is based on maximum-likelihood methods, while in the present study an approach based on a parsimony analysis was employed (Nei & Gojobori, 1986). (ii) Another reason could be the sample selection of both studies. Obenauer et al. (2006) computed on the basis of 284 avian FLUAV segment 2 sequences, while in the present study datasets consisting of up to 250 virus sequences from different hosts were used.
Evidence of expression and mitochondrial localization of porcine FLUAV PB1-F2 suggests a similar function for the porcine und human PB1-F2 polypeptides. Induction of apoptosis in influenza virus-infected cells is well established and NS1 is an important pro-apoptotic protein. It induces expression of interferon 
/
genes and inhibits the activation of NF-B (for a recent review, see Ludwig et al., 2006). Assuming a complex interplay leading to virus replication, activation of the apoptotic cascade and induction of the antiviral effect against influenza virus, the role of the pro-apoptotic PB1-F2 polypeptide in the viral life cycle remains to be elucidated.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Castrucci, M. R., Donatelli, I., Sidoli, L., Barigazzi, G., Kawaoka, Y. & Webster, R. G. (1993). Genetic reassortment between avian and human influenza A viruses in Italian pigs. Virology 193, 503–506.[CrossRef][Medline]
Castrucci, M. R., Campitelli, L., Ruggieri, A., Barigazzi, G., Sidoli, L., Daniels, R., Oxford, J. S. & Donatelli, I. (1994). Antigenic and sequence analysis of H3 influenza virus haemagglutinins from pigs in Italy. J Gen Virol 75, 371–379.
Chen, W., Calvo, P. A., Malide, D., Gibbs, J., Schubert, U., Bacik, I., Basta, S., O'Neill, R., Schickli, J. & other authors (2001). A novel influenza A virus mitochondrial protein that induces cell death. Nat Med 7, 1306–1312.[CrossRef][Medline]
Chen, G. W., Yang, C. C., Tsao, K. C., Huang, C. G., Lee, L. A., Yang, W. Z., Huang, Y. L., Lin, T. Y. & Shih, S. R. (2004). Influenza A virus PB1-F2 gene in recent Taiwanese isolates. Emerg Infect Dis 10, 630–636.[Medline]
Fu, Y. X. & Li, W. H. (1993). Statistical tests of neutrality of mutations. Genetics 133, 693–709.[Abstract]
Gibbs, J. S., Malide, D., Hornung, F., Bennink, J. R. & Yewdell, J. W. (2003). The influenza A virus PB1-F2 protein targets the inner mitochondrial membrane via a predicted basic amphipathic helix that disrupts mitochondrial function. J Virol 77, 7214–7224.
Gregory, V., Lim, W., Cameron, K., Bennett, M., Klimov, A., Hall, H., Cox, N., Hay, A. J. & Lin, Y. (2001). Infection of a child in Hong Kong by an influenza A H3N2 virus closely related to viruses circulating in European pigs. J Gen Virol 82, 1397–1406.
Guan, Y., Shortridge, K. F., Krauss, S., Li, P. H., Kawaoka, Y. & Webster, R. G. (1996). Emergence of avian H1N1 influenza viruses in pigs in China. J Virol 70, 8041–8046.[Abstract]
Karasin, A. I., Carman, S. & Olsen, C. W. (2006). Identification of human H1N2 and human–swine reassortant H1N2 and H1N1 influenza A viruses among pigs in Ontario, Canada (2003 to 2005). J Clin Microbiol 44, 1123–1126.
Kimura, M. (1983). Neutral Theory of Molecular Evolution. Cambridge, MA: Cambridge University Press.
Krieg, R. (2006). Entwicklung, Struktur-/Eigenschaftsoptimierung und Anwendung neuartiger multifunktioneller Chromogene und Fluorochrome für den histochemischen Nachweis von peroxidatischer Aktivität und die in vivo Markierung zellulärer Organellen. Eine transmissionsmikroskopische, fluoreszenzoptische und ein-/multiphotonenlasermikroskopische Untersuchung. Habilitationsschrift, Medizinische Fakultät der Friedrich-Schiller-Universität Jena, Germany.
Kumar, S., Tamura, K. & Nei, M. (2004). MEGA3: integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Brief Bioinform 5, 150–163.
Lamb, R. A. & Krug, R. M. (1996). Orthomyxoviridae: the viruses and their replication. In Fields Virology, 3rd edn, pp. 1353–1395. Edited by B. N. Fields, D. M. Knipe & P. M. Howley. Philadelphia: Lippincott–Raven Publishers.
Ludwig, S., Pleschka, S., Planz, O. & Wolff, T. (2006). Ringing the alarm bells: signalling and apoptosis in influenza virus infected cells. Cell Microbiol 8, 375–386.[CrossRef][Medline]
Marozin, S., Gregory, V., Cameron, K., Bennett, M., Valette, M., Aymard, M., Foni, E., Barigazzi, G., Lin, Y. & Hay, A. (2002). Antigenic and genetic diversity among swine influenza A H1N1 and H1N2 viruses in Europe. J Gen Virol 83, 735–745.
McDonald, J. H. & Kreitman, M. (1991). Adaptive protein evolution at the Adh locus in Drosophila. Nature 351, 652–654.[CrossRef][Medline]
Miyata, T. & Yasunaga, T. (1980). Molecular evolution of mRNA: a method for estimating evolutionary rates of synonymous and amino acid substitutions from homologous nucleotide sequences and its application. J Mol Evol 16, 23–36.[CrossRef][Medline]
Nei, M. & Gojobori, T. (1986). Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol Biol Evol 3, 418–426.[Abstract]
Obenauer, J. C., Denson, J., Mehta, P. K., Su, X., Mukatira, S., Finkelstein, D. B., Xu, X., Wang, J., Ma, J. & other authors (2006). Large-scale sequence analysis of avian influenza isolates. Science 311, 1576–1580.
Ross, M. F., Kelso, G. F., Blaikie, F. H., James, A. M., Cocheme, H. M., Filipovska, A., Da Ros, T., Hurd, T. R., Smith, R. A. & Murphy, M. P. (2005). Lipophilic triphenylphosphonium cations as tools in mitochondrial bioenergetics and free radical biology. Biochemistry (Mosc) 70, 222–230.[CrossRef][Medline]
Rozas, J., Sanches-DelBarrio, J. C., Messeguer, X. & Rozas, R. (2003). DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19, 2496–2497.
Schmidtke, M., Zell, R., Bauer, K., Krumbholz, A., Schrader, C., Suess, J. & Wutzler, P. (2006). Amantadine resisitance among porcine H1N1, H1N2, and H3N2 influenza A viruses isolated in Germany between 1981 and 2001. Intervirology 49, 286–293.[CrossRef][Medline]
Schrader, C. & Süss, J. (2003). Genetic characterization of a porcine H1N2 influenza virus strain isolated in Germany. Intervirology 46, 66–70.[CrossRef][Medline]
Schrader, C. & Süss, J. (2004). Molecular epidemiology of porcine H3N2 influenza A viruses isolated in Germany between 1982 and 2001. Intervirology 47, 72–77.[CrossRef][Medline]
Tajima, F. (1989). Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123, 585–595.
Yamada, H., Chounan, R., Higashi, Y., Kurihara, N. & Kido, H. (2004). Mitochondrial targeting sequence of the influenza A virus PB1-F2 protein and its function in mitochondria. FEBS Lett 578, 331–336.[CrossRef][Medline]
Yang, Z. (1997). PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci 13, 555–556.
Zamarin, D., Garcia-Sastre, A., Xiao, X., Wang, R. & Palese, P. (2005). Influenza virus PB1-F2 protein induces cell death through mitochondrial ANT3 and VDAC1. PLoS Pathog 1. doi:10.1371/journal.ppat.0010004
Received 12 July 2006;
accepted 19 October 2006.
This article has been cited by other articles:
|
|
 |
|
  J. R. Powe and W. L. Castleman Canine Influenza Virus Replicates in Alveolar Macrophages and Induces TNF-{alpha} Vet. Pathol., November 1, 2009; 46(6): 1187 - 1196. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Ma, Y. Feng, D. Liu, and G. F. Gao Avian influenza virus, Streptococcus suis serotype 2, severe acute respiratory syndrome-coronavirus and beyond: molecular epidemiology, ecology and the situation in China Phil Trans R Soc B, September 27, 2009; 364(1530): 2725 - 2737. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. M. Wise, A. Foeglein, J. Sun, R. M. Dalton, S. Patel, W. Howard, E. C. Anderson, W. S. Barclay, and P. Digard A Complicated Message: Identification of a Novel PB1-Related Protein Translated from Influenza A Virus Segment 2 mRNA J. Virol., August 15, 2009; 83(16): 8021 - 8031. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | MICROBIOLOGY | J GEN VIROL |
| J MED MICROBIOL | ALL SGM JOURNALS | |
