|Home | About | Journals | Submit | Contact Us | Français|
BK polyomavirus (BKPyV) is a human pathogen that causes polyomavirus-associated nephropathy and hemorrhagic cystitis in transplant patients. Gangliosides and caveolin proteins have previously been reported to be required for BKPyV infection in animal cell models. Recent studies from our lab and others, however, have indicated that the identity of the cells used for infection studies can greatly influence the behavior of the virus. We therefore wished to re-examine BKPyV entry in a physiologically relevant primary cell culture model, human renal proximal tubule epithelial cells. Using siRNA knockdowns, we interfered with expression of UDP-glucose ceramide glucosyltransferase (UGCG), and the endocytic vesicle coat proteins caveolin 1, caveolin 2, and clathrin heavy chain. The results demonstrate that while BKPyV does require gangliosides for efficient infection, it can enter its natural host cells via a caveolin- and clathrin-independent pathway. The results emphasize the importance of studying viruses in a relevant cell culture model.
BK Polyomavirus (BKPyV) is a human pathogen that was first isolated from an immunosuppressed renal transplant patient in 1971 (Gardner et al., 1971). BKPyV infection is ubiquitous in the human population, with over 80% of adults worldwide testing seropositive (Egli et al., 2009). Primary infection with and seroconversion to BKPyV usually occurs during early childhood (Kean et al., 2009), after which the virus establishes a persistent infection in the urinary tract with periodic urinary shedding (Ahsan and Shah, 2006). BKPyV infection is asymptomatic in immunocompetent individuals. However, viral reactivation in immunosuppressed transplant patients causes two major human diseases, polyomavirus-associated nephropathy (PVAN) and hemorrhagic cystitis (HC) (Jiang et al., 2009a). PVAN is one of the major causes of graft failure after kidney transplant. Up to 10% of renal transplant recipients suffer from PVAN, and approximately 50% of PVAN patients will experience allograft loss (Kuypers, 2012). In a recent study, 16.6% of the allogeneic hematopoietic cell transplant patients suffered from HC, and BKPyV was detected in 90% of the HC patients (Lunde et al., 2015). HC can cause significant morbidity with symptoms including excessive blood loss and dysuria. Taking into account that roughly 100,000 kidney and hematopoietic cell transplantation occur annually (Ramaswami et al., 2011), and that the only treatment is to reduce immunosuppression in PVAN patients, thereby increasing the likelihood of graft rejection, understanding the life cycle of BKPyV and establishing a detailed virus infection model will undeniably benefit future research and help to reveal potential drug targets.
BKPyV is small non-enveloped icosahedral DNA virus about 45 nm in diameter (Gardner et al., 1971). Each mature BKPyV particle contains a circular double-stranded genome that is approximately 5 kb in size. The early region of the BKPyV viral genome codes for early proteins including large tumor antigen (TAg), which has multiple functions essential to viral replication and is widely used as a readout to evaluate viral infection. The late region, on the other hand, codes for proteins expressed later in the viral life cycle, including the major capsid protein VP1 and the minor capsid proteins VP2 and VP3. BKPyV does not encode any DNA polymerase; thus the host DNA replication machinery is indispensable for viral replication (Jiang et al., 2012; Verhalen et al., 2015). In order to fully access the DNA replication machinery, BKPyV must traffic from outside the host cell and deliver its viral genome to the cell nucleus.
The current model for the BKPyV life cycle begins with viral attachment to the cell surface via an interaction between the major capsid protein VP1 and BKPyV receptors, b-series gangliosides (GD1b/GT1b) (Low et al., 2006; Neu et al., 2013). Specific co-receptors for BKPyV have not been identified, but data suggest that an N-linked glycoprotein with (2,3)-linked sialic acid may function as a co-receptor (Dugan et al., 2005). Caveolin-dependent endocytosis has been previously implicated in BKPyV entry into host cells, and gangliosides are commonly enriched with lipid rafts where caveolin-dependent endocytosis frequently takes place (Lingwood and Simons, 2010). After initial attachment to receptors, viral particles enter caveolin coated plasma membrane invaginations named caveolae. The caveolae coat protein, caveolin, has been reported to further guide endocytosis during BKPyV entry (Eash et al., 2004; Moriyama et al., 2007), and internalized individual BKPyV particles have been detected in tight-fitting vesicles (Maraldi et al., 1975). Disturbing caveolin-dependent endocytosis through removal of cholesterol from the plasma membrane or transfecting a dominant negative caveolin 1 mutant into the cells significantly inhibited BKPyV infection (Eash et al., 2004; Moriyama et al., 2007), while disturbing actin polymerization did not affect the BKPyV endocytic process (Eash and Atwood, 2005). After internalization, it is believed that BKPyV-containing vesicles fuse with the endosome in the same way as other polyomaviruses (Engel et al., 2011; Liebl et al., 2006; Querbes et al., 2006), and our laboratory demonstrated that acidification of the late endosome/lysosome is essential for infection (Jiang et al., 2009b). BKPyV then traffics to the ER along microtubules (Jiang et al., 2009b; Moriyama and Sorokin, 2008). After trafficking to the ER, the BKPyV capsid partially disassembles in the ER lumen, egresses from the ER via the ERAD pathway, and enters the cytosol before it translocates into the nucleus using the importin α/β pathway (Bennett et al., 2013; 2015).
Caveolin-dependent endocytosis and clathrin-dependent endocytosis are two major well-studied pathways that form small endocytic vesicles (Mayor and Pagano, 2007); in addition, there are also several caveolae- and clathrin-independent pathways that generate similar vesicles. Caveolin proteins are a family of cholesterol binding proteins that is comprised of three members (Murata et al., 1995), caveolin 1, caveolin 2, and caveolin 3. Caveolin 1 and caveolin 2 are ubiquitously expressed in cells including epithelial cells, while caveolin 3 is specifically expressed in muscle cells. Caveolin proteins were named because they all serve as caveolae coat proteins (Rothberg et al., 1992); however, only caveolin 1 has been demonstrated to be required for caveolae formation, as caveolae formation was abrogated in caveolin 1-deficient but not caveolin 2-deficient murine cells (Drab et al., 2001; Razani et al., 2002). In addition to BKPyV, SV40 and murine polyomavirus have also been thought to take advantage of caveolae-dependent endocytosis (Pelkmans et al., 2001; Richterová et al., 2001). On the other hand, JC polyomavirus (JCPyV) infects cells via clathrin-dependent endocytosis (Pho et al., 2000). The clathrin protein complex is a triskelion comprised of three heavy chains and three light chains (Fotin et al., 2004). Clathrin complexes are not capable of directly interacting with the plasma membrane. Adaptor proteins are required to recruit clathrin complexes to clathrin pits on the membrane, where the legs of the clathrin triskelions further interdigitate to form a lattice. The clathrin coated pit pinches off from the plasma membrane with help of dynamin to form clathrin-coated vesicles. Silencing clathrin heavy chain expression with siRNA has been shown to impair clathrin coated pit formation (Hinrichsen et al., 2003).
Our laboratory has demonstrated that SV40 traffics differently in monkey kidney CV-1 cells and primary human renal proximal tubule epithelial (RPTE) cells, the natural host cell for BKPyV replication (Bennett et al., 2013). Because much of the current model of BKPyV entry and trafficking is built on research using African green monkey kidney cells, our findings with SV40 drew our attention to the possibility that BKPyV may traffic differently in RPTE cells. In order to determine whether there are such differences, we have re-examined the BKPyV entry process in RPTE cells. By silencing caveolin 1, caveolin 2, clathrin heavy chain and UDP-Glucose Ceramide Glucosyltransferase (UGCG), which catalyzes the first step in ganglioside synthesis, with siRNA, we show that BKPyV infection requires gangliosides but enters RPTE cells through a caveolin- and clathrin-independent pathway.
Our lab has previously shown that BKPyV interacts with liposomes containing gangliosides GD1b and GT1b, and that adding these gangliosides to non-permissive LNCaP cells makes them permissive for BKPyV infection (Low et al., 2006). In order to continue examination of the role of gangliosides during BKPyV infection of human RPTE cells, an siRNA pool targeting UDP-glucose ceramide glucosyltransferase (UGCG) was transfected into human RPTE cells. UGCG catalyzes the first glycosylation step of its substrate, ceramide, and transforms it into cerebroside, which is required for all ganglioside synthesis (Figure 1A). Galactose, N-acetyl-galactosamine, and sialic acid are then added to cerebroside in order to synthesize gangliosides. Because a quality UGCG antibody was not available, RT-qPCR was performed to confirm efficient knockdown of UGCG. The RNA samples were harvested at two days post transfection, and UGCG mRNA expression was normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA. Compared to the no siRNA and the non-targeting siRNA controls, UGCG mRNA levels were reduced by more than 90% (Figure 1B).
We next tested the effect of knocking down UGCG on BKPyV infection. Human RPTE cells were transfected with no siRNA, a non-targeting siRNA pool, or a UGCG siRNA pool. Cells were infected with BKPyV at a MOI of 0.5 IU/cell, and at 48 hours post infection cell lysates were collected, resolved by SDS-PAGE, blotted, and assayed for TAg and GAPDH. The siRNA pool targeting UGCG reduced viral infection (Figure 1C). To further confirm that GD1b and GT1b are receptors for BKPyV infection in human RPTE cells, a rescue assay was performed. RPTE cells were transfected with the UGCG siRNA as described above. 10 hours prior to infection, gangliosides GM1, GD1b, or GT1b were added to the media, during which time the gangliosides incorporate into the cell membranes. Free gangliosides were removed by aspirating off the media and washing the cells with fresh media immediately before viral infection. Cells were infected and TAg expression was visualized by Western blotting. When siRNA that did not target UGCG was used and the gangliosides were subsequently added to the cells, the infection level stayed the same (Figure 1C). When the UGCG level was reduced, adding GM1, the receptor for SV40, did not restore infection as measured by TAg expression. However, when UGCG was reduced and GT1b or GD1b was added back, infection returned to the same level as the untransfected control. These results confirm the previous finding that gangliosides GD1b and GT1b serve as receptors for BKPyV entry in human RPTE cells.
After examining receptor used by BKPyV in RPTE cells, we next wanted to see if the virus would follow a caveolin-dependent pathway. Caveolae have been reported to guide BKPyV entry (Eash et al., 2004). In order to test whether BKPyV infects RPTE cells via the same endocytic pathway, caveolin 1 or caveolin 2 siRNA pools were transfected into cells. Each siRNA pool contained four unique siRNAs targeting the designated human mRNA. At 48 hours post-transfection protein lysates were collected, resolved by SDS-PAGE, and probed for caveolin 1 or caveolin 2. Compared to the negative controls, samples with caveolin 1 or caveolin 2 siRNA transfection showed corresponding protein expression reduction, suggesting that the siRNA transfection was successful. However, an obvious off-target effect was observed with caveolin 1-transfected RPTE cells. The caveolin 1 siRNA pool knocked down not only caveolin 1, but also caveolin 2 (Figure 2B). To eliminate the off target effect from the caveolin 1 pool, the four siRNAs from this pool were transfected into RPTE cells individually to determine which siRNAs would knock down caveolin 1 but not caveolin 2. Analyzing protein samples from cells transfected with the individual siRNAs showed that siRNA #4 was the only siRNA that was capable of knocking down caveolin 1 without affecting caveolin 2 (Figure 2A).
To test the role of caveolin 1 or caveolin 2 during viral entry, caveolin 1 siRNA #4 or the caveolin 2 siRNA pool were transfected into RPTE cells. Transfection with no siRNA and the non-targeting siRNA pool were used as negative controls, and the siRNA pool targeting UGCG or an siRNA targeting TAg were chosen as positive controls. Cells were cultured for two days after transfection in order to deplete the targeted proteins and then infected with BKPyV. Two days post infection, cell lysates were collected and TAg and caveolin protein expression was evaluated by Western blotting. The Western blotting confirmed that the caveolin 1 siRNA #4 and the caveolin 2 siRNA pool successfully reduced caveolin 1 or caveolin 2 protein levels respectively (Figure 2B). The siUGCG pool and the siRNA targeting TAg both reduced TAg expression as expected. Surprisingly, knocking down caveolin did not increase or decrease BKPyV infection. Because caveolin was not completely depleted in the knockdown cells, we wished to assay TAg expression at an earlier time point, to ensure that we were not missing subtle effects that might be masked once the viral genome begins to be amplifed. RPTE cells were infected at an MOI of 5 IU/cell and protein lysates were collected at one day post infection, the earliest time point at which we can detect TAg, and prior to the onset of DNA replication (Low et al 2004). However, we still did not observe any decreased TAg expression in cells in which the two caveolins were knocked down (Figure 2C). These results suggest that BKPyV infection is independent of caveolin 1 or caveolin 2 in RPTE cells.
Caveolin- and clathrin-dependent endocytic pathways are two major pathways that form small vesicles (Mayor and Pagano, 2007). Clathrin-dependent endocytosis has been reported to be essential for JCPyV entry (Pho et al., 2000). Since caveolins were not required for BKPyV entry into RPTE cells, we next examined if the clathrin-dependent pathway plays a role in BKPyV infection. To deplete clathrin, a siRNA pool targeting clathrin heavy chain was introduced into RPTE cells. BKPyV infection was evaluated as previously described. Compared to the no siRNA control and the non-targeting siRNA control, clathrin heavy chain was knocked down efficiently (Figure 3A). However, viral infection was not affected, indicating that BKPyV does not infect human RPTE cells by a clathrin-dependent pathway, consistent with the previous report from Moriyama et al (Moriyama et al., 2007).
Finally, we tested the possibility that both caveolin- and clathrin-dependent pathways could be involved during BKPyV endocytosis, and inhibiting one pathway can be compensated by the other pathway. To test this scenario, we transfected siRNA pools targeting clathrin heavy chain and the two caveolins into RPTE cells, and viral infection was evaluated. Similar to the single knockdowns, BKPyV infection was not affected comparing to non-targeting control by knocking down caveolin 1, caveolin 2 and clathrin heavy chain at the same time (Figure 3B).
BKPyV directly causes allograft failure among at least 1% of kidney transplant patients worldwide. Moreover, BKPyV causes painful HC in hematopoietic cell transplant patients. Unfortunately, specific antiviral drugs are not currently available, and the only management available is withdrawing immunosuppression in PVAN patients, which risks acute rejection, or palliative care in HC patients. Due to the fact that BKPyV uses the host DNA synthesis machinery for its genome replication, one possible effective strategy to protect cells from infection would be targeting critical host factors required by BKPyV before it delivers its DNA genome into the nucleus of the host cell. Therefore, understanding the trafficking pathway that BKPyV undergoes is not only important for understanding viral biology but also potentially important for identifying drug targets.
The trafficking pathways of polyomaviruses have been extensively studied since the initial isolation of murine polyomavirus in 1953 (Gross, 1953). Polyomaviruses have been shown to use three pathways to infect host cells: caveolin-dependent (Eash et al., 2004; Moriyama et al., 2007), clathrin-dependent (Pho et al., 2000), or caveolin- and clathrin-independent (Damm et al., 2005; Liebl et al., 2006). With BKPyV, most of the initial trafficking studies were carried out using non-human animal cell models because of a lack of relevant human cell models at the time. Those studies indicated that BKPyV infects monkey cells via a caveolin-dependent pathway (Eash et al., 2004). However, several pieces of evidence caused us to examine whether the BKPyV entry process might be cell type- or species-specific. First, SV40, murine polyomavirus (MPyV), and cholera toxin B subunit had been originally considered to enter their host cells via a caveolin-dependent pathway (Orlandi and Fishman, 1998; Parton et al., 1994; Pelkmans et al., 2001; Richterová et al., 2001), but subsequent studies showed that SV40, MPyV, and cholera toxin are able to enter caveolin 1-deficient cells (Damm et al., 2005; Gilbert et al., 2003; Liebl et al., 2006; Shogomori and Futerman, 2001; Torgersen et al., 2001). Second, our lab showed that SV40 trafficking is cell type-dependent (Bennett et al., 2013). Third, an in vitro study showed that the 3-hydroxy-3-methyl-glutaryl coenzyme A (HMG-CoA) reductase inhibitor Pravastatin, which inhibits cholesterol synthesis, prevents BKPyV infection by depleting caveolin 1 protein in RPTE cells (Moriyama et al., 2008); however, a clinical trial using Pravastatin failed to protect patients from PVAN at the maximum effective dose (Gabardi et al., 2015). Lastly, a study focused on BKPyV trafficking in human kidney cells supported the previous animal model that BKPyV entry is caveolin-dependent (Moriyama et al., 2007); however, we were unable to reproduce the results of Moriyama et al. by exactly following their reported protocol. In fact, our caveolin depletion was at least two fold more efficient than theirs, and viral infection was evaluated as early as one day post infection versus five days post infection as they tested. We therefore do not know the reason for this discrepancy.
These findings inspired us to re-examine BKPyV entry into RPTE cells. Our results demonstrate that BKPyV does behave differently when infecting RPTE cells as compared to monkey cells. By transfecting siRNAs targeting caveolin 1 or caveolin 2, we reduced caveolin 1 or caveolin 2 expression in RPTE cells without affecting BKPyV infection. We also tested if BKPyV infects RPTE cells using clathrin-coated vesicles, as JCPyV does in the human glial cell line SVG (Pho et al., 2000). Similar to the caveolin knockdown, there was no obvious difference in infection between clathrin heavy chain-depleted cells versus control cells. In order to eliminate the possibility that BKPyV enter RPTE cells via both caveolin- and clathrin-dependent pathways, the two caveolins and clathrin heavy chain were knocked down together, and when we challenged the transfected cells with BKPyV, no change in BKPyV infection was observed. In addition, our UGCG knockdown data confirmed previous findings from our lab that gangliosides GD1b and GT1b serve as receptors for BKPyV in primary RPTE cells. We saw a reproducible slight increase in infection in cells transfected with the non-targeting siRNA pool: we cannot explain this phenomenon.
Several caveolin- and clathrin-independent pathways that could assist BKPyV entry have been reported, such as RhoA/Rac1 (Lamaze et al., 2001), Cdc42 (Chadda et al., 2007; Sabharanjak et al., 2002), ARF6 (Radhakrishna and Donaldson, 1997), and flotillin (Glebov et al., 2006) mediated endocytosis. Considering that polyomavirus enters tight fitting vesicles after endocytosis (Damm et al., 2005; Drachenberg et al., 2003; Maraldi et al., 1975), and actin polymerization is not required for BKPyV entry (Eash and Atwood, 2005) (L.Z., data not shown), we speculate that BKPyV is more likely to infect RPTE cells by a yet uncharacterized endocytic pathway, as most of the endocytic pathways listed above are associated with actin polymerization (Burridge and Wennerberg, 2004; Chinnapen et al., 2012; Doherty and McMahon, 2009; Ilangumaran and Hoessli, 1998).
In addition to protein-dependent endocytic pathways, it is possible that BKPyV enters host cells using a lipid-mediated endocytosis pathway in which cholesterol and gangliosides alone may be sufficient to initiate entry. An in vitro assay showed that artificial liposomes with cholesterol and gangliosides, called giant unilamellar vesicles, can form caveolae-like vesicles without the addition of any host proteins (Bacia et al., 2005), and a later study showed that SV40 was able to induce deep invaginations on the surface of these vesicles (Ewers et al., 2010), suggesting that entry could be vesicle coat protein-independent. In that scenario, the virus-containing tight fitting vesicles might be formed by multiple direct interactions between gangliosides and the VP1 capsid protein, with cholesterol stabilizing the vesicle membrane invagination. Further studies are required to further define the BKPyV endocytosis process.
In summary, we have demonstrated that BKPyV enters its natural host cell via a caveolin- and clathrin-independent pathway, and we have confirmed that gangliosides GD1b/GT1b serve as receptors. Additional studies will be required to determine the role of additional cellular proteins in the viral entry and trafficking processes.
Primary renal proximal tubule epithelial cells (RPTE) (Lonza) were grown in renal epithelial basal growth medium (REBM) (Lonza) with renal epithelial cell growth medium (REGM) supplement (Lonza) at 37°C with 5% CO2 in a humidified incubator.
siGENOME siRNA pools and individual siRNAs were purchased from Dharmacon: Non-targeting control (D-001206-14); Caveolin 1 (M-003467-01); Caveolin 2 (M-01095800); UGCG (M-006441-02); Caveolin 1 #1 (D-003467-01); Caveolin 1 #2 (D-003467-02); Caveolin 1 #3 (D-003467-03); Caveolin 1 #4 (D-003467-05); Clathrin heavy chain (M-004001-00). siRNA targeting large T antigen was custom synthesized with the sequence (5’ AUCUGAGACUUGGGAAGAGCAU 3’), which corresponds to the natural BKPyV 5p miRNA (Seo et al., 2008). siRNAs were rehydrated at 20 µM according to the Basic siRNA Resuspension protocol from Dharmacon. RPTE cells were reverse transfected according to the Lipofectamine RNAiMAX (Thermo Fisher Scientific) manual. Briefly, transfection complexes were prepared by mixing 1 µl of 20 µM siRNA with 400 µl of diluted transfection reagent (0.7% RNAiMAX reagent v/v in REBM/REGM without antibiotics) in each well of a 12 well plate. The complexes were incubated at room temperature for 20 min before adding 70,000 cells suspended in 400 µl REBM/REGM without antibiotics to each well.
BKPyV (Dunlop) was cultured, purified with a cesium chloride linear gradient, and titered as described previously (Abend et al., 2007; Jiang et al., 2009b). RPTE cells were infected as follows at 2 days post transfection. Cells were pre-chilled for 15 min at 4°C. Purified viruses were diluted to 87,500 IU/ml in REBM/REGM. 400 µl of the diluted virus were added to the wells and incubated at 4°C for 1 hour with shaking every 15 minutes to distribute the inoculum over the entire well. The plate was transferred to 37°C after the 1-hour incubation.
Gangliosides GM1, GD1b, and GT1b (Matreya) were kindly provided by Billy Tsai (University of Michigan). Gangliosides were rehydrated at 1 mM in cell culture grade water. 10 hours prior to infection, gangliosides were added directly to the media at 1 µM. At the time of infection, media with gangliosides was removed and cells were washed with fresh media to remove the unincorporated gangliosides.
Cells were lysed at 48 hours post infection with E1A buffer (50 mM HEPES [pH 7], 250 mM NaCl, and 0.1% NP-40, with inhibitors: 5 µg/ml PMSF, 5 µg/ml aprotinin, 5 µg/ml leupeptin, 50 mM sodium fluoride and 0.2 mM sodium orthovanadate added right before use). Protein concentration was quantified with the Bradford assay (Bio-Rad).
Protein samples were separated on 12% SDS-PAGE gels. After electrophoresis, the proteins were transfer to a nitrocellulose membrane (pore size 0.2 µm) with a Bio-rad Trans-Blot Cellin Towbin transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol), at 60 V, overnight. Membranes were blocked with 2% nonfat milk in PBS-T buffer (144 mg/L KH2PO4, 9 g/L NaCl, 795 mg/L Na2HPO4, pH 7.4, 0.1% Tween 20) for 1 hour. Membranes were probed with primary and secondary antibodies diluted in 2% milk in PBS-T as follows: TAg (pAb416) at 1:5,000 dilution (Harlow et al., 1986); GAPDH (Abcam ab9484) at 1:10,000; caveolin 1 (Santa Cruz SC-894) at 1:20,000; caveolin 2 (Cell signaling #8522) at 1:5,000; Clathrin heavy chain (Thermo Fisher Scientific, MA1-065) was kindly provided by Christiane Wobus (University of Michigan), used at 1:10,000, horseradish peroxidase (HRP)-conjugated ECL sheep anti-mouse (GE healthcare NA931V) at 1:5,000; and HRP-conjugated ECL donkey anti-rabbit antibody (GE healthcare NA934V) at 1:5,000. Protein bands were further visualized with HRP substrate (Millipore, WBLUF0100) and exposure to films.
At 48 hours post transfection, cellular RNA was harvested with TRIzol RNA isolation reagent (Thermo Fisher Scientific) and further purified according to the manual for the Zymo Research Direct-zol RNA MiniPrep kit. 100 ng of RNA was treated with 1 unit of DNase I (Promega) and cDNA was prepared according to the SuperScript II Reverse Transcriptase (Thermo Fisher Scientific) manual. qPCR was performed on the cDNA with either of the following primer pairs: GAPDH (5’ GCCTCAAGATCATCAGCAAT 3’) and (5’ CTGTGGTCATGAGTCCTTCC 3’), UGCG (5’ AGACACCTGGGAGCTTGCTA 3’) and (5’ TTCGTCCTCTTCTTGGTGCT 3’). UGCG or GAPDH expression in each cDNA sample was quantified in 25 µl reactions and in triplicate. Each reaction was comprised of 2.5 µl of cDNA, 12.5 µl of Power SYBR green PCR master mix (Applied Biosystems), 300 nM of each primer, and nuclease free water (Promega). Amplification was performed using the iCycler iQ5 real-time detection system (Bio-Rad) with the following PCR protocol settings: 2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C and 30 s at 57°C. SYBR green signal intensity was read immediately after each 57°C extension.
We thank Heather Manza for help with this work, and Billy Tsai and Christiane Wobus for critical review of the manuscript and sharing reagents with us. This work was supported by NIH grant AI060584 awarded to M.J.I., and a Rackham Graduate Student Research grant awarded to L.Z.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.