Benzo(a)pyrene diolepoxide-DNA adducts detected by synchronous fluorescence spectrophotometry.

Using benzo(a)pyrene (BP) as a model carcinogen we are currently applying a fluorescence technique to detect the very low levels of carcinogen-DNA adducts in human populations due to environmental exposure. In synchronous fluorescence spectrophotometry for detection of BP-diol epoxide-DNA, excitation and emission wavelengths are scanned simultaneously with a fixed wavelength difference (delta lambda) of 34 nm. Compared to conventional fluorescence methods only one peak emerges because excitation and emission peaks have to match delta lambda to show. Because of the quenching effect of DNA, samples are hydrolyzed by acid. After this, BP-diol epoxide (BPDE)- -modified DNA gives a peak at the same wavelength and of the same fluorescence yield as BP-tetrols. When DNA from peripheral blood lymphocytes of 44 coke oven workers were analyzed, 10 had a sharp peak at 379. Among 36 coke oven workers from another factory, 4 had detectable levels of adducts. A much smaller percentage of samples was positive in a group of aluminum plant workers. We have also found BPDE-DNA adducts in DNA from pulmonary alveolar macrophages and peripheral blood lymphocytes from tobacco smokers and some of the nonsmokers.


Introduction
Because polycyclic aromatic hydrocarbons (PAH) have highly fluorescent aromatic nucleus, fluorescence has been used to study the structural features of PAH-DNA complexes (1)(2)(3)(4). This has also led to the idea of biologic monitoring of PAH exposure by measuring PAH-DNA adducts in human tissues by fluorescence (5) ( Table 1). The main difficulty is the low level of environmental exposure compared to animal experiments (6), leading to minute amounts of DNA adducts formed. Highly sensitive assays are thus needed.
The first important improvement to the conventional fluorescence method was the photon-counting device developed by Vigny and Duquesne (7). It increased the signal-to-noise ratio so that very weak fluorescence could  be detected, for instance one 7-bromomethylbenz(a)anthracene molecule could be detected in 20,000 nucleotides (7). Another improvement was synchronous fluorescence spectrophotometry (SFS). The principle was presented by Lloyd (8) and developed further by Vo-Dinh (9,10). The main application of the method has been in analytical chemistry in differentiating between PAHs in oil mixtures (10,11). Conventional excitation and emission spectra are often complicated, having many peaks. By scanning excitation and emission synchronously with a fixed wavelength difference (A X), in some cases only one peak emerges. This is because a signal is seen only when A X matches the interval between one absorption and one emission band (9). Low temperature can increase the fluorescence yield and has also been used in detection of carcinogen-DNA adducts; Ivanovic et al. (12) were able to detect one benzo(a)pyrene (BP) EXCITATION 300 350 residue per 50,000 nucleotides with conventional fluorescence spectrophotometry after freezing the sample in 1:1 ethylene glycol at 77°K. Low temperature with synchronous scanning improves the sensitivity by the order of magnitude compared to low temperature alone (13). The fluorometric HPLC assay by Rahn et al. (14) is sensitive enough to detect BP at 1 BP residue per 107 nucleotides.

Results and Discussion
Synchronous scanning is possible, ifthe excitation and emission monochromators can be locked together and then scanned with a fixed wavelength difference (A X). Our first experiments were done with Perkin-Elmer 44B fluorescence spectrophotometer attached to an Ortec photon counter ( Fig. 1) (15). For the synchronous scanning we used AX 34 nm (Fig. 2). The correlation curve between the fluorescence yield at 379 nm of emission of in vitro-modified (BPDE)-DNA (± +)-r7,t8-dihydroxy-t9, 10-epoxy-7,8,9, 10-tetrahydrobenzo(a)pyrene and the concentration of BP moieties in the solution was linear (Fig. 2). The minimum amount we could detect was about 200 fmole/mL. From some ofthe human DNA samples we could detect the same peak with AX 34 nm as from BPDE-DNA (Fig. 3).
The main part of the studies was done with a Perkin-Elmer fluorescence spectrophotometer 650-40 with a Perkin-Elmer 3600 data station (16). This machine gives more accuracy, the possibility of three-dimensional scanning, and mathematical manipulation of the curves, e.g., derivative spectroscopy, and higher sensitivity. The linear correlation curve between the fluorescence intensity and concentration of BP moieties was confirmed. Because DNA quenches fluorescence of BP molecules covalently bound to it (2,17), we wanted to compare the fluorescence yields of (r7,t8,t9,clO-tetra- hydroxy-7,8,9, 10-tetrahydrobenzo(a)pyrene (BP-tetrol) to that of BPDE-DNA. The differences we saw between (BP-tetro) and BPDE-DNA were close to those reported by Rahn et al. (14), BP-tetrol being more fluorescent than BPDE-DNA (Fig. 4). Because of this strong effect and possibly different fluorescence yields of different adducts (18), the quantitation seemed impossible without total hydrolysis of the BPDE-DNA after the isolation and purification of DNA. This seemed to be best gained by hydrolyzing in HCl at 90°C (14,16). Three hours at 90°C in 0.1 M HCI was enough to increase the fluorescence yields of BPDE-DNA to the same level with BP-tetrol (Fig. 4). We are currently applying this fluorescence assay and enzyme immunoassays to study DNA isolated from peripheral blood lymphocytes (PBL) from coke oven workers and aluminum plant workers and also to compare DNA from alveolar macrophages and PBLs from smokers to that from nonsmokers. There are some earlier encouraging reports where immunoassays have been used (19,20).
In one sample of 44 coke oven workers (21), we found about 65% positives at 380 nm with A A 34 nm by SFS. Of the 10 persons who had a sharp peak at 379 to 380 nm, either after 1 hr or 3 hr heating at 90°C in 0.1 M HCI, one was a nonsmoker, three were ex-smokers, and six were smokers. We had enough DNA from some of the coke oven workers to measure adducts by USERIA (ultrasensitive enzymatic radioimmunoassay) (22,23). Some of the samples positive by SFS were negative by USERIA and vice versa. Although the reasons for the discrepancies are not totally clear yet, we might speculate that they could be explained in part by the fact that USERIA with polyclonal rabbit antiserum detects BPDE-DNA, whereas BP-GUA and BP-tetrols are poorly detected by this antiserum (23), and some amount of hydrolysis is already caused by DNA isolation and more by storage. By both of the methods smokers had the highest mean adduct levels. In another sample of 36 coke oven workers, four had a sharp peak at 379 nm after HCl treatment (unpublished results). Of 30 aluminum plant workers only one of the smokers was positive (16). Pulmonary alveolar macrophage and peripheral blood lymphocyte DNAs from both some of the healthy smokers and nonsmokers show a peak implying BPDE-DNA adducts (unpublished results).

Conclusion and Perspectives
We are able to detect less than 1 BP moiety per 107 nucleotides in in vitro modified BPDE-DNA by fluorescence using synchronous scanning with AX 34 nm after hydrolyzing the DNA in 0.1 M HCI at 900C. According to preliminary results, the method should be usable for biologic monitoring in exposed humans, but verification of the quantitation and specificity of the method need in vivo animal experiments.
Further developments of the approach include derivative spectroscopy, which is a routine, powerful, and useful technique to aid in resolution of complex spectra (24). An interesting method is three-dimensional scanning (Fig. 5) by which the best AX for new adducts/ compounds might be found. A contour map of the threedimensional figure makes it possible to find sites of peaks with great accuracy and might provide a kind of "finger print" in complicated cases (Harris and