Modulation of bacterial cell size and growth rate via activation of a cell envelope stress response

Title:

odulation of

bacterial

cell size

and growth rate via activation of

a cell

envelope

stress response

Authors:

Amanda Miguel

, Matylda Zietek

Handuo Shi

Anna Sueki

, Lisa

Maier

Jolanda Verheul

, Tanneke den Blaauwen

David

Valen

Athanasios Typas

†

Kerwyn Casey Huang

†

epartment

of Bioengineering, Stanford University, Stanford, CA 94305, USA

enome Biology

Unit, EMBL Heidelberg, Meyerhofstraße 1, 69117 Heidelberg,

Germany

Department of

Microbiology and Immunology, Stanford University School of

Medicine, Stanford, CA 94305, USA

Collaboration for joint PhD degree between EMBL and Heidelberg

University,

Faculty of Biosciences, Germany

Faculty of Natural Sciences, Mathematics

and Computer Science

Swammerdam

Institute for Life Sciences

, U

niversity of Amsterdam, The Netherlands

Department of Biology, California Institute of Technology, Pasad

ena, CA 91125,

USA

Chan Zuckerberg Biohub, San Francisco, CA 94158

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Short title:

Rcs activation

reduces

cell size and growth rate

Keywords:

Rcs

hosphorelay

;

cell shape

;

FtsZ

;

IgaA; RcsF; A22;

cell division

;

morphogenesis

, growth rate

These

authors contributed equally.

†

Correspondence:

typas@embl.de

and

kchuang@stanford.edu

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Abstract

Fluctuating conditions and

diverse stresses

are typical in natural

environments.

In response, cells mount

complex responses

across multiple scales

including

adjusting their shape to withstand stress

In enterobacteria,

Rcs phosphorelay

is activated

cell envelope

damage

and by changes

periplasmic dimensions

and

cell width

Here

investigated the physiological

and morphological

consequences

Rcs activation

scherichia

coli

in the absence of stresses

, using

an inducible version of RcsF that mislocalizes to the inner membrane,

RcsF

Expression of

RcsF

immediately reduced

cellular

growth rate

and the added

length per cell cycle in a manner that was directly dependent on induction levels,

but

independent of

Rcs

induced capsule

production

At the same time, cells

increased intracellular concentration of the cell division protein

FtsZ

, and

decreased the distance between

division rings in filamentous cells.

Depletion of

the Rcs

negative

regulator IgaA

phenocopied

RcsF

duction

, indicating

that

IgaA

essen

due to

growth

inhibition

in its absence

However,

A22

treatment

did not affect

growth

rate or

FtsZ

intracellular

concentration

, despite activating

the Rcs system

These

findings

suggest that

the effect of Rcs

activation

FtsZ

levels

mediated

indirectly

through growth

rate changes

, and

highlight

feedback

among

the

Rcs

stress response,

growth dynamics

and

cell

size

control

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Introduction

The e

nvironment plays a

key

role in determining the physiological

and physical

state of a bacterial cell. In natur

al conditions

, bacteria are exposed to varied and

often stressful environments that can include changes to

pH, temperature, and

nutrient availabili

ty,

as well as the

presence

of harmful chemicals

such as

antibiotics.

The

myriad

ways that bacteria respond to

such

environmental

changes

include

modulation of growth rate, protein composition, cell envelope

modification, and cell

size and

shape

[1, 2]

. M

any of these

changes in behavior

are governed by complex signaling pathways, each with

their own set of

triggers

hese triggers can overlap; f

or example, as

Escherichia coli

cells

encounter

starvation

conditions,

they

express

the

master transcription

regulator RpoS

which

activate

programs

involved in

osmotic, oxidative,

and envelope

stress

[3,

Due to the overlap of these

stress

responses, i

remains

an open question

as to

how

many

response

pathway

specifically

alter

cell

ular

physiology

The

Rcs phosphorelay is a stress response pathway that responds to damage

the cell envelope

Gram

negative bacteria

[5, 6]

. In

coli

, deletion

of genes

the Rcs pathway

causes

sensitiv

ity

to cell shape

perturbing

drugs

such as A22

and mecillinam

[7]

suggesting that the Rcs system

may

play

yet unidentified

role

helping cells adjust to shape changes

he Rcs system

senses envelope stress

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

via

the outer membrane lipoprotein RcsF

nder normal growth conditions

, RcsF

transpo

rted t

o the outer membrane

and ultimately

surface exposed

[8, 9]

However,

when its transport to the outer membrane or surface exposure is

perturbed due to envelope stress,

RcsF

remains in the periplasm,

free to interact

with the inner membrane protein IgaA

[8, 10, 11]

, and thereby to activate the Rcs

system

[12]

When

activated

, the Rcs system

regulat

the expression of genes

control

ling many functions,

including

production

of the

exopolysaccharide

colanic acid

[13]

A previous study suggested

that the Rcs system

transcription

ally

regulat

cell division proteins,

further supporting

a connection

between

Rcs activation

and

cell shape

[14]

mutant variant of RcsF that

localizes to the inner membrane

(

RcsF

)

constitutively activate

the Rcs pathway

[8, 15]

providing

a straightforward mechanism

for induction of the Rcs system in

the absence of environmental stress

nvironment

al changes

can have varied impacts on bacterial cell shape and size

features

that

are connected

with

behavior

s such as motility, adhesion, and

immune evasion

[16]

, all of which are regulated by the Rcs pathway

[6]

The

utri

tional

content of the environment

also

strongly

affects

E. coli

and

Salmonella

enterica

cell size, with faster growing cells

having larger volume

[2, 17]

transitions between nutrient

poor and

nutrient

rich envi

ronments,

E. coli

cells

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

rapidly adjust their length and width within

hour

[18, 19]

By contrast

, during

steady

state e

xponential grow

th,

bacteria robustly maintain their cell shape

and

size

[20]

This cell

size homeostasis results from

an adder mechanism of g

rowth

[21, 22]

which cells on average add a fixed volume

∆

during each cell cycle

[23]

However, t

molecular

mechanisms by which

cells modulate their cell size

during nutrient changes

and how they are coupled to growth rate remain largely

unknown.

Bacteria

cell size and

shape is determined by th

e cell wall, a

single

macromolecule composed of glycan strands cross

linked by peptides

[24]

During

cell

elongation,

the actin homolog

MreB

controls the spatial pattern of

cell

wall

synthesis

[25, 26]

and is essential for maintaining rod

like shape in

many bacteria such as

E. coli

[25]

epletion of MreB

[27]

or depolymerization

MreB

the

small molecule

A22

results in cell rounding and eventual lysis

[28]

100

Additionally,

sub

inhibitory

concentrations of A22

cause

cell width

to increase

101

and cell length

to decrease

[29]

without affecting cell wall composition

[30]

102

During

cell

division,

cell

wall

synthesis localizes

a ring at midcell initialized

103

by the

essential

and highly conserved

tubulin

homolog FtsZ

[31, 32]

The

104

concentration of

FtsZ

is negatively

correlate

with growth rate

(and hence cell

105

size)

[33

35]

and

overexpression

results in decreased

cell length

[36]

Moreover,

106

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

FtsZ overexpression can restore viability

to cells depleted of

MreB

[27]

107

suggesting that FtsZ

levels

can impact

both

cell

shape

regulation and survival

108

109

Here, we investigate

the physiological consequences of Rcs activation in

both

110

the absence

d presence

of cell

envelope damage. We show

that RcsF

111

induction reduce

growth rate and cell length

increase

FtsZ concentration, and

112

reduce

the distance between

division rings in filamentous cells.

Depletion of

113

IgaA resulted in

similar

phenotypes

, indicating that they are general to Rcs

114

activation in the absence of envelope stresses.

In fact, cells treated with A22

115

maintained

growth rate

and

FtsZ concentration

even though

the Rcs system

was

116

activated

Thus,

FtsZ

levels

and cell length

are likely

downstream of the

growth

117

rate decreases

caused by

Rcs

activation

Altogether, our results

show that the

118

nature of an activating stress can affect the phenotypic consequences of a

119

pathway.

120

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Results

121

122

Induction

of a constitutive

ly active

RcsF

mutant

reduces growth rate and cell

123

length

124

A22 and mecillinam treatment make

E. coli

rounder, increasing cell width and

125

decreasing cell length

[29]

, and activate the Rcs pathway

[8]

. As we show in

126

accompanying study, increases in cell width are linked to Rcs activation via

127

altering the periplasmic size

[37]

However,

in such conditions it is hard to access

128

whether

Rcs activation

itself

affects cellular dimensions and growth rate.

129

ecouple the cue from the response

, we

reasoned we should control Rcs activation

130

ectopically and

the

RcsF m

utant

RcsF

, which

was

previously shown to

131

result in

constitutive

Rcs activation due to its mislocalization to the inner

132

membrane where it alleviates repression of the Rcs signaling by IgaA

[8]

133

134

To quantify growth

behavior

and

cellular dimensions during exponential growth

135

express

RcsF

in DH300 ∆

rcsF

cells from

a low

copy

plasmid

(Table S1). We

136

kept cells in steady state and

deplete

the Rcs proteins produced during stationary

137

phase

by continuously diluting RcsF

cells (Methods)

At the same time

138

added various concentrations of IPTG

and monitored

population

growth rate

, cell

139

size, and

Rcs

activation

levels

beta

galactosidase activity

the promoter of

140

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

rprA

, a small RNA expressed

when

the

Rcs

pathway is

activa

ted

[38]

Growth rate

141

decreased and

rprA

expression increased as a function of IPTG concentration, and

142

both saturated at 5

10 μM IPTG

(Fig. 1A).

Concomitantly,

cells decreased in cell

143

length (Fig. 1

Final yield in stationary phase also decreased with increasing

144

concentrations of IPTG (Fig. S1).

Thus

, ectopic

Rcs activation

in the absence of

145

envelope stress slows down

growth

decreases cell size

, and reduces total biomass

146

production

147

148

To determine the

dynamics

of Rcs

activation on

single

cells

and to avoid

149

confounding effects of cell

shape changes in optical density measurements, we

150

performed time

lapse imaging in a microfluidic flow cell (Methods) that

enabled

151

precise control

of induction

timing

via

switching to medium supplemented with

152

IPTG

. In this case,

utilized ∆

rcsF

strains

with a common

plasmid

expressi

153

wild

type

RcsF (

RcsF

) or RcsF

from an IPTG

inducible promoter (or empty

154

vector), and a second

plasmid

carrying

msfGFP

under

the

rprA

promoter

[38]

155

Cells were grown for 10 min

in LB before

15 μM IPTG

was supplemented to reach

156

full Rcs activa

tion

RcsF

(Fig. 1

)

Upon RcsF

induction, cell length

(Fig. 1

)

157

and

single

cell instantaneous

growth rate

(measured as 1/

V dV

, where

is cell

158

volume, Fig. 1

)

decreased

steadily

, while the empty

vector control

strain

159

remained largely unaffected (Fig.

D,E

The

Rcs

pathway

was fully activated in

160

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

RcsF

induced cells (Fig. 1

RcsF

induced cells exhibited a small reduction in

161

cell length (Fig. 1

) and

growth rate

(

Fig.

)

, consistent with the observation that

162

RcsF

induced cells exhibited

msfGFP

levels intermediate between

empty vector

163

and RcsF

cells (Fig. 1

)

164

165

e tracked cell lineages

in our time

lapse data to determine whether the volume

166

added during the cell cycle was affected by Rcs activation. Upon IPTG addition,

167

the length added over the course of a cell cycle

∆

(a proxy for added volume) in

168

RcsF

cells equilibrated at 30

40% lower

than

that of empty

vector or RcsF

cells

169

(Fig. 1

), consistent with their constant decrease in lower mean cell length (Fig.

170

) and lower growth rate (Fig. 1

). Interestingly,

cell

cycle duration increased by

171

40%

min of induction (Fig.

Taken together, these data

indicate that

172

Rcs activation

alters growth rate, cell size

and cell

cycle timing, motivating us to

173

further investigate the link between the division machinery and Rcs activation

174

175

Cell shape and gr

owth rate changes are not due to c

nic acid production

176

Production of the e

xopolysaccharide colanic acid is regulated by the Rcs

pathway

177

[39]

. In our time

lapse imaging experiments, RcsF

induced cells became

highly

178

separated over time (Fig.

1C,2A

), which we hypothesized was due to the

179

production of colanic acid.

determine

whether colanic acid productio

n was

180

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

responsible for the reductions in cell length

(Fig. 1D)

and growth rate

(Fig. 1E)

181

upon Rcs activation

we deleted

wcaJ

the most upstream

gene

in the colanic acid

182

synthesis pathway

that

encod

the

initial lipid carrier.

Knocking out colanic acid

183

bio

synthesis through disruption early in the pathway avoids

buildup of

184

intermediates

, which have been shown to

change cell shape due to titrating

185

precursor

flux away from cell

wall

synthesis

[40]

We performed time

lapse

186

imaging in a microfluidic flow cell of RcsF

induced

∆

wcaJ

cells. ∆

wcaJ

cells

187

remained closely packed (Fig.

)

, confirming that the cell separation is indeed

188

due to colanic acid

. Moreover, ∆

wcaJ

cells showed a similar decrease in mean cell

189

length

(Fig. 2B)

and growth rate (Fig.

)

as colanic acid

producing cells

(Fig.

190

1D,E).

Therefore, colanic acid production is not the cause of the cell shape and

191

growth rate changes

upon

Rcs

induction.

192

193

Constitutive Rcs activation

increases FtsZ

intracellular concentration and FtsZ

194

localization to division sites

195

A previous study

sho

wed

that RcsB, the primary activat

or of Rcs

regulated

genes,

196

can bind

upstream of one of the many

ftsZ

promoters, specifically the one that lies

197

also upstream of the preceding

ftsA

[14]

, and activate

fts

suggest

ing

that

the

198

Rcs

system may

regulate expression of the division machinery.

To assess this

199

scenario

more precisely,

quantif

ied

expression of the division machinery

upon

200

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Rcs induction,

and

expression to

changes in growth rate and cell length. To

201

quantify total

cellular

FtsZ

concentration as well as

the concentration specifically

202

within the

FtsZ

ring

, w

e introduce

an FtsZ

msfGFP

sandwich

fusion into the

203

chromosome at the native

ftsZ

locus in

E. coli

DH300 ∆

rcsF

cells

with each RcsF

204

plasmid variant

Cells with the

FtsZ

msfGFP

fusion

are

viable

[41]

nd have

205

similar growth rate as that of cells with nati

ve FtsZ

[42]

imaged cells on

206

agarose

pads with

M9+0.04% glucose (

o alleviate the high autofluorescence of LB

)

207

after

supplement

ing

them

with

0 or

15 μM IPTG.

msfGFP intensity (

mean total

208

fluorescence normalized to cell volume

)

was higher in

RcsF

induced compared

209

to uninduced cells

(Fig.

), indicating that Rcs activat

ion increased FtsZ

210

expression

, and

RcsF

induced cells

exhibited a ~50% increase in

FtsZ

ring

211

intensi

ty relative to uninduced cells (Fig. 3B). During time

lapse imaging in a

212

microfluidic flow cell, FtsZ

ring intensity increased

over time

in RcsF

induced

213

cells

, concurrent with

decrease in mean cell length (Fig.

Across a population

214

of single cells,

cell length and FtsZ

ring intensity followed

a tightly constrained

215

inverse relationship

with both

and

15 μm IPTG (Fig.

). Thus, we inferred that

216

decreases in cell length were

coupled to

increase

in FtsZ ring intensity.

217

218

To verify the increase in FtsZ concentration, we directly measured FtsZ protein

219

levels with antibodies. We analyzed wild

type, ∆

rcsF

strains with the empt

y and

220

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

RcsF

plasmids, and a ∆

rcsF

∆

rcsB

strain with the RcsF

plasmid (the latter to

221

avoid cell size changes caused by ectopic Rcs induction). Strains were diluted and

222

passaged multiple times to ensure exponential growth, and then treated with 0, 2,

223

or 1

5 μM IPTG for >90 min. From Western blotting (Fig. 3E), we

calculated

the

224

ratio of FtsZ to RecA levels (as

control) and normalized to that of wild

type. Basal

225

expression of RcsF

slightly increased the FtsZ/RecA ratio, and addition of 2 or 15

226

μM IPTG le

d to a further increase of ~50%. This increase required the activation of

227

Rcs

controlled genes by RcsB (Fig. 3E). Thus, Rcs activation leads to an increase in

228

FtsZ levels.

229

230

Rcs induction

decreases the separation between

FtsZ rings in filamentous cells

231

Normally dividing cells harbor a single FtsZ ring positioned at midc

ell,

232

regardless of cell length

[33]

o determine whether FtsZ localization

namics

233

and local variations in FtsZ concentration

ere

affected by RcsF

induction

and

234

connected with reductions in cell length

, we

treated

cells

with

cephalexin, a beta

235

lactam that inhibits the divisio

specific transpeptidase FtsI

[43]

These

236

filament

cells

had multiple FtsZ rings that

were easily

identified based on the

237

peaks in per

ipheral fluorescence

(Methods)

[42]

After growth

agarose

238

pads with

IPTG

kymographs

indicated

that Rcs

induced

cells

had

239

more closely spaced FtsZ

rings

than

empty

vector

cells

(Fig.

). Indeed,

RcsF

240

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

induced

cells

exhibited

an increase

number of rings per unit length (Fig.

)

241

and

decrease

distance between FtsZ rings (Fig.

)

compared

with

empty

242

vector or RcsF

induced

cells

Thus,

Rcs

induction increases the number of

243

FtsZ

ring

s per unit length,

thereby defining shorter

cell

ular units even in the

244

absence of cytokinesis

245

246

IgaA depletion

mimics RcsF

induction

247

To determine whether the changes in growth rate (Fig. 1E), cell length

(Fig. 1D),

248

and FtsZ concentration (Fig. 3) were general features of Rcs activation as

249

opposed to an artefact of the RcsF

mutant, we decided to use a different way to

250

induce the Rcs system, this time

using

the essential protein IgaA, which is an

251

inner mem

brane inhibitor of the Rcs system

[10]

. We utiliz

ed a strain expressing

252

the

igaA

L523A

allele from an arabinose

inducible promoter, with the native

igaA

253

deleted. This strain, akin to the previously used

igaA

L643P

allele

[8, 44]

exhibits

254

lower level

cs rep

ression

and hence the effects of depletion are observed

255

more rapidly than

depletion

wild

type

gaA

To determine whether IgaA

L523A

256

depletion

and the consequent activation of the Rcs system

affect

FtsZ

257

concentration, we transduced the FtsZ

msfGFP sandwich fusion onto the

258

chromosome of the

igaA

L523A

inducible strain

; all measurements were

perfor

med

259

in strains with the FtsZ

msfGFP sandwich fusion

monitored growth of

the

260

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

strain

via absorbance

in L

B supplemented

with either 0.2% arabinose, to

261

maintain normal growth, or 0.2% glucose, to deplete

IgaA

L523A

ith arabinose,

262

the IgaA

L523A

strain

grew

similar

to wild

type

E. coli

, while depletion with

263

glucose

substantially reduced growth rate (

1.76

0.17

in arabinose v

ersus

264

1.15

0.20

in glucose;

< 10

test)

(Fig. 5A,

), similar to the

decrease in

265

growth rate

upon RcsF

induction

(Fig.

266

267

We next examined the IgaA

L523A

strain during depletion using single

cell

268

microscopy. When spotted onto agarose pads with 0.2% glucose, cells

physically

269

separated

during growth

and division

signifying col

nic acid production

due to

270

Rcs activation

(Fig.

ean cell

length (Fig.

nd instantaneous growth rate

271

(Fig.

)

decreased in a manner

similar

RcsF

induction (Fig.

1C,D

Size

272

normalized

FtsZ

msfGFP

fluorescence intensity through

ut the cell

(Fig.

)

and

273

FtsZ

ring intensity (Fig.

)

increased coincident with the

decrease in lengt

(Fig.

274

)

, such that total fluorescence remained approximately constant (Fig. 5G).

275

sum,

the phenotypes that emerge

during RcsF

induction

were

closely

276

mimicked by IgaA

L523A

depletion

suggesting

that they are general properties of

277

cs activation

in the absence of

envelope

stress

278

279

A22 activates RcsF without

affecting growth rate or

FtsZ concentration

280

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

To investigate the role of the Rcs system in the response to A22 at the cellular

281

level,

we used time

lapse microscopy to analyze the response of

wild

type

cells

282

a range of

A22

concentrations

(Fig.

. As expected, mean c

ell width

283

increased upon exposur

e to

A22

in a dose

dependent

manner

(Fig.

[30]

284

However,

despite

rapid

activation of the Rcs system by A22

[37]

growth rate

was

285

essentially maintained for the first hour (Fig.

)

, consistent with previous

286

studies showing that

A22 does not affect growth rate

[45]

287

288

This growth rate maintenance

is markedly

distinct from

other Rcs

activating

289

perturbations such as

RcsF

induction (Fig. 1D)

IgaA

L523A

depletion (Fig.

D).

290

Moreover, there was little increase

in the volume

normalized FtsZ

msfGFP

291

intensity (Fig.

)

, as both

total fluorescence

and cell width

increased

together:

292

fold

for the former (Fig. 6E) and ~75% for the latter (Fig. 6B)

at the highest A22

293

concentration

These phenotypes are also distinct

from the increase in FtsZ

294

concentration (Fig. 3A,E, 5E) observed for Rcs activation in the absence of

295

envelope stresses.

Thus,

surprisingly

A22

somehow

reverses the growth

296

inhibition

that normally occurs under

Rcs activation

, and prevents

upregulation

297

of FtsZ

concentration

298

299

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Consistent with the

maintenance of FtsZ concentration

(Fig.

)

despite

300

activation of the Rcs system

under A22 treatment

[37]

the trajectories of

cell

301

width (Fig.

) and growth rate (Fig.

were quan

titatively similar in

A22

302

treated wild

type and

∆

rcsF

cells, indicating that the

morphological

effects of A22

303

treatment

(cell widening)

are independent of

the

Rcs

pathway

Consistent with

304

this finding, wild

type and ∆

rcsF

cells exhibited similar shape trajectories during

305

recovery from A22 (Fig.

Given that FtsZ concentration varies with nutrient

306

determined growth rate

[33]

these data suggest that

the dynamics of cellular

307

dimensions and FtsZ are

downstream of the growth

rate effects of

Rcs activation.

308

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Discussion

309

In this study, we

showed

that activation of the Rcs system

in the absence of

310

envelope stress

adjusts the added length (

) per cycle and decouples elongation

311

and cell

cycling timing

, leading to shorter cells

(Fig. 1

)

This

phenotype

312

achieved by up

regulation of FtsZ

levels

and FtsZ ring formation

A,B,E

)

313

which

likely

are

consequence

the slowdown in growth due to

Rcs

activation

314

[33]

Although

the up

regulation in FtsZ protein levels may

be due

in part

to the

315

effect of Rcs on

ftsAZ

transcription

[14]

how FtsZ increases it

mid

cell localization

316

remains to be elucidated. Overall,

our data highlight the strong coupling between

317

growth rate and FtsZ levels,

the latter of which determines cell length in

a highly

318

stereotyped

manner

(Fig. 3D)

319

320

The

growth

rate

decrease

and upregulation of FtsZ

in RcsF

induced

cells (Fig.

321

)

is a direct result of Rcs activation, as IgaA depletion

had similar effect

322

Thus, these phenotypes are likely to be general conseq

ences of Rcs

323

activation in the absence of envelope stresses.

However,

A22 treatment did not

324

affect growth rate (Fig.

or FtsZ concentration (Fig.

despite activation of the

325

Rcs pathway

[37]

. Moreover, A22

treatment

resulted in

length

decreases

326

similar to RcsF

induction

even in the absence of RcsF

(Fig.

)

indicat

ing

that

327

there

are Rcs

independent

mechanisms for coupling the cell cycle to cellular

328

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

dimensions

[17, 46]

, and/or there may be additional signals that buffer or alter the

329

Rcs response

during cell elongation stress

Together, these findings indicate that

330

the results of activating a stress response can be highly dependent on context.

They

331

also provide a plausible explanation for why

igaA

deletions are non

viable in

332

enterobacteria, unless Rcs is also knocked out

[8, 44]

As we

have

show

, the

333

resulting constant activation

of the Rcs pathway

leads to cells becoming very

334

small, likely tuning

below a level that supports growth. This ability to transition

335

to a no

growing sta

e may be used

enterobacteria to enter a persistent state in

336

host, as IgaA is one of the most down

regulated proteins in non

divi

337

Salmonella

cells in host tissues

[47]

Interestingly, LPS

stress that induces

338

Rcs independently of cell size cues

[37]

cause

transient Rcs activation that fixes

339

the damage

[48, 49]

340

341

The relationship between

the

Rcs

pathway

and cell division has

long

the

342

subject of

spec

ulat

ion, with

overexpression of

RcsB and RcsF

previously

shown to

343

suppress the

division defect of

ftsZ84

mutant

[50, 51]

Further analysis of the

344

FtsZ promoter sequence identified an RcsB binding site in

ftsAZ

that

enables

345

transcription of

ftsZ

upon Rcs induction

[14]

Our findings present a more nuanced

346

pers

pective on

Rcs

mediated regulation of the division machinery

in which

347

increases in FtsZ

concentration

occur only

for conditions

in whic

growth rate

348

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

decrease

such as

RcsF

induction (Fig.

)

IgaA depletion (Fig.

but not

349

A22 treatment (Fig.

)

Taken together,

our findings

reinforce previous studies

350

linking the elongation and division machineries

[42, 52, 53]

and

highlight the

351

importance of growth rate in determining cell size.

352

353

Understanding how the Rcs system enacts physiological changes

in bacterial

cells

354

has important implication

for

the general response of

bacteria

stress

. One

355

significant takeaway

from

our

study is that

cell

shape and transcriptional

changes

356

can

result from

a global change in growth rate rather than

direc

ly from

stress

357

response pathway

activation

That

cells maintain growth rate during A22

358

treatment in the face of

Rcs

activation suggests that other

responses

induced by

359

A22 interfere with some of the downstr

eam Rcs signaling.

It will be interesting to

360

probe the extent to which activation of

stress response

pathways

generally

lead to

361

changes in growth rate

;

ur findings predict that activation of growth

inhibiting

362

pathways will also cause

ftsZ

upregulation.

reover

, changes in growth rate

363

could lead to cross protection of other stresses

[54]

; for example, growth i

364

minimal medium alleviates the essentiality of MreB

[55]

. Ultimately, our strategy

365

of decoupling Rcs activation from the activating stresses should be a powerful

366

strategy for

connecting response pathways to their ensuing

the physiological

367

phenotypes

368

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Methods

369

Strains,

plasmids, and media

370

All strains and plasmids are in Table S1.

E. coli

MG1655 and derivatives of it were

371

used in all experiments. We used P1 transduction to introduce

ftsZ

msfGFP

into

372

the chromosome at the native

ftsZ

locus of DH300 ∆

rcsF

cells, and introduced the

373

pNG162

empt

y, pNG162

WT, and pNG162

IM plasmids harboring the RcsF

374

variants, along with the pTrcHis2A plasmid, which carries

lacI

from a high copy

375

plasmid to shut down basal expression from the pNG162 plasmid

[8]

376

377

Cells were

normally cultivated in LB Lennox (10 g/L tryptone, 5 g/L yeast extract,

378

5 g/L NaCl)

or EZ

RDM

[56]

[57]

media for imaging.

Antibiotics

were

379

added to the culture

when needed to maintain the plasmid

380

381

Growth

rate

measurements

for batch culture

382

For Fig. 1A

and S1

, o

vernight cultures were grown at 37

°C in LB

(

Lennox

383

formulation)

with

appropriate antibiotics

. Cells were diluted to O

578

=0.001, and

384

grown for ~3

until O

578

0.1. Cells were

then re

diluted to O

578

=0.025 in fresh

385

LB with 0 to

μM IPTG.

enable

growth

rate quantification

, O

578

was

386

measured throughout the experiment, and once O

578

=0.3 was reached, cells

387

were diluted to O

578

=0.025. This

passaging

cycle was

maintained for 6 to 9

388

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Growth rate was calculated from O

578

measurements using

≥3

measurements

389

from the range with OD

578

<0.4 by

fitt

ing

to an exponential.

390

391

For the measurements in Fig. 5A,

overnight cultures

in LB with 100 μg/mL and

392

0.2% arabinose

were inoculated into 200 μ

of fresh media supplemented with

393

100 μg/mL of ampicillin and 0.2% of arabinose or glucose

in a clear 96

well plate.

394

The plate was covered with an optical film, with small holes poked at t

he side of

395

each well to allow aeration. Incubation and OD measurements were performed

396

with an Epoch 2 plate reader (BioTek) at appropriate temperatures with

397

continuous shaking and OD

600

measured at 7.5

min intervals.

rowth rate was

398

calculated as the slope

of ln(OD) with respect to time after smoothing using a

399

moving average filter of window size five.

400

401

Imaging

acquisition

on agarose pads

402

For

fixed time

point

experiments, cells were diluted 1:5000 from an overnight

403

culture

. For experiments in Fig. 3A,B, Fig

. 4B,C,

and

Fig. 5E

G, diluted cells were

404

grown for 3 h to OD~0.

1, then diluted 1:10 into fresh medium with appropriate

405

inducers

For experiments in Fig. 6A,B,D,E

and

diluted

ells were

grown

406

to OD=0.4, then

diluted 1:200 in 0, 0.25, 0.5, 1, 2,

5 μg/ml A22

. In these

407

experiments, a small aliquot of cells

(~ 1 μL)

was placed onto agarose pads with

408

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

M9+0.04% glucose every 30 min.

Phase contrast and GFP fluorescence i

mages

409

were acquired as

quickly

as possible t

o avoid cell shape changes due to the

410

medium change.

411

412

For time

lapse

experiments

, cells were diluted 1:5000 from an overnight culture.

413

For experiments in Fig. 3C, 4A, and 5B

, a

fter 3

the

1:5000

dilution

, cells

414

were

further

diluted 1:10 and

spotted onto EZ

RDM pads

(Fig. 3

)

LB pads with

415

8 μM IPTG and 10 μg/mL cephalexin (Fig. 4

)

, or

LB pads with 0.2%

416

arabinose/glucose

(Fig. 5B

For experiment in Fig. 6C, a

t OD

0.2, cells were

417

diluted 1:10 onto LB pads with 1% agarose and 0, 0.25, 0.5, 1

, 2,

5 μg/ml A22.

418

Cells were imaged under phase contrast and fluorescenc

The agarose pads

419

were maintained at 37 °C using a

heat

environmental chamber

(Haison Tech)

420

Phase contrast and f

luorescence i

mages were

acquired

every

min

All phase

421

contrast and fluorescence images were collected

on a Nikon Ti

E epifluorescence

422

microscope

using μManager v. 1.4

[58]

423

424

Imaging in m

icrofluidic

flow cells

425

For imaging experiments in Fig.

and Fig. 2,

ells

were diluted 1:500

from

426

overnight cultures

and grown for 3.5

. Cells were then diluted to approximately

427

0.001 and placed in a

ell

sic

flow cell

chamber

pre

warmed

37 °C

Once

428

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

introduced

into the imaging chamber, cells were grown in LB for 20

min

before

429

switching to

LB supplemented with IPTG.

430

431

uantification

of FtsZ fluorescence levels and spatial distrib

ution

432

All phage contrast images were

segmented using

Morphometrics

[26]

433

434

FtsZ rings were detected based on peaks in intensity along the contour.

FtsZ ring

435

intensity was calculated as the

difference

between

the

max

imum and minimum

436

the fl

uor

escence

peak times the

width of the peak,

after

background

437

subtract

ion

[42]

luorescence

intensity

was calculated as the sum of al

l pixels in

438

the fluorescence channel within the contour of the cell divided by

the

area of

439

contour,

after

background subtract

ion

For data in Fig. 4

FtsZ ring distances

440

were calculated

based on adjacent

fluorescence

peaks as

extracted

from the

441

midline of

the cells

442

443

Western blot

quantification

444

After incubation with IPTG

for at least

1.5 h, 1 mL of

exponentially growing

445

culture was harvested, and lysed and solubilized by boiling in Laemmli buffer

446

for 5

min

at 95

°C. Samples were diluted and normalized

based on

578

prior to

447

SDS

PAGE analysis.

Proteins were separated on

15% SDS

PAGE

gel

, and

448

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

transferred on PVDF membran

es (Immobilon

P).

embranes were blocked with

449

5% skim milk in

TBS

T (

50 mM Tris

HCl [pH 7.6], 0.15 M NaCl, and 0.1%

450

Tween20). TBS

T was used in all subsequent steps of the immunoblotting

451

procedure.

nti

FtsZ (1:1000, Acris) and anti

RecA (loading control,

1:1000,

452

Abcam) rabbit antisera were used as primary antibodies.

embranes were

453

incubated with secondary antibodies conjugated with horseradish peroxidase

454

(HRP) diluted in 5% skim milk in TBS

T (1:10,000, GE healthcare). Labelled

455

proteins were detected via

enhanced chemiluminescence (Pierce ECL Western

456

Blotting Substrate, Thermo Scientific) and exposed on X

ray films (Kodak Biomax

457

Mr1).

458

459

he built

in gel analysis tool in

FIJI

[59]

was used to quantify FtsZ

and RecA

460

levels

from a horizontal rectangle including

the

relevant

bands.

461

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Acknowledgements

462

The auth

ors thank the Huang and Typas labs for useful discussions. This work

463

was supported by a National Science Foundation Graduate Research Fellowship

464

(to A.M.), an ARCS Fellowship (to A.M.),

a James McDonnell Postdoctoral

465

Fellowship (to H.S.)

EMBL core funding

and a DFG grant

(TY 116/2

for

466

SPP1617 (to A.T.),

NIH Director’s New Innovator Award DP2OD006466 (to

467

K.C.H.), NSF CAREER Award MCB

1149328

and

grant EF

2125383

(to K.C.H.),

468

and the Allen Discovery Center at Stanford on Systems Modeling of Infection (to

469

K.C.H.). K.C.H. is a Chan Zuckerberg Biohub Investigator. This work was also

470

supported in part by the National Science Foundation under Grant PHYS

471

1066293 and the

hospitality of the Aspen Center for Physics.

472

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Figure Legends

473

474

475

Figure 1

RscF activation

in the absence of envelope stress

alters cell

length

476

control.

477

In batch culture,

∆

rcsF

cells expressing a

constitutively

active RcsF inner

478

membrane

(RcsF

)

variant show a decrease in

steady

state

growth rate

479

dependent on the level of RcsF

expression

Error bar

represent the

480

standard deviation of three

replicate experiments.

481

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

The

length of

RcsF

induced cells

in batch culture decease

with increasing

482

IPTG

concentration

Top: representative cells. Bottom: violin plots of

the

483

distribution

of cell lengths at each concentration

, with

>500 cells

in each

484

condition

485

fter induction of RcsF

with

μM IPTG

, RcsF

cells

decreased

486

in length relative to ∆

rcsF

cells with an empty vector and became physically

487

separated.

488

RcsF

induction at

result

ed in a gradual reduction in

mean cell length

489

(

solid

black

curve,

right), while ∆

rcsF

cells (empty vector, left

) maintained

490

mean length

and RcsF

resulted in

slight

length

reduction (middle)

491

Gray, blue, and purple lines are trajectories of individual cells.

492

Instantaneous growth rates averaged across the populations in (C) show

493

that RcsF

induction gradually reduced growth rate, while ∆

rcsF

cells

494

maintained growth rate and RcsF

resulted in an intermediate growth

rate

495

reduction.

496

Mean msfGFP fluorescence averaged across the populations in (C) show

497

that Rcs was activated to a high and inter

mediate level in RcsF

and RcsF

498

cells, respectively, but not in ∆

rcsF

cells.

499

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint

Cell length added over the course of each cell cycle ∆

averaged across the

500

populations in (C)

stabilized to a lower ∆

after induction of RcsF

501

compared to ∆

rcsF

or RcsF

duced cells.

502

Division interval averaged across the populations in (C) shows that RcsF

503

induction

increased the time required for division, but not immediately

504

after induction

505

The copyright holder for this preprint (which

this version posted July 27, 2022.

;

https://doi.org/10.1101/2022.07.26.501648

doi:

bioRxiv preprint