ARTICLE IN PRESS

Soil Biology & Biochemistry 39 (2007) 445–453 www.elsevier.com/locate/soilbio

Losses of glomalin-related soil protein under prolonged arable cropping: A chronosequence study in sandy soils of the South African Highveld Anne C. Pregera,, Matthias C. Rilligb, Annika R. Johnsb, Chris C. Du Preezc, Ingo Lobed, Wulf Amelunga a

Institute of Crop Science and Resource Conservation–Soil Science and Soil Ecology, University of Bonn, Nussallee 13, 53115 Bonn, Germany b Microbial Ecology Program, Division of Biological Sciences, The University of Montana, Missoula, MT 59812, USA c Department of Soil, Crop and Climate Science, University of the Free State, P.O. Box 339, Bloemfontein 9300, Republic of South Africa d Department River Ecology, UFZ Centre for Environmental Research Leipzig-Halle, Bru¨ckstr. 3a, 39114 Magdeburg Received 14 March 2006; received in revised form 5 August 2006; accepted 12 August 2006 Available online 18 September 2006

Abstract Residues of arbuscular mycorrhizal fungi (AMF) may be important for agroecosystem functioning due to their ability to promote soil aggregation, especially in coarse textured soils with little biomass input and low capacity to conserve soil organic matter (SOM). Our aim was to assess the fate of AMF residues with prolonged arable cropping in coarse textured soils in a subtropical savannah assuming that glomalin-related soil protein (GRSP), especially the MAb32B11-immunoreactive fraction, mainly constitutes material of AMF origin. In three agroecosystems on the South African Highveld, surface soils were sampled. The former grassland soils had a history of up to 98 yr of cropping. We measured four GRSP fractions: Bradford-reactive soil protein (BRSP) and immunoreactive soil protein (IRSP), and easily extractable fractions of both. The primary grassland sites exhibited generally low contents of SOM and low GRSP contents. Prolonged arable land use of former grassland soils reduced the content of GRSP further. The decline could be described with a monoexponential function with rate constants ranging from 0.04 to 0.41 yr1. Depending on the GRSP fraction, steady-state conditions were reached after 11–92 yr on a level of 39–69% of the initial contents. We conclude that even though GRSP fractions had the same hypothesized origin, they comprised pools with different stability or replacement rate. Easily extractable IRSP was lost most rapidly. In contrast to carbon, nitrogen and microbial residue dynamics, GRSP contents were not reduced below a certain steady-state level, despite potentially negative management effects on AMF, such as tillage, inclusion of fallows into crop rotation and fertilization with inorganic phosphorus. The steady-state GRSP contents coincided with low, but steady agroecosystem yields under the given cropping management. r 2006 Elsevier Ltd. All rights reserved. Keywords: Soil organic matter; Glomalin; Arbuscular mycorrhizal fungi; Biomarker; Arable land use

1. Introduction Drylands cover about 41% of Earth’s land surface, a quarter of these areas being used as cropland (Millennium Ecosystem Assessment, 2005). Sustainable cropping in these regions largely depends on soil organic matter (SOM) for the maintenance of soil fertility (Olsson and Ardo¨, 2002; Zingore et al., 2005). Coarse textured soils, however, rapidly lose SOM upon conversion of grassland Corresponding author. Tel.: +49 228 73 9369; fax: +49 228 73 2782.

E-mail address: [email protected] (A.C. Preger). 0038-0717/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.soilbio.2006.08.014

or woodland to prolonged arable cropping, as can be seen on the South African Highveld (e.g. Lobe et al., 2001; Zingore et al., 2005). In particular, disruption of soil aggregates promote mineralization of formerly physically protected SOM and exacerbated wind erosion (Davidson and Ackerman, 1993; Feller and Beare, 1997; Lobe et al., 2001). Hence, losses of SOM coincide with losses of aggregate binding agents, such as residues of bacteria and fungi (Amelung et al., 2002). The latter work indicated that fungal-derived glucosamine was less rapidly lost upon prolonged cropping on the South African Highveld than bacteria-derived amino sugars. This suggests that soil fungi

ARTICLE IN PRESS 446

A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453

and their residues may play a key role in agroecosystem stability (Bethlenfalvay and Schu¨epp, 1994). Yet, analyses of glucosamine fail to differentiate between saprotrophs and biotrophic mycorrhizal fungi. There is increasing evidence that especially arbuscular mycorrhizal fungi (AMF) are crucial for the regulation of SOM contents in grasslands and arable lands under semiarid and arid climates (Miller and Jastrow, 1992; Varma, 1998). AMF contribute to SOM directly by producing extraradical hyphae (Miller and Kling, 2000; Johnson et al., 2002), and indirectly by enhancing plant growth (Varma, 1998). Extraradical hyphae support physical protection of SOM by promoting soil aggregation (e.g. Tisdall and Oades, 1982; Jastrow et al., 1998; Rillig and Mummey, 2006), which is especially important in sandy and loamy soils (Miller and Jastrow, 2000) where other binding agents are scarce. There, hyphal contribution to physical protection should gain a higher relative importance in comparison to soils with higher clay contents. In general, AMF are reduced by cropping (Miller and Lodge, 1997; Rillig, 2004), mainly due to mechanical disturbances by tillage (Wright et al., 1999; Wright and Anderson, 2000; Kabir, 2005) and due to changes of host plants in crop rotation systems (Johnson and Pfleger, 1992; Wright and Anderson, 2000). Such effects are pronounced if these systems include fallow periods as commonly established in semiarid areas for water conservation during dry seasons (Allen, 1989; Kabir, 2005). Host plants can profit from enhanced phosphorus (P) provision by AMF and allocate carbon (C) to the fungal partner in return; hence, the availability of P to plant roots is another critical factor for the C supply of AMF. Thus, P fertilization can lead to a reduction of AMF abundance or arbuscular mycorrhiza (AM) formation (e.g. Abbott et al., 1984; Eom et al., 1999; Ma¨der et al., 2000), and monitoring inorganic P (Pi) contents might be useful in elucidating AM performance. Studies on the impact of arable land use on AMF in semiarid regions, however, are scarce (Johnson et al., 2004) and have been restricted to the restoration of abandoned fields (Allen, 1989; Miller and Jastrow, 1992; Roldan et al., 1997). We are not aware of long term or chronosequence studies on the effect of arable cropping on AMF in subtropical, semiarid climate. This lack of knowledge makes it difficult to integrate AMF into sustainable management strategies in these areas. In order to understand long lasting effects of AM on soil properties and SOM conservation, a parameter is needed that allows an integration of AM development over several years. Extraradical hyphal lengths are the most direct parameter of AMF quantification in soils (Leake et al., 2004). However, hyphal longevity is relatively short, with calculated turnover times of 5–6 days (Staddon et al., 2003), or at the most several weeks (Steinberg and Rillig, 2003; Olsson and Johnson, 2005). Due to this rapid turnover, hyphal length integrates AMF activity only over a short period compared to some mycelium products

(Steinberg and Rillig, 2003). To follow AMF development over longer time scales, the determination of glomalin as stable, specific biomarker has been suggested (Wright and Upadhyaya, 1996; Johnson et al., 2004). Glomalin is a putative gene product of AMF. The glomalin gene has recently been completely sequenced from the AMF Glomus intraracides and shows homology to stress-related proteins (Gadkar and Rillig, 2006). The operationally defined fractions of glomalin obtained from soils are termed glomalin-related soil protein (GRSP; Rillig, 2004). This nomenclature reflects the fact that other heat-stable proteins of non-AMF origin may be contained in the extracts (Rosier et al., 2006). Thus defined, the contents of GRSP respond to agricultural land use changes (Rillig et al., 2003). The proteinaceous substance has a turnover time of many years (Rillig et al., 2001), and is highly correlated with the percentage of water-stable aggregates (e.g. Wright and Upadhyaya, 1998; Rillig et al., 2002; Rillig, 2004). Following the hypothesis that GRSP are largely of AMF origin (the evidence for this is reviewed in Rillig (2004), the constraints of the methodology shown in Rosier et al. (2006)), we use different GRSP fractions to elucidate their response to conversion of grassland into long-term cropping systems. Based on recent results (Rosier et al., 2006), emphasis is placed on the immunoreactive (IR) fractions since confidence is currently highest that these represent material of largely AMF origin. Samples were taken from our previous chronosequence studies on the South African Highveld (e.g. Lobe et al., 2001; Amelung et al., 2002; Lobe et al., 2005). 2. Materials and methods 2.1. Samples Composite samples (0–20 cm) were taken from eight arable Plinthustalf (Soil Survey Staff, 1998) sites with different cropping durations (2–98 yr) and one adjacent permanent native grassland site in each of three agroecosystems close to Harrismith, Kroonstad and Tweespruit in the Free State Province of South Africa in June/July 1998. The altitude of the sites ranged between 1350 and 1800 m above sea level. All three agroecosystems lie in a summerrainfall region and are comparable concerning soil, slope and climatic conditions (Table 1) (Lobe et al., 2001). Major crops were wheat (Triticum aestivum), maize (Zea mays), and occasionally sunflower (Helianthus sp.) according to Lobe et al. (2005). The sites were ploughed regularly to a depth of about 20 cm (except for the 30- and 68-yrold fields in Harrismith with a plowing depth of 40 cm), and inorganic fertilizer was applied regularly according to local recommendations (maize 50–70 kg N, 10–25 kg P, 0–10 kg K ha1 yr1; wheat 20–50 kg N, 10–20 kg P, 0–15 kg K ha1 yr1; sunflower: 20–50 kg N, 10–20 kg P, 2–6 kg K ha1 yr1). The crop rotation cycle included fallow periods of up to 6 months every 1–2 yr. The

ARTICLE IN PRESS A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453

447

Table 1 Climatic data and soil properties of the three agroecosystems at Harrismith, Kroonstad, and Tweespruit (data from Lobe et al. (2001) and Stallforth (unpublished)) Area

Clay (%)

MDC (yr)

MAT (1C)

MAP (mm)

pH (H2O)

pH (KCl)

CECpot (mmolc kg1)

Pi (mg kg1)

Harrismith Kroonstad Tweespruit

13–19 10–15 10–16

90 98 90

13.8 16.6 16.0

625 563 516

4.56–5.72 5.24–6.82 5.36–6.30

3.76–4.61 4.02–5.53 4.21–5.15

63–130 42–84 49–120

63.5–138.9 33.9–101.8 92.3–194.8

MDC ¼ maximum duration of cultivation, MAT ¼ mean annual temperature, MAP ¼ mean annual precipitation, CECpot ¼ potential cation exchange capacity, Pi ¼ inorganic phosphorus.

grassland sites were used for grazing by either cattle or sheep between 1 and 3 months per year and had a stocking density of 0.4 large stock units per hectare at Harrismith, 0.5 at Kroonstad and 0.6 at Tweespruit. Further details were described by Lobe et al. (2001) and Amelung et al. (2002). Pooled (n ¼ 5 per site) samples of fine earth were stored under air-dried conditions. The land use change from primary grassland to cropland led to a minor compaction of the soil. Correction for bulk density differences resulted in reductions of the grassland element and biomarker contents of 0.7–1.3% (Lobe et al., 2001; Amelung et al., 2002; Brodowski et al., 2004). 2.2. Biochemical analysis Bulk soil C and N contents were reported by Lobe et al. (2001); glucosamine contents were taken from Amelung et al. (2002). Total P and Pi contents of the bulk soil were determined in a sequential extraction (Hedley et al., 1982, slightly modified) by summation of total P and Pi of each fraction (Stallforth, unpublished data). Prior to GRSP analysis, soil samples were sieved o0.25 mm. Four different GRSP pools were distinguished depending on the extraction conditions and the chosen quantification method (Wright and Upadhyaya, 1998; Rillig, 2004). Autoclaving for 30 min at pH 7.0 with 20 mM citric acid yielded easily extractable (i.e. prefix EE-) fractions. For total fractions (i.e. no prefix), the same samples were subjected to sequential cycles of autoclaving for 60 min each, with 50 mM citric acid at pH 8.0. After each cycle, the supernatant was removed and replaced with new extractant. This procedure was followed until the extract was only slightly yellow, as described previously (Wright and Upadhyaya, 1998). The protein contents in the extracts were determined with a Bradford assay (easily extractable Bradford-reactive soil protein (EE-BRSP) and Bradfordreactive soil protein (BRSP); Rillig, 2004) and immunoreactivity with an enzyme-linked immuno-sorbent assay (ELISA) using the monoclonal antibody MAb32B11 (Wright and Upadhyaya, 1996) (easily extractable immunoreactive (MAb32B11) soil protein (EE-IRSP) and immunoreactive soil protein (IRSP); Rillig, 2004). Assays for EE-BRSP, BRSP and EE-IRSP were performed in duplicate. IRSP contents were determined once for each agroecosystem.

2.3. Models The dynamics of the four GRSP fractions were described by a mono-exponential decline model (Lobe et al., 2001; Zingore et al., 2005). This decline function reaches a limit value for the GRSP content under cropping, below which the content does not drop. Ecologically, this limit value signifies a dynamic equilibrium or steady state where GRSP build up balances its degradation: Gt ¼ Ge þ ðG 0  G e ÞexpðktÞ,

(1)

where G is the absolute content of a certain GRSP fraction (g kg1 soil) at a particular time, t (y), of cultivation (Gt), at steady state under cultivation (Ge), under primary grassland (G0); k is a rate constant (yr1, kX0). The point of time when steady state is reached was termed te and approximated as the time when the given content differed less than 1% (of the limit value) from steady state. Apart from the mono-exponential decline function, a bi-exponential model was tested for the description of GRSP dynamics, as C and N decline under cropping had been found to described best bi-exponentially (Lobe et al., 2001). The ratio of the GRSP fractions to other soil parameters was used to compare both the degree and the rate of decline of the contents given above and below the fraction bar. This permits examination of the time course of enrichment or depletion of GRSP relative to other soil parameters. The ratio of the Bradford fractions to C, N and glucosamine, respectively, showed dynamics that could be described by a mono-exponential model rising to a dynamic equilibrium or steady-state value: Qt ¼ Q0 þ ðQe  Q0 Þð1  expðkQ tÞÞ,

(2)

where Q signifies the quotient of a certain GRSP fraction to C and glucosamine, respectively, at a particular time, t (yr), of cultivation (Qt), at steady state under cultivation (Qe), under primary grassland (Q0); kQ is a rate constant (yr1, kX0). The point of time at which the steady-state ratio is reached was termed teQ and approximated as the time when the given ratio differed less than 1% (of the limit value) from steady-state ratio. The ratio of GRSP to Pi was described with Eq. (1).

ARTICLE IN PRESS A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453

448

Table 2 Model parameters for the exponential decline of the contents of the glomalin fractions EE-BRSP, BRSP, EE-IRSP and IRSP according to Eq. (1)

EE-BRSP BRSP EE-IRSP IRSP

k (yr1)

G0 (g kg1 soil)

Ge (g kg1 soil)

Ge/G0

te (yr)

R2

0.04 0.07 0.41 0.08

0.96 1.75 0.46 0.79

0.66 1.17 0.26 0.31

0.69 0.67 0.57 0.39

92.4 53.4 10.5 64.0

0.85** 0.91*** 0.80** 0.81**

(0.02) (0.02) (0.24) (0.04)

(0.04) (0.06) (0.04) (0.08)

(0.05) (0.05) (0.02) (0.06)

Standard errors created by parameter fitting of the means are given in parentheses. EE-BRSP (Easily extractable Bradford-reactive soil protein), BRSP (Bradford-reactive soil protein), EE-IRSP (Easily extractable immunoreactive (MAb32B11) soil protein), IRSP (Immunoreactive (MAb32B11) soil protein). k ¼ rate constant, G0 ¼ content under primary grassland, Ge ¼ content at steady state under cultivation, te ¼ time of steady state (see 2.3 Models);  Po0:01;  Po0:001.

2.4. Statistical analysis

2.5 EE-BRSP (R2 = 0.85) BRSP (R2 = 0.91)

2.0 GRSP (g kg-1 soil)

All models were fitted to the arithmetic means calculated for the three agroecosystems using Sigma Plot 9.01 for Windows (Systat Software Inc., Richmond, USA; determination of initial parameters, 200 iterations, step size 1, and a tolerance of 1 e10, Marquardt-Levenberg algorithm). The mono-exponential GRSP decline model applied also to individual agroecosystems, except for the Kroonstad agroecosystem, where the fit was not significant due to high data variation. The parameters te, teQ, Ge/G0 and Qe/Q0 were estimated from model fits. Standard errors were calculated by parameter fitting. Linear two-sided Pearson correlations were calculated with SPSS 11.0 (SPSS Inc., Chicago, USA).

1.5

1.0

0.5

0.0 0

20

40

40

80

100

Duration of cultivation (yr)

(A) 1.2

EE-IRSP (R2 = 0.80)

3. Results

The native grassland sites on the Highveld contained, on average, 0.98 g EE-BRSP kg1 soil, 1.75 g BRSP kg1 soil, 0.46 g EE-IRSP kg1 soil and 0.75 g IRSP kg1 soil. Converting grassland to arable land resulted in exponential losses of GRSP with rate constants ranging from 0.04 to 0.41 yr1 (Table 2). However, after 11–92 yr steady state was reached for all four GRSP fractions (Fig. 1, Table 2). EE-IRSP levels reached steady state much faster than the other GRSP fractions. Overall, the steady-state contents accounted for 57–69% of initial EE-BRSP, BRSP and EE-IRSP contents. The loss was most pronounced for IRSP, the steady-state contents being only 39% of the initial IRSP content (Table 2). The fits of the mono-exponential function used to describe the dynamics of the four GRSP fractions were highly significant for all GRSP fractions (R2 ¼ 0:8020:91, po0:01; Table 2). The bi-exponential model used by Lobe et al. (2001) for C and N dynamics was over-parameterized; the curve fits were therefore not shown. Among all samples, the contents of EE-BRSP, BRSP and IRSP were highly linearly correlated (R2 ¼ 0:7620:92,

GRSP (g kg-1 soil)

3.1. GRSP contents in native grassland and decline under cropping

1.0

IRSP (R2 = 0.81)

0.8 0.6 0.4 0.2 0.0 0

(B)

20

40

60

80

100

Duration of cultivation (yr)

Fig. 1. Contents of the four GRSP fractions (A: EE-BRSP and BRSP, B: EE-IRSP and IRSP) as affected by cultivation period, shown as the means of the agroecosystems Harrismith, Kroonstad and Tweespruit. Horizontal bars represent the span of cultivation period among the three agroecosystems, vertical bars represent the standard error. The solid line represents the fit of the mono-exponential model according to Eq. (1). See Table 2 for significance levels. GRSP (glomalin-related soil protein), EE-BRSP (Easily extractable Bradford-reactive soil protein), BRSP (Bradford-reactive soil protein), EE-IRSP (Easily extractable immunoreactive (MAb32B11) soil protein), IRSP (Immunoreactive (MAb32B11) soil protein).

po0:05), but there was no significant linear relationship between the two IR-fractions and the two easily extractable fractions.

ARTICLE IN PRESS A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453

3.2. Relation of GRSP to other soil parameters With prolonged cultivation, the contents of C, N and glucosamine have been reported to decline (Lobe et al., 2001; Amelung et al., 2002). All four GRSP fractions showed significant positive linear correlations to C, N and glucosamine as biomarker for total fungal residues (R2 ¼ 0:7920:91, po0:05). Absolute Pi contents and the relative amounts of Pi, i.e. Pi/total P, increased under cropping (not shown). For BRSP and IRSP contents, significant negative correlations with Pi, contents were found (R2 ¼ 0:81, po0:01 and R2 ¼ 0:71, po0:05, respectively). When relating GRSP contents to the relative portion of Pi, all four GRSP fractions were significantly negatively correlated with Pi/total P (R2 ¼ 0:7720:94, po0:05). To better illustrate the differences in reaction times of our chemical parameters, we plotted the ratios of the contents of BRSP, EE-BRSP and EE-IRSP against those of C and glucosamine (Fig. 2) and the ratios of all GRSP

fractions to Pi (Fig. 3). The changes in the ratios between the Bradford fractions and the other soil parameters were described best by mono-exponential relationships (Eq. (2); Fig. 2A and C; Fig. 3; Table 3), whereas the ratios of EEIRSP to C, N and glucosamine could not be described well with Eq. (2); statistical tests yielded non-significant results (p40:05). These ratios rose linearly with R2 ¼ 0:67 (po0:01) for EE-IRSP/C and R2 ¼ 0:54 (po0:05) for EE-IRSP/glucosamine (Fig. 2B and D; EE-IRSP/N (R2 ¼ 0:69, po0:01), not shown). The ratios of IRSP to C, N and glucosamine did not follow a significant linear or exponential trend and are therefore not shown. The ratios of EE-BRSP and BRSP to C, N and glucosamine rose exponentially to a maximum (Qe =Q0 41; Table 3). In contrast, GRSP/Pi ratios of all four fractions declined exponentially (Fig. 3; Qe =Q0 o1; Table 3). For the ratios of BRSP and EE-BRSP to N and glucosamine, steady-state values were reached after 8–18 yr. In case of the ratios BRSP/C and EE-BRSP/C, it took more than 35–49 yr to reach the maximum enrichment. The 0.14

EE-BRSP/C 0.4

BRSP/C

(R2=

(R2

0.81)

EE-IRSP/C (g protein g-1 carbon)

(EE-)BRSP/C (g proteing-1 carbon)

0.5

= 0.69)

0.3

0.2

0.1

0

20

40

60

80

0.10 0.08 0.06 0.04 0.02

100

Duration of cultivation (yr)

(A)

0

20

40

60

80

100

Duration of cultivation (yr)

(B) 1.6

7 EE-BRSP/Glucosamine (R2 = 0.84) 6

1.4

BRSP/Glucosamine (R2 = 0.81) EE-IRSP/Glucosamine

(EE-)BRSP/Glucosamine

EE-IRSP/C = 0.0292 + 0.0005 t (R2= 0.67, p < 0.01)

0.12

0.00

0.0

5 4 3 2 1

1.2 1.0 0.8 0.6 0.4 EE-IRSP/Glucosamine = 0.7697 + 0.0037 t (R2= 0.54, p < 0.05)

0.2 0.0

0 0 (C)

449

20

40

60

Duration of cultivation (yr)

80

100

0 (D)

20

40

60

80

100

Duration of cultivation (yr)

Fig. 2. Ratio of the EE-BRSP and BRSP content to the content of (A) carbon and (C) glucosamine and ratio of EE-IRSP content to the content of (B) carbon and (D) glucosamine as affected by cultivation period, shown as the means of the agroecosystems Harrismith, Kroonstad and Tweespruit. Horizontal bars represent the span of cultivation period among the three agroecosystems, vertical bars the standard error. The solid line represents the fit of the mono-exponential model according to Eq. (2). See Table 3 (A, C) or text (B, D) for significance levels. EE-BRSP (Easily extractable Bradfordreactive soil protein), BRSP (Bradford-reactive soil protein), EE-IRSP (Easily extractable immunoreactive (MAb32B11) soil protein), t (cropping duration given in yr).

ARTICLE IN PRESS A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453

450

(EE-)BRSP/Pi (g protein g-1 Pi)

35 EE-BRSP/Pi (R2= 0.96)

30

BRSP/Pi (R2= 0.93)

25 20 15 10 5 0 0

20

40

60

80

100

Duration of cultivation (yr)

(A)

(EE-)IRSP/Pi (g protein g-1 Pi)

18 16

EE-IRSP/Pi (R2 = 0.90)

14

IRSP/Pi (R2= 0.93)

12 10 8 6 4 2 0 0

(B)

20

40

60

80

100

Duration of cultivation (yr)

Fig. 3. Ratio of the EE-BRSP and BRSP content to the content of inorganic P (A) and ratio of the EE-IRSP and IRSP content to the content of inorganic P (B) as affected by cultivation period, shown as the means of the agroecosystems Harrismith, Kroonstad and Tweespruit. Horizontal bars represent the span of cultivation period among the three agroecosystems, vertical bars the standard error. The solid line represents the fit of the mono-exponential model according to Eq. (1). See Table 3 for significance levels. EE-BRSP (Easily extractable Bradford-reactive soil protein), BRSP (Bradford-reactive soil protein), EE-IRSP (Easily extractable immunoreactive (MAb32B11) soil protein), IRSP (Immunoreactive (MAb32B11) soil protein).

steady-state ratio of all four GRSP fractions to Pi was reached after 21–61 yr. 4. Discussion 4.1. GRSP levels in native grassland and decline under cropping The semiarid climate and the coarse soil texture on the South African Highveld limit net primary production (Beukes et al., 2004). As a result, there is little C allocation to AMF, and GRSP contents were low in the native grassland compared to a range of native and arable soils (Wright and Upadhyaya, 1998; Nichols and Wright, 2004). When looking at the four different fractions of GRSP, it has to be considered that other heat-stable proteins of non-

AMF origin may be contained in the BRSP extracts (Rosier et al., 2006). However, Rosier et al. (2006) showed that the coextraction of non-AMF-protein in the two Bradford fractions is less pronounced in soils with a low carbon content, such as the given sandy soils of the Highveld. There are large uncertainties concerning the C and N contents of BRSP depending on extraction procedure and examined soil. Hence, without knowing the actual C and N contents of the GRSP fractions extracted under this study it is not reasonable to calculate the amount of soil organic C or N bound in GRSP. All crops planted on the South African Highveld are arbuscular mycorrhizal. Therefore, it is reasonable to assume that the exponential decline of GRSP during prolonged cropping (Fig. 1) is a net curve influenced by degradation of grassland-derived GRSP and by production of new GRSP during cropping: When AMF are negatively affected by cropping due to tillage, inclusion of fallow periods into crop rotations and Pi provided by fertilization and SOM mineralization, production rates of GRSP under cropping might be reduced in comparison to primary grassland. At the same time, GRSP decomposition proceeds; it might even be enhanced due to the cropping induced disturbances, as deduced from breakdown of aggregates and losses of old savannah N reserves (Brodowski et al., 2004). As the presented curves are net developments, it is not possible to deduce any information on the turnover time from the curve itself. That the dynamics of the four GRSP fractions were all described best by the same kind of mono-exponential model indicates on the one hand that each GRSP fraction acted as one pool under cultivation; no division into pools of different dynamics was possible with the given data solution. On the other hand, together with the significant correlation between the Bradford and the more specific IR fractions this might indicate that all GRSP fractions had a similar hypothesized main source, namely AMF. At least, it shows that potentially coextracted non-AMF-derived proteins in the two Bradford fractions respond in the same way to prolonged cropping like AMF derived proteins which are better represented by the two IR fractions. Although the dynamics of all four GRSP fractions can be described by one kind of model, there are pronounced differences in decline rates between EE-IRSP and the other three fractions and in the relative steady-state values (Ge/ G0; Table 2) between IRSP and the other fractions. The faster dissipation of EE-IRSP classifies this fraction as a labile GRSP pool, indicating a more rapid turnover, slower replacement or a shift into another operationally defined pool (Nichols and Wright, 2004). The EE-BRSP dissipated least rapidly. Some EE-BRSP might have been re-provided by mobilization of BRSP, or produced in preference to other GRSP fractions by the crop-associated AMF, but the dynamics of this fraction might also be more strongly influenced by coextraction of heat-stable proteins being of

ARTICLE IN PRESS A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453

451

Table 3 Model parameters for the exponential changes of the ratios of the glomalin fractions EE-BRSP, BRSP, EE-IRSP and IRSP to C, N, Gluc (glucosamine) (according to Eq. (2)), and Pi (according to Eq. (1)), respectively

EE-BRSP/C BRSP/C EE-BRSP/N BRSP/N EE-BRSP/Gluc BRSP/Gluc EE-BRSP/Pi BRSP/Pi EE-IRSP/Pi IRSP/Pi

kQ (yr1)

Q0

0.11 0.08 0.22 0.19 0.31 0.43 0.08 0.09 0.23 0.09

0.08 0.15 0.81 1.48 1.60 2.91 15.30 27.15 6.20 10.74

(0.05) (0.06) (0.11) (0.15) (0.15) (0.24) (0.02) (0.03) (0.08) (0.03)

Qe (0.01) (0.03) (0.09) (0.20) (0.20) (0.33) (0.67) (1.64) (0.49) (0.82)

0.16 0.27 1.29 2.15 2.79 4.65 6.34 10.76 2.49 2.55

(0.01) (0.02) (0.04) (0.10) (0.09) (0.13) (0.48) (1.09) (0.23) (0.55)

Qe/Q0

teQ (yr)

R2

2.00 1.82 1.59 1.45 1.74 1.60 0.41 0.40 0.40 0.24

35.2 49.0 16.8 17.9 11.9 8.4 58.7 53.4 21.4 61.3

0.81** 0.69* 0.80** 0.63 n.s. 0.84** 0.81** 0.96*** 0.93*** 0.90** 0.93***

Standard errors created by parameter fitting of the means are given in parentheses. EE-BRSP (Easily extractable Bradford-reactive soil protein), BRSP (Bradford-reactive soil protein), EE-IRSP (Easily extractable immunoreactive (MAb32B11) soil protein), IRSP (Immunoreactive (MAb32B11) soil protein). k ¼ rate constant, Q0 ¼ ratio under primary grassland (g protein g1 C, N, glucosamine protein, Pi respectively), Qe ¼ ratio at steady state under cultivation (g protein g1 C, N, glucosamine protein, Pi respectively), teQ ¼ time of steady-state ratio (see 2.3. Models)  Po0:05;  Po0:01;  Po0:001; n.s. ¼ non significant.

non-AMF origin than the other GRSP fractions (see Rosier et al., 2006). Our findings can be reconciled with the studies on GRSP decomposition (Steinberg and Rillig, 2003). In such studies, the relative decomposition rates increased in the order EEBRSPoBRSPoIRSP. Our rate constant data are grouped in identical order (Table 2). However, in the study of Steinberg and Rillig (2003), EE-IRSP contents even increased in the course of incubation, indicating that another part of GRSP could either be converted into easily extractable pools or that additional binding sites for the antibody became exposed. The rapid dissipation of EEIRSP in the Highveld cropland soils, however, makes a significant conversion of IRSP to EE-IRSP unlikely. 4.2. Relations to other soil parameters As C and N, as well as amino sugars and amino acids and GRSP follow a trend of decline under prolonged cropping, the presented positive correlations between GRSP and the other SOM-related parameters are not unexpected and have been shown for soil organic C by Nichols and Wright (2005) as well. The mono-exponential decline to a steady state of the four GRSP fractions is in contrast to the declines of C and N (Lobe et al., 2001), and microbial residues like amino sugars (Amelung et al., 2002) which all could be described better by a bi-exponential function declining to complete depletion. For those SOM fractions, the bi-exponential model indicated the presence of two pools of distinctively different decay stability (Lobe et al., 2001; Amelung et al., 2002), probably due to different microlocations of SOM in the soil and hence, different protection mechanisms. Even though the different extractability of the determined GRSP fractions promotes the assumption that there might be glomalin pools in soils with different stabilities,

the difference is not pronounced enough to be seen at the given data solution. This may also be due to the fact that the easily extractable fractions in these soils make up a high percentage of the total pools (see Fig. 1). This is probably attributable to the sandy texture of the soils. The ratios of the Bradford fractions to C, N and glucosamine contents can be described best with a function leading to a steady value (Fig. 2A, C). The latter reflects a steady state, where balanced inputs and outputs of the parameters under consideration lead to stable ratios. In contrast, the more AMF-specific EE-IRSP fraction did not develop into a stable ratio with C, N and the fungal biomarker. This is in line with the finding that the dynamics of the three latter parameters can be described best with a bi-exponential model with continued decline even after 90 yr of cropping (Lobe et al., 2001; Amelung et al., 2002), while EE-IRSP contents stabilize on a certain level (Table 2, Fig. 1). This leads to continued enrichment of EE-IRSP under cropping. However, IRSP could not be shown to change its ratio to C, N and glucosamine significantly with time, indicating that its decline is more or less proportional to the decline of the other three parameters. The ratios EE-BRSP/glucosamine and BRSP/glucosamine reached their maximum relative enrichment of GRSP earlier than the ratios related to C and N. This reflected a faster dissipation of glucosamine in the arable soils compared to C and N. Glucosamine is a biomarker for both saprotrophic and biotrophic, i.e. mycorrhizal fungi. Its depletion in comparison to EE-BRSP, BRSP and EEIRSP (Fig. 2C, D) does not necessarily indicate that the fungal community shifted to more AM-fungi. The dynamics of GRSP and glucosamine are to a large extent influenced by their possibly different biochemical recalcitrance, if not by different production rates of GRSP and glucosamine, both being unknown.

ARTICLE IN PRESS 452

A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453

As cropping resulted in increased Pi contents, probably due to SOM mineralization and Pi fertilization, GRSP fractions were depleted in comparison to Pi. The depletion eventually reached a steady level due to more or less balanced in- and outputs for both Pi and the respective GRSP fraction (Fig. 3). Comparisons to aggregate dynamics (unpublished data) revealed that the four GRSP fractions did not seem to be influenced by the cropping induced breakdown of larger macroaggregates in the Highveld soils: EE-IRSP declined much faster than macroaggregates with a diameter 42000 mm, which exhibited a te of 26–27 yr (Lobe, 2003), while the other GRSP fractions declined much more slowly (Table 2). Lobe (2003) found that, while macroaggregates of 2000–2800 and 2800–8000 mm decreased, the amount of mesoaggregates/small macroaggregates of 250–2000 mm, large microaggregates of 53–250 mm, and small microaggregates of 20–53 mm increased during cropping. This, in turn, could be an effect of GRSP or vice versa, an effect of aggregates on GRSP stabilization. After 90 yr of cropping to the Highveld, C and N contents as well as those of microbial residues were still declining without having reached steady state (Lobe et al., 2001; Amelung et al., 2002). This finding is in contrast to the GRSP fractions, which reached a steady content within the period evaluated. The steady state was reached despite the stresses posed on AMF due to agricultural management practices. Crop yields at the sites under study declined with prolonged cropping, but also reached a steady state on a low level of ca. 3t ha1 for maize and 1.6t ha1 for wheat after 34–35 yr (within the time span from the early 20th century until 1998; Lobe et al., 2005). It seems reasonable that steady GRSP contents coincide with steady crop yields, because both yields and AMF biomass are directly related to net primary production of the prevailing crops. Notably, despite prolonged arable land use and concomitant disturbances such as ongoing wind erosion (Lobe et al., 2001), the dry agroecosystems on the South African Highveld have not yet reached the threshold at which they are turned into desertified badlands. In any case, it may be dangerous to stress these agroecosystems further. To restore soil functions, the careful introduction of conservational cropping practices like minimum or mulch tillage may be considered, or even, the reconversion to grassland which can be used for pasture (Du Preez, 2003). Successful examples for such practices and land use changes can already be found on the Highveld. Acknowledgements This work was financially supported by the German Research Foundation and by the US National Science Foundation (M.C.R.). We are grateful to all the South African farmers who allowed us to sample their fields. We thank Deon Maree, Leon Strachan, De Villiers Bosman, Frikkie Swanepoel and Donny Holmes for organizing the

contacts with the farmers, Nico Kroese for providing climatic data, and Ralf Stallforth for providing P data. We are grateful to Sonja Brodowski for comments on the manuscript. We also thank her and Elmarie Kotze for assistance in South Africa during sampling. We thank Sara Wright for providing MAb32B11.

References Abbott, L.K., Robson, A.D., De Boer, G., 1984. The effect of phosphorus on the formation of hyphae in soil by the vesicular-arbuscular mycorrhizal fungus, Glomus fasciculatum. New Phytologist 97, 437–446. Allen, M.F., 1989. Mycorrhizae and rehabilitation of disturbed arid soils: processes and practices. Arid Soil Research 3, 229–241. Amelung, W., Lobe, I., Du Preez, C.C., 2002. Fate of microbial residues in sandy soils of South African Highveld as influenced by prolonged arable cropping. European Journal of Soil Science 53, 29–35. Bethlenfalvay, G.J., Schu¨epp, H., 1994. Arbuscular mycorrhizas and agrosystem stability. In: Gianinazzi, S., Schu¨epp, H. (Eds.), Sustainable agriculture and natural ecosystems. Birkha¨user, Basel, pp. 117–131. Beukes, D.J., Bennie, A.T.P., Hensley, M., 2004. Optimizing soil water balance components for sustainable crop production in dry areas of South Africa. In: Rao, S.C., Ryan, J. (Eds.), Challenges and strategies for dryland agriculture. Crop Science Society of America. CSSA special publ. No. 32, Madison, WI., pp. 291–313. Brodowski, S., Amelung, W., Lobe, I., Du Preez, C.C., 2004. Losses and biogeochemical cycling of soil organic nitrogen with prolonged arable cropping in South African Highveld–evidence from D- and L-amino acids. Biogeochemistry 71, 17–42. Davidson, E.A., Ackerman, I.L., 1993. Changes in soil carbon inventories following cultivation of previously untilled soils. Biogeochemistry 20, 161–193. Du Preez, C.C., 2003. Sustainable land use and soil quality: organic matter as an indicator. South African Journal for Natural Science and Technology 22, 106–112. Eom, A.H., Hartnett, D.C., Wilson, G.W.T., Figge, D.A.H., 1999. The effect of fire, mowing and fertilizer amendment on arbuscular mycorrhizas in tallgrass prairie. American Midland Naturalist 142, 55–70. Feller, C., Beare, M.H., 1997. Physical control of soil organic matter in the tropics. Geoderma 79, 69–116. Gadkar, V., Rillig, M.C., 2006. The arbuscular mycorrhizal fungal protein glomalin is a putative homolog of heat shock protein 60. FEMS Microbiology Letters 263, 93–101. Hedley, M.J., Stewart, J.W.B., Chauhan, B.S., 1982. Changes in inorganic and organic soil phosphorus fractions induced by cultivation practices and by laboratory incubations. Soil Science Society of America Journal 46, 970–976. Jastrow, J.D., Miller, R.M., Lussenhop, J., 1998. Contributions of interacting biological mechanisms to soil aggregate stabilization in restored prairie. Soil Biology & Biochemistry 30, 905–916. Johnson, N.C., Pfleger, F.L., 1992. Vesicular-arbuscular mycorrhizae and cultural stress. In: Bethlenfalvay, G.J., Linderman, R.G. (Eds.), Mycorrhizae in sustainable agriculture. ASA special publ. no. 54. Agron. Soc. Am., Crop Sci. Soc. Am. and Soil Sci. Soc. Am., Madison, WI, pp. 71–99. Johnson, D., Leake, J.R., Ostle, N., Ineson, P., Read, D.J., 2002. In situ 13 CO2 pulse-labelling of upland grassland demonstrates a rapid pathway of carbon flux from arbuscular mycorrhizal mycelia to the soil. New Phytologist 153, 327–334. Johnson, C.K., Wienhold, B.J., Doran, J.W., Drijber, R.A., Wright, S.F., 2004. Linking microbial-scale findings to farm-scale outcomes in a dryland cropping system. Precision Agriculture 5, 311–328.

ARTICLE IN PRESS A.C. Preger et al. / Soil Biology & Biochemistry 39 (2007) 445–453 Kabir, Z., 2005. Tillage or no-tillage: impact on mycorrhizae. Canadian Journal of plant science 85, 23–29. Leake, J., Johnson, D., Donnelly, D., Muckle, G., Boddy, L., Read, D., 2004. Networks of power and influence: the role of mycorrhizal mycelium in controlling plant communities and agroecosystem functioning. Canadian Journal of Botany 82, 1016–1045. Lobe, I., 2003. Fate of organic matter in sandy soils of the South African Highveld as influenced by the duration of arable cropping. PhD Thesis. Bayreuther Bodenkundliche Berichte 79, Bayreuth. Lobe, I., Amelung, W., Du Preez, C.C., 2001. Losses of carbon and nitrogen with prolonged arable cropping from sandy soils of the South African Highveld. European Journal of Soil Science 52, 93–101. Lobe, I., Bol, R., Ludwig, B., Du Preez, C.C., Amelung, W., 2005. Savanna-derived organic matter remaining in arable soils of the South African Highveld long-term mixed cropping: evidence from 13C and 15 N natural abundance. Soil Biology & Biochemistry 37, 1898–1909. Ma¨der, P., Edenhofer, S., Boller, T., Wiemken, A., Niggli, U., 2000. Arbuscular mycorrhizae in a long-term field trial comparing low-input (organic, biological) and high-input (conventional) farming systems in a crop rotation. Biology and Fertility of Soils 31, 150–156. Miller, R.M., Jastrow, J.D., 1992. The role of mycorrhizal fungi in soil conservation. In: Bethlenfalvay, G.J., Linderman, R.G. (Eds.), Mycorrhizae in Sustainable Agriculture. ASA special publ. no. 54. Agron. Soc. Am., Crop Sci. Soc. Am. and Soil Sci. Soc. Am., Madison, WI., pp. 29–44. Miller, R.M., Jastrow, J.D., 2000. Mycorrhizal fungi influence soil structure. In: Kapulnik, Y., Douds, D.D. (Eds.), Arbuscular Mycorrhizas: Physiology and Function. Kluwer, Dordrecht, pp. 3–18. Miller, R.M., Lodge, D.J., 1997. Fungal Responses to disturbance: agriculture and forestry. In: Wicklow, D.T., So¨derstro¨m, B.E. (Eds.), The Mycota IV: Environmental and Microbial Relationships. Springer, Berlin, pp. 65–84. Miller, R.M., Kling, M., 2000. The importance of integration and scale in the arbuscular mycorrhizal symbiosis. Plant and Soil 226, 295–309. Nichols, K.A., Wright, S.F., 2004. Contributions of fungi to soil organic matter in agroecosystems. In: Magdoff, F., Weil, R.R. (Eds.), Soil Organic Matter in Sustainable Agriculture. CRC Press, Boca Raton, pp. 179–198. Nichols, K.A., Wright, S.F., 2005. Comparison of glomalin and humic acid in eight native US soils. Soil Science 170, 985–997. Olsson, L., Ardo¨, J., 2002. Soil carbon sequestration in degraded semiarid agroecosystems—perils and potentials. Ambio 31, 471–477. Olsson, P.A., Johnson, N.C., 2005. Tracking carbon from the atmosphere to the rhizosphere. Ecology Letters 8, 1264–1270. Rillig, M.C., 2004. Arbuscular mycorrhizae, glomalin, and soil aggregation. Canadian Journal of Soil Science 84, 355–363.

453

Rillig, M.C., Mummey, D.L., 2006. Tansley review. Mycorrhizas and soil structure. New Phytologist 171, 41–53. Rillig, M.C., Wright, S.F., Nichols, K.A., Schmidt, W.F., Torn, M., 2001. Large contribution of arbuscular mycorrhizal fungi to soil carbon pools in tropical forest soils. Plant and Soil 233, 167–177. Rillig, M.C., Wright, S.F., Eviner, V., 2002. The role of arbuscular mycorrhizal fungi and glomalin in soil aggregation: comparing effects of five plant species. Plant and Soil 238, 325–333. Rillig, M.C., Ramsey, P.W., Morris, S., Paul, E.A., 2003. Glomalin, an arbuscular-mycorrhizal fungal soil protein, responds to land-use change. Plant and Soil 253, 293–299. Roldan, A., Garcia, C., Albaladejo, J., 1997. AM fungal abundance and activity in a chronosequence of abandoned fields in a semiarid Mediterranean site. Arid Soil Research Rehabilitation 11, 211–220. Rosier, C.L., Hoye, A.T., Rillig, M.C., 2006. Glomalin related soil protein: assessment of current detection and quantification tools. Soil Biology & Biochemistry 38, 2205–2211. Soil Survey Staff, 1998. Keys to Soil Taxonomy, 8th ed. Pocahontas Press, Blacksburg, VA. Staddon, P.L., Ramsey, C.B., Ostle, N., Ineson, P., Fitter, A.H., 2003. Rapid turnover of hyphae of mycorrhizal fungi determined by AMS microanalysis of 14C. Science 300, 1038–1040. Steinberg, P.D., Rillig, M.C., 2003. Differential decomposition of arbuscular mycorrhizal fungal hyphae and glomalin. Soil Biology & Biochemistry 35, 191–194. Tisdall, J.M., Oades, J.M., 1982. Organic matter and water-stable aggregates in soils. Journal of Soil Science 33, 141–164. Varma, A., 1998. Functions and application of arbuscular mycorrhizal fungi in arid and semi-arid soils. In: Varma, A., Hock, B. (Eds.), Mycorrhiza: Structure, Function, Molecular Biology and Biotechnology. Springer, Berlin, pp. 521–556. Wright, S.F., Anderson, R.L., 2000. Aggregate stability and glomalin in alternative crop rotations for the central Great Plains. Biology and Fertility of Soils 31, 249–253. Wright, S.F., Upadhyaya, A., 1996. Extraction of an abundant and unusual protein from soil and comparison with hyphal protein of arbuscular mycorrhizal fungi. Soil Science 161, 575–586. Wright, S.F., Upadhyaya, A., 1998. A survey of soils for aggregate stability and glomalin, a glycoprotein produced by hyphae of arbuscular mycorrhizal fungi. Plant and Soil 198, 97–107. Wright, S.F., Starr, J.L., Paltineanu, I.C., 1999. Changes in aggregate stability and concentration of glomalin during tillage management transition. Soil Science Society of America Journal 63, 1825–1829. Zingore, S., Manyame, C., Nyamugafata, P., Giller, K.E., 2005. Longterm changes in organic matter of woodland soils cleared for arable cropping in Zimbabwe. European Journal of Soil Science 56, 727–736.