To inhibit hSpt5 expression in a cultured human cell line using RNAi, siRNA targeting an hSpt5 sequence from position 407 to 427 relative to the start codon was designed (Figure 1A). Magi cells were transfected with this hSpt5 duplex siRNA using Lipofectamine (Invitrogen). To evaluate the effects of hSpt5 RNAi, total cell lysates were prepared from siRNA-treated cells harvested at various time points after transfection. hSpt5 mRNA or protein levels were analyzed by RT-PCR or western blot using anti-hSpt5 antibodies, respectively. Cells transfected with hSpt5 siRNA had significantly lowered hSpt5 mRNA (Figure 1B, lane 3) and protein expression (Figure 1C, lane 3), indicating that siRNA-mediated silencing of hSpt5 had occurred successfully. hSpt5 knockdown was consistently between ~85–90%. This knockdown effect was dependent on the presence of the 21-nt siRNA duplex harboring a sequence complementary to the mRNA target. As shown in Figures 1B and 1C, mock-treated (no siRNA) (lane 1), single-stranded antisense hSpt5 siRNA (lane 2), or mismatched hSpt5 duplex siRNA (lane 4) containing two nucleotide mismatches between the target mRNA and siRNA antisense strand at the putative cleavage site of the target mRNA (Figure 1A) did not affect hSpt5 mRNA or protein levels. These results showed that hSpt5 knockdown was specific to duplex siRNA exactly complementary to the hSpt5 mRNA target. In evaluating either mRNA or protein levels, human Cyclin T1 (hCycT1) was used as an internal control, showing that the effects of hSpt5 siRNA were specific to hSpt5 and did not effect hCycT1 mRNA or protein expression (Figure 1B and 1C, lower panel). Taken together, these results demonstrated that hSpt5 knockdown was sequence specific and led to significantly decreased hSpt5 mRNA and protein levels.

Having established that hSpt5 could be knocked down using RNAi, the kinetics of hSpt5 knockdown were examined. To perform kinetic experiments, hSpt5 duplex siRNA, single-stranded antisense hSpt5 siRNA, or mismatch duplex hSpt5 siRNA were transfected into Magi cells. Cell lysates were collected at various time points to assay for protein levels during hSpt5 knockdown. Immunoblot analysis using anti-hSpt5 antibodies revealed the timing of gene suppression and persistence of hSpt5 RNAi effects in Magi cells during the time course experiment (Figure 2). hSpt5 knockdown was first observed between 30–42 h post-transfection, with maximum knockdown (~85–90% knockdown) occurring at 42–66 h post transfection (Figure 2, lane 8–14). Protein levels gradually recovered to normal levels between 66–90 h (data not shown), indicating that the effects of hSpt5 siRNA did not last indefinitely. Neither single-stranded antisense siRNA (Figure 2, lanes 1–7) nor mismatched duplex siRNA (Figure 2, lanes 15–21) affected hSpt5 protein levels throughout the duration of the time course. These results indicated that hSpt5 knockdown by RNAi occurred after 30 h and these knockdown effects were specific to duplex siRNA targeting hSpt5.

Knowing that the kinetics of hSpt5 peaked at 42–54 h post-transfection, we were able to evaluate the viability of cells during hSpt5 knockdown experiments over varied time intervals. Cell viability was assessed using trypan blue exclusion at various times after a single transfection of various siRNAs. As shown in Figure 3, during the 66 h time course experiment, the number of non-viable hSpt5 knockdown cells (yellow line) observed was comparable to mock-treated cells (no siRNA; dark blue line). Cells transfected with single-stranded antisense hSpt5 siRNA (purple line) or mismatched hSpt5 duplex siRNA (light blue line) that did not show hSpt5 knockdown also showed minimal changes in cell viability.

hSpt5 has been shown to interact with the human mRNA capping enzyme (HCE) and this interaction enhances capping enzyme guanylylation and mRNA capping several fold 40. Since Spt5, Tat, and CTD associate with the capping apparatus to stimulate capping 363739404142, we planned to separately define the role of HCE and hSpt5 in Tat transactivation by using RNAi to specifically knockdown HCE expression. HCE knockdown was confirmed by RT-PCR (data not shown). In contrast to hSpt5 knockdown cells, HCE knockdown cells showed a significant increase in cell death (Figure 3, red line) over the course of the knockdown experiment. These results indicated that HCE is an essential enzyme for cell viability and growth. Simialr findings showing that RNA capping was essential for metazoan viability have also been previously reported using RNAi in C. elegans 58. These results indicated that hSpt5 knockdown was not lethal to human cells, while a much more stringent requirement for HCE expression was essential for cell viability.

To examine whether hSpt5 was required for HIV-1 Tat transactivation in vivo, Tat transactivation during hSpt5 knockdown in Magi cells was monitored. Magi cells are a HeLa cell line harboring a stably integrated single copy of the HIV-1 5' LTR-β-galactosidase gene. These cells are also genetically programmed to express the CD4 receptor for HIV-1 infection (59; see below). In this experiment, Magi cells were co-transfected with Tat expression plasmid pTat-RFP and hSpt5 duplex siRNA. Co-transfection with Tat siRNA was used as a positive control for inhibition of Tat transactivation while single-stranded antisense hSpt5 siRNA and mismatched siRNA were used as negative controls. Tat transactivation and protein levels were evaluated by harvesting cells 48 h post transfection, which was within the timeframe that hSpt5 knockdown peaked. Expression of HIV-1 Tat-RFP under the control of the CMV early promoter was confirmed by western blot using anti-RFP antibody and by measuring RFP fluorescence using a fluorescence spectrophotometer (data not shown). In addition, immunoblot analysis confirmed that hSpt5 siRNA specifically inhibited hSpt5 protein expression in the absence and presence of HIV-1 Tat protein in Magi cells (data not shown).

Tat-RFP enhances the expression of genes that are under the control of the HIV-1 5' LTR promoter. In this experiment, Tat transactivation was measured by assaying the β-galactosidase activity resulting from expression of the β-galactosidase gene under the HIV-1 5' LTR promoter. To quantify the effects of various siRNAs on HIV-1 Tat transactivation, the ratio between β-galactosidase activity in cells transfected with pTat-RFP (with or without siRNAs) and mock-treated cells not transfected with pTat-RFP was determined. The results of this quantitation are shown in Figure 4. In Magi cells, Tat-RFP strongly stimulates the expression of β-galactosidase, represented by a 13-fold increase in Tat transactivation (Figure 4, lane 1). On the other hand, Tat transactivation was strongly inhibited in cells transfected with Tat siRNA (~90% knockdown; Figure 4, lane 5), as previously shown 51. Tat transactivation was similarly inhibited when cells were transfected with hSpt5 duplex siRNA, exhibiting only ~30% of the Tat transactivation observed with Tat-RFP alone (Figure 4, lane 3). Neither antisense hSpt5 siRNA nor mismatched hSpt5 siRNA (Figure 4, lane 4) showed any effect on Tat transactivation. These results indicated hSpt5 knockdown caused by siRNA specifically targeting hSpt5 mRNA inhibited HIV-1 Tat transactivation in human cells. These results strongly supported an important role for hSpt5 in Tat transactivation in vivo and suggested that RNAi of hSpt5 had the potential to inhibit HIV-1 replication.

To evaluate the effect of hSpt5 knockdown on HIV-1 replication, a double siRNA transfection protocol was used to maximize the knockdown efficiency of hSpt5 during HIV-1 infection. Magi cells were transfected with siRNA directed against hSpt5. Cells mock transfected without siRNA, or transfected with single-stranded antisense hSpt5 siRNA or mismatch hSpt5 siRNA were used as negative controls. Transfection with Vif or Nef siRNAs was used as a positive control 20. 24 h after the first transfection, a second siRNA transfection identical to the first was performed. 24 h later, doubly transfected cells were infected with various concentrations of HIV_NL-GFP, an infectious molecular clone of HIV-1. Knockdown of hSpt5 protein levels was then evaluated 48 h post infection in doubly transfected cells. An even larger decrease in hSpt5 protein levels was observed in doubly transfected cells (~95% knockdown) as compared to singly transfected cells (~85–90% knockdown; Supplementary Figure 1, compare lanes 4 and 10), suggesting that more robust knockdown of gene expression can be achieved using this double transfection method.

HIV-1 Tat-mediated transactivation of the 5' LTR occurring in cells infected with virus led to β-galactosidase production, which was also quantified 48 h post-infection. In this single-cycle replication assay for evaluating HIV-1 replication, β-gal activity reflected the activity of reverse transcriptase and viral replication of varying amounts of viral inoculum. Therefore, changes in β-gal activity could be directly correlated to changes in the efficacy of HIV-1 replication. The positive siRNA control targeting HIV-1 Vif showed decreased levels of β-gal activity and HIV-1 replication, as shown previously (Figure 5; 47). Double-stranded siRNA directed against hSpt5 showed a similar decrease in β-gal activity when compared with Vif knockdown. This observed decrease was equivalent to the β-gal activity measured when using 32 times less viral inoculum with mock-treated cells (Figure 5), indicating that hSpt5 knockdown had significantly reduced HIV-1 replication. p24 levels were also monitored during these experiments and decreased in the context of hSpt5 knockdown (data not shown), supporting the conclusion that hSpt5 knockdown has a negative effect on the HIV-1 life cycle. Control experiments using hSpt5 single-stranded antisense or mismatched duplex siRNA duplexes showed no antiviral activities. In addition, no significant toxicity or cell death was observed during these experiments, suggesting that hSpt5 knockdown was not lethal even in the context of HIV-1 infection. These results demonstrated that hSpt5 silencing using RNAi modulated HIV-1 replication and firmly established an important role for hSpt5 in Tat transactivation and HIV-1 replication in vivo.

Twenty-one nucleotide siRNAs were chemically synthesized as 2' bis(acetoxyethoxy)-methyl ether-protected oligos (Dharmacon, Lafayette, CO). Synthetic RNAs were deprotected, annealed and purified using standard protocols provided by the manufacturer. Formation of duplex RNA was confirmed by 20% non-denaturing polyacrylamide gel electrophoresis (PAGE). Sequences of siRNA duplexes were designed as described previously 46 and subjected to a BLAST search against the NCBI EST library to ensure that only the desired genes were targeted.

Magi (

m

a

g

i

Total cellular mRNA was prepared from HeLa cells transfected without siRNA or with hSpt5 or control siRNAs using a Qiagen RNA mini kit, followed by an oligotex mRNA mini kit (Qiagen). RT-PCR was performed using a SuperScript One-Step RT-PCR kit with platinum Taq (Invitrogen) and 40 cycles of amplification. Each RT-PCR reaction included 100 ng total cellular mRNA, gene-specific primer sets for hSpt5 and hCycT1 amplification (0.5 μM for each primer), 200 μM dNTP, 1.2 mM MgSO₄ and 1U of RT/platinum Taq mix. Primer sets for hSpt5 produced 2.6 kb products while hCycT1 produced 1.8 kb products. RT-PCR products were resolved on a 1% agarose gel and viewed by ethidium bromide staining. Forward and reverse primer sequence for amplifying SPT5 were 5'-ATGTCGGACAGCGAGGACAGC-3' (nts 1–21) and 5'-TGTACATGGCCGGCGTCCC-3' (nts 2638–2656), respectively. Forward and reverse primer sequences for amplifying hCycT1 are 5'-GCAACAAGTTCAAGATCTGGTCAT-3' (nts 381–404) and 5'-

CCCGGGGGATCC

siRNA treated cells were harvested as described above and lysed in 1X reporter lysis buffer (Promega). After centrifugation to remove cellular debris, concentrations of proteins in lysates were determined using a Dc protein assay kit (Bio-Rad). Proteins in 30 μg of total cell lysates were fractionated by 10% SDS-PAGE, transferred onto a polyvinylidene difluoride membrane (PVDF membrane, Bio-Rad), and immunoblotted with antibodies against hSpt5 (Pharmingen) and hCycT1 (Santa Cruz Biotech). Protein content was visualized by a BM chemiluminescence Blotting Kit (Roche Molecular Biochemicals). The blots were exposed to X-ray film (Kodak MR-1) for various times (between 1 s and 5 min).

Magi cells were harvested 48 h after transfection with Tat-RFP plasmids in the absence or presence of siRNAs. Cell lysates were prepared and quantified as described above. To perform standard β-galactosidase assays, 120 μg of cell lysates were mixed in 150 μl of 1X reporter lysis buffer and 150 μl of 2X β-galactosidase assay buffer (Promega), and incubated at 37°C for 30 min. To stop the reaction, 500 μl of 1 M sodium carbonate was added to the mixture and mixed well by vortexing briefly. Absorbance of the reaction mixture was read immediately at 420 nm. The amount of Tat-RFP protein was determined using a fluorescence spectrophotometer (Photon Technology International). 300 μg of cell lysates was subjected to the spectrophotometer with slit widths set at 4 nm for both excitation and emission wavelengths as described previously 4664. Fluorescence of Tat-RFP in the cell lysate was detected by exciting at 558 nm and recording the emission spectrum from 568 nm to 650 nm; the spectrum peak at 583 nm represents the maximum fluorescence intensity of Tat-RFP. Tat transactivation was determined by calculating the ratio of β-galactosidase activity (absorbance at 420 nm) of the pTat-RFP transfected cells to that of cells without pTat-RFP plasmid treatment. The inhibitory effect of siRNA treatment was determined by normalizing Tat-transactivation activity to the amount of Tat-RFP protein (represented by RFP fluorescence intensity) in the presence and absence of siRNA.

HeLa-CD4-LTR/β-gal indicator (Magi) cells 59 were plated in 24-well plates (7.5 × 10⁵ cells per well) and transfected with siRNAs as previously described 47. siRNA (60 pmol) was transfected into cells using oligofectamine (2 μl, Invitrogen) for 3 h in serum-free DMEM (GIBCO). Cells were rinsed twice and top-layered in 500 μl of DMEM-10% FBS. 24 h after the first transfection, a second identical siRNA transfection was performed. 24 h after the second transfection, cells were trypsinized and seeded in 96-well microtiter plates (4 × 10⁴ cells per well), incubated 3 h and infected with HIV_NL-GFP, an infectious molecular clone of HIV-1. HIV-1 virions (normalized to RT activity in cpm) were added in doubling dilutions to duplicate wells. 48 h post infection, cells were harvested to quantify β-galactosidase activity and protein levels.