|Home | About | Journals | Submit | Contact Us | Français|
Single-stranded conformation polymorphism (SSCP) analysis and heteroduplex mobility assays (HMAs) were used to identify and genotype enteric adenoviruses (EAd). The results were compared to those of restriction endonuclease assays, species-specific PCRs, and direct nucleotide sequence analyses. Of the 31 stool samples tested, 15 isolates were identified as EAd and 7 were identified as nonenteric Ad by all methods. An agreement of 100% was found between the SSCP and HMA results.
Human adenoviruses are classified into the Mastadenovirus genus of the Adenoviridae family. So far, 51 different serotypes of human adenoviruses have been described, and they are classified into six species (formerly subgenera) (A to F) (4, 10). Adenovirus serotypes 40 and 41 are the unique members of the species F, and they are referred to as enteric adenoviruses (EAd) because of their association with infantile gastroenteritis.
Single-stranded conformation polymorphism (SSCP) analysis and heteroduplex mobility assays (HMAs) have been described as alternative methodologies for the identification of genetic variants and the genotyping of viral strains (2, 3, 5, 8, 12, 14). Here we describe two protocols for using SSCP and HMA analyses for EAd identification and genotyping directly from stool samples and compare their efficacies to those of other molecular methods.
The 33 stool samples included in this study were obtained from Brazilian children under 5 years of age with acute diarrhea, were collected between February 1998 and December 2002, and were previously identified as adenovirus-positive samples by an enzyme immunoassay (11). Prototype strains of Ad serotypes 1-31, 40, and 41 were used as reference strains. Clinical specimen and adenovirus prototype DNAs were extracted by use of DNAzol according to the manufacturer's instructions (Invitrogen, Carlsbad, Calif.). PCRs were performed as previously described (1), using a pair of adenovirus-specific primers that amplified a conserved region of the hexon gene, generating a fragment of 301 bp. Aliquots of 5 to 15 μl of PCR products containing approximately 1 μg of DNA were digested for 2 to 3 h with 5 U of SalI or HinfI endonuclease, according to the manufacturer's recommendations (Promega, Madison, Wis.), to identify species F adenoviruses and to discriminate Ad40 and Ad41 (1). Twenty-two strains were identified as EAd (17 Ad40 and 5 Ad41 strains), seven strains were identified as nonenteric adenoviruses (NEAd), and four strains gave nonconclusive results (Table (Table11).
PCR amplification using species F-specific primers (AdF1 and AdF2) that generate a 541-bp (Ad40) or 586-bp (Ad41) fragment, was performed as described by Xu et al. (13). Seventeen isolates were identified as EAd, with 11 identified as Ad40 (541 bp) and 6 identified as Ad41 (586 bp), and 16 samples were negative by PCR (Table (Table11).
For SSCP analysis, aliquots of 10 μl of PCR products (301 bp) containing approximately 1 μg of DNA were added to 15 μl of formamide buffer (95% formamide, 20 mM EDTA, and 0.05% bromophenol blue) and 3 μl of 3 M NaOH. The DNA was denatured at 95°C for 8 min and then was rapidly cooled on ice (12). The denatured products were subjected to electrophoresis in a 12% polyacrylamide gel containing 0.5% polyethylene glycol at 60 V for 16 h at room temperature. The DNA bands were visualized by silver staining (7). Initially, several different adenovirus prototypes were analyzed by SSCP analysis. This assay showed that each reference amplicon had a unique electrophoresis mobility shift pattern, allowing the differentiation of distinct adenovirus serotypes (Fig. (Fig.1A).1A). Thirty-one clinical specimens were then tested by SSCP analysis, and 21 DNA samples showed a migration pattern identical to that of Ad40 or Ad41. Based on this result, the samples were identified as EAd (16 Ad40 and 5 Ad41 isolates). Ten clinical specimens showed electrophoresis mobility shift patterns that were distinct from those of EAd, and therefore these samples were characterized as NEAd (Table (Table1;1; Fig. Fig.1B).1B). Two samples could not be analyzed by this method due to the small amount of product generated by PCR.
HMA analysis was carried out by mixing, in separate tubes, 5 μl of each reference (Ad3, Ad40, and Ad41) PCR product (301 bp) containing approximately 0.5 μg of DNA with 5 μl of a sample PCR product and 10 μl of water. The DNA was denatured at 95°C for 5 min and then kept in an ice bath for 10 min (12). The resultant heteroduplexes and homoduplexes were subjected to electrophoresis in a 6% polyacrylamide gel containing 0.5% polyethylene glycol at 110 V for 75 min at room temperature. DNAs were silver stained and analyzed visually (7). Usually, nonhybridized PCR amplicons were also applied in the gel to differentiate spurious PCR fragments from the heteroduplex bands. Initially, several different adenovirus prototypes were hybridized with Ad40 or Ad41 to verify that they would yield distinct migration patterns (Fig. (Fig.2A).2A). Since the formation of heteroduplexes was demonstrated between viruses bearing different serotype specificities, we proceeded with the analysis of the clinical samples. PCR products from 31 specimens were analyzed by HMA: the formation of homoduplexes with the Ad40 reference strain was observed for 16 samples, which were therefore identified as Ad40, and the formation of homoduplexes with the Ad41 reference strain was observed for 5 samples, which were therefore identified as Ad41. Heteroduplexes with Ad40 and Ad41 were observed for 10 samples, which were therefore identified as NEAd. Two samples could not be analyzed by this method due to the small amount of product generated by PCR (Table (Table1;1; Fig. Fig.2B).2B). Identical results were obtained by SSCP and HMA.
The 301-bp PCR-amplified fragment of the hexon region of nine randomly selected Brazilian adenovirus strains, typed by SSCP and HMA analyses as Ad40, Ad41, or NEAd, was sequenced and then analyzed by using the DNASTAR program. Comparisons of the partial sequences of the hexon gene of several adenovirus strains (i) showed that the amplicons used for the SSCP and HMA analyses, although derived from a conserved region of the hexon, allowed discrimination among different adenovirus species, as previously demonstrated (1), and (ii) confirmed the identification of all nine Brazilian adenovirus strains, in total agreement with the SSCP and HMA methodologies.
Restriction endonuclease digestion assays have been described as an important tool for EAd identification (1, 6, 9). However, these methods may present some difficulties of interpretation since their efficacy depends upon the recognition of specific cleavage sites by the endonucleases in the viral genome. When this method was applied in this study, we found that four samples displayed a restriction pattern different from that expected for the enzymes used, and therefore these results were considered inconclusive. On the other hand, two isolates that were identified by this methodology as EAd were characterized as NEAd in the other assays. This could have been due to a failure of the other assays to identify these strains, or an insertion could have occurred that created a cleavage site in the viral DNA, inducing false-positive results.
The set of primers for species F adenovirus detection is quite specific and is able to discriminate serotypes 40 and 41, although rare isolates of Ad41 may be misclassified as Ad40 due to a deletion in the sequence of such strains (13). In this study, the species F-specific primer did not amplify adenovirus sequences from five samples that were positive by the generic PCR. Therefore, these primers may not be suitable for identification of adenoviruses directly from clinical samples as originally reported (13).
To our knowledge, this is the first time that SSCP and HMA analyses have been used as straightforward methods for the identification and genotyping of EAd directly from stool samples. Our results suggest that these methods can be used to identify and type EAd strains. However, further evaluation with a larger number of clinical samples is needed to validate the sensitivity, specificity, and practicality of these methods.
The nucleotide sequences of all nine Brazilian strains were deposited in GenBank under the accession numbers AY375454, AY375455, AY375456, AY375457, AY375458, AY375459, AY375460, AY375461, and AY375462.
This work was partially supported by CNPq, FINEP, FAPERJ, FUJB, and CAPES, Brazil, and TWAS, Italy.
We thank Nailini Singh-Naz, Children's Hospital National Medical Center (CHNMC), Washington, D.C., for kindly providing the EAd strains.