Ubiquitin carboxyl-terminal hydrolase 1 (UCHL1) is a potential tumour suppressor in prostate cancer and is frequently silenced by promoter methylation
© Ummanni et al; licensee BioMed Central Ltd. 2011
Received: 15 May 2011
Accepted: 14 October 2011
Published: 14 October 2011
We have previously reported significant downregulation of ubiquitin carboxyl-terminal hydrolase 1 (UCHL1) in prostate cancer (PCa) compared to the surrounding benign tissue. UCHL1 plays an important role in ubiquitin system and different cellular processes such as cell proliferation and differentiation. We now show that the underlying mechanism of UCHL1 downregulation in PCa is linked to its promoter hypermethylation. Furthermore, we present evidences that UCHL1 expression can affect the behavior of prostate cancer cells in different ways.
Methylation specific PCR analysis results showed a highly methylated promoter region for UCHL1 in 90% (18/20) of tumor tissue compared to 15% (3/20) of normal tissues from PCa patients. Pyrosequencing results confirmed a mean methylation of 41.4% in PCa whereas only 8.6% in normal tissues. To conduct functional analysis of UCHL1 in PCa, UCHL1 is overexpressed in LNCaP cells whose UCHL1 expression is normally suppressed by promoter methylation and found that UCHL1 has the ability to decrease the rate of cell proliferation and suppresses anchorage-independent growth of these cells. In further analysis, we found evidence that exogenous expression of UCHL1 suppress LNCaP cells growth probably via p53-mediated inhibition of Akt/PKB phosphorylation and also via accumulation of p27kip1 a cyclin dependant kinase inhibitor of cell cycle regulating proteins. Notably, we also observed that exogenous expression of UCHL1 induced a senescent phenotype that was detected by using the SA-ß-gal assay and might be due to increased p14ARF, p53, p27kip1 and decreased MDM2.
From these results, we propose that UCHL1 downregulation via promoter hypermethylation plays an important role in various molecular aspects of PCa biology, such as morphological diversification and regulation of proliferation.
Keywordsprostate cancer UCHL1 ubiquitin system tumour suppression signalling
List of abbreviations
Ubiquitin Carboxyl-Terminal Hydrolase 1
Prostate cancer (PCa) is the most common type of cancer found in men and is among the leading causes of cancer death in the western world . The specific causes of prostate cancer remain poorly understood . Recently, our group identified differentially expressed proteins which are significantly deregulated in PCa predicting their role in initiation and progression of PCa . Among those proteins several members of the ubiquitin system have shown an altered expression. Ubiquitination of proteins has emerged as one of the most versatile post-translational modifications, regulating a diverse arrays of cellular processes . Ubiquitination plays a central role in degradation of proteins both through proteasomal targeting and by lysosomal degradation. In recent years, it became clear that deubiquitination is a crucial process in multiple intracellular signaling pathways, resulting in putative oncogenic or tumor suppressive functions . Deubiquitination of proteins is catalyzed by a set of enzymes known as deubiquitinases (DUBs). In the human genome approximately one hundred human DUBs are known so far classified into five categories: ubiquitin specific proteases (USP), ubiquitin C-terminal hydrolases (UCH), ovarian tumour proteases (OTU), Josephins and the Jab1/MPN/MOV34 metalloenzymes [5, 6].
Ubiquitin C-terminal hydrolase L1 (UCHL1), a member of the UCH class of DUBs, is one of the most well studied DUBs, and was identified in our prostate cancer protein profiling study [3, 5, 7]. Although, previous data demonstrate a putative role of UCHL1 in different tumor types, the exact oncogenic mechanism remains unclear. Deregulation of UCHL1 has been observed in solid tumors such as pancreatic cancer , non-small cell lung cancer , colorectal cancer , osteosarcoma , and oesophageal cancer . Furthermore, it has been reported that UCHL1 overexpression is associated with tumour progression, size and invasiveness . In gallbladder cancer UCHL1 is overexpressed due to hypomethylation of its promoter and the enhanced activity of the gene correlates with metastasis . In contrary, promoter hypermethylation leading to silencing of UCHL1 has been reported in progression of squamous cell carcinoma as well as gastric cancer and in pancreatic cancer cell lines [14–16]. Recent reports demonstrated that UCHL1 plays a key role in dissemination of non-small cell lung cancer  and an association of UCHL1 with β-catenin signaling pathway . Functional genomics studies revealed that siRNA mediated downregulation of UCHL1 regulates expression of several genes which are involved in multiple cellular processes such as apoptosis, cell proliferation and migration . Mutations in the UCHL1 gene have been shown to be associated with Parkinson's disease rather than cancer, for which differential expression appears to be more common. Expression profiling data from various tumour types reported that UCHL1 is either up- or downregulated due to promoter hypo- or hypermethylation depending on the type of malignant tissue. Li et al. showed that UCHL1 promotes tumour suppressor p53 signaling and is silenced due to its promoter methylation in nasopharyngeal carcinoma .
In our previous proteomic profiling study, we have identified a list of differentially expressed proteins in cancer containing several proteins that are known to be dysregulated in prostate cancer . Among them we identified UCHL1 as being downregulated in PCa compared to surrounding histological normal tissue or benign prostate epithelium. It has been reported that UCHL1 is deregulated in multiple types of tumours and the precise mechanism of the downregulation and function of UCHL1 in prostate cancer progression have not been investigated before. Therefore, the main objective of the present study was the functional characterization of UCHL1 in prostate cancer progression. The methylation status of UCHL1 promoter in tissue samples and the effect of altered UCHL1 expression on different cellular events were examined to determine the role of UCHL1 expression in PCa. We found that UCHL1 is downregulated in PCa due to promoter hypermethylation and demonstrated that UCHL1 has tumour suppressor activity in LNCaP cells.
Clinical samples, ethics statement and protein extraction
Prostate tissue samples were obtained from the University Medical Center Hamburg Eppendorf after informed consent. The study was approved by the local ethics committee of the University Hospital Eppendorf, Hamburg. For expression profiling whole prostates were collected after radical prostatectomy from patients with elevated PSA values and preoperative pathological examination performed at Martini Clinics, Hamburg, Germany. Patients received no preoperative therapy. After radical prostatectomy samples were frozen in liquid nitrogen until use. Tumor and benign areas were marked on the sections. We employed manual micro dissection method to obtain pathologically characterized materials for gene and protein expression profiling. The corresponding areas on the remaining blocks were sliced out with sharp knife, embedded in Tissue-tek® and stored at -80°C until use for total protein extraction. Protein preparation has been described previously .
RNA and DNA extraction from tissue sections
After surgical removal of the prostate, tissue samples were immediately taken with a 6 mm punch biopsy instrument (Stiefel, Wächtersburg, Germany) from areas that were suspected to contain tumor foci based on information obtained from the preoperative systematic biopsies. Tissue biopsy was immediately immersed in RNAlater (Qiagen, Hilden, Germany), stored overnight at ambient temperature and frozen at -20°C until use. For nucleic acid isolation, the specimen was thawed at room temperature and immediately washed two times each of 5 minutes in 10 ml ice-cold sterile PBS-buffer to elute most of the RNAlater from the tissue. Cryo sections were taken by fixing the tissue using Tissue-Tek® (Sakura, Netherlands) followed by freezing in a cryo-microtome and stained with haematoxylin and eosin (H&E) and analyzed by pathologists. Tissues were only enrolled into the study if at least 70% of cells were epithelial prostate tumor cells. Then, 10-15 subsequent unstained sections were transferred to a cryo tube for RNA and DNA isolation and the final section was again H&E stained and analyzed by pathologists. In parallel, normal prostate tissues were collected from tumor free areas and processed in the same way as tumor samples. Adjacent cryo sections were used for DNA and RNA extraction from the same prostate tissue specimens. Total RNA and DNA were extracted using the All Prep DNA/RNA Mini kit (Qiagen) according to the manufacturer's instructions. The quantity of the DNA and total RNA was checked using the Nanodrop and RNA quality by Bioanalyzer. Samples with low RNA quality (RIN < 6) were excluded from further analysis.
The PCa cell lines LNCaP and DU145 were purchased from DSMZ (Braunschweig, Germany) and cultivated in RPMI1640 (Invitrogen) supplemented with 10% fetal bovine serum, 100 units/mL) penicillin and streptomycin as recommended by suppliers Phoenix amphotrophic packaging cells were grown in DMEM with 10% fetal bovine serum (FBS) and penicillin/streptomycin. Cells were regularly tested for mycoplasma contamination using the MycoAlert Kit (Cambrex Bio Science Rockland, Inc., Rockland, ME, USA).
Bisulfite treatment and methylation specific PCR (MSP)
DNA isolation from prostate tissues was performed using Qiagen all prep kit according to supplier's protocol. The DNA concentration was measured by nanodrop spectrophotometer (Peqlab, Germany). Approximately 1 μg DNA was sodium bisulfite-modified and subjected to MSP with primers specifically recognizing the unmethylated or the methylated sequence of UCHL1. MSP primers for the UCHL1 gene were adapted from previous publication . Primers used for MSP are mentioned in additional files (Additional file 1). The PCR was run for 35 cycles with an annealing temperature of 56°C. Normal DNA from peripheral blood was treated in vitro with SssI methyltransferase (New England Biolabs, Beverly, MA) in order to generate in vitro methylated DNA (IVD) that served as a positive control for methylated alleles. PCR products were separated on 2.5% agarose gels and visualized by ethidium bromide staining.
For quantitative analysis of regional DNA methylation, pyrosequencing was used. Following PCR amplification of bisulfite-converted DNA using primer sequences (see Additional file 1) the final biotin-labeled PCR products were captured by Streptavidin Sepharose HP (Amersham Biosciences). PCR products bound on the beads were purified and made single-stranded in a Pyrosequencing Vacuum Prep Tool (Pyrosequencing Inc.). The forward sequencing primers were annealed to single-stranded PCR products and pyrosequencing was performed using the PSQ HS 96 Pyrosequencing system (Biotage AB). Quantification of cytosine methylation was performed using the PSQ HS96A 1.2 software package.
RNA isolation and quantitative real time PCR
Quantitative real time PCR for analysis of transcriptional levels of UCHL1 was performed in 48 benign and 45 tumour samples using SYBR Green. RNA isolation and cDNA synthesis carried out according to standard protocols. Quanti Tect primers for UCHL1, p53, MDM2 and p27Kip1 and GAPDH (housekeeping gene) were purchased directly from Qiagen, Germany. Quantitative real time PCR was performed in thermal cycler (Stratagene, Germany) using Dynamo Flash SYBR Green qPCR kit (Finnzymes, Finland) under optimized cycling conditions. PCRs for the target and housekeeping genes were performed in triplicates and mean relative expression levels were reported. To obtain statistical significance data obtained were analyzed by unpaired student t-test and p value < 0.05 was considered as significant. For semiquantitative UCHL1 RTPCR we used the cloning primers and for RPLP0 primers refer additional files (Additional file 1).
Cloning strategy for UCHL1 overexpression
An UCHL1 protein expressing recombinant vector was generated by cloning the coding region of the human UCHL1 (Accession number NM_004181) cDNA derived from prostate tissue into the pMSCV-Puro vector (Clontech, Palo Alto, CA, USA). UCHL1 coding sequence (CDS) was amplified from total cDNA by PCR using Phusion DNA Polymerase (Finnzymes Oy, Finland). After digestion of the PCR product and pMSCV vector with BglII/XhoI enzymes (Fermentas GmbH, Germany), ligation of the PCR product with linear vector resulted in recombinant pMSCV-UCHL1 construct. The sequence of the cloned PCR fragment was confirmed by DNA sequencing (MWG Operon).
Virus production and infection of target cells
Phoenix amphotrophic packaging cells were transfected with either empty or pMSCV-UCHL1 vector using CaCl2 transfection method. The transfection mixture was prepared by mixing 15 μg of plasmid DNA and 125 mM Cacl2 in 1 ml of HBS. The DNA precipitate was added drop wise into the cell culture medium containing 25 μm of chloroquine. After 12 h of transfection, medium was replaced with fresh medium and further incubated for 12 h. The virus containing medium was collected, filtered by 0.45 μm sterile filters directly on the target cells at around 50% of confluence. The cells were fed with fresh medium to continue another round of virus collection. Both target and packaging cells were continued to grow. 12 h later, two more infection cycles were repeated. After three cycles of infection, the target cells were grown in normal cultivation medium for 24 hrs and selected for integration of the target gene and/or puromycin at a concentration of 2 μg/ml until all cells died in control dishes. The colonies appeared with resistance to puromycin were propagated further and verified for overexpression of UCHL1 in LNCaP cells.
Cell proliferation assay
For cell proliferation assays, cells were plated at a density of 1.5 × 105 cells/well in 6-well format in complete growth medium. Cells were allowed to grow under optimal culture conditions over a period of 0 to 8 days. Cells were harvested by trypsinization and counted using Vi-Cell Cell counter (Beckaman Coulter GmbH, Germany). The growth rate was shown by plotting the mean total number of cells from triplicate experiments vs. growth time in days. Each experiment was performed in triplicate wells and repeated 3 times. The significance of difference in growth between UCHL1 positive and mock LNCaP cells was calculated using student t-test.
Detection for SA-ß-galactosidase was performed as described elsewhere . Briefly, LNCaP cells were harvested at sub confluent density and fixed with 2% PFA and 0.25% glutaraldehyde in PBS supplemented with 1 mM MgCl2 (pH 6.0) and incubated in a staining solution containing potassium cyanide/X-gal in PBS/MgCl2 (pH 6.0) at 37°C over night. Slides were analyzed using an Axioplan microscope (10 × magnifications) (Carl Zeiss AG, Germany). 200 cells/triplicate were analysed for a positive staining.
Colony formation assay
Effect of UCHL1 expression on anchorage-independent LNCaP cell growth was analyzed by soft agar assays. In soft agar assay, bottom agar was prepared by mixing 1% of agarose (Bacto Agar: Becton, Dickinson, Sparks, MD) with 2 × RPMI 1640 with 10% FBS in 6-well plates at 37°C to achieve final concentration of 0.5% of agar. After solidifying the bottom agar, 1 × 104 cells were mixed with cultivation medium and agar solution to obtain a final concentration of 0.35% agar. The mixture was spread on the surface of pre prepared base agar plates immediately. The culture medium was replenished every 3 days with fresh medium. After 14 days of incubation, plates were stained with 0.005% crystal violet solution until colonies turned purple color. After washing excess stain solution colonies were photographed and counted under a light microscope. Each experiment was performed in triplicates and repeated 3 times.
Proteasomal activity assay
Cells were lysed in M-PER buffer with complete protease inhibitor cocktail without EDTA (Roche, Germany). Protein concentration was measured by BCA method (Thermo Fischer Scientific, Germany) according to supplier's protocol. For proteasomal activity assay, 10 μg of total protein diluted to final volume of 50 μl with incubation buffer (5 mM DTT, 0.5 mMEDTA, 20 mM HEPES, 0.1 mg/ml ovalbumin in ddH2O, pH 7.8). Protein lysate prepared in incubation buffer were pre incubated for 2 h at 4°C. For blank control, incubation buffer alone was included in each assay. After incubation, the substrate suc-LLVY-AMC (Calbiochem) was added to the incubation mixtures to achieve final of concentration of 60 μM in final volume of 100 μl. The assay plate was incubated at 37°C for 1 h in the dark before measuring the proteasomal activity in fluorescent spectrophotometer (Mitras LB 940, Berchthold Technology, TN, US) at 355 and 460 nm. Each experiment was performed in triplicates and repeated 3 times.
Commercially available antibodies against protein targets of interest were purchased. Protein extracts prepared in M-PER (Pierce) with protease and phosphatase inhibitors were separated by 4-12% Bis-Tris-NuPAGE in Nupage running buffer and electrophoretically transferred onto PVDF membrane (Millipore). Blocking was carried out in 1 × Rotiblock solution (Roth Chemicals) followed by incubating the membrane with primary antibodies purchased all from cell signaling except anti UCHL1 from Millipore diluted at 1:1000 in 3% BSA in TBST overnight at 4°C. Excess antibodies were removed by washing with NaCl-Tris-Tween 20. Incubation with secondary antibody conjugated to horseradish peroxidase [anti-(mouse IgG) or anti-(rabbit IgG) from cell signaling, diluted 1:5000 in 1 × Rotiblock] was performed for 1 h at room temperature. After three washes, the reaction was developed by the addition of LumiGLO substrate (Thermo). The emitted light was captured on X-ray film (GE Healthcare).
Cell culture experiments were carried out in triplicates and repeated three times. Data points were expressed as mean of triplicates and median of repeated experiments. Graph Pad PRISM Version 5.0 statistics program was used to test significance of the results with Mann-Whitney test or t-test and p values less than 0.05 with 95% confidence interval were considered as significant.
UCHL1 expression and promoter DNA methylation status in PCa patients
Expression of UCHL1 in PCa cell lines
To assess the physiological effects of UCHL1 expression on prostate cancer cells, UCHL1 protein producing constructs were generated and the target gene was transferred into LNCaP cells by retroviral transduction. The expression of UCHL1 mRNA and protein in LNCaP cells were verified by RTPCR and western blotting (Figure 3) using an anti-UCHL1 antibody for western blotting. The UCHL1 transduced cells showed clear overexpression of UCHL1 mRNA and protein.
UCHL1 activity on LNCaP cell proliferation, cellular senescence and anchorage-independent growth
UCHL1 overexpression leads to K63 specific de ubiquitination in LNCaP cells
UCHL1 suppress anchorage- independent growth and cell proliferation via AKT phosphorylation and stabilizing p53 levels
Exogenous expression of UCHL1 results in accumulation of p27Kip1 and suppressed Cyclin A expression
To understand the impact of UCHL1 accumulation on cell cycle control of LNCaP cells, we examined the expression and phosphorylation of retinoblastoma proteins (Rb) predominantly regulated by p21 as well as Cyclin E and Cyclin D1 proteins associated with p27Kip1. Notably, we found no change in levels of either total or phosphorylated Rb in cells with a p27Kip1 accumulation. Interestingly, Cyclin D1 which promotes Cyclin A expression is significantly reduced on protein level due to high p27Kip1 activity in UCHL1 overexpressing LNCaP cells where as Cyclin E remains unchanged (Figure 7D). In line with Cyclin A the mRNA of Cyclin D was not significantly regulated in UCHL1 overexpressing cells. (Figure 7E).
These results together suggest that the accumulation of p27Kip1 as a result of UCHL1 overexpression might interfere with cell cycle progression and thus influence cell growth.
UCHL1 downregulation in PCa is associated with epigenetic modifications
In the present study, we analyzed the mechanism of UCHL1 downregulation in PCa and the role of UCHL1 as tumour suppressor in LNCaP prostate cancer cells. Our proteomic analysis of prostate tissue samples highlighted significant downregulation of UCHL1 in cancer compared to the surrounding benign tissue . In the current study, we analyzed whether downregulation of UCHL1 in PCa is due to promoter hypermethylation. Methylation specific PCR and pyrosequencing results demonstrated that UCHL1 suppression in primary prostate tumour tissues is linked to its promoter DNA methylation. This hypermethylation is correlated with low UCHL1 transcripts and protein levels in cancer compared to adjacent benign prostate epithelium.
UCHL1 has tumour suppressor role PCa via p53 accumulation
Our results have shown that UCHL1 inhibited LNCaP cell growth. Furthermore, UCHL1 significantly suppressed anchorage independent growth in soft agar. In cancer cells, anchorage independent growth and the metastatic potential have been linked to the AKT/PI3 kinase pathway , that the tumor suppressor p53 as an upstream regulator of AKT/PI3 kinase pathway directly suppress PTEN a tumour suppressor and negative regulator of AKT pathway [25–27]. In our results we found that there was a significant increase in p53 in UCHL1 positive cells, but no influence of UCHL1 on PTEN expression (data not shown). It is therefore intriguing to speculate that the observed suppression of tumorigenesis by UCHL1 is due an effect of p53 on activated Akt/PI3 kinase pathway regardless of PTEN expression. In p53 signaling pathway, the tumor suppressor p14ARF inhibits the E3 ligase MDM2 activity, which in turn inhibits p53 degradation. Furthermore p53 can act as a transcription factor attenuates MDM2 function by suppressing transcription of MDM2 [28, 29]. Our results also show that p14ARF is upregulated in UCHL1 positive cells possibly explaining the observed suppression of MDM expression.
Therefore we would speculate that UCHL1 controls p53 activity indirectly via multiple mechanisms. Together, these finding are in line with the results of Tokumaru et al. describing tumor suppressor function of UCHL1 in head and neck squamous cell carcinoma . The cyclin dependent kinase inhibitor p21 is the major transcriptional target of p53 and is required for p53 dependant cell cycle arrest [31–33]. p21 regulates cell cycle progression at the S phase by direct inhibition of cyclin E/CDK2 and cyclin D/CDK4 complexes activity. We could not demonstrate an influence of p53 on the expression of p21 in UCHL1 positive LNCaP cells. We therefore suppose that the accumulated p53 targets AKT phosphorylation and thereby cell survival, rather than p21 in LNCaP cells. Consistent with our observation that p53 and p14ARF are upregulated in UCHL1 overexpressing cells, we detected a p27Kip1 upregulation and a senescent phenotype by using the SA-ß-gal assay. The importance of p27kip1 stabilisation for induction of a senescent phenotype in murine prostatic intraepithelial neoplasia has been shown by Majumder et al . Furthermore, p53 seems to be involved in radiation-induced senescence . Therefore, we propose that UCHL1 might represent a regulator of senescence induction in prostate carcinogenesis.
UCHL1 exhibits influence on cell cycle regulator proteins via p27Kip1
UCHL1 may be involved in multiple cellular processes. In lung cancer cells, UCHL1 interacts with p27Kip1 . There is an inverse relationship between UCHL1 and p27Kip1 expression in many cell lines . p27Kip1 is a multifunctional protein involved in regulation of cell proliferation and apoptosis . In our hands we observed that UCHL1 overexpression positively regulates p27Kip1 levels in LNCaP cells. However, the precise role of UCHL1 in regulation of p27Kip1 is not well described so far. Earlier reports showed evidence that p27Kip1 is degraded through the ubiquitin-proteasome pathway resulting in decreased cellular concentration . Caballero et al. demonstrated that UCHL1 might be involved in nuclear localization of p27Kip1 mediated via JAB1 in lung cancer cells .
To get more insights about the role of UCHL1 via p27Kip1 accumulation, we investigated the mechanism by which proliferation and anchorage-independent growth is suppressed by the UCHL1 in LNCaP. To inhibit cell cycle progression, p27Kip1 needs to be located in the nucleus and interacts with cyclin dependant kinases (CDKs) and/or cyclins to block their activity. As a result, activation of Rb by phosphorylation is blocked and the Rb protein binds to E2F resulting in a positive regulation of cell cycle progression. Guandagno et al. have shown that the expression of cyclin A is critical for anchorage-dependent progression of the cell cycle in NIH3T3 and NRK normal rodent fibroblast cell lines and exogenous expression of cyclin A induced anchorage-independent growth in NRK cells . These findings are consistent with our result, indicating that the downregulation of Cyclin A is correlated with the suppression of anchorage-independent growth of LNCaP cells by UCHL1.
It has also been reported that p27Kip1 inhibits the expression of Cyclin A. p27Kip1-/- knockout in mice is associated with increased cell proliferation due to deficient Cyclin D inhibition . Consistent with these findings we could also demonstrate downregulation of Cyclin D1 in UCHL1 positive LNCaP cells. One of the best known substrates of cyclin D is the Rb tumor suppressor protein. Phosphorylation of Rb is performed by Cyclin D/Cyclin E and Rb remains phosphorylated throughout S, G2 and M phases during cell cycle progression. Zhu et al have shown that phosphorylation of the Rb protein is anchorage-dependent in NIH3T3 and human fibroblast cells. However, in the same report they have also shown convincingly Rb independent regulation of Cyclin A expression in NRK suspension cultures . In our findings we observed that both the expression of Rb or phosphorylation of Rb and Cyclin E expression were not affected by UCHL1 expression in LNCaP cells. Therefore we assume that downregulation of cyclin A in p27Kip1 accumulated LNCaP cells by UCHL1 overexpression is caused by an Rb-independent pathway.
In summary, we showed that UCHL1 is suppressed in prostate cancer patients by promoter hypermethylation. Restoration of UCHL1 in LNCap cells in which UCHL1 silenced by its promoter methylation could activate p53 signaling axis via reduced AKT phosphorylation and accumulation of p27 kip1 cell cycle inhibitor there by suppressing cellular growth (Figure 4). Thus, our study supplement substantially the current knowledge of the tumor suppressor functions of UCHL1 in cancer progression and postulate that UCHL1 hypermethylation could be a potential molecular marker for prostate cancer need to be evaluated in large number of patient cohort.
This work was supported by the German Federal Ministry for Education and Research in the framework of the Program for Medical Genome Research (Grants: 01GS0890, 01GS0892, 01GS0189).
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.