RNA Analysis

2.2.4.1 RT-PCR

2.2.4.1.1

mRNA extraction

Following RNA extraction as described in section 2.2.2.5, mRNA was extracted from the total cellular RNA using the Qiagen Oligotex mRNA extraction kit as

described in the manufacturer’s instructions. Briefly, 250µg of RNA in 250µl of DEPC-H2O was added to 250µl of Buffer OBB and 15µl of Oligotex solution and mixed. The RNA was incubated at 70°C for 3 minutes to denature any secondary structure before incubating at 20°C for 10 minutes to allow the poly A tail of the mRNA to bind to the Oligotex particles. The RNA was centrifuged at 16110 g for 2 minutes in an eppendorf 5415 R bench top centrifuge and the supernatant removed. The Oligotex pellet was resuspended in 400µl Buffer OW2 and transferred to a spin column. The column was centrifuged at 16110 g for 1 minute and the flow through discarded. The column was transferred to a new 1.5ml eppendorf tube and 400µl Buffer OW2 applied to the column. The RNA was again centrifuged at 16110 g for 1 minute and the flow through discarded. The spin column was again transferred to a new 1.5ml eppendorf tube and 20µl of Buffer OEB preheated to 70°C was added and the mRNA resuspended. The mRNA was centrifuged at 16110 g for 1 minute before a second 20µl of hot Buffer OEB was added to maximise the mRNA yield. The mRNA was centrifuged again at 16110 g for 1 minute and the eluted mRNA stored at -20°C.

2.2.4.1.2

DNase-1 treatment for RT-PCR

After RNA extraction as described in section 2.2.2.5, any remaining DNA in the RNA extract was removed by DNase-1 digestion using the Promega RQ1 DNase-1 kit as described in the manufacturer’s instructions. 1U of RQ1 was used per 1µg of RNA. Reaction buffer to a final concentration of 1X was added to the reaction and the digestion carried out for 1 hour at 37°C. 1µl of DNase-1 stop solution was added per µg of RNA and incubated at 65°C for 10 minutes to stop the

2.2.1.6 to remove enzyme and excess salts from the sample that may inhibit downstream applications.

2.2.4.1.3

Reverse Transcription

After DNase-1 digestion, the RNA was reverse transcribed into cDNA using the Superscript III kit (Invitrogen) following the manufacturer’s instructions. 1µl of 10mM dNTP mix and 1µl of Oligo dT20 (50µM) primer was added to 1µg of RNA and incubated at 65°C for 5 minutes to disrupt any secondary structures. The reactions were placed on ice and 2µl 10X RT buffer (200mM Tris-HCl (pH 8.4), 500mM KCl), 4µl 25mM MgCl2, 2µl 0.1M DTT, 1µl 40U/µl RNase OUT and 1µl (1U/µl) of either Superscript III or DEPC-treated H2O (for the RT- controls) were added. The reactions were incubated at 50°C for 50 minutes before being incubated at 85°C for 5 minutes to stop the reaction. The samples were placed on ice and 2U (1µl) RNase H was added and incubated at 37°C for 20 minutes to degrade the initial mRNA strands.

2.2.4.1.4

PCR

2µl of cDNA (~100µg) or 1µl (~0.5µg) of miniprep DNA was amplified in a PCR reaction using Taq DNA polymerase (Invitrogen). Reactions were carried out in 50µl volumes with 38.1µl DEPC treated H2O, 5µl 10X PCR Buffer (200mM Tris-HCl (pH 8.4), 500mM KCl), 1.5µl 50mM MgCl2, 1µl 10mM dNTPs, 1µl gene specific forward primer (10µM), 1µl gene specific reverse primer (10µM), 2µl cDNA and 0.4µl 5U/µl Taq DNA polymerase. The final concentration for the PCR reaction was 1X PCR Buffer, 200nM Primers, 200µM dNTPs, 1.5mM MgCl2 and 2U Taq DNA

polymerase. PCRs were carried out on a Hybaid PCR Express machine with a general program consisting of:

95°C – 2 minutes 95°C – 30 seconds 55°C – 30 seconds 72°C – 1 minute 72°C – 5 minutes 55°C – 5 minutes 72°C – 10 minutes 4°C – hold

PCR products were resolved on acrylamide gels as described in section 2.2.1.9.1.

2.2.4.2 qRT-PCR

2.2.4.2.1

DNase-1 Treatment for qPCR

After RNA extraction as described in section 2.2.2.5, any remaining DNA in the RNA extract was removed by DNase-1 digestion using the Ambion TURBO DNase-1 kit as described in the manufacturer’s instructions. RNA was diluted to 50ng/µl in a 30µl volume. 3µl 10X reaction buffer and 1µl of 2U/µl TURBO DNase was added to the reaction and the digest carried out for 30 minutes at 37°C. 3.5µl of Inactivation Reagent was added and incubated at room temperature for 5 minutes to stop the reaction. The RNA was then centrifuged at 13 200 rpm for

25-30 cycles

1.5 minutes in an eppendorf 5415 R bench top centrifuge and the supernatant recovered to a fresh tube. The centrifugation step was repeated and the recovered RNA stored at -20°C.

2.2.4.2.2

Reverse transcription

After DNase-1 digestion the RNA was reverse transcribed into cDNA using the Affinityscript qPCR cDNA synthesis kit (Stratagene) following the manufacturer’s instructions. 10µl of 2X Mastermix, 2µl of Oligo dT20 primer, 1µl random primers (0.1µg/µl) and either 1µl Affinityscript RT/RNase Block enzyme was added to 6µl of DNase-1 treated RNA and incubated with the following program on the Hybaid PCR Express machine: 25°C – 5 minutes 42°C – 45 minutes 95°C – 5 minutes 4°C - hold

2.2.4.2.3

qPCR

After reverse transcription or DNA extraction, cDNA or DNA was amplified using an Applied Biosystems 7500 qPCR machine. For some targets, for example many human genes, probe and primer sets are commercially available and have been optimised and shown to have close to 100% efficiency. Therefore when possible these Taqman predesigned gene expression assays were ordered. However for other targets, for example viral genes, there are no commercially available assays, therefore these assays had to be custom designed and synthesised. The Stratagene Brilliant qPCR mastermix was used for qPCR reactions. Reactions

were prepared as follows: 12.5µl of 2X Mastermix, 2.25µl of 10µM forward primer, 2.25µl of 10µM reverse primer, 0.5µl of 5µM probe, 0.5µl of 15µM reference dye, 5µl of DEPC-water and 2µl cDNA or 100ng DNA. If a Taqman predesigned gene expression assay was being used then the reactions are

prepared as follows: 12.5µl of 2X Mastermix, 1.25µl gene expression assay, 0.5µl of 15µM reference dye, 10.75µl of DEPC-water and 2µl (~15ng) cDNA or 100ng DNA. The amplification protocol consisted of 50°C for 2 minutes, 95°C for 10 minutes, 40 cycles at 95°C for 15 seconds followed by 60°C for 1 minute. To calculate the relative efficiencies of probe and primer sets, cDNA or plasmid standard curves were generated and the efficiency calculated from the slope of the line using the formula Efficiency = 10(-1/slope)-1. For probe and primer sets to be used in a relative quantification calculation the efficiencies must be within +/- 10% of each other. The predesigned assays were all guaranteed to be 100% efficient therefore no standard curve was performed. Relative quantification was not used with these assays. The relative efficiencies for my custom E6 isoform probe and primer sets were extremely varied out with +/- 10% of each other and could not be used for relative quantification.

2.2.4.3 DNA sequencing

Sequencing of PCR products or cloned vectors was carried out on an ABI PRISM 3130XI Genetic Analyzer machine. All sequencing reactions were carried out using the Applied Biosystems Big Dye Mix and 5X sequencing buffer. The DNA to be sequenced was amplified as follows: 2 µl Big Dye Mix, 4 µl 5X Sequencing Buffer, 2 µl 10 µM sequencing primer, 6 µl ddH2O and 6 µl (~3µg) of DNA from a

standard miniprep. The products were amplified using the PCR Express machine as follows 95°C - 10 sec 50°C - 5 sec 60°C - 4 min 4°C - hold

Products were purified using the EdgeBio Spin column purification method. The columns were centrifuged for 2 minutes at 3K in an eppendorf 5415 R bench top centrifuge to remove storage solution. The collection tube was discarded and the spin column inserted into a fresh collection tube. The PCR sequencing reaction was transferred onto the centre of the gel in the spin column and the column centrifuged for 2 minutes at 3K in an eppendorf 5415 R bench top centrifuge. The DNA was dried under vacuum for 15-20 minutes and the pellet resuspended in 25 µl of Hi-Di Formamide (Applied Biosystems).

3 Results 1

3.1 Introduction

A large body of scientific evidence has led to the conclusion that E6 and E7 are the major oncoproteins of the high risk HPVs. However, one problem with our current knowledge is that the high risk virus E6 and E7 mRNAs can be

alternatively spliced yet it is not known which E6/E7 mRNA isoform(s) contribute to the transformed phenotype. Interestingly, low risk virus E6/E7 mRNAs appear not to be alternatively spliced suggesting that high risk HPV oncogenes have evolved to encode alternative protein isoforms that contribute to the virus replication cycle in some way but also may be involved in cancer progression. In particular, HPV16 E6 can be alternatively spliced with the production of at least four isoforms. The first reports of E6 being alternatively spliced came in 1986 when Smotkin and Wettstein identified two spliced isoforms of HPV16 E6 after RT-PCR analysis of RNA isolated from transformed cells and tumour biopsies (Smotkin & Wettstein, 1986). In this study they identified the largest spliced product, now called E6*I as being the predominant isoform in transformed cells. It was suggested that this spliced product in fact allows for efficient translation of E7 protein. The initiation codon for E7 is only two bases downstream from the termination codon of E6. Therefore the hypothesis is that there is not enough RNA between the two open reading frames to allow for efficient

translation re-initiation. Splicing of the RNA and termination of the E6 isoform protein at an earlier stop codon allows more space between the E6 isoform termination codon and the E7 start codon. However, further studies have since

demonstrated that E7 is translated as efficiently from E6/E7 bicistronic mRNAs as from spliced E6 mRNAs (Stacey et al., 1995). Furthermore, the assumption from these earlier studies was that, as E6*I is the predominant isoform in cervical tumour cells, the isoform may have tumour promoting functions. However, a number of studies over many years have suggested that this is

unlikely to be true. As will be discussed, E6*I appears to have opposing functions to E6 with regards to a number of cellular interacting partners.

In 2006, Tang et al. reported the possibility of four mRNA splice isoforms being produced from the open reading frames encoding E6 and E7 (Figure 3.1) (Tang et al., 2006). One splice donor site was located in the E6 open reading frame with two splice acceptor sites in the E6 open reading frame and one in the E7 open reading frame (Tang et al., 2006). The two splice acceptor sites in the E6 coding region were those already identified in previous studies: E6*I and E6*II. A third acceptor site was identified in the E7 region and the putative isoform was named E6*X. There have been comparatively few reported functions for E6*I compared to the number of reports of full length E6 functions and no reported functions for E6*II and E6*X.

The first report of any function for HPV16 E6*I came from Shirasawa et al. in 1994 where they showed that while E6 has a repressive effect on the P97 major early promoter, low levels of E6*I could in fact trans-activate the promoter, indicating that E6*I may have completely separate functions to E6 (Shirasawa et al., 1994). The next report of E6*I function came from studying the effects of E6

Figure 3.1 Diagram of HPV16 E6 isoforms. Schematic diagram of the E6/E7

isoforms with the splice donor (SD) and acceptor (SA) sites labelled and the nucleotide numbers given.

and E6*I on p53. As already stated, E6 targets p53 for proteasomal degradation to inhibit growth arrest and apoptosis in HPV-infected cells. Shalley et al. (1996) and Pim et al. (1997) studied HPV16 E6 and HPV18 E6 respectively and found that the spliced E6 product acts to protect p53 from the degradation induced by full length E6 proteins (Shally et al., 1996, Pim et al., 1997). Pim et al. (2007) also demonstrated that expression of E6*I in CaSki cervical cancer cells decreased the levels of cellular proliferation similar to expression of p53 in CaSki cells (Pim et al., 1997). This finding suggests that E6*I, at least for HPV18, is in fact acting to suppress the transformed phenotype of the cells.

E6 has been shown to have anti-apoptotic effects, therefore many of the studies into E6*I function have analysed levels of apoptosis. E6*I was shown to have an opposing effect from E6 full length on procaspase 8. Filippova et al. (2007) reported that the full length E6 protein could bind to procaspase 8 and

accelerate its degradation thus preventing U2OS cells from entering apoptosis (Filippova et al., 2007). In contrast, although the smaller E6*I isoform could still bind procaspase 8, it stabilised it thus making the cells susceptible to apoptosis (Filippova et al., 2007). The authors explained this seemly paradoxical

interaction of the two E6 isoforms with procaspase 8 by suggesting that the ratio of the two isoforms may be the important factor in determining procaspase 8 levels and therefore the cells’ potential to enter apoptosis (Filippova et al., 2007). Furthermore, the procaspase 8 binding domains of E6 and E6*I have been identified and are not the same, perhaps explaining why two similar proteins can have such diverse effects (Filippova et al., 2007). This suggests that there could be a very delicate balance of E6 expression in HPV-infected cells. Similar results

have been demonstrated for FADD and TNF R1 where both the full length E6 and E6*I have opposing effects on stabilising or destabilising the molecules upon protein-protein interaction (Filippova et al., 2002, Filippova et al., 2004, Filippova et al., 2009). The targeting of these apoptosis effector molecules shows that E6 very tightly regulates the extrinsic apoptosis pathways. In these studies E6 and E6*I were shown to be able to form complexes which may also be important in controlling E6 action (Filippova et al., 2009). The levels of E6 in the cell were also reported to be important as low levels of E6 and E6*I make cells resistant to TNF induced apoptosis (Filippova et al., 2009). E6-mediated control over apoptosis pathways is extremely complex and it appears to require a very delicate balance of E6 levels and isoform ratios.

Pim et al. (1997) also demonstrated a similar scenario when studying the E6- mediated degradation of PDZ domain-containing proteins. They showed that E6*I could direct the degradation of the tumour suppressor hDlg as efficiently as E6 full length and that the two proteins appeared to cooperate to allow efficient degradation (Pim et al., 1997). Again the ratio between the two isoforms was reported to be important for regulating the function of E6 (Pim et al., 1997).

Finally, a recent report showed that splicing of E6 is regulated by the EGF pathway. Rosenberger et al. (2010) demonstrated the requirement for EGF for production of full length E6 transcripts and that a reduction in EGF levels altered the E6 splicing pattern into producing mainly E6*I transcripts (Rosenberger et al., 2010). This finding is very interesting as the authors suggested that perhaps

full length E6 is required initially to avoid apoptosis after infection (Rosenberger et al., 2010). There is a concentration gradient of EGF in the differentiating epithelium and the levels of EGF are highest in the basal epithelial layer. Rosenberger et al propose that as the infected cell moves up through the

epithelium the requirement for full length E6 may change and E7 production, to keep the infected cell cycling, becomes more important. Therefore the EGF levels drop, E6*I transcripts becomes more predominant and translation of E7 is enhanced (Rosenberger et al., 2010). This hypothesis is still controversial

because there is conflicting evidence that E6*I allows enhanced E7 production. However, this study provides the first suggestion of a switch between isoform production being important for the virus life cycle. It is also important to note that the E6*I transcript is the dominant isoform in transformed cells and the levels of E6*I have been reported to be indicative of severity of the clinical lesion (Kosel et al., 2007). This may also fit with our current knowledge of E6 splicing control: if lesions express higher levels of E6*I and if the original hypothesis of alternative splicing of E6 providing enhanced translation of E7 is true, and there is a proliferative advantage of expressing high E7 levels, this may explain why E6*I levels are high in lesions.

As already mentioned there have so far been no reported functions for E6*II and E6*X. In my opinion, however, due to the vast number of E6 protein-interacting partners already identified, I think it is entirely possible that one or more of the E6 isoforms may be involved in some of the functions attributed to the E6

protein. For example all isoforms are spliced from the same splice donor site therefore all proteins have the same N-terminus. Given the presence of the

same N-terminus in all E6 isoforms these could still carry out any functions

mapped to this domain in the full length E6 protein, for example part of the first zinc finger domain is retained in the E6 isoforms however it will be unable to form the loop finger structure, therefore E6AP could still bind if the structure is not important for binding. All isoforms lack the entire C-terminus of E6

including the second zinc finger domain and the whole PDZ domain. Splicing in all isoforms occurs at splice donor 226 which is equivalent to amino acid 41 meaning that the first CXXC motif at amino acid 30 is present in all isoforms but the second at amino acid 63 is absent. This means the zinc finger structure cannot form; therefore it is likely that interactions dependent on this first zinc finger domain will also be lost. E6*I is a small protein and after splicing only two additional amino acids are added after the splice junction before meeting a frameshift termination codon. Therefore it is likely that only interactions attributed to full length E6 N-terminus (amino acids 1-41) will be able to be fulfilled by E6*I. However as E6*I still conatins the L1 and S1 domains (Figure 1.5) it could retain the capacity to bind full length E6 and block folding between the N-terminus and C-Terminus of E6. This could explain why E6*I can

antagonise full length E6 function if there is a dominant negative effect of E6*I binding to full length E6. The same could be true of all isoforms as they contain the same portion of the N-terminus. This however does not explain any

transformation properties of the isoforms and at present these properties cannot be determined or mapped to the structures of the isoforms. E6*II has an

additional 22 amino acids added after the splice junction before meeting a stop