Predominance and Circulation of Enteric Viruses in the Region of Greater Cairo, Egypt

Sample collection and preparation.Stool specimens were collected at the Tabarak Hospital Cairo and at dispensaries in Gizzeh, Egypt, from 1-month-old to 18-year-old patients who visited the clinics for acute gastroenteritis from March 2006 through February 2007. The patients were routinely checked for symptoms of dehydration. In the event of severe dehydration, the patient was hospitalized for intravenous rehydration until improvement of his status. The stool specimens were routinely tested in Cairo for the presence of bacterial infections and fecal leukocytes. Two hundred thirty clinical specimens from sporadic cases of gastroenteritis for which the bacterium and leukocyte tests were negative were selected for further analysis. During the cold season (September 2006 though February 2007), 73% of these samples (n = 169) were collected; the mean age of the patients was 1.7 years, ranging from 1 month to 18 years of age. For the warm season (March through August 2006), 61 samples were collected; the mean age of the patients was 10.6 years, ranging from 1 year to 17 years old. Virological analysis was performed at the National Reference Center for Enteric Viruses in Dijon, France. For each patient, a 10% stool suspension was prepared in phosphate-buffered saline. The stool suspension was used for antigen detection by an enzyme-linked immunosorbent assay. Additionally, 500 to 1,000 μl of the stool suspension was used for the extraction of nucleic acids (DNA and RNA) using a Nuclisens Easy MAG system (bioMérieux, Marcy l'Etoile, France) according to the manufacturer's instructions. RNA was eluted in a final volume of 50 μl and used for NoV detection and typing of the virus-positive samples. The NoV strains E872 (Farmington-like), E1057 (Hunter-like), E1501 (2006a-like), E2703 (2006b-like), and E1267 (2006b-like), which originated from outbreaks of gastroenteritis that occurred in France between 2002 and 2008, were used as reference strains.

Detection and typing of rotaviruses, astroviruses, adenoviruses, and NoVs.The stool specimens were first screened for the presence of group A rotavirus, astrovirus, and adenovirus types 40 and 41 by an enzyme-linked immunosorbent assay, as described previously (31). For the rotavirus-positive samples, the G and P genotypes were determined by reverse transcription-PCR (RT-PCR) using genotype-specific primers, as described previously (12, 14). For rotavirus strains that could not be typed by PCR, the G genotype was determined by sequence analysis of the entire VP7 gene segments, amplified by RT-PCR with primers VP7-F (nucleotides 49 to 71) and VP7-R (nucleotides 914 to 933), as previously reported (17), and primers Beg 9 and End 9 (14). To genotype the adenovirus-positive samples, the primer set Adv-Hex1 DEG/Adv-Hex2 DEG was used to amplify a portion of the hexon gene (1). The nucleotide sequences of the PCR amplicons were determined and compared to those of reference strains. The human astrovirus genotype was determined by sequence analysis of the PCR products obtained with primers Mon244 and Mon245 located in the capsid coding region, as described previously (25).

For NoV detection, the RNA was extracted and purified as described above, and 3 μl was used for each RT-PCR assay. The RT and PCR steps were performed in a single tube by using a Qiagen (Hilden, Germany) one-step RT-PCR kit according to the manufacturer's instructions. The primers JV12 and JV13 were used to amplify a conserved region of the polymerase domain (38). For GGI and GGII NoVs, a conserved region of the ORF2 gene was amplified by using primer sets G1SKF/G1SKR and G2SKF/G2SKR, respectively, as described previously (19). The NoVs detected in the stool samples were genetically characterized by nucleotide sequencing of the PCR products. The genotypes were determined by alignments with reference sequences from GenBank.

Amplification and cloning of the entire sequences of ORF2 and ORF3 of GGII.4 NoVs.Because GGII.4 NoVs were predominant, we determined the entire sequences of ORF2 and ORF3 from these strains for phylogenetic analysis. Three microliters of the extracted RNA was denatured by incubation at 68°C for 10 min and then chilled on ice. The primer pairs FW1/RT5 and FW5/RT7 (Table 1) were used to amplify ORF2 and ORF3, respectively. The denatured RNA and 10 μM of each primer were used in a final volume of 50 μl for RT amplification using a Titan one-tube RT-PCR kit (Roche Applied Biosystems) according to the manufacturer's instructions. RT-PCR was performed at 50°C for 60 min; followed by 10 cycles at 94°C for 30 s, 50°C for 30 s, and 68°C for 2 min 40 s; followed by 40 cycles at 94°C for 30 s, 55°C for 30 s, and 68°C for 2 min 40 s, with an increment of 5 s per cycle for the elongation phase; and a final runoff at 68°C for 10 min. The amplicons were analyzed on a 1% agarose gel for the presence of the fragments corresponding to ORF2 and ORF3, respectively. The PCR products were purified from the gel and cloned into the pGEM-T easy vector (Promega France, Charbonnières-les-Bains, France) prior to sequencing.

Nucleotide sequencing and phylogenetic analysis.To genotype NoVs, astroviruses, and adenoviruses, the PCR products were directly sequenced. For the GII.4 NoV strains, the recombinant constructs containing the ORF2 and ORF3 inserts were sequenced in both directions with T7 primer (5′-TAATACGACTCACTATAGGG-3′) and SP6 primer (5′-TATTTAGGTGACACTATAG-3′), in addition to the newly designed primers listed in Table 1. The sequencing reaction was performed by using an ABI Prism big dye terminator cycle sequencing ready reaction kit v1.1 (Applera, France) and an ABI 3100 automated sequencer (PE Biosystems). The nucleotide sequences were edited with the CodonCode Aligner program version 2.0.4 (Dedham, MA). The sequences were aligned using Clustal W2 (Heidelberg, Germany).

A set of 67 (ORF2) and 37 (ORF3) representative strains from GenBank and this study was used to construct phylogenetic trees based upon amino acid sequences. The alignment of the sequences was generated using Clustal W from the MEGA 4.0 package (34). The percentage of identity between isolates or clusters was calculated using MEGA 4.0 software. The percentage of identity was determined according to the number of nucleotide changes per site. The consensus trees were each generated from 1,000 replicates using the neighbor-joining (NJ) method for clustering (MEGA 4.0 software). To study evolution between strains, a minimum spanning tree (MST) from the Bionumerics package (Applied Maths BVBA, Sint-Martems-Latem, Belgium) was constructed using the default setting. The MST was calculated according to amino acid changes, and unlike the NJ method, the MST takes into account the position of each variation. Of note, for this type of tree, it is assumed that each amino acid change is unique and that reverse mutations do not occur. The MST was based upon 193 complete ORF2 amino acid sequences of GGII.4 from GenBank and this study (the list is available upon request). Version 3.5.1 of Simplot software was used to detect putative recombination events between ORF2 and ORF3. The ORF2 and ORF3 plasmid constructions for the Cairo strains are available upon request.

Nucleotide sequence accession numbers.The NoV nucleotide sequences reported here have been deposited in GenBank under accession numbers EU876882 (Cairo 2), EU876883 (Cairo 3), EU876884 (Cairo 4), EU876885 (Cairo 5), EU876886 (Cairo 6), EU876887 (Cairo 7), EU876888 (Cairo 8), EU876889 (Cairo 9), EU876890 (E1057 Dijon), EU876891 (E2703 Dijon), EU876892 (Cairo 1), EU876893 (Cairo 10), EU876894 (E1501 Dijon), EU876895 (E1627 Dijon), and FJ538900 (E872 Dijon).

Sample collection and virus detection.Two hundred thirty stool specimens from cases of gastroenteritis were selected for further investigation for enteric viruses based upon the absence of pathogenic bacteria and fecal leukocytes. Eighty percent (n = 184) of the 230 patients presented mild (n = 133) or severe (n = 51) dehydration. Half of the 230 cases were positive for at least one enteric virus (Table 2), and were all associated with mild (n = 64) or severe (n = 51) dehydration. The bulk of the positive specimens were detected during the cold season (88.7%, n = 102). Thirteen of the 61 tested specimens were positive for enteric viruses during the warm season. Ten percent of the viral infections were mixed viral infections. Sixty and 86% of the positive samples were taken from patients younger than 1 and 5 years of age, respectively. For patients over the age of 5 years, only 22% of cases of gastroenteritis were viral infections, and no mixed infections were observed. Rotaviruses and NoVs were the predominant viruses detected during the survey and accounted for 87.8% of viral infections, including mixed infections. The adenovirus was the third cause of viral diarrhea in our study and was detected in all age groups. Human astrovirus was detected in five patients; three of these were mixed infections. All cases of severe dehydration (n = 51) were associated with either rotavirus (n = 32) or NoV (n = 10) monoinfections or mixed infections (rotavirus-NoV [n = 5], rotavirus-astrovirus [n = 3], or rotavirus-NoV-adenovirus [n = 1]). Severe dehydration occurred mostly in children under 1 year old.

Typing.For group A rotaviruses, the P and G genotypes were determined by multiplex RT-PCR and sequencing (Table 3). The P genotype was determined for all strains. Three P genotypes were found, P[4], P[8], and P[6], which accounted for 51.3, 42.1, and 13.2% of the rotavirus strains (including the mixed infection), respectively. Four G genotypes were detected. G2 and G1 genotypes predominated and accounted for 53.9 and 19.7% of the rotavirus strains, respectively. G9 and the rare G12 genotypes were also detected, and each accounted for 9.2% of rotavirus strains. The G genotype could not be determined for six rotavirus strains. The rotavirus strains G2P[4] and G1P[8] were the most predominant and accounted for 40.8 and 17.1% of rotavirus strains, respectively. All but one G9 strain showed a P[8] genotype, and the G12 rotaviruses were associated mainly with P[6]. All of the G9- and G12-related strains were detected in children under the age of 5 years. The G12 rotavirus isolates were all associated with severe dehydration. Three rotavirus strains, which were detected in 8-, 11-, and 12-year-old patients, were P typed but were not typeable for the G type. These data suggested that these G types might be unusual variants infecting older children and could not be amplified by the typing primers that we used.

NoVs accounted for 27% of the enteric viruses based upon our PCR results. Twenty-nine percent and 71% of the strains belonged to NoV GGI and GGII, respectively (Table 4). Among the GGII NoVs, GGII.4 strains predominated, representing 45% of all of the NoV isolates. For one sample, GGII.4 and GGII.15 were detected in the polymerase and the capsid regions, respectively, which might suggest the presence of both NoVs in the stool specimen. The newly characterized GGIIb recombinant was also detected often (n = 7) and was associated with the GGII.3 capsid genotype for four samples. Sequencing the ORF1-ORF2 junction of one isolate confirmed that the GGIIb NoV in the polymerase was indeed GGII.3 in the capsid, as described previously (2). For the GGI NoVs, GGI.1 and GGI.9 accounted for two-thirds of the GGI NoV strains. Of note, the percentage of GGI NoVs could have been underestimated during the study since they were detected with only one set of primers.

Phylogenetic analysis of GGII.4 NoVs.For several years, GGII.4 NoVs have been the most predominant NoVs detected in Europe and North America during outbreaks of gastroenteritis (8, 20). To determine whether similar GGII.4 NoVs were circulating in Egypt, the entire nucleotide sequences of ORF2 were determined for the GGII.4 isolates that were found in Cairo. Given the similarity between the sequences in the capsids of the 2006a and 2006b strains from this study and those of the GGII.4 NoVs isolated in Europe (e.g., France), it can be suggested that similar strains circulate among Mediterranean countries (Fig. 1). Moreover, we detected three new GGII.4 NoV strains, belonging to none of the known variants. The tree constructed from the amino acid sequences using either the NJ (Fig. 1) or the maximum-parsimony (data not shown) method clearly showed that the three new variants clustered together to form a new subgroup, the Cairo cluster. For all but the Bristol and the Cairo variants, we observed at least 95% nucleotide identity within each cluster (Table 5). The percentage of nucleotide identity between subgroups ranged from 88.4% (Bristol and 2006b) to 96.1% (Hunter and 2006a). The average amino acid identity between variants was 94.5% and ranged from 92.1% (Bristol and Sakai) to 97.7% (Hunter and 2006a) (Table 5). Previous studies showed the increasing diversity of NoVs within GGII.4 strains (4, 20, 21, 22, 33). To determine the minimum number of amino acid changes between strains, an MST was constructed based upon the complete amino acid sequences of ORF2 using a set of 193 sequences corresponding to the complete ORF2 of Bristol, US95/96, Farmington, Hunter, Sakai, 2006a, 2006b, and Cairo variants (Fig. 2). The MST is an alternative tool to study the relationship between closely related NoV strains. It is important to mention that a complete set of sequences, including those of outlying strains (e.g., Erfurt [Fig. 2]), must be used to construct a valid MST. Overall, the lineages and variants found for the MST matched those observed for the NJ tree (Fig. 1). The MST was built with 484 amino acid changes. The MST suggested that 2006a and 2006b might be directly related to the former Hunter and Farmington variants, respectively, as previously described (33). The tree showed that the Hunter, Farmington, and Sakai variants might all be related to the US95/96 variants via the Lanzhou isolate, which might be considered an intermediate strain. The MST showed that the three strains belonging to the Cairo group were distant from the Sakai variants (data not shown) and the US95/96 (Fig. 2) isolates by the same number of amino acid changes (n = 23).

• Open in new tab
• Download powerpoint

FIG. 1.

Phylogenetic tree based upon the amino acid sequences of ORF2 from 67 representative strains from GenBank and GGII.4 isolates from this study. The groups of variants and outlier strains are indicated to the right of the tree. Bootstrap values are percentages of 1,000 iterations. The scale is indicated at the bottom and corresponds to the number of amino acid substitutions per amino acid residue. The GenBank accession numbers of the strains are indicated in parentheses. The NoV isolates sequenced during this study are indicated in bold.

• Open in new tab
• Download powerpoint

FIG. 2.

MST based upon the alignment of 193 complete amino acid sequences of ORF2 of GGII.4 NoVs from GenBank and this study. The numbers of amino acid changes between strains are indicated on the connecting lines. The groups of variants have arbitrarily been defined by six or fewer amino acid changes. The number of sequences for each group of variants is indicated in parentheses. Identical sequences are represented by color-coded circles, which are scaled according to member count. The white and light-gray circles represent one and two sequences, respectively. For dark-gray and black circles, the number of identical sequences is indicated inside each circle. The isolates Erfurt, Lanzhou, Houston, and Ast6139 are indicated by the initials E, L, H, and A, respectively.

In the last part of the study, we analyzed the complete GGII.4 ORF3 sequences to determine whether ORF3 presented the same genetic distribution as ORF2. For the E1057, Cairo 10, Miami Beach, and Yuri strains, we observed the deletions of 1 (position 145), 2 (positions 145 and 146), 6 (positions 206, 207, and 209 through 212), and 7 (positions 150 through 153 and 155 through 157) amino acid residues, respectively, located in the variable region of ORF3, as described previously (7). For E1057 and Cairo 10, the presence of the deletions was confirmed by direct sequencing of the PCR products from the stool specimens (data not shown). Despite the less available sequence information for ORF3, 37 representative sequences from this study and GenBank were used to construct a phylogenetic tree. We observed clustering of the variants similar to that previously observed for ORF2 (Fig. 3). However, the Cairo variants were related to the US95/96 and Bristol variants in ORF3, while they were closely related to 2006b in ORF2 (Table 5 and Fig. 1). Additionally, the Lanzhou and Houston strains did not cluster together in ORF3. The variations between ORF2 and ORF3 that we observed might suggest an intragenotypic recombination event(s) between groups of variants, which may have been impossible to detect because of the strong homology between isolates (data not shown).

• Open in new tab
• Download powerpoint

FIG. 3.

Phylogenetic tree based upon the amino acid sequences of ORF3 from 37 strains from GenBank and GGII.4 isolates from this study. The variants are indicated to the right of the tree. Strains with deleted amino acids in ORF3 are indicated by an asterisk. The bootstrap values and scale are described in the legend for Fig. 1. The GenBank accession numbers of the strains are indicated in parentheses. The NoV isolates sequenced during this study are indicated in bold.