BiogeochemistryD O I 10.1007/S10533-006-9065-Z

O R IG IN A L P A P E R

Increases in soil respiration following labile carbon additions linked to rapid shifts in soil microbial community composition

Cory C. Cleveland • Diana R. Nemergnt Steven K. Schmidt • Alan R. Townsend

Received: 16 June 2006 / Accepted: 10 November 2006 © Springer Science+Business Media B.V. 2006

Abstract O rganic m atter decom position and soil C O 2 efflux are bo th m ediated by soil m icroorgan­isms, bu t the po ten tia l effects of tem poral varia­tions in m icrobial com m unity com position are not considered in m ost analytical m odels of these two im portan t processes. H ow ever, inconsistent relationships betw een rates of hetero troph ic soil respiration and abiotic factors, including tem per­ature and m oisture, suggest th a t m icrobial com ­m unity com position m ay be an im portan t regulator of soil organic m atter (SOM ) decom position and C O 2 efflux. W e perform ed a short-term (12-h) laboratory incubation experim ent using tropical rain forest soil am ended w ith either w ater (as a control) or dissolved organic m atter (D O M ) leached from native plant litter, and analyzed the effects of the treatm ents on soil respiration and m icrobial com m unity com position. The la tte r was

C. C. Cleveland (E3) ■ D. R. Nemergut ■A. R. TownsendINST AAR: A n E arth and Environm ental Sciences Institute, University of Colorado, 450 UCB/1560 30th Street, Boulder, CO 80303, USA e-mail: Cory.Cieveiand@Coiorado.Edu

D. R. NemergutEnvironm ental Studies Program, University of Colorado, 397 UCB, Boulder, CO, USA

S. K. Schmidt ■ A. R. TownsendD epartm ent of Ecology and Evolutionary Biology,University of Colorado, 334 UCB, Boulder, CO, USA

determ ined by constructing clone libraries of sm all-subunit ribosom al R N A genes (SSU rR N A ) extracted from the soil at the end of the incubation experim ent. In contrast to the subtle effects of adding w ater alone, additions of D O M caused a rapid and large increase in soil C O 2 flux. D O M - stim ulated C O 2 fluxes also coincided w ith p ro ­found shifts in the abundance of certain m em bers of the soil m icrobial com m unity. O ur results suggest th a t natu ral D O M inputs may drive high rates of soil respiration by stim ulating an opportu ­nistic subset of the soil bacterial com m unity, particularly m em bers of the G am m aproteobacte- ria and Firm icutes groups. O ur experim ent indi­cates tha t variations in m icrobial com m unity com position may influence SOM decom position and soil respiration rates, and em phasizes the need for in situ studies of how natu ral variations in m icrobial com m unity com position regulate soil biogeochem ical processes.

Keywords C arbon cycle ■ D ecom position ■ Dissolved organic m atter ■ D O M ■ M icrobial com m unity com position ■ O rganic m atter ■ Soil respiration

Introduction

O rganic m atter decom position is a fundam ental process, regulating rates of ne t carbon (C)

^ Springer

Biogeochemistry

Storage and nu trien t cycling in terrestria l eco­systems. D ecades of research have inform ed the developm ent of ecosystem m odels tha t describe decom position and hetero troph ic soil respiration rates prim arily as functions of abiotic environ­m ental factors including tem peratu re, m oisture, and organic m atter substrate quantity and quality (M eentem eyer 1978; Swift et al. 1979; Scott-D enton et al. 2003; D avidson and Janssens 2006; Scott-D enton et al. 2006). H ow ever, the po ten tia l effects of tem poral variations in m icro­bial com m unity structure on these processes are seldom explicitly acknow ledged (but see Fang et al. 2005). Instead, m ost conceptual and em pirical decom position m odels trea t the m icro­bial com m unity as a “ black box” in which m icroorganism s act as passive catalysts of decom position reactions of substrates whose rates vary as functions of tem pera tu re and m oisture (R aich and Schlesinger 1992; Lloyd and Taylor 1994; P arton et al. 1994).

Simple soil respiration-clim ate relationships are convenient for m odels attem pting to describe ecosystem C O 2 fluxes, bu t m any studies show contradictory responses of soil respiration to variations in clim ate (D avidson and Janssens2006). Soil m oisture correlates w ith soil resp ira­tion rates in some sites (e.g., B uchm ann 2000; Schw endenm an et al. 2003), bu t only weakly predicts ra tes of soil respiration in others (Scott- D en ton et al. 2003; C leveland and Tow nsend 2006; M onson et al. 2006). The inability of clim ate to adequately explain variation in soil respiration rates suggests tha t soil C O 2 fluxes are regulated by a m ore com plex set of variables and interactions.

For exam ple, hetero troph ic C O 2 fluxes also depend on the availability of soluble, labile C sources, and m icrobial respiration rates are tightly linked to the chem istry and am ount of organic m atter entering the soil (Schlesinger and A ndrew s 2000; W ardle et al. 2004). In addition, th ree em erging lines of evidence suggest that short- and long-term variations in soil m icrobial com m unity com position (e.g., Schm idt et al.2007) could also influence soil respiration rates. F irst, m icrobial com m unity structure varies sig­nificantly w ithin soil types— both spatially and tem porally—^with research indicating th a t w ithin

a single site, rhizosphere m icrobial com m unity com position and activity vary by p lant species and across seasons (G rayston et al. 1998; Bardg- e tt e t al. 1999; W ardle et al. 2004; C arney and M atson 2005; Schm idt et al. 2007). Second, m anipulative experim ents have shown that specialized organism s respond to certain C com pounds. Padm anabhan et al. (2003) found tha t specific and unique soil bacterial assem ­blages w ere responsible for the decom position of individual soil organic com pounds, including glucose, naphthalene, phenol, and caffeine. Finally, the types of m icroorganism s w ithin a site th a t respond to C inputs vary th rough tim e, suggesting th a t tem poral shifts in com m unity structure may directly affect soil respiration rates following C inputs. For exam ple, in a subalpine forest soil, C inputs consistently stim ulated m em bers of the B etapro teobacteria , bu t the specific types of soil B etapro teobacteria that responded in sum m er w ere genetically and physiologically distinct from those responding in w inter (M onson et al. 2006).

These studies and others support proposed links betw een m icrobial com m unity com position and ecosystem function (A dam s and W all 2000; H ooper e t al. 2000; W ardle e t al. 2004). U nfo r­tunately , relatively little is know n about the types of m icroorganism s tha t decom pose native soil organic m atter (SO M ), or how variations in m icrobial activity in response to labile C inputs are re la ted to soil m icrobial com m unity com po­sition. H ere, we asked a simple question: in an ecosystem w here high rates of soil respiration are positively re la ted to seasonal increases in C availability (C leveland and Tow nsend 2006), could seasonal increases in soil respiration be re la ted to variations in m icrobial com m unity com position? To begin to address this question, we perform ed a soil incubation experim ent using soil collected from a m ature, lowland tropical rain forest. Previous research from the study site revealed a strong, positive relationship betw een labile C availability and soil respiration rates (C leveland and Tow nsend 2006), and we hypoth­esized tha t dissolved organic m atter (D O M )- stim ulated increases in soil respiration coincide w ith rapid, m easurable shifts in soil m icrobial com m unity com position.

^ Springer

Biogeochemistry

Materials And Methods

Study sites and sam pling regim e

Soil and litter sam ples w ere collected from a prim ary tropical rainforest on the O sa Peninsula in southw estern C osta R ica (see C leveland et al. 2006 for com plete site description). Soils at the site are highly w eathered, nu trien t poor Ultisols. The site lies w ithin the tropical w et low land forest bioclim ate, w ith an average annual tem peratu re of 26.5°C (H oldridge et al. 1971). A verage annual rainfall exceeds 5,000 m m (C leveland and Tow n­send 2006), bu t like m ost tropical rainforests, this area experiences a dry season (from January to M arch on the O sa Peninsula). D uring the 3-m onth dry season, m ore than half of the annual litter falls on the forest floor (>500 g C m ^; Cleveland and Tow nsend 2006), providing a large pool of soluble, highly decom posable C (e.g., D on and K albitz 2005) tha t m ay fuel high ra tes of soil respiration during the early rainy season (C leve­land and Tow nsend 2006). In this experim ent, our objective was to sim ulate a D O M pulse leached from litter layer during a low-intensity ra in ­storm — an event tha t occurs alm ost daily in this site during the early rainy season (C leveland and Tow nsend 2006).

W e collected 8 x 10 cm surface soil cores from each of ten 25 m^ plots in the dry season using a hand corer (see C leveland et al. 2006 for soil physical and chem ical characteristics). Im m edi­ately after collection, soil sam ples w ere trans­ported on ice in an insulated cooler to the laboratory (within 72 h of collection), and stored at ~4°C until analysis. Soil sam ples w ere gently hand-hom ogenized (to 4 mm) to rem ove roots and other organic debris. Following hom ogeniza­tion, sam ples w ere split in to paired experim ental samples. A ll incubation experim ents w ere initi­ated w ithin 96 h of soil sam pling to avoid artifacts incurred during long-term storage.

Soil incubation experim ent

To sim ulate an episodic D O M leaching event, we added soluble organic m ateria l leached from freshly fallen litter sam ples to soil. A b ou t 100 g of mixed, air-dried p lant litter ob tained from

litter traps in the dry season (litter harvested at 2-week intervals th roughout the year; C leveland and Tow nsend 2006) w ere extracted in 1 L of de-ionized w ater for 12 h at 22°C, pre-filtered to 0.45 f im and sterile filtered to 0.22 fim. The C concentration of the resulting leachate (650 mg/L) was m easured using a Shimadzu T O C 5050A to ta l organic carbon analyzer (Shim adzu C orporation , Kyoto, Japan). Sterility of the leachate was confirm ed using Biolog m icrotiter plates (B iolog Inc., H ayw ard, CA, U SA ); no color developm ent was observed after 24, 48 or 72 h, indicating th a t the leachate was free of all viable bacteria. A pproxim ately 25 g of field-m oist soil (-15 g dry weight) w ere placed in 1 L glass vessels and pre-incubated at 26°C (±1°C) for 24 h in a Precision 815 low- tem pera tu re incubator (Precision Scientific, W inchester, V A , U SA ).

A fter pre-incubation and equilibration at 26°C, all vessels w ere aera ted and sam ples ( N = 10 replicates per trea tm en t) received -5 m L doses (i.e., an am ount to bring each soil sam ple to -50% of w ater holding capacity) of D O M (-225 fig D O M -C g“ soil; -hDOM). T o separate w ater (soil wet-up) versus C -addition effects on m icrobial com m unity activity and com position, we also am ended a parallel set of sam ples w ith -5 m L of de-ionized w ater ( N = 10 replicates; -HH2 O). Following trea tm en t additions, glass vessels w ere im m ediately sealed with caps equipped with rubber septa for gas sam pling, incubated at 26°C in darkness for 12 h, and sam pled for C O 2 a t 2, 4, 7, 9, and 12 h using glass gas-tight syringes. Sam ple C O 2 concentrations w ere determ ined at each tim e poin t using a Shim adzu G C -14 gas chrom atograph equipped w ith a therm al conduc­tivity detector.

D N A extraction and analysis

O ur strategy was to sam ple soil m icrobial com ­m unity com position w hen soil respiration achieved m axim um rates. Previous experim ents showed tha t following C additions to these soils, m icrobial respiration ra tes w ere highest after -1 2 h of incubation at 26°C (C leveland et al.2002). T herefore, after the 12-h headspace C O 2

sam pling was com plete, -5 g of soil from each

^ Springer

Biogeochemistry

m icrocosm w ere rem oved to sterile conical tubes and preserved at -80°C until D N A extraction.

Soil D N A was extracted from soil sam ples using a m odification of the protocol described by M ore et al. (1994). Briefly, 0.5 g of soil from each of the +H 2 O and +D O M sam ples w ere agitated in 1 m L phosphate extraction buffer using a bead bea te r and centrifuged at 14,000g for 10 min. The resulting supernatan t was rem oved to a fresh tube and incubated on ice for 5 m in with 200 f i h of 7.5 M am m onium acetate, centrifuged at 14,000g for 3 m in, and the supernatan t transferred to a fresh tube. Sam ples w ere then extracted w ith an equal volum e of phenol : chloroform : isoamyl alcohol (25:24:1), D N A was precip ita ted with 200 f i h of isopropanol overnight at -20°C and w ashed with 1 m L of 70% ethanol. Following extraction, the individually extracted D N A sam ­ples from each trea tm en t w ere pooled and puri­fied over Sepharose 4B packed colum ns (Sigma, St. Louis, M O , U SA ) as described in Jackson et al. (1997).

PC R , cloning, and sequencing

A pproxim ately 30 ng of D N A w ere amplified w ith universal bacterial 16S rD N A prim ers [8f and 1492r (Lane 1991)]. E ach reaction included 400 uM of each prim er, 200 f iM of each dNTP, 1.25 U of T aq D N A polym erase (Prom ega, M adison, W I, U SA ) in T aq D N A polym erase buffer [2.5 mM M gC li (Prom ega)j. A fter an initial denaturation step a t 94°C (1 m in), P C R was perform ed using 30 cycles at 94°C for 1 min, 58 ± 5°C for 30 s and 72°C for 2.5 m in w ith a term inal 10 m in extension at 72°C. R esulting PC R products w ere gel purified using Q IA quick gel purification colum ns (Q iagen, V alencia, CA, U SA ), ligated into the vector T O P O 2.1 (Invitro- gen Inc., Carlsbad, CA , U SA ) and transform ed into Escherichia coli following the m anufactu rer’s instructions.

T ransform ants w ere inoculated into 96-well deep-dish plates containing 1.5 m L L uria broth (1% NaCl, 1% tryptone, 0.5% yeast extract) and 50 mg m L“ ampicillin. C ultures w ere agitated at 200g for 14 h at 37°C, plasm id D N A was extracted and purified (Sam brook and Russell 2001), and pelleted D N A sam ples w ere air-dried

(20 min.) and re-suspended in 50 m L T ris-H C l pH 8.5. Inserted 16S rR N A genes w ere PC R amplified from the plasm ids using the prim ers M 13F and M 13R (Invitrogen Inc.), and purified 16S rR N A genes w ere sequenced with the T7 prom oter prim er and the M13-9 prim er using the BigD ye T erm inator Cycle Sequencing kit V ersion3.0 (P E Biosystem s, F oster City, CA, U SA ) according to the m anufactu rer’s directions. Sequencing products w ere analyzed at the Iowa State U niversity D N A Sequencing Facility.

D ata analysis

A ll soil respiration data w ere tested for hom o- scedasticity (L evene’s test for equal variances), norm ality, and skew edness (SPSS, Chicago, IE, U SA ). Significant differences in m axim um respi­ra tion rates following D O M additions in the incubation experim ent w ere tested w ith a paired-sam ples t-test (a = 0.05; Snedecor and Cochran 1989).

D N A sequences w ere edited in Sequencher 4.1 (G ene Codes Co., A nn A rbor, M I, U SA ) and subjected to B LA ST searches (A ltschul et al. 1990). The 16S rR N A gene sequences w ere also subjected to chim era check in R D P (Cole et al.2003) and B ellerophon (H uber et al. 2004). These sequences w ere aligned in D r. Phil H ugenholtz’s 16S rR N A A R B database (http://rdp8.cm e.m su. edu/htm Palignm ents.htm l) in A R B (Ludwig et al.2004). Phylogenetic affiliations w ere assigned based on bo th the A R B alignm ent and nearest B LA ST m atches. O perational taxonom ic unit (O T U ) designations w ere assigned using D O - T U R (Schloss and H andelsm an 2005) w ith a distance-based phylogenetic tree constructed in A R B . L ineage-per-tim e plots w ere also con­structed in D O T U R (Schloss and H andelsm an2005). D etailed phylogenies of the G am m aprote- obacteria and the Firm icutes w ere constructed by aligning close relatives from the A R B database and B L A ST m atches w ith our sequences in A R B . A lignm ents w ere used to generate a neighbor- joining phylogenetic tree in P A U P (Swofford 2001) using the distance optim ality criterion. A ll trees w ere subjected to bootstrap analyses (1,000 replicates) using bo th the parsim ony and neighbor-joining optim ality criterion.

^ Springer

Biogeochemistry

Results

Soil respiration rates

Dissolved organic m atter additions (3,375 fig C as D O M ) rapidly stim ulated respiration in the soil m icrocosm s (Fig. 1). A fter the 12-h incubation, soil respiration rates in the -hDOM sam ples w ere 35.1 ± 1.7 fig C O 2 -C g“ hr^, versus only9.1 ± 1.2 fig C O 2 - C g“ h“ in the -HH2 O sam ples (Fig. lA ) . Twelve hours after the D O M addition, the -hDOM sam ples had respired -100% m ore CO 2 than the control soil sam ples (2,981 ± 245 fig C O 2 - C vs. 1,503 ± 80 fig C O 2 - C in -hDOM and -HH2 O sam ples, respectively; Fig. IB ). In contrast, the soil w et-up (w ater only) had little effect on soil respiration rates. Im m ediately following w ater additions, soil respiration rates in -HH2 O sam ples increased slightly (Fig. lA ) , bu t rates of

40

35

co 25

20Q.{/}a>CC

3500

3000O

C\i 2500

2000

1500Q .{/}<1> 1000

o 500

0 2 4 6 8 1 0 1 2 1 4

T im e (h)

Fig. 1 (A) Soil respiration rates; and (B) total CO 2

respired in soil incubations following water (open circles) or dissolved organic m atter (DOM', open squares) addi­tions (3,375 /ig C as DOM ) to tropical rain forest soil. Soil respiration rates are expressed on a soil mass (dry -weight) basis. Values represent the mean of ten replicates per treatm ent, and error bars are ±1 SE

soil respiration in the -HH2 O sam ples did no t differ significantly over the 12-h incubation (Fig. lA ) .

Soil m icrobial com m unity com position

W e analyzed a to tal of 102 sequences from the -HH2 O ( N = 46) and -hDOM sam ples ( N = 56), and there w ere som e notable sim ilarities in com m unity com position betw een the two tre a t­m ents. O verall, broad-scale (i.e., division- and sub-division-level) diversity was sim ilar in both sets of sam ples, w ith sequences w ithin nine and ten m ajor bacterial clades rep resen ted in the -HH2 O and -hDOM sam ples, respectively (Fig. 2). Sequences re la ted to several of these clades w ere found in bo th the -HH2 O and -hDOM soil samples, including the A cidobacteria, Firm icutes, A lphapro teobacteria , B etapro teobacteria , D elta- p ro teobacteria , V errucom icrobia, and the Plan- ctom ycetes. H ow ever, m ore detailed com parisons of the com m unities revealed that, in spite of the sim ilarities betw een the two trea tm ents at b road ­er phylogenetic levels, the types of organisms w ithin the -HH2 O and -hDOM com m unities w ere very different. For exam ple, we com pared the actual num ber of unique sequences (>3% 16S rR N A gene sequence difference) in the two com m unities over various genetic distances with the theoretical m axim um num ber of possible unique sequences (i.e., if there was no overlap in the two com m unities), and the theoretical m inim um num ber of possible unique sequences (i.e., if each sequence p resen t in one com m unity was also presen t in the o ther com m unity; Fig. 3). R esults indicated tha t a t genetic distances of less than 10% , the -HH2 O and -hDOM com m unities w ere distinct from each other.

L ineage per tim e (LPT) curves, in which the num ber of unique m icrobial phylotypes is p lo tted as a function of genetic distance, provide an index of the am ount of phylogenetic diversity w ithin a com m unity. L PT plots for these com m unities also revealed profound differences betw een the -HH2 O and -hDOM soils. The -HH2 O curves show ed a gradual decline, suggesting th a t phylogenetic diversity was relatively high. In contrast, in the -hDOM sam ples, LPT curves show ed an initial, relatively steep decline, indicating tha t m any of the individual phylotypes in the -hDOM samples

^ Springer

Biogeochemistry

Fig. 2 Relative abundance of 16S rR N A gene phylotypes in (A) +H 2 O; and (B) +DOM samples. Colors represent division or sub-division- level clades(Proteobacteria) and the black lines represent unique phylotypes (i.e., >3% different from all other sequences)

- Verrucomicrobia

Pianctomycetesr D e l ta p r o te o b a c t e r i i

ChiorofiexiNitrospiraiacteroidetes

Acidobacteria

- Firmicutes - A lp h a p r o t e o b a c t e r ia

- B e t a p r o t e o b a c t e r ia

^ D e i t a p r o t e o b a c t e r i a /— C a n d id a te D iv is io n T M 7 A . Verrucomicrobia

Pianctomycetes Acidobacteria

B e t a p r o t e o b a c t e r ia - A lp h a p r o t e o b a c t e r ia

- G a m m a p r o t e o b a c t e r ia

Firmicutes

100

90

80

70

60CODH 500

40

30

20

10

0

• com bined data ' theoretica l m aximum ‘ theoretica l m inimum

0.1 0.2

Genetic D istance

0.3

Fig. 3 Estim ated number of unique phylotypes in the combined (+DOM and +H 2 O) libraries (circles), com­pared to the theoretical minimum num ber of phylotypes (i.e., if all sequences in one library were found in the other library; triangles) and the theoretical maximum num ber of phylotypes (i.e., if none of the sequences present in one library were found in the other library; squares). This plot indicates that although there was some division-level sequence overlap in the +DOM and +H 2 O libraries, the sequences in the +H 2 O and +DOM libraries were very different

w ere closely rela ted , and th a t the depth of phylogenetic diversity in the -hDOM sam ples was lower than in the -HH2 O sam ples (data not shown).

O ur sam pling and analysis of m icrobial com ­m unity com position in the soil sam ples was not exhaustive, therefo re it is difficult to m ake defin­itive sta tem ents about the abundance of specific bacterial lineages or overall m icrobial diversity.

bu t our results do show som e clear and im portan t differences. For exam ple, perhaps the m ost strik­ing difference betw een -hDOM and control soil com m unities was the increase in abundance of non-A cidobacteria sequences in the -hDOM soil. Sequences re la ted to the A cidobacteria clade w ere num erically dom inant in bo th the -HH2 O and -hDOM sam ples, but they represen ted nearly 60% of the sequences in the contro l library and only about 30% in the -hDOM library (Fig. 2). In addition, in the control samples, non-A cidobac- teria l sequences included m em bers of nine m ajor clades, and the frequencies of individuals w ithin those groups w ere fairly equally represented; no single non-A cidobacterial clade rep resen ted m ore than 11% of the sequences, and m ost individual clades com prised <4% of the sequences. In contrast, in the -hDOM samples, non-A cidobac- teria l sequences w ere dom inated by individuals from the G am m aproteobacteria and Firm icutes groups, which com prised 34 and 16% of the -hDOM sequences, respectively. In the -HH2 O samples, phylotypes from the Firm icutes group w ere p resen t in small num bers (-4 % ), bu t the sequences w ere no t closely re la ted to the Firm i- cute sequences in the -HH2 O sam ples (data not shown). Sequences rela ted to the G am m apro teo­bacteria w ere no t detected in the -HH2 O samples.

Discussion

Soil m oisture conten t can lim it decom position and soil respiration ra tes (e.g., H ousm an et al.

^ Springer

Biogeochemistry

2006; Potts et al. 2006). H ow ever, in this tropical rain forest site, soil m oisture (to 10 cm depth) rem ains relatively constant th roughout the year, even during the 3-m onth dry season w hen p re­cipitation is <100 m m /m onth (C leveland and Tow nsend 2006). In this experim ent, w ater-only additions did no t significantly enhance rates of soil respiration (Fig. la ) . The lack of a strong CO 2 response to the soil w et-up indicates that even though soil sam ples w ere collected in the dry season, soil m oisture was no t lim iting to h e te ro ­trophic respiration at th a t tim e. This result is consistent w ith previous data from this site showing tha t soil m oisture does no t correlate strongly w ith rates of soil respiration or o ther soil processes (C leveland and Tow nsend 2006).

A lthough seasonal increases in rainfall do not drive significant variation in soil m oisture, they do facilitate the m ovem ent of D O M from litte r to soil, driving significant dry-to-w et season in­creases in soil respiration (C leveland et al. 2006; Cleveland and Tow nsend 2006). In this experi­m ent, m icrobial respiration responded rapidly to dissolved organic C additions in the -hD O M sam ples (Fig. 1). M any o ther studies showing rapid m icrobial responses to labile C additions have added relatively high concentrations of sim ple, pure, C-rich com pounds tha t stim ulate m icrobial activity, bu t tha t m ay not necessarily be analogous to native organic m ateria l th a t is presen t in natu ral ecosystems. Thus, while the high rates of m icrobial respiration we observed in response to C additions (in general) are not surprising, the rapid response im m ediately fol­lowing additions of relatively dilute, litter-leached native D O M is notew orthy for several reasons.

First, leached D O M is a com plex, he teroge­neous suite of C com pounds tha t vary in com plexity and decom posability (N eff and A sner 2001; C leveland et al. 2004), yet it still stim ulated rapid rates of m icrobial respiration in our incu­bation. In addition, previous in situ D O M additions elicited similarly high-respiration rates, suggesting th a t rap id respiratory responses to episodic D O M inputs may be im portan t in regulating ecosystem function in this site (C leve­land et al. 2002). N ext, while the concentration of the D O M added here (-650 mg L“^) was higher than typical tropical rain forest throughfall or soil

solution concentrations (Schw endenm ann and V eldkam p 2005), it did represen t realistic con­centrations of D O M through the litter layer (N eff and A sner 2001; C leveland et al. 2004). Previous research at this site show ed th a t as m uch as 50% of dry season litte r C is susceptible to leaching in the first 60 days of the rainy season (Cleveland et al. 2006). If so, a typical daily 10 m m rainfall event in the early rainy season and subsequent leaching of the dry season litter pool (250 mg C m “^) could generate a pulse of D O M of -200 mg to the soil surface. The increase in soil respiration tha t we observed following addi­tions of native, leached D O M suggests that episodic D O M fluxes in situ may also rapidly stim ulate high rates of soil respiration.

R ap id increases in m icrobial activity following pulses of labile C inputs are com m on, bu t do such increases rep resen t a com m unity-w ide response, or are they accom plished via shifts in the abun­dance of certain m icrobial taxa? O ur results provide com pelling evidence tha t increases in soil respiration following C inputs m ay be dom inated by a subset of soil m icrobes, w ith D O M inputs causing rap id shifts in soil bacterial com m unity com position. In the -HH2 O sam ples, as is com m on in m any soils, the A cidobacteria was the num er­ically dom inant group, bu t the abundances of individuals w ithin o ther division-level groups were relatively evenly distributed. H ow ever, at the peak of soil respiration, m icrobial com m unity structure in the -hDOM soil was m arkedly different. In particular, C additions appeared to strongly stim ­ulate organism s w ithin the G am m aproteobacteria and Firm icutes clades.

The dom inance of organism s w ithin the G am ­m apro teobacteria and Firm icutes groups, and their relative increase in abundance in response to C additions, suggests th a t m em bers of these groups quickly responded to the additions of organic m aterial added in the -hD O M soil, leading to rap id increases in soil respiration. D etailed phylogenetic analyses of the G am m a­pro teobacteria and Firm icutes sequences re ­vealed tha t they are re la ted to the order E nterobacteria les and the genus Bacillus, respec­tively (Figs. 4, 5), groups tha t include an array of both pathogenic and sym biotic bacteria. In te r­estingly, m em bers of these orders have been

^ Springer

Biogeochemistry

f— Shigella flexnerl

I * Escherichia collSalmonella typhlm urlum

L Salm, Salmonella enterica

^.K le b s ie lla pneum oniae

Klebsie lla CBNU-34-2 P solubilizer

Pantoea agglom erans P solubilizer_|L Pantoe

L EnterobaE nterobacter aerogenes P solubilizer

Kluyvera Intermedlus P solubilizer

Pasteurellales

Alterom onadales

Legionellales

A lterom onadales

Methylococcales

^ — 0.01 substitutions/site

Fig. 4 Neighbor-joining phylogenetic tree showing repre­sentative sequences from the clones related to the Enterobacteriales in the +DOM samples. Rectangles represent sequences from this study, and the size of the rectangle symbolizes the num ber of sequences obtained. Reference taxa were selected based on nearest BLAST matches. Tree is rooted with Rhodobaeter sphaeroides (X53853), Bartonella baeilliformis (X60042) and Rhizo- bium hainanensis (U71078). Asterisk indicates a parsimony and distance bootstrap (1,000 replicates) value of 80 or higher. Accession numbers: Shigella flexnerl (X80679), Eseheriehia eoli (X80724), Salmonella typhlmurlum (X80681), Salmonella enterlea (AL627282), Klebsiella pneumoniae (X87276), Klebsiella sp. CBNU-34-2 (AY335553), Pantoea agglomerans (AY335552), Enterob- aeter aerogenes (AY335552), Kluyvera Intermedlus (AF297470), Vibrionales [Vibrio natrlegens (X74714), Vibrio eholerae (X74696), Photobaeterlum hlstamlnum (D25308)], Pasteurellales [Alteromonas maeleodll (X82145), Chromohalobaeter marlsmortul (X87221)], Leg­ionellales [Eeglonella gormanll (Z32639), Eeglonella pneu­mophila (M59157)], Alteromonadales [Franelsella tularensls (L26086), Thlomlerosplra erunogena (L40810)], Methylococcales [Methyloeoeeus eapsulatus (M29023), Methyloeoeeus sp. (X72769)]

identified as poten tia l phosphate solubilizers, capable of accessing calcium-, iron- and alum i­num -bound phosphate m inerals (R odriguez and F raga 1999; C hung et al. 2005; Souchie et al.2006). This is notew orthy given previous w ork at this site indicating th a t the decom position of

■[j— —T » I-----

Lactobacillus agilis Enterococcus faecalis

S treptococcus

I— B acillus c irculans I— Bacillus m egaterlum P so lub ilizer1■I

B acillus firm us P so lub lllzer

— Saccharococcus thermophllus

Fllobaclllus

G racilibacillus d ipsosaurl G raclllbaclllus haloto lerans

I

P ontlbaclllus

Lentlbaclllus

Oceanobaclllus

• Exiguobacterlum

Paenlbaclllus

________ G eobacillus

Urelbaclllus

Ureaplasm a

M ycoplasm a

0.01 substitu tions/s ite

soluble C in soil is strongly phosphorus (P) lim ited (C leveland et al. 2002; C leveland and Tow nsend 2006). W hile by no m eans definitive, the presence of relatives of know n P-solubilizing organism s in the -hD O M sam ples suggests an interesting hypothesis: th a t the ability to solubi­lize P may give the E n terobacteria les and the Bacillus a com petitive advantage in these P-poor soils, allowing them to rapidly grow in response to episodic C inputs.

To our know ledge, our data are am ong the first to show th a t increases in soil respiration following native D O M additions to soils correspond with a rapid shift in m icrobial com m unity com position. Y et, rapid shifts in m icrobial com position in response to changing environm ental conditions and/or resources have been observed by m icrobial ecologists for some tim e. M ore than 75 years ago, W inogradsky (1925) observed increases in zymog- enous organism s following substrate additions, with a re tu rn to m ore “stab le ,” slow growing com m unities following substrate depletion. M ore

^ Springer

Biogeochemistry

■^Fig. 5 Neighbor-joining phylogenetic tree showing repre­sentative sequences from the clones related to the genus Bacillus in the +DOM samples. Rectangles represent sequences from this study, and the size of the rectangle symbolizes the num ber of sequences obtained. Reference taxa were selected based on nearest BLAST matches. Tree is rooted with Trichodesmium sp. (X70767), Nostoc punctiforme (AF027655) and Gloeothece membranacea (X78680). Asterisk indicates a parsimony and distance bootstrap (1,000 replicates) value of 80 or higher. Acces­sion numbers: Laetobaeillus agilis (M58803), Enterococcus faecalis (AE016830), Streptococcus [Streptococcus pneu­moniae (X58312), Streptococcus macedonicus (Z94012)], Bacillus circulans (D78312), Bacillus megaterlum (D16273), Bacillus firm us (D16268), Saccharococcus ther- mophilus (L09227), Filobacillus [Filobacillus salipiscarius (AB194046), Filobacillus milensis (AJ238042)], Graciliba­cillus dipsosaurl (AB101591), Gracilibacillus halotolerans (AF036922), Pontibacillus [Pontibacillus chungwhensis (AY603362), Pontibacillus marinus (AY603977)], Lenti- bacillus [Fentibacillus kapialis (AY667493), Fentibacillus salarius (AB231905)], Oceanobacillus [Oceanobaclllus chironomi (DQ468656) Oceanobacillus profundus (DQ386635)], Exiguobacterium [Exiguobacterlum auran- tiacum (X70316), Exiguobacterlum sp. (X86064)], Paeni- bacillus [Paenibacillus validus (M77489), Paenibacillus azotofxans (D78318)], Geobacillus [Geobacillus thermo- glucosidasius (AJ781265)], Ureibacillus [Ureibacillus koreensis (DQ348072), Ureibacillus rudaensis (DQ348071)], Ureaplasma [Ureaplasma canigenitalium (D78648), Ureaplasma parvum (L08642)], Mycoplasma [Mycoplasma capricolum (U26046), Mycoplasma mycoides (U26044)]

recently, F ierer et al. (2007) suggested an ecolog­ical classification of bacteria tha t p roposed that the abundance of copiotrophic and oligotrophic bacteria in soil is regulated by the abundance of labile C substrates, and tha t C decom position rates (which include hetero troph ic soil resp ira­tion) vary directly w ith copiotrophic bacterial abundance. For exam ple, in low C environm ents, oligotrophic groups like the A cidobacteria were m ost dom inant, bu t in high C soils, copiotrophic groups like the P ro teobacteria, and B acteroidetes increased in relative abundance (F ierer e t al.2007). In our experim ent, we observed a similar pattern . D O M additions caused a relative decrease in A cidobacterial sequences, and relative increases in G am m aproteobacteria and Firm icutes groups. O ur results support a m odel in which abundant bacterial phyla may be ecologi­cally classified according to their physiology, and in which the abundance of these ecologically defined groups (i.e., copiotrophic and ohgotrophic

groups) may be strongly regulated by C availability (F ierer et al. 2007). H ow ever, our results also suggest tha t m icrobial com m unity structure and com position may vary rapidly in response to short-term changes in C availability. In o ther words, while low-soil C availability (on average) may indeed favor oligotrophic organism s, natu ral and frequent increases in C availability in m any ecosystems may prom ote short-term variations in com m unity com position, and favor the tem porary dom inance of copiotrophic organisms.

The increased abundance of G am m apro teo­bacteria and Firm icutes sequences (Fig. 2) fol­lowing D O M additions suggests th a t m em bers of these two divisions may be largely responsible for the observed increases in soil respiration in our incubation experim ent, contributing to a growing body of research showing tha t m em bers of the these clades may dom inate m any im portan t b io­geochem ical processes in natu ral ecosystems. The im portance of the G am m aproteobactria , in p ar­ticular, in soil C cycling has also been shown in other studies. Padam anbhan et al. (2003) showed tha t phylotypes w ithin the G am m aproteobacteria clade w ere responsible for the decom position of pulses of suite of organic substrates including naphthalene, caffeine, and glucose. A m olecular characterization of phenol-degrading bacteria also showed tha t close relatives of both the G am m aproteobacteria and the Firm icutes w ere im portan t phenolic com pound m ineralizers in soil (A bd-E l-H aleem et al. 2002). These studies, am ong others, dem onstrate th a t m em bers of the G am m aproteobacteria and Firm icutes divisions are particularly adep t at responding to increases in labile soil C, and tha t in systems w here natu ral pulses of diverse C com pounds are com m on, these groups may be disproportionately im por­tan t to the decom position (and hence respiration) of labile C com pounds.

Traditionally, ecosystem m odels have focused on the role of abiotic factors in regulating organic m atter decom position rates. H ow ever, such sim ­plifications are often inadequate for explaining observed variations in soil respiration (D avidson and Janssens 2006; M onson et al. 2006; Sacks et al.2006). O ur results provide evidence of strong links betw een variations in bo th m icrobial com m unity com position and soil respiration, and suggest that

^ Springer

Biogeochemistry

a com plete understanding of decom position dynamics and soil respiration (and perhaps o ther soil biogeochem ical processes) in natu ral ecosys­tem s may require a m ore thorough understanding of tem poral shifts in m icrobial com m unity com ­position. W hile results from a short-term labora­tory experim ent w ith lim ited sam ple num bers may no t accurately describe m icrobial dynam ics in situ, they do provide bo th justification and a fram e­w ork for fu rther research investigating links betw een soil m icrobial com m unity structure and function in the natu ral environm ent. M any forces, including land use change (B ornem ann and T rip­le tt 1997; Nusslein and T iedje 1999), variations in nu trien t fertility (B ardgett e t al. 1999), and clim ate variability (M onson et al. 2006), have been shown to drive either short-term variations or perhaps long-term or perm anen t changes in m icrobial com m unity com position. U nderstand ­ing how natu ral and disturbance-driven processes affect m icrobial com m unity dynamics, and how these, in turn, in teract w ith abiotic factors to regulate soil biogeochem ical processes, m ay be im portan t in generating m ore accurate predictions of the response of terrestria l ecosystem s to envi­ronm ental change.

Acknowledgments We thank H. and M. Michaud of the D rake Bay Wilderness Camp for providing field access and logistical support in Costa Rica, and Francisco Campos and the Organizacion para Estudios Tropicales (OBT) and the Ministerio de Ambiente y Energia (M INAE) in Costa Rica for assisting with research permits and logistics. We are grateful to R. Ley for valuable discussions and insight during the experimental design and data analysis, and Sarah Flood Page for assistance with sequence editing and analysis. W e also thank K. Kalbitz and two anonymous reviewers for valuable comments on the manuscript. This work was supported by N ational Science Foundation (NSF) G rant DEB-0089447 (to A.R.T. and S.K.S.) and by NSF G rant DEB-0515744 (to A.R.T. and C.C.C.)

References

Abd-EI-Haleem D, M oawad H, Zaki EA, Zaki S (2002) M olecular characterization of phenol-degrading bac­teria isolated from different Egyptian ecosystems. Microb Ecol 43:217-224

Adams GA, Wall D H (2000) Biodiversity above and below the surface of soils and sediments: linkages and implications for global change. Bioscience 50: 1043-1049

Altschul SF, Gish W, Miller W, Myers EW, Lipman D J (1990) Basic local alignment search tool. J Mol Biol 215:403-410

Bardgett RD, Lovell RD, Hobbs PJ, Jarvis SC (1999) Seasonal changes in soil microbial communities along a fertility gradient of tem perate grasslands. Soil Biol Biochem 31:1021-1030

Borneman J, Triplett EW (1997) M olecular microbial diversity in soils from eastern amazonia: Evidence for unusual microorganims and microbial population shifts associated with deforestation. AppI Environ Microbiol 63:2647-2653

Buchmann N (2000) Biotic and abiotic factors controlling soil respiration rates in Picea abies stands. Soil Biol Biochem 32:1625-1635

Carney KM, Matson PA (2005) Plant communities, soil microorganisms, and soil carbon cycling: Does alter­ing the world belowground m atter to ecosystem functioning? Ecosystems 8:928-940

Chung H, Park M, Madhaiyan M, Seshadri S, Song J, Cho H, Sa T (2005) Isolation and characterization of phosphate solubilizing bacteria from the rhizosphere of crop plants of Korea. Soil Biol Biochem 37:1970- 1974

Cleveland CC, Townsend AR, Schmidt SK (2002) Phos­phorus limitation of microbial processes in moist tropical forests: Evidence from short-term laboratory incubations and field experiments. Ecosystems 5:680- 691

Cleveland CC, Neff JC, Townsend A R, H ood E (2004) Composition, dynamics and fate of leached dissolved organic m atter in terrestrial ecosystems: results from a decomposition experiment. Ecosystems 7:275-285

Cleveland CC, Reed SC, Townsend A R (2006) N utrient regulation of organic m atter decomposition in a tropical rain forest. Ecology 87:492-503

Cleveland CC, Townsend A R (2006) N utrient additions to a tropical rain forest drive substantial soil carbon dioxide losses to the atmosphere. Proc Natl Acad Sci U SA 103:10316-10321

Cole JR , Chai B, M arsh TL, Farris RJ, W ang Q, Kulan SA, Chandra S, M cGarrell DM, Schmidt TM, Garrity GM, Tiedje JM (2003) The Ribosom alDatabase Project (RDP-II): previewing a new autoaligner that allows regularupdates and the new prokaryotic taxonomy. Nucleic Acids Res 31:442-443

Davidson EA, Janssens lA (2006) Tem perature sensitivity of soil carbon decomposition and feedbacks to climate change. N ature 440:165-173

D on A, Kalbitz K (2005) Amounts and degradability of dissolved organic carbon from foliar litter at different decomposition stages. Soil Biol Biochem 37:2171-2179

Fang C, Smith P, Smith JU , M oncrief JB (2005) Incorpo­rating microorganisms as decomposers into models to simulate soil organic m atter decomposition. G eoder­ma 129:139-146

Fierer N, Bradford MA, Jackson RB (2007) Towards an ecological classification of soil bacteria. Ecology (in press)

^ Springer

Biogeochemistry

Grayston SJ, W ang S, Campbell CD, Edwards A C (1998) Selective influence of plant species on microbial diversity in the rhizosphere. Soil Biol Biochem 30:369-378

Holdridge LR, G renke WC, Hatheway WH, Liang T, Tosi JA (eds) (1971) Forest Environments in Tropical Life Zones: a pilot study. Pergamon, Oxford

H ooper DU, Bignell D E, Brown VK, Brussaard L, Dangerfleld JM, Wall DH, W ardle DL, Coleman DC, Ciller KE, Lavelle P, V an der Putten W H, De R uiter PC, Rusek J, Silver WL, Tiedje JM, W olters V (2000) Interactions between aboveground and below- ground biodiversity in terrestrial ecosystems: Patterns, mechansims, and feedbacks. Bioscience 50:1049-1061

Housman DC, Naumburg E, Huxman TE, Charlet TN, Nowak RS, Smith SC (2006) Increases in desert shrub productivity under elevated carbon dioxide vary with water availability. Ecosystems 9:374-385

H uber T, Faulkner G, Hugenholtz P (2004) Bellerophon: a program to detect chimeric sequences in multiple sequence alignments. Bioinformatics 20:2317-2319

Jackson CR, H arper IP , Willoughby D, Roden BE, Churchill PF (1997) A simple, efficient method for the separation of humic substances and D N A from environmental samples. AppI Environ Microbiol 63:4993-4995

Lane D J (1991) 16S/23S rR N A Sequencing. In: Stacke- brandt E, Goodfellow M (eds) Nucleic acid tech­niques in bacterial systematics. Wiley, Chichester, pp 115-175

Lloyd J, Taylor JA (1994) On the tem perature dependence of soil respiration. Funct Ecol 8:315-323

Ludwig W, Strunk O, W estram R et al (2004) ARB: a software environment for sequence data. Nucleic Acids Res 32:1363-1371

M eentemeyer V (1978) Macroclimate and lignin control of litter decomposition rates. Ecology 59:465^72

Monson RK, Lipson DL, Burns SP, Turnipseed AA, Delany AC, Williams MW, Schmidt SK (2006) W inter forest soil respiration controlled by climate and microbial community composition. N ature 439:711-714

More MI, Herrick JB, Silva MC, Ghiorse WC, M adsen EL (1994) Q uantitative cell-Iysis of indigenous m icroor­ganisms and rapid extraction of microbial D N A from sediment. AppI Environ Microbiol 60:1572-1580

Neff JC, A sner C P (2001) Dissolved organic carbon in terrestrial ecosystems: synthesis and a model. Ecosys­tems 4:29-48

Niisslein K, Tiedje JM (1999) Soil bacterial community shift correlated with change from forest to pasture vegetation in a tropical soil. AppI Environ Microbiol 65:3622-3626

Padm anabhan P, Padm anabhan S, D eRito C et al (2003) Respiration of ^^C-Iabeled substrates added to soil in the field and subsequent 16S rR N A gene analysis of ^^C-Iabeled soil DNA. AppI Environ Microbiol 69:1614-1622

Parton WJ, Schimel DS, Cole CV, Ojima DS (1994) A general model for soil organic m atter dynamics:

sensitivity to litter chemistry, texture and management. In: Bryant RB, A rnold RW (eds) Quantitative modeling of soil forming processes. Soil Science Society of America Special Publication 39, Madison, pp 147-167

Potts DL, Huxman TE, Cable JM, English NB, Ignace DD, Eilts JA , Mason M l, W eltzin IF , Williams D C(2006) A ntecedent moisture and seasonal precipita­tion influence the response of canopy-scale carbon and water exchange to rainfall pulses in a semi-arid grassland. New Phytol 170:849-860

Raich JW, Schlesinger W H (1992) The global carbon dioxide flux in soil respiration and its relationships to vegetation and climate. Tellus 44B:81-99

Rodriguez H, Fraga R (1999) Phosphate solubilizing bacteria and their role in plant growth promotion. Biotechnol Adv 17:319-339

Sacks WJ, Schimel DS, Monson RK, Braswell BH (2006) M odel-data synthesis of diurnal and seasonal CQ 2

fluxes at Niwot Ridge, Colorado. Global Change Biol 12:240-259

Sambrook J, Russell DW (2001) M olecular cloning: a laboratory manual. Cold Spring H arbor Laboratory Press, Cold Spring H arbor

Schlesinger W H, Andrews JA (2000) Soil respiration and the global carbon cycle. Biogeochemistry 48:7-20

Schloss PD, Handelsm an J (2005) Introducing D Q TU R , a computer program for defining operational taxonomic units and estimating species richness. AppI Environ Microbiol 71:1501-1506

Schmidt SK, Costello EK, Nemergut D R, Cleveland CC, R eed SC, W eintraub MN, Meyer AF, M artin AM(2007) Microbial turnover and seasonal succession drive biogeochemical cycles in the alpine. Ecology (in press)

Schwendenmann L, Veldkamp E, Brenes T, Q ’Brien JJ, Mackensen J (2003) Spatial and tem poral variation in soil CQ 2 efflux in an old-growth neotropical rain forest. La Selva, Costa Rica. Biogeochemistry 64:111- 128

Schwendenmann L, Veldkamp E (2005) The role of dissolved organic carbon, dissolved organic nitrogen, and dissolved inorganic nitrogen in a tropical wet forest ecosystem. Ecosystems 8:339-351

Scott-Denton LE, Sparks KL, Monson R K (2003) Spatial and tem poral controls of soil respiration rate in a high-elevation, subalpine forest. Soil Biol Biochem 35:525-534

Scott-Denton LE, Rosenstiel TN, Monson R K (2006) Differential controls by climate and substrate over the heterotrophic and rhizospheric components of soil respiration. Global Change Biol 12:205-216

Snedecor GW, Cochran W C (1989) Statistical Methods.Iowa State University, Ames

Souchie EL, Saggin-Junior QJ, Silva EM R et al (2006) Communities of P-soIubilizing bacteria, fungi and arbuscular mycorrhizal fungi in grass pasture and secondary forest of Paraty, RJ-Brazil. A n Acad Braz Cienc 78:183-193

^ Springer

Biogeochemistry

Swift MJ, Heal OW, Anderson JM (1979) Decomposition W ardle DA, Bardget RD, Klironomos JN, Setala H, Van in terrestrial ecosystems. University of California, der Putten WH, Wall D H (2004) Ecological linkagesBerkeley between aboveground and belowground biota. Sci-

Swofford D L (2001) Phylogenetic analysis using parsi- ence 304:1629-1633mony (*and other methods), 4th edn. Sinauer Asso- W inogradsky SN (1925) Etudes sur la microbiologie duciates, Sunderland, M A sol. I. Sur la method. A nn Inst Pasteur 40:455-520

^ Springer