key: cord-0333687-dgqzdqrd
authors: Mathieu, Cyrille; Figueira, Tiago Nascimento; Decker, Amanda R.; Ferren, Marion; Bovier, Tiziana F.; Jurgens, Eric M.; Marcink, Tara C.; Moscona, Anne; Porotto, Matteo
title: Measles fusion complexes from central nervous system clinical isolates: decreased interaction between hemagglutinin and fusion proteins
date: 2021-06-19
journal: bioRxiv
DOI: 10.1101/2021.06.18.449082
sha: 84f73432387e3b8cc695c9220564af203834484f
doc_id: 333687
cord_uid: dgqzdqrd

Measles virus (MeV) viral entry is mediated by a fusion complex comprised of a receptor binding protein (hemagglutinin, H) and a fusion protein (F). The wild type H/F complex requires interaction with specific proteinaceous receptors (CD150/SLAM and nectin-4) in order to be activated. In contrast the H/F complexes isolated from viruses infecting the central nervous system (CNS) do not require a specific receptor. A single amino acid change in the F protein (L454W) was previously identified in two patients with lethal sequelae of MeV CNS infection, and the F bearing this mutation mediates fusion even without the H protein. We show here that viruses bearing the L454W fusion complex are less efficient than wt virus at targeting receptor expressing cells and that this defect is associated with a decreased interaction between the H and the F proteins. Importance Measles (Mev) infection can cause serious complications including measles inclusion body encephalitis (MIBE) and subacute sclerosing panencephalitis (SSPE). MIBE and SSPE are relatively rare but lethal. We have shown that the fusion complex of CNS adapted clinical samples can spread in the absence of known receptor. We now provide evidence that HRC mutations leading to CNS adaptation come at a cost to the efficiency of viral entry. One Sentence Summary Measles CNS adapted fusion complexes have altered H/F interaction.

We previously suggested that the ability of HRC mutants to fuse in the absence of 101 known receptors could be at least partly due to reduced interaction between H and F(34). The 102 interaction between F and H and the impact of mutations in the HRC domain (i.e., L454W, 103 T461I and N462 may be affected by thermal instabilityobscuring the effect of the specific 104 mutations --we included the highly unstable mutant F-S262R (11, 12, 33, 35) which interacts 105 efficiently with H. H-F interaction was assessed with both the cleavable F proteins ( Fig. 2A ) 106 and the cleavage site mutant (CSM) version of these F proteins (Fig. 2B) , to distinguish 107 between effects on the F 0 and the F 1 form of F. Results are presented as the average of 3 108 separate experiments ± SD (Fig.2C ). Densitometric analysis was used to normalize the results 109 to total protein content and to convert the data to graphs (average of 3 separate experiments ± 110 standard deviation; Fig. 2C ). There is significant reduction in the amount of F 1 co-111 immunoprecipitated with H for the HRC hyperfusogenic mutants (i.e. , L454W, T461I and 112 N462K) compared to the wt F (P = 0.0023, P = 0.0030 and P = 0.0014, respectively; two-way 113 ANOVA with multiple comparisons against wt F, corrected with the Dunnett hypothesis test). 114

For the other hyperfusogenic mutant, F S262R, there was more co-immunoprecipitated F 1 115 protein when compared to the HRC mutants. There were no differences in co-116 immunoprecipitation of F 0 . Thus, mutations in HRC associated with a hyperfusogenic 117 phenotype decrease the ability of H to interact with the active F 1 form of F. 118 119 Viral evolution leads to compensatory mutation 120

MeV bearing the L454W F protein, when grown at 37 o C in cell culture, acquired a 121 compensatory mutation (E455G) that re-balanced F's stability and dependence on H and 122 cellular receptor for mediating fusion (12). The E455G mutation is shown in Fig. 3A , alone 123 and in combination with L454W. Tryptophan (W454) contains a bulky aromatic sidechain when compared to leucine (L454), while glutamic acid (E455) is larger than glycine (G455). 125

We determined whether the E455G mutation restored H-F interaction (Fig. 3) . H-F 126 interaction for wt ,L454W, E455G, and L454W/E455G F proteins is shown in Fig. 3 . 127

Densitometric analysis was used to normalize the results to total protein content and to 128 convert the data to graphs (average of 3 separate experiments ± SD; Fig. 3C ). The wt versus 129 the L454W F was significantly different (as seen in Fig.2 

One of the consequences of the loss of interaction between F and H is the loss of 138 protection of F from spontaneous triggering(37). Additionally, hyperfusogenic MeV F HRC 139 mutants proteins are less stable in their pre-fusion state, as measured by their temperature 140 sensitivity, compared to the wt F protein. The hyperfusogenic F proteins mediate fusion even 141 in the absence of either CD150/SLAM or nectin-4 receptors(10, 34). We have shown that this 142 decreased F stability, however, results in viruses that are inactivated at lower temperatures 143 than wt virus (10), because F transitions to its post-fusion state more readily, rendering the 144 viral particle non-infectious. We asked whether the destabilizing mutations in F of the 145 neuropathogenic variants affect viral entry into SLAM-expressing cells (38). To do so, we 146 used an assay developed for a related paramyxovirus, human parainfluenza virus 3 (HPIV3). 147

For HPIV3 we have shown that the kinetics of F protein activation and viral entry modulate 148 the potency of HR derived peptides(37, 39), and therefore this modulation can be used as a tool to quantitate F activation and entry. Applying this strategy to MeV, we determined the 150 sensitivity of MeV F to inhibition by MeV HR derived peptides, using a previously described 151 dimeric HR lipopeptide, HRC4. In Fig. 4A mechanism of ongoing activation of the fusion process by receptor-engaged H could explain 169 our observations for MeV. We showed that constant interaction with F by receptor-engaged H 170 may allow wt H-F to fuse even in the presence of F-targeted anti-MeV peptides (40). As 171 above, we also previously showed that for HRC peptide fusion inhibitors, adding a lipid 172 moiety improves antiviral potency over time (40) . In this case a cholesterol-conjugated dimeric peptide (HRC4) was inhibitory as long as 6h after in vitro (41). Without a lipid moiety 174 (HRC1), peptide was effective only at early time points after infection. 175

We hypothesize that the loss of of H-F interaction affects both the triggering of HRC 176 mutants and their ability to fuse in the presence of inhibitor. In Fig. 4B we assessed weak 177 peptide inhibitor (HRC1) for inhibition of wt and mutant F fusion. The assay measures the 178 fusion of cells that express viral envelope glycoproteins (MeV IC323 H/F) with cells that 179 express the MeV receptor SLAM. HRC1 peptide inhibited fusion at early time points (1h, Fig.  180 4 B). As expected, this inhibition decreased below 70% after 3h and below 40% for the wt F 181 and S262R F respectively after 6h. In contrast, for the H/L454W F fusion complex, HRC1 182 peptide inhibited over the time course of this fusion assay, significantly better than for the wt 183 complex (**p value, Mann-Whitney U-Test). A parallel assay was performed with the HRC4 184 peptide, which completely inhibited fusion of all MeV Fs (Fig. S1) . properties in the context of the L454W-bearing F , and when introduced singly into the F 228 protein led to an extremely stable F(12) . We speculate that introducing W454 into the 229 prefusion F stalk results in local destabilization that enhances F activation, even in the 230 absence of receptor. The nearby mutation G455 may alter the structure sufficiently to restore 231 the "wt" properties. When the viral quasispecies formed by L454W and L454W/E455G F 232 bearing viruses is cultivated in human brain organoids , the L454W/E455G F bearing virus is 233 eliminated within 10 days confirming that the L454W F bearing virus is fit for growth in CNS 234 tissues(12). The virus bearing the S262R mutation in F is neuropathogenic in vivo (11, 35), 235 and we observed no significant differences in H-F interaction. A S262G mutation in F (along 236 with several other mutations) was found in virus from a SSPE clinical case (8). Thus, while 237 destabilizing the F protein significantly improves MeV's ability to spread in the brain, the loss 238 of H-F interaction may not be necessary for CNS adaptation. 239

MeV H-F interaction is key for the steps after F insertion in the target membrane, to 240 complete fusion. The viruses bearing F mutations that lead to CNS spread are more 241 susceptible to fusion inhibitory peptides(10). We attribute this, at least in part, to the fact that 242 the H does not continue to activate F after initial triggering. 243

In Fig. 5 we propose models of viral infection with wt MeV and the CNS adapted 244 variants, incorporating the findings in this work as well as work from others. A common 245 pattern for neuropathogenic F variants is decreased stability of F (as assessed by sensitivitiy to 246 heat) and a fusion complex that mediates cell-to-cell fusion in the absence of either 247 CD150/SLAM or nectin-4(8, 10, 12, 34, 44) . These properties confer an infectivity cost, since 248 they are significantly slower to complete the fusion process after inital triggering. These 249 features together with lower viability (10) make these viruses adapt to cell-to cell transmission 250 after initial infection but also likely mean that they are less transmissible. (unpaired T-test) between the levels of F 0 that were co-294 immunoprecipitated for each allele. There was also no statistical difference between F(wt) and 295 either allele that contained E455G (single and double mutant) for F 1 levels. However, the 296 difference between the F(wt) and F(L454W) F 1 that co-immunoprecipitated with H was 

A 418 dangerous measles future looms beyond the COVID-19 pandemic

SLAM (CDw150) is a cellular receptor 421 for measles virus

423 Tumor cell marker PVRL4 (nectin 4) is an epithelial cell receptor for measles virus

Adherens junction 428 protein nectin-4 is the epithelial receptor for measles virus

Measles virus, immune control, and persistence

432 Measles

Subacute 434 sclerosing panencephalitis in South African children following the measles outbreak 435 between

South African Measles Epidemic Shows That Hyperfusogenic F Proteins 440 Contribute to Measles Virus Infection in the Brain

Molecular characterisation 442 of virus in the brains of patients with measles inclusion body encephalitis (MIBE)

Measles Virus Bearing Measles Inclusion 447 Body Encephalitis-Derived Fusion Protein Is Pathogenic after Infection via the 448 Respiratory Route

New Insights into 450 Measles Virus Brain Infections

Molecular Features of the

Measles Virus Viral Fusion Complex That Favor Infection and Spread in the Brain

Measles virus 457 receptor SLAM (CD150)

Measles virus receptors. Current 459 topics in microbiology and immunology

Measles virus hemagglutinin: structural 461 insights into cell entry and measles vaccine

Structure of the 463 uncleaved ectodomain of the paramyxovirus (hPIV3) fusion protein

Paramyxovirus membrane fusion: 466 lessons from the F and HN atomic structures

Structure of the 468 parainfluenza virus 5 F protein in its metastable, prefusion conformation

Viral membrane fusion

Paramyxovirus fusion and entry: multiple paths to a 472 common end

Structural and mechanistic studies of 474 measles virus illuminate paramyxovirus entry

Structures and mechanisms of 476 viral membrane fusion proteins: multiple variations on a common theme

Design of potent inhibitors of HIV-1 entry from the gp41 482 N-peptide region

Molecular determinants of antiviral potency of paramyxovirus entry inhibitors

Peptide inhibitors of flavivirus entry 487 derived from the E protein stem

Hendra and nipah infection: pathology, models and potential 489 therapies

A general strategy to endow 492 natural fusion-protein-derived peptides with potent antiviral activity

Henipavirus mediated membrane 495 fusion, virus entry and targeted therapeutics

The Current Status and 499 Challenges in the Development of Fusion Inhibitors as Therapeutics for HIV-1 500 Infection. Current pharmaceutical design

502 Capturing a fusion intermediate of influenza hemagglutinin with a cholesterol-503 conjugated peptide, a new antiviral strategy for influenza virus

Measles Virus Spread between Human Neurons Is Dependent on Hemagglutinin and 507 Hyperfusogenic Fusion Protein

Measles fusion machinery is dysregulated in neuropathogenic variants. mBio 6

Mutant 511 fusion proteins with enhanced fusion activity promote measles virus spread in human 512 neuronal cells and brains of suckling hamsters

Structures of the prefusion 515 form of measles virus fusion protein in complex with inhibitors

Spring-loaded model revisited: paramyxovirus fusion requires 519 engagement of a receptor binding protein beyond initial triggering of the fusion 520 protein

In Vivo 523 Efficacy of Measles Virus Fusion Protein-Derived Peptides Is Modulated by the 524 Properties of Self-Assembly and Membrane Residence

Kinetic dependence of paramyxovirus entry inhibition

Prevention of measles virus infection by intranasal 529 delivery of fusion inhibitor peptides

531 Fatal measles virus infection prevented by brain-penetrant fusion inhibitors

Broad spectrum 535 antiviral activity for paramyxoviruses is modulated by biophysical properties of fusion 536 inhibitory peptides

Measles virus fusion machinery activated by 538 sialic acid binding globular domain

Measles 540 virus mutants possessing the fusion protein with enhanced fusion activity spread 541 effectively in neuronal cells, but not in other cells, without causing strong 542 cytopathology

Peptides: MeV F derived fusion inhibitor peptides HRC1 and HRC4 were previously 353 described (40). Briefly 36aa long peptides derived from the heptad repeat region at the C-354 terminal of the MeV F protein were synthesized (using the wt sequence or the L454W 355 sequence). Monomeric unconjugated (HRC1) or dimeric cholesterol conjugated (HRC4) 356 forms of the peptides were used in this study. MeV IC323-EGFP-F T461I, MeV IC323-EGFP-F N462K all these viruses are from(34)) was 367 used to infect sub-confluent VERO-SLAM cells in 6 well plates (100pfu/well) for 2h at 32°C. 368MeV HRC4 dimeric fusion inhibitory peptide (1µM) was added to the medium at time points 369 from 15 to 120 min after the beginning of infection. After 2h of incubation with the virus, 370 medium was replaced with medium containing Avicel. Viral titers were assessed after 3 days 371 of incubation at 32°C by immune staining. 372

complementation-based fusion assay was performed as described previously(34). Briefly, 374293T cells transiently transfected with the constructs indicated above and the omega reporter 375 subunit were incubated for the indicated period with cells coexpressing viral glycoproteins 376 and the alpha reporter subunit in presence or not of MeV F HRC derived peptide (40). 377

expressing MV H_Y17H HPIV3_T193A chimerae (43) protein expression was synchronized with cycloheximide (Sigma, 0.1 mg/mL), followed by 397 membrane protein cross-linking with 3,3'-dithiobis(sulfosuccinimidyl propionate) (Sigma; 1 398 mM), at low temperature. The cross-linking reaction was quenched with 20 mM Tris, 150 399 mM NaCl, pH 7.5. Cells were lysed with 50 mM HEPES, 100 mM NaCl, 0.05 g/mL dodecyl 400 maltoside, pH 7.5, supplemented with complete protease inhibitor cocktail (Roche). Lysates 401 were centrifuged at 16000 g for 10 min to remove nuclei and cell debris, and the supernatant 402 was collected for immunoprecipitation and total protein content analysis. MeV H-6xHis 403 protein was immunoprecipitated from cell lysates using Dynabeads® (Thermo, 1 mg/mL) 404 coated with a 6xHis tag-specific antibody (mouse monoclonal, Thermo, MA1-21315). 405For Fig.2 : Co-immunoprecipitated and cell lysate proteins were analyzed by western blotting, 406 using primary antibodies specific for MeV F HRC (rabbit polyclonal, Genscript, 503028-1) 407 and 6xHis tag (rabbit polyclonal, Thermo, PA1-983B), followed by an HRP-conjugated anti-408 rabbit secondary antibody (Kindle Biosceiences, R1006). Western blots were developed using 409 the SuperSignal West Femto substrate (Thermo) and imaged on a KwikQuant TM Imager UV 410 (Kindle Biosciences). 411 For Fig. 3 : Co-immunoprecipitated and cell lysate proteins were analyzed by western 412 blotting, using primary antibodies specific for MeV F HRC (rabbit polyclonal, Genscript, 413 503028-1) and 6xHis tag (rabbit polyclonal, Thermo, PA1-983B), followed by 414 WesternBreeze Chromogenic Immunodetection Protocol for detection. 415 416 Literature cited 417