Biphasic regulation of A20 gene expression during human cytomegalovirus infection
© Gu et al.; licensee BioMed Central Ltd. 2014
Received: 25 April 2014
Accepted: 28 June 2014
Published: 8 July 2014
The A20 ubiquitin-editing enzyme is a target of nuclear factor kappa B (NF-κB) and also plays a key role in regulating the NF-κB signaling pathway. NF-κB activity is increased during human cytomegalovirus (HCMV) infection and HCMV appears to be adapted to this change. To better understand the regulation of NF-κB signaling during HCMV infection, we investigated how A20 expression is controlled during HCMV infection.
The expression level of A20 in human fibroblast cells infected with HCMV or UV-inactivated virus (UV-HCMV) was measured by immunoblot analysis, cell staining, and quantitative real-time PCR. Changes of histone modifications on the A20 promoter were determined by chromatin immunoprecipitation assays. Lentiviral vectors were used to knockdown A20 in fibroblast cells.
A20 expression was increased at early times after HCMV infection. This increase of the A20 protein level was promoted by viral gene expression under low viral load conditions. The viral IE1 protein, which is known to activate NF-κB, increased the A20 promoter activity through the upstream NF-κB sites in reporter assays, suggesting that IE1 is at least partly involved in A20 induction. Analysis of A20 expression with a high viral load demonstrated that the A20 regulation by HCMV was biphasic; both A20 protein and mRNA levels were increased at the early stage of infection, but decreased at the late stage. Under high viral load conditions, A20 upregulation was more profound with UV-HCMV than with HCMV, indicating a role of the viral gene product(s) in limiting A20 induction. Consistently, more histone modifications for euchromatin were found on the A20 promoter during UV-HCMV infection than with HCMV infection. A20 knockdown by shRNA reduced HCMV growth.
These results suggest that the biphasic regulation of A20 expression may be important for productive HCMV infection.
The nuclear factor-kappa B (NF-κB) signaling pathway controls inflammation, the immune response, and cell survival during viral infections. The regulation of NF-κB signaling during human cytomegalovirus (HCMV) infection is complicated. The pro-viral role of NF-κB for HCMV gene expression is suggested by the presence of the NF-κB sites on the viral major immediate-early (IE) promoter, and the IE1 protein autoregulates its own promoter through upstream NF-κB sequences . HCMV appears to activate NF-κB at different levels. For example, NF-κB is highly induced in HCMV-infected fibroblasts . Binding of viral glycoproteins to cell receptors activates NF-κB in fibroblasts and monocytes [3, 4]. In addition, NF-κB upregulation is mediated by viral IE proteins through induction of the cellular factor, Sp1, which activates the NF-κB p105/p50 and p65 subunit promoters [3, 5]. The role of IE1 in the induction of NF-κB is demonstrated in vascular smooth muscle cells, where IE1 selectively induces nuclear RelB and p50 among NF-κB/Rel factor , particularly through Jun kinase and c-Jun/Fra-2 AP-1 complexes for RelB induction .
Since NF-κB activity leads to an increase in antiviral inflammatory cytokine expression, an antiviral role for NF-κB signaling during viral infection is expected. When NF-κB activation is blocked, cytokines that inhibit HCMV are no longer produced in culture supernatants, and overexpression of IκB kinase β (IKKβ) inhibits HCMV replication . Thus, viruses appear to establish several strategies to interfere with innate immunity by preventing NF-κB pathway activation . IE2 inhibits tumor necrosis factor α-induced NF-κB DNA binding, leading to suppression of the IFN-β promoter .
A20, a ubiquitin-editing enzyme that belongs to the OTU domain family of deubiquitinases (DUBs), is induced by NF-κB and also acts as a negative regulator of NF-κB signaling, playing a critical role in terminating NF-κB responses to various stimuli . A20 DUB activity cleaves K63-linked poly-ubiquitin chains from RIP1  and TRAF6 . In addition, A20 harbors atypical zinc finger-dependent K48-specific E3 ubiquitin ligase activity that adds K48-linked poly-ubiquitin chains to RIP1, inducing its degradation . A20 also promotes proteasomal degradation of ubiquitin E2 conjugating enzymes, Ubc13 and UbcH5, involved in NF-κB signaling . By targeting IκB kinase γ (IKKγ), A20 also non-catalytically inhibits TNFR-mediated signaling . The importance of A20 in NF-κB regulation is shown in in vivo studies demonstrating that A20-deficiency in mice results in excess NF-κB activity and increased inflammation in several organs [13, 16]. Furthermore, A20 is reported to stabilize NF-κB-inducing kinase (NIK), promoting the transition from canonical to non-canonical NF-κB activation for proper immune system control .
It has been shown that A20 is induced in human monocytes after HCMV infection and that the binding of viral glycoproteins to cellular receptors plays a role in this induction . In the present study, we investigated how A20 expression is regulated in human fibroblasts after HCMV infection. We showed that A20 expression is biphasically regulated during HCMV infection. A20 expression was initially increased at early times of infection but inhibited at later stages. We also found that the viral IE1 protein activates the A20 promoter through the upstream NF-κB biding sites, suggesting a role of IE1 in the initial A20 activation. Analysis of A20 expression with a high viral load demonstrated that the newly synthesized viral gene product(s) epigenetically limit the upregulation of A20 transcription. We also found that A20 knockdown reduces HCMV growth, suggesting that the biphasic regulation of A20 expression may be important for productive HCMV infection.
Effect of HCMV infection on A20 expression
Evaluation of the IE1-mediated NF-κB activation on A20 promoter activity
Biphasic regulation of A20 expression by HCMV at a high viral load
Evidence for epigenetic regulation of the A20 promoter during HCMV infection
To further investigate whether A20 transcription is epigenetically regulated, we examined changes in histone modification in the A20 promoter region after infection with HCMV and UV-HCMV. The results of ChIP assays performed at 24 and 48 h showed that the level of histone H4 acetylation, a marker for euchromatin, was slightly reduced by HCMV compared to mock-infection, whereas histone H4 acetylation was increased by UV-HCMV (Figure 3C). In addition, we measured the level of histone H3 tri-methylation, a marker for heterochromatin, and both HCMV and UV-HCMV showed increased levels as the infection progressed, but HCMV showed a more profound increase than UV-HCMV at 48 h (Figure 3D). These ChIP assay results indicate that HCMV infection led to more heterochromatin-favored histone modifications than did UV-HCMV at this high MOI, implicating a role for the viral gene product(s) in these changes. Taken together, the results of the qRT-PCR assays and the ChIP assays in virus-infected cells suggest that the expression of A20 is epigenetically controlled to limit its upregulation by the viral gene product(s) during HCMV infection.
Role of A20 regulation in HCMV growth
In this study, we evaluated the impact of HCMV infection on A20 expression, which plays a key role in regulating NF-κB signaling. We found that the A20 expression is increased upon HCMV infection. This is consistent with a previous report showing that the binding of HCMV glycoproteins to cellular receptors induces A20 transcription in human monocytes . We also found that IE1 activates the A20 promoter. The direct activation of A20 promoter by IE1 is not surprising, since IE1 induces NF-κB activity and the A20 promoter contains two NF-κB binding sites. Thus, IE1 can exert opposite effects on NF-κB activity; it can increase NF-κB activity by directly activating the NF-κB subunit production [3, 5–7] and also reduces NF-κB activity by indirectly inducing A20 expression. We found that A20 was moderately increased by IE1 expression alone and localized largely in the cytoplasm, whereas it was increased to a higher level by HCMV infection and accumulated at the cytoplasm and to a lesser extent the nucleus. Therefore, it is likely that other viral genes are also implicated in A20 upregulation and intracellular localization. Similar opposite effects on NF-κB activity have been observed during Kaposi’s sarcoma-associated herpesvirus (KSHV) infection . The KSHV vFLIP protein promotes NF-κB activation by directly binding to NEMO and deregulates it by indirectly inducing A20 expression. A20 induction by vFLIP is thought to inhibit vFLIP-induced NF-κB activation. Therefore, both HCMV and KSHV appear to have a bi-directional regulatory mechanism to control NF-κB activity.
Notably, we found that the A20 regulation during HCMV is biphasic. The initial A20 upregulation was observed at the early phases of infection with both low and high MOIs. This biphasic regulation of A20 was more apparent with high MOIs, where the A20 expression was initially upregulated at the early phases of infection but downregulated during the late phases and these differences were controlled at the level of transcription. Importantly, the upregulation of A20 was much higher and sustained in UV-HCMV infection than in HCMV infection. Therefore, it appears that the viral gene products may be involved in limiting A20 upregulation. Intriguingly, we found that the effects of the newly synthesized viral gene product(s) on A20 regulation were different with different viral loads; viral gene expression promoted A20 upregulation with a low MOI, whereas it contributed to A20 downregulation with a high MOI. Further studies are necessary to investigate how A20 expression is differently regulated by different viral loads on a single cell level.
The exact role of A20 regulation during HCMV infection is not clear. Given the critical role of A20 in terminating canonical NF-κB signaling, HCMV may regulate A20 expression for elaborate modulation of the NF-κB activity during infection in a timely manner. The initial increase of A20 level after HCMV infection may just reflect the increase of NF-κB activity upon virus infection and this negative feedback regulation is to control excessive NF-κB activity, which may be harmful by provoking a robust immune response in vivo. This regulation appears to be necessary for the efficient progress of viral infection, since A20 depletion reduced viral growth by affecting the accumulation of IE2 and subsequent viral early and late proteins. Notably, the effect of A20 depletion on HCMV growth was more profound at a lower MOI, suggesting that reducing the NF-κB activity by A20 induction may be more critical for viral replication under low MOI conditions. Besides acting as a terminator in canonical NF-κB signaling, A20 also activates the non-canonical NF-κB pathway through NIK stabilization . Therefore, it is complicated to expect the output NF-κB activity after A20 induction during HCMV infection. Further studies on the exact role of A20 during HCMV infection are warranted.
In this study, we show that the A20 expression is biphasically regulated during HCMV infection. The A20 level was increased during the early phases of HCMV infection, but this induction was reduced at the late phases. Although A20 has been shown to be induced during HCMV entry, we also found that IE1 activates the A20 promoter through the upstream NF-κB sites. We also suggest that the downregulation of A20 induction at the late stages of infection may involve epigenetic controls of A20 transcription and that the newly produced viral gene product(s) may facilitate this process. Furthermore, we show that the A20 expression is required for efficient HCMV growth, highlighting the importance of the fine regulation of A20 gene expression in the virus replication cycle. This A20 regulation may affect NF-κB activity during HCMV infection or be critical in regulating unknown A20 targets or signaling that may be involved in HCMV growth.
Materials and methods
Cells and viruses
Human foreskin fibroblast ( HF) and 293 T cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum in a 5% CO2-humidified incubator at 37°C. The cell growth medium also contained 100 units/ml of penicillin and 100 μg/ml of streptomycin. The stock for the Towne virus and the production of UV-inactivated virus (UV-HCMV) were previously described . Infectious center assays for the measurement of viral titers were also previously described .
pSG5-based expression plasmids for HA-IE1 and HA-IE2 were described previously [20, 21]. The A20 cDNA was cloned into a pENTR vector (Invitrogen). The retroviral vectors expressing A20 or IE1 were produced by transferring the A20 or IE1 cDNA from ENTR vectors to the pMIN-based destination vector using the LR Clonase (Invitrogen). Dr. Kun-Sang Chang (The University of Texas M. D. Anderson Cancer Center, Houston, USA) provided an A20(-95/+9)-luciferase reporter construct containing the A20 minimum promoter with two NF-κB sites [22, 23]. In addition, Dr. Rivka Dikstein (The Weizmann Institute of Science, Rehovot, Israel) provided a set of pA20(-229/+12)-luciferase reporter plasmids containing wild-type or mutant NF-κB (mNF-κBs) sites. In these plasmids, the wild-type A20 promoter contained six Sp1 sites and two NF-κB sites, whereas the mNF-κBs A20 promoter contained mutant NF-κB sites . pNFκB-Luc, a luciferase reporter plasmid that contains a synthetic promoter (herpes simplex virus thymidine kinase TATA-like promoter) with four copies of NF-kB sites from the MHC class II gene, was purchased from Clontech.
Lentiviral and retroviral vectors
Lentiviral vector pLKO.1-TRC control expressing a non-hairpin control RNA (shC) was purchased from Addgene. Lentiviral vectors (on a pLKO.1 background) expressing shRNAs against A20 (shA20) (TRCN50959, 50961, and 218517) were purchased from Open Biosystems. Lentiviral stocks were prepared by cotransfecting 293 T cells with the lentiviral vectors together with packaging plasmids pCMV-DR8.91 expressing the gag-pol, tat, and rev proteins of human immunodeficiency virus (HIV) and pMD-G expressing the envelope G protein of vesicular stomatitis virus (VSV) . For retroviral stocks, 293 T cells were cotransfected with the retroviral vectors (MIN, or MIN-IE1) together with packaging plasmids pHIT60 (Gag-Pol) and pMD-G (VSV-G)  using Omicsfect reagents (Omicsbio). Viral supernatants were harvested 48 h after transfection. HF cells were transduced by viral vectors in the presence of polybrene (7.5 μg per ml), and then were selected with puromycin (1 μg per ml) for lentiviral evctors or with G418 (0.5 mg/ml; Calbiochem) for retroviral vectors. For efficient knockdown A20, three lentiviral vectors expressing three different shRNAs were used together to transduce HF cells.
HF cells were transfected via electroporation using the Neon Transfection System (Invitrogen). For each reaction, 3 × 105 cells were suspended in 100 μl of resuspension (R) buffer and mixed with expression plasmids (up to 4 μg) in a 1.5-ml tube. After electroporation at 1,300 V and 40 ms, the cells were plated in six-well plates or chamber slides. 293 T cells were transfected using Omicsfect reagents (Omicsbio) according to the manufacturer’s instructions.
Immunoblot analysis and antibodies
The standard procedures for immunoblot analysis were described previously . The mouse monoclonal antibody (MAb) for A20 was purchased from Imgenex. Mouse MAbs for IE1/IE2, p52 (UL44), and pp28 (UL99), rat anti-HA MAb (3 F10) conjugated to peroxidase, and mouse MAb against β-actin were described previously . Rabbit polyclonal Ab (PAb) raised against purified His-IE1 were described previously [20, 21]. Mouse MAb for IκBα was purchased from AbFrontier. Secondary Abs for immunoblot analysis were obtained from Jacson ImmunoResearch Laboratories, Inc.
Indirect immunofluorescence assay (IFA)
Cells were fixed for 5 min in 1% paraformaldehyde and permeabilized in cold 0.2% Triton X-100, or fixed for 5 min in cold and rehydrated in cold PBS and then incubated with appropriate Abs in PBS for 1 h at 37°C and then incubated with fluorescein isothiocyanate (FITC)-labeled or rhodamine/red X-coupled donkey immunoglobulin G (IgG) (Jackson ImmunoResearch Laboratories, Inc.). For double-labeling, two different Abs were incubated together. The slides were examined and photographed with a Carl Zeiss Axiophot microscope.
RNA isolation and quantitative real-time reverse-transcription PCR (RT-PCR)
Total RNA was isolated from 2 × 105 cells using TRIzol reagent (Invitrogen) and a MaXtract High Density Tube (Qiagen). cDNAs were synthesized using the random hexamer primers in the SuperScript III system (Invitrogen). Quantitative real-time PCR was performed using the SYBR green PCR core reagents (Applied Biosystems) and ABI Prism SDS software. The primers used to amplify A20 were 5’-CACACAAGGCACTTGGATCC-3’ (forward) and 5’-TCCCCAGGAGTCCGTGCAGC-3’ (reverse).
Chromatin immunoprecipitation (ChIP) assays
ChIP assays were carried out using a kit (Upstate Biotechnology, Inc.). HF cells (6 × 106) mock-infected or infected with HCMV were fixed with 1% formaldehyde for 10 min at 24 h after infection and then lysed with the lysis buffer provided in the kit. ChIP assays were performed with 5 μg of anti-acetyl-histone H4 Ab (Millipore), anti-trimethyl-histone H3 (K9) Ab (Millipore), or control IgG. One-sixth of the lysates were reserved for quantitation of the DNA present in different samples prior to immunoprecipitation. Relative changes in precipitated DNA were calculated via PCR and agarose gel electrophoresis. The primer sequences used for amplification of the A20 promoter region were 5’-CAGCCCGACCCAGAGAGTCAC-3’ (forward) and 5’-CTCCGGGCCCCGCGATCC-3’ (reverse). The PCR program used involved 30 amplification cycles described as follows: 95°C for 30 s, 62°C for 30 s, and 72°C for 45 s.
Luciferase reporter assay
Cells were lysed using three freeze-thaw steps in 100 μl of 0.25 M Tris–HCl (pH 7.9) containing 1 mM dithiothreitol. Subsequent procedures were performed as previously described . A TD-20/20 luminometer (Turner Designs) was used for the 10-s assay of the photons produced.
We thank Kun-Sang Chang and Rivka Dikstein for providing us with their reporter plasmids. This research was supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2012R1A2A2A01002551).
- Cherrington JM, Mocarski ES: Human cytomegalovirus ie1 transactivates the alpha promoter-enhancer via an 18-base-pair repeat element. J Virol 1989, 63: 1435-1440.PubMedPubMed CentralGoogle Scholar
- Kowalik TF, Wing B, Haskill JS, Azizkhan JC, Baldwin AS Jr, Huang ES: Multiple mechanisms are implicated in the regulation of NF-kappa B activity during human cytomegalovirus infection. Proc Natl Acad Sci U S A 1993, 90: 1107-1111.PubMedPubMed CentralView ArticleGoogle Scholar
- Yurochko AD, Kowalik TF, Huong SM, Huang ES: Human cytomegalovirus upregulates NF-kappa B activity by transactivating the NF-kappa B p105/p50 and p65 promoters. J Virol 1995, 69: 5391-5400.PubMedPubMed CentralGoogle Scholar
- Yurochko AD, Huang ES: Human cytomegalovirus binding to human monocytes induces immunoregulatory gene expression. J Immunol 1999, 162: 4806-4816.PubMedGoogle Scholar
- Yurochko AD, Mayo MW, Poma EE, Baldwin AS Jr, Huang ES: Induction of the transcription factor Sp1 during human cytomegalovirus infection mediates upregulation of the p65 and p105/p50 NF-kappaB promoters. J Virol 1997, 71: 4638-4648.PubMedPubMed CentralGoogle Scholar
- Jiang HY, Petrovas C, Sonenshein GE: RelB-p50 NF-kappa B complexes are selectively induced by cytomegalovirus immediate-early protein 1: differential regulation of Bcl-x(L) promoter activity by NF-kappa B family members. J Virol 2002, 76: 5737-5747.PubMedPubMed CentralView ArticleGoogle Scholar
- Wang X, Sonenshein GE: Induction of the RelB NF-kappaB subunit by the cytomegalovirus IE1 protein is mediated via Jun kinase and c-Jun/Fra-2 AP-1 complexes. J Virol 2005, 79: 95-105.PubMedPubMed CentralView ArticleGoogle Scholar
- Eickhoff JE, Cotten M: NF-kappaB activation can mediate inhibition of human cytomegalovirus replication. J Gen Virol 2005, 86: 285-295.PubMedView ArticleGoogle Scholar
- Le Negrate G: Viral interference with innate immunity by preventing NF-kappaB activity. Cell Microbiol 2012, 14: 168-181.PubMedView ArticleGoogle Scholar
- Taylor RT, Bresnahan WA: Human cytomegalovirus IE86 attenuates virus- and tumor necrosis factor alpha-induced NFkappaB-dependent gene expression. J Virol 2006, 80: 10763-10771.PubMedPubMed CentralView ArticleGoogle Scholar
- Ma A, Malynn BA: A20: linking a complex regulator of ubiquitylation to immunity and human disease. Nat Rev Immunol 2012, 12: 774-785.PubMedPubMed CentralView ArticleGoogle Scholar
- Wertz IE, O'Rourke KM, Zhou H, Eby M, Aravind L, Seshagiri S, Wu P, Wiesmann C, Baker R, Boone DL, Ma A, Koonin EV, Dixit VM: De-ubiquitination and ubiquitin ligase domains of A20 downregulate NF-kappaB signalling. Nature 2004, 430: 694-699.PubMedView ArticleGoogle Scholar
- Boone DL, Turer EE, Lee EG, Ahmad RC, Wheeler MT, Tsui C, Hurley P, Chien M, Chai S, Hitotsumatsu O, McNally E, Pickart C, Ma A: The ubiquitin-modifying enzyme A20 is required for termination of Toll-like receptor responses. Nat Immunol 2004, 5: 1052-1060.PubMedView ArticleGoogle Scholar
- Shembade N, Ma A, Harhaj EW: Inhibition of NF-kappaB signaling by A20 through disruption of ubiquitin enzyme complexes. Science 2010, 327: 1135-1139.PubMedPubMed CentralView ArticleGoogle Scholar
- Skaug B, Chen J, Du F, He J, Ma A, Chen ZJ: Direct, noncatalytic mechanism of IKK inhibition by A20. Mol Cell 2011, 44: 559-571.PubMedPubMed CentralView ArticleGoogle Scholar
- Lee EG, Boone DL, Chai S, Libby SL, Chien M, Lodolce JP, Ma A: Failure to regulate TNF-induced NF-kappaB and cell death responses in A20-deficient mice. Science 2000, 289: 2350-2354.PubMedPubMed CentralView ArticleGoogle Scholar
- Yamaguchi N, Oyama M, Kozuka-Hata H, Inoue J: Involvement of A20 in the molecular switch that activates the non-canonical NF-small ka. CyrillicB pathway. Sci Rep 2013, 3: 2568.PubMedPubMed CentralView ArticleGoogle Scholar
- Sakakibara S, Espigol-Frigole G, Gasperini P, Uldrick TS, Yarchoan R, Tosato G: A20/TNFAIP3 inhibits NF-kappaB activation induced by the Kaposi's sarcoma-associated herpesvirus vFLIP oncoprotein. Oncogene 2013, 32: 1223-1232.PubMedPubMed CentralView ArticleGoogle Scholar
- Kim YE, Lee JH, Kim ET, Shin HJ, Gu SY, Seol HS, Ling DP, Lee CH, Ahn JH: Human cytomegalovirus infection causes degradation of Sp100 proteins that suppress viral gene expression. J Virol 2011, 85: 11928-11937.PubMedPubMed CentralView ArticleGoogle Scholar
- Huh YH, Kim YE, Kim ET, Park JJ, Song MJ, Zhu H, Hayward GS, Ahn JH: Binding STAT2 by the acidic domain of human cytomegalovirus IE1 promotes viral growth and is negatively regulated by SUMO. J Virol 2008, 82: 10444-10454.PubMedPubMed CentralView ArticleGoogle Scholar
- Kim ET, Kim YE, Huh YH, Ahn JH: Role of noncovalent SUMO binding by the human cytomegalovirus IE2 transactivator in lytic growth. J Virol 2010, 84: 8111-8123.PubMedPubMed CentralView ArticleGoogle Scholar
- Wu WS, Xu ZX, Chang KS: The promyelocytic leukemia protein represses A20-mediated transcription. J Biol Chem 2002, 277: 31734-31739.PubMedView ArticleGoogle Scholar
- Krikos A, Laherty CD, Dixit VM: Transcriptional activation of the tumor necrosis factor alpha-inducible zinc finger protein, A20, is mediated by kappa B elements. J Biol Chem 1992, 267: 17971-17976.PubMedGoogle Scholar
- Ainbinder E, Amir-Zilberstein L, Yamaguchi Y, Handa H, Dikstein R: Elongation inhibition by DRB sensitivity-inducing factor is regulated by the A20 promoter via a novel negative element and NF-kappaB. Mol Cell Biol 2004, 24: 2444-2454.PubMedPubMed CentralView ArticleGoogle Scholar
- Pham TH, Kwon KM, Kim YE, Kim KK, Ahn JH: DNA sensing-independent inhibition of herpes simplex virus 1 replication by DAI/ZBP1. J Virol 2013, 87: 3076-3086.PubMedPubMed CentralView ArticleGoogle Scholar
- Tavalai N, Papior P, Rechter S, Leis M, Stamminger T: Evidence for a role of the cellular ND10 protein PML in mediating intrinsic immunity against human cytomegalovirus infections. J Virol 2006, 80: 8006-8018.PubMedPubMed CentralView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.