|Home | About | Journals | Submit | Contact Us | Français|
We have developed a light-upon-extension (LUX) real-time PCR assay for detection, quantification, and genogrouping of group A rotavirus (RV), the most common cause of acute gastroenteritis in children. The LUX system uses a fluorophore attached to one primer and having a self-quenching hairpin structure, making it cost-effective and specific. We designed genogroup-specific primers having different fluorophores, making it possible to differentiate between the two main genogroups of human group A RVs. The assay was applied on clinical stool specimens from Sweden and Central America (n = 196) and compared to immunological and conventional PCR assays. The genogrouping ability was further validated against a subset of clinical specimens, which had been genogrouped using monoclonal antibodies. Our real-time PCR assay detected and quantified all positive specimens (n = 145) and exhibited higher sensitivity than immunological assays and conventional PCR. The assay exhibited a wide dynamic range, detecting from 5 to >107 genes per PCR, resulting in a theoretical lower detection limit of <10,000 viruses per gram of stool. No cross-reaction was observed with specimens containing norovirus, sapovirus, astrovirus, or adenovirus. In total, 22 (15%) of the positive clinical specimens were identified as genogroup I, 122 (84%) were identified as genogroup II, and 1 specimen was found to contain a mix of both genogroups. All genogroup I-positive specimens were associated with capsid glycoprotein 2 (G2). No significant difference in viral load was found between genogroups or geographic region. The detection and quantification, combined with the genogrouping ability, make this assay a valuable tool both for diagnostics and for molecular epidemiological investigations.
Rotavirus (RV) is the most important cause of severe diarrheal illness in infants and young children worldwide (28, 29, 39), causing more than 600,000 deaths annually (28). RV is classified into 7 different serogroups (A through G), based on the antigenic specificity of the middle-layer protein of the virus (VP6) (8), with most human RV infections caused by group A viruses (18, 19). Within group A RVs, subgroups (SGs) have historically been defined as SGI, SGII, SGI+II, and SG non-I, non-II according to reactivity of monoclonal antibodies against the VP6 protein (13). Recently, based on molecular analysis of the VP6 gene, only two groups (termed genogroups) were distinguished in group A human rotaviruses (genogroup I, SGI; genogroup II, SGII, SGI+II, SG non-I, non-II). Genogroup II is the most prevalent in humans (3, 17, 37), while genogroup I is more commonly isolated in animals (34). RVs are furthermore classified into G and P types, based on gene analysis of the genes encoding the two outer capsid proteins (VP7 and VP4) (12, 16). Although reassortment events often occur, there is evidence for independent segregation of RV genes; e.g., G1P is associated with genogroup II, and G2P is associated with genogroup I (17, 20). The genogroups have also been associated with a difference in severity, with genogroup I infections causing less severe symptoms than genogroup II infections (32, 36).
Subgrouping of the VP6 protein of RV, using enzyme-linked immunosorbent assay (ELISA), has often been applied in epidemiological studies (9, 13, 15, 17, 24, 33, 38). However, the unreliability of serological methods to characterize rotavirus due to antigenic drift and cross-reactivity is a well-known problem (6, 14). There is a reverse transcription-PCR (RT-PCR) method available for genogrouping of the VP6 gene (17). This method, however, requires an additional sequencing step which may be time-consuming and costly. A PCR assay for direct genogrouping has not, to our knowledge, been reported previously. The VP6 protein is also the primary target for RV diagnosis using ELISA (12, 18, 30). There is a limited number of RV real-time PCR assays reported, many targeting the VP6 gene, such as the SYBR green assay (31) and the TaqMan assay (22). A TaqMan-based real-time PCR assay targeting the nonstructural protein 3 gene has also been reported (27). These and other studies (2, 11) demonstrate that real-time PCR is superior in sensitivity compared to electron microscopy or immunological methods for detection and quantification of RV.
The light-upon-extension (LUX) technique applied in this work uses a fluorophore attached near the 3′end of one of the primers, constructed to form a hairpin loop, rendering fluorescence quenching capability. Incorporation of the labeled primer into the double-stranded PCR product dequenches the fluorophore, resulting in an increase of the fluorescence signal (23), and also enables the use of melting curve analysis. This technique simplifies PCR kinetics, and the hairpin oligonucleotides prevent primer-dimer formation and mispriming (25). Thus, the LUX system offers high levels of sensitivity and specificity, without the use of a probe or a quencher molecule (26).
In this study, we have developed and explored a LUX real-time PCR assay for simultaneous detection, quantification, and genogrouping of rotavirus A. A real-time assay will enable a sensitive detection of rotavirus for use in diagnostics, and the quantification and genogrouping ability will be valuable in epidemiological surveys. To our knowledge, this is the first assay of its kind developed for RV. The assay was applied on clinical specimens of different geographical origins and compared to immunological methods and other PCR-based assays in order to validate broad detection ability. We further used the assay to collect and compare epidemiological data from different areas in Sweden and Central America.
Stool specimens (n = 75) were received from the Department of Microbiology at UNAN-León University, Nicaragua. The specimens were collected from RV outbreaks in Nicaragua in 2005 (4) and Honduras in 2006 and from a survey of acute diarrhea cases in Nicaragua, 2002-2003. Furthermore, stool specimens were collected from the Microbiological Laboratory, Ryhov County Hospital (n = 46), Jönköping, Sweden, and from the Department of Clinical Microbiology, Linköping University Hospital (n = 34), Linköping, Sweden, all of them originating from patients investigated for RV infection during 2006 and 2007. RV-positive stool specimens were further received from a molecular epidemiological study of RV in Uppsala, Sweden, during 1981 (33), of which a subset had been characterized by PAGE and monoclonal antibodies targeting the VP6 protein (13) used here to assess primer cross-reactivity and genogroup specificity (n = 41).
Stool suspensions (10% [vol/vol] phosphate-buffered saline [PBS]) from Central America were first clarified by centrifugation at 4,000 × g for 3 min. Viral RNA was then extracted from 140 μl of the stool suspension using a QIAamp viral RNA minikit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. RNA was eluted with 60 μl AVE buffer (Qiagen, Hilden, Germany) and stored at −80°C. RNA extraction from stool specimens in Jönköping was performed using the EZ1 robot (Qiagen, Hilden, Germany), and the RNA was stored at −80°C. RNA extraction from stool suspensions (10% [vol/vol] PBS) from Linköping and Uppsala was performed using the BioRobot M48 (Qiagen, Hilden, Germany), and the RNA was stored at −80°C until further use.
Briefly, for reverse transcriptase (RT) PCR, 28 μl of the extracted RNA was mixed with 50 pmol of Pd(N)6 (GE-Healthcare, Uppsala, Sweden), denatured at 97°C for 5 min, and quickly chilled on ice for 2 min, followed by the addition of one Illustra Ready-To-Go RT-PCR bead (GE-Healthcare, Uppsala, Sweden) and RNase-free water to a final volume of 50 μl. The RT reaction was carried out for 30 min at 42°C to produce the cDNA later used for real-time PCR, and the reaction mixture was stored at −20°C.
G and P genotyping was performed on the Linköping and Jönköping RV-positive specimens as described by Iturriza-Gómara et al. (16). This method is a modification of the original methods described by Gouvea et al. (12) for G typing and Gentsch et al. (10) for P typing, and it includes modified G and P primers (16). The Central American specimens had been genotyped in previous studies (4, 7).
All complete sequences found (n = 125) of the VP6 gene from GenBank at the start of the study were aligned using ClustalW (35), followed by phylogenetic analysis with the neighbor-joining and Kimura 2-parameter methods using the MEGA 4.0 software (21). A distinct division of two separate genogroups of group A RV was found (data not shown), which has also been shown elsewhere (20). These alignment data were used for the design of genogroup-specific primers.
Using the alignment data for the complete VP6 genes (n = 125), LUX primers were designed, and their properties were calculated with the software OligoAnalyzer 3.0 (Integrated DNA Technologies, Coralville, IA) and Oligo Property Calculator 3.1 (Northwestern University, Chicago, IL). The forward primer (VP6fw) was designed to be specific for both genogroups, while the two different reverse primers (VP6rvI and VP6rvII) were designed to be specific for RV genogroups I and II, respectively. The reverse primers were subsequently attached with different fluorophores, 6-carboxy-4′,5′-dichloro-2′,7′-dimethoxyfluorescein (JOE) for genogroup I and 6-carboxyfluorescein (FAM) for genogroup II (Table (Table1)1) . They differ in their emission wavelengths (FAM at 518 nm, JOE at 548 nm), thus enabling the assay to distinguish between the two genogroups using different emission filters. Investigation of primer cross-specificity was done using the BLAST software (1).
A 270-bp fragment spanning the nucleotides 1017 to 1284 in the VP6 gene of the RV genome was amplified with PCR from a clinical RV A genogroup II strain, using the primers SGIIfw (5′-GAA TCA GTG CTT GCG GA-3′) and SGIIrv (5′-AGA GTC TGA ATG ACT TGA-3′) (this study). The PCR product was cloned into a pPCR-Script Amp SK(+) vector and transformed into XL10-Gold Kan ultracompetent cells, using the Stratagene PCR-Script Amp cloning kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. After overnight incubation, plasmid DNA was extracted and purified using the Qiagen plasmid miniprep kit (Qiagen, Hilden, Germany) according to manufacturer′s instructions. The DNA concentration was determined using the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE).
Designed PCR primers without fluorophores were evaluated using conventional PCR with cDNA from patient material from both Central America and Sweden which had previously been diagnosed as RV positive. Different annealing temperatures (50 to 58°C) were tested in order to optimize the reaction. The PCR mixture contained 2.5 μl cDNA from the RT reaction, 2.5 μl of 10× PCR buffer, 1 μl of 50 mM MgCl2, 2 μl of 2.5 mM deoxynucleotide triphosphate (dNTP), 0.5 μl of 10 pmol μl−1 of each primer, 0.125 μl of 5 U μl−1 Taq polymerase (Invitrogen, Carlsbad, CA), and RNase-free water up to a final volume of 25 μl. The PCR was carried out on a PCT-100 programmable thermal controller (MJ Research, Waltham, MA). The reaction mixture was preheated at 95°C for 5 min, followed by 35 thermal cycles of 15 s at 94°C, 30 s at 55°C, and 30 s at 72°C and thereafter a final extension step at 72°C for 10 min. The PCR products were analyzed on a 2% agarose gel and visualized by ethidium bromide staining and UV transillumination.
Two μl of cDNA from the RT-PCR was added to an 18-μl reaction mixture, consisting of 10 μl Platinum Quantitative PCR SuperMix-UDG (Invitrogen, Carlsbad, CA), 0.04 μl ROX reference dye (Invitrogen, Carlsbad, CA), 0.2 μl of 10 pmol μl−1 of each reverse LUX primer, 0.4 μl of 10 pmol μl−1 of unlabeled forward primer, and 7.16 μl RNase-free water, to a total volume of 20 μl. The real-time PCR was performed on a 7500 fast real-time PCR system (Applied Biosystems, Foster City, CA). An initial uracil DNA glycosylase contamination protection step was performed at 50°C for 2 min, followed by 2 min at 95°C and then 40 cycles of 15 s at 95°C, 30 s at 55°C, and 30 s at 72°C. Melting curve analysis was performed immediately after PCR completion, by heating at 95°C for 15 s, followed by cooling to 60°C for 1 min and subsequent heating to 95°C at 0.8°C min−1 with continuous fluorescence recording. Melting temperatures were determined on all samples using the Sequence Detection Software version 1.3.1 (Applied Biosystems, Foster City, CA) and visualized by plotting the negative derivatives against temperature. A selection of PCR products was furthermore analyzed with electrophoresis on a 2% agarose gel to verify that they had the expected length.
The RV LUX real-time PCR assay was compared to a PCR-based method targeting the VP7 gene, described by Iturriza-Gómara et al. (14). This PCR assay was applied on all specimens from Linköping and Jönköping. The specimens from Jönköping were additionally tested using a commercial immunochromatographic test (Combi-Strip Rota- and Adenovirus Diagnostic C-1004; BioConcept, Uppsala, Sweden), and those from Linköping were additionally tested with a commercial ELISA (IDEIA Rotavirus, ref. K6020; Oxoid, Copenhagen, Denmark). The RV-positive specimens from Uppsala and Stockholm had been verified positive previously using RNA electrophoresis (36). The specimens from Central America were tested with commercial immunological tests (IDEIA Rotavirus, ref. K6020 [Oxoid, Copenhagen, Denmark]; ELISA kit for Rotavirus [Cypress Diagnostic, Langdorp, Belgium]; or Rota-Strip C-1001 [BioConcept, Gembloux, Belgium]) and genotyped as described by Iturriza-Gomara et al. (16).
Using the reference DNA in a serial dilution, we determined the detection limits, efficiency, and dynamic range of the real-time PCR assay. We estimated the amount of viruses in the starting material of the Central American and Linköping specimens, measured as genome equivalents per gram of stool, based on the assumption of a 100% yield. At Ryhov Hospital, Jönköping, only the number of genes per PCR reaction was calculated, since the exact amount of stool sample was not estimated before RNA extraction.
To assert the ability of the assay to detect a simultaneous infection of the two genogroups present in different concentrations, we diluted cDNA from clinical specimens of genogroups I and II (n = 4) 10- to 1,000,000-fold. These dilutions were subsequently mixed in combinations from high to low concentrations and run in the real-time PCR assay we developed to determine the maximum concentration difference for detection of mixed infection of genogroups I and II.
Nucleotide sequencing was performed by Macrogen Inc. (Seoul, South Korea) on the reference DNA in order to verify it. The sequencing reaction was based on BigDye chemistry, using M13 forward and reverse primers as sequencing primers.
The detection limit of the real-time PCR assay was determined to ≤5 gene copies per PCR, corresponding to ≤10,000 RV per gram stool, using a dilution series of reference DNA. The lowest quantification limit was estimated to 10 gene copies per PCR, and the upper quantification limit was estimated to 1 × 107 gene copies per PCR, corresponding to ~3.8 × 1010 RV per gram stool (see Fig. S1 in the supplemental material). As the reference DNA belonged to genogroup II, we also investigated its applicability as a reference for quantification of genogroup I. Using clinical specimens of genogroup I in serial 10-fold dilutions, we found that the real-time PCR efficiency was comparable to the real-time PCR efficiency when genogroup II was used as the reference DNA. Thus, we concluded that only one set of reference DNA for the standard curve is needed for quantification of both genogroups.
Primer cross-reactivity with clinical specimens positive for norovirus genogroups I and II, sapovirus, adenovirus, and astrovirus was not observed, nor was cross-reactivity between the RV genogroup I-specific reverse primers and RV genogroup II-positive specimens. The genogroup II-specific primer occasionally cross-reacted weakly with genogroup I specimens, yielding an easily distinguishable melting curve profile which was lower than the detection limit of reference DNA (5 plasmid copies). The melting temperature range for the genogroup I amplicons was 77.1 to 78.9°C ( = 78.2; 95% confidence interval [CI], 78.1 to 78.4), and for the genogroup II amplicons it was 77.2 to 79.5°C ( = 78.7; 95% CI, 78.5 to 78.8) (see Fig. S1 in the supplemental material).
We applied the LUX real-time PCR assay on 65 RV-positive specimens from Central America that previously had been G and P typed as described by Iturriza-Gomara et al. (16). This was done in order to verify that the assay was able to detect the most common genotypes and also to investigate the genogroup distribution in relation to combinations of G and P types. Of 65 specimens, 14 (22%) belonged to genogroup I and 50 (77%) belonged to genogroup II. One specimen contained a mix of both genogroups (Table (Table2).2). The genogroup I specimens were associated with G type 2. Ten ELISA RV-negative specimens were also included, and all of them yielded negative results in the real-time PCR.
From the 46 specimens collected at Ryhov County Hospital, 24 specimens were positive with our RV LUX real-time PCR assay (Table (Table3)3) . The Combi-Strip detected 18 of these 24 specimens, and the VP7-PCR detected 22. The remaining 22 specimens were negative by all three methods. The 22 specimens positive in the VP7-PCR were assigned to genogroup II, while 2 specimens were assigned to genogroup I, using the LUX real-time PCR assay. The genogroup I-positive specimens were determined to be G2P (Table (Table3).3). From the 34 specimens collected at the Department of Clinical Microbiology, Linköping University Hospital, in total 15 were positive with all three methods (Table (Table4).4). One of them was determined to be genogroup I, while the remaining 14 belonged to genogroup II. The rest of the 19 specimens were negative by all three methods. The genogroup I-positive specimens were determined to be G2P (Table (Table4).4). Furthermore, all the previously subgrouped RV specimens collected in 1981 from Uppsala (33) (n = 41) were found positive with the real-time PCR. Of these, five specimens belonged to genogroup I and 36 specimens to genogroup II, all in accordance with previous observations.
In summary, out of 196 clinical specimens from Sweden and Central America investigated in this study, the LUX real-time assay we developed detected more (n = 145) RV-positive specimens than other methods (Table (Table5)5) . The majority of RVs was found to belong to genogroup II (n = 122; 84%), while 22 specimens were determined to be genogroup I, using the genogrouping ability of the LUX real-time PCR assay. One specimen was found to contain a mixture of both genogroups.
We quantified and compared RV concentrations between specimens of RV genogroups I and II from Central America and RV concentration from Swedish specimens from Linköping. No significant difference regarding genogroups or geographic location was found when comparing the number of viruses in feces. The geometric mean concentration of RV from Central America was 1.0 × 109 RV g feces−1 (8.3 × 108 for genogroup I and 1.1 × 109 for genogroup II), and from Sweden it was 5.2 × 108 RV g feces−1 (see Fig. S2 in the supplemental material). We furthermore investigated the ability of our RV real-time PCR assay to detect and quantify mixed infections with the two genogroups. Our experiments showed that if the two genogroups were present in concentrations differing up to 1,000-fold, a detection of mixed infection could be performed (data not shown); however, it could not be accurately quantified if concentrations differed more than 10-fold.
We have established and evaluated a novel, highly sensitive and specific LUX real-time PCR assay, able to broadly detect and quantify group A RV. Using genogroup-specific reverse primers with different fluorophores, the assay also detects and distinguishes between genogroups I and II. We have applied the assay on specimens from Sweden and Central America, showing that the assay can successfully be used on specimens from different geographic regions. By using the assay, we also obtained valuable epidemiological information about viral load and genogroup distribution in different settings.
The real-time PCR assay we have developed can be used for quantification of the viral load in clinical specimens, with a broad dynamic range of 101 to 107 gene copies. A broad quantification range is useful since the concentrations of RV can vary manyfold between different specimens, e.g., when sampling is performed at different time points after the onset of illness. Quantification can also enable the use of cutoff values in order to distinguish between clinical and subclinical infections, as suggested by Phillips et al. (30). Also, if the assay is to be applied on environmental samples to measure RV concentration, e.g., in wastewater, a broad quantification range is valuable. Using known copy numbers of reference DNA, the assay can detect ≤5 gene copies per PCR, which corresponds to ~10,000 RV per gram of stool. Logan et al. (22), using a TaqMan assay, reported a detection limit of 10 gene copies per PCR using plasmid DNA for group A RV, and Schwarz et al. (31), using SYBR green, reported a detection limit of ~10 RNA transcripts per PCR. The design of short amplicons for the real-time PCR assays developed in this study will likely result in a more efficient amplification and higher sensitivity. A low detection limit is especially useful when sampling occurs late after the disappearance of symptoms, when the viral load in stool is likely to be low.
We observed that the RV LUX real-time PCR assay can detect the most common human RV G and P types (Tables (Tables2,2, ,3,3, and and4),4), and it was shown to be more sensitive than the immunological assays and conventional PCR methods used in this study. In the clinical specimens obtained from Jönköping, Sweden (n = 46), the immunochromatographic (Combi-Strip) test detected only 75% and the ELISA test 92% of the RV-positive specimens compared to the LUX real-time PCR. All specimens that were positive in the LUX real-time PCR were verified by at least one other method or by genotyping (Table (Table3),3), verifying that the higher detection rate of the LUX real-time PCR assay was not due to false positives. As seen in Table Table3,3, one reason why the immunochromatographic and conventional PCR tests failed to detect some specimens may be that the concentration of viruses was too low. Another reason may be that the antigenic and genomic sites targeted by these tests are less conserved, since the concentration of RV in some of the undetected specimens was high (Table (Table33).
Previous real-time PCR assays targeting the VP6 gene have focused on detection of group A RV and not on distinguishing between genogroups (22, 31). To our knowledge, this is the first real-time PCR assay developed to distinguish between genogroups I and II of human RVs. The LUX assay uses only three oligonucleotide primers with a low level of degeneration, making the assay sensitive while still maintaining genogrouping ability. The lack of probe, compared to TaqMan assays, also makes it less prone to mismatch errors due to eventual genetic drift of the rotavirus. The genogrouping ability gives additional important information regarding immunology and epidemiology to RV, and this is obtained instantly without the need for post-PCR processing.
The genogrouping primers were first tested against a subset of specimens that have previously been subgrouped by PAGE and monoclonal antibodies (33, 36). Then, we compared the distribution of genogroups with the G and P types of the RV specimens and found that all RV genogroup I-positive specimens were associated with G2P or G2 with unknown P type, with one genogroup I-positive specimen not possible to genotype. The independent genomic segregation of genogroup I and G2P specificities has been shown before (17, 20), which confirms the ability of the LUX assay to accurately determine genogroup specificity in unknown specimens. Identification of a specimen that contained a mix of genogroup I and II viruses identified as G2P and G9P, which is in concordance with common RV genomic segregation patterns (17), demonstrates the ability of the assay to detect mixed infections of both genogroups. Also noteworthy is the finding of the two uncommon P RV types in the Central American specimens, two specimens to which G type could not be assigned. One of them belonged to genogroup I, and the other one belonged to genogroup II.
Genogroup II infections are more common than genogroup I infections in humans, and some reports have proposed that genogroup I viruses replicate less efficiently than genogroup II viruses, although the data are inconclusive (32, 36). We addressed this by quantifying the amount of virus in feces, and we compared the viral load for different genogroups and also between geographic regions. No significant differences were observed between genogroups or between geographic regions (see Fig. S2 in the supplemental material). The different genogroups were thus not shed in different amounts, something which has earlier been described for norovirus infections (5). This indicates, but does not prove, similar replication efficiencies between the genogroups. The genogroups could, however, be found in similar amounts in stool while having different replication efficiencies, e.g., depending on time of sampling. The specimens from Sweden contained slightly fewer viruses per gram of stool than the Central American specimens, but since the RNA extraction assays in these two geographical areas were performed with different methods, no general conclusions can be drawn from this.
We found that genogrouping of specimens containing a mix of both genogroups could be achieved when concentrations differed up to 1,000-fold between genogroups. We furthermore observed, using spiked specimens, that if the mixed genogroup concentrations differed up to 10-fold, quantification of both genogroups could be performed simultaneously. We quantified the number of RVs in the specimen containing a mix of both genogroups with monoplex real-time PCR, using only one primer pair, and found that the measured viral load did not differ compared to measurements using multiplex reactions with all three primers simultaneously. However, when a mixed genogroup infection is detected, it would be advisable to do a monoplex real-time PCR for each genogroup to ensure an accurate viral load estimation, if such is needed.
The genogroup I-specific reverse primer was on no occasion observed to cross-react with genogroup II-positive material, either from reference samples with high concentrations or from clinical specimens. The genogroup II-specific reverse primer, on occasion, weakly cross-reacted with genogroup I-positive material if it was presented in high concentration. This was, however, easily discarded as a cross-reaction due to the unusually low melting curve profile, which was well below the detection limit of reference DNA (5 copies). This was regarded as a cutoff signal and thus interpreted as negative in our assay.
In an earlier study, we described a real-time PCR assay for detection and differentiation between norovirus genogroups I and II, using the same LUX technique (26). The amplicons in the norovirus assay show a nonoverlapping interval of melting temperatures with the real-time PCR assay for RV in this study. Thus, it would be possible to use these primers for RV, as well as for norovirus, in a multiplex assay, allowing simultaneous detection and differentiation between norovirus genogroups I and II and group A RV, with melting curve analysis and/or different fluorophores. Since the symptoms of rotavirus- and norovirus-induced diarrhea can be very similar in children, such an approach would considerably lower the screening costs, while enabling a fast determination of the pathogen.
The novel LUX real-time PCR assay described is time-saving, simple, and cost-effective, since there is no need for expensive probes or quencher molecules, and it can be used on most real-time PCR platforms. The assay can be used for detection and quantification of human group A RV, and the use of different fluorophores enables the assay to directly genogroup the specimens without additional sequencing or other post-PCR processing.
This study was supported by the Swedish Research Council for the Environment, Agricultural Sciences and Spatial Planning (grant 245-2004-1821) and by the Swedish Research Council (grant 10392).
We thank Felix Espinoza, Annabelle Ferrera, Britt Åkerlind, and Andreas Matussek for kindly providing patient material.
Published ahead of print on 10 March 2010.
†Supplemental material for this article may be found at http://jcm.asm.org/.