In Situ Amplification Using Universal Energy Transfer-labeled Primers

Materials and Methods

For HPV analysis by in situ PCR, CaSki (500 copies HPV-16 per cell) and SiHa (one copy of HPV-16 per cell) were grown to confluence and fixed for 1-3 days in 10% buffered formalin as previously described (Nuovo et al. 1995). Three different samples of 2000 cells each were spotted on silanecoated slides to allow direct comparison of different test variables. Also tested were formalin-fixed, paraffin-embedded tissues from four cervical low-grade squamous intraepithelial lesions (SILs), which were demonstrated to contain HPV-16 using individual HPV probes and varying stringency as previously described (Nuovo 1997). A negative control included HeLa cells (HPV-18-positive) and three cervical tissue specimens from women with no clinical or pathological evidence of HPV infection. For hepatitis C analysis, five paraffin-embedded liver biopsy specimens shown to be positive for hepatitis C RNA by RT in situ PCR were used. A negative control included two liver biopsy specimens that contained hepatitis E and G, respectively, and were negative for hepatitis C by RT in situ PCR.

The samples were digested in 2 mg/ml pepsin for 30-90 min. The protocol used for detection of HPV-16 DNA by in situ amplification has been described (Nuovo et al. 1991). Briefly, the amplifying solution contained 1 × PCR buffer (Perkin Elmer; Norwalk, CT), 4.5 mM MgCl₂, and 200 μM dNTP. In the reactions with one-tailed target-specific primer, 0.05 μM tailed primer, 0. 5 μM untailed target specific primer, and 0.5 μM UniPrimer were used. In the reactions in which both target-specific primers had tails, their concentration was 0.05 μM, and the concentration of the UniPrimer was 0.5 μM. The sequences of the primers used in this study are provided in Table 1. For comparative purposes, the same sequence of HPV-16 primers without the AmpliFluor-specific tails were synthesized with one molecule of either digoxigenin or biotin per oligomer. In addition, another set of HPV-16 primers was generated with three biotin molecules per oligomer, but these primers were 40-mers (Genosys Biotechnology; The Woodlands, Texas), which was needed to be able to incorporate the three biotin moieties. The concentration of the biotin- or digoxigenin-labeled primers was 1 μM. After 35 cycles, the slides were washed for 5 min in 0.2 × SSC at 60C, then counterstained in nuclear fast red. For comparison purposes, PCR in situ hybridization was also done on the samples using unlabeled primers and the biotinlabeled full-length HPV-16 probe as previously described (Nuovo et al. 1991).

OPEN IN VIEWER

Table 1

Sequences of the primers used in this study

Target	Primer sequence a
Prostate-specific antigen	AS-CCC TCA GAA GGT GAC CAA GTT CAT
	S-GGT GTA CAG GGA AGG CCT TTC GGG AC
	AS-tailed CCT GCA GGC TGA GGT GAA GGT GAC CAA GTT CAT
Chlamydia	AS-GTA CTA GAG GAC TTA CCT CTT CCC
	S-tailed CCT GCA GGC TGA GGT CTG TAA CAA CAA GTC AGG TT
Human P16 gene	AS-CAG AGG GTG GGG CGG ACC GC
	S-tailed CCT GCA GGC TGA GGT CCC GGG CCG CGG CCG TGG
HPV-16 b	AS- GGA CAC AGT GGC TTT TGA CAG TTA
	AS-tailed CCT GCA GGC TGA GGT GAC CCA CAG GAG CGA CC
	S-tailed CCT GCA GGC TGA GGT GGA CAC AGT GGC TTT TGA CAG TTA
HIV-1	AS-ATA ATC CAC CTA TCC CAG TAG GAG AAT
	S-TTT GGT CCT TGT CTT ATG TCC AGA ATG C
	S-tailed CCT GCA GGC TGA GGT CCT ATC CCA GTA GGA GAA AT
Hepatitis C c	AS-tailed CCT GCA GGC TGA GGT CCCAACACTACTCGGCTA (JH51)
	S-tailed CCT GCA GGC TGA GGT AGTCTTGCGGCCGCAGCGCCAAATC (JH52)
	S-tailed CCT GCA GGC TGA GGT TCCGCGGCCGCACTCCACCATGAATCACTCCCC (JH93)

a

AS, anti-sense; S, sense.

b

HPV primers correspond to the E6 ORF (Nuovo et al. 1991).

c

Hepatitis C primers per Han et al. (1991).

Our protocol for the detection of hepatitis C RNA using RT in situ PCR has been described (Nuovo et al. 1993a). Briefly, after pepsin digestion for 30 min, the amplifying solution containing the enzyme rTth, 0.05 μM of each of the tailed primers JH51, JH52, and JH92, plus 0.5 μM UniPrimer (Table 1) was placed over the tissue sections. After incubation at 60C for 30 min, 20 cycles of 94C and 60C were used, followed by the high-stringency wash.

Solution-phase PCR was performed at four different settings using (a) two conventional linear primers, (b) one conventional linear primer and energy transfer-labeled hairpin primer with the 3′ sequence complementary to the target, (c) one conventional linear primer, one tailed primer, and a UniPrimer, or (d) two tailed primers and a UniPrimer. Amplification was performed in 20 μl as described previously (Nazarenko et al. 1997) with 0.5 μM conventional primers and UniPrimer, 0.05-0.1 μM tailed primers, and 1 U of Taq DNA polymerase. Thermal cycling was performed with 5-min denaturation at 94C, followed by 35-40 cycles: 20 sec at 95C, 30 sec at 55C, and 1 min at 72C, and was completed with a final 5-min extension at 72C.

A Shimadzu RF-5000 spectrofluorophotometer and Victor fluorescent platereader were used to measure the fluorescence. For the spectrofluorophotometer, 2.5-5 μl of the reaction mixture was diluted to 600 μl with 20 mM Tris-HCl, pH 8.5, 50 mM NaCl, 2 mM MgCl₂, and emission at 516 nm with 490-nm excitation was detected in the 10 × 3 cuvette. For the platereader, the detection was performed directly in closed 0.2-ml PCR tubes.

Results

In Situ PCR with the UniPrimer System

To validate the specificity of in situ PCR using the HPV-16-specific UniPrimer system, the Caski (500 copies HPV-16/cell) and HeLa (HPV-18-positive) cells were examined. HPV-16 DNA was readily detected in the CaSki cells by in situ PCR using direct incorporation of the UniPrimer with the HPV-specific primers. No signal was evident in the HeLa cells. Equivalent results were obtained using one- or two-tailed target-specific primers (Table 1) or with primers that contained one biotin or digoxigenin molecule per oligomer, reflecting the high target copy number in these cells (data not shown).

We next studied the conditions that were needed to detect low-copy targets using in situ PCR and labeled primers. SiHa (1 copy HPV-16/cell) cells were used. No signal was evident when the primers labeled with either one biotin or one digoxigenin molecule were used. A signal was seen in 10% of the SiHa cells when the primers that contained three biotin moieties per oligomer were used. This compares to a detection rate of 100% if PCR in situ hybridization is used (Figure 2). A signal was evident in 10% of the SiHa cells when the UniPrimer system with one tailed primer was used (Figure 2). When both HPV-specific primers were tailed, the percentage of SiHa cells with detectable signal increased to 100% (Figure 2). The latter signal was lost when either the Taq polymerase or magnesium was omitted from the amplifying solution. Therefore, in situ PCR using two-tailed primers with the UniPrimer was as sensitive as PCR in situ hybridization using a labeled probe, being able to reliably detect one target copy per cell.

OPEN IN VIEWER

Figure 2

Detection of HPV-16 DNA and hepatitis C RNA using the UniPrimer system. (A) The HPV-16 data. A signal was seen in most cells after in situ PCR for the viral DNA if both HPV-16 primers were tailed and used with the consensus UniPrimer (1). The signal was lost if Taq polymerase was omitted or if HPV 18 specific primers were used (2). A strong signal was evident using a digoxigenin-labeled probe and PCR in situ hybridization (3; signal due to NBT/BCIP). In contrast, a weak signal was seen if primers labeled with three biotin molecules each were directly incorporated during in situ amplification and the hybridization step was omitted (4). Hepatitis C PCR-amplified cDNA was seen in the cytoplasm and perinuclear membrane with the UniPrimer system and tailed nested primers in this liver biopsy (B). No DNase step was needed. No signal was seen in the negative controls, which included a liver biopsy that was negative for hepatitis C and positive for hepatitis G (inset).

Next, four cervical SILs, each of which contained HPV-16 from low copy, detectable only after in situ hybridization, to high copy numbers were examined with the two tailed primers (i.e., each target-specific primer labeled with sequence whose complement could hybridize with the UniPrimer) and compared to the data generated on serial sections from PCR in situ hybridization (with a labeled probe) and in situ PCR using the primers with three biotin molecules per oligomer. A strong signal was evident in each of the four cervical tissues that localized to the dysplastic cells only, which confirms its specificity, and was equivalent when PCR in situ hybridization with a labeled probe was compared to in situ PCR with the two-tailed UniPrimer system. In each case, the number of positive cells was greater compared to standard in situ hybridization. However, the number of positive cells decreased to below that noted for standard in situ hybridization when in situ PCR was done using the HPV-16-specific primers that contained three biotin molecules per primer (data not shown).

Next, we studied the utility of the UniPrimer system for RT in situ PCR. The comparison was made between RT in situ PCR after DNase digestion and direct incorporation of the labeled nucleotide vs RT in situ PCR with no DNase digestion and in which the equivalent primers were modified for use with the UniPrimer system. Five liver biopsies that contained 5-50 or more infected hepatocytes were studied, as well as two liver tissues negative for hepatitis C (and positive for hepatitis E and G, respectively). In each of the five cases, an equivalent number of hepatocytes were positive using either RT in situ PCR with DNase digestion and digoxigenin dUTP or RT in situ PCR without DNase digestion and the UniPrimer system (Figure 2). With respect to the UniPrimer system, no signal was evident if the primers were omitted or in the two negative controls (Figure 2). A strong signal was evident with the digoxigenin-labeled nucleotide when the primers and DNase step were omitted, owing to DNA repair, as described previously (Nuovo et al. 1993a, 1994). Another advantage of the UniPrimer system was that protease digestion time did not have to be optimized for each tissue as is necessary with RT in situ PCR when one uses DNase digestion and a labeled nucleotide. With the UniPrimer system, 30 min of digestion with pepsin was adequate for each tissue compared to RT in situ PCR, in which DNase digestion was followed by direct incorporation of the labeled nucleotide, where optimal protease digestion was 30-90 min. As described previously, suboptimal protease digestion in this setting invariably leads to a false-positive signal due to the persistence of DNA repair (Nuovo et al. 1993a; Nuovo 1997).

Solution-phase PCR with the UniPrimer

First, the sensitivity of solution-phase PCR using standard primers was compared to that using the energy transfer-labeled primers with the HIV-1 gag region as the model system. Three different primer sets were tested: (a) two conventional linear primers, (b) one conventional linear primer and energy transfer-labeled hairpin primer with the 3′ sequence complementary to the target, and (c) one conventional linear primer, one tailed primer, and the UniPrimer. To compensate for the first cycles in which UniPrimer was not incorporated, three more cycles were performed in PCR with the UniPrimer compared to the other two systems. The results in Figure 3 for HIV-1 demonstrate that the UniPrimer-based system yielded the PCR product of the expected size and that the yield is comparable with that for the regular primers and the energy transferlabeled target-specific hairpin primers.

OPEN IN VIEWER

Figure 3

Detection of HIV-1 DNA. Lanes 1 and 2, regular PCR primers; Lanes 3 and 4, linear forward primer and HIV-1-specific AmpliFluor reverse primer; Lanes 5 and 6, regular forward primer, tailed reverse primer, and AmpliFluor UniPrimer. PCR was performed for 37 (Lanes 1-4) or 40 cycles (Lanes 5 and 6) with 10⁴ molecules of HIV-1 target (Lanes 1, 3, and 5) or without DNA target (Lanes 2, 4, and 6). Gel assays of the reactions before (A) and after (B) ethidium bromide staining and the measurement of fluorescent intensities (C) were done as described in Materials and Methods.

Next, we studied whether the use of two-tailed primers had any effect on the specificity of solution-phase PCR. In these analyses, both specific primers have the same tail that dictates the synthesis of the corresponding sequence which is complementary to the initial region of the UniPrimer, which permits the latter to incorporate into both ends of the PCR product. Therefore, both tailed target-specific primers are used in concentrations one tenth of that of the Uni-Primer, because they are necessary only during the first step of the reaction. It is possible that this may affect primer oligomerization, a common PCR pathway that can interfere with target-specific amplification (Nuovo et al. 1993b, 1994). It was indeed noted that the primer oligomerization band observed on the gel after 45 cycles of HPV-16 amplification using one tailed primer disappears when the two tailed primers are applied (data not shown).

Discussion

The main finding of this study was that one could detect one copy per cell with in situ PCR using two tailed primers and the corresponding Uni-Primer sequence. This improved sensitivity with in situ PCR may relate to decreased primer oligomerization, as demonstrated by solution-phase PCR with the corresponding Uni-Primer system. The use of labeled primers enables one to do target-specific incorporation for detection of DNA targets with hot-start in situ PCR without the need for a hybridization step. This is especially important with hybridization using an oligoprobe, because of the narrow window between signal and background and the concomitant potential for weakened signal and/or background (Nuovo 1997). The hot-start maneuver is needed to inhibit mispriming. With respect to RT in situ PCR, the advantages of using a labeled primer system compared to a labeled nucleotide are twofold. One avoids the need for overnight DNase digestion and the need to determine the optimal protease digestion time for each tissue. These are important advantages for paraffin-embedded tissues as they relate to the issue of residual DNA repair which, in our experience, is the most common cause of failure and false-positive results with RT in situ PCR (Nuovo et al. 1993b, 1994). These advantages make the in situ detection of DNA or RNA targets by PCR simpler and more rapid to do, with less potential for false-positive results. Moreover, the use of a fluorescein reporter system allows for simple co-labeling experiments with either a different fluorescent label, such as rhodamine or coumarin, or a chromagen such as NBT/BCIP, in which one could detect the two different labels by switching from darkfield to brightfield microscopy.

The foundation of the AmpliFluor UniPrimer system is the hairpin structure located on its 5′ end. Two moieties are attached to the stem sequence of the hairpin: a fluorophore (fluorescein) on the 5′ end, and a quencher (DABSYL) on the opposite side of the hairpin stem. DABSYL is a nonfluorescent chromophore whose absorption spectrum overlaps the emission spectrum of fluorescein. When stimulated by light of peak wavelength 488 nm, fluorescein emits fluorescence of peak wavelength 516 nm. However, when DABSYL is located sufficiently close to the donor fluorophore (less than 100 Angstroms), the energy from the excited fluorophore can be transferred to DABSYL and dissipated as heat. The hairpin structure brings the fluorophore and quencher into close proximity when the primer is free in solution, providing efficient quenching of the donor fluorophore. In the early steps of amplification, the extension of the labeled target-specific primers 1 and 2 will yield an amplicon for which the 3′ end of each strand is complementary to a corresponding region of the UniPrimer. During the amplification reaction, the hairpin structure of the UniPrimer is unfolded and copied and the fluorescein and DABSYL are no longer close enough to permit quenching. Instead, a fluorescent signal is emitted. To minimize the competition between the tailed primers and the UniPrimer, the former are used in a much lower concentration. As a result, the majority of the PCR product has the incorporated hairpin primer and generates the fluorescent signal.

An amplification detection system based on incorporation of energy transfer-labeled primers is an efficient way to eliminate carryover contamination and to simplify the high throughput detection procedure (Nazarenko et al. 1997). The high signal-to-noise ratio provided by the conformational transition of the labeled primers on their incorporation into the PCR product permits the quantitative results not only in real time but also in endpoint format. Modification of the energy transfer system presented here utilizes the universal primer labeled with energy transfer moieties, providing additional flexibility to the method mainly by decreasing the cost of synthesis of fluorescent primers: the same batch of the UniPrimer can be used in combination with several sequence-specific primers. Although a fluorescent platereader is the most efficient way to detect the fluorescent signal with solution-phase PCR, the product can also be visualized by placing the tube directly against a UV transilluminator image analysis system and photography with a mounted camera using a D540/40 filter.

For any PCR-based system, the absence of side reactions that may generate a background signal is crucial. One of the common PCR artifacts is primer oligomerization (Nuovo et al. 1994), which can create background in the UniPrimer system if the 3′ end of any oligonucleotide hybridizes to the single-stranded tail of the labeled hairpin primer by unfolding of the hairpin secondary to extension of the oligonucleotide. The traditional solution to the problem would be another design of the sequence-specific primer which will eliminate the complementarity with the UniPrimer, or using the hot-start maneuver (Nuovo et al. 1991; Nazarenko et al. 1997). Another and more general solution is doing PCR with two-tailed sequence-specific primers. When both sequence-specific primers have the same tail and are used at concentrations much lower than the concentration of the UniPrimer, no primer oligomerization-based background is generated. This result agrees with a recently published model (Brownie et al. 1997) which used a similar approach to eliminate primer-dimer problems. Another advantage of this approach, in which both target-specific primers are tailed, is a stronger signal, because both ends of the amplification product will be labeled. This was demonstrated by in situ PCR, in which the percentage of SiHa cells with detectable signal increased from 10 to 95% when the number of tailed primers was changed from one to two.

The UniPrimer-based methodology described here circumvents a major problem with the in situ PCR technique using paraffin-embedded tissues. Specifically, the heating to 65C for 4 hr that is obligatory in paraffin-embedded tissues induces DNA nicks (Nuovo et al. 1994). These nicks invariably lead to a false-positive signal when in situ PCR is done using a labeled nucleotide, which becomes incorporated in the DNA synthesized by Taq polymerase in a process analogous to nicktranslation (Nuovo 1997). The inability of digoxigeninor biotin-labeled primers to reliably detect one DNA copy per cell with in situ PCR probably reflects the relatively low specific activity of these moieties. Whether this problem can be circumvented by improvements in the detection system requires further study. At this stage, the UniPrimer-conjugated fluorescein-based signal with two tailed primers permits one target copy detection, either DNA or RNA, with low background.