Integrative Biomedical Research

Introduction

Epstein Barr virus (EBV) also known as human herpesvirus 4 (HHV-4) is a ubiquitous and oncogenic gammaherpesvirus that has infected over 95% of the worldwide population ranging from asymptomatic to infectious mononucleosis (Womack & Jimenez, 2015). EBV is implicated in the pathogenesis of Hodgkin lymphoma, nasopharyngeal carcinoma, gastric cancer, and numerous malignancies in individuals with inherited or acquired immunodeficiency (Matthew & Razelle, 2004). EBV infects B cells and resides in memory B cells in healthy people to establish a life-long persistence in the human host in asymptomatic individuals and does not cause disease. EBV persistently infects memory B-cells due to the prevalence of the CD21 receptor present on the surface of B-cells (Sugano et al., 1997). CD21 receptor acts as the major cellular receptor for EBV as CD21 interacts with the EBV glycoprotein (Busse et al., 2010). Epstein Barr virus nuclear antigen 2 (EBNA-2) is a gene encoded by EBV that is essential for the growth transformation process and a major determinant of the differences between EBV-1 and EBV-2 subtype in lymphocyte growth transformation (Cohen et al., 1989). 

In this study, the EBV EBNA2 gene that was used as reference samples is from the C666-1 cell line, an undifferentiated nasopharyngeal carcinoma (NPC) from a subclone of its parental cell line, C666, derived from an NPC xenograft of southern Chinese origin where the GenBank accession number is KC617875 (Cheung et al., 1999). This cell line consistently maintains EBV in long-term culture, providing an excellent in vitro model for EBV and NPC studies (Cheung et al., 1999). Maximum likelihood is the technique used for this study in the estimation of evolutionary trees from nucleic acid sequence data which is not likely to give misleading results if rates of evolution differ in different lineages. In addition, it also allows testing of assumptions about evolutionary rate constancy via likelihood ratio tests and gives a rough indication of the error in tree estimates (Felsenstein, 1981). Studying the diversity of EBV EBNA2 in various disease is pivotal as this gene could act as a transcription factor in activating crucial downstream elements. Elucidating the differences of EBNA2 across diseases could provide valuable insights on its functioning.

Materials and Methods

FASTA Sequence Searching

The FASTA sequences of the EBNA-2 region for Epstein-Barr virus (EBV) were found, retrieved, and downloaded from the National Center for Biotechnology Information (NCBI) database. The EBV FASTA sequences were identified for the country or region, type, disease-associated and isolate by using the NCBI accession code. The downloaded FASTA sequences were then combined into 1 sequence data file before being loaded.

Creating Multiple Sequence Alignments

An alignment was created from nucleotide sequence data that will be imported into the alignment editor. The Alignment Explorer was launched by selecting the Align | Edit/Build Alignment menu command. Create New Alignment was selected and Ok clicked. The button labelled DNA was clicked. To align sequences contained in a sequence data file, first, the unaligned sequences, which is the combined FASTA sequence data file are added into the Alignment Explorer by clicking selecting the Data | Open | Retrieve Sequences from the File menu command. Then, the Edit | Select All menu command was selected to select all sites for every sequence in the data set. Next, the Alignment | Align by ClustalW menu command was selected to align the selected sequences data using the ClustalW algorithm. The OK button was clicked to run ClustalW analysis at default settings. Lastly, the current alignment session was saved by selecting the Data | Save Session menu command.

Model Analysis and Phylogenetic Tree Construction

Once the sequence has been aligned, the Data | Phylogenetic Analyses icon on the tab is clicked. The aligned sequence was then analysed by clicking on Model | Find Best DNA/Protein Models (ML) to look for the best-fits nucleotide models with the maximum likelihood statistical method. Next, the model with the lowest BIC scores (Bayesian Information Criterion), which is the Hasegawa-Kishino-Yano with gamma-distributed and invariant sites model was selected to create a phylogenetic tree. Out of 3 types of phylogenetic trees, which are Maximum Likelihood, Neighbour-joining and Minimum evolution tree, Maximum Likelihood was used. The phylogenetic trees were constructed by using original and bootstrap data. To create phylogenetic trees, first, the Phylogeny | Construct/Test Maximum Likelihood Tree menu command was selected from the main MEGA window launch bar. The Analysis Preferences window will appear, and the test of phylogeny was set as none. Then, the nucleotide substitution type was clicked. Nucleotide was clicked for substitution type and the Hasegawa-Kishino-Yano model was selected. Next, the rate among sets was set at gamma distributed with invariant sites (G+I) and the number of discrete gamma categories was set at 5. The OK button was clicked to construct a phylogenetic tree. For the bootstrap phylogenetic tree, a test of phylogeny was set to bootstrap, and the bootstrap value was set to 100.

Result & Discussion

Gene accession number, phenotype, and country mined for 11 samples and 1 reference sample of Epstein-Barr virus with Epstein-Barr virus nuclear antigen 2 (EBNA-2) gene were retrieved from the NCBI database (table 1). EBNA-2 sequence from 4 diseases which includes Nasopharyngeal Carcinoma, Burkitt Lymphoma, gastric cancer, and Hodgkin Lymphoma were used. These samples are respectively from 6 different countries namely China, Japan, Kenya, South Korea, Poland, and the United Kingdom.

The accession number of 1 reference sample and 11 samples are listed in the first column. In the second listed the disease associated and the third column listed the country the the EBV is isolated.

Mutation sites of each EBV EBNA-2 gene according to the mutation type which are missense, silent, insertion, deletion and unknown mutation were identified and significant mutations except for silent mutation were categorised and shown in table 2.1 to 2.4 and Figure 1. Missense mutations are significant mutation that causes changes in the amino acid sequence, while silent mutation does not change or affect the amino acid sequence. Insertion and deletion of nucleotides may give rise to different results on amino acids. Lastly, there is also an unknown mutation identified.

Missense mutation

The nucleotide sites changed for each sample when compared with reference gene KC617875 are listed in the first column.

The accession number of samples that have the corresponding mutation is listed in the second column. In the third column, the amino acid changes that are caused by the missense nucleotide mutation are listed.

Table 1: List of EBV EBNA-2 genes from different phenotypes and countries.

Table 2.1: Missense mutation sites of each EBV EBNA-2 gene by comparing to reference sample (KC617875)

Table 2.2: Deletion mutation sites of each EBV EBNA-2 gene by comparing to reference sample (KC617875)

Table 2.3: Insertion mutation sites of each EBV EBNA-2 gene by comparing to reference sample (KC617875)

Table 2.4: Unknown mutation sites of each EBV EBNA-2 gene by comparing to reference sample (KC617875)

Deletions

The nucleotide sites deleted for each sample when compared with reference gene KC617875 are listed in the first column. The accession number of samples that have the corresponding mutation is listed in the second column.

In the third column, the amino acid change or deletion caused by the nucleotide deletion mutation is listed.

Insertions

The nucleotide sites inserted for each sample when compared with reference gene KC617875 are listed in the first column.

The accession number of samples that have the corresponding mutation is listed in the second column. In the third column, the amino acid insertion that is caused by the nucleotide insertion mutation is listed.

Unknown mutation

The nucleotide sites that have unknown mutations for each sample when compared with reference gene KC617875 are listed in the first column. The accession number of samples that have the corresponding mutation is listed in the second column.

In the third column, the amino acid change or deletion that is caused by the nucleotide deletion mutation is listed.

Substitution Model Analysis and Phylogenetic Tree Construction

Nucleotide substitution models were analysed by using sample nucleotide sequences. The analysis is done by using an automatically created neighbour joining tree with the maximum likelihood method and all the nucleotide sites were used. The model analysed are combinations of 5 substitution models General Time Reversible (GTR), Hasegawa-Kishino-Yano (HKY), Tamura-Nei (TN93), Tamura 3-parameter (T92), Kimura 2-parameter (K2) or Jukes-Cantor (JC) with 3 rates among sites, gamma distributed (G), has invariant sites (I) or gamma distributed with invariant sites (G+I) (Makarova et al., 2012). Bayesian Information Criterion (BIC) scores and Akaike Information Criterion corrected (AICc) values are arranged from lowest to highest. While log-likelihood values (InL) are arranged from highest to lowest (Banos, 2010). The model is arranged from the most desirable to undesirable.

The BIC value, AICc values and lnL value are the 3 criterion that are used to determine the desirability of the model, in which a lower BIC and AICc value correlates to a better fit (Vrieze, 2012).  Out of 24 types of different nucleotide substitution models listed (table 3), Hasegawa-Kishino-Yano with gamma-distributed and invariant sites, the HKY+G+I model is the best model to be used as the BIC value and AICc values are the lowest among all model listed. In addition, the model has high InL value.

Maximum likelihood phylogenetic trees constructed by using normal and bootstrap data. The tree branched into 2 main clades with a branch length scale of 0.002 from a common ancestor where one clade is closely related to reference sample KC617875, this clade includes MK540241, MK540314, MK540313 and MK540359. Another clade is distantly related to KC617875, which includes AY961628, KC207813, LN824204, KC440851, MG021307, KC207814 and MG021308. Sample from closely related clades are mostly isolated from NPC sample in China except for MK540241 isolated from Burkitt Lymphoma sample in China. In distantly related clades, most of the samples are isolated from gastric cancer, Burkitt Lymphoma and Hodgkin Lymphoma except for AY961628 which were isolated form NPC. The bootstrap confidence values are shown on the node of each branch. Bootstrap value was set to 100 in which this indicates that the phylogenetic tree is constructed 100 times using the HKY+G+I model. Based on the generated trees, a particular branch occurring 100 times indicates high confidence rate, where this means that the occurrence of that particular branch is highly probable. The higher the value the more probable that the branch being real (Efron, Halloran, and Holmes 1996).

Phylogenetic constructed by using EBV EBNA-2 gene from 1 reference sample and 11 samples by using the maximum likelihood method and Hasegawa-Kishino-Yano with gamma-distributed and invariant sites, the HKY+G+I model is selected. All samples are branched according to their divergence and the bootstrap value of each branch is shown to indicate the desirability of the branch. It could be observed that the first major clade has a bootstrap value of more than 90 corresponding to occurrence rate of > 0.9. This indicates that high probability of the first major clade indicating diverging lineage between samples resulting in two major clades. Interesting observation to be made here is that the EBNA2 sequence from KC617875 and AY961628 are distally related even though the EBV sequence are from the same disease (NPC). Moreover, the MK540241 EBNA2 sequence share similar lineage between most of the EBV in NPC, indicating high similarity between the EBNA2 sequence in Burkitt’s lymphoma (BL) and NPC. Bootstrap value of 100 of the second major clade (from top) indicates total occurrence. This could also predict that EBNA2 sequence in KC207814 from BL and MG021308 from gastric cancer (GC) share a close similarity, it is interesting to note here also that geographical region also varies in these two samples, the BL sample was from Kenya while GC sample was from Poland. The gastric cancer sample isolated from the United Kingdom is the most diverged with 5 unique mutation sites identified.

4 types of known mutation and 1 unknown mutation have been identified from multiple sequence alignments performed by comparing 11 samples with reference sample KC617875. Out of 4 types of known mutation, which are missense, silent, deletion and insertion, missense mutation was most frequently identified, which causes change in encoded amino acids. However, the effect of these amino acid changes is not well known. For insertion mutation, which is of the lowest frequency, was identified only once in KC207814 and MG021307 respectively, where insertion of nucleotide CTC has been identified in-between nucleotide sequence 633 to 637 and caused insertion of amino acid leucine at position 212. EBNA-2 gene of the C666-1 cell line was used as a reference as C666-1 consistently maintains EBV in long-term culture, providing an excellent in vitro model for EBV EBNA-2 comparative study (Cheung et al., 1999). The use of the EBNA-2 gene from the C666-1 cell line enables the comparison of variation of the EBNA-2 gene in NPC samples with the EBNA-2 gene of other diseases which are Burkitt Lymphoma, gastric cancer, and Hodgkin Lymphoma.

The nucleotide deletion identified in NPC sample (AY961628), Burkitt Lymphoma (KC207813, KC20814) and gastric cancer (MG021307, MG021308) which is the deletion of amino acid from position 63 to 95, and amino acids change at 62, 66, 67 and 68 have a similar scenario with previous research of Harada et al. (2001) where the deletion occur from position 59 to 95. Deletion within this range did not affect the self-association of EBNA-2 in vitro or in vivo but inhibited the ability to maintain higher-order structures in non-denaturing gels (Harada et al., 2001). Since EBNA2 acts as viral transcription factor, the stability of protein structures are crucial for transcription factor (TF) functioning as disruption in the structure could impair DNA binding capacity which could lead to a lower efficiency in TF functioning (Wu et al., 1996; Krieger et al., 2022). 

Next, the deletions identified in AY961628, KC207813, KC207814, MG021307 and MG021308 from amino acids ranging from position 63 to 95 were also reported previously. Yalamanchili and group reported that the deletion of amino acids from position 2 to 88 resulted in minor effects on primary B lymphocyte transformation efficiency (Yalamanchili et al., 1996).  In addition, the deletion of amino acids from position 358 to 359 were also validated experimentally where amino acids from position 333 to 425 were identified to be deleted and this resulted in the impairment of lymphocyte transformation (Cohen et al., 1991). Besides, the deletion of amino acids from position 358 to 359 was also suggested to influence the RG domain of the EBNA-2 gene (Wang et al., 2012). As mentioned earlier, the changes in amino acid sequence could affect crucial structure EBNA2 (Wu et al. 1996; Krieger et al. 2022). This could lead to lower activity which could result in impairment of lymphocyte transformation, or these changes would result in insignificant changes thus not affecting EBNA2 crucial function. Further research is needed to truly elucidate the actual mechanism.

Conclusion

The divergence of the ENBA-2 gene increases concerning geographical regions when compared to the EBNA2 sequence of EBV in C666-1 NPC. Three major clades indicate high variation of the EBNA2 sequence across disease samples. Although differences were observed, some sequences showed close lineage especially with EBNA2 from BL, this includes sequence from GC and NPC, despite originating from different diseases. Thus, is it worth to further investigate on the mechanistic events that could arise from these variations which could help in further advancing the knowledge of EBV contribution towards disease development.  

Author Contributions

L.W.Z., R.B.S.M.N.M conceptualized and designed the study. L.W.Z., A.A., M.Y.M collected and analyzed the data, drafted the manuscript. All authors critically reviewed and approved the final version.

Acknowledgment

The authors acknowledged the Ministry of Higher Education, Malaysia for supporting this work through the Fundamental Research Grant Scheme Project Code (FRGS/1/2021/SKK03/USM/02/3) and for the facilities provided by the Universiti Sains Malaysia.

References