Interlaboratory Evaluation of a Method for Quantification of Norovirus RNA as an Alternative Use for ISO 15216-1:2017 to Conduct Japan Baseline Survey of Oysters

Abstract

In this study, we aimed to investigate the standard method used for quantification of norovirus in oysters in Japan for the provisional adaptation of the method as an alternative to ISO 15216-1:2017, to conduct a Japan baseline survey of norovirus in oysters. For this purpose, the method provided by the Japan Committee for Standardization of Virus Detection in Food was subjected to an interlaboratory study to determine the performance characteristics of the standard method used in Japan. As a result, the theoretical limit of quantification for norovirus GI and GII in oysters by the standard method used in Japan was expected to be 1.92 and 1.85 log₁₀ copies/g, respectively. The repeatability standard deviations (Sr) were 0.26 and 0.30 log₁₀ copies/g for GI and GII, respectively, and the reproducibility standard deviations (S_R) were 0.47 and 0.44 log₁₀ copies/g for GI and GII, respectively. Through the interlaboratory study, we specified several critical points to obtain scientifically reliable results by using the standard method used in Japan. Especially, necessity for application of using process control virus was the most crucial point that needed to be improved. In addition, there are many participating laboratories that could not handle dilution of standard and quantify or detect the viruses in the test samples. To ensure scientifically reliable test result, capacity building of laboratories and implementation of proficiency testing should be considered for future tasks in combination with an application of process control materials in the method. On the assumption that the problems revealed in this study will be solved, the standard method used in Japan would be suitable for use in Japan baseline survey of norovirus in oysters, which will contribute to the international action against norovirus in oysters, led by the EU.

Introduction

Oysters contaminated with noroviruses pose a particular risk to human health because they are routinely consumed raw (Le Guyader et al., 2000). The European Food Safety Authority (EFSA) panel on Biological Hazards (BIOHAZ) recommended that risk managers should consider establishing an acceptable limit for norovirus in oysters to be harvested and placed on the market (EFSA, 2012).

A European project, launched in 2004, to develop a standardized method to detect norovirus in a variety of food matrices, including oysters, resulted in the publication of a two-part technical specification for determination of the viruses in food matrices in 2013 (International Organization for Standardization/Technical Specification [ISO/TS]) (ISO/TS 15216-1:2013). To replace the technical specification with the full, validated EN/ISO standard, the performance characteristics of the method were determined in an interlaboratory study (Lowther et al., 2019), and the method has been adopted for the European baseline survey of norovirus in oysters to estimate the European prevalence of norovirus in raw oysters (EFSA, 2016).

A problem identified so far in Japan is the lack of reliable methods for the detection and quantification of norovirus in oysters, except for ISO 15216-1:2017. Because of the unavailability of commercial products (e.g., BioMerieux NucliSens^® miniMAG or easyMAG), which were used for the optimization of ISO 15216-1:2017, there are no private food analysis agencies that adopt ISO 15216-1:2017 to quantify norovirus in food for regulatory purposes in Japan.

Although we also recognize that ISO 15216-1:2017 is able to reproduce by using not only commercial kits but also combining other individual materials, lack of proficiency testing system makes it difficult to assess whether the test result obtained by conducting ISO 15216-1:2017 reproduced by ourselves will be scientifically reliable. Moreover, most of private food analysis agencies in Japan traditionally perform quantification of norovirus in oysters without using a process control virus, because they do not recognize the importance of process control in norovirus testing. Or even if there is recognition of the importance to manage a process of testing, they have no facilities, apparatus, and experiences to prepare process control materials by themselves as described in ISO 15216-1:2017.

In contrast, there is the standard method adopted in Japan by private food analysis agencies for voluntary inspection of norovirus in oyster to determine whether a certain harvested oyster lot is suitable for releasing oysters intended for raw consumption. Thus, in this study, we aimed to clarify the performance characteristics of the standard method used in Japan for considering the alternative application of the method to our future task, so-called Japan baseline survey of norovirus in oysters, to estimate the prevalence of norovirus in raw oysters.

Materials and Methods

Preparation of test samples for the interlaboratory study

Oysters were purchased directly from Momonoura producer of oyster consolidated company in Miyagi Prefecture, Japan. The digestive tissues (DTs) were dissected from 480 oysters and pooled after homogenization using a glass homogenizer. Homogenized DTs (HDTs) were subjected to RNA extraction and real-time PCR (RT-PCR) analysis to investigate whether the oysters were naturally contaminated with norovirus.

The norovirus GI and GII strains used in this study were purified from stool specimens of patients with acute infectious gastroenteritis in Miyagi Prefecture, Japan. This study used both GI.4 and GII.17, which were revealed using RT-PCR and sequencing analysis. Norovirus quantification was conducted according to the method described as below.

Three different levels of test samples were prepared for the interlaboratory study. Briefly, for test samples B and C, HDTs were mixed with both norovirus GI- and GII- diluted solution in phosphate-buffered saline (PBS), to obtain 1000 or 5000 copies of norovirus per 1 g of HDTs. Then, the resulting HDTs were diluted by nine times volume of PBS. The final mixture was divided into 5 mL as test samples B and C. Virus dosing was calculated to provide detectable levels of 16 and 80 copies/5 μL of complementary DNA (cDNA) at contamination levels of test samples B and C, respectively. For test sample A, HDTs, which was not spiked with norovirus, were mixed with PBS so that 1 g of HDTs was diluted by nine times volume of PBS. Then, the mixture was divided into 5 mL as test sample A as unspiked material. All test samples were frozen at −80°C until use.

Levels of norovirus contamination in the test samples

To evaluate the levels of contamination prepared in the test samples, more than 10% of the test samples were randomly selected from each group (20, 10, and 20 for A, B, and C, respectively), and were tested in duplicates by the expert laboratory, according to the Food Analysis Performance Assessment Scheme (FAPAS) guidelines (2002), introducing homogeneity tests. One-way analysis of variance (ANOVA) was used to confirm homogeneity of norovirus concentration in each test sample.

To confirm the stability of the test samples, 10 samples each were randomly selected from each group and tested in duplicates 10 weeks after preparation. The rationale behind 10 weeks is because we designed the interlaboratory test to be completed at least within 10 weeks. Then, one-way ANOVA was used to analyze the variance of norovirus in each test sample group.

Interlaboratory study

An interlaboratory study was designed in accordance with the Association of Official Agricultural Chemists (AOAC) international Methods Committee Guidelines for Validation of Microbiological Methods for Food and Environmental Surfaces (AOAC, 2012). Six frozen anonymized test samples (duplicate samples for each contamination level) were distributed to each of the 14 participating laboratories. The expert laboratory where they prepared test samples did not participate in the interlaboratory study. All participating laboratories used the same kit and materials as described in the next section, “Standard Method Used in Japan.” Apparatus for real-time polymerase chain reaction (PCR) was not coordinated between participating laboratories. Each participating laboratory tested six samples within 5 weeks after distribution. Laboratories were permitted to store frozen test samples before testing, but were instructed to return data to the MAFF (Tokyo, Japan) by a specified date, 6 weeks from sample reception.

Standard method used in Japan

The test samples were incubated with α-amylase (017-26371; Fujifilm Wako Pure Chemical, Osaka Japan) at a final concentration of 2.5 mg/mL for 1 h at 37°C, and vortexed every 15 min. After centrifugation at 8000 × g for 20 min, the supernatant was recovered. To concentrate the virus, polyethylene glycol 6000 (Fujifilm Wako Pure Chemical) and sodium chloride (Fujifilm Wako Pure Chemical) were added at a final concentration of 12% and 5.8%, respectively. Eighteen hours after incubation at 4°C, the samples were centrifuged at 8000 × g for 20 min. After removing the supernatant, the pellet was resuspended in 250 μL of sodium dodecyl sulfate (SDS) Tris-Glycine buffer (25 mmol/L Tris, 192 mmol/L glycine, 0.1% [w/v] SDS; Nippon Gene, Toyama, Japan). The resulting solution was transferred into a 1.5 mL tube and centrifuged at 8000 × g for 5 min at 4°C. The supernatant was immediately used for RNA extraction.

Viral RNA was extracted from 200 μL of viral suspension using a High Pure Viral RNA kit (Roche Diagnostics K.K., Tokyo, Japan) with recombinant DNase I (Roche Diagnostics K.K.), according to the manufacturer's instructions, with slight modification that 8 μL of synthetic RNA (Nippon Gene) was added to 400 μL of the binding buffer to promote the recovery of RNA. RNA was eluted by 50 μL of the elution buffer. cDNA synthesis of the first strand with random primers was performed using a High-Capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific, Tokyo, Japan), according to the manufacturer's instructions.

TaKaRa quantitative PCR (qPCR) Norovirus (GI/GII) Typing Kit (Takara-Bio, Kusatsu, Japan) was used for quantification of norovirus in samples, according to the manufacturer's instructions with slight a modification that the volume of the template cDNA used for quantification was changed from 2.5 to 5 μL. All amplification of the test samples was performed in duplicate, while that of dilution series of the positive controls from 10¹ copies/reaction to 10⁶ copies/reaction was performed in triplicate. The positive control in the TaKaRa qPCR Norovirus (GI/GII) Typing Kit (Takara-Bio) was plasmid DNA encoding target sequences for the primers and probe set designed based on the previous report (Kageyama et al., 2003) (Table 1).

Table 1.

Primes and Probes for Norovirus

GI
COG1F	5′-CGY TGG ATG CGN TTY CAT GA-3′
COG1R	5′-CTT AGA CGC CAT CAT CAT TYA C-3′
RING1-Ta(a)	5′-FAM-AGA TYG CGA TCY CCT GTC CA-TAMRA 3′
RING1-TP(b)	5′-FAM-AGA TCG CGG TCT CCT GTC CA-TAMRA 3′
GII
COG2F	5′-CAR GAR BCN ATG TTY AGR TGG ATG AG-3′
ALPF	5′-TTT GAG TCC ATG TAC AAG TGG ATG CG-3′
COG2R	5′-TCG ACG CCA TCT TCA TTC ACA-3′
RING2AL-TP	5′-FAM-TGG GAG GGS GAT CGC RAT CT-TAMRA 3′

Generation of method performance characteristics: theoretical limit of quantification, repeatability, and reproducibility standard deviations

The theoretical limit of quantification (tLOQ) and repeatability and reproducibility standard deviations (Sr and S_R) were determined using the data generated in the interlaboratory study. First, the standard curve was created from the results for dilution series of the positive control in the TaKaRa qPCR Norovirus (GI/GII) Typing Kit (Takara-Bio) by plotting the Cq values obtained against log₁₀ concentration. Then, R ² (where R is Pearson's correlation coefficient), slope, and intercept parameters were determined from the standard curve. The Cq values were not averaged from triplicate reactions before plotting. The same acceptable criteria of standard curves from ISO 15216-1:2017 were adopted for this study. Among the entire data set returned from the individual participant laboratory, standard curve with R ² values of <0.980, or where the slope was not between −3.10 and −3.60 (corresponding to amplification efficiencies of ∼90% to 110%), was excluded from data analysis as outlier.

The tLOQ was calculated based on the assumption that at least 1 copy of target genome was in a one reaction mixture of real-time PCR. Based on this assumption, it was theoretically necessary to contain more than 62.5 copies of the target genome in the 1 g of DTs, if recovery of the target genome was assumed as 100%. However, 100% recovery was not practically feasible. Then, the values were adjusted by using recovery of norovirus from test samples. For calculation of recovery of norovirus, quantification values of norovirus in each test sample were divided by amount of norovirus added in each test sample. For example, when a real-time PCR value that comes from a sample C (expected norovirus dosing of sample C was 3.70 log₁₀ copies/g) was 3.18 log₁₀ copies/g, recovery of norovirus was calculated as around 85.9%. The Sr and S_R were calculated according to the formulae provided in the AOAC guidelines (AOAC, 2012).

Results

Levels of norovirus contamination in the test samples

Levels of contamination for norovirus GI and GII in the test samples are shown in Tables 2 and 3. Norovirus was not detected in test sample A, which was not spiked with norovirus. Variance analysis was conducted to confirm homogeneity of norovirus within test samples B and C. As a result, no significant difference (p > 0.05) was observed within each contamination level, suggesting appropriate levels of homogeneity of norovirus in individual groups. The test results for GI obtained from test samples B and C were not significantly different (p > 0.05) between the groups, suggesting that the contamination level of norovirus GI between test samples B and C was almost equivalent. On the other hand, the test results for GII obtained from test samples B and C were significantly different (p < 0.02) between both groups. Contamination levels in samples A, B, and C were not affected by frozen storage for 10 weeks and were not significantly (p > 0.05) different from the initial level after freezing (data not shown).

Table 2.

Levels of Contamination for Norovirus GI in Oyster Test Samples

	Contamination level
	Low (A)	Medium (B)	High (C)
No. of samples tested	20	10	20
Mean value (log₁₀ copies/g)	ND	2.47^a	2.42^b,c
SD of value (log₁₀ copies/g)	—	0.15	0.21

p-Value of variance analysis in test sample B was 0.24.

p-Value of variance analysis in test sample C was 0.74.

p-Value of variance analysis between test sample B and C was p = 0.47.

ND, not detected; SD, standard deviation.

Table 3.

Levels of Contamination for Norovirus GII in Oyster Test Samples

	Contamination level
	Low (A)	Medium (B)	High (C)
No. of samples tested	20	10	20
Mean value (log₁₀ copies/g)	ND	2.37^a	2.90^b,c
SD of value (log₁₀ copies/g)	—	0.26	0.23

p-Value of variance analysis in test sample B was 0.72.

p-Value of variance analysis in test sample C was 0.52.

p-Value of variance analysis between test sample B and C was p < 0.02.

ND, not detected; SD, standard deviation.

Interlaboratory study

Results were submitted from all 14 participants. However, 6 out of 14 laboratories for GI and GII reported that the slope obtained from standard curve was not between −3.10 and −3.60. Thus, the test results from these six laboratories were excluded from analysis as outliers. No laboratory reported that R ² values obtained from standard curve was <0.980. As shown in Tables 4 and 5, it was observed that several laboratories failed to obtain quantification values from test samples B and/or C, even though the data set from these laboratories met acceptable criteria of the standard curves. Substitution of quantitative values was not accepted in this study. Thus, remaining valid values were used for the analysis.

Table 4.

Interlaboratory Study Results for Norovirus GI in Oyster Samples

	Contamination level
	Negative (A)	Medium (B)	High (C)
Total data sets^a	14	14	14
No. of outliers^b	6	6	6
No. of laboratories after elimination outliers	8	8	8
No. of laboratories that failed to obtain quantification values of test sample B and/or C	—	1	2
No. of laboratories that submitted test result adopted for calculation	7	7	6
Average of the slope of the standard curves^c	−3.44	−3.44	−3.44
Average of the R ² values^c	0.994	0.994	0.994
Average of the intercept values^c	41.4	41.4	41.4
Average value (log₁₀ copies/g)	ND	2.31	2.79
SD of value (log₁₀ copies/g)	—	0.53	0.25
Repeatability SD (Sr; log₁₀ copies/g)	—	0.28	0.24
Reproducibility SD (SR; log₁₀ copies/g)	—	0.48	0.46

All 14 participating laboratories submitted test results.

Laboratories that reported that the slope is not between −3.10 and −3.60 were excluded from data analysis as outlier.

Values from laboratory-recorded outliers were excluded from calculation.

ND, not detected; SD, standard deviation.

Table 5.

Interlaboratory Study Results for Norovirus GII in Oyster Samples

	Contamination level
	Negative (A)	Medium (B)	High (C)
Total data sets^a	14	14	14
No. of outliers^b	6	6	6
No. of laboratories after elimination outliers	8	8	8
No. of laboratories that failed to obtain quantification values of test sample B and/or C	—	4	3
No. of laboratories that submitted test result adopted for calculation	8	4	5
Average of the slope of the standard curves^c	−3.40	−3.40	−3.40
Average of the R ² values^c	0.990	0.990	0.990
Average of the intercept values^c	40.7	40.7	40.7
Average value (log₁₀ copies/g)	ND	2.64	3.24
SD of value (log₁₀ copies/g)	—	0.61	0.21
Repeatability SD (Sr; log₁₀ copies/g)	—	0.46	0.13
Reproducibility SD (SR; log₁₀ copies/g)	—	0.41	0.47

All 14 participating laboratories submitted test results.

Laboratories that reported that the slope is not between −3.10 and −3.60 were excluded from data analysis as outlier.

Values from laboratory-recorded outliers were excluded from calculation.

ND, not detected; SD, standard deviation.

The average values obtained at each contamination level for GI were 2.31 and 2.79 log₁₀ copies/g, and those for GII were 2.64 and 3.24 log₁₀ copies/g, test sample B and C, respectively (Tables 4 and 5). The recovery rates of norovirus obtained at each contamination level for GI were 76.0% and 75.5%, and those for GII were 88.1% and 87.6%, test sample B and C, respectively. The tLOQ for norovirus GI and GII in oysters was estimated to be 1.92 and 1.85 log₁₀ copies/g, respectively. The Sr for GI were 0.28 and 0.24 log₁₀ copies/g, and those for GII were 0.46 and 0.13 log₁₀ copies/g, test sample B and C, respectively (Tables 4 and 5). The S_R for GI were 0.48 and 0.46 log₁₀ copies/g, and those for GII were 0.41 and 0.47 log₁₀ copies/g, test sample B and C, respectively (Tables 4 and 5). The average Sr and S_R obtained at each contamination level for GI were 0.26 and 0.47, and those for GII were 0.30 and 0.44, respectively.

Discussion

Because of limitations to implement test system by using ISO 15216-1:2017 under appropriate quality management system in Japan, this study was conducted to investigate the standard method used for quantification of norovirus in oysters in Japan for provisional adaptation as an alternative to ISO 15216-1:2017. For this purpose, acceptability of the method was evaluated based on the LOQ, Sr, and S_R reported in the validation study for ISO 15216-1:2017 by Lowther et al. (2019). The LOQ for norovirus GI and GII in oysters by using ISO 15216-1:2017 was 1.53 and 1.72 log₁₀ copies/g, respectively (Lowther et al., 2019), whereas the LOQ by using the standard method in Japan was not determined in this study. Thus, in this study, we tried to estimate the tLOQ for norovirus GI and GII in oysters by the standard method used in Japan. First, the tLOQ for the standard method used in Japan was calculated as 1.80 log₁₀ copies/g, if the recovery of norovirus is 100%. However, it needs to be addressed that the tLOQ is probably much higher because 100% recovery is unattainable. In this study, the recovery rate of norovirus GI and GII was estimated as about 76% and 88%, respectively. If these recovery rates would be taken it into account for estimation of the tLOQ, the tLOQ for norovirus GI and GII in oysters by the standard method used in Japan was expected to be 1.92 and 1.85 log₁₀ copies/g, respectively. Thus, the tLOQ for norovirus GI and GII in oysters by using the standard method used in Japan seems to be slightly higher than the LOQ of ISO 15216-1:2017 reported by Lowther et al. (2019). The lower LOQ in ISO 15216-1:2017 may come from a relatively higher volume of template RNA used for one real-time PCR reaction, compared with the volume of cDNA used for that of the standard method in Japan. The standard method in Japan uses reverse transcription with random primers before the real-time PCR step that may cause dilution of the template genome. In addition, the complexity of the process for the standard method in Japan may influence the higher values of the tLOQ. The sample preparation procedure adopted for ISO 15216-1:2017 is simple, since RNAs are directly subjected to one-step RT-PCR that will expect higher recovery of norovirus genome. Since the practical LOQ of the standard method used in Japan remains unknown, further study should be to clarify the LOQ of the method for further consideration of the performance characteristics.

To our most concern, exogenous process control materials were not used in this study. Therefore, we were not able to specify the possible reasons why several laboratories could not quantify norovirus GI and/or GII in samples B and/or C, which was quantifiable by expert laboratory as determined by homogeneity and stability test, despite meeting criteria of acceptability of the standard curve. In fact, stability of spiked norovirus in oyster tissues conducted by expert laboratory was not affected by frozen storage for 10 weeks. In our preliminary study to assess the PCR inhibition by using exogenous internal positive control plasmid (IPC; Nippon Gene) in the PCR reaction, there were 15 out of 508 wild oyster samples that the difference of the Cq value of IPC due to the presence or absence of sample-derived cDNA (ΔCq) was >2.00 (unpublished data). Regarding the recovery rate of an exogenous process control virus (Feline calicivirus [FCV] strain F9), there were 188 out of 510 wild oyster samples with an extraction efficiency of <1% (unpublished data). Median recovery rate of FCV in samples in which FCV was recovered by more than 1% was 5.5% (unpublished data). Thus, recovery of target would be more affective to the quantification values compared to PCR inhibition. In this study, it is likely that norovirus in the sample has not been successfully recovered, which might be caused by artificial factor in conducting the procedure of the method. To ensure scientifically reliable test result, capacity building of laboratories and implementation of proficiency testing should be considered for future tasks in combination with an application of process control materials in the method. ISO 15216-1:2017 was reported to produce the average Sr and S_R of 0.23 and 0.50 log₁₀ copies/g (Lowther et al., 2019) for norovirus, respectively, suggesting that the performance of the standard method in Japan would be acceptable to a certain extent. Since the Sr and S_R for each virus and contamination level were based on two observations from each laboratory, we recognized that more observations would be needed to provide robust data.

Conclusions

To complement a number of subjects that emerged from this study, future tasks will be undertaken as follows: (1) determination of practical LOQ for the Japan method, (2) preparing quality control data, and (3) determination of a compatibility for test results obtained using the Japan method according to ISO 15216-1:2017.

Footnotes

Acknowledgment

The whole study was carried out as part of the surveillance/monitoring program of microbiological hazards in the MAFF.

Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

[AOAC] Association of Official Agricultural Chemists. AOAC International methods committee guidelines for validation of microbiological methods for food and environmental surfaces. 2012. Available at: http://members.aoac.org/aoac_prod_imis/AOAC_Docs/StandardsDevelopment/AOAC_Validation_Guidelines_for_Food_Microbiology-Prepub_version.pdf accessed February 10, 2021 .

[EFSA] European Food Safety Authority. Scientific opinion on norovirus (NoV) in oysters: Methods, limits and control options. EFSA J, 2012; 10:2500.

[EFSA] European Food Safety Authority. Technical specifications for a European baseline survey of norovirus in oysters. EFSA J, 2016; 14:4414.

[ISO 15216-1:2017] Microbiology of the food chain—Horizontal method for determination of hepatitis A virus and norovirus using real-time RT-PCR—Part 1: Method for quantification.

[ISO/TS 15216-1:2013] Microbiology of food and animal feed—Horizontal method for determination of hepatitis A virus and norovirus in food using real-time RT-PCR—Part 1: Method for quantification.

Kageyama

, Kojima

, Shinohara

, Uchida

, Fukushi

, Hoshino

, Takeda

, Katayama

. Broadly reactive and highly sensitive assay for Norwalk-like viruses based on real-time quantitative reverse transcription-PCR. J Clin Microbiol, 2003; 41:1548–1557.

Le Guyader

, Haugarreau

, Miossec

, Dubois

, Pommepuy

. Three-year study to assess human enteric viruses in shellfish. Appl Environ Microbiol, 2000; 66:3241–3248.

Lowther

, Bosch

, Butot

, Ollivier

, Mäde

, Rutjes

, Hardouin

, Lombard

, In't Veld

, Leclercq

. Validation of EN ISO method 15216 - Part 1—Quantification of hepatitis A virus and norovirus in food matrices. Int J Food Microbiol, 2019; 288:82–90.