Antiviral activity of Bignoniaceae species occurring in the State of Minas Gerais (Brazil): part 1

Authors

  • G.C. Brandão,

    1.  Laboratório de Fitoquímica, Departamento de Produtos Farmacêuticos, Faculdade de Farmácia, Universidade Federal de Minas Gerais – UFMG, Belo Horizonte MG, Brazil
    Search for more papers by this author
  • E.G. Kroon,

    1.  Laboratório de Vírus, Departamento de Microbiologia – ICB, Universidade Federal de Minas Gerais – UFMG, Belo Horizonte MG, Brazil
    Search for more papers by this author
  • J.R. Dos Santos,

    1.  Laboratório de Vírus, Departamento de Microbiologia – ICB, Universidade Federal de Minas Gerais – UFMG, Belo Horizonte MG, Brazil
    Search for more papers by this author
  • J.R. Stehmann,

    1.  Laboratório de Taxonomia Vegetal, Departamento de Botânica – ICB, Universidade Federal de Minas Gerais – UFMG, Belo Horizonte MG, Brazil
    Search for more papers by this author
  • J.A. Lombardi,

    1.  Departamento de Botânica,Instituto de Biociências de Rio Claro, Universidade Estadual Paulista – UNESP, Rio Claro, SP, Brazil
    Search for more papers by this author
  • A. Braga de Oliveira

    1.  Laboratório de Fitoquímica, Departamento de Produtos Farmacêuticos, Faculdade de Farmácia, Universidade Federal de Minas Gerais – UFMG, Belo Horizonte MG, Brazil
    Search for more papers by this author
    • Part of G. C. Brandão’s Doctorate Thesis at the Programa de Pós-Graduação em Ciências Farmacêuticas (PPGCF), Faculdade de Farmácia/UFMG, Belo Horizonte, MG, Brazil.


Alaide Braga de Oliveira, Laboratório de Fitoquímica, Departamento de Produtos Farmacêuticos, Faculdade de Farmácia, Universidade Federal de Minas Gerais – UFMG, Av. Antônio Carlos, 6627, CEP 31270-901 Belo Horizonte MG, Brazil. E-mails: alaide.braga@pq.cnpq.br; alaidebraga@terra.com.br

Abstract

Aims:  To evaluate the antiviral activity of Bignoniaceae species occurring in the state of Minas Gerais, Brazil.

Methods and Results:  Ethanol extracts of different anatomical parts of bignoniaceous plant species have been evaluated in vitro against human herpesvirus type 1 (HSV-1), vaccinia virus (VACV) and murine encephalomyocarditis virus (EMCV) by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. A total of 34 extracts from 18 plant species selected according to ethnopharmacological and taxonomic criteria were screened. Fifteen of the 34 extracts (44·1%) have disclosed antiviral activity against one or more of the viruses assayed with EC50 values in the range of 23·2 ± 2·5–422·7 ± 10·9 μg ml−1.

Conclusions:  Twelve of the 34 extracts (35·3%) might be considered promising sources of antiviral natural products, as they have shown EC50 ≤ 100 μg ml−1. The present screening discloses the high potential of the Bignoniaceae family as source of antiviral agents.

Significance and Impact of the Study:  Active extracts were identified and deserve bioguided studies for the isolation of antiviral compounds and studies on mechanism of action.

Introduction

There is a need to develop more effective and safe antiviral agents to overcome viral resistance and problems of toxicity to clinically available anti-infective drugs, and a multidisciplinary investigation into plant-derived natural products is a valid strategy. Indeed, a survey of natural products as source of new drugs has disclosed its utility in the discovery and development process of different classes of new therapeutic agents for the treatment of human diseases (Newman and Cragg 2007).

A significative number of medicinal plant extracts is reported to exhibit high activity against herpes simplex viruses (Jassim and Naji 2003), and a great structural variety of natural products has been identified as inhibitors of several pathogenic viruses (Chattopadhyay and Naik 2007; Chattopadhyay and Khan 2008)). Furthermore, hydromethanolic extracts from 54 Brazilian medicinal plants used in folk medicine to treat infections were screened for antiviral properties against five different viruses of human herpesvirus (HSV-1, HSV-2, poliovirus type 2, adenovirus type 2 and VSV (vesicular stomatitis virus)). Fifty-two per cent of the plant extracts exhibited antiviral activity against one or more tested viruses (Simões et al. 1999). These results provide evidences on the validity of ethnopharmacological screenings of Brazilian plants for antiviral activity.

In the present study, plants belonging to the Bignoniaceae family were selected on the basis of ethnopharmacological and taxonomic criteria, which are valid exploited strategies for a screening programme searching for bioactive natural products (Verpoorte 1998). This report concerns the preparation of extracts and their in vitro evaluation for antiviral activity. The colorimetric [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] MTT assay was used to evaluate the cytotoxicity of the extracts in Vero cells and the susceptibility of a RNA virus, murine encephalomyocarditis virus (EMCV, Picornaviridae) and two DNA viruses, human herpesvirus type1 (HSV-1, Herpesviridae) and vaccinia virus (VACV, Poxviridae).

Human herpes viruses (HSV-1 and HSV-2) are pathogenic to humans, causing recurrent infections, especially in the case of highly susceptible adults. Acyclovir is the most commonly used drug for the treatment of HSV infections although drug resistant strains have frequently arose following therapeutic treatment with this drug and other nucleosides (De Clercq et al. 2001).

VACV has a wide host range and infects mammalian cells of many different species. Re-emergence of human VACV infections (Trindade et al. 2007a,b) as well as the threat that variola virus, the aetiological agent of smallpox, might be used in warfare or terrorism, has motivated the search for measures to control or treat smallpox and poxvirus infections, in general. VACV inhibitors represent a paradigm for the chemotherapy of poxvirus infections (De Clercq 2001).

EMCV is a group of closely related virus species with a wide host range. EMCV infections are associated with sporadic cases and outbreaks of myocarditis and encephalitis in domestic pigs, in nonhuman primates and in other mammalian species. Few cases of human EMCV infections have been documented, the most recent were in 2004 from two febrile patients in Peru (Oberste et al. 2009). EMCV is used as a model for RNA viruses, especially from the Picornaviridae family, as it represents a safe animal model to test antiviral drugs (Mujtaba et al. 2006).

Materials and methods

Plant material

Plant material was collected in the state of Minas Gerais, Brazil. Voucher species are deposited into the herbarium of UFMG (BHCB), Belo Horizonte, Brazil, and was dried in an air-circulating oven at 45°C for 48 h. The different anatomical plant parts were separated, ground and exhaustively extracted by percolation with ethanol 92·8° GL. The percolates from each plant species were combined, and the solvent was completely removed under reduced pressure in a rotavapor. All the extracts were chemically characterized by thin-layer chromatography (TLC) on silica gel plates, using appropriate eluents and chromogenic reagents for different classes of natural products (Wagner et al. 1984) as well as by high performance liquid chromatography coupled to a diode array detector (HPLC-DAD), with online registration of the UV spectra of the constituents (data not shown).

Cell culture and virus

Vero cells (ATCC CCL-81) were cultured in Dulbecco’s modified Eagle’s medium (DMEM), at 37°C, in 5% CO2 atmosphere, supplemented with 5% FBS, in the presence of 50 μg ml−1 gentamicin, 100 U ml−1 penicillin and 5 μg ml−1 fungizone. HSV-1 was a clinical isolate obtained from the Laboratory of Virus, UFMG, Belo Horizonte, Brazil. EMCV and VACV Western Reserve strains were kindly donated by Dr I. Kerr (London Research Institute, UK) and Dr C. Jungwirth (University of Würzburg, Germany), respectively.

The propagation of HSV-1 and VACV was performed on Vero cells, while L929 cells were used for EMCV for better viral titre. Stock viruses were prepared from supernatants of infected cells and stored at −70°C. The virus titre was estimated from cytopathogenicity and expressed as 50% tissue culture infective dose per millilitre (TCID50 ml−1), i.e. the virus dilution required to infect 50% of inoculated cell cultures (Reed and Muench 1938). However, for all the antiviral experiments, the multiplication of the three viruses was performed in Vero cells.

Cytotoxicity assay

Vero cells were exposed to different concentrations of extracts for 48 and 72 h. After incubation, cell viability was assessed by the MTT assay (Merck solution 2 mg ml−1 in PBS) (Twentyman and Luscombe 1987). Each sample was assayed in four replicates for concentrations ranging from 500 to 0·125 μg ml−1. The cytotoxicity of each extract was expressed as CC50, i.e. the concentration of extract that reduced cell growth by 50%.

Antiviral assay

Vero cells were seeded on 96 well plates. Extracts were dissolved in DMSO, and twofold dilutions in MEM concentrations were prepared. Virus suspensions and four different concentrations of the extracts, at noncytotoxic concentrations, were simultaneously added to the plates. Controls for cytotoxicity (uninfected treated cells), cells (uninfected untreated cells), virus (infected untreated cells) and positive controls (acyclovir/Calbiochem and α-2a interferon/Bergamo) were run in parallel during each experiment. The antiviral activity was evaluated by the MTT colorimetric assay (Betancur-Galvis et al. 1999) and is expressed as EC50 that is the concentration of extract required to inhibit the cytopathic effect to 50% of the control value.

Statistical analysis

The mean values ± standard deviations are representative of four independent experiments. For the determination of CC50 and EC50 values, nonlinear regressions of concentration–response curves were used. Statistical analysis of the data was carried out by the Graph-Pad Prisma V.3 program (GraphPad Software Inc., San Diego, CA).

Selectivity index

The selectivity index (SI) is calculated as the ratio between the 50% cytotoxic concentration (CC50) of a compound or extract and the 50% effective antiviral concentration (EC50) (SI = CC50/EC50).

Results

A total of 34 ethanol extracts from 18 Bignoniaceous plant species were assayed for their antiviral activity against HSV-1, VACV and EMCV. Data on the collected plant species, their voucher numbers at the BHCB herbarium (UFMG) and the yields of ethanol extractives are shown in Table 1. In this table, plants that were selected according to documented ethnopharmacological informations are signalled by A, while B denotes those that were taxonomically selected.

Table 1.   Bignoniaceae species assayed for antiviral activity, voucher numbers and ethanol extractives
Extract numberFamily/species (selection method)Voucher numberPlant partExtractives (%)
  1. A, ethnopharmacological selection; B, taxonomic selection.

  2. *synonymy; extractives: % yield of dry ethanol extract (w) in relation to plant material (w).

 Bignoniaceae
 1Adenocalymma trifoliatum Laroche (B)BHCB 24763Leaves23·7
 2Anemopaegma chamberlaynii (Sims) Bureau & K. Schum. (B)BHCB 104332Leaves8·0
 3  Stems12·0
 4Anemopaegma setilobum A.H. Gentry (B)BHCB 88946Leaves16·0
 5  Stems17·0
 6Arrabidaea brachypoda (DC.) Bureau (B)BHCB 23402Leaves28·2
 7(Fridericia platyphylla (Cham.) L.G. Lohmann)* Stems and fruits14·0
 8Arrabidaea craterophora (DC.) Bureau (B)BHCB 23404Leaves13·9
 9(Fridericia craterophora (DC.) L.G. Lohmann)* Stems10·0
10Arrabidaea formosa (Bureau) Sandwith (B)BHCB 23885Leaves25·1
11(Fridericia formosa (Bureau) L.G. Lohmann)* Stems13·0
12  Fruits18·1
13Arrabidaea pulchra(Cham.) Sandwith (B) (Cuspidaria pulchra (Cham.) L.G. Lohmann)*BHCB 21573Leaves24·5
14Arrabidaea sceptrum (Cham.) Sandwith (A)BHCB 21879Leaves26·9
15(Cuspidaria sceptrum (Cham.) L. Lohmann)* Stems16·2
16Macfadyena unguis-cati (L.) A.H.Gentry (A)BHCB 21871Leaves12·2
17  Stems8·0
18Manaosella cordifolia (DC.) A.H. Gentry (B)BHCB 45727Leaves30·0
19  Stems12·0
20Mussatia prieurei (DC.) Bureau ex K. Schum. (B)BHCB 21857Leaves27·3
21  Stems17·3
22Pyrostegia venusta Miers (A)BHCB 22486Leaves15·2
23  Stems7·3
24Sparattospema leucanthum (Vell.) K. Schum. (A)BHCB 24756Leaves22·0
25Stizophyllum perforatum Miers (B)BHCB 104334Leaves8·2
26  Stems10·0
27Tabebuia aurea (Silva Manso) Benth. & Hook. f. ex S. Moore (A)BHCB 111069Stems12·3
28Tabebuia cassinoides (Lam.) DC. (A)BHCB 112989Leaves10·0
29  Stems10·0
30Zeyheria montana (Vell.) Mart. (A) (Zeyheria digitalis (Vell). Hoehne)BHCB 22609Leaves21·0
31  Stems6·1
32Zeyheria tuberculosa (Vell.) Bureau ex Verl. (A)BHCB 27265Leaves20·0
33  Fruits23,5
34  Stems11·0

The cytotoxicity to Vero cells and the in vitro antiviral assays were carried out by the MTT method. Plant extracts, in concentrations ranging from 500 to 0·125 μg ml−1, were first screened for their cytotoxicity to Vero cells (Table 2). Pronounced cytotoxicity (CC50 < 50 μg ml−1) was observed for leaf extracts of 4 of the 34 extracts tested: Arrabidaea brachypoda, Arrabidaea sceptrum, Sparattosperma leucanthum and Tabebuia cassinoides. For 12 of the 34 extracts assayed, the CC50 values ranged from 120·8 ± 1·0 to 387·0 ± 6·9 μg ml−1, four of the extracts have shown CC50 > 200 μg ml−1, and no cytotoxicity was observed up to 500 μg ml−1 for 14 of them.

Table 2.   Antiviral activity (EC50) of ethanol extracts from Bignoniaceae species against HHV-1, EMCV, VACV, cytotoxicity on Vero cells (CC50) and selectivity index (SI)
Extract numberSpeciesPlant partVero cells CC50 (μg ml−1)HHV-1 EC50 (μg ml−1)SIVACV-WR EC50 (μg ml−1)SIEMCV EC50 (μg ml−1)SI
  1. A, ethnopharmacological selection; B, taxonomic selection; NA, not active; EMCV, encephalomyocarditis virus; VAVC, vaccinia virus.

  2. *40 μg ml−1 with 70–90% cytopathic effect inhibition.

 1Adenocalymma trifoliatum (B)Leaves>500NA NA NA 
 2Anemopaegma chamberlaynii (B)Leaves165·1 ± 2·6NA NA NA 
 3 Stems387·0 ± 6·9NA NA NA 
 4Anemopaegma setilobum (B)Leaves>200138·1 ± 9·3>1·477·5 ± 4·7>2·6NA 
 5 Stems>500168·2 ± 12·8>3·0115·7 ±12·7>4·395·2 ± 8·4>5·2
 6Arrabidaea brachypoda (B)Leaves27·9 ± 3·9NA NA NA 
 7 Stems +Fruits>500NA 92·3 ± 2·2>5·4290·2 ± 22·3>1·7
 8Arrabidaea craterophora (B)Leaves297·3 ± 14·8NA NA 68·2 ± 10·74·4
 9 Stems355·5 ± 0·7188·1 ± 21·01·9NA NA 
10Arrabidaea formosa (B)Leaves>50082·2 ± 5·7>6·185·9 ± 2·2>5·8195·1 ± 25·2>2·6
11 Stems>50090·2 ± 6·7>5·558·54 ± 1·5>8·5322·5 ± 14·4>1·5
12 Fruits>500148·5 ± 1·7>3·4253·8 ± 5·5>2·0138·8 ± 10·3>3·6
13Arrabidaea pulchra (B)Leaves>500232·1 ± 22·5>2·2248·9 ± 29·1>2·0NA 
14Arrabidaea sceptrum (A)Leaves39·7 ± 5·4NA NA NA 
15 Stems>500375·3 ± 13·5>1·340·6 ± 2·1>12·3NA 
16Macfadyena unguis-cati (A)Leaves>500NA NA NA 
17 Stems>500NA NA NA 
18Manaosella cordifolia (B)Leaves203·6 ± 6·2NA NA NA 
19 Stems384·0 ± 7·6NA NA NA 
20Mussatia prieurei(B)Leaves>500NA 422·7 ± 10·9>1·2NA 
21 Stems>500NA 258·9 ± 23·7>1·9NA 
22Pyrostegia venusta (A)Leaves>200NA NA NA 
23 Stems>200NA NA NA 
24Sparattosperma leucanthum (A)Leaves22·5 ± 0·3NA NA NA 
25Stizophyllum perforatum (B)Leaves>500NA NA 331·3 ± 2·5>1·5
26 Stems>500338·7 ± 21·1>1·554·4 ± 2·3>9·2NA 
27Tabebuia aurea (A)Stems>200NA NA NA 
28Tabebuia cassinoides (A)Leaves29·9 ± 1·1NA NA NA 
29 Stems333·1 ± 10·6NA NA NA 
30Zeyheria montana (A)Leaves120·8 ± 1·0NA NA NA 
31 Stems320·7 ± 8·2NA NA NA 
32Zeyheria tuberculosa (A)Leaves168·9 ± 2·481·8 ± 7·22·123·2 ± 2·57·3NA 
33 Stems241·1 ± 19·5NA NA NA 
34 Fruits272·5 ± 9·8NA NA NA 
 Acyclovir  40*     
 α-2a-Interferon    2·5 × 102 1·5 × 102 

The extracts were evaluated for their effect against HSV-1, VACV and EMCV, in four noncytotoxic concentrations. Fifteen of the 34 extracts (41·2%) have shown some activity against one or more of the three viruses assayed, and the EC50 values ranged from 23·2 ± 2·5 to 422·7 ± 10·9 μg ml−1. In terms of the botanical species assayed, the rate of positive results is 50% (9 of the 18 species). The following parameters are used to ascribing the antiviral extracts effects: for EC50 ≤ 50 μg ml−1, the sample is considered active, 50 ≥ EC50 ≥ 100 μg ml−1 corresponds to a moderate activity, low activity is related to EC50 > 100 μg ml−1, while inactivity (NA) corresponds to no observed changes in the level of cytopathic effects for the four tested product concentrations (Table 2).

In the assays against HSV-1, 10 of the 34 extracts have shown antiviral activity (29·4%) (Table 2). The three more active extracts were obtained from Arrabidaea formosa (leaves and stems) and Zeyheria tuberculosa (leaves), which disclosed moderate activity. Low activity (EC50 from 138·1 ± 9·3 to 375·3 ± 13·5 μg ml−1) was observed for seven of the ten active extracts: Anemopaegma setilobum leaves and stems, Arrabidaea craterophora stems, A. formosa fruits, Arrabidaea pulchra leaves, A. sceptrum stems and Stizophyllum perforatum stems. The 10 HSV-1 active extracts were obtained from seven Bignoniaceae species and, therefore, in terms of botanical species, the rate of positive results is 38·9%.

Zeyheria tuberculosa leaves and A. sceptrum stems were the most active extracts against VACV disclosing EC50 of 23·2 ± 2·5 and 40·6 ± 2·1 μg ml−1, respectively. Ten of the 34 extracts have shown moderate to low anti-VACV activity with EC50 values in the range of 54·4 ± 2·3 to 422·7 ± 10·9 μg ml−1. Therefore, the rate of active extracts was 35·3% (12 of 34 extracts) and 44·4% for active species (8 of 18). (Table 2).

Only two extracts have disclosed moderate activity against EMCV: A. craterophora leaves (68·2 ± 10·7 μg ml−1) and Anemopaegma setilobum leaves (95·2 ± 8·4 μg ml−1). Five extracts have shown low activity (EC50 values: 138·8 ± 10·3 to 331·3 ± 2·5 μg ml−1) that were obtained from A. brachypoda (stems + fruits), A. formosa (leaves, stems and fruits) and Stizophyllum perforatum (leaves). EMCV was the less susceptible of the three viruses assayed with positive results for 7 of the 34 extracts (20·6%) (Table 2).

Eight of the 18 Bignoniaceae species assayed are reputed to medicinal use for treating symptoms that could be associated with virus infections and are signalized as A, while the remaining 10 species (B) were selected because either they are closely related to the ethnopharmacological ones or just because they belong to the Bignoniaceae family. Antiviral activity has been observed for only two of the eight ethnopharmacologically selected species and seven were from the ten taxonomically selected group. The SI values calculated for the 15 active extracts were in the range of 1·4–12·3 and are shown in Table 2.

Discussion

In the present screening for antiviral activity, we have identified nine active plant species belonging to the Bignoniaceae, a family that occurs widely in South America and is of interest because of its ethnobotany and economic botany, including the popular and/or traditional uses for the treatment of different diseases (Gentry 1992). Chemically, bignoniaceous plants are characterized by the presence of flavonoids, terpenoids, quinones, mainly naphthoquinones, lignans and simple aromatic compounds (Cipriani et al. 2007). Naphthoquinones occur in several species of the family (Rao and Kingston 1982; Oliveira et al. 1990) and disclose different biological effects, including antiviral, anticancer, cytotoxic, anti-inflammatory, antimicrobial, antifungal, antimalarial and antiparasitic activities (Gómez-Catellanos et al. 2009).

In general, in vitro antiviral assays are based on the inhibition of the cytopathic effect in a virus cell culture. In these assays, the activity is expressed by the 50% end point titration technique (Vlietinck et al. 1995; Cos et al. 2006). However, over the last two decades, a colorimetric assay, in which the MTT, an yellow dye, is reduced by viable cells to a formazan (purple) has been frequently used. This assay is semi-automated, rapid and requires only a small amount of test sample and directly assesses cell viability (Takeuchi et al. 1991; Betancur-Galvis et al. 1999). In the present screening, the cytotoxicity and the in vitro antiviral assays were carried out by the MTT method.

The pronounced cytotoxicity observed on Vero cells for the leaves extracts of 4 of the 34 extracts which were tested (A. brachypoda, A. sceptrum, S. leucanthum and T. cassinoides) might be related to cytotoxic naphthoquinones, that have been isolated from T. cassinoides (Rao and Kingston 1982), and/or to flavonoids occurring in A. brachypoda. Antifungal activity has been reported for A. brachypoda flavonoids (Alcerito et al. 2002) that have been recently reported to disclose enhanced cytotoxicity on human leukaemia cells (Plochmann et al. 2007). To date, there are no chemical or biological data on S. leucanthum and A. sceptrum.

Five of the nine Bignoniaceae species, which have disclosed antiviral activity, belong to the genus Arrabidaea (A. brachypoda, A. craterophora, A. formosa, A. pulchra and A. sceptrum). The four remaining species are from the genera Anemopaegma (Anemopaegma setilobum), Mussatia (Mussatia prieurei), Stizophyllum (Stizophyllum perforatum) and Zeyheria (Z. tuberculosa). Only two of these nine species, A. brachypoda and Z. tuberculosa, have been previously investigated. A. brachypoda afforded antifungal and cytotoxic flavonoids (Alcerito et al. 2002; Plochmann et al. 2007). Antimicrobial flavonoids are described for Z. tuberculosa (Bastos et al. 2009) besides cytotoxic naphthoquinones (Oliveira et al. 1990). This is the first report on the antiviral activity of these two species, while no report on chemistry and/or biological effects was found for the remaining seven species. Two of the 15 active extracts (Z. tuberculosa leaves and A. sceptrum stems) have been the only ones to disclosing EC50 < 50 μg ml−1 (23·2 ± 2·5 and 40·6 ± 2·1 μg ml−1, respectively), both of them against VACV. This is a relevant result once there is a need to develop anti-vaccinia drugs.

Along with flavonoids, triterpenes were detected by TLC in the nine active species evaluated. Xanthones were shown, by HPLC-DAD, to occur in Anemopaegma setilobum, A. formosa and Stizophyllum perforatum (data not shown), and naphthoquinones were detected only in Z. tuberculosa. Representatives of these four classes of natural products have been reported as inhibitors of several viruses (Lagrota et al. 1983; Perez-Sacau et al. 2003; Chattopadhyay and Naik 2007) and they could be responsible for the effects of the active extracts in which they were detected.

When considering the total of assays carried on that has been 102 (34 extracts against three viruses), close results have been observed when the viral inactivation effectiveness of extracts from leaves and stems is compared, 11·8% (12 of 102 assays) and 13·7% (14 of 102 assays), respectively, which is coherent with their chemical similarity in terms of the principal components, triterpenes and flavonoids (data not shown).

Concerning the SI, highly selective drugs should show SI > 10, while reasonably potent ones would have 3 < SI < 10. Furthermore, in terms of antiviral efficacy, SI is considered the most important parameter (De Clercq et al. 2001). Reasonable selectivity indices (SI > 4) were displayed by four of the ten active extracts against HSV-1, seven of 12 with anti-VACV activity and three of the seven active against EMCV. The highest SI values have been calculated for A. sceptrum stems (SI = 12·3) and Stizophyllum perforatum stems (SI > 9·2), both of them against VACV. Therefore, these extracts deserve further investigations because of the re-emergence of vaccinia human infections and its potential use as a biological warfare or for terrorism (De Clercq 2001; Trindade et al. 2007a,b).

In general, the plant extracts have shown different profiles towards the viruses tested. Only four extracts, Anemopaegma setilobum (stems) and A. formosa (leaves, stems and fruits), have inhibited the multiplication of the three viruses assayed and they have shown moderate to low activity (Table 2). The SI values for the corresponding 12 assays are favourable (SI > 3) although for three of them, the values could not be precisely calculated because the exact CC50 was not determined. No report was found on the chemistry or biological effects of both of these two species whose antiviral activity is also reported for the first time.

The differences on the susceptibilities of the viruses assayed towards the extracts evaluated might be related to their structure, as the presence of envelope or type of receptor-binding proteins involved in the adsorption to the cell. The two DNA viruses (HSV-1 and VACV) are quite big enveloped viruses; meanwhile, the most resistant virus in this study, EMCV, is a quite small nonenveloped RNA virus. Besides the difference in the nuclei acid composition, the three viruses disclose distinct multiplication cycles. Therefore, different characteristics of these three viruses probably play important role in viral drug susceptibility and will certainly reflect also on the mechanisms of action of the extracts/constituents, which deserve to be further investigated. It is interesting to notice that more positive results have been observed for the pair of the two DNA viruses (HSV-1 and VACV) what seems to be consequence of the differences that were pointed out.

When considering the total of assays carried on that has been 102 (34 extracts against three viruses), close results have been observed when the viral inactivation effectiveness of extracts from leaves and stems is compared: 11·8% (12 of 102 assays) and 13·7% (14 of 102), respectively.

It is well known that the virus titre of a culture can influence the EC50 in the in vitro antiviral assays. It is considered that titres up to 105 TCID50 ml−1 markedly increases the sensitivity of the test system (Cos et al. 2006). In the present screening, assays were run against virus cultures with 2·5 × 106 TCID50 ml−1 (HSV-1) and 1·0 × 106 TCID50 ml−1 (VACV and EMCV). Therefore, the positive results may be considered strong proofs of activity. Besides, it might be expected that lower EC50 values would be expected if the experiments had been carried on with lower viral cultures titres.

In conclusion, the present screening is the first to demonstrate the in vitro antiviral activity of these 18 plant species, pointing out the potential of Bignoniaceae taxa as sources of antiviral agents. The high rate of positive results for taxonomically than ethnopharmacologically selected species is quite surprising and supports the validity of combining these two approaches for selecting plants to be screened for antiviral activity, at least for taxa belonging to the Bignoniaceae family.

Acknowledgements

To CNPq (Brazil) for financial support and research fellowships (A.B.O. and E.G.K.) and to FAPEMIG (Brazil) for a Doctorate fellowship (G.C.B.).

Ancillary