ASS1 inhibits liver cancer by promoting CAD ubiquitination and reversing the urea cycle and pyrimidine synthesis imbalance

Abstract

Background:

The urea cycle and pyrimidine synthesis occur mainly in the liver and undergo opposite changes during hepatocarcinogenesis. Argininosuccinate synthase 1 (ASS1) and carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase (CAD) are key enzymes in the urea cycle and pyrimidine synthesis, respectively, and compete for the common substrate, aspartate. Moreover, ASS1 is lowly expressed in certain cancers, while CAD is highly expressed. However, the role of ASS1 and CAD in liver cancer still remains unclear.

Methods:

ASS1 and CAD expression in liver cancer were detected by tissue microarrays. Overexpression of ASS1 and CAD was achieved via lentivirus methods. All in vitro experiments were conducted in cells. The interactions of ASS1 and CAD were detected by co-immunoprecipitation (Co-IP) and GST-pull down. The in vivo study was conducted in a BALB/c nude mouse model. Intracellular metabolites were detected by LC–MS/MS.

Results:

ASS1 was lowly expressed in liver cancer, while CAD was highly expressed. In patients with recurrent liver cancer, ASS1 and CAD were significantly negatively correlated. Moreover, liver cancer patients with low ASS1 expression and high CAD expression had a poor prognosis. ASS1 and CAD interacted directly and promoted CAD ubiquitination through STUB1. In addition, Overexpression of CAD attenuated the tumor-suppressive effect of ASS1 in liver cancer cells. Pyrimidine supplementation enhanced the growth of liver cancer cells with ASS1 overexpression.

Conclusions:

ASS1 deficiency causes an imbalance in the urea cycle and pyrimidine synthesis in liver cancer. ASS1 directly controls the ubiquitination of CAD via STUB1, rather than just competing with aspartate, thereby suppressing liver cancer. Thus, ASS1 has potential as a druggable target in liver cancer.

INTRODUCTION

According to the International Agency for Research on Cancer (IARC), there were 20 million new cancer cases globally in 2022.¹ Among these, liver cancer was the sixth most commonly diagnosed cancer and the third leading cause of cancer death worldwide in 2022.¹ The development of liver cancer is strongly influenced by external environmental factors, such as hepatitis B/C viruses, alcohol, and aflatoxin.² In East Asia, such as China, Japan, and Korea, liver cancer is most prevalent and deadly,³ primarily because the hepatitis virus is pandemic in these areas.⁴,⁵ Although surgical approaches and percutaneous ablative therapies are recommended as treatment options for curing liver cancer, these options are not available for advanced liver cancer patients.⁶,⁷ Thus, more effective targets and drugs for liver cancer are needed.

Metabolic reprogramming is a hallmark of tumors, and ⁸ metabolic alterations in these pathways provide important nutrient sources for tumor cells. Liver is the most important organ⁹ and contains a number of specialized metabolic pathways, such as the urea cycle and pyrimidine synthesis.¹⁰,¹¹ ASS1 is the rate-limiting enzyme of the urea cycle, catalyzing the production of argininosuccinate from citrulline and aspartate. Aspartate is also a key amino acid in pyrimidine synthesis; it is catalyzed by carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase (CAD) to ultimately produce pyrimidine, which is essential for cell proliferation.¹² In the development of liver cancer, both the urea cycle and pyrimidine synthesis undergo metabolic reprogramming. While the genes essential for the urea cycle, including carbamoyl-phosphate synthetase 1 (CPS1), ornithine transcarbamylase (OTC), ASS1, and arginase 1 (ARG1), are considerably downregulated in liver cancer¹³,¹⁴,¹⁵,¹⁶—a mechanism used in tumor cells to obtain an adequate source of nitrogen and ammonia, especially under nutritionally deficient conditions¹⁷—the genes involved in pyrimidine synthesis, including CAD and uridine monophosphate synthase (UMPS), are significantly upregulated in liver cancer,¹⁸,¹⁹ which is consistent with increased pyrimidine levels in liver cancer.²⁰ These studies highlight the delicate balance between the urea cycle and pyrimidine synthesis.

ASS1 is a critical urea cycle enzyme.²¹ ASS1-deficient tumors are positively correlated with metastasis, poor prognosis, and adverse clinical outcomes in patients.²² The CpG island of ASS1 has been shown to be methylated in cancer, resulting in a reduction in ASS1 expression, and consistent with this, shorter survival rates have been reported in glioma patients with ASS1 methylation.²³ The Rabinovich research showed that loss of ASS1 in cancer increases cytoplasmic aspartate levels. Increased aspartate increases the phosphorylation of S6K1 through the mTOR pathway, thereby activating CAD and promoting the increase of pyrimidine.²⁴ This suggests that silencing ASS1 may control pyrimidine levels in tumor cells, and ASS1 may be essential to reversing the urea cycle and pyrimidine imbalance. Thus, focusing on ASS1 in the development of antitumor drugs may yield novel tumor drugs.

CAD is the initiating enzyme for pyrimidine synthesis and catalyzes the first 3 steps in pyrimidine synthesis.¹⁷ CAD is highly expressed in the liver, where pyrimidine synthesis is very active.¹² N-(phosphonacetyl)-L-aspartate (PALA) can effectively inhibit CAD activity and has a suitable antitumor effect at the cellular level.²⁵ However, the inefficacy and high toxicity of PALA in phase II trials resulted in the termination of the study.²⁶

Here, we determined that ASS1 is downregulated in liver cancer, while CAD is upregulated. In addition, liver cancer patients with low ASS1 expression and high CAD expression had a very poor prognosis. We further demonstrated that ASS1 and CAD interacted directly and promoted CAD ubiquitination. Thus, activating ASS1 is a potential strategy not only for reversing the imbalance in pyrimidine synthesis and the urea cycle but also for suppressing liver cancer.

METHODS

Cell culture

The 293T cell line and 5 liver cancer cell lines (HCCLM3, HepG2, Hep3B, SMMC7721, and MHCC97L) were utilized. These cell lines were procured from ATCC. The cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS, Vivacell) and 1% penicillin–streptomycin (Vivacell) and cultured in a humidified incubator at 37 °C with 5% CO₂. MG132 (10 mM, Cat. #: S2619, Selleck), BaF-A1 (250 nM, Cat. #: S1413, Selleck), and cycloheximide (CHX; 100 μg/mL, Cat. #: S7418, Selleck) were used in experiments.

Cell viability

For MTT assays, cells were plated into 96-well plates. Subsequently, 5 mg/mL of MTT solution was added to each well. Then, the medium was replaced with 100 μL of DMSO. The absorbance was measured by a Multiskan Reader (Thermo Fisher Scientific, USA) at 570 nm. For the colony assay, cells were seeded into 12-well plates. After fixing the cells with 4% paraformaldehyde, the plates were stained with 0.1% crystal violet overnight. The plates were dried, and then the colonies were photographed.

Real-time qPCR

Total RNA was extracted with FreeZol Reagent (Cat. #: R711-01, Vazyme), and the RNA was reverse transcribed into cDNA using the Reverse Transcription Kit (Cat. #: R212-01, Vazyme). HiScript II Q RT SuperMix (Cat. #: R222-01, Vazyme) was used in the qPCR reactions. GAPDH served as the reference gene. The primer sequences used are as follows: GAPDH (F: CGACCACTTTGTCAAGCTCA; R: GGGTCTTACTCCTTGGAGGC), ASS1 (F: GAGGATGCCTGAATTCTACA; R: GTTGGTCACCTTCACAGG), and CAD (F: AGTGGTGTTTCAAACCGGCAT; R: CAGAGGATAGGTGAGCACTAAGA).

Western blot analysis

Protein lysates were prepared from cells or tissues using RIPA buffer (Cat. #: V900854, VETEC) supplemented with 1 mM PMSF and protease inhibitors (Cat. #: S8830, Sigma-Aldrich). The protein concentration was quantified using a BCA assay kit (Cat. #: P0010, Beyotime), and the lysates were resolved by 10% SDS-PAGE. The bands were incubated successively for 5% milk, the primary antibody, and finally the secondary antibody. The antibodies used are as follows: ASS1 (1:1000/1:3000, Cat. #: 70720, Cell Signaling Technology), CAD (1:500, Cat. #: 93925, Cell Signaling Technology), GAPDH (1:5000, Cat. #: ABP50163, Abbkibe), HA (1:1000, Cat. #: AE036, Abclonal), Flag (1:1000, Cat. #: AE005, Abclonal), Ub (1:1000, Cat. #: 10201-2-AP, Proteintech), STUB1 (1:1000, Cat. #: 55430-1-AP, Proteintech), and Goat Anti-Rabbit IgG (1:10000, Cat. #: A21020, Abbkibe).

Immunoprecipitation–tandem mass spectrometry

The Flag-ASS1 was inserted into the pcDNA3.1 plasmid (Addgene). Flag-ASS1 plasmid was transfected into 293T cells for 48 hours. Total protein extraction was performed by immunoprecipitation (IP) buffer (Cat. #: P0013, Beyotime). Subsequently, 5 mg of the total protein was incubated with anti-Flag magnetic beads (Cat. #: L1303, Biolinkedin) with gentle inversion. The protein was colored with a silver-staining kit (Cat. #: P0017S, Beyotime). Based on the staining results, several specific bands were cut off and identified by immunoprecipitation–tandem mass spectrometry (IP–MS/MS) (Huada Gene Research Institute).

Endogenous immunoprecipitation

Total protein extraction was performed using IP lysis solution (P0013, Beyotime). Protein A/G beads (Cat. #: XE351456, Thermo Fisher Scientific) were conjugated with anti-ASS1 and anti-CAD antibodies through overnight incubation with rotation to facilitate binding. Subsequently, the conjugated antibodies were introduced into the protein lysate, followed by washing with IP lysis solution. The samples were then analyzed via western blotting.

Immunohistochemistry and hematoxylin and eosin staining

Following tissue procurement, the samples were immersed in 4% paraformaldehyde for fixation. Subsequently, they were processed into paraffin-embedded sections. The antibodies employed in this study were anti-ASS1 and anti-CAD. The staining outcomes were evaluated by ImageJ. A liver cancer tissue microarray (Cat. #: AF-LvcSur2201, Aifang Biological) was purchased from Aifang Biological. A) The study complies with both the Declaration of Helsinki and the Istanbul Declaration, B) This experiment was approved by the Life Sciences Ethics Committee of Changsha Yaxiang Biotechnology under the ethical approval number: Csyayj2024065, C). All liver cancer patients signed informed consent forms.

Lentiviral packaging and stable strain construction

293T cells were cultured and plated into 6-cm dishes. Upon reaching ~80% confluency, LVCV-CAD, psPAX2, and pMD2.G were transfected into the cells. After culturing for 48 hours, the viral supernatant was filtered through a 0.45-μM filter (Cat. #: BS-PES-45, Biosharp). Liver cancer cells were then incubated with the viral supernatants, and positive cells were selected with puromycin and validated by western blot. The stable cells that were successfully validated were used for subsequent experiments. The sequences of shRNA: ASS1 (shRNA#1: CCGGCTCAGGCTGAAGGAATATCATCTCGAGATGATATTCCTTCAGCCTGAGTTTTTG; shRNA#2: CCGGCCATCCTTTACCATGCTCATTCTCGAGAATGAGCATGGTAAAGGATGGTTTTTG); CAD (shRNA#1: CCGGCGAATCCAGAAGGAACGATTTCTCGAGAAATCGTTCCTTCTGGATTCGTTTTTG; shRNA#2: CCGGGCTCCGAAAGATGGGATATAACTCGAGTTATATCCCATCTTTCGGAGCTTTTTG); STUB1 (shRNA#1: CCGGCCCAAGTTCTGCTGTTGGACTCTCGAGAGTCCAACAGCAGAACTTGGGTTTTTG; shRNA#2: CCGGGAAGAGGAAGAAGCGAGACATCTCGAG ATGTCTCGCTTCTTCCTCTTCTTTTTG).

Immunofluorescence

HCCLM3 cells were plated in a 6-well dish for 24 hours. Cell-climbing slices were placed in a 6-well dish. Subsequently, we fixed with 4% paraformaldehyde for 20 minutes after washing with phosphate-buffered saline (PBS), 5% bovine serum albumin (BSA) was added for occlusion for 30 minutes. After washing by PBS, ASS1 antibody was added, overnight at 4 °C after washing by PBS containing 0.1% Tween-20 (PBST) for 3 times, TYR-570 staining solution was added, at room temperature for 1 hour after washing by PBST for 3 times, CAD antibody was added, overnight at 4 °C after washing by PBST for 3 times, ASS1 antibody was added, at room temperature for 1 hour after washing by PBST for 3 times, TYR-520 staining solution was added, at room temperature for 1 hour. After 3 washes with PBST, add TYR-520 staining solution at room temperature for 1 hour. After 3 washes with PBST, add DAPI to avoid light staining for 10 minutes. ASS1-TYR-570 (λ_ex550 nm/λ_em570 nm) and CAD-TYR-520 (λ_ex490 nm/λ_em520 nm) were detected by laser confocal microscopy (Eclipse Ti, Nikon).

Mouse xenograft tumor models

Six-week-old male BALB/c nu/nu mice (GemPharmatech Co., Ltd.) were housed in a specific pathogen-free environment.A) This study was performed and report these in accordance with the ARRIVE statement, B) The animal experiments in this study were approved by the Animal Ethics Committee of Hunan Normal University (ethical approval number: D20230053). Nude mice were randomly assigned into 4 groups (n=6 per group): (1) Negative control—NC (infected with blank lentivirus), (2) Overexpression (oe)-ASS1 (infected with ASS1 lentivirus), (3) oe-CAD (infected with CAD lentivirus), (4) oe-ASS1-CAD (infected with CAD and ASS1 lentiviruses successively). Each nude mouse was injected with 5×10⁶ HCCLM3 cells at the root of their thighs. When the xenograft grew to ~50 mm³, tumor volumes were measured every 7 days using calipers and calculated using the following equation: a×b ²/2, where “a” is the length and “b” is the width.

Ubiquitination assays

HCCLM3 was treated with 10 μM MG132 for 8 hours. The ubiquitination of CAD proteins was detected by western blot. The HA-Ub plasmid, Flag-CAD-pcDNA3.1, and Myc-ASS1-pcDNA3.1 were transfected into 293T cells for 48 hours. Total protein was extracted by IP lysis solution (Cat. #: P0013, Beyotime). Flag agarose magnetic beads (Cat. #: L1303, Biolinkedin) were added to the protein buffer and detected by western blot.

GST pulldown assays

The GST pulldown assays were modified from a previously published study.²⁷ ASS1 and CPSII were cloned into the PGX-4T-1 and PET-28a vectors, respectively. The GST-ASS1, GST-CPSII, His-ASS1, and His-CPSII plasmids were transfected into Escherichia coli BL21 competent cells (Cat. #: CD601-02, Transgen). Purified GST-ASS1/CPSII protein was immobilized on Magne GST Glutathione Particles (Promega). Then, purified His-ASS1/CPSII protein was added and incubated overnight at 4 °C with slow rotation. Lysis buffer was added to the magnetic beads and washed for 5 minutes. When the magnetic beads were fully adsorbed to the side of the centrifuge tube, the supernatant was removed. The above operation was repeated 3 times. The magnetic beads were resuspended with 1×Loadding Buffer and placed in boiling water for 10 minutes. Finally, the expression of ASS1 and CAD proteins was detected by western blot.

LC–MS/MS analysis of metabolites

Metabolite analysis by LC–MS/MS was performed as previously published with modifications.²⁸ HCCLM3 was harvested and washed. The metabolites in cells were extracted with 100 μL lysis buffer. Samples were centrifuged at 13,000 rpm at 4 °C for 10 minutes, and the supernatants were collected. The supernatant was lyophilized, and then the cellular powder was dissolved in formic acid. The metabolite content was detected using the AB SCIEX QTRAP 6500+LC–MS/MS system (USA), with Analyst (version 1.6.3, AB Sciex) and MultiQuant (version 3.0.2, AB Sciex) software. Firstly, we used standard samples of urea, aspartic acid, UMP, CMP, and TMP for mass spectrometry to find the peak position of each substance in the LC/MS measurement software. Subsequently, the samples were tested on the machine, and the area corresponding to the best peak was obtained as the content of the substance. The relative content of the substance is obtained by comparison with the NC group.

Statistics

All the data are expressed as the mean ± SD from at least 3 independent experiments. All cell experiments were at least 3 replicates, animal experiments were 5 per group, and metabolite assays were 6 replicates per group. Statistical significance testing is indicated in the figure legends. Statistical analysis was conducted with GraphPad Prism 9, and p<0.05 was considered statistically significant (*p<0.05; **p<0.01; ***p<0.001).

RESULTS

ASS1 protein level is inversely correlated with CAD in liver cancer

Under physiological conditions, the liver is the primary site for the synthesis of urea and pyrimidines. It plays a big role in energy supply and waste metabolism in the human body.²⁹ To further investigate the expression of ASS1 and CAD in liver cancer tissues, immunohistochemistry (IHC) staining was conducted on a liver cancer tissue microarray, which comprised 80 samples of adjacent normal and 164 samples of tumor tissues. In liver cancer tissues, the expression of ASS1 and CAD showed no significant correlation with patient age, gender, T stage, or N stage. However, the expression of CAD shows a significant positive correlation with the TNM staging of liver cancer patients. The detailed clinical information is in Table 1. The protein level of ASS1 was significantly reduced in liver cancer, whereas CAD exhibited elevated expression (Figures 1A, B). Consistently, in recurrent liver cancer, ASS1 expression was lower, while CAD expression was higher (Figures 1C, D). Moreover, the expression of ASS1 and CAD showed a significant negative correlation in recurrent liver cancer (Figure 1E) and 5 liver cancer cell lines (Figure 1F). As such, we were interested in whether the expression of ASS1 and CAD also differs in liver cancer prognosis. Indeed, liver cancer patients with low ASS1 expression (Figure 1G) or high CAD expression (Figure 1H) had a poor prognosis. Moreover, liver cancer patients with a low expression of ASS1 and high CAD expression had the worst prognosis (Figure 1I). We then obtained the mRNA expression levels of ASS1 and CAD in liver cancer from the Kaplan–Meier Plotter (https://www.kmplot.com/). We found that patients with low ASS1 expression had poor overall survival (OS) (Figure 1J), whereas those with low CAD expression had good OS (Figure 1K). Furthermore, we statistically analyzed the ratio of CAD to ASS1 in relation to liver cancer prognosis. Patients with a high ratio of CAD to ASS1 had poor OS (Figure 1L). Collectively, ASS1 and CAD are inversely correlated in liver cancer. Both low ASS1 expression and high CAD expression are risk factors for liver cancer, and the cumulative consequences of the above risk factors are even more serious.

TABLE 1.

ASS1 and CAD were associated with clinical information in liver cancer

	ASS1 scores			CAD scores
Clinical characteristics	Low	High	p	Low	High	p
Age
≤54	54	34	0.8730	45	43	>0.9999
>54	45	31		38	38
Genders
Male	68	51	0.5080	63	61	0.7255
Female	24	21		23	17
T stage
T1+T2	54	30	0.1747	50	34	0.5210
T3+T4	60	20		52	28
N stage
N0	51	56	0.8712	68	49	0.1236
N1+N2	28	29		21	26
TNM stage
I–II	59	41	0.7439	63	43	0.0217a
III–IV	40	24		23	35

^a The statistical method for this experiment was the Fisher exact test, p<0.05.

Abbreviations: ASS1, argininosuccinate synthase 1; CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase.

FIGURE 1.

ASS1 and CAD expression in liver cancer. (A, B) A tissue microarray with liver adjacent normal tissue (n=80) and tumor tissue (n=164) was stained by IHC using anti-ASS1 and anti-CAD antibodies; scale bar=125 µm. The p values are based on a Student t test; ***p<0.001. (C, D) ASS1 and CAD expression in recurrent (n=87) and non-recurrent (n=77) liver cancer tissue. (E) Correlation between ASS1 and CAD in recurrent liver cancer. The p value is based on a simple linear regression; p=0.0036. (F) Detection of ASS1 and CAD expression by western blot in HCCLM3, MHCC97L, Hep3B, SMMC7721, and HepG2 cell lines. The grey scale statistics and quantitative analysis using ImageJ. The p value is based on a simple linear regression; p=0.0482. (G) Statistical association of ASS1 with OS in liver cancer. (H) Statistical association between CAD and OS through liver cancer. The date from the liver cancer tissue microarray. (I) Combined analysis of ASS1 and CAD on the prognosis in liver cancer (ASS1 low expression and CAD low expression: 54, ASS1 low expression and CAD high expression: 58; ASS1 high expression and CAD high expression: 35; ASS1 high expression and CAD low expression: 17). The p values are based on a Log-rank (Mantel–Cox) test. Comparing with the ASS1 high-expression group and CAD low-expression group, *p<0.05 and **p<0.01; comparing with the ASS1 low-expression group and CAD low-expression group, ##p<0.01. (J) Statistical association of ASS1 with OS in liver cancer patients (n=790). Prognostic information of liver cancer patients was obtained from a Kaplan–Meier Plotter. (K) Statistical association of CAD with OS in liver cancer patients (n=364). (L) Statistical association between CAD/ASS1 and OS (n=364). Abbreviations: ASS1, argininosuccinate synthase 1; CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase; IHC, immunohistochemistry; OS, overall survival.

CAD interacts directly with ASS1

In clinical outcomes, there was a significant negative correlation between ASS1 and CAD in liver cancer. Thus, we hypothesized that they might regulate each other. To further explore the relationship between ASS1 and CAD, we assessed whether CAD was a protein that interacts with ASS1. We transfected the Flag-ASS1 plasmid into 293T cells for 48 hours. From the silver-staining image (Figure 2A), we found several unique bands, one of which has a molecular weight bigger than 180 kDa (red arrows). Interestingly, 1 specific band was identified as CAD protein by LC–MS (Figure 2B). To validate the above results, Flag-CAD or HA-ASS1 plasmid was transfected into 293T and HCCLM3 cells, and we found that ASS1 and CAD interacted in both 293T and HCCLM3 cells (Figures 2C, D). In addition, we also performed Co-IP experiments directly with anti-ASS1 and anti-CAD in 293T (Figure 2E) and HCCLM3 (Figure 2F) cells, and the results were consistent with the above results. Next, confocal microscopy (Figure 2G) showed colocalization of ASS1 (red) and CAD (green) in HCCLM3 cells. To further characterize the specific structural domains involved in the interaction between CAD and ASS1, we truncated ASS1 into 2 segments (ASS1-C and ASS1-N) and CAD into 4 segments [glutamine amidotransferase (GATase), carbamoyl-phosphate synthetase II (CPSII), dihydroorotase (DHOase), and aspartate transcarbamoylase (ATCase)].³⁰,³¹ Then, we labeled the ASS1 segments with an HA-tag and the CAD segments with a Flag-tag and performed a Co-IP. Western blot analysis verified that the CPSII domain of CAD interacts with the ASS1-C domain of ASS1 (Figures 2H, I). Then, GST-ASS1 and GST-CPSII were cloned into the prokaryotic expression vector pET-28a for recombinant protein expression. In line with our previous results, we found that the CPSII domain of CAD interacts with ASS1 (Figure 2J). Taken together, our results show that ASS1 directly interacts with CAD.

FIGURE 2.

Interaction between ASS1 and CAD proteins. (A) Detection of specific protein bands interacting with ASS1 by silver stain. Flag-ASS1 plasmid was transfected into 293T cells for 48 hours. Red arrows point to specific bends. (B) The specific bands identified by silver staining (about 200, 100, and 55 kDa) were cut out and identified by LC–MS/MS. (C) Validation of HA-ASS1 and CAD protein interactions in 293T and HCCLM3 cells by Co-IP and western blot. Anti-HA agarose magnetic beads were used for Co-IP. (D) Validation of Flag-CAD and ASS1 protein interactions in 293T and HCCLM3 cells by Co-IP and western blot. Anti-Flag agarose magnetic beads were used for Co-IP. (E, F) Validation of ASS1 and CAD protein interactions in 293T and HCCLM3 by Co-IP and western blot. Anti-ASS1 and anti-CAD were used. (G) The localization of ASS1 and CAD was confirmed by confocal laser microscopy; scale bar=10 µm; red: ASS1, green: CAD, and blue: DAPI. (H) ASS1 and its segments (ASS1-C, ASS1-N) pulled down the CAD protein by Co-IP. CAD, ASS1, and its segments were detected in whole cell lysis (WCL) and IP buffer (HA) by western blot. (I) CAD and its segments (GATase, CPSII, DHOase, ATCase) pulled down ASS1 protein by Co-IP. (J) The interaction between ASS1 and GST-CPSII or GST-ASS1 and CPSII. GST was used as a negative control. Abbreviations: ASS1, argininosuccinate synthase 1; CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase; Co-IP, co-immunoprecipitation; GST, glutathione S-transferase; LC–MS/MS, liquid chromatography–tandem mass spectrometry.

ASS1 promotes ubiquitination of CAD proteins

The previous results showed that ASS1 can directly interact with CAD proteins. In order to determine the regulatory circuit between ASS1 and CAD, we first knocked down and overexpressed ASS1 in liver cancer cells. The results showed that CAD protein levels exhibited an opposite trend with changes in ASS1 expression, but the mRNA levels were unchanged (Figures 3A, C and Supplemental Figure S1, http://links.lww.com/HC9/C58). Furthermore, changes in CAD expression had no effect on ASS1 protein levels (Supplemental Figure S2, http://links.lww.com/HC9/C58). These results indicated that ASS1 serves as an upstream regulatory protein for CAD. To further investigate the effect of ASS1 on the stability of CAD proteins, we constructed HCCLM3 and HepG2 cell lines that stably overexpressed ASS1 using lentiviral infection. After treatment with CHX (100 μg/mL), protein samples at various time points (0, 3, 6, 12, and 24 h) were detected by western blot. The half-life of CAD protein became shorter after overexpressing ASS1 in HCCLM3 (Figure 3B) and HepG2 cells (Figure 3D). The main pathways that affect protein stability are the lysosomal and ubiquitination pathways.³²,³³ Thus, we treated HCCLM3 cells overexpressing ASS1 with 250 nM Baf-A1 to block lysosomal pathway. The results showed that Baf-A1 did not inhibit the reduction of CAD proteins caused by overexpressing ASS1 (Supplemental Figure S3, http://links.lww.com/HC9/C58). However, after HCCLM3 and HepG2 cells stably overexpressing ASS1 were treated with the ubiquitination inhibitor MG132 (10 μM) for 8 hours, the levels of CAD protein were only slightly decreased (Figure 3E). We also examined the ubiquitination level of CAD proteins in liver cancer cells treated with MG132. The results showed that overexpressing ASS1 promoted ubiquitination of CAD proteins in HCCLM3 cells and HepG2 cells (Figure 3F). In addition, Flag-CAD, Myc-ASS1, and HA-Ub proteins were expressed stably in cells, and we found that Flag-CAD was able to pull down more HA-Ub after overexpressing ASS1 (Figure 3G). Taken together, overexpressing ASS1 can promote the ubiquitination of CAD proteins.

FIGURE 3.

ASS1 mediates the ubiquitination of CAD. (A, C) Detection of ASS1 and CAD protein in HCCLM3 and HepG2 cells after overexpressing ASS1 through western blot and qPCR. (B, D) Detection of ASS1 and CAD protein in HCCLM3 and HepG2 cells through western blot. Calculating the half-life of CAD proteins. HCCLM3 and HepG2 cells stably overexpressing ASS1 through lentiviral. HCCLM3 and HepG2 were harvested at 0, 3, 6, 12, and 24 hours after being treated with 100 μg/mL CHX. The bands were analyzed in grayscale with ImageJ. The p values are based on a Student t test; **p<0.01, and ***p<0.001. (E) Detection of ASS1 and CAD protein in HCCLM3 and HepG2 cells via western blot. HCCLM3 and HepG2 stably overexpressing ASS1 were treated with ubiquitination inhibitor MG132 (10 μM) for 8 hours. (F) Detection of the ubiquitination of the CAD protein by western blot. CAD and ubiquitin antibodies were used. (G) The expression of HA-Ub bound to CAD was detected by western blot. Anti-Flag agarose magnetic beads were used for Co-IP. Cells were treated with MG132 for 8 hours. Abbreviations: ASS1, argininosuccinate synthase 1; CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase; CHX, cycloheximide; Co-IP, co-immunoprecipitation.

STUB1 is essential for ASS1 to promote ubiquitination of CAD

To further explore the cause of CAD ubiquitination, we transfected the Flag-CAD plasmid into 293T cells for 48 hours and found several potential interacting proteins via silver staining (Figure 4A). These specific bands were cut out and characterized by LC–MS. One potential CAD-interacting protein we found was STUB1 (Figure 4B), which is an E3 ubiquitin–protein ligase. It has been reported that STUB1 binds to CAD and promotes its ubiquitinated degradation. Based on these results, we speculate that STUB1 may play a key role in ASS1-induced ubiquitination degradation of CAD proteins.³⁴ Then, we verified the interactions between STUB1 and CAD by Co-IP experiments. The result showed that STUB1 and CAD interacted in 293T (Figure 4C). In addition, we also transfected the Myc-STUB1 or HA-ASS1 plasmid into 293T cells. However, ASS1 and STUB1 did not interact (Figure 4D).

FIGURE 4.

The role of STUB1 in the ASS1-induced ubiquitination of CAD protein. (A) Detection of specific protein bands interacting with CAD by silver stain. (B) The specific bands identified by silver staining (about 100, 70, 55, and 36 kDa) were identified via LC–MS/MS. (C) Validation of Myc-ASS1 and Flag-CAD protein interactions in HCCLM3 cells by Co-IP and western blot. (D) Validation of Myc-STUB1 and HA-ASS1 protein interactions in HCCLM3 cells by Co-IP and western blot. (E) Detection of STUB1 and CAD protein interactions after overexpressing ASS1 by Co-IP. (F) STUB1 was knocked down in 293T cells through sh-SUTUB1#1 and sh-STUB1#2, and the ubiquitination of CAD protein was assessed by western blot. ASS1, CAD, GAPDH, STUB1, and ubiquitin antibody was used. Abbreviations: ASS1, argininosuccinate synthase 1; CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase; Co-IP, co-immunoprecipitation; LC–MS/MS, liquid chromatography–tandem mass spectrometry.

To further determine the role of STUB1 in the ubiquitination of CAD proteins by ASS1, we quantified the STUB1 proteins dragged by CAD after overexpressing ASS1. The result showed that CAD protein could pull down more STUB1 proteins after overexpressing ASS1 (Figure 4E). Then, we knocked down STUB1 expression in 293T cells through shRNA. At this point, we transfected the HA-ASS1 plasmid into 293T cells. The result showed that the ubiquitination of CAD protein caused by overexpressing ASS1 was reduced after knocking down STUB1 (Figure 4F).

Summarizing the above results, the ubiquitination-related protein STUB1 was identified by silver staining and LC–MS. After a series of Co-IP experiments, we demonstrated that CAD could interact with the STUB1 protein. Meanwhile, ASS1 could promote the interaction of CAD and STUB1 proteins. Taken together, STUB1 is a key interacting protein with CAD protein and is essential for ASS1 to promote ubiquitination of CAD.

ASS1 suppresses liver cancer growth by inhibiting CAD

To investigate the function of ASS1 and CAD, we modulated the expression of ASS1 and CAD in HCCLM3 and HepG2 cells. Overexpressing ASS1 inhibited the proliferation of liver cancer cells (Figures 5A, B), while overexpressing CAD had the opposite effect (Figures 5C, D). ASS1 knockdown accelerated liver cancer cell proliferation, while knocking down CAD did the opposite (Supplemental Figure S4, http://links.lww.com/HC9/C58). Next, we constructed stable liver cancer cell lines overexpressing ASS1 or CAD. For liver cancer cells with overexpressed CAD, the inhibitory effect of ASS1 was reduced (Figures 5E, F and Supplemental Figure S5, http://links.lww.com/HC9/C58). In mouse experiments, we prepared 5 female nude mice per group and injected 5×10⁶ HCCLM3 cells into the inner thighs of the mice. Animal experiments are performed according to the process in Figure 5G. The xenograft tumors were arranged and photographed (Figure 5H) and then weighed (Figure 5I). Based on the measured data, the volume of the xenograft tumor was calculated (Figure 5J). The weight and size of tumors in xenografts became smaller in the oe-ASS1 group, but larger in the oe-CAD group. However, the weight and size of tumors in xenografts in the oe-ASS1-CAD group were minimally reduced compared to the oe-CAD group. Then, we randomly selected three xenograft tumors for the western blot. The results confirmed that ASS1 and CAD were still overexpressed in xenograft tumors (Figure 5K). Taken together, ASS1 is a tumor suppressor and CAD is an oncogene in liver cancer. ASS1 inhibits liver cancer cell proliferation by inhibiting CAD.

FIGURE 5.

ASS1 regulates liver cancer cell proliferation through inhibiting CAD. (A) The proliferation of HCCLM3 and HepG2 overexpressing ASS1 was detected by MTT at days 1, 3, and 5. (B) The clone-forming ability of HCCLM3 and HepG2 cells overexpressing ASS1 was assessed through a 10-day clone formation assay. The clone numbers of HCCLM3 and HepG2 cells overexpressing ASS1 were counted by ImageJ. (C) The proliferation of HCCLM3 and HepG2 overexpressing CAD was detected by MTT at days 1, 3, and 5. (D) The clone-forming ability of HCCLM3 and HepG2 cells overexpressing CAD was assessed through a 10-day clone formation assay. The clone numbers of HCCLM3 and HepG2 cells overexpressing CAD were counted by ImageJ. The p values are based on a Student t test; **p<0.01 and ***p<0.001. (E) The proliferation of HCCLM3 and HepG2 was detected by MTT at days 1, 3, and 5. The 4 groups in this experiment were: (1) oe-NC (infected with blank lentivirus); (2) oe-ASS1 (infected with ASS1 lentivirus); (3) oe-CAD (infected with CAD lentivirus); and (4) oe-ASS1-CAD (infected with CAD and ASS1 lentiviruses successively). (F) The clone-forming ability of HCCLM3 and HepG2 cells was assessed via a 10-day clone formation assay. The number of HCCLM3 and HepG2 clones overexpressing ASS1 was counted by ImageJ. The p values are based on a Student t test as compared to the oe-NC group, **p<0.01 and ***p<0.001; compared to the oe-CAD group, ##p<0.01. (G) The growth process of transplanted tumors in nude mice (n=5). (H) Photographs of xenograft tumors in nude mice. (I) Growth curves of xenograft tumors in nude mice. Tumors were measured every 7 days for 7 weeks. (J) The xenograft tumors were weighed by scales—5 nude mice per group. The p values are based on a Student t test as compared to the oe-NC group; **p<0.01 and ***p<0.001; compared to the oe-CAD group, ##p<0.01. (K) Detection of ASS1 and CAD expression in xenograft tumors by western blot. Abbreviations: ASS1, argininosuccinate synthase 1; CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase; NC, negative control, oe, overexpression.

ASS1 inhibits liver cancer by suppressing CAD-mediated pyrimidine synthesis

Considering that ASS1 and CAD share the substrate aspartate,¹² we hypothesized that ASS1 inhibits liver cancer growth by regulating aspartate flow. We detected the contents of primidine and urea in 5 liver cancer cells by LC–MS. In liver cancer cells, pyrimidine content decreased with the increase of ASS1 expression, while urea content decreased with the increase of CAD expression (Figure 6A). After ASS1 was overexpressed in HCCLM3 and HepG2, the cellular levels of aspartate and pyrimidines became significantly lower, while the levels of urea were significantly increased (Figure 6B). The converse was true when ASS1 was knocked down (Supplemental Figures S6, http://links.lww.com/HC9/C58). However, the regulation of pyrimidine and urea by ASS1 was reduced with CAD overexpression (Figure 6C).

FIGURE 6.

ASS1 inhibits CAD-mediated pyrimidine synthesis. (A) Detection of pyrimidine (UMP, uridine 5′-monophosphate; CMP, cytidine 5′-monophosphate; TMP, thymidine 5′-monophosphate) and urea in liver cancer (HCCLM3, MHCC97L, Hep3B, SMMC7721, and HepG2) by LC–MS. (B) Detection of aspartate, pyrimidine, and urea in HCCLM3 and HepG2 cells with ASS1 overexpression by LC–MS. (C) Detection of aspartate and pyrimidine in HCCLM3 and HepG2 cells by LC–MS. Liver cancer cells were infected with oe-ASS1 and oe-CAD lentivirus. There were 6 repetitions for each group. The p values are based on a Student t test. Compared to the oe-NC group, **p<0.01; compared to the oe-CAD group, #p<0.05 and ##p<0.01. (D) The proliferation of HCCLM3 and HepG2 cells overexpressing ASS1 was detected by MTT. Pyrimidine (UMP:CMP:TMP=1:1:1) concentrations used were 0, 0.25, 0.5, 1, 2, 4, and 8 μg/mL. (E) The proliferation of HCCLM3 and HepG2 overexpressing ASS1 was detected by MTT. Liver cells continued to be cultured for 24 hours in aspartate-deficient medium (ASP−). Aspartate concentrations were 0, 0.25, 0.5, 1, 2, 4, and 8 μg/mL. (F) The proliferation of HCCLM3 and HepG2 cells with ASS1 knocked down was detected by MTT. Liver cells were cultured for 24 hours in ASP-medium. The p values are based on a Student t test; **p<0.01 and ***p<0.001. Abbreviations: ASS1, argininosuccinate synthase 1; CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase; LC–MS, liquid chromatography–mass spectrometry; NC, negative control, oe, overexpression.

To further determine that ASS1 inhibits liver cancer by suppressing CAD-mediated pyrimidine synthesis, we selected to use aspartate-deficient (ASP-medium) for the aspartate rescue experiment. Aspartate supplementation restored the proliferation of HCCLM3 and HepG2 cells cultured with ASP-medium (Supplemental Figures S7A, B, http://links.lww.com/HC9/C58). The effect plateaued at 4 μg/mL of aspartate in HCCLM3 cells and at 6 μg/mL of aspartate in HepG2 cells. Moreover, supplementation with pyrimidine promoted proliferation only in liver cancer cells cultured in ASP-medium (Supplemental Figures S7C, D, http://links.lww.com/HC9/C58). The above results indicate that sufficient aspartate is converted to pyrimidines for cell proliferation. Pyrimidine and aspartate were then used to rescue the proliferation of cells with overexpressed ASS1. Pyrimidine facilitated liver cancer cell growth overexpressing ASS1 in a concentration-dependent manner (Figure 6D). Aspartate had a minimal salvage effect on liver cancer cells overexpressing ASS1 (Figure 6E). However, aspartate facilitated liver cancer cell growth with ASS1 knocked down (Figure 6F). Taken together, ASS1 inhibits liver cancer by suppressing CAD-mediated pyrimidine synthesis.

DISCUSSION

The imbalances in urea and pyrimidine synthesis had been identified early in tumors.¹⁸,³⁵ The urea cycle is blocked in most tumors to secure sufficient nitrogen for cell growth.³⁶ The main provider of NH3 is glutamine.³⁷ In breast cancer cells, ammonia can be effectively integrated into several amino acids, such as proline, aspartic acid, glutamate, and alanine, thereby promoting tumor growth.³⁸ In order to reduce toxicity, ammonia is rapidly condensed with bicarbonate ions to form carbamyl phosphate, catalyzed by CPSI.³⁹ CPSI is a mitochondrial protein that is homologous to the CPSII domain of CAD protein.⁴⁰ CPSII is associated with initiation specificity in pyrimidine synthesis, directly using glutamine as a nitrogen donor to synthesize carbamyl phosphate.⁴⁰ Generally, the mitochondrial membrane blocks carbamyl phosphate from passing unless its concentration is excessively high. The spilled carbamyl phosphate can be used by CAD to synthesize pyrimidine.⁴¹ This suggests that the urea cycle initiation enzyme CPSI may play an important role in the maintenance of pyrimidine pools. Moreover, aspartate is a common substrate of ASS1 and CAD, and its flow direction affects cell growth. These investigations have revealed that the urea cycle and pyrimidine synthesis are closely interconnected.

In our study, there was a strong link between ASS1 and CAD. Tissue microarray results showed that ASS1 is lowly expressed in liver cancer, while CAD is highly expressed. In recurrent patients, ASS1 and CAD were significantly negatively correlated. Moreover, liver cancer patients with low ASS1 and high CAD expression had a very poor prognosis. In liver cancer tissues and cell lines, ASS1 was positively correlated with nitrogen metabolism and negatively correlated with pyrimidine metabolism, whereas CAD showed opposite results. The above results suggested that there is mutual regulation between the ASS1-dominated urea cycle and CAD-mediated pyrimidine synthesis.

Further, we found that ASS1 and CAD interact directly. Then, we altered the expression of ASS1 and CAD in vivo. The results showed that ASS1 promoted CAD protein degradation, but CAD did not promote ASS1 degradation. This suggests that ASS1 acts as an upstream protein to directly regulate CAD protein degradation. However, ASS1, as a protease, did not have the ability to degrade proteins on its own. We identified that CAD interacts with STUB1, which is a co-chaperone protein and functional E3 ubiquitin ligase, containing a tetratricopeptide repeat and U-box.³⁴,⁴² In the field of oncology, STUB1 enhanced the chemotherapy sensitivity of glioma by promoting the ubiquitination of LRPPRC.⁴³ It has also been reported that STUB1 can promote the ubiquitination of BCAT1 to inhibit the progression of glioma.⁴⁴ STUB1 was essential for ASS1 to promote ubiquitination of CAD. However, ASS1 did not interact with STUB1, indicating that ASS1 does not bridge between STUB1 and CAD by itself. Since ASS1 and CAD share a common substrate, we hypothesized that ASS1 interacts with CAD to steal aspartate, which leads to an excess of ubiquitinated CAD protein by STUB1. We will study this further in future studies.

Under physiological conditions, the urea cycle and pyrimidine synthesis take place primarily in the liver.⁹ ASS1 inhibited the proliferation of liver cancer by suppressing CAD both in vitro and in vivo. Cell proliferation was consistent with modifications in the intracellular pyrimidine content. Moreover, the rescue effect of pyrimidine supplementation was much better than that of aspartate for the proliferation of liver cancer cells with overexpressed ASS1, indicating that ASS1 regulates pyrimidine synthesis in liver cancer cells by controlling CAD, rather than just competing with aspartate.

In summary, this study found that an imbalance in the urea cycle and pyrimidine synthesis was prevalent in liver cancer. ASS1, a key enzyme of the urea cycle, played a crucial role in this imbalance. ASS1 directly promoted the ubiquitination of CAD via STUB1, rather than just competing with aspartate, thereby reversing the imbalance in the urea cycle and pyrimidine synthesis. Thus, ASS1 is a potential druggable target in liver cancer.

Supplementary Material

hc9-9-e0769-s001.docx ^{(1.1MB, docx)}

DATA AVAILABILITY STATEMENT

Data that has been analyzed and published in this research is available upon request to the corresponding author.

AUTHOR CONTRIBUTIONS

Conceptualization: Zhengnan Ming and Zhiyong Luo. Data curation: Jijia Li. Formal analysis: Zizheng Zou. Funding acquisition: Suyou Liu and Zhiyong Luo. Investigation: Xiyuan Hu. Methodology: Zhengnan Ming. Project administration: Mingquan Liu. Resources: Ling Chen. Software: Tiao Luo. Supervision: Junli Luo and Dayou Ma. Validation: Zhengnan Ming and Wensong Luo. Visualization: Jiang Zhou and Xiaohe Liu. Writing—original draft: Zhengnan Ming. Writing—review and editing: Tiao Luo.

FUNDING INFORMATION

This work was financially supported by the grants from the Natural Science Foundation of Hunan Province (2025JJ30044), the National Natural Science Foundation of China (81973710, 82271907, 82272720), the Central South University Research Programme of Advanced Interdisciplinary Studies (2023QYJC032), and the postgraduate innovation project of Central South University of China (2023ZZTS0025).

CONFLICTS OF INTEREST

The authors have no conflicts to report.