This is a sel -a chi ed e sion o an o iginal a icle. This e sion
may di e om he o iginal in pagina ion and ypog aphic de ails.
Au ho (s):
Ti le:
Yea :
Ve sion:
Copy igh :
Righ s:
Righ s u l:
Please ci e he o iginal e sion:
CC BY 4.0
h ps://c ea i ecommons.o g/licenses/by/4.0/
Sys ema ic Compa ison o Epidemic and Non-Epidemic Ca bapenem Resis an Klebsiella
pneumoniae S ains
© 2021 he Au ho s
Published e sion
Koskinen, Ka a iina; Pen inen, Ree a; Ö mälä-Odeg ip, Anni-Ma ia; Giske,
Ch is ian G.; Ke ola, Ta mo; Jalas uo i, Ma i
Koskinen, K., Pen inen, R., Ö mälä-Odeg ip, A.-M., Giske, C. G., Ke ola, T., & Jalas uo i, M.
(2021). Sys ema ic Compa ison o Epidemic and Non-Epidemic Ca bapenem Resis an Klebsiella
pneumoniae S ains. F on ie s in Cellula and In ec ion Mic obiology, 11, A icle 599924.
h ps://doi.o g/10.3389/ cimb.2021.599924
2021
Sys ema ic Compa ison o Epidemic
and Non-Epidemic Ca bapenem
Resis an Klebsiella pneumoniae
S ains
Ka a iina Koskinen
1
, Ree a Pen inen
2
, Anni-Ma ia Ö mälä-Odeg ip
3
,
Ch is ian G. Giske
3,4
, Ta mo Ke ola
1
and Ma i Jalas uo i
1
*
1
Depa men o Biological and En i onmen al Science, Nanoscience Cen e , Uni e si y o Jy a
skyla
, Jy a
skyla
, Finland,
2
Depa men o Biology, Uni e si y o Tu ku, Tu ku, Finland,
3
Di ision o Clinical Mic obiology, Depa men o Labo a o y
Medicine, Ka olinska Ins i u e , S ockholm, Sweden,
4
Depa men o Clinical Mic obiology, Ka olinska Uni e si y Hospi al,
S ockholm, Sweden
O e he pas ew decades, ex ensi ely d ug esis an (XDR) esis an Klebsiella
pneumoniae has become a no able bu den o heal hca e all o e he wo ld. Especially
ca bapenemase-p oducing s ains a e p oblema ic due o hei capabili y o wi hs and
e en las eso an ibio ics. Some sequence ypes (STs) o K. pneumoniae a e significan ly
mo e p e alen in hospi al se ings in compa ison o o he equally esis an s ains. This
p o okes he ques ion whe he o no he e a e pheno ypic cha ac e is ics ha may
ende ce ain K. pneumoniae mo e sui able o epidemic dispe sal be ween pa ien s,
hospi als, and di e en en i onmen s. In his s udy, we selec ed se en epidemic and non-
epidemic ca bapenem esis an K. pneumoniae isola es o ex ensi e sys ema ic
cha ac e iza ion o pheno ypic and geno ypic quali ies in o de o iden i y po en ial
ac o s ha p ecede o eme ge om epidemic success ulness. S udied cha ac e is ics
include g ow h a es and densi ies in di e en condi ions (media, empe a u e, pH,
esou ce le els), ole ance o alcohol and d ough , inhibi ion be ween s ains, abili y o
compensa e pH, as well as a ious genomic ea u es. O e all, he e a e clea di e ences
be ween isola es, ye , only d ough ole ance was ound o no ably associa e wi h non-
epidemic K. pneumoniae s ains. We u he epo a p elimina y s udy on he po en ial o
con ol K. pneumoniae ST11 wi h an an imic obial componen p oduced by a non-
epidemic K. pneumoniae. This componen ini ially es ic s bac e ial g ow h, bu s able
esis ance de elops apidly in i o.
Keywo ds: XDR Klebsiella pneumoniae, ex ended-spec um be a-lac amase, epidemic, an ibio ic
esis ance, i ulence
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999241
Edi ed by:
Rodnei Dennis Rossoni,
Sao Paulo S a e Uni e si y,
B azil
Re iewed by:
Ta iana Amabile De Campos,
Uni e si y o B asilia, B azil
Jona as Ra ael De Oli ei a,
Anhembi Mo umbi Uni e si y, B azil
*Co espondence:
Ma i Jalas uo i
ma i.jalas uo i@jyu.fi
Special y sec ion:
This a icle was submi ed o
Bac e ia and Hos ,
a sec ion o he jou nal
F on ie s in Cellula
and In ec ion Mic obiology
Recei ed: 28 Augus 2020
Accep ed: 08 Janua y 2021
Published: 23 Feb ua y 2021
Ci a ion:
Koskinen K, Pen inen R,
Ö mälä-Odeg ip A-M,
Giske CG, Ke ola T and Jalas uo i M
(2021) Sys ema ic Compa ison o
Epidemic and Non-Epidemic
Ca bapenem Resis an Klebsiella
pneumoniae S ains.
F on . Cell. In ec . Mic obiol. 11:599924.
doi: 10.3389/ cimb.2021.599924
ORIGINAL RESEARCH
published: 23 Feb ua y 2021
doi: 10.3389/ cimb.2021.599924
INTRODUCTION
Klebsiella pneumoniae is a G am-nega i e bacillus causing
oppo unis ic in ec ions ou side o he gas oin es inal ac
(Podschun and Ullman, 1998). Common condi ions include
pneumonia, u ina y ac in ec ions, wound in ec ions, and less
o en li e abscess, meningi is, and sep icemia. Some o he s ains
ci cula ing in clinical se ings a e also showing inc easingly
i ulen pheno ypes (Pomako a e al., 2012;Shon e al., 2013).
These s ains a e o en cha ac e ized by hype muco iscosi y when
cul i a ed on aga pla es and hey a e mo e esilien agains killing
by se um o phagocy osis (Ca alan-Naje a e al., 2017). Mo eo e ,
ex ensi ely d ug- esis an (XDR) among K. pneumoniae is
inc easing e y apidly compa ed o many o he p io i y
pa hogens (Wo ld Heal h O ganiza ion, 2014). In pa icula ,
in ec ions caused by K. pneumoniae s ains which ha e
de eloped esis ance agains newe gene a ions o b-lac ams,
such as ca bapenems, can be haza dous and o en li e-
h ea ening. These ca bapenemase genes a e ound o be
abundan in bac e ia o igina ing om hospi al en i onmen s
al hough he e a e no able egional di e ences. Ye , sequence
yping o he pa hogens indica e ha esis an K. pneumoniae
s ains can also dissemina e globally be ween hospi als.
WHO has classified ca bapenemase-p oducing K. pneumoniae
as an u gen h ea (Wo ld Heal h O ganiza ion, 2017). O en, K.
pneumoniae isola es a e yped by u ilizing pa ial sequences om
se en housekeeping genes. These genes a e pa o he co e
genome and hence unlikely o be ho izon ally ans e ed
be ween di e en K. pneumoniae s ains. As such, sequence
yping p o ides a udimen a y app oach o iden i y gene ic
simila i y among isola es o di e en sou ces o o igin. I
appea s ha ce ain sequence ypes (STs) ha e been mo e
success ul in dispe sing be ween hospi als compa ed o o he
equally esis an s ains. NDM-1 me allo-b-lac amase p oducing
ST11 and 14 ha e been no ed o be esponsible o epidemics in
a ious coun ies (Yong e al., 2009;Pi ou e al., 2015;Samuelsen
e al., 2017). Klebsiella pneumoniae ca bapenemase (KPC)
p oducing ST258 ha e e en been e e ed as hype epidemic
clone (Bowe s e al., 2015) and i s epidemic po en ial has been
u he in es iga ed in se e al me a-analyses (Dau zenbe g e al.,
2016). ST512, a single-locus a ian o ST258, is also highly
associa ed in epidemics globally (Con e e al., 2016). ST147 is
also haza dous wi h nume ous i ulence and esis ance genes
(Tu on e al., 2018). Despi e o he no ion ha ce ain STs
appea o be mo e p one o in e -hospi al dispe sal, i is s ill
unclea wha quali ies alongside o pa hogenici y, i any, may be
esponsible o his epidemic success. Majo i y o he su eillance
a emp s ocus on esis ance p ofiles and geno ypic ea u es (Giske
e al., 2012). Genomic da a also accumula es apidly as whole
genome sequences o many s ains ha e become a ailable (Hol
e al., 2015). Ye , he pheno ypic cha ac e is ics o di e en ly
success ul K. pneumoniae STs a e a ely s udied in de ail, o he
pheno ypic analysis ocus on specific ai ssuchas
hype muco iscousi y (Ca alan-Naje a e al., 2017). Compa ison
o he pheno ypes o epidemic and non-epidemic s ains could
po en ially e eal meaning ul in e ac ions be ween bac e ia and
hei en i onmen ha con ibu e o he epidemic sp ead o
XDR s ains.
In his s udy, we selec ed 14 ca bapenem esis an K.
pneumoniae s ains isola ed om pa ien s hospi alized in USA,
Sweden, UK, G eece, o India (Ki chel e al., 2009;Samuelsen
e al., 2009;Ki chel e al., 2010;Samuelsen e al., 2011;Vading
e al., 2011;Giske e al., 2012;Hasan e al., 2014). Fi e o hese
STs a e con inuously being de ec ed in hospi als in mul iple
coun ies and can be conside ed as epidemiologically success ul
o epidemic (Table 1). The es o he STs ha e made only
seldom appea ances and ha e a ely i a all dispe sed o o he
hospi als and may hence be conside ed as non-epidemic STs.
He e, we sys ema ically de e mined and measu ed po en ially
ele an cha ac e is ics o hese s ains in o de o e eal
di e ences ha may co ela e and pe haps pa ly explain he
epidemic success.
MATERIALS AND METHODS
S ains
All 14 s udied Klebsiella pneumoniae s ains (see Table 1) we e
Illumina sequenced a Ka olinska Ins i u e , Sweden. Genome
sequences can be ound om GenBank unde BioP ojec id
PRJNA680903. Sequences o EKP24 and NKP2 we e also
PacBio sequenced in Uni e si y o Helsinki, Finland.
Sequenced genomes we e anno a ed by Rapid Anno a ion
using Subsys em Technology (RAST, h ps:// as .nmpd .o g)
and bo h secu e and po en ial p o ein coding genes we e
mapped and he dis ibu ion o p o ein amilies we e
compa ed be ween he s ains. Resis ance genes we e iden ified
wi h ResFinde (h ps://cge.cbs.d u.dk/se ices/ResFinde /),
p ophages wi h P ophage Finde (h ps://omic ools.com/
p ophage-finde - ool), and CRISPR- egions wi h CRISPR-
finde (h ps://c isp .i2bc.pa is-saclay. ). An algo i hm was
w i en o iden i y unanno a ed sho open eading ames
(ORFs) om he genome files (Supplemen a y File 1). The
algo i hm scans he genome o ORFs ha ha e a po en ial
ibosome binding si e ups eam o he s a codon and is no
o e lapping wi h anno a ed genes.
G ow h Expe imen s
G ow h densi ies o each s ain we e measu ed in empe a u e o
+37°C and oom empe a u e wi h 230 pm shaking o wi hou
shaking. Cells we e g own o e nigh in 5 ml o LB media (+37°C,
230 pm) and hen ans e ed in o 5 ml o esh LB media in
1:5,000 a io. F esh cul u es we e g own in expe imen al se ings
o +37°C and 230 pm o RT and 0 pm o 20 h and g ow h
densi ies we e calcula ed as colony o ming uni s (c u)/ml. As a
s anda d ini ial liquid cul u e o he g ow h cu e expe imen s
all he s ains we e cul u ed in 5 ml o 100% LB, +37°C, and 230
pm o e nigh and hen ans e ed in o expe imen al se ings.
In o de o es he e ec o shaking, he ini ia ing cul u es we e
p epa ed hen ans e ed in o 5 ml o LB in 1:100 a io and
g own in expe imen al se ings o +37°C and 230 pm and +37°C
Koskinen e al. Epidemic and Non-Epidemic Klebsiella pneumoniae
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999242
and 0 pm. In g ow h cu e expe imen s he e ec o di e en
media and a ying concen a ions and composi ions o nu ien s
on g ow h was de e mined o each s ain. One hund ed pe cen
LB was used h oughou he expe imen s unless men ioned
o he wise. G ow h cu es we e measu ed in 10 and 1% LB,
100% BHI, and 100% o pu e DMEM by dilu ing he ini ial
cul u e in 1:100. G ow h cu es we e measu ed a +37°C, 595 nm
wa eleng h wi h Mul iscan FC (The mo Scien ific) o 20 h in
5 min in e als and maximum g ow h and a e age g ow h a e
we e calcula ed.
Su i al in Acidic pH and
Compensa ion Capaci y
Bac e ial cells’abili y o ole a e acidic su ounding pH and
capabili y o compensa e i by me abolism was measu ed o each
s ain in pH 3–7. Ini ial cul u es we e g own in 5 ml LB pH 7 a
+37°C and 210 pm o e nigh . Each s ain was hen ans e ed
in o 5 ml LB o ei he pH 3, pH 4, pH 5, pH 6, o pH 7 in 1:100
a io and cul u ed in +37°C and 210 pm. In pH 5–7 cul u es
we e g own o 90 h and g ow h densi ies we e calcula ed by
pla ing in 16, 24, and 90 h. Then 1.5 ml o cul u e was fil e ed
h ough 0.2 mm and supe na an pH was measu ed wi h Basic
pH Me e (Den e Ins umen s) in 24 and 90 h. Cul u es in pH
3–4 we e sho ened in o 24 h expe imen and g ow h densi ies
we e calcula ed in 16 and 24 h and supe na an pH was measu ed
in 24 h. G ow h cu es we e measu ed by dilu ing he ini ial pH 7
cul u es in 1:100 a io in o LB pH o 3–7 and g ow h cu es we e
measu ed a +37°C, 595 nm wa eleng h o 20 h in
5 min in e als.
C oss-S ain In e ac ions
In aim o s udy he dynamics all he s ains we e cul u ed
sepa a ely, and hei me abolic p oduc s sec e ed in o
su oundings we e es ed agains o he s ains in c oss-s ain
inhibi ion expe imen s. C oss-s ain in e ac ions we e es ed by
collec ing he media a e o e nigh cul u ing a +37°C and 210
pm. O e nigh cul u es we e cen i uged fi s wi h 7,000 × g o
4 min, and he supe na an was cen i uged again wi h 10,000 × g
o 1 min. Each s ain was c oss-pla ed wi h all he supe na an s
(o e nigh , +37°C), and he inhibi ion o he g ow h o each
s ain was obse ed.
Supe na an Inhibi ion and P ophages
A 4-week e olu iona y expe imen was designed o s udy
appea ance, pe sis ence, and e e sibili y o pu a i e colicin E3
esis ance in sensi i e EKP24 s ain. Fo he fi s 2 weeks EKP24
was cul u ed wi h (n = 5) and wi hou (n = 5) colicin E3 in 10%
LB media supplemen ed wi h ei he NKP2 (con aining colicin
E3) o EKP24 (no con aining colicin E3) supe na an fil a e
(0.2 mm) in 1:4 a io. Ten pe cen LB media was supplemen ed
TABLE 1 | K. pneumoniae isola es used in he s udy.
Isola e ST om
da abase
Isola ion
loca ion
No. o
CRISPR
loci
No. o p o-
phage
egions
Re e ence Vi ulence
genes
Capsule ypes Be a-lac amase
NKP01 1534 37 USA 0 10 Ki chel
e al., 2009
m k K15K17K50K51K52 blaTEM-1B, blaKPC-2, blaSHV-11
NKP02 10924 334 USA 0 4 Ki chel
e al., 2009
m K N/A blaKPC-3, blaOKP-B-4, blaOXA-9,
blaTEM-1A
EKP03 70165 14 USA 3 10 Ki chel
e al., 2009
i p, yu,
yb , k u,
m k
K2 blaTEM-1A, blaSHV-28, blaKPC-3
EKP05 70708 258 USA 0 7 Ki chel
e al., 2009
m k N/A
(K15K17K50K51K52)
blaOXA-9, blaKPC-3, blaSHV-12
EKP08 2008025 11 USA 0 11 Ki chel
e al., 2009
m k K13 blaKPC-2, blaSHV-11
EKP10 AO-8053 512 Sweden
(Is ael*)
010Samuelsen
e al., 2009
m k N/A blaTEM-1A, blaOXA-9, blaSHV-11,
blaKPC-3
EKP11 AO-15200 147 Sweden
(G eece*)
26Samuelsen
e al., 2011
m k K64,K14 blaSHV-11, blaVIM-1
NKP18 VPKP389 36 A hens,
G eece
110Hasan
e al., 2014
i p, m k,
yu, yb
k27 blaSHV-129, blaVIM-26
NKP20 VPKP229 17 A hens,
G eece
08Hasan
e al., 2014
i p, m k,
yb
k25 blaSHV-129, blaVIM-1
EKP22 N6 14 UK 2 6 Giske e al.,
2012
m k, yb ,
k u
k2 blaCTX-M-15, blaSHV-11, blaTEM-1A,
blaOXA-1, blaNDM-1, blaOXA-9
EKP24 ED502873 11 Sweden 0 6 Giske e al.,
2012
i p, m k,
yu, yb
N/A
(K15K17K50K51K52)
blaSHV-11, blaCTX-M-15, blaNDM-1
NKP25 N12 231 UK 1 4 Giske e al.,
2012
k u, m k K51 blaSHV-1, blaTEM-1B, blaNDM-1,
blaOXA-1
NKP28 B357 43 UK 1 8 Giske e al.,
2012
m k, k u K30 blaCTX-M-15, blaDHA-1, blaCMY-6,
blaSHV-11, blaOXA-9, blaNDM-1,
blaTEM-1A
NKP30 IR34 624 Chennai,
India
13Giske e al.,
2012
m k K12, K29 blaTEM-1B, blaDHA-1, blaNDM-1,
blaSHV-36, blaCTX-M-15, blaOXA-1
*Isola es associa ed wi h impo om he coun y.
Koskinen e al. Epidemic and Non-Epidemic Klebsiella pneumoniae
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999243
wi h 25 mg/ml o kanamycin and 150 mg/ml o ampicillin. A e 2
weeks EKP24 cul u ed wi h he p esence o colicin E3 we e
di ided in o wo se s o samples (bo h n = 5). O he se was
con inued wi h colicin E3 exposu e as desc ibed ea lie . In he
o he se o samples, colicin E3 con aining supe na an was
eplaced wi h EKP24 supe na an . Cul u es we e e eshed in
1:100 a io h ee imes and samples s o ed once a week wi h
glyce ol a −80°C. De elopmen and pe sis ence o colicin E3
esis ance was de e mined by pla ing. Samples we e aken a he
beginning o he expe imen , be o e di ision o colicin E3
exposed EKP24, and a he end o he expe imen and we e
used o DNA ex ac ion. DNA was isola ed wi h DNeasy Blood
& Tissue Ki (Qiagen) and sequenced wi h Illumina HiSeq. The
obse ed eads we e mapped o o iginal PacBio-sequenced
genome o NKP2 (desc ibed abo e) in o de o de ec he
gene ic a ian s. The a ian s de eloped unde he exposu e o
colicin E3 we e iden ified by fil e ing ou hose a ian s ha we e
al eady p esen in he beginning o he expe imen . The gene ic
analysis was pe o med wi h CLC Genomics Wo kbench
11 (Qiagen).
In e ac ions be ween he pu a i e colicin E3 p oducing NKP2
and suscep ible EKP24 bac e ial cells we e also obse ed wi h
con ocal mic oscopy. Then 200 ml o 1% LB-aga was placed in o
a chambe o eigh -chambe ed ibidi®ibiT ea m-Slide (Ibidi
GmbH) co e ed wi h CID lid o µ-dishes (Ibidi GmbH), and
3ml o NKP2 and EKP24 we e injec ed unde he aga on o
opposi e sides. Encoun e o he wo s ains was isualized wi h
Nikon AR1 lase scanning con ocal mic oscope wi h 60× wa e
imme sion objec i e and using Gal ano scanne .
P oduc ion o pu a i e colicin E3 was u he s udied by
g owing NKP2 s ain in di e en medias. NKP2 was g own in
LB concen a ions o 100, 10, and 1%, LB wi hou yp one, 100%
Shieh (Song e al., 1988) and in 100% DMEM in +37°C and 200–
230 pm o e nigh . NKP2 cul u es we e fil e ed h ough 0.2 mm
fil e and colicin E3 p esence was de e mined by pla ing he
supe na an wi h suscep ible EKP24 s ain.
Alcohol Exposu e
All s udied s ains we e exposed o mul iple concen a ions o
e hanol (20, 50, 75, and 90%) and hei abili y o su i e he
exposu e we e measu ed wi h spec oscopy. In 30 s exposu e
expe imen all he s ains we e fi s g own in 5 ml o LB (+37°C,
230 pm, o e nigh ) and hen ans e ed in 1:10 a io in o esh
LB and le o g ow o e nigh a +37°C on 96-well pla e 100 ml
pe well o o m biofilm. On he ollowing day media was gen ly
emo ed and eplaced wi h 200 ml o e hanol in concen a ions o
20, 50, 75, and 90%. A e 30 s incuba ion in RT e hanol was
eplaced wi h 100 ml o LB and g ow h a +37°C was measu ed in
595 nm wa eleng h o 20 h in 5 min in e als. In e hanol
e apo a ion expe imen , cul u es o 96-well pla e we e p epa ed
as desc ibed ea lie bu olume was lowe ed in o 35 ml pe well.
Nine y-six-pla e cul u es we e incuba ed o e nigh a +37°C
wi hou he lid o le he excess media e apo a e. D ied
biofilms we e exposed o 50 ml o e hanol (ei he 20, 50, 75, o
90%) and le o ully e apo a e be o e addi ion o 200 mlo LB
pe well. G ow h cu e measu emen s was pe o med as
ea lie desc ibed.
D ough Tole ance
The capabili y o su i e o e long-las ing d ough in a oom ai
humidi y was es ed by cul u ing he s ains on 96-well pla e by
ans e ing o e nigh g own cul u e (5 ml LB, +37°C, 210 pm)
in 1:10 a io o LB. One hund ed mic oli e s pe well was used
wi h ou eplica es o each s ain. Pla es we e incuba ed a
+37°C o 3 days in o de o g ow biofilm. A e 3 days, pla es
we e eloca ed o RT and lids we e emo ed o ensu e o al
e apo a ion o media. A e 12 days, 2 mon hs and 6 mon hs in
d ough , 200 ml o esh LB was added in o each well and he
g ow h cu es we e measu ed a +37°C, 595 nm wa eleng h, 20 h
in 5 min in e als.
Mo phological Cha ac e iza ion
Mo phological cha ac e is ics o he colonies we e analyzed. In
o de o ge single colonies, all s ains we e cul u ed in 5 ml LB
in +37°C, 210 pm o e nigh . Cul u es we e dilu ed in o 10
−6
in
wa e and pla ed on o LB-aga pla es. Pla es we e incuba ed a
+37°C o o e nigh and he colonies we e pho og aphed. S ains
EKP5, EKP3 and EKP22 we e ound o p oduce anslucen
colonies along wi h he adi ional colonies. These anslucen
colonies we e u he cul i a ed by ans e ing one colony o a
esh LB-aga pla e and incuba ed a +37°C o e nigh o
ans e ed in o liquid cul u e o 5 ml LB and cul u ed a +37°C
and 210 pm o e nigh be o e pla ing in 10
−6
dilu ion on o he
new LB-pla es. G ow h densi ies we e calcula ed om hese pla es.
F om he same liquid cul u e used o he g ow h densi y defini ion,
1:100 dilu ions we e made in o LB and g ow h cu es we e
measu ed in +37°C, 595 nm wa eleng h, 20 h in 5 min in e als.
S a is ical Analyses
To explo e i epidemic o non-epidemic s ains can be
cha ac e ized by hei capabili ies, we pe o med disc iminan
analysis using MASS package (Venables & Ripley, 2002)inR
( e sion. 3.3.2). E ec s o indi idual a iables on disc iminan
unc ion we e es ed by eg essing p edic ed alues agains
o iginal a iables. O e all pe o mance o disc iminan
unc ion was add essed by Bayesian logis ic eg ession o
epidemic s a us agains p edic ed alues o disc iminan
unc ion using S an wi h R (McEl ea h, 2016).
RESULTS
Selec ion o S ains and Genomic Analysis
We selec ed 14 K. pneumoniae s ains o de ailed pheno ypic
and genomic analysis in an a emp o iden i y cha ac e is ics
ha may po en ially associa e wi h epidemic STs. The s ains
we e abb e ia ed ei he as EKP o NKP o Epidemic and Non-
epidemic K. pneumoniae, espec i ely, and he s ain numbe
was de i ed om an in e nal naming sys em. The s ains and
hei key genomic ai s a e lis ed in Table 1. No e ha wo
epidemic STs a e ep esen ed wice (ST14 and ST11), bu hei
gene ic ea u es di e om one ano he and we e hus selec ed
o pheno ypic s udies in o de o e alua e whe he he
pheno ypes o di e en s ains o a single ST a e simila .
Koskinen e al. Epidemic and Non-Epidemic Klebsiella pneumoniae
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999244

Epidemic and non-epidemic STs a e no g ouping oge he when
he genomic egions used o sequence yping a e used o in e
gene ic ela ionship (Figu e 1). As such, he epidemic s ains do
no appea o sha e a common ances o ha di e ged om non-
success ul s ains. The e o e, epidemic success is no likely o be
linked o a single e ically inhe i ed (gene ic) ai , which
e ol ed once. This howe e does no exclude he possibili y
ha ai s p eceding epidemic sp ead a e ans e ed ho izon ally
be ween s ains o di e en STs. In iguingly, s ains EKP3 and
EKP22 (bo h ST14) and EKP8 and EKP24 (bo h ST11) a e
pheno ypically di e en ega dless o he same ST
(Supplemen a y Figu e 1), some imes being e en he opposi e
pheno ypic ex emes ou o all s ains. Colony mo phologies o
all s ains a e highly simila (Supplemen a y Figu e 2).
The genomes o all he s ains we e anno a ed and hei
o e all gene con en s compa ed. Based on he anno a ion,
epidemic and non-epidemic s ains appea o be gene ally
uni o m me abolically and unc ionally, hence p o iding no
appa en genome-le el design di e ences o explain epidemic
quali ies (Table 2). Nei he he p esence o absence o CRISPR
sys em o he numbe o CRISPR loci appea o be linked wi h
epidemic success ulness. Also, he numbe o mobile elemen s
such as p ophages o plasmids do no associa e specifically wi h
ei he g oup. The mos ob ious po en ially explana o y ea u es,
i.e. i ulence genes and an ibio ic esis ance genes, a e simila
be ween epidemic and non-epidemic K. pneumoniae despi e o
di e ences among indi idual s ains (Tables 1 and 2). We
u he specula ed ha some gene ally o e looked ea u es
such as sho open eading ames (ORFs) o leng h 30 o 150
nucleo ides could possibly be linked wi h epidemic quali ies.
These genes a e a ely iden ified as coding egions wi h
au oma ed anno a ion algo i hms despi e o he ac ha hey
a e some imes ansc ibed and ansla ed and may eflec ecen
adap a ions o new li e s a egies o specific condi ions ( ha may
be ela ed o epidemic sp ead). We p epa ed an algo i hm o
ex ac all sho ORFs which a e p eceded by a (nea -)pe ec
ibosome biding si e and which do no o e lap wi h exis ing
anno a ed ORFs (Py hon code is a ailable in Supplemen a y File
1). On a e age, app oxima ely 200 unanno a ed ORFs we e
ex ac ed om he sequences. Ye , while pu a i e sho genes
a e common, hei coun is simila in epidemic and non-
epidemic s ains (da a no shown).
Pheno ypic Quali ies o he S ains
I is possible ha epidemic sp ead selec s o o is p eceded by
specific pheno ypic ai s. These ai s may no necessa ily be
linked wi h any pa icula gene ic ea u e as he e may be se e al
mu a ional pa hways o acqui e he quali y and hence, hey may
be di ficul o iden i y wi h gene ic o genomic compa isons. As
such, we lis ed a numbe o measu able pheno ypes ha may be
linked wi h epidemic success. The ai s, hei specula ed
associa ion wi h epidemic dispe sal, and he a iables used in
his s udy a e lis ed in Table 3. The o iginal da a om
measu emen s is a ailable in Supplemen a y File 2.
FIGURE 1 | Phylogene ic dis ance based on Sequence Type sequences.
Koskinen e al. Epidemic and Non-Epidemic Klebsiella pneumoniae
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999245
TABLE 2 | Genomic analysis o indi idual s ains based on RAST-anno a ion.
EKP03 EKP05 EKP08 EKP10 EKP11 EKP22 EKP24 NKP01 NKP02 NKP18 NKP20 NKP25 NKP28 NKP30
Cell Wall and Capsule 217 230 218 230 234 214 220 216 196 214 217 217 193 227
Capsula and ex acellula
polysacch ides
38 54 42 54 57 38 41 38 42 36 37 37 39 49
G am-nega i e cell wall componen s 91 90 91 90 90 38 90 89 66 90 88 92 69 92
Cell wall and capsule—no
subca ego y
88 86 85 86 87 85 89 89 88 87 88 88 85 86
Vi ulence, Disease, and De ense 148 150 141 155 156 154 136 146 143 154 144 168 134 129
Adhesion 77777777777777
Toxins and supe an igens 00000000000000
Bac e iocins, ibosomally
syn hesized an ibac e ial pep ides
12 12 12 12 12 12 12 12 12 12 12 12 12 12
Resis ance o an ibio ics and oxic
compounds
121 123 114 128 129 127 109 119 116 127 117 141 107 102
Vi ulence, disease, and de ense 00000000000000
De ec ion 00000000000000
In asion and in acellula esis ance 88888888888888
Phages, P ophages,
T ansposable elemen s,
Plasmids
47 51 62 57 81 49 44 29 78 78 66 19 77 8
Phage amily-specific subsys ems 00000000000000
T ansposable elemen s 00400007000050
Phages, p ophages 47 50 57 57 80 48 43 20 77 77 66 19 70 8
Phages, p ophages, ansposable
elemen s, plasmids—no
subca ego y
01101112110010
Pa hogenici y islands 00000000000000
Gene ans e agen 00000000000000
Plasmid ela ed unc ions 00000000000000
Memb ane T anspo 231 267 214 246 289 346 238 293 247 310 270 325 279 234
P o ein sec e ion sys em, Type II 19 19 19 19 19 19 19 19 19 19 19 19 19 19
ABC anspo e s 78 72 65 72 78 79 60 71 68 69 78 71 71 75
P o ein sec e ion sys em, Type VII
(Chape one/Ushe pa hway, CU)
20 25 20 25 19 20 25 25 19 20 20 28 26 22
P o ein ansloca ion ac oss
cy oplasmic memb ane
77777777777777
P o ein sec e ion sys em, Type V 00000000002002
P o ein sec e ion sys em, Type I 50000500050000
Ca ion anspo e s 24 24 22 24 24 24 24 24 23 24 24 25 22 22
P o ein sec e ion sys em, Type III 00000000000000
P o ein sec e ion sys em, Type VI 19 17 16 17 14 19 15 14 0 14 15 14 13 22
P o ein sec e ion sys em, Type VIII
(Ex acellula nuclea ion/p ecipi a ion
pa hway, ENP)
00000000000000
P o ein and nucleop o ein sec e ion
sys em, Type IV
21 68 28 46 92 137 49 97 73 114 67 125 87 29
I on Acquisi ion and Me abolism 80 77 67 76 74 80 68 76 74 67 74 70 69 67
Side opho es 18 21 20 20 18 18 20 20 22 18 18 18 17 19
I on acquisi ion and me abolism—
no subca ego y
62 56 47 56 56 62 48 56 52 49 56 52 52 48
I on anspo 00000000000000
Mo ili y and Chemo axis 12 10 9 10 11 13 8 10 11 10 11 10 10 10
Magne o axis 00000000000000
Mo ili y and chemo axis—no
subca ego y
12 10 9 10 11 13 8 10 11 10 11 10 10 10
Flagella mo ili y in P oka yo a 00000000000000
Social mo ili y and nonflagella
swimming in bac e ia
00000000000000
Regula ion and Cell Signaling 173 167 165 167 168 173 175 178 174 180 170 172 171 174
Quo um sensing and biofilm
o ma ion
13 13 13 13 13 13 13 13 13 13 13 13 13 13
Regula ion o i ulence 98888888888888
P og ammed cell dea h and oxin-
an i oxin sys ems
16 18 15 17 17 18 23 23 17 19 20 18 19 24
DNA Me abolism 155 131 126 140 173 157 145 151 138 144 143 133 142 142
CRISPs 70007700000007
Koskinen e al. Epidemic and Non-Epidemic Klebsiella pneumoniae
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999246
Supplemen a y Figu e 1 summa izes he esul s o epidemic
and non-epidemic s ains.
We ca ied ou a disc iminan analysis o he measu ed
pheno ypic da a. The analysis associa ed se e al pheno ypic
measu emen s wi h disc iminan unc ion. The disc iminan
unc ion is a combina ion o linea e ec s o a iables ha gi e
bes sepa a ion o he da a o dis inc classes, in his s udy
epidemic and non-epidemic s ains. This app oach e ealed
highly significan (pos e io alues did no o e lap wi h ze o)
likelihood o a gi en ai o belong o ei he epidemic o non-
epidemic g oup ( esol ed wi h Bayesian logis ic eg ession
be ween p edic ed alues o disc iminan unc ion and
epidemic s a us, Bayesian R
2
= 45%, Figu e 2). Reg essing
p edic ed alues o disc iminan unc ion agains o iginal
a iables indica ed especially s ong ole o a ious
measu emen s o d ough ole ance in disc iminan unc ion
(Figu e 3) and non-epidemic s ains wi h high disc iminan
unc ion sco e we e clea ly mo e d ough ole an .
C oss-Inhibi ion and An ibac e ial
Po en ial o a Pu a i e Colicin
We u he s udied he c oss-s ain inhibi ion gi en ha he
epidemic success ulness could eme ge om he abili y o
epidemic s ains o supp ess non-epidemic s ains du ing
sp ead be ween hospi als o hos s (Figu e 4). The inhibi ion,
when de ec ed, was de e mined wi h a dilu ion se ies o be ei he
due o molecula ac i i y o p ophage induc ion (dilu ed phages
p oduce dis inc plaques unlike inhibi ing molecules). Pai wise
inhibi ion was in equen and no gene al pa e n be ween
epidemic and non-epidemic s ains was iden ifiable.
As a cu iosi y, we selec ed one c oss-inhibi ing s ain pai o
mo e de ailed analysis. NKP2 p oduces a componen in o i s
medium ha inhibi s EKP24 (ST11). Gi en he wide dispe sal o
ST11 K. pneumoniae and i s associa ion wi h NDM-1 encoding
plasmids and hype i ulence (Gu e al., 2017), he inhibi ing
ac o could p o ide a possible way o con ol hese s ains.
Howe e , he s ain used in his s udy is no hype i ulen and
he e o e assessing di ec applicabili y agains hype i ulen
s ains was no conduc ed. Genomic compa ison o NKP2 and
EKP24 e ealed he p esence o genes o Colicin E3 in NKP2
ha we e absen om EKP24. Hence, Colicin E3 was
hypo hesized o be he inhibi ing molecule. Co-cul u ing o
hese s ains in he same medium demons a es ha EKP24 is
unable o mul iply. We u he s udied he adap a ion o EKP24
o he con inuous p esence o he hypo hesized Colicin E3 by
se ially cul u ing EKP24 in he p esence o NKP2 medium
ex ac o 4 weeks (n = 5). These cul u es we e e eshed h ee
imes a week. A e 2 weeks, we emo ed he selec ion om fi e
TABLE 3 | S udied cha ac e is ics and hei hypo hesized associa ion wi h epidemic capabili y.
T ai /quali y Va iables/ ac o s
used in his s udy
Rele ance
Vi ulence genes Numbe / ype Numbe o i ulence genes may di ec ly a ec he s ain’s po en ial o cause in ec ions
An ibio ic esis ances Numbe / ype An ibio ic esis ance can comp omise ea men , hus causing p olonged in ec ions and inc ease he ime du ing
which bac e ia dispe se
G ow h a e/densi y Max g ow h a e ( ),
Max g ow h densi y
(K)
Fas e and mo e dense g ow h may inc ease he bac e ial load in he su ounding en i onmen and hus i s epidemic
po en ial
G ow h empe a u e 22°C ( oom
empe a u e), 37°C
G ow h di e ences in oom empe a u e s 37°C may eflec ade-o s in wi hin- and ou side-hos en i onmen s and
hus i s adap a ion o he en i onmen s he hos
G ow h in di e en
media
LB, DMEM, BHI Po en ial o g ow in di e en nu ien en i onmen s may p o ide bac e ia mo e oppo uni ies o p oli e a e in
al e na i e habi a s and hus p o ide possibili ies o su i e ou side he hos
G ow h in di e en
nu ien le els
1, 10, 100% L Po en ial o g ow in a ying nu ien le els may p o ide ad an age in di e en en i onmen s and hence a ec i s
dispe sal o new hos s
G ow h in a ying pH pH 3, pH 4, pH 5, pH
6, pH 7
Bac e ia may be exposed o di e en pH in he en i onmen and he hos (phagocy osis, skin, gas oin es inal ac )
and he sensi i i y o pH may dec ease he changes o dispe sal
Po en ial o compensa e
su ounding pH
pH 3, pH 4, pH 5 Po en ial o modi y he su ounding mic oen i onmen may play a c ucial ole in he bac e ial chances o adap o
fluc ua ing en i onmen al pH and hence a ec i s dispe sal
Resilience in E OH 20, 50, 75, 90% E OH Su i al in he p esence o alcohol con aining sani ize s may di ec ly a ec he pe sis ence o he s ain in he
en i onmen and hence influence i s po en ial o ge ansmi ed
Su i al in he absence
o wa e
14 days, 2 mon hs, 6
mon hs
The po en ial o wi hs and d ough can inc ease he imespan du ing which pa hogen emains iable in hospi al
en i onmen and hence a ec i s changes o ge ansmi ed o new hos s
Reco e y a e d ough 14 days, 2 mon hs, 6
mon hs
Fas e eco e y a e d ough can p o ide bac e ia inc eased po en ial o colonize o in ec new hos s
G ow h in mixed/
spa ially s uc u ed
popula ion
0 pm, 230 pm G ow h as a di e en ly s uc u ed popula ion may play a ole in a ious s ages o in ec ion and pe sis ence in he
en i onmen and may he e o e a ec he s ain’s po en ial o dispe se
Numbe o plasmids Numbe /Inc- ype Plasmids o en ca y genes ha benefi he bac e ium in specific condi ions and hence hei numbe may be ela ed
o su i al in a ious condi ions inside and ou side he hos
Genomic p ophages Numbe /Inc- ype Ac i a ion o p ophages may cause in ec ions in compe ing K. pneumoniae s ains and may he e o e p o ide he
p ophage ca ying s ain an ad an age in si ua ions whe e se e al s ains occupy he same en i onmen
S ain-specific inhibi ion Pai wise inhibi ion P oduc ion o bac e iocins o o he an imic obials may inhibi he g ow h o compe ing s ains and hus may hinde
he po en ial o sensi i e s ains o dispe se in o en i onmen s o hos s wi h o he K. pneumoniae s ains
Sho ORFs in he
genome
Numbe o ORFs Numbe o sho open eading ames in he genome may eflec bac e ial adap i e his o y and hence may be linked
o epidemic po en ial
Koskinen e al. Epidemic and Non-Epidemic Klebsiella pneumoniae
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999247
addi ional eplica es. Resis ance o NKP2 medium (wi h
hypo hesized Colicin E3) eme ged al eady du ing he fi s
cul u e ans e , and i emained s able e en a e he emo al
o selec ion. Re-sequencing o h ee pu a i e Colicin E3 esis an
samples e ealed a p e alen mu a ion in Ae obac in side opho e
ecep o Iu A. This mu a ion was absen in a cul u e ha was no
exposed o NKP2 ex ac . Iu A has been shown o se e as a
ecep o o cloacin DF13 (Van Tiel-Menk eld e al., 1982),
which is homologous o Colicin E6 and E3 (Aku su e al., 1989).
As such, we hypo hesize ha NKP2 ex ac apidly selec ed o
Iu A mu an s. Also, i is wo h no ing ha NKP2 medium was
no obse ed o inhibi he g ow h o he o he ST11 s ain EKP8,
hence showing na ow ac i i y. Al oge he , he hypo hesized
Colicin E3 does no appea o p o ide e ficien an imic obial
ac i i y agains d ug- esis an K. pneumoniae s ains e en when
he a ge ed s ain is ini ially sensi i e o he colicin.
DISCUSSION
K. pneumoniae has become one o he p io i y d ug- esis an
pa hogens in hospi al se ings wo ldwide. Some K. pneumoniae
STs a e mo e p e alen compa ed o o he s, which p o okes he
ques ion whe he he e a e quali ies in hese gene ically ela ed
g oups ha ha e made hem mo e po en o dispe sal. He e, we
s udied mul iple pheno ypic and geno ypic cha ac e is ics o
se en epidemic and non-epidemic ca bapenem esis an K.
pneumoniae isola es ha eme ged om di e en pa s o he
wo ld. None o he specific gene ic quali ies associa ed uni o mly
wi h epidemic o non-epidemic s ains. O e all, his again
sugges s ha sequence yping is no an op imal app oach o
in e ing quali ies o indi idual pa hogens. In o he wo ds, he e
a e K. pneumoniae s ains ha a e ela i ely di e en om one
ano he o hei specific cha ac e is ics while s ill g ouping
oge he in ST-analyses. Pheno ypic cha ac e iza ion e ealed
indica ion ha , while gene ic di e ences we e minu e based on
ST-analyses, pheno ypic di e ences ha sepa a e epidemic and
non-epidemic K. pneumoniae s ains do exis . Such appa en ly
con o e sial esul could eme ge i pheno ypes a e s ongly
dic a ed by gene ic di e ences o he han indica ed by co e
genome-based ST-analyses o e en by epigene ic modifica ions
(Casadesus and Low, 2006).
While mos o he s udied ac o s could no explici ly help
explain epidemic quali ies, he s ong associa ion o d ough
ole ance wi h epidemically non-success ul s ains could gi e
some insigh s on he dissemina ion. All he isola es in his s udy
we e able o wi hs and 6 mon hs o d yness. In e es ingly, non-
epidemic s ains gene ally eco e ed as e and in o highe densi y
a e he d ough , which p esen s he possibili y ha non-epidemic
s ains may ha e quali ies ha p o ide hem oppo uni ies o
cause in ec ions in specific cases, o example, a e long- e m
esidence on su aces (K ame e al., 2006). Bac e ia ha e mul iple
ways o p o ec i sel du ing he d ough . Fo example, e ec i e
biofilm o ma ion is a majo p o ec ion mechanism in bac e ial
FIGURE 2 | P edic ed Linea disc iminan alues in epidemic ( ed) and non-epidemic (g een) K. pneumoniae s ains.
Koskinen e al. Epidemic and Non-Epidemic Klebsiella pneumoniae
F on ie s in Cellula and In ec ion Mic obiology | www. on ie sin.o g Feb ua y 2021 | Volume 11 | A icle 5999248