Flagellated bacteria have superb self-propulsion capabilities and are able to effectively move through highly viscous fluid and semi-solid (porous) environments. This innate aptitude has been harvested for whole-cell actuation of bio-hybrid microrobotic systems with applications in directed transport and microassembly. In this work, we present the biomanufacturing of Nanoscale Bacteria-Enabled Autonomous Delivery Systems (NanoBEADS) by controlled self-assembly and investigate the role of nanoparticle load on the dynamics of their self-propulsion in aqueous environments. Each NanoBEADS agent is comprised of spherical polystyrene nanoparticles assembled onto the body of a flagellated Escherichia coli bacterium. We demonstrate that the NanoBEADS assembly configuration is strongly dependent upon the nanoparticles to bacteria ratio. Furthermore, we characterized the stochastic motion of the NanoBEADS as a function of the quantity and size of the nanoparticle load and computationally analyzed the effect of the nanoparticle load on the experienced drag force. We report that the average NanoBEADS swimming speed is reduced to 65% of the free-swimming bacteria speed (31 μm/s) at the highest possible load. NanoBEADS can be utilized as single agents or in a collaborative swarm in order to carry out specific tasks in a wide range of applications ranging from drug delivery to whole cell biosensing.

Rapid progress in micro and nanofabrication technologies over the past two decades has led to the development of sophisticated microrobotic systems that are envisioned for a wide variety of potential applications such as in-vivo diagnosis and treatment of diseases, manipulation and assembly of micro/nanoscale objects, and environmental monitoring.1–3 However, numerous challenges in microrobotics remain insurmountable, including on-board power storage, self-actuation, and embedded intelligence.4 Whole-cell actuators, such as flagellated bacteria, offer an effective solution to a few of the aforementioned challenges. These microorganisms utilize chemicals naturally present in their native environments as sources of energy.5 Using that energy, they self-propel in fluidic and semi-solid environments by alternating periods of run and tumble motions in a stochastic manner.6,7 They also have the ability to sense and respond to physical and chemical stimuli in their immediate microenvironment exceedingly well.8–10 Arguably, bacteria remain the most advanced microrobotic systems to date, due not only to their unmatched ability to sense and react to their microenvironment but also their ability to communicate amongst themselves. These same characteristics have motivated many researchers to develop bacteria-based microrobots,11–15 where bacteria are used as the living component of bio-hybrid microrobotic systems in order to achieve specific tasks such as micro/nano-scale cargo transport and manipulations.11,13,15–18

In this work, we investigate biomanufacturing and self-propulsion dynamics of Nanoscale Bacteria-Enabled Autonomous Delivery Systems (NanoBEADS). Each NanoBEADS agent is comprised of a peritrichously flagellated Escherichia coli bacterium, with all the advantages described earlier, coupled with polystyrene nanoparticles via a biotin-streptavidin link. These bio-hybrid robotic agents are capable of carrying nanoscale load for the “targeted” delivery of drugs in otherwise inaccessible regions of the body, including poorly vascularized regions of solid tumors.19,20 A complete understanding of the NanoBEADS assembly (i.e., nano-loads attachment rate and configuration) and self-propulsion dynamics (i.e., speed, drag effects) is crucial to achieving efficacious therapeutic outcome. In this work, controlled self-assembly of nanoparticle load onto the bacterial cell was examined. Furthermore, the self-propulsion dynamics of the NanoBEADS as a function of the load size and quantity in a chemically isotropic aqueous environment was analyzed. E. coli MG1655m, a derivative of E. coli MG1655 from the K-12 family with increased motility was used in all experiments.21 Bacterial culture from a single colony was incubated overnight in 10 ml of fresh Luria-Bertani (LB) Broth (1% w/v of tryptone, 0.5% w/v of NaCl, and 0.5% w/v of yeast extract) in an incubator shaker (30 °C, 150 rpm). A 100 μl aliquot of the overnight culture was inoculated in 10 ml of fresh LB and was grown to OD600 of 0.5 (∼2.5 × 108 CFU/ml). A 1 ml aliquot of the culture was then centrifuged at low speed (1700 × g) for 5 min at room temperature and suspended in 1 ml of motility buffer (0.01M potassium phosphate, 0.067M sodium chloride, 10−4M ethylenediaminetetraacetic acid (EDTA), 0.01M glucose, and 0.002% (v/v) Tween-20). Bacteria were washed twice in the motility buffer and incubated with 10 μg/ml of biotin-conjugated goat polyclonal anti-lipid A LPS antibody (Thermo Scientific, Waltham, MA). The suspension was gyrated on a vortex shaker for 1 h at 600 rpm to facilitate antibody attachment to the bacterial cell. The bacterial suspension was then centrifuged at low speed (1700 × g) for 5 min at room temperature to remove the unbound antibody from the solution and was then resuspended in 50 μl of motility buffer. Streptavidin-coated carboxylate polystyrene nanoparticles (109 nm diameter and 390 nm diameter, Bangs laboratories, Fishers, IN) were agitated with biotinylated antibody-coated bacteria at specified ratios for 30 min. NanoBEADS were constructed through formation of streptavidin-biotin complex between the streptavidin coated nanoparticles and bacteria that were coated with biotin-conjugated antibody. The antibody used in this work was raised against the “O” antigens only present on the cell membrane, which restricts the attachment of the biotin-conjugated antibody (and thus streptavidin coated nanoparticles) to the cell surface, as shown in Fig. 1. Microscopy images and time-lapse videos of the NanoBEADS were captured using a Zeiss AxioObserver Z1 inverted microscope equipped with an AxioCam mRM camera and a 63 × oil objective. The recorded images and videos were obtained with Zen Blue 2012 software (Zeiss Microscopy) at 12.4 frames per second. Scanning electron micrographs of the NanoBEADS were acquired using a FEI Quanta 600 FEG environmental scanning electron microscope.

FIG. 1.

Representative NanoBEADS configurations, (a) schematic of a NanoBEADS agent showing streptavidin-biotin bond between the bacterium and the nanoparticles, (b)-(c) optical microscopy images of 390 nm particles attached to bacteria, (d) and (e) SEM images of 390 nm particles attached to bacteria, (f)–(i) SEM images of 109 nm particles attached to bacteria. All scale bars are 500 nm.

FIG. 1.

Representative NanoBEADS configurations, (a) schematic of a NanoBEADS agent showing streptavidin-biotin bond between the bacterium and the nanoparticles, (b)-(c) optical microscopy images of 390 nm particles attached to bacteria, (d) and (e) SEM images of 390 nm particles attached to bacteria, (f)–(i) SEM images of 109 nm particles attached to bacteria. All scale bars are 500 nm.

Close modal

The process of NanoBEADS biomanufacturing by controlled self-assembly can be described through a theoretical framework developed by Smoluchowski based on aggregation kinetics.22 Bacteria and nanoparticles were mixed and agitated under uniform and laminar shear conditions in order to facilitate self-assembly of the nanoparticles onto the bacterial cell. According to the Smoluchowski theory, the total number of collisions for orthokinetic aggregation (Jbp) and the associated rate coefficient (kbp) can be respectively determined by

Jbp=kbp×Nb×Np,
(1)
kbp=G6×(db+dp)3,
(2)

where Nb and Np are the number density of bacteria (b) and particles (p), G is the fluid shear rate, db and dp are, respectively, the diameters for bacteria (approximately 1 μm) and nanoparticles (390 nm). As illustrated in Eq. (2), the effect of the nanoparticle size is tremendous and larger particles have a tendency to favor more collisions. Moreover, for smaller nanoparticles, viscous effects that arise from the hydrodynamic interactions between bacteria and nanoparticles during orthokinetic aggregation may prevent collision.22 Because of the low Reynolds number regime present during aggregation, particles tend to follow a curved path near the body of bacteria. This leads to larger separation distances between the nanoparticle and the body of the bacterium, thus limiting collision between the two. It is important to note that in biomanufacturing of NanoBEADS for drug delivery applications, there will be a trade-off between the larger particle sizes that favor aggregation and increase the yield of formation of NanoBEADS, and the smaller particle sizes that are favored for transvascular and interstitial transport. The fluid shear rate, G, and the number density of bacteria and nanoparticles also affect the aggregation outcome. We kept the shear rate, bacteria size, and particle size constant and investigated the effect of number density on aggregation and NanoBEADS formation. Three ratios of nanoparticles to bacteria (10:1, 50:1, and 100:1) were investigated. The attachment yield for the different ratios used is illustrated in Fig. 2. It can be seen that the occurrence rate of nanoparticle attachment to bacteria can be improved through higher initial nanoparticles to bacteria ratio. At 10:1 ratio, nearly 70% of bacteria have no nanoparticle load while at 50:1 ratio, almost 70% bacteria have more than 4 particles attached.

FIG. 2.

Rate of occurrence of 390 nm particles attachment to bacteria at various nanoparticles-bacteria construction ratio (n20).

FIG. 2.

Rate of occurrence of 390 nm particles attachment to bacteria at various nanoparticles-bacteria construction ratio (n20).

Close modal

The motile behavior of the NanoBEADS was studied to determine the effect of quantity and size of the nanoparticle load on the self-propulsion of bacteria. Two particle sizes of 390 nm and 109 nm were used. Optical microscopy does not permit a precise count of the number of nanoparticles attached to the body of a bacterium moving in an aqueous environment; thus, we report aggregate speed values for NanoBEADS at various construction ratios. The average speed and the standard deviation for free swimming bacteria, motile 109 nm NanoBEADS, and motile 390 nm NanoBEADS are summarized in Fig. 3. In all analysis, NanoBEADS with an average speed above 5 μm/s were considered motile. As seen, free-swimming bacteria have swimming speed of 31.0 ± 5.5 μm/s. This average speed value is in accordance with previously reported values for E. coli.23 NanoBEADS, irrespective of the size and number of the attached nanoparticles, have lower average speeds. The lower speeds can be attributed to the attachment of nanoparticles to the body of the bacteria and the increased overall drag force experienced by the NanoBEADS. To theoretically illustrate the impact of nanoparticle load on NanoBEADS self-propulsion, three-dimensional computational fluid dynamics (CFD) models of freely-swimming bacteria as well as NanoBEADS were developed in COMSOL® multiphysics. The propulsion force of a single bacterium moving at the experimentally measured average speed of 31.0 μm/s was found to be 0.42 pN. Given NanoBEADS low Reynolds number (Re < 10−4), and significant dominance of viscous effects over inertia, we assume that for each NanoBEADS agent, Fpropulsion=FDrag. According to the Stokes' law, the drag force at low Reynolds number regime can be theoretically defined as FDrag=CDV, where V is the propulsion speed and the damping coefficient, CD, is a function of shape and size of the object, and the fluid property. In a COMSOL® multiphysics model of the NanoBEADS, the effect of the nanoparticle position on the bacteria body was investigated to determine the attachment location that leads to the maximum damping coefficient value (data not shown). For all the NanoBEADS configurations analyzed, the nanoparticle attachment location was chosen to obtain the maximum drag force on the NanoBEADS. The computational results for the relationship between the damping coefficient and the number of nanoparticles attached to the bacterium are illustrated in Fig. 4. It can be seen that the damping coefficient linearly increases with increased number of nanoparticle load. This contributes to an increase in the drag force and a decrease in average swimming speed of NanoBEADS. For instance, for the 390 nm NanoBEADS, a weighted average of the damping coefficient for each construction ratio of 10:1, 50:1, and 100:1 was determined, using the occurrence rate data from Fig. 2 in conjunction with the damping coefficient data from Fig. 4. The weighted average for each loading ratio scenarios was obtained using the following equation:

CD,avg=i=04+(Occurrenceratei×CD,i).
(3)

The corresponding average velocities were determined using the Stoke's law in conjunction with the calculated propulsion force of a single bacterium (0.42 pN) and the average damping coefficient. Using this method, the average speed values for each construction ratio scenario were found to be, respectively, 29.4 μm/s for the 10:1 ratio, 25.3 μm/s for the 50:1 ratio, and 25.6 μm/s for the 100:1 ratio. The calculated average speeds for each of the construction ratio scenarios of the 390 nm NanoBEADS are in good (on average within 23%) agreement with the experimental speed values shown in Fig. 3. The close match between experimental and theoretical speed data suggests that this theoretical framework can be used to provide an approximate prediction of the speed of NanoBEADS at other nanoparticle size and construction ratios without the need for experimental characterization of self-propulsion. The similar speeds recorded of 109 nm and 390 nm NanoBEADS may be attributed to the clumping of 109 nm particles during the self-assembly process. Similar patterns of agglomeration in 109 nm diameter or smaller nanoparticles in biological media have been previously reported.24,25 Degree of flocculation is dependent upon a variety of factors including the size of the particles, the concentration of particles, the surface chemistry and charge density, and the protein coating of the nanoparticles. The aggregates formed have sizes that are close to larger nanoparticles (i.e., 390 nm particles), which leads to similar kinematic behavior of the NanoBEADS (Fig. 1(f)–1(i)). Additionally, it should be noted that even though the use of membrane-specific antibody limits the attachment of the nanoparticles to bacterial cell, nanoparticle attachment could still affect the motility of NanoBEADS. This may be attributed to the fact that E. coli are peritrichous bacteria with random flagella distribution on their body. Given the size of the flexible hook (∼59 nm) in the flagellar assembly,26 it is plausible that one or more of the bacterial flagellar structures may directly interact with nanoparticles, leading to disturbance in the flagellar bundle formation and compromised motility.

FIG. 3.

Average and standard deviation of the NanoBEADS speed as a function of nanoparticle size and construction ratio (n = 33, p < 0.0001).

FIG. 3.

Average and standard deviation of the NanoBEADS speed as a function of nanoparticle size and construction ratio (n = 33, p < 0.0001).

Close modal
FIG. 4.

Damping coefficient of NanoBEADS as a function of the number and size of the attached nanoparticles.

FIG. 4.

Damping coefficient of NanoBEADS as a function of the number and size of the attached nanoparticles.

Close modal

In summary, we report the biomanufacturing of NanoBEADS of various assembly configurations and investigation of the role nanoparticles to bacteria ratio on the controlled self-assembly outcome. We show that an increase in the nanoparticles to bacteria ratio leads to a higher yield in the nanoparticle-bacteria self-assembly and formation of NanoBEADS with larger loads. Furthermore, we experimentally and computationally characterized the stochastic motion of the NanoBEADS as a function of the quantity and size of the nanoparticle load and demonstrated that smaller size nanoparticles will not significantly deteriorate the self-propulsion of bacteria; the damping coefficient for the 109 nm NanoBEADS with 4 nanoparticles attached to the bacterial cell is only 2.5% higher than the damping coefficient experienced by a free-swimming bacterium with no load. This is especially valid if the location of nanoparticles attachment does not interfere with the flagellar assembly of the bacterium. However, larger size nanoparticles or larger loads of smaller nanoparticles cause a self-propulsion speed reduction of up to 35%. In future, we will continue our efforts on enhancing various aspects of our biomanufacturing process to further reduce the agglomeration of nanoparticles during the NanoBEADS construction process. NanoBEADS can be employed to dramatically improve transport (targeted delivery) of larger quantity (multiple nanoparticles can self-assemble onto bacteria) of therapeutic drugs to solid tumors.

The authors would like to thank colleagues in the MicroN BASE laboratory at Virginia Tech, especially Ali Sahari and Zhou Ye for their assistance with the experiments and electron microscopy. We also thank Birgit Scharf from the Department of Biological Sciences at Virginia Tech for generously providing E. coli MG1655m. This project was partially supported by the National Science Foundation (IIS-117519) and Virginia Tech's Institute for Critical Technology and Applied Sciences (ICTAS).

1.
A.
Forgione
,
Surg. Oncol.
18
,
121
(
2009
).
2.
F.
Ullrich
,
C.
Bergeles
,
J.
Pokki
,
O.
Ergeneman
,
S.
Erni
,
G.
Chatzipirpiridis
,
S.
Pané
,
C.
Framme
, and
B. J.
Nelson
,
Invest. Ophthalmol. Vis. Sci.
54
,
2853
(
2013
).
3.
P.
Valdastri
,
P.
Corradi
,
A.
Menciassi
,
T.
Schmickl
,
K.
Crailsheim
,
J.
Seyfried
, and
P.
Dario
,
Rob. Auton. Syst.
54
,
789
(
2006
).
4.
B. J.
Nelson
,
I. K.
Kaliakatsos
, and
J. J.
Abbott
,
Annu. Rev. Biomed. Eng.
12
,
55
(
2010
).
6.
H.
Berg
,
Phys. Today
53
(
1
),
24
(
2000
).
7.
H.
Berg
,
Random Walk in Biology
(
Princeton University Press
,
1983
).
9.
H. C.
Berg
and
L.
Turner
,
Nature
278
,
349
(
1979
).
10.
Y.
Magariyama
and
S.
Kudo
,
Biophys. J.
83
,
733
(
2002
).
11.
G.
Martel
,
S.
Tremblay
,
C.
,
Ngakeng
, and
S.
Langlois
,
Appl. Phys. Lett.
89
,
233904
(
2006
).
12.
B.
Behkam
and
M.
Sitti
,
Appl. Phys. Lett.
90
,
023902
(
2007
).
13.
M.
Steager
,
E.
Kim
,
C.
Patel
,
J.
Bith
,
S.
Naik
,
C.
Reber
, and
L.
Kim
,
Appl. Phys. Lett.
90
,
263901
(
2007
).
14.
R.
Fernandes
,
M.
Zuniga
,
F. R.
Sassine
,
M.
Karakoy
, and
D. H.
Gracias
,
Small
7
,
588
(
2011
).
15.
D.
Akin
,
J.
Sturgis
,
K.
Ragheb
,
D.
Sherman
,
K.
Burkholder
,
J.
Robinson
,
A.
Bhunia
,
S.
Mohammed
, and
R.
Bashir
,
Nat. Nanotechnol.
2
,
441
(
2007
).
16.
M. A.
Traoré
,
A.
Sahari
,
B.
Behkam
, and
M. A.
Traore
,
Phys. Rev. E
84
,
61908
(
2011
).
17.
B.
Sahari
,
A.
Headen
, and
D.
Behkam
,
Biomed. Microdevices
14
,
999
(
2012
).
18.
A.
Sahari
,
M. A.
Traore
,
B. E.
Scharf
, and
B.
Behkam
,
Biomed. Microdevices
16
(
5
),
717
(
2014
).
19.
N. S.
Forbes
,
Nat. Rev. Cancer
10
,
785
(
2010
).
20.
J. M.
Pawelek
,
K. B.
Low
,
D.
Bermudes
,
D.
Pawelek
,
J. M.
Low
, and
K. B.
Bermudes
,
Lancet Oncol.
4
,
548
(
2003
).
21.
F. R.
Blattner
,
Science
277
,
1453
(
1997
).
22.
J.
Gregory
,
Particles in Water: Properties and Processes
(
CRC Press
,
2005
), p.
200
.
23.
N. C.
Darnton
,
L.
Turner
,
S.
Rojevsky
, and
H. C.
Berg
,
J. Bacteriol.
189
,
1756
(
2007
).
24.
A.
Albanese
and
W. C.
Chan
,
ACS Nano
5
(
7
),
5478
(
2011
).
25.
S. C.
Hayden
,
G.
Zhao
,
K.
Saha
,
R. L.
Phillips
,
X.
Li
,
O. R.
Miranda
,
V. M.
Rotello
,
M. a
El-Sayed
,
I.
Schmidt-Krey
, and
U. H. F.
Bunz
,
J. Am. Chem. Soc.
134
,
6920
(
2012
).
26.
A.
Sen
,
R. K.
Nandy
, and
A. N.
Ghosh
,
J. Electron. Microsc.
53
,
305
(
2004
).