A Founder Effect Led Early SARS-CoV-2 Transmission in Spain

Abstract

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) whole-genome analysis has identified five large clades worldwide which emerged in 2019 (19A and 19B) and in 2020 (20A, 20B, and 20C). This study aimed to analyze the diffusion of SARS-CoV-2 in Spain using maximum-likelihood phylogenetic and Bayesian phylodynamic analyses. The most recent common ancestor (MRCA) of the SARS-CoV-2 pandemic was estimated to have emerged in Wuhan, China, around 24 November 2019. Phylogenetic analyses of the first 12,511 SARS-CoV-2 whole-genome sequences obtained worldwide, including 290 from 11 different regions of Spain, revealed 62 independent introductions of the virus in the country. Most sequences from Spain were distributed in clades characterized by a D614G substitution in the S gene (20A, 20B, and 20C) and an L84S substitution in ORF8 (19B) with 163 and 118 sequences, respectively, with the remaining sequences branching in 19A. A total of 110 (38%) sequences from Spain grouped in four different monophyletic clusters of clade 20A (20A-Sp1 and 20A-Sp2) and 19B clade (19B-Sp1 and 19B-Sp2) along with sequences from 29 countries worldwide. The MRCAs of clusters 19A-Sp1, 20A-Sp1, 19A-Sp2, and 20A-Sp2 were estimated to have occurred in Spain around 21 and 29 January and 6 and 17 February 2020, respectively. The prevalence of clade 19B in Spain (40%) was by far higher than in any other European country during the first weeks of the epidemic, probably as a result of a founder effect. However, this variant was replaced by G614-bearing viruses in April. In vitro assays showed an enhanced infectivity of pseudotyped virions displaying the G614 substitution compared with those having D614, suggesting a fitness advantage of D614G.IMPORTANCE Multiple SARS-CoV-2 introductions have been detected in Spain, and at least four resulted in the emergence of locally transmitted clusters that originated not later than mid-February, with further dissemination to many other countries around the world, and a few weeks before the explosion of COVID-19 cases detected in Spain during the first week of March. The majority of the earliest variants detected in Spain branched in the clade 19B (D614 viruses), which was the most prevalent clade during the first weeks of March, pointing to a founder effect. However, from mid-March to June 2020, G614-bearing viruses (clades 20A, 20B, and 20C) overcame D614 variants in Spain, probably as a consequence of an evolutionary advantage of this substitution in the spike protein. A higher infectivity of G614-bearing viruses than D614 variants was detected, suggesting that this substitution in SARS-CoV-2 spike protein could be behind the variant shift observed in Spain.

Keywords: COVID-19; Europe; SARS-CoV-2; Spain; phylodynamics; phylogeography.

FIG 1

Origin and phylogeny of SARS-CoV-2 sequences from Spain. (A) Whole-genome sequencing of 61 samples from different locations (red circles) was carried out, and 229 other sequences from Valencia, Madrid, and Barcelona were retrieved from GISAID. (B) Phylogenetic tree of SARS-CoV-2 whole-genome sequences (n = 12,511) inferred by FastTree software v2.1.11 using a generalized time-reversible (GTR) model of nucleotide evolution. Sequences included in clades that emerged in 2019 (orange) and 2020 (green) are differentiated. The sequences from Spain are highlighted in pink to differentiate the 62 independent introductions observed in the country. The four monophyletic clusters with a probable origin in Spain (19B-Sp1, 19B-Sp2, 20A-Sp1, and 20A-Sp2) are indicated.

FIG 2

Maximum clade credibility (MCC) tree of cluster 19B-Sp1. Branch colors indicate the most probable location of the MRCA, and node labels indicate the posterior probability supporting the estimated MRCA location. Node support values are indicated by node size (only nodes with PP of ≥0.8 are considered well supported). The scale axis represents estimated dating of the MRCA for each cluster, and label spacing defines exactly 9.16 days from the most recent sample included in the analysis.

FIG 3

Maximum clade credibility (MCC) tree of cluster 19B-Sp2. Branch colors indicate the most probable location of the MRCA, and node labels indicate the posterior probability supporting the estimated MRCA location. Node support values are indicated by node size (only nodes with PP of ≥0.8 are considered well supported). The scale axis represents estimated dating of the MRCA for each cluster, and label spacing defines exactly 9.16 days from the most recent sample included in the analysis.

FIG 4

Maximum clade credibility (MCC) tree of cluster 20A-Sp1. Branch colors indicate the most probable location of the MRCA, and node labels indicate the posterior probability supporting the estimated MRCA location. Node support values are indicated by node size (only nodes with PP of ≥0.8 are considered well supported). The scale axis represents estimated dating of the MRCA for each cluster, and label spacing defines exactly 9.16 days from the most recent sample included in the analysis.

FIG 5

Maximum clade credibility (MCC) tree of cluster 20A-Sp2. Branch colors indicate the most probable location of the MRCA, and node labels indicate the posterior probability supporting the estimated MRCA location. Node support values are indicated by node size (only nodes with PP of ≥0.8 are considered well supported). The scale axis represents estimated dating of the MRCA for each cluster, and label spacing defines exactly 9.16 days from the most recent sample included in the analysis.

FIG 6

Maximum clade credibility (MCC) tree of the cluster including one sequence of the Contamines-Monjoie transmission chain (EPI-ISL-410486). Branch colors indicate the most probable location of the MRCA, and node labels indicate the posterior probability supporting the estimated MRCA location. Node support values are indicated by node size (only nodes with PP of ≥0.8 are considered well supported). The scale axis represents estimated dating of the MRCA for each cluster, and label spacing defines exactly 9.16 days from the most recent sample included in the analysis. The red star indicates the sequence associated with the Contamines-Monjoie cluster, and “V” indicates the presence of the G251V mutation in ORF3a protein.

FIG 7

Founder effect of D614-variants in Spain and selective advantage of D614G substitution. (A) D614/G614 prevalence over time in Spain according to the data available on 8 July 2020. Increased infectivity of D614G pseudoviruses. VeroE6, Caco-2, and 293T cells expressing human ACE2 were infected with equal amounts (10 ng of Gag p24) of D614 or G614 pseudoviruses. Viral infectivity was determined by measuring luciferase activity in the cell lysates 48 h postinfection and is expressed as percent infectivity relative to that of D614 virus. Results are means and standard deviations for an experiment performed in triplicate. Results of a representative experiment out of three independent assays is shown. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01 (unpaired two-tailed Student's t test).