Dieses PDF/A-Dokument wurde maschinell aus der approbierten Originalversion erzeugt. Die Originalversion finden Sie an der Universitätsbibliothek der Veterinärmedizinischen Universität, Wien

Establishment and Validation of Real-Time PCR Assays for the Quantification of Different DNAForms of Feline Immunodeficiency Virus

Master Thesis submitted to attain the degree of Master of Science MSc.

Submitted by Matthias Hofer Bal90°C, (2) primer annealing at 50°C to 75°C, (3) primer extension at 72°C 78°C. A typical PCR-reaction consists of 30 - 50 cycles (Powledge et al., 2004).

The disadvantage of PCR often lies in the post-PCR handling steps that are laborious and prone to cross-contamination. Post-PCR steps include the agarose-gel electrophoresis with ethidium bromide detection. Southern blot or PCR-ELISA, respectively. Real-time PCR eliminates post-PCR steps by simultaneous amplification and detection without opening the tube and therefore is referred to be a closed or homogenous system. Real-time PCR is more precise than conventional PCR, minimizes the possibility of cross-contamination between samples due to the closed system and enables fast and continuous data collection. However, amplicon size determination still requires the opening of the tubes. Furthermore, higher startup costs and expenses are needed compared to conventional PCR (outlined in Mackay et al., 2002 and Houghton et al., 2006).

30

Introduction

1.3.1. Amplicon detection During real-time PCR, amplification is monitored by recording changes in samples fluorescence. Both, specific and unspecific detection methods are used. All establish a link between amplification and increase in fluorescence (see Figure 1.7).

1.3.1.1.

DNA-binding fluorophores

This is the earliest and simplest mechanism, but still in use. DNA-binding fluorogenic molecules including Ethidium bromide, BEBO, YOYO-1 and SYBR-green can intercalate into double stranded DNA and then fluoresce under exposure to light with a suitable wavelength. SYBR green is the most commonly used intercalator. The bound SYBR green exhibits 1000fold more fluorescence than free dye and the binding affinity is 100-times higher than that of ethidium bromide. In general, intercalation happens independent from the DNA sequence, thus this type of detection is unspecific (Figure 1.7a). That is why unwanted amplification products like primer-dimers are also detected. Therefore, melting curve analysis of the amplicon should be considered. If it shows two or more peaks, there is evidence that not only one amplicon was generated. However, DNA-binding fluorophores have the advantage that no specific probe has to be designed. That reduces time and costs. Furthermore, amplicon size is irrelevant for detection (Wilhelm and Pingoud, 2003).

/. 3.1.2.

Hybridisation probes: ligtit cycler probes

This method uses fluorescence resonance energy transfer (FRET) for amplicon detection. FRET is the transfer of excitation energy from one fluorophore to the other - from dipole to dipole - that share overlapping emission and excitation spectra. Two fluorophore-labelled probes are used in one assay. The donor probe has a 3' label, while the acceptor probe is labelled on the 5' end. In case that both probes are not bound, only the donor-fluorophore is exited. During the annealing step of the real-time PCR, both probes specifically bind onto the template in close proximity. As a result, the 5' donor fluorophore transfers its energy on the 5' acceptor fluorophore (Figure 1.7b). The emerging fluorescence of the acceptor is detected and correlates with amplification. A disadvantage can be the use of Taq-polymerase that can partly hydrolyze the probes through its endonucleolytic activity. The resulting higher signal-to-noise ratio can be prevented by the use of other polymerases (Wilhelm and Pingoud, 2003).

31

Introduction

/. 3.1.3.

Linear oligoprobes: 5' nuclease probes/ hydrolysis probes/ TaqMan probes

This method to detect amplification was described first in 1991 for the use of radiolabelled probes (Holland et al., 1991). In 1993, the method was improved by the use of a dualfluorophore labelled probe (Lee et al., 1993). This amplicon-detection method uses the 5' - 3' exonuclease function of the Taq-polymerase and a single sequence specific oligonucleotide probe, which binds to the template between the primers, before phmer-annealing. The probes are labelled with a quencher fluorescent dye and a reporter fluorescent dye, respectively. When both fluorophores are bound on the probe, they are in close proximity and the quencher subsequently absorbs the reporter's energy by FRET. During amplification, the probe is cleaved by the Taq polymerase and the reporter is separated from the quencher, resulting in liberated reporter fluorescence (Figure 1.7c). This fluorescence is measured and is proportional to amplification. 5' nuclease probes should be 20-40 nucleotides in length, should have a GC content of 40-60%, should not include single nucleotide runs and should not have repeated sequence motifs or overlapping regions with the primers. Common quenchers include fluorescing and non-fluorescing quenchers (NFQ). Additionally, all obtained results can be normalized to the passive internal reference fluorophore ROX, in order to adjust for non-PCR related fluctuations in fluorescence (Mackay et al., 2002). In conclusion, the advantage of TaqMan probes is the specificity of amplicon detection. However, TaqMan probes are more expensive and probe design can be time consuming and challenging (Wilhelm and Pingoud, 2003).

/. 3.1.4.

tiairpin oligoprobes: molecular beacon probes

Molecular beacon probes are a variation of dual labelled oligoprobes. They also have a reporter and a quencher fluorophore at the probe's ends that can inhibit reporter fluorescence due to FRET. The probe is designed in such a manner, that only a part in the middle of the probe is homologous to the template, while the terminal 10 to 15 nucleotides are self-complementary. Thus, the free probe builds a stem-loop and facilitates FRET due to close proximity of the fluorophores (Figure 1.7d). In order to bind to the template, the probe must open the stem-loop structure and thus releases reporter fluorescence (Tyagi and Kramer, 1996).

/. 3.1.5.

Self fluorescing amplicon: sunrise primers

The sunrise primers work similar to molecular beacon probes, except that the fluorescent label is incorporated into the PCR product. At its 5' end, the dual-fluorophore labelled sunrise -32-

Introduction

primer forms a hairpin structure, which maintains close proximity of reporter and quencher, inducing FRET. The sunrise primer acts as forward primer and is elongated during real-time PCR. This single stranded DNA amplicon serves as template for the reverse primer. During elongation, the hairpin structure is opened and reporter fluorescence is released (Figure1.7e). The amplicon detection method is not highly specific, because also primer dimers lead to opening of the stem-loop (Nazarenko et al., 1997).

/. 3.1.6.

Self fluorescing amplicon: scorpion primers

The dual fluorophore labelled scorpion primer is structurally similar to the molecular beacon probe or the sunrise primer. Like the molecular beacon probe, but unlike the sunrise primer, the hairpin region is complementary to a part of the amplicon. The 3' end of the scorpion primer acts as fonward primer and a single stranded amplicon is elongated, which can then serve as template for the reverse primer. However, the stem loop region is complementary to a target sequence further downstream, subsequently binds to it and thus separates the fluorophores and induces the fluorescence signal of the reporter (Whitcombe et al., 1999) (Figure 1.7f).

33

Introduction

Figure 1.7. Different types of amplicon detection in a real-time PCR assay: (Wilhelm and Pingoud, 2003) (D) donor, (A) acceptor, (R) reporter, (Q) quencher (a) DNA binding fluorophores, (b) hybridization probes (c) 5' nuclease probes, (d) molecular beacon probes, (e) sunrise primers, (f) scorpion primers

1.3.2. Quantification Real-time PCR offers a wide dynamic range of quantification of 7-8 logarithmic decades with a high precision (

The

correlation

coefficient (r^) of the standard

I" u

curve, the slope of the standard

O 20

i e

axis. The mean and the standard

curve and the

15

real-time

PCR

efficiency (E) are listed in the box.

10 S'

1 E*02

1 E+03

log copy number

After deciding for the HeLa^ standard, the assay setup had to be tested and improved. There was the question, which primers to use in the first round PCR. Former tests (with the HeLa integration standard) showed that using a fon/vard and a reverse Alu primer - as described in the literature - resulted in a worse standard curve and an impaired detection of integrated DNA, compared to a setup with only one Alu primer (data not shown). Therefore, we tested 4 possible setups for the first round PCR of the nested real-time PCR assay for integrated DNA with the HeLa^ integration standard (Figure 3.7): a setup with both Alu primers (+Alu) (Figure 3.7a), a setup without Alu primers (-Alu) (Figure 3.7b), a setup only with the Alu forward -72

Results

primer (AluF) (Figure 3.7c) and a setup with only the Alu reverse primer (AluR) (Figure 3.7d). The AluR setup showed the best standard curve of all approaches with a correlation of 0.996 and a slope of -3.46. The efficiency of the real-time PCR reaction was 95 % (Figure 3.7d).

a) +Alu setup r^: 0.9808 slope: -4.011 E: 0.7755 i 30 E

Figure 1,5

3.7.

standard

Comparison

curves

of

of

l-ieLa^

standard in different setups of the integrated DNA real-time PCR assay: The standard curves of (a) 1.E+02

the assay with both AluF and

log copy number

AluR primer, (b) the assay with no

b) -Alu setup

Alu primer, (c) the assay with the AluF primer, (d) and the assay

r^: 0.9930 slope: -3.372 E: 0.9795

40

-___^^ 35 -

with

the

generated

AluR

primer

from the

are

(threshold

a

1i

cycle) CT values of the dilution '•

.25-

&

series of HeLa2 standard. The log copy number of the template input

2 20

is shown on the x-axis and the CT values on the y-axis. The mean

10

and the standard deviation of the

s

duplicates

are

indicated.

The

0

correlation coefficient (r^) of the

1 ,E+02

log copy number

standard curve, the slope of the

c) Alu-F setup

standard curve and the real-time PCR efficiency (E) are listed in

36 «30

F

—~~^~^*

•

a 25

the box

r^: 0.9298 slope: -3.642 E: 0.8818

40

"•——.

~—*-—-^ ^^~~"~~*'~~—.^

•D 20

^

'

S 16 ••

10'

s 0 1.E+02

log copy number

73

•"—-^-._^

Results

d) Alu-R setup •to ~

r2: 0.9960 slope: -3.461 E: 0.9450

40

35 w

^ 30

E

3 C « 25

• ••»

u >> u

s

"^~—-. «

•O 20 -

«

o

£ W £ 15 £ +rf

—

»

10 -

5 -

_

0

1.E+00

1.E+01

1 .E+02

1.E+03

1.E+04

log copy number Figure 3.7 continued: description In first part of ttie figure.

As described earlier, the principle of the assay additionally depends on the subtraction of the -Alu approach from one of the +Alu approaches in order to subtract background amplification from linearly amplified unintegrated DNA. This is crucial, because the forward primer in the first step of the nested integrated DMA assay can produce single stranded DNA from all viral DNA forms. These single stranded fragments are detected in the second step of the integrated DNA assay. This false positive detection is measured by a -Alu approach and consequently is subtracted from the +Alu approach in order to quantify only integrated DNA. Concluding, all the different +Alu test setups were also analyzed for usability concerning subtraction from the -Alu test setup. The +Alu setup showed higher CTs than the CTs of the AluR- and the Alu-F-setup respectively, which implied that the setup with both Alu primers detected less integrated DNA (Figure 3.8). Concluding, again the AluR setup leaded to the best result, because the subtraction showed a constant CT-differences of about 9 cycles, while the subtraction of the AluF or the +Alu setup led to more variable CT-differences (Figure 3.8).

Concluding, the AluR setup in combination with the HeLa^ standard seemed to be best suited to quantify integrated DNA. The assay shows a linear range of quantification over 3 logarithmic decades. Compared to the other assays, the range of quantification is smaller. 74

Results

which is explained by the use of an 1:4 dilution series of the standard compared to 1:10 dilution series by the other assays.

CT difference from -Alu setup -+Alu-«-AluF ---AluR

O •o

12

3

4

5

6

Standard dilutions Figure 3.8. Comparison of the different +Alu setups of the integrated DNA assay: CT-difference of standard dilutions: The +Alu setups (+Alu, AluF and AluR) and the -Alu setup of the real-time nested PCR for integrated DNA were performed as described in the text. The 6 standard dilutions of the HeLa2 standard resulted in different CTs depending on the setup. The CTs of each +Alu setup have to be subtracted from those of the -Alu setup as described in the text. The 6 standard dilutions are plotted on the x-axis. The subtracted ACT values (between the different +Alu setups and the -Alu setup) are plotted on the y-axis. The AluR setup shows the highest and most constant ACT, which makes this setup superior to the other tested setups.

3.1.3. 2-LTR circle assay The real-time PCR assay for quantification of 2-LTR circles uses primers and probe for amplification of the U5-U3 junction that is unique in 2-LTR circles (Figure 3.9). The assay has already been created at the institute (Figure 3.9.) (Steinrigl, unpublished). Thus, the probe as well as the primers had already been designed and tested for usability.

U3

'

R

'

U5

I

I

U3

I

R

h-

U5

f p r Figure 3.9. 2-LTR circle assay: The 2-LTR circle assay is shown with primers (f and r) and probe (p) on a 2-LTR circle fragment with both LTRs

75-

Results

The standard for this assay was a 1:10 dilution series of plasmid p2LTRsense that contains a FIV 2-LTR-junction. The standard was created and tested for usability with the primers and probe for this assay (Figure 3.10). The standard could be used for quantification, because it had a good correlation of 0.99 and a slope of -3.6. The standard shows a linear range of amplification over 7 logarithmic decades.

45 40

r2: 0.9903

35

slope: -3.596 E: 0.8971

E

30

3 C « 25 o

>• o o

20

K 15 10 5 -

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

log copy number Figure 3.10. Standard curve from 2-LTR circle standard for quantification of 2-LTR circles: The standard curve is generated from the (threshold cycle) CT values of the dilution series of p2LTRsense. The log copy number of the template input is shown on the x-axIs and the CT values on the y-axIs. The mean and the standard deviation of the duplicates are indicated. The correlation coefficient (r^) of the standard curve, the slope of the standard curve and the real-time PCR efficiency (E) are listed in the box.

3.1.4. 1-LTR circle assay 1-LTR circles have no unique sequences that could be used to place primers and probe. As a consequence, primers that bind in the area surrounding the LTR must be used: the reverse primer in the gag and the forward primer in the e/7i/(Figure 3.11). These primers specifically amplify circular episomal DNA, while the 1-LTR circles can be distinguished from the 2-LTR circles by amplicon size (Figure 3.11). First, the primers were designed and tested for usability before designing the complete assay including a probe. Furthermore, once the assay with the primer is established, there would also be the possibility to use SYBR green for real-time PCR detection. 76

Results

Figure 3.11. 1-LTR circle assay: Tlie two different primer pairs that were designed to detect 1-LTR circles are stiown together in this figure. Both primer pairs can amplify parts of 1-LTR circles - as shown in the figure - but can also amplify 2-LTR circles with a 355 bp longer amplicon as described in the text. f1 and r1 are fonward and reverse primer that bind adjacent to the LTR and produce an amplicon of 459 bp from a 1-LTR circle and an 814 bp amplicon from a 2-LTR circle. f2 and r2 are forward and reverse primers that produce a larger amplicon of 1510 bp from a 1-LTR circle and an 1865 bp amplicon from a 2-LTR circle.

A first primer pair was designed with primers binding just outside of the LTR (Figure 3.11: fl/rl): a 1-LTR circle amplicon would be 459 bp long, while a possible amplicon from a 2LTR circle would have a size of 814 bp. The primers were tested in a conventional PCR on the following templates: DNA extracts of CrFK cells infected with CT25ein, plasmids pCT25ein and pCT5efs and a negative control. The PCR on the infected DNA extracts showed the correct band at 459 bp. However, also the plasmid controls showed a band of the same size (Figure 3.13a). Theoretically, plasmids should lead to larger amplicons because they contain two LTRs and various plasmid-backbone elements like ampicilin resistance genes between both primer binding sites (Figure 3.13a). Thus, the next step was to identify the binding sites of the 1-LTR primers on one of the tested plasmids. Therefore, the plasmid pCT5efs was sequenced with both 1-LTR primers (F1/R1) and the resulting sequences were aligned with the sequence of pCTSefs. The result indicated that the primers bound to the expected sites: the reverse primer bound to the region between gag and CMV-LTR hybrid and elongated a fragment over the complete CMVLTR hybrid sequence that ended upstream of the CMV-LTR hybrid and downstream of the ampicilin resistance gene. The forward primer bound upstream the 3'LTR and elongated a fragment over the total 3'LTR that ended in the ColEI origin (Figure 3.12d: F/R). Thus, we concluded that primers annealed at the correct sites. For identification of the unexpected PCR-amplicons, obtained after using pCT5efs or pCT25ein as templates, the pCT5efs-amplicon was purified out of the agarose gel and sequenced with both 1-LTR primers. The results are shown in Figure 12e. The reverse primer produced an amplicon spanning over the complete LTR and ending upstream of the 3'LTR (Figure 3.12e: R2). The fonA^ard primer unfortunately produced only a 61 bp long fragment that represents a piece of the LTR (Figure 3.12e: F2). Upon repetition of sequencing, the fon^/ard primer again only yielded a 206 bp long fragment of the amplicon, -77-

Results

which again only aligns with the LTR. The reverse primer again sequenced an amplicon including the LTR and a part of the sequence upstream of the LTR (figure 3.12: F3/R3). Concluding, the results sufficiently showed that the PCR-amplicon, obtained from pCTSefs, contains one complete LTR, equal to the expected amplicon that should be obtained from a 1-LTR circle. However, generation of such an amplicon should not be possible, according to the maps of either pCTSefs or pCT5efsD66V. However, the sequencing results also showed that the primers bind correctly so that the reason for that unexpected amplicon is probably not due to misbinding of the primer.

a)

b) '•nv

3' LTR'

C0IEI origin Amplcillin res.

CMV-LTR

F primer

gag'

R primer

C)

3' LTR'

C0IEI origin'

F primer

d) n—I-

•nv

3' LTR' F primer

C0IEI origin Ampicillin res.

F

CMV-LTR R

gag'

R primer

e) ••nv

3' LTR'

C0IEI origin"

F primer F2

R2

"•nv

3" LTR' F primer

Col El origin'

F3

R3 Figure 3.12. Sequencing of plasmids and 1-LTR circle amplicons: (a) to (c) shows different views and fragments of the plasmid pCTSefs that was used for sequencing and alignment: (a) whole plasmid (b) LTR regions and backbone elements of pCTSefs (c) 3'LTR plus adjacent sequences (d) The plasmid was sequenced with the 1LTR circle primers: alignment between pCTSefs and the F primer amplicon (F) and the R primer amplicon (R).

78

Results (e) The PCR product of the 1-LTR circle assay was sequenced with the 1-LTR circle primers: alignment between pCTSefs and the F primer amplicon (F2) and the R primer amplicon (R2) (f) The PCR product of the 1-LTR circle assay was sequenced with the 1-LTR circle primers: alignment between pCTSefs and the F primer amplicon (F3) an the R primer amplicon (R3).

Since similar 1-LTR circle assays have been described in the literature, another attempt for unique 1-LTR quantification/detection was made by adopting the amplification conditions from reported work (Jacque and Stevenson, 2006). For that, a new primer-pair, amplifying a larger amplicon of 1510 bp, was designed (Figure 3.11: f2/r2). This time, primers were tested on infection extracts of HeLa cells infected with CTSefs and on different plasmid concentrations. The PCR on the FIV infected DNA extracts showed the correct band at 1510 bp. Unfortunately, the plasmid controls again showed a band of the same size. Consequently, conditions of the PCR reaction, in particular the annealing temperature, were altered for further testing. The result showed little influence of the annealing temperature on the PCR reaction, because pCTSefs leads to a 1510 bp amplicon at all annealing temperatures. Lower concentrations of pCT5efs lead to amplification only at 65°C (Figure 3.13b).

Figure 3.13. Conventional PCR of 1-LTR circle assay: (a) A 1-LTR circle PCR was performed with different templates and the first designed primer pair that produces a 459 bp amplicon from 1-LTR circles (indicated with an arrow). Templates: lane 1: infection CT25ein IN+ 8h p.i., lane 2: infection CT25ein IN+ 144h p.i., lane 3: plasmid pCT25ein, lane 4: plasmid pCTSefs, lane 4: negative control; (b) Four different 1-LTR circle PCR were performed with the

79-

Results second designed primer pair that produces a 1510 bp amplicon from 1-LTR circles (indicated with an arrow). Different annealing temperatures were used (shown in the boxes below the lanes). Templates: lane 1: infection CT5efs IN+ 8h p.i., lane 2: infection CTSefs IN+ 144h p.i., lane 3: negative control, lane 4: plasmid pCTSefs 1 ng lane 5: plasmid pCTSefs 1 pg;

We then argued that the observed amplicon might also be due to recombination in the bacteria used for plasmid propagation. In order to test this possibility a new primer pair was designed that amplifies a similar 1500 bp fragment from the e/7/gene. A SYBR green realtime PCR with a dilution series of the plasmid pCTSefs as template was performed including one approach with the newly designed e/7i 0.05). This supports the suggestion that most of total viral DNA is integrated at these time points. However, the 2-LTR circle amounts show a stable 2-LTR circle fraction of 5% among the total viral DNA at these time points.

The 2-LTR circle assay showed that 2-LTR circles peak 24 hours after infection and decrease until 144 hours in case of both viral vectors. The same kinetics were already obtained for HIV-1 (Butler et al., 2001; Butler et al., 2002). The decrease of 2-LTR circles after the peak at 24 hours, seen in our study as well as in HIV-1 studies, is still under discussion. It could be due to cell division; in this case, 2-LTR circles are diluted out, because they do not contain any replicative element (Butler et al., 2001; Butler et al., 2002; Pierson et al., 2002). Others argue that 2-LTR circles are labile and instable DNA forms (Sharkey et al., 2000; Sharkey et al., 2005). A long-time in vitro study showed stable 2-LTR circles and therefore even non-specific integration of 2-LTR circles was suggested (Brüssel and Sonigo, 2003). The reason for the 2-LTR circle decrease in this study cannot be revealed with certainty: it could be due to instability, or due to dilution since the cell number was steadily rising from 24 hours post infection onwards (data not shown). Further investigations, for example by using cell cycle arresting agents like aphidicolin, could give more concrete information about 2-LTR stability.

-100

Discussion

The 2-LTR circle assay showed no difference between the integration deficient and the integration proficient vectors, neither in l