key: cord-0829751-p7r91vx8
authors: Wang, Wenqiang; Han, Guan-Zhu
title: Ancient Adaptative Evolution of ACE2 in East Asians
date: 2021-07-24
journal: Genome Biol Evol
DOI: 10.1093/gbe/evab173
sha: 7dd30333c2e86cc72e061eee0056909e1e147068
doc_id: 829751
cord_uid: p7r91vx8

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been posing an unprecedented challenge to global public health. SARS-CoV-2 and several other coronaviruses utilize angiotensin-converting enzyme 2 (ACE2) as their entry receptors. The ACE2 gene has been found to experience episodic positive selection across mammals. However, much remains unknown about how the ACE2 gene evolved in human populations. Here, we use population genetics approaches to investigate the evolution of the ACE2 gene in 26 human populations sampled globally. We find the ACE2 gene exhibits an extremely low nucleotide diversity in the East Asian populations. Strong signals of selective sweep are detected in the East Asian populations, but not in the other human populations. The selective sweep in ACE2 is estimated to begin in East Asian populations ∼23,600 years ago. Our study provides novel insights into the evolution of the ACE2 gene within human populations.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the pathogen that causes Coronavirus Disease 2019 , has been posing an unprecedented threat to global public health, resulting in more than 179 million infections and 3.90 million deaths (https://www.who.int/emergencies/diseases/novel-coronavirus-2019; as to June 26, 2021). To date, at least seven coronaviruses have been known to infect humans, including human coronavirus (HCoV)-229E, HCoV-NL63, HCoV-OC43, HCoV-HKU1, Middle East respiratory syndrome coronavirus (MERS-CoV), SARS-CoV, and SARS-CoV-2 (Drosten et al. 2003; Pene et al. 2003; Vabret et al. 2003; van der Hoek et al. 2005; Woo et al. 2005; Zaki et al. 2012; Zhu et al. 2020 ). Among them, four coronaviruses, namely HCoV-229E, HCoV-NL63, HCoV-OC43, and HCoV-HKU1, are globally endemic and usually cause mild to moderate respiratory infections (Gaunt et al. 2010) . Three novel coronaviruses, namely SARS-CoV, MERS-CoV, and SARS-CoV-2, with high case fatality rate have emerged in human populations in the past two decades (Petrosillo et al. 2020) .

Binding of cellular surface protein known as viral receptor by spike (S) protein is the initial step of coronavirus infection. To date, the receptors of seven HCoVs have been identified: HCoV-OC43 and HCoV-HKU1 interact with 9-O-acetylsialic acid to invade host cells (Vlasak et al. 1988; Huang et al. 2015) ; HCoV-229E and MERS-CoV utilize amino-peptidase N (ANPEP) and dipeptidyl-peptidase 4 (DPP4) as their receptors, respectively (Yeager et al. 1992; Widagdo et al. 2016 ); SARS-CoV, HCoV-NL63, and SARS-CoV-2 utilize angiotensinconverting enzyme 2 (ACE2) as their receptors to initialize viral infection (Li et al. 2003; Hofmann et al. 2005; Zhou et al. 2020) . ACE2 and its homolog ACE are the main enzymes of renin-angiotensin system (RAS), which is a key regulator of maintaining blood pressure homeostasis and fluid and salt Significance SARS-CoV-2 utilizes angiotensin-converting enzyme 2 (ACE2) as its entry receptor, but little is known about the evolution of the ACE2 gene in human populations. We detect strong signals of selective sweep in the East Asian populations, but not in other human populations. Our study provides novel insights into the evolution of ACE2 within the human populations.

balance (Tipnis et al. 2000; Bosso et al. 2020) . The octapeptide angiotensin (ANG) II, an intermediate produced by ACE in RAS, promotes vasoconstriction, inflammatory, and fibrotic properties (Gaddam et al. 2014) . ACE2 can convert Ang II to Ang (1-7), which mediates vasodilatation (Warner et al. 2004; Gaddam et al. 2014 ). Thus ACE2 acts as a negative regulator in the RAS to counterbalance the effects of ACE (Kuba et al. 2006) . The human ACE2 gene encodes an 805 amino acid protein with a highly conserved active zinc-binding motif His-Glu-X-X-His (HEXXH motif) and is located on chromosome X (Hamming et al. 2007 ).

Viral receptors, essentially "normal" host cellular proteins hijacked by viruses, have been thought to be subject to two conflicting directional forces, namely negative selection due to functional constraints to maintain their own cellular function and positive selection due to everchanging host-virus arms race (Wang et al. 2020) . The ACE2 gene has been found to experience episodic positive selection across mammals, including bats and primates (Demogines et al. 2012; Damas et al. 2020; Wang et al. 2020) . Many positively selected sites overlap the interaction interface between ACE2 and coronaviruses, implying that ancient recurrent evolutionary arms races occurred between mammals and coronaviruses (Demogines et al. 2012; Damas et al. 2020; Wang et al. 2020) . Our previous study indicates that the ACE2 gene is likely to have experienced local adaptation in the Chinese Han population (Wang et al. 2020 ). However, much remains unknown about the evolution of the ACE2 gene in human populations. In this study, we used population genetics approaches to explore the evolution of the ACE2 gene in a total of 26 human populations sample globally, and detected strong signals of selective sweep in East Asian populations.

Here, we employed a series of population genetics approaches to analyze the evolution of the ACE2 gene in 26 human populations sampled globally, including seven African populations, four American populations, five East Asian populations, five European populations, and five South Asian populations (Sudmant et al. 2015) . Initially, we found that the five East Asian populations, namely CDX (Chinese Dai in Xishuangbanna, China), CHS (Han Chinese South), CHB (Han Chinese), JPT (Japanese in Tokyo, Japan), and KHV (Kinh in Ho Chi Minh City, Vietnam), exhibit much lower nucleotide diversity than the other human populations ( fig. 1A ), raising the possibility that the ACE2 gene underwent natural selection in East Asian populations. To investigate the possibility of natural selection shaping the evolution of ACE2 in human populations, we first employed the neutrality test to detect selection signals. We found that the ACE2 gene displays a D value of significantly less than 0 in all the five East Asian populations, CDX (D ¼ À1.75; P ¼ 0.009 and P ¼ 0.028 for coalescence simulations under consistent population size and population growth, respectively), CHB (D ¼ À1.93; P ¼ 0.006 and P ¼ 0.015), CHS (D ¼ À1.74; P ¼ 0.013 and P ¼ 0.031), JPT (D ¼ À1.66; P ¼ 0.022 and P ¼ 0.040), and KHV (D ¼ À1.82; P ¼ 0.009 and P ¼ 0.021). In contrast, no significant signal of natural selection was detected in other human populations ( fig. 1B) . Moreover, we also merged the population genetic data based on the continents (AFR, African; AMR, Ad Mixed American; EAS, East Asian; EUR, European; SAS, South Asian). The continent-level analyses show the ACE gene of the East Asian super-population exhibits the lowest nucleotide diversity, and signal of selective sweep was only detected in the East Asian super-population (D ¼ À2.233; P < 0.001) (supplementary table S1, Supplementary Material online). These results suggest that the East Asian populations might have experienced a common selective sweep in the past.

We also performed integer neighbor joining network analyses to investigate haplotype structures of ACE2 among different human populations. The haplotype networks for all the five East Asian populations display a star-like structure, and many haplotypes of relatively low frequency are connected with the main haplotypes with short branches (supplementary fig. S1 , Supplementary Material online), supporting selective sweep occurring in East Asian populations. Unlike East Asian populations, the ACE2 haplotypes in non-East Asians appear to be more scattered with long step length (supplementary fig. S1 , Supplementary Material online). Statistical analyses based on haplotypes also show that five East Asian populations exhibit the lowest nucleotide diversity and D values of significantly less than 0 (supplementary table S2, Supplementary Material online). Moreover, we merged the population genetic data according to the continent of populations for selection analysis. We found that the pattern of the main haplotype surrounding excess low-frequency haplotypes became more pronounced in the East Asian populations (supplementary fig. S1 and table S1, Supplementary Material online). These results further confirm that the ACE2 gene of East Asian populations is likely to have underwent a selective sweep.

We also employed the integrated haplotype score (iHS) statistic to detect evidence of selective sweep in the ACE2 gene in human populations. We found that four East Asian populations possess a high proportion of SNPs with extremely high jiHSj values within the ACE2 gene (80% in CDX, 69% in CHB, 50% in JPT, 80% in KHV; fig. 2 fig. 2 ). Moreover, we also merged the population genetic data based on the continents. The continent-level analyses show that 60% of SNPs within the ACE2 gene display extremely high jiHSj values in East Asian populations, whereas no SNP with extremely high jiHSj values was identified in the human populations from the other continents (supplementary fig. S2 , Supplementary Material online). Taken together, our results further support that the ACE2 gene might have experienced a unique selective sweep in the East Asian populations.

To date the selective sweep occurring in the ACE2 gene in the East Asian populations, we used a method based on ancestral recombination graph to estimate the starting time of selection for each SNPs of the ACE2 gene. The selective sweep in ACE2 was estimated to begin in East Asian populations approximately 844 generations ago ($23,630 years, with generation time of humans to be 28 years) ( fig. 3 ).

In this study, we found strong signal of selective sweep in the ACE2 gene in East Asian populations. The selective sweep was estimated to begin near 25,000 years ago. What drove the adaptive evolution of ACE2 in East Asian populations? ACE2 is the receptor for at least three human coronaviruses, including SARS-CoV, SARS-CoV-2, and HCoV-NL63 (Li et al. 2003; Hofmann et al. 2005; Zhou et al. 2020) . The arms race between viruses and hosts has been found to drive the accelerated adaptive evolution in viral receptors (Wang et al. 2020) . Previous studies show positively selected sites detected in bat ACE2 genes overlap almost perfectly to its interaction interface with SARS-CoV and HCoV-NL63, indicating ACE2 may have been utilized by coronaviruses for millions of years (Demogines et al. 2012) . Therefore, we propose that an ACE2-utilizing virus epidemic is likely to occur in East Asia around 25,000 years ago and might have been circulating for some time, which might have driven the adaptive evolution of ACE2 in the common ancestor of East Asians. Consistent with this possibility, unique signals of selective sweep in East Asians were also detected in 42 other proteins that interact with coronaviruses, and the selective sweep was estimated to start around 25,000 year ago (Souilmi et al. 2021) . However, it is hard to find direct evidence for the occurrence of the ancient virus epidemic. Moreover, the possibility that nonviral factors have driven the adaptive evolution of the ACE2 gene in East Asians cannot be unambiguously excluded. Nevertheless, our study provides novel insights into the evolution of ACE2 in human populations.

The genetic variation data of the ACE2 gene from 26 human populations were retrieved from the pilot 3 phase of the 1,000 Genomes Project (Sudmant et al. 2015) . Considering that ACE2 is located on the X chromosome, only female population genetic data were utilized in this analysis to avoid the conflict caused by the coexistence of different ploidies. The nucleotide diversity (p) and Tajima's D values of the ACE2 gene were estimated using the DNASP v6, and the statistical significance of the neutrality test is evaluated based on coalescent simulations with constant population size and no recombination (Rozas et al. 2017) . For those populations with significant D values, we further conducted 1,000 coalescence simulations using the msms software (Ewing and Hermisson 2010 ) under the demographic model of population growth (Gravel et al. 2011 ) and with the recombination rate inferred from AAmap (Hinch et al. 2011) . The command used for simulation is: msms -N 10,000 -ms 50 1,000 -s 50 -r 23 41,000 -G 38.698 -eN 0.0315 0.2955 -oTPi. Tajima's D is considered to be different from 0, only when P values are significant in both simulations.

We also merged the population genetic data based on the five continents (AFR, African; AMR, Ad Mixed American; EAS, East Asian; EUR, European; SAS, South Asian) and performed neutrality test. AFR includes ESN ( 

Haplotype data were transformed from VCF files in the 1,000 Genomes Project. We reconstructed the haplotype network of human populations based on the integer neighbor joining method implemented in PopART v1.7 (Leigh and Bryant 2015) . Statistic based on haplotypes was also performed using PopART v1.7 (Leigh and Bryant 2015) .

Selscan was performed to evaluate iHS by scanning all the SNPs of chromosome X, and only biallelic SNPs with the lowest frequency greater than 0.05 were used in the analysis (Szpiech and Hernandez 2014) . After normalizing the unstandardized iHS scores separately in each population with 10 equally sized allele frequency bins by norm v1.3.0, we calculated the absolute value of these data, given that both the significantly high positive and negative values represent long haplotype homozygosity, and their difference is limited to whether ancestral alleles or derived alleles maintain long haplotype homozygosity. We ranked all the SNPs on chromosome X according to the jiHSj scores to delimit the top 5% region, and inferred whether SNPs of the ACE2 gene are enriched in the plateau area. We also merged the population genetic data based on the five continents and performed iHS test.

An ancestral recombination graph approach implemented in Relate was utilized to estimate the starting time of the selective sweep (Speidel et al. 2019) . The recombination rate data used for the analysis were extracted from Hinch's research (Hinch et al. 2011) . Relate calculates the P value based on the frequency of the derived alleles and the number of lineages in the general coalescent tree at specific generation to quantify the selection of each SNP. We identified the SNP with the lowest P value in the ACE2 gene, and the lowest P value must be less than 10 À3 to ensure the reliability of the selection. After identified the SNP with lowest P value, rs2106809, we calculated the selection time of the ACE2 gene based on the formula Time ¼ (upper generation þ lower generation)/ 2Âgeneration time, where the upper generation represents the generation with the most significant P value, and the lower generation represents the generation with the least significant P value but less than 0.05.

Supplementary data are available at Genome Biology and Evolution online.

This work was supported by the National Natural Science Foundation of China (31922001). 

The two faces of ACE2: the role of ACE2 receptor and its polymorphisms in hypertension and COVID-19

Broad host range of SARS-CoV-2 predicted by comparative and structural analysis of ACE2 in vertebrates

Evidence for ACE2-utilizing coronaviruses (CoVs) related to severe acute respiratory syndrome CoV in bats

Identification of a novel coronavirus in patients with severe acute respiratory syndrome

MSMS: a coalescent simulation program including recombination, demographic structure and selection at a single locus

ACE and ACE2 in inflammation: a tale of two enzymes

Epidemiology and clinical presentations of the four human coronaviruses 229E, HKU1, NL63, and OC43 detected over 3 years using a novel multiplex real-time PCR method

Demographic history and rare allele sharing among human populations

The emerging role of ACE2 in physiology and disease

The landscape of recombination in African Americans

Human coronavirus NL63 employs the severe acute respiratory syndrome coronavirus receptor for cellular entry

Human coronavirus HKU1 spike protein uses Oacetylated sialic acid as an attachment receptor determinant and employs hemagglutinin-esterase protein as a receptor-destroying enzyme

Angiotensin-converting enzyme 2 in lung diseases

PopART: full-feature software for haplotype network construction

Angiotensin-converting enzyme 2 is a functional receptor for the SARS coronavirus

Coronavirus 229E related pneumonia in immunocompromised patients

COVID-19, SARS and MERS: are they closely related?

DNA sequence polymorphism analysis of large data sets

An ancient coronavirus-like epidemic drove adaptation in East Asians from 25,000 to 5,000 years ago

A method for genome-wide genealogy estimation for thousands of samples

An integrated map of structural variation in 2,504 human genomes

Selscan: an efficient multithreaded program to perform EHH-based scans for positive selection

A human homolog of angiotensin-converting enzyme: cloning and functional expression as a captopril-insensitive carboxypeptidase

An outbreak of coronavirus OC43 respiratory infection in Normandy, France

Croup is associated with the novel coronavirus NL63

Human and bovine coronaviruses recognize sialic acid-containing receptors similar to those of influenza C viruses

Host-virus arms races drive elevated adaptive evolution in viral receptors

Angiotensin-converting enzyme-2: a molecular and cellular perspective

Differential expression of the Middle East respiratory syndrome coronavirus receptor in the upper respiratory tracts of humans and dromedary camels

Clinical and molecular epidemiological features of coronavirus HKU1 associated community acquired pneumonia

Human aminopeptidase N is a receptor for human coronavirus 229E

Isolation of a novel coronavirus from a man with pneumonia in Saudi Arabia

A pneumonia outbreak associated with a new coronavirus of probable bat origin

A novel coronavirus from patients with pneumonia in China