Ink-Jet Printing: Perfect Tool for Cantilever Array Sensor Preparation for Microbial Growth Detection

The use of microcantilever arrays for microbial growth detection provides a rapid and reliable technique for monitoring growth in industrial and clinical applications. Improving the reproducibility and sensitivity of this technique is of great importance. Ink-jet printing has been successfully used in microfabrication and biofabrication due to its high precision; however, only a few microbe-based applications have been reported. Here we demonstrate the advantages of its use for microcantilever based-growth sensing. Four microbial strains Escherichia coli , Staphylococcus aureus subsp. aureus , Pseudomonas aeruginosa , and Candida albicans were deposited and successfully grown on agarose-coated cantilevers by ink-jet printing. When compared to the capillary-coating method, ink-jet printing demonstrated more controlled cell deposition on cantilevers. The e ﬀ ect of various conditions on cell morphology was also investigated.


Introduction
The bioengineering fields of genomics, drug screening, tissue engineering, regenerative medicine, and various biosensor applications rely on biofabrication techniques which combine living and nonliving components deposited in a controlled manner.Bioprinting provides a rapid and reliable alternative to traditional methods giving the required spatial resolution for these applications [1].Although biological structures are considered fragile, many studies have proven that ink-jet printing is feasible for a broad spectrum of biological material [2][3][4][5][6], and even more complex living systems such as cells can be deposited by this technique [7,8].Drop-on-demand bioprinters use different mechanisms to force the "bio-ink" through a microfluidic chamber to the output orifice [1].Thermal ink-jet printheads use a heating element to raise the temperature in the reservoir creating a bubble which then forces a small amount of ink through the output orifice.In a piezoelectric printhead, a piezo actuator is supplied with short electrical pulses which are converted into short pressure pulses.These pulses propagate through the liquid.Due to inertia forces and supported by surface tension, a small volume of liquid breaks off the liquid column to form a droplet.The droplet then flies freely with a velocity of 2-3 m/s.A pressure-driven printhead uses a pressure source behind the reservoir to produce a force on the liquid.A gate opens to allow ink to flow through the orifice [1].It is clear that each of these approaches could damage the biomaterial by heat, electric field gradient, or pressure shock, and therefore, the range of working parameters must be optimised.Depending on the application, other issues such as nozzle clogging by cells or particles, aseptic printing, and aerosol formation have to be taken into account.Despite these problems, the spatial precision of the printed biomaterial renders these techniques highly advantageous for biosensing applications.Computer-controlled deposition offers a rapid functionalisation technique for micron-sized sensors such as cantilevers, quartz crystal microbalance (QCM), or surface plasmon resonance-(SPR) based sensing.
In this paper, we focus on microcantilever array functionalisation for cell-based sensor applications.Cantilever arrays operated in dynamic mode allow microbial growth analysis of various microbial cultures faster than conventional culturing techniques [9,10].This biological sensor is based on oscillating cantilevers, where additional mass loading onto the cantilever surface results in a change of its resonance frequency.Ramos et al. demonstrated [11] that the response of nanomechanical resonators depends on both the position and the stiffness of adsorbed biological material on the cantilever surface.This is important for the achievable sensitivity of the technique.Here we introduce the use of ink-jet printing for the functionalisation of cantilevers with industry-relevant microbial species for growth detection.C) for 110 min in order to reach a logarithmic growth rate.One mL of each culture was precipitated by centrifugation (10 000 rpm, 5 min, Genofuge 16 M, Techne) and resuspended in 0.5 mL of the tested medium (LB, 10% LB, ME, 10% ME, or deionised water).Cell concentrations of the cultures were determined using cell counting chambers and optical microscopy.

Cantilever Preparation.
Cantilever arrays were obtained from the IBM Zurich Research Laboratory (Zurich, Switzerland).Each array has 8 cantilevers with a length of 500 µm, width of 100 µm, and thickness of 2-7 µm with a pitch of 250 µm.In preparation for microbial coating, the arrays were cleaned using O 2 plasma (0.3 mbar, 3 min, PICO Barrel Asher: Diener electronic GmbH and Co. KG, Ebhausen, Germany) and silanised for 45 min (1% (3-glycidyloxypropyl)trimethoxysilane, 1% N-ethyldiisopropylamine in water-free toluene).The epoxy-silanised arrays were washed in waterfree toluene (2 × 15 min) and dried in N 2 and subsequently coated with 1% (wt/vol) agarose-water solution (SeaKem Gold Agarose, Bioconcept; NH, USA) by using preheated (>100 • C) glass microcapillaries (King Precision Glass Inc., Calif, USA).The pH of the melted agarose was adjusted to pH 11.9 by adding 2 M NaOH to the solution.The cantilevers were incubated in the agarose-filled capillaries for 5 seconds.Using this method a thin agarose layer is formed on the cantilever surface.The boundary of this gel layer is visible on the images of each cantilever.

Ink-Jet Spotting.
The MD-P-801 autodrop dispensing system (Microdrop, Norderstedt, Germany) is comprised of a piezo-driven pipette (AD-K-501) and is mounted on a three axis micropositioning system which has an accuracy of 1 µm in all directions (Figure 1).The nozzle diameter used was 50 µm.The volume of liquid deposited is not only dependant on its viscosity but also on the voltage and the pulse length applied on the piezo actuator.In order to deposit one droplet per pulse, these parameter settings have to be adjusted for each individual suspension prior to  deposition.The final spot diameter depends on the surface tension of the printed material and on the surface on which it is deposited.By using a low pulse frequency in our experiments we allowed each droplet to soak into the agarose surface before the next droplet landed on the surface.With this method, aerosol formation could be prevented, and; therefore, no cross-contamination between the cantilevers is expected.The micropipette was cleaned 3 times with 70% ethanol solution and twice with 96% ethanol solution before use.These cleaning steps were sufficient to avoid nozzle clogging or cross-contamination.Cell suspensions prepared as mentioned previously were loaded into 96-well microtiter plates (Sterilin Ltd., Norwick, UK).Prior to filling the micropipette, solutions in the microtiter plate reservoirs were mixed in order to provide an even cell concentration in the nozzle.Droplets of the different suspensions were spotted on each cantilever independently.To avoid crosscontamination of the lever surfaces the nozzle was cleaned with ethanol between each tested suspension.

Scanning Electron Microscopy (SEM) Sample Preparation.
The cell-functionalised cantilever arrays were exposed to relatively dry room conditions (20-40% RH) for less than 30 minutes.For growth testing, they were placed in high Prior to SEM imaging, cantilever arrays were dried at 24 • C and 35% RH for at least 24 hrs.Cantilever arrays were immersed in 5% glutaraldehyde solution (0.05 M phosphate buffer, pH 7) and incubated at room conditions with gentle shaking for 3-4 h.After fixation, excess glutaraldehyde was removed by washing in phosphate buffer (0.05 M, pH 7) for 10 min.This step was repeated 6 times.Samples were subsequently dehydrated using 10-30-50-70-90-100-100% ethanol solution (10 minutes for each step).After drying, 10 nm Pd was deposited on the samples (Cressington Sputter Coater 208 HR Watford, United Kingdom).SEM images (SEM-ULTRA, Carl Zeiss, Germany) were taken at 5 kV, with an aperture size of 30,000 µm using the lens detector.

Results and Discussion
Microcantilever arrays were previously used to detect the viable growth of E. coli [10,12] and Aspergillus niger (A.niger) [9,13].In these studies individual glass microcapillaries were used to seed cells on individual agarosefunctionalised cantilevers.The deposition of A. niger spores was performed by functionalising the cantilever surface with a uniform distribution of Aspergillus-specific antibodies.The seeded microbial species were mixed with a culture medium which provided essential nutrition for microbial growth.After immersion of the levers in the suspension, nutrition is stored in the porous agarose layer.During growth microbes assimilate nutrition and water from the nutritive layer and humid air.In order to compare the coating efficiency of ink-jet printing to that of the previously used capillary coating method, two cantilever arrays were prepared for growth testing (Figure 2).Agarose-coated cantilevers were immersed to full, half, and 1/5 of their lengths in glass capillaries filled with E. coli suspension for 6, 4, and 2 minutes, respectively, (Figure 2(a)).On a second agarosecoated array the same E. coli suspension was deposited by spotting at different points along the cantilever.At each position, 10 drops of cell suspension were dispensed with high accuracy (Figure 2(b)).SEM images of E. coli colonies were taken after 24 h growth.As a control, a second set of arrays were prepared as described above; however, they were not incubated at optimal growth conditions (non growth controls).Capillary coating with these relatively large structures (2 µm) results in the accumulation of cells at the capillary open end.This can be explained by the flow of cells with the liquid and evaporation at the front of the tube (Figure 2(c)).Compared to capillary functionalisation, the localised deposition of cells can be easily controlled by ink-jet printing.The sharp edge of the drop boundary visible on the SEM image (Figure 2(b)) shows that the positional accuracy of the delivered droplet is highly reproducible.Deposition of 180 droplets with 3 different dilutions of microbe suspension shows that the average droplet size and the average cell number per droplet correlate with the dilution of the samples (Table 1).There was no increase in the cell number per drop which indicates that cells are not sedimenting in the nozzle during deposition.The optimal cell number to avoid nozzle clogging in the case of bacteria is under 5-6 × 10 8 cfu and under 1.2-1.4× 10 6 cfu in the case of C. albicans.It is important to mention that before glutaraldehyde fixation of samples, salt and medium residues covered the cantilevers indicating good diffusion of the nutrition along the full length of the lever.The washing and dehydration steps after glutaraldehyde fixation were sufficient to remove these residues without the removal of the deposited cells (Figure 3).We postulate that the thin liquid layer present at >90% relative humidity may act to keep the   in solution on the cantilever surface which aids in this distribution.Therefore, in an actual growth measurement, at high humidity, no salt-carbohydrate-protein crystal layer is formed which would have an effect on the mechanical properties of these sensors and the growth characteristics of the seeded microorganisms.This provides an opportunity for initial cells to spread along the surface even if nutrition is originally available only at the spotting site at the beginning of the experiment.A similar distribution of nutrition is not seen on the non-growth control cantilever which was not incubated at elevated humidity levels.In some experiments Figures 4(c Each array was prepared twice, and one of each was used as a non-growth control.SEM images of these experiments show the viability and multiplication of the tested organisms (Figure 6).In each case, elevated cell numbers can be observed on the arrays exposed to growth conditions (high humidity and elevated temperature).Candida albicans is a yeast form microfungi with typical cell dimensions of 4-5 µm in diameter.At certain conditions, mostly related to limited nutrition, they are able to grow pseudohypha where the cells have not separated after cell division.This morphological switch is often studied as it has a major role in C. albicans pathogenicity [14].In our tests, C. albicans cells were seeded in the budding-yeast form in 100% ME suspension and also 10% ME suspension.Figure 6(a) shows that in the suspension with full medium some cells are developing filamentous structures (pseudohyphae) but also budding-yeast form can also be observed.It is likely that in this special microenvironment the pseudohyphae play a part in the spread of Candida along the cantilevers.All these processes resulting in the colonisation of the surface might cause changes in the cantilevers mechanical properties, and; hence, influence growth detection measurements which are originally based on the detection of a mass increase on the cantilever surface.For some biosensing techniques, high salt or sugar content of microbial culturing medium can disturb the measurement although decreasing the available nutrients in the environment can cause changes in growth rate or morphology of the cells.In other experiments C. albicans (Figure 7(a)) and S. aureus (Figure 7(b)) cells were deposited on agarose-coated cantilevers with diluted medium.We found that decreasing nutrition in the used suspension results in a decrease in viability and also altered the morphology of the cells.Cells started to accumulate extracellular biomaterial which resembled biofilm formation (Figure 7).Bacteria and Candida colonies usually form biofilms on various surfaces including silicon in order to help the bacterial population to survive in certain environments.However, further investigation has to be done to understand this process in the case of microcantilevers.It is expected that biofilm formation will influence the spring constant (k) of a cantilever.It is also important to mention that the use of deionised water for cell deposition resulted in the centering of bacterial cells with respect to droplet positions, while the use of more viscous suspensions, such as culture medium, resulted in the accumulation of cells around the rim of the liquid droplet, thus, allowing for high accuracy when positioning microorganisms within a droplet (data not shown).

Conclusion
It was successfully proved in our experiments that different types of microbial cells are able to be deposited on microcantilever surfaces while retaining their viability.Ink-jet spotting was used for cantilever functionalisation with viable cells, thus, revealing its potential as a superior alternative for sensor applications and real-time growth measurements.
There is no need for specific antibodies facilitating tight binding during deposition of cells.In the case of motile microorganisms the fluid film allows cells to be freely distributed, and, in these cases the introduction of antibodies could keep the species in place.Compared to the previously reported capillary-coating technique ink-jet printing has been shown to be a more accurate way to position cells, and hence, to optimise the detection of mass changes during microcantilever-based growth detection.

Figure 1 :
Figure 1: (a) Illustration of autodrop ink-jet spotting system for cantilever array functionalisation.Piezo-driven micropipette is mounted on a positioning system which has a positional accuracy of 1 µm in all directions.(b) Viable E. coli cells are able to be inoculated by spotting on an agar plate in a controlled manner.The developing colonies form the CRANN Logo after an incubation of 18 hours at 35 • C. Scale bar 5 mm.The different optical appearance of the individual E. coli colonies spotted is a result of different rates of metabolism which lead to different colouring.

Figure 2 :
Figure 2: (a) SEM images of capillary-coated cantilevers for full (top), half (medium), and 1/5 (bottom) immersion in E. coli suspension.(b) Ten drops of E. coli suspension were deposited at different positions along the cantilevers' longitudinal axis.Cantilever arrays in images (a) and (b) were incubated for 24 h at optimal growth conditions before SEM imaging.The width of the cantilevers is 100 µm.The used suspension (2.49 × 10 8 cells/mL) was prepared in 10% LB.(c) SEM image showing accumulation of E. coli cells at the front end of capillary tube after functionalisation.Scale bar 10 µm.

Figure 3 :
Figure 3: Optical microscope images of cantilevers with widths of 100 µm.Ten drops of E. coli suspension (2.49 × 10 8 cells/mL) were deposited at different positions on the cantilevers.(a) Cantilever array was incubated at growth conditions before image was taken.(b) Array was incubated at growth conditions then glutaraldehyde fixation (see Materials and Methods section) of the sample was carried out.(c) Array was dried at room conditions after deposition of cells (nongrowth control) (d) Array was prepared as a nongrowth control, then glutaraldehyde fixation was carried out.

Figure 4 :
Figure 4: SEM images of cantilevers demonstrating accuracy of drop positioning and cell distribution after 24 h incubation at growth conditions.The width of cantilevers is 100 µm in each case.Ten drops of cell suspension were deposited at the front end of the cantilevers.(a) Imprints of series of droplets on agarose-coated cantilever, after growth of P. aeruginosa.(b) Ten perfectly spotted drops of S. aureus (2.92×10 8 cells/mL, 10% LB) cells on an agarosecoated cantilever.One spot was deliberately deposited off-position.(c) S. aureus (3.28 × 10 8 cells/mL, LB) growth on cantilever surface.(d) P. aeruginosa (1.84×10 8 cells/mL, 10% LB) growth on cantilever surface.Cells are growing over the clamping point of the cantilever.

Figure 5 :
Figure 5: SEM images of cantilevers showing C. albicans propagation on the surface after 24 h incubation at growth conditions.The width of cantilevers is 100 µm in each case.Ten drops of cell suspensions were deposited at the front end of the cantilevers.(a, c) C. albicans (1.36 × 10 6 cells/mL, 10% ME) before growth (a) after growth (c).(b, d): C. albicans (10 6 cells/mL, ME) before growth (b) after growth (d).

Figure 7 :
Figure 7: SEM images of biofilm-like structures of C. albicans (a) and S. aureus (b) on a cantilever surface.C. albicans cells were deposited in 10% ME medium (10 6 cells/mL) and incubated at growth conditions for 24 h.S. aureus cells were deposited in 10% LB medium (2.92 × 10 8 cells/mL) and incubated at growth conditions for 24 h.
) and 4(d), cells were observed outside the spotting area after growth of the culture which cannot be explained by an off-position droplet deposited during inkjet printing (Figure 4(b)).Concentric imprints caused by individual droplets on the agarose layer are often visible on SEM images showing an accurate centre position of the droplets (Figure 4(a)).The tested C. albicans (Figure 5) or S. aureus (Figure 4(c)) cannot actively move away from their original position while E. coli or P. aeruginosa (Figure 4(d)) are motile bacteria enabling cell propagation especially if a thin liquid layer is available.This propagation is more likely if 100% culturing medium is used for microbe deposition instead of with diluted medium (10% LB or 10% ME).Dilution of the growth medium decreased the growth rate of microbes.In the case of C. albicans, the use of less cells per droplet (5-10 cells/droplet) in full medium still resulted in the complete coverage of the cantilever surface with cells (Figure 5(d)) when compared to the non-growth control (Figure 5(b)).Spotting 50-100 cells in diluted medium (Figures 5(a), 5(c)) did not allow the organism to grow out of their initial spotted position.In several experiments, four microbial strains were tested in terms of viability after inkjet printing and grown on the surfaces of microcantilevers.

Table 1 :
Calculation of drop volume after deposition of 180 droplets of microbe suspensions.>95% RH) at 35 • C. For nongrowth controls, they were left to dry completely at room conditions.High humidity during growth testing was provided by a small water reservoir placed closed to the arrays in a closed Petri dish.