CA9 Silencing Promotes Mitochondrial Biogenesis, Increases Putrescine Toxicity and Decreases Cell Motility to Suppress ccRCC Progression

Carbonic anhydrase IX (CA9), a pH-regulating transmembrane protein, is highly expressed in solid tumors, and particularly in clear cell renal cell carcinoma (ccRCC). The catalytic mechanisms of CA9 are well defined, but its roles in mediating cell migration/invasion and survival in ccRCC remain to be determined. Here, we confirmed that the mRNA expression of CA9 in ccRCC was significantly higher than that in para-carcinoma tissues from analysis of the datasets in The Cancer Genome Atlas. CA9 knockdown upregulated oxidative phosphorylation-associated proteins and increased mitochondrial biogenesis, resulting in the reversal of the Warburg phenotype and the inhibition of cell growth. Our study revealed that CA9 knockdown upregulated mitochondrial arginase 2 (ARG2), leading to the accumulation of putrescine, which suppressed ccRCC proliferation. Surfaceomics analysis revealed that CA9 knockdown downregulated proteins associated with extracellular matrix (ECM)—receptor interaction and cell adhesion, resulting in decreased cell migration. CA9 silencing also downregulated amino acid transporters, leading to reduced cellular amino acids. Collectively, our data show that CA9 knockdown suppresses proliferation via metabolic reprogramming and reduced cell migration, reaffirming that CA9 is a potential therapeutic target for ccRCC treatment.

Herein, we manipulated CA9 expressions by knockdown and overexpression in ccRCC cells and systematically analyzed effects of CA9 on promoting cell survival and migration. We found that CA9 silencing resulted in the accumulation of putrescine and increased mitochondrial biogenesis, leading to decreased cell proliferation. We carried out a quantitative surfaceomics analysis and found that CA9 silencing downregulated amino acid transporters and proteins associated with cell motility.

CA9 Knockdown Inhibits Cell Growth in ccRCC Cells
To confirm that CA9 is overexpressed in tumor tissues compared with para-carcinoma tissues in kidney renal clear cell carcinoma (KIRC), we analyzed the transcriptome data of paired tumor samples and normal tissues in The Cancer Genome Atlas (TCGA) ( Figure 1A). We found that CA9 was significantly higher in KIRC tissues than in normal tissues ( Figure 1B). To examine the effects of CA9 on ccRCC progression, we knocked down (KD) or overexpressed (OE) CA9 in 786-O and 769-P to establish stable cell lines. Two different short hairpin RNAs (shRNAs) against CA9 were used to establish stable CA9 knockdown cell lines, designated CA9 KD1 and KD2, and cells expressing nontargeting shRNA were used as negative control. Cells were transduced with the lentivirus vector encoding human CA9 to stably overexpress CA9, and were designated 786-O-CA9-OE, while the cells transduced with the lentivirus vector encoding the empty pLVX-IRES-ZsGreen1 cassette were used as the control, designated 786-O-plvx. The expression of CA9 in 786-O and 769-P cells was confirmed by quantitative real-time PCR (qPCR) and Western blotting, revealing that CA9 expression was reduced by more than 90% and 60% in 786-O-CA9-KD and 769-P-CA9-KD cells, respectively, compared with control cells, while it was increased more than 60-fold in 786-O-CA9-OE cells ( Figure 1C,D and Figure S1). Cell Counting Kit-8 (CCK-8) assays showed that CA9 knockdown inhibited cell growth ( Figure 1E,F), while CA9 overexpression promoted cell proliferation ( Figure 1G). These results demonstrate that high CA9 expression promoted ccRCC cell proliferation, while low expression inhibited cell growth.  Table S1. (B) The mean mRNA level of CA9 in ccRCC tissues (num(T) = 523) and normal tissues (num(N) = 72) from TCGA data. (C) mRNA expression of CA9 decreased in 786-O-CA9-KD cells compared with control cells, measured by quantitative real-time PCR (qPCR). ACTB was used as a control. (D) Western blotting of CA9 revealed that the expression of CA9 was reduced in 786-O-CA9-KD cells. βactin was used as a control. (E) Knockdown of CA9 in 786-O cells inhibited cell growth compared with control cells. (F) Knockdown of CA9 in 769-P cells inhibited cell growth. (G) Overexpression of CA9 in 786-O cells promoted cell proliferation. Cell proliferation curves measured by the Cell Counting Kit-8 (CCK-8) assays. Significance was calculated by Student's t-test. **** p < 0.0001, ** p < 0.01, * p < 0.05; n = 3, mean ± SEM.

CA9 Knockdown Increases Mitochondrial Biogenesis and Reverses Warburg Metabolic Phenotypes
To elucidate CA9 function in the progression of ccRCC, quantitative proteomics analysis was carried out to characterize the proteome changes after CA9 knockdown in 786-O cells. Most experiments were conducted using the CA9-KD2 cells in the present study. The percentage variations corresponding to 88% coverage were taken as the threshold cut-off [44]; thus, proteins with ratios ≥ 1.33 or ≤ 0.75 and p-values < 0.05 were regarded as the upregulated or downregulated proteins, respectively, in our results (Figure 2A,B). A total of 6513 proteins were identified in biological triplicates including 748 downregulated proteins and 726 upregulated proteins in 786-O-CA9-KD cells compared with control cells (Tables S2 and S3). Differentially expressed proteins (DEPs) after CA9 knockdown in 786-O cells were analyzed by Kyoto Encyclopedia of Genes and Genomes (KEGG) BlastKOALA [45]. The downregulated proteins were enriched in pathways of carbohydrate metabolism, cofactors and vitamins metabolism, and nucleotide metabolism ( Figure 2C). The genetic information processing pathways enriched by the downregulated proteins were mainly focused on  Table S1. assays. Significance was calculated by Student's t-test. **** p < 0.0001, ** p < 0.01, * p < 0.05; n = 3, mean ± SEM.

CA9 Knockdown Increases Mitochondrial Biogenesis and Reverses Warburg Metabolic Phenotypes
To elucidate CA9 function in the progression of ccRCC, quantitative proteomics analysis was carried out to characterize the proteome changes after CA9 knockdown in 786-O cells. Most experiments were conducted using the CA9-KD2 cells in the present study. The percentage variations corresponding to 88% coverage were taken as the threshold cut-off [44]; thus, proteins with ratios ≥1.33 or ≤0.75 and p-values <0.05 were regarded as the upregulated or downregulated proteins, respectively, in our results (Figure 2A (Tables S2 and S3). Differentially expressed proteins (DEPs) after CA9 knockdown in 786-O cells were analyzed by Kyoto Encyclopedia of Genes and Genomes (KEGG) BlastKOALA [45]. The downregulated proteins were enriched in pathways of carbohydrate metabolism, cofactors and vitamins metabolism, and nucleotide metabolism ( Figure 2C). The genetic information processing pathways enriched by the downregulated proteins were mainly focused on ribosome biogenesis, DNA repair and recombination, and chromosome and transcription factors. Interestingly, most mitochondrial proteins were upregulated in CA9 knockdown cells ( Figure 2D). Ingenuity Pathway Analysis (IPA) further showed that proteins related to oxidative phosphorylation (OXPHOS), fatty acid β-oxidation, and mitochondrial L-carnitine shuttle were upregulated, while proteins associated with the pentose phosphate pathway (PPP) were downregulated in CA9 knockdown cells ( Figure 2E). Furthermore, the arginine metabolism-related pathways and the pathways related to cell motility, including p21-activated kinase (PAK) signaling, paxillin signaling, integrin signaling, Ras-related C3 botulinum toxin substrate (RAC) and actin cytoskeleton signaling, were also enriched in DEPs ( Figure 2E). These results indicate that CA9 expression regulated cell metabolism and cell motility. ribosome biogenesis, DNA repair and recombination, and chromosome and transcription factors. Interestingly, most mitochondrial proteins were upregulated in CA9 knockdown cells ( Figure 2D). Ingenuity Pathway Analysis (IPA) further showed that proteins related to oxidative phosphorylation (OXPHOS), fatty acid β-oxidation, and mitochondrial L-carnitine shuttle were upregulated, while proteins associated with the pentose phosphate pathway (PPP) were downregulated in CA9 knockdown cells ( Figure 2E). Furthermore, the arginine metabolism-related pathways and the pathways related to cell motility, including p21-activated kinase (PAK) signaling, paxillin signaling, integrin signaling, Ras-related C3 botulinum toxin substrate (RAC) and actin cytoskeleton signaling, were also enriched in DEPs ( Figure 2E). These results indicate that CA9 expression regulated cell metabolism and cell motility.  As shown in Figure 3A,B, 16 OXPHOS-related proteins and 28 mitochondrial ribosomal proteins were upregulated in CA9-KD cells; thus, we inferred that mitochondrial biogenesis was upregulated after CA9 knockdown. As predicted, CA9 knockdown increased MitoTracker staining intensity and enhanced the expressions of key factors in mitochondrial biogenesis, including peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), nuclear factor erythroid 2-related factor 2 (NRF2), and mitochondrial transcription factor 1 (TFAM) ( Figure 3C,D and Figure S2A-C) in 786-O and 769-P cells [46,47]. Furthermore, mitochondrial DNA (mtDNA) content, which is proportional to the number of mitochondria [46,48], was also increased after CA9 knockdown in 786-O and 769-P cells ( Figure S2D). Reactive oxygen species (ROS) are byproducts of electron transport chain activity, and we reasoned that an increase in the mitochondrial biogenesis could enhance the cellular ROS level after CA9 knockdown [49]. Indeed, the ROS level was higher in CA9 knockdown cells detected by CellRox Deep Red Reagent ( Figure S2E). Moreover, CA9 knockdown resulted in downregulation of glycolytic enzymes such as hexokinase-2 (HK2), glucose-6-phosphate 1-dehydrogenase (G6PD), ATP-dependent 6-phosphofructokinase (PFKM), and phosphoglycerate mutase 1 (PGAM1). Consistently, levels of glucose-6-phosphate (G6P), lactate, Uridine-5 -diphosphate glucose (UDP-G), and 6-phosphpogluconate (6PG) were decreased after CA9 knockdown ( Figure S3). These results demonstrated that CA9 knockdown increased mitochondrial biogenesis and reversed the Warburg metabolic phenotypes. As shown in Figure 3A,B, 16 OXPHOS-related proteins and 28 mitochondrial ribosomal proteins were upregulated in CA9-KD cells; thus, we inferred that mitochondrial biogenesis was upregulated after CA9 knockdown. As predicted, CA9 knockdown increased MitoTracker staining intensity and enhanced the expressions of key factors in mitochondrial biogenesis, including peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), nuclear factor erythroid 2related factor 2 (NRF2), and mitochondrial transcription factor 1 (TFAM) ( Figure 3C,D and Figure  S2A-C) in 786-O and 769-P cells [46,47]. Furthermore, mitochondrial DNA (mtDNA) content, which is proportional to the number of mitochondria [46,48], was also increased after CA9 knockdown in 786-O and 769-P cells ( Figure S2D). Reactive oxygen species (ROS) are byproducts of electron transport chain activity, and we reasoned that an increase in the mitochondrial biogenesis could enhance the cellular ROS level after CA9 knockdown [49]. Indeed, the ROS level was higher in CA9 knockdown cells detected by CellRox Deep Red Reagent ( Figure S2E). Moreover, CA9 knockdown resulted in downregulation of glycolytic enzymes such as hexokinase-2 (HK2), glucose-6-phosphate 1-dehydrogenase (G6PD), ATP-dependent 6-phosphofructokinase (PFKM), and phosphoglycerate mutase 1 (PGAM1). Consistently, levels of glucose-6-phosphate (G6P), lactate, Uridine-5′diphosphate glucose (UDP-G), and 6-phosphpogluconate (6PG) were decreased after CA9 knockdown ( Figure S3). These results demonstrated that CA9 knockdown increased mitochondrial biogenesis and reversed the Warburg metabolic phenotypes.

CA9 Knockdown Increases Putrescine Production to Inhibit Cell Growth
To confirm proteomics results suggesting that CA9 silencing reprogramed cellular metabolism, we performed metabolomics analysis in five biological replicates to investigate the effects of CA9 knockdown on cellular metabolism in 786-O cells. Metabolomics analysis showed that CA9 knockdown increased the cellular contents of 24 metabolites while it decreased 127 metabolites (Tables S4 and S5). Interestingly, we found that putrescine was increased seven-fold in CA9 knockdown cells compared with control cells ( Figure 4A). Putrescine is generated from ornithine decarboxylation catalyzed by ornithine decarboxylase (ODC) and ornithine production is dependent on arginine metabolism [50,51]. Indeed, the metabolomics results show that CA9 knockdown decreased the abundance of arginine, whereas it increased the contents of putrescine, ornithine, and citrulline ( Figure 4B).

CA9 Knockdown Increases Putrescine Production to Inhibit Cell Growth
To confirm proteomics results suggesting that CA9 silencing reprogramed cellular metabolism, we performed metabolomics analysis in five biological replicates to investigate the effects of CA9 knockdown on cellular metabolism in 786-O cells. Metabolomics analysis showed that CA9 knockdown increased the cellular contents of 24 metabolites while it decreased 127 metabolites (Tables S4 and S5). Interestingly, we found that putrescine was increased seven-fold in CA9 knockdown cells compared with control cells ( Figure 4A). Putrescine is generated from ornithine decarboxylation catalyzed by ornithine decarboxylase (ODC) and ornithine production is dependent on arginine metabolism [50,51]. Indeed, the metabolomics results show that CA9 knockdown decreased the abundance of arginine, whereas it increased the contents of putrescine, ornithine, and citrulline ( Figure 4B).  Then, isotope tracing with 13 C 6 -arginine was performed to determine the effects of CA9 knockdown on putrescine production from arginine. Five carbons in arginine are converted to ornithine, and then four carbons are incorporated into putrescine during putrescine synthesis ( Figure 4C). 13 C 6 -arginine isotope tracing revealed that M+5 ornithine, M+5 citrulline, and M+4 putrescine increased after CA9 knockdown, suggesting putrescine synthesis was enhanced ( Figure 4D). Arginase 2 (ARG2), a mitochondrial enzyme catalyzing the conversion of arginine to ornithine, was significantly upregulated after CA9 knockdown identified by quantitative proteomics data and Western blotting. ODC catalyzing ornithine decarboxylation was also upregulated, while argininosuccinate synthase 1 (ASS1) and argininosuccinate lyase (ASL) catalyzing the conversion of citrulline to arginine were downregulated after CA9 knockdown ( Figure 4E,F and Figure S4). The expression of the enzymes involved in the metabolism of arginine in 769-P cells is consistent with that in 786-O cells ( Figure S4). These results demonstrated that putrescine synthesis from arginine was increased by CA9 knockdown in ccRCC cells. Putrescine was used to treat 786-O cells to examine its effects on cell growth. The viability of 786-O cells after putrescine treatment for 24 h was detected by CCK-8 kit, and putrescine was found to induce cell death with an IC50 (the half maximal inhibitory concentration) of 16.6 mM ( Figure 4G), which is consistent with previous research [52].

CA9 Silencing Downregulates Amino Acid Transporters and Proteins Associated with Cell Motility in ccRCC Cells
Metabolomics analysis also showed that most amino acids were reduced in CA9 knockdown cells ( Figure 5A). To explore whether CA9 silencing downregulates amino acid transporters, we performed a surfaceomics analysis to investigate changes in membrane proteins between 786-O-CA9-KD cells and control cells. Briefly, surface proteins were biotinylated by sulfo-NHS-SS-Biotin and enriched by streptavidin conjugated sepharose, followed by in-gel digestion and liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis ( Figure 5B). A total of 239 differentially expressed membrane proteins were identified, including 146 downregulated proteins and 93 upregulated proteins (Tables S6 and S7). We found that six amino acid transporters were downregulated in CA9 knockdown cells, including sodium-coupled neutral amino acid transporter 1 (SLC38A1), neutral amino acid transporter A (SLC1A4), 4F2 cell-surface antigen heavy chain (SLC3A2), sodium-and chloride-dependent taurine transporter (SLC6A6), volume-regulated anion channel subunit LRRC8E (LRRC8E), and neutral amino acid transporter B(0) (SLC1A5) ( Figure 5C). These results indicate that the intake of amino acids was decreased in CA9 knockdown cells. In addition, the phosphorylation level of mammalian target of rapamycin (mTOR)-S2448 was decreased and the mTOR pathway was inactivated in 786-O-CA9-KD cells, detected by Western blotting ( Figure 5D). Taking these results together, CA9 knockdown decreased cellular amino acid levels and protein synthesis, which contributed to suppressed cell proliferation in 786-O cells.
Furthermore, KEGG BlastKOALA analysis showed that the expression of most surface proteins involved in extracellular matrix (ECM)-receptor interaction and cell adhesion was decreased in CA9 knockdown cells, suggesting that CA9 silencing decreased cell migration ( Figure 5E). The differentially expressed membrane proteins were also analyzed by Gene Ontology with DAVID 6.8 [53]. Consistent with the KEGG analysis, the downregulated surface proteins were found to be mainly involved in cell motility, cell migration, cell adhesion, and extracellular matrix organization ( Figure 5F). These results were also consistent with proteomics showing that pathways associated with cell motility, including PAK signaling, paxillin signaling, integrin signaling, RAC and actin cytoskeleton signaling, were enriched in the DEPs between 786-O-CA9-KD cells and control cells. To confirm the surfaceomics results, we measured the migration rates of CA9 knockdown cells by a wound healing assay, and found that CA9 silencing significantly inhibited the migration of 786-O and 769-P cells ( Figure 5G and Figure S5). cytoskeleton signaling, were enriched in the DEPs between 786-O-CA9-KD cells and control cells. To confirm the surfaceomics results, we measured the migration rates of CA9 knockdown cells by a wound healing assay, and found that CA9 silencing significantly inhibited the migration of 786-O and 769-P cells ( Figure 5G and Figure S5).

Discussion
ccRCC displays the typical Warburg phenotype with enhanced glycolysis and decreased oxidative phosphorylation. Our previous studies demonstrated that proteostasis in ccRCC is highly correlated to this metabolic phenotype. For example, SIRT3 (NAD-dependent protein deacetylase

Discussion
ccRCC displays the typical Warburg phenotype with enhanced glycolysis and decreased oxidative phosphorylation. Our previous studies demonstrated that proteostasis in ccRCC is highly correlated to this metabolic phenotype. For example, SIRT3 (NAD-dependent protein deacetylase sirtuin-3, mitochondrial) expression is lower in cancer tissues compared to the pericarcinous tissues, and SIRT3 overexpression enhances mitochondrial functions while it inhibits cell growth [54]. As the major chaperonin for mitochondrial homeostasis, HSP60 (60 kDa heat shock protein, mitochondrial) knockdown switches the mitochondrial function from ATP production to biosynthesis to facilitate cell proliferation in ccRCC cells [32].
In contrast, CA9 was highly expressed in ccRCC and we proposed that CA9 silencing enhanced oxidative phosphorylation and decreased cell growth. This was confirmed by results presented herein in which CA9 silencing increased mitochondrial biogenesis, reversed the Warburg phenotype, and inhibited cell growth, whereas CA9 overexpression promoted cell proliferation.
Moreover, CA9 is a carbonic anhydrase, catalyzing the reversible hydration of carbon dioxide to maintain the intracellular pH (pHi) homeostasis. We investigated the effects of CA9 knockdown on pHi. We found that CA9 knockdown resulted in a reduction in pHi in 786-O and 769-P cells ( Figure S6), consistent with previous research [55]. It has been shown that CA9 enhances tumor invasion and migration by promoting extracellular matrix degradation through a pH-dependent mechanism [37]. In our study, we focused on the effects of CA9 on cellular metabolism in ccRCC. It has been found that an acidic pHi inhibits activities of glycolytic enzymes such as lactate dehydrogenase and phosphofructokinase 1 [56][57][58]. Although the effects of pHi on glycolytic enzymes need to be further investigated, our results suggest that metabolic changes we observed may be linked to a reduction of intracellular pH.
We carried out multi-omics analysis and revealed the global effects of CA9 silencing on cellular processes. We found that CA9 silencing activated oxidative phosphorylation, fatty acid oxidation, and mitochondrial biogenesis, as confirmed by fluorescent MitoTracker analysis. Consequently, ARG2 was upregulated as a mitochondrial protein, while ASS and ASL were downregulated, leading to putrescine accumulation which is toxic to ccRCC cells. This was validated by results showing that putrescine inhibited cell growth. However, the link between CA9 expression and ARG2 upregulation needs to be further explored. Putrescine, spermidine, and spermine are polyamines, which are vital to many cellular processes. ODC, catalyzing the decarboxylation of ornithine, is the first rate-limiting enzyme in polyamine biosynthesis and produces putrescine, which is a precursor of spermidine and spermine [50]. Polyamines can bind anionic complexes such as DNA, RNA, and proteins at physiological pH to modulate protein synthesis, gene expression, protection from oxidative damage, and maintenance of the structure of cellular macromolecules [59]. Polyamines play complex roles in pathological and physiological processes, such as angiogenesis and aging [59,60]. Increased polyamine can promote pancreatic ductal adenocarcinoma and colon cancer cell proliferation [61][62][63]. Conversely, excess polyamines and their by-products, such as aldehydes and ROS, are toxic to cells. Polyamines contribute to uremic toxins and liver damage [64,65]. Ochock et al., found that increased polyamines were toxic to ccRCC, and speculated that naturally occurring tumor hypoxia and constitutive HIF activity might prefer ccRCC cells with lower endogenous polyamine synthesis [52]. Thus, ccRCC cells are sensitive to putrescine accumulation. Our work revealed that CA9 knockdown increased arginine-dependent putrescine production to inhibit cell growth. Furthermore, the polycationic nature of putrescine may be changed after CA9 knockdown due to a reduction in pHi. However, the molecular mechanisms underlying polyamine toxicity remain elusive and need to be further explored.
Surfaceomics showed that CA9 silencing downregulated multiple amino acid transporters, resulting in the decreased cellular amino acids levels as confirmed by metabolomics analysis. It is worth mentioning that CA9 silencing also downregulated solute carrier family 4 member 4 (SLC4A4), a bicarbonate cotransporter, indicating that bicarbonate ions may regulate the expressions of these surface proteins.
Clear cell renal cell carcinoma is radioresistant, which is a challenge for its clinical treatment. Previous research found that inhibition of CA9 increases sensitivity of renal cell carcinoma to ionizing radiation [39]. The combined treatment of CA9 inhibitor and radiotherapy is a potential treatment option for ccRCC. Based on the results presented in this study in which CA9 knockdown downregulated programmed cell death 1 ligand 1 (PD-L1), we proposed that CA9-inhibition may enhance immunotherapy efficacy. Furthermore, CA9 inhibition increased mitochondrial biogenesis, which may enhance the efficacy of glutaminase inhibition for ccRCC treatment [32]. CA9 is a transmembrane protein and highly expressed in solid tumors. Zhu et al., found that CA9 aptamer-functionalized targeted nanobubbles can be applied to ultrasound molecular imaging of tumor cells, and can improve the accuracy of early diagnosis [66]. Minn et al., synthesized [ 64 Cu] XYIMSR-06, a dual-motif CA9 ligand, and found it can image ccRCC by positron emission tomography [67]. Therefore, molecular imaging reagents targeting CA9 can improve diagnostic accuracy.

Cell Culture
Human clear cell renal cell carcinoma cell lines 786-O and 769-P, and human kidney cell line 293T were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). 786-O and 769-P cells were cultured in 1640 medium (Wisent, Nanjing, China). 293T cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) medium (Wisent, Nanjing, China). The culture medium was supplemented with 1% penicillin/streptomycin (Wisent, Nanjing, China) and 10% fetal bovine serum (FBS) (Wisent, Nanjing, China). Cells were cultured in a 37 • C humidified incubator containing 5% CO 2 .

Establishment of Stable CA9-Knockdown and CA9-Overexpression Cell Lines
Two small hairpin RNA sequences of shRNA1 and shRNA2 were chosen for targeting CA9. A scrambled non-silencing shRNA without homologous sequences in the human genome was used as a negative control. The oligonucleotides were annealed and inserted into plasmid pLL3.7 to generate pLL3.7-CA9 vectors and a control vector. The vectors were transient transfected into 293T cells along with packing plasmids for lentivirus production using Lipofectamine™ 3000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Lentivirus particles were harvested after 60 h and then transfected 786-O and 769-P cells with polybrene when cells reached 60% confluence. A single GFP (green fluorescent protein) -positive cell was sorted by a flow cytometer into a single well of a 96-well plate to generate the monoclonal stable cell line. The clone with intense and uniform GFP expression was selected and confirmed the efficiency of CA9 knockdown by Western blotting. For the overexpression of CA9, human CA9 cDNA was amplified from 786-O cells and cloned into the plasmid pLVX-IRES-ZsGreen1 to create the pLVX-CA9-IRES-ZsGreen1 vector. The blank plasmid pLVX-IRES-ZsGreen1 was negative control (plvx). The transfection and cell line sorting procedures were same as the construction of knockdown cell lines. Detailed primer sequences are listed in Table S8.

Quantitative Real-Time PCR (qPCR)
Total RNA was isolated from cells with Trizol reagent (TIANGEN, Beijing, China). cDNA was synthesized from 2 µg total RNA using the reverse transcription kit (Promega, Madison, WI, USA) according to the manufacturer's instructions. Total DNA was isolated from cells using DNAzol ® (MRC, Cincinnati, OH, USA) according to the manufacturer's instructions. qPCR was performed with the Roche LightCycler 96 System (Roche, Basel, Switzerland) using SYBR green reaction mixture (CWBIO, Beijing, China) according to the manufacturer's instructions. ACTB was used as an internal control for mRNA quantification. mtDNA content was measured by the relative abundance of human cytochrome oxidase 1 (human mtCO1) and β-globin was used as an internal control [48]. Primers were acquired from the Primer-Blast [68] and are listed in Table S8.

Cell Proliferation Assay with Cell Counting Kit-8 (CCK-8)
Cells were seeded in 96-well plates with 2000 cells per well. Cell proliferation rate was detected with the CCK-8 (Dojindo, Kumamoto, Japan) according to the manufacturer's instructions. CCK-8 reagent was added into cells and cells were incubated in a cell incubator with 5% CO 2 for 1.5 h at 37 • C. Then, absorbance at 450 nm was measured by a microplate reader (Bio-Rad, Hercules, CA, USA). For detecting survival rates of 786-O cells treated with putrescine, cells were seeded into 96-well plates with 3000 cells per well. Three wells were included in each group. After seeding for 5 h, cells were treated with putrescine (0, 1.95, 7.81, 15.63, 31.25, 62.5, 125, and 250 mM) for 24 h. Then CCK-8 reagent was added into treated cells, followed by incubation for 1.5 h, then absorbance at 450 nm was measured.

Quantitative Proteomics Analysis
Proteins were extracted from cells with 8 M urea in phosphate buffer saline (PBS) (Wisent, Nanjing, China), 1X protease inhibitor cocktail. Then, cells were sonicated for 2 min and centrifugated with a collection of the supernatant. Protein concentrations were measured using the BCA protein assay kit. In-solution digestion was performed. A total of 100 µg of protein extracted from each cell line was reduced with 5 mM dithiothreitol (DTT) at room temperature and alkylated with 12.5 mM iodoacetamide (IAM) in the dark at room temperature. Then, proteins were diluted to 1.5 M urea with PBS and digested with trypsin (Promega, Madison, WI, USA) overnight at 37 • C. The tryptic peptides were desalted using Sep-Pak desalting columns (Waters, Milford, MA, USA) and then desalted peptides were labeled with tandem mass tags (TMT) 6-plex reagents (Thermo Fisher Scientific, Waltham, MA, USA). The combined TMT-labeled peptides were desalted by Sep-Pak columns and separated with a UPLC 3000 system (Thermo Fisher Scientific, Waltham, MA, USA) with an XBridgeTM BEH300 C18 column (Waters, Milford, MA, USA) at a flow rate of 1 mL/min. The mobile phase A was H 2 O (pH 10) and the mobile phase B was 98% acetonitrile (pH 10). Peptides were separated with a gradient elution including an increase from 8% to 18% phase B for 30 min; 18% to 32% phase B for 22 min. Forty-eight fractions were dried by speedvac and recombined to 12 fractions. The fractions were dissolved in 20 µL of 0.1% (v/v) formic acid (FA) and analyzed by LC-MS/MS.

Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) Analysis
Labeled peptides were separated with a high performance liquid chromatography (HPLC) system (Thermo Fisher Scientific, Waltham, MA, USA), which was connected to a Orbitrap Fusion LUMOS Tribrid mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). The mass spectrometer used the data-dependent acquisition mode with the Xcalibur 3.0 software (Thermo Fisher Scientific, Waltham, MA, USA). The parameters were as follows: a single full-scan mass spectrum was done in the Orbitrap (350-1550 m/z, 120,000 resolution) and the automatic gain control (AGC) target was 2 × 10 5 ; acquisition settings for MS/MS spectra were 30,000 for resolution with the AGC target of 2 × 10 5 and maximum injection time was 60 ms; the isolation window width was 0.7 Da; and the normalized collision energy for dissociation was 35%.

Peptide and Protein Identification
The UniProt human database (16 June 2017; 89,105 sequences) was used by the Sequest HT in Proteome Discoverer (PD) 2.1 software (Thermo Fisher Scientific, Waltham, MA, USA) to search the MS/MS spectra. The searching process used the following criteria: full tryptic specificity required, two missed cleavages tolerance, fixed modifications were carbamidomethylation and TMT 6-plex, variable modification was oxidation; 10 ppm tolerance for precursor ion mass for all MS acquired in the Orbitrap mass analyzer; and fragment ion mass tolerance was 20 mmu for all MS2 spectra acquired in the Orbitrap. The searched data were further processed with the percolator function in Proteome Discoverer to filter with 1% peptide false discovery rate (FDR). Relative protein quantification was carried out using PD 2.1 according to the intensities of reporter ions per peptide.

Detection of Intracellular pH
The pHi was measured using the Intracellular pH Detection Kit (BestBio, Shanghai, China) according to manufacturer's instructions. Briefly, cells were washed once with Hank's Balanced Salt Solution (HBSS) (BestBio, Shanghai, China) containing 20 mM HEPES (Solarbio, Beijing, China) and stained with 5 uM BBcellProbe in HBSS containing 20 mM HEPES for 45 min at 37 • C. After washing with HBSS containing 20 mM HEPES twice, fluorescence intensity of cells at 640 nm was analyzed by a BD FACS Aria II Flow Cytometer (BD Biosciences, Franklin Lakes, NJ, USA). A pH calibration standard curve was obtained using the nigericin as described in a previous study [69]. The pHi values were obtained based on the standard curve.

Metabolomics Analysis
Metabolomics analysis was performed as previously described [70]. Briefly, the cells were washed twice with cold PBS and incubated with pre-chilled 80% methanol (−80 • C) for 1 h at −80 • C. Then, the cells were scraped in 80% methanol on dry ice and centrifuged for 5 min. The extracted metabolites in the supernatant were concentrated completely to dryness with a lyophilizer and the protein concentrations of centrifuged pellets were used for normalization. The dried metabolites were dissolved in 80% methanol and analyzed by LC-MS/MS. We used the TSQ Quantiva™ Triple Quadrupole Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) with positive/negative ion switching for targeted metabolites quantitative analysis. The Q-Exactive Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) was chosen for untargeted metabolites profiling. Metabolites were identified based on the retention time and the accurate mass measured with <5 ppm mass accuracy. Quantitative information of metabolites was analyzed by TraceFinder.

Isotope Tracing Metabolomics
We used arginine-and lysine-free RPMI1640 medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) for metabolic labeling assay. The medium was supplemented with 10% dialyzed FBS (Biological Industries, Beit Haemek, Israel) and 1% penicillin/streptomycin. This medium also lacked lysine, which was added back to standard 1640 concentration (0.2 mM). For cell labeling, cells were cultured in 10 cm dishes with that medium supplemented with 1.1 mM 13 C 6 -arginine (Cambridge Isotope Laboratories, Tewksbury, MA, USA) for 12 h. In parallel with this, an unlabeled culture that was supplemented with equal concentration of unlabeled arginine was used to identify unlabeled metabolites. Then, the metabolites were extracted for mass spectroscopy analysis as described in Metabolomics analysis.

Wound Healing Assay
Cells were cultured in 6-well plates until 95% confluence. A sterile 200 µL pipette tip was used to scratch the cells to generate a uniform wound in the well. The well was replenished with 1640 medium without FBS [71]. Closures of these wounds were imaged at 0, 12, and 24 h with a Nikon microscope (Olympus Corporation, Tokyo, Japan). The wound area was quantitatively evaluated with ImageJ 1.52k software (National Institutes of Health, Bethesda, MD, USA).

Surfaceomics Analysis
To isolate cell surface proteins,~1 × 10 7 cells were washed three times with PBS and biotinylated in the culture plate with 0.25 mg/mL of sulfo-NHS-SS-Biotin (APExBIO, Houston, TX, USA) in PBS for 10 min at room temperature. The residual sulfo-NHS-SS-Biotin reagent was quenched with 10 mM lysine. Cells were lysed in RIPA lysis buffer supplemented with Protease Inhibitor Cocktail and 1% SDS (v/v). Then, cells were sonicated for 2 min and centrifugated with collection of the supernatant. Protein concentrations were measured using the BCA protein assay kit. Samples were diluted with PBS to 0.1% SDS followed by the isolation of biotinylated proteins with streptavidin conjugated sepharose (GE Healthcare, Chicago, IL, USA) for 4 h at 4 • C according to the manufacturer's instructions. The sepharose was washed five times and boiled with loading buffer (with DTT) (Solarbio, Beijing, China). The supernatant was separated on the SDS-PAGE followed by in-gel digestion and label-free analysis. Briefly, the gel was excised, reduced with 25 mM DTT at 55 • C and alkylated with 55 mM IAM in the dark at room temperature. In-gel digestion was carried out with trypsin overnight at 37 • C. The tryptic peptides were extracted twice using 50% acetonitrile aqueous solution with 0.1% FA. The extracts were dried by speedvac and dissolved in 20 µL of 0.1% FA for LC-MS/MS analysis. The MS/MS spectra were searched against the UniProt human database processed by the Sequest HT search engine in PD 2.2. The searching process used the following criteria: full tryptic specificity required, two missed cleavages tolerance, fixed modifications were carbamidomethylation and TMT 6-plex, variable modification was oxidation; 10 ppm tolerance for precursor ion mass for all MS acquired in the Orbitrap mass analyzer; and 20 mmu tolerance for fragment ion mass for all MS2 spectra acquired in the Orbitrap. The searched data were further filtered with the percolator function in the Proteome Discoverer software based on 1% FDR.

Statistical Methods
Data statistical analysis was carried out with GraphPad Prism 6.0 software (GraphPad, La Jolla, CA, USA). Significance was determined by the two-tailed Student's t-test. p-values below 0.05 were considered significant.

Conclusions
In summary, the present study demonstrated that CA9 knockdown increased mitochondrial biogenesis and putrescine production, and decreased the expression of surface proteins associated with amino acid transport and cell motility, leading to the reduced cell proliferation and migration in ccRCC. Our results shed light on CA9 functions in ccRCC progression, and reaffirm that CA9 is a potential therapeutic target for ccRCC.