Aldolase B attenuates clear cell renal cell carcinoma progression by inhibiting CtBP2

Abstract

Aldolase B (ALDOB), a glycolytic enzyme, is uniformly depleted in clear cell renal cell carcinoma (ccRCC) tissues. We previously showed that ALDOB inhibited proliferation through a mechanism independent of its enzymatic activity in ccRCC, but the mechanism was not unequivocally identified. We showed that the corepressor C-terminal-binding protein 2 (CtBP2) is a novel ALDOB-interacting protein in ccRCC. The CtBP2-to-ALDOB expression ratio in clinical samples was correlated with the expression of CtBP2 target genes and was associated with shorter survival. ALDOB inhibited CtBP2-mediated repression of multiple cell cycle inhibitor, proapoptotic, and epithelial marker genes. Furthermore, ALDOB overexpression decreased the proliferation and migration of ccRCC cells in an ALDOB-CtBP2 interaction-dependent manner. Mechanistically, our findings showed that ALDOB recruited acireductone dioxygenase 1, which catalyzes the synthesis of an endogenous inhibitor of CtBP2, 4-methylthio 2-oxobutyric acid. ALDOB functions as a scaffold to bring acireductone dioxygenase and CtBP2 in close proximity to potentiate acireductone dioxygenase-mediated inhibition of CtBP2, and this scaffolding effect was independent of ALDOB enzymatic activity. Moreover, increased ALDOB expression inhibited tumor growth in a xenograft model and decreased lung metastasis in vivo. Our findings reveal that ALDOB is a negative regulator of CtBP2 and inhibits tumor growth and metastasis in ccRCC.

Keywords

ALDOB / kidney cancer / cell proliferation

1 Introduction

Clear cell renal cell carcinoma (ccRCC) is a common tumor of the urinary system, and its global incidence has increased annually [1]. The occurrence of ccRCC is closely related to von Hippel–Lindau (VHL) gene inactivation [2]. The VHL protein is critical for the ubiquitination of hypoxia-inducible factor-1 (HIF-1), and the loss of VHL activity leads to the accumulation of HIF-1, which causes the dysregulation of a series of glycolysis-, angiogenesis-, and tumor cell survival-related genes [3]. Several studies conducted in recent years have shown that some glycolytic enzymes are localized to multiple cell compartments and perform multiple nonmetabolic functions [4].

Aldolase is an enzyme in the glycolytic pathway that catalyzes the cleavage of fructose 1,6-biphosphate (FBP) to generate glyceraldehyde 3-phosphate and dihydroxyacetone phosphate (which simultaneously catalyzes the reverse reaction). Human aldolase is expressed as three isoenzymes: aldolase A (ALDOA) is mainly present in muscle tissue and red blood cells; ALDOB is mainly present in the kidney and liver; and aldolase C (ALDOC) is expressed at the highest levels in the central nervous system. ALDOA and ALDOC show 81% sequence identity, and the sequence differences between ALDOB and other aldolases are greater, with only 70% sequence identity. ALDOB also participates in fructose metabolism, catalyzing the production of fructose 1-phosphate (F1P) to generate dihydroxyacetone phosphate and glyceraldehyde. Additionally, aldolase may be involved in glucose sensing [5]. ALDOB potentiates p53-mediated inhibition of glucose 6-phosphate dehydrogenase (G6PD) through the ALDOB–G6PD–p53 complex [6].

Our previous metabolic studies revealed that, in contrast to other glycolytic enzymes that are upregulated in tumors [7–9], ALDOB, the major aldolase subtype expressed in renal tubular cells, is significantly downregulated in ccRCC. Downregulation of ALDOB causes the accumulation of upstream metabolites and prevents oxidative stress-induced injury [10]. However, we also discovered that ALDOB mutants without enzymatic activity could inhibit the proliferation of ccRCC cells through a mechanism independent of their metabolic activity. The underlying mechanisms are unclear.

In the present study, oncogenic transcriptional corepressor C-terminal binding protein 2 (CtBP2) was identified as a novel ALDOB-interacting protein. ALDOB recruited acireductone dioxygenase (ADI1) and impaired transcriptional repression mediated by CtBP2. This study implicates ALDOB as a negative regulator of tumor progression and provides additional mechanistic insights into the regulatory effect of CtBP2.

2 Materials and methods

2.1 Human tissue samples

Tissue samples were obtained from the Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine. The investigation was conducted in accordance with ethical standards and was approved by the Institutional Review Board at the authors’ institutions (2020SQ192).

2.2 Cell culture and transfection

The 786-O, Caki-1, 769-P, OSRC-2, and HK-2 cell lines were obtained from the Type Culture Collection of the Chinese Academy of Sciences, Shanghai, China. All these cells were cultured in RPMI-1640 (Thermo Fisher Scientific, USA) supplemented with 10% FBS. In terms of transfection, cells were grown to 60%–70% confluence and then transiently transfected with plasmids or short interfering RNAs (siRNAs) using Lipofectamine 3000 or RNAiMax transfection reagent (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. The sequences of the siRNAs targeting siCTBP2 were 5′-GUGAUCGUGCGGAUAGGCAGU-3′ for siCTBP2 #1 and 5′-CACCCUGCUCUACAAUGUUGC-3′ for siCTBP2 #2. The sequences of the siRNAs targeting ADI1 were 5′-UAGUUUAUCUUUGCAUAUGGU-3′ for siADI1 #1 and 5′- AUUUGGUAGUUUAUCUUUGCA-3′ for siADI1 #2.

2.3 Expression constructs

Human ALDOB, CtBP2, and ADI1 cDNAs were amplified from HK-2 cells and subcloned into modified pIRESpuro and pCDH-CMV-MCS-EF1-puro vectors. The deletion and point mutant constructs of ALDOB or CtBP2 were generated with a KOD-Plus mutagenesis kit (TOYOBO).

2.4 Antibodies

Human anti-ALDOB, anti-tubulin, and anti-Lamin B1 antibodies were purchased from Proteintech (Wuhan, China), and the human anti-CtBP2 antibody was purchased from Cell Signaling Technology (Beverly, MA, USA). Human anti-Myc and anti-Flag antibodies were purchased from Sigma (St. Louis, MO, USA). A human anti-HA antibody was purchased from Millipore (Darmstadt, Germany). A human anti-ADI1 antibody was purchased from Abcam (Cambridge, UK).

2.5 Quantitative RT–PCR

Total RNA was extracted from cells or tissue using TRIzol (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. Total RNA was reverse transcribed into cDNA using PrimeScript™ RT reagent (Takara Bio Inc., Japan). Quantitative real-time PCR (qRT–PCR) was performed using a SYBR green Premix Ex Taq^TM kit (TaKaRa Bio Inc., Japan) according to the manufacturer’s instructions. The sequences of the primers are presented in Table S1.

2.6 Chromatin immunoprecipitation (ChIP)

Chromatin immunoprecipitation (ChIP) was performed as previously described with minor modifications [11]. Briefly, pooled treated cells were cross-linked with formaldehyde, lysed, sonicated, and cleared; Protein G Dynabeads were incubated with 1 μg of control IgG or anti-Flag (Sigma–Aldrich, USA) antibody for 3 h. ChIP samples were incubated with antibody-Dynabead Protein G complexes at 4 °C for 3 h. DNA was purified from input and ChIP samples using 10% Chelex-100 (Bio-Rad, USA). The purified DNA was quantified using real-time polymerase chain reaction (PCR) with CDH1 (E-cadherin), PTEN, CDKN1A (p21), and BAX primers that were previously validated to detect the corresponding human promoter region [12–15]. The sequences of the primers are presented in Table S2.

2.7 Luciferase assay

A CDH1 (E-cadherin) promoter-luciferase reporter containing the region from −427 to +53 was constructed using PCR and inserted into a pGL3-luciferase (luc) vector (Protégé). The indicated cells were transfected with the luciferase reporter and β-galactosidase using Lipofectamine 3000 (Thermo Fisher Scientific, USA). The total amount of transfected DNA was normalized using pcDNA3. β-Galactosidase was used as a control to normalize transfection efficiency. After 48 h, the cells were harvested and assayed using a luciferase reporter assay system (Promega, USA) and a beta-Gal assay kit (Clontech, USA) according to the manufacturers’ instructions.

2.8 Coimmunoprecipitation

For coimmunoprecipitation of the Flag-tagged proteins, cell lysates were immunoprecipitated with anti-FLAG M2 beads (Sigma–Aldrich, USA) overnight at 4 °C. For coimmunoprecipitation of endogenous proteins or other-tagged proteins, the cell lysates were precleared with Protein A/G beads (Sigma–Aldrich, USA) and incubated with the indicated antibody overnight at 4 °C. The immunocomplexes were then incubated for 2 h at 4 °C with Protein A/G beads. After centrifugation at 1000 rpm for 30 s, the pellets were collected, washed five times with lysis buffer (pH 7.4 Triton X-100 (0.8%), 150 mM Tris-HCl, and 400 mM NaCl), resuspended in sample buffer, and further analyzed on SDS–PAGE gels.

2.9 Prediction of protein–protein interactions

Protein–protein interactions were predicted using the mask multiple parallel convolutional neural network DeepTrio [16]. The DeepTrio models and training data are deposited at the website of Github. The model was run on a server with an AMD EPYC 7552, 128 GB of memory and GTX3090. Protein sequential information and the accession identifiers were uploaded onto the server and inputted into the deep learning model on Ubuntu 20.04. The detailed process of deep learning prediction is provided in the Supplementary Methods.

2.10 Protein preparation and glutathione S-transferase (GST) pulldown assay

Recombinant GST and GST-CtBPs proteins were prepared as previously described [17,18]. Another Flag-tagged protein was affinity purified on M2-anti-Flag agarose. The purified protein was added to glutathione Sepharose 4B (GE Healthcare)-conjugated GST fusion proteins. After rotation for 2 h at 4 °C, the beads were washed four times with lysis buffer. The proteins were eluted and subjected to Western blot (WB) analysis using the indicated antibodies.

2.11 Immunofluorescence staining

Briefly, the tissue slices were first fixed, washed three times with phosphate-buffered saline (PBS), and incubated overnight at 4 °C with the following primary antibodies: anti-ALDOB (Proteintech Group, USA) and anti-CtBP2 (Novus Biologicals, USA). All images were photographed at the same exposure time with a Nikon C2 Plus confocal microscope.

2.12 Aldolase enzyme activity assays

The substrate cleavage rate was determined spectrophotometrically by measuring NADH oxidation at 340 nm and 30 °C in a coupled assay with α-glycerol-phosphate dehydrogenase/triose phosphate isomerase (GDH/TIM) and FBP or F1P as substrates [19]. Activity assays were performed in a final volume of 200 µL of a mixture of the cell extract and the appropriate substrate concentrations of 0.2 mM NADH, 0.5 mM EDTA, 100 mM Tris/HCl, pH 7.5, and 10 µg/mL GDH/TIM.

2.13 ADI1 enzyme activity assays

ADI1 enzymatic activity was measured as previously described [20,21]. The experiment was performed in assay buffer containing 50 mM HEPES, pH 7.0, and 1 mM MgCl₂. The consumption rate of the substrate desthio-acireductone by ADI1 was monitored at 308 nm. A blank control was used to monitor the nonenzymatic decay of desthioacireductone. The average consumption rates were calculated by selecting the linear portion and calculating the linear fit in this region after correction for the nonenzymatic reaction rate.

2.14 Cell proliferation

Cells were plated in 6-well plates at a density of 4000 cells/well; the day of seeding was considered day 1. Cell numbers were counted on days 2, 3, and 4 using the trypan blue exclusion method.

For bromodeoxyuridine (BrdU) analysis, cells were pulsed with 10 μM BrdU (BD Biosciences, USA) for 5 h, trypsinized, fixed, and labeled using a BrdU flow kit (BD Biosciences, USA) according to the manufacturer’s instructions. Seven-amino actinomycin D (7-AAD) was used as a DNA content indicator. The cells were analyzed using flow cytometry with a FACSCalibur flow cytometer (BD Biosciences, USA).

2.15 Cell death assay

ccRCC cells were seeded on 6-well culture plates. After attachment for 12 h, the cells were treated with 8 nM paclitaxel. After incubation for the indicated periods, the attached cells were trypsinized, and both attached and unattached cells were harvested. The rate of cell death was measured using the trypan blue exclusion assay as previously described [22].

2.16 Animal studies

All animal experiments complied with the ARRIVE guidelines and were conducted in accordance with the UK Animals (Scientific Procedures) Act, 1986 and associated guidelines, EU Directive 2010/63/EU, for animal experiments. All animal experiments were conducted using protocols approved by the Institutional Animal Care and Use Committee of Shanghai General Hospital (2021AWS0154, 2022AW006). For establishment of the lung metastasis mouse model, stable ALDOB^WT/ALDOB^A318E and luciferase expression in Caki-1 cells were achieved by infecting parental cells with the corresponding lentivirus. Six-week-old male nude mice were injected with 5 × 10⁵ of these cells resuspended in PBS via the tail vein. After 40 days, mice were anesthetized with isoflurane gas, and metastatic tumors were visualized using a bioluminescent imaging system (PerkinElmer IVIS Spectrum). The animals were sacrificed, and the number of lung tumors was counted and plotted as shown.

For the subcutaneous xenograft tumor model, six-week-old male nude mice were ordered from Shanghai SLAC Laboratory Animal Co. Ltd. and subjected to subcutaneous tumorigenesis. Briefly, 1 × 10⁷ of the indicated cells were resuspended in 200 μL of PBS and subcutaneously injected into the flanks of nude mice. Four weeks after injection, the tumors were excised and weighed.

2.17 Nonnegative Matrix Factorization (NMF) analysis

The NMF molecular classification model was constructed using the rNMF [23] software package. Read counts of RNAseq data were extracted, and the genes with a low standard deviation (standard deviation < 0.5) were removed, as described in the literature [24]. The brunet algorithm was applied with κ = 3.

2.18 Statistical analysis

Statistical analyses were performed using R version 3.2.3 software. Pearson’s correlation analyses were performed between the CtBP2-to-ALDOB expression ratio and the expression levels of the indicated mRNA or protein. A median cutoff of the CtBP2-to-ALDOB expression ratio was used in the Kaplan–Meier analysis survival analysis. Independent experiments were performed at least three times. The P values were 2-sided, and P < 0.05 was considered statistically significant.

3 Results

3.1 ALDOB interacts with CtBP2

In a previous study, we found that the catalysis-defective ALDOB K147A mutant inhibited the proliferation of ccRCC cells, but the rate was reduced compared with that induced by wild-type ALDOB (Fig.1) [25]. We used the state-of-the-art deep learning-based DeepTrio tool to filter a panel of oncogenes and tumor suppressors to determine their capacity to interact with ALDOB and to obtain mechanistic insights into the function of ALDOB (Table S3) [16]. In particular, we were intrigued by the top ranked candidate, CtBP2, a transcriptional corepressor known to be involved in cancer initiation, progression, and metastasis [26–28]. Here, we determined the interaction between ALDOB and CtBP2 by performing coimmunoprecipitation (Co-IP). Flag-ALDOB coimmunoprecipitated with Myc-CtBP2 in 293T cells (Fig.1). Endogenous ALDOB and CtBP2 proteins also coimmunoprecipitated with each other (Fig.1 and 1D). We speculated that this candidate protein might explain the nonmetabolic function of ALDOB in ccRCC.

We assessed the interactions of recombinant proteins in vitro to determine whether the interaction between ALDOB and CtBP2 was direct and whether other members of the aldolase family interacted with members of the CtBP family. Pull-down assays revealed that GST-CtBP2 strongly interacted with Flag-ALDOB, whereas GST-CtBP1 showed a weak interaction with Flag-ALDOB (Fig.1). However, we did not detect Flag-tagged ALDOA or ALDOC in the immunoprecipitates of GST-CtBP2 (Fig.1). Furthermore, ALDOB and CtBP2 were colocalized in the nucleus (Fig.1).

Because the results of experiments assessing the function of CtBP2 have been primarily based on prostate, gastric, and lung cancers [26,27,29], we validated the role of CtBP2 in ccRCC proliferation and migration (Fig.1‒1J and Fig. S1). CtBP2 knockdown impaired the proliferation and migration of ccRCC cells.

3.2 Identification of the mutual binding regions of ALDOB and CtBP2

Two types of proteins are known to bind to CtBPs. In one category, classic proteins containing the -Pro-X-Asp-Leu-Ser- (PXDLS) motif are mostly transcriptional repressors or histone modification enzymes [30]. In the other category, nonclassical proteins, such as ALDOB, do not contain the PXDLS motif, do not have a common binding site, and have diverse functions [31].

We generated a number of ALDOB truncation mutants to determine the domains in ALDOB that were critical for binding CtBP2 (Fig.1). We cotransfected all ALDOB fragments into 293T cells and performed a Co-IP experiment. ALDOB regions consisting of amino acids 0–67 and 290–364 contained domains that coimmunoprecipitated with CtBP2 (Fig.1).

We generated fragments containing the N-terminal PXDLS domain, central dehydrogenase domain, and C-terminal extension of CtBP2 to analyze which domains interacted with ALDOB. We cotransfected these CtBP2 fragments with ALDOB in 293T cells and performed a Co-IP experiment. The CtBP2 dehydrogenase domain coimmunoprecipitated with ALDOB (Fig.1). This finding suggested that the dehydrogenase domain of CtBP2 interacted with ALDOB.

3.3 ALDOB binding to CtBP2 is specifically inhibited by fructose 1,6-bisphosphate

Because the interactions between ALDOB and some of its protein partners are regulated by metabolites such as fructose 1,6-bisphosphate (FBP), we sought to determine whether FBP affected the interaction of ALDOB and CtBP2 [5,32]. Less ALDOB protein coimmunoprecipitated with the CtBP2 protein after FBP treatment (Fig.1). Similarly, less endogenous CtBP2 protein coimmunoprecipitated with ALDOB after FBP treatment (Fig.1). In contrast to FBP, neither fructose 6-phosphate nor NADH affected the interaction between ALDOB and CtBP2 (data not shown). These findings suggested that ALDOB binding to CtBP2 was specifically inhibited by FBP.

3.4 The CtBP2-to-ALDOB expression ratio correlates with the expression of CtBP2 target genes and is associated with shorter survival

We aimed to further extend our observations to clinicopathologically relevant settings, and an analysis of a public data set (GSE53757) showed significantly lower expression of ALDOB and higher expression of CtBP2 in ccRCC tissues than in normal kidney tissues (Fig. S2A‒S2C). Interestingly, we observed a greater degree of separation in the CtBP2-to-ALDOB expression ratio between normal kidney tissues and ccRCC tissues (Fig. S2A). We validated these results in human ccRCC samples (Fig.2‒2C). Furthermore, the CtBP2-to-ALDOB expression ratio correlated with the mRNA level of the CtBP2-repressed target genes CDH1, CDKN1C, BIK, and BAX in both the human ccRCC samples and another public data set (GSE73731) (Fig.2‒2G, Fig. S2D‒S2G). Similarly, the expression ratio of CtBP2 and ALDOB correlated with the mRNA levels of the CtBP2 target genes CDH1 and CDKN1C in the ccRCC data sets of The Cancer Genome Atlas (TCGA) (Fig.2 and 2I). The CtBP2-to-ALDOB expression ratio also correlated with p21 and Bax protein levels in the TCGA ccRCC data sets (Fig.2 and 2K).

As we previously reported that ccRCC was classified into three distinct clusters (subtypes) based on the unsupervised classification algorithm NMF (Fig.2), we compared the CtBP2-to-ALDOB expression ratio in these ccRCC subtypes using the TCGA database. Interestingly, the ccRCC subtypes were characterized by the CtBP2-to-ALDOB expression ratio (Fig.2). Cluster 1, which was associated with the worst prognosis, showed the highest CtBP2-to-ALDOB expression ratio, while cluster 3, which was associated with the best prognosis, showed the lowest CtBP2-to-ALDOB expression ratio (Fig.2). We further explored the direct relationship between the expression ratio of CtBP2 and ALDOB and the prognosis of patients with ccRCC by performing a survival analysis based on the CtBP2-to-ALDOB expression ratio. Kaplan–Meier ccRCC overall and disease-free survival curves are shown in Fig.2 and 2P. A high CtBP2-to-ALDOB expression ratio was associated with a significantly higher risk of disease progression (P = 0.00021) and shorter survival (P = 1.2e‒6) (Fig.2 and 2P ).

3.5 ALDOB releases CtBP2-mediated repression of target genes

We upregulated ALDOB expression and performed RT–qPCR assays of CtBP2 target genes to confirm the role of ALDOB in regulating CtBP2 transcriptional activity. Overexpression of ALDOB restored (upregulated) the transcription of CDH1, PTEN, CDKN1A, CDKN2C, and BAX in both Caki-1 and 769-P cells (Fig.3 and 3B). In contrast, we observed no alterations in the enzymatic activity of aldolase in CtBP2-depleted Caki-1 or OSRC-2 cells (Fig.3). Taken together, these data suggested that ALDOB might negatively regulate the corepressor activity of CtBP2.

We subsequently investigated the mechanism by which ALDOB counteracted the effects of CtBP2. Overexpression of ALDOB neither changed the total level of the CtBP2 protein nor altered the localization of CtBP2 (Fig.3 and 3E). We further investigated the effects of ALDOB on CtBP2 transcriptional activity and DNA binding. For the luciferase reporter assay, a reporter plasmid carrying the E-cadherin promoter, a known direct target of CtBP2-mediated transcriptional repression [33], was transfected into 786-O cells stably expressing Ctrl and ALDOB (Fig.3). In 786-O cells expressing ALDOB, the activity of the CDH1 promoter was substantially upregulated (Fig.3). However, in the context of CtBP2 knockdown, ALDOB did not significantly alter CDH1 transcription. Moreover, ChIP analysis of CtBP2 revealed that ALDOB overexpression resulted in reduced binding of CtBP2 to the promoter regions of the CDH1, PTEN, CDKN1A, and BAX genes (Fig.3).

Furthermore, after considering the identified binding regions of ALDOB and CtBP2 and the reported ALDOB mutations, we screened ALDOB mutants K147A, K230A, R304A, and A318E [6,34]. We found that the ALDOB R304A and A318E mutants exhibited deficient interactions with CtBP2 (Fig.3). More importantly, the A318E mutant exhibited normal aldolase enzymatic activities (Fig.3); R304A has 50% enzymatic activity [6]. The ALDOB A318E mutant partially abolished the repressive effect of ALDOB on CtBP2 transcriptional repression (Fig.3 and 3L). Together, these data revealed that ALDOB functioned as a negative regulator of CtBP2.

3.6 ALDOB negates CtBP2-driven cell proliferation and migration in an ALDOB-CtBP2 interaction-dependent manner

An increase in CtBP2 expression levels reportedly promotes cell proliferation and metastatic potential in several cancers [26,27,29]. Based on the effect of aldolase binding to CtBP2, we hypothesized that ALDOB reversed CtBP2-mediated oncogenic phenotypes. A comparison with control cells revealed the diminished proliferative capacity of ALDOB^WT-expressing cells (Fig.4–4C). Compared with ALDOB^WT, the CtBP2 binding-defective ALDOB^A318E mutant was less effective in suppressing the proliferation of ccRCC cells (Fig.4–4C).

The BrdU incorporation assay showed that the percentages of Caki-1 and 786-O cells stably expressing ALDOB in S phase were decreased (Fig.4 and 4E). The ALDOB A318E mutation partially abolished the repressive effect of ALDOB on BrdU incorporation (Fig.4 and 4E)

Although ALDOB alone did not alter the baseline apoptosis level, ALDOB overexpression slightly increased the percentage of apoptotic cells induced by low levels of staurosporine (Fig.4 and 4G). The A318E mutation failed to abolish the CtBP2-induced antiapoptotic effect (Fig.4 and 4G). However, overexpression of ALDOB did not increase the percentage of dead cells induced by paclitaxel (Fig. S3).

One of the most characteristic effects of CtBP2 on the cellular phenotype is mediated by its promotion of cell migration and tumor metastasis via the epithelial-to-mesenchymal transition [35]. We further examined the effects of ALDOB on ccRCC cell migration and metastasis. ALDOB inhibited ccRCC migration, as evidenced by the results of wound healing and Transwell assays (Fig.4–4K); this finding was consistent with the increased expression of the epithelial marker E-cadherin. The A318E mutation of ALDOB, which blocked the interaction of ALDOB with CtBP2, partially abolished the effect of ALDOB on migration (Fig.4–4K).

3.7 ALDOB inhibits CtBP2 by recruiting acireductone dioxygenase

Another member of the aldolase family, ALDOA, has been reported to interact with acireductone dioxygenase (ADI1) in the methionine salvage pathway [36]. ADI1 catalyzes the synthesis of a precursor of methionine, 4-methylthio 2-oxobutyric acid (MTOB), which is the most powerful endogenous inhibitor of CtBP2. We showed that ALDOB also directly interacted with ADI1 through Co-IP and GST pulldown assays (Fig.5 and 5B). The N terminus of ALDOB was responsible for binding ADI1 (Fig.5). The GST pulldown assay showed that GST-CtBP2 did not directly interact with ADI1 (Fig.5). ALDOB acted as a scaffold protein bridging between ADI1 and CtBP2 (Fig.5). The addition of FBP inhibited the formation of the CtBP2/ALDOB/ADI1 ternary complex (Fig.5). ADI1 did not compete with CtBP2 for binding ALDOB (Fig.5).

Knockdown of ADI1 or the catalysis-defective E94A mutant abolished the effect of ALDOB on CtBP2-mediated transcriptional repression (Fig.5–5K). The addition of MTOB partially restored the inhibitory activity of CtBP2 (Fig.5 and 5H). The catalysis-defective ALDOB K147A mutant did not significantly alter ADI1-mediated inhibition of CtBP2, indicating that the scaffolding function of ALDOB did not rely on its enzyme activity (Fig.5 and 5M). These findings indicated that ALDOB recruited ADI1 and linked ADI1 and CtBP2 together to potentiate the MTOB-mediated inhibition of CtBP2 by protecting MTOB from consumption by competing reactions catalyzed by other enzymes.

Furthermore, the A318E mutation of ALDOB did not block the interaction between ADI1 and ALDOB, but the A318E mutation of ALDOB diminished the level of CtBP2 protein that coimmunoprecipitated with ADI1 protein (Fig.5 and 5O).

We investigated the effects of ALDOB expression on the growth of xenograft tumors in nude mice to verify these in vitro findings. As expected, compared with control cells, ccRCC cells stably expressing ALDOB developed significantly smaller tumors (Fig.5). Moreover, an A318E mutation of ALDOB that blocked the interaction of CtBP2 with ALDOB also separated CtBP2 and ADI1 and partially restored the growth of xenograft tumors in vivo (Fig.5).

Caki-1 cells engineered to stably express firefly luciferase were infected with lentiviruses carrying an empty vector, ALDOB^WT, and ALDOB^A318E mutant to investigate the role of ALDOB in ccRCC cell dissemination and metastasis in vivo. Mouse models of metastasis were established via a tail vein injection of these cells. Compared with mice in the control group, Caki-1 cells stably expressing ALDOB produced fewer and smaller lung metastasis nodules, while the lung metastases in nude mice injected with the A318E mutant cells were significantly restored (Fig.5 and 5R). Collectively, these data indicated that the migration, metastasis, and proliferation of ccRCC are negatively regulated by ALDOB via its scaffolding function that bridges CtBP2 and ADI1.

4 Discussion

ccRCC is characterized by a reprogramming of aerobic glycolysis, impaired mitochondrial bioenergetics and oxidative phosphorylation, as well as dysregulated lipid metabolism [37–42]. In cross-cancer studies, an analysis of the metabolic gene expression profile showed that ccRCC is also characterized by the concerted downregulation of many metabolic genes compared with other malignancies [43]. Fructose-1,6-bisphosphatase 1 (FBP1) has been reported to be uniformly depleted in ccRCC, and FBP1 reportedly inhibits renal carcinoma progression through multiple mechanisms [44]. Another substantially depleted metabolic gene in ccRCC, ALDOB, plays an unknown role, and its underlying mechanism(s) of action remains obscure. We report in this study that ALDOB physically interacts with CtBP2 in ccRCC cells and recruits ADI1, which inhibits the transcriptional repression mediated by CtBP2 and sequentially inhibits ccRCC proliferation and metastasis (Fig.6).

The effect of ALDOB seems to depend on the context. ALDOB induces apoptosis in colon adenocarcinoma and inhibits hepatic cancer intrahepatic metastasis and lung metastasis [45,46]. However, ALDOB is required for the growth of colon cancer liver metastasis, but not for that of primary colon tumors or liver metastasis [47].

Indeed, in the present study, we found that the function of ALDOB also depends on the metabolic context. The interaction of ALDOB with CtBP2 is abolished in the presence of a high level of FBP. A similar finding has been observed for ALDOB interactions with members of the TRPV family and NKCC2 [32,48]. Aldolase binding is modulated by the presence of its products, which indicates complex physiologic regulation in cells in response to different metabolic environments. As shown in our previous study, ccRCC simultaneously downregulates ALDOB and steadily increases FBP concentrations. This study may also suggest that both the depletion of ALDOB and high levels of FBP contribute to the active oncogenic function of CtBP2. Since we previously reported that ALDOB knockdown neither increases cell proliferation nor inhibits cell apoptosis, the loss of function test of ALDOB was not performed in this study. This finding may be attributed to the already low level of ALDOB in ccRCC cells.

Integrated lipidomics-transcriptomics research revealed a reprogramming of fatty acid metabolism in ccRCC [40]. Depletion of ALDOB contributes to lipogenesis in hepatocellular carcinogenesis [49]. Further research is needed to determine whether ALDOB is involved in lipid metabolism in ccRCC.

CtBP1 and CtBP2, which are transcriptional corepressors, inhibit the transcription of a variety of tumor suppressor genes, and both play overlapping and unique roles in tumorigenesis and development [26,50,51]. CtBP1 promotes breast cancer growth by inhibiting SIRT4 transcription [52]. CtBP2 promotes the progression of prostate cancer by inhibiting multiple cosuppressors (NCOR, RIP140, etc.) that inhibit the transcription of AR. Therefore, CtBP might represent a potential therapeutic target for malignant tumors [52]. In the present study, CtBP2 knockdown abolished the effect of ALDOB on the transcription of CDH1 (Fig.3), consistent with the role of CtBP2 as an executor.

Notably, CtBPs are regulated by several metabolites. MTOB, an intermediate in the methionine salvage pathway, is capable of antagonizing CtBP-mediated transcriptional regulation. However, MTOB is normally quickly converted to methionine by mitochondrial tyrosine aminotransferase. The molecular glue ALDOB is very likely to enable the passing of MTOB, an intermediary metabolite of ADI1, directly to CtBP2 without its release into solution. Furthermore, MTOB consumes high-affinity NADH in the enzymatic reaction of CtBPs and generates NAD ⁺ , which has 100 times lower affinity than NADH [53]. Disabling the protein interaction of ALDOB by mutation almost completely abolishes this proximity-induced inhibitory effect of CtBP2.

References

[1] Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin 2020; 70(1): 7–30

[2] Mitchell TJ, Turajlic S, Rowan A, Nicol D, Farmery JHR, O’Brien T, Martincorena I, Tarpey P, Angelopoulos N, Yates LR, Butler AP, Raine K, Stewart GD, Challacombe B, Fernando A, Lopez JI, Hazell S, Chandra A, Chowdhury S, Rudman S, Soultati A, Stamp G, Fotiadis N, Pickering L, Au L, Spain L, Lynch J, Stares M, Teague J, Maura F, Wedge DC, Horswell S, Chambers T, Litchfield K, Xu H, Stewart A, Elaidi R, Oudard S, McGranahan N, Csabai I, Gore M, Futreal PA, Larkin J, Lynch AG, Szallasi Z, Swanton C, Campbell PJ, Consortium TRR. Timing the landmark events in the evolution of clear cell renal cell cancer: TRACERx renal. Cell 2018; 173(3): 611–623.e17

[3] Schödel J, Grampp S, Maher ER, Moch H, Ratcliffe PJ, Russo P, Mole DR. Hypoxia, hypoxia-inducible transcription factors, and renal cancer. Eur Urol 2016; 69(4): 646–657

[4] Lv L, Lei Q. Proteins moonlighting in tumor metabolism and epigenetics. Front Med 2021; 15(3): 383–403

[5] Zhang CS, Hawley SA, Zong Y, Li M, Wang Z, Gray A, Ma T, Cui J, Feng JW, Zhu M, Wu YQ, Li TY, Ye Z, Lin SY, Yin H, Piao HL, Hardie DG, Lin SC. Fructose-1,6-bisphosphate and aldolase mediate glucose sensing by AMPK. Nature 2017; 548(7665): 112–116

[6] Li M, He X, Guo W, Yu H, Zhang S, Wang N, Liu G, Sa R, Shen X, Jiang Y, Tang Y, Zhuo Y, Yin C, Tu Q, Li N, Nie X, Li Y, Hu Z, Zhu H, Ding J, Li Z, Liu T, Zhang F, Zhou H, Li S, Yue J, Yan Z, Cheng S, Tao Y, Yin H. Aldolase B suppresses hepatocellular carcinogenesis by inhibiting G6PD and pentose phosphate pathways. Nat Can 2020; 1(7): 735–747

[7] Huang J, Kong W, Zhang J, Chen Y, Xue W, Liu D, Huang Y. c-Myc modulates glucose metabolism via regulation of miR-184/PKM2 pathway in clear-cell renal cell carcinoma. Int J Oncol 2016; 49(4): 1569–1575

[8] Courtney KD, Bezwada D, Mashimo T, Pichumani K, Vemireddy V, Funk AM, Wimberly J, McNeil SS, Kapur P, Lotan Y, Margulis V, Cadeddu JA, Pedrosa I, DeBerardinis RJ, Malloy CR, Bachoo RM, Maher EA. Isotope tracing of human clear cell renal cell carcinomas demonstrates suppressed glucose oxidation in vivo. Cell Metab 2018; 28(5): 793–800.e2

[9] Wettersten HI, Aboud OA, Lara PN Jr, Weiss RH. Metabolic reprogramming in clear cell renal cell carcinoma. Nat Rev Nephrol 2017; 13(7): 410–419

[10] Wang J, Wu Q, Qiu J. Accumulation of fructose 1,6-bisphosphate protects clear cell renal cell carcinoma from oxidative stress. Lab Invest 2019; 99(6): 898–908

[11] Lu Y, Li Y, Liu Q, Tian N, Du P, Zhu F, Han Y, Liu X, Liu X, Peng X, Wang X, Wu Y, Tong L, Li Y, Zhu Y, Wu L, Zhang P, Xu Y, Chen H, Li B, Tong X. MondoA-thioredoxin-interacting protein axis maintains regulatory T-cell identity and function in colorectal cancer microenvironment. Gastroenterology 2021; 161(2): 575–591.e16

[12] Martinelli P, Carrillo-de Santa Pau E, Cox T, Sainz B Jr, Dusetti N, Greenhalf W, Rinaldi L, Costello E, Ghaneh P, Malats N, Büchler M, Pajic M, Biankin AV, Iovanna J, Neoptolemos J, Real FX. GATA6 regulates EMT and tumour dissemination, and is a marker of response to adjuvant chemotherapy in pancreatic cancer. Gut 2017; 66(9): 1665–1676

[13] Ganaie AA, Beigh FH, Astone M, Ferrari MG, Maqbool R, Umbreen S, Parray AS, Siddique HR, Hussain T, Murugan P, Morrissey C, Koochekpour S, Deng Y, Konety BR, Hoeppner LH, Saleem M. BMI1 drives metastasis of prostate cancer in Caucasian and African-American men and is a potential therapeutic target: hypothesis tested in race-specific models. Clin Cancer Res 2018; 24(24): 6421–6432

[14] Kouwenhoven EN, van Heeringen SJ, Tena JJ, Oti M, Dutilh BE, Alonso ME, de la Calle-Mustienes E, Smeenk L, Rinne T, Parsaulian L, Bolat E, Jurgelenaite R, Huynen MA, Hoischen A, Veltman JA, Brunner HG, Roscioli T, Oates E, Wilson M, Manzanares M, Gómez-Skarmeta JL, Stunnenberg HG, Lohrum M, van Bokhoven H, Zhou H. Genome-wide profiling of p63 DNA-binding sites identifies an element that regulates gene expression during limb development in the 7q21 SHFM1 locus. PLoS Genet 2010; 6(8): e1001065

[15] Lu C, Yang D, Sabbatini ME, Colby AH, Grinstaff MW, Oberlies NH, Pearce C, Liu K. Contrasting roles of H3K4me3 and H3K9me3 in regulation of apoptosis and gemcitabine resistance in human pancreatic cancer cells. BMC Cancer 2018; 18(1): 149

[16] Hu X, Feng C, Zhou Y, Harrison A, Chen M. DeepTrio: a ternary prediction system for protein-protein interaction using mask multiple parallel convolutional neural networks. Bioinformatics 2021; 38(3): 694–702

[17] Sakamoto T, Weng J S, Hara T, Yoshino S, Kozuka-Hata H, Oyama M, Seiki M. Hypoxia-inducible factor 1 regulation through cross talk between mTOR and MT1-MMP. Mol Cell Biol. 2014; 34(1): 30–42

[18] Li X, Eberhardt A, Hansen JN, Bohmann D, Li H, Schor NF. Methylation of the phosphatase-transcription activator EYA1 by protein arginine methyltransferase 1: mechanistic, functional, and structural studies. FASEB J 2017; 31(6): 2327–2339

[19] SantamariaREsposito GVitaglianoLRaceVPaglionico IZancanLZagariASalvatore F. Functional and molecular modelling studies of two hereditary fructose intolerance-causing mutations at arginine 303 in human liver aldolase. Biochem J 2000; 350 Pt 3(Pt 3): 823–828

[20] Zhang Y, Heinsen MH, Kostic M, Pagani GM, Riera TV, Perovic I, Hedstrom L, Snider BB, Pochapsky TC. Analogs of 1-phosphonooxy-2,2-dihydroxy-3-oxo-5-(methylthio)pentane, an acyclic intermediate in the methionine salvage pathway: a new preparation and characterization of activity with E1 enolase/phosphatase from Klebsiella oxytoca. Bioorg Med Chem 2004; 12(14): 3847–3855

[21] Deshpande AR, Wagenpfeil K, Pochapsky TC, Petsko GA, Ringe D. Metal-dependent function of a mammalian acireductone dioxygenase. Biochemistry 2016; 55(9): 1398–1407

[22] Wang J, Tan M, Ge J, Zhang P, Zhong J, Tao L, Wang Q, Tong X, Qiu J. Lysosomal acid lipase promotes cholesterol ester metabolism and drives clear cell renal cell carcinoma progression. Cell Prolif 2018; 51(4): e12452

[23] AlexandrovBSIlievFLStanevV VesselinovVAlexandrov LB. rNMF 1.0: Robust Nonnegative matrix factorization with kmeans clustering and signal shift. 2016

[24] Guinney J, Dienstmann R, Wang X, de Reyniès A, Schlicker A, Soneson C, Marisa L, Roepman P, Nyamundanda G, Angelino P, Bot BM, Morris JS, Simon IM, Gerster S, Fessler E, De Sousa E Melo F, Missiaglia E, Ramay H, Barras D, Homicsko K, Maru D, Manyam GC, Broom B, Boige V, Perez-Villamil B, Laderas T, Salazar R, Gray JW, Hanahan D, Tabernero J, Bernards R, Friend SH, Laurent-Puig P, Medema JP, Sadanandam A, Wessels L, Delorenzi M, Kopetz S, Vermeulen L, Tejpar S, Tejpar S. The consensus molecular subtypes of colorectal cancer. Nat Med 2015; 21(11): 1350–1356

[25] He X, Li M, Yu H, Liu G, Wang N, Yin C, Tu Q, Narla G, Tao Y, Cheng S, Yin H, Yin H. Loss of hepatic aldolase B activates Akt and promotes hepatocellular carcinogenesis by destabilizing the Aldob/Akt/PP2A protein complex. PLoS Biol 2020; 18(12): e3000803

[26] Takayama K, Suzuki T, Fujimura T, Urano T, Takahashi S, Homma Y, Inoue S. CtBP2 modulates the androgen receptor to promote prostate cancer progression. Cancer Res 2014; 74(22): 6542–6553

[27] Wang DP, Gu LL, Xue Q, Chen H, Mao GX. CtBP2 promotes proliferation and reduces drug sensitivity in non-small cell lung cancer via the Wnt/β-catenin pathway. Neoplasma 2018; 65(6): 888–897

[28] Paliwal S, Ho N, Parker D, Grossman SR. CtBP2 promotes human cancer cell migration by transcriptional activation of tiam1. Genes Cancer 2012; 3(7–8): 481–490

[29] Dai F, Xuan Y, Jin JJ, Yu S, Long ZW, Cai H, Liu XW, Zhou Y, Wang YN, Chen Z, Huang H. CtBP2 overexpression promotes tumor cell proliferation and invasion in gastric cancer and is associated with poor prognosis. Oncotarget 2017; 8(17): 28736–28749

[30] Quinlan KGR, Verger A, Kwok A, Lee SHY, Perdomo J, Nardini M, Bolognesi M, Crossley M. Role of the C-terminal binding protein PXDLS motif binding cleft in protein interactions and transcriptional repression. Mol Cell Biol 2006; 26(21): 8202–8213

[31] Riefler GM, Firestein BL. Binding of neuronal nitric-oxide synthase (nNOS) to carboxyl-terminal-binding protein (CtBP) changes the localization of CtBP from the nucleus to the cytosol: a novel function for targeting by the PDZ domain of nNOS. J Biol Chem 2001; 276(51): 48262–48268

[32] Benziane B, Demaretz S, Defontaine N, Zaarour N, Cheval L, Bourgeois S, Klein C, Froissart M, Blanchard A, Paillard M, Gamba G, Houillier P, Laghmani K. NKCC2 surface expression in mammalian cells: down-regulation by novel interaction with aldolase B. J Biol Chem 2007; 282(46): 33817–33830

[33] Zhao LJ, Kuppuswamy M, Vijayalingam S, Chinnadurai G. Interaction of ZEB and histone deacetylase with the PLDLS-binding cleft region of monomeric C-terminal binding protein 2. BMC Mol Biol 2009; 10(1): 89

[34] Lu M, Ammar D, Ives H, Albrecht F, Gluck SL. Physical interaction between aldolase and vacuolar H⁺-ATPase is essential for the assembly and activity of the proton pump. J Biol Chem 2007; 282(34): 24495–24503

[35] Lee S, Hong S, Kim S, Kang S. Ataxin-1 occupies the promoter region of E-cadherin in vivo and activates CtBP2-repressed promoter. Biochim Biophys Acta 2011; 1813(5): 713–722

[36] Wan C, Borgeson B, Phanse S, Tu F, Drew K, Clark G, Xiong X, Kagan O, Kwan J, Bezginov A, Chessman K, Pal S, Cromar G, Papoulas O, Ni Z, Boutz DR, Stoilova S, Havugimana PC, Guo X, Malty RH, Sarov M, Greenblatt J, Babu M, Derry WB, Tillier ER, Wallingford JB, Parkinson J, Marcotte EM, Emili A. Panorama of ancient metazoan macromolecular complexes. Nature 2015; 525(7569): 339–344

[37] Lucarelli G, Loizzo D, Franzin R, Battaglia S, Ferro M, Cantiello F, Castellano G, Bettocchi C, Ditonno P, Battaglia M. Metabolomic insights into pathophysiological mechanisms and biomarker discovery in clear cell renal cell carcinoma. Expert Rev Mol Diagn 2019; 19(5): 397–407

[38] Ragone R, Sallustio F, Piccinonna S, Rutigliano M, Vanessa G, Palazzo S, Lucarelli G, Ditonno P, Battaglia M, Fanizzi FP, Schena FP. Renal cell carcinoma: a study through NMR-based metabolomics combined with transcriptomics. Diseases 2016; 4(1): 7

[39] Bianchi C, Meregalli C, Bombelli S, Di Stefano V, Salerno F, Torsello B, De Marco S, Bovo G, Cifola I, Mangano E, Battaglia C, Strada G, Lucarelli G, Weiss RH, Perego RA. The glucose and lipid metabolism reprogramming is grade-dependent in clear cell renal cell carcinoma primary cultures and is targetable to modulate cell viability and proliferation. Oncotarget 2017; 8(69): 113502–113515

[40] Lucarelli G, Ferro M, Loizzo D, Bianchi C, Terracciano D, Cantiello F, Bell LN, Battaglia S, Porta C, Gernone A, Perego RA, Maiorano E, Cobelli O, Castellano G, Vincenti L, Ditonno P, Battaglia M. Integration of lipidomics and transcriptomics reveals reprogramming of the lipid metabolism and composition in clear cell renal cell carcinoma. Metabolites 2020; 10(12): 509

[41] Bombelli S, Torsello B, De Marco S, Lucarelli G, Cifola I, Grasselli C, Strada G, Bovo G, Perego RA, Bianchi C. 36-kDa annexin A3 isoform negatively modulates lipid storage in clear cell renal cell carcinoma cells. Am J Pathol 2020; 190(11): 2317–2326

[42] Lucarelli G, Rutigliano M, Sallustio F, Ribatti D, Giglio A, Lepore Signorile M, Grossi V, Sanese P, Napoli A, Maiorano E, Bianchi C, Perego RA, Ferro M, Ranieri E, Serino G, Bell LN, Ditonno P, Simone C, Battaglia M. Integrated multi-omics characterization reveals a distinctive metabolic signature and the role of NDUFA4L2 in promoting angiogenesis, chemoresistance, and mitochondrial dysfunction in clear cell renal cell carcinoma. Aging (Albany NY) 2018; 10(12): 3957–3985

[43] Gatto F, Nookaew I, Nielsen J. Chromosome 3p loss of heterozygosity is associated with a unique metabolic network in clear cell renal carcinoma. Proc Natl Acad Sci USA 2014; 111(9): E866–E875

[44] Li B, Qiu B, Lee DSM, Walton ZE, Ochocki JD, Mathew LK, Mancuso A, Gade TPF, Keith B, Nissim I, Simon MC. Fructose-1,6-bisphosphatase opposes renal carcinoma progression. Nature 2014; 513(7517): 251–255

[45] Tao QF, Yuan SX, Yang F, Yang S, Yang Y, Yuan JH, Wang ZG, Xu QG, Lin KY, Cai J, Yu J, Huang WL, Teng XL, Zhou CC, Wang F, Sun SH, Zhou WP. Aldolase B inhibits metastasis through ten-eleven translocation 1 and serves as a prognostic biomarker in hepatocellular carcinoma. Mol Cancer 2015; 14(1): 170

[46] Lian J, Xia L, Chen Y, Zheng J, Ma K, Luo L, Ye F. Aldolase B impairs DNA mismatch repair and induces apoptosis in colon adenocarcinoma. Pathol Res Pract 2019; 215(11): 152597

[47] Bu P, Chen KY, Xiang K, Johnson C, Crown SB, Rakhilin N, Ai Y, Wang L, Xi R, Astapova I, Han Y, Li J, Barth BB, Lu M, Gao Z, Mines R, Zhang L, Herman M, Hsu D, Zhang GF, Shen X. Aldolase B-mediated fructose metabolism drives metabolic reprogramming of colon cancer liver metastasis. Cell Metab 2018; 27(6): 1249–1262.e4

[48] Li M, Zhang CS, Zong Y, Feng JW, Ma T, Hu M, Lin Z, Li X, Xie C, Wu Y, Jiang D, Li Y, Zhang C, Tian X, Wang W, Yang Y, Chen J, Cui J, Wu YQ, Chen X, Liu QF, Wu J, Lin SY, Ye Z, Liu Y, Piao HL, Yu L, Zhou Z, Xie XS, Hardie DG, Lin SC. Transient receptor potential V channels are essential for glucose sensing by aldolase and AMPK. Cell Metab 2019; 30(3): 508–524.e12

[49] Liu G, Wang N, Zhang C, Li M, He X, Yin C, Tu Q, Shen X, Zhang L, Lv J, Wang Y, Jiang H, Chen S, Li N, Tao Y, Yin H. Fructose-1,6-bisphosphate aldolase B depletion promotes hepatocellular carcinogenesis through activating insulin receptor signaling and lipogenesis. Hepatology 2021; 74(6): 3037–3055

[50] Chan CB, Liu X, Jang SW, Hsu SIH, Williams I, Kang S, Chen J, Ye K. NGF inhibits human leukemia proliferation by downregulating cyclin A1 expression through promoting acinus/CtBP2 association. Oncogene 2009; 28(43): 3825–3836

[51] Barroilhet L, Yang J, Hasselblatt K, Paranal RM, Ng SK, Rauh-Hain JA, Welch WR, Bradner JE, Berkowitz RS, Ng SW. C-terminal binding protein-2 regulates response of epithelial ovarian cancer cells to histone deacetylase inhibitors. Oncogene 2013; 32(33): 3896–3903

[52] Wang L, Zhou H, Wang Y, Cui G, Di LJ. CtBP maintains cancer cell growth and metabolic homeostasis via regulating SIRT4. Cell Death Dis 2015; 6(1): e1620

[53] Fjeld CC, Birdsong WT, Goodman RH. Differential binding of NAD⁺ and NADH allows the transcriptional corepressor carboxyl-terminal binding protein to serve as a metabolic sensor. Proc Natl Acad Sci USA 2003; 100(16): 9202–9207

RIGHTS & PERMISSIONS

Higher Education Press