UV light absorption parameters of the pathobiologically implicated bilirubin oxidation products, MVM, BOX A, and BOX B.

The formation of the bilirubin oxidation products (BOXes), BOX A ([4-methyl-5-oxo-3-vinyl-(1,5-dihydropyrrol-2-ylidene)acetamide]) and BOX B (3-methyl-5-oxo-4-vinyl-(1,5-dihydropyrrol-2-ylidene)acetamide), as well as MVM (4-methyl-3-vinylmaleimide) were synthesized by oxidation of bilirubin with H2O2 without and with FeCl3, respectively. Compound identity was confirmed with NMR and mass spectrometry (MS; less than 1 ppm, tandem MS up to MS4). UV absorption profiles, including λmax, and extinction coefficient (ε; estimated using NMR) for BOX A, BOX B, and MVM in H2O, 15% CH3CN plus 10 mM CF3CO2H, CH3CN, CHCl3, CH2Cl2, and 0.9% NaCl were determined. At longer wavelengths, λmax's for 1) BOX A were little affected by the solvent, ranging from 295-297 nm; 2) BOX B, less polar solvent yielded λmax's of lower wavelength, with values ranging from 308-313 nm, and 3) MVM, less polar solvent yielded λmax's of higher wavelength, with values ranging from 318-327 nm. Estimated ε's for BOX A and BOX B were approximately 5- to 10-fold greater than for MVM.

Specifications Table

Value of the data

First report (to our knowledge) of UV absorption profile, including λmax, of MVM in solvents relevant to detection in biologic/pathobiologic samples.

Comparison of UV absorption profiles of MVM with BOX A and BOX B.

First report (to our knowledge) of BOX B extinction coefficient (ε; estimated using NMR), along with comparison to BOX A and MVM estimated ε’s in different solvents, along with MS at less than 1 ppm and tandem MS up to MS4.

Novel methodology to increase MVM yield through FeCl3 inclusion in oxidation reaction mixture.

Data will potentially assist in the detection and determination of these BOXes in pathobiologies associated with elevated bilirubin.

Data

The bilirubin oxidation products (BOXes), MVM (4-methyl-3-vinylmaleimide), along with BOX A ([4-methyl-5-oxo-3-vinyl-(1,5-dihydropyrrol-2-ylidene)acetamide]) and BOX B (3-methyl-5-oxo-4-vinyl-(1,5-dihydropyrrol-2-ylidene)acetamide), have been implicated in the deleterious effects associated with subarachnoid hemorrhage (SAH; [1], [2], [3], [4], [5]). The detection method utilized to determine the presence of these compounds is UV absorption associated with reversed phase-HPLC [1]. However, reports (to our knowledge) of the UV absorption profile and/or λmax of MVM have not been reported for the solvent utilized in their detection (H2O/CH3CN), but are limited to CH3OH [6], [7]. Also, reports of these absorption characteristics are limited (to our knowledge) for BOX A to H2O and CH3CN, and for BOX B to H2O [1], [8]. Further, extinction coefficients (ε) for MVM and BOX A are limited (to our knowledge) to CH3OH and CH3CN, respectively [6], [7], [9], and are lacking for BOX B. Thus, it is anticipated that the present data will assist in the detection and quantitative determination of BOXes levels in biologic samples from SAH, as well as in other pathobiologies associated with elevated bilirubin.

UV absorption

UV absorption spectra of BOX A, BOX B and MVM were determined in CHCl3, CH2Cl2, CH3CN, 15% CH3CN plus 10 mM CF3CO2H, H2O, and 0.9% NaCl (Fig. 1, Table 1). At longer wavelengths, BOX A λmax's were little affected by the solvent, ranging from 295–297 nm (Fig. 1, Table 1). With BOX B, less polar solvent yielded λmax's of lower wavelength, with values ranging from 308–313 nm (Fig. 1, Table 1). With MVM, less polar solvent yielded λmax's of higher wavelength, with values ranging from 318–327 nm (Fig. 1, Table 1). These λmax values corresponded to previously reported λmax's at longer wavelengths, as limited to the following solvents: BOX A of 300 nm in H2O and 295 nm in CH3CN [1], [2], BOX B of 310 nm in H2O [1], and MVM of 317 and 319 nm in CH3OH [6], [7].

Fig. 1

Absorption spectra for BoxA, BoxB, and MVM. Spectra were determined for Box A (A,B), Box B (C,D,) and MVM (E-H). In A,C,E, and G: CHCl3 (---), CH2Cl2 (), and CH3CN (___). In B, D, F, and H: 0.9% NaCl (__), H2O () and 15% CH3CN/85% H2O with 10 mM CF3CO2H (TFA; (---)).

Table 1

Solvent Effects on λmaxa and εb.

Solvent	Box A		Box B		MVM
	λmax	ε	λmax	ε	λmax	ε
CHCl3	296	13,000	313	24,200	318	2660
CH2Cl2	297	12,200	312	24,300	317	2820
CH3CN	295	10,600	308	22,200	314	2290
CH3CN (15%) + TFA (10 mM)	297	11,900	309	22,400	327	2150
H2O	297	11,000	308	19,000	327	2130
NaCl (0.9%)	297	11,600	308	21,000	327	2100

nm

L/mol•cm

Extinction coefficients (ε)

Calculated ε’s for BOX A, BOX B, and MVM at their respective λmax's in CHCl3, CH2Cl2, CH3CN, 15% CH3CN plus 10 mM CF3CO2H, H2O, and 0.9% NaCl, ranged from 10,600-13,000, 19,000–24,200, and 2,100-2,820 L/mol-cm respectively (Table 1). The ε determined using the actual amount of Z-BOX A (complete chemical synthesis), at λmax 295 in CH3CN, was 17,000 L/mol-cm [9]. Thus, the present Box A ε likely represents a low estimate (Table 1). The estimated MVM ε (Table 1) is similar to that reported for MVM at λmax 317 and 319 nm in CH3OH of 2,300 and 2,290 L/mol-cm [6], [7].

Experimental design, materials and methods

Synthesis

Bilirubin solubilization was performed at room temperature in an aluminum foil wrapped vessel due to the reported light sensitivity of BOX A, BOX B, and MVM [1], [8], [10]. One or more 50 mg portions of bilirubin were incubated in 25 ml 0.2 M NaOH(aq) with occasional vortexing over 24–72 h [1], [10]. The dark red bilirubin solution was then buffered by addition of 7.5 ml of 0.5 M Tris base before neutralization with 0.4 ml of 12.3 M HCl(aq) to pH 7.0. Overtitration of the dark red solution to lower pH resulted in a green solution. The neutralized (pH 7) buffered bilirubin solution was immediately used for oxidation with H2O2. With prolonged storage, bilirubin precipitated from this supersaturated solution.

As performed under dim ambient light and in an unlit fume hood (and with dim ambient light) the neutral buffered solution (now in 0.1 M TrisHCl, pH 7.0, 0.15 M NaCl) was oxidized for 24 h with 8% H2O2 (final concentration). For MVM synthesis, 0.5 M FeCl3 was added (novel procedure) to the bilirubin solution prior to H2O2 and the oxidation allowed to proceed for 10 min. Each aqueous reaction mixture (about 45 ml per 50 mg bilirubin) was extracted twice with 6 ml CHCl3 or CH2Cl2 (recoveries of BOX A, BOX B, and MVM were similar with CHCl3 and CH2Cl2) and the combined organic phase extracted once with 1 ml water, evaporated to ~2 ml at <50 °C and atmospheric pressure, transferred to microfuge tubes, and evaporated to near dryness. Additional ~2 ml aliquots of extract were repeatedly added, each followed by evaporation to near dryness. The final addition of washed extract was evaporated to dryness and reconstituted in 1 ml 1% CH3CN(aq) for purification by reversed phase (RP)-HPLC.

Purification by RP-HPLC

RP-HPLC (0.1 cm light path; Shimadzu LC-10AT, Shimadzu Scientific Instruments, Columbia, MD) was used for both purification and analysis of the bilirubin oxidation products. As performed under dim light, organic solvent extracts of BOX A and BOX B, as well as MVM, reconstituted in 1% CH3CN(aq) were diluted as necessary into the RP-HPLC starting buffer of 2% CH3CN:98% H2O (v:v) containing 10 mM M CF3CO2H. Injections of 1.0–1.5 ml were made onto a Vydac 218TP C-18 5 µm column (250 × 4.6 mm) with guard column equilibrated with 2% CH3CN containing 0.01 M CF3CO2H. The guard column was necessitated by the detection of a small amount of residual H2O2 in the CHCl3 and CH2Cl2 extracts of the bilirubin-H2O2 reaction mixtures. An attempt to remove the H2O2 with CH3CH2OH (molar CH3CH2OH:H2O2 ratio 1.5:1) addition to the reaction mixtures actually caused a 10-fold increase in the amount of substrate detected as H2O2. H2O2 was not detected in RP-HPLC-purified oxidation products.

The column was eluted (1 ml/min) with a continuous gradient of 0.5% CH3CN/min (2% to 18% CH3CN) over 32 min, followed by steeper gradients and higher CH3CN concentration for washing the system between runs. Eluates were monitored from 210–350 nm using a diode array spectrophotometer and flow cell and were collected in aluminum foil wrapped test tubes.

RP-HPLC of the combined products of bilirubin-H2O2 reaction mixtures with and without Fe3+ yielded three peaks with retention times at 26.0, 28.7, and 31.2 min, respectively (Fig. 2). These retention times corresponded to eluting CH3CN concentrations of 12.8, 14.4, and 15.6% (v/v), respectively. UV absorption at other retention times was not detected at 297, 310, and 327 nm, i.e., at the longer wavelength λmax's of the compounds with 26.0, 28.7, and 31.2 retention times, respectively, as well as at 223 nm (Fig. 1, Fig. 2; Table 1), indicative of a purified preparation. This relative order of retention time of MVM, BOX A, and BOX B differs from that which a laboratory previously reported, which was BOX A, BOX B, and then MVM [1], [5]. While this difference in relative order of retention time may be due to differences in column properties, it should also be considered that the present inclusion of CF3CO2H in the solvent resulted in ion pairing with BOX A and BOX B (retention times of the ion pairs would be increased as compared to the non-paired species).

Fig. 2

HPLC traces of BOX A, BOX B, and MVM. Standards were prepared and subjected to RP-HPLC (Materials and Methods). Absorption was determined at 223 nm (upper trace) and at the respective λmax's of BOX A, BOX B, and MVM, which were 297 (second most upper trace), 310 (third most upper trace), and 327 nm (lower trace), respectively (see Fig. 1 and Table 1).

From the bilirubin-H2O2 oxidation in the absence of Fe3+, the ratio of MVM:BOX A:BOX B (BOX A set at absorption unity) formed at their respective λmax's (Fig. 1, Fig. 2, Table 1) was 0.10 ± 0.03:1.0:0.95 ± 0.05, respectively (mean ± SE; n = 5). Several minor peaks were also observed (determined prior to further RP-HPLC purification). Incubation at times shorter or longer than 24 h did not result in additional MVM formation. Yields after purification of BOX A and BOX B were ~1% each, based on starting material and measured by UV spectroscopy (calculated with ε’s as described below; Table 1).

From the bilirubin-H2O2 oxidation in the presence of Fe3+, the ratio of MVM:BOX A:BOX B (MVM set at absorption unity) formed at their respective λmax's (Fig. 1, Fig. 2, Table 1) was 1.0:0.05 ± 0.01:0.04 ± 0.01, respectively (mean ± SE; n = 5). Several minor peaks were also observed (determined prior to further RP-HPLC purification). Incubation for 1, 5, 30, 45, and 60 min did not increase BOX A and BOX B formation while MVM formation was reduced. The reaction yielded ~5% MVM, based on starting material and measured by UV spectroscopy (calculated with ε; Table 1). Increased MVM formation with Fe3+ inclusion in the bilirubin-H2O2 reaction mixture is consistent with the dependency of MVM formation following H2O2 oxidation of ferriprotoporphyrin IX on the chelated iron [11] as well as the oxidation of bilirubin by CrO3 [6].

Present yields are generally consistent with earlier reports of <5% and 4% formation of BOX A, BOX B and MVM [1], [10]. While one of these reports [1] also demonstrated significant MVM synthesis (MVM:BOX A:BOX B = 2.8:1:0.9; determined at 320 nm and as presently calculated with BOX A set to unity), the increased MVM formation was possibly due to a somewhat greater H2O2 concentration in the reaction mixture with bilirubin (13% H2O2). On the other hand, highly variable amounts of MVM were formed by oxidation of bilirubin with ~10% H2O2 [10]. Hydrogen peroxide oxidation of biliverdin instead of bilirubin did not increase the yield of MVM.

Stability

After purification, BOX A, BOX B, and MVM samples shielded with aluminum foil from light were stable for at least 6 mo at −20 °C and for 24 h at room temperature in 14.6% CH3CN (eluting solvent), as determined by RP-HPLC; i.e., no loss of compound or detection of additional absorption peaks through the UV absorption spectrum. Removal of the aluminum foil and exposure of BOX A, BOX B, and MVM (in 14.6% CH3CN in a clear polypropylene microfuge tube) to ambient light for 24 h decreased recovery by 10%, 15%, and 5%, respectively, and the appearance of peaks at 18.3 min and 20.8 min with a ratio 1.13:1, and with λmax's of 288 and 296 nm, respectively.

UV absorption spectrometry

UV spectra were performed in a SpectraMax M5 (Molecular Devices, Sunnyvale, CA, USA).

1H-NMR

For compound identification and ε determinations, analytic samples of BOX A, BOX B, and MVM (CHCl3 extraction) were loaded onto a C18 separation cartridge (Sep-Pak), washed with 1 ml D2O and eluted with 1.5 ml, 80% CD3CN (in D2O). Samples were then evaporated to dryness under N2 and reconstituted in 1 ml CD3CN. BOX A, BOX B, and MVM chemical shifts and coupling constants were determined on a DMX-500. Extinction coefficients (ε) at the respective λmax's for BOX A, BOX B, and MVM were determined by titration in CH3OH (3.49 ppm singlet) and integration of signals relative to CH3OH under conditions of long recycle delay, and determination of UV absorption. 1H-NMR spectroscopy yielded chemical shifts and coupling constants for BOX A, BOX, B, and MVM consistent with previous reports (Fig. 3; Table 2; [1], [6], [9]).

Fig. 3

Structures of BOX A, BOX B, and MVM and correspondingH-NMR data. Structures of BOX A and BOX B depicted as Z regio-isomers (after Seidel et al., 2015) and MVM with 1H-NMR data from Table 2.

Table 2

1H-NMR chemical shifts and coupling constants for synthetic compoundsa.

	Box A	Box B	MVM
Chemical shiftsb
δ vinyl-CH	6.56	6.59	6.58
δ vinyl-CH2 cis	5.70	6.27	6.27
δ vinyl-CH2 trans	5.67	5.52	5.65
δ -CH3	1.96	2.06	1.97
δ -CONHR	9.71	9.52	7.20
δ =CH-CONH2	5.65	5.59	n/a
Coupling constantsc
3J trans vinyl	17.9	17.9	17.8
3J cis vinyl	11.7	11.7	11.5
2J gem vinyl	<1.5	2.1	1.7

Spectra collected at 500 MHz in CD3CN.

Chemical shifts (δ) in ppm downfield of TMS, referenced to CHD2CN at 1.940 ppm.

Coupling constants (J) in Hz.

MS

Samples for MS were prepared by evaporation of compounds in aqueous CH3CN to dryness in an N2 stream at 40 °C, followed by reconstitution in 10% CH3CN/90% H2O containing 0.2% HCO2H. Lyophylization was avoided due to apparent loss of compounds. Samples obtained from RP-HPLC were infused into a Thermo Scientific LTQ-FT™ hybrid MS consisting of a linear ion trap and a Fourier transform ion cyclotron resonance (FT-ICR) MS. The standard electrospray ionization (ESI) source was operated in a profile mode for both positive and negative ions as indicated (Fig. 4, Table 3). The only possible elemental composition at 2 ppm mass error, but also even at 5 ppm, for 0–10 nitrogen, 0–15 oxygen, 0–30 carbons, and 0–60 hydrogens are those of BOX A and BOX B (as assessed for the positive ion mode), and for MVM (as assessed in both the positive and negative ion mode; Fig. 4, Table 3; consistent with 1,6,9,10). With MVM as the protonated molecular ion, the observed mass was m/z 138.05498 with a mass error of 180 ppb. For MVM, MS also suggested the apparent presence of the plastic antioxidant/stabilizer 1,10-bis(2,2,6,6-tetramethyl-4-piperidinyl-decanedioate), resulting from the (initial) carrying out of the FeCl3-bilirubin-H2O2 oxidation in a polypropylene vessel (“2. Experimental design, materials and methods; 2.1. Synthesis”; subsequent oxidations were performed in glass containers).

Fig. 4

Mass spectrometry (MS) of BOX A, BOX B, and MVM. MSn for BOX A (A and B), BOX B (C and D), and MVM (E). Corresponding group loss and error are indicated in Table 3.

Table 3

FT-ICR.

	Product Ion	Error (ppb)	Loss
Box A
MS2	162.05485	−649	NH3
MS3	134.05995	−675	CO
MS4	106.06507	−527	CO
Box B
MS2	162.05485	−649	NH3
MS3	134.05997	−526	CO
	106.06506	−621	2CO
	91.05418	−515	[ion is C7H7+]
	144.04430	−628	H2O
	116.04940	−653	CO, H2O
MS4	106.06506	−621	CO
	79.05419	−344	[ion is C6H7+]
	91.05419	−445	[ion is C7H7+]

Subject area	Chemistry
More specific subject area	Bilirubin oxidation products detection
Type of data	Table, figure
How data was acquired	NMR, mass spectroscopy, UV spectrometry, HPLC
Data format	Raw, analyzed
Experimental factors	Oxidation of bilirubin, extraction with chloroform
Experimental features	Bilirubin oxidation products BOX A, BOX B, and MVM were synthesized by the oxidation of bilirubin, purified by HPLC and UV absorption profiles and extinction coefficients determined
Data source location	Cincinnati, OH USA
Data accessibility	The data are accessible within the article.