Characterization of conserved residues in the mammarenavirus matrix protein Z using novel Lassa virus life cycle modelling assays

TL;DR

Researchers developed a new system to study Lassa virus life cycle and identified key amino acids in the virus's matrix protein that are crucial for its replication and interactions.

Contribution

A novel trVLP system was developed to model the LASV life cycle outside high-containment labs, revealing essential conserved residues in the Z protein.

Findings

Residues L71 and P72 are essential for Z protein's role in inhibiting viral RNA synthesis.

Residues L71-T73 mediate interaction of Z with the viral nucleoprotein (NP), supported by residue G2.

Abstract

Lassa virus (LASV) is the causative agent of Lassa fever, which causes thousands of deaths every year within the endemic regions of West Africa. Despite its discovery over half a century ago, our knowledge regarding basic steps in the LASV life cycle remains limited. In order to simplify studying basic principles of the LASV life cycle, we developed a novel transcription- and replication-competent virus-like particle (trVLP) system that can model the whole virus life cycle using only authentic LASV components but without the need for a maximum containment laboratory. As a proof-of-concept we used this system, together with classical minigenome systems, to determine the functional contributions of highly conserved amino acids within the mammarenavirus matrix protein Z. We could demonstrate that residues L71 and P72 are essential for the role of Z in inhibiting viral RNA synthesis, with residues R16, D22, K68 and T73 also contributing to this function. Additionally, we showed that L71-T73 mediate interaction of Z with the viral nucleoprotein (NP), which is also supported by residue G2. Given the critical role of these interactions in the virus life cycle, understanding these amino acids’ functions will help to define potential targets for the development of urgently needed countermeasures.

The online version contains supplementary material available at 10.1038/s41598-026-40023-6.

Figures

Establishment and optimization of the LASV minigenome assay. (A) Schematic of the LASV monocistronic minigenome assay. The monocistronic minigenome is derived from a version of the LASV S segment in which the ORFs were deleted and the GPC ORF was replaced by a reporter gene (i.e. nanoluciferase), as indicated. Transfection of cells with plasmids encoding for the minigenome as well as the T7 polymerase leads to initial transcription of the minigenome. Additional transfection of plasmids encoding the LASV nucleoprotein (NP) and polymerase (L) enables secondary transcription and replication of th

Alignment of Z amino acid sequences from different mammarenaviruses. Sequences of representative mammarenaviruses from across different LASV lineages as well as other members of both the Old World and New World arenaviruses were aligned using Clustal Omega. Sequences are arranged by clades, with the black box containing Old World mammarenaviruses and grey boxes each representing one clade of the New World mammarenaviruses, as indicated (i.e. A, B. C or recombinant A/B). For the different LASV isolates, lineage numbers are indicated by Roman numerals. The RING domain is shown in a red box and l

Role of conserved amino acids in LASV Z in RNA synthesis inhibition. (A) Titration of Z. Huh7 cells were transfected with all the components for a LASV minigenome assay in addition to increasing amounts of either pCAGGS-Z or pCAGGS-Z-myc. The dashed line shows the background of the assay as determined by a control lacking the viral polymerase (-L). Normalized reporter activities with standard deviations of three independent experiments are shown. (B) Expression levels of Z mutants. Huh7 cells were transfected with plasmids encoding wildtype (WT) Z-myc or Z mutants with a c-myc-tag, as indicate

Establishment of the LASV transcription and replication-competent virus-like particle (trVLP) assay. (A) Schematic of the LASV trVLP system. The multicistronic minigenome is derived from the LASV S segment, in which the NP and GPC ORFs were replaced with a GPC-T2A-nanoluciferase and Z ORF, respectively. Transfection of producer (p0) cells with the plasmid containing the multicistronic minigenome as well as the T7 polymerase leads to initial transcription of the minigenome. Additional transfection of plasmids encoding the LASV nucleoprotein (NP) and polymerase (L) enables replication and second

Role of conserved amino acids in LASV Z in budding and RNP incorporation. (A) Influence of Z-mutants on trVLP assay reporter activity in producer (p0) cells. Huh7 cells were transfected with all the components for a trVLP assay including multicistronic minigenomes encoding the different LASV Z-mutants, as indicated. As negative controls either a minigenome lacking the Z ORF (ΔZ) was used, or the minigenome was omitted entirely (-mg). Nanoluciferase reporter activity was normalized to FF activity. (B) Influence of Z-mutants on trVLP assay reporter activity in target (p1) cells. Huh7 cells (pre-

Role of conserved amino acids in Z in NP-Z interaction. (A) CoIP analysis of NP-Z interaction. Myc-tagged LASV Z mutants were overexpressed together with HA-tagged LASV NP in 293T cells. The cells were lysed and the proteins subjected to co-immunoprecipitation (CoIP) using anti-myc antibodies. Input and CoIP samples were analyzed via SDS-PAGE and Western blot. Z was detected using an anti-myc antibody and NP was detected using an anti-HA antibody. (B, C) Quantification of band intensities. Band intensities in input (B) and CoIP (C) samples were quantified. CoIP band intensities for NP were nor

Full text

Introduction

Lassa virus (LASV) is an Old World mammarenavirus and the causative agent of Lassa fever, a hemorrhagic fever disease endemic to West Africa^1^. Estimates suggest that LASV leads to 300,000-500,000 cases of Lassa fever annually with approx. 5,000 deaths, thus representing a notable disease burden in the endemic region^2^.

LASV is an enveloped virus with a bi-segmented, single-stranded negative-sense RNA genome^3^. Each segment contains two open reading frames (ORFs) in an ambisense arrangement, which are then flanked by terminal untranslated regions (UTRs) and separated by an intergenic region (IGR) (Fig. 1A). The UTRs at the genome ends are crucial for regulating viral RNA replication and gene expression, while the IGR has a hairpin-like structure and regulates transcription^4–6^. The S segment encodes the nucleoprotein (NP) and glycoprotein precursor (GPC), whereas the L segment encodes the viral RNA-dependent RNA polymerase (L) and matrix protein (Z)^3,7^. In the viral particle, the genome is encapsidated by NP and associated with L, which together form the ribonucleoprotein complex (RNP), while Z interacts with NP and forms the matrix layer beneath the envelope membrane^8^. GPC is post-translationally cleaved into three subunits: GP1, GP2 and SSP (stable signal peptide), which remain associated and are further assembled into heterotrimers on the virion surface^9–12^.

Fig. 1. Establishment and optimization of the LASV minigenome assay. (A) Schematic of the LASV monocistronic minigenome assay. The monocistronic minigenome is derived from a version of the LASV S segment in which the ORFs were deleted and the GPC ORF was replaced by a reporter gene (i.e. nanoluciferase), as indicated. Transfection of cells with plasmids encoding for the minigenome as well as the T7 polymerase leads to initial transcription of the minigenome. Additional transfection of plasmids encoding the LASV nucleoprotein (NP) and polymerase (L) enables secondary transcription and replication of the minigenome. A plasmid encoding Firefly luciferase (FF) measures transfection efficiency, cell viability and plasmid driven gene-expression, and is used as a normalization control. **(B) **Cell line testing. Huh7, BsrT7/5 and 293T cells were transfected with all components for a LASV or JUNV minigenome assay. As a control, the viral polymerase was omitted (-L). Numbers above each pair of -L and + L samples indicate the difference between these samples (in log RLU) for that cell line. **(C) **Introduction of decoy ORFs. Huh7 cells were transfected with all components for a JUNV or LASV minigenome assay using plasmids encoding the minigenomes with (1L2L) or without (w/o) decoy ORFs. **(D) and (E) **Titration of NP and L. Huh7 cells were transfected with all components for a LASV minigenome assay, but with increasing amounts of pCAGGS-NP (D) or pCAGGS-L (E). Normalized reporter activities with standard deviations from three (C-E) or four (B) independent experiments are shown.

Z fulfills multiple roles in the viral life cycle. Its primary function is mediating budding of virus particles at the plasma membrane, which for many mammarenaviruses, including LASV, is a process that can also be observed in the absence of other viral proteins^7,13,14^. Additionally, Z regulates viral RNA synthesis in a dose-dependent manner by interacting with L^15,16^. Several important regions of the protein related to these functions have previously been described. In particular, the myristoylation of residue G2 is essential for both the interaction of Z with GP and for the insertion of Z into the membrane in order to mediate budding of new virions, and recent results indicate that non-myristoylated Z undergoes proteasomal degradation^14,17–19^. Furthermore, Z has a central RING (Really Interesting New Gene) zinc finger domain. For the related Lymphocytic Choriomeningitis Virus (LCMV) the RING domain is involved in the downregulation of viral RNA synthesis, with other regions supporting this function^16,20^. Moreover, data for the New World arenaviruses Tacaribe virus (TCRV) and Junín virus (JUNV) show an involvement of the RING domain in the interaction of Z with both NP and L^21–23^. Finally, two C-terminal late domains in LASV Z are required for efficient budding by interacting with components of the cellular endosomal sorting complex required for transport (ESCRT)^7,13^.

Consistent with their involvement in essential functions of Z, the G2 myristoylation site, RING domain and late domain motifs are all highly conserved among mammarenavirus matrix proteins from different clades and lineages. In addition, several other amino acids are rather well conserved across the genus of mammarenaviruses, and some previous efforts were undertaken to identify their functional relevance^24^. However, these studies were performed using LASV Z together with RNP components from the heterologous LCMV, and thus needed to be confirmed in authentic LASV systems. Therefore, in order to be able to systematically study the relevance of these conserved amino acids using only LASV components, we developed life cycle modelling systems for LASV that allow analysis of all steps of the virus life cycle.

Life cycle modelling systems offer the unique advantage of investigating the life cycle of highly pathogenic viruses without the need for maximum containment facilities^25^. More importantly, in contrast to infection experiments, these systems facilitate the targeted investigation of specific steps in the viral life cycle. To accomplish this, life cycle modelling systems use minimal versions of the viral genome, known as minigenomes, in which some or all ORFs have been removed and replaced by one (or more) reporter ORF(s). The 3’ and 5’ UTRs and, in the case of arenaviruses, also the IGR are retained since they contain cis-acting regulatory elements required for viral RNA synthesis. Transfection of cells with the plasmids encoding the minigenome and the viral RNP proteins (i.e. NP and L) leads to viral RNA synthesis and, ultimately, expression of the reporter protein, thereby modelling viral RNA synthesis and protein expression. An advancement on such a minigenome system is the transcription- and replication-competent virus-like particle (trVLP) system, which can model the whole viral life cycle^21,26^. This can be achieved using a multicistronic minigenome, which additionally encodes Z and GP - the viral proteins necessary for budding and formation of particles^27,28^, which can then be used for the infection of new target cells.

In this study we describe the establishment and optimization of a LASV trVLP system. We then use this system, together with the LASV minigenome system, to characterize highly conserved amino acids in the matrix protein Z, and show that many of these residues play important roles in the LASV life cycle.

Results

Generation and optimization of the LASV minigenome assay. In a first step, we generated a T7-driven luciferase-expressing minigenome system based on LASV (strain Josiah) and optimized the assay conditions for robust activity. The minigenome design was conceptually similar to previously established T7-driven minigenome systems for LASV (strain AV) and JUNV^15,29,30^, and used the same optimized plasmid backbone previously described for JUNV^31^. To generate the monocistronic minigenome, the ORFs of the S segment were removed and the GPC ORF was replaced with a nanoluciferase reporter gene (nluc) (Fig. 1A). The minigenome and the plasmids encoding the T7 polymerase and the viral RNP proteins (i.e. NP and L) were transfected into three different cell lines to determine the optimal cell line for this assay: the human Huh7 and 293T cells, as well as the hamster cell line BsrT7/5, which has often been used for these systems (Fig. 1B)^32^. As a control, the viral polymerase was omitted (-L). A previously established JUNV minigenome assay was used as a comparison to judge assay performance. For JUNV, we confirmed prior results that transfection of 293T cells results in low reporter activities, especially when compared to BsrT7/5 cells^31^. However, the JUNV minigenome assay worked robustly in Huh7 cells, resulting in the largest dynamic range (i.e. difference in reporter activity between the positive (+ L) and negative (-L) controls) from among the tested cell lines. Results for our LASV minigenome system were similar and showed high levels of reporter activity in all three cell lines. Specifically, the differences between the samples containing (+ L) and lacking polymerase (-L) ranged from 125-fold (293T) to 402-fold (Huh7). Notably, while the LASV assay showed a larger dynamic range in 293T cells than the JUNV assay, both systems showed the best performance in Huh7 cells, and consequently all subsequent assays were performed in these cells.

To further improve the minigenome assay performance by reducing the background reporter activity, we inserted decoy ORFs in a 3’-5’ orientation both downstream of the T7 terminator (site 1) and between the HDV ribozyme and the T7 terminator (site 2). These decoy ORFs increase the distance between putative cryptic promotor elements in the vector backbone and the reporter gene encoded as part of the minigenome, leading to less unspecific expression of the reporter protein. This approach has previously been successfully used to increase the dynamic range of the JUNV minigenome assay^31^, which was used as control (Fig. 1C). We could confirm these results and observed a similar impact on the LASV system, where introduction of decoy ORFs led to a very robust assay with a much higher dynamic range (i.e. 3021-fold). All further experiments were, therefore, performed using the minigenome with decoy ORFs.

For the initial experiments we used the same plasmid amounts for the expression of NP and L described for the JUNV minigenome assay. To determine the optimal amounts of these expression plasmids in the LASV minigenome assay, we titrated NP and L in individual experiments (Fig. 1D-E). Based on these results, the optimal plasmid amounts were found to be 125 ng of pCAGGS-NP and 500 ng of pCAGGS-L, which were the same amounts we used before based on our experience with the JUNV system.

Identification of conserved amino acid residues in LASV Z for mutational analysis. In order to identify conserved residues for analysis with respect to their relevance for different functions of the matrix protein we generated multiple sequence alignments of diverse mammarenavirus Z proteins, including members of all LASV lineages, as well as various other Old World arenaviruses and representatives of all New World arenavirus clades (Fig. 2). Based on this analysis, amino acids G2, G27, L71 and P72 were selected for further study, as these residues are conserved among (almost) all mammarenaviruses examined. Further, amino acids R16, P21, D22, P28, K68 and T73 were included since they were conserved among most of the Old World mammarenaviruses, and in the case of R16 and T73 also some New World arenaviruses. Amino acid positions G2 and L71 served as controls, since G2 is already well-known to be required for myristoylation of Z, resulting in membrane association, which is an essential prerequisite for budding^14,17,18^, while L71 has been suggested to mediate interaction with L and mutations at this site are lethal in virus rescues^33^. We exchanged the identified amino acids to alanine for subsequent analysis of their impact on various functions of Z in the LASV life cycle.

Fig. 2. Alignment of Z amino acid sequences from different mammarenaviruses. Sequences of representative mammarenaviruses from across different LASV lineages as well as other members of both the Old World and New World arenaviruses were aligned using Clustal Omega. Sequences are arranged by clades, with the black box containing Old World mammarenaviruses and grey boxes each representing one clade of the New World mammarenaviruses, as indicated (i.e. A, B. C or recombinant A/B). For the different LASV isolates, lineage numbers are indicated by Roman numerals. The RING domain is shown in a red box and late domains are indicated by turquoise boxes. Amino acids that were mutated in this study are indicated at the top of the alignment in orange and the sequences in which these residues are conserved are correspondingly highlighted.

Effects on RNA synthesis inhibition. To investigate the influence of the selected amino acid positions in Z on RNA synthesis inhibition we overexpressed Z in LASV minigenome assays. Since our expression constructs are c-terminally myc-tagged to allow detection via Western blot, we first titrated wildtype Z with and without myc-tag in LASV minigenome assays. The results of this comparison show that the introduction of the myc-tag does not have any influence on Z-mediated inhibition of RNA synthesis (Fig. 3A). Further, we examined the expression of the myc-tagged mutants and found that they are expressed in equal amounts when transfected into Huh7 cells, allowing us to exclude that any observed differences in RNA synthesis inhibition are due to differences in expression levels between the mutants (Fig. 3B and C, Supplemental Fig. S1). We then assessed the effects of the introduced mutations by transfecting either 3.2 ng, 12.5 ng or 100 ng of each of the respective Z-encoding plasmids in LASV minigenome assays (Fig. 3D-F). Our data show that several mutations led to increased reporter expression, indicating a decreased inhibition of RNA synthesis by the respective mutants. When compared to wildtype, mutation of L71 or P72 led to a severe loss of RNA synthesis inhibition at the highest levels of Z (i.e. 100 ng), with significant differences compared to wildtype Z already apparent even at low concentrations of Z. This was expected for residue L71, since this residue has previously been reported to be important for the interaction of Z with L, whereas a critical role for P72 in regulation of RNA synthesis had not previously been suggested^33^. In addition, several other Z mutants showed significant impairments in RNA inhibition at higher concentrations, specifically mutations at positions R16, D22, K68 and T73. This suggests that while L71 and P72 play a dominant role in the regulation of RNA synthesis inhibition, R16, D22, K68 and T73 also contribute to the inhibitory effect of Z on this process. However, one potential caveat with this experiment is the fact that the Z mutants were expressed in context of active RNPs, and it has been previously shown that this can further influence Z expression levels^33^. To control for this, we additionally expressed the Z mutants in context of minigenome assays using HA-tagged NP, and assessed the Z expression levels in HA-positive cells by FACS analysis (Fig. 3G). Interestingly, we observed for the Z mutants L71A and P72A that there was indeed a lower expression in context of active RNPs. However, the extent of this downregulation in expression (3.2-fold in case of L71A, and 3.0-fold in case of P72A) does not explain the observed differences in downregulation of reporter activity (681-fold for L71A and 384-fold for P72A in case of 100 ng transfected expression plasmid), suggesting that these mutants really are impaired in this ability (Fig. 3H).

Fig. 3. Role of conserved amino acids in LASV Z in RNA synthesis inhibition.** (A) **Titration of Z. Huh7 cells were transfected with all the components for a LASV minigenome assay in addition to increasing amounts of either pCAGGS-Z or pCAGGS-Z-myc. The dashed line shows the background of the assay as determined by a control lacking the viral polymerase (-L). Normalized reporter activities with standard deviations of three independent experiments are shown. **(B) **Expression levels of Z mutants. Huh7 cells were transfected with plasmids encoding wildtype (WT) Z-myc or Z mutants with a c-myc-tag, as indicated, and protein expression levels were determined via SDS-PAGE and Western blot. Z was detected using an anti-myc antibody and GAPDH was used as a loading control. A representative Western blot from four independent experiments is shown. **(C) **Quantification of band intensities. Band intensities of Z in Western blots from four independent experiments were quantified. Normalized mean Z band intensities with standard deviations from four independent experiments are shown. Asterisks indicate p values from a one-way analysis of variance (ns p > 0.05; ****p ≤ 0.0001) compared to WT samples. **(D-F) **Influence of Z-mutants on RNA synthesis inhibition. Huh7 cells were transfected with all the components for a LASV minigenome assay as well as (C) 3.2 ng, (D) 12.5 ng or (E) 100 ng of Z WT or Z mutants, as indicated. As controls, cells were transfected with empty vector instead of Z either in the presence or absence of the viral polymerase (i.e. -L/+L). Means of normalized reporter activities with standard deviations from three independent experiments per transfection condition are shown. Statistical significance was calculated compared to Z-WT using one-way analysis of variance with asterisks indicating the obtained p values (ns p > 0.05; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001: ****p ≤ 0.0001). **(G) **Influence of vRNPs on Z expression. Cells were transfected with myc-tagged Z-WT or Z mutants and all the components for a minigenome assay, but including HA-tagged NP. 48 h p.t. cells were harvested and subjected to FACS analysis. Cells were gated for HA-positive cells, and the mean fluorescence intensity of the anti-myc signal was determined. Means and standard deviations of two independent experiments are shown. **(H) **Influence of selected Z-mutants on RNA synthesis inhibition as a function of the amount of transfected plasmid. Data from panels D-F are shown with the reporter activity plotted on the Y-axis, and the amounts of transfected plasmids on the X-axis, together with nonlinear regression curves for these data.

Establishment of a trVLP assay. In order to study budding-related functions of Z, we further established a LASV trVLP assay, which can model the whole viral life cycle including particle morphogenesis and budding. To this end, we cloned an S segment-based minigenome that encodes a GPC-T2A-nluc in the position normally occupied by the NP ORF, and Z in the position normally occupied by the GPC ORF (Fig. 4A). Using this approach, the Z ORF remains in the same orientation as in the viral genome to maintain temporally regulated expression of Z. After transfecting producer (p0) cells with this multicistronic minigenome together with the viral RNP proteins, the trVLPs produced by these cells were used to infect target (p1) cells that had been pre-transfected with the plasmids for the RNP proteins. We first tested the assay in different cells lines (i.e. Huh7 and 293T) in order to identify the most suitable cell line for this assay. Performing the assay in 293T cells resulted in minimal reporter activity over background in p0 cells (Fig. 4B). Consistent with this, we detected only very limited reporter activity in p1 cells, suggesting there is no significant trVLP production occurring (Fig. 4B). In contrast, performing the assay in Huh7 cells resulted in reporter activity that was 16-fold over background in p0 and also resulted in reporter activity in p1 cells 1480-fold over background, indicating robust generation of LASV trVLPs in this cell type.

Fig. 4. Establishment of the LASV transcription and replication-competent virus-like particle (trVLP) assay.** (A) **Schematic of the LASV trVLP system. The multicistronic minigenome is derived from the LASV S segment, in which the NP and GPC ORFs were replaced with a GPC-T2A-nanoluciferase and Z ORF, respectively. Transfection of producer (p0) cells with the plasmid containing the multicistronic minigenome as well as the T7 polymerase leads to initial transcription of the minigenome. Additional transfection of plasmids encoding the LASV nucleoprotein (NP) and polymerase (L) enables replication and secondary transcription of the minigenome, resulting in expression of Z and GPC as well as the nanoluciferase reporter. A plasmid encoding Firefly luciferase (FF) is used as normalization control to assess transfection efficiency, plasmid-driven gene expression and cell viability. The p0 cells produce transcription and replication competent virus like particles (trVLPs) that contain the viral minigenome RNA and can be used to infect target (p1) cells that have been pre-transfected with NP, L and FF. This facilitates further replication and transcription of the minigenome in these cells. The normalized reporter activity in p1 cells reflects assembly and budding of trVLPs in the p0 cells, entry of the trVLPs in p1 cells, as well as replication, secondary transcription and protein expression in both p0 and p1 cells. **(B) **Testing of the LASV trVLP assay in different cell lines. 293T and Huh7 cells were transfected with all components for a LASV trVLP assay. As a control, the viral polymerase was omitted in p0 (-L). The differences between negative (-L) and positive (+ L) controls for each cell line and in each passage (i.e. p0 or p1) are shown as log RLUs above each pairing. The normalized means and standard deviations from five independent experiments are shown.

Effects on budding and incorporation of RNPs into viral particles. Having successfully established the LASV trVLP system, we next used this assay to investigate the impact of the conserved amino acids in Z on viral budding and RNP incorporation into trVLPs by introducing each of the mutations into the multicistronic minigenome. In p0 cells we found that none of the mutations affected reporter activity (Fig. 5A), indicating that the selected amino acids influence neither replication nor secondary transcription of the multicistronic minigenome. However, in p1 cells we observed significantly reduced reporter activity for several mutants when compared to the wildtype (Fig. 5B). Specifically, the mutations G2A, R16A, D22A, K68A, L71A, P72A and T73A all showed varying degrees of reduction in the reporter activities seen in p1. The strongest effects were observed for G2A, D22A and L71A, which in the case of G2A and L71A is consistent with previous reports describing them to be essential for Z myristoylation, as a prerequisite for membrane association and ultimately budding, and for virus rescue, respectively, although in case of L71A we cannot completely rule out that differences in expression level in context of active vRNPs might also play a role^14,18,33^. Nevertheless, since none of the mutations had an influence on reporter activity in p0 cells, these data suggest that these residues all might contribute to robust budding and/or RNP incorporation into viral particles.

Fig. 5. Role of conserved amino acids in LASV Z in budding and RNP incorporation. (A) Influence of Z-mutants on trVLP assay reporter activity in producer (p0) cells. Huh7 cells were transfected with all the components for a trVLP assay including multicistronic minigenomes encoding the different LASV Z-mutants, as indicated. As negative controls either a minigenome lacking the Z ORF (ΔZ) was used, or the minigenome was omitted entirely (-mg). Nanoluciferase reporter activity was normalized to FF activity. **(B) **Influence of Z-mutants on trVLP assay reporter activity in target (p1) cells. Huh7 cells (pre-transfected with LASV NP and L) were infected with trVLPs produced by the corresponding p0 cells shown in (A). **(C) **Impact of conserved amino acids in Z on VLP production. Myc-tagged LASV Z mutants were overexpressed in Huh7 cells. The released VLPs were purified via ultracentrifugation over a sucrose cushion and subsequently treated with proteinase K. As a negative control, VLPs were left untreated, and as positive control VLPs were additionally treated with Triton X100. Samples were analyzed via SDS-PAGE and Western blot and Z was detected using an anti-myc antibody. A representative blot from four independent experiments is shown. **(D) **Quantification of band intensities. The band intensities of samples treated with Proteinase K in (C) were determined and normalized to a lysate expression control (not shown). Normalized means with standard deviations from four independent experiments are shown (A, B, D). Asterisks indicate p values from a one-way analysis of variance (ns p > 0.05; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001) compared to WT samples.

To further test this and more precisely define the functions of these amino acids, we next investigated whether they directly impair viral morphogensis and budding. Since expression of LASV Z alone can lead to the formation of virus-like particles (VLPs) in cell culture^7,13^, we overexpressed myc-tagged Z mutants in Huh7 cells and quantified the amount of VLPs released into the supernatant (Fig. 5C-D; Supplemental Fig. S2). For this purpose, VLPs were purified via ultracentrifugation over a sucrose cushion before being subjected to Western blot analysis. Further, to ensure that we only detect Z that is released inside of VLPs, and not protein released from the cells through other mechanisms, we additionally performed proteinase K protection assays. Consistent with what has been previously reported^14,18^, the G2A mutant showed no release of Z in the form of VLPs (i.e. in the proteinase K treated sample), thereby confirming that Z-G2A (due to a defect in Z myristoylation, which is a prerequisite for budding) is budding-deficient. In contrast, while some of the other mutants showed a slight reduction in Z release in VLPs compared to wildtype Z, these differences were limited in magnitude and failed to reach statistical significance, suggesting that none of these other amino acids strongly impacts the budding of viral particles.

Effects on interaction of LASV Z with NP. Since these conserved amino acids in Z did not impair viral budding, we next assessed whether they are important for NP-Z interaction, which has been suggested to recruit the RNP complexes into budding virions^34^. To analyze this, we overexpressed myc-tagged Z and HA-tagged NP in 293T cells and performed co-immunoprecipitation (CoIP) assays by precipitating Z and analyzing the amount of co-precipitated NP. While the amount of precipitated Z was comparable for all mutants, we observed marked differences in the amount of co-precipitated NP for several of the mutants (Fig. 6; Supplemental Fig. 3). Specifically, mutations G2A, L71A, P72A and T73A strongly impaired the interaction of Z with NP, suggesting that for these mutants the lack of activity observed in the trVLP assay p1 cells is due to a lack of NP-Z interaction, leading to the generation of particles that lack RNPs and, therefore, cannot productively infect target cells.

Fig. 6. Role of conserved amino acids in Z in NP-Z interaction.** (A) **CoIP analysis of NP-Z interaction. Myc-tagged LASV Z mutants were overexpressed together with HA-tagged LASV NP in 293T cells. The cells were lysed and the proteins subjected to co-immunoprecipitation (CoIP) using anti-myc antibodies. Input and CoIP samples were analyzed via SDS-PAGE and Western blot. Z was detected using an anti-myc antibody and NP was detected using an anti-HA antibody. **(B, C) **Quantification of band intensities. Band intensities in input (B) and CoIP (C) samples were quantified. CoIP band intensities for NP were normalized to the respective Z band intensities. Individual and mean NP band intensities from four independent experiments are shown. Asterisks indicate p values from a one-way analysis of variance (ns p > 0.05; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001) compared to WT samples.

Discussion

In this study we describe the establishment and optimization of a monocistronic LASV minigenome system based on the Josiah reference strain, as well as the establishment of a LASV trVLP assay that can be used to investigate not only LASV RNA synthesis and protein expression, but the whole viral life cycle outside of maximum containment facilities. While minigenome systems for LASV have been previously established^15,35^, we demonstrated that their performance can be significantly improved by the selection of an appropriate cell line, as well as by implementing strategies to reduce background reporter expression from cryptic promoters. Further, while some systems modelling the whole LASV life cycle outside of maximum containment have been previously reported, these were either heterologous systems (i.e. using a combination of LCMV and LASV components), or used stable cell lines to supply the RNPs^7,24,36^. By applying the same principle that was used to generate a bicistronic minigenome-based trVLP assay for New World arenaviruses^27,28^ we have now generated a LASV trVLP system. This approach minimizes the artificial steps involved to the encapsidation of “naked” minigenome RNA (“illegitimate encapsidation”) in p0 and the expression of RNP proteins from plasmids, which are common limitations to most of these lifecycle modelling systems. Importantly, all viral components in this system are derived from LASV, which alone ensures a more authentic approach to modelling of the LASV life cycle. This is further improved upon by the regulated expression of Z and GPC directly from the minigenome, which also recapitulates the biphasic protein expression strategy natural to arenaviruses (at least for the gene in the positive-sense locus, which in this case was Z). Compared to other trVLP approaches that rely on a second later transfection of Z and GPC from plasmids (i.e. after the plasmids encoding the minigenome and RNP proteins are expressed) to counteract the strong inhibitory effect of Z on viral RNA synthesis, this approach is more robust and also needs fewer plasmids, which makes it easier to handle. Further, since all viral proteins are expressed from either the minigenome or expression plasmids, the system still offers the flexibility to modify individual proteins, for instance to analyze mutants, as we have done here. Ultimately, this means that investigating the LASV life cycle outside of maximum containment laboratories with the LASV trVLP assay represents a more authentic approach and offers greater experimental flexibility compared to the previously available systems.

To demonstrate the power of these systems for investigating virus biology, and particularly to analyze not only viral RNA synthesis (using the minigenome assay), but also other steps like budding and nucleocapsid incorporation (using the trVLP assay), we applied these tools to systematically study the roles of highly conserved amino acids in the mammarenavirus matrix protein Z (results are summarized in Table 1). We found that residues L71 and P72 are essential for the inhibition of viral RNA synthesis by Z, with residues R16, D22, K68 and T73 also contributing to this function. This is in agreement with previous data obtained using a heterologous minigenome assay to analyze different LASV Z mutants in the context of LCMV RNP proteins^24^. While the authors demonstrated similar results for most residues, they also reported an effect of the G27A and K68A mutants, which we do not observe in the homologous LASV system. This might be explained by a limited inter-species compatibility between the L and Z proteins of different mammarenaviruses, which makes further contact sites necessary to support the interaction between these proteins. Another previous study using a LASV-based minigenome showed a loss of inhibition for Z with mutations in residues L71 and P72, which is similar to our results, while additionally implicating T73 as being equally important^37^. However, this study did not include any of the other residues our assays now show to be important contributors to this function. These findings also support what is known from structural data for other mammarenaviruses such as TCRV and LCMV, where it has been shown that Z interacts with L, and that the inhibitory effect of Z on RNA synthesis is dependent on this interaction^16,23^. For both viruses the binding site has been mapped to the RING domain, although the TCRV data additionally strongly suggest an involvement of the C-terminus of Z in this interaction^16,22,23^. This is then also broadly consistent with our data suggesting the involvement of not only L71 and P72, but also other residues from the C-terminus of Z (i.e. R68 and T73), although our data also implicate residues R16 and D22 from the Z protein N-terminus. More recently, cryo-electron microscopy structures of the LASV Z:L complex have also mapped the proteins’ interaction site to the RING domain of Z^38^. However, these structural data implied that the L71/P72 region (as part of the C-terminus) is not directly involved in the Z-L interaction^38^. One possible explanation for these apparently disparate results is that Z may be present in different conformations depending on its binding partner(s), something which is supported by structural folding predictions for this protein^39^. Specifically, these predictions show that the C-terminal arm of the protein, in which the L71/P72 residues are located, remains rather flexible and mediates interactions with multiple binding partners through structural changes, possibly including the RING domain. Some of the modelled structures further suggest that hydrophobic intra-protein interactions in Z, including at position L71, could be required to stabilize the protein structure. Therefore, we propose that mutations L71A and P72A impose structural constraints that limit the efficient interaction between Z and L, and thereby prevent inhibition of viral RNA synthesis. This is supported by the previous finding that mutating L72 in LCMV, which is analogous to L71 in LASV, leads to a loss of this interaction and prohibits virus rescue^33^. These constraints could potentially be linked to the oligomerization behavior of LASV Z, as P72 has previously been shown to be important for modulating the oligomerization activity of Z, with a P72A mutant of Z showing a much higher tendency to oligomerize than WT Z^40^. At the same time, recent data has shown that, at least for LCMV, oligomerization of Z is not required for its inhibitory activity on viral RNA synthesis^41^, giving rise to the possibility that it is actually the oligomerization of Z that might prohibit Z from interacting with L - an idea that is further supported by the fact that many of the amino acids involved in L-Z interaction are not accessible for interaction in the oligomeric form of Z^38,40^.

Table 1. Summary of the observed phenotypes for each LASV Z mutant. For each mutant it is indicated whether the mutant shows no significant (+/-), a statistically significant but modest (-) or a statistically significant and strong (--, i.e. > 10-fold) reduction in activity compared to wildtype.G2R16P21D22G27P28K68L71P72T73RNA synthesis inhibition3.2 ng+/-+/-+/-+/-+/-+/-+/---+/-12.5 ng+/----+/-------+/-100 ng+/--+/----+/--------trVLP assay: p1 activity---+/---+/-+/--------Budding activity--+/-+/-+/-+/-+/-+/-+/-+/-+/-Z-NP interaction--+/-+/-+/-+/-+/-+/----

In addition to studying the role of Z in regulating viral RNA synthesis, our LASV trVLP system also allowed us to study its contributions to additional functions like budding and RNP incorporation. These assays showed decreased reporter activity in target cells of the trVLP assays upon mutation of residues G2, R16, D22, K68, L71, P72 or T73. For G2A this can be explained by a loss of myristoylation leading to decreased budding activity, which is a well-known phenotype of this mutant^14,18^. Further, for the residue L79 in JUNV Z, which is homologous to LASV Z L71, previous data have also indicated its importance for Z-NP interaction^21,24^. Consistent with these reports, mutation of L71 in LASV Z also resulted in a profound reduction in trVLP infectivity. In contrast, mutations in the neighboring P72 and T73 appeared to also contribute to this, but to a much lesser extent. Interestingly, other than G2, no other residues showed significantly decreased budding activity upon mutation, indicating impairment of another step related to trVLP formation. For the mutants L71A, P72A and T73A we could show that this was related to an impaired ability to interact with NP, which is critical for the recruitment of nucleocapsids to budding sites and their subsequent incorporation into budding particles. Surprisingly, we found that interaction with NP was also impaired in the G2A mutant, indicating that the decreased reporter activity seen with G2A in trVLP assays is not only due to its impaired budding activity, but that even where budding is still possible, the incorporation of RNPs (and thus genomes) into the budding particles is impaired. These observations exhibit significant parallels to what has been reported for other mammarenaviruses. For instance, Pichinde virus (PICV) cannot be rescued with mutations in any of the analogous residues to G2, L71 or P72, or with simultaneous mutations of L71-K74, further emphasizing the importance of these residues in the mammarenavirus life cycle^37^. The same study additionally described reduced incorporation of NP into VLPs for mutations in the L71-T73 region of LASV Z, which our data suggest is most likely due to decreased Z-NP interaction^37^. Overall, these findings support that the L71-T73 region of LASV Z forms a critical part of the Z-NP interaction site, and as such contributes to nucleocapsid recruitment into budding particles.

A recent study using Z and NP of a different LASV lineage showed that residues R16 and R74 are also important for Z-NP interaction^34^. While we did not test R74 (since it is not highly conserved among all mammarenaviruses), our interaction data suggest that the residues neighboring R74 (L71-T73) are indeed part of the interaction site with NP, and thus additional participation of R74 in this interaction seems plausible. However, when we examined the effect of mutation at R16, we did not observe any impairment of Z-NP interaction, although this mutant was still severely impaired in the trVLP assay. Taken together, these data suggest that R16 is indeed crucial for the incorporation of RNPs into viral particles, but in a way that is not modelled in our NP-Z CoIP-based interaction assay.

For two residues, D22 and K68, neither our experiments nor the currently published data can provide an explanation of how their mutation impairs the viral life cycle. Our data clearly show that reporter activity in p1 cells is affected by mutation of these residues, however, neither budding nor interaction of Z with NP are impacted. Additionally, it has previously been shown that the Z-GP interaction is not affected by either of these mutations^24^. However, previous data from JUNV indicate that interactions between GP and the Z-NP complex could be important for the efficient incorporation of GP into infectious particles^21^. Therefore, we speculate that while direct interaction of Z with other viral proteins is not affected, the interplay between Z, NP and GP could be impaired in a way that leads to deficiencies in particle morphogenesis that impact particle infectivity. Further investigations to describe and understand all relevant interaction sites between these viral factors, as well as additional structural data, could help to resolve this in the future. Alternatively, it remains possible that uncoating of the virus in infected cells could be affected by mutation of these amino acids, which would be reflected in the trVLP assay, but not mechanistically clarified by our follow-up experiments. The mechanisms involved in arenavirus uncoating have so far not been described in detail and, therefore, it is currently not possible to evaluate the impact of the identified mutations on specific functions related to this process. Lastly, it is possible that interactions with cellular components are altered by the introduced mutations, and indeed, Z is known to interact with several host cell factors^7,42–45^. However, the binding sites for these interactions have so far all been mapped to either the RING or the late domains, and their functions are thought to involve the regulation of budding and translation, which seem not to be influenced by D22 or K68. Nevertheless, we cannot exclude that D22 and K68 might still be involved in so far undescribed interactions of Z with host cell proteins.

In addition to these open questions, which will have to be addressed in future studies, there are some methodological limitations to this study. The minigenome used for expression of Z is based on the S-segment of LASV, whereas during virus infection Z is expressed from the L-segment. We cannot exclude that this might result in differences in expression due to different promoter strengths in those two genome segments, although this would affect the expression of Z and the Z-mutants to a similar extent. Further, the interaction studies are currently based solely on CoIP data, and additional orthogonal approaches (e.g. a BRET assay) could increase confidence in these findings. Also, it will be interesting in the future to attempt probing Z-NP interactions by CoIP in the context of active vRNPs, i.e. when Z is expressed from a minigenome, as recent work has shown that the presence of active vRNPs can influence the expression levels/stability of specific Z mutants^33^. However, this will be technically challenging since regulated Z expression from minigenomes results in much lower levels of expression than transient transfection approaches.

In conclusion, we have optimized the classical LASV monocistronic minigenome system to achieve a high dynamic range and thereby allow highly sensitive investigation of factors impacting viral RNA synthesis. At the same time, we report the development of a bicistronic trVLP assay system that models the entire viral life cycle without the need for a maximum containment facility using only authentic LASV components, while at the same time retaining experimental flexibility to modify viral proteins components for analysis. Using these systems we have demonstrated that the majority of the amino acid residues in Z that are highly conserved across different mammarenavirus species are of critical importance for the LASV life cycle. We confirmed that residue G2 is essential for budding and further reveal its importance for Z-NP interaction. Further, we showed that residues L71/P72/T73 are also involved in NP binding, as well as demonstrating the functional relevance of R16, which had previously been shown to interact with NP^34^. Several residues were also found to be involved in inhibiting viral RNA synthesis, with L71 and P72 being essential for this function of Z. Furthermore, while residues D22 and K68 are clearly important for the functions of Z in the viral life cycle, the underlying mechanism remains elusive. Resolving the roles of these (and other) individual residues in specific virus functions at a mechanistic level will lead to a more detailed understanding of how virus-virus and virus-host interactions affect the viral life cycle, and in doing so offer important future potential for the targeting of these interaction interfaces as a part of antiviral therapeutic strategies.

Material & methods

Cells

Human hepatocarcinoma (Huh7; kindly provided by Stephan Becker, Philipps University, Marburg, Germany) and human embryonic kidney 293T cells (Collection of Cell Lines in Veterinary Medicine CCLV-RIE 1018) were cultivated and maintained in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific) supplemented with 100 U/mL penicillin and 100 µg/mL streptomycin (Thermo Fisher Scientific), as well as 1x GlutaMAX (Thermo Fisher Scientific) and 10% fetal bovine serum (FBS; Thermo Fisher Scientific). BsrT7/5 cells (Syrian golden hamster kidney cells, CCLV-RIE 0583, kindly provided by Stefan Finke, Friedrich-Loeffler-Institut)^32^ were cultivated in MEM Glasgow with 100 U/mL penicillin and 100 µg/mL streptomycin (Thermo Fisher Scientific), as well as 1x GlutaMAX (Thermo Fisher Scientific) and 10% newborn calf serum. BsrT7/5 cells were treated with 1 mg/mL Geneticin in every other passage. All cell cultures were grown at 37 °C and 5% CO_2_.

Plasmids

Expression plasmids for T7-polymerase, JUNV-NP, JUNV-L as well as the T7-driven JUNV minigenome plasmid have been previously described^31^. The LASV (strain Josiah) NP, Z and L ORFs were cloned from viral RNA in order to generate pCAGGS-LASV-NP, -Z and -L, respectively. For generation of the minigenome plasmid, an empty LASV S segment lacking both ORFs was synthesized by Thermo Fisher Scientific (GeneArt Gene Synthesis) and inserted into the previously published pAmp plasmid^31^. Subsequently, the nanoluciferase ORF was inserted into the GPC locus of this construct to generate pAmp-LASV-S-ΔNP_ΔGP-nluc. For the generation of the multicistronic minigenome, LASV Z was cloned into the GPC locus, while GPC was fused to nluc via a T2A-linker and inserted into the NP locus to generate pAmp-LASV-S_ΔNP-GP-T2A-nluc_ΔGP-Z. The expression plasmids for the different c-myc tagged Z mutants as well as the bicistronic minigenomes harboring Z mutations were generated by site-directed mutagenesis.

Antibodies

For CoIP experiments, an anti-myc antibody (mouse anti-c-myc, clone 9E10; Thermo Fisher Scientific, MA1-980-1MG) was used. The proteins were detected in Western blot analysis using the primary antibodies anti-myc (rabbit; Thermo Fisher Scientific, PA1-981) and anti-HA (chicken; Abcam, ab9111) with secondary antibodies against rabbit (IRDye^®^ 680RD, goat anti-rabbit; Licor, 926-68071) and chicken (IRDye^®^ 800CW, donkey anti-chicken; Licor, 926-32218). For Western blot analysis of the proteinase K protection assay, anti-myc (rabbit; Thermo Fisher Scientific, PA1-981) and anti-GAPDH (mouse; Santa Cruz, sc-47724) were used as primary antibodies with secondary antibodies against rabbit (Alexa Fluor 790, goat anti-rabbit; Dianova, 111-655-144) and mouse (IRDye^®^ 680RD, goat anti-mouse; Licor, 926-68070). For FACS analysis, a FITC-conjugated anti-myc antibody (clone SH1-26E7, Miltenyl Biotec) and an APC-conjugated anti-HA antibody (clone GG8-1F3.3.1, Miltenyl Biotec) were used.

LASV and JUNV minigenome assay

For minigenome assays, Huh7, BsrT7/5 or 293T cells were seeded in 12-well plates and transfected using Transit LT-1 (Mirus Bio) on the next day with 125 ng pCAGGS-T7, 125 ng pCAGGS-NP, 500 ng pCAGGS-L and 125 ng pAmp-S-ΔNP_ΔGP-nluc or pAmp-S-ΔNP_ΔGP-nluc_1L2L as well as 12.5 ng pCAGGS-luc2 (encoding firefly luciferase for normalization), as previously described for the JUNV minigenome system^31^. At 24 h post transfection, medium was replaced with DMEM containing 5% FBS and another 24 h later cells were harvested. To this end, cell culture supernatant was removed and cells were lysed in 200 µL 1x Lysis Juice (PJK) for 10 min. After clearing the lysate via centrifugation at 10,000 x g for 3 min, reporter activity was determined by adding 40 µL lysate to 40 µL Nano Glo luciferase assay reagent (Promega) or 40 µL Beetle Juice (PJK) in black opaque 96-well plates. Samples were then measured in a GloMax Discover (Promega) multiplate reader. Nanoluciferase activity was normalized to firefly luciferase activity.

Clustal Omega alignment

A multiple sequence alignment of the amino acid sequences of different mammarenavirus matrix proteins was performed in Geneious Prime v2021.0.1 (Biomatters) using the Clustal Omega 1.2.2 algorithm^46^. The following amino acid sequences were used in the analysis (NCBI reference sequence or GenBank accession numbers): LASV Pinneo (AIT17834.1), LASV CSF (AZI96290.1), LASV NL (AAO59510.1), LASV Josiah (AEY85214.1), LASV AV (CCA30312.1), LASV Togo (AVN98159.1), LNKV (Lunk virus, YP_006858708.1), Ngerengere virus (WDW20697.1), Ryukyu virus (YP_009508473.1), DANV (Dandenong virus, ABY20731.1), LCMV Armstrong (Lymphocytic Choriomeningitis virus, KY514257.1), LCMV WE (AB627954.1), Alxa virus (ATY47649.1), Souris virus (AJI43720.1), BITV (Bitu virus, YP_010840417.1), OKAV (Okahandja virus, YP_009141009.1), MWV (Merino Walk virus, ADD63339.1), LORV (Loei River virus, NC_038365.1), WENV (Wenzhou virus, AJA91492.1), Kitale virus (QHB13140.1), MRTV (Mariental virus, AKH39837.1), MORV (Morogoro virus, YP_003090216.1), MOPV AN20410 (Mopeia virus, AAV54106.1), MOPV Mozambique (ABC71136.1), KWAV (Kwanza virus, YP_010840423.1), LUNV (Luna virus, BAL03413.1), MOBV (Mobala virus, NC_007904.1), DHWV (Dhati Welel virus, QLJ57223.1), Gairo virus (AIK25570.1), Lijiang virus (AWM11452.1), Ippy virus (ABC71142.1), Solwezi virus (YP_009505807.1), LUJV (Lujo virus, YP_002929492.1), TAMV (Tamiami virus, YP_001911117.1), BCNV (Bear Canyon virus, YP_001649224.1), WWAV (Whitewater Arroyo virus, YP_001911119.1), PARV (Parana virus, YP_001936027.1), PICV (Pichinde virus, AF427517.1), PIRV (Pirital virus, YP_025092.1), ALLV (Allpahuayo virus, YP_001649213.1), FLEV (Flexal virus, YP_001936023.1), GTOV (Guanarito virus, NP_899220.1), TCRV (Tacaribe virus, NP_694847.1), JUNV Candid#1 (Junin virus, AY819707.2), JUNV Romero (AY619640.1), MACV (Machupo virus, NP_899214.1), AMAV (Amapari virus, YP_001649217.1), CPXV (Cupixi virus, YP_001649219.1), SABV (Sabia virus, ABY59837.1), APOV (Apare virus, AUD40060.1), Chapare virus (YP_001816784.1), LATV (Latino virus, YP_001936025.1), OLVV (Oliveros virus, YP_001649215.1).

FACS analysis

For FACS analysis, minigenome assays were performed as described above, but expressing HA-tagged NP instead of WT-NP, and additionally transfecting 250 ng of myc-tagged Z or Z mutants. 48 h p.t. cells were harvested in 1 ml PBS and spun down for 5 min at 350 x g and 4 °C prior to staining using the Foxp3/Transcription Factor Staining Buffer Set (Invitrogen) together with an FITC-conjugated anti-myc antibody (clone SH1-26E7, Miltenyl Biotec) at a dilution of 1:100 and an APC-conjugated anti-HA antibody (clone GG8-1F3.3.1, Miltenyl Biotec) at a dilution of 1:50. Samples were analyzed using a BD Biosciences FACSymphony A3, and gated for APC (HA) positive cells, after which the mean fluorescence intensity in the FITC (myc) channel was determined.

LASV trVLP assay

For p0 cells, 293T or Huh7 cells were seeded in 12-well plates and transfected using Transit LT-1 (Mirus Bio) on the next day with 125 ng pCAGGS-T7, 125 ng pCAGGS-NP, 500 ng pCAGGS-L and 125 ng pAmp-LASV-S_ΔNP-GP-T2A-nluc_ΔGP-Z as well as 12.5 ng pCAGGS-luc2 (for normalization) according to the manufacturer’s protocol and as has been previously described for the TCRV trVLP assay^27^. Different versions of the multicistronic minigenome were used encoding each of the Z-mutants. The medium was replaced with 2 mL DMEM containing 5% FBS 24 h post transfection and another 48 h later both the supernatant, containing newly generated trVLPs, and the cell lysates were harvested. To this end, 1.8 mL of supernatant was cleared of cell debris by centrifugation at room temperature for 5 min at 800 x g before the cleared supernatant was used to infect the target cells (p1). For measuring reporter activity, the p0 cells were lysed in 200 µL 1x Lysis Juice (PJK) for 10 min before the lysate was cleared by centrifugation and luciferase activity determined using the same protocol described above for the minigenome assay. The p1 cells were reverse transfected and seeded into 12-well plates 24 h prior to trVLP infection with 125 ng pCAGGS-NP, 500 ng pCAGGS-L and 12.5 ng pCAGGS-luc2. For the infection, cell culture supernatant was removed and cells were infected for 1 h at 37 °C and 5% CO2 with 1.5 mL of the cleared p0 supernatant. Afterwards, the inoculum was removed and 2 mL DMEM containing 5% FBS were added to the p1 cells. At 72 h post infection the p1 cells were harvested and analyzed for luciferase activity as described for the p0 cells. Nanoluciferase activity was normalized to firefly luciferase activity.

Proteinase K protection assay

Huh7 cells were seeded in 12-well plates and transfected 24 h later using Transit LT-1 with pCAGGS based expression plasmids encoding c-myc-tagged LASV Z mutants according to the manufacturer’s protocol (Mirus Bio). For each Z mutant two wells were transfected. At 24 h post transfection the supernatant was exchanged and 48 h post transfection the samples were harvested. To this end, the supernatants of the duplicate wells were pooled and cleared of cell debris twice by centrifugation for 5 min at 800 x g, after which cleared supernatants were transferred into new sample tubes before subjecting them to further ultracentrifugation. For the cell lysates, cells were washed in phosphate-buffered saline (PBS) and pelleted by centrifugation. Pellets were resuspended in 90 µL PBS and 30 µL 4x SDS sample buffer before boiling the samples at 95 °C for 5 min.

For the proteinase K protection assay, VLPs were purified via ultracentrifugation over a sucrose cushion. For this, the supernatant was underlaid with 20% sucrose in PBS and centrifuged for 2 h at 129,000 x g and 4 °C. The supernatant was discarded and pellets containing the VLPs were resuspended in 130 µL PBS. Next, 40 µL of the resuspended VLPs were incubated for 1 h at 37 °C with 12 µL of either PBS or 7.2 µL PBS and 4.8 µL proteinase K (150 µg/mL, Thermo Fisher Scientific) or 7.2 µL PBS containing 0.1% Triton X100 and 4.8 µL proteinase K. The samples were boiled for 5 min at 99 °C before adding 20 µL 4x SDS sample buffer and then boiled for another 15 min. The lysates, as well as the proteinase K-treated supernatant samples were analyzed via SDS-PAGE followed by Western blot. The blots were analyzed using an Odyssey CLx (Li-Cor) imaging system and mean band intensities were quantified in Image Studio Light (Li-Cor). The band intensities of Z determined in the treated VLPs samples were normalized to the corresponding Z band intensities in the lysate expression control blot. Additionally, the results were normalized to the mean band intensities across all replicates.

Co-immunoprecipitation

293T cells were seeded in 6-well plates and transfected one day later using Transit LT-1 (Mirus Bio) with pCAGGS-based expression plasmids encoding HA-tagged LASV NP and c-myc-tagged LASV Z mutants according to the manufacturer’s protocol. 24 h post transfection the supernatant was exchanged and 48 h post transfection CoIP was performed as described previously^47^. Briefly, the cells were washed and then lysed in 1% NP-40 buffer containing cOmplete protease inhibitor (Roche). The cells were lysed using a 22G needle (Becton Dickinson) and then rotated for 2 h at 4 °C. Afterwards, the lysates were cleared by centrifugation and 150 µL (corresponding to 20% of the amount used for immunoprecipitation) of the cleared lysate was subjected to acetone precipitation in order to generate the input control. The remaining 750 µL of the cleared lysate was subjected to CoIP by incubation for 10 min with anti-c-myc coupled magnetic beads (Protein G Dynabeads, Thermo Fisher Scientific). The bead-antibody-antigen-complexes were then washed three times in PBS containing 0.02% Tween-20. Finally, the beads were transferred to new sample tubes and resuspended in 1x SDS sample buffer and boiled for 10 min at 99 °C. The input and CoIP samples were analyzed using SDS-PAGE followed by Western blot with band intensities determined as described above. For the input control, the raw NP band intensities were quantified, while for the CoIP the NP band intensities were normalized to Z. Further, the results were normalized to the mean band intensities across all replicates to account for experimental variability.

Statistics

Statistical analysis was performed using one-way analysis of variance (ANOVA) with Dunnett’s multiple comparisons test in GraphPad Prism 9.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1