Selective Quantification by 2D HSQC₀ Spectroscopy of Thiocoraline in an Extract from a Sponge-Derived Verrucosispora sp

Abstract

We recently developed 2D ¹H-¹³C HSQC₀ approach to quantify individual chemicals in complex mixtures. The HSQC₀ approach has been implemented in phase-cycled and gradient selective versions. As in quantitative 1D NMR, the normalized integrated signal intensities in HSQC₀ are proportional to the concentrations of individual chemicals in the mixture. We applied the HSQC₀ approaches to selectively quantify thiocoraline present at a level of 1% w/w in an extract from a Verrucosispora sp. isolated from the sponge Chondrilla caribensis f. caribensis. We expect that this approach can be used to quantify other natural products of interest in extracts without prior purification.

The field of natural products has seen a constant increase in the application of quantitative NMR methods over the past 40 years.¹ Quantitative NMR methods have been used, among other applications, for biosynthetic studies,² quantification of species in complex mixtures,³ and regulation of reference materials.⁴ Quantitative NMR offers advantages over more sensitive MS-based methods in being unbiased and not subject to variability in ionization. Quantitative 1D proton NMR (qHNMR) has become a routine analytical tool because of its universality, sensitivity, precision, and nondestructive nature.¹ Furthermore, the proportionality of the integrated intensities of proton resonances to the overall number of proton spins in the mixture makes qHNMR suitable for quantification of individual compounds in mixtures.^{1, 5} However, 1D qHNMR has shortcomings for signals that are overlapped⁶ as often is the case with complex mixtures of natural products.

Our recent extension of quantitative NMR to 2D ¹H-¹³C spectroscopy overcomes the overlap problem by taking advantage of peak dispersion in the two dimensions. The cross-peaks in standard 2D ¹H-¹³C NMR spectra, such as HSQC (homonuclear single quantum correlation), are not proportional to concentration as the result of several factors. We have shown that peak intensities that do scale with concentration can be determined by an approach called time-zero HSQC, or HSQC₀.⁷ We have developed two versions of this experiment: phase-cycled HSQC₀⁷ and gradient-selective HSQC₀.⁸ As in 1D qHNMR, the integrated signal intensities in the virtual HSQC₀ spectrum are directly proportional to concentrations of individual chemicals in the mixture. With the HSQC₀ approach, it is not necessary to relate measured intensities to data from the pure compound at known concentration; concentrations can be determined instead by reference to peak intensities of internal standards added at known concentration, for example, TMS ((CH₃)₄Si) for samples in CDCl₃, or DSS (4,4-dimethyl-4-silapentane-1-sulfonic acid) for samples in aqueous solution. In theory, the HSQC₀ approach can be used to quantify all individual chemicals in the mixture without isolation or purification.

We illustrate here the application of the HSQC₀ approach to selectively quantify thiocoraline (1). Thiocoraline (1), which was first isolated in 1997 from the fermentation broth of Micromonospora marina,^{9, 10} has shown potent cytotoxicity in lung, breast, colon, renal, and melanoma cancer cells,^10–12 and in vivo efficacy against human carcinoma xenografts.¹³ For this study, thiocoraline (1) was quantified in an extract from a Verrucosispora sp. isolated from the sponge Chondrilla caribensis f. caribensis.¹⁴ The isolation of thiocoraline (1) and related analogs from a Verrucosispora sp. was recently published.¹⁵

As described previously,⁷ a series of HSQC_i spectra can be acquired with one, two, or three repetitions of the basic HSQC building block. The time-zero 2D ¹³C HSQC spectrum (HSQC₀) is then obtained by linear regression extrapolation to zero repetitions (time zero). For gradient-selective HSQC_i spectra,

$$ln(A_{i,n})=ln(2A_{0,n})+i\times ln(f_{A,n}·\frac{1}{2})$$ (1)

and for phase-cycled HSQC_i spectra,

$$ln(A_{i,n})=ln(A_{0,n})+i\times ln(f_{A,n})$$ (2)

where f_A,n is the amplitude attenuation factor through each HSQC block specific for peak n; A_i,n and A_0,n are the amplitudes of peak n in HSQC_i spectra and the extrapolated virtual HSQC₀ spectra; i is the number of times of repetition of the basic HSQC building block. Although the HSQC₀ approach was designed initially for determining the absolute concentrations of individual compounds in metabolite mixtures, it also can be used to selectively quantify natural products of interest in an extract.

To evaluate the reproducibility of our measurements, two NMR samples were prepared from extracts from a Verrucosispora sp. Sample #1 contained 8.0 mg of the dry extract plus 2.59 μg TMS (final concentration 58.7 μM), and Sample #2 contained 4.0 mg of the dry extract plus 3.63 μg TMS (final concentration 82.3 μM). Note, the ratios of extract to TMS were intentionally different in the two samples. 40 μL of 4% Cr(AcAc)₃ in CDCl₃ was added to each sample as a relaxation enhancing agent, and CDCl₃ was then added to a final volume of 500 μL for each sample. Independent HSQC_i data sets were collected for each sample in both constant-time phase-cycled mode and non-constant-time gradient selective mode.

Cross-peaks assigned⁸ to the (13, 13′), (14, 14′), and (16, 16′) methyl groups of thiocoraline (1) (Figure 1) were chosen for quantifying thiocoraline (1) in the extract because of their characteristic and relatively stronger intensities. Intensities of these peaks, together with that of TMS, were manually integrated as described previously⁷ (Tables 1–3).

Figure 1.

(A) 700 MHz 2D constant-time phase-cycled HSQC₁ spectrum of the NMR sample with 8.0 mg extract and 58.7 μM TMS as concentration reference in CDCl₃. The cross-peaks of thiocoraline (1) used for the concentration measurement together with that of TMS methyl groups are shown in the expanded panels (B) 13, 13′ of thiocoraline (1), (C) TMS, and (D) 14, 14′ and 16, 16′ of thiocoraline (1).

Table 1.

Manually integrated peak volume from constant-time phase-cycled HSQC_i of sample #1

	csC	csH	HSQC1	HSQC2	HSQC3	Corr. Coef.	Scaling factor	A0	# of H	Normalized A0	Ave./Stdev	w/w % of thiocoraline
13,13′	15.166	2.137	2.55E+09	1.63E+09	1.41E+09	−0.9592	0.7436	3.26E+09	6	5.43E+08	5.81E+08
14,14′	30.773	3.059	2.63E+09	2.00E+09	1.33E+09	−0.9936	0.7111	3.78E+09	6	6.30E+08	7.69%	1.12%
16,16	30.627	3.013	2.31E+09	1.69E+09	1.10E+09	−0.9959	0.6901	3.41E+09	6	5.69E+08	Mol. of TMS
TMS	19.999	0	1.96E+09	1.72E+09	1.17E+09	−0.9617	0.7726	2.65E+09	12	2.21E+08	2.93584E-08

Table 3.

Manually integrated peak volume from constant time phase-cycled HSQC_i of sample #2

	csC	csH	HSQC1	HSQC2	HSQC3	Corr. Coef.	Scaling factor	A0	# of H	Normalized A0	Ave./Stdev	w/w % of thiocoraline
13,13′	15.153	2.139	6.98E+08	5.13E+08	4.51E+08	−0.9731	0.8038	8.43E+08	6	1.40E+08	1.73E+08
14,14′	30.754	3.059	7.98E+08	4.44E+08	4.07E+08	−0.9193	0.7142	1.03E+09	6	1.71E+08	19.15%	0.90%
16,16	30.608	3.011	7.97E+08	4.42E+08	3.06E+08	−0.9912	0.6196	1.24E+09	6	2.07E+08	Mol. of TMS
TMS	19.997	0	2.11E+09	1.60E+09	1.24E+09	−0.9997	0.7666	2.74E+09	12	2.29E+08	4.11471E-08

NMR sample #2: 82.3 μM TMS (3.63 μg in 500 μL) was added as concentration reference into the NMR sample solution with 4.0 mg extract. Molecular weights of TMS and thiocoraline are 88.22 and 1157.41, respectively.

As shown in Figure 2, the peak intensities of the (13, 13′), (14, 14′) and (16, 16′) methyl groups of thiocoraline (1) and that of methyl groups of TMS in HSQC₀ were extrapolated by linear regression according to Eq. (1) or Eq. (2). As predicted, the slopes of the linear regressions of the gradient-selective data (Figure 2B) are steeper due to the scaling factor of 0.5 (Eq. 1) than those for the constant-time data (Figure 2A, C; Eq. 2). The intercepts from the linear regressions that were obtained using the gradient-selective pulse program in Figure 2B are larger because of the signal enhancement due to the linear recombination of the time-domain NMR data recorded in the echo-antiecho mode before Fourier transformation. The extrapolated HSQC₀ peak intensities, A₀, are shown in Tables 1–3 along with the values normalized by the corresponding number of attached protons. The averaged normalized peak intensities in HSQC₀ from the cross-peaks of the (13, 13′), (14, 14′), and (16, 16′) methyl groups of thiocoraline (1) were used to calculate the final quantity of thiocoraline (1) in the extract by reference to the TMS reference standard:

$$w/w(\%of\mathit{Thio}.)=\frac{\mathit{Ave}.\mathit{Norm}.A_0(\mathit{Thio}.)}{\mathit{Norm}.A_0(\mathit{TMS})}·\mathit{Mol}.(\mathit{TMS})·M.W.(\mathit{Thio}.)/\mathit{Weight}(\mathit{dry}\mathit{extracts})$$ (3)

Figure 2.

Extrapolations of the peak intensities of 13, 13′ (◆), 14, 14′, (■) and 16, 16′ (▲) methyl groups of thiocoraline and that of the methyl groups of TMS (×) to yield HSQC0 intensities from (A) constant-time phase-cycled HSQCi of sample #1, (B) Non-constant-time gradient-selective HSQCi of sample #1, (C) constant time phase-cycled HSQCi of sample #2.

As shown in Tables 1–3, the quantities of thiocoraline (1) in the extract of a Verrucosispora sp. were determined to be 1.12%, 1.17%, and 0.90% from the constant time phase-cycled HSQC_i, non-constant time gradient selective HSQC_i of sample #1, and constant time phase-cycled HSQC_i of sample #2, respectively. Thus, the quantity of thiocoraline (1) in the extract of a Verrucosispora sp. is estimated to be 1.0 % (w/w). This calculated concentration equates to ~8.5 mg of thiocoraline (1) produced per liter of medium cultured, in agreement with actual quantities of thiocoraline (1) isolated in the laboratory.

In this paper, we demonstrated how to use the newly developed HSQC₀ NMR approach to selectively quantify thiocoraline (1) in an extract from a Verrucosispora sp. isolated from the sponge Chondrilla caribensis f. caribensis without any purification. The HSQC₀ approach can be used to quantify a variety of natural products of interest at low concentrations (μM) in complex extracts even if they are difficult to isolate or purify. It can be used to standardize herbal extracts through quantification of the standard ingredients. Furthermore, it can be used in the selection or optimization of media for the production of (isotope labeled) compounds and for optimization of protocols for compound isolation and purification.

EXPERIMENTAL SECTION

Biological Material

Sponge specimens were collected on February 10, 2010, in the Florida Keys (24° 39′ 17.90″, 81° 17′ 51.09″). A voucher specimen for Chondrilla caribensis f. caribensis (FLK-10-4-24) is housed at the University of Wisconsin-Madison. Strain WMMA107 was isolated from the sponge Chondrilla caribensis f. caribensis (FLK-10-4-24) and was identified as a marine actinomycete, Verrucosispora sp. Details of the cultivation, purification and 16S ribosomal DNA sequencing^{16, 17} of the Verrucosispora sp. have been reported.¹⁵

Fermentation and Extraction

Strain WMMA107 was fermented in 25 × 150 mm culture tubes (4 × 10 mL) in medium ASW-A (20 g soluble starch, 10 g glucose, 10 g peptone, 5 g yeast extract, 5 g CaCO₃ per liter of artificial seawater) for one week at 28 °C. A 250 mL baffled flask containing 50 mL medium ASW-A with Diaion HP20 (4% by weight) was inoculated with 2 mL from culture tube and shaken at 200 RPM at 28 °C for one week. Filtered HP20 was washed with H₂O and extracted with 50 mL acetone for 2 hrs.

NMR Sample Preparation

The dry extract (40.0 mg) was dissolved in 3 mL CHCl₃. 600 μL (Sample #1) and 300 μL (Sample #2) aliquots were removed and dried (8.0 and 4.0 mg, respectively). Standard solutions of TMS were prepared by dilutions of a stock solution of TMS in CDCl₃ (0.147 mM). The NMR samples were prepared by dissolving Sample #1 and Sample #2 in 500 μL of CDCl₃ containing TMS at concentrations of 58.7 μM and 82.3 μM, respectively. 40 μL of 4% Cr(AcAc)₃ in CDCl₃ was added as a relaxation enhancing agent to each sample. NMR-grade CDCl₃ (99.8% D, Cambridge Isotopes Laboratory) was used for both samples.

NMR Experiments

All NMR data were collected at 25 °C on a Bruker Avance III 700 MHz spectrometer equipped with 5 mm QCI probe, with radio-frequency pulses applied on ¹H and ¹³C at 2.5 ppm and 23 ppm, respectively. GARP ¹³C decoupling used a field strength of γB₂ = 2.5 kHz. 2048 × 42 complex data points with spectral width of 16 ppm and 20 ppm, respectively, were collected along the ¹H and ¹³C dimensions, with 64 scans per FID and an inter-scan delay of 2.6 s (longer than 5 times of the longest proton T₁), resulting in a total acquisition time of 4 h 15 min for each HSQC_i. Measurement of proton T₁ values was conducted by using a standard inversion recovery pulse sequence (180°- τ −90°). For non-constant-time gradient-selective HSQC_i, the strength of the pulsed field gradients applied along the z-axis were: g₁ = 80%; g₂ = 20.1%; g₃ = 70%; g₄ = 17.5875%, g₅ = 60%; g₆ = 15.075%; g₁₃ = 70%; g₁₄ = 85%, g₁₅ = 15%; g₆ = 90% of the maximum of 53 G/cm; all with a duration of 1 ms followed by a gradient recovery period of 200 μs. Quadrature detection in the ¹³C (t₁) dimension was achieved by echo-antiecho by flipping the polarity of gradient g₁. For constant-time phase-cycled HSQC_i, the constant time T was set to 12 ms without applying gradients g₁ to g₆. Quadrature detection in the ¹³C (t₁) dimension was achieved using States-TPPI applied to the phase φ₂. Intensities of relevant peaks are manually integrated as described before⁷ and the data are shown in Table 1–3.

Table 2.

Manually integrated peak volume from Non-constant-time gradient-selective HSQC_i of sample #1

	csC	csH	HSQC1	HSQC2	HSQC3	Corr. Coef.	Scaling factor	A0	# of H	Normalized A0	Ave./Stdev	w/w % of thiocoraline
13,13′	15.16	2.136	2.64E+09	8.92E+08	3.82E+08	−0.9975	0.3804	6.67E+09	6	1.11E+09	1.26E+09
14,14′	30.77	3.059	2.70E+09	1.04E+09	3.08E+08	−0.9976	0.3377	8.35E+09	6	1.39E+09	11.16%	1.17%
16,16	30.615	3.012	2.58E+09	1.07E+09	3.27E+08	−0.9964	0.3560	7.63E+09	6	1.27E+09	Mol. of TMS
TMS	19.996	0	2.10E+09	7.77E+08	3.01E+08	−0.9999	0.3786	5.50E+09	12	4.59E+08	2.93584E-08

NMR sample #1: 58.7 μM TMS (2.59 μg in 500 μL) was added as concentration reference into the NMR sample solution with 8.0 mg extract. Molecular weights of TMS and thiocoraline are 88.22 and 1157.41, respectively.