Antiviral activity of ethanol extract of Geranii Herba and its components against influenza viruses via neuraminidase inhibition

Influenza viruses are a serious threat to human health, causing numerous deaths and pandemics worldwide. To date, neuraminidase (NA) inhibitors have primarily been used to treat influenza. However, there is a growing need for novel NA inhibitors owing to the emergence of resistant viruses. Geranii Herba (Geranium thunbergii Siebold et Zuccarini), which is edible, has long been used in a variety of disease treatments in Asia. Although recent studies have reported its various pharmacological activities, the effect of Geranii Herba and its components on influenza viruses has not yet been reported. In this study, Geranii Herba ethanol extract (GHE) and its component geraniin showed high antiviral activity against influenza A strain as well as influenza B strain, against which oseltamivir has less efficacy than influenza A strain, by inhibiting NA activity following viral infection in Madin–Darby canine kidney cells. Thus, GHE and its components may be useful for the development of anti-influenza drugs.

inhibitory effects of GHe on nA activity. NA inhibitors play an important role in preventing the spread of influenza infection via inhibition of the enzyme function of NA, the surface glycoprotein of influenza virus, by attaching to its active site 11 . Accordingly, the active site of NA is a good target for the development of anti-influenza drugs. This study investigated the potential effects of GHE on influenza virus NA activity.  , and influenza type B (E) were added to the indicated concentrations of GHE and oseltamivir carboxylate (OTC). Fluorescence was measured using fluorescence spectrophotometry (excitation, 365 nm and emission, 415-445 nm). Bar graph (mean ± SEM) statistics were determined by three experiments' data using one-way ANOVA with Tukey's posthoc test, ***P < 0.001; **P < 0.01. n.s.: not significant, compared with the (GHE untreated) samples.
(B/Korea/72/2006) was significantly reduced with GHE and oseltamivir carboxylate (Fig. 1B-E). In particular, treatment with GHE (250 μg/mL) had significant effects on the NA activity of H3N2 and H1N1. We further assessed the NA activity of GHE using chemiluminescent-based neuraminidase inhibition (NI) assays. The results of this assessment confirmed that GHE inhibits NA activity in influenza A virus H3N2 similar to that demonstrated by the results of fluorescent-based NI assay ( Supplementary Fig. 1A,B). Moreover, GHE exhibited 3.1-12-fold increase in NA inhibition against influenza type B strain whereas influenza B strain was much less susceptible (13-32-fold) to oseltamivir carboxylate than influenza A strain ( Fig. 1B-E). The results suggest that GHE has an additional inhibitory effect on the influenza virus release stage by inhibiting the NA of A/PR/8/34, H3N2, H1N1, and type B in a dose-dependent manner.

GHE inhibited the infection of influenza virus in MDCK cells. To investigate if GHE inhibits influ-
enza A virus infection in MDCK cells, we examined viral replication in GHE-treated MDCK cells (100 or 200 μg/ mL) infected with H1N1 (Fig. 2). We observed that GHE-treated MDCK cells had significantly increased cell survival rate compared to the cells exposed only to H1N1 ( Fig. 2A), indicating that treatment with GHE reduces viral replication in MDCK cells. Moreover, GHE-treated MDCK cells showed reduced green fluorescent protein (GFP) expression levels compared to the untreated cells with high GFP expression levels upon infection with A/ PR/8/34-GFP at 24 h (Fig. 2B). Additionally, flow cytometry analysis using fluorescence detection and plaque and NS-1 (F) were analyzed using quantitative real-time polymerase chain reaction and normalized to β-actin mRNA levels. Bar graph (mean ± SEM) statistics were determined by three experiments' data using one-way ANOVA with Tukey's post-hoc test, ***P < 0.001, compared with the CON (GHE untreated) samples. ### P < 0.001, compared with the cell only sample. GHE reduced the expression of influenza A virus (H1N1) proteins NP (nuclear protein) and NA in infected MDCK cells. The reduction of NP (G) and NA (H) proteins in MDCK cells were observed with fluorescence microscopy using the influenza A viral proteins NP-and NA-specific antibodies. MDCK cells were also stained with DAPI (blue), and the merged images represent NP and NA (red). Influenza H1N1 virus protein levels (NP, PA, M1, M2, PB1, PB2, HA, and NA) in MDCK cell lysates were detected using Western blotting, and β-actin was analyzed as an internal control (I). Full-length blots are presented in Supplementary Fig. 2. reduction assay showed that GHE effectively inhibits viral replication in MDCK cells (Fig. 2C,D). At the highest concentration (200 μg/mL) of GHE, viral titers were reduced by 4.8 log 10 TCID 50 /mL at 48 h post infection ( Supplementary Fig. 1D). We further confirmed that GHE also inhibited viral growth in MDCK cells infected by influenza A virus at low multiplicity of infection (MOI) conditions (0.1 and 0.01) using NA-XTD influenza neuraminidase assay ( Supplementary Fig. 1E,F). These results indicate that GHE-treated cells exhibited significantly reduced cell death and viral load following infection with influenza virus as compared to untreated cells.

GHE reduced mRNA and protein levels of influenza a virus.
We investigated if GHE inhibited the synthesis of influenza A HA and NS-1 mRNA. A/PR/8/34-infected MDCK cells were harvested after treatment with GHE at a concentration of 100 or 200 μg/mL, and the relative mRNA expression levels of the viral genes at 24 h were evaluated using quantitative real-time polymerase chain reaction (qRT-PCR). HA and NS-1 mRNA synthesis was completely inhibited in GHE-treated MDCK cells versus untreated MDCK cells at 24 h (Fig. 2E,F).
Furthermore, we evaluated the effect of GHE on the expression of influenza A virus proteins, such as NP (nucleoprotein) and NA (neuraminidase), using immunofluorescence analysis in GHE-treated MDCK cells at 24 h after infection with H1N1. The expression of influenza A virus proteins NP and NA was inhibited in GHE-treated MDCK cells (100 and 200 μg/mL) compared to that in the untreated cells upon infection with H1N1 at 24 h (Fig. 2G,H). In addition, the effect of GHE on the expression of influenza A virus protein NP was evalu-
The inhibition of NA activity was confirmed using the NA-Fluor ™ influenza neuraminidase assay. The nine components (at 100 μM) were added to the viruses A/PR/8/34, H1N1, and H3N2. Oseltamivir carboxylate was used as a positive control. We noted that the treatment with GN demonstrated the highest NA-inhibitory activity against all viruses ( Fig. 3C-E). Furthermore, we evaluated the dose-dependent NA inhibition activity of GN and observed that GN is effective on both influenza type B and A strains, particularly H1N1 ( Fig. 4A-D).
To investigate if the nine aforementioned GHE components inhibited influenza viral infection in MDCK cells, we examined viral replication in these cells treated with the nine components (at 100 μM) and infected with H1N1. We observed that GN treatment increased cell survival rate after viral infection (Fig. 4E,F), indicating that treatment with GN reduces viral replication in MDCK cells. In addition, GN-treated MDCK cells showed reduced GFP expression levels compared to the untreated cells, which displayed high GFP expression levels upon infection with PR8-GFP at 24 h (Fig. 4G). Moreover, the reduction of viral replication was confirmed using flow cytometry analysis with fluorescence detection of PR8-GFP and plaque reduction assay (Fig. 4H,I). Moreover, after 24 h of infection with A/PR/8/34-GFP, GN-treated A549 cells showed reduced GFP expression levels compared with the untreated cells with high GFP expression levels ( Supplementary Fig. 3A). We investigated if GN inhibits the synthesis of influenza A M1 and NS-1 mRNA. A/PR/8/34-GFP-infected A549 cells were harvested after treatment with GN at a concentration of 10 or 100 μM, and the relative mRNA expression levels of the viral genes at 24 h were evaluated using qRT-PCR. M1 and NS-1 mRNA synthesis completely inhibited in GN-treated A549 cells versus that in untreated A549 cells at 24 h ( Supplementary Fig. 3B,C). Additionally, NA-XTD ™ assay using cell-based virus growth inhibition assay showed that GN effectively inhibited viral replication in A549 cells ( Supplementary Fig. 3D). We also investigated the effect of GN on the expression of influenza A virus proteins NP and NA using immunofluorescence analysis 24 h after infection with H1N1. The NP and NA expression was inhibited in MDCK cells treated with GN (10 and 100 μM) compared to that in the untreated cells upon infection with H1N1 at 24 h (Fig. 4J). In addition, the effect of GN on the expression of influenza A virus protein NP was evaluated using immunofluorescence analysis in GN- www.nature.com/scientificreports www.nature.com/scientificreports/ These results suggest that among the constituents of GHE, GN is a major effective substance for inhibiting influenza virus infection via NA inhibition.

Protein-ligand docking simulation and pharmacophore analysis of the components in GHE.
Since oseltamivir carboxylate inhibits the release of progeny virions from infected host cells by binding to NA, we investigated the binding affinity of two components in GHE to H1N1 NA (PDB code: 3TI6) by using the protein-ligand docking simulation with AutoDock Vina; these components were effective in NA inhibition as per above experiments. The predicted binding affinity between H1N1 NA and oseltamivir carboxylate was −6.6 kcal/mol   www.nature.com/scientificreports www.nature.com/scientificreports/ (Fig. 5A). The binding affinity for GN, which is present at high concentrations in GHE, was −9.9 kcal/mol (Fig. 5B), indicating that it strongly binds to H1N1 NA corresponding to the resultant inhibition of H1N1 and A/ PR/8/34 (GN). Despite the higher affinity of GN observed using docking simulation, in reality, the NA-inhibitory activity of oseltamivir is higher than that of GN, which means that more intensive studies on the relationships between activity and binding structure, such as binding direction and interactions, are required to predict the NA-inhibitory activity. Nevertheless, we further investigated the interactions between the residues of H1N1 NA and GN to assess the NA-inhibitory activity through pharmacophore analysis using LigPlot + software. Oseltamivir carboxylate is predicted to bind to H1N1 NA via six hydrogen bonds and seven hydrophobic interactions, involving the critical residues R225 and E277, which contribute to the creation of a hydrophobic pocket where the hydrophobic pentyl ether side chain of oseltamivir gets occupied by the rotation of E277 toward R225 37 . GN can form eight hydrogen bonds and nine hydrophobic interactions with the residues of H1N1 NA, including R225 and E277. The docking simulation between GHE and NA suggests that GN inhibits NA activity by binding to the hydrophobic pocket of NA with high affinity (Fig. 5B).

Discussion
Geranii Herba (GH), belonging to the family of Geraniaceae, is a perennial plant found in Asia (Korea, China, and Japan) and has been traditionally used as anti-diarrhetic drug in East Asia 27 . GH is listed on Korean pharmacopeia and Korean Food Standards Codex. It means that GH is safe to use for additives in foods. Although there is no report on toxicity studies of GH, GH would have a low toxicity because GH allow to use food additives according to Korean Food Standards Codex. When we consider that many effective drug candidates are dropped because of their toxicities, GH would be good drug candidate. Recently, it has been studied that GH has anti-obesity effects via improvement of lipid metabolism 27 and anti-inflammatory effects by Nrf2 activation 29 .
Despite evidence of its many medicinal uses against obesity, diarrhea, and dysentery, there have been no reported antiviral effects of GH, including therapies for the treatment of influenza virus infections.
In this study, we found that GHE inhibits the infection of influenza viruses by NA suppression and increased the survival of MDCK cells infected with influenza viruses. In addition, HA and NS-1 were not detected in qRT-PCR of A/PR/8/34-infected MDCK cells, suggesting that GHE has an antiviral mechanism for preventing the export of replicated viruses from infected cells. We further identified the constituent compounds present in GHE using HPLC and confirmed that GN exhibits antiviral effects. In addition, we investigated that the amounts of identified phytochemicals ranged from 1.2 to 65.2 mg/g in GHE extract, and kaempferitrin and GN were found to be the most abundant at 65.2 and 39.4 mg/g, respectively (Table 1).
Based on these experimental results, the NA-inhibiting properties of the nine GHE components were investigated using NA inhibition assay as well as docking simulation and interaction analysis. Among the GHE constituents, GN showed the most promising NA-inhibitory effect. GN is one of the ellagitannins which contain up to 10% of GH. GN is present in different parts of various types of plants and exhibits antiviral activity against HSV-1 and -2 30 , HIV-1 38 , HBV 39 , human enterovirus 71, and dengue virus type-2 40 . However, to date, there has been no report on the specific uses of GN in the inhibition of the NA activity of influenza viruses.
We investigated the mechanism of NA-inhibitory action of GN. We performed in silico modeling to investigate whether GN binds directly to the receptor NA protein of influenza virus. Pharmacophore assays were then used to identify key NA protein inhibitor interactions and common properties with the known binding compound oseltamivir carboxylate. Docking simulations performed with AutoDock Vina predicted the binding affinity of GN and oseltamivir carboxylate with the NA protein, which showed that GN is a potential NA inhibitor. This is the first report showing that GHE has anti-influenza virus activity. In addition, we showed that GN had greater anti-influenza activity than the other constituent compounds of GHE, suggesting that GN was an important active component of GHE. Interestingly, GN is known to be hydrolyzed to corilagin, gallic acid, ellagic acid, and acid after 1 h of decoction 35 . Bastian et al. showed that GN increase until 10 min into decoction, after that, it reduced rapidly. So, when we use a GN as an active compound in Oriental clinic, GH should be decocted for 10 min to maximize extraction of GN and minimize of its hydrolysis. Although we showed some evidences to demonstrate the antiviral effects of GHE and its compounds, further research is required to identify additional pharmacological mechanisms of GHE and to investigate GN as prophylactic and therapeutic agents against influenza viral infection through in vivo experiments. Additionally, further comparative analysis between GN structure and NA sequence depending on influenza virus types may provide a more interesting interpretation for the anti-influenza activity of GN. In the early 2000s, NA inhibitor (e.g. oseltamivir and zanamivir)-resistant influenza viruses emerged following mutations in E119, H274, R292, and N294 of NA 41,42 ; thus, the research regarding the efficacy of natural products, including GHE and GN, against these resistant viruses will be also of interest.
In conclusion, this study was the first to demonstrate the efficacy of GHE against influenza viruses and proposes an alternative to the existing anti-influenza drugs. GHE inhibited the infection of influenza virus by NA suppression and increased the survival of MDCK cells infected with influenza virus. GHE and its components, including GN, may be useful as a therapeutic or prophylactic agent for restricting viral replication via NA at 37 °C with 5% CO 2 . GN reduced the expression of influenza A virus (H1N1) proteins NP and NA in infected MDCK cells to virus (J). The reduction of NP and NA proteins in MDCK cells were observed with fluorescence microscopy using the influenza A viral proteins NP-and NA-specific antibodies. MDCK cells were also stained with DAPI (blue), and the merged images represent NP and NA (red). Bar graph (mean ± SEM) statistics were determined by three experiments' data using one-way ANOVA with Tukey's post-hoc test, ***P < 0.001; *P < 0.05, compared with the CON (GHE untreated) samples. ###  www.nature.com/scientificreports www.nature.com/scientificreports/ inhibition. In addition, GHE and its components could be a useful starting material for the development of potent pharmaceutical drugs for selected influenza virus infections.

preparation of GHe. GHE was purchased from the NIKOM (National Development Institute of Korean
Medicine, Gyeongsan, South Korea). The freeze-dried extract powder was dissolved in DMSO and centrifuged at 12,000 rpm for 20 min to remove the insoluble residues, following which the supernatant was stored in desiccators at 4 °C until further use.

MTS assay. Cell viability was determined by using the CellTiter 96 ® AQueous One Solution Cell Proliferation
Assay (Promega, Madison, WI, USA), according to the manufacturer's instructions. MDCK cells (1 × 10 4 cells/ well) were seeded into 96-well plates and GHE was added to the wells at concentrations of 0-400 μg/mL. After 48 h, MTS solutions were added to each well and the cells were incubated for additional 2 h. Subsequently, absorbance at 490 nm was recorded using a GloMax ® Explorer Multimode Microplate Reader (Promega, Madison, WI, USA). The values of MTS assay were represented by the mean ± SEM of four independent experiments. Antiviral assay. The inhibition of viral replication was assayed as previously described 25 . Briefly, MDCK cells were cultured in 96-well plates (2 × 10 4 cells/well) for 16 h. Then differing GHE concentrations (100 or 200 μg/mL) were added to H1N1 [multiplicity of infection (MOI) = 1] and A/PR/8/34-GFP (MOI = 1) and the mixtures were incubated at 37 °C for 1 h. MDCK cells were infected with these mixtures at 37 °C for 2 h. Afterwards, the virus was removed, and cells were washed three times with PBS, and the medium was replaced by complete DMEM. The cells were incubated for 48 h at 37 °C with 5% CO 2 . The reduction of viral cytopathic effect was determined by measuring cell viability measured using MTS assay as described above 25 . Influenza virus GFP expression was measured under a fluorescence microscope (Olympus, Tokyo, Japan) following 24 h of viral infection. In addition, antiviral activities of GHE components were evaluated at 100 μM concentration by the same method described above. The values of antiviral assay were represented by the mean ± SEM of four (GHE) and three (GHE components) independent experiments. Plaque reduction assay. For the plaque reduction assay 45 , we used a slightly modified version of the previously used plaque reduction assay. Briefly, MDCK cells were cultured in 12-well plates (5 × 10 5 cells/well) for 18 h. Subsequently, A/PR/8/34 (MOI = 1) was mixed with different concentrations of GHE (100 and 200 μg/mL) and GN (10 and 100 μM), and the mixtures were incubated at 37 °C for 1 h. Then, MDCK cells were infected with these mixtures at 37 °C for 2 h. Subsequently, the virus was removed, the cells were washed three times with PBS, and the MDCK monolayer was covered with 1.5% agarose with 2X complete DMEM and TPCK trypsin at 2 mg/ml. The entire MDCK monolayer along with agarose was cultivated for 3 days at 37 °C with 5% CO 2 , fixed with 10% formalin, stained with 1% crystal violet solution, and the plaques were then counted. Analysis of GFP expression using flow cytometry. MDCK cells were cultured in 12-well plates (5 × 10 5 cells/well) for 18 h. Next, A/PR/8/34 (MOI = 1) was mixed with different concentrations of GHE (100 and 200 μg/mL) and GN (10 and 100 μM), and the mixtures were incubated at 37 °C for 1 h. Then, MDCK cells were infected with these mixtures at 37 °C for 2 h. Subsequently, the virus was removed, and the cells were washed three times with PBS and the medium was replaced by complete DMEM. The cells were incubated for 24 h at 37 °C with 5% CO 2 . The reduction of viral infection effect was determined by measuring GFP expression using flow cytometry. MDCK cells were harvested and resuspended in 1 mL of PBS containing 2% FBS and fixed in suspension with 4% paraformaldehyde. The cells were washed three times with PBS and stored at 4 °C until analysis with a CytoFLEX flow cell counter (Beckman). We analyzed the data using FlowJo software. ni assay. The NI assay was performed using an NA-Fluor ™ Influenza Neuraminidase Assay and NA-XTD ™ Influenza Neuraminidase Assay Kit (Applied Biosystems, Foster City, CA, USA) as per the manufacturer's instructions with slight modifications 22,24 . For the NA-Fluor ™ influenza neuraminidase assay, GHE was added to the assay buffer in 96-well plates at concentrations of 0-500 μg/mL for A/PR/8/34, H3N2, H1N1, and influenza type B viruses. A/PR/8/34, H1N1, H3N2, or influenza type B in assay buffer were added to the GHE-containing wells and incubated at 37 °C. Oseltamivir was considered as the positive control in the assay. After 30 min, NA-Fluor Substrate was added to each well and incubated for additional 2 h, followed by recording the fluorescence (excitation: 365 nm; emission: 415-445 nm) with a fluorescence spectrophotometer (Promega, Madison, WI., USA). Samples treated with only GHE or its components were used as negative controls. In addition, the NA activities for 100 μM of GHE components were evaluated by the same method described above. As a positive control, NA activity of oseltamivir carboxylate was measured in a range of 0 to 10,000 nM. The values of NI assay were represented by the mean ± SEM of three independent experiments.
For the NA-XTD ™ influenza neuraminidase assay, GHE was added to the assay buffer in 96-well plates at the concentrations of 0-500 μg/mL for H3N2. Subsequently, H3N2 in assay buffer were added to GHE-containing wells and incubated at 37 °C. Oseltamivir was considered as the positive control in the assay. After 20 min, NA-XTD ™ substrate was added to each well and incubated for additional 30 min. Then NA-XTD ™ accelerator was added to each well, followed by recording the luminescence using a luminescence plate reader (Promega, Madison, WI., USA). qRt-pcR. The qRT-PCR assay was conducted as previously described 25 . Briefly, total RNA was extracted from H1N1-infected MDCK cells treated or untreated with GHE (100 or 200 μg/mL) using an RNeasy Mini kit (Qiagen, Hilden, Germany). qRT-PCR was carried out using an AccuPower ® 2× Greenstar qPCR Master Mix (Bioneer, Daejeon, South Korea) and a CFX96 Touch Real-Time PCR System (Bio-Rad, Hercules, CA, USA) according to the manufacturers' instructions 25 . The primer sense and antisense sequences used were as follows: HA (5′-ttgctaaaacccggagacac-3′ and 5′-cctgacgtatttgggcact-3′); NS1 (5′-gcgatgccccattccttg-3′ and 5′-atccgctccactatctgctttc-3′) and canine β-actin (5′-tgccttgaagttggaaaacg-3′ and 5′-ctggggcctaatgttctcaca-3′). The values of qRT-PCR were represented by the mean ± SEM of three independent experiments. Docking simulation and interaction analysis. The components of GHE and the positive control oseltamivir carboxylate were docked onto the predefined binding pocket of the H1N1 NA crystal structure (PDB code: 3TI6) retrieved from the Protein Data Bank (www.rcsb.org) using AutoDock Vina integrated with UCSF Chimera-alpha v1.13. After docking simulation, the lowest energy scoring binding mode for each compound was selected. The hydrogen bonding and hydrophobic interactions between H1N1 NA and each compound were investigated with LigPlot + v1.4.5. Amino acid residues involved in the interactions were indicated with green (H-bonds) and red (hydrophobic interactions).
Immunofluorescence staining. For the immunofluorescence analysis, we used a slightly modified version of the previously used immunofluorescence analysis method 26 . Briefly, MDCK cells were cultured in 4-well tissue culture slides (5 × 10 4 cells/well) for 16 h. Subsequently, H1N1 (MOI = 1) and A/PR/8/34-GFP (MOI = 10) were mixed with different concentrations of GHE (100 and 200 μg/mL) and GN (10 and 100 μM), and the mixtures were incubated at 37 °C for 1 h. MDCK and A549 cells were infected with these mixtures at 37 °C for 2 h. Thereafter, the virus was removed, and the cells were washed three times with PBS and were cultured in a CO 2 incubator at 37 °C for 24 h. The cells were then washed three times with cold PBS and fixed with 4% paraformaldehyde in PBS and 1% Triton X-100 for 10 min each at room temperature. After blocking, the fixed cells were incubated overnight at 4 °C with NP-and NA-specific antibodies, washed three times (5 min per wash) with TBS, and incubated with Alexa Fluor 568 goat anti-rabbit IgG antibody (1:1,000; Life Technologies, Eugene, OR, USA) and washed three times (5 min per wash) with TBS. Next, the cells were incubated with DAPI for 10 min and measured using fluorescence microscopy.
Western blot analysis. MDCK cells were cultured in 6-well plates (1 × 10 6 cells/well) for 18 h. Then, H1N1 was mixed with different concentrations of GHE (100 and 200 μg/mL), and the mixtures were incubated at 37 °C for 1 h. MDCK cells were infected with these mixtures at 37 °C for 2 h. Afterwards, the virus was removed, the cells were washed three times with PBS, and the medium was replaced by complete DMEM. After 24 h, the cells were harvested and subjected to western blotting using whole cell extracts 26 . The PVDF membrane was then blocked with 5% BSA in TBS-T buffer for 1 h and overnight at 4 °C with primary anti-PA, -HA, -NA, -NP, -PB1, -PB2, -M1, -M2, -NS-1, -NS-2, and -β-actin antibodies (1:1,000 dilution). Primary antibodies were washed three times (5 min per wash) with TBS-T buffer and incubated with HRP-conjugated secondary antibodies (1:5,000 dilution) at room temperature for 1 h; the relative intensities of protein bands were measured using Image J program 26 . The experiment was carried out three times independently, and similar results were obtained each time.
Statistical analysis. Data are expressed as mean ± SEM. Differences in the mean values between the treatment and control groups were determined to be statistically significant by using one way ANOVA was performed with Tukey's post-hoc test for multiple comparisons. Analyses were performed using GraphPad PRISM software ® Version 5.02 (GraphPad, La Jolla, CA, USA). P < 0.05 was considered to denote statistical significance.

Data Availability
The datasets in this study are available from the corresponding author on reasonable request.