Molecular mechanisms of human coronavirus NL63 infection and replication

Highlights • HCoV-NL63 being the most prevalent human coronaviruses after HCoV-OC43.• HCoV-NL63 cause upper and lower respiratory tract infections mainly in young children.• HCoV-NL63 shares a common host cell virus receptor (ACE2) with SARS-like coronaviruses.• HCoV-NL63 is a safe surrogate to study disease mechanisms and develop therapeutic interventions against SARS-like-CoV.

Patients affected with HCoV-NL63, primarily develops into mild to moderate respiratory disease symptomatically featuring fever, cough, and runny nose, or into croup and pneumonia (Bastien et al., 2005;Forster et al., 2004;König et al., 2004;Vabret et al., 2005;van der Hoek et al., 2005). HCoV-NL63, like other CoVs, was thought to be a zoonotic pathogen with an origin in bats (Huynh et al., 2012;Tao et al., 2017). Bats not only work as a reservoir but also as a home for interspecies genetic recombination in the development of new genotypes or new emerging viruses (Al-Khannaq et al., 2016;Tao et al., 2017). These genetic changes cause HCoV-NL63 to have a seasonal/repeated infection pattern every year, although with low frequencies of about 2% and self-limiting respiratory infections (Al-Khannaq et al., 2016;Chiu et al., 2005;Kiyuka et al., 2018;Sipulwa et al., 2016).
HCoV-NL63 shares a common host cell virus receptor (ACE2) with SARS-CoV and SARS-CoV-2 (Hofmann et al., 2005). It is also known that both HCoV-NL63 and SARS-like CoVs infect ciliated respiratory cells, although with different efficiency, using the same virus receptor Banach et al., 2009;Zhu et al., 2020). Working with SARS-like CoVs require access to a high-containment level biosafety laboratory (BSL-3), while HCoV-NL63 work can be performed in BSL-2 facilities. Therefore, HCoV-NL63 can be used as one of the safe surrogates for comparative studies on viral infectivity and replication capacity, disease mechanisms and therapeutic interventions against the emerged SARS-CoV-2. This prompted us to review the current research towards the mechanism of infection and replication of HCoV-NL63.
The recombination between different HCoV-NL63 strains or with other CoVs have resulted in different clades (A, B, C) and seven subclades (A1, A2, A3, B, C1, C2, and C3) described so far. The present review includes an updated phylogenetic analysis on the complete gene sequences of HCoV-NL63 (Fig. 2). The internal variability is mainly based on the amino (N)-terminal domain of the spike gene and the nsp2/ nsp3 sequence of the ORF1a (Arden et al., 2005;Chiu et al., 2005;Dominguez et al., 2012;Minosse et al., 2008;Moës et al., 2005;Wang et al., 2020), while the nucleotide sequence from 1b gene is rather conserved (Arden et al., 2005). The different genotypes co-circulate as a mixture of variants strains (Chiu et al., 2005;Minosse et al., 2008), with a prevalence of ~1-3% among symptomatic patients with respiratory disease, with genotype A associated with most of the hospitalizations due to life-threatening acute respiratory disease (Arden et al., 2005;Dominguez et al., 2012;Minosse et al., 2008). The highly prevalent circulation of HCoV-NL63 might result in widespread, newly emerging subgenotypes showing more severe respiratory symptoms .

Virus isolates for research purposes
Different HCoV-NL63 sequences have been submitted to the Gen-Bank database, but few studies reported virus isolation and propagation in vitro.  was the first isolating the virus ("Amsterdam-1 isolate") from a nasopharyngeal swab of a 7-month-old child in January 2003 in Amsterdam, Netherlands (van der Hoek et al., 2004). One month later, Fouchier (2004) reported the isolation of HCoV-NL63 from a nasal swab obtained from a 8-month-old child in Rotterdam, Netherlands (Fouchier et al., 2004). Later, Lednicky (2013) isolated the virus from a contaminated stock of renal proximal tubule epithelial cells (RPTEC) in 2004 (Lednicky et al., 2013), while Beau De Rochars (2017) isolated the virus from blood samples of four children of Haiti in (Beau De Rochars et al., 2017. Most recently, in Japan, Komabayashi (2021) isolated the virus from six frozen nasopharyngeal swabs collected between 2012 and 2020 (Komabayashi et al., 2021). Among all the isolates described in the literature, only the Amsterdam-1 isolate originally described  is broadly available for the research community via BEI Resources Repository (NIAID, NIH: HCoV-NL63, NR-470), a global provider of materials for infectious disease research, which might limit the pathogenic comparison between studies.
Phylogenetic studies from sequenced viral genomes reported that three genotypes (A, B, C) with probable different disease severity (Dominguez et al., 2012;Shao et al., 2022;Wang et al., 2020). The greatest variability was found in the receptor binding domain (RBD) of the S protein, followed by nsp3. Interestingly, a single mutation within the S protein (I507L) was identified in the subgenotype C3, which was associated with enhancing virus entry in cell culture .

Genomic organization
Like other members of the order Nidovirales, HCoV-NL63 is an enveloped single-stranded positive-sense RNA virus (27,553 bases in size) that is capped and polyadenylated . Following the typical Nidovirus genome organization: 5 ′ -ORF1a-ORF1b-spike (S)-ORF3-envelope (E)-membrane (M)-nucleocapsid (N)-polyadenylated Tail-3 ′ , the HCoV-NL63 genome includes a large ORF 1ab (two-thirds of the genome) on the 5 ′ terminal encoding the proteases and most proteins (non-structural proteins, nsp1-nsp16) necessary for controlling gene expression and replication. The ORF 1a overlaps with the 1b gene region, which encodes an RNA-dependent RNA polymerase . HCoV-NL63 ORF1a/1b contains a putative elaborated pseudoknot structure (ribosomal frameshifting element between ORF 1a and ORF1b) that triggers a -1 ribosomal frameshift to translate the complete 1ab polyprotein (Namy et al., 2006;Pyrc et al., 2004). Likewise, the 3 ′ terminal contains the regions for structural proteins (one-third of the genome) S, E, M, and N (van der Hoek et al., 2006). The HCoV-NL63 genome encodes for only one accessory N-glycosylated protein known as ORF3, which is expressed from a distinct subgenomic RNA, one of at least 6 distinct mRNAs . The ORF3 gene has a unique nucleotide composition and appears as a U-rich and A-poor region within the genome, indicating a recent transfer event from another viral or cellular origin .

Virion structure
Members of family Coronaviridae are roughly spherical with virion size ranging from 120 to 160 nm in diameter, and envelop showing petal shaped surface projections of the S protein decorating the surface of the virion (ICTV, 2020). The viral envelope is supported by the M protein and contains a small amount of the E protein (Liu and Inglis, 1991;Schoeman and Fielding, 2019;Yu et al., 1994). Inside the viral envelope, the genome is bound by the nucleocapsid N protein and forms a helical symmetric nucleocapsid (Masters, 2006).

S protein
The HCoV-NL63 S protein mediates both cell attachment and membrane fusion; it is the major determinant of host and tissue tropism and may contribute to viral pathogenesis by activating the endoplasmic reticulum (ER) stress response (Chan et al., 2006;Zheng et al., 2006). The S protein is a large type I single-chain transmembrane glycoprotein, which forms a homotrimer with a molecular weight of 128-160 KDa before glycosylation and 150-200 KDa after N-linked glycosylation (Holmes et al., 1981;Rottier et al., 1981). The HCoV-NL63 S protein is a class I fusion protein similar to influenza virus hemagglutinin and the HIV-1 Env glycoprotein gp120/gp41 (Bosch et al., 2003;Weis et al., 1988;Weissenhorn et al., 1997). As a class I viral fusion protein, the S protein forms homotrimer and is cleaved by host proteases into the N-terminal subunit (S1; bulb) for receptor binding and the carboxyl-terminal subunit (S2; stalk) for membrane fusion. Disulfide bonds modify the exterior ectodomain of the S protein, whereas the conserved cysteine residues in the very short cytosolic tail are modified by palmitoylation (Fung and Liu, 2018). The variable S1 subunit constitutes the receptor-binding domain (RBD) and contains a unique 179-amino-acid domain not present in other CoVs (Li et al., 2007;van der Hoek et al., 2004). The conserved S2 subunit contains a membrane-spanning region (transmembrane and cytoplasmic domains); 2 generally conserved heptad repeat regions (HR1 and HR2) forming a 6-helix bundle that is critical during virus fusion; and the fusion peptide, which is similar to other class I fusion proteins . A total of 58 whole genome sequences of HCoV-NL63 were used for phylogenetic analysis using Geneious Prime® 2021.2.2. Genomes were aligned with MAFFT (v7.450) and Neighbor Joining tree was constructed with PhyML (3.3.20180621) with 1000 Bootstrap. Three clades (A, B and C) were identified with using genome NCBI accession DQ445911.1 as root. Subclades were identified with bootstrap value of over 99% where Clades A and C were clustered into 3 subclades each. FigTree phylogenetic drawing tool (V1.4.4) was used to build the tree.

M protein
The M protein (25-30 KDa) is the most abundant structural protein, embedding in the envelope via 3 transmembrane domains (Masters, 2006). The M protein forms homodimer and interacts with other viral structural proteins to orchestrate the assembly of the CoV particle. In HCoV-NL63 and most other CoVs including the related HCoV-229E, the short N-terminal ectodomain of M protein is modified by N-linked glycosylation (Dea et al., 1990;Hogue and Nayak, 1990;Naskalska et al., 2018;Stern and Sefton, 1982;Utiger et al., 1995). However, in some animal CoVs, like the mouse hepatitis virus (MHV) and the bovine coronavirus (BCoV) or in HCoV-OC43, this ectodomain is modified by O-linked glycosylation (Deregt et al., 1987;Holmes et al., 1981;Lapps et al., 1987;Mounir and Talbot, 1992). The M protein may also contribute to viral pathogenesis. For example, the retinoic acid-inducible gene 1 (RIG-I)-dependent induction of type I interferon (IFN) is observed in cells overexpressing the M protein of SARS-CoV but not HCoV-HKU1 (Siu et al., 2014b).

E protein
The E protein is a small (8-12 KDa) integral membrane protein found in low amounts in the virion (Corse and Machamer, 2000;Liu and Inglis, 1991). Current evidence on avian infectious bronchitis coronavirus (IBV) strongly suggests that the E protein adopts an N-ecto/C-endo topology with one transmembrane domain (Corse and Machamer, 2000;Yuan et al., 2006). In SARS-CoV, the E protein is modified by N-linked glycosylation, and 3 cysteine residues in its endodomain are modified by palmitoylation Yuan et al., 2006). Additionally, different studies have shown that the E protein of both SARS-CoV and IBV forms homopentamers with ion channel (IC) activity (viroporins) (Nieto-Torres et al., 2014;Wilson et al., 2004), where the IC activity could modulate the process of virion release and contribute to viral pathogenesis. Although the deletion of the E gene is not lethal for SARS-CoV, the resultant mutant virus was severely defective in virion morphogenesis and attenuated in vivo, compared with the wild-type virus (DeDiego et al., 2007).

N protein
The N protein is a multidomain, multifunctional protein essential for viral replication and a number of cellular processes, including RNA packaging, viral genome replication, and evasion of the immune response Szelazek et al., 2017;Zhou et al., 2008). The N-terminal domain is responsible for nucleic acid binding, and the C-terminal domain is involved in protein oligomerization. Interestingly, as with SARS-CoV, the HCoV-NL63 N protein is not translocated in the nucleus of infected cells (Zuwala et al., 2015).
The N protein (43-50 KDa) forms a dimer and binds to the genomic RNA in a beads-on-a-string fashion, forming a helically symmetric nucleocapsid (Chang et al., 2006;Fan et al., 2005;Tan et al., 2006). As with other CoVs, the sgRNA coding the HCoV-NL63 N protein is the most abundant among the different viral proteins . In SARS-CoV and other CoVs, the N protein is phosphorylated by cellular kinases such as glycogen synthase kinase 3 (GSK3) and ataxia-telangiectasia mutated and Rad3-related (ATR) kinase (Fang et al., 2013;Wu et al., 2009a). Other modifications such as SUMOylation, ADP-ribosylation, and proteolytic cleavage by caspases have also been demonstrated in the N proteins of some CoVs (Eléouët et al., 2000;Grunewald et al., 2018;Li et al., 2005). Moreover, the N protein of some CoVs (i.e., IBV, SARS-CoV) can affect cell cycle progression, cytoskeleton organization, gene transcription, and apoptosis induction in infected cells (Harrison et al., 2007;Surjit et al., 2006Surjit et al., , 2004Zhao et al., 2006;Zhou et al., 2008). Contrastingly, no significant alteration of cell cycle progression has been reported in HCoV-NL63 (Zuwala et al., 2015).
Multiple crystallization attempts of the full-length HCoV-NL63 N protein have been unsuccessful because the rigid β-sheets within this region restrict its plasticity compared to those of the interrupted β-structures in SARS-CoV and MHV (Szelazek et al., 2017). The significance of the aforementioned differences for nucleic acid binding remains unknown. The HCoV-NL63 nucleoprotein forms oligomers via its C-terminal domain (CTD) and binds nucleic acids via its N-terminal domain (NTD) (Zuwala et al., 2015). The CTD exists as a dimer in solution and tends to aggregate in solution, which may reflect further steps of nucleocapsid assembly (Zuwala et al., 2015). Nevertheless, in the case of the HCoV-NL63 CTD, interaction is weak if present at all (Zuwala et al., 2015). The HCoV-NL63 NTD contains a large, positively charged groove implicated in RNA binding. Each monomer within the structure coordinates several sulfate ions derived from the crystallization buffer. Sulfate ions chemically resemble phosphate moieties present within the structure of RNA and are therefore likely to be located at the same binding sites (Szelazek et al., 2017). HCoV-NL63 NTD has little or no specificity for particular sequences, or even the type of nucleic acid (i.e., RNA vs. DNA) (Zuwala et al., 2015). Interestingly, in the case of the swine alphacoronavirus TGEV, the RNA-binding ability of the N protein is important, not only for genome encapsidation but also for discontinuous transcription and polymerase template switching (Mateos-Gómez et al., 2011;Zúñiga et al., 2010).

ORF3 accessory protein
Poorly characterized overall, the N-glycosylated ORF3 protein has been detected within the ER/Golgi intermediate secretory compartment (ERGIC), where CoV assembly and budding occur. HCoV-NL63 ORF3 is incorporated into virions colocalized with E and M proteins in the same compartment and facing towards the extracellular space. This further suggests an important function, particularly in virus assembly and/or budding from infected cells (Müller et al., 2010).
The HCoV-NL63 S protein shares more than 50% of amino acid (aa) identity with the HCoV-229E, and both viruses belong to the same genus, but they use different viral receptors (Pöhlmann et al., 2006). Conversely, HCoV-NL63 S protein shares only 25% and 17.1% of aa sequence identity with SARS-CoV and SARS-CoV-2, respectively. However, these 3 viruses use ACE2 as a host receptor with different binding affinities (Brielle et al., 2020;Smith et al., 2006). The RBD of the HCoV-NL63 S-protein has 3 non-linear binding motifs located at the C-terminus S1 subunit (aa 476-616), and these in conjunction with distinct aa from the central regions of the CTD of S1 interact with ACE2 Lin et al., 2008;Pöhlmann et al., 2006). These regions largely overlap based on sequence homology and show comparable binding affinities with SARS-CoV RBDs (Li et al., 2007;Lin et al., 2008), probably because of the linear and monomeric nature of both RBDs, indicating that in the full-length S-protein of HCoV-NL63, the non-linear RBD has weaker interaction than linear RBD from SARS-CoV (Mathewson et al., 2008;Wong et al., 2004;Wu et al., 2011). Initial proteolytic studies of the S2 fusion core identified an α-helical domain consisting of a trimer of the HR segments N57 and C42. The resolved crystal structure of this trimer complex shows distinctive high-affinity conformations of interacting cross-sectional layers of 6 helices. It has been suggested that the larger HR regions of the alphacoronavirus may be required to prime the S proteins for the fusion-activating conformational changes during entry of the virus .
Variations in aa composition and glycosylation are crucial for CoV infectivity and interaction with the host immune system. In HCoV-NL63, a deletion of 18 aa from the C-terminus of the S protein, corresponding to an intracellular retention signal, has been shown to enhance (1.5 times) its accumulation and facilitating virus entry. Further deletion of residue 29 has also been shown to enhance the amount of S protein on both the cell surface and virion yet reducing virus entry by 25%, suggesting that residues 19-29 may contribute to membrane fusion. A 29 aa-deletion mutant had a defect in anchoring on the plasma membrane, which led to a dramatic decrease of S protein in virion and virus entry; a total of 15 residues (Y498, V499, V531, G534, G537, D538, S540, G575, S576, E582, W585, Y590, T591, V593, and G594) within the RBD were necessary for receptor binding and virus entry (Lin et al., 2011). They probably form 3 receptor-binding motifs, being the 3rd motif highly conserved between HCoV-NL63 and SARS-CoV (Lin et al., 2011;Wu et al., 2009b). Ren et al. (2022) reported that ACE2 polymorphisms could impact susceptibility to infection by SARS-CoV-2, SARS-CoV, and HCoV-NL63. For example, ACE2 variants D355N and G352V restrict the S protein-ACE2 interaction and limit infection, where HCoV-NL63 S protein binds to ACE2 less efficiently (Ren et al., 2022).
Furthermore, the C-terminal region (aa 153-226) of the HCoV-NL63 M protein is responsible for attaching heparan sulfate proteoglycans to target cells (Milewska et al., 2014;Naskalska et al., 2019), and C-type lectin receptors like DC-SIGN could work as co-receptors . Most recently, genome-wide CRISPR screens in Vero-E6 identified HMGB1 as a novel regulator of ACE2 involved in virus entry of HCoV-NL63, SARS-CoV, and SARS-CoV-2. This nuclear protein translocates to the cytoplasm under cellular stress conditions (Wei et al., 2021).
(2) Virus-ACE2 interaction triggers receptor mediated endocytosis, with acidification in the endosome. (3) Spike protein priming is carried out with TMPRSS2 in the early endosome or cathepsins in the late endosome, and viral membranes fuse with the endosome to release the viral genome into the cytoplasm. (4) Nucleocapsid protein is degraded to release the uncoated viral genome ready for translation with the ribosomes to produce the two viral polyproteins (pp) 1a and 1ab through ribosome − 1 frameshifting.
(5) Viral polyproteins induce autoproteolytically cleavage with their proteases nsp3 and nsp5. (6) Membrane-spanning non-structural proteins (nsp) nsp3 and nsp4 mediates the zippering of the rough endoplasmic reticulum (RER) membrane to generate double membrane structures (DMS). (7) Virus genome and nsps initiates the viral replication and transcription within the DMS, releasing the sgRNA outside to translate the viral structural proteins. (8) N protein accumulated in inclusion bodies (IB), while other structural proteins are processed into the ERGIC. Viral genome release from DMS is coated with N protein and directed to ERGIC and Golgi where new virions are assembled and packaged into vesicles. (9) Virions can be released either via exocytosis or through cell lysis.
PACE4, PC4, PC5, PC7), which activates MERS-CoV, and SARS-CoV-2 but not SARS-CoV (Hoffmann et al., 2020a;Millet and Whittaker, 2014;Shang et al., 2020), are less likely to exert an influence on HCoV-NL63 entry, given that the S protein lacks a furin motif (arginine-rich residues) to be activated . The role of other non-membrane-associated proteases like elastase or coagulation factor Xa, which enhance SARS-CoV infectivity through cleavage of the S protein (Du et al., 2007;Matsuyama et al., 2005), has not been investigated in HCoV-NL63 infections.
Generally, once the virions are internalized into endosomes, the interaction between viral N protein and cellular host proteins is required to release the viral genome from its capsid, which has not yet been demonstrated for HCoV-NL63 (Fig. 3). It has been proposed for HCoV-229E and IBV that valosin-containing protein (p97) would help in the maturation of early endosomes and decomposition of the nucleocapsid (Wong et al., 2015). The mechanism of the subsequent transduction response after ACE2 engagement in relation to HCoV-NL63 pathogenesis has not been investigated.
Following translation, the HCoV-NL63 polyprotein contains two nsp with proteolytic domains: nsp3 and nsp5 . These proteases autocatalytically cleavage the polyproteins into 16 nsp . The nsp3 of HCoV-NL63, HCoV-229E, HCoV-OC43, and HCoV-HKU1 contains two proteolytic domains, known as papain-like protease 1 (PLP1pro), which autocleavage the polyprotein at the nsp1/nsp2 site to release nsp1, and PLP2pro, that cleaves polyproteins at the nsp2/nsp3 and nsp3/nsp4 sites to release nsp2 and nsp3, respectively . Moreover, although less efficient than PLP1, PLP2 can also autocleavage the polyprotein at the nsp1/nsp2 site (Lim and Liu, 1998;Ziebuhr et al., 2007). Other CoVs like MERS-CoV or SARS-CoV contain only one proteolytic domain (PLpro) within the nsp3 . The second protease nsp5, known as the main protease (Mpro) or 3-Chymotrypsin-like protease (3CLpro), is a highly conserved endopeptidase with a serine-like domain responsible for cleaving and releasing the remaining 13 non-structural proteins (Piñón et al., 1999;Pyrc et al., 2005;Ziebuhr et al., 2000). The HCoV-NL63 Mpro is active when forming homodimers containing 3 domains each, where the catalytic site (Cys144 and His 41) is located in a cleft formed between Domain I and Domain II, while Domain III allows the homodimer formation (Wang et al., 2016). Further analysis of substrate specificities of Mpro of HCoV-NL63 found histidine at position P1 instead of glutamine present in other CoVs (Wang et al., 2016). Overall, CoV 3C-like protease substrate profiling has identified glutamine preference for the P1 position, leucine at the P2 position, basic residues at the P3 position, small hydrophobic residues at the P4 position, and small residues at the P1 ′ and P2 ′ positions (Chuck et al., 2011).
A genome-wide CRISPR knockout screen in Huh-7.5 found TMEM41B as a pan-CoV host factor required for a post-entry step in the CoV life cycle. This poorly studied ER transmembrane protein was the only gene implicated in autophagy. The specific host factors of HCoV-NL63 rely on a core set of host chromatin regulators like EP300, KMD6A, KMT2D, MED23, MED24, MEN1, PAXIP1, and SETDB1, which were proposed to reprogram the host transcriptome for successful infection (Schneider et al., 2021).
Nsp3 is a multifunctional protein involved in viral replication (Lei et al., 2018;Neuman et al., 2008). Among its eight domains, the ubiquitin-like domain (Ubl1) facilitates RNA synthesis through binding nsp3 and RTC to the viral genome (Serrano et al., 2007) via a serine-and arginine-rich linker region of the residual N protein (Hurst et al., 2013;Keane and Giedroc, 2013), and promotes the initiation complex Zúñiga et al., 2007).

Replication and transcription
As in IBV, MERS-CoV, and SARS-CoV infection, HCoV-NL63 RNA synthesis occurs within replication compartments , and the replication is mediated by the common negative-strand intermediates to make either viral genome for new virions and sgRNAs for coding structural proteins through a continuous or discontinuous strategy, respectively   (Fig. 3). Both processes require the formation of the RTC assembled by the recruitment of many nsps and then binding to the 3 ′ -end of the viral genome within the replication compartments (Züst et al., 2008). The precise mechanism of how the RTC switches to generate a viral genome or sgRNAs in HCoV-NL63 replication has not yet been elucidated. However, a proposed model in TGEV has suggested that N protein facilitates sgRNAs transcription (Zúñiga et al., 2010) and that the viral genome synthesis in MHV requires phosphorylation of the viral nucleocapsid by the host GSK-3 with the help of a helicase (Wu et al., 2014).
The generation of sgRNAs in alphacoronavirus is initiated with the formation of short RNA primers (~6 nucleotides) by nsp7 and nsp8, which are essential co-factors for the viral RNA polymerase (nsp12) (Xiao et al., 2012). To maintain fidelity on CoV genomes like SARS-CoV, the exoribonuclease nsp14 functions as a proofreading enzyme (Minskaia et al., 2006), highly enhanced by nsp10 (Bouvet et al., 2012;Ma et al., 2015). Together, these proteins are part of the RTC, which recognizes specific sequences on the viral RNA to conduct replication or transcription. The RTC recognizes the anti-leader sequence at the 3 ′ -end and starts synthesizing the negative-strand intermediates but pauses at transcription regulatory sequences (TRSs) on the genome (leader-TRS) for making templates for new genomes, or between them (body-TRS), to generate templates for sgRNAs . These negative-strand intermediates are synthesized at very low amounts, at around only 1% of the positive strands like in TGEV (Sethna et al., 1991), and they contain both polyuridylate and anti-leader sequences. Additionally, nsp9 might stabilize newly forming RNA strands to avoid their degradation (Egloff et al., 2004;Sutton et al., 2004).
As for MHV and SARS-CoV-2, once newly formed HCoV-NL63 RNA strands are synthesized, nsp3 mediates the formation of a molecular pore in -infected cells (Wolff et al., 2020). Viral pores allow controlled release of viral RNA transcripts from replication compartments to the cytosol for protein translation. Finally, the translation of the viral proteins is initiated, and the HCoV-NL63 S protein is incorporated into virions through microtubule interaction near the nucleus (Rüdiger et al., 2016), while the RNA-binding N protein exhibits a tendency to aggregate into the cytoplasm, suggesting virion assembly (Zuwala et al., 2015) (Fig. 3).
HCoV-NL63 is hardly more isolated than other CoVs with variable replication kinetics , and mild cytopathic effect (CPE) if present, which can be visualized throughout the cell culture, with a refractive, rounding appearance and vacuolization, followed by detachment (Fouchier et al., 2004;Herzog et al., 2008;Lednicky et al., 2013;Milewska et al., 2014;Orenstein et al., 2008;Schildgen et al., 2006;Smith et al., 2006;van der Hoek et al., 2004). In these studies, the first signs of CPE start between 24-72 h post-inoculation (hpi) and peak between 72 and 96 hpi, with around 25 -50% of infected cells and an estimated yield of ~10 5 TCID 50 /mL. Trypsin treatment showed no enhancement of the infection, and infected freeze-thawed cells were thought to have less infectivity than supernatant alone (Lednicky et al., 2013).
Differentiated human bronchial epithelial cell cultures of HCoV-NL63 revealed a polarized apical release of new virions , with viral proteins nsp3 and nsp4 localized in discrete foci of ciliated cells showing punctuated and perinuclear distribution Banach et al., 2009), but also found at the base of the cilia. Immature virions are packaged in vesicles near the rough ER (RER), associated with double-membrane vesicles or areas of granular nucleocapsid material (Lednicky et al., 2013). Assembly seems to occur through sequential budding of virions from a nucleocapsid lined membrane, a pinching off, and then envelope acquisition, with further accumulation of mature virion-laden vesicles into RER or Golgi cisternae that might rupture and release virions into the cytosol, resulting in cell death through either necrosis or apoptosis (Orenstein et al., 2008;Banach et al., 2009).

Infection models in vivo
Precedents of in vivo infection models for HCoV-NL63 are scarce. In a previous study, Bently et. al. (2021) reported the use of K18-hACE2 mice. The virus was initially propagated in vitro in Caco-2 cells and inoculated intranasally in 8-10 weeks-old mice. The inoculation triggered a classical acute inflammatory response in lungs, and active viral propagation that reached plateau levels at 4 days post-inoculation. Moreover, a more recent study reported the use of immunocompromised (STAT1-/-) or sensitized (IFNAR-/-) mice transduced with the viral receptor hACE2 to generate an infection model for HCoV-NL63 (Liu et al., 2023). This study described that BALB/c mice were more susceptible to HCoV-NL63 infection than C57BL/6 mice, and both models were capable of produce specific CD8+ T cells against the virus N protein.

Virus-host interactions: induction and evasion of host innate immunity
The structural analysis of HCoV-NL63 S protein revealed that the RBD is buried or masked by glycosylation in a native trimer conformation, which might protect the virus from host immune recognition (Walls et al., 2016). However, when recognized by antigen-presenting cells, HCoV-NL63 can trigger a strong cytokine and chemokine mediated immune response, which self-limits the spread of the infection (Lister, 2009).
The S protein of HCoV-OC43, HCoV-HKU1, or SARS-CoV triggers ER stress, and activates unfolded protein response (UPR) in virus-infected cells (Favreau et al., 2009;Siu et al., 2014a). This signaling pathway eventually leads to apoptosis to limit viral replication (DeDiego et al., 2011;Favreau et al., 2009), although not yet studied in HCoV-NL63 (Chan et al., 2006). Apoptosis is also mediated by MAP kinases during CoV infection with HCoV-229E, MERS-CoV, and SARS-CoV (Kono et al., 2008;Mizutani et al., 2005;Yeung et al., 2016). The role of autophagy in CoV replication is not clear since virus can manipulate cell machinery for DMV formations . Infected cells with TGEV activates autophagy cells to protect from oxidative stress and avoid apoptosis (Zhu et al., 2016), but promotes viral replication in certain cases like in PEDV infection (Guo et al., 2017). The activation of autophagy depends on the activation of the microtubule-associated protein 1A/1B-light chain 3 (LC3) by the HCoV-NL63 PLP2 transmembrane protein (TM), independently of the catalytic site . Although the catalytic site has a robust deubiquitinating action on STING, RIG-I, TBK1, and IRF-3 to avoid IFN signaling activation, it is not required to antagonize the STING-mediated activation of IRF-3 Clementz et al., 2010;Nicholson et al., 2008;Sun et al., 2012). PLP2-TM interacts with Beclin1 to inhibit STING and impair the autophagosomes' maturation through interference with lysosome fusion . HCoV-NL63 PLP2-TM also promotes IFN inhibition through ubiquitination of the cellular oncoprotein MDM2. The stabilized form of MDM2 induces proteasomal degradation of p53, which is responsible for activating IRF7 (Yuan et al., 2015).

Antiviral strategies
Effective antiviral treatment is required when HCoV-NL63-infected patients undergo clinical complications. Despite the fact that intravenous administration of immunoglobulins seemed to be protective , they might not be available for high numbers of patients during outbreaks. Thus, different strategies have been proposed to suppress HCoV-NL63 replication by either targeting the virus or enhancing host antiviral mechanisms (Table 2).

Receptor blockers
The similarities between RBD of HCoV-NL63 and SARS-CoV have revealed a cross-inhibition of the ACE2 transduction by the S protein . For instance, in one study, a linear peptide (YKYRYL) derived from the SARS-CoV RBD was able to block the interaction of HCoV-NL63 S protein with the ACE2 receptor, being the KYR motif critical for binding (Struck et al., 2012). Other studies have demonstrated that the interaction of HCoV-NL63 S protein and ACE2 could be prevented by using N-(2-hydroxypropyl)− 3-trimethylammonium chitosan chloride (HTCC), a chitosan derivative with antisense oligonucleotides targeting residues of ACE2 responsible for its interaction with RBD (Milewska et al., 2013(Milewska et al., , 2016 or with phenolic compounds like caffeic acid from plant extract (Weng et al., 2019). The phytochemical compound Thymoquinone from Nigella sativa binds to ACE2 to block the entry of HCoV-NL63, SARS-CoV, and SARS-CoV-2 . In another example, methylene blue inhibits the interaction between SARS-CoV-2 and HCoV-NL63 spike protein and ACE2, IL-2R, TNFR, or CD40, including delta (B.1.617.2) variant through entry and replication inhibition (Chuang et al., 2022).
The use of splice-switching antisense oligonucleotides modulates alternative isoforms of ACE2 that limit infection (Rehman and Tabish, 2020). Alternately, the inhibition of ER glucosidases with iminosugars impairs the N-glycosylation and transduction of ACE2, but it also affects ACE2 expression and virion production (Zhao et al., 2015).
The heptad repeats (HR1 and HR2) of the S protein are critical during the fusion step of virus entry in conjunction with host proteases like TMPRSS2 for S protein priming. Therefore, they make attractive targets for the development and assessment of entry inhibitors. Heptad repeats' (HRs') interaction with the spike protein is also required to promote viral membrane fusion. Heptad repeat-derived peptides from HR2 are designed to block HR interaction , while the selective cysteine protease inhibitor-camostats antagonize TMPRSS2 (Hoffmann et al., 2020b;Kawase et al., 2012). Host defense peptides (HDPs) are typically 12-50 aa length expressed in neutrophils and mucosa, and serve as the first line of defense against foreign pathogens. Brilacidin, a synthetic HDP, demonstrates broad-spectrum antiviral against HCoV-OC43, − 229E, -NL63, and SARS-CoV-2 but not influenza or enterovirus  through blocking virus attachment and early entry targeting heparan sulfate proteoglycans (HSPGs).
With regard to replication blockers, Stenoparib is an inhibitor of cellular poly (ADP-ribose) polymerase (PARP) that blocks SARS-CoV-2 and HCoV-NL63 replication by reducing virus entry and complete inhibition of plaque formation. In addition, this drug showed a synergistic effect with remdesivir with more than 90% virus suppression on HCoV-NL63 (Stone et al., 2021).

Other antiviral strategies
Among the host factors, the interferon-inducible transmembrane (IFITM) proteins inhibit the entry of most CoVs, including HCoV-NL63, HCoV-229E, and SARS-CoV, probably through cholesterol accumulation on late endosomes (Huang et al., 2011;Wrensch et al., 2014). However, it has the opposite effect in other CoVs like HCoV-OC43 . These differences are explained by differential response of S proteins with specific motifs in IFITMs that control viral entry .
Additionally, immunization with spike proteins before SARS-CoV-2 immunization impedes the generation of SARS-CoV-2 neutralizing antibodies in mice, thus, they amplify cross-reactive antibodies that are non-neutralizing antibodies . Inconsistencies in humoral immunity response is that prior studies did not examine the level of human common coronaviruses antibodies in the same individual before and after SARS-CoV-2 infection.
In connection to cell mediated immunity, a cross-reactive T cell response to SARS-CoV-2 and HCoV-NL63 and OC43 was demonstrated in a cohort of convalescent patients. However, they also displayed an increased number of CD4+ T cells than unexposed group throughout the 9-month study period (Wirsching et al., 2022).
Beyond the cross-protection between coronaviruses, other studies reported a decrease on the antibody (IgG, and IgA) levels to respiratory syncytial virus and influenza virus was reported during COVID-19 lockdowns (Grobben et al., 2022). In this study, decreased antibody levels to common respiratory viruses in human milk was also observed during COVID-19 pandemic, which translated in lower passive immunity in children.

General conclusions
This article constitutes a comprehensive review on the infection mechanism and replication of HCoV-NL63, compiling the current HCoV-NL63-related research related to virus entry and replication mechanism, including virus attachment, endocytosis, genome translation, and replication and transcription, in comparison with other coronaviruses.
Moreover, cumulative knowledge on the susceptibility of different cells to HCoV-NL63 infection in vitro, was also reviewed, which is essential for successful virus isolation and propagation, and contribute to address different scientific questions from basic science to the development and assessment of diagnostic tools, and antiviral therapies. Finally, this review discussed different antiviral strategies that have been explored to suppress replication of HCoV-NL63, and other related human coronaviruses, by either targeting the virus or enhancing host antiviral mechanisms.

Declaration of Competing Interest
Authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability
Data will be made available on request.