Acetylation, Methylation and Allysine Modification Profile of Viral and Host Proteins during Influenza A Virus Infection

Protein modifications dynamically occur and regulate biological processes in all organisms. Towards understanding the significance of protein modifications in influenza virus infection, we performed a global mass spectrometry screen followed by bioinformatics analyses of acetylation, methylation and allysine modification in human lung epithelial cells in response to influenza A virus infection. We discovered 8 out of 10 major viral proteins and 245 out of 2280 host proteins detected to be differentially modified by three modifications in infected cells. Some of the identified proteins were modified on multiple amino acids residues and by more than one modification; the latter occurred either on different or same residues. Most of the modified residues in viral proteins were conserved across >40 subtypes of influenza A virus, and influenza B or C viruses and located on the protein surface. Importantly, many of those residues have already been determined to be critical for the influenza A virus. Similarly, many modified residues in host proteins were conserved across influenza A virus hosts like humans, birds, and pigs. Finally, host proteins undergoing the three modifications clustered in common functional networks of metabolic, cytoskeletal, and RNA processes, all of which are known to be exploited by the influenza A virus.


Introduction
The influenza virus is a globally prevalent human respiratory pathogen, with recurring epidemic and pandemic potentials. Practically, the influenza A virus (IAV) will be impossible to eradicate because of its intrinsic evolving nature, zoonotic potential, and broad host range which includes migratory birds. Hence, there will be a constant need to develop new vaccines and antiviral drugs. Like all viruses, IAV is an obligate intracellular pathogen and exploits the host machinery to multiply and cause flu [1]. IAV utilises all three parts: the plasma membrane, cytoplasm, and nucleus, of the host cell to complete its life cycle. For this, IAV employs viral proteins and exploits host proteins to execute multiple crucial functions, such as anterograde and retrograde intracellular trafficking, RNA processing, and antagonism of host antiviral response. Many of these functions involve numerous protein-protein, protein-nucleic acid, and protein-lipid interactions. Evidently, these interactions are known to be regulated by multiple protein modifications, which occur co-or post-translationally and can be reversible and irreversible. Further, the dysregulation or imbalance of protein modifications is a hallmark of various human diseases [2,3]. Therefore, it is conceivable that the modifications of host proteins that are involved in IAV infection play an important role in their function. Furthermore, the magnitude of the modifications of such proteins is expected to change in response to IAV infection and, in addition to host proteins, IAV proteins also are expected to undergo those modifications.
Indeed, the modifications like phosphorylation, glycosylation, and ubiquitination have been described to differentially occur on both viral and host proteins and be significant that in addition to acetylation, both viral and host proteins are modified by different lysine modifications such as methylation and allysine during IAV infection.
To identify a repertoire of both viral and host proteins that are differentially modified by acetylation, methylation and allysine during IAV infection, we employed a global profiling approach by liquid chromatography-coupled tandem mass spectrometry (LC-MS/MS) followed by functional annotation of modified proteins using various bioinformatics tools. Here, we report that many viral and host proteins are indeed modified on multiple amino acid residues by three modifications during IAV infection. Furthermore, most of the modified residues are conserved and critical for the function of those viral and host proteins. Finally, many of the identified host proteins undergoing modifications are already known to play a critical role in IAV infection, indicating the significance of these modifications.

Cells, Viruses, and Infection
Human lung alveolar epithelial cells, A549 were grown and maintained in minimum essential medium (MEM) supplemented with 10% foetal bovine serum, 1% L-glutamine and 1% penicillin-streptomycin (Life Technologies, Carlsbad, CA, USA) under 5% CO 2 atmosphere and at 37 • C. Influenza virus A/Puerto Rico/8/1934 (H1N1) was propagated in 10 days old embryonated chicken eggs and titrated in Madin Darby canine kidney (MDCK) cells. For infection, confluent A549 cell monolayers, prewashed twice with serum-free MEM, were incubated with virus inoculum (multiplicity of infection of 1.0) in serum-free MEM for 1 h at 35 • C. The inoculum was removed, cells were washed once with serum-free MEM, replenished with fresh serum-free MEM, and incubated at 35 • C. After 24 h, the culture medium and the cells were harvested separately. The medium was titrated to measure the amount of released viral progeny and confirm a productive infection.

In-Gel Digestion
Cells were lysed in a lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.5% sodium dodecyl sulfate, and 1X protease inhibitor cocktail). The protein amount per sample was quantified using the Pierce™ BCA protein assay kit (ThermoFisher Scientific, San Jose, CA, USA). Equal amounts of protein were loaded and resolved on 15% SDS-PAGE along with SeeBluePlus 2 pre-stained molecular weight standards (Life Technologies, Carlsbad, CA, USA). Proteins were prefixed in-gel with 50% methanol and 10% glacial acetic acid in water, then stained with Coomassie Blue R-250 solution (Sigma-Aldrich, St Louis, MO, USA) for 1 h and de-stained overnight with 40% methanol and 10% glacial acetic acid in water. An image of the gel was acquired on an Odyssey Fc imaging system using Image Studio software version 5.0 (Li-COR, Lincoln, NE, USA). Each gel lane was cut into 10 molecular weight fractions and proteins were in-gel digested with trypsin using an automated digestion procedure (Intavis AG, Koln, Germany) described elsewhere [30]. Eluates containing the peptides were dried by centrifugal vacuum concentration.

Filter-Aided Sample Preparation (FASP)
Cells were lysed in a lysis buffer containing 40 mM Tris base, 1 mM EDTA, 0.5 M TCEP, 0.5% SDS, 0.5% sodium deoxycholate, 0.1% NP-40 and 1X protease inhibitor cocktail. The lysates were centrifuged at 16,000× g for 30 min and the supernatants were further processed by the FASP method described elsewhere [31]. The resulting protein amount was quantified by Bradford assay and digested with trypsin overnight. Peptides were dried and concentrated as above.

Mass Spectrometry Analysis
The dried peptides were resolubilized in 5% acetonitrile (ACN) and 0.1% formic acid (FA) in water and analysed using the Ultimate 3000 nano-flow uHPLC-System (Dionex Co, San Jose, CA, USA) in-line coupled to the nanospray source of LTQ-Orbitrap XL mass spectrometer (ThermoFisher Scientific, San Jose, CA, USA). The peptides were separated on in-house packed emitter tip columns (75 µm ID fused silica tubing filled with C-18 material to a length of 12 cm) at a flow rate of 400 nL/min. The liquid chromatography gradient gradually increased from 5% (v/v) ACN, 0.1% (v/v) FA to 45% ACN, 0.1% (v/v) FA and then to 95% (v/v) ACN, 0.1% (v/v) FA. Full mass spectra were acquired in the Orbitrap analyser in a mass range between m/z 400-2000 and a mass resolution of 60,000 at m/z 400. The five strongest signals were selected for collision-induced dissociation-MS2 acquisition in the LTQ ion trap at a normalised collision energy of 35%. At least two technical replicates of each of the biological replicates were analysed.

Data Analysis
Raw data were pooled from three independent biological replicates and was analysed using the Proteome Discoverer software (version 2.2, ThermoFisher Scientific, San Jose, CA, USA) and searched against a customised database containing human and influenza virus A/Puerto Rico/8/1934 (H1N1) reference amino acid sequences (55,836 entries downloaded on 23 June 2018 from the NCBI server, https://www.ncbi.nlm.nih.gov/) using the SEQUEST HT search engine (ThermoFisher Scientific, San Jose, CA, USA). The search was set up for full tryptic peptides with a maximum of 3 missed cleavages. The precursor mass tolerance threshold was fixed at 10 ppm with a maximum fragment mass error of 0.8 Da. Carbamidomethyl (C) was set as a static modification and acetylation (K, S), lysine > allysine (K), methylation (K, R) as dynamic peptide modifications. N-terminal acetylation with the dynamic loss of the N-terminal methionine was set as a dynamic protein modification. Proteins were considered identified only when at least two significant peptide hits were detected. Peptide hits were considered significant when passing the adjusted score threshold at a strict False Discovery Rate (FDR) of q < 0.01 as calculated by the percolator algorithm [32]. Protein modifications were considered only if a given modification in a peptide was identified in at least two biological replicates.

Bioinformatics Analysis
All protein sequences analysed here were extracted from NCBI databases and aligned by BioEdit (version 7.2.5) using ClustalW multiple sequence alignments. The conservation of amino acid residues was visualised using WebLogo (version 2.8.2). The 3-dimensional (3D) structure of viral proteins, acquired from the RCSB protein data bank (PDB) was visualized and analysed by PyMOL (version 2.3.4, Schrödinger LLC, New York, NY, USA). The acetylation site prediction was performed using ASEB web software (http: //bioinfo.bjmu.edu.cn/huac/predict_p/, accessed on 28 February 2020). The STRING database (version 11.0) was used to generate functional association networks of modified host proteins at a confidence score of ≥ 0.7 (high confidence). The resultant networks were further analysed by Cytoscape (version 3.8.0) for densely connected regions using the Molecular Complex Detection (MCODE) algorithm. The four highest-ranked clusters were selected as the most significant clusters.

Results
Human lung alveolar epithelial cells, A549 were either mock-infected or infected with influenza virus A/Puerto Rico/8/1934 (H1N1) strain at the multiplicity of infection (MOI) of 1.0. After 24 h, the culture medium and the cells were harvested separately. The medium was titrated to detect viral progeny and confirm productive infection (Supplementary Figure S1A). The cells were lysed, and the lysates were processed to prepare LC-MS/MS samples using either in-gel (Supplementary Figure S1B) or in-solution digestion procedures (Supplementary Figure S1C). The raw data obtained from three independent biological replicates were pooled and searched against a customised database comprising 55,836 influenza virus A/Puerto Rico/8/1934 (H1N1) and human reference sequences using Proteome Discoverer software and SEQUEST HT search engine with appropriate protein modification filters.

Eight Viral Proteins
Were Modified by Acetylation, Methylation and/or Allysine IAV genome encodes 10 major viral proteins and at least 7 accessory viral proteins [33]. We detected all 10 major IAV proteins: hemagglutinin (HA), matrix proteins (M1 and M2), nucleoprotein (NP), neuraminidase (NA), NS1 and NS2, and RNA polymerase subunits PA, PB1, and PB2 with a sequence coverage ranging from 89% to 30% (Supplementary Table S1). Out of these, eight proteins: HA, M1, NP, NS1, NS2, PA, PB1, and PB2 were found to be modified by acetylation, methylation or allysine (Table 1). In some viral proteins, individual modifications occurred on multiple amino acid residues. Furthermore, most of the viral proteins were modified by at least two modifications on different residues, but the same residues were also detected to be alternatively modified by different modifications (Table 1).
M1, which forms a shell under the envelope of IAV particles and plays critical roles during IAV assembly, was found to be modified by all three modifications, i.e., methylation, acetylation and allysine (Table 1). Methylation occurred on lysine (K) 95, K98, arginine (R) 160, K230, and K242; whereas acetylation was detected on serine (S) 195, S196, and S207 and K95; and allysine was found on K35, K98, and K230 (Table 1, in red). Interestingly, three residues: K95, K98 and K230 were found to be differentially modified by either methylation, acetylation or allysine (Table 1, asterisks). To understand the significance of these modifications, we aligned at least 800 M1 sequences from 39 different IAV subtypes, ranging from H1N1 to H13N6 (Supplementary Table S2), and found that, except S207, all identified modified residues in M1 were highly conserved across IAV subtypes ( Figure 1A). Furthermore, R160 was also conserved in M1 of all three main types of influenza viruses, i.e., IAV, influenza B virus (IBV) and influenza C virus (ICV); whereas K35 and K95 were conserved within M1 of IAV and IBV ( Figure 1B). Next, we determined the location of modified residues in the M1 structure, which comprises 12 α-helices and several loop regions but no β-strands [34,35]. The allysine K35 was located on the loop between α2 and α3 helices, whereas the acetylated S195 and S196 were located on the loop between α10 and α11 helices ( Figure 1C). The rest of the modified residues were located on different α-helices ( Figure 1C). Particularly, K95 and K98 were located on "helix six", which is critical for various M1 functions, including in IAV assembly and particle shape and integrity [36]. Finally, we used Acetylation Set Enrichment Based (ASEB) web software [37] to predict whether some of the identified acetylated sites in M1 can be acetylated or deacetylated by host lysine acetyltransferases (KAT) or deacetylases, respectively. The ASEB analysis revealed that acetylation at K95 can be deacetylated by sirtuin-1 (SIRT1) deacetylase ( Figure 1D). Red indicates modified amino acids; * indicates amino acids with more than one modification. NP, an RNA-binding protein that plays a critical role in IAV RNA transcription and replication, was found to be methylated and acetylated, predominantly on arginine and serine, respectively (Table 1). Methylation was detected on R150, R246, R317, K325, R416, and R422, whereas acetylation occurred on S274, S283, S287, S326, S403, and K325. Like in M1, one residue, K325 was found to be either methylated or acetylated. The alignment of 1,124 NP polypeptide sequences from 42 IAV subtypes (Supplementary Table S2) revealed that four of the methylated residues, R150, R246, R317, and R416, and three of the acetylated residues, S274, S287, and S326, as well as dual-modified K325, were highly conserved ( Figure 2A). Out of these, R150, R416, and K325 were also conserved within NP of IAV, IBV and ICV. Whereas R246, R317, S326, and S403 were conserved within the NP of IAV and IBV ( Figure 2B). Next, we located the modified residues in the NP structure, which is composed of multiple α-helices, β-strands and loop regions [38]. We found that methylated R150, R246, R317, and R416 and acetylated S274 and S403 were all located in the loop regions ( Figure 2C). The rest of the modified residues were situated on α-helices: S283 and S287 on α12 helix, K325 and S326 on α15 helix, and R422 on α18 helix ( Figure 2C). Evidently, R150, K325, and R416 are indispensable for IAV growth. The R150 and K325 are critical for NP to bind the viral RNA [39] and incorporate it into virus particles [40], respectively; whereas, R416 is important for NP to oligomerise [41]. In addition, R422 also contributes to the NP oligomerisation [42] as well as NP's interaction with host heterogeneous nuclear ribonucleoprotein (hnRNP) A2/B1, which is involved in IAV RNA synthesis [43]. Markedly, we also found several hnRNPs being modified in response to the IAV infection (described below).
PA, PB1, and PB2, the three IAV RNA polymerase subunits, were also discovered to be modified (Table 1). PA was methylated as well as acetylated on both K102 and K104, and acetylated on S631. In contrast, PB1 and PB2 were N-terminally acetylated and allysine (on K718), respectively. In PA, differentially modified K102 and K104 were highly conserved ( Figure 3A) across 800 sequences from 48 IAV subtypes (Supplementary Table S2), though these residues were not conserved within PA of IAV, IBV or ICV ( Figure 3B). Furthermore, ASEB analysis predicted that K102 can be acetylated by acetyltransferases GCN5/PCAF and K104 can be deacetylated by SIRT1 ( Figure 3C). In PB2, allysine K718 was also highly conserved ( Figure 3D) across 986 sequences from 48 IAV subtypes (Supplementary Table S2), but not within the PB2 of IAV, IBV or ICV ( Figure 3E). In the PA structure, K102 and K104 were located in the loop region of its critical N-terminal endonuclease domain ( Figure 3F), which also binds PB2 [44]. Furthermore, K102 is vital for PA endonuclease activity, particularly for its mRNA cap-binding activity [45] and consequently for IAV growth [46]. The S631 was located on the β9 strand ( Figure 3F) within the PB1-binding site of the PA C-terminal domain [47]. In the PB2 structure, K718 was situated in the loop region of the C-terminal "DPDE" domain ( Figure 3G). This domain contains a nuclear localisation signal (NLS) and binds to the host nuclear import protein, importin α to facilitate PB2 nuclear import [48].  NS1 and NS2, two main IAV non-structural proteins were found to be acetylated at the N-terminus (Table 1). Additionally, NS1, which is the main interferon antagonist, was also methylated on R193 and modified by allysine on K110. Both K110 and R193 were highly conserved ( Figure 4A) across 998 NS1 sequences belonging to 48 IAV subtypes (Supplementary Table S2). Furthermore, they were also conserved within the NS1 of IAV and IBV ( Figure 4B). In the NS1 structure, both K110 and R193 were part of the critical NS1 effector domain [49] and located on β-strands: K110 on β2-strand and R193 on β7-strand ( Figure 4C). HA, the receptor-binding protein of IAV that is critical for virus entry, was found to be modified by methylation on R91, K252, and R269 and allysine on K62 (Table 1). HA is the most rapidly evolving protein of IAV. Therefore, the conservation analysis was performed using 480 sequences from only H1 subtypes (Supplementary Table S2). We found that R269 was largely conserved within HA of IAV subtypes ( Figure 5A). However, K252 was highly conserved across HA of IAV, IBV, and ICV ( Figure 5B). The structural analysis revealed that all modified residues were part of the HA1 globular subunit and located in its loop regions ( Figure 5C). As HA1 is further divided into fusion, vestigial esterase and receptor-binding domains [50], the methylated R91 and R269 were situated in the vestigial esterase domain and K252 in the receptor-binding domain, whereas the allysine K62 was situated in the fusion domain ( Figure 5C).

Two Hundred and Forty-Five Host Proteins Were Modified at 300 Sites by Methylation, Acetylation, and/or Allysine in Response to IAV Infection
Our proteomic screen identified a total of 2280 host proteins (Supplementary Table S3), out of which, 245 were discovered to be selectively modified by methylation, acetylation and/or allysine in IAV-infected cells. However, the total number of modified sites was 300 because, like viral proteins, some host proteins were modified: (1) by the same modification on multiple residues, (2) by more than one modification, or (3) on the same residues by different modifications (Supplementary Table S4). In addition, 18 proteins were discovered to be modified in both uninfected and infected cells and 54 proteins were selectively modified only in uninfected cells. Since our aim here was to identify the host proteins that were selectively modified in response to IAV infection, these proteins were excluded from further analysis.

Furthermore, 16 of 34 proteins were methylated on both lysine and arginine residues (Supplementary
Acetylated proteins. We identified 98 host proteins to be uniquely acetylated at 108 sites in infected cells (Supplementary Table S4). Out of these, 20 proteins were acetylated on 24 lysine and 27 on 31 serine residues, and 53 proteins were acetylated at the N-terminus. Most of them were acetylated on one lysine or serine residue; only 8 were acetylated on two lysine or serine residues. Further, as mentioned above, some of the acetylated proteins were also methylated or allysine (Supplementary Table S4).
All 98 proteins that were either N-terminally acetylated or acetylated on lysine and serine residues were included in functional association networks analyses using the STRING database ( Figure 7A). This analysis processed 77 of the acetylated proteins and identified their involvement in viral processes (GO = 0016032) (FDR = 0.0201) and IRES-dependent viral translation initiation (GO = 0075522) (FDR = 0.0289). The MCODE algorithm identified three major interconnected clusters of identified acetylated proteins. The highest-ranked cluster 1 (MCODE score: 5) included SF3B2, poly (rC) binding protein 2 (PCBP2), NHP2L1 and hnRNPH2 ( Figure 7B), all of which have been implicated in mRNA splicing. The SF3B2 was acetylated on conserved K672 and K680, and PCBP2 was acetylated on conserved S84, whereas the hnRNPH2 was acetylated at the N-terminus (Supplementary Tables S4 and S5). As mentioned above, IAV utilises the host mRNA splicing machinery, particularly to process its M and NS gene transcripts to produce multiple proteins [67]. Further, PCBP2, which regulates antiviral response, has been identified as the target of miRNA-like small RNA encoded by IAV H5N1 subtype [68]. Cluster 2 (MCODE score: 4) contained 3 tubulins, all acetylated on S364 and associated chaperone T-complex protein 1, acetylated at N-terminus ( Figure 7B) (Supplementary Table S4). Cluster 3 included actin, acetylated on highly conserved K52 and S54 and actin-binding proteins: actinin, acetylated on highly conserved K682 and K684; laminin, acetylated on conserved S2164; and associated kinase serine/threonine-protein kinase PAK2, acetylated on S42 ( Figure 7B; Supplementary Tables S4 and S5). The significance of the proteins in both clusters 2 and 3 in IAV infection has been described above and elsewhere [69]. In support of this data, ASEB analysis predicted that K672 of SF3B2 and K52 of actin can be acetylated by GCN5/PCAF acetyltransferase ( Figure 7C). In addition, ASEB analysis also predicted that the conserved K540 of polyadenylate-binding protein 4 that was identified to be acetylated here, can be acetylated and deacetylated by CBP/p300 and SIRT1, respectively ( Figure 7C).
Individual acetylated proteins (Supplementary Tables S4 and S5) not included in any cluster have also been implicated in IAV infection. Farnesyl pyrophosphate synthase (FDPS), acetylated on S73, is the target of anti-IAV host factor viperin and is a pro-IAV factor [70]. Serine/threonine-protein kinase mTOR, acetylated on conserved K1068, is a component of mammalian target of rapamycin (mTOR) complex and known to promote IAV infection [71]. The N-terminally acetylated eukaryotic translation initiation factor 4GI (eIF4GI) is recruited by IAV NS1 for preferential translation of viral proteins [49] and nucleophosmin interacts with viral ribonucleoprotein (vRNP-a complex of IAV NP, RNA, and RNA polymerase subunits PA, PB1, PB2) and contributes to viral RNA synthesis [72]. Peptidyl-prolyl cis-trans isomerase A or cyclophilin A (PPIA), acetylated at the N-terminus, restricts IAV infection by targeting M1 [73]. Finally, importin 4 (IPO4) and exportin 2 (CSE1L), both acetylated at the N-terminus, are the members of host nuclear transport machinery which is critical for IAV replication [1].

Discussion
The role and significance of lysine acetylation and lysine/arginine methylation have been extensively studied in histones modifications and their influence on gene expression. However, these modifications are now known to occur in many non-histone proteins and contribute to, for example, their subcellular localisation and organelle retention (Nterminal acetylation) [2], intracellular transport (lysine acetylation) [2], and role in RNA processes such as splicing (arginine methylation) [3]. Evidently, both viral and host proteins with these characteristics are involved in IAV infection [1], and, in this study, we have identified such proteins with those modifications. Further, the identified host proteins were distributed across intracellular compartments and the plasma membrane of the host cell (Supplementary Figure S2). M1 and NP proteins carried the most modifications of all IAV proteins, with multiple modifications on the same and different amino acid residues. M1 and NP perform multiple functions during the IAV life cycle, from viral entry to RNA synthesis to assembly. To execute these functions, they shuttle between the cytoplasm and nucleus and to the plasma membrane and interact with various host as well as viral proteins and RNAs [74]. Potentially, the modifications on M1 and NP identified herein are crucial to conduct those functions. In M1, K95 was methylated as well as acetylated and K98 was methylated as well as allysine. Both these residues are located on M1 "helix six", a region of positively-charged amino acids containing an NLS and responsible for multiple M1 functions, including binding to vRNP and lipid membrane [36]. Therefore, it is conceivable that different modifications on K95 and K98 dynamically influence the interactions of M1 with vRNP and lipid membrane. The NP, an RNA-binding protein, was methylated mostly on arginine residues (R150, R246, R317, R416 and R422). Evidently, the arginine methylation of RNA-binding proteins such as hnRNPs-several of which have also been found here to be selectively methylated on arginine in infected cells, is common and regulates their nucleocytoplasmic shuttling and role in transcription and mRNA splicing [75,76]. Therefore, methylation of R150, which resides in the NP groove region that binds viral RNA [39], may influence NP's role in viral transcription and vRNP nuclear import/export. Furthermore, methylation of K317 and R416 may contribute to NP-NP oligomerisation by governing the salt bridge formation with residue E369 [77] and E339 [78], respectively. Evidently, methylation of lysine influences the conformational changes in HSP90 through salt bridge formation [79]. In addition, NP was also acetylated on multiple serine residues (S274, S283, S287, S326, and S403), of these S274 is part of the NP nuclear export signal 3 (NES3) which facilitates the export of vRNP from the nucleus to the cytoplasm [74,80]. The role of serine acetylation in nuclear transport is yet to be identified, but lysine acetylation does regulate nuclear export [81,82]. Furthermore, N-terminal acetylation of the nuclear transport proteins, exportin 2 and importin 4 in infected cells detected here, may also regulate vRNP nuclear transport.
Likewise, identified modifications of other IAV proteins also potentially influence their already known functions. For example, methylation and acetylation of viral RNA polymerase subunit PA on both K102 and K104 in the endonuclease domain potentially regulate PA binding to viral RNA [45]. Indeed, methylation as well as acetylation of HIV-1 Tat on two adjacent lysine residues (K50 and K51) is known to regulate its binding to RNA [83]. The allysine of PB2 on K718 in the NLS domain may determine its diverse binding activity to NLS-binding pockets of importin α isoforms [84]. Similarly, allysine and methylation of NS1 in the effector domain (on K110 and R193, respectively) potentially regulate its binding to host cleavage and polyadenylation specificity factor 30 (CPSF30) and poly(A)-binding protein II (PABII) and facilitate the sequestering of host mRNAs. Furthermore, these modifications may also facilitate the interaction of NS1 with eIF4GI for preferential translation of IAV proteins [49]. Incidentally, eIF4GI was also found to be N-terminally acetylated in the infected cells. Finally, the methylation of HA on R269, which is part of a crucial "tetrad" salt bridge, potentially influences its irreversible conformational change at acidic pH in endosomes during IAV entry [85].
On host proteins, methylation was the most abundant modification followed by acetylation and allysine in infected cells. Nevertheless, the methylated, acetylated, and allysine host proteins in infected cells were commonly annotated to three protein-protein interaction clusters implicated in metabolic (glycolysis, TCA cycle, lipid metabolism), cytoskeletal, and RNA processes. Evidently, these three processes are known to be involved in IAV infection and their proteins interact with viral proteins in various stages of the IAV life cycle [51,54,67,86]. Further, the identified modifications have been known to influence the function of some of those proteins. For example, methylation of GAPDH is associated with its enhanced catalytic activity in pancreatic cancer [87]; acetylation alters the activity and stability of pyruvate kinase M in cancer [88]; acetylation regulates the stability of actin filaments and microtubules, hence intracellular transport [2], including of the IAV components [18]; and, finally, methylation regulates the mRNA splicing [75,76].
In conclusion, our data point to strong alterations in the landscape of methylation, acetylation and allysine of both viral and host proteins in response to IAV infection. The modifications in viral proteins occurred on conserved residues in critical domains, whereas the modified host proteins belonged to important cellular pathways that are known to be exploited by IAV. Furthermore, SIRT1 and GCN5/PCAF (predicted to deacetylate and acetylate, respectively, some acetylated viral and host proteins identified here) have been described to play an antiviral and proviral role, respectively, during influenza virus infection [19,89]. Therefore, the significance of these modifications on some viral and host proteins vis-à-vis IAV infection is evidently clear, but on others, it seems to be rather complex. Experimental validation of the functional significance of the identified modifications, on both conserved and non-conserved residues will shed further light on it.
This study had some limitations. We were unable to detect known lysine acetylation sites that were previously reported to be present on viral NP and NS1 [20,21,90]. However, these reports used selective samples that were enriched either by immunoprecipitation, ectopic overexpression, or ultracentrifugation to detect the lysine acetylation sites. In our view, the enrichment and targeted sample strategies are narrow and selective and will confound the detection of other modifications. Our goal was to detect the modifications that enriched and occurred specifically in response to IAV infection and to gain insights into possible crosstalk between different modifications. Most of the modifications are kinetically dynamic in response to a stimulus like infection, and the turnover of lysine acetylation is faster than methylation [91]. For similar reasons, the mutagenesis of modified residues will unlikely reveal the significance of a modification on a protein, because the same residue can undergo dynamic and alternative multiple modifications as we have seen here. The alternative experimental approaches [5,92] that navigate these challenges (using a model of primary human respiratory cells/tissues) will need to be applied in tandem to reveal and validate the full extent of global protein modifications and their significance during IAV infection.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/v13071415/s1. Figure S1: Influenza A virus (IAV) titer and polypeptide profile of infected cells used for mass spectrometry, Figure S2: Subcellular localization of the identified modified host proteins based on their GO term, Table S1: Detected IAV proteins and their modifications, Table  S2: Number of viral protein sequences and IAV subtypes used for alignments, Table S3: List of total number of identified host proteins, Table S4: List of total number of identified host proteins undergoing the modifications and their modified sites, Table S5: Conservation of the modified sites in host proteins across human, pig, chicken, dog, horse, and mouse.

Data Availability Statement:
The data presented in this study are available as supplementary material or from the authors upon request.