Introductions and evolutions of SARS-
CoV-2 strains in Japan
Reitaro Tokumasu
1
, Dilhan Weeraratne
2
, Jane Snowdon
2
, Laxmi Parida
3
, Michiharu Kudo
1
, Takahiko
Koyama
3
1
IBM Research- Tokyo, Tokyo, Japan
2
IBM Watson Health, Cambridge, MA 02142, USA
3
IBM TJ Watson Research Center, Yorktown Heights, NY 10598, USA
Correspondence to Takahiko Koyama (email: tkoyama@us.ibm.com, postal address: 1101 Kitchawan Rd,
Yorktown Heights, NY 10598, USA
Key Words: SARS-COV-2, variant, COVID-19, Japan, quarantine, strains
. CC-BY-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 14, 2021. ; https://doi.org/10.1101/2021.02.26.21252555doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
Abstract
COVID-19 caused by SARS-CoV-2 was first identified in Japan on January 15
th
, 2020, soon after the
pandemic originated in Wuhan, China. Subsequently, Japan experienced three distinct waves of the
outbreak in the span of a year and has been attributed to new exogenous strains and evolving existing
strains. Japan engaged very early on in tracking different COVID-19 strains and have sequenced
approximately 5% of all confirmed cases. While Japan has enforced stringent airport surveillance on
cross-border travelers and returnees, some carriers appear to have advanced through the quarantine
stations undetected. In this study 30493 genomes sampled in Japan were analyzed to understand the
strains, heterogeneity and temporal evolution of different SARS-CoV-2 strains. We identified 12 discrete
strains with a substantial number of cases with most strains possessing the spike (S) D614G and
nucleocapsid (N) 203_204delinsKR mutations. 155 distinct strains have been introduced into Japan and
39 of them were introduced after strict quarantine policy was implemented. In particular, the B.1.1.7
strain, that emerged in the United Kingdom (UK) in September 2020, has been circulating in Japan since
late 2020 after eluding cross-border quarantine stations. Similarly, the B.1.351 strain dubbed the South
African variant, P.1 Brazilian strain and R.1 strain with the spike E484K mutation have been detected in
Japan. At least 14 exogenous B.1.1.7 sub-strains have been independently introduced in Japan as of late
March 2021, and these strains carry mutations that give selective advantage including N501Y,
H69_V70del, and E484K that confer increased transmissibility, reduced efficacy to vaccines and possible
increased virulence. Furthermore, various strains, which harbor multiple variants in the PCR primers and
the probe developed by National Institute of Infectious Disease (NIID), are emerging. It is imperative
that the quarantine policy be revised, cross-border surveillance reinforced, and new public health
measures implemented to mitigate further transmission of this deadly disease and to identify strains
that may engender resistance to vaccines.
. CC-BY-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 14, 2021. ; https://doi.org/10.1101/2021.02.26.21252555doi: medRxiv preprint
Introduction
SARS-CoV-2, the etiological agent of COVID-19 was first identified in Wuhan, China in late 2019 before
rapid worldwide transmission in the first quarter of 2020. In just a year, the number of confirmed cases
exceeded 148 million globally with 3.1 million deaths as of April 27
th
, 2021
1
The actual number of
infections are likely much higher by accounting for asymptomatic cases and mild disease that do not get
tested and under reporting due to social stigma and discrimination associated with COVID-19 infections
2
.
The novel coronavirus quickly spread to neighboring Japan, with a confirmed case of an individual on Jan
15, 2020 with travel history to Wuhan
3
. Soon after the identification of the first case, a major initial
transmission catastrophe was averted when the government decreed the Diamond Princess cruise ship
with infected patients to be anchored off the coast of Japan and mandated a 14-day quarantine for all
passengers
4
. However, the number of confirmed infections has surpassed 572,000 with over 10
thousand reported fatalities in Japan
1
.
The COVID-19 pandemic has exerted an unprecedented stress on global health systems and has created
a ripple effect touching every rubric of human life. Particularly the impact on healthcare (including
mental health) and the underlying social, political, psychological and economic disruptions have had
profound ramifications. In Japan, the dichotomy between the public health safety and societal and
economic dynamics forced the postponement of the much-awaited summer Olympics and Paralympics
games in Tokyo. Nonetheless, Japan implemented a stringent surveillance process at airports and
seaports at the very early stages of the outbreak to monitor and quarantine travelers and repatriates
with COVID-19 infection. In short, the rigorous surveillance process required a negative COVID-19 test
prior to boarding, saliva antigen testing at the cross-border port-of-entry and, if positive, polymerase
chain reaction (PCR) confirmation and sequencing at the National Institute of Infectious Diseases. While
the surveillance process has been largely successful with 2392 patients detected and intercepted at
quarantine stations by end of March 2021
5
, there appear to be some patients harboring exogenous
strains with different haplotypes of the virus who were undetected at the port-of-entry. Japan has been
vested and engaged from the beginning of the pandemic to monitor genomic changes in SARS-CoV-2
and has sequenced remarkable 30493 genomes which are approximately 5% of all confirmed cases.
Notwithstanding the rigor of the public health measures, three discrete waves of the disease have been
observed in Japan and it’s plausible that different viral strains may have contributed to each spike.
Mutations are inevitable as viruses evolve as a mechanism to cope with selective pressure and confer
selective advantage. COVID-19 is the first pandemic to occur after inexpensive sequencing technologies
became widely available. SARS-CoV-2 accumulates mutations at the rate of 1.0 x 10-3 (per
site/genome/year), that corresponds to 2.5 mutations in a month
6
. However, as number of the
infections increases, the number of variants increase proportionally. SARS-CoV-2 has accumulated and
established multiple mutations within a year from the first published report of D614G in early April
2020
7
. New strains with spike N501Y mutation such as B.1.1.7 from England, B.1.351 from South Africa,
P.1 from Brazil and R.1 have recently emerged
8
. Previously, N501Y has been functionally characterized
and was reported to cause higher binding with ACE2
9
and another study reported that B.1.1.7 carrying
H69_V70del in addition to N501Y possess a higher infectivity rate of 75%
10
. Furthermore, B .1.351, P. 1
. CC-BY-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 14, 2021. ; https://doi.org/10.1101/2021.02.26.21252555doi: medRxiv preprint
and R.1 strains contain spike E484K and appear to confer reduced efficacy to the currently available
vaccines
11,12
.
In this work we have evaluated the publicly available SARS-CoV-2 genomes in Japan to elucidate
different viral strains that were exogenously introduced, to understand community transmission
patterns and to delineate founder strains that further evolve within a community.
. CC-BY-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 14, 2021. ; https://doi.org/10.1101/2021.02.26.21252555doi: medRxiv preprint
Methods
Between 1 February 2020 and 5 April 2021, genomes were downloaded from the publicly available viral
sequencing repositories, Global Initiative on Sharing All Influenza Data (GISAID)
13
, National Center for
Biotechnology Information (NCBI), and NGDC Genome Warehouse, and the National Microbiology Data
Center (NMDC); of the 939410 genomes downloaded, 30493 were from specimens collected in Japan.
Low quality genomes with gaps and ambiguous bases over 50 base pairs in length (excluding at the start
and the end of the genomes) were discarded. After completing quality control, variant analysis was
performed on 638543 including 29679 Japanese genomes with the method described previously
6
. In
brief, genomes were first aligned to the reference genome NC_045512 using The European Molecular
Biology Open Software Suite (EMBOSS) needle
14
with open gap penalty of 100 to filter out spurious
frameshifts. Next, differences with the reference genome were extracted and annotated using gene
definitions of SARS-CoV-2 as different variant types including missense, synonymous, non-coding and
indels. All the obtained variants including global cases are stored in Supplemental Table S1.
Subsequently, hierarchical clustering was performed on Japanese domestic cases to organize strains
with similar haplotypes to construct a variant graph (Figure 1B).
Haplotype defining major strains were extracted from the variant graph in Figure 1B. Numbers of
confirmed cases and deaths in Japan obtained from Our World in Data
15
were shown in Figure 2A.
Monthly occurrences of each strain defined by the haplotype and its active period with evolutionary
relationships are illustrated in Figure 2B.
A candidate parent of a particular strain was identified by the maximum variant approach using the
haplotype of the query strain
16
. A parental strain should have a haplotype which is a subset of the query
strain. Among ancestors obtained in the previous step, the closest ancestor is the one with maximal
number of matches between haplotypes. The resulting parent and child relationships for all Japanese
domestic cases are obtained in Supplemental Table S2.
Data on monthly COVID-19 positive cases at airport quarantine centers in Japan and number of
passengers though immigration were obtained from a report released by the Japanese Government
5,17
(Figure 4A). For each strain, we sought to identify a plausible parent including exogenous strains, which
share more common variants than any domestic strains. Number of exogenous strains for each month is
represent ed in Figure 4B. Among the exogen ous strains, strains whose all probably ancestors were
found after April 15th were considered to have advanced through quarantine stations undetected (Table
2).
Strains which NIID_N qPCR probe fails to detect were identified by counting mutations in primers and
probe sites. If any given strain has mutations involving multiple bases in the probe or in the primers, the
strain was considered undetectable (Table 3).
. CC-BY-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted May 14, 2021. ; https://doi.org/10.1101/2021.02.26.21252555doi: medRxiv preprint