Retraction: Human Cystathionine-β-Synthase Phosphorylation on Serine227 Modulates Hydrogen Sulfide Production in Human Urothelium

The authors cooperated with the editorial follow up by providing the western blot images underlying Fig 5C, along with replicate blots for Fig 5C. The corresponding author has stated that the height and width of the image panels were reset during figure preparation, resulting in a horizontally stretched appearance of the bands. An editorial evaluation of the provided images, to the extent allowed by the image resolution, could not fully resolve the concerns about the images in Fig 5C. In light of the unresolved concerns about the western blot panels of Fig 5C, the PLOS ONE Editors retract this article. REVB, EM, DE, AR, FF, GC, GR, and RS did not agree with retraction. ED, AI, and VM did not respond.

at serine 227. Incubation of human urothelium or T24 cell line with 8-Bromo-cyclic-guanosine monophosphate (8-Br-cGMP) but not dibutyryl-cyclic-adenosine monophosphate (d-cAMP) causes an increase in hydrogen sulfide production. This result is congruous with the finding that PKG is robustly expressed but PKA only weakly present in human urothelium as well as in T24 cells. The cGMP/PKG-dependent phosphorylation elicited by 8-Br-cGMP is selectively reverted by KT5823, a specific PKG inhibitor. Moreover, the silencing of cystathionine-βsynthase in T24 cells leads to a marked decrease in hydrogen sulfide production either in basal condition or following 8-Br-cGMP challenge. In order to identify the phosphorylation site, recombinant mutant proteins of cystathionine-β-synthase in which Ser32, Ser227 or Ser525 was mutated in Ala were generated. The Ser227Ala mutant cystathionine-β-synthase shows a notable reduction in basal biosynthesis of hydrogen sulfide becoming unresponsive to the 8-Br-cGMP challenge. A specific antibody that recognizes the phosphorylated form of cystathionine-β-synthase has been produced and validated by using T24 cells and human urothelium. In conclusion, human cystathionine-β-synthase can be phosphorylated in a PKGdependent manner at Ser227 leading to an increased catalytic activity.

Introduction
Hydrogen sulfide (H 2 S) is the third member of the gasotransmitter family that also includes nitric oxide and carbon monoxide [1]. H 2 S is primarily generated by two pyridoxal 5'-phosphate dependent-enzymes, cysthationine-β-synthase (CBS) and cysthationine-γ-lyase (CSE), which employ cysteine and homocysteine as substrates [2][3][4]. The H 2 S pathway is involved in many physiological and pathological processes in different organs and apparatuses in animal experimental models and in human [5][6][7][8]. Recently, a role for the L-cysteine/H 2 S pathway has been proposed in human urogenital tract [9][10][11][12][13][14][15]. Within this context, we have previously demonstrated that CBS and CSE are both expressed in the human bladder and that either an H 2 S donor or L-cysteine relaxes human bladder strips. In addition, a prominent role for CBS in the human bladder has also been postulated [10]. It has also been shown that the incubation of full human thickness bladder samples with 8-Bromo-cyclic-guanosine monophosphate (8-Br-cGMP), a stable analogue of cGMP, or dibutyryl-cyclic-adenosine monophosphate (d-cAMP), a stable analogue of cAMP, triggered a significant increase in H 2 S production [10]. These data suggested that the mechanism underlying this effect could involve post-translational modifications. Several enzymes and protein are regulated through post-translational modifications [16][17][18]. In the current literature a post-translational activation mechanism for CBS has been demonstrated. Indeed, CBS activity is regulated by post-translational modifications through a small ubiquitin-like modifier protein which is correlated with the localization of CBS in the nucleus leading to diminished catalytic activity [19,20]. Post-translational activation of CBS in response to oxidative stress has also been demonstrated [21] but the molecular mechanism(s) and its pertinence to H 2 S generation are not yet known.
In the recent literature, urothelium has been shown to play an important role in the physiopathology of human bladder. The urothelium is the epithelium lining the surface of the urinary bladder, in close contact with the urine. Urothelial cells are specialized to detect both physical and chemical stimuli and their transducer role is enhanced by their close proximity to the urothelium of afferent and efferent autonomic nerves [22]. To date, the functions ascribed to the urothelium include control of permeability, immune responses and cell-cell communication and which seem to play a pivotal role in responding to injuries and infections [23]. Starting from our previous findings, by using human urothelium and T24 human urothelial cell line, herein we define that i) CBS is the main enzyme involved in generating H 2 S ii) the activation of the cGMP/PKG pathway leads to CBS phosphorylation at Ser227, thereby increasing its activity.

Human tissue
Full thickness bladder dome samples were obtained from patients, aged 61-73 years, affected with benign prostatic hyperplasia that underwent open prostatectomy. All patients presented urodynamic obstruction and large prostate volume (>80 ml). We excluded patients affected by bladder stones, urinary infections, detrusor areflexia and a history of urothelial cancer. The experimental protocol was approved by the Local Ethical Committee (School of Medicine and Surgery, University of Naples Federico II, via Pansini, 5; 80131, Naples, Italy). All patients were informed of all procedures and gave their written consent. The samples were cleared of adherent tissue and urothelium and detrusor were carefully dissected, separated and immediately frozen [10].
quantization. Final preparation of RNA was considered DNA-and protein-free if the ratio between readings at 260/280 nm was 1.7. Isolated mRNA was reverse-transcribed by use of iScript Reverse Transcription Supermix for quantitative real-time RT-PCR (Bio-Rad, Milan, Italy). The quantitative RT-PCR was carried out in CFX384 real-time PCR detection system (Bio-Rad, Milan, Italy) with specific primers by the use of SYBR Green master mix kit (Bio-Rad, Milan, Italy). Samples were amplified simultaneously in triplicate in one-assay run with a non-template control blank for each primer pair to control for contamination or primerdimers formation, and the ct value for each experimental group was determined. The housekeeping gene (ribosomal protein S16) was used as an internal control to normalize the ct values, using the 2 -ΔCt formula [24].

CBS Silencing
To stably silence CBS, T24 cells were seeded the day before transfection in a 60 mm plate. Cells were transfected 24 hours later, with 2 μg of a CBS short hairpin (sh)RNA expressing vector purchased from Santa Cruz Biotechnology (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Transfected cells were selected with 1 μg/ml of puromycin (Sigma-Aldrich) for 7 days and then the CBS depletion was evaluated by western blotting using anti-CBS (Santa Cruz Biotechnology, DBA, Milan, Italy). The cell clone with lowest expression level was selected and used for further experiments. A short hairpin non-silencing construct was used as control.

Mutagenesis
The WT-CBS cDNA, was cloned into a version of the eukaryotic expression vector pcDNA4/ HisMax C (Invitrogen, Life Technology, Monza, Italy) containing the HA epitope. The constructs containing the substitution of single Ser with Ala, i.e. TCC/GCC mutation, were obtained by PCR site-directed mutagenesis using the QuickChange 1 Lightning Site-Directed Mutagenesis Kit (Agilent Technology, Agilent Technology, Milano, Italy) and the WT-CBS cDNA as template as previously described [27]. All constructs were verified by DNA sequencing.
Immunoprecipitation and western blotting T24 cells were starved for 16 hours, incubated with 100 μM of 8-Br-cGMP or d-cAMP (Tocris, UK) for 5, 15 or 30 min and then used for immunoprecipitation assay. Samples were analyzed as previously described [28,29]. The eluted proteins were used in cell free kinase assays, phosphorylation assay in T24 cells and western blotting analysis.

Phosphorylation assay in T24 cells
Starved T24 cells were incubated with PKG inhibitor, KT5823 (10 μM), or PKA inhibitor, KT5720 (10 μM) and then treated with 8-Br-cGMP (100 μM) or d-cAMP (100μM) for 5, 15 or 30 min respectively. Next, CBS was specifically immunoprecipitated from protein extracts by using antibodies against the endogenous protein. Immunoprecipitated proteins were separated by SDS-PAGE and the presence of a phosphorylated CBS was evaluated by western blotting using anti-phospho-(Ser/Thr) (Cell Signaling Technology, DBA, Milan, Italy).
Starved T24 cells were transiently transfected with constructs encoding for HA-CBS WT or HA-CBS mutants. After transfection (24h), cells were treated or not with 100 μM of 8-Br-cGMP for 15 min. Samples were collected and WT or mutant proteins were specifically immunoprecipitated by using antibodies against the epitope HA. Immunoprecipitated proteins were separated by SDS-PAGE and the presence of the proteins and their phosphorylated form was investigated by western blotting analysis using anti-phospho-(Ser/Thr).

Generation and characterization of specific antibody against phosphorylated CBS
Peptide from amininoacids 225-234 (NAS 227 NPLAHYD) containing a single phosphorylated Ser at position 227 has been selected to obtain polyclonal antibodies in rabbits, namely anti-pCBS Ser227 (PRIMM srl, Milano, Italy). The purified antibody was used at 1:400 dilutions for western blotting.

Statistical analysis
All data were calculated and expressed as mean±SE. The results were analyzed using t Student' test or analysis of variance (ANOVA) followed by Bonferroni, as needed. A p value <0.05 was considered significant.

Human tissue
CBS and CSE are expressed in human urothelium. Human urothelium expresses both CBS and CSE as demonstrated by western blot (Fig 1A) and quantitative RT-PCR (Fig 1B). Human urothelium homogenates generate detectable amount of H 2 S in basal condition (i.e. in presence of the vehicle). The incubation with L-cysteine (the substrate) significantly increases H 2 S production (p<0.01, Fig 1C).
8-Br-cGMP but not d-cAMP elevates H 2 S levels in human urothelium. 8-Br-cGMP, a stable analogue of cGMP, significantly increases H 2 S generation compared to the vehicle (p<0.05, Fig 1D). This effect is reverted by KT5823, a selective inhibitor of PKG (p<0.05, Fig  1D). Conversely, H 2 S levels are not modified by the stimulation with d-cAMP, a stable analogue of cAMP ( Fig 1D). PKA blockade, operated by the selective inhibitor, KT5720, does not modify H 2 S production ( Fig 1D).
Human urothelial T24 cell line CBS and CSE are expressed in human urothelial T24 cell line. T24 cell line expresses CBS and CSE (Fig 2A). The incubation of T24 cell homogenate with the vehicle generates detectable amount of H 2 S (Fig 2B). Following incubation with L-cysteine, H 2 S synthesis is significantly increased in comparison to the vehicle (p<0.05, Fig 2B).
8-Br-cGMP but not d-cAMP elevates H 2 S levels in human urothelial T24 cell line. T24 cells treated with 8-Br-cGMP show a significant increase in H 2 S production as compared to vehicle peaking at 15 min of incubation (p<0.01 and p<0.05, at 15 and 30 min respectively, Fig 2C). The PKG inhibitor, KT5823, abolishes the 8-Br-cGMP-induced effect (p<0.001, Fig  2C). Incubation of T24 cells with either d-cAMP or KT5720, the PKA inhibitor, does not affect H 2 S production ( Fig 2C).
CBS silencing. T24 cells were depleted of CBS by stably transfecting shRNA specific for CBS. The clone with the lowest expression level e.g. clone F, namely CBSΔT24 cells, was selected for the subsequent experiments ( Fig 3A). The H 2 S production observed in T24 cell is virtually abolished in CBSΔT24 cells (p<0.05, Fig 3B).
CBS phosphorylation. T24 cell incubation with 8-Br-cGMP causes an increase of pCBS at the 15 and 30 min time points (p<0.001 and p<0.05 respectively, Fig 4A). The maximum effect is reached after 15 min of incubation with 8-Br-cGMP. Treatment with the PKG inhibitor, KT5823, abolishes CBS phosphorylation induced by 8-Br-cGMP (p<0.001 and p<0.01 at 15 and 30 min respectively, Fig 4A). In contrast, neither d-cAMP nor KT5720, the PKA inhibitor, modifies the levels of pCBS (Fig 4B).
PKG and PKA expression in human urothelial T24 cell, urothelium, detrusor and full thickness bladder. Western blot analysis clearly showed the expression of PKG in all examined samples i.e. human full thickness bladder, urothelium and detrusor, as well as in T24 cells (Fig 4C). Conversely a robust signal for PKA is only detected in the human detrusor muscle and in the full thickness bladder homogenates but not in human urothelial T24 cells or in human urothelium (Fig 4C).
PKG-mediated phosphorylation of Ser227 and Ser525 of CBS in cell free assay. Among the sites predicted by computational analysis (GPS 2.1, Group-based Prediction System) [35], we chose those with the highest score e.g. Ser32, Ser227 and Ser525 (Fig 5A). Three constructs encoding for mutated forms of CBS in which Ser32, Ser227 or Ser525, mutated in Ala namely HA-CBS S32A, HA-CBS S227A and HA-CBS S525A, were generated. An high significant reduction of labeling with 32 P of HA-CBS S227A and HA-CBS S525A but not HA-CBS S32A proteins compared with HA-CBS WT is observed (p<0.001, Fig 5B), as revealed by cell free phosphorylation assay.
PKG   (p<0.001, Fig 5C). Of note, the PKG/cGMP-induced phosphorylation of HA-CBS S227A and HA-CBS S525A is significantly and markedly reduced when compared with HA-CBS WT (p<0.01, Fig 5C). The phosphorylation levels of HA-CBS S227A and HA-CBS S525A mutants are still increased as compared to their matched control (p<0.001, Fig 5C).
The mutation of Ser227 impairs PKG-mediated increase of H 2 S production. CBSΔT24 cells were transiently transfected with constructs HA-CBS WT, HA-CBS S227A or HA-CBS S525A mutants and treated with 8-Br-cGMP or vehicle. Ectopic expression of HA-CBS WT or HA-CBS mutants was assayed by western blotting using anti-HA (Fig 6A). Homogenates of HA-CBS WT generate detectable amount of H 2 S; as expected, 8-Br-cGMP causes a significant increase in H 2 S production (p<0.05, Fig 6B). Of note, mutation of Ser227 led to two effects: i) a significant reduction of H 2 S basal production (p<0.05) ii) the abrogation of 8-Br-cGMPmediated increase in H 2 S production (p<0.001) both compared to WT (Fig 6B). In contrast, mutation of Ser525 does not affect H 2 S production either in basal condition or following 8-Br-cGMP stimulation. In fact, the H 2 S level in these cells is similar to that observed in homogenates from cells transfected with HA-CBS WT.
Characterization of the antibody for pCBS at Ser227. A peptide of 10 aminoacids corresponding to residues 225 through 234 of CBS including phosphorylated Ser227 was used to obtain rabbit polyclonal antibody. The phosphopeptide-specific CBS antibody (anti-pCBS Ser227 ) efficiently recognizes CBS protein in human urothelium and in T24 cells (Fig 7A  and 7B). Of note, a more robust pCBS specific signal is detected by anti-pCBS Ser227 either in human urothelium homogenates (Fig 7A) or in T24 cell lysates treated with 8-Br-cGMP ( Fig  7B and 7C). No difference in signal intensity is appreciable in the same samples incubated with the anti-CBS (Fig 7A and 7B). No other proteins were detected (data not shown). Specificity of the antibody for the phospho-epitope of CBS was further evaluated in T24 cells transiently transfected with constructs HA-CBS WT or HA-CBS S227A. As shown in Fig  7D, the immunoprecipitated Ser227Ala mutant CBS protein was not recognized by the anti-pCBS Ser227 . This latter results demonstrates that the presence of phosphorylation at Ser227 site is necessary for the detection of CBS protein by this specific antibody after 8-Br-cGMP challenge.

Discussion
We have previously shown that incubation of human bladder tissue with the stable analogues of cGMP or cAMP, i.e. 8-Br-cGMP or d-cAMP, leads to increased H 2 S production [10]. It is well established that cGMP and cAMP activate PKG or PKA respectively, which in turn phosphorylate proteins thereby triggering signal transduction [36]. On this basis, we have hypothesized that the effect triggered by 8-Br-cGMP or d-cAMP involves phosphorylation of CBS, leading to an increased catalytic activity.
In order to address this issue, we focused our attention on the urothelium. The choice of the urothelium was driven by the following reasons i) the important role played by the urothelium    As a first step we defined whether CBS and/or CSE were involved in driving the H 2 S pathway in human urothelium, comparing this profile with T24 cell line. We found that both human urothelium and T24 cells express CBS and CSE and efficiently convert L-cysteine into H 2 S. Having determined that both enzymes are expressed and are catalytically active, we assessed whether incubation with 8-Br-cGMP or d-cAMP increases H 2 S production. We have demonstrated that 8-Br-cGMP but not d-cAMP causes a significant increase in H 2 S production in both urothelium and T24 cells. The specificity of the effect of 8-Br-cGMP was confirmed by the reversion operated by KT5823, a selective inhibitor of PKG. These results differed from our previous findings where both 8-Br-cGMP and d-cAMP caused an increase in H 2 S production in full human bladder thickness (detrusor muscle plus urothelium) [10]. In order to clarify this issue, we investigated the distribution/expression of PKA versus PKG in the human bladder. Western blot analysis for PKG and PKA in human urothelium, detrusor muscle, whole bladder and T24 cells clearly show that urothelium as well as T24 cells strongly express PKG as opposite to PKA. This finding explains the lack of the effect of d-cAMP observed in human urothelium and urothelial T24 cells. Having well characterized that human urothelium and T24 cell line act similarly, we have sought to investigate CBS phosphorylation by stimulating T24 cells with 8-Br-cGMP. A clear time-dependent signal of CBS phosphorylation was observed in CBS isolated from T24 cells stimulated with 8-Br-cGMP. The phosphorylation reached the maximum effect at 15 min declining thereafter. CBS phosphorylation in presence of KT5823, a PKG inhibitor, was blunted. In agreement with the finding that both human urothelium and T24 cells have a very low level of PKA, d-cAMP did not trigger CBS phosphorylation. These data strongly indicate that CBS phosphorylation is PKG-dependent. Due to the lack of selective CBS inhibitors [37], in order to further characterize the role of CBS-derived H 2 S in urothelium, we stably silenced CBS in T24 cells. CBS silencing markedly reduces H 2 S production, confirming the major role played by CBS in triggering the H 2 S production. In summary, we have demonstrated that i) T24 cell line mimics human urothelium behaviour ii) cGMP/PKG pathway activation triggers CBS phosphorylation iii) CBS-derived H 2 S is the major source of H 2 S in T24 cells. Through a computational approach, we searched for possible PKG-phosphorylation sites in CBS protein, focusing on the aminoacid sequence known as PKG target consensus, distinct from sequence predicted as PKA-phosphorylation sites. Predicted PKG-phosphorylation sites in CBS were Ser32, Ser227, and Ser525. To experimentally validate these Ser as sites of phosphorylation by PKG, we performed a phosphorylation assay by using proteins from T24 cells transiently transfected with plasmids expressing HA-CBS WT or HA-CBS mutants (HA-CBS S32A, HA-CBS S227A, HA-CBS S525A). The mutation of Ser227 and Ser525 but not Ser32 causes a drastic reduction of CBS phosphorylation. Furthermore, HA-CBS S227A and HA-CBS S525A mutants display a significantly reduced level of phoshorylation when incubated with 8-Br-cGMP. Taken together, these findings show that among predicted PKGphosphorylation sites, only Ser227 and Ser525 were experimentally validated as sites of phosphorylation by PKG.
To assess the contribution of each phosphorylation sites to the enzyme activity and ultimately to the extent of H 2 S production, we transiently transfected cells lacking CBS (CBSΔT24) with HA-CBS S525A, HA-CBS S227A mutants or HA-CBS WT as a control. Treatment with 8-Br-cGMP did not induce a significant production of H 2 S in the S227A mutant as opposite to S525A mutant. Indeed, the production of H 2 S in S525A mutant is comparable to that induced by 8-Br-cGMP in CBSΔT24 cells transfected with HA-CBS WT. Therefore only the phosphorylation at Ser227 in CBS protein increased H 2 S production.
In order to further confirm our data we generated a specific antibody against the phosphorylated form. The anti-pCBS Ser227 recognizes the CBS phosphorylated form in both urothelium and T24 cell line. Challenge of T24 cell with 8-Br-cGMP causes a clear increase in the signal of the anti-pCBS Ser227 confirming and validating our previous findings. The specificity of the antibody was also confirmed by the finding that Ser227Ala mutant is not recognized by the anti-pCBS Ser227 . Thus the phosphorylation at Ser227 site is necessary for recognition of CBS protein by this specific antibody. Interestingly even though Ser525 is also phosphorylated in a PKGdependent manner, mutation at this residue does not affect H 2 S production. Ser 525 lies within a non-consensus CBS domain that is described as a regulatory region target of the allosteric modulator S-adenosyl methionine. Therefore, it is reasonable to postulate that phosphorylation of Ser 525 may be implicated in the modulation of additional features of CBS protein not as yet determined.
Although additional experiments will be required to understand the control of CBS function by phosphorylation, it is evident that this covalent modification increases CBS activity. Such an outcome may include conformational changes which could allow more efficient binding to substrates or allosteric modulators and regulation of potential binding partners. Herein we demonstrates that in human urothelium as well as in T24 cells CBS-dependent H 2 S generation involves cGMP/PKG mediated phosphorylation of CBS at Ser227.
In the current literature it is widely accepted that the urothelium influences the contractile state of detrusor smooth muscle [23,38]. Indeed, in several studies performed with human or animal tissues, it has been demonstrated that removal of the urothelium increases the contractile response to several different agonists [39,40]. Thus urothelium constitutively releases agents that modulate muscle contractility thereby contributing in a dynamic manner to bladder function homeostasis. Since cGMP/PKG dependent CBS phosphorylation occurs in human urothelium and regulates H 2 S production, we suggest that this mechanism dynamically contributes to bladder homeostasis by controlling detrusor muscle contractility [10].
Finally, the results of this study may not solely be not confined to the bladder since they introduce the concept that changes in H 2 S production driven by CBS, often ascribed in the relevant literature to a generic change in activity, could involve CBS phosphorylation.

Conclusions
In conclusion, we have demonstrated that i) incubation of T24 cells with 8-Br-cGMP causes a time-dependent increase in CBS phosphorylation reverted by selective inhibition of PKG ii) in T24 cells stably silenced for CBS, H 2 S production is markedly reduced thus supporting a relevant contribution of CBS in triggering H 2 S production in urothelial cells iii) CBS phosphorylation by PKG at Ser227 increases endogenous H 2 S production confirming that, CBS activity can be selectively enhanced in a cGMP/PKG-dependent manner in human urothelium.
Finally, the antibody, directed against the phosphorylated form of CBS, may represent a useful tool to define the pathophysiological role of the H 2 S pathway, in any pathology where CBS is reputed to play a role.