S U P P L E M E N T A R T I C L E
Recombinant Marburg Virus Expressing EGFPAllows Rapid Screening of Virus Growth andReal-time Visualization of Virus Spread
Kristina Maria Schmidt,1,2,3,a Michael Schumann,3,a,b Judith Olejnik,1,2,3 Verena Krahling,3 and Elke Muhlberger1,2,3
1Department of Microbiology, Boston University School of Medicine, and 2National Emerging Infectious Diseases Laboratories Institute, Boston,Massachusetts; and 3Department of Virology, Philipps University of Marburg, Germany
The generation of recombinant enhanced green fluorescent protein (EGFP)–expressing viruses has
significantly improved the study of their life cycle and opened up the possibility for the rapid screening of
antiviral drugs. Here we report rescue of a recombinant Marburg virus (MARV) expressing EGFP from an
additional transcription unit (ATU). The ATU was inserted between the second and third genes, encoding
VP35 and VP40, respectively. Live-cell imaging was used to follow virus spread in real time. EGFP expression
was detected at 32 hours postinfection (hpi), and infection of neighboring cells was monitored at 55 hpi.
Compared to the parental virus, production of progeny rMARV-EGFP was reduced 4-fold and lower protein
levels of VP40, but not nucleoprotein, were observed, indicating a decrease in downstream protein expression
due to the insertion of an ATU. Interestingly, EGFP concentrated in viral inclusions in infected cells. This was
reproduced by transient expression of both EGFP and other fluorescent proteins along with filovirus
nucleocapsid proteins, and may suggest that a general increase in protein synthesis occurs at viral inclusion
sites. In conclusion, the EGFP-expressing MARV will be a useful tool not only to monitor virus spread and
screen for antiviral compounds, but also to investigate the biology of inclusion body formation.
Marburg virus (MARV) and the closely related Ebola
virus (EBOV) belong to the filovirus family and cause
a severe hemorrhagic fever in humans, with mortality
rates up to 90%. Currently, there is no approved vaccine
or antiviral treatment.
Filoviruses have a nonsegmented negative-sense RNA
genome encoding 7 structural proteins. Four of these
proteins constitute the nucleocapsid complex, containing
the nucleoprotein (NP), the viral polymerase L, the
polymerase cofactor VP35, and the viral protein VP30
in close association with the viral genome (for review,
see [1]). Cytoplasmic inclusions, which are thought to
represent active sites of viral replication, are present as
large aggregates in filovirus-infected cells. These in-
clusions are formed by all 4 nucleocapsid proteins, with
NP being the driving force for aggregation due to self-
assembly of NP [2, 3]. NP interacts with VP35, VP30,
and L, either directly or via a linker protein, thereby
redirecting the nucleocapsid proteins into cytoplasmic
aggregates [4–8].
Rescue systems to recover infectious virus from full-
length complementary DNA (cDNA) clones have been
established for both MARV and EBOV [9–13]. These
techniques were used to generate recombinant forms of
EBOV, derived from isolates of the Zaire ebolavirus
(ZEBOV) species, containing the enhanced green
fluorescent protein (EGFP) gene within an additional
transcription unit (ATU), which provide a sensitive
and quantitative readout for antiviral drug screening
assays and virus spread studies [14–21]. EGFP was
Potential conflicts of interest: none reported.Presented in part: XIV International Conference on Negative Strand Viruses,
Brugge, Belgium, 20–25 June 2010; and New England Regional Center ofExcellence (NERCE) Sixth Annual Retreat, Newport, RI, 14–15 November 2010.
aK. M. S. and M. S. contributed equally to this work.bPresent address: CSL Behring GmbH, Emil-von-Behring-Straße 76, 35041
Marburg, Germany.Correspondence: Elke Muhlberger, PhD, Department of Microbiology, Boston
University School of Medicine, 72 East Concord St, Boston, MA 02118([email protected]).
The Journal of Infectious Diseases 2011;204:S861–S870� The Author 2011. Published by Oxford University Press on behalf of the InfectiousDiseases Society of America. All rights reserved. For Permissions, please e-mail:[email protected] (print)/1537-6613 (online)/2011/204S3-0016$14.00DOI: 10.1093/infdis/jir308
Recombitant Marburg Virus Encoding EGFP d JID 2011:204 (Suppl 3) d S861
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
efficiently expressed over 10 passages, confirming the stability
of the EBOV constructs [15].
In this study, we rescued a recombinant MARV from a clone
containing an ATU encoding EGFP. This clone allows for the
visualization of MARV spread in infected cells and was used to
assess the localization of EGFP and nucleocapsid proteins in
infected cells.
MATERIALS AND METHODS
Cell Lines and VirusesVero E6 (African green monkey kidney), HT-1080 (human fi-
brosarcoma), and U2OS (human osteosarcoma) were main-
tained in Dulbecco’s modified Eagle’s medium supplemented
with penicillin (50 units/mL), streptomycin (50 mg/mL), and
10% fetal calf serum (FCS). MARV strain Musoke and re-
combinant Marburg viruses were propagated in Vero E6 cells as
described previously [9]. All work with infectious MARV was
performed under biosafety level 4 conditions at the Institute of
Virology, Philipps University of Marburg, Marburg, Germany.
Generation of an Infectious MARV Clone Expressing EGFPThe MARV strain Musoke cDNA clone pMARV(1) described
in [9] was used as a template to insert an ATU encoding EGFP
between the second and third genes. The intergenic region be-
tween VP35 and VP40 genes spanning 5 nucleotides (CTATG)
was mutated by in vitro mutagenesis, generating an AvrII re-
striction site (CCTAGG; inserted or substituted nucleotides
underlined). The AvrII restriction site was then used to insert the
ATU consisting of the EGFP open reading frame (ORF) flanked
by authentic MARV transcription start and stop signals [22].
Virus rescue was performed as previously described [10].
Stable integration of the ATU in the viral genome was verified
by reverse transcription–polymerase chain reaction (RT-PCR).
Vero E6 cells were infected with rMARV-EGFP and total RNA
was isolated from cells and supernatants at 6 days postinfection
(dpi) using TRIZOL reagent (Invitrogen). The isolated RNA was
subjected to RT-PCR (OneStep RT-PCR, Qiagen) using primers
flanking a 362–base pair (bp) PCR fragment of the EGFP gene.
TransfectionsHT-1080 cells, grown on glass coverslips, were transfected using
FuGeneHD (Roche), and U2OS cells were transfected using
TransIT-LT1 (Mirus) according to the suppliers’ protocols.
Unless otherwise stated, cells were transfected with 50 ng ex-
pression plasmid for EGFP, or red fluorescent proteins TagRFP,
DsRed, or mCherry in the absence of or along with plasmids
encoding NP (500 ng) and VP35 (500 ng).
Immunofluorescence Analysis of Infected Cells105 Vero E6 cells per well of a 6-well plate were infected with
rMARV-EGFP at a multiplicity of infection (MOI) of 0.05. At 2
and 5 dpi, cells were fixed in 4% (w/v) paraformaldehyde for at
least 24 hours and permeabilized with a mixture of acetone and
methanol (1:1, v/v) for 5 minutes at 220�C. As primary anti-
bodies, a rabbit antiserum directed against the nucleocapsid
complex of MARV (anti-NC antiserum) or a goat anti-MARV
antiserum were used. Antibody binding was visualized by using
Alexa Fluor 568-conjugated and Alexa Fluor 594-labeled sec-
ondary antibodies (Invitrogen). In addition, the cells were
stained with 100 ng/mL 4#-6-diamidino-2-phenylindole (DAPI)
for 10 minutes.
Virus titration was performed by counting foci of infected cells.
Vero E6 cells were infected with recombinantMARV at anMOI of
0.05. Supernatants were collected at 2 and 6 dpi, purified by low-
speed centrifugation, and 500 lL of the supernatants was used for
infection of 105 Vero E6 cells per well of a 6-well plate. Cells were
fixed and permeabilized at 2 dpi as described above. Staining of
infected cells was performed using the anti-NC antiserum. Foci of
infected cells were counted by UV fluorescence microscopy.
Immunofluorescence Analysis of Transfected CellsHT-1080 or U2OS cells were transfected as described above and
subjected to immunofluorescence analysis at 1 day post trans-
fection (dpt). Cells were fixed with 4% (w/v) paraformaldehyde
and permeabilized with 0.1% (v/v) Triton X100. A MARV anti-
NC rabbit antiserum was used to detect MARV proteins. For the
detection of ZEBOV and Reston ebolavirus (REBOV) proteins,
a goat anti-ZEBOV serum that cross-reacts with REBOVNP was
used. Alexa Fluor 594–conjugated antibodies were used for vi-
sualization. The cell nuclei were stained with DAPI.
Western Blot AnalysisVero E6 cells seeded in 6-well plates were infected with recombi-
nant MARV at an MOI of 0.05. At 2 and 5 dpi, cells were scraped
into 200 lL radioimmunoprecipitation assay (RIPA) buffer (20
mM Tris–HCl, pH 7.5; 150 mMNaCl; 10 mM EDTA; 0.1% (w/v)
SDS; 1% (v/v) Triton X100; 1% (v/v) deoxycholate; 10 mM
iodacetamide) and subjected toWestern blot analysis usingmouse
monoclonal antibodies directed against EGFP (B-2; Santa Cruz
Biotechnology), MARV NP, MARV VP40, or b-actin (Abcam).
As a secondary antibody, an IRDye800-conjugated antibody was
used (Rockland). Protein bands were quantified using an Odyssey
imaging system (LI-COR) and standardized to b-actin.
Live-Cell ImagingVero E6 cells were infected with rMARV-EGFP at an MOI of
0.05 in a l-Dish35mm (Ibidi). At 1 hpi, the inoculum was re-
placed by GIBCO Leibovitz’s L-15 Medium (Invitrogen) con-
taining 20% (v/v) FCS. The cell monolayer was analyzed with
a DM16000B Leica inverted fluorescence microscope. EGFP
fluorescence and phase contrast images were captured every
hour for a period of 9 days. Images were taken with a 203
objective. A Zeiss Axiovert 200 M inverted microscope was used
for live-cell imaging of the transfected cells. Fluorescence and
phase contrast images were taken with a 403 objective.
S862 d JID 2011:204 (Suppl 3) d Schmidt et al
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
RESULTS
Rescue of Recombinant Marburg Virus Expressing EGFPAn additional ATU encoding EGFP was inserted into a full-
length antigenomic cDNA plasmid pMARV(1) of MARV strain
Musoke [9, 23] between the second and third genes, encoding
VP35 and VP40, respectively (Figure 1A).
Successful rescue of the MARV containing the EGFP gene
(rMARV-EGFP) was confirmed by RT-PCR, detection of EGFP
expression in infected cells, Western blot analysis, and immu-
nofluorescence analysis. Stable integration of the EGFP gene in
the viral genome was verified by amplification of a 362-bp PCR
fragment of the EGFP gene using template RNA isolated from
rMARV-EGFP-infected cells (Figure 1B) and by sequencing.
ExpressionofEGFP in livingVeroE6cells infectedwith rMARV-
EGFPwas analyzedbyphase contrast andfluorescencemicroscopy.
Foci formation of green fluorescent cells was initiated at 1–2 dpi
without inducing visible cytopathic effects (CPE). The initial signs
of CPE were observed at 5 dpi, when EGFP was detected in
clusters of infected cells (Figure 1C). These data show that
rMARV-EGFP productively infects susceptible cells and can be
used as a sensitive marker to visualize virus spread over time.
To assess the replication efficiency of rMARV-EGFP com-
pared with recombinant wild-type virus (recMARV; described
in [9]), supernatant fluids of Vero E6 cells infected with either
virus were collected at 6 dpi and used for infection of Vero E6
cells. At 2 dpi, cells were subjected to immunofluorescence
analysis using an antiserum directed against MARV nucleo-
capsid proteins (anti-NC antiserum). Foci of infected cells were
counted by fluorescence microscopy (Figure 1D). Progeny virus
production of rMARV-EGFP was reduced approximately 4-fold
compared with wild-type virus.
To further address this, we compared the protein expression of
both viruses. Vero E6 cells were infected as described above, har-
vested at 2 and 5 dpi, and lysates were analyzed by quantitative
Western blot analysis using antibodies directed against EGFP,
MARVNP,MARV VP40, and actin. While viral proteins could be
readily detected at 2 dpi, EGFP accumulated to detectable levels
only at 5 dpi, which might be due to differences in the sensitivity
of the used antibodies (Figure 1E). NP levels were similar at all
time points, whereas VP40 levels were reduced in rMARV-EGFP-
infected cells at 2 dpi and to a lesser extent at 5 dpi. Since the EGFP
gene is located downstream of NP and upstream of VP40 gene
(Figure 1A), the reduced VP40 expression in rMARV-EGFP
indicates that the presence of the ATU causes a decrease in
downstream protein expression, thereby explaining the slightly
growth-restricted phenotype of rMARV-EGFP.
Cell-to-Cell Spread of MARV-EGFP Observed by Live-CellImagingNext we examined the spread of rMARV-EGFP in cell
culture by live-cell imaging. Vero E6 cells were infected with
rMARV-EGFP and the cell monolayer was analyzed by collect-
ing EGFP fluorescence and phase contrast photomicrographs.
Photomicrographs of 25 different positions were captured every
hour from 1 hpi for a period of 9 days (Supplementary Video;
online only). Single infected cells expressing EGFP were ob-
served at 26 hpi. In most of the infectious centers, EGFP ex-
pression in neighboring cells was detected 20–30 hours later with
a mean value of 24.6 hours, which correlates with the MARV
Musoke replication cycle of approximately 21 hours [24].
However, in some infectious centers, EGFP expression in sur-
rounding cells was observed as late as 48 hpi. Intriguingly, even
late in infection, EGFP fluorescence was not homogenously
distributed throughout the monolayer but restricted to in-
dividual foci, suggesting that virus spread occurred by direct
cell-to-cell contact rather than by release of viral particles.
Typically, individual infected cells were observed early in in-
fection, and later on, the infection spread to cells in close
proximity to the primarily infected cell (Figure 1C and
Supplementary Video). In addition, virus spread was promoted
by viral replication in actively dividing cells (Figure 2A and
Supplementary Video).
After the first signs of CPE appeared at 5 dpi, the cell
monolayer began to disintegrate at 6–7 dpi, followed by cell
rounding and blebbing of EGFP-expressing cells, which corre-
lates with impending cell death (Figure 2B, arrows). Some, but
not all, fluorescent cells formed large intracytoplasmic vacuoles
resembling vacuolated degenerating cells as described for non-
apoptotic forms of cell death [25] (Figure 2C).
Higher magnification of infected cells revealed that
EGFP was homogenously distributed in the nucleus and in
the cytoplasm, but unexpectedly was also observed in
intracytoplasmic aggregates (Figure 1C, bottom panel). Since
MARV infection leads to the formation of inclusions in in-
fected cells, we examined whether the EGFP aggregates were
localized with nucleocapsid-derived inclusions. Therefore, at
2 and 5 dpi, rMARV-EGFP-infected cells were examined by
indirect immunofluorescence using anti-NC antiserum rec-
ognizing the nucleocapsid proteins. EGFP autofluorescence
was assessed in parallel. Cytoplasmic EGFP aggregates colo-
calized with MARV-induced inclusions (Figure 3A). In-
terestingly, immunofluorescence analysis revealed infected
cells that were stained with the virus-specific antiserum but
lacked detectable EGFP expression at 5 dpi, indicating that
immunodetection using virus-specific antibodies is more
sensitive than EGFP detection. To exclude the possibility of
‘‘cross-talk’’ or nonspecific binding of antibodies, rMARV-
EGFP-infected cells were stained with a goat anti-MARV an-
tiserum that predominantly recognizes the MARV surface
protein GP. Surface staining of infected cells was observed
with the GP-specific antibody (Figure 3B, middle panels, red
staining). However, green fluorescent inclusions were also
visible (Figure 3B, left panels).
Recombitant Marburg Virus Encoding EGFP d JID 2011:204 (Suppl 3) d S863
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
EGFP Accumulates in Nucleocapsid Protein-Derived InclusionBodies
To further investigate the nature of the EGFP-positive viral in-
clusions, EGFP was expressed in the absence or presence of
MARV nucleocapsid proteins NP and VP35, and cells were
monitored by live-cell imaging. We used U2OS cells for the
transfection experiments because these cells are large and flat,
resulting in high-quality images. U2OS cells were transfected
Figure 1. Characterization of recombinant MARV containing an ATU encoding EGFP. A, Scheme of rMARV-EGFP genome. The EGFP coding sequence isflanked by conserved MARV transcription start and stop signals. The EGFP ORF was inserted between the VP35 and VP40 gene via a newly createdAvr II restriction site within the intergenic region (IR). The intergenic region spanning 5 nucleotides (CTATG) was altered to CCTAGG. B, Detection ofrecombinant genomes in supernatants and cell lysates of rMARV-EGFP-infected Vero E6 cells by RT-PCR. RT-PCR was conducted using primers binding inthe EGFP ORF. Cellular RNA from Vero E6 cells transiently expressing EGFP was used as a positive control. C, Fluorescence microscopy of rMARV-EGFP-infected cells. Living cells were analyzed by phase contrast and fluorescence microscopy. Images were collected at 2 and 5 dpi. Bottom panel showsrMARV-EGFP-infected cell at higher magnification. Inclusions are indicated by an arrow. D, Comparison of progeny production of recombinant wild-typeMARV (recMARV) and rMARV-EGFP. Vero E6 were seeded on glass coverslips and infected with recMARV or rMARV-EGFP. At 2 dpi, cells were subjectedto immunofluorescence analysis using a MARV-specific antibody, and foci of infected cells were counted. The experiment was performed in triplicate andthe bars represent mean values, including standard deviations. E, Quantitative Western blot analysis of virus protein and EGFP levels in Vero E6 cellsinfected with recombinant wild-type recMARV or rMARV-EGFP. Assays were performed in triplicate and standard deviations are shown.
S864 d JID 2011:204 (Suppl 3) d Schmidt et al
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
with 500 ng or 50 ng EGFP expression plasmids along with
plasmids encoding MARV NP and VP35 genes. When 50 ng of
EGFP plasmid was used for transfection, EGFP accumulated in
the nucleus but was also observed in the cytoplasm of transfected
cells, where it was concentrated in inclusion-like aggregates
surrounded by homogenously distributed protein (Figure 4A).
We also observed cells in which EGFP was not concentrated in
intracytoplasmic inclusions. However, since it was not possible
to verify the expression of the nucleocapsid proteins in the live-
cell imaging studies, it is not clear whether these cells expressed
NP and VP35. When 500 ng of EGFP plasmid was used for
transfection, it was difficult to distinguish between concentrated
EGFP and nonspecifically distributed EGFP due to the high
intensity of overexpressed EGFP. Inclusions were only observed
in few cells with lower EGFP expression (Figure 4A). EGFP
inclusions were not detected in cells expressing EGFP in the
absence of NP and VP35 (Figure 4A, bottom panels).
Next, the distribution of EGFP in transfected cells was
analyzed by indirect immunofluorescence. Since concentrated
EGFP could not be differentiated from the homogenous
nonspecific distribution when 500 ng EGFP plasmid was used
for transfection, U2OS cells were transfected with 50 ng EGFP
plasmid along with plasmids for NP and VP35 stained
with anti-NC antiserum. As shown in the upper panels of
Figure 4B, EGFP was distributed in a punctate pattern and
colocalized with intracytoplasmic nucleocapsid inclusions.
Interestingly, the amount of intracytoplasmic homogenously
distributed EGFP was reduced compared with the live-cell
imaging data, which might be due to fixation and/or per-
meabilization effects or due to the fact that EGFP is constantly
Figure 2. Time-lapse fluorescent microscopy of rMARV-EGFP spread. Vero E6 cells were infected with rMARV-EGFP at an MOI of 0.05, and EGFPfluorescence and phase contrast images were captured every hour for a period of 9 days. A, Cell division of infected cells. B and C, Cytopathic effects atlate stages of infection. Blebbing cells are indicated by arrows. Time points postinfection when images were taken are indicated.
Recombitant Marburg Virus Encoding EGFP d JID 2011:204 (Suppl 3) d S865
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
being formed in the live cells, leading to limited bleaching. To
exclude the possibility of antibody cross-reactivity, antibody
staining was omitted. Fluorescence analysis revealed that
EGFP was distributed in large cytoplasmic aggregates (Figure 4B,
middle panels). In contrast, EGFP was homogenously distri-
buted when expressed in the absence of NP and VP35
(Figure 4B, bottom panels). The transfection experiments were
repeated using HT-1080 cells with a similar outcome (data not
shown).
To analyze whether the association with filovirus inclusions is
restricted to EGFP or can also be observed with other fluorescent
proteins, we expressed MARVNP and VP35 proteins along with
various fluorescent proteins from a range of different taxa and
exhibiting different physicochemical features. We selected
TagRFP and DsRed as genealogically different proteins, which
share about 20% amino acid sequence identity with EGFP
[26, 27]. In addition, the monomeric mCherry derivative of the
tetrameric DsRed was included [28]. Intriguingly, each of the
fluorescent proteins colocalized with nucleocapsid-derived
inclusions, when coexpressed with NP and VP35 (Figure 5A).
Similar to EGFP, fluorescent inclusions were also observed when
antibody staining was omitted (Figure 5A, right panels). As
a control, the examined fluorescent proteins were expressed in
the absence of NP and VP35, and were found to be homoge-
neously distributed in the cells (data not shown).
Next, we addressed the question of the observed relocalization
of EGFP in virus-derived inclusions was specific for MARV or
could also be observed with other filovirus species. Therefore,
EGFP was coexpressed with either REBOV or ZEBOV NP and
VP35 proteins. Cells were subjected to immunofluorescence
analysis at 1 dpi using an anti-ZEBOV antiserum that cross-
reacts with REBOV NP. Figure 5B shows that EGFP colocalized
with both ZEBOV and REBOV inclusions, demonstrating that
EGFP accumulation in inclusion bodies is not restricted to
MARV. These data demonstrate that the accumulation of
coexpressed proteins in inclusion bodies is neither restricted to
MARV nor EGFP, but occurs irrespective of filoviral species or
fluorescent proteins used in the assay.
Figure 3. Fluorescence microscopy analysis of EGFP and immunohistochemically labeled viral proteins in rMARV-EGFP-infected cells. Vero E6 cellswere infected with rMARV-EGFP and subjected to immunofluorescence analysis at 2 and 5 dpi. Antibodies were directed against (A) intracellular viralproteins or (B) viral surface proteins. Antibody staining is indicated by red color; EGFP autofluorescence, green; and DAPI staining of the nuclei, blue.
S866 d JID 2011:204 (Suppl 3) d Schmidt et al
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
DISCUSSION
Here we report the generation of a recombinant MARV ex-
pressing EGFP from an ATU inserted between the VP35 and
VP40 genes. We chose this position for the insertion of the ATU
to avoid altering the balance between NP (first gene product)
and VP35 (second gene product), since previous results with
a MARV minigenome system suggested that the ratio of NP to
VP35 is critical for efficient replication and transcription [5].
EGFP has been expressed from the closely related ZEBOV from
different positions in the genome, and insertion of the ATU
between the NP and VP35 genes did not lead to significant
growth defects in cell culture. However, the virus was attenuated
in a STAT-1 knockout mouse model [14, 15]. Similar effects
were observed with a ZEBOV variant containing the ATU be-
tween the VP30 and VP24 genes (fifth and sixth genes).
This virus showed no or mild growth defects in cell culture
depending on the cell line used for propagation, was moderately
attenuated in the mouse model, and was severely attenuated in
a nonhuman primate model [15]. A recombinant ZEBOV in
which the ATU was added between the VP35 and VP40 genes
(second and third genes) could be rescued and propagated in
cell culture [16]. Taken together, these data show that ZEBOV
tolerates the addition of a foreign gene at different positions,
although the insertion of extra nonviral genetic material may
lead to reduced virulence in animal models.
Although replication of the recombinant MARV expressing
EGFP was reduced 4-fold in cell culture, it was successfully used
to monitor viral spread in living cells. Our data suggest that virus
spread in the infectious centers occurred predominantly
through cell-to-cell-contact. Release of viral particles in MARV-
infected cells takes place at filamentous protrusions, the filo-
podia [29]. Since filopodia act as sensory cellular organelles to
explore the extracellular environment, including neighboring
Figure 4. Accumulation of EGFP in MARV inclusions formed by NP and VP35. U2OS cells were transfected with an EGFP expression construct alone oralong with plasmids encoding MARV NP and MARV VP35, as indicated. A, Live-cell imaging of transfected cells. EGFP autofluorescence is shown ingreen. Intracytoplasmic EGFP aggregates are indicated by arrows. B, Cells were stained using an antiserum directed against MARV nucleocapsid proteins(anti-NC; red) and DAPI (blue).
Recombitant Marburg Virus Encoding EGFP d JID 2011:204 (Suppl 3) d S867
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
cells [30], it has been suggested that MARV particles may bud
into adjacent cells via filopodia-mediated cell-to-cell contact
[29]. Besides cell-to-cell-contact, virus replication in actively
dividing cells seems to be an important mechanism of MARV
spread in cell culture. Cell division was not inhibited by MARV
infection, indicating that MARV does not interfere with cell
cycle progression. The collected data clearly illustrate the
strength of live-cell imaging.
Toward the end of the observation period of 9 days, cell
rounding, blebbing, then detachment of infected cells was ob-
served. Some of the infected cells formed large intracytoplasmic
vacuoles. While blebbing is associated with both apoptotic and
necrotic cell death [31], vacuolization of dying cells has been
described for nonapoptotic cell death, such as necrosis or au-
tophagy [25], suggesting that MARV-infected cells might not
undergo apoptosis late in infection. Although the induction of
apoptosis in the context of filovirus infection has been observed
in bystander cells, there are conflicted data on the induction of
apoptosis in infected cells [32–36].
Surprisingly, we found that EGFP accumulates in filoviral
inclusions. A similar observation was reported for some mem-
bers of the nucleorhabdoviruses, where green fluorescent pro-
tein colocalized with viral nucleocapsid protein in loci within
and around the nuclei [37]. In contrast, EGFP was found to be
homogenously distributed in the nuclei and cytoplasm of cells
infected with EGFP-expressing measles virus, which also pro-
duces intracytoplasmic inclusions [38, 39]. The intracellular
distribution of EGFP was examined by live-cell imaging and
immunofluorescence analysis using infected and transfected
cells. In both infected and transfected cells, the fluorescence
intensity of homogenously distributed EGFP surrounding the
intracytoplasmic EGFP aggregates was higher in living cells,
making it difficult to distinguish between EGFP aggregates and
nonspecific-distributed EGFP. Moreover, when large amounts
Figure 5. Accumulation of coexpressed fluorescent proteins (FPs) in nucleocapsid protein-induced inclusions. A, Fluorescence microscopy analysis ofHT-1080 cells expressing TagRFP (red), DsRed (orange), or mCherry (pink), along with MARV NP and VP35. Cells were either stained with anti-NCantiserum (green) and DAPI (blue) or DAPI alone (right panels). B, Fluorescence microscopy analysis of HT-1080 cells coexpressing EGFP along with ZEBOV(upper panel) or REBOV (lower panel) NP and VP35 proteins. Cells were stained with DAPI (blue) and an antibody detecting the nucleocapsid proteins ofboth virus species (red).
S868 d JID 2011:204 (Suppl 3) d Schmidt et al
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
of EGFP plasmid were used for transfection (500 ng), the
punctate pattern of EGFP was not observed in cells coexpressing
MARV NP and VP35, suggesting that the EGFP aggregates were
masked when overall EGFP expression was high, illustrating the
importance of making observations in cells that do not express
too much of a protein. In fixed and permeabilized cells, punctate
EGFP was clearly visible in both infected cells and transfected
cells coexpressing NP and VP35, indicating that the intensity of
EGFP autofluorescence was reduced by the treatment of the
cells.
Intriguingly, EGFP did not only colocalize with MARV in-
clusions but also with ZEBOV and REBOV inclusions formed by
NP and VP35. In addition, fluorescent proteins other than EGFP
also accumulated in MARV inclusions. These data indicate that
the accumulation of ectopic proteins in filoviral inclusions is
most likely not mediated by direct protein–protein interaction,
and future studies are planned to elucidate the underlying
mechanisms. For now, the observed colocalization of ectopic
fluorescent proteins with filovirus inclusions may be useful to
investigate nucleocapsid maturation and transport in infected
cells without tagging viral proteins.
In conclusion, this study describes the generation and
characterization of an EGFP-containing MARV, which will be
a useful tool for imaging-based antiviral drug screening as-
says. High-throughput screening assays based on the expres-
sion of EGFP are currently the preferred approaches for
testing potential therapeutic compounds against EBOV in-
fection [19], and the availability of an EGFP-expressing
MARV facilitates the development of similar assays for
MARV. Furthermore, we demonstrated that rMARV-EGFP
can be used to study important steps of the viral replication
cycle in living cells, including virus spread and infection-
related morphological changes. Since EGFP and other fluores-
cent proteins colocalize with the viral inclusions, the putative
sites of viral replication, rMARV-EGFP could also be used to
study the temporal and spatial regulatory steps involved in the
formation of these virus-derived intracytoplasmic structures.
Future studies are planned to use rodent-adapted recombinant
Marburg viruses expressing EGFP or luciferase as valuable tools
for the development of whole-animal imaging assays and
pathogenesis studies.
Supplementary Data
Supplementary video available at The Journal of Infectious Diseases
online.
Funding
This work was supported by the Manchot Foundation (to K. M. S. and J.
O.); by funds from the German Research Foundation (SFB 535); by National
Institutes of Health (NIH; grants AI082954 and AI057159; New England
Regional Center of Excellence–Kasper, subaward 149047-0743); and by start-
up funds from Boston University.
Acknowledgments
The authors are grateful to O. Dolnik, Philipps University of Marburg,
Germany, for rescue of rMARV-EGFP, J. Connor, Boston University,
Boston, MA, for his help with carrying out live-cell imaging, and W. P.
Duprex, Boston University, for critical review of the manuscript. We also
thank N. Kedersha and P. Anderson, Brigham And Women’s Hospital,
Harvard Medical School, Boston, MA. for providing the U2OS cells, V.
von Messling, University of Quebec, Quebec, Canada, for the mCherry
plasmid, and S. Becker, Philipps University of Marburg, for kindly pro-
viding anti-MARV antiserum, anti-EBOV antiserum, and anti-MARV
VP40 monoclonal antibody.
References
1. Muhlberger E. Filovirus replication and transcription. Future Virol
2007; 2:205–15.
2. Kolesnikova L, Muhlberger E, Ryabchikova E, Becker S. Ultrastruc-
tural organization of recombinant Marburg virus nucleoprotein:
comparison with Marburg virus inclusions. J Virol 2000; 74:
3899–904.
3. Mavrakis M, Kolesnikova L, Schoehn G, Becker S, Ruigrok RW.
Morphology of Marburg virus NP-RNA. Virology 2002; 296:300–7.
4. Becker S, Rinne C, Hofsass U, Klenk H-D, Muhlberger E. Interactions
of Marburg virus nucleocapsid proteins. Virology 1998; 249:406–17.
5. Muhlberger E, Lotfering B, Klenk H-D, Becker S. Three of the four
nucleocapsid proteins of Marburg virus, NP, VP35, and L, are sufficient
to mediate replication and transcription of Marburg virus-specific
monocistronic minigenomes. J Virol 1998; 72:8756–64.
6. Moller P, Pariente N, Klenk HD, Becker S. Homo-oligomerization of
Marburgvirus VP35 is essential for its function in replication and
transcription. J Virol 2005; 79:14876–86.
7. DiCarlo A, Moller P, Lander A, Kolesnikova L, Becker S. Nucleocapsid
formation and RNA synthesis of Marburg virus is dependent on two
coiled coil motifs in the nucleoprotein. Virol J 2007; 4:105.
8. Hartlieb B, Muziol T, Weissenhorn W, Becker S. Crystal structure of
the C-terminal domain of Ebola virus VP30 reveals a role in tran-
scription and nucleocapsid association. Proc Natl Acad Sci USA 2007;
104:624–9.
9. Enterlein S, Volchkov V, Weik M, et al. Rescue of recombinant Mar-
burg virus from cDNA is dependent on nucleocapsid protein VP30.
J Virol 2006; 80:1038–43.
10. Krahling V, Dolnik O, Kolesnikova L, et al. Establishment of fruit bat
cells (Rousettus aegyptiacus) as a model system for the investigation of
filoviral infection. PLoS Negl Trop Dis 2010; 4:e802.
11. Volchkov VE, Volchkova VA, Muhlberger E, et al. Recovery of in-
fectious Ebola virus from complementary DNA: RNA editing of the GP
gene and viral cytotoxicity. Science 2001; 291:1965–9.
12. Neumann G, Feldmann H, Watanabe S, Lukashevich I, Kawaoka Y.
Reverse genetics demonstrates that proteolytic processing of the Ebola
virus glycoprotein is not essential for replication in cell culture. J Virol
2002; 76:406–10.
13. Theriault S, Groseth A, Neumann G, Kawaoka Y, Feldmann H. Rescue
of Ebola virus from cDNA using heterologous support proteins. Virus
Res 2004; 106:43–50.
14. Towner JS, Paragas J, Dover JE, et al. Generation of eGFP expressing
recombinant Zaire ebolavirus for analysis of early pathogenesis events
and high-throughput antiviral drug screening. Virology 2005;
332:20–7.
15. Ebihara H, Theriault S, Neumann G, et al. In vitro and in vivo char-
acterization of recombinant Ebola viruses expressing enhanced green
fluorescent protein. J Infect Dis 2007; 196(Suppl 2):S313–22.
16. Martinez MJ, Biedenkopf N, Volchkova V, et al. Role of Ebola virus
VP30 in transcription reinitiation. J Virol 2008; 82:12569–73.
17. Prins KC, Delpeut S, Leung DW, et al. Mutations abrogating VP35
interaction with double-stranded RNA render Ebola virus avirulent in
guinea pigs. J Virol 2010; 84:3004–15.
Recombitant Marburg Virus Encoding EGFP d JID 2011:204 (Suppl 3) d S869
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021
18. Michelow IC, Dong M, Mungall BA, et al. A novel L-ficolin/mannose-
binding lectin chimeric molecule with enhanced activity against Ebola
virus. J Biol Chem 2010; 285:24729–39.
19. Panchal RG, Kota KP, Spurgers KB, et al. Development of high-content
imaging assays for lethal viral pathogens. J Biomol Screen 2010; 15:755–65.
20. Radoshitzky SR, Dong L, Chi X, et al. Infectious Lassa virus, but not
filoviruses, is restricted by BST-2/tetherin. J Virol 2010; 84:10569–80.
21. Saeed MF, Kolokoltsov AA, Albrecht T, Davey RA. Cellular entry of
Ebola virus involves uptake by a macropinocytosis-like mechanism and
subsequent trafficking through early and late endosomes. PLoS Pathog
2010; 6e1001110.
22. Muhlberger E, Trommer S, Funke C, Volchkov V, Klenk H-D, Becker S.
Termini of all mRNA species of Marburg virus: sequence and secondary
structure. Virology 1996; 223:376–80.
23. Enterlein S, Schmidt KM, Schumann M, et al. The Marburg virus 3#noncoding region structurally and functionally differs from that of
Ebola virus. J Virol 2009; 83:4508–19.
24. Muhlberger E. Genome organization, replication, and transcription
of filoviruses. In: Feldmann H, Klenk H-D, eds. Ebola and Marburg
viruses: molecular and cellular biology. Wymondham, UK: Horizon
Scientific Press, 2004; 1–12.
25. Overmeyer JH, Kaul A, Johnson EE, Maltese WA. Active ras triggers
death in glioblastoma cells through hyperstimulation of macro-
pinocytosis. Mol Cancer Res 2008; 6:965–77.
26. Merzlyak EM, Goedhart J, Shcherbo D, et al. Bright monomeric red
fluorescent protein with an extended fluorescence lifetime. Nat
Methods 2004; 4:555–7.
27. Chudakov DM, Matz MV, Lukyanov S, Lukyanov KA. Fluorescent
proteins and their applications in imaging living cells and tissues.
Physiol Rev 2010; 90:1103–63.
28. Shaner NC, Campbell RE, Steinbach PA, Giepmans BN, Palmer AE,
Tsien RY. Improved monomeric red, orange and yellow fluorescent
proteins derived from Discosoma sp. red fluorescent protein. Nat
Biotechnol 2004; 22:1567–72.
29. Kolesnikova L, Bohil AB, Cheney RE, Becker S. Budding of Marburg-
virus is associated with filopodia. Cell Microbiol 2007; 9:939–51.
30. Mellor H. The role of formins in filopodia formation. Biochim Biophys
Acta 2010; 1803:191–200.
31. Barros LF, Kanaseki T, Sabirov R, et al. Apoptotic and necrotic blebs in
epithelial cells display similar neck diameters but different kinase de-
pendency. Cell Death Differ 2003; 10:687–97.
32. Baize S, Leroy EM, Georges-Courbot MC, et al. Defective humoral
responses and extensive intravascular apoptosis are associated with
fatal outcome in Ebola virus-infected patients [see comments]. Nat
Med 1999; 5:423–6.
33. Bradfute SB, Braun DR, Shamblin JD, et al. Lymphocyte death
in a mouse model of Ebola virus infection. J Infect Dis 2007;
196(Suppl 2):S296–304.
34. Bradfute SB, Swanson PE, Smith MA, et al. Mechanisms and con-
sequences of ebolavirus-induced lymphocyte apoptosis. J Immunol
2010; 184:327–35.
35. Geisbert TW, Hensley LE, Gibb TR, Steele KE, Jaax NK, Jahrling PB.
Apoptosis induced in vitro and in vivo during infection by Ebola and
Marburg viruses. Lab Invest 2000; 80:171–86.
36. Gupta M, Spiropoulou C, Rollin PE. Ebola virus infection of human
PBMCs causes massive death of macrophages, CD4 and CD8 T cell
sub-populations in vitro. Virology 2007; 364:45–54.
37. Goodin M, Yelton S, Ghosh D, Mathews S, Lesnaw J. Live-cell imaging
of rhabdovirus-induced morphological changes in plant nuclear
membranes. Mol Plant Microbe Interact 2005; 18:703–9.
38. Duprex WP, McQuaid S, Hangartner L, Billeter MA, Rima BK. Ob-
servation of measles virus cell-to-cell spread in astrocytoma cells by
using a green fluorescent protein-expressing recombinant virus. J Virol
1999; 73:9568–75.
39. Duprex WP, McQuaid S, Roscic-Mrkic B, Cattaneo R, McCallister C,
Rima BK. In vitro and in vivo infection of neural cells by a recombinant
measles virus expressing enhanced green fluorescent protein. J Virol
2000; 74:7972–9.
S870 d JID 2011:204 (Suppl 3) d Schmidt et al
Dow
nloaded from https://academ
ic.oup.com/jid/article/204/suppl_3/S861/2192363 by guest on 20 D
ecember 2021