ABSTRACT
The Pseudallescheria boydii complex, comprising environmental pathogens with Scedosporium anamorphs, has recently been subdivided into five main species: Scedosporium dehoogii, S. aurantiacum, Pseudallescheria minutispora, P. apiosperma, and P. boydii, while the validity of some other taxa is being debated. Several Pseudallescheria and Scedosporium species are indicator organisms of pollution in soil and water. Scedosporium dehoogii in particular is enriched in soils contaminated by aliphatic hydrocarbons. In addition, the fungi may cause life-threatening infections involving the central nervous system in severely impaired patients. For screening purposes, rapid and economic tools for species recognition are needed. Our aim is to establish rolling circle amplification (RCA) as a screening tool for species-specific identification of Pseudallescheria and Scedosporium. With this aim, a set of padlock probes was designed on the basis of the internal transcribed spacer (ITS) region, differing by up to 13 fixed mutations. Padlock probes were unique as judged from sequence comparison by BLAST search in GenBank and in dedicated research databases at CBS (Centraalbureau voor Schimmelcultures Fungal Biodiversity Centre). RCA was applied as an in vitro tool, tested with pure DNA amplified from cultures. The species-specific padlock probes designed in this study yielded 100% specificity. The method presented here was found to be an attractive alternative to identification by restriction fragment length polymorphism (RFLP) or sequencing. The rapidity (<1 day), specificity, and low costs make RCA a promising screening tool for environmentally and clinically relevant fungi.
INTRODUCTION
During the last decade, physiological and morphological identification and screening methods for filamentous fungi have been supplemented with molecular techniques. Molecular siblings and genotypes within a single species are being described, which are morphologically identical but may differ in ecology (27), virulence (12), metabolic ability (12), and antifungal susceptibility (1). Strains belonging to siblings within Scedosporium and Pseudallescheria (e.g., Scedosporium dehoogii, Pseudallescheria minutispora) have a “dual ecology,” able to cause human infection on the one hand but usable for bioremediation on the other (22).
Commonly used methods for species distinction today are sequencing of ribosomal and structural genes, restriction fragment length polymorphism (RFLP), and PCR-based tools. These methods have been applied to common and clinically important species and genera (10), but for many fungi outside the well-known species spectrum, only a limited number of diagnostic tools are available. In addition, methods may be relatively expensive and time-consuming. Thus, there is a need for techniques that have high specificity, reproducibility, and sensitivity, are suitable for high throughput, have low complexity, and are cost-effective.
A recent method fulfilling these requirements is rolling circle amplification (RCA). RCA is very fast and requires limited infrastructure (33). Thus far, the technique has only occasionally been applied to fungal pathogens (14, 15, 38). Model fungi used in these papers showed relatively high interspecific diversity, with species being clearly distinct from each other at the molecular level. In the present paper, we aim to verify whether RCA is applicable for the distinction of sets of fungal species which differ by only a few nucleotides or by a single nucleotide located in a marker gene widely used for identification, i.e., the rRNA gene internal transcribed spacer (ITS) region.
We selected a group of fungal opportunists belonging to the ascomycete genus Pseudallescheria and its anamorph Scedosporium. The P. boydii complex now comprises five species, while the validities of some sibling species are still being debated (3–6). The taxonomic entities differ in ecology and in their ability to cause infections in humans (12). Consistency in these preferences was abstracted from a large data set representing isolates from humans, animals, and the abiotic environment (16). Pseudallescheria/Scedosporium species were chosen as model fungi because the group comprises emerging human opportunists that have been underdiagnosed in patient populations such as those suffering from cystic fibrosis (CF). At isolation from CF sputum, these fungi may be outcompeted by Aspergillus and Candida species, thus requiring selective media for their cultural detection (24). Pseudallescheria species have been recognized as main agents of near-drowning-related brain infection (13, 31) and are otherwise involved in a wide array of opportunistic infections in immunocompromised as well as healthy individuals. Strains affiliated with S. dehoogii tend to accumulate in human-dominated environments, particularly in those with industrial or agricultural pollution (12). The species are resistant to most commonly used antifungal drugs and therefore are refractory to therapy (30). Some species are difficult to identify (3, 5, 6). The most common species of this group, P. boydii and P. apiosperma, are very close in their ITS regions, differing only by a few nucleotides. The molecular screening technique used should therefore perform at a low level of diversity.
MATERIALS AND METHODS
Strains studied.The set of strains analyzed (677 in total) comprised 648 isolates of Pseudallescheria and Scedosporium (P. apiosperma [375 strains], S. dehoogii [110 strains], P. boydii [90 strains], S. aurantiacum [37 strains], P. minutispora [14 strains], P. ellipsoidea [1 strain], P. fusoidea [3 strains], and S. prolificans [18 strains]) and 29 distantly related strains belonging to the genera Petriella, Lophotrichus, Parascedosporium, and Graphium. Isolates were taken from the collections of (i) the Institute for Microbiology, University of Innsbruck, Austria (JK), (ii) the Robert Koch-Institute, Berlin, Germany (RKI), (iii) the CBS-KNAW Fungal Biodiversity Centre, Utrecht, The Netherlands (CBS), (iv) the Institute for Hygiene and Microbiology, Brussels, Belgium (IHEM), and (v) the Faculty of Medicine, Universitat Rovira i Virgili, Reus, Spain (FMR). Strain details are given on the Web page www.scedosporium-ecmm.com under the “Protocol” tab “Supplementary Materials.” The set contained all ex-type strains available from the genera Pseudallescheria and Scedosporium and included reference isolates recommended by Guarro et al. (9). GenBank data were used only if the identity of the isolate could be verified, i.e., when the respective strain was part of the studied set mentioned above.
DNA extraction and identification.DNA extraction and purification were performed as previously described (20, 29). Briefly, fungal material was mechanically disrupted using glass beads (425- to 620-μm diameter; Sigma, Zwijndrecht, The Netherlands) and vortex adapters for Vortex-Genie 2 (MoBio Laboratories, Carlsbad, CA), subsequently purified with cetyltrimethyl ammonium bromide (CTAB)-sodium dodecyl sulfate (SDS) buffer (Fluka, Zwijndrecht, The Netherlands). After purification of the lysate, DNA was precipitated with ice-cold isopropanol (Sigma). The DNA was pelleted by centrifugation at room temperature for 10 min at 20,000 × g (Eppendorf C5410, Hamburg, Germany), washed with ice-cold 70% ethanol (Sigma), and air dried. The DNA pellet was eluted in 50 μl Tris-EDTA buffer (Sigma) (20, 29). DNA extracts were measured with a Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Landsmeer, The Netherlands) to determine DNA concentrations. Ribosomal internal transcribed spacer regions were amplified using the primer pairs ITS4 and ITS5 (Invitrogen, Bleiswijk, The Netherlands) (18, 36). Amplification was performed in an ABI PRISM 2720 (Applied Biosystems, Foster City, CA) thermocycler as follows: 95°C for 4 min, followed by 35 cycles consisting of 95°C for 45 s, 52°C for 30 s, and 72°C for 2 min, with a delay at 72°C for 7 min. The BT2 fragment of the β-tubulin gene, known to be usable for distinction of species in Pseudallescheria and Scedosporium (3), was sequenced for a random selection of 359 strains representing all species under study by using Bt2a and Bt2b as primers (Invitrogen) (7). The annealing temperature was 58°C; sequencing was performed with a BigDye Terminator cycle sequencing kit (Applied Biosystems) using an ABI Prism 3730XL sequencer.
Phylogenetic reconstruction.One representative of each recognized ITS genotype was used to reconstruct a maximum parsimony tree using the Cipres portal (26). PAUP (Sinauer Associates, Sunderland, MA) was selected with the following parameters: heuristic search and tree bisection-reconnection (TBR), using Scedosporium prolificans as the outgroup to root the tree. To obtain a maximum parsimony tree, 200 ratchet replicates were used. Bootstrap values were calculated. The tree was created with Mega 4.0, and the bootstrap minimal concordance was set to 80.0%.
Padlock probe design.Padlock probes were designed on the basis of 677 ITS sequences of strains of Pseudallescheria and Scedosporium and relatives in Petriella, Lophotrichus, Parascedosporium, and Graphium. ITS sequences were exported to the automatic alignment package Muscle (www.ebi.ac.uk/Tools/muscle/index.html) to identify informative nucleotide polymorphisms within and between species, each species containing its ex-type isolate; the alignment was verified manually. Forty-three genotypes were detected and trimmed. A single sequence of each genotype was selected to create a genotype alignment file, which was used to design the species-specific padlock probes. The ex-type strains of all species analyzed (see Table 2 below; indicated with T) were among the sequences selected. For each species-specific site, a padlock probe was designed. Padlock probes were analyzed for tertiary structure, annealing characteristics, and palindromes in Primer select (Lasergene software; DNASTAR, Madison, WI). To enhance specific binding, the 5′-end probe binding arm was constructed with a minimum secondary structure and a melting temperature (Tm) of ≥63°C. The ligation temperature was 63°C. Species-specific sites are those regions of the padlock probes that bind to the target DNA (Fig. 1) (15, 32). To increase specificity, the 3′-end binding arm (Fig. 1) was designed with a Tm of between 48 and 53°C (29). More information on padlock probe design and structure is provided in Fig. 1 (also see Table 2). The 5′ terminal end of padlock probe p was phosphorylated (Fig. 1; also see Table 2). The species-specific 5′ site X (Fig. 1, left, underlined capital letters) is located downstream of the phosphorylated end, which is connected to the padlock probe core. The 3′ end of the padlock probe core is connected to the species-specific site Y (Fig. 1, right, underlined capital letters). Specificity was verified against sequences from the in-house database of Pseudallescheria and Scedosporium maintained at CBS for research purposes and validated by reference to ex-type strains (supplementary material, http://www.scedosporium-ecmm.com) and sequences in GenBank.
Design of an RCA padlock probe exemplified by the S. aurantiacum (AuraRCA) probe. RCA primer binding sites are in bold capital letters. Linker regions are in bold lowercase letters. Species-specific sites of the RCA padlock probe are underlined and capitalized. Target DNA of S. aurantiacum ITS amplicon is in gray and capitalized. The numbers 68 and 101 indicate the positions where the padlock probe binds to ITS amplicons of S. aurantiacum target DNA (counted from the first nucleotide of ITS1, excluding gaps).
Ligation.The ligation master mix consisted of 1 μl of purified ITS amplicon (2 to 5 pmol/μl) mixed with 2 U Pfu DNA ligase (Promega, Leiden, The Netherlands) and 10 μl of padlock probe solution (200 pmol/μl padlock probe [Invitrogen] in 20 mM Tris-HCl [pH 7.5] [Sigma], 20 mM KCl [Sigma], 10 mM MgCl2 [Sigma], 0.1% IGEPAL [(octylphenoxy-)polyethoxyethanol] [Sigma], 0.01 mM rATP [Sigma], 1 mM dithiothreitol [DTT] [Sigma]). Ligation started with 5 min of predenaturation at 94°C and then five cycles as follows: 94°C for 30 s, 4 min at 60°C. Normally the RCA reaction can be performed with one denaturation cycle followed with one ligation cycle, but 5 cycles were used to improve the yield of RCA products according to the work of Kong et al. (15). RCA reaction products were visualized on common agarose gels.
Exonucleolysis.Unbound padlock probes and ITS templates were removed by exonucleolysis in 20-μl volumes containing 10 U exonuclease I and 10 U exonuclease III (New England BioLabs, Leusden, The Netherlands) for 30 min at 37°C. Lytic activity was stopped by incubation at 94°C for 30 min.
RCA reaction.Rolling circle reactions were run in triplicate in total volumes of 25 μl with 2 μl of the enzyme-treated ligation product as the template. RCA mix contained 4 U Bst DNA polymerase (Promega), 200 μM deoxynucleoside triphosphate mix (Gentaur, Brussels, Belgium), 5 pmol RCA-primer 1b (Invitrogen), and 5 pmol RCA-primer 2c (Invitrogen) dissolved in demineralized water. The RCA reaction mix was incubated for 60 min at 65°C. RCA products were run on 1.0% agarose gels for 5 min at 50 V and then for 40 min at 100 V. A reaction without DNA served as the negative control, and as the speed marker we used DNA Smart Ladder (Eurogentec, Maastricht, The Netherlands). Double-stranded DNA products were visualized on 1.0% agarose (Invitrogen) gels containing ethidium bromide (Sigma) as a smear illuminating under UV exposure, while negative reactions showed no illumination.
Sensitivity and detection limit.ITS amplicons were compared with Smart Ladder (Eurogentec, Maastricht, The Netherlands) to determine DNA concentrations. Copy numbers were calculated with an online tool based on Avogadro's number (http://www.uri.edu/research/gsc/resources/cndna.html). For calculation in general, an amplicon length of 560 bp was assumed.
RESULTS
Forty-three ITS genotypes were recognized, and one representative of each (Table 1) was used to create a parsimony tree (Fig. 2). All species except P. boydii and P. apiosperma were statistically supported (bootstrap values of >80%). Two supported subgroups (α and β) within S. dehoogii differed by a maximum of 6 nucleotides (nt) in ITS, but bilocus analysis (ITS/BT2) suggested occasional recombination (7 strains [JK291, JK298, JK316, JK162, JK310, FMR 8532, and JK125] out of a total of 69). Within the P. apiosperma ITS, several supported subgroups (α, β, and γ) (Fig. 2) were found, but these were not confirmed in bilocus analysis (ITS/BT2). Hence, the total number of species recognized in this data set on the P. boydii complex was five.
List of 43 strains representative of ITS genotypesa
Parsimony tree constructed using the ITS genotype set of Pseudallescheria and Scedosporium species (n = 43). Intraspecific ITS variability is reflected by the number of genotypes per species and their mutual distance. Bootstrap-supported subgroups of S. dehoogii are marked with α and β, while supported subgroups within P. apiosperma are marked with α, β, and γ.
Species-specific padlock probes were designed for all five species above, which were affiliated with Pseudallescheria and Scedosporium (Table 2). For P. minutispora, the padlock probe (MinuRCA) was based on a single, unique transversion, while for P. boydii (BoyRCA), a 1-nucleotide indel was the unique feature (Table 2, bold). All remaining padlock probes (AuraRCA, DehoαRCA, DehoβRCA, ProRCA, and ApioRCA) were based on two or more nucleotides. The high intraspecific ITS sequence variability of S. dehoogii required two padlock probes. RCA reactions were performed with each species listed in Table 3: P. boydii, P. apiosperma, S. dehoogii, P. minutispora, and S. aurantiacum, with S. prolificans included for comparison.
Details of species-specific sites and padlock probe core of Pseudallescheria and Scedosporium RCA probesa
Evaluation of specificity of padlock probesa
The nucleotides characterizing P. boydii and P. minutispora (Table 2, bold) were detected by RCA; the remaining species, differentiated by combinations of mutations (Table 2), were also invariably recognized. All ITS padlock probes thus were found to be species specific, detecting all species based on molecular analysis within Pseudallescheria and Scedosporium, despite low interspecies diversity and high intraspecies variability. Positive RCA reactions were visualized by UV irradiation as ladderlike, strongly illuminating smears on the gel (Fig. 3). Negative controls were strictly without fluorescence. RCA reactions were subsequently performed without a preceding exonucleolytic step. Positive reactions were analogous to those obtained with exonucleolysis. The concordance of RCA results with expected species identification based on ITS sequence data was found to be 100%. No false-negative or false-positive reactions were found with any of the padlock probes; nor were they found with any Pseudallescheria and Scedosporium species (Fig. 3) or any more distantly related fungus. The lower detection limit of fungal RCA was established to be 107 copies when RCA products were visualized on agarose gel.
Proof of species specificity of RCA padlock probes and intraspecific variation of RCA response. Amplification and subsequent fluorescent banding were seen only with appropriate template-probe mixtures (empty lanes denote the absence of signals with unmatched template-probe mixtures). The species-specific probes are labeled as listed in Table 2 (ProRCA, S. prolificans; MinuRCA, S. minutispora; AuraRCA, P. aurantiacum; DehoαRCA, S. dehoogii α; DehoβRCA, S. dehoogii β; BoyRCA, S. boydii). Lanes: L, DNA Smart Ladder; 1 to 7, RCA reaction with DNA of P. aurantiacum (JK226) (lane 1), P. minutispora (JK305) (lane 2), S. dehoogii α (JK16) (lane 3), S. dehoogii β (JK63) (lane 4), S. prolificans (CBS 114.90) (lane 5), P. apiosperma (CBS 117407) (lane 6), and P. boydii (JK36) (lane 7); 8, negative control (reaction without DNA). Reactions for testing intraspecificity within P. apiosperma and ITS RCA padlock probe ApioRCA: lane A, P. apiosperma (CBS 116899); lane B, P. apiosperma (CBS 119697); lane C, P. apiosperma (CBS 101718); lane D, P. apiosperma (dH 18395); lane E, P. apiosperma (IHEM 21799); lane F, P. apiosperma (CBS 118234); lane G, P. apiosperma (CBS 116779); lane H, negative control (reaction without DNA).
DISCUSSION
Modern molecular taxonomy increasingly leads to the recognition of sibling species which had been indistinguishable with classical tools. Molecular entities deviate from each other by concordant mutations in independent genes and may be regarded as valid species when they have, e.g., different ecological preferences or different physiological abilities (12). Significant taxonomic developments have taken place in Pseudallescheria during the past 5 years. Scedosporium apiospermum was for a long time considered to be the anamorph of Pseudallescheria boydii. At present, P. boydii (anamorph S. boydii) and S. apiospermum (teleomorph P. apiosperma) are judged to represent two different species (5). Pseudallescheria apiosperma appears to be the prevalent clinical species. Various genetic clusters close to P. apiosperma have subsequently been described as separate taxa on the basis of multilocus analyses, mating experiments, morphology, and physiology (3, 5, 6, 25). Scedosporium aurantiacum was segregated from P. apiosperma on the basis of mating experiments and of marked physiological differences such as the production of an extracellular, canary yellow-to-reddish pigment on nutrient-rich agar media (3). The molecular species Scedosporium dehoogii is particularly found in aliphatic hydrocarbon-polluted environments (11, 23) (M. Eggertsberger, unpublished data). Pseudallescheria minutispora differed from other Pseudallescheria species by morphology and sequence data (3).
Given the near absence of classical criteria and the fact that some of the species, such as P. boydii and P. apiosperma, also have closely similar ITS sequences, recognition of these species necessitates the use of molecular techniques with high resolution power. In the present study, we used ITS sequences where possible as a basis for RCA, because in Pseudallescheria and Scedosporium this region can reliably be amplified. Moreover, no paralogues are amplified with this region. All accepted Pseudallescheria and Scedosporium species can be distinguished by ITS sequencing, although the difference between P. boydii and P. apiosperma is very small, with a minimum ITS distance of 0.3%. Two species (P. minutispora, P. boydii) were recognizable by one conserved nucleotide difference, whereas the remaining taxa were distinguishable by combinations of mutations (Table 2).
Rolling circle amplification is very economical since it can be performed under isothermal conditions in a water bath. Because of its simplicity, speed, robustness, cost-effectiveness, and specificity, the technique is promising for epidemiological high-throughput investigations. In the present study, RCA was proven to provide reproducible results down to single nucleotides (Table 2, P. minutispora and P. boydii, bold). Functional padlock probes were designed on the basis of alignments validated by inclusion of sequences derived from ex-type material of each species and verified against all currently known genotypes. Each species was characterized by one or more fixed mutations that were absent from related organisms. The underlying database contained sequences with a wide geographical representation and with a wide range of sources of isolation. Our alignment contained 677 ITS sequences derived from strains of Pseudallescheria and Scedosporium (http://www.scedosporium-ecmm.com), as well as from more distantly related taxa belonging to the genera Petriella, Graphium, Lophotrichus, Parascedosporium, and Petriellopsis.
RCA has been applied to the identification of clinically relevant fungi (14, 15, 19, 38). Circularized oligonucleotides, also called padlock probes, have been demonstrated to be more sensitive for the detection of target sequences than conventional primers (10), although gel-based detection of RCA still needs some detection optimalization (28). Zhou et al. (38) published primers for the identification of S. prolificans and the former S. apiospermum (sensu lato) based on sequences deposited in GenBank. Their S. prolificans padlock was confirmed in our validated database. It is predictive, since this species has a low intraspecific variability (30, 31). However, with P. boydii/S. apiospermum, Zhou et al. (38) disregarded the recent molecular subdivision of this complex (3, 5, 6). The padlock probes of their “S. apiospermum” (sensu lato) proved to be specific neither for the newly circumscribed species P. apiosperma (anamorph S. apiospermum) nor for P. boydii (anamorph S. boydii). This leads us to the conclusion that for reliable diagnostics of less common fungi, such as Scedosporium species, GenBank is insufficient because the majority of accessions have not been updated according to the latest taxonomic standards.
Molecular species are currently segregated on the basis of multilocus analyses rather than on rRNA gene ITS alone. After comparison with BT2 data, all species of Pseudallescheria and Scedosporium proved to be recognizable with ITS. Internal transcribed spacer is more convenient for high-throughput purposes than BT2, because it is amplified unambiguously. In the present study, we therefore maintained a preference for ITS as a diagnostic marker. Scedosporium strains are slow-growing fungi compared to other saprobic opportunists such as Aspergillus (7 to 10 days versus 2 to 4 days, respectively, to produce sporulating colonies). This is a possible explanation of the underdiagnosis of Scedosporium species in patients with cystic fibrosis, where they colonize the lungs concomitantly with A. fumigatus (17). Culture-independent detection is needed since Scedosporium strains have reduced susceptibilities to many antifungal agents. Early diagnosis is the key to successful antifungal treatment (8, 21, 35). At present, no comprehensive data exist on mortality rates attributable to the newly circumscribed Pseudallescheria and Scedosporium species in different patient groups (e.g., transplant recipients or cystic fibrosis patients). Given the differential ecological preferences of species (12), these differences are likely to exist.
The design of a generic ITS primer for the entire genus Pseudallescheria proved to be impossible, but the large-subunit (LSU) primers proposed by Wedde et al. (34) were verified and found to be usable. LSU primers were checked in the CBS in-house research database of 677 sequences, with the result that all Pseudallescheria and Scedosporium strains could be detected and no mismatches with other fungi were found.
For screening and epidemiological purposes, time- and cost-effectiveness are essential. Thus far, RCA has not been tested directly on clinical specimens. To assess the sensitivity in clinical practice, further studies are needed. RCA in Pseudallescheria comprises the following workload: (i) DNA extraction from sample pure cultures (60 to 180 min depending on extraction method), (ii) ITS amplification (60 to 120 min, depending on number of amplification cycles), (iii) ligation of padlock probes (less than 15 min), (iv) exonucleolytic reaction (60 min), (v) RCA reaction (60 min), and (vi) electrophoresis (15 to 45 min). RCA thus takes between 4.5 h and 8 h. We proved that exonucleolysis can be omitted because background signals caused by nonligated DNA were insignificantly higher than with inclusion of the lytic reaction, as was also found by Tong et al. (32). The exclusion of an exonuclease reaction shortens the procedure by a further 60 min.
Concentrations of amplicons were not determined quantitatively prior to RCA reaction. This influenced yield of RCA products as can be seen on agarose gel (Fig. 3, lanes B, F, and G). In these lanes, the yield was lower than expected, but the reaction was clearly different from that seen for the negative control (lane H). The intensity of reactions can also be influenced by the quality of the padlock probes, e.g., the presence of nonphosphorylated probe material competing with phosphorylated probes. Since RCA is based on the principle of hybridization and ligation, limitations similar to those encountered in using these techniques may be expected.
In conclusion, RCA is a very fast (less than 1 working day), specific (down to the single-nucleotide level), and economical (no additional equipment required) tool for fungal screening (2, 37), but it is less useful for the identification of single strains when padlock probes are not in stock. The method may be an attractive alternative to currently available PCR- and RFLP-based screening. Potential areas of application are in clinical microbiological assessment of CF patients whose lungs are recurrently colonized by Scedosporium species, but also in environmental microbiology, where Scedosporium species are differentially associated with environments polluted by aliphatic and aromatic hydrocarbons. Another advantage of RCA is that it can be performed under isothermal conditions with minimal reagents and no generation of false-positive results, which may be encountered in PCR-based assays (33).
ACKNOWLEDGMENTS
We thank Bert Gerrits van den Ende and Regine Horré for their suggestions and comments on the manuscript. We thank the reviewers for their suggestions and comments to improve the manuscript. We are indebted to Ferry Hagen, Rolf Boeston, Corné Klaassen, and Jacques Meis for providing DNA extracts.
The work of Mohammad Javad Najafzadeh was financially supported by Mashhad University of Medical Sciences, Mashhad, Iran.
All experiments on living material of Scedosporium and Pseudallescheria were performed according to the International Biosafety policies (biosafety level 2). The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
FOOTNOTES
- Received 17 May 2011.
- Accepted 24 October 2011.
- Accepted manuscript posted online 4 November 2011.
- Copyright © 2012, American Society for Microbiology. All Rights Reserved.
