Quantitative ubiquitylomics reveals the ubiquitination regulation landscape in oral adenoid cystic carcinoma

Abstract Adenoid cystic carcinoma (ACC) is an extremely rare salivary gland tumor with a poor prognosis and needs attention on molecular mechanisms. Protein ubiquitination is an evolutionarily conserved post-translational modification (PTM) for substrates degradation and controls diverse cellular functions. The broad cellular function of ubiquitination network holds great promise to detect potential targets and identify respective receptors. Novel technologies are discovered for in-depth research and characterization of the precise and dynamic regulation of ubiquitylomics in multiple cellular processes during cancer initiation, progression and treatment. In the present study, 4D label-free quantitative techniques of ubiquitination proteomics were used and we identified a total of 4152 ubiquitination sites in 1993 proteins. We also performed a systematic bioinformatics analysis for differential modified proteins and peptides containing quantitative information through the comparation between oral ACC (OACC) tumor with adjacent normal tissues, as well as the identification of eight protein clusters with motif analysis. Our findings offered an important reference of potential biomarkers and effective therapeutic targets for ACC.


Introduction
Adenoid cystic carcinoma (ACC) is discovered a long time ago with location in the major and minor salivary glands [1,2] and other organs [3], which can be divided into tubular form (Grade I), cribriform (Grade II) and solid form (Grade III) [4]. Clinical staging had more significant importance than histological grade on prognosis [5]. The ten-year survival rates are about 73% (Stage I), 43% (Stage II) and 15% (Stage III and IV) [4]. Studies on oral ACC (OACC) are even more uncommon and insufficient.
Protein post-translational modifications (PTMs) exist in both eukaryotes and prokaryotes at one or more sites [27]. More than 200 types of PTMs are identified in humans [28]. Ubiquitination is the covalently conjugation in proteins by conserved small protein ubiquitin and ubiquitin-like (Ubl) proteins either as a monomer or as a polyubiquitin chain. [29] Ubiquitin-activating protein E1, ubiquitin-conjugating protein E2 and ubiquitin-ligase E3 are involved to form a transient reaction, which can be reversed by deubiquitylating enzymes (DUBs) [30][31][32]. The modified substrates are degraded after binding to a multi-subunit protease complex [29]. The dynamic changes of ubiquitination form an enzymatic and complete ordered system to control subcellular processes [33], which affects pathophysiological states in cancer under various conditions [34,35]. Both oncogenes and tumor-suppressor genes undergo ubiquitination [36]. Several studies mentioned the potential of ubiquitination Downloaded from http://portlandpress.com/bioscirep/article-pdf/doi/10.1042/BSR20211532/918230/bsr-2021-1532.pdf by guest on 06 August 2021 Bioscience Reports. This is an Accepted Manuscript. You are encouraged to use the Version of Record that, when published, will replace this version. The most up-to-date-version is available at https://doi.org/10.1042/BSR20211532 related proteins as OACC biomarkers. From Nanostring nCounter miRNA assay, ubiquitin-like modifier activating enzyme 2 (UBA2) was identified increased in primary and recurrent tumors than normal tissue, revealing the potential connection between UBA2 and tumor recurrence and metastasis [37]. The expression of Ubiquitin-specific protease 22 (USP22) in salivary adenoid cystic carcinoma (SACC) was higher in the tumor group than in the adjacent normal group, which was associated with a poor prognosis [38]. The low expression of the tumor suppressor gene cylindromatosis (CYLD), which has deubiquitinating enzyme activity, corrected with salivary gland tumor progression through NF-κB pathway [39]. Notwithstanding, there is lack of relevant research on the molecular mechanisms and related pathogenesis of ubiquitination network in OACC.
A global and comprehensive information about ubiquitin system is difficult due to the challenge for high-throughput analysis [40]. Another obstacle is that these revisable modifications can be easily lost or not easily detected during the experiment.
The mass spectrometry (MS) based proteomic approaches have been widely used in qualitative and quantitative analysis of cellular biology [41]. The organic combination of non-standard quantitative and MS based proteomic technology can effectively identify ubiquitination substrates and modification network in OACC.
In our project, 4D label free quantitative ubiquitination proteomics was carried out.
A total of 4152 ubiquitination sites were identified on 1993 proteins, in which 1648 loci of 859 proteins contained quantitative information. We also conducted a systematic bioinformatics analysis, including protein annotation, functional classification, functional enrichment and cluster analysis based on functional enrichment proteins. The proteomic methodologies in our work illustrating ubiquitination landscape in OACC can be applied to the search and identification of novel molecular biomarkers, provide valuable information for diagnosis and help discovering novel therapeutic anti-cancer strategies.  Table Ⅰ.

Protein extraction
Appropriate number of samples were added with liquid nitrogen to grind to powder and added with 4 times of powder lysis buffer (1% Triton X-100, 1% protease inhibitor, 50 μm PR-619, 3 μm TSA, 50 mm NAM) for ultrasonic pyrolysis. After centrifugation in 12000 g for 10 min, the supernatant was transferred to a new centrifuge tube. The protein concentration was determined by BCA kit.

Trypsin digestion
TCA was slowly added in each sample, followed by precipitating at 4 ℃. The supernatant was discarded after centrifugation in 4500 g for 5 min. The precipitate was washed with pre-cooled acetone, dried in the air and dissolved by 200 mm TEAB buffer. Trypsin was added and the reaction was maintained overnight. After the incubation by dithiothreitol at 56 ℃ for 30 min, iodoacetamide (IAA) was added. The samples were then incubated at room temperature for 15 min in dark.

Mass spectrometry
The digested peptides were dissolved by liquid chromatography mobile phase A (containing 0.1% formic acid and 2% acetonitrile) and separated by NanoElute ultra performance liquid chromatography system. The flow rate was 450.00nl/min, and the gradient of mobile phase B (containing 0.1% formic acid and 100% acetonitrile) was set as follows: 6%-22%: 0-43 min; 22%-30%: 43-56 min; 30%-80%: 56-58 min; 80%: 58-60 min. The peptides were separated by Ultra Performance Liquid Chromatography (UPLC) and injected into Capillary Ion Source for ionization to be analyzed by Tims-TOF Pro mass spectrometry. The voltage of ion source was 2.0 kV, 6 and the peptide parent ion and its secondary fragments were detected and analyzed by high resolution TOF. The scanning range of secondary MS is set to 100-1700. The data acquisition mode is PASEF. After the collection of a first-order mass spectrum, the second-order mass spectrum with charge number of parent ion in the range of 0-5 was collected 10 times in PASEF mode, and the dynamic exclusion time of tandem mass spectrometry scanning was set to 30s.

Systematic profiling of protein ubiquitination in OACC samples
In order to globally reveal the involvement of ubiquitin in the progression and regulation of OACC, we performed 4D label free quantitative ubiquitination proteomics study through comparing OACC tumor samples (OACC_T) with the adjacent normal samples (OACC_N) in 4 patients who had not received any drug treatment before operation. The identification data were filtered as localization probability > 0.75. After MS analysis and database search, a total of 63282 secondary spectra were obtained, of which 15172 were available. The relative quantitative value was obtained according to the intensity of the modified site between different samples.
According to this method, 7956 peptides and 4116 modified peptides were identified.

Functional classification of differentially ubiquitinated proteins (DUPs)
In order to better study the ubiquitinated proteins in OACC, we firstly investigated the subcellular location of DUPs. Nearly half of the ubiquitinated proteins located in cytoplasm (n=196, 43.36%), followed by plasma membrane (n=73, 16.15%) and nucleus (n=65, 14.38%) (Fig. 1c). We also performed the up-and down-subcellular

Functional enrichment analysis of DUPs
We conducted the functional enrichment analyses with GO annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis.
Followed by phagocytic vesicle membrane, cytosolic ribosome and ribosomal subunit were most differed in CC category. In MF category, the most differentially ubiquitinated proteins were concentrated in insulin-like growth factor receptor  (Table SⅢ).

Cluster analysis of differential modified proteins
In order to test the rationality and accuracy of the identified Fig. 2b, we used cluster analysis to aggregate the proteins according to the trend of expression. In order to find the correlation of protein functions in different fold change, we divide DUPs into four clusters according to OACC_T/OACC_N ratio, which were called Q1 (Ratio < 0.5), Q2 (Ratio between 0.5-0.667), Q3 (Ratio between 1.5-2) and Q4 (Ratio > 2) (Fig. 3a).
In GO enrichment analysis, chromaffin granule membrane, cytosolic small ribosomal subunit, myelin sheath and polysome differed the most in each cluster by cellular component category. In molecular function category of each cluster, the most enriched were acid-ammonia (or amide) ligase activity, neurexin family protein binding, ion channel binding and ADP binding. In addition, filopodium assembly, mitochondrial RNA metabolic process, positive regulation of stress fiber assembly and sarcomere organization were the top regulated in biological process category in each cluster    (Table Ⅲ).

Motif analysis of protein modification
Protein motif analysis calculates the regular trend of amino acid sequences in the region of modification sites. This kind of analysis can find the sequence characteristics, so as to speculate or determine the modification related enzymes. We identified 8 conserved ubiquitination motifs analyzed by Motif-X (Fig. 4b). The enrichment of specific amino acids neighboring the ubiquitination sites were exhibit (Fig. 4c).

Discussion
ACC of the salivary glands performs the properties of slow-growing, local and/or distant spread, nodal positivity and high-mortality, as well as high rate of occurrence and metastasis [43,44]. Several specific prognostic factors had been identified the The ubiquitination modification plays a significant role in cancer pathology [46].
Mass spectrometry can be used to analyze the role of protein PTMs in human diseases, and PTM-based protein variants can be explored as deeply as possible.
Ribosomal protein S27a (RPS27A) is the top differentially up-regulated protein in our identification. RPS27A performs multi-function in ribosome biogenesis and protein PTMs, contributing to progression of leukemia or solid tumors [47,48]. A ubiquitin fused RPS27A protein (Uba80) was reported related to apoptotic cell death and over-expressed in colon and renal cancer [49][50][51][52]. However, the molecular mechanism of RPS27A-related ubiquitination in tumors remains to be studied.
Phosphoglycerate dehydrogenase (PHGDH) is the rate limiting enzyme of de novo serine biosynthesis pathway, which is closely related to the occurrence and development of many kinds of tumors [53,54]. The serine synthesis during cancer progression was suppressed when PHGDH went through Parkin-related ubiquitination and degradation [55,56]. PHGDH is also a ubiquitination substrate of RNF5 in the study of breast cancer cells [57]. The PHGDH ubiquitination in OACC has not been reported. TMEM87A, also named as Elkin1, is important in cell-cell adhesin and metastasis with limited studies [58]. The insulin receptor (INSR) is a key regulator in metabolic homeostasis through diverse signal pathways including PI3K/AKT and MAPK [59]. Although phosphorylation is critical in INSR-dependent signal cascade, the ubiquitin/proteasome system modulate degradation of transducers in this pathway [60,61]. The biological function of these targets in OACC and their combination with metabolic abnormalities and immune regulation will drive us to further study how gene changes modulate the behavior of cancer cells, so as to unlock more effective treatment methods.
In addition, ATP1A1, TUBA1B and ITGA9 are the top down-regulated according to OACC_T/OACC_N Ratio. As a membrane bound ion pump, Na+/K+-ATPase shows tissue-specific profile [62]. The overexpression of α1 subunit of Na + /K + -ATPase (ATP1A1) were observed in esophageal squamous cell carcinoma [63], non-small cell lung cancer and hepatocellular carcinoma, contributing to cancer proliferation and migration [64,65]. However, ATP1A1 were significantly down-regulated in prostate cancer [66,67], colorectal cancer [68]and renal cell carcinoma [69]. The function and PTM regulation of ATP1A1 in OACC is not yet fully clear. Tubulin alpha 1b (TUBA1B) belongs to cytoskeleton compartment with a central function in cell shape maintenance and cellular process regulation, especially in cell division [70]. The higher expression of TUBA1B and poor prognosis were reported in hepatocellular carcinoma [71]. Interesting, the ubiquitination at different sites of TUBA1B shows different up-and down-regulation trends in tumor tissues (Table II). This suggests that TUBA1B ubiquitination regulation may have different biological functions. With the continuous deepening of cytoskeleton related research, the importance of microtubules in tumor metastasis has begun to become prominent.
The integrin subunit α9 (ITGA9) belongs to integrin protein family and a partner of β1 subunit facilitating the interaction of cell-cell and cell-extracellular matrix [72].
Depletion of ITGA9 suppressed breast cancer progression and metastasis through GSK3/β-catenin pathway [73]. The decreased expression of ITGA9 in lung cancer indicated potential genetic and epigenetic regulation mechanism [74]. Besides TUBA1B and ITGA9, other cell cytoskeleton and cell adhesion associated proteins were also identified in our results, including ACTA2, MCAM, ACTG1, VIM and CDH11. Due to the high recurrence and metastasis of OACC, research on the modification of cytoskeleton and cell adhesion will help to further clarify the mechanism of OACC metastasis.
Considering the hub genes of interaction regulation by Cytoscape analysis, we found they all belong to ribosomal proteins. Ribosomal proteins are abnormally expressed in a variety of tumors, which affect the apoptosis, aging, growth, invasion, drug and radiation resistance [75]. The expression level of ribosomal protein has become a potential indicator of tumor diagnosis, treatment and prognosis [76]. Protein ubiquitination plays a very important role in cellular processes such as subcellular localization, growth, apoptosis and metabolism. The ubiquitination modification omics based on mass spectrometry has been used for disease biomarkers and pathogenic mechanism analysis [77,78]. In this project, we studied the ubiquitin However, whether the selected differential protein can be used as a therapeutic target for OACC still needs more detailed in vivo and in vitro experiments to verify, so as to better carry out precise treatment for OACC, add further knowledge specifically for patients, and explore more promising combination therapy.

Corresponding Authors
Donge Tang

Ethics approval and consent to participate
The study was carried out following the Declaration of Helsinki for experiments involving humans and was approved by the Medical Ethics Committee of Shenzhen People's Hospital.

Data availability
All data included in this study are available upon request by contact with the corresponding author.

Competing interest
The authors declare no competing financial interests or personal relationships that could affect the work published in this paper.