Lysine-222 succinylation reduces lysosomal degradation of lactate dehydrogenase a and is increased in gastric cancer

The tumor samples in this study were taken from seven patients with primary gastric adenocarcinoma who had not received chemotherapy or any other treatment before surgery and were surgically removed with the patient’s informed consent. The patients were from the Second Affiliated Hospital of Nanjing Medical University (Jiangsu, China). This study was approved by the Hospital Ethics Committee of the Second Affiliated Hospital of Nanjing Medical University and carried out in accordance with the principles of the Declaration of Helsinki.

The corresponding tumor and adjacent normal tissues obtained from the patients were used to quantify dynamic changes of lysine succinylome in PTM Biolabs (Hangzhou, China). The experiment included TMT labeling, HPLC fractionation, affinity enrichment, and mass spectrometry-based quantitative proteomics. Intensive bioinformatic analysis was then carried out to annotate the quantifiable targets, including protein annotation, functional classification, functional enrichment, functional enrichment-based cluster analysis, etc.

Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and transcribed into cDNA using the reverse transcription kit (Takara Biotechnology, Beijing, China) following the manufacturers’ instructions. RNA was analyzed using real-time qPCR with SYBR Green PCR Master mix (Roche Applied Science, Mannheim, Germany) on a StepOnePlus™ real-time PCR System (Applied Biosystems, Foster City, CA, USA). The relative gene expression was normalized to GAPDH. Specific primer sets used for this assay included LDHA (forward: GGA TCT CCA ACA TGG CAG CCT T, reverse: AGA CGG CTT TCT CCC TCT TGC T), and GAPDH (forward: TTG CCA TCA ATG ACC CCT TCA, reverse: CGC CCC ACT TGA TTT TGG A).

The tissues were fixed in 10% formalin, and then embedded in paraffin. The specific steps were performed as previously described [13]. Rabbit polyclonal antibodies against the human LDHA K222-succinylated peptide, PDLGTDK (suc) DKEQWK, were generated from two rabbits (#1 and #2) at ChinaPeptides Co., Ltd. (Shanghai, China). For immunohistochemical staining, 5-μm-thick serial sections were used to prepare the slides. Antigen retrieval was performed with 10 mM citrate antigen retrieval solution (CW Biotech, Beijing, China). Anti-LDHA (#3582S, Cell Signaling Technology, Danvers, MA, USA) and anti-LDHA-K222suc antibodies was used at a dilution of 1:400. The immunostaining index was based on the staining intensity and percentages of positively stained tumor cells. The intensity was scored between 0 and 3: 0 (no staining), 1 (weak staining), 2 (moderate staining), and 3 (strong staining). The percentages of positively stained tumor cells were defined as 0 (< 10%), 1 (10 –30%), 2 (31 –50%), and 3(> 50% of positive cells) [25]. The staining index was then calculated as the score of staining intensity multiplied by the percentage of positively stained tumor cells. Tumor stage was classified according to the seventh edition of the UICC/AJCC TNM staging system.

AGS, HGC27, B16-F10, and 293 T cells were purchased from Shanghai Cafa Biological Technology Co. Ltd. (Shanghai, China), tested negative for mycoplasma, and authenticated by Genetic Testing Biotechnology Corporation (Suzhou, China) using short tandem repeat markers. The cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, Grand Island, NY, USA) and 1% penicillin/streptomycin (Thermo Scientific, Waltham, MA, USA), and maintained in a humidified incubator at 37 °C with 5% CO₂. The following reagents were used to treat the cells: MG132 (10 mM, #S2619, Selleck, Shanghai, China), cycloheximide (CHX, 10 μg/ml, #HY-12320, Selleck, Shanghai, China), Bafilomycin A1 (Baf-A1, 20 nM, #S1413, Selleck, Shanghai, China) and chloroquine (50 μM, #C6628, Sigma, St. Louis, MO, USA).

The stable cell lines were generated using lentivirus system. Briefly, the genes were cloned into the specific vector pLJM1, and then transfected into HEK293T cells using Lipofectamine 3000 reagent. Lentiviral supernatants were harvested from HEK293T cells and mixed together with 8 μg/mL of polybrene to increase the infection efficiency. The infected cancer cells were then selected in culture media containing 2 μg/ml of puromycin for 2 weeks.

AGS cells (5 × 10⁵) were seeded and cultured in 12-well plates in the presence or absence of BafA1 for 24 h. Then, the cells were fixed with 4% paraformaldehyde for 15 min, washed three times with PBS for 5 min, and permeabilized with 0.1% Triton x-100 for 10 min. After blocking for 1 h in 3% bovine serum albumin (BSA) at 37 °C, the cells were incubated overnight at 4 °C with primary antibodies anti-LDHA (#3582S, Cell Signaling Technology, Danvers, MA, USA) and anti-SQSTM1/p62 (#88588S, Cell Signaling Technology). Subsequently, the secondary antibodies were added for 1 h at 37 °C followed by counterstaining with DAPI. The cells were then observed and photographed under the confocal microscope (Olympus FV-1000).

CPT1A shRNA and control shRNA plasmids were purchased from Shanghai Genechem Co., Ltd. (Shanghai, China) and used as before [13]. The shRNA sequences for CPT1A are as follows: (#1) TAG CCT TTG GTA AAG GAA T, (#2) ATG TTA CGA CAG GTG GTT T, (#3) CAA CGA TGT ACG CCA AGA T. The transfections were performed with Lipofectamine 3000. The protein samples were collected for WB detection after transfection for 24 h.

The co-immunoprecipitation (co-IP) assay was performed as described before [13]. In brief, cells were lysed in co-IP buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X-100, and 1 mM EDTA) containing protease inhibitors (Roche Applied Science, Mannheim, Germany) on ice for 30 min. Then, the cells were centrifuged, and the supernatant was collected, followed by incubation with primary antibodies and GammaBind Plus Sepharose (#17088601; GE Healthcare, Logan, UT, USA) with gentle rocking overnight at 4 °C. The next day, the mixture was pelleted, washed six times with cold 1× co-IP buffer, and then analyzed by western blotting.

The Duolink® PLA assay was performed as indicated in the manual. In brief, AGS cells were treated as indicated and stained with mouse anti-SQSTM1 and rabbit anti-LDHA antibodies as described for the immunofluorescent staining. Duolink® PLA was then performed using the anti-rabbit PLUS (#DUO92002, Sigma, St. Louis, MO, USA) and anti-mouse MINUS (#DUO92004, Sigma, St. Louis, MO, USA) probes. Following probe incubation, ligation, and amplification, the cells were observed and photographed under the confocal microscope (Olympus FV-1000).

Total proteins were extracted from GC tissues or cells using RIPA lysis buffer containing protease inhibitor cocktail (Roche Applied Science, Mannheim, Germany). The lysates were mixed with SDS loading buffer, boiled for 8 min, resolved by SDS-PAGE, and then transferred to PVDF membranes (Millipore, Bedford, MA, USA). After blocking with 5% nonfat milk, the membranes were incubated with primary antibodies: anti-LDHA (#3582S, Cell Signaling Technology), anti-HA (clone 3F10, #11867423001, Roche), anti-FLAG M2 (#F1804; Sigma), anti-His TAG (#12698S, Cell Signaling Technology), pan succinyl-lysine antibody (#PTM-401; PTM Bio, Hangzhou, China), anti-β-actin (#4970; Cell Signaling Technology), or anti-CPT1A antibody (#12252; Cell Signaling Technology).

A total of 800 AGS cells stably expressing LDHA or LDHA variants were seeded in 6-well plates, cultured for about 14 days. Then, the cells were fixed with 70% methanol and stained with Giemsa solution. Colonies containing more than 50 cells were considered as survivors.

The cell invasion assay was performed in a 24-well Transwell Chamber (Costar, Corning, NY, USA) coated with Matrigel (BD Pharmingen, San Jose, CA, USA). AGS cells (2 × 10⁵ /200 μl) were cultured in the upper compartment in serum-free medium. In the lower compartment, 10% complete medium was added. After incubation at 37 °C for 24 h, the cells were fixed with 4% paraformaldehyde, stained by crystal violet, and then photographed under a microscope.

Cells were seeded and cultured in a 6-well plate until a confluent monolayer was formed. A sterile plastic tip was used to scratch on the monolayer of cells. Pictures were taken with a microscope at the specified timepoints to observe the migration distance. Migration was quantified as a percentage of wound closure.

Male nude mice (4–6 weeks old) were purchased from Model Animal Research Center of Nanjing University (Nanjing, Jiangsu, China). All animal studies were approved by the Nanjing Medical University Ethics Review Board. Approximately 5 × 10⁶ cells stably expressing Flag-LDHA, Flag-K222R, or K222E were subcutaneously injected into the nude mice. The tumor tissues were removed after 4 weeks, and the mice were euthanized. Tumor volume was calculated as width × length × (width + length)/2. LDHA levels were examined by western blotting. The lactate levels were measured by lactic acid assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) followed by the manual.

The mouse model of melanoma lung metastasis was employed following a previously established method [13, 26]. In brief, B16-F10 cells were administered to the mice by tail-vein injection (2 × 10⁵ cells/mouse in 200 μl DMEM) (n = 5 in each group). Lung melanoma metastases were determined by the number of colonies that appeared as black dots on the pleural surface.

In order to study the effect of protein PTM in GC, we selected seven patients with primary GC who had not received any treatment before surgery, and surgically extracted their tumor tissues and adjacent normal tissues for the study. Here, we sought to analyze protein succinylation in GC samples. The tumor tissues and adjacent normal tissues of seven patients were digested and dissociated by trypsin. Using TMT labeling and affinity enrichment followed by high-resolution LC-MS/MS analysis, quantitative lysine succinylome analysis was performed. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://​proteomecentral.​proteomexchange.​org) via the iProX partner repository [27] with the dataset identifier PXD018376. Altogether, 503 lysine succinylation (Ksuc) sites in 303 proteins were identified, among which 331 sites in 188 proteins were quantified. Based on the quantification of lysine succinylation in tumor versus normal tissue, we set a ratio > 1.3 as upregulation and < 0.77 as downregulation thresholds. The identified modified proteins are found in various cell organelles and are involved in several metabolic pathways (Fig. 1a). The GO analysis of upregulated and downregulated proteins, biological processes, and cellular components involved are shown in Additional file 1 (Fig. S1a-d). Numerous differentially expressed proteins and Ksuc sites were obtained in each group, including 56 upregulated and 127 downregulated sites (Fig. 1b). Further, analysis of the succinylation sites revealed that the Ksuc motif showed a strong preference for proline (P) or aspartic acid (D) in the + 1 positions, as well as for glutamic acid (E) in the + 2 position (Fig. 1c). The samples were divided into four groups according to the quantitative ratio of lysine succinylation sites between tumor and adjacent normal tissues: Q1 (< 0.667), Q2 (0.667 ~ 0.769), Q3 (1.3 ~ 1.5), and Q5 (> 1.5). The results of KEGG pathway enrichment cluster analysis on the four groups are shown in Fig. 1d. Functions involved in biological processes and cellular components were also analyzed and are shown in Additional file 1 (Fig. S1c and d), respectively. Protein-protein interaction network analysis performed on all up- and downregulated proteins is shown in Fig. 1e.

LDHA is highly expressed in several tumors, including GC (Fig. 2a), as showed on the GEPIA website (http://​gepia.​cancer-pku.​cn/​index.​html) (Additional file 2: Fig. S2a, b), and is closely related to the prognosis and staging of tumors. Therefore, we compared the expression of LDHA in cancer and adjacent normal tissues of the seven patients. We found that the mRNA level of LDHA was significantly higher in cancer tissues than in adjacent normal tissues (Fig. 2b). Correspondingly, we found a significant increase in LDHA in tumor tissues at the protein level (Fig. 2c). Accordingly, the lactate level was significantly increased in tumor tissues than that in adjacent normal tissues (Fig. 2d). Mass spectrometry analysis showed that LDHA was succinylated on K222 in GC tissues, where the level of LDHA succinylation was 1.42 times higher than in adjacent normal tissues (Fig. 2e).

Next, we generated K222 succinylation-specific antibodies using a human K222-succinylated LDHA peptide (PDLGTDKsucDKEQWK) as the antigen. In addition, we synthesized and constructed the LDHA full-length plasmid (Flag-LDHA) and the mutant plasmids Flag-K222R (mimicking deletion) and Flag-K222E (mimicking the negatively charged succinyl-lysine modification). Both the pan anti-succinyl-lysine antibody and K222 succinylation-specific antibodies (#1 and #2) recognized the Flag-LDHA protein but not the K222 mutants (Fig. 2f). The level of K222suc-LDHA was significantly higher in GC tissues than in adjacent normal tissues, as indicated by western blot (Fig. 2g). Similarly, as shown by IHC, the levels of LDHA and K222suc-LDHA were higher in GC tissues than in adjacent normal tissues (Fig. 2h, i, respectively). However, the staining of K222suc-LDHA seems to be much stronger than that of LDHA (Fig. 2h, i).

Intriguingly, in the three-dimensional structure diagram of LDHA homotetramer (PDB: 6MV8, http://​www.​rcsb.​org/​3d-view/​6MV8), K222 is highly exposed and extends outside (Fig. 2j). Succinylation at this site could therefore affect the binding capacity and protein-protein interactions, thereby influencing the structure and function of the protein.

Next, the expression of LDHA was examined in 293 T cells and four GC cell lines. LDHA level was the lowest in AGS cells (Fig. 3a), indicating that these cells could be useful for exogenous LDHA expression. The main protein degradation pathways in the cell include the ubiquitin-proteasome system (UPS) and the autophagy-lysosome system. To investigate whether LDHA can be degraded upon ubiquitination, Flag-LDHA was overexpressed in AGS cells, which were then treated with protein synthesis inhibitor cycloheximide (CHX) or proteasome inhibitor MG132. After CHX treatment, Flag-LDHA expression decreased significantly over time (Fig. 3b). However, LDHA protein levels showed no obvious changes after treatment with MG132 (Fig. 3c). There are numerous ubiquitination sites in human LDHA protein, as shown on the Phosphosite website (https://​www.​phosphosite.​org), which also include K222 site [28, 29]. To determine whether K222 succinylation affects the ubiquitination of LDHA, Flag-LDHA, Flag-K222R, or Flag-K222E in combination with HA-Ub (Ub), HA-Ub (K48), or HA-Ub (K63) plasmids were co-transfected into AGS cells. Flag-LDHA protein was pulled down by Flag antibody and ubiquitination was measured using HA antibody. The results showed that LDHA was ubiquitinated, mainly by K63-linked ubiquitin (Fig. 3d). K222 mutations, including the succinylation mimic K222E, did not affect the ubiquitination of LDHA (Fig. 3e-h).

Two inhibitors of the autophagy lysosomal pathway, chloroquine and Bafilomycin A1 (BafA1), were used to treat Flag-LDHA-overexpressing AGS or HGC27 cells. Flag-LDHA was expressed stably in AGS and HGC27 cells after transfection (Fig. 4a). Importantly, Flag-LDHA expression significantly increased after treatment with either BafA1 (Fig. 4b) or chloroquine (Fig. 4c). These results indicate that LDHA is indeed degraded by the autophagy lysosomal pathway. The ubiquitin proteasome pathway and autophagy lysosomal pathway are not completely independent from each other. There are a lot of cross-links between them such as sequestosome 1 (SQSTM1). Therefore, to investigate whether LDHA could interact with SQSTM1, AGS or 293 T cells were co-transfected with Flag-LDHA and His-SQSTM1 for 24 h prior to co-IP with either anti-Flag antibody or control IgG. Western blot analysis revealed that LDHA can interact with SQSTM1 (Fig. 4d). The interaction was further confirmed by in situ proximity ligation assay (PLA) (Fig. 3e). As shown in the figure, red PLA signals were visible in control-treated AGS cells (upper), but were much stronger in BafA1-treated AGS cells (lower). The co-localization of LDHA and SQSTM1 was also observed by immunofluorescence (magnification 800x in Fig. 4f) and indicated by the Pearson’s coefficient (Fig. 4g) and the Mander’s coefficients (Fig. 4h). In BafA1-treated AGS cells, Flag-LDHA expression in autophagy lysosomes significantly increased, and its localization considerably overlapped with that of SQSTM1, further demonstrating that LDHA is degraded by the lysosomal pathway.

Exogenous Flag-K222E or Flag-K222R LDHA proteins were relatively stable in AGS or HGC27 cells after gene transfection (Fig. 5a, Fig. S3a). To investigate the effect of K222suc on the stability of LDHA, Flag-K222E or Flag-K222R plasmid-transfected cells were treated with CHX, MG132, chloroquine, or BafA1. Flag-K222E protein expression was relatively stable even after CHX treatment (Fig. 5b, upper; Fig. S3b), while that of Flag-LDHA (Fig. 3b) or Flag-K222R (Fig. 5b, lower) was significantly reduced. Interestingly, the addition of MG132 resulted in an increase in Flag-K222E protein levels (Fig. 5c, upper; Fig. S3c), but it did not affect the levels of either Flag-LDHA (Fig. 3c) or Flag-K222R (Fig. 5c, lower). However, treatment with chloroquine or BafA1 increased the protein levels of Flag-K222E (Fig. 5d, e, upper; Fig. S3d), Flag-LDHA (Fig. 4b, c), and Flag-K222R significantly (Fig. 5d, e, lower). These results suggest that Flag-K222E accumulated in cells and that K222suc could inhibit the lysosomal degradation of LDHA.

Consistently, Flag-K222E could still bind to SQSTM1 (Fig. 5f). However, the interaction of Flag-LDHA with SQSTM1 was much stronger than that of Flag-K222E (Fig. 5g). Furthermore, K63-linked ubiquitination promoted the binding of LDHA to SQSTM1 (Fig. 5h), while K48-linked ubiquitination had no such effect (Fig. 5i). Neither K63 nor K48 promoted the interaction between Flag-K222E and SQSTM1 (Fig. 5j, k, respectively). To further verify this, Flag-LDHA, His-SQSTM1, and either GFP or HA-UbK63 were co-transfected into AGS cells. Upon co-transfection of Flag-LDHA and SQSTM1 with K63, the protein level of Flag-LDHA decreased (Fig. 5l), while upon co-transfection with GFP, the protein level remained stable (Fig. 5m). Instead, the protein levels of Flag-K222E dramatically increased upon co-transfection of Flag-K222E and SQSTM1 with either K63 or GFP (Fig. 5n, o). In AGS cells co-transfected with Flag-LDHA, SQSTM1, and K63, Flag-LDHA accumulated upon treatment with BafA1 (Fig. 5p). These results show that K63-linked ubiquitination of LDHA leads to its degradation by promoting its binding to SQSTM1, whereas succinylation on K222 inhibits LDHA degradation by reducing its interaction with SQSTM1.

Next, we sought to determine whether KAT2A and CPT1A could promote K222 succinylation of LDHA. KAT2A did not interact with either LDHA or K222E mutant (Fig. 6a, b). Remarkably, co-IP results indicated that CPT1A could bind to both LDHA (Fig. 6c, d) and K222E mutant (Fig. 6e). More importantly, overexpression of CPT1A increased the K222 succinylation of LDHA (Fig. 6f), but not of K222E mutant (Fig. 6g). To verify this effect, we knocked down CPT1A with three shRNAs, and chose the second shRNA based on its efficiency (Fig. 6h). Consequently, the knockdown of CPT1A resulted in the reduction of LDHA succinylation (Fig. 6i).

To analyze the association of LDHA K222suc and prognosis of patients with GC, the specific K222suc antibody was employed to conduct tissue microarray-based IHC. Tumor tissues and adjacent normal tissues were scored according to the IHC results. K222suc was considered upregulated in samples with a ratio > 1.3 in tumor/normal tissue (Table 1). Interestingly, we found that the proportion of LDHA with high K222 succinylation was much higher in female patients than in male patients (Fig. 7a). Moreover, among all patients, those with high succinylation of LDHA K222 had lower overall survival (OS) (Fig. 7b). We also analyzed the OS based on other key clinicopathological factors and according to LDHA K222suc levels. Significant differences were found in pathological grade III-IV, T3-T4, and clinical stage III-IV tumors (Fig. 7c-e). Interestingly, although the tumors with higher LDHA K222 succinylation were more likely to metastasize to lymph nodes (Table 1), the OS did not significantly differ in patients with a higher degree of lymphatic metastasis (N2-N3) (Fig. 7f).

To explore the effect of LDHA K222suc on the function of GC cells, we established AGS cells stably expressing Flag-LDHA, Flag-K222R, and Flag-K222E using lentiviruses (Fig. 7g). As determined by wound-healing assays, AGS cells expressing Flag-K222E had increased ability to migrate (Fig. 7h). Similarly, transwell and plate colony formation experiments showed that invasion and proliferation also increased in AGS cells expressing Flag-K222E, but did not significantly change in cells expressing Flag-LDHA and Flag-K222R (Fig. 7i, j). To address the biological significance of K222 succinylation in vivo, we performed mouse melanoma lung metastasis experiments using B16-F10 cell lines overexpressing WT, K222R or K222E mutant LDHA. Our data show that K222E mutant LDHA-overexpressing B16-F10 cells caused more lung metastasis than WT or K222R LDHA-overexpressing cells (Fig. 7k).

Furthermore, a xenograft model was adopted to determine the effect of Flag-K222E on tumor growth. We inoculated AGS cells stably expressing Flag-LDHA, Flag-K222R, and Flag-K222E subcutaneously in nude mice to observe the formation and the size of tumors. AGS cells expressing Flag-K222E formed larger subcutaneous tumors than AGS cells expressing Flag-LDHA or Flag-K222R, and there was no significant difference between the tumors from AGS cells expressing Flag-LDHA and Flag-K222R (Fig. 7l). Moreover, the level of LDHA and lactate were significantly higher in the tumors from Flag-K222E group as shown in Fig. 7m and n, respectively. Together, these data indicated that K222E, which mimics K222suc, promoted tumor cell invasion, metastasis, and growth.