ANALYTICALBIOCHEMISTRY
185,353-358
(1990)
Nonradioactive DNA Detection on Southern Blots by Enzymatically Triggered Chemiluminescence Denise Pollard-Knight,*91 Adrian C. Simmonds,* A. Paul Schaap,? Hashem Akhavan,? and Michael A. W. Brady* *Corporate Research, Amersham international plc, Pollards Wood Luboratories, White Lion Road, Amershum, Bucks, HP7 9LL, United Kingdom, and TDepartment of Chemistry, Wayne State University, Detroit, Michigan 48202
Received
August
24,1989
A chemiluminescent reaction based on the deprotection of a phosphorylated phenyl dioxetane by alkaline phosphatase has recently been described (Schaap, A. P., 1988, J. Biolumin. Chemilumin. 2, 253). Light output is enhanced by intermolecular energy transfer to a micelle-solubilized fluorophore. This system is applied here to the detection of DNA probes on Southern blots. Enzyme solution assays which give an indication of sensitivity show that using this substrate 100 fg (0.7 amol) alkaline phosphatase can be detected on a luminescence plate reader (200 ms reading time). In a model Southern blotting system 180 fg HindIII digested X DNA was detected on film with homologous biotinylated DNA and a streptavidin-alkaline phosphatase complex. The single copy genes mos and r&-l, representing targets of 4.2 and 2.4 pg target DNA respectively, have also been detected in Southern-blotted human genomic DNA. A delay in reaching a plateau level of light output which is dependent on pH is observed but signal continues for at least 7 days. Typically, 12-h exposures to X-ray film were performed but once a steady-state light output had been achieved this time could be reduced to 2 h by preflashing film. This detection system represents a sensitive nonradioactive method, which is applicable not only to Southern blots but also to Northern and Western blots and any assay in which alkaline phosphatase is the label. o 1990 ACTdemic
Press,
Inc.
Nucleic acid hybridization is commonly used to detect specific DNA sequences. Under conditions of appropriate stringency, complementary strands bind with high affinity and specificity. Frequently, detection of a single copy gene within human genomic DNA is required. On ’ To whom
correspondence
should
0003-2697/90 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form
be addressed.
Inc. reserved.
average this may represent 1000 bases from a total of 3 X 10’ bases. Since the samples of DNA available for analysis are often small (~10 pg) this can constitute levels of target DNA of between 1 and 5 pg (1 to 5 X 10-l’ mol). Amplification systems such as the polymerase chain reaction may be used to amplify the target nucleic acid resulting in improved sensitivity but quantification of target material is not usually possible. A highly sensitive detection system removes the requirement for target amplification. Radioactive detection methods are most frequently employed, in particular the incorporation of 32Pinto nucleic acid probes with detection limits down to 1 to 10 fg (1 to 10 X lO-‘l mol) target DNA (1). However, sensitive nonradioactive methods are increasingly being sought. Enzymes are widely used in detection systems for protein or nucleic acid since by their turnover they provide an inherent amplification mechanism. Conventional calorimetric methods lack the necessary sensitivity, while the detection of light from chemiluminescent reactions by film, phototube-based luminometers, or CCD cameras is increasingly being used. For example, the horseradish peroxidase/luminol-enhanced chemiluminescent system has recently been applied to clinical analyte determination in a microtiter plate format (2), to DNA dot blots (3), and to Southern/Northern blotting (15). A chemically triggered system based on probes labeled with acridinium esters can also be used (4). This relies on the protection from chemical attack of the acridinium ester label by the target DNA complementary to the probe in the double-stranded form. An alternative enzymatically triggered chemiluminescent system based on 1,2dioxetane breakdown has also recently been described. Derivatives for use with a wide variety of enzymes including alkaline phosphatase and /3-galactosidase have been prepared (56). After deprotection by the enzyme, the unstable dioxetane intermediate produced decomposes with concomitant emission of light. 353
354
POLLARD-KNIGHT
The 1,2dioxetanes were also developed as chemiluminescent compounds triggerable by chemical reagents and the effect on stability of various chemical substituents was determined (7). One of the dioxetanes investigated carried a phenolic substituent and had a silyloxy protecting group. The protecting group could be chemically released by fluoride ions in organic solvent and this produced an unstable dioxetane phenoxide. The quantum yield of light released by the breakdown of this intermediate in DMSO’ was 0.25 (7). If this protecting group is phosphate and this is released by a phosphatase in aqueous medium the quantum yield is reduced to 1.3 X 10m5. This can be increased to 4.8 X low3 by addition of a detergent-solubilized derivative of fluorescein through the process of intermolecular energy transfer (8). The phosphodioxetane substrate can be triggered by alkaline phosphatase, a stable enzyme which has a high turnover (K,,, equal to 4100 s-l). A recent report by Clyne et al. (9) has demonstrated the application of this substrate system to a microtiter plate-based DNA probe sandwich assay for Chlumydia. In this paper we describe its application to DNA detection on Southern blots. Less than 1 pg DNA can be detected and the light output is maintained for up to 7 days. MATERIALS
Solution
AND
METHODS
Assays
Solution assays were performed in black 96-well microtiter plates (Dynatech Laboratories, Inc.). Aliquots of calf intestinal alkaline phosphatase, enzyme immunoassay grade (Boehringer), were added to the substrate solution supplied by Lumigen, Inc. (Detroit, MI) and contained 0.33 mM 4-(phenyl-3-phosphate)-4-methoxyspiro[l,2-dioxetane-3,2’-adamantane] (Lumigen PPD), cetyltrimethylammonium bromide, 5-(hexadecanoyl)aminofluorescein, 0.8 IYIM magnesium acetate, and 0.75 M 2-amino-2-methyl-1-propanol, pH 9.1. This solution can be stored at 4°C for at least 3 months. For pH dependence experiments, the individual components were mixed in buffer prepared by mixing appropriate volumes of the free base and the hydrochloride to give the required pH. Luminescence from the plate was read on a microtiter plate luminescence reader developed at Amersham with an integration time of 200 msec/well. X Southern Blots Loadings of 100, 10, and 2 pg X DNA, digested with Hind111 (New England Biolabs) were run on 0.8% agarose gels and after depurination and denaturation trans2Abbreviations used: DMSO, dimethyl sulfoxide; roindolyl phosphate; NBT, nitroblue tetrazolium; (phenyl-3-phosphate)-4-methoxyspiro[l,2-~oxe~ne-3,2’-a~mantane] ,
BCIP, bromochloLumigen PPD, 4-
ET
AL.
ferred to nitrocellulose by capillary blotting and fixed by baking (1). The blots were probed overnight with 100 ng/ ml homologous DNA, biotinylated by nick-translation using biotinylated dUTP. The hybridized biotinylated probe was detected with alkaline phosphatase-streptavidin complex and by soaking in the luminescent substrate (as described for solution assays) for 30 to 60 min. They were then wrapped in Saranwrap and exposed to Kodak Ortho G (green sensitive) X-ray film. Green sensitive film was used as the light is emitted at 525 nm. Exposure times were as indicated in the text but varied between 2 and 16 h. Comparisons were made with colorimetric substrate containing 0.4 mM nitroblue tetrazolium (Sigma) and 0.4 mM bromochloroindolyl phosphate (Sigma) in 0.1 M Tris, pH 9.5, 0.1 M sodium chloride, and 10 mM magnesium chloride. Blots were incubated in substrate overnight (12-16 h) although most of the colored product appeared within 2 to 4 h. All methods other than the luminescent detection were as in the protocols supplied with the Amtutor kit from Amersham. Human
Genomic Southern Blots
Loadings of 10 and 5 pg human genomic DNA restricted with EcoRI were run on agarose gels and transferred onto nitrocellulose as for the X blots. Probes for the protooncogenes mos and raf-1 (Amprobes, Amersham) biotinylated by nick-translation were hybridized overnight against the genomic blots at concentrations of 100 rig/ml. Luminescent detection was as described for the X blots. Where indicated, preflashing of film was performed to give an optical density change of 0.1 at 600 nm. RESULTS
Solution
Assays
As an indication of sensitivity, dilutions of alkaline phosphatase were added to the luminescent substrate solution. The results demonstrate that 100 fg (0.7 amol) alkaline phosphatase is detectable on this Amersham plate reader with 200 ms integration (Fig. 1). Using the more sensitive gallium arsenide photocathode photomultiplier in the photon counting mode, 0.0016 amol alkaline phosphatase has been detected. The profile of the light output shows a delay in attaining a steady state which is dependent on the alkaline phosphatase concentration. The delay can be reduced by increasing pH as shown in Fig. 2. At pH 10 the steady state is attained in 100 min but at higher pHs the light output is significantly decreased. X Southern Blots To determine the sensitivity of this detection system for Southern blotting a model system of X DNA digested with HindIII was used to provide restriction fragments
CHEMILUMINESCENT
DNA
lOOpg (700amol) +
Iw
(7fmol) %-
DETECTION
ON
SOUTHERN
BLOTS
low (70amoQ
IOOfg (0.7a mol)
X
A
LIGHT OUTPUT - arbitrary units
3000
2000
* /”
w’
*--
*--*
*-*-*-*
-*-
1000
0
0
50
100
150
200
250
TIME (minutes) FIG. 1.
Light output measured on addition components as under Materials and Methods.
of 10 ~1 alkaline phosphatase solution to 100 81 PPD-containing substrate. Light output was measured on a microtiter plate luminescence reader (counts
in known quantities (see Fig. 3a). The DNA was transferred to nitrocellulose so that any losses experienced in this process were taken into account, unlike dotting labeled probes directly onto the membrane. Using the luminescent substrate 280 fg target DNA could be detected in the 2-pg track after an initial 12-h exposure to green sensitive film (Fig. 3b), compared with 1.4 pg detected in the lo-pg track by the calorimetric method (Fig. 3~). On Day 2 180 fg DNA was detectable in the 2-pg track by the luminescent method in an overnight exposure and signal was still strong on the seventh day although sensitivity was decreased to 280 fg (Fig. 4). Over the course of this period, there was no need to add fresh substrate to the blots. The optimum exposure time in this experiment was 16 h. By increasing the pH of the substrate to 9.7 the exposure time can be significantly reduced but there does not appear to be any significant increase in overall sensitivity.
Human
Genomic Southern Blots
The chemiluminescent system can be used to detect single copy genes in human genomic DNA as shown in Fig. 5. Probes complementary to the proto-oncogenes mos and raf-1 were hybridized to Southern-blotted human genomic DNA. The probes are contained within the
Buffers and other per 200 ms).
plasmid vector pSP6. The target restriction fragments are 2.8 and 2.9 kb, respectively, providing 4.2 and 2.4 pg (5.5 and 5.7 X 10-l’ mol) target DNA in the 5-pg loading tracks. After a 12-h exposure to film the signal is clearly visible with little background. Film suffers from reciprocity failure at the low light levels: activated silver crystals can decay back to the ground state. In order to overcome this, preflashing can be used which lowers the threshold of light required to produce a stable activated crystal. This itself results in a small change in optical density of the film but not enough to fog it. However, the difference in detection with the preflashed film is shown in Fig. 6. Using 2-h exposures the signal from blots probed for the mos gene is more intense on preflashed film with little effect on background. DISCUSSION
Lumigen PPD is a novel substrate for alkaline phosphatase which we have applied to DNA detection on Southern blots. The advantage of using an enzyme as the label rather than a chemically triggerable luminescent substrate is that the enzymatic turnover provides a significant amplification mechanism. In the case of alkaline phosphatase with PPD the turnover rate (KC,) is 4100 s-l. This provides an extremely sensitive detection
356
POLLARD-KNIGHT
pH 9.1 *
pH 9.4 +
ET
AL.
pH 9.7
pH 10.0 m
X
pH 10.3 A
arbitrarv units
600
400
200
0 0
50
100
150
200
250
TIME (minutes)
pH 9.1 3c
pH 9.4 +
pH 9.7
pH 10.0
X
q
OUTPUT - arbitrary units b
300
200
100
0 100
150
200
250
TIME (minutes) FIG. 2.
The pH dependence of light output measured on addition phosphatase to 100 ~1 PPD-containing substrate. Buffers and other on a microtiter plate luminescence reader (counts per 200 ms).
of 10 ~1 solution containing (a) 1 pg (7 amol) components as under Materials and Methods
or (b) 100 fg (0.7 amol) alkaline Light intensity was measured
CHEMILUMINESCENT
DNA
DETECTION
ON
SOUTHERN
BLOTS
357
a Total DNA loaded (pg)
2.00
10.0
100
Target DNA per band (pg)
0.96 0.38 0.28 0.18 0.10 0.08
4.8 1.9 1.4 0.9 0.5 0.4
48 19 14 9 5 4
FIG. 3. (a) Quantities of DNA obtained on digesting X phage with Hind111 followed by separation of the fragments by agarose gel electrophoresis. The three tracks relate to bands obtained from total DNA loadings of 2,10, and 100 pg. (b) Chemiluminescent detection of Southern-blotted X DNA HindI restriction fragments using biotinylated homologous probe, biotinylated alkaline phosphatase-streptavidin complex, and PPDcontaining substrate. Buffers and other components as under Materials and Methods. Total DNA loaded in the three tracks was 2,10, and 100 pg. The blot was exposed to film for 12 h after the initial incubation in substrate. (c) As in b but calorimetric detection incubating in NBT/ BCIP substrate over a 12-h period.
method, as indicated by the detection of 0.7 amol (100 fg) alkaline phosphatase on a luminescence plate reader in solution. Nonenzymatic dephosphorylation does not appear to be a significant problem with this substrate. The delay in attaining a steady-state light output results from the long lifetime of the dephosphorylated dioxetane intermediate in the presence of micelles (t”2 = 37 min). This lifetime can be decreased by increasing the pH, for example to pH 9.7, resulting in a reduction in the delay. In applying this technique to Southern blotting sensitivity relies not only on the absolute alkaline phospha-
tase detection limits but also on the efficiency of labeling DNA. We have used biotinylated probes labeled by nicktranslation followed by binding of biotinylated alkaline phosphatase-streptavidin complex. Another equally
FIG. 5.
FIG. 4.
Chemiluminescent detection of X DNA ZfindIII restriction fragments with biotinylated probes, biotinylated alkaline phosphatase-streptavidin complex, and PPD-containing substrate, as in Fig. 3b. Total DNA loaded in each track was (1) 100 pg, (2) 10 pg, and (3) 2 pg. Blots were exposed to film for 12 h on (a) Day 1, (b) Day 2, and (c) Day 7.
Detection of (a) mos and (b) raf-1 proto-oncogenes in Southern-blotted human genomic DNA restricted with EcoRl. Biotinylated probes were detected with biotinylated alkaline phosphatase-streptavidin complex and PPD-containing substrate. Buffers and other components as under Materials and Methods. Total DNA loaded in the two tracks was (1) 10 pg and (2) 5 pg. The blots were exposed to film for 12 h.
358
POLLARD-KNIGHT
FIG. 6. Detection of the mos proto-oncogene in a Southern blot of human genomic DNA (as in Fig. 5). (a) A 2-h exposure to unflashed film. (b) A 2-h exposure to film preflashed to an optical density of 0.1 at 600 nm.
efficient method for incorporation of biotin into DNA is random primer labeling (10,ll). Other indirect methods of labeling DNA include the use of antibody conjugates with hapten-derivatized nucleotides and chemical labeling methods. Direct enzyme labeling methods are also available (e.g., (12)) and alkaline phosphatase-labeled oligonucleotides can also be made (e.g., (13)). We have not compared the efficiencies of the different labeling methods. A major advantage of the PPD-containing substrate is that signal is generated over a long time period, at least 7 days on blots without addition of fresh substrate, allowing reexposure of the blot to film to attain the optimal signal:noise ratio. This is the result of the high stability of alkaline phosphatase. In addition, the products of the reaction do not seem to cause appreciable enzyme inhibition as do the products of the NBT/BCIP reaction which may form insoluble precipitates around the enzyme (14). In contrast with substrates which form soluble products, the signal from PPD is extremely well localized, with resolution comparable to that from 32P. Since the half-life of the dioxetane intermediate is 37 min in the micellar system, diffusion of the micelle-solubilized intermediate must in some way be prevented. It is not yet clear how this occurs.
ET
AL.
This paper describes the use of film to detect light output from blots. This is inexpensive and widely available, especially in laboratories equipped for using 32P. We have also used cooled and intensified CCD cameras which have a linear response to light over several orders of magnitude and can be used to integrate signal over long periods. In conclusion, we have shown that the PPD/fluorescent micelle system can be used with an alkaline phosphatase label for highly sensitive DNA detection on Southern blots. The substrate has a long shelf life: we have stored it in association with fluorescent micelles at 4°C for 3 months with no change in performance. We have used it to detect 180 fg target DNA in a model system and the equivalent of 2.4 pg target DNA in a single copy gene analysis of 5 pg DNA. It can be used for any application in which alkaline phosphatase is the label, e.g., Western blots, Northern blots, and Southern blots. This paper describes work carried out on nitrocellulose membranes. We are now applying the technique to nylon membranes where shorter exposure times to film are required.
REFERENCES 1. Southern,
E. M. (1975)
J. Mol.
Biul.
S&503-517.
2. Edwards, J. C. (198’7) Znt. Clin. Prod. Reu. Jan/Feb, 30-37. 3. Matthews, J. A., Batki, A., Hynds, C., and Kricka, L. J. (1985) Anal. Biochem. 15 1,205-209. 4. Arnold, L. J., Hammond, P. W., Wiese, W. A., and Nelson, N. C. (1989) Clin. Chm. 35,1588-1594. 5. Schaap, A. P., Sandison, M. D., and Handley, hedron L&t. 28.1159-1162. Chemilumin. 6. Schaap, A. P. (1988) J. Biolumin.
R. S. (1987)
Tetra-
2,253.
I. Schaap, A. P., Chen, T.-S., Handley, R. S., DeSilva, B. P. (1987) Tetrahedron L&t. 28,1155-1158.
R., and Giri,
8. Schaap, A. P., Akhavan, H., and Romano, L. J. (1989) Clin. Chem. 35,1863-1864. 9. Clyne, J. M., Running, J. A., Sanchez-Pescador, R., Besemer, D., Stempien, M., Schaap, A. P., Stephens, R. S., and Urdea, M. S. (1988) J. Biolumin. Chemilumin. 2,193. 10. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 132, 6-13. 11. Feinberg, A. P., and Vogel&in, B. (1984) Anal. Biochem. 137, 266-267. 12. Renz,
M., and Kurz,
C. (1984)
Nucleic
Acids
13. Jablonski, E., Moomaw, E. W., Tullis, (1986) Nucleic Acids Res. 14,6115-6128.
Res.
R. H.,
12,3435-3444. and
Ruth,
J. L.
14. Yasue, H., and Awata, T. (1988) Anal. Biochem. 169,410-414. 15. Pollard-Knight, D., Read, C. A., Downes, M. J., Howard, L. A., Leadbetter, M. R., Pheby, S. A., McNaughton, E., Syms, A., and Brady, M. A. W. (1990) Anal. Biochem. 185,84-89.
185,353-358
(1990)
Nonradioactive DNA Detection on Southern Blots by Enzymatically Triggered Chemiluminescence Denise Pollard-Knight,*91 Adrian C. Simmonds,* A. Paul Schaap,? Hashem Akhavan,? and Michael A. W. Brady* *Corporate Research, Amersham international plc, Pollards Wood Luboratories, White Lion Road, Amershum, Bucks, HP7 9LL, United Kingdom, and TDepartment of Chemistry, Wayne State University, Detroit, Michigan 48202
Received
August
24,1989
A chemiluminescent reaction based on the deprotection of a phosphorylated phenyl dioxetane by alkaline phosphatase has recently been described (Schaap, A. P., 1988, J. Biolumin. Chemilumin. 2, 253). Light output is enhanced by intermolecular energy transfer to a micelle-solubilized fluorophore. This system is applied here to the detection of DNA probes on Southern blots. Enzyme solution assays which give an indication of sensitivity show that using this substrate 100 fg (0.7 amol) alkaline phosphatase can be detected on a luminescence plate reader (200 ms reading time). In a model Southern blotting system 180 fg HindIII digested X DNA was detected on film with homologous biotinylated DNA and a streptavidin-alkaline phosphatase complex. The single copy genes mos and r&-l, representing targets of 4.2 and 2.4 pg target DNA respectively, have also been detected in Southern-blotted human genomic DNA. A delay in reaching a plateau level of light output which is dependent on pH is observed but signal continues for at least 7 days. Typically, 12-h exposures to X-ray film were performed but once a steady-state light output had been achieved this time could be reduced to 2 h by preflashing film. This detection system represents a sensitive nonradioactive method, which is applicable not only to Southern blots but also to Northern and Western blots and any assay in which alkaline phosphatase is the label. o 1990 ACTdemic
Press,
Inc.
Nucleic acid hybridization is commonly used to detect specific DNA sequences. Under conditions of appropriate stringency, complementary strands bind with high affinity and specificity. Frequently, detection of a single copy gene within human genomic DNA is required. On ’ To whom
correspondence
should
0003-2697/90 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form
be addressed.
Inc. reserved.
average this may represent 1000 bases from a total of 3 X 10’ bases. Since the samples of DNA available for analysis are often small (~10 pg) this can constitute levels of target DNA of between 1 and 5 pg (1 to 5 X 10-l’ mol). Amplification systems such as the polymerase chain reaction may be used to amplify the target nucleic acid resulting in improved sensitivity but quantification of target material is not usually possible. A highly sensitive detection system removes the requirement for target amplification. Radioactive detection methods are most frequently employed, in particular the incorporation of 32Pinto nucleic acid probes with detection limits down to 1 to 10 fg (1 to 10 X lO-‘l mol) target DNA (1). However, sensitive nonradioactive methods are increasingly being sought. Enzymes are widely used in detection systems for protein or nucleic acid since by their turnover they provide an inherent amplification mechanism. Conventional calorimetric methods lack the necessary sensitivity, while the detection of light from chemiluminescent reactions by film, phototube-based luminometers, or CCD cameras is increasingly being used. For example, the horseradish peroxidase/luminol-enhanced chemiluminescent system has recently been applied to clinical analyte determination in a microtiter plate format (2), to DNA dot blots (3), and to Southern/Northern blotting (15). A chemically triggered system based on probes labeled with acridinium esters can also be used (4). This relies on the protection from chemical attack of the acridinium ester label by the target DNA complementary to the probe in the double-stranded form. An alternative enzymatically triggered chemiluminescent system based on 1,2dioxetane breakdown has also recently been described. Derivatives for use with a wide variety of enzymes including alkaline phosphatase and /3-galactosidase have been prepared (56). After deprotection by the enzyme, the unstable dioxetane intermediate produced decomposes with concomitant emission of light. 353
354
POLLARD-KNIGHT
The 1,2dioxetanes were also developed as chemiluminescent compounds triggerable by chemical reagents and the effect on stability of various chemical substituents was determined (7). One of the dioxetanes investigated carried a phenolic substituent and had a silyloxy protecting group. The protecting group could be chemically released by fluoride ions in organic solvent and this produced an unstable dioxetane phenoxide. The quantum yield of light released by the breakdown of this intermediate in DMSO’ was 0.25 (7). If this protecting group is phosphate and this is released by a phosphatase in aqueous medium the quantum yield is reduced to 1.3 X 10m5. This can be increased to 4.8 X low3 by addition of a detergent-solubilized derivative of fluorescein through the process of intermolecular energy transfer (8). The phosphodioxetane substrate can be triggered by alkaline phosphatase, a stable enzyme which has a high turnover (K,,, equal to 4100 s-l). A recent report by Clyne et al. (9) has demonstrated the application of this substrate system to a microtiter plate-based DNA probe sandwich assay for Chlumydia. In this paper we describe its application to DNA detection on Southern blots. Less than 1 pg DNA can be detected and the light output is maintained for up to 7 days. MATERIALS
Solution
AND
METHODS
Assays
Solution assays were performed in black 96-well microtiter plates (Dynatech Laboratories, Inc.). Aliquots of calf intestinal alkaline phosphatase, enzyme immunoassay grade (Boehringer), were added to the substrate solution supplied by Lumigen, Inc. (Detroit, MI) and contained 0.33 mM 4-(phenyl-3-phosphate)-4-methoxyspiro[l,2-dioxetane-3,2’-adamantane] (Lumigen PPD), cetyltrimethylammonium bromide, 5-(hexadecanoyl)aminofluorescein, 0.8 IYIM magnesium acetate, and 0.75 M 2-amino-2-methyl-1-propanol, pH 9.1. This solution can be stored at 4°C for at least 3 months. For pH dependence experiments, the individual components were mixed in buffer prepared by mixing appropriate volumes of the free base and the hydrochloride to give the required pH. Luminescence from the plate was read on a microtiter plate luminescence reader developed at Amersham with an integration time of 200 msec/well. X Southern Blots Loadings of 100, 10, and 2 pg X DNA, digested with Hind111 (New England Biolabs) were run on 0.8% agarose gels and after depurination and denaturation trans2Abbreviations used: DMSO, dimethyl sulfoxide; roindolyl phosphate; NBT, nitroblue tetrazolium; (phenyl-3-phosphate)-4-methoxyspiro[l,2-~oxe~ne-3,2’-a~mantane] ,
BCIP, bromochloLumigen PPD, 4-
ET
AL.
ferred to nitrocellulose by capillary blotting and fixed by baking (1). The blots were probed overnight with 100 ng/ ml homologous DNA, biotinylated by nick-translation using biotinylated dUTP. The hybridized biotinylated probe was detected with alkaline phosphatase-streptavidin complex and by soaking in the luminescent substrate (as described for solution assays) for 30 to 60 min. They were then wrapped in Saranwrap and exposed to Kodak Ortho G (green sensitive) X-ray film. Green sensitive film was used as the light is emitted at 525 nm. Exposure times were as indicated in the text but varied between 2 and 16 h. Comparisons were made with colorimetric substrate containing 0.4 mM nitroblue tetrazolium (Sigma) and 0.4 mM bromochloroindolyl phosphate (Sigma) in 0.1 M Tris, pH 9.5, 0.1 M sodium chloride, and 10 mM magnesium chloride. Blots were incubated in substrate overnight (12-16 h) although most of the colored product appeared within 2 to 4 h. All methods other than the luminescent detection were as in the protocols supplied with the Amtutor kit from Amersham. Human
Genomic Southern Blots
Loadings of 10 and 5 pg human genomic DNA restricted with EcoRI were run on agarose gels and transferred onto nitrocellulose as for the X blots. Probes for the protooncogenes mos and raf-1 (Amprobes, Amersham) biotinylated by nick-translation were hybridized overnight against the genomic blots at concentrations of 100 rig/ml. Luminescent detection was as described for the X blots. Where indicated, preflashing of film was performed to give an optical density change of 0.1 at 600 nm. RESULTS
Solution
Assays
As an indication of sensitivity, dilutions of alkaline phosphatase were added to the luminescent substrate solution. The results demonstrate that 100 fg (0.7 amol) alkaline phosphatase is detectable on this Amersham plate reader with 200 ms integration (Fig. 1). Using the more sensitive gallium arsenide photocathode photomultiplier in the photon counting mode, 0.0016 amol alkaline phosphatase has been detected. The profile of the light output shows a delay in attaining a steady state which is dependent on the alkaline phosphatase concentration. The delay can be reduced by increasing pH as shown in Fig. 2. At pH 10 the steady state is attained in 100 min but at higher pHs the light output is significantly decreased. X Southern Blots To determine the sensitivity of this detection system for Southern blotting a model system of X DNA digested with HindIII was used to provide restriction fragments
CHEMILUMINESCENT
DNA
lOOpg (700amol) +
Iw
(7fmol) %-
DETECTION
ON
SOUTHERN
BLOTS
low (70amoQ
IOOfg (0.7a mol)
X
A
LIGHT OUTPUT - arbitrary units
3000
2000
* /”
w’
*--
*--*
*-*-*-*
-*-
1000
0
0
50
100
150
200
250
TIME (minutes) FIG. 1.
Light output measured on addition components as under Materials and Methods.
of 10 ~1 alkaline phosphatase solution to 100 81 PPD-containing substrate. Light output was measured on a microtiter plate luminescence reader (counts
in known quantities (see Fig. 3a). The DNA was transferred to nitrocellulose so that any losses experienced in this process were taken into account, unlike dotting labeled probes directly onto the membrane. Using the luminescent substrate 280 fg target DNA could be detected in the 2-pg track after an initial 12-h exposure to green sensitive film (Fig. 3b), compared with 1.4 pg detected in the lo-pg track by the calorimetric method (Fig. 3~). On Day 2 180 fg DNA was detectable in the 2-pg track by the luminescent method in an overnight exposure and signal was still strong on the seventh day although sensitivity was decreased to 280 fg (Fig. 4). Over the course of this period, there was no need to add fresh substrate to the blots. The optimum exposure time in this experiment was 16 h. By increasing the pH of the substrate to 9.7 the exposure time can be significantly reduced but there does not appear to be any significant increase in overall sensitivity.
Human
Genomic Southern Blots
The chemiluminescent system can be used to detect single copy genes in human genomic DNA as shown in Fig. 5. Probes complementary to the proto-oncogenes mos and raf-1 were hybridized to Southern-blotted human genomic DNA. The probes are contained within the
Buffers and other per 200 ms).
plasmid vector pSP6. The target restriction fragments are 2.8 and 2.9 kb, respectively, providing 4.2 and 2.4 pg (5.5 and 5.7 X 10-l’ mol) target DNA in the 5-pg loading tracks. After a 12-h exposure to film the signal is clearly visible with little background. Film suffers from reciprocity failure at the low light levels: activated silver crystals can decay back to the ground state. In order to overcome this, preflashing can be used which lowers the threshold of light required to produce a stable activated crystal. This itself results in a small change in optical density of the film but not enough to fog it. However, the difference in detection with the preflashed film is shown in Fig. 6. Using 2-h exposures the signal from blots probed for the mos gene is more intense on preflashed film with little effect on background. DISCUSSION
Lumigen PPD is a novel substrate for alkaline phosphatase which we have applied to DNA detection on Southern blots. The advantage of using an enzyme as the label rather than a chemically triggerable luminescent substrate is that the enzymatic turnover provides a significant amplification mechanism. In the case of alkaline phosphatase with PPD the turnover rate (KC,) is 4100 s-l. This provides an extremely sensitive detection
356
POLLARD-KNIGHT
pH 9.1 *
pH 9.4 +
ET
AL.
pH 9.7
pH 10.0 m
X
pH 10.3 A
arbitrarv units
600
400
200
0 0
50
100
150
200
250
TIME (minutes)
pH 9.1 3c
pH 9.4 +
pH 9.7
pH 10.0
X
q
OUTPUT - arbitrary units b
300
200
100
0 100
150
200
250
TIME (minutes) FIG. 2.
The pH dependence of light output measured on addition phosphatase to 100 ~1 PPD-containing substrate. Buffers and other on a microtiter plate luminescence reader (counts per 200 ms).
of 10 ~1 solution containing (a) 1 pg (7 amol) components as under Materials and Methods
or (b) 100 fg (0.7 amol) alkaline Light intensity was measured
CHEMILUMINESCENT
DNA
DETECTION
ON
SOUTHERN
BLOTS
357
a Total DNA loaded (pg)
2.00
10.0
100
Target DNA per band (pg)
0.96 0.38 0.28 0.18 0.10 0.08
4.8 1.9 1.4 0.9 0.5 0.4
48 19 14 9 5 4
FIG. 3. (a) Quantities of DNA obtained on digesting X phage with Hind111 followed by separation of the fragments by agarose gel electrophoresis. The three tracks relate to bands obtained from total DNA loadings of 2,10, and 100 pg. (b) Chemiluminescent detection of Southern-blotted X DNA HindI restriction fragments using biotinylated homologous probe, biotinylated alkaline phosphatase-streptavidin complex, and PPDcontaining substrate. Buffers and other components as under Materials and Methods. Total DNA loaded in the three tracks was 2,10, and 100 pg. The blot was exposed to film for 12 h after the initial incubation in substrate. (c) As in b but calorimetric detection incubating in NBT/ BCIP substrate over a 12-h period.
method, as indicated by the detection of 0.7 amol (100 fg) alkaline phosphatase on a luminescence plate reader in solution. Nonenzymatic dephosphorylation does not appear to be a significant problem with this substrate. The delay in attaining a steady-state light output results from the long lifetime of the dephosphorylated dioxetane intermediate in the presence of micelles (t”2 = 37 min). This lifetime can be decreased by increasing the pH, for example to pH 9.7, resulting in a reduction in the delay. In applying this technique to Southern blotting sensitivity relies not only on the absolute alkaline phospha-
tase detection limits but also on the efficiency of labeling DNA. We have used biotinylated probes labeled by nicktranslation followed by binding of biotinylated alkaline phosphatase-streptavidin complex. Another equally
FIG. 5.
FIG. 4.
Chemiluminescent detection of X DNA ZfindIII restriction fragments with biotinylated probes, biotinylated alkaline phosphatase-streptavidin complex, and PPD-containing substrate, as in Fig. 3b. Total DNA loaded in each track was (1) 100 pg, (2) 10 pg, and (3) 2 pg. Blots were exposed to film for 12 h on (a) Day 1, (b) Day 2, and (c) Day 7.
Detection of (a) mos and (b) raf-1 proto-oncogenes in Southern-blotted human genomic DNA restricted with EcoRl. Biotinylated probes were detected with biotinylated alkaline phosphatase-streptavidin complex and PPD-containing substrate. Buffers and other components as under Materials and Methods. Total DNA loaded in the two tracks was (1) 10 pg and (2) 5 pg. The blots were exposed to film for 12 h.
358
POLLARD-KNIGHT
FIG. 6. Detection of the mos proto-oncogene in a Southern blot of human genomic DNA (as in Fig. 5). (a) A 2-h exposure to unflashed film. (b) A 2-h exposure to film preflashed to an optical density of 0.1 at 600 nm.
efficient method for incorporation of biotin into DNA is random primer labeling (10,ll). Other indirect methods of labeling DNA include the use of antibody conjugates with hapten-derivatized nucleotides and chemical labeling methods. Direct enzyme labeling methods are also available (e.g., (12)) and alkaline phosphatase-labeled oligonucleotides can also be made (e.g., (13)). We have not compared the efficiencies of the different labeling methods. A major advantage of the PPD-containing substrate is that signal is generated over a long time period, at least 7 days on blots without addition of fresh substrate, allowing reexposure of the blot to film to attain the optimal signal:noise ratio. This is the result of the high stability of alkaline phosphatase. In addition, the products of the reaction do not seem to cause appreciable enzyme inhibition as do the products of the NBT/BCIP reaction which may form insoluble precipitates around the enzyme (14). In contrast with substrates which form soluble products, the signal from PPD is extremely well localized, with resolution comparable to that from 32P. Since the half-life of the dioxetane intermediate is 37 min in the micellar system, diffusion of the micelle-solubilized intermediate must in some way be prevented. It is not yet clear how this occurs.
ET
AL.
This paper describes the use of film to detect light output from blots. This is inexpensive and widely available, especially in laboratories equipped for using 32P. We have also used cooled and intensified CCD cameras which have a linear response to light over several orders of magnitude and can be used to integrate signal over long periods. In conclusion, we have shown that the PPD/fluorescent micelle system can be used with an alkaline phosphatase label for highly sensitive DNA detection on Southern blots. The substrate has a long shelf life: we have stored it in association with fluorescent micelles at 4°C for 3 months with no change in performance. We have used it to detect 180 fg target DNA in a model system and the equivalent of 2.4 pg target DNA in a single copy gene analysis of 5 pg DNA. It can be used for any application in which alkaline phosphatase is the label, e.g., Western blots, Northern blots, and Southern blots. This paper describes work carried out on nitrocellulose membranes. We are now applying the technique to nylon membranes where shorter exposure times to film are required.
REFERENCES 1. Southern,
E. M. (1975)
J. Mol.
Biul.
S&503-517.
2. Edwards, J. C. (198’7) Znt. Clin. Prod. Reu. Jan/Feb, 30-37. 3. Matthews, J. A., Batki, A., Hynds, C., and Kricka, L. J. (1985) Anal. Biochem. 15 1,205-209. 4. Arnold, L. J., Hammond, P. W., Wiese, W. A., and Nelson, N. C. (1989) Clin. Chm. 35,1588-1594. 5. Schaap, A. P., Sandison, M. D., and Handley, hedron L&t. 28.1159-1162. Chemilumin. 6. Schaap, A. P. (1988) J. Biolumin.
R. S. (1987)
Tetra-
2,253.
I. Schaap, A. P., Chen, T.-S., Handley, R. S., DeSilva, B. P. (1987) Tetrahedron L&t. 28,1155-1158.
R., and Giri,
8. Schaap, A. P., Akhavan, H., and Romano, L. J. (1989) Clin. Chem. 35,1863-1864. 9. Clyne, J. M., Running, J. A., Sanchez-Pescador, R., Besemer, D., Stempien, M., Schaap, A. P., Stephens, R. S., and Urdea, M. S. (1988) J. Biolumin. Chemilumin. 2,193. 10. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 132, 6-13. 11. Feinberg, A. P., and Vogel&in, B. (1984) Anal. Biochem. 137, 266-267. 12. Renz,
M., and Kurz,
C. (1984)
Nucleic
Acids
13. Jablonski, E., Moomaw, E. W., Tullis, (1986) Nucleic Acids Res. 14,6115-6128.
Res.
R. H.,
12,3435-3444. and
Ruth,
J. L.
14. Yasue, H., and Awata, T. (1988) Anal. Biochem. 169,410-414. 15. Pollard-Knight, D., Read, C. A., Downes, M. J., Howard, L. A., Leadbetter, M. R., Pheby, S. A., McNaughton, E., Syms, A., and Brady, M. A. W. (1990) Anal. Biochem. 185,84-89.
