expression of long anti hiv 1 hairpin rnas for the generation of multiple sirnas advantages and limitations

Expression of Long Anti-HIV-1 Hairpin RNAs for the Generation of Multiple siRNAs: Advantages and Limitations Masayuki Sano1,2, Haitang Li1, Mahito Nakanishi2 and John J Rossi1,3 1 Divis

Expression of Long Anti-HIV-1 Hairpin RNAs

for the Generation of Multiple siRNAs:

Advantages and Limitations

Masayuki Sano1,2, Haitang Li1, Mahito Nakanishi2 and John J Rossi1,3

1 Division of Molecular Biology, Beckman Research Institute of the City of Hope, Duarte, California, USA; 2 Biotherapeutic Research Laboratory,

National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan; 3 Division of Molecular Biology, Graduate School of Biological Science, Beckman Research Institute of the City of Hope, Duarte, California, USA

Promoter expressed long-hairpin RNAs (lhRNAs) that

can be processed into multiple small interfering RNA

(siRNAs) are being considered as effective agents for

treating rapidly mutating viruses such as human

immu-nodeficiency virus (HIV) In the present study, we have

generated human U6 promoter-driven lhRNAs of 50, 53,

and 80 base pairs (bp) targeting contiguous sequences

within the tat and rev genes of HIV-1 and evaluated the

efficacy of these lhRNAs as well as their processing in

cells By using multiple G:U mismatches in the stems, we

have been able to readily incorporate the long-hairpin

structures into a lentiviral vector transduction system

Here we show that such long hairpins can be stably and

functionally expressed for a long term in HIV-1

suscep-tible T cells, where they provide potent inhibition of HIV

replication against both non-mutant and mutant

vari-ants of HIV-1 Our studies provide strong support for the

use of the G:U wobble pair containing lhRNAs to

gener-ate multiple siRNAs from a single transcript, but we also

show that lhRNAs of 80 bp may be the upper size limit

for effectively producing multiple, functional siRNAs

Received 18 May 2007; accepted 1 August 2007; published online

28 August 2007 doi: 10.1038/sj.mt.6300298

INTRODUCTION

RNA interference (RNAi) is a eukaryotic regulatory mechanism

that uses double-stranded RNA (dsRNA) molecules as triggers to

direct homology-dependent, post-transcriptional gene silencing.1

Small interfering RNAs (siRNAs), are approximately 21–22 base

pairs (bp) long; dsRNA molecules have characteristic 2 nucleotide

(nt) 3′ overhangs that allow them to be recognized by the

enzy-matic machinery of RNAi which leads to homology-dependent

degradation of the target messenger RNA (mRNA).2 SiRNAs are

produced from cleavage of longer dsRNA precursors by the RNase

III endonuclease Dicer.3,4 In human cells, Dicer substrate

recogni-tion appears to be mediated via interacrecogni-tions with either one or

both accessory RNA binding proteins, the TAR-RNA binding

protein and PACT, which are presumably involved in the hand

off of siRNAs to the RNA-induced silencing complex.5,6 Cleavage-assisted removal of the passenger-strand leads to the activation

of the RNA-induced silencing complex,7,8 where a catalytic endo-nuclease, AGO-2, cleaves the target mRNA.9,10

Synthetic siRNAs and expressed short hairpin RNAs (shRNAs) have been used to target virtually all of the human

immunodefi-ciency virus (HIV)-encoded RNAs in cell lines, including tat, rev,

gag, pol, nef, vif, env, vpr, and the long terminal repeat.11–15 Other studies have shown a host of other viruses, including the hepati-tis B virus, hepatihepati-tis C virus, poliovirus and respiratory syncytial virus to be targeted by RNAi (recently reviewed in ref 16) Despite the early successes of RNAi-mediated inhibition of HIV-encoded RNAs in cell lines, targeting the virus directly presents a substan-tial challenge in clinical applications since the high viral mutation rate often leads to the outgrowth of mutants in the siRNA target sequences.17–19 In some cases, a single nucleotide mutation within the target site is sufficient to abolish the silencing efficacy of siR-NAs or shRsiR-NAs.20,21 As with the conventional drugs used in the treatment of HIV infection, the best approach for preventing or delaying the emergence of escape mutants is to use multiplexed anti-HIV RNAi triggers or combinations of anti-HIV-1 RNAi approaches with other antiviral genetic elements.22,23

Use of a long hairpin RNA (lhRNA) is an attractive approach

in controlling HIV-1 replication since a single long transcript can

in theory be processed into multiple siRNAs It has recently been reported that lhRNAs of more than 50 bp, harboring multiple C

to U (or A to G) mutations within the sense strand, can specifi-cally inhibit gene expression without triggering type I interferon (IFN) gene expression.24 Multiple targeting can be achieved from

a single long-hairpin precursor, suggesting that multiple siRNAs

can be processed from the long hairpins in vivo This strategy has

been applied in the inhibition of the hepatitis C virus and hepa-titis B virus as well as in HIV-1 replication.25–27 Despite reports of transient target knockdown efficacy from previous studies, before this particular study was undertaken, the long-term expression and inhibitory function of lhRNAs in stably transduced cells had not been demonstrated Moreover, we also show that long hairpin containing lentiviral vectors can be packaged to generate vectors with titers similar to that of non-hairpin containing vectors, an

Correspondence: John J Rossi, Division of Molecular Biology, Graduate School of Biological Science, Beckman Research Institute of the City of Hope,

1450 E Duarte Rd., Duarte, California 91010, USA E-mail: jrossi@coh.org

important consideration in gene therapy applications The

pro-cessing patterns of siRNAs from long hairpins have been

exam-ined in transiently transfected cells27 but no attempts to monitor

expression have been made in stable transduction settings We

therefore set out to investigate the patterns of processing and

functionality of expressed lhRNAs in both transiently transfected

and stably transduced cells so as to better understand the

mecha-nisms of action of expressed lhRNAs for use as anti-HIV agents

In this study we show that lhRNAs generate a gradient of

siR-NAs through Dicer-mediated cleavage, suggesting a limitation in

the maximum lengths of the hairpin structures for use in vivo

Nevertheless, the stably expressed lhRNA targeting a tat/rev

com-mon exon can suppress HIV-1 replication for the long term

with-out inducing type I IFN gene expression Moreover, the lhRNA

is capable of effectively inhibiting the replication of an HIV-1

strain harboring a mutation in the tat/rev common exon, whereas

a 21 bp shRNA targeting this exon failed to inhibit replication

Given these attributes, we believe that there is a strong rationale

for employing long hairpins in the treatment of HIV infection,

and in other applications requiring simultaneous expression of

two or three siRNAs

RESULTS

Efficient inhibition of HIV-1 gene expression

by expressed lhRNAs

We initially constructed vectors expressing either shRNAs or

lhRNAs that target the essential viral regulatory genes, tat/rev

(site I) and rev (site II), under the control of the human U6

pro-moter (Figure 1a) These target sequences have been previously

identified as highly effective sites for RNAi mediated inhibition

of HIV-1 replication.12 As controls, shRNAs (21 bp) harboring G:

U wobble pairings were constructed for both the site I-tat/rev and

site II rev targets To create the lhRNAs, we simply extended the

21 base sequences used in the shRNAs to be complementary to

the targets 3′ of the 21-mer target sites This generated a 53 nt (site

I-tat/rev) as well as 50 or 80 nt (site II-rev) hairpins Appropriate

U or G residues were inserted within the sense strands to create

G:U-U:G wobble pairings with the antisense strand (Figure 1b)

To examine the RNAi activity of the lhRNAs, we first tested them

using the non-infectious HIV-1 pNL4-3.Luc proviral DNA as a

target This construct harbors the firefly luciferase gene in the

place of the nef gene as well as frame-shift mutations within the

vpr and env genes The pNL4-3.Luc plasmid, a separate plasmid

encoding Renilla luciferase and the vectors expressing either

shRNAs or lhRNAs were cotransfected into HEK293 cells The

ratios of firefly luciferase versus Renilla luciferase activities were

determined 24, 48 and 72 hours after transfection The shRNAs

and lhRNAs targeting either site I or site II strongly inhibited the

expression of firefly luciferase at each time point, with the

lhR-NAs inhibiting luciferase expression to approximately the same

extent as the shRNAs at 72 hours post-transfection (Figure 2a)

These results confirmed that the G:U wobble pairs within both

the shRNAs and lhRNAs do not impede RNAi Next, to examine

the potential anti-HIV-1 inhibitory activities of the lhRNAs, we

co-transfected each of the lhRNA expression vectors as well as

the shRNA vectors with infectious HIV-1 pNL4-3 proviral DNA

into HEK293 cells Two days post-transfection, supernatants

were collected from the cell cultures and measurements of HIV-1 p24 viral antigen levels were carried out All of the shR-NAs and lhRshR-NAs potently inhibited HIV-1 replication in these co-transfections (Figure 2b) We next tried to find out if the lhR-NAs were inducing a type I IFN responsive gene expression in HEK293 cells To do this, we transfected the cells with shRNA

or lhRNA expression vectors and monitored the expression of

IFN-β, p56 and MxA mRNAs by quantitative real-time

poly-merase chain reaction (PCR) assays 48 h post-transfection Our results demonstrate that the lhRNAs do not induce expression of

the IFN-β, p56 and MxA mRNAs in this cell line (Figure 2d) To verify that HEK293 cells are capable of IFN mediated responses, cells were transfected with poly I:C, resulting in strong activa-tion of the IFN responsive genes (Figure 2c)

Expression profiles of siRNAs produced from lhRNAs

We next examined the profiles of siRNAs produced from each

of the lhRNAs using Northern blotting analyses HEK293 cells were transfected with either shRNA or lhRNA expression con-structs and total RNAs were isolated, either 24 hours (site I) or

48 hours (site II) after transfection The RNAs were fraction-ated by electrophoresis and transferred to nylon membranes for probing with 32P labeled oligos complementary to the antisense strands of each of the putative siRNAs that could be derived from

the hairpins Each of the probes used was non-overlapping and

was complementary to a 21 nt sequence (Figure 1b) As shown

to specific siRNAs For each of the lhRNAs, the strongest signal was obtained for the first 21 base sequence, with subsequent loss of intensity for the second siRNA and a greater reduction in signal intensity for the third siRNA produced from the 80 base lhRNA To exclude the possibility that the poor hybridization

LTR

nef rev

vif

tat

a

b

Site I

Site II

80 50 21 53

21

SI probe 1

SI probe 2 SII probe 1 SII probe 2 SII probe 3

Figure 1 Schematic representation of human immunodeficiency virus

RNA target sites, site I and site II, which are in the HIV-1 tat/rev common

exon and rev exon, respectively (b) The sequences and predicted

struc-tures of short- and long-hairpin RNAs directed against site I and site II The superscripts in the sequence of the sense strands of the hairpin RNAs indicate altered nucleotides, which generate G:U wobble base-parings in the transcript The regions complementary to probes used in Northern blotting analysis are indicated by thick bars LTR, long terminal repeat.

efficiency of probe 3 to the complementary region was the

rea-son for the weak signal, we designed two other probes, which

partially overlapped probe 3 and covered additional sequences

(Supplementary Figure S1a) Northern analyses revealed no

significant differences in signal intensities between these probes

(Supplementary Figure S1a) Finally, we have carried out in vitro

Dicing reactions using the lhRNA substrates (Supplementary

Figure S5) For these experiments, we incubated in vitro

tran-scribed long hairpins with recombinant human Dicer

over-night The Diced RNAs were electrophoresed and blotted onto

a nylon membrane The polarity that we observe in cell culture

for processing of the third position of the 80 base lhRNA was

observed in vitro Taken together these results strongly support

the conclusion that processing of the hairpins by the Dicer is

directional, beginning at the 5′ end of the duplex and proceeding

towards the loop Moreover, the gradient of siRNAs produced

b50,000 40,000 30,000

20,000 10,000

0 Control SI-2

1 SI-53 SII-21 SII-50

a 120

100

80

60

40

20

0

Control SI-21 SI-53 SII-21 SII-50 SII-80

c10,000

1,000

100

10

1

Control Poly I:C

IFN-� p56 MxA

d3.0 2.5 2.0

1.5 1.0 0.5 0 Control SI-21 SI-53 SII-21 SII-50 SII-80

IFN-� p56 MxA

Figure 2 Inhibition of human immunodeficiency virus (HIV)-1 gene

expression and monitoring for type I interferon (IFN) responses

(a) Effects of short- and long hairpin RNA (lhRNA) expression vectors

on the expression of pNL4-3.Luc transcripts The pNL4-3.Luc and Renilla

luciferase expression plasmids were used along with the individual

lhRNA and short hairpin RNA constructs in cotransfections of HEK293

cells Following the transfections, luciferase activities were determined

at the indicated time points The luciferase activities of cells transfected

with the empty vector were set at 100% for each time point, and relative

luciferase activities of cells transfected with hairpin RNA expression

vec-tors are indicated Values represent the means with standard deviations

from three replicate experiments (b) Inhibition of HIV-1 p24 expression

by lhRNAs HEK293 cells were cotransfected with pNL4-3 infectious

pro-viral DNA along with the U6 empty vector or U6 hairpin RNA expression

vectors Two days after transfection, supernatants were collected and

p24 levels were determined The transfections were performed in

dupli-cate and results are shown as the average values (c, d) The effects of

lhRNA constructs on the levels of IFN-inducible messenger RNA (mRNA)

HEK293 cells in each well of a 12-well plate were transfected with 0.2 µg

poly I:C or 1 µg empty vector or hairpin-expression vector Twenty-four

(poly I:C) or forty-eight (vector) hours after transfection, total RNA was

isolated and the relative expression levels of IFN-β, p56 and MxA mRNAs

were quantified by real-time polymerase chain reaction assays The levels

of the IFN regulated mRNAs were normalized with the HPRT1 RNA The

mRNA levels of cells transfected with the empty vector was set at 1, and

relative mRNA levels are indicated The mean values and standard

devia-tions from four (IFN-β) or three (p56 and MxA) replicate experiments are

presented.

psiCHECK-SI wt mut-1 mut-2

wt mut-1 mut-2 psiCHECK-SII

b

140 120 100

80 60 40 20

SI-wt SI-mut-1 SI-mut-2 SII-wt SII-mut-1 SII-mut-2

c

a SI probe1

U6

Control SI-21SI-53 Control SI-21 SI-53 Control SII-21 SII-50 SII-80 Control SII-21 SII-50 SII-80 Control SII-21SII-50SII-80

*

SI probe2 SII probe1 SII probe2 SII probe3

Figure 3 Processing of long-hairpin RNAs into multiple small inter-fering RNAs (siRNAs) (a) Detection of siRNAs in HEK293 cells

trans-fected with each hairpin RNA expression vector Twenty micrograms of total RNA were electrophoresed in an 8% polyacrylamide-7M urea gel The separated RNAs were blotted onto nylon membranes and hybrid-ized with the specific probes as illustrated in Figure 1b RNA extracted from cells transfected with the empty vector was used as a negative control U6 RNAs were probed as a loading standard Bands that cor-respond to siRNAs are indicated by an arrowhead Asterisks indicate

non-specific bands (b) Sequences of DNA inserts used to create the

reporter constructs Each synthesized DNA was inserted into the cloning site in the 3′UTR of the Renilla luciferase gene in the psiCHECK-2

plas-mid Nucleotide changes from wild-type are underlined (c) Inhibition of

reporter gene expression by the various hairpin constructs Each of the reporter constructs and hairpin RNA expression vectors were cotrans-fected into HEK293 cells, and luciferase activities were determined The luciferase activities determined from cells transfected with the empty plasmid were set at 100% The mean and standard deviations from three replicate experiments are presented.

from these constructs also suggests that the polarity of Dicer

processing could be a rate limiting step for the production of

multiple siRNAs from a single long precursor, especially those

longer than 80 bp

To investigate whether the individual siRNAs (processed from

the lhRNA precursors) had indeed functioned in triggering RNAi,

we constructed psiCHECK reporter plasmids harboring target

sequences complementary to the full length of the 50 bp lhRNAs,

as well as targets in which the sequences complementary to the

first siRNA were mutated (Figure 3b) These sequences were

inserted within the 3′UTR of the Renilla luciferase gene for

func-tional assays Each luciferase reporter plasmid was co-transfected

into HEK293 cells with the vector expressing either the shRNA or

lhRNA, and the ratios of Renilla luciferase versus firefly luciferase

activities were determined 24 hours post-transfection The 21 bp

shRNAs targeting site I and site II reduced the luciferase

activ-ity from psiCHECK-SI-wt and psiCHECK-SII-wt, respectively,

but did not downregulate the luciferase from the other reporter

plasmids harboring mutations in the target region (Figure 3c) In

contrast, the lhRNAs targeting site I and site II, inhibited

lucif-erase expression from both the wt- and mut-1-reporter

plas-mids, indicating that at least two different species of functional

siRNAs could be generated during the processing of the lhRNAs

Importantly, each lhRNA was less efficient at knocking down

lucif-erase activity from the psiCHECK-mut-1 than the psiCHECK-wt,

consistent with our observations of a gradient of siRNAs being

produced from the hairpins The lhRNAs showed no reduction of

luciferase activities from mut-2 reporter plasmids, reaffirming the

specificity of the siRNAs for their cognate, fully complementary

targets (Figure 3c)

Northern blotting analysis revealed that the 80 bp lhRNA

pro-duced marginal levels of siRNAs from the region near the loop

these siRNAs contributed to the target knockdown, we once again

used the psiCHECK reporter assay We constructed reporter

plas-mids with different short targets (Supplementary Figure S2a)

and monitored the efficacy of each siRNA produced from the

pro-cessing of the 80 bp lhRNA The inhibition of luciferase

expres-sion from the plasmid containing the sequence complementary

to the third siRNA was 12%, whereas the inhibitory activities of

luciferase expressed from the plasmids containing the sequences

complementary to the first or second siRNAs were 97 and 91%

respectively (Supplementary Figure S2b) The 50 bp lhRNA also

efficiently inhibited luciferase expression from a plasmid that

contained target 2 These results suggest that Dicer processing

rates decline somewhat dramatically after the second siRNA is

processed

Construction of lentiviral vectors expressing lhRNAs

To evaluate the anti-HIV-1 activity of the lhRNAs in acute,

long-term HIV-1 challenges, the genes encoding the lhRNAs or the

shRNAs were inserted into the pHIV-7-GFP lentiviral vector The

vectors were packaged and titered as described in the Materials

and Methods section Following a 200-fold concentration of

culture supernatant, approximately 3 × 108 transduction units/

ml was achieved for vectors harboring lhRNAs as well as

shR-NAs Thus, the inclusion of a long hairpin did not significantly

affect the packaging and transduction efficiencies of the lentiviral vector

The packaged vectors were used to transduce human CEM T cells and the transduction efficiencies were determined by fluo-rescence activated cell sorting analyses of green fluorescent pro-tein (GFP) expression At a multiplicity of infection of 10, the transduction efficiency for each of the hairpin constructs was approximately 90% (data not presented) To confirm that shR-NAs or lhRshR-NAs were expressed in transduced cells, total RNA was extracted from the transduced cells and probed for siRNA expression by Northern blotting Probe 1 for either sI or sII, as shown in Figure 1b, yielded signals that corresponded to the cog-nate siRNAs in each of the lhRNA transduced cell lines, whereas probes 2 and 3 yielded signals in RNA preps from the lhRNA transduced cells but not the shRNA transduced cells (Figure 4a) Only weak hybridization signals corresponding to precursors were observed for each of the probes, suggesting that the hairpin precursor processing reaction is efficient in these stably trans-duced CEM T cells

We next examined whether the lhRNAs were capable of

inducing the expression of the IFN responsive genes,

oligoadenylate synthetase 1 (OAS1) and p56 in the transduced

SI probe1

U6

Vecto r Vector Vector Vector Vector SI-21 SI-53 SI-21 SI-53 SII-21 SII-50 SII-80 SII-21 SII-50 SII-80 SII-21SII-50SII-80

SI probe2 SII probe1 SII probe2 SII probe3

a

100

90 70 50 30 10 0

CEM IFN VecSI-21 SI-53 SII-21 SII-50 SII-80

b

0 4 8 10 14 18

CEM IFN Vec SI-21SI-53 SII-21 SII-50 SII-80

c

Figure 4 Processing of long-hairpin RNAs (lhRNAs) and measure-ments of type I interferon (IFN) response gene expression in stably transduced CEM T cells (a) Detection of small interfering RNAs

pro-duced from lhRNAs Total RNA was extracted from CEM T cells trans-duced with the lentiviral vectors harboring short hairpin RNA or lhRNA genes and Northern blotting analyses were performed with the specific probes shown in Figure 1b RNA prepared from cells transduced with the pHIV-7 lentiviral vector-backbone was used as a negative control U6 RNAs were probed as loading controls Bands that correspond to siRNA

are indicated by the arrowhead (b) Effects of lhRNAs on the levels of

OAS1 messenger RNA (mRNA) Total RNA was isolated from the

vec-tor-backbone- or hairpin RNA-transduced CEM T cells and the level of

OAS1 was detected by real-time polymerase chain reaction The level

of OAS1 mRNA was normalized to GAPDH mRNA An RNA extract from cells treated with 1,000 U/ml of IFN-α for 24 hours was used as a

posi-tive control The mRNA level of untransduced CEM T cells was set at 1, and the relative levels of mRNA are indicated The means and standard

deviations from four replicate experiments are presented (c) Effects of

lhRNAs on the levels of p56 mRNA The experimental procedures and

values are the same as in (b).

CEM T cells Quantitative real-time PCR assays revealed that

stably expressed lhRNAs do not induce expression of the OAS1

and p56 mRNAs (Figure 4b and c) To verify that CEM T cells are

capable of IFN mediated responses, CEM cells were treated with

IFN-α (1,000 U/ml), resulting in potent activation of the IFN

responsive genes Other IFN-inducible mRNAs, MxA, and RIG-I,

had no appreciable increases in CEM cells expressing lhRNAs

(Supplementary Figure S3a and b).

Effective suppression of acute HIV-1 infection

in cells transduced with lhRNAs

We next sought to demonstrate the antiviral activity of the lhRNA

CEM T cells transduced with either the lhRNA or the shRNA,

targeting sI-tat/rev, were challenged with HIV-1 IIIB (Figure 5a)

or a variant harboring a mutation in the tat/rev common exon,

which corresponds to the first siRNA produced from the lhRNA

of inhibition of viral RNA production (relative to the controls)

in both the wild-type and mutant viral challenges, whereas the

shRNA that targeted tat/rev site I was only effective against

the wild-type virus (Figure 5b) CEM T cells transduced with either the 50 bp or 80 bp lhRNA that targeted sII-rev also

con-ferred resistance to virus infection up to 14 days after infection

(Supplementary Figure S4) Importantly, both the lhRNAs that

targeted sII-rev showed more potent inhibition than the shRNA

To evaluate the anti-HIV-1 activity of the lhRNA under conditions

of long-term incubation, we challenged CEM T cells transduced

with either the shRNA or the lhRNA, that targetted sI-tat/rev,

with HIV-1 IIIB at a multiplicity of infection of 0.001 As shown

repli-cation for over 6 weeks, whereas cells transduced with the single shRNA had a viral replicactive breakthrough on approximately day 41 The branched chain DNA assay used for these analyses quantitatively measures viral RNA production, thereby providing

a stringent analysis of the antiviral activities of these constructs

DISCUSSION

It has been largely assumed that the application of long dsRNAs as RNAi triggers was only appropriate in undifferentiated embryonic stem cells and embryonic carcinoma cells due to their lack of IFN-inducible gene expression.28,29 However, recent studies suggested that endogenously expressed long dsRNAs and lhRNAs do not activate the IFN pathway, perhaps due to their being rapidly pro-cessed into siRNAs by Dicer.24,25,30–33 In the present study, we have shown that transiently or stably expressed lhRNAs with several G:U mismatches are processed into two or more siRNAs which can effectively suppress replication of HIV-1 We have demonstrated that the ability to produce two siRNAs from a single transcript has distinct advantages over a single siRNA, in conferring long-term repression of HIV-1 replication and inhibition of a virus harbor-ing a mutation to one of the two siRNAs

10 9

10 8

10 7

10 6

10 5

10 4

10 3

10 2

10

1

Days post-infection

10 9

10 8

10 7

10 6

10 5

10 4

10 3

10 2

10

1

Days post-infection

b

a

Wt Mut Vector SI-21 SI-53

Vector SI-21 SI-53

Figure 5 Inhibition of human immunodeficiency virus (HIV)-1

infec-tion in CEM T cells expressing anti-tat/rev hairpin RNAs (a) CEM T

cells transduced with the vector-backbone or each of the hairpin RNA

expression vectors were infected with HIV-1 IIIB at a multiplicity of

infec-tion of 0.01 Culture supernatants were collected on the days indicated

and viral RNA levels were determined by a branched DNA assay (b) (top)

The sequences of the site I target region of the 21 base pair short hairpin

RNA in wild-type (Wt) and mutant (Mut) strains of HIV-1 IIIB Nucleotide

mutations in the mutant strain are indicated by underlining (Bottom)

Transduced CEM T cells challenged with the mutant HIV-1 IIIB isolate.

10 9

10 8

10 7

10 6

10 5

10 4

10 3

10 2

10 1

7 14 21 28 35 41 48 Days post-infection

Vector SI-21 SI-53

Figure 6 Long-term Inhibition of human immunodeficiency virus

(HIV)-1 replication in CEM T cells expressing an anti-tat/rev long

hairpin RNA (lhRNA) CEM T cells transduced with the

vector-back-bone, a short hairpin RNA or a lhRNA targeting the tat/rev common

exon were infected with HIV IIIB at a multiplicity of infection of 0.001 Culture supernatants were collected on the days indicated and viral RNA levels were determined with the branched DNA assay system.

The use of G:U wobble pairings for the cloning and

expres-sion of lhRNAs may have several advantages over perfectly

complementary lhRNAs For instance, the presence of wobble

pairing (which in DNA is G:T) prevents the formation of

ther-modynamically stable hairpin structures which often result in

deletions during plasmid propagation in Escherichia coli This

feature also facilitates sequencing of these constructs, which is

often a problem with perfectly complementary hairpin

struc-tures Previously published results from our lab and others show

that an shRNA sequence encoded in a lentiviral vector results

in reduced packaging titers.22,34 In the present study we have

routinely obtained titers that are comparable to the backbone

vector, perhaps owing to the G:U wobbles which lower the

ther-modynamic stabilities of the hairpins, thereby allowing more

efficient reverse transcription Another possible advantage of

using the G:U wobble pairing in the stem is that it may

abro-gate RNAi mediated cleavage of the vector transcripts, triggered

by siRNAs produced from the hairpins, during the packaging

reaction We thus believe that the incorporation of G:U wobble

pairings in the sense strand of both shRNAs and lhRNAs has

several advantages when the hairpins are going to be

incorpo-rated into lentiviral or retroviral vectors Most importantly, our

stably integrated lhRNA expression cassettes produce efficacious

levels of multiple siRNAs and confer long-term suppression of

HIV replication (Figures 4a , 5a and 6)

It has been suggested that endogenously expressed lhRNAs

do not activate the expression of type I IFN inducible genes in

cultured cell lines.24,26,27 However, in some cases, some genes

have been shown to be up-regulated by perfectly complementary

lhRNAs, whereas lhRNAs with G:U mismatches attenuated such

activation.24,26 Since G:U wobble parings most probably change

the thermodynamic properties of hairpin structures, these may

evade recognition by dsRNA-binding proteins such as PKR

Recently, it was reported that the cytoplasmic helicase proteins

RIG-I and MDA5 recognize single and dsRNAs bearing a

triphosphate, leading to type I IFN expression.35,36 These proteins

sense infection from a variety of viruses and trigger antiviral

responses.37 Interestingly, RIG-I can also recognize chemically

synthesized blunt-ended 27 nt siRNAs.38 Since stably expressed

lhRNAs with G:U wobble parings did not induce the

activa-tion of IFN-inducible genes (Figure 4b and c , Supplementary

Figure S3a and b), it is unlikely that these lhRNAs are recognized

by RIG-I and MDA5 Further analyses will be required to

exam-ine the potential activation of the IFN pathway in the presence of

these long RNAs in immune cells such as monocytes,

plasmacy-toid dendritic cells and non-plasmacyplasmacy-toid dendritic cells

Consistent with previous reports, we observed that lhRNAs

produce multiple active siRNAs We have also observed by both

functional assays and Northern analyses that the processing of

lhRNAs is processive and produces a gradient of siRNAs from

the stem towards the loop The reduction in the amount of the

third position siRNA in our constructs suggests a limitation in

the useful length of lhRNAs (Figure 3a) This processive

pro-cessing was observed in transient transfections, and in stably

transduced cells (Figure 4a) as well as in vitro (Supplementary

Figure S5) A recent report has demonstrated that the first 21 bp

siRNA produced from a 62 bp lhRNA showed stronger silencing

efficacy than the second siRNA in transiently transfected cells27

although the mechanism for this was not elaborated upon We observed a similar polarity in target knockdown by the siRNAs

produced from an 80 bp lhRNA (Supplementary Figure S2b)

Nevertheless, our 53 bp tat/rev targeting-lhRNA, which produces

two distinct siRNAs, provided a significant advantage over a sin-gle shRNA in both longevity of inhibition of HIV and the ability

to inhibit a variant harboring two point mutations in the target sequence complementary to the first siRNA (Figure 5b)

Sequence-specific target degradation by siRNAs is usually negatively affected by even a single point mutation, although the degree of silencing efficacy varies depending on the position of nucleotide mismatches A single mutation at position 2 or 12 relative to the 5′-end of the sense strand has been shown to be tol-erated in suppression of HIV-1.20,21 However, two point mutations

at both positions 2 and 12 completely abolished the activity of our shRNA (Figure 5b) It has recently been reported that a human T-cell line expressing an integrase-targeting lhRNA showed limited anti-HIV activity 6 days post infection.39 In contrast, our lhRNA strongly inhibited HIV-1 replication for several weeks (Figures 5a

and 6) By targeting contiguous sequences in the tat/rev common

exon, mutants that arise should also have a selective disadvantage because the reading frames for these two proteins are different, and silent third-position codon changes for one gene will very often result in an amino acid alteration for the other gene, thereby affecting expression of that gene

Overall, our studies demonstrate the potential advantage of using lhRNAs over shRNAs for long-term control of HIV-1 rep-lication Nevertheless, there are limitations to this approach, the most significant of which is the polariy of processing by Dicer

as it proceeds along the hairpin structure, rendering hairpins longer than 80 or so bp less useful than those below this length

A final consideration is that producing multiple siRNAs from

a single transcript could in some cases result in competition for endogenous RNA-induced silencing complex components, thereby impacting on the functionality of the multiple siRNAs themselves as well as the micro RNA pathway Strand selectivity

is another potential concern, but this can be controlled by ther-modynamic properties of the ends of the siRNAs produced from the lhRNAs.40,41 Competition among the lhRNA produced siR-NAs with each other and/or the cellular micro RsiR-NAs will depend upon the sequences of the lhRNAs and how effectively they compete for the RNA-induced silencing complex.42 The number

of integrated copies and the strength of the promoters used to transcribe the lhRNAs are also important factors for controlling potential siRNA mediated toxicities

MATERIALS AND METHODS

Plasmid construction Plasmids that contain a human U6 promoter were described previously 43 To create 21, 50 and 53 bp-hairpin RNA expres-sion vectors, the hybridized sense and antisense strands were inserted

within a BspMI site downstream of the U6 promoter To create the 80 bp

lhRNA expression vector, the DNA was amplified by PCR using

prim-ers with BspMI sites Following digestion, the double-stranded DNA was

inserted immediately downstream of the U6 promoter Sequences used for shRNAs and lhRNAs were as follows (sense mutations indicated in lower case and loop sequence underlined): SI-21: 5′-GCGGAGAtAGC GAtGAAGAGtTTCAAGAGAGCTCTTCGTCGCTGTCTCCGCTTT

TT-3′, SI-53: 5′-GCGGAGAtAGCGAtGAAGAGtTCATtAGAACAGTt

AGACTtATCAAGtTTCTCTTCAAGAGAGAGAAGCTTGATGAGTC

TGACTGTTCTGATGAGCTCTTCGTCGCTGTCTCCGCTTTTT-3′,

SII-21: GCCTGTGtCTCTTCAGtTACtTTCAAGAGAGGTAGCTGAAG

AGGCACAGGCttttt-3′, SII-50: 5′-GCCTGTGtCTCTTCAGtTACtACCG

tTTGAGAGgCTTAtTCTTGgTTGTATTCAAGAGATACAATCAAGAG

TAAGTCTCTCAAGCGGTGGTAGCTGAAGAGGCACAGGCTTTTT-

3′, SII-80: 5′-GCCTGTGtCTCTTCAGtTACtACCGtTTGAGAGgCTTA

tTCTTGgTTGTAAtGAGGgTTGTGGAgCTTCTGGGgCGtAGGTTCA

AGAGACCTGCGTCCCAGAAGTTCCACAATCCTCGTTACAATCA

AGAGTAAGTCTCTCAAGCGGTGGTAGCTGAAGAGGCACAGGC-

3′ Construction of the lentiviral vector was described previously 44 Each

hairpin RNA expression cassette was excised from the U6-based vectors

with BglII and BamHI or EcoRI and BamHI followed by filling in of the

digested sites with DNA polymerase I (Klenow), and the fragments were

ligated into either the unmodified BamHI site or the blunt-ended BamHI

site of pHIV-7-GFP Luciferase reporter plasmids were constructed by

inserting the double stranded target sequences into the BglII and NotI

sites of the psiCHECK-2 plasmid (Promega, Madison, WI).

Cell culture HEK293, 293T and HT1080 cells were cultured in Dulbecco’s

modified Eagle’s medium supplemented with 10% fetal bovine serum

The human T-cell line CEM was cultured in the Roswell Park Memorial

Institute’s 1640 medium, supplemented with 10% fetal bovine serum.

Luciferase assays HEK293 cells were grown to approximately 70–80%

confluency in 24-well plates and cotransfected with 100 ng of the

lucifer-ase reporter plasmid, 100 ng of the hairpin RNA expression vector and, in

the case of the pNL4-3.Luc plasmid, 1 ng of a renilla luciferase expression

plasmid (Promega, Madison, WI) All transfections were performed with

Lipofectamine 2000 (Invitrogen) At the designated time points,

lucif-erase activities were analyzed with a Dual Luciflucif-erase system (Promega,

Madison, WI).

Northern blotting Total RNAs were extracted and purified with RNA

STAT-60 (TEL-TEST “B”, Friendswood, TX) according to the

manufac-turer’s instructions Twenty micrograms of total RNA per lane were loaded

in an 8% polyacrylamide denaturing gel Following electrophoresis, bands

of RNA were electro-transferred to a Hybond-N membrane (Amersham

Bioscience, Piscataway, NJ) The membrane was probed with γ-32 P-labeled

synthetic oligonucleotides complementary to the antisense sequences of

hairpin RNAs.

Lentiviral vector production The packaging system used has been

described previously 45 293T cells were cultured until they reached 80%

confluency in a 100-mm culture dish Fifteen micrograms of the lentiviral

vectors, 15 µg of pCHGP-2, 5 µg of pCMV-G, and 5 µg of pCMV-rev were

cotransfected into 293T cells using the calcium phosphate precipitation

procedure Six hours after transfection, the culture medium was replaced

The culture supernatants were collected 42 hours after transfection The

supernatants were pooled together, passed through a 0.45-µm-pore-size

filter, concentrated by ultracentrifugation, and stored at –80 °C until use

Vector titers were determined by transduction of HT1080 cells and assayed

for enhanced GFP expression using flow cytometry The vectors were free

of replication-competent lentivirus as determined by both genomic PCR

and p24 antigen assays.

Transduction of target cells For transduction of CEM T cells, 2 × 10 5

cells were placed in a 15-ml centrifuge tube with 1 ml culture medium and

4 µg/ml polybrene in the presence of the lentiviral vector at a multiplicity

of infection of 10 Following centrifugation at 2,000 rpm for 30 minutes,

the cells were transferred into a 24-well culture plate and after 24 hours the

culture medium was replaced Transduced cells were sorted by means of

fluorescence activated cell sorting before use, when the transduction

effi-ciency was less than 90% (based on enhanced GFP expression).

RNA expression analysis Expression of human mRNAs encoding, IFN-β,

MxA, p56, OAS1 and RIG-I were determined by a real-time PCR using 2×

iQ SyberGreen Mastermix (Bio-Rad, Hercules, CA) and specific primer sets for these genes 46 RNA STAT-60 was used to extract total RNA Residual DNA was digested with DNase I (Roche, Nutley, NJ), and then comple-mentary DNA was produced using 2 µg of total RNA, moloney murine leukemia virus reverse transcriptase (Invitrogen, Carlsbad, CA) and 100 ng

of random primers in a 25 µl reaction. HPRT1 or GAPDH expressions were

used as internal controls and for normalization of the PCR data.

HIV-1 challenge One million CEM T cells were infected with wild-type or mutant strains of HIV-1IIIB at a multiplicity of infection of 0.01

or 0.001 After overnight incubation, the cells were washed three times with Hank’s balanced salts solution and cultured using the Roswell Park Memorial Institute’s 1640 medium with 10% fetal bovine serum At des-ignated time points, culture supernatants were collected and analyzed for viral RNA levels by a branched DNA (bDNA) assay using the QuantiGene Reagent System (Panomics, Fermont, CA), according to the manufactur-er’s instructions.

HIV-1 antiviral assays Cells were cotransfected with HIV-1 pNL4-3 pro-viral DNA and each hairpin RNA expression vector in 1:1 wt/wt ratio using Lipofectamine Plus (Invitrogen, Carlsbad, CA), according to the manu-facturer’s instructions Culture supernatants were collected 2 days after transfection and analyzed for HIV-1 p24 antigen using an enzyme-linked immunosorbent assay (Beckman Coulter, Fullerton, CA) The p24 values were calculated using a Dynatech MR5000 enzyme-linked immunosor-bent assay plate reader (Dynatech Lab, Chantilly, VA).

In vitro Dicer cleavage of lhRNAs Template DNAs, including the 21, 50

or 80 bp-hairpin sequences, were prepared by PCR of the various sII-rev hairpin plasmids The primers for producing the T7 transcription units were: upper (5′-TAATACGACTCACTATAGCCTGTGTCTCTTCAGTT ACT-3′) and lower (5′-AAGCCTGTGCCTCTTCAGCTACC) Once the PCR products were produced, the corresponding RNAs were transcribed

in vitro using the MEGAshortscript T7 kit (Ambion, Austin, TX),

accord-ing to the manufacturer’s instructions One microgram of transcript was incubated with 1 U of a recombinant human Dicer (Genlantis) for 16 hours

at 37 °C The products of the reaction were resolved in an 8% polyacryl-amide denaturing gel After electrophoresis, Northern blotting was per-formed as described in the Northern blotting analysis section.

ACKNOWLEDGMENTS

This research was supported by National Institutes of Health (NIH) AI29329; AI42552 and HL07470 The authors thank Kazunari Taira for helpful discussion, Ming-Jie Li (Department of Neuroscience, Washington University Medical School, St Louis, MO) and Hideo Akashi (National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan) for technical support and NIH AIDS Research and Reference Reagent Program for providing the pNL4-3.Luc.R-E-plasmid.

SUPPLEMENTARy MATERIAL

Figure S1 Expression of siRNAs from the 80bp lhRNA.

Figure S5 In vitro Dicer cleavage of shRNAs and lhRNAs.

Figure S2 Efficacy of individual siRNAs produced from lhRNAs.

Figure S3 Effects of lhRNAs on the levels of MxA and RIG-I mRNA in

transduced cells.

Figure S4 Inhibition of HIV-1 infection in CEM T cells expressing

sII-rev hairpin RNAs.

REFERENCES

1 Almeida, R and Allshire, RC (2005) RNA silencing and genome regulation

Trends Cell Biol 15: 251–258.

2 Elbashir, SM, Lendeckel, W and Tuschl, T (2001) RNA interference is mediated by

21- and 22-nucleotide RNAs Genes Dev 15: 188–200.

3 Bernstein, E, Caudy, AA, Hammond, SM and Hannon, GJ (2001) Role for a bidentate

ribonuclease in the initiation step of RNA interference Nature 409: 363–366.

4 Zhang, H, Kolb, FA, Jaskiewicz, L, Westhof, E and Filipowicz, W (2004) Single

processing center models for human Dicer and bacterial RNase III Cell 118: 57–68.

5 Chendrimada, TP, Gregory, RI, Kumaraswamy, E, Norman, J, Cooch, N, Nishikura, K

et al (2005) TRBP recruits the Dicer complex to Ago2 for microRNA processing and

gene silencing Nature 436: 740–744.

6 Lee, Y, Hur, I, Park, SY, Kim, YK, Suh, MR and Kim, VN (2006) The role of PACT in the

RNA silencing pathway EMBO J 25: 522–532.

7 Matranga, C, Tomari, Y, Shin, C, Bartel, DP and Zamore, PD (2005) Passenger-strand

cleavage facilitates assembly of siRNA into Ago2-containing RNAi enzyme complexes

Cell 123: 607–620.

8 Rand, TA, Petersen, S, Du, F and Wang, X (2005) Argonaute2 cleaves the anti-guide

strand of siRNA during RISC activation Cell 123: 621–629.

9 Liu, J, Carmell, MA, Rivas, FV, Marsden, CG, Thomson, JM, Song, JJ et al (2004)

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

10 Song, JJ, Smith, SK, Hannon, GJ and Joshua-Tor, L (2004) Crystal structure of

Argonaute and its implications for RISC slicer activity Science 305: 1434–1437.

11 Jacque, JM, Triques, K and Stevenson, M (2002) Modulation of HIV-1 replication by

RNA interference Nature 418: 435–438.

12 Lee, NS, Dohjima, T, Bauer, G, Li, H, Li, MJ and Ehsani, A et al (2002)

Expression of small interfering RNAs targeted against HIV-1 rev transcripts in human

cells Nat Biotechnol 20: 500–505.

13 Martinez, MA, Clotet, B and Este, JA (2002) RNA interference of HIV replication

Trends Immunol 23: 559–561.

14 Novina, CD, Murray, MF, Dykxhoorn, DM, Beresford, PJ, Riess, J, Lee, SK et al (2002)

siRNA-directed inhibition of HIV-1 infection Nat Med 8: 681–686.

15 Coburn, GA and Cullen, BR (2002) Potent and specific inhibition of human

immunodeficiency virus type 1 replication by RNA interference J Virol 76: 9225–9231.

16 Leonard, JN and Schaffer, DV (2006) Antiviral RNAi therapy: emerging approaches

for hitting a moving target Gene Ther 13: 532–540.

17 Boden, D, Pusch, O, Lee, F, Tucker, L and Ramratnam, B (2003) Human

immunodeficiency virus type 1 escape from RNA interference J Virol

77: 11531–11535.

18 Das, AT, Brummelkamp, TR, Westerhout, EM, Vink, M, Madiredjo, M, Bernards, R

et al (2004) Human immunodeficiency virus type 1 escapes from RNA

interference-mediated inhibition J Virol 78: 2601–2605.

19 Westerhout, EM, Ooms, M, Vink, M, Das, AT and Berkhout, B (2005) HIV-1 can

escape from RNA interference by evolving an alternative structure in its RNA genome

Nucleic Acids Res 33: 796–804.

20 Pusch, O, Boden, D, Silbermann, R, Lee, F, Tucker, L and Ramratnam, B (2003)

Nucleotide sequence homology requirements of HIV-1-specific short hairpin RNA

Nucleic Acids Res 31: 6444–6449.

21 Sabariegos, R, Gimenez-Barcons, M, Tapia, N, Clotet, B and Martinez, MA (2006)

Sequence homology required by human immunodeficiency virus type 1 to escape

from short interfering RNAs J Virol 80: 571–577.

22 Li, MJ, Kim, J, Li, S, Zaia, J, Yee, JK, Anderson, J et al (2005) Long-term inhibition of

HIV-1 infection in primary hematopoietic cells by lentiviral vector delivery of a triple

combination of anti-HIV shRNA, anti-CCR5 ribozyme, and a nucleolar-localizing TAR

decoy Mol Ther 12: 900–909.

23 Unwalla, HJ, Li, HT, Bahner, I, Li, MJ, Kohn, D and Rossi, JJ (2006) Novel Pol II fusion

promoter directs human immunodeficiency virus type 1-inducible coexpression of a

short hairpin RNA and protein J Virol 80: 1863–1873.

24 Akashi, H, Miyagishi, M, Yokota, T, Watanabe, T, Hino, T and Nishina, K et al (2005)

Escape from the interferon response associated with RNA interference using vectors

that encode long modified hairpin-RNA Mol Biosyst 1: 382–390.

25 Konstantinova, P, de Vries, W, Haasnoot, J, ter Brake, O, de Haan, P and Berkhout, B

(2006) Inhibition of human immunodeficiency virus type 1 by RNA interference using

long-hairpin RNA Gene Ther 13: 1403–1413.

26 Watanabe, T, Sudoh, M, Miyagishi, M, Akashi, H, Arai, M, Inoue, K et al (2006)

Intracellular-diced dsRNA has enhanced efficacy for silencing HCV RNA and

overcomes variation in the viral genotype Gene Ther 13: 883–892.

27 Weinberg, MS, Ely, A, Barichievy, S, Crowther, C, Mufamadi, S, Carmona, S et al (2007) Specific inhibition of HBV replication in vitro and in vivo with expressed long

hairpin RNA Mol Ther 15: 534–541.

28 Wianny, F and Zernicka-Goetz, M (2000) Specific interference with gene

function by double-stranded RNA in early mouse development Nat Cell Biol

2: 70–75.

29 Billy, E, Brondani, V, Zhang, H, Muller, U and Filipowicz, W (2001) Specific interference with gene expression induced by long, double-stranded RNA

in mouse embryonal teratocarcinoma cell lines Proc Natl Acad Sci USA

98: 14428–14433.

30 Paddison, PJ, Caudy, AA and Hannon, GJ (2002) Stable suppression of gene

expression by RNAi in mammalian cells Proc Natl Acad Sci USA 99: 1443–1448.

31 Diallo, M, Arenz, C, Schmitz, K, Sandhoff, K and Schepers, U (2003) Long endogenous dsRNAs can induce complete gene silencing in mammalian cells and

primary cultures Oligonucleotides 13: 381–392.

32 Tran, N, Raponi, M, Dawes, IW and Arndt, GM (2004) Control of specific gene expression in mammalian cells by co-expression of long complementary RNAs

FEBS Lett 573: 127–134.

33 Strat, A, Gao, L, Utsuki, T, Cheng, B, Nuthalapaty, S, Mathis, JM et al (2006)

Specific and nontoxic silencing in mammalian cells with expressed long dsRNAs

Nucleic Acids Res 34: 3803–3810.

34 Sliva, K and Schnierle, BS (2006) Stable integration of a functional shRNA

expression cassette into the murine leukemia virus genome Virology

351: 218–225.

35 Hornung, V, Ellegast, J, Kim, S, Brzozka, K, Jung, A, Kato, H et al (2006)

5′-Triphosphate RNA is the ligand for RIG-I Science 314: 994–997.

36 Pichlmair, A, Schulz, O, Tan, CP, Naslund, TI, Liljestrom, P, Weber, F et al (2006)

RIG-I-mediated antiviral responses to single-stranded RNA bearing 5′-phosphates

Science 314: 997–1001.

37 Kato, H, Takeuchi, O, Sato, S, Yoneyama, M, Yamamoto, M, Matsui, K et al (2006)

Differential roles of MDA5 and RIG-I helicases in the recognition of RNA viruses

Nature 441: 101–105.

38 Marques, JT, Devosse, T, Wang, D, Zamanian-Daryoush, M, Serbinowski, P,

Hartmann, R et al (2006) A structural basis for discriminating between self and

nonself double-stranded RNAs in mammalian cells Nat Biotechnol 24: 559–565.

39 Nishitsuji, H, Kohara, M, Kannagi, M and Masuda, T (2006) Effective suppression

of human immunodeficiency virus type 1 through a combination of short- or

long-hairpin RNAs targeting essential sequences for retroviral integration J Virol

80: 7658–7666.

40 Khvorova, A, Reynolds, A and Jayasena, SD (2003) Functional siRNAs and miRNAs

exhibit strand bias Cell 115: 209–216.

41 Schwarz, DS, Hutvagner, G, Du, T, Xu, Z, Aronin, N and Zamore, PD (2003)

Asymmetry in the assembly of the RNAi enzyme complex Cell 115: 199–208.

42 Castanotto, D, Sakurai, K, Lingeman, R, Li, H, Shively, L, Aagaard, L et al (2007) Nucleic Acids Res (in press).

43 Miyagishi, M and Taira, K (2003) Strategies for generation of an siRNA expression

library directed against the human genome Oligonucleotides 13: 325–333.

44 Gasmi, M, Glynn, J, Jin, MJ, Jolly, DJ, Yee, JK and Chen, ST (1999) Requirements for efficient production and transduction of human immunodeficiency virus type 1-based

vectors J Virol 73: 1828–1834.

45 Li, MJ, Bauer, G, Michienzi, A, Yee, JK, Lee, NS, Kim, J et al (2003) Inhibition of

HIV-1 infection by lentiviral vectors expressing Pol III-promoted anti-HIV RNAs

Mol Ther 8: 196–206.

46 Robbins, MA, Li, M, Leung, I, Li, H, Boyer, DV, Song, Y et al (2006) Stable expression

of shRNAs in human CD34+ progenitor cells can avoid induction of interferon

responses to siRNAs in vitro Nat Biotechnol 24: 566–571.