HT1080 human fibrosarcoma (Japanese Collection of Research Bioresources Cell Bank, Osaka, Japan), 293T (RIKEN BioResource Center, Tsukuba, Japan), A549 human lung adenocarcinoma (RIKEN BioResource Center) (37-40), PANC1 human pancreatic adenocarcinoma (RIKEN BioResource Center) and MDA-MB-231 human breast adenocarcinoma cell lines, which was gifted by Professor Masakazu Toi (Kyoto University Graduate School of Medicine, Kyoto, Japan), were all cultured in Dulbecco’s modified Eagle’s medium (DMEM; Nissui Pharmaceutical Co., Ltd, Tokyo, Japan) supplemented with 6% (v/v) fetal bovine serum (Cosmo Bio Co. Ltd., Tokyo, Japan), 100 U/ml penicillin G, 100 mg/l kanamycin, 600 mg/l L-glutamine, and 2.25 g/l NaHCO₃ at 37°C in a humidified incubator with 5% CO₂. The LoVo human colon adenocarcinoma cell line was cultured in RPMI-1640 medium (Nissui Pharmaceutical Co., Ltd) supplemented with 10% (v/v) fetal bovine serum, 105 U/ml penicillin G, 105 mg/l kanamycin, 314 mg/l L-glutamine, and 2.25 g/l NaHCO₃ at 37°C in a humidified incubator with 5% CO₂.

Total RNA was extracted from THP1 cells using RNA extraction buffer (38% (w/w) Phenol, 0.8 M guanidine thiocyanate, 0.4 M ammonium thiocyanate, 0.1 M sodium acetate, 5% (v/v) glycerol), and cDNA was prepared from 2 µg of total RNA with the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems; Thermo Fisher Scientific, Inc.). The resulting cDNA was used for polymerase chain reaction (PCR) amplification with PrimeSTAR Max DNA Polymerase (Takara Bio Inc., Shiga, Japan). The sequences of the primers used to amplify the human Rspo2 cDNA, were: 5′-GAACCCTCCAGTTCCTAGACTTTGA GAGGCGTCTC-3′ (forward) and 5′-ATTGGTTAGCTCTGT CTGTAGCTAGGAAGACGCTG-3′ (reverse). PCR was performed for 35 cycles with an annealing temperature of 60°C. The PCR product contained a 5′ untranslated region (UTR), the coding sequence for the Rspo2 gene, and a 3′UTR. To obtain the coding region of the Rspo2 gene, a second PCR was performed with the same reagents/conditions as the first PCR, and using the following primers: 5′-TTTTCTCGAGATGCAGTTTCGCCTTTTCTCCTTTG-3′ (forward) and 5′-TTTTGCGGCCGCTTGGTTAGCTCTGTCTGTAGC-3′ (reverse). The obtained cDNA was subcloned into pCI-neo (Promega Corporation, Madison, WI, USA) and CSII-CMV-MCS-IRES2-Bsd vectors (RIKEN BioResource Center). To introduce the C-terminal myc-his6 tag, PCR was performed with primers that had Myc and his6 codons. The sequences of the tags were as follows: Myc, 5′-GAACAAAAACTCATCTCAGAAGAGGATCTG-3′; and his6, 5′-CATCATCACCATCACCAT-3′. Certain tryptophan residues in Rspo2 were substituted with alanine residues by PCR site-directed mutagenesis using the overlap extension technique with PrimeSTAR Max DNA Polymerase. The sequences of primers that were used for the mutagenesis, the number of cycles, and the annealing temperatures were as follows: W150A, 5′-GTGAAGTTGGTCATGCGAGCGAATGGGGAA-3′ (forward) and 5′-GTTCCCCATTCGCTCGCATGACCAACTTCA-3′ (reverse), 35 cycles, 63°C; and W153A, 5′-GAGCGAAGCGGGAACTTG-3′ (forward) and 5′-CAAGTTCCCGCTTCGCTC-3′ (reverse), 35 cycles, 63°C. The resulting cDNAs were cloned into the XhoI/NotI restriction sites of pCI-neo and CSII-CMV-MCS-IRES2-Bsd vectors for mammalian cell expression.

Green fluorescent protein (GFP) cDNA was amplified from pAcGFP1-N1 (Takara Bio Inc.) with PrimeSTAR Max DNA Polymerase and subcloned into CSII-CMV-MCS-IRES2-Bsd. The sequences of the gene-specific primers, the number of cycles, and the annealing temperature were as follows: 5′-TT TTCTCGAGATGGTGAGCAAGGGCGCCGAGCTG-3′ (forward) and 5′-TTTTGCGGCCGCTCACTTGTACAGCTCATCCATGC-3′ (reverse), 35 cycles, 63°C. The resulting cDNA was cloned into the XhoI/NotI restriction sites of CSII-CMV-MCS-IRES2-Bsd for mammalian cell expression.

HT1080 cells were grown in a 6-well plate at 50-60% confluence and transfected with 2 µg pCI-neo vectors using the Lipofectamine 3000 transfection kit (Thermo Fisher Scientific, Inc.), according to the manufacturer’s protocol, in order to establish stable cell lines that expressed wild-type or mutant Rspo2-myc-his6. At 72 h post-transfection, cells were treated with 400 µg/ml G418 (Wako Pure Chemical Industries, Ltd., Osaka, Japan) for selection of stable cells. Clonal cells that expressed high levels of myc-his6-tagged wild-type and double mutant W150A/W153A (termed herein 2WA) Rspo2 were designated HT1080-Rspo2-MH and HT1080-Rspo2/2WA-MH cells, respectively. Mixed population of W150A-expressing cells were designated HT1080-Rspo2/W150A-MH. Cells that were transfected with the empty pCI-neo vector were termed HT1080-neo (41).

To establish Rspo2-overexpressing A549, PANC1, MDA- MB-231, and LoVo cell lines, the CSII-CMV-MCS-IRES2-Bsd plasmids were transfected into 293T cells using a Lentivirus High Titer Packaging Mix (Takara Bio Inc.) for lentivirus production. After 6 h of transfection, the cells were washed, and fresh medium was added. After an additional 48 h of culture, the conditioned media, containing lentivirus, were collected, and each cell line was infected with the lentivirus media. Following infection, A549, PANC1, MDA-MB-231 and LoVo cells were selected with 20, 10, 7.5 and 15 µg/ml Blasticidin S (Wako Pure Chemical Industries, Ltd.), respectively. Cells that expressed high levels of myc-his6-tagged wild-type and 2WA mutant Rspo2, and GFP which was used for control expression, were used for experiments. Rspo2-overexpressing A549 cells were followed by limiting dilution method to establish clonal cells.

For the western blot analysis, a slightly modified version of a previously described method was used (42-44). Cells were cultured, and then lysed in buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 1% sodium deoxycholate, and 1 mM PMSF) at 4°C with sonication. The lysates were centrifuged at 15,300 × g for 10 min, and the amount of protein in each lysate was measured by Coomassie Brilliant Blue (CBB) G-250 staining (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Loading buffer (350 mM Tris-HCl pH 6.8, 30% glycerol, 0.012% bromophenol blue, 6% SDS, and 30% 2-mercaptoethanol) was added to each lysate. The samples were then boiled for 5 min and electrophoresed on 12.5% SDS-polyacrylamide gels. Proteins were transferred to polyvinylidene fluoride membranes and, following blocking with 5% skim milk at room temperature for 30 min, immunoblotted with anti-c-Myc (9E10 hybridoma culture supernatant; Developmental Studies Hybridoma Bank, Iowa City, IA, USA) or anti-α-tubulin (cat. no. T5168; Merck KGaA, Darmstadt, Germany). Signals were detected with enhanced chemiluminescence using Western Lightning Plus-ECL reagent (PerkinElmer, Inc., Waltham, MA, USA) or Immobilon Western Chemiluminescent HRP substrate (Merck KGaA), and exposed to RX-U films (Fujifilm, Tokyo, Japan) in a dark room. The 9E10 hybridoma culture supernatant was diluted 1:50 in a washing buffer (19.8 mM Tris-HCl pH 7.6, 137 mM NaCl, and 0.1% Tween-20) containing 5% skim milk.

Cells were washed with PBS twice, and then cultured in serum-free DMEM for 24 h with or without 50 µg/ml soluble heparin. The conditioned media was collected, and cell lysates were prepared as described above. The conditioned media was concentrated using Ni-nitrilotriacetic acid (Ni-NTA) agarose (Roche Diagnostics, Mannheim, Germany) for 2 h at 4°C. Then, the Ni-NTA agarose was collected, washed with PBS twice, and eluted with buffer I (900 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH2PO4, and 500 mM imidazole). After being eluted from the Ni-NTA agarose, the amount of protein in each sample of conditioned media was estimated from the total amount of protein in each set of cell lysates by CBB staining. Based on the inverse ratio of each cell lysate concentration, additional buffer I was added to each sample to establish uniform concentrations among the conditioned media. Loading buffer was added to the conditioned media and cell lysates, and boiled for 5 min. Then, the proteins were separated by 12.5% SDS-PAGE and analyzed by immunoblotting with anti-c-Myc and anti-α-tubulin antibodies, as aforementioned.

To purify recombinant Rspo2, a slightly modified version of a previously described method was used (35). Clonal HT1080 or A549 cells that expressed high levels of wild-type Rspo2-MH were established and cultured for 24 h in serum-free medium that contained 25 µl/ml heparin sepharose 6 fast flow (GE Healthcare Life Sciences, Little Chalfont, UK). Then, the heparin sepharose was collected, washed with PBS twice, and eluted with buffer II (900 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, and 5 mM imidazole). The eluted samples were concentrated using Ni-NTA agarose for 2 h at 4°C. The Ni-NTA agarose was collected, washed with buffer II, and eluted with buffer I by incubating for 5 min at room temperature followed by boiling for 5 min. The eluted samples were further concentrated on a Vivaspin 500 (3000 MWCO; Sartorius AG, Göttingen, Germany). The eluates were electrophoresed on an SDS-polyacrylamide gel, and the protein bands were visualized by CBB staining.

To identify the C-mannosylation sites, an ultrasensitive Q-Exactive nanoLC-MS/MS system was used. Purified Rspo2 samples were subjected to 12.5% SDS-PAGE. After CBB staining, the visible band was excised and de-stained. Following reduction and S-carboxymethylation with iodoacetic acid, in-gel digestion was performed using trypsin (TPCK-treated; Worthington Biochemical, Worthington, OH, USA) and an endoproteinase Asp-N (sequencing-grade; Roche Diagnostics, Basel, Switzerland). The digestion mixture was separated on a nanoflow LC (Easy nLC; Thermo Fisher Scientific, Inc.) using a nano-electrospray ionization spray column (NTCC analytical column; C18, φ75 µm × 100 mm, 3 µm; Nikkyo Technology Co., Ltd., Tokyo, Japan) with a linear gradient of 0-66% buffer B (100% acetonitrile and 0.1% formic acid) and a flow rate of 300 nl/min over 20 min, coupled on-line to a Q-Exactive mass spectrometer (Thermo Fisher Scientific, Inc.) that was equipped with a nanospray ion source that had no gas assistance. The mass spectrometer was operated in positive-ion mode, and the MS/MS spectra were acquired using an inclusion list that contained double or triple charged peptide ions with or without the C-mannosylation (¹⁴⁶EVGHWSEWGTCSR¹⁵⁸ unmodified, m/z=796.3333 or 531.2247; mono-C-mannosylation, m/z=877.3598 or 585.2423; di-C-mannosylations, m/z=958.3862 or 639.2599). The MS/MS chromatograms of the y5 ion (m/z=581.2350±10 ppm) of the listed peptides and their MS/MS spectra were constructed using Qual Browser Thermo Xcalibur 3.1.66.10 (Thermo Fisher Scientific, Inc.).

Cells were grown on coverslips in a 6-well plate, washed twice with PBS, fixed in 4% paraformaldehyde for 10 min, and permeabilized with 0.1% Triton X-100 for 10 min at room temperature. After being blocked with 3% bovine serum albumin for 30 min, the cells were incubated with anti-c-Myc at room temperature for 1 h. Alexa Fluor488-conjugated anti-mouse immunoglobulin (Ig)G (cat. no. A11029; Thermo Fisher Scientific, Inc.) was used as the secondary antibody at room temperature for 1 h. To determine the localization of the Golgi apparatus and the endoplasmic reticulum (ER), cells were incubated with a rabbit polyclonal anti-Golgi reassembly stacking protein 1 (GRASP65; cat. no. sc-30093; Santa Cruz Biotechnology, Inc., Dallas, TX., USA) and a rabbit polyclonal anti-calnexin (cat. no. 2433; Cell Signaling Technology, Inc., Danvers, MA, USA) antibody at room temperature for 1 h each. Alexa Fluor568-conjugated anti-rabbit IgG (cat. no. A11036; Thermo Fisher Scientific, Inc.) was used as the secondary antibody at room temperature for 1 h. After being washed twice with PBS, the cells were incubated with 2 µg/ml Hoechst 33258 (Polysciences, Inc., Warrington, PA, USA) for 10 min to stain the nuclei. Cells were then washed again three times with PBS and observed under a FLUOVIEW FV10i confocal laser-scanning microscope (Olympus Corporation, Tokyo, Japan). The 9E10 hybridoma culture supernatant towards c-Myc was diluted 1:4, while the anti-GRASP65 and anti-calnexin antibodies were diluted 1:100 in PBS containing 3% bovine serum albumin. The secondary antibodies were diluted 1:500 in PBS containing 3% bovine serum albumin.

To assess Wnt/β-catenin signaling activity, a TOP/FOP luciferase reporter system was used. Each cell (4×10⁴ cells/well) was seeded in 24-well plates and cultured overnight. Using the Lipofectamine 3000 transfection kit, the cells were transiently transfected with 800 ng canonical Wnt signaling reporter Super 8×TopFlash or mutant reporter Super 8×FopFlash (cat. nos. 12456 and 12457; Addgene, Watertown, MA, USA) (45) and 20 ng phRL-TK (Promega Corporation) in the presence of 30% Wnt3a-conditioned medium from L-Wnt3a cells (cat. no. CRL-2647; American Type Culture Collection, Manassas, VA, USA), as previously described (46). After 24 h, the cells were lysed, and firefly and Renilla luciferase activities were measured using infinite 200Pro multifunctional microplate reader (Tecan Group Ltd., Männedorf, Switzerland). TOPFlash and FOPFlash activities were normalized to that of Renilla.

HT1080 and A549 cells (2×10⁵ cells/well) were seeded into 24-well plates and cultured for 24 h to ~90% confluence, and then wounded using a yellow pipette tip. After being washed twice with PBS to remove floating cells, the cells were cultured in serum-free DMEM for 12 h (HT1080) or 96 h (A549). Photographs were captured of four independent areas, and the migrated areas were quantified using ImageJ 1.51 software (National Institutes of Health, Bethesda, MD, USA) (47).

To assess migration, 6.5-mm transwell inserts with 8.0 µm pore polycarbonate membranes were used (cat. no. 3422; Corning, Inc., Corning, NY, USA). Cells were washed twice with serum-free DMEM to remove serum and seeded into the upper chambers at 2×10⁴ cells/chamber; the lower chambers were filled with growth medium. The cells were incubated for 3 h, and then non-migrated cells on the upper surfaces of the membranes were carefully removed with a cotton swab. Migrated cells were fixed and stained using Diff-Quick solution (Sysmex Corporation, Kobe, Japan). Photographs were capture from four independent areas and migration was quantified as the average number of migrated cells per area (48).

Data were expressed as the mean ± standard deviation of three independent experiments. Statistical analyses were performed using one-way analysis of variance with Tukey’s post-hoc test. Analysis was performed with GraphPad Prism 8 statistical software (GraphPad Software, Inc., La Jolla, CA, USA). P<0.05 was considered to indicate a statistically significant difference.

Human Rspo2 has two consecutive C-mannosylation consensus sequence sites from Trp¹⁵⁰ (W¹⁵⁰-S-E-W¹⁵³-G-T-C) in the TSR1 domain (Fig. 1A). To determine whether these tryptophans are C-mannosylated, a Rspo2-overexpressing HT1080 clonal cell line was established (HT1080-Rspo2-MH). Because the R-spondin family proteins bind to cell surface heparan sulfate proteoglycans (14,49), to confirm that Rspo2 is secreted in the conditioned medium, cells were treated with heparin. As a result, the levels of secreted Rspo2 were increased (Fig. 1B). Therefore, sequential affinity chromatography was performed using heparin sepharose and Ni-NTA agarose to purify the recombinant Rspo2 (Fig. 1C).

The resulting purified sample was digested with trypsin and Asp-N, and the subsequent mixture of peptides was subjected to LC-MS. LC-MS analysis indicated a triply charged form of the peptide ¹⁴⁶EVGHWSEWGTCSR¹⁵⁸, corresponding to un-, mono-, and di-mannosylated forms (Fig. 2A). The expected values of m/z for each peptide are 531.22, 585.24 and 639.26, respectively, and each peptide was detected at the expected m/z, suggesting that Rspo2 exists in un-, mono-, and di-mannosylated forms. In addition, to identify the C-mannosylation sites of Rspo2, LC-MS/MS analysis was performed on the aforementioned peptides (Fig. 2B). The un- and mono-mannosylated peptides had the same spectral pattern until the y8 ions. This result suggested that one mannose was modified in ¹⁴⁶EVGHW¹⁵⁰. Based on previous studies, there is no indication that any hexose modification occurs in those peptides, except at Trp¹⁵⁰. Thus, it is most likely that Trp¹⁵⁰ is mannosylated. Then, the di-mannosylated peptide was identified from the y1 ion to the y5 ion by LC-MS/MS analysis, indicating that two mannoses are modified in ¹⁴⁶EVGHWSEW¹⁵³. Considering that un- and mono-mannosylated peptides were identified from the y1 ion to the y8 ion, it was concluded that Trp¹⁵³ was modified. In addition, Trp¹⁵⁰ was likely modified by one mannose residue; one- and two-hexose glycosylation at Glu, Val, Gly, His, and Ser have not been reported. Furthermore, to verify whether Rspo2 is also C-mannosylated in other cell types, a Rspo2-overexpressing clonal cell line was generated in A549 cells (A549-Rspo2-MH) and recombinant Rspo2 was purified. As presented in Fig. S1, Rspo2 was also C-mannosylated in A549 cells. Altogether, Rspo2 was demonstrated to exist in three forms: un-mannosylated, Trp¹⁵⁰-mannosylated, and Trp¹⁵⁰- and Trp¹⁵³-mannosylated. Therefore, it was concluded that both Trp¹⁵⁰ and Trp¹⁵³ were C-mannosylation sites of Rspo2 in human tumor cells.

Glycoproteins often regulate their secretion by glycosylation. Because Rspo2 is secreted into the extracellular environment, we examined whether C-mannosylation of Rspo2 regulates its secretion. To this end, a C-mannosylation-defective mutant Rspo2-overexpressing HT1080 clonal cell line was generated (HT1080-Rspo2/2WA-MH) by substituting the two Trps, Trp¹⁵⁰ and Trp¹⁵³, with Ala. This cell line was confirmed by western blot analysis (Fig. 3A).

Firstly, western blotting was performed to compare secretory levels between wild-type and 2WA Rspo2. The results indicated that the secretory levels of the C-mannosylation-defective 2WA mutant Rspo2 decreased compared with wild-type Rspo2 (Fig. 3B). Because secretory proteins must be able to move from the ER to the Golgi apparatus (50), it was hypothesized that the defect in C-mannosylation of Rspo2 may have affected its intracellular localization, resulting in less secretion and increased accumulation of Rspo2/2WA in the cell lysate. Using antibodies against calnexin, an ER marker, and GRASP65, a marker for the Golgi apparatus, immunofluorescence staining was performed in order to examine the localization of wild-type and 2WA localization in the ER or Golgi apparatus. As expected, whereas wild-type Rspo2 localized to the Golgi apparatus, Rspo2/2WA accumulated in the ER (Fig. 3C and D). These data indicated that C-mannosylation of Rspo2 was important for its transportation from the ER to the Golgi apparatus.

R-spondin proteins, including Rspo2, have been reported to be agonists of Wnt/β-catenin signaling (2,4-8). Because N-glycosylation of Wnt3a is necessary for its function and subsequent signaling (51), it was examined whether C-mannosylation of Rspo2 affects its agonistic activity in the Wnt/β-catenin signaling pathway. To measure Wnt/β-catenin signaling activity, a luciferase reporter assay was used. TOPFlash, which has seven tandem repeats of the TCF/LEF binding motifs upstream of the firefly luciferase gene, was transfected into each Rspo2-overexpressing HT1080 clonal cell line, and luciferase activity was measured. Transfected cells were cultured in Wnt3a-containing medium for 24 h. Although both wild-type and 2WA Rspo2-overexpressing cell lines displayed increased TOPFlash activity compared with control cells, the activity in wild-type cells was increased by 27.3-fold compared with only 7.8-fold in 2WA cells (Fig. 3E). Thus, the C-mannosylation-defective Rspo2/2WA had reduced TOPFlash activity compared with the wild-type Rspo2.

To confirm the role of C-mannosylation on Rspo2 functions in cancer, three additional human tumor cell lines were used, namely A549, PANC1 and MDA-MB-231. First, wild-type and 2WA mutant Rspo2-overexpressing A549, PANC1 and MDA-MB-231 cell lines were established, and then the secretory levels of these proteins were assessed. Notably, the secretory levels of the C-mannosylation-defective 2WA mutant Rspo2 were increased compared with wild-type Rspo2 in A549, PANC1 and MDA-MB-231 cells (Fig. 4A). Next, the effect of C-mannosylation on Rspo2-mediated activation of Wnt/β-catenin signaling was examined in these cell lines. The results demonstrated that the TOPFlash activity of 2WA mutant Rspo2-overexpressing cells was reduced compared with wild-type Rspo2-overexpressing cells in PANC1 and MDA-MB-231 cell lines (Fig. 4B), similar to the aforementioned results in HT1080 cells. No difference in TOPFlash activity was observed, however, in Rspo2/wild-type- and Rspo2/2WA-overexpressing A549 cells (Fig. 4B). Therefore, the present results suggested that C-mannosylation may be important for the Wnt/β-catenin signaling-enhancing activity of Rspo2 in HT1080, PANC1 and MDA-MB-231 cells.

Because, Rspo2 contributes to cancer progression (18-20), it was next examined whether C-mannosylation of Rspo2 may regulate cancer cell migration in vitro. Confluent Rspo2-overexpressing HT1080 clonal cells were subjected to wound-healing assays, and 12 h post-scratching, wound closure areas were quantified. As presented in Fig. 5A, the overexpression of wild-type Rspo2 significantly increased cell migration compared with control cells, whereas overexpression of 2WA mutant Rspo2 had little effect on migration. Similar results were obtained with transwell migration assays (Fig. 5B).

To confirm that C-mannosylation of Rspo2 may mediate migration of other cancer cell lines, Rspo2-overexpressing A549 clonal cells were used. Using a wound-healing assay, cell migration ability was demonstrated to be increased in A549-Rspo2/wild-type, compared with control cells, but was unchanged in A549-Rspo2/2WA cells (Fig. 5C). These results indicate that C-mannosylation of Rspo2 regulated the migration of HT1080 and A549 cells. In addition, the present findings demonstrated that a defect in the C-mannosylation of Rspo2 abrogated its ability to mediate cancer cell migration.

Although the present results demonstrated that Rspo2 is C-mannosylated at Trp¹⁵⁰ and Trp¹⁵³, Rspo2 was observed in un-mannosylated (4%), Trp¹⁵⁰-mannosylated (94%), and Trp¹⁵⁰- and Trp¹⁵³-mannosylated (2%) forms (Fig. 2A). Therefore, it can by hypothesized that single C-mannosylation at Trp¹⁵⁰ may be more important for Rspo2 functions, compared with the di-mannosylated form. Therefore, Trp¹⁵⁰-C-mannosylation-defective mutant Rspo2-overexpressing HT1080 cells were generated (HT1080 Rspo2/W150A-MH) and compared with the other Rspo2-overexpressing HT1080 cells for their ability to activate Wnt/β-catenin signaling. The results demonstrated that Rspo2/W150A had reduced TOPFlash activity compared with the wild-type Rspo2, but exhibited comparable activity with the Rspo2/2WA mutant (Fig. S2). These results suggested that C-mannosylation at Trp¹⁵⁰ may be more important than Trp¹⁵³ for Rspo2 functions.