ATP11C mutation is responsible for the defect in phosphatidylserine uptake in UPS-1 cells.

Type IV P-type ATPases (P4-ATPases) translocate phospholipids from the exoplasmic to the cytoplasmic leaflets of cellular membranes. We and others previously showed that ATP11C, a member of the P4-ATPases, translocates phosphatidylserine (PS) at the plasma membrane. Twenty years ago, the UPS-1 (uptake of fluorescent PS analogs) cell line was isolated from mutagenized Chinese hamster ovary (CHO)-K1 cells with a defect in nonendocytic uptake of nitrobenzoxadiazole PS. Due to its defect in PS uptake, the UPS-1 cell line has been used in an assay for PS-flipping activity; however, the gene(s) responsible for the defect have not been identified to date. Here, we found that the mRNA level of ATP11C was dramatically reduced in UPS-1 cells relative to parental CHO-K1 cells. By contrast, the level of ATP11A, another PS-flipping P4-ATPase at the plasma membrane, or CDC50A, which is essential for delivery of most P4-ATPases to the plasma membrane, was not affected in UPS-1 cells. Importantly, we identified a nonsense mutation in the ATP11C gene in UPS-1 cells, indicating that the intact ATP11C protein is not expressed. Moreover, exogenous expression of ATP11C can restore PS uptake in UPS-1 cells. These results indicate that lack of the functional ATP11C protein is responsible for the defect in PS uptake in UPS-1 cells and ATP11C is crucial for PS flipping in CHO-K1 cells.


RESULTS AND DISCUSSION
A nonsense mutation is found in ATP11C gene in UPS-1 cells UPS-1, a mutant cell line of CHO-K1 (hereafter CHO), is defective in nonendocytic uptake of NBD-PS ( 20 ). Prior to the experiments described below, we confi rmed that the uptake of NBD-PS is defective in UPS-1 cells. To this end, we assayed the fl ippase activities by incubating CHO and UPS-1 cells in the presence of NBD-PS, -PE, -PC, or -SM at 15°C, followed by extraction with fatty acid-free BSA of fl uorescent phospholipids that were unincorporated or retained in the exoplasmic leafl et of the plasma membrane. As shown in Fig. 1A , B , uptake of NBD-PS was lower in UPS-1 cells than in parental CHO cells, whereas uptake of NBD-PE, -PC, or -SM did not change. These results are consistent with those of an original report ( 20 ).

Flippase assay
Incorporation of NBD-phospholipids was analyzed by fl ow cytometry as described ( 13 ). In brief, CHO-K1 or UPS-1 cells were detached from dishes in PBS containing 5 mM EDTA and then harvested by centrifugation. Cells   The mutation site was located at 205 nucleotides upstream from the junction between exon 19 and exon 20 ( Fig. 2C ). Nonsense-mediated mRNA decay (NMD) is a well-characterized posttranscriptional quality control mechanism to ensure transcription fi delity ( 28 ). NMD can be activated when a nonsense codon appears more than 50-55 nucleotides upstream from an exon-exon junction. Thus, in UPS-1 cells, the ATP11C mRNA might be degraded by the NMD mechanism, and thus the level of ATP11C was dramatically reduced ( Fig. 2A, B ).
We previously showed that expression of ATP11A and ATP11C in HeLa cells increases the PS-and PE-fl ipping activities, although the PE-fl ipping activity of ATP11C is lower than that of ATP11A ( 13 ). Because PE-fl ipping activity was not signifi cantly affected in UPS-1 cells ( Fig. 1A, B ), ATP11C is primarily responsible for PS-fl ipping in CHO cells. The residual PS-fl ipping activity observed in UPS-1 cells might be due to the presence of ATP11A or other fl ipping activities ( Fig. 1A ). ATP11C appears to be a critical protein for PS fl ipping at the plasma membrane in multiple cell types: 1 ) some phenotypes of the ATP11C mutant mice cannot be suppressed by the presence of endogenous ATP11A ( 17,18 ), and 2 ) degradation of ATP11C by caspase is responsible for PS exposure to the outer leafl et of the plasma membrane in apoptotic cells ( 19 ).

Exogenous expression of ATP11C complements the defect in PS-fl ipping activity in UPS-1 cells
Next, we asked whether PS-fl ipping activity can be recovered by exogenous expression of ATP11C in UPS-1 cells. To this end, we expressed C-terminally hemagglutinin (HA) -tagged ATP11C or its ATPase-defi cient Glu-to-Gln mutant (E184Q) ( 13 ) by infection of recombinant retrovirus and subjected the infected cells to the fl ippase assay. ATP11C and the ATP11C(E184Q) mutant were expressed at comparable levels, as confi rmed by immunoblot analysis ( Fig. 3C ). By expressing ATP11C but not ATP11C(E184Q), the PS-fl ipping activity was signifi cantly increased, relative to vector-infected UPS-1 cells ( Fig. 3A, B ), whereas the PE-fl ipping activity was not changed. When a UPS-1 cell line stably expressing ATP11C was examined (clone 16), PS-fl ipping activity was approximating to that in parental CHO cells ( Fig. 4A , B ). We confi rmed ATP11C expression in clone 16 by immunoblot analysis ( Fig. 4C ) and its localization to the plasma membrane by immunofl uorescence ( Fig. 4D ). Thus, in CHO cells, ATP11C is crucial for PS fl ipping, but not PE fl ipping. Because there was a subtle, but signifi cant, increase in the PE-fl ipping activity in stably ATP11C-expressing UPS-1 cells ( Fig. 4A ), exogenous ATP11C may also fl ip NBD-PE as we described previously ( 13 ). lines) in the presence of the indicated NBD-lipids. C: Expression level of ATP11C in UPS-1 cells (clone 16) was analyzed by immunoblotting with antibodies against HA and ␤ -tubulin (as an internal control). D: Cells were fi xed, permeabilized, and immunostained with antibody against Na/K-ATPase (a marker for the plasma membrane, PM) and anti-HA antibody followed by Alexa488-conjugated anti-rabbit and Cy3-conjugated anti-rat antibodies, respectively. Bar indicates 10 m. E: RT-PCR was performed using total RNA isolated from indicated cells. A primer set targeting the 3 ′ UTR of the Chinese hamster ATP11C was designed. ATP8B1, ATP8B2, and ATP10A preferentially fl ip NBD-PC ( 13,14 ). We also previously showed that CDC50A is required for the plasma membrane localization of ATP11A and ATP11C ( 12 ). Knockdown by RNA interference or clustered regularly interspaced short palindromic repeat/ CRISPR-associated protein-9 nuclease (CRISPR/Cas9) knockout of CDC50A dramatically decreases PS-fl ipping activity in several cell types (our unpublished observations) ( 19 ). Moreover, ATP11C defi ciency also decreases PSfl ipping activity (17)(18)(19). Therefore, we hypothesized that the defect in the uptake of PS in UPS-1 cells might result from a defect in plasma membrane-localizing PS-fl ipping P4-ATPases (ATP11A and ATP11C) or a defect in CDC50A, which is required for delivery of P4-ATPases to the plasma membrane. To examine this hypothesis, we fi rst performed RT-PCR to measure the mRNA levels of ATP11A, ATP11C, and CDC50A in CHO and UPS-1 cells ( Fig. 2A ). All primer sets were designed to include at least one predicted intron to avoid amplifi cation from genomic DNA. As shown in Fig. 2A , the mRNA level of ATP11C was substantially lower in UPS-1 cells (U) than in CHO cells (C). We confi rmed the low level of ATP11C mRNA in UPS-1 cells using three independent primer sets ( Figs. 2A and 4E ). By contrast, the mRNA levels of ATP11A and CDC50A were comparable between UPS-1 and CHO cells ( Fig. 2A ). We also examined the mRNA levels of ATP11A, ATP11C, and CDC50A by qRT-PCR. As shown in Fig. 2B , the mRNA level of ATP11C was much lower in UPS-1 cells than in the parental CHO cells. By contrast, the ATP11A and CDC50A mRNA levels in UPS-1 cells were comparable to those in CHO cells.
Next, we examined a possibility that the coding sequence of ATP11C is changed in UPS-1 cells. To this end, we cloned nine cDNA fragments for ATP11C from UPS-1, which altogether cover the entire coding sequence of ATP11C and its C-terminal splicing variant, and sequenced at least four independent clones for each fragment. Interestingly, we found a nonsense mutation in the region covered by exon 19 ( Fig. 2C ). Direct sequencing of a chromosomal DNA region (362 bp) encompassing exon 19 and the preceding intron indicated that the ATP11C gene in UPS-1 cells carried a G to A homozygous mutation in exon 19, resulting in a change in the Trp664 codon (TGG) to a nonsense codon (TGA) ( Fig. 2C, D ). Although chromosomal heterogeneity has been observed in CHO cell lines ( 26,27 ), the raw data of direct sequencing of the chromosomal DNA showed a single peak at the mutation site ( Fig. 2C ), indicating a homozygous mutation in the ATP11C gene. We cannot exclude a possibility that the locus of ATP11C gene in UPS-1 cells is hemizygous. These results indicate that the defect of PS uptake in UPS-1 cells is caused by the lack of the functional ATP11C protein, although we could not examine the ATP11C protein level because an antibody, In order to exclude possible contamination of parental CHO cells in the ATP11C-expressing UPS-1 cells, we performed RT-PCR analysis using a set of primers targeting the 3 ′ -UTR of the Chinese hamster ATP11C. The endogenous mRNA level of ATP11C did not change among the exogenous ATP11C-or ATP11C(E184Q)-expressing UPS-1 cells ( Fig. 4E ). Thus, the recovery of PS-fl ipping activity in UPS-1 cells is due to the exogenous expression of ATP11C rather than contamination of parental CHO cells.
Based on these results, we conclude that the defect in PS uptake observed in UPS-1 cells is ascribed to the lack of functional ATP11C. It is likely that ATP11C is a major P4-ATPase involved in PS fl ipping in CHO cells. ATP11C-defi cient mice exhibit anemia, hyperbilirubinemia, abnormal differentiation of B cells, and hepatocellular carcinoma ( 17,18,29 ). Hopefully, UPS-1 cells will be useful for understanding the physiological and pathophysiological roles of ATP11C.