Pe asi e ansc ip ion enhances he accessibili y o
H-NS–silenced p omo e s and gene a es bis abili y in
Salmonella i ulence gene exp ession
Na a Figue oa-Bossi
a
,Ma
ıa An onia S
anchez-Rome o
b
, Pa icia Ke bo iou
a
, Delphine Naquin
a
, Cla a Mendes
a
, Philippe Bouloc
a
,
Josep Casades
us
c
, and Lionello Bossi
a,1
Edi ed by Thomas Silha y, P ince on Uni e si y, P ince on, NJ; ecei ed Feb ua y 18, 2022; accep ed June 1, 2022
In Esche ichia coli and Salmonella, many genes silenced by he nucleoid s uc u ing p o-
ein H-NS a e ac i a ed upon inhibi ing Rho-dependen ansc ip ion e mina ion.
This esponse is poo ly unde s ood and di ficul o econcile wi h he iew ha H-NS
ac s mainly by blocking ansc ip ion ini ia ion. He e we ha e analyzed he basis o
he up- egula ion o H-NS–silenced Salmonella pa hogenici y island 1 (SPI-1) in cells
deple ed o Rho-co ac o NusG. E idence om gene ic expe imen s, semiquan i a i e
50 apid amplifica ion o complemen a y DNA ends sequencing (5’RACE-Seq), and
ch oma in immunop ecipi a ion sequencing (ChIP-Seq) shows ha ansc ip ion o igi-
na ing om spu ious an isense p omo e s, when no s opped by Rho, elonga es in o a
H-NS–bound egula o y egion o SPI-1, displacing H-NS and ende ing he DNA
accessible o he mas e egula o HilD. In u n, HilD’s abili y o ac i a e i s own an-
sc ip ion igge s a posi i e eedback loop ha esul s in ansc ip ional ac i a ion o he
en i e SPI-1. Significan ly, single-cell analyses e ealed ha his mechanism is la gely
esponsible o he coexis ence o wo subpopula ions o cells ha ei he exp ess o do
no exp ess SPI-1 genes. We p opose ha cell- o-cell di e ences p oduced by s ochas ic
spu ious ansc ip ion, combined wi h eedback loops ha pe pe ua e he ac i a ed
s a e, can gene a e bimodal gene exp ession pa e ns in bac e ial popula ions.
pe asi e ansc ip ion jH-NS jsilencing jpa hogenici y islands jbis abili y
T ansc ip omic analyses o Esche ichia coli bac e ia exposed o an inhibi o o ansc ip-
ion e mina ion ac o Rho ha e led o he ecogni ion ha a majo ac i i y o Rho
and i s co ac o NusG in g owing cells is de o ed o genome-wide supp ession o ubiq-
ui ous an isense ansc ip ion genome-wide (1, 2). Thus, indi ec ly, hese findings a e
also pe inen o ansc ip ion ini ia ion, as hey un eil he exis ence o a high-le el
spu ious ansc ip ional noise appa en ly cu bed by he e mina ion ac i i y o Rho.
This no ion gained momen um wi h he demons a ion ha E. coli genes con ain a
mul i ude o in agenic p omo e s in bo h sense and an isense o ien a ions (3–5). The
phenomenon is pa icula ly d ama ic in genomic egions hough o o igina e om
ho izon al ans e whose ypically highe adenine and hymine (AT) con en ma ches
he sequence composi ion o he a e age bac e ial p omo e (6). The disp opo ionally
high numbe o p omo e -like sequences in AT- ich DNA can ac ually be a sou ce o
oxici y by causing RNA polyme ase i a ion (7). The abo e s udies ha e an addi ional
common denomina o : They implica e he nucleoid s uc u ing p o ein H-NS. On he
one hand, he si es o in agenic Rho-dependen e mina ion colocalized wi h he
egions bound by H-NS (1); on he o he hand, H-NS was shown o play he majo ole
o silencing he spu ious in agenic p omo e s (5). These findings enewed he in e es in
unde s anding he hidden complexi y o H-NS–RNA polyme ase in e ac ions (8).
A small abundan p o ein, H-NS oligome izes along he DNA upon binding o
high-a fini y AT- ich nuclea ion si es and sp eading coope a i ely o adjacen sequences
h ough lowe -a fini y in e ac ions (9–11). The oligome iza ion p ocess gene a es a
highe -o de supe helical s uc u e hough o con ibu e o DNA condensa ion (12).
The esul ing nucleop o ein filamen s e ec i ely coa he DNA and he eby hampe
p omo e ecogni ion by RNA polyme ase (13). In addi ion, H-NS can ep ess an-
sc ip ion h ough he o ma ion o b idged o looped DNA s uc u es ha ap RNA
polyme ase in he open complex (14, 15) o ac as oadblocks agains ansc ip elonga-
ion (16, 17). In pa icula , b idged bu no linea H-NS filamen s ha e been shown o
p omo e Rho-dependen ansc ip ion e mina ion by inc easing ansc ip ional paus-
ing in i o (17). H-NS has gained conside able a en ion since he disco e y o i s ole
as a xenogeneic silence . Due o i s a fini y o AT- ich DNA, H-NS p e e en ially
binds o and p e en s he exp ession o sequences acqui ed h ough ho izon al ans e
Significance
In bac e ia, a ac ion o
ansc ip ional noise o igina es
om p omo e -like sequences
ound in bo h o ien a ions wi hin
ansc ip ional uni s. This wo k
shows ha ansc ip ion om
spu ious p omo e s can
elie egenesilencing by H-NS, a
bac e ial ch oma in p o ein ha
oligome izes along he DNA
o ming filamen s and b idged
s uc u es. Upon in ading a pa ch
o oligome ized H-NS, elonga ing
ansc ip ion complexes “unzip”
he nucleop o ein filamen ,
making he DNA accessible o
egula o y p o eins and RNA
polyme ase. In Salmonella, his
mechanism igge s a posi i e
eedback loop ha ac i a es a
i ulence egula o y p og am. The
occu ence o wo subpopula ions
o cells ha , al hough gene ically
iden ical, ei he exp ess o do no
exp ess i ulence genes sugges s
ha s ochas ic spu ious
ansc ip ion can play a ole in he
gene a ion o bis abili y in
bac e ia.
Au ho con ibu ions: N.F.-B., M.A.S.-R, J.C. and L.B.
designed esea ch; N.F.-B., M.A.S.-R, P.K., C.M. and L.B.
pe o med esea ch; D.N. and P.B. con ibu ed new
eagen s/analy ical ools; N.F.-B, M.A.S.-R., D.N. and L.B.
analyzed da a; and N.F.-B., J.C. and L.B. w o e he pape .
The au ho s decla e no compe ing in e es .
This a icle is a PNAS Di ec Submission.
Copy igh © 2022 he Au ho (s). Published by PNAS.
This a icle is dis ibu ed unde C ea i e Commons
A ibu ion-NonComme cial-NoDe i a i es License 4.0
(CC BY-NC-ND).
See online o ela ed con en such as Commen a ies.
1
To whom co espondence may be add essed. Email:
[email p o ec ed].
This a icle con ains suppo ing in o ma ion online a
h p://www.pnas.o g/lookup/suppl/doi:10.1073/pnas.
2203011119/-/DCSupplemen al.
Published July 18, 2022.
PNAS 2022 Vol. 119 No. 30 e2203011119 h ps://doi.o g/10.1073/pnas.2203011119 1o 9
RESEARCH ARTICLE
|
MICROBIOLOGY
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
(18, 19). In doing so, H-NS p o ec s he bac e ium agains he
oxici y o o eign DNA (7, 20, 21) and allows he e olu ion o
mechanisms o coop ing newly acqui ed unc ions and egula -
ing hei exp ession (22, 23). Indeed, he as majo i y o
H-NS–silenced genes a e igh ly egula ed and exp essed only
unde a limi ed se o condi ions. Ac i a ion o H-NS–silenced
genes ypically esul s om he binding o he ac ion o egula-
o s able o displace H-NS (24–26). Unlike classical gene ac i-
a ion, H-NS coun e silencing exhibi s conside able flexibili y
in he spa ial a angemen o he egula o p o ein ela i e o
he p omo e (27).
In Salmonella en e ica, H-NS silences mos o he genes ha
con ibu e o i ulence, including Salmonella pa hogenici y
islands (SPIs) ha a e specifically ac i a ed in he en i onmen
o he in ec ed hos (18, 19). SPI ac i a ion occu s in he o m
o a hie a chical and empo al egula o y cascade ha begins
wi h he exp ession o SPI-1, a 44-kb island encoding a ype III
sec e ion sys em (T3SS) ha deli e s e ec o p o eins p omo -
ing in es inal coloniza ion and epi helial cell in asion (28, 29).
Se e al lines o e idence sugges ha he p ocess is ini ia ed by
HilD, an SPI-1–encoded A aC- ype ansc ip ional egula o
ha ac i a es he exp ession o a second mas e egula o , HilA,
which in u ns ac i a es he T3SS along wi h he p oduc o a
hi d egula o y gene, in F (30). Ac ing as a hub in eg a ing
di e se en i onmen al and physiological signals, HilD is i sel
egula ed a mul iple le els including messenge RNA (mRNA)
s abili y (31), mRNA ansla ion (32, 33), and p o ein ac i i y
(33, 34). Howe e , cen al o he egula o y cascade is he abil-
i y o HilD o ac i a e i s own syn hesis. HilD binds o an
ex ended egion ups eam o he hilD p omo e in i o (35)
and in i o (36). The p esence o his egion among he DNA
agmen s bound by H-NS in ch oma in immunop ecipi a ion
(ChIP) expe imen s sugges s ha HilD s imula es ansc ip ion o
i s own gene by an agonizing H-NS (37). In e es ingly, SPI-1
exhibi s a bis able exp ession pa e n cha ac e ized by he p esence
o wo subpopula ions o cells ha ei he exp ess o do no exp ess
SPI-1 genes (38–43). In labo a o y cul u es, he SPI-1
OFF
popula-
ion as ly p edomina es; howe e , SPI-1
ON
cells a e con inuously
p oduced and pe sis o se e al gene a ions (42, 44) despi e he
fi ness cos associa ed wi h he syn hesis and assembly o he
T3SS, which esul s in g ow h e a da ion (45). Re a ded g ow h,
howe e , makes he SPI-1
ON
subpopula ion ole an o an ibio ics
(46). SPI-1
OFF
cells also benefi om bis abili y: Inflamma ion
igge ed by he T3SS o SPI-1
ON
cells leads o he p oduc ion o
eac i e oxygen species in phagocy es. Such chemicals p oduce e -
a hiona e upon oxida ion o endogenous sul u compounds, and
e a hiona e espi a ion con e s a g ow h ad an age o Salmonella
o e compe ing species o he in es inal mic obio a (47). Fu he -
mo e, SPI-1
OFF
cells can in ade he in es inal epi helium, a capac-
i y ha may benefi he popula ion as a whole by coun e ing
in asion by a i ulen mu an s (38, 39, 42).
We ecen ly ound ha inhibi ing Rho-dependen ansc ip-
ion e mina ion, by mu a ion o h ough he deple ion o Rho
co ac o NusG, causes massi e up- egula ion o many Salmonella
i ulence genes including all majo SPIs (16). The magni ude
and he span o hese e ec s sugges ed ha hey we e no p o-
duced locally bu eflec ed he ac i a ion o a global egula o y
esponse. This led us o u n ou a en ion o HilD. The wo k
desc ibed below confi med he HilD in ol emen and p o-
ided new insigh on he in e play be ween ansc ip ion elon-
ga ion and bac e ial ch oma in. In pa icula , ou da a sugges
ha H-NS–bound egions a e no comple ely impe meable o
RNA polyme ase. Occasional spu ious ansc ip ion ini ia ion
e en s wi hin hese egions igge a elay cascade whe eby
elonga ing ansc ip ion complexes, i no s opped by Rho,
dislodge H-NS oligome s, making mo e p omo e s accessible
o RNA polyme ase and egula o y p o eins. In addi ion, hese
findings suppo a model o he mechanism unde lying SPI-1
bis abili y.
Resul s
Mos o he Salmonella Response o NusG Deple ion Is HilD-
Media ed. In all analyses desc ibed below, NusG deple ion is
achie ed in s ains wi h he sole copy o he nusG gene used o
a phage p omo e unde he con ol o an a abinose-inducible
ep esso (16). In he p esence o a abinose (ARA), ac i a ion
o he ep esso gene causes he nusG gene o be u ned o and
i s p oduc o be p og essi ely deple ed. Al hough NusG is
essen ial o Salmonella iabili y (48), he ea men is no le hal
since esidual NusG syn hesis is su ficien o suppo g ow h.
In ac , g ow h is nea ly una ec ed by ARA un il bac e ia en e
ea ly s a iona y phase. A his poin he g ow h a e becomes
significan ly educed, appa en ly as a side e ec o he s ong
ac i a ion o SPIs (45).
To assess he possible ole o HilD in he esponse o NusG
deple ion, we measu ed he exp ession o lacZ ansla ional
usions o h ee SPI genes: in B (SPI-1) sseE (SPI-2), and sopB
(SPI-5) in a hilD
+
s ain and in a s ain in which he hilD gene
is eplaced by a e RA casse e. ARA exposu e elici ed a sha p
inc ease in he exp ession o all h ee usions in bo h hilD
+
and
hilD
backg ounds; howe e , he changes in he hilD
+
s ain
occu ed wi hin a ange be ween 10- o 50- old highe han in
he ΔhilD:: e RA mu an (SI Appendix, Fig. S1), poin ing o
he HilD in ol emen in he ARA-media ed ac i a ion o SPIs.
To examine his esponse a he single-cell le el, we cons uc ed
in- ame usions o supe olde g een fluo escen p o ein (GFP
SF
)
o HilD- egula ed genes. Two usions, in hilA and in B,we e
ob ained by inse ing he g p
SF
open eading ame in he a ge
gene; a hi d usion, also in hilA, was made by concomi an ly
dele ing a 28,266-bp segmen spanning nea ly he en i e SPI-1
po ion on he 30side o he usion bounda y. As ini ial expe i-
men s showed no significan di e ences in he beha io s o he
h ee s ains, only he s ain wi h he 28-kb SPI-1 dele ion (hilA::
g p
SF
ΔK28) was used o subsequen analyses.
NusG Deple ion P omo es SPI-1 Bis abili y. Cells ca ying
hilA::g p
SF
ΔK28 display a ypical bis able pheno ype cha ac e -
ized by he p esence o wo subpopula ions o bac e ial cells, o
which only one subpopula ion shows GFP fluo escence (Fig.
1A). Significan ly, g ow h in he p esence o ARA causes he
a io be ween hilA
ON
and hilA
OFF
cells o inc ease d ama ically
in ea ly s a iona y phase (Fig. 1A). Flow cy ome ic measu e-
men s show he inc ease o be mo e han 10- old (Fig. 1Band
SI Appendix, Fig. S2). SPI-1 bis abili y has been linked o he
sel -ac i a ing na u e o hilD exp ession and i is hough o
eflec cell- o-cell a iabili y in HilD le els (32, 33, 41). In line
wi h his model, hilA
ON
cells a e no longe de ec ed in a s ain
ca ying a 309-bp in- ame dele ion emo ing he DNA bind-
ing mo i in he ca boxyl- e minal domain o HilD (Fig. 1C
and SI Appendix, Fig. S2). These esul s sugges ha NusG
deple ion allows a la ge ac ion o cells o each he HilD
au oac i a ion h eshold. Consis en wi h his conclusion, RNA
quan ifica ion by RT-qPCR shows ha he ARA ea men
causes a la ge inc ease in hilD ansc ip ion when HilD is unc-
ional bu a smalle inc ease in he hilDΔ309 mu an (Fig. 2A).
The analysis e eals ha HilD is also needed o he exp ession
o he adjacen p gH gene (Fig. 2B). A fi s sigh , his may
2o 9 h ps://doi.o g/10.1073/pnas.2203011119 pnas.o g
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
seem su p ising since p gH is hough o be ac i a ed by HilA,
no by HilD (37, 49), and he s ain used in Fig. 2Bca ies
he hilA::g p
SF
ΔK28 allele which emo es o e wo- hi ds o
he hilA sequence. Howe e , we no e ha he usion e ains he
N- e minal 112 amino acids domain o HilA p e iously impli-
ca ed in p gH p omo e ecogni ion (50), sugges ing ha he
HilA-GFP chime a e ains he abili y o ac i a e p gH.
Pe asi e T ansc ip ion Ac i a es he hilD P omo e . The cen-
al ole o HilD in he esponse o NusG deple ion was u he
co obo a ed by he obse a ion ha ARA ea men s imula es
HilD binding o he hilD p omo e egion (Fig. 2C), a esponse
ha co ela es wi h he ac i a ion o he hilD and p gH p omo e s
(Fig. 2Dand E). This las se o da a was ob ained pe o ming
semiquan i a i e 50 apid amplifica ion o complemen a y DNA
(cDNA) ends (50RACE) gene a ed by empla e swi ching e e se
ansc ip ion (51). In his me hod, e e se ansc ip ion is p imed
by a gene-specific p ime and ca ied ou in he p esence o a
empla e-swi ching oligonucleo ide (TSO). The 50ends o RNAs
a e defined by he posi ion he deoxycy idine epea s ( ypically 3
o 4) ha a e added by e e se ansc ip ase when i eaches he
end o he RNA and swi ches o he TSO (51). Use o p ime s
ca ying Illumina adap o s o he PCR s ep allows o he analy-
sis o be pe o med by high- h oughpu sequencing (RACE-Seq).
He e, p ime s we e designed o de ec ansc ip ion ini ia ion ak-
ing place a he p ima y hilD and p gH p omo e saswellasa
h ee seconda y p omo e s p e iously iden ified by K €oge e al.
(52) (SI Appendix,Fig.S3). Read summa iza ion a each o he
ansc ip ion s a si es (TSSs) showed ha ARA exposu e causes
he numbe o ansc ip s ini ia ing a hilD and p gH p ima y
TSSs o inc ease 7- old and 13- old, espec i ely (Fig. 2Dand E).
The mos likely explana ion o his e ec is ha NusG deple ion
allows ansc ip ion complexes o med ou side he p gH-hilD p o-
mo e egion o in ade his egion, displacing H-NS and igge -
ing HilD au ogenous ac i a ion. P ime suspec s o his e ec a e
he seconda y hilD p omo e s whose ac i i y con ibu es o he
hilD mRNA inc ease (SI Appendix,Fig.S4A). Howe e , he
expe imen does no dis inguish whe he he inc ease in he num-
be o eads associa ed wi h he seconda y TSSs is due o a la ge
p opo ion o ansc ip s eaching he RT p ime si e (and hus
suscep ible o e e se ansc ip ion) in ARA- ea ed cells o eflec s
an inc ease in p omo e ac i i y. I he la e we e ue, he impli-
ca ion would be ha he seconda y p omo e s a e hemsel es ac i-
a ed by ansc ip ion ini ia ing elsewhe e, p esumably u he
ups eam, and hey elay he e ec o he hilD p omo e . To
add ess his poin , we pe o med pa allel qPCR measu emen s
using p ime s pai s annealing a p oximal and dis al posi ions el-
a i e o he RT p iming si e. Resul s showed ha a significan
ac ion o ansc ip s en e ing he hilD coding sequence in ARA-
ea ed cells ini ia e as a as o e 1,400 bp ups eam om he
hilD p omo e (compa e ed ba s be ween PCR-1 and PCR-2 in
Fig. 2F), hus conside ably ups eam ela i e o he seconda y
p omo e s. I is he e o e concei able ha he elonga ion o hese
o e lapping ansc ip s may ac i a e he seconda y p omo e s.
In e es ingly, his class o ansc ip s is also de ec able in un ea ed
cells(g eenba sinFig.2F) albei a a e y low le el. A simila
end is obse ed in he opposi e s and whe e a ac ion o p gH
RNAs o igina es om an i-sense hilD ansc ip ion (Fig. 2G).
To u he assess he con ibu ion o ups eam ansc ip ion
o hilD p omo e ac i i y, he SPI-1 segmen ex ending om
he le bounda y o he island up o a posi ion 610 bp
ups eam o he hilD main TSS was dele ed and eplaced by a
casse e comp ising he e R gene and he Te R- ep essed P
e A
p omo e (Fig. 3A). The ansc ip ional esponses o ARA and
o he P
e A
induce , anhyd o e acycline (AHTc), alone o
combined, we e analyzed by RT-qPCR and 50RACE-Seq. ARA
was s ill able o ac i a e ansc ip ion o hilD and p gH in he
new backg ound (Fig. 3Band C); howe e , ac i a ion was
mo e mode a e han seen abo e wi h he pa en al s ain (com-
pa e o Fig. 2Aand B, espec i ely); u he mo e, no significan
changes we e obse ed a he le el o he hilD and p gH p i-
ma y TSSs (Fig. 3Dand E). These findings co obo a e he
idea ha he egion dele ed in he new cons uc con ibu es o
he ampli ude o he ARA e ec s. Ac i a ing P
e A
wi h AHTc
s imula es hilD and p gH ansc ip ion (Fig. 3Band C); he e
oo, howe e , he e ec emains limi ed and unde ec able by
50RACE-Seq a he p ima y TSSs (Fig. 3Dand E). In con as ,
Fig. 1. NusG deple ion enhances HilD-dependen SPI-1 bis abili y. The s ains
used, MA14302 (hilD
+
) and MA14561 (hilDΔ309), ca y a hilA-g p
SF
ansla ional
gene usion (hilA::g p
SF
ΔK28) and a ch omosomal P
Tac
p omo e -mChe y gene
usion in he ARA-inducible-NusG deple ion backg ound. (A)Rep esen a i e
image o MA14302 cells g own a 37 °C o ea ly s a iona y phase isualized by
fluo escence mic oscopy unde 100×magnifica ion. (Band C)Rep esen a i e
flow cy ome y analysis o cells om s ains MA14302 (B) and MA14561 (C)
g own as in A.TheGFPfluo escence in ensi y dis ibu ion was examined in
s ains ca ying g p ansla ional usions. The ull geno ypes o MA14302 and
MA14561 a e shown in SI Appendix,TableS1. Fo he cons uc ion o hilA::
g p
SF
ΔK28 and hilDΔ309 by λ ed ecombinee ing, DNA p ime s (lis ed in SI
Appendix,TableS2)we eusedasde ailedinSI Appendix,TableS3.
PNAS 2022 Vol. 119 No. 30 e2203011119 h ps://doi.o g/10.1073/pnas.2203011119 3o 9
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
when AHTc and ARA a e used conjoin ly, ansc ip ion o
bo h hilD and p gH genes is s ongly ac i a ed (Fig. 3Band C)
and an ini ia ion bu s is obse ed a he le el o bo h p ima y
p omo e s (Fig. 3Dand E). Significan ly, his bu s co ela es
wi h inc eased occupancy o he hilD p omo e by he HilD
p o ein (Fig. 3F). The seconda y p omo e s exhibi a simila
o e all esponse, which is, howe e , cha ac e ized by p onounced
sca e in he indi idual A a/AHTc ea men s (SI Appendix,Fig.
S4B). The o e all s eng h o he hilD esponse can be co ela ed
wi h he de ec ion o highe le els o P
e A
ansc ip ionin he
p esence o ARA (SI Appendix,Fig.S4C).
In he backg ound o he hilA-g p
SF
usion (which dele es
he igh wo- hi ds o SPI -1), he dele ion gene a ed by he
e R-P
e A
inse ion cons i u es a minimal sys em wi h only wo
SPI-1 genes, hilD and pphB, emaining in ac . Significan ly,
his s ain s ill exhibi s he HilD-dependen bis able pheno ype,
sugges ing ha HilD is no only equi ed bu also su ficien o
bis abili y (Fig. 4Aand B). G ow h in a medium supplemen ed
wi h AHTc a ec s he basal hilA
ON
/hilA
OFF
a io only ma gin-
ally unless ARA is also p esen , in which case he as majo i y
o he cell popula ion swi ches o he hilA
ON
s a us (Fig. 4A
and Band SI Appendix, Fig. S5). To confi m ha he ARA
Fig. 2. NusG deple ion induces HilD-dependen ac i a ion o hilD and p gH p omo e s. (Aand B) Quan ifica ion o hilD mRNA (A) and p gH mRNA (B) om
s ains MA14302 (hilD
+
) and MA14561 (hilDΔ309) g own o ea ly s a iona y phase in he absence o in he p esence o 0.1% ARA. RNA was quan ified by
wo-s ep RT-qPCR. C alues we e no malized o he C alues de e mined o ompA mRNA. T ansc ip le els a e shown ela i e o hose o un ea ed
MA14302, se as 1. (C) Measu emen o HilD p o ein binding o he hilD p omo e . HilD-bound DNA was isola ed by ChIP om s ain MA14363 (ca ying a
ch omosomal hilD-3xFLAG usion) and quan ified by eal- ime PCR (ChIP-qPCR). C alues we e no malized o he alues o a ka E gene e e ence. Resul s
a e p esen ed as a ios be ween he alues measu ed in cells g own in ARA-supplemen ed medium and he alues om un ea ed cells. (Dand E) The 50
RACE-Seq analysis o hilD and p gH p omo e ac i i y, espec i ely. RNA om s ain MA14302 was e e se- ansc ibed in he p esence o a TSO. The esul ing
cDNA was used as empla e o semiquan i a i e PCR wi h p ime s ca ying Illumina adap e sequences a hei 50ends. Amplified DNA was subjec ed o
high- h oughpu sequencing. Read coun s we e no malized o hose measu ed a he ompA p omo e . Resul s shown ep esen he a ios be ween he no -
malized coun s om ARA- ea ed cells and hose om un ea ed cells. (Fand G) Con ibu ion o dis al ansc ip ion o hilD and p gH RNA le els, espec i ely.
RNA was e e se- ansc ibed wi h p ime s annealing inside he p omo e -p oximal po ion o hilD o p gH (p edic ed ansc ip s a e depic ed as wa y lines).
The esul ing cDNAs (s aigh lines) we e used o qPCR amplifica ion wi h p ime s annealing close o (qPCR-1) o a he away om (qPCR-2) he RT p iming
si e. C alues we e no malized o he C alues de e mined o ompA mRNA. T ansc ip le els a e shown ela i e o hose o un ea ed MA14302 cells, se
as 1. All he da a in his figu e o igina e om h ee o mo e independen expe imen s (wi h e o ba s indica ing SDs). S a is ical significance was de e -
mined by unpai ed wo- ailed S uden es s wi h Welch’s co ec ion o unequal a iances (*P≤0.05; **P≤0.01; ***P≤0.001). In Fand G, he calcula ed
P alues o he di e ences be ween un ea ed samples (g een ba s) we e 0.0002 (F) and <0.0001 (G). The P alues o he ARA- ea ed samples ( ed ba s)
we e 0.0108 (F) and 0.0025 (G). The oligonucleo ides used as p ime s in he abo e expe imen s a e lis ed in SI Appendix, Table S4. Fu he expe imen al
de ails a e p o ided in Ma e ials and Me hods.
4o 9 h ps://doi.o g/10.1073/pnas.2203011119 pnas.o g
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
e ec s depend on ansc ip ion o igina ing ups eam o hilD,
and no by an al e na i e, uniden ified mechanism s imula ing
hilD exp ession when NusG is deple ed, we cons uc ed a s ain
ca ying he s ong Rho-independen ansc ip ion e mina o
om he his idine ope on a enua o egion (Te m
hisL
) imme-
dia ely downs eam om P
e A
in he hilA::g p
SF
ΔK28 back-
g ound (SI Appendix, Fig. S6Aand B). Du ing cons uc ion, a
clone displaying s ong g een fluo escence on a pla e supple-
men ed wi h ARA and AHTc was iden ified. Sequence analysis
e ealed ha his isola e ha bo s a dele ion emo ing six ou o
he nine epea ed Us a he 30end o Te m
hisL
. Bo h he s ain
wi h he wild- ype Te m
hisL
inse and he Δ6U de i a i e we e
used o bis abili y assays. Resul s showed ha Te m
hisL
abol-
ishes all e ec s o ARA on hilA OFF/ON a ios bo h in he
absence and in he p esence o AHTc (SI Appendix, Fig. S6C
and E). In e es ingly, he Δ6U dele ion e e ses his pa e n,
causing abou hal o he cell popula ion o swi ch o he hilA
ON s a us in he p esence o ARA alone and i ually he en i e
popula ion o swi ch ON in he p esence o ARA and AHTc
combined (SI Appendix, Fig. S6Dand F). These esul s p o ide
conclusi e e idence ha ansc ip ion o igina ing mo e han
600 bp ups eam o hilD’s p ima y TSS is solely esponsible o
he e ec s o NusG deple ion on hilD exp ession. Since hilD
seconda y p omo e s a e all loca ed downs eam om Te m
hisL
,
hese da a also suppo he idea ha hey play no di ec ole in
he ARA-induced ac i a ion o he p ima y p omo e .
Viewing SPI-1 Up-Regula ion a he Ch oma in Le el. In pa al-
lel wi h he abo e s udies, we sough o de e mine whe he NusG
deple ion a ec ed he binding o H-NS o SPI-1 and o he geno-
mic islands. Fo his pu pose, ChIP coupled wi h high- h oughpu
sequencing (ChIP-Seq) was pe o med in s ains ca ying he
NusG- ep essible allele and an epi ope- agged e sion o H-NS.
Examina ion o he ChIP-Seq p ofiles in he SPI-1 sec ion o he
genome showed a succession o peaks and alleys consis en wi h
he p esence o mul iple con iguous pa ches o oligome ized
H-NS sepa a ed by segmen s wi h li le o no H-NS bound (Fig.
5A). Supe imposing he p ofiles om cells g owing in he absence
o in he p esence o ARA e eals small bu none heless app ecia-
ble di e ences in he le els o DNA agmen s bound by H-NS.
One can see ha a numbe o peaks sh ink as a esul o he ARA
ea men (Fig. 5A). In e es ingly, he mos conspicuous changes
a e de ec ed in he hilD-hilA and in F-in H sec ions o SPI-1,
co esponding o he loca ions o main egula o y hubs (30). In
con as , no changes a e obse ed a he a - igh end o SPI-1
[pigA-pphB segmen (53)] o in he cen al po ion o he island.
Fig. 3. O e lapping ansc ip ion igge s HilD-dependen ac i a ion o hilD and p gH p omo e s in NusG-deple ed cells. (A) Schema ic diag am showing
he gene o ganiza ion on he 50side o hilD in s ains MA14358 (hilD
+
) and MA14569 (hilDΔ309). Bo h s ains con ain a e R-P
e A
casse e ha eplaces a
10,828-bp segmen o SPI-1 and places he P
e A
p omo e 610 bp ups eam o he main hilD TSS. (Band C) Quan ifica ion o hilD mRNA (B) and p gH mRNA
(C) in cells g own o ea ly s a iona y phase in he p esence o absence o ei he ARA o AHTc, o in he p esence o bo h. RNA was quan ified by wo-s ep
RT-qPCR. C alues we e no malized o he C alues de e mined o ompA mRNA. T ansc ip le els a e shown ela i e o hose o un ea ed MA14358, se
as 1. (Dand E) The 50RACE-Seq analysis o hilD and p gH p omo e ac i i y, espec i ely. RNA om s ain MA14358 g own unde he di e en condi ions was
p ocessed as desc ibed in he legend o Fig. 2 Dand E. Resul s shown ep esen he a ios be ween he no malized ead coun s om ea ed cells and hose
om un ea ed cells. (F) Measu emen o HilD p o ein binding o he hilD p omo e . HilD-bound DNA was isola ed by ChIP om s ain MA14505 (ca ying a
ch omosomal hilD-3xFLAG usion) and quan ified by eal- ime PCR (ChIP-qPCR). C alues we e no malized o he alues o a ka E gene e e ence. Resul s
a e p esen ed as a ios be ween he alues measu ed in cells g own in a medium supplemen ed wi h ARA and AHTc and hose om un ea ed cells. All he
da a in his figu e o igina e om h ee o mo e independen expe imen s (wi h e o ba s indica ing SDs). S a is ical significance was de e mined by
unpai ed wo- ailed S uden ’s es s wi h Welch’s co ec ion o unequal a iances (ns, P>0.05; *P≤0.05; **P≤0.01; ***P≤0.001). The oligonucleo ides
used as p ime s in he abo e expe imen s a e lis ed in SI Appendix, Table S4.
PNAS 2022 Vol. 119 No. 30 e2203011119 h ps://doi.o g/10.1073/pnas.2203011119 5o 9
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
Read dep h quan ifica ion confi med he p ofile changes. ARA
ea men lowe s H-NS binding in he hilD-hilA and in F-in H
in e als by 34% and 47%, espec i ely, while ha ing no e ec
in he pigA-pphB egion (Fig. 5B). Likewise, no app eciable di e -
ences a e obse ed a he p oV locus (54) (Fig. 5B). ChIP-Seq
analysis was also pe o med in he s ain ca ying he e R-P
e A
-
and hilA::g p
SF
ΔK28-associa ed dele ions, compa ing unchallenged
cells o cells g own in he p esence o bo h ARA and AHTc (Fig.
5C). Somewha su p isingly, he double ea men caused only
a 20% educ ion o H-NS binding in he hilD-hilA in e al (Fig.
5D). In bo h abo e analyses, isual inspec ion o he p ofiles
a ound o he H-NS–bound loci known o be up- egula ed in
NusG-deple ed cells (16) ailed o e eal app eciable di e ences.
Finding ha ansc ip ional changes p oduce compa a i ely small
o unde ec able al e a ions in H-NS binding is no no el (24, 26)
and sugges s ha H-NS–DNA complexes exis dynamically and ap-
idly e o m a e he passage o ansc ip ion elonga ion complexes.
Discussion
This s udy was aimed a unde s anding why impai ing Rho-
dependen ansc ip ion e mina ion by deple ion o Rho-
co ac o NusG elie es H-NS silencing o SPIs. We show ha
NusG deple ion igge s a posi i e eedback loop ha gene a es
and main ains HilD, he mas e egula o o he Salmonella i -
ulence egula o y cascade. Accumula ion o HilD is p ima ily
esponsible o H-NS coun e silencing in NusG-deple ed cells.
This is di ec ly demons a ed o SPI-1, SPI-2, and SPI-5
Fig. 4. O e lapping ansc ip ion p omo es HilD-dependen bis abili y o SPI-1 exp ession. S ains MA14358 (hilD
+
) and MA14569 (hilDΔ309) ca y he
e R-P
e A
casse e (Fig. 3A) combined wi h hilA::g p
SF
ΔK28 and he ch omosomal P
Tac
-mChe y in he ARA-inducible-NusG deple ion backg ound. Cells g own
o ea ly s a iona y phase unde he indica ed condi ions we e used o single-cell analysis by flow cy ome y. GFP fluo escence was measu ed and he dis i-
bu ion o hilA
OFF
and hilA
ON
cells is shown in hea maps. (A) MA14358. (B) MA14569.
Fig. 5. NusG deple ion a ec s H-NS binding o specific po ions o SPI-1. (A) Rep esen a i e ChIP-Seq p ofiles om NusG-deple able s ain MA13748 (hns-
3xFLAG) g own o ea ly s a iona y phase in he p esence o absence o 0.1% ARA. (B) Read dep h quan ifica ion in he sec ions amed by he dashed ec an-
gles in Aand in he n dF-p oV in e genic egion. Read dep h alues (de e mined by he bedco ool o he Sam ools sui e) we e no malized o he alues
om he en i e genome. The esul s shown ep esen he a ios be ween he no malized alues om IP samples and hose om inpu DNA. (C) Rep esen a-
i e ChIP-Seq p ofiles om NusG-deple able s ain MA14513 (Δ[si A-p gH]:: e R-P
e A
,hilA::g p
SF
ΔK28, hns-3xFLAG) g own o ea ly s a iona y phase in he
p esence o absence o ARA +AHTc. (D) Read dep h quan ifica ion in he in e als amed by he dashed ec angles in C. Read dep h was calcula ed and
no malized as in B. The da a in Band C ep esen he means om h ee independen ChIP-Seq expe imen s (wi h e o ba s indica ing SDs).
6o 9 h ps://doi.o g/10.1073/pnas.2203011119 pnas.o g
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
genes, bu i seems likely ha he HilD in ol emen may ex end
o mos , i no all, islands and isle s up- egula ed in NusG-
deple ed cells (16). No e ha al hough SPI-1 and SPI-2 a e
gene ally ac i a ed in esponse o sha ply di e en cues, HilD-
media ed c oss- alk allows exp ession o SPI-2 genes unde condi-
ions unusual o his island, no ably in ich medium (55, 56).
By oligome izing along he DNA, H-NS silences no only
bona fide p omo e s a he 50end o genes bu also a ple ho a o
spu ious in agenic p omo e s ha “in es ”A/T- ich ho izon ally
acqui ed DNA (5–7). Finding ha he inhibi ion o Rho o
NusG causes widesp ead sense and an isense ansc ip ion o
H-NS-silenced genes (1, 16) sugges s ha H-NS–bound DNA is
suscep ible o ansc ip ional in asion and ha Rho ( ec ui ed by
NusG) ac s o p e en elonga ion o in ading ansc ip ion com-
plexes. Va ious lines o e idence sugges ha H-NS–bound
egions a e no o ally impe meable o RNA polyme ase. Exis-
ence o se e al e y sho ansc ip s ini ia ing om wi hin
H-NS–associa ed loci was p e iously in e ed om a genome-
wide analysis o TSSs in E. coli (4). Mo e ecen ly, pa allel ChIP-
Seq TSS mapping expe imen s showed a clea TSS being used
ups eam o he E. coli ydbCD ope on, e en when H-NS was
p esen , bu no ull-leng h mRNA (6). Finally, in ou p e ious
wo k, we ound ha a e R-P
e A
casse e placed only 57 bp away
om he H-NS nuclea ion si e in he leuO p omo e egion o
Salmonella no mally esponds o AHTc induc ion (al hough lead-
ing o LeuO syn hesis only when NusG is deple ed) (16). Elon-
ga ing h ough a pa ch o oligome ized H-NS, RNA polyme ase
can dislodge H-NS and allow o he RNA polyme ase molecules
o gain access o no mally silenced p omo e s, hus u he con-
ibu ing o ansc ip ional noise (57, 58). The da a p esen ed
he e show ha ansc ip ional “noise”can be con e ed in o a
ue egula o y “melody”i he ac i a ed p omo e di ec s he
syn hesis o a posi i e au o egula o . In he model schema ized in
Fig. 6, we posi ha a ansc ip ion complex o med a a spu ious
p omo e (“Px”), i no s opped by Rho, “unzips” he H-NS
nucleop o ein filamen in he hilD p omo e egion, igge ing a
posi i e eedback loop ha esul s in HilD accumula ion and
concomi an de ep ession o bo h hilD and p gH.An isense an-
sc ip ion om inside hilD (no shown o simplici y in Fig. 6)
may con ibu e o des abiliza ion o he H-NS–DNA complex.
No e ha ansc ip ion may no need o a el all he way o he
p omo e sequence in o de o cause H-NS dissocia ion. Due o
he mul icon ac na u e o he H-NS–DNA in e ac ion (54),
dis up ion o con ac s a he edge o he oligome iza ion pa ch
could be su ficien o des abilize he en i e egion. By linking
hilD ac i a ion o a s ochas ic and likely in equen ansc ip ion
e en —i.e., ini ia ion a a spu ious p omo e o ead h ough o a
Rho-dependen e mina o by a spu ious ansc ip — he model
can explain he bis abili y in he exp ession o HilD- egula ed
loci and sugges s ha he equency o hese e en s may se SPI-1
ON/OFF subpopula ion a ios du ing no mal g ow h. The
HilD/H-NS in e play in he egula ion o SPI-1 bea s analogy
wi h he mechanism egula ing he exp ession o he locus o
en e ocy es e acemen (LEE) o en e opa hogenic E. coli.He e
oo, exp ession is cha ac e ized by a bis able esponse (59), sug-
ges ing ha he in e play be ween H-NS and egula o y p o eins
(Le in his case) may cons i u e an elemen al p emise o bis abil-
i y. Whe he LEE egula ion esponds o pe asi e ansc ip ion
is cu en ly unknown.
Euka yo ic genomes, including he p edominan noncoding
ac ion o human genomes, a e pe asi ely ansc ibed and his
p ocess s ongly impac s gene egula ion and ch oma in s uc u e
(60, 61). In p oka yo es, a ious po en ial oles o pe asi e an i-
sense ansc ip ion in gene egula ion and genome e olu ion we e
conside ed (62), bu o da e such oles ha e emained hypo he i-
cal. Da a p esen ed he e show ha he elonga ion o pe asi e
ansc ip s in o H-NS–DNA complexes can ac as a coun e
silencing mechanism modula ing a egula o y esponse. Al hough
mos o hee ec swe eobse edunde condi ionso impai ed
ansc ip ion e mina ion, low-le el ead h ough ansc ip s we e
de ec ed in unchallenged cells, sugges ing ha hei e ec s (e.g.,
bis abili y) a e exe ed du ing no mal g ow h. This s udy adds
elonga ion o pe asi e ansc ip s o he se o mechanisms ha
p oduce ansc ip ional noise (63) and p o ides a model o
unde s and he molecula basis o SPI-1 bis abili y, which has
emained a long-s anding mys e y in Salmonella biology. The
model fi s well in he iew ha s ochas ic cell- o-cell di e ences
pe pe ua ed by eedback loops can gene a e pheno ypic lineages
(64, 65).
Ma e ials and Me hods
S ains and Cul u e Condi ions. All s ains used in his wo k a e de i ed
om S. en e ica se o a Typhimu ium s ain LT2 (66). S ains and hei geno ypes
a e lis ed in SI Appendix,TableS1. Bac e ia we e ou inely cul u ed in lysogeny
b o h (LB: yp one 10 g/L, yeas ex ac 5 g/L, NaCl, 5 g/L) a 37°C o , occasion-
ally, a 30 °C when ca ying empe a u e-sensi i e plasmid eplicons. Typically,
bac e ia we e g own o e nigh in s a ic 2-mL cul u es (14-mm-diame e ubes),
subcul u ed by 1:200 dilu ion he nex day (20 mL cul u e in 125 mL E len-
meye flasks) and g own wi h 170 pm shaking. Fo g ow h on pla es, LB was
solidified by he addi ion o 1.5% Di co aga . When needed, an ibio ics (Sigma-
Ald ich) we e included in g ow h media a he ollowing final concen a ions:
chlo amphenicol, 10 μg/mL; kanamycin monosul a e, 50 μg/mL; sodium ampi-
cillin, 100 μg/mL; spec inomycin dihyd ochlo ide, 80 μg/mL; e acycline hyd o-
chlo ide, 25 μg/mL S ains we e cons uc ed by gene alized ansduc ion using
he high- equency ansducing mu an o phage P22, HT 105/1 in -201 (67) o
by he λ- ed ecombinee ing echnique implemen ed p e iously (68); 3xFLAG
Fig. 6. Model o ac i a ion o hilD and p gH p omo e s by o e lapping
ansc ip ion. (A) A spu ious ansc ip ion ini ia ion e en occu s a he
edge o a pa ch o oligome ized H-NS (o ange ci cles). T ansc ip elonga ion
h ough bound H-NS is p e en ed by NusG-media ed ec ui men o Rho
ac o (s op sign). (B) Occasionally, he ansc ip eludes Rho e mina ion
and p og esses along he DNA dislodging H-NS in on o i s pa h. This
ac ion opens a kine ic window du ing which RNA polyme ase (g een o als)
can bind o p omo e s ha become exposed, including hilD seconda y p o-
mo e s (no shown) and he p ima y hilD p omo e . (C) Ac i a ion o he
hilD p omo e leads o an inc ease in he le els o HilD p o ein (blue double
o als), which, upon binding o he hilD egula o y egion, u he s imula es
hilD ansc ip ion and p o ein p oduc ion. (D) This locks he sys em in a
posi i e eedback loop: Accumula ion o HilD leads o mo e hilD ansc ip-
ion and mo e HilD p o ein made. Th ough HilA (no shown) i also esul s
in high-le el ansc ip ion o he p gH gene. Di e gen ansc ip ion u he
enhances he accessibili y o addi ional spu ious p omo e sequences, u -
he con ibu ing o unaway ansc ip ion ac i a ion.
PNAS 2022 Vol. 119 No. 30 e2203011119 h ps://doi.o g/10.1073/pnas.2203011119 7o 9
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
epi ope usions we e cons uc ed as desc ibed (69) o by wo-s ep sca less ecom-
binee ing. The la e p ocedu e in ol ed he use o ipa i e selec able coun e
selec able casse es (condi ionally exp essing he ccdB oxin gene) amplified
om in-house-de eloped plasmid empla es. Oligonucleo ide used as p ime s
o amplifica ion (ob ained om Sigma-Ald ich o Eu ofins) a e lis ed in SI
Appendix, Table S2. Thei asso men o he cons uc ion o he ele an alleles
used in his s udy is shown in SI Appendix, Table S3. PCR-amplified agmen s
o be used o ecombinee ing we e p oduced wi h high-fideli y Phusion poly-
me ase (New England Biolabs). Cons uc s we e e ified by colony-PCR using Taq
polyme ase ollowed by DNA sequencing (pe o med by Eu ofins-GATC Bio ech).
Fluo escence Mic oscopy. Bac e ial cul u es g own o e nigh in LB a 37 °C
we e dilu ed 1:200 in o 2 mL o he same medium wi h o wi hou 0.1% ARA
and/o 0.4 μg/mL AHTc (in 14-mm-diame e ubes) and g own o 4 h a 37 °C
wi h shaking (170 pm). Cells we e hen ha es ed by cen i uga ion (2 min a
12,000 ×g), washed once in phospha e-bu e ed saline (PBS), and used imme-
dia ely o mic oscopic examina ion. Images we e cap u ed wi h a Leica DM
6000 B mic oscope (CTR 6500 d i e con ol uni ) equipped wi h a EBQ 100
lamp powe uni and fil e s o phase con as , GFP, and mChe y de ec ion
(100×oil imme sion objec i e). Pic u es we e aken wi h a Hamama su C11440
digi al came a and p ocessed wi h Me amo ph so wa e.
Flow Cy ome y. Flow cy ome y was used o moni o exp ession o ansla-
ional GFP usions. Da a acquisi ion was pe o med using a Cy omics FC500-MPL
cy ome e (Beckman Coul e ) and da a we e analyzed wi h FlowJo X e sion
10.0.7 so wa e (T ee S a , Inc.). S. en e ica cul u es we e washed and esus-
pended in PBS o fluo escence measu emen . Fluo escence alues o 100,000
e en s we e compa ed wi h he da a om he epo e less con ol s ain, hus
yielding he ac ion o ON and OFF cells.
RNA Ex ac ion and Quan i ica ion by RT-qPCR. O e nigh bac e ial cul-
u es in LB we e dilu ed 1:200 in he same medium—o in LB supplemen ed
wi h 0.1% ARA o 0.4 μg/mL AHTc, o bo h d ugs whe e app op ia e—and g own
wi h shaking a 37 °C o op ical densi y a 600 nm (OD
600
)=0.7 o 0.8. Cul-
u es (4 mL) we e apidly spun down and esuspended in 0.6 mL ice-cold REB
bu e (20 mM sodium ace a e, pH 5.0, and 10% suc ose). RNA was pu ified by
sequen ial ex ac ion wi h ho acid phenol, phenol-chlo o o m 1:1 mix u e and
chlo o o m. Following o e nigh e hanol p ecipi a ion a 20 °Candcen i uga-
ion, he RNA pelle was esuspended in 20 μLo H
2
O. Th ee samples we e p e-
pa ed om independen biological eplica es o each s ain and condi ion. RNA
yields, measu ed by Nanod op eading, ypically anged be ween 2 and 3 μg/μL.
The RNA p epa a ions we e used o fi s -s and DNA syn hesis wi h he New
England Biolabs P o oSc ip II Fi s S and DNA syn hesis ki , ollowing he manu-
ac u e ’sspecifica ions. B iefly, RNA (1 μg) was combined wi h 2 μL o a mix u e
o wo p ime s (5 μM each), one annealing in he p omo e p oximal po ion o
he RNA o be quan ified(p ime AI41 o hilD o p ime AI48 o p gH), he o he
annealing o a simila posi ion in he e e ence RNA (p ime AJ33 o ompA)in
an 8 μLfinal olume. A e 5 min a 65 °C and a quick cooling s ep on ice, ol-
umes we e b ough o 20 μLby headdi iono 10μL o P o oSc ip II Reac ion
Mix (2×)and2μL o P o oSc ip II Enzyme Mix (10×). Mixes we e incuba ed o
one hou a 42 °C ollowed by a 5-min enzyme inac i a ion s ep a 80 °C.
Samples we e hen used o eal- ime qPCR as desc ibed in SI Appendix,
Supplemen a y Ma e ials and Me hods.
50RACE-Seq Analysis. RNA 50-end analysis was ca ied ou by empla e-
swi ching e e se ansc ip ion (51) coupled o PCR. Ini ially, we applied his
echnique on RNA p e ea ed wi h Vaccinia i us capping enzyme as epo ed
p e iously (70). Howe e , hese ini ial es s indica ed ha he capping s ep is
unnecessa y; he e o e, his s ep was subsequen ly omi ed. F om ha poin on,
we ollowed he p o ocol desc ibed by he Templa e Swi ching RT Enzyme
Mix p o ide (New England Biolabs) wi h a ew modifica ions (SI Appendix,
Supplemen a y Ma e ials and Me hods). The syn hesized cDNA was amplified by
PCR wi h p ime s ca ying Illumina adap e s a hei 50ends. Se e al PCRs we e
ca ied ou in pa allel wi h a common o wa d p ime (AJ38, annealing o he
TSO) and a e e se p ime specific o he egion being analyzed and ca ying a
ea men -specific index sequence (see example in SI Appendix,Fig.S3). Reac-
ions we e se up acco ding o New England Biolabs PCR p o ocol o Q5 Ho
S a High-Fideli y DNA polyme ase in a final olume o 50 μL(using1μLo he
abo e cDNA p epa a ion pe eac ion). The numbe o amplifica ion cycles
needed o ep oducible semiquan i a i e measu emen s, de e mined in ial
expe imen s, was chosen o be 25 o he ompA e e ence, 30 o he p ima y
hilD and p gH p omo e s, and 35 o he seconda y hilD p omo e s and he P
e A
p omo e . The PCR p og am was as ollows: ac i a ion: 98°C o 30 s; amplifica-
ion(25o 30o 35cycles):98°C o 10 s; 65 °C o 15 s; 72 °C o 30s;final
s age: 72 °C o 5 min. P oduc s om pa allel PCRs we e mixed in equal ol-
umes; mixes o igina ing om he amplifica ion o sepa a e egions we e pooled
and he pools subjec ed o high- h oughpu sequencing. The p ocedu e was
implemen ed a leas once, occasionally wice, wi h each o he independen
RNA p epa a ions. The coun s o eads con aining he TSO sequence posi ioned
a he TSSs analyzed he e, each no malized o he coun s o eads con aining he
TSO posi ioned a he ompA TSS, we e used o calcula e he a ios be ween he
ac i i y o a p omo e unde a gi en ea men ela i e and i s ac i i y in
un ea ed cells. The aw da a om RACE-Seq expe imen s we e deposi ed in o
A ayExp ess unde he accession numbe E-MTAB-11419.
ChIP-Seq Analysis. O e nigh bac e ial cul u es we e dilu ed 1:100 in LB o in
LB supplemen ed wi h 0.1% ARA o 0.1% ARA +0.4 μg/mL AHTc and g own a
37 °C oanOD
600
o 0.7 o 0.8. A his poin 1.6 mL o 37% o maldehyde (Al a
Aesa ) we e added o 30 mL o cul u e and he cul u e was incuba ed o 30 min
a oom empe a u e wi h gen le agi a ion. This was ollowed by he addi ion o
6.8 mL o a 2.5 M glycine solu ion and u he 15-min incuba ion wi h gen le
agi a ion a oom empe a u e. Cells we e cen i uged and he pelle esus-
pended in 24 mL o TBS bu e (50 mM T isHCl, pH 7.4, and 150 mM NaCl).
These s eps we e epea ed once and he cells cen i uged again. Cells we e hen
p ocessed o ChIP as p e iously desc ibed (71) and adap ed he e o Salmonella
(see SI Appendix, Supplemen a y Ma e ials and Me hods). The aw da a om all
ChIP-Seq expe imen s we e deposi ed in o A ayExp ess unde he accession
numbe E-MTAB-11386.
S a is ics, Rep oducibili y, and Bioin o ma ic Analyses. See SI Appendix,
Supplemen a y Ma e ials and Me hods.
Da a A ailabili y. ChIP-Seq da a (72) and RACE-Seq da a (73) ha e been
deposi ed in A ayExp ess.
ACKNOWLEDGMENTS. We a e g a e ul o Vicky Lioy o ad ice on ChIP
expe imen s, o Lau en Ku as o ad ice on RT-qPCR expe imen s and o
Modes o Ca ballo, Lau a Na a o, and C is ina Reyes (Se icio de Biolog
ıa, CIT-
IUS, Uni e sidad de Se illa) o help wi h he flow cy ome y analysis. We
hank he High- h oughpu Sequencing Co e Facili y o he I2BC (Gi -su -
Y e e) o lib a y p epa a ion and sequencing (ChIP-Seq and RACE-Seq) and
he ICGex NGS pla o m o he Cu ie Ins i u e (Pa is) o gene a ing some o
he sequence da ase s (RACE-Seq). This s udy was suppo ed by he Cen e
Na ional de la Reche che Scien ifique, by he Agence Na ionale de la Reche che
(ANR-15-CE11-0024-03), and by g an PID2020-116995RB-I00 om he Minis-
e io de Ciencia e Inno aci
on o Spain - Agencia Es a al de In es iga
ıon and he
Eu opean Regional Fund).
Au ho a filia ions:
a
Uni e si
e Pa is-Saclay, CEA, CNRS, Ins i u de Biologie In
eg a i e de
la Cellule (I2BC), 91190 Gi -su -Y e e, F ance;
b
Depa amen o de Mic obiolog
ıayPa asi olog
ıa,
Facul ad de Fa macia, Uni e sidad de Se illa, 41012 Se illa, Spain; and
c
Depa amen o de
Gen
e ica, Facul ad de Biolog
ıa, Uni e sidad de Se illa, 41012 Se illa, Spain
1. J. M. Pe e s e al., Rho and NusG supp ess pe asi e an isense ansc ip ion in Esche ichia coli.
Genes De . 26, 2621–2633 (2012).
2. J. M. Pe e s e al., Rho di ec s widesp ead e mina ion o in agenic and s able RNA ansc ip ion.
P oc. Na l. Acad. Sci. U.S.A. 106, 15406–15411 (2009).
3. L. E wille , J. Buswell, E. Yigi , I. Schildk au , A no el en ichmen s a egy e eals unp eceden ed
numbe o no el ansc ip ion s a si es a single base esolu ion in a model p oka yo e and he
gu mic obiome. BMC Genomics 17, 199 (2016).
4. V. V. Panyuko , O. N. Ozoline, P omo e s o Esche ichia coli e sus p omo e islands: Func ion and
s uc u e compa ison. PLoS One 8, e62601 (2013).
5. S. S. Singh e al., Widesp ead supp ession o in agenic ansc ip ion ini ia ion by H-NS. Genes
De . 28, 214–219 (2014).
6. D. Fo es , E. A. Wa man, A. M. E kelens, R. T. Dame, D. C. G ainge , Xenogeneic silencing
s a egies in bac e ia a e dic a ed by RNA polyme ase p omiscui y. Na . Commun. 13, 1149
(2022).
8o 9 h ps://doi.o g/10.1073/pnas.2203011119 pnas.o g
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
7. L. E. Lambe e e al., Ho izon ally acqui ed AT- ich genes in Esche ichia coli cause oxici y by
seques e ing RNA polyme ase. Na . Mic obiol. 2, 16249 (2017).
8. R. Landick, J. T. Wade, D. C. G ainge , H-NS and RNA polyme ase: A lo e-ha e ela ionship?
Cu . Opin. Mic obiol. 24,53–59 (2015).
9. C. J. Do man, H-NS: A uni e sal egula o o a dynamic genome. Na . Re . Mic obiol. 2, 391–400
(2004).
10. C. J. Do man, J. C. Hin on, A. F ee, Domain o ganiza ion and oligome iza ion among H-NS-like
nucleoid-associa ed p o eins in bac e ia. T ends Mic obiol. 7, 124–128 (1999).
11. F. C. Fang, S. Rimsky, New insigh s in o ansc ip ional egula ion by H-NS. Cu . Opin. Mic obiol.
11, 113–120 (2008).
12. S. T. A old, P. G. Leona d, G. N. Pa kinson, J. E. Ladbu y, H-NS o ms a supe helical p o ein sca old
o DNA condensa ion. P oc. Na l. Acad. Sci. U.S.A. 107, 15728–15732 (2010).
13. C. Ueguchi, T. Mizuno, The Esche ichia coli nucleoid p o ein H-NS unc ions di ec ly as a
ansc ip ional ep esso . EMBO J. 12, 1039–1046 (1993).
14. R. T. Dame, C. Wyman, R. Wu m, R. Wagne , N. Goosen, S uc u al basis o H-NS-media ed
apping o RNA polyme ase in he open ini ia ion complex a he nB P1. J. Biol. Chem. 277,
2146–2150 (2002).
15. M. Shin e al., DNA looping-media ed ep ession by his one-like p o ein H-NS: Specific
equi emen o Esigma70 as a co ac o o looping. Genes De . 19, 2388–2398 (2005).
16. L. Bossi e al., NusG p e en s ansc ip ional in asion o H-NS-silenced genes. PLoS Gene . 15,
e1008425 (2019).
17. M. V. Ko lajich e al., B idged filamen s o his one-like nucleoid s uc u ing p o ein pause RNA
polyme ase and aid e mina ion in bac e ia. eLi e 4, e04970 (2015).
18. S. Lucchini e al., H-NS media es he silencing o la e ally acqui ed genes in bac e ia. PLoS Pa hog.
2, e81 (2006).
19. W. W. Na a e e al., Selec i e silencing o o eign DNA wi h low GC con en by he H-NS p o ein in
Salmonella.Science 313, 236–238 (2006).
20. S. S. Ali, B. Xia, J. Liu, W. W. Na a e, Silencing o o eign DNA in bac e ia. Cu . Opin. Mic obiol.
15, 175–181 (2012).
21. C. J. Do man, H-NS, he genome sen inel. Na . Re . Mic obiol. 5, 157–161 (2007).
22. S. S. Ali e al., Silencing by H-NS po en ia ed he e olu ion o Salmonella.PLoS Pa hog. 10,
e1004500 (2014).
23. K. Higashi e al., H-NS acili a es sequence di e sifica ion o ho izon ally ans e ed DNAs du ing
hei in eg a ion in hos ch omosomes. PLoS Gene . 12, e1005796 (2016).
24. J. C. Pe ez, T. La ifi, E. A. G oisman, O e coming H-NS-media ed ansc ip ional silencing o
ho izon ally acqui ed genes by he PhoP and SlyA p o eins in Salmonella en e ica.J. Biol. Chem.
283, 10773–10783 (2008).
25. D. M. S oebel, A. F ee, C. J. Do man, An i-silencing: O e coming H-NS-media ed ep ession
o ansc ip ion in G am-nega i e en e ic bac e ia. Mic obiology (Reading) 154, 2533–2545
(2008).
26. W. R. Will, D. H. Bale, P. J. Reid, S. J. Libby, F. C. Fang, E olu iona y expansion o a egula o y
ne wo k by coun e -silencing. Na . Commun. 5, 5270 (2014).
27. W. R. Will, W. W. Na a e, F. C. Fang, In eg a ed ci cui s: How ansc ip ional silencing and
coun e -silencing acili a e bac e ial e olu ion. Cu . Opin. Mic obiol. 23,8–13 (2015).
28. J. R. Elle meie , J. M. Slauch, Adap a ion o he hos en i onmen : Regula ion o he SPI1 ype III
sec e ion sys em in Salmonella en e ica se o a Typhimu ium. Cu . Opin. Mic obiol. 10,24–29
(2007).
29. J. E. Gal
an, Salmonella in e ac ions wi h hos cells: Type III sec e ion a wo k. Annu. Re . Cell De .
Biol. 17,53–86 (2001).
30. C. D. Elle meie , J. R. Elle meie , J. M. Slauch, D. Hil, C. Hil, HilD, HilC and R sA cons i u e a eed
o wa d loop ha con ols exp ession o he SPI1 ype h ee sec e ion sys em egula o hilA in
Salmonella en e ica se o a Typhimu ium. Mol. Mic obiol. 57, 691–705 (2005).
31. J. L
opez-Ga ido, E. Pue a-Fe n
andez, J. Casades
us, A euka yo ic-like 30un ansla ed egion in
Salmonella en e ica hilD mRNA. Nucleic Acids Res. 42, 5894–5906 (2014).
32. C. C. Hung e al., Salmonella in asion is con olled h ough he seconda y s uc u e o he hilD
ansc ip . PLoS Pa hog. 15, e1007700 (2019).
33. D. P
e ez-Mo ales e al., An incohe en eed o wa d loop o med by Si A/Ba A, HilE and HilD is
in ol ed in con olling he g ow h cos o i ulence ac o exp ession by Salmonella Typhimu ium.
PLoS Pa hog. 17, e1009630 (2021).
34. J. R. G enz, J. E. Co Chubiz, P. Thap awa , J. M. Slauch, HilE egula es HilD by blocking DNA
binding in Salmonella en e ica Se o a Typhimu ium. J. Bac e iol. 200, e00750-17 (2018).
35. I. N. Olekhno ich, R. J. Kadne , DNA-binding ac i i ies o he HilC and HilD i ulence egula o y
p o eins o Salmonella en e ica se o a Typhimu ium. J. Bac e iol. 184, 4148–4160 (2002).
36. C. Smi h, A. M. S inge , C. Mao, M. J. Palumbo, J. T. Wade, Mapping he egula o y ne wo k o
Salmonella en e ica se o a Typhimu ium in asion. MBio 7, e01024-16 (2016).
37. M. Kala a is, J. M. Slauch, Long-dis ance e ec s o H-NS binding in he con ol o hilD exp ession
in he Salmonella SPI1 locus. J. Bac e iol. 203, e0030821 (2021).
38. M. Acke mann e al., Sel -des uc i e coope a ion media ed by pheno ypic noise. Na u e 454,
987–990 (2008).
39. M. Dia d e al., S abiliza ion o coope a i e i ulence by he exp ession o an a i ulen pheno ype.
Na u e 494, 353–356 (2013).
40. I. Hau e o , M. J. P oenc¸a, J. C. Hin on, Single-copy g een fluo escen p o ein gene usions allow
accu a e measu emen o Salmonella gene exp ession in i o and du ing in ec ion o mammalian
cells. Appl. En i on. Mic obiol. 69, 7480–7491 (2003).
41. S. Saini, J. R. Elle meie , J. M. Slauch, C. V. Rao, The ole o coupled posi i e eedback in he
exp ession o he SPI1 ype h ee sec e ion sys em in Salmonella.PLoS Pa hog. 6, e1001025 (2010).
42. M. A. S
anchez-Rome o, J. Casades
us, Con ibu ion o SPI-1 bis abili y o Salmonella en e ica
coope a i e i ulence: Insigh s om single cell analysis. Sci. Rep. 8, 14875 (2018).
43. M. C. Schlumbe ge e al., Real- ime imaging o ype III sec e ion: Salmonella SipA injec ion in o
hos cells. P oc. Na l. Acad. Sci. U.S.A. 102, 12548–12553 (2005).
44. M. A. S
anchez-Rome o, J. Casades
us, Single cell analysis o bis able exp ession o pa hogenici y
island 1 and he flagella egulon in Salmonella en e ica. Mic oo ganisms 9, 210 (2021).
45. A. S u m e al., The cos o i ulence: Re a ded g ow h o Salmonella Typhimu ium cells exp essing
ype III sec e ion sys em 1. PLoS Pa hog. 7, e1002143 (2011).
46. M. A noldini e al., Bis able exp ession o i ulence genes in Salmonella leads o he o ma ion o
an an ibio ic- ole an subpopula ion. PLoS Biol. 12, e1001928 (2014).
47. B. S eche e al.,Salmonella en e ica se o a yphimu ium exploi s inflamma ion o compe e wi h
he in es inal mic obio a. PLoS Biol. 5, 2177–2189 (2007).
48. L. Bossi, A. Schwa z, B. Guillema de , M. Boud illain, N. Figue oa-Bossi, A ole o Rho-dependen
pola i y in gene egula ion by a noncoding small RNA. Genes De . 26, 1864–1873 (2012).
49. C. P. Los oh, C. A. Lee, The HilA box and sequences ou side i de e mine he magni ude o
HilA-dependen ac i a ion o P(p gH) om Salmonella pa hogenici y island 1. J. Bac e iol. 183,
4876–4885 (2001).
50. R. A. Daly, C. P. Los oh, Gene ic analysis o he Salmonella ansc ip ion ac o HilA.
Can. J. Mic obiol. 54, 854–860 (2008).
51. M. G. Wul e al., Non- empla ed addi ion and empla e swi ching by Moloney mu ine leukemia
i us (MMLV)-based e e se ansc ip ases co-occu and compe e wi h each o he . J. Biol. Chem.
294, 18220–18231 (2019).
52. C. K €
oge e al., An in ec ion- ele an ansc ip omic compendium o Salmonella en e ica Se o a
Typhimu ium. Cell Hos Mic obe 14, 683–695 (2013).
53. N.A.Le miniaux,K.D.MacKenzie,A.D.S.Came on,Salmonella pa hogenici y island 1 (SPI-1): The
e olu ion and s abiliza ion o a co e genomic ype h ee sec e ion sys em. Mic oo ganisms 8, 576 (2020).
54. E. Bou a igues, M. Buckle, C. Badau , A. T a e s, S. Rimsky, H-NS coope a i e binding o
high-a fini y si es in a egula o y elemen esul s in ansc ip ional silencing. Na . S uc . Mol. Biol.
14, 441–448 (2007).
55. V. H. Bus aman e e al., HilD-media ed ansc ip ional c oss- alk be ween SPI-1 and SPI-2. P oc.
Na l. Acad. Sci. U.S.A. 105, 14591–14596 (2008).
56. L. C. Ma
ınez, M. M. Banda, M. Fe n
andez-Mo a, F. J. San ana, V. H. Bus aman e, HilD induces
exp ession o Salmonella pa hogenici y island 2 genes by displacing he global nega i e egula o
H-NS om ss AB. J. Bac e iol. 196, 3746–3755 (2014).
57. A. A. Ranga ajan, K. Schne z, In e e ence o ansc ip ion ac oss H-NS binding si es and ep ession
by H-NS. Mol. Mic obiol. 108, 226–239 (2018).
58. J. T. Wade, D. C. G ainge , Waking he neighbou s: Dis up ion o H-NS ep ession by o e lapping
ansc ip ion. Mol. Mic obiol. 108, 221–225 (2018).
59. H. Leh e al., Bac e ial-ch oma in s uc u al p o eins egula e he bimodal exp ession o he locus o
en e ocy e e acemen (LEE) pa hogenici y island in en e opa hogenic Esche ichia coli. MBio 8,
e00773-17 (2017).
60. J. Soude , F. S u z, Regula ion o gene exp ession and eplica ion ini ia ion by non-coding
ansc ip ion: A model based on eshaping nucleosome-deple ed egions: Influence o pe asi e
ansc ip ion on ch oma in s uc u e. BioEssays 41, e1900043 (2019).
61. K. S uhl, T ansc ip ional noise and he fideli y o ini ia ion by RNA polyme ase II. Na . S uc . Mol.
Biol. 14, 103–105 (2007).
62. J. T. Wade, D. C. G ainge , Pe asi e ansc ip ion: Illumina ing he da k ma e o bac e ial
ansc ip omes. Na . Re . Mic obiol. 12, 647–653 (2014).
63. R. Sil a-Rocha, V. de Lo enzo, Noise and obus ness in p oka yo ic egula o y ne wo ks. Annu. Re .
Mic obiol. 64, 257–275 (2010).
64. M. Kae n, T. C. Els on, W. J. Blake, J. J. Collins, S ochas ici y in gene exp ession: F om heo ies o
pheno ypes. Na . Re . Gene . 6, 451–464 (2005).
65. M. A. S
anchez-Rome o, J. Casades
us, Wadding on’s landscapes in he bac e ial wo ld. F on .
Mic obiol. 12, 685080 (2021).
66. K. Lilleengen, Typing o Salmonella yphimu ium by means o bac e iophage. Ac a Pa hol.
Mic obiol. Scand. 77,2–125 (1948).
67. H. Schmiege , Phage P22-mu an s wi h inc eased o dec eased ansduc ion abili ies. Mol. Gen.
Gene . 119,75–88 (1972).
68. K. A. Da senko, B. L. Wanne , One-s ep inac i a ion o ch omosomal genes in Esche ichia coli K-12
using PCR p oduc s. P oc. Na l. Acad. Sci. U.S.A. 97, 6640–6645 (2000).
69. S. Uzzau, N. Figue oa-Bossi, S. Rubino, L. Bossi, Epi ope agging o ch omosomal genes in
Salmonella.P oc. Na l. Acad. Sci. U.S.A. 98, 15264–15269 (2001).
70. F. Liu, K. Zheng, H. C. Chen, Z. F. Liu, Capping-RACE: A simple, accu a e, and sensi i e 50RACE
me hod o use in p oka yo es. Nucleic Acids Res. 46, e129 (2018).
71. V. S. Lioy, F. Bocca d, Con o ma ional s udies o bac e ial ch omosomes by high- h oughpu
sequencing me hods. Me hods Enzymol. 612,25–45 (2018).
72. Figue oa-Bossi e al., Da a om “Compa ing he ch omosome binding p ofiles o nucleoid
s uc u ing p o ein H-NS in he p esence o absence o condi ions elici ing he deple ion o
ansc ip ion e mina ion p o ein NusG in Salmonella.”h ps://www.ebi.ac.uk/bios udies/
a ayexp ess/s udies/E-MTAB-11386. Deposi ed 1 Feb ua y 2022.
73. Figue oa-Bossi e al., Da a om “Measu ing he e ec s o pe asi e ansc ip ion on p omo e
ac i i y in Salmonella Pa hogenici y Island 1 by semiquan i a i e 5’RACE-Seq.”h ps://www.ebi.ac.
uk/bios udies/a ayexp ess/s udies/E-MTAB-11419. Deposi ed 1 Feb ua y 2022.
PNAS 2022 Vol. 119 No. 30 e2203011119 h ps://doi.o g/10.1073/pnas.2203011119 9o 9
Downloaded om h ps://www.pnas.o g by UNIVERSIDAD DE SEVILLA BIBLIOTECA CENTRAL on May 30, 2023 om IP add ess 150.214.182.235.
