Introduction

Human noroviruses (NoVs), members of the family Caliciviridae, are now recognized as the most common cause of outbreaks of nonbacterial gastroenteritis worldwide [1, 2]. NoV infections are especially associated with the ingestion of contaminated water [3–6], food [7, 8], and oysters [9, 10]. Since a low infectious dose is enough to develop illness in humans, the virus is highly contagious [11, 12]. It is estimated that noroviruses cause about 900,000 episodes of gastroenteritis that require a clinic visit and 64,000 hospitalizations among children <5 years of age residing in high-income countries each year. In developing countries, noroviruses are estimated to cause 200,000 child deaths each year [13].

NoVs have been detected in persons of all ages, but more frequently in children <5 years of age [13]. Clinical features associated with noroviruses include nausea, vomiting, abdominal cramps, myalgias, headache, fever, chills, sore throat, and non-bloody diarrhea [14–16]. Although NoV transmission occurs year-round in most parts of the world, generally, a cold weather peak was found in The Netherlands, England and Wales, Japan, the US, Australia, Canada, and Denmark [17, 18].

The norovirus genome is positive-sense, single-stranded RNA, approximately 7.7 kb in length [19]. There exists considerable diversity among norovirus strains circulating in different geographical locations and time periods. Five genogroups (GI through GV) of NoVs have been assigned from the molecular characterization of complete capsid gene sequences, which are further subdivided into 27 genotypes [20–22]. Human disease has been associated with GI, GII, and GIV, of which genetic cluster GII.4 are the most predominant strains [23–27]. Recently, a new GII.4 variant with mutation in the polymerase gene circulating in Europe was responsible for 50% of outbreaks in the United Kingdom [14, 28]. Similarly, a novel recombinant GII.3 strain was identified in 2006 in 44% of the clinical samples in Japan [29].

Reverse transcription–polymerase chain reaction (RT-PCR) is currently the most widely used technique for the detection of norovirus from stool, water, and food [30–32]. Several primer pairs have been described that were deduced from highly conserved polymerase gene [33–35]. However, none of the reported conventional RT-PCR assays were able to detect all strains. Sequence analysis of the polymerase region of a wide range of virus strains indicated that this region was also variable, with nucleotide identity as low as 53% between strains of different genogroups and 60–64% within genogroups [36]. In recent years, real-time RT-PCR methodology has emerged as a potentially important diagnostic procedure [37, 38].

Diarrhea is an important public health concern in developing countries, including Bangladesh. Over 100,000 diarrhea patients are admitted at the Dhaka hospital of the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B) each year, and stool samples are investigated in the hospital surveillance system for the presence of different enteric pathogens (Salmonella typhi, non-typhi Salmonella, Shigella dysenteriae, Shigella flexneri, Shigella boydii, Shigella sonnei, Vibrio cholerae O139, Vibrio cholerae O1, and group A rotavirus). It was observed that a huge proportion of samples was left undiagnosed each year. It is possible that other diarrhea-causing agents, such as norovirus, adenovirus, sapovirus, astrovirus, group B rotavirus, group C rotavirus, Aichi virus, etc., which were not included in the hospital surveillance system could have been involved in causing diarrhea hospitalization. To investigate the burden of these unknown pathogens, we aimed to detect noroviruses in these undiagnosed specimens. We also characterized the Bangladeshi NoV strains by sequencing their polymerase genes and compared them with other globally circulating strains to reveal their genetic relationships.

Methods

RT-PCR

Viral RNA was extracted using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. A quantity of 140 µl of stool specimen was taken for extraction and final elution was done with 50 µl RNase-free water. RT-PCR was carried out using the OneStep RT-PCR Kit (Qiagen, Hilden, Germany). Forward primers Mon 431 (5’-TGG ACI AGR GGI CCY AAY-3’) and Mon 432 (5’-TGG ACI CGY GGI CCY AAY-3’) and reverse primers Mon 433 (5’-GAA YCT CAT CCA YCT GAA-3’) and Mon 434 (5’-GAA SCG CAT CCA RCG GAA-3’) were used to amplify 213-nucleotide-long segments from the ORF1 of the norovirus genome [33]. The reaction was carried out with an initial reverse transcription step at 50°C for 30 min with an activation step at 95°C for 15 min, followed by 40 cycles of amplification (30 s at 94°C, 30 s at 52°C, 30 s at 72°C) and a final extension of 10 min at 72°C in a thermal cycler. PCR products were run on a 1.5% ethidium bromide-stained agarose gel and visualized under UV light.

Nucleotide sequencing

The PCR products were purified with the QIAquick PCR Purification Kit (Qiagen, Hilden, Germany) and sequenced using the dideoxynucleotide chain termination method with the ABI PRISM® BigDye Terminator Cycle Sequencing Reaction Kit (Applied Biosystems, Foster City, CA) on an automated sequencer. The same forward primers which were used in the PCR amplification were used as sequencing primers.

DNA and protein sequence analysis

The chromatogram sequencing files were inspected using Chromas 2.23 (Technelysium, Queensland, Australia) and consensus sequences were prepared using SeqMan II (DNASTAR, Madison, WI). Nucleotide and amino acid sequence similarity searches were performed using the National Center for Biotechnology Information (NCBI; National Institutes of Health, Bethesda, MD) BLAST (basic local alignment search tool) server on the GenBank database, release 173.0 [39]. Multiple sequence alignments were calculated using ClustalX 1.81 [40]. Sequences were manually edited in the GeneDoc version 2.6.002 alignment editor [41]. Phylogenetic analyses were conducted using the MEGA version 4.1 software package [42]. The dendrograms were constructed using the neighbor-joining method.

Nucleotide sequence accession numbers

The nucleotide sequences reported in this paper were submitted to GenBank using National Center for Biotechnology Information (NCBI; Bethesda, MD) Sequin version 9.2 and were assigned under accession numbers GU370930–GU370937.

Data analysis

The data were analyzed using SPSS for Windows release 11.5.1 (SPSS Inc., Chicago, IL).

Results

Detection of norovirus

From January 2004 to December 2005, a total of 4,407 diarrhea patients attended the Dhaka hospital of the ICDDR,B. The hospital surveillance system tested 2% samples for several pathogens (i.e., group A rotaviruses, Salmonella typhi, non-typhi Salmonella, Shigella dysenteriae, Shigella flexneri, Shigella boydii, Shigella sonnei, Vibrio cholerae O139, and Vibrio cholerae O1). Any of these pathogens were detected in 2,468 (56%) samples and the remaining 1,939 (44%) were undiagnosed. From these undiagnosed samples, every tenth stool specimen (n = 189) was selected for the presence of NoV by conventional RT-PCR. Norovirus RNA was detected in 37 samples (19.6%), of which 18 were detected in 2004 (n = 100) and 19 in 2005 (n = 89).

Seasonality

The virus was isolated throughout the year and no clear seasonality was observed for NoV infection, as depicted in Fig. 1. However, comparatively high numbers were detected in August–September (rainy season) and December (winter).

Fig. 1
figure 1

Seasonality of noroviruses in Bangladesh

Full size image

Age distribution

NoV was detected in both children and adult patients, with the age ranging from 1.8 months to 50 years. The median age of the norovirus-positive patients was 2 years and the mean age was 14 years. Figure 2 shows the age distribution of the NoV-infected patients. Remarkably, children less than 2 years of age (49% of the NoV-positive patients) and elderly people more than 18 years of age (43% of the NoV-positive patients) were more likely to be infected by NoV. Only three patients were between 2 and 18 years old.

Fig. 2
figure 2

Age distribution of norovirus-infected patients among hospitalized patients in Bangladesh

Full size image

Clinical features

Among the NoV-positive patients, 62% were male. Clinical features of the norovirus-infected patients are shown in Table 1. The illness caused by norovirus infection was mild and was mainly characterized by diarrhea and vomiting. Vomiting was the most common feature and was observed in 83.8% of patients. Fever (19.1%) and abdominal cramp (37.8%) were not frequent in these patients. Severe dehydration caused by norovirus gastroenteritis was found in about 20% of patients and most of them were in the elderly age group (6 out of 7 severely ill patients). Diarrhea was generally moderate; 73% of patients had diarrhea for less than 3 days and 35% of patients had more than ten stools over a 24-h period. Since NoV caused mild dehydration in our study, most of the patients were treated by oral rehydration solution (ORS) (70.3%) only. For severe cases, intravenous fluid (IV) was given additionally to the rest of the patients.

Table 1 Clinical presentations of NoV-infected patients
Full size table

Genetic characterization

The RNA samples which produced sufficient PCR-amplified products (n = 17) were sequenced using the dideoxynucleotide chain termination method. Mon 431 and Mon 432 primers produced at least 100 nucleotide bases from the 5’-end of the polymerase gene. Nucleotide similarity searches were done using the NCBI BLAST server on GenBank database release 130.0 [39]. Similarity searches indicated that three (17.6%) were NoV genogroup GI and 14 (82.4%) were genogroup GII (Table 2). Among the GI strains, two were most similar to genotype GI.1 and one was GI.3. Among the GII strains, four were GII.4 (28.6%) and nine were GII.7 (64.3%). One sample was found to be mixed with GI and GII genogroups.

Table 2 Genotype distribution of noroviruses in hospitalized patients
Full size table

BLAST analysis indicated that the GI Bangladeshi strains Dhaka54 and Dhaka67 were most similar to Indian V16/06/IND (94% nucleotide and 100% amino acid identity), Russian 6836/Chelyabinsk/RUS (96% nucleotide and 100% amino acid identity), and GI.1 reference US NoV strain West Chester/2001/USA (96% nucleotide and 100% amino acid identity). Another GI isolate, Dhaka23, had 98% nucleotide and 100% amino acid identity with Russian strain 11227/NizhnyNovgorod/RUS. Nucleotide identities with other strains were low (less than 85%). An Indian strain Kolkata/L8775/2006/IND had 80 and 98% identities with Dhaka23 based on nucleotides and amino acids, respectively.

The GII strains isolated in Bangladesh were divided into two genotypes, GII.4 and GII.7. Based on amino acid sequences, isolates Dhaka58 and Dhaka100 were most similar to GII.4 Indian strain V1699/07/IND, as well as strains from all over the world (100% identity). The most similar strain based on nucleotide identity was Canadian Manitoba/4205/2003/ CAN. On the other hand, Dhaka55 and Dhaka97 were most similar to GII.7 East Asian norovirus strains. Absolute identity on amino acid sequences was identified with Japanese Saitama U25, Chinese Shanxi/50106/2006, and Thai strain Mc17/2002/Th. Isolate Dhaka42 was placed a little farther from them and was most similar to Brazilian strain 5037/2001/Bra (99% nucleotide and 100% amino acid identity).

Three GI sequences and five GII sequences were included in the phylogenetic trees, along with globally circulating NoV strains (Figs. 3 and 4). GI Bangladeshi strains belonged to two different clusters. Dhaka54 and Dhaka67 were closely related to a US strain WestChester of GI.1 lineage. The third Bangladeshi GI strain, Dhaka23, did not cluster closely with any genotype, although the mostly related genotype was GI.3. Bangladeshi GII NoVs clustered with two different genotypes. Dhaka58 and Dhaka100 were closely related to the GII.4 cluster and Dhaka97 and Dhaka55 were closely related to the GII.7 cluster.

Fig. 3
figure 3

Phylogenetic tree based on partial nucleotide sequences of the polymerase gene of norovirus genogroup GI. The tree was constructed by the neighbor-joining method. The numbers adjacent to the nodes represent the percentages of bootstrap support (of 1,000 replicates) for the clusters to the right of the nodes. Bootstrap values lower than 75% are not shown. Bangladeshi strains are in bold. BGD, Bangladesh; USA, United States of America; JPN, Japan; CAN, Canada; NZL, New Zealand; BRA, Brazil; GBR, Great Britain; DEU, Germany; DJI, Djibouti; FRA, France. The GenBank accession numbers were: West Chester (AY502016), Saitama T23cGI (AB112094), Saitama KU19dGI (AB058528), Saitama T24cGI (AB112098), Saitama T2GI (AB112103), Little Rock316 (AF414405), NoV69 (EF078281), DjiboutiVdG50 (EF190919), ICB1969C2 (DQ386978), Wyoming (AY210317), PontdeRoide682 (EF529736), Berlin/92617.04-3-BA2 (DQ340078), Baltimore/277 (AF414404), Saitama T24dGI (AB112099), NoV748 (EF078287), BCCDC04003 (DQ452547), Wisconsin (AY502008), Queen’s Arms/Leeds (AJ313030), Appalachicola Bay/318 (AF414406), Cardrona (EF527258), Berlin/92674.04-5-BA2 (DQ340080), ICB1918 (DQ386948), New Orleans/266 (AF414402), Berlin/90068.04-8-BA4 (DQ340083), Steinbach/EG (AF473567), and an outgroup sequence from a GII norovirus Saitama T24eGII (AB112306) is included

Full size image
Fig. 4
figure 4

Phylogenetic tree based on partial nucleotide sequences of the polymerase gene of norovirus genogroup GII. The tree was constructed by the neighbor-joining method. The numbers adjacent to the nodes represent the percentages of bootstrap support (of 1,000 replicates) for the clusters to the right of the nodes. Bootstrap values lower than 75% are not shown. Bangladeshi strains are in bold. AUS, Australia; NLD, The Netherlands; IRE, Ireland; VIE, Vietnam; ITA, Italy; HUN, Hungary; CHN, China. Reference strains of GII genotypes were collected from GenBank and the accession numbers of the reference strains are as follows: Hawaii (U07611), PC301 (AF414421), MOH99 (AF397156), NO306 (AF414422), SU16 (AB039778), SU25 (AB039780), SW918 (AB074893), SU1 (AB039775), HLL314 (AF414420), Tiffin (AY502010), CSE1 (AY502009), Manitoba4205 (DQ463420), OCS980730 (AB089871), Sydney348-97O (DQ078829), Gwynedd273 (AF414409), 691 (AF493210), Terneuzen70 (EF126964), BCCDC04007 (DQ452543), ICB2162 (DQ386959), NoV809 (EF078293), Rhyl440 (DQ822040), Carlow (DQ415279), Kunming146 (DQ304651), NV/Osaka/F140 (AB258331), Minato/N1-14 (AB233474), Burwash Landing331 (AF414425), Dresden174-pUS-NorII (AY741811), Miami81 (AF414416), and Vietnam 026 (AF504671), with an outgroup GI Baltimore/277 (AF414404)

Full size image

Discussion

From 1993 to 2004, among diarrhea patients less than 5 years of age who attended the Dhaka hospital of the ICDDR,B, n = 18,544 were tested for several pathogens. Rotavirus constituted 33%, bacterial agents, including V. cholerae, Shigella, and Salmonella, 21%, and parasites 2% [43]. Remarkably, a major proportion (44% of the samples tested) was undiagnosed. Our results show that NoVs constituted 19.6% of these undiagnosed samples and are a common diarrhea-causing agent which necessitates hospitalization. Overall, the prevalence of NoV in all pathogen-positive and -negative samples (n = 440) was 8.4%. The overall NoV detection rate of 8.4% in our study is similar to other studies which report NoV prevalence in the range 6–19% [44–52]. The detection rate in our study, however, might be underestimated because the primer set that we used might not be able to detect all of the different strains. It is noteworthy that designing one set of primers to detect all NoV strains with equal efficiency is difficult because the mutation rate in the NoV genome is very high [53, 54].

Norovirus was found in both children and adults in our study. Similarly, reports from all over the world also indicated that NoVs could infect both children and adults [55–58]. Clinical features of the NoV-infected patients were characterized by diarrhea, abdominal cramp, vomiting, and fever. Although hospitalized, the illness of most of our NoV-infected patients was not severe, which was supported by the previous investigations [59–61]. Treatments of the patients were carried out by ORS in most cases and a few severe patients were treated with IV.

It is established that GI, GII, and GIV are the major norovirus genogroups isolated worldwide causing human gastroenteritis [20]. In the present study, we also identified norovirus genogroups GII and GI, although GII strains were predominant over GI strains. Genogroup GI was present in only small numbers in our study period, which is also similar to other studies conducted during the same period [11, 24, 54, 62–64]. It is interesting to note that a previous study conducted in Bangladeshi children less than 38 months of age found only the GII.4 genotype [59]. The GII.4 strains identified in the present study were also found in infants and children (aged 4 to 12 months). No GIV strain was detected in our study, although this genogroup has been circulating in humans worldwide [20, 65].

One of the goals of this study was to compare Bangladeshi strains with other noroviruses isolated all over the world to reveal their origin and genetic relationships. Phylogenetic analysis indicated that the Bangladeshi strains did not cluster with strains from a particular region; instead, strains from different countries clustered in the same branch. This indicates that the global dispersion of the viruses occurred between different geographical locations. GI strain Dhaka23 was an interesting strain, since it was placed in a separate branch distantly related to other GI genotypes (Fig. 3). Since we analyzed a small portion of the polymerase gene, complete genome sequence analysis would be required for detailed characterization of this strain.

The study had several limitations. The sample size was small and was collected over a 24-month period. Longer, longitudinal studies are required to address issues such as norovirus seasonality and genetic variability and to monitor the spread and persistence of the various genotypes circulating in Bangladesh. In addition, we sequenced a small proportion of the norovirus genes which might have an impact on genotyping of the strains described in this paper. The use of short conserved sequences, although successful for the diagnosis of norovirus infection, should be used with caution for classification and phylogenetic analyses. Therefore, an expanded genetic study is required to genotype the strains accurately, as well as to detect new and novel strains, and further analysis with full genome sequences might be more helpful. Since the study included only hospitalized patients, community-based studies are required in order to investigate the true burden of the disease caused by this pathogen. In conclusion, this study provides information for the future epidemiological studies of noroviruses, which is required for the control and prevention of diarrheal diseases, as well as norovirus vaccine development programs.