Heterogeneity in Spatial P-distribution and Foraging Capability by Zea mays: Effects of Patch Size and Barriers to Restrict Root Proliferation within a Patch

Abstract

• Background and Aims Localized proliferation of roots in nutrient-enriched patches seems to be an adaptive response in many plants, but its function is still debatable. To understand the efficiency and limitation of foraging behaviour, the impact of patch size and the presence or absence of a barrier to root proliferation within phosphorus (P)-enriched patches was examined.

• Methods In pots filled with P-poor soil, six treatments of heterogeneous P supply were prepared: three patch sizes with or without a root barrier between patches. In addition, a homogeneous P supply treatment was also prepared. Irrespective of these treatments, each pot received the same total amount of P. Maize (Zea mays) was grown in each pot for 45 d in a greenhouse.

• Key Results P content and biomass were greatest in plants grown in the largest patch due to successful root proliferation, and were higher in the presence of a root barrier. Interestingly, plants preferentially developed adventitious nodal roots projecting from the stem into the P-enriched soil, particularly in the largest patch with a root barrier. Removal of the barrier reduced the P-uptake capacity per unit root surface area or volume in P-enriched patches, revealing that the P-uptake capacity per root can be suppressed even in P-rich soil if other portions on the root axis encounter P-poor conditions.

• Conclusions The results suggest that the efficiency of root morphological plasticity is largely determined by the size of the P-enriched patch. Furthermore, the results imply a novel aspect of P-uptake physiology that roots in heterogeneous P cannot demonstrate their potential capacity, as would be observed in roots encountering P continuously; this effect is probably mediated by an internal root factor.

INTRODUCTION

Phosphorus (P) is a limiting factor for crop yield in >30 % of the world's arable land (Vance et al., 2003), where a P fertilizer is often an essential requirement in order to obtain a sufficient yield. However, most of the P in such a fertilizer—often 90 % or more—is not taken up by the crop, but rather is retained by the soil in sparingly soluble forms (Stevenson and Cole, 1999). As a result, P fertilizer has been used at rates that far exceed the demands of the crops. Recently, attention has focused on the effects of excessive use of P fertilizer, not only in an economic but also an environmental context. Excessive application of P can cause eutrophication of lakes and rivers (Sharpley, 1995), while there are concerns about the depletion of commercially available phosphate rock, which is the most useful raw material of P fertilizers. It is therefore imperative that ways to improve the efficiency of P fertilizers for promoting plant growth are explored.

Regardless of the amounts of nutrients available, plant productivities can vary according to their spatial distribution. For example, a spatially heterogeneous supply of nutrients has been reported to result in superior uptake and growth of plants compared with a homogeneous supply (Birch and Hutchings, 1994; Fransen et al., 1998). Given a heterogeneous supply of compost, plant biomass was shown to vary according to the soil volume fertilized (Wijesinghe and Hutchings, 1997). Similar consequences were also observed when maize (Zea mays) was supplied with P in heterogeneous patches; the largest P patch resulted in the highest level of P uptake and the greatest biomass, independent of the P-input levels (Yano and Kume, 2005). Evidence is now accumulating that shows plants are strongly affected by heterogeneous, rather than homogeneous, resource distribution, even if the total resource supply remains the same (Hutchings and John, 2004).

Root plasticity plays a key role in obtaining heterogeneous resources. Plant roots can respond to a heterogeneous nutrient supply with great flexibility in terms of both morphology and physiology (for a review, see Hodge, 2004, 2006; Hutchings and John, 2004). For example, proliferation of lateral roots occurred on the mother root axis when P or nitrogen (N) was supplied locally (Drew, 1975). Similarly, root proliferation of higher-order laterals occurred following the localized colonization of first-order laterals by a symbiotic arbuscular mycorrhizal fungus (Yano et al., 1996), which delivers mainly labile P (Bolan, 1991; Shibata and Yano, 2003) and N in the ammonium form only (Tanaka and Yano, 2005) from soil to host plant. Prior to such morphological changes, root physiological parameters for the uptake of solutes were also found to be elevated in nutrient-enriched microenvironments (Drew and Saker, 1975; Jackson et al., 1990; van Vuuren et al., 1996).

However, there are various costs associated with phenotypic plasticity (DeWitt et al., 1998). If nutrient availability is short-lived or the nutrient patch only supplies limited resources for a plant, then the cost of the investment in new roots may not be repaid (Fitter, 1994). In such a case, plant growth may be depressed more with root plasticity than without it. Indeed, such a biomass depression of maize plants grown in smaller P-patches has been observed previously (Yano and Kume, 2005), in which more roots were distributed outside the patches (excessive proliferation), compared with plants grown in larger P-patches. If roots were not allowed to proliferate excessively outside P-patches, then the depressed growth might be recovered. To address this question, an attempt was made to restrict root proliferation within a P-patch by placing a root barrier between the patches. In the current study, the impact of P-patch size in the presence or absence of the root barrier was examined in order to understand the efficiency and limitation of the foraging behaviour.

MATERIALS AND METHODS

Plastic pots (160 mm in diameter, 200 mm deep) were filled with Andosol (2·5 kg of dry soil per pot) that contained a very low level of P (0·1 mg Truog-P kg⁻¹ dry soil). As shown in Fig. 1, (NH₄)₂SO₄ (0·5 g N pot⁻¹) and KCl (0·4 g K pot⁻¹) were only supplied to a quarter section of the soil column within each pot, which was completely isolated from the rest of the soil column by the insertion of thin plastic plates. The remainder of the column received superphosphate (0·2 g P pot⁻¹) in seven patterns as follows: homogeneous P and heterogeneous P in three different patch sizes [large (L), medium (M) and small (S)] in the presence (B) or absence (NB) of root barriers (Fig. 1). The plastic plate barriers (65 mm wide, 200 mm deep, 0·5 mm thick) were vertically installed into the pot, and the gaps between the plate and the pot were sealed with heavy clay to prevent root penetration. Each patch was filled with either P-enriched soil (0·2 g P kg⁻¹ dry soil) or P-poor soil (non-amended soil) so as to be alternately arranged. The pots without barriers contained a stainless-steel wire net (6-mm mesh size, 1·0 mm thick) which was initially sandwiched with plastic plates to fix each patch but the plates were pulled out before transplanting. As for homogeneous P supply, soil was amended with P at 0·1 g P kg⁻¹ dry soil to make the total P-input equivalent. Each pot had a central opening (30 mm diameter) to enable roots at the centre to become elongated without blockage.

Fig. 1.

Overview of the study design. Heterogeneous P was supplied in pots of different-sized patches; an isolated section contained N and K (NK-enriched soil), and the remaining soil was P-enriched or P-poor. The side view of the pots depicts the different rooting behaviours assumed in the presence (B) or absence (NB) of a root barrier. Each plant was transplanted to the centre of the pot, where the roots were able to randomly elongate in a radial direction, except in the NK section. The control treatment was a uniform supply of P. Each treatment contained the same total amount of nutrient per pot.

Seeds of Z. mays L. ‘Robust 30-71’ (Takii Seed Co. Ltd, Kyoto, Japan) were sown in nursery beds filled with sandy soil, and grown without fertilizer in a greenhouse. Uniform seedlings that developed a third leaf were transplanted into the pots described above. A cork borer (20 mm diameter, 100 mm long) was used to remove soil from the centre of each pot, into which space the seminal root system of each seedling was placed, and then the soil returned. Two or three adventitious roots from each seedling were placed into the NK section. Each plant was grown in a greenhouse for 45 d with irrigation of the whole soil surface at 3-d intervals. The soil moisture content was maintained at 40 % (w/w) by weighing each pot. The pots were arranged in a randomized complete block design with four replicates.

Shoots were harvested after the growth period. The soil column was taken from each patch using a knife, and the roots were gently washed free of soil under running water above a 2-mm sieve. The root sample collected from each patch was divided into two sub-samples: one for dry weight (d. wt) and one for morphological measurements (the latter was preserved in formalin : acetic acid : 70 % ethanol (FAA) at a ratio of 1 : 1 : 18, v/v/v). The number of nodal roots in each patch was counted on the stem base remaining at the centre of each pot, and the stem base was then included as part of the shoot. The shoot and root samples were dried at 80 °C for 48 h, weighed and then ground using a vibrating mill. The tissue was digested with nitric acid and perchloric acid, and the P concentration was determined using a colorimetric method (Watanabe and Olsen, 1965).

The sub-samples of roots preserved in FAA were stained with 0·2 % (w/v) methylene blue for 48 h, and then sandwiched between two transparent OHP (overhead projector) sheets taking care to prevent any overlap between the roots. A digitized image was acquired using a desktop scanner at 300 dpi and 256 greyscale. Using a macro program (Kimura et al., 1999), the root length and width were simultaneously measured on the images using the NIH Image software (version 1·60), and were used to calculate the root surface area and volume.

The data were analysed by two-way analysis of variance (ANOVA), in which the variation sources consisted of the patch size and the root-barrier treatments. A paired-sample t-test was also used to compare the root parameters inside and outside the P-enriched patches.

RESULTS

Plant growth and P acquisition

The initially uniform growth of the maize plants at transplantation became more varied over the 45-d treatment period for spatial P distribution. Figure 2 shows the dry weight (d. wt) accumulation in shoots and roots following each treatment. Significant differences in total d. wt were observed between plants depending upon the presence or absence of a barrier and upon patch size. These effects were additive, as the interaction between them was not significant. Similar results were also obtained for the shoot and root d. wts.

Fig. 2.

The effect of P patch size (L, M or S), and the presence (B) or absence (NB) of a root barrier, on the Z. mays biomass of the shoot and root, 45 d after transplantation. Data are shown as the mean ± s.e. (n = 4). The grey horizontal band indicates the mean ± s.e. (n = 4) of the control treatment that supplied P uniformly.

Dry weight increased in the presence of root barriers, and the increases were of a similar extent in each size of patch, while larger patches achieved a higher d. wt than smaller patches in the presence or absence of root barriers. All of the treatments with root barriers (B–L, B–M and B–S) resulted in a greater total d. wt than that of the control [8·66 ± 0·59 g (mean ± s.e.) for plants supplied with homogeneous P]. By contrast, only the largest patch without a barrier (NB–L) achieved a d. wt above the control value. NB–M achieved a similar d. wt to the control, while the d. wt of NB–S was lower than that of the control. There were no significant differences in shoot d. wt : root d. wt ratio between the treatments.

Figure 3 shows the P content of the harvested plants. Overall, the results were similar to those shown in Fig. 2, but there was no significant difference in the root P content between patches of different sizes (P = 0·9778). All of the treatments, including those with root barriers as well as the largest patch without a barrier, resulted in a higher P content than that of the control (12·1 ± 1·7 mg). The P content of both the NB–M and NB–S treatments varied within the control limits. A linear regression between the d. wt and P contents of the whole plants was obtained across the seven treatments (y = 0·441x + 3·617, r² = 0·857), indicating that P was the limiting factor for overall plant growth.

Fig. 3.

The effect of P patch size (L, M or S), and the presence (B) or absence (NB) of a root barrier, on the Z. mays P content of the shoot and root 45 d after transplantation. Data are shown as the mean ± s.e. (n = 4). The grey band indicates the mean ± s.e. (n = 4) of the control treatment that supplied P uniformly.

Root growth

Table 1 shows the root growth in terms of d. wt and total length. As the interaction between the two main effects was not significant for any of the parameters, each of the main effects could be regarded as additive. The presence of a barrier to the whole root system increased the d. wt significantly, while the differences due to patch sizes were less clear (P = 0·0565). In terms of the total root length of the whole root system, the effect of patch size, but not the presence of a root barrier, was significant; a longer root length was achieved with a larger patch size.

Table 1.

The effects of P-patch size and a root barrier on the root d. wt and length of Z. mays 45 d after transplantation

	Root d. wt (g plant−1)			Root length (m plant−1)
Treatment	Whole	NK section	P section	Whole	NK section	P section
P uniform	2·29 ± 0·07	0·94 ± 0·08	1·35 ± 0·14	308 ± 22	106 ± 17	202 ± 22
NB–L	3·18 ± 0·35	1·20 ± 0·10	1·98 ± 0·26	400 ± 28	128 ± 10	272 ± 22
NB–M	2·76 ± 0·45	1·12 ± 0·17	1·64 ± 0·28	327 ± 43	92 ± 14	234 ± 29
NB–S	2·14 ± 0·29	0·85 ± 0·06	1·29 ± 0·24	241 ± 26	56 ± 11	185 ± 18
B–L	4·04 ± 0·38	1·75 ± 0·26	2·29 ± 0·24	403 ± 31	139 ± 10	263 ± 25
B–M	3·40 ± 0·40	1·46 ± 0·12	1·93 ± 0·31	340 ± 37	117 ± 12	223 ± 26
B–S	3·16 ± 0·33	1·47 ± 0·14	1·69 ± 0·32	340 ± 40	130 ± 25	209 ± 15
Probability according to two-way ANOVA
Barrier	P = 0·0013	P < 0·0001	P = 0·1550	P = 0·1898	P = 0·0066	P = 0·9291
Patch size	P = 0·0565	P = 0·1566	P = 0·0932	P = 0·0150	P = 0·0339	P = 0·0229
Interaction	P = 0·8741	P = 0·6501	P = 0·9821	P = 0·3336	P = 0·0980	P = 0·7084

B and NB indicate the presence or absence of the root barrier, respectively.

L, M and S represent the size of the P patch (large, medium and small, respectively).

Data are shown as the mean ± standard error (n = 4). Values inside parentheses are the ratio to the mean of the P-uniform treatment.

The whole root system was sub-divided into those that developed in an NK section or in a P section. For roots grown in an NK section, both d. wt and the length were significantly increased in the presence of a root barrier. Roots grown in a P section showed a similar d. wt following all of the treatments, although the root length was reduced in smaller patches. The root barrier inserted into a P section did not significantly affect total root length. However, both the root barrier and the patch size were shown to have an effect on root growth within the P section as described below.

Root responses in patches

The P-treated area was further divided into P-enriched and P-poor patches. The measure of importance was whether plants could respond to the patch, resulting in a contrast in root growth within each plant. A paired-sample t-test was therefore performed to examine the significance of the contrasts on root d. wt, total root length and the number of nodal roots that penetrated each patch (Fig. 4). Significant increases in both d. wt and root length were observed in P-rich patches compared with P-poor patches following all but one of the six heterogeneous treatments (i.e. the effect of the B–M treatment on d. wt). It was therefore apparent that plants receiving any of the treatments were able to respond to the local supply of P, resulting in a contrast between P-rich and P-poor patches in each plant (Fig. 4B, C).

Fig. 4.

The effect of P patch size (L, M or S), and the presence (B) or absence (NB) of a root barrier, on the contrast of Z. mays root growth inside (shaded columns) and outside (open columns) the patch 45 d after transplantation. The number of nodal roots (A), root d. wt (B) and root length (C) in P-enriched and P-poor soil. Data are shown as the mean ± s.e. (n = 4). * and ** indicate significant contrasts between P-enriched and P-poor patches at P = 0·05 and 0·01, respectively, by paired-sample t-test.

To investigate the effect of P-rich and P-poor patches on the nodal roots of each plant, the number of nodal roots at the basal part near the stem was counted, to avoid duplicating a single nodal root axis across both patches. Interestingly, the B–L plants elongated their nodal roots approx. 3-fold preferentially into P-rich rather than P-poor patches. This preference, albeit at a lower magnitude, was also observed following the NB–L and B–S treatments (Fig. 4A). Thus, the plants responded to the local supply of P by elongating their adventitious roots from the stems.

ANOVA detected significant effects of the treatments, patch size and root barrier, for roots in P-enriched patches only and not for ones in P-poor patches (Fig. 4). Specifically, root d. wt in P-enriched patches was altered by both treatments, while root length in them was affected only by the patch size. As for the number of nodal roots, a significant effect of the barrier was detected in P-enriched patches.

The P-uptake capacity was estimated per unit morphological parameter of the roots in the P-rich patch. To eliminate an effect of varied root thickness, Table 2 shows the capacity not only per root d. wt and root length, but also per root surface area and root volume. Significantly higher P uptake per unit measure of root was achieved in the presence of the root barrier, but patch size had no effect on the P uptake. The ratio to the mean of the P-uniform control ranged from 1·3 to 1·8 for the treatments without a barrier, and from 1·7 to 2·4 for the treatments with a barrier.

Table 2.

The effect of P-patch size and a root barrier on the root P-uptake capacity in P-enriched soil

	P-uptake capacity
Treatment	Per root d. wt (μg mg−1)	Per root length (μg mm−1)	Per root surface area (μg mm−2)	Per root volume (μg mm−3)
P uniform	8·7 ± 1·1	0·058 ± 0·006	0·040 ± 0·004	0·263 ± 0·030
NB–L	14·5 ± 0·7 (1·7)	0·098 ± 0·010 (1·7)	0·069 ± 0·006 (1·7)	0·462 ± 0·032 (1·8)
NB–M	12·0 ± 0·9 (1·4)	0·072 ± 0·011 (1·2)	0·051 ± 0·007 (1·3)	0·346 ± 0·042 (1·3)
NB–S	13·5 ± 2·1 (1·5)	0·080 ± 0·006 (1·4)	0·055 ± 0·005 (1·4)	0·381 ± 0·041 (1·4)
B–L	15·1 ± 1·2 (1·7)	0·133 ± 0·007 (2·3)	0·087 ± 0·003 (2·2)	0·525 ± 0·017 (2·0)
B–M	19·3 ± 3·9 (2·2)	0·141 ± 0·023 (2·4)	0·092 ± 0·014 (2·3)	0·555 ± 0·077 (2·1)
B–S	17·7 ± 1·2 (2·0)	0·127 ± 0·007 (2·2)	0·089 ± 0·004 (2·2)	0·589 ± 0·031 (2·2)
Probabilities according to two-way ANOVA
Barrier	P = 0·0224	P < 0·0001	P < 0·0001	P = 0·0003
Patch size	P = 0·9002	P = 0·5923	P = 0·6545	P = 0·5956
Interaction	P = 0·2636	P = 0·3669	P = 0·2741	P = 0·1937

B and NB indicate the presence or absence of a root barrier, respectively.

L, M and S represent the size of the P patch (large, medium and small, respectively).

Data are shown as the mean ± s.e. (n = 4).

DISCUSSION

The effect of patch size

It was evident that the spatial distribution of P in the soil affected plant P acquisition, even though the same amount of P was available for the individual plants. The larger P patches caused enhanced P acquisition (Fig. 3), resulting in the production of a greater biomass (Fig. 2). These results were consistent with those of a previous report (Yano and Kume, 2005), although a different spatial distribution design was employed. Based on these results, supplying P fertilizer to a larger patch would be expected to achieve more efficient plant growth compared with smaller patches and/or a homogeneous P supply.

The observed differences in P acquisition among the treatments could be attributed to the foraging abilities of the roots in P-enriched soil, because the background soil contained a negligible amount of P (0·1 mg Truog-P kg⁻¹ dry soil). Numerous studies have demonstrated that the plastic responses of roots are important in foraging for heterogeneously distributed soil nutrients, and that both morphological and physiological responses are involved in root plasticity (for reviews, see Hodge, 2004, 2006; Hutchings and John, 2004). The current results show that the patch size strongly affected morphological characteristics of the plants such as root length in P-enriched patches (Table 1 and Fig. 4), but did not affect physiological parameters such as the P-uptake rate per root (Table 2). Thus, at least for maize, the successful acquisition of immobile P depends upon morphological root plasticity, which is strongly determined by the patch size irrespective of the absence or presence of a root barrier.

The effect of a root barrier

It had been predicted previously that the effect of patch size might be eliminated once a barrier against external root proliferation was in place. However, the current results did not support this prediction, as the effect of patch size on d. wt was significant for the treatments both with and without a barrier (Fig. 2), as well as the P content (Fig. 3). Looking at the root d. wt and root length in the P-rich patches (Fig. 4), they increased even in the treatments with the barrier, particularly in the case of the B–L plants in which preferential nodal root elongation into the P-rich zone was clearly observed (Fig. 4A). Probably due to this, the root barrier could not eliminate the effect of patch size.

The number of adventitious nodal roots of maize can be affected by the P status of the whole plant (Pellerin et al., 2000). The current results revealed that such a developmental response could occur more locally within an individual plant, strongly suggesting systemic control at the stem base where the nodal roots develop. Presumably, the stem base receives and responds to a signal from the roots exploring the P-enriched soil. While Bonser et al. (1996) found that P availability changed the orientation of the basal root-growth angle with respect to gravity, the present finding is a novel aspect of plant plasticity, in that the plants had the potential to preferentially elongate their nodal roots horizontally from the stem into a P-rich zone.

Interestingly, enhanced P acquisition was not caused by the increase in root length in P-enriched patches (Fig. 4C)—it is more likely to have been achieved with the increase in P uptake capacity per unit root length or volume (Table 2). This was surprising, given that the P content of each enriched patch was similar between all six heterogeneous treatments. In terms of the P content per soil volume, moreover, each P-enriched patch for the six heterogeneous treatments was 2-fold higher than the P-uniform control. Based on linear relationships between P added into soils and resin-exchangeable P (Kovar and Barber, 1988), it might therefore be expected that the physiological parameters of P uptake in plants receiving the six treatments could be 2-fold higher than the control. This was, indeed, observed for those treatments that included a root barrier, but not for those without a barrier.

It therefore appears that the potential capacity of the P-uptake rate per root is reduced in the absence of a root barrier. In addition, it is likely that such a physiological depression was based on an internal root factor, as the external soil environment of all the roots in the current study was enriched with P to a similar extent. In the absence of a root barrier, a root axis is able to explore both P-rich and P-poor zones; by contrast, if a root barrier is present, it continuously encounters P-rich soil (Fig. 1). This difference can affect aspects of the plant's internal environment, such as the xylem sap composition in the root axis. For example, the xylem of a root axis branching into both P-rich and P-poor zones would have a more dilute P concentration and associated signals than that of a root axis branching into P-rich soil only.

P uptake physiology of roots in P-enriched patches is strongly dependent on the photosynthetic activity of the shoot (Jackson and Caldwell, 1992). Even in a single root, physiological aspects such as proton flux can vary locally along the axis, and further the local flux strongly depends on the photosynthetic activity (Rao et al., 2000, 2002). These studies suggest that a diluted signal in the root xylem might reduce photosynthate partitioning to roots that explored both P-rich and P-poor zones (NB treatments) compared with those that were contained entirely within a P-rich zone (B treatments), resulting in a systemic depression of P-uptake capacity in the absence of a root barrier.

Conclusions

It has been demonstrated that the success of root foraging, which is reliant upon root plasticity, largely depends upon the size of the P patch. This is based primarily on the difference in the root volume that is available for exploration within each patch. The observed increase in root volume was particularly apparent in the largest patches. This occurred even in the presence of root barriers, in which the nodal root elongated primarily on the inside, rather than the outside, of the patch. In addition, the P-uptake capacity per unit volume was lower in the absence of the barrier than in its presence, probably by means of an internal root factor rather than the external soil environment.