Physiochemical Properties of Caulobacter crescentus Holdfast: A

Aug 7, 2013 - Department of Chemistry, Indiana University, Bloomington, Indiana 47405, United States. §. Department of Natural Sciences, University o...
0 downloads 0 Views 5MB Size
Subscriber access provided by DUESSELDORF LIBRARIES

Article

Physiochemical Properties of Caulobacter crescentus Holdfast: A Localized Bacterial Adhesive Cecile Chantal Berne, Xiang Ma, Nicholas A. Licata, Bernardo Ruegger Almeida Neves, Sima Setayeshgar, Yves V. Brun, and Bogdan Dragnea J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/jp405802e • Publication Date (Web): 07 Aug 2013 Downloaded from http://pubs.acs.org on August 20, 2013

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

1

Physiochemical Properties of Caulobacter crescentus Holdfast: a Localized

2

Bacterial Adhesive

3

Cécile Berne †, Xiang Ma ‡, Nicholas A. Licata §, Bernardo R.A. Neves /, Sima

4

Setayeshgar //, Yves V. Brun* †, Bogdan Dragnea* ‡

5 6



Department of Biology, Indiana University, Bloomington, IN 47405, USA

7



Department of Chemistry, Indiana University, Bloomington, IN 47405, USA

8

§

Department of Natural Sciences, University of Michigan-Dearborn, Dearborn, MI 48128,

9

USA

10

/

11

//

Universidade Federal de Minas Gerais, Belo Horizonte, MG 30123-970, Brazil. Department of Physics, Indiana University, Bloomington, IN 47405, USA

12

13

* Address correspondence to: [email protected] and [email protected]

14 15

16

KEYWORDS. Caulobacter crescentus, bacterial bioadhesive, atomic force microscopy,

17

dynamic force spectrometry, adhesion kinetics

18 1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

19

20

Page 2 of 43

ABSTRACT

To colonize surfaces, the bacterium Caulobacter crescentus employs a polar

21

polysaccharide, the holdfast, located at the end of a thin, long stalk protruding from the

22

cell body. Unlike many other bacteria which adhere through an extended extracellular

23

polymeric network, the holdfast footprint area is tens of thousands times smaller than

24

that of the total bacterium cross-sectional surface, making for some very demanding

25

adhesion requirements. At present, the mechanism of holdfast adhesion remains poorly

26

understood. We explore it here along three lines of investigation: a) the impact of

27

environmental conditions on holdfast binding affinity, b) adhesion kinetics by dynamic

28

force spectroscopy, and c) kinetic modeling of the attachment process to interpret the

29

observed time-dependence of the adhesion force at short and long time scales. A picture

30

emerged in which discrete molecular units called adhesins are responsible for initial

31

holdfast adhesion, by acting in a cooperative manner.

32

33 34 35

INTRODUCTION Biological adhesives found in the bacterial world are an abundant, yet mostly

36

untapped, source of adhesives with varied composition and properties. They hold

37

promise for industrial and medical applications, offering impressive performance in their

38

natural context: they enable attachment to a broad variety of surfaces, share desirable

39

properties, such as sustainability, biodegradability and biocompatibily, and yield a much-

40

reduced impact on the environment compared to their synthetic counterparts 1.

2 ACS Paragon Plus Environment

Page 3 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Most bacteria are found attached to surfaces where they form large groups of

41 42

cells, called biofilms. A critical step in biofilm formation is the initial single cell

43

attachment, which proceeds from a reversible stage, often mediated by proteinaceous

44

appendages like flagella and pili, to an irreversible stage mediated by polysaccharide

45

adhesins 2,3. After biofilm maturation, a dense extracellular matrix mainly composed of

46

polysaccharides usually encloses the cells and helps to maintain adhesive properties in

47

a broad range of aqueous environments and on a variety of surfaces 4. Earlier studies

48

attempted to describe the viscoelastic properties of biofilms in terms of

49

phenomenological parameters 5,6 and have shed some light on the mechanisms that

50

mediate the transition from the reversible to the irreversible stage of bacterial adhesion 7-

51

11

52

atomic force microscopy (AFM) tips 13, or quartz crystal microbalances 14 to study the

53

development of bacterial adhesion forces on different surfaces. Additionally, single-

54

molecule force microscopy has enabled direct measurement of the elasticity of single

55

biomacromolecules, including bacterial polysaccharides involved in adhesion 15,16. Most

56

of these studies were done on bacterial species which having a mixture of

57

polysaccharides of different composition and structure and appendages such as pili and

58

flagella distributed around the cell surface. However, a few bacterial species of the

59

Alphaproteobacteria group adhere to surfaces using a discrete, microscopic patch of

60

polysaccharide-based adhesive 17,18.

61

. Other studies have used flow displacement systems 12, bacterial cells immobilized on

Such localized anchor points have received much less attention in the past

62

probably mainly due to the difficulty of adhesion measurements at sub-micron scale.

63

Nevertheless, their study is timely because it is reasonable to believe that they must be

64

coping with mechanical stress differently than extended adhesive films. In this work, we

65

quantitatively explore the physiochemical properties responsible for adhesion in a

66

member of this group, Caulobacter crescentus. 3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

67

C. crescentus synthesizes its holdfast adhesin during the differentiation of the

68

motile swarmer cell into a sessile stalked cell (Figure 1A). The holdfast patch (~100 nm

69

diameter) responsible for permanent adhesion to surfaces is found at the tip of a thin

70

cylindrical stalk-like extension (~ 1 µm long) of the cell envelope 19-23. The holdfast has

71

mechanical properties characteristic of an elastic gel 24 and outperforms the strongest

72

biological and commercial glues, with a force of adhesion exceeding 68 N/mm2,

73

sufficient to resist to a variety of stresses including fluid flow and capillary forces 25. The

74

holdfast elastic modulus was estimated at approximately 2.5 x 104 Pa 26, comparable to

75

other biological gels such as collagen or gelatin matrices 26,27.

76

Page 4 of 43

Specific binding of wheat germ agglutinin lectin to the holdfast and its sensitivity

77

to lysozyme indicate that holdfast contains ß-1,4 N-acetylglucosamine polymers 24,28.

78

However, its detailed composition and structure remain largely unknown, due to its

79

strong adhesiveness and inherent insolubility. While oligomers of ß-1,4 N-

80

acetylglucosamine confer gel-like properties on holdfasts 24 and may play a major role in

81

holdfast elastic properties, available data strongly suggest the existence of additional

82

adhesive components in the holdfast 24,28.

83

In this study, we report the first analysis of the development of adhesive forces in a

84

microscopic holdfast anchor through measurements of the time-dependence of holdfast

85

rupture forces on a variety of substrates. Having access to a Caulobacter mutant which

86

sheds holdfast free of cellular components 29, we were able to determine that holdfast

87

morphology is dependent on the surface to which it is bound and that holdfast affinity for

88

a substrate is modulated by hydrophobic interactions and depends on buffer ionic

89

strength and pH. In addition, we evaluated the maximum tensile strength of pure

90

holdfast, free of interference from cellular components, in relation to the nature of the

91

substrate to which it was attached. To this end, we employed Dynamic Force

92

Spectroscopy (DFS) to measure rupture forces between the holdfast and the substrate. 4 ACS Paragon Plus Environment

Page 5 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

93

This method provides both the spatial and temporal resolution required for bridging the

94

molecular and mesoscopic scales at which crucial phenomena take place. We found that

95

holdfast adhesion is strongly time-dependent, involving transformations on multiple time

96

scales. We further demonstrate that the observed time-dependence is well described by

97

a kinetic rate model of adhesin-surface interaction coupled to diffusion of molecular

98

adhesins within the bulk of the holdfast. Our DFS results also show that the initial

99

adhesion of holdfast to surfaces is dependent on the substrate hydrophobicity and

100

roughness. Finally, our data suggest that the GlcNac polymers present in the holdfast

101

and the holdfast anchor proteins are not likely to be major players in the adhesion

102

mechanism, and that cooperative contributions from discrete adhesive units within the

103

holdfast are dominantly responsible for initial adhesion. These findings provide a

104

framework for future molecular mechanistic studies and for comparison of bacterial

105

holdfast properties with the more extensively studied case of adhesive extracellular

106

matrices.

107 108

EXPERIMENTAL SECTION

109

Bacterial strains and growth conditions. The main strain used in this study was

110

Caulobacter crescentus CB15 ∆hfaB (YB4251) 29, a mutant strain from C. crescentus

111

CB15 wild-type (YB135). This mutant has a clean deletion of the hfaB gene and

112

therefore does not synthesize HfaB, one of the holdfast anchor proteins. This strain still

113

produces a holdfast, but is unable to anchor it to the cell envelope. As a consequence,

114

the newly synthesized holdfast is shed in the culture medium and on surfaces 29.

115

Another C. crescentus CB15 mutant, ∆hfsH (YB2198), was used to study the role

116

of deacethylation in adhesion efficiency. Indeed, this mutant is lacking the gene hfsH,

117

encoding a deacetylase that affects both cohesive and adhesive properties of the 5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

118

holdfast 30. C. crescentus ∆hfsH produces smaller holdfasts compared to the wild-type

119

and the ∆hfaB strains. These fully acetylated holdfasts are not anchored properly to the

120

cell envelope and are shed in the medium 30.

121

C. crescentus strains were grown at 30˚C in minimal M2 medium supplemented

122

with 0.2% glucose (M2G) 31 or in low phosphate HIGG medium 32 containing 120 µM

123

phosphate (for stalk sample preparations).

124

Page 6 of 43

Escherichia coli TRMG (MG1655 csrA::kan), a strain that overproduces the

125

polysaccharide PGA (poly-ß-1,6- N-acetylglucosamine polymer) and releases it in the

126

culture medium 2 was grown in LB medium at 37˚C with constant shaking (150 rpm), to

127

maximize PGA production and release 2.

128 129

PGA purification. PGA was purified from E. coli TRMG stationary phase cultures (24 h

130

at 37˚C), as described previously 2. Cells were harvested by centrifugation and the

131

supernatant (8 ml) was concentrated using MWCO 3,000 Centricon units (Millipore) to

132

500 µl final.

133 134

Glass treatment. A hydrophobic treatment was performed on 12 mm glass coverslips

135

(#26020, Ted Pella Inc.). Coverslips were incubated with a 1:1 3-trimethoxysilyl propyl

136

methacrylate (3-TMSM, Acros Organics): anhydrous dimethylformamide (Acros

137

Organics) mixture for 2 h, rinsed twice using 100% acetone and then air-dried.

138

Hydrophobic coverslips were used within a day of treatment.

139 140

Purified holdfast affinity assay. Purified holdfast affinity assays were performed as

141

described previously 33, with few modifications. C. crescentus ∆hfaB cells were grown to

142

late exponential phase (OD600 of 0.6 – 0.8) and cells were pelleted by centrifugation (30 6 ACS Paragon Plus Environment

Page 7 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

143

min at 4,000 g). The supernatant contains free holdfasts shed by the cells. 100 µl of

144

purified holdfasts in solution were spotted on a 12 mm borosilicate glass coverslip

145

(#26020, Ted Pella Inc.), previously glued to a microscope glass slide, and incubated for

146

4 h at room temperature in a saturated humidity chamber. After incubation, the slides

147

were rinsed with dH2O to remove unbound material. Holdfasts were visualized by

148

labeling using AlexaFluor 488 (AF488) conjugated Wheat Germ Agglutinin (WGA)

149

(Molecular Probes). WGA binds specifically to the N-acetylglucosamine residues of the

150

holdfast 28. AF488-labeled WGA (50 µl at 5 µg/ml) was added to the rinsed coverslips

151

and incubated in the dark for 20 min at room temperature. Slides were then rinsed with

152

dH2O, toped with a large glass coverslip (24 x 50 mm) and sealed with nail polish.

153

Holdfast attachment to the coverslips was visualized by epifluorescence microscopy

154

using a Nikon Eclipse 90i and a Photometrics Cascade 1K EMCCD camera. Fluorescent

155

holdfasts were quantified using ImageJ analysis software 34: microscopy 16-bit pictures

156

were manually thresholded using the B/W default setting and fluorescent particles were

157

automatically analyzed with the ImageJ built in function.

158 159

pH sensitivity assays. Purified holdfast binding assays under different pH conditions

160

were performed in 100 mM citrate-phosphate or sodium-acetate buffers (from pH 2.6 to

161

pH 6), 100 mM phosphate or Tris buffers (from pH 6 to pH 8) and N-cyclohexyl-3-

162

aminopropanesulfonic acid (CAPS) buffers (from pH 8 to pH 12). 50 µl of purified

163

holdfasts in suspension were mixed with 50 µl 100 mM buffer and incubated on glass

164

coverslips for 4 h at room temperature in a humid chamber, as described above. Bound

165

holdfast labeling, imaging and quantification were performed as described above. PGA

166

binding assays were run under the same conditions, but incubated for 24 h instead of 4

167

h, to maximize binding. 7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

168

Page 8 of 43

To determine if the low binding efficiency of holdfasts at low pH (< 6) or high pH

169

(> 8) was due to a physical or a chemical modification of the holdfasts, 50 µl purified

170

holdfasts were incubated in suspension with 25 µl of 100 mM buffers at different pH for 2

171

h at room temperature (1st incubation). 50 µl of new buffer were added to the samples to

172

modify their pH, and the samples were allowed to bind to coverslips for 2 h, as described

173

above (2nd incubation). Bound holdfast labeling, imaging and quantification were

174

performed as described above.

175 176

Sample preparation for Atomic Force Microscopy analysis. Early exponential phase

177

grown C. crescentus ∆hfaB cells (OD600 of 0.3 – 0.4) were diluted to an OD of 0.1 in

178

M2G and spotted on a 10 x 10 mm piece of freshly cleaved mica. Samples were

179

incubated at room temperature in a humid chamber. After overnight incubation, the mica

180

was thoroughly rinsed with sterile dH2O to remove all cells and debris. A 100 µl aliquot of

181

sterile dH2O was placed on the surface for DFS experiments.

182

Typically, an AFM image was taken from the holdfast-covered mica surface prior

183

to the experiment. To cover the AFM silicon nitrite tip with holdfast, the tip was placed in

184

contact with a holdfast present on the mica surface for 90 s. Using a trigger force of 5 nN

185

insured maximal tip penetration (down to the substrate). This procedure was repeated

186

several times, until a part of the holdfast present on the mica had been transferred to the

187

AFM tip. A second AFM image was then taken to ensure that part of the holdfast was

188

missing from the surface and was therefore attached to the tip. To confirm holdfast

189

attachment to the tip, the holdfast-loaded tip was moved above a clean mica surface

190

while maintained in dH2O, and a force-displacement curve was recorded to ensure a

191

significant force due to the holdfast coating the tip.

8 ACS Paragon Plus Environment

Page 9 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

192 193

The Journal of Physical Chemistry

The same procedure was followed to coat the AFM tip using purified PGA previously immobilized on a clean mica surface.

194 195

Scanning Electron Microscopy (SEM). The typical area occupied by and the thickness

196

of holdfast attached to the AFM tip were determined using a FEG environmental SEM

197

(Quanta 600F, FEI). Samples were first fixed with 2.5% (v/v) electron microscopy grade

198

glutaraldehyde (Ted Pella, Inc.) in 10 mM phosphate buffer pH 7 for 1.5 h. The samples

199

were then treated with a series of ethanol dehydration steps (30%, 50%, 70%, 90% and

200

100% (v/v), 15 minutes each) and dried using a critical point dryer (Blazers CPD 030).

201

Uncoated samples were affixed to a metal stub with double-stick conductive carbon tape

202

(Electron Microscopy Sciences) and then visualized under secondary electron mode.

203 204

AFM analysis. AFM AC mode images and force-displacement curves were obtained

205

using a Cypher AFM (Asylum Research). Measurements were performed in sterile dH2O

206

at room temperature using gold-coated silicon nitride Biolever cantilevers (Frequency f0

207

= 13 kHz, spring constant k = 0.006 N/m, Olympus Inc.). Spring constants were

208

measured from the thermal noise spectrum of the cantilevers. Force measurements

209

were performed using tips previously coated with holdfasts as described above, using a

210

trigger force of 500 pN if not stated otherwise.

211 212

RESULTS AND DISCUSSION

213 214 215 216

Holdfast on polar and non-polar surfaces Previous biophysical studies on holdfast adhesion have been performed using only one type of surface: borosilicate glass 24,25. To determine if capillary forces may

9 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 43

217

influence significantly initial adhesion, we investigated the morphology of holdfasts

218

bound to two surfaces of different polar character: hydrophilic mica and hydrophobic

219

highly-ordered pyrolitic graphite. We used AFM to image holdfasts at 16 h after

220

attachment (Figure 2A). Height, diameter and contact angles were thus determined for

221

~200 particles (Figure 2). Holdfast height varied from 5 to 100 nm on both surfaces, with

222

a few holdfasts reaching up to 160 nm (Figure 2B). The average height was 30.6 ± 2.4

223

nm and 21.5 ± 0.9 on mica and graphite, respectively. The average holdfast footprint

224

diameter was also substrate dependent (Figure 2C). Holdfasts attached to mica had

225

diameters from 30 to 280 nm, with an average of 90.2 ± 2.7 nm, while holdfasts attached

226

to graphite ranged from 45 to 440 nm, with an average of 119.2 ± 4.1 nm. The average

227

contact angles were 52.6 ± 1.3 ˚ and 38.9 ± 2 ˚ on mica and graphite, respectively

228

(Figure 2 D-F). Since contact angles reflect the relative strength of holdfast-liquid,

229

substrate-liquid, and holdfast-substrate interaction, results in Figure 2 suggest that the

230

graphite-holdfast interaction is stronger than mica-holdfast interaction. Note that both

231

graphite and mica substrate preparations yield atomically-flat surfaces, thus minimizing a

232

possible role played by roughness.

233

Another substrate of interest is glass. Having established that the holdfast

234

contact angle showed marked differences between graphite and mica, we measured the

235

binding affinity of holdfasts to clean and non-polar adsorbate (3-TMSM) coated glass

236

surfaces. In these experiments, purified holdfasts in suspension were allowed to bind to

237

the two types of surfaces, and the amount of surface-deposited holdfasts were quantified

238

using fluorescently-labeled wheat germ agglutinin (WGA), a lectin specific for N-

239

acetylglucosamine residues present in the holdfast 28 (Figure 3A). Figure 3C shows that

240

the binding affinity to hydrophobic 3-TMSM-treated glass seems somewhat smaller (~ 55

241

%) than that of clean glass, which seems to disagree with the finding above that

10 ACS Paragon Plus Environment

Page 11 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

242

attractive capillary forces are stronger on hydrophobic substrates. One explanation that

243

would reconcile these apparently contradictory results is the possibility of a time-

244

dependent curing process, which occurs at slower time scales than those characteristic

245

of capillary interactions. Indeed, we will discuss later in the paper the independent

246

evidence for such processes.

247

To determine if the only identified component, N-acetylglucosamine plays a role

248

in the dependence of adhesion efficiency on substrate polarity, we measured the affinity

249

of PGA (a poly-ß-1,6- N-acetylglucosamine polymer purified from E. coli TRMG 2) for the

250

two types of glass substrates (Figure 3B). Due to very low binding affinity, PGA samples

251

had to be incubated for 16 h to provide measurable coatings (instead of 4 h for holdfast

252

samples). PGA affinity assays exhibited no significant variation as a function of substrate

253

(Figure 3C). These results suggest that the holdfast adhesion mechanism is not

254

dominated by the N-acetylglucosamine adhesive properties and is likely to involve

255

additional components. This hypothesis is also supported by the fact that we had to

256

incubate the PGA samples four times longer than the holdfast ones to obtain

257

measurable coverages.

258

Nevertheless, as we are showing in the following, N-acetylglucosamine plays an

259

important albeit indirect role. Thus, a recent study showed that, in C. crescentus, a

260

mutation in the hfsH gene encoding a deacetylase acting on the holdfast, affects both

261

cohesive and adhesive properties of the holdfast 30. Partial deacetylation of N-

262

acetylglucosamine in holdfast should leave free amine residues in place of acetyl

263

groups, thereby changing the charge of the polysaccharide. Biologically, the free amine

264

group could be useful to covalently link an adhesin or for crosslinking. Indeed, binding

265

affinity of holdfasts produced by the ∆hfsH deacetylase mutant is drastically decreased

266

with around 30-40% of holdfasts attached compared to deacetylated ∆hfaB holdfasts 11 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 43

267

(Figure 3C). This result is in agreement with previous studies 30 and strongly suggests

268

that N-acetylglucosamine deacetylation is crucial for holdfast adhesive properties. The

269

phenomenon is reminiscent of the effect of deacetylation of chitin, a long chain polymer

270

of N-acetylglucosamine, to produce adhesive chitosan 35,36. Note that binding affinity of

271

∆hfsH holdfasts is not significantly different on the hydrophobic and hydrophilic glass

272

(Figure 3C), similarly to PGA.

273

In summary, these results suggest that (i) N-acetylglucosamine is not solely

274

responsible for holdfast adhesion; rather, active components, here referred to as

275

adhesins, dispersed in the holdfast bulk, generate the stronger adhesion and may be

276

responsible for the observed differential surface response, and (ii) N-acetylglucosamine

277

deacetylation mediated by HfsH is important for establishing a cohesive network, and

278

possibly interconnecting adhesins. Further experiments should focus on elucidating the

279

nature of the putative adhesins described in this work; one possibility would be for the

280

adhesin to be a protein or peptide, comparable to bacterial fimbriae protein subunits 37 ,

281

the mussel Mytilus edulis foot proteins 38, or gingipain adhesin peptides 39.

282 283

284

Ionic strength and pH affect holdfast affinity In order to further identify characteristics of adhesin-surface interaction, we

285

investigated the possible role of electrostatics interactions between substrate and

286

holdfast. Thus, purified holdfast binding to glass at different NaCl concentrations was

287

quantified using fluorescence labeling (Figure 4A). PGA binding affinity was found to be

288

insensitive to added salt (Figure 4B), indicating that the adhesive properties of the N-

289

acetylglucosamine molecules were not affected by ionic strength. Similarly, the adhesive

290

properties of fully acetylated holdfasts from the C. crescentus ∆hfsH mutant holdfasts

291

were not affected by ionic strength (Figure 4C). 12 ACS Paragon Plus Environment

Page 13 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

292

The Journal of Physical Chemistry

In stark contrast to PGA and holdfasts produced by the ∆hfsH mutant, binding

293

affinity on hydrophilic clean glass of deacetylated holdfasts decreased visibly with

294

increasing the NaCl concentration. On 3-TMSM-treated glass, surface coverage was

295

insensitive to salt concentration. This behavior points to the occurrence of attractive

296

interactions between charged or polar groups of deacetylated holdfast and the polar

297

glass surface. Since the principal mechanism by which glass and silica surfaces acquire

298

a charge in contact with water is the dissociation of silanol groups, glass is negatively

299

charged at close to neutral pH. At the same time amines in the deacetylated holdfast are

300

positively charged. Silanols can be gradually deprotonated in aqueous solution by

301

adjusting pH. Thus, to further confirm the origin of the electrostatic interaction, holdfast

302

binding assays were next performed in solutions at different pH (Figure 5A).

303

For clean glass, binding affinity rapidly increased from acidic to neutral pH, with

304

roughly 15% and 40% of surface binding at pH 2 and pH 5 respectively, to reach 100%

305

at pH 6.5 to 7.5. Under the same conditions, the binding affinity for 3-TMSM treated

306

glass remained approximately constant (55 to 75% for pH ranging from 2 to 7.5). This

307

observation supports the hypothesis that ionization of silanol groups, which is at least

308

partly suppressed on the 3-TMSM treated glass surface, is responsible for the observed

309

electrostatic interaction. At the same time, PGA binding is insensitive to ionic strength

310

(Figure 4B), therefore other moieties than N-acetylglucosamine and carrying positive

311

charges must be involved from the holdfast side.

312

At pHs higher than 8 and for both types of surfaces, holdfast binding affinity

313

dropped steeply and at the same rate (Figure 5A). PGA binding is greater at acidic pH

314

and the maximal binding affinity of PGA on clean glass occurred around pH 5-6 (Figure

315

5B), whereas neutral pH (6.5 to 7.5) was optimal for holdfast binding efficiency (Figure

316

5A). At basic pH, PGA binding efficiency decreased drastically, being completely 13 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 43

317

abolished at pH higher than 8 (Figure 5B). For fully acetylated ∆hfsH holdfasts, maximal

318

binding was achieved at pH 4-5 and decreased steadily with increasing pH (Figure 5C).

319

When purified holdfasts were incubated on clean glass at pH 5 for 2 h and subsequently

320

adjusted the pH at 7 for an additional 2 h, binding affinity was partially restored (75%,

321

compared to 40% if the total incubation was performed at pH 5, Figure 5A). In contrast,

322

the effects of basic pH were irreversible (Figure 5D). Since the drop of affinity at basic

323

pH occurs for both holdfast and PGA on both surfaces and is irreversible (Figure 5D) we

324

hypothesize that the decrease in affinity at basic pH may involve degradation of the N-

325

acetylglucosamine matrix, likely through base hydrolysis 40.

326 327 328

Time-dependent adhesion studies by dynamic force spectroscopy The smaller angle of contact on hydrophobic surfaces (Figure 2 D-E) suggested

329

stronger holdfast/surface interactions on this type of substrate and therefore the

330

possibility of hydrophobic interactions between adhesins and substrate. However, affinity

331

results indicated more frequent binding to clean glass than to TMSM-coated glass. A

332

hypothesis that could reconcile these facts is the existence of a curing process that may

333

occur after adsorption. This hypothesis is supported by previous work, which indicated

334

that the individual holdfast footprint on the surface increases with time as it is

335

synthesized after initial surface contact 7, suggesting that the holdfast is initially in a fluid

336

state and stops spreading after reaching a 60-200 nm footprint 41. Moreover, surface-

337

holdfast bonds are extremely strong for samples incubated overnight (~ 68 N nm-2) 25,

338

but possibly much weaker initially, thus allowing the organism to explore its environment

339

before binding irreversibly.

14 ACS Paragon Plus Environment

Page 15 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Since prior to this work it was not known what the initial adhesion forces may be,

340 341

we measured rupture forces after initial holdfast adhesion, taken within seconds of

342

contact by gradually increasing incubation (dwell) times by liquid-cell dynamic force

343

spectrometry (DFS) 42. In these experiments, AFM tips were coated with a layer of

344

holdfast, as described in the experimental section (Figure 6A). The surface area of the

345

AFM tip covered with holdfast was analyzed by SEM (Figure 6B) and estimated at ~10-8

346

mm2.

347

The types of surfaces studied were different in terms of both hydrophobic

348

character and microscopic roughness: 1) mica (hydrophilic, atomically smooth,

349

homogeneous surface chemistry), 2) non treated clean glass (hydrophilic,

350

microscopically rough, heterogeneous surface chemistry), 3) 3-TMSM-treated

351

borosilicate glass (hydrophobic, microscopically rough, heterogeneous surface

352

chemistry) and 4) graphite (hydrophobic, atomically smooth, homogeneous surface

353

chemistry). Figures 6C and 6D represent typical force-displacement curves recorded by

354

DFS. Retraction curves exhibit a negative deflection dip, resulting from the adhesion

355

interaction as the tip is pulled back. The lowest point on the retraction curve corresponds

356

to the rupture force. Two kinds of curves were observed: curves with a single adhesion

357

event (Figure 6C) and curves with numerous local minima corresponding to multiple

358

partial rupture events (Figure 6D). In each case, the area enclosed between the negative

359

deflection curve and abscissa represents the work of adhesion. Retraction and extension

360

curves (red and blue respectively, Figure 6C-D) overlap completely between the trigger

361

and the contact point. Thus, holdfast behaved as an elastic medium for the force loading

362

rate (~ 1µm s-1) and magnitude range (0.1 - 1 nN) used in this study.

363

It is important to note that separation at rupture occurs at the contact interface

364

between holdfast and the substrate. Several lines of evidence support this idea: First, 15 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 43

365

direct SEM inspection of the AFM tip after DFS experiments show no visible loss of

366

holdfast material. Second, the cantilever resonance (in air, where quality factor is high)

367

did not change significantly before and after adhesion, indicating that the total mass

368

remained constant within the measurement error (~ 10-15 g) 43. Third, as shown in the

369

following section, the bond strength is much smaller initially than that after long contact

370

times (as in the case of tip/holdfast interface) making it much more likely that rupture will

371

occur at the substrate/holdfast interface. Finally, no significant loss of adhesion could be

372

detected after subsequent measurements using the same coated tip under similar

373

conditions.

374

The work of adhesion corresponding to the initial phases of interaction for the

375

four tested substrates is presented in Figure 7A. Clearly, the work of adhesion increased

376

with the hydrophobic character of the substrate. Graphite stands out with almost two

377

orders of magnitude greater work of adhesion than the other substrates. On graphite, the

378

time to onset of the rapidly increasing phase is shorter than the minimum measurable

379

time of 0.01 s. Long dwelling time adhesion to 3-TMSM treated glass is also significantly

380

stronger than adhesion to untreated glass. However, the time to onset of the rapidly

381

increasing phase is longer on 3-TMSM-treated glass than on graphite. Substrate

382

roughness does not seem to play a major role in initial adhesion. Mica, which is

383

hydrophilic and flat (0.2 nm rms), has a similar work of adhesion with glass, which is also

384

hydrophilic but microscopically rough (4.0 nm rms).

385

If we compare the work of adhesion, using the entire data set of force-

386

displacement curves (single and multi peak curves), with the maximal rupture force data

387

(Figure 7B), we observe the same trend for strength of adhesion as a function of

388

different surfaces. This trend indicates that adhesion strength increases with time on all

389

surfaces but the kinetics are different. Table 1 shows the maximum adhesion force per 16 ACS Paragon Plus Environment

Page 17 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

390

unit area on various surfaces. To find these estimates we have used the maximal force

391

determined at 90 seconds of dwell time by DFS (Figure 7B) and an average contact area

392

between the holdfast-covered AFM tip and the surface of 10-8 mm2 (Figure 6B). As for

393

the work of adhesion, the maximal adhesion force depended on the surface: the more

394

hydrophobic the substrate, the higher the adhesion force.

395

Contact area can be varied in principle by adjusting the maximal compression

396

force (the trigger point) acting on the tip/holdfast complex (Figure 7C). However, for all

397

surfaces and for trigger forces between 250 pN and 5 nN, the work of adhesion and

398

maximal force measurements remained constant within experimental error, which means

399

contact area was constant, the tip likely being in contact with the substrate. However, on

400

both hydrophobic surfaces, the work of adhesion increased with the trigger point force

401

above a threshold value of about 50 pN and then remained relatively unchanged. For

402

hydrophilic surfaces (clean glass and mica) the work of adhesion no such trigger force

403

threshold was observed. Note that the existence of a threshold force may prompt a

404

cooperative interaction between hypothetical adhesins since if the adhesins were

405

interacting with surface sites in a non-correlated manner we would expect in all cases a

406

gradual increase of the work of adhesion as a function of trigger force (due to contact

407

area expansion).

408

Qualitative examination of the force-extension curves revealed that roughly 60%

409

of them contained multiple rupture events. The existence of both single and multiple

410

rupture events highlights the underlying complexity of adhesion through multiple surface

411

bonds. A statistical analysis of the magnitudes of rupture forces and extensions in the

412

DFS force displacement curves was performed (Figure 8). Figure 8A shows the

413

distribution of rupture events by visual identification, while the histogram in Figure 8B

414

was derived from algorithms designed to extract these automatically (Supporting 17 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 43

415

Information). The distribution was fitted with a function consisting of six Gaussian peaks

416

each with a mean corresponding to a particular integer (n = 1-6) multiple of a

417

characteristic rupture force, and with identical widths, plus a constant “background", thus

418

3 fit parameters. The data are well-described by the fit function for a characteristic

419

rupture force parameter of 29.7 ± 0.6 pN. The first three peaks are present at high

420

significance while the others are present at roughly 1.5 - 2 standard deviation level.

421

Therefore, assuming that the high significance peaks in Figure 8A-B are associated with

422

one, two, and three adhesins, we deduce that the initial adhesion occurs through

423

discrete interactions, each carrying approximately 30 pN force. These values are within

424

the range of those found for some small proteins, like ankyrin 44 or dystrophin 45 for

425

example, or polymers, such as polystyrene 46 or polyethylene oxide 47. In contrast,

426

overall single bond rupture force measurements performed on various polysaccharides

427

are an order of magnitude higher than the value obtained here 48,49.

428

In addition, the distribution of extension values corresponding to the rupture

429

events shown above is illustrated in Figure 8C. The most probable extension between

430

rupture events is observed to be approximately 2 nm. This is well above the z-noise of

431

the AFM (~ 0.3 nm) within detection bandwidth. Based upon these data, we suggest that

432

main initial adhesion is likely to occur through adhesin/surface interactions, each contact

433

being capable of 2 nm extension before rupturing. It is worth noting here that in DFS

434

experiments, rupture occurs via thermally assisted escape across an activation barrier

435

that diminishes with applied force. Hence, measured forces are not a sole property of the

436

bound complex but also depend on the loading rate 50. Here, however, the loading rate

437

was held fixed, and we expect that the distribution of forces will vary somewhat for

438

different pulling velocities.

439 440

The smallest average number of rupture events per force-displacement curve is approximately 1 and occurs on atomically flat graphite (Figure 9). The largest average 18 ACS Paragon Plus Environment

Page 19 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

441

number of rupture events per force-displacement curve occurs for glass (both clean and

442

hydrophobic). One possible explanation is that cooperativity of adhesion postulated for

443

graphite in relation with the results of Figure 7C may be manifesting here as well. Thus,

444

in a cooperative bonding scenario, rupture of one adhesion-surface bond is quickly

445

followed by neighboring adhesins as in a zipper. On a morphologically heterogeneous

446

surface such as glass, other interactions (such as rapid spatial variations of interfacial

447

tension) may interrupt adhesin linkage and cooperativity. In this case, the result would

448

be a multiplication of rupture events per contact area.

449

We now return to the question of why in Figure 7 the non-polar character of 3-

450

TMSM glass manifests itself ~10 s after contact with the substrate while it is practically

451

instantaneous for graphite? A possible explanation that correlates well with the idea of

452

relatively sparse, mobile adhesins that organize in cooperative units is illustrated in

453

Figure 10. For a given initial distribution of nonpolar adhesins on the surface of the

454

holdfast proximal to the substrate, these adhesin molecules are able to bind immediately

455

to the homogeneous non-polar graphite. In contrast, a period of time may be required for

456

their rearrangement into domains, allowing effective binding to corresponding patches

457

on a heterogeneous glass surface. Enhancement of overlapping between glass (fixed)

458

and holdfast adhesin (mobile) non-polar patches would occur by diffusion and refolding

459

of the latter. Future experiments performed on surfaces with controlled heterogeneity

460

would be able to further substantiate this description. At this point, we provide a coupled

461

reaction-diffusion model of adhesin diffusion within the holdfast matrix and its reaction

462

with the substrate (see below), which reproduces well the observations.

463

The maximum adhesion force per unit area reported in this study (Table 1) was

464

three orders of magnitude lower than the force reported in previous work 25. We

465

hypothesize that the main reason for this difference is due to difference in the amount of

466

time the holdfast has adhered to the surface. In the Tsang et al. work, the holdfast was 19 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 43

467

in contact with the substrate for days before the pulling measurements were made 25. In

468

our DFS experiments, the holdfast spent approximately one hour on the tip after

469

application compared to an instrument-limited maximum measurement time of 90 s on

470

the surface. Indeed, examination of the force dependence on the dwell time (Figure 7)

471

clearly indicates that holdfast adhesion is strongly time-dependent.

472

A parsimonious reaction-diffusion model given by coupled diffusion of adhesin

473

within the holdfast matrix and its multi-step surface attachment kinetics (Supplemental

474

Information) can indeed describe the dependence of the adhesion force on dwell time in

475

the current DFS experiments as well as reconcile these results with those of Tsang et al.

476

25

477

of the rupture force is determined by surface-adsorbed adhesins, which have not yet

478

undergone the irreversible transition to the surface bound form. The time to onset of

479

weak adhesion is determined by the rates of diffusion of adhesin within the holdfast

480

mass and its reversible association with the substrate, in contrast with the time to onset

481

of the strong adhesion at later times determined by the relatively slower rate of

482

irreversible association with the surface.

(Figure 11A). Within the framework of this model, for short dwell times, the magnitude

483

Following previous works 51,52, we model the rupture geometry as shown in

484

Figure 11B, with parallel surface bonds, coupled through the holdfast to the AFM tip. The

485

dependence of the rupture force on the number of surface-associated adhesin species is

486

assumed to be linear in the scaling regime of loading rate relevant for the experiments

487

reported here 51,52. Within the general framework of this model, for short dwell times, the

488

magnitude of the rupture force is determined by the number of surface-adsorbed

489

adhesins that have not yet undergone an irreversible transition to the surface bound

490

form. The time to onset of weak adhesion is determined by the rates of diffusion of

491

adhesin within the holdfast mass and its reversible association with the substrate, in

20 ACS Paragon Plus Environment

Page 21 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

492

contrast with the longer time to onset of strong adhesion determined by the relatively

493

slower rate of irreversible association with the surface.

494

While this model represents a possible biophysical mechanism with plausible

495

parameter values leading to the separation of time scales for weak and strong adhesion,

496

we additionally consider two related, alternative mechanisms (Supporting Information). i)

497

First, we have quantitatively analyzed the slow diffusion limit of the current reaction-

498

diffusion model with a modified, single-step surface kinetic scheme, where the short and

499

long time scales for adhesion are given by the rates of surface adsorption and bulk

500

diffusion, respectively. A small rate of diffusion of the adhesin could result from rescaling

501

of the bare diffusion constant due to strong adhesin binding to the holdfast

502

polysaccharide matrix. ii) Second, we considered the possibility of cross-linking of the

503

holdfast matrix over time, either by the putative adhesin or side chains of the N-

504

acetylglucosamine polymer matrix itself. This is shown schematically in Figure 11C,

505

leading to stiffening of the holdfast and resulting in a more uniform distribution of an

506

externally applied load on surface bonds. Given the load dependence of the dissociation

507

constant for adhesin-surface binding 53,54, where the unbinding rate increases

508

exponentially with applied load, we hypothesize that when the holdfast is less stiff, an

509

applied load is more likely to be concentrated on a few bonds leading to a greater

510

probability of their rupture. Consequently, a larger load is distributed among the

511

remaining surface bonds, resulting in a cascade of multiple ruptures with shorter rupture

512

time and therefore smaller rupture force. In contrast, with a stiffer holdfast, the applied

513

load transferred to each surface bond and hence the load-dependent dissociation rate is

514

smaller. The rate of holdfast stiffening in the natural environment is expected to be slow

515

to provide sufficient time for the cell to detach from an inhospitable surface, consistent

516

with the onset of strong adhesion at longer times. While these models suggest different

517

mechanisms by which holdfast adhesion strength could evolve from its initial values to 21 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 43

518

considerably higher values over time, they all rely on multiple step kinetics, in absence of

519

which data could not be fit. Further experiments seeking to identify the underlying

520

mechanisms for the observed kinetics will be the scope of future studies.

521 522 523

CONCLUSION In conclusion, an expanded ensemble of biophysical characteristics of bonding

524

development in an isolated microscopic bioadhesive was investigated. We found that: (i)

525

holdfast binding affinity is modified by environmental conditions and the nature of the

526

substrate; (ii) holdfast adhesion varies on multiple time scales; and (iii) a kinetic model

527

can describe the observation of time-dependence of the adhesion force on short and

528

long time scales.

529

Together, our results suggest adhesion is initiated through discrete, cooperative

530

events, with a magnitude of force suggestive of single molecules. The number of these

531

initial surface interactions is enhanced on non-polar substrates. We propose that that the

532

initial adhesive properties of the C. crescentus holdfast are dominated by a yet to be

533

identified adhesin molecule acting in concert and present within a polysaccharide matrix

534

composed of N-acetylglucosamine multimers.

535

Biologically, being able to modulate the strength and the timing of the adhesion

536

process as a function of environmental cues is vital for bacteria. Indeed it has been

537

suggested that permanent adhesion of newborn cells is prevented when environmental

538

conditions deteriorate, allowing their dispersion and the formation of a new colony where

539

the growth conditions are more favorable 33. Uncovering the chemical nature of adhesins

540

as well as mechanisms underlying interactions of the adhesins within the holdfast matrix

541

in response to environment will be critical not only for understanding the remarkable

542

biology of adhesion, but also to modulate its properties for various applications. Indeed,

543

Caulobacter holdfast has all the desired properties for a valuable bioadhesive: it adheres 22 ACS Paragon Plus Environment

Page 23 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

544

to a wide variety of surfaces under aqueous conditions and its adhesion is time-

545

dependent, probably involving a natural curing mechanism, leading to a currently

546

unequalled adhesion strength of 68 N /mm2.

547 548

549

ACKNOWLEDGMENTS We thank members of the Brun laboratory for comments on the manuscript. We

550

thank Prof. T. Romeo (Department of Microbiology and Cell Science, University of

551

Florida) for providing the E. coli TRMG strain and C. Dufort who participated in the

552

preliminary stages of this study. This work was supported by National Institutes of Health

553

Grant GM102841 and by a grant from the Indiana METACyt Initiative of Indiana

554

University (funded in part through a major grant from the Lilly Endowment, Inc.) to YVB

555

and BD, and by a National Science Foundation Grant PHY-0645652 to SS. BRAN

556

acknowledges financial support from CNPq, Brazil.

557 558 559

SUPPORTING INFORMATION Theoretical Methods; Supplementary data about (i) Quantitative Analysis of the

560

Dependence of the Rupture Force on Dwell Time: Reaction-diffusion model of adhesin-

561

surface association and (ii) Analysis of Rupture Event Force Distributions;

562

Supplementary Tables S1 and S2; Supplementary Figures S1 to S7; Supplementary

563

References. This information is available free of charge via the Internet at

564

http://pubs.acs.org

565 566

REFERENCES

567 568

(1) Rehm, B. H. Bacterial Polymers: Biosynthesis, Modifications and Applications. Nat Rev Microbiol 2010. 8, 578-592.

23 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619

Page 24 of 43

(2) Itoh, Y.; Rice, J. D.; Goller, C.; Pannuri, A.; Taylor, J.; Meisner, J.; Beveridge, T. J.; Preston, J. F., 3rd; Romeo, T. Roles of Pgaabcd Genes in Synthesis, Modification, and Export of the Escherichia Coli Biofilm Adhesin Poly-Beta-1,6-N-Acetyl-DGlucosamine. J Bacteriol 2008. 190, 3670-3680. (3) Beloin, C.; Roux, A.; Ghigo, J. M. Escherichia Coli Biofilms. Curr Top Microbiol Immunol 2008. 322, 249-289. (4) Flemming, H. C.; Wingender, J. The Biofilm Matrix. Nat Rev Microbiol 2010. 8, 623-633. (5) Galy, O.; Latour-Lambert, P.; Zrelli, K.; Ghigo, J. M.; Beloin, C.; Henry, N. Mapping of Bacterial Biofilm Local Mechanics by Magnetic Microparticle Actuation. Biophys J 2012. 103, 1400-1408. (6) Shaw, T.; Winston, M.; Rupp, C. J.; Klapper, I.; Stoodley, P. Commonality of Elastic Relaxation Times in Biofilms. Phys Rev Lett 2004. 93, 098102. (7) Li, G.; Brown, P. J. B.; Tang, J. X.; Xu, J.; Quardokus, E. M.; Fuqua, C.; Brun, Y. V. Surface Contact Stimulates the Just-in-Time Deployment of Bacterial Adhesins. Mol Microbiol 2012. 83, 41-51. (8) O'Toole, G. A.; Kolter, R. Flagellar and Twitching Motility Are Necessary for Pseudomonas Aeruginosa Biofilm Development. Mol Microbiol 1998. 30, 295-304. (9) Pratt, L. A.; Kolter, R. Genetic Analysis of Escherichia Coli Biofilm Formation: Roles of Flagella, Motility, Chemotaxis and Type I Pili. Mol Microbiol 1998. 30, 285-293. (10) Watnick, P. I.; Kolter, R. Steps in the Development of a Vibrio Cholerae El Tor Biofilm. Mol Microbiol 1999. 34, 586-595. (11) Yildiz, F. H.; Visick, K. L. Vibrio Biofilms: So Much the Same yet So Different. Trends Microbiol 2009. 17, 109-118. (12) Vadillo-Rodriguez, V.; Busscher, H. J.; Norde, W.; de Vries, J.; van der Mei, H. C. Atomic Force Microscopic Corroboration of Bond Aging for Adhesion of Streptococcus Thermophilus to Solid Substrata. J Colloid Interface Sci 2004. 278, 251-254. (13) Boks, N. P.; Busscher, H. J.; van der Mei, H. C.; Norde, W. Bond-Strengthening in Staphylococcal Adhesion to Hydrophilic and Hydrophobic Surfaces Using Atomic Force Microscopy. Langmuir 2008. 24, 12990-12994. (14) Olsson, A. L.; van der Mei, H. C.; Busscher, H. J.; Sharma, P. K. Novel Analysis of Bacterium-Substratum Bond Maturation Measured Using a Quartz Crystal Microbalance. Langmuir 2010. 26, 11113-11117. (15) Francius, G.; Alsteens, D.; Dupres, V.; Lebeer, S.; De Keersmaecker, S.; Vanderleyden, J.; Gruber, H. J.; Dufrene, Y. F. Stretching Polysaccharides on Live Cells Using Single Molecule Force Spectroscopy. Nat Protoc 2009. 4, 939-946. (16) Marszalek, P. E.; Dufrene, Y. F. Stretching Single Polysaccharides and Proteins Using Atomic Force Microscopy. Chem Soc Rev 2012. 41, 3523-3534. (17) Brown, P. J.; Hardy, G. G.; Trimble, M. J.; Brun, Y. V. Complex Regulatory Pathways Coordinate Cell-Cycle Progression and Development in Caulobacter Crescentus. Adv Microb Physiol 2009. 54, 1-101. (18) Tomlinson, A. D.; Fuqua, C. Mechanisms and Regulation of Polar Surface Attachment in Agrobacterium Tumefaciens. Curr Opin Microbiol 2009. 12, 708-714. (19) Poindexter, J. S. The Role of Calcium in Stalk Development and in Phosphate Acquisition in Caulobacter Crescentus. Arch Microbiol 1984. 138, 140-152. (20) Gonin, M.; Quardokus, E. M.; O'Donnol, D.; Maddock, J.; Brun, Y. V. Regulation of Stalk Elongation by Phosphate in Caulobacter Crescentus. J Bacteriol 2000. 182, 337-347. (21) Wagner, J. K.; Setayeshgar, S.; Sharon, L. A.; Reilly, J. P.; Brun, Y. V. A Nutrient Uptake Role for Bacterial Cell Envelope Extensions. Proc Natl Acad Sci U S A 2006. 103, 11772-11777. 24 ACS Paragon Plus Environment

Page 25 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669

The Journal of Physical Chemistry

(22) Bodenmiller, D.; Toh, E.; Brun, Y. V. Development of Surface Adhesion in Caulobacter Crescentus. J Bacteriol 2004. 186, 1438-1447. (23) Entcheva-Dimitrov, P.; Spormann, A. M. Dynamics and Control of Biofilms of the Oligotrophic Bacterium Caulobacter Crescentus. J Bacteriol 2004. 186, 8254-8266. (24) Li, G.; Smith, C. S.; Brun, Y. V.; Tang, J. X. The Elastic Properties of the Caulobacter Crescentus Adhesive Holdfast Are Dependent on Oligomers of NAcetylglucosamine. J Bacteriol 2005. 187, 257-265. (25) Tsang, P. H.; Li, G.; Brun, Y. V.; Freund, L. B.; Tang, J. X. Adhesion of Single Bacterial Cells in the Micronewton Range. Proc Natl Acad Sci U S A 2006. 103, 57645768. (26) Alipour-Assiabi, E.; Li, G.; Powers, T. R.; Tang, J. X. Fluctuation Analysis of Caulobacter Crescentus Adhesion. Biophys J 2006. 90, 2206-2212. (27) Sato, H.; Kataoka, N.; Kajiya, F.; Katano, M.; Takigawa, T.; Masuda, T. Kinetic Study on the Elastic Change of Vascular Endothelial Cells on Collagen Matrices by Atomic Force Microscopy. Colloids Surf B Biointerfaces 2004. 34, 141-146. (28) Merker, R. I.; Smit, J. Characterization of the Adhesive Holdfast of Marine and Freshwater Caulobacters. Appl Environ Microbiol 1988. 54, 2078-2085. (29) Hardy, G. G.; Allen, R. C.; Toh, E.; Long, M.; Brown, P. J.; Cole-Tobian, J. L.; Brun, Y. V. A Localized Multimeric Anchor Attaches the Caulobacter Holdfast to the Cell Pole. Mol Microbiol 2010. 76, 409-427. (30) Wan, Z.; Brown, P. J.; Elliott, E. N.; Brun, Y. V. The Adhesive and Cohesive Properties of a Bacterial Polysaccharide Adhesin Are Modulated by a Deacetylase. Mol Microbiol 2013. (31) Johnson, R. C.; Ely, B. Isolation of Spontaneously Derived Mutants of Caulobacter Crescentus. Genetics 1977. 86, 25-32. (32) Poindexter, J. S.; Hagenzieker, J. G. Novel Peptidoglycans in Caulobacter and Asticcacaulis Spp. Journal of bacteriology 1982. 150, 332-347. (33) Berne, C.; Kysela, D. T.; Brun, Y. V. A Bacterial Extracellular DNA Inhibits Settling of Motile Progeny Cells within a Biofilm. Mol Microbiol 2010. (34) Abramoff, M. D.; Magelhaes, P. J.; Ram, S. J. Image Processing with ImageJ. In Biophotonics International, 1994; Vol. 11; pp 36-42. (35) Singla, A. K.; Chawla, M. Chitosan: Some Pharmaceutical and Biological Aspects - an Update. J Pharm Pharmacol 2001. 53, 1047-1067. (36) Sorlier, P.; Denuziere, A.; Viton, C.; Domard, A. Relation between the Degree of Acetylation and the Electrostatic Properties of Chitin and Chitosan. Biomacromolecules 2001. 2, 765-772. (37) Kline, K. A.; Fälker, S.; Dahlberg, S.; Normark, S.; Henriques-Normark, B. Bacterial Adhesins in Host-Microbe Interactions. Cell Host & Microbe 2009. 5, 580-592. (38) Silverman, H. G.; Roberto, F. F. Understanding Marine Mussel Adhesion. Mar Biotechnol (NY) 2007. 9, 661-681. (39) Boisvert, H.; Duncan, M. J. Translocation of Porphyromonas Gingivalis Gingipain Adhesin Peptide A44 to Host Mitochondria Prevents Apoptosis. Infect Immun 2010. 78, 3616-3624. (40) Chebotok, E. N.; Novikov, V. Y.; Konovalova, I. N. Depolymerization of Chitin and Chitosan in the Course of Base Deacetylation. Russian Journal of Applied Chemistry 2006. 79, 1162-1166. (41) Li, G.; Brun, Y. V.; Tang, J. X. Holdfast Spreading and Thickening During Caulobacter Crescentus Attachment to Surfaces. BMC Microbiol 2013. 13, 139. (42) Hinterdorfer, P.; Garcia-Parajo, M. F.; Dufrene, Y. F. Single-Molecule Imaging of Cell Surfaces Using near-Field Nanoscopy. Acc Chem Res 2012. 45, 327-336.

25 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701

Page 26 of 43

(43) Ilic, B.; Craighead, H. G.; Krylov, S.; Senaratne, W.; Ober, C.; Neuzil, P. Attogram Detection Using Nanoelectromechanical Oscillators. Journal of Applied Physics 2004. 95, 3694-3703. (44) Lee, G.; Abdi, K.; Jiang, Y.; Michaely, P.; Bennett, V.; Marszalek, P. E. Nanospring Behaviour of Ankyrin Repeats. Nature 2006. 440, 246-249. (45) Bhasin, N.; Law, R.; Liao, G.; Safer, D.; Ellmer, J.; Discher, B. M.; Sweeney, H. L.; Discher, D. E. Molecular Extensibility of Mini-Dystrophins and a Dystrophin Rod Construct. J Mol Biol 2005. 352, 795-806. (46) Li, I. T.; Walker, G. C. Single Polymer Studies of Hydrophobic Hydration. Acc Chem Res 2012. (47) Liu, K.; Song, Y.; Feng, W.; Liu, N.; Zhang, W.; Zhang, X. Extracting a Single Polyethylene Oxide Chain from a Single Crystal by a Combination of Atomic Force Microscopy Imaging and Single-Molecule Force Spectroscopy: Toward the Investigation of Molecular Interactions in Their Condensed States. J Am Chem Soc 2011. 133, 32263229. (48) Rief, M.; Oesterhelt, F.; Heymann, B.; Gaub, H. E. Single Molecule Force Spectroscopy on Polysaccharides by Atomic Force Microscopy. Science 1997. 275, 1295-1297. (49) Marszalek, P. E.; Oberhauser, A. F.; Pang, Y. P.; Fernandez, J. M. Polysaccharide Elasticity Governed by Chair-Boat Transitions of the Glucopyranose Ring. Nature 1998. 396, 661-664. (50) Dudko, O. K.; Filippov, A. E.; Klafter, J.; Urbakh, M. Beyond the Conventional Description of Dynamic Force Spectroscopy of Adhesion Bonds. Proc Natl Acad Sci U S A 2003. 100, 11378-11381. (51) Seifert, U. Rupture of Multiple Parallel Molecular Bonds under Dynamic Loading. Physical Review Letters 2000. 84, 2750-2753. (52) Seifert, U. Dynamic Strength of Adhesion Molecules: Role of Rebinding and SelfConsistent Rates. Europhys. Lett. 2002. 58, 792. (53) Bell, G. I. Models for the Specific Adhesion of Cells to Cells. Science 1978. 200, 618-627. (54) Evans, E.; Ritchie, K. Dynamic Strength of Molecular Adhesion Bonds. Biophys J 1997. 72, 1541-1555.

702 703 704

FIGURE LEGEND

705 706

Figure 1. (A) Caulobacter crescentus cell cycle. This dimorphic bacterium starts its life

707

as a motile swarmer cell with a single polar flagellum and pili at the same pole. The

708

swarmer cell is unable to initiate DNA replication and eventually differentiates into a

709

replication-competent stalked cell by shedding its flagellum, retracting pili, and

710

synthesizing the holdfast followed by a stalk at the same pole. The stalked cell initiates

26 ACS Paragon Plus Environment

Page 27 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

711

DNA replication, and divides to give rise to a new swarmer cell. (B) AFM picture of a C.

712

crescentus stalked cell. (C) AFM picture of shed holdfasts attached to a mica surface.

713 714

Figure 2. (A-C) Holdfast size distribution. (A) AFM picture of purified holdfasts attached on a

715

mica surface. Height (B) and diameter (C) distribution of holdfasts. on mica (white bars) and

716

graphite (black bars). n = 191 and 212 for mica and graphite samples respectively. (D-F)

717

Contact angle of holdfast on mica (D) and graphite (E). (F) Relative contact angle

718

distributions on mica (white bars) and graphite (black bars) as described in the experimental

719

section. n = 96 for mica and 72 for graphite.

720 721

Figure 3: B/W thresholded image from AF488-WGA labeled WT holdfasts (A), and PGA (B).

722

(C) Surface attachment to clean (white bars) and 3-TMSM treated (black bars) glass.

723

Results are expressed as percent of WT holdfasts attached to clean glass (for all holdfasts

724

experiments) or as percent of surface coverage of PGA on clean glass. The bars represent

725

the average of at least 10 fields of view for four indepedent replicates and the error bars

726

represent the standard error.

727 728

Figure 4: Surface attachment to clean (white symbols, dashed lines) and 3-TMSM-treated

729

(black symbols, solid lines) glass, using different concentrations of NaCl, as described in the

730

experimental section. (A) Purified holdfasts, (B) PGA and (C) ∆hfsH holdfasts.

731 732

Figure 5: Surface attachment to clean (white symbols, dashed lines) and 3-TMSM-treated

733

(black symbols, solid lines) glass, using different pH. (A) WT holdfasts, (B) PGA and (C)

734

∆hfsH holdfasts. (D) Surface attachment of purified holdfasts to clean (white bars) and 3-

27 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

735

TMSM-treated (black bars) glass, using sequential different pH, as described in the

736

experimental section.

Page 28 of 43

737 738

Figure 6: (A) Schematics for holdfast adhesion to the AFM tip as described in the main text.

739

(B) Scanning electron micrographs of a pristine tip and the tip after holdfast attachment (top

740

row). Close up of the pristine tip (bottom left) and the difference image between a tip before

741

and after holdfast loading (bottom right). The holdfast-covered area is in white (bottom right).

742

(C-D) Representative force-displacement curves collected from mica at 0.5 nN trigger force

743

and 2 s dwell time (red: extension. blue: retraction curve). Single-peak retraction curve (C),

744

(multi-peak retraction curve (D). The light gray area represents the work of adhesion.

745

T=trigger point, C = contact point, R = rupture point.

746 747

Figure 7: Experimental data for work of adhesion (A) and maximal force (B) for four

748

different surfaces as a function of dwell time (fixed trigger point = 500 pN). (C) Work of

749

adhesion as a function of the trigger point (fixed dwell time = 2 s). Data are expressed as an

750

average of 4-16 independent replicates and the error bars represent the standard error. Mica

751

is shown in red, graphite in black, clean glass in blue and 3-TMSM treated glass in green.

752 753

Figure 8: (A-B) Rupture event histograms analysis. The lower panel of each image

754

displays the histogram of rupture-event forces, corresponding to the visual (A) and

755

algorithmic (20 pN threshold) (B) identification of rupture events in the AFM force-

756

displacement data. The corresponding best-fit function comprising six Gaussians plus

757

uniform background is overlaid. The fits were performed over the range 25 to 250 pN.

758

The upper panel of each image displays the normalized residuals, computed as

28 ACS Paragon Plus Environment

Page 29 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

759

(data−fit)/fit. (C) Distribution of the extension between rupture events for the AFM force-

760

extension curves containing multiple rupture events. The multiple broad peaks are due

761

to the finite step size during tip retraction.

762 763

Figure 9: Frequency of number of single rupture events in force displacement curves

764

exhibiting multiple rupture events. n = 24 for all surfaces (mica in red, graphite in black,

765

clean glass in blue and 3-TMSM treated glass in green).

766 767

Figure 10: Proposed mechanism relying on the diffusion of sparse adhesion molecules

768

towards the contact interface and their binding to surface hydrophobic domains. On

769

graphite, the entire substrate is hydrophobic. In this case, reaching the interface is

770

sufficient to create a bond. On chemically heterogeneous substrates such as 3-TMSM-

771

treated glass, the additional step of interfacial diffusion of adhesins towards surface-

772

bound hydrophobic domains adds a lag time to adhesion.

773 774

Figure 11. (A) The rupture force on mica as a function of dwell time, as measured by

775

DFS (solid squares), and in the model (line). The parameter values of the fit are: h = 60

776

nm, k = 7.2 × 10−6 nm2 s−1, ka = 120 nm s−1, kd = 2.0 × 10−4 s−1, S0 = 1.0 nm−2, P0 = 4.6 ×

777

10−3 nm−3,  = 5000 s , δf1 = 7 pN , δf2 = 134 pN. The mean force as measured by

778

Tsang et al. 25 is shown as a red triangle (reduced by a factor of 4, which corresponds to

779

 /  2. (B-C). Schematic diagram of holdfast surface attachment. (B) In modeling the

780

total adhesion force, attachment occurs via multiple parallel bonds, each assumed to be

781

a Hookean spring with spring constant,  . The number of initial bonds with the surface,

782

, depends on the dwell time, and the load is assumed to be distributed (approximately) 29 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 43

783

uniformly among them. The holdfast is modeled as an elastic element with spring

784

constant  , which is coupled to the AFM cantilever arm with spring constant  , pulled

785

at constant speed . (C) Cross-linking of the holdfast matrix, achieved by the putative

786

adhesin or through chemical modification of N-acetylglucosamine, leads to stiffening of

787

the holdfast and increase in the rupture force.

788 789

30 ACS Paragon Plus Environment

Page 31 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

790

The Journal of Physical Chemistry

Table 1: Maximal adhesion forces per unit area on the surfaces

791 Surface

Maximal calculated adhesion force (N/mm2)

Mica

0.05

Clean glass

0.08

3-TSMS treated glass

0.13

graphite

0.66

792

31 ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1 A Swarmer&cell& Pili&

Flagellum&

9 Mo

) tyle s e if Se le)l

) Predivisional&cell& yle t s fe e)li l i s s Stalked&cell&

Stalk&

Holdfast)

Reversible)a2achment) B

Irreversible)a2achment) C

ACS Paragon Plus Environment

Page 32 of 43

Page 33 of 43

Figure 2 A

B

C 30

80

Frequency (%)

60 Frequency (%)

40

20

10

20

Diameter (nm)

0 42

0 34

0 26

0 18

10

20

0

0 16

0

E

20

F

15 Frequency (%)

10

5

Contact angle (°)

ACS Paragon Plus Environment

80

70

60

50

40

30

20

0 10

D

14

0

12

80

10

60

40

0

Height (nm)

0

0

0

20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The Journal of Physical Chemistry

Figure 3 C 100

75

*

50

25

ld

fa

st

s

PG A Δh f

sH

ho

ol

df

as

ts

0

th

B

w

A

Holdfast attached (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 43

ACS Paragon Plus Environment

Page 35 of 43

Figure 4 B 100 Surface coverage (%)

100

75

50

25

0 0

C Holdfast attached (% of WT holdfasts)

A

Holdfast attached (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

5

10

15

20

NaCl concentration (mM)

25

75

50

25

0 0

5

10

15

20

25

NaCl concentration (mM)

ACS Paragon Plus Environment

100

75

50

25

0 0

5

10

15

20

NaCl concentration (mM)

25

The Journal of Physical Chemistry

Figure 5 A

B

250

Surface coverage (%)

75

50

25

0 0

200 150 100 50

2

4

6

8

10

0 0

12

2

4

6

8

10

12

pH

pH

C

D 100

Holdfast attached (%)

100

75

50

25

50

2nd incubation (on glass)

0

1st incubation (in suspension)

pH 7

ACS Paragon Plus Environment

pH 4

pH 10.5

pH 10

12

pH 7

10

pH 5

pH

8

pH 10

6

pH 7

4

pH 10

2

pH 7

25

pH 5

0 0

75

pH 5

Holdfast attached (%)

100

Holdfast attached (% of WT holdfasts)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 43

Page 37 of 43

The Journal of Physical Chemistry

Figure 6

1 2 A 3 AFM tip 4 5 6 7 Holdfast 8 9 10 11 Strong contact between the AFM tip and the Figure S2bound to the surface 12 holdfast 13 14 a 15 B 16 1500 nm! 17 18 19 20 21 22 23 125 nm! 24 25 26 27 28 29 30 C D 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Transfer of holdfast from the surface to the AFM tip

b

Force measurement using the holdfast-coated AFM tip on a new surface

1500 nm!

r 125 nm!

ACS Paragon Plus Environment

R

The Journal of Physical Chemistry

Figure 7 A

B 10-15

10-14

10-17

10-18

Mica Graphite Clean glass TMSM-treated glass

0.1

1 Dwelling time (s)

10

100

Work of adhesion (N.m)

10-16

10-19 0.01

C 104

Maximal force (pN)

Work of adhesion (N.m)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 43

10-16

103

10-18

102 0.01

0.1

1

10

100

Dwell time (s)

ACS Paragon Plus Environment

10-20

101

102

103

Trigger point (pN)

104

Page 39 of 43

Figure 8

Residual

0

-2

40

40

0 0

50

100 150 200 Rupture force (pN)

250

80

0

-2

20

C

2

60 Count

B 2

Events

Residual

A

Events

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

20

0 0

40

20

50

100 150 200 Rupture force (pN)

250

ACS Paragon Plus Environment

0 0

2 4 6 8 10 Extension between rupture events (nm)

The Journal of Physical Chemistry

Figure 9

20

Mica Graphite Clean glass 3-TMSM-treated glass

15 Frequency (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

10

5

0 0

2

4

6

8

10

Number of rupture events

ACS Paragon Plus Environment

Page 40 of 43

Page 41 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Figure 10

Homogenous graphite

Heterogenous patchy glass

ACS Paragon Plus Environment

The Journal of Physical Chemistry

Figure 11 A 104

Maximal force (nN)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

102

100

10-2 10-2

100

102

104

106

Dwell time (s)

B

C

ACS Paragon Plus Environment

Page 42 of 43

Page 43 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

The Journal of Physical Chemistry

Caulobacter Crescentus

ACS Paragon Plus Environment