Targeting host DEAD-box RNA helicase DDX3X for treating viral infections

Abstract

DDX3X or DDX3, a member of the DEAD (asp, glu, ala, asp) box RNA helicase family of proteins, is a multifunctional protein, which is usurped by several viruses and is vital to their production. To date, 18 species of virus from 12 genera have been demonstrated to be dependent on DDX3 for virulence. In addition, DDX3 has been shown to function within 7 of 10 subcellular regions that are involved in the metabolism of viruses. As such, due to its direct interaction with viral components across most or all stages of viral life cycles, DDX3 can be considered an excellent host target for pan-antiviral drug therapy and has been reported to be a possible broad-spectrum antiviral target. Along these lines, it has been demonstrated that treatment of virally infected cells with small molecule inhibitors of DDX3 blunts virion productions. On the other hand, DDX3 bolsters an innate immune response and viruses have evolved capacities to sequester or block DDX3, which dampens an innate immune response. Thus, enhancing DDX3 production or co-targeting direct viral products that interfere with DDX3’s modulation of innate immunity would also diminish virion production. Here we review the evidence that supports the hypothesis that modulating DDX3’s agonistic and antagonistic functions during viral infections could have an important impact on safely and efficiently subduing a broad-spectrum of viral infections.

1. Introduction

Antiviral therapies based on small molecule drugs have been developed into effective alternatives to vaccine resistance or for use in combination with vaccinations. Two broad categories of such therapeutics are direct-acting antiviral agents (DAAs) and those that are host-targeted antivirals (HTAs). An advantage to HTAs is the finding of host components that are required by a few or many viruses for cellular virion production and thus the potential of broad-spectrum antiviral agents may be achieved by targeting these components (Kumar et al., 2020). Given this, here we review a possible HTA therapy directed against a host DEAD (asp, glu, ala, asp)-box RNA helicase, DDX3X or DDX3, which has the potential of providing broad-spectrum antiviral therapy. We will focus on small molecule based HTA therapies directed against DDX3 including an inhibitor developed in our lab, RK-33, that has been designed to be a competitive inhibitor of DDX3’s ATPase site and has been demonstrated to abrogate its RNA helicase function (Bol et al., 2015; Heerma van Voss et al., 2018; Xie et al., 2016).

2. DDX3 is a multifunctional protein

DDX3 belongs to the DEAD-box ATPase-dependent RNA helicase family of proteins (Song and Ji, 2019; Soto-Rifo and Ohlmann, 2013) that are characterized by two highly conserved Rec-A like core domains with the defining conserved motifs of this family of helicases: Q, Walker I, Ia, Ib, Walker II, III, IV, V, & VI, which are involved with ATPase/RNA binding helicase activity (Fig 1). Contrary to these domains are the amino and carboxy regions of the same proteins that vary greatly in length and sequence between proteins and exhibit a lack of any regions of homology or similarity (Fig 2). These latter domains imbues each member of this family with distinct functions that are independent of unwinding RNA (Bourgeois et al., 2016). DDX3 shuttles from the cytoplasm to the nucleus (Brennan et al., 2018) and has been demonstrated in be involved with a diversity of functions in both compartments (Bourgeois et al., 2016). Examples, in broad terms, of DDX3 functions are: translation activation of nuclear and mitochondrial mRNAs (Geissler et al., 2012; Heerma van Voss et al., 2018; Lee et al., 2008; Soto-Rifo et al., 2012; Waldron et al., 2019); transport, storage, and translation inhibition of mRNA (stress granules) (Gaete-Argel et al., 2019; Heaton et al., 2019; Shih et al., 2012); nuclear export of mRNA (Heaton et al., 2019); mRNA splicing (Bourgeois et al., 2016; Soto-Rifo and Ohlmann, 2013); modulation of transcription (Bourgeois et al., 2016; Fullam et al., 2018; Heaton et al., 2019; Khadivjam et al., 2017; Schroder, 2011); lipid hemostasis (Tsai et al., 2017); regulation of endoplasmic reticulum stress (Adjibade et al., 2017); modulation of epigenetic modifications (Chen et al., 2017); and innate immune system activation (Heaton et al., 2019; Oshiumi et al., 2010b; Taschuk and Cherry, 2020). Given this broad array of functions and the fact that viruses require the functions of cellular components to reproduce it should perhaps not be considered remarkable that a multifunctional protein such as DDX3 is essential for the replication and assembly of several species of viruses, as detailed below.

Fig 1. Illustration of the nine highly conserved protein motifs within the Rec-A like domains of DEAD box RNA helicases.

Protein sequences of DDX3, DDX5, DDX17, DDX43, DDX53, DDX47, DDX54, DDX6, DDX18, DDX41, DDX19, and DDX59 were obtained from the NCBI database as Fasta formats. In all cases of multiple isoforms, isoform 1 was used. Alignment was done in ClustalX2 with default settings (Gap opening: 10, Gap Extension: 0.2, and protein weight matrix: Gonnet series) and shading in GENEDOC. In all cases the sequences are represented by single letter amino acid code. The conserved consensus motif sequences above their respective shaded alignments are given with abbreviations: x = any amino acid and b = branched chain amino acids. Designations of each motif (Q, Walker I, Ia, Ib, Walker II, III, IV, V, & VI) are above the consensus sequences and those involved in ATP binding/metabolism and RNA binding are indicated by arrows.

Fig 2. Amino and carboxy termini sequences of the DEAD box RNA helicases represented in Fig 1.

(A) Amino-termini sequences and (B) carboxy-termini sequences. In either case, no homology or similar sequence motifs were found during the alignment of the sequences as described in Fig 1.

3. Viruses interact with multiple cellular pathways associated with DDX3

It is important to note that viral infections dramatically disrupt normal cellular functions and cellular homeostasis (Glingston et al., 2019) by diverting a cell’s functioning solely towards virion production. Table 1 presents a list, in broad terms, of the major subcellular regions and organelles that DDX3 is associated with and have been identified as regions that viruses are required to interact with throughout their life cycles (Glingston et al., 2019; Ji and Li, 2020; Kumar et al., 2020).

Table 1:

Examples of DDX3’s functional association with cellular regions that are required for viral entry, processing, and egress from cells.

Subcellular Region	Associated Function	References
Plasma Membrane	TRPV4 Mediated Ca2+ Influx Releases DDX3 to Nuclear Translocation & Inhibition of Viral Nuclear Export/Translation	(Donate-Macian et al., 2018; Ji and Li, 2020; Kumar et al., 2020)
Nucleus	Exportin/Viral mRNA Nuclear Export	(Bourgeois et al., 2016; Frohlich et al., 2016; Heaton et al., 2019; Ji and Li, 2020; Kumar et al., 2020)
	Regulation of Epigenetic Modifiers	(Chen et al., 2017; Kumar et al., 2020)
	Transcription Viral Modulators	(Fullam et al., 2018; Schroder, 2011)
Cytoplasm	Translation Viral mRNA	(Geissler et al., 2012; Han et al., 2020; Lee et al., 2008; Soto-Rifo et al., 2012; Su et al., 2018; Waldron et al., 2019)
Lipid	Synthesis/Metabolism	(Li et al., 2013; Tsai et al., 2017)
MLOs	mRNA Storage	(Gaete-Argel et al., 2019; Shih et al., 2012)
MIT	Translation of MIT Protein/Modulation of OXPOS/ROS Generation	(Heerma van Voss et al., 2018)
ER	Translation of ER Protein	(Adjibade et al., 2017)

Abbreviations: ER – Endoplasmic Reticulum; MIT – Mitochondria; MLO – Membraneless Organelles (Stress Granules & Processing Bodies); TRPV4 – Transient Receptor Potential Vanilloid 4.

As noted (Table 1) and illustrated in Fig 3, in line with its multifaceted functions in cells, DDX3 has been found to be associated with 7 subcellular regions/organelles that are associated with a virus’ entry into a cell, nuclear/cytoplasmic processing/replication/translation/transcription, lipid envelope assembly, energy needs (mitochondrial), and egress from the cell. This is a very important finding as it indicates that DDX3 can be associated with many stages of viral life cycles, which increases the barrier of drug resistance, i.e., DDX3 is a multifunctional protein and viruses have evolved to utilize or block many of these functions making DDX3 essential for sustained virion production. Hence, if functions of DDX3 that are essential for multiple stages of virion production were abrogated (targeted) then the necessity of the co-evolution of multiple viral components that interact with DDX3 in a manner that allows a virus to remain fit for efficient propagation would be predicted to be slim, i.e., development of viral resistance to drugs targeting DDX3 would be improbable.

Fig 3. Illustrative summary of the multiple sub-compartments that DDX3 functions in and the association of viral components with the identical compartments.

A second arm of our rationale for selecting DDX3 for HTA therapy has been summarized in this section, i.e., once a cell has been infected by a virus it can no longer be considered as normal as it will have been converted into a virion factory with alterations to its gene expression repertoire and hence its transcriptome and proteome (Table 1 & (Claus and Liebert, 2014; El-Bacha and Da Poian, 2013; Gaete-Argel et al., 2019; Garcia-Sastre, 2017; Glingston et al., 2019; Hoffmann et al., 2017; Ji and Li, 2020; Kumar et al., 2020; Lange et al., 2019; Mohamed et al., 2020; Tanner et al., 2014). As such, targeting the functions DDX3, which has become an aberrantly required agent of virion production, may minimal impact the viability of normal uninfected cells. We hypothesize that a parallel with transformed cancer cells that present abnormal transcriptomes and proteomes can be considered. For example, there is convincing evidence to indicate that many viruses not only interact with various host proteins (Fung and Liu, 2019; Lim et al., 2016; van Hemert et al., 2008) but also replicate only in the S-phase of the cell cycle as viral replication proteins are regulated by cellular factors that are activated only in the S-phase (Schang, 2003). This has a direct association with DDX3, as DDX3 expression levels are important for the transition from G1 to S-phase of cell division (Ariumi, 2014) and a hallmark of cancer is that a large portion of cells in a tumor are rapidly dividing cells that have been tracked by scoring for the established division cycle protein Ki-67 (Pathmanathan and Balleine, 2013; Scholzen and Gerdes, 2000; Yang et al., 2018). Along these lines, we have demonstrated that in a number of cancer types, which have DDX3 is an essential component of the aggressiveness of these cancers and inhibition of DDX3 with a small molecule inhibitor, RK-33, kills cancer cells with no toxic effects on normal cells (Bol et al., 2015; Heerma van Voss et al., 2018; Xie et al., 2016).

4. DDX3 as a broad-spectrum antiviral target

To date, as listed in Table 2, 18 species of viruses grouped into 12 genera and 11 families have been reported as requiring DDX3 for efficient virion production and or sustained evasion of innate immunity (see references in Table 2). We searched for a few of the basic defining characteristics of these viruses in an attempt to identify common characteristics, such as type of genome (RNA or DNA), reliance on receptor entry into cells, enveloped or not, cytoplasmic or nuclear replication, 5’ 7-methylguanylate capped mRNA, and 3’ poly(A) appended mRNA, which might explain this broad reliance on DDX3. However, it is apparent that viruses with broadly different basic defining characteristics are represented in Table 2. For example, although most are single-stranded (ss) RNA viruses, double-stranded (ds) DNA viruses are also represented along with examples of segmented RNA viruses. The majority are packaged with a lipid envelop but three are not (Smith and Smith, 2019; Smyth and Martin, 2002), and 17 use protein specific receptor mediated cellular entry, but one does not (Edinger et al., 2014). For 13 of these, genome replication occurs exclusively in the cytoplasm, 3 replicate in the nucleus, and 2 replicate in a cytoplasmic stage coupled to a nuclear stage (Table 2). Obviously, the phylogenetic diversity represented in Table 2 explains the diversities enumerated here. In summary, the evidence presented in Table 1 & 2 indicates why DDX3 is an excellent candidate for HTA therapies across a broad and diverse spectrum of virus.

Table 2:

Viruses that require DDX3 for efficient sustained virion production.

					mRNA
Species (Genus)	Genome	Env.	P.S.R. M.E.	Cyto. or Nucl.	5′ Cap	3′ poly-A	References
HIV (Lentivirus)	2x +ssRNA	+	+	Cyto/Nucl	+	+	a
WNV (Flavivirus)	+ssRNA	+	+	Cyto	+	+	b
DENV (Flavivirus)	+ssRNA	+	+	Cyto	+	+	b
ZIKV (Flavivirus)	+ssRNA	+	+	Cyto	+	+	(Donate-Macian et al., 2018; Yang et al., 2020a)
JEV (Flavivirus)	+ssRNA	+	+	Cyto	+	+	(Li et al., 2014; Valiente-Echeverria et al., 2015)
HCV (Hepacivirus)	+ssRNA	+	+	Cyto	−	−	(Brai et al., 2016; Chatel-Chaix et al., 2013; Donate-Macian et al., 2018; Geissler et al., 2012; Li et al., 2013; Oshiumi et al., 2010a; Tsai et al., 2017; Valiente-Echeverria et al., 2015; Villareal et al., 2015)
EV-A71 (Enterovirus)	+ssRNA	−	+	Cyto	−	+	(Han et al., 2020; Su et al., 2018)
CV-B (Enterovirus)	+ssRNA	−	+	Cyto	−	+	(Quaranta et al., 2020)
NLV (Norovirus)	+ssRNA	−	+	Cyto	−	+	(Vashist et al., 2012)
HPIV-3 (Paramyxovirus)	-ssRNA	+	+	Cyto	+	+	(Heaton et al., 2019; Yang et al., 2020a)
RSV (Pneuovirus)	-ssRNA	+	+	Cyto	−	+	(Yang et al., 2020a)
FLUAV (α-Influenzavirus)	8x -ssRNA	+	−	Nucl	+	+	(Thulasi Raman et al., 2016)
LASV (Mammarenavirus)	2x +/−ssRNA	+	+	Cyto	+	−	(Loureiro et al., 2018)
JUNV (Mammarenavirus)	2x +/−ssRNA	+	+	Cyto	+	−	(Loureiro et al., 2018)
LCMV (Mammarenavirus)	2x +/−ssRNA	+	+	Cyto	+	−	(Loureiro et al., 2018)
HBV (Orthohepadnavirus)	dsDNA	+	+	Cyto/Nucl	+	+	(Megahed et al., 2020)
HSV-1 (Simplexvirus)	dsDNA	+	+	Nucl	+	+	(Khadivjam et al., 2017)
HCMV (Simplexivirus)	dsDNA	+	+	Nucl	+	+	(Cavignac et al., 2015; Lenarcic et al., 2015)
VACV (Orthopoxvirus)	dsDNA	+	+	Cyto	+	+	(Oda et al., 2009)

^{a -}(Brai et al., 2016; Brai et al., 2020b; de Bisschop et al., 2019; Frohlich et al., 2016; Gringhuis et al., 2017; Valiente-Echeverria et al., 2015; Yasuda-Inoue et al., 2013);

^{b -}(Brai et al., 2016; Brai et al., 2019a; Geissler et al., 2012; Valiente-Echeverria et al., 2015; Yang et al., 2020a);

^{c -}(Brai et al., 2016; Chatel-Chaix et al., 2013; Donate-Macian et al., 2018; Geissler et al., 2012; Li et al., 2013; Oshiumi et al., 2010a; Tsai et al., 2017; Valiente-Echeverria et al., 2015; Villareal et al., 2015);

Abbreviations: +ssRNA – positive sense single-stranded RNA; -ssRNA – negative sense single-stranded RNA; 2x – two genomic segments; 8x – eight genomic segments, +/−ssRNA – ambisense single-stranded RNA; HIV – Human Immunodeficiency virus; WNV – West Nile virus; DENV – Dengue virus; ZIKV – Zika virus; JEV – Japanese Encephalitis virus; HCV – Hepatitis C virus; EV-A71 – Enterovirus A71 virus; CV-B – Coxsackie virus group B; NLV – Norwalk virus; HPIV-3 – human Parainfluenza Type-3; RSV – Respiratory Syncytial virus; FLUAV – Influenza A virus; LASV – Lassa virus; JUNV – Junin virus; LCMV – Lymphocytic Choriomeningitis virus; HBV – Hepatitis B virus; HSV-1 – Herpes Simplex Type-1 virus; HCMV – Human Cytomegalovirus; VACV – Vaccinia virus; Env. – lipid envelope; P.S.R.M.E. – protein specific receptor mediated endocytosis; Cyto. or Nucl. – cytoplasmic or nuclear replication; 5’ Cap – 7-methylguanylate cap.

5. Small molecule inhibitors of DDX3 abrogate a broad-spectrum of viral infections.

The viral species listed in Table 2 continue to have a devastating impact on human health by causing periodic deadly epidemics; e.g., West Nile virus (WNV), Dengue virus (DENV), Zika virus (ZIKA), Japanese Encephalitis virus (JEV), Influenza A virus (FLUAV), Lassa virus (LASV), Junin virus (JUNV), and Lymphocytic Choriomeningitis virus (LCMV), and persistent chronic illnesses; e.g., Human Immunodeficiency virus (HIV), Hepatitis C virus (HCV), and Human Cytomegalovirus (HCMV). As such, safe effective alternative antiviral treatments that overcome drug resistance as well as robust vaccines that control these infections are actively being sought (Brai et al., 2016; Ji and Li, 2020; Kumar et al., 2020). Along these lines, a goal is to develop safe HTA therapies as monotherapies or combination therapies, i.e., combinations of two or more HTA drugs or combinations of HTA therapeutics with DAA drugs, which can overcome drug resistance (Ji and Li, 2020; Kumar et al., 2020; Yang et al., 2020b). Given that DDX3 has been found as a common required host mediator for the efficient sustained production of the viruses listed in Table 2 and that vaccines or DAA antiviral therapies that are robust enough, i.e., are not limited by drug resistance, to effectively manage and ideally eliminate these viruses are lacking has brought many investigators to consider or actively pursue the targeting of DDX3 as a broad-spectrum antiviral strategy (Brai et al., 2020a; Brai et al., 2016; Brai et al., 2019a; Brai et al., 2020b; Brai et al., 2019b; Kukhanova et al., 2020; Quaranta et al., 2020; Schroder, 2011; Valiente-Echeverria et al., 2015; Yang et al., 2020a). For example, Botta and colleagues (Brai et al., 2020a; Brai et al., 2016; Brai et al., 2019a; Brai et al., 2020b; Brai et al., 2019b) have reported the development of several small molecule inhibitors that targets DDX3’s RNA binding site (Fig 4A). Using such inhibitors, these researchers have repeatedly demonstrated very effective anti-viral targeting of DDX3 in cells infected with HIV, HCV, DENV, and WNV (all positive-sense single stranded RNA viruses (+ssRNA – Table 2), which were non-toxic, i.e., in vitro half-maximum cytotoxic concentrations (CC₅₀) of ~200 x the half-maximal effective concentrations (EC₅₀), which were in the sub- to low-micromolar range as well as being safe in vivo with no brain, liver or kidney toxicity. Nevertheless, this strategy of targeting DDX3’s RNA binding site did not show anti-viral activity against negative-sense ssRNA (-ssRNA) viruses (Quaranta et al., 2020). However, a small molecule, originally developed as an anti-cancer therapeutic (Bol et al., 2015; Heerma van Voss et al., 2018; Xie et al., 2016), RK-33 (Fig 4B), that binds DDX3’s ATP binding site and inhibits its ATPase dependent RNA helicase activity, has been demonstrated to effectively (EC₅₀ in the low micromolar range) abrogate virion production in both +ssRNA (DENV, WNV, and ZIKA) and -ssRNA (RSV and hPIV-3) infections (Yang et al., 2020a) (Table 2). Thus, both strategies: targeting RNA binding or ATPase activity, have produced similar results except that the RK-33 based targeting has the advantage of increasing the spectrum of viruses that can be suppressed. RK-33’s development was based on a rational design approach and has been demonstrated as being specific for DDX3 (Bol et al., 2015). For instance, RK-33 does not bind other members of the DEAD box RNA helicase family, despite the highly similar ATP binding sites across this family (Bourgeois et al., 2016). Moreover, the specificity of RK-33 to DDX3 was independently verified by Yang et al. (Yang et al., 2020a) using highly accurate physical biochemical methodologies: analytical ultracentrifugation and isothermal titration calorimetry, demonstrating that RK-33 binds DDX3 within the Walker I motif that contributes to ATP-binding (Fig 1), resulting in inhibition of DDX3’s ATPase and RNA unwinding activities. RK-33’s specificity is important to note because an objection to targeting the ATP binding pocket has been raised as it was argued such drugs would exhibit deleterious off target effects due to the similarity of ATP binding pockets across several classes of ATP binding proteins (Brai et al., 2016; Riva and Maga, 2019). Although, it has not been demonstrated that RK-33 will not bind any of the many DEAD box helicases or any of the ATP binding domain proteins, our animal studies have indicated that RK-33, even at significantly higher dose than the therapeutic window, had minimal or no toxicity, which indicates that it likely has minimal deleterious off target effects (Bol et al., 2015).

Fig 4. Examples of small molecule inhibitors of DDX3.

(A) Drugs that bind at the RNA binding site (Brai et al., 2016). (B) Drugs that are competitive inhibitors ATP binding at the ATP binding site (Bol et al., 2015).

Given the repeated evidence from independent laboratories that targeting DDX3’s RNA helicase activity has provided an effective safe abrogation of the production of a diversity of viruses indicates that further investigations aimed at developing drugs that block DDX3’s necessary interactions with viral propagation is warranted. As described above, to date, the focus has been on DDX3’s RNA helicase activity by either directly targeting the RNA binding site or its ATP binding site. However, the DEAD box family of proteins has diverse and generally non-overlapping functions, which are defined by their binding partners (Bourgeois et al., 2016; Shih et al., 2012; Soto-Rifo and Ohlmann, 2013). There is strong evidence that DDX3 interacts with several binding partners, such as chromosome region maintenance-1 (CRM1) (de Bisschop et al., 2019; Frohlich et al., 2016), poxvirus protein K7 (Oda et al., 2009), hepatocyte nuclear factor-4 (HNP4) (Tsai et al., 2017), small heterodimer partner (SHP) (Tsai et al., 2017), Tank-binding kinase1 (TBK1) (Khadivjam et al., 2017), calmodulin (CaM) (Khadivjam et al., 2017), transient receptor potential vanilloid-4 (TRPV-4) (Donate-Macian et al., 2018), ribosomal protein L13 (Han et al., 2020), dengue virus capsid protein (Kumar et al., 2017), eukaryotic translation initiation factor 4E (eIF4E) (Shih et al., 2012), eIF4F (Soto-Rifo et al., 2012), eIF3 (Geissler et al., 2012), 80s ribosomes (Geissler et al., 2012), p53 (Chen et al., 2017), retinoic acid inducible gene-1 (Oshiumi et al., 2010b), protein phosphatase 2A-C (PP2A-C) (Wang et al., 2017), and interferon-β promotor stimulator-1 (IPS-1) (Wang et al., 2017), through either its amino or carboxy termini, which thus, provides opportunities for the design of small molecule regents that block these interactions and inhibit those viruses that depend on the functions that such binding partners impart to their production.

6. DDX3’s function in innate immunity provides another anti-viral targeting strategy

DDX3 has been shown to be an integral component of an innate immune response to viral infections and, as such, is antiviral (Fullam et al., 2018; Heaton et al., 2019; Oshiumi et al., 2010b; Schroder, 2011; Wang et al., 2017); e.g., it contributes to the mediation of viral detection and functions in the induction of the type I interferon (IFN-I) pathway (Schroder, 2011). The necessity of an innate immune response to viral infections has led some investigators to caution against the targeting of DDX3 as an antiviral strategy as this might also comprise an innate immune response (Schroder, 2011). However, several counter points to this rationale can be made. First, viruses have evolved to block DDX3’s innate immunity functions or sequester DDX3 in subcellular virion producing regions, which provides an in situ evasion of an immune response (Gringhuis et al., 2017; Oda et al., 2009; Oshiumi et al., 2010a; Taschuk and Cherry, 2020). Second, the antiviral strategy targeting of DDX3 has been, as described above, based on the targeting of DDX3’s helicase activity, i.e., ATP or RNA binding sites, but these activities are independent and dispensable to DDX3’s innate immune antiviral functions (Khadivjam et al., 2017) and thus, these targeted strategies would be unlikely to affect innate immunity while simultaneously abrogating virion production. Third, during arenaviral infections DDX3 suppresses the IFN-I pathway (Loureiro et al., 2018) without being blocked or sequestered. However, the same report presented evidence that DDX3 promotes arenavirus replication and transcription in a helicase, i.e., ATPase/RNA binding, dependent manner. This latter finding is in line with all other reported evidence (as reviewed here) that DDX3 is generally directly utilized by viruses for virion production and in the case of arenaviruses the suppression of DDX3 mediated innate immunity is brought about in a manner other than the blocking or sequestering of DDX3 but with the same result.

It can be concluded that the most effective targeting of DDX3 would require drugs that do not solely target the ATP or RNA binding sites but also target domains outside of these with the aim of rectifying DDX3’s antiviral innate immunity activity. This concept of a dual drug approach can be extended to a combination drug approach where combinations of HTA antiviral agents might be used or DDX3 centered HTAs combined with DAAs could be advantageous (Kumar et al., 2020).

7. Future research

7.1. Rationale for a potential role for DDX3 in SARS-CoV-2 pathogenesis

Evolutionary studies have revealed that SARS-CoV-2 is closely related to SARS-CoV with a nearly identical genome size and organization as well as replication (Coronaviridae Study Group of the International Committee on Taxonomy of, 2020; Petrosillo et al., 2020; Stammler et al., 2011; van den Born et al., 2005; Yang and Leibowitz, 2015; Yoshimoto, 2020). SARS-CoV has been extensively studied and it is known that its replication initially proceeds through a 5’ cap-dependent translation using the host’s translation complexes (Cencic et al., 2011). As such, SARS-CoV-2 also has a 5’ capped mRNA genome that requires cap-dependent translation (Gordon et al., 2020). As pointed out above, there is strong evidence that DDX3 forms functional complexes with several eukaryotic translation initiation co-factors including eIF4E, eIF4E, eIF4G and eIF3 (Geissler et al., 2012; Lee et al., 2008; Shih et al., 2012; Soto-Rifo et al., 2012) and has been demonstrated as contributing a required function in promoting translational as a part of the 80S translation initiation complex (Geissler et al., 2012). Moreover, this is not the only example of a potential overlap between DDX3’s function and subcellular locations with those of SARS-CoV-2 virion processing/production. A recent detailed study (Gordon et al., 2020) of SARS-CoV-2 resulted in a comprehensive depiction of the SARS-CoV-2 interactome along with the associated subcellular compartments and organelles that fits the general schema laid out in Table 1 and illustrated in Fig 3. Thus, SARS-CoV-2 interactome maps reveal signaling pathways and organelles, which, as noted above, have been demonstrated as being involved with the functional and regulatory activities of DDX3 (Table 1 and Fig 5). Moreover, the same study found that the SARS-CoV-2 interactome was highly similar to that of WNV along with a high similarity with Zika and HIV and these three viruses usurp DDX3 during their infection cycles (Table 2). It is also noteworthy, in this context, that two members of the DEAD family, DDX1 and DDX5, have been shown to participate in viral replication of SARS-CoV ((Chen et al., 2009; Wu et al., 2014; Xu et al., 2010)) and further, that DDX3 has been demonstrated as binding with DDX5 forming a complex that functions in shuttling mRNP export from the nucleus to the cytoplasm (Choi and Lee, 2012; Riva and Maga, 2019). Consequently, although not yet demonstrated, given the structural similarity of DEAD box family members, it is not unreasonable to hypothesize that an antiviral strategy in SARS-CoV-2 infections might be achieved by targeting DDX3 to diminish or abolish cellular virion production of SARS-CoV-2.

Fig 5. Association of DDX3 in SARS-Cov2 host protein-protein interactome map.

7.2. Rationale for the consideration that DDX3 centered HTA strategy might be extended to other viral infections

The viruses listed in Table 2 cover a broad diversity of characteristics and some of these will be common to several other human viral infections. This is important to note because the rationale discussed in this review need not be limited to only the viruses listed in Table 2. We suggest that continued molecular characterization of the motifs/domains of DDX3 and its viral partners (Oda et al., 2009) will provide the capability to use rational design strategies to develop safe highly specific drugs that block the advantages that DDX3 supplies to virion production. Given DDX3’s critical function in inducing an innate immune response (Fullam et al., 2018; Meier-Stephenson et al., 2018; Schroder, 2011; Taschuk and Cherry, 2020), the development of reagents that bolster this activity and subvert the viral programs that thwart it is warranted.

• Human host DEAD box RNA helicase, DDX3, has been demonstrated as a necessary component for the replication of several viruses.

• Specific targeting of multifunctional DDX3 has the potential of providing a broad-spectrum antiviral therapy strategy.

• DDX3 associated sub-cellular organelles have been identified in SARS-CoV-2 interactome maps.

• Functional identification of antiviral and proviral activities of DDX3 can be exploited for defined virus treatment.

• Small molecule inhibitors of DDX3 have been demonstrated to reduce viral titers and could have clinical applications as potential host-targeted antivirals.

Abbreviations:

b

branched chain amino acid

x

any amino acid