Constant up-regulation of BiP/GRP78 expression prevents virus-induced apoptosis in BHK-21 cells with Japanese encephalitis virus persistent infection
© Lyoo et al.; licensee BioMed Central. 2015
Received: 5 November 2014
Accepted: 18 February 2015
Published: 26 February 2015
Persistent infection of the Japanese Encephalitis Virus (JEV) has been reported in clinical cases, experimental animals, and various cell culture systems. We previously reported the establishment of spontaneous JEV persistent infection, assisted by defective interfering particle accumulation and/or attenuated helper viruses, in BHK-21 cells devoid of virus-induced apoptosis, cBS6-2 and cBS6-3. However, cell-specific factors may play important roles in controlling JEV replication and have never been assessed for this specific phenomenon. Recent evidence suggests that viruses have evolved various mechanisms to cope with endoplasmic reticulum stress signaling pathways for their efficient amplification and transmission, including the unfolded protein response (UPR).
To identify the host cell factors that affect JEV persistence, we investigated the expression of essential UPR factors in cBS6-2 and cBS6-3 cells. Of the selected UPR factors tested, the most noticeable deviations from those of the normal BHK-21 cells with JEV acute infection were as follows: the suppression of C/EBP homologous binding protein (CHOP) and the constant up-regulation of immunoglobulin binding protein (BiP) expression in cBS6-2 and cBS6-3 cells. In JEV acute infection on normal BHK-21 cells, silencing CHOP expression through specific siRNA blocked cell death almost completely. Meanwhile, depletion of BiP by specific siRNA unlocked CHOP expression in cBS6-2 and cBS6-3 cells, resulting in massive cell death. Fulminant apoptotic cell death for both cell clones on tunicamycin treatment revealed that the JEV persistently infected cells still contained functional arms for cell fate decisions.
BHK-21 cells with JEV persistent infection strive against virus-induced apoptosis through constant up-regulation of BiP expression, resulting in the complete depletion of CHOP even with apparent virus amplification in the cells. Accordingly, the attenuation of virus replication as well as the modifications to cell metabolism could be additional factors contributing to the development of JEV persistent infection in mammalian cells.
Viruses have evolved a wide range of strategies to persist in their hosts. It remains a challenge to understand the mechanisms whereby viral persistence is established and maintained, especially viral persistence within a cell or group of cells. Mechanisms by which RNA virus persistence is initiated and maintained usually involve two virus-specific factors: the generation of defective interfering (DI) particles or temperature-sensitive mutation of wild-type virus [1,2]. Research suggests that host factors involved in the control of persistent infection relate to elements of innate immunity in Morbillivirus  and cellular protein synthesis in Reovirus .
Protein synthesis and folding occurs in the endoplasmic reticulum (ER). Mammalian cells have evolved many sophisticated signaling pathways to monitor any abnormality, including the accumulation of misfolded proteins; these pathways are known as the unfolded protein response (UPR) . These signaling pathways monitor the ER’s capacity to refold and/or remove abnormally folded proteins and to make cell-fate decisions according to the homeostatic balance [6,7]. In all known animal cells, the following are known to be activated to initiate the UPR: three ER-localized transmembrane UPR transducers, inositol requiring kinase 1 (IRE1), double-stranded RNA-activated protein kinase-like kinase (PERK), and activating transcription factor 6 (ATF6) . Under basal conditions, these three sensors are associated with immunoglobulin binding protein (BiP), also known as GRP78, which is a chaperone of the heat shock protein 70 family. Each branch operates parallel with a particular target downstream and contributes to both cell-protective and cell-death pathways [6,7]. Under severe or chronic ER stress, the UPR switches its mode of action toward apoptosis. C/EBP homologous binding protein (CHOP), also known as growth arrest and DNA damage-inducible protein 153 (GADD153), is the pro-apoptotic transcription factor that plays an important role in regulating cell death after ER stress [9,10]. Several molecular mechanisms of CHOP-induced apoptosis have been cited, such as compromised alteration of Bcl-2 family proteins [11,12].
A variety of viruses induce ER stress and the UPR, having evolved various mechanisms to cope with the UPR . West Nile virus modulates all three arms of the UPR and induces numerous apoptotic responses, including induction of CHOP expression . Modulation of the UPR by the West Nile virus is regulated differentially along with its replication cycle . Similar to other flaviviruses, the dengue virus also induces the three arms of the UPR and CHOP expression. However, activated CHOP does not induce its downstream apoptotic markers, such as suppression of anti-apoptotic protein Bcl-2 and activation of caspase-3 or caspase-9 [16,17]. In addition, studies of the hepatitis C virus have shown that both viral structural (envelope) and non-structural (NS2) proteins can induce ER stress and the UPR activation with up-regulation of BiP and CHOP [18,19].
Japanese encephalitis virus (JEV), a member of the Flaviviridae, is the causative agent of encephalitis in humans and can be transmitted by persistently JEV-infected mosquitoes . Viral persistence in the nervous systems of JEV-infected patients has been shown in approximately 5% of JEV cases, suggesting that JEV persistence may contribute to neural sequelae after acute infection . Though JEV is usually cytolytic for susceptible cells, persistent infection of JEV has been established in various cell cultures, including baby hamster kidney (BHK)-21 [22-28] as well as in a mouse model . The underlying mechanisms for JEV persistence in cultured cells are not clearly described. We have previously demonstrated spontaneous establishment of persistent JEV infection in BHK-21 cells via serial undiluted passages without any supplemental treatment . Examples of supplemental treatment include Bcl-2 overexpression  or indirect infection with supernatants from persistently JEV-infected mosquito cells . Our previous study suggested that DI particle accumulation and helper virus attenuation are possible mechanisms for the development and maintenance of JEV persistence in BHK-21 cells. Nonetheless, labile cellular factors that may play a role in JEV persistence have never been identified in this system. JEV also induces ER stress and the UPR, and studies suggest that the activation of the UPR is a major cause of JEV-induced apoptosis . The UPR has never been assessed in JEV persistent infection, however.
In this study, we utilized two persistently JEV-infected BHK-21 cell clones (previously reported in ) to identify the cell-specific factors of the UPR involved in JEV persistence in mammalian cells. These cell clones seldom, if ever, undergo apoptosis, while JEV replicates actively within. We observed that there was no CHOP expression at any time, but a significant amount of BiP expression was constant in these cells. Knockdown of BiP expression resulted in CHOP induction and subsequent cell death. Because the level of JEV amplification in these cells was not low enough to hold an apoptotic process, we suggest that the readjustment of BiP expression in host cells could be one of key factors involved in cell fate decision under viral persistence.
Results and discussion
BHK-21 cells with JEV persistent infection avoided virus-induced apoptosis even with active virus replication
Constant expression of elevated amounts of BiP and complete depletion of CHOP were associated with the survival of the persistently infected BHK-21 cells
Phosphorylation of eIF2α often leads to inhibition of protein translation in general, but the translation of ATF4 is promoted by p-eIF2α . This phenomenon was reconfirmed in this experiment as the expression of ATF4 increased gradually 6 hr p.i. in JEV acutely infected cells (Figure 2A). CHOP, the key mediator of ER stress-induced apoptosis, was also induced and accumulated gradually along with the ATF4 activation. In contrast, for the JEV persistently infected cell clones, the expression of CHOP was not detected at all, even in the presence of ATF4 for its transcription (Figure 2A). Unlike cBS6-2, impaired expression of ATF4 in cBS6-3 was repeatedly noticed in several independent experiments. It was noteworthy that a much higher level of BiP was maintained in both JEV persistently infected cells throughout the culture period.
To further investigate whether the JEV persistently infected cell clones kept the UPR pathway intact, naïve BHK-21 cells and the two cell clones with JEV persistent infection were treated with 0.5 μg tunicamycin mL−1 for 24 hr. Compared to the DMSO-treated control, both tunicamycin-treated normal BHK-21 cells and cBS6-2 cells showed PERK-eIF2α-ATF4 pathway activation followed by CHOP induction (Figure 2B). Most of the cells treated with tunicamycin, including cBS6-2 and cBS6-3 cells, succumbed to apoptotic cell death within 24 hr (data not shown). Although CHOP was also clearly induced, expression of p-PERK or ATF4 in cBS6-3 cells was not comparable to the normal BHK-21 cells or cBS6-2 cells on tunicamycin treatment for unknown reasons. This observation suggests that cBS6-3 cell clones may utilize another pathway to induce CHOP expression, perhaps involving IRE1 activation. These differences in the activation process of the BiP-PERK arm between cBS6-2 and cBS6-3 imply that individual cells comprising a cell batch with JEV persistent infection could have their own unique modification in cellular physiology to avoid the virus-induced apoptosis.
Taken together, these results suggest that the JEV persistently infected cells avoid fulminant apoptosis by maintaining a constant, highly-elevated level of BiP, which results in the complete suppression of CHOP induction.
Silencing CHOP expression blocked JEV-induced apoptosis without the serious intervention of viral replication
The level of JEV NS3 protein in the cells transfected with siCHOP and then infected with JEV was slightly lower than that of the other cells transfected with scramble siRNA (Figure 3A). This observation aligns with the result that the virus titer obtained from the CHOP-silenced cells was slightly compromised compared to those from the control cells at 48 hr p.i. (Figure 3C). These results are also consistent with a report noting that up-regulation of CHOP during JEV infection plays a key role in virus-induced apoptosis . As infectious bronchitis virus-induced apoptosis was suppressed and virus replication inhibited in the CHOP-knockdown cells , the complete blockage of CHOP expression in cBS6-2 or cBS6-3 cells might intervene in JEV replication and therefore assist in the development of JEV persistent infection. However, a clear explanation for this observation is presently beyond the scope of this study.
Silencing BiP expression in JEV persistently infected cells led to CHOP induction, followed by severe reduction in cell viability
In this study of JEV persistently infected cells compared to an acute infection in normal BHK-21 cells, the most distinguishable aspects regarding the UPR factors were the complete depletion of CHOP and the constant up-regulation of BiP expression (Figure 2A and Additional file 1: Figure S1). Some previous studies also showed that BiP overexpression attenuates ER stress signaling and is protective against apoptosis . Therefore, we decided to assess the implications of the constant expression of BiP for maintaining cell viability against virus-induced apoptosis in the JEV persistently infected cells.
In conclusion, BHK-21 cells with JEV persistent infection strive against virus-induced apoptosis through constant up-regulation of BiP expression, a key chaperone involved in ER stress. As demonstrated in tunicamycin treatment, these cells maintained their capacity to decide their own cell death fate by inducing CHOP, although some of the UPR factors relaying the BiP-PERK-ATF4 arm appeared to be impaired.
In a previous study, we successfully established JEV persistent infection in several mammalian cells in the presence of DI particle generation . Certain modifications in the genetic makeup of helper JEV could also play a role in the development and maintenance of the viral persistent infection . Therefore, the observations made in this experiment demonstrate that both the compromised virus replication capacity and certain cellular factors, such as the UPR factors, also participate in establishing viral persistent infection in mammalian cells.
Our study highlights the importance of certain host cell factors in ER stress-signaling pathways for JEV persistence by utilizing an in vitro model for the first time. This work provides new insight into the complex mechanism of viral persistence and potentially contributes to developing useful agents and tools for therapeutic intervention in the clinical sequelae of Japanese encephalitis.
Materials and methods
Cells and virus
Baby hamster kidney-21 (BHK-21; Korea Cell Line Bank) cells were maintained in a minimum essential medium (MEM; Gibco) containing 5% FBS (Gibco) and 100 units of penicillin-streptomycin (Gibco) mL−1. The persistently JEV-infected BHK-21 cells have been described previously . Two cell clones, cBS6-2 and cBS6-3, were chosen for use. All cells were grown at 37°C in a 5% CO2 incubator. JEV K94P05 strain (provided by the Korean National Institute of Health) was employed throughout this study. Propagation of the virus was carried out in BHK-21 at 37°C in MEM supplemented with 5% FBS for 72 hr. After infection, the virus-containing supernatant was collected and centrifuged to remove cell debris, and then stored at −72°C.
The virus was inoculated on a monolayer of BHK-21 cells in 35 mm plates; an overlay medium was applied containing 5% FBS, 1% penicillin-streptomycin (10,000 U), 15% 5× MEM, 61% 1× MEM, and 1% agarose. After 3 to 4 days incubation, the cells were fixed with 3.7% formaldehyde in phosphate buffered saline (PBS) for 2 hr and stained with 0.1% crystal violet. To titrate the intracellular virus particles, the infected cells were washed with PBS, trypsinized, and resuspended in 1 ml of MEM. After three times of freeze and thaw cycles, the cell debris was pelleted before collecting the virus-containing supernatant. The virus titer was determined by a plaque-forming assay.
To analyze apoptosis, an annexin V-fluorescein isothiocyanate (FITC) and PI double-staining method (MEBCYTO Apoptosis kit; MBL) was used according to the manufacturer’s protocol. After the adherent cells were harvested, they were re-suspended in binding buffer, and 5 μL annexin V-FITC and 1.5 μL PI were added to the cell samples. The mixtures were incubated for 15 min in the dark at room temperature and then analyzed by flow cytometry (BD FACSCalibur).
Antibodies and reagent
Tunicamycin (Sigma-Aldrich) was dissolved in DMSO. Rabbit anti-NS3 antibody was kindly provided by Professor Radhakrishnan Padmanabhan (Georgetown University, USA). Antibodies against p-PERK, total PERK, p-eIF2α, total eIF2α, BiP, and caspase-3 were purchased from Cell Signaling Technology. Antibodies against CHOP and Bcl-2 were purchased from Santa Cruz Biotechnology. Anti-ATF4 antibody and Anti-β-Actin antibody were sourced from Abcam and NeoMarkers, respectively. HRP-conjugated goat anti-mouse antibody and HRP-conjugated goat anti-rabbit antibody were obtained from Molecular Probes and Invitrogen, respectively.
The BHK-21 cells, cBS6-2 and cBS6-3, were seeded in 6-well plates and grown to 50% confluence. siCHOP, siBiP, and scramble siRNA were purchased from Santa Cruz. Transfection of siRNA was conducted using jetPRIME™ (Polyplus-transfection) according to the manufacturer’s instructions. The BHK-21 cells were infected with JEV at 24 hr post-transfection. The cells and the supernatant were harvested at 48 hr post-infection, and the PI cell clones were harvested at 40 hr post-transfection for further analysis. The collected supernatant was used for the quantification of the viral production, and viable cells were counted using trypan blue exclusion.
SDS-PAGE and Western blot analysis
For total protein extraction, virus- or mock-infected cells in monolayers were washed with cold PBS and then lysed in ice-cold M-PER buffer (Pierce) with a cocktail of protease inhibitors (Roche). Proteins were separated with 7.5% or 12% gradient PAGE using the Gradi-Gel™ gradient analysis kit (Elpis biotech); they were subsequently transferred to PVDF membrane (Millipore). The membrane was blocked with 5% skim milk in TBST. Primary antibodies were incubated at 4°C with membrane overnight in 3% skim milk in TBST or 3% BSA in TBST. After primary incubation, the membrane was washed in TBST once for 5 min and three times for 10 min; it then was incubated with secondary antibodies in 3% skim milk in TBST at room temperature for 2 hr. The membrane was washed again three times in TBST for 10 min, and the proteins were detected with an enhanced luminol-based chemiluminescent detection kit (AbFrontier) according to the manufacturer’s instructions.
The intensities of bands from Western blot analysis were quantified using the ImageJ program (National Institutes of Health) according to the developer’s instructions.
The data were presented as mean ± standard deviations (SD) of three independent experiments. The differences between groups were assessed by Student’s t test. A P value < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS.
We would like to thank Professor Radhakrishnan Padmanabhan (Georgetown University, USA) for providing flavivirus NS3-specific antibody.
- Rima RK, Martin SJ. Persistent infection of tissue culture cells by RNA viruses. Med Microbiol Immunol. 1976;162:89–119.View ArticlePubMedGoogle Scholar
- Mahy BW. Strategies of virus persistence. Br Med Bull. 1985;41:50–5.PubMedGoogle Scholar
- Rima BK, Duprex WP. Molecular mechanisms of measles virus persistence. Virus Res. 2005;111:132–47.View ArticlePubMedGoogle Scholar
- Dermody TS. Molecular mechanisms of persistent infection by reovirus. Curr Top Microbiol Immunol. 1998;233:1–22.PubMedGoogle Scholar
- Lin JH, Walter P, Yen TS. Endoplasmic reticulum stress in disease pathogenesis. Annu Rev Pathol. 2008;3:399–425.View ArticlePubMed CentralPubMedGoogle Scholar
- Walter P, Ron D. The unfolded protein response: from stress pathway to homeostatic regulation. Science. 2011;334:1081–6.View ArticlePubMedGoogle Scholar
- Woehlbier U, Hetz C. Modulating stress responses by the UPRosome: a matter of life and death. Trends Biochem Sci. 2011;36:329–37.View ArticlePubMedGoogle Scholar
- Malhotra JD, Kaufman RJ. The endoplasmic reticulum and the unfolded protein response. Semin Cell Dev Biol. 2007;18:716–31.View ArticlePubMed CentralPubMedGoogle Scholar
- Wang XZ, Kuroda M, Sok J, Batchvarova N, Kimmel R, Chung P, et al. Identification of novel stress-induced genes downstream of chop. EMBO J. 1998;17:3619–30.View ArticlePubMed CentralPubMedGoogle Scholar
- Oyadomari S, Mori M. Roles of CHOP/GADD153 in endoplasmic reticulum stress. Cell Death Differ. 2004;11:381–9.View ArticlePubMedGoogle Scholar
- Scull CM, Tabas I. Mechanisms of ER stress-induced apoptosis in atherosclerosis. Arterioscler Thromb Vasc Biol. 2011;31:2792–7.View ArticlePubMed CentralPubMedGoogle Scholar
- Tabas I, Ron D. Integrating the mechanisms of apoptosis induced by endoplasmic reticulum stress. Nat Cell Biol. 2011;13:184–90.View ArticlePubMed CentralPubMedGoogle Scholar
- He B. Viruses, endoplasmic reticulum stress, and interferon responses. Cell Death Differ. 2006;13:393–403.View ArticlePubMedGoogle Scholar
- Medigeshi GR, Lancaster AM, Hirsch AJ, Briese T, Lipkin WI, Defilippis V, et al. West Nile virus infection activates the unfolded protein response, leading to CHOP induction and apoptosis. J Virol. 2007;81:10849–60.View ArticlePubMed CentralPubMedGoogle Scholar
- Ambrose RL, Mackenzie JM. West Nile virus differentially modulates the unfolded protein response to facilitate replication and immune evasion. J Virol. 2011;85:2723–32.View ArticlePubMed CentralPubMedGoogle Scholar
- Pena J, Harris E. Dengue virus modulates the unfolded protein response in a time-dependent manner. J Biol Chem. 2011;286:14226–36.View ArticlePubMed CentralPubMedGoogle Scholar
- Umareddy I, Pluquet O, Wang QY, Vasudevan SG, Chevet E, Gu F. Dengue virus serotype infection specifies the activation of the unfolded protein response. Virol J. 2007;4:91.View ArticlePubMed CentralPubMedGoogle Scholar
- Chan SW, Egan PA. Hepatitis C virus envelope proteins regulate CHOP via induction of the unfolded protein response. FASEB J. 2005;19:1510–2.PubMedGoogle Scholar
- von dem Bussche A, Machida R, Li K, Loevinsohn G, Khander A, Wang J, et al. Hepatitis C virus NS2 protein triggers endoplasmic reticulum stress and suppresses its own viral replication. J Hepatol. 2010;53:797–804.View ArticleGoogle Scholar
- Misra UK, Kalita J. Overview: Japanese encephalitis. Prog Neurobiol. 2010;91:108–20.View ArticlePubMedGoogle Scholar
- Ravi V, Desai AS, Shenoy PK, Satishchandra P, Chandramuki A, Gourie-Devi M. Persistence of Japanese encephalitis virus in the human nervous system. J Med Virol. 1993;40:326–9.View ArticlePubMedGoogle Scholar
- Schmaljohn C, Blair CD. Persistent infection of cultured mammalian cells by Japanese encephalitis virus. J Virol. 1977;24:580–9.PubMed CentralPubMedGoogle Scholar
- Gavrilov VI, Deryabin PG, Lozinsky TF, Loghinova NV, Karpova EF, Zhdanov VM. Continuous mouse brain cell lines chronically infected with Japanese encephalitis virus. J Gen Virol. 1974;24:293–300.View ArticlePubMedGoogle Scholar
- Schmaljohn CS, Blair CD. Clonal analysis of mammalian cell cultures persistently infected with Japanese encephalitis virus. J Virol. 1979;31:816–22.PubMed CentralPubMedGoogle Scholar
- Shah PS, Gadkari DA. Persistent infection of porcine kidney cells with Japanese encephalitis virus. Indian J Med Res. 1987;85:481–91.PubMedGoogle Scholar
- Liao CL, Lin YL, Shen SC, Shen JY, Su HL, Huang YL, et al. Antiapoptotic but not antiviral function of human bcl-2 assists establishment of Japanese encephalitis virus persistence in cultured cells. J Virol. 1998;72:9844–54.PubMed CentralPubMedGoogle Scholar
- Park SY, Choi E, Jeong YS. Integrative effect of defective interfering RNA accumulation and helper virus attenuation is responsible for the persistent infection of Japanese encephalitis virus in BHK-21 cells. J Med Virol. 2013;85:1990–2000.View ArticlePubMedGoogle Scholar
- Yoon SW, Lee SY, Won SY, Park SH, Park SY, Jeong YS. Characterization of homologous defective interfering RNA during persistent infection of Vero cells with Japanese encephalitis virus. Mol Cells. 2006;21:112–20.PubMedGoogle Scholar
- Mathur A, Arora KL, Rawat S, Chaturvedi UC. Persistence, latency and reactivation of Japanese encephalitis virus infection in mice. J Gen Virol. 1986;67(Pt 2):381–5.View ArticlePubMedGoogle Scholar
- Tsai KN, Tsang SF, Huang CH, Chang RY. Defective interfering RNAs of Japanese encephalitis virus found in mosquito cells and correlation with persistent infection. Virus Res. 2007;124:139–50.View ArticlePubMedGoogle Scholar
- Su HL, Liao CL, Lin YL. Japanese encephalitis virus infection initiates endoplasmic reticulum stress and an unfolded protein response. J Virol. 2002;76:4162–71.View ArticlePubMed CentralPubMedGoogle Scholar
- Liao CL, Lin YL, Wang JJ, Huang YL, Yeh CT, Ma SH, et al. Effect of enforced expression of human bcl-2 on Japanese encephalitis virus-induced apoptosis in cultured cells. J Virol. 1997;71:5963–71.PubMed CentralPubMedGoogle Scholar
- Blazquez AB, Escribano-Romero E, Merino-Ramos T, Saiz JC, Martin-Acebes MA. Stress responses in flavivirus-infected cells: activation of unfolded protein response and autophagy. Front Microbiol. 2014;5:266.View ArticlePubMed CentralPubMedGoogle Scholar
- Jheng JR, Ho JY, Horng JT. ER stress, autophagy, and RNA viruses. Front Microbiol. 2014;5:388.View ArticlePubMed CentralPubMedGoogle Scholar
- Ron D. Translational control in the endoplasmic reticulum stress response. J Clin Invest. 2002;110:1383–8.View ArticlePubMed CentralPubMedGoogle Scholar
- Liao Y, Fung TS, Huang M, Fang SG, Zhong Y, Liu DX. Upregulation of CHOP/GADD153 during coronavirus infectious bronchitis virus infection modulates apoptosis by restricting activation of the extracellular signal-regulated kinase pathway. J Virol. 2013;87:8124–34.View ArticlePubMed CentralPubMedGoogle Scholar
- Gu X, Li K, Laybutt DR, He ML, Zhao HL, Chan JC, et al. Bip overexpression, but not CHOP inhibition, attenuates fatty-acid-induced endoplasmic reticulum stress and apoptosis in HepG2 liver cells. Life Sci. 2010;87:724–32.View ArticlePubMedGoogle Scholar
- Zinszner H, Kuroda M, Wang X, Batchvarova N, Lightfoot RT, Remotti H, et al. CHOP is implicated in programmed cell death in response to impaired function of the endoplasmic reticulum. Genes Dev. 1998;12:982–95.View ArticlePubMed CentralPubMedGoogle Scholar
- Marciniak SJ, Yun CY, Oyadomari S, Novoa I, Zhang Y, Jungreis R, et al. CHOP induces death by promoting protein synthesis and oxidation in the stressed endoplasmic reticulum. Genes Dev. 2004;18:3066–77.View ArticlePubMed CentralPubMedGoogle Scholar
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.