Characterization of Norovirus-Associated Traveler’s Diarrhea

Background. Traveler’s diarrhea is the most common medical complaint of international visitors to developing regions. Previous ﬁndings suggested that noroviruses (NoVs) are an underappreciated cause of traveler’s diarrhea. Methods. In the present study, we sought to deﬁne the presence of NoVs in 320 acute diarrheic stool samples collected from 299 US students who traveled to Guadalajara, Cuernavaca, or Puerto Vallarta, Mexico, during the period from 2007 through 2008. Conventional and quantitative real-time polymerase chain reaction assays were used to detect and determine NoV loads in stool samples. NoV strains were characterized by puriﬁcation of viral RNA followed by sequencing of the viral capsid protein 1 gene. Sequences were compared using multiple sequence alignment, and phylogenetic trees were generated to evaluate the evolutionary relatedness of the viral strains associated with cases of traveler’s diarrhea. Results. Twelve

and healthcare settings [10][11][12][13]; in developing countries, malnourished children or those without access to effective healthcare are at a higher risk of suffering significant morbidity and mortality [6].
Traveler's diarrhea (TD) is the most common medical complaint of international visitors to tropical and semitropical regions [14][15][16], and although TD is most often caused by bacteria [17], travelers have been previously identified as a group at risk of contracting NoVassociated diarrhea [18][19][20]. NoVs have been shown to be the most commonly identified enteric pathogen (after diarrheagenic Escherichia coli) in studies on travelers to Mexico [19,20].
Our research team monitors groups of US students traveling to Mexico as a model of TD, and previous findings suggested that NoVs were an underappreciated cause of TD marked by a continuous and diverse presence of viruses belonging to GI and GII [18][19][20]. An explanation as to why travelers are highly affected by GI strains is still missing, and little is known about naturally-occurring infection in locations where NoVs are endemic, including populations of international travelers to endemic areas. The aims of this study were to determine the frequency of NoV-associated TD in US students traveling to Mexico, characterize the strains responsible for causing disease, determine the cDNA viral load, and characterize the clinical symptoms associated with NoV-associated TD.

Study design and case definitions.
Adults of at least 18 years of age who were experiencing TD acquired in Cuernavaca, Guadalajara, or Puerto Vallarta, Mexico, during the years of 2007 and 2008 were evaluated for the presence of NoVs in diarrheal stools. TD was defined as the passage of у3 unformed stools within a 24-h period associated with у1 of the following symptoms: abdominal pain or cramps, excessive gas or flatulence, nausea, vomiting, fever (temperature, у100ЊF [у37.8ЊC]), fecal urgency, blood and/or mucus in the stool, or tenesmus. Once TD was diagnosed, a stool sample was obtained; it was verified as being unformed stool by a clinic staff member and submitted to the microbiology laboratory for study. Exclusion criteria included previous enrollment in our study, a duration of diarrhea for 172 h, and the presence of a clinically important underlying illness other than diarrhea. For safety reasons, women who were pregnant or breast feeding were excluded from our study because most of the students enrolled in epidemiologic studies also participate in clinical trials. In addition, patients were excluded if they had taken any antimicrobial agent effective against diarrheal pathogens within the prestudy week or any antidiarrheal agent within 12 hours of enrollment. All patients provided written informed consent. The study was approved by the University of Texas Health Science Center Committee for the Protection of Human Subjects.
Stool samples. All study participants were requested to submit stool samples to the local travel clinics within 48 h after onset of diarrhea. Additional stool samples were collected from participants during any subsequent acute diarrheal episodes. Stool samples were processed as reported elsewhere [17]. Aliquots of diarrheal stools were stored at Ϫ20ЊC and shipped on dry ice to our laboratory at the Center for Infectious Diseases, University of Texas, School of Public Health, Houston, Texas, for norovirus studies.
Collection of clinical information. Participants recorded on weekly diary cards the presence or absence of a number of gastrointestinal symptoms, along with the number, time, and characteristics of all stools passed during their stay in Mexico. Study subjects rated their signs and/or symptoms on the basis of the following severity scores: 0, no symptoms; 1, mild symptoms (tolerable, without interfering with normal activities); 2, moderate symptoms (distressing, forcing changes in normal activities); 3, severe symptoms (incapacitating, prohibiting performance of normal activities).

Detection and characterization of NoVs.
A 10% (w/v) stool suspension was prepared with 0.01 mol/L phosphate-buffered saline and clarified by centrifugation at 3000 rpm (2000g) for 10 min. Viral RNA was extracted from 200 mL of the 10% stool suspension with a QIAamp viral RNA kit (Qiagen) according to the manufacturer's instructions, and the RNA was stored at Ϫ80ЊC until used.
Reverse transcription (RT) was carried out in 50 mL of a reaction mixture containing 75 pmol of random hexamers (Thermo Fisher Scientific), 5 U of AMV-RT (Life Sciences), 1ϫ AMV-RT buffer (Life Sciences), 10 mmol/L of dNTP mix, 10 U of Protector RNase Inhibitor (Promega), and 10 mL of RNA extract. The reaction was performed for 2 h at 43ЊC, and the enzyme was inactivated at 75ЊC for 5 min. The resulting cDNA was stored at Ϫ20ЊC until used in polymerase chain reaction (PCR) assays.
A set of degenerate primers [21] was used for initial detection of NoVs. The PCR reaction mixture contained 5 mL of template cDNA, 5 U of AmpliTaq (Applied Biosystems), 1ϫ AmpliTaq buffer I (Applied Biosystems), 10 mmol/L of dNTP mix, and 200 nmol/L of each primer. A genogrouping PCR assay was performed on positive samples using primers G1SKF, G1SKR, G2SKF, and G2SKR [22]. All PCR products were separated by 2% agarose gel electrophoresis and visualized under UV lamp after ethidium bromide staining.
Nucleotide sequence and phylogenetic analysis. PCR products resulting from the amplification with the GSK primer set were used to sequence the ORF1-ORF2 junction region of all GI NoVs and GII NoVs identified. PCR products were purified using a MinElute PCR purification kit (Qiagen) according to the manufacturer's instructions. Viral DNA was sent to SeqWright (Houston, TX) for sequencing. Sequences were analyzed using BioEdit, version 7.0.9.0, and aligned with Clustal W 2.0.10 (http://www.ebi.ac.uk/clustalw) with default parameters. Phylogenetic trees were constructed using the neighborjoining and Poisson correction methods [23,24] by use of MEGA, version 4.0 (http://www.megasoftware.net) [25]. The significance of the taxonomic relationships was obtained by bootstrap analysis (1000 replications) [26]. The reference strains described by Zheng et al [27] were used in the assignment of genotypes. The NoV sequences identified in the 2007-2008 Mexico cases were compared with those found in earlier studies in Mexico during 1998 and 2004 [19,28] to determine relatedness of the viruses in 1 region over time.
Quantitative real-time PCR. The quantitative real-time PCR was carried out in triplicate on an ABI StepOne Real-Time PCR system (Applied Biosystems) using purified NoV GI or GII cDNA plasmids as standards and following the protocol described by Kageyama et al [29]. Amplification data were analyzed with StepOne Real-Time PCR System software, version 1.0 (Applied Biosystems). was P ! .05 considered to be statistically significant. Overall, 17 travelers experienced recurrent diarrhea, and the maximum number of samples corresponding to discrete episodes was 2 samples per subject. Travelers ranged in age from 19 to 45 years of age (mean age, 24 years; median age, 22 years); 72% of travelers were female, and 96% were white.

Sample collection.
Prevalence of NoV among TD patients. In the present study, 30 (9.4%) of 320 stool samples collected had detectable levels of NoV RNA, and 27 (9%) of 302 travelers were infected by NoV. Most travelers had single bouts of NoV diarrhea, and only 3 travelers experienced 11 episode of diarrhea in which NoV was recovered (Table 1).
NoV infection was more prevalent among travelers to Cuernavaca than among travelers to Guadalajara (15.8% vs 3.6% of travelers; ) Clinical symptoms associated with NoV-associated TD. Clinical data were available on 23 of the 27 travelers who experienced NoV-associated TD. The majority of travelers recorded clinical symptoms representing a typical TD episode. Data on 3 travelers who had second independent episodes of diarrhea were excluded from the analysis because most travelers received treatment drugs after initial episodes and because this patient population also takes part in pharmaceutical clinical trials. Among the recorded symptoms, median scores suggest that abdominal cramping, flatulence, and vomiting were experienced at a mild level, whereas moderate levels of nausea and fecal urgency were recorded. None of the subjects reported experiencing fever during the illness. NoV-associated TD in the NoV single infection group (6 subjects) was characterized by a mild level of flatulence and moderate-to-severe levels of nausea, abdominal cramping, vomiting, and fecal urgency, whereas all symptoms were recorded to be at a mild level in the NoV coinfection group (11 subjects). We found that cases of NoV single infection were associated with more severe vomiting, compared with cases of NoV coinfection with more than 1 pathogen ( ). Ninety percent of the stool samples col-P p .047 lected from ill travelers with NoV-associated TD were watery; the remaining 10% had a soft consistency.
Phylogenetic analysis of nucleotide sequences was carried out for NoV cases by season and location. All GI NoV-associated TD cases were identified during 2007. A single GI.8 strain was found to cause disease in 9 different travelers in 2 cities during the same season. Only 1 positive sample was recovered from a traveler during spring. Four different GII.4 strains were associated with 6 NoV-associated TD cases in travelers to Cuernavaca. In Guadalajara, 4 travelers had GI NoV-associated TD (3 GI.8 and 1 GI.3), and 1 traveler had a GII virus (GII.5). All 11 episodes of diarrhea observed during the winter months were associated with viruses belonging to GII, whereas all 12 of the GI strains identified were isolated from patients who had traveled during the summer months of 2007. All the NoV-  [30]. Sequences listed as TD (for traveler's diarrhea), followed by sample number, location, and year, were taken from Ko et al [19], and the sequences listed by accession numbers DQ220751-DQ220757 were taken from Garcia et al [28]. Percentage bootstrap values are shown at the branch nodes. The scale indicates nucleotide substitutions per site. The partial capsid sequence alignment of the GI.1 strain described in the present study showed a high degree of similarity with a subset of GI.1 viruses previously identified among a population of travelers to Guadalajara in 2004 [19]. Less similarity was observed with other GI.1 strains identified among a pediatric population living in a periurban location located along the eastern perimeter of Mexico City [28] (Figure 2), which shows strain diversity across time.
NoV cDNA viral loads among affected travelers. The median cDNA viral load of NoV GI (for 11 subjects) detected in the fecal samples was genome copies per gram 10 1.73 ϫ 10 (range, genome copies per gram) (Figure 7 13 3.76 ϫ 10 -1.18 ϫ 10 3). The median cDNA viral load of NoV GII (for 12 subjects) detected in the fecal samples was genome copies per 7 3.43 ϫ 10 gram (range, genome copies per gram) 5 9 3.26 ϫ 10 -8.93 ϫ 10 ( Figure 3). The median GI NoV fecal shedding was found to be 500-fold higher than that of GII ( . These results suggest that NoVs may play a role P p .691 in the development of TD. Furthermore, we found no association between severity of clinical symptoms and intensity of cDNA viral load (data not shown).

DISCUSSION
In line with our previous results showing that NoVs are a significant agent of TD [19,20], the study reported herein has shown a prevalence rate of 9% for travelers to Mexico experiencing NoV-associated diarrhea. Although the prevalence rate observed here is lower than that reported in previous years [18][19][20], the present study represents a multiseasonal analysis of travelers during 2 consecutive years in 3 different locations. The role of NoVs in travelers' illnesses could be even greater, because stool samples collected and analyzed in the present study were only from travelers with acute diarrhea and not from travelers who had a vomiting illness and no diarrhea in the presentation of their gastroenteritis. This observation should be taken into consideration for further studies of the true NoV prevalence among ill travelers, because the onset of vomiting and the absence of diarrhea can be seen after NoV infection [31].
Consistent with previous reports in which NoV infection had been referred to as a "winter-vomiting disease" [32], the highest prevalence of NoV infections occurred from December to February, which is considered the winter season in Mexico, and GII.4 strains were responsible for 55% of the cases, which correlates with previously published data showing that winter outbreaks were primarily due to GII.4 strains [33,34]. However, it is important to mention that it was only during the winter of 2008 in Cuernavaca that we observed a prevalence rate of NoV that was higher than the prevalence rates observed in other seasons. Interestingly, all the NoV-associated TD episodes observed during the winter months were associated with viruses belonging to GII, whereas all the GI strains identified were isolated from patients who had traveled during the summer months of 2007.
As observed in the past (unpublished data) and in this present study, travelers to Cuernavaca are usually more commonly affected by NoV-associated diarrhea than those who travel to Guadalajara. A possible explanation for this finding might be that Cuernavaca is closer to Mexico City, the most populated city in Mexico, and it is occupied by a large population of people who commute to Mexico City on a daily basis. Guadalajara, on the other hand, is a larger urban city inhabited by people who usually do not commute to other cities to work, thus decreasing the possibility of introducing new pathogens into the community.
In line with previous reports [18,19], we observed a strong and diverse presence of GI viruses despite the predominance of GII.4 viruses in most outbreaks [35][36][37][38][39][40][41][42]. Although our results demonstrate a great diversity of NoV strains affecting travelers (12 [40%] of 30 samples positive for GI NoV and 18 [60%] of 30 samples positive for GII NoV), an outbreak caused by a single GI.8 virus affected 9 different travelers during the summer months of 2007 in both Cuernavaca and Guadalajara, which means that 75% of the NoV-associated TD cases were attributed to GI.8 viruses.
In our study, we show that the median cDNA viral load of NoV GI is 500-fold higher than that of GII in fecal samples of travelers with NoV-associated TD. Differences in viral loads may be caused by experimental and host variables, including time of sample collection after infection, affinity of annealing of primers and probes to diverse viral genome sequences, age of the affected person, host immune response, and course of infection. However, on the basis of our results, we speculate that the increased cDNA viral load observed in NoV-GI samples had an impact on the transmission of these strains among susceptible hosts by the fecaloral route, partially explaining the continuous presence of GI strains in travelers to these locations, as previously reported by our group and others [18,19].
The majority of clinical symptoms reported by patients with NoV-associated TD did not differ from the symptoms that have been reported by patients with TD caused by other common bacterial pathogens [19,43,44]. Nonetheless, vomiting was more often associated with NoV single infections.
Our study demonstrates that the vast diversity of NoVs found among travelers to Mexico play an important role in travelers' health, posing a challenge for vaccine development. Prevention methods such as vaccination should focus on eliciting crossreacting antibodies that will protect individuals against a diverse population of NoVs, because natural infection could indicate the presence of cocirculating strains, as shown in our study.