•Phosphorus (P)-limited plants produce higher amounts of root phosphatases, but research has mostly focused on phosphomonoesterases (PMEs). Because phosphate diesters can form a significant proportion of organic P in wetlands, we aimed to determine whether wetland plants produce both root PMEs and root phosphodiesterases (PDEs), and, if so, what factors influence activities of these enzymes.
•We measured the activities of root PMEs and PDEs colorimetrically in a wide range of macrophytes from natural and P-enriched wetlands. Hydrolyzable P in sediments was analyzed using commercially available PMEs and PDEs.
•In all species, both root PMEs and PDEs were always present, and their activities were closely correlated. Sedges and broadleaved emergents had the highest activity of both enzymes, while those of floating-leaved plants were the lowest. Redundancy analysis revealed close association between root enzymes and the proportion of monoesterase- and diesterase-hydrolyzable dissolved unreactive P. Both enzymes were positively correlated with root tissue N : P ratio.
•Both plant and sediment traits were important when explaining differences in enzyme activities. Although the activities are related to ambient P regime, the relationship was not close enough to use root enzymes as reliable predictors of dissolved unreactive P that is hydrolyzed by sediment phosphomono- and diesterases.
Plants possess various functional traits that reflect adaptations to local environments. In phosphorus (P)-limited environments, the traits related to root P foraging strategies are critical for plant survival. P is an essential nutrient available to plants as the orthophosphate anion (Pi). The availability of Pi in natural ecosystems is often extremely low and plants have to rely on accessing Pi from various organic phosphates. The organic P in soils is usually dominated by a mixture of phosphate monoesters such as mononucleotides and inositol phosphates, and phosphate diesters, which are mainly nucleic acids and phospholipids, in varying proportions to each other (Turner, 2008). Enzymes known as phosphatases hydrolyze the orthophosphate anion (Pi) from orthophosphate mono- and diesters (Duff et al., 1994; Richardson et al., 2009). Phosphate monoesters are the most available form of organic P in the soil, being weakly sorbed and requiring only hydrolysis by phosphomonoesterase (PME), a common plant enzyme, to release Pi for plant uptake (Condron & Tiessen, 2005). Phosphate diesters are considered less available than phosphate monoesters because they must be hydrolyzed by both PME and phosphodiesterase (PDE) to release Pi for plant uptake (Turner, 2008). Although abundant in various biological materials and therefore constituting most of the organic P inputs to soils, phosphate diesters usually do not accumulate in terrestrial ecosystems (Condron & Tiessen, 2005) because of rapid degradation (Bowman & Cole 1978). However, there is increasing evidence that phosphodiesters can accumulate in cold or wet soils (Leake & Miles, 1996; Makarov et al., 2002; Turner & Newman, 2005) as a result of their higher recalcitrance under these conditions (Condron et al., 1990; Cheesman et al., 2010). Extra-cellular phosphatases in the soil are, apart from plants, also produced by a variety of other organisms, most notably the rhizosphere bacterial community and mycorrhizal fungi (Neumann & Römheld, 2007). In plants, the sites of extracellular phosphatase activity are usually located on the outer surfaces of epidermal cells and root apical meristems (Vance et al., 2003). Phosphatase induction during P deficiency is a universal response to P limitation in higher plants (Phoenix et al., 2003; Raghothama & Karthikeyan, 2005; Rejmánková & Snyder, 2008).
Phosphomonoesterases on root surfaces, in water, or in soils have been studied much more frequently than PDEs, despite the widespread occurrence of PDE (Turner & Haygarth, 2005; Whitton et al., 2005; Yamaguchi et al., 2005; Ellwood et al., 2008), which is a discrepancy caused by the focus of researchers on agriculturally important plants growing in soils where phosphomonoesters are the dominant organic P form. The available data based on agricultural crops show a much higher ratio of PMEs : PDEs, with PDE activities in some cases an order of magnitude lower than PME activities (Asmar & Gissel-Nielsen, 1997).
The question then arises as to whether plants from environments with high phosphodiester concentration in the soil, such as wetlands, produce a higher ratio of PDE : PME to utilize this source of P. The information on root PDE from wetland systems is virtually nonexistent, although PDE in some phytoplankton and moss species was found to have comparable activity to PME (Yamaguchi et al., 2005; Ellwood et al., 2008).
There is a growing interest in the applications of phosphatases as early indicators of environmental change (Newman et al., 2003; Martinez-Crego et al., 2006; Ellwood et al., 2008). The phosphatase enzymes have been used to characterize functional classes of organic P compounds in soil water (Turner et al., 2002; Büneman, 2008). Conceivably, root enzymes of wetland plants might play a similar role for indication of phosphodiester presence in the sediments. In addition, differences in enzyme production among different macrophytes would add further evidence supporting the resource partitioning theory, which states that coexisting plants use different forms of soil P to minimize competition. At the same time, differential production could help to explain the ecological success of species that produce PDE in P-limited environments, where phosphodiesters form a substantial proportion of plant-available organic P (Turner, 2008).
The main goal of our study was to determine whether a selection of wetland plants produce root PDE. If yes, we aimed to determine whether the patterns of root enzyme production in terms of PMEs, PDEs, and their ratio, are better explained by: plant taxonomy, plant functional groups; plant characteristics in terms of P concentration, N : P ratio, and root type; or site characteristics in terms of proportion of phosphomono- and diesters in the soil organic pool, salinity, and nitrogen (N) concentration. The functional groups were a priori defined based on growth form. We predicted that high root phosphatase activities would be found only in plants from regions limited in P (Rejmánková & Snyder, 2008). Of those, plants growing in sediments with a higher proportion of P diesters were expected to show higher activities of root PDEs.
Materials and Methods
The majority of PME and PDE activities were measured on roots of wetland plants collected in early March 2008 from the wetlands located in the lowlands of northern Belize, in the region of the Yucatan Peninsula. Most of these marshes are limestone-based and all are strongly P-limited (Rejmánkováet al., 1996). Thus, vegetation in this area is suitable to study plant adaptations for P-deficiency. Previous research showed a response of PME to increased P input to these wetland ecosystems (Rejmánková & Snyder, 2008) but did not provide any information on PDE activities. In addition to samples from natural marshes, we also collected samples of Eleocharis and Typha from the long-term nutrient enrichment plots located in a subset of these marshes. The nutrient enrichment experiment is described in Rejmánkováet al. (2008). Briefly: there were 10 × 10 m plots, treatments with additions of P, N, N&P and a control. All plots were established in 15 marshes in 2001. The nutrients were applied in 2001, 2002 and 2005.
To provide a comparison with plants from a region that is not limited by P, we included a six-species data set from the Cosumnes River reserve located in the California Central Valley, USA (Hammersmark et al., 2005). The reserve includes both natural and restored wetlands, and it is surrounded by agricultural lands. The agriculture and river flood water contribute to relatively high nutrient input to the reserve (Rejmánková & Tekel, 2000).
Plant selection and root material
Plant species were selected to represent the diversity of growth forms and taxonomic groups present. Root samples were collected from permanently flooded marshes, with a minimum of three different locations for each species. Exceptions were Pontederia cordata in Belize and Echinochloa crus-gali in California data, which were collected from two locations only. Three plants with at least a part of their root system intact and soil attached were collected at each location and kept in a cool environment for a maximum of 24 h. For each sample, three 0.075–0.15 g FW subsamples of subapical roots of diameter < 1 mm and with root hairs were washed in distilled water, towel-dried, placed in 4 ml of buffer, and assayed for the enzyme activity. After the enzyme assay, the individual root segments were again towel-dried, and their length, surface area, and diameter were measured using WIN/Mac RHIZO interactive image capture and analysis software (Regent Instruments Inc., Ottawa, Canada). The root segments were then dried at 70°C and weighed. Root phosphatase activity was expressed on both a dry weight (DW) and surface area basis, because of large differences in the specific root area (Rejmánková & Macek, 2008).
Dry root tissue was ground and assayed for total N with a Perkin Elmer HCN analyzer, San Jose, CA, USA. Total P was measured spectrophotometrically using ascorbic acid reduction of phosphomolybdate complex after combustion and subsequent acid digestion of separate dry tissue subsamples (McNamara & Hill, 2000).
The enzyme activities were determined using para-nitrophenyl phosphate (pNPP; Fluka 71768, Buchs SG, Switzerland) and bis para-nitrophenyl phosphate (bis-pNPP; Aldrich 123943; Sigma-Aldrich Corp., St. Louis, MO, USA) as substrates for root PME and PDE, respectively. In the presence of the respective enzymes, the substrates pNPP and bis-pNPP are hydrolyzed into a colorimetrically analyzable product para-nitrophenol (pNP). Each sample was measured as three subsamples. Following the 5 min preincubation period in 4.5 ml of pH 8 Modified Universal Buffer (MUB; 0.1 M), 0.5 ml of the corresponding substrate was added, the samples were mixed and then incubated for 30 min at 30°C. The final substrate concentrations, 10 mM for pNPP and 5 mM for bis-pNPP, were selected after the preliminary determination of saturation concentration for each substrate. In all cases we used substrate concentrations more than twice the Michaelis constant, Km, determined as 2.74 and 2.03 mM for pNPP and bis-pNPP, respectively. The reaction was terminated by adding 0.5 ml of a solution composed of 1.1 M NaOH (100 mM final), 27.5 mM EDTA (2.5 mM final) and 0.55 M K2HPO4 (50 mM final) (Turner et al., 2001), and the absorbance was read at 410 nm.
Characterization of soil P
To assess the amount of hydrolyzable phosphomonoesters and diesters in the soil, we added commercially available PME and PDE enzymes to filtered soil-water slurry and compared values of dissolved reactive phosphorus (DRP) initially and after addition. We followed the method described by Turner et al. (2002). Briefly: 10 g wet weight soil was placed in a 50 ml centrifuge tube with 25 ml distilled water with 0.25 ml of 0.1 M sodium azide to prevent microbial interference and shaken for 1 h at high speed. Following centrifugation, samples were filtered through 0.45 μm membrane filters. The filtrate was analyzed for DRP and, after a persulfate digestion (Turner et al., 2002), for total P. Dissolved unreactive P (DUP, in ml) was calculated as DUP = TP – DRP. PME from Escherichia coli (Sigma P-4254) and PDE from Crotalus atrox venom (Sigma P-4506) dissolved in 0.1 M Tris (pH 8) with 2 mM MgCl2 were used for the enzyme assays. One set of two replicates for each sample had PME added and one set had a combination of PME and PDE added. The samples were incubated at 37°C for 16 h. All P analyses were performed on the Lachat FIA system (Hach Company, Loveland, CO, USA) spectrophotometrically using ammonium molybdate.
Functional groups definition
The functional groups were delineated based on Sculthorpe’s (1967) classification of aquatic macrophytes into growth forms, and Boutin & Keddy’s (1993) functional group classification of wetland plants modified for our wetland types. The following groups were distinguished: sedges (Cyperaceae), cattails (Typhaceae), grasses (Poaceae), broadleaved emergents (Alismataceae and Pontederiaceae), floating-leaved plants (Nymphaceae), small shrubs (Fabaceae and Solanaceae) and mangroves (Combretaceae and Rhizophoraceae).
Bacterial cells from the rhizoplane as well as the inner surfaces of the root were counted in order to determine if there was any relationship between bacterial abundance and enzyme production. For the surface-attached bacteria, vials with roots and aliquots of Milli-Q water were sonicated for 10 min in a water bath. The liquid with detached bacteria was then transferred into a different vial, while the root was washed with Milli-Q water and sonicated for an additional 1 min. The procedure was repeated twice and suspensions of detached bacteria pooled.
In order to retrieve bacteria from the inner surfaces, one half of each of the above root samples was used to estimate the FW : DW ratio; the other was hand-homogenized with sterile sand, using a mortar and pestle. Ten milliliters of Milli-Q water was added to the homogenate, which was subsequently sonicated in a water bath for 5 min, centrifuged and decanted. Two successive washes of the pellet by 0.5 ml of Milli-Q were performed.
Sample aliquots were stained with 4′,6-diamidino-2-phenylindole (DAPI) and filtered onto 0.2 mm membrane filters. Bacterial cells were counted under the epifluorescence microscope (magnification ×1000) using a counting grid. Results were normalized to root area.
Statistical analyses were performed using Statview (SAS Institute Inc, 1998, Cary, NC, USA). Redundancy analysis (RDA) was performed in the CANOCO package (ter Braak & Šmilauer, 2002). Redundancy analysis enables one to relate and visualize the relationship between two multivariate data sets, and can be considered as an extension of multivariate regression for a multivariate response variable (see Lepš & Šmilauer, 2003). To evaluate the effect of soil factors on the complex of plant attributes factors, we used the Monte Carlo test with 499 random permutations. The effects of soil variables on root enzyme activities were evaluated using stepwise multiple regressions. The RDA, stepwise regression, and various correlations were calculated using the species means.
Root surface enzymes
In all species, both root PMEs and PDEs were always present (Table 1), and their activities were closely correlated (R2 = 0.87; P < 0.0001). PME and PDE activities ranged from 19 to 189 μmol cm−2 h−1 and 24 to 301 μmol cm−2 h−1, respectively. Cladium jamaicense, Pontederia cordata, and Cyperus articulatus showed the highest activities, while Typha domingensis, Nymphaea ampla, and Rhizophora mangle had the lowest. The monocots displayed on average c. 30% higher activities for both PDEs and PMEs compared with dicots. To evaluate the differences among the functional groups, we included only species from Belize natural wetlands. Nearly all species from the California Cosumnes wetland complex, regardless of functional or taxonomic type, showed much lower enzyme activities, and so did the plants from Belize P-enriched plots (Table 1). Functional groups sedges, broadleaved emergents, and small shrubs had the highest activity of both enzymes, while enzyme activities of floating-leaved plants, mangroves, and grasses were low (Fig. 1).
Table 1. Root and shoot nitrogen (N) and phosphorus (P) concentration, activities of root phosphomonoesterase (PME) and phosphodiesterase (PDE), and the specific root area (SRA)
PDE : PME
SRA (cm2 g−1 DW)
(μmol cm−2 h−1)
n, number of location; standard deviation shown in parentheses.
E. cellulosa (+P)
T. domingensis (+P)
California – wetland
The rhizoplane and inner surface bacteria were positively correlated (R2 = 0.3; P = 0.002). Neither PMEs nor PDEs showed any significant correlation with either rhizoplane or inner surface bacteria (data not shown).
Relationship with plant and sediment attributes
To relate and visualize the relationship among enzyme activities with corresponding root attributes, soil attributes, and individual plant species, we conducted an RDA. This revealed a close association between root enzymes (both PMEs and PDEs) and the proportion of soil P hydrolyzable by PMEs and PDEs (Fig. 2). The majority of species from the Cyperaceae family were associated with the same ordination space. All the species from Cosumnes wetland complex, regardless of their taxonomy, were grouped in the right side of the ordination diagram, all associated with high root P content and a high proportion of unreactive soil P (DUP). The Monte Carlo test with forward selection of variables indicated significant conditional effects of sediment DUP (P = 0.004), and sediment extractable N (P = 0.002) on plant attributes, specifically root phosphatases.
To learn more about the relationship between root enzyme activities and plant and soil characteristics, we conducted the stepwise regression using the same variables as in the RDA ordination. For root PME as the dependent variable, the stepwise regression kept root N : P and proportion of diesterase-hydrolyzable P (P-di) as positively correlated, and soil salinity and soil DRP as negatively correlated with PME activity. Similarly, root PDE was positively correlated with root N : P and proportion of P-di, but contrary to PME, the stepwise regression eliminated soil salinity and DRP while including a negative correlation with soil N (Table 2). While root PME was positively correlated with both monoesterase- and diesterase-hydrolyzable P, it was more closely correlated with the diester pool (PME = 34.72 + 5.63 × P-di; R2 = 0.553).
Table 2. Results of stepwise regression analysis of dependence of root phosphomonoesterase and phosphodiesterase on plant and soil variables (% DUP, proportion of dissolved unreactive P hydrolyzed by phosphodiesterase; DRP, dissolved reactive P)
Standardized partial regression coefficient
Root N : P
Root N : P
There is an asymptotic relationship between activities of both root enzymes and root tissue P. After the threshold of c. 1 mg P g−1 DW, the activities stay low and stable (Fig. 3a). This is contrary to root N : P which has a positive relationship with both enzymes (PME: R2 = 0.48, P = 0.0005; PDE: R2 =0.56, P = 0.0001) (Fig. 3b). Both root enzymes are positively correlated with the proportion of both soil monoesterase- and diesterase-hydrolyzable DUP (Fig. 4), but the relationship is stronger for diesterase-hydrolyzable DUP.
Do wetland plants produce root PDE?
We found PDE activity in all tested species and usually in amounts closely comparable to that of PME. Because neither PMEs nor PDEs showed any significant correlation with either rhizoplane or inner surface bacteria, we assumed that roots, rather than root-associated bacteria were the main source of phosphatase activities measured. The PDE : PME ratio ranged from 0.7 to 2.0, that is, PDE activity was almost equal to, or higher than, PME activity. This finding contrasts strongly with the limited published data on root diesterases of other angiosperms. Activities of root diesterases of Hordeum vulgare, Secale cereale and Triticum aestivum were c. five- to 10-fold lower than those of PMEs (Asmar & Gissel-Nielsen, 1997; George et al., 2008). PDE c. 10-fold lower than PME was also reported from the rhizosphere of Lolium perenne and Pinus radiata (Chen et al., 2008). More information is available on PDE activity in nonvascular phototrophs, specifically cyanobacteria, eukaryotic algae, and mosses (Whitton et al., 1991, 2005; Yamaguchi et al., 2005; Singh et al., 2006; Ellwood et al., 2008). Generally, PDE activity in these organisms has been found to be almost equally widespread and of similar magnitude to PME activity. This discrepancy could be explained by the conditions in which the organisms occur. Cereal crops usually grow in soils where monoesters prevail, while more diesters are present in environments where mosses and cyanobacteria occur.
Are there any consistent patterns in enzyme production among wetland macrophytes related to taxonomic or functional groups?
In the natural marshes, both enzymes showed higher activities in representatives from the sedge family. High phosphatase activity is not surprising, because, while sedges are typical dominants in wetland environments regardless of their trophic status, there is a subset of Cyperaceae including species such as Cladium jamaicense, Eleocharis cellulosa, Lepidosperma spp., Costularia xyridoides, Schoenus nigricans, Caustis blakey, etc., that are found in marshes extremely limited by P (Bakker et al., 2005, Rejmánková, 2005; Playsted et al., 2006). Higher phosphatase activites in Cladium and Eleocharis (Cyperaceae) than in Typha (Typhaceae) were reported previously (Kuhn et al., 2002; Rejmánková & Macek, 2008). Cattails, grasses, floating-leaved macrophytes, and mangroves consistently showed low activities of both enzymes (Fig. 5). (Note: while the figure only plots PME, the same applies to PDE because the activities of the two enzymes are closely correlated as explained in the first paragraph of the Results section). It seems that the differences are given by the functional role and ecophysiology of the plants: some plant groups that have life strategies requiring a higher tissue P, that is, strong competitors characterized by a rapid growth such as Typha, have generally lower phosphatase activities. We measured root surface phosphatases, but intracellular enzymes in these plants can also be quite active in recycling P. Increased available P in the environment leads to a corresponding increase in root P and lowering of the enzyme activity. The change in enzyme activity is more pronounced in Eleocharis than in Typha, among others, because root phosphatase activity in Typha is never very high, even at low P availability (Fig. 5).
There was no clear pattern in the ratio of PDE : PME among the taxonomic or functional groups. The ratio fluctuated around the mean of 1.4 (Table 1). It does not seem that different wetland macrophytes differ in use of mono- or diesters as their P source, and our data do not support a resource partitioning hypothesis for soil P (Turner, 2008).
Factors explaining differences in phosphatase activities
Both plant and sediment traits proved to be important when explaining differences in activities of both enzymes. Phosphorus concentration of plant tissue is known to be an important predictor of phosphatase activities (Turner et al., 2003; Martinez-Crego et al., 2006; Rejmánková & Macek, 2008). Here, we found the N : P ratio to be the best predictor of both enzyme activities (Table 2). Similar results for PME in grass species were reported by Olde Venterink & Güsewell (2010), and for mosses by Christmas & Whitton (1998) and Ellwood et al. (2008). Generally, activities of both PME and PDE increased when the N : P ratio rose above 10. Accordingly, PME in the fungal component of a Cladonia lichen increased with increasing N : P ratio of a thallus (Hogan et al., 2010). In systems not limited by N, such as the Belizean wetlands, the main determinant of N : P ratio is the tissue P concentration, which is dependent on the availability of P in the sediments. Phosphatase production can proceed with increasing N : P ratios as long as the N content is sufficient for enzyme synthesis (Wang et al., 2007).
Questions that we were not able to answer
Do these plants hydrolyze phospho-, mono-, and diesters from the soil organic-P pool, or rather do they salvage any organic P that the plant itself is releasing either through root leakage, or from dead cells? Eleocharis and Typha release large amounts of dissolved organic carbon (DOC) in the form of root exudates (H. Šantrůčková, unpublished. However, there is little known about the possible leakage of organic P from roots and the ecological significance of this phenomenon. Especially under conditions of P deficiency, the leakage of P-esters can have a detrimental effect on the plant P economy. The increase in production of extracellular phosphatases and the subsequent increase in exudate recycling would reduce these costs for the plant.
Correlations between the PME and PDE activities and the proportion of diesterase-hydrolyzable DUP in our study were not very strong, indicating that the enzymes are probably capable of using both organic P sources, that is, the soil organic P and organic P leaked from roots.
It is interesting that PME activity was more positively correlated with the diester than with the monoester pool in the soil. While it is commonly assumed that PME and PDE are two distinct enzymes, it is becoming more apparent that alkaline phosphatases produced by various organisms are capable of catalyzing both reactions on the same molecule, and that such promiscuity is an inherent property of many enzymes (Babtie et al., 2010). This catalytic versatility has been suggested to play a role in enabling organisms to survive changes in the ambient environment (Kim & Copley, 2007), and this may help to explain the above-mentioned relationship between PME and DUP.
We are aware of the limitations of studying a system with such a complex chemistry. Most notably, there is a limitation by scale. We are comparing soil characteristics of more-or-less bulk soil, while there are most certainly steep gradients between the root surface and the rhizosphere (Hinsinger et al., 2009; Richardson et al., 2009). This issue of scale may be one of the reasons why we did not find a sufficiently tight relationship between the root enzyme activities and the ambient P regime to use root enzymes as reliable predictors of sediment phospho-, mono-, and diester concentrations. We have, however, confirmed that plants in P-limited wetland environments display PDE activities of the same magnitude as PME. We believe that this indicates the ecological importance of plant PDE production in wetlands and that this topic warrants further investigation.
We thank to Jenise Snyder for help with root scanning, Jaroslava Kubešová for help with field sampling and the three anonymous reviewers for their useful comments. Our research was supported by NSF grant no. 0089211.