Rational Design of Peptide Vaccines for the Highly Lethal Nipah and Hendra Viruses

The Nipah virus disease is a lethal infection that has led to 40% to 75% fatalities in Malaysia, Bangladesh and India. The reports of human-to-human transmission documented in Bangladesh has raised the specter of pandemic potential and has caused the World Health Organization to list the Nipah virus as one of the pathogens to be considered for development of drugs and vaccines on urgent basis, neither of which exist against the Nipah virus as of now, although many proposals have been made and trials initiated. Given that there are established country-specific differences in the virus’ effects and fatalities, meeting the sudden need for a vaccine in case of an epidemic will require design, development and preparation for a peptide vaccine. Thus, we propose a protocol for creating peptide vaccines that can be tailor-made for these specific countries, an approach which is being advocated for the first time. Here, we analyze the surface proteins, Fusion protein and Glycoprotein, of the strains currently affecting the three countries on a large scale and determine the specific country-based epitope differences.


Introduction
A lethal zoonotic paramyxovirus with a possibility of pandemic potential is affecting wide 23 swathe of tropical populations from Philippines to India to Madagascar [1]. First isolated from a 24 patient in Kampung Sungai Nipah village of Malaysia in 1998 [2], the Nipah virus (NiV) caused where the infection is reported to have spread through contaminated fruits and date palm sap [3]. 30 The infection has become since an annual occurrence there with occasional infections in 31 neighbouring India [4], and lately in Kerala [5], far removed from Bangladesh. In 2014, the 32 disease was also identified in the Philippines where a number of horses, and some people in 33 contact with the horses, died from the infections [6]. Similar viruses, not yet human infecting, 34 have been found in bats in Africa from Madagascar to Ghana [1], which indicates that the 35 geographical range of the infectious disease could expand in future. 36 Nipah was determined to be closely related to the Hendra virus (HeV), first identified in horses 37 in Australia, the two viruses forming a subspecies, the Henipavirus, in the Paramyxoviridae 38 family, although the HeV is not known to infect humans the way the NiV has been seen to do, 39 but has led to some deaths in Australia. The virus causes respiratory and encephalitic disorders in 40 symptomatic individuals leading to death in severe cases; in Malaysia the case-fatality ratio 41 (CFR) was about 40%, but in Bangladesh and India the CFR was between 70% and 100%. The 42 pathogenicity of the two strains of NiV seem to be different, with the Bangladesh strain being 43 more lethal. This was also borne out by experiments with ferrets [7] and African green monkeys 44 [8], showing that any post-infection therapies need to be specific to the strains involved. 45 The disease is transmitted through physical contact with bodily secretions and excretions of 46 infected animals, but more importantly, human-to-human infections through exhaled droplets 47 from respiratory organs have been known to occur in about 50% of the cases in the Bangladesh 48 infections [9]. So far with NiV infections being in localized remote villages, the current spread of 49 the highly lethal viral infections is as yet limited in numbers. However, RNA viruses being prone 50 to rapid mutational changes, it could be a matter of time before the virus may mutate to a more 51 efficient human-to-human transmissibility, which, coupled with the density of population in  paramyxoviruses such as measles and mumps provide a reason to expect that a vaccine against 78 the henipaviruses will also lead to positive results, although the extreme lethality of the Nipah 79 virus remains a cause for concern. Of the various types of vaccines that can be developed, live  situated G protein as the most appropriate target since it is responsible for attachment to the 115 ephrin receptor in host cells. Since the Nipah viral sequence showed very strong, long stretches 116 of conserved domains that matched HeV's conserved stretches also, it was natural to club the 117 two together to design a common vaccine that matched all our criteria. Using our protocol, we 118 identified conserved segments on the viral protein which are sufficiently surface exposed, 119 determined their epitope potential in the host population and tested for auto-immune threats. We 120 report here on the results for the peptide targets for the Malaysian, Bangladeshi and Indian 121 (Keralite) populations as a ready guide for developing appropriate vaccines. Such overall 122 country/community-wise epitope search is, to the best of our knowledge, being reported for the 123 first time.

125
We downloaded all the available complete sequences of the Nipah virus glycoprotein (20) and 126 the Hendra virus glycoprotein (15) and the Nipah glycoprotein structure, 2VWD, from the NCBI 127 databases.

128
For numerical comparison of protein sequences, we used a sequence representation described 129 earlier [27] where in a rectangular co-ordinate system of 20-dimensions each amino acid is 130 associated with a particular direction and the protein sequence is plotted by taking successive 131 steps in the directions dictated by the amino acids. This generates a plot in an abstract space, but 132 we can define a weighted center of mass and a protein graph radius, p R . The special property of 133 p R is that in each case when two sequences result in the same graph radius the sequences are 134 found to be identical (28). Using a n-amino acid window (n < N, the total length of the protein Hendra viruses are RNA type which therefore is comparatively less stable, we fixed window 159 length for this exercise at 12 amino acids. 160 The protein sequences are next mapped for solvent accessibility through an Average Solvent  To see whether our selected peptide regions can act as epitopes and can able to elicit necessary 181 immune response, we use another epitope-prediction tool ABCpred server 182 (http://www.imtech.res.in/raghava/abcpred/ABC_submission.html) [36]. These tools indicate 183 binding affinities of B-cell epitopes (with 66% accuracy). 184 We also needed to analyze the role of human MHC polymorphism in the Nipah disease severity.   Table 1.  Using the 2VWD crystal structure from the Protein Data Base (PDB) we determined that regions 206 3 and 4 above are peptides that are highly surface exposed (Fig.2); while we show the peptides 207 on one monomer of the dimeric structure, it is clear that those are not covered by the neighboring 208 protein. Unfortunately, since the PDB structure starts from aa number 187, we could not visually 209 demonstrate that regions 1 and 2 are also surface exposed, but on the basis of the ASA profile  (Table 1) with a latitude 234 of 4 to 5 residues to ensure good binding with the HLA alleles while remaining within the 235 requirements of reasonably well conserved regions and high ASA; references to the crystal 236 structure ensures the peptides are still surface exposed even if they are slightly adjusted from the 237 regions as per Table 1.

238
The HLA-DRB alleles, which are the dominant types, are shown in Table 2a   between. All regions show one or more alleles with good binding, region 4 just outside the 10-244 percentile rank for all 3 countries. All peptides of regions 3 and 4 (aa numbers 300+) were found 245 to be surface exposed except for last two residues of Malaysian region 3 peptide (Table 2a) Bangladesh viral sequence and variation there also in case of the Malaysian sample.

252
The HLA DP/DQ allele profiles are available for India but not for Bangladesh and Malaysia. The 253 results for Indian HLA alleles as determined through IEDB are given in Table 2b for the peptide the best available for the surface exposed conserved segments we are interested in.   The peptide sequences we determined to be conserved surface exposed with good epitope 300 potential were all tested for auto-immune threats by doing a protein-protein BLAST but no 301 homology was found for any human proteins.  the Nipah virus as mentioned in the Introduction, these are all "one size fits all" type of vaccines.

321
What we are advocating here are peptide vaccines that can be tailor-made to suit individual 322 countries' populations, which will be entirely possible on a "manufacturing" kind of scenario, 323 something that is not possible, or would take an inordinately long time, with live-attenuated or 324 VLP vaccines. If nothing else, peptide vaccines should be tested and got ready to combat any