1

DISCLAIMER

This paper was submitted to the Bulletin of the World Health Organization and was posted to

the Zika open site, according to the protocol for public health emergencies for international

concern as described in Christopher Dye et al. (http://dx.doi.org/10.2471/BLT.16.170860).

The information herein is available for unrestricted use, distribution and reproduction

in any medium, provided that the original work is properly cited as indicated by the Creative

Commons Attribution 3.0 Intergovernmental Organizations licence (CC BY IGO 3.0).

RECOMMENDED CITATION

Corman VM, Rasche A, Baronti C, Aldabbagh S, Cadar D, Reusken CBEM et al. Clinical

comparison, standardization and optimization of Zika virus molecular detection [Submitted].

Bull World Health Organ E-pub: 19 Apr 2016. doi: http://dx.doi.org/10.2471/BLT.16.175950.

Clinical comparison, standardization and optimization of Zika virus molecular detection

Victor M. Corman,a Andrea Rasche,a Cecile Baronti,b Souhaib Aldabbagh,a Daniel Cadar,c Chantal B.E.M. Reusken,d Suzan D. Pas,d Abraham Goorhuis,e Janke Schinkel,f Richard Molenkamp,f Beate M. Kuemmerer,a Tobias Bleicker,a Sebastian Brünink,a Monika Eschbach-Bludau,a Anna M. Eis-Hübinger,a Marion P. Koopmans,d Jonas Schmidt-Chanasit,c Martin P. Grobusch,e Xavier de Lamballerie,b Christian Drostena & Jan Felix Drexlera aInstitute of Virology, University of Bonn Medical Centre, Bonn 53127, Germany

bAix Marseille Université, IRD French Institute of Research for Development, EHESP French School of

Public Health, EPV UMR_D 190 "Emergence des Pathologies Virales", France

cBernhard Nocht Institute for Tropical Medicine, WHO Collaborating Centre for Arbovirus and

Hemorrhagic Fever Reference and Research, Hamburg, Germany

dErasmus MC, Department of Viroscience, 3000 CA, Rotterdam, the Netherlands

eCenter of Tropical Medicine and Travel Medicine, Department of Infectious Diseases, Academic

Medical Center, University of Amsterdam, 1100 DD Amsterdam, the Netherlands

fClinical Virology Laboratory, Department of Medical Microbiology, Academic Medical Center, Amsterdam, the Netherlands (JS, RM)

Correspondence to Jan Felix Drexler (e-mail: drexler@virology-bonn.de)

(Submitted: 18 April 2016 – Published online: 19 April 2016)

2

One sentence summary: Very low Zika virus loads in blood and urine from patients and low

sensitivity of several published assays imply invalid test results during the current outbreak

and demand highly accurate diagnostics.

What was already known about the topic concerned:

The Zika virus (ZIKV) has been known since the 1950ies. There are six published real time

RT-PCR-based protocols (qPCR), several of which are widely used for virus detection in the

context of the current outbreak in the Americas. Data on analytical sensitivity and

compatibility with current ZIKV outbreak strains is not available for most of these assays and

the comparability of these assays between laboratories remains unknown.

What new knowledge the manuscript contributes:

Several assays may be of limited utility for patient diagnostics during the current outbreak

because of low sensitivity and incompatibility with ZIKV outbreak strains. ZIKV loads were

low irrespective of sample type, implying patients may go undiagnosed during the current

outbreak due to limited assay sensitivity. The novel control RNA generated in this study

allowed uniform ZIKV quantification and will prove useful for patient characterization in

multicentric studies on ZIKV pathogenesis.

3

Abstract

Objective

Molecular Zika virus (ZIKV) detection is key to patient diagnostics during the current

outbreak. Here, we address standardization and diagnostic performance of widely used real-

time RT-PCR (qPCR) protocols for ZIKV detection.

Methods

Two novel qPCR protocols covering the currently known ZIKV genetic variability were

analyzed together with all six published qPCR protocols. The performance of all assays was

compared using a newly constructed universal control RNA (ucRNA) that contains the target

regions of all compared assays on one strand of synthetic RNA.

Findings

Up to 10 oligonucleotide mismatches with ZIKV outbreak strains existed in published qPCR

protocols. The analytical sensitivity of most assays was around 5 copies per reaction, whereas

three assays showed a 3-250-fold decreased sensitivity. The novel ucRNA enabled uniform

ZIKV quantification, whereas comparisons of PCR threshold cycles (CT-values) resulted in up

to 20-fold misquantification between protocols. Mean ZIKV loads in 33 outbreak samples

were 104 RNA copies/mL of blood (range; 10

2-4x10

5) and 5x10

3 RNA copies/ mL of urine

(range; 4x102-5.9x10

4) within two weeks after symptom onset.

Conclusion

Several ZIKV qPCR protocols show limited sensitivity and incompatibility with ZIKV

outbreak strains. ZIKV infection results in low virus concentrations close to the technical

limit of detection irrespective of sample type, implying that 20%-80% of patients may go

undiagnosed due to limited sensitivity of molecular tests. We provide updated protocols for

ZIKV detection that are suitable for all ZIKV strains. The ucRNA will enable coordinated

implementation of ZIKV molecular diagnostics across regions and within multicentric clinical

trials.

4

Introduction

The arthropod-borne Zika virus (ZIKV, genus Flavivirus, family Flaviviridae) was first

identified in 1947 in Uganda (1). Only sporadic human cases were reported prior to the 2007

outbreak in the Micronesian Yap islands from which 49 confirmed and 59 probable cases

were reported (2, 3). ZIKV infections are frequently asymptomatic or show only mild clinical

symptoms, including fever, arthralgia and rash (2, 4). However, severe neurological

complications, including the Guillain-Barré syndrome were reported from previous outbreaks

(5, 6). The current outbreak is additionally associated with fetal malformations (7-10).

In Latin America and the Caribbean, ZIKV infection cannot be reliably diagnosed by clinical

presentation because the co-circulating dengue virus (DENV) and chikungunya virus

(CHIKV) cause similar symptoms. Serology is challenging because of the cross-reactivity of

antibodies caused by endemic flaviviruses including DENV, St Louis encephalitis, and West

Nile virus (WNV) (4, 11, 12). Reliable detection of ZIKV is key to investigations of ZIKV

epidemiology and pathogenesis (13). Because of potential association with neurological

fetopathies, ZIKV infection should ideally be diagnosed already in the first trimester of

pregnancy when neurological development takes place (14). Direct detection of ZIKV is also

key to investigations of alternative transmission routes such as semen and blood donations.

There are six widely used real-time RT-PCR assays for ZIKV detection (11, 15-17). An

additional novel real-time RT-PCR assay has been recommended by the Pan American health

association (PAHO) for the current outbreak (13).

It is unclear which type of clinical specimens is most suitable for ZIKV detection.

Investigations of small series of patients suggest that ZIKV is present in blood only a few

days after infection. According to these studies, a generally low level of viral loads may

further complicate ZIKV detection (11, 18). ZIKV detection in saliva may be more sensitive

than detection in blood, but shedding in saliva and blood appeared to be equally short-lived

(19). Urine, semen and saliva were reported to be positive for ZIKV RNA for 2 weeks and

longer, and could thus be useful non-invasive materials for diagnosis and clinical studies (18,

20, 21).

Here we determined viral load profiles in blood and urine, provide comparative laboratory

data for published real-time RT-PCR tests, generate quantitative controls and project a high

risk of false negative ZIKV test results.

5

Materials and Methods

Clinical specimens

Clinical specimens were obtained from routine diagnostics sent for investigation of ZIKV or

DENV to the University of Bonn Medical Centre, Bonn, Germany, the Bernhard-Nocht

Institute for Tropical Medicine, Hamburg, Germany, the Academic Medical Centre,

Amsterdam, the Netherlands and the Erasmus Medical Centre, Rotterdam, the Netherlands.

Virus quantification and characterization

DENV RNA quantification and flavivirus typing were done as described previously (22)(23).

Quantitative controls were generated as described previously (24). The universal control RNA

(ucRNA) was custom designed as a gBlocks fragment with a T7 promotor sequence

(Integrated DNA Technologies, Leuven, Belgium) and in-vitro transcribed as described

before (24). All individual IVT and the ucRNA allowed highly comparable quantification of

ZIKV RNA with a mean 2-fold deviation of results (maximum deviation, 6-fold), suggesting

the ability to use all of these controls to generate comparable results even upon usage of

different real-time RT-PCR methods in different laboratories. For all other experiments,

ZIKV RNA was generally quantified using reaction conditions exemplified in

Supplementary Figure S1.

The ucRNA offers advantages to laboratories operating different real-time RT-PCR assays,

but bears the same risk of laboratory contamination as full viral RNA. In contrast to full viral

RNA, potential cases of laboratory contamination with the ucRNA can be proven by two

highly sensitive real-time RT-PCR marker assay variants designed to specifically detect the

ucRNA at lower limits of detection that were comparable to ZIKV-specific assays with 4.3

(95% confidence interval (CI), 2.9-10.9) and 3.3 (95% CI, 2.4-6.5) copies per reaction,

respectively. These marker assays contain detection probes that target the overlap of two

joined genomic target domains, which do not naturally occur in the full ZIKV genome.

Accordingly, the two assays showed no detection of full ZIKV RNA even upon using high-

titred cell culture isolates (106-10

9 copies/mL). Oligonucleotide sequences of the two marker

assays were: Marker1-rtF, GCATCCAGCCAGAGAATCTG; Marker1-rtR,

CAATAACGGCTGGATCACACTC; Marker1-rtP,

TGCTGTCAGTTCACTCAAGGTTAGAGA-Black Hole Quencher-1 (BHQ-1) and Marker2-

rtF, CTTGACAATATTTACCTCCAAGATG; Marker2-rtR,

GTTGCTTTTCGCTCCAGAGAC; Marker2-rtP, FAM-

CATAGCCTCGCTCTCTACACATGAGA-BHQ1.

6

Results

All published real-time RT-PCR assays used in ZIKV studies until April 1st were used in this

study. These published assays target the membrane (M), envelope (E), NS1, NS2b, NS3 and

NS5 domains of the ZIKV genome (Figure 1A) (11, 13, 15-17). Two additional assays

targeting the E and NS1 domains were designed for this study. To enable comparative testing

of all existing assays, we generated individual quantified in-vitro transcripts (IVT) for the

genomic target regions of all assays (IVT I-V, Figure 1A). In addition, we joined all target

domains into a quantitative universal calibrator RNA (ucRNA, Figure 1A). This quantifiable

genome mimic enables stoichiometrically exact analyses of the lower limits of detection of all

assays in comparison, without the need to rely on full viral RNA that cannot be quantified. All

controls generated for this study are based on a current ZIKV outbreak strain from Brazil

(GenBank accession no. KU321639). Table 1 provides details on oligonucleotide sequences

and the IVT to be used for each assay. All quantitative controls can be acquired free of charge

for non-commercial purposes via the European Virus Archive (EVA, see Acknowledgment

for details)

Real-time RT-PCR sensitivity can be affected by nucleotide mutations in the binding sites of

primers and probes (28). So far, the genetic variability of the ZIKV Asian lineage, including

virus strains causing the current American outbreak is lower than that of the African lineage.

Genomic variation within the known Asian lineage ZIKV strains so far is limited to around

2% nucleotide differences across the genome (Figure 1B). However, mutations do not occur

evenly across viral genomes and up to 10 nucleotide mismatches between the sequences of

the published assay oligonucleotides and the Asian lineage already exist, with up to 5

mismatches in individual primers or probes (Figure 2A). Due to the various mismatches

observed with published assays (Table 1), we designed two novel assays covering the

currently known ZIKV genetic variability in two different genomic domains. These novel

assays showed only up to 3 potential mismatches per assay (Figure 2B) and were designed to

avoid mismatches in the most critical 3’-terminal regions of oligonucleotides which affect

primer binding most (28). The novel NS1-based assay was additionally designed to allow

cross-detection of Spondweni virus, the closest relative of ZIKV. This is because regions

conserved between related virus taxa can be expected to allow less variation than other

genomic regions.

Specific detection of ZIKV is crucial to avoid false-positive test results. We evaluated all

assays on 37 high-titered flavivirus cell culture isolates, covering the majority of the

mosquito-borne flaviviruses (Supplementary Figure S2 and Table S1). None of the

7

published assays detected the co-circulating alphavirus CHIKV or any other flavivirus tested.

Despite very high concentrations of viral RNA an Asian ZIKV strain-specific E-based assay

did not detect the African ZIKV lineage likely because of nucleotide mismatches, (16). As

intended, the novel NS1-based assay presented in this study cross-detected the Spondweni

virus, as well as Kokobera and Jugra virus, which do not occur in humans affected by the

current outbreak (12).

The limit of detection (LOD) is a standardized measure to assess each assay´s performance

for qualitative detection. Data on analytical sensitivity including LOD are not available for

most of the published assays. To determine exactly comparable LODs based on the present

outbreak strain, the ucRNA was used to assess the sensitivity of all assays. As shown in Table

2 and Supplementary Figure S3, all but three assays showed comparably high analytical

sensitivities of around 5 copies per reaction for this standardized target molecule. An NS1-

and an NS2b-based assay showed about two- to three-fold higher LODs (13, 16), whereas an

NS3-based assay showed a high LOD of 1,373 copies per reaction (17). To exemplify the

clinical impact of the LODs of real-time RT-PCR assays, we extrapolated the assay sensitivity

to viral loads in clinical samples. Even highly sensitive assays with an LOD of 5 copies per

reaction reach a detection limit in the range of 103 copies per mL, whereas an LOD of 1,000

copies per reaction implies a detection limit in the range of 106 copies per mL (Table 3).

In addition to the qualitative detection of viral RNA, virus quantification is an important tool

to investigate the risk of transmission through different biological specimens, viral response to

antiviral therapy and the impact of viral concentrations on pathogenesis. Several previous

studies reported ZIKV viral load data in form of threshold cycle (CT) values (15, 18, 29).

However, CT values are highly variable and may cause misleading comparisons of viral loads

between studies (30, 31). Laboratory conditions such as PCR instruments and reagents can

greatly influence CT values. We explored the variability of CT values by testing our two new

assays under diverse reaction conditions involving reagents by different suppliers on different

real-time PCR instruments. Even upon a variation of only two variables, the same virus target

concentration could yield CT values that differed from each other by up to 4.3 cycles, which

corresponds to about 20-fold deviations in viral load results (Figure 3).

To obtain more information on the relative utility of blood or urine for diagnostics of acute

ZIKV infection, the novel assays were used to quantify ZIKV loads in 33 clinical materials

from patients sampled during the current outbreak (24 patients in total). Matched urine and

blood specimens taken on the same day from the same patient were available from six patients

sampled 2-12 days after symptoms onset. In three patients, urine viral loads were equivalent

8

to or lower than those in blood, whereas in the other three patients urine viral loads were 10-

100-fold higher than those in blood (Figure 4).

ZIKV loads in all available samples (12 blood and 21 urine specimens) showed low viral

loads (Figure 5A and Supplementary Figure S4) and were comparable to viral loads in

blood from patients sampled during the ZIKV outbreak in Micronesia in 2007 (11) (Figure

5A). A combined dataset comprising the data from the 2007 outbreak provided by Lanciotti et

al. (11) and this study resulted in mean ZIKV loads of 104 RNA copies per mL of blood

(range; 102-4x10

5) and 5x10

3 RNA copies per mL of urine (range; 4x10

2-5.9x10

4). These viral

loads were determined in clinical specimens sampled during comparable intervals with 11

days after symptom onset for urine and 12 days after symptom onset for blood specimens,

respectively and did not differ significantly (t-test, p=0.26). Within the combined ZIKV

dataset, nine of 41 specimens contained viral loads of 2.5x103 RNA copies per mL or lower,

leading to an estimated risk of false-negative test results of 20% even with highly sensitive

assays. Using assays with an LOD of 100 copies per reaction, the rate of false-negative test

results grows up to 80% (Figure 5B). Many of the laboratories conducting ZIKV testing in

countries affected by the outbreak have long-standing experience with detection and

quantification of the co-circulating DENV. To compare the risk of false-negative test results

between ZIKV and DENV, we additionally quantified 38 DENV clinical specimens. Here,

mean viral loads were 5x105 RNA copies per ml (range; 5x10

2-5x10

8), i.e., DENV loads in

blood were generally about 100-fold higher virus than ZIKV loads (t-test, p=0.03) (Figure

5C). Accordingly, the technical risk of false-negative results can be estimated to be 10-fold

lower for highly sensitive DENV assays with an LOD of 5 copies per reaction and 4-fold

lower for assays with an LOD of 100 copies per reaction compared to ZIKV (Figure 5D).

Discussion

Highly sensitive molecular testing is a key element in the response to the current ZIKV

outbreak. The present study provides guidance for the choice of methodology and introduces

optimized ZIKV detection assays along with an available molecular calibrator that enables the

comparison of results between laboratories and studies. The finding of low ZIKV loads

forecasts a high proportion of false-negative test results during clinical application of real-

time RT-PCR.

The results of assay comparisons suggest that several published real-time RT-PCR assays

may be of limited utility for clinical diagnostics during the current ZIKV outbreak. One NS3-

based assay that was intended for virus typing and differentiation should not be used for

9

clinical ZIKV diagnostics because of its low sensitivity (17). Several other assays present

features potentially limiting their utility within the current outbreak, including limited access

to specific probe formats (15), relatively lower analytical sensitivity (13) and high numbers of

potential mismatches with members of the Asian ZIKV lineage (16). Our novel assays may be

more robust against genetic variation in ZIKV, but real-time monitoring of all assay

oligonucleotide binding regions is required during the current situation. In summary, the

Lanciotti E-, the Bonn E- and the Bonn NS1-based assays are highly sensitive according to

our data and show limited mismatched genomic positions. These assays can thus be reliably

used at the present knowledge of ZIKV variability, preferably in combination of at least two

assays to increase clinical sensitivity.

The low ZIKV loads we detected in urine and blood samples are in agreement with the few

previous studies reporting quantitative data (11, 18). Contrary to preliminary data from 6

patients from French Polynesia (18), no significant difference in virus loads between urine

and blood specimens could be observed. Our data thus did not support urine as the generally

more suitable clinical specimen to detect ZIKV. However, ZIKV RNA seems to remain

detectable in urine and semen for a longer time period than in blood (18, 20, 21). Therefore,

our results support real-time RT-PCR testing of at least two different clinical specimens for

ZIKV diagnostics, ideally including blood and urine. Unfortunately, we could not

comparatively evaluate saliva specimens in this study.

While commercial diagnostic real-time RT-PCR reagents for ZIKV detection are becoming

available, in-house formulations are still widely used in the affected region because of limited

resources (13, 24). These assays are difficult to standardize and compare, but will often

constitute the only technical resource available in laboratories. Transfer of essential reagents

with coordinated implementation of protocols and practical skills can permit accurate real-

time RT-PCR diagnostics in resource-limited settings (24, 25, 32, 33). Among the most

essential contributions to technology transfer is the provision of standardized RNA reagents

that can be shipped internationally without biosafety concerns. Research consortia and public

health structures can use these reagents to establish a technical basis for test implementation,

as demonstrated, e.g., for SARS and MERS coronavirus (34, 35). These viruses were novel at

the time of emergence, and diagnostic tests could be defined along with the provision of

reagents. In the case of ZIKV, a long-known agent invading a new region, a multitude of test

formulations has already been available, so that assay standardization can only work by

provision of a reference reagent that is universally applicable in all assays. Our ucRNA

reagent can be used with any current published in-house assay to continuously control for

10

sensitivity, without the need to rely on full virus RNA. Quantitative comparability between

studies will enable relative estimates of transmission risks associated with blood donations as

well as solid organ transplants and other body fluids than blood, e.g., semen or saliva. These

data may also assist studies on ZIKV pathogenesis, since higher viral loads have been

associated with severe clinical courses of DENV infection (36) and with prolonged severe

arthralgia in CHIKV infection (37).

Low ZIKV loads imply a high proportion of false-negative test results in countries affected by

the outbreak. To date, only 3% of 199,922 suspected ZIKV cases could be confirmed in the

PAHO region (38). Certainly, the difficulties to manage the high number of diagnostic

requests in resource-limited settings contribute most to the low number of confirmed cases,

consistent with the results of a recent study conducted in Puerto Rico, in which 20% of

suspected ZIKV cases could be confirmed using molecular and serologic tools (39). However,

our data imply that several thousand patients presenting with low ZIKV loads may have gone

undiagnosed by molecular testing, contributing to the low rate of confirmed cases and

highlighting the need of combined ultrasensitive molecular and serologic testing.

The need for ultrasensitive molecular ZIKV detection additionally applies to blood safety in

endemic countries. ZIKV has been detected in 3% of blood donors in previous outbreaks (40)

and transfusion-associated transmission has already been reported from Brazil (41). Our

comparison of blood viral loads and real-time RT-PCR sensitivity suggest a risk of false-

negative results during pooled and even individual blood donor screening. This estimated risk

is consistent with several cases of transfusion- and solid organ transplantation-associated

transmission of WNV in the US, another mosquito-borne flavivirus showing relatively low

viral loads that led to false-negative real-time RT-PCR results in pooled blood donor testing

previously (42).

Most importantly, the observed association of ZIKV and fetal malformations demand reliable

ZIKV diagnostics of pregnant women. A comparison of the current sensitivity and viral load

data suggest that molecular testing during pregnancy may preferentially diagnose highly

viremic women, which may influence estimates of the manifestation index of congenital

disease, if congenital disease correlates with viral load. The low PCR sensitivity caused by

low ZIKV loads implies a limited capacity of molecular protocols to exclude ZIKV detection

in highly affected regions, requiring additional serological testing to exclude ZIKV infection

during pregnancy in multicentric cohort studies investigating ZIKV pathogenesis.

In summary, our data emphasize the need for highly sensitive protocols in molecular

diagnostic testing for ZIKV infection. In addition to an appropriate choice of methodology,

11

clinical sensitivity can be increased by testing several specimens per patient, by using more

than one real-time RT-PCR target in the laboratory and combined molecular and serological

testing (11). The novel assays provided through this study, as well as the ucRNA reagent used

as a universal quantitative calibrator and positive control can ensure high sensitivity and good

comparability of qualitative and quantitative diagnostic results in public health laboratories

and clinical studies.

12

Acknowledgement

The controls described in this paper can be ordered at the following links to the European

Virus Archive (EVA): Zika virus IVT I, http://www.european-virus-

archive.com/Portal/produit.php?ref=1598&id_rubrique=9; Zika virus IVT II,

http://www.european-virus-archive.com/Portal/produit.php?ref=1599&id_rubrique=9; Zika

virus IVT III, http://www.european-virus-

archive.com/Portal/produit.php?ref=1600&id_rubrique=9; Zika virus IVT IV,

http://www.european-virus-archive.com/Portal/produit.php?ref=1601&id_rubrique=9; Zika

virus IVT V, http://www.european-virus-

archive.com/Portal/produit.php?ref=1602&id_rubrique=9; Zika virus universal calibrator

RNA 1.0, http://www.european-virus-

archive.com/Portal/produit.php?ref=1603&id_rubrique=9.

We thank Janett Wieseler, Sandra Junglen, and Annemiek van der Eijk for their support. This

study was supported by funding from the European Commission through the Horizon 2020

project EVAg (European Virus Archive goes global), grant agreement number 653316, and

the framework program (FP)7 project PREPARE (Platform for European Preparedness

Against (Re-)emerging Epidemics), grant agreement number 602525.

13

References

1. Dick GWA, Kitchen SF, Haddow AJ. Zika Virus (I). Isolations and serological specificity.

Transactions of The Royal Society of Tropical Medicine and Hygiene. 1952 September 1,

1952;46(5):509-20.

2. Duffy MR, Chen TH, Hancock WT, Powers AM, Kool JL, Lanciotti RS, et al. Zika virus

outbreak on Yap Island, Federated States of Micronesia. N Engl J Med. 2009 Jun

11;360(24):2536-43.

3. Fauci AS, Morens DM. Zika Virus in the Americas--Yet Another Arbovirus Threat. N

Engl J Med. 2016 Feb 18;374(7):601-4.

4. Gatherer D, Kohl A. Zika virus: a previously slow pandemic spreads rapidly through

the Americas. J Gen Virol. 2015 Dec 18.

5. Broutet N, Krauer F, Riesen M, Khalakdina A, Almiron M, Aldighieri S, et al. Zika Virus

as a Cause of Neurologic Disorders. N Engl J Med. 2016 Mar 9.

6. Cao-Lormeau VM, Blake A, Mons S, Lastere S, Roche C, Vanhomwegen J, et al.

Guillain-Barre Syndrome outbreak associated with Zika virus infection in French Polynesia: a

case-control study. Lancet. 2016 Feb 29.

7. Calvet G, Aguiar RS, Melo AS, Sampaio SA, de Filippis I, Fabri A, et al. Detection and

sequencing of Zika virus from amniotic fluid of fetuses with microcephaly in Brazil: a case

study. Lancet Infect Dis. 2016 Feb 17.

8. Mlakar J, Korva M, Tul N, Popovic M, Poljsak-Prijatelj M, Mraz J, et al. Zika Virus

Associated with Microcephaly. N Engl J Med. 2016 Feb 10.

9. Brasil P, Pereira JP, Jr., Raja Gabaglia C, Damasceno L, Wakimoto M, Ribeiro Nogueira

RM, et al. Zika Virus Infection in Pregnant Women in Rio de Janeiro - Preliminary Report. N

Engl J Med. 2016 Mar 4.

10. Rasmussen SA, Jamieson DJ, Honein MA, Petersen LR. Zika Virus and Birth Defects -

Reviewing the Evidence for Causality. N Engl J Med. 2016 Apr 13.

11. Lanciotti RS, Kosoy OL, Laven JJ, Velez JO, Lambert AJ, Johnson AJ, et al. Genetic and

serologic properties of Zika virus associated with an epidemic, Yap State, Micronesia, 2007.

Emerg Infect Dis. 2008 Aug;14(8):1232-9.

12. Charrel R.N., Leparc-Goffart I., Pas S., de Lamballerie X, Koopmans M, C. R. State of

knowledge on Zika virus for an adequate laboratory response Bull World Health Organ.

2016;E-pub: 10 Feb 2016.

13. Waggoner JJ, Pinsky BA. Zika Virus: Diagnostics for an Emerging Pandemic Threat. J

Clin Microbiol. 2016 Feb 17.

14. Zika Epidemiological Update – 8 April 2016. PAHO/WHO. 2016.

15. Faye O, Faye O, Diallo D, Diallo M, Weidmann M, Sall AA. Quantitative real-time PCR

detection of Zika virus and evaluation with field-caught mosquitoes. Virol J. 2013;10:311.

16. Pyke AT, Daly MT, Cameron JN, Moore PR, Taylor CT, Hewitson GR, et al. Imported

zika virus infection from the cook islands into australia, 2014. PLoS Curr. 2014;6.

17. Tappe D, Nachtigall S, Kapaun A, Schnitzler P, Gunther S, Schmidt-Chanasit J. Acute

Zika virus infection after travel to Malaysian Borneo, September 2014. Emerg Infect Dis.

2015 May;21(5):911-3.

18. Gourinat AC, O'Connor O, Calvez E, Goarant C, Dupont-Rouzeyrol M. Detection of

Zika virus in urine. Emerg Infect Dis. 2015 Jan;21(1):84-6.

19. Musso D, Roche C, Nhan TX, Robin E, Teissier A, Cao-Lormeau VM. Detection of Zika

virus in saliva. J Clin Virol. 2015 Jul;68:53-5.

20. Barry A, Pasco H, Babak A, Sarah L, Daniel C, Emma JA, et al. Detection of Zika Virus in

Semen. Emerging Infectious Disease journal. 2016;22(5).

14

21. Barzon L, Pacenti M, Berto A, Sinigaglia A, Franchin E, Lavezzo E, et al. Isolation of

infectious Zika virus from saliva and prolonged viral RNA shedding in a traveller returning

from the Dominican Republic to Italy, January 2016. Euro Surveill. 2016;21(10).

22. Drosten C, Gottig S, Schilling S, Asper M, Panning M, Schmitz H, et al. Rapid detection

and quantification of RNA of Ebola and Marburg viruses, Lassa virus, Crimean-Congo

hemorrhagic fever virus, Rift Valley fever virus, dengue virus, and yellow fever virus by real-

time reverse transcription-PCR. J Clin Microbiol. 2002 Jul;40(7):2323-30.

23. Moureau G, Temmam S, Gonzalez JP, Charrel RN, Grard G, de Lamballerie X. A real-

time RT-PCR method for the universal detection and identification of flaviviruses. Vector

borne and zoonotic diseases (Larchmont, NY. 2007 Winter;7(4):467-77.

24. Drexler JF, Kupfer B, Petersen N, Grotto RM, Rodrigues SM, Grywna K, et al. A novel

diagnostic target in the hepatitis C virus genome. PLoS Med. 2009 Feb 10;6(2):e31.

25. Drosten C, Panning M, Drexler JF, Hansel F, Pedroso C, Yeats J, et al. Ultrasensitive

monitoring of HIV-1 viral load by a low-cost real-time reverse transcription-PCR assay with

internal control for the 5' long terminal repeat domain. Clin Chem. 2006 Jul;52(7):1258-66.

26. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Molecular Evolutionary

Genetics Analysis version 6.0. Mol Biol Evol. 2013 Dec;30(12):2725-9.

27. Simmonds P. SSE: a nucleotide and amino acid sequence analysis platform. BMC Res

Notes. 2012;5:50.

28. Drexler JF, de Souza Luna LK, Pedroso C, Pedral-Sampaio DB, Queiroz AT, Brites C, et

al. Rates of and reasons for failure of commercial human immunodeficiency virus type 1 viral

load assays in Brazil. J Clin Microbiol. 2007 Jun;45(6):2061-3.

29. Hamel R, Dejarnac O, Wichit S, Ekchariyawat P, Neyret A, Luplertlop N, et al. Biology

of Zika Virus Infection in Human Skin Cells. J Virol. 2015 Sep;89(17):8880-96.

30. Mackay IM. Real-time PCR in the microbiology laboratory. Clin Microbiol Infect. 2004

Mar;10(3):190-212.

31. Donald CE, Qureshi F, Burns MJ, Holden MJ, Blasic JR, Jr., Woolford AJ. An inter-

platform repeatability study investigating real-time amplification of plasmid DNA. BMC

Biotechnol. 2005;5:15.

32. Stocker A, Souza BF, Ribeiro TC, Netto EM, Araujo LO, Correa JI, et al. Cosavirus

infection in persons with and without gastroenteritis, Brazil. Emerg Infect Dis. 2012

Apr;18(4):656-9.

33. Panning M, Charrel RN, Donoso Mantke O, Landt O, Niedrig M, Drosten C.

Coordinated implementation of chikungunya virus reverse transcription-PCR. Emerg Infect

Dis. 2009 Mar;15(3):469-71.

34. Corman VM, Eckerle I, Bleicker T, Zaki A, Landt O, Eschbach-Bludau M, et al.

Detection of a novel human coronavirus by real-time reverse-transcription polymerase chain

reaction. Euro Surveill. 2012;17(39).

35. Drosten C, Doerr HW, Lim W, Stohr K, Niedrig M. SARS molecular detection external

quality assurance. Emerg Infect Dis. 2004 Dec;10(12):2200-3.

36. Vaughn DW, Green S, Kalayanarooj S, Innis BL, Nimmannitya S, Suntayakorn S, et al.

Dengue viremia titer, antibody response pattern, and virus serotype correlate with disease

severity. J Infect Dis. 2000 Jan;181(1):2-9.

37. Hoarau JJ, Jaffar Bandjee MC, Krejbich Trotot P, Das T, Li-Pat-Yuen G, Dassa B, et al.

Persistent chronic inflammation and infection by Chikungunya arthritogenic alphavirus in

spite of a robust host immune response. J Immunol. 2010 May 15;184(10):5914-27.

38. Cumulative Zika suspected and confirmed cases reported by countries and territories

in the Americas, 2015-2016. Updated as of 7 April 2016. PAHO/WHO. 2016.

15

39. Thomas DL, Sharp TM, Torres J, Armstrong PA, Munoz-Jordan J, Ryff KR, et al. Local

Transmission of Zika Virus - Puerto Rico, November 23, 2015-January 28, 2016. MMWR Morb

Mortal Wkly Rep. 2016;65(6):154-8.

40. Musso D, Nhan T, Robin E, Roche C, Bierlaire D, Zisou K, et al. Potential for Zika virus

transmission through blood transfusion demonstrated during an outbreak in French

Polynesia, November 2013 to February 2014. Euro Surveill. 2014;19(14).

41. Anonymous. Maintaining a safe and adequate blood supply during Zika virus

outbreaks - Interim guidance (WHO/ZIKV/HS/16.1). February 2016 [cited 2016, February

23rd]; Available from:

42. Dodd RY, Foster GA, Stramer SL. Keeping Blood Transfusion Safe From West Nile Virus:

American Red Cross Experience, 2003 to 2012. Transfus Med Rev. 2015 Jul;29(3):153-61.

16

Table 1. Assay oligonucleotides and potential nucleotide mismatches with Zika virus strains

Assay Target* No. potential

nucleotide

mismatches

(all ZIKV)

No. potential

nucleotide

mismatches

(ZIKV Asian

lineage)

Forward primer

sequence (5‘ →→→→ 3‘)

Probe sequence

(5‘ →→→→ 3‘)

Reverse primer

sequence (5‘ →→→→ 3‘)

Control* Reference

Lanciotti M M/E

(939-

1,015)

19 7 TTGGTCATGATA

CTGCTGATTGC

CGGCATACAGC

ATCAGGTGCAT

AGGAG

CCTTCCACAAA

GTCCCTATTGC

ucRNA;

IVT I

(811-1,500)

(11)

Lanciotti E E

(1,190-

1,266)

18 4 CCGCTGCCCAA

CACAAG

AGCCTACCTTG

ACAAGCAGTCA

GACACTCAA

CCACTAACGTTC

TTTTGCAGACAT

ucRNA;

IVT I

(11)

Bonn E E

(1,188-

1,316)

0 0 AGYCGYTGYCC

AACACAAG

CCTMCCTYGAY

AAGCARTCAGA

CACYCAA

CACCARRCTCCC

YTTGCCA

ucRNA;

IVT I

This study

Pyke E E

(1,326-

1,397)

28 7 AAGTTTGCATG

CTCCAAGAAAA

T

ACCGGGAAGAG

CATCCAGCCAG

A

CAGCATTATCCG

GTACTCCAGAT

ucRNA;

IVT I

(16)

Pyke NS1 NS1

(3,433-

3,498)

13 7 GCACAATGCCC

CCACTGT

TTCCGGGCTAA

AGATGGCTGTT

GGT

TGGGCCTTATCT

CCATTCCA

ucRNA;

IVT II

(3,145-

3,739)

(16)

Bonn NS1 NS1

(3,385-

3,495)

4 3 CRACYACTGCA

AGYGGAAGG

ATGGTGCTGYA

GRGARTGCACA

ATGC

GCCTTATCTCCA

TTCCATACC

ucRNA;

IVT II

This study

PAHO NS2b NS2b/N

S3

(4,538-

4,628)

11 4 CTGTGGCATGA

ACCCAATAG

CCACGCTCCAG

CTGCAAAGG

ATCCCATAGAG

CACCACTCC

ucRNA;

IVT IV

(4,246-

4,882)

(13)

Tappe NS3 NS3

(6,012-

6106)

15 10 TGGAGATGAGT

ACATGTATG

CTGATGAAGGC

CATGCACACTG

GGTAGATGTTGT

CAAGAAG

ucRNA;

IVT III

(5,770-

6,370)

(17)

Faye NS5 NS5

(9,376-

9,477)

6 3 AARTACACATA

CCARAACAAAG

TGGT

CTYAGACCAGC

TGAAR

TCCRCTCCCYCT

YTGGTCTTG

ucRNA;

IVT V

(9,100-

9,696)

(15)

*nucleotide position according to GenBank acc. KU321639. Locked nucleic acid bases in the Faye et al. probe

are underlined.

17

Table 2. Analytical sensitivity of real-time PCR tests

Assay 95% lower limit of detection

[copies per reaction, confidence

interval]

Lanciotti M 3.2 [2.2-8.3]

Lanciotti E 4.1 [2.7-11.4]

Bonn E 2.1 [1.4-8.0]

Pyke E 5.3 [3.0-25.7]

Pyke NS1 12.1 [5.9-78.5]

Bonn NS1 3.1 [2.3-5.8]

PAHO NS2b 17.0 [12.3-30.9]

Tappe NS3 1,377.3 [860-5,162]

Faye NS5 4.5 [8.0-43.9]

Table 3. Extrapolation of analytical sensitivity to clinical viral loads

Technical sensitivity

(copies/µl eluate)

Technical sensitivity

(copies/reaction#)

Viral load upon 100 µL

input volume eluted in

100 µL (copies/mL)

Viral load upon 140 µL

input volume eluted in 70

µL (copies/mL)*

1 5 5.0x103 2.5x10

3

5 25 2.5x104 1.25x10

4

10 50 5.0x104 2.5x10

4

20 100 1.0x105 5.0x10

4

250 1,000 1.0x106 5.0x10

5

#using 5 µL of eluted RNA per PCR reaction

*corresponding to standard input and elution volumes in the commonly used Qiagen Viral

RNA Mini kit (Qiagen, Hilden, Germany).

18

Figure legends

Figure 1. Genomic locations of Zika virus real-time RT-PCR tests and controls

A, Zika virus genomic representation (GenBank accession no. KU321639), with real-time RT-PCR assays identified below by respective first

authors or location and corresponding control in vitro transcripts (IVT) and parts of the universal control RNA (ucRNA) identified above. Genomic

regions not containing published assays so far, but potentially useful for future assay design due to genomic conservation within the Asian Zika

virus lineage were also included in the ucRNA. UTR, untranslated region. B, Genomic identity plot of all Zika virus polyprotein sequences

characterized to at least 80% available at GenBank by April 13th

, 2016. Similarity plots were done using SSE V1.2 (27). A sliding window of 200

and a step size of 40 nucleotides were used.

19

Figure 2. Variable genomic positions under assay oligonucleotide sites

For the African and Asian lineage, 100% consensus sequences were generated and mapped to respective

PCR primers and probes. Consensus calculations included all complete and partial Zika virus sequences

available at GenBank by April 7th

, 2016 (n=259 entries). Variable sites within the oligonucleotide binding

sites of the two ZIKV lineages are highlighted in orange, variable sites between the two lineages in grey,

mismatches by asterisks and red color within the oligonucleotide sequences. Y=C/T, R=A/G, M=A/C,

B=C/G/T, S=G/C, W=A/T, H=A/C/T, D=A/G/T, N=A/C/T/G, K=G/T, V=A/C/G. Locked nucleic acids in

the probe of the Faye et al. NS5-based assay are underlined, the deletion under the forward primer boxed.

Potential mismatches below oligonucleotides are indicated in red and by asterisks. Highlighted in grey,

variable site African vs. Asian Zika virus lineage.

20

Figure 3. Threshold cycle variation using different reaction conditions and thermocyclers

Symbols identify reaction mix and thermocycler yielding different threshold cycle (CT) values (y-axis)

compared to the standard protocol and thermocycler (x-axis). Comparison of CT values used the Bonn E- and

NS1-based assays using either the Superscript III One-Step RT-PCR kit (Thermo Fischer) or the Qiagen

One-Step RT-PCR kit (Qiagen) on a Roche LightCycler 480 and LightCycler 2.0 a Qiagen Rotorgene HQ

and an Applied Biosystems 7500 thermocycler. Reference conditions refer to the usage of Life

Technologies SuperScript III One-Step enzyme mix and a Roche LC480 thermocycler.

21

Figure 4. Zika virus loads in paired clinical specimens

Zika virus loads plotted per type of clinical specimen. Each color corresponds to an individual patient. Paired

urine and blood specimens were taken on same days and within first 10 days after symptom onset.

22

Figure 5. Zika and Dengue virus loads and risk of false-negative test results

A, Zika virus loads in blood and urine. Lines in plots show mean viral loads. Data from Lanciotti et al. were

reported in (11). All blood specimens included in the analysis were sampled on day 2-10 after symptom

onset, urine specimens were sampled on day 2-12 after symptom onset. For details on time of sampling see

Supplementary Figure S4. B, Projection of false-negative test results according to different lower limits of

detection for Zika virus, according to a 2:1 input vs. elution volume (e.g., 140 µL blood eluted in 70 µL) as

in (11) and a 100% extraction efficacy. C, Dengue virus loads in blood. D, Projection of false-negative test

results according to different lower limits of detection for dengue virus, using identical parameters as in B.

Supplementary Table S1. Viruses used for specificity testing Virus family Genus Virus Species Virus strain TCID50/mL Flaviviridae Flavivirus Aroa virus Aroa virus 1.32E+06

Aroa virus Bussuquara virus 4.18E+02

Aroa virus Iguape virus 1.32E+05

Bagaza virus Bagaza virus 2.89E+05

Banzi virus Banzi virus 1.32E+03

Bouboui virus Bouboui virus 1.32E+05

Dengue virus Dengue 1 5.78E+04

Dengue virus Dengue 2 6.30E+03

Dengue virus Dengue 3 7.09E+03

Dengue virus Dengue 4 8.93E+06

Edge Hill virus Edge Hill virus 1.32E+03

Ilheus virus Ilheus virus 1.32E+04

Ilheus virus Rocio virus 2.89E+05

Japanese encephalitis virus

Japanese encephalitis virus

3.15E+06

Jugra virus Jugra virus 6.18E+04

Kedougou virus Kedougou virus 1.32E+04

Kokobera virus Kokobera virus 1.32E+04

Koutango virus Koutango virus 4.18E+05

Modoc virus Modoc not available

Murray Valley encephalitis virus

Alfuy virus 1.32E+05

Murray Valley encephalitis virus

Murray Valley Encephalitis Virus

4.18E+04

Saboya virus Saboya virus 5.26E+04

Sepik virus Sepik virus 4.18E+04

St Louis encephalitis virus

Saint Louis Encephalitis Virus

1.32E+05

Tembusu virus Tembusu virus 4.18E+04

Tick-borne encephalitis virus

Tick-borne encephalitis virus

1.16E+06

Uganda S virus Uganda S virus 1.32E+04

Usutu virus Usutu virus 4.18E+05

Wesselsbron virus Wesselsbron virus 4.18E+03

West Nile virus West Nile virus 3.36E+06

Yellow fever virus Yellow fever virus 1.47E+07

Zika virus Zika virus 1.32E+04

no approved species Spondweni virus 1.32E+04

no approved species Cell fusing agent virus 3.40E+05

no approved species Culex Flavi 7.00E+04

no approved species Kamiti River Virus 3.40E+06

no approved species Lammi Virus 1.32E+04

no approved species Niénokoué 6.45E+06

Togaviridae Alphavirus Chikungunya virus Chikungunya virus 3.1E+06

SupplementaryFigureS1.Benchprotocolforreal‐timeRT‐PCRassays

Professor Dr. med. Jan Felix Drexler Fon: 0228. 287-11697 Fax: 0228. 287-19144 drosten@virology-bonn.de Dr. med. Victor Corman Fon: 0228. 287-13590 corman@virology-bonn.de Universitätsklinikum Bonn Sigmund-Freud-Str. 25 53105 Bonn

Real time RT-PCR for Zika virus

Bonn NS1 and Bonn E assay Example formulation: Thermo Fisher SuperScriptIII OneStep RT-PCR System with Platinum Taq DNA Polymerase

*non-acetylated. This component is only necessary if using glass capillary LightCycler. Can be replaced with water in plastic vessel machines such as ABI 7500, LC 480, etc. Primers / probe: Bonn_NS1 _FWD CAACYACTGCAAGYGGAAGG Bonn_NS1 _P 6-FAM-ATGGTGCTGYAGRGARTGCACAATGC-BHQ Bonn_NS1_ REV GCCTTATCTCCATTCCATACC Bonn_ E_FWD AGYCGYTGYCCAACACAAG Bonn _E_P 6-FAM-CCTMCCTYGAYAAGCARTCAGACACYCAA-BHQ Bonn _E_REV CACCARRCTCCCYTTGCCA

Reference: Positive Control: We will provide controls through the European Virus Archive (EVA) http://global.european-virus-archive.com

25µl Cycler: MasterMix: single rxn, μl H2O (RNAse free) 1.4 50°C 15’ MgSO4(50mM) 0.4 95°C 3’ 2x Reaction mix 12.5 BSA (1 mg/ml)* 1 95°C 15’’ Fwd primer (10 µM) 1.5 56°C 20’’ 45x Rev primer (10 µM) 1.5 72°C 15“ Probe (10 µM) 0.7 SSIII/Taq EnzymeMix* 1 40°C 30’’ 20

' = minutes; " = seconds Template RNA 5

Supplementary Figure S2. Genetic relationship of representative flaviviruses.

Maximum Likelihood phylogeny of the genus Flavivirus. Filled circles at nodes indicate bootstrap supports above 75% (1,000 replicates). Viruses are identified by name and GenBank accession number. Viruses included in specificity testing are highlighted with an asterisk. Viruses known to occur in humans are given in red, viruses co-circulating with ZIKV in the current outbreak are given in bold. Vector associations are indicated to the right. Tick-borne flaviviruses were collapsed for graphical reasons and included Gadgets Gully virus (DQ235145); Kyasanur Forest disease virus (AY323490); Alkhurma hemorrhagic fever virus (AF331718); Langat virus (AF253419); Louping ill virus, Spanish subtype (DQ235152); Louping ill virus, Turkish sheep encephalitis virus subtype (DQ235151); Louping ill virus, Greek goat encephalitis virus subtype (DQ235153); Omsk hemorrhagic fever virus (AY323489); Powassan virus (L06436); Royal Farm virus (DQ235149); Tick-borne encephalitis virus, European subtype (U27495); Tick-borne encephalitis virus, Far Eastern subtype (JN229223); Tick-borne encephalitis virus, Siberian subtype (L40361); Meaban virus (DQ235144); Saumarez Reef virus (DQ235150); Tyuleniy virus (DQ235148); Kadam virus (DQ235146) . ML phylogenies were calculated using a complete deletion option, a HKY nucleotide substitution model and 1,000 bootstrap replicates in MEGA6 (26).

Supplementary Figure S3. Analytical sensitivity of Zika virus real-time RT-PCR tests.

Probability of detection (y-axis) is plotted against ucRNA concentration per reaction in 8 parallel test

samples (x-axis), with data points representing the observed fraction of positive results in parallel

experiments. Solid line, predicted proportion of positive results at a given RNA input concentration;

dashed lines, 95% confidence limits for the prediction. In each panel, assay name, 95% lower limit of

detection and confidence intervals are indicated. Probit analyses were done using SPSS V22 (IBM,

Ehningen, Germany).

Supplementary Figure S4. Zika virus loads in clinical specimens per day after symptom onset.

Zika virus loads in different types of clinical specimens plotted per day after symptom onset. Each color corresponds to an individual patient. Datum points to the right represent unknown onset of symptoms.