Murine cytomegalovirus infection of cultured mouse cells induces ...

Journal of General Virology (2012), 93, 1537–1547

DOI 10.1099/vir.0.041822-0

Murine cytomegalovirus infection of cultured mouse cells induces expression of miR-7a Alexandra Dittmer and Klaus Fo¨rstemann Correspondence

Gene Center, Ludwig Maximilian University, Feodor Lynen Strasse 25, 81377 Munich, Germany

Klaus Fo¨rstemann [email protected]

Received 6 February 2012 Accepted 16 March 2012

One goal of virus infection is to reprogramme the host cell to optimize virus replication. As part of this process, viral microRNAs (miRNAs) may compete for components of the miRNA/small interfering RNA pathway, as well as regulate cellular targets. Murine cytomegalovirus (MCMV) has been described to generate large numbers of viral miRNAs during lytic infection and was therefore used to analyse the impact of viral miRNAs on the host-cell small-RNA system, as well as to check for sorting of viral small RNAs into specific Argonaute (Ago) proteins. Deep-sequencing analysis of MCMV-infected cells revealed that viral miRNAs represented only ~13 % of all detected miRNAs. All previously described MCMV miRNAs with the exception of miR-m88-1* were confirmed, and for the MCMV miR-m01-1 hairpin, an additional miRNA, designated miR-m01-13p, was found. Its presence was confirmed by quantitative real-time PCR and Northern blotting. Deep sequencing after RNA-induced silencing complex (RISC) immunoprecipitation with antibodies specific for either Ago1 or Ago2 showed that all MCMV miRNAs were loaded into both RISCs. The ratio of MCMV to mouse miRNAs was not increased after immunoprecipitation of Ago proteins. Viral miRNAs therefore did not overwhelm the host miRNA processing system, nor were they incorporated preferentially into RISCs. Three mouse miRNAs were found that showed altered expression as a result of MCMV infection. Downregulation of miR-27a, as described previously, could be confirmed. In addition, miR-26a was downregulated, and upregulation of miR-7a dependent on viral protein expression could be observed.

INTRODUCTION MicroRNAs (miRNAs) are non-coding RNAs of around 21–23 nt that repress cognate mRNAs, often referred to as target mRNAs. They are transcribed as long hairpincontaining precursors, which are first processed into a stem–loop structure by the RNase III Drosha and, after export to the cytoplasm, are further processed by a second RNase III, called Dicer (Bartel, 2004; He & Hannon, 2004; Lee et al., 2002). One strand of the resulting short duplex is then incorporated into a member of the Argonaute (Ago) protein family, the functional component within the RNAinduced silencing complex (RISC) (Bartel, 2004; Du & Zamore, 2005; Meister & Tuschl, 2004). Once loaded, the mature miRNA can remain stable with a mean half-life of several days (Gantier et al., 2011), whilst the second strand is rapidly degraded. Functionally, miRNAs act at the posttranscriptional level, thus adding another possible means of regulating gene expression for a number of biological processes including the cell cycle, metabolism and

The NCBI Sequence Read Archive accession number determined in this study is GSE36919. Three supplementary tables are available with the online version of this paper.

041822 G 2012 SGM

development (Biggar & Storey, 2011; Malumbres, 2011; Sayed & Abdellatif, 2011).

&

Several viruses, the majority from the family Herpesviridae, encode their own miRNAs and use the host-cell miRNA pathway both to regulate their own life cycle, (e.g. the switch from latent to lytic infection) and to manipulate the host system in a non-immunogenic manner (PlaisanceBonstaff & Renne, 2011; Tuddenham & Pfeffer, 2011). Host miRNAs can also be co-opted for virus replication, such as miR-122 which activates hepatitis virus C translation (Jopling et al., 2005). Conversely, host miRNAs can have antiviral functions by targeting the viral genome or downregulating essential components for infection. Vesicular stomatitis virus is directly targeted by miR-24 and miR-93 (Otsuka et al., 2007), and RIG-I-inducible miR-23b inhibits infections by minor-group rhinoviruses through downregulation of the receptor VLDLR (Ouda et al., 2011). In turn, miRNAs, which may have antiviral effects, can be downregulated by viral factors, e.g. miR-100 and miR-101, which regulate mTOR signalling components and are downregulated by human cytomegalovirus (HCMV) infection (Wang et al., 2008). It has previously been shown that murine cytomegalovirus (MCMV) encodes 18 pre-miRNAs that are highly expressed during lytic infection using a small-RNA cloning

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

Printed in Great Britain

Bueno

1537

A. Dittmer and K. Fo¨rstemann

approach (Buck et al., 2007; Do¨lken et al., 2007). Analysis of cellular miRNAs revealed only slight changes, indicating that even high levels of viral miRNA have no global impact on the host miRNA system. The only exception found was miR-27a, which was rapidly downregulated post-transcriptionally and was shown to have antiviral properties (Buck et al., 2010). Interestingly, miR-27a is also targeted for degradation by the herpesvirus saimiri non-coding RNAs HSUR 1 and 2 (Cazalla et al., 2010). In this study, we wanted to investigate further the impact of MCMV infection on the mouse miRNA system. To this end, we constructed small-RNA libraries suitable for deep sequencing of both mock- and MCMV-infected cells, as well as after RISC immunoprecipitation (IP), to check for expression and also loading of both viral and mouse miRNAs into the functional component. We found that viral miRNAs only comprised ~13 % of all miRNAs late after infection. A novel MCMV mature miRNA, m01-1-3p, was detected, indicating that our sequencing covered the small-RNA profile at least as comprehensively as previous reports. After RISC IP, the viral miRNA component was not increased, suggesting that viral miRNAs were not preferentially loaded. Whilst most cellular miRNAs remain unchanged following MCMV infection, we could confirm downregulation of miR-27a. In addition, we found two more host miRNAs that are influenced by virus infection, miR-26a, which was moderately downregulated in fibroblasts but not in epithelial cells, and miR-7a, which was strongly upregulated in both cells types. Neither virus entry nor viral RNA production were sufficient for miR-7a upregulation, indicating that synthesis of a viral protein is required.

RESULTS Analysis of mouse miRNAs in Solexa libraries of MCMV-infected cells To look at MCMV miRNA production and to analyse the impact of MCMV infection on the host miRNA system, we generated small-RNA libraries from NIH 3T3 fibroblasts infected with MCMV at m.o.i. of 1 at 72 h post-infection

(p.i.) or from non-infected cells. After removal of adaptor sequences and length selection (17–28 nt), 3.9 million reads for mock-infected and 2.8 million reads for MCMVinfected cells were obtained (Table 1). With no mismatches allowed, 78.6 % of the reads from mock-infected cells and 75.6 % of the reads from MCMV-infected cells could be mapped to either the mouse or the MCMV genome. The majority of reads from total RNA, from both mock- or MCMV-infected cells, were cellular miRNAs (70.76 and 52.89 %, respectively). rRNAs and tRNAs accounted only for a small fraction of the reads (1.22 % in mock and 2.23 % in MCMV-infected cells). Of the annotated mouse miRNAs, ~30 % could be detected in NIH 3T3 cells with at least ten reads (see Table S1, available in JGV Online). Whilst most miRNAs showed similar read numbers for both mock- and MCMV-infected cells, some outliers were noted (Fig. 1a). miRNAs with a more than tenfold difference and represented by at least 1000 reads were let-7g, miR-7a, miR-26a, miR-27a, miR-125b-3p and miR-199a-5p. We used qPCR to verify these results. Total RNA was isolated from mock- or MCMV-infected cells at 72 h p.i. and qPCR was performed for the 56 most abundant miRNAs (Table S2). Whilst most miRNAs were expressed at a slightly lower level in MCMVinfected cells (possibly an artefact of normalization), three miRNAs showed a greater than fourfold difference (Fig. 1b). As already demonstrated by Buck et al. (2010), miR-27a was drastically reduced in MCMV-infected cells, as was miR-26a, although to a lesser degree. Interestingly, one mouse miRNA, miR-7a, was highly upregulated (Table 2). Analysis of MCMV miRNAs in deep-sequencing libraries of MCMV-infected cells Only 7.83 % of the total reads in MCMV-infected cells, accounting for 13 % of all miRNA reads, mapped to viral miRNAs (Table 1). This is in contrast to previous observations (Do¨lken et al., 2007) where a small-RNA cloning approach showed the majority of miRNAs to be of viral origin at 72 h p.i. This discrepancy was not due to differences in the m.o.i., as the same protocol was used (m.o.i.51 plus centrifugal enhancement). Furthermore, a higher m.o.i. (2.5 and 5) did not lead to more abundant viral miRNAs, as determined by Northern blotting (Fig. S1).

Table 1. Distribution of small-RNA reads in different libraries

Viral miRNA Viral pre-miRNAs Mouse miRNA Mouse pre-miRNAs Mouse tRNAs Mouse rRNAs Mouse and MCMV genome Total number of reads

1538

Mock-infected cells

MCMV-infected cells

Ago1 IP

Ago2 IP

Fraction of miRNAs

0.00 % 0.00 % 70.76 % 76.13 % 0.64 % 0.58 % 77.05 % 3 908 549

7.83 % 9.86 % 52.89 % 58.68 % 1.51 % 0.72 % 71.91 % 2 850 603

1.35 % 1.62 % 21.53 % 24.62 % 1.85 % 47.97 % 74.72 % 4 741 416

0.80 % 1.28 % 36.93 % 40.95 % 1.10 % 32.73 % 75.85 % 3 675 192

12.82 % – 87.18 % – – – – 1 731 426

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

Journal of General Virology 93

Impact of MCMV infection on the host miRNA system

(b)

20

miR-7a

15

10

miR-26a

5 miR-27a

0

5

10

15

20

10

10– Ct values MCMV

Read numbers log2 MCMV

(a)

8 miR-7a 6

4 miR-26a 2

miR-27a

0

2

Read numbers log2 mock

4

For the m01-1 stem–loop, a new miRNA was found, designated m01-1-3p (Fig. 2a; structure calculated using Mfold, Zuker, 2003), following previous conventions (Do¨lken et al., 2007; Pfeffer et al., 2005). It was detected in almost equal numbers to the miRNA m01-1 on the opposite strand (1511 vs 1758 reads), which in this study is referred to as m01-1-5p. This miRNA was also observed in a recent study of miRNAs associated with tagged Ago2 following MCMV infection (Libri et al., 2012). Its expression could be confirmed by qPCR by 24 h p.i. at slightly lower levels than that of m01-1-5p (Fig. 2b) and showed similar levels at both 48 and 72 h p.i., whereas m01-1-5p increased in expression up to 72 h p.i. Expression of both m01-1-3p and m01-1-5p could also be confirmed by Northern blotting (Fig. 2c). Analysis of Solexa libraries after RISC IP To check for sorting of viral miRNAs into RISCs, we used RNA isolated after RISC IP with antibodies specific for either Ago1 or Ago2 to generate small-RNA libraries. To show that the antibodies used were specific, GFP-tagged Ago1 and Ago2 were precipitated with anti-GFP beads and detected by Western blotting with the antibodies against Table 2. Fold differences compared with mock infection for miR-7a, miR-27a and miR-26a

miR-7a miR-27a miR-26a

http://vir.sgmjournals.org

12.95 0.03 0.10

8

10– Ct values mock

All previously annotated MCMV miRNAs (Do¨lken et al., 2007) could be accounted for, with the exception of m881*, which was not found in any library (Table S3). Whilst some reads also mapped to non-hairpin regions of the MCMV genome (0.67 % of reads), these reads were spread over the entire genome at low numbers and did not show any additional clusters that could be folded into stem–loop structures (data not shown).

Solexa reads for total RNA

6

qPCR data for total RNA 12.67 0.08 0.22

10

Fig. 1. Mouse miRNAs in mock-infected versus MCMV-infected cells. (a) Read numbers (log2) from mock- and MCMV-infected Solexa libraries. (b) Quantitative real-time PCR (qPCR) analysis of total RNA isolated from NIH 3T3 cells infected with MCMV (m.o.i.51) at 72 h p.i. and from mock-infected cells. Ct values (n53) were normalized to controls. Ct values are shown in reverse order to display lower to higher expression levels, comparable with panel (a). miR-7a, miR-26a and miR-27a are indicated by arrows.

Ago1 and Ago2 that were used for the RISC IP. No crossreaction in Western blot detection between Ago1 and Ago2 was observed. For the commercially available Ago1 antibody, the reverse reaction proved to be efficient and specific (Fig. 3a). The Ago2 antibody has been characterized previously by Zhu et al. (2010). The suitability of both antibodies for immunoprecipitation was also verified by qPCR. Mouse miRNAs were measured after IP using antibodies against Ago1, Ago2 or myc and compared with input values (Fig. 3b). Whilst the Ago1 antibody was more efficient than the Ago2 antibody, both clearly produced a higher recovery level than the control anti-myc antibody. After sequencing of the Ago-associated small RNAs, 4.7 million reads and 3.6 million reads for the Ago1 and Ago2 libraries, respectively, were mapped against both mouse and MCMV sequences (Table 1). In contrast to what was observed for total RNA, a large number of reads mapped to mouse rRNAs, indicating that the RISC IP co-precipitated ribosomes presumably associated with target mRNAs. All detected miRNAs of viral or endogenous origin were loaded in both Ago1- and Ago2-containing RISCs to a similar extent (linear regression of Ago1 vs Ago2 loading: R250.88 for mouse miRNAs, R250.89 for MCMV miRNAs, Fig. 3c and Table S3). Viral miRNAs accounted for only 6 and 2 % (Ago1 IP and Ago2 IP, respectively) of all loaded miRNAs as opposed to the 13 % in total RNA. The qPCR data after IP with either Ago1 or Ago2 antibodies also showed essentially equal values for Ago1 and Ago2 loading (Fig. 3d) and was performed for the same mouse miRNAs as in Fig. 1(b), as well as all MCMV miRNAs (Table S2). A comparison of the qPCR results between input and IP did not confirm the reduced incorporation rate of MCMV miRNAs, possibly due to experimental noise, which may well have been within the range of the observed difference. Conservatively, we concluded that viral miRNAs are not preferentially incorporated into RISC. Kinetics of miR-7a and miR-26a expression changes To analyse the expression of miR-7a and miR-26a over the time course of infection in fibroblasts as well as in a different cell line, RNA was isolated from NIH 3T3 fibroblasts and

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

1539

A. Dittmer and K. Fo¨rstemann

(a)

(b) 12

m01-1-3p m01-1-5p

miR-m01-1 hairpin

15– Ct norm. to U6

10

m01-1-3p

m01-1-5p

8 6 4 2

24 h p.i.

48 h p.i.

72 h p.i.

m01-1-3p

m01-1-5p

U6 snRNA

TCMK-1 epithelial cells at 24, 48 and 72 h p.i. and from mock-infected cells. We then performed Northern blotting to detect miR-7a and miR-26a; miR-27a served as a control for a differentially expressed miRNA and miR-16 was used for normalization as its level does not change following MCMV infection (Fig. 4a). Quantitative analysis of three independent Northern blots showed that miR-7a was slightly upregulated at 24 h p.i. but was strongly upregulated from 48 h p.i. onwards in both NIH 3T3 and TCMK-1 cells. In contrast, miR-26a displayed a steady downregulation in NIH 3T3 cells but was marginally upregulated in TCMK-1 cells (Fig. 4b). The qPCR results for miRNA expression in TCMK-1 cells confirmed these results (data not shown). Therefore, miR26a downregulation appeared to be cell-type dependent.

i. h

p.

p. 72

48

h

p. h 24

M oc k

i.

i.

(c)

Fig. 2. Expression of MCMV miR-m01-1. (a) Schematic of the m01-1 hairpin with the novel m01-1-3p sequence GUCAGACUUUAUUCUCUUCUCU. (b) qPCR analysis of total RNA isolated from NIH 3T3 cells infected with MCMV (m.o.i.51) at 24, 48 and 72 h p.i. (c) Northern blots probed with m01-1-3p, m01-15p and U6 small nuclear RNA (snRNA). RNA was isolated from NIH 3T3 cells infected with MCMV (m.o.i.51) at 24, 48 and 72 h p.i. or mock infected.

miRNA changes was performed at an m.o.i. of 10 at 24 h p.i., as later time points could not be analysed due to the toxicity of CHX. As described previously, miR-27a was rapidly downregulated under these conditions, but no change in miR-7a and miR-26a levels occurred (Fig. 5c, d). The upregulation of miR-7a could be the result of de novo transcription or an extended half-life of the miRNA. Mapping of the Solexa libraries to the two miR-7a-1 and miR-7a-2 precursors revealed that only miR-7a-1 was transcribed. The miR-7a-1 hairpin is located in an intron of the hnrnpk gene, whilst miR-7a-2 is intergenic. qPCR analysis of hnrnpk mRNA showed no upregulation during MCMV infection (data not shown), indicating that pri-miR-7a transcription was specifically induced and was not a secondary consequence of a change in expression of the host gene.

Requirements of miR-7a and miR-26a regulation

Effects of miRNA mimics or inhibitors on virus replication

Downregulation of miR-27a is a post-transcriptional event mediated through a viral RNA that contains a single miR27-binding site (Buck et al., 2010; Libri et al., 2012; Marcinowski et al., 2012). For miR-26a, a similar mechanism is possible, whilst there are presumably different requirements for the upregulation of miR-7a. UV-inactivated MCMV did not induce changes in miRNA levels either by qPCR or by Northern blot analysis at an m.o.i. of 1 at 72 h p.i. (Fig. 5a, b), indicating that virus entry is not sufficient. Treatment with PAA, which specifically blocks the viral DNA polymerase, showed that amplification of the viral genome is necessary for the full extent of upregulation, consistent with a late effect (Fig. 5a, b).

The strong upregulation of miR-7a may influence MCMV infection similar to the antiviral effect demonstrated for miR-27a (Buck et al., 2010). To test for the effect of miR-7a inhibition and overexpression on virus replication in cell culture, we transfected NIH 3T3 cells with mimics or inhibitors for miR-7a, miR-26a and miR-27a, as well as with a corresponding unrelated control. Cells were then infected with MCMV and virus yield was determined (Fig. 6a, b). We did not observe significant changes relative to the control for any miRNA inhibitor or mimic. Therefore, the tested miRNAs do not appear to significantly modulate MCMV replication in cell culture as assayed by our approach.

Treatment with cycloheximide (CHX) blocks protein synthesis but allows viral RNAs to be transcribed. Analysis of 1540

To verify that the mimic and inhibitor were functional upon transfection, NIH 3T3 cells were co-transfected with 50 nM mimic miR-7, mimic plus inhibitor miR-7 or a

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

Journal of General Virology 93

pE G FP pE G FP pE -Ag o G FP 2 -A go 1

Impact of MCMV infection on the host miRNA system

(a) IP - GFP Input

(b) 25

130 kDa 30 kDa

20 40– Ct values IP

WB -Ago1

15

WB -Ago2

10

pE G FP pE G FP pE -Ag o G FP 2 -A go 1

WB -GFP

myc Ago1 Ago2

Input

5 0 0

130 kDa 30 kDa

5 10 15 20 40– Ct values input

25

IP -Ago1 IP -Ago2 (d)

(c)

15

Mouse miRNAs MCMV miRNAs

15

15- Ct-Values IP Ago2

Read numbers (log2) IP Ago2

20

10

5

0

5

10

15

20

Read numbers (log2) IP Ago1

MCMV miRNAs

10

5

0 0

Mouse miRNAs

0

5

10

15

15– Ct values IP Ago1

Fig. 3. Analyses of RISC IP. (a) NIH 3T3 cells were transfected with GFP-Ago1, GFP-Ago2 or GFP alone for 48 h and IP was performed on an equal amount of cell lysate with GFP-Trap beads or Ago1 and Ago2 antibodies. Western blots were performed with antibodies against Ago1, Ago2 or GFP alone. Input was 10 % of the immunoprecipitate. (b) qPCR data of mouse miRNAs immunoprecipitated with anti-Ago1, anti-Ago2 or anti-myc versus input. Arrows indicate controls (U6, MettRNA). (c, d) Solexa read numbers (log2) (c) and PCR Ct values (d) normalized to controls for both mouse and MCMV miRNAs loaded into Ago1 versus Ago2. Ct values for (b) and (d) are shown in reverse order for easy comparison with (c).

control together with a reporter plasmid containing four perfect miR-7a-binding sites or an unrelated sequence in the 39 untranslated region (UTR) of GFP. The miR-7a mimic reduced GFP expression by 96 % in comparison with the control (Fig. 6c). If both the miR-7a mimic and the miR-7a inhibitor were transfected, GFP expressed was only reduced by 73 % (Fig. 6c). Neither the miR-7a mimic nor the mimic plus inhibitor combination had any effect on the control plasmid. Thus, both the miR-7a mimic and the inhibitor had readily observable and specific effects in a miR-7a reporter system. We determined the transfection efficiency to be 55±5 % (mean±SD, n57) in our http://vir.sgmjournals.org

experiments with flow cytometry using the control reporter. As the GFP expression plasmid must reach the nucleus whereas the mimic and inhibitor act in the cytoplasm, we assumed that the transfection efficiency was higher for these compounds. This has been observed in Drosophila S2 cells where transfection efficiencies with plasmids using lipid transfection reagents do not exceed 70 %, but a stably integrated, clonal miRNA reporter can be de-repressed in essentially all cells following transfection of an inhibitor (Fo¨rstemann et al., 2007). Thus, we estimated that a large fraction of the cells was loaded with the miRNA mimic or inhibitor in our experiments.

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

1541

h

p. i.

72

h

p. i.

p. i. 48

24

72

h

M oc k

p. i. h

h

(a)

48

M oc k 24 h p.

i.

p. i.

A. Dittmer and K. Fo¨rstemann

miR-7a

miR-27a

miR-26a

miR-16 NIH-3T3

TCMK-1

(b)

2.0

miR-7a NIH 3T3

i. p.

p. h

h

M

72

oc

h

k

p.

i. p.

p. 48

h

p. 24

h

k oc M

h

0.0

72

0

i.

0.5

i.

4

i.

1.0

48

8

miR-27a NIH 3T3 miR-26a NIH 3T3 miR-27a TCMK-1 miR-26a TCMK-1

1.5

Fold change

12

i.

Fold change

miR-7a TCMK-1

24

16

Fig. 4. miR-7a and miR-26a expression kinetics. (a) Northern blots probed with miR7a, miR-27a, miR-26a and miR-16. RNA was isolated from NIH 3T3 or TCMK-1 cells infected with MCMV (m.o.i.51) at 24, 48 and 72 h p.i. or mock infected. (b) Quantitative analysis of Northern blots (n53) for upregulated miRNAs (left graph) and downregulated miRNAs (right graph) in both NIH 3T3 and TCMK-1 cells. Signal intensity was normalized to miR-16 and fold change compared with mock values was calculated.

DISCUSSION

Mapping of the reads obtained at 72 h p.i. showed that only ~13 % of all miRNA reads were of viral origin (Table 1). This is in contrast to the observations made by Do¨lken et al. (2007), who demonstrated 61 % viral miRNAs late after infection using a small-RNA cloning approach. It has

In this study, we used a deep-sequencing approach to gain further insight into the small RNAs expressed in mouse fibroblasts during MCMV infection.

MCMV/UV

(a)

PA V C

M

M M

M

oc M

C

k

V/ M M

C M

C

M

k oc M

4

V/

UV

(b)

V

MCMV

A

MCMV/PAA

10

miR-7a Fold change

3 U6 2

1

M C M

M

C

M

V

V/

C (d)

M

MCMV

k/

MCMV+CHX

C

H

X

H

X

m

m

iR

-1 6

7a -2 iR

-2

4

oc

(c)

m

m

iR

iR

-7

a

6a

0

Fold change

3 miR-7a 2 U6 1

1542

m

iR

-1

6

7a m

iR

-2

-2 iR m

m

iR

-7

a

6a

0

Fig. 5. miR-7a regulation requires protein synthesis. (a) RT-PCR quantification of the indicated miRNAs for NIH 3T3 cells at 72 h p.i. infected with UV-inactivated MCMV, MCMV treated with 100 mg phosphonoacetic acid (PAA) ml”1 or untreated MCMV (m.o.i.51). The fold change relative to mock-infected cells are shown. (b) Northern blot analysis of miR7a accumulation at 72 h p.i. with UV-inactivated (left panel) or PAA-treated (right panel) MCMV. U6 snRNA served as a loading control. (c) RT-PCR quantification relative to mock infection of the indicated miRNAs for NIH 3T3 cells at 24 h p.i. (m.o.i.510) with or without CHX treatment (100 mg ml”1). Mock cells also received CHX treatment. (d) Northern blot analysis of miR-7a accumulation at 24 h p.i. (m.o.i.510) with or without CHX treatment. U6 snRNA served as a loading control.

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

Journal of General Virology 93

Impact of MCMV infection on the host miRNA system

70

Cells infected (%)

60

(c) I miR-7 I miR-27a miR-7 reporter

I miR-26a

50

I nc

Mean GFP fluorescence

(a)

MCMV

40 30 20 10 0

24 h p.i.

48 h p.i.

72 h p.i.

96 h p.i.

72 h p.i.

96 h p.i.

Control reporter

120 100 80 60 40 20 0

Mimic

(b) 70 60

M miR-7 M miR-27a

Cells infected (%)

M miR-26a

50

M nc MCMV

40 30 20 10 0

24 h p.i.

48 h p.i.

been suggested that MCMV miRNAs may outcompete and thereby disrupt global cellular miRNA expression, given that Kaposi’s sarcoma-associated herpesvirus miRNAs also are highly overexpressed (Umbach & Cullen, 2010) and adenoviral virus-associated RNAs have been implicated as competitive inhibitors for the miRNA pathway components Exportin-5, Dicer and RISC (Andersson et al., 2005; Lu & Cullen, 2004; Xu et al., 2007). Our numbers concerning MCMV miRNAs both in total RNA and after RISC IP, however, did not confirm this hypothesis. Using a higher m.o.i. in Northern blot experiments did not lead to higher expression levels of the viral miRNA m01-4, which was the most abundant miRNA in the study by Do¨lken et al. (2007) (Fig. S1). We could confirm expression of all 27 mature MCMV miRNAs described by Do¨lken et al. (2007) except for m881* (Table S3). However, there were more than 100 reads of a small fragment that mapped 12 nt upstream and partially overlapping with m88-1* (data not shown). Furthermore, expression of the other three MCMV miRNAs that had only been described by one single clone could easily be verified. We therefore cannot confirm m88-1* as a MCMV miRNA based on our data. We found one additional miRNA on the m01-1 hairpin, whose presence could be confirmed by qPCR and Northern blotting (Fig. 2). The newly designated m01-1-3p was expressed at similar levels to m01-1-5p. This miRNA was also described in a recent deep-sequencing study (Libri et al., 2012) but was not verified by alternative detection methods. In addition, a small number of reads were distributed along the entire MCMV genome, but there were no further clusters that could be folded into hairpin structures and the majority of these reads were only 18–20 nt. http://vir.sgmjournals.org

Mimic+inhibitor

Fig. 6. Cellular miRNAs do not appreciably affect MCMV replication in cell culture. (a, b) Yield assay for NIH 3T3 cells transfected with 50 nM inhibitor (I) (a) or mimic (M) (b) for miR7a, miR-26a, miR-27a, a control miRNA (nc) or blank. Cells were infected at 24 h after transfection at an m.o.i. of 1 with GFP–MCMV, and supernatant was collected over a period of 4 days and used for re-infection of fresh NIH 3T3 cells. The proportion of cells infected in this second round was quantified by flow cytometry (GFP expression) of 10 000 cells. (c) Control of miR-7 mimic and inhibitor efficiency. NIH 3T3 cells were co-transfected with 50 nM mimic, mimic plus inhibitor or control miRNA together with the reporter pEGFP-miR-7 containing four miR-7a-binding sites in the 39 UTR of GFP or pEGFP control. GFP-positive cells were quantified by flow cytometry and set relative to scramble miRNA.

Read numbers for the miRNAs varied over three orders of magnitude (Table S3), which is in the same range as the numbers from the previous small-RNA cloning approach (Do¨lken et al., 2007). However, abundance based on the available numbers from both datasets varied considerably. The most abundant MCMV miRNA in our deep-sequencing library, m108-1, which represented ~30 % of all viral miRNA reads, was detected at a lower frequency with the classical cloning strategy, whereas the opposite was true for M23-2. In addition, m01-2* came up with fivefold more reads than the proper m01-2 miRNA. Recently, Raabe et al. (2011) investigated the small-RNA transcriptome of the human pathogen Vibrio cholerae and noted considerable differences depending on the technique employed. They identified several steps in library preparation that might introduce bias and advocated caution in comparisons of absolute numbers between alternatively processed RNA samples. Alon et al. (2011) also reported that multiplexing strategies with bar codes introduced through adaptor ligation, as used in our protocol, can confer significant bias in miRNA expression profiles. As small-RNA cloning and sequencing by traditional means is equally prone to these biases but, in addition, has limitations due to the low sequence coverage, differences in the abundance rank of single miRNAs can be expected and some miRNAs such as m01-1-3p may be missed entirely. However, this does not explain the large gap between the cumulative abundance of viral miRNAs in both datasets (64 vs 13 %). Our qPCR data confirmed the notion that MCMV miRNAs do not constitute the majority of the small-RNA population at 72 h p.i. (e.g. see Fig. 3d). Furthermore, increasing the m.o.i. did not result in a substantially increased production of viral miRNAs (Fig. S2). HCMV miRNAs as measured by

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

1543

A. Dittmer and K. Fo¨rstemann

a deep-sequencing approach accounted for 30 % of all miRNAs at an m.o.i. of 3 (Stark et al., 2012). Even in this case, the viral miRNAs did not appear to overwhelm the host miRNA system. The effectors of miRNA-mediated repression are the RISCs, in which the Ago proteins (Pfeffer et al., 2005) associate with the small RNAs. Mammalian cells encode four Ago-like subfamily proteins (Ago1–Ago4) with mostly redundant functions that also associate equally with miRNAs, even though only Ago2 can cleave target mRNAs (Azuma-Mukai et al., 2008; Liu et al., 2004; Meister et al., 2004; Su et al., 2009). In contrast to Drosophila, where small RNAs are actively sorted based on differences in structure (Czech et al., 2009; Fo¨rstemann et al., 2007; Ghildiyal et al., 2010; Okamura et al., 2009; Tomari et al., 2007), there seems to be no specific sorting mechanism in mammalian cells. However, it has also been reported that Ago1 and Ago2 but not Ago3 and Ago4 possess strand-dissociating activity of miRNA duplexes and function as RNA chaperones (Wang et al., 2009a) and that, at least for endogenous small hairpin RNAs, thermodynamic stability greatly influences the loading process (Gu et al., 2011). To check to what extent viral miRNAs are loaded into RISCs, we performed RISC IP with non-cross-reactive antibodies against Ago1 or Ago2 (Fig. 3a) and then generated small-RNA libraries as well as qPCR data. All MCMV miRNAs could be in found in both Ago1- and Ago2-containing RISCs (Table S3). Whilst Solexa reads showed some differences (Fig. 3b), qPCR values correlated well between Ago1 and Ago2 (Fig. 3c), showing that MCMV miRNAs are also similarly loaded into both Agos. For the host miRNAs, we could show that Solexa read numbers correlated well between mock- and MCMVinfected cells, although a few miRNAs showed a greater than tenfold discrepancy (Fig. 1a). The same is also true for both HCMV lytic and latent infections, where only a few cellular miRNAs are altered substantially (Poole et al., 2011; Stark et al., 2012). qPCR and analysis of Northern blots confirmed the regulation of miR-7a and miR-26a in NIH 3T3 fibroblasts, but whilst miR-7a was strongly upregulated in epithelial TCMK-1 cells, miR-26a was marginally upregulated rather than downregulated (Fig. 4b). MCMV can infect many different cell types in vivo, and results by Buck et al. (2010) indicated that miR-27a regulation is similarly pervasive. However, our findings with miR-26 indicated that cell type-specific effects can also occur. Experiments with UV-inactivated virus, PAA and CHX (Fig. 5) showed that miR-7a upregulation depends on virus replication and protein synthesis. Many examples of changes in cellular miRNA expression caused by virus infection are known, and both induction and repression of miRNAs can be due directly to viral factors or indirectly to an activated host-cell response. For example, papillomavirus E6 destabilizes the tumour suppressor p53, a known miR-34a 1544

transactivator (Wang et al., 2009b), whilst Epstein–Barr virus LMP1 activates the miR-146a promoter, leading to modulation of lymphocyte signalling pathways (Cameron et al., 2008). In some cases, however, the cause of miRNA regulation remains unclear, such as the suppression of miR17-5p and miR-20a by human immunodeficiency virus type 1 (HIV-1) infection (Triboulet et al., 2007), which targets a potential co-factor of the HIV-1 Tat transactivator, and of miR-100 and miR-101 by HCMV (Wang et al., 2008), which are part of the TOR signalling pathway. The viral protein responsible for miR-7a upregulation needs to be determined in further experiments, although the requirement for virus replication to amplify the effect of miR-7a argues for a process driven by high levels of viral protein rather than triggering of a cellular gene expression switch. So far, there are no confirmed targets for mouse miR-7a. Human miR-7, which is identical to mouse miR-7a, has been widely implicated as a tumour suppressor in lung cancer, glioblastoma and breast cancer cells, mostly through downregulation of the epidermal growth factor receptor (EGFR) (Kefas et al., 2008; Lee et al., 2011; McInnes et al., 2012; Xiong et al., 2011). Furthermore, a recent report by Melnick et al. (2011) showed that inhibition of EGFR pathway signalling attenuates MCMV-induced pathogenesis in organ culture. However, EGFR expression in NIH 3T3 cells transfected with a miR-7a mimic or infected with MCMV showed no detectable difference in Western blot analysis (Fig. S2). Human miR-7 has been shown to regulate several critical signalling pathways, for example by targeting focal adhesion kinase, which is essential for the Ras-ERK 1/2 signalling pathway (Wu et al., 2011), and by targeting PIK3CD, mTOR and p70S6K to regulate the PI3K/Akt pathway (Fang et al., 2012). Interestingly, whilst upregulation of miR-7 was not described following HCMV infection (Stark et al., 2012; Wang et al., 2008), other cellular miRNAs that interfere with the mTor pathway were affected (Wang et al., 2008). We could not demonstrate an effect of the miR-7a mimic or inhibitor on MCMV replication efficiency in cell culture, despite the fact that we could demonstrate their functionality in a reporter system (Fig. 6). The potential significance of miR-7 induction may therefore be related to virus–host interactions beyond the replication and assembly steps that are accessible in cell culture. We also could not reproduce the moderate antiviral effect of miR-27a that was reported previously (Buck et al., 2010). However, this is in line with a screen of large deletion mutants of MCMV, where viruses missing the region containing the m169 transcript responsible for miR-27 degradation showed no attenuation in NIH 3T3 fibroblasts (Marcinowski et al., 2012; Mohr et al., 2008). Technical differences (yield assay vs plaque assay with different cell lines used for the titre measurements and differences in transfection and infection protocols, as well as different m.o.i. between our study and the experiments by Buck et al., 2010) may also be responsible for this discrepancy. Our yield assay could resolve a twofold

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

Journal of General Virology 93

Impact of MCMV infection on the host miRNA system

difference in viral titre within technical replicates (data not shown), but we certainly cannot exclude the possibility that more subtle effects of miR-7a or miR-27a inhibitors/ mimics have remained undetected in our assay. As miR-7a is induced following virus infection, cell lines derived from miR-7a knockout mouse embryos should be more informative but are currently unavailable. One of the future challenges will be to study the importance of cellular miRNAs for controlling virus infections in vivo. Viral mutants with an impaired ability to induce miR-7a expression may be a tool for this purpose, but this will probably also blunt other virus-induced changes in gene expression. A more specific tool could be an engineered virus that expresses a miR-7a inhibitor (a so-called miRNA ‘sponge’) (Ebert et al., 2007).

METHODS Cell lines and viruses. NIH 3T3 and TCMK-1 cells were grown in

Dulbecco’s minimal essential medium supplemented with 10 % FBS, 100 U penicillin ml21 and 0.1 mg streptomycin ml21. Wild-type MCMV (Smith strain) with an inserted GFP gene under the control of a CMV promoter was generated as described by Wagner et al. (1999). Virus stocks were prepared in M2-10B4 cells as described by Do¨lken et al. (2007) and titres were determined by counting GFP-positive cells. Cells were infected at a calculated m.o.i. of 1 (CHX assays m.o.i.510) with centrifugal enhancement (30 min, 800 g). This infection protocol was referred to as an effective m.o.i. of 10 in the study by Do¨lken et al. (2007). The efficiency of infection was monitored by GFP fluorescence assessed by flow cytometry or fluorescence microscopy. RNA preparation and Northern blotting. RNA was extracted using

Trizol reagent (Invitrogen) and Northern blotting was performed on 20 mg total RNA, as described previously (Fo¨rstemann et al., 2005). DNA probes were 59-labelled with [c-32P]ATP and polynucleotide kinase (Fermentas). Blots were analysed and quantified by phosphorimaging using a Typhoon 9400 Imager (Amersham Biosciences) and Multi Gauge v3.0 (Fujifilm). Immunoprecipitation and immunoblotting. Cell lysis, immuno-

by Czech et al. (2008) (a detailed protocol is available on request). Solexa sequencing was carried out at Fasteris (Plan-Les-Ouates, Switzerland) and the MPI for Molecular Genetics (Berlin). Output files were mapped onto the target sequences using BOWTIE (Langmead et al., 2009) with the option 2n0 to force selection of perfectly matching sequences only. Pre-processing of sequences and analysis of the BOWTIE output files were carried out using PERL scripts (available on request). tRNA, rRNA and MCMV genome sequences were obtained from GenBank (http://www.ncbi.nlm.nih.gov/Genbank), and miRNA sequences were extracted from miRBase (http://www. mirbase.org/). miRNA quantification by qPCR. RNA from cell lysates (1 mg per

reaction) and immunoprecipitates (entire precipitate) was reverse transcribed using a miScript Reverse Transcription kit (Qiagen). The obtained cDNA (1 ml of a 1 : 5 dilution) was then PCR amplified using a miScript SYBR Green PCR kit (Qiagen) in a 10 ml reaction with the mature miRNA sequence as forward primer (Table S2). PCRs were subjected to 15 min of 94 uC hot-start enzyme activation, and 40 cycles of 94 uC denaturation for 20 s, 55 uC annealing for 30 s and 70 uC elongation for 30 s on an ABI Prism 7000 SDS. Ct values were normalized either to U6 snRNA as an endogenous reference or to the mean value of five different controls [glyceraldehyde 3-phosphate dehydrogenase (GAPDH), U6 snRNA, 5.8S rRNA, Met-tRNA and Val-tRNA) as indicated in figure legends. Values were compared directly or fold change was calculated using the 22DDCt method (Livak & Schmittgen, 2001). Viral yield assay and miRNA transfections. NIH 3T3 cells were

seeded at 56104 cells per well and transfected with miRNA mimics or inhibitors (50 nM) using Fugene HD Transfection Reagent. Cells were incubated 24 h prior to infection with GFP–MCMV. Supernatant from the infected cells was collected over a period of 4 days p.i., and viral yield was determined by re-infection of NIH 3T3 cells and quantification of GFP expression by flow cytometry (BD FACScan flow cytometer; Becton Dickinson). For controls of mimic and inhibitor effects, oligonucleotides comprising four repeats of a perfect miR-7a-binding site or four repeats of an unrelated sequence were cloned into pEGFP-N1 using the NotI site in the GFP 39 UTR yielding pEGFP-mir-7-reporter and pEGFP-control-reporter.

ACKNOWLEDGEMENTS

precipitation and Western blotting were performed essentially as described by Hartig & Fo¨rstemann (2011). For RISC IP, 50 ml Protein G Plus/Protein A agarose (Calbiochem) was incubated overnight with 2.5 ml anti-Ago1 mAb (clone 2A7; Wako Chemicals) or 50 ml antiAgo2 mAb (clone 6F4; Zhu et al., 2010). As a negative control, an anti-c-myc antibody was used (clone 9E10; Santa Cruz Biotechnology). Beads were washed three times with 500 ml lysis buffer [30 mM HEPES (pH 7.4), 100 mM KAc, 2 mM MgAc, 1 mM DTT] containing 1 % (v/v) Triton X-100 (Sigma), incubated with 1 mg cell extract for 1 h at 4 uC and washed again.

We thank Lars Do¨lken for MCMV inoculum as well as M2-10B4 and TCMK-1 cells, Gunther Meister for the 6F4 anti-Ago2 antibody and all members of the Nationales Genomforschungsnetz (NGFN) consortium on ‘Pathogenic roles of herpesviral miRNAs’, as well as Susanne Bailer for stimulating discussions and critical comments on the manuscript. This work was supported by the NGFN funded through the BMBF (FKZ: 01GS0801). K. F. is the recipient of a Human Frontiers Science Program Career Development Award.

For Western blotting, Ago1 and Ago2 were cloned into pEGFP-N1 (Clontech) and transfected into NIH 3T3 cells using Fugene HD Transfection Reagent (Roche). GFP fusion constructs were precipitated using GFP-Trap_A beads (Chromotek) or anti-Ago1 and antiAgo2 antibody as described above. Primary antibodies were diluted 1 : 1000 for anti-Ago1, 1 : 10 for anti-Ago2 and 1 : 500 for anti-GFP antibody (Santa Cruz Biotechnology). For detection of EGFR in transfected or infected cells EGFR rabbit mAb (clone C74B9; Cell Signalling) was used at a 1 : 1000 dilution.

REFERENCES

Generation of small-RNA libraries. Small RNAs were isolated, and

Azuma-Mukai, A., Oguri, H., Mituyama, T., Qian, Z. R., Asai, K., Siomi, H. & Siomi, M. C. (2008). Characterization of endogenous

libraries for Solexa sequencing were prepared essentially as described http://vir.sgmjournals.org

Alon, S., Vigneault, F., Eminaga, S., Christodoulou, D. C., Seidman, J. G., Church, G. M. & Eisenberg, E. (2011). Barcoding bias in high-

throughput multiplex sequencing of miRNA. Genome Res 21, 1506– 1511. Andersson, M. G., Haasnoot, P. C., Xu, N., Berenjian, S., Berkhout, B. & Akusja¨rvi, G. (2005). Suppression of RNA interference by

adenovirus virus-associated RNA. J Virol 79, 9556–9565.

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

1545

A. Dittmer and K. Fo¨rstemann human Argonautes and their miRNA partners in RNA silencing. Proc Natl Acad Sci U S A 105, 7964–7969.

Gu, S., Jin, L., Zhang, F., Huang, Y., Grimm, D., Rossi, J. J. & Kay, M. A. (2011). Thermodynamic stability of small hairpin RNAs highly

Bartel, D. P. (2004). MicroRNAs: genomics, biogenesis, mechanism,

influences the loading process of different mammalian Argonautes. Proc Natl Acad Sci U S A 108, 9208–9213.

and function. Cell 116, 281–297. Biggar, K. K. & Storey, K. B. (2011). The emerging roles of

microRNAs in the molecular responses of metabolic rate depression. J Mol Cell Biol 3, 167–175.

Hartig, J. V. & Fo¨rstemann, K. (2011). Loqs-PD and R2D2 define independent pathways for RISC generation in Drosophila. Nucleic Acids Res 39, 3836–3851.

Buck, A. H., Santoyo-Lopez, J., Robertson, K. A., Kumar, D. S., Reczko, M. & Ghazal, P. (2007). Discrete clusters of virus-encoded

He, L. & Hannon, G. J. (2004). MicroRNAs: small RNAs with a big

microRNAs are associated with complementary strands of the genome and the 7.2-kilobase stable intron in murine cytomegalovirus. J Virol 81, 13761–13770.

Jopling, C. L., Yi, M., Lancaster, A. M., Lemon, S. M. & Sarnow, P. (2005). Modulation of hepatitis C virus RNA abundance by a liver-

specific microRNA. Science 309, 1577–1581.

Buck, A. H., Perot, J., Chisholm, M. A., Kumar, D. S., Tuddenham, L., Cognat, V., Marcinowski, L., Do¨lken, L. & Pfeffer, S. (2010). Post-

Kefas, B., Godlewski, J., Comeau, L., Li, Y., Abounader, R., Hawkinson, M., Lee, J., Fine, H., Chiocca, E. A. & other authors (2008). microRNA-

transcriptional regulation of miR-27 in murine cytomegalovirus infection. RNA 16, 307–315.

7 inhibits the epidermal growth factor receptor and the Akt pathway and is down-regulated in glioblastoma. Cancer Res 68, 3566–3572.

Bueno, M. J. & Malumbres, M. (2011). MicroRNAs and the cell cycle.

Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. (2009).

Biochim Biophys Acta 1812, 592–601.

Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol 10, R25.

Cameron, J. E., Yin, Q., Fewell, C., Lacey, M., McBride, J., Wang, X., Lin, Z., Schaefer, B. C. & Flemington, E. K. (2008). Epstein–Barr virus

role in gene regulation. Nat Rev Genet 5, 522–531.

Lee, Y., Jeon, K., Lee, J.-T., Kim, S. & Kim, V. N. (2002). MicroRNA

latent membrane protein 1 induces cellular microRNA miR-146a, a modulator of lymphocyte signaling pathways. J Virol 82, 1946–1958.

maturation: stepwise processing and subcellular localization. EMBO J 21, 4663–4670.

Cazalla, D., Yario, T. & Steitz, J. A. (2010). Down-regulation of a host

Lee, K. M., Choi, E. J. & Kim, I. A. (2011). microRNA-7 increases

microRNA by a Herpesvirus saimiri noncoding RNA. Science 328, 1563–1566.

radiosensitivity of human cancer cells with activated EGFR-associated signaling. Radiother Oncol 101, 171–176.

Czech, B., Malone, C. D., Zhou, R., Stark, A., Schlingeheyde, C., Dus, M., Perrimon, N., Kellis, M., Wohlschlegel, J. A. & other authors (2008). An endogenous small interfering RNA pathway in Drosophila.

Libri, V., Helwak, A., Miesen, P., Santhakumar, D., Borger, J. G., Kudla, G., Grey, F., Tollervey, D. & Buck, A. H. (2012). Murine

Nature 453, 798–802.

cytomegalovirus encodes a miR-27 inhibitor disguised as a target. Proc Natl Acad Sci U S A 109, 279–284.

Czech, B., Zhou, R., Erlich, Y., Brennecke, J., Binari, R., Villalta, C., Gordon, A., Perrimon, N. & Hannon, G. J. (2009). Hierarchical rules

Liu, J., Carmell, M. A., Rivas, F. V., Marsden, C. G., Thomson, J. M., Song, J.-J., Hammond, S. M., Joshua-Tor, L. & Hannon, G. J. (2004).

for Argonaute loading in Drosophila. Mol Cell 36, 445–456.

Argonaute2 is the catalytic engine of mammalian RNAi. Science 305, 1437–1441.

Do¨lken, L., Perot, J., Cognat, V., Alioua, A., John, M., Soutschek, J., Ruzsics, Z., Koszinowski, U., Voinnet, O. & Pfeffer, S. (2007). Mouse

cytomegalovirus microRNAs dominate the cellular small RNA profile during lytic infection and show features of posttranscriptional regulation. J Virol 81, 13771–13782. Du, T. & Zamore, P. D. (2005). microPrimer: the biogenesis and

function of microRNA. Development 132, 4645–4652. Ebert, M. S., Neilson, J. R. & Sharp, P. A. (2007). MicroRNA sponges:

competitive inhibitors of small RNAs in mammalian cells. Nat Methods 4, 721–726. Fang, Y. X., Xue, J. L., Shen, Q., Chen, J. & Tian, L. (2012). miR-7

Livak, K. J. & Schmittgen, T. D. (2001). Analysis of relative gene

expression data using real-time quantitative PCR and the 22DDCt method. Methods 25, 402–408. Lu, S. & Cullen, B. R. (2004). Adenovirus VA1 noncoding RNA can inhibit small interfering RNA and microRNA biogenesis. J Virol 78, 12868–12876. Marcinowski, L., Tanguy, M., Krmpotic, A., Ra¨dle, B., Lisnic´, V. J., Tuddenham, L., Chane-Woon-Ming, B., Ruzsics, Z., Erhard, F. & other authors (2012). Degradation of cellular mir-27 by a novel,

highly abundant viral transcript is important for efficient virus replication in vivo. PLoS Pathog 8, e1002510.

inhibits tumor growth and metastasis by targeting the PI3K/AKT pathway in hepatocellular carcinoma. Hepatology [Epub ahead of print].

McInnes, N., Sadlon, T. J., Brown, C. Y., Pederson, S., Beyer, M., Schultze, J. L., McColl, S., Goodall, G. J. & Barry, S. C. (2012). FOXP3

Fo¨rstemann, K., Tomari, Y., Du, T., Vagin, V. V., Denli, A. M., Bratu, D. P., Klattenhoff, C., Theurkauf, W. E. & Zamore, P. D. (2005).

and FOXP3-regulated microRNAs suppress SATB1 in breast cancer cells. Oncogene 31, 1045–1054.

Normal microRNA maturation and germ-line stem cell maintenance requires Loquacious, a double-stranded RNA-binding domain protein. PLoS Biol 3, e236.

Meister, G. & Tuschl, T. (2004). Mechanisms of gene silencing by

Fo¨rstemann, K., Horwich, M. D., Wee, L., Tomari, Y. & Zamore, P. D. (2007). Drosophila microRNAs are sorted into functionally distinct

Argonaute complexes after production by Dicer-1. Cell 130, 287–297. Gantier, M. P., McCoy, C. E., Rusinova, I., Saulep, D., Wang, D., Xu, D., Irving, A. T., Behlke, M. A., Hertzog, P. J. & other authors (2011).

Analysis of microRNA turnover in mammalian cells following Dicer1 ablation. Nucleic Acids Res 39, 5692–5703.

double-stranded RNA. Nature 431, 343–349. Meister, G., Landthaler, M., Patkaniowska, A., Dorsett, Y., Teng, G. & Tuschl, T. (2004). Human Argonaute2 mediates RNA cleavage

targeted by miRNAs and siRNAs. Mol Cell 15, 185–197. Melnick, M., Abichaker, G., Htet, K., Sedghizadeh, P. & Jaskoll, T. (2011). Small molecule inhibitors of the host cell COX/AREG/EGFR/

ERK pathway attenuate cytomegalovirus-induced pathogenesis. Exp Mol Pathol 91, 400–410.

Ghildiyal, M., Xu, J., Seitz, H., Weng, Z. & Zamore, P. D. (2010).

Mohr, C. A., CIˆcIˆn-SaIˆn, L., Wagner, M., Sacher, T., Schnee, M., Ruzsics, Z. & Koszinowski, U. H. (2008). Engineering of cytomegalo-

Sorting of Drosophila small silencing RNAs partitions microRNA* strands into the RNA interference pathway. RNA 16, 43–56.

virus genomes for recombinant live herpesvirus vaccines. Int J Med Microbiol 298, 115–125.

1546

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

Journal of General Virology 93

Impact of MCMV infection on the host miRNA system Okamura, K., Liu, N. & Lai, E. C. (2009). Distinct mechanisms for

microRNA strand selection by Drosophila Argonautes. Mol Cell 36, 431–444.

(2007). Suppression of microRNA-silencing pathway by HIV-1 during virus replication. Science 315, 1579–1582. Tuddenham, L. & Pfeffer, S. (2011). Roles and regulation of

Otsuka, M., Jing, Q., Georgel, P., New, L., Chen, J., Mols, J., Kang, Y. J., Jiang, Z., Du, X. & other authors (2007). Hypersusceptibility to

microRNAs in cytomegalovirus infection. Biochim Biophys Acta 1809, 613–622.

vesicular stomatitis virus infection in Dicer1-deficient mice is due to impaired miR24 and miR93 expression. Immunity 27, 123–134.

Umbach, J. L. & Cullen, B. R. (2010). In-depth analysis of Kaposi’s

Ouda, R., Onomoto, K., Takahasi, K., Edwards, M. R., Kato, H., Yoneyama, M. & Fujita, T. (2011). Retinoic acid-inducible gene I-

sarcoma-associated herpesvirus microRNA expression provides insights into the mammalian microRNA-processing machinery. J Virol 84, 695–703.

inducible miR-23b inhibits infections by minor group rhinoviruses through down-regulation of the very low density lipoprotein receptor. J Biol Chem 286, 26210–26219.

Wagner, M., Jonjic, S., Koszinowski, U. H. & Messerle, M. (1999).

Pfeffer, S., Sewer, A., Lagos-Quintana, M., Sheridan, R., Sander, C., Gra¨sser, F. A., van Dyk, L. F., Ho, C. K., Shuman, S. & other authors (2005). Identification of microRNAs of the herpesvirus family. Nat

Wang, F.-Z., Weber, F., Croce, C., Liu, C.-G., Liao, X. & Pellett, P. E. (2008). Human cytomegalovirus infection alters the expression of

Methods 2, 269–276. Plaisance-Bonstaff, K. & Renne, R. (2011). Viral miRNAs. Methods

Mol Biol 721, 43–66. Poole, E., McGregor Dallas, S. R., Colston, J., Joseph, R. S. & Sinclair, J. (2011). Virally induced changes in cellular microRNAs

maintain latency of human cytomegalovirus in CD34+ progenitors. J Gen Virol 92, 1539–1549.

Raabe, C. A., Hoe, C. H., Randau, G., Brosius, J., Tang, T. H. & Rozhdestvensky, T. S. (2011). The rocks and shallows of deep RNA

sequencing: examples in the Vibrio cholerae RNome. RNA 17, 1357– 1366. Sayed, D. & Abdellatif, M. (2011). MicroRNAs in development and

disease. Physiol Rev 91, 827–887. Stark, T. J., Arnold, J. D., Spector, D. H. & Yeo, G. W. (2012). High-

resolution profiling and analysis of viral and host small RNAs during human cytomegalovirus infection. J Virol 86, 226–235. Su, H., Trombly, M. I., Chen, J. & Wang, X. (2009). Essential and

overlapping functions for mammalian Argonautes in microRNA silencing. Genes Dev 23, 304–317.

Systematic excision of vector sequences from the BAC-cloned herpesvirus genome during virus reconstitution. J Virol 73, 7056–7060.

cellular microRNA species that affect its replication. J Virol 82, 9065– 9074. Wang, B., Li, S., Qi, H. H., Chowdhury, D., Shi, Y. & Novina, C. D. (2009a). Distinct passenger strand and mRNA cleavage activities of

human Argonaute proteins. Nat Struct Mol Biol 16, 1259–1266. Wang, X., Wang, H.-K., McCoy, J. P., Banerjee, N. S., Rader, J. S., Broker, T. R., Meyers, C., Chow, L. T. & Zheng, Z.-M. (2009b).

Oncogenic HPV infection interrupts the expression of tumorsuppressive miR-34a through viral oncoprotein E6. RNA 15, 637–647. Wu, D.-G., Wang, Y.-Y., Fan, L.-G., Luo, H., Han, B., Sun, L.-H., Wang, X.-F., Zhang, J.-X., Cao, L. & other authors (2011). MicroRNA-7

regulates glioblastoma cell invasion via targeting focal adhesion kinase expression. Chin Med J (Engl) 124, 2616–2621. Xiong, S., Zheng, Y., Jiang, P., Liu, R., Liu, X. & Chu, Y. (2011).

MicroRNA-7 inhibits the growth of human non-small cell lung cancer A549 cells through targeting BCL-2. Int J Biol Sci 7, 805–814. Xu, N., Segerman, B., Zhou, X. & Akusja¨rvi, G. (2007). Adenovirus

virus-associated RNAII-derived small RNAs are efficiently incorporated into the RNA-induced silencing complex and associate with polyribosomes. J Virol 81, 10540–10549.

Tomari, Y., Du, T. & Zamore, P. D. (2007). Sorting of Drosophila small

Zhu, J. Y., Strehle, M., Frohn, A., Kremmer, E., Ho¨fig, K. P., Meister, G. & Adler, H. (2010). Identification and analysis of expression of novel

silencing RNAs. Cell 130, 299–308.

microRNAs of murine gammaherpesvirus 68. J Virol 84, 10266–10275.

Triboulet, R., Mari, B., Lin, Y.-L., Chable-Bessia, C., Bennasser, Y., Lebrigand, K., Cardinaud, B., Maurin, T., Barbry, P. & other authors

Zuker, M. (2003). Mfold web server for nucleic acid folding and

http://vir.sgmjournals.org

hybridization prediction. Nucleic Acids Res 31, 3406–3415.

Downloaded from www.microbiologyresearch.org by IP: 185.107.94.33 On: Tue, 05 Dec 2017 10:19:27

1547