The Conformations of 17β-Estradiol (E2) and 17α-Estradiol as Determined by Solution NMR

Abstract

The conformational structures of the hormone 17β-estradiol (E2) and the epimeric 17α-estradiol determined by solution NMR spectroscopy and restrained molecular dynamics calculations found a single low energy conformation.

Introduction

Estrogens are gonadal steroidal hormones that have important roles in reproductive function.^1,2 These amphipathic steroids are quite insoluble in an aqueous environment,^3,4 but circulate via the bloodstream⁵ signaling to a number of tissues including breast, ovaries, uterus, and brain.⁶ Estrogens initiate rapid nongenomic signaling events at cell membranes,^2,7-10 readily diffuse across membranes interacting with nuclear estrogen receptors that regulate gene expression,^6,11,12 and also act at mitochondrial membranes.¹³ Estrogen receptors are activated by the hormone 17β-estradiol (E2) but not by the 17α-estradiol isomer (Figure 1).^14,15 Recently, the structure of the estrogen hormone 17β-estradiol (E2) as determined by NMR and X-ray was reviewed.¹⁶ In addition to the low energy C-ring chair conformation that corresponded well to crystal structures, a second conformation with a twisted boat C-ring was proposed for E2 in dimethylsulfoxide solution.¹⁶

Figure 1.

The hormone 17β-estradiol (E2) and the isomeric antiestrogen 17α-estradiol.

In conjunction with our ongoing studies of steroids interacting with phospholipid bilayer model membranes,^17,18 we have also characterized the solution structures of 17β-estradiol (E2) as well as 17α-estradiol by NMR and restrained molecular dynamics calculations. Our data clearly demonstrate that for both E2 and the anomeric antiestrogen 17α-estradiol, there is only one conformation that exists in solution.

Results and discussion

We first assigned the proton resonances of 17β-estradiol (E2) as well as 17α-estradiol by analysis of their 1D ¹H-, ¹³C-, and 2D COSY, HSQC, and NOESY spectra recorded on a Bruker AVANCEII 700MHz NMR spectrometer (see supplementary data). All spectra were referenced to the DMSO-d₆ multiplets at 2.47 ppm and 39.50 ppm. The 2D HSQC experiment was particularly useful for distinguishing resonances of H_11β, H₈, H_15β and H_7α for E2 that were between 1.19 and 1.27 ppm (see Figure 2). We have also, for the first time, distinguished the two benzylic geminal 6α and 6β protons by analyzing the J couplings of this rather complicated ABXY pattern. Our analysis was further confirmed by comparing the ¹H- spectrum of E2-6β-d (Figure 3) which was synthesized in our lab according to a previously reported procedure.¹⁹

Figure 2.

¹H-¹³C HSQC spectra of 17β-estradiol (E2, left) and 17α-estradiol (right).

Figure 3.

¹H NMR of E2 (a) and E2-6β-d (b) in DMSO-d₆ solution. (a) The benzylic C-6 protons of E2 appeared as an ABXY pattern with the H_6β (ddd) downfield of the H_6α (ddd). (b) The deuterated analog, which contained approximately 10% non-deuterated E2, appeared as a broadened doublet with a 0.03 ppm upfield shift due to the isotope effect.

The observed proton chemical shifts, carbon chemical shifts, and the NOE intensities of key crosspeaks are tabulated for E2 and 17α-estradiol in Tables 1 and 2, respectively. Their observed J-couplings are listed in Table 3. Strong trans coplanar diaxial couplings were observed for J_6β,7α, J_7α,8, J_8,9, J_8,14, J₉,_11β, J_11β,_12α, and J_14,15β. The chair conformation of the C-ring was clearly evidenced by the characteristic large J_8,9, J_8,14, J_9,11β, and J_11β,12α couplings. The C-18 methyl group gave NOEs with H₈, H_11β, H_12β, H_15β and H_16β for both 17β-estradiol (E2) and 17α-estradiol, and NOE of the C-18 methyl with H₁₇ for 17α-estradiol was also observed (Figure 4). We did not, however, observe the reported¹⁶ NOE between H_12β and H_15β for E2 which was due to their mis-assignment of the proton resonances. Interestingly, the NOESY of H₁ with H_11α in the plane of the aromatic ring gave a negative NOE, an unusal effect.^20,21 This aromatic H₁ proton had the expected positive NOEs with H₉ and H_11β, below and above the plane, analogous to H₄ NOEs with H_6α and H_6β.

Table 1. 17β-Estradiol (E2).

Position	Chemical Shift (ppm)		1H-1H NOEs*
	δH	δC
1	7.00	126.0	2(m); 9 (w); 11β(w)
2	6.47	112.7	1(m)
3	—	154.8
4	6.39	114.9	6α(m); 6β(w)
5	—	137.1
6α	2.65	29.1	4(m); 7α(w)
6β	2.68		4(w); 7β(w); 8(w)
7α	1.19	26.9	6α(w); 7β(s); 9(m)
7β	1.74		6β(w); 7α(s); 8(m)
8	1.24	38.7	7β(m); 18(m)
9	2.03	43.5	7α(m); 11α(w); 14(w)
10	—	130.4
11α	2.19	26.0	9(w); 11β(s); 12α(w)
11β	1.27		11α(s); 18(m)
12α	1.13	36.5	11α(w); 12β(s); 14(w); 17(w)
12β	1.80		12α(s);18(w)
13	—	42.8
14	1.07	49.5	9(w); 12α(w); 15α(m)
15α	1.55	22.7	14(m); 15β(s); 16α(w); 16β(w)
15β	1.21		15α(s); 16α(w); 18(m)
16α	1.85	29.8	15α(w); 15β(w); 16β(s); 17(m)
16β	1.34		15α(w); 16α(s); 17(w); 18(w)
17	3.48	80.0	12α(w);16α(m); 16β(w)
18	0.63	11.2	8(m); 11β(m); 12β(w); 15β(m); 16β(w)

^*NOE intensities were categorized as “strong” (s), “medium” (m) and “weak” (w), and were converted into upper limit distance constraints of 2.7, 3.5, and 5.0 Å in the molecular dynamics calculations, respectively.

Table 2. 17α-Estradiol.

Position	Chemical Shift (ppm)		1H-1H NOEs*
	δH	δC
1	7.02	126.0	2(m); 9(w); 11β(w)
2	6.46	112.6	1(m)
3	—	154.8
4	6.39	114.8	6α(m); 6β(w)
5	—	137.1
6α	2.65	29.2	4(m); 7α(w)
6β	2.68		4(w); 7β(w); 8(w)
7α	1.26	27.8	6α(w); 7β(s); 9(m)
7β	1.78		6β(w); 7α(s); 8(m)
8	1.22	38.8	7β(m); 18(m)
9	2.02	43.3	7α(m); 11α(w); 12α(w); 14(w)
10	—	130.4
11α	2.23	26.0	9(w); 11β(s); 12α(w)
11β	1.29		11α(s); 18(m)
12α	1.71	31.4	9(w); 11α(w); 12β(s); 14(m)
12β	1.40		12α(s); 17(w); 18(w)
13	—	45.0
14	1.51	47.2	9(w); 12α(m); 15α(m)
15α	1.67	23.8	14(m); 15β(s); 16α(w); 16β(w)
15β	1.12		15α(s); 16α(w); 18(w)
16α	1.34	32.0	15α(w); 15β(w); 16β(s); 17(w)
16β	2.01		15α(w); 16α(s); 17(m); 18(w)
17	3.54	77.9	12β(w); 16α(w); 16β(m); 18(m)
18	0.58	16.9	8(m); 11β(m); 12β(w); 15β(w); 16β(w); 17(m)

^*NOE intensities were categorized as “strong” (s), “medium” (m) and “weak” (w), and were converted into upper limit distance constraints of 2.7, 3.5, and 5.0 Å in the molecular dynamics calculations, respectively.

Table 3. Observed coupling constants for 17β- and 17α-estradiol.

	J (Hz)		J (Hz)
J1,2	8.5	J1,2	8.4
J2,4	2.7	J2,4	2.6
J6α,6β	17.1	J6α,6β	17.1
J6α,7α	6.3	J6α,7α	6.3
J6α,7β	2.4	J6α,7β	2.4
J6β,7α	11.6	J6β,7α	11.3
J6β,7β	6.1	J6β,7β	6.2
J7α,7β	12.3	J7α,7β	12.6
J7α,8	12.0	J7α,8	12.0
J7β,8	2.3	J7β,8	2.5
J8,9	11.2	J8,9	11.2
J8,14	12.4	J8,14	12.2
J9,11α	4.3	J9,11α	4.3
J9,11β	11.2	J9,11β	11.0
J11α,11β	13.5	J11α,11β	13.4
J11α,12α	4.1	J11α,12α	4.3
J11α,12β	2.9	J11α,12β	2.7
J11β,12α	12.8	J11β,12α	13.2
J11β,12β	3.8	J11β,12β	4.0
J12α,12β	12.7	J12α,12β	13.0
J14,15α	7.4	J14,15α	7.4
J14,15β	10.8	J14,15β	10.8
J15α,15β	12.1	J15α,15β	12.1
J15α,16α	9.4	J15α,16α	9.5
J15α,16β	3.4	J15α,16β	2.9
J15β,16α	5.8	J15β,16α	6.6
J15β,16β	11.5	J15β,16β	12.2
J16α,16β	13.4	J16α,16β	14.3
J16α,17α	8.9	J16α,17β	0.0
J16β,17α	8.4	J16β,17β	5.8

Figure 4.

NOESY spectra of 17β-estradiol (E2, left) and 17α-estradiol (right).

Our NOE data in Tables 1 and 2 were used as distance restraints to determine the conformations of 17β-estradiol (E2) and 17α-estradiol, respectively. The calculations were performed using Macromodel in the Schrodinger software package. In order to fully sample the entire conformational space, stochastic dynamics simulations with NOE distance restraints were carried out at 1000K with a time step of 1.0 fs. Conformations were recorded every 250,000 steps for a total of 20 trajectories. Each conformer was further subject to conjugate gradient minimization until the maximum derivative was less than 0.05 kJ/mol. Figure 5a depicts the 20 conformers of 17β-estradiol (E2) and Figure 5b depicts the 20 conformers of 17α-estradiol. For each compound, all 20 conformers were closely superimposable over the entire backbone. The solution conformations of the anomeric steroids were in good agreement with reported X-ray structures^22-26 and included the following features. The aromatic A-ring was planar. The unsaturated B-ring had trans coplanar diaxial orientations for H_6β with H_7α and for H₈ with H₉ and was a half-chair (Figure 5c). The C-ring was a chair conformation with trans coplanar diaxial orientations for H_11β with H₉ and H_12α as well as for H₈ with H₉ and H₁₄ (Figure 5d). Thus, the conformations of the ABC rings of both anomers were nearly identical. The five-membered D-rings were both C-13 β-envelopes with only 3° of distortion (C18-C13-C17 was 110.3° for E2 and 107.3° for the α-anomer, Figures 5e and 5f, respectively) due to the steric interaction between the C-18 methyl group and the C-17 hydroxyl group of E2.

Figure 5.

(a) The 20 low-energy conformers of 17β-estradiol (E2). (b) The 20 low-energy conformers of 17α-estradiol. (c) The unsaturated B-ring half-chair showing trans coplanar diaxial orientations for H_6β with H_7α and for H₈ with H₉. (d) The C-ring chair with trans coplanar diaxial orientations for H_11β with H₉ and H_12α as well as for H₈ with H₉ and H₁₄. (e) The D-ring for 17β-estradiol (E2). (f) The D-ring for 17α-estradiol.

The conformational structures of the hormone 17β-estradiol (E2) and the anomeric antiestrogen 17α-estradiol have now been clearly established in DMSO solution using a combination of NMR spectroscopy and restrained molecular dynamics calculations. We have found that there is only one low energy conformation for theses steroids with a chair conformation for their C-rings. We have also, for the first time, assigned all resonances including the two benzylic geminal H_6α and H_6β protons. We are currently using these structural details to study the interactions of these compounds with bicelle model membranes.

Abbreviations

COSY

correlation spectroscopy

E2

17β-estradiol

HSQC

heteronuclear single quantum coherence

NMR

nuclear magnetic resonance

NOE

nuclear Overhauser effect

NOESY

nuclear Overhauser enhancement spectroscopy