The extremely radioresistant Gram positive bacterium Deinococcus radiodurans strain UWO 288 was selected as template for sol-gel imprinting due to its coccoidal cells organized in pairs and tetrads [35]. Beside radioresistance, D. radiodurans has another interesting feature relevant to the imprinting process, namely its outer membrane structure containing a regular surface protein array (RS) or so-called hexagonally packed intermediate layer (HPI) [42,44]. According to Thompson and Murray [43] blebs from the outer membrane of this bacterium are shed as large vesicles from approximately 5% of the cell population during growth. Elution of entrapped D. radiodurans cells into thin SG films (∼1 μm), casted on microscope slides, revealed paired/tetrad like cavities similar to D. radiodurans cells morphology (Figure 1). Figure 1 (A, B and C) shows the morphology of D. radiodurans cells entrapped into sol-gel film observed with different microscopes. Interestingly, it was found that sole Gram staining procedure is enough to cause elution of the imprinted cells (at the moment under study to comprehend this phenomenon) leaving clear paired or tetrad cavities as observed with SEM (Figure 1D).

Sol-gel imprinted films were also subjected to acridine orange (AO) staining to be used with epifluorescent microscopy (Figure 2). AO staining of cavities formed following elution of imprinted D. radiodurans revealed staining of residual components entrapped in a sol-gel layer (Figure 2A). As already mentioned earlier, these components are apparently membrane blebs shed by this bacterium during the imprinting process. A similar phenomenon was observed with other bacteria such as Escherichia coli CN₁₃, Sphaerotilus natans and Bacillus subtilis (data not shown). In order to test the affinity of D. radiodurans planktonic cells in suspension towards imprinted SG film, an imprinted probe was immersed into a fresh D. radiodurans suspension (Figure 2B) against a non-imprinted sol-gel film as control (Figure 2C). The imprinted SG film was selectively covered with adsorbed D. radiodurans only onto imprinted sites compared with control that showed only few sparse non-specific adsorbed cells (Figure 2C).

The imprinting specificity for D. radiodurans cells was further substantiated by SEM micrographs (Figure 3). Imprinted SG film exposed to fresh D. radiodurans suspension clearly revealed adsorption and location of these bacteria pairs/tetrads solely in formed cavities.

As already mentioned, it was assumed that following the imprinting procedure a number of membrane surface components of entrapped/imprinted bacteria are left in the sol-gel matrix. In our opinion, these molecular components facilitate adsorption/attachment of free cells in suspension onto imprinted SG films. A schematic potential explanation of this process is presented in Figure 4. Briefly, a morphological defined bacterium (e.g., rod shaped) is entrapped into a sol-gel thin film (approx. 1–1.5 μm) partially or completely (Figure 4, first stage). Some of the SEM micrographs revealed bacterial cells dipped into the SG layer and covered with SG material as a very thin cap, while others were just sunk without any cover layer. Release of entrapped bacteria revealed, beside cell morphology, certain membrane components left entrapped into the SG formed cavity (Figure 4, second stage). It should be mentioned, that SG cover of entrapped bacteria during imprinting process is very fragile and it is easily removed by elution procedure leaving a clear morphological imprint. Based on these characteristics, target microorganisms are easily adsorbed onto an imprinted probe for recognition and further analysis, while discriminating between various morphologies and cell surface components (Figure 4, stage 3).

D. radiodurans residual bacterial surface components left in SG films were further analyzed electrophoretically for the presence of protein/glycoproteins against non-imprinted SG film and whole cell lysate (Figure 5).

Controls were: whole cell (diluted to the equivalent total protein concentration as imprinted sample), and a negative control of non-imprinted dissolved SG films. The imprinted SG films revealed protein bands corresponding to whole cell lysate, but at lower intensity. As expected, no protein bands were seen with non-imprinted SG films. These quantitative results, clearly confirmed the previous microscopic observations (Figure 2) that residual membrane proteins had been entrapped in the SG film imprinted cavities.

In order to assess the versatility of whole cell imprinting method in sol-gel films, bacteria such as E. coli CN₁₃, S. natans, B. subtilis and oocysts of the protozoan parasite Cryptosporidium parvum were used as templates (Figures 6–9). In all imprinted SG films, a well defined morphology of these microorganism species was visible, including molecular fingerprints left by imprinted cells. This characteristic can be clearly seen in Figure 9B, where Cryptosporidium parvum oocysts were stained with FITC antibodies and examined with an epifluorescent microscope. As these antibodies are specific to oocysts’ outer layers, empty imprinted cavities were stained only circumferentially revealing outer layer residuals entrapped into SG film. Exposure of imprinted SG films revealed excellent reinclusion of planktonic cells in suspension only onto formed cavities without any nonselective adsorption on non-imprinted sites. Non-imprinted SG control films revealed extremely low adsorption (close to zero) of the different microorganisms that may be explained by elevated hydrophobicity of prepared films. In spite of SG-films hydrophobicity, subsequent to imprinting process, cells adsorption ability increased dramatically providing the evidence that template cavities changed their surface chemistry due to residual outer layer entrapment during the process as already shown [12].

An interesting imprinting feature was observed with rod shaped bacteria, in contrast with spherical ones (e.g., D. radiodurans). Geometrically obvious, spheres (cocci) entrapped into SG films will form the same cavity shape no matter the orientation of their deposition, however rod shaped bacteria can form two shaped cavities: rectangular and square, depending on SG film deposition of rod bacteria (see Figure 8A). However, new planktonic bacteria reinclusion revealed correct redirection of rods according to imprinted outline (Figure 8B).

To test sensitivity, adsorption of bacteria in suspension (D. radiodurans and E. coli CN₁₃ separately) onto certain imprinted SG film area, was measured (Figure 10). At 10³ to 10⁴ CFU/mL a 10% recovery of these bacteria was observed which seems to be the threshold of this method sensitivity. Increasing target planktonic bacteria numbers improved adsorption onto SG films to approx. 90%. Adsorption volume of 100 mL was slightly more sensitive compared to 5 mL. Future studies will be performed in order to increase sensitivity through better contact between planktonic cells and imprinted film.

To determine selectivity, two experiments were carried out: (1) Differentiation between different bacterial morphologies-in which D. radiodurans SG imprinted films were exposed to a mixture of D. radiodurans, E. coli CN13 and B. subtilis in suspension (10⁸ CFU/mL each) for 30 minutes. The SG film screened under SEM and epifluorescent microscope revealed sole adsorption of D. radiodurans cells onto the formed cavities (data not shown); (2) Differentiation between same bacterial morphologies-in which E. coli CN₁₃ SG imprinted film was exposed to B. subtilis in suspension (10⁸ CFU/mL) for 30 minutes. Both bacteria have a rod like morphology with a size difference (B. subtilis is larger in comparison with E. coli). The exposed SG films revealed adsorption of E. coli CN₁₃ cells only, as observed by specific fluorescent antibodies staining with no adsorption onto non-imprinted control SG films (data not shown). B. subtilis imprinted films exposed to E. coli did not reveal any adsorption of the last (data not shown).

Bacteria such as D. radiodurans, S. natans, E. coli CN₁₃, B. subtilis and oocysts of protozoan parasite Cryptosporidium parvum were used as imprinting templates due to their morphology and size variations.

D. radiodurans UWO 288 bacterium (ATCC13939) (kindly donated by Prof. R.G. Murray, University of Western Ontario, Canada) has a morphology of spherical cells (Ø = 0.5–3.5 μm) organized in pairs or tetrads. D. radiodurans was grown in TGY medium at 36 °C for 36–48 hours as already described [35].

S. natans bacterium (from our collection, primarily isolated from corrosion deposits of a cooling water heat exchanger at Haifa’s Refinery Plant) is a Gram negative (G-) rod filamentous iron oxidizing bacteria with a cell length from 2 to 12 μm, depending on growth conditions. Sphaerotilus natans was grown in Winogradsky medium containing (g/L): NH₄NO₃ (0.5), NaNO₃ (5), 0.66 K₂HPO₄·3H₂O (0.66), MgSO₄ (5), CaCl₂ (0.1), ferric ammonium citrate (10) and incubated at 28 °C for 4–5 days. Under these conditions, bacterium length was 2–2.5 μm.

E. coli CN₁₃ (a nalidixic acid resistant G- E. coli C strain) [36] and B. subtilis (G+) (Difco,USA) with cell sizes of 0.5 × 1.5 μm and 0.7 × 3 μm, respectively, are both rod-shaped bacteria selected in order to assess selectiveness towards different bacterial species with similar morphology. E. coli CN₁₃ was grown on R₂A agar (Acumedia, USA) at 37 °C for 48 hours and B. subtilis in nutrient broth (Difco, USA), both shaken (50 rpm) at 30 °C for 24 hours. All experimental bacterial species stocks contained 10⁷ to 10⁹ CFU/mL.

Protozoan parasite Cryptosporidium parvum oocysts with spherical pea-like morphology and an average size of 5.5 μm were purchased from Waterborne Co. (New Orleans, USA). Stock oocysts, (10⁸/mL) were washed three time with sterile saline before use.

Sol-gel films (SG) were prepared on standard microscope glass slides previously cleaned according to a modified procedure [16]. Briefly, glasses were soaked for 20–30 min successively in 1:1:1 acetone-ethanol-chloroform mixture; soap solution; Piranha solution—H₂SO₄:H₂O₂ (4:1), then thoroughly rinsed in double-distilled water followed by drying at 100 °C for 2–4 hours.

The imprinting procedure was as follows: aged sol-gel stock solution (0.5 mL) was added to bacterial suspension (3 mL, 10⁹ to 10¹⁰ CFU/mL) and stirred for 8 to 10 minutes. Subsequently a glass slide was drop-coated with 0.25 mL of this experimental mixture. In order to obtain a thin and uniform film, the glass slide was instantly positioned vertically to drain excess solution. With Cryptosporidium parvum oocysts, 1 mL suspension (1.2 × 10⁷ oocysts) was mixed with 166.7 μL of the sol-gel stock solution stirred for the same time interval and slide spread. Slides coated with sol-gel films containing bacteria/oocysts were placed into a desiccator and dried over night.

Removal of immobilized bacterial cells from dried sol-gel film was performed by soaking the coated slides for 30–40 min in 96% ethanol and washed with sterile double-distilled water. Slides were further Gram stained (successively 3 min in crystal violet, iodine, 70% ethanol, safranin for 3 minutes each) (Bactolab Diagnostics, Israel) to visually detect cavities formed by entrapped bacteria. It is important to note that in many cases immobilized bacterial cells dropped just after Gram staining procedure without ethanol treatment. Removal of Cryptosporidium oocysts was done by exposure of sol-gel coated slides to a NaCl solution (10% w/v) for 4 hours. Control glass slides were coated with sol-gel films without bacteria (as non-imprinted films) and rinsed with suspension media (0.85% NaCl + 0.2% Tween 80).

Readsorption/reinclusion tests were carried out with SG film coated slides (imprinted and non-imprinted) by exposure to 10⁷–10⁹ CFU/mL of an appropriate bacterial suspension (5 and 100 mL volume) for 30 min with gentle stirring. Following exposure, slides were gently washed with sterile triple-distilled water and then examined under an epifluorescent microscope (×1,000, Zeiss Axiolab, filters 450–490 excitation and FT510-LP520 emission), confocal laser scanning microscope (CLSM) (DM5500 Q, Leica, Germany) at ×630 magnification, SEM (LEO-982 Geminate FEG-HRSEM, Germany and a dual-beam focused ion beam FIB, FEI Strata 400-S, USA) and AFM (AFM/STM, PicoScan, Molecular Imaging, USA). Slides for epifluorescence microscopy were previously immersed in a glutaraldehyde solution (2%) for 10 minutes and subsequently for additional 10 minutes into acridine orange (AO) solution (0.01% AO in 0.1 M phosphate buffer, pH 7.0, w/v) before observation. SEM samples were dehydrated and critical-point dried, fixed on aluminum stubs, sputter coated with gold-palladium and examined with a scanning electron as already described [37]. AFM samples were performed at 20 °C in phosphate buffer (10mM, pH 7) KH₂PO₄ and contact mode images were taken in constant force mode with applied force of ∼1 nN, and scan rate between 1 and 2.5 Hz [38].

Cryptosporidium parvum oocyst enumeration in suspension was determined by monoclonal antibodies staining and epifluorescence. Briefly, 500 μL of oocyst suspension was placed in an Eppendorf standard microtest tube (2 mL volume). To this suspension, a volume of 250 μL fluorescein isothiocyanate (FITC)-conjugated monoclonal antibody (Crypto-A-Glo Kit, Waterborne Co., USA) was added and the mixture was incubated for 30 min at 36 °C. Following incubation, the stained sample was subjected to three rinses with phosphate buffered saline (pH 7.0, 0.1 mM) and centrifugally washed (4,000 × g) for removal of unattached antibodies. Then, 100-μL aliquots were placed on six-well chamber slides and fixed with 100% methanol. After methanol vaporization at 36 °C, each well containing stained oocyst sample was covered with 35 μL of mounting medium (2%) (1,4-diazabicyclo-2.2.2-octane, DABCO, Aldrich, USA) in glycerol and a coverslip. The entire well was scanned with an epifluorescence microscope (AxioLab, Zeiss) at ×1,000 magnification. Imprinted and reincluded oocysts were directly stained with FITC-conjugated monoclonal antibody, rinsed with phosphate buffer (0.01 M, pH7), dried and observed with epifluorescent microscope as mentioned above. Pea-like oval particles with suture on its envelope, 4 to 6 μm in diameter and with brilliant apple-green fluorescence glow, were identified as C. parvum oocysts.