BTOMEDlCAL CHROMATOGRAPHY, VOL. 6,55-58 (1992)

Peroxyoxalate Chemiluminescence Detection of Condensates of Malondialdehyde with Thiobarbituric Acids Using a Flow System Kenichiro Nakashima,* Masahiko Nagata, Masakatsu Takahashi and Shuzo Akiyama School of Pharmaceutical Sc~enceq,Nagasaki University. 1-14 Bunkyo-mdchi, Nagasaki 852, Japan

The peroxyoxalate chemiluminescence(C1,) detection method for the evaluation of the CL intensity of malondialdehyde(MDA) condensates with seven 2-thiobarbituric acid derivatives is described. The method consists of a flow injection technique together with a CL detection system using bis(2,4,6-trichlorophenyl)oxalate(TCP0) and hydrogen peroxide as chemiluminogenic reagents. Linear correlations between CL intensity and concentration are obtained for pmol levels of condensates. Among the condensates, 1,3-diethyl-2-thiobarbituricacid(DETBA)MDA shows the largest CL intensity. High pcrformmance liquid chromatography (HPLC)/CL detection of DETBA-MDA and 1,3-diphenyI-2-thiobarbituricacid(DPTBA)-MDA using a mixture of TCPO and hydrogen peroxide in acetonitrile as a postcolumn reagent solution is also described. The detection limits for DETBA-MDA and DPTBA-MDA are 20 and 200 fmol, respectively, per 20 pL injection at a signal-to-noise ratio of 2. This HPLC/CL detection system was applied to the determination of MDA in rat brains by using DETBA as a fluorescent derivatizing reagent.

INTRODUCTION Malondialdehyde(MDA) derived from peroxidized polyunsaturated lipids is well known as an index of lipoperoxide, which has been implicated in aging (Frei et al., 1988), mutagcnesis (Mukai and Goldstein, 1976; Shamberger ef al., 1979) and carcinogcnesis (Basu and Marnett, 1983). The 2-thiobarbituric acid(TBA) method has been commonly used for MDA assay (Yagi, 1976; Buege and Aust, 1978; Uchiyama and Mihara, 1978; Wong et al., 1987). The method consists of the spectrophotometric or fluorornetric measurement of the red-coloured reaction product of MDA with TBA. 'To improve the sensitivity. 1,3-diphenyl-2thiobarbituric acid(DPTBA) was used instead of TBA as a derivatizing reagent for MDA (Nakashima et al., 1983; Yoshimura et a l . , 1988). However, TBA or DPTBA react with several compounds other than MDA to give similarly coloured pigments. Consequently, there are some examples where the methods using TBA derivatives have overestimated the MDA value (Matsuo et al., 1984; Csallany et al., 1983). To counteract this defect, high performance liquid chromatographic(HPLC) separation of MDA-TBA condensates has been recommended (Bird et al., 1983; Therasse and Lemonnier, 1987). In a previous paper, we synthesixd condensates of 13 different barbituric acid derivatives with MDA and studied their spectral properties (Nakashima et al., 1983). Among these, the DPTBA condensate showed the largest absorption intensity. A sensitive and selective HPLC/visible detection method for DPTBA-MDA condensate was devcloped (Nakashima et al., 1984, 1985). Recently it has been reported that HPLC with peroxyoxalate chemiluminescencc( CL) detection is an excellent method for the sensitive determination of * Author

to whom correspondence should be addressed

0269-3879 /Y2/02OO55-O4 $05.00

01992 by John Wiley & Sons, Ltd.

fluorescent compounds (Imai, 1986, Imai et al., 1987; Baeyens et al., 1989). This paper describes our study of peroxyoxalate CL of fluorescent TBA-MDA condensates by a flow injection system in order to find a suitable reagent for MDA among the TBAs. Furthermore, HPLC/CL detection of 1.3-diethyl-2-thiobarbituric acid(DETBA) or DPTBA condensates with MDA was studied, using bis(2,4,6-trichlorophenyl) oxakdte(TCP0) and hydrogen peroxide as postcolumn cherniluminogenic reagents, in order to obtain fundamental data for the HPLC/CL assay of MDA with DETBA as a fluorescent derivatizing reagent. Subsequently, this HPLC/CL detection system was successfully applied to the determination of MDA in rat brains.

EXPERIMENTAL Apparatus. A flow diagram for the measurement of CL is shown in Fig. I . The flow injection system consists of three Shimadzu (Kyoto. Japan) Model LC-(,A HPLC pumps, a Shimadru Model SCL-6A system controller, a Rheodyne

TEA

a

@

+

Waste

Figure 1. Flow diagram for the measurement of CL. P: pump; M: mixing T; I: injector; D: detector; R: recorder; DC,: delay coil ( 1 0 0 0 0.5 ~ mm, i.d.1; DC2:delay coil (100 x 0.5 mm, i.d.1; TCPO: 0.15 mM in CHJN (1.0 mllmin); H,02: 4.5 mM in CH,CN (1.0 rnllmin); TEA: 3 mM in CH,CN (1.0 mllmin).

Rrrriued 17 Drcember 1990 Accepted 15 February I991

K. NAKASHIMA EZ A L .

56

(Cotati, CA, USA) Type 7125 injection valve with a 20 pL loop, an ATTO Model AC-2220 luminomonitor with a 60 pL spiral flow cell and a Pantos (Tokyo, Japan) Model U-228 recorder. Stainless-steel tubing of 0.5 mm in diameter was used for the mixing coils. A flow diagram for HPLClCL detection is shown in Fig. 2. The apparatus are the same as those described in Fig. 1 except that two HPLC pumps and an analytical column are used. For the analytical column, Tosoh (Tokyo, Japan) TSK-ODS-SOTM (150X4.6mm i d . ; 5 pm) was used. A mobile phase (A) was used for the separation of an authentic sample of DETRA-MDA or DPTBA-MDA and a mobile phase (B) was used for the separation of DETBA-MDA in rat brains. Fluorescence spectra were measured on a Hitachi Model 650-10s fluorescence spectrophotometer.

Table 1. Fluorescence properties of the MDA condensates with TBA derivatives R'

Ex

Fluorescence Em

RFP

nm in EtOH)

Compounds

533 534 534 536 542 540 538

1

2 3 4

5 6 7 a

R'

549 552 552 552 558 557 555

1.oo 1.61 1.88 1.75 1.14 1.62 0.77

Relative fluorescence intensity. RFI of 1 was arbitrarily taken a s

1.oo.

Chemicals. Synthesk of TBA-MDA condensates: Condensates were synthesized according to the literature (Shepherd, 1964; Nakashima et al., 1983). Fluorescence properties of the condensates are summarized in Table 1. Standard stock solutions (5 x M) of each codensate were prepared by dissolving in ethanol for the measurement of CL intensity by flow injection system. For HPLCKL detection, the condensates were dissolved in a mobile phase (A) to give a W 4M solution. Each standard solution was appropriately diluted with ethanol or a mobile phase before use. Procedure j o r assay of MDA in rat brains: To 20 pL of 2% rat brains homogenate (or MDA standard solution) were added 100 pL of 0.2 M phosphate buffer (pH 2.0) and 100 pL 10 mM DETBA in acetonitrile in a 3 mL screw-capped vial. The well-mixed solution was incubated at 95 "C for 30 min in a heating block, chilled in tap water and then 780 pL of acetonitrile was added. After vortex mixing, the mixture was centrifuged at 15OOg for 10min. A 20pL aliquot of the supernatant obtained was injected into the HPLC column. A stock standard solution of MDA was prepared by dissolving 1,1,3,3-tetraethoxypropane(TEP) in water (5 X M) and was stable for two weeks at 4 "C. A working solution was prepared by appropriately diluting the stock solution before use. The rat brains used were of Wistar rat (7 weeks old). The fresh brains obtained were divided into seven regions and homogenated. A calibration graph was prepared with a MDA standard solution, and chromatograms were evaluated on the peak height of MDA condensate. TCPO, imidazole and 30% hydrogen peroxide were analytical grade and obtained from Wako Pure Chemicals (Tokyo, Japan). TEP and triethylamine(TEA) were purchased from Tokyo Kasei Kogyo (Tokyo, Japan). Acetonitrile was HPLC grade (Wako) and filtered through a membrane filter (0.45 pm) before use. All other chemicals were analytical grade

RESULTS AND DISCUSSION Comparison of chemiluminescence intensities of TBA-MDA condensates

Generally, the fluorescence intensity of the compound affects the CL intensity. Therefore, fluorescence properties of the TBA-MDA condensates were firstly examined. As shown in Table 1, there is no remarkable difference in wavelength and intensity of fluorescence among the condensates. CL intensities of various condensates were then measured using the flow injection system (Fig. l). The CL reaction took place in non-aqueous acetonitrile solution. The CL reaction conditions used are the same as those reported previously (Nakashima et al., 1990). The calibration curves for the condensates were prepared and the relative CL intensity (RCI) was calculated from the slope of each calibration curve with the RCI of 1 taken as 1.0 (Table 2). The DETBA-MDA condensate (6) gave the largest RCI. It therefore seemed that DETBA was the most sensitive derivatizing reagent for MDA. Though 3 showed the largest RFT, its RCI was smaller than that of 6. The reason is not clear, but it may be caused by differences in conditions, such as solvents, pH, etc. A representative recorder response with 6 is shown in Fig. 3. By comparison. MDA condensates with barbituric acids showed

Table 2. Calibration curves and relative chemiluminescence intensities of condensates Compounds

1

2 3 Waste

Figure 2. Flow diagram for HPLC/CL detection. Re: 1.0 mM TCPO+50 mM H,O, in CH,CN (1.0 mL/min); mobile phase (A): 5 rnM imidazole buffer (pH 8): CH,CN [ l : l , v/v, containing 0.1 M NaCl (1.0 mL/min)]; mobile phase (6): 10 mM imidazole buffer (pH 7.5): CHJN [ l : l , v/v, containing 0.1 mM tetrabutylammonium bromide (1 .O mL/rnin)]; column: Tosoh ODs-8OTM (150 x 4.6 mrn, id.); DC: delay coil (2000 X 0.5 mrn, i.d.1. For other conditions see Fig. 1.

4

5 6 7

Equationa

RClb

Y=O.34X- 0.05 Y= 0.68X- 0.48 Y= 0.89X+ 0.09 Y- 0.71 X+ 0.22 Y= 0.58X- 0.45 Y= 1.28X+ 1.16 Y= 0.73X- 0.40

1.oo 2.00 2.62 2.09 1.70 3.76 2.15

Y= peak height (cm),X = concentration of condensate (pmol/ injection). Relative chemilurninescence intensity. RCI of 1 was arbitrarily taken as 1.00.

a

57

PEKOXYOXALATE CII DETECTION OF TBA-MDA CONDENSATES

very low CL intensities (no data was shown). Thus it is clear that the thiocarbonyl group at the 2-position on the barbituric ring is needed for strong CL.

(B)

f

(A)

HPLCKL detection of MDA condensates with DETBA and DPTBA HPLC/CL detection was examined for the DETBA-MDA condensate. DPTBA-MDA, which had been previously examined as the reaction product of DPTBA and MDA by HPLC colorimetric detection, was also examined to compare with DETBA-MDA. Generally, three pumps are used for the peroxyoxalate CL detection system as shown in Fig. 1. However, a mixture of hydrogen peroxide and aryl oxalate in acetonitrile could be used as postcolumn reagents for HPLC/CL detection. This has the advantage that the system can reduce the number of pumps as shown in Fig. 2. In this case, the stability of the solution of chemiluminogenic reagent is very important. Thus, the stability of the mixture of hydrogen peroxide and TCPO in acetonitrile as a function of peak height of 6 was measured. As a result, at least for 6 h, constant peak heights were obtained and 90% of the original peak height was maintained even after 12h. These results agreed well with those of Tmaizumi et al. (1989). For the HPLC separation of 6, 5 rnhf irnidazole buffer and acetonitrile containing 0.1 M NaCl was used as a mobile phase (A). Representative chromatograms for 6 and 7 are shown in Fig. 4. Under these conditions, linear relationships between the peak height and the concentration were obtained up to 500 fmol for 6 and 5 pmol for 7.The detection limits for 6 and 7 were 20 and 200 fmol, respectively, per injection at a signal-tonoise ratio of 2. This result suggests that a highly sensitive HPLC/CL assay of MDA could be possible.

12,o

r

0

i

0

4

12

8

t i me/m i n

4

8 time/rnin

12

Figure 4. HPLC chromatograms for MDA condensates with DETBA and DPTBA. (A) DETBA-MDA, 400fmol; (B) DPTBA-MDA, 1.O pmol. For other experimental conditions see text.

HPLC/CL assay of MDA in rat brains DETBA was suggested as being the preferred derivatizing reagent for MDA with HPLC/CL detection. Thus the reactivity of MDA with DETBA was examined and it was found that the yield of the condensate increased with a increase in reaction temperature or time (Fig. 5). It was also found that a lower acidic pH gave a higher yield (Fig. 6). As a result, the reaction of MDA with DETBA was carried out at 95 "C for 30 min using a phosphate buffer (pH2.0). The reaction product obtained was stable at 4°C for at least 24h. For the HPLC separation, 10mM imidazole buffer and acetonitrile containing 0.1 mM tetrabutylammonium bromide was used as a mobile phase (B).

c

c

biank

A.

9,a

6-0 pmo l / i n j ec t ion

3,O

Figure 3. Recorder response of chemiluminescence with the DETBA-MDA condensate. For experimental conditions see text.

"0

20

40

60 80 Tirne/mln

100

120

Figure 5. Effect of reaction temperature and time on the reaction yield. MDA solution, 2.5 x M: (0)95 "C,(0) 60 "C. For other conditions see text.

58

K. NAKASHIMA ET A I , .

0

I

1

2

* 3

I

4

5

6

I

PH

t ime/mi n

Figure 6. Effect of pH of phosphate buffer on the reaction yield. MDA solution, 2.5 x 10 M. For other conditions see text.

Tetrabutylammonium bromide improved the stability of the base line and the shape of the peak. A calibration graph was linear up to 500frnol (r=0.997). However, a tiny impurity peak was observed from the reagent blank which overlapped unfavourably with t h e peak derived from MDA and the peaks could not be separated from each other under any separation conditions used. The detection limit for MDA was 50 fmol at a peak height ratio of signal-to-blank of 2. The reproducibility of the peak height for MDA (200 fmol) was 3.6% (n=5). The percentage recovery of MDA was calculated from the slopes of calibration graphs obtained from standard solutions and the spiked cerebral cortex homogenate was 93.7%. The MDA in rat brains was determined with four rats. Amounts of MDA (nmol/g wet tissue) obtained were as follows: cerebral cortex (47.3-1 12.11, midbrain (66.1 f 16.71, cerebellum (35.2&8.3), hippocampus (52.6& S . l ) , hypothalamus (66.5 k 21.1), striatum (86.6 -124.3) and pons/medulla oblongata (59.9 k 20.2). A typical chromatogram for MDA in rat brians is given in Fig. 7.

t ime/min

Figure 7. HPLC chromatograms for MDA-DETBA condensate in rat midbrain. (A) Rat midbrain; 03) reagent blank. For experimental conditions see text.

CONCLUSION By using a flow injection/CL detection system, the CL intensities of seven different TBA-MDA condensates were measured and it was found that DETBA-MDA yielded the highest CL intensity. This means that DETBA is the most sensitive dcrivatizing reagent for MDA determination using TBA derivatives with CL detection. HPLC/CL Detection of DETBA-MDA was also examined and it was found that 20fmol of DETBA-MDA could be detected. On the basis o f these facts, HPLC/CL method for the determination of MDA by using DETBA as a label was examined. Consequently, a sensitive method was developed and successfully applied to the determination of MDA in rat brains. This method should be useful for clinical studies.

REFERENCES

Baeyens, W. R. G., Nakashirna, K., Imai, K, Ling, B. L. and Tsukamoto, Y. (1989). J. Pharm. Biorned. Anal. 7, 407. Basu, A. K. and Marnett, L. J. (1983). Carcinogenesis 4, 331. Bird, R . P., Hung, S. S. O., Hadley. M. and Draper, H. H. (1983). Anal. Biochem. 128, 240. Buege, J. A. and Aust, S. D. (1978). Methods Enzymol. 52, 302. Csallany, A. S., Guan, M. D., Manwaring, J. D. and Addis, P. B. (1984). Anal. Biochem. 142, 277. Frei, B.. Yamamoto, Y., Niclas, D. and Ames, B. N. (19881. Anal. Biochem. 175, 120. Imai, K. (1986). Methods Enzymol. 133, 435. Irnai, K., Nishitani, A. and Tsukarnoto, Y. (1987). Chromatographia 24, 77. Irnaizumi, N., Hayakawa, K., Miyazaki, M. and Imai, K. (1989). Analyst 114, 161. Matsuo, T., Yoneda, T. and Itoo, S. (1984). Agric. Biol. Chem. 48, 1631. Mukai, F. H. and Goldstein, B. D. (1976). Science 191, 868. Nakashima, K., Ando, T., Nakarnura, K. and Akiyama, S. (1983).

Chem. Pharm. Bull. 31, 2523. Nakashima, K.. Ando, T. and Akiyama, S. (1984). Chern. Pharm. Bull. 32, 1654. Nakashima, K., Ando, T., Nakamizo, T. and Akiyarna, S. (1985). Chern. Pharm. Bull. 33, 5380. Nakashirna, K., Maki, K., Akiyama, S. and Imai, K. (1990). Biorned. Chromatogr. 4, 105. Sharnberger, R. J., Corlett, C. L., Beaman, K. 0. and Kasten, B. L. (1979). Mutat. Res. 66, 349. Shepherd, R. G. (1964). J. Chem. SOC.1964, 4410. Therasse, J. and Lernonnier, F. (1987). J. Chromatogr. 413, 237. Uchiyama, M. and Mihara, M. (1978). Anal. Biochem. 86, 271. Wong, S. H. Y., Knight, J. A., Hopfer, S. M., Zaharia, O., Leach, C. N. Jr. and Sunderrnann, W. Jr. (1987). Clin. Chem. 33, 214. Yagi, K. (1976). Biochem. Med. 15, 212. Yoshimura, Y., Koike, S., Tanaka, H.,Tamura, K., Ohsawa, K., Imaeda, K., Akiyama, S. and Nakashima, K. (1988). Anal. Sci. 4, 207.