-
1
Title: Niclosamide inhibits SARS-CoV2 entry by blocking
internalization through pH-dependent
CLIC/GEEC endocytic pathway
Authors:
Chaitra Prabhakara1,*, Rashmi Godbole1,2,*, Parijat Sil1,*,
Sowmya Jahnavi1,*, Thomas S van Zanten1,#,
Dhruv Sheth1,#, Neeraja Subhash1, Anchal Chandra1, Vijay Kumar
Nuthakki4, Theja Parassini
Puthiyapurayil3, Riyaz Ahmed4, Ashaq Hussain Najar4, Sai Manoz
Lingamallu3,5, Snigdhadev Das1,
Bhagyashri Mahajan1, Praveen Vemula3, Sandip B Bharate4,
Parvinder Pal Singh4, Ram Vishwakarma4,
Arjun Guha3, Varadharajan Sundaramurthy1 and Satyajit Mayor1
Affiliations:
1- National Centre for Biological Sciences (TIFR), Bengaluru,
India
2- University of Trans-Disciplinary Health Sciences and
Technology (TDU), Bengaluru, India
3- Institute for Stem Cell Science and Regenerative Medicine
(inSTEM), Bengaluru, India
4- CSIR - Indian Institute of Integrative Medicine, Jammu,
India
5- Manipal Academy of Higher Education (MAHE), Madhav Nagar,
Manipal, Karnataka, India
* and # contributed equally
Correspondence to: [email protected]
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
mailto:[email protected]://doi.org/10.1101/2020.12.16.422529
-
2
Abstract: 1
Many viruses utilize the host endo-lysosomal network to infect
cells. Tracing the endocytic itinerary of 2
SARS-CoV2 can provide insights into viral trafficking and aid in
designing new therapeutic targets. 3
Here, we demonstrate that the receptor binding domain (RBD) of
SARS-CoV2 is internalized via the 4
clathrin and dynamin-independent, pH-dependent CLIC/GEEC (CG)
endocytic pathway. Endosomal 5
acidification inhibitors like BafilomycinA1 and NH4Cl, which
inhibit the CG pathway, strongly block 6
the uptake of RBD. Using transduction assays with SARS-CoV2
Spike-pseudovirus, we confirmed that 7
these acidification inhibitors also impede viral infection. By
contrast, Chloroquine neither affects RBD 8
uptake nor extensively alters the endosomal pH, yet attenuates
Spike-pseudovirus entry, indicating a 9
pH-independent mechanism of intervention. We screened a subset
of FDA-approved acidification 10
inhibitors and found Niclosamide to be a potential SARS-CoV2
entry inhibitor. Niclosamide, thus, 11
could provide broader applicability in subverting infection of
similar category viruses entering host 12
cells via this pH-dependent endocytic pathway. 13
Keywords: SARS-CoV2 entry, CLIC/GEEC endocytosis,
Spike-pseudovirus, Endosomal acidification 14
inhibitors, Niclosamide, Chloroquine 15
Introduction: 16
Coronaviruses (CoVs) are a group of related enveloped RNA
viruses of which two alpha CoVs (229E 17
and NL63) and four beta CoVs (OC43, HKU, SARS, and MERS) are
known to cause respiratory tract 18
infections in humans. The recent emergence of SARS-CoV2 and its
rapid spread across the world has 19
posed a global health emergency 1. While several therapeutic
strategies are currently being used to 20
alleviate the respiratory symptoms of patients infected with
SARS-CoV2 2,3, there is limited 21
understanding of the cell biology of viral entry as well as the
availability of drugs which target this 22
process. A search for antivirals affecting the endocytic entry
of viruses is particularly exciting as 23
infections from multiple related viruses can be controlled
through the inhibition of a common step. 24
Virus entry into host cells is a multistep process. A key step
in successful invasion is the release of viral 25
genomic content into the host cell cytoplasm. To achieve this,
viruses bind to specific cell surface 26
receptors and subsequently undergo membrane fusion either
directly at the plasma membrane or 27
following endocytic uptake. While fusion directly at the plasma
membrane is well established for HIV 28
and Influenza virus infections 4,5, both alternatives of entry
are feasible for CoV infections depending 29
on the availability of receptors and proteases at the host cell
surface. Different CoVs interact with a 30
range of specific receptors for entry. For instance, CoV 229E
binds to CD13 (aminopeptidase N) 6, 31
CoVs OC43 and HKU1 recognize 9-O-acetylated sialic acids 7, MERS
uses DPP4/CD26 8 and CoVs 32
NL63 9, SARS-CoV 10 and SARS-CoV2 11 interact with angiotensin
converting enzyme 2 (ACE2). 33
Although ACE2 is a well studied receptor, other receptors for
SARS-CoV2 are being discovered 12–16. 34
Additionally, CoVs require proteolytic processing of the viral
envelope spike protein by host cell 35
proteases to gain entry 17,18. Therefore, these viruses can
directly fuse at the cell surface if the Spike 36
protein is cleaved by a cell surface serine protease like
TMPRSS2 11,19, or utilize an endo-lysosomal 37
route for fusion, where the Spike protein is primed by cysteine
protease Cathepsins 11,20–22. 38
The role of the endo-lysosomal network appears to be crucial in
delivering these viruses to acidic 39
compartments. Cathepsins function optimally in a low pH
environment 17,23. Inhibitors of acidification 40
which increase the pH of endosomal compartments significantly
reduce the infection of spike 41
pseudotyped as well as native MERS-CoV, SARS-CoV, SARS-CoV2
viruses 11,24–27. Supporting this 42
view, drugs inhibiting the maturation of late endosome to
lysosome, Apilimod and YM201636, also 43
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
3
reduce MERS, SARS-CoV, SARS-Cov2 infection 24,28,29. These
studies emphasize the importance of a 44
pH-dependent endocytic route in viral entry and infection.
However, the endocytic routes preferred by 45
SARS-CoV2 for host cell entry are largely unknown. 46
Multiple endocytic pathways operate at the cell surface 30,31.
One of these, the clathrin and dynamin 47
independent CLIC/GEEC (CG) endocytic pathway 32, is of
particular interest here as uptake through 48
this pathway is known to be pH-dependent. Vacuolar ATPases
(V-ATPases), which actively pump 49
protons into the endocytic compartments 33, play a crucial role
in the formation of CG endosomes as 50
established using genetic and pharmacological perturbations
34,35. By contrast, uptake through clathrin 51
mediated endocytosis (CME) remains unaltered upon V-ATPase
perturbation 34. The homotypic fusion 52
of nascent CG endosomes (called CLICs – clathrin-independent
carriers) forms highly acidic early 53
endosomal compartments of the CG pathway (called GEECs – GPI
anchored protein enriched early 54
endosomal compartments) with an estimated luminal pH of 6.0 36.
Thus, GEECs could provide a 55
conducive environment for viral uncoating and membrane fusion.
Additionally, V-ATPases and 56
ARP2/3 complex, both imperative for CG endocytosis 37, are
identified as host factors necessary for 57
SARS-CoV2 viral infection in genome-wide loss of function screen
38. Interestingly, Adeno-associated 58
virus (AAV2) hijacks the CG pathway for infection 39 and
SARS-CoV has also been reported to enter 59
cells through a clathrin and dynamin independent endocytic
pathway 26. These observations prompted 60
us to study the role of CG endocytosis in the context of
SARS-CoV2 entry and infection. 61
In this report, we show that the receptor binding domain (RBD)
of SARS-CoV2 Spike protein is 62
endocytosed through the CG pathway and its uptake is sensitive
to pharmacological perturbations of 63
the CG pathway. RBD uptake, similar to CG cargo uptake, is
strongly inhibited by acidification 64
inhibitors such as BafilomycinA1 and NH4Cl. Inhibitors of
endosomal acidification also blocked 65
infection by SARS-CoV2 Spike-pseudoviruses. Extending our
observations, we conducted a targeted 66
screen using a subset of FDA-approved drugs which are known to
interfere with endosomal 67
acidification. We identified Niclosamide as a promising
candidate that inhibits RBD uptake, Spike-68
pseudovirus infection and, in combination, potentiates the
effects of Hydroxychloroquine. We suggest 69
that Niclosamide could be a feasible start point for developing
small molecule entry inhibitors to 70
mitigate SARS-CoV2 infection. 71
Results: 72
Generation of SARS-CoV2 probe to study its endocytosis itinerary
73
The Spike (S) protein of SARS-CoV2 plays crucial roles in
mediating viral entry to cells. S protein 74
binds to the receptors on the host cell surface through the S1
subunit which harbours the receptor 75
binding domain (RBD) and aids in membrane fusion through the S2
subunit 40. To explore the 76
trafficking route of SARS-CoV2 in human cells, we purified RBD
protein, following a previously 77
established method 41, and generated fluorescently labelled RBD
using NHS-ester chemistry (Figure 78
S1A, S1B, Methods). We chose human adenocarcinoma gastric cells
(AGS cells) as a model system to 79
study RBD uptake as the cell line exhibits multiple endocytic
routes 42 and is also permissive to infection 80
by SARS-CoV2 Spike-pseudovirus (Figure S6E). We tested the
specificity of the labelled RBD probe 81
in AGS cells transiently overexpressing myc-tagged ACE2 and
found that more RBD was bound to 82
cells overexpressing ACE2 (Figure S1C). We observed a positive
correlation between the amount of 83
RBD endocytosed and surface levels of ACE2 (Figure S1D, S1E),
supporting the notion that ACE2 is 84
one of the cell surface receptors of RBD 11. 85
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
4
RBD is internalized via CG endocytosis and RBD uptake is
sensitive to CG Pathway inhibitors 86
We employed the methodology of tracking RBD uptake along with
cargoes specific to CME 87
(transferrin) and CG (10kDa dextran) endocytic pathway to
determine the endocytic route taken up by 88
RBD (Figure 1A). AGS cells without overexpression of ACE2 also
support RBD uptake suggesting 89
that there is sufficient endogenous receptor expressed in these
cells. At 10 minutes post internalization, 90
transferrin endosomes of the CME pathway are distinct from
dextran endosomes of the CG pathway. 91
At these times, internalized RBD is colocalized with endosomes
containing the CG cargo but not the 92
CME cargo (Figure S2A, S2B). At 30 minutes post internalization,
as well, this segregation remains. 93
While a small fraction of RBD endosomes were colocalized with
endosomes containing both transferrin 94
and dextran, a large fraction of RBD endosomes were localized to
compartments uniquely marked by 95
dextran (Figure 1B, 1C; compare % RBD with transferrin and
dextran). This suggests that the itinerary 96
of uptake of RBD is similar to CG cargo and different from CME
cargo. 97
Towards determining the trafficking route of RBD, we examined
the effect of inhibitors of CG pathway 98
on RBD, dextran and transferrin uptake in AGS cells. Cells were
subjected to a brief pre-treatment with 99
different inhibitors (30 minutes), followed by a pulse of RBD,
dextran and transferrin (30 minutes) in 100
the presence of these inhibitors (Methods). CG pathway is
regulated by small GTPases - CDC42 43, 101
Arf1 44 and GEF of Arf1, GBF1 45. Inhibitors that block the
function of these regulators affect the 102
formation of CG endosomes without altering uptake through the
CME pathway. The inhibitor AN96, 103
which is a stable analog of LG-186 46,47, targets GBF1 and
specifically affects the CG pathway (Godbole 104
et al., Manuscript in preparation). We observed that AN96
treatment reduced both RBD and dextran 105
uptake but had minimal effects on the amount of transferrin
internalized (Figure 1D, 1E). We also 106
observed that the peri-nuclear transferrin recycling endosomal
pool was redistributed throughout the 107
cytoplasm upon treatment with AN96 without affecting the net
amount of transferrin internalized. 108
Another CG pathway inhibitor, ML141 (CDC42 inhibitor) 47, also
significantly decreased both dextran 109
as well as RBD uptake (Figure S2E, S2F). 110
Blocking of the CG pathway often results in the redistribution
of CG cargo towards CME pathway 42. 111
Therefore, using high-resolution imaging, we assessed the fate
of the RBD endosomes that continued 112
to be internalized upon treatment with AN96. We observed that an
increased fraction of internalized 113
RBD endosomes colocalized with transferrin in AN96 treated cells
compared to that of control (Figure 114
S2C, S2D(i)). Similarly, increased co-occurrence was observed in
the fraction of dextran endosomes 115
associated with transferrin endosomes on comparing the control
with AN96 treated cells (Figure 116
S2D(iii)). Blocking the CG pathway results in altered
trafficking itinerary of RBD and increases its 117
association with transferrin. This suggests that RBD could be
redirected to be internalized via the CME 118
upon blocking the CG pathway. 119
Since 10kDa dextran marks both CG cargo as well as larger
endocytic compartments like those derived 120
from macropinocytosis 48, we tested if macropinocytosis plays
any role in RBD uptake. Several viruses 121
utilize macropinocytosis pathway as an entry route into cells
49. Macropinocytosis is dependent on 122
amiloride-sensitive Na+/H+ exchangers 50. Upon treatment with
Amiloride, we found no alteration in 123
the uptake of RBD, dextran and transferrin confirming that
macropinocytosis does not play a role in 124
RBD trafficking in AGS cells (Figure S2G, S2H). Together, the
co-localization studies and 125
pharmacological inhibition experiments strongly suggest that RBD
uptake occurs via the CG Pathway 126
and is inhibited by specific blockers of the CG pathway. 127
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
5
RBD and CG uptake is blocked by endosomal acidification
inhibitors – BafilomycinA1 and NH4Cl 128
Given the relevance of acidification in both formation of CG
endosomes 34,36 and in the context of viral 129
infection 51, we focused on studying the role of acidification
inhibitors on the uptake of RBD. We 130
checked the effect of BafilomycinA1 (BafA1), a specific
inhibitor of V-ATPase 52, on RBD, dextran 131
and transferrin uptake in AGS cells. Treatment with 200nM BafA1
strongly reduced both RBD and 132
dextran uptake (Figure 1F, 1G) and enhanced the normalized
transferrin uptake (Figure 1H, 1I). This 133
could be because BafA1 also retards the transferrin recycling
from the recycling endosomes 53 and 134
thereby increasing the net amount of transferrin internalized
within cells as observed. A dose-dependent 135
reduction in RBD and dextran uptake and an increase in
transferrin uptake was seen when cells were 136
treated with a higher concentration of BafA1 (Figure S3A, S3B
(i)). We examined the effect of NH4Cl, 137
a weak base known to alter endosomal acidification 54, on the
uptake of these 3 cargoes. We observed 138
similar results as with BafA1 (Figure S3A, S3B (i)), thus
re-establishing our earlier 34 finding that uptake 139
via the CG pathway is pH sensitive and blocking acidification
results in reduced CG uptake. 140
Towards understanding the mechanism of action for acidification
inhibitors in bringing about these 141
changes in trafficking, we assessed their effect on two
parameters – numbers of endosomes (Figure S3B 142
(ii)) and per-endosome intensity in the presence/absence of
inhibitor (Figure S3B (iii)). We observed 143
that both BafA1 and NH4Cl reduced the total number of RBD and
dextran endosomes without affecting 144
the per-endosome intensity. However, while the total number of
transferrin endosomes remained 145
unchanged, the per-endosome intensity of transferrin increased
with BafA1 and NH4Cl treatment. This 146
indicates that the reduction in RBD and dextran is likely due to
a block in the entry while an increase 147
in per-endosome transferrin intensity could be because of a
block in the formation of recycling 148
endosome carriers, as proposed earlier. 149
We studied the effect of BafA1 on RBD uptake in cells
overexpressing myc-tagged ACE2 receptor by 150
measuring the uptake of RBD normalized to the surface ACE2
levels. We observed that BafA1 strongly 151
affects the normalized RBD uptake (Figure S3C, S3D). HEK-293T
cells, which is also permissive to 152
Spike-pseudovirus transduction (Figure S6C), showed similar
inhibition of RBD and dextran uptake, 153
and increase in transferrin uptake with BafA1 (Figure
S10A-S10D). 154
RBD is localized to acidic compartments 155
Internalized cargoes can be recycled along with the bulk
membrane 55 or directed towards degradation 156
with the fluid phase 56. Typically, transferrin bound to its
receptor marks the early sorting and recycling 157
endosomes and lysotracker labels the acidic degradative
compartments within a cell 31. At 30 minutes 158
of pulse with RBD, dextran and transferrin, while a small
fraction of RBD (~36%) associated with 159
transferrin, the majority of RBD (~84%) co-localized with
dextran suggesting that RBD is directed 160
predominantly towards the degradation route rather than the
recycling route (Figure 1B, 1C). The 161
lysotracker labelling showed highly acidic tubular compartments
with significant co-localization with 162
RBD. At 30 minutes of pulse with RBD, around 55% of RBD
co-localized with lysotracker marked 163
compartments. At longer time points (3 hours) of pulse with RBD,
an even increased proportion of RBD 164
(85%) associated with compartments marked by lysotracker,
confirming that RBD is trafficked to acidic 165
compartments (Figure 2A, 2B). 166
BafilomycinA1 and NH4Cl alter the pH of acidic endosomal
compartments 167
We next focused on determining the change in endosomal pH
brought about by various inhibitors within 168
the acidic compartments populated by RBD. Cells were labelled
with pH-sensitive (FITC) and pH-169
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
6
insensitive (TMR) dextran for 2 hours and chased for 1 hour with
or without inhibitors (Figure 2C, 170
Methods). The above pulse and chase durations were chosen to
allow accumulation of labelled dextran 171
in late endosomes and lysosomal compartments (co-labelled by
Lysotracker, data not shown). 172
Additionally, since the acidification inhibitors also have
inhibitory roles in the early steps of CG 173
endocytosis as discussed in the previous section, to evaluate
their effect on endosomal pH, cells were 174
incubated with inhibitors only during the chase. While the ratio
of the fluorescence of these probes is 175
used to estimate endosomal pH by comparing the ratio with the
calibration curve 57 (Figure S4A, S4B, 176
Methods), quantifications of the endosomal intensities and the
endosomal number of TMR dextran aids 177
in understanding the effect of various drugs on late endosomal
trafficking. 178
Treatment of cells with acidification inhibitors showed an
increase in endosomal pH. The average pH 179
of the late endosomes in control cells was 5.8. The pH of these
compartments increased to 6.2 and 7.1 180
in the presence of BafA1 200nM and 400nM respectively.
Incubation with NH4Cl also resulted in 181
increasing the pH of these endosomes to 6.6 (Figure 2D, 2E,
S4C). While BafA1 marginally changed 182
the TMR intensity per endosome, NH4Cl greatly increased the TMR
intensity indicating that NH4Cl 183
also brings about the fusion of endosomes (Figure S4D). All the
acidification inhibitors also reduced 184
the numbers of endosomes (Figure S4D) and this effect was most
prominent with NH4Cl wherein the 185
endosomes were organized close to the perinuclear region (Figure
S4C). The spatial pH maps show the 186
distribution of pH of endosomes within a cell. Cells treated
with BafA1 400nM and NH4Cl showed a 187
homogenous distribution of endosomes with increased pH similar
to the respective cell averages. BafA1 188
200nM cells, on the other hand, showed heterogeneity in
endosomal pH with some endosomes depicting 189
high pH while others were closer to the average (Figure 2F).
190
To assess the effect of BafA1 on the pH of early time point
endosomes, AGS cells were labelled with 191
FITC and TMR dextran for 20 minutes and chased for 10 minutes
with or without BafA1 for the entire 192
duration of pulse and chase (Figure S4E). While the total amount
of dextran uptake is not affected 193
significantly, the endosomal FITC intensity and the endosomal
ratio of FITC/TMR, which can be 194
considered as a proxy for endosomal pH, show a robust increase
with BafA1 treatment (Figure S4F). 195
This indicates that BafA1 also affects the endosomal pH of early
time point endosomes. 196
Chloroquine treatment does not affect RBD uptake and minimally
alters endosomal pH 197
Chloroquine, a diprotic weak base, is expected to accumulate in
acidic compartments and neutralize 198
lysosomal pH 58. While, mounting evidence shows that Chloroquine
and its analogs can inhibit the 199
infection by several viruses such as Ebola, Dengue, Chikungunya,
HIV, etc 59, many studies point 200
towards differences between the mode of action of Chloroquine
and acidification inhibitors – BafA1 201
and NH4Cl 60,61. We, therefore, tested the effect of Chloroquine
on RBD, dextran and transferrin uptake 202
to verify if it behaves like BafA1. We found that upon treatment
with Chloroquine, the uptake of neither 203
RBD nor transferrin was altered significantly (Figure S5A, S5B).
Dextran uptake was marginally higher 204
upon treatment with Chloroquine (Figure S5B). 205
The effect of Chloroquine in changing the endosomal pH of late
endosomes using FITC/TMR ratio as 206
a proxy for endosomal pH was assessed. At different
concentrations of Chloroquine tested, the 207
endosomal pH was only minimally increased (Figure S5C, S5D). We
also observed that both FITC and 208
TMR endosomal intensities increased with the concentration of
Chloroquine. To confirm our results, 209
we used another method to estimate endosomal pH. FITC has a
pH-sensitive (488nm) and a pH-210
insensitive excitation (450nm) 54. We used the 488/458
excitation ratio of FITC dextran as a readout of 211
pH and found that this ratio also showed only a small albeit
significant increase with Chloroquine when 212
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
7
compared to control cells, unlike the increase brought about by
NH4Cl (Figure S5E, S5F). This 213
observation could explain the lack of an endocytic effect on
RBD, dextran or transferrin uptake upon 214
treatment with Chloroquine. 215
Designing Spike-pseudovirus transduction assay to specifically
address the effects of inhibitors on viral 216
entry 217
To ascertain that the observations made using the RBD, as a
proxy for viral entry, are valid in the context 218
of a viral entity decorated by the SARS-CoV2 Spike protein
itself, we generated SARS-CoV2 Spike-219
pseudotyped lentiviral particles (Spike-pseudovirus), following
a previously established methodology 220 62. mCherry fluorescent
protein expression was used as a reporter for assessing viral
infection (Figure 221
S6A, Methods). The expression of the Spike protein in the
pseudovirus particles was verified by a 222
western blot using the antibody against the C-terminal Strep-tag
of Spike, which revealed bands 223
corresponding to both the S2 fragment as well as the full-length
protein (Figure S6B). The infection 224
specificity of the pseudovirus was validated by infection of
human (HEK-293T) versus mouse (NIH-225
3T3) cells, where the latter showed lower infectivity,
consistent with a lack of a bonafide hACE2 226
receptor to bind the Spike protein (Figure S6C). Independently,
a competition experiment was 227
conducted to check the effect of excess soluble RBD on
transduction of Spike-pseudovirus in HEK-228
293T (Methods). The transduction efficiency was reduced in the
presence of soluble RBD, indicating 229
that the Spike-pseudovirus competes for the same binding sites
as RBD (Figure S6D). However, the 230
inhibition was not complete, possibly since even high
concentration of free RBD in the solution cannot 231
compete with the high effective concentrations on
Spike-pseudoviruses, augmented even more by their 232
trimeric configuration that facilitates multi-valent
interactions 63. 233
Since our experiments were aimed at understanding the entry
mechanism of Spike-pseudovirus, we 234
designed the transduction assays with shorter pseudovirus
incubation time and followed the infection 235
efficiency by tracing reporter gene expression at a later time
point. We characterized the transduction 236
efficiency of the pseudovirus as a function of its MOI and time
of incubation to obtain an optimum MOI 237
and incubation time (Figure S6E). Transduction efficiency
measured across the tested regime suggested 238
4 or 8 hours of incubation at 0.5 MOI to be optimal to achieve
at least >1000 positive cells (per well of 239
a 96-well assay plate) with reasonably low viral load and
incubation time. 240
Spike-pseudovirus transduction is reduced by endosomal
acidification inhibitors and Chloroquine 241
If the pseudovirus expressing the full-length spike mimicked the
same trafficking pathway for entry as 242
RBD, we reasoned that the transduction efficiency would be
reduced upon treatment with inhibitors 243
affecting RBD uptake. SARS-CoV2 Spike-pseudovirus transduction
efficiency has been reported to be 244
reduced upon treatment with NH4Cl, BafA1 and Chloroquine 24,64.
However, we wanted to specifically 245
explore the actions of these inhibitors at the initial stages of
infection. Therefore, to address this, we 246
pre-treated cells with drugs for an hour followed by the
addition of Spike-pseudovirus in the presence 247
of the drug for 2, 4 or 8 hours. Both virus and drug were
removed thereafter, and cells were incubated 248
with fresh media either in the absence or presence of a minimal
concentration of the drug as indicated 249
(Figure 3A, Methods). This design was chosen to reduce long-term
toxicity of the inhibitors to the cells 250
and minimize any secondary effects on the translational
processes of the reporter gene post entry. 251
Infection, or transduction efficiency, is reported as the
normalized percentage of transduction compared 252
to corresponding control and cell viability is measured in terms
of nuclei number normalized to control. 253
We tested the effect of NH4Cl and BafA1 on the Spike-pseudovirus
transduction assay in AGS cells 254
upon treatment with the 20mM NH4Cl or 50nM BafA1. At the end of
designated time points, the 255
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
8
pseudovirus containing media along with the NH4Cl and BafA1 was
removed and replenished with 256
fresh media alone. We observed a significant reduction of
Spike-pseudovirus transduction with NH4Cl 257
and BafA1 compared to the corresponding controls with no
significant difference in cell viability at the 258
end of both 4 and 8 hours (Figure 3B-3E). NH4Cl also shows
robust reduction even upon 2 hours of 259
incubation (Figure 3E). HEK-293T cells also exhibited a similar
inhibition of transduction upon 260
treatment with 20mM NH4Cl or 50nM BafA1 (Figure S10 E-G).
261
Although Chloroquine did not alter RBD uptake or increase the pH
of the endocytic compartments 262
significantly, we observed a marked reduction of viral
transduction with 50µM Chloroquine treatment 263
in AGS cells (Figure 3D, 3E) and 10µM Chloroquine treatment in
HEK-293T cells (Figure S10E-G), 264
where the drug was removed along with the Spike-pseudovirus at
the end of 8 hours of incubation. 265
Upon testing whether long term incubation of Chloroquine (as
used in this assay), results in changes in 266
endosomal pH or RBD uptake, we observed no alterations in the
assessed phenotypes in AGS cells 267
(Figure S5G, S5H). This suggests a distinct pH-independent
mechanism of intervention by 268
Chloroquine, functioning at the initial stages of infection.
269
RBD uptake is reduced upon treatment with AN96 and ML141, albeit
to a lesser extent compared to 270
the effect of BafA1 and NH4Cl. Therefore, we assessed the effect
of AN96 and ML141 on Spike-271
pseudovirus transduction in AGS cells, using the 2 hours and 8
hours format of the assay, respectively. 272
The AN96 concentration was reduced to a non-toxic level of 1µM
after removal of the virus. With this 273
experimental design, 5 different concentrations of AN-96 were
tested, and we observed no reduction in 274
normalized percentage transduction even at the highest
concentration of 25µM (Fig. 3F and G), with 275
no compromise on cell viability. This was consistent with the
observation that RBD is rerouted and 276
associates more with transferrin. In case of ML141, we treated
the cells with 5µM of the drug and 277
observed no difference in normalized percentage transduction
compared to the control (Figure S6F, 278
S6G). Partial inhibition of uptake may not strongly manifest in
our pseudovirus assay, as the read-out 279
is all or none and is not sensitive to the number of virus
particles entering the cells. Our findings suggest 280
that inhibitors that affect both RBD uptake and neutralize
acidic endosomes could be one of the 281
strategies used to impede Spike-pseudovirus transduction.
282
Identifying FDA-approved drugs functioning similar to BafA1 and
NH4Cl 283
Armed with the knowledge on the mode of action of acidification
inhibitors in reducing the uptake of 284
RBD, increasing the pH of endosomes and abrogating the infection
of Spike-pseuodovirus, we screened 285
a small subset of FDA-approved drugs with the potential to alter
the pH of endosomes (Figure 4A). We 286
selected a panel of 6 drugs which includes those acting on
Na+/K+ ATPase (Omeprazole, 287
Esomeprazole, Pantoprazole, SCH-28080, Lansoprazole) and a
protonophore that disrupts proton 288
gradient (Niclosamide). We developed a quantitative high
throughput screening pipeline for testing 289
these drugs in both endocytic assay as well as pH estimation
assay in AGS cells. The screen was carried 290
out at a concentration of 10µM for all drugs. 291
Of the 6 drugs tested in the endocytosis assay, Niclosamide
showed the strongest effect on the uptake 292
of the 3 probes (RBD, dextran and transferrin) similar to what
we observed for acidification inhibitors. 293
Niclosamide treated cells showed reduced RBD and dextran uptake
and increased transferrin uptake 294
(Figure 4B, 4C). It is interesting to note that while the other
proton pump inhibitors had minimal effects 295
on RBD or dextran uptake at the concentration tested, Omeprazole
and Pantoprazole showed a 296
significant increase in transferrin uptake (Figure 4C, S7A).
This suggests that these two drugs could 297
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
9
specifically act on the transferrin containing endosomes and not
in the compartments of relevance for 298
RBD and dextran uptake, while Niclosamide inhibits the RBD and
dextran uptake. 299
Of the 6 drugs tested in the late endosomal pH estimation assay,
Niclosamide also showed the strongest 300
neutralization effect on the pH of acidic endosomes (Figure 4E)
by increasing the endosomal ratio of 301
FITC/TMR (Figure S7C). The other drugs had minimal effects on
the pH of late endosomes at the 302
concentration tested (Figure 5E, S7B). The spatial pH maps of
Niclosamide treated cells show an 303
increase in pH in the majority of endosomes within the cell
(Figure 5D). Niclosamide increased the 304
FITC endosomal intensity and reduced the numbers of endosomes
(Figure S7C) similar to the effect of 305
BafA1 on these endosomal trafficking parameters. 306
Omeprazole and other proton pump inhibitors are prodrugs which
are used for treating Gastro-307
esophageal reflux disease (GERD) 65. They are activated by low
pH, bind covalently to H+/K+ ATPase 308
and inhibit the enzymatic function 66. We tested the hypothesis
if these drugs could also similarly block 309
the proton pumps in the late endosomes and thus increase the
endosomal pH 67,68. Earlier studies have 310
indicated that Omeprazole 69, Lansoprazole 70, and Pantoprazole
71, neutralize the endosomal pH only 311
when used at very high concentrations (> 1mM) in EMT-6 and
MCF-7 cells. However, the plasma 312
concentration of these proton pump inhibitors varies between
1–23µM 65. Thus, at least in the 313
concentration range of relevance, we find no effect of these
drugs on the acidification of endosomes 314
and the uptake of RBD. 315
Niclosamide functions as an acidification and entry inhibitor
316
Niclosamide is an anti-helminthic FDA-approved drug and has been
in use since the 1960s (Ditzel, 317
1967). Many recent studies show that Niclosamide has broader
clinical applications and has also been 318
identified as an antiviral against SARS-CoV, human Rhinovirus,
Influenza viral, Dengue virus 72,73. As 319
Niclosamide emerged as a potential drug candidate in both the
RBD endocytosis as well as endosomal 320
pH neutralization screens, we investigated the dose-dependent
role of Niclosamide in reducing RBD 321
uptake, neutralizing endosomal pH and inhibiting
Spike-pseudovirus infection. We found that 322
Niclosamide reduced both RBD and dextran uptake, as well as
increase transferrin uptake in a dose-323
dependent manner (1–25µM) (Figure 5A, S8A). We observed
Niclosamide’s effect on RBD 324
endocytosis even at concentrations as low as 1µM. On analyzing
the effect of Niclosamide on 325
endosomal numbers and intensity, we found that Niclosamide
increased the endosomal intensity of 326
transferrin endosomes and reduced the number of RBD and dextran
endosomes (Figure S8B). These 327
effects are remarkably similar to the effects observed with
acidification inhibitors – BafA1 and NH4Cl. 328
We also confirmed the inhibitory effect of Niclosamide on RBD
and dextran uptake in another cell line 329
– HEK-293T (Figure S10A-D), and on normalized RBD uptake in AGS
cells overexpressing ACE2 330
(Figure S3C, S3D). 331
Further, we also observed a dose-dependent effect of Niclosamide
on neutralizing the pH of late 332
endosomes, with neutralization effects seen even at 2.5µM
(Figure 5B, 5C). The dose-response effect 333
is seen on the ratio of endosomal FITC/TMR as well as other
endosomal trafficking parameters - FITC 334
and TMR endosomal intensities and numbers of endosomes (Figure
S8C). The spatial pH maps of cells 335
also show a gradual shift of endosomal pH from acidic to neutral
pH with different doses of Niclosamide 336
(Figure 5B), especially at 2.5µM wherein some endosomes within
the cell are still acidic while some 337
others are neutralized. Towards evaluating the effect of
Niclosamide on the pH of early time point 338
endosomes, AGS cells were labelled with FITC and TMR dextran for
20 minutes and chased for 10 339
minutes with or without Niclosamide for the entire duration of
pulse and chase (Figure S4E). Unlike 340
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
10
BafA1, while Niclosamide reduced the net uptake of dextran,
similar to BafA1, Niclosamide also 341
increased the endosomal FITC intensity and endosomal FITC/TMR
ratio of early time point (30 342
minutes) endosomes (Figure S4F), indicating that Niclosamide
neutralizes the pH of these endosomes 343
as well. 344
We assessed the effect of different concentrations (0.1-10 µM)
of Niclosamide on Spike-pseudovirus 345
entry in AGS cells, using the experimental strategy designed to
assess virus entry as described before. 346
At the end of 4 and 8 hours of viral incubation, the pseudovirus
containing media along with the 347
Niclosamide was removed and all treatments were replenished with
media containing a reduced 348
concentration of 0.1µM Niclosamide. We observed a strong
dose-dependent reduction of transduction 349
efficiency as a function of increasing Niclosamide concentration
at both viral incubation durations with 350
negligible toxicity (Figure 5D, S9Ai and Aii). IC50 of ~1.27µM
was estimated on fitting a sigmoidal 351
function to the dataset obtained for 8 hours of viral incubation
(Figure 5E, Methods). Together, all three 352
assays conclude that Niclosamide can act as acidification and
entry inhibitor. 353
Enhancing the inhibition of infectivity: A combination strategy
with Niclosamide and 354
Hydroxychloroquine 355
A combinatorial approach of drugs with varying mechanisms of
inhibition works as an effective therapy 356
to combat infection 74. Given that Niclosamide exhibits a short
half-life 75, has poor bio-availability 357
(~10 %) 76 and our observations indicate moderate IC50 for
inhibition of Spike-pseudovirus transduction, 358
we tested if the action of Niclosamide can be enhanced in the
presence of another FDA-approved drug 359
known to be effective against SARS-CoV2 infection. Since
published reports and commonly practiced 360
treatments against SARS-CoV2 infection employ Hydroxychloroquine
(HCQ), a less toxic variant of 361
Chloroquine 77, we tested the effect of HCQ on altering late
endosomal pH and Spike-pseudovirus 362
transduction assay. Like Chloroquine, cells treated with 50µM
HCQ also minimally altered the late 363
endosomal pH (Figure S9B, S9C). However, we observed the
pseudovirus transduction to be markedly 364
reduced at HCQ concentrations of 50µM and 25µM (Figure S9D, S9E)
and only modestly reduced at 365
the concentrations of 10µM or lower in a dose-dependent manner
(See the first box plot in Figures S9 366
Fi, Fii and Fiii). To assess the synergistic effect of the two
drugs, we chose a concentration range with 367
the maximum concentrations of 10µM HCQ and 5µM Niclosamide. The
2-dimensional dose-response 368
map shown in Figure 5G summarizes the effect of the two drugs on
transduction. We observed an 369
augmented reduction in infection when HCQ was used at a
concentration of 10µM along with varying 370
concentrations of Niclosamide compared to where HCQ was used at
0, 2 and 5µM (Figure S9Fi, Fii 371
and Fiii). These results indicate an additive effect on
inhibition of pseudovirus transduction when 372
effective concentrations of HCQ is added along with effective
concentrations of Niclosamide (Figure 373
5F, 5G). Thus, Niclosamide could potentially enhance the
efficacy of the plethora of treatments 374
currently being used to combat SARS-CoV2 infection. 375
Discussion: 376
Understanding the molecular mechanisms of viral entry into
target cells is critical to design effective 377
treatments and prevention strategies against infection.
Employing various methodologies, we report for 378
the first time that fluorescently labelled RBD of SARS-CoV2
enters cells through a pH-dependent CG 379
pathway. High-resolution quantitative imaging approaches enabled
us to detect the localization of RBD 380
to acidic compartments. Endosomal acidification inhibitors that
affect the uptake of CG cargo also 381
inhibit RBD uptake. Complementing our observations with RBD, we
show that infection by Spike-382
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
11
pseudovirus is also dependent on endosomal acidification.
Further, by employing a targeted drug screen, 383
we have identified Niclosamide as a potential inhibitor against
SARS-CoV2 entry (Figure 6). 384
The choice of viral entry into host cells is influenced by cell
surface interacting partners and co-factors 385 11,25. Although
ACE2 has been identified as the receptor for SARS-CoV2, other
receptors are being 386
uncovered. These include Neuropilin 12,13, CD147 14, Heparan
Sulphate proteoglycans 15 and HDL 387
scavenger receptors 16. Additionally, the highly glycosylated
nature of Spike protein could also confer 388
the ability to interact with yet unidentified receptors. These
virus-receptor interactions could potentially 389
dictate the endocytic route employed by the virus. This is
exemplified by our observation that although 390
RBD uptake is reduced upon blocking the CG pathway, residual RBD
re-routes towards the CME and 391
enables pseudovirus infection. Re-routing could presumably be
due to binding to different receptors 392
that could follow alternative internalization routes. Whether
the Spike-pseudovirus follows routes of 393
entry like RBD, can be addressed with tractable pseudoviruses or
synthetic virus-like particles. 394
However, recent genome wide screens 38,78 indicating the
importance of cholesterol homeostasis in 395
SARS-CoV2 infection are consistent with a cholesterol senstitive
CG endocytic route 79 in entry. 396
Knockouts of genes affecting cholesterol biosynthesis (SCAP,
MBTSP1, MBTSP2) not only reduced 397
infection of native SARS-CoV2 but also of Spike-pseuodviruses
indicating that cellular cholesterol is 398
necessary for efficient Spike mediated entry of SARS-CoV2 78.
399
Known inhibitors of endosomal acidification, BafilomycinA1 and
NH4Cl, play an important role in 400
neutralizing acidic lysosomes and thus subverting viral membrane
fusion and entry of several viruses 401 11,24–27. Here, we report
that these inhibitors also play a more upstream role by inhibiting
the endocytosis 402
of RBD itself. Both these treatments inhibited the uptake of CG
cargo and RBD, reduced Spike-403
pseudovirus infection and drastically elevated endosomal pH. It
is interesting to note that the inhibition 404
of acidification in addition to dramatically reducing CG uptake
did not cause re-trafficking of RBD 405
through another endocytic pathway, as was observed for other CG
inhibitors. This suggests that the 406
acidification inhibitors could negatively influence the
RBD-receptor interactions at the cell surface 407
along with further ramifications of blocking the CG pathway.
408
These observations encouraged us to screen a subset of
FDA-approved compounds known to affect 409
endosomal acidification: proton-pump inhibitors (Omeprazole,
Lansoprazole, Pantoprazole, 410
Esomeprazole, SCH-28080), and protonophore (Niclosamide). Of all
the 6 compounds tested only 411
Niclosamide inhibited CG cargo and RBD uptake, elevated
endosomal pH and concomitantly inhibited 412
Spike-pseudovirus infection, all in a dose-dependent manner with
an IC50 of 1.27 µM in AGS cells. 413
Among several mechanisms of action 75, Niclosamide disrupts
proton gradient across mitochondrial 80 414
and endosomal 72 membranes. The elevated endosomal pH brought
about by Niclosamide was shown 415
to inhibit human rhinovirus infection 72. Additionally,
Niclosamide has been identified as an anti-viral 416
agent against SARS 81, Dengue 73, MERS 82 and more recently
proposed for SARS-CoV2 (with IC50 of 417
0.28 µM in Vero cells) 83, although the mechanism of action
remained unknown. In contrast, the proton 418
pump inhibitors used in our study failed to interfere with RBD
uptake. This could be because they 419
remained inactive 65 or the concentrations tested predominantly
affect H+/K+ ATPases, while mM 420
concentrations are required to inhibit V-ATPases 67. Along these
lines, studies show that proton pump 421
inhibitors inhibit Ebola-pseudovirus 84, SARS-CoV and SARS-CoV2
85 infection only when used 422
beyond achievable plasma concentrations 65. 423
Surprisingly, Chloroquine did not affect RBD uptake and only
marginally raised the endosomal pH. 424
However, it caused a strong inhibition of Spike-pseudovirus
infection. This strongly suggests that 425
Chloroquine could be operating in the initial steps of viral
infection but post endocytosis 64, as observed 426
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
12
with RBD uptake. Chloroquine is likely to function in many
pH-independent ways to inhibit SARS-427
CoV2 infections, distinct from Niclosamide. For example, by
altering terminal glycosylation of ACE2 428 86; via its activity as
a zinc ionophore affecting ACE2 activation 87,88; by interacting
with ER resident 429
Sigma receptors that initiates cell stress response 89; by its
ability to strongly bind a viral protease 430
essential for Spike activation 90. At this time, the exact
mechanism(s) by which Chloroquine inhibits 431
SARS-CoV2 entry remains unclear. 432
In conclusion, our study reports the high capacity CG pathway as
a potential endocytic route for SARS-433
CoV2. We further show that endosomal acidification is critical
for SARS-CoV2 entry and infection and 434
can be a promising therapeutic target as observed by the results
seen with Niclosamide, BafilomycinA1 435
and NH4Cl. This study also paves way for large-scale screens to
repurpose FDA-approved drugs as 436
acidification inhibitors and scrutinize for more
Niclosamide-like drugs that might have better 437
bioavailability or can be used in combination with other
antiviral drugs. Moreover, the methods 438
described in our study can effectively be extended and better
represented with clinical isolates of viruses 439
to assess their infective journey in primary cells that
represent the more natural hosts for infection. 440
Materials and Methods: 441
Cell lines, constructs, and antibodies: See supplementary
methods for more details 442
Chemicals and reagents: 443
Niclosamide and AN96 were chemically synthesized and proton pump
inhibitors, Esomeprazole and 444
Pantoprazole, were extracted from commercially available tablets
as detailed in the Supplementary 445
Methods. The other proton pump inhibitors, Lansoprazole and
SCH-28080, were obtained from the 446
LOPAC®1280 library, and Omeprazole was procured from Sigma
(O104). 447
Endocytosis assays: 448
AGS or HEK-293T cells were plated in 35mm coverslip bottom
dishes and processed after 48 hours at 449
60-70% confluency. Cells were washed twice with HEPES buffer
(wash and imaging buffer 450
composition: 150mM NaCl, 20mM HEPES, 5mM KCl, 1mM CaCl2, 1mM
MgCl2, 2mg/ml Glucose, 451
pH 7.5) at 37C. Endocytosis was monitored using fluorescently
labelled RBD (Alexa/Atto 488, 452
10g/ml), 10kDa TMR-dextran (1mg/ml) and/or Iron-loaded
Transferrin (10g/ml, Alexa 647) in 453
serum-free medium for indicated time points at 37C. Endocytosis
was stopped using ice-cold wash 454
buffer and cells were subsequently fixed with 2.5%
paraformaldehyde (PFA) for 20 minutes at room 455
temperature (RT). Cells were then washed and imaged. For
inhibitor experiments, cells were pre-treated 456
with various inhibitors (AN96 25M, ML141 50M, Amiloride 1mM,
BafA1 200nM or 400nM, NH4Cl 457
30mM) and respective controls in serum-free medium for 30
minutes at 37C and inhibitors were 458
maintained during endocytic assays. 459
To measure normalized transferrin or normalized RBD uptake
(Figures 1H-1I, S3C-S3D), cell surface-460
bound probes after the endocytic pulse with transferrin or RBD
were stripped using two washes with 461
ice-cold ascorbate buffer (160mM sodium ascorbate, 40mM ascorbic
acid, 1mM MgCl2, 1mM CaCl2, 462
pH 4.5), followed by three washes with ice-cold wash buffer at
4 °C. Cells were then fixed with ice-463
cold 2.5% PFA for 5 mins at 4 °C and 15 minutes at RT.
Transferrin receptor (TfR) was labelled by 464
incubating cells with fluorescently labelled anti-hTfR (OKT-9)
for 2 hours at RT. To label surface 465
ACE2, fixed cells were blocked with 10mg/ml bovine serum albumin
(30 minutes) followed by 466
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
13
incubation with anti-myc primary antibody (1 hour) and secondary
antibody (45 minutes) in blocking 467
buffer at RT. Cells were then washed and imaged. 468
pH estimation assays: 469
For estimating the pH of late endosomes, cells were pulsed with
pH-sensitive 10kDa FITC-dextran 470
(1mg/ml) and pH-insensitive 10kDa TMR-dextran (1mg/ml) for 2
hours in serum-free media, chased 471
for 1 hour in the presence of inhibitors or control and imaged
live. The above pulse and chase times 472
were chosen to allow the accumulation of labelled dextran in
acidic late endosomal and lysosomal 473
compartments (co-labelled with Lysotracker, data not shown). To
estimate the endosomal pH, the ratio 474
of FITC to TMR fluorescence was computed and compared to a pH
calibration curve (Figures S4A-475
S4B) which was generated by equalizing the endosomal pH to that
of an external buffer. After the pulse 476
with FITC and TMR-dextran and chase, cells were incubated with
5µg/ml nigericin containing buffers 477
of different pH for 10 minutes and imaged to evaluate FITC/TMR
ratios for each pH. 478
For estimating the pH of late endosomes using the 488/458
excitation ratio of FITC-dextran (Figures 479
S5E-S5F), cells were pulsed with FITC-dextran at 1mg/ml for 2
hours, followed by chase in the 480
presence or absence of inhibitors and imaged live. 481
For estimating the FITC/TMR ratio of early endosomes (Figures
S4E-S4F), cells were incubated with 482
pH-sensitive 10kDa FITC-dextran (1mg/ml) and pH-insensitive
10kDa TMR-dextran (1mg/ml) for 20 483
minutes, chased for 10 minutes and imaged live. Throughout the
pulse and chase duration, the cells 484
were incubated in serum-free media with control (0.2%DMSO) or
BafA1 400nM or Niclosamide 485
10µM. 486
Spike-pseudovirus transduction assays: 487
AGS/HEK-293T cells were plated in optical bottom 96-well plates.
36 hours post-plating, when cell 488
numbers were ~4000, transduction was carried out at indicated
MOIs. For inhibitor treatment, cells 489
were pre-incubated with indicated concentrations of NH4Cl/
BafA1/ CQ/ Niclosamide/ HCQ/ AN96/ 490
ML141, for 1 hour. This was followed by addition of the
Spike-pseudoviruses in presence or absence 491
of the inhibitors. At the end of 2/4/8hours, media containing
pseudoviruses and inhibitors was removed, 492
and cells were washed once with drug-free media. This was
followed by addition of media with or 493
without inhibitor: NH4Cl, BafA1 and CQ were removed from the
media; Niclosamide, AN96 and HCQ 494
were maintained at a low concentration of 100nM, 1µM and 500nM
respectively. This was done to 495
assess the effects of the inhibitors at the initial stages of
inhibition, minimize long-term toxicity to the 496
cells as well as to avoid effects on the translational processes
of the reporter gene post entry. After 60 497
hours, cells were fixed, nuclei were labelled with Hoescht and
assessed for transduction efficiency based 498
on mCherry reporter expression. In the case of HEK-293T cells
(Figure S10G), MTT cell viability assay 499
was performed to check toxicity (assay described in
Supplementary Methods). 500
Imaging and Analysis: 501
a. Endocytic and pH estimation assays 502
For 35mm dish-based endocytic experiments, fixed samples were
imaged using confocal microscopy 503
(Olympus FV3000, 20X/0.85NA objective) to image RBD, dextran and
transferrin endosomes with Z 504
sections of 1µm. Maximum intensity projected images were used
for further analysis. Cell ROIs were 505
drawn and features such as cell mean intensity in each channel
was extracted. 506
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
14
For high-throughput endocytic and pH estimation experiments,
automated imaging (Spinning disc, 507
Phenix Perkin Elmer, 40XW/1.1NA objective) was used to image
nucleus along with RBD, dextran 508
and transferrin (for endocytosis) or FITC and TMR dextran (for
pH) with Z sections of 1µm each. For 509
both assays, cell profiler based pipeline was used to segment
cells, nucleus and endosomes and extract 510
features as described in supplementary methods. For pH
calibration, the mean of the endosomal ratio 511
distributions at different extracellular pH was fit to a
sigmoidal equation. For both assays, custom 512
MATLAB routines were used to estimate the endosomal intensities,
the number of endosomes and cell 513
mean intensities. In addition, for pH assays, endosomal ratio
(FITC/TMR) and endosomal pH (using 514
the calibration curve) for each endosome was computed. As the
endosomal intensity distribution within 515
cells is a heavy right-tailed distribution, median endosomal
intensity for each probe for each cell was 516
estimated. The distributions of cell mean intensity/endosomal
intensities/numbers of endosomes per 517
cell per treatment (for endocytosis) and endosomal
intensities/ratio/pH per cell per treatment (for pH) 518
is represented in each quantification. 519
For 488/458 endosomal ratio estimation experiments, live imaging
was done using confocal microscopy 520
(Zeiss LSM 780, 40X/1.4NA objective). Excitation lasers 488nm,
458nm were used and emission was 521
detected using a spectral detector (490nm-560nm). Images were
processed as described above to 522
estimate endosomal intensities and endosomal ratios per cell.
523
b. Colocalization analysis - Confocal microscopy (Olympus
FV3000, 60X/1.42NA objective) with Z 524
sections of 0.4µm each was employed to image cells across all
channels. A MATLAB routine was 525
written to extract colocalization indices. For each cell,
endosomes in each channel were segmented 526
based on threshold values. The segmentation in each channel was
made finer using morphological 527
operations (dilation followed by erosion). Segmented endosomes
were considered for colocalization 528
analysis. Manders’ coefficients and Pearson’s correlation
coefficients were computed as described 529
before 91. 530
c. Pseudovirus transduction assays - Automated imaging
(Widefield, Phenix, 10X/0.3NA objective) of 531
96 well assay plates was used to image nucleus as well as
mCherry positive cells. A cell profiler based 532
pipeline was used to segment nucleus and extract features, as
described in supplementary methods. 533
Approximately 50,000 nuclei (cells) were scored for each
treatment. A MATLAB routine was written 534
to estimate the % transduction. Mean intensities of the
segmented nucleus in the nuclei channel and the 535
mCherry channel for each nucleus across all fields were
extracted. Each assay plate included “No-536
Virus” negative control. This control was used to estimate the
background intensities of the mCherry 537
channel within each segmented nucleus. The median of this
distribution was considered the background. 538
All nuclei with mCherry intensities of at least 1.8 - 2.2 times
(empirically determined) the intensity of 539
the background were considered positive. For each field, the
fraction of positive nuclei to the total 540
number of nuclei was determined. The mean of % transduction
across all fields for each treatment was 541
calculated. The % transduction was normalized to that of the
control and is represented in all the 542
quantifications. The total number of nuclei for each treatment
is also represented to understand the 543
effect of the toxicity of drugs. 544
Statistical methods and hypothesis testing: 545
All statistical tests between control and treatment were
performed in MATLAB using Wilcoxon rank-546
sum test and the p-value of the hypothesis testing and the
number of repeats is indicated in figure 547
legends. In the entire manuscript, ***, **, * and ns indicate
p-value of Wilcoxon rank-sum test < 0.001, 548
-
15
magnitude of differences as well as the sample size. For large
sample size, as in case of our high 550
throughput experiments, the impact of random error in
measurement will be reduced and the larger 551
magnitude of difference between the control and treatment will
be associated with a much smaller p-552
value. 553
References: 554
1. Zhu, N. et al. A Novel Coronavirus from Patients with
Pneumonia in China, 2019. N. Engl. J. 555
Med. 382, 727–733 (2020). 556
2. Wang, X. & Guan, Y. COVID‐19 drug repurposing: A review
of computational screening 557
methods, clinical trials, and protein interaction assays. Med.
Res. Rev. med.21728 (2020) 558
doi:10.1002/med.21728. 559
3. Magro, G. COVID-19: Review on latest available drugs and
therapies against SARS-CoV-2. 560
Coagulation and inflammation cross-talking. Virus Research vol.
286 198070 (2020). 561
4. Chan, D. C. & Kim, P. S. HIV entry and its inhibition.
Cell vol. 93 681–684 (1998). 562
5. Danieli, T., Pelletier, S. L., Henis, Y. I. & White, J.
M. Membrane fusion mediated by the 563
influenza virus hemagglutinin requires the concerted action of
at least three hemagglutinin 564
trimers. J. Cell Biol. 133, 559–569 (1996). 565
6. Shirato, K., Kanou, K., Kawase, M. & Matsuyama, S.
Clinical Isolates of Human Coronavirus 566
229E Bypass the Endosome for Cell Entry. (2016)
doi:10.1128/JVI.01387-16. 567
7. Hulswit, R. J. G. et al. Human coronaviruses OC43 and HKU1
bind to 9-O-acetylated sialic 568
acids via a conserved receptor-binding site in spike protein
domain A. Proc. Natl. Acad. Sci. U. 569
S. A. 116, 2681–2690 (2019). 570
8. Raj, V. S. et al. Dipeptidyl peptidase 4 is a functional
receptor for the emerging human 571
coronavirus-EMC. Nature 495, 251–254 (2013). 572
9. Hofmann, H. et al. Human coronavirus NL63 employs the severe
acute respiratory syndrome 573
coronavirus receptor for cellular entry. Proc. Natl. Acad. Sci.
102, 7988–7993 (2005). 574
10. Li, W. et al. Angiotensin-converting enzyme 2 is a
functional receptor for the SARS coronavirus. 575
Nature 426, 450–454 (2003). 576
11. Hoffmann, M. et al. SARS-CoV-2 Cell Entry Depends on ACE2
and TMPRSS2 and Is Blocked 577
by a Clinically Proven Protease Inhibitor. Cell 181, 271-280.e8
(2020). 578
12. Cantuti-Castelvetri, L. et al. Neuropilin-1 facilitates
SARS-CoV-2 cell entry and infectivity. 579
Science (80-. ). 370, eabd2985 (2020). 580
13. Daly, J. L. et al. Neuropilin-1 is a host factor for
SARS-CoV-2 infection. Science (80-. ). 370, 581
eabd3072 (2020). 582
14. Wang, K. et al. SARS-CoV-2 invades host cells via a novel
route: CD147-spike protein. (2020) 583
doi:10.1101/2020.03.14.988345. 584
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
16
15. Clausen, T. M. et al. SARS-CoV-2 Infection Depends on
Cellular Heparan Sulfate and ACE2. 585
Cell 183, 1043-1057.e15 (2020). 586
16. Wei, C. et al. HDL-scavenger receptor B type 1 facilitates
SARS-CoV-2 entry. Nat. Metab. 587
(2020) doi:10.1038/s42255-020-00324-0. 588
17. Brix, K. Host cell proteases: Cathepsins. in Activation of
Viruses by Host Proteases 249–276 589
(Springer International Publishing, 2018).
doi:10.1007/978-3-319-75474-1_10. 590
18. Laporte, M. & Naesens, L. Airway proteases: an emerging
drug target for influenza and other 591
respiratory virus infections. Current Opinion in Virology vol.
24 16–24 (2017). 592
19. Glowacka, I. et al. Evidence that TMPRSS2 Activates the
Severe Acute Respiratory Syndrome 593
Coronavirus Spike Protein for Membrane Fusion and Reduces Viral
Control by the Humoral 594
Immune Response. J. Virol. 85, 4122–4134 (2011). 595
20. Simmons, G. et al. Characterization of severe acute
respiratory syndrome-associated 596
coronavirus (SARS-CoV) spike glycoprotein-mediated viral entry.
Proc. Natl. Acad. Sci. U. S. 597
A. 101, 4240–4245 (2004). 598
21. Simmons, G. et al. Inhibitors of cathepsin L prevent severe
acute respiratory syndrome 599
coronavirus entry. Proc. Natl. Acad. Sci. 102, 11876–11881
(2005). 600
22. Huang, I. C. et al. SARS coronavirus, but not human
coronavirus NL63, utilizes cathepsin L to 601
infect ACE2-expressing cells. J. Biol. Chem. 281, 3198–3203
(2006). 602
23. Turk, V. et al. Cysteine cathepsins: From structure,
function and regulation to new frontiers. 603
Biochimica et Biophysica Acta - Proteins and Proteomics vol.
1824 68–88 (2012). 604
24. Ou, X. et al. Characterization of spike glycoprotein of
SARS-CoV-2 on virus entry and its 605
immune cross-reactivity with SARS-CoV. Nat. Commun. 11, (2020).
606
25. Dittmar, M. et al. Drug repurposing screens reveal FDA
approved drugs active against SARS-607
Cov-2. (2020) doi:10.1101/2020.06.19.161042. 608
26. Wang, H. et al. SARS coronavirus entry into host cells
through a novel clathrin- and caveolae-609
independent endocytic pathway. Cell Res. 18, 290–301 (2008).
610
27. Qian, Z., Dominguez, S. R. & Holmes, K. V. Role of the
Spike Glycoprotein of Human Middle 611
East Respiratory Syndrome Coronavirus (MERS-CoV) in Virus Entry
and Syncytia Formation. 612
PLoS One 8, 76469 (2013). 613
28. Riva, L. et al. A Large-scale Drug Repositioning Survey for
SARS-CoV-2 Antivirals. bioRxiv 614
Prepr. Serv. Biol. (2020) doi:10.1101/2020.04.16.044016. 615
29. Kang, Y. L. et al. Inhibition of PIKfyve kinase prevents
infection by Zaire ebolavirus and SARS-616
CoV-2. Proc. Natl. Acad. Sci. U. S. A. 117, 20803–20813 (2020).
617
30. Thottacherry, J. J., Sathe, M., Prabhakara, C. & Mayor,
S. Spoiled for Choice: Diverse Endocytic 618
Pathways Function at the Cell Surface. Annu. Rev. Cell Dev.
Biol. 35, 55–84 (2019). 619
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
17
31. Elkin, S. R., Lakoduk, A. M. & Schmid, S. L. Endocytic
pathways and endosomal trafficking: a 620
primer. Wiener Medizinische Wochenschrift vol. 166 196–204
(2016). 621
32. Sabharanjak, S., Sharma, P., Parton, R. G. & Mayor, S.
GPI-anchored proteins are delivered to 622
recycling endosomes via a distinct cdc42-regulated
clathrin-independent pinocytic pathway. 623
Dev. Cell 2, 411–423 (2002). 624
33. Forgac, M. Structure, function and regulation of the
vacuolar (H + )-ATPases. FEBS Lett. 440, 625
258–263 (1998). 626
34. Gupta, G. D. et al. Population distribution analyses reveal
a hierarchy of molecular players 627
underlying parallel endocytic pathways. PLoS One 9, (2014).
628
35. Dey, G. et al. Exploiting cell-to-cell variability to detect
cellular perturbations. PLoS One 9, 629
(2014). 630
36. Kalia, M. et al. Arf6-independent GPI-anchored
protein-enriched early endosomal 631
compartments fuse with sorting endosomes via a
Rab5/phosphatidylinositol-3′-kinase- 632
dependent machinery. Mol. Biol. Cell 17, 3689–3704 (2006).
633
37. Sathe, M. et al. Small GTPases and BAR domain proteins
regulate branched actin 634
polymerisation for clathrin and dynamin-independent endocytosis.
Nat. Commun. 9, 1835 635
(2018). 636
38. Daniloski, Z. et al. Identification of Required Host Factors
for SARS-CoV-2 Infection in Human 637
Cells. Cell 0, (2020). 638
39. Nonnenmacher, M. & Weber, T. Adeno-associated virus 2
infection requires endocytosis 639
through the CLIC/GEEC pathway. Cell Host Microbe 10, 563–576
(2011). 640
40. Tai, W. et al. Characterization of the receptor-binding
domain (RBD) of 2019 novel coronavirus: 641
implication for development of RBD protein as a viral attachment
inhibitor and vaccine. Cell. 642
Mol. Immunol. 17, 613–620 (2020). 643
41. Stadlbauer, D. et al. SARS‐CoV‐2 Seroconversion in Humans: A
Detailed Protocol for a 644
Serological Assay, Antigen Production, and Test Setup. Curr.
Protoc. Microbiol. 57, (2020). 645
42. Sathe, M. et al. Small GTPases and BAR domain proteins
regulate branched actin 646
polymerisation for clathrin and dynamin-independent endocytosis.
Nat. Commun. 9, (2018). 647
43. Chadda, R. et al. Cholesterol-sensitive Cdc42 activation
regulates actin polymerization for 648
endocytosis via the GEEC pathway. Traffic 8, 702–17 (2007).
649
44. Kumari, S. & Mayor, S. ARF1 is directly involved in
dynamin-independent endocytosis. Nat. 650
Cell Biol. 10, 30–41 (2008). 651
45. Gupta, G. D. et al. Analysis of endocytic pathways in
Drosophila cells reveals a conserved role 652
for GBF1 in internalization via GEECs. PLoS One 4, e6768 (2009).
653
46. Boal, F. et al. LG186: An Inhibitor of GBF1 Function that
Causes Golgi Disassembly in Human 654
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
18
and Canine Cells. Traffic 11, 1537–1551 (2010). 655
47. Thottacherry, J. J. et al. Mechanochemical feedback control
of dynamin independent 656
endocytosis modulates membrane tension in adherent cells. Nat.
Commun. 9, 1–14 (2018). 657
48. Berthiaume, E. P., Medina, C. & Swanson, J. A. Molecular
size-fractionation during endocytosis 658
in macrophages. J. Cell Biol. 129, 989–998 (1995). 659
49. Mercer, J. & Helenius, A. Virus entry by
macropinocytosis. Nature Cell Biology vol. 11 510–660
520 (2009). 661
50. Masereel, B., Pochet, L. & Laeckmann, D. An overview of
inhibitors of Na+/H+ exchanger. 662
European Journal of Medicinal Chemistry vol. 38 547–554 (2003).
663
51. Bayer, N. et al. Effect of Bafilomycin A1 and Nocodazole on
Endocytic Transport in HeLa 664
Cells: Implications for Viral Uncoating and Infection. J. Virol.
72, 9645–9655 (1998). 665
52. Bowman, B. J., McCall, M. E., Baertsch, R. & Bowman, E.
J. A model for the proteolipid ring 666
and bafilomycin/concanamycin-binding site in the vacuolar ATPase
of Neurospora crassa. J. 667
Biol. Chem. 281, 31885–31893 (2006). 668
53. Presley, J. F., Mayor, S., McGraw, T. E., Dunn, K. W. &
Maxfield, F. R. Bafilomycin A1 669
treatment retards transferrin receptor recycling more than bulk
membrane recycling. J. Biol. 670
Chem. 272, 13929–13936 (1997). 671
54. Ohkuma, S. & Poole, B. Fluorescence probe measurement of
the intralysosomal pH in living 672
cells and the perturbation of pH by various agents. Proc. Natl.
Acad. Sci. 75, 3327–3331 (1978). 673
55. Mayor, S., Presley, J. F. & Maxfield, F. R. Sorting of
membrane components from endosomes 674
and subsequent recycling to the cell surface occurs by a bulk
flow process. J. Cell Biol. 121, 675
1257–1269 (1993). 676
56. Dunn, K. W., McGraw, T. E. & Maxfield, F. R. Iterative
fractionation of recycling receptors 677
from lysosomally destined ligands in an early sorting endosome.
J. Cell Biol. 109, 3303–3314 678
(1989). 679
57. Murphy, R. F., Powers, S. & Cantor, C. R. Endosome pH
measured in single cells by dual 680
fluorescence flow cytometry: rapid acidification of insulin to
pH 6. J. Cell Biol. 98, 1757–1762 681
(1984). 682
58. Homewood, C. A., Warhurst, D. C., Peters, W. & Baggaley,
V. C. Lysosomes, pH and the anti-683
malarial action of chloroquine [9]. Nature vol. 235 50–52
(1972). 684
59. Li, C. et al. Chloroquine, a FDA-approved Drug, Prevents
Zika Virus Infection and its 685
Associated Congenital Microcephaly in Mice. EBioMedicine 24,
189–194 (2017). 686
60. Mauthe, M. et al. Chloroquine inhibits autophagic flux by
decreasing autophagosome-lysosome 687
fusion. Autophagy 14, 1435–1455 (2018). 688
61. Lu, S., Sung, T., Lin, N., Abraham, R. T. & Jessen, B.
A. Lysosomal adaptation: How cells 689
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
19
respond to lysosomotropic compounds. PLoS One 12, e0173771
(2017). 690
62. Crawford, K. H. D. et al. Protocol and reagents for
pseudotyping lentiviral particles with SARS-691
CoV-2 spike protein for neutralization assays. Viruses 12,
(2020). 692
63. Yan, R. et al. Structural basis for the recognition of
SARS-CoV-2 by full-length human ACE2. 693
Science (80-. ). 367, 1444–1448 (2020). 694
64. Wang, M. et al. Remdesivir and chloroquine effectively
inhibit the recently emerged novel 695
coronavirus (2019-nCoV) in vitro. Cell Research vol. 30 269–271
(2020). 696
65. Shin, J. M. & Kim, N. Pharmacokinetics and
pharmacodynamics of the proton pump inhibitors. 697
Journal of Neurogastroenterology and Motility vol. 19 25–35
(2013). 698
66. Shin, J. M., Cho, Y. M. & Sachs, G. Chemistry of
covalent inhibition of the gastric (H+, K +)-699
ATPase by proton pump inhibitors. J. Am. Chem. Soc. 126,
7800–7811 (2004). 700
67. Sabolic, I., Brown, D., Verbavatz, J. M. & Kleinman, J.
H+-ATPases of renal cortical and 701
medullary endosomes are differentially sensitive to Sch-28080
and omeprazole. Am. J. Physiol. 702
- Ren. Fluid Electrolyte Physiol. 266, (1994). 703
68. De Milito, A., Lucia Marino, M. & Fais, S. A Rationale
for the Use of Proton Pump Inhibitors 704
as Antineoplastic Agents. Curr. Pharm. Des. 18, 1395–1406
(2012). 705
69. Lee, C. M. & Tannock, I. F. Inhibition of endosomal
sequestration of basic anticancer drugs: 706
Influence on cytotoxicity and tissue penetration. Br. J. Cancer
94, 863–869 (2006). 707
70. Yu, M., Lee, C., Wang, M. & Tannock, I. F. Influence of
the proton pump inhibitor lansoprazole 708
on distribution and activity of doxorubicin in solid tumors.
Cancer Sci. 106, 1438–1447 (2015). 709
71. Patel, K. J., Lee, C., Tan, Q. & Tannock, I. F. Use of
the Proton Pump Inhibitor Pantoprazole to 710
Modify the Distribution and Activity of Doxorubicin: A Potential
Strategy to Improve the 711
Therapy of Solid Tumors. (2013)
doi:10.1158/1078-0432.CCR-13-0128. 712
72. Jurgeit, A. et al. Niclosamide Is a Proton Carrier and
Targets Acidic Endosomes with Broad 713
Antiviral Effects. PLoS Pathog. 8, (2012). 714
73. Jung, E. et al. Neutralization of Acidic Intracellular
Vesicles by Niclosamide Inhibits Multiple 715
Steps of the Dengue Virus Life Cycle In Vitro. Sci. Rep. 9, 1–12
(2019). 716
74. Hung, I. F. N. et al. Triple combination of interferon
beta-1b, lopinavir–ritonavir, and ribavirin 717
in the treatment of patients admitted to hospital with COVID-19:
an open-label, randomised, 718
phase 2 trial. Lancet 395, 1695–1704 (2020). 719
75. Chen, W., Mook, R. A., Premont, R. T. & Wang, J.
Niclosamide: Beyond an antihelminthic 720
drug. Cellular Signalling vol. 41 89–96 (2018). 721
76. Lin, C. K. et al. Preclinical evaluation of a nanoformulated
antihelminthic, niclosamide, in 722
ovarian cancer. Oncotarget 7, 8993–9006 (2016). 723
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
20
77. Liu, J. et al. Hydroxychloroquine, a less toxic derivative
of chloroquine, is effective in inhibiting 724
SARS-CoV-2 infection in vitro. Cell Discov. 6, 6–9 (2020).
725
78. Wang, R. et al. Genetic screens identify host factors for
SARS-CoV-2 and common cold 726
coronaviruses. Cell 0, (2020). 727
79. Chadda, R. et al. Cholesterol-sensitive Cdc42 activation
regulates actin polymerization for 728
endocytosis via the GEEC pathway. Traffic 8, 702–717 (2007).
729
80. Park, S. J., Shin, J. H., Kang, H., Hwang, J. J. & Cho,
D. H. Niclosamide induces mitochondria 730
fragmentation and promotes both apoptotic and autophagic cell
death. BMB Rep. 44, 517–522 731
(2011). 732
81. Wu, C. J. et al. Inhibition of severe acute respiratory
syndrome coronavirus replication by 733
niclosamide. Antimicrob. Agents Chemother. 48, 2693–2696 (2004).
734
82. Ko, M. et al. Screening of FDA-approved drugs using a
MERS-CoV clinical isolate from South 735
Korea identifies potential therapeutic options for COVID-19.
(2020) 736
doi:10.1101/2020.02.25.965582. 737
83. Jeon, S. et al. Identification of antiviral drug candidates
against SARS-CoV-2 from FDA-738
approved drugs. Antimicrob. Agents Chemother. (2020)
doi:10.1101/2020.03.20.999730. 739
84. Barclay, W., Long, J., Wright, E., Molesti, E. &
Temperton, N. Antiviral therapies against Ebola 740
and other emerging viral diseases using existing medicines that
block virus entry. 741
F1000Research 4, 1–11 (2015). 742
85. Bojkova, D. et al. SARS-CoV-2 and SARS-CoV differ in their
cell tropism and drug sensitivity 743
profiles. 1–28 (2020) doi:10.1101/2020.04.03.024257. 744
86. Vincent, M. J. et al. Chloroquine is a potent inhibitor of
SARS coronavirus infection and spread. 745
Virol. J. 2, 69 (2005). 746
87. Turner, A. J., Hiscox, J. A. & Hooper, N. M. ACE2: From
vasopeptidase to SARS virus receptor. 747
Trends in Pharmacological Sciences vol. 25 291–294 (2004).
748
88. Hecel, A. et al. Zinc(II)—The Overlooked Éminence Grise of
Chloroquine’s Fight against 749
COVID-19? Pharmaceuticals 13, 228 (2020). 750
89. Gordon, D. E. et al. A SARS-CoV-2 protein interaction map
reveals targets for drug 751
repurposing. Nature 583, 459–468 (2020). 752
90. Braz, H. L. B. et al. In silico study of azithromycin,
chloroquine and hydroxychloroquine and 753
their potential mechanisms of action against SARS-CoV-2
infection. Int. J. Antimicrob. Agents 754
56, 106119 (2020). 755
91. Aaron, J. S., Taylor, A. B. & Chew, T.-L. Image
co-localization-co-occurrence versus 756
correlation. (2018) doi:10.1242/jcs.211847. 757
758
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
21
Acknowledgements: 759
We thank Nevan Krogan (UCSF, USA) for Spike expression
construct, Florian Krammer (Mt. Sinai, 760
USA) for secreted RBD expression construct, Biocon, Ltd (India)
and Raghavan Varadharajan (IISc, 761
India) for providing purified RBD for initial experiments,
Minhaj Sirajuddin (inSTEM, India) for 762
mCherry expression plasmid, Vinoth Kumar (inSTEM, India) for
discussions on Spike-pseudovirus 763
characterization and Mylan Laboratories (India) for providing
HCQ. We thank the Central Imaging and 764
Flow Cytometry Facility (CIFF) and the Screening Facility at
NCBS, for imaging and high content 765
screening. We sincerely thank Abrar Bhat (for help with RBD
purification), Chandrima Patra (for help 766
with imaging), Greeshma Pradeep (for help with microscopes),
Sarayu Beri (for help with cell culture) 767
and all members of SM lab for exciting discussions on this
project. We acknowledge the technical, 768
administrative and hospitality staff at NCBS for tremendous help
especially during the national 769
lockdown. We acknowledge NCBS-TIFR graduate fellowship (for CP
and PS), UGC graduate 770
fellowship (for RG) and NCBS postdoctoral fellowship (for SJ).
TSvZ acknowledges EMBO 771
postdoctoral fellowship (ALTF 1519-2013) and NCBS Campus
fellowship; AC thanks India Alliance 772
DBT – Wellcome Trust Early career fellowship (IA/E/15/1/502339);
SM acknowledges J.C. Bose 773
Fellowship from DST, Government of India, and India Alliance DBT
– Wellcome Trust Margdarshi 774
fellowship (IA/M/15/1/502018). CP, PS and SJ received support
from Margadarshi fellowship grant 775
(IA/M/15/1/502018). 776
Author Contributions: 777
CP, RG, PS, AG, VS and SM conceived the study. CP, RG, PS, SJ,
AG, VS and SM designed the 778
experiments. Endocytosis and pH assays were done by RG and CP,
Spike-pseudovirus assays were 779
done by PS and SJ. CP, TSvZ and DS acquired and analyzed data.
CP, RG, PS, SJ, TSvZ, DS, AG, VS 780
and SM interpreted the data. RG, NS, SML, DS and CP conducted
high-throughput experiments. AC 781
and PS prepared Spike-pseudovirus. VKN and SBB synthesized
Niclosamide. RA, AHN, PPS and RV 782
synthesized AN96. TPP and PV extracted Esomeprazole and
Pantoprazole. CP, SD, BM and PS purified 783
and labelled RBD. CP, RG, PS, SJ, TSvZ and SM wrote the
manuscript with comments from AC, AG 784
and VS. 785
preprint (which was not certified by peer review) is the
author/funder. All rights reserved. No reuse allowed without
permission. The copyright holder for thisthis version posted
December 16, 2020. ; https://doi.org/10.1101/2020.12.16.422529doi:
bioRxiv preprint
https://doi.org/10.1101/2020.12.16.422529
-
22
Main Figure Legends: 1
Figure 1: RBD uptake is sensitive to CG Pathway inhibitors in
AGS cells 2
A: Schematic describing endocytic pathways at the plasma
membrane with specific cargoes for each 3
endocytic pathway: transferrin (CME Cargo) and 10kDa dextran (CG
Cargo). AN96, ML141 and 4
BafA1 specifically affect the uptake of CG cargoes. 5
B, C: AGS cells were pulsed with RBD, dextran and transferrin
for 30 minutes and imaged at high 6
resolution after fixation. Images are shown in B and
quantification of Manders’ co-occurrence 7
coefficient is shown in C. This compares the fraction of RBD
endosomal intensity with transferrin or 8
dextran (p-value < e-06), transferrin endosomal intensity
with dextran or RBD (p-value < e-05) and 9
dextran endosomal intensity with transferrin or RBD (p-value =
0.18). RBD is more co-localized to 10
dextran endosomes. Number of cells = 10. White arrow represents
endosomes containing RBD, dextran 11
and transferrin. Yellow arrow represents endosomes with RBD and
dextran without transferrin. Dashed 12
white line in B represents the approximate cell boundary. 13
D, E: AGS cells were pretreated with Control (0.6% DMSO) or AN96
25µM for 30 minutes and pulsed 14
with RBD, dextran and transferrin for 30 minutes with or without
the inhibitor. Treatment with AN96 15
reduces RBD (p-value < e-19) and dextran (p < e-44) uptake
while minimally alters transferrin uptake 16
(p = 0.02). Images are shown in D and quantification in E.
Numbers of cells > 100 for each treatment. 17
F, G: AGS cells were treated with Control (0.2% DMSO) or BafA1
200nM for 30 minutes and pulsed 18
with RBD and dextran for 30 minutes with or without the
inhibitor. Treatment with BafA1 reduces 19
RBD (p-value < e-33) and dextran (p-value < e-18) uptake.
Images are shown in F and quantification 20
in G. Numbers of cells > 100 for each treatment. 21
H, I: AGS cells were treated with Control (0.2% DMSO) or BafA1
200nM for 30 minutes and pulsed 22
with RBD and transferrin for 30 minutes with or without the
inhibitor. The surface transferrin receptor 23
(TfR) was labelled after fixation. Treatment with BafA1 reduces
RBD uptake (p-value < e-27) and 24
increased normalized transferrin uptake (p-value < e-03).
Images are shown in H and quantification in 25
I. Numbers of cells > 80 for each treatment. 26
Data (E, G, I) is represented as a scatter with box plot. Black
dots represent per-cell data points. Box 27
plot represents the distribution (25% to 75% percentile) with
the red line indicating the median and red 28
dot indicating the mean of the distribution. Whiskers represent
distribution up to 1.5 times interquartile 29
range and + indicates outliers beyond the whiskers. In the
entire manuscript, ***, **, * and ns indicate 30
p-value of Wilcoxon rank-sum test < 0.001,
-
23
more Lysotracker positive compartments have RBD (p-value <
e-06). Each condition has >12 cells. 39
Dashed white line in A represents approximate cell boundary.
40
C: Schematic describing the experimental protocol for estimating
the pH of endosomes by ratiometric 41
measurements using pH-sensitive (FITC) and pH-insensitive (TMR)
dextran. 42
D-F: AGS cells were pulsed with FITC and TMR dextran for 2
hours, chased for 1 hour with BafA1 43
200nM/400nM, NH4Cl 30mM or control and imaged live. Endosomal pH
is increased upon addition of 44
acidification inhibitors (p-values < e-118 for BafA1 200nM,
< e-122 for BafA1 400nM, < e-223 for 45
NH4Cl). Images along with pH maps are shown in D (and in S4C)
and quantification in E (and in S4D). 46
Enlarged regions of pH maps indicated by white boxes are shown
in F. Box plot in E represents the 47
distribution of medians of each repeat which is denoted by red
dots. Violin plot indicates all the data 48
points from repeats. Colour bar in F corresponds to the
estimated endosomal pH. Control1 is 0.2% 49
DMSO, Control2 is 0.4% DMSO and Control3 is 0% DMSO. Number of
repeats ≥ 3 for each treatment 50
and each repeat has >80 cells. 51
Scale bar: 20µm (A) and 40µm (D). 52
Figure 3: BafA1, NH4Cl and Chloroquine affect Spike-pseudovirus
infection 53
A: Schematic describing the experimental protocol for SARS-CoV2
Spike-pseudovirus transduction 54
assay. 55
(B-I): AGS cells were pre-incubated with the inhibitors (BafA1
50nM, NH4Cl 20mM, CQ 50µM and 56