10018–10033 Nucleic Acids Research, 2021, Vol. 49, No. 17 Published online 20 August 2021
https://doi.org/10.1093/nar/gkab731
Cis regulation within a cluster of viral microRNAs
Monika Vilimova1,†, Maud Contrant1,†, Ramy Randrianjafy1, Philippe Dumas2,
Endrit Elbasani3, Pa¨ivi M. Ojala3, Se´bastien Pfeffer 1,*,† and Aure´lie Fender 1,*,†
1Universite´ de Strasbourg, Architecture et Re´activite´ de l’ARN, Institut de Biologie Mole´culaire et Cellulaire du CNRS,
2 alle´e Konrad Roentgen, 67084 Strasbourg, France, 2Institut de Ge´ne´tique et Biologie Mole´culaire et Cellulaire
(IGBMC), Department of Integrated structural Biology, 1 rue Laurent Fries, BP10142, 67404 Illkirch-Graffenstaden,
France and 3Translational Cancer Medicine Research Program, P.O. Box 63 (Haartmaninkatu 8), FIN-00014
University of Helsinki, Finland
Received November 19, 2020; Revised August 06, 2021; Editorial Decision August 09, 2021; Accepted August 10, 2021
ABSTRACT
MicroRNAs (miRNAs) are small regulatory RNAs in-
volved in virtually all biological processes. Although
many of them are co-expressed from clusters, little is
known regarding the impact of this organization on
the regulation of their accumulation. In this study, we
set to decipher a regulatory mechanism controlling
the expression of the ten clustered pre-miRNAs from
Kaposi’s sarcoma associated herpesvirus (KSHV).
We measured in vitro the efficiency of cleavage of
each individual pre-miRNA by the Microprocessor
and found that pre-miR-K1 and -K3 were the most ef-
ficiently cleaved pre-miRNAs. A mutational analysis
showed that, in addition to producing mature miR-
NAs, they are also important for the optimal expres-
sion of the whole set of miRNAs. We showed that this
feature depends on the presence of a canonical pre-
miRNA at this location since we could functionally
replace pre-miR-K1 by a heterologous pre-miRNA.
Further in vitro processing analysis suggests that
the two stem-loops act in cis and that the cluster is
cleaved in a sequential manner. Finally, we exploited
this characteristic of the cluster to inhibit the expres-
sion of the whole set of miRNAs by targeting the pre-
miR-K1 with LNA-based antisense oligonucleotides
in cells either expressing a synthetic construct or
latently infected with KSHV.
INTRODUCTION
Kaposi’s sarcoma herpes virus (KSHV) or Human herpes
virus 8 is a gammaherpesvirus associated with cancers such
as Kaposi’s sarcoma, B-lymphomas or the proliferative dis-
order Castelman disease. Its genome is a ∼165 kb dsDNA
molecule that encodes >90 open reading frames (ORFs) as
well as 25 mature microRNAs (miRNAs) (1). KSHV estab-
lishes lifelong persistent infection with a restricted expres-
sion of viral genes. However, a small percentage (<3%) of
cells support lytic replication and under certain conditions
KSHV can reactivate from latency to lytic replication. A dy-
namic balance between the latent and lytic phases of KSHV
replication is critical to establish a successful virus infection,
maintain latency, and is involved in pathogenic effects such
as tumorigenesis (reviewed in (2)).
Interestingly, all KSHV precursor (pre-)miRNAs are ex-
pressed on the same polycistronic transcript, which is as-
sociated with latency (3–5). Ten of them (pre-miR-K1 to
-K9 and miR-K11) are clustered within an intron of ∼4 kb
between ORF71 (v-FLIP) and the kaposin genes, and are
expressed under the control of a latent promoter (6). Pre-
miR-K10 and pre-miR-K12 localize within the ORF and
the 3′UTR of the kaposin gene, respectively, and are con-
trolled by both latent and lytic promoters (6,7).
KSHV miRNAs are able to regulate the expression of
both viral and cellular genes that are essential to virus infec-
tion and associated diseases. Abundantly expressed during
the latent phase, they directly participate in its maintenance,
for instance by repressing directly, or indirectly through tar-
geting of NF-kB pathway, the replication and transcrip-
tion activator (RTA), which is crucial for viral reactivation
(8–10). They also promote tumorigenesis by modulating
apoptosis, angiogenesis or cell cycle (e.g. (11–14)). Finally,
KSHV miRNAs also enhance immune evasion and viral
pathogenesis by regulating host immune responses (e.g. (15–
18). See also (19) for a recent review on KSHV miRNA
functions.
*To whom correspondence should be addressed. Tel: +33 3 88 41 70 60; Fax: +33 3 88 60 22 18; Email: s.pfeffer@ibmc-cnrs.unistra.fr
Correspondence may also be addressed to Aure´lie Fender. Email: a.fender@ibmc-cnrs.unistra.fr
†The authors wish it to be known that, in their opinion, the first two and last two authors should be regarded as Joint First and Joint Last Authors.
Present addresses:
Maud Contrant, French Agency for Food, Environmental and Occupational Health & Safety, Laboratory of Ploufragan-Plouzane´-Niort, Unit of Viral Genetics and
Biosafety, Ploufragan, France.
Endrit Elbasani, Orion Corporation, Orion Pharma, Tengstro¨minkatu 8, 20360 Turku, Finland.
C© The Author(s) 2021. Published by Oxford University Press on behalf of Nucleic Acids Research.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License
(http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work
is properly cited. For commercial re-use, please contact journals.permissions@oup.com
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10019
Although we now know of numerous functions of KSHV
miRNAs due to active research in the field, we still do not
have a precise understanding of the regulation of expres-
sion of these key viral factors. In animals, miRNA bio-
genesis is a multi-step process including two maturations
by RNase III enzymes. MiRNA genes are generally tran-
scribed by RNA polymerase II as a long primary transcript
(pri-miRNA) of several kilobases that can contain one or
several miRNA precursor hairpins (pre-miRNA). First, the
pri-miRNA is processed in the nucleus by the Micropro-
cessor, comprising the RNAse III type enzyme Drosha and
its cofactor DGCR8. After export into the cytoplasm, the
pre-miRNA is further processed by another RNase III en-
zyme, Dicer, associated with TRBP. The final result is a du-
plex of miRNAs (5p and 3p) from which one of the strands
is preferentially incorporated into an Argonaute protein to
form the RISC complex, which can then be directed toward
target mRNAs (reviewed in (20)). Alternative pathways of
miRNA biogenesis exist such as Drosha-independent pro-
cessing of mirtrons or Dicer-independent Ago2-dependent
miR-451 cleavage (21–24).
About 25–40% of human miRNAs are found in clus-
ters (25,26). There are usually two to three miRNAs in
a cluster. However, a few larger clusters were also de-
scribed such as the conserved mammalian pri-miR-17–92
that contains 6 members, or the imprinted C19MC that
contains 46 tandemly-repeated pre-miRNA genes (27–29).
Co-expression may be essential as it was shown that many
clustered miRNAs regulate common biological processes
as is the case for KSHV miRNAs (e.g. (18)). Even though
clustered miRNAs are co-transcribed, the resulting ma-
ture miRNAs are found at different levels in the cell (e.g.
(30,31)). This suggests that complex regulation events occur
downstream of the transcription. Two independent stud-
ies from Zeng and Orom laboratories revealed the key im-
portance of maturation by the Microprocessor to explain
the global level of cellular miRNAs (32,33). Maturation
of pri-miRNAs by the Microprocessor is controlled by se-
quence and structural features of miRNA hairpin, defining
the basal level of pre-miRNAs excised. In addition, protein
cofactors may interact with specific motifs or structure of
the stem-loop and thus modulate the Microprocessor ac-
tivity (see (34–36). Recently, several studies demonstrated
interdependent processing in the context of bicistronic pri-
miRNA where an optimal miRNA hairpin assists the pro-
cessing of a neighboring suboptimal one (37–42). Such in-
terdependency has not yet been documented in the case of
larger miRNA clusters.
Previously, we demonstrated the importance of RNA sec-
ondary structure of the long primary transcript containing
the ten intronic miRNAs from KSHV (pri-miR-K10/12)
for the accumulation of mature miRNAs (31). Here, we
show cis regulation within this large viral miRNA cluster.
We observed that the ten miRNA hairpins from the KSHV
intronic cluster are processed in vitro by the Microproces-
sor with different efficiencies. Intriguingly, high processing
levels of miR-K1 and miR-K3 hairpins were not consistent
with the low level of accumulation of their mature miRNAs
in infected cells (31), suggesting that these miRNA hairpins
could serve other purposes than solely producing mature
miRNAs. Indeed, specific deletion of pre-miR-K1 or pre-
miR-K3 within the cluster significantly reduce the expres-
sion of the remaining miRNAs in the cell. Moreover, only
the pre-miRNA feature is sufficient to support such regu-
latory mechanism since replacement of pre-miR-K1 by the
heterologous pre-Let-7a-1 restores expression of clustered
miRNAs to the wt level. Further experiments of in vitro
processing assays using pri-miRNA fragments (mimicking
cleavage of pre-miR-K1 or pre-miR-K3) suggest that reg-
ulation may occur before pre-miR-K1 and -K3 are cut by
the Microprocessor or that processing of the cluster is se-
quential. Finally, we developed an antisense strategy based
on Locked Nucleic Acid (LNA) oligonucleotides to post-
transcriptionally downregulate the expression of the whole
KSHV miRNA cluster. Using an LNA targeting either the
5p or 3p of the pre-miR-K1 sequence, wemanaged to signif-
icantly reduce the levels of clusteredKSHVmiRNAs in cells
transfected with a synthetic construct. We also showed that
in KSHV-infected cells, the levels of neosynthesized miR-
NAs derived from the cluster dropped significantly upon
LNA targeting of pre-miR-K1, indicating that this could be
a useful strategy to block the entire cluster in infected cells.
MATERIALS AND METHODS
Cells and media
HEK293FT-rKSHV cells were generated by infecting
HEK293FT cells with concentrated rKSHV.219 virus (43)
in the presence of 8g/ml polybrene (Sigma) and by spinoc-
ulation (800 g for 30 min at room temperature). Puromycin
selection was applied to select for rKSHV.219 infected cells.
The virus stock used to infect HEK293FT cells was gen-
erated by treating iSLK.219 cells (44) with 1g/ml doxy-
cycline and 1.35mM sodium butyrate and collecting virus
particles 48 h post-reactivation by ultracentrifugation.
Adherent HEK293Grip and HEK293FT-rKSHV cell-
lines were cultured in a humidified 5% CO2 atmosphere at
37◦C in DMEM medium containing 10% fetal calf serum
(FCS). In addition, HEK293FT-rKSHV were grown with
2,5 g/ml puromycin (for viral genome maintenance).
RNA preparation
Total RNA was extracted from cells using TRIzol reagent
(Invitrogen, Thermo Fisher Scientific).
Pre-miRNAs, wt and mutants of pri-miR-K10/12, de-
rived from BCBL-1 cell line, were transcribed from PCR-
generatedDNA templates carrying a T7 promoter (see Sup-
plementary Table S1 for primers). In vitro RNA synthesis
was done by T7 RNA polymerase (Ambion). Pre-miRNAs
were purified on denaturant polyacrylamide gel and long
pri-miR-K10/12 derived transcripts (up to ∼3 kb) were
salt purified using Monarch® PCR and DNA cleanup kit
(NewEnglandBioLabs). After acidic phenol extraction and
ethanol precipitation, the RNAs were pelleted and recov-
ered in MilliQ water.
Northern blot analysis
RNAs were resolved on a 8% urea-acrylamide gel, trans-
ferred on a nylon membrane (Amersham Hybond-NX,
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
10020 Nucleic Acids Research, 2021, Vol. 49, No. 17
GE-Healthcare Life Sciences), crosslinked to the mem-
brane by chemical treatment at 60◦C using 1-ethyl-3-[3-
dimethylaminopropyl]carbodiimide hydrochloride (EDC)
(Sigma) for 1 h 30 min. MiRNAs and pre-miRNAs were
detected with specific 5′-32P labeled oligonucleotides (Sup-
plementary Table S1). The signals were quantified using a
Fuji Bioimager FLA5100. miR-16 was probed as a loading
control.
In vitro Drosha miRNA processing assays
Drosha and DGCR8 were overexpressed in 10-cm Petri
Dish Hek293Grip cells using pCK-Drosha-Flag and pCK-
Flag-DGCR8. After 48 h, cells were washed with ice-cold
PBS, centrifuged and pellet was resuspended in 120 l ice-
cold lysis buffer (20mMTris–HCl pH 8.0, 100mMKCl, 0.2
mM EDTA, 0.5 mM DTT, 5% glycerol and mini-complete
EDTA-free protease inhibitor (Roche)). The cell suspen-
sion was sonicated during 5 min at high amplitude, 30 s on
and 30 s off using Bioruptor™ UCD-200 (Diagenode), cen-
trifuged for 10 min at 10 000 g, 4◦C, and the supernatant
was used for in vitro processing assays.
500 or 1000 fmol of in vitro transcribed wt or mutant pri-
miR-K10/12 RNAs were denatured 3 min at 95◦C, cooled
on ice 3 min and folded in 1× structure buffer provided
by Ambion during 30 min at 37◦C. Processing assays were
performed in 30 l containing 15 l of total protein ex-
tract (10 g/l) or 15 l of lysis buffer, 6.4 mM MgCl2,
30 U Ribolock (Thermo Scientific™, Thermo Fisher Sci-
entific). Just after addition of 170 l elution buffer (2%
SDS, 0.3 M sodium acetate), reaction was terminated by
acidic phenol extraction followed by ethanol precipitation
with 5 g glycogen. After resuspension in formamide load-
ing buffer, cleavage products were analyzed by northern
blot. For quantification, in vitro transcribed and gel purified
pre-miRNAs and synthetic miRNA oligonucleotides (IDT)
were loaded at increasing concentration (from 1.5 to 25
fmol). A standard curvewas generated by plotting the signal
intensity against the amount of pre-miRNAs loaded and
was used to calculate the absolute amount of pre-miRNAs
produced by in vitro Drosha processing.
Kinetic analysis
The experimental cleavage curves show different rates and
different fractions of cleavage of each pre-miR. In addition,
the experimental cleavage curves most often show a max-
imum followed by a slight decrease of the amount of pre-
miR,which requires to introduce a secondary cleavage event
of the newly formed pre-miR (either from a contaminant
RNase, or from Dicer). The model in use is thus:
pri − miR
kDrosha−−−−→ pri − miR(i )
kRNAse−−−−→ miR(i )? (1)
At first, it was attempted to model the rates of enzy-
matic cleavage (kDrosha and kRNAse) according to Michaelis–
Menten kinetics, but it turned out to be inefficient due to
a very large correlation of the parameters KM and Vmax.
We finally used the following simplest possible mathemat-
ical model:
dKi
dt
= k+i ( fi R0 − Ki − K∗i ) − k−i Ki ;
dK∗i
dt
= k−i Ki (2)
with R0 the total concentration of the pri-miRNA, Ki the
time-dependent concentration of the ith pre-miR, fi the
fraction of R0 used by Drosha to produce Ki , and Ki* the
time-dependent concentration of the secondary-cleavage
product (miR(i) in Equation (1)). Since Drosha act differ-
ently on the pri-miR to produce each pre-miR, it is neces-
sary to replace kDrosha with a particular cleavage rate Ki+
for each pre-miR and, similarly, it is necessary to replace
kRNASe with ki– for each secondary-cleavage rate. This sim-
ple model, therefore, does not try in any way to differentiate
situations wherein a particular pre-miR would be obtained
either from the first cleavage of the full pri-miR, or from
subsequent cleavage of a fragment of it. The previous linear
differential Equation (2) are readily integrated, which gives
the cleaved fraction Yi = Ki(t)/R0 necessary to fit the ith
experimental curve:
Yi = Ki (t)R0 = fi k
+
i
e−k
−
i t − e−k+i t
k+i − k−i
(if k+i = k−i ) (3)
Yi = Ki (t)R0 = fi k
+
i t e
−k+i t (if k+i = k−i ) (3′)
The results obtained with this model are in Figure 1 and
Supplementary Figure S2.
Mutagenesis and miRNA expression in human cells
Plasmid pcDNA-K10/12, derived from pcDNA5 (Invitro-
gen) and containing the wild type (wt) pri-miR-K10/12 (14)
was mutated using the Phusion site-directed mutagenesis
kit (Thermo Scientific) and transformation of Escherichia
coli DH5alpha strain. Positive clones were identified by se-
quencing (GATC Biotech, France). Four deletion mutants
were designed: K1, K3, K7 and K9 where the re-
spective pre-miRNA sequence, was deleted. K1-Let7 cor-
responds to K1 mutant where the pre-Let7a-1 was in-
serted in place of pre-miR-K1 (see Figure 2). Expression
tests were conducted as follow: 2 g of plasmids (wt or
mutated) were used to transfect HEK293Grip cells in 6-
well/plate. Total RNAwas collected after 48 h and miRNA
expression was analyzed by northern blot, using 9 g of to-
tal RNA and standard protocol. miR-16 was probed as a
loading control and used for signal normalization.
Antisense LNA treatments
HEK293Grip cells in 6-well/plate were transfected with 1
g of wt pcDNA-K10/12 in combination with 20 nMLNA
oligonucleotide (see Supplementary Table S1 for sequence).
Total RNA was collected after 48 h and miRNA expres-
sion was analyzed by northern blot, using 10 g of total
RNA and standard protocol. miR-16 was probed as a load-
ing control and used for signal normalization.
HEK293FT-rKSHV were seeded in 12-well/plates.
When approximately 50–60% confluent, they were trans-
fected with 20 nM LNA by using Lipofectamine 2000
reagent (Invitrogen). On the next day, they were detached
by 50 l of 0,05% Trypsin-EDTA (Gibco), resuspended in
fresh medium and half of the cells were transferred into a
new well and allowed to seed. On the next day they were
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10021
transfected again with 20 nM LNA. They were collected
by direct lysis in Trizol reagent (Ambion) one day after the
second transfection.
4sU metabolic labeling, neosynthesized RNA pull-down and
RT-qPCR analysis
The experimental procedure was adapted from the protocol
developed by the Nicassio laboratory (45). One day prior to
LNA transfection, 5 million HEK293FT-rKSHV cells were
seeded in a 10 cm culture dish. 50nMLNAwere transfected
using Lipofectamine 2000 (Invitrogen) in total culture vol-
ume of 10 ml. After 24 h, culture medium was collected, fil-
tered (45 m) and 4sU was added to reach final concentra-
tion of 300 M. The medium was then transferred back to
the cells which were allowed to incorporate the 4sU during
3 h (incubation at 37◦C, 5% CO2, in dark). After that, cells
were detached in ice-cold PBS and collected by centrifuga-
tion. RNAwas extracted by TRIzol reagent (Invitrogen) by
using 3 ml of the reagent per dish.
70 g of total RNA were biotinylated by incubation
with 160 l of EZ-Link HPDP-Biotin (Thermo Scientific,
1 mg/ml in DMF), and biotinylation buffer (final concen-
tration 10 mM Tris pH 7.4, 1 mM EDTA) in total volume
of 490 l during 2 h at 25◦C. Following PCI (Roth) extrac-
tion and isopropanol precipitation, the RNA was washed
with EtOH 75% and dissolved in 80 l of RNase-free wa-
ter. BiotinylatedRNAwas then pulled-down onDynabeads
MyOne Streptavidin T1 (Invitrogen) by using 80l of beads
per condition. The beads were first washed twice in buffer
A (80 l of 100 mM NaOH, 50 mM NaCl), then once in
buffer B (100 mM NaCl) and finally resuspended in 160 l
of buffer C (2 M NaCl, 10 mM Tris pH 7.5, 1 mM EDTA,
0.1% Tween-20). The volume of RNA was increased to 160
l and added to the beads. After 15 min of rotation on a
wheel at room temperature, the beads with captured RNA
were washed three times in 320 l of buffer D (1 M NaCl,
5 mM Tris pH 7.5, 0.5 mM EDTA, 0.05% Tween-20). The
RNA was then eluted with 160 l of elution buffer (10 mM
EDTA in 95% formamide) by heating at 65◦C for 10 min.
Trizol-LS (Invitrogen) and chloroform were used for eluted
RNA extraction and after addition of 1.5 V of EtOH 100%
to the aqueous phase, RNA was recovered on miRNeasy
columns (Qiagen) in final volume of 30 l.
3 l of purified RNA were reverse-transcribed by using
TaqMan™ MicroRNA Reverse Transcription Kit (Applied
Biosystems) and a pool of eight specific stem-loop primers
(miR-K1, miR-K3, miR-K4, miR-K11, miR-16, let-7a1,
miR-92, U48, 0.5 l each). RT reaction was then diluted
twice and 1 l used to perform qPCR in total volume of
10 l, by using TaqMan™ Universal Master Mix II, no
UNG (Applied Biosystems) and 0.5 l of individual Taq-
Man miRNA assays (Applied Biosystem). qPCR was real-
ized on CFX96 Touch Real-Time PCR Detection System
(Biorad). Analysis of inputRNAwas performed in the same
way on 100 ng of total RNA. In order to determine the
amount of neosynthesized miRNAs relative to the input,
we first calculated the enrichment of miRNA levels in the
pull-down relative to the input after normalizing the data
to Let-7. This ratio was then compared between the specific
treatment (LNA @K1*) and the control treatment (LNA
Ctrl), which was arbitrarily set to 1.
Primary transcript pri-miR-K10/12 was analyzed after
prior treatment with DNase I (Invitrogen) or TURBO™
DNase (Invitrogen). 5 l of purified or 1 g of input RNA
was treated and subsequently reverse-transcribed (using
1
4 of reaction volume for non-reverse-transcribed control)
with Superscript IV (ThermoFisher Scientific) according to
the manufacturer protocol. cDNA was diluted twice be-
fore using 1 l for quantitative PCRMaxima SYBRGreen
qPCRMasterMix (Thermo Scientific). The same approach
than formaturemiRNAswas used to determine the amount
of neosynthesized pri-miRNAs except that the data were
normalized to the CYC1 mRNA instead of Let-7.
RT-qPCR analysis of the primary transcript in
HEK293Grip cells transfected with LNA was performed
according to the same procedure, however by diluting the
cDNA 10 times.
miRNA expression in HEK293FT-rKSHV cells trans-
fected with LNAs without metabolic labeling was measured
similarly to the 4sU-samples, except for the reverse tran-
scription step, which was performed individually for each
RT stem-loop primer (0.5 l) and without diluting the re-
sulting cDNA. 100 ng of total RNA was used for each RT
reaction.
Primer sequences are indicated in the Supplementary Ta-
ble S1.
RESULTS
Clustered KSHV pre-miRNAs are processed in vitro by the
Microprocessor with different efficiencies
In this work, we focused on the polycistronic feature of the
KSHV intronic pri-miRNA containing ten miRNA hair-
pins (miR-K1 to miR-K9, and miR-K11), that we referred
to as pri-miR-K10/12 (Supplementary Figure S1). Our pre-
vious results showed that pri-miR-K10/12 adopts a well-
organized 2D structure, composed ofmultiple hairpins with
allmiRNAsequences found in stem-loops. Interestingly, the
secondary structural features of miRNA stem-loops corre-
late to some extent with the cellular abundance of mature
miRNAs. Indeed, optimally folded stem-loops tend to lead
to more abundant miRNAs. Moreover, we demonstrated
that the structural context of miRNA hairpins within the
primary transcript is important since swapping miRNA
stem-loops or expressing miRNAs individually results in
differential miRNA accumulation in cells (31).
To further understand the mechanism behind the regula-
tion of expression of polycistronic KSHV miRNAs, we as-
sessed the in vitro processing efficiency of the different pre-
miRNAs within the cluster. To do so, we took advantage
of in vitro processing assays using total extracts obtained
from cells over-expressing Drosha and DGCR8 and in vitro
transcribed pri-miR-K10/12 (∼3.2 kb) (Figure 1A). Accu-
mulation levels of all pre-miRNAs from the cluster were an-
alyzed at different time points using quantitative northern
blot analysis (Figure 1B). Figure 1 and Supplementary Fig-
ure S2 show the results obtained from two independent ex-
periments (Exp#1 and Exp#2).
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
10022 Nucleic Acids Research, 2021, Vol. 49, No. 17
Figure 1. Kinetic analysis of KSHV clustered pre-miRNAs maturation in vitro by the Microprocessor. (A) Overview of in vitro processing assays starting
with synthesis ofDNA template containing aT7 promoter by PCR from the pcDNA-K10/12 plasmid and in vitro transcription to generate pri-miR-K10/12
containing the 10 KSHV clustered pre-miRNAs. In vitro processing assays was performed by incubating pri-miR-K10/12 with total protein extract of
HEK293Grip cells overproducing Drosha and DGCR8. Pre-miRNAs production was monitored along the time and quantified by northern blot analysis.
(B) Northern blot analysis of the time course of in vitro processing assays using 500 fmol of in vitro transcribed pri-miR-K10/12 and HEK293Grip cells
total protein extract where Drosha and DGCR8 were overexpressed. In vitro transcribed pre-miRNAs and synthetic RNA oligonucleotides were loaded
at increasing concentration as standards. (C) Cleavage curves were obtained after plotting pre-miRNA product, in percentage of initial pri-miR-K10/12
substrate, according to time. The fits were obtained with the model involving three free parameters per curve (compare with Supplementary Figure S3 for
the more stringent model with two free parameters per curve). (D) Processing efficiencies (left panel) and cleavage rate (right panel) were plotted in respect
to miRNA hairpins showing variation among the clustered pre-miRNAs. The error bars come from standard procedures used to fit the experimental
curves by minimizing the residuals between the experimental points and their theoretical estimates. Data are from Exp#1 (see Supplementary Figure S2
for Exp#2).
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10023
In the conditions used, all pre-miRNAs were produced
from the unique pri-miR-K10/12 substrate. Pre-miRNAs
were the major products, and, in some cases, we also ob-
served accumulation of mature miRNAs due to residual
Dicer activity in the total protein extract. Experimental
cleavage curves were fitted using a simple kineticmodel with
three free parameters per curve (Figure 1C). This led to
excellent agreement compared to a more stringent model
with only two free parameters per curve (see Materials and
Methods and Supplemental Method). The numerical re-
sults are shown in Supplementary Table S2. Interestingly,
we noticed that the sum of all pre-miRNAs is (114 ± 8) %
and (110± 17) % for Exp#1 and Exp#2, respectively, which
is 10-fold lower that the maximum possible value of 1000
% if pri-miRNA gave rise to all ten possible pre-miRNAs.
Since these two sums are not significantly different from
100 %, this suggests that each pri-miR-K10/12 would be
cleaved only once and would produce only one particular
pre-miRNA. However, we set our experimental conditions
such that the initial concentration of substrate will be high
enough to ensure that we measured the maximum rate of
cleavage for each miRNA stem-loop. In that respect, the
Microprocessor was probably saturated by the full-length
pri-miR-K10/12 at the disadvantage of cleaved RNA frag-
ments. As a consequence, we overlook the possibility to ob-
serve several rounds of cleavage.
Comparison of processing efficiencies fi and of the cleav-
age rate constant k+i for each KSHV pre-miRNA showed
important differences between pre-miRNAs as illustrated
in Figure 1C and Supplementary Figure S2C. Indeed, av-
erage accumulation levels vary from 1.4 % for pre-miR-
K2 up to 36 % for pre-miR-K3 products (Table 1); cleav-
age rate constants ranged from 0.027 min−1 for pre-miR-
K7 to 0.23 min−1 for pre-miR-K8 in Exp#1 (Figure 1 and
Supplementary Table S2). Kinetic parameters were not av-
eraged due to a higher activity of the Microprocessor in
Exp#1 versus Exp#2. Drosha/DGCR8 concentration was
estimated to be about three-fold greater in Exp#2 based
on the relative values of cleavage rates in the two experi-
ments. The agreement between the two experiments is rather
good for the processing efficiencies fi (correlation coeffi-
cient = 0.95) and lower for the kinetic constants of cleav-
age k+i (correlation coefficient = 0.52) (Supplementary Fig-
ure S4). We can nevertheless conclude that there are signif-
icant differences of processing efficiencies between the dif-
ferent KSHV pre-miRNAs. To exclude that the observed
differences might have arisen from various stability of the
different pre-miRNAs and not their processing efficiencies,
we assessed the stability of two pre-miRNAs that appeared
well-processed in our experiment (pre-miR-K1 and -K8)
and two pre-miRNAs that are less well-processed (pre-miR-
K11 and -K7) (Supplementary Figure S5). Briefly, in vitro
transcribed pre-miRNAs were incubated in total cellular
extract from HEK293Grip cells and pre-miRNAs stability
was measured at various time points by northern blot anal-
ysis. Overall, we did not observe any striking difference in
the decay rate between them, which could account for the
variation measured in our in vitro cleavage assays.
We previously determined hairpin optimality features for
each KSHV pre-miRNA (31), which we decided to up-
date in order to take into account novel features that have
Table 1. Correlation between KSHV miRNA hairpin processing effi-
ciency with their hairpin optimality and cellular abundance of their cor-
responding mature miRNAs
KSHV
miRNAs
Hairpin
optimalitya
Processing
efficiencies
fi (%)
Cellular
abundance
in BCBL-1
(%)b Correlation/comment
K3 + 36.0 ± 5.7 8.15 No/over-processed
K1 + 16.0 ± 1.4 3.03 No/over-processed
K4alt +c 13.5 ± 2.1 18.64 Yes
K8 + 12.0 ± 1.4 2.10 No/over-processed
K11 + 10.3 ± 1.0 23.14 No/sub-processed
K9 - 7.4 ± 0.5 6.70 Yes
K6 + 4.5 ± 2.3 14.25 No/sub-processed
K7 + 3.6 ± 0.2 22.03 No/sub-processed
K2 - 1.4 ± 0.5 1.15 Yes
K5 - 2.8 / 11.0 0.80 No/no conclusion
The miRNAs are ranked according to their processing efficiencies (n = 2). We
defined well or moderately processed pre-miRNAs as accumulating above 10%
(percentage are in bold). MiR-K5 hairpin was not ranked since processing effi-
ciencies were too distant in the two experiments.
In grey are emphasized the values of fi parameter that correlate with one or
the two other parameters, allowing to define miRNA hairpin as over-processed,
processed accordingly to hairpin optimality andmiRNA level or sub-processed.
In the case of miR-K5, no conclusion could be drawn due to variation of data
between experiments.
aHairpin optimality takes into account presence of primary and secondary fea-
tures required for optimal processing by the Microprocessor as described pre-
viously (31) and in Supplementary Table S3.
bRelative cellular abundance of viral miRNAs in BCBL-1 is expressed in per-
centage and abundantmiRNAswere defined as above 10% (in bold) (from (31)).
For miR-K4, -K6 and -K9 hairpins, percentage were obtained after addition of
miRNA abundance of 5p and 3p arms.
cIn the case of miR-K4 hairpin, whereas experimentally determined structure
was not optimal, it could be manually folded into an alternative and more op-
timal structure (see Supplementary Table S3).
been published since (46,47). In particular, we calculated the
Shannon entropy for eachmiRNAhairpin as highlighted in
Supplementary Figure S6. The data obtained indicated that
overall a lower Shannon entropy could be observed along
the stem in well-processed miRNA hairpins, with the excep-
tion of stem-loop miR-K11 and miR-K9, as described pre-
viously (47). Supplementary Table S3 summarizes the up-
dated optimality features observed forKSHVpre-miRNAs.
This allowed us to compare processing efficiencies fi of
miRNA hairpins with their hairpin optimality feature and
their corresponding miRNA abundance in infected BCBL-
1 cells (as previously determined (31)) (Table 1). MiRNA
hairpins were ranked from the best to the worst substrates,
which allowed us to define two groups, i.e. the well or mod-
erately processed ( fi ≥ 10 %, miR-K1, -K3, -K4, -K8 and
-K11) and the less efficiently processed (miR-K2, -K6, -K7
and -K9). The miR-K5 hairpin was not ranked since the
results were too discordant between the two experiments.
Overall, well processed miRNA hairpins corresponded to
optimally folded stem-loops. However, with the exception
of miR-K4, this did not correlate with the level of accumu-
lation of mature miRNAs in infected cells (31). Thus, miR-
K1 and -K3 are embedded within hairpins that are optimal
and the best processed by the Microprocessor in vitro (16
and 36% respectively), whereas the level of mature miRNAs
is quite low (∼3 and ∼8%, respectively). Accordingly, they
were defined as over-processed. On the opposite, miR-K11
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
10024 Nucleic Acids Research, 2021, Vol. 49, No. 17
hairpin is only processed at ∼10% although it is the most
abundant miRNA in cells, representing ∼23% of viral clus-
tered miRNAs. This hairpin was thus defined as being sub-
processed. The same held true for miR-K6 and -K7 hair-
pins. Of the remaining hairpins, only miR-K2, -K4 and -K9
showed a good correlation between their optimality feature,
processing efficiencies and cellular abundance.
Altogether, our kinetic analysis therefore shows different
processing efficiencies of KSHV miRNA hairpins within
the polycistronic pri-miR-K10/12, which did not fully cor-
relate with the fact that the miRNA hairpin was optimal
or not or with the accumulation level of the respective ma-
ture miRNAs in infected cells. These discrepancies may re-
flect complex regulation of their biogenesis. Indeed, sub-
processing of miRNA hairpins may be explained by the re-
quirement of additional elements such as protein cofactors
that may be absent or present at low levels in our in vitro
assay. In addition, we also observed cases of over-processed
miRNAs, such asmiR-K1 andmiR-K3,which suggests that
processing by the Microprocessor may serve here another
purpose than solely producing mature miRNAs.
Deletion of pre-miR-K1 or pre-miR-K3 globally impairs the
expression of the remaining miRNAs from the cluster
To further study the processing of the KSHV miRNA clus-
ter, we used a plasmid allowing expression of the pri-miR-
K10/12 sequence driven by a CMV promoter (14). Al-
though it seems that there is a better accumulation of miR-
NAs at the 5′ extremity of the cluster, thewild type construct
gives rise to all ten miRNAs and their relative expression
level is close to what can be measured in latently infected
BCBL1 cells (Supplementary Figure S7).
To investigate the potential other role of miR-K1 and -
K3 processing we generated mutant constructs in which we
deleted individually pre-miR-K1 or pre-miR-K3 sequences
within the polycistronic pri-miR-K10/12. Other miRNA
sequences from the cluster were unchanged. We then as-
sessed the impact of these deletions on the expression of
the remaining clustered miRNAs in the cell. As negative
controls, we deleted pre-miR-K7 and pre-miR-K9, that
are located in the middle and at the 3′ end of the clus-
ter, respectively, and are not well processed in vitro by the
Microprocessor. The resulting mutants were named K1,
K3, K7 and K9 (Figure 2A). These were expressed in
HEK293Grip cells and the accumulation levels of all ma-
ture miRNAs from the cluster were assessed by northern
blot analysis (Figure 2B and C).
Interestingly, the expression of all miRNAs in the clus-
ter was globally and drastically decreased compared to the
wt construct in the K1 and K3 mutants, whereas it was
only moderately or mostly unaffected in theK7 and K9
mutants. K1 construct led to miRNA levels significantly
reduced down to ∼28% (3.6-fold) and ∼52% (1.9-fold) for
miR-K3 and miR-K7, respectively, when compared to the
wt plasmid. All the miRNAs from the cluster were neg-
atively affected, whatever the distance between pre-miR-
K1 and the impacted miRNAs. For example, the farthest
miR-K9 was even more impacted than the closest miR-K2
(∼38% (2.6-fold) versus ∼49 % (2-fold), respectively) (Fig-
ure 2C).Deletion of pre-miR-K3within the cluster was even
more deleterious for the expression of the rest of the clus-
teredmiRNAs. Indeed, miRNA levels were decreased down
to∼6% (16.7-fold) (miR-K5) or∼18% (5.6-fold) (miR-K2).
In the case of K7 mutant, a moderately negative impact
was observed for some miRNAs with the most affected be-
ing miR-K11, which level was reduced down to ∼54% (1.9-
fold). In the case of miR-K11, this may be due to a local
effect of pre-miR-K7 deletion on the folding and/or pro-
cessing of the adjacent miR-K11 hairpin.
Altogether, our results suggest that deletion of pre-miR-
K1 or -K3 within pri-miR-K10/12 globally and drastically
impacts the expression of the remaining miRNAs from the
cluster in cells. This global and severe impact is specific to
pre-miR-K1 and pre-miR-K3 since it was not observed for
two other pre-miRNAs in the cluster.
Expression of clustered miRNAs can be rescued by replacing
pre-miR-K1 by a heterologous pre-miRNA
The analysis of deletion mutants indicates that pre-miR-
K1 and -K3 appear to be required for the optimal expres-
sion of KSHV clusteredmiRNAs. This may be due either to
(i) their ability to recruit the microprocessor, as stem-loop
structures, or (ii) specific sequences that establish tertiary
contacts within the cluster or that are recognized by protein
cofactors.
To investigate which hypothesis should be favored, we in-
serted a heterologous pre-miRNA sequence in lieu of pre-
miR-K1 into theK1mutant and measured the expression
of the rest of the clustered miRNAs. We chose the pre-Let-
7a-1 sequence since both its primary sequence and its sec-
ondary structure, especially in its apical loop, is very dif-
ferent from those of pre-miR-K1, therefore lowering the
possibility to form the same tertiary contacts or to recruit
the same cofactors. The resulting mutant K1-Let7 con-
struct was expressed in HEK293Grip cells and the expres-
sion of miRNAs was assessed by northern blot analysis
(Figure 2B, C). Whereas pre-miR-K1 was as expected not
produced from themutant construct, Let-7a expression was
increased ∼2-fold when compared to the endogenous ex-
pression, showing that Let-7a within the context of the clus-
ter was fully recognized and processed by the Microproces-
sor. The level of accumulation of all the KSHV miRNAs
was measured and compared to the wt construct. Interest-
ingly, expression of all of them was restored almost to the
wt level, showing that replacing pre-miR-K1 with a heterol-
ogous pre-miRNA is sufficient for optimal production of
the other miRNAs in the cluster. In conclusion, our results
suggest that a pre-miRNA structure at this position within
the cluster is sufficient to optimize the expression of KSHV
miRNAs from this construct.
In vitro, mimicking initial cleavage of miR-K1 or miR-K3
hairpins impacts the processing of only few miRNA hairpins
from the cluster
Pre-miR-K1 or pre-miR-K3 are necessary for the optimal
expression of the other miRNAs from the cluster, and at
least for pre-miR-K1, this is independent of primary se-
quence but rather relies on the presence of a miRNA stem-
loop structure. According to that observation, we hypothe-
sized that cleavage of pre-miR-K1 or pre-miR-K3 by the
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10025
Figure 2. Mutational analysis reveals cis regulation within KSHVmiRNA cluster. (A) Schematic view of pri-miR-K10/12 wt or mutant constructs used in
the study. KSHV or hsa Let-7a-1 miRNA hairpins are represented by grey and black bars, respectively. Cytomegalovirus (CMV) promoter is shown. (B)
Northern blot analysis of the accumulation of mature miRNAs, after overexpression of wt and mutant constructs in Hek293Grip cells (n= 3). MiR-16 was
probed as a loading control. Dotted lines indicate where the blot was cut. (C) Histogram showing the relative expression of the clustered miRNAs from
the mutated pri-miR-K10/12 constructs compared to the wt. Error bars were obtained from three independent experiments and P-values were obtained
using unpaired t tests comparing wt versus mutant for each miRNA. ns: non-significant, *P < 0.05, **P < 0.01, ***P < 0.001.
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
10026 Nucleic Acids Research, 2021, Vol. 49, No. 17
Microprocessor may help to favor the processing of the
other miRNAs from the cluster. If true, an RNA mimick-
ing the initial cut of one or the other of these two pre-
miRNAs would lead to a better processing of the remaining
pre-miRNAs.
To test this, we performed in vitro processing of two dif-
ferent in vitro transcribed RNA mimics, namely cut-K1
and cut-K3. Figure 3A gives a schematic view of the RNA
molecules. Briefly, cut-K1 is composed of anRNA fragment
containing pre-miR-K2 to pre-miR-K9 and including pre-
miR-K11. Its 5′ end starts just downstream of the 3′ end of
pre-miR-K1 3p arm, as it would be after Microprocessor
cleavage. Cut-K3 RNA is the combination of two indepen-
dent RNA fragments. One comprises pre-miR-K1 and pre-
miR-K2 and its 3′ end finishes just upstream of the 5′ end
of pre-miR-K3 5p arm. The second embeds pre-miR-K4 to
pre-miR-K9, including pre-miR-K11, and its 5′ end starts
just downstream of the 3′ end of pre-miR-K3 3p arm. These
constructs were incubated with total protein extracts from
cells over-expressing Drosha and DGCR8 and pre-miRNA
products were analyzed by northern blot analysis (Figure
3B). We analyzed pre-miRNAs based on their proximity
(pre-miR-K1, -K3 and -K4) or not (pre-miR-K6, -K7, -K11
and -K8) to the deleted pre-miRNAs and the fact that their
levels obtained in our kinetic analysis were low compared
to the corresponding mature miRNA levels measured in in-
fected cells (miR-K6, -K7 and -K11). Pre-miR-K4 and -K8
were chosen as controls since they were both efficiently pro-
cessed.
Cut-K1RNAgave rise to all the tested pre-miRNAs from
the cluster, with the exception of pre-miR-K1 as expected.
Whereas most of them are produced to similar levels as
from the wt transcript, pre-miR-K3 and to a minor extent
pre-miR-K4 accumulates significantly ∼1.6- and ∼1.3-fold
more respectively when compared to wt condition.
RNA mimicking the cleavage of pre-miR-K3 led to sig-
nificantly more pre-miR-K11 (∼1.5-fold) and to a milder
extent more pre-miR-K8 (∼1.3-fold) whereas pre-miR-K1,
-K4 and -K6 levels were unchanged. Pre-miR-K7 showed
a small increase (∼1.25-fold) but this was not statistically
significant. Whereas cut-K1 showed rather local effect, cut-
K3 increased the processing levels of pre-miRNAs at long
distances.
Altogether, our results show that initial processing of pre-
miR-K1 or pre-miR-K3 does not dramatically improve in
vitro the overall maturation by the Microprocessor of the
other miRNAs within the cluster. On the contrary, it affects
the processing of only few pre-miRNAs. These results may
emphasize the necessity of miR-K1 or miR-K3 hairpins to
be an integral part of the cluster to exert a cis-regulatory
function and/or a sequential processing of the different pre-
miRNAs.
Blocking pre-miR-K1 cleavage by an antisense LNA oligonu-
cleotide phenocopies its deletion
The use of antisense oligonucleotides has been described
as an efficient approach to suppress miRNA function by
sponging the mature miRNA (48). Interestingly, Hall et al.
described that antisense LNA can also inhibit miRNAmat-
uration steps (49). Indeed, an LNA oligonucleotide target-
ing the liver specific miR-122 also binds to pri-miR-122 and
pre-miR-122, invading the stem-loop structure and hinder-
ing recognition by the Microprocessor and Dicer. This may
account for ∼30% of the total inhibition of miR-122 activ-
ity. We therefore decided to use a similar strategy to block
the processing of miR-K1 hairpin in order to downregulate
the whole cluster. The LNA oligonucleotide that we used
consists of 20 nt fully complementary tomaturemiR-K1-5p
arm and contains 8 LNA residues in the middle part (from
nt 8 to nt 15) (Supplementary Table S1). Using a similar
oligonucleotide, Gao and colleagues managed to efficiently
suppress miR-K1 activity (9). In their study, they did not
assess whether this was solely due to the sponging effect
of mature miRNA, or whether this also decreased miRNA
biogenesis.
The previously described construct containing theKSHV
miRNAs cluster was transfected in HEK293Grip cells to-
gether with an LNA targeting miR-K1, namely LNA@K1,
or a control LNA (Supplementary Table S1). We then mea-
sured the levels of mature miRNAs from the entire cluster
by northern blot analysis (Figure 4A, B). As expected, miR-
K1 accumulation was strongly decreased (∼3.7-fold) upon
treatment with LNA@K1 compared to treatment with the
control LNA. Interestingly, the levels of all the other miR-
NAs within the cluster were also negatively affected (∼2.2-
to almost 6-fold decrease) (Figure 4B). As a control, en-
dogenous Let-7a was not affected, since its level was un-
changedwhatever the LNA treatment. Thus, inhibiting pro-
cessing by antisense LNA targeting miR-K1 phenocopies
the impact of K1 mutant on the expression of the clus-
tered miRNAs.
Since LNAs also have the capacity to bind DNA, there is
a possibility that LNA@K1 could interfere with transcrip-
tion of the miRNA cluster and thus explain such a global
effect. We therefore performed RT-qPCR to evaluate the
levels of pri-miR-K10/12 in control LNA and LNA@K1
conditions. Figure 4C shows that pri-miR-K10/12 level was
not affected by LNA@K1 treatment, ruling out a possible
inhibition at the transcription step.
In conclusion, we were able to negatively affect the ex-
pression of the 10 clustered miRNAs of KSHV solely by
targeting miR-K1 sequence. Since it does not interfere with
transcription, this downregulation most likely occurs at the
post-transcriptional level, probably by interfering with the
Microprocessor recognition and/or cleavage of pre-miR-
K1.
Targeting miR-K1 inhibits the expression of the cluster in in-
fected cells
So far, we have demonstrated that the whole cluster can
be downregulated by using one single molecule targeting
miR-K1 sequence. However, our experimental settings did
not mirror natural conditions of KSHV infection, since
the cluster was expressed from a plasmid. In order to test
whether this cis regulation exists in a context closer to phys-
iological infection, we decided to apply our antisense LNA
strategy in HEK293FT cells carrying recombinant KSHV
genomes (HEK293FT-rKSHV) (43). Similar to physiologi-
cal conditions, rKSHV in cells remainsmostly in latent state
and it produces all viral miRNAs, even though their global
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10027
Figure 3. In vitro maturation assays using RNA mimicking miR-K1 and miR-K3 hairpins cleavage. (A) Schematic view of in vitro transcribed RNAs
used in the study. Cut-K1 and cut-K3 RNAs mimic cleavage products by Drosha/DGCR8 of pre-miR-K1 and pre-miR-K3, respectively. As a result,
cut-K3 is composed of 2 RNA fragments. (B) Northern blot analysis of pre-miRNAs produced after 45 min incubation of 1000 fmol of in vitro transcribed
RNAs with HEK293Grip cells total protein extract where Drosha and DGCR8 were overexpressed (right part of blot) or lysis buffer (left part of blot).
Dotted lines indicate where the blot was cut. (C) Histogram showing the relative level of pre-miRNAs compared to the wt. Error bars were obtained from
three independent experiments and p-values were obtained using unpaired t tests comparing wt versus mutant for each pre-miRNA. ns: non-significant,
*P < 0.05, **P < 0.01, ***P < 0.001.
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
10028 Nucleic Acids Research, 2021, Vol. 49, No. 17
Figure 4. Antisense LNA targetingmiR-K1 inhibits expression of othermiRNAs from the cluster. (A) Northern blot analysis of the accumulation ofmature
miRNAs, after overexpression of wt plasmid in HEK293Grip cells during 48h (n= 3), with 20 nM control LNA (Ctrl LNA) or antisense LNA to miR-K1
(LNA@K1) treatment. Let-7a and miR-16 were probed as a control of miRNA expression and as a loading control, respectively. (B) Histogram showing
the relative expression of the different miRNAs upon treatment with LNA@K1 compared to control LNA. Error bars were obtained from 3 independent
experiments and p-values were obtained using unpaired t test with ns: non-significant, *P < 0.05, **P < 0.01, ***P < 0.001. (C) The expression of the
KSHV miRNA primary transcript pri-miR-K10/12 was measured by RT-qPCR in total RNA samples from (B) and normalized to GAPDH.
expression is much lower, probably due to the number of
viral episomes per cell.
By using the same antisense oligonucleotide as previously
described (LNA@K1) in these cells, we were not able to
see any global downregulation of the cluster, with the ex-
ception of miR-K1(data not shown). We hypothesized that
most of the LNAs transfected into the cells were probably
sponged by the mature miR-K1 molecules already abun-
dantly present in the cell, as opposed to the situation where
miRNAs are expressed from a plasmid concomitantly with
the LNA treatment. Thus, only a limited amount of the
LNA would be available for the microprocessor inhibition
and the potential impact on other miRNAs is too low to be
measured.
To cope with this situation, we designed a newLNA com-
plementary to the opposite strand of pre-miR-K1 stem loop
(LNA@K1*). Given that this new molecule will not be se-
questrated by the mature miR-K1, more oligonucleotides
can reach the nucleus and interfere with pre-miR-K1 pro-
cessing. In addition, this would also confirm that the effect
we observe on the expression of the cluster is dependent on
the processing event of pre-miR-K1 and not on a down-
stream function of the mature miR-K1.
First, to show that LNA@K1* can indeed interfere with
processing of miRNA from the cluster, we performed the
experiment in HEK293Grip cells by co-transfecting the pri-
miR-K10/12 construct with either LNA@K1* or a control
LNA and assessed the expression of mature miR-K1, -K4
and -K11 by RT-qPCR. Upon treatment with LNA@K1*,
the accumulation of all three miRNAs dropped substan-
tially compared to the treatment with the control LNA,
while the level of Let-7a remained unchanged (Figure
5A). As previously, a potential impact of LNA treatment
on cluster transcription was verified and only a mild de-
crease (30%) of pri-miR-K10/12 could be observed between
control and LNA@K1* conditions (Figure 5B). We thus
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10029
Figure 5. Inhibition of pre-miR-K1 processing impacts the expression of other viral miRNAs in infected cells. (A) Levels of mature miRNAs in
HEK293Grip cells co-transfected with pri-miR-K10/12 expression plasmid and LNAs complementary to miR-K1* or control LNA. The analysis was
performed on total RNA extracted 48 h post-transfection and miRNA levels were normalized to U48. (B) Measure of pri-miR-K10/12 expression in
samples from (A), GAPDH was used as a normalizer. (C) Levels of mature miRNAs in HEK293FT-rKSHVcells transfected twice with 20nM of LNAs
complementary to miR-K1* or control LNA. U48 was used as a normalizer. (D) Accumulation of neosynthesized miRNAs in HEK293FT-rKSHV cells
transfected with either LNA complementary to miR-K1* or control LNA. 3 h of metabolic labelling with 300 M 4sU was performed 24 h after LNA
transfection and levels of mature miRNAs were measured in total RNA (input) and in isolated newly synthesized fraction (pull-down). To account for
variation in pull-down efficiencies, enrichment of miRNA levels in pull-down over input were determined after normalizing to Let-7 levels. (E) Accumula-
tion of neosynthesized pri-miR-K10/12 measured in the samples from (D). The same approach was used to determine the enrichment of pri-miRNA levels
in the pull-down over input except that CYC1 was used to normalize the data instead of Let-7. Mature miRNAs and pri-miR-K10/12 in all experiments
were quantified by RT-qPCR. All results are displayed relative to control samples set to 1. Bars represent mean ± s.e.m of three (A, B, C) or five (D, E)
independent experiments. Statistical significance was verified by unpaired t test with ns: non-significant, *P < 0.05, ***P < 0.001.
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
10030 Nucleic Acids Research, 2021, Vol. 49, No. 17
confirmed the LNA@K1*-mediated post-transcriptional
inhibition occurring during miRNA processing. However,
transfection of this LNA oligonucleotide into HEK293FT-
rKSHV cells resulted only in a mild impact on KSHV
miRNAexpression, as demonstrated formiR-K1, -K4 and -
K11.While only miR-K11 levels were significantly reduced,
miR-K1 and -K4 did not decrease in a statistically signifi-
cant manner (Figure 5C). This might be explained by a dif-
ferential stability of the different viral miRNAs in infected
cells. In contrast to ectopically expressed cluster, infected
cells already contain a certain amount of mature viral miR-
NAs and their differential turnover would directly influence
the sensitivity of our assay. If the half-lives of miR-K1 and
-K4 are longer than the half-life of miR-K11, then it might
prove difficult to assess the impact of inhibiting their pro-
cessing. As a solution, we set to measure the accumula-
tion of newly synthesized miRNAs. LNA-transfected cells
were therefore incubatedwith 4-thiouridine (4sU), which al-
lowed the isolation of novel transcripts via selective biotiny-
lation and pull-down on streptavidin beads. Due to the vari-
ation between the efficiency of individual pull-downs, viral
miRNAs were analyzed in total RNA (input) as well as in
isolated pull-down fraction and their enrichment in pull-
down over input was expressed relative to Let-7a. Following
two different experimental protocols, we observed a consis-
tent and significant reduction in neosynthesized miR-K1,
-K3, -K4 and -K11 upon treatment with LNA@K1* (Fig-
ure 5D and Supplementary Figure S8). In addition, we veri-
fied by RT-qPCR that the LNA had no impact on the levels
of neosynthesized pri-miRNA transcript (Figure 5E). Thus,
we have shown that the expression of the KSHV miRNA
cluster can be inhibited by using one single oligonucleotide
targeting pre-miR-K1 and that this phenomenon indeed
takes place also within KSHV-infected cells. Although at
this stage, we cannot formally conclude that what is impor-
tant for the cis-regulation is the processing event or the pres-
ence of a stem-loop, our results clearly point toward the im-
portance of pre-miR-K1.
DISCUSSION
In this study, we explored a complex layer for miRNA bio-
genesis regulation in which the expression of miRNA hair-
pins within a large miRNA cluster is interdependent. We
showed that two miRNA hairpins, namely miR-K1 and
-K3 hairpins, within the intronic KSHV miRNA cluster
were required for the optimal expression of the remain-
ing miRNAs. Indeed, their deletion within an expression
plasmid drastically diminished clustered miRNAs expres-
sion in cell. Similarly, antisense LNAs that bind to the pre-
miR-K1 hairpin and inhibit its processing by impeding the
recognition and/or cleavage by the Microprocessor led to
global downregulation of the cluster. This strategy also al-
lowed to decrease viral miRNAs in the more natural con-
text of infected cells. Furthermore, our data showed that
the pre-miRNA feature per se, at least for pre-miR-K1, is
responsible for this regulation since pre-miR-K1 could be
replaced by a heterologous pre-miRNA. Altogether, these
results indicate that miR-K1 and miR-K3 hairpins are im-
portant cis-regulatory elements for the expression of the
KSHV clustered miRNAs. Previously, the Krueger labo-
ratory produced bacmid constructions containing KSHV
genome deleted from individual KSHV miRNAs and they
also noticed a decrease in expression of other viral miRNAs
in mutants deleted of miR-K1 and miR-K3 (50). Here, we
explain these observations through the cis-regulatory func-
tion of these two pre-miRNAs.
Interdependency of clustered miRNA hairpins was doc-
umented previously for different miRNAs and in diverse
species (37–40,40,41). However, it has so far only been
studied for small clusters of two or three miRNA hairpins
where a helper hairpin assists the processing of a neighbor-
ing suboptimal hairpin. Here, we show cis regulation for
the first time within a large cluster of 10 miRNA hairpins.
Thus, ‘assisted’ miRNA hairpins are not all proximal to the
helper hairpins. In addition, they are not necessarily sub-
optimal as demonstrated by our 2D structure probing data
published previously (31). So, in the case of KSHV miR-
NAs cluster, cis regulation might result from a more com-
plex mechanism. One possibility could rely on the recruit-
ment of the Microprocessor by miR-K1 and miR-K3 hair-
pins, inducing its local concentration to re-initiate further
maturation events on the same polycistronic pri-miRNA.
From a conceptual point of view, this might be compared
to ribosome re-initiating translation on a same mRNA.
However, from a mechanistic point of view, Microproces-
sor would not scan the pri-miRNA but rather cycle from
the cleaved miRNA hairpin and relocate on promiscuous
miRNA hairpin. Two alternate but not mutually exclusive
models may help this repositioning. One model would be
that the globular and compact 3D structure of pri-miRNA
may per se help the Microprocessor to relocate. Indeed,
structural study performed on pri-miR-17–92 shows such
organization (30). We previously determined the 2D struc-
ture of KSHV miRNA cluster using SHAPE method (31).
Although we could not conclude on a compact arrange-
ment of the viral pri-miRNA, we did observe numerous
stem–loops, containing or not miRNAs, that could partici-
pate to long-distance interaction to maintain such compact
3D structure. A second model would involve protein cofac-
tors able to assist in the recruitment of the Microproces-
sor on the neighboring miRNA hairpin. Very recently, it
was shown that ERH and SAFB2 can interact with the Mi-
croprocessor and their ability to dimerize may even medi-
ate multimerization of the Microprocessor and allow its si-
multaneous binding to several hairpins (37,40,42,51). How-
ever, this model remains to be clearly established. For ex-
ample, the role of ERH and SAFB2 proteins may be spe-
cific to only certain miRNA clusters and not all clustered
miRNAs may depend on such assistance. Indeed, using
CRISPR/Cas9 editing, Lataniotis et al showed that editing
of miR-195 led to a decrease of its neighboring miR-497,
whereas no such interdependency was observed for miR-
106–25 or miR-17–92 clusters (39). Another study, based
on genetically engineered mice, also showed that deletion
of any miRNA from the cluster miR-17–92 did not alter
the expression of the other miRNAs (52). In the case of the
KSHV cluster, pre-miR-K1 and pre-miR-K3 may interact
with high affinity with a protein cofactor, potentializing fur-
ther interaction with the other pre-miRNAs. This protein
cofactor might then recruit or improve the Microprocessor
activity.
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10031
Another mechanism may rely on specific structural con-
strains that could be resolved after cleavage of miR-K1
and miR-K3 hairpins, rendering the other miRNA hair-
pins more accessible to the Microprocessor. Indeed, previ-
ous studies reported that the globular fold of the pri-miR-
17–92 may autoregulate its processing (30,53,54). Chaulk
et al. demonstrated that the compact fold of the cluster is
adopted through a specific tertiary contact between pre-
miR-19b and a non-miRNA hairpin resulting in reduced
miR-92a expression whereas this inhibitory effect could be
abolished by disrupting this contact (54). However, our data
obtained from our in vitro processing assays of RNA mim-
icking initial cleavage of miR-K1 or miR-K3 hairpins show
neither global nor drastic improvement of other miRNA
hairpins maturation. This suggests that miR-K1- and miR-
K3-dependent regulation may rather occur in cis, when the
two hairpins are still present in the cluster. Cleavage of the
cluster may also happen in a hierarchical way similarly to
pri-miR-17–92 (55). In that case, the restricted impact of
pre-miR-K1 or pre-miR-K3 cleavage may reflect that pro-
cessing of the KSHV cluster occurs in different steps having
each downstream additive positive effect. Unfortunately,
the experimental design of our in vitro processing assays did
not allow to follow complete processing of the cluster. Thus,
we probably observed only the first cleavage events.
Another intriguing aspect of this work is the fact that de-
spite their efficient processing by the Microprocessor, the
levels of mature miR-K1 and -K3 are low in infected cells.
It would be interesting to know if these two pre-miRNAs
are as well efficiently exported into the cytoplasm and/or
processed by Dicer and how the excess of pre-miRNAs is
eliminated from the cell. It was shown previously thatMCP-
1-induced protein-1 (MCPIP1), a suppressor of miRNA
biogenesis and involved in immunity, could directly cleave
KSHV pre-miRNAs through its RNase domain (56,57).
However, while their results show that pre-miR-K1 and -K3
can be bound and cleaved by MCPIP1, almost all remain-
ing KSHV pre-miRNAs are prone to be degraded as well,
suggesting that another protein or mechanism might be in-
volved in selective decay of excessive pre-miR-K1 and –K3.
KSHV miRNAs are involved in a multitude of functions
related to the viral life cycle, immune escape, and patho-
genesis (reviewed by e.g. (58,59)). Latent infection is one
of the biggest hurdles in the treatment of KSHV-related
pathologies. Therefore, the role of viral miRNAs in latency
maintenance is of great importance. Several groups have
shown that artificial reduction of levels of particular KSHV
miRNAs can lead to higher viral reactivation. For exam-
ple, the main transactivator of lytic cycle RTA is directly
regulated by at least two miRNAs, miR-K7-5p and miR-
K5 (10,60). In addition, miR-K1 indirectly controls latency
maintenance by downregulation of the inhibitor of NF-
kB pathway, IkBa (9). Furthermore, many of their protein
targets validated to date are involved in pathways impor-
tant in oncogenesis (61). Interestingly, miR-K11 presents
the same seed sequence as a well-known oncomiR miR-
155 and both miRNAs target a common subset of genes
(62,63). TheGao’s and Renne’s groups have shown that sev-
eral KSHVmiRNAs participate to the transforming poten-
tial of KHSV by targeting cell growth and survival path-
ways (64,65). Targeting several, if not all the KSHV miR-
NAs can therefore represent a valuable therapeutic option.
Recently, Ju et al. have proposed a therapeutic strategy
based on LNA-modified oligonucleotides complementary
to miR-K1, -K4 and -K11 coupled to carbon dots for better
intracellular delivery (66). While they use a combination of
three different molecules, our results suggest that blocking
the processing of one single miRNA might lead to a global
decrease of the entire miRNA cluster. Given that about 25
to 40% of all human miRNAs are embedded in clusters
(25,26), and given the growing body of evidence that they
are implicated in disease, the ability to suppress their expres-
sion as a whole might be of importance for future therapies.
DATA AVAILABILITY
All data generated or analyzed during this study are in-
cluded in this published article (and its supplementary file).
Requests for material should be made to the corresponding
authors.
SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.
ACKNOWLEDGEMENTS
The authors would like to thankmembers of the Pfeffer lab-
oratory for discussion. They also would like to thank Prof
Narry Kim for the kind gift of plasmids expressing flag-
tagged Drosha and DGCR8.
Author contributions: A.F., S.P., M.C. and M.V. conceived
the project. M.C., M.V., S.P. and A.F. designed the work.
M.C., M.V., R.R. and A.F. performed the experiments and
analyzed the results. P.D. performed the kinetic analysis.
E.E. and P.M.O. generated the HEK293FT-rKSHV cells.
A.F. and S.P. coordinated the work and S.P. assured fund-
ing. M.V., P.D., S.P. and A.F. wrote the manuscript with in-
put from the other authors. All authors reviewed and ap-
proved the final manuscript.
FUNDING
This work was funded by the European Research Council
[ERC-CoG-647455 RegulRNA] and was performed in the
Interdisciplinary Thematic Institute IMCBio, as part of the
ITI 2021–2028 program of the University of Strasbourg,
CNRS and Inserm; IdEx Unistra [ANR-10-IDEX-0002];
SFRI-STRAT’US project [ANR 20-SFRI-0012]; EUR IM-
CBio [IMCBio ANR-17-EURE-0023] under the frame-
work of the French Investments for the Future Program as
well as from the previous LabexNetRNA [ANR-10-LABX-
0036]. It also received funding from the FrenchMinister for
Higher Education, Research and Innovation (PhD contract
to M.V.). Funding for open access charge: European Re-
search Council [ERC-CoG-647455 RegulRNA].
Conflict of interest statement.None declared.
REFERENCES
1. Ganem,D. (2007) Kaposi’s sarcoma-associated herpesvirus. In:
Knipe,D.M., Howley,P.M., Griffin,D.E., Lamb,R.A., Martin,M.A.,
Roizman,B. and Straus,S.E. (eds). Fields Virology. Lippincott,
Williams & Wilkins, Philadelphia, Vol. 5, pp. 2847–2888.
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
10032 Nucleic Acids Research, 2021, Vol. 49, No. 17
2. Aneja,K.K. and Yuan,Y. (2017) Reactivation and lytic replication of
Kaposi’s sarcoma-associated herpesvirus: an update. Front.
Microbiol., 8, 613.
3. Cai,X., Lu,S., Zhang,Z., Gonzalez,C.M., Damania,B. and
Cullen,B.R. (2005) Kaposi’s sarcoma-associated herpesvirus
expresses an array of viral microRNAs in latently infected cells. Proc.
Natl. Acad. Sci. U.S.A., 102, 5570–5575.
4. Pfeffer,S., Sewer,A., Lagos-Quintana,M., Sheridan,R., Sander,C.,
Gra¨sser,F.A., van Dyk,L.F., Ho,C.K., Shuman,S., Chien,M. et al.
(2005) Identification of microRNAs of the herpesvirus family. Nat.
Methods, 2, 269–276.
5. Samols,M.A., Hu,J., Skalsky,R.L. and Renne,R. (2005) Cloning and
identification of a microRNA cluster within the latency-associated
region of Kaposi’s sarcoma-associated herpesvirus. J. Virol., 79,
9301–9305.
6. Cai,X. and Cullen,B.R. (2006) Transcriptional origin of Kaposi’s
sarcoma-associated herpesvirus microRNAs. J. Virol., 80, 2234–2242.
7. Lin,Y.-T., Kincaid,R.P., Arasappan,D., Dowd,S.E.,
Hunicke-Smith,S.P. and Sullivan,C.S. (2010) Small RNA profiling
reveals antisense transcription throughout the KSHV genome and
novel small RNAs. RNA N. Y., 16, 1540–1558.
8. Bellare,P. and Ganem,D. (2009) Regulation of KSHV lytic switch
protein expression by a virus-encoded microRNA: an evolutionary
adaptation that fine-tunes lytic reactivation. Cell Host Microbe, 6,
570–575.
9. Lei,X., Bai,Z., Ye,F., Xie,J., Kim,C.-G., Huang,Y. and Gao,S.-J.
(2010) Regulation of NF-kappaB inhibitor IkappaBalpha and viral
replication by a KSHV microRNA. Nat. Cell Biol., 12, 193–199.
10. Lu,F., Stedman,W., Yousef,M., Renne,R. and Lieberman,P.M. (2010)
Epigenetic regulation of Kaposi’s sarcoma-associated herpesvirus
latency by virus-encoded microRNAs that target Rta and the cellular
Rbl2-DNMT pathway. J. Virol., 84, 2697–2706.
11. Gottwein,E. and Cullen,B.R. (2010) A human herpesvirus
microRNA inhibits p21 expression and attenuates p21-mediated cell
cycle arrest. J. Virol., 84, 5229–5237.
12. Li,W., Hu,M., Wang,C., Lu,H., Chen,F., Xu,J., Shang,Y., Wang,F.,
Qin,J., Yan,Q. et al. (2017) A viral microRNA downregulates
metastasis suppressor CD82 and induces cell invasion and
angiogenesis by activating the c-Met signaling. Oncogene, 36,
5407–5420.
13. Liu,X., Happel,C. and Ziegelbauer,J.M. (2017) Kaposi’s
sarcoma-associated herpesvirus microRNAs target GADD45B to
protect infected cells from cell cycle arrest and apoptosis. J. Virol., 91,
e02045-16.
14. Suffert,G., Malterer,G., Hausser,J., Viilia¨inen,J., Fender,A.,
Contrant,M., Ivacevic,T., Benes,V., Gros,F., Voinnet,O. et al. (2011)
Kaposi’s sarcoma herpesvirus microRNAs target caspase 3 and
regulate apoptosis. PLoS Pathog., 7, e1002405.
15. Boss,I.W. and Renne,R. (2010) Viral miRNAs: tools for immune
evasion. Curr. Opin. Microbiol., 13, 540–545.
16. Gallaher,A.M., Das,S., Xiao,Z., Andresson,T., Kieffer-Kwon,P.,
Happel,C. and Ziegelbauer,J. (2013) Proteomic screening of human
targets of viral microRNAs reveals functions associated with immune
evasion and angiogenesis. PLoS Pathog., 9, e1003584.
17. Nachmani,D., Stern-Ginossar,N., Sarid,R. and Mandelboim,O.
(2009) Diverse herpesvirus microRNAs target the stress-induced
immune ligand MICB to escape recognition by natural killer cells.
Cell Host Microbe, 5, 376–385.
18. Ramalingam,D. and Ziegelbauer,J.M. (2017) Viral microRNAs target
a gene network, inhibit STAT activation, and suppress interferon
responses. Sci. Rep., 7, 40813.
19. Qin,J., Li,W., Gao,S.-J. and Lu,C. (2017) KSHV microRNAs: tricks
of the devil. Trends Microbiol., 25, 648–661.
20. Bartel,D.P. (2018) Metazoan MicroRNAs. Cell, 173, 20–51.
21. Cheloufi,S., Dos Santos,C.O., Chong,M.M.W. and Hannon,G.J.
(2010) A dicer-independent miRNA biogenesis pathway that requires
Ago catalysis. Nature, 465, 584–589.
22. Cifuentes,D., Xue,H., Taylor,D.W., Patnode,H., Mishima,Y.,
Cheloufi,S., Ma,E., Mane,S., Hannon,G.J., Lawson,N.D. et al. (2010)
A novel miRNA processing pathway independent of Dicer requires
Argonaute2 catalytic activity. Science, 328, 1694–1698.
23. Okamura,K., Hagen,J.W., Duan,H., Tyler,D.M. and Lai,E.C. (2007)
The mirtron pathway generates microRNA-class regulatory RNAs in
Drosophila. Cell, 130, 89–100.
24. Ruby,J.G., Jan,C.H. and Bartel,D.P. (2007) Intronic microRNA
precursors that bypass Drosha processing. Nature, 448, 83–86.
25. Altuvia,Y., Landgraf,P., Lithwick,G., Elefant,N., Pfeffer,S.,
Aravin,A., Brownstein,M.J., Tuschl,T. and Margalit,H. (2005)
Clustering and conservation patterns of human microRNAs. Nucleic
Acids Res, 33, 2697–2706.
26. Kabekkodu,S.P., Shukla,V., Varghese,V.K., D’ Souza,J.,
Chakrabarty,S. and Satyamoorthy,K. (2018) Clustered miRNAs and
their role in biological functions and diseases. Biol. Rev. Camb.
Philos. Soc., 93, 1955–1986.
27. Bortolin-Cavaille´,M.-L., Dance,M., Weber,M. and Cavaille´,J. (2009)
C19MC microRNAs are processed from introns of large Pol-II,
non-protein-coding transcripts. Nucleic Acids Res., 37, 3464–3473.
28. Olive,V., Li,Q. and He,L. (2013) mir-17-92: a polycistronic oncomir
with pleiotropic functions. Immunol. Rev., 253, 158–166.
29. Seitz,H., Royo,H., Bortolin,M.-L., Lin,S.-P., Ferguson-Smith,A.C.
and Cavaille´,J. (2004) A large imprinted microRNA gene cluster at
the mouse Dlk1-Gtl2 domain. Genome Res., 14, 1741–1748.
30. Chaulk,S.G., Thede,G.L., Kent,O.A., Xu,Z., Gesner,E.M.,
Veldhoen,R.A., Khanna,S.K., Goping,I.S., MacMillan,A.M.,
Mendell,J.T. et al. (2011) Role of pri-miRNA tertiary structure in
miR-17∼92 miRNA biogenesis. RNA Biol., 8, 1105–1114.
31. Contrant,M., Fender,A., Chane-Woon-Ming,B., Randrianjafy,R.,
Vivet-Boudou,V., Richer,D. and Pfeffer,S. (2014) Importance of the
RNA secondary structure for the relative accumulation of clustered
viral microRNAs. Nucleic Acids Res., 42, 7981–7996.
32. Conrad,T., Marsico,A., Gehre,M. and Orom,U.A. (2014)
Microprocessor activity controls differential miRNA biogenesis In
Vivo. Cell Rep., 9, 542–554.
33. Feng,Y., Zhang,X., Song,Q., Li,T. and Zeng,Y. (2011) Drosha
processing controls the specificity and efficiency of global microRNA
expression. Biochim. Biophys. Acta, 1809, 700–707.
34. Creugny,A., Fender,A. and Pfeffer,S. (2018) Regulation of primary
microRNA processing. FEBS Lett., 592, 1980–1996.
35. Michlewski,G. and Ca´ceres,J.F. (2019) Post-transcriptional control of
miRNA biogenesis. RNA N. Y., 25, 1–16.
36. Treiber,T., Treiber,N. and Meister,G. (2019) Regulation of
microRNA biogenesis and its crosstalk with other cellular pathways.
Nat. Rev. Mol. Cell Biol., 20, 5–20.
37. Fang,W. and Bartel,D.P. (2020) MicroRNA clustering assists
processing of suboptimal microRNA hairpins through the action of
the ERH protein.Mol. Cell, 78, 289–302.
38. Haar,J., Contrant,M., Bernhardt,K., Feederle,R., Diederichs,S.,
Pfeffer,S. and Delecluse,H.-J. (2016) The expression of a viral
microRNA is regulated by clustering to allow optimal B cell
transformation. Nucleic Acids Res., 44, 1326–1341.
39. Lataniotis,L., Albrecht,A., Kok,F.O., Monfries,C.A.L., Benedetti,L.,
Lawson,N.D., Hughes,S.M., Steinhofel,K., Mayr,M. and
Zampetaki,A. (2017) CRISPR/Cas9 editing reveals novel
mechanisms of clustered microRNA regulation and function. Sci.
Rep., 7, 8585.
40. Shang,R., Baek,S.C., Kim,K., Kim,B., Kim,V.N. and Lai,E.C. (2020)
Genomic clustering facilitates nuclear processing of suboptimal
Pri-miRNA Loci.Mol. Cell, 78, 303–316.
41. Truscott,M., Islam,A.B.M.M.K. and Frolov,M.V. (2016) Novel
regulation and functional interaction of polycistronic miRNAs. RNA,
22, 129–138.
42. Kwon,S.C., Jang,H., Shen,S., Baek,S.C., Kim,K., Yang,J., Kim,J.,
Kim,J.-S., Wang,S., Shi,Y. et al. (2020) ERH facilitates microRNA
maturation through the interaction with the N-terminus of DGCR8.
Nucleic Acids Res., 48, 11097–11112.
43. Vieira,J. and O’Hearn,P.M. (2004) Use of the red fluorescent protein
as a marker of Kaposi’s sarcoma-associated herpesvirus lytic gene
expression. Virology, 325, 225–240.
44. Myoung,J. and Ganem,D. (2011) Generation of a
doxycycline-inducible KSHV producer cell line of endothelial origin:
maintenance of tight latency with efficient reactivation upon
induction. J. Virol. Methods, 174, 12–21.
45. Marzi,M.J. and Nicassio,F. (2018) Uncovering the stability of mature
miRNAs by 4-thio-uridine metabolic labeling.Methods Mol. Biol.,
1823, 141–152.
46. Fang,W. and Bartel,D.P. (2015) The menu of features that define
primary MicroRNAs and enable de novo design of MicroRNA genes.
Mol. Cell, 60, 131–145.
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021
Nucleic Acids Research, 2021, Vol. 49, No. 17 10033
47. Rice,G.M., Shivashankar,V., Ma,E.J., Baryza,J.L. and Nutiu,R.
(2020) Functional atlas of primary miRNA maturation by the
microprocessor.Mol. Cell, 80, 892–902.
48. Lanford,R.E., Hildebrandt-Eriksen,E.S., Petri,A., Persson,R.,
Lindow,M., Munk,M.E., Kauppinen,S. and Orum,H. (2010)
Therapeutic silencing of MicroRNA-122 in primates with chronic
hepatitis C virus infection. Science, 327, 198–201.
49. Gebert,L.F.R., Rebhan,M.A.E., Crivelli,S.E.M., Denzler,R.,
Stoffel,M. and Hall,J. (2014) Miravirsen (SPC3649) can inhibit the
biogenesis of miR-122. Nucleic Acids Res., 42, 609–621.
50. Jain,V., Plaisance-Bonstaff,K., Sangani,R., Lanier,C., Dolce,A.,
Hu,J., Brulois,K., Haecker,I., Turner,P., Renne,R. et al. (2016) A
toolbox for herpesvirus miRNA research: Construction of a complete
set of KSHV miRNA deletion mutants. Viruses, 8, 54.
51. Hutter,K., Lohmu¨ller,M., Jukic,A., Eichin,F., Avci,S., Labi,V.,
Szabo,T.G., Hoser,S.M., Hu¨ttenhofer,A., Villunger,A. et al. (2020)
SAFB2 enables the processing of suboptimal stem-loop structures in
clustered primary miRNA transcripts.Mol. Cell, 78, 876–889.
52. Han,Y.-C., Vidigal,J.A., Mu,P., Yao,E., Singh,I., Gonza´lez,A.J.,
Concepcion,C.P., Bonetti,C., Ogrodowski,P., Carver,B. et al. (2015)
An allelic series of miR-17∼92–mutant mice uncovers functional
specialization and cooperation among members of a microRNA
polycistron. Nat. Genet., 47, 766–775.
53. Chakraborty,S., Mehtab,S., Patwardhan,A. and Krishnan,Y. (2012)
Pri-miR-17-92a transcript folds into a tertiary structure and
autoregulates its processing. RNA N. Y., 18, 1014–1028.
54. Chaulk,S.G., Xu,Z., Glover,M.J.N. and Fahlman,R.P. (2014)
MicroRNA miR-92a-1 biogenesis and mRNA targeting is modulated
by a tertiary contact within the miR-17∼92 microRNA cluster.
Nucleic Acids Res., 42, 5234–5244.
55. Donayo,A.O., Johnson,R.M., Tseng,H.-W., Izreig,S., Gariepy,A.,
Mayya,V.K., Wu,E., Alam,R., Lussier,C., Jones,R.G. et al. (2019)
Oncogenic biogenesis of pri-miR-17∼92 reveals hierarchy and
competition among polycistronic MicroRNAs.Mol. Cell, 75,
340–356.
56. Happel,C., Ramalingam,D. and Ziegelbauer,J.M. (2016)
Virus-Mediated alterations in miRNA factors and degradation of
viral miRNAs by MCPIP1. PLoS Biol., 14, e2000998.
57. Kook,I. and Ziegelbauer,J.M. (2021) Monocyte chemoattractant
protein-induced protein 1 directly degrades viral miRNAs with a
specific motif and inhibits KSHV infection. Nucleic Acids Res., 49,
4456–4471.
58. Gottwein,E. (2012) Kaposi’s sarcoma-associated herpesvirus
microRNAs. Front. Microbiol., 3, 165.
59. Ramalingam,D., Kieffer-Kwon,P. and Ziegelbauer,J.M. (2012)
Emerging themes from EBV and KSHV microRNA targets. Viruses,
4, 1687–1710.
60. Lin,X., Liang,D., He,Z., Deng,Q., Robertson,E.S. and Lan,K. (2011)
miR-K12-7-5p encoded by Kaposi’s sarcoma-associated herpesvirus
stabilizes the latent state by targeting viral ORF50/RTA. PloS One, 6,
e16224.
61. Wong,J.P. and Damania,B. (2017) Modulation of oncogenic signaling
networks by Kaposi’s sarcoma-associated herpesvirus. Biol. Chem.,
398, 911–918.
62. Gottwein,E., Mukherjee,N., Sachse,C., Frenzel,C., Majoros,W.H.,
Chi,J.-T.A., Braich,R., Manoharan,M., Soutschek,J., Ohler,U. et al.
(2007) A viral microRNA functions as an orthologue of cellular
miR-155. Nature, 450, 1096–1099.
63. Skalsky,R.L., Samols,M.A., Plaisance,K.B., Boss,I.W., Riva,A.,
Lopez,M.C., Baker,H.V. and Renne,R. (2007) Kaposi’s
sarcoma-associated herpesvirus encodes an ortholog of miR-155. J.
Virol., 81, 12836–12845.
64. Moody,R., Zhu,Y., Huang,Y., Cui,X., Jones,T., Bedolla,R., Lei,X.,
Bai,Z. and Gao,S.-J. (2013) KSHV microRNAs mediate cellular
transformation and tumorigenesis by redundantly targeting cell
growth and survival pathways. PLoS Pathog., 9, e1003857.
65. Gay,L.A., Stribling,D., Turner,P.C. and Renne,R. (2021) Kaposi’s
sarcoma-associated herpesvirus MicroRNA mutants modulate cancer
hallmark phenotypic differences in human endothelial cells. J. Virol.,
95, e02022-20.
66. Ju,E., Li,T., Liu,Z., da Silva,S.R., Wei,S., Zhang,X., Wang,X. and
Gao,S.-J. (2020) Specific inhibition of viral MicroRNAs by carbon
dots-mediated delivery of locked nucleic acids for therapy of
virus-induced cancer. ACS Nano, 14, 476–487.
D
ow
nloaded from
 https://academ
ic.oup.com
/nar/article/49/17/10018/6355881 by guest on 10 D
ecem
ber 2021