Nucleotide Sequence Requirements at the 5′ End of the Influenza A Virus M RNA Segment for Efficient Virus Replication

Abstract

The mechanism by which the influenza A virus genome is packaged into virions is not fully understood. The coding and noncoding regions necessary for packaging of the viral RNA segments, except for the M segment, have been identified. Here, we delineate the M segment regions by incorporating a reporter viral RNA into virions and by generating viruses possessing mutations in the regions. We found that, like the other segments, the M segment coding regions are essential for virion incorporation and that the nucleotide length rather than the nucleotide sequence of the 5′ end of the coding region is important.

The genome of influenza A virus consists of eight negative-stranded RNA segments (20). For efficient virus replication, all eight viral RNA (vRNA) segments must be packaged into progeny viruses. Recently, for all segments except the M segment, we and other groups demonstrated that the coding sequences (1-5, 11, 14-16, 19, 21) as well as the noncoding regions (13) at both the 3′ and 5′ ends of the vRNA segments are required for their efficient packaging into virions. For the M segment, Hutchinson et al. (7) demonstrated that synonymous mutations introduced into the highly conserved codons located in the terminal coding regions lead to defects in virion assembly and viral genome packaging. However, the exact nucleotide (nt) sequences of the M vRNA coding region critical for genome packaging remain unknown. Here, we defined the packaging signal for the M segment by using virus-like particles (VLPs) that possessed green fluorescent protein (GFP)-encoding M vRNAs and mutant viruses whose M vRNAs encoded M2 protein in which the carboxy (C)-terminal 11 amino acids were deleted, allowing us to modify this region without affecting the amino acid sequence.

To identify the packaging signal of the M vRNA, we generated infectious VLPs by A/WSN/33 (H1N1; WSN) virus-based reverse genetics (17) using plasmids for mutant M vRNAs that possessed the GFP gene flanked by various lengths of the coding and/or noncoding regions (Fig. 1). The packaging efficiency of each test M vRNA into infectious VLPs was measured as previously described (19) (Fig. 1). Briefly, we infected Madin-Darby canine kidney (MDCK) cells with the VLPs, detected virus-infected cells with an anti-WSN virus polyclonal antibody, and counted the number of WSN viral antigen-expressing (representing the total number of infectious VLPs) and GFP-expressing (representing the number of VLPs that possessed the test M vRNA) cells at 24 h postinfection. As the 3′ or 5′ coding region was extended, both the M vRNA packaging efficiency and the VLP number gradually increased except for those of M(0)GFP(0) (the numbers in parentheses indicate the nucleotide numbers for the M coding regions). These results indicate that, like the other segments (2-4, 10, 14, 16, 19, 21), both ends of the M vRNA coding region support efficient packaging of the vRNA segment into virions.

FIG. 1.

Identification of domains in the 3′ or the 5′ coding region of the M vRNA important for genome packaging and virion formation. (A) Schematic diagrams of test M vRNAs. The numbers in parentheses indicate the nucleotide numbers for the M coding regions. The noncoding and coding regions are represented by gray and green bars, respectively, while the dashed lines indicate nucleotides deleted from the M coding regions. The GFP open reading frame (green bar) was inserted in frame into the M open reading frame. All mutants are shown in the negative-sense orientation. (B) Efficiencies of genome packaging and virion formation. To generate WSN strain-based infectious VLPs possessing the test M vRNAs given in panel A, plasmids for each test M vRNA were transfected into 293T cells with seven plasmids for the remaining vRNAs and six plasmids for the expression of PB2, PB1, PA, NP, M1, and M2. Forty-eight hours posttransfection, the supernatants were transferred to MDCK cells. The numbers of WSN viral antigen-expressing (representing the total number of infectious VLPs) and GFP-expressing (representing the number of VLPs that possessed the test M vRNA) cells at 24 h postinfection were determined, and the packaging efficiency of each test M vRNA into VLP was calculated by dividing the number of M vRNA-positive VLPs by the total number of VLPs. The numbers of VLPs with and without test M vRNAs are shown by green and gray bars, respectively. The results shown are representative of three independent experiments.

To further define the nucleotide sequences of the M vRNA coding region critical for virus replication, we focused on the coding sequence at the 5′ end. Previously, we have shown that an 11-amino-acid deletion at the M2 C terminus (Mstop) (Fig. 2A) does not severely affect virus replication in cell culture (8). The Mstop vRNA harbors a 29-nt sequence (at positions 979 to 1007) that normally encodes part of the M2 C terminus but in this mutant becomes noncoding due to the double stop codons introduced immediately upstream of the 29 nt. We exploited this 29-nt sequence at the 5′ end of the Mstop vRNA to further define the requirements in this region for genome packaging without further altering the amino acid sequence of the virus.

FIG. 2.

Construction of mutant M vRNAs and their packaging efficiencies into virions. (A) Schematic diagrams of wild-type and mutant M vRNAs. The nucleotide sequences at the 5′ end (positions 972 to 1013) of wild-type, Mstop, Mrdm, Mstop-cc, M-27U, Mdel, and Mscr-1, -2, and -3 vRNAs are shown. For the construction of a plasmid pool for Mrdm vRNA, we performed inverse PCR with plasmid for Mdel vRNA as a template and primers containing randomized nucleotides at positions 979 to 1007. The self-ligated PCR products were used to transform Escherichia coli DH5α, resulting in 5.0 × 10³ transformants. The InvivoGen small interfering RNA Wizard (http://www.sirnawizard.com/scrambled.php) was used to create the randomized sequences of the Mscr-1, -2, and -3 vRNAs. The noncoding and coding regions are represented in gray and black, respectively. Stop codons and the altered nucleotides from the authentic sequence are shown in red and blue, respectively. Note that the amino acid sequences of M2 encoded on the listed M vRNAs are identical except for the wild-type M vRNA. (B) Transcription/replication efficiencies of wild-type and mutant M vRNAs. An RNA polymerase I-driven plasmid for M vRNA synthesis was transfected into 293T cells together with four plasmids for the expression of PB2, PB1, PA, and NP, which are necessary and sufficient for vRNA transcription and replication. Twenty-four hours posttransfection, the expression of M1, M2, and β-actin (as an internal control) in the cell lysates was analyzed by Western blotting with monoclonal antibodies specific for each protein. The results shown are representative of three independent experiments.

To investigate whether a specific nucleotide(s) of the 29-nt sequence in the Mstop vRNA plays a critical role in virus replication, a plasmid library for mutant M vRNAs that possessed randomized 29-nt sequences (Mrdm) (Fig. 2A) was constructed and used for reverse genetics. The transfectant supernatant was subjected to plaque assays on MDCK cells. Although the pool size of the Mrdm vRNA (1.5 × 10³; see legend to Fig. 2 for details) was much smaller than the total possible sequence variation (4²⁹ = 10¹¹), this approach allowed us to find any sequence trend(s) in this region that favored efficient virus replication.

We analyzed 38 plaque-purified viruses and found 25 different nucleotide sequences of M vRNAs (Table 1). None of these 25 sequences were identical to the sequence found in the original Mstop vRNA. All of the viruses possessing these different nucleotide sequences in their 29-nt sequence grew efficiently (10⁷ to 10⁸ PFU/ml) and maintained the sequences after three passages, with the exception of clone number 13. In this clone, we found a uracil inserted at the 5′ end of the randomized nucleotides in the M vRNA after three passages (data not shown). No other mutations were found in these mutant M vRNAs. The ratios of adenine and uracil were much higher than those of guanine and cytosine in these mutant M vRNAs; however, this AT-rich tendency was likely an artifact, since a similar nucleotide bias was also found in the plasmid pool used to generate the Mrdm vRNA (Table 2).

TABLE 1.

Nucleotide sequences of viruses possessing M vRNAs derived from a pool of mutant M vRNA segments^a

Mstop	Nucleotide sequence at position 972-1007 (3′ to 5′)	No. of samplesb
	ACUCACUCAGUAAAACAGUUGUAUCUCGACCUCAUU
1	ACUCACUACCCUCUUUUUCCCGAUGUGCUUUUUAUU	1
2	ACUCACUACUAGAUUACUAACUAAAUUAAAGCUAUU	2
3	ACUCACUUUUACAAUUUUUUAUCACAAAUUUCACUU	1
4	ACUCACUUAACACUUUCAACCAUGACUAAGCUUAUU	4
5	ACUCACUCACUAACAUUAUAAUUACUCCAGACUAUU	2
6	ACUCACUCCUUAAUUACCAAAAAUGCAAUUUGUAUU	5
7	ACUCACUCAAUGGCUGCAAUACUUUUCAUCACCUUU	1
8	ACUCACUAUACAAAAUCCUUUCUCUGAUUAAGUAUU	1
9	ACUCACUACGCAUUAUAUCAUAUCAUCGUAAACAUU	2
10	ACUCACUUAAUACAUUAGAUCUUUUUCGAGUACAUU	1
11	ACUCACUACCUAUUAUUACAUAAUAUUUUGCAUAUU	2
12	ACUCACUAAUCCACAUCACUACAUUCCUCUUUCCUU	1
13	ACUCACUUCUUAAUCAUAAUAUUUACUAGUCAACAU	1
14	ACUCACUUAAUUUAUAUAUAUUUAAUAUCUAGUAUU	2
15	ACUCACUUUCACAAUAUUUUAUACACUCAAUACAUU	1
16	ACUCACUACCCAAUUUUCAUCUAUCUGUCUAGUGUU	1
17	ACUCACUCAAUUUAAAUUUAGGUUCCUUUUUGUCUU	1
18	ACUCACUCCGAUUUAAUCAAUUUCGAAUGCUAGAUU	1
19	ACUCACUAUUAUAUAUUAAACCCUCUACAGACUCUU	1
20	ACUCACUUAUAAACCUAAUUAAUCUACUAGAUUAUU	2
21	ACUCACUUUUACAGUAGCAUUUUUCCUCACACCAUU	1
22	ACUCACUAACUUUUAUGUACGUCAUUACAAAUUAUU	1
23	ACUCACUCCAAUCAAGACCUUUAAAACACACCACUU	1
24	ACUCACUGCAUCAAGACGUCAACACUAGUUAACAUU	1
25	ACUCACUACAUGAUUAGUAAUGAAAGAUUAUACCUU	1

^aReverse-transcription PCR products from mutant viruses were sequenced directly. Nucleotides identical to the wild-type are roman, stop codons are underlined, and the conserved uracils at the 5′ end (see text for details) are bold.

^bNumber of clones possessing the indicated sequence.

TABLE 2.

Nucleotide biases found at nucleotide positions 979 to 1007 in plasmids and in vRNAs from rescued viruses

Nucleotide	% of nucleotides for:
	Randomized pool (at plasmid level)a	Selected clones (at vRNA level)b	Wild typeb
A	33.4	34.6	31
U	38.5	39.8	31
G	9	6.4	13.8
C	19	19.2	24.1

^aEscherichia coli DH5α was transformed with a plasmid pool for Mrdm vRNA. Plasmids for Mrdm vRNAs were recovered from 22 randomly selected colonies to determine the nucleotide sequences. The percentages of nucleotides corresponding to the nucleotide sequences at the 5′ end (positions 979 to 1007) of the M vRNAs in the selected plasmids are shown.

^bThe percentages of nucleotides at positions 979 to 1007 of M vRNAs derived from 38 plaque-purified and wild-type viruses are shown.

Analyses of the 25 sequences found in the 29-nt regions of the mutant M vRNA segments neither identified consensus sequences nor predicted secondary structures, except for two uracils at the 5′ end (Table 1). The 5′ end of influenza vRNAs contains a series of uracils, the so-called poly(U) signal sequence (9, 12). This nucleotide sequence, which allows the addition of the poly(A) tail at the 3′ end of the viral mRNAs, plays a role in vRNA transcription (9, 12). For WSN M vRNA, which we used as a template in this study, the poly(U) sequence comprises two uracils in stop codons and an additional four uracils (Fig. 2A). The two uracils at the 5′ end of the M vRNAs conserved among the viral clones obtained from the randomized pool suggest that a series of six uracils are needed for the efficient transcription of M vRNA for virus replication. To test this possibility, we prepared a plasmid for Mstop-cc vRNA (Fig. 2A) in which two cytosines replaced the two conserved uracils at the 5′ end of the coding sequence. To assess the transcription/replication efficiency of the Mstop-cc vRNA, the plasmid was transfected into 293T cells together with four plasmids for the expression of PB2, PB1, PA, and NP, which are necessary and sufficient for vRNA transcription and replication (18). Twenty-four hours posttransfection, the expression of M1 and M2 in the cell lysates was analyzed by Western blotting with monoclonal antibodies specific for each protein (Fig. 2B). Under these conditions, the protein expression of the Mstop-cc vRNA was much lower than that of the wild-type M and Mstop vRNAs, demonstrating that these two conserved uracils likely support the efficient transcription/replication of the M vRNAs for virus replication.

Our finding of only limited conservation of the sequence in the 29-nt region in the M vRNA among our viral clones indicated that the nucleotide sequence in this region can be highly diverse without disrupting efficient virus replication, except for the two uracils at the 5′ end. To confirm the relaxed stringency of the nucleotide sequence at the 5′ end of the Mstop vRNA for efficient virus replication, we designed a mutant vRNA containing uracils at all positions from 979 to 1005 (Fig. 2A). However, the transcription/replication efficiency of this mutant vRNA was quite low (Fig. 2B).

We therefore designed four M vRNA mutants whose noncoding 27-nt regions for the M2 C terminus (except for the last two uracils) were deleted (Fig. 2A) or replaced with randomized sequences (Fig. 2A). The nucleotide compositions of the 27-nt sequence regions of these latter three M vRNAs were the same as those of the Mstop vRNA, but the nucleotide sequences were randomized. The transcription/replication efficiencies of these mutant M vRNAs were comparable to those of the wild-type M and Mstop vRNAs (Fig. 2B). While infectious viruses possessing the Mscr-1, -2, and -3 and Mstop vRNAs were readily generated by reverse genetics, we could not rescue infectious virus that possessed the Mdel vRNA. Although all of the viruses that possessed the scrambled sequences in their M vRNAs grew more slowly than the Mstop virus, their maximum titers were comparable to those of the Mstop virus (Fig. 3A). These results indicate that the nucleotide length, rather than the nucleotide sequence, at the 5′ end of the Mstop vRNA is critical for virus replication.

FIG. 3.

Characterization of mutant viruses possessing a scrambled sequence in their M vRNAs. (A) Growth comparison of Mstop and Mscr-1, -2, and -3 viruses. MDCK cells were infected with each virus at a multiplicity of infection of 0.01. The culture supernatants harvested at the indicated times postinfection were subjected to plaque assay on MDCK cells. Error bars indicate the standard deviations of three independent experiments. (B) Packaging efficiencies of HA, NP, and M vRNAs into wild-type, Mstop, and Mscr-1, -2, and -3 viruses. MDCK cells were infected with the indicated viruses at a low (<0.1) multiplicity of infection. Twelve hours postinfection, the cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min and M1-, NP-, or HA-expressing cells were identified with anti-M1, -NP, or -HA monoclonal antibodies; virus-infected cells were identified with an anti-WSN virus polyclonal antibody. The packaging efficiency of each vRNA was determined by dividing the numbers of monoclonal antibody-positive cells (representing the number of infectious virions that possessed M, NP, or HA vRNA, respectively) by the number of polyclonal antibody-positive cells (representing the number of all infectious virions). Error bars indicate the standard deviations of three independent experiments. (C) Formation efficiency of infectious virions possessing wild-type, Mstop, Mdel, and Mscr-1, -2, and -3 viruses. The indicated viruses were generated in 293T cells by reverse genetics using a plasmid for the NA vRNA that possesses a mutation from lysine to arginine at the C terminus. MDCK cells were infected with the transfectant viruses and fixed with 4% paraformaldehyde in PBS, and virus-infected cells were identified with an antiserum to WSN virus. Error bars indicate the standard deviations of three independent experiments.

To determine the packaging efficiency of each M vRNA into virions, MDCK cells infected with these mutant viruses at a low (<0.1) multiplicity of infection were fixed at 12 h postinfection and M1-, NP-, or hemagglutinin (HA)-expressing cells were identified with anti-M1, -NP, or -HA monoclonal antibodies, respectively; virus-infected cells were identified with anti-WSN virus polyclonal antibody. The packaging efficiencies were calculated by dividing the number of monoclonal antibody-positive cells by the number of polyclonal antibody-positive cells (Fig. 3B). The packaging efficiencies of the NP and HA vRNAs into all viruses were greater than 90%. On the other hand, those of the M vRNAs into the Mscr-1, -2, and -3 viruses (72%, P = 0.03; 79%, P = 0.17; and 79%, P = 0.37, respectively) were lower than that into the Mstop virus (82%); thus, a statistically significant difference was observed only with the Mscr-1 virus. These results indicate that the nucleotide sequence scrambling somewhat reduced the packaging efficiency of the M vRNA into infectious virions.

To elucidate the effect of the nucleotide sequence scrambling on infectious virion formation, we generated, by reverse genetics, Mscr-1, -2, and -3 viruses whose NA vRNA encoded arginine at the carboxyl-terminal position 453 instead of lysine. This mutation prevents trypsin-independent cleavage of WSN virus HA via the interaction between NA and plasminogen (6) and multiple cycles of virus replication in plasmid-transfected cells. To determine the efficiency of infectious virion formation, we titrated the transfectant viruses in the supernatant of the plasmid-transfected cells (Fig. 3C). The Mscr-1, -2, and -3 virus titers were more than 10-fold lower than those of the wild-type WSN and Mstop viruses, indicating that the nucleotide sequence scrambling reduced the efficiency of infectious virion formation.

In summary, here we demonstrated that both ends of the coding regions of the M vRNA are required for its efficient packaging into virions as has been shown for the other seven vRNAs (2-4, 10, 14, 16, 19, 21). We also found that the 27-nt sequence at the 5′ end of the coding region in the M vRNA can accommodate highly diverse sequences compromising virus replication to only a limited extent and that the length, rather than the sequence, of nucleotides in this region is important for virus replication.

Synonymous mutations introduced into the highly conserved M2 codons at positions 90 to 92 (nucleotide positions 981 to 989 of M vRNA) (7) caused a more severe reduction in virus titer in MDCK cells compared with that found with our Mscr-1, -2, and -3 viruses. However, the results of both studies indicate that the nucleotide sequence of the 5′ end of the coding region in the M vRNA, but not the amino acid sequence of the corresponding gene product, affects virus replication efficiency. These findings further our understanding of the packaging mechanism of influenza A virus vRNA segments into virions.