key: cord-335652-v98gv5uf
authors: Salazar, Cecilia; Díaz-Viraqué, Florencia; Pereira-Gómez, Marianoel; Ferrés, Ignacio; Moreno, Pilar; Moratorio, Gonzalo; Iraola, Gregorio
title: Multiple introductions, regional spread and local differentiation during the first week of COVID-19 epidemic in Montevideo, Uruguay
date: 2020-05-10
journal: bioRxiv
DOI: 10.1101/2020.05.09.086223
sha: 
doc_id: 335652
cord_uid: v98gv5uf

Background After its emergence in China in December 2019, the new coronavirus disease (COVID-19) caused by SARS-CoV-2, has rapidly spread infecting more than 3 million people worldwide. South America is among the last regions hit by COVID-19 pandemic. In Uruguay, first cases were detected on March 13 th 2020 presumably imported by travelers returning from Europe. Methods We performed whole-genome sequencing of 10 SARS-CoV-2 from patients diagnosed during the first week (March 16th to 19th) of COVID-19 outbreak in Uruguay. Then, we applied genomic epidemiology using a global dataset to reconstruct the local spatio-temporal dynamics of SARS-CoV-2. Results Our phylogeographic analysis showed three independent introductions of SARS-CoV-2 from different continents. Also, we evidenced regional circulation of viral strains originally detected in Spain. Introduction of SARS-CoV-2 in Uruguay could date back as early as Feb 20th. Identification of specific mutations showed rapid local genetic differentiation. Conclusions We evidenced early independent introductions of SARS-CoV-2 that likely occurred before first cases were detected. Our analysis set the bases for future genomic epidemiology studies to understand the dynamics of SARS-CoV-2 in Uruguay and the Latin America and the Caribbean region.

In December 2019, a new Coronavirus disease (COVID-19) was detected in Wuhan, China. Its causative agent, known as SARS-CoV-2, has spread rapidly causing a global pandemic of pneumonia affecting more than 3M people with more than 200,000 deaths to date [1] . Global spread of SARS-CoV-2 was primarily driven by international movement of people, since travel-associated cases of COVID-19 were reported outside of China as early as mid-January 2020 [2] . This global health emergency has deployed international efforts to apply genomic epidemiology to track the spread of SARS-CoV-2 in real time. The recent development of targeted sequencing protocols by the ARTIC Network [3] , open sharing of genomic data through the GISAID (www.gisaid.org) database and straightforward bioinformatic tools for viral phylogenomics [4] , provides the opportunity to reconstruct global spatio-temporal dynamics of the COVID-19 pandemic with unprecedented comprehensiveness and resolution.

South America was one of the last regions in the world to be hit by COVID-19. Indeed, first cases were confirmed in Brazil around 2 months after its emergence in China. Then, SARS-CoV-2 rapidly emerged in neighboring countries like Uruguay, that reported first cases in Montevideo, its capital city, on March 13 th .

Population of Uruguay is among the smallest in South America (~3.5M), and is mainly concentrated in Montevideo and its metropolitan area (~56% of the total population). Also, the biggest international airport of the country is placed in this metropolitan area and concentrates more than 90% of international arrivals.

These characteristics makes Montevideo a suitable place to apply genomic surveillance to uncover epidemiological patterns during the early phase of the COVID-19 local outbreak in Uruguay. We therefore aimed to characterize the spatio-temporal dynamics of SARS-CoV-2 by sequencing around 10% of cases occurred during the first week of outbreak in Montevideo, allowing us to identify transmission patterns, geographic origins and genetic variation among local strains.

Ethics statement. Residual de-identified nasopharyngeal samples were remitted to the Institut Pasteur Montevideo, that has been validated by the Ministry of Health of Uruguay as an approved center providing diagnostic testing for COVID-19. All samples were de-identified before receipt by the study investigators.

Sample collection, processing and sequencing. Nasopharyngeal swabs were obtained from patients residing in Montevideo. No other location data was available to researchers to avoid patient identification. Swabs were placed in virus transport media (BD Universal Viral Transport Medium) immediately after collection.

Total RNA extraction was performed directly from samples (500 μL) using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA). An in-house OneStep RT-qPCR assay based on TaqMan probes targeting the N gene and the Open Reading Frame 1b region were used to test for the presence of SARS-CoV-2 RNA.

We selected a total of 10 positive samples with cycle-threshold (CT) less than 30 for downstream wholegenome sequencing. Sequencing of SARS-CoV-2 positive samples was performed according to the primalseq [5] approach on the MinION sequencing platform (Oxford Nanopore Technologies, United Kingdom) using the V2 primer pools. Sequencing libraries were prepared based on the PCR tiling of COVID-19 virus protocol (ARTIC Network) using the Ligation Sequencing Kit (SQK-LSK109) and Native Barcoding Expansion (EXP-NBD114) (Oxford Nanopore Technologies, United Kingdom), with the following modifications: i) RNA was reverse-transcribed using Superscript II reverse transcriptase (Invitrogen Life Technologies, Carlsbad, CA, USA) and random hexamer primers following the manufacturer's instructions, ii) Phusion Hot Start II High-Fidelity PCR Master Mix (Thermo Scientific) was used for PCR amplification and, iii) Blunt/TA ligase (New England Biolabs, USA) was used to ligate barcodes to each sample. To detect crosscontamination between samples, a no template sample was added from a patient tested negative for SARS-CoV-2 from the RNA purification step.

A total of 15 ng of pooled samples was loaded onto a MinION R9.4.1 flow cell and sequenced for 5 hours generating ~2 million reads of average quality of 12. The RAMPART software from the ARTIC Network (https://github.com/artic-network/rampart) was used to monitor the sequencing run in real time, estimate coverage across samples and check barcoding. Subsequently, basecalling was performed Guppy software v3.2.1 using the high accuracy module. Consensus genomes were generated and variants were called using the ARTIC Network bioinformatic pipeline (https://artic.network/ncov-2019/ncov2019bioinformatic.sop.html). Amplicons that were not sequenced or whose depth was less than 20x were not included in the consensus sequences and these positions were represented by "N" stretches. See Supplementary Table S1 .

Phylogenetic and spatio-temporal analysis. To investigate the origin, dynamics and genetic variation of SARS-CoV-2 in Uruguay, we added our 10 genomes to a dataset of 2,443 complete (>29,000 bp.), highquality genomes available from GISAID (https://www.gisaid.org) on April 2 nd 2020. A list of sequences and acknowledgements to the originating and submitting labs is presented in Supplementary Table S2 . We analyzed this dataset using the augur toolkit version 6.4.2 [3] . Briefly, genomes were aligned using MAFFT [6] and a phylogeny was reconstructed with IQ-TREE [7] . Estimation of ancestral divergence times and geographic locations for internal nodes of the tree, and identification of branch-specific or clade-defining mutations were obtained using augur and TreeTime [8] . Code for performing these analyses can be found at http://github.com/giraola/covid-19-uruguay and results can be visualized at https://nextsrain.org/community/ giraola/covid19-uruguay. genomes (hereinafter referred as cluster C3-UY) were related each other and were placed within clade B (Fig. 2) . These clusters were supported by specific, non-homoplasic mutations (Table 1) . Together, these results indicate three independent introductions of SARS-CoV-2 into Uruguay.

Early SARS-CoV-2 circulation and regional spread. To uncover the spatio-temporal dynamics of SARS-CoV-2 in Uruguay, we performed a phylogeographic reconstruction. The most probable ancestral location for C1-UY was Australia, however, very similar viruses have been also sequenced from the United Kingdom (Fig. 3) . The last common ancestor of C1-UY in Australia dated back to Mar 5 th (Confidence Interval (CI):

Feb 28 th -Mar 13 rd ). C2-UY was embedded in a cluster of viruses sequenced in Canada, the United States, Australia and Iceland whose emergence dated back to February 28 th (CI: Feb 25 th -Mar 4 th ) (Fig. 4) . C3-UY likely originated from European viruses circulating in Spain since Jan 21 th (CI: Jan 7 th -Feb 5 th ). The last common ancestor of C3-UY in Uruguay was tracked to Mar 4 th (CI: Feb 23 rd -Mar 11 th ) (Fig. 5) . Two viruses from Chile were intermingled between viruses from Spain and Uruguay, indicating regional circulation of these variants. Together, these results indicate that SARS-CoV-2 could be circulating in the country previous to its first detection in Mar 13 th .

Local clusters and genetic differentiation. To further characterize Uruguayan SARS-CoV-2, we identified genetic differences among genomes. C1-UY ancestor likely located in Australia presented one nonsynonymous mutation C3096T which affected the ORF1ab protein (S955L) and genome UY-10 harbored one additional synonymous mutation A29731G. C2-UY ancestor likely located in Canada was characterized by one synonymous mutation A24694T and the last common ancestor to UY-7 and UY-8 presented another synonymous change A11782G. Additionally, genome UY-8 presented one synonymous mutation C17470T.

C3-UY and Chilean genomes displayed one synonymous mutations C17470T which distinguished them from the original Spanish lineage. Additionally, C3-UY differentiated from Chilean genomes by one synonymous mutation C25521T. Specifically, among C3-UY genome UY-5 presented a single synonymous mutation C6310T and genome UY-6 presented one non-synonymous mutation C7635A in the ORF1ab protein (T2457K). Overall, we observed mutations within Uruguayan clusters supporting rapid local differentiation (Table 2) .

Using portable, low-cost sequencing approaches based on the MinION platform (Oxford Nanopore Technologies) we were able to generate whole-genome sequences of SARS-CoV-2 just 24 hours after receiving the samples in the lab. This allowed us to determine the most plausible spatio-temporal scenario during the first week since COVID-19 was detected in Uruguay. Consequently, we show that real-time molecular epidemiology responses can be implemented locally as previously done by other countries [9] .

Our results uncovered SARS-CoV-2 genomes derived from three independent international origins as early as 3 days after the disease was reported in Uruguay. International air transport has likely played a central role in the rapid spread of SARS-CoV-2, contributing to the current COVID-19 pandemic. According to air transport statistics from the World Bank (https://data.worldbank.org) and the National Government We estimated these introductions could have occurred as early as mid-February, around 1 month before first cases were reported in Uruguay in March 13 th . This is in line with observations from other geographic regions, for example, in New York first cases were confirmed in March but a recent study suggested that initial virus introductions could be traced back to February 20 th [10] . Also, the virus was likely introduced in Europe by a traveler that arrived from Wuhan to France, but the recent identification of a group of ill Chinese tourists in France previous to this case supports earlier introductions [11] . Indeed, SARS-CoV-2 has been proposed to circulate in humans even before its first detection in December 2019 [12] .

Together, our results highlight the importance of active genomic surveillance during the ongoing COVID-19 pandemic to guide informed decisions, based on assessing the epidemiological behavior of SARS-CoV-2 in real time. The application of rapid genomic epidemiology during local outbreaks allows to identify main routes of virus introduction and estimate spatio-temporal dynamics which can be used to improve contingency measures. Remarkably, tracking the emergence of local genetic variants like those we observed in Montevideo, is important to evaluate the sensibility of molecular diagnosis and potential impacts on future anti-viral strategies. Subsequent generation of a more comprehensive genomic dataset from the ongoing outbreak in Uruguay will allow us to identify domestic transmission patterns and spatio-temporal characterization of local clusters. Additionally, setting up coordinated efforts to generate genomic data in South America will allow us to perform integrative analyses to uncover SARS-CoV-2 dynamics at the continental level. 

An interactive web-based dashboard to track COVID-19 in real time

Novel Coronavirus (2019-nCoV) Situation Report -1

Multiplex PCR method for MinION and Illumina sequencing of Zika and other virus genomes directly from clinical samples | Nature Protocols

Nextstrain: real-time tracking of pathogen evolution

An amplicon-based sequencing framework for accurately measuring intrahost virus diversity using PrimalSeq and iVar

MAFFT Multiple Sequence Alignment Software Version 7: Improvements in Performance and Usability

IQ-TREE: A Fast and Effective Stochastic Algorithm for Estimating Maximum-Likelihood Phylogenies

TreeTime: Maximum-likelihood phylodynamic analysis

Real-time, portable genome sequencing for Ebola surveillance

Introductions and early spread of SARS-CoV-2 in the New York City area

Early Release -Early Introduction of Severe Acute Respiratory Syndrome Coronavirus 2 into Europe

A Genomic Perspective on the Origin and Emergence of SARS-CoV-2

Acknowledgements. We thank Josh Quick and the ARTIC Network for providing SARS-CoV-2 sequencing primers and support with sequencing protocols. We also thank to everyone who openly shared their genomic