Functional characterization of SLC26A3 c.392C>G (p.P131R) mutation in intestinal barrier function using CRISPR/CAS9-created cell models

Background Congenital chloride diarrhea (CCD) in a newborn is a rare autosomal recessive disorder with life-threatening complications, requiring early diagnostics and treatment to prevent severe dehydration and infant mortality. SLC26A3 rs386833481 (c.392C>G; p.P131R) gene polymorphism is an important genetic determinant of CCD. Here, we report the influence of the non-synonymous SLC26A3 variant rs386833481 gene polymorphism on the function of the epithelial barrier and the potential mechanisms of these effects. Results We found that P131R-SLC26A3 increased dysfunction of the epithelial barrier compared with wild type SLC26A3 in human colonic Caco-2 and mouse colonic CMT-93 cells. When P131R-SLC26A3 was subsequently reverted to wild type, the epithelial barrier function was restored similar to wild type cells. Further study demonstrated that variant P131R-SLC26A3 disrupts function of epithelial barrier through two distinct molecular mechanisms: (a) decreasing SLC26A3 expression through a ubiquitination pathway and (b) disrupting a key interaction with its partner ZO-1/CFTR, thereby increasing the epithelial permeability. Conclusion Our study provides an important insight of SLC26A3 SNPs in the regulation of the epithelial permeability and indicates that SLC26A3 rs386833481 is likely a causative mutation in the dysfunction of epithelial barrier of CCD, and correction of this SNP or increasing SLC26A3 function could be therapeutically beneficial for chronic diarrhea diseases.


Introduction
Globally, diarrhea is a leading cause of death among all ages (1.31 million deaths in 2015), as well as a leading cause of the disability-adjusted life years (DALYs) because of its disproportionate impact on young children (71.59 million DALYs, range from 66.44 million to 77.21 million), especially in developing countries [1]. The etiology and underlying molecular pathogenesis of diarrhea are complicated and not fully understood though it is increasingly recognized that genetic predisposition may play a significant role in an individual's susceptibility to chronic diarrhea [2,3]. Understanding the genetic contributions to disease biology can help identify at-risk individuals, guide more effective personalized treatment approaches, and illuminate new targets and pathways for therapeutic development and intervention.
Chronic diarrhea diseases can be classified into inflammatory, malabsorptive, osmotic, secretory and motility disorders [4]. Congenital chloride diarrhea (CCD-OMIM 214700) is an autosomal recessive disorder characterized by life-long, severe diarrhea with intestinal Cl − malabsorption. Postnatal clinical diagnosis is based on the presentation of dehydration and failure to thrive in the setting of hypokalemic metabolic alkalosis, with acidic stool pH and elevated stool chloride (> 90 mM) measured after normalization of systemic volume status and serum electrolytes. Untreated disease leads to chronic systemic volume depletion, nephrocalcinosis and impaired renal function sometimes progressing to end-stage renal disease. Additional clinical manifestations later in life have included intestinal inflammation, hyperuricemia, inguinal hernia, and impaired male fertility [3].
CCD is caused by mutations in the gene encoding SLC26A3 [3,5,6], with 21 exons spanning ~ 38 kb on chromosome7q31.1. SLC26A3 is a Cl − /HCO 3 − exchanger that contributes to intestinal fluid absorption and enterocyte acid/base balance [7,8], which has been unequivocally demonstrated to be a Cl − /HCO 3 − exchanger with a 2:1 transport stoichiometry [9,10]. The transport function of SLC26A3 is thought to play an important role in Cl − absorption and HCO 3 − secretion in the colon and perhaps in the pancreas [11,12].
Single nucleotide polymorphisms (SNPs) are the most common type of genetic variation among humans. Some SNPs have been proven to be directly associated with human diseases. SNPs that lead to amino acid substitutions in proteins are of particular interest because they are responsible for nearly half of the known genetic variations related to inherited diseases in human [13,14]. Previous studies have found a number of SNPs in SLC26A3, including the damaging missense mutation rs386833481 (c.392C>G; p.P131R), from patients with CCD [3]. Whether any of these SNPs is a causative mutation has been unproven. Furthermore, patients with diarrhea associated with inflammatory bowel disease (IBD), either ulcerative colitis (UC) or Crohn's disease (CD), exhibit reduced SLC26A3 expression [15,16]. Xiao et al. [17] previously identified that SLC26A3 deficiency is associated with the absence of a firmly adherent mucus layer and mucus barrier impairment in mice. This change in mucus layer renders SLC26A3 −/− mice susceptible to dextran sulfate sodium (DSS)-induced colitis. These studies imply reduced SLC26A3 expression, leads to increased dysfunction of the epithelial barrier. However, little is known about whether these genetic variants could lead to the dysfunction of the epithelial barrier and about the potential mechanisms of these effects.
SLC26A3 interacts with cystic fibrosis transmembrane conductance regulator (CFTR) and they reciprocally regulate each other through binding of the R domain of CFTR and the STAS domain of SLC26A3 [18,19]. Meanwhile, there is an increased permeability of the small intestine both in CF humans and in CF mice (Cftr knockout mouse model) [20], and CFTR interacts with ZO-1 to regulate tight junctions [21]. The importance of both SLC26A3 and CFTR functions in the physiology of tight junctions (TJs) is supported by their molecular interaction. These findings prompted us to study whether SNPs in SLC26A3 disturb its normal interaction with ZO-1/ CFTR and increase intestinal epithelial permeability.
In this study, we dissected the functional consequences of the P131R variant and SLC26A3 expression level on intestinal epithelial permeability and functionally characterized the interaction between SLC26A3 SNP encoded protein or WT SLC26A3 protein and ZO-1/CFTR in human colonic Caco-2 cells. Further, we evaluated the therapeutic potential of correcting this SNP mutation of SLC26A3 by testing the function of epithelial barrier of Caco-2 cells. Our study provides solid evidence that SLC26A3 SNP rs386833481 (c.392C>G; p.P131R) is a likely causative mutation in the dysfunction of epithelial barrier of CCD. Our biochemical study has also provided a lead to the underlying molecular mechanism.

Construction of the P131R-SLC26A3 genetic variant
Based on analysis of public databases, we identified an exonic SNP in the human SLC26A3 gene from patients with CCD. The SLC26A3 genetic variant (rs386833481) changes the DNA from a cytosine (C) to a guanine (G) base and an amino acid change from Proline (P) to Arginine (R) at its amino acid sequence position 131 (Fig. 1a). In this study, the SLC26A3 rs386833481 is referred to as P131R-SLC26A3. The P131R mutation was predicted to be "deleterious" and "damaging" by Provean (score − 7.32; cutoff: − 2.5) and Sift (score 0.001; cutoff: 0.05) web server tools for predicting the functional effect of amino acid substitutions. Amino acid residue P131 resides within the polytopic transmembrane domain of SLC26A3 (Fig. 1b). Although the membrane domains of SLC26 polypeptides are of unknown topographical disposition, hydropathy profiling has predicted a location for P131 at the putative transmembrane span3. This residue is conserved among SLC26A3 orthologs in primates, rodents, goat, sheep, dog, horse, rabbit and zebrafish (Fig. 1c). Until now, there is little information and indication of this SLC26A3 genetic variant being linked to human diarrhea susceptibility. To further explore whether the SLC26A3 genetic variant alters its function and expression, we adapted an HDRmediated modification strategy using the CRISPR/ Cas9 system in both human (Caco-2, Fig. 1d) and murine colonic epithelial (CMT-93, Fig. 6a) cell lines. After the SLC26A3 c.392C>G (p.P131R) mutation was generated in both cell lines, they went though a weeklong puromycin selection for a single clone that carries the exact mutation. TaqMan SNP Genotyping (Fig. 1e) and Sanger Sequencing (Fig. 1f ) both were used to validate the accurate construction of P131R-SLC26A3. These results indicated that we successfully recreated SLC26A3 SNP rs386833481 (c.392C>G; p.P131R), providing the foundation for functional analysis of its effect on intestinal epithelial cell permeability.

P131R-SLC26A3 weakens the epithelial barrier and augments TNF-α-induced damage
To determine the role of P131R-SLC26A3 in epithelial barrier function, we also upregulated SLC26A3 expression by transfecting Caco-2 cells with either SLC26A3-pCS6 or the empty vector pCS6 control (Fig. 2b, c). A GFP expression vector was used to monitor transfection efficiency (Fig. 2a). Previous work showed that SLC26A3 expression is down-regulated in a TNF-α overexpressing mouse model and that TNF-α can affect the expression of tight junction proteins [22,23]. We therefore measured transepithelial electric resistance (TEER) in P131R-SLC26A3, SLC26A3-overexpressing and normal Caco-2 cells. Upon TNF-α treatment, P131R-SLC26A3 cells showed significantly lower TEER values compared with normal cells. Consistent with these results, the TEER value in SLC26A3-overexpressing (SLC26A3 OE ) cells was higher than that in control cells (Fig. 2d). The maximum TEER induction are 0.85 ± 0.006 vs. 0.94 ± 0.003 in P131R-SLC26A3 vs. WT (P < 0.001), as well as 0.99 ± 0.003 vs. 0.95 ± 0.001 in SLC26A3 OE vs. control cells (P < 0.001), respectively. There was no significant difference between WT and control cells (P = 0.70) (Fig. 2e). These results indicated that P131R-SLC26A3 weakened the epithelial barrier and augmented TNF-α-induced damage. Further, overexpression of SLC26A3 prevented TNF-α induced epithelial barrier dysfunction.

Correction of P131R-SLC26A3 to WT restored the epithelial barrier function
To investigate if normal function of SLC26A3 could be restored by changing the P131R-SLC26A3 sequence back to a WT sequence, we employed CRISPR/Cas9 gene editing using a novel ssODN in Caco-2 cells. We designed an ssODN that coded for the WT-SLC26A3, but utilized unique codons for the three amino acid sequence (F-P-I) that spanned the wild-type Proline. This allowed us to differentiate the newly constructed WT gene from the original gene. Sanger sequencing validated the correction of P131R-SLC26A3 to WT-SLC26A3 (Fig. 4a). In order to investigate the function of SLC26A3 corrected P131R-SLC26A3 (RWT) cells, we exposed confluent monolayers of Caco-2 cells to 150 mM NaCl. The treatments provoked reduction in TEER in P131R-SLC26A3 cells relative to WT-SLC26A3 cells that indicated a significant decrease in epithelial barrier function (Fig. 4b).
However, when SLC26A3-P131R was reversed back to wild type a similar TEER and epithelial barrier function was observed in the corrected cells (Fig. 4b). The maximum TEER inductions were 0.43 ± 0.02 vs. 0.72 ± 0.01 in P131R-SLC26A3 vs. WT (P < 0.001) and 0.69 ± 0.04 vs. 0.72 ± 0.01 in RWT vs. WT (P = 0.27) (Fig. 4c). These effects were transient and did not induce significant cellular loss, because TEER values recovered after withdrawal of the osmotic challenge. These results indicated that reverting P131R-SLC26A3 to WT can restore the epithelial barrier function.

ZO-1/CFTR mediates the epithelial barrier dysfunction induced by TNF-α and osmotic stress and P131R-SLC26A3 promotes SLC26A3 ubiquitination
Since TJ disruption is considered a vital event in the pathogenesis of intestinal inflammation, and there is increased permeability of the small intestine both in CF humans and in CF mice (Cftr knockout mouse mode) [20], we explored the physiological functions of P131R-SLC26A3 on the TJ protein ZO-1 and CFTR. Endogenous co-IP assays revealed that ZO-1 was immune-precipitated by the SLC26A3 antibody. In addition, SLC26A3 and ZO-1 protein levels are also decreased in P131R-SLC26A3 cells. Moreover, the interaction between SLC26A3 and ZO-1/CFTR are both decreased (Fig. 5a, b). To examine the detailed mechanism by which P131R-SLC26A3 induced lower levels of SLC26A3, we detected SLC26A3 degradation changes. As shown in Fig. 5c, d, P131R-SLC26A3 in Caco-2 cells resulted in increased ubiquitination of SLC26A3, which was intensified by TNF-α treatment. Correction of P131R-SLC26A3 to WT displayed similar ubiquitination status as the original WT. These results indicated that ZO-1/CFTR mediated the epithelial barrier dysfunction induced by TNF-α and osmotic stress, and that P131R-SLC26A3 further promoted SLC26A3 ubiquitination.

Construction and function of Slc26a3 P131R genetic variant on murine colonic epithelial cells
To lay the foundation for in vivo experiments we investigated the role of P131R-Slc26a3 in murine epithelial barrier function. We adapted a similar HDR-mediated modification strategy using the CRISPR/Cas9 system to edit the Slc26a3 gene sequence in a murine colonic epithelial cell line (CMT-93, Fig. 6a). CMT-93 cells were transiently transfected with the mixture of two sgRNA constructs and an ssODN. Sanger sequencing assay (Fig. 6b) validated the construction of P131R-Slc26a3. To characterize the role of P131R-Slc26a3 in murine epithelial barrier function, we measured TEER in P131R-Slc26a3 and WT CMT-93 cells. Upon 150 mM NaCl treatment, P131R-Slc26a3 cells showed lower TEER values compared with WT cells (Fig. 6c). The maximum TEER inductions were 0.38 ± 0.01 in P131R-Slc26a3 vs. 0.67 ± 0.02 in WT (P < 0.001) (Fig. 6d). Notably, similar results were obtained when we constructed P131R-Slc26a3 and analyzed effect of this SNP on permeability in murine epithelial cells.
In summary, variant P131R-SLC26A3 increases the epithelial permeability and disrupts function of SLC26A3 through two distinct molecular mechanisms: (a) Fig. 4 Functional analysis of the restored SLC26A3 function in reverted cells. a Sanger sequencing for confirming the correction of P131R SNP. We used a novel ssODN that codes for the same amino acids of WT SLC26A3, but uses unique codons for the F-P-I triAA sequence flanking the wild-type Proline that was changed to Arginine and now back to Proline. b TEER data from ECIS analysis of wild-type Caco-2 cells, Caco-2 cells containing the P131R and corrected cells (RWT). c Results from each group are presented as mean ± SD of three samples from three separate experiments. When SLC26A3-P131R was reversed back to wild type with the corrective donor templates by delivering Cas9/sgRNA vectors, a similar TEER and epithelial barrier function was observed in reverted cells. *P < 0.05, **P < 0.01, ***P < 0.001 decreasing SLC26A3 expression through ubiquitination pathway and (b) disrupting a key interaction with ZO-1/ CFTR, thereby increasing the epithelial permeability and induced epithelial barrier dysfunction.

Discussion
In this study, we employed the CRISPR/Cas9 genomic editing tool to create human colonic epithelial Caco-2 cells containing the SLC26A3 SNP rs386833481 (c.392C>G; p.P131R) and then reverted the SNP back to its wild type sequence to investigate its effects on intestinal epithelial cell permeability. SNP rs386833481 was identified in patients with congenital chloride diarrhea (CCD), but its functional consequence was unknown. We have provided several lines of solid evidence that SNP rs386833481 caused the increased permeability in intestinal epithelial cells, indicating it is a likely causative mutation for diarrhea. We have also demonstrated that this mutation caused increased ubiquitination mediated degradation of SLC26A3, leading to decreased protein levels of SLC26A3. Our findings, along with other reports [27], demonstrated that SLC26A3 overexpression enhances intestinal epithelial cell barrier function and may explain why SNP rs386833481 mutation caused increased intestinal epithelial cell permeability. Our study is the first to supply the evidence that SLC26A3 SNP rs386833481 (c.392C>G; p.P131R) is a likely causative mutation for diarrhea and has also provided a molecular mechanism underlying this observation.
Here, we investigated the influence of P131R-SLC26A3 on the epithelial barrier, and the mechanisms of regulation in human colonic epithelial cells (Caco-2). Functional analysis showed that P131R-SLC26A3 was associated with increased dysfunction of the epithelial barrier induced by TNF-α (Fig. 2) and osmotic stress (Fig. 3a), while overexpression of SLC26A3 protected the epithelial barrier against TNF-α (Fig. 2). Moreover, P131R-SLC26A3 was involved in the [Cl − ] i decrease induced by Tenidap (Fig. 3b). Correction of P131R-SLC26A3 prevented the NaCl-mediated alteration of epithelial barrier (Fig. 4) and TJ protein to a certain extent compared with the normal control (Fig. 5a),   Fig. 5 Identification of SLC26A3 as a CFTR and ZO-1 binding partner and effects on their interaction. a Endogenous IP assays of Caco-2 WT, P131R and RWT cell lysates showing that SLC26A3 co-precipitated with ZO-1 and CFTR, as well as that the interaction between SLC26A3 and ZO-1 is significantly decreased in P131R-SLC26A3 cells (P < 0.05). IgG was used as a negative control, GAPDH was used as an input control. b Densitometry analysis of endogenous IP assays. c P131R promotes SLC26A3 ubiquitination. WT, P131R and RWT cells were untreated (left panel) or stimulated with 100 ng/ml of TNF-α for 20 h (right panel), and cell lysates were assayed for immunoprecipitation (IP), and levels of ubiquitinated SLC26A3 were assessed by immunoblotting (IB). d Ratio of ubiquitination of protein between control and TNF-α treated groups. *P < 0.05, **P < 0.01, ***P < 0.001 suggesting that P131R-SLC26A3 might be critical for development of chronic diarrhea diseases caused by impaired epithelial barrier associated with disruption of TJ proteins. We further presented two mechanisms through which chronic diarrhea-risk-associated variant at 7q31.1 lead to increased dysfunction of the epithelial barrier by lower levels and activity of SLC26A3: (a) P131R-SLC26A3 reduced SLC26A3 expression through an enhanced ubiquitination mediated degradation pathway, and (b) a disrupted interaction with ZO-1/ CFTR protein (Fig. 5), resulting in increased epithelial permeability and induced epithelial barrier dysfunction. Both mechanisms point to reduced function of SLC26A3 as a mechanism for disease pathogenesis.
We also recreated this SNP and investigated its influence on the function of epithelial barrier in murine colonic epithelial cells (CMT-93). The result similarly indicated that P131R-Slc26a3 caused increased intestinal epithelial cell permeability induced by osmotic stress (Fig. 6). Our ongoing studies are pursuing recreation and correction of this point mutation in an in vivo mouse model by the AAV-CRISPR system to evaluate its utility for therapeutic development in chronic diarrhea.

Conclusions
In conclusion, our work using state of the art in vitro approaches has demonstrated that P131R-SLC26A3 (rs386833481) is a causal mutation in CCD, and correction of this SNP or increasing SLC26A3 function could be therapeutically beneficial for chronic diarrhea diseases. This is the first report defining function of the known SLC26A3 genetic variant. We corrected that variant, the mutant P131R allele, using the CRISPR/Cas9 mediated homologous recombination, and demonstrated restored normal epithelial barrier functionality of the corrected allele in the human colonic epithelial cells. Together with previous studies, which efficiently deliver the CRISPR components in vivo [28][29][30][31], this work provides a potential strategy for future gene therapy in patients with chronic diarrhea disease. Thus, our study is important in the elucidation of functional and biological consequences of SLC26A3 rs386833481, a likely therapeutic target in congenital chloride diarrhea, and with applicability to complex IBD, which exhibits reduced SLC26A3 expression and harbors some SNPs in its SLC26A3 [32,33].

CRISPR target sequence design
Guide sequences for CRISPR/Cas9 gene editing were designed as previously detailed [34], chemically synthesized, and RNase-Free HPLC purified by Integrated DNA Technologies (Coralville, IA, USA). Single-strand ODN was chemically synthesized and standard desalted by Integrated DNA Technologies (Coralville, IA, USA). All sequences are listed in Table 1.
We designed 2 guide RNAs (a forward and a reverse) that flank the SNP and a unique ssODN for a human epithelial cell line (Human sgRNA1, Human sgRNA2 and Human ssODN) and for a mouse epithelial cell line (mouse sgRNA1, mouse sgRNA2 and mouse ssODN), respectively. Each gRNA has a Top and Bottom oligo for cloning. For reverting the SNP to WT, we used different Forward sgRNA and a novel ssODN (Human sgRNA1a, Human sgRNA2 and Human ssODN-a) that code for the same amino acids as of WT SLC26A3, but use unique codons for the F-P-I triAA sequence flanking the wildtype Proline that was changed to Arginine and now back to Proline.

SNP models in human Caco-2 cells and murine CMT-93 cells
A total of 200,000 cells were seeded in 6-well plate overnight in the regular growth medium, so that they would be 80-90% confluent at the time of transfection. One hour prior to transfection, media was removed and 750 µl of pre-warmed reduced serum OptiMEM Table 1 Primer sequence for sgRNA cloning   (2) MQAE fluorescence without cells was recorded, and the Tenidap were added. No significant fluorescence changes were observed after performing both controls.

Co-immunoprecipitation
For co-IP assays, Caco-2 cells were lysed on ice with nondenaturing lysis buffer for 1 min, then were scraped and gently transferred into a chilled microcentrifuge tube. The cells were mixed on a rotary mixer for 30 min at 4 °C. After centrifugation, the concentration of supernatants were assayed by BCA assay and incubated overnight at 4 °C with the SLC26A3 antibody (Catalog#: GTX34204, GeneTex, Irvine, CA, USA). We used 500 µg protein and 2 µg SLC26A3 antibody in 500 µl Lysis Buffer containing the protease inhibitor for WT, P131R and RWT at the same time. After antibody binding, add 25 µl of protein A/G Sepharose ® beads slurry to each tube and incubate for 1 h at 4 °C. The beads were then washed three times with 1× wash buffer, and the precipitates were eluted with sample buffer, separated by 7.5% SDS/PAGE, and analyzed by immunoblotting.
We had validated all antibodies in Caco-2 cells, CMT-93 cells and HCT116 cells, which are all colonic epithelial cells.

Statistical analyses
Statistical analyses were carried out using the Sigma Stat (ver.4.0, SysTest Software, Inc., San Jose, CA). All data were expressed as mean ± SD (standard deviation) of at least three independent experiments. Two group comparisons were done by an unpaired Student's t test. Differences between groups were considered statistically significant at P < 0.05.