Modulation of Endosome Function, Vesicle Trafficking and Autophagy by Human Herpesviruses

Human herpesviruses are a ubiquitous family of viruses that infect individuals of all ages and are present at a high prevalence worldwide. Herpesviruses are responsible for a broad spectrum of diseases, ranging from skin and mucosal lesions to blindness and life-threatening encephalitis, and some of them, such as Kaposi’s sarcoma-associated herpesvirus (KSHV) and Epstein–Barr virus (EBV), are known to be oncogenic. Furthermore, recent studies suggest that some herpesviruses may be associated with developing neurodegenerative diseases. These viruses can establish lifelong infections in the host and remain in a latent state with periodic reactivations. To achieve infection and yield new infectious viral particles, these viruses require and interact with molecular host determinants for supporting their replication and spread. Important sets of cellular factors involved in the lifecycle of herpesviruses are those participating in intracellular membrane trafficking pathways, as well as autophagic-based organelle recycling processes. These cellular processes are required by these viruses for cell entry and exit steps. Here, we review and discuss recent findings related to how herpesviruses exploit vesicular trafficking and autophagy components by using both host and viral gene products to promote the import and export of infectious viral particles from and to the extracellular environment. Understanding how herpesviruses modulate autophagy, endolysosomal and secretory pathways, as well as other prominent trafficking vesicles within the cell, could enable the engineering of novel antiviral therapies to treat these viruses and counteract their negative health effects.


Introduction
Cell entry is a critical step for initiating the replication cycle of viruses and for the evasion of host antiviral responses. Similarly, the egress of newly synthesized infectious particles from the infected cell is a determinant step for virus perpetuation and spreading onto other cells and organisms [1][2][3][4]. Because viruses are obligate intracellular pathogens, the manipulation of host cellular molecular determinants and processes is key for their perpetuation [5][6][7][8]. To achieve this, viruses have undergone natural selection over time and co-evolved with their hosts, allowing them to harness fundamental cellular processes in their favor. Among the relevant cellular processes that impact viral infection and yield from infected cells are those related to membrane trafficking, such as endolysosomal and º Figure 1. Mechanisms used by human herpesviruses to enter host cells. (A) Clathrin-mediated entry for herpesviruses in general. Herpesviruses induce clathrin oligomerization, together with the recruitment of host factors and adapter proteins, which also induce the polimerization of actin filaments and lastly the invagination of the cell membrane, thanks to the GTPase activity of dynamin. (B) Cell entry mechanisms exploited by human herpesviruses type 1 and type 2 (HSV-1 and HSV-2). The binding of viral glycoproteins B and D to their ligands triggers dimerization of the gH/gL complex, which causes the activation of the fusogenic activity of gB, enabling the release of the viral capsid into the cytosol. On the other hand, viral capsid delivery through the endosomal pathway has also been described, which is associated with endosomes having a low pH. (C) The mechanism used by varicella-zoster virus (VZV) for entering cells is also mediated by the gH/gL complex, but clathrin-mediated entry has also been described for this virus. (D) The Epstein-Barr virus (EBV) has been described to enter the cell through its binding to the EPHA2 receptor. Alternatively, other entry pathways exist, such as that mediated by BMRF-2 binding to integrins, causing gH/gL complex activation and gB-mediated fusion of membranes; lipid raft-mediated endocytosis; or viral entry through micro-and micropinocytosis. (E) The human cytomegalovirus (HCMV) mainly enters the cell via the activation of the gH/gL complex, triggering gB-mediated membrane fusion, but a low pH-dependent endocytosis mechanism has also been reported, involving a viral protein complex binding to host OR14I1. (F) The entry mechanisms exploited by human herpesviruses 6 (HHV-6; HHV-6A and HHV-6B) mainly occurs by endocytosis through lipid rafts. (G) The only entry mechanism described so far for the human herpesvirus 7 (HHV-7) is mediated by a fusogenic process mediated by the gH/gL complex. No endocytic pathway has been described yet for this virus. (H) Finally, Kaposi's sarcoma-associated herpesvirus (KSV) infects the cells through a gB-mediated membrane fusion pathway, clathrin-mediated endocytosis and may also induce micro-and macropinocytosis.
Although varicella-zoster virus (VZV) entry into target cells is not fully understood, it is presumed to occur similarly to HSV-1 and HSV-2 through a membrane fusion process, in which several envelope glycoproteins participate [64]. To carry on the initial attachment event, gB of VZV has been reported to bind to myelin-associated glycoprotein [65]. Other undetermined viral envelope proteins are also predicted to interact with the mannose-6phosphate (M6P) receptor [66]. The interaction of gB of VZV with MAG causes gH/gL to act as the core fusogenic complex, promoting viral membrane fusion with the host lipid bilayer [65].
Alternatively, VZV may enter the cell through an endocytosis process mediated by clathrin, in which the tyrosine-based endocytosis motif (YNKI) in the cytoplasmatic tail of gH is required [67]. Interestingly, this sequence has been shown to be the main component responsible for regulating the fusogenic activity of gH and cell-to-cell [68].
Infection with EBV virus seems to have a preferable tropism towards epithelial and B lymphocytes [69]. Recently, ephrin receptor A2 (EPHA2) was identified as a key host factor, required as the fusion receptor for membrane fusion-dependent viral entry into epithelial cells, whereas knocking out integrins did not affect EBV infectivity in HEK293 cells [70]. Previously, it was described in epithelial cells that EBV BMRF-2 protein interacts with αv or β1 integrins to facilitate basolateral infection [71]. After BMRF-2 interaction with the abovementioned integrins, EBV uses the gH/gL complex to interact either with the αvβ6 or αvβ8 integrins as receptors to trigger a conformational change in gB for membrane fusion and virus internalization to occur [72]. Interestingly, EBV has also been reported to use an alternative route for entering nasopharyngeal epithelial cells, which is lipid raftmediated endocytosis or micropinocytosis [73]. Interestingly, the route used for infecting oral epithelial cells is transcytosis, through apical macropinocytosis, whereas viral exit occurs via basolateral clathrin-mediated endocytosis [71].
On the other hand, Epstein-Barr virus (EBV) infection of B cells involves endocytosis through the fusion of the viral membrane with endosomal membranes [70,74]. The viral glycoprotein gp350 binds to CD21 on the B cell surface as an initial attachment [75] and thereafter glycoprotein gp42 forms a stable complex with the gH/gL complex, promoting them to bind to the human leukocyte antigen class II (HLA-II) and triggering the fusogenic activity of gB with consequent viral and host membranes fusing, allowing the entry of viral components into the cell [76,77].
Human cytomegalovirus (HCMV) entry into host cells bears a resemblance with HSVs in regards to the fusion process, with glycoprotein gB mediating viral attachment to heparan sulfate glycosaminoglycans [78,79]. Nevertheless, HCMV gB has also been shown to have the ability to engage other receptors, such as tetherin [80], platelet-derived growth factor-α receptor (PDGF-Rα) [81] and integrins such as αvβ3 [82,83]. The gH-gL complex also participates in HCMV entry by triggering virus and cell membrane fusion through gB. However, in endothelial and epithelial cells, but not fibroblasts, this complex requires the formation of a pentamer with three other viral proteins, namely, UL128, UL130 and UL131, to be functional [84,85]. Furthermore, it has been shown that HCMV can mediate its entry into epithelial and endothelial cells through the interaction of the viral fusion protein complex with the host receptor OR14I1, followed by a low-pH dependent-endocytosis process [86,87].
The distinct members of the roseoleovirus genus human herpesvirus 6, namely, HHV-6A and HHV-6B, display differences compared to the other herpesviruses detailed above. These viruses use a different configuration of glycoproteins for cell attachment and entry during their lytic cycles [88]. Although they also use gB for attaching to heparan sulfates, gH forms a complex with gL and gQ, which oversees the recognition and attachment of the virus to the ubiquitous receptor molecules CD46 for HHV-6A, and CD134 for HHV-6B [89]. This sequence of molecular interactions then triggers the fusion of the viral and cellular membranes [89][90][91][92][93]. Current evidence suggests that following cell surface binding, HHV-6A and HHV-6B glycoprotein complexes gH-gL-gQ mediate cell entry primarily via endocytosis in T cells (T lymphoblastoid cell line) [94,95], seemingly through lipid rafts [96]. Consequently, it has been reported that endocytosed virions travel through the cytosol, where, by unknown means, the capsids suffer de-envelopment, which grants them access to the external nuclear membrane in the cell [97]. Human herpesvirus 7 uses a similar entry mechanism to HHV-6A and HHV-6B, with the viral proteins gB and U100 both involved in the attachment of the virus to heparan sulfate proteoglycans on the cell surface [98,99], or alternatively to CD4 by means of a yetunknown viral protein [100]. Notably, an alternative mechanism of entry via endocytosis has not been described yet for HHV-7 [98,99].
Finally, Kaposi's sarcoma-associated herpesvirus (KSHV) also uses glycoproteins to promote viral entry through clathrin-coated mediated endocytosis. In order to do this, KSHV requires the viral gH/gL complex to bind to the ephrin-binding site of the EPHA2 receptor [101]. Consistent with this notion, an ephrin antagonist applied to epithelial or endothelial cells inhibits viral entry [102]. KSHV may also bind to the EPHA4 receptor in epithelial cells with higher efficiency than the EPHA2 receptor [103]. Notably, KSHV has also been observed to preferentially enter endothelial cells via micropinocytosis [104] by inducing tyrosine phosphorylation of c-Cbl to promote the translocation of virus receptors, such as the integrin αVβ3, into lipid rafts [105]. Furthermore, in early infection events, the adaptor protein c-Cbl interacts with actin and ubiquitinates the myosin light chain IIA, thus enabling bleb formation, which in turn allows for macropinocytosis-mediated viral entry [106].
Overall accumulating evidence reveals multiple entry mechanisms into cells by herpesviruses, with more than one type of entry for single viruses in many cases ( Figure 1). Moreover, there is evidence that some herpesviruses can trigger actin remodeling, which causes ruffling of the host plasma membrane and macropinocytosis for endocytic uptake [73,84,107,108]. In these cases, virus-containing vesicles consequently acidify, mature and fuse with late endosomes or lysosomes [39]. Nevertheless, it remains to be elucidated which signals determine that a herpesvirus prefers to use the endocytic pathway for entry or membrane fusion processes. Interestingly, the essential viral molecular determinants participating in membrane fusion and virus entry for some herpesviruses, such as VZV and HHV-7, are still unknown.

Autophagosome Function
Autophagy consists of a cellular process that promotes the elimination of damaged or senescent organelles and proteins to maintain cellular homeostasis and prevent metabolic deregulation [5,27]. Furthermore, autophagy is important in cellular antiviral responses because it can help cells to control virus takeover, promote cell survival, favor antigen presentation, regulate the initiation of inflammatory cytokine responses and target pathogens for degradation [109]. Some of the immune-modulatory aspects of autophagy are accomplished by Toll-like receptors (TLRs), pathogen recognition receptors (PRRs), which recognize pathogen-associated molecular patterns (PAMPs) from the virus that are present in endosomes and belong to the autophagosome pathway, causing signaling cascades with different outcomes [110]. Importantly, autophagy participates in providing the cell with a source of virus-derived antigens that can be loaded onto major histocompatibility molecules (MHC) that are involved in antigen presentation to immune cells [111]. This is achieved by autophagosomes fusing with multivesicular compartments containing MHC class I (MHC-I) molecules that enable antigen presentation to CD8+ T cells [112]. Furthermore, fusion of autophagosomes with MHC class II (MHC-II)-containing compartments allows the infected cell to perform type-2 cross-presentation, in which case intracellular antigens are targeted for presentation to CD4+ T cells [113].
During autophagy, small lipid bilayers that enclose diverse contents in the cytosol are recruited to form an autophagosome [114]. The engulfed contents are then targeted for degradation by delivering them to lysosomes [115]. The sequential steps that mediate the synthesis of autophagosomes relate to a series of enzymatic reactions involving, among others, autophagy-related (ATG) proteins [116]. In mammalian cells, the initiation of autophagosomes begins with crescent-shaped pre-autophagosome buds, termed phagophores, that hail from the endoplasmic reticulum (ER) and require ATG13 to form a complex with ULK1 and ULK2 [115]. Afterwards, nucleation occurs due to ULK1 phosphorylation, which in turn activates Beclin-1, the ATG6 orthologue in mammals that (along with vacuolar protein sorting 15 and PIK3C3/Vps34) forms a membrane-bound complex that acts as a class III phosphatidylinositol-3 kinase (PI3K) [117,118]. This PI3K complex phosphorylates phosphatidylinositol into phosphatidylinositol-3-phosphate (PIP3), which is recruited to the budding autophagosome membrane [119]. This, in turn, leads to an elongation step; as the phagophore becomes larger the complex recruits microtubule-associated light chain 3 (LC3) and other ATG proteins, such as ATG5, ATG12 and ATG16 [120]. Afterwards, LC3 is primed by E1-like enzyme ATG7 for cleavage by protease ATG4 and converts it into LC3-I [121,122]. Finally, autophagosome maturation is achieved by the action of an E2-like enzyme, ATG3, which allows LC3-I lipidation and its transformation into LC3-II through a ubiquitin-like mechanism [123]. LC3-II is the main membrane-bound protein that characterizes mature autophagosomes capable of fusing with lysosomes, with which the content is combined to allow the degradation of organelles and proteins by acidic proteases, such as cathepsins [124,125]. Rab7, SNARE proteins, ATG14 and ATG6 also play roles in autophagosome-lysosome fusion, with the latter also participating to some extent in endosomal membrane trafficking [118,126,127]. Furthermore, autophagosomes can fuse with endosomes to create intermediary membrane structures termed amphisomes, which have been described to form by reactive oxygen species generated by NADPH oxidase-containing endosomes that recruit LC3-containing organelles [128].
Depending on the herpesvirus species, the autophagy pathway may be suppressed in infected cells to promote virus fitness and replication (e.g., HSV-1 and HSV-2), or may be induced to harness its components in order to promote viral replication and yield [27]. Further details on how these viruses modulate autophagy are described below.

Herpesvirus Modulation of Autophagosome Function
Human herpesviruses have been described to have the capacity to selectively modulate autophagosome formation and generate different cellular outcomes thereof ( Figure 2) [5,110,129]. In macrophages, HSV-1 infection causes the formation of autophagosomes, which can process viral proteins prior to their delivery to the proteasome, thus enhancing presentation in MHC-I molecules [130]. On the contrary, HSV-1 ICP34.5 binds to beclin-1, which results in autophagy inhibition in neurons, conferring this virus with neurovirulence properties in mouse models [131]. Furthermore, ICP34.5 has been demonstrated to have an antagonistic function against PKR-induced eIF2α phosphorylation, which aids in preventing autophagy [132] and degrading HSV-1 viral proteins [133]. Through ICP34.5, the inhibition of autophagosome function in dendritic cells allows HSV-1 to prevent viral antigen presentation in MHC-II molecules, thus reducing the capacity of DCs to stimulate and induce the proliferation of CD4+ T cells [134]. ICP34.5 of HSV-2 shares common features with its homolog in HSV-1, such as promoting neurovirulence [89], but for the former, this viral protein allows basal autophagy to occur, which seems to be required for the infection of human corneal epithelial cells [135]. Nevertheless, autophagy is key for enabling protection against HSV-2, as knocking out ATG5 in DCs causes mice to succumb to HSV-2 intravaginal infection [136]. At present, it is currently unknown how viral promotion of the persistence of basal autophagy supports HSV-2 infection and replication. cumb to HSV-2 intravaginal infection [136]. At present, it is currently unknown how viral promotion of the persistence of basal autophagy supports HSV-2 infection and replication. Afterward, ULK1 phosphorylation activates the beclin-1 protein and triggers the recruitment of proteins that finally recruit the class-III phosphatidylinositol-3 kinase to the autophagosome membrane. The activation of beclin-1 is inhibited by the viral proteins BCL2, TRS1 and ICP34.5 of Kaposi's sarcoma-associated herpesvirus (KSHV), human cytomegalovirus (HCMV) and herpes simplex virus (HSV), respectively. (C) The early autophagosome recruits the LC-3 protein, a process that is enhanced by VZV, which triggers the maturation of the autophagosome due to the lipidation of LC-3 into LC-3-II. The latter step has been shown to be enhanced by KSHV. (D) Ultimately, autophagy concludes with fusion of a mature autophagosome with a lysosome. This process has been shown to be inhibited by VZV and the protein RAB7A by KSHV.
On the other hand, VZV and EBV potentiate the formation of autophagosomes in virus-infected cells [137]. Indeed, VZV readily induces LC3B (LC3 isoform)-containing autophagosome formation in human skin vesicles, which has been reported to help VZV-related glycoprotein biosynthesis and yield increased viral titers [138]. However, although VZV does not prevent autophagosome formation, it can inhibit late-stage autophagic fluxes, presumably by inhibiting autophagosome and lysosome fusion, which prevents viral degradation and increases viral titers [139]. Nevertheless, it remains yet to On the other hand, VZV and EBV potentiate the formation of autophagosomes in virus-infected cells [137]. Indeed, VZV readily induces LC3B (LC3 isoform)-containing autophagosome formation in human skin vesicles, which has been reported to help VZVrelated glycoprotein biosynthesis and yield increased viral titers [138]. However, although VZV does not prevent autophagosome formation, it can inhibit late-stage autophagic fluxes, presumably by inhibiting autophagosome and lysosome fusion, which prevents viral degradation and increases viral titers [139]. Nevertheless, it remains yet to be determined what mechanism VZV uses to promote autophagy in infected cells [140,141].
In turn, EBV possesses a latent-lytic switch, which during latent infection activates host PERK and IRE-1 unfolded protein response (UPR) elements while regulating the proliferation of virus-infected B cells [137,142]. Alternatively, the EBV oncoprotein EBNA3C increases autophagy markers in infected cells through the upregulation of tumor suppressor Cells 2021, 10, 542 9 of 23 genes, which results in increased survival in transformed B cells [143]. However, increased EBV replication and escape from Burkitt's lymphoma cells occurs when autophagy is inhibited due to the early lytic gene products BZLF1 and BRLF1 [144]. EBNA1 was observed to accumulate in autophagosomes prior to lysosomal processing for antigen presentation in MHC-II during B cell latent infection [145]. EBV uses an unknown mechanism to downregulate reactive oxygen species (ROS) production in infected cells, with a concomitant reduction in mitochondrial biogenesis, which impairs monocyte differentiation and triggers their apoptosis [146].
Other human herpesviruses appear to mediate mixed-effects over autophagy-related processes. HCMV can primarily induce the autophagosome pathway while inhibiting the degradation of its own newly-synthetized viral proteins by impairing lysosome fusion, promoting the accumulation of LC3 vesicles as a reservoir for viral particle assembly [147]. Indeed, HCMV directs the accumulation of lipidated LC3 and LC3 homologs GABARAPL1and GATE16-containing vesicles towards the viral assembly complex, along with fragmented Golgi vesicles, in order to aid the cytoplasmic envelopment of viral particles [148]. Moreover, in fibroblasts, HCMV infection inhibited autophagosome formation, either dependently or independently of mTOR signaling [149]. Since mTOR complex 1 is a downstream effector of PI3-kinase signaling, this factor participates in the negative regulation of autophagosome formation. Ultimately, HCMV has been reported to inhibit the completion of autophagy due to the interaction of the TRS1 viral protein with beclin-1, which blocks autophagosome formation [147].
Finally, KSHV has dual effects over autophagy. It has been reported that the viral protein named replication and transcription activator (RTA), which is involved in lytic cycle reactivation from latency, promotes autophagy to facilitate KSHV lytic replication. In addition, KSHV RTA is capable of increasing LC3 lipidated proteins and the number of autophagic vacuoles [150]. Interestingly, during KSHV infection, mTORC1 was required for the synthesis of RTA, but it was unable to affect autophagosome formation or control autophagic flux [151]. Moreover, when the lytic cycle is induced in vitro in latently-infected lymphoma B cells, KSHV, similarly to EBV, promoted autophagy to enhance their replication, while blocking the last autophagic steps by downregulating RAB7A, preventing the fusion of autophagosomes with lysosomes [150,152]. Conversely, KSHV has been reported to have the ability to suppress autophagy by expressing a homolog of FLICE-inhibitor protein vFLIP, which binds to ATG3 and impairs the lipidation and activation of LC3, which is required for autophagosome vacuole formation [153]. Moreover, KSHV encodes a homolog of BCL2, which helps the virus inhibit beclin-1 binding to PI3 kinase complex, thus preventing autophagosome formation [154]. KSHV has also proven to inhibit autophagy in monocytes by de-phosphorylating JNK2, altering the calpains-calpastatin balance and increasing the calpain activity responsible for the cleavage of ATG5 [129]. Furthermore, KSHV also inhibits autophagy in differentiated DCs by hyper-phosphorylating STAT3 [129]. Finally, it is noteworthy to mention that the murine gammaherpesvirus 68 (MHV-68), which is closely related to EBV and KSHV, has been reported to display enhanced reactivation from latency by autophagy-related genes [155].
Overall, autophagy inhibition would help these viruses avoid being recognized by immune components that rely on autophagy-derived class-II antigen presentation [156].
Taken together, although most herpesviruses may lean towards inhibiting autophagosome formation, as it likely represents a threat for virus viability and fitness in the infected cells, other herpesviruses may opt to induce the formation of autophagosomes ( Figure 2). Recently, it was reported that HHV-6A promotes autophagy, whereas HHV-6B inhibits autophagy, which was related to their ability to differently modulate the unfolded protein response, impacting cell survival [157]. However, further studies are needed to determine the existing interplay between HHV-6A and HHV-6B viral determinants in autophagosome formation. The effect of HHV-7 infection over autophagy remains to be determined, as well as the potential linkage of autophagy deregulation by this virus and the early onset and development of neurodegenerative diseases [158,159].

Lysosomal and Golgi-Sorting Vesicles during Herpesvirus Exit
Herpesviruses have been reported to harness vesicular trafficking of the endosomal sorting complex required for transport (ESCRT), responsible for the biogenesis of multivesicular bodies (MVB) [160]. The ESCRT pathway involves the recognition of ubiquitinated cargo and the sorting of membrane vesicles, while vesicles are sculpted to acquire new morphologies and functions such as deforming endosomal membranes to produce MVB [161]. The ESCRT machinery is composed of a set of five protein complexes that are sequentially recruited to late endosomes [162]. Furthermore, it has been described that all ESCRT complexes, namely complexes-0, -I, -II, -III, and the complex formed between Vps4 and Vps-associated protein VTA1, are required to structure the MVB that can then direct cargo to be targeted for lysosomal degradation or exosome formation [163]. Interestingly, MVB plays an important role in transporting glycoproteins to the plasma membrane once they exit the trans-Golgi network [164].

Herpesvirus Modulation of the Lysosomal Pathway
Lysosomal vesicles are a significant portion of the recycling mechanism that processes damaged and senescent macromolecules inside acidic membrane compartments from the cytosol and operates downstream of endosomal trafficking [165]. Importantly, this pathway is also used by the cell for targeting incoming viral particles to degradation as a host antiviral mechanism [166,167].
However, HSV-1 and HCMV can survive within lysosomes by increasing the pH of this compartment in order to inactivate lytic enzymes, such as hydrolases, and at the same time by partially acquiring an envelope from these organelles, forming a budding virion (Figure 3) [168]. Moreover, in a neuronal oxidative stress model, HSV-1 was seen to impair some functional features of lysosomes that ranged from an increase in the lysosome load to a reduction in hydrolase activity, as well as the maturation of lysosome cathepsin, along with the inhibition of EGFR-mediated endocytosis degradation [169]. For VZV, it has been reported that intracellular viral uptake by CHO cells is inhibited by lysomotropic agents, which suggests a role for low-pH-dependent endocytic and lysosomal-mediated virus entrance [170].
Additionally, herpesviruses have also been shown to exploit the lysosomal pathway to promote viral maturation (Figure 3). Viral particle maturation may require the budding of virions through intermediary membrane vesicles that compose the interface of the endosomal-lysosomal pathway to travel within the cells [171]. For instance, HSV-1 glycoprotein B has been found in multivesicular bodies (MVB) coming from late endosomes [172]. Of particular interest is the fact that HHV-6A, HHV-6B and HHV-7 all use the viral protein U21 in order to bind to MHC class I molecules and to direct these host proteins needed for antigen presentation to T cells to endolysosomal compartments favoring the evasion of immune system components [173].
Altogether, some herpesviruses can harness the cellular vesicles formed during lysosomal degradation or utilize them to direct glycoproteins to the surface membrane, mediate their transit within the cells and finally mediate the synthesis of new virions. Nevertheless, additional studies are needed in order to better understand the viral determinants that are responsible for modulating lysosomal pathways within cells infected with herpesviruses.  The corresponding viral proteins are transported to the cell membrane and then internalized into endosomes. This process has been described to be enhanced by means of the RAB5A protein by HSV-1. Clathrin-coated vesicles are transported into the cytosol, where the adapter protein 1 (AP-1) has been shown to interact with the VP22 and ORF9b proteins of HSV-2 and VZV, respectively. HCMV modifies the TGN, forming a barrel-like structure at the cis and trans sides of the Golgi apparatus.
Additionally, herpesviruses have also been shown to exploit the lysosomal pathway to promote viral maturation (Figure 3). Viral particle maturation may require the budding of virions through intermediary membrane vesicles that compose the interface of the endosomal-lysosomal pathway to travel within the cells [171]. For instance, HSV-1 glycoprotein B has been found in multivesicular bodies (MVB) coming from late endosomes [172]. Of particular interest is the fact that HHV-6A, HHV-6B and HHV-7 all use

The Role of the Golgi Apparatus in Herpesvirus Virion Maturation
After herpesviruses replicate their genome and assemble their capsids within the nucleus, they are exported from this compartment for further maturation in the cytoplasm. This maturation process involves the addition of several tegument proteins to the capsids, as well as capsid envelopment. The latter requires the transport of viral glycoproteins through vesicles. Herpesvirus glycoproteins converge in the endoplasmic reticulum and then travel to the host cell membrane, before being re-internalized to combine with mature capsids [164]. After the viral glycoproteins reach the Golgi apparatus, they are immersed into the trans-Golgi network (TGN) and are exported to the plasma membrane ( Figure 3) [174,175]. Afterward, these proteins return from the cell surface to the cytosol via early endosomes [176] and are combined with viral capsids, thereby creating newly infectious virions within these membranous compartments [177]. This endocytosis step is at least partly mediated by Rab GTPases, such as RAB5A and RAB11, which enable the plasma membranes with viral glycoproteins entering the cell to form endocytic tubular membranes and provide viral capsids with an envelope, as seen in the case of HSV-1 [178]. Regarding anterograde and retrograde endosomal transport of viral particles in neurons, PC12 cells have been described to have viral capsids co-localizing with the trafficking regulator RAB5A, as well as the nerve growth factor NTRK1 (TrkA), which caused a higher expression of the amyloid precursor protein (APP) and altogether could account for the negative implications of some herpesviruses, such as HSV-1 in Alzheimer's disease [179]. Interestingly, HSV-1 has been shown to induce the fragmentation of the Golgi network upon lytic infection in some cell types, such as Vero and HEp-2 cells [180]. Importantly, the fragments of the Golgi apparatus retained their glycotransferase activities required for O-glycosylation of extracellular viral particles [181].
On the other hand, the tegument protein ORF9b of VZV and its HSV-2 homolog VP22 have been found to interact with adaptor protein 1 (AP-1), which participates in clathrin-coated intracellular vesicle-mediated transport, as well as in defining the cargo of proteins shuttling between endosomes and the trans-Golgi network, and likely contributes to the secondary envelopment stage needed for producing mature herpesvirus virions [182]. Of particular interest was the discovery of the intricate structure of the HCMV assembly complex. This virus reorganizes in an inward barrel architecture at the cis and trans sides of the Golgi network, on top of early endosomal vesicles, and is surrounded by MVBs and recycling endosomes [183]. This process is believed to allow the orchestrated envelopment of herpesviruses before their exit from the cell [184,185]. Finally, the nucleocapsids of HHV-6A and HHV-6B are released from the nucleus into the cytoplasm through a deenvelopment step, where they acquire tegument proteins [186]. This process allows the viruses to egress with a viral membrane ridden with glycoproteins from the trans-Golgi network [187].
Taken together, numerous herpesviruses have evolved strategies for surviving the harsh conditions within lysosomes, and some for benefiting from the ESCRT machinery to travel through MVBs and into the Golgi network [185,188,189]. The modulation of these cellular trafficking pathways underscores the steps required by human herpesviruses to complete effective replication cycles, which result in mature virions that are ready for molecular signals for exiting the infected cell [190].

Exocytosis Vesicles Hijacked by Herpesviruses
To allow the egress of newly-synthesized infectious herpesvirus particles from the infected cells, the viral capsids that are assembled in the nucleus must travel through the perinuclear space, through the inner nuclear membrane (INM) and outer nuclear membrane (ONM) to reach the cytoplasm, followed by their envelopment within vesicles containing the viral glycoproteins that decorate the viral envelope [175,191]. After an initial envelopment at the INM, the enveloped capsids are released into the perinuclear space, from which the virus fuses to the ONM and releases into the cytosol the viral capsids [192,193]. These capsids are then reenveloped in vesicles that belong to the trans-Golgi network, or in the case of HCMV, enter secretory pathway-derived vesicles from the virus assembly complex (vAC) for their extracellular release as enveloped virions within exocytosis vesicles (Figure 4) [184,194]. itial envelopment at the INM, the enveloped capsids are released into the perinuclear space, from which the virus fuses to the ONM and releases into the cytosol the viral capsids [192,193]. These capsids are then reenveloped in vesicles that belong to the trans-Golgi network, or in the case of HCMV, enter secretory pathway-derived vesicles from the virus assembly complex (vAC) for their extracellular release as enveloped virions within exocytosis vesicles (Figure 4) [184,194].  The HSV-1 Us9 protein interacts with capsids in the endoplasmic reticulum (ER), the Golgi apparatus and the cytosol in order to promote the anterograde transport of viral components through microtubules. Us3 protein has also been shown to stabilize and activate microtubules and CLASP complexes to promote vesicle transport. The interactions of viral gD with host M6P and its receptors promote the export of viruscontaining late endosomes. On the other hand, viral proteins have been shown to co-localize with CD63 or CD63 and MHC-II, suggesting a potential immune-modulatory role for late endosomal trafficking. (C) VZV has been reported to use M6P and its receptors to promote viral export. Additionally, this virus appears to be exocytosed in single-membrane vesicles from the autophagosomal pathway, which contain LC-3 and Rab II. (D) EBV shuttles LPMI-and miRNA-containing vesicles derived in exosomal traffic modulation. (E) Exit of HHV-6 occurs through the exosomal pathway, using multivesicular bodies (MVBs). (F) Finally, KSHV recycles clathrin-coated endosomal vesicles, used during cell entry, for viral exit using the viral protein ORF65.

Modulation of Exocytosis by Herpesvirus
Interestingly, several viral glycoproteins that are not found in fully developed and mature extracellular virions are involved in capsid egress from the perinuclear space. Among these proteins are pUL31, which is a nuclear matrix-associated phosphoprotein; pUL34, a nuclear membrane-associated phosphoprotein; and Us3, a serine/threonine kinase that interacts with lamins A and C to disrupt the nuclear lamina in order to promote the envelopment of nascent viral capsids [192,195,196]. More recently, it has been reported that depletion of the host vesicle-associated membrane protein-associated protein B (VAPB) caused the accumulation of viral capsids in the nucleus, which supports its involvement in primary envelopment during nuclear egress [197]. Furthermore, the HSV-1 envelope protein US9 has been shown to associate with capsids in the cytoplasm, the endoplasmic reticulum and the Golgi apparatus, and to function in concert with gE/gI during anterograde transport within axons in neurons by promoting both the loading of capsids and glycoprotein-containing vesicles onto kinesin motors and microtubules [198].
Notably, several viral molecules participate in the active transport of newly synthesized virions through the exocytic route. Similar to the interaction of tegument proteins in the transport of capsids during cell entry, UL36 and UL37 proteins may be involved in plus-end capsid transport towards the cell surface [199]. For example, UL20, an envelope protein that is highly conserved in alphaherpesviruses, has been implicated in syncytia formation, which is required for cell-to-cell infection, and also for cytoplasmic envelopment within the trans-Golgi network (TGN) and for virion transport from the TGN to the plasma membrane or extracellular space, therefore promoting the egress of infectious virions [200]. Importantly, it has been found that the function of UL20 of HSV depends on its localization at the Golgi apparatus, causing it to be palmitoylated by host GODZ, which allows the proper trafficking of UL20 and gK to exocytosis vesicles [201]. Here, they would participate, along with gM, in incorporating other glycoproteins such as gD and gH/gL into mature virions [202].
Additionally, herpesviruses can stabilize microtubules to enhance virus release [203]. For instance, the HSV-1 protein US3 has been reported to stabilize microtubule structures and be necessary for viral spread by activating cytoplasmic linker-associated proteins (CLASPs), a type of plus-end tracking protein (+TIP) that operates in microtubule nucleation at the Golgi apparatus and captures microtubules near the membrane bilayer [204]. HSV-1 employs a mechanism in which glycoproteins such as gD are modified with mannose-6-phosphate [205] residues that cause their export within late endosomes containing M6P receptors through the TGN, and which culminates in viral assembly complexes for virus egress [205].

Modulation of Endosomal and Exosomal Pathways by Herpesviruses
Herpesviruses have also evolved molecular mechanisms to interfere with the endosomalexosomal pathways of infected cells (Figure 4). For instance, gB of HSV-1 is frequently found in extracellular vesicles and has been shown to co-localize with the late host endosomal marker CD63 and bind to MHC-II molecules (HLA-DR in human cell lines) in such a way to delay its endosomal trafficking, which may have immunomodulatory effects [206,207]. Furthermore, exosomal targeting, but not MHC-II interference, is mostly conserved in other gB-bearing alphaherpesviruses [208]. This finding may prompt investigations into how the exosomal release of gB improves viral fitness while roaming in the infected host. On the other hand, VZV utilizes an amalgam of endosomal and autophagosomal pathways to perform exocytosis of complete viral particles, as observed by electron microscopy, which shows virions in single-membraned vesicles carrying both Rab11 and LC3, which resembled amphisomes [209]. Alternatively, VZV has been shown to have the ability to bind to the mannose-6-phosphate receptor in late endosomes as a mechanism to egress from infected melanoma cells or fibroblasts, regardless of the autophagosome pathway [210]. A recent report indicates that the VZV glycoprotein M (gM) contains two tyrosine-XX-bulk hydrophobic amino acid domains and one dileucine trafficking-domain associated with trans-Golgi sorting vesicles, which implies that gM is targeted to endosomal-lysosomal compartments to help viral maturation [211]. Interestingly, manipulating these motifs in VZV mutants did not impair virus assembly or egress, but rather increased its virulence in the skin in a humanized severe combined immune deficient (SCID) mouse model [211]. The actual mechanism by which the glycoprotein M is responsible for increased virulence is currently unknown.
EBV employs the exosomal pathway to introduce several microRNAs and LMP-1 in extracellular vesicles, capable of altering the latent-lytic cycle of an infected cell [212,213]. The latter feature likely promotes the communication between epithelial cells undergoing lytic replication to direct the maintenance of a latently-infected B cell [213]. To this end, CD63 has been found to promote LMP-1 exosomal trafficking and enhance exosome production [214].
Meanwhile, as discussed above, HCMV has an organized assembly complex in which after capsids acquire their tegument layer, they pass through the TGN to acquire an envelope and likely obtain the necessary glycoproteins from early endosomes, which are ultimately released from recycling endosomes or Golgi-derived membranes [215,216].
Regarding HHV-6, previous reports indicate that viral glycoproteins co-localize with CD63, a marker of late endosomes that is found in MVB vesicles, suggesting that the infectious viruses are released by the cell via the exosomal pathway. Noteworthily, it is currently unknown what processes are involved during HHV-7 virion exit [187].
In addition to playing an active role in KSHV during entry, actin filaments also undergo rearrangements that drive the trafficking of viral proteins throughout the cytosol of endothelial cells, as seen with ORF65, through clathrin-coated early and recycled endosomes during secondary envelopment steps [107].
Altogether, at present there is relatively scarce information available regarding the multiple processes by which herpesvirus may prefer to leave infected cells and how viral-derived exosomes affect neighboring cells. For example, it is unknown whether most herpesviruses possess a mechanism for egressing similarly to HSV-1, using a UL20/gK/gM complex directing viral glycoproteins to exocytosis vesicles to assemble mature virions [203]. Therefore, additional studies are needed to elucidate how herpesviruses egress infected cells and to explore the pros and cons of these routes.

Concluding Remarks
Herpesviruses use and modulate several host intracellular membrane trafficking pathways in order to elicit productive infection, effective virus replication and synthesis of new infectious viral particles and, finally, virion exit from the infected cells. Although significant progress has been made regarding the understanding of the interplay between herpesviruses and their host, namely at the cellular level, increased knowledge is required in this respect, which could shed light on novel strategies for their control or inhibition. Particularly, much remains to be studied in relation to viruses such as HHV-6 and HHV-7, regarding what viral components mediate and modulate host membrane vesicle trafficking, their interactions and the key factors in the infectious cycles of these viruses. In particular, future work may focus on how intracellular viral particles are differentially directed to specific cell membranes or sites within the cell for exocytosis or cell-to-cell spread.
Additional studies are needed in order to determine the roles of each of the viral and host factors that mediate the processes involving intracellular host membranes that are modulated during the replication cycle of herpesviruses. Again, knowledge in this regard should contribute to the understanding of critical interactions that could be potentially targeted by different drugs to limit viral replication. Furthermore, we foresee that new studies providing insights into the workings and regulations of the vast intracellular transport machinery and how it assists viral infections will be valuable in the long-term for harnessing herpesviruses and limiting the diseases they produce. Funding: Authors of this work are supported by the Millennium Institute on Immunology and Immunotherapy (P09/016-F; ICN09_016, ANID, Chile) and FONDECYT grants #1190864, #1161525, #1170964 and #1191300 from the Agencia Nacional de Investigación y Desarrollo (ANID, Chile).

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Conflicts of Interest:
The authors declare no conflict of interest.