L 77 - 2020-21 - Endeligt svar på spørgsmål 186: Spm. om der er andre sundhedsfaglige kilder, der har udtrykt samme bekymring og slået tvivl om en fremtidig vaccines effekt, til sundheds- og ældreministeren

Miljø- og Fødevareudvalget 2020-21
L 77
Offentligt

Sundheds- og Ældreudvalget 2020-21

SUU Alm.del - Bilag 105

Offentligt

Dato 9. november 2020

Sagsnr. 2020110983

Implications of mutations in the spike protein of Danish mink-de-

rived SARS-CoV-2 isolates for vaccines and monoclonal antibody

therapeutics.

Purpose:

In this text, the data received from SSI regarding neutralization of a mink-derived SARS-CoV-2 isolate

carrying 4 simultaneous mutations in the spike protein by convalescent sera from patients early in the

Danish epidemic is evaluated, and risks for vaccine efficacy posed by such mink-derived virus strains

assessed.

Summary:

Danish mink-derived SARS-CoV-2 isolates have been reported to carry a number of mutations in the

spike protein, with some isolates having 4 simultaneous spike protein mutations.

Some of these mutations have been described as being mink-specific, i.e., as having originated in

mink.

However, data from GISAID (https://www.gisaid.org/) , the COVID-19 viral genome analysis pipeline at

Los Alamos, as well as scientific publications shows that viruses currently circulating in humans already

carry spike protein mutations which are identical or very similar to the spike protein mutations described

in Danish mink-derived viruses, albeit at very low frequency (<0.3% of circulating viruses in humans).

Therefore, the spike protein mutations observed in Danish mink-derived viruses should not be termed

mink-specific.

The higher frequency of certain spike protein mutations in mink-derived viruses than in human-derived

viruses is most likely caused by the virus' adaptation to replication in mink. Thus, the fitness of the mink-

derived viruses in humans is most likely reduced.

Only 2 of the mutations in the mink-derived viruses occur at spike protein sites targeted by neutralizing

antibodies:



The receptor-binding site of the spike protein harboring the Y453F mutant is the most critical of

these, as it is able to induce particularly potent neutralizing antibody responses.

However, the Y453F mutation is conservative, and not expected to have major impact

on receptor or antibody binding.

In agreement with this, the Y453F spike mutation has been shown to have mild positive

effect on binding of the spike protein to the human ACE2 receptor (1,8-fold increase in

binding, likely within experimental error).



The N-terminal domain of the spike protein is known also to be the target for neutralizing anti-

bodies, and the deletions of amino acids 69 and 70 in mink-derived viruses may impact antibody

binding.

However, residues 69 and 70 in human-derived viruses already carry radical mutations.

In short, based on initial, theoretical/bioinformatical evaluation of the mutations in the spike protein of

Danish mink-derived viruses as outlined above, the following can be concluded:



These mutations do not seem highly likely to raise new issues for vaccine development which

do not already exist, due to the well-known and relatively well understood currently ongoing

mutation and evolution of SARS-CoV-2 in humans.



The spike protein mutations are evaluated as not being likely to have substantial impact the

efficacy of first-generation vaccines.

C:\Users\depesl\AppData\Local\Microsoft\Windows\INetCache\Content.Outlook\SJ3CMJMF\LMST - 2020.11.10 Risk

assesment Mink strains (003).DOCX

L 77 - 2020-21 - Endeligt svar på spørgsmål 186: Spm. om der er andre sundhedsfaglige kilder, der har udtrykt samme bekymring og slået tvivl om en fremtidig vaccines effekt, til sundheds- og ældreministeren

In a preliminary experiment, SSI compared the neutralizing activity of 9 human convalescent sera

against a mink-derived SARS-CoV-2 isolate containing 4 simultaneous changes in the spike protein

(Δ69-70,

Y453F, I692V and M1229I) with the neutralizing activity of the same sera against a control

SARS-CoV-2 strain without these spike mutations, using a monkey kidney cell line (Vero cells).

On average, these 9 human sera exhibited 2,6-fold reduction in neutralizing activity against the mink-

derived SARS-CoV-2 strain containing 4 simultaneous changes in the spike protein, compared to the

control virus without these changes. As a rule of thumb, in this type of assay, a 2.6-fold difference would

typically be considered to have, at best, low biological significance due to the analytical method varia-

bility/inaccuracy, and certainly require a large number of replicates and additional controls to document

with acceptable confidence.

Thus, while the serum neutralization data is preliminary and requires follow up (see detailing of data

strength/weakness and recommendation for follow-up work in main text), the SSI data nevertheless

suggests that the mutations observed in the spike protein of mink-derived viruses are unlikely to mark-

edly impact vaccine efficacy. This DKMA conclusion based on the SSI neutralization data is in agree-

ment with the theoretical/bioinformatic evaluation of the mutations in the spike protein of Danish mink-

derived viruses outlined above.

In short, for the reasons above, and with reservations for the preliminary nature of the available data,

the following can be concluded :



The spike protein mutations seen in Danish mink-derived SARS-CoV-2 isolates are evaluated

as not being likely to have substantial impact the efficacy of first-generation vaccines and mon-

oclonal antibody therapeutics.



Specifically, at the population level, the likelihood of spike protein mutations seen in mink-de-

rived SARS-CoV-2 isolates causing vaccine failure is considered to be very low (with the reser-

vation for the preliminary nature of currently available data mentioned below).



However, it cannot be ruled out that vaccine failure may occur in individual vaccinees infected

with such isolates, particularly vaccinees mounting low neutralizing antibody responses to start

with.



This is a preliminary risk assessment, based on available data; further laboratory analyses and

scientific studies are needed to better understand the impact of the changes in the spike protein

of mink-derived viruses for the transmissibility and neutralization of such viruses in human hosts.

Mechanisms behind and biological consequences of mutations

in spike protein of mink-derived viruses:

Danish mink-derived SARS-CoV-2 isolates have been reported to contain a number of mutations in

the spike protein, with some isolates containing 4 simultaneous mutations (Δ69-70,

Y453F, I692V,

M1229I; ref 16).

Some of these mutations have been described as being mink-specific, i.e., as having originated in

mink (for example Y453F; ref 13 and 16).

However, data from GISAID, the COVID-19 viral genome analysis pipeline at Los Alamos as well as

scientific publications (14, 15) shows that the Y453F and M1229I were and are already present in

SARS-CoV-2

viruses circulating in humans, and as regards the Δ69-70

and I692V, viruses already cir-

culating in humans have comparable changes (e.g. I692F mutation and a variety of radical mutations

at spike positions H69 and V70 in human-derived viruses)(table 1). It should also be mentioned that

combinations of 2 mutations have already occurred in the spike protein of human-derived viruses, of-

ten including the D614G mutation which confers higher transmissibility between humans (8).



By human-derived viruses is meant SARS-CoV-2 isolates from cases of human-human virus

transmission, unrelated to the mink outbreaks in the Netherlands and Denmark in march and

august 2020, respectively (6, 13, 16, 18).



it is recognized that such so-called human-derived viruses very recently crossed into humans

from an animal reservoir, likely bats.

Mink were likely infected from humans, and represent naïve and susceptible hosts, with virus spread

between mink farms having been rapid. For all these reasons, the mutations in the spike protein of

mink-derived viruses are unlikely to represent the result of selective pressure from the immune re-

sponses in mink.

More likely, for the reasons above, the mutations seen in the spike protein of mink-derived viruses

were caused by the following 2 non-exclusive mechanisms:



Founder effects: Mutations in the spike protein of human viruses were carried over into mink.

This is in fact supported by the available data: The D614G mutation which confers

higher transmissibility of virus between humans arose in human-derived virus in China

and perhaps Germany around February 2020, and within 3 months outcompeted the

original D614 became the dominant virus form globally. The global spread of the

D614G mutant through the period February-May 2020 coincided temporally with the

disease outbreaks in mink in the Netherlands and Denmark (March and August 2020,

respectively; ref 6, 13, 18). This appears to be mirrored in an increased frequency of

the D614G mutant in mink-derived virus in Denmark compared to the Netherlands (ta-

ble 1).



Adaption to a new host: Mink as well as ferrets are known to be susceptible to SARS-CoV-2

infection, but less so than humans (1, 6, 7). Thus, it is expected that some adaptive changes

may occur in the spike protein and other viral proteins after transmission from humans to

mink.

In human-derived viruses, most mutations in the spike protein, particularly in the re-

ceptor-binding part of the protein, reduce the binding to human ACE2 receptor and/or

reduce spike protein stability, i.e., have negative effects on virus fitness and transmis-

sibility between humans (8, 9). Further, multiple mutations in spike have a more se-

vere negative effect on ACE2 binding and/or spike stability than single mutations (9).

Finally, some mutations in the spike protein increase the binding to the human ACE2

receptor, and viruses carrying these mutations are already circulating between hu-

mans; yet, the frequency of such ACE2-binding-enhancing mutations in naturally cir-

culating viruses is very low (<0.3%) and has remained stable over time, most likely

because the SARS-CoV-2 virus already has the biologically maximally desirable bind-

ing to human ACE2 receptor, acquired already or soon after the species jump to hu-

mans, i.e. at the start of the pandemic (8, 9).

This is not surprising, given that the binding affinity between the spike protein of

SARS-CoV-2 and human ACE2 is in the low pico-molar range (already very strong

binding to human ACE2, 10-fold stronger than for SARS-Cov-1; ref 9), and in experi-

ments intentionally selecting S protein mutants for increased binding to human ACE2,

this already high binding affinity could be only minimally improved (< 10-fold improve-

ments in binding affinity; ref 9).

Thus, for the reasons above, adaptive changes are expected to increase virus fitness

in mink, but are likely to be detrimental to virus fitness in humans.

In short:



It does not appear that SARS-CoV-2 spike protein has acquired unique mutations in mink.



For this reason, the spike protein mutations observed in mink-derived viruses should not be

termed mink-specific.



Thus, the spike protein mutations observed in mink-derived viruses do not appear to raise

new issues for vaccine and monoclonal antibody development which do not already exist due

to the well-known and relatively well understood currently ongoing mutation and evolution of

SARS-CoV-2 in humans (see for example refs 3, 8, 9, 10, 11, 14, 15, 17).

Only 2 of the mutations in the mink-derived viruses occur at spike protein sites targeted by neutralizing

antibodies (table 1):



The receptor-binding site of the spike protein harboring the Y453F mutant is the most critical

of these, as it is able to induce particularly potent neutralizing antibody responses (9).



However, the Y453F mutation is conservative, and not expected to have major impact

on receptor or antibody binding.

In agreement with this, the Y453F spike mutation having been shown to have mild

positive effect on binding of the spike protein to the human ACE2 receptor (1,8-fold

increase in binding, likely within experimental error; ref 9).

The N-terminal domain of the spike protein is known to also be the target for neutralizing anti-

bodies, and the deletions of amino acids 69 and 70 in mink-derived viruses may impact anti-

body binding.

However, residues 69 and 70 in human-derived viruses already carry radical muta-

tions (table 1).

In short:



The several mutations observed in the spike protein mink-derived viruses are already present

in human-derived viruses circulating between humans.



For this reason, and the reasons detailed in the text above, the mutations observed in the

spike protein of mink-derived viruses are evaluated as not being likely to have substantial im-

pact the efficacy of first-generation vaccines and monoclonal antibody therapeutics.



This assessment is based on available data, which is limited.



Thus, unquestionably, studies are warranted and required to better understand the properties

and pathogenicity of mink derived viruses, as well as the implications of the changes present

in the spike protein of mink-derived viruses for SARS-CoV-2 serodiagnostics, monoclonal anti-

body therapeutics and vaccines in development.

Table 1: Summary of aminoacid mutations in spike protein of mink-derived SARS-CoV-2 in

Denmark and the Netherlands, based on publicly available data

Mutation in

Frequency Location of mutation

Spike mutation Comments

spike protein

of mutation in spike protein

already present

of mink-de-

in human-de-

rived virus,

rived viruses ?

compared to

human-derived

virus

Δ69-70

(dele-

* Not known N-terminal domain.

H69D, H69Q,

N-terminal domain is sur-

tion of aminoac-

H69R, H69Y,

face-exposed, and

ids 69 and 70)

V70A, V70F and known to be target for

V70I mutations neutralizing antibodies in

present in hu-

humans.

man-derived vi-

ruses (rare,

<0.3%).

Y453F

5/13 NL se- Receptor-binding do-

Yes (rare,

Receptor-binding domain

quences.

main.

<0.3%).

is surface-exposed. Most

8/12 DK se-

of the known highly po-

quences.

tent neutralizing antibod-

ies target the receptor-

binding domain.

The Y453F mutation is

conservative, and not ex-

pected to have major im-

pact on receptor or anti-

body binding.

This is in agreement with

the Y453F spike muta-

tion having been shown

Mutation in

Frequency Location of mutation

spike protein

of mutation in spike protein

of mink-de-

rived virus,

compared to

human-derived

virus

Spike mutation Comments

already present

in human-de-

rived viruses ?

D614G

7/13 NL se- D-domain of S1 subunit

quences.

(carboxy-terminal part of

12/12 DK

S1 subunit)

sequences.

Yes (currently

dominant in hu-

man epidemic,

globally).

I692V

* Not known Seven aminoacids

downstream from furin

cleavage site between

S1 and S2 subunits (i.e.

in N-terminus of S2 sub-

unit).

I692 F mutation

in human-de-

rived viruses

(rare, <0.3%).

to have mild positive ef-

fect on binding of the

spike protein to the hu-

man ACE2 receptor (1,8-

fold increase in binding,

likely within experimental

error; ref 9).

Mutant considered to be

associated with higher

transmissibility between

humans, and higher sus-

ceptibility to neutraliza-

tion by antibodies.

The apparent increase in

frequency of the mutant

in Danish samples mir-

rors its increased domi-

nance in human-derived

viruses through the pe-

riod when the mink out-

breaks occurred in the

Netherlands and Den-

mark (march through au-

gust 2020; see ref. 17

and references therein).

Spike protein location

not surface-exposed, un-

likely to be target for

neutralizing antibodies.

Mink mutation is more

conservative than mutant

already present in hu-

mans.

Spike protein location

not surface-exposed ,

unlikely to be relevant for

antibody neutralization of

virus.

M1229I

* Not known Transmembrane part

Yes (rare,

<0.3%).

In addition,

M1229L and

M1229/ muta-

tions are also

seen in humans

(also rare,

<0.3%).

The data in the table is based on publicly available spike protein sequences and information from SSI

(16).

Information regarding whether spike protein mutations in mink-derived viruses are already present in

human-derived viruses is from GISAID, the COVID-19 viral genome analysis pipeline at Los Alamos,

and scientific publications (14, 15).

Mutations are reported in single-letter aminoacid code, using the following convention: Residue in

SARS-CoV-2, Wuhan-Hu-1, followed by residue number in the spike protein, followed by residue in

mutates spike protein.

Accession numbers:



Reference spike protein sequence, human-derived SARS-CoV-2, Wuhan-Hu-1 (first whole-genome SARS-CoV-2

sequence): Accession

number YP 009724390.

Mink-derived spike protein sequences from the Netherlands: Accession numbers QJS39496, QJS39507, QJS39519, QJS39531,

QJS39543, QJS39555, QJS39567, QJS39579, QJS39591, QJS39603, QJS39615, QJS39627, and QJF11995 (13 in total).

Mink-derived spike protein sequences from Denmark: QNJ45106, QNJ45118, QNJ45130, QNJ45142, QNJ45154, QNJ45166,

QNJ45178, QNJ45190, QNJ45202, QNJ45214, QNJ45226 and QNJ45238 (12 in total).

* These mutations do not appear in the publicly available sequences.

Initial experimental work done at SSI to evaluate the implica-

tions of spike protein mutations in mink-derived viruses for

countermeasure development:

The data discussed in the following is from the first serum neutralization experiment of mink-derived

virus with human convalescent sera, carried out by Dr Anders Fomsgaard and Maria Magdalena Las-

sauniére (SSI report dated 02.11.2020; ref 16).

The neutralizing activity of human convalescent sera against a mink-derived SARS-CoV-2 isolate con-

taining 4 simultaneous changes in the spike protein (Δ69-70,

Y453F, I692V and M1229I; see table 1)

was compared to the neutralizing activity of the sera against a control SARS-CoV-2 strain without

these spike mutations. Serial dilutions of convalescent sera were pre-incubated with fixed amounts of

mink-derived or control virus for 1h, followed by inoculation on Vero cells (African green monkey kid-

ney cell line) and incubation for 24h. Neutralization titers were defined as serum dilutions causing 50%

reduction in the levels of nucleocapsid protein production (NT

values), determined by immunostain-

ing of fixed cells (exact experimental and analysis details lacking).

A total of 9 human convalescent sera were tested, all from patients early in the Danish epidemic.

In the following, all interpretations of the SSI data represent evaluation done at the DKMA.

The NT

titers of the 9 human convalescent sera against the control SARS-CoV-2 strain ranged from

39 to >1280 (table 2). Four of these sera had neutralization titers < 80, considered to be low titers, and

for the group as a whole, the geometric mean titer was 168 (table 1). This wide range of neutralization

titers as well as relatively high frequency of low-end NT50 values has been reported by others for SARS-

CoV-2 convalescent sera, and was as such not surprising (2, 4, 5). However, with the reservation that

neutralization assays are not standardized and neutralization titer values difficult to compare between

datasets, these patient sera appeared to exhibit lower neutralization titers than expected in vaccinees.

For the mink-derived strain, neutralization titers ranged from 5 to >1280, with a geometric mean of 64

(table 2).

Thus, on average, these 9 human sera exhibited a 168/64=2.6-fold reduction in neutralizing activity

against the mink-derived SARS-CoV-2 strain containing 4 simultaneous changes in the spike protein,

compared to the control virus without these changes.

Neutralizing titers against control virus exhibited a good linear correlation with neutralization titers

against mink-derived virus (figure 1, R2>0,9), with inclinations of the regression lines were 0,74 to 0,97

(figure 2).

Based on the regression equations shown in figure 2, a neutralization titer of for example 160 against

human-derived virus corresponds to a neutralization titer of 0.75 x 160 - 41 = 79 (figure 2 panel A) to

0.98 x 160 - 76 = 80 (figure 2 panel B) against mink-derived virus.

Thus, based on regression analysis, human sera exhibited 160/80 = 2-fold to 160/79 = 2-fold reduction

in neutralizing activity against the mink-derived SARS-CoV-2 strain containing 4 simultaneous changes

in the spike protein, compared to the control virus without these changes, in agreement with the results

based on simple geometric group means detailed above.

Neutralization assays are known to exhibit a relatively high level of variability, and as a simplified, gen-

eral rule of thumb, differences in neutralization titers of 10-fold or more are considered to be high and

likely biologically significant (see for example ref. 3 and 8), whereas the biological significance of neu-

tralization titer differences of less than 10-fold are less likely to have biological significance, and differ-

ences of 2 fold or less would typically be considered to have at best low biological significance, and

certainly require a large number of replicates and additional controls to document with acceptable con-

fidence (8).

Thus, the observed reduction in mean neutralization activity of human convalescent sera against mink-

derived virus (average 2,6-fold reduction for group of 9 convalescent sera) is considered to be of low

significance, given the known variability of neutralization assays, and preliminary nature of the data

(see section on data strength and recommendations for follow-up experiments below).

Sera which were low-titered against the control virus exhibited the largest reductions in titer against the

mink-derived virus, and this correlation was significant (figure 1; P=0,0182). Thus, the potential risk of

failure of COVID-19 vaccines against mink-derived cluster 5 strains is likely highest in vaccinees exhib-

iting low vaccine responses to start with.

At the same time, of the 4 sera with titers <80 (sera 1 through 4, table 2), neutralization of mink virus

varied from 7% to 72% of the neutralization of control virus (table 2), and in this (admittedly very small)

sub-group of sera with low neutralization titers, the correlation between neutralization of human and

mink virus was less clear (figures 1 and 2; not detailed). This is not surprising, because the distribution

of circulating polyclonal neutralizing antibodies in patients against different parts of the S protein (for

example balance between amounts of circulating polyclonal neutralizing antibodies against N-terminal

domain versus receptor-binding domain) is known to be idiosyncratic amongst patients (3, 4, 5), and

also, affinities of neutralizing polyclonal anti-S protein antibodies differ between patients (4); these idio-

cyncrasies in circulating polyclonal neutralizing antibody responses are likely due to factors such as

subtle differences in how the spike protein was presented to patient B-cells, age of patients, MHC-II

haplotype of patients, etc.

In short:



The serum neutralization data is preliminary and has several shortcomings requiring follow up

(see detailing and recommendation for follow-up work above).



Nevertheless, the data suggests that the mutations observed in the spike protein of mink-de-

rived viruses are unlikely to impact vaccine efficacy.



This conclusion based on evaluation of SSI data is in agreement with the conclusion drawn

based on bioinformatic analysis and theoretical considerations in the section above, titled

"Mechanisms behind and biological consequences of mutations in spike protein of mink-derived

viruses".

Table 2: Raw serum neutralization data.

Patient

Neutralization titer (NT

)

Control virus iso- Mink-derived cluster 5 virus

late

isolate

186

>1280

846

606

186

291

177

* Descriptive statistics, all 9 sera:

mean:

335

median:

186

geometric

168

mean:

Titer ratio

mink virus/control virus

0,72

0,07

0,17

0,45

0,36

1,00

0,72

0,29

0,61

251

0,49

0,45

0,38

Descriptive statistics, excl. patient 6:

mean:

217

122

median:

127

0,42

0,40

geometric

142

0,36

mean:

The data is from the first neutralization experiment, carried out by Dr Anders Fomsgaard and Maria

Magdalena Lassauniére (SSI report dated 02.11.2020; ref 16).

* The patient 6 serum with a titer of >1280 was assigned the value 1280.

The geometric mean is usually used in evaluating results from serological assays such as this; for con-

venience, mean, median and geometric mean values are all shown.

Figure 1: Relationship between magnitude of anti-SARS-CoV-2 antibody response in patients

against homologous (human-derived) virus, and failure to neutralize mink-derived virus

The inclination of the regression line was significantly different from 0 (P=0,182; GraphPad QuickCalc software,

available online https://www.graphpad.com/quickcalcs/linear2/)

Figure 2: Ability of convalescent human sera to neutralize mink-derived SARS-CoV-2 cluster 5

isolate with 4 simultaneous changes in spike (Δ69-70,

Y453F, I692V, M1229I)

Panel A, serum with a titer of >1280 is excluded.

Panel B, serum with a titer value of >1280 is included, and assigned a titer value of 1280.

Strength of data from initial studies, and recommendations for

follow-up experiments:

All interpretations of the SSI data below represent evaluation done at the DKMA.

The data has the following shortcomings:



The cell culture system used for neutralization assays is permissive for SARS-CoV-2 replication

(6), but is non-human, and represents a cell type which is not a primary target for infection in

humans (Vero cells; African green monkey kidney cell line). The ACE2 receptor expressed by

Vero cells has been reported to differ by approx. 5% from human ACE2 protein at the amino

acid level (1), and while Vero cells are often used for in vitro studies of SARS-CoV-2 virus,

typically human cell lines are also included (6), and results obtained using Vero cells and human

cells may differ (6).



For neutralization results to be valid, identical amounts of mink-derived and control virus should

be used, and these 2 viruses should exhibit identical or very comparable replication kinetics in

Vero cells. It is stated in the SSI report that same amounts of mink-derived and control virus



were used, but it is not known which methods were used to quantitate these 2 virus isolates.

Comparison of replication kinetics in Vero cells for the 2 viruses is not provided.

With the reservation that neutralization assays are not standardized and neutralization titer val-

ues are difficult to compare between datasets, the patient serum panel appeared to be low

titered compared to expected neutralization titers in vaccinees.

The serum neutralization data represents a single experiment, the number of replicate determi-

nations is not reported, and the overall variability of the neutralization assay is not known; such

assays are known to be quite variable.

The technical performance of the neutralization assay is not known; specifically, the cutoff level

between positive and negative sera, and the linearity (dose-response) of the assay are of inter-

est.

For the reasons above, the strength of the data is evaluated as low (preliminary results), as is also

acknowledged in the SSI report.

For initial confirmatory follow-up experiments, the following is recommended:



Clinical data should be provided for patient sera (e.g. patient age, and time interval since infec-

tion).



Sera which are more representative of vaccinees should be included.



More virus strains should be included.



More technical information should be provided for the variability of the neutralization assay,

controls used, and the performance of the assay (see above).



The virus neutralization/serology work should be expanded to include determination of the af-

finity of the mutated spike protein in mink-derived cluster 5 strain(s) for human ACE2 receptor.

It should also be mentioned here that even the relatively modest and focused follow-up work outlined

above represents a relatively large effort.

Reference list:

(1) Olivieri ER, Heller LK, Gillim-Ross L, Wentworth DE. Analysis of SARS-CoV receptor activity of

ACE2 orthologs. Adv Exp Med Biol. 2006;581:277-280. doi:10.1007/978-0-387-33012-9_46

(2) Robbiani DF et al. Convergent antibody responses to SARS-CoV-2 in convalescent individuals.

Nature. 2020 Aug;584(7821):437-442. doi: 10.1038/s41586-020-2456-9. Epub 2020 Jun 18. PMID:

32555388; PMCID: PMC7442695.

(3) Weisblum Y et al. Escape from neutralizing antibodies by SARS-CoV-2 spike protein variants. bio-

Rxiv [Preprint]. 2020 Jul 22:2020.07.21.214759. doi: 10.1101/2020.07.21.214759. Update in: Elife.

2020 Oct 28;9: PMID: 32743579; PMCID: PMC7386497.

(4) Barnes CO et al. Structures of Human Antibodies Bound to SARS-CoV-2 Spike Reveal Common

Epitopes and Recurrent Features of Antibodies. Cell. 2020 Aug 20;182(4):828-842.e16. doi:

10.1016/j.cell.2020.06.025. Epub 2020 Jun 24. PMID: 32645326; PMCID: PMC7311918.

(5) Zhang BZ et al. Mining of epitopes on spike protein of SARS-CoV-2 from COVID-19 patients. Cell

Res. 2020 Aug;30(8):702-704. doi: 10.1038/s41422-020-0366-x. Epub 2020 Jul 1. PMID: 32612199;

PMCID: PMC7327194.

(6) Molenaar RJ et al. Clinical and Pathological Findings in SARS-CoV-2 Disease Outbreaks in

Farmed Mink (Neovison vison). Vet Pathol. 2020 Sep;57(5):653-657. doi:

10.1177/0300985820943535. Epub 2020 Jul 14. PMID: 32663073.

(7) Heller LK, Gillim-Ross L, Olivieri ER, Wentworth DE. Mustela vison ACE2 functions as a receptor

for SARS-coronavirus. Adv Exp Med Biol. 2006;581:507-10. doi: 10.1007/978-0-387-33012-9_90.

PMID: 17037586; PMCID: PMC7123197.

(8) Li Q et al. The Impact of Mutations in SARS-CoV-2 Spike on Viral Infectivity and Antigenicity. Cell.

2020 Sep 3;182(5):1284-1294.e9. doi: 10.1016/j.cell.2020.07.012. Epub 2020 Jul 17. PMID:

32730807; PMCID: PMC7366990.

(9) Starr TN et al. Deep Mutational Scanning of SARS-CoV-2 Receptor Binding Domain Reveals Con-

straints on Folding and ACE2 Binding. Cell. 2020;182(5):1295-1310.e20.

doi:10.1016/j.cell.2020.08.012

(10) GISAID (Global Initiative on Sharing Avian Influenza Data), https://www.gisaid.org/

(11) COVID-19 viral genome analysis pipeline (Los Alamos), https://cov.lanl.gov/content/index

(12) Mutations in spike putatively linked to outbreak at Danish mink farms

GISAID, accessed 08.11.2020, https://www.gisaid.org/references/gisaid-in-the-news/mutations-in-

spike-putatively-linked-to-outbreak-in-danish-mink-farms/

(13) Besvarelse vedr. Epidemiologisk udredning i SARS-COV-2 smittede minkfarme. Dansk Veterinær

Konsortium, journalnummer FVST 2020-14-81-01705; journalnummer KU 061-0115/20-3680, journal-

nummer SSI 20/06443

(14) Baum A et al. Antibody cocktail to SARS-CoV-2 spike protein prevents rapid mutational escape

seen with individual antibodies. Science. 2020 Aug 21;369(6506):1014-1018. doi: 10.1126/sci-

ence.abd0831. Epub 2020 Jun 15. PMID: 32540904; PMCID: PMC7299283.

(15) Dearlove B et al. A SARS-CoV-2 vaccine candidate would likely match all currently circulating

variants. Proc Natl Acad Sci U S A. 2020 Sep 22;117(38):23652-23662. doi:

10.1073/pnas.2008281117. Epub 2020 Aug 31. PMID: 32868447; PMCID: PMC7519301

(16) Assessing cross-neutralization activity for SARS-CoV-2 spike mutants. SSI 02. November 2020,

Anders Fomsgaard, professor & Maria Magdalena Lassauniére, senior researcher

(17) MAKING SENSE OF CORONAVIRUS MUTATIONS. Commentary, Ewen Callaway. 174 | Nature

| Vol 585 | 10 September 2020 | Corrected 16 September

(18) Oreshkova N, Molenaar RJ, Vreman S, et al. SARS-CoV-2 infection in farmed minks, the Nether-

lands, April and May 2020. Euro Surveill. 2020;25(23):2001005. doi:10.2807/1560-

7917.ES.2020.25.23.2001005