Tracking Coronavirus Variants Shows Early Spread of the P.1 (Gamma) Variant in Southern Brazil

The World Health Organization (WHO) declared COVID-19 a public health emergency of international interest in 2020 (1). In fact, since the emergence of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), many different lineages of SARS-CoV-2 have been reported worldwide (2). The causative agent of COVID-19 proved to present high genetic diversity as it adapted and evolved quickly (2).

In Brazil, a Variant of Concern (VOC) P.1, also known as Gamma (3), emerged at the beginning of December 2020 and was associated with the second wave of infections in Manaus, the capital of the state of Amazonas (AM), North Brazil (4, 5). The P.1 lineage was first detected in four travelers returning to Japan from AM at the beginning of January 2021 (6). As soon as it was recognized as an emerging lineage, P.1 was confirmed in six Brazilian states in late February 2021 (4). According to Faria et al. (4), the common ancestor of the P.1 lineage emerged around November 15, 2020, in Manaus, 3–4 weeks before the first confirmed cases of this lineage in the region. Similarly, Naveca et al. (5), using Bayesian reconstruction and phylogeographic analyses, estimated that VOC P.1 arose in Manaus around late November 2020 and rapidly spread to other municipalities and neighboring countries.

In the state of Rio Grande do Sul (RS), Southern Brazil, the first report of the P.1 lineage was of a clinical specimen obtained on February 1st, 2021, from a patient in the city of Gramado who had no history of traveling or contact with any returning travelers (7). Subsequent to this first report, Silva et al. (8) described an infection caused by P.1 on November 30, 2020, in RS, 3 months earlier than the first evidence of this lineage in the state. However, the authors inferred that P.1 probably only started to circulate in RS later because they were unable to identify any other P.1 sequences between November 2020 and January 2021 in the region (8). Noteworthy, there was a high incidence of P.1 lineage in RS in February 2021, which coincided with a significant increase in hospitalizations (9).

A phylogenetic analysis indicated that the VOC P.1 descended from the B.1.1.28 lineage (4), which together with the B.1.1.33 lineage were the first dominant lineages in Brazil in 2020 (5). These lineages were gradually replaced by the P.2 Variant of Interest (VOI), which was later replaced by the P.1 lineage (4). The replacement by P.1 was an abrupt process in AM, since it became the dominant lineage in the region in less than 2 months (5).

The P.1 lineage was defined by 17 amino acid changes (including 10 in the Spike protein), three deletions, four synonymous mutations, and one four-base-pair nucleotide insertion compared to its immediate ancestor B.1.1.28 (4). The P.1 lineage-defining mutations in the Spike protein, especially in the receptor-binding domains (RBD) such as K417T, E484K and N501Y, are of concern because they may enhance ACE2 affinity and contribute to antibody evasion (10). Recently, new sublineages (P.1.2 to P.1.8) that have emerged from P.1 have been reported in Brazil (11, 12), which indicates a continuous evolution of the virus and the community spread in the country.

The aim of this study was to evaluate the SARS-CoV-2 lineages circulating in RS state, southernmost Brazil, from March 2020 to May 2021. We also investigated the genetic events associated with the emergence of the VOC P.1 and its sublineages in the region.

In this study, we analyzed 200 SARS-CoV-2 genomes from RS, Southern Brazil, from 2020 and 2021, mainly from Porto Alegre city (capital of RS) and its metropolitan region. Our data demonstrate that the lineages of SARS-CoV-2 circulating in 2020 were gradually replaced by the VOC P.1 which has become the most prevalent lineage in RS since February 2021 (9), as well as in some other Brazilian regions (4, 5, 12).

According to our results, most of the P.1 sequences carried the 22 lineage-defining mutations, including the three mutations in the Spike protein, RBD domain (K417T, E484K and N501Y) associated with evasion of the immune system and greater transmissibility (10). Moreover, all the P.1 sequences accumulated other important mutations, such as P314L in the ORF1ab and R203K/G204R in the Nucleocapsid gene. Furthermore, 41 (38.7%) sequences presented the D614G mutation in the Spike gene and, after its first description in February 2020, it has been identified in all VOCs (14). These mutations, in addition to the 5′UTR:241 mutation, appear to dominate the most frequent nonlineage-defining mutations worldwide and their stability suggests virus adaptation to humans (15).The phylogenetic tree of the P.1 sequences obtained from this study revealed a high degree of genetic diversity (Fig. 1). The sequences were grouped in several distinct clades, suggesting multiple introductions of this lineage in Southern Brazil. It was possible to identify specific mutations for five defined subclades, most of them rarely found in deposited genomes (less than 15 occasions; accessed on https://nextstrain.org/↗ 20 July 2021) and whose significance is unknown so far. However, the comparative phylogenetic tree indicated that six sequences (6_LABRESIS, 90_LABRESIS, 101_LABRESIS, 195_LABRESIS, 198_LABRESIS, and 211_LABRESIS) were clustered together in subclade 4, assigned as P.1.2 (Fig. 1). The P.1.2 sublineage was identified in Southern Brazil in February 2021 and later it was also detected in other Brazilian regions and other countries (11). This sublineage evolved from parental P.1 and, according to our data, its occurrence increased over time (Fig. S1; Table S1). The P.1.2 sublineage carries some different mutations in comparison to the P.1 lineage and it was considered a Variant of Interest (VOI) due to the mutation D155Y in the ORF3a, which results in a low viral affinity to the host caveolin-1 protein, thereby avoiding cell apoptosis and extending the asymptomatic phase of infection (11, 16). As reported recently, the pattern of accumulated mutations in some P.1 sublineages seems to increase the transmissibility of the virus. However, it does not necessarily increase its capability to evade immune system among vaccinated population. Consequently, these sublineages could replace parental P.1 mainly in the community with low vaccination coverage (12).

The VOC P.1 was first detected in December 2020 in Manaus city (Amazonas state), and in January 2021 the cases due to the P.1 variant represented 85.4% (41/48) of the identified SARS-CoV-2 sequences in the city (4). As soon as it was recognized as an emergent lineage in late February 2021, P.1 had been confirmed in six Brazilian states (4). The prevalence of the P.1 variant increased extremely rapid by January 2021 in Brazil and other countries (6, 12, 17). In February 2021, it reached nearly 40% of sequenced cases in South America. In the United States, the VOC P.1 first appeared in mid-January. During the last 2 weeks in April 2021, the P.1 variant represented an estimated 5.0% of U.S. SARS-CoV-2 infections, demonstrating an emergence of the P.1 VOC in the country (18). According to WHO, the P.1 (Gamma) variant has already reached 91 different countries (14).

In the RS state, the first report of the VOC P.1 was in the late January 2021 (6). After this first description in RS, Silva et al. reported that the P.1 lineage had already been circulating in the state in late November 2020 but was not disseminated among individuals in the region (7). Moreover, Naveca et al. (5) reported that the P.1 and P.1-like variants share a ancestor dated from August 2020 and suggested that P.1 was circulating in Amazonas before the outbreak (late December 2020), even though this variant was not found in specimens collected between October and November 2020.

Our findings demonstrated that the divergence time of the oldest P.1 sequence identified in Brazil (Fig. S3; dated from around 10 July, 2020) is greater than the oldest specimen from AM (AM-FIOCRUZ-20842882CA; Fig. S3; dated from around 24 November, 2020). The same result was observed for the two oldest P.1 sequences from RS, both collected in November 2020 (164_LABRESIS and RS-LMM38991), which were obtained from patients with no history of travel or contact with returning travelers from Manaus (8, 19). In fact, according to the molecular clock, the common ancestor of these sequences from RS is dated to mid-June/July, earlier than those previously reported from AM (4, 5).

Besides, the time-scaled inference (Fig. 2) provided evidence of at least 3 major genetic diversity events, related to the evolution of P.1, and suggests that the sequences with the longest divergence time from different states of Brazil share a ancestor with those from AM. However, we observed that the 164_LABRESIS and RS-FIOCRUZ-7363 sequences from RS share an ancestral node of unknown origin, which could be from AM or even from RS. These findings corroborate the idea that the P.1 lineage was introduced on multiple genetic events in the country (5) and could have arisen in other Brazilian states before Amazonas.

In conclusion, our results demonstrated a high degree of genetic diversity among the P.1 genome sequences, which suggests the continuous evolution and community spread of the virus. Although the first VOC P.1 outbreak was reported in Manaus city, the lineage had already been circulating in Southern Brazil before November 2020. Nevertheless, it is well known that the P.1 spread faster in Manaus although the reasons for this are unclear. Thereby, the monitoring of virus evolution is an important tool to understand the emergence and dissemination of the old and new SARS-CoV-2 lineages, in particular the VOCs.