dynamic regulation of me1 phosphorylation and acetylation ......molecular cell article dynamic...

Article

Dynamic Regulation of ME
1 Phosphorylation andAcetylation Affects Lipid Metabolism and ColorectalTumorigenesis
Graphical Abstract

Highlights

d PGAM5 dephosphorylates ME1 at S336 to permit increased

acetylation by ACAT1 at K337

d ME1 S336 phosphorylation and K337 acetylation aremutually

exclusive

d ME1 K337 acetylation drives NADPH generation, lipogenesis,

colorectal tumorigenesis

d PGAM5 and ME1 are dysregulated and activated by

b-catenin/TCF1 in CRC

Zhu et al., 2020, Molecular Cell 77, 1–12January 2, 2020 ª 2019 Elsevier Inc.https://doi.org/10.1016/j.molcel.2019.10.015

Authors

Yahui Zhu, Li Gu, Xi Lin, ...,

Edward V. Prochownik,

Chengpeng Fan, Youjun Li

[email protected]

In Brief

Zhu et al. demonstrate that the b-catenin/

TCF1 complex transcriptionally activates

PGAM5 and ME1 in colorectal cancer.

PGAM5 dephosphorylates ME1 at S336,

thus permitting increased acetylation by

ACAT1 at the adjacent K337. This causes

ME1 dimerization and activation thereby

enhancing NADPH generation,

lipogenesis, and the promotion of

colorectal tumorigenesis.

mailto:[email protected].�cn

https://doi.org/10.1016/j.molcel.2019.10.015

Please cite this article in press as: Zhu et al., Dynamic Regulation of ME1 Phosphorylation and Acetylation Affects Lipid Metabolism and ColorectalTumorigenesis, Molecular Cell (2019), https://doi.org/10.1016/j.molcel.2019.10.015

Molecular Cell

Article

Dynamic Regulation of ME1 Phosphorylationand Acetylation Affects Lipid Metabolismand Colorectal TumorigenesisYahui Zhu,1,2,7 Li Gu,1,2,7 Xi Lin,1,2 Cheng Liu,1,2 Bingjun Lu,1,2 Kaisa Cui,1,2 Feng Zhou,3,4 Qiu Zhao,3,4

Edward V. Prochownik,5 Chengpeng Fan,6 and Youjun Li1,2,8,*1Hubei Key Laboratory of Cell Homeostasis, College of Life Sciences, Wuhan University, Wuhan 430072, China2Medical Research Institute, School of Medicine, Wuhan University, Wuhan 430071, China3Department of Gastroenterology, Zhongnan Hospital of Wuhan University School of Medicine, Wuhan 430071 China4Hubei Clinical Center and Key Laboratory for Intestinal and Colorectal Diseases, Wuhan 430071, China5Division of Hematology/Oncology, Children’s Hospital of Pittsburgh of UPMC, The Department of Microbiology andMolecular Genetics, The

Pittsburgh Liver Research Center and The Hillman Cancer Center of UPMC, The University of Pittsburgh Medical Center, Pittsburgh,PA 15224, USA6Department of Biochemistry and Molecular Biology, School of Basic Medical Sciences, Wuhan University, Wuhan 430071, China7These authors contributed equally8Lead Contact*Correspondence: [email protected]


SUMMARY

PGAM5 is a mitochondrial serine/threonine phos-phatase that regulates multiple metabolic pathwaysand contributes to tumorigenesis in a poorlyunderstood manner. We show here that PGAM5 inhi-bition attenuates lipid metabolism and colorectaltumorigenesis in mice. PGAM5-mediated dephos-phorylation of malic enzyme 1 (ME1) at S336 allowsincreased ACAT1-mediated K337 acetylation, lead-ing to ME1 dimerization and activation, both of whichare reversed by NEK1 kinase-mediated S336 phos-phorylation. SIRT6 deacetylase antagonizes ACAT1function in a manner that involves mutually exclusiveME1 S336 phosphorylation and K337 acetylation.ME1 also promotes nicotinamide adenine dinucleo-tide phosphate (NADPH) production, lipogenesis,and colorectal cancers in which ME1 transcripts areupregulated andME1 protein is hypophosphorylatedat S336 and hyperacetylated at K337. PGAM5 andME1 upregulation occur via direct transcriptionalactivation mediated by b-catenin/TCF1. Thus, thebalance between PGAM5-mediated dephosphoryla-tion ofME1 S336 and ACAT1-mediated acetylation ofK337 strongly influences NADPH generation, lipidmetabolism, and the susceptibility to colorectaltumorigenesis.

INTRODUCTION

Colorectal cancer (CRC) is the third most common malignant

disease and the fourth leading cause of cancer death worldwide

(Torre et al., 2015). Altered metabolism is a common hallmark of

cancer cells in general, and CRC in particular, given that rapid

tumor growth is dependent on changes in aerobic glycolysis,

glutaminolysis, and lipidmetabolism (Cotte et al., 2018; Hanahan

and Weinberg, 2011; Hao et al., 2016). Most tumors display

elevated fatty acid (FA) synthesis and show increases in the

transcripts and enzymes that encode the enzymes in these

pathways (Finicle et al., 2018). Lipogenesis inhibitors are

currently undergoing clinical trials for cancer therapy (Makinosh-

ima et al., 2018; Zadra et al., 2019). However, very little is

known about the regulation of these enzymes by post-transla-

tional modification and how they contribute, if at all, to CRC

induction.

Malic enzymes (ME) participate in the reactions that link

anabolic and catabolic branches of metabolism. The cyto-

plasmic form of ME, ME1, provides one source of nicotinamide

adenine dinucleotide phosphate (NADPH) that is used for FA

synthesis and glutamine production (Dey et al., 2017; Jiang

et al., 2013). ME1 supports tumor growth through its participa-

tion in various tumor-related signaling pathways, particularly

those involving TP53, Myc, and KRAS. For example, in senes-

cent and cancer cells, TP53 inhibits NADPH production and lipo-

genesis by decreasing ME1 expression via direct binding to ME1

promoter (Jiang et al., 2013). Depletion of ME1 inhibits mutant

KRAS and APC mutant-driven CRC growth by decreasing

NADPH production and lipogenesis (Fernandes et al., 2018; Lu

et al., 2018; Shen et al., 2017). NADPH supports redox defenses

and reductive biosynthesis in tumor cells (Wen et al., 2015). ME1

thus stands at an important crossroad that links normal and

neoplastic proliferation. However, little is known concerning

ME1’s post-translational regulation.

Cyclic phosphorylation/dephosphorylation and acetylation/

deacetylation play critical roles in the regulation of many essen-

tial proteins during different cellular processes including meta-

bolism, carcinogenesis, epigenetic modeling, and survival

(Robitaille et al., 2013; Shan et al., 2014; Yang et al., 2012).

Recent studies have revealed that K321 acetylation of the

Molecular Cell 77, 1–12, January 2, 2020 ª 2019 Elsevier Inc. 1

mailto:[email protected]



pyruvate dehydrogenase A1 (PDHA1) subunit inhibits PDH activ-

ity by recruiting the inhibitory PDH kinase 1 (PDK1). Conversely,

the stimulatory PDH phosphatase (PDP1) K202 acetylation in-

hibits PDH function by dissociating its substrate PDHA1, both

of which are critical through promoting glycolysis in tumor

growth (Fan et al., 2014; Shan et al., 2014). The epidermal growth

factor receptor (EGFR) stimulates the activity of acetyl-CoA

acetyltransferase (ACAT1) and stimulates tumor growth by phos-

phorylating the ACAT1 Y407 residue and promoting its tetrame-

rization (Fan et al., 2016). The acetylation and phosphorylation of

metabolic enzymes also may affect their stability, as evidenced

by the P300/CBP-associated factor (PCAF)-mediated acetyla-

tion of cell death-inducing DFFA-like effector C (CIDEC) (Lin

et al., 2013; Qian et al., 2017).

PGAM family member 5 (PGAM5) belongs to the phospho-

glycerate mutase family, and in contrast to other family

members, functions as an atypical serine/threonine protein

phosphatase rather than as a phosphoglycerate mutase.

PGAM5 is involved in regulating apoptotic and necroptotic path-

ways as well as mitochondrial turnover by inducing mitophagy

after mitochondrial damage (He et al., 2017; Panda et al.,

2016; Ramachandran and Jaeschke, 2017; Wang et al., 2012).

PGAM5 is also upregulated in hepatocellular carcinoma and

promotes its growth (Cheng et al., 2018).

Here, we report that PGAM5 is upregulated in patients with

CRC and functions by dephosphorylating ME1 at S336. This

lead to a reciprocal increase in ME1 acetylation by ACAT1 at

the adjacent K337 residue and increases in ME1’s dimerization

and enzymatic activity, both of which are necessary to promote

colorectal tumorigenesis. In this manner, ME1 S336 phosphory-

lation and K337 acetylation are shown to be mutually antago-

nistic. These findings reveal that an important step in the

metabolic reprogramming that accompanies CRC tumorigen-

esis involves the reciprocal PGAM5-mediated dephosphoryla-

tion and ACAT1-mediated acetylation of adjacent residues on

ME1 that dictate its inherent metabolic activity.

RESULTS

PGAM5 Is Upregulated in CRCNumerous kinase inhibitors have been used to treat a variety of

cancers including CRC (Qi et al., 2018). However, the selective

targeting of protein phosphatases remains a relatively underde-

veloped area. To study the role of phosphatases in colorectal

tumorigenesis and to identify potential therapeutic targets, we

first examined the expression of transcripts encoding 324 known

phosphatases in 382CRCs from the TCGA database. The results

indicated that 54 phosphatase genes were commonly upregu-

lated in CRCs (1.6-fold threshold) (Figure 1A). We next analyzed

the protein expression of these genes in CRCs from the Human

Protein Atlas database and found that, in 10 of 12 tumor sam-

ples, only PPM1G and PGAM5 were overexpressed (Figure 1A).

To demonstrate a functional role for these two phosphatases in

CRC, we knocked down their expression in CRC cells and found

that depleting PGAM5 significantly inhibited CRC cell growth,

whereas depleting PPM1G had no effect (Figures 1B and S1A).

Additional studies revealed that CRCs contained the highest

levels of PGAM5 RNA and protein among evaluable tumors

2 Molecular Cell 77, 1–12, January 2, 2020

(Figures S1B and S1C). A fuller analysis among 33 different tu-

mor types from the TCGA and GTEx datasets showed that

PGAM5 transcripts were also upregulated in most other types

of cancer (Figure S1D). PGAM5 RNA levels were also upregu-

lated in virtually all tumors from 32 matched normal colorectal

and CRC tissues from the TCGA database (Figure S1E).

PGAM5 RNA was also highly upregulated in 31 pairs of matched

normal:CRC tissues fromWuhan Union Hospital (Figure 1C) and

in CRCs from several GSE databases (Figure S1F). Consistently,

PGAM5 RNA and protein levels gradually increased in the colon

of C57BL/6 mice after azoxymethane/dextran sulfate sodium

(AOM/DSS) treatment (Figures 1D and 1E). Finally, PGAM5

was also significantly upregulated in the colon of ApcMin/+ mice

(Figures 1F and 1G).

We next investigated the effects of PGAM5 modulation on

CRC growth. ShRNA-mediated inhibition of PGAM5 significantly

retarded in vivo xenograft growth, and this defect could be

rescued by the expression of wild-type (WT) PGAM5 but not

by a phosphatase-dead H105A mutant (Figures S1G–S1K).

These results indicate that PGAM5 is highly expressed in CRC

and necessary to sustain optimal levels of tumor growth.

PGAM5DephosphorylatesME1 to EnhanceME1ActivityTo better study how PGAM5 contributes to CRC tumorigenesis,

we overexpressed PGAM5 in HEK293 cells and then performed

coimmunoprecipitation (coIP) studies followed by SDS-PAGE in

order to identify key PGAM5-associated proteins. ME1 was

identified as one of the most prominent interactors (Figure 2A)

and in reciprocal experiments, PGAM5 was identified as a major

ME1 interactor (Figure S2A). The endogenous interactions be-

tween PGAM5 and ME1 were confirmed by coIP experiments

in CRC cells (Figures 2B and S2B). Collectively, these results

show that PGAM5 interacts with ME1 in CRC cells.

Because PGAM5 is a phosphatase, we examined the effect of

PGAM5 on ME1 phosphorylation in HCT116 cells. PGAM5 inhi-

bition significantly increased ME1 serine phosphorylation, and

this defect could be rescued by expressing WT PGAM5 but

not the phosphatase-dead H105A mutant (Figure 2C). Finally,

PGAM5 promoted ME1 dimerization (Figures S2C and S2D).

ME1 is an important source of NADPH and plays an important

role in lipogenesis (Jiang et al., 2013). Therefore, we quantified

ME1 activity and lipid metabolism in stable cells and CRC xeno-

grafts. The results showed that PGAM5 depletion inhibited ME1

activity, NADPH production, and lipogenesis but had no effect

on the activity of the mitochondrial ME isoform, ME2 (Figures

2D, 2E, and S2E–S2K). Depletion of PGAM5 also decreased

glycolysis and the extracellular acidification rate (ECAR), which

reflects overall glycolytic flux. Further, it also augmented the ox-

ygen consumption rate (OCR), an indirect indicator of mitochon-

drial oxidative respiration (Figures 2F, S2L, and S2M). Together,

these data demonstrate that PGAM5 directly interacts with ME1,

promotes its dimerization, and enhances its ability to generate

NADPH and promote lipid synthesis.

NEK1 Phosphorylates ME1 at Serine 336 to Inhibit ME1ActivityTo identify the kinase(s) responsible for ME1 phosphorylation,

we analyzed the results of mass spectrometry derived from

Figure 1. PGAM5 Is Upregulated in CRC

(A) Summary of bioinformatic screening results in CRCs. Among the 324 phosphatase genes examined, 53 were upregulated (tumor/normal >1.6-fold) in CRCs

based on transcript levels in the TCGA database. Only PPM1G and PGAM5 protein were highly expressed in 10 of 12 CRC tumors from in The Human Protein

Atlas datasets.

(B) MTT assays in CRC cells after transfection with PPM1G and PGAM5 shRNAs.

(C) PGAM5 expression in 31 pairs of CRCs from Wuhan Union Hospital.

(D and E) Relative mRNA (D) and protein (E) levels of PGAM5 during CRC induction with AOM/DSS.

(F and G) Relative mRNA (F) and protein levels (G) of PGAM5 in CRCs from ApcMin/+ mice compared with WT mice.

Data were presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.


the above-described ME1 IPs and identified 13 peptides

derived from NIMA1-related kinase (NEK1) (Figure S2N, left).

Interestingly, NEK1 and PGAM5 transcript levels were

negatively correlated (Figure S2N, right). CoIP experiments

confirmed an association between endogenous ME1 and

NEK1 (Figure S2O).

NEK1 depletion drastically decreased ME1 serine phosphory-

lation. NEK1 depletion could be rescued by the enforced expres-

sion of WT NEK1 but not by the T162A mutant (Figure 2G).

To identify the actual serine residue(s) of ME1 phosphorylated

by NEK1, we used PhosphoSite Plus to predict candidate phos-

phorylation sites and mutated these to alanines. We found that

NEK1 was unable to phosphorylate the ME1 S336A mutant

whereas other mutants were phosphorylated at normal levels

(Figure 2H). Consistent with the idea that S336 is the site

of NEK1-mediated phosphorylation and PGAM5-mediated

dephosphorylation, the co-expression of PGAM5 did not affect

the phosphorylated status of the ME1 S336A mutant (Fig-

ure S2P). The S336 of ME1 was also found to be highly

conserved among different species (Figure S2Q).

To examine the cross-talk between PGAM5 and NEK1 on the

serine phosphorylation of ME1, we co-expressed or co-depleted

NEK1 and PGAM5 in CRC cells. The studies showed that

PGAM5 decreased NEK1-mediated phosphorylation of ME1

whereas NEK1 increased phosphorylation (Figures 2I and

S2R). Collectively, these observations demonstrate that ME1

S336 is phosphorylated and dephosphorylated by NEK1 and

PGAM5, respectively, and they compete for modification of

this site.

ME1 S336 Phosphorylation Abolishes Its ActivityTo determine the functional consequences of ME1 S336 phos-

phorylation, we co-expressedWTME1 or S336A andS336Dmu-

tants with HA- and Flag-tags. CoIP experiments demonstrated

that the ME1 S336A mutant was able to efficiently dimerize,

whereas S336D was not (Figure S2S). S336A also strongly pro-

moted tumor growth in vivo and increased glycolysis, NADPH

production, and lipid synthesis in a CRC xenograft, whereas

these activities were significantly attenuated in cells expressing

S336D (Figures 2J and S2T–S2Z). Together, these data

Molecular Cell 77, 1–12, January 2, 2020 3

Figure 2. ME1 Is Phosphorylated and Dephosphorylated by NEK1 and PGAM5 at S336, Respectively

(A) Scatterplots displaying comparisons in peptide identities from coIP of control and PGAM5 expressing cells. Cellular extracts from HEK293 cells expressing

Flag-PGAM5 or control vector were immuno-purified on anti-Flag affinity columns. The eluates were resolved by SDS-PAGE and Coomassie blue staining. The

bands were retrieved and analyzed by mass spectrometry.

(B) Endogenous interaction between PGAM5 and ME1.

(C) Depletion of PGAM5 increases serine phosphorylation of ME1. This is rescued by co-expressing WT PGAM5 but not the H105A mutant.

(D) ME1 activity in xenograft from indicated cells.

(E) NADPH levels in xenograft from indicated cells.

(F) Extracellular acidification rate (ECAR) (top) and oxygen consumption rate (OCR) (bottom) analyses of HTC116 cells from (C). The time at which a given

compound was added is indicated.

(G) Depletion of NEK1 reduces serine phosphorylation of ME1. This was rescued with WT NEK1 but not with the inactive T162A mutant.

(H) NEK1 phosphorylates ME1 at S336.

(I) PGAM5 antagonizes NEK1-mediated serine phosphorylation of ME1.

(J) ECAR (top) and OCR (bottom) analyses of HCT116 cells expressing the phosphorylation site mutant of ME1.

Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.


demonstrate that ME1 S336 phosphorylation strongly attenu-

ated multiple ME1-mediated activities.

ACAT1 Acetylates ME1 at K337During the course of our studies, we observed that depleting

PGAM5 expression decreased ME1 acetylation while ectopic

PGAM5 had the opposite effect (Figure 2C). ME1 acetylation

was also increased by depleting NEK1 (Figure 2G). A review

of our above-described mass spectrometry analysis of ME1

coIPs identified acetyl CoA transferase (ACAT1) as the most

abundant acetyltransferase and thus the most likely ME1 ace-

tylase (Figure S3A). We confirmed both the endogenous

interactions between ACAT1 and ME1 (Figures 3A and S3B).

Furthermore, depleting ACAT1 expression in CRC cells

dramatically decreased ME1 acetylation without affecting its

protein level whereas ectopic ACAT1 had the opposite effect

(Figure 3B). To determine the impact of ACAT1 on ME1 dimer-

ization, coIP experiments were performed in HEK293 cells

stably expressing or depleted for ACAT1. We observed

that ectopic ACAT1 expression promoted ME1 dimerization


(Figure S3C) whereas ACAT1 depletion prevented this

(Figure S3D).

Five putative acetylation sites of ME1 were predicted in the

Protein Lysine Modifications Database. We mutated each indi-

vidual lysine (K) to an arginine (R) and found that K337R was

associated with a significant reduction of ME1 acetylation, indi-

cating that ME1 was likely acetylated at this site (Figure 3C).

Ectopic ACAT1 expression proved capable of acetylating all

ME1 mutants except K337R (Figure 3C). These data demon-

strate that ACAT1 acetylates ME1 at K337, which is also a highly

evolutionarily conserved residue (Figure 3D). ME1 K337R

showed a marked loss of dimerization potential whereas the

acetylation mimetic mutation K337Q demonstrated enhanced

dimerization in HCT116 ME1 knockout cells (Figure 3E). Collec-

tively, these results suggest that ACAT1-mediated K337 acetyla-

tion positively regulates ME1 dimerization.

ME1 Is Deacetylated by SIRT6 at K337To identify candidate ME1 deacetylase(s), we used nicotinamide

(NAM) and Trichostatin A (TSA) to selectively inhibit members of

Figure 3. ME1 Is Acetylated and Deacetylated at K337 by ACAT1 and SIRT6, Respectively

(A) Endogenous interactions between ACAT1 and ME1.

(B) ACAT1 inhibition decreases ME1 acetylation.

(C) ACAT1 acetylates ME1 at K337.

(D) Alignment of K337 and adjacent amino acid sequence of ME1 among different species.

(E) ME1 acetylation at K337 enhances its dimerization.

(F) ME1 acetylation is increased by treatment with NAM but not TSA in HCT116.

(G) Interaction between endogenous SIRT6 and ME1.

(H) Depletion of SIRT6 enhances ME1 acetylation in HCT116 cells. Acetylation is reduced by the co-expression of WT SIRT6 but not by the H133Y mutant.

(I) SIRT6 antagonizes ACAT1-mediated ME1 acetylation in HCT116 cells.


the SIRT and HDAC families, respectively. ME1 acetylation

dramatically increased in cells treated with NAM, but not TSA

(Figure 3F), indicating that a SIRT family member was likely

responsible for ME1 deacetylation. To confirm this finding, we

co-expressed ME1 and individual members of the SIRT family

in HEK293 cells and found that ME1 acetylation was reduced

only by SIRT6 (Figure S3E). Endogenous interactions between

ME1 and SIRT6 were documented by coIP experiments (Figures

3G and S3F). SIRT6 depletion increased ME1 acetylation (Fig-

ure 3H). Ectopic expression of SIRT6 decreased acetylation of

WT ME1 and other three mutants, but not K337R mutant (Fig-

ure S3G). Taken together, these data indicate that SIRT6 deace-

tylates ME1 at K337.

To determine the relationship between ACAT1 and SIRT6

on ME1 acetylation, we co-expressed or co-depleted ACAT1

and SIRT6 in CRC cells. The results showed that SIRT6

decreased ACAT1-mediated ME1 acetylation (Figures 3I

and S3H).

ME1 S336 Phosphorylation and K337 Acetylation AreCompetitively AntagonisticTo better understand the relationship between PGAM5-medi-

ated S336 dephosphorylation and ACAT1-mediated K337

acetylation of ME1 in ME1 KO CRC cells, we expressed WT

ME1, S336A, or S336D. The acetylation of ME1 S336A was

dramatically enhanced whereas the acetylation of S336D

was reduced (Figure 4A). Conversely, ectopic expression of

the ME1 K337R acetylation mutant dramatically enhanced

its phosphorylation, whereas ectopic expression of ME1

K337Q, acetylation mimic mutant, had the opposite effect

(Figure 4B).

We next determined the role of ACAT1 on the acetylation of

ME1 S336A and S336D and the role of NEK1 on phosphorylation

of ME1 K337R and K337Q. We found that ACAT1 dramatically

enhanced the acetylation of the ME1 S336A, but not S336D (Fig-

ure 4C). Similarly, NEK1 co-expression increased the phosphor-

ylation of ME1 K337R but not K337Q (Figure 4D).

We next determined whether ME1’s phosphorylation status

affected its interaction with ACAT1 and found that ME1 phos-

phorylation decreased the interaction between ACAT1 and

ME1 (Figure 4E). Similarly, ME1 acetylation decreased the inter-

action between NEK1 and ME1 (Figure 4F). Furthermore, we

compared the subcellular localization of ME1 S336A versus

S336D and K337R versus K337Q. Subcellular fractionation

studies showed ME1 S336D and K337R to be mainly localized

to mitochondria, while S336A and K337Q localized mainly to

the cytosol (Figure S4A). These data demonstrate that the

mechanism by which NEK1 and ACAT1 mutually antagonize

one another’s activities involves the apparent physical

displacement of one protein by the other at their individual

substrate recognition residues, namely S336 and K337,

respectively.


Figure 4. ME1 Acetylation and Phosphorylation Are Mutually Antagonistic

(A) Ectopic expression of ME1 S336A and S336D mutants increase and decrease its acetylation, respectively.

(B) Ectopic expression of ME1 S337R and S337Q mutant increase and decrease its serine phosphorylation, respectively.

(C) Ectopic ACAT1 expression promotes the acetylation of the ME1 S336A mutant but not the S336D mutant.

(D) Ectopic NEK1 expression increases the phosphorylation of the ME1 K337R mutant but not the K337Q mutant.

(E) ME1 mutants S336D and S336A, respectively inhibit and enhance the interaction between ACAT1 and ME1.

(F) ME1 K337R and K337Q mutants, respectively promote and inhibit the interaction with NEK1.

(G) ECAR (left) and OCR (right) analyses of ME1 acetylation site mutants in ME1-KO HCT116 cells.

(H) ME1 activity in lysates from indicated xenografts.

(I) Typical images of colon tumors expressing the indicated ME1 proteins 120 days after AOM/DSS treatment.

(J) Colon tumor numbers (left) and average tumor volume (right) in mice from (I).

(K) ME1 activity in CRCs of mice from (I).

Data were presented as mean ± SD. **p < 0.01, ***p < 0.001.

(L) Heatmap displays colon lipid species as detected via mass spectrometry. Scale bar represents fold change versus Ctrl. FFA, free fatty acids; PA, phos-

pholipids alcohol; PC, phosphatidylcholine; PG, phosphatidyl glycerin; SM, sphingomyelin; Cer, ceramide; PS, phosphatidylserine; PI, phosphatidylinositol;

CerP, ceramide phosphate; LPC, lyso-phosphatidylcholine; CE, cholesterol ester.

(J and K) Significance was performed using Wilcoxon signed rank test. The horizontal lines in the boxplots represent the median, the boxes represent the in-

terquartile range, and the whiskers represent the minimal and maximal values.


ME1 K337 Acetylation Promotes CRC GrowthNext, we explored the biological consequences of ME1 K337

acetylation. Ectopic expression of ME1 acetylation mimicked

K337Q but not K337R, significantly increased CRC in vivo

growth, glycolysis, ME1 enzymatic activity, NADPH production,

and lipid synthesis (Figures 4G, 4H, and S4B–S4H). These results

demonstrate that ME1 K337 acetylation promotes CRC growth

and metabolic re-programming.

ME1 S336 Dephosphorylation and K337 AcetylationPromotes AOM/DSS-Induced Colorectal Tumorigenesisin MiceThe AOM/DSS-induced CRC mouse model was used to eval-

uate the effects of post-translational modification of ME1 on

colorectal tumorigenesis. C57/B6 mice were injected intra-peri-


toneally with AOM, followed by injections of DSS and adenoviral

vectors encoding WT or mutant forms of ME1. Mice were sacri-

ficed 120 days after AOM injection and ME1 expression was

confirmed by western blot (Figures S4I and S4J). Compared to

mice injected with a control adenoviral vector, those expressing

WT ME1 developed more and larger CRCs (Figures 4I, 4J, and

S4K) and showed high levels of NADPHproduction and lipid syn-

thesis (Figures 4K, 4L, and S4L–S4T). In contrast, animals ex-

pressing the ME1 S336D and K337R mutants had similar tumor

burdens and all previously identified activities (Figures 4I–4L and

S4K–S4T). Unexpectedly, mice expressing the S336A and

K337Q mutants developed more numerous and larger CRCs

(Figures 4I, 4J, and S4K) and had higher levels of the above-

described activities (Figures 4K, 4L, and S4L–S4T). These

results reveal the critical and reciprocal roles of ME1 S336

Figure 5. Depletion of PGAM5 Attenuates Chemically Induced CRC Tumorigenesis in Mice

(A) Typical images of colon tumors from adeno-shCtrl- and adeno-shPgam5-treated mice 120 days after AOM/DSS treatment.

(B) Colon tumor numbers (left) and average tumor volumes (right) in mice from (A).

(C) ME1 activities in CRCs of mice from (A).

(D and E) NADPH (D), FA (E, left), and TG (E, right) levels in CRCs of mice from (A).

(F) Heatmap showing tumor lipid species from adeno-shPgam5-treated mice 120 days after AOM/DSS treatment as detected via mass spectrometry. Scale bar

represents fold change versus shCtrl.

Data were presented as mean ± SD.


dephosphorylation and K337 acetylation in CRC development

and lipid metabolism in mice.

Depletion of PGAM5 Attenuated Spontaneous andChemically Induced Colorectal Tumorigenesis in MiceTo test the therapeutic potential of manipulating PGAM5 in vivo,

we asked whether abrogation of its expression would inhibit tu-

mor growth in the CRC mouse model. Concentrated adenovirus

encoding shPgam5 was delivered intraperitoneally into AOM/

DSS-treated WT mice and tumor burdens were assessed on

day 120. Depletion of PGAM5 was confirmed by immunoblot

(Figure S5A). Compared with mice treated with a control adeno-

viruses (shCtrl), these animals showed a significantly reduced

tumor burden throughout the colon and decreased tumor prolif-

eration (Figures 5A, 5B, S5B, and S5C). Abrogation of PGAM5

also significantly reduced all previously identified activities

(Figures 5C–5F and S5D–S5K).

Adenomatous polyposis coli (APC) is the most frequently

mutated gene in sporadic CRC cancer in humans and is nearly

always mutated in familial adenomatous polyposis coli (FAP).

ApcMin/+ mice contain a germline Apc mutation, which predis-

poses to spontaneous colonic and small intestinal polyposis

and serves as a model for human FAP (Peuker et al., 2016).

qRT-PCR and western blot experiments showed that PGAM5

expression was significantly increased in ApcMin/+ mice (Figures

1F and 1G). To investigate the role of PGAM5 in spontaneous

tumor development, we treated the ApcMin/+ mice with

shPgam5-encoding adenovirus. This dramatically suppressed

spontaneous colorectal and small intestine tumorigenesis (Fig-

ures S5L–S5O) and reduced ME1-related activities (Figures


Figure 6. PGAM5 and ME1 Are Transcriptionally Activated by b-Catenin/TCF1

(A) Schematic diagram showing the location of LEF1/TCF1 binding sites in thePGAM5 promoter from upstream�10 kb to downstream +10 kb. 1–3 are amplicons

of CHIP-PCR. TSS, transcription start site.

(B) Dose-dependent inhibition of PGAM5 and ME1 mRNA levels by the LEF1/TCF1 inhibitor iCRT14 in HCT116 and SW480 cells.

(C) Dose-dependent inhibition of PGAM5 and ME1 protein by iCRT14 in HCT116 and SW480 cells.

(D and E) Relative mRNA (D) and protein levels (E) in HCT116 and SW480 CRC cells stably transfected with the indicated combinations of vectors.

(F) qPCR ChIP analyses of b-catenin binding to PGAM5 and ME1 promoter regions in HCT116 cells.

(G) Luciferase activity of different vectors in PGAM5 and ME1 promoter regions with b-catenin inhibition in HCT116 cells.

Data were presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.


S5P–S5T). Depletion of Pgam5 thus significantly inhibits tumor

growth and lipid metabolism in both spontaneous and AOM/

DSS-induced colorectal tumorigenesis in mice.

PGAM5 and ME1 Are Transcriptionally Activated byb-Catenin/TCF1 ComplexThe previous bioinformatics results showed PGAM5 andME1 to

be dramatically upregulated in CRC (Figures S1D–S1F and S6A)

but without evidence for gene amplification (Figures S6B and

S6C). Because mutational activation of the APC/WNT/b-catenin

signaling is pivotal during CRC development (Peuker et al.,

2016), we tested whether WNT/b-catenin signaling was respon-

sible for PGAM5 and ME1 overexpression in CRC. To this end,

we analyzed the putative transcription factor binding sites within

the PGAM5 and ME1 promoter regions (�10 to +10 kb) in the

Motifmap website. TCF/LEF emerged as the top candidate

and its candidate binding sites were largely conserved in

PGAM5 and ME1 promoters (Figures 6A and S6D). Supporting

the functional relevance of this, both PGAM5 and ME1 RNA

and protein were highly downregulated following treatment of


CRC cells with iCRT14, a potent inhibitor of the b-catenin-

TCF/LEF interaction (Figures 6B and 6C). Depletion of b-catenin

had the similar effect (Figures 6D and 6E). Meanwhile, the pro-

tein levels of PGAM5, ME1, and acetylation of ME1 were

dramatically increased, and NEK1 and phosphorylation of

ME1 were decreased in NCM460 cells stimulated with Wnt3a

(Figure S6E).

We next tested whether b-catenin directly transactivates

PGAM5 and ME1, using chromatin immunoprecipitation

(ChIP) assays. The results showed that b-catenin occupied

the promoters of both PGAM5 and ME1 in HCT116 cells (Fig-

ures 6F and S6F). To confirm this, luciferase reporters contain-

ing PGAM5 and ME1 promoter elements were shown to be

responsive to ectopic b-catenin expression. Moreover, short

hairpin RNA (shRNA)-mediated b-catenin suppression dramat-

ically decreased luciferase activity driven by WT promoters,

but not by those with mutant b-catenin binding sites (Figures

6G and S6G). Thus, b-catenin activation facilitates tumor

growth via the direct upregulation of PGAM5 and ME1 in

CRC patients.

Figure 7. ME1 Acetylation and Phosphorylation are Up-regulated and Down-regulated Respectively in Human CRCs

(A) Correlation of PGAM5 (A, left) and ME1 (A, right) protein levels with b-catenin protein levels from CRC samples.

(B–D) Relative protein levels of PGAM5 (B, left), ACAT1 (B, right), P-Ser ME1 (C, left), Ac-ME1 (C, right), ME1 (D). The results shown are from the same tissues

shown in Figure S7A.

(E) Correlation of Ac-ME1 protein levels with P-Ser ME1 protein levels from CRC samples.

(F) Correlation of P-Ser ME1 (F, left) and Ac-ME1 (F, right) protein levels with PGAM5 protein levels from CRC samples.

(G) Correlation of P-Ser ME1 (G, left) and Ac-ME1 (G, right) protein levels with NEK1 protein levels from CRC samples.

(H) Correlation of Ac-ME1 (H, left) and P-Ser ME1 (H, right) protein levels with ACAT1 protein levels from CRC samples.

(I) Correlation of Ac-ME1 (I, left) and P-Ser ME1 (I, right) protein levels with SIRT6 protein levels from CRC samples.

(A, E–I) Each circle is an individual sample. R, Spearman correlation coefficient.

(J) Schematic summary. b-catenin/TCF1 transcriptionally upregulates PGAM5 and ME1. PGAM5-mediated ME1 dephosphorylation and ACAT1-mediated ME1

acetylation increase ME1 activity to promote lipid metabolism and CRC development. Conversely, NEK1-mediated ME1 phosphorylation and SIRT6-mediated

ME1 deacetylation repress these processes.


ME1 S336 Phosphorylation and K337 Acetylation AreDownregulated and Upregulated in Human CRC,RespectivelyTo investigate the clinical relevance of b-catenin, PGAM5, NEK1,

ACAT1, SIRT6, phosphorylation, and acetylation of ME1, we

collected 22 CRC samples (T) with adjacent normal colon tissues

(N) and performed quantitative immunoblotting (Figure S7A). Our

findings showed that b-catenin levels positively correlated with

PGAM5 and ME1 in the above samples (Figure 7A). Both

PGAM5 and ACAT1 proteins were also significantly increased

in these CRC samples (Figure 7B), whereas NEK1 and SIRT6

were decreased (Figure S7B). ME1 phosphorylation was

deceasedwhereas its acetylation and total levels were increased

(Figures 7C and 7D). Overall, ME1 phosphorylation and acetyla-

tion were negatively correlated (Figure 7E), whereas PGAM5was

negatively correlated with phosphorylation and positively corre-

lated with acetylation of ME1 (Figure 7F). NEK1 positively corre-

lated withME1 phosphorylation but negatively correlated with its

acetylation and PGAM5 expression (Figures 7G and S7C).

ACAT1 was positively correlated withME1 acetylation and nega-

tively correlated with phosphorylation ofME1 (Figure 7H). Finally,

SIRT6 was negatively correlated with acetylation of ME1

and ACAT1, but positively correlated with ME1 phosphorylation

(Figures 7I and S7D). Collectively, these data are consistent with

the proteomic and biochemical findings discussed above and

support the idea that post-translational modification of ME1

plays a crucial role in human CRC development.

DISCUSSION

The regulation of signaling pathways by protein kinases and

phosphatases is necessary to achieve well-coordinated and

limited cellular outcomes in response to opposing environmental

cues (Robitaille et al., 2013; Yang et al., 2012). In many cases, ki-

nase/phosphatase complex signaling is both temporally and

spatially dynamic in order to permit fine control over substrate

phosphorylation status (Ben-Sahra et al., 2013; Fan et al.,

2014; Gao et al., 2019). Such complexes may be necessary for



cancer development and can therefore be considered as poten-

tial therapeutic targets. Protein kinase inhibitors have long been

used to treat a variety of tumors, including CRC, and in some

cases have achieved remarkable therapeutic outcomes and sur-

vival benefits (Liao et al., 2017; Mason et al., 2017). In contrast,

the clinical application of phosphatase inhibitors has been signif-

icantly more modest, thus presenting a significant therapeutic

opportunity.

The current study reveals a molecular mechanism underlying

ME1 regulation in highly proliferative CRC cells and suggests

that targeting ME1’s post-translational modifications represents

a promising therapeutic strategy. Our findings point to the exis-

tence of a dynamic equilibrium between ME1 phosphorylation

and acetylation, representing inactive and active forms, respec-

tively (Figure 7J). In normal proliferating cells, such as those

stimulated by WNT, ME1 is dephosphorylated by PGAM5 and

acetylated by ACAT1, leading to active ME1 dimerization. More-

over, this mechanism may be hijacked in human cancer cells,

where oncogenic WNT/b-catenin signaling is commonly upregu-

lated. These results suggest a molecular mechanism by which

metabolic enzymes are regulated via competing phosphoryla-

tion/acetylation reactions in response to normal and oncogenic

signaling.

PGAM5 normally resides on the mitochondrial outer mem-

brane and has been implicated in a variety of cellular activities

related to the control of signal transduction pathways (He

et al., 2017; Wang et al., 2012). PGAM5 has also been identified

in the cytoplasm where it participates in the activation of WNT/

b-catenin signaling (Bernkopf et al., 2018; Rauschenberger

et al., 2017). This study indicates that PGAM5 dramatically

enhances lipid metabolism and CRC growth through its direct

interaction with and dephosphorylation of ME1 (Figure 7J).

These results suggest that the targeting of PGAM5 phosphatase

activity could be of therapeutic benefit in CRC and other cancers

with high-level PGAM5 activity (Figure S1D).

The mutual acetylation and deacetylation of metabolic en-

zymes also play important roles in tumorigenesis (Fan et al.,

2016; Lin et al., 2013), but their precise function in CRC devel-

opment is largely unknown. In this study, we identified ACAT1

and SIRT6 as the ME1 acetylase and deacetylase, respec-

tively. Recently ACAT1 has been reported to enhance tumor

growth by increasing the acetylation of pyruvate dehydroge-

nase phosphatase 1 (PDP1) in a manner that is dependent

upon Y381 phosphorylation status (Fan et al., 2014). The

phosphorylation of ACAT1 itself on Y407 dramatically in-

creases tumor growth by regulating its activity and tetrameri-

zation (Fan et al., 2016). SIRT6 plays a key role in regulating

lipid metabolism and tumorigenesis. For example, SIRT6 pro-

motes DNA repair and its loss leads to abnormalities in mice

that overlap with aging-associated degenerative processes

(Ghosh et al., 2018; Kugel et al., 2016). Meanwhile, SIRT6 de-

acetylates and significantly increases the activity of nicotin-

amide phosphoribosyltransferase (NAMPT). NAMPT is the

rate-limiting enzyme of the NAD+ salvage pathway and also

plays important roles in aging (Sociali et al., 2019). Together

with these findings, our results suggest that ACAT1 and

SIRT6 regulate lipid metabolism and CRC tumorigenesis by

dynamically regulating ME1 acetylation (Figure 7M). Therefore,


ACAT1 inhibitors and/or SIRT6 agonists also merit consider-

ation for the treatment of CRC.

Reprogramming metabolism is a major hallmark of cancer

(Hanahan and Weinberg, 2011). Altered lipid synthesis in partic-

ular plays an extremely critical role in tumor growth. In rapidly

proliferating tumor cells, for example, a high level of lipid synthe-

sis provides precursors for membrane biogenesis and steroid

hormones. Many studies have shown that lipid uptake and

de novo lipid synthesis are markedly increased in tumor cells

(Cheng et al., 2017; Cotte et al., 2018). ME1 is an NADP+-depen-

dent enzyme that generates NADPH for fatty acid biosynthesis

via the reversible oxidative decarboxylation of malate and the

production of pyruvate, thereby linking glycolysis and the TCA

cycle. ME1 upregulation has been observed in a wide spectrum

of cancers, including gastric cancer, liver cancer, and CRC (Fer-

nandes et al., 2018; Lu et al., 2018; Wen et al., 2015). In this

study, we have demonstrated two inter-dependent post-transla-

tional modifications for ME1 by which S336 phosphorylation in-

hibits enzymatic activity by disrupting its ACAT1-dependent

acetylation on the adjacent K337 residue and subsequent ME1

dimerization. In a reciprocal manner, K337 acetylation increases

ME1 activity by blocking S336 phosphorylation and promoting

dimerization (Figure 7M). Both S336 and K337 reside on the

surface of ME1 (https://www.rcsb.org/3d-view/1PJ4/1), thus

potentially allowing ready accessibility to each of the four post-

translational modifying enzymes. ME1 interacts with PGAM5

and ACAT1 in mitochondria but not the cytosol. This indicates

that ME1 shuttles between mitochondria and cytosol and might

be regulated by different post-translational modifications in

these compartments. Furthermore, the binding of phosphory-

lated ME1 to PGAM5 but not NEK1, and the binding of

acetylated ME1 to SIRT6 but not ACAT1, suggests that phos-

phorylated but unacetylated ME1 is located to mitochondria

whereas acetylated but unphosphorylated ME1 is located to

the cytosol. Consistent with this, the results of our subcellular

localization studies showedME1 S336D and K337R to bemainly

localized to mitochondria whereas S336A and K337Q were

mainly cytosolic. Whether the binary and mutually exclusive

choice between phosphorylation and acetylation of these two

residues is based on steric hindrance, electrostatic repulsion,

or some combination of the two remains to be determined as

does whether either of the modifications of these residues leads

to structural alterations that preclude alteration by the other

modifying enzymes. In pigeons, ME1 S346, which corresponds

to human ME1 S336, is reported to be necessary for binding

NADP+ (Chang and Tong, 2003). Once ME1 S336 is phosphory-

lated, the electronegativity of S336 is greatly enhanced, and the

active center of acetylase is unable to bind K337 in the optimum

conformation and orientation. As a result, the original hydrogen

bond between ME1 and NADPH is interrupted, resulting in a

greatly weakened bond between ME1 and NADP+. As K337

acetylation increases, the hydrophobicity of ME1 weakening

the binding of the kinase to adjacent S336 is enhanced as the

active site of its kinase could not bind the substrate in the opti-

mum conformation and orientation. As the hydrophobicity of

the ME1 protein increases once its acetylation, the possibility

of dimer turning over into tetramer may increase. Such questions

can likely be resolved by further structural studies.

https://www.rcsb.org/3d-view/1PJ4/1


Reciprocal S336 phosphorylation and K337 acetylation regu-

late NAPDH production and lipid synthesis largely by affecting

ME1 activity. These findings indicated that post-translational

modifications of ME1 are likely to be dynamic and thus subject

to significant change during the course of tumor development

and/or in response to environmental changes that temporarily

favor tumor growth or quiescence. Small molecule inhibitors

that impair ME1 acetylation and/or dephosphorylation may

thus represent viable future strategies for the treatment of CRC.

STAR+METHODS

Detailed methods are provided in the online version of this paper

and include the following:

d KEY RESOURCES TABLE

d LEAD CONTACT AND MATERIALS AVAILABILITY

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

B Cell Culture and Stable Cell Lines

B Human CRC Samples

d METHOD DETAILS

B Plasmids and Constructs

B Co-IP and Immunoblot

B Metabolic Assays

B Extracellular Acidification Rate (ECAR) and Oxygen

Consumption Rate (OCR) Assays

B ROS Detection in Intracellular and Tumor Tissues

B Reporter, CHIP, and qPCR Assays

B Animal Experiments

B IHC Assays

B Lipidomics Assays

B Bioinformatic Analysis

d QUANTIFICATION AND STATISTICAL ANALYSIS

d DATA AND CODE AVAILABILITY

SUPPLEMENTAL INFORMATION

Supplemental Information can be found online at https://doi.org/10.1016/j.

molcel.2019.10.015.

ACKNOWLEDGMENTS

The authors thank Dr. Jinxiang Zhang (Wuhan Union Hospital, Wuhan, China)

for providing CRC patient samples. This work was supported by grants from

the National Key R&D Program of China (2016YFC1302300), National Nature

Science Foundation of China (81772609, 81802782, 81902843), Medical Sci-

ence Advancement Program (Basic Medical Sciences) of Wuhan University

(TFJC2018005), and China Postdoctoral Science Foundation (2019T120682,

2019T120681).

AUTHOR CONTRIBUTIONS

Y.Z. and Y.L. designed the study. Y.Z. and L.G. performed most of the exper-

iments. X.L. constructed plasmids. K.C. performed bioinformatics analyses.

C.L. and B.L. performed some animal experiments. F.Z. and Q.Z. provided re-

agents and clinical assistance. E.V.P. and C.F. provided reagents and concep-

tual advice. All authors discussed the results. Y.Z., E.P., and Y.L. wrote the

manuscript with comments from all authors.

DECLARATION OF INTERESTS

The authors declare no competing interests.

Received: January 30, 2019

Revised: May 14, 2019

Accepted: October 10, 2019

Published: November 14, 2019

REFERENCES

Ben-Sahra, I., Howell, J.J., Asara, J.M., and Manning, B.D. (2013). Stimulation

of de novo pyrimidine synthesis by growth signaling through mTOR and S6K1.

Science 339, 1323–1328.

Bernkopf, D.B., Jalal, K., Br€uckner, M., Knaup, K.X., Gentzel, M., Schambony,

A., and Behrens, J. (2018). Pgam5 released from damaged mitochondria in-

duces mitochondrial biogenesis via Wnt signaling. J. Cell Biol. 217,

1383–1394.

Chang, G.G., and Tong, L. (2003). Structure and function of malic enzymes, a

new class of oxidative decarboxylases. Biochemistry 42, 12721–12733.

Cheng, J., Qian, D., Ding, X., Song, T., Cai, M., Dan Xie, Wang, Y., Zhao, J., Liu,

Z., Wu, Z., et al. (2018). High PGAM5 expression induces chemoresistance by

enhancing Bcl-xL-mediated anti-apoptotic signaling and predicts poor prog-

nosis in hepatocellular carcinoma patients. Cell Death Dis. 9, 991.

Cheng, L., Zhu, Y., Han, H., Zhang, Q., Cui, K., Shen, H., Zhang, J., Yan, J.,

Prochownik, E., and Li, Y. (2017). MicroRNA-148a deficiency promotes hepat-

ic lipid metabolism and hepatocarcinogenesis in mice. Cell Death Dis.

8, e2916.

Cotte, A.K., Aires, V., Fredon, M., Limagne, E., Derangere, V., Thibaudin, M.,

Humblin, E., Scagliarini, A., de Barros, J.P., Hillon, P., et al. (2018).

Lysophosphatidylcholine acyltransferase 2-mediated lipid droplet production

supports colorectal cancer chemoresistance. Nat. Commun. 9, 322.

Dey, P., Baddour, J., Muller, F., Wu, C.C., Wang, H., Liao, W.T., Lan, Z., Chen,

A., Gutschner, T., Kang, Y., et al. (2017). Genomic deletion of malic enzyme 2

confers collateral lethality in pancreatic cancer. Nature 542, 119–123.

Fan, J., Shan, C., Kang, H.B., Elf, S., Xie, J., Tucker, M., Gu, T.L., Aguiar, M.,

Lonning, S., Chen, H., et al. (2014). Tyr phosphorylation of PDP1 toggles

recruitment between ACAT1 and SIRT3 to regulate the pyruvate dehydroge-

nase complex. Mol. Cell 53, 534–548.

Fan, J., Lin, R., Xia, S., Chen, D., Elf, S.E., Liu, S., Pan, Y., Xu, H., Qian, Z.,

Wang, M., et al. (2016). Tetrameric Acetyl-CoA Acetyltransferase 1 Is

Important for Tumor Growth. Mol. Cell 64, 859–874.

Fernandes, L.M., Al-Dwairi, A., Simmen, R.C.M., Marji, M., Brown, D.M.,

Jewell, S.W., and Simmen, F.A. (2018). Malic Enzyme 1 (ME1) is pro-oncogenic

in ApcMin/+ mice. Sci. Rep. 8, 14268.

Finicle, B.T., Jayashankar, V., and Edinger, A.L. (2018). Nutrient scavenging in

cancer. Nat. Rev. Cancer 18, 619–633.

Gao, T., Li, M., Mu, G., Hou, T., Zhu, W.G., and Yang, Y. (2019). PKCz

Phosphorylates SIRT6 to Mediate Fatty Acid b-Oxidation in Colon Cancer

Cells. Neoplasia 21, 61–73.

Ghosh, S., Wong, S.K., Jiang, Z., Liu, B., Wang, Y., Hao, Q., Gorbunova, V.,

Liu, X., and Zhou, Z. (2018). Haploinsufficiency of Trp53 dramatically extends

the lifespan of Sirt6-deficient mice. eLife 7, e32127.

Gu, L., Zhu, Y., Lin, X., Li, Y., Cui, K., Prochownik, E.V., and Li, Y. (2018).

Amplification of Glyceronephosphate O-Acyltransferase and Recruitment of

USP30 Stabilize DRP1 to Promote Hepatocarcinogenesis. Cancer Res. 78,

5808–5819.

Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of cancer: the next gener-

ation. Cell 144, 646–674.

Hao, Y., Samuels, Y., Li, Q., Krokowski, D., Guan, B.J., Wang, C., Jin, Z., Dong,

B., Cao, B., Feng, X., et al. (2016). Oncogenic PIK3CA mutations reprogram

glutamine metabolism in colorectal cancer. Nat. Commun. 7, 11971.

He, G.W., G€unther, C., Kremer, A.E., Thonn, V., Amann, K., Poremba, C.,

Neurath, M.F., Wirtz, S., and Becker, C. (2017). PGAM5-mediated pro-

grammed necrosis of hepatocytes drives acute liver injury. Gut 66, 716–723.




http://refhub.elsevier.com/S1097-2765(19)30793-2/sref1


























































Jiang, P., Du, W., Mancuso, A., Wellen, K.E., and Yang, X. (2013). Reciprocal

regulation of p53 and malic enzymes modulates metabolism and senescence.

Nature 493, 689–693.

Kugel, S., Sebastian, C., Fitamant, J., Ross, K.N., Saha, S.K., Jain, E.,

Gladden, A., Arora, K.S., Kato, Y., Rivera, M.N., et al. (2016). SIRT6

Suppresses Pancreatic Cancer through Control of Lin28b. Cell 165,

1401–1415.

Liao, S., Davoli, T., Leng, Y., Li, M.Z., Xu, Q., and Elledge, S.J. (2017). A genetic

interaction analysis identifies cancer drivers that modify EGFR dependency.

Genes Dev. 31, 184–196.

Lin, R., Tao, R., Gao, X., Li, T., Zhou, X., Guan, K.L., Xiong, Y., and Lei, Q.Y.

(2013). Acetylation stabilizes ATP-citrate lyase to promote lipid biosynthesis

and tumor growth. Mol. Cell 51, 506–518.

Lu, Y.X., Ju, H.Q., Liu, Z.X., Chen, D.L., Wang, Y., Zhao, Q., Wu, Q.N., Zeng,

Z.L., Qiu, H.B., Hu, P.S., et al. (2018). ME1 Regulates NADPH Homeostasis

to Promote Gastric Cancer Growth and Metastasis. Cancer Res. 78,

1972–1985.

Makinoshima, H., Umemura, S., Suzuki, A., Nakanishi, H., Maruyama, A.,

Udagawa, H., Mimaki, S., Matsumoto, S., Niho, S., Ishii, G., et al. (2018).

Metabolic Determinants of Sensitivity to Phosphatidylinositol 3-Kinase

Pathway Inhibitor in Small-Cell Lung Carcinoma. Cancer Res. 78, 2179–2190.

Mason, J.M., Wei, X., Fletcher, G.C., Kiarash, R., Brokx, R., Hodgson, R.,

Beletskaya, I., Bray, M.R., and Mak, T.W. (2017). Functional characterization

of CFI-402257, a potent and selective Mps1/TTK kinase inhibitor, for the treat-

ment of cancer. Proc. Natl. Acad. Sci. USA 114, 3127–3132.

Panda, S., Srivastava, S., Li, Z., Vaeth, M., Fuhs, S.R., Hunter, T., and Skolnik,

E.Y. (2016). Identification of PGAM5 as a Mammalian Protein Histidine

Phosphatase that Plays a Central Role to Negatively Regulate CD4(+) T

Cells. Mol. Cell 63, 457–469.

Peuker, K., Muff, S., Wang, J., K€unzel, S., Bosse, E., Zeissig, Y., Luzzi, G.,

Basic, M., Strigli, A., Ulbricht, A., et al. (2016). Epithelial calcineurin controlsmi-

crobiota-dependent intestinal tumor development. Nat. Med. 22, 506–515.

Qi, X.M., Wang, F., Mortensen, M., Wertz, R., and Chen, G. (2018). Targeting

an oncogenic kinase/phosphatase signaling network for cancer therapy.

Acta. Pharm. Sin. B. 8, 511–517.

Qian, H., Chen, Y., Nian, Z., Su, L., Yu, H., Chen, F.J., Zhang, X., Xu, W., Zhou,

L., Liu, J., et al. (2017). HDAC6-mediated acetylation of lipid droplet-binding

protein CIDEC regulates fat-induced lipid storage. J. Clin. Invest. 127,

1353–1369.

Ramachandran, A., and Jaeschke, H. (2017). PGAM5: a new player in immune-

mediated liver injury. Gut 66, 567–568.


Rauschenberger, V., Bernkopf, D.B., Krenn, S., Jalal, K., Heller, J., Behrens, J.,

Gentzel, M., and Schambony, A. (2017). The phosphatase Pgam5 antagonizes

Wnt/b-Catenin signaling in embryonic anterior-posterior axis patterning.

Development 144, 2234–2247.

Robitaille, A.M., Christen, S., Shimobayashi, M., Cornu, M., Fava, L.L., Moes,

S., Prescianotto-Baschong, C., Sauer, U., Jenoe, P., and Hall, M.N. (2013).

Quantitative phosphoproteomics reveal mTORC1 activates de novo pyrimi-

dine synthesis. Science 339, 1320–1323.

Shan, C., Kang, H.B., Elf, S., Xie, J., Gu, T.L., Aguiar, M., Lonning, S., Hitosugi,

T., Chung, T.W., Arellano, M., et al. (2014). Tyr-94 phosphorylation inhibits py-

ruvate dehydrogenase phosphatase 1 and promotes tumor growth. J. Biol.

Chem. 289, 21413–21422.

Shen, H., Xing, C., Cui, K., Li, Y., Zhang, J., Du, R., Zhang, X., and Li, Y. (2017).

MicroRNA-30a attenuates mutant KRAS-driven colorectal tumorigenesis via

direct suppression of ME1. Cell Death Differ. 24, 1253–1262.

Sociali, G., Grozio, A., Caffa, I., Schuster, S., Becherini, P., Damonte, P.,

Sturla, L., Fresia, C., Passalacqua, M., Mazzola, F., et al. (2019). SIRT6 deace-

tylase activity regulates NAMPT activity and NAD(P)(H) pools in cancer cells.

FASEB J. 33, 3704–3717.

Torre, L.A., Bray, F., Siegel, R.L., Ferlay, J., Lortet-Tieulent, J., and Jemal, A.

(2015). Global cancer statistics, 2012. CA Cancer J. Clin. 65, 87–108.

Wang, Z., Jiang, H., Chen, S., Du, F., and Wang, X. (2012). The mitochondrial

phosphatase PGAM5 functions at the convergence point of multiple necrotic

death pathways. Cell 148, 228–243.

Wen, D., Liu, D., Tang, J., Dong, L., Liu, Y., Tao, Z., Wan, J., Gao, D., Wang, L.,

Sun, H., et al. (2015). Malic enzyme 1 induces epithelial-mesenchymal transi-

tion and indicates poor prognosis in hepatocellular carcinoma. Tumour Biol.

36, 6211–6221.

Yang, W., Zheng, Y., Xia, Y., Ji, H., Chen, X., Guo, F., Lyssiotis, C.A., Aldape,

K., Cantley, L.C., and Lu, Z. (2012). ERK1/2-dependent phosphorylation and

nuclear translocation of PKM2 promotes the Warburg effect. Nat. Cell Biol.

14, 1295–1304.

Zadra, G., Ribeiro, C.F., Chetta, P., Ho, Y., Cacciatore, S., Gao, X., Syamala,

S., Bango, C., Photopoulos, C., Huang, Y., et al. (2019). Inhibition of de novo

lipogenesis targets androgen receptor signaling in castration-resistant pros-

tate cancer. Proc. Natl. Acad. Sci. USA 116, 631–640.

Zhu, Y., Gu, L., Li, Y., Lin, X., Shen, H., Cui, K., Chen, L., Zhou, F., Zhao, Q.,

Zhang, J., et al. (2017). miR-148a inhibits colitis and colitis-associated tumor-

igenesis in mice. Cell Death Differ. 24, 2199–2209.



















































































STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies

PGAM5 abcam Cat# ab126534; RRID: AB_11127076

PGAM5 Santa Cruz Cat# sc-515880

ACAT1 zen-bio Cat# 513273

ACAT1 OriGene Cat# AM50027PU-N

NEK1 abcam Cat# ab229489

NEK1 Santa Cruz Cat# sc-398813

SIRT6 abcam Cat# ab191385

SIRT6 Santa Cruz Cat# sc-517196

Acetylated-Lysine Antibody CST Cat# 9441; RRID: AB_331805

Acetylated-Lysine Antibody Protein gene Cat# 2067

phosphorylation-Serine Antibody Santa Cruz Cat# sc-81514; RRID: AB_1128624

phosphorylation-Threonine Antibody CST Cat# 9381; RRID: AB_330301

ME1 Santa Cruz Cat# sc-100569; RRID: AB_2143832

ME1 Proteintech Cat# 16619-1-AP; RRID: AB_2143821

b-Catenin Abcam Cat# ab6302; RRID: AB_305407

c-Myc Santa Cruz Cat# A2814

HA OriGene Cat# AF4911

Flag abcam Cat# ab205606

Flag SIGMA Cat# F3165; RRID: AB_259529

b-actin Santa Cruz Cat# H1914

Chemicals and Reagent

PMSF Calbiochem 80055-380

Triton X-100 Thermo Fisher Scientific 28314

Trypsin-EDTA Gibco 15400054

Fetal Bovine Serum (FBS) Atlanta Biologicals S11550 LOT#M15135

Penicillin/Streptomycin Gibco 15140-122

DMSO Sigma-Aldrich D2650

Oligomycin Sigma-Aldrich 75351

FCCP Sigma-Aldrich C2920

Antimycin A Sigma-Aldrich A8674

Rotenone Sigma-Aldrich R8875

Wnt3a abcam ab23327

TRIzol ThermoFisher Scientific 15596026

Immobilon Western Chemiluminescent HRP substrate EMD Millipore WBKLS0500

T4 DNA Ligase NEB M0202L

Protein G Sepharose Sigma-Aldrich P3296

Dolbecco’s Modification of Eagle’s Medium Corning Life Sciences 10-013-CV

Critical Commercial Assays

BCA Assay Thermo Fisher Scientific 23225

Triglyceride Quantification Colorimetric Assay Kit BioVision K952

Fatty Acids Assay Kit BioVision K956

XF Glycolysis Stress Test Kit Agilent Technologies 103020

XF Cell Mito Stress Test Kit Agilent Technologies 103015

Pyruvate Colorimetric Assay kit Biovision K609

(Continued on next page)

Molecular Cell 77, 1–12.e1–e5, January 2, 2020 e1

Continued

REAGENT or RESOURCE SOURCE IDENTIFIER

Lactate Colorimetric Assay Kit II Biovision K627

ATP Colorimetric Assay kit Biovision K354

NADP/NADPH Quantification Colorimetric Kit BioVision K347-100

OxiselectTM In vitro ROS/RNS assay kit Cell Biolabs STA-347

Deposited Data

Raw images This paper and Mendeley Data https://doi.org/10.17632/b2jgrw648t.1

Experimental Models: Cell Lines

Human embryonic kidney HEK293T cells (N/A) ATCC CRL-3216

Human colorectal carcinoma HCT116 cells (male) ATCC CCL-247

Human colorectal adenocarcinoma SW480 cells (male) ATCC CCL-228

Human colorectal carcinoma RKO cells (N/A) ATCC CRL-2577

Human colorectal adenocarcinoma DLD1 (male) ATCC CCL-221

Colon normal epithelial cell line NCM460 (N/A) San Antonio N/A

Human normal epithelial cell line RPE (N/A) ATCC CRL-2240

Human normal fibroblast cell line HFF (male) ATCC SCRC-1041

Human colorectal adenocarcinoma HT-29 (female) ATCC HTB-38

Experimental Models: Organisms/Strains

C57B6/J mice NRCMM N/A

Recombinant DNA

Phage-PGAM5 and its mutans This paper N/A

HA-PGAM5 This paper N/A

Phage-NEK1 and its mutans This paper N/A

HA-NEK1 This paper N/A

HA-b-Catenin This paper N/A

Phage-ACAT1 and its mutans This paper N/A

HA-ACAT1 This paper N/A

HA-SIRT6 This paper N/A

Flag-SIRT6 and its mutans This paper N/A

Flag-ME1 and its mutans This paper N/A

HA-ME1 and its mutans This paper N/A

Myc-ME1 This paper N/A

Ad-ME1 and its mutans This paper N/A

Oligonucleotides

See Table S1 for qPCR primer list This paper N/A

shRNAs targeting sequence, see Table S1 This paper N/A

Software

ImageJ NIH https://imagej.nih.gov/ij/

Graphpad Prism 7 Graphpad https://www.graphpad.com/

scientificsoftware/prism/

GEPIA N/A http://gepia.cancer-pku.cn/detail.php

The Human Protein Atlas N/A https://www.proteinatlas.org/

Oncomine N/A www.oncomine.org


LEAD CONTACT AND MATERIALS AVAILABILITY

Further information and requests for resoures and reagents should be directed to and will be fulfilled by the Lead Contact, Youjun Li

([email protected]).

All unique/stable reagents generated in this study are available from the Lead Contact with a completed Materials Transfer

Agreement.

e2 Molecular Cell 77, 1–12.e1–e5, January 2, 2020

mailto:[email protected]

https://doi.org/10.17632/b2jgrw648t.1

https://imagej.nih.gov/ij/

https://www.graphpad.com/scientificsoftware/prism/

https://www.graphpad.com/scientificsoftware/prism/

http://gepia.cancer-pku.cn/detail.php

https://www.proteinatlas.org/

http://www.oncomine.org


EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell Culture and Stable Cell LinesHuman embryonic kidney HEK293 cells, human retinal pigmented epithelium (RPE) and human foreskin fibroblast (HFF) cells immor-

talized with hTERT, normal colonic epithelial cells NCM460, human colon tumor-derived cell lines HCT116, SW480, RKO, HT29,

DLD1 were obtained from the American Type Culture Collection (Rockville, MD, USA). All cells were regularly authenticated by short

tandem repeat analysis, tested for absence ofMycoplasma contamination andwere usedwithin 5 passages after thawing. They were

routinely cultured as previously described (Zhu et al., 2017).

Stable cell lines expressing the indicated vectors were generated by retroviral or lentiviral transduction in the presence of 8 mg/mL

polybrene followed by selection with blasticidin, puromycin or G418 for at least 10 days. Stable cell lines were examined for the

expression of the indicated vectors by real-time PCR or western blot.

Human CRC SamplesCRC samples were used as previously described (Zhu et al., 2017). Written informed consent was obtained from all patients at the

Union Hospital in Wuhan, China. The studies were conducted in accordance with Declaration of Helsinki and approved by the review

board of Wuhan University. The diagnoses of all samples were confirmed by histological review.

METHOD DETAILS

Plasmids and ConstructsThe open reading frame of human PGAM5, NEK1, ACAT1, SIRT6 and ME1 were amplified and cloned into pHAGE-CMV-MCS-

PGK-33 Flag and pCMV-HA vectors. Mutations in the PGAM5, NEK1 and ME1 cDNA sequences were generated by overlap exten-

sion PCR. Human PGAM5, NEK1, ACAT1, SIRT6 and ME1 shRNA vectors were obtained from GENECHEM (Shanghai, China).

Adenovirus was obtained from Yan Wang (College of Life Science, Wuhan University). Primers and shRNAs used in this study are

listed in Tables S1 and S2.

Co-IP and ImmunoblotTransfected cells were lysed in 1 mL lysis buffer [20 mM Tris (pH 7.4), 300 mM NaCl, 1% Triton, 1 mM EDTA, 10 mg/ml aprotinin,

10 mg/ml leupeptin, and 1 mM PMSF]. Sepharose beads were washed three times with 1 mL lysis buffer containing 150 mM

NaCl. Co-IP and immunoblot analysis was performed as described (Zhu et al., 2017). To detect ME1 phosphorylation and acetylation

levels, immunoprecipitations were performed using ME1 antibody. This was followed by immunoblot using pan serine phosphory-

lation and acetylation antibodies.

Metabolic AssaysThe levels of NADPH, triglyceride and fatty acids in cultured cells were determined using a NADP+/NADPH quantification Kit, Triglyc-

eride Assay Kit and Fatty Acid Assay Kit (from BioVision, Milpitas, CA, USA) respectively, following the manufacturer’s instructions.

ME1 and ME2 activity were determined using cytosolic extracts and lysate of mitochondria as described, respectively (Jiang

et al., 2013).

For ATP quantification, cells were collected and extracted in ATP Assay Buffer (Biovision). The cells were centrifuged at 1.23 104

rpm for 5 min and the supernatant was used for ATP assay. The reaction mixture was incubated in the dark for 30 min at room tem-

perature, and measured at 570 nm in a microplate reader. Values were normalized to cell number.

For lactate production, cells were plated into 12-well plates and incubated in DMEM with 10% FBS overnight. The medium was

removed, cells were incubated with DMEM without FBS for 1 hr and the supernatant was used for lactate assays (Biovision). The

reaction mixture was incubated in the dark for 30 min at room temperature. The lactate levels were measured at 450 nm in a micro-

plate reader and normalized to cell number. For the measurement of tumor lactate levels, 10 mg of tissue was homogenized in Assay

Buffer (Biovision). Samples were centrifuged and the soluble fractions were measured and normalized to protein concentrations.

For pyruvate kinase activity assays, cells were harvested and extracted with 4 volume of the Pyruvate Kinase Assay Buffer

(Biovision). The cells were centrifuged at 1.23 104 rpm for 10 min and the resultant supernatents were assayed for 20 min according

to the directions of the supplier. Results were normalized to the cell number.

Isotope experiments were performed as previously described (Gu et al., 2018). Briefly, 3H-acetic acid (1.59 Ci/mmol, Perkin Elmer,

37 KBq/ml) was added to the wells and incubated at 37�C overnight. Then cells were disrupted with lysis buffer (100mM Tris-HCl pH

7.5, 150 mM NaCl, 2% SDS, 10% glycerol, 12.5 mM EDTA, 1% Triton X-100 and 0.5% NP-40) and incubated at 37�C for 30 min.

Organic extractions were performed in chloroform following the protocol of Bligh andDyer and lipid-soluble products weremeasured

by scintillation counting (Tri-Carb 2910 TR, Perkin Elmer).

Extracellular Acidification Rate (ECAR) and Oxygen Consumption Rate (OCR) AssaysThe Extracellular Acidification Rate (ECAR) and Oxygen Consumption Rate (OCR) were measured using the Seahorse XFe24

Extracellular Flux Analyzer (Seahorse Bioscience). Experiments were performed according to the manufacturer’s instructions.



Briefly, 33 104 cells per well were seeded into a Seahorse XF 24 cell culture microplate overnight. Following incubation in the assay

medium (XF-Base medium supplemented plus 2 mM L-glutamine) for 1 hr at 37�C (non-CO2 incubator), ECAR was measured at the

basal level and in response to sequential injection of glucose (10mM), oligomycin (1 mM) and 2-DG (50mM) (XFGlycolysis Stress Test

Kit, Agilent Technologies), respectively. OCR was measured following incubation in the assay medium (XF Base Medium supple-

mented with 10 mMD-glucose, 1 mM sodium pyruvate and 2mM L-glutamine) for 1 hr at 37�C (non-CO2 incubator) at the basal level

and in response to injection of oligomycin (1 mM), FCCP (2 mM), and antimycin/rotenone (0.5 mM) (XF Cell Mito Stress Test Kit, Agilent

Technologies), respectively. Data were analyzed by Seahorse XF24 Wave software. OCR is reported in pmols/minute and ECAR in

mpH/minute. The results were normalized to cell number.

ROS Detection in Intracellular and Tumor TissuesIntracellular ROS were measured by detecting dichlorodihydrofluorescein, a cleavage product of 20, 7’-dichlorodihydrofluoresceindiacetate (H2DCF-DA, Sigma). Cultured cells were washed with PBS and incubated with 10 mM H2DCF-DA at 37�C for 30min.

Then cells were dissociated after 30min culture in phenol red-free medium. Fuorescence wasmonitored with a plate reader (Fluostar

OPTIMA, BMG LABTECH).

For the measurement of ROS in xenograft tumors, OxiSelect TM in vitro ROS assay kit (Cell Biolabs, cat. STA-347) was used.

Briefly, 20 mg of tumor tissue was homogenized in 1 mL of cold PBS and then centrifuged at 104 x g for 5 min to remove the insoluble

particles. 50 mL of supernatant wasmixedwith the same volume of catalyst. Following a 5min incubation at room temperature, 100 mL

DCFH-DiOxyQ probe solution was added to the mixture to measure the total free radical population. The sample were incubated for

30 min and analyzed with a fluorescence plate reader at excitation/emission = 480/530 nm. A standard curve using H2O2 was gener-

ated to quantify the free radical content in the tumor lysate samples.

Reporter, CHIP, and qPCR AssaysHuman PGAM5 and ME1 promoter sequences were inserted into pGL3-basic luciferase vector (Promega). Luciferase assays were

performed as described previously (Zhu et al., 2017).

Chromatin immuno-precipitation (CHIP) was performed as described previously (Zhu et al., 2017). In brief, intracellular protein-

DNA complexes were cross-linked with 1% formaldehyde, sonicated and subjected to chromatin-conjugated IP using specific an-

tibodies. After reversal of cross-links, precipitated DNA was purified and analyzed by qPCR with the primers indicated in Table S3.

ForqRT-PCR, totalRNAwas isolatedusingTrizol (Invitrogen,Carlsbad,CA) followedbyDNase (ThermoScientific) treatment.Reverse

transcription was performed with a cDNA Synthesis Kit (Promega) and qPCRwere performed using SYBRGreen master mix (Bio-Rad,

Hercules, CA) using standard protocols. All primer sequences used are shown in Table S4. b-actin was used as an internal control.

Animal ExperimentsAll animal studies were approved by the Animal Care Committee of Wuhan University. For xenograft experiments, four-week-old

male BALB/c nude mice were purchased from Model Animal Research Center (Nanjing, China) and maintained in microisolator ca-

ges. Detailed procedures were described previously (Zhu et al., 2017).

AOM/DSS-induced CRCswere generated following the previously described procedures (Zhu et al., 2017). Briefly, eight-week-old

female C57BL/6 mice were injected intraperitoneally with 10mg/kg AOM (Sigma-Aldrich). Seven days later, mice were given drinking

water containing 3% DSS (MP Biomedicals, Santa Ana, CA, USA) for 7 days followed by 2 weeks regular drinking water. Then mice

were fed three rounds with 2.5% DSS water for 1 week and sacrificed on day 120.

For administration of adenovirus to AOM/DSS-induced CRC mice, female C57BL/6 mice were randomly distributed into 9 groups

of 10 animals each. The first group was treated with adenovirus expressing Ad-shCtrl and two groups of mice were treated with

adenovirus expressing shPgam5. Six additional groups of mice were treated with adenovirus expressing the control vector or WT

ME1 or the indicate ME1 mutants. Concentrated viruses in a volume of 100 mL were delivered intraperitoneally twice per week for

3 weeks. At day 120, tumor burden was evaluated after sacrificing the mice (Peuker et al., 2016; Zhu et al., 2017).

For administration of adenovirus in ApcMin/+ mice model, female ApcMin/+ mice were randomly distributed into 3 groups (10 mice

each) 3 wks after birth. The first group of mice was treated with adenovirus expressing shCtrl and the other two groups of mice were

treated with adenovirus expressing shPgam5. On week 20, tumor burden was assessed. Visible tumors larger than 1 mm were

measured and enumerated. Colon tissue was harvested for RNA and protein assays as well as pathological studies (Peuker et al.,

2016; Zhu et al., 2017).

IHC AssaysIHC assayswasperformed as described previously (Zhu et al., 2017). Briefly, colon sectionswere fixed in 10% formalin and embedded

in paraffin according to standard protocols. Colons were sectioned and stained with hematoxylin and eosin (H&E). Ki-67 staining was

performed according to the supplier’s protocol (mouse monoantibody, 1:200, Sc-25280; Santa Cruz Biotechnology).

Lipidomics AssaysColon sections were quickly dissected and snap frozen in liquid nitrogen with a pre-cooled Wollenberger clamp. Frozen samples

were ground in liquid nitrogen and then placed into 1 mL of 0.3 M KOH in 90% methanol at 80�C for 1 hr in a 2 mL glass vial for

e4 Molecular Cell 77, 1–12.e1–e5, January 2, 2020


saponification. Formic acid (0.1 mL) was added for neutralization. The saponified fatty acids were extracted with 0.5 mL of hexane

and vortexed, and the top organic layer was transferred to a new glass vial. Samples were then dried under a stream of N2 and dis-

solved in 1mL of isopropanol:methanol (1:1, v/v) solution for LC-MS analysis. Separation was performed by reversed-phase ion-pair-

ing chromatography on a C8 column coupled to negative-ion mode, full-scan LC-MS at 1-Hz scan time and 100,000 resolving power

(stand-alone orbitrap; Thermo Fischer Scientific).

Bioinformatic AnalysisCRC datasets were downloaded from the Cancer Genome Atlas (TCGA) data portal (http://tcga-data.nci.nih.gov). Gene expression

was assessed and Kaplan Meier curves were analyzed in human CRC tissues from the TCGA RNA-seq dataset (n = 378). CRC sam-

ples were assigned to two groups based on gene expression level using the minimum p value approach.

QUANTIFICATION AND STATISTICAL ANALYSIS

Experimental data were analyzed using the unpaired, two-tailed Student’s t test, Wilcoxon signed rank test, or Mann-Whitney U-test

and the correlation was analyzed using a Spearman rank correlation test. Kaplan Meier curves for survival were analyzed with

GraphPad software using the log-rank test. Results are represented as the mean ± SD and p < 0.05 was considered statistically

significant.

DATA AND CODE AVAILABILITY

The original data have been deposited to Mendeley Data and are available at https://doi.org/10.17632/b2jgrw648t.1.


http://tcga-data.nci.nih.gov

https://doi.org/10.17632/b2jgrw648t.1

dynamic regulation of me1 phosphorylation and acetylation ......molecular cell article dynamic...

Documents