ARTICLE pubs.acs.

org/Langmuir

Antimicrobial Sand via Adsorption of Cationic Moringa oleifera Protein

Huda A. Jerri, Kristin J. Adolfsen, Lauren R. McCullough, Darrell Velegol, and Stephanie B. Velegol*,

Department of Chemical Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States Department of Civil and Environmental Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States ABSTRACT: Moringa oleifera (Moringa) seeds contain a natural cationic protein (MOCP) that can be used as an antimicrobial occulant for water clarication. Currently, the main barrier to using Moringa seeds for producing potable water is that the seeds release other water-soluble proteins and organic matter, which increase the concentration of dissolved organic matter (DOM) in the water. The presence of this DOM supports the regrowth of pathogens in treated water, preventing its storage and later use. A new strategy has been established for retaining the MOCP protein and its ability to clarify and disinfect water while removing the excess organic matter. The MOCP is rst adsorbed and immobilized onto sand granules, followed by a rinsing step wherein the excess organic matter is removed, thereby preventing later growth of bacteria in the puried water. Our hypotheses are that the protein remains adsorbed onto the sand after the functionalization treatment, and that the ability of the antimicrobial functionalized sand (f-sand) to clarify turbidity and kill bacteria, as MOCP does in bulk solution, is maintained. The data support these hypotheses, indicating that the f-sand removes silica microspheres and pathogens from water, renders adhered Escherichia coli bacteria nonviable, and reduces turbidity of a kaolin suspension. The antimicrobial properties of f-sand were assessed using uorescent (live-dead) staining of bacteria on the surface of the f-sand. The DOM that can contribute to bacterial regrowth was shown to be signicantly reduced in solution, by measuring biochemical oxygen demand (BOD). Overall, these results open the possibility that immobilization of the MOCP protein onto sand can provide a simple, locally sustainable process for producing storable drinking water.

INTRODUCTION Moringa oleifera is often called the miracle tree. Its nutrientrich leaves can be harvested and eaten, and the seeds can be pressed to extract oil for biodiesel.1 The tree is prevalent in equatorial lands,2 which often coincide with the regions of the world suering from malnutrition, insucient energy resources, and waterborne disease.3 Treating water that contains pathogenic bacteria, viruses, protozoa, parasites, and sediment is a signicant challenge in many developing nations.4 This paper focuses on the use of seeds from the Moringa oleifera tree (hereafter called Moringa) to produce potable water.5 The antimicrobial, occulant properties of dried Moringa seeds are well established in the literature.611 In addition to removing turbidity and killing microorganisms, the Moringa seeds have even been reported to remove dyes and detergents,12 arsenic and cadmium,13,14 and color agents15 from aqueous solutions. The Moringa seeds contain a natural, cationic protein (MOCP) that acts as a occulant, decreasing the turbidity and removing negatively charged particles, including bacteria.16 The antimicrobial functionality of the MOCP stems from positively charged, glutamine-rich sections surrounding a helixloophelix motif containing a hydrophobic proline. This portion of the protein chain acts like a molecular knife: the positively charged section is electrostatically drawn to the bacterial cell wall while the hydrophobic loop penetrates and disturbs the wall.17 The main barrier to using Moringa for clarifying water has been the presence of excess organic matter left in the water from the seed.2
r XXXX American Chemical Society

This organic matter provides sucient food, measured by biochemical oxygen demand (BOD), to support new bacterial growth. The bacterial regrowth requires additional water treatment, which is wasteful and time-consuming, and limits the ability to store treated water. Previous studies have addressed this issue by purifying,16 synthesizing, or expressing the MOCP as a recombinant peptide.17,18 Although the isolated or synthesized protein maintains its ability to clean water, the processes are too elaborate and expensive to be used in places where clean water is scarce. In this work, MOCP is electrostatically adsorbed onto anionic silica particles or sand, forming a MOCP-functionalized sand, or f-sand. Since the sand can be mixed with water, but also rapidly removed from water by simple sedimentation, the remaining organics that cause bacterial growth are readily removed by rinsing. A key question is whether the adsorbed, immobilized protein retains antimicrobial activity while simultaneously removing turbidity in bulk solution. In order to examine this question, we measured the removal of various negatively charged colloids, the removal and viability of DH5 Escherichia coli bacteria, and the reduction of turbidity of a kaolin solution in the presence of MOCP-functionalized f-sand. Confocal laser scanning microscopy (CLSM) was used in conjunction with bacterial viability
Special Issue: Bioinspired Assemblies and Interfaces Received: Revised: September 30, 2011 November 16, 2011

dx.doi.org/10.1021/la2038262 | Langmuir XXXX, XXX, 000000

Langmuir staining to assess, qualitatively, the antimicrobial activity of the functionalized sand. Additional studies were performed to assess antimicrobial and turbidity removal ecacy of f-sand after long-term dehydration and storage.

ARTICLE

MATERIALS AND METHODS

Materials. Sodium phosphate (Na2HPO4), potassium phosphate (KH2PO4), and sodium chloride (NaCl) were purchased along with cationic polyelectrolyte poly(allylamine hydrochloride) (PAH, Mw = 70 000), fused white quartz sand (5070 mesh, corresponding to 250 to 500 m diameter particles) and kaolin (Al2Si2O5(OH)4) from SigmaAldrich Chemicals, USA, and used without further purification. Silica particles (3.01 m diameter, 9.8% solids) and monodisperse, surfactantfree, sulfate-functionalized 60 nm diameter fluorescent yellow-green polystyrene latex nanoparticles and 1 m fluorescent carboxylated latex microparticles were purchased from Interfacial Dynamics Corporation (Portland, OR). Whole, tree-dried Moringa seeds of the PKM variety were purchased from ECHO seed bank. These seeds were crushed (including shell and wings) prior to use to increase surface area and create millimeter-sized particles using a mortar and pestle. Millipore Milli-Q deionized (DI) water (resistivity greater than 18.2 M 3 cm) was used in all experiments. Protein Solubilization. The water-soluble portion of the seed was extracted, including the MOCP and other proteins present in Moringa seeds, into an aqueous phase. Approximately 1.0 g of crushed seed powder (4 seeds) was suspended in 40 mL of either phosphate buffered saline (PBS) buffer (10 mM Na2HPO4, 1.37 mM KH2PO4, 1 mM NaCl) or DI water in 50 mL polypropylene centrifuge tubes. Each Moringa seed (shell and wings intact) weighed 0.25 ( 0.05 g. In order to allow the protein to dissolve in the water, the centrifuge tubes were placed horizontally on a slow roller for 1 h at room temperature. The tubes were then turned upright, causing the heavy matter to settle to the bottom. The supernatant, now containing MOCP, was collected, and this concentrated serum was used unfiltered. Cationic Moringa Protein (MOCP) Adsorption and Particle Mapping. Monodisperse 3.01 m silica particles were used as model
particles to study electrostatic adsorption of MOCP to microscale sand. The adsorption of the cationic protein onto the anionic silica microparticles was determined by taking potential measurements (Zeta Potential Analyzer, ZetaPALS, Brookhaven Instruments) before and after the silica microparticles were treated with the MOCP solution. Unlike large heterogeneous sand granules, which settle rapidly, the silica microparticles remained suspended in solution, thereby facilitating measurement of accurate potentials. Silica particles (60 L, 9.8% solids) were pipetted into 10 mL of the MOCP supernatant (25 mg/mL crushed seed in DI water). This solution was rolled for 1 h in 15 mL centrifuge tubes. The samples were then centrifuged in a standard bucket centrifuge for 5 min at 725 g to form a pellet of coated particles at the bottom of the tube. The supernatant was removed; the particles were resuspended in DI water; and the tube was again centrifuged. This rinsing step was repeated three times to adequately remove excess BOD from solution. After the third centrifugation step, the samples were resuspended in solutions of various ionic strengths, and the potentials were measured. The distribution of surface charge over the particle surfaces was assessed using uorescent nanoparticle mapping. The Moringa-treated silica microparticles were incubated with 50 L of 60 nm uorescent yellow-green particles in 10 mL of DI water and left to roll overnight. The nanoparticles were added in excess and given adequate time to adsorb electrostatically to the underlying cationic coating before centrifuging and rinsing with DI water. The samples were rinsed twice prior to imaging, and then confocal and dierential interference contrast (DIC) optical microscopy images were captured using an Olympus
B

Figure 1. Process for immobilizing Moringa protein onto particles. Schematics are shown for both silica microparticles (ad) and the corresponding macroscale sand (eh, to create BOD-free f-sand). A core anionic silica microparticle (a) or sand granule (e) in a batch process is incubated with Moringa extract, which contains MOCP (b,f). Excess Moringa solution is removed via centrifugation and resuspension in DI for microparticles (c), or by sedimentation and rinsing with water for sand (g), resulting in core particles with adsorbed MOCP. The MOCP-coated particles can be incubated with anionic uorescent nanoparticles to tag and visualize the coating (d), whereas the f-sand (h) is ready to be used for water clarication. Fluoview 1000 confocal laser scanning microscope with 10, 20, and 60 oil objectives at the Huck Institute of the Life Sciences Center Cytometry Facility at Penn State University. Moringa-functionalized sand, or f-sand, was produced using steps similar to those used to treat the silica (see Figure 1). For the initial experiments, 1.2 g of quartz sand (rinsed three times with DI water prior to use) was incubated with 9 mL of MOCP supernatant solution (from 0.9 seeds) for 1 h on a roller, followed by ve rinsing steps with 10 mL each of DI water. Further experiments showed that similar results were obtained even when 0.15 seeds were used, indicating possible saturation. Centrifugation was not required for sand, which settled naturally in a few seconds due to gravity. Thorough rinsing was performed to ensure that excess BOD was not trapped in the interstices of the settled, closely packed sand. To test the capturing ability of f-sand for anionic particles, 50 L of 1 m uorescent carboxylated particles (negatively charged) were added to 1.2 g of the f-sand in 9 mL of DI water, and this mixture was rolled for 5 h before imaging. As a control, sand was also functionalized with a cationic polyelectrolyte, PAH. A 20 M solution of PAH was made in 30 mM KCl and 9 mL of this solution was incubated with 1.2 g of sand for 1 h prior to rinsing and imaging. Antimicrobial Studies. In order to test the antimicrobial properties of the f-sand, the nonpathogenic E. coli bacteria strain DH5 was grown in lysogeny (Luria) broth (LB) at 37 C for 12 h to a concentration of 108 bacteria/mL. An equal volume of media supernatant was then exchanged for PBS via a centrifugation and resuspension step. PBS solution was used in place of DI for all steps of sand functionalization and control experiments with sand granules to minimize lysing of the bacteria and to maintain staining consistency. 1.2 g of either f-sand or bare sand was incubated with 9 mL of the bacterial solution (108 per mL) in 15 mL centrifuge tubes and rolled at room temperature for 16 h prior to staining and imaging. Experiments were performed both with and without rinsing away excess unadhered bacteria with PBS. Samples were stained with a BacLight dual stain kit (Invitrogen, Cat No. L7012) to determine the viability of adhered and unadhered bacteria. BacLight employs two nucleic acid stains: green
dx.doi.org/10.1021/la2038262 |Langmuir XXXX, XXX, 000000

Langmuir
fluorescent SYTO 9 stain and red fluorescent propidium iodide stain. These stains differ in their ability to penetrate healthy bacterial cells. SYTO 9 dye can penetrate all bacteria while propidium iodide can penetrate only bacteria with damaged membranes. Bacteria that are alive and have intact membranes appear green; nonviable bacteria with compromised membranes appear red. Staining was also performed on f-sand and bare sand without bacteria to assess the degree of background fluorescence and confirm that actual bacteria were being stained. BOD Determination. BOD5 (5-day biochemical oxygen demand) is a measure of how much oxygen is used when bacteria biodegrade organic matter in an aqueous solution after 5 days. This measurement was used to compare the BOD per crushed seed when that seed was added directly to water, versus the BOD released from f-sand. The hypothesis was that much less BOD would be released from the f-sand. A Hach BODTrak was used to determine the BOD5 of the Moringa seed supernatant, as well as the BOD5 of the water in contact with rinsed f-sand. The standard operating procedure defined in the Hach BODTrak Manual was used in all BOD experiments. A BOD5 range of 070 mg/L for the final prepared samples was chosen. Each bottle contained 5 mL of wastewater seed solution from the Pennsylvania State University wastewater treatment plant. This wastewater solution contained the microorganisms required for the BOD5 measurements. Dilution solution was made using a nutrient buffer pillow. For crushed Moringa seed measurements, 1 mL of supernatant was added to 355 mL of dilution solution. To measure the BOD5 in solution around f-sand, 5 g of f-sand was added to 355 mL of dilutant. Bulk Turbidity Removal Studies. A model turbid solution of kaolin was used to test the ability of f-sand to flocculate turbidity. This solution was made by adding 5.0 g of kaolin powder to 1.0 L of water, and stirring for 1 h before allowing the slurry to settle for 24 h. This stock solution has previously been shown to have a turbidity of 105 NTU.16 Two stock solutions were made from this supernatant with mass concentrations of 0.54 mg/mL and 0.84 mg/mL. A Helios ThermoSpectronic UVvis spectrophotometer was used to relate absorbance measurements to concentration via predetermined calibration curves. To study removal of kaolin from the more concentrated stock solution (Abs = 0.95 at = 650 nm), bare sand and f-sand were incubated with kaolin stock solutions in plastic centrifuge tubes, and gently rolled. The centrifuge tubes were inverted, and the sample supernatant was removed, tested using disposable cuvettes, and returned to the tubes to continue rolling between consecutive measurements. Protein Isolation and Identification. To confirm that MOCP was indeed present on the surface of f-sand, the protein was eluted off the surface by increasing the ionic strength of the surrounding solution. First, 35 g of f-sand was prepared and rinsed thoroughly 10 times with 10 mL each. The f-sand was transferred to a glass Petri dish, and 0.30 M NaCl was added. The mixture was agitated for 10 s and then was left standing for 10 min. The 0.30 M NaCl solution was removed by pipetting. Then this entire procedure was repeated with 0.60 M NaCl solution. Only the 0.60 M NaCl revealed significant absorbance at 280 nm, indicating that the protein was eluted with 0.60 M NaCl, but not with 0.3 M NaCl. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDSPAGE) analysis was performed on the samples at the Proteomics and Mass Spectrometry Core Facility at Penn State. After eluting o the proteins with the 0.6 M NaCl, the solutions were dialyzed using 3000 Da cellulose membranes to isolate the MOCP protein, which has a molecular weight of 6500 Da. A 12% Mini-PROTEAN TGX TM Precast Gel was then used in a Mini-PROTEAN electrophoresis cell. There was a single distinct band at 6500 Da from the solution of 0.60 M NaCl, and no other bands were visible on the gel, indicating that MOCP had been adsorbed onto the f-sand, and no other protein had been adsorbed. The purity of the adsorption might be attributed to the facts that (1) the MOCP is cationic, and therefore adsorbs electrostatically,
C

ARTICLE

Figure 2. Negatively charged uorescent nanoparticles adsorbing onto Moringa-treated silica particles. The 60 nm FYG (green) nanoparticles eectively tagged and labeled the underlying MOCP coating surrounding the 3.01 m silica particles, and remained well-adhered to the particles after rinsing.

Figure 3. Microscale f-sand eectively occulates bacteria. (a) Bare silica particles cause no signicant occulation of bacteria. (b) Moringafunctionalized 3.01 m silica microparticles have signicant occulating ability, in addition to antimicrobial properties, in the presence of bacteria. Even after immobilization onto the silica particles, the MOCP retains occulant and antimicrobial properties. The images were taken using CLSM with a 60 oil immersion objective. and (2) since the MOCP is only 6500 Da, it has a high diusion coecient and reaches the surface before other proteins. This 6500 Da band was digested and analyzed using mass spectroscopy. The isoelectric point was determined to be 10.8, and the amino acid sequence was consistent with the amino acid sequence of MOCP found at the NCBI (National Center for Biotechnology). MOCP is also known as MO2,1 and Flo peptide.19

RESULTS AND DISCUSSION The occulant and antimicrobial functionality of the MOCP has been explained in the literature by its molecular structure.18,17 The Moringa seeds contain a water-soluble cationic dimer protein with a molecular weight of 13 000 Da (each monomer has a molecular weight of 6500). The potential of the protein in water is +6 mV with an isoelectric point between 10 and 11.16 Suarez et al. used recombinant and synthetic forms of MOCP to
dx.doi.org/10.1021/la2038262 |Langmuir XXXX, XXX, 000000

Langmuir

ARTICLE

Figure 4. Microparticle adherence onto f-sand. Negatively charged uorescent microparticles, comparable to bacteria in size, adhered to the f-sand and uorescently tagged the underlying cationic coating. f-sand granules labeled with 1.0 m carboxylated polystyrene particles (shown purple) were imaged using CLSM and a 10 objective.

isolate the peptide chain responsible for both the occulation and antimicrobial functionality.18,17 They found that aggregation of bacterial cells required a positively charged, glutamine-rich portion. Bacterial membrane damage occurs due to a helix loophelix structural motif with a hydrophobic proline residue. This peptide was active against Pseudomonas aeruginosa and Streptococcus pyogenes, acting as a type of molecular knife. Adsorbing the MOCP to the 3.01 m silica particles in DI water lowered the magnitude of the potential, from 58 ( 1.5 mV to 1.8 ( 3.4 mV (i.e., essentially neutral). The cationic proteins neutralized the charge without overcompensating and reversing it. Upon introducing negatively charged, uorescent nanoparticles and bacteria to these functionalized microspheres, we observed the adherence of the nanoparticles and bacteria shown in Figures 2 and 3. Figure 2 shows microscale f-sand coated with the anionic nanoparticles, which stuck to the cationic sites on the silica particle surface. As expected, a control experiment showed that no anionic nanoparticles adhered to bare silica microspheres. Figure 3 shows that MOCP-functionalized microspheres cause occulation of the E. coli bacteria. With evidence that positively charged proteins were coating the surface of the silica and that negative particles (including bacteria) were adhering to the Moringa-functionalized silica microspheres, the transition was made from 3.01 m silica particles to sand granules, approximately 250500 m in size. The potential could not be measured directly for the fastsettling sand, and so the bare sand was crushed to a small size and found to have a potential of 36 ( 0.6 mV. The noncrushed sand was immersed in a solution of MOCP, in order to create f-sand. To map the charge distribution on the f-sand particles, negatively charged particles were utilized. 1.0 m uorescent, carboxylated polystyrene latex (PSL) particles were chosen for this task. The samples were analyzed using CLSM to demonstrate the ability of the f-sand to capture small negatively charged microparticles (Figure 4). Bare sand showed no adhered anionic particles. Figure 4 shows there are parts of the heterogeneous sand surfaces that are positively charged due to the adsorption of cationic proteins. These MOCP-coated regions facilitate the adsorption of anionic colloids electrostatically. In practice, it might be useful that the MOCP does not cover all surface sites on the sand. The sites having no adsorbed MOCP may remain anionic, and therefore can capture cationic particulates from suspension. This gives the f-sand dual functionality, with an ability to capture both positive and negative particles. To test the ability of the f-sand to remove and kill bacteria, E. coli bacteria strain DH5 at a concentration of 108 per mL was
D

Figure 5. Ability of bare sand and f-sand to capture and kill bacteria. (a) With the untreated sand control (bare sand), the DH5 E. coli bacteria were alive (green) in solution, and also when directly contacting the sand surface. The black regions show the positions of the sand grains. Sand alone neither adsorbs nor kills bacteria. (b) The f-sand surface captured and killed the DH5 E. coli bacteria (red uorescence), while the surrounding bacteria in solution remained viable. The red outline surrounding the functionalized sand granules indicates a coating of adhered, dead bacteria. These images were taken at 10 magnication with CLSM.

Figure 6. Microscopic view of the capturing ability of f-sand. (a) Bare sand captures no bacteria at its surface, which is indicated by the superimposed blue dotted line (drawn using the DIC image), and in the bulk solution the bacteria remain alive (green). (b) f-sand readily captures and kills bacteria on the sand surface. The individual rodshaped bacteria were resolved using CLSM and a 60 oil immersion objective.

Figure 7. Bacteria remain stuck to the f-sand. f-sand was incubated with bacteria (4 h), and then vigorously rinsed prior to staining to demonstrate the robustness of the MOCP coating and layer of tightly bound, unviable bacteria.

incubated with the f-sand. At this excess concentration, we expected the surface of the f-sand to become saturated with bacteria. BacLight live/dead staining was used to examine the viability of the bacteria. Around bare sand (Figure 5a) there were almost exclusively live (green) bacteria. The sand shows up as black since it is nonuorescent. There were no bacteria adhered to the bare sand surface. By contrast, the image of f-sand with
dx.doi.org/10.1021/la2038262 |Langmuir XXXX, XXX, 000000

Langmuir

ARTICLE

Figure 8. Control experiment conrming red uorescence is directly attributable to nonviable bacteria, and the fact that MOCP coating does not passively uoresce. Simultaneous staining and imaging of a mixture of f-sand granules separately incubated with bacteria (three red uorescent sand grains on the left) and f-sand incubated solely in buer (one dim sand grain on the right) clearly reveals the eectiveness of the f-sand in attracting and killing bacteria. The nonviable bacteria remain bound, and do not transfer to adjacent f-sand granules with available binding sites.

Figure 9. Eectiveness of PAH cationic polymer in capturing and killing bacteria. The cationic polyelectrolyte PAH electrostatically adsorbs to the sand surface, and demonstrates some antimicrobial benet as shown by the slight red glow on the surface of a polymerfunctionalized sand granule. Twenty micromolar PAH in 30 mM KCl was used to functionalize the sand prior to a 4 h incubation with bacteria, BacLight staining, and CLSM with a 60 oil immersion objective.

bacteria (Figure 5b) revealed dead (red) bacteria covering the surface of the f-sand and live bacteria in the surrounding solution. A closer look at the surface of f-sand reveals that the f-sand captured and killed the bacteria. Figure 6 shows individual rodshaped bacteria, now stained red (compromised membranes), and they remain adhered to the sand. The bacteria that were adhered to the surface of the f-sand remained tightly bound, even after vigorous rinsing with PBS solution (Figure 7). This implies that bacteria would not be released back into the water once the concentration of bacteria in solution is changed. As a further control to check for autouorescence or bacterial transfer, two f-sand samples were prepared separately. One of the samples was incubated with bacteria for a 4 h period before gentle rinsing with PBS solution. The other sample was not exposed to bacteria initially. Then equal quantities of the two f-sand samples were mixed, and the samples were stained in the same manner as those in Figures 57. In Figure 8 there are three f-sand grains covered with dead bacteria and one f-sand grain with no bacteria present. Figure 8 shows that the bacteria remain stuck on the surface of the functionalized sand and do not transfer to neighboring particles throughout the process. Additionally, the simultaneous staining of the two f-sand samples with and without bacteria reveals that the MOCP coating does not uoresce, and conrms that the uorescent red shell is directly attributed to nonviable adhered bacteria. The precise mechanism for this antimicrobial surface is still somewhat unclear. In solution, the cationic proteins found in Moringa seeds are initially attracted to the bacteria through electrostatic forces. Once adhered, the protruding hydrophobic loop on the protein may penetrate the bacterial cell wall and destroy the bacteria.17 On the surface of silica, that portion of the protein would have to be exposed to the solution for the mechanism to remain the same. Similar antimicrobial peptides, however, have been shown to lose their antimicrobial properties once adsorbed to a surface due to a lack of mobility.20 Once they are tethered, these peptides retain some activity with greater activity corresponding to longer tethers.21 To try to elucidate something about the mechanism, the MOCP coating was compared to another cationic polymer, PAH. PAH was adsorbed to the sand, and the surface was exposed to bacteria (Figure 9). The surface of the PAH-coated sand became covered in dead bacteria, although it appears there are fewer bacteria on the PAH-coated sand surface compared to
E

Figure 10. Signicant turbidity decrease and removal of kaolin from water using f-sand. 83% of kaolin was removed from a 6 mL solution containing 5.0 mg of kaolin and 2 g of f-sand in 1 h. A total of 0.40 g of raw Moringa seeds was used to make the f-sand. 95% error bars are calculated from six experiments ran in triplicate. Inset. Turbid kaolin solution becomes clear following treatment with f-sand. (a) 0.54 mg/mL kaolin suspension treated with 2 g of f-sand. (b) 0.54 mg/mL kaolin solution treated with bare sand. The kaolin solution treated with f-sand is signicantly less turbid than the samples treated with bare sand and the stock solution.

MOCP-treated f-sand surfaces. This suggests that the destruction of the bacterial wall is due, at least in part, to interactions between the cationic PAH polymer on the surface and the anionic cell wall. Previous studies have shown that bacteria initially adhere more rapidly to a cationic surface but show slower surface growth, once adhered.22,23 This is particularly true for Gramnegative strains.24 This may be because the strong interaction between the bacteria and the surface prevent elongation of the adhered bacterium. In this case, we are measuring the ability of the BacLight stain to penetrate the cell membrane. Red stain will only enter a compromised bacterial membrane, indicating that
dx.doi.org/10.1021/la2038262 |Langmuir XXXX, XXX, 000000

Langmuir the bacterial wall is being compromised by adhesion to the surface. It is emphasized that the functionalization of sand with synthetic materials may not be possible or sustainable in developing countries, due to the unavailability and expense associated with the chemicals. Since anionic particles (including bacteria) can adsorb to f-sand, a macroscopic experiment was performed to determine the kaolin capturing and removal ability of f-sand from a model turbid kaolin solution. For the model turbid solution of kaolin, 2 g of either bare sand or f-sand were incubated with 6 mL of a kaolin suspension (0.54 mg of kaolin per 1 mL of water) and imaged. A qualitative examination of samples of f-sand (Figure 10, inset a) and bare sand (Figure 10, inset b) with the kaolin stock solution in Figure 10 reveals that the f-sand claried the solution much more than the bare sand. The f-sand drastically reduced the turbidity of the solution, transitioning the sample from opaque white to clear. The image was taken 10 h after the timed experiment began, and kaolin removal was enhanced by gently rolling the sand with the kaolin solution. Quantitatively, the macroscopic experiment showed the same signicant kaolin removal as the qualitative visual observations (Figure 10). A stock solution of 0.84 mg/mL kaolin was used, with removal measured using UV/vis spectrophotometry. The f-sand signicantly decreased the concentration of kaolin in solution, removing 83% of the kaolin from the original solution in 1 h. For these experiments, 40 mL of Moringa supernatant was made using 1 g of crushed Moringa seed (approximately 4 seeds). Two grams of f-sand was then created using 2.5 mL of this supernatant (0.0625 g of seed). These 2 g of f-sand were used to treat 6 mL of kaolin suspension, an equivalent of 40 seeds per liter of cleaned water. Further studies should enable us to approach the recommended dosage of one seed per liter of water used for Moringa cleaning in bulk solution.25 To test the adaptability and longevity of the f-sand, both dried and wet f-sand were stored and then tested for turbidity removal. The f-sand was spread on a at surface to dry overnight. The dry samples were stored in capped tubes for up to 4 months at room temperature. Two grams of wet f-sand was stored for 22 days in capped plastic tubes with 10 mL of clean DI water. After the storage period, a kaolin removal experiment was conducted. The eectiveness of the f-sand to remove turbidity was not altered when the f-sand was stored either wet or dry (data not shown). After verifying that the MOCP adsorbed onto f-sand retained its antimicrobial and turbidity removing properties, tests were performed to determine whether f-sand added signicantly less BOD to the treated water than the standard treatment with Moringa seeds. In the standard Moringa seed water treatment used in the eld,25 one seed (0.25 g) is used to treat 1 L of water (although one study found that turbidity is removed using as few as 0.2 seeds per 1 L of water16). To follow the bulk treatment method, we crushed one seed in 1 L of water and measured 2090 ( 250 ppm of BOD5, or 2090 ( 250 mg of BOD5 from one seed. With the f-sand method, we rst made 5.0 g of rinsed f-sand from 0.6 seeds. The f-sand was used to treat 30 mL of water. The total release of BOD in the water was 1.5 ( 0.6 mg. In short, the f-sand treatment removed almost all the BOD originating from the raw seed, which will greatly improve the ability to store water cleaned with f-sand.

ARTICLE

order to improve the quality of their water.26 These women were unknowingly immobilizing proteins for water purication, although it is now known in part the mechanisms by which the seed acts. It is shown in this paper that the active cationic proteins from Moringa can be isolated and immobilized to create f-sand, which can in turn remove turbidity and kill bacteria in water. One possible long-term outcome is that these results will lead to the development of a water treatment process for villages in the developing world. The use of Moringa to treat water would circumvent the use and expense of chemicals such as alum and chlorine, which can lead to toxic byproduct and are sometimes unavailable. The direct use of naturally dried seeds eliminates the need to boil water, thereby reducing the energy expenditure of nding and burning fuels. Compared to treating water with only Moringa seeds (and no sand) in the bulk, the key advantage of using f-sand is the elimination/reduction of organic matter upon which microorganisms can feed, resulting in clean water that can be stored for longer times. In order to apply this technology to communities that need it most, a few key points must be addressed. Here we have specically demonstrated that immobilized MOCP has the ability to capture and compromise the membranes of lab-grown E. coli from a concentrated bacterial solution (108 per mL). Our goal here was to test the hypothesis that the MOCP still maintains antimicrobial properties once adsorbed to sand. However, future work needs to be done to quantify the amount of bacterial removal as a function of seeds added, concentration of bacteria, and available surface area. We expect at least two limitations to the ability of the f-sand to capture bacteria (or any negatively charged particle). One limitation is available surface area; once the surface has become saturated, no more bacteria can adsorb. Second, the bacterial adsorption may become transport limited if the concentration of bacteria or f-sand is low. One way around the transport limitation is to incorporate the f-sand in a sand lter to minimize the transport distance to the nearest sand grain. We are currently working on a process to regenerate the f-sand surface. In addition, although it has not been tested, we are condent that starting with dirty water (rather than DI water) to produce the f-sand may decrease the eectiveness of f-sand to clean water, but not eliminate the eectiveness. These results indicate that Moringa protein can be quickly and simply adsorbed to form natural and eective antimicrobial f-sand. A potential batch process for water treatment might require only raw water, Moringa seed, and sand, which are locally sustainable components in many parts of the world.

AUTHOR INFORMATION
Corresponding Author

*E-mail: sbvelegol@engr.psu.edu.

CONCLUSIONS It is reported that thousands of years ago, women from the Nile region rubbed their clay water pots with Moringa seed in
F

ACKNOWLEDGMENT We thank Dan Nold and Rick Bates for introducing us to the opportunities with Moringa. Siela Maximova graciously provided the original Moringa seeds, and we thank her and Mark Guiltinan for their numerous useful discussions concerning Moringa. Patrick Cirinos input regarding the mechanism by which the MOCP kills bacteria has been invaluable as we learned more about the antimicrobial properties of immobilized MOCP. Je Larsen is greatly appreciated for his advice on bacteria handling and for providing the bacteria used in these experiments.
dx.doi.org/10.1021/la2038262 |Langmuir XXXX, XXX, 000000

Langmuir We thank the National Science Foundation (CBET Grant #0651611) and the EPA (P3 Sustainability Student Competition Award, Grant # SU834740) for funding this work. We also thank the Penn State Department of Chemical Engineering Undergraduate Research Scholarship for providing funding for two undergraduate students working on this project. We also thank the Huck Institute of Life Sciences Cytometry Facility for use of the Olympus Fluoview 1000 confocal laser scanning microscope (through a project funded in part under a grant with the Pennsylvania Department of Health using Tobacco Settlement Funds), and specically thank Nicole Zembower for her continued support with staining and imaging. Additionally, we would like to thank Danny Hoover for help with separating and characterizing the protein, as well as the sta at the Penn State Proteomics and Mass Spectrometry Core Facility for assisting with gel electrophoresis, mass spectrometry, and protein identication, specically Edward Kaiser and Tatiana Laremore. Further thanks go to Matthew Langston of the Chemistry Department for running SDS-PAGE for us.

ARTICLE

(20) Bagheri, M.; Beyermann, M.; Dathe, M. Antimicrob. Agents Chemother. 2009, 53 (3), 11321141. (21) Haynie, S. L.; Crum, G. A.; Doele, B. A. Antimicrob. Agents Chemother. 1995, 39 (2), 301307. (22) Liu, C. X.; Zhang, D. R.; He, Y.; Zhao, X. S.; Bai, R. J. Membr. Sci. 2010, 346, 121130. (23) Van Loosdrecht, M. C. M.; Lyklema, J.; Norde, W.; Zehnder, A. J. B. Microbiol. Rev. 1990, 54 (1), 7587. (24) Gottenbos, B.; Grijpma, D. W.; van der Mei, H. C.; Feijen, J.; Busscher, H. J. J. Antimicrob. Chemother. 2001, 48, 713. (25) Lea, M. Bioremediation of Turbid Surface Water Using Seed Extract from Moringa oleifera Lam. (Drumstick) Tree. Current Protocols in Microbiology [Online] 2010, 1G.2.11G.2.14. http://www.currentprotocols. com/protocol/mc01g02 (accessed Nov 15, 2011). (26) Fernndez, R. The Marango Makes Magic in Water. Envo. a [Online] 1994. 156 http://www.envio.org.ni/articulo/1778 (accessed Nov 15, 2011).

REFERENCES
(1) Olson, M. E. Introduction to the Moringa Family in The Miracle Tree. In The Multiple Attributes of Moringa; Fuglie, L.J., Ed.; Technical Centre for Agricultural and Rural Cooperation, and Church World Service: Wageningen, The Netherlands and New York, USA, 2001; pp 1128. (2) National Research Council, Moringa. In Lost Crops of Africa: Vegetables; National Academy Press: Washington, DC, 2006; Vol II, pp 246267. (3) Foidl, N.; Makkar, H. P. S.; Becker, K. The potential of Moringa oleifera for agricultural and industrial uses in The Miracle Tree. In The Multiple Attributes of Moringa; Fuglie, L.J., Ed.; Technical Centre for Agricultural and Rural Cooperation, and Church World Service: Wageningen, The Netherlands and New York, USA, 2001; pp 4576. (4) Gadgil, A. Annu. Rev. Energy Environ. 1998, 23, 253286. (5) Bhuptawat, H.; Folkard, G. K.; Chaudhari, S. J. Hazard. Mater. 2007, 142, 477482. (6) Jabeeb, R.; Shahid, M.; Jamil, A.; Ashraf, M. Pak. J. Bot. 2008, 40 (4), 13491359. (7) Ndabigengesere, A.; Narasiah, K. S. Water Res. 1998, 32 (3), 781791. (8) Ghebremichael, K. A.; Gunaratna, K. R.; Henrisksson, H.; Brumer, H.; Dalhammar, G. Water Res. 2005, 39, 23382344. (9) Muyibi, S. A.; Evison, L. M. Water Res. 1995, 29 (12), 26892695. (10) Muyibi, S. A.; Alfugara, A. M. Int. J. Environ. Stud. 200, 60 (6), 617626. (11) Miller, S. M.; Fugate, E. J.; Craver, V. O.; Smith, J. A.; Zimmerman, J. B. Environ. Sci. Technol. 2008, 42, 42744279. (12) Beltran-Heredia, J.; Sanchez-Martin, J.; Delgado-Regalado, A.; Jurado-Bustos, C. J Hazard. Mater. 2009, 170, 4350. (13) Kumari, P.; Sharma, P.; Srivastava, S.; Srivastava, M. M. J. Ind. Microbiol. Biotechnol. 2005, 32, 521526. (14) Sharma, P.; Kumari, P.; Srivastava, S.; Srivastava, M. M. J. Hazard. Mater. 2006, 97 (2), 299305. (15) Prasad, R. K. J. Hazard. Mater. 2009, 165 (13), 804811. (16) Ndabigengesere, A.; Narasiah, K. S.; Talbot, B. G. Water Res. 1995, 29, 701710. (17) Suarez, M.; Haenni, M.; Canarelli, S.; Fisch, F.; Chodanowski, P.; Servis, C.; Michielin, O.; Freitag, R.; Moreillon, P.; Mermod, N. Antimicrob. Agents Chemother. 2005, 49 (9), 38473857. (18) Suarez, M.; Entenza, J. M.; Doerries, C.; Meyer, E.; Bourquin, L.; Sutherland, J.; Marison, I.; Moreillon, P.; Mermod, N. Biotechnol. Bioeng. 2003, 81 (1), 1320. (19) The National Center for Biotechnology Information Home Page. http://www.ncbi.nlm.nih.gov/protein/AAB34890.1 (accessed on November 14, 2011).
G
dx.doi.org/10.1021/la2038262 |Langmuir XXXX, XXX, 000000