lable at ScienceDirect

Soil Biology & Biochemistry 43 (2011) 749e759

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lb io

Contrasting effects of nitrogen limitation and amino acid imbalanceon carbon and nitrogen turnover in three species of Collembola

Thomas Larsen a,b,*, Marc Ventura b,c, Diane M. O’Brien d, Jakob Magid e,Bente Aa. Lomstein f, John Larsen g,h

a Leibniz-Laboratory for Radiometric Dating and Stable Isotope Research, Christian-Albrechts-Universität zu Kiel, Max-Eyth-Str. 11-13, 24118 Kiel, GermanybBiogeodynamics and Biodiversity Group, Centre for Advanced Studies of Blanes (CEAB), Spanish Research Council (CSIC), Accés Cala Sant Francesc 14, 17300 Blanes, Catalonia, Spainc Institut de Recerca de l’Aigua, Universitat de Barcelona, Av. Diagonal, 684, 08034 Barcelona, Catalonia, Spaind Institute of Arctic Biology and Department of Biology and Wildlife, University of Alaska Fairbanks, Fairbanks, AK 99775-7000, USAeDepartment of Agriculture and Ecology, Faculty of Life Sciences, University of Copenhagen, Thorvaldsensvej 40, 1871 FC, DenmarkfDepartment of Biological Sciences, Section for Microbiology, University of Aarhus, Building 1540, NyMunkegade, 8000 Aarhus C, DenmarkgDepartment of Integrated Pest Management, Faculty of Agricultural Science, University of Aarhus, Forsøgsvej 1, 4200 Slagelse, DenmarkhCentro de Investigaciones en Ecosistemas, Universidad Nacional Autónoma de México, Apartado Postal 27-3 Santa María de Guido, C.P. 58090 Morelia, Michoacán, Mexico

a r t i c l e i n f o

Article history:Received 6 July 2010Received in revised form3 December 2010Accepted 7 December 2010Available online 23 December 2010

Keywords:Energy and nutrient budgetsDiet qualityIsotope patterns of amino acidsNutritional resourcesSoil detritivoresStable isotopesTissue replacement

* Corresponding author. Biogeodynamics and BioAdvanced Studies of Blanes (CEAB), Spanish ResearcSant Francesc 14, 17300 Blanes, Catalonia, Spain. Tel.: þ337806.

E-mail address: natursyn@gmail.com (T. Larsen).

0038-0717/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.soilbio.2010.12.008

a b s t r a c t

Soil animal detritivores play an important role in facilitating decomposition processes but little infor-mation is available on how the quality of dietary resources affects their stoichiometry of carbon (C)nitrogen (N) and phosphorus (P), and turnover of C and N. This study investigated how a fungal diet,Fusarium culmorum, with a low N content and imbalanced amino acid (AA) composition affected thephysiology of three soil-dwelling collembolans (Folsomia candida, Protaphorura fimata and Proisotomaminuta) in comparison to a control diet, Saccharomyces cerevisiae, with a high N content and balanced AAcomposition. We compared the elemental composition of animals, their growth rates and tissuereplacement of C and N.We alsomeasured the individual AA d13C to investigate the extent that Collembolamay rely on endogenous sources to compensate for scarcity of essential AAs. The results showed thatanimal’s N content tracked closely the composition of their diets, decreasing from around 10 to 7% N fromthe high to low N diet. They also had a significant increase of C and a decrease of P. P. fimata was lessaffected than F. candida and P. minuta. The total incorporation of C and N in the animals due to growth andtissue replacement decreased from 11e17 to 6e12% DM d�1 on the high and low N diet respectively withP. fimata experiencing the smallest change. Essential AAs d13C did not always match perfectly betweenCollembola species and their diets; particularly on the low N diet. Isotope patterns of AAs indicate thatbacteria may have been the alternative source of essential AAs. While the results of this study cannot beextrapolated directly to the dynamics of Collembola populations in the field, they serve to demonstratetheir flexibility in adapting physiologically to the temporal and spatial patchiness of the soil environment.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

The role of consumers in total ecosystem function has tradition-ally been viewedwithin the context of energy flow (Lindeman,1942;Odum, 1957; O’Neill, 1968) allowing ecologists to use a commoncurrency for estimating how consumers influence ecosystem

diversity Group, Centre forh Council (CSIC), Accés Cala34 972336101; fax: þ34 972

All rights reserved.

function (Kitchell et al., 1979; Visser et al., 1981). The concept ofenergy flows has been widely used to characterize and understandinteractions in the soil food web (Petersen and Luxton, 1982; deRuiter et al., 1995; Lavelle et al., 1997) demonstrating that whilethe autotrophs are responsible for determining the amount of carbonthat enters the system, animal detritivores play a crucial role ingoverning the availability of nutrients required for plant produc-tivity. Detritivores primarily exert their influence by microbialgrazing and by masticating and translocating organic matter(Ingham et al., 1986; Lussenhop, 1992; Coleman, 2008), and theseactivities are inmost ecosystems regulated by bottom-up control, i.e.by the quality and availability of dietary resources (Wardle, 2002;White, 2008; Pollierer et al., 2010).

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759750

The stoichiometry of dietary carbon (C), nitrogen (N) and phos-phorus (P) is in combination with the composition of biochemicalcompounds key for the nutritional value of the diet (Sterner andElser, 2002; Moe et al., 2005). The majority of dietary C, N and P issupplied to cell metabolism in the form of amines, phosphate ions,sugars and other simple compounds. In addition, animals also needessential complex compounds from the diet that they cannotsynthesize from simple precursors. Quantitatively, the mostimportant essential biochemical compounds are the essentialamino acids and fatty acids. While the requirements and dietaryrouting of fatty acids in soil detritivores have received considerableattention in recent years (e.g. Ruess et al., 2002, 2007; Evans et al.,2003; Chamberlain et al., 2004, 2005; Pollierer et al., 2010) onlylittle attention has been paid to the amino acids (Rothstein, 1963;Anderson et al., 1983; Pokarzhevskii et al., 1997). This is surprisingsince amino acids (AAs) unlike other nutrients must comprisea large proportion of the diet for the consumer to achievemaximumgrowth and fecundity (e.g. Lii et al., 1975; Bowen,1984; Kleppel andBurkart, 1995; Ramsay and Houston, 1998; Guisande et al., 1999;O’Brien et al., 2003; Boëchat and Adrian, 2006). It is likely thatdietary AA composition also affects physiology and food preferenceof soil detritivores, as is the case for dietary N content (Booth andAnderson, 1979; Bengtsson et al., 1985; Lavy and Verhoef, 1996a;Hogervorst et al., 2003; Jensen et al., 2006) and for other inverte-brates (Ventura and Catalan, 2010). The link between N and AAs isimportant because the majority of animal’s N content is in the formof AAs (e.g. Ventura, 2006) and because the consumer is morehomeostatic than their food in essential AAs (Sterner and Elser,2002; Anderson et al., 2004). Comparison between organisms andtheir dietary AA composition has therefore been proposed asa stoichiometric argument for obtaining knowledge of the mecha-nisms underlying N limitation (Cowgill et al., 1986; Anderson et al.,2004; Ventura and Catalan, 2010).

Detritus-based systems are widely considered to be extremelynutrient limited, (Schulten and Schnitzer, 1997; Moe et al., 2005),and detritivores have therefore adapted different behavioural andphysiological strategies to copewith temporal and spatial variationsof dietary resources (Cross et al., 2003). One strategy is ingestion ofthe most digestible components in soil such as microorganismsinvolved in decomposition. Another strategy is ingestion of lessnutritious components such as senescent plantmaterialmixedwithmicroorganisms living on these substrates (Plante et al., 1990).A feeding strategy on low nutrient diets requires morphologicaladaptations such large guts (Christiansen, 1964) and is demandingin terms of metabolic investments for maintaining a rich gut florafor digesting hardly decomposable materials (Sibly and Calow,1986). Detritivores foraging on high nutrient diets are not onlybetter adapted to sequester limiting nutrients by selective feedingbut also more prone to stoichiometric mismatches between theirown demands and dietary supply (Fagan and Denno, 2004). Severalauthors have proposed that scarcity of a specific nutrient in avail-able food also can impact their physiology such as growth (Fink andVon Elert, 2006) and reproduction (Færøvig and Hessen, 2003), andconsequently reduce consumer demands for that nutrient (Elseret al., 2000b, 2006; Kay et al., 2005). It has also been found thatsome animals are able to adjust their body C:N:P to the quality ofresources (Elser et al., 2000a; Cross et al., 2003; Martinson et al.,2008) but it is not clear to what degree soil detritivores are able toadjust their stoichiometry to dietary N limitations and how itimpacts their turnover of C and N.

To investigate how limitations in dietary N and AA imbalancesaffect the physiology of soil detritivores we investigated howa fungal diet (Fusarium culmorum) with a low N content andimbalanced AA composition affected body composition of C, N andP and turnover of C and N in three soil-dwelling Collembola species

(Folsomia candidaWillem 1902; Protaphorura fimata Gisin 1952 andProisotoma minuta Tullberg 1871) and compared them with thosefed a control diet (Saccharomyces cerevisiae) with a high N contentand balanced AA composition. While N-containing sugars such aschitin can constitute up to 45% of the organic fraction of fungal cellwalls and structural membranes (Roberts, 1992) and be an impor-tant N source for Collembola (Borkott and Insam, 1990), this studyfocuses particularly on AAs as an N source for following reasons:chitin has a lower N content than AAs (6.6% vs. 16.5% N) and is insoil a less abundant N pool than the AAs(Schulten and Schnitzer,1997), which in fungi typically constitute the major fraction ofthe total N (Kurkela et al., 1980). Moreover, essential AAs cannot bebiosynthesized by metazoans and need to be provided from thediet or endosymbionts (Nation, 2002). The Collembola speciesselected for our study have previously been shown to have differentdietary preferences. P. fimata has the largest dietary breadth of thethree species and is able to function as a root herbivore as well asa detritivore (Ulber, 1983; Endlweber et al., 2009) in contrast toF. candida and P. minuta, that are thought to feed primarily on fungiand other highly nutritious sources (Harasyme and Sinha, 1974;Jørgensen et al., 2003; Larsen et al., 2007, 2008). The differencesbetween the species were also demonstrated by Larsen et al.(2009b) who found that P. fimata primarily allocated nutrients togrowth contrary to P. minuta that primarily allocated nutrients tometabolism and reproduction.

In this study we quantified the incorporation of C and N inCollembola by changing the isotopic (13C and 15N) composition oftheir diets after they had reached sexual maturity and followedgrowth, body composition of C, N and P, and the rate of isotopicchange in their tissues during a period of 28 days. This designallowed us to partition between C and N allocations to growth andtissue replacement (Larsen et al., 2009b) and test the hypothesisthat F. candida and P. minuta would decrease their rates of growthand tissue replacement relatively more than P. fimata on the low Ndiet. We also measured d13C of the AAs in the animals and theirdiets to determine whether the AA carbon skeletons were derivedfrom the diet or were synthesized in the animals after assimilation.Similar d13C values between diet and animals would suggest thatthe AA carbon skeletons were primarily derived from the diet,whereas differences would indicate de novo synthesis. It has notbeen previously documented that gut bacteria in Collembolaprovide AAs to their hosts. For this reason, we hypothesized thatthe d13C values of amino acids essential for insects would matchclosely between Collembola and their diets (Sang and King, 1961;Davis, 1975; Nation, 2002; O’Brien et al., 2002).

2. Material and methods

2.1. Experimental diets and animals

The three Collembolan species, F. candida, P. fimata and P. minutawere obtained from a stock of established laboratory cultures raisedon commercial freeze-dried S. cerevisiae (De Danske SpritfabrikkerA/S, DK) for several (>10) generations. During the experiment theanimals were raised on laboratory cultured fungi, S. cerevisiae asa high N reference diet and F. culmorum as a low N diet. F. culmorumwas cultured in a broth medium with a low concentration ofnitrogen to alter the C:N ratio: 1000 mg KH2PO4 (VWR), 1000 mgKNO3 (VWR), 500 mg MgSO4$7H2O (VWR), 500 mg KCl, 26.00 g D

(þ)-glucose, 2.4 g Potato Dextrose broth (Difco, Denmark), 1000 mldistilled water. We also cultured an isotopically enriched batch ofF. culmorum by making the following substitutions of KNO3 andglucose: 250 mg D-glucose-13C6 99 atom%, 25.75 g D(þ)-glucose.The growth medium for unlabelled S. cerevisiae was: 1.6 g yeastsynthetic media with AAs (FlukaeSigma Aldrich, Denmark), 5.0 g

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759 751

KNO3, 1.5 g (NH4)2SO4 (VWR, Denmark), 0.5 g KCL (VWR), 20.0 g D

(þ)-glucose (Sigma Aldrich) in 1000 ml distilled water. Thefollowing substitutions for glucose were made for the labelledS. cerevisiae medium: 0.250 g D-Glucose-13C6 99 atom% (SigmaAldrich), 19.750 g D(þ)-glucose. All ingredients were of laboratorygrade. After growth, the liquid cultures were filtered with Advantac5C filter paper, freeze-dried and stored at �18 �C until feeding. Wemeasured the elemental content and AA composition of the dietsand stock cultures of animals reared on commercial yeast prior tostarting the feeding experiment. The C content of the fungal dietsranged between 40.7 and 44.2% (Table 1). The N content was6.2e6.4% in the low N F. culmorum diet and 9.4e10.2% in the high NS. cerevisiae diet. In stock cultures of Collembola the N content wasbetween 9.8 and 11.2% indicating that the high N diet matched theN nutritional requirements of the animalsmore than the lowN diet.The P content was more than two-fold higher in the two diets(1.61e1.89%) than in the three Collembola species (0.62e0.77%). SeeTable 1 for elemental composition and bulk isotope values of thediets. To investigate the protein quality of the two diets prior to thediet change study we also measured their AA imbalance relative tothe stock culture Collembola. Throughout the paper we distinguishbetween essential and non-essential AAs according to the defini-tion by Nation (2002) for insects. The Euclidean distances for theessential AAs between animal tissue and diet were 4.8e6.8% for thelow N diet and 1.1e2.9% for the high N diet indicating thatthe overall AA composition of the high N diet was more similar tothe animals than those of the low N diet (Fig. 1, see also Table S1 inSupporting Information for composition of the AAs). For the low Ndiet, isoleucine, leucine andmethionine in particular were deficientin the diet relative to the three collembolans (15e60%) and for thehigh N diet methionine in particular was deficient (25e50%) (Fig.1).Amino acid N constituted between 48 and 77% of the total Ncontent in the studied animals (Table 1).

2.2. Diet change procedures

Eggs collected from laboratory animals were hatched and rearedfor two generations at 20 �C in Petri dishes (Ø ¼ 6 cm) with plasterof Paris (CaSO4) on either of the following four diets: 1) unlabelledS. cerevisiae, 2) labelled S. cerevisiae, 3) unlabelled F. culmorum and4) labelled F. culmorum. The animals were fed twice aweek and left-over food was removed. Incubation boxes and utensils for handlingwere kept separate throughout the experiment. The third genera-tion was synchronized to the age of 30e35 days and used for thediet isotope change experiment by transferring clusters of 40e80individuals to new Petri dishes (n¼ 6 for each diet change) or to tincapsules for analysis (n ¼ 2). The animals previously reared onunlabelled diets were fed labelled diets and vice versa. Two Petridishes from each treatment were subsequently taken out forelemental and stable isotope analysis 7,14 and 28 days after the diet

Table 1Stable isotopes values of diets used in the diet change experiment (n ¼ 3), elemental comcommercial S. cerevisiae (n ¼ 3) and the fraction amino acid N of DM and relative to totCollembola.

Species d13C d15N Bulk

%C

UD F. culmorum �7.5 � 0.1 2.0 � 0.5 42.7 � 0.2S. cerevisiae �7.0 � 0.1 �3.4 � 0.1 40.7 � 0.0

LD F. culmorum 348.6 � 0.7 368.6 � 8.0 42.7 � 0.6S. cerevisiae 360.4 � 0.3 702.3 � 2.8 44.2 � 1.0

SC F. candida 52.5 � 1.2P. fimata 53.2 � 0.6P. minuta 48.4 � 0.3

change. At each sampling day, food was removed from thesubstrates 6 h before sampling to ensure that the collembolansemptied their guts. Eggs laid during the incubation were removedfrom the substrates and discarded. The dry mass (DM) of individualanimals was determined by counting each batch of animals andplacing them into pre-weighed tin capsules before drying them at50 �C for 24 h and weighing them on a Sartorius MC210 microscale.

2.3. Analyses of elements, amino acids and isotopes

The dry weight of the samples for bulk 13C, C and N analysesranged between 0.15 and 0.80 mg and was determined at the UCDavis Stable Isotope Facility using a PDZ Europa ANCA-GSLelemental analyzer interfaced to a PDZ Europa 20e20 isotope ratiomass spectrometer (Sercon Ltd., Cheshire, UK). The working stan-dard for C was beet (Beta vulgaris L.) sucrose with a d13C valueof �23.83&, calibrated against NIST SRM 8539 and NIST SRM 8542standards. During isotope analysis at UC Davis 18 samples werelost due to clogging up of aluminum capsules in the elementalanalyzer. Elemental P analysiswas carried out inpre-weighed Tefloncapsules with 0.3 mg dry mass sample and determined by acid-persulphate digestion followed by phosphate analysis using theammonium molibdate method (Grasshoff et al., 1983). Thecomposition of AAs was determined on stock cultures of laboratoryanimals reared on commercial yeast and on unlabelled F. culmorumand S. cerevisiae. For the analysis we used 50 mg freeze-driedsamples and were analysed at Section for Microbiology, Universityof Aarhus following the procedure by Lomstein et al. (2006)with thefollowing modifications: L-amino butyric acid was used as aninternal standard and buffer A was modified. Buffer A used in thepresent studywas composed of 8.16 g sodiumacetate trihydrate and15ml tetrahydrofurane, to which Milli-Q water was added to a finalvolume of 2 l. The pH was adjusted to 6.7 with 6 M HCl. Isotopecomposition of the AAs was determined on 1e3 mg dried samples(n ¼ 3) at Alaska Stable Isotope facility following the proceduredescribed by Larsen et al. (2009a). Hydrolysed samples werederivatized with acidified isopropyl and N-trifluoroacetate, whichwere injected into an Agilent 6890N gas chromatograph (GC) withan autosampler and separated on an HP Ultra-1 column (Agilent).The GC was interfaced with a Finnigan Delta Plus XP isotope ratiomass spectrometer (IRMS) via the GC-III combustion (C) interface(Thermo-Finnigan, Waltham, Massachusetts, USA). We obtainedgood chromatographic peaks (Fig. 2) for ten AAs. We categorizedessential and non-essential AAs according to Nation (2002). Thenon-essential AAs we were able to analyse were (Fig. 2): alanine(Ala), asparagine/aspartic acid (Asx), glutamine/glutamic acid (Glx),glycine (Gly) and proline (Pro); and the essential AAs were:isoleucine (Ile), leucine (Leu), phenylalanine (Phe), threonine (Thr)and valine (Val). The samples were derivatized and analysedconcurrently with AA standards of known AA d13C. We used the

position relative to dry mass (DM) of diets and stock culture Collembola raised onal N in the SC animals. UD ¼ unlabelled diet, LD ¼ labelled diet, SC ¼ Stock culture

Amino acids

%N %P %N of DM %N of total

6.2 � 0.4 1.88 � 0.079.4 � 0.0 1.51 � 0.066.4 � 0.1 1.90 � 0.03

10.2 � 0.1 1.70 � 0.099.9 � 0.7 0.62 � 0.10 5.86 59.29.8 � 0.2 0.66 � 0.20 7.51 76.7

11.2 � 0.2 0.77 � 0.08 5.41 48.3

Fig. 1. Amino acid composition measured prior to the feeding experiment of the low(A: Fusarium culmorum) and high (B: Saccharomyces cerevisiae) nitrogen diets relativeto Collembola from stock cultures fed commercial S. cerevisiae.

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759752

known d13C values from the standards to calculate correction factorsspecific to each amino acid that account for carbon addition andfractionation during derivatization (Silfer et al., 1991; O’Brien et al.,2002). Reproducibility varied among AAs, with standard deviationsof three replicates ranging from �0.2 to �1.5& AA�1.

2.4. Calculations

The isotopic ratios are reported with units of per mil (&)difference with VPDB (13C/12C ¼ 0.011180) and atmospheric N

Fig. 2. GC-C-IRMS chromatogram of a Collembola sample. (A) Ratio of 45 (13C16O2) to44 (12C16O2) voltages over time. (B) Voltage over time of ion mass 44, the mostabundant CO2 ion form.

(15N/14N ¼ 0.0036765) as reference values. Isotopic difference inanimals relative to their diets is described as

3 ¼�1000þ dw1000þ dc

� 1�*1000 (1)

where dw and dc are the isotopic ratios of animal tissue and diet,respectively.

Isotopic incorporation into the animals followed an asymptoticexponential model akin to that used in previous diet change studies(Fry and Arnold, 1982; Larsen et al., 2009b)

dwðtÞ ¼ dn � ðdn � diÞ*e�ts (2)

where dn is the asymptotic isotope ratio approached after extendedgrowth, di is the isotope ratio of tissues before diet change and s isthe average retention time of a given element (del Rio andAnderson-Sprecher, 2008). The half-live of the elements, which isthe time period it takes to reach the midpoint of the initial andequilibriumvalues can be foundmultiplying s by Ln(2). The fractionof tissue that has the isotope value of the new diet after extendedgrowth is called the mixing fraction (b) and was calculated as

b ¼ dn � dide � di

(3)

where de is the isotope ratio of animals that lived on the old diet fortwo generations. We calculated the incorporation rate of newelements relative to the total body mass (T) (mixing and non-mixing fraction combined) as T ¼ b/s.

To assess the relative contributions of growth and tissuereplacement to the incorporation of elements into Collembola wefirst used a massebalance model (Frazer et al., 1997) in which theinitial elemental pool was diluted by adding newly assimilatedelements

dgðtÞ ¼ dc þ ðdi � dcÞ*ðmi=mtÞ (4)

where mi and mt are the dry mass of animals at the time of dietchange and at sampling (7, 14 and 28 days), respectively. Fromequation (4) it can be inferred that during extended growth dg willapproach an asymptotic value lower than dc. We estimated theasymptotic value and the rate by which new elements are incor-porated due to growth by fitting an asymptotic exponential model(similar to equation (2)) to the dg values. Next we modelled therelative contribution of tissue replacement (dr) simply by sub-tracting the isotopic dilution due to growth (dg) from total isotopicisotopic change (dw). We subsequently fitted dr as a function of timeto an asymptotic exponential model similar to equation (2).

The AA imbalance in tissue relative to diet was calculatedaccording to Conceicao et al. (2003)

rAAj ¼AAjc � AAjw

AAjw(5)

where AAj is a given AA in percentage of total AAs (w/w) with thesubscripts c andw representing diet and animal tissue, respectively.

Euclidean distance (dejk) was used to determine the overalldistance in the relative AA composition between consumers andtheir diets

deij ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPn

i¼1ðXi � YiÞ2n

s(6)

where X and Y are the relative abundances of the AAi of theconsumer and the diet respectively, and n is the number of AAs. Ahigher overall AA difference between animals and diets is signified

Table 2The effects of the S. cerevisiae and F. culmorum diets on elemental composition andratios in unlabelled and labelled treatments (n ¼ 3e4) tested with two-way ANOVAwith species nested within diet. “SS” signifies sum of squares, “df” degrees offreedom and “F” the F ratio, “CI” for confidence intervals (ns; not significant). SeeFig. 3 for post-hoc tests on elemental composition and ratios.

Element Effect SS df F CI (%)

%C Diet 78.1 1 139.3 �99.99Diet:Species 21.3 4 9.51 �99.99Residuals 11.2 20

% N Diet 34.2 1 69.7 �99.99Diet:Species 4.54 4 2.32 nsResiduals 9.80 20

%P Diet 0.139 1 73.1 �99.99Diet:Species 0.171 4 22.5 �99.99Residuals 0.0323 17

C:N Diet 44.1 1 72.4 �99.99Diet:Species 2.24 4 3.68 �95Residuals 12.2 20

C:P Diet 16000 1 97.3 �99.99Diet:Species 17100 4 26.1 �99.99Residuals 2790 17

N:P Diet 27.5 1 14.8 �99Diet:Species 45.6 4 6.14 �99Residuals 31.6 17

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759 753

by a higher Euclidean distance (Legendre and Legendre, 1998). Fordetermining the AA d13C dissimilarities between consumers andtheir diets we calculated the absolute distance (dajk)

daij ¼Pn

i¼1 jXi � Yijn

(7)

where X and Y are the individual AA d13C values of the consumerand the diet, respectively.

2.5. Statistical analyses

All statistical analyses and modelling were performed using R,version 2.9.1 (R-Development-Core-Team, 2009). All treatmentswere tested for variance homogeneity before applying ANOVA orstudent t-test. All comparison tests between species in terms ofelemental content and ratios performed with two-way nestedANOVA nesting species within diet treatments. Non-linear func-tions were fitted by non-linear least squares minimization andlinear functions were fitted by linear modelling. Lack-of-fit wastested by comparing the functions with general ANOVA modelsusing a likelihood ratio test. To assess whether functions from twotreatments were significantly different, we first normalized theirdatasets to an uniform scale and tested the residual sum of squaresof individual functions against the residual sum of squares ofcombined functions using Fischer’s F test (Motulsky and Ransnas,1987). To investigate whether any variations in d13C patterns ofessential AAs between Collembola and their fungal diets wereindicative of de novo synthesis in Collembola we applied thediscriminant method described by Larsen et al. (2009a) for cate-gorizing the origin of essential AAs in unknown samples to eitherfungi, bacteria and plants based on their d13C patterns. The lineardiscriminant function of the R package “MASS” (Venables andRipley, 2002) was used for the linear discriminant analysis, the Rpackage “cluster” was used for computing pairwise dissimilarities(distances) for AAs (composition or d13C) between Collembola andtheir diets, and the Satterthwaite approximationwas used to derivestandard errors of pooled samples.

3. Results

3.1. Elemental composition and isotopic spacing

The two diets affected elemental composition and ratios of allthe three species significantly (P < 0.01, Table 2). The interactionbetween diet and species was significant for all elements and ratios(P < 0.01) except that N was not significant (Table 2). This indicatesthat the Collembola species reacted differently to the two dietsexcept in terms of N content that all reacted similarly by decreasingtheir N content from 9.3e10.1% N on the high N diet to 6.5e7.7% Non the lowN diet (Fig. 3). It was consistent that all three Collembolaspecies had the highest %C and the lowest %N and %P on the low Ndiet but with F. candida having the largest differences between thetwo diets in terms of C and P (P < 0.05). F. candida also had thelargest differences in C:N and C:P ratios between the two dietscompared to P. fimata and P. minuta (P < 0.05). However, N:P wasequal between the two diet treatments for F. candida but not forP. fimata and P. minuta for which the ratios were lowest on the lowN diet (P < 0.05). Carbon relative to N and P was highest on the lowN diet (P < 0.05).

The isotope values of Collembola fed the labelled diets for twogenerations varied considerably between treatments and species(see Fig. S2-A and S2-B). The three Collembola species wereenriched in 13C relative to the low N diet (between 3 and 27& 313C)but depleted relative to the high N diet (between �28 and �10&

313C) with F. candida having the largest offset values relative to thediets (see Fig. S2-A). In terms of 15N, Collembola were greatlydepleted relative to the high N diet (between �66 and �47& 315N)with P. minuta having the largest offset between tissue and diet (seeFig. S2-B). P. minutawas also greatly depleted relative to the low Ndiet (�60& 315N). F. candida and P. fimata were almost identical tothe low N diet with offset values <5& 315N. The different isotopeoffsets between Collembola and their diets indicate that thelabelled fungi were not labelled homogenously and that each Col-lembola species incorporated the unlabelled and labelled elementsdifferently.

3.2. Growth and elemental turnover

All three Collembola species reached a significantly higher bodymass on the high than low N diet at the first (onset of sexualmaturity) and last sampling occasions (28 days later) (P � 0.01%,Table 3). The adult growth rates (% d�1) were likewise significantlylower on the low than high N diet for F. candida and P. minuta butsignificantly higher for P. fimata (Table 3). However, the dailygrowth increments (mg d�1) for adult P. fimata were significantlyhigher on the high than low N diet due to the differences in initialbodymass at the first sampling occasion (Table 3). The daily growthincrements during the 28 days after diet change were for F. candidaon average 2.5 times higher on the high than low N diet, forP. fimata 1.5 times higher and for P. minuta 8 times higher (Table 3).

The incorporation rates of C and N were higher on the high thanlow N diet for F. candida and P. minuta but not for P. fimata wherethe rates were in a similar range on both diets (Table 4). On the lowN diet the C incorporation rates were 6.4e7.7% C d�1 for F. candida,11.9% C d�1 for P. fimata, and 5.8% C d�1 for P. minuta. The corre-sponding rates on the high N diet were 13.7e14.8% C d�1 forF. candida, 10.7% C d�1 for P. fimata, and 16.3% C d�1 for P. minuta.The C and N incorporation rates for P. minutawere not significantlydifferent from each other on both diets and P. fimata on the low Ndiet (Table 4). However, the incorporation rate of N compared to Cwas significantly lower (P > 0.001) on the high N diet for P. fimata(8.6 vs. 10.7% N d�1) and significantly higher for F. candida (7.0e8.1vs. 6.4e7.7% N d�1). The incorporation rates of C due to growthranged between 2.3 and 3.6% C d�1 on the low N diet and 4.1 and5.4% C d�1 on the high N diet (Table 4). The values for N were in

0

10

20

30

40

50

60ab c b c b

%C

diet

0

2

4

6

8

10

12

a

ba

b

a

b

%N

diet

0.0

0.5

1.0

1.5

2.0

ab b c

b b

%P

diet

F. candida P. fimata P. minuta

0

2

4

6

8

10

12 a

b

c

b

c

b

C:N

diet

0

50

100

150

200

250

300 a

b cb

cb

C:P

diet

0

10

20

30

40

a aba c

b

N:P

diet

F. candida P. fimata P. minuta

F. culmorum S. cerevisiae

Fig. 3. Elemental composition and ratios (by atoms) of Collembola raised for two generations on either the low (Fusarium culmorum) or high (Saccharomyces cerevisiae) nitrogen diet(n ¼ 4e8). Different letters signify significant differences (ANOVA, P < 0.05).

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759754

a similar range. The incorporation rates due to growth for bothF. candida and P. minuta were 1.5e2 times larger on the high thanlow N diet. For P. fimata the rates were slightly higher on the highthan low N diet (4.1 vs. 3.6% C d�1). The fraction of growth relativeto total incorporation ranged between 27 and 41% for C and 24 and45% for N with the growth fractions being higher on the low thanhigh N diet (C: 6e8 points and N: 5e10 points).

We estimated the incorporation rates of C due to tissue turnoverto 3.6e8.8% C d�1 on the lowN diet and 7.0e12.0% C d�1 on the highN diet (Table 4). For N the rates were similar, 4.3e6.9% N d�1 on thelow quality diet and 5.0e14.0% N d�1 on the high N diet. The leastdifference between diet treatments was found for P. fimata with

Table 3Collembolan dry mass at the first (M0, 30e35 days old) and the final (M28, 28 days afincrements and rates under the two diets, F. culmorum (Fc) and S. cerevisiae (Sc). The sighighest.

Collembola M0 (mg) M28 (mg)

Fc Sc Fc

F. candida 11.0 � 0.8 22.3 � 2.0x 14.4 � 0.4P. fimata 15.6 � 1.3 31.5 � 2.2x 23.2 � 0.9P. minuta 2.8 � 0.2 8.2 � 0.6x 3.5 � 0.2

almost equal tissue replacement rates being the most pronounceddifference for P. minutawith 3 times lower tissue turnover rates onthe low than high N diet.

3.3. Stable isotope composition of individual amino acids

The d13C dissimilarities between animals and their diets weregreater for the non-essential AAs than the essential AAs (Fig. 4). Onthe high N diet, the mean d13C dissimilarities for all three Collem-bola species were between 1.2 and 2.1& for the essential AAs andbetween 5.5 and 6.7& for the non-essential AAs. The mean d13Cdissimilarities between the essential AAs were more pronounced

ter diet change) sampling occasions (mean � sd, n ¼ 3e4) and their daily growthnificance codes (y�5%; z�1%; x�0.1%) are placed where the values are significantly

Growth (mg d�1) Growth (% d�1)

Sc Fc Sc Fc Sc

32.8 � 1.3x 0.12 0.38x 1.11 1.69y

42.9 � 1.9x 0.27 0.40z 1.74x 1.2813.6 � 0.5x 0.024 0.19x 0.87 2.34x

Table 4Parameters of isotopic change in Collembola (A) partitioned into growth (B) and tissue replacement (C) with residence times (s), mixing fractions (b) and change relative tototal body mass (T) (means � SE, n ¼ 6e8 for each fit). Abbreviations: U; unlabelled diet, L; labelled diet, CI; Confidence intervals, ns; not significant.

Species Diet Diet change d13C d15N CI (%)

s (days) b (% of total DM) T (% DM d�1) s (days) b (% of total DM) T (% DM d�1) d13C vs. d15N

A F. candida F. culmorum U / L 10.5 � 1.2 81.0 � 4.1 7.7 � 0.9 11.5 � 1.0 93.2 � 3.6 8.1 � 0.7 �99.99a

A L / U 12.1 � 1.4 77.4 � 3.6 6.4 � 0.7 13.0 � 1.4 91.5 � 4.3 7.0 � 0.8A S. cerevisiae U / L 6.4 � 0.5 86.8 � 2.0 13.7 � 1.1 6.6 � 0.5 87.0 � 2.1 13.1 � 1.0 �99.99a

A L / U 6.2 � 0.2 92.1 � 0.8 14.8 � 0.5 5.4 � 0.2 91.4 � 0.7 17.0 � 0.5A P. fimata F. culmorum L / U 6.2 � 1.1 73.5 � 3.6 11.9 � 2.1 7.2 � 1.0 72.6 � 3.0 10.2 � 1.3 nsA S. cerevisiae U / L 5.8 � 0.4 62.2 � 1.1 10.7 � 0.7 9.2 � 0.7 78.6 � 2.0 8.6 � 0.6 �99.99A P. minuta F. culmorum L / U 14.4 � 3.6 83.9 � 10.4 5.8 � 1.5 13.4 � 4.2 86.7 � 13.0 6.5 � 2.0 nsA S. cerevisiae U / L 5.5 � 0.3 89.3 � 1.8 16.3 � 1.0 5.4 � 0.4 90.8 � 2.1 16.9 � 1.3 ns

B F. candida F. culmorum U / L 10.7 � 6.3 28.8 � 8.4 2.7 � 1.5 10.7 � 6.3 26.6 � 8.4 2.5 � 1.4B L / U 9.1 � 4.4 23.7 � 5.4 2.6 � 1.2 9.1 � 4.4 23.7 � 5.4 2.6 � 1.2B S. cerevisiae U / L 6.5 � 2.8 32.8 � 4.4 5.0 � 1.9 6.5 � 2.9 30.8 � 4.4 4.7 � 1.9B L / U 8.6 � 3.8 33.9 � 5.0 4.0 � 1.6 8.6 � 3.8 33.9 � 5.0 4.0 � 1.6B P. fimata F. culmorum L / U 10.1 � 2.8 36.4 � 8.5 3.6 � 3.3 10.1 � 5.3 36.4 � 8.5 3.6 � 3.3B S. cerevisiae U / L 6.9 � 2.2 27.8 � 2.5 4.1 � 1.7 6.9 � 2.2 26.7 � 2.5 3.9 � 1.7B P. minuta F. culmorum L / U 8.9 � 3.6 20.2 � 3.0 2.3 � 0.9 8.9 � 3.6 20.2 � 3.0 2.3 � 0.9B S. cerevisiae U / L 7.9 � 1.6 42.5 � 5.3 5.4 � 2.9 7.9 � 1.6 40.9 � 5.3 5.2 � 2.7

C F. candida F. culmorum U / L 10.3 54.1 5.2 11.8 66.3 5.6C L / U 14.1 52.8 3.7 15.4 69.6 4.5C S. cerevisiae U / L 6.3 55.8 8.9 6.7 55.6 8.3C L / U 5.3 56.5 10.8 4.2 58.5 14.0C P. fimata F. culmorum L / U 4.1 36.8 8.8 5.3 36.7 6.9C S. cerevisiae U / L 5.2 36.2 7.0 10.7 53.7 5.0C P. minuta F. culmorum L / U 18.0 64.2 3.6 15.8 67.5 4.3C S. cerevisiae U / L 4.1 49.8 12.0 4.0 51.4 12.7

a For F. candida the U / L and L / U treatments were normalized and pooled before comparing elements.

Fig. 4. Isotopic spacing (3) of individual amino acids in Collembola relative to their low(A: Fusarium culmorum) and high (B: Saccharomyces cerevisiae) nitrogen diets.

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759 755

on the low N diet (between 2.1 and 3.5&), and similar for the non-essential AAs (between 5.6 and 7.0&). In terms of individualessential AAs, Ile, Leu, Phe and Val were enriched or similar in d13Crelative to both diets (Fig. 4, F. culmorum 313C < 6.2&, S. cerevisiae313C < 3.0&). The effect of diets was contrasting for Thr; especiallyfor the high N diet where d13C in Thr was depleted in F. candida(�2.5&) and enriched in P. fimata (3.4&) and P. minuta (3.6&).To investigate whether variations in d13C patterns between Col-lembola and their diets were indicative of de novo synthesis by gutbacteria we used the AA 13C library by Larsen et al. (2009a) that canbe used to predict whether essential AAs in unknown samplesoriginate from either fungi, bacteria and plants. Each of these taxaimparts characteristic and unique patterns of d13C variation duringsynthesis of amino acids. The linear discriminant model based onthe 33 samples from Larsen et al. (2009a) classified the two fungaldiets and the three Collembola species on the high N diet as fungiwith 100% probability (Fig. 5). On the low N diet, P. fimata andF. candida classified with 99.9 and 100% probability as fungi, andP. minuta as 95.6% probability as fungus and 4.4% probability asbacteria.

4. Discussion

4.1. Stoichiometric composition

The elemental and amino acid N differences of the two dietsaffected the physiology of all three Collembola species. Notably,body content of N in Collembola tracked composition of their dietsdecreasing from around 10% N on the high N diet to around 7% N onthe low N diet, which is below the 8e13.7% N reported in otherstudies (Teuben and Verhoef, 1992; Lavy and Verhoef, 1996b;Jensen et al., 2006; Larsen et al., 2009b). The high C:N ratios forall three species on the low N diet point to that their storagecompounds such as lipids and fatty acids increased relative to theircontent of proteins, nucleic acids, chitin and nucleotides (Ventura,2006). This form of nutritional compensation is known to occur

Fig. 5. Predicting classification of amino acid d13C from Collembola (Folsomia candida,Protaphorura fimata and Proisotoma minuta) and their low (Fc: Fusarium culmorum)and high (Sc: Saccharomyces cerevisiae) nitrogen diets based on the d13C amino acidlibrary by Larsen et al. (2009a) of 33 fungi (n ¼ 13), bacteria (n ¼ 10) and plant(n ¼ 10) samples (filled circles) with the following five essential amino acids:isoleucine, leucine, phenylalanine, threonine and valine. The dotted lines representconfidence ranges at P ¼ 0.5, and the symbols for Collembola were tagged either withFc or Sc to indicate the animal’s diet. The two diets were not tagged due to their closeproximity. See Table S3 in Supporting Information for the linear discriminantcoefficients.

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759756

in several insects (Sterner and Elser, 2002) and Collembola feedingon lowN diets (Lavy and Verhoef, 1996a; Haubert et al., 2004; Ruesset al., 2004). The animals compensate for low levels of dietary N byconsuming more food to maintain a relatively constant amount ofproteins and as a consequence accumulate more fat. However, it isalso possible that the collembolans not only compensated byaccumulating more fat but also decreased their amount of proteins,nucleic acids, chitin or nucleotides. The variations in N:P ratiosbetween Collembola on the low N diet does not unequivocallysupport the growth rate hypothesis, which states that differences inorganismal C:N:P ratios are caused by differential allocations toRNA necessary to meet protein synthesis demands for growth andmetabolic turnover (Sterner and Elser, 2002). F. candidamaintainedsimilar N:P ratios on both diets while significantly decreasing itsgrowth rate and tissue change rate. In contrast, P. fimata decreasedits N:P ratio on the F. culmorum diet while maintaining similartissue change rates on either diet. These contrasting responsesamong the three species during nutritional stress could beexplained by different metabolic expenditures associated withcompensatory feeding, digestion and reproduction. The changes inbody C:N:P in conjunction with the effects on growth rates andtissue replacement indicate that animals were nutritionallystressed on the low N diet, and that the animals were not able tomake up for elemental and biochemical N imbalances of the diet bycompensatory feeding. However, the study showed that the threeCollembola species have the ability to adapt to nutrient poor dietsby regulating their stoichiometry and still maintain basic physio-logical functions such as reproduction and growth, albeit at lowerrates. Ecophysiological selection of metabolic pathways adapted tothe large fluctuations in food quality and availability in the soilenvironment may explain why Collembola appear to be among theleast homeostatic animals reported in the literature (Sterner andElser, 2002; Acquisti et al., 2009).

4.2. Effect of diets on growth and incorporation of carbonand nitrogen

The diets affected C and N turnover in Collembola as their ratesdecreased from 9e17% DM d�1 on the high N diet to 6e12% DM d�1

on the low N diet. P. minuta and F. candida were the species mostaffected by the low N diet and we observed almost no effect forP. fimata in the turnover data. This confirms our hypothesis thatF. candida and P. minuta would decrease the rates of tissuereplacement relatively more than P. fimata on the low N diet.However, Collembola might react differently to dietary imbalancesdepending on their life stage, as juvenile P. fimata grew to half thesize on the low compared to the high N diet. This response to dietmay be caused by high nutritional requirements or low capacity fordigestion and assimilation during the juvenile phase. The residencetimes of C and N of animals fed the high N diet were in a similarrange to those found by Chamberlain et al. (2004) for F. candida andby Larsen et al. (2009b) for P. minuta and P. fimata. The totalincorporation rates for P. minuta in our study were somewhathigher than in the study by Larsen et al. (2009b) (16e17 vs. 13e14%DM d�1), which was most likely due to the faster growth rates ofP. minuta in this study.

We partitioned incorporation of C and N to growth and tissuereplacement and found that growth accounted for less than half oftotal elemental incorporation for both diet treatments (24e45%)with allocation to growth being higher on the low than high N dietfor all three species. For F. candida and P. minuta the tissuereplacement rates were much higher (two- to three-fold) on thehigh than low N diet and we speculate that this is linked to a rela-tively lower fecundity on the low N diet. We did not monitoroviposition but nutrient limitation has been found to decrease

fecundity in other taxa such as nematodes (Klass, 1977) and fruitflies (Grandison et al., 2009). A lower fecundity would lowermetabolic costs associated with egg manufacturing as well as allo-cation fromassimilated nutrients to eggs. That Collembola allocatedmore C and N to tissue replacement than growth is consistent witha previous study on P. fimata and P. minuta (Larsen et al., 2009b) andtwo studies on cold water ectotherms, the Atlantic salmon (Salmosalar) (Jardine et al., 2004) and the Artic amphipod (Onisimus litor-alis) (Kaufman et al., 2008) but contrastswithfindings forfish larvae(Herzka and Holt, 2000), whitefish (Hesslein et al., 1993), goby(Maruyama et al., 2001) and zooplankton (Ventura and Catalan,2008). Similar ratios for tissue turnover relative to growth have toour knowledge not yet been published for other soil animal detri-tivores, but Ostle et al. (2007) found that Collembola incorporatedrhizosphere carbon at a much higher rate than mites, enchytraeidsand earthworms, and the authors ascribed this to feeding prefer-ences as well high metabolic rates for Collembola. Our study showsthat Collembola have relatively high rates ofmetabolism, i.e. C andNreplacement even on a low quality diet, which underlines thatCollembola in their natural environment play a proportionallylarger role for decomposition processes thanwhat can be attributedby their relatively modest biomass.

4.3. Amino acid imbalances

In soil, essential AAs in collembolan diets may be limited inperiods with low microbial activity. To investigate the extent thatCollembola may rely on endogenous sources to compensate forscarcity of essential AAs, we compared individual AA d13C valuesbetween Collembola and two diets to evaluatemetabolic reworkingof AAs in the animals after assimilation. For both diets, the Dd13C ofAAs between diet and animals was much larger for the non-essential than the essential AAs demonstrating a high degree of denovo synthesis for the non-essentials, which agrees with results

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759 757

from other animal studies (Edgar Hare et al., 1991; Howland et al.,2003; O’Brien et al., 2003, 2005; McCullagh et al., 2008; McMahonet al., 2010). However, contrary to our hypothesis essential AAs d13Cdid not always match perfectly between the diet and Collembolaspecies, and this mismatch was greater for Collembola on the lowNdiet. Relative to the low N diet, Ile, Leu, and Phe were enriched incollembolan tissues between 2.7 and 6.2& (Fig. 5). In contrast, forCollembola on the high N diet essential AAs d13C matched those ofthe diet, with the exception of Thr (Fig. 5). This suggests thepossibility of endogenous production of essential AAs, particularlyon the unbalanced diet. Symbiotic bacteria in the gut are the mostlikely provider of these essential AAs as indicated by the slight shiftof essential AA d13C values towards a bacterial pattern on theimbalanced diet (Fig. 5). Provision of AAs from gut bacteria has notyet been documented in Collembola but it is well known that thediversity of obligate symbiotic bacteria in Collembola is large(Thimm et al., 1998; Berg et al., 2004; Czarnetzki and Tebbe, 2004)and that bacteria play a major role for enhancing degradation ofhardly digestible molecular compounds such as chitin (Borkott andInsam, 1990). Thus, it is possible that digestion of gut bacteria orbacterial exudates can be a minor source of essential AAs for theanimals but it cannot ruled out either that bacterial infection of thediets after feeding accounted for the isotopic mismatch. Sincethe isotopic offsets for the essential AAs between the three specieson the imbalanced AA diet were very similar, it suggests thatsymbiotic production of essential AAs do not account for the higherrates of growth and metabolism by P. fimata than F. candida andP. minuta. The reasons for the isotopic fractionation of essential AAsbetween the animals and their diets are unclear and need to beexplored further.

5. Conclusions

The low N diet affected the physiology of all three Collembolaspecies. Notably, body content of N tracked composition of theirdiets decreasing from around 10% N on the high N diet to around7% N on the low N diet along with a significant increase of N:P fortwo of the species. P. fimata was less affected than F. candida andP. minuta in terms of growth and C and N turnover on the low Ndiet. Essential AAs d13C did not always match perfectly betweenthe diet and Collembola species, and this mismatch was greaterfor Collembola on the imbalanced AA diet with the low N content.Our classification of d13C AA patterns according to the uniqueisotope patterns created by fungi, bacteria and plants during AAsynthesis indicates that bacteria may have been the alternativesource of essential AAs. However, the reasons for the isotopicfractionation of essential AAs between the animals and their dietsare unclear and need to be explored further. Our physiologicaldata point to that P. fimata in particular is evolutionary adapted towithstand greater stoichiometric changes in tissue with thetrade-off that it has a lower maximum metabolic rate on a highquality diet than the two other species. While the results of thisstudy cannot be extrapolated directly to the dynamics of Col-lembola populations in the field, they serve to demonstrate thatthey are very flexible in terms of adapting physiologically to thetemporal and spatial patchiness of the soil environment. More-over, the rates of C and N replacement on the low and high Ndiets are useful for assessing the quantitative contribution ofCollembola to soil decomposition processes under differentscenarios.

Acknowledgements

This study was conducted with support from DARCOF (DanishResearch Centre for Organic Farming). Partial support to TL and MV

was also provided by the Spanish Ministry of Science and Innova-tion (Ref. CGL2010-14841/BOS). We thank P.H. Krogh for advice andthe National Environmental Research Institute in Silkeborg forproviding the facilities for the study.

Appendix. Supplementary data

Supplementary data related to this article can be found online atdoi:10.1016/j.soilbio.2010.12.008.

References

Acquisti, C., Kumar, S., Elser, J.J., 2009. Signatures of nitrogen limitation in theelemental composition of the proteins involved in the metabolic apparatus.Proceedings of the Royal Society B-Biological Sciences 276, 2605e2610.

Anderson, R.V., Gould, W.D., Woods, L.E., Cambardella, C., Ingham, R.E.,Coleman, D.C., 1983. Organic and inorganic nitrogenous losses by microbivorousnematodes in soil. Oikos 40, 75e80.

Anderson, T.R., Boersma, M., Raubenheimer, D., 2004. Stoichiometry: linkingelements to biochemicals. Ecology 85, 1193e1202.

Bengtsson, G., Ohlsson, L., Rundgren, S., 1985. Influence of fungi on growth andsurvival of Onychinrus armatus (Collembola) in a metal polluted soil. Oecologia68, 63e68.

Berg, M.P., Stoffer, M., van den Heuvel, H.H., 2004. Feeding guilds in Collembolabased on digestive enzymes. Pedobiologia 48, 589e601.

Boëchat, I.G., Adrian, R., 2006. Evidence for biochemical limitation of populationgrowth and reproduction of the rotifer Keratella quadrata fed with freshwaterprotists. Journal of Plankton Research 28, 1027e1038.

Booth, R.G., Anderson, J.M., 1979. The influence of fungal food quality on the growthand fecundity of Folsomia candida (Collembola: Isotomidae). Oecologia 38,317e323.

Borkott, H., Insam, H., 1990. Symbiosis with bacteria enhances the use of chitin bythe springtail, Folsomia candida (Collembola). Biology and Fertility of Soils 9,126e129.

Bowen, S.H., 1984. Detrital amino-acids and the growth of Sarotherodon mossam-bicus e a reply. Acta Hydrochimica Et Hydrobiologica 12, 55e59.

Chamberlain, P.M., Bull, I.D., Black, H.I.J., Ineson, P., Evershed, R.P., 2004. Lipidcontent and carbon assimilation in Collembola: implications for the use ofcompound-specific carbon isotope analysis in animal dietary studies. Oecologia139, 325e335.

Chamberlain, P.M., Bull, I.D., Black, H.I.J., Ineson, P., Evershed, R.P., 2005. Fatty acidcomposition and change in Collembola fed differing diets: identification oftrophic biomarkers. Soil Biology and Biochemistry 37, 1608e1624.

Christiansen, K., 1964. Bionomics of Collembola. In: Smith, R.F. (Ed.), Annual Reviewof Entomology. Annual Reviews, Inc., Paulo, Alto, pp. 147e178.

Coleman, D.C., 2008. From peds to paradoxes: linkages between soil biota and theirinfluences on ecological processes. Soil Biology and Biochemistry 40, 271e289.

Conceicao, L.E.C., Grasdalen, H., Ronnestad, I., 2003. Amino acid requirements of fishlarvae and post-larvae: new tools and recent findings. Aquaculture 227,221e232.

Cowgill, U.M., Emmel, H.W., Hopkins, D.L., Takahashi, I.T., Parker, W.M., 1986.Variation in chemical composition, reproductive success and body weight ofDaphnia magna in relation to diet. Internationale Revue der gesamten Hydro-biologie und Hydrographie 71, 79e99.

Cross, W.F., Benstead, J.P., Rosemond, A.D., Wallace, J.B., 2003. Consumer-resourcestoichiometry in detritus-based streams. Ecology Letters 6, 721e732.

Czarnetzki, A.B., Tebbe, C.C., 2004. Diversity of bacteria associated with Collem-bola e a cultivation-independent survey based on PCR-amplified 16S rRNAgenes. FEMS Microbiology Ecology 49, 217e227.

Davis, G.R.F., 1975. Essential dietary amino-acids for growth of larvae of yellowmealworm, Tenebrio molitor L. Journal of Nutrition 105, 1071e1075.

de Ruiter, P.C., Neutel, A.M., Moore, J.C., 1995. Energetics, patterns of interactionstrengths, and stability in real ecosystems. Science 269, 1257e1260.

del Rio, C.M., Anderson-Sprecher, R., 2008. Beyond the reaction progress variable:the meaning and significance of isotopic incorporation data. Oecologia 156,765e772.

Edgar Hare, P., Fogel, M.L., Stafford, T.W., Mitchell, A.D., Hoering, T.C., 1991. Theisotopic composition of carbon and nitrogen in individual amino acids isolatedfrommodern and fossil proteins. Journal of Archaeological Science 18, 277e292.

Elser, J.J., Fagan, W.F., Denno, R.F., Dobberfuhl, D.R., Folarin, A., Huberty, A.,Interlandi, S., Kilham, S.S., McCauley, E., Schulz, K.L., Siemann, E.H.,Sterner, R.W., 2000a. Nutritional constraints in terrestrial and freshwater foodwebs. Nature 408, 578e580.

Elser, J.J., Sterner, R.W., Gorokhova, E., Fagan, W.F., Markow, T.A., Cotner, J.B.,Harrison, J.F., Hobbie, S.E., Odell, G.M., Weider, L.J., 2000b. Biological stoichio-metry from genes to ecosystems. Ecology Letters 3, 540e550.

Elser, J.J., Watts, T., Bitler, B., Markow, T.A., 2006. Ontogenetic coupling of growthrate with RNA and P contents in five species of Drosophila. Functional Ecology20, 846e856.

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759758

Endlweber, K., Ruess, L., Scheu, S., 2009. Collembola switch diet in presence of plantroots thereby functioning as herbivores. Soil Biology and Biochemistry 41,1151e1154.

Evans, C.J., Evershed, R.P., Black, H.I.J., Ineson, P., 2003. Compound-specific stableisotope analysis of soil mesofauna using thermally assisted hydrolysis andmethylation for ecological investigations. Analytical Chemistry 75, 6056e6062.

Færøvig, P.J., Hessen, D.O., 2003. Allocation strategies in crustacean stoichiometry:the potential role of phosphorus in the limitation of reproduction. FreshwaterBiology 48, 1782e1792.

Fagan, W.F., Denno, R.F., 2004. Stoichiometry of actual vs. potential predator-preyinteractions: insights into nitrogen limitation for arthropod predators. EcologyLetters 7, 876e883.

Fink, P., Von Elert, E., 2006. Physiological responses to stoichiometric constraints:nutrient limitation and compensatory feeding in a freshwater snail. Oikos 115,484e494.

Frazer, T.K., Ross, R.M., Quetin, L.B., Montoya, J.P., 1997. Turnover of carbon andnitrogen during growth of larval krill, Euphausia superba Dana: a stable isotopeapproach. Journal of Experimental Marine Biology and Ecology 212, 259e275.

Fry, B., Arnold, C., 1982. Rapid 13C/12C turnover during growth of brown shrimp(Penaeus aztecus). Oecologia 54, 200e204.

Grandison, R.C., Piper, M.D.W., Partridge, L., 2009. Amino-acid imbalance explainsextension of lifespan by dietary restriction in Drosophila. Nature 462,1061e1064.

Grasshoff, K., Ehrhardt, M., Kremling, K., Almgren, T., 1983. Methods of SeawaterAnalysis, second rev. and extended ed. Verlag Chemie, Weinheim, xxviii, 419 pp.

Guisande, C., Maneiro, I., Riveiro, I., 1999. Homeostasis in the essential amino acidcomposition of the marine copepod Euterpina acutifrons. Limnology andOceanography 44, 691e696.

Harasyme, L., Sinha, R.N., 1974. Survival of springtails Hypogastrura tullbergi andProisotoma minuta on fungal and bacterial diets. Environmental Entomology 3,965e968.

Haubert, D., Häggblom, M.M., Scheu, S., Ruess, L., 2004. Effects of fungal food qualityand starvation on the fatty acid composition of Protaphorura fimata (Collem-bola). Comparative Biochemistry and Physiology B e Biochemistry & MolecularBiology 138, 41e52.

Herzka, S.Z., Holt, G.J., 2000. Changes in isotopic composition of red drum (Sciae-nops ocellatus) larvae in response to dietary shifts: potential applications tosettlement studies. Canadian Journal of Fisheries and Aquatic Sciences 57,137e147.

Hesslein, R.H., Hallard, K.A., Ramlal, P., 1993. Replacement of sulfur, carbon, andnitrogen in tissue of growing broad whitefish (Coregonus nasus) in response toa change in diet traced by d34S, d13C and d15N. Canadian Journal of Fisheries andAquatic Sciences 50, 2071e2076.

Hogervorst, R.F., Dijkhuis, M.A.J., van der Schaar, M.A., Berg, M.P., Verhoef, H.A.,2003. Indications for the tracking of elevated nitrogen levels through the fungalroute in a soil food web. Environmental Pollution 126, 257e266.

Howland, M.R., Corr, L.T., Young, S.M.M., Jones, V., Jim, S., Van der Merwe, N.J.,Mitchell, A.D., Evershed, R.P., 2003. Expression of the dietary isotope signal inthe compound-specific d13C values of pig bone lipids and amino acids. Inter-national Journal of Osteoarchaeology 13, 54e65.

Ingham, E.R., Trofymow, J.A., Ames, R.N., Hunt, H.W., Morley, C.R., Moore, J.C.,Coleman, D.C., 1986. Trophic interactions and nitrogen cycling in a semiaridgrassland soil. 2. System responses to removal of different groups of soilmicrobes or fauna. Journal of Applied Ecology 23, 615e630.

Jardine, T.D., MacLatchy, D.L., Fairchild, W.L., Cunjak, R.A., Brown, S.B., 2004. Rapidcarbon turnover during growth of Atlantic salmon (Salmo salar) smolts in seawater, and evidence for reduced food consumption by growth-stunts. Hydro-biologia 527, 63e75.

Jensen, T.C., Leinaas, H.P., Hessen, D.O., 2006. Age-dependent shift in response tofood element composition in Collembola: contrasting effects of dietarynitrogen. Oecologia 149, 583e592.

Jørgensen, H.B., Elmholt, S., Petersen, H., 2003. Collembolan dietary specialisationon soil grown fungi. Biology and Fertility of Soils 39, 9e15.

Kaufman, M.R., Gradinger, R.R., Bluhm, B.A., O’Brien, D.M., 2008. Using stableisotopes to assess carbon and nitrogen turnover in the Arctic sympagicamphipod Onisimus litoralis. Oecologia 158, 11e22.

Kay, A.D., Ashton, I.W., Gorokhova, E., Kerkhoff, A.J., Liess, A., Litchman, E., 2005.Toward a stoichiometric framework for evolutionary biology. Oikos 109, 6e17.

Kitchell, J.F., O’Neill, R.V., Webb, D., Gallepp, G.W., Bartell, S.M., Koonce, J.F.,Ausmus, B.S., 1979. Consumer regulation of nutrient cycling. Bioscience 29,28e34.

Klass, M.R., 1977. Aging in nematode Caenorhabditis elegans e major biological andenvironmental factors influencing life span. Mechanisms of Ageing andDevelopment 6, 413e429.

Kleppel, G.S., Burkart, C.A., 1995. Egg production and the nutritional environment ofAcartia tonsa e the role of food quality in copepod nutrition. Ices Journal ofMarine Science 52, 297e304.

Kurkela, R., Koivurinta, J., Kuusinen, R., 1980. Non-protein nitrogen compounds inthe higher fungi e a review. Food Chemistry 5, 109e130.

Larsen, T., Gorissen, A., Krogh, P.H., Ventura, M., Magid, J., 2007. Assimilationdynamics of soil carbon and nitrogen by wheat roots and Collembola. Plant andSoil 295, 253e264.

Larsen, J., Johansen, A., Larsen, S.E., Heckmann, L.H., Jakobsen, I., Krogh, P.H., 2008.Population performance of collembolans feeding on soil fungi from differentecological niches. Soil Biology and Biochemistry 40, 360e369.

Larsen, T., Taylor, D., Leigh, M.B., O’Brien, D., 2009a. Stable isotope fingerprinting:a novel method for identifying plant, fungal or bacterial origins of amino acids.Ecology 90, 3526e3535.

Larsen, T., Ventura, M., Damgaard, C., Hobbie, E.A., Krogh, P.H., 2009b. Nutrientallocations and metabolism in two collembolans with contrasting reproductionand growth strategies. Functional Ecology 23, 745e755.

Lavelle, P., Bignell, D., Lepage, M., Wolters, V., Roger, P., Ineson, P., Heal, O.W.,Dhillion, S., 1997. Soil function in a changing world: the role of invertebrateecosystem engineers. European Journal of Soil Biology 33, 159e193.

Lavy, D., Verhoef, H.A., 1996a. Effects of food quality on growth and body compo-sition of the collembolan Orchesella cincta. Physiological Entomology 21, 64e70.

Lavy, D., Verhoef, H.A., 1996b. Spatiotemporal variation in body composition andcold tolerance of soil arthropods. Pedobiologia 40, 529e540.

Legendre, P., Legendre, L., 1998. Numerical Ecology, second ed. Elsevier, New York,853 pp.

Lii, G.Y., Garlich, J.D., Rock, G.C., 1975. Protein and energy utilization by the insect,Argyrotaenia velutinana (walker) fed diets containing graded levels of an aminoacid mixture. Comparative Biochemistry and Physiology Part A: Physiology 52,615e618.

Lindeman, R.L., 1942. The trophic-dynamic aspect of ecology. Ecology 23, 399e418.Lomstein, B.A., Jorgensen, B.B., Schubert, C.J., Niggemann, J., 2006. Amino acid

biogeo- and stereochemistry in coastal Chilean sediments. Geochimica etCosmochimica Acta 70, 2970e2989.

Lussenhop, J., 1992. Mechanisms of microarthropod-microbial interactions in soil.Advances in Ecological Research 23, 1e33.

Martinson, H.M., Schneider, K., Gilbert, J., Hines, J.E., Hamback, P.A., Fagan, W.F.,2008. Detritivory: stoichiometry of a neglected trophic level. EcologicalResearch 23, 487e491.

Maruyama, A., Yamada, Y., Rusuwa, B., Yuma, M., 2001. Change in stable nitrogenisotope ratio in the muscle tissue of a migratory goby, Rhinogobius sp., in a naturalsetting. Canadian Journal of Fisheries and Aquatic Sciences 58, 2125e2128.

McCullagh, J., Gaye-Siessegger, J., Focken, U., 2008. Determination of underivatizedamino acid d13C by liquid chromatography/isotope ratio mass spectrometry fornutritional studies: the effect of dietary non-essential amino acid profile on theisotopic signature of individual amino acids in fish. Rapid Communications inMass Spectrometry 22, 1817e1822.

McMahon, K.W., Fogel, M.L., Elsdon, T.S., Thorrold, S.R., 2010. Carbon isotope frac-tionation of amino acids in fish muscle reflects biosynthesis and isotopicrouting from dietary protein. Journal of Animal Ecology 79, 1132e1141.

Moe, S.J., Stelzer, R.S., Forman, M.R., Harpole, W.S., Daufresne, T., Yoshida, T., 2005.Recent advances in ecological stoichiometry: insights for population andcommunity ecology. Oikos 109, 29e39.

Motulsky, H.J., Ransnas, L.A., 1987. Fitting curves to data using nonlinear-regression ea practical and nonmathematical review. Faseb Journal 1, 365e374.

Nation, J.L., 2002. Insect Physiology and Biochemistry. CRC Press, Boca Raton, 485 pp.O’Brien, D.M., Fogel, M.L., Boggs, C.L., 2002. Renewable and nonrenewable

resources: amino acid turnover and allocation to reproduction in lepidoptera.Proceedings of the National Academy of Sciences of the United States ofAmerica 99, 4413e4418.

O’Brien, D.M., Boggs, C.L., Fogel, M.L., 2003. Pollen feeding in the butterfly Heliconiuscharitonia: isotopic evidence for essential amino acid transfer from pollen toeggs. Proceedings of the Royal Society B-Biological Sciences 270, 2631e2636.

O’Brien, D.M., Boggs, C.L., Fogel, M.L., 2005. The amino acids used in reproduction bybutterflies: a comparative study of dietary sources using compound-specificstable isotope analysis. Physiological and Biochemical Zoology 78, 819e827.

Odum, H.T., 1957. Trophic structure and productivity of Silver Springs, Florida.Ecological Monographs 27, 55e112.

O’Neill, R.V., 1968. Population energetics of millipede Narceus americanus (Beau-vois). Ecology 49, 803e809.

Ostle, N., Briones, M.J.I., Ineson, P., Cole, L., Staddon, P., Sleep, D., 2007. Isotopicdetection of recent photosynthate carbon flow into grassland rhizospherefauna. Soil Biology and Biochemistry 39, 768e777.

Petersen, H., Luxton, M., 1982. A comparative analysis of soil fauna populations andtheir role in decomposition processes. In: Petersen, H. (Ed.), QuantitativeEcology of Microfungi and Animals in Soil and Litter. Oikos, pp. 287e388.

Plante, C.J., Jumars, P.A., Baross, J.A., 1990. Digestive associations between marinedetritivores and bacteria. Annual Review of Ecology and Systematics 21,93e127.

Pokarzhevskii, A.D., Zaboyev, D.P., Ganin, G.N., Gordienko, S.A., 1997. Amino acids inearthworms: are earthworms ecosystemivorous? Soil Biology and Biochemistry29, 559e567.

Pollierer, M.M., Scheu, S., Haubert, D., 2010. Taking it to the next level: trophictransfer of marker fatty acids from basal resource to predators. Soil Biology andBiochemistry 42, 919e925.

Ramsay, S.L., Houston, D.C., 1998. The effect of dietary amino acid composition onegg production in blue tits. Proceedings of the Royal Society of London Series B-Biological Sciences 265, 1401e1405.

R-Development-Core-Team, 2009. R: a Language and Environment for StatisticalComputing. R Foundation for Statistical Computing, Vienna, Austria.

Roberts, G.A.F., 1992. Chitin Chemistry. Macmillan, London, xvi, 350 pp.Rothstein, M., 1963. Nematode biochemistry-III. Excretion products. Comparative

Biochemistry and Physiology 9, 51e59.Ruess, L., Haggblom, M.M., Garcia Zapata, E.J., Dighton, J., 2002. Fatty acids of fungi

and nematodesdpossible biomarkers in the soil food chain? Soil Biology andBiochemistry 34, 745e756.

T. Larsen et al. / Soil Biology & Biochemistry 43 (2011) 749e759 759

Ruess, L., Häggblom, M.M., Langel, R., Scheu, S., 2004. Nitrogen isotope ratios andfatty acid composition as indicators of animal diets in belowground systems.Oecologia 139, 336e346.

Ruess, L., Schutz, K., Migge-Kleian, S., Haggblom, M.M., Kandeler, E., Scheu, S., 2007.Lipid composition of Collembola and their food resources in deciduous foreststands e implications for feeding strategies. Soil Biology and Biochemistry 39,1990e2000.

Sang, J.H., King, R.C., 1961. Nutritional requirements of axenically culturedDrosophila melanogaster adults. Journal of Experimental Biology 38, 793e809.

Schulten, H.R., Schnitzer, M., 1997. The chemistry of soil organic nitrogen: a review.Biology and Fertility of Soils 26, 1e15.

Sibly, R.M., Calow, P., 1986. Physiological Ecology of Animals: an EvolutionaryApproach. Blackwell Scientific, Oxford, ix, 179 pp.

Silfer, J.A., Engel, M.H., Macko, S.A., Jumeau, E.J., 1991. Stable carbon isotope analysisof amino acid enantiomers by conventional isotope ratio mass spectrometryand combined gas chromatography/isotope ratio mass spectrometry. Anal.Chem. 63, 370e374.

Sterner, R.W., Elser, J.J., 2002. Ecological Stoichiometry: The Biology of Elementsfrom Molecules to the Biosphere. Princeton University Press, Princeton, xxi, 439pp.

Teuben, A., Verhoef, H.A., 1992. Direct contribution by soil arthropods to nutrientavailability through body and fecal nutrient content. Biology and Fertility ofSoils 14, 71e75.

Thimm, T., Hoffmann, A., Borkott, H., Munch, J.C., Tebbe, C.C., 1998. The gut of thesoil microarthropod Folsomia candida (Collembola) is a frequently changeable

but selective habitat and a vector for microorganisms. Applied and Environ-mental Microbiology 64, 2660e2669.

Ulber, B., 1983. Einfluss von Onychirurus fimatus Gisin (Collembola, Onychiuridae)und Folsomia fimetaria L. (Collembola, Isotomidae) auf Pythium ultimum Trow.einen Erreger des Wurzelbrandes der Zuckerrübe. In: Lebrun, P., André, H., DeMedts, A., Grégoire-Wibo, C., Wauthy, G. (Eds.), Trends in Soil Biology, Louvain-la-neuve (Belgium), pp. 261e268.

Venables, W.N., Ripley, B.D., 2002. Modern Applied Statistics with S, fourth ed.Springer, New York, xi, 495 pp.

Ventura, M., Catalan, J., 2008. Incorporating life histories and diet quality in stableisotope interpretations of crustacean zooplankton. Freshwater Biology 53,1453e1469.

Ventura, M., Catalan, J., 2010. Variability in amino acid composition of alpinecrustacean zooplankton and its relationship with nitrogen-15 fractionation.Journal of Plankton Research, fbq066.

Ventura, M., 2006. Linking biochemical and elemental composition in freshwaterand marine crustacean zooplankton. Marine Ecology-Progress Series 327,233e246.

Visser, S., Whittaker, J.B., Parkinson, D., 1981. Effects of collembolan grazing onnutrient release and respiration of a leaf litter inhabiting fungus. Soil Biologyand Biochemistry 13, 215e218.

Wardle, D.A., 2002. Communities and Ecosystems: Linking the Aboveground andBelowground Components. Princeton University Press, Princeton, NJ, vii, 392 pp.

White, T.C.R., 2008. The role of food, weather and climate in limiting the abundanceof animals. Biological Reviews 83, 227e248.