The use of polymerase chain reaction generated nucleotide sequences as probes for hybridization

Molecular and Cellular Probes (1990) 4, 485-495

The use of polymerase chain reaction generated nucleotide sequences as probes for hybridization Sean P. Scully * M. E . Joyce, N . Abidi and M . E . Bolander Orthopedic Research Unit, NIAMS NIH, Bethesda, MD 20892, USA (Received 17 April 1990, Accepted 13 May 1990)

In this report a rapid, simple and economical means of preparing a cDNA probe of specified length, sequence and specific activity is described . The process involves the use of a polymerase chain reaction to incorporate radiolabelled nucleotides into a single stranded or double stranded cDNA sequence . A pair of oligonucleotide primers are synthesized, flanking the sense and antisense strands of a selected sequence . The primers are then used with a cloned DNA fragment or a cellular source of RNA or DNA as a template to amplify the specific gene sequence . The sequence to be used as a probe is selected from the known sequence using free energy calculations of the secondary structure . The calculation of free energy predicts regions of stable secondary structure which may hinder transcription and thus are to be avoided . By selecting the distance between primers the probe length can be controlled to allow adequate probe permeability into tissue samples . The specific region of the gene sequence can be chosen to differentiate between closely related sequences by avoiding areas of homology. Altering the concentration of a radiolabelled nucleotide allows direct control of probe specific activity . The use of asymmetric PCR allows the preferential generation of an antisense single stranded cDNA sequence for a higher sensitivity in the detection of low abundance mRNA . This report highlights the advantage of this technique in generating probes for in situ hybridization . However, any technique that relies on homology for detection of sequences, such as Northern and Southern blotting could also utilize this technique .

KEYWORDS: cDNA probes, polymerase chain reaction, in situ hybridization .

INTRODUCTION The technique of in situ hybridization utilizes labelled cDNA or RNA probes to localize gene expression to individual cells in fixed tissue samples . This technique has been useful in identifying the cells that are actively expressing structural and regulatory genes in complex tissues such as a growth plate and embryo . The success of in situ hybridization is highly dependent on the choice of an appropriate * Current address : Box 3000, Duke University Medical Center, Durham, NC 27710, USA 0890-8508/90/060485+11 $03 .00/0

© 1990 Academic Press Limited 485

486

S. P. Scully et al.

radioactive probe and detection system . Current techniques utilize oligonucleotide, RNA, or cDNA probes to detect the expression of gene sequences intracellularly .' Early procedures for synthesis of radiolabelled probes utilized in vivo synthesis with incorporation of radiolabelled metabolic precursors . Later with the advent of molecular biology the techniques of nick translation and random primer extension were used to synthesize probes from cloned DNA fragments . These probes had increased specific activities but were fragmented and heterogeneous in nature . More recently, in vitro synthesis has yielded RNA probes with high specific activity but that are contaminated with varying amounts of vector associated sequence . This method requires a cloned DNA fragment in an appropriate expression vector . Although these improvements in generating labelled probes have greatly enhanced the techniques that rely on hybridization, several critical probe characteristics are still not under direct control . One such characteristic is the length of the probe, thought to be critical for tissue permeability in in situ hybridization . A second characteristic not under direct control is the specific region of the gene sequence to be used as a probe . The choice of an appropriate sequence becomes important when differentiating between closely related sequences, or conversely when taking advantage of homology between species . The commercial introduction of polymerase chain reaction (PCR) allows further improvement in generating cDNA probes with control over the probe length, region of gene sequence, specific activity and single or double stranded character . PCR is used to amplify a segment of DNA that lies between two regions of known sequence . Oligonucleotides are used as primers for a series of reactions catalysed by the Taq polymerase . The reaction cycles through denaturation, annealing of primers and DNA synthesis . Because the reaction is cycled repeatedly, one round of amplification serves as a template for the next amplification resulting in a logarithmic amplification . This procedure can be completed in several hours and at a decreased cost compared to other methods of probe preparation .

MATERIALS AND METHODS Free energy calculations The sequence of the rat prepro-al(II) cartilage collagen mRNA (type II collagen) was retrieved from Genebank (accession number K02804) . Using the FOLD program the free energy of the RNA secondary structure was calculated according to the method of Zucker . 2 The data generated by FOLD was then plotted using the

companion program SQUIGGLES .' The software package was supplied by the Genetics Computer Group at the University of Wisconsin Biotechnology Center and was executed on a VAX computer system .

RNA/DNA purification A crude preparation of RNA was extracted from cartilage homogenate using a guanidinium thiocyanate/chloroform/phenol method as described .' Purified RNA was isolated from both rat fracture callus and liver using caesium chloride density

PCR generated nucleotide sequences

487

gradient centrifugation .' A purified DNA fraction was isolated from the caesium chloride fraction after RNA purification by phenol/chloroform extraction . A pUC 9 plasmid containing the cloned amino terminal portion of the pro-001) chain of cartilage collagen was a gift of Dr Y . Yamada.'

Synthesis of oligonucleotide primers Oligonucleotides were synthesized by an automated Applied Biosystems DNA synthesizer model 380 (Foster City, CA) using a phosphoramidite method .' The synthesized oligonucleotides were used directly as primers without further purification .

Reverse transcription and polymerase chain reaction Purified total cellular RNA from rat fracture callus or the crude preparation of RNA extracted from cartilage was reverse transcribed using a previously published procedure.' Briefly, a mixture containing dNTPs (10 mm), oligonucleotide primer (2 . 5 11M), 100 U µl - ' MMLV reverse transcriptase (BRL, Gaithersburg, MD) and 5 gg RNA was incubated for 1 h at 37°C . These reaction products, or the recombinant plasmid containing the type 11 collagen cDNA sequence, were then amplified with PCR . The reaction mixture contained 125 gM deoxynucleotide-5'-triphosphate, dATP, dGTP, dTTP and dCTP ; 2 . 5 lam oligonucleotide primers; 2. 5 U Taq polymerase and buffer (Perkin Elmer Cetus, Norwalk, Conn .), with either 400 pg pUC 9 plasmid, reverse transcription products or 2 . 6 lag fracture callus DNA . The reaction mixture was then cycled through annealing (50 ° C X 150 s), extension (72 ° C X 210 s) and denaturation (94°C X 90 s) steps, in a Coy Model 50 thermocycler (Ann Arbor, MI) for 30 cycles . Radiolabelled probes were generated by including 5 gCi of 35 S-dCTP (1300 Ci mmol -1 , Amersham, Arlington Heights, IL) with varying concentrations of unlabelled dCTP, 10 nM-125 gm. For reactions involving asymmetric amplification, the downstream primer was included at 2 . 5 gM while the upstream primer was added after serial dilution in TE buffer . Polymerase chain reaction amplification was carried out as above . The amplification products were evaluated by electrophoresis on 10% polyacrylamide gels . The incorporation of radiolabel into the amplified cDNA probe was performed by adding a constant amount of radiolabelled CTP to a varying concentration of unlabelled CTP . To remove unincorporated nucleotides after PCR amplification, the reaction mixture was diluted to 1 ml with TE and placed in a Centricon-100 filter (Amicon, Danvers, MA) . The filter was centrifuged at 3200 X g for 30 min . The retentate was diluted to 1 ml and recentrifuged . The resultant retentate, approximately 40 gl, was then frozen at -80 ° C prior to use . The purified reaction products were evaluated by gel electrophoresis and autoradiography . Specific activity values were calculated from n-scintillation counting of the purified reaction products and DNA concentration as determined by optical density at 260 nm .

488

S. P. Scully et al.

Dot-blot hybridization Purified total cellular RNA extracted from fracture callus and liver was blotted on nylon membrane (Hybond-N, Amersham, Arlington Heights, IL) at concentrations of 4 to 404 gg per 100 gl using a previously published procedure .' Hybridization of denatured radiolabelled probe, 6 X 10$ cpm gg - ' was carried out overnight at 45°C . The membrane was washed and autoradiography was then performed to analyse the hybridizaton .

In situ hybridization Newborn rat forequarters were fixed in 4% neutral buffered formalin for 24 h at 4°C . The samples were then paraffin embedded and 5 gm sections were placed on aminosilane treated glass slides (Digene, College Park, MD) . In situ hybridization was performed with the DNA Hybridization Kit (Oncor, Gaithersburg, MD) with a modified protocol . Briefly, paraffin was removed from the sections with xylene and ethanol . Sections were then treated with protease K (Sigma, St Louis, MO), 10 gg ml - ' for 15 min at room temperature . Sections were next treated with acetic anhydride and Tris-glycine buffers . The tissue was then post fixed in 4% paraformaldehyde for 5 min at room temperature . Twenty microlitres of radiolabelled cDNA (6 x 10$ cpm gg - ') was incubated with 180 iii of hybridization mix II (Oncor) at 90°C for 10 min . The probe mixture was then cooled to 45°C and applied to the prepared sections . The hybridization mixture was then covered with a coverslip and sealed with rubber cement . Hybridization was performed at 45°C overnight . The coversliips were next removed, and the sections washed with 50% formamide/2 X SSC at 45°C . The sections were then dehydrated in ethanol and dried . The dried slides were dipped in Kodak NTB2 emulsion (1 :1 DDW) and exposed at 4°C for 10 days . The emulsion was developed with Kodak D-19 for 3 min at room temperature and fixed with 30% sodium thiosulphate for 5 min .

RESULTS The calculated free energy of the type II collagen mRNA sequence from residue 1 to 521 was - 203 . 8 Kcal mole - ' . z The predicted secondary structure was plotted using the companion program, SQUIGGLES (Fig . 1) . With this information primers were selected that would amplify a probe approximately 200 by in length, avoid sequences of significant secondary structure and avoid complementary sequences in the primers so as to avoid 'primer-dimers' .9 The primers that were chosen for the type II collagen probe were oligonucleotide residues (140-159) on the sense strand and (295-314) on the antisense strand . The oligonucleotide primers described above were then used in PCR reactions to amplify a cDNA fragment from a variety of cellular templates (Fig . 2), including cloned plasmid collagen II cDNA, genomic DNA, ssDNA transcribed from purified cellular RNA and crude RNA extract . The reaction products, displayed after polyacrylamide gel electrophoresis and ethidium bromide staining, demonstrate the amplification of a 174 by fragment in each case . Digestion of the amplified fragment

PCR generated nucleotide sequences

489

OrO\ (A

6,

o

O N

er

a

O â

w 01, 50

r) N O Fig. 1 . Secondary structure of type if collagen mRNA . The gene sequence of rat prepro-al(II) cartilage collagen (bases 1 to 521) was obtained from Genebank . The computer package FOLD supplied by the University of Wisconsin Computer group was used to calculate the conformation energy of this sequence . The companion program SQUIGGLES was then used to plot the secondary structures

with Hae III resulted in restriction fragments of the size predicted by sequence analysis . Sequencing of the reaction product by the Sanger dideoxy sequencing method proved amplification of the selected collagen II sequence (data not shown) . The inclusion of radiolabelled dCTP in the polymerase chain reaction resulted in the synthesis of a radiolabelled amplified product that can be used as a hybridization probe . The specific activity of the probe can be adjusted by decreasing the concentration of unlabelled dCTP . The specific activity and the amount of probe synthesized are plotted as a function of the log of the CTP concentration (Fig . 3) . These results indicate that there is a concentration range over which the specific activity can be increased to 8 x 108 cpm µg -1 without a significant adverse effect on the amount of probe generated . Dot-blots were performed to evaluate the sensitivity and specificity of the hybridization with this probe (Fig . 4) . This is important because there remains a possibility that the labelled amplification product would not be complementary to the target sequence or that the primer sequences would not be unique matches . Dot-blots of serially diluted fracture callus RNA, known to contain the type II collagen message, were probed with a cDNA probe synthesized by PCR amplification (specific activity 6 x 10 8 cpm .tg - ') . It can be seen that a specific signal is detected in fracture callus RNA at a dilution of 64-fold, which is approximately 0 . 125 µg RNA (Fig . 4a) . An equal amount of total liver RNA, known not to contain

S. P. Scully et al.

490

a

b

C

d

174bp-

Fig . 2. Amplification of the selected type II collagen sequence . Oligonucleotide primers were used to amplify a specific segment of the type II collagen sequence from either cloned cDNA, a ; purified total cellular RNA from fracture callus, b ; purified DNA extracted from fracture callus, c ; or crude RNA extract from cartilage, d . The amplification products were analysed by polyacrylamide gel electrophoresis and ethidium bromide staining .

(CTP) Fig . 3 . Probe specific activity . Amplification by PCR of the type II collagen sequence was performed using a constant amount of radioactive dCTP and a varying total dCTP concentration . Amplified cDNA was separated from unincorporated dNTP. The specific activities of the synthesized probes were calculated by a-scintillation counting and optical density at 260 nm . The amount of probe synthesized and the specific activity are plotted as a function of the log of the dCTP concentration .

PCR generated nucleotide sequences

a

b

e

d

491

e

Fig. 4. Probe specificity and sensitivity . The synthesized collagen II cDNA probe was analysed by hybridization to dot-blots of fracture callus and liver RNA . Lane a demonstrates serial two-fold dilutions of 4 µg of fracture callus RNA . In lane b, 4 µg of liver RNA is serially diluted two-fold in a similar fashion . In the remaining lanes, c-e, serial two-fold dilutions were performed of fracture callus RNA (4 µg) combined with increasing amounts of liver RNA (4 jig, 40 µg, 400 µg) . The autoradiogram was developed after 24 h exposure .

type II collagen mRNA, was serially diluted and probed in a similar fashion . There is no significant hybridization indicating the absence of both a type II collagen message and non-specific hybridization (Fig . 4b) . To further examine specificity, increasing amounts of liver RNA were combined with a constant amount of fracture callus RNA . The diminution of the type II collagen signal results from dilution of the type II collagen message with liver RNA and saturation of the binding capacity of the nylon membrane, further demonstrating that no non-specific hybridization occurs with the PCR generated probe . The use of a single stranded probe has the advantage of eliminating the competing non-productive hybridization that occurs between complementary strands in a dsDNA probe . An asymmetric PCR reaction can be used to generate a

S. P. Scully et al.

492

a

b

a

a

dsDNA -

ssDNA-

Fig . 5 . ssDNA probes synthesized by asymmetric PCR . Amplification of the type II collagen sequence was carried out with decreasing concentrations of the upstream primer . Lane a contains 2 gm, b : 0. 2 gm, c : 20 nM, d : 2 nM and e: 200pM . Radiolabelled dCTP, 5 jCi, was included in the reactions . The reaction products were analysed by polyacrylamide gel electrophoresis and subsequent autoradiography .

single stranded probe, limiting the concentration- of the sense primer results in a preferential synthesis of an antisense ssDNA strand (Fig . 5) . The dependence of ssDNA synthesis on probe concentration was shown by amplification reactions using serial dilutions of the sense primer . The autoradiogram of the polyacrylamide gel demonstrates a diminishing band at the position corresponding to a dsDNA type II collagen fragment and an increasing band at the position of the ssDNA fragment . A section of newborn rat humerus was prepared as described in Materials and Methods . In situ hybridization was performed using a PCR generated dsDNA probe for type II collagen (Fig . 6). The dsDNA probe was labelled with 35S-dCTP resulting in a specific activity of 6 X 108 cpm µg - ' . The regions of the articular cartilage and growth plate show marked hybridization of the probe to proliferating chondrocytes . Consistent with other studies decreased hybridization is seen as cells mature into hypertrophic chondrocytes .

DISCUSSION

The standard approaches to the generation of radiolabelled probes for detection of sequence homology in techniques such as Southern, Northern and in situ hybridizations are both expensive and labour intensive . The process has involved cloning the gene of interest, with subsequent radiolabelling . Using these procedures, neither the specific region of homology nor the length of the probe are easily controlled . In this report we have described the use of the polymerase chain reaction as a means of generating a radiolabelled cDNA probe that requires less time than other tech-

PCR generated nucleotide sequences

493

Fig. 6. In situ hybridization . A cDNA probe for type II collagen was produced by PCR . This probe was used to localize regions of collagen II gene expression in the newborn rat humerus by in situ hybridization . The photographic image is originally a 1000 x magnification of the developed emulsion using differential interference contrast .

niques and allows direct control over critical probe parameters . The only requirement is a known gene or amino-acid sequence, the availability of oligonucleotide synthesis and a template sequence . The template can be obtained from cellular DNA, cloned cDNA, purified RNA or crude RNA extract . The ability of PCR to amplify a sequence from a relatively small amount of cellular material has been previously demonstrated . 10 We have demonstrated that with the appropriate choice of primers, specific probes for a desired gene can be reproducibly generated and used in subsequent blot hybridizaton and in situ hybridization . The dependence of hybridization on probe length has been demonstrated and thought to reflect tissue permeability ." It has been shown that shorter probe length gave higher hybridization signals . This is more pronounced when the tissue was fixed in cross-linking reagents such as glutaraldehyde . With shorter probe length, however, the limiting factor becomes the lowered melting temperature and the ability of the hybridization to withstand the stringency conditions required to lower background non-specific binding . For these reasons it has been suggested that

494

S . P. Scully et al.

optimal probe length should be approximately 200 bp . The ability to directly control the length of PCR generated probes is particularly attractive in this regard . Differential detection of mRNA that have a high degree of homology is difficult . Using methods presented in this report, the probe sequence can be directly controlled by primer selection to avoid regions of homology . For example, closely related isoforms of a growth factor family could be differentiated by taking advantage of regions of sequence divergence . The ability to directly control probe sequence could likewise be used as an advantage when a sequence is known in one species and a probe is desired for a second species . Using this approach oligonucleotide primers that have sequence homology could be used to synthesize an appropriate probe from the second species . The ability to detect low abundance mRNA is critically dependent upon probe specific activity . The specific activity of a PCR generated probe depends solely upon the specific activity of the reaction mixture ." Reaction mixtures for radiolabelling typically include three unlabelled nucleotides at concentrations greater than the K m and one isotopically labelled nucleotide slightly below the K m . This results in an increased proportion of radiolabel incorporated although at a submaximal reaction rate ." The limiting factor in this approach is a diminution in amplification as the dNTP falls below approximately 1 µM . Radiolabelled probes of high specific activity are rapidly degraded by radiochemical decay . The ability to program the therrnocycler to synthesize the probe overnight for use the subsequent morning, results in minimal probe degradation . In this report we have performed in situ hybridization with a dsDNA probe for the relatively abundant collagen II mRNA . These double stranded cDNA probes have the disadvantage of hybridization between complementary strands for each other rather than the intracellular mRNA . The ability to generate an antisense ssDNA probe avoids this competition, strengthening the hybridization signal . An additional advantage of ssDNA probes in addition to heightened signal, is the ability to generate a sense strand ssDNA probe that can be used to evaluate non-specific binding . This non-specific control would have all the same properties of the antisense probe except homology to the mRNA of interest . Single strand DNA synthesis can be accomplished by asymmetric PCR in which the reaction conditions can be manipulated to promote the preferential synthesis of ssDNA . 13 In the current report this is optimal at a ratio of 10 2_103 :1 in agreement with previous reports ." During the initial cycles dsDNA is preferentially synthesized . However, when the limiting primer has been consumed ssDNA will be produced . Initially dsDNA will be amplified in an exponential fashion, and during the latter phase ssDNA will be synthesized at a linear rate . The ssDNA and dsDNA can be subsequently separated by gel electrophoresis or anion exchange chromatography ." The results of the in situ hybridization featured in this report are consistent with earlier reports using nick translated cDNA clones . 15 The use of PCR technology simplifies the tedious and technically demanding procedure of probe preparation preceding all hybridizations . The ability to directly control specific activity, probe length, specific gene sequence and the generation of ssDNA probes are distinct advantages . Furthermore, the convenience of the rapid synthesis permits the use of a probe with minimal radiochemical decay .

PCR generated nucleotide sequences

495

ACKNOWLEDGEMENTS The authors would like to thank Dr L . Love for careful reading of this manuscript and introduction to the technique of in situ hybridization. Ms A . Heydeman is gratefully acknowledged for her expert technical assistance. The study was funded in part by grants from the American Academy of Orthopaedic Surgery and the Orthopaedic Research and Education Foundation .

REFERENCES 1 . Tecott, L . H ., Eberwine, J . H ., Barchas, J. D . & Valentino, K . L . (1987). Methodological considerations in the utilization of in situ hybridization . In In Situ Hybridization . New York : Oxford University Press . 2 . Zucker, M. & Stiegler, P . (1981) . Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information. Nucleic Acid Research 9 (1), 133-48 . 3 . Pallansch, L ., Beswick, H ., Talian, T . & Zelenka, P . (1990). Use of an RNA folding algorithm to choose regions for amplification of the polymerase chain reaction . Analytical Biochemistry (in press) . 4 . Chomczynski, P . & Sacchi, N . (1987) . A single step method of RNA isolation by acid guanidiniumphenol-chloroform extraction . Analytical Biochemistry 162, 156-9 . 5 . Nemeth, G . G., Heydemann, A . & Bolander, M. E . (1989) . Isolation and analysis of ribonucleic acids from skeletal tissues . Analytical Biochemistry 183, 301-4 . 6 . Kohono, K ., Martin, G . & Yamada, Y . (1984). Isolation and characterization of a cDNA clone for the aminoterminal portion of the Pro-al(II) chain of cartilage collagen . journal of Biological Chemistry 259,13668-73 . 7 . Beaucage, S . L. & Caruthers, M . H . (1981) . Deoxynucleoside phosphoramites-a new class of key intermediates of deoxypolynucleotide synthesis . Tetrahedron Letters 22, 1859-62 . 8 . Boden, S . D ., Joyce, M. E ., Oliver, B . Heydemann, A. & Bolander, M . E. (1989) . Estrogen receptor mRNA expression in callus during fracture healing in the rat . Calcified Tissue International 45(5), 324-5 . 9 . Saiki, R. K . (1989) . The design and optimization of the PCR . In PCR Technology (Erlich, H ., ed .). Stockton Press . 10 . Rappolee, D . A ., Mark, D ., Banda, M . J . & Werb, Z . (1988). Wound macrophages express TGF-a and other growth factors in vivo: analysis of mRNA phenotyping . Science 241, 708-12 . 11 . Moench, T. R ., Gendelman, H . E ., Clements, J . E ., Narayan, O . & Griffin, D . E . (1985). Efficiency of in situ hybridization as a function of probe size and fixation technique . Journal of Virological Methods 11,119-30 . 12 . Melton, D . A ., Krieg, P . A,, Rebaggliati, M . R., Maniatis, T., Zinn, K . & Green, M . R . (1984) . Efficient in vitro synthesis of biologically active RNA and RNA hybridization in probes from plasmids containing a bacteriophage . Nucleic Acids Research 12, 7035-56. 13 . Gyllensten, U . B . & Erlich, H . A . (1988) . Generation of single stranded DNA by the polymerase chain reaction and its application to direct sequencing of the HLA-DQA locus . Proceedings of the National Academy of Sciences (USA) 85, 7622-56 . 14 . Sambrook, J ., Fritsch, E . F . & Maniatis, T. (1989) . In Molecular Cloning, A Laboratory Manual, 2nd edn . New York: Cold Spring Harbor Laboratory Press, Cold Spring Harbor . 15 . Sanberg, M . & Vuorio, E. (1987) . Localization of types I, II, and III collagen mRNAs in development human skeletal tissues by in situ hybridization . journal of Cell Biology 104, 1077-84 .