Photoaffinity-Based Chemical Proteomics Reveals 7-Oxocallitrisic Acid Targets CPT1A to Trigger Lipogenesis Inhibition

Abstract

One of the natural terpenoids isolated from Resina Commiphora, 7-oxocallitrisic acid (7-OCA), has lipid metabolism regulatory activity. To uncover its lipogenesis inhibition mechanism, we developed a photoaffinity and clickable probe based on the 7-OCA scaffold and performed chemical proteomics to profile its potential cellular targets. It was found that 7-OCA could directly interact with carnitine palmitoyl transferase 1A (CPT1A) to promote its activity to reduce lipid accumulation. The present work reveals our understanding of the mode of lipid mebabolism regulation by abietic acids and provides new clues for antiobesity drug development with CPT1A as a main target.

Natural products have always been a rich source of new drugs for the treatment of various diseases. Terpenoids are the most numerous and structurally diverse natural products, many of which have been found in Chinese herbal medicines.¹ Abietic acids as diterpene compounds have special chemical structures and exhibit important biological effects such as antitumor, anti-inflammatory, antiobesity, antibacterial, and insecticide properties.² Despite the diverse biological roles of abietic acids, their targets remain unknown to date.

Our previous studies revealed structurally diverse terpenoids from Resina Commiphora.³ Among them, 7-oxocallitrisic acid (7-OCA) (Figure 1A), one of the abietic acids, was isolated in a considerable amount and is actually distributed widely in other natural resources such as Juniperus Chinensis,⁴Armillaria mellea,⁵ and the soil of the Japanase red pine forest floor.⁶ It is therefore necessary to explore its biomedical potential. Inspired by the traditional applications of Resina Commiphora(7) and with a lipid regulation assay in hand, the effects of 7-OCA on lipid regulation were thus evaluated. We chose the AML12 cell line preincubated with oleic acid (OA) as a biological model, since it is a normal mouse liver cell line and has been widely used for lipid regulation studies.⁸ In detail, AML12 cells were treated by 0.15 mM OA dissolved in BSA to induce the deposition of lipid droplets and then treated with two concentrations of 7-OCA for 24 h with DMSO as a control group. Then the cells were stained by Oil Red O (ORO) dye. The results showed that the lipid accumulation was considerably suppressed by 7-OCA at 20 and 40 μM (Figure 1B). The total triglyceride and total cholesterol determinations were also performed with the 7-OCA or DMSO treatment, and a certain decrease of the total triglyceride and a negligible decrease of the total cholesterol were observed (Figure 1C). These findings indicated that 7-OCA could be a potential lead compound for lipogenesis inhibition.

Figure 1.

(A) Structures of 7-OCA, SZU002, and SZU076. (B) Oil Red O staining of the lipid accumulation with treatments of 7-OCA, SZU002, and SZU076 (20 and 40 μM) in OA-treated AML12 cells for 24 h and quantification via ImageJ. (C) Enzymatic determination of the total triglyceride and total cholesterol with treatments of 7-OCA, SZU002 and SZU076 (40 μM) in OA-treated AML12 cells for 24 h. All data are expressed as mean ± SEM, n = 3. ^####p-value <0.0001 and ^#p-value <0.05 versus BSA + DMSO; ***p-value <0.001, **p-value <0.01, and *p-value <0.05 versus OA + DMSO.

However, the action mechanism of 7-OCA on the reduction of lipid accumulation has not been clarified. To address this question, we need to figure out the direct targets of 7-OCA in lipid metabolism systems. Activity-based protein profiling (ABPP)⁹ has emerged as a powerful strategy in the field of chemical proteomics to capture and identify proteins with specific enzymatic activities in native biological systems.¹⁰ Recently, ABPP has also been implemented to identify converse-small molecules as selective enzyme activators.¹¹ Using this strategy, target identification of many natural products has been accomplished in past decades. For instance, an artemisinin-based probe was developed to identify the target proteins of artemisinin and to study artemisinin–target interaction mechanisms.¹² Recently, Wang’s group revealed that a major flavonoid component, baicalin, could activate hepatic carnitine palmitoyl transferase 1 (CPT1) to ameliorate diet-induced obesity with a photoaffinity-based protein profiling method.¹³

In order to assess reversible 7-OCA-interacting proteins in living cells, we herein report the development of a cell-permeable photoaffinity-based probe SZU076 and a negative control probe SZU002 based on the parental 7-OCA (Figure 1A). The design of probes SZU002 and SZU076 required modifications that were introduced as small as possible to avoid affecting aromaticity and chirality. We therefore prepared a probe scaffold that combined a small alkyne handle (L1) or a “minimalist” diazirine photo cross-linker (L2).¹⁴ Based on literature reports, the linker was incorporated into the carboxylic acid group, which can tolerate a variety of modifications with no or little effect on the bioactivity of similar natural products,¹⁵ producing the probes through a one-step acid–amine coupling reaction in 56% and 80% yields (Schemes S1 and S2, respectively). Furthermore, the alkyne group allows modifications with azide-containing biotin or fluorophores by copper-catalyzed azide–alkyne cycloaddition (CuAAC).¹⁶

With the probes in hand, we first investigated their toxicity in living cells and activity on the reduction of lipid accumulation in AML12. The cell viability results showed no difference among a series of concentrations of 7-OCA and probes (Figure S1). The extent of lipid accumulation was also quantified by ORO staining after the treatment of the probes. The results proved that SZU002 and SZU076 exhibited abilities comparable to that of the original scaffold 7-OCA (Figure 1B). SZU002 and SZU076 were able to significantly decrease the accumulation of lipid droplets, indicating that the retention of bioactivity could be caused by the parental scaffold in probe’s structure. The considerable decrease in the total triglyceride and very slight decrease in the total cholesterol in the cells after incubation with 40 μM SZU002 or SZU076 (Figure 1C) are consistent with the phenotypes observed in ORO staining.

Next, to discover the direct targets of 7-OCA we first evaluated their labeling efficiency in a gel-based ABPP profile in living cells by fluorescence. After incubation of AML12 cells with a series of concentrations of SZU002 and SZU076 for 6 h each, the samples were irradiated for 10 min under 365 nm UV light to induce photo cross-linking. Then, the labeled proteins were conjugated with an azide-cyanine 5 (Cy5-N₃) or azide-biotin (Biotin-N₃) tag by CuAAC (Figure 2A). The in-gel fluorescence and Western blot results revealed distinct and dose-dependent labeling for the photoaffinity probe SZU076 and no labeling for SZU002 (Figure 2B and C), confirming that the interactions of the negative probe SZU002 with target proteins are noncovalent. In parallel, a competitive experiment was set up in which 20 μM parental molecule 7-OCA was mixed with SZU076 and incubated for for 24 h or 20 μM parental molecule 7-OCA was preincubated for 24 h before SZU076 incubation for 12 h to determine whether the probe and the parental compound compete for the same binding targets. From the in-gel visualization, multiple proteins were found to be blocked by 7-OCA (Figure 2D), and the positions of the bands were consistent with the SZU076 direct-labeling targets, indicating that the probe was indeed interacting with the same proteins as the parental molecule.

Figure 2.

(A) Photoaffinity-based pull-down procedure. (B and C) Dose response of probe labeling and (D) competitive ABPP labeling of cell lysates with probe SZU076 by preincubation or a mixture of 7-OCA. (E) Target profiles are shown by a volcano plots as the log₂ ratio of SZU076 to the SZU002 treatment against the log₁₀p-value (p-value ≤0.05 and log₂ ratio ≥1.5). (F) Functional cluster enrichment and potential target protein selectivity were performed by the gene oncology (GO) term and KEGG pathway analysis. (G) Western blot was used to analyze the pull-down sample with the target protein antibody.

We then performed an affinity pull-down experiment, which was followed by mass spectrometry-based quantitative proteomics. AML12 cells were preincubated with OA and incubated with DMSO, probe SZU002, or probe SZU076 in a medium for 6 h, and UV irradition was performed for 10 min at 365 nm. Cells were lysed at 4 °C by sonication, and the soluble and insoluble fractions were separated by centrifugation. Protein fractions were “click” reacted with Biotin-N₃, enriched on streptavidin magnetic beads, and released either by tryptic digest for the LC-MS/MS analysis or by a sample buffer containing 2% SDS at a high temperature for the Western blot analysis of the labeled proteins. The MS data files were analyzed with the MaxQuant program using a label-free quantitative proteomics strategy. MS/MS spectra were searched in the Andromeda search engine against the Mus musculus Uniprot database. The statistics analysis was mainly performed using R package DEqMS,¹⁷ which was developed on top of Limma. Total proteins were identified by filtering duplicates of each condition after the removal of the contaminant proteins. The filtered proteins were ranked in corresponding volcano plots as a log₂ ratio of the SZU076 treatment to the SZU002 treatment against statistical significance log₁₀p-value. In total, 125 proteins (shown as red and blue dots) (Figure 2E), with a p-value ≤0.05 and a log₂ ratio ≥1.5, were identified as potential targets selectively bound to SZU076 (Figure S2). We next used ClueGO to perform a gene oncology (GO) functional enrichment analysis. It showed that many proteins were located in the mitochondria and revealed top-ranking functional clusters of lipid metabolism (Figure S2). There were 5 proteins (ACSL1, ACOX2, CPT1A, ECHS1, and PCK2) in the in the GO term of the lipid metabolism. Then we performed a KEGG pathway analysis with these five proteins, and three subpathways (fatty acid degradation, the PPARs (peroxisome proliferator-activated receptors) signaling pathway, and the adipocytokine signaling pathway) were shown (Figure 2F). It was known that ACSL1 (acyl-CoA synthetase long chain family member 1), CPT1A (carnitine palmitoyl transferase 1A), and ECHS1 (enoyl-CoA hydratase 1) were directly involved in the lipid-chain degradation process.¹⁸ However, ACOX2 (peroxisomal acyl-coenzyme A oxidase 2) oxidizes the CoA esters of the bile acid intermediates di- and trihydroxycholestanoic acids, and PCK2 (phosphoenolpyruvate carboxykinase 2) just catalyzes the conversion of oxaloacetate to phosphoenolpyruvate.¹⁹ Therefore, we focused on CPT1A, ACSL1, and ECHS1 proteins involved in the “fatty acid degradation” pathway to further identify their functions and finally proved CPT1A as the main target of 7-OCA. Additionally, the central role of the target proteins in a cluster of lipid-related process pathway was CPT1A, a key protein that acts as a rate-limiting enzyme in mitochondrial lipid β-oxidation and slows the lipid accumulation process.²⁰ Moreover, we also found one called protein fatty acid desaturase 3 (FADS3),²¹ which was involved in the pathway of sphingolipid metabolism in the protein list and was also detected recently.

To confirm target–probe interactions identified in the proteomics data, we analyzed whether these proteins were present in pull-down on-beads samples using Western blot analysis. The results demonstrated that CPT1A, ACSL1, ECHS1, and FADS3 could be covalently captured by the photoaffinity probe SZU076 but not probe SZU002 in an UV-dependent manner (Figures 2G and S3). Since CPT1A is highly relevant to the observed phenotype of reduced lipid accumulation, we continued to study its functional ability to mediate the 7-OCA fatty acid-reducing effect. Small interfering RNA (siRNA) against CPT1A was introduced to knock down its expression, and 7-OCA’s effect of reducing lipid accumulation was almost abolished, as shown by the ORO staining and triglyceride quantitative assay experiments (Figure 3). The total cholesterol content was not changed by the 7-OCA treatment after CPT1A siRNA was introduced, in accordance with the phenotype observed in primary experiments. The reason might be that the lipid droplets in cells induced by OA belong to neutral lipid, which is mainly composed of triglyceride. Meanwhile, we introduced etomoxir (ETO),²² a potent and irreversible inhibitor of CPT1A, to validate the effect of 7-OCA. It showed that preincubation with ETO in living cells lead to the alteration of 7-OCA activity to reduce the lipid accumulation (Figure 3), which was consistent with results of the CPT1A RNA interference. We also investigated the knock down of ACSL1, ECHS1, and FADS3 by siRNA transfection in AML12 cells. However, 7-OCA’s effect of reducing lipid accumulation still worked (Figure S4). This indicated that ACSL1, ECHS1, and FADS3 might serve as the off-targets of our probe in lipid metabolism effects. Therefore, we hypothesized that 7-OCA might promote the lipid-reducing effect mainly by the positive regulation of CPT1A activities to accelerate fatty acid metabolism.

Figure 3.

(A) Oil Red O staining of the lipid accumulation with a treatment of 7-OCA (40 μM) in OA-treated AML12 cells for 24 h after either the RNA inference of CPT1A or a pretreatment of 20 μM ETO compared with the negative control (N.C) and quantification via ImageJ. (B) Enzymatic determination of the amount of triglyceride and total cholesterol with a treatment of 7-OCA (40 μM) in OA-treated AML12 cells for 24 h after either the RNA inference of CPT1A or a pretreatment of 20 μM ETO compared with the N.C. All data are expressed as mean ± SEM, n = 3. ^####p-value <0.0001, ^###p-value <0.001, and ^#p-value <0.05 versus BSA + DMSO; **p-value <0.01 versus OA + DMSO.

To further determine whether 7-OCA can activate CPT1A directly, we established a quantitative enzymatic assay, which uses l-carnitine and palmitoyl-CoA as the substrates and detects the production of CoA-SH using the general thiol reagent 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB), to measure the CPT1A activity (Figure 4). The strategy for the activity measurement was modified from a method described by Lee’s group.²³ For the activity assay, AML12 cells were pretreated with either ETO or DMSO for 6 h and then treated with 7-OCA for 24 h in the medium containing 0.15 mM OA. The lysates were prepared by sonication in a buffer composed of 0.5 M Tris, 0.15 M KCl, and a 1 × protease inhibitor cocktail (Roche) using sonication. Activity was defined as the nmol CoA-SH released per minute per milligram of protein. The protein concentration of the cell homogenates was determined using the Bradford method. In this assay, lysates of 20 and 40 μM 7-OCA-treated cells showed 1.5-fold and 1.6-fold increases in CPT1A activity, respectively, compared with the DMSO group (Figure 4A). However, there was no significant difference between the absence and presence of 20 and 40 μM 7-OCA in the ETO pretreatment group since ETO blocked the activity of CPT1A remarkably. The reason probably is that the irreversible inhibition of ETO leads to 7-OCA being incapable of interacting with CPT1A. Additionally, these results are consistent with the conclusions of previous lipid metabolism phenotypic experiments. Further Western blot results confirmed that the 7-OCA treatment did not affect the protein expression of CPT1A (Figure 4B). Therefore, it were suggested that pretreatment with the CPT1A inhibitor ETO could induce 7-OCA insensitivity to regulate CPT1A.

Figure 4.

(A) Enzyme activity assay of CPT1A with a pretreatment of either ETO or DMSO and a treatment of 7-OCA (20 and 40 μM) in OA-treated AML12 cells for 48 h. (B) Western blot analyses of the protein expression of CPT1A with a treatment of 7-OCA (20 and 40 μM) in OA-treated AML12 cells for 48 h. All data are expressed as mean ± SEM, n = 3. **p-value <0.01 vs OA + DMSO.

In summary, our present research describes that a photoaffinity and clickable activity-based probe SZU076 and a negative control probe SZU002, which were prepared from one abietic acid scaffold. The probes and their prototype 7-OCA a showed similar bioactivity of reducing lipid accumulation in cells. Through chemical proteomics studies, we discovered that 7-OCA directly targets CPT1A and activates it to effect lipogenesis inhibition. These results could be valuable clues for understanding the antiobesity bioactivities of abietic acids and their derivatives. These findings also imply that 7-OCA (abundant in nature) based derivatives might be good candidates for leading compound discovery in terms of lipid regulation through targeting proteins like CPT1. Last but not the least, our work could also provide the basis for the further identification of other bioactive targets of terpenoids.

Glossary

ABBREVIATIONS

ABPP

activity-based protein profiling

ACSL1

acyl-CoA synthetase long chain family member 1

AML12

α-mouse liver 12

BSA

bovine serum albumin

CPT1A

carnitine palmitoyl transferase 1A

CuAAC

copper-catalyzed azide–alkyne cycloaddition

Cy5

cyanine 5

DMSO

dimethyl sulfoxide

DNTB

5,5′-dithio-bis(2-nitrobenzoic acid

ECHS1

enoyl-CoA hydratase 1

ETO

etomoxir

FADS3

fatty acid desaturase 3

GO

gene ontology

KEGG

kyoto encyclopedia of genes and genomes

OA

oleic acid

7-OCA

7-oxocallitrisic acid

ORO

Oil Red O dye

PPARs

peroxisome proliferator-activated receptors

UV

ultraviolet

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.1c00316.

• Additional information about the synthesis procedure and biological assays (PDF)

• Proteomics data (XLSX)

Author Contributions

^‡ These authors contributed equally.

The research was supported by National Key Research and Development Program of China (2018YFA0902703 and 2017YFA0503900), the National Natural Science Foundation of China (21807106, 31800694, 81525026, and 31971354), the Natural Science Foundation of Guangdong Province (2018A030310035), the Shenzhen Science and Technology Innovation Committee (JCYJ20180302145723601, JCYJ20170818164014753, JCYJ20180507182250795, and JCYJ20200109114003921), the Natural Science Research Project of Shenzhen University (2018019), and the SZU Top Ranking Project (860/00000210 and 827/000568).

The authors declare no competing financial interest.