Validation of trifluoromethylphenyl diazirine cholesterol analogues as cholesterol mimetics and photolabeling reagents

Kathiresan Krishnan; Mingxing Qian; McKenna Feltes; Zi-Wei Chen; Sarah Gale; Lei Wang; Yusuke Sugasawa; David E Reichert; Jean E Schaffer; Daniel S Ory; Alex S Evers; Douglas F Covey

doi:10.1021/acschembio.1c00364

. Author manuscript; available in PMC: 2022 Jan 7.

Published in final edited form as: ACS Chem Biol. 2021 Aug 6;16(8):1493–1507. doi: 10.1021/acschembio.1c00364

Validation of trifluoromethylphenyl diazirine cholesterol analogues as cholesterol mimetics and photolabeling reagents

Kathiresan Krishnan ¹, Mingxing Qian ¹, McKenna Feltes ², Zi-Wei Chen ^3,⁶, Sarah Gale ², Lei Wang ³, Yusuke Sugasawa ³, David E Reichert ^4,⁶, Jean E Schaffer ², Daniel S Ory ², Alex S Evers ^1,^2,^3,⁶, Douglas F Covey ^1,^3,^5,^6,^*

PMCID: PMC8740335 NIHMSID: NIHMS1765311 PMID: 34355883

Abstract

Aliphatic diazirine analogues of cholesterol have been used previously to elaborate the cholesterol proteome and identify cholesterol binding sites on proteins. Cholesterol analogues containing the trifluromethylphenyl diazirine (TPD) group have not been reported. Both classes of diazirines have been prepared for neurosteroid photolabeling studies and their combined use provided information that was not obtainable with either diazirine class alone. Hence, we prepared cholesterol TPD analogues and used them along with previously reported aliphatic diazirine analogues as photoaffinity labeling reagents to obtain additional information on the cholesterol binding sites of the pentameric ion channel GLIC. We first validated the TPD analogues as cholesterol substitutes and compared their actions with those of previously reported aliphatic diazirines in cell culture assays. All the probes bound to the same cholesterol binding site on GLIC, but with differences in photolabeling efficiencies and residues identified. Photolabeling of mammalian (HEK) cell membranes demonstrated differences in the pattern of proteins labeled by the two classes of probes. Collectively, these date indicate that cholesterol photoaffinity labeling reagents containing an aliphatic diazirine or TPD group provide complementary information and will both be useful tools in future studies of cholesterol biology.

INTRODUCTION

Cholesterol (1, Figure 1) is arguably the most important sterol in mammalian biology. It has a structural role whereby it modulates membrane properties and is crucial for the formation of microdomain signaling platforms for membrane proteins.^1–4 It is an enzyme substrate as a precursor for a wide range of steroids such as the steroid sex hormones, oxysterols, and bile steroids.⁵ It also has functional effects as a sterol with specific binding sites on proteins such as SCAP (sterol regulatory element binding protein cleavage-activating protein), the oncogene smoothened and nicotinic acetylcholine receptors.^6–9 Because of its very low solubility in water, cholesterol is trafficked in the cytosol by non-vesicular means, such as those that include sterol-binding proteins (e.g. StARD4 and StARD5), and within organelles through hydrophobic hand-offs, such as occurs via NPC1 and NPC2 (Niemann-Pick carrier proteins 1 and 2).¹⁰ Additional proteins that interact with cholesterol were identified in a proteomic study that utilized the cholesterol photoaffinity labeling (PAL) analogue trans-sterol (2, Figure 1).¹¹ Specific binding sites for cholesterol on voltage-dependent anion channel 1 and Gloeobacter ligand-gated ion channel (GLIC) have also been determined using the cholesterol PAL LKM38 (3, Figure 1).^12,13 The cholesterol PAL reagents used in these studies contain an aliphatic diazirine as the photoreactive group. Another potential class of cholesterol PAL analogues is the class containing a trifluoromethylphenyl diazirine (TPD) as the photoreactive group.¹⁴ Previous photolabeling studies using both classes of PAL analogues to identify neurosteroid binding sites on a GABA_A receptor provided information that was not obtainable with either diazirine class alone,^15,16 suggesting that photolabeling studies to identify binding sites for cholesterol on proteins will benefit from having both aliphatic diazirine and TPD cholesterol analogues available for use.

In this study, we report the synthesis of four TPD cholesterol analogues (Figure 1, compounds 5–8), characterize them for their ability to mimic the biological effects of cholesterol and their efficiency and specificity in photolabeling cholesterol binding sites on the bacterial pentameric ligand-gated ion channel, GLIC. We compare the results obtained with these new TPD cholesterol analogues to those of two previously studied aliphatic diazirine analogues (Figure 1, compounds 3 and 4) to elucidate the advantages that each type of diazirine has for proteomic and binding site identification studies. Finally, we photolabeled HEK (human embryonic kidney) cell membranes to determine if there were different patterns of protein photolabeling by the two classes of photolabeling reagents.

RESULTS

Chemistry.

The TPD group is large and it seemed unlikely that attaching this group anywhere on cholesterol (1) other than where the cholesterol iso-octyl group is attached would give an analogue useful for either biochemical or cell biological applications. Notably, less bulky structural modifications of the cholesterol side chain have given the useful cholesterol PAL probes trans-sterol (2) and LKM38 (3).^11,17 MQ182 (4) is another PAL probe with a modified C-17 substituent that was prepared and found to be a useful oxysterol analogue.¹⁸ Given the demonstrated utilities of PALs 2–4, we chose to attach the TPD group at C-17 by either an ether linkage as in KK226 (5), KK231 (6), KK237 (7) or by incorporating it into an alkyl chain as in MQ238 (8).

KK226 (5) was prepared from steroid 9¹⁹ in three steps (Scheme 1). NaBH₄ reduction of steroid 9 gave 17β-hydroxysteroid 10. O-alkylation of steroid 10 with ICH₂TPD gave steroid 11 and removal of the MOM protecting group from steroid 11 gave KK226 (5).

KK231 (6) was prepared from steroid 12²⁰ in seven steps (Scheme 2). NaBH₄ reduction of steroid 12 gave 17β-hydroxysteroid 13 and the Δ⁴-3-ketone group of steroid 13 was converted to the dienol acetate of steroid 14 which upon NaBH₄ reduction gave steroid 15. MOM protection of the 3β-OH group of steroid 15 gave steroid 16 and hydrolysis of the acetate group of steroid 16 gave steroid 17. O-alkylation of steroid 17 with ICH₂TPD gave steroid 18 and removal of the MOM protecting group from steroid 18 gave KK231 (6).

KK237 (7) was prepared from steroid 19²¹ in eight steps (Scheme 3). Diethyl malonate was added to steroid 19 and then treated with DABCO at reflux in p-xylene to obtain steroid 20. The 3β-acetate was hydrolyzed to obtain steroid 21 and the resultant 3β-OH group was protected subsequently as a silyl ether to obtain steroid 22. The carboethoxy group was then reduced with LiAlH₄ to obtain steroid 23. The primary alcohol group of steroid 23 was oxidized to the aldehyde group and immediately treated with Bestmann-Ohira reagent²² to obtain the alkyne group of steroid 24. Removal of the 17-ketal group from steroid 24 using PTSA in acetone also removed the silyl ether protecting group and the 3β-OH group had to be protected again as the silyl ether to obtain steroid 25. NaBH₄ reduction of the 17-ketone group of steroid 24 gave 17β-hydroxysteroid 26. Finally, O-alkylation of steroid 26 with ICH₂TPD and removal of the silyl ether protecting group using TBAF gave KK237 (7).

MQ238 (8) was prepared from steroid 9 in 10 steps (Scheme 4). Following a synthetic strategy developed by Wicha and Bal for the preparation of side-chain oxysterols of cholesterol, ²³ steroid 9 was subjected to a Wittig reaction to obtain steroid 27. Hydrogenation of steroid 27 and removal of the MOM group gave, after separation of a minor side-product in which the Δ⁵-double bond was also reduced, steroid 28 (the purification procedure for steroid 28 is described in detail in the Methods). The 3β-OH of steroid 28 was protected as a silyl ether to obtain steroid 29 which was subsequently alkylated with ICH₂TPD to obtain steroid 30. Reduction of the carboethoxy group of steroid 30 gave steroid 31 and oxidation of its 21-hydroxyl group gave steroid 32. Treatment of steroid 32 with Bestmann-Ohira reagent give steroid 33 and removal of the 3β-silyl ether protecting group using TBAF gave MQ238 (8).

Cellular Actions of Cholesterol PALs.

Cholesterol levels in the endoplasmic reticulum (ER) regulate cellular cholesterol levels through transcriptional regulation.²⁴ When ER cholesterol levels are elevated, expression of key genes involved in the cholesterol synthesis and uptake pathways are suppressed. One of these gene targets is HMG-CoA synthase (HMGCS), a key enzyme for the de novo synthesis of cholesterol and isoprenoids as the second enzyme of the mevalonate pathway. Addition of exogenous cholesterol leads to elevated ER cholesterol and subsequent suppression of HMGCS expression (Figure 2). Similarly, compounds KK226, KK231, KK237, MQ238 and MQ182 were as effective as cholesterol in suppressing HMGCS expression. The previously characterized aliphatic diazirine probe, LKM38, was also shown to suppress HMGCS mRNA expression as effectively as cholesterol in U2OS-SRA (U2OS-scavenger receptor type A cells).¹⁷

Figure 2. — U2OS-SRA cells were treated with 10 μM cholesterol, KK226, KK231, KK237, MQ238, or MQ182, or an equal volume of ethanol (V) in cholesterol starvation media. After 14 h RNA was harvested for quantification of HMGCS RNA relative to Rplp0 (ribosomal protein lateral stalk subunit P0) by qPCR (quantitative PCR) . Values are normalized to the cholesterol treated condition. Bars: mean + SE, n = 3 experiments/condition. *, P<0.05 vs. vehicle, #, P<0.05 vs. cholesterol, unpaired t-test.

Similar to cholesterol, oxysterols like 25-hydroxycholesterol, also act at the ER to suppress expression of HMGCS and other cholesterol synthesis and uptake genes.²⁵ Unlike cholesterol, 25-hydroxycholesterol and several other oxysterols are potent agonists of liver X receptors (LXR). To determine if KK226, KK231, KK237, MQ238 and MQ182 effects on HMGCS expression can function as oxysterol analogues we evaluated their ability to activate LXR by monitoring expression of ABCA1, a cholesterol transporter and LXR target (Figure 3). Analogue activity was compared to that of 25-hydroxycholesterol. None of the TPD-containing analogues (KK226, KK231, KK237, MQ238) significantly increased ABCA1 expression relative to vehicle treatment. The aliphatic diazirine probe MQ182 had a weak stimulatory effect similar to that of cholesterol. Similar to the TPD-containing analogues, the previously studied aliphatic diazirine probe LKM38 was not an LXR agonist.¹⁷

Figure 3. — U2OS-SRA cells were treated with 10 μM cholesterol, KK226, KK231, KK237, MQ238, or MQ182, or an equal volume of ethanol (V) in cholesterol starvation media. After 14 h, RNA was harvested for quantification of ABCA1 RNA relative to Rplp0 by qPCR. Values are normalized to the cholesterol treated condition. Bars: mean + SE, n = 3 experiments/condition. *, P<0.05 vs. vehicle, #, P<0.05 vs. cholesterol, unpaired t-test.

As a final characterization assay, we tested the five aforementioned analogues and LKM38 for their ability to support growth when provided as the sole sterol source for U2OS-SRA cells made auxotrophic for cholesterol by growth in cholesterol-deficient media in the presence of an HMG-CoA reductase inhibitor. Two of the TPD-containing analogues, KK231 and KK237, did not support cell growth and were toxic to the cells. Two of the TPD-containing analogues, KK226 and MQ238, supported cell growth as effectively as cholesterol. The two aliphatic diazirines, MQ182 and LKM38, supported cell growth but were not as robust as cholesterol in doing so. The LMK38 activity shown here was similar to that reported earlier.¹⁷

Probe photolabeling of the Intact Ion Channel GLIC by MS.

To compare the specificity and efficiency with which the various cholesterol analogue photolabeling reagents label cholesterol binding sites on a model protein, we photolabeled GLIC, a bacterial pentameric ligand-gated ion channel in which we have previously identified neurosteroid and cholesterol binding sites.^13,26 Notably, bacteria do not contain cholesterol and the cholesterol binding sites on GLIC are likely occupied by structurally similar hopanoids.²⁷ Purified GLIC was photolabeled, with 100 μM LKM38, MQ182, MQ238, KK226, KK231, or KK237 followed by intact protein MS.²⁶

Deconvolution of the spectra identified a feature of 36,535 Da representing the intact non-photolabeled GLIC monomer (Fig. 5A). An additional feature of 36,932Da (labeled with *) was observed in the LKM38 photolabeled GLIC sample. The difference of 397 Da corresponds to the addition of one LKM38 molecule after nitrogen loss to the GLIC monomer (Fig. 5A). Under the conditions employed for photolabeling (100 μM LKM38, 5 min of UV irradiation), each GLIC monomer was labeled with an efficiency of 57% (Fig. 5B red). Similarly, 100 μM MQ182 photolabeled GLIC with an observed stoichiometry of one per monomer (Fig. 5A), and a photolabeling efficiency of 30% (Fig. 5B blue). A feature representing a singly labeled GLIC monomer was also observed with the TPD-containing cholesterol analogue photolabeling reagents KK226, KK231 and KK237. In contrast, no photolabeled feature was detected by this technique with MQ238. Overall, the TPD-containing cholesterol analogues photolabeled GLIC with lower efficiency than the reagents containing an aliphatic diazirine labeling group. The observed photolabeling efficiencies for KK226, KK231 and KK237 were 3%, 16% and 11%, respectively (Figure 5B).

Figure 5. — A) Deconvoluted intact protein mass spectra of non-photolabeled GLIC and GLIC photolabeled with 100 μM LKM38, MQ182, MQ238, KK226, KK231 and KK237. The non-photolabeled GLIC monomer has a MW of 36,535 Da. The photolabeled GLIC monomer with addition of the adduct weight of each photolabeling analogue is marked with *. B) The bar graph shows the photolabeling efficiency of each of the cholesterol analogue photolabeling reagents from three replicate experiments. Bars represent mean + standard deviation. C) Fluorescence image of an SDS-PAGE showing GLIC photolabeled with each of the cholesterol analogue photolabeling reagents (100 μM) and coupled to TAMRA-azide via click chemistry (top). The SyproRuby stained gel (below) shows that each lane was loaded with a similar amount of GLIC. D) The bar graph shows the densitometric analysis of the TAMRA fluorescence in the labeled GLIC bands from three replicate experiments.

Fluorescence Analysis of Photolabeled GLIC.

We also examined the efficiency of photolabeling utilizing the alkyne moiety on the various cholesterol analogue photolabeling reagents to attach a fluorescent tag via click chemistry. In these experiments, purified GLIC (10 μg) was photolabeled with each of the photolabeling reagents and coupled to TAMRA-azide (5-carboxytetramethylrhodamine-azide) via the Huisgen copper-catalyzed cycloaddition reaction. The protein was analyzed by SDS-PAGE and photolabeled bands were visualized and quantitated using fluorescence imaging. Imaging (Figure 5C) showed fluorescent bands of GLIC monomer at ~35 kDa indicating photolabeling by LKM38, MQ182, KK231, KK237 and MQ238. For MQ182, photolableing was prevented by a 10-fold excess of either cholesterol or cholesterol hemisuccinate, consistent with labeling of a specific cholesterol binding site (Supplemental Figure S1). As described below, all of the PAL reagents labeled the same site on GLIC and we infer that this MQ182 sterol protection experiment provides a specificity control for all the probes. Additionally, cholesterol protection of photolabeling of this site by LKM38 was shown previously.¹³ Quantitation of the fluorescent bands at 35 kDA (Figure 5D) gave similar results to the analysis of GLIC labeling using intact protein MS. The highest fluorescence density was observed with LKM38 followed by MQ182. The three TPD-containing analogues (KK231, KK237 and MQ238) showed lower density. No fluorescent band was observed for GLIC labeled with KK226, consistent with the absence of an alkyne group for the click reaction. SyproRuby staining of GLIC protein indicated that similar amounts of GLIC were loaded in each lane of the gel (Figure 5C). Quantitative densitometry of TAMRA fluorescence from three replicate experiments is shown in Figure 5D.

Identification of Sites of Cholesterol Analogue Adduction Using Middle-down MS.

Two LKM38 photolabeled residues were identified in GLIC in our previous studies: Glu272 in TM3 and Phe317 in TM4.¹³ To determine the residues photolabeled by MQ182, MQ238, KK226, KK231 and KK237, GLIC was photolabeled with 100 μM concentration of each reagent, digested with trypsin and analyzed by middle-down LC-MS as previously described.²⁶ Using this method we obtained 100% sequence coverage of the GLIC TMDs (Supplementary Figure S2). The two photolabeling reagents with an aliphatic diazirine labeling group (LKM38 with a C7 diazirine and MQ182 with a diazirine in the aliphatic tail) labeled the same residues: Glu272 in TM3 and Phe317 in TM4. In contrast, a cluster of adjacent residues in TM4 (Val302, Phe303 and Leu304) were identified as the adducted residues of the four TPD-containing photolabeling reagents (KK226, KK231, KK237 and MQ238). While no photolabeled feature was detected for MQ238 in the intact protein MS analysis, middle-down MS provides more sensitive detection of protein photolabeling thereby allowing identification of the site of MQ238 adduction. Consistent with the intact protein MS data, the efficiency of photolabeling was higher for LKM38 and MQ182 than for KK226, KK231, KK237 and MQ238. Notably the aliphatic diazirine reagents label two spatially disparate residues whereas the TPD reagents label only a single residue. This discrepancy is likely due to the difference in efficiency of labeling between the aliphatic diazirine reagents and the TPD-containing reagents. The residues labeled by the TPD-containing reagents are proximal (see below) to the Glu272 residue labeled with high efficiency by LKM38 and MQ182 (See Table 1 for photolabeling efficiencies at each labeled residue. Note that the modest discrepancies between the efficiencies of labeling determined by intact protein MS and middle down MS may be attributable to effects of the adduct on chromatographic peptide recovery or peptide ionization in the middle down analyses). The TM4 residue (Phe317) labeled by the aliphatic diazirines is labeled with lower efficiency, suggesting either lower affinity binding or less efficient photochemical adduction. The less efficient TPD-containing cholesterol PAL reagents may thus not label residues near Phe317 sufficiently to exceed the lower limit of detection. (The product ion spectra and fragmentation tables used to identify the adducted residues for all photolabeling reagents are shown in Supplementary MS2 Spectra Table S1 and Supplementary Figures S3 – S10).

Table 1.

Residue level photolabeling efficiency of cholesterol analogue photolabeling reagents using middle-down MS analysis.

Cholesterol analogues (photolabeled residues)	KK226 (Val302/Phe303)	KK231 (Val302)	KK237 (Leu304)	MQ238 (Leu304)	LKM38 (Phe317)	LKM38 (Glu272)	MQ182 (Phe317)	MQ182 (Glu272)
photolabeling efficiency (%, mean ± STD)	4.24 ± 1.06	16.66 ± 1.16	5.35 ± 3.04	0.30 ± 0.23	0.09 ± 0.03	34.05 ± 1.51	2.17 ± 0.46	18.0 ± 1.69
n	4	2	2	2	2	3	2	2

Open in a new tab

Mapping of the Cholesterol Photolabeling Sites on the GLIC Structure.

The X-ray structure of GLIC (PDB 4F8H) was used to visualize the location of the residues labeled with high efficiency in TM3 and TM4 (Table 1). As shown in Fig. 6, Glu272 (colored green), the residue labeled by LKM38 and MQ182, is located on the cytoplasmic end of TM3 (right, orange), facing TM4. Within the same subunit, the TPD-containing analogues labeled Val302 (yellow), Phe303 (purple), and Leu304 (blue), located in TM4 (left, orange). These two groups of photolabeled residues face each other within the same subunit, suggesting that they represent two sides of an intrasubunit cholesterol binding site, analogous to the pregnenolone sulfate binding site observed in the structure of an α₁ GABA_A subunit/GLIC chimeric protein.²⁸

Figure 6. — Docking poses of cholesterol (A), LKM38 (B) and KK231 (C). The TM3 and TM4 domains are labeled as orange and TM1 and TM2 domains are white. The Glu272 residue in TM3 that is photolabeled by LKM38 and MQ182 is shown in green. The Val302, Phe303 and Leu304 residues in TM4 photolabeled by the TPD-containing analogues are shown in yellow, purple and blue respectively. The * and * indicate the location of the diazirine photolabeling groups on LKM38 and KK231 respectively. Docking of cholesterol to a frame containing all the transmembrane domains yielded a cluster of poses. The preferred pose is shown in Figure 6A. Cholesterol is located between TM3 and TM4 of the same subunit, with the C3-hydroxyl group pointing to the cytoplasmic end of the TMDs and the cholesterol *iso*-octyl tail pointing to the center of the TMDs. C7 of cholesterol is adjacent to Glu272 in TM3 and the aliphatic tail is proximate to Val302 and Phe303 in TM4. These results are consistent with an intrasubunit cholesterol binding site between TM3 and TM4 with the sterol A ring facing toward the cytoplasm. LKM38 and KK231 were also docked to the same GLIC model. They both docked in the intrasubunit site with an orientation similar to cholesterol (Figures. 7B, C). The C7 diazirine (*) of LKM38 faces the labeled Glu272, and the TPD group in the aliphatic tail of KK231 (*) faces the labeled residues Val302 and Phe303.

To determine the potential utility of the cholesterol analogue photolabeling reagents for identifying cholesterol binding proteins in proteomic studies, we labeled membranes from HEK cells with a 10 μM concentration of each of the photolabeling reagents. Photolabeled proteins were coupled to TAMRA-azide using the Huisgen cycloaddition reaction and the labeled protein bands were visualized with fluorescent imaging following SDS-PAGE (Figure 7). Multiple bands were labeled by LKM38, MQ182, KK237 and MQ238. As expected, no labeled bands were observed with KK226, due to the absence of the alkyne required for coupling to TAMRA-azide. Labeled bands were also absent with KK231, which labels purified GLIC, but is toxic to live cells. Each of the photolabeling reagents showed a distinct labeling pattern, consistent with their structural differences. The aliphatic diazirine reagents (LKM38, MQ182) labeled a similar set of protein bands, as did the TPD-containing reagents (KK237, MQ238). While all four reagents label some common bands, there are numerous bands labeled by the TPD reagents that are not labeled by the aliphatic diazirines (grey arrows) and a few bands labeled by the aliphatic diazirines that are not labeled by the TPD-reagents (red arrows). There were also some marked differences in the efficiency with which specific bands were labeled (band intensity) by the two aliphatic diazirines (e.g. Figure 7; red arrow at ~38 kDa) and the two TPD reagents (e.g. Figure 7; grey arrow at ~55 kDa).

Figure 7. — 75 μg of HEK cell membrane protein was incubated with 10 μM concentration of each PAL and irradiated with UV light. The photolabeled proteins were coupled to TAMRA dye, separated by SDS-PAGE and visualized with fluorescence imaging. No labeled protein bands were observed with KK226 or KK231. The aliphatic diazirine PALs, LKM38 and MQ182, labeled a common set of bands with some differences in intensity of labeling (e.g. lower red arrow). The TPD-PALs, KK237 and MQ238, also labeled a common set of bands with some differences in intensity of labeling. There were notable differences between the bands labeled by the aliphatic diazirine PALs and the TPD-PALs. Red arrows illustrate bands labeled by the aliphatic diazirine PALs, but not the TPD reagents; grey arrows indicate bands preferentially labeled by the TPD-PALs.

CONCLUSION

Discussion.

A previously observed difference in the types of amino acids preferentially photolabeled by aliphatic and trifluoromethylphenyl diazirines of neurosteroids was the motivation for preparing cholesterol TPD reagents. Upon photoactivation, rearrangement of aliphatic diazirines to aliphatic diazo compounds is a favorable reaction that competes with nitrogen loss and carbene formation.¹⁴ These aliphatic diazo compounds will only react with nucleophilic amino acids. This reactivity profile together with the long-lived (relative to rates of diffusion) lifetimes of diazo compounds allows these reactive intermediates to reorient in a binding site until an amino acid with a nucleophilic side chain is encountered. This likely explains why in our previous studies with aliphatic diazirines of neurosteroids in which the diazirine group was located in different locations on the neurosteroid all of these PALs reacted with the same binding site glutamate residue.²⁶ By contrast, we found that TPD neurosteroids reacted with hydrophobic residues and did not reorient in their binding sites. Photolabeling of hydrophobic residues is attributable to the highly reactive carbene group and the maintenance of binding orientation to the carbene’s short lifetime (relative to rates of diffusion). Under the photolabeling conditions used, diazo compounds formed during photoactivation of the TPD group would be unreactive.²⁹ Based on these previous results with the neurosteroid PALs, we expected that the successful preparation of cholesterol TPD PALs would expand upon the information obtainable with the previously prepared aliphatic cholesterol PAL reagents.

To obtain the TPD containing PAL analogues the cholesterol iso-octyl side chain was replaced with a side chain containing the TPD group. We initially chose to accomplish this by incorporating a steroid 17β-hydroxyl group into an ether linkage to a side chain containing the TPD group. This strategy afforded the cholesterol PAL analogue KK226 which was then characterized for its ability to act as a cholesterol mimic.

Relative to cholesterol, neither the ether oxygen nor the bulk of the TPD group adversely affected the ability of KK226 to suppress HMGCS expression. In this regard, the actions of KK226 were similar to those of the aliphatic diazirines LKM38 and MQ182. Next, we assessed the activity of KK226 as an oxysterol analogue instead of a cholesterol analogue. The presence of the oxygen atom in the ether linkage of the side chain of KK226 did not confer oxysterol activity for this analogue as it failed to act as an LXR agonist by inducing ABCA1 expression. By contrast, the aliphatic diazirine MQ182, which has the same ether linkage in its side chain, had some activity as an LXR agonist. This likely explains why MQ182 behaves like an oxysterol modulator of the NMDA class of glutamate receptors.¹⁸ As reported earlier LKM38, which does not have an ether linked side chain, is not an LXR agonist.¹⁷

In the U2OS-SRA cell line made auxotrophic for cholesterol, KK226 was as effective as cholesterol in maintaining cell viability. KK226 was also better able to support cell growth than the aliphatic diazirines LKM38 and MQ182. It is notable that LKM38, which has a side chain more similar to that of cholesterol than the ether linked side chains of KK226 and MQ182, was not the best analogue for maintaining cell viability. The diazirine substituent at the C7 position in LKM38 is the reason for this outcome as the LKM38 analogue without the diazirine group is as effective as cholesterol in maintaining cell viability.¹⁷

PALs containing an alkyne group enable click chemistry to be performed with these probes, thereby increasing their utility as imaging agents, proteomic analysis tools and photolabeling site identification reagents.^11,15,30,31 Hence, we further modified KK226 to incorporate an alkyne group in the probe. We installed a propargyl group either at steroid position C19 (KK231) or C7 (KK237). Both of these probes had activities similar to KK226 in the HMGCS and LXR cellular expression assays, but unlike KK226, they were toxic to cells in the cholesterol auxotrophy assay. This cellular toxicity of KK231 and KK237 limits the utility of these PALs for in vivo studies (e.g. cholesterol trafficking studies or live cell proteomic studies) where cell viability is an important consideration. The inability of KK231 to photolabel proteins in HEK cell membranes, suggests it does not orient in membranes in the same way as cholesterol. It further suggests that its cellular toxicity may be based on membrane disruption.

In view of the cellular toxicity of KK231 and KK237 we concluded that it was unlikely that placing an alkyne substituent anywhere but in the C17 side chain would yield a clickable TPD PAL that would not be toxic to cells. For this reason, we prepared MQ238. Elimination of the ether linkage in this analogue enabled replacement of what would be the 21-methyl group of cholesterol with an ethynyl group. A −CH₂TPD group is attached though a carbon-carbon bond to what would normally be C20 in cholesterol. As desired, MQ238 suppressed HMGCS expression, was not an LXR agonist and was as efficacious as cholesterol in the auxotrophy assay. Overall, MQ238 satisfies all of the criteria we used to judge its suitability for cholesterol trafficking studies.

GLIC, an ion channel with transmembrane cholesterol binding sites previously identified with LKM38,¹³ was photolabeled with the TPD reagents KK226, KK231, KK237, MQ238 and aliphatic diazirines MQ182 and LKM38 to compare the specificity and efficiency of the two diazirine classes. Except for KK226, which does not contain an alkyne group, photolabeling of intact GLIC was detectable by a gel assay after a dye was attached covalently to the probes by the click reaction. Photolabeling of GLIC was analyzed by two LC-MS techniques. Initially, intact protein MS was used to determine the stoichiometry and efficiency of photolabeling of intact GLIC. With the exception of MQ238, for which photolabeling was not detected by this technique, the probes in the study photolabeled GLIC with a stoichiometry of one steroid adducted per monomer of GLIC. The aliphatic diazirines LKM38 and MQ182 photolabeled GLIC with higher efficiencies (57% and 30%, respectively) than the TPD probes KK226 (3%), KK231 (16%), KK237 (11%). It is not possible to fully explain the lower photolabeling efficiency of the TPD probes, but the identity of the photolabeled amino acid residues suggests that the formation of diazo compounds during photoactivation of the aliphatic diazirines LKM38 and MQ182 is a major factor underlying their greater labeling efficiencies. Inherent differences in probe affinity at these cholesterol binding sites may also be a contributing factor.

Photolabeled GLIC was digested with trypsin and the GLIC transmembrane domains were analyzed by middle-down LC-MS to identify the photolabeled amino acid residues. The aliphatic diazirines LKM38 and MQ182 both photolabeled Glu272 in TM3 with high efficiency (34 and 18% respectively) and Phe317 in TM4 with much lower efficiency (0.09 and 2.2% respectively). The ten to 100 fold higher efficiency of photolabeling of Glu272 (compared to Phe317) is likely due to the nucleophilic displacement of a diazo group generated during photoactivation of the aliphatic diazirine group by the Glu272 carboxylic acid group. Photolabeling of Phe317 results solely from carbene photolabeling since the aromatic ring of Phe is not a nucleophile. The observation that LKM38 labels Glu 272 with higher efficiency than does MQ182, whereas MQ182 labels Phe317 with higher efficiency than LKM38 may be related to the preferred orientation of the probes in the two binding pockets (see below) adjacent to the labeled residues.

Neither Glu272 nor Phe317 was photolabeled with the TPD probes KK226, KK231, KK237 or MQ238. Instead, all of these probes labeled the clustered hydrophobic residues Val302/Phe303 (KK226), Val302 (KK231) and Leu304 (KK237 and MQ238) in TM4. Since the TPD group is similarly located in all four of these probes, it is not surprising that they all photolabel the same cluster of hydrophobic residues. Photolabeling of these non-nucleophilic residues can only occur via a reactive carbene. Differences in photolabeling efficiencies between the four TPD probes can be attributed to probe affinity for the cholesterol binding site and possibly chemical activity of the probes in the detergent micelles used in the labeling experiments. The markedly lower photolabeling efficiency of MQ238 relative to that of the other three TPD probes suggests that the ethynyl group in MQ238 decreases its affinity for the intrasubunit cholesterol binding site formed by TM3 and TM4 in GLIC.

Docking studies performed with the X-ray structure of GLIC present a coherent picture of the photolabeling results, with the aliphatic diazirine of LKM38 photolabeling the adjacent Glu272 in TM3 and the TPD probes photolabeling the adjacent cluster of hydrophobic amino acid residues in TM4. MQ182, in contrast, labeled Glu272, despite the predicted position of its diazirine labeling group near Val302/Phe303. The fact that LKM38 and MQ182 both label Glu272 suggests that the aliphatic diazirines generate a photoreactive intermediate that is relatively long-lived and can diffuse in its binding site to label the nearest nucleophilic amino acid.³² Photolabeling of Phe317 in TM4 by LKM38 and MQ182 was not modeled, but photolabeling of this amino acid may indicate occupancy of a second cholesterol binding pocket near the exoplasmic surface of TM4.

The new cholesterol analogue PALs contain an alkyne to facilitate their use in proteomic studies to identify cholesterol-binding proteins and cholesterol tracking studies. To assess their utility in proteomic studies, HEK cell membranes were photolabeled and adducted proteins were coupled to a fluorescent tag via alkyne-azide click chemistry. Patterns of protein labeling were visualized with fluorescent imaging of SDS-PAGE gels. No fluorescent bands were observed with KK226, which does not contain an alkyne, or KK231, which may be membrane disruptive. Each of the PALs (LKM38, MQ182, KK237 and MQ238) showed unique patterns of protein labeling with differences in both the bands labeled and band intensity. The aliphatic diazirine PALs showed a pattern of labeling that was more similar to each other than to the TPD reagents and vice versa. These data indicate that the cholesterol binding proteins labeled by various PALs are dependent on both the specific structure of the reagent and the photochemical mechanisms of labeling.

In summary, we have prepared four new cholesterol PALs containing the TPD group and compared their actions with two previously reported cholesterol PALs containing aliphatic diazirines. All probes in both chemical classes mimicked cholesterol’s ability to suppress HMGCS in U2OS-SRA cells. Except for MQ182, which had some oxysterol-like agonist activity at LXR receptors in these cells, the compounds did not induce ABCA1 expression, indicating that HMGCS suppression was not due to the compounds acting as oxysterol agonists at LXR receptors. Two of the TPD probes (KK231 and KK237) were toxic to cells. Their toxicity might be due to their effects on the integrity of the plasma membranes of the cells as the location of the propargyl groups in these analogues might interfere with lipid packing in membranes. Biophysical lipid packing studies will be needed to address this possibility.

We investigated the photolabeling of GLIC by the compounds and found differences in the residues photolabeled by the TPD and aliphatic diazirine PALs. We observed a strong preference for photolabeling of Glu272 in TM3 by the aliphatic diazirines and for photolabeling of hydrophobic residues in TM4 by the TPD containing PALs. We hypothesize that rearrangement of the photoactivated aliphatic diazirines to reactive diazo compounds contributed to their efficient covalent modification of Glu272. We found that photolabeling efficiency of the TPD containing PALs was less than that of the aliphatic diazirines. Most notably MQ238, which appeared to be the best cholesterol mimetic in the cell-based assays and photolabeled multiple proteins in HEK cell membranes, was very inefficient at photolabeling GLIC, an effect we attributed to its low affinity for cholesterol binding sites on GLIC caused by the position of the ethynyl group in this probe.

Differences in the photochemistry of cholesterol TPD and aliphatic diazirines led to complimentary information obtained with the two classes of PALs. The aliphatic diazirines had a preference for labeling an acidic amino acid. This may limit their ability in future studies to identify all cholesterol binding proteins since an acidic amino acid or other nucleophilic residue may be absent from some cholesterol binding pockets. Additionally, the long life-time of diazo photo-intermediates may limit their precision in identifying residues in cholesterol binding pockets. By contrast, the TPD reagents labeled hydrophobic residues in the cholesterol binding pocket and the short life-time of the photo-generated carbene more precisely pinpointed their orientation in the binding pocket. Since binding pockets for cholesterol on different proteins are non-identical, it can be anticipated that a family of cholesterol PALs containing both the aliphatic diazirine and the TPD group will be necessary to define and characterize the cholesterol proteome. This assertion is supported by the labeling of HEK membranes which showed reagent-specific protein labeling patterns with substantial differences between the aliphatic diazirine and TPD PAL reagents.

Outlook for future studies.

Cholesterol PAL reagents vary in their protein binding specificity and efficiency, their ability to precisely map binding residues in a cholesterol binding pocket and as biological mimics of cholesterol. As such, there is no ideal cholesterol-analogue PAL. The aliphatic diazirine and TPD reagents described here each have unique attributes for proteomic, cholesterol trafficking or binding site mapping studies. Because of their different reactivity and specificity profile, the TPD-alkyne cholesterol analogue PALs, KK226, KK231, KK237 and MQ238 represent a substantial addition to the armamentarium of previously reported cholesterol PALs.³³ Photolabeling of hydrophobic residues by these new reagents may lead to the identification of additional cholesterol binding proteins that have not been identified previously because of the absence of nucleophilic residues in their cholesterol binding sites.

METHODS

Synthetic procedures. (3β,17β)-17-[[4-[3-(Trifluoromethyl)-3H-diazirin-3-yl]phenyl]methoxy]-androst-5-en-3-ol (KK226, 5).

Steroid 11 (80 mg, 0.15 mmol) dissolved in MeOH containing 10-15% dry HCl (12 mL) was stirred at room temperature for 7 h. The reaction was made basic by addition of aqueous satd. NaHCO₃ and extracted into CH₂Cl₂ (3 x 70 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 15-30% EtOAc in hexanes) to give KK226 (5) as an off-white solid (58 mg, 79%): ¹H NMR (400 MHz, CDCl₃) δ 7.37 (d, J = 8.1 Hz, 2H), 7.17 (d, J = 8.1 Hz, 2H), 5.35 (b s, 1H), 4.55 (s, 2H), 3.52 (m, 1H), 3.40 (t, J = 8.3 Hz, 1H), 2.40–0.90 (m), 1.03 (s, 3H), 0.85 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 141.3, 140.8, 127.9, 127.4 (2 x C), 126.4 (2 x C), 122.1 (q, J = 275 Hz), 121.3, 88.6, 71.7, 70.8, 51.5, 50.2, 42.9, 42.3, 37.7, 37.2, 36.5, 31.7, 31.6, 31.5, 27.8, 23.4, 20.7, 19.4, 11.7.

(3β,17β)-10-(2-Propyn-1-yl)-17-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methoxy]-estr-5-en-3-ol (KK231, 6).

Steroid 18 (63 mg, 0.11 mmol) dissolved in MeOH containing 10-15% dry HCl (12 mL) was stirred at room temperature for 7 h. The reaction was made basic by addition of aqueous satd. NaHCO₃ and the product was extracted into CH₂Cl₂ (60 mL x 3). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 15-30% EtOAc in hexanes) to give KK231 (6) as an off-white solid (44 mg, 76%): ¹H NMR (400 MHz, CDCl₃) δ 7.37 (d, J = 8.2 Hz, 2H), 7.17 (d, J = 8.2 Hz, 2H), 5.55 (b s, 1H), 4.55 (s, 2H), 3.56 (m, 1H), 3.39 (t, J = 8.2 Hz, 1H), 2.60–0.85 (m), 0.91 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 141.3, 137.6, 127.9, 127.4 (2 x C), 126.4 (2 x C), 124.1, 122.1 (q, J = 275 Hz), 88.6, 83.0, 71.4, 70.9, 70.8, 52.0, 51.1, 43.1, 42.3, 39.1, 38.0, 35.9, 32.0, 31.7, 31.0, 29.7, 27.9, 23.4, 22.1, 21.8, 11.9.

(3β,7β,17β)-7-(2-Propyn-1-yl)-17-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methoxy]-androst-5-en-3-ol (KK237, 7).

NaH (60% suspension in mineral oil, 400 mg, 10 mmol), steroid 26 (40 mg, 0.09 mmol) and ICH₂TPD (200 mg, 0.61 mmol) in THF (10 mL) were refluxed for 13 h. The reaction was cooled in an ice bath, the excess sodium hydride was eliminated by careful addition of small portions of water from a syringe over a period of 15 min. Water (20 mL) followed by aqueous satd. NH₄Cl (20 mL) were added and the product was extracted into EtOAc (3 x 60 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give the O-alkylation product containing the 3-(methoxymethyoxy) group. This product was partially purified by flash column chromatography (silica gel first with hexanes, followed by 2-15% EtOAc in hexanes) to give an oil (70 mg) that was dissolved in THF (8 mL). (n-Bu)₄NF (1 M in THF) was added and the reaction was refluxed for 2 h. The reaction mixture was cooled, water (30 mL) was added and the product was extracted into EtOAc (3 x 30 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 5-30% EtOAc in hexanes) to give steroid KK231 (7, 32 mg, 68%) as an oil: ¹H NMR (400 MHz, CDCl₃) δ 7.36 (d, J = 8.4 Hz, 2H), 7.17 (d, J = 8.0 Hz, 2H), 5.30 (b s, 1H), 4.72 (b s, 1H), 4.54 (q, J = 12 Hz, 2H), 3.53 (b s, 1H), 3.37 (t, J = 6 Hz, 1H), 2.50–0.80 (m), 1.02 (s, 3H), 0.82 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 141.8, 141.2, 127.4 (2 x C), 127.1, 126.4 (2 x C), 125.3, 88.2, 83.2, 71.4, 70.8, 69.6, 52.1, 50.7, 43.9, 42.0, 40.4, 37.9, 37.1, 35.9, 35.8, 31.7, 28.0, 25.8, 24.9, 21.2, 19.4, 12.0.

(3β,17β)-17-[[[4-(3-Trifluoromethyl-3H-diazirin-3-yl)-phenyl]methyl]-[(1S)-prop-2-ynyl]]-androst-5-en-3-ol (MQ238, 8).

(n-Bu)₄NF (1.0 M in THF, 2.0 mL, 2 mmol) was added to steroid 33 (145 mg, 0.232 mmol) in THF (10 mL) at room temperature. After 16 h, the THF was removed under reduced pressure and the residue was purified by flash column chromatography (silica gel eluted with 25% EtOAc in hexanes) to give MQ238 (8) as a solid (115 mg, 96%): ¹H NMR (400 MHz, CDCl₃) δ 7.23 (d, J = 8.2 Hz, 2H), 7.19 (d, J = 8.2 Hz, 2H), 5.27–5.26 (m, 1H), 3.47–3.41 (m, 1H), 2.86–2.82 M, 1H), 2.47–0.71 (m), 0.93 (s, 3H), 0.69 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 141.4, 140.9, 129.8 (2 x C), 126.9, 126.0 (2 x C), 121.4, 86.3, 72.0, 71.7, 56.4, 53.0, 50.1, 42.6, 42.2, 39.9, 38.7, 37.2, 36.5, 35.7, 36.5, 35.7, 31.9, 31.6, 28.7, 23.9, 20.8, 19.3, 12.1.

(3β,17β)-3-(Methoxymethoxy)-androst-5-en-17-ol (10).

(3β,17β)- 3-(Methoxymethoxy)-androst-5-en-17-one (9) was prepared according to a literature procedure³⁴ and dissolved in stirred cold (0 °C) THF/EtOH (1:4, 10 mL). NaBH₄ (76 mg, 2 mmol) was added in portions and the reaction was allowed to warm to room temperature. After 4 hr, satd. aqueous NH₄Cl (10 mL) was added and the product was extracted into CH₂Cl₂ (3 x 60 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an off-white solid which was purified by flash column chromatography (silica gel eluted with 20-35% EtOAc in hexanes) to give a white solid (317 mg, 95%): ¹H NMR (400 MHz, CDCl₃) δ 5.36 (b s, 1H), 4.69 (s, 2H), 3.65 (t, J = 8.2 Hz, 1H), 3.41 (m, 1H) 3.38 (s, 3H), 2.50–0.90 (m), 1.03 (s, 3H), 0.77 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 140.8, 121.4, 94.7, 81.9, 76.9, 55.2, 51.3, 50.3, 42.7, 39.5, 37.3, 36.8, 36.6, 31.9, 31.5, 30.5, 28.9, 23.4, 20.6, 19.4, 11.0.

3-[4-(3β,17β)-[[(3-Methoxymethoxy)-androst-5-en-17-yl]oxymethyl]phenyl]-3-trifluoromethyl-3H-diazirine (11).

A mixture of NaH (60% suspension in mineral oil, 400 mg, 10 mmol), steroid 10 (100 mg, 0.3 mmol) and 3-[4-(iodomethyl)phenyl]-3-(trifluoromethyl)-3H-diazirine (ICH₂TPD, 200 mg, 0.61 mmol, prepared as described previously¹⁵) in THF (20 mL) was refluxed for 13 h. The reaction mixture was cooled in an ice bath, the excess NaH was eliminated over a period of 15 min by very careful addition of small portions of water using a syringe. Water (20 mL) followed by satd. aqueous NH₄Cl (20 mL) was added and the product was extracted into EtOAc (3 x 60 mL). The combined organic extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give a crude product which was purified by flash column chromatography (silica gel eluted first with hexanes, followed by 2-15% EtOAc in hexanes) to give product 11 as an oil (80 mg, 50%): ¹H NMR (400 MHz, CDCl₃) δ 7.37 (d, J = 7.8 Hz, 2H), 7.16 (d, J = 7.8 Hz, 2H), 5.35 (b s, 1H), 4.69 (s, 2H), 4.55 (s, 2H), 3.39–3.36 (m, 2H), 3.37 (s, 3H), 2.44–0.90 (m), 1.03 (s, 3H), 0.84 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 141.3, 140.8, 127.9, 127.4 (2 x C), 126.4 (2 x C), 122.1 (q, J = 275 Hz), 121.4, 94.7, 81.6, 76.8, 71.5, 70.8, 55.1, 51.5, 50.3, 42.9, 39.5, 37.8, 37.2, 36.8, 31.7, 31.5, 28.9, 27.8, 23.4, 20.7, 19.4, 11.7.

(17β)-17-Hydroxy-10-(2-propyn-1-yl)-estr-4-en-3-one (13).

NaBH₄ (4 mg, 0.11 mmol) in EtOH (3 mL) was added in portions to a stirred, cold (−5 to 0 °C) solution of steroid 12 (100 mg, 0.32 mmol, prepared as described previously²⁰ in 20% THF in EtOH (10 mL) and the reaction was monitored by TLC for the complete consumption of the 17-ketone. Aqueous satd. NH₄Cl (5 mL) was added and the product was extracted into CH₂Cl₂ (3 x60 mL). The combined organic layers were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent removed under reduced pressure to give an off-white solid which was purified by flash column chromatography (silica gel eluted with 20-35% EtOAc in hexanes to give known steroid 13³⁴ as a white solid (100 mg, 99%): ¹HNMR (400 MHz, CDCl₃) δ 5.85 (s, 1H), 3.61 (t, J = 8.2 Hz, 1H), 2.80–0.85 (m), 0.76 (s); ¹³C NMR (100 MHz, CDCl₃) δ 199.6, 167.4, 125.6, 81.5, 81.2, 71.9, 54.3, 50.5, 42.9, 41.6, 36.5, 36.0, 34.6, 34.5, 33.1, 31.8, 30.4, 23.3 (2 x C), 21.0, 11.1.

(17β)-10-(2-Propyn-1-yl)-estra-3,5-diene-3,17-diol, 3,17-diacetate (14).

Trimethylchlorosilane (0.63 mL, 5 mmol) was added to steroid 13. (100 mg, 0.31 mmol) and NaI (750 mg, 5 mmol) in acetic anhydride (4 mL) at 0 °C and the reaction was stirred at room temperature for 4 h. The reaction was cooled in an ice bath, triethylamine (3mL) in Et₂O (10 mL) was slowly added. Aqueous satd. NaHCO₃ (20 mL) was added and the reaction was stirred for 1 h at room temperature. The biphasic solution was extracted with Et₂O (3 x 50 ml) and the combined organic layers were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent removed under reduced pressure to give the crude product as an oil which was passed through a short silica gel column eluted with 10-15% EtOAc in hexanes to give steroid 14 as an oil (110 mg, 89%): ¹H NMR (400 MHz, CDCl₃) δ 5.69 (s, 1H), 5.50 (b s, 1H), 4.62 (t, J = 8.2 Hz, 1H), 2.60–0.80 (m), 2.14 (s, 3H), 2.05 (s, 3H), 0.89 (s); ¹³C NMR (100 MHz, CDCl₃) δ 171.2, 169.3, 147.5, 137.1, 125.4, 116.4, 83.1, 82.6, 71.0, 51.5, 48.9, 42.6, 37.4, 36.9, 31.7, 31.7, 31.1, 27.5, 24.8, 23.5, 22.3, 21.7, 21.2, 21.1, 12.2.

(3β,17β)-10-(2-Propyn-1-yl)-estr-5-ene-3,17-diol, 17-acetate (15).

NaBH₄ (93 mg, 2.5 mmol) was added in portions to a stirred solution of steroid 14 (110 mg, 0.28 mmol) in 20% THF in EtOH (10 mL) and the reaction was stirred for 17 h at room temperature. Aqueous satd. NH₄Cl (15 mL) was added and the product was extracted into CH₂Cl₂ (3 x 60 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent removed under reduced pressure to give to give a viscous liquid which was purified by flash column chromatography (silica gel eluted with 20-35% EtOAc in hexanes) to give steroid 15 as a colorless liquid (65 mg, 66 %): ¹H NMR (400 MHz, CDCl₃) δ 5.55 (b s, 1H), 4.59 (t, J = 8.2 Hz, 1H), 3.54 (m, 1H), 2.58–0.80 (m), 0.87 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 171.2, 137.5, 124.0, 83.0, 82.7, 71.3, 70.9, 51.6, 50.9, 42.5, 42.2, 39.1, 36.9, 35.9, 31.9, 31.6, 30.9, 27.5, 23.5, 22.1, 21.6, 21.2, 12.2.

(3β,17β)-3-(Methoxymethoxy)-10-(2-propyn-1-yl)-estr-5-en-17-ol, 17-acetate (16).

Methoxymethyl chloride (0.1 mL, 1.3 mmol) was added to a stirred solution of steroid 15 (65 mg, 0.18 mmol) and Hunig’s base (1mL) in CH₂Cl₂ (5 mL) and the reaction was stirred for 6 h at room temperature. Aqueous satd. NaHCO₃ (10 mL) was added and the product was extracted into CH₂Cl₂ (3 x 40 mL). The combined extracts were dried over anhydrous Na₂SO₄, filtered and the solvent removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 10-20% EtOAc in hexanes) to give steroid 16 an oil (68 mg, 94%): ¹H NMR (400 MHz, CDCl₃) δ 5.55 (b s, 1H), 4.69 (s, 2H), 4.60 (t, J = 8.2 Hz, 1H), 3.45 (m, 1H), 3.37 (s, 3H), 2.58–0.80 (m), 0.87 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 171.2, 137.5, 124.1, 94.7, 83.0, 82.7, 76.5, 70.9, 55.2, 51.6, 51.0, 42.5, 39.5, 39.4, 36.9, 36.0, 32.0, 31.0, 29.0, 27.5, 23.5, 22.1, 21.6, 21.2, 12.2.

(3β,17β)-3-(Methoxymethoxy)-10-(2-propyn-1-yl)-estr-5-en-17-ol (17).

K₂CO₃ (276 mg, 2 mmol) was added to steroid 16 (68 mg, 0.17 mmol) dissolved in MeOH (5 mL) and the reaction was stirred at room temperature or 14 h. Water (50 mL) was added and the product was extracted into CH₂Cl₂ (3 x 50 mL). The combined extracts were dried over anhydrous Na₂SO₄, filtered and the solvent removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 20-30% EtOAc in hexanes) to give steroid 17 as an oil (60 mg, 99%): ¹H NMR (400 MHz, CDCl₃) δ 5.55 (b s, 1H), 4.68 (s, 2H), 3.64 (t, J = 8.2 Hz, 1H), 3.45 (m, 1H), 3.37 (s, 3H), 2.59–0.80 (m), 0.83 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 137.5, 124.2, 94.7, 83.0, 81.8, 76.5, 70.9, 55.2, 51.8, 51.2, 42.9, 39.5, 39.4, 36.8, 36.0, 32.2, 31.0, 30.5, 29.0, 23.3, 22.1, 21.7, 11.2.

3-[4-(3β,17β)-[(3-Methoxymethoxy)-10-(2-propyn-1-yl)-estr-5-en-17-yloxymethyl]phenyl]-3-(trifluoromethyl)-3H-diazirine (18).

NaH (60% suspension in mineral oil, 400 mg, 10 mmol), steroid 17 (60 mg, 0.17 mmol) and ICH₂TPD (200 mg, 0.61 mmol) in THF (20 mL) were refluxed for 13 h. The reaction was cooled in an ice bath and the excess NaH was eliminated by very careful addition of small portions of water from a syringe over 15 min. Water (20 mL) and aqueous satd. NH₄Cl (20 mL) were added and the product was extracted into EtOAc 3 x 60 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent removed under reduced pressure to give the product as an oil which was purified by flash column chromatography (silica gel eluted first with hexanes and then with 5-15% EtOAc in hexanes) to give steroid 18 as an oil (63 mg, 66 %): ¹H NMR (400 MHz, CDCl₃) δ 7.36 (d, J = 8.2 Hz, 2H), 7.16 (d, J = 8.2 Hz, 2H), 5.55 (b s, 1H), 4.69 (s, 2H), 4.55 (s, 2H), 3.50–3.32 (m, 2H), 3.37 (s, 3H), 2.60–0.85 (m), 0.91 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 141.3, 137.5, 127.9, 127.4 (2 x C), 126.4, 126.4, 124.1, 122.1 (q, J = 275 Hz), 94.7, 88.6, 83.0, 76.5, 70.9, 70.8, 6.19, 52.1, 51.2, 43.1, 39.6, 39.4, 38.0, 35.9, 32.0, 31.0, 29.7, 29.0, 27.9, 23.4, 22.1, 21.8, 11.9.

(3β,7α)-3-(Acetyloxy)-7-bromo-androst-5-en-17-one, cyclic 1,2-ethanediyl acetal (19).

Steroid 19 (300 mg, 0.45 mmol) was prepared as described previously.²¹ This steroid is not stable upon standing and was immediately converted to steroid 20.

(3β,7β)-17,17-[1,2-Ethanediylbis(oxy)]-3-(acetyloxy)-androst-5-ene-7-acetic acid, ethyl ester (20).

Diethyl malonate (1.52 mL, 10 mmol) in THF (5 mL) was added to a NaH suspension (60% in mineral oil, 400 mg, 10 mmol) in THF (10 mL) over a period of 30 min at 0 °C. Freshly prepared steroid 19 (300 mg, 0.45 mmol) in THF (4 mL) was added and the reaction was stirred at room temperature for 48 h. Aqueous satd. NH₄Cl was added and the product was extracted into EtOAc. The combined extracts were dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give the crude C7 diethyl malonate addition product (350 mg) which was converted to steroid 20 without further purification. This addition product and 1,4-diazabicyclo[2.2.2]octane (1.12 g, 10 mmol) in p-xylene (20 mL) was heated at reflux for 16 h. The p-xylene was removed under reduced pressure and the resulting mixture was purified by flash column chromatography (silica gel eluted with 5-25% EtOAc in hexanes) to give steroid 20 as a viscous liquid (130 mg, 63%): ¹H NMR (400 MHz, CDCl₃) δ 5.17 (b s, 1H), 4.56 (m, 1H), 4.09 (q, J = 7.1 Hz, 2H), 3.84 (m, 4H), 2.62–0.80 (m), 2.00 (s, 3H), 1.22 (t, J = 7.1 Hz, 3H), 0.96 (s, 3H), 0.86 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 172.6, 170.4, 139.8, 125.9, 118.6, 73.4, 65.1, 64.5, 60.1, 51.2, 50.3, 46.6, 40.5, 38.0, 37.8, 36.8, 36.7, 35.7, 34.3, 30.3, 27.8, 25.1, 21.3, 20.8, 19.0, 14.4, 12.2.

(3β,7β)-17,17-[1,2-Ethanediylbis(oxy)]-3-hydroxy-androst-5-ene-7-acetic acid, ethyl ester (21).

Steroid 20 (120 mg, 0.26 mmol) and K₂CO₃ (276 mg, 2 mmol) in EtOH (6 mL) were refluxed for 4 h. Water was added and the product was extracted into CH₂Cl₂ (3 x 80 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give a viscous liquid which was purified by flash column chromatography (silica gel eluted with 15-30% EtOAc in hexanes) to give steroid 21 as a viscous liquid (101 mg, 93%): ¹H NMR (400 MHz, CDCl₃) δ 5.12 (b s, 1H), 4.08 (q, J = 7.1 Hz, 2H), 3.83 (m, 4H), 3.47 (m, 1H), 2.62–0.80 (m), 1.21 (t, J = 7.1 Hz, 3H), 0.93 (s, 3H), 0.84 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 172.8, 140.9, 124.9, 118.6, 71.0, 65.1, 64.4, 60.1, 51.2, 50.4, 46.6, 41.9, 40.5, 37.9, 37.0, 36.7, 35.6, 34.3, 31.5, 30.3, 25.0, 20.8, 19.0, 14.4, 14.2.

(3β,7β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-17,17-[1,2-ethanediylbis(oxy)]-androst-5-en-7-acetic acid, ethyl ester (22).

Steroid 21 (101 mg, 0.24 mmol), t-butyldimethylsilyl chloride (151 mg, 1 mmol) and imidazole (68 mg. 1 mmol) in DMF (4 mL) were stirred at room temperature for 15 h. Aqueous satd. NH₄Cl was added and the product was extracted into EtOAc (3 x 50 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by column chromatography (silica gel eluted with 5-20% EtOAc in hexanes) to give steroid 22 as a liquid (126 mg, 90%): ¹H NMR (400 MHz, CDCl₃) δ 5.13 (b s, 1H), 4.11 (q, J = 7.1 Hz, 2H), 3.88 (m, 4H), 3.47 (m, 1H), 2.62–0.80 (m), 1.24 (t, J = 7.1 Hz, 3H), 0.95 (s, 3H), 0.88 (s, 9H), 0.86 (s, 3H), 0.05 (s,6H) ; ¹³C NMR (100 MHz, CDCl₃) δ 172.8, 141.5, 124.6, 118.6, 72.0, 65.1, 64.5, 60.1, 51.2, 50.4, 46.6, 42.6, 40.6, 38.0, 37.2, 36.8, 35.7, 34.4, 32.1, 30.4, 25.9 (3 x C), 25.2, 20.9, 19.1, 18.2, 14.4, 14.3, −4.6 (2 x C).

(3β,7β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-hydroxyethyl)-androst-5-en-17-one, cyclic 1,2-ethanediyl acetal (23).

LiAlH₄ (2 M in THF, 0.5 ml, 1 mmol) was added to a cold (0 °C) solution of steroid 22 (126 mg, 0.24 mmol) in Et₂O and the reaction was allowed to warm to room temperature and continued for 2 h. The reaction mixture was diluted with Et₂O (20 mL), cooled to 0 °C and water (0.3 mL) was added dropwise. After 5 min, 5 M aqueous NaOH (2 mL) was added and the mixture was stirred for 30 min. Water (1 ml) was added again and the mixture was stirred for 1 h. The Et₂O was decanted, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 20-40% EtOAc in hexanes) to give steroid 23 as an oil (106 mg, 91%): ¹H NMR (400 MHz, CDCl₃) δ 5.11 (b s, 1H), 3.80 (m, 4H), 3.62 (m, 2H), 3.43 (m, 1H), 2.25–0.80 (m), 0.87 (s, 3H), 0.80 (s, 9H), 0.79 (s, 3H), −0.03 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 141.2, 125.6, 118.8, 72.1, 65.1, 64.5, 60.7, 51.6, 50.8, 46.7, 42.7, 38.1, 37.9, 37.3, 36.7, 35.7, 34.4, 32.2, 30.5, 25.9 (3 x C), 25.2, 20.9, 19.0, 18.2, 14.5, −4.6 (2 x C).

(3β,7β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-propyn-1-yl)-androst-5-en-17-one, cyclic 1,2-ethanediyl acetal (24).

Steroid 23 (105 mg, 0.22 mmol) and Dess-Martin reagent (212 mg, 0.5 mmol) in CH₂Cl₂ were stirred at room temperature for 4 h. CH₂Cl₂ (40 ml) was added and washed with aqueous satd. NaHCO₃, brine, dried over anhydrous Na₂SO₄, filtered and the solvent removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 10-20% EtOAc in hexanes) to give the 7-acetaldehyde intermediate as an oil (90 mg, 88 %) which was immediately dissolved in THF (2 ml) and MeOH (4 mL). Bestmann-Ohira reagent (0.150 ml, 1 mmol) and K₂CO₃ (138 mg, 1mmol) were added and the mixture was stirred at room temperature for 40 h. Water was added and the product was extracted into EtOAc (3 x 30 ml). The combined extracts were dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 10-15% EtOAc in hexanes) to give steroid 24 as an oil (75 mg, 78 %): ¹H NMR (400 MHz, CDCl₃) δ 5.25 (b s, 1H), 3.90 (m, 4H), 3.47 (m, 1H), 2.40–0.80 (m), 1.95 (t, J = 2.3 Hz, 1H), 1.00 (s, 3H), 0.89 (s, 9H), 0.87 (s, 3H), −0.06 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 142.5, 124.9, 118.7, 83.2, 72.1, 69.6, 65.2, 64.5, 51.1, 50.4, 46.7, 42.5, 40.4, 37.2, 36.3, 35.9, 34.3, 32.2, 30.4, 25.9 (3 x C), 25.1, 25.0, 20.9, 19.4, 18.2, 14.5, −4.6 (2 x C).

(3β,7β)- 3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-propyn-1-yl)-androst-5-en-17-one (25).

Steroid 24 (70 mg, 0.14 mmol) and p-toluenesulfonic acid (10 mg) in acetone (5 mL) were stirred at room temperature for 4 h. Aqueous satd. NaHCO₃ (5 mL) was added and the product was extracted into EtOAc (3 x 50 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give a steroid in which both the alcohol and ketone protecting groups were removed. This steroid was dissolved in DMF (4 mL) and t-butyldimethylsilyl chloride (150 mg, 1mmol) and imidazole (78 mg, 1 mmol) were added and the reaction was stirred at room temperature for 15 h. Aqueous satd. NH₄Cl was added and the product was extracted into EtOAc (3 x 50 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by flash column chromatography (silica gel eluted with 5-20% EtOAc in hexanes) to give steroid 25 as a liquid (50 mg, 81 %): ¹H NMR (400 MHz, CDCl₃) δ 5.29 (b s, 1H), 3.48 (m, 1H), 2.50–0.80 (m), 1.98 (t, J = 2.3 Hz, 1H), 1.02 (s, 3H), 0.91 (s, 3H) 0.89 (s, 9H), 0.02 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 220.8, 142.7, 124.3, 82.8, 72.0, 69.9, 52.4, 50.7, 48.3, 42.5, 39.8, 37.1, 36.0, 35.9, 35.8, 32.1, 31.6, 25.9 (3 x C), 25.1, 24.1, 20.8, 19.4, 18.2, 13.9, −4.6 (2 x C).

(3β,7β,17β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-propyn-1-yl)-androst-5-en-17-ol (26).

NaBH₄ (19 mg, 0.5 mmol) was added to steroid 25 (50 mg, 0.11 mmol ) in EtOH (5 mL) and THF (1 mL) and the reaction was stirred at room temperature for 6 h. Water and aqueous satd. NH₄Cl were added and the product was extracted into CH₂Cl₂ (3 x 60 mL). The combined extracts were washed with brine, dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give an oil which was purified by flash column chromatography to give steroid 26 as a viscous oil (40 mg, 80%): ¹H NMR (400 MHz, CDCl₃) δ 5.27 (b s, 1H), 3.62 (t, J = 6.8 Hz, 1H), 3.48 (m, 1H), 2.40–0.80 (m), 1.95 (t, J = 2.3 Hz, 1H), 1.00 (s, 3H), 0.90 (s, 9H), 0.78 (3H, s) 0.06 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 142.6, 124.8, 83.2, 81.5, 72.2, 69.6, 51.9, 50.7, 43.7, 42.6, 40.6, 37.2, 36.9, 36.1, 35.9, 32.2, 30.6, 25.9 (3 x C), 25.8, 25.0, 21.1, 19.4, 18.2, 11.2, −4.6 (2 x C).

(3β,17E)-3-(Methoxymethoxy)-pregna-5,17(20)-dien-21-oic acid, ethyl ester (27).

Sodium ethoxide generated in situ from sodium metal (1.15 g, 50 mmol) in anhydrous EtOH (40 mL) was added dropwise to a stirred solution of steroid 9 (3.46 g, 10.4 mmol) and triethyl phosphonoacetate (10.4 ml, 52 mmol) in anhydrous EtOH (100 mL) under N₂ at 35-40 °C. After addition, the reaction was refluxed for 16 h. Ethanol was removed under reduced pressure and water was added. The product was extracted into CH₂Cl₂ (150 mL x2), washed with brine (50 mL x2), dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure. The residue was purified by flash column chromatography (silica gel eluted with 5-10% EtOAc in hexanes) to give steroid 27 as a solid (3.28 g, 78%): ¹H NMR (400 MHz, CDCl₃) δ 5.53 (d, J = 2.0 Hz, 1H), 5.35–5.34 (m, 1H), 4.68 (s, 2H), 4.17 (q, J = 7.0 Hz, 2H), 3.45–3.23 (m, 1H), 3.36 (d, J = 0.8 Hz, 3H), 2.85–2.81 (m, 2H), 2.38–0.87 (m, 20H), 1.03 (s, 3H), 0.83 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 176.1, 167.4, 140.7, 121.3, 108.6, 94.7, 76.8, 59.5, 55.1, 53.8, 50.2, 46.0, 39.5, 37.2, 36.8, 35.1, 31.6, 31.5, 30.4, 28.8, 24.4, 20.9, 19.3, 18.2, 14.3.

(3β)-3-Hydroxypregn-5-en-21-oic acid, ethyl ester (28).

PtO₂ (80 mg) was added to steroid 27 (3.28 g, 8.16 mmol) in EtOAc (150 mL) and the flask was evacuated and charged with H₂ three times. The hydrogenation was carried out under 40 psi at room temperature. After 4 h, the mixture was filtered through Celite and washed with EtOAc (200 mL). Solvent was removed under reduced pressure and the residue (3.28 g) was dissolved in EtOH (100 mL) and stirred with dry HCl generated by the addition of AcCl (6 mL) for 2 h at room temperature. Water was added and the product was extracted into CH₂Cl₂ (2 x 150 mL). Solvent was removed under reduced pressure and the residue was purified by flash column chromatography (silica gel eluted with 10% EtOAc in hexanes) to afford steroid 28 containing 5-10% of a steroid in which the Δ⁵-double bond was also reduced (2.82 g). To separate the two steroids and obtain pure steroid 28 the following procedure was used.

The steroid mixture (2.82 g) was dissolved in Et₂O (100 mL), cooled to 0 °C and Br₂ in HOAc (5 mL) was added until a brown color persisted. Aqueous Na₂S₂O₃ was added until the reaction became colorless. The steroids were extracted into EtOAc (150 mL), washed with NaHCO₃ (2 x 50 mL), brine (50 mL), dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure. The residue, a mixture of the saturated steroid and 5,6-dibromide, were purified by flash column chromatography (silica gel eluted with 10-25% EtOAc in hexanes) to obtain the 5,6-dibromide (3.65 g) which was subsequently converted back to steroid 28 by the following procedure.

Zn dust (13.65 g, 210 mmol) was added to the 5,6-dibromide (2.65 g, 7.02 mmol) dissolved in HOAc (50 mL) and EtOAc (75 mL) at room temperature. After 16 h, the mixture was filtered through Celite and washed with EtOAc (200 mL). Solvent was removed under reduced pressure and the residue was purified by flash column chromatography (silica gel eluted with 10-25% EtOAc in hexanes) to give steroid 28 as a solid (2.15 g, 74%): ¹H NMR (400 MHz, CDCl₃) δ 5.33–5.32 (m, 1H), 4.12 (q, J = 7.4 Hz, 2H), 3.53–3.46 (m, 1H), 2.41–0.91 (m), 0.99 (s, 3H), 0.59 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 173.9, 140.7, 121.4, 71.5, 60.1, 55.5, 50.2, 46.7, 42.0, 41.8, 37.2, 37.1, 36.4, 35.2, 31.8, 31.8, 31.3, 28.0, 24.5, 20.7, 19.3, 14.1, 12.3.

(3β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-pregn-5-en-21-oic acid, ethyl ester (29).

t-Butyldimethylsilyl chloride (300 mg, 2 mmol) and imidazole (204 mg, 3 mmol) were added to steroid 28 (410 mg, 1.14 mmol) in DMF (20 mL) at room temperature. After 16 h, water (50 mL) was added and the product was extracted into EtOAc (200 mL). The extract was washed with water (2 x 50 mL), dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure. The residue was purified by flash column chromatography (silica gel eluted with 5% EtOAc in hexanes) to give steroid 29 (578 mg, 96%): ¹H NMR (400 MHz, CDCl₃) δ 5.32–5.31 (m, 1H), 4.14 (q, J = 7.0 Hz, 2H), 3.51–3.45 (m, 1H), 2.39–0.92 (m), 1.00 (s, 3H), 0.89 (s, 9H), 0.61 (s, 3H), 0.06 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 173.9, 141.6, 121.0, 72.6, 60.1, 55.6, 50.4, 46.8, 42.8, 41.9, 37.4, 37.2, 36.6, 35.3, 32.1, 32.0, 31.9, 28.1, 25.9 (3 x C), 24.6, 20.8, 19.4, 18.2, 14.2, 12.4, −4.6 (2 x C).

(3β,20S)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-20-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methyl]-pregn-5-en-21-oic acid, ethyl ester (30).

LiN(i-Pr)₂ (2.0 M in THF, 1.5 mL, 3 mmol) and hexamethylphosphoramide (0.53 mL, 3 mmol) were added to steroid 29 (518 mg, 1.09 mmol) in THF (20 mL) at −78 °C. After 1 h, ICH₂TPD (533 mg, 1.6 mmol) in THF (5 mL) was added at −78 °C and the reaction was continued at room temperature for 2 h. Water was added and the product was extracted into EtOAc (2 x 150 mL), dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure. The residue was purified by flash column chromatography (silica gel eluted with 5% EtOAc in hexanes) to give steroid 30 as an oil (220 mg, 30%): ¹H NMR (400 MHz, CDCl₃) δ 7.18 (d, J = 8.2 Hz, 2H), 7.09 (d, J = 8.2 Hz, 2H), 5.32–5.31 (m, 1H), 3.96–3.82 (m, 2H), 3.52–3.44 (m, 1H), 2.91–0.90 (m), 0.99 (s, 3H), 0.89 (s, 9H), 0.73 (s, 3H), 0.06 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 174.7, 141.6, 141.3, 129.3 (3 x C), 127.0, 126.4 (2 x C), 120.9, 72.6, 59.8, 56.1, 52.7, 50.1, 49.7, 42.8, 42.1, 38.2, 37.4, 37.3, 36.5, 32.0, 31.9, 31.8, 27.3, 25.9 (3 x C), 23.9, 20.8, 19.4, 18.2, 13.9, 12.0, −4.6 (2 x C).

(3β,20S)- 3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-20-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methyl]-pregn-5-en-21-ol (31).

DIABAL (1.0 M in hexanes, 1.0 mL, 1 mmol) was slowly added to steroid 30 (220 mg, 0.33 mmol) in CH₂Cl₂ (20 mL) at −78 °C. After 1.5 h, MeOH (0.5 mL) was added and stirred at −78 °C for 20 min. Aqueous satd. potassium sodium tartrate (20 mL) was added and stirred at room temperature for 1 h. The product was extracted into CH₂Cl₂ (2 x100 mL). The combined extracts were washed with water (2 x 50 mL), dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure. The residue was purified by flash column chromatography (silica gel eluted with 20% EtOAc in hexanes) to give steroid 31 as an oil (190 mg, 92%): ¹H NMR (400 MHz, CDCl₃) δ 7.20 (d, J = 8.2 Hz, 2H), 7.12 (d, J = 8.2 Hz, 2H), 5,26–5.25 (m, 1H), 3.56–3.30 (m, 3H), 2.75–0.86 (m), 0.94 (s, 3H), 0.83 (s, 9H), 0.67 (s, 3H), 0.00 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 143.1, 141.5, 129.9 (3 x C), 126.5, 126.3 (2 x C), 121.0, 72.6, 61.0, 56.7, 50.1, 50.0, 45.0, 42.8, 42.2, 39.1, 37.3, 36.6, 35.9, 32.0, 31.9, 31.8, 28.5, 25.9 (3 x C), 24.2, 21.0, 19.4, 18.2, 12.2, −4.6 (2 x C).

(3β,20S)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-20-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methyl]-pregn-5-en-21-al (32).

Dess-Martin reagent (424 mg, 1 mmol) was added to steroid 31 (190 mg, 0.30 mmol) and NaHCO₃ (300 mg) in CH₂Cl₂ (20 mL) at room temperature. After 2 h, water was added and the product was extracted into CH₂Cl₂ (2 x 150 mL). The combined extracts were dried over anhydrous Na₂SO₄, filtered, the solvent was removed under reduced pressure and the residue was purified by flash column chromatography (silica gel eluted with 15% EtOAc in hexanes) to give steroid 32 as an oil (181 mg, 95%): ¹H NMR (400 MHz, CDCl₃) δ 9.56 (d, J = 5.1 Hz, 1H), 7.15 (d, J = 8.6 Hz, 2H), 7.09 (d, J = 8.6 Hz, 2H), 5.33–5.31 (m, 1H), 3.51–3.44 (m, 1H), 2.92–0.57 (m), 0.99 (s, 3H), 0.89 (s, 9H), 0.73 (s, 3H), −0.09 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 204.6, 141.6, 140.6, 129.4 (3 x C), 127.2, 126.6 (2 xC), 120.8, 72.5, 56.6, 56.2, 51.3, 50.1, 42.8, 42.2, 38.5, 37.3, 36.6, 35.0, 32.0, 31.8, 31.7, 27.4, 25.9 (3 x C), 24.0, 20.7, 19.4, 18.3, 12.9, −4.6 (2 x C).

3-[(2S)-4-[2-[(3β,17β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-androst-5-en-17-yl]-but-3-ynyl]-phenyl]- 3-(trifluoromethyl)-3H-diazirine (33).

Bestmann-Ohira reagent (0.23 mL, 1.5 mmol) and K₂CO₃ (328 mg, 2.4 mmol) were added to steroid 32 (181 mg, 0.29 mmol) in MeOH/THF (5 mL/5 mL) at room temperature. After 16 h, water was added and the product was extracted into EtOAc (2 x 100 mL). The combined extracts were dried over anhydrous Na₂SO₄, filtered and the solvent was removed under reduced pressure to give a residue which was purified by flash column chromatography (silica gel eluted with 15% EtOAc in hexanes) to give steroid 33 as an oil (145 mg, 80%): ¹H NMR (400 MHz, CDCl₃) δ 7.32 (d, J = 8.2 Hz, 2H), 7.27 (d, J = 8.2 Hz, 2H), 5.33–5.32 (m, 1H), 3.52–3.46 (m, 1H), 2.95–2.91 (m, 1H), 2.57–0.80 (m), 1.01 (s, 3H), 0.90 (s, 9H), 0.77 (s, 3H), 0.07 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 141.7, 141.4, 129.9 (3 x C), 127.0, 126.1 (2 x C), 120.9, 86.4, 72.6, 72.0, 56.5, 53.0, 50.2, 42.8, 42.6, 39.9, 38.8, 37.4, 36.6, 35.7, 32.1, 31.9, 28.7, 25.9 (3 x C), 24.0, 20.8, 19.4, 18.2, 12.1, −4.6 (2 x C).

Gene Expression Assays for HMGCS and ABCA1.

Gene expression assays were performed as previously described.¹⁷ In brief, U2OS-SRA cells were treated in lipoprotein deficient medium containing 20 μM lovastatin and 50 μM mevalonate and supplemented with 10 μM cholesterol, KK226, KK231, KK237, MQ238, MQ182, 3μM 25-hydroxycholesterol, or an equal volume of ethanol. After 14 h RNA was collected in Trizol (Ambion) and prepared for quantitative PCR. Relative quantification of HMG-CoA synthase (HMGCS) or ABCA1 transcript abundance was calculated using the ddCT method using Rplp0 as an endogenous control.

Cholesterol auxotroph viability assays.

Cholesterol auxotrophy assays were performed as previously described.¹⁷ In brief, U2OS-SRA cells were plated, pulsed with methyl-β-cyclodextrin, washed, then incubated in lipoprotein deficient medium containing 20 μM lovastatin and 50 μM mevalonate and supplemented with 10 μM cholesterol, KK226, KK231, KK237, MQ238, MQ182, or LKM38, or an equal volume of ethanol. After 72 hours, cell viability was assayed using CellTiterGlo Luminescent Cell Viability Assay (Promega) according to the manufacturer’s instructions.

Expression and purification of GLIC from E. coli.

GLIC was expressed and purified as previously described. Briefly, pET26-MBP-GLIC, a gift from Raimund Dutzler, was expressed in OverExpress™ C43(D3) Escherichia coli, grown in Terrific broth supplemented with 10% Glycerol at 37 °C and induced with 0.2 mM isopropyl-1-thio-β-d-galactopyranoside at 18 °C. Cells were pelleted and lysed in buffer A (20 mM Tris pH 7.5, 100 mM NaCl, protease inhibitors and DNase), and the membrane proteins were collected by centrifugation at 400,000 g for 45 min. The membrane proteins were solubilized in buffer containing 1% n-dodecyl-β-d-maltoside (DDM). GLIC was purified with amylose resin and eluted with 40 mM maltose in buffer containing 0.02% DDM. His-MBP-GLIC was digested with HRV 3c to remove the affinity tag and further purified with size exclusion chromatography (SEC) on a Sephadex 200 Increase 10/300 column. Purified protein was eluted from SEC in 50 mM Tris pH 7.5, 150 mM NaCl, and 0.02% DDM.

Cell culture and membrane preparation.

Cell culture and membrane preparation followed a previously described protocol.¹⁵ Briefly, HEK T-Rex™-293 (ThermoFisher) cells were maintained in DMEM/F-12 50/50 medium containing 10% fetal bovine serum (Takara, Mountain View, CA), penicillin (100 units/ml), streptomycin (100g/ml), and blatscidine (2μg/ml) in a humidified atmosphere containing 5% CO2. Cells were grown to 70-80% confluency, harvested with cell scrapers and collected by centrifugation at 21,000g at 4 °C for 5 min. The cells were homogenized with a glass mortar Teflon pestle for 10 strokes on ice. Membranes were collected by centrifugation at 34,000g (30 min at 4 °C) and re-suspended in 10 mM potassium phosphate buffer, 100mM KCl, pH 7.5. Protein concentration was determined using a micro-BCA protein assay and stored at −80 °C.

Photolabeling and MS analysis.

Purified GLIC (70 μg) was exposed to UV light in the presence of 100 μM cholesterol photolabeling analogues for 5 min. Photolabeled protein (20 μg) was used for middle-down MS and 50 μg was used for intact protein MS analysis. For middle-down MS sample preparation, BioRad Biospin 6 columns were used for removing salts. The proteins were then reduced, alkylated and digested with trypsin at 4 °C for one week. The resultant peptides were analyzed with an OrbiTrap ELITE mass spectrometer (Thermo Fisher Scientific) as previously described.²⁶ The data generated by middle-down MS was searched with the adduct weight of the various cholesterol analogue photolabeling reagents (i.e. minus N₂) against a database containing GLIC sequence using PEAKS studio software. For intact protein MS analysis, the photolabeled protein was precipitated with CH₃Cl:MeOH:H₂O in a 1:1:1 ratio and resuspended in formic acid. Formic acid was diluted to <5% with 4:4:1 CH₃Cl:MeOH:H₂O and directly injected into a Thermo™ Orbitrap Elite mass spectrometer. A spray voltage of 4 kV, capillary temperature of 320 °C, and insource dissociation of 30 V was used. Full spectra were acquired in the linear trap quadrupole. Spectra were deconvoluted using Unidec software.³⁵

Protein photolabeling, click cycloaddition reaction and gel analysis.

Purified GLIC protein (5 μg in 0.02% DDM, 100 mM NaCl, 20 mM Tris pH 7.5 ) was incubated with a 100 μM concentration of each of the cholesterol analogue photolabeling reagents for 1 h at 4 °C. For the experiment examining competitive prevention of MQ182 labeling, GLIC (5 μg in 0.01% DDM, 0.004% NP-40, 100 mM NaCl, 20 mM Tris, pH 7.5) was incubated with 1 μM MQ182 in the presence or absence of cholesterol hemisuccinate (CHS; 3,10, 30 μM) or cholesterol (3,10 μM). The protein samples were then transferred to a quartz cuvette, placed in a photoreactor at a distance of 7 cm from the source and irradiated for 5 min. The photoreactor uses a 450-watt Hanovia medium pressure mercury lamp as the light source. A cold water jacket cooled the lamp, and the light was filtered through a 1.5-cm-thick saturated copper sulfate solution to absorb all light <315 nM. SDS was then added to a final concentration of 2% and the proteins were incubated with TAMRA-azide in the dark at room temperature overnight in a solution containing (in mM) 2.5 sodium ascorbate, 0.25 Tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine, 2.5 CuSO₄. The reaction was terminated by addition of 5X-SDS loading buffer (10% sodium dodecyl sulfate, 25% 2-mercaptoethanol, 30% glycerol, 62.5 mM Tris, pH 6.8). The proteins were then loaded onto a 4-20% Tris-glycine SDS gel. For SyproRuby protein staining, the SDS-PAGE gel was fixed at RT with 50% methanol and 7% acetic acid for 30 min, incubated with SyproRuby overnight and destained with 10% methanol and 7% acetic acid for 4 h. Fluorescence was visualized with a GE Typhoon FLA 9000 gel scanner and densitometry was analyzed with Image J software. For labeling HEK cell membranes, the various cholesterol PALs were incubated with 75 μg of membrane protein in a volume of 100 μl for 1 hour at 4°C. Following UV irradiation (as described above for GLIC) membranes were collected by centrifugation (20,000g for 45 min) and solubilized in 2% SDS/PBS for 1 hour at room temperature. Insoluble material was removed by centrifugation (21,000g for 30 min) prior to click chemistry and SDS-PAGE (as described above for GLIC).

Computational Docking.

The X-ray structure of GLIC (PDB 4F8H) was used as the starting structure for a short molecular dynamics simulation. All non-protein atoms were removed from the structure, the receptor was then oriented into a membrane bilayer using the OPM web server (https://opm.phar.umich.edu), then protonated using H++ (http://biophysics.cs.vt.edu). The receptor simulation was then setup (embedded into a POPC bilayer, fully solvated with water, and ionic strength set to 0.15M KC) using CHARMM-GUI (http://charmm-gui.org). A 30 ns molecular dynamics trajectory was obtained and the final 10 ns were clustered for docking studies using AutoDock Vina. A docking box was built to include the known intersubunit binding site with dimensions of 26 x 26 x 26 Å.

Supplementary Material

Chem. Biol.Supplemental

NIHMS1765311-supplement-Chem__Biol_Supplemental.pptx^{(3.4MB, pptx)}

Figure 4. — U2OS-SRA cells were treated with 10 μM cholesterol, KK226, KK231, KK237, MQ238, MQ182, or LKM38, or an equal volume of ethanol (V) in cholesterol starvation media. After 72 h cell viability was determined by CellTiterGlo assay. Values are normalized to V. Bars: mean + SE, n = 3-4 experiments/condition. *, P<0.05 vs. vehicle, #, P<0.05 vs. cholesterol, unpaired t-test. $, P<0.05 paired t-test.

Acknowledgements

Research was supported by National Institutes of Health grants R01 MH110550 (D.F.C.), R01 HL067773 (J.E.S., D.S.O., D.F.C), R01 GM108799 (A.S.E., D.F.C), T32 HL134635 (M.F.), F31 HL142167 (M.F.) and The Taylor Family Institute for Innovative Psychiatric Research (D.F.C., A.S.E., Z-W.C, D.E.R.).

Footnotes

Supporting Information

Supplementary Figures S1 – S10, MS2 Spectra Table S1 and ¹H NMR and ¹³C NMR spectra of PAL reagents 5-8 (PDF).

REFERENCES

1.Yeagle PL (1985) Cholesterol and the cell membrane, Biochim. Biophys. Acta, Rev. Biomembr 822, 267–287. [DOI] [PubMed] [Google Scholar]
2.Yeagle PL (1991) Modulation of membrane function by cholesterol, Biochimie 73, 1303–1310. [DOI] [PubMed] [Google Scholar]
3.Lingwood D, and Simons K (2010) Lipid Rafts As a Membrane-Organizing Principle, Science (Washington, DC, U. S.) 327, 46–50. [DOI] [PubMed] [Google Scholar]
4.Rog T, and Vattulainen I (2014) Cholesterol, sphingolipids, and glycolipids: What do we know about their role in raft-like membranes?, Chem. Phys. Lipids 184, 82–104. [DOI] [PubMed] [Google Scholar]
5.Miller WL, and Auchus RJ (2011) The molecular biology, biochemistry, and physiology of human steroidogenesis and its disorders, Endocr. Rev 32, 81–151. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Brown MS, Radhakrishnan A, and Goldstein JL (2018) Retrospective on Cholesterol Homeostasis: The Central Role of Scap, Annu. Rev. Biochem 87, 783–807. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Kowatsch C, Woolley RE, Kinnebrew M, Rohatgi R, and Siebold C (2019) Structures of vertebrate Patched and Smoothened reveal intimate links between cholesterol and Hedgehog signalling, Curr. Opin. Struct. Biol 57, 204–214. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Mosqueira A, Camino PA, and Barrantes FJ (2018) Cholesterol modulates acetylcholine receptor diffusion by tuning confinement sojourns and nanocluster stability, Sci. Rep 8, 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Radhakrishnan A, Sun L-P, Kwon HJ, Brown MS, and Goldstein JL (2004) Direct binding of cholesterol to the purified membrane region of SCAP: Mechanism for a sterol-sensing domain, Mol. Cell 15, 259–268. [DOI] [PubMed] [Google Scholar]
10.Cologna SM, and Rosenhouse-Dantsker A (2019) Insights into the molecular mechanisms of cholesterol binding to the NPC1 and NPC2 proteins, Adv. Exp. Med. Biol 1135, 139–160. [DOI] [PubMed] [Google Scholar]
11.Hulce JJ, Cognetta AB, Niphakis MJ, Tully SE, and Cravatt BF (2013) Proteome-wide mapping of cholesterol-interacting proteins in mammalian cells, Nat. Methods 10, 259–264. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Cheng WWL, Budelier MM, Sugasawa Y, Bergdoll L, Queralt-Martin M, Rosencrans W, Rostovtseva TK, Chen Z-W, Abramson J, Krishnan K, Covey DF, Whitelegge JP, and Evers AS (2019) Multiple neurosteroid and cholesterol binding sites in voltage-dependent anion channel-1 determined by photo-affinity labeling, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 1864, 1269–1279. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Budelier MM, Cheng WWL, Chen Z-W, Bracamontes JR, Sugasawa Y, Krishnan K, Mydock-McGrane L, Covey DF, and Evers AS (2019) Common binding sites for cholesterol and neurosteroids on a pentameric ligand-gated ion channel, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 1864, 128–136. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Brunner J, Senn H, and Richards FM (1980) 3-Trifluoromethyl-3-phenyldiazirine. A new carbene generating group for photolabeling reagents, J. Biol. Chem 255, 3313–3318. [PubMed] [Google Scholar]
15.Chen Z-W, Bracamontes J, Budelier MM, Germann AL, Shin DJ, Krishnan K, Qian M, Manion B, Cheng WWL, Reichert DE, Akk G, Covey DF, and Evers AS (2019) Multiple functional neurosteroid binding sites on GABAA receptors, PLoS Biology 17, e3000157. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Sugasawa Y, Cheng WWL, Bracamontes J, Chen Z-W, Wang L, Germann AL, Pierce SR, Senneff TC, Krishnan K, Reichert DE, Covey DF, Akk G, and Evers AS (2020) Site-specific effects of neurosteroids on GABAA receptor activation and desensitization, eLife 9, e55331. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Feltes M, Moores S, Gale SE, Krishnan K, Mydock-McGrane L, Covey DF, Ory DS, and Schaffer JE (2019) Synthesis and characterization of diazirine alkyne probes for the study of intracellular cholesterol trafficking, Journal of lipid research 60, 707–716. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Chen DM, Ziolkowski L, Benz A, Qian M, Zorumski CF, Covey DF, and Mennerick S (2018) A clickable oxysterol photolabel retains NMDA receptor activity and accumulates in neurons., Front. Neurosci 12, 923. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Kaji Y, Koami T, Nakamura A, and Fujimoto Y (2000) Synthesis of 3-epi-6,7-dideoxyxestobergsterol A, Chem. Pharm. Bull 48, 1480–1483. [DOI] [PubMed] [Google Scholar]
20.Covey DF, Hood WF, and Parikh VD (1981) 10β-propynyl-substituted steroids. Mechanism-based enzyme-activated irreversible inhibitors of estrogen biosynthesis, J Biol Chem 256, 1076–1079. [PubMed] [Google Scholar]
21.Duval D, Desfosses B, and Emiliozzi R (1980) Preparation of dehydroepiandrosterone, testosterone and progesterone antigens through 7-carboxymethyl derivatives: characteristics of the antisera to testosterone and progesterone, Steroids 35, 235–249. [DOI] [PubMed] [Google Scholar]
22.Dhameja M, and Pandey J (2018) Bestmann-Ohira Reagent: A Convenient and Promising Reagent in the Chemical World, Asian J. Org. Chem 7, 1502–1523. [Google Scholar]
23.Wicha J, and Bal K (1978) Synthesis of 21-hydroxycholesterol and 25-hydroxycholesterol from 3β-hydroxyandrost-5-en-17-one. A method for the stereospecific construction of sterol side-chains, J. Chem. Soc., Perkin Trans 1, 1282–1288. [Google Scholar]
24.Goldstein JL, DeBose-Boyd RA, and Brown MS (2006) Protein sensors for membrane sterols, Cell (Cambridge, MA, U. S.) 124, 35–46. [DOI] [PubMed] [Google Scholar]
25.Gill S, Chow R, and Brown AJ (2008) Sterol regulators of cholesterol homeostasis and beyond: The oxysterol hypothesis revisited and revised, Prog. Lipid Res 47, 391–404. [DOI] [PubMed] [Google Scholar]
26.Cheng WWL, Chen Z-W, Bracamontes JR, Budelier MM, Krishnan K, Shin DJ, Wang C, Jiang X, Covey DF, Akk G, and Evers AS (2018) Mapping two neurosteroid-modulatory sites in the prototypic pentameric ligand-gated ion channel GLIC, J. Biol. Chem 293, 3013–3027. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Barrantes FJ, and Fantini J (2016) From hopanoids to cholesterol: Molecular clocks of pentameric ligand-gated ion channels, Prog. Lipid Res 63, 1–13. [DOI] [PubMed] [Google Scholar]
28.Laverty D, Thomas P, Field M, Andersen OJ, Gold MG, Biggin PC, Gielen M, and Smart TG (2017) Crystal structures of a GABAA-receptor chimera reveal new endogenous neurosteroid-binding sites, Nat. Struct. Mol. Biol 24, 977–985. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Mix KA, and Raines RT (2015) Optimized Diazo Scaffold for Protein Esterification, Org. Lett 17, 2358–2361. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Jiang X, Shu H-J, Krishnan K, Qian M, Taylor AA, Covey DF, Zorumski CF, and Mennerick S (2016) A clickable neurosteroid photolabel reveals selective Golgi compartmentalization with preferential impact on proximal inhibition, Neuropharmacology 108, 193–206. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Peyrot SM, Nachtergaele S, Luchetti G, Mydock-McGrane LK, Fujiwara H, Scherrer D, Jallouk A, Schlesinger PH, Ory DS, Covey DF, and Rohatgi R (2014) Tracking the Subcellular Fate of 20(S)-Hydroxycholesterol with Click Chemistry Reveals a Transport Pathway to the Golgi, J. Biol. Chem 289, 11095–11110. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Das J (2011) Aliphatic Diazirines as Photoaffinity Probes for Proteins: Recent Developments, Chem. Rev. (Washington, DC, U. S.) 111, 4405–4417. [DOI] [PubMed] [Google Scholar]
33.Gimpl G, and Gehrig-Burger K (2011) Probes for studying cholesterol binding and cell biology, Steroids 76, 216–231. [DOI] [PubMed] [Google Scholar]
34.Johnston JON, Wright CL, Holbert GW, and Benson HD (1990) Enzyme inactivation by potential metabolites of an aromatase-activated inhibitor (MDL 18,962), J. Enzyme Inhib 4, 137–142. [DOI] [PubMed] [Google Scholar]
35.Marty MT, Baldwin AJ, Marklund EG, Hochberg GKA, Benesch JLP, and Robinson CV (2015) Bayesian Deconvolution of Mass and Ion Mobility Spectra: From Binary Interactions to Polydisperse Ensembles, Anal. Chem. (Washington, DC, U. S.) 87, 4370–4376. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Chem. Biol.Supplemental

NIHMS1765311-supplement-Chem__Biol_Supplemental.pptx^{(3.4MB, pptx)}

[R1] 1.Yeagle PL (1985) Cholesterol and the cell membrane, Biochim. Biophys. Acta, Rev. Biomembr 822, 267–287. [DOI] [PubMed] [Google Scholar]

[R2] 2.Yeagle PL (1991) Modulation of membrane function by cholesterol, Biochimie 73, 1303–1310. [DOI] [PubMed] [Google Scholar]

[R3] 3.Lingwood D, and Simons K (2010) Lipid Rafts As a Membrane-Organizing Principle, Science (Washington, DC, U. S.) 327, 46–50. [DOI] [PubMed] [Google Scholar]

[R4] 4.Rog T, and Vattulainen I (2014) Cholesterol, sphingolipids, and glycolipids: What do we know about their role in raft-like membranes?, Chem. Phys. Lipids 184, 82–104. [DOI] [PubMed] [Google Scholar]

[R5] 5.Miller WL, and Auchus RJ (2011) The molecular biology, biochemistry, and physiology of human steroidogenesis and its disorders, Endocr. Rev 32, 81–151. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Brown MS, Radhakrishnan A, and Goldstein JL (2018) Retrospective on Cholesterol Homeostasis: The Central Role of Scap, Annu. Rev. Biochem 87, 783–807. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] 7.Kowatsch C, Woolley RE, Kinnebrew M, Rohatgi R, and Siebold C (2019) Structures of vertebrate Patched and Smoothened reveal intimate links between cholesterol and Hedgehog signalling, Curr. Opin. Struct. Biol 57, 204–214. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Mosqueira A, Camino PA, and Barrantes FJ (2018) Cholesterol modulates acetylcholine receptor diffusion by tuning confinement sojourns and nanocluster stability, Sci. Rep 8, 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Radhakrishnan A, Sun L-P, Kwon HJ, Brown MS, and Goldstein JL (2004) Direct binding of cholesterol to the purified membrane region of SCAP: Mechanism for a sterol-sensing domain, Mol. Cell 15, 259–268. [DOI] [PubMed] [Google Scholar]

[R10] 10.Cologna SM, and Rosenhouse-Dantsker A (2019) Insights into the molecular mechanisms of cholesterol binding to the NPC1 and NPC2 proteins, Adv. Exp. Med. Biol 1135, 139–160. [DOI] [PubMed] [Google Scholar]

[R11] 11.Hulce JJ, Cognetta AB, Niphakis MJ, Tully SE, and Cravatt BF (2013) Proteome-wide mapping of cholesterol-interacting proteins in mammalian cells, Nat. Methods 10, 259–264. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R12] 12.Cheng WWL, Budelier MM, Sugasawa Y, Bergdoll L, Queralt-Martin M, Rosencrans W, Rostovtseva TK, Chen Z-W, Abramson J, Krishnan K, Covey DF, Whitelegge JP, and Evers AS (2019) Multiple neurosteroid and cholesterol binding sites in voltage-dependent anion channel-1 determined by photo-affinity labeling, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 1864, 1269–1279. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Budelier MM, Cheng WWL, Chen Z-W, Bracamontes JR, Sugasawa Y, Krishnan K, Mydock-McGrane L, Covey DF, and Evers AS (2019) Common binding sites for cholesterol and neurosteroids on a pentameric ligand-gated ion channel, Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 1864, 128–136. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.Brunner J, Senn H, and Richards FM (1980) 3-Trifluoromethyl-3-phenyldiazirine. A new carbene generating group for photolabeling reagents, J. Biol. Chem 255, 3313–3318. [PubMed] [Google Scholar]

[R15] 15.Chen Z-W, Bracamontes J, Budelier MM, Germann AL, Shin DJ, Krishnan K, Qian M, Manion B, Cheng WWL, Reichert DE, Akk G, Covey DF, and Evers AS (2019) Multiple functional neurosteroid binding sites on GABAA receptors, PLoS Biology 17, e3000157. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Sugasawa Y, Cheng WWL, Bracamontes J, Chen Z-W, Wang L, Germann AL, Pierce SR, Senneff TC, Krishnan K, Reichert DE, Covey DF, Akk G, and Evers AS (2020) Site-specific effects of neurosteroids on GABAA receptor activation and desensitization, eLife 9, e55331. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.Feltes M, Moores S, Gale SE, Krishnan K, Mydock-McGrane L, Covey DF, Ory DS, and Schaffer JE (2019) Synthesis and characterization of diazirine alkyne probes for the study of intracellular cholesterol trafficking, Journal of lipid research 60, 707–716. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Chen DM, Ziolkowski L, Benz A, Qian M, Zorumski CF, Covey DF, and Mennerick S (2018) A clickable oxysterol photolabel retains NMDA receptor activity and accumulates in neurons., Front. Neurosci 12, 923. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Kaji Y, Koami T, Nakamura A, and Fujimoto Y (2000) Synthesis of 3-epi-6,7-dideoxyxestobergsterol A, Chem. Pharm. Bull 48, 1480–1483. [DOI] [PubMed] [Google Scholar]

[R20] 20.Covey DF, Hood WF, and Parikh VD (1981) 10β-propynyl-substituted steroids. Mechanism-based enzyme-activated irreversible inhibitors of estrogen biosynthesis, J Biol Chem 256, 1076–1079. [PubMed] [Google Scholar]

[R21] 21.Duval D, Desfosses B, and Emiliozzi R (1980) Preparation of dehydroepiandrosterone, testosterone and progesterone antigens through 7-carboxymethyl derivatives: characteristics of the antisera to testosterone and progesterone, Steroids 35, 235–249. [DOI] [PubMed] [Google Scholar]

[R22] 22.Dhameja M, and Pandey J (2018) Bestmann-Ohira Reagent: A Convenient and Promising Reagent in the Chemical World, Asian J. Org. Chem 7, 1502–1523. [Google Scholar]

[R23] 23.Wicha J, and Bal K (1978) Synthesis of 21-hydroxycholesterol and 25-hydroxycholesterol from 3β-hydroxyandrost-5-en-17-one. A method for the stereospecific construction of sterol side-chains, J. Chem. Soc., Perkin Trans 1, 1282–1288. [Google Scholar]

[R24] 24.Goldstein JL, DeBose-Boyd RA, and Brown MS (2006) Protein sensors for membrane sterols, Cell (Cambridge, MA, U. S.) 124, 35–46. [DOI] [PubMed] [Google Scholar]

[R25] 25.Gill S, Chow R, and Brown AJ (2008) Sterol regulators of cholesterol homeostasis and beyond: The oxysterol hypothesis revisited and revised, Prog. Lipid Res 47, 391–404. [DOI] [PubMed] [Google Scholar]

[R26] 26.Cheng WWL, Chen Z-W, Bracamontes JR, Budelier MM, Krishnan K, Shin DJ, Wang C, Jiang X, Covey DF, Akk G, and Evers AS (2018) Mapping two neurosteroid-modulatory sites in the prototypic pentameric ligand-gated ion channel GLIC, J. Biol. Chem 293, 3013–3027. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R27] 27.Barrantes FJ, and Fantini J (2016) From hopanoids to cholesterol: Molecular clocks of pentameric ligand-gated ion channels, Prog. Lipid Res 63, 1–13. [DOI] [PubMed] [Google Scholar]

[R28] 28.Laverty D, Thomas P, Field M, Andersen OJ, Gold MG, Biggin PC, Gielen M, and Smart TG (2017) Crystal structures of a GABAA-receptor chimera reveal new endogenous neurosteroid-binding sites, Nat. Struct. Mol. Biol 24, 977–985. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] 29.Mix KA, and Raines RT (2015) Optimized Diazo Scaffold for Protein Esterification, Org. Lett 17, 2358–2361. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R30] 30.Jiang X, Shu H-J, Krishnan K, Qian M, Taylor AA, Covey DF, Zorumski CF, and Mennerick S (2016) A clickable neurosteroid photolabel reveals selective Golgi compartmentalization with preferential impact on proximal inhibition, Neuropharmacology 108, 193–206. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Peyrot SM, Nachtergaele S, Luchetti G, Mydock-McGrane LK, Fujiwara H, Scherrer D, Jallouk A, Schlesinger PH, Ory DS, Covey DF, and Rohatgi R (2014) Tracking the Subcellular Fate of 20(S)-Hydroxycholesterol with Click Chemistry Reveals a Transport Pathway to the Golgi, J. Biol. Chem 289, 11095–11110. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Das J (2011) Aliphatic Diazirines as Photoaffinity Probes for Proteins: Recent Developments, Chem. Rev. (Washington, DC, U. S.) 111, 4405–4417. [DOI] [PubMed] [Google Scholar]

[R33] 33.Gimpl G, and Gehrig-Burger K (2011) Probes for studying cholesterol binding and cell biology, Steroids 76, 216–231. [DOI] [PubMed] [Google Scholar]

[R34] 34.Johnston JON, Wright CL, Holbert GW, and Benson HD (1990) Enzyme inactivation by potential metabolites of an aromatase-activated inhibitor (MDL 18,962), J. Enzyme Inhib 4, 137–142. [DOI] [PubMed] [Google Scholar]

[R35] 35.Marty MT, Baldwin AJ, Marklund EG, Hochberg GKA, Benesch JLP, and Robinson CV (2015) Bayesian Deconvolution of Mass and Ion Mobility Spectra: From Binary Interactions to Polydisperse Ensembles, Anal. Chem. (Washington, DC, U. S.) 87, 4370–4376. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Validation of trifluoromethylphenyl diazirine cholesterol analogues as cholesterol mimetics and photolabeling reagents

Kathiresan Krishnan

Mingxing Qian

McKenna Feltes

Zi-Wei Chen

Sarah Gale

Lei Wang

Yusuke Sugasawa

David E Reichert

Jean E Schaffer

Daniel S Ory

Alex S Evers

Douglas F Covey

Abstract

INTRODUCTION

Figure 1. Structures of cholesterol and PAL analogues of Cholesterol.

RESULTS

Chemistry.

Scheme 1.

Scheme 2.

Scheme 3.

Scheme 4.

Cellular Actions of Cholesterol PALs.

Figure 2. Functionalized cholesterol probes suppress SREBP2-target gene expression.

Figure 3. Functionalized cholesterol probes do not activate LXR-target gene expression.

Probe photolabeling of the Intact Ion Channel GLIC by MS.

Figure 5. Intact protein MS analysis of GLIC photolabeled with cholesterol photolabeling analogues.

Fluorescence Analysis of Photolabeled GLIC.

Identification of Sites of Cholesterol Analogue Adduction Using Middle-down MS.

Table 1.

Mapping of the Cholesterol Photolabeling Sites on the GLIC Structure.

Figure 6. Docking poses of cholesterol and selected photolabeling analogues in a TM3-4 intrasubunit site of GLIC.

Figure 7. Photolabeling of HEK cell membranes with cholesterol analogue PAL reagents.

CONCLUSION

Discussion.

Outlook for future studies.

METHODS

Synthetic procedures. (3β,17β)-17-[[4-[3-(Trifluoromethyl)-3H-diazirin-3-yl]phenyl]methoxy]-androst-5-en-3-ol (KK226, 5).

(3β,17β)-10-(2-Propyn-1-yl)-17-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methoxy]-estr-5-en-3-ol (KK231, 6).

(3β,7β,17β)-7-(2-Propyn-1-yl)-17-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methoxy]-androst-5-en-3-ol (KK237, 7).

(3β,17β)-17-[[[4-(3-Trifluoromethyl-3H-diazirin-3-yl)-phenyl]methyl]-[(1S)-prop-2-ynyl]]-androst-5-en-3-ol (MQ238, 8).

(3β,17β)-3-(Methoxymethoxy)-androst-5-en-17-ol (10).

3-[4-(3β,17β)-[[(3-Methoxymethoxy)-androst-5-en-17-yl]oxymethyl]phenyl]-3-trifluoromethyl-3H-diazirine (11).

(17β)-17-Hydroxy-10-(2-propyn-1-yl)-estr-4-en-3-one (13).

(17β)-10-(2-Propyn-1-yl)-estra-3,5-diene-3,17-diol, 3,17-diacetate (14).

(3β,17β)-10-(2-Propyn-1-yl)-estr-5-ene-3,17-diol, 17-acetate (15).

(3β,17β)-3-(Methoxymethoxy)-10-(2-propyn-1-yl)-estr-5-en-17-ol, 17-acetate (16).

(3β,17β)-3-(Methoxymethoxy)-10-(2-propyn-1-yl)-estr-5-en-17-ol (17).

3-[4-(3β,17β)-[(3-Methoxymethoxy)-10-(2-propyn-1-yl)-estr-5-en-17-yloxymethyl]phenyl]-3-(trifluoromethyl)-3H-diazirine (18).

(3β,7α)-3-(Acetyloxy)-7-bromo-androst-5-en-17-one, cyclic 1,2-ethanediyl acetal (19).

(3β,7β)-17,17-[1,2-Ethanediylbis(oxy)]-3-(acetyloxy)-androst-5-ene-7-acetic acid, ethyl ester (20).

(3β,7β)-17,17-[1,2-Ethanediylbis(oxy)]-3-hydroxy-androst-5-ene-7-acetic acid, ethyl ester (21).

(3β,7β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-17,17-[1,2-ethanediylbis(oxy)]-androst-5-en-7-acetic acid, ethyl ester (22).

(3β,7β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-hydroxyethyl)-androst-5-en-17-one, cyclic 1,2-ethanediyl acetal (23).

(3β,7β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-propyn-1-yl)-androst-5-en-17-one, cyclic 1,2-ethanediyl acetal (24).

(3β,7β)- 3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-propyn-1-yl)-androst-5-en-17-one (25).

(3β,7β,17β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-7-(2-propyn-1-yl)-androst-5-en-17-ol (26).

(3β,17E)-3-(Methoxymethoxy)-pregna-5,17(20)-dien-21-oic acid, ethyl ester (27).

(3β)-3-Hydroxypregn-5-en-21-oic acid, ethyl ester (28).

(3β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-pregn-5-en-21-oic acid, ethyl ester (29).

(3β,20S)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-20-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methyl]-pregn-5-en-21-oic acid, ethyl ester (30).

(3β,20S)- 3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-20-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methyl]-pregn-5-en-21-ol (31).

(3β,20S)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-20-[[4-[3-(trifluoromethyl)-3H-diazirin-3-yl]phenyl]methyl]-pregn-5-en-21-al (32).

3-[(2S)-4-[2-[(3β,17β)-3-[[(1,1-Dimethylethyl)dimethylsilyl]oxy]-androst-5-en-17-yl]-but-3-ynyl]-phenyl]- 3-(trifluoromethyl)-3H-diazirine (33).

Gene Expression Assays for HMGCS and ABCA1.

Cholesterol auxotroph viability assays.

Expression and purification of GLIC from E. coli.

Cell culture and membrane preparation.

Photolabeling and MS analysis.

Protein photolabeling, click cycloaddition reaction and gel analysis.

Computational Docking.

Supplementary Material

Figure 4. Trifluoromethylphenyl diazirine and alkyl diazirine sterols with side chain alkyne modifications support cholesterol auxotroph cell growth.

Acknowledgements

Footnotes

REFERENCES

Associated Data

Supplementary Materials

ACTIONS