Mutagenesis

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (105) Request Permissions Google Scholar Articles by Phillips, D.H. Articles by Castegnaro, M. Search for Related Content PubMed PubMed Citation Articles by Phillips, D.H. Articles by Castegnaro, M. Social Bookmarking          What's this?

Mutagenesis, Vol. 14, No. 3, 301-315, May 1999
© 1999 UK Environmental Mutagen Society/Oxford University Press

Standardization and validation of DNA adduct postlabelling methods: report of interlaboratory trials and production of recommended protocols

D.H. Phillips³, M. Castegnaro¹ on behalf of the trial participants^,2

Institute of Cancer Research, Haddow Laboratories, Cotswold Road, Sutton, Surrey SM2 5NG, UK and ¹ International Agency for Research on Cancer, 150 cours Albert-Thomas, 69372 Lyon Cedex 08, France

	Abstract

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
The aim of this project was to devise and test improved protocols of the ³²P-postlabelling assay for the detection of carcinogen–DNA adducts. The intention was to reverse the drift of different investigators using increasingly divergent experimental conditions. This would lead to a more standardized assay that can be used in future applications by different investigators for the monitoring of human exposure to genotoxic agents, permitting more meaningful comparisons between different studies or between different participants in the same study. As part of this process, there was perceived to be a need for carcinogen-modified DNA standards of known levels of adducts for use as positive controls, as standards for normalization of results with unknown samples and to assist interlaboratory comparisons. The preparation of characterized DNA standards modified by benzo[a]pyrene (BaP), a polycyclic aromatic hydrocarbon (PAH), 4-aminobiphenyl (ABP), an aromatic amine, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), a heterocyclic amine, and N-methyl-N-nitrosourea (MNU), a methylating agent yielding DNA containing O⁶-methylguanine, was carried out. A critical appraisal of all aspects of the ³²P-postlabelling procedure and investigations to examine the influence of a number of key variations on the assay were conducted. There followed testing of a consensus protocol in a first interlaboratory trial involving 25 participants in Europe and the USA, conducted on the prepared synthetic DNA standards, the assessment of interlaboratory variability and the reasons for it. Revision of the protocols was followed by further testing in a second interlaboratory trial in which liver DNA from mice treated with BaP or ABP were assayed together with the synthetic DNA standards. Adduct levels were found to be significantly lower by ³²P-postlabelling than by ³H incorporation. A recommended set of procedures has been developed for the detection and quantitation of DNA adducts formed by PAHs, aromatic amines and methylating agents. These trials have led to a much clearer idea as to what are the critical features and procedures of the ³²P-postlabelling assay and there is a set of standard DNA samples for use in quality control and against which biological samples can be normalized. Use of these standards and procedures has reduced interlaboratory variability in quantitation of DNA adducts.

	Introduction

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
The ³²P-postlabelling assay is an ultrasensitive method for detecting carcinogen–DNA adducts that is applicable to a very wide range of DNA lesions (Gupta et al., 1982; Beach and Gupta, 1992; Randerath and Randerath, 1994; Phillips, 1997). It has been applied to a large number of studies in which human exposure to genotoxic agents has been investigated. However, it is a complex assay and experimental conditions can vary widely between different investigators. There have been no generally agreed protocols or positive controls available to investigators.

In advance of the first International Conference on Postlabelling Methods for DNA Adducts, held in Lyon in 1992, an interlaboratory trial was conducted, with 15 laboratories participating. The trial was carried out on four DNA samples distributed to each of the participants, who were instructed to assay the samples using methods and conditions currently in use in his/her laboratory. The results revealed good qualitative agreement between laboratories, but significant quantitative differences (Phillips and Castegnaro, 1993). Furthermore, it was apparent that widely differing conditions (for example, enzyme and substrate concentrations and chromatography conditions) were in use, many laboratories having introduced minor, and in some cases major, modifications to the originally published method (Gupta et al., 1982).

The current project was initiated in July 1994 and the European participants were joined by the three following national institutes from the USA: the Environmental Protection Agency (EPA), the National Centre for Toxicological Research (NCTR) and the National Cancer Institute (NCI). These three institutes had already created a `Biomonitoring Group' whose task was to develop DNA adduct standards which could be used for interlaboratory trials on a single method or on comparison of methods and to test the stability of these standards under different storage conditions (e.g. in solution at –20 or –80°C or in liquid nitrogen or dry at the same temperatures).

The main objectives of this project have been to standardize the available ³²P-postlabelling protocols, to identify the various methodological problems and to provide guidelines for the use of ³²P-postlabelling in detecting carcinogen–DNA adducts. This would enable the procedure to be applied rationally to monitoring human exposure to environmental carcinogens in epidemiological studies. These objectives were to be achieved through testing of protocols and analysis of standards and unknown samples in interlaboratory trials.

	Chronology

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
First workshop, 20–21 October 1994, Lyon, France
Prior to this meeting, all those attending (see list of Trial participants) completed a questionnaire on the current postlabelling conditions used in their laboratories. The purpose of the workshop was: (i) to discuss the current status of the assay for the analysis of bulky DNA adducts; (ii) to discuss the reasons for modifications made to the original protocol (Gupta et al., 1982); (iii) to design a common test protocol. Where agreement for the conditions to use in the trial could not be reached, it was agreed to subject these issues to a pre-trial, involving four laboratories, to establish the optimum conditions. Once the results of this exercise had been carried out, the procedure would be incorporated into the test protocol for the first trial.

For the analysis of O⁶-methylguanine (O⁶-MeG), the method of Povey and Cooper (1995) was available and it was decided to use it as such.

At a separate meeting of the `Biomonitoring Group', it was agreed that standard adducted DNA samples would be prepared that contained: (i) a polycyclic aromatic hydrocarbon (benzo[a]pyrene, BaP); (ii) an aromatic amine (4-aminobiphenyl, ABP); (iii) a heterocyclic amine (2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine, PhIP); (iv) methylated bases (by modification with N-methyl-N-nitrosourea, MNU). [See Preparation of DNA adduct standard reference material.]

First mini-trial, winter 1994–summer 1995
A number of variables in the postlabelling assay were investigated by four participating laboratories (those of R.A.Baan, J.Lewtas, M.Castegnaro and D.Phillips) to resolve issues of controversy that arose at the first workshop. The DNA samples used for these investigations were those that had been used in the 1992 interlaboratory trial (Phillips and Castegnaro, 1993) and those prepared for the first interlaboratory trial of the current project (see below).

Preparation of standard adducted DNA samples, spring/summer 1995
DNA adduct standards for the first trial were prepared in the laboratories of P.Fu, F.F.Kadlubar, F.A.Beland and D.H.Phillips (see Preparation of DNA adduct standard reference material).

First interlaboratory trial, autumn 1995
The trial was carried out on the four standard adducted DNA samples, plus the control DNA, using the trial protocol. Investigators also used their own laboratory protocols on the same samples and the results reported were compared.

Second workshop, 8–9 December 1995, Lyon, France
At the second workshop the results of the first trial were considered. Reasons for interlaboratory variability were sought. Refinements to the protocol were discussed. Areas of outstanding uncertainty were identified, answers to which were to be sought in a second mini-trial before the second main trial was carried out. Agreement was reached that the second trial should involve testing of a set of `unknown' samples alongside the DNA standards, whose levels of modification were now known to investigators.

Second mini-trial, spring 1996
This was carried out in three laboratories (those of J.Lewtas, M.Castegnaro and D.H.Phillips) to resolve outstanding issues concerning the most appropriate assay conditions. The protocol was revised, where appropriate, as a result.

Preparation of samples for second interlaboratory trial, summer 1996
DNA isolated from the livers of mice that had been treated with BaP, 4-aminobiphenyl or solvent control was prepared by F.A.Beland. A sample of methylated DNA was prepared by A.Povey and D.Cooper.

Second interlaboratory trial, spring 1997
The second trial involved using the revised trial protocol to examine the DNA samples prepared from mouse liver (see above) and the methylated DNA sample. Investigators were instructed to analyse the standard adducted DNA samples used in the first trial and, using values for their level of modification provided, to normalize the values obtained with the unknown samples. Another laboratory that had developed a low volume, low ATP modification of the postlabelling assay was invited to participate using its own protocol, in order to see what concordance there was with the trial protocol.

Third workshop, 30–31 May 1997, Lyon, France
The results of the second trial were reported and considered. The protocol was scrutinized and discussed at length. A questionnaire was planned for all participants to complete giving further details of their experimental procedures and calculations. Consideration was also given to determination of the true levels of adduction in the standards (originally prepared for the first trial) and the samples prepared for the second trial. Preliminary results were presented using other, non-postlabelling methods of adduct determination. The protocols were finalized and a set of additional recommendations drafted. Further trials and collaborations, beyond the scope of the present project, were planned.

	Design of the trial protocols

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
Prior to the first workshop, each participant was requested to send the detailed methodology being used for the analysis of bulky DNA adducts together with justifications and illustrations of any modification/improvement made to the original method. These documents were analysed and summaries were prepared to initiate the discussions during the meeting. With regard to the methods of analysis of small nucleotides, there was a very restricted choice and one method was selected for analysis of alkyl adducts. This method has been transcribed in an ISO-like style and amended for homogeneity of conditions of storage/reaction with the method for determination of bulky adducts.

In analysing the conditions used in the various collaborating laboratories, a very wide variability was found, as illustrated in the examples below.

To check for DNA purity and quantitation, the following methods were proposed: measurement of A₂₆₀:A₂₈₀ ratio, full UV spectrum, methods using colorimetry after preparation of a DNA standard curve, HPLC/UV analysis of normal nucleotides after hydrolysis and ³²P-postlabelling of normal nucleotides after hydrolysis.

For the hydrolysis of DNA, the micrococcal nuclease (MN) range used was found to be 20–250 mU/µg DNA (6.25–133 mU/µl solution) and that of spleen phosphodiesterase (SPD) 0.28–12.5 mU/µg DNA (0.2–1.3 mU/µl solution). Three digestion methods were proposed: (i) hydrolysis by MN/SPD at fixed concentrations; (ii) digestion by MN at pH 9, then addition of acetate buffer, pH 5, before addition of SPD; (iii) a second addition of MN/SPD after a certain reaction time.

For labelling of the adducts the following range of conditions were found to be in use. The buffer varied both in concentration and pH, the amount of ATP used varied from 7 to 500 µCi (equivalent to 2.4–167 pmol) and the polynucleotide kinase (PNK) from 2–11.7 U, i.e. 0.14–1 U/µl solution (in the NP1 enrichment method) and 0.2–0.6 U/µl solution (in the butanol enrichment method).

A wide variability was also found in the adduct quantification techniques. Some laboratories relied on the activity of the ATP quoted by the supplier and did not check it, others did. Some used quantification of DNA by measurement of the absorbance at 260 nm, whereas others quantified it by HPLC/UV or postlabelling after hydrolysis.

With this degree of variation in the assay, these conditions served as a basis for discussion and the following conclusions were drawn which were then used as the basis of the agreed protocol.

Summary of discussions (agreed protocol)
In the first phase the laboratories were to analyse the standard DNA samples that were being prepared. Each laboratory would know the order of magnitude of the level of modification but not the exact level. [Note: All values given below are concentrations in final solution at each phase.]

Quantitation of DNA. This should be by using absorption at 260 nm but should measure the full spectrum to see what impurities there are in terms of protein. Labelling of normal nucleotides will also be used as a test of purity of the DNA.

DNA hydrolysis. Enzymes should be dialysed and added together to the DNA for a 4 h digestion. Final concentration of solvents and reagents: MN, 60 mU/µl of solution; SPD, 1 mU/µl; sodium succinate, 20 mM, pH 6; calcium chloride, 5 mM. MN and SPD can be stored in separate vials or as a mixture, but at –20°C in small aliquots to avoid freezing and thawing problems.

Enrichment
Butanol extraction (see Gupta, 1985). The method described here is a modification of the original method in which the volume was made to 50 µl and the extractions did not use butanol saturated with water and water saturated with butanol. Tetrabutylammonium chloride (TBA), 1 mM; ammonium formate, 10 mM; volume prior to extraction, 250 µl. Addition of the buffer and of the butanol should be simultaneous. Butanol should be redistilled, two extractions should be performed with butanol saturated with water, then two back washes of the pooled butanol phase with water saturated with butanol. Termination should be performed by adding Tris–HCl to the butanol phase, so that the final concentration after evaporation to dryness and reconstitution in 16 µl water is 60 mM.

Difficulties in labelling in 1 day were discussed. It was decided that the pre-trial group should investigate the possibility of stopping the procedure on the first day at the point where the digest/extracted adducts were at `dryness before reconstitution' and continuing the following day.

Nuclease P1 digestion (see Reddy and Randerath, 1986). The method described here is a modification of the original method, as indicated below. NP1, 0.5 µg/µg DNA; ZnCl₂, 0.1 mM (0.03 mM in the original method) final concentration; sodium acetate, pH 5, 40 mM final concentration. Reaction 30 min (40 min in the original method) at 37°C. Termination of reaction with Tris (the pH will be determined by the pre-trial group) to a final concentration of 80 mM (59 mM in the original method).

Labelling. PNK, 0.5 U/µl final solution; ATP, 3 µM; apyrase will not be used; bicine, 20 mM; spermidine, 0.5–1 mM; MgCl₂, 10 mM; dithiothreitol (DTT), 10 mM; reaction 30 min at 37°C. [Note. DTT is not stable on long-term storage and this instability will inhibit the reaction.]

Separation of adducts. Prewashing of the plates is not essential; it is recommended to use a range of phosphate concentrations in D1 to ensure that adducts remain at the origin; D2, the use of which is described in Gupta et al. (1982), is no longer required; D3 and D4 should be adjusted to spread the spots over the plates; D5 should be used for low count samples.

Quantitation. The method should be based on a method which quantifies normal nucleotides. Background should always be quantified and subtracted.

Consensus was not reached on the following five points.

• Can the DNA digest be stored after hydrolysis? How?
  
• Can the DNA digest/adducts be stored after the NP1 treatment or butanol extraction? How?
  
• Should the butanol enrichment procedure be performed at 4°C?
  
• How many butanol extractions/back washes are optimal?
  
• Which pH should be used for labelling when using a PNK with phosphatase activity, 7.5, 8.0, 8.5, 9.0 or 9.5?
  

The sub-trial group received the two available bulky DNA adduct standards (DNA modified by BaP and ABP) to clarify these points. As a result of this activity the following was found.

• The DNA digest could be safely stored overnight at –80°C after hydrolysis by MN/SPD.
  
• The nuclease P1-enriched solution can be stored overnight at –80°C after addition of buffer. Similarly, after evaporation of butanol and transfer into buffer, the butanol-enriched extract can be stored overnight at –80°C.
  
• There was no advantage in performing butanol extractions at 4°C. It is therefore sufficient to carry out this procedure at room temperature.
  
• Two butanol extractions were found to be sufficient for the ABP–DNA adducts.
  
• The pH optimum for postlabelling using PNK that may possess some phosphatase activity was found to be between pH 8.5 and 9. Lower labelling efficiency was observed below pH 8.5, which could be due to phosphatase activity. If phosphatase-free PNK is used, then the labelling efficiency at pH 7 is equal to that at pH 9.
  

Method for O⁶-MeG
For a preliminary trial of methylated adducts, a published protocol by Povey and Cooper (1995) was adopted.

	Preparation of DNA adduct standard reference material

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
Four DNA adduct standards were prepared at a modification level of ~1 adduct/10⁶ normal nucleotides (see Table I). For this, 10 g of calf thymus DNA were purified and aliquots distributed to the four laboratories who had agreed to prepare these standards. The O⁶-MeG-containing DNA was prepared essentially as described by Lawley and Shah (1973) and the ABP-modified DNA as described by Beland et al. (1999); the BPDE-modified DNA was prepared according to the general method of Zhan et al. (1995) and the PhIP-modified DNA by the method of Lin et al. (1992).

View this table: [in this window] [in a new window]   Table I. Carcinogen-adducted DNA standards

 
These standards have been tested for stability after 1 and 6 months and then at 6 monthly intervals. At present, stocks (in the form of 1 mg aliquots) are held by the laboratories that prepared them.

The standards were used in both interlaboratory trials and have also been supplied on request to several other investigators not involved in the trials, for use as standards and positive controls. These uses include wider comparisons of different methods for DNA adduct determination, as proposed by the `Biomonitoring Group'; others methods currently being investigated are immunoassays (ELISA and DELFIA), fluorescence and GC-MS. Details of the preparation and further analysis of the ABP–DNA standard have been published elsewhere (Beland et al., 1999). Investigators wishing to obtain any of the standards for research use should contact the respective person listed in Table I.

A fifth DNA standard that had been planned, containing etheno adducts (etheno-dG, etheno-dA and etheno-dC) was prepared in another laboratory, but difficulties were experienced with the characterization and quantitation of the adducts in this sample. These problems were not resolved in time by the laboratory involved and thus plans to conduct an interlaboratory trial of postlabelling of this DNA had to be abandoned.

	First interlaboratory trial

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
Analysis of bulky adducts
Each laboratory participating in this part of the study received four samples (BaP-modified DNA, PhIP-modified DNA, ABP-modified DNA and the unmodified calf thymus DNA which was used to prepare the modified DNA). In addition, they received one vial of a single batch of PNK (obtained from Epicentre Technologies). Each laboratory was requested to analyse the samples with the provided PNK and the PNK currently used in their laboratory.

Analysis of O⁶-MeG
Each laboratory participating in this part of the study received two samples (O⁶- and N⁷-methylated DNA and the unmodified calf thymus DNA which was used to prepare the modified DNA). In addition each laboratory received the necessary reagents and equipment not commercially available to carry out the experiment (i.e. one vial containing a solution of 100 µmol/l of deoxycorfomycin, one vial containing a solution of 120 µmol/l O⁶-MerG, one vial containing a solution of 400 µmol/l deoxyinosine 3'-monophosphate and eight immunoaffinity columns).

Collection and analysis of the results from the first trial
All laboratories were requested to fill in a form which had been provided at the same time as the samples/equipment and to send their results to the coordinators. These data were collated, qualitatively and quantitatively analysed and sent back to each participant for checking and study before the group evaluation meeting.

Group evaluation of the results
The second workshop was convened to evaluate the data accumulated concerning the DNA adduct reference materials, to review in detail the data and determine at which step, and how, the methods could be improved and to decide the future steps in the study.

Quantification and stability of reference material
The reference adducted DNA samples were prepared with ³H-labelled test compounds and were designed to contain the adducts of interest at a level of ~1 adduct/10⁶ nucleotides. The level of adduction in each case was determined by enzyme hydrolysis and HPLC separation of adducted nucleosides which were assayed by liquid scintillation counting. This measurement of adduct levels was not made known to the participants until after the first trial had been completed. The BaP–DNA, ABP–DNA and O⁶-MeG–DNA samples have been monitored at several intervals since they were made and have been found to be very stable when stored in solution at –20 (freezer), –75 (freezer) and –140°C (liquid nitrogen).

Qualitative analysis of the results
Each laboratory was requested to provide copies of the chromatograms obtained for each adduct analysis. The overall evaluation of the appearance of these chromatograms is summarized below.

Control DNA. Most participants obtained chromatograms that were essentially devoid of radioactive spots. In a number of cases faint or small `background spots' were seen with reasonable consistency, enabling them to be discounted from analysis of the modified DNA samples. They were in any case much weaker than, for example, the BaP–DNA adduct spots. One participant obtained background radioactivity on chromatograms in all experiments using the supplied PNK, but this observation was not confirmed by other participants. A few participants obtained diffuse areas of radioactivity close to the origin that would have precluded accurate identification or quantification of adducts had they migrated in this area.

BaP–DNA. All participants obtained a major spot, due to the N² guanine substitution by anti-benzo[a]pyrene 7,8-dihydrodiol 9,10-oxide (anti-BPDE). In all cases this appeared as a spot in the lower left-hand quarter of the chromatograms, with roughly equal migration in the x and y directions (D2 and D3, respectively). The only qualitative differences concerned whether or not additional adduct spots were observed. In roughly equal proportions of results no other adduct spots were observed or a very minor spot was observed, partially resolved from the major spot with a greater R_f in both directions, or both this spot and an additional faster migrating spot, again very faint relative to the major spot, were observed (making three spots in all).

PhIP–DNA. The principal PhIP–DNA adduct, a C-8 guanine adduct, migrated with reasonable mobility in D2 but lesser mobility in D3. Its intensity varied from strong to very faint. In some cases, it was the only adduct spot observed, whereas in other cases an additional spot above and to the right of the aforementioned spot was observed, which was frequently more diffuse. In other cases a cluster of spots was observed at this same position. Some laboratories obtained similar patterns with both nuclease P1 enrichment and with butanol extraction enrichment; some laboratories failed to detect adducts with butanol enrichment and some laboratories (more) failed to detect adducts with nuclease P1 enrichment.

The qualitative variations observed may be the result of incomplete digestion of the DNA prior to labelling. Earlier studies have suggested that the faster migrating cluster of spots contains PhIP-modified dinucleotides or oligonuclotides, the presence of which is inversely proportional to the C-8 guanine adduct spot (Pfau et al., 1994).

ABP–DNA. For this DNA sample, only butanol enrichment was attempted, because of the unsuitability of nuclease P1 digestion as an enrichment method for aromatic amine adducts. In most laboratories a single adduct spot was observed. In just a few laboratories (two), additional very minor spots were observed in some analyses. One participant failed to produce a clear adduct pattern (high background near origin).

Quantitative analysis of the results
Table IIA and B present the statistical evaluation of the data for the trial protocol and for the participants' own protocols. The type of PNK used in each laboratory is indicated. When compared with the 1992 collaborative study (Phillips and Castegnaro, 1993), the results obtained for the analysis of bulky adducts were not significantly improved. As demonstrated in Table IIB, one source of variability in the results could have been related to the source of PNK used.

View this table: [in this window] [in a new window]   Table IIA. Statistical evaluation of the results of the first interlaboratory trial: results from trial method, using PNK provided

 

View this table: [in this window] [in a new window]   Table IIB. Statistical evaluation of the results of the first interlaboratory trial: all results

 
With the BaP–DNA samples and using the PNK provided, the mean values with the nuclease P1 and butanol methods were 93 and 77%, respectively, of the value of 111 adducts/10⁸ nucleotides obtained from measurement of the ³H label in the DNA samples (Table IIA). Some laboratories were closer with the trial method than with their own method, others obtained similar values with both methods and others were closer with their own method. With the trial method, the distributions of the results are partially Gaussian, tailing to the high value for both the nuclease P1 and butanol enrichment methods. For both enrichment methods, the median is ~20% lower than the mean. When investigators used their own methods with the PNK provided there was a trend to values 30% lower than that obtained by ³H incorporation, but mean and median values were much closer and there was less interlaboratory variation.

With the ABP–DNA samples, all values obtained were significantly lower than the value obtained by ³H incorporation (62 adducts/10⁸ nucleotides), with one exception where the value was close to that value. The only significant difference between this laboratory's method and others was the source of PNK and this effect was investigated in the subsequent mini-trial. The reason for the low values obtained by the other participants is not clear. There was a similar distribution of values as seen with the BaP–DNA sample, with a high coefficient of variation (~91%) with the trial method, however, in contrast to the BaP–DNA sample, this was better than when the investigators had used their own methods (~115%).

For the PhIP–DNA samples, all values were significantly lower than that calculated from ³H incorporation (350 adducts/10⁸ nucleotides). Using the PNK provided, the distribution of results was clearly non-Gaussian with high variability mainly when using the nuclease P1 enrichment technique (coefficient of variation ~100% with both the trial method and the investigators' own methods). Less variability was obtained with the butanol extraction technique, with coefficients of variation of 76 and 46%, respectively, for the trial method and the investigators' own methods. The quantitative results along with the qualitative findings suggest that PhIP–DNA adducts are difficult to detect reliably by the trial and investigators' own postlabelling methods. Recent investigations by one participant have indicated a 5–10% recovery of adduct by TLC, while a 70–80% recovery is achievable by HPLC. Reasons for the poor recovery on TLC have yet to be elucidated.

For many participants the analysis of O⁶-MeG was being attempted for the first time. There was therefore some variability in results, presented in Table III, with some laboratories apparently failing to detect the methylated adduct and others obtaining values close to the expected value obtained by HPLC analysis of the DNA hydrolysate and quantitation by scintillation counting of the adduct peak. Accordingly, a repeat trial by these participants was planned for the next part of the study (see below).

View this table: [in this window] [in a new window]   Table III. Statistical evaluation of results of the first interlaboratory trial of O6-MeG 32P-postlabelling

 
Evaluation of the data and conclusions
The above data served as a basis for discussion of the method. During the discussion, it became clear that some laboratories deviated in part from the prescribed protocols which might have explained part of the variability in the results. In the light of subsequent experience, however, these deviations were found to be trivial.

The methods were re-discussed point by point and several steps were found to require further clarification/investigation. These are listed below.

• A suggestion to reduce the volume of hydrolysis by evaporating the DNA solution in a Speedvac and dissolving it in the solution of buffer/MN/SPD. This implies re-calculation of all the quantities of enzymes used throughout the experiment and investigation of intra- and interindividual reproducibility.
  
• Evaluation of the influence of the number of back washes of butanol extracts with water on the labelling efficiency of ABP–DNA adducts. This will include analysis of the remaining normal nucleotides after each wash.
  
• For the nuclease P1 treatment, a test of the possibility to perform treatment without the addition of pH 5 buffer (especially for PhIP).
  
• Comparison of relative adduct labelling (RAL) for BaP and ABP by labelling adducts and normal nucleotides with a phosphatase-free PNK and PNKs from different sources. Comparison with those obtained by HPLC analysis of normal nucleotides.
  
• A check for dephosphorylation products (on TLC mapping of normal nucleotides) when using PNK with different phosphatase activity.
  

These were to be investigated by the mini-trial group. When full agreement was obtained in the participating laboratories, the trial protocol was revised to incorporate these developments. The results of this second mini-trial are summarized as follows.

• A new protocol using reduced volumes as a consequence of evaporating the DNA solution to dryness prior to digestion was devised and found to give identical results to the previous procedure.
  
• Increasing the number of back washes of butanol extracts above two was liable to reduce the recovery of adducts, even though it removed more of the normal nucleotides, i.e. a 30-fold reduction between back washes 2 and 4.
  
• Omitting the change of pH to 5 for the nuclease P1 digestion was found to reduce recovery of adducts by >50%, so adding buffer to lower the pH for this digestion is essential.
  
• When quantifying normal labelled nucleotides, it is necessary to run a blank using water instead of DNA digest and to subtract the value obtained as background. With some PNKs, this value could be as much as 50% of that of the normal nucleotide count.
  
• The choice of PNK did not greatly affect the RAL value when adduct quantitation was based on labelling of normal nucleotides. However, the adduct levels were over-estimated when the calculation was based on quantitation of normal nucleotides by HPLC.
  
• Some workers have noticed a fast-migrating radioactive spot in the separation of normal nucleotides, which may correspond to dephosphorylation products produced by PNK. This needs to be investigated.
  

Prospects for the second interlaboratory trial
The following samples were planned to be prepared for the second interlaboratory trial: (i) mouse liver DNA from animals treated with two different doses of benzo[a]pyrene; (ii) mouse liver DNA from animals treated with two different doses of 4-aminobiphenyl; (iii) control mouse liver DNA from untreated animals; (iv) an `unknown' DNA sample containing O⁶-MeG.

In this second trial participants were to adhere strictly to the trial protocol. By using the DNA standards from the first trial alongside, results obtained with the new samples could be presented both as absolute measurements of adduct levels and as values normalized to those of the standards. The value to be used for the level of modification of these standards, provided to the participants, were: BaP–DNA, 111 adducts/10⁸ nucleotides; ABP–DNA, 62 adducts/10⁸ nucleotides; O⁶-MeG, 60 adducts/10⁸ nucleotides. When the ABP–DNA sample was analysed by HPLC/MS, a lower estimate of 19 adducts/10⁸ nucleotides was obtained (Beland et al., 1999).

The samples were to be produced using radiolabelled test compounds in order to provide an independent measurement of DNA adduct levels, as before.

The methods investigated were to include the revised methods for the bulky adducts, in particular the pre-tested protocol using small total volumes of reaction mixture, and a further trial of the method for O⁶-MeG taking into account the experience gained by participants in the first phase of the trial, many of whom had been using the method for the first time. In addition, other researchers in Europe and the USA, outside the group currently conducting the ³²P-postlabelling trial, were to analyse these samples by other methods, including immunochemical (ELISA) and physicochemical (mass spectrometry and fluorescence spectroscopy) in order to provide further validation of both the ³²P-postlabelling methods of quantitation and of their own methods.

	Preparation of carcinogen-modified DNA for the second interlaboratory trial

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
Samples of DNA containing bulky adducts were prepared by F.A.Beland as follows.

BaP–DNA.
Four mice each were injected i.p. with 0.5 or 1.0 mg [7,8-³H]BaP in trioctanoin (100 µl). After 24 h, the mice were killed, their livers pooled and nuclei isolated, from which DNA was prepared. Approximately 1 mg was counted after treatment with DNase.

ABP–DNA.
Four mice each were injected with 0.1 or 1.0 mg 4-ABP in trioctanoin (100 µl). DNA was isolated from their livers as described above for BaP–DNA.

O⁶-MeG.
A sample of DNA containing O⁶-MeG was prepared by G.Margison and provided by A.Povey and D.Cooper. It was essentially prepared in a similar way to the synthetic standard, i.e. by reaction of DNA with [³H]MNU.

	Second interlaboratory trial

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
Analysis of bulky adducts.
Each laboratory participating in this part of the study received five samples (two BaP-modified DNA from mouse liver, two ABP-modified DNA from mouse liver and one `unmodified' DNA from liver of untreated mice). In addition, they received one vial of a single batch of PNK from a single source (Epicentre Technologies). Each laboratory was requested to analyse the samples with the provided PNK. In each experiment, analyses were to be carried out in duplicate. Separate experiments were to be carried out on different days.

Analysis of O⁶-MeG.
Each laboratory participating in this part of the study received the `unknown' methylated DNA sample together with immunoaffinity columns and reagents as previously described for the first interlaboratory trial. Participants were requested to carry out three independent determinations.

Analysis of bulky adducts and of O⁶-MeG.
As previously, some laboratories participated in all studies and received all the above-mentioned samples and equipment.

Reporting and analysis
All participants were requested to fill in a form which had been provided at the same time as the samples/equipment and to send their results to the coordinators. These data were collated, qualitatively and quantitatively analysed and distributed at the third workshop.

Qualitative analysis of the results
The chromatograms obtained were similar in appearance to those seen in the first interlaboratory trial.

Control DNA. Of the 19 laboratories that participated in the analysis of bulky adducts, seven produced chromatograms of control DNA with no radioactive spots visible. The remainder obtained varying patterns of very weak spots or background radioactivity. In all cases the patterns were extremely faint and readily distinguishable from the much stronger adduct spots seen in the test samples. Two of the laboratories demonstrated that some of these faint spots were related to source and batch of some of the PNK.

BaP–DNA. The major spot, due to N² guanine substitution by anti-BPDE was seen by all participants, both with the mouse liver DNA samples and with the synthetic standard. One or two participants detected a very minor additional spot in the mouse liver samples. Most participants detected an additional spot or spots (very minor) in the analysis of the standard.

ABP–DNA. In most laboratories a single major adduct spot was observed (in one case this appeared to be a cluster of several spots. In about four cases, additional very minor spots were observed. One participant failed to detect the adduct in the synthetic standard, but not in the mouse liver samples.

O⁶-MeG. The number of participants in this part of the study was small, but there was good qualitative agreement, with most reporting chromatograms which showed the adduct spot due to this lesion. Although separation of the adduct was satisfactory, in many cases problems were encountered achieving adequate resolution and location of the internal standard dIp. A schematic chromatogram is presented in Appendix 2 (Figure A2.1). One way to better locate the internal standard may be to carry out a short autoradiographic exposure to locate the dIp spot, followed by a longer exposure to locate the O⁶-methyldeoxyguanosine (O⁶-MedG) adduct spot.

View larger version (14K): [in this window] [in a new window]   Fig. A2.1. 2-D chromatogram of 32P-postlabelled O6-MepdG

 
Quantitative analysis of the results
Table IV presents the uncorrected and corrected data for the trial protocol. The average uncorrected values per experiment and the average corrected values per experiment are also presented in this table, together with a summary statistical evaluation of the results.

View this table: [in this window] [in a new window]   Table IV. Statistical evaluation of the results of the second interlaboratory trial

 
BaP–DNA. When compared with the first interlaboratory trial, no apparent improvement in the uncorrected results obtained with both the low and high modified samples is noticeable from the coefficients of variation. However, there is distinct improvement in the shape of the distribution curves, with mean and median values much closer than in the first trial. For the corrected values, the improvement in both the shape of the curves and in the statistics is noticed with a reduction of 20% in the coefficient of variation for both the low and high BaP samples. Based on the incorporation of ³H, the levels of adducts in the low and high BaP–DNA mouse liver samples were estimated to be 86.4 and 137.6 adducts/10⁸ nucleotides, respectively. Investigators obtained values with the synthetic standard that were in a range that spanned the expected value (from measurement of ³H incorporation) of 111 adducts/10⁸ nucleotides. However, the average value for the high adducted BaP–DNA sample was 16.1% of the expected value and for the low adducted sample it was 13.2%. However, these values are close to the estimates obtained by mass spectrometric analysis (Pastorelli et al., 1998) of the samples (Table V). It should be noted that the standard was an in vitro sample from the reaction of BPDE with DNA, whereas the test samples were from the livers of mice treated with the parent PAH. Further analysis, which is in progress, will be needed to resolve this issue of the apparent lower labelling efficiency of the in vivo samples. Overall, the second trial achieved a significant improvement in interlaboratory reproducibility in the analysis of BaP-modified DNA.

View this table: [in this window] [in a new window]   Table V. Comparison of the DNA adduct levels obtained by different methods for the samples of the second interlaboratory trial

 
ABP–DNA. When compared with the first interlaboratory trial, a clear improvement in the uncorrected results obtained with both the low and high modified samples is noticeable from the coefficients of variation, which are 18% less than previously. In addition, there is distinct improvement in the shape of the distribution curves, with mean and median values closer than in the first trial. For the corrected values, there was an improvement in both the shape of the curve and in the statistics, with an additional reduction of 20% in the coefficient of variation for both the low and high ABP samples. Based on the incorporation of ³H, the levels of adducts in the low and high ABP–DNA mouse liver samples were estimated to be 70.4 and 444.8 adducts/10⁸ nucleotides, respectively. As had been observed in the previous trial, participants obtained recoveries that were very much lower than these expected values. As the same poor recovery was obtained with the synthetic standard, normalizing to this sample produced higher values for the mouse liver DNA samples. When the ABP–DNA standard is based on the HPLC/MS analysis of Beland et al. (1999) (19 adducts/10⁸ nucleotides), lower values are calculated than when the value of the standard is based on ³H incorporation (62 adducts/10⁸ nucleotides). The lower estimates are reasonably close to the estimates obtained by mass spectrometry (Table V).

As one participant had reported that when PhIP–DNA (analysed in the first interlaboratory trial) was compared using HPLC and TLC, a much better recovery of ³²P-labelled adducts was obtained using HPLC, the inference was that adducts might be lost due to their mobility in the D1 stage of the TLC. To determine whether this might explain the low recovery of the ABP–DNA adducts, the time and conditions for this stage of the procedure were obtained from each participant and investigated for trends. Although some participants used wicks during D1 that were outside the chromatography tanks, there was no significant difference from the values obtained by other workers who used wicks kept inside the tanks. Also, those participants using higher concentrations of phosphate for D1 (2.3 or 1.7 M, compared with 1.0 or 0.9 M) did not obtain higher recoveries. Also, in one laboratory in which TLC and HPLC were compared for the recovery of adducts in the high ABP sample, similar levels were obtained by both methods. These results, therefore, do not provide an explanation for the overall low recovery in these trials. Despite these outstanding issues, the second trial represented an improvement in interlaboratory reproducibility over the first trial. In addition, pending resolution of some discrepancies with the ABP–DNA standard (Beland et al., 1999), the values compare well with those obtained by mass spectrometry (Table V) using the method of Martoni et al. (1998), which in turn are very close to the values obtained by Beland et al. by mass spectrometry.

O⁶-MeG. Only a limited number of results were obtained with this assay and some participants experienced problems with either the chromatographic separation or the new batch of immunoaffinity columns. In addition, some participants carried out all their analyses on the same day and could therefore only provide a single (average) corrected value. From the available data (n = 10 + an outlier) the mean corrected value is 206.7 adducts/10⁸ nucleotides and the SD is 76.45 (coefficient of variation 37%). This result is an improvement on the outcome of the first interlaboratory trial (coefficient of variation 62 or 106%; Table III). However, several laboratories could not meet the objective of producing results, which means that further effort towards standardization is necessary.

The level of modification of O⁶-MeG in the unknown sample, calculated from ³H incorporation, was reported to be 134 adducts/10⁸ nucleotides. Thus the mean value obtained in the trial is 50% higher than the expected value.

	Conclusions and perspectives

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
The primary objectives of this project were to standardize the ³²P-postlabelling assay and to improve interlaboratory reproducibility. These goals have been achieved in large measure through the conducting of two major interlaboratory trials of test samples and two smaller trials investigating detailed methodological conditions. After the second interlaboratory trial, there was an improved distribution of results for adducts of an aromatic amine (ABP) and a 40% reduction in variability compared with the first interlaboratory trial. For adducts of a PAH (BaP) there was also an improvement in distribution of results and at least a 20% reduction in variability. For the measurement of O⁶-MeG, an improvement in the coefficient of variation of 25% (or 70% if 0 values from the first trial are included) was achieved. These calculations were based on values corrected to those obtained with synthetic standards.

Thus, not only was a reduction in variability achieved, but the degree of variability compares favourably with that achieved in other interlaboratory trials, such as that reported by Friesen and Garren (1982). In that study, the level of contamination of peanut and corn meal with aflatoxins was determined in 121 laboratories. When participants analysed individual aflatoxins by their own methods, the coefficients of variation ranged from 44 to 93%, depending on the toxin and the matrix; these values can be compared with those of the first postlabelling trial with investigators' own methods (Table IIB). When investigators in the aflatoxin trial used three different test protocols, the coefficients of variation were 41–79, 51–136 and 31–76%, respectively; these values compare well with the results of the second postlabelling trial given in Table IV.

In addition, this enterprise has resulted in the production of a series of DNA adduct standards which will be of use to the scientific community. However, there are still some uncertainties about the true levels of adducts in these materials and further investigations will continue through collaboration of the trial participants and others (Beland et al., 1999). Although there are discrepancies between estimates by ³H incorporation and by ³²P-postlabelling, this is part of a wider issue which will require the application of other methods for adduct measurement. Such studies have been initiated. Preliminary mass spectrometry results obtained with the ABP–DNA standard were also lower than the ³H incorporation values. Similarly, mass spectrometric analysis of the BaP–DNA standard and mouse liver samples gave lower values than obtained by ³H incorporation, but the first estimates by MS of the O⁶-MeG-containing DNA standard was 4-fold higher; however, the corrected value of the second trial sample is close to that obtained by HPLC. Other methods currently being applied include immunochemical techniques, HPLC/fluorescence, ³⁵S-postlabelling and HPLC of DNA hydrolysates. This is the first occasion on which many different methods of DNA adduct determination have been compared in a thorough and systematic approach. The outcome of this exercise will be a much clearer understanding not only of ³²P-postlabelling, but of all methods currently available.

The recommended protocol for bulky carcinogen–DNA adducts is given in Appendix 1 and that for O⁶-MeG, based on the method of Povey and Cooper (1995), is given in Appendix 2. Investigators wishing to enquire about the availability of materials for analysis of O⁶-MeG should contact A.Povey (apovey{at}fs1.scg.man.ac.uk). It should be borne in mind that the method for bulky adducts was developed from trials of a PAH (BaP) adduct and an aromatic amine (ABP) adduct. Optimum conditions for other classes of adducts may well differ, as suggested by the failure in these trials to obtain reproducible results with PhIP–DNA adducts. Nevertheless, the present protocol should form a starting point for method development. Investigators who develop a different ³²P-postlabelling assay for bulky adducts should verify that they obtain comparable results with their own method to those obtained with the recommended protocol and should examine carefully and critically the consequences of any variations introduced.

	Appendix 1. Protocol for 32P-postlabelling anaysis of bulky DNA adducts

Top Abstract Introduction Chronology Design of the trial... Preparation of DNA adduct... First interlaboratory trial Preparation of carcinogen... Second interlaboratory trial Conclusions and perspectives Appendix 1. Protocol for... Appendix 2. Protocol for... References

 
The original description of the ³²P-postlabelling method for bulky DNA adducts is by Gupta et al. (1982).

I. Dissolution and Storage of DNA
DNA sample should be dissolved in 0.15 mmol/l sodium citrate and 1.5 mmol/l sodium chloride, which do not interfere with further analysis, at a concentration of 2 µg/µl [Note. The presence of EDTA, even at 0.1 mmol/l, may generate problems during hydrolysis/labelling.]

The dissolution will be preferentially performed in glass tubes. Plastic tubes may result in DNA damage and adduct losses. If the latter are used, they should be washed with water and ethanol, then dried before use. Tubes with scratches must not be used.

Before storage, DNA will be separated into small aliquots since it has been observed that repeated freezing and thawing of DNA solutions tends to lead to loss of adducts.

Storage must be performed at –70°C or colder.

II. DNA purity and quantitation
To test the purity of DNA and to quantitate it, two complementary methods may be used.

Before hydrolysis. A full spectrum should be taken between 220 and 320 nm (using, if possible, two different spectrophotometers). The maximum absorption should be at 258–259 nm for DNA not contaminated with RNA. A shift of this maximum towards 250 nm indicates contamination with RNA. A shoulder at 280 nm indicates contamination with proteins. In either case, DNA must be re-purified before proceeding with the hydrolysis.

After hydrolysis. Two methods can be used which will allow: (i) quantitation of normal nucleotides; (ii) a check for RNA contamination. These are: labelling of an aliquot of the digest and TLC separation of normal nucleotides; HPLC analysis of an aliquot of the digest. The former method has been selected for the following reasons.

• It allows for checking of protein contamination (a tailing spot at the origin is an indication of protein contamination and will generate an increased background on the 2-dimensional TLC separation of the adducts).
  
• It allows for correction of the efficiency of hydrolysis.
  
• It allows for correction of batch-to-batch variability in the activity of the ATP provided by the supplier.
  

[Note. This does not imply that HPLC is not a suitable method for DNA quantitation. Postlabelling was only selected by the group in view of the time required for HPLC analysis. Another advantage of postlabelling is that it removes the need for accurate determination of the specific activity of ATP.]

III. DNA hydrolysis
Where possible, a reference DNA sample of known level of modification should be analysed alongside the test samples to allow for correction of values (see Section XI). [Note. The hydrolysate prepared at this stage is to be used for both the analysis of adducts (Sections IV–VIII) and of normal nucleotides (Section IX).]

Reagents. Micrococcal nuclease (from Staphylococcus aureus; Sigma N3755); phosphodiesterase (from calf spleen; Boehringer 108251). [Note. Both MN and SPD solutions must be dialysed (MN, because it may contain oligonucleotides which can interfere with analysis of normal nucleotides; SPD, because it is in solution in (NH₄)₂SO₄ which will inhibit the labelling). Dialysis is performed for 3x3 h in milliQ water or equivalent at +4°C. When dialysing, no air should be left in the tube, otherwise this leads to a change in volume and concentration.]

• Dissolve MN in water at a concentration of ~450 U/ml. After dialysis adjust to 300 U/ml.
  
• Dialyse SPD solution at 4 U/ml. Adjust final concentration to 2.5 U/ml.
  
• Buffer solution, pH 6: Na succinate, 50 mmol/l; CaCl₂, 12.5 mmol/l.
  

[Note 1. Solutions of enzymes can be stored separately or as ready-to-use mixture, at ~20°C and in small aliquots to avoid freezing and thawing problems.] [Note 2. If SPD is used beyond the expiration date (3 months), concentration can be increased by ~30–50% by increasing the volume of SPD added and reducing the volume of water.]

Procedure.

• Take an equivalent of 10 µg DNA solution and evaporate to dryness in a Speedvac evaporator.
  
• Add: 1.9 µl MN solution (final concentration, 60 mU/µl); 3.8 µl SPD solution (final concentration, 1 mU/µl); 3.8 µl buffer (final concentration of Na succinate, 20 mmol/l; final concentration of CaCl₂, 5 mmol/l).
  
• Shake on a vortex shaker to dissolve DNA and allow to react for 4 h at 37°C.
  
• Remove 2 µl for dilution and analysis of normal nucleotides (see Section IX).
  

[Note 1. Digestion time can be adapted to tissue DNA and type of adducts looked at. Long digestion time can lead to losses of adducts.] [Note 2. The DNA digest can be stored frozen overnight as such for subsequent butanol extraction and labelling on the following day or after treatment with nuclease P1 and stopping of the reaction with the Tris base solution (see next section).]

IV. Nuclease P1 enrichment procedure
Reagents. Sodium acetate buffer, pH 5, 0.42 mol/l; ZnCl₂, 2.1 mmol/l; nuclease P1, 1 mg/305 µl sodium acetate 10 mmol/l; Tris base solution (unbuffered), 0.92 mol/l.

Procedure.

• To the remaining 7.5 µl of hydrolysate add: 1 µl sodium acetate buffer (final concentration 40 mmol/l); 1 µl ZnCl₂ solution (final concentration 0.2 mmol/l); 1 µl nuclease P1 solution (final concentration 0.31 µg/µl).
  
• Allow to react for 30 min at 37°C, then stop the reaction by addition of 1 µl Tris solution.
  

V. Butanol enrichment procedure
Reagents. TBA chloride solution 10 mmol/l; ammonium formate solution, pH 3.5, 11.8 mmol/l; Tris–HCl, pH 9.5, 0.66 mol/l; butanol (redistilled, saturated with water just before use)

Procedure.

• To the remaining 7.5 µl of hydrolysate, add 212 µl of ammonium formate solution and 25 µl TBA chloride solution and extract immediately with 250 µl butanol (saturated with water) by shaking on a vortex shaker (full speed) for at least 1 min.
  
• Centrifuge to separate layers and collect top layer (butanol).
  
• Extract a second time with 250 µl butanol (saturated with water), centrifuge, collect butanol and pool with previous extract.
  
• To pooled butanol extracts, add 400 µl water (saturated with butanol) and shake for at least 1 min in a vortex shaker (full speed).
  
• Centrifuge to separate layers and discard aqueous washes (bottom layer).
  
• Repeat washing of the butanol phase with 400 µl water (saturated with butanol).
  
• Add 1 µl of Tris–HCl solution to butanol layer.
  
• Evaporate butanol to dryness in a Speedvac at room temperature.
  
• Take residue in 100 µl butanol by vortex shaking, evaporate to dryness again, then take up in 11.5 µl water by vortex shaking.
  

[Note 1. The extractions and back extractions can be performed at room temperature.] [Note 2. The samples can be labelled immediately or stored at –80°C in the butanol phase or after transfer into water.]

VI. Labelling of the adducts
Reagents. Polynucleotide kinase (PNK) (with or without phosphatase activity). Buffer solution: bicine, 180 mmol/l; magnesium chloride, 90 mmol/l; DTT, 90 mmol/l; spermidine, 4.5 mmol/l, pH 9 (when using PNK with phosphatase activity) [Note. Bicine buffer can be prepared in advance but should be stored as small aliquots at –80°C as repeated freezing and thawing and storage at temperatures >–80°C can cause DTT decomposition and loss of labelling efficiency]. ATP, home-made or from a supplier. [Note 1. If ATP is home-made, ³²Pi should be obtained in water and not in HCl solution. When it is delivered in HCl solution, the HCl content varies from batch to batch and this can affect the efficiency of the ATP synthesis.] [Note 2. Whether home-made or purchased, the activity of the ATP must be verified.]

Procedure.

• To 11.5 µl solution from the nuclease P1 digestion or butanol extraction enrichment mix, add 2 µl of bicine buffer solution and 4.5 µl of a mix containing 54 pmol of ATP and 9.0 U PNK. [Note. The final concentrations of the reagents will be as follows: bicine, 20 mmol/l; MgCl₂, 10 mmol/l; DTT, 10 mmol/l; spermidine, 0.5 mmol/l; PNK, 0.5 U/µl; ATP, 3 µmol/l (in case there is very high activity in ATP, cold ATP may be added).]
  
• Allow to react for 30 min at 37°C. [Note. Apyrase should not be added to stop the experiment.]
  
• Spot the whole sample (i.e. 18 µl) onto a TLC sheet.
  

VII. Test for efficiency of nuclease P1 or butanol enrichment techniques

• Wash the bottom of the tube with ~50 µl water.
  
• Shake for 30 s in a vortex shaker and centrifuge to ensure no contamination of the lid.
  
• Spot ~5 µl on a PEI–cellulose TLC plate.
  
• Chromatograph in 250 mM (NH₄)₂SO₄, 40 mM NaH₂PO₄.
  

[Note 1. The phosphate content of the chromatography solvent can be adjusted for batch-to-batch variability of the TLC plates in order to allow separation of dT bisphosphate and ³²Pi.] [Note 2. Poor efficiency of nuclease P1 treatment or butanol enrichment will be demonstrated by appearance on the autoradiogram of the spots of the four normal nucleotides after ~20–30 min exposure. In the case of absence of the ATP spot, the sample should be discarded (see Fig. A1.1).]

View larger version (8K): [in this window] [in a new window]   Fig. A1.1. Schematic representation of migration of normal nucleotides on PEI–cellulose developed in 250 mM (NH4)2SO4, 40 mM NaH2PO4.

 
VIII. TLC separation of adducted nucleotides

• Spot the entire sample on the TLC plate for chromatography clean-up of the adducts in D1.
  
• Prewashing the TLC plates is not essential. It may be performed to remove the yellow colour from some plates, a colour which will increase the background, especially at the solvent front. [Note. Two types of washes are currently performed. (i) Overnight chromatography with deonised water either in an open TLC tank or in a closed tank overnight followed by at least 1 h with the lid removed; (ii) washes with water, then with methanol, in a shaking bath.]
  
• D1. It is recommended to test a range of phosphate concentrations to ensure that adducts remain at the origin. [Note 1. For bulky adducts 1–1.7 M will be enough; for smaller adducts >2 M may be necessary. Up to 2.8 M can be used to retain polar adducts (lowering the pH will be necessary). For information on a possible approach, see Laws et al. (1994).] [Note 2. It is recommended to perform an autoradiography after D1 to ensure quality of the clean-up. Presence of a tailing peak at the origin indicates contamination of the sample with proteins which will increase the background in D2 and D3 (former D3 and D4, respectively).] [Note 3. Wash the plate in deionised water after chromatography by shaking gently for ~5 min each in two successive baths. Then allow to dry.]
  
• Former D2. This chromatography step is not required.
  
• D2 and D3 (formerly D3 and D4). The solvents should be adjusted to result in adequate mobility of the spots over the TLC plate. [Note 1. If transfer chromatography from D1 to D2 (formerly D3) is used, the capacity of the donor and the acceptor plates must be equivalent. If the donor is thicker than the acceptor, this can cause streaking. If the reverse, this can cause spot compression.] [Note 2. When using Li formate in D2 (formerly D3), it should be prepared as follows. For, for example, 3 mol/l Li formate, pH 3.5, use 3 mol/l formic acid and adjust to pH 3.5 with lithium hydroxide.] [Note 3. To avoid the problems