Improvement of an Automatic HPLC System for Nitropolycyclic Aromatic Hydrocarbons: Removal of an Interfering Peak and Increase in the Number of Analytes

全文

(1)

Introduction

Airborne particulates contain not only carcinogenic polycyclic aromatic hydrocarbons (PAHs), such as benzo[a]pyrene (BaP),1 but also nitropolycyclic aromatic hydrocarbons (NPAHs), such as dinitropyrenes, whose mutagenicity is much stronger.2,3 However, the atmospheric concentrations of NPAHs are much lower than those of PAHs. Therefore, an ultrasensitive analytical method is necessary to determine atmospheric NPAHs. Two highly sensitive methods for NPAHs are high- performance liquid chromatography with fluorescence detection (HPLC/FLD) and gas chromatography with mass spectrometry (GC/MS).4–6 However, the sensitivities of these methods are not high enough to detect NPAHs that typically occur at low concentrations, such as the very strongly mutagenic 1,3-, 1,6- and 1,8-DNPs.

We previously developed an HPLC method with chemiluminescence detection (HPLC/CLD) for NPAHs.7 The detection limits were two orders of magnitude lower than those by HPLC/FLD or GC/MS. Utilizing this method, we showed that automobiles, especially diesel-engine vehicles, were the main source of atmospheric 1,3-, 1,6-, 1,8-DNPs and 1-NP in major commercial cities in Japan,8–11and that the contribution of coal combustion for domestic heating, power plants and industries might be larger in Vladivostok.12 We also showed that several NPAHs, such as 2-nitropyrene (NP) and 2- nitrofluoranthene, were formed in the atmosphere.13,14

However, the analyses required tedious preparation procedures including cleaning up the NPAH samples and chemically reducing them to the corresponding aminopolycyclic aromatic hydrocarbons (APAHs). To simplify the NPAH analyses, we developed an automatic HPLC system by adding on-line clean-up, a reducer and concentrator columns.15 This system reduced the analysis time by about 2 h, which were previously required for the clean-up and reduction processes.

However, this system detected only four kinds of NPAHs (1,3-, 1,6-, 1,8-DNPs and 1-NP), which were responsible for about one-third of the total direct-acting mutagenicity of diesel-engine exhaust particulate extracts. Moreover, a peak originating from an ascorbic acid solution used in the on-line concentration process interfered with the quantitative analysis of trace levels of 1,6-DNP. In this study, an ODS column was introduced just after the pump for the ascorbic acid solution to remove the interfering peak. Also, the conditions for loading the reduced fraction of NPAHs were optimized so as to increase the number of detectable analytes.

Materials and Methods

Chemicals

1,3-, 1,6-, 1,8-DNPs, 1-, 4-NPs, 6-nitrochrysene (6-NC), 7- nitrobenz[a]anthracene (7-NBaA), 3-nitroperylene (3-NPer) and 2-fluoro-7-nitrofluorene (FNF, internal standard) were purchased from Aldrich Chemical Company (Milwaukee, WI, USA). 1-NPer, 2- and 3-nitrofluoranthenes (NFRs) were purchased from Chiron AS (Trondheim, Norway), ChemSyn Laboratories (Kansas, USA) and Wako Pure Chemical 2003 © The Japan Society for Analytical Chemistry

Improvement of an Automatic HPLC System for Nitropolycyclic Aromatic Hydrocarbons: Removal of an Interfering Peak and Increase in the Number of Analytes

Ning T

ANG

,* Akira T

ORIBA

,** Ryoichi K

IZU

,*

,

** and Kazuichi H

AYAKAWA

*

,

**

*Graduate School of Natural Science and Technology, Kanazawa University, 13-1, Takara-machi, Kanazawa 920–0934, Japan

**Pharmaceutical Sciences, Kanazawa University, 13-1, Takara-machi, Kanazawa 920–0934, Japan

An automatic HPLC system for analyzing nitropolycyclic aromatic hydrocarbons (NPAHs, nitroarenes) in airborne particulates was previously described (Anal. Chim. Acta, 2001, 445, 20). Some problems with this system were that it generated a peak originating from an ascorbic acid solution that elutes at a retention time close to that of 1,6- dinitropyrene (DNP), and that it was able to analyze only 1,3-, 1,6-, 1,8-DNPs and 1-nitropyrene (1-NP). Here, we describe an improved system that effectively removes the interfering peak by introducing an ODS column just after the pump for the ascorbic acid solution, and which is capable of analyzing several additional compounds (2-, 4-NPs, 2-nitrofluorene, 6-nitrochrysene, 7-nitrobenz[a]anthracene, 3-nitroperylene and 6-nitrobenzo[a]pyrene etc.). The improved sensitivities were achieved by concentrating the compounds in a benzene–ethanol extract from airborne particulates, by increasing the loading time of the sample solution from 20 to 38 min, and by increasing the flow rate of an ascorbic acid solution from 1.3 to 1.8 mL/min.

(Received September 25, 2002; Accepted November 27, 2002)

To whom correspondence should be addressed.

E-mail: hayakawa@p.kanazawa-u.ac.jp

(2)

Industries (Osaka, Japan), respectively. 2-NP was kindly provided by Prof. A. Hirayama of the Laboratory of Public Health, Kyoto Pharmaceutical University. Figure 1 shows the structures of these NPAHs. All other chemicals used were obtained from commercial sources.15

HPLC system

Figure 2 shows a schematic diagram of the improved HPLC system, in which the dashed-line box indicates the part newly added to the previous system. The system consisted of four Shimadzu (Kyoto, Japan) LC-10A pumps (pump1 – 4), one Sanuki (Tokyo, Japan) DMX-2000 pump with two heads (pump 5), a Shimadzu SIL-10A auto sample injector, a Shimadzu DGU-14 degasser, a Shimadzu CLD-10A chemiluminescence detector, a Shimadzu SCL-10A system controller and a Shimadzu C-R4A integrator. The clean-up column (4.6 i.d. × 150 mm), concentrator column (4.6 i.d. × 30 mm), separator column (4.6 i.d. ×250 mm), guard columns 1 (4.6 i.d. ×30 mm) and 2 (4.6 i.d. ×50 mm) were packed with Cosmosil 5C18-MS (Nacalai Tesque, Kyoto, Japan), and the reducer column (4.0

i.d. ×10 mm) was a Nitroarene Reactor Column (Shimadzu).

The reducer column was kept at 80˚C in an HIC-6A column oven (Shimadzu) and guard columns 1 and 2, the clean-up, the concentrator and the separation columns were kept at 20˚C in a CTO-10AC column oven (Shimadzu). All other conditions were the same as those given in our previous report.15

The HPLC system was operated as follows. After the sample solution (airborne particulate extracts) was introduced into the HPLC system by the auto sample injector, the NPAHs were separated from interfering substances on the clean-up column.

The interfering substances which elute faster from the clean-up column were discarded. Then, by changing the position of the switching valve (from the solid line to the dotted line in Fig. 2), the APAHs, which were produced by passing the NPAHs through the reducer column, were concentrated on the concentrator column by adding the ascorbic acid solution from pump 2 to the effluent from the reducer column to increase the water concentration. After all analytes (APAHs) were eluted from the clean-up column and concentrated on the concentrator column completely, the APAHs were eluted into the separator Fig. 1 Structures of the analyzed NPAHs. For abbreviations, see text.

Fig. 2 Schematic diagram of the proposed NPAH analysis system.

(3)

column with a mixture of imidazole buffer (from pump 3) and acetonitrile (from pump 4) by changing the position of the switching valve (from the dotted line to the solid line in Fig. 2).

Then, the APAHs are separated on the separator column and detected by chemiluminescence.

Sampling and pretreatment of airborne particulates

Airborne particulates were collected by a 123VL high-volume air sampler (Kimoto Electric Company Ltd., Osaka, Japan) with a 2500QAT-UP quartz fiber filter (8″ ×11″, Pallflex Products, Putnam, CT, USA) set on a sidewalk 1 m from a heavy traffic road in Kanazawa, Japan for 24 h at a flow rate of 1.3 m3/min.

Thirty square centimeters of the filter (containing about 4 mg airborne particulates) were cut into small pieces and placed in a flask. After adding an FNF solution as an internal standard, NPAHs were extracted ultrasonically twice with benzene/ethanol (3:1, v/v); the solution was then filtered with No. 6 filter paper (Toyo Roshi Kaisha, Ltd., Tokyo, Japan) and a 0.45 µm HLC-Disk membrane filter (Kanto Chemical Co., Inc., Tokyo, Japan). The filtrate was washed with a sodium hydroxide solution, a sulfuric acid solution and water, and was then evaporated to dryness. The residue was dissolved in 1 mL of 75% ethanol–0.02 M acetic acid–sodium acetate buffer (pH 5.5). This sample solution was filtered with a 0.45 µm membrane filter again, and an aliquot of this solution was injected into the HPLC system.

Results and Discussion

Improvement of the HPLC system

When crude benzene–ethanol extracts were injected directly into the reducer column of the previous HPLC system, several interfering peaks were observed around the retention times of the three DNPs. Figure 3 shows chromatograms of benzene–ethanol extracts from airborne particulates collected in Kanazawa. A large peak (retention time of 7 min) was observed without guard column 2 (Fig. 3A). This peak interfered with the quantification of trace levels of 1,6-DNP (retention time of 8 min). The height of the interfering peak depended on the loading time of the eluate from the reducer column onto the concentrator column, suggesting that the interfering substance originated from the ascorbic acid solution. In order to remove this interfering peak, several kinds of columns were placed just after the ascorbic acid solution pump (Fig. 2). As a result, the interfering peak was effectively removed by a Cosmosil 5C18-

MS (4.6 i.d. ×50 mm) and the determination of trace levels of 1,6-DNP became possible (Fig. 3B). This column was effective for over 30 days.

The 15 NPAHs, including the internal standard (FNF), eluted from the reducer cokumn between 20.5 and 56 min (Fig. 4).

FNF was the first to elute at 20.5 – 24 min and 3-NPer was the last at 52 – 56 min. In the previous HPLC system,15the NPAHs eluted at 20 – 40 min. These included not only 1,3-, 1,6-, 1,8- DNPs, 1-NP and FNF, but also 2-NF, 2-, 4-NPs, 2-, 3-NFRs, 7- NBaA and 1-NPer. However, 6-NC, 3-NPers and 6-NBaP, bound more strongly to the reducer column, and little if any of these compounds entered the concentrator column. Therefore, to analyze these NPAHs simultaneously, the loading time of the new system was set from 20 to 58 min. When the eluate from the reducer column after 58 min was introduced into the concentrator column, there was no peak (see Fig. 6C). Thus, this range (20 – 58 min) was able to cover not only these 15 NPAHs, but also other NPAHs having 3 to 5 rings. On the other hand, some peaks were observed around the retention times of the three DNPs when eluting from the reducer column before 20 min was introduced into the concentrator column.

However, these peaks were also observed without the reducer column, suggesting that these substances were not NPAHs.

The volume of the eluate from the reducer column that was introduced into the concentrator column was increased from 4.0 to 7.6 mL, suggesting that it was necessary to increase the water concentration in the concentrator column to concentrate the 15 NPAHs completely. Figure 5 shows the relative peak heights of three NPAHs as a function of the flow rate of the ascorbic acid solution when the loading time was set from 20 to 58 min. The peak height of 3-NPer, which had the strongest retention, was constant at different flow rates of the ascorbic acid solution.

However, when the flow rate was lower than 1.4 mL/min, the peak heights of 1,6-DNP and FNF, whose retentions were weaker than that of 3-NPer, became smaller. The peak heights of 1,6-DNP and FNF were highest when the flow rate was 1.8 mL/min. Thus, the flow rate of the ascorbic acid solution was set at 1.8 mL/min in a following experiment.

Fig. 3 Chromatograms of extracts from airborne particulates by the proposed HPLC system with/without guard column 2 (ODS, 4.6 i.d.

×50 mm). A, without guard column 2; B, with guard column 2.

Fig. 4 Elution times of NPAHs from the reducer column. For HPLC conditions, see text.

(4)

Accuracy

The NPAHs targeted for analysis in this study were 1,3-, 1,6-, 1,8-DNPs, 2-NF, 1-, 2-, 4-NPs, 6-NC, 7-NBaA, 6-NBaP and 3- NPer, which were major NPAHs in airborne particulate extracts.

The detection limits, quantitation limits, ranges of the calibration curves and relative standard deviations (n = 3) of the new system were investigated. As shown in Table 1, when the sample injection volume was 100 µL, the detection limits (S/N = 3) were 1 fmol for the DNPs, 10 fmol for 1-NP, 150 fmol for 2- NP, 30 fmol for 4-NP, 1000 fmol for 2-NF, 150 fmol for 6-NC, 30 fmol for 7-NBaA, 10 fmol for 3-NPer and 15 fmol for 6- NBaP. The relative standard deviations (n = 3) were less than 5%. The calibration curves were linear (r2> 0.995) from 3 to 100 fmol DNPs, from 30 to 1000 fmol 1-NP, from 500 to 3000 fmol 2-NP, from 100 to 3000 fmol 4-NP, from 3000 to 20000 fmol 2-NF, from 500 to 3000 fmol 6-NC, from 100 to 1000 fmol 7-NBaA, from 30 to 1000 fmol 3-NPer and from 50 to 150 fmol 6-NBaP, respectively. The detection limits of DNPs and 1-NP by the new system were at the same levels of those by the previous system.

Application to airborne particulates

Figure 6 shows chromatograms of benzene–ethanol extracts from airborne particulates collected in Kanazawa. The peaks of DNPs, 1-, 2-, 4-NPs, 2-NF and 7-NBaA detected in chromatogram (A) using the previous loading time (20 – 40 min) were as high as those in chromatogram (B) using the present loading time (20 – 58 min). However, the peak height of

6-NC was lower, and neither 3-NPer nor 6-NBaP was detected in chromatogram (A) using the previous loading time (20 – 40 min). Thus, the new system appeared to be more efficient than the proposed system.

The atmospheric concentrations of 1,3-, 1,6-, 1,8-DNPs, 1-, 2-, 4-NPs, 2-NF, 6-NC, 7-NBaA, 3-NPer and 6-NBaP, calculated from chromatogram (B), were 2, 2, 2, 320, 730, 3, 610, 100, 63, 27 and 45 fmol/m3, respectively. In addition, several unknown peaks were observed in chromatogram (B). It is well known that peroxyoxalate-chemiluminescence detection is highly sensitive and selective for APAHs. Therefore, these peaks, which were detected only after on-line reduction, might have originated from NPAHs. The heights of five unknown peaks (a, b, c, d and e) detected in chromatogram (A) were the same as those in (B), suggesting that the retention times of these compounds on the concentrator column were in the range of 20 to 40 min. The heights of the unknown peaks (f, g, j and k) were higher in chromatogram (A) than those in (B), suggesting that the retention times of these compounds on the concentrator column were around the retention time of 6-NC (39 – 44 min).

Fig. 5 Effect of flow rate of an ascorbic acid solution on the peak heights of NPAHs. Loading time 20 – 58 min.

Table 1 Validations of the proposed method

a. Detection limit, S/N = 3. b. Quantitation limit, S/N = 10. c. Relative standard deviation, n = 3.

Compound Range/fmol r2 DLa/fmol QLb/fmol RSDc, %

1,3-DNP 3 – 100 0.9993 1 3 1.5

1,6-DNP 3 – 100 0.9986 1 3 3.1

1,8-DNP 3 – 100 0.9991 1 3 1.5

2-NF 3000 – 20000 0.9963 1000 3300 3.2

2-NP 500 – 3000 0.9952 150 500 4.0

4-NP 100 – 3000 0.9897 30 100 3.5

1-NP 30 – 1000 0.9964 10 30 2.4

6-NC 500 – 3000 0.9994 150 500 2.9

7-NBaA 100 – 1000 0.9982 30 100 2.3

3-NPer 30 – 1000 0.9983 10 30 2.1

6-NBaP 50 – 500 0.9992 15 50 2.2

Fig. 6 Chromatographic comparison of airborne particulate extract by previous and proposed HPLC systems. A: airborne particulate extract; loading time, 20 – 40 min. B: airborne particulate extract;

loading time, 20 – 58 min. C: airborne particulate extract; loading time, 58 – 100 min. D: NPAH standard; loading time; 20 – 58 min.

For atmospheric concentrations of NPAHs, see text. For concentrations of standard NPAHs: 0.8 pM (DNPs); 7.5 pM (1-, 2- NPs, 7-NBaA); 150 pM (2-NF); 22 pM (2-NP, 6-NC); 3.7 pM (3- NPer, 6-NBaP). Injection volume: 100 µL.

(5)

The unknown peaks (h, i, l and m) were not detected in chromatogram (A), but were detected in chromatogram (B), suggesting that the retention times of these compounds on the concentrator column were in the range of 44 to 58 min. Thus, by using the retention times obtained with the two different loading times it is possible to identify some of the unknown peaks. Furthermore, in the new system, no peaks were detected in chromatogram (C) when the loading time was 58 to 100 min.

These results suggested that no NPAHs can be detected with chemiluminescence after 58 min, and that the loading time of the proposed system (20 – 58 min) is suitable for the quantitative analysis of NPAHs extracted from airborne particulates.

Acknowledgements

This work was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture and by funds from the Ministry of Environment, Japan.

The authors thank Prof. A. Hirayama of Laboratory of the Public Health, Kyoto Pharmaceutical University for providing 2-NP.

References

1. J. R. Landolph, J. C. Bartholomew, and M. Calvin, Cancer Res., 1976, 36, 4143.

2. H. Rosenkranz and R. Mermelstein, Mutation Res., 1983, 144, 217.

3. H. Tokiwa and Y. Ohnishi, Crit. Rev. Toxicol., 1986, 17, 23.

4. K. Hayakawa, Biomed. Chromatogr., 2000, 14, 397.

5. S. B. Tejada, R. B. Zweiginger, and J. E. Jr. Sigsby, Anal.

Chem., 1986, 58, 1827.

6. D. Schuetzle, T. L. Riley, T. J. Prater, T. M. Harvey, and D.

F. Hunt, Anal. Chem., 1982, 54, 265.

7. K. Hayakawa, R. Kitamura, M. Butoh, N. Imaizumi, and M. Miyazaki, Anal. Sci., 1991, 7, 573.

8. K. Hayakawa, T. Murahashi, M. Butoh, and M. Miyazaki, Environ. Sci. Technol., 1995, 29, 928.

9. T. Murahashi, M. Miyazaki, R. Kakizawa, Y. Yamagishi, M. Kitamura, and K. Hayakawa, Jpn. J. Toxicol. Environ.

Health, 1995, 41, 328.

10. H. Kakimoto, M. Kitamura, Y. Matsumoto, S. Sakai, F.

Kanoh, T. Murahashi, K. Akutsu, R. Kizu, and K.

Hayakawa, J. Health Sci., 2000, 46, 5.

11. H. Kakimoto, H. Yokoe, Y. Matsumoto, S. Sakai, F.

Kanoh, T. Murahashi, K. Akutsu, A. Toriba, R. Kizu, and K. Hayakawa, J. Health Sci., 2001, 47, 385.

12. N. Tang, M. Tabata, V. F. Mishukov, V. Sergienko, A.

Toriba, R. Kizu, and K. Hayakawa, J. Health Sci., 2002, 48, 30.

13. K. Hayakawa, T. Murahashi, K. Akutsu, T. Kanda, N.

Tang, H. Kakimoto, A. Toriba, and R. Kizu, Polycycl.

Aromat. Comp., 2000, 20, 179.

14. T. Murahashi, R. Kizu, H. Kakimoto, A. Toriba, and K.

Hayakawa, J. Health Sci., 1999, 45, 244.

15. K. Hayakawa, K. Noji, N. Tang, A. Toriba, R. Kizu, S.

Sakai, and Y. Matsumoto, Anal. Chim. Acta, 2001, 445, 20.

Updating...

参照

関連した話題 :