Detecting Bacterial Surface Organelles on Single Cells Using

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Detecting Bacterial Surface Organelles on Single Cells using Optical Tweezers Johan Zakrisson, Bhupender Singh, Pontus Svenmarker, Krister Wiklund, Hanqing Zhang, Shoghik Hakobyan, Madeleine Ramstedt, and Magnus Andersson Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.5b03845 • Publication Date (Web): 18 Apr 2016 Downloaded from http://pubs.acs.org on April 19, 2016

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.

Langmuir 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 38

Langmuir

Tuesday, February 23, 2016 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

1

Detecting Bacterial Surface Organelles on Single Cells using

2

Optical Tweezers

3 4 5 6 7 8 9 10 11 12 13 14 15

Johan Zakrisson,† Bhupender Singh,† Pontus Svenmarker,† Krister Wiklund,† Hanqing Zhang,†Shoghik Hakobyan,‡ Madeleine Ramstedt, ‡ and Magnus Andersson†,*

16

Bacterial cells display a diverse array of surface organelles that are important for a range of

17

processes such as: intercellular communication, motility and adhesion leading to biofilm

18

formation, infections and bacterial spread. More specifically, attachment to host cells by

19

Gram-negative bacteria are mediated by adhesion pili, which are nm wide and µm long

20

fibrous organelles. Since these pili are significantly thinner than the wavelength of visible

21

light, they cannot be detected using standard light microscopy techniques. At present, there is

22

no fast and simple method available to investigate if a single cell expresses pili while keeping

23

the cell alive for further studies. In this study, we present a method to determine the presence

24

of pili on a single bacterium. The protocol involves imaging the bacterium to measure its size,

25

followed by predicting the fluid drag based on its size using an analytical model, and

26

thereafter oscillating the sample while a single bacterium is trapped by an optical tweezer to

27

measure its effective fluid drag. Comparison between the predicted and the measured fluid

28

drag thereby indicate the presence of pili. Herein, we verify the method using polymer coated

29

silica microspheres and Escherichia coli bacteria expressing adhesion pili. Our protocol, can

30

in real time and within seconds assist single cell studies by distinguishing between piliated

31

and non-piliated bacteria.

32

†

Department of Physics, Umeå University, 901 87 Umeå, Sweden

‡

Department of Chemistry, Umeå University, 901 87 Umeå, Sweden

*Corresponding author: Magnus Andersson, Department of Physics, Umeå University, Phone: +46 90 786 6336, e-mail: [email protected] Keywords: optical tweezers, E. coli, pili, fimbriae, microsphere

Abstract

1 ACS Paragon Plus Environment

Langmuir


1

1. Introduction

2

Bacterial cells have a complex surface architecture with both nano- and micrometer long

3

surface organelles. These surface organelles are important for cellular functions such as;

4

adhesion, motility, transfer of genetic material, intercellular communication etc.1–3 In

5

particular, adhesion organelles expressed by bacteria are key factors for the biofilm formation

6

and for the initial attachment to host cells as well as for maintaining contact during the first

7

stages of bacterial colonization. A well-studied example is pili of pathogenic Escherichia coli

8

(E. coli), which are essential for initiation of urinary tract and gastrointestinal infections.4,5 In

9

particular, uropathogenic and enterotoxigenic E. coli, (UPEC) and (ETEC), respectively, are

10

known to express a plethora of coilable pili,6 which help the bacteria to resist shearing forces

11

in various host organs by reducing the instantaneous force on the adhesin.7 It is suggested that

12

challenging this switch by means of; proteins, chemicals and antibodies that bind to pili

13

epitopes is a possible approach of preventing bacterial adhesion.8,9 Detecting the presence of

14

pili and understanding its role in pathogenesis are thus important in the mission to develop

15

new drugs against antibiotic resistant bacteria, new antibacterial surfaces, and in vaccine

16

development.10–13

17

Detecting bacterial pili can be achieved by microscopic and molecular-biology methods.

18

Since the width of pili is in the nanometer range, it is usually not possible to visualize them

19

using quick diagnostic techniques such as light microscopy or dynamic light scattering.

20

Localization of pili is possible through the use of fluorescence microscopy techniques, while

21

electron microscopy (EM) together with atomic force microscopy provide high-resolution

22

images of pili or even smaller organelles.9,14–17 EM along with mathematical modelling have

23

been used to demonstrate the morphological states of pili together with other important

24

fundamental properties such as; numbers of pili on a cell, length of pili, rigidity, etc.18–20 In

25

addition, nuclear magnetic resonance spectroscopy (NMR) and Western blot analysis are

26

often used for indirect assessment of pili expression.9,21 2 ACS Paragon Plus Environment

Page 2 of 38

Page 3 of 38

Langmuir


1

One major drawback of the above-mentioned methods is that they are invasive, making it

2

impossible to perform any further analysis on the cells after detection of pili. In addition, most

3

of these methods cannot be applied to investigate the physiological and biomechanical

4

features of pili, i.e., relative extension length, unwinding force, bond kinetics, and elasticity

5

response. However, these features can in turn be assessed using force spectroscopy

6

techniques.22 Force spectroscopy offers a powerful and sensitive method to understand pili

7

mechanics and physiology at the single cell and the single organelle levels.23–25 However, the

8

expression of pili in a bacterial population ranges from high to none, which makes the

9

probability of capturing a piliated single cell for investigation extremely low,26 and thus these

10

types of experiments become tedious. Therefore, there is a need of a non-invasive method that

11

can distinguish piliated from non-piliated bacteria.

12

E. coli are rod-shaped bacterial cells with a typical length of 1-3 µm that express hundreds

13

of peritrichous pili that are 1-3 µm long and 7-10 nm wide.18,27 A high amount of pili

14

protruding out from the cell surface will increase the surface area to mass per bacterium. In a

15

fluid environment, this additional surface area causes an increased drag on the bacterium

16

compared to a “bald” bacterium of the same size that does not express any pili. Using this

17

phenomenon, we developed a simple method that can rapidly identify bacteria expressing pili.

18

Here we report that using a bright-field microscopy image together with a quick oscillation of

19

the sample with a trapped non flagellated bacterium in an optical tweezers instrument provide

20

enough information to differentiate bald bacteria from piliated bacteria. An illustration of the

21

proposed method is shown in Figure 1. Since this is a quick and non-invasive method for cell

22

sorting, the cells can directly be used for downstream applications. For examples:

23

investigating the mechanical properties of pili; expression analysis in response to various

24

external stimuli and drugs such as pilicides; bacterial pathogenesis; attachment of nano-scale

25

animate or inanimate objects on the cell surface and biofilm formation. In addition, combined

26

with flow chambers, the method could be further developed for large scale cell sorting. 3 ACS Paragon Plus Environment

Langmuir


1 2

Figure 1

3


Page 4 of 38

Page 5 of 38

Langmuir


1

2. Materials and Methods

2

2.1 Bacterial cultures

3

E. coli strain HB10128 is a non-piliated strain, which we here refer to as having bald cells. The

4

HB101 cells were transformed with plasmids pHMG93 and pAZZ50. The transformed strain

5

HB101/pHMG93 expressed P pili 29, whereas the HB101/pAZZ50 strain expressed SII pili.30

6

Expression of pili was confirmed using scanning electron microscopy. The HB101 strain was

7

grown on Lysogeny Agar (LA), and strains HB101/pHMG93 and HB101/pAZZ50 were

8

grown on LA supplemented with 100 µg/ml ampicillin. All strains were grown at 37°C and

9

passaged once before harvesting in 1x phosphate buffered saline, pH 7.4 (1xPBS).

10

2.2 Scanning Electron Microscopy (SEM)

11

Bacterial cells were harvested and fixed in 2% paraformaldehyde for 15 minutes at room

12

temperature. The fixed samples were then cast onto poly-L-lysine coated glass slides,

13

dehydrated in series of graded ethanol (70, 80, 90, 95, 100%), critical point dried and coated

14

with an iridium layer of 5 nm. Thereafter, the samples were examined by field-emission SEM

15

(Zeiss Merlin, GmbH) using in-lens secondary electron detector at a beam accelerating

16

voltage of 4 kV and a probe current of 150 pA.

17

2.3 Chemicals

18

Silica microspheres with a diameter of 2.0 µm (Catalog Code SS04N, Manufacturer Lot

19

Number: 7829 – Bang Laboratories); 3-Sulfopropylmethacrylate, K salt (SPM) (CAS:31098-

20

21-1,

21

Copper(I)chloride (CuCl) (CAS: 7758-89-6, Sigma Aldrich); Copper(II)chloride (CuCl2)

22

(CAS: 7447-39-4, Sigma Aldrich); Initiator – (3-Trimethoxylsilyl)propyl 2-bromo-2-

23

methylpropionate (CAS: 314021-97-1, Fluorochem Lt, UK).

24

2.4 Deposition of silane initiator

25

In a small centrifuge tube, a volume of 960 µL of methanol, 15 µL of deionized (DI) water

26

and 9 µL of acetic acid were added to a volume of 25 µL of silica microsphere water

Sigma

Aldrich);

2,2'-Bipyridyl

(BiPy)

(CAS:


366-18-7,

Sigma-Aldrich);

Langmuir


1

suspension. The mixture was shaken for 1 hour. A volume of 20 µL of initiator was added to

2

the mixture and the centrifuge tube was placed on a rocker for 20-24 h. The tube was

3

removed from the rocker, centrifuged and the solvent was aspirated out. A volume of 1-2 mL

4

of methanol was added and the silica microspheres were re-dispersed using an ultrasonic bath.

5

We repeated the last step 5 times.

6

2.5 SPM Polymerization

7

The silica microspheres grafted with initiator31,32 were sonicated in a volume of 2 mL of

8

methanol to disperse and were thereafter degassed by bubbling with nitrogen gas (N2 gas). In

9

a round-bottomed flask, 8.6 g of SPM monomer was dissolved by stirring in a solution of

10

methanol (8 ml) and water (5 ml) and degassed with N2 gas for 15 min. To the monomer

11

solution, 325 mg of BiPy and 7.4 mg of CuCl2 were added and the mixture was degassed by

12

purging N2 gas for 30 min prior to the addition of 82.4 mg of CuCl. The polymerization

13

solution was left for 20 min under N2 gas. Thereafter, the reaction solution was syringed over

14

to the microspheres and left under N2 gas for 120 minutes. The polymerization was stopped

15

by adding DI water and bubbling with air through the solution until it turned green or blue.

16

The mixture was shaken with DI water and centrifuged at least 5 times. This process is

17

estimated to have given a thickness of the polymer brush coating on the microspheres

18

corresponding to around 100 nm.

19

2.6 Assay preparation

20

Bacteria were suspended in 1xPBS to a concentration suitable for single cell analysis. Silica

21

microspheres were similarly suspended in Milli-Q water. A sample chamber was prepared by

22

adding two pieces of double-sided Scotch adhesive tape (product no 34-8509-3289-7, 3M)

23

spaced 5 mm apart, on a 24.0 x 60.0 mm coverslip (no.1, Knittel Glass). A 20.0 x 20.0 mm

24

coverslip (no.1, Knittel Glass) was thereafter gently positioned on top of the tape, forming a

25

5.0 x 20.0 x 0.1 mm chamber. The chamber was infused with either the bacterial suspension

26

or the silica microsphere suspension by adding a few µl of suspension at one of the openings 6 ACS Paragon Plus Environment

Page 6 of 38

Page 7 of 38

Langmuir


1

allowing capillary forces to fill the chamber. To avoid drying of the sample, the chamber was

2

sealed by vacuum grease (DOW CORNING®) at the open ends of the chamber. The sample

3

was thereafter mounted in a sample holder that was fixed to a piezo-stage (Physik Instrument,

4

P-561.3CD stage) in the optical tweezers (OT) instrumentation. The temperature was

5

measured using a thermocouple in the sample chamber to 23.0 °C ± 0.1 °C. Moreover, the

6

suspension viscosity was assumed to only vary with temperature, thus, the viscosity was set to

7

0.932 mPas ± 0.002 mPas.

8

2.7 Optical tweezers system and measurement procedure

9

The OT was built around an inverted microscope (Olympus IX71, Olympus, Japan) equipped

10

with a high numerical aperture oil immersion objective (model: UplanFl 100X N.A. = 1.35;

11

Olympus, Japan) and a 1292 x 964 pixel camera with a cell size of 3.75 x 3.75 µm (model:

12

StingRay F-125, Allied Vision) as described in.33 Figure 2A shows a schematic illustration of

13

the OT setup used in this work, which stands in a temperature controlled room with

14

computers and controllers isolated from the room to reduce noise and vibrations. A

15

continuous wave Nd:YVO4 laser (model: Millennia IR) operating at 1064 nm was used for

16

trapping a single bacterium or silica microsphere. The trap laser was merged with a probe

17

laser (low power HeNe-laser operating at 632.8 nm) using a polarizing beam splitter cube

18

(PBSC). The light from the probe laser was refracted by the trapped object, collected by the

19

condenser and imaged onto a 2D position sensitive detector (PSD, L20 SU9, Sitek Electro

20

Optics, Sweden). The light illuminating the PSD was converted to a photocurrent and

21

thereafter converted to a voltage that was sent to a programmable low pass filter (SR640,

22

Stanford research systems), before collected by a computer and processed with an in-house

23

LabVIEW program. To minimize the amount of noise in the setup and to optimize the

24

measured time series, we used the Allan variance method described in.34

25

The following steps were carried out to detect pili on a single bacterial cell. A video

26

sequence of bright field images were initially acquired to find the geometrical dimensions of 7 ACS Paragon Plus Environment

Langmuir


1

the bacterium. The image that provided the largest parameter values, i.e., the largest length of

2

the long and short axis, L and w, were chosen as illustrated in Figure 1. A representative

3

image of a cell with its dimensions marked with the green arrows is shown in Figure 2B.

4

These parameters were thereafter used to calculate the model diameter, which is described in

5

detail in the theory and experimental procedure section. Prior to a measurement we

6

determined the pixel to length conversion factor (1 pixel = 37.8 ± 0.2 nm, mean ± standard

7

error (SE) on the mean, and the standard deviation (SD) was 0.7 nm, n = 11).

8

The bacterium was thereafter trapped while the sample chamber was oscillated to

9

measure its effective diameter. In the optical trap, a non-spherical object will orient itself with

10

its long axis, L, along the optical axis, as illustrated in Figure 2C. To determine its effective

11

diameter with respect to its long axis the sample was oscillated at 32 Hz with an amplitude of

12

108 nm as previously described35. The sampling frequency was chosen as a multiple of 2 x

13

(where x is a positive integer) of the oscillation frequency. We used a sampling frequency of

14

131 072 Hz ( x = 17 ), and averaged 32 consecutive data sets that were acquired for 0.25 s

15

each, giving a total acquisition time of 8 s. To reduce the influence of the wall during the

16

measurement procedure, the measurement was performed at a height of 10 µm, since the wall

17

influence at this height will be less than 3% for a spherical particle with diameter of 1 µm.36

18

To also ensure a reliable calibration, we chose cells that had a straight body shape, i.e., cells

19

with an clear body curvature were discarded to avoid dissipation of energy due to rotational

20

movement of the cell in the trap.35 Two power spectra of trapped bacteria are shown in Fig.

21

S1.


Page 8 of 38

Page 9 of 38

Langmuir


1 2

Figure 2

3


Langmuir

Page 10 of 38


1

3 Theory and experimental procedure

2

3.1 Drag forces on spherical and non-spherical objects

3

The proposed method assumes that the presence of pili increases the fluid drag on a piliated

4

cell in comparison to a bald cell with the equal cell membrane surface area. To calculate the

5

drag force acting on the membrane of an E. coli bacterium, we use a mathematical model with

6

input parameters assessed from a 2D-image of the cell. The cells used in this work have a

7

cylindrical shape with spherical caps, rather than a spherical or ellipsoidal shape, as quantified

8

in the supplementary materials section, Fig. S2-S3. Fitting both an ellipse and projection of a

9

cylinder with spherical caps to the segmented area of bacteria in SEM micrographs resulted in

10

an average of 89.84 % and 94.84 % agreement (n = 15), respectively. Therefore, we used an

11

empirical model that can handle irregular shapes, like a cylinder with spherical caps, with

12

good performance.37 For cells that have ellipsoidal shape, we suggest to use the analytical

13

drag force model described by Happel and Brenner.36

14

For a non-spherical object, e.g., a rod-shaped bacterial cell, the drag force at low Reynolds

15

number can be defined as described by Leith.38 The drag force on a non-spherical object

16

using this model is given by,

17

FD = 3πηVd Mod ,

18

where η is the dynamic viscosity of the surrounding liquid, V is the velocity of the sphere

19

relative to the fluid and d Mod is the model diameter defined by

20

d Mod =

21

where An is the projected area of the object, see figure 3A, and As is the surface area of the

22

object.

(1)

2 An 2 As + , 3 π 3 π

(2)

23

In the present work, we model cells as cylinders with spherical caps, as depicted in Figure

24

3B, implying that the projected area will alter depending on the orientation of the object. For a


Page 11 of 38

Langmuir


1

trapped cell moving through a fluid with a velocity perpendicular to its cylinder axis, as is

2

shown in Figure 1B and the rightmost cell in Figure 3A, the projected area and surface area

3

becomes An = π w2 4 + w ( L − w ) and As = π wL respectively, where L and w, the objects

4

dimensions, are defined in Figure 1A. The model diameter of the cell is thus given by,

5

d Mod =

6

For a spherical object, i.e., where w = L , Equation (3) reduces as expected to dMod = w and the

7

drag force models given by Equation (1) becomes identical to the standard Stokes’ drag force

8

model for a sphere.

2 3

w2 w 2 + ( L − w) + wL . 4 π 3

(3)

9 10

Figure 3

11


Langmuir

Page 12 of 38


1

3.2 Measuring the effective diameter using optical tweezers

2

To compare the model diameter defined in Equation (2) with the measured effective diameter

3

of a trapped object we use force-measuring optical tweezers. The effective diameter was

4

found by analyzing the power spectral density for the position data of a trapped object. The

5

power spectral density of the trapped object was assessed from the Langevin equation, which

6

describes the Brownian motion of a trapped object and is given by,

7

mx&& ( t ) + γ 0 x& ( t ) + kx ( t ) = (2 k BT γ 0 )1 2 ξ ( t ) ,

8

where m is the mass of the particle, γ 0 is the viscous drag coefficient (assuming a laminar

9

flow) and is given by γ 0 = 3πη d Eff , k is the trap stiffness, x is the position of the trapped

10

object relative to the center of the trap, kB is Boltzmann constant, T is the temperature, ξ ( t )

11

is the normalized white noise with a zero mean value, and dEff is the effective diameter of the

12

object. In an overdamped system, the inertia of a particle can be neglected, i.e., the first term

13

in Equation (4). The power spectral density is given by the square of the time-Fourier

14

transform divided by the measurement time as,

15

P( f ) =

16

where D is the diffusion constant defined by D = k BT γ 0 , f is the frequency, and f c is the

17

corner frequency given by fc = κ ( 2πγ 0 ) .39 Since the position of the trapped particle was

18

measured in Volt we can also define D = β 2 DV , where D V is the measured diffusion

19

constant and β is the conversion factor relating the position of the object in the sample plane

20

to the measured voltage signal. By fitting equation (5) to the power spectra density of the

21

position data, we determined D V . To find the conversion factor we introduced a small

22

oscillation of the sample while recording the position of the trapped object,35 yielding,

D

π

2

(

f 2 + fc2

)

(4)

,

(5)


Page 13 of 38

Langmuir


D

A2

δ ( f − f drive ) ,

1

P( f ) =

2

where A is the oscillation amplitude and fdrive is the oscillation frequency. The conversion

3

factor is given by,

4

β = Wth Wex .

5

where W th and Wex is the theoretical and experimental power spectrum amplitude,

6

respectively. The theoretical amplitude is given by the second term in Equation (6), including

7

the delta function. The experimental amplitude Wex is given by the height at the oscillation

8

frequency relative to the fit using Equation (5) divided by the measurement time for a single

9

power spectrum, measured at the oscillation frequency. Thus, we obtain a relation for the

10

effective diameter of the trapped object, see Figure 1C, in terms of the experimentally

11

measured parameters, T, η, β and DV given by,

12

d Eff =

(

π 2 f 2 + f c2

+

) 2 (1 + fc2

2 f drive

)

(6)

(7)

k BT . 3πηβ 2 DV

(8)

13


Langmuir


1

4 Results and discussion

2

4.1 Verification of the method using polymer coated silica

3

microspheres

4

The fundamental hypothesis in the present work is that surface organelles, even if they are

5

very small, affect the fluid velocity around an object. This change in the fluid velocity will

6

increase the drag force on the cell, which is possible to detect using OT instrumentation. To

7

verify this hypothesis and the theory described above, we performed well controlled

8

experiments with pure silica microspheres, and silica microspheres coated with a thin polymer

9

brush layer consisting of poly(3-sulfopropymethacrylate) (SPM).31,32 Polymer brushes are

10

polymer chains tethered to a surface by one of their chain ends. For these reference

11

microspheres the brushes were synthesized using surface initiated atom transfer radical

12

polymerization, which gives rise to dense brushes with well controlled polymer thickness.40,41

13

The SPM brush is highly hydrophilic with contact angles smaller than 10 degrees.42 In water

14

solution these types of highly charged brushes swell and become completely hydrated leading

15

to a thickness in water that is 2-7 times thicker than the dry thickness.41,43 However, the exact

16

swelling depends on several factors such as ionic strength and solvent quality. The exact

17

thickness of the swollen brush on the reference microspheres was not directly quantified,

18

however, it can be estimated from the thickness of a corresponding brush on a flat surface.

19

Polymerization on a curved surface, such as a microsphere, will be less affected by crowding,

20

implying that the brush thickness on the microsphere would be slightly thicker than on a

21

corresponding flat surface. For the SPM coated microspheres used in this study the

22

polymerization time was 120 minutes, which normally produces a dry thickness of 100 nm on

23

a flat silica surface. In pure water this brush layer is expected to swell to at least 200 nm.

24

The SPM brush is here assumed to mimic surface expressed organelles on a spherical cell

25

as illustrated in Figure 4A. Both types of microspheres were imaged dried using a scanning

26

electron microscope (SEM) to visualize their surface. A representative image of a pure silica 14 ACS Paragon Plus Environment

Page 14 of 38

Page 15 of 38

Langmuir


1

microsphere is shown in Figure 4B together with a SPM brush coated microsphere in Figure

2

4C. Individual polymer chains and their length cannot be resolved and measured with SEM

3

due to the resolution of the SEM and since the polymer brush collapses and compacts when

4

dry. However, the image shows that the coated microsphere has an evenly distributed SPM

5

coating. According to the manufacturer, the diameter of the pure microspheres were 2.0 µm,

6

however, the microsphere size variation was not provided. To measure the microsphere size

7

variation, we used both SEM micrographs and bright-field images. Microspheres in the SEM

8

images showed a high degree of clustering wherefore we manually measured the microsphere

9

size variation, which was found to be 2.0 % (n = 25). On the bright-field images, however, we

10

applied a circle detection algorithm and calculated the microsphere size including the size

11

variation.44 This approach is briefly explained in the caption of Figure S4. It was found that

12

the average size for pure microspheres were 1.989 ± 0.003 µm (n = 138, mean ± SE, SD =

13

0.038 µm).

14

We also quantified the size of the SPM coated microspheres using bright-field images to

15

1.997 ± 0.007 µm (n = 34, mean ± SE, SD = 0.039). A comparison of the size of the pure and

16

SPM coated microspheres showed no significant difference (p = 0.26). Thus, neither of the

17

imaging methods could detect the presence of surface polymers.

18

To determine if the SPM coating could be detected using the proposed OT method we

19

first measured the effective diameter of the pure silica microspheres as a reference. The mean

20

effective diameter was assessed to 1.96 µm ± 0.02 µm (n = 25, mean ± SE, SD = 0.08 µm), as

21

presented in Figure 4D. For an individual measurement, the propagated relative errors in the

22

effective diameter considering the error on temperature, viscosity and the two measured

23

parameters β and DV , were 0.03 %, 0.2 %, 7 %, 4 %, respectively.

24

Next, the effective diameter of the SPM coated silica microspheres was assessed. The

25

highly hydrophilic SPM brush, here in its hydrated form in MilliQ water, was expected to

26

create an additional fluid drag on the microspheres. This was confirmed by the OT method 15 ACS Paragon Plus Environment

Langmuir


1

which resulted in an effective diameter of 2.61 µm ± 0.04 µm (n = 25, mean ± SE, SD = 0.19),

2

see Figure 4D. The increase in the effective diameter is in-line with the expected thickness of

3

the SPM brush. To evaluate the performance for surface polymer detection, the ratio of the

4

effective diameter d Eff and the model diameter d Mod , where the latter is given by Equation

5

(3) with the mean value of the diameter assessed for pure microspheres 1.989 ± 0.003 µm,

6

was calculated. The propagating error for d Mod was estimated to 3.0 % based on the

7

segmentation and fitting procedure, see supplementary materials section Propagating error

8

when quantitating the bacterial shape. The ratio d Eff / d Mod is shown in Figure 4E using a

9

kernel density estimate. The kernel density was calculated using a normal distribution with

10

100 points.

11

For pure microspheres, the ratio dEff/dMod is expected to be one, whereas higher values

12

indicate the presence of surface coating. The mean of the ratio, for the pure and coated silica

13

microspheres were 0.980 ± 0.008 (mean ± SE, SD = 0.039) and 1.31 ± 0.02 (mean ± SE, SD =

14

0.096), respectively. A statistical test (t-test: p 1.5 is required, see Figure 6B.

15

Several studies have reported that absorption of light in optical traps can damaged trapped

16

cells.45–47 We took this into consideration when designing the assay since we wanted the cells

17

to be alive after trapping. The laser was therefore run at a low laser power in the sample

18

plane, 100 mW, and cells were trapped for a short time period, 8 seconds. This gives an

19

estimated energy dose of

Detecting Bacterial Surface Organelles on Single Cells Using

Recommend Documents