The surface of polytetrafluoroethylene (PTFE) was modified using plasma immersion ion implantation (PIII) with the aim of improving its ability to immobilize yeast. The density of immobilized cells on PIII-treated and -untreated PTFE was compared as a function of incubation time over 24 h. Rehydrated yeast cells attached to the PIII-treated PTFE surface more rapidly, with higher density, and greater attachment strength than on the untreated surface. The immobilized yeast cells were removed mechanically or chemically with sodium hydroxide and the residues left on the surfaces were analysed with Fourier transform infrared spectroscopy-attenuated total reflection (FTIR-ATR) and X-ray photoelectron spectroscopy (XPS). The results revealed that the mechanism of cell attachment on both surfaces differs and a model is presented for each. Rapid attachment on the PIII-treated surface occurs through covalent bonds of cell wall proteins and the radicals on the treated surface. In contrast, on the untreated surface, only physisorbed molecules were found in the residue and lipids were more highly concentrated than proteins. The presence of lipids in the residue was found to be a consequence of damage to the plasma membrane during the rehydration process and the increased cell stress was also apparent by the amount of Hsp12 in the protein residue. The immobilized yeast cells on PIII-treated PTFE were found to be as active as yeast cells in suspension.
Ethanol is becoming an increasingly popular liquid biofuel and is produced industrially in large quantities from agricultural feedstock (for example, cane sugar and corn starch) using fermentation processes. Batch fermentation using yeast is common in current ethanol production [1,2]. Enzyme poisoning , labour costs of batch renewal, water usage and waste water management have been identified as problems in batch fermentation processes. A continuous process maintained in a steady-state condition can offer a higher yield, lower risk of contamination, easy operation and hence lower cost. The economics of continuous processes are improved by immobilizing the yeast cells on a solid support structure . Cells immobilized on a suitable carrier show advantages for ethanol production over cells freely floating in solution including yield, ethanol tolerance and enzyme reuse [4–6]. In continuous fermentation, there are requirements for yeast cells to remain attached in the flow of liquid and therefore remain in the fermentor as well as to tolerate an adequate constant level of ethanol fermentation product. Attempts have been made to immobilize yeast cells on polymers, glass beads, and stainless steel wire or to capture them in polymer-gels [7–10].
In a review by Cassidy et al. , yeast immobilization was categorized into six common methods: flocculation, physical adsorption to surfaces, covalent bonding to a carrier, cross-linking to cells, encapsulation in a polymer-gel and entrapment in a matrix. Of these, only physical adsorption, covalent bonding, encapsulation and entrapment enable immobilization to a surface. Encapsulation in polymer-gel showed a good yield of ethanol. Nagashima et al.  reported pilot plant operations with yeast cells immobilized in calcium alginate for 4000 h with a productivity of 2.4 kl ethanol per day. The disadvantages of this and other encapsulation methods arise from the need for nutrients to diffuse into the gel and for ethanol and CO2 to diffuse out of the gel, frequently causing damage to the gel and cell loss . Physical adsorption onto a surface is a simple method of immobilization in which yeast cells form a biofilm on the supporting surface. This surface adsorption can overcome the disadvantages of the encapsulation method but typically the adhesion of the cells is not strong enough to resist the flow rates required in a continuous process. There are reports of work on the immobilization of yeast on natural materials with a rough surface as a shelter for yeast cells and their progeny to improve their resistance to the flow [6,13]. Cell adhesion by the formation of covalent bonds between molecular components of the cell exterior and the immobilizing surface promises a more stable and secure immobilization than physical adsorption. Jirku & Turkova  reported covalent immobilization of yeast cells on modified hydroxyalkyl methacrylates with glutaraldehyde as a coupling linker between the support and the cells. They reported difficulties associated with the large number of steps in the linker process. Also, a minimum of 4 days of incubation was required to get complete coverage with yeast cells.
Here, we present a new method of immobilizing yeast cells on polymer surfaces. We investigate cell attachment of rehydrated Saccharomyces cerevisiae on polytetrafluoroethylene (PTFE) surfaces before and after treatment with plasma immersion ion implantation (PIII). PIII achieves ion implantation into the surface by accelerating ions from plasma in which the sample is immersed. Such ion implantation creates active radicals in PTFE , and surfaces were observed to strongly immobilize proteins [16,17]. A recent report  shows that this immobilization occurs through covalent bonds formed with mobile unpaired electrons associated with the radicals. To investigate the immobilization of yeast cells on the treated and untreated surfaces, we used Fourier transform infrared spectroscopy-attenuated total reflection (FTIR-ATR), X-ray photoelectron spectroscopy (XPS) and mass spectroscopy to study the residues on the surface after cell removal. Immobilization of rehydrated cells on polystyrene (PS) was used to confirm the results obtained on PTFE. A study of the immobilization of cells inoculated from the dry culture was also undertaken. A model for the mechanism of attachment that explains the observations is presented. Finally, batch fermentation was conducted to compare the activity of the immobilized cells on PTFE with cells in suspension in terms of producing ethanol.
2. Material and methods
2.1. Plasma immersion ion implantation treatment
PIII was carried out in an inductively coupled, radio frequency nitrogen plasma. Samples used were PTFE film 0.2 mm thick (LS 362817 VCS, Goodfellow Cambridge Ltd) and PS film 0.25 mm thick (LS 351053, Goodfellow Cambridge Ltd). The samples were placed on the surface of a conductive sample holder, electrically connected to a conducting mesh placed approximately 5 cm in front of the holder. The sample holder and mesh were placed in the chamber and the nitrogen pressure was maintained at 2 × 10−3 torr during operation. A radio frequency plasma was generated by applying radio frequency power of 100 W. During treatment, the polymer surface was modified by a flux of nitrogen ions to ion fluences of 2.5–5.0 × 1015 ions cm–2. A bias of 20 kV was applied to the sample holder and mesh in pulses of 20 µs duration at a frequency of 50 Hz over total treatment times of 200 or 400 s.
2.2. Water contact angle and surface energy
The wettability of the samples 1 h after PIII treatment was measured by Kruss contact angle DS10 equipment using the sessile drop method. Surface energy was calculated using the Owens–Wendt–Rabel–Kaelble method with contact angles of water and diiodo methane. Results were taken as an average of five measurements on each sample.
2.3. Yeast cell preparation
Rehydrated cells: 0.1 g of dry S. cerevisiae (YSC2, Sigma Aldrich) was rehydrated with 50 ml milli-Q water and centrifuged at 2000g for 5 min. The pellet was washed again by suspending in 50 ml milli-Q water and re-centrifuging for another 5 min. The pellet was then suspended in 50 ml PBS 0.1 M to obtain a cell density of approximately 5–6 × 107 cells ml–1. This washing step was undertaken to ensure that no free protein was left in the incubation suspension. (Protein assays were carried out by FTIR-ATR and sodium dodecyl sulphate (SDS) polyacrylamide gel electrophoresis (PAGE) with Sypro Ruby stain. The minimum amount of protein detectable on the surface by amide A, I and II lines is 0.083 mg m–2 and is limited by the noise level of the spectra. This amount corresponds to a coverage of 1.7 per cent of the surface in the case of a 5 nm thick protein layer. For SDS-PAGE with Sypro Ruby stain, the detection limit is 0.0125–0.05 μg ml–1.)
Cultured yeast cells were prepared as described by Guillemot et al.  with some modifications. A few dry yeast cells were added into 100 ml of YEPD broth in a 500 ml Erlenmeyer flask and incubated overnight at 30°C and 200 r.p.m. One millilitre of this culture was transferred to another 500 ml Erlenmeyer flask with 100 ml of YEPD broth and incubated for 48 h at 30°C and 200 r.p.m. After incubation, cultured cells were centrifuged and washed twice with milli-Q water as described already and finally dissolved in PBS and adjusted to get the same cell density.
The polymer film after PIII treatment was stored in a laboratory environment for an hour and cut into smaller size samples (1.6 × 1.1 cm2) for incubation with yeast. Untreated polymer film was also cut into the same size samples for incubation. Two millilitres of yeast suspension was transferred to wells of a 12-well PS plate containing samples (treated and untreated) and incubated for 15, 30, 60, 120, 240, 360 and 1440 min at room temperature (regulated at approx. 22°C). Loosely bound cells on samples were then removed by dipping the samples twice in each of 10 beakers containing milli-Q water and the remaining cells were fixed with 2 ml of 3 per cent formaldehyde for 30 min. After fixation, samples were washed twice with 2 ml milli-Q water and stained with crystal violet (0.4%) for 30 min. Extra crystal violet was washed off four times with 2 ml milli-Q water each and samples were observed under an optical microscope (Axioplan 2 Imaging, 10× magnification). Nine images were taken on each sample at different positions and these images were subjected to cell counting using ImageJ software (v. 1.43). Results presented were the average of three different replicates.
2.5. Characterization of yeast cell residues on polymer surface after cell removal
This experiment was conducted to investigate the mechanism of adhesion of S. cerevisiae on polymers by studying the residue left on polymer surfaces after the cells were removed. Two PIII treatment conditions (400 s for PS and 200 s for PTFE) were tested with short and long incubation times of 1 and 24 h, respectively. Two methods were used to remove yeast cells:
— attached cells on polymers after incubation were rubbed off using gloves (Nisense, powder free nitril examination gloves) and samples were rinsed again with running milli-Q water and dried for one day in a desiccator. Samples were checked by optical microscopy to ensure that no cells remained on the surfaces. The PTFE samples were then subjected to FTIR-ATR, XPS and mass spectroscopy analysis. The PS samples were analysed by FTIR-ATR to confirm the results; and
— attached cells on polymers were also removed by soaking in sodium hydroxide (3%) solution while shaking at 120 r.p.m. for 24 h. Sodium hydroxide is a strong cleaning agent that can dissolve oil, fat and protein by interfering with physical adsorption forces but not with covalent bonds. After this treatment, samples were rinsed with milli-Q water and dried in a desiccator before being analysed by FTIR-ATR.
FTIR-ATR analysis: spectra from the sample surfaces were recorded using a Digilab FTS 7000 FTIR spectrometer with germanium crystal ATR accessory (Harrick Inc., USA) and an incident angle of 45°. Five hundred scans were done on each analysis with a resolution of 4 cm−1.
XPS analysis: XPS (Specs-XPS, mode XP-50 High Performance Twin Anode with Focus 500 Ellipsoidal Crystal Monochromator and PHOIBOS 150 MCD-9 analyser) was used to characterize elements on the sample's surface. Binding energies (within a range of 50–1200 eV) of photoelectrons emitted upon irradiation with Kα X-rays produced by a dual anode were used to identify the elements present. Areas of element peaks were calculated using CASA XPS and divided by relative sensitivity factors for each element. The concentration of each element was calculated as an atomic percentage.
To investigate the mechanism of the protein attachment to PTFE surfaces, samples (untreated and PIII treated, 1 h incubation; untreated and PIII treated, 24 h incubation) were washed with 2 per cent SDS solution for an hour at 100°C. SDS denatures secondary and non-sulphide-linked tertiary structures of proteins and is used widely in biochemistry, especially in SDS-PAGE gels to separate proteins. The SDS washing protocol used in this paper satisfies the conditions established in previous studies [18,20,21] to test for the covalent attachment of proteins to surfaces. SDS breaks non-covalent interactions between proteins and the surface, unfolding proteins and releasing them to the solution. Only those proteins that are covalently bound to the surface will remain. XPS data of sample surfaces after SDS washing were compared with XPS data of surfaces before washing to determine the nature of protein attachment to the PTFE surfaces.
Mass spectroscopy analysis: rehydrated cells and cultured cells were incubated for an hour with PIII-treated PTFE. The protein residues on polymer surfaces after cell removal were digested by sequencing grade trypsin. The digested peptides were separated by high-performance liquid chromatography (HPLC) and were analysed by LTQ-FT Ultra mass spectrometer (Thermo Electron, Bremen, Germany). MS/MS spectra were searched against NCBI database using Mascot (v. 2.3). The exponential modified protein abundance index (emPAI) values of common proteins were obtained from the Mascot reports and normalized to the total protein emPAI amount to understand the relative abundance of each protein. emPAI values are used widely in proteomics to estimate the relative amount of protein [22,23]. Subsequently, the relative abundance of each protein was normalized to the total adsorbed protein amount using the FTIR-ATR amide I peak signal in order to quantify the relative amount of each adsorbed protein.
2.6. Activity testing
In order to confirm the viability of immobilized yeast cells on PTFE, batch fermentation by immobilized cells was compared with fermentation by cells in suspension. Two beakers each containing 200 ml of fermentation medium (glucose (150 g l–1), KH2PO4 (2 g l–1), (NH4)2SO4 (2 g l–1), MgSO4 × 7H2O (1 g l–1) and yeast extract (2 g l–1)) were autoclaved at 120°C for 20 min. Four pieces of PTFE (6 × 3.5 cm2) were incubated with rehydrated yeast for 15 min and washed as described above before loading into a stainless steel rack in a beaker with fermentation medium. The stainless steel rack maintains a gap between the PTFE sheets and keeps them vertical in the liquid. Cell density was counted on a small sample incubated at the same time to determine the total number of immobilized cells on the PTFE. The same amount of rehydrated yeast cells was added to the other beaker. The fermentation was carried out in an oven with temperature regulated at 30°C. One millilitre of fermentation medium was taken at the beginning of fermentation and twice every day thereafter up to 6 days for glucose (glucose assay kit GAGO-20 from Sigma) and ethanol (gas chromatography) analysis. At the end of the experiment, one PTFE sample was fixed with glutaraldehyde and osmium tetroxide and subjected to SEM imaging.
3.1. Comparison of rehydrated yeast cell attachment on untreated and plasma immersion ion implantation-treated polytetrafluoroethylene
After 15 min of incubation, a white layer of yeast cells was observed on the polymer surfaces. This layer was easily removed by dipping the samples in milli-Q water, leaving only a thin film of cells attached to the polymer surface. Figure 1 shows images of stained rehydrated yeast cells on PTFE after various incubation times. There is a significant difference in the density of adhered cells on the PIII-treated and on the untreated surfaces. On untreated PTFE surfaces, only a few cells were found in the first 15 min. More cells attached to the surface as incubation time increased but clumps were observed. The clumps may form as a result of cell moving during fixing and staining steps. From 6 h the number of clumped cells had reduced and no clumps were observed 24 h. Compared with untreated samples, PIII-treated samples had a more even distribution of yeast cells. High cell density, which was achieved after 15 min of incubation, remained approximately stable over the following 24 h.
Figure 2 shows a graph of the cell density as a function of time on the untreated and PIII-treated PTFE surfaces over a 24 h incubation. Cell attachment on the untreated surface increased in the first 6 h and saturated after 24 h while cell attachment on PIII-treated surface reached the highest density during the first 15 min of incubation and remained stable thereafter. Compared with untreated PTFE, PIII-treated PTFE can attach yeast cells with approximately twice the density. For treated PS surfaces a similar behaviour was observed although the cell density was somewhat lower than that observed on treated PTFE at 10 400 ± 3400 cells mm–2. As was true with PTFE, untreated PS attached a lower saturated cell density of 6200 ± 2200 cells mm–2 at all incubation times.
3.2. Surface characterization
It has been suggested that the wettability of a surface plays a role in cell attachment [24,25]. Klotz et al.  found that the number of Candida albicans attached on a polymer surface was proportional to the water contact angle of that surface. In the present experiment, water contact angles and surface energies of the treated and untreated PTFE were not significantly different while the PS showed a significant reduction in water contact angle with PIII treatment (table 1). Therefore, water contact angle does not appear to be the only factor in the different cell attachment observed on untreated and PIII-treated surfaces. The surface energy of untreated PTFE used in the experiment is 12.6 mN m–1, which is significantly lower than the surface energy of PTFE of 20 mN m–1 reported by Kondyurin et al. . An unchanged water contact angle after PIII treatment has not been observed previously in our group. Kondyurin et al.  used the same PIII treatment with PTFE 20 μm thick (Halogen, Perm, Russia) and found the water contact angle reduced from 120° to 90°. Bax et al.  used PTFE 0.1 mm thick (Goodfellow) and observed a water contact angle change from 114.4 ± 3° to 93 ± 3°. The resistance to change of the water contact angle of the PTFE used in our experiment is believed to be due to the presence of low-molecular-weight components in the material we are using (Goodfellow, 0.2 mm thick). Such low molecular weight components are mobile and diffuse to the surface from the bulk . After PIII treatment, the low-molecular-weight components diffuse to the surface and reduce the measured surface energy as observed previously in plasticized polymers . Although the PTFE materials used in the earlier experiments are too thin to allow accurate cell counting, the same trends of cell density on untreated and PIII-treated surfaces were observed as reported in this work.
The infrared spectra of PTFE before and after 200 s of PIII treatment are shown in figure 3. Compared with the untreated sample, the PIII-treated sample has new absorption lines at 1881 and 1723 cm−1, corresponding to C=O vibrations, and at 1657 cm−1, corresponding to C=C vibrations . These peaks indicate that the PTFE surfaces were carbonized and oxidized as a result of PIII treatment. The peaks at 3300 and 2922 cm−1 are attributed to O–H and C–H vibrations, respectively. Because they appear in both spectra, these are assumed to come from species adsorbed on the polymer surfaces. The peaks near 2400 cm−1 are associated with an overtone of C–F vibrations.
In order to investigate the interaction of cells with the surfaces, the residue remaining after the removal of all the attached cells was analysed by FTIR-ATR spectroscopy. Figure 4 shows the spectra after 1, 2 and 24 h incubations with yeast cells. The spectra shown are obtained by subtracting spectra taken before incubation. The presence of protein residue, on all samples, is confirmed by the strong peaks at 1650 cm−1 (amide I), 1540 cm−1 (amide II) and a broad, weaker peak at 3300 cm−1 (amide A). The intensities of the amide peaks are consistently stronger for PIII-treated samples than for untreated samples at the same incubation times (figure 4a–c). New lines are observed at 2920 and 2850 cm−1 (assigned to CH2 symmetric and asymmetric stretching vibrations); at 1742 cm−1 (assigned to the C=O stretch of ester carbonyl groups); and at 1471 cm−1 (interpreted as bending vibrations in CH2 groups). The CH2 and ester C=O groups are associated with the presence of lipids. These lines are stronger for untreated samples than for PIII-treated samples, showing that the adsorbed residue on the untreated samples contains a higher proportion of lipids. Table 2 summarizes intensities of the amide I (1650 cm−1), ester C=O (1742 cm−1) and CH2 (2920 cm−1) lines, normalized to the C–F peak at 1205 cm−1.
Our method of cell removal by rubbing the surface under running milli-Q water avoids contamination from chemicals but also has some disadvantages. Yeast cells on PTFE-untreated surface incubated for a short time can be rubbed off easily, but those with long incubation time are more difficult to remove. The strong rubbing action required may remove not only yeast cells but also yeast residues that are adsorbed on polymer surface. Therefore, we cannot compare the absolute amount of the residues on samples with different incubation times. However, the ratios between protein (amide I peak at 1654 cm−1) and lipid (1735 cm−1) on the same surface can be compared. A fourfold reduction of this ratio on untreated PTFE was observed as a function of incubation time. This ratio was 0.42 at 1 h incubation and reduced to 0.1 at 24 h.
Figure 4d shows a spectrum taken from untreated PTFE after incubation with cultured (as opposed to rehydrated) cells for 24 h. In this case, amide peaks are observed, but the CH2 and ester C=O peaks associated with the presence of lipids are absent.
For further investigation of the residues, PTFE surfaces before and after incubation in yeast and subsequent cell removal were analysed by XPS. Figure 5 shows the percentage of carbon, fluorine, oxygen and nitrogen, detected on the sample surfaces by XPS: carbon and fluorine are the two primary elements of PTFE. Oxygen and nitrogen were found on both PIII-treated samples as a result of nitrogen plasma treatment and oxidation after the treatment. The PIII treatment also increased carbon and reduced fluorine owing to carbonization and defluorination of the surface .
After cell attachment, the oxygen concentration increases for untreated PTFE by 2 per cent and for the PIII-treated surface by 4–6%. The nitrogen concentration increases for untreated PTFE by 0.5 per cent and for the PIII-treated surface by 2 per cent. This observation is in accordance with the higher cell density on PIII samples and supports the conclusions made from the earlier-mentioned FTIR data about the existence of proteins on sample surfaces. The increase in nitrogen and oxygen concentration on the PTFE surfaces can be attributed to protein and lipid molecules adsorbing on the surface. The proteins and lipids also partially mask the polymer surface, reducing the detected fluorine concentration after incubation.
After SDS washing, the oxygen and nitrogen concentrations were significantly reduced on the untreated PTFE samples for both 1 h (figure 6b) and 24 h incubations (figure 6d), indicating that proteins were removed from these samples. In contrast, oxygen and nitrogen remain on PIII-treated samples, showing that SDS cannot remove the proteins on these sample surfaces. From this, we can conclude that proteins on the untreated PTFE were physically adsorbed, while those on PIII-treated surfaces were covalently bonded. The use of this SDS washing protocol as a test of covalency has been demonstrated in literature over a range of surfaces of different surface energies [16,29,30].
Attempts to wash the cells and residues away with sodium hydroxide were made to confirm the mechanism of attachment. Figure 7 shows the FTIR-ATR spectra from residue remaining on untreated and PIII-treated PTFE after sodium hydroxide washing. Residues on untreated PTFE are removed with sodium hydroxide solution, while those on PIII-treated PTFE remain on the surface. Spectral lines associated with vibrations of the CH2 and ester C=O groups of lipids are not observed after washing. Only amide peaks remain on PIII-treated PTFE surface.
To investigate the source of the lipids, the FTIR-ATR spectrum of a rehydrated yeast cell wall (figure 8a) was recorded by placing a thin layer of a solution containing rehydrated yeast cells on the germanium ATR crystal (with a contact angle of 64°) and letting it dry. The spectra clearly show the three amide peaks (1540, 1650 and 3300 cm−1) associated with proteins and O–H group (broad peak at 3600 − 3000 cm−1), C–H group (3000 − 2800 cm−1) and glycogen bend (1045 cm−1) associated with mannose. No ester carbonyl vibration was observed in the cell wall spectra, indicating that there is no lipid on the outside of the cell wall. A FTIR-ATR spectrum from an untreated PTFE sample incubated vertically in a plastic tube with rehydrated yeast suspension for 24 h (so as to attach no cells) showed absorptions associated with a small amount of protein but no lipid (figure 8b). This indicates that there is no lipid excretion into the medium during incubation.
Figure 8c shows the normalized emPAI values obtained from the mass spectroscopy analysis of the trypsin-digested protein residues for the PIII-treated rehydrated and cultured cells, which are indicative of differences in the amount of each protein on the surface. The greatest difference between the rehydrated and cultured cells was observed to be in the amount of Hsp12. The rehydrated cells have four-times more Hsp12 on the surface than do the cultured cells.
3.3. Activity testing
Figure 9 shows the glucose consumption and ethanol production of immobilized cells compared with cells in suspension over 6 days of batch fermentation. The levels of glucose and ethanol are the same in both beakers, indicating that immobilized yeast and suspended yeast are equally viable. The SEM-image-captured immobilized yeasts on the PTFE surface after 6 days of fermentation (figure 10) shows that the immobilized yeast is proliferating. Daughter cells are observed on top of the original immobilized cells.
The results suggest that the mechanism of cell attachment is different on the PIII-treated and -untreated surfaces. On the PIII-treated surface, yeast cells were observed to attach more rapidly, with higher cell density and attachment strength than on the untreated surface. The distribution of cells was even across the surface without clumping (figure 1). FTIR (figures 4 and 7) and XPS (figure 6) analyses showed that the residues on the PIII-treated surface consisted of proteins and a small amount of lipids. With FTIR, the residue was found to be similar for both PS and PTFE. The protein-to-lipid ratio was always greater than one throughout the incubation (table 2). The lipids were physically adsorbed on the PIII-treated surface as they were not detected after washing with sodium hydroxide. On the other hand, the proteins remained both after sodium hydroxide and SDS washes, which illustrates that a significant fraction of the protein residue is covalently bound to the surface.
On the untreated surface, cell attachment occurred much slower than on PIII-treated surface. A gradual increase in cell density and attachment strength was observed with incubation time. At 15 min and 1 h incubation times, only few cells were resistant to mild milli-Q rinsing and were observed on the surface (figure 1). The cell density increased from 3 to 6 h, but clumping of the cells in the images illustrate that the cell attachment was too weak to withstand the fixing and staining process. At 24 h, no clumps were observed showing the highest attachment and cell density. The cell density saturated at half the level of the PIII-treated surface. Lipids were found with a high concentration compared with proteins for both short and long incubation times (table 2). Protein residues were washed off after SDS and sodium hydroxide treatment, indicating physical adsorption of proteins on untreated surface (figures 6 and 7). With longer incubation time, the protein-to-lipid ratio reduced from 0.42 to 0.1 (table 2) and at the same time points, the cell density was observed to increase.
The composition of the yeast cell wall plays an important role in the attachment process. Cell organelles are protected by three layers, which are (from outside to inside): the cell wall (100–200 nm), the periplasmic space (3.5–4.5 nm) and the plasma membrane (7 nm) . The cell wall of S. cerevisiae consists of approximately 90 per cent carbohydrate, 5–10% protein and 1–2% glucosamine. The cell wall has two layers: an inner of β-1,3 and β-1,6 glucan (which form a complex with chitin), and an outer layer of mannoproteins (which determine the surface properties of the yeast cell) . The protein content in the cell wall mannoproteins is approximately 4–5%, with the remaining mass consisting of protein-linked mannose-containing carbohydrate . We propose that these mannoproteins form the covalent bonds with the free radicals on the PIII-treated surface or adsorb on the untreated surface as yeast cells landed there. However, reports in the literature about the presence and amount of lipid on the cell wall are rare.
The source of lipid was investigated further with three different experiments (figure 8). Figure 8a shows that absorptions characteristic of lipids were not observed when intact cultured cells were placed on a germanium crystal surface and investigated with ATR-FTIR, which implies that the lipid concentration in the cell wall is less than the detection limit. ATR-FTIR analysis uses an evanescent field that decreases exponentially with distance to a depth of approximately 1 μm from the detection surface and therefore is most sensitive to the outermost cell surface. Figure 8b shows that the lipids were also not detected on surfaces that were not directly exposed to cells, demonstrating that the lipids were not secreted into the solution and adsorbed from there. From these two results, we can infer that the source of the lipids is from inside the yeast cell, particularly from the plasma membrane where lipids are abundant.
Beker et al.  found that the phospholipid membrane was damaged during the dehydration process. When free and bound water molecules were removed, the phospholipid molecules were found to have changed their orientation in the bilayer. At rehydration, some of the phospholipid molecules could not return to the original bilayer structure of the membrane. To investigate the influence of plasma membrane damage caused by the dehydration process, cultured cells were prepared that had intact membranes. The cultured cells showed no lipid residue on the untreated PTFE surface with FTIR (figure 4d), which contrasts to the residue from the rehydrated cells.
Stress from the dehydration/rehydration process was evident in the protein residue left by rehydrated cells when investigated with mass spectroscopy. Among the stress response proteins, Hsp12 is found to be significantly higher in rehydrated cells than in cultured cells. Hsp12 is a small hydrophilic protein expressed under stress by the cells in both the cell wall and the plasma membrane. Sales et al.  found that Hsp12 protected the lipid membrane integrity against desiccation or in the presence of ethanol. The higher level of Hsp12 in rehydrated cells is an indication of the requirement for the cell protective effect so that it can be inferred that the original dry cells have experienced much more stressful conditions than cultured cells. These results suggest that the plasma membrane damage is a consequence of the dehydration/hydration process.
A model illustrating the differences in the mechanism of attachment on the PIII-treated and -untreated surfaces is proposed (figure 11). On the PIII-treated surface, we propose that the initially adsorbed proteins from the cell wall are covalently immobilized by reaction with radicals diffusing from the treated surface layer. PIII treatment is known to create radicals in the treated surface layer, which react with oxygen in the air, modifying the chemical properties of the polymer surface , and are able to covalently bind proteins . The yeast cell wall has a lipid membrane on the inner side and mannoproteins on the outermost layer [31,32]. The observation of protein characteristic vibrations in the FTIR spectra (figures 4 and 7) confirms the presence of proteins on the outermost layer of the yeast cell wall. Upon contact with proteins, the radicals in the treated polymer, which have migrated to the surface, covalently bind the protein molecules . The presence of covalent bound proteins on the treated surface has been confirmed by XPS observation of the surface after washing with SDS (figure 6). On the untreated surface, we propose that both proteins and lipids play a role in cell attachment through the hydrophobic interactions because both are present in the FTIR spectra (figure 4).
The cell attachment mechanism on the PIII-treated surface has many advantages for continuous flow processes. High initial cell density can shorten fermentation time (lag phase and exponential phase) and give higher ethanol productivity . Initial cell density also has a particularly important role in those fermentations where cells do not proliferate. The surface immobilization allows yeast cells to directly contact with nutrients in the medium and it reduces carbon dioxide pressure more quickly (as opposed to the encapsulation method). The covalent attachment on the PIII-treated polymer is expected to help yeast cells to resist the flow better than physical adsorption. Both of these features are very important for continuous operation, as cells and their buds are required to remain in the continuous flow fermentation medium with stable density. The model can also be expanded to other types of yeast or microorganisms that have cell wall proteins that can react with free radicals on the PIII-treated surface.
The PIII treatment of polymers is a new convenient and effective method for yeast cell immobilization. The PIII-treated polymer has the capacity to immobilize yeast cells quickly and with high cell density. Those immobilized cells were found as active as cells in suspension. We observed higher levels of protein on the PIII-treated surfaces after cell removal and showed that the protein was covalently immobilized while physical forces were responsible for protein and lipid retention on the untreated surfaces. The covalent attachment is proposed to be via the formation of covalent bonds between cell wall proteins and radicals in the activated polymer surface. On the other hand, the increased attachment of rehydrated yeast cells on untreated PTFE with incubation time was found to be associated with the increased concentration of a lipid residue from the plasma membrane that was damaged during the dehydration process. The presence of stress associated with the rehydration/dehydration process was confirmed by mass spectroscopy.
We thank Dr Neil Nosworthy for SDS-PAGE analysis, Dr Leo Philips, Mr Fernando Barasoain and Ms Natalia Kislova for the ethanol assay and Dr Cenk Kocer, Dr Daniel Bax for their helpful advice. The provision of laboratory facilities by Professor Cris dos Remedios is gratefully acknowledged. We also thank Australian Centre for Microscopy and Microanalysis at university of Sydney and Ms Ling Zhong from UNSW for the mass spectroscopy measurement. This study was financially supported by the Australian Research Council.
- Received May 2, 2012.
- Accepted May 21, 2012.
- This journal is © 2012 The Royal Society