Ribosomal protein S3 is a protein component of the ribosome, which is a cellular structure that plays a central role in protein synthesis. It has extra-ribosomal functions, so called UV endonuclease III, to repair the DNA damage by UV irradiation (1-5). This enzyme cleaves UV irradiated DNA and AP (apyrimidic/apurinic) DNA, by cleaving the phosphodiester bond within a cyclobutane pyrimidine dimer and 3’ of AP site via a lyase mechanism (1, 5, 6). A previous study showed that rpS3 is secreted in cancer cell lines, and the level of secretion increased as the malignancy of the tumor cells increased (7, 8). RpS3 protein and mRNA expression levels are higher in both leukemia patients and leukemia cell lines than in their normal counterparts (8, 9). These results suggest that rpS3 can be used as an indicator of tumorigenesis. Therefore, in this study, we investigated the possible role of rpS3 as a key target of anti-cancer drug resistance.

In addition to its multiple extra-ribosomal functions, rpS3 possesses several post-translational modification motifs. Although many roles of these modifications are yet to be found, several different forms of modifications have been found both in vitro and in vivo, including phosphorylation, neddylation, sumoylation, and ubiquitination, thus imparting this protein versatile functions through ribosomal heterogeneity (8). Among these, rpS3 phosphorylation has been well-studied in conjunction with its functions of DNA repair and secretion (7, 8, 10-12).

RpS3 was also identified as a mammalian DNA repair endonuclease III, which has AP endonuclease/lyase activity and is involved in base excision repair (13). In addition, rpS3 is involved in apoptosis in cases of severe DNA damage in immune cells (12). These functions are important for maintaining cellular homeostasis and are regulated through several pathways. In our previous study, we found that rpS3 exists in the promyelocytic leukemia (PML) body within the nucleus and is modified by SUMO-1 (14). Hence, rpS3 may play a major role in B-cell growth and differentiation and be associated with acute myeloid leukemia along with PML. In this study, we examined the correlation between post-translational modifications and signaling systems of rpS3 that can integrate control pathways, such as protein synthesis, DNA damage recovery, and cell death.

The Lyn is a member of the Src family of tyrosine kinases and is mainly expressed in hematopoietic tissues (15, 16). It plays an important role in the regulating various cellular processes, such as cell growth, survival, migration, and immune response (17-20). This protein contains SH1 (kinase domain), SH2, and SH3 domains and a unique N-terminal sequence; the latter is dispensable for membrane localization via attachment of fatty acid myristate. The SH3 domain binds to proline-rich motifs, such as PXXP, and the SH2 domain binds to specific phosphotyrosine motifs. In addition, the SH2 and SH3 domains are implicated in substrate binding and the regulation of kinase activity via intramolecular interactions (18). Studies using cell fractionation and confocal microscopy have demonstrated that Lyn kinase also exists in the nucleus (19). Nuclear Lyn, but not other Src family tyrosine kinases, are activated after genotoxic treatment, such as 1-β-D-arabinofuranosylcytosine (Ara-C), ionizing irradiation, and adriamycin (20). Recently, BCR-ABL fusion proteins have been reported to activate Src tyrosine kinases, such as Lyn, Hck, and Fgr, to induce leukemia (21). Lyn has been studied mainly in the growth and differentiation processes of B lymphocytes (22); however, many studies have been recently conducted on its activation caused by DNA damage (23). Studies on protein synthesis have revealed various regulatory mechanisms of ribosomal S6 kinase, and some studies on regulatory mechanisms for other ribosomal proteins have also been revealed (24).

P-glycoprotein (p-gp) or multidrug resistance protein (MDR1) is a type of protein that is expressed in the membranes of certain cells, including cells in the blood-brain barrier and the cells lining the small intestine. It is a member of the ATP-binding cassette (ABC) family of transporters and is responsible for pumping drugs and other xenobiotics out of cells, making them less effective. ATP-binding cassette (ABC) transporters are known to have multiple drug binding sites, which allow them to transport a diverse range of substrates, including drugs such as doxorubicin (25). High expression levels of ABCB1 (also known as P-glycoprotein or P-gp) are commonly associated with drug resistance in cancer cells (26), both acute myeloid leukemia (AML) and acute lymphocytic leukemia (ALL) (27).

The multidrug resistance protein 1 (MDR1) is involved in the extrusion of a wide variety of drugs and xenobiotics from cells, leading to decreased intracellular drug accumulation and, ultimately, reduced drug efficacy (28). MDR1 has also been implicated in resistance to certain anticancer, anti-inflammatory, and antiviral drugs, among others (26). P-gp is known to be regulated by post-translational modifications such as phosphorylation and N-glycosylation (29).

Doxorubicin (DOX) is a chemotherapy drug that works by inhibiting DNA replication and inducing DNA damage in cancer cells. However, the effects of doxorubicin are often limited by overexpression of P-glycoprotein (P-gp) (30). DOX is widely used and highly effective for the treatment of breast cancer, sarcoma, leukemia, and lymphoma (31). Inhibition of p-gp as a method to reverse MDR in patients with cancer has been extensively reviewed; however, the results have generally been disconcerting (32).

Acute PML and chronic myeloid leukemia are associated with the activation of Src tyrosine kinases; thus, investigating the modification of the rpS3 protein could provide important insights into cancer treatment.

Herein, we found that rpS3 interacts with Lyn and that this interaction elicits the phosphorylation of rpS3. Furthermore, we investigated the effect of Lyn on rpS3 phosphorylation by adding various kinase inhibitors. We also demonstrated that the phosphorylation of rpS3 by Lyn regulates p-gp translation.

RpS3 is phosphorylated after the interaction with Lyn

Our previous studies using a yeast two-hybrid system revealed that Lyn is a putative binding partner of rpS3 (14). To demonstrate the interaction of the tyrosine kinase Lyn with rpS3, we performed a co-immunoprecipitation assay using overexpression of pEGFPc1-Lyn. HEK293T cells were transfected with either a blank GFP vector or GFP vector carrying Lyn protein. As shown in Fig. 1A, the interaction between rpS3 and GFP-Lyn was verified (lane 2); however, the interaction between rpS3 and lysate was not verified (lane 1). Next, we investigated whether this interaction occurs between endogenous Lyn and rpS3. We confirmed specific interactions between endogenous proteins in MPC-11 cells. Conversely, endogenous rpS3 was precipitated with antibodies against endogenous Lyn. These results indicated that rpS3 has a strong physical interaction with Lyn in vitro (Fig. 1A).

As Lyn is a well-known protein tyrosine kinase, we tested the possibility of rpS3 phosphorylation by Lyn. For this purpose, we performed an immunocomplex kinase assay with an anti-Lyn antibody, and recombinant His-tagged rpS3 protein was used as a substrate. Immunoprecipitates were incubated with acid-treated enolase as a positive control. As shown in Fig. 1B, autophosphorylation of Lyn demonstrated that rpS3 protein inhibits the kinase activity of Lyn. As expected, recombinant His-tagged rpS3 protein was phosphorylated by Lyn. Collectively, these results suggest that rpS3 protein (lane 3) can be phosphorylated in vitro as much as enolase (lane 2), which is the positive substrate of Lyn and acts as a regulator of Lyn.

RpS3 interacts with SH3 domain of Lyn

To determine the region that is crucial for the interaction of Lyn with rpS3, a fine deletion analysis of the N- and C-termini was performed. For this purpose, we generated GST-tagged Lyn deletion constructs, as described in Fig. 2A. All the GST-tagged constructs were transferred into HEK293T cells, and GST pull-down assays were performed to target rpS3 (Fig. 2B). Furthermore, GST-tagged Lyn deletion mutants were co-transfected with a plasmid bearing GFP-tagged rpS3 into HEK293T cells, and an in vitro binding assay was performed (Fig. 2C). These results demonstrate that rpS3 binds to the SH3 domain (63-123, lane 3) and SH1 domain of Lyn (240-500, lane 5) but not to the unique N-terminal domain (1-62, lane 2) and SH2 domain (128-226, lane 4). As the binding affinity of the SH1 domain was stronger than that of SH3, we suggest that the SH1 domain of Lyn is the major binding domain with rpS3 (Fig. 2C). Furthermore, to examine the role of the SH3 domain of Lyn, we co-transfected the GFP-tagged SH3 domain of Lyn and rpS3 into Cos-1 cells (Supplementary Fig. 1). Although the wild type Lyn interacted very weakly, the SH3 domain of Lyn strongly interacted with rpS3 (lanes 3 and 4).

The RpS3 protein is highly basic and contains two domains: the KH domain and the S3-C terminal domain. We tested which region of rpS3 interacted with the Lyn, N-, and C-terminal deletion constructs, as described in Fig. 2D (33). Each deletion mutant was co-expressed with Lyn in HEK293T cells, and immunoprecipitation analysis was performed using cell lysates. Interestingly, rpS3 appeared to interact with Lyn via both N-terminal and C-terminal domains (Fig. 2E, lanes 5, 6, 7, 9; Open arrowheads). This result implies that there are more than one binding sites in the rpS3. More specifically, the Lyn-binding region of rpS3 lies in the N-terminal (rpS3: 1-85, lane 7) and C-terminal regions (rpS3: 159-242, lane 3). Overall, these results suggest that the association between Lyn and rpS3 occurs by direct interaction of the Lyn SH3 domain with the C-terminal region of rpS3 (159-242, lane 3), or it may occur by the interaction of the Lyn SH1 domain with the N-terminal region of rpS3 (1-85, lane 7). The binding activity of rpS3-P (159-242, Fig. 2E, lane 3; GST-rpS3 C-terminal, Supplementary Fig. 2, lane 9) was equal to that of the wild type. These results demonstrated that the binding region of rpS3 for Lyn is within the C-terminus of the rpS3 protein. The C-terminal region (Fig. 2E, lane 3 and Supplementary Fig. 2, lane 9) contains three putative proline-rich motifs, such as PxxP, that can bind to the SH3 domain of Lyn, which raises the possibility that the SH3 domain of Lyn would take part in the interaction.

Tyrosine 167 of rpS3 is a critical residue for the phosphorylation by Lyn in vitro

Human rpS3 has several phosphorylation target sites, mostly serine/threonine residues (31, 32). To identify the tyrosine residue phosphorylated by Lyn, we generated a series of rpS3 mutations in which each tyrosine was substituted with phenylalanine (Fig. 3A). Subsequently, an immunocomplex kinase assay was performed (Fig. 3B). Substitution of all six tyrosine residues with phenylalanine (KS2, lane 4) abrogated Lyn-mediated phosphorylation. Interestingly, the double substitution mutant (KS6, Lane 8) also abrogated phosphorylation. Based on these results, Y166 and Y167 were assumed to be important for phosphorylation at rpS3. We generated appropriate single substitution mutants (Fig. 3C) for further investigation. The results of this analysis revealed that Y166 was weakly phosphorylated, while the Y167 residue was not phosphorylated. Therefore, we concluded that Lyn-induced rpS3 phosphorylation was eliminated by Y167F mutation of rpS3 and confirmed that the Y167 residue of rpS3 is the critical phosphorylation site for Lyn.

We tested the effects of DOX, an anti-cancer drug that inhibits topoisomerase II, on the association of Lyn with rpS3. When GFP-rpS3 was treated with DOX, the interaction between rpS3 and Lyn increased (Supplementary Fig. 3, lower panel, lane 4). The results of this study suggested that DOX plays an important role in the interaction between rpS3 and Lyn.

Lyn may be activated by genotoxic agents, such as Ara-C (1-beta-D-Arabinofuranosylcytosine), adriamycin (DOX), and ionizing radiation (23). Therefore, we investigated whether the phosphorylation of rpS3 increases after treatment with genotoxins. As expected, the higher the concentration of Ara-C, the higher was the level of phosphorylated rpS3 (Supplementary Fig. 4, lane 3). However, the result may not be owing to a specific phosphorylation by Lyn; hence, we subsequently studied the effect of a global and specific inhibitor of Src family tyrosine kinases such as PP2 and genistein. Overall, rpS3 phosphorylation decreased in a dose-dependent manner. Thus, these results suggest that the phosphorylation of rpS3 is due to Src tyrosine family kinases, such as Lyn.

Increase of MDR1 expression in DOX-resistant cells

The levels of MDR1 mRNA and p-gp in the DOX-resistant MPC-11 cell line (MPC11-DOX) were significantly higher than that in the MPC-11 parent cells (Supplementary Fig. 5A). In addition, the level of tyrosine phosphorylation of rpS3 in MPC11-DOX cells increased significantly compared to that in MPC-11 parent cells (Supplementary Fig. 5B, C, lane 6). These results confirmed that phosphorylation of the tyrosine residue of rpS3 was increased in MPC11-DOX cells. DOX induction of p-gp expression occurred owing to the phosphorylated status of rpS3. Although a DOX-resistant cell line was established, there was little change in the ribosome content of this cell line through the ribosomal profile, particularly in polysomes (Supplementary Fig. 5D). We examined the translation levels of c-Jun NH2-Terminal kinase and rpS3 proteins in non-ribosome and ribosome fractions (34) and compared the levels of MDR1 mRNA in parent cells (M, M’) and MDR11-DOX cells (D, D’) using reverse transcription polymerase chain reaction (Fig. 4A). The results demonstrated that in the ribosomal protein fractions, MDR1 mRNA was higher in MPC11-DOX cells (D, Fig. 4A, lane 2) than in parental cells (M, Fig. 4A, lane 1). MPC11-DOX cells with enhanced resistance to DOX (D’, Fig 4A, lane 4) showed a marked increase in MDR1 mRNA compared to the original MPC11-DOX cells (D, Fig. 4A, lane 2). Altogether, these results suggest that selection for drug resistance, by long-term exposure to drugs, led to the establishment of a boosted DOX-resistant cell line D’, in which the translational block of MDR1 mRNA was overcome such that the mRNA was translated and p-gp was overexpressed (35). Based on these results (Fig. 3C), we performed in vivo quantitative measurements of p-gp synthesis following control rpS3-WT or Y166D and Y167D as substrates (phosphomimetic form) in MPC11-DOX cells using [S³⁵]-methionine labeling (Fig. 4B). P-gp of Y166D (lane 3) was similar to that of the wild type. In the case of Y167D (lane 4), p-gp expression level was three-fold higher than that of the wild type (lane 2). This suggests that protein synthesis (i.e., translation) of p-gp is upregulated by the phosphorylation of rpS3 by Lyn.

RpS3 is an essential constituent of the 40S subunit of the ribosome and is involved in protein translation. However, the extra-ribosomal functions of many ribosomal proteins have not yet been elucidated. RpS3 has been extensively studied for its functions outside of the translational machinery, and multiple roles in DNA damage repair and apoptosis have been reported (1, 8, 12).

Overexpression of rpS3 in colorectal cancers has been reported, and the relationship between rpS3 and cancer development has long been suspected. In this regard, we studied the relationship between leukemia-related carcinogenesis and the modification of rpS3. Based on the regulatory properties of Lyn in tumorigenesis and its biochemical relationships with various binding partners, we hypothesized that the relationship between rpS3 and Lyn could be functionally involved in the drug resistance of MPC11 cells. The antitumor effect of DOX primarily involves topoisomerase-mediated double-strand DNA breaks with the subsequent triggering of DNA damage-associated cell cycle arrest and apoptosis pathways. In this study, we showed the interaction between the rpS3 protein and Lyn, demonstrating the phosphorylation of rpS3 by Lyn (Fig. 1). The SH3 domain of Lyn contains a module that can bind to the proline-rich motif (PxxP) (25-27). Although three putative proline-rich motifs that lie in the C-terminus of rpS3 are not exactly matched with the conserved proline-rich motifs, our results showed that the putative proline-rich motifs of rpS3 interacted with the SH3 domain of Lyn (Fig. 2C). However, to identify the binding sites between the rpS3 and SH3 domains of Lyn, further studies need to be carried out, such as point mutations in three putative proline-rich motifs. Furthermore, we showed that the SH1 domain of Lyn interacted with rpS3 (Fig. 2C).

The results of kinase assay also revealed that rpS3 is directly phosphorylated by Lyn, and the level of phosphorylation is comparable to that for enolase, a typical phosphorylation target of Lyn (Fig. 1B). In addition, we demonstrated that rpS3 inhibits the autophosphorylation of Lyn. Collectively, our data demonstrated that rpS3 is a substrate of Lyn kinase. Lyn phosphorylates 166Y and 167Y in rpS3 (Fig. 3C), and Y167 is a critical residue for rpS3 function.

Lyn is activated by various agents that arrest DNA replication or damage (20, 34). Herein, we demonstrated that the phosphorylation of rpS3 was increased by treatment with genotoxic agents and decreased by the addition of a specific inhibitor of Src family tyrosine kinases, such as PP2 (Supplementary Fig. 4). In addition, we showed that phosphorylation was completely abolished by the addition of general tyrosine kinase inhibitors, such as genistein.

Although further studies are needed to clarify the precise role of Lyn phosphorylation on rpS3, we suggest that phosphorylation may affect the repair activity of rpS3.

RpS3 is subjected to various post-translational modifications, such as phosphorylation, sumoylation, methylation, and ubiquitination (8, 11, 13, 33, 36, 37). These modifications also lead to ribosome heterogeneity, resulting in differential translational products (38).

Various modifications and changes in the expression level of the rpS3 protein occur when cells are under various types of stress. We confirmed that phosphorylation of the tyrosine residue of rpS3 was increased in response to genotoxic stress, and p-gp, with the phospho-mimic form of Y167D, was higher than that of the wild type (Fig. 4B). This suggests that p-gp protein synthesis is regulated by the phosphorylation of rpS3 by Lyn via ribosome heterogeneity.

In conclusion, our study identified the interaction between Lyn and rpS3 and the phosphorylation of rpS3 by activated Lyn. This phosphorylation of rpS3 by Lyn increased the translation of p-gp, a gene product of MDR1 (multi-drug resistance gene 1). This phosphorylation of rpS3 by Lyn may result in increased drug resistance in cancer cells.

Further detailed information is provided in the Supplementary Information.

This study was supported in part by Korea University Grant, NRF-2020R1A2C2100803, 2019S1A5A2A03050121 and NRF-2021R1A6A1A10045235.

The authors have no conflicting interests.