Colloidal suspensions of clay or titanium dioxide nanoparticles can inhibit leaf growth and transpiration via physical effects on root water transport

Authors


P. M. Neumann. Fax: +972 4 8228898; e-mail: agpetern@tx.technion.ac.il

ABSTRACT

A laboratory investigation was conducted to determine whether colloidal suspensions of inorganic nanoparticulate materials of natural or industrial origin in the external water supplied to the primary root of maize seedlings (Zea mays L.) could interfere with water transport and induce associated leaf responses. Water flow through excised roots was reduced, together with root hydraulic conductivity, within minutes of exposure to colloidal suspensions of naturally derived bentonite clay or industrially produced TiO2 nanoparticles. Similar nanoparticle additions to the hydroponic solution surrounding the primary root of intact seedlings rapidly inhibited leaf growth and transpiration. The reduced water availability caused by external nanoparticles and the associated leaf responses appeared to involve a rapid physical inhibition of apoplastic flow through nanosized root cell wall pores rather than toxic effects. Thus: (1) bentonite and TiO2 treatments also reduced the hydraulic conductivity of cell wall ghosts of killed roots left after hot alcohol disruption of the cell membranes; and (2) the average particle exclusion diameter of root cell wall pores was reduced from 6.6 to 3.0 nm by prior nanoparticle treatments. Irrigation of soil-grown plants with nanoparticle suspensions had mostly insignificant inhibitory effects on long-term shoot production, and a possible developmental adaptation is suggested.

INTRODUCTION

World production and industrial uses of engineered materials comprised of nanosized particles with diameters ≤100 nm are rapidly increasing. Because nanoparticles may form relatively stable colloidal suspensions in water, there are increasing concerns that their unregulated accumulation in environmental water sources could have toxic effects on plants, animals and humans (e.g. Yang and Watts 2005; Murashov 2006; Nel et al. 2006; UNEP 2007; Blaser et al. 2008). Another possible danger is that suspended colloid particles might interact physically with cell walls or cell membranes, thereby reducing their permeability to external water sources. Indeed, several reports suggest that solutions or suspensions of nanosized materials such as large polymer molecules, China ink pigments or gold nanoparticles might reduce water flow into plant cells and tissues by accumulating in their cell walls (Neumann 1987; Chazen, Hartung & Neumann 1995; Shane, McCully & Canny 2000; Choat et al. 2003; Ranathunge et al. 2004; Proseus & Boyer 2005). Soil waters routinely contain natural organic and inorganic materials with particle sizes in the nanometer range. For example, colloidal suspensions of inorganic clays can be found in soil waters at concentrations of up to 1 g L−1 (DeNovio, Saiers & Ryan 2004). Associations between such clays and low-molecular weight contaminants can facilitate undesirable contaminant transport through the soil (Whalley & Mullins 1991; DeNovio et al. 2004). Soil clays can also have well-known beneficial effects on plant growth, which are related to their mineral nutrient binding capacity (Jenny & Overstreet 1938; Croker et al. 2004). We report here an investigation into the overlooked possibility that colloidal suspensions of clays, or other nanomaterials, may have additional and adverse effects on the porosity of root cell walls. More specifically, these materials may accumulate at the external root surfaces of transpiring plants, thereby reducing root hydraulic conductivity and plant availability of external water sources.

The maintenance of a root water transport capacity which can meet ongoing water usage for evaporation and growth in the shoots is essential in transpiring plants. The water supply capacity of plant root systems is the product of the total area of roots available for water transport and root-specific hydraulic conductivity. Any reductions in the water supply capacity of the root system, through inhibitory environmental effects on root hydraulic conductivity or root size, may result in stressful shoot responses such as increased xylem tension, xylem embolism, leaf growth inhibition, stomatal closure, associated reductions in transpiration and photosynthesis, wilting and eventual plant death by desiccation (cf. Tyree & Sperry 1988; McCully 1999; Sack & Holbrook 2006; Neumann 2008).

In order to flow through higher plants, external water sources and any materials dissolved or suspended therein must first pass through the cell walls of the outer epidermal layer of intact roots. Subsequent radial passage to the root xylem tissues likely involves parallel transport through both apoplastic cell wall pathways and symplastic cell-to-cell pathways (Steudle & Peterson 1988). The epidermal cell walls through which water must pass during its initial entry into roots are based on porous networks of polysaccharide fibre matrices (Carpita & Gibeaut 1993). The pores which penetrate these fibre matrices have reported mean diameters in the range of 3–8 nm, although larger pores can also be present (cf. Carpita et al. 1979; Baron-Epel, Gharyal & Schindler 1988; Chesson, Gardner & Wood 1997). Cell walls at root surfaces should therefore act as filters which allow preferential passage through the pores of water molecules and small solutes, but hinder or block the passage of colloidal materials with particle diameters greater than 8 nm. We hypothesized that the presence of colloidal suspensions of nanoparticles of natural or industrial origin in the water supplied to plant roots could inhibit root hydraulic conductivity, leaf growth and transpiration.

To test these hypotheses, we carried out a laboratory investigation of the extent to which the presence of colloidal suspensions of inorganic nano-particulate materials of natural or industrial origin in the external water supply could interfere with external water transport through the intact surfaces of excised maize (Zea mays L.) primary roots with intact apices. In addition, we investigated the possibility of associated effects, caused by reductions in water supply, on leaf growth and transpiration in whole seedlings. Suspensions of inorganic bentonite clay, with minimum plate dimensions of constituent particles ranging from 1 to 60 nm, and maximum dimensions of circa 6000 nm, were used to measure effects of water-borne particles of natural origin. Suspensions of inorganic TiO2 nanoparticles (Degussa P-25, Evonik Degussa GmbH, Essen, Germany) with mean diameters of 30 nm were used to measure the effects of a more uniform, industrially produced material. Both materials were found to inhibit root hydraulic conductivity, leaf growth and transpiration in maize seedlings.

MATERIALS AND METHODS

Plant material

Seeds of Z. mays L. (cv. 32p75; Milchan Brothers, Ramat Gan, Israel) were grown essentially as in Chazen et al. (1995). The seeds were germinated in the dark at 27 °C on filter paper wetted with 0.4 mm CaC12. After 2 d, the germinated seedlings were transferred to a growth chamber at 27 ± 2 °C. Relative humidity (RH) in the growth chamber varied between 35% by day and 60% at night. Light in the growth chamber, provided by mixed incandescent/fluorescent lamps during a 12 h photoperiod, was at 35 W m−2 photosynthetically active radiation (PAR). One-centimetre-long roots were inserted through narrow holes in polystyrene floats on trays containing 8 L of a well-aerated 0.1 strength modified Hoagland solution, so that only the primary root developed in the solution.

Root hydraulic conductivity

A 6 cm section of primary root with intact apex was excised, and 1 cm at the basal (cut) end was fitted into glass capillary tubes with scale marks. Silicon grease and Parafilm (Parafilm Company, Menasha, WI, USA) wrapping were used to carefully seal the junction. The remaining 5 cm of root with intact apex protruded from the capillary and was immersed vertically in 1 L lightproof containers (to prevent any photo-oxidative interaction between UV light and TiO2). A row of up to 10 roots spaced at 1.5 cm intervals, to minimize contact between roots, was inserted in each container. The containers were filled with continuously stirred solutions of 0.1 mm CaCl2 without or with additions of Na bentonite clay (Spectrum Chemical Company, Gardena, CA, USA, prepared and purified according to the protocol in Parfit & Greenland 1970) or of TiO2 (P-25; Degussa Company) each at 0.3 or 1 g L−1, pH 6.8 and 27 °C. Flow from the external water source through the intact epidermal tissues of each exposed root was followed by measuring the rise of water menisci produced in the protruding capillaries after pressurizing the root container to 10 KPa. Hydraulic conductivity was generally based on flow rates assayed over intervals of 10 min. Roots which gave pressurized flow rates >0.3 µL min−1 in the first 30 min were considered leaky and discarded. The 0.3 µL min−1 cut-off was based on previous observations showing that aberrant transport of high-molecular weight dextran blue (0.4 g L−1) polymer solutions through to the glass capillary accompanied these high flow rates; dextran blue did not penetrate through the cell walls of the exposed epidermal tissues of non-leaky root capillary set-ups. Thus, water and suspended colloids could only enter by passing through the epidermal layers of the intact apical root sections.

For killed root assays, the protruding 5 cm of roots fitted to capillaries as above was contacted with hot (80 °C) ethanol for 1 min and gently rinsed with water prior to the flow assay. Killed roots which gave pressurized flow rates >0.55 µL min−1 were considered leaky and discarded. Here, too, the cut-off rate was based on preliminary trials with dextran blue solutions. Axial hydraulic conductivity of the xylem was found to increase from 2 cm behind the tip (not shown). Thus, only the 3-cm-long exposed region of more mature tissue starting 2 cm behind the tip was assumed to contain open xylem and was taken to represent the effective root region for water uptake in the calculation of hydraulic conductivity. Hydraulic conductivity was calculated as cubic metre solution transported per square metre effective root area per second per megapascal applied pressure.

Leaf and root growth rates

Elongation of emerging primary leaves of uniform maize seedlings was tracked at 5–10 h intervals for 48 h by following the length increases. Growth rates were based on measurements made during the period of linear elongation. Primary root elongation rates were based on increases in the length of ink-marked roots as in Fan & Neumann (2004) over 3 d. The roots grew in well-aerated and stirred 0.1 strength modified Hoagland solution.

Transpiration

Transpiration rates were assayed gravimetrically using whole seedlings selected for uniform size. These had open first and second leaves, and an emerging third leaf. The seedling roots were loosely sealed into lightproof plastic vials which were three-fourths filled with 0.1 mm CaCl2 with or without appropriate colloidal suspensions at 1 g L−1. Plants with their foliage removed and the remaining stump capped with Parafilm were used as controls. Transpiration was assayed by following weight loss for 3 h with an electronic balance, subtracting control values and dividing by leaf surface area. A factor relating 1 g of leaf fresh weight to 65.1 cm2 of open leaf area was determined for 30 plants and used to convert leaf fresh weight of assayed plants to leaf area (circa 13 cm2 per plant). Oxygen levels in the root media at the beginning and end of 3 h transpiration assays were measured using a dissolved O2 probe (Cyberscan, DO 300; Eutech Instruments Europe BV, Nijkerk, the Netherlands). Similar decreases of circa 20% were measured after 3 h in solutions with or without nanoparticle additions.

Root cell wall pore size

The mean diameters of the pores limiting particulate transport through the cell walls of living maize roots were determined by observation of cytorrhysis (i.e. root collapse) (Carpita et al. 1979). This was induced by sequential exposure to solutions of PEG molecules with hydrodynamic diameters ranging from 0.9 to 7.2 nm (Kuga 1981). The concentrations of PEG 200, 1500, 4000, 6000 and 10 000 (57.84, 180.71, 204.17, 201.62 and 210.54 g PEG per 1000 g water, respectively) were fixed using the calibration of Money (1989) to give the same solution water potential of −0.7 MPa for all PEG treatments. Batches of 10 roots infused for 5 h with control solution (0.1 mm CaCl2), or roots treated with colloids for 4 h after 1 h in control solution, were gently rinsed and transferred to Petri dishes containing 50 mL of PEG 200 solution, and then transferred at 50 min intervals to PEG solutions with higher molecular weights and hydrodynamic diameters. When the external PEG molecules were too large to penetrate the pores in the cell walls, an osmotically induced efflux of cell water resulted in inward collapse and root cytorrhysis. Flattened, ribbon-like roots could then be observed through a binocular microscope, and average pore diameters were calculated.

Long-term effects on potted plants

Plants were grown on a local clay soil in 1 L pots with drainage holes at base and sides. Each treatment consisted of three replicate pots with 6 plants per pot. The pots were sub-irrigated at 48 h intervals by standing in suspensions of bentonite or TiO2, at 1 g L−1 of 0.1 strength modified Hoagland solution. Shoots were excised at the base and weighed after 6 weeks of treatment.

Statistics

Differences between individual treatment and control values were evaluated by paired two-sample t-tests using the analysis tool box in Excel.

RESULTS

Nanoparticle effects on root hydraulics

The assays of water flow and hydraulic conductivity were based on the use of excised roots in a custom-built system which allowed simultaneous assay of up to 10 roots. The variability between the flow rates of individual roots and different batches of roots could therefore be reduced by simultaneously assaying relatively large numbers of roots from the same batch. Moreover, the sensitivity of the assay was such that the same set of roots could be assayed before, during and after exposure to colloid suspensions, all within 5 h. Rates of pressurized water transport through the intact epidermal tissues of the excised apical root sections and out through the cut base were found to remain linear for 5 h (Fig. 1). All the hydraulic assays on excised roots were therefore completed within this period.

Figure 1.

Flow through excised roots remains linear for 5 h. A 1 cm length at the basal cut end of a 6-cm-long apical section of the primary root of maize seedlings was sealed into a marked capillary tube. The protruding 5 cm, with intact apex, was immersed vertically in a continuously stirred solution of 0.1 mm CaCl2 in a custom-built container for up to 10 roots spaced at 1.5 cm intervals. The external solution was pressurized to 10 KPa, and rate of water transport through the intact root epidermis and onwards into the capillary was followed over time. Means ± SE, n = 10. Points joined by computer-fitted linear regression.

The kinetics of inhibitory effects on water flow through the roots and, hence, on root hydraulic conductivities, after additions of two concentrations of colloidal suspensions of either bentonite or TiO2 are shown in Fig. 2. The upper traces in each graph suggest that additions of both materials at 0.3 g L−1 caused only marginal reductions in hydraulic conductivity. In contrast, the lower traces show that at the higher concentrations of 1 g L−1, both bentonite and TiO2 rapidly decreased root hydraulic conductivities in an apparently progressive manner. In summary, nanoparticle suspensions had flow inhibitory effects on roots which appeared to be concentration dependent and progressive.

Figure 2.

Kinetics of apparently progressive decreases in root hydraulic conductivity with increases in external concentrations and exposure times to colloidal bentonite or TiO2. Upper and lower lines in each graph represent colloid concentrations of 0.3 and 1.0 g L−1, respectively. Other details as in Fig. 1. Means ± SE, n = 7–10.

A summary of the pooled mean values of hydraulic conductivities from three separate experiments is presented in Table 1. It shows the inhibitory effects of colloid additions on root hydraulic conductivity, and degrees of recovery by the same roots after colloid removal. The fact that the inhibitory effects of relatively short exposures of live roots to nanoparticles were not completely reversed suggests that some of the particles became irreversibly attached to cell walls at the root surface during the infiltration treatments.

Table 1.  Rapid inhibitory effects of external nanoparticle additions on hydraulic conductivity of live or killed roots, and reversal by transfer to particle-free solution
 Root hydraulic conductivity (m s−1 MPa−1 . 10−7)
ControlBentonite addedBentonite removedControlTiO2 addedTiO2 removed
  1. Pooled means ± SE for 20–25 roots from three experiments in which either bentonite or TiO2 suspensions were added and then removed from the roots. Killed roots are after hot alcohol disruption of symplastic membrane barriers to flow. Hydraulic conductivities based on flow over last 20 min of 40 min of control infusion, 70 min of colloid infusion and 30 min of infusion without colloids, respectively. Different letters in rows indicate that treatment means differed from the appropriate control values at P < 0.05. Other details as in Fig. 1.

Live11.45 ± 0.85a6.86 ± 0.75b8.50 ± 0.73b15.95 ± 1.40A10.66 ± 0.85B12.10 ± 0.75B
Killed19.16 ± 1.28a16.05 ± 0.86b19.16 ± 1.28a20.16 ± 2.03A15.95 ± 1.51B18.01 ± 1.57B

Further evidence for a physical interaction between root cell walls and colloidal particles was provided by experiments in which the effects of bentonite or TiO2 on flow through hot alcohol extracted root ‘ghosts’ were assayed. The hot alcohol treatment was expected to disrupt lipid-based cell membrane barriers to symplastic flow while leaving the polysaccharide polymers of the cell walls and apoplastic flow relatively intact. The results for killed roots in Table 1 show that killed roots had, as might be expected, higher hydraulic conductivities than roots with intact membranes. The complete recovery of hydraulic conductivity after bentonite removal from killed roots, as compared with live roots, may reflect additional effects of the hot alcohol treatment on cell wall pore dimensions and/or surface characteristics. More importantly, these results show that colloidal suspensions of both bentonite and TiO2 could inhibit the hydraulic conductivities of the killed roots in a similar manner to live roots. Thus, a presumably physical interaction between colloidal particles and root cell walls appeared to be involved in inhibiting root hydraulic conductivity.

In order to further investigate the phenomenon of root clogging by nanoparticle suspensions, the mean pore diameters of root cell walls were assayed after 4 h of flow inhibitory exposure to the particles. We reasoned that the accumulation of colloidal particles filtered out at the cell wall surfaces might reduce the mean diameter of pores in the cell wall matrix. Table 2 shows clearly that prior treatment with either bentonite or TiO2 induced reductions in the mean pore diameters measured in the roots. In summary, colloidal nanoparticles of both bentonite and TiO2 suspended in the water flowing into roots appeared to attach to root cell wall surfaces, thus reducing effective cell wall pore diameters and root hydraulic conductivities.

Table 2.  Prior root treatment with nanomaterials reduces cell wall pore diameters
RootCell wall pore diameter (nm)
  1. Following flow assay for 5 h, roots were rinsed briefly in 0.1 mm CaCl2 to remove loosely attached materials, and then assayed for cell wall pore diameter as in Materials and methods (mean ± SE, n = 10). Different letters in columns indicate treatment means were significantly different from control values, P < 0.05.

Control6.6 ± 0.3a
TiO23.0 ± 0.2b
Bentonite3.1 ± 0.2b

Whole-plant effects

If root-to-shoot supplies of water fall below shoot requirements for optimal functioning, the shoot may exhibit symptoms of water stress such as stomatal closure, which reduces transpirational water losses and growth inhibition, which reduces water requirements for cell expansion. Both transpiration and leaf growth in maize plants are sensitive to reductions in root water supply (e.g. Chazen & Neumann 1994; Voisin et al. 2006), and external additions of colloidal suspensions of bentonite and TiO2 had clear inhibitory effects on the hydraulic conductivity of maize primary roots. Because primary roots are a vital part of the pathway of water transport into maize seedlings, it was of interest to determine whether reductions in water supply caused by additions of suspensions of bentonite or TiO2 at 1 g L−1 to the hydroponic media bathing the primary roots of intact seedlings could affect shoot transpiration and leaf growth. The kinetics of a rapid onset of inhibitory effects on transpiration induced by bentonite or TiO2 additions to the root medium are shown in Fig. 3. The pooled means of three experiments of the type shown in Fig. 3 revealed that after 3 h, transpiration was significantly reduced from the control values (P < 0.05) by exposures to either bentonite or TiO2 (Table 3). The same table shows that mean rates of primary leaf growth also showed small, but significant (P < 0.05), reductions in response to these treatments. Conversely, the rates of primary root elongation were little affected during 3 d treatments. Thus, there was no evidence for any short-term acclimation via nanoparticle-stimulated increases in root area. More importantly, the roots retained a healthy appearance despite continuous contact with colloidal suspensions of bentonite or TiO2. This suggests that the colloids assayed here did not act to produce their stressful effects on the shoots of maize seedlings via growth inhibitory or toxic effects on the root.

Figure 3.

Rapid inhibition of transpiration following external treatment of roots of whole seedlings with colloidal suspensions of bentonite or TiO2. Shoot transpiration was followed after additions of suspensions of bentonite or TiO2 at 1 g L−1 to the root media of intact hydroponic maize seedlings at the three leaf stages (means ± SE, n = 5).

Table 3.  Bentonite and TiO2 in the root media inhibit transpiration and leaf growth, but not root growth in hydroponic maize seedlings
 ControlTiO2 addedBentonite added
  1. Values are the pooled mean ± SE for three similar experiments comprised of five replicate plants each for transpiration assay, 18–23 replicate plants each for leaf growth assay and 16 plants each for root growth assay. Different letters in horizontal rows indicate significant differences between each treatment mean and the control, P < 0.05.

Transpiration (mg H2O cm−2 h−1)10.90 ± 1.70a8.10 ± 0.90b7.55 ± 0.60b
Leaf GR (mm h−1)1.00 ± 0.01a0.85 ± 0.03b0.95 ± 0.02b
Root GR (mm h−1)2.87 ± 0.14a2.65 ± 0.12a2.79 ± 0.12a

Despite the clear inhibitory effects of nanoparticle treatments on primary leaf growth in hydroponic maize seedlings, an investigation of possible long-term shoot growth effects on potted maize plants grown for 6 weeks in a clay soil repeatedly irrigated with nutrient solutions containing either bentonite or TiO2 at 1 g L−1 revealed only minor inhibitory effects as compared with control treatments without nanoparticles (Table 4). These effects, with the exception of the bentonite effect on shoot dry weight, were not statistically significant.

Table 4.  Shoot growth of potted maize plants after 6 weeks in a heavy soil irrigated with suspensions.of either bentonite or TiO2
 Shoot fresh weight (g)Shoot dry weight (g)
  1. Plants were grown on a local clay soil in 1 L pots with drainage holes at base and sides, and sub-irrigated at 48 h intervals with suspensions (1 g L−1) of bentonite or TiO2 in 0.1 strength nutrient solution. Treatments consisted of three replicate pots with 6 plants per pot. Shoots were excised at soil level and weighed after 6 weeks of treatment. Means ± SE. Identical letters in the shoot fresh weight column indicate that neither bentonite nor TiO2 treatments caused significant reductions relative to control value (P < 0.05, Student's t-test). Different letters in the shoot dry weight column indicate that bentonite treatment caused a significant small reduction.

Control16.08 ± 1.01a2.94 ± 0.20a
Bentonite15.18 ± 1.22a2.53 ± 0.14b
TiO215.10 ± 1.74a2.74 ± 0.17a

DISCUSSION

Mechanism of flow inhibition

A mean root cell wall pore diameter of 6.6 nm was found for maize primary roots. This value falls within the range of values reported by others using isolated cells or cell wall powders from different tissues and plants (e.g. Carpita et al. 1979; Baron-Epel et al. 1988; Chesson et al. 1997). Note that the reported minimum plate dimensions of approximately 60 × 1 nm for bentonite clay particles (Whalley & Mullins 1991; Cadene et al. 2005) and the mean 30 nm diameter of the TiO2 nanoparticles (Ohno et al. 2001) are far greater than cell wall pore dimensions of 6.6 nm found in maize roots. Thus, when present in the external solution, such particles would likely be too large to effectively penetrate through the cell wall pores in the outer epidermal layers of the roots. Particle accumulation at the wall interface might then reduce wall conductivity through a build-up of interfacial viscosity or other boundary layer effects. For example, particles rejected at the surface of the roots may physically limit root water transport by forming surface ‘cake’ layers that decrease the hydraulic conductivity of the epidermal cell walls. The fact that both bentonite and TiO2 showed some irreversible binding to the cell walls of live roots supports this suggestion. Moreover, inhibition of flow resulting from formation of surface cake layers is well known to occur whenever colloidal solutions are filtered through synthetic nano-filtration membranes (e.g. Lee, Cho & Elimelech 2005; Boussu et al. 2007). Finally, the finding that colloidal suspensions of external colloids reduced water flow through the maize root cell walls remaining after hot alcohol extraction and the associated disruption of phospholipid membrane barriers, clearly shows that interactions between external colloid particles and root cell walls alone can reduce root hydraulic conductivity.

Some previous investigations of interactions between large polymer molecules or colloid particles and plant cell walls are consistent with our findings. Thus, Baron-Epel et al. (1988) were able to follow the diffusion of fluorescent-labelled dextrans with diameters of up to 4.6 nm through the walls of cultured soy bean cells. Dextrans and proteins with larger diameters showed hindered or zero passage. Similarly, Chazen et al. (1995) showed that solutions of polyethylene glycol 6000 (with a mean hydrodynamic diameter of 5.3 nm) caused an additional, non-osmotic inhibition of pressurized water flow through killed roots.

More recently, Ranathunge et al. (2004) showed that 60 min of pressurized infusion of a colloidal suspension of China ink particles with diameters of circa 50 nm through the outer part of rice roots could reduce water flow by 25 to 30%. Proseus & Boyer (2005) used confocal laser microscopy to track the pressurized movement of aqueous suspensions of gold nanoparticles of varying diameters, into or across algal cell walls. They showed that gold nanoparticles with circa 10 nm diameter could not penetrate through algal walls, even when 0.5 MPa pressure was applied. Similar results were obtained in investigations of the thin internal walls of xylem boundary pits by Shane et al. (2000) and Choat et al. (2003). These studies did not, however, reveal the inhibitory effects on water flow and shoot development of external nanomaterials supplied to the intact surfaces of plant roots.

Physiological and environmental implications

This appears to be the first report to establish that colloidal suspensions of nano-particulate materials in the root media can reduce the hydraulic conductivity of maize primary roots and induce symptoms of water stress (reduced transpiration and leaf growth) in the shoots of single-rooted young seedlings. These symptomatic responses were rapidly initiated and appeared to involve physical rather than toxic interactions between nanoparticles and roots. Thus, both physical and toxic effects of nanoparticles should be considered when evaluating their potential environmental impacts on plants. For example, as fresh water availability decreases, recycled waste waters which can contain high concentrations of suspended colloids and dissolved biopolymers will be increasingly used for crop irrigation in drought-prone regions of the world (Tal 2006; Neumann 2008). It is conceivable that the use of such recycled waters, or waters contaminated with industrial nanoparticle wastes, will cause significant root clogging and thereby limit root water uptake under field conditions.

Because of the short-term inhibition of leaf growth rates by bentonite and TiO2 in the hydroponic seedling experiments, we expected to see a major inhibition of shoot growth in soil-grown plants regularly irrigated with suspensions of bentonite or TiO2 for 6 weeks. Albeit, only minor inhibitions of shoot fresh and dry weights were apparent and most of these were not statistically significant. This is consistent with a report by Croker et al. (2004). They showed that additions of mineral nutrient-enriched bentonite to tropical sandy soils at rates equivalent to 40 tons per hectare consistently increased shoot growth of potted sorghum (Sorghum bicolor) plants. In these experiments, however, any inhibitory effects of the bentonite on shoot growth could have been negated by growth stimulatory responses to associated increases in mineral nutrient availability.

One hypothetical explanation for the absence of major long-term inhibitory effects of the nanoparticle treatments on the shoot growth of potted maize plants is that passage through the soil effectively filtered out the colloidal nanoparticles suspended in the water supply, thereby decreasing their effective concentrations and rates of accumulation at the root surfaces of the transpiring plants. Note that the colloid concentration of 1 g L−1 used in the experiments represents the high end of reported soil water concentrations (DeNovio et al. 2004) and that lower concentrations were much less effective in inhibiting root hydraulic conductivity in short-term assays (cf. Fig. 2). Nevertheless, this explanation seems unlikely because the high mobility of colloidal suspensions of particles through various soil types has been repeatedly confirmed (e.g. Noack, Grant & Chittleborough 2000; DeNovio et al. 2004).

An alternative explanation for the observed lack of major long-term inhibitory effects of nanoparticle treatments on shoot growth relates to the fact that the overall water supply capacity of plant root systems is determined by the specific hydraulic conductivity of the individual roots and the total number of roots, which increases with time. Thus, the time-dependent development of an extensive root system with adventitious roots, fine roots and root hairs could provide an excess capacity for water transport, at least under well-watered conditions. For example, an extensive root system which, because of nanoparticle clogging of individual roots was operating at only 80% of unhindered hydraulic capacity might still be able to provide the full water requirements of the transpiring shoot. This suggestion is consistent with root-trimming experiments indicating that the intact root system in wheat plants had a water supply capacity which exceeded the water demands of the shoots (Vysotskaya et al. 2004). Moreover, the fact that plants can grow successfully in clay soil environments under natural conditions also suggests that they are able to adapt to any adverse effects of colloids in the water supply. Conceivably, the development of large root systems with an excess of water transport capacity facilitates plant adaptation to partial root clogging by the colloidal nanoparticles found naturally in soil waters.

In conclusion, this report shows for the first time that the presence in the external water supplies of hydroponic maize seedlings of colloidal suspensions of inorganic nanomaterials of natural or industrial origin can lead to accumulation at the cell wall surfaces of the primary root and consequent inhibition of cell wall pore size, water transport capacity, leaf growth and transpiration. The developmental production of large root systems with an excess of water transport capacity may represent a plant adaptation which minimizes the potentially stressful effects of root clogging by nanoparticles naturally present in soil solutions.

ACKNOWLEDGMENTS

Research supported by Loewengart and FMW funds. We thank Prof. Robert Armon for a gift of Degussa P-25 nanoparticles, and the reviewers for constructive suggestions.

Ancillary