N-terminal VP1 truncations favor T=1 norovirus-like particles

Noroviruses cause immense sporadic gastroenteritis outbreaks worldwide. Emerging genotypes, which are divided based on VP1 sequence, further enhance this public threat. Self-assembling properties of the human norovirus major capsid protein VP1 are crucial for using virus-like particles (VLPs) for vaccine development. However, there is no vaccine available yet. Here, VLPs from different variants produced in insect cells are characterized in detail using a set of biophysical and structural tools. We are using native mass spectrometry, gas-phase electrophoretic mobility molecular analysis and proteomics to get clear insights into particle size, structure, composition as well as stability. Generally, noroviruses have been known to form mainly T=3 particles. Importantly, we identify a major truncation in the capsid proteins as a likely cause for the formation of merely T=1 particles. For vaccine development, particle production needs to be a reproducible, reliable process. Understanding the underlying processes in capsid size variation will help to produce particles of a defined capsid size presenting antigens consistent with intact virions. Next to vaccine production itself, this would be immensely beneficial for bio-/nano-technological approaches using viral particles as carriers or triggers for immunological reactions.

generally are known to assemble into VP1 180-mers with T=3 icosahedral symmetry. However, VP1 66 60-mers of T=1 symmetry have been described as byproducts of hNoVLP production coexisting 67 with other particle sizes and independent of the expression system (19). Recently, also VP1 240-68 mers of T=4 symmetry have been described so far only for GII.4 variants expressed in insect cells as 69 well as in plants (20,21). In studies on virions of different norovirus variants T=3 as well as T=1

79
In order to characterize hNoVLPs in detail and gain more insights into size determination, we 80 extended our previous native MS studies with a set of biophysical methods. Next to charge 81 detection mass spectrometry (CDMS) for mass determination of heterogeneous particle populations 82 and proteomics, we are using nano electrospray gas phase electrophoretic mobility molecular 83 analysis (nES GEMMA) (26), especially suited to measure high-mass particles at low concentrations 84 ((27, 28). Notably, this fast technique allows measurements at low ionic strength and with less 85 concentrated sample.

86
We describe particle preparations from insect cells with sample batches, where T=1 only 87 particles are detected. A major VP1 truncation was identified in all particle preparations forming 88 these T=1 particles. We find that this size-limitation is genogroup and genotype-independent and 89 cannot be rescued in different buffer conditions. This provides great implications for vaccine design 90 and other applications of bio-nanoparticles, where size-homogeneity is highly favored.

96
After transfection of a bacmid containing the recombinant VP1 gene in Sf9 insect cells and 97 incubation for 5-7 days, the culture medium was collected and centrifuged for 10 min at 3,000 rpm 98 at 4°C. Subsequently, Hi5 insect cells were infected with recovered baculovirus and incubated for 5 99 days. After centrifuging the culture medium for 10 min at 3,000 rpm at 4°C and then 1 h at 6,500 100 rpm at 4°C, VLPs in the supernatant were concentrated by ultracentrifugation at 35

176
After entering a heated metal capillary, ions are transmitted using various ion optics to a dual 177 hemispherical deflection energy analyzer, which selects ions with energies centered on 100 eV/z.

178
Subsequently, ions enter a modified cone trap where they oscillate back and forth in a charge 179 detection cylinder for 100 ms. Single ion masses were binned to generate mass spectra. Mass spectra 180 were analyzed by fitting Gaussian peaks with Origin software (OriginPro 2016).

181
Gas-phase electrophoresis was performed on a nES GEMMA instrument (TSI Inc, Shoreview, MN, 182 USA) consisting of a nES aerosol generator (model 3480) including a 210 Po α-particle source, an 183 electrostatic classifier (model 3080) with a nano differential mobility analyzer (nDMA) and an n-184 butanol based ultrafine condensation particle counter (model 3025A). Briefly, particle-size 185 determination is a function of the particles' trajectory in the nDMA chamber. The trajectory of a 186 size-specific particle is based on the sheath flow of particle-free ambient air and an orthogonal 187 electric field applied. Therefore, with a constant high laminar sheath flow of air and a variable 188 electrical field, only specific particle sizes can successfully be transported to the particle counter

196
The corresponding EMD size range was scanned for 120 s. Subsequently, the applied voltage was

T=1 capsid formation is genotype-independent 242
Additional hNoVLPs are investigated to pinpoint whether the truncation seen in GII.4 Saga 243 causes T=1 formation. Norovirus particle polymorphism has been described as putatively genotype-244 dependent. Therefore, we extend our sampling to GII.10 Vietnam and GII.17 Saitama (Figure 1). In 245 line with GII.4 Saga measurements, most abundant peak distributions are assigned to VP1 60-mers 246 for both variants. Notably, more acceleration energy compared to GII.4 Saga is needed to gain 247 charge-resolution for VP1 60-mer peaks, which indicates increased VP1 heterogeneity in these 248 samples (Supplement Figure S1). Furthermore, T=1 ions in GII.10 Vietnam show tailing with a non-249 resolved shoulder peak, indicating either aggregation or a further low-intensity assembly of slightly 250 higher mass. In GII.17 Saitama mass spectra, heterogeneity is even more prominent as multiple

280
To provide further evidence of truncation influence, we compare two GI.1 West Chester 281 batches. Batch 1 is identical to the sample used in our previous work (25). In the second batch, no 282 T=3 particles are detected at neutral pH using native MS (Figure 2). The main peak distribution is 283 assigned to T=1 particles, which is accompanied by a low-intense shoulder peak comparable to 284 GII.10 Vietnam. At increased acceleration voltage, the T=1 ions release VP1 monomers. A close-up 285 of these monomers shows that there are at least two subspecies present. This directly contributes to 286 heterogeneity and therefore low peak resolution of higher-mass species. The dominating VP1 287 species is assigned to 52760 ± 10 Da, or the theoretical VP1 mass lacking 40 N-terminal aa.

288
Proteomics data, which hint to subspecies with minor truncations, is consistent with other variants 289 tested in this study (Table 1). In our previous study, we could identify GI.1 West Chester VP1 290 monomers with the major species lacking only three amino acids forming mainly T=3 particles (25).

291
Taken together, we can assume that with the VP1 N-terminus of GI.1 West Chester lacking three       Figure S3). For all samples, particle-size patterns are also consistent at pH 5, 359 although with lower particle counts and with increased background noise. Taken together, this 360 indicates that T=1 particles are highly stable resisting alkaline pH and T=3 particle formation cannot 361 be rescued by changing solution conditions.     Figure S4). In GII.17 Saitama, CDMS helps to elucidate

494
However, in a follow-up study, GI.4 Chiba mutants were shown to form 23 nm or T=1 particles, 495 putatively due to freezing and thawing of preparations or pH-dependent processes (52). Previously,

496
T=4 particles were identified in hNoVLP preparations (20,21). Interestingly, one study included 497 GII.2 Snow Mountain virus forming T=1 particles (21). Here, residues 1 to 46 were not covered in 498 electron density maps. Hence, truncation as origin of small particles similar to our observations 499 cannot be excluded.

500
Next to particle size distribution, the influence of solution pH was investigated. T=1 particles, 501 as well as higher-mass assemblies in GII.17 Saitama, were found to be pH-independent. Moreover,

502
in preparations forming mainly T=1 particles, like GI.1 West Chester batch 2, T=1 particles showed 503 increased stability in alkaline conditions. Therefore, truncated VP1 is able to build particles with 504 increased stability. This implies great advances for bionanotechnology as especially in approaches 505 using VLPs as carrier particles need to be stable independently of environmental conditions. The contribution of the N-terminus to pH stability suggests a way to obtain S-particles of increased 507 stability by truncation.

508
There is no hNoV vaccine available yet and hNoVLP size polymorphism could contribute to 509 this circumstance. Therefore, N-terminally truncated particles have great potential to be beneficial 510 as they imply size homogeneity. Furthermore, these particles have intact protrusions, which are 511 necessary for antigen recognition and putatively enable other immunological approaches like 512 antigen presentation. However, whether the altered orientation, and therefore interaction between 513 dimeric protrusions affects antibody raising, needs to be investigated. Additionally, increased 514 stability would likely allow simplified and prolonged storage. Our results indicate that such small 515 particles from truncated VP1 can be produced independent of genotype by introducing N-terminal 516 deletion mutants.