Novel Functional Motifs of the Cell Entry Glycoprotein D for Oncolytic Herpes Simplex Viruses

Background: Oncolytic herpes simplex virus (oHSV) have been proved effective and safe to treat tumors. Glycoprotein D (gD) has been engineered for targeting cancer cells and de-targeting normal cells successfully, however, the effectiveness and safety of oHSVs still need to be improved. Method: Here we sequenced the DNA encoding gD of our recently isolated new strain HSV-1-LXMW and compared the gD amino acid sequence with the gDs of other 7 HSV-1 and 3 HSV-2 strains. Results: Phylogenetic analysis revealed that HSV-1-LXMW is evolutionarily close to HSV-1-Patton and -KOS strains. The gD amino acid sequence alignment identified 19 conserved and 8 variable regions. We further predicted 10 new motifs in HSV gD for the first time and identified motif differences in HSV-1 and HSV-2. We summarized the gD-engineered oHSVs and found that some of the newly identified gD motifs are actually functional. Conclusion: Our results shed light on HSV gD biology and provided new directions for future gD functional studies and engineering in order to make better oHSVs.


Background
Tumors contain heterogeneous cancer cells, such as tumor stem cells, and tumor stromal cells, immune cells in the tumor microenvironment [1]- [6]. The infection of herpes simplex virus 1 and 2 (HSV-1 and -2) is spread worldwide. The infection of HSV-1 often causes oral lesion and is frequently diagnosed in children, while the HSV-2 causes genital lesion and sexually transmitted diseases [7]. HSV in not only a human pathogen, but HSV mutants named oncolytic herpes simplex virus (oHSV) can be used to treat cancer [8] [9] [10]. Because the infection of HSV can induce rapid and efficient cell killing, HSVs have a wide host range, and the virus replication and infection can be controlled by effective anti-herpetic agents, HSVs have been widely used as oHSVs [11]- [18]. Targeting cancer cells without infecting normal cells is still a big challenge in cancer treatment [19]. Oncolytic viruses provided a new way to retarget to the specific cancer cells [20]. For example, Jeeninga et al. used the characteristics that human immunodeficiency virus 1 (HIV-1) infects T lymphocytes specifically to construct HIV-1 derived oncolytic viruses to treat the T-cell acute lymphoblastic leukemia (T-ALL) [21]. Schneider et al. engineered measles virus to enter cells expressing specific receptors for epidermal growth factor or the insulin-like growth factor 1, and trigger cell death [22]. T-VEC is engineered by deletion of the genes γ34.5 and US11 (ICP47) and insertion of the gene encoding human granulocyte macrophage colony-stimulating factor (GM-CSF) [15]. In 2015, the US Food and Drug Administration (FDA) proved that T-VEC as the first oHSV to treat advanced inoperable malignant melanoma [16]. HSV envelope glycoproteins were also engineered to target the special receptors expressed on tumor cells [23].
The process of virus entering into the host cells includes recognition of virus receptors, triggering the fusion process and fusion execution [24]. This process needs the glycoprotein D (gD), heterodimer glycoprotein H/glycoprotein L (gH/gL) and glycoprotein B (gB), as well as their cognate receptors [25] [26] [27] [28] ( Figure 1(A)). gD is a glycoprotein that can interact with one of its three receptors, herpes virus entry mediator (HVEM), nectin-1, and 3-O-sulfated heparan sulfate (3-OS HS), thus inducing the initiation of entry process [29]- [34]. Receptor binding leads to the conformational change of gD, which will trigger the virus fusion process after activation of the heterodimer gH/gL and gB [35] ( Figure 1(A)), thus gD is the major tropism determinant.
The gD protein has been divided into three parts, including the extracellular, transmembrane and cytoplasmic domains [36] [37] [38]. gD is made up of 394 AA and becomes a mature form (369 AA) after cleavage of the 25 AA signal sequence at the N-terminal and the transmembrane domain is between AA 317 and 339 [39]. The gD co-crystal structure revealed that the HVEM binding site in the gD is located at amino acids 1 -37 [40]. The nectin-1 biding site is discontinuous, and the critical residues have been identified at AA 34,38,215,222,and 223 [41] [42] [43]. The N-terminus (residues 1 -260) contains the receptors  binding sites, while the C-terminus (residues 260 -310) acts as the pro-fusion domain (PFD) that can be used to induce virus infectivity and fusion process [44]. A quantity of mutations in gD can abolish the binding of HVEM and 3-OS HS with gD at the same time, so these two receptors binding sites in gD are at least overlapping in part [45]. However, several mutations aborted the HVEM binding to the gD but not nectin-1, so the nectin-1 binding site is discrete from HVEM but partially overlapping, and its location is downstream of residue 32 [ [48]. The interaction of gD with HVEM or nectin-1 can lead to displacement of the binding of C-terminus and N-terminus, then lead to the structural change of gD [44] [49]. In the process of virus entry into host cells, AA 61-218 don't code executable meanings and the interaction of the ligand with its receptor doesn't need to change the structure of gD [50]. Consistent with reports, mutations the amino acids between 25 -27 can abolish the binding of gD with HVEM and increase the affinity of gD with ncctin-1 [33] [45] [51]. The majority of oHSV given through systemic route were cleared by neutralization antibodies and other immune effects [52]. The intralesional injection of oHSV limited the oHSV usage in clinics. The effectiveness and safety of oHSV still need to be improved. Here we sequenced the DNA encoding gD of our recently isolated new strain HSV-1-LXMW, compared the gD amino acid sequence with the other 10 gDs by phylogenetic analysis, amino acid sequence alignment, and motif predictions. Our results shed light on HSV gD biology and provided new options for future engineering of gD in order to make improved oHSVs.

HSV Genomic DNA Sequencing Analysis
A new HSV strain from an oral herpes lesion of a patient in Beijing, named as HSV strain LXMW, was isolated earlier [53]. The HSV DNA sequence identification was carried out by genomic DNA sequencing as described earlier [53]. Briefly, we isolated HSV DNA in our laboratory, high-quality genomic DNA (500 ng) was submitted to Beijing Genomics Institute (BGI, http://www.genomics.cn) for sequence analysis and Burrows-Wheeler Aligner (BWA) software was used for alignment. The genomic sequences of other 10 HSV strains were obtained from NCBI Reference Database (Table 1).

Phylogenetic Analysis of the gD Amino Acids in HSV-1 and HSV-2
For phylogenetic analysis of the gD amino acid sequences of 11 HSV strains, the software MEGA7 was used. The evolutionary history was calculated by means of the Maximum Likelihood method option based on the General Time Reversible model with "completely deletion". The bootstrap consensus tree is taken to represent the evolutionary history of the taxa analyzed. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates are collapsed. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pair wise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting

Alignment of gD Amino Acid Sequences of HSV-1 and HSV-2 Strains
The genomic sequences used for analysis were the same as above. The online software EMBL-EBI (https://www.ebi.ac.uk) was used to align the gD amino acid sequences. Amino acid differences were marked.

Prediction of the gD Secondary Structure, Motifs and Function Domains
The online software UCL-CS Bioinformatics (http://bioinf.cs.ucl.ac.uk/) was used to predict the secondary structure. The stringent conditions were set as default by the online program. The online software Phobius (https://phobius.sbc.su.se/) was used to predict the gD function domain. We predicted the conserved domain of gD in LXMW using the online software NCBI (https://www.ncbi.nlm.nih.gov), and the motifs were predicted using the online software MEME (http://meme-suite.org/).

gD DNA Sequencing of Our New Strain HSV-1-LXMW
The sequence of gD was determined as follows: 138401ATGGGGGGGGCTGCC

Translation of HSV-1-LXMW gD DNA Sequence into Amino Acids
We used the DNAMAN software to translate the DNA sequence to amino acids; the translation result is shown as follow:

Phylogenetic Analysis Showed HSV-1-LXMW Is Close to HSV-1-Patton and -KOS
To understand the evolutionary relationship of our HSV strain with other HSV-1 and HSV-2 strains, a phylogenetic analysis was performed. The gD DNA sequence of HSV-1-LXMW together with 7 HSV-1 strains (17, F, H129, RH2, SC16, Patton and KOS) and 3 HSV-2 strains (G, HG52 and SD90e) ( Table 1) were translated to amino acid sequences and analyzed. Our gD is highly similar to HSV-1 strains, but there is no significant similarity between the sequences of our HSV-1-LXMW and other HSV-2 strains. Our data supported that HSV-1-LXMW is an HSV-1 strain. Both the phylogenetic tree data (Figure 2(A)) and neighbor network data (Figure 2(B)) showed the presence of four groups of clustering structures. Our new strain HSV-1-LXMW isolated in Beijing, China is close to strains HSV-1-Patton in New York, US and HSV-1-KOS in Princeton, US, and far from the strain HSV-1-17 in University of Kansas, USA ( Figure 2). The data showed a mean distance of approximately 87% among the strains tested collectively.

The gD Amino Acid Sequence Alignment Identified 19 Conserved and 8 Variable Regions
To understand if the gD amino acid sequences are conserved among HSVs and thus likely to be biologically functional, the sequence alignments of the 11 above-described sequences of HSVs were performed (Figure 3). We found that all the HSV-1 gD have 394 amino acids, while there are 393 amino acids in HSV-2 strains SD90e and HG52, and 347 amino acids in HSV-2 strain G ( Figure  3). Our results showed that there were fewer mutations among HSV-1 strains in the gD amino acids sequences, but there are more variations between HSV-1 and HSV-2. Our results identified that there are 2 amino acids of HSV-1-LXMW   strains RH2, H129, SC16 and F, and all the HSV-2 strains (black boxes in Figure 3). The conserved domains among all the gD amino acids of 11 HSVs are marked as numbers 1 -19 and underlined in red and the variable amino acids are marked as numbers 1 -8 underlined in black. The conserved domains predicted by NCBI software are marked in gray. The motifs predicted by the online software are marked by color boxes.

Prediction of the Conserved Domains, Motifs and Secondary Structures in gD
The conserved domains of HSV predicted by NCBI software. We used the online NCBI conserved domain database (https://www.ncbi.nlm.nih.gov/) to predict the conserved domains of the gDs of all the 11 HSVs, and the result is marked in gray (Figure 3). We found that the conserved domain in HSV-1 strains 17, Patton, KOS, LXMW and HSV-2 strain HG52, sd90e are the same (located at AA 82-206), however, the conserved domains in HSV-1 strain H129, RH2, SC16, F are the same (located in AA 82-206 and 343-373), the conserved domain in HSV-2 strain G is located in AA 36-106.

The Prediction of gD Function Domains
To predict the function domains of gD, we used online software Phobius (https://phobius.sbc.su.se/). We found that the gD function domains of HSV-1 strain 17 and LXMW are the same (Figure 4). The signal peptide is located at

The Prediction of 10 gD Motifs
To predict the gD motifs, we used the online software MEME (http://meme-suite.org/). The motif amino acids are marked in Figure 5(A) and the predicted new motifs are named as 1-10 in Figure 5 that the motifs of gD in 7 HSV-1 strains 17, H129, RH2, SC16, Patton, KOS, and LXMW are the same. HSV-2 strains HG52 and SD90e have the same motifs. The main motif differences between HSV-1 and HSV-2 are located in the motifs 8 and 9. The motifs of gD in HSV-2 strain G is different from any others. All the HSV-2 gD lack the motif 9 that is located in the signal peptide of AA 1-15 ( Figure 3 and Figure 5). The HSV stain G also lacks the motif 7 that is located in the receptor binding region. The motif 8 in strains HG52 and SD90e is located at AA 328-342 and the motif 8 in HSV-1 is located at AA 329-343. The motif 8 is near to the transmembrane domain and overlapped with the pro-fusion domain located at AA 250/260-310 (AA 275/285-335 in the alignment in Figure 3) [54]. The motif 5 is also located in the profusion domain. In addition, all the gDs in HSV-2 lack the AA 329 in the motif 8. Most of the gD amino acid mutations between HSV-1 and HSV-2 are located in the variable regions 1, 6 and 7, which overlapped respectively with the signal peptide and transmembrane domains, and with motifs 9, 8 and 6 ( Figure 3 and Figure 5).

The Prediction of gD Secondary Structures
We predicted the gD secondary structures of HSV-1 strain 17, LXMW, and HSV-2 strain HG52 using the online software UCL-CS Bioinformatics (http://bioinf.cs.ucl.ac.uk/). The prediction results are shown in Figure 6. According to the results in HSV-1 strain 17, AA 327-337 form disordered protein binding sites, while in HSV-1 strain LXMW, this site locates at AA 327-338. Both 2 sites are located in the pro-fusion domain and the motif 8.

The gD-Engineered oHSVs
We summarize the oHSV engineered on the gD to retarget the HSVs to cancer cells (Figure 1(B) and Table 2). Zhou et al. first engineered gD to retarget to the IL13 α2 receptor (IL13Rα2) and the viruses can enter cells via the IL13Rα2 receptor [23]. Then Kamiyama et al. engineered the gD in HSV-1 to retarget to the urokinase plaminogen activator receptor (uPAR) using the same strategy [55].  [58]. oHSV retargeted to the human epidermal growth factor receptor (EGFR) or human carcinoembryonic antigen (CEA) also have been explored [59]. Most of the engineering was done in the motif 7. We first compared the relationship of engineering sites in gD with conserved regions, domains and motifs.

Discussion
Studies on HSV gD molecular biology will aid future development of more effective and safer oHSVs. In this article, for the first time, we did systematic comparative sequence analysis of gDs from 8 HSV-1 and 3 HSV-2 strains, including our recently isolated new strain HSV-1-LXMW. These comparisons include phylogenetic analysis, aminoet acid sequence alignment, prediction of functional domains, motifs and secondary structures, and reported oHSVs. We identified 19 conserved regions, 8 variable regions and 10 new motifs in HSV gD and correlated the gD engineering sites with the newly identified motifs, for the first time. These findings have important implications on HSV gD biology and making better oHSVs.
The phylogenetic analysis showed that the gD amino acid sequence in HSV-1-LXMW isolated in China is close to HSV-1 strains Patton and KOS in USA, while our earlier study showed ICP27 is close to HSV-1 strains Patton and H129  [53]. This suggests that our strain is revolutionarily closer to strain Patton. Our identified relatively conserved arginine cluster (RRR residues 365-367) in the cytoplasmic domain was reported to be required for the efficient induction of plasma membrane projection and viral final envelopment, a function important to the viral replication and cell-cell spread [60]. Our literature search identified only one study, which reported that the putative HSV gD domains, motifs or their functions [60]. The only one such article reported that the putative cholesterol-binding motif that overlaps HRR is not essential for the gBsyn phenotype and cell-to-cell spread in HSV-1 strain KOS [61]. Future studies may focus on the functional analysis of our newly identified HSV gD domains and motifs.
Our phylogenetic analysis showed a mean distance of approximately 87% among the strains tested collectively. An earlier study reported that the gDs of HSV-1 and HSV-2 show 82% distinctiveness, and are supposed to be similar in structure and function [36]. Several studies have further indicated that the gDs of HSV-1 and HSV-2 have similar affinities for nectin-1 and are interchangeable in virus infection and soluble gD entry inhibition [51] [62] [63]. A report showed that the HSV-2 subunit vaccine containing gD can provide protection in HSV-1 genital infections but not preventing HSV-2 infection [64]. The authors suggest that this phenomenon probably is caused by the 89% gD amino acids homology from HSV-1 and -2 [64]. Among the gD receptors, nectin-1 plays a more important role in the process of HSV-2 infection at least in murine models [65] [66]. Although we still cannot find any reports about engineering the gD in HSV-2 to retarget the virus to cancer cells, there are some reports about oHSV-2 [67] [68] [69]. HSV-2 is reported to have more potent ability to destroy tumor cells, more potent virulence in causing necrotizing stromal keratitis, encephalitis, and produce higher titers in animal models [70]. Future engineering of HSV-2 gD may be pursued.

Conclusions
The ideal route of delivery of oHSV is systemic route, but the production of neutralization antibody limits this application [52] [71]. The intralesional injection of oHSV limits oHSV usage in clinics. To improve oHSV vectors, we can engineer the epitopes in gD to reduce the production of neutralization antibody [71]. In addition, engineered oHSVs can enter only the cancer cells, but the clinical usage grade viruses should be produced in normal cells. gD can be engineered to retarget to two different receptors, so that the virus can enter both cancer and special normal cells [72].
Through the understanding of HSV gD, we can engineer the gD to detarget from normal cells and retarget to cancer cells more precisely and efficiently, in order to make better oHSVs.

Funding
This work was partly supported by grants from the National Natural Science Foundation of China (81872412 to XHW, 81772223 to RBX, 81602303 to XY, 31700736 to WXW). We thank Hubei Province Natural Science Foundation of China (2016CFB180 to WXW), Foundation of Health and Family Planning Commission of Hubei Province (WJ2016Y07 to WXW), Hubei Province Scientific and Technological Research Project (Q20171306 to XWW), Jingzhou Science and Technology Development Planning Project (JZKJ15063 to WXW).

Ethics Statement
The studies involving human participants were reviewed and approved by Yangtze University. The patients/participants provided their written informed consent to participate in this study.

Conflicts of Interest
No conflict of interest exists in the submission of this manuscript, and manuscript is approved by all authors for publication. I would like to declare on behalf of my co-authors that the work described was original research that has not been published previously, and not under consideration for publication elsewhere, in whole or in part. All the authors listed have approved the manuscript that is enclosed.