Stoichiometry of soil enzyme activity at global scale

Ecology Letters, (2008) 11: 1252–1264 REVIEW AND SYNTHESIS 1 Robert L. Sinsabaugh, * Christian L. Lauber,1 Michael N. Weintraub,2 Bony Ahmed,3 Steven D. Allison,4 Chelsea Crenshaw,1 Alexandra R. Contosta,5 Daniela Cusack,6 Serita Frey,5 Marcy E. Gallo,1 Tracy B. Gartner,7 Sarah E. Hobbie,8 Keri Holland,9 Bonnie L. Keeler,8 Jennifer S. Powers,10 Martina Stursova,1 Cristina Takacs-Vesbach,1 Mark P. Waldrop,11 Matthew D. Wallenstein,12 Donald R. Zak13 and Lydia H. Zeglin1 doi: 10.1111/j.1461-0248.2008.01245.x Stoichiometry of soil enzyme activity at global scale Abstract Extracellular enzymes are the proximate agents of organic matter decomposition and measures of these activities can be used as indicators of microbial nutrient demand. We conducted a global-scale meta-analysis of the seven-most widely measured soil enzyme activities, using data from 40 ecosystems. The activities of b-1,4-glucosidase, cellobiohydrolase, b-1,4-N-acetylglucosaminidase and phosphatase g)1 soil increased with organic matter concentration; leucine aminopeptidase, phenol oxidase and peroxidase activities showed no relationship. All activities were significantly related to soil pH. Specific activities, i.e. activity g)1 soil organic matter, also varied in relation to soil pH for all enzymes. Relationships with mean annual temperature (MAT) and precipitation (MAP) were generally weak. For hydrolases, ratios of specific C, N and P acquisition activities converged on 1 : 1 : 1 but across ecosystems, the ratio of C : P acquisition was inversely related to MAP and MAT while the ratio of C : N acquisition increased with MAP. Oxidative activities were more variable than hydrolytic activities and increased with soil pH. Our analyses indicate that the enzymatic potential for hydrolyzing the labile components of soil organic matter is tied to substrate availability, soil pH and the stoichiometry of microbial nutrient demand. The enzymatic potential for oxidizing the recalcitrant fractions of soil organic material, which is a proximate control on soil organic matter accumulation, is most strongly related to soil pH. These trends provide insight into the biogeochemical processes that create global patterns in ecological stoichiometry and organic matter storage. Keywords C : N : P ratio, cellobiohydrolase, ecological stoichiometry, leucine aminopeptidase, peroxidase, phenol oxidase, phosphatase, soil enzyme activity, soil organic matter, b-1,4-glucosidase, b-1,4-N-acetylglucosaminidase. Ecology Letters (2008) 11: 1252–1264 1 8 que, NM, 87131, USA 2 Department of Environmental Sciences, University of Toledo, Minnesota, 1987 Upper Buford Circle, St Paul, MN 55108, USA 9 Department of Ecology, Evolution and Marine Biology, Uni- Toledo, OH 43606-3390, USA versity of California, Santa Barbara, Santa Barbara, CA 93106, Department of Biology, University of New Mexico, Albuquer- Department of Ecology, Evolution and Behavior, University of 3 USA 85281, USA 10 4 and Soil, Water & Climate, University of Minnesota, 1987 Upper School of Life Sciences, Arizona State University, Tempe, AZ Departments of Ecology and Evolutionary Biology and Earth Departments of Ecology, Evolution & Behavior, Plant Biology System Science, University of California, Irvine, CA 92697, USA Buford Circle, St Paul, MN 55108, USA 5 11 Hampshire, Durham, NH 03824, USA 6 Department of Environmental Science, Policy and Manage- Menlo Park, CA 94025, USA 12 Natural Resource Ecology Laboratory, Colorado State Univer- Department of Natural Resources, University of New United States Geological Survey, 345 Middlefield Rd, MS 962, ment, University of California, Berkeley, Berkeley, CA 94720, USA sity, Campus Delivery 1499, Fort Collins, CO 80523-1499, USA 7 13 gram, Carthage College, 2001 Alford Park Drive, Kenosha, WI Arbor, MI 48109-1115, USA 53140, USA *Correspondence: E-mail: [email protected] Department of Biology and the Environmental Science Pro-  2008 Blackwell Publishing Ltd/CNRS School of Natural Resources, University of Michigan, Ann Review and Synthesis INTRODUCTION Terrestrial soils contain the largest reservoir of organic carbon in the biosphere (c. 1800 Pg). Mineralization of this organic matter by heterotrophic microorganisms affects global carbon and nutrient cycles, plant production and atmospheric composition. The proximate agents of soil organic matter (SOM) decomposition are extracellular enzymes that deconstruct plant and microbial cell walls and reduce macromolecules to soluble substrates for microbial assimilation (Burns 1978; Burns & Dick 2002). In the context of global nutrient cycles, these enzymes catalyse processes that are antipodal to C-fixation by ribulose bisphosphate carboxylase and N-fixation by nitrogenase. Extracellular enzyme activity (EEA) in soils has been studied for more than a century with a goal of understanding the biochemistry of decomposition and nutrient cycling (Skujins 1978). Soil EEA has also been studied in relation to ecosystem responses to global change and other disturbances (e.g. Lipson et al. 2005; Sinsabaugh et al. 2005; Finzi et al. 2006). The most widely assayed enzymes are those involved in the degradation of cellulose and lignin, the most abundant components of plant litter (Allison et al. 2007). Other commonly measured enzymes hydrolyze proteins, chitin and peptidoglycan, which are the principal reservoirs of organic N (Caldwell 2005). Extracellular phosphatases are of interest for their role in mineralizing P from nucleic acids, phospholipids and other ester phosphates (Turner et al. 2002; Toor et al. 2003). The structural heterogeneity of biopolymers requires the interaction of several classes of enzymes to reduce them to constituent monomers available for microbial consumption (Ljungdahl & Eriksson 1985; Kirk & Farrell 1987; Sinsabaugh 2005). However, most studies of soil EEA are limited to enzymes that catalyse the production of the terminal monomers, because the kinetics are easier to study and the reactions produce assimilable products (Allison et al. 2007). Because EEA mediates microbial nutrient acquisition from organic matter, these activities are commonly interpreted as indicators of microbial nutrient demand (Olander & Vitousek 2000; Schimel & Weintraub 2003; Caldwell 2005; Moorhead & Sinsabaugh 2006). This demand is determined by the elemental stoichiometry of microbial biomass in relation to environmental nutrient availability. Stoichiometric constraints on biomass composition are evident for phytoplankton (Redfield 1958), terrestrial plants (McGroddy et al. 2004) and animals (Sterner & Elser 2002) as well as soil microbial biomass (Cleveland & Liptzin 2007). However, within all of these groups, there is variation in biomass stoichiometry among ecosystems that can be related to constraints imposed by local nutrient availability. For example, large-scale variation in the C : N : P ratios of Stoichiometry of soil enzyme activity 1253 plant foliage is consistent with observations that tropical forests are generally more P-limited than high-latitude forests, which tend to be N-limited (McGroddy et al. 2004; Reich & Oleksyn 2004). This pattern arises because high rates of weathering in tropical regions lead to the loss of rock-derived nutrients, such as P, while episodic glaciations in high-latitude regions limit the accumulation of N (Walker & Syers 1976; Vitousek & Howarth 1991). Microbial biomass composition does not follow a latitudinal trend but does vary in relation to ecosystem type (Cleveland & Liptzin 2007). Because EEA links environmental nutrient availability with microbial production, large-scale patterns in EEA may reveal the constraints on microbial biomass stoichiometry and enzyme relationships to SOM composition. Large-scale EEA patterns may also provide insights into the biochemical controls on soil carbon storage. Because EEA catalyses rate-limiting steps in organic matter degradation, correlations between rates of plant litter decomposition, microbial production and EEA are frequently observed (Andersson et al. 2005; Sinsabaugh et al. 2005; Weintraub et al. 2007; Waldrop & Harden 2008). However, the contribution of these relationships to the global distribution of SOM has not been evaluated. Despite thousands of published studies, technological limitations and lack of standardized protocols have precluded a comparative analysis of the magnitude and distribution of soil EEA in relation to global climatic and edaphic gradients. During the past decade, protocols that combine the use of fluorigenic substrates with high throughput microplate technology have come into general use (Sinsabaugh et al. 1997; Marx et al. 2001). As a result, we can now assemble a comparative database of soil EEA potentials for 40 ecosystems. These data reveal unexpected stoichiometric constraints on the functional organization of microbial communities and the dynamics of SOM accumulation. METHODS Data description Soil, excluding surface litter, was collected at each site to depths of 5–20 cm, and assayed for the potential activities of one or more extracellular enzymes. Hydrolytic enzymes were assayed using substrates linked to a methylumbelliferyl fluor; oxidative enzymes were assayed colorimetrically using L-3,4-dihydroxyphenylalanine (Table 1). Activities were calculated in units of nmol h)1 g)1 dry mass and nmol h)1 g)1 SOM. Samples were incubated at 20 ± 2 C, except for Niwot Ridge samples, which were incubated at 15 ± 2 C. To approximate ambient soil pH (Table 2), acidic soils were assayed at pH 5 by suspending  2008 Blackwell Publishing Ltd/CNRS 1254 R. L. Sinsabaugh et al. Review and Synthesis Table 1 Soil enzymes assayed for potential Enzyme EC Abbreviation Substrate b-1,4-glucosidase Cellobiohydrolase b-N-acetylglucosaminidase EC 3.2.1.21 EC 3.2.1.91 EC 3.2.1.14 BG CBH NAG Leucyl aminopeptidase EC 3.4.11.1 LAP Acid (alkaline) phosphatase Phenol oxidase Peroxidase EC 3.1.3.1 EC 1.10.3.2 EC 1.11.1.7 AP POX PER 4-MUB-b-D-glucoside 4-MUB-b-D-cellobioside 4-MUB-N-acetyl-ßD-glucosaminide L-Leucine-7-amido4-methylcoumarin 4-MUB-phosphate L-3,4-dihydroxyphenylalanine L-3,4-dihydroxyphenylalanine and H2O2 activity EC, enzyme commission classification; MUB, methylumbelliferyl. c. 1 g soil in 100 mL of 50 mM sodium acetate buffer; alkaline soils were assayed at pH 8 using 50 mM sodium bicarbonate buffer. The database includes activities for the seven-most widely measured soil enzymes from 40 ecosystems. The number of cases per ecosystem (the number of locations sampled · the number of sampling dates), ranges from 4 to 169, for a total of 1154 cases (Table 3). Metadata for all sites are appended as Supporting information. Enzyme description b-1,4-Glucosidase (BG) and cellobiohydrolase (CBH) are enzymes that contribute to the degradation of cellulose and other beta-1,4 glucans (Ljungdahl & Eriksson 1985). The principal function of BG is hydrolysis of cellobiose to glucose, but many of these enzymes are active against other substrates as well. CBH hydrolyzes cellobiose dimers from the non-reducing ends of cellulose molecules. b-N-acetylglucosaminidase (NAG) plays a role in the degradation of chitin and other b-1,4-linked glucosamine polymers that are analogous to the role of BG in cellulose degradation (Sinsabaugh 2005). Leucine aminopeptidase (LAP) hydrolyzes leucine and other hydrophobic amino acids from the N terminus of polypeptides. There are other classes of aminopeptidases, but assays of environmental samples generally show the greatest activities towards leucine- and alanine-linked substrates, so LAP activity is broadly used as an indicator of peptidase potential (Sinsabaugh & Foreman 2001; Stursova et al. 2006). Phosphatases (alkaline and acid, AP) hydrolyze phosphomonoesters, and in some cases phosphodiesters, releasing phosphate (Turner et al. 2002; Toor et al. 2003). The degradation of polyphenols (e.g. lignin, tannin and their degradation products) is an oxidative process (Kirk & Farrell 1987). Two classes of enzymes have a large role. Phenol oxidases (POX, e.g. laccases) have Cu-containing prosthetic groups with redox potentials sufficient to extract electrons from phenolic groups (Mayer  2008 Blackwell Publishing Ltd/CNRS & Staples 2002). Peroxidases (PER, e.g. lignin peroxidase, Mn peroxidase) have Fe-containing haeme prosthetic groups that use H2O2 or secondary oxidants to degrade aromatic compounds (Dorán & Esposito 2000; Hofrichter 2002). Statistical analysis Univariate and multivariate (enter-removal) linear regression analyses were used to relate mean ecosystem EEA (Table 3) to variation among ecosystems in mean annual temperature (MAT), mean annual precipitation (MAP), soil pH and SOM concentration (Table 2). A principal components analysis that included data from 24 ecosystems was used to reduce the seven enzyme variables to two factors. The remaining 16 sites had missing data for one or more enzyme activities. Mean factor values with 95% confidence intervals were calculated for each ecosystem to graphically display largescale patterns in the distribution of soil EEA. Ratios of ln(BG) : ln(AP) and ln(BG) : ln(NAG + LAP) activities were calculated for all cases. These indices, measures of the enzymatic resources directed towards acquisition of organic P and organic N relative to C, were used to test for functional convergence in soil EEA distributions across ecosystems and compare relative nutrient demand in relation to climatic gradients. RESULTS The potential activities g)1 dry soil of four enzymes, BG, CBH, NAG and AP, varied across ecosystems in relation to SOM concentration (R2: 0.55, 0.42, 0.49 and 0.60 respectively; Fig. 1, Table 4). Five enzyme activities had significant but weaker univariate relationships with bulk soil pH (R2: CBH 0.12, NAG 0.31, LAP 0.28, AP 0.36, POX 0.17; Table 4). Links to climate parameters were more tenuous: CBH and NAG were correlated with MAT; BG, LAP and AP were correlated with MAP (Table 4). Multiple regressions that Abbreviation Location Cover ⁄ Biome Soil SOM (%) pH MAT (C) MAP (mm) Crested Butte, CO Konza Prairie LTER, KS Kruger National Park, SA Ukulinga, SA Niwot Ridge LTER, CO Niwot Ridge LTER, CO Duke Forest, NC Oak Ridge National Lab, TN Sevilleta LTER, NM Sevilleta LTER, NM Sevilleta LTER, NM Manistee National Forest, MI Manistee National Forest, MI Manistee National Forest, MI Delta Junction, AK Delta Junction, AK Delta Junction, AK McMurdo Dry Valley, ANT Arctic LTER, AK Arctic LTER, AK Cedar Creek, MN Cedar Creek, MN Cedar Creek, MN Cedar Creek, MN Cedar Creek, MN UT Arboretum, OH Fuller Preserve, OH Harvard Forest, MA Chicago Botanical Garden, IL CB KNZ KRNP UKL NWT ⁄ S NWT ⁄ P DF ORNL SEV ⁄ G SEV ⁄ C SEV ⁄ J MNF ⁄ O MNF ⁄ MO MNF ⁄ M DJ ⁄ A DJ ⁄ S DJ ⁄ H MCM ARC ⁄ T ARC ⁄ S CDR ⁄ H CDR ⁄ M CDR ⁄ A CDR ⁄ O CDR ⁄ P UTA FP HFR CBG N39 W107 N39 W97 S24 E32 S30 E29 N40 W105 N40 W105 N36 W79 N36 W84 N34 W107 N34 W107 N34 W107 N44 W85 N44 W85 N44 W85 N63 W145 N63 W145 N63 W145 S77 E163 N69 W149 N69 W149 N45 W93 N45 W93 N45 W93 N45 W93 N45 W93 N41 W83 N41 W8 N42 W71 N42 W88 Sagebrush shrubland Tall grass prairie Tall grass prairie Tall grass prairie Spruce ⁄ fir forest Lodgepolepine forest Loblolly pine forest Sweetgum forest Grama grassland Creosote shrubland Juniper shrubland Black ⁄ White oak forest Sugar maple ⁄ red oak forest Sugar maple ⁄ basswood forest Aspen forest Black Spruce forest Herbaceous sere Cold desert Tundra ⁄ tussock Tundra ⁄ shrub Forb ⁄ grass grassland Sugar maple ⁄ basswood forest Bigtooth aspen forest Pin oak forest White pine forest Oak forest Oak forest Mixed deciduos forest Maple forest 8.4 8.2 6.5 13.0 28.8 28.7 5.0 2.8 1.8 2.4 3.7 2.5 2.7 4.5 26.7 19.7 14.5 0.55 94 81 2.0 3.3 2.5 2.0 1.2 12.0 12.0 12.7 11.3 8.0 5.6 5.1 5.1 5.0 5.0 5.0 6.0 8.2 7.5 7.2 3.9 4.1 5.6 5.3 5.2 5.5 8.7 4.5 4.9 6.0 5.3 5.6 5.1 5.6 6.4 6.9 4.3 6.4 1.1 13.0 22.9 17.6 )3.7 )3.7 15.5 14.2 13.2 13.2 13.2 9.7 9.7 9.7 )1.9 )1.9 )1.9 )22.4 )9 )9 6.7 6.7 6.7 6.7 6.7 10 10 7.6 9.3 650 835 550 694 930 930 1140 1390 250 250 250 890 890 890 311 311 311 100 330 330 801 801 801 801 801 856 856 1100 935 Barro Colorado Monument, Panama Luquillo LTER, PR Luquillo LTER, PR Luquillo LTER, PR Luquillo LTER, PR CAP LTER, AZ CAP LTER. AZ Kitty Todd, OH Southview savannah, OH Secor Metropark, OH Ohio University, OH BCNM LUQ ⁄ M LUQ ⁄ P LUQ ⁄ C LUQ ⁄ LM CAP ⁄ D CAP ⁄ U KT SS SM OU N9 W80 N18 W66 N18 W66 N18 W66 N18 W66 N33 W112 N33 W112 N41 W83 N41 W83 N41 W83 N39 W82 Lowland tropical forest Montane tropical forest Palm forest Cloud forest Lower montane forest Sonoran desert Urban Sonoran desert Mesic tallgrass prairie Mesic tallgrass prairie Oak ⁄ maple forest Oak ⁄ maple ⁄ ash forest Alfisol Udic arguistoll Alfisol Plinthic paleustalf Inceptisol Inceptisol Acidic hapludult Aquic hapludult Thermic halpocalcid Thermic halpocalcid Mesic halpocalcid Entic halporthod Typic halporthod Typic halporthod Pergelic cryaquepts Pergelic cryaquepts Pergelic cryaquepts Anhyorthels ⁄ Anhyoturbels Typic aquaturbel Aquic umborthel Udipsamments Udipsamments Udipsamments Udipsamments Udipsamments Aeric haplaquept Typic argiaquoll Typic Dystrocrept Oxyaquic Hapludalf ⁄ Aeric Epiaqualf Oxisol Aquic tropohumults Aquic tropohumults Aquic tropohumults Aquic tropohumults Aridisol Aridisol Typic Haplaquolls Typic Udipsamments Typic Haplaquolls Hapludalfs 9.4 10.5 23.6 25.9 14.7 2.2 3.1 10.4 3.3 7.4 9.8 5.5 5.0 4.9 4.9 5.0 7.7 7.6 7.2 6.4 6.6 5.9 27 19.6 19.0 18.9 21.0 17 17 10 10 10 10 2600 3137 4172 3237 3500 250 250 900 900 900 900 SOM, soil organic matter; MAT, mean annual temperature; MAP, mean annual precipitation. Stoichiometry of soil enzyme activity 1255  2008 Blackwell Publishing Ltd/CNRS Site Review and Synthesis Table 2 Study locations and ecosystem characteristics 1256 R. L. Sinsabaugh et al. Review and Synthesis Table 3 Potential soil EEA across ecosystems shown as mean values (nmol h)1 g SOM)1) with coefficients of variation Site n BG CB KNZ KRNP UKL NWR ⁄ S NWR ⁄ P DF ORNL SNWR ⁄ G SNWR ⁄ C SNWR ⁄ J MNF ⁄ O MNF ⁄ MO MNF ⁄ M DJ ⁄ A DJ ⁄ S DJ ⁄ H MCM ARC ⁄ T ARC ⁄ S CDCR ⁄ H CDCR ⁄ M CDCR ⁄ A CDCR ⁄ O CDCR ⁄ P UTA FP HF CBG BCNM LUQ ⁄ M LUQ ⁄ P LUQ ⁄ C LUQ ⁄ LM CAP ⁄ D CAP ⁄ U OH ⁄ KT OH ⁄ SS OH ⁄ SM OH ⁄ OU GLOBAL 45 24 12 12 169 53 45 18 87 27 27 39 39 39 4 4 4 44 40 40 36 18 18 36 36 40 30 12 24 8 3 3 3 3 10 10 24 12 48 10 1154 1850 3320 2670 2650 4290 3790 3450 4490 2740 2280 1820 1480 2310 3690 3160 3360 4850 119 3940 1940 6810 4640 5840 7080 9200 4240 5940 NA 4450 1920 367 355 176 986 1080 1180 1160 981 1410 4560 3320 CBH (33) (18) (16) (24) (116) (115) (54) (24) (109) (172) (50) (67) (46) (41) (70) (37) (18) (139) (93) (33) (49) (31) (39) (48) (67) (40) (20) (53) (33) (27) (29) (11) (14) (50) (33) (92) (37) (63) (41) (70) NA 1010 525 794 2290 2120 757 1040 438 395 278 450 641 1050 NA NA NA NA 2090 881 1080 1070 1190 1560 1960 1510 2700 NA 1540 512 41 35 33 164 67 77 214 1480 337 1710 942 (18) (18) (47) (149) (105) (53) (34) (104) (192) (64) (51) (37) (40) (74) (60) (106) (41) (59) (51) (92) (57) (24) (49) (46) (53) (8) (14) (26) (94) (48) (97) (63) (91) (58) (78) NAG LAP 377 2530 1210 2750 3190 3330 3340 2100 114 200 145 1450 1110 1450 2750 2760 1570 NA 2080 1670 2920 1870 2680 3210 3090 2340 2490 NA 2080 1560 189 167 108 433 42 42 705 4820 845 2320 1740 763 205 62 72 125 115 121 187 5730 7590 4380 131 192 396 47 34 31 3920 637 340 NA NA NA NA NA 2490 3930 NA 2010 158 42 25 11 79 5770 7360 652 777 336 666 1450 (36) (25) (18) (31) (100) (127) (46) (23) (81) (187) (47) (58) (73) (90) (68) (59) (43) (66) (56) (66) (31) (24) (50) (60) (57) (28) (39) (40) (36) (33) (44) (42) (53) (46) (104) (50) (91) (25) (70) AP (94) (32) (20) (40) (180) (135) (70) (44) (44) (135) (56) (69) (47) (29) (57) (76) (163) (88) (108) (56) (77) (38) (66) (108) (20) (20) (30) (28) (55) (44) (85) (69) (183) (52) (156) 988 6180 4200 3700 5440 5270 9460 9130 2490 2140 2670 2440 4070 4110 2550 3850 2820 3070 NA NA 9820 5870 8250 13 000 14 200 5960 5400 NA 5120 11 500 6132 1210 3400 4720 983 1000 2100 15 800 1850 5200 5300 POX (74) (12) (18) (33) (107) (120) (77) (27) (47) (184) (39) (45) (27) (29) (33) (30) (20) (100) (75) (31) (28) (67) (80) (39) (40) (48) (27) (3) (38) (25) (20) (43) (34) (91) (29) (61) (21) (71) 414 000 3880 2220 11 700 13 900 6150 12 800 17 500 1 016 000 383 000 299 000 1110 386 40 ND 37 700 ND 210 000 NA NA 424 1370 4780 6020 2880 54 500 48 600 9700 8230 ND 24 000 5400 1775 17 100 8140 8910 7 33 200 16 500 1400 70 600 PER (50) (77) (144) (162) (170) (188) (136) (69) (109) (178) (70) (207) (305) (323) (129) (89) (299) (260) (184) (184) (285) (81) (89) (57) (275) (87) (93) (90) (75) (129) (147) (187) (137) (157) (316) (266) NA 3120 209 1500 1540 3750 163 000 412 000 3 056 000 842 000 807 000 31 000 33 300 9570 11 500 6830 ND 217 000 NA NA 93 400 54 100 69 800 96 700 148 000 36 000 69 700 50 400 34 200 5200 44 800 60 600 29 400 65 000 71 800 34 100 12 5400 10 600 15 700 178 000 (149) (346) (168) (222) (208) (101) (64) (119) (82) (57) (159) (147) (192) (120) (200) (107) (61) (41) (30) (53) (36) (96) (58) (65) (81) (55) (32) (50) (78) (43) (35) (51) (233) (346) (153) (142) (293) n, number of cases (number of experimental units sampled · number of sampling dates); NA, not assayed; ND, not detected. Enzyme abbreviations given in Table 1. included edaphic and climatic variables accounted for 50– 70% of the variation in activity among ecosystems, except for PER (22%). Within these regression models, soil pH was a significant variable for all enzymes except CBH. Extracellular enzyme activity potentials are also commonly presented as specific activities (i.e. activity g)1 SOM) to analyse and compare the dynamics of decomposition (Table 3). Because of their strong covariance with SOM, the mean-specific activities of BG, CBH, NAG and AP varied  2008 Blackwell Publishing Ltd/CNRS by less than an order of magnitude across ecosystems and showed similar coefficients of variation (CV, 70–78%, Table 3). Specific LAP, POX and PER activities, which showed stronger relationships with soil pH (Fig. 2), varied more widely across ecosystems with CVs of 156%, 266% and 293%, respectively (Table 3). On average, spatiotemporal variation in specific EEA within ecosystems was lower than the variation among ecosystems (mean within ecosystem CV: BG 50%, CBH 64%, AP 49%, NAG 57%, LAP 71%, Review and Synthesis Stoichiometry of soil enzyme activity 1257 Figure 1 Natural logarithm of mean extracellular enzyme activity g)1 dry soil by site in relation to natural logarithm of soil organic matter concentration (%). Linear regressions are shown for the four enzymes with statistically significant relationships (P < 0.05). R2 values for BG, CBH, AP and NAG are 0.55, 0.42, 0.60 and 0.49 respectively. Slopes are 0.98, 0.96, 0.80 and 1.13 respectively. Enzyme abbreviations given in Table 1.  2008 Blackwell Publishing Ltd/CNRS 1258 R. L. Sinsabaugh et al. Review and Synthesis Table 4 Regression statistics relating ln(EEA g)1 soil dry mass) to climatic and edaphic variables BG CBH NAG LAP AP POX PER SOM MAT MAP pH Multiple 0.55* 0.42* 0.49 –* 0.60* –* – – 0.46* 0.31* –* – – – 0.22 – – 0.20* 0.18 – –* –* 0.12 0.31* 0.28* 0.36* 0.17* –* 0.56 0.56 0.57 0.70 0.63 0.50 0.22 Values are R2 statistics for significant (P < 0.05) linear regressions. Multiple is R2 statistics for multiple linear regressions (stepwise removal) of ln(EEA g)1 DM) as f(SOM, MAT, MAP, pH). Abbreviations and units given in Tables 1 and 2. POX regressions exclude five sites with anomalous undetectable values; PER regressions exclude two sites with anomalous undetectable values (Table 3). *Variables that make significant (t-test, P < 0.05) contributions to the multiple linear regressions. POX 158%, PER 116%). However, within ecosystems, the CV for hydrolytic activities covaried with the number of observations (R2: BG 0.29, CBH 0.29, AP 0.23, NAG 0.17, LAP 0.23, P < 0.05), so the full magnitude of spatiotemporal variation within many of the ecosystems represented may be underestimated. Variation in oxidative activities, though greater than that of hydrolytic activity, was not correlated with sampling effort (R2: POX 0.05, PER 0.006, P > 0.05). The specific activities of all seven enzymes had statistically significant relationships with soil pH within multiple linear regression models, and all but BG and CBH also showed significant univariate regressions with pH (Fig. 2, Table 5). Relationships with climate variables were weaker: only three specific activities (BG CBH, LAP) had significant relationships with MAP; two (CBH, NAG) had significant relationships with MAT (Table 5). Multiple regression models that included the two climatic (MAP and MAT) and soil pH captured 17% (AP) to 70% (LAP) of between ecosystem variances in EEA (Table 5). Principal components analysis (PCA) of data from 24 ecosystems reduced the seven enzyme variables to two factors that captured 80% of the variation. Ordination of ecosystems by these factors showed two discrete distributions (Fig. 3). Arid and semiarid sites, which generally have low SOM and alkaline soil pH, varied primarily in relation to factor 2 (32% of variance, positively correlated with LAP, POX and PER). Wetter ecosystems, which generally have acidic soil pH, varied principally along factor 1 (46% of variance, positively correlated with BG, CBH, NAG and AP). No sites showed high activity for both sets of variables.  2008 Blackwell Publishing Ltd/CNRS Estimates of C : N : P composition for soil and soil microbial biomass converge on 186 : 13 : 1 and 60 : 7 : 1 respectively (Cleveland & Liptzin 2007). However, nutrient acquisition effort, as indicated by the potential activities of the hydrolytic enzymes that generate readily consumed products from the largest soil pools of organic C, N, and P (i.e. cellulose, protein, chitin, peptidoglycan and sugar phosphates), may be more equitably distributed. The ratio ln(BG) : ln(NAG + LAP), an indicator of potential C : N acquisition activity averaged 1.02 ± 0.20 (SD); the corresponding C : P ratio, represented by the ratio of ln(BG) : ln(AP) activity, was 0.95 ± 0.15 (Fig. 4). By these indicators, the ratio of C : N : P acquisition activity is c. 1 : 1 : 1. Although lignin, tannin and other aromatic components of plant and microbial biomass are mineralized within the soil profile, they are not primary carbon sources for any major group of soil microorganisms, so POX and PER activities are not included in this acquisition ratio. At the ecosystem scale, enzymatic indicators of relative nutrient availability showed patterns in relation to climatic gradients. The mean enzymatic C : P acquisition ratio for ecosystems declined in relation to both MAT (R2 = 0.33) and MAP (R2 = 0.71), suggesting that P availability declines relative to C as soil-weathering intensity increases (Fig. 5). Using a multiple linear regression [enzymatic C : P ratio = f(MAT, MAP)], the C : P acquisition ratio is predicted to decrease from 1.1 to 0.7 across latitudinal gradients in weathering intensity (R2 = 0.74, F = 47, MAT coefficient: )0.00239, MAP coefficient: )0.0000704, intercept: 1.044) The mean C : N acquisition ratio for ecosystems was not related to MAT, but did show a weak positive relationship with MAP (R2 = 0.16, Fig. 5). Because the C : P and C : N acquisition ratios had opposing trends in relation to MAP, the N : P acquisition ratio was negatively related to MAP (R2 = 0.58, Fig. 5). DISCUSSION The description of soil EEA on a global scale provides a frame of reference for comparing ecosystems and an opportunity to relate the soil microbial community function to global patterns of microbial biomass composition, nutrient dynamics and SOM storage. Our analysis documents that the most commonly measured extracellular enzyme activities show different ranges of variation and different distributions in relation to ecosystem variables, yet converge on a common pattern linked to the stoichiometry of microbial growth. For BG, CBH and AP, activity g)1 dry soil tracked SOM content. LAP, POX and PER activities varied widely but generally increased with soil pH, while NAG activity was strongly related to both SOM (positively) and soil pH (negatively). When specific activities (i.e. activity g)1 SOM) Review and Synthesis Stoichiometry of soil enzyme activity 1259 Figure 2 Natural logarithm of mean extracellular enzyme activity g)1 soil organic matter by site in relation to soil pH. Linear regressions are shown for the five enzymes with statistically significant relationships (P < 0.05). R2 values for POX, PER, NAG, LAP and AP are 0.21, 0.10, 0.23, 0.62 and 0.16 respectively. Regression slopes are: POX 0.91, PER 0.63, NAG )0.54, LAP 1.25 and AP )0.25. Enzyme abbreviations given in Table 1. are compared, all enzymes show a statistically significant relationship to soil pH in either univariate or multivariate models with weak negative trends for BG, CBH and AP, a strong negative trend for NAG and strong positive trends for LAP, POX and PER. Whether EEA is expressed g)1 of soil or g)1 SOM, soil pH emerges as the variable most  2008 Blackwell Publishing Ltd/CNRS 1260 R. L. Sinsabaugh et al. Review and Synthesis Table 5 Regression statistics relating ln(EEA g)1 SOM) to climatic and edaphic variables BG CBH NAG LAP AP POX PER MAT MAP pH Multiple – 0.39* 0.21* – – – – 0.19* 0.27* – 0.30* – – – –* –* 0.23* 0.62* 0.16* 0.21* 0.10* 0.40 0.53 0.50 0.70 0.17 0.43 0.25 Values are R2 statistics for significant (P < 0.05) linear regressions. Multiple is R2 statistics for multiple linear regressions (stepwise removal) of ln(EEA g)1 SOM) as f(MAT, MAP, pH). Abbreviations and units given in Tables 1 and 2. POX regressions exclude five sites with anomalous undetectable values; PER regressions exclude two sites with anomalous undetectable values (Table 3). *Variables that make significant (t-test, P < 0.05) contributions to the multiple linear regressions. Figure 3 Ordination of 24 ecosystems based on potential soil extracellular enzyme activity g)1 organic matter using principal components analysis (varimax rotation). Factor 1 (46% of variance) is correlated with BG (r = 0.89), CBH (0.84), NAG (0.92) and AP (0.84). Factor 2 (31% of variance) is correlated with LAP (0.85), POX (0.83) and PER (0.74). The vertical grouping represents arid and semiarid ecosystems with soil pH > 7. The horizontal grouping represent ecosystems with relatively high precipitation and soil pH < 7. Values shown are means with 95% confidence intervals. Enzyme abbreviations given in Table 1. closely linked to ecosystem variation (Tables 4 and 5). These patterns resemble recent findings that microbial diversity in soil and other systems also follow pH gradients (Baath & Anderson 2003; Fierer & Jackson 2006; Singh et al. 2006; Cookson et al. 2007).  2008 Blackwell Publishing Ltd/CNRS Figure 4 Ratio of ln(BG) : ln(NAG + LAP), an indicator of potential C : N acquisition activity, in relation to the ratio ln(BG) : ln(AP), an indicator of potential C : P acquisition activity. The centroid is 0.95 ± 0.15 (SD) for C : P and 1.02 ± 0.20 for C : N values > 1.2 for either ratio constrain values of the complementary ratio. The regression C : N = 0.75 (C : P) + 0.31 has an R2 value of 0.28, n = 929. The association of pH and EEA reflects interactions at multiple scales of organization. Soil pH has direct biochemical effects on the activity of extracellular enzymes immobilized in the soil matrix. Glycosidases have pH optima c. 5 ± 1. POX, lignin peroxidases and most proteases (metallo-proteases, serine proteases) have optima of 8 ± 1. Extracellular phosphatases are produced in acid and alkaline active forms by various taxa. At the ecosystem scale, soil pH reflects climatic controls on soil weathering and plant community composition, which may affect the large-scale distribution of EEA through changes in nutrient availability and SOM composition, as well as microbial community composition. These interactions over multiple levels of organization generate global patterns that are not observed at the ecosystem scale. The most conspicuous of these is the distribution of oxidative activity. Basidiomycetes produce a variety of extracellular oxidative enzymes and are generally considered to be the most efficient degraders of lignin (Rabinovich et al. 2004; Baldrian 2006). These organisms are most abundant in mid- to high-latitude forests where the dominant plants have high lignin concentrations and the soil is acidic. Within these ecosystems, POX and PER activities tend to increase with secondary succession (Sinsabaugh et al. 2005). This trend is evident for the MNF and DJ ecosystems (Table 3). But at the global scale, this biome-specific trend is overwhelmed by the inclusion of arid alkaline soils, which have near optimal pH for POX and PER activities and edaphic conditions that may promote enzyme stability (Stursova & Sinsabaugh 2008), even though basidiomycetes are relatively uncommon (Porras-Alfaro et al. 2008). Review and Synthesis Stoichiometry of soil enzyme activity 1261 Figure 5 Mean ecosystem ratios of C : P, C : N and N : P acquisition activity, as indicated by ratios of ln(BG) : ln(AP), ln(BG) : ln(LAP + NAG) and ln(LAP + NAG) : ln(AP) respectively, in relation to mean annual temperature and mean annual precipitation. Data from the McMurdo Dry Valleys are excluded from the C : P and C : N graphs because BG activities are extremely low. Enzyme abbreviations are listed in Table 1. Regression statistics for C : P vs. MAT: n = 36, R2 = 0.33, F = 16.7, P < 0.001, a = )0.0070; for C : P vs. MAP: n = 36, R2 = 0.71, F = 84.2, P < 0.001, a = )0.000080; for C : N vs. MAP: n = 38, R2 = 0.16, F = 6.61, P = 0.014, a = 0.000035; for N : P vs. MAP: n = 37, R2 = 0.58, F = 47.5, P < 0.001, a = )0.00011. The regressions for C : N and N : P vs. MAT were not statistically significant (P > 0.1).  2008 Blackwell Publishing Ltd/CNRS 1262 R. L. Sinsabaugh et al. The oxidative degradation of lignin, tannin and other aromatic components of plant litter is generally considered the rate-limiting step in decomposition (Meentemeyer 1978; Fog 1988). Freeman et al. (2001), for example, proposed that POX activity was the proximate control on organic matter mineralization, and thereby CO2 efflux, in high-latitude peats (histosols), and that regional climate warming could release constraints imposed by low oxygen availability on the activity of these immobilized enzymes, leading to net losses of SOM. Our EEA analyses suggest a broader context for POX and PER in SOM storage. Despite low rates of primary production, SOM content is greatest in highlatitude ecosystems where POX and PER activities in soil are physicochemically constrained by low pH, low temperature and low oxygen availability caused by soil flooding. SOM contents are lowest in arid ecosystems, which also have low rates of primary production, where alkaline pH increases the solubility of polyphenols and optimizes POX and PER activities (Collins et al. 2008). Another global pattern in the distribution of EEA is the convergence of C : N : P acquisition potentials, as measured by ln(BG) : ln(LAP + NAG) : ln(AP) activities (Fig. 3). Across ecosystems, BG activity was most strongly correlated with the abundance of SOM. Despite its low pH optimum, specific BG activity varied only weakly with soil pH, presumably because cellulose and other b-1,4-glucan polymers dominate the organic matter inputs of vegetated ecosystems. Declines in specific activity as a result of higher soil pH are counteracted by increased enzyme expression. The role of plant litter in controlling BG activity is suggested by data from the McMurdo Dry Valleys of Antarctica where there are no plants, soil pH is high and the specific activity of BG is only 3% of the global average (Table 3). As indicators of organic N acquisition from amino acids and amino sugars, LAP and NAG showed similar ranges of activity but inverse relationships to soil pH (Fig. 2). As a result, the sum of LAP + NAG was similar across ecosystems. AP activity, like BG, varied across ecosystems largely in relation to SOM abundance. Soil pH had little effect on specific AP activity, presumably because both acid and alkaline active enzymes are produced. Because of these trends, specific C, N and P acquisition potentials generally showed a consistent stoichiometry across ecosystems, even though the component activities had different relationships with environmental variables. The consistency of stoichiometric relationships across ecosystems is unexpected because experimental manipulations within ecosystems show that C, N and P acquisition activities can be modulated by inorganic nutrient availability (Olander & Vitousek 2000; Sinsabaugh et al. 2002; Stursova et al. 2006), following resource allocation models based on the premise that cellular resources directed towards N and P  2008 Blackwell Publishing Ltd/CNRS Review and Synthesis acquisition reduce resources available for C acquisition (Sinsabaugh & Moorhead 1994; Allison et al. 2007). The convergence of the C : N : P acquisition ratio on a global scale shows that the plasticity of these relationships is constrained. The C : N and C : P acquisition ratios increase colinearly to a maximum value of 1.2 (Fig. 4). Values > 1.2 for either ratio occur only when the other ratio remains < 1.2 and such instances occurred in < 3% of the cases in our dataset. Thus, the C : N : P acquisition ratio appears to be an integral feature of soil microbial community function, linking environmental nutrient availability to the C : N : P stoichiometry of microbial biomass (Cleveland & Liptzin 2007). Although nutrient acquisition ratios are constrained by stoichiometry, variance at the ecosystem scale follows largescale biogeochemical patterns. Biogeochemical theory predicts that soil N availability should be highest in tropical ecosystems, while P availability should be greatest in mid- to high-latitude ecosystems (Walker & Syers 1976; Martinelli et al. 1999). P is a rock-derived nutrient that may be lost due to leaching or occlusion in mineral particles in highly weathered soils. N, an atmospherically derived nutrient, tends to be scarce in areas that have experienced recent glaciation. These large-scale trends are apparent in the elemental C : N : P ratios of plants (McGroddy et al. 2004; Reich & Oleksyn 2004). Our analyses show that they are also reflected in the enzymatic ratios of C : N : P acquisition by soil microbial communities. While individual enzyme activities were not strongly linked to climatic variables, mean ecosystem C : P acquisition ratios declined as MAT and MAP increased, indicating that soil microbial communities direct more effort to acquiring and cycling P relative to processing C in more weathered soils (Fig. 5). The C : N acquisition ratio was much less responsive to climatic gradients. However, the trend towards greater ratios, indicative of higher relative N availability, with increasing MAP is consistent with biogeochemical predictions. The contrast between the robust C : P acquisition relationship with climate measures and the weak C : N acquisition relationship extends patterns observed at the ecosystem scale. In both aquatic and terrestrial ecosystems, an inverse relationship between extracellular phosphatase activity and relative P availability is a general phenomenon, reflecting the role of P in energy metabolism. Because most extracellular phosphatases will hydrolyze phosphate from a wide range of substrates, it is relatively easy to measure this potential with a single assay. N acquisition from organic matter is more complex than P acquisition. N is distributed among several classes of polymers as well as humic molecules, so N acquisition strategies are linked to the C-substrate preferences of particular taxa (McGill & Cole 1981; Manzoni et al. 2008). In the context of decomposition models, three broad strategies Review and Synthesis can be defined, each assigned to a guild of organisms: opportunists consume labile proteins, decomposers need external N inputs to decompose lignocellulose and miners use oxidative enzymes to breakdown humus for C and N (Moorhead & Sinsabaugh 2006). Given the diversity of N acquisition strategies and their conflation with carbon acquisition, it is not surprising that studies that compare soil EEA responses to experimental N amendment produce mixed results. In some ecosystems, POX and PER (e.g. Frey et al. 2004; Sinsabaugh et al. 2005), LAP (e.g. Stursova et al. 2006) or NAG activities (e.g. Olander & Vitousek 2000; DeForest et al. 2004) decrease with N amendment, but others studies find no response (e.g. Zeglin et al. 2007). As a result, large-scale relationships between particular enzyme activities and measures of N availability are likely to be weaker than relationships between P availability and phosphatase activity. Our analyses indicate that the enzymatic potential for hydrolyzing the labile components of SOM is tied to substrate availability, soil pH and the stoichiometry of microbial nutrient demand. The enzymatic potential for oxidizing the recalcitrant fractions of soil organic material, which is a proximate control on SOM accumulation, is most strongly related to soil pH. These trends provide insight into the biogeochemical processes that create global patterns in ecological stoichiometry and organic matter storage. ACKNOWLEDGEMENTS This paper is the product of a workshop funded by the National Science Foundation Long Term Ecological Research Network Office. REFERENCES Allison, S.D., Gartner, T., Holland, K., Weintraub, M. & Sinsabaugh, R.L. (2007). Soil enzymes: linking proteomics and ecological process. In: Manual of Environmental Microbiology. ASM Press, 3rd Edition, Washington D. C., pp. 704–711. Andersson, M., Kjøller, A. & Struwe, S. (2005). Microbial enzyme activities in leaf litter, humus and mineral soil layers of European forests. Soil Biol. Biochem., 36, 1527–1537. Baath, E. & Anderson, T.H. (2003). Comparison of soil fungal ⁄ bacterial ratios in a pH gradient using physiological and PLFA-based techniques. Soil Biol. Biochem., 35, 955–963. Baldrian, P. (2006). Fungal laccases-occurrence and properties. FEMS Microbiol. Rev., 30, 215–242. Burns, R.G. (1978). Soil Enzymes. Academic Press, New York. Burns, R.G. & Dick, R.P. (2002). Enzymes in the Environment: Activity, Ecology and Applications. Marcel Dekker, New York. Caldwell, B. (2005). Enzyme activities as a component of soil biodiversity: a review. Pedobiologia, 49, 637–644. Cleveland, C.C. & Liptzin, D. (2007). C:N:P stoichiometry in soil: is there a ‘‘Redfield ratio’’ for the microbial biomass? Biogeochemistry, 85, 235–252. Stoichiometry of soil enzyme activity 1263 Collins, S.L., Sinsabaugh, R.L., Crenshaw, C., Green, L.E., PorrasAlfaro, A., Stursova, M. et al. (2008). Pulse dynamics and microbial processes in aridland ecosytems. J. Ecol., 96, 413–420. Cookson, W.R., Osman, M., Marschner, P., Abaye, D.A., Clark, I., Murphy, D.V. et al. (2007). Controls on soil nitrogen cycling and microbial community composition across land use and incubation temperature. Soil Biol. Biochem., 39, 744–756. DeForest, J.L., Zak, D.R., Pregitzer, K.S. & Burton, A.J. (2004). Anthropogenic NO3- deposition alters microbial community function in northern hardwood forests. Soil Sci. Soc. Am. J., 68, 132–138. Dorán, N. & Esposito, E. (2000). Potential applications of oxidative enzymes and phenoloxidase-like compounds in wastewater and soil treatement: a review. Appl. Catal. B Environ., 6, 83–99. Fierer, N. & Jackson, R.B. (2006). The diversity and biogeography of soil bacterial communities. Proc. Natl Acad. Sci. U.S.A., 103, 626–631. Finzi, A.C., Sinsabaugh, R.L., Long, T.M. & Osgood, M.P. (2006). Microbial community responses to atmospheric CO2 enrichment in a Pinus taeda forest. Ecosystems, 9, 215–226. Fog, K. (1988). The effect of added nitrogen on the rate of decomposition organic matter. Biol. Rev., 63, 433–462. Freeman, C., Ostle, N. & Kang, H. (2001). An enzymic ÔlatchÕ on a global carbon store – a shortage of oxygen locks up carbon in peatlands by restraining a single enzyme. Nature, 409, 149. Frey, S.D., Knorr, M., Parrent, J.L. & Simpson, R.T. (2004). Chronic nitrogen enrichment affects the community structure and function of the soil microbial community in temperate hardwood and pine forests. For. Ecol. Manage., 196, 159–171. Hofrichter, M. (2002). Review: lignin conversion by manganese peroxidase (MnP). Enzyme Microb. Technol., 30, 454–466. Kirk, T.K. & Farrell, R.L. (1987). Enzymatic ‘‘combustion’’: the microbial degradation of lignin. Annu. Rev. Microbiol., 41, 465– 505. Lipson, D.A., Wilson, R.F. & Oechel, W. (2005). Effects of elevated atmospheric CO2 on soil microbial biomass, activity, and diversity in a Chaparral ecosystem. Appl. Environ. Microbiol., 71, 8573–8580. Ljungdahl, L.G. & Eriksson, K.-E. (1985). Ecology of microbial cellulose degradation. Adv. Microb. Ecol., 8, 237–299. Manzoni, S., Jackson, R.B., Trofymow, J.A. & Porporato, A. (2008). The global stoichiometry of litter nitrogen mineralization. Science, 321, 684–686. Martinelli, L.A., Piccolo, M.C., Townsend, A.R., Vitousek, P.M., Cuevas, E., McDowell, W. et al. (1999). Nitrogen stable isotopic composition of leaves and soil: tropical versus temperate forests. Biogeochemistry, 46, 45–65. Marx, M.C., Wood, M. & Jarvis, S.C. (2001). A microplate fluorimetric assay for the study of enzyme diversity in soils. Soil Biol. Biochem., 33, 1633–1640. Mayer, A.M. & Staples, R.C. (2002). Laccase: new functions for an old enzyme. Phytochemistry, 60, 551–565. McGill, W.B. & Cole, C.V. (1981). Comparative aspects of cycling of organic C,N, S and P through soil organic matter. Geoderma, 26, 267–286. McGroddy, M.E., Daufresne, T. & Hedin, L.O. (2004). Scaling of C:N:P stoichiometry in forests worldwide: implications of terrestrial Redfield-type ratios. Ecology, 85, 2390–2401. Meentemeyer, V. (1978). Macroclimate and lignin control of litter decomposition rates. Ecology, 59, 465–472.  2008 Blackwell Publishing Ltd/CNRS 1264 R. L. Sinsabaugh et al. Moorhead, D.L. & Sinsabaugh, R.L. (2006). A theoretical model of litter decay and microbial interaction. Ecol. Monogr., 76, 151–174. Olander, L.P. & Vitousek, P.M. (2000). Regulation of soil phosphatase and chitinase activity by N and P availability. Biogeochemistry, 49, 175–190. Porras-Alfaro, A., Lipinski, K., Herrera, J., Natvig, D.O. & Sinsabaugh, R.L. (2008). Diversity and distribution of soil fungal communities in a semiarid grassland. FEMS Microbiol. Ecol., in press. Rabinovich, M.L., Bolobova, A.V. & VasilÕchenko, L.G. (2004). Fungal decomposition of natural aromatic structures and xenobiotics: a review. Appl. Biochem. Microb., 40, 1–17. Redfield, A. (1958). The biological control of chemical factors in the environment. Am. Sci., 46, 205–221. Reich, P.B. & Oleksyn, J. (2004). Global patterns of plant leaf N and P in relation to temperature and latitude. Proc. Natl Acad. Sci. U.S.A., 101, 11001–11006. Schimel, J.P. & Weintraub, M.N. (2003). The implications of exoenzyme activity on microbial carbon and nitrogen limitation in soil: a theoretical model. Soil Biol. Biochem., 35, 549–563. Singh, B.K., Munro, S., Reid, E., Ord, B., Potts, J.M., Paterson, E. et al. (2006). Investigating microbial community structure in soils by physiological, biochemical and molecular fingerprinting methods. Eur. J. Soil Sci., 57, 72–82. Sinsabaugh, R.L. (2005). Fungal enzymes at the community scale. In: The Fungal Community, 3rd edn (eds Dighton, J., Oudermans, P. & White, J.). CRC Press, New York, pp. 237–247. Sinsabaugh, R.L. & Foreman, C.M. (2001). Activity profiles of bacterioplankton in a eutrophic river. Freshw. Biol., 46, 1–12. Sinsabaugh, R.L. & Moorhead, D.L. (1994). Resource allocation to extracellular enzyme production: a model for nitrogen and phosphorus control of litter decomposition. Soil Biol. Biochem., 26, 1305–1311. Sinsabaugh, R.L., Findlay, S., Franchini, P. & Fischer, D. (1997). Enzymatic analysis of riverine bacterioplankton production. Limnol. Oceanogr., 42, 29–38. Sinsabaugh, R.L., Carreiro, M.M. & Repert, D.A. (2002). Allocation of extracellular enzymatic activity in relation to litter composition, N deposition, and mass loss. Biogeochemistry, 60, 1–24. Sinsabaugh, R.L., Gallo, M.E., Lauber, C., Waldrop, M. & Zak, D.R. (2005). Extracellular enzyme activities and soil carbon dynamics for northern hardwood forests receiving simulated nitrogen deposition. Biogeochemistry, 75, 201–215. Skujins, J. (1978). History of abiontic soil enzyme research. In: Soil Enzymes (ed. Burns, R.G.). Academic Press, New York. pp. 1–50. Sterner, R.W. & Elser, J.J. (2002). Ecological Stoichiometry: The Biology of Elements from Molecules to the Biosphere. Princeton University Press, Princeton, NJ.  2008 Blackwell Publishing Ltd/CNRS Review and Synthesis Stursova, M. & Sinsabaugh, R.L. (2008). Stabilization of oxidative enzymes in desert soil may limit organic matter accumulation. Soil Biol. Biochem., 40, 550–553. Stursova, M., Crenshaw, C. & Sinsabaugh, R.L. (2006). Microbial responses to long term N deposition in a semi-arid grassland. Microb. Ecol., 51, 90–98. Toor, G.S., Condron, L.M., Di, H.J., Cameron, K.C. & CadeMenum, B.J. (2003). Characterisation of organic phosphorus in leachate from a grassland soil. Soil Biol. Biochem., 35, 1317–1323. Turner, B.L., McKelvie, I.D. & Haygarth, P.M. (2002). Characterisation of water extractable soil organic phosphorus by phosphatase hydrolysis. Soil Biol. Biochem., 34, 27–35. Vitousek, P.M. & Howarth, R.W. (1991). Nitrogen limitation on land and in the sea: how can it occur? Biogeochemistry, 13, 87–115. Waldrop, M.P. & Harden, J.H. (2008). Interactive effects of wildfire and permafrost on microbial communities and soil processes in an Alaskan black spruce forest. Glob. Chang. Biol., doi: 10.1111/j.1365-2486.2008.01661.x. Walker, T.W. & Syers, J.K. (1976). The fate of phosphorus during pedogenesis. Geoderma, 15, 1–19. Weintraub, M.N., Scott-Denton, L.E., Schmidt, S.K. & Monson, R.K. (2007). The effects of tree rhizodeposition on soil exoenzyme activity, dissolved organic carbon, and nutrient availability in a subalpine forest ecosystem. Oecologia, 154, 327–338. Zeglin, L.H., Stursova, M., Sinsabaugh, R.L. & Collins, S.L. (2007). Microbial responses to nitrogen addition in three contrasting grassland ecosystems. Oecologia, 296, 65–75. SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article: Appendix S1 Metadata. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. Editor, Johannes Knops Manuscript received 8 May 2008 First decision made 10 June 2008 Second decision made 7 August 2008 Manuscript accepted 19 August 2008