Anti-herpes activity of the total alkaloid fraction from the branches of Fusaea longifolia (Annonaceae)

ABSTRACT The herpes simplex virus type 1 (HSV-1) and type 2 (HSV-2) occur worldwide. Infections caused by these viruses have great public health importance due to the growing resistance to the first-choice drug, acyclovir, especially in immunosuppressed patients. Alkaloids derived from species of Annonaceae have been reported as antiviral agents against HSV and others viruses. Within this context, we evaluated the antiviral activity of the total alkaloid fraction (TAF) extracted from the branches of Fusaea longifolia (Aubl.) Saff. (Annonaceae), a species native to the Amazon region, against the HSV-1 and HSV-2 viruses. The antiviral activity was evaluated through the plate reduction assay and the mode of action was investigated by a set of other assays. The TAF was active against the HSV-2 strain 333 and against the HSV-1 strains KOS and 29R (acyclovir resistant), with selectivity index values (SI = 50% cytotoxic concentration/50% effective concentration) of 5, 4 and 3, respectively. In the preliminary study of the anti-HSV-2 mode of action, TAF showed viral inhibitory effects if added up to 12 h post-infection, had virucidal activity and did not present viral inhibition in pre-treatment. Our results showed that the TAF exhibited anti-HSV activity. Regarding HSV-2, TAF acted after the viral infection and had virucidal activity. A mass spectrometry analysis revealed the presence of nine alkaloids in the TAF that had previously been reported for Annonaceae, including liriodenine, lysicamine and isoboldine, which have been described as potential anti-HSV-1 agents.


INTRODUCTION
The herpes simplex viruses (HSV) type 1 (HSV-1) and type 2 (HSV-2) are distributed worldwide and produce lifelong infections with frequent reactivations throughout life (James et al. 2020).HSV-2 infection is the most common cause of recurrent genital ulcer disease (Looker et al. 2020) and is often related to negative psychosocial consequences such as shame, anxiety and depression (Treml et al. 2020).HSV-1 normally causes orolabial infection, but has been increasingly implicated as a cause of genital herpes (Looker et al. 2020).In 2016, it was estimated that 491 million people aged 15-49 years were infected by HSV-2 and that 3,752 million people were infected by HSV-1, at any site, in 0-49-year-olds.That data corresponds to a global prevalence of 13.2% and 66.6% of the world population, respectively (James et al. 2020).
HSV infections have great importance to public health due to the increasing problem of resistance to first-line drugs, acyclovir and its derivatives (Treml et al. 2020;Anton-Vazquez et al. 2020).Furthermore, epidemiological evidence suggests that HSV-2 infection also increases susceptibility to HIV infection, elevating it even further as a public health concern (Schiffer and Gottlieb 2019;Hopkins et al. 2020).Therefore, the search for new antiviral agents, with different mechanisms of action, is a fundamental and necessary issue (Sharifi-Rad, et al. 2018;Benassi-Zanqueta et al. 2019).
Although the family has been widely investigated from the phytochemical point of view, and is considered a promising source of bioactive alkaloids (Lebouef et al. 1982), some Annonaceae species remain unexplored.One example is Fusaea longifolia (Aubl.)Saff., which is known in the Amazon region as araticum, ata or biritá, and is still in the early stages of phytochemical investigation (Lebouef et al.1982;Tavares et al. 2005;Rios et al. 2011).Fusaea longifolia is used to treat ailments by the Urubu-Kaapor indigenous people (Rios et al. 2011), and the essential oil of its aerial parts showed trypanocidal activity against amastigote and trypomastigote forms of Trypanosoma cruzi (Bay et al. 2019).
In this context, the present study aimed to evaluate the antiviral activity of the total alkaloid fraction from the branches of F. longifolia against the HSV-1 and HSV-2 viruses, as well as to perform the dereplication of the alkaloids from this sample.

Alkaloid extraction
The branches were dried in an oven with air circulating at 45 °C for seven days.The ground plant material (52.63 g) was humidified with 5% ammonium hydroxide (NH 4 OH) and subjected to exhaustive extraction by maceration with 95% ethanol, ratio of 1:5 (w/v).The extractive solution was concentrated under vacuum to obtain the crude ethanolic extract (CEE).The CEE (5.12 g) was put under mechanical stirring with 3% HCl and then filtered over Celite R .The obtained filtrate (acid solution) was alkalized with NH 4 OH (pH 9-10) and extracted with chloroform until negative to Dragendorff's reagent.The chloroform phase was concentrated under vacuum and provided 1.0361 g of a total alkaloid fraction (TAF) (Tavares et al. 2005).

Fractionation of the total alkaloid fraction
Part of the total alkaloid fraction (0.51 g) obtained was solubilized in chloroform and incorporated into 1 g of silica gel 60 (70-230 mesh, Merck) previously treated with a 10% sodium bicarbonate solution (Costa et al. 2010).After eliminating the solvent, the TAF was submitted to chromatographic fractionation in a chromatography column (CC) with 15 g of silica gel 60 (70-230 mesh, Merck) previously treated with 10% sodium bicarbonate.The alkaloids were eluted using a polarity gradient with n-hexane/ dichloromethane/ethyl acetate/methanol), affording 170 fractions (15 mL each).The eluted fractions were analyzed and pooled according to the chromatographic profile obtained by thin layer chromatography (TLC), after examination under ultraviolet (UV) light (254 and 365 nm) and revealed with Dragendorff's reagent, affording seven groups: A (57.0 mg), B (18.6 mg), C (13.7 mg), D (29.9 mg), E (33.1 mg), F (92.1 mg) and G (134.4 mg).The groups were purified using preparative TLC eluted with chloroform:methanol (95:05, v/v) and UV light at 254 nm and Dragendorff's reagent for visualization.Sample recovery was performed using dichloromethane, chloroform, methanol and/or mixtures of these solvents.

Mass spectrometry analysis
The APCI-MS and APCI-MS/MS analyses were performed as following: the stock solutions of the samples A to G were prepared with methanol (HPLC grade) at 1.0 mg mL -1 , diluted to 10.0 µg mL -1 , and analyzed using direct infusion via an instrument syringe pump (10.0 µL min -1 ) in a TSQ Quantum Access mass spectrometer (Thermo Scientific, San Jose, CA, USA) equipped with an atmospheric pressure chemical ionization source (APCI) and programmed to operate in positive acquisition mode (Lima et al. 2020).

Cells and viruses for in vitro assay
All biological assays were performed using Vero E6 cells, cultivated in Leibovitz medium (L-15), supplemented with 5% fetal bovine serum (FBS) and 1% PSA (10,000 IU mL -1 of penicillin G, 10,000 µg mL -1 of streptomycin and 25 µg mL -1 of amphotericin B), and maintained at 37 °C.The following viruses were used: Herpes simplex virus type 2 (HSV-2), 333 strain, from the Department of Clinical Virology, University of Goteborg, Sweden, and the Herpes simplex virus type 1 (HSV-1) KOS and 29R strains, which are acyclovir (AVC) sensitive and resistant, respectively, from the Laboratory of Applied Virology at the Federal University of Santa Catarina, Brazil.The Viruses were propagated in Vero E6 cells.Viral stocks were stored at -80°C and titrated based on plaque forming units (PFU) counted by plaque assay as previously described by Burleson et al. (1992).4,5-dimethylthiazol-2-yl) -2,5-diphenyltetrazolium bromide] assay (Mosmann 1983), with minor modifications.The Vero E6 cells (1.8 x10 5 cells mL -1 ) were cultured in 96well microplates at 37 °C for 24 h until confluence of the cell monolayer.Several concentrations of the TAF (200 µg mL -1 to 0.78 µg mL -1 ) or acyclovir (225.21 µg mL -1 to 0.88 µg mL -1 ), 1:2 ratio were added to the cells and incubated at 37 °C, for 48 h.After that period, the supernatants were removed and 50 µL of MTT solution (1mg mL -1 ) was added per well for 4 h at 37 °C.Then the supernatant was removed and 100 µl of dimethyl sulfoxide (DMSO) was added in each well.All experiments were performed in duplicate with three plates.The absorbances were read in a spectrophotometer (Biolisa Reader R792, Bioclin, China) at 495 nm.The percentages of cell viability (X%), in relation to the untreated cell control were obtained by the following formula: X% = sample Abs x 100/untreated cell control Abs, where Abs corresponds to the absorbances.The 50% cytotoxic concentration (CC 50) was defined as the concentration that reduced 50% of cell viability when compared to the untreated cell control.

Viral plaque reduction assay
This test was performed according to Kuo et al. (2001), with some modifications.Briefly, confluent cell monolayers (1.8 x10 5 cells mL -1 ) in 24-well plates were infected with approximately 50 PFU of each virus strain for 1 h at 37 ºC.The viral suspension was removed and treated with different concentrations of the TAF (50.00 to 3.12 µg mL -1 , ratio 1:2) or acyclovir (positive control) (50.00 µg mL -1 to 0.05 µg mL -1 , ratio 1:10), prepared in a solution of L-15 medium (2X) containing 1.5% carboxymethylcellulose (CMC).After 72 hours or 48 hours for HSV-1 and HSV-2, respectively, the medium was removed, cells were fixed and stained with naphthol blue-black and plaques were counted.Experiments were performed in duplicate with three plates.The EC 50 was calculated as the effective concentration that inhibited the formation of viral plaques by 50% when compared to untreated controls.The selectivity index (SI=CC 50 /EC 50 ) was calculated.

Pre-treatment
This assay was performed according to Bertol et al. (2011).The Vero E6 cells confluent layer was treated with different concentrations of TAF (50 to 3.12 µg mL -1 , ratio 1:2) and incubated for 3 h at 37 ºC.After this period, the samples were removed, the cell monolayers were washed with PBS and infected with 50 PFU of HSV-2 for 1 h at 37 ºC.After this period, the medium was removed, the infected cells were overlaid with L-15 containing 1.5% CMC, incubated at 37 ºC for 48 h, and treated as described earlier for the plaque reduction assay.Experiments were performed in duplicate with three plates.

Virucidal activity
This test was performed according to Kuo et al. (2001), with some modifications.Mixtures of equal volumes of TAF at different concentrations (100 to 6.25 µg mL -1 , ratio 1:2) and 50 PFU of HSV-2 were co-incubated for 30 minutes at 37 °C.After this time, the residual infectivity was determined by plaque reduction assay as described above.Untreated viral suspensions were used as infection controls.Experiments were performed in duplicate with three plates.EC 50 and SI values were calculated as described above.

TAF effect on viral adsorption
This assay was performed according to Bertol et al. (2011) with some modifications.Confluent Vero E6 cells were prechilled at 4 ºC for 30 minutes and exposed to a mixture of equal volumes of TAF at different concentrations (100 to 6.25 µg mL -1 , ratio 1:2) and 50 PFU of HSV-2 were incubated at 4 ºC for 2 hours.Viral controls were treated with culture medium only.Unabsorbed virus and remaining samples were removed by washing with cold PBS and cells were overlaid with L-15 containing 1.5% CMC, incubated at 37 ºC for 48 ACTA AMAZONICA h, and treated as described earlier for plaque reduction assay.One of the viral controls was washed with a citrate buffer (pH = 3.0), because, at this pH, the viral particles adsorbed on the plasma membranes are destabilized, thus promoting virus release and elimination.Experiments were performed in duplicate with three plates.

Time-of-addition effect on viral replication cycle
This assay was performed according to Bertol et al. (2011) with some modifications.Vero E6 cells (1.8 x10 5 cells mL -1 ) were infected with 50 PFU of HSV-2 and incubated for 1h at 37 ºC.The TAF was prepared at the maximum non-toxic concentration (50 µg mL -1 ) and was added at different times: simultaneously with the viral suspension (time zero) and at intervals of 1, 2, 4, 6, 8, 10, 12, 18 and 24 h post-infection.After 48 h of incubation and cells were treated as described earlier for plaque reduction assay.Experiments were performed in triplicate with three plates.

Data analysis
The dates expressed as mean values ± standard deviation.CC 50 and EC 50 values were estimated from concentration effect curves after non-linear regression analysis by using Microsoft Excel®.Student´s t test (p < 0.05) was carried out to compare SI values of TAF and acyclovir (positive control) for each virus separately, using GraphPad Prism software (version 5.0 for Windows).

Antiviral evaluation of TAF fractions
The MTT assay to determine TAF cytotoxicity resulted in a CC 50 of 72.9 ± 0.6 µg mL -1 (Table 1) and a maximum nontoxic concentration of 50.0 µg mL -1 , which was used in the antiviral activity assays.Acyclovir was not cytotoxic at the maximum concentration tested of 225.21 µg mL -1 .
In the pretreatment assay, no inhibition of HSV-2 replication was observed.In the virucidal assay, the complete inactivation of HSV-2 infectivity was observed, with a EC 50 of 3.13 ± 18.86 µg mL -1 and a SI of 23.In the viral adsorption assay, the TAF showed 53.71 ± 14.10% of viral inhibition at the highest concentration tested (50.0 µg mL -1 ), not being effective in other concentrations.In the time-of-addition test, the TAF (50.0 µg mL -1 ) inhibited the replication of the HSV-2 virus when the fraction was added up to 12 h post-infection and decreased thereafter (Figure 1).

HSV-2 333 strain
HSV-1 KOS strain  ACTA AMAZONICA alkaloids, such as carbon monoxide (CO) (28 Da) and a methyl radical (•CH 3 ) (15 Da) loss (Silva et al. 2017).At m/z 276, two subsequent losses of 28 Da (m/z 276 → m/z 248 and m/z 248 → m/z 220) were observed (Table 2).This fragmentation pattern is consistent with the oxoaporphine alkaloid liriodenine (Figure 2 [1]) (Silva et al. 2017;Lima et al. 2020) The MS 2 spectra of the ions at m/z 296, 282, and 328 (Supplementary Material, Figure S6-S8) presented key fragmentation patterns commonly observed in aporphine alkaloids (Lima et al. 2020).The MS 2 spectrum of the protonated molecule at m/z 296 presented an initial loss of 17 Da (NH 3 ) (m/z 296 → 279), followed by subsequent losses of 30 Da ( The ions at m/z 312 and 342 present in fractions F and C, respectively, were tentatively identified using fragmentation pattern analysis (Supplementary Material, Figure S9 and S10) and comparison with the literature.The MS 2 spectrum of the ion at m/z 312 presented an initial loss of 17 Da (m/z 312 → 295), followed by competitive losses of 15 Da (m/z 295 → 280) and 31 Da (m/z 295 → 244), consistent with the aporphine alkaloid O-methylisopiline [8] (Lima et al. 2020).Finally, at m/z 342, there was a high mass loss, commonly observed in tetrahydroprotoberberine alkaloids, with a base peak at m/z 178, which is in accordance with the structure of isocorypalmine [9] (Lima et al. 2020).

DISCUSSION
Cytotoxicity assessment prior to performing antiviral evaluations is important because it allows to determine that an observed reduction in virus infectivity is due to an effect of the samples on the virus, and not to a toxic effect on the host cells (Vlietinck et al. 1997;Schnitzler 2019).It also allows the determination of the maximum noncytotoxic concentration (MNTC) of a sample, which is used in antiviral assays (Pritchett et al. 2014;Schnitzler 2019).The MTT assay is one of the most widely exploited approaches for assessing cell viability and drug cytotoxicity (Stockert et al. 2018).
The F. longifolia TAF showed antiviral activity against both HSV-1 strains [KOS and 29R (acyclovir resistant)], as well as against the HSV-2 (strain 333) presenting SI values of 4, 3 and 5, respectively.The effect of TAF was considerably less potent then that of acyclovir against HSV-1 KOS and HSV-2 333, however, it showed antiherpes activity against HSV-1 29R, implying that the target of this fraction is probably different from those of acyclovir.
Although the TAF has shown activity against the HSV-1 strains that are sensitive (KOS) and resistant (29R) to acyclovir, we decided to study the antiviral mechanism of the TAF against HSV-2, which presented the highest selectivity index (SI) of 5.In antiviral activity studies, samples must be selective against the virus and not against the cells.The  ACTA AMAZONICA SI represents the relationship between the 50% cytotoxic concentration and the 50% effective concentration (Amoros et al. 1992), and is a widely accepted criterion used to express the in vitro efficacy of a sample in terms of inhibiting virus replication (Pritchett et al. 2014).Antiviral agents can be virucidal, causing an irreversible inactivation of the viral particle (virion) before it reaches the target cell, or they can interfere in the different stages of virus replication, from virus adsorption to the host cell until its release, or with virus binding, modifying cell membrane components, when incubated with host cells, before viral infection (Santos et al. 2013;Schnitzler 2019).Agents with virucidal capacity can damage the virion protein capsid, alter viral integrity, or penetrate into the virion and destroy the viral genome (Galabov 2007).Infections with HSV-2, and also with HSV-1, result in recurrent genital lesions (Looker et al. 2020).Thus, a strategy for the treatment and transmission prevention of these human viral infections is the development of topical formulations with compounds that can permanently inactivate viral particles (Ekblad et al. 2010;Mishra et al. 2018).
The synthesis of HSV viral gene products is carried out in three sequential phases, immediate (α), early (β) and late (γ).In the immediate phase, 2-4 hours after infection, the transcriptional machinery is activated and the proteins ICP4, ICP22, ICP27 and ICP47 are produced, which are fundamental for the production of all encoded viral proteins (Santos et al. 2013).In the early phase, which reaches its peak between 5-7 hours after infection, enzymes for viral replication are produced, e.g.thymidine kinase and viral DNA polymerase.In the late phase, until 12 hours after infection, are synthetized the structural components of the virions.The whole HSV cycle in cell lines takes about 18 to 20 hours (Santos et al. 2013).For these reasons, the prospection of the antiviral mechanism of action tested the TAF at different stages of the HSV-2 viral infection.
The pre-treatment of Vero E6 cells tested the effect of TAF on cells before they were infected with HSV-2, demonstrating no effect on virus replication, and therefore showing that the TAF has no in vitro prophylactic effect against HSV-2 infection.Yet, the TAF caused complete inactivation of HSV-2 infectivity at the lower EC 50 value (< 3.1 ± 18.9 µg.mL -1 and SI > 23), i.e., the TAF showed direct action on viral particles, and significantly decreased their infectious capacity, maintaining 76 ± 6% of HSV-2 inhibition at the lowest concentration tested (3.1 µg mL -1 ).The adsorption assay assessed whether the TAF was able to inhibit virus binding to the host cell membrane, one of the early steps of viral infection, with only 53.7 ± 14.1% of viral inhibition at the highest concentration tested (50 µg mL -1 ) and no effectiveness in other concentrations.
The time-of-addition assay, which evaluated the antiviral activity in different stages of HSV-2 replication, the TAF inhibited almost 100% HSV-2 replication between 0 and 12 h post-infection, decreasing thereafter.These findings suggested that, the alkaloid(s) present in the TAF can interfere in the three sequential phases of the HSV-2 infection: immediate, early and late.New studies are necessary to investigate the effect of TAF on HSV-2 gene expression in all these stages.Annonaceae alkaloids have already been described as having an antiviral activity (Padma et al. 1998;Yu et al. 2019).Of the aporphine alkaloids found in the TAF, isoboldine and nuciferine (Boustie et al. 1998) were active in vitro against the human poliovirus, with an SI > 14.Nuciferine was suggested as a prototype for the development of anti-HIV drugs (Kashiwada et al. 2005).Liriodenine, lysicamine and isoboldine were described as having antiviral activities against the HSV-1 virus strain H 29S, with SI of 6.9, 4.8 and 3.5, respectively (Montanha et al. 1995).
We showed that the TAF is active against the HSV-1 strain KOS (SI = 4), and the HSV-1 strain 29R (resistant to acyclovir) (SI = 3).This is the first report on the pharmacological activity of F. longifolia and the presence of the alkaloids liriodenine, lysicamine and isoboldine in its TAF, which are likely linked to its antiviral activity against the HSV-1 and HSV-2 viruses.Further studies should identify which alkaloids are involved in the antiviral activity and and their mechanisms of action and/or potential synergism.

CONCLUSIONS
The total alkaloid fraction (TAF) of Fusaea longifolia branches showed antiviral activity against the HSV-2 virus (strain 333) and the HSV-1 virus (strains KOS and 29R, the latter being acyclovir resistant).Against HSV-2, the TAF acted after the viral infection, with virucidal activity.Mass spectrometry identified the presence of nine alkaloids in the TAF, including liriodenine, lysicamine and isoboldine, which have already been described as having anti-HSV-1 activity.

Figure 1 .
Figure 1.Effect of time-of-addition of the total alkaloid fraction (TAF) from Fusaea longifolia (50.0 μg mL -1 ) on HSV-2 (333 strain) replication determined by plaque reduction assay.Results are plotted as % of viral inhibition in comparison to the untreated controls.Columns are the mean and bars the standard deviation of three independent experiments performed in duplicate.

Figure 2 .
Figure 2. Structures of oxoaporphine and aporphine alkaloids present in the fractions A-G of Fusaea longifolia extract.Credit: F.M.A. Silva.

Table 2 .
Mass spectrometry data of the alkaloids tentatively identified in fractions A-G of the Fusaea longifolia branch extract.Class: Oxoaporphine (O), aporphine (A) or tetrahydroprotoberberine (T) alkaloid.MS/MS: Main product ions obtained in QqQ-MS/MS analysis.