Collection Sites

Figure 1

Figure 1. Locations of study areas, Minas Gerais state, Brazil. A) Locations of the 3 municipalities where collections were performed: Sabará, Serro, and Rio Pomba. Inset shows location of Minas Gerais state in...

We collected small mammals in 3 areas of Minas Gerais. Brazil was chosen because of its history of BV outbreaks and its different biomes and conservational status (Figure 1, panel A). The 3 municipalities where collections were performed were Sabará, Serro, and Rio Pomba.

Sabará is a city located in an anthropic area situated in the transition from savannah (the Cerrado biome) to the Atlantic Forest. The study site (19°53′9′′S, 43°48′45′′W) was delimited on the grounds of a former educational institution in a previous study (Figure 1, panel B). Three sampling transects were demarcated: 1 in savannah vegetation with intense anthropogenic disturbance and the other 2 in forest vegetation (with 1 of the 2 having more disturbance than the other) (Figure 1, panel B). In each transect, 15 sampling points were established with 2 live traps in each, 20 m apart, where captures of small mammal took place.

Figure 2

Figure 2. Location of collection sites and biomes represented in each, Minas Gerais state, Brazil. A) Collection site 1 in Serro. B) Collection site 2 in Serro. C) Peridomicile collection areas in Rio...

In Rio Pomba, the field site (21°16′29′′S, 43°10′45′′W) has characteristic Atlantic Forest vegetation. Animal trapping was performed in the area around the Instituto Federal de Educação, Ciência e Tecnologia (Figure 2, panels A, B). Animal collections were performed in forest, pasture, and peridomicile areas (Figure 2, panels A–G). In each transect, 10 traps were placed at a distance of 10 m from each other and in alternating positions (on forest floor or on tree trunks).

Serro, a city whose economy is based on milk and cheese production, has seen many cases of BV since 2005 (23). The capture of small mammals was carried out in 2 farms (Figure 2, panels C, D). The study site (180°36′21.16′′S, 43°23′12.89′′W) is situated in the Cerrado biome and has some intersections of Atlantic Forest. Animal collections were carried out in forest, pasture, and peridomicile areas (Figure 2, panels A–G), and traps were positioned in the same manner as in Rio Pomba.

Animal Trap and Sample Collection

Captures lasted from April 2011 through May 2012 for Sabará (12 campaigns) and from September 2012 through September 2013 in Serro (5 campaigns) and Rio Pomba (6 campaigns). Small mammals were captured in size-selective live cages by using pineapple chunks and cotton balls soaked in cod liver oil as baits. Each sampling section lasted for 4 nights, and baits were replaced after 2 nights. After capture, animals were anesthetized with ketamine (70 mg/kg) and xylazine (12 mg/kg) for serum collection. Animals were weighed, measured for size, and visually evaluated for clinical signs of disease, such as skin lesions. For organ collection, animals were euthanized by intracardiac injection of 3 times the anesthetic dose according to guidelines of the American Society of Mammalogists (24). Collections were authorized by the Environment Ministry of Brazil through the SISBIO system (license no. 20807–2).

Biosafety

All collections were performed by trained professionals (either veterinarians or biologists) according to US Centers for Disease Control and Prevention recommendations (25). During animal manipulation, personal protective equipment (disposable coveralls, surgical gloves, goggles, and N98 masks) was used.

DNA Extraction from Organs

In addition to serum, which was tested by real-time PCR (rPCR) targeting the C11R viral growth factor gene without DNA extraction, liver was the chosen organ for rPCR trials. The organs were macerated with mortar and pestle after liquid nitrogen was added, and DNA was extracted with PureLink Genomic DNA Mini Kit (Invitrogen, Carlsbad, CA, USA) as recommended by the manufacturer. The same protocol was applied for other organs tested, including intestine, bladder, heart, gonads (ovary/testicles), bone marrow, spleen, lung, diaphragm, and kidney.

Cells and Virus

A VACV Western Reserve strain was used as a positive control. BSC-40 cells were grown in Eagle’s minimum essential medium (Invitrogen) supplemented with 5% fetal bovine serum (Cultilab, São Paulo, Brazil); 25 mg/mL Fungizone (amphotericin B) (Cristália, São Paulo, Brazil); 500 U/mL penicillin; and 50 mg/mL gentamicin (Schering-Plough, São Paulo, Brazil).

rPCR Assays

All rPCR experiments were performed in 48-well plates in Step One machines (Applied Biosystems, Foster City, CA, USA) by using SYBR Green Master Mix (Applied Biosystems). DNA from liver samples was diluted in water for a final use concentration of 10 ng/μL and 50 ng/μL. For serum samples, a 1:10 or 1:100 dilution was performed, and samples were tested without previous DNA extraction. For both liver and serum samples, amplification of the C11R gene was tested, and liver samples were additionally tested for amplification of the A56R hemagglutinin gene (primer sequences available upon request). For C11R, an amplicon of 82 bp and a melting temperature of 74.99°C were expected, and for A56R, a sequence of 160 bp and a melting temperature of 74.41°C were expected. All reactions had a final volume of 10 μL, and samples were tested in duplicates. Reaction steps comprised initial DNA denaturation at 95°C for 10 min, 40 cycles of denaturation (95°C for 15 s), annealing/extension (60°C for 60 s), and a melting curve (95°C for 15 s, 60°C for 60 s, and 95°C for 15 s). Samples were considered positive when melting temperature varied only up to 1°C compared with a positive control (10 ng of DNA extracted from purified VACV Western Reserve strain) and had amplification in duplicate or for >1 target. Samples with a single amplification were retested and considered equivocal when no more amplification was observed.

Nucleotide Sequencing and Sequence Analyses

Positive samples that could be reamplified (A56R-positive) or amplified by a conventional PCR targeting C11R (C11R-positive) (26) were chosen for sequencing. For A56R, product from the previous reaction was reamplified in a conventional PCR reaction using 1 μL of the first reaction as input and 0.2 nmol/L (A56R) rPCR primers. PCR cycling for A56R gene consisted of 10 min at 95°C for denaturation, 30 cycles of denaturation (95°C for 10 min), annealing (60°C for 60 s), extension (72°C for 60 s), and a final extension of 10 min at 72°C. Products with single bands were directly sequenced, and products with multiple bands had the target gene extracted from acrylamide gels stained with SYBR Gold Nucleic Acid Gel Stain (Invitrogen) and had its DNA purified. Nucleotide sequencing was performed by dideoxy method in an ABI3130 platform (Applied Biosystems), and sequence quality was analyzed by using Sequence Scanner Software 1.0 (Applied Biosystems). Sequences were aligned with other reference sequences from the BLAST nucleotide database (http://blast.ncbi.nlm.nih.gov/Blast.cgi) by using MEGA 6.0; the same program was used for identity matrix construction (27).

Plaque-Reduction Neutralization Test

The plaque-reduction neutralization test (PRNT) protocol has been described previously by Geessien Kroon et al. (28). Samples were considered positive when a reduction of >50% in virus plaque numbers was observed.

ELISA

ELISA was performed for rodent blood samples following a protocol also described previously (28). For each plate, 1 positive control (serum from Mus musculus experimentally infected with VACV-Guarani P1) (29) and 3 negative controls (serum of noninfected M. musculus) were added. Cutoff was established as the mean of negative controls optical density units plus 3 times their SD. Samples with an optical density 10% above or below the cutoff were considered equivocal.

Interaction Networks

Interaction networks are useful to help with understanding of virus-host dynamics. Each species is represented by a vertex, and the link between 2 vertices represents the interaction between 2 different species, making it possible to analyze their interdependence (30). The networks also show which species have the higher number of positive samples for virus detection and the area where they were collected. By using data from VACV-positive small mammal species, we created weighted networks with the program Pajek 4.07 (31). Accordingly, adjacency matrices were generated for each study area in which hosts were represented by lines and VACV by columns.

rPCR Amplification of VACV DNA from Free-living Small Mammals

A total of 325 animals had their samples tested by rPCR targeting the C11R gene, the A56R gene, or both. Of these animals, 21 (6.4%) tested positive (i.e., amplification in duplicates or in >1 sample/target) and 58 (17.8%) equivocal (Technical Appendix Table 1). The cycle thresholds varied from 28.42 to 39.33. From the total animals tested by rPCR, 114 had samples available for all tests (C11R targeted in liver and serum and A56R in liver). One animal was positive in the 3 tests performed, 11 in 2 tests, and 5 in only 1 test (data not shown). For the remaining positive animals, >1 assays could not be performed, and 1 sample type was positive in 1 test.

Of all the animals from the different collection sites in Sabará, 11/48 (22.9%) rodents and 3/76 (3.9%) marsupials were positive by rPCR. For Serro, no marsupials were positive but 4/25 (16.0%) rodents were positive. For Rio Pomba, 2/137 (1.4%) of rodents and 1/18 (5.5%) of marsupials were positive (Technical Appendix Table 1).

Four rPCR-positive animals (2 rodents and 2 marsupials) were selected for viral DNA detection in different organs by rPCR targeting the C11R and A56R genes. Positivity was found for heart, spleen, intestines, bladder, lungs, kidneys, and gonads (Technical Appendix Table 2). No amplification was observed in any bone marrow or diaphragm samples tested.

OPV Antibodies in Serum from Free-living Small Mammals Tested by PRNT

PRNT tests were performed in a total of 314 serum samples, and from these, 33 were considered positive, corresponding to 10.5% of the animals. The reduction percentages varied from 50.5% to 95.6%. For the Sabará collection, positivity was 9.0% (10/111), 14.3% (6/42) for rodents and 5.8% (4/69) for marsupials. For Serro, positivity was 4.2% (2/47), and only rodent samples were positive, corresponding to 8.0% (2/25) of rodents tested. For Rio Pomba, positivity was 13.4% (21/156), 14.3% (20/139) for rodents and 5.8% (1/17) for marsupials (Technical Appendix Table 1).

OPV Antibodies in Serum from Free-living Small Mammals Tested by ELISA

ELISA tests were performed on 189 rodent serum samples; a control serum for marsupials was not available. Of the animals tested, 19/189 (10.0%) were positive and 11/189 (5.8%) equivocal (Technical Appendix Table 1). By location, 3/35 (8.5%) animals from Sabará, 9/25 (36.0%) from Serro, and 7/129 (5.4%) from Rio Pomba were positive.

Sequencing

Two PCR amplicons, amplified from Sabará animals, were sequenced with C11R primers and resulted in sequences of 168 bp that were aligned with VACVs in Brazil and other OPVs (Technical Appendix Figure 1). These 2 sequences had 100% similarity with each other; similarity with VACVs in Brazil ranged from 98.2% to 100% and with CPXV from 87.3% to 89.1%, whereas similarity with VARV was 94.5% and with MPXV 90.9% (data not shown). For A56R, sequencing was performed in positive rPCR samples and resulted in 6 sequences of 102 bp. When compared with OPV sequences, 2 samples from Sabará had an 18-nt deletion shared by Brazil VACV group I, whereas the other 4 samples from Sabará, Serro, and Rio Pomba did not have that deletion, being more similar to Brazil VACV group II and other OPVs (Technical Appendix Figure 2).

Geographic and Species Distribution of Positivity

For all areas studied, 336 animals belonging to 18 genera had their samples tested by rPCR, PRNT, and/or ELISA, and 65 (19.3%) were positive in >1 of these tests. Total positivity was 18.1% for Sabará, 25.5% for Serro, and 18.5% for Rio Pomba (Technical Appendix Table 1). A higher positivity was observed for rodents (25.7%) than for marsupials (7.6%) (data not shown).

Species identified among the test-positive rodents were Calomys sp., Akodon sp., Necromys lasiurus, Trinomys setosus, Cerradomys subflavus, Oligoryzomys sp., Nectomys squamipes, Mus musculus, and Rattus rattus. For marsupials, the positive animals were characterized as species/genera Didelphis sp., and Caluromys philander (Technical Appendix Table 3).

Figure 3

Figure 3. Interaction networks for vaccina virus among small mammals in Sabará (A), Serro (B), and Rio Pomba (C) in Minas Gerais state, Brazil. The square represents vaccinia virus. Circles represent small mammal...

Test-positive animals were captured in all sample areas in Sabará (savannah and forest) and Serro (pasture, forest, and peridomicile areas), whereas test-positive animals in Rio Pomba were captured in pasture and forest. The interaction network for Sabará illustrates that 4 species had positive samples, with 3 of them found either in forest or savannah and the other in both areas. N. lasiurus (the hairy-tailed bolo mouse) had the highest number of positive samples in Sabará (Figure 3, panel A). For Serro, evidence of VACV circulation was found in 6 species; of these, 2 were captured in forest, 2 in pasture, and 2 in peridomicile areas. The species with the largest number of positive samples in Serro was T. setosus (the hairy Atlantic spiny rat; Figure 3, panel B). For Rio Pomba, 5 genera were positive, including Akodon sp. mice captured in pasture and forest, Calomys tener and N. lasiurus mice captured in forest, and C. philander and Didelphis aurita opossums captured in pasture. The species with the highest number of positive animals in Rio Pomba was C. tener (the delicate vesper mouse), followed by N. lasiurus and Akodon sp. mice (Figure 3, panel C).