PARP-1 and YY1 Are Important Novel Regulators of CXCL12 Gene Transcription in Rat Pancreatic Beta Cells

Despite significant progress, the molecular mechanisms responsible for pancreatic beta cell depletion and development of diabetes remain poorly defined. At present, there is no preventive measure against diabetes. The positive impact of CXCL12 expression on the pancreatic beta cell prosurvival phenotype initiated this study. Our aim was to provide novel insight into the regulation of rat CXCL12 gene (Cxcl12) transcription. The roles of poly(ADP-ribose) polymerase-1 (PARP-1) and transcription factor Yin Yang 1 (YY1) in Cxcl12 transcription were studied by examining their in vitro and in vivo binding affinities for the Cxcl12 promoter in a pancreatic beta cell line by the electrophoretic mobility shift assay and chromatin immunoprecipitation. The regulatory activities of PARP-1 and YY1 were assessed in transfection experiments using a reporter vector with a Cxcl12 promoter sequence driving luciferase gene expression. Experimental evidence for PARP-1 and YY1 revealed their trans-acting potential, wherein PARP-1 displayed an inhibitory, and YY1 a strong activating effect on Cxcl12 transcription. Streptozotocin (STZ)-induced general toxicity in pancreatic beta cells was followed by changes in Cxcl12 promoter regulation. PARP-1 binding to the Cxcl12 promoter during basal and in STZ-compromised conditions led us to conclude that PARP-1 regulates constitutive Cxcl12 expression. During the early stage of oxidative stress, YY1 exhibited less affinity toward the Cxcl12 promoter while PARP-1 displayed strong binding. These interactions were accompanied by Cxcl12 downregulation. In the later stages of oxidative stress and intensive pancreatic beta cell injury, YY1 was highly expressed and firmly bound to Cxcl12 promoter in contrast to PARP-1. These interactions resulted in higher Cxcl12 expression. The observed ability of PARP-1 to downregulate, and of YY1 to upregulate Cxcl12 promoter activity anticipates corresponding effects in the natural context where the functional interplay of these proteins could finely balance Cxcl12 transcription.


Introduction
Type 1 diabetes (T1D) is a multifactorial disease believed to be of immunological origin, precipitated by infections and environmental factors in genetically predisposed individuals. The hallmark of T1D is selective death of pancreatic insulinproducing beta cells resulting from attack by mononuclear cells. The maintenance of an appropriate number of pancreatic beta cells remains a viable interventive measure in diabetes. Detection of novel beta cell growth factors will provide crucial information for strategies that could compensate for depletion and defects of beta cell functioning.
The chemokine (C-X-C motif) ligand 12 (CXCL12) or stromal cell-derived factor-1 (SDF-1) belongs to the CXC group of chemokines. CXCL12 was discovered as a pre-B cell growthstimulating factor [1,2]. The CXCL12 is a ligand of two transmembrane receptors, chemokine (C-X-C motif) receptor 4 (CXCR4) and chemokine (C-X-C motif) receptor 7 (CXCR7) [3,4]. An antidiabetogenic potential of CXCL12 was recently revealed in vitro and in vivo. Transgenic mice that overexpress CXCL12 in their beta cells are resistant to apoptosis and diabetes.
It was shown that CXCL12 stimulates pancreatic beta cell survival by preventing apoptosis via activation of the prosurvival kinase Akt and the resulting upregulation of antiapoptotic protein Bcl-2 and phosphorylation of the proapoptotic protein Bad [5]. Also, beta cell injury induces CXCL12 expression, and the secreted CXCL12 causes the dedifferentiation of adjacent alpha cells into pro-alpha cells. This is an initial step in transdifferentiation of alpha to beta cells [6,7]. The process of transdifferentiation in the pancreas is of particular interest, since T1D results from an insufficient number of functional beta cells. Furthermore, the human gene for CXCL12 is located on chromosome 10q11.1, near to the T1D susceptibility locus IDDM10, indicating that CXCL12 gene variants could contribute to diabetes development. Analysis of single nucleotide polymorphisms in the CXCL12 gene revealed that the CXCL12-39A variant is associated with the early onset of T1D in some populations [8,9].
Since, CXCL12 is proven to be important in pancreatic islet survival, we aimed to advance knowledge concerning the regulation of rat CXCL12 gene (Cxcl12) transcription, and for the first time we focused on two transcription factors; the poly(ADP-ribose) polymerase-1 (PARP-1) and the ubiquitous transcription factor Yin Yang 1 (YY1).
PARP-1 is a multifunctional nuclear enzyme involved in the regulation of a variety of nuclear processes, including cell death [10], replication and differentiation [11], telomere activity [12], energy balance for cellular processes [13] and transcription [14]. Aside from binding to DNA breaks [15], PARP-1 binds to specific DNA sequences, thereby regulating transcription of its own [16,17] and other genes such as Mcat1 [18], Pax-6 [19], MHC II [20], Cxcl1 [21], Reg [22] and Bcl-6 [23]. The enzymatic activation of PARP-1 implies the transfer of ADP-ribose moieties to acceptor proteins in the nucleus (heteromodification), including itself (automodification) [24]. With regard to diabetes, PARP-1 deficiency provides protection from experimentally induced diabetes. Namely, PARP-1 knockout (PARP2/2) mice were shown to be resistant to streptozotocin (STZ) -induced diabetes, restoring normal blood glucose and pancreatic islet structure [25]. Moreover, PARP-1 inhibition protects against autoimmune beta cell destruction in NOD mice via induction of apoptosis of isletinfiltrating leukocytes [26].YY1 is a ubiquitous zinc finger transcription factor that can initiate, activate or repress transcription [27]. When the DNA binding motif of YY1 occurs downstream from the transcription start site, it often overlaps with the Kozak sequence [28]. YY1 is originally identified as a DNA-binding nuclear matrix protein with the ability to bind DNA sequences possessing an unwinding propensity [29,30]. Subsequent transient poly(ADP-ribosyl)ation of YY1 reduces its DNA binding affinity [31]. Functional relations between YY1 and PARP-1 are also relevant in cases where enzymatic PARP-1 activity modulates transcription [32]. YY1 is involved in the regulation of several genes responsible for cellular functions governing cellular stability [33]. YY1 participates in the regulation of Parp-1 [34], the chemokine receptor Cxcr4 [35] and the antiinflamatory cytokine IL-4 [36] that protects against diabetes development [37].
The aim of our study was to investigate the molecular mechanisms that regulate gene transcription of CXCL12, a potential beta cell growth factor. Our results revealed two novel regulators of the CXCL12 gene and elucidated their influence on Cxcl12 transcription. Furthermore, our investigation clarified Cxcl12 promoter regulation in the basal state and during STZinduced pancreatic beta cell injury.

Cell Viability Assay
Rin-5F wt and clone #1 cell viability was estimated by the 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) viability assay. Cells were cultured in a 96-well plate and treated with increasing concentrations of STZ (0.1-15 mM) for 24 h. After removing the medium, 200 ml of MTT (Sigma, M5655) at a concentration of 0.5 mg/ml in RPMI was added to each well. Cells were incubated for 2 h in the dark and the resultant formazan crystals were dissolved in dimethyl sulfoxide. The absorbance was measured at 570 nm. Cell viability was expressed in percentages after comparison with control cells that were assumed to be 100% viable.

Comet Assay
The levels of DNA damage after increasing times of STZ treatment were estimated by the alkaline Comet assay according to Singh et al. [38]. Rin-5F wt cells were grown in 6-well plates and exposed to 5 mM STZ (Sigma) for increasing times (0.5, 1, 3 and 6 h). Control and STZ-treated cells were collected in PBS. The collected cells (10 ml) were mixed with low-melting agarose (0.75%) and applied to a microscope slide. Cells were lysed for 2 h at 4uC in lysis buffer (2.5 M NaCl, 100 mM EDTA, 10 mM Tris, pH 10, 1% Triton X-100). After lysis, microscope slides were incubated for 30 min at 4uC in electrophoresis buffer (300 mM NaOH, 1 mM EDTA, pH 13.0) in order to denature DNA. Damaged DNA fragments were separated at 10 V for 30 min at 4uC. The slides were washed in neutralization buffer (0.4 M Tris-HCl, pH 7.4) and stained with SYBR Green I (1:10,000 dilution; SYBR; Sigma-Aldrich). Images were analysed with TriTekCo-metScoree Freeware version 1.5 (http://www.AutoComet.com).

Preparation of Protein Fractions
Cell protein fractions were prepared with the ProteoJET Mammalian Cell Lysis Reagent and ProteoJET Cytoplasmic and Nuclear Protein Extraction Kit (Fermentas).

Electrophoretic Mobility-shift Assay (EMSA)
Rat genomic DNA was extracted from the RIN-5F wt cell line. The Cxcl12 promoter (739 bp) was amplified using biotinylated PCR primers: upstream 5'-biotin-CAGCACAGCCCTACGT-TAGA- 39 and downstream 5'-biotin-ACAGAGCTGC-GAGCCTTGCC-39. The PCR products were purified using QIAquick Gel Extraction Kit (Qiagen). EMSA was performed in a binding buffer containing 6.25 mM MgCl 2, 10% glycerol, 2.5 mM EDTA, 2.5 mM DTT, 250 mM NaCl and 50 mM Tris-HCl (pH 7.5). The nuclear lysate (20 mg) was incubated with binding buffer for 15 min at room temperature. Subsequently, 100 ng of biotinylated DNA fragments were added and incubation was carried out at 37uC for 30 min. Poly(dIdC) (1 mg) was used as a competitor DNA in each binding reaction. For supershift experiments, 1 mg of anti-PARP antibodies (H-250-Santa Cruz; R&D and C2-10-ALEXIS Biochemicals), and 1 mg of anti-YY1 antibody (H-414-Santa Cruz) were added to the protein mixture and incubated at 37uC for 30 min. EMSA was also performed using recombinant PARP-1. The binding reaction containing the binding buffer, 100 ng of biotinylated Cxcl12 promoter, 100 ng of recombinant PARP-1 (ALEXIS Biochemicals) and 1 mg of poly(dIdC) was incubated for 30 min at 37uC. Reaction mixtures were separated by non-denaturing electrophoresis on a 1% agarose gel at 80 V for 2-3 h in 16Tris-borate-EDTA buffer (16TBE) at room temperature. The biotinylated DNA ends conjugated with streptavidin-alkaline phosphatase prevent PARP-1 binding to the DNA ends and allows post-EMSA DNA visualization by the DuoLux chemiluminescent substrate (Ultra-SNAP detection system; Vector Laboratories) according to the manufacturer's instruction.

Chromatin Immunoprecipitation (ChIP)
Rin-5F wt cells were treated with 5 mM STZ for 0.5 and 6 h. Control and STZ-treated cells were fixed with 1% formaldehyde (Lach-ner) for 10 min at room temperature, according to the ChIP-IT Express protocol (ActiveMotif). Chromatin was sheared on ice with 20 pulses. Each pulse consisted of sonication for 20 s, followed by a 30 s rest on ice. Sheared chromatin yielded a 200-500 bp smear. Immunoprecipitation was performed using 3 mg of the following antibodies: two different anti-PARP-1 antibodies

Reporter Gene Constructs
Rat genomic DNA extracted from Rin-5F wt cells was PCRamplified following standard procedures. To amplify the Cxcl12 promoter we used the following primers: upstream 59-GGTCGA-TACTAGTTTGTAAAGACACCAATGACC-39 and downstream 59-CCTAAGCCTCGAGTGGGCGG-GAGGGCGCGCCGGAGGCT-39. The amplified Cxcl12 promoter was cloned in the pMDICluc construct using restriction enzymes SpeI and XhoI. In the generated pCXCL12luc construct, the Cxcl12 promoter drives the firefly luciferase gene. The ampicillin gene served as a selection marker. The resulting pCXCL12luc construct was sequenced in both directions using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems).

Transient Transfection
The day before transfection, cells were plated in 24-well plates, being 70% confluent after 24 h. Rin-5F (wt and clone #1) and NIH3T3 (PARP-1 +/+ and PARP-1 2/2 ) cells were transfected with pCXCL12luc and pMDICluc constructs. pMDICluc is a control construct in which the firefly luciferase gene is under the control of the CMV promoter. PARP-1 +/+ and PARP 2/2 fibroblasts were cotransfected with both pCXCL12luc and the PARP-1 expression construct pECV PARP. PARP-1 +/+ fibroblasts were cotransfected with both pCXCL12luc and the pcDNA3.1FLAGYY1 plasmid containing the FLAG-tagged human YY1 coding sequence. In each transfection reaction the Renilla luciferase-construct, served as a normalizing transfection control for firefly luciferase. Transfection experiments were performed using Lipofectamine TM 2000 (Invitrogen), according to the manufacturer's instructions. Plasmid DNA (0.5 mg) and the Renilla luciferase-construct (0.065 mg) were diluted in cell culture medium. Lipofectamine TM 2000 (2 ml) was dissolved in the medium and incubated at room temperature for 5 min. The diluted DNA and cationic lipid were combined and incubated for 20 min. Cell culture medium (without antibiotics) supplemented with the plasmid-lipid complexes was added to each well. After 5 h incubation, the medium was replaced by complete medium and incubation was continued for 24 h. The ratio of the luciferase activity units obtained for each cell line transfected with pCXCL12luc and pMDICluc was normalized by dividing the firefly signal by the Renilla signal. The activity of the Cxcl12 promoter was expressed relatively to activity of control CMV promoter.

Dual-luciferase Reporter Assay System
Luciferase activity was measured with the Dual-LuciferaseH Reporter Assay System according to the manufacturer's recommendations (Promega Corporation). Cells were lysed 24 h after transfection and firefly luciferase activity was measured immediately after adding LAR II reagent. Afterwards, Stop & GloH reagent was introduced and Renilla luciferase activity was recorded.

RNA Isolation and Real-time RT-PCR (RT-qPCR)
Rin-5F wt cells were exposed to 5 mM STZ for increasing times (0.5, 1, 3 and 6 h). In some experiments cells were pretreated with 50 mM 3AB for 30 min, followed by 6 h STZ treatment. Total RNA from Rin-5F control (wt and clone #1) and STZ-treated wt cells was extracted using the GeneJET RNA Purification Kit (Thermo Fisher Scientific). For cDNA synthesis, 1 mg of the total RNA was treated with DNAse I and reverse-transcribed with RevertAid First Strand cDNA Synthesis Kit (Fermentas) using oligo(dT) primers. For RT-qPCR the Maxima SYBR Green/ ROX qPCR Master Mix (Fermentas) was used. mRNA levels were quantitatively determined with an ABI Prism 7000 Sequence Detection system (Applied Biosystems). The fragments were amplified using the following primer sets: upstream 59-GATTCTTTGAGAGCCATGTC-39 and downstream 59-GTCTGTTGTTGCTTTTCAGC-39 for rat CXCL12 gene; upstream 59-ATGAACGCCAAGGTCGTGGT-39 and downstream 59-GGGCACAGTTTGGAGTGTTG-39 for human CXCL12 gene; upstream 59-CTGACTGGTACTTTGGGAAA-39 and downstream 59-GGAACACCACCATCCACAGG-39 for rat CXCR4 gene; upstream 59-CTGGTGGACATTGT-GAAAGG- 39 and downstream 59-TCTGCCTTCTGCTCAGTTTC-39 for rat PARP-1 gene; upstream 59-GCCAGCCGAGATCGTGGAAC-39 and downstream 59-GATCATGGGCGGGTGGTGGT-39 for rat YY1 gene; upstream 59-AGATTACTGCCCTGGCTCCT-39 and downstream 59-ACATCTGCTGGAAGGTGGAC-39 for rat bactin gene. The real-time PCR program for quantitative RT-PCR was comprised of an initial step at 50uC for 2 min, followed by an initial denaturation step at 95uC for 10 min and a subsequent 2-step PCR program at 95uC for 15 s and 60uC for 60 s for 40 cycles. Negative controls lacking the template were used in all RT-qPCR reactions. The expression levels of target genes were related to the averaged expression level of rat b-actin as the housekeeping gene.

Statistical Analysis
The results are expressed as means 6 SEM of triplicate data. Student's t test was used to determine the significance of differences between two groups; p,0.05 was considered significant.

Increased Presence of CXCL12 Improves Pancreatic Beta Cell Survival during Oxidative Stress Induced by a Diabetogenic Stimulus
To confirm the prosurvival potential of CXCL12, the RIN-5F rat pancreatic beta cell line (wt) and its counterpart possessing a stably integrated human gene for CXCL12 (clone #1) were exposed to increasing concentrations of STZ (Fig. 1A). Monitoring cell viability revealed that increasing STZ concentrations were toxic to proportionally more wt than clone #1 cells. STZ concentration of 5 mM induced death in about 42% and 6% of wt and clone #1 cells, respectively. This STZ concentration was taken as EC 50 in all further experiments. The viability data indicates that the overexpressed CXCL12 gene and increased presence of CXCL12 protein in the cell culture medium (inset in Fig. 1A) exerted a prosurvival effect on pancreatic beta cells through autocrine and paracrine signalling. RT-qPCR analysis of the rat and human genes encoding for CXCL12 revealed no significant differences in rat Cxcl12 expression between wt and clone #1 cells while high expression of human Cxcl12 was observed in clone #1 cells, confirming its stable genomic integration (Fig. 1B). The existence of paracrine and autocrine signalling was confirmed by the high expression of Cxcr4 in clone #1 compared to wt cells (Fig. 1B).
The Comet assay was used to assess the levels of STZ-induced DNA damage (Fig. 1C). A considerable percentage (90-95%) of wt and clone #1 cells was without DNA tails ( Fig. 1C) which reflects the absence of DNA breaks. A 30 min treatment with 5 mM STZ was accompanied by increasing amounts of DNA in the Comet tails due to increased DNA damage. Estimation of the tail moments confirmed that DNA damage was significantly less pronounced in clone #1 than in wt cells.

Characterization of the Cxcl12 Gene Promoter Region
To elucidate the molecular mechanisms that regulate Cxcl12 transcription in pancreatic beta cells, we defined the Cxcl12 promoter sequence by computer analysis (Genomatix, Munich). The Genomatix prediction of the Cxcl12 promoter revealed a 739 bp long sequence. The prediction covered sequences between positions 21520023 and 21520762 (NCBI Reference Sequence: NC_005103.3). The predicted Cxcl12 promoter possesses a noncanonical TATA box (from 225 to 220 bp) with a cytosine instead of adenosine at the second position of the TATA box, as reported [39]. It also contains an initiator element (Int; from +13 to +20 bp), in agreement with the Int-consensus (KYAY*TCYYY) surrounding the secondary transcription start site. Downstream from the Int element lies a putative Kozak sequence (GCCATGG) containing the initiation ATG codon ( Fig. 2A), consistent with its vertebrate consensus (RCCATGG) [40].
PARP-1 and YY1 are Part of the Transcription Machinery that Regulates Cxcl12 Expression in vitro and in vivo PARP-1 and YY1 binding affinity toward the Cxcl12 promoter was examined by EMSA. Several nucleoprotein complexes were formed between the Cxcl12 promoter and nuclear proteins isolated from wt cells (Fig. 3A; lane 2). Super-shift analysis with different anti-PARP-1 and anti-YY1 antibodies revealed that PARP-1 ( Fig. 3A; lanes 3, 4, 5) and YY1 (lane 6) proteins were present in the nucleoprotein complexes.
ChIP analysis was employed to determine whether the protein-DNA interactions at the Cxcl12 promoter detected by EMSA also occur in vivo (Fig. 3B). To examine protein binding to the promoter at a higher resolution, the 739 bp promoter was divided into 3 fragments (Fig. 2B): fragment 1 (2546 to 2299 bp); fragment 2 (2319 to 254 bp); fragment 3 (274 to +193 bp). PARP-1 and YY1 binding to all three Cxcl12 promoter fragments was observed (Fig. 3B). Each promoter fragment contains at least one PARP-1 and YY1 binding site ( Fig. 2A, B). The putative Kozak sequence in the Cxcl12 promoter overlaps with the YY1 binding motif in the third promoter fragment, indicating that the third fragment contains the translation start site. The chromatin-associated Sp1 transcription factor, analysed as a positive Cxcl12 promoterbinding protein [39], was also found at the Cxcl12 promoter ( Fig. 3B; lane 4) which correlates with its multiple binding sites within Cxcl12 promoter ( Fig. 2A).

PARP-1 is an Inhibitor and YY1 a Strong Activator of CXCL12 Gene Transcription
To analyse Cxcl12 promoter activity, the promoter was cloned into a luciferase expression vector (pCXCL12luc) (Fig. 4A). Transfection of wt and clone #1 cells with pCXCL12luc showed slight decrease in Cxcl12 promoter activity in clone #1 cells. However, this reduction in promoter activity was not statistically significant, indicating that Cxcl12 does not influence its own expression (Fig. 4B). To examine the effect of PARP-1 on Cxcl12 transcription, promoter activity was examined in NIH3T3 fibroblasts (PARP +/+ ) and PARP-1 knockout fibroblasts (PARP 2/2 ) (Fig. 4C). Activity of the Cxcl12 promoter was expressed relative to the activity of the control CMV promoter. Functional analysis using the luciferase assay showed significantly enhanced (2-fold) Cxcl12 promoter activity in PARP 2/2 compared to PARP +/+ fibroblasts (Fig. 4C). This was confirmed when a PARP-1 expression construct (pEVC PARP) was introduced into PARP +/+ and PARP 2/2 cells which caused reduced expression of pCXCL121luc (Fig. 4E, F, respectively). These results indicate that PARP-1 downregulates Cxcl12 promoter activity.
To explore the influence of YY1, NIH3T3 fibroblasts were transfected with pCXCL12luc and cotransfected with the pcDNA3.1FLAGYY1 expression vector containing a YY1 expression unit (Fig. 4D). Cxcl12 promoter activity was enhanced 196fold when YY1 was overexpressed. This result suggests that YY1 strongly upregulates Cxcl12 promoter activity. YY1 (inset to

STZ-induced Toxicity in Pancreatic Beta Cells is Accompanied by Changed Regulation of Cxcl12 Promoter Activity
During the early stage of oxidative stress, i.e. after a 30 min treatment of wt cells with 5 mM STZ (EC 50 ), the mRNA levels for Cxcl12, Parp-1 and Yy1 decreased slightly (but not significantly) below the respective basal mRNA values (Fig. 5B, C, D; bar 2). To analyse the later stages of oxidative stress, wt cells were treated with 5 mM STZ for increasing times (1, 3, and 6 h). DNA damage was monitored by the Comet assay (Fig. 5A). The level of Cxcl12 transcription was estimated using RT-qPCR (Fig. 5B). As can be seen in Fig. 5A, DNA damage increased with the length of exposure of cells to STZ. Estimation of the amount of transcribed Cxcl12 at the respective time points confirmed that exposure to 5 mM STZ for 6 h caused a peak in Cxcl12 mRNA synthesis ( Fig. 5B; column 5), suggesting that when DNA is extensively damaged (Fig. 5A; column 5), Cxcl12 transcription is induced. The extended STZ treatment was accompanied by similar patterns of increased Yy1 and Parp-1 mRNAs (Fig. 5C, D). The Parp-1 mRNA level increased slightly after 6 h of STZ exposure, but remained below the control value ( Fig. 5C; column 5), while the Yy1 mRNA level was significantly above the control level ( Fig. 5D; column 5), suggesting that long-term exposure of pancreatic beta cells to STZinduced diabetogenic stimulation was accompanied by increased Yy1 transcription. The mRNA levels estimated by RT-qPCR were accompanied with analysis of the protein levels at the same time points using immunoblot analysis (Fig. 5E). Increased amount of YY1, due to its ability to upregulate Cxcl12 transcription, may be responsible for higher Cxcl12 expression also observed after 6 h of STZ treatment.
PARP-1 and YY1 binding to the Cxcl12 promoter after shortand long-term cell exposure to STZ was examined by the ChIP assay. Chromatin was isolated from Rin-5F wt control cells and cells exposed to 5 mM STZ for 30 min (early stage of oxidative stress) and 6 h (later stage of oxidative stress) (Fig. 6). The ChIP experiment revealed that the STZ treatment for 30 min lowered PARP-1 affinity for fragment 3 while its affinity for Cxcl12 fragments 1 and 2 remained as in the control. As a result of the 30 min exposure to STZ, YY1 exhibited very low binding to fragments 1 and 3 and no binding to fragment 2 (Fig. 6). The STZ treatment for 6 h revealed opposite binding pattern for PARP-1 and YY1 if compare to 30 min STZ treatment. PARP-1 was bound to Cxcl12 promoter fragment 1 but displayed no binding affinity to promoter fragments 2 and 3 (Fig. 6). As expected, YY1 displayed increased binding affinity for all three Cxcl12 promoter fragments, which resulted in the increased CXCL12 gene expression during extended (6 h) pancreatic beta cell injury.

Influence of PARP-1 Inhibition on Cxcl12 Promoter Regulation
PARP-1 enzymatic activity was chemically inhibited using 3AB, a general PARP-1 enzymatic activity blocker. Treatment of wt cells with 50 mM 3AB had no apparent effect on cell viability, while in the cells treated with 5 mM STZ we observed slight improvement in cell survival (Fig. 7A). When wt cells were treated with STZ together with 3AB, the disappearance of the necrotic PARP-1 fragment indicates that the necrotic pathway was turned off as a consequence of the inhibition of PARP-1 enzymatic activity (Fig. 7B). Besides pronounced PARP-1 apoptotic fragment, PARP-1 inhibition was accompanied with the activation of the main effector caspase 3 (Fig. 7B).
RT-qPCR analysis did not reveal significant changes in PARP-1 gene transcription in the presence of STZ, either alone or together with 3AB. Cxcl12 mRNA levels exhibited pronounced differences (Fig. 7C). The treatment with STZ led to upregulation of Cxcl12 while treatment with STZ and 3AB downregulated Cxcl12 transcription (Fig. 7C). We suggest that intensive binding of nonautomodified PARP-1 to the Cxcl12 promoter partially suppressed promoter activity. The ability of non-automodified PARP-1 to bind the Cxcl12 promoter was proven in an EMSA experiment using recombinant (non-automodified) PARP-1 (Fig. 7D; lane 2). In addition, we observed that YY1 was not ADP-ribosylated, in both physiological and STZ-compromised conditions (Fig. 7E) that allows its binding to Cxcl12 promoter and subsequent induction of the Cxcl12 expression in the later stage of oxidative stress.

Discussion
The positive impact of the elevated expression of the chemokine CXCL12 on prosurvival and proliferative phenotype in pancreatic islet cells has been observed recently [5]. Based on our experimental results and information from other studies [5,6,7], we confirmed importance of CXCL12 as a pancreatic beta cell prosurvival factor. Going a step further, we clarified the role of PARP-1 and YY1 in the regulation of the Cxcl12 transcription, suggesting that transcriptional activation of the Cxcl12 promoter certainly depends on the finely balanced functional interplay of these proteins.
For the first time we defined and analysed the promoter sequence of the rat CXCL12 gene using computer analyses (Genomatix, Munich). Up to now, transcriptional regulation of the Cxcl12 promoter has been described in human [39,41,42,43,44] and mouse [45,46] cells. Their results on mammalian Cxcl12 promoter characterization and our data analysis show that Cxcl12 promoter contains a non-canonical TATA box and a downstream Int element for the initiation of transcription [39,45] that can function as an alternative promoter in eukaryotic genes that lack the classical TATA box [47]. In the rat Cxcl12 promoter, Int is located 31 bp downstream from the non-canonical TATA box, while in the human this distance is 26 bp [39]. The Kozak sequence (CCATGG) with the contained ATG initiation codon was designated as a translation start site [40]. In human and mouse Cxcl12 promoters, Kozak sequence is not particularly defined so far [39,40]. We observed the high degree of similarity between human and rodent Kozak sequence, although the Kozak sequence is not strictly conserved in eukaryotic mRNAs [48]. As in the mouse Parp-1 promoter [17] we also observed overlap of the YY1 core-binding motif (ATG) with the Kozak sequence (GCCATGG) in the rat Cxcl12 promoter. This is in agreement with the colocalization of the YY1 motif and the translation start site in many human promoters [28]. Our analysis of the organization of the rat Cxcl12 promoter Figure 3. PARP-1 and YY1 binding affinity toward the Cxcl12 promoter. (A) EMSA was performed with end-protected (biotinylated) fragments to permit PARP-1 binding exclusively to DNA-internal motifs. Super-shift analysis was performed using anti-PARP-1 and anti-YY1 antibodies as indicated. (B) ChIP analysis was performed with PARP-1 and YY1 antibodies. PARP-1 and YY1 binding was verified for each Cxcl12 promoter fragment; fragment 1 is 246 bp; fragment 2 is 265 bp; fragment 3 is 268 bp. For immunoprecipitation, RNA pol II served as a positive, and IgG as a negative control. In the PCR reaction the positive control was genomic DNA (input); a water-only (blank) was the negative control. doi:10.1371/journal.pone.0059679.g003 also revealed the presence of GC-rich sequences in the 5'-flanking region, as reported previously [39]. Ubiquitous expression of the Cxcl12, except in blood cells, is consistent with the presence of the GC-rich sequence in the 5'-flanking region and the non-canonical TATA box, common features of housekeeping genes, as we also observed for the mouse PARP-1 gene [17]. Up to now several transcription factors involved in the regulation of CXCL12 gene transcription have been described in human and mouse. The transcription factors of the early B cell factor (EBF) family [45], Sp1 [39], STAT3 [46], c-myb [43], C/ EBPb [41,42] and Ets-related molecule [49] were defined as promoter activators while transcription factors Foxf1 [50] and p53 [51] were identified as a promoter suppressors. The 1 kb long proximal human Cxcl12 promoter possesses 15 putative Sp1 binding motifs, of which 6 are important for basal human Cxcl12 expression [39]. Our computer analysis confirmed the presence of 8 putative Sp1 binding sites within the Cxcl12 promoter, all surrounding the non-conventional TATA box. In our study, transcription factor-binding site analysis showed several putative binding sites within the Cxcl12 promoter for the following transcription factors: C/EBPb, C/EBPa, FOXO3a, HMGI/Y, p53, STAT3 and NF-kB. EMSA experiments and ChIP assays confirmed binding affinities for some of these transcription factors that will be published elsewhere. Although PARP-1 has been characterized as a transcription factor only recently, an increasing number of reports have been published, indicating that PARP-1 can also bind to DNA in a sequence-specific manner. While several PARP-1 binding motifs have been published (reviewed in [17,18,21,22,23]), a DNA binding consensus sequence for PARP-1 is still not available in the existing computer analysis tools. For that reasons we analysed the previously published PARP-1 DNA binding sequences [17,22].
Our transfection experiments clearly show that PARP-1 has a pivotal role in partial suppression of the Cxcl12 promoter, allowing for its constitutive expression. During the oxidative stress, due to PARP-1 automodification and the negative charge of ADP-ribose polymers, PARP-1 is detached from the Cxcl12 promoter thus allowing for enhanced Cxcl12 transcription. This momentum of PARP-1 release from the Cxcl12 promoter allows for increased gene transcription and places PARP-1 in the position of a weak suppressor, enabling basal promoter activity. The suppressive effect of PARP-1 on gene transcription was observed for the Tracp gene in pre-osteoclastic cells [52] and for its own gene in mouse fibroblasts [17]. Our observation is in agreement with Amiri et al. [53] who revealed the suppressive effect of PARP-1 on CXCL1 gene expression. In contrast, Nirodi et al. [21] reported that PARP-1 may act as a coactivator of CXCL1 gene transcription. Besides Cxcl12, PARP-1 is also involved in the regulation of several other diabetes-related genes. PARP-1 acts as a corepressor for the Foxo1 gene, which could play an important role in proper cell proliferation and in the response to oxidative stress [54]. Furthermore, Akiyama et al. [22] demonstrated that PARP-1 forms the active complex for Reg transcription with some nuclear proteins, and that complex formation was stabilized when PARP-1 was not automodified.
The CXCL12 gene was initially considered to be constitutively expressed. However, it was recently established that its transcription is induced by cell injury [6], in response to cytokines and cell confluence [41] and during cell growth arrest and hypoxia [55]. Recently, Liu et al. [6] observed short-term influence of exogenous CXCL12 on the induction of its own gene expression. In our experiments, long-term (chronic) presence of CXCL12 in clone #1 cells has no autoregulatory transcriptional potential. Under conditions of YY1 overexpression, which is followed by its increased binding to Cxcl12 promoter, we observed highly increased Cxcl12 promoter activity. This lends further support to the concept that the Cxcl12 promoter is inducible, i.e., characterized by a low basal activity that greatly increases upon induction. Baumeister et al. [56] reported similar pattern of transcriptional regulation by YY1 concerning Grp78 gene promoter regulation. The authors proposed that YY1 has no effect on Grp78 promoter basal activity, however in cells undergoing ER stress YY1 strongly enhances Grp78 promoter induction. The observation that both YY1 [35] and CXCL12 genes [55] are upregulated by hypoxia points to the interrelatedness of their responses to stress signals. In contrast to its role in Cxcl12 transcription, YY1 represses the activities of Cxcr4 and Cxcr7 promoters through its binding to upstream regions on their promoters [35]. As a potential element of a negative feed-back mechanism controlling the CXCL12/ CXCR4 axis, this action represents an additional level of control of Cxcl12 expression by YY1. Therefore, we believe that YY1 could be a major regulator which helps beta cells to transcribe important proteins that help coping with oxidative stress. In addition, strong transcriptional induction of the prosurvival chemokine could be the pancreatic cell's response to severe oxidative stress.
We have summarized the apparently opposing effects of PARP-1 and YY1 on Cxcl12 transcription in a model presented in Fig. 8. Our findings suggest that in the physiological state Cxcl12 transcription is supported by the binding of non-automodified PARP-1 and YY1 to the Cxcl12 promoter. We hypothesize that under basal conditions when PARP-1 activity in pancreatic beta cells is negligible, the bound non-automodified PARP-1 offsets the elevation in Cxcl12 transcription that occurs in the presence of YY1 (Fig. 8A). In response to cytotoxic signals, as PARP-1 enzymatic activity rises, PARP-1 automodification progressively lowers PARP-1 affinity for the Cxcl12 promoter, since PARP-1 automodification blocks the ability of PARP-1 to bind to DNA [57] (Fig. 8B). This has an overall permissive effect on Cxcl12 transcription. However, short-lasting exposure to STZ-induced diabetogenic signals was accompanied by a decrease in YY1 binding to DNA and consequent removal of its stimulatory effect on Cxcl12 expression (Fig. 8B). Prolonged exposure to cytotoxic signals was followed by significantly enhanced PARP-1 automodification and decreased binding affinity for Cxcl12 promoter (Fig. 8C). In contrast, prolonged pancreatic beta cell injury is followed by increased YY1 binding to Cxcl12 promoter that will cause upregulation of the CXCL12 gene expression (Fig. 8C). It would appear that at relatively low levels of cytotoxicity, an increase in Cxcl12 transcription that would result from the lifting of transcriptional suppression by PARP-1, is tempered by the withdrawal of YY1 stimulation. In contrast, severe beta cell injury was associated with increased Yy1 expression that was followed by increased binding of YY1 protein to the Cxcl12 promoter. The promoter assumed a more relaxed conformation due to a major increase in PARP-1 activity with resulting ADP-ribosylation of chromatin proteins, and the complete lifting of transcriptional suppression as a result of extensive PARP-1 automodification. The sum effect of these events is the induction of Cxcl12 expression and subsequent prosurvival actions of CXCL12.

Concluding Remarks
Several transcription factors involved in beta cell functioning, differentiation, proliferation and survival have been identified so far (reviewed in [58]). Novel drugs, which enhance the expression of key transcription factors, as in the case of Pdx-1 [59], could restore beta cell functions in diabetic patients. We believe that the present study is particularly relevant since two transcription factors that have been identified as important regulators of Cxcl12 transcription could in the future serve as focal points for targeting Cxcl12 expression. An open challenge is to find a way of inducing and/or suppressing transcription factors that participate in the transcriptional regulation of genes that could improve beta cell functioning. This strategy could advance the treatment of diabetes. In the basal state, non-automodified PARP-1 (acting as a transcriptional inhibitor) and YY1 (acting as a transcriptional activator) strongly bind to the Cxcl12 promoter, enabling its transcription (Fig. 3, 4). (B) DNA damage induced by STZ (diabetogenic-like) treatment causes PARP-1 binding to the DNA breaks, leading to a net increase in PARP-1 activity and consequently to poly(ADP-ribosyl)ation of PARP-1 and associated chromatin proteins. After 30 min of STZ treatment, DNA is moderately damaged (Fig. 5A), which results in weak poly(ADP-ribosyl)ation so that most of PARP-1 molecules remain attached to the promoter, however, as YY1 partially dissociates from the promoter (Fig. 6), Cxcl12 transcription is still at the basal level (Fig. 5B). (C) After 6 h of STZ treatment, the DNA is severely damaged (Fig. 5A), causing intensive poly(ADP-ribosyl)ation of both PARP-1 and associated chromatin proteins and PARP-1 dissociation from the promoter, with resulting opening of the chromatin. Severe beta cell injury induces YY1 expression (Fig. 5D, E). Furthermore, the open chromatin structure enables intense YY1 binding to the promoter (Fig. 6), upregulation of Cxcl12 expression (Fig. 5B, E) and consequent increased cell survival. Treatment with PARP-1 inhibitor (3AB) causes reduced poly(ADP-ribosyl)ation and intense PARP-1 binding to the promoter, resulting in decreased Cxcl12 transcription (Fig. 7C). doi:10.1371/journal.pone.0059679.g008 dizin, Berlin, Germany), for providing a expression vector for FLAGtagged human YY1 and to Professor Dr. Juergen Bode (Department of Experimental Hematology, Hannover Medical School, Hannover, Germany) for constant scientific advising and friendly support through the years.