Rapid detection of pathogens using lyotropic liquid crystals.

Lyotropic liquid crystals play an important role in many biological environments, such as micelles, liposomes, and phospholipid bilayers of cell membranes. In this work, we explore the performance of lyotropic liquid crystals as biosensors for macromolecules, proteins and whole microorganisms in hydrophilic media, i.e., the natural media where these specimens exist. The aim is to detect specific targets employing simple, unpowered sensors that can be used in the field, with minimum additional equipment. A number of different structures have been explored. The novelty in this work is the inclusion of a new optical effect, flow enhanced amplification, that allows for the semiquantitative detection of microscopic targets in lyotropic liquid crystal cells using the naked eye only.

microwaves [5]. Additionally, LCs show an extraordinary sensitivity to external electric and magnetic fields, as well as changes in boundary surface conditions which may be transduced into optical signals, visible between in polarized light.
Liquid crystal biosensors usually take advantage of this sensitivity. A typical LC biosensor detects the presence of small molecular compounds by adsorbing them on an aqueous/lipidic interface. LCs can also detect microparticles of a certain size. In either case, the visual signal usually arises from disordered LC regions of an otherwise ordered LC, due to the presence of the target. The ordered LC can be made dark between crossed polarizers, while disordered regions show up as bright colored spots.
An interesting feature of LC sensors is amplification [6]. The disorder propagates to an area considerably larger than the original target size, making it possible to observe submicron-or micron-sized targets with a standard polarized microscope; in some cases, amplification can be high enough to allow defects to be visible at naked eye. Different LC materials and procedures have been proposed for biosensing of DNA [7], enzymes [8], antigens [9] and even full microorganisms like viruses and bacteria [10], as well as biologically relevant compounds such as environmental pollutants and gases [11].
Most proposed LC biosensors are based on nematic LCs [12] and other thermotropic liquid crystals, like cholesterics [13,14], blue-phase [15], and smectics [16]. All these LCs feature a common issue, their immiscibility in aqueous media. As many biological samples are solved or suspended in water, biosensing with thermotropic LCs requires alternative solutions to achieve interactions between targets and LCs [17]. Free-standing LC droplets [18,19] or surfaces suspended in sub-mm grids [20] have been proposed to procure an interface for interactions between the hydrophobic LC layer and the aqueous sample being examined [21]. The LC spontaneously orients homeotropically (perpendicular) to the freestanding surface for surface tension. Adsorption of targets on the interface modifies such orientation, and the change can be observed in a microscope between crossed polarizers.
A different straightforward approach to overcome these difficulties is to use lyotropic, rather than thermotropic, liquid crystals. Lyotropic liquid crystals (LLC) are stable and soluble in aqueous media since they mostly derive from amphiphilic molecules. LLC phases exist in a range of temperatures, and within a range of concentrations of the material in a solvent, usually water. LLCs are ubiquitous in biological media, forming for example the lipidic bilayer of cell membranes. However, their interest as materials for technical applications is largely overtaken by thermotropic liquid crystals, particularly calamitic nematics. Nevertheless, water based lyotropics are the preferred alternative in some biological applications, in general, those whose actual working conditions imply the use of hydrophilic media [22]. LLCs have been proposed, for example, as biomimetic vehicles for delivery of sparingly soluble drugs or medical contrast agents [23].
The use of LLCs for biosensing of microparticles and microorganisms is known for some time [24]. In this case, the aqueous solvent where the targets are suspended is mixed with the LLC -taking into account the concentration range where the desired lyotropic phase is formed. A cell is formed by two transparent surfaces (glass, rigid polymers) with a small gap (some µm) between them. The inner cell surfaces are conditioned in advance to induce a certain orientation on the LLC. The element can be integrated in a microfluidic system. In the absence of targets, surface preconditioning makes the LLC to orient as the surfaces dictate. If targets are present, they may induce defects in the overall orientation; these defects can be visualized between crossed polarizers in a microscope.
Visualizing director defects is possible when the object in the medium is larger than the extrapolation length [17] where K is the distortion elastic constant and W is the anchoring strength of the LLC with the microparticle. Typically [10], K is in the range of 1-10 pN, while W is in the range of 10-1000 µJ/m 2 ; therefore b is in the µm-tenths of µm range, which is appropriate for detection of bacteria. The technique has become commercial [25], allowing fast detection of microorganisms. Nevertheless, specific sensing cannot be achieved with this procedure. To achieve specific sensing it is necessary, after adding the alignment layer, to functionalize at least one of the inner surfaces of the cell with some reagent specific for a given target, e.g., an antibody, and to include a washing process in the sampling protocol. Bound microorganisms are then revealed as defects in alignment. An alternative procedure is to coat only one of the plates with the alignment layer and to employ the other plate as functionalized surface. This requires the alignment anchoring to be strong enough for the sample to be aligned with just one surface.
In this work, the optimization of LLC biosensors for selective detection of microorganisms has been undertaken. Several aspects have been considered: seeking alternatives of antibodies for selective detection, providing good alignment conditioning for LLCs, and increasing the amplification, with the ultimate goal of performing the test without optical aids, i.e. with naked eye.

Experimental
The optimization of LLC biosensors includes the selection of binding agents, manufacturing of alignment layers, and enhancing amplification. Anyhow, the aim of this work is to evaluate the performance of LLCs in biodetection; this will wrap up the whole set of activities.

LLC cell manufacturing
LLC cell geometry is the same as in standard LC cells with a few differences. Two parallel glass plates separated 10 µm were assembled using a photocurable adhesive. Opposite ends of the cell were left open to allow filling by capillarity or microfluidics. Spacers were only added to the adhesive layer framing the cell, since any microparticle inside the cell would lead to a false positive.
One of the inner surfaces was functionalized either with an aptamer or an antibody so that a specific pathogen would be bound to the surface. The other surface was used to induce an alignment to the LLC. The alignment was homogeneous -i.e., parallel to the surfaces-and oriented along the flow direction when the cell is filling up.

Lyotropic liquid crystals
Two LLCs have been chosen for the experiments: Sunset Yellow FCF (disodium 6-hydroxy-5-[(4-sulphonatophenyl)azo]naphthalene-2-sulphonate) and Cromolyn (sodium cromoglycate) (Sigma-Aldrich). These two materials are among the most studied LLCs. Both are watersoluble salts of organic acids, and possess a nematic (N) phase at room temperature in the concentration range 27% < N < 35% w:w for Sunset Yellow (SSY) and 13% < N < 17% w:w for Cromolyn. Despite Cromolyn is the most used and studied, the work ultimately focused on SSY since it showed several advantages: a wider nematic range, a lower variation of this range with temperature, and a lower viscosity, which resulted essential for dynamic studies shown below.

Aptamers
Most of the work on LLCs has been carried out employing aptamers [26]. Specific aptamers (80-120 bases) for Meningococcus (the original target), and foodborne pathogens such as Salmonella, Listeria, and Campylobacter have been prepared within the project by the iterative SELEX method [27]. Legionella, prepared outside the project, has been employed as well. Once sequenced and replicated, aptamers were bound to one inner surface of the LLC cell, making the surface functional to specifically trap a target -in our case, a pathogen. Details of fabrication and characterization of these aptamers are beyond the scope of this work.

Alignmen
Finding a surf Standard poly nor polyamid suitable align published else

Amplifica
Increasing the affected, henc devices. Inde threshold, ho threshold pro reduces to a method, and reduces as we Consequen significantly accomplished

Static pre
We have test inorganic lay evaporation re was achieved best option. T all samples w After deve pathogens usi Detection has water), and bo not intended screenings fo through a full nt layers face conditioni yimides employ des (e.g., Nylon nment layers f ewhere [28,29] ation e amplification ce giving posi eed, larger are owever, above duces a signif mere low-cos allowing the u ell. ntly, the main higher. Nevertheless, our rates of false positives with aptamers are still unacceptable for actual working biosensors; this issue is under study. The results of this work have been obtained employing either antibodies or artificial samples with just one, or a few, kinds of microorganisms. Figure 2 shows preliminary results testing the detection procedure. A 10µm thick SSY/SiO 2 cell was filled with a clean solution and with a doped solution. No surface functionalization has been performed. The targets were silica microspheres of 2.5 µm diameter. Targets are clearly visible as bright spots in a dark field (the residual brightness of the undoped sample is due to the automatic gain of the camera; actually both pictures are equally dark). The visible field width is about 1 mm. The estimated size of the spots is 20-25 µm; therefore, the amplification factor is about 8-10. Similar results have been published by other authors [10].
The targets, or the agglomerates in the case of functionalized surfaces, produce a disorder in the neighboring LC molecules since they must adapt to the new anchoring condition imposed by the target [24]. Disorder propagates a certain distance while the influence of the anchoring energy fades out. Indeed, the distance would depend on several external parameters, such as LLC concentration and temperature, the solution's ionic strength and, obviously, the physical characteristics of the targets and their size. However, for a given target and a specific set of experimental conditions, the distorted area, hence the amplification factor, is apparently fixed.

Measurements under flow: wake formation
In the above scenario the LLC organization and orientation are static, i.e., performed once the cell has been filled up with the mixture of target and LLC solutions. The situation is drastically modified if the LLC is studied under a controlled flow. If the sample is homogeneous, and the imposed orientation is parallel to the flowing direction, the LLC spontaneously orients upon flowing. This phenomenon [30] was known since the dawn of liquid crystals and is due to the anisotropy of LC viscosity. As seen in is that dynam experimental 600, i.e., almo Figure 4   The cell on the left side shows the long wakes, as Legionella pneumophila is detected where it has been bound to the surface, while the cell on the right shows little or no wakes given that Mouse ABs do not bind Legionella. The few wakes observed in the negative cells are attributed to surface defects or microparticles and non-specific interactions during functionalization of the surface. Anyhow, there is a relevant difference between both samples to the naked eye. The size of the active area is 18 × 8 mm; the close-ups on the right side of the figure show the central area of the cells (6 × 5mm).

Conclusions
A new procedure for detection of macromolecules and microorganisms has been described. The procedure is simple, inexpensive and semiquantitative. Different targets can be explored, even in the same cassette. The procedure may be useful for point-of-care fast tests carried out by non-specialists. The remarkably high amplification factor achieved in dynamic monitoring allows in most cases to detect the target by naked eye, further simplifying the required experimental setup.
Lyotropic liquid crystals reveal as a strong alternative for liquid crystal-mediated biosensors, especially in those cases were a hydrophilic medium must be used, like milk or animal waste in food chains or sea water in aquaculture. They are also a good candidate for biosensors utilized in body fluids like saliva, lymph, urine, plasma, etc.
At present, the tested procedures have shown a significant number of false positives, while false negatives are nearly absent in correctly functionalized cells. It may be argued that false positives are less relevant than false negatives. Indeed, the test can be used in many applications as a simple screening to select those samples or patients that should be analyzed in an eventual control at a healthcare center. In these cases, a certain number of false positives, if not excessive, can be acceptable, while false negatives must be avoided in any circumstances.

Disclosures
The authors declare that there are no conflicts of interest related to this article.