Pax6 Directly Down-Regulates Pcsk1n Expression Thereby Regulating PC1/3 Dependent Proinsulin Processing

Background Heterozygous paired box6 (Pax6) mutations lead to abnormal glucose metabolism in mice older than 6 months as well as in human beings. Our previous study found that Pax6 deficiency caused down-expression of prohormone convertase 1/3 (Pcsk1), resulting in defective proinsulin processing. As a protein cleaving enzyme, in addition to its expression, the activity of PC1/3 is closely related to its function. We therefore hypothesize that Pax6 mutation alters the activity of PC1/3, which affects proinsulin processing. Methodology/Principal Findings Using quantitative RT-PCR, western blot and enzyme assay, we found that PC1/3 C-terminal cleavage and its activity were compromised in Pax6 R266Stop mutant mice, and the expression of Pcsk1n, a potent inhibitor of PC1/3, was elevated by Pax6 deficiency in the mutant mice and MIN6 cells. We confirmed the effect of proSAAS, the protein encoded by Pcsk1n, on PC1/3 C-terminal cleavage and its activity by Pcsk1n RNAi in MIN6 cells. Furthermore, by luciferase-reporter analysis, chromatin immunoprecipitation, and electrophoretic mobility shift assay, we revealed that Pax6 bound to Pcsk1n promoter and directly down-regulated its expression. Finally, by co-transfecting Pax6 siRNA with Pcsk1n siRNA, we showed that Pax6 knock-down inhibited proinsulin processing and that this effect could be rescued by proSAAS down-regulation. These findings confirm that Pax6 regulates proinsulin processing partially through proSAAS-mediated PC1/3 processing and activity. Conclusions/Significance Collectively, the above experiments demonstrate that Pax6 can directly down-regulate Pcsk1n expression, which negatively affects PC1/3 C-terminal cleavage and activity and subsequently participates in proinsulin processing. We identified proSAAS as a novel down-regulated target of Pax6 in the regulation of glucose metabolism. This study also provides a complete molecular mechanism for the Pax6 deficiency-caused diabetes.


Introduction
Paired box 6 (Pax6) is a transcription factor and a member of the paired-box gene family [1,2]. It plays a critical role in regulating the development of the eyes, central nervous system and pancreas [3,4,5,6]. The Pax6-null mouse exhibits defects in multiple organs and therefore cannot survive after birth [4,7,8]. In addition, mice with heterozygous Pax6 mutations develop diabetes [6,7], suggesting Pax6 also regulates glucose metabolism.
Using conditional gene-inactivation approach, people showed that Pax6 was actually not required for the specification, formation, or survival of pancreatic beta cells. Instead, it is essential for the late stage differentiation and maturation of pancreatic beta cells [9]. In the previous report, we found that Pax6 directly regulated the expression of prohormone convertase 1/3 (PC1/3) encoded by the gene Pcsk1, a serine protease essential for proinsulin processing. The down-regulation of Pcsk1 led to defective proinsulin processing and abnormal glucose metabolism in Pax6 mutant mice and in PAX6-deficient patients as well [10]. Mice with a disruption of Pcsk1 exhibited growth retardation and several hormone precursors processing defects including proinsulin [11,12]. This conclusion was further supported by a human pedigree study, which reported that the PCSK1 mutation exhibited phenotypes including, but not limited to severe obesity, abnormal glucose homeostasis, elevated plasma proinsulin, low insulin level [13]. Human PCSK1 deficiency also caused severe malabsorptive refractory neonatal diarrhea and shared most phenotypes with the first case including reactive hypoglycemia, and elevated circulating levels of certain prohormones [14].
ProSAAS, the protein encoded by Pcsk1n, is a potent inhibitor of PC1/3 [15,16,17]. ProSAAS is expressed by broad neuronal distribution in the central nervous system, while in the periphery it is an excellent marker of endocrine cells [18]. ProSAAS was found responsible for inhibition of PC1/3 activity and C-terminal cleavage [15,16,17,19]. As reported, the activity of 66 kDa PC1/3 can be 40-folds higher than that of 87 kDa PC1/3, thus C-terminal cleavage was considered an important factor affecting the enzyme activity of PC1/3 [20]. Since there was an elevated fasting blood glucose level in adult Pcsk1n transgenic mice and impaired glucose tolerance in the Cpe fat/fat background transgenic mice [21], we sought to investigate whether the Pcsk1n expression is altered in the Pax6 mutant mice, and if so, whether the alteration would affect PC1/3 C-terminal cleavage and activity, thereby contributing to the abnormal proinsulin processing and onset of diabetes in these Pax6 mutant mice. In addition, to confirm the repressive function of Pax6 in regulation of gene expression as proposed by others [22,23,24], we also investigated whether Pax6 direct down-regulation of ProSAAS expression.

Results
We isolated pancreatic islets from either Pax6 R266Stop mutants or normal sibling mice, and measured PC1/3 activity and its C-terminal processing. Data from 3 experiments of western blot analysis is shown in Fig. 1 A. The 87 kDa/66 kDa PC1/3 ratio was elevated in the islets of the Pax6 mutant mice shown in Fig. 1 B. The results indicate that the mutant mice have defective PC1/3 processing in Pax6 mutated islets. In addition to investigating PC1/3 C-terminal cleavage, we also investigated the activity of PC1/3 in the islets of the Pax6 mutant mice. Considering the previously discovered fact that the expression of PC1/3 will be reduced in the islets of the mutant mice [10], the change of PC1/3 activity should be measured only when the amount of PC1/3 is equivalent. When the same amount of cleaved PC1/3 was used in enzyme activity analysis (Fig. 1C, upper row), the activity of PC1/3 was decreased in the mutant mice, as shown in Fig. 1 C. ProSAAS was reported to play roles in inhibiting PC1/ 3 activity and decreasing C-terminal cleavage of PC1/3 [19]. Therefore, our results indicate that proSAAS may be involved in the regulation of PC1/3 activity in Pax6 mutant mice.
To investigate whether the expression of Pcsk1n was affected in the pancreatic islets of the Pax6 mutant mice, we measured Pcsk1n expression by qRT-PCR and western blot. Indeed, Pcsk1n was significantly elevated in the pancreatic islets of Pax6 mutant mice as measured by qRT-PCR ( Fig. 2A) and western blot (Fig. 2B) in comparison with the results of the pancreatic islets of normal mice, while Pax6 was decreased. Furthermore, both Pcsk1n mRNA (Fig. 2C) and proSAAS proteins (Fig. 2D) were significantly elevated upon the decrease in Pax6-knocked down in MIN6 cells. The graphs in Fig. 2B and 2D were quantification for scanning western blot bands of Pax6 and proSAAS relative to b-actin, respectively. These observations showed that decreased Pax6 led to an increase of proSAAS production in the pancreatic islets of Pax6 mutant mice and in MIN6 cells, which indicates Pax6 downregulates Pcsk1n expression in islets.
To investigate the possibility that the decrease of proSAAS promotes PC1/3 processing and activity, we transfected Pcsk1n siRNA in MIN6 cells. Results showed that the 87 kDa/66 kDa PC1/3 ratio had decreased (Fig. 3 A and B), contrary to its increase in Pax6 mutant mice. In Fig. 3A & 3B the knock-down of Pcsk1n was shown to affect the cleavage of PC1/3 pro-form. Fig. 3C shows that Pcsk1n not only affects the cleavage of PC1/3 pro-form but also the activity of PC1/3. When the final amount of 66 kD PC1/3 was the same, PC1/3 activity was increased in Pcsk1n knock-down cells. This observation indicates that proSAAS actually inhibits the activity of PC1/3 by affecting its cleavage as well as its enzyme activity. Because the substrate used in the enzyme assay can be hydrolyzed not only by PC1/3 but also by PC2 and furin, it is necessary to investigate the specificity of the enzyme assay. The PC1/3 specific inhibitor Ac-LLRVKR-NH 2 , a peptide derived from proSAAS, was added to the enzyme assay reaction system according to Apletalina et.al and Cameron et.al [25,26]. As shown in Fig. 3D, Ac-LLRVKR-NH2 dramatically blocked the enzyme activity to a lower level with Pcsk1n and scrambled siRNA. This is because Ac-LLRVKR-NH 2 almost blocked all the PC1/3 activity, but with scrambled siRNA, some PC1/3 activity remained. The blocked enzyme activity shows that PC1/3 plays a major role in the enzyme activity assay system. This observation indicates that reduced enzyme activity in the islets of Pax6 mutants is mainly due to the deficiency of PC1/3 activity and the elevation of its inhibitor proSAAS. We also observe diminished PC1/3 expression level after Pcsk1n knock-down, which is consistent with reports that proSAAS co-expression increases synthesis and secretion of PC1/3 in CHO cells [19]. However, whether or not Pax6 affected PC1/3 C-terminal cleavage and its activity through regulating the Pcsk1n expression level remains to be revealed.
To answer this question, we first analyzed the distribution of proSAAS and Pax6 in the pancreatic islets as well as in MIN6 cells by immunofluorescence staining. The results indicated that not only Pax6 co-localized with proSAAS ( Fig. 4A & 5A), but Pax6 also co-localized with insulin in pancreatic islets and MIN6 cells (Fig 4 B & 5B). Furthermore, we investigated the co-localization of proSAAS and endocrine hormones, such as insulin and glucagon. ProSAAS was expressed in b-cells (Fig 4 D, 5C), though some acells was also stained with proSAAS (Fig 4 C). We noticed that in b-cells the signal of insulin was lower as the signal of proSAAS was higher and vice versa (Fig 4 D). Similar results were observed in MIN6 cells (Fig. 5C). This observation showed that the expression of proSAAS might get involved in down-regulating insulin processing.
To test the hypothesis that Pax6 regulates Pcsk1n expression in islets, we first performed a luciferase reporter assay using Pcsk1n promoter activity reporter system. Wild-type Pax6 but not the R266Stop mutation moiety led to a dose-dependent decrease of Pcsk1n promoter activity in HEK293 cells (Fig. 6A); knock-down of Pax6 led to elevation of Pcsk1n promoter activity in MIN6 cells ( Fig. 6 B).To identify the Pax6 binding site in the Pcsk1n promoter, we investigated a 2000 bp up-stream sequence of Pcsk1n promoter from the transcription start site, as pGl 3 -full-length promoter, based on Genomatix MatInspector database [27]. The bioinformatics analysis results suggested a putative binding site: sequences from 21433 to 21415 bp containing a Pax6 core binding sequence TTCATGCTTG [28] from the transcriptional start site. Based on the ChIP analyses, Pax6 was able to bind to the Pcsk1n promoter within (21599 to 21350 bp) region ( Fig. 7 A, upper panel). This binding was specific since the other regions including the sequence of 24116 to 23526 bp did not show Pax6 binding ( Fig. 7 A, lower panel).
To further certify the binding, a long oligonucleotide containing the candidate binding site of the sequence from 21433 to 21415 bp was used as a labeled probe in an EMSA. When the labeled oligonucleotide representing the promoter sequence from 21441 to 21408 bp was incubated with nuclear extract prepared from MIN6 cells, a protein-probe complex was detected. When 100-fold molar excess of competitive unlabeled DNA was added, the formation of protein-probe complex failed (Fig. 7B). Moreover, the addition of Pax6-specific antibody, but not the normal IgG control, caused a supershift band (Fig. 7B). These findings indicate the Pax6 binding site is located at 33 base pairs within the region of 21441 to 21408 bp from the transcriptional start site of the Pcsk1n promoter. This is consistent with the bioinformatics analysis.
To functionally confirm this Pax6 binding site for regulation of Pcsk1n expression, we performed truncated luciferase reporter analysis for Pcsk1n promoter activity. Pcsk1n transcriptional activity was found significantly decreased when Pax6 was over-expressed using pGl 3 -full-length promoter system. On the other hand, Pax6 no longer affected the Pcsk1n promoter activity when the region of 2999 to 21319 bp was deleted (Fig. 7C).
Knock-down of Pax6 led to up-regulation of proSAAS (Fig. 2,  Fig. 8A). Knock-down of Pcsk1n elevated PC1/3 C-terminal cleavage and activity (Fig. 3, Fig 8 A, B and C). This is consistent with the fact that PC1/3 C-terminal cleavage and activity in Pax6 mutant pancreatic islets is decreased (Fig. 1). In addition to Pax6 RNAi, Pcsk1n and Pax6 double knock-down could decrease the proSAAS production (Fig. 8A), and partially rescue the decreased PC1/3 C-terminal cleavage ( Fig. 8A and B) and activity (Fig. 8C) induced by Pax6 knock-down alone. Since PC1/3 C-terminal cleavage and activity is closely related to proinsulin processing, as indicated by true insulin/total insulin ratio, we further investigated the concentration of true insulin and total insulin in conditional medium of MIN6 cells by ELISA. The double knock down of Pax6 and Pcsk1n partially recovered the true insulin/total insulin ratio,

Discussion
Using Pax6 heterozygous R266Stop mutant mice as an animal model, we showed in previous study that Pax6 regulated glucose metabolism via up-regulation of PC1/3, an essential enzyme for proinsulin processing. In this follow-up study, we revealed that Cterminal cleavage and activity of PC1/3 were also compromised in pancreatic islets of Pax6 mutant mice. On the other hand proSAAS, a protein responsible for inhibition of PC1/3 Cterminal cleavage [19] and activity [15,16,17,19], was significantly increased in Pax6 mutant mice. We further proved that Pax6 bound directly to the Pcsk1n promoter and down-regulated Pcsk1n expression. Jointly, the above results reveal a more complete mechanism for Pax6 deficiency related diabetes. Specifically, Pax6 affects proinsulin processing and glucose metabolism through upregulation of PC1/3 and down-regulation of proSAAS. This study also identifies proSAAS as a mediator between Pax6 and PC1/3 in pancreatic islets.
It has been reported that over-expression of proSAAS reduced the rate of processing of the endogenous prohormone proopiomelanocortin at a PC1/3-specific site [16]. Our study proves that proSAAS also regulates proinsulin processing via inhibiting Cterminal cleavage and activity of PC1/3 in pancreatic cells for the first time. It has been reported that Pcsk1n transgenic mice exhibited elevated fasting serum glucose [21], which supports the finding that proSAAS functions as a proinsulin processing related molecule in blood glucose regulation. The study of fetal mice with a Pcsk1n mutation suggests that proSAAS-derived peptides could inhibit PC1/3 in an embryonic brain, but the PC1/3 activity was not affected by the absence of Pcsk1n in adult mice [29]. This finding in neuropeptide processing is not consistent with our results in islets. Our study focuses on proinsulin processing in islets of Pax6 mutant mice, in which the expression of Pcsk1n is elevated. Differing from the Pcsk1n mutant mice reported by Morgan et al., our finding is more consistent with the impaired fasting glucose of Pcsk1n transgenic mice [21]. In support of our findings, Fortenberryet.al has also found that full-length proSAAS inhibits PC1/3 activity in CHO/PC1/3 cells [19]. It seems that in the central nerve system, proSAAS functions as neuropeptides regulating body weight and other behaviors, but in the endocrine system, proSAAS regulates PC1/3 maturation and activity, though its regulation seems to be modest. Relatively little research has focused on proSAAS's endocrinal functions, and the reason for its distinguished roles remains to be revealed.
Our previous study has reported that the proinsulin/insulin ratios in Pax6 mutant mice older than 6 months were significantly higher than those in wild-type controls. These mice exhibit impaired glucose tolerance. In this study, the Pax6 mutant mice at the age of 4 months showed elevated Pcsk1n expression. The aging related regulation suggests an effect of the accumulation of heterozygous mutation in mice.
In a cells of pancreatic islets Pax6 up-regulates PC2 and its inhibitory chaperon 7B2 [30] and in b cells we find Pax6 upregulates PC1/3 but down-regulates Pcsk1n. Pcsk1n derived peptides show structural and functional homology with 7B2 [15,19], and this difference in regulation by Pax6 may correspond with different functions of proSAAS and 7B2. Although both proSAAS and 7B2 potently inhibit PCs, their intracellular interactions with PCs differ significantly. The N-terminal domain of proSAAS has no effect on stabilization of PC1/3 activity, different from the effect of the N-terminal domain of 7B2 on PC2 [19].
While most research reveals an up-regulation effect of Pax6, we identify a novel down-regulated target of Pax6, proSAAS, which participates in the regulation of proinsulin processing and glucose metabolism. There have been a few down-regulated targets of Pax6 identified. VEGFA, a growth factor and MMP2, a cellular matrix protein, involved in tumor formation and metastasis are found repressed by Pax6 [22,23]. Olig2, a transcription factor critical for glial cell fate determination during forebrain development, is also directly repressed by Pax6 [24].Our finding of the down regulation of Pcsk1n by Pax6 provides a potential target for improving proinsulin processing and attenuating Pax6 deficiency-caused abnormal glucose metabolism. with Pcsk1n siRNA for 48 h, and protein was collected to analyze the PC1/3 C-terminal cleavage with western blots. B, ratio of densitometry scanning of 87 kDa and 66 kDa PC1/3 bands was analyzed using MIN6 cells transfected with Pcsk1n siRNA (black bar) or scrambled siRNA (white bar) (n = 3). C, enzyme activity was analyzed using MIN6 cells transfected with Pcsk1n siRNA (black circles) or scrambled siRNA (white circles) with equivalent protein (n = 3). D, the specificity of enzyme activity measured was tested by PC1/3 specific inhibitor Ac-LLRVKR-NH 2 using MIN6 cells. Black squares, Pcsk1n siRNA; Black circles, scrambled siRNA; White squares, Pcsk1n siRNA + inhibitor; White circles, scrambled siRNA + inhibitor. A representative experiment is shown; the experiment was repeated three times. Each value represents the mean 6 SE. *p,0.05. doi:10.1371/journal.pone.0046934.g003 Pax6 Regulates Pcsk1n Expression PLOS ONE | www.plosone.org The role of Pax6 in proinsulin processing has been constantly followed by other researchers. For instance, Gosmainet.al has found that Pax6 represents a key component of the transcriptional network in a cells [31] and b cells [32], regulating a series of molecules including PC1/3. This result supports our findings above. In summary, our study reveals a novel regulating pathway of Pax6-proSAAS-PC1/3 activity in proinsulin processing as part of glucose metabolism regulation.

Animals
Animals using were as previously described (11). Briefly, a stopcondon mutation at 266 aa of Pax6 was led by injecting the chemical mutagen N-ethyl-N-nitrosourea. After breeding with normal mice, the mutation was transmitted to newborns. Based on the small-eye phenotype, mice with heterozygous Pax6 gene mutation can be selected. Mice were maintained in a pathogenfree facility under a 12 h light-dark cycle and were given free access to food and water. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health [34]. The protocol was approved by the Peking University animal ethics committee, and all efforts were made to minimize suffering. All data relevant not specifically mentioned were collected after 6 months of birth.

Isolation and Purification of Mouse Pancreatic Islets
Mouse pancreatic islets were isolated and purified following Hori's method [35]. Briefly, pancreases were digested with type V collagenase (Sigma-Aldrich, St Louis, MO, USA), and centrifuged (2,700 g) in PBS. Then individual islets were handpicked under a dissecting microscope, without bringing attached acinar, vascular or ductal tissues.

Enzyme Assays
Enzyme assays were conducted using 400 mmol/l Pyr-Glu-Arg-Thr-Lys-Arg-methylcoumarin amide (Peptide Institute, Osaka, Japan) as a substrate in an assay previously described [36]. Protein of islets or MIN6 cells were subjected with substrate in 100 mmol/ l sodium acetate, pH 5.0, containing 5 mmol/l CaCl 2 , 0.1% Triton X-100, and a protease inhibitor mixture composed of 1 mM trans-epoxysuccinic acid, 1 mmol/l pepstatin, 280 mmol/l TPCK, and 140 mmol/l TLCK. All assays were performed at 37uC in a 96-well plate fluorometer with an excitation wavelength of 380 nm and an emission wavelength of 460 nm. The total volume was 50 ml. When used, 120 nM of PC1/3 specific inhibitor Ac-LLRVKR-NH 2 (synthesized by Bioss. Inc, Beijing, China) was pre-incubated with protein at room temperature for 30 min prior to the addition of the substrate.The activity was measured duplicate or triplicate every 5 min for a period of 2.5 h, and the rates of hydrolysis over this time were calculated; the first 10 min were excluded from the rate calculation. Enzyme activity is given in fluorescence units per minute (FU/min) where one FU corresponds to 8.6 pmol of methylcoumarin amide.

Luciferase Reporter Assay
Mouse Pcsk1n promoter (21999 bp to 0 bp) and serial deletions were cloned from mouse genomic DNA and inserted into pGL 3promoter luciferase reporter vectors. The primers were listed as follows: forward primer for 21999Luc (21999 bp to 0 bp) 59-

Chromatin Immunoprecipitation Assays
Chromatin immunoprecipitation (ChIP) assays were performed in MIN6 cells. Briefly, the chemical-linked complexes of genomic siRNA or Pax6 and Pcsk1n siRNA, and aliquots of conditional medium were applied to ELISA Kits for total insulin or true insulin detection (n = 3). True insulin, the mature form of insulin, which was derived from proinsulin cleaved by PC1/3 and PC2; total insulin, the sum of true insulin and proinsulin. White bars, value for transfected with scrambled siRNA; Black bars, value for transfected with Pax6 siRNA; Grey bars, value for transfected with both of Pax6 and Pcsk1n siRNA. A representative experiment is shown; the experiment was repeated three times. Each value represents the mean 6 SE. *p,0.05, **p,0.01. doi:10.1371/journal.pone.0046934.g008 DNA and Pax6 protein were immunoprecipitated with antibodies specific to Pax6 (Chemicon, Rosemont, Illinois, USA). The DNA was amplified with LA taq (TAKARA, Shiga, Japan) after ChIP using primers of Pcsk1n promoter 21599,21350 bp (forward  primer 59-TGTGCCATCTCTCCAGCCCAGC-39, reverse  primer  59-ATCTAGCAGCTAGATAATGTCT-39), and 24116,23526 bp (forward primer 59-TAGACCAG-GATGGCCTCAAACTAAAA-39, reverse primer 59-AGCCC-TAAACCCTTACTTCTGACCT-39.

Electrophoretic Mobility Shift Assay
Electrophoretic Mobility Shift Assay (EMSA) was performed with LightShift Chemiluminescent EMSA Kit (Pierce, Rockford, USA) following the manufacturer's protocols. Briefly, 5 micrograms of nuclear extract from MIN6 cells were added in a reaction mixture.
The oligonucleotides used were: 59-TTTCTTGGTGAACTCCAAGCATGAACGTGCTCTA-39. Probes were synthesized with biotin labeled at the 59 terminal and then annealed. For supershift experiments, nuclear extracts were preincubated with antibody for 10 min on ice before the addition of biotin-labeled DNA probe and subsequently incubated at room temperature for 10 min. Competition was carried out with a 100fold excess of the unlabeled probe. All reaction mixtures were loaded onto a 6% nondenaturing polyacrylamide gel, and after electrophoresis, the gel was exposed to X-ray film for 1 min.

ELISA
MIN6 cells at required time point were washed twice with PBS and once with DMEM-High, and incubated with 1 mL DMEM-High per well (6 well plate) for 4 h. The conditioned medium was centrifuged and appropriately diluted, and 10 ml aliquots were subjected for ELISA. True insulin and total insulin were measured with an ELISA kit from Shibayagi Company (Ishihara Shibukawashi, Gumma, Japan) according to the manufacturer's instructions. True insulin is the mature form of insulin that was cleaved by PC1/3 and PC2, and total insulin is the sum of true insulin and proinsulin. Thus the ratio of true insulin/total insulin indicates proinsulin processing. Since the cross-reactivity of true insulin and proinsulin is less than 5% according to the manufacturer's description, the calculated ratios of true insulin over total insulin in mice are likely to be not precise, but are acceptable for the present purposes of comparison.

Statistical Analysis
Statistical analysis was performed using an independent samples t test. Differences were considered statistically significant at p,0.05.