Occurrence and degradation of peptidoglycan inaquatic environments

Niels O.G. J�rgensen a;�, Ramunas Stepanaukas b;1, Anne-Grethe U. Pedersen c,Michael Hansen a, Ole Nybroe a

a Department of Ecology, Royal Veterinary and Agricultural University, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmarkb Department of Ecology/Limnology, Lund University, S-22362 Lund, Sweden

c Department of Microbial Ecology, University of Aarhus, Ny Munkegade, DK-8000 Aarhus C, Denmark

Received 1 December 2002; received in revised form 21 March 2003; accepted 26 April 2003

First published online 27 August 2003

Abstract

Mechanisms controlling microbial degradation of dissolved organic matter (DOM) in aquatic environments are poorly understood,although microbes are crucial to global nutrient cycling. Bacterial cell wall components may be one of the keys in understanding thepresence of slowly degrading DOM in nature. We found that dominant components of bacterial cell walls (D-amino acids (D-AA),glucosamine (GluA) and diaminopimelic acid (DAPA)) comprised up to 11.4% of the dissolved organic nitrogen in 50 diverse riversentering the Baltic Sea. Occurrence of DAPA, a characteristic component of Gram-negative (G3) bacteria, in the rivers suggests thatG3 bacteria rather than Gram-positive (Gþ) were the major source of the cell wall material. In laboratory studies, the degradation ofwhole bacterial cells, cell wall material and purified peptidoglycan was studied to characterize degradation of cell wall material by naturalaquatic bacteria. Addition of whole killed G3 and Gþ bacteria to cultures of estuarine bacteria demonstrated fragmentation and loss ofcell structure of the Gþ bacteria, while the G3 bacteria maintained an intact cell shape during the entire 69-day period. In anotherexperiment, estuarine bacteria degraded 39^69% of GluA, D-AA and DAPA in added cell wall material of a representative G3 bacterialspecies during 8 days, as compared to a 72^89% degradation of GluA, D-AA and DAPA in cell material of a Gþ bacterial species. Whencultures of estuarine bacteria were enriched with purified G3 and Gþ peptidoglycan (1 mg l31), at least 49% (G3) and 58% (Gþ) of D-AAin the peptidoglycan was degraded. No major changes in GluA were obvious. Interpretation of the results was difficult as a portion of thepurified peptidoglycan was of similar size to the bacteria and could not be differentiated from cells growing in the cultures. Addition ofthe purified peptidoglycan stimulated the bacterial growth, and after 6 days the cell density in the enriched cultures was 4-fold higher thanin the controls. A regrowth of bacteria after addition of L-broth at 105 days caused a 50- to 75-fold increase in dissolved D-AA and GluA.Most of the D-AA and GluA were taken up during the following 10 days, indicating that cell wall constituents are dynamic compounds.Our results show that a variable portion of peptidoglycan in G3 and Gþ bacteria can be degraded by natural bacteria, and thatpeptidoglycan in G3 bacteria is more resistant to bacterial attack than that in Gþ bacteria. Thus, the presence of cell wall constituents innatural DOM may reflect the recalcitrant nature of especially G3 peptidoglycan.@ 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.

Keywords: Peptidoglycan; Bacterial cell wall ; Baltic river; Degradation; D-amino acid; Glucosamine; Diaminopimelic acid

1. Introduction

An important function of bacteria in nature is the deg-radation of dissolved organic matter (DOM), but bacteriathemselves also produce DOM components. Certain bac-

terial products have been found recalcitrant to microbialdecomposition, including membrane proteins [1], lipo-some-like particles formed after protozoan grazing [2],and unidenti¢ed organic molecules rich in sugars and ami-no acids (AA) [3]. In addition, bacterial cell wall compo-nents have been found to constitute a major fraction ofrecalcitrant DOM in the ocean [4].The cell wall of Gram-positive (Gþ) bacteria consists of

a thick and uniform peptidoglycan layer, forming 90% ofthe cell wall. In contrast, Gram-negative (G3) bacteriahave a complex, multilayered cell wall structure with a

0168-6496 / 03 / $22.00 @ 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.doi :10.1016/S0168-6496(03)00194-6

* Corresponding author. Tel. : +45-3528-2625; Fax: +45-3528-2606.E-mail address: nogj@kvl.dk (N.O.G. J�rgensen).

1 Present address: Savannah River Ecology Laboratory, Aiken,SC 29803, USA.

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

FEMS Microbiology Ecology 46 (2003) 269^280

www.fems-microbiology.org

relatively thin inner peptidoglycan layer (constitutes 10%of the cell wall) and an outer membrane of lipopoly-saccharides and proteins [5]. Peptidoglycan consists ofstrands of alternating L-1,4-linked N-acetylglucosamineand N-acetylmuramic acid units, cross-linked by shortpeptides [5]. AA in these peptides include D-isomers ofalanine, aspartate, glutamate, and serine, which havebeen used as markers of peptidoglycan in environmentalstudies.In marine waters and sediments, D-AA have been found

to make up 10^30% of the hydrolyzable AA, indicatingthat peptidoglycan may comprise a relatively large fractionof DOM [4,6,7]. Considering that 6 10% of all AA inliving marine bacteria are D-isomers (unpublished), andthat bacterial biomass in pelagic waters and sedimentsmakes up 6 5% of the total organic carbon pool2, a pro-portion of D-isomers s 10% in this organic matter indi-cates a 20- to 60-fold enrichment with D-AA. Most likely,this increase is caused by a selective biological degradationof L-AA relative to D-AA.The enrichment of DOM with AA that are speci¢c or

typical to bacterial cell walls may indicate a reduced deg-radation of cell wall components, but the enhanced con-centrations do not provide information on the mechanismscontrolling the degradation of, e.g. peptidoglycan. To shedlight on relations between occurrence of cell wall AA anddegradation of peptidoglycan, we surveyed occurrence ofvarious peptidoglycan components in 50 rivers discharginginto the Baltic Sea and studied degradation of cell wallcomponents by estuarine bacteria. The Baltic rivers repre-sent diverse freshwater ecosystems on a broad geograph-ical scale [8]. The rivers were analyzed for content ofglucosamine (GluA), diaminopimelic acid (DAPA), andD-isomers of alanine, aspartate, glutamate, and serine.DAPA is a peptide component of peptidoglycan in G3

bacteria but is typically replaced by lysine in Gþ bacteria[5].Microbial degradation of bacterial cell wall material was

analyzed in laboratory cultures of growing estuarine bac-teria that were amended with whole ultraviolet (UV)-killedbacteria, crude cell material or puri¢ed peptidoglycan. Cellwalls of both Gþ and G3 bacteria were included in theexperiments due to the di¡erence in structure and compo-sition of their cell walls. These di¡erences may cause apreferential degradation of intact cell walls and speci¢ccell wall components in Gþ relative to G3 bacteria. Thiswas supported by our results, which indicated that bothwhole

cells, cell walls and peptidoglycan were more resistant tomicrobial degradation in G3 than in Gþ bacteria.

2. Materials and methods

2.1. Sampling and analysis of nitrogen compounds

Water samples were collected in July 1999 at the mouthof 50 rivers discharging into the Baltic Sea. The drainagebasin of the Baltic Sea includes intensive agriculture atthe southern part, while the northern part is almost en-tirely covered by boreal forest [8]. The water sampleswere ¢ltered through 0.45-Wm ¢lters to remove particulatematerial. Details on sampling, water chemistry and con-centrations of nutrients are given in Stepanaukas et al.[8]. Dissolved organic nitrogen (DON) was measured astotal dissolved nitrogen by chemiluminescence after con-version to NOx and subtraction of dissolved inorganicnitrogen (ammonium and nitrate) [9]. Water samples foranalysis of dissolved D-AA (D-Asp, D-Glu, D-Ser andD-Ala), GluA and DAPA were hydrolyzed by a micro-wave-assisted vapor-phase technique [10] before high per-formance liquid chromatography (HPLC) analysis. Forquanti¢cation of D-AA and GluA, the hydrolyzed materialwas derivatized with o-phthaldialdehyde and N-isobutyryl-L-cysteine as a chiral agent [11] and separated on a WatersXTerra RP18 3.5-Wm particle column (Waters Corpora-tion, Milford, USA) at a £ow rate of 0.7 ml min31. Mobilephases were (A) aqueous solution of 5 mM Na2H2PO4,45 mM sodium acetate trihydrate and 7.5% methanol atpH 6.4, and (B) 100% methanol [12]. The derivatives wereeluted with the following gradient: T0 min (100% A, 0% B),T27 min (50% A, 50% B), T30 min (10% A, 90% B) andT33 min (100% A, 0% B). The o-phthaldialdehyde andN-isobutyryl-L-cysteine mixture forms £uorescent deriva-tives with GluA but not with N-acetylglucosamine. Butsince the hydrolysis converted N-acetylglucosamine toGluA, all N-acetylglucosamine in peptidoglycan was in-cluded in GluA. DAPA was detected and quanti¢ed as£uorescent primary amines after derivatization witho-phthaldialdehyde [13,14]. A standard mixture containingLL, DD and meso DAPA eluted as two peaks, which wereidenti¢ed as meso and LL+DD DAPA. Dissolved combinedamino acids (DCAA) in the rivers and the bacterial cul-tures (see below) were measured in 0.45-Wm ¢ltered sam-ples that were hydrolyzed as above. The DCAA were de-tected and quanti¢ed by HPLC using the procedure forDAPA.

2.2. Bacterial cultures

Bacterial strains used in the experimental work includedPseudomonas sp. (G3), Erythrobacter sp. (G3) and Bacil-lus sp. (Gþ). The three strains were isolated from coastalor estuarine waters in Denmark and identi¢ed by 16S

2 Assuming (1) a bacterial density of 109 bacteria l31, each with aC content of 10313 g, and a maximum of 2 mg dissolved organic carbon(DOC) l31 in pelagic waters, and (2) 109 bacteria g31 and a maximum of5% organic C in marine sediments. Data compiled from various sources.

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280270

rDNA sequencing by L. Frette, National EnvironmentalResearch Institute, Roskilde, Denmark.

2.3. Degradation of whole bacterial cell walls

Morphological changes of dead Bacillus and Erythro-bacter cells due to degradation by natural estuarine bac-teria were studied during a 2- and a half-month incubationperiod. The two bacterial strains were grown in 10% liquidZobell medium [15] until stationary phase, washed threetimes in phosphate-bu¡ered saline (PBS; 10 mMNaH2PO4 in 0.9% NaCl at pH 7.5), and killed by expo-sure to UV light for 15 min (Philips 57413 P/30 30 Wlamp). Reinoculation of subsamples of the UV-treatedcells to 10% Zobell medium con¢rmed that all cells werekilled by the light exposure. Killed cells were added toeach of triplicate 500 ml 0.8-Wm ¢ltered water from Ros-kilde Fjord, Denmark, to a ¢nal density of 3.0U106 ml31.The enrichment cultures were incubated at 20‡C on a hor-izontal shaker in the dark operating at 100 rpm. Samplesfor counting of all bacteria (estuarine bacteria and addeddead bacteria [16]) and the added dead bacteria (immuno-chemical detection, see below) were collected every thirdday.Polyclonal antibodies against Bacillus sp. and Erythro-

bacter sp. were raised by immunizing rabbits with wholeBacillus cells as in [17] or with proteinase K-digestedErythrobacter cells as in [18]. Cells immobilized on0.2-Wm pore-size polycarbonate ¢lters were detected byimmuno£uorescence microscopy essentially as in [17] usingthe primary antibodies in dilutions of 1:1000. Fluoresceinisothiocyanate (FITC)-conjugated swine anti-rabbit immu-noglobulins (Dako, Glostrup, Denmark) diluted 1:20 wereused as the secondary antibody. Bacterial cells on the ¢l-ters were recorded by a Leica TCS4d confocal laser scan-ning microscope at 1000U magni¢cation as in [19].

2.4. Degradation of crude cell wall material

Crude cell wall material of Pseudomonas sp., Erythro-bacter sp. and Bacillus sp. was obtained from culturesgrown in 10% liquid Zobell medium. The cells were har-vested in the late exponential phase, washed twice in PBSand centrifuged at 8000Ug for 5 min. Finally the cellswere sonicated (Sanyo Soniprep 150, 15 Wm amplitude)for six times 3 min in an ice bath, and subsequently auto-claved for 10 min to kill remaining living cells.The sonicated cell material was ¢ltered through 0.45-Wm

¢lters to remove larger cell fragments. In addition to cellwall material, the ¢ltrates probably also included variouscytoplasmatic substances. Cell ¢ltrates corresponding to6U109 Pseudomonas sp., 7.1U109 Erythrobacter sp. and11.3U109 Bacillus sp. were each added to triplicate cul-tures of 250 ml Roskilde Fjord water (16 psu salinity). Thewater was ¢ltered through 0.8-Wm membrane ¢lters beforethe addition of cell ¢ltrate. As controls served triplicate

250 ml 0.8-Wm ¢ltered estuarine water without additionof cell ¢ltrate. All cultures were incubated for 8 days at20‡C on a horizontal shaker in the dark at approximately100 rpm.The 0.8-Wm ¢ltration of the Roskilde Fjord water was

performed to remove bacterivorous protists. The propor-tion of estuarine bacteria removed by the ¢ltration wasnot measured, but when comparing the initial density of1.2U106 ml31 in the 0.8-Wm ¢ltrate with the typical bac-terial density in the fjord during the sampling period inJune (6^9U106 ml31 ; unpublished), the ¢ltration mayhave removed about 85% of the ambient bacteria.Samples for counting estuarine bacteria were taken dai-

ly, and total bacterial counts were obtained by epi£uores-cence microscopy after staining with acridine orange [16].Water samples for analysis of peptidoglycan componentswere collected at days 0 and 8. The water samples were0.45-Wm ¢ltered, hydrolyzed and analyzed by HPLC forcontent of D-AA, GluA and DAPA as above. At day 0some estuarine bacteria may have been included in the0.45-Wm ¢ltrates. At the end of the incubations, when sig-ni¢cantly larger cells occurred in cultures, the ¢ltratesmost likely did not include estuarine bacteria [20].

2.5. Degradation of puri¢ed peptidoglycan

Pseudomonas sp. and Bacillus sp. were grown in liquidZobell medium and harvested in the late exponentialphase. The cells were washed twice in PBS bu¡er andcentrifuged at 8000Ug for 5 min before extraction of pep-tidoglycan according to Pelz et al. [21]. Brie£y, the extrac-tion included (1) treatment with sodium dodecyl sulfate(SDS) at 100‡C for 1 h for cell lysis, (2) sonication for1 min for further cell destruction, (3) digestion with tryp-sin to remove membrane proteins, and (4) extraction oflipids with organic solvents of decreasing polarity. Thepuri¢ed peptidoglycan was analyzed for content of AA,D-AA, GluA and DAPA by HPLC as above.Microbial degradation of the puri¢ed peptidoglycan was

studied by addition of the peptidoglycan to 0.8-Wm ¢lteredestuarine water as described in Section 2.4. Triplicate cul-tures of 400 ml 0.8-Wm ¢ltered water received peptidogly-can corresponding to 1 mg l31. All cultures were incubatedfor 116 days on a horizontal shaker at room temperature(20^25‡C) at 100 rpm in the dark. At day 105, 50 WlL-broth (10 g tryptone l31, 5 g yeast extract l31, 10 gNaCl l31 and 1 g glucose l31) was added to the culturesto stimulate bacterial growth in an attempt to promotedegradation of possible residual peptidoglycan in the cul-tures. Samples for bacterial counts and for analysis ofamino compounds in the cultures were taken at variabletime intervals.The collected water was analyzed for content of AA and

GluA in un¢ltered samples (including dissolved and par-ticulate matter) and in 0.45-Wm ¢ltered samples (dissolvedfraction). Analysis of peptidoglycan components in un¢l-

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280 271

tered and ¢ltered water was performed since a portion ofthe puri¢ed peptidoglycan was found to consist of larger(s 0.45 Wm) cell fragments. This was observed by epi£uo-rescence microscopy of acridine orange-stained peptido-glycan. The staining produced weakly £uorescent cell ma-terial that ranged from almost intact cell size to minutefragments. The particulate structure of a portion of thepuri¢ed peptidoglycan implies that 0.45-Wm ¢ltered waterfrom the estuarine cultures did not include all added pep-tidoglycan. Therefore both un¢ltered and 0.45-Wm ¢lteredwater was analyzed to follow the degradation of peptido-glycan.

2.6. Comparison of peptidoglycan components in bacterialcell walls and in the Baltic rivers

The composition of AA, D-AA, GluA and DAPA inpuri¢ed peptidoglycan (Section 2.5) was related to thecomposition of peptidoglycan components in the Balticrivers to estimate whether a selective degradation of pep-tidoglycan had occurred in the rivers.

3. Results and discussion

3.1. Peptidoglycan components in 50 Baltic rivers

The concentration of D-AA and GluA in the Baltic riv-ers increased with the concentration of DON, while a dif-ferent relationship emerged between DON and DAPA.D-AA ranged from 50 nM (most of the Swedish rivers)to 300 nM N (highest level in the Estonian and Latvianrivers) (Fig. 1A). Relative to the concentration of totalDCAA in the rivers, the four D-AA on the averagemade up 5.5% (data not shown). The concentrations ofGluA exceeded the amounts of D-AA and ranged from500 (Swedish rivers) to about 2000 nM N (Finnish rivers)(Fig. 1B). For DAPA, one group of samples (13 rivers;half of the Finnish and the three Estonian rivers) hadconcentrations above 200 nM and appeared to vary inde-pendently of the DON concentrations (Fig. 1C). In theremaining rivers, DAPA concentrations were below150 nM and correlated negatively with DON (r2 = 0.605).

The exclusion of DAPA concentrations s 150 nM in theregression was chosen in an attempt to locate a relation-ship between DAPA and DON.The mechanisms causing the apparent lack of correla-

tion between DAPA and DON (DAPA concentrationss 150 nM) or the inverse correlation (DAPA concentra-tions 6 150 nM) are unknown. Obviously the degradationor preservation of this peptidoglycan-speci¢c compounddi¡ers from D-AA and GluA in natural waters. Possibly,

Fig. 1. Concentrations of dissolved D-AA (sum of D-Asp, D-Glu, D-Serand D-Ala; A), GluA (B), and DAPA (sum of meso and LL+DD iso-mers; C) relative to DON in 0.45-Wm ¢ltered water from 50 rivers dis-charging into the Baltic Sea. Numerals in the text box in A refer to thenumber of rivers sampled in each country. The linear regressions werecalculated from all shown concentrations, except in C where DAPAconcentrations s 150 nM (data points within the dotted line) were ex-cluded. The samples were collected at the mouth of the rivers in July1999. Details on sampling, water chemistry and concentrations of nu-trients are given in Stepanaukas et al. [8]. MeansT 1 S.D. shown. n=3.

Table 1Proportion of D-AA, DAPA and GluA at low and high DON concentrations in 50 Baltic rivers and in puri¢ed peptidoglycan of Pseudomonas sp. andBacillus sp.

Low DON (5 WM N) High DON (55 WM N) Relative composition of peptidoglycan inPseudomonas sp. and Bacillus sp.

Relative composition of DON in the Balticrivers vs. composition of peptidoglycan

% of DON Ratio % of DON Ratio Ratio Ratio

D-AA 0.37 T 0.06 1 0.51T 0.21 1 1 1DAPA 1.0T 0.35 2.7 0.08T 0.06 0.22 0.083^0.066 32.5 (low DON)^3.33 (high DON)GluA 0.28 T 0.03 7.6 10.8T 4.10 29.2 0.178^0.406 42.7 (low DON)^71.9 (high DON)

Mean concentrations including 95% con¢dence interval ranges of D-AA, DAPA and GluA in the rivers were determined from the linear regressions inFig. 1. Only DAPA concentrations 6 150 nM were included in the regression. Content of DAPA and GluA in the rivers and the bacteria was normal-ized to D-AA. The relative composition of peptidoglycan components was based on mean values of each bacterium in Table 2.

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280272

dissolved DAPA, alone or associated with other peptido-glycan components, reacts with organic substances insome rivers and forms biologically refractory complexes.Although DAPA has been found preserved in ancient stro-matolites [22], there are no reports of DAPA being recal-citrant to biological degradation and thus accumulate innature.D-AA, GluA, and DAPA together comprised from 1.65

to 11.4% of DON in the rivers, indicating that bacterialcell wall components make up a sizable fraction of fresh-water DON (regressions in Fig. 1, Table 1). This contri-bution of cell wall components is probably an underesti-mate of the share of peptidoglycan components in DON,since peptidoglycan also contains L-AA and other N com-pounds such as muramic acid. On the other hand, chitin(structural polymer in invertebrate exoskeletons and fun-gal cell walls [23]) may provide an additional source ofGluA to the rivers.The present amounts of GluA were 10^30 times higher

than concentrations published for high molecular mass(HMM) DOM (1^100 nm size) in the oceans [24]. How-ever, relative to HMM DON, GluA in the oceanic samplesmade up about 0.5^5%, which agrees with the range mea-sured in the Baltic rivers (0.28^10.8%; Table 1). Thepresent DON pools were not separated into di¡erent mo-lecular size classes, but assuming that most GluA residedin riverine HMM, e.g. in cell wall fragments, there is aclose agreement between the riverine and the oceanic en-vironments. The similarity may indicate that there was aminor terrestrial in£uence on the production of GluA inthe rivers, e.g. due to a mainly bacterial origin of GluA.The abundance of GluA in the rivers may also supportthat GluA (and galactosamine) are important bacterialderived components of refractive DOM, as has been sug-gested for marine environment [3].When normalized to D-AA, the content of DAPA and

GluA was 3.33- to 32.5-fold and 42.7- to 71.9-fold higher,respectively, than in pure peptidoglycan of the bacteriaBacillus sp. and Pseudomonas sp. (Table 1). Thus, com-pared to the composition of AA in intact peptidoglycan(Tables 1 and 2), DON in the rivers was signi¢cantly en-

riched with DAPA and GluA, relative to D-AA. In thiscomparison we did not consider a potential input of chi-tin-derived GluA, as well as selective degradation andabiotic sorption of peptidoglycan components may haveoccurred in the rivers. These processes could, togetherwith biological degradation, have in£uenced the observedenrichment with GluA and DAPA in the rivers.Previous analyses of dissolved AA enantiomers in nat-

ural waters are few and indicate a variable occurrence. Inmarine waters, D-AA have been found either to be insig-ni¢cant (estuarine colloidal and particulate material [25])or constitute a major portion of high molecular DON(oceanic environments [4]). No data on peptidoglycancomponents in freshwater are available, but the presentstudy indicates that peptidoglycan components may ac-count for a sizable portion of riverine DON.Peptidoglycan in the Baltic rivers may originate from

aquatic as well as terrestrial environments, as D-AA havebeen shown to be released during £ooding of soil [26]. Thebiological availability of peptidoglycan components in thepresent rivers was not tested, but previous studies indicatethat only a small fraction of riverine D-AA is degradableby microorganisms [26]. A recalcitrant nature of bacterialcell wall material is supported by the occurrence of non-degraded peptidoglycan components in the Baltic rivers.

3.2. Integrity of whole bacterial cell walls

The riverine samples con¢rm that peptidoglycan and itsconstituents are not completely degraded by microorgan-isms, as has previously been observed in marine environ-ments [4,7]. In order to determine to which extent bacterialcell walls and their peptidoglycan components are degrad-able by natural aquatic bacteria, the persistence of wholebacterial cell walls, cell wall material and puri¢ed peptido-glycan were studied in cultures of actively growing estua-rine bacteria.The fate of whole, UV-killed cells of Erythrobacter sp.

(G3) and Bacillus sp. (Gþ), which were added to the es-tuarine cultures, was determined by immuno£uorescencemicroscopy. A progressive fragmentation of Bacillus sp.cell walls was observed during the 69-day period (Fig. 2,left panel). Cell fragments began to aggregate into amor-phous structures after 3 days, and only clumps of celldebris remained present after 69 days. In contrast, thecell wall structure of Erythrobacter sp. remained intactduring the entire incubation period (Fig. 2, right panel).During the experiment the total number of bacteria in

the estuarine cultures (estuarine bacteria and stainableErythrobacter sp. and Bacillus sp. cells) increased fromabout 4U106 cells ml31 at start to 15U106 cells ml31

(amendment with Bacillus sp.) and 9.6U106 cells ml31

(amendment with Erythrobacter sp.) at day 8. At day 69,cell densities of about 2U106 ml31 (Bacillus sp.) and3.5U106 ml31 (Erythrobacter sp.) were found (data notshown).

Table 2Content of four D-AA, GluA and DAPA relative to the total D- andL-AA content in peptidoglycan of Bacillus sp. and Pseudomonas sp.

% of all L+D-AA

Bacillus sp. Pseudomonas sp.

D-Asp 1.68 T 0.37 1.91 T 0.21D-Glu 6.74 T 0.34 1.33 T 0.09D-Ser 1.06 T 0.20 0.82 T 0.08D-Ala 5.55 T 0.76 0.87 T 0.26GluA 6.10 T 0.56 0.88 T 0.11DAPA 0.99 T 0.04 0.41 T 0.07

The total L- and D-AA concentrations were determined with the OPAderivatization procedure and do not include proline and cysteine.MeanT 1 S.D. shown. n=3

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280 273

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280274

The experiment showed that the estuarine bacteria rap-idly degraded the cell walls of Bacillus sp., but did notdegrade Erythrobacter sp. cell walls to a comparable ex-tent. The employed polyclonal antibodies target severalcell wall components. Therefore, the immuno£uorescencereaction monitors morphological changes rather than di-rect degradation of peptidoglycan during the incubation.However, peptidoglycan is responsible for the strength andshape of the Gþ cell wall [5]. Consequently the morpho-logical changes of the Bacillus sp. cells shown in Fig. 2most likely re£ect degradation or severe alteration of pep-tidoglycan. We speculate that lipopolysaccharides andproteins in the G3 Erythrobacter sp. outer membranemight protect its peptidoglycan against enzymatic attack,or that G3 peptidoglycan is not as readily degraded asGþ peptidoglycan.The application of UV treatment to kill the bacteria was

chosen as this procedure was expected to in£ict the leastpossible damage to the cell walls. Light-induced e¡ects onbacteria appear restricted to nucleic acids [27], but it can-not be excluded that the UV light had some e¡ects ondegradability of the cell wall components. Peptidoglycanin Bacillus sp. spores has been found to resist extensiveUV radiation but the peptidoglycan structure in sporesand whole bacterial cells may not be similar [28].In nature, morphologically intact but dead bacterial

cells originate from mortality that only causes minor orno damage on the cell wall, e.g. starvation, predatory bac-teria, antibiosis and possibly viral infection [29]. Bacterial‘ghosts’ or nucleoid-less bacteria make up a substantialfraction (61^88%) of the total bacterial numbers in somemarine waters [30]. Empty ‘cell sacs’ have also been ob-served in a marine sediment [31]. Our study suggests thatcell walls of mainly G3 bacteria may retain their structuralintegrity for prolonged periods and thus contribute to the‘ghost’ population in the environment.

3.3. Degradation of crude cell wall material

The higher stability of G3 than of Gþ cells shown inFig. 2 may indicate that G3 peptidoglycan better resistedmicrobial degradation than did Gþ peptidoglycan. To testthis hypothesis, cultures of estuarine bacteria wereamended with suspensions of ultrasonicated and 0.45-Wm¢ltered cell wall material of Pseudomonas sp. and Eryth-robacter sp. (G3) and Bacillus sp. (Gþ). The amendmentsgave rise to a signi¢cant growth of estuarine bacteria.Hence, the initial density of 1.2U106 cells ml31 increasedduring the 8-day incubation to 54, 65 and 101U106 bac-teria ml31 in the Pseudomonas sp., Erythrobacter sp. andBacillus sp. amended cultures, respectively (Fig. 3A). Inthe controls the ¢nal density was 7.0U106 bacteria ml31.

Simultaneous with the cell growth, the added peptido-glycan was reduced. After 8 days, 72% of D-AA and 89%of GluA were degraded in the cultures amended with Ba-cillus sp., while 51 and 45% (D-AA), and 69 and 65%(GluA) were degraded in the Pseudomonas sp. and Eryth-robacter sp. cultures, respectively (Fig. 3B and C). Simi-larly, DAPA was reduced by 77% in the cultures amendedwith Bacillus sp., but only by 44 and 39% in the cultureswith Pseudomonas sp. and Erythrobacter sp. (Fig. 3D). AllAA concentrations were measured in the dissolved(6 0.45-Wm) fraction. The results indicate that peptidogly-can components derived from G3 cell walls were moreresistant to microbial degradation than components ofGþ cell walls. No changes in D-AA, GluA or DAPAwere found in the control cultures without enrichmentwith cell material.

6

Fig. 2. Confocal microscopy recordings of immunochemical slides showing morphological changes of dead Bacillus sp. (Gþ) and Erythrobacter sp. (G3)during a 69-day incubation in cultures of estuarine bacteria.

Fig. 3. Growth of estuarine bacteria and changes in concentrations ofspeci¢c peptidoglycan components after enrichment with crude cell wallmaterial of Pseudomonas sp. (G3), Erythrobacter sp. (G3) and Bacillussp. (Gþ) during an 8-day period. Increase in bacterial density (A) anduptake of four D-AA (B), GluA (C) and meso and LL+DD DAPA (D)are shown in the panels. Means T 1 S.D. shown. n=3 (one sample fromeach culture).

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280 275

In addition to D-AA, GluA and DAPA, the cell wallmaterial also included DCAA, originating from bacterialproteins and the estuarine water. The initial DCAA con-centrations were 7-fold higher (mean value) than the D-AAconcentrations (Table 3). At day 8, the DCAA concentra-tions were only 2.8-fold higher (mean value) than theD-AA, indicating that DCAA were taken up to a largerextent than D-AA by the bacteria. In the controls, theDCAA were reduced by 40%. The concentration ofDON in the estuarine water was 37 WM N (data notshown). Changes in DON due to addition of peptidogly-can and during the experiments were not followed.The amount of Bacillus sp. cell wall material added to

the cultures was 2.5- to 3-fold higher than that of the othertwo bacteria (exempli¢ed by the concentrations of D-AA,GluA and DAPA at day 0). This di¡erence in cell wallmaterial may have in£uenced the rate of degradation, butit appeared not to a¡ect the overall degradation of D-AA,GluA and DAPA, as comparable, residual concentrationsof these amino compounds were present in the three cul-tures at day 8. Thus, for testing degradability of G3 andGþ cell wall components, the di¡erent amounts of cellwall material seemed not to introduce any experimentalbiases.Interestingly, ultrasonication released about 80% of the

total DAPA content in Bacillus sp. into the 6 0.45-Wmfraction, while less than 30% of DAPA in Pseudomonassp. and Erythrobacter sp. was released. This demonstratesthat DAPA is more ¢rmly integrated into the G3 lipo-polysaccharide^protein cell wall matrix, than in the pepti-doglycan-dominated Gþ cell wall.The relative high abundance of DAPA in Bacillus sp.

(made up about 1% of all peptidoglycan AA; Table 2) wasunexpected as this AA according to literature is a typicalcomponent of G3 peptidoglycan and is replaced by lysinein Gþ bacteria [5]. However, DAPA does occur in pepti-doglycan of some Gþ bacteria [32] and is used as a chemo-taxonomic marker in Actinobacteria (Gþ bacteria; [33]).The occurrence of DAPA in Gþ bacteria in natural envi-ronments is unknown. However, the low degradation ofG3 peptidoglycan observed in this study and the domi-nance of G3 bacteria in aquatic environments [34] bothsupport that the abundant amounts of DAPA in the Balticrivers most likely originated from G3 bacteria.

3.4. Degradation of puri¢ed peptidoglycan

The previously shown modi¢cations of whole bacterialcell walls and the degradation of crude cell wall materialsuggest that Gþ peptidoglycan is less resistant to microbialdegradation than that of G3 bacteria. To test if pure pep-tidoglycan of Gþ bacteria actually is more labile than thatof G3 bacteria, puri¢ed peptidoglycan of Pseudomonas sp.and Bacillus sp. was added to cultures of estuarine bacte-ria. This experiment was followed for 116 days. Resultsfrom the nutrient addition on day 105 are given after pre-sentation of results from the initial 105-day period.

3.4.1. Changes in bacterial growth and peptidoglycancomponents between days 0 and 105

The addition of peptidoglycan signi¢cantly stimulatedgrowth of the estuarine bacteria. After 6 days, the bacte-rial abundance had increased from initially 0.15 to2.4U106 cells ml31 (controls without peptidoglycan) and0.85 to 8.3 and 9.9U106 cells ml31 in the cultures withpeptidoglycan of Pseudomonas sp. and Bacillus sp., respec-tively (Fig. 4A). For unknown reasons, the initial bacterialdensity was lower in the controls than in the peptidogly-can-enriched cultures. However, the about 4-fold higherbacterial abundance in the enriched cultures than in thecontrols at day 6 demonstrates that peptidoglycan sus-tained a larger bacterial growth. At 105 days, the bacterialabundance in the peptidoglycan-enriched cultures had de-creased to 1.1^1.2U106 cells ml31 and 0.3U106 cells ml31

in the controls.The initial amounts of D-AA in the un¢ltered water

(about 1000 nM; Fig. 4B) and 0.45-Wm ¢ltered water(75^150 nM; Fig. 4C) demonstrate that the majority ofpeptidoglycan-derived AA resided in the particulate frac-tion (s 0.45 Wm). The un¢ltered water includes the puri-¢ed peptidoglycan, naturally occurring peptidoglycan inthe fjord water, and peptidoglycan in the estuarine bacte-ria; the 0.45-Wm ¢ltered water includes the corresponding,dissolved peptidoglycan pools (expected to be free of bac-teria) in the cultures. During the following 4 days, theamount of D-AA in the un¢ltered water declined, reducingthe concentration by 510 nM (Bacillus sp.) and 390 nM(Pseudomonas sp.). Concentration changes in the controlswere not statistically signi¢cant (Ps 0.05; t-test).

Table 3Initial and ¢nal concentrations of D-AA and DCAA during degradation of crude cell wall material (days 0 and 8) and puri¢ed peptidoglycan (days 0and 25a)

Degradation of crude cell wall material (Fig. 3) Degradation of pure peptidoglycan (Fig. 4)

Pseudomonas sp. Erythrobacter sp. Bacillus sp. Control Pseudomonas sp. Bacillus sp. Control

Day 0 D-AA 889T 57 1 245T 231 2 950T 352 219T 17 86T 21 140T 21 65T 27DCAA 5 994T 741 7 618T 521 21 063T 1921 1 633T 102 1 147T 82 1 459T 153 1 063T 118

Day 8 or 25 D-AA 392T 68 611T 129 845T 120 258T 48 168T 31 142T 23 112T 32DCAA 1 447T 187 1 390T 215 2 274T 321 981T 157 1 529T 193 1 568T 179 1 352T 83

All concentrations are given in nM concentrations of 0.45-Wm ¢ltered samples. S.D. shown; n=3.aEnd of the initial sampling period.

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280276

The reduction of the added peptidoglycan re£ects bac-terial uptake of D-AA, but the actual amounts taken upcannot be determined as the concentration changes includeboth reduction of the added and naturally occurring pep-tidoglycan, and the increase in peptidoglycan in the grow-ing estuarine bacteria. The reduction in D-AA concentra-tions corresponded to a decline of 39% (Pseudomonas sp.)and 49% (Bacillus sp.) of the initial peptidoglycan D-AA.Between days 4 and 25, D-AA in the un¢ltered sampleseither increased due to the ongoing bacterial production(Bacillus sp.) or remained largely unchanged (Pseudomo-nas sp. and controls).In the 0.45-Wm ¢ltered water, an initial reduction of

D-AA of 140^90 nM was observed for Bacillus sp. pepti-doglycan, but otherwise D-AA were released and increasedthe concentrations from initially about 90 to 170 nM onday 25 (Fig. 4C). On days 9, 16 and 25, concentrations ofD-AA were 16^27% higher in the cultures with Pseudomo-nas sp. than with Bacillus sp. peptidoglycan. The additionof peptidoglycan to the cultures also included DCAA, buta higher DCAA concentration was only found in the Ba-cillus sp. amendment (27% more DCAA at day 0) (Table3). During the 25-day incubation period, the DCAA con-centration slightly increased (Pseudomonas sp. cultures andcontrol) or remained unchanged (Bacillus sp. cultures).The results indicate that there was no net uptake ofDCAA in the estuarine water. The DON concentrationin the estuarine water was 49 WM N, but changes in con-centrations were not measured in the experiment.Presence of another peptidoglycan component, GluA, in

the un¢ltered samples increased 2.2-fold (controls) and upto 7.6-fold (Bacillus sp.) during the initial 25 days, fol-lowed by a decline until day 105 (Fig. 4D). The initialamounts of GluA were rather similar in the controls(230 nM) and in the cultures enriched with peptidoglycan(250^350 nM), suggesting that the added peptidoglycanhad a low content of GluA. Changes in concentrationsof dissolved GluA in the 6 0.45-Wm ¢ltered samplesshowed that a production as well as an uptake of GluAoccurred. During the initial 16 days, the amount of dis-solved GluA increased 2.0- to 3.75-fold (concentrationrange of 120^550 nM), but was followed by a decrease(Fig. 4E).

3.4.2. Changes in bacterial growth and peptidoglycancomponents between days 105 and 116

The addition of growth medium (L-broth) on day 105caused a considerable regrowth of the estuarine bacteria.The cell density increased from about 106 cells ml31 (orbelow) to 0.7^2.5U108 cells ml31 on day 116 (Fig. 4A).Simultaneous with the bacterial growth, the concentra-tions of D-AA in the un¢ltered water increased from about500 nM to levels of 3000^5000 nM on day 107 (Fig. 4B).These D-AA were reduced by 43^51% on day 112 in theBacillus sp. and control cultures ; no changes occurred inthe Pseudomonas sp. cultures. Similarly, in the 6 0.45-Wm

Fig. 4. Growth of estuarine bacteria and changes in concentrations offour D-AA (D-Asp, D-Glu, D-Ser and D-Ala) and GluA in un¢ltered and0.45-Wm ¢ltered water after enrichment with puri¢ed peptidoglycan ofPseudomonas sp. (G3) and Bacillus sp. (Gþ) (equivalent to 1 mg l31)during a 116-day period. The cultures were enriched with L-broth atday 105. Increase in bacterial density (A), D-AA in un¢ltered (B) and0.45-Wm ¢ltered water (C), and GluA in un¢ltered (D) and 0.45-Wm ¢l-tered water (E) are shown in the panels. Means T 1 S.D. shown. n=3(one sample from each culture).

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280 277

¢ltered water, the D-AA content increased from 50^100nM at day 105, to 3500^5000 nM at day 107 (Fig. 4C).However, these D-AA were signi¢cantly reduced the fol-lowing days, concurrent with the increase in bacterial den-sity.The nutrient enrichment also led to a signi¢cant raise in

GluA. In the un¢ltered water, the concentration of GluAincreased 20- to 42-fold after day 105 (Fig. 4D). In con-trast to D-AA in the un¢ltered water, there were no indi-cations of a bacterial utilization of ‘un¢ltered’ GluA in thecultures. In the 6 0.45-Wm fraction, the concentration ofGluA increased to about 15 000 nM (Fig. 4E). Interest-ingly, largely all (77^90%) of this GluA was assimilatedby the growing bacteria during the following 9 days. As-suming a bacterial carbon content of 10313 g cell31 and acarbon content of GluA of 72 ng C nM31, the uptake ofdissolved GluA corresponded to approximately 10% of thenet bacterial C production.

3.4.3. Degradation and production of peptidoglycancomponents in the 116-day experiment

The main conclusions from the enrichment with pepti-doglycan are (1) that about half of the D-AA in peptido-glycan were degradable, (2) that D-AA of peptidoglycanfrom the Gþ Bacillus sp. were assimilated to a larger ex-tent than D-AA of the G3 Pseudomonas sp., and (3) thatmore D-AA of Pseudomonas sp. than of Bacillus sp. accu-mulated in the 6 0.45-Wm fraction during the initial25-day period. These observations support that Gþ cellwalls, including their peptidoglycan, are more labile thancell walls of G3 bacteria. The almost identical amounts ofadded peptidoglycan of the two bacteria were re£ected insimilar bacterial abundances within the initial 25 days. Theinitial concentrations of dissolved D-AA and GluA weremore variable, but obviously they did not in£uence thebacterial production within the same period.Our experiments do not reveal whether peptidoglycan

can be totally degraded by bacterial processes, as theadded peptidoglycan could not be size-separated frompeptidoglycan in living bacteria. However, the residualconcentrations of dissolved D-AA (about 100 nM) andGluA (about 200 nM) after 105 days of incubation wereclose to the ambient concentrations before the addition ofpeptidoglycan. This suggests that most of the peptidogly-can in the dissolved fraction, being added or producedduring the incubation (Fig. 4C and E), was degraded.The residual amounts of D-AA and GluA after 105 days

show that portions or constituents of peptidoglycan areresistant to bacterial degradation. Possibly, due to theirpolymer structure, these compounds are not easily de-graded by bacterial enzymes, such as hydrolases, as hasbeen suggested to explain the presence of peptidoglycan innatural environments [35]. The semilabile properties ofpeptidoglycan observed in our study are supported bylong turnover times (10^167 days) recently observed forradiolabelled G3 peptidoglycan in marine waters [36].

The study further demonstrated a slower degradation ofthe polysaccharide than the peptide moiety of G3 pepti-doglycan. If the similar is true of freshwater environments,a more recalcitrant nature of saccharide components inpeptidoglycan may explain the higher abundance ofGluA than of D-AA in the rivers (Fig. 1).The release of dissolved D-AA and GluA following the

regrowth of bacteria after enrichment with L-broth showsthat fast-growing bacteria may release cell wall material,the process being deliberate or accidental. The subsequentre-assimilation of D-AA and GluA con¢rms that bacteriacan take up extracellular peptidoglycan components.These D-AA and GluA most likely resided in intact orfragments of peptidoglycan, as a chromatographic analysisdemonstrated that they did not occur as free monomers(data not shown). Once degraded to free molecules, e.g. bybacterial ectoenzymes, both D-AA and GluA are taken upby bacterial populations [37^39].An additional process that might have in£uenced the

AA pools in the cultures is virus-induced lysis of the bac-teria. Weinbauer and Peduzzi [40] observed an increase indissolved AA following addition of virus-rich material tonatural marine cultures. If lysogenic cells, or viruses andtheir bacterial hosts, were present in the peptidoglycan-enriched cultures, production of cell fragments from bac-terial lysis, also including peptidoglycan components, mayhave contributed to the observed increase of both dis-solved D-AA and GluA during the initial 25-day period.The importance of viral infection on the degradation ofpeptidoglycan in natural waters has not been studied, butits signi¢cance may be underestimated and deserves fur-ther analysis.

3.5. Enzymatic reactions and peptidoglycan degradation

In contrast to the addition of whole bacterial cells andcrude peptidoglycan to cultures of estuarine bacteria (Figs.2 and 3), the addition of puri¢ed peptidoglycan did notinclude additional proteins (as structural constituents orenzymes), lipids or other cell components. The signi¢canceof these non-peptidoglycan products on the degradation ofpeptidoglycan in whole cells and crude cell material can-not be estimated from the present experiments. Possibly,non-peptidoglycan compounds may have served as nu-trients and stimulated the bacterial growth and with thisthe activity of enzymes involved in degradation of pepti-doglycan.Enzymatic processes involved in uptake of extracellular

peptidoglycan are still unidenti¢ed. Among bacterial en-zymes known to be active in peptidoglycan degradationare hydrolases that break covalent bonds in peptidogly-can. These enzymes are typically located in the cell wall[41,42]. Another process for peptidoglycan hydrolysis isfound in bacteriophages that produce lytic enzymes whichcreate small openings in peptidoglycan [43]. If similar cell-bound enzyme systems are responsible for degradation of

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280278

extracellular peptidoglycan, a physical contact between thebacterial cells and peptidoglycan is necessary, but themechanisms are unknown.An alternate pathway for degradation of peptidoglycan

may involve a suite of extracellular enzymes such as pep-tidases (for removal of the peptide interbridges) and lyso-zyme-like enzymes (for breakage of the L(1-4) bonds be-tween N-acetylglucosamine and N-acetylmuramic acid).Both peptidases and L(1-4)-cleaving enzymes are com-monly found in natural waters [44,45].Another aspect that may in£uence the enzymatic attack

on peptidoglycan is composition of the extracted peptido-glycan. In our study, the cell wall material was digestedwith trypsin to remove membrane proteins [21]. Otherstudies, e.g. Nagata et al. [36], did not include an enzy-matic treatment of the cell wall material. We did not testto which extent trypsin actually removed membrane pro-teins or perhaps even modi¢ed peptidoglycan. If a physicalcontact between extracellular peptidoglycan and cell-bound peptidoglycan hydrolases is required, presence ofmembrane proteins may in£uence the enzymatic activity.The present lack of standard procedures for extraction ofpeptidoglycan restricts a comparison of results on degra-dation of peptidoglycan and deserves more attention.

3.6. Pre¢ltering and bacterial grazing

In an attempt to eliminate protist grazing in the degra-dation experiments, estuarine bacteria in 0.8-Wm ¢lteredrather than natural, un¢ltered water was used in the cul-tures. When present in bacterial laboratory cultures, bac-terivorous protists may signi¢cantly reduce bacterial pop-ulations [46]. In the present experiments, utilization of cellwall material by the estuarine bacteria was expected toresult in increased bacterial densities. However, with graz-ing protists present, the bacterial density might not mirrorthe actual bacterial production. The 0.8-Wm pre¢ltrationwas estimated to exclude about 85% of the bacteria, in-cluding mainly attached and large bacteria. This meansthat bacteria in the cultures represent only a minor portionof the natural populations and that the observed degrada-tion of cell wall material may not be directly comparableto the degradation at in situ conditions. The ¢ltration mayintroduce a major bias if attached and large bacteria aredominant organisms in the degradation of cell wall mate-rial. This was not tested in the present experiments.

3.7. Conclusions and perspectives

The occurrence of peptidoglycan components in the Bal-tic rivers, the ocean [4], and marine sediments [7,47] cor-roborates that peptidoglycan is not completely degradableby microorganisms and contributes to the pool of recalci-trant organic matter. Our studies on degradation of pep-tidoglycan support that a fraction of peptidoglycan, exem-pli¢ed by D-AA and GluA, is resistant to bacterial

degradation. Results from the experiments further indicatethat cell walls of G3 bacteria are more resistant to degra-dation than those of Gþ bacteria. This may be related tothe di¡erences in peptidoglycan structure between G3 andGþ bacteria.The present knowledge on occurrence and cycling of

peptidoglycan in natural environments is very restrictedand requires additional research. Our observations on pep-tidoglycan components in the Baltic rivers and the degra-dation of cell wall material demonstrate that althoughpeptidoglycan-derived components are not dominant ele-ments of DOM, they may be important in shaping thecomposition and reactivity of natural DOM.

Acknowledgements

We thank R.E. Jensen for technical assistance, L. Frettefor donating bacterial cultures, and professors D. Burdige,C. Klausen and M.A. Moran for valuable comments onthe manuscript. The study was supported by Danish Nat-ural Science Research Council, MISTRA, the foundationOscar och Lili Lamms Minne, the EU TMR research net-work ‘Wetland Ecology and Technology’ and the SwedishInstitute.

References

[1] Tanoue, E. (1995) Detection of dissolved protein molecules in oceanicwaters. Mar. Chem. 51, 239^252.

[2] Nagata, T. and Kirchman, D.L. (2001) Bacterial mortality: A path-way for the formation of refractory DOM? In: Microbial Biosys-tems: New Frontiers. Proceedings of the 8th International Symposi-um on Microbial Ecology Vol. 1, pp. 153^158.

[3] Ogawa, H., Amagai, Y., Koike, I., Kaiser, K. and Benner, R. (2001)Production of refractory dissolved organic matter by bacteria. Sci-ence 292, 917^920.

[4] McCarthy, M.D., Hedges, J.I. and Benner, R. (1998) Major bacterialcontribution to marine dissolved organic nitrogen. Science 281, 231^234.

[5] Madigan, M.T., Martinko, J.M. and Parker, J. (2000) BROCK Biol-ogy of Microorganisms, 9th Edn., pp. 1^991. Prentice-Hall, UpperSaddle River, NJ.

[6] Lee, C. and Bada, J.L. (1977) Dissolved amino acids in the equatorialPaci¢c, the Sargasso Sea, and Biscayne Bay. Limnol. Oceanogr. 22,502^510.

[7] Pedersen, A.G.U., Thomsen, T.R., Lomstein, B.Aa. and J�rgensen,N.O.G. (2001) Bacterial in£uence on amino acid enantiomerization ina coastal marine sediment. Limnol. Oceanogr. 46, 1358^1369.

[8] Stepanauskas, R., J�rgensen, N.O.G., Eigaard, O.R., Zvikas, A.,Tranvik, L.J. and Leonardson, L. (2002) Summer inputs of riverinenutrients to the Baltic Sea: Bioavailability and eutrophication rele-vance. Ecol. Monogr. 72, 579^597.

[9] Burdige, D.J. and Zheng, S.L. (1998) The biogeochemical cycling ofdissolved organic nitrogen in estuarine sediments. Limnol. Oceanogr.43, 1796^1813.

[10] J�rgensen, N.O.G. and Jensen, R.E. (1997) Determination of dis-solved combined amino acids using microwave-assisted hydrolysisand HPLC precolumn derivatization for labeling of primary and sec-ondary amines. Mar. Chem. 57, 287^297.

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280 279

[11] Bru«ckner, H., Haasmann, S., Langer, M., Westhauser, T., Wittner,R. and Godel, H. (1994) Liquid-chromatographic determination ofD-amino and L-amino acids by derivatization with o-phthaldialdehydeand chiral thiols. J. Chromatogr. A 666, 259^273.

[12] Mopper, K. and Furton, K.G. (1999) Extraction and analysis ofpolysaccharides, chiral amino acids, and SFE-extractable lipidsfrom marine POM. Geophys. Monogr. 63, 151^160.

[13] Lindroth, P. and Mopper, K. (1979) High performance liquid chro-matographic determination of subpicomole amounts of amino acidsby precolumn £uorescence derivatization with o-phthaldialdehyde.Anal. Chem. 51, 1667^1674.

[14] J�rgensen, N.O.G., Kroer, N., Co⁄n, R.B., Yang, X.H. and Lee, C.(1993) Dissolved free amino acids, combined amino acids, and DNAas sources of carbon and nitrogen to marine bacteria. Mar. Ecol.Prog. Ser. 98, 135^148.

[15] ZoBell, C. (1946) Marine Microbiology; a Monograph on Hydro-bacteriology. Chronica Botanica Co., Waltham, MA.

[16] Hobbie, J.E., Daley, R.J. and Jasper, S. (1977) Use of nucleopore¢lters for counting bacteria by £uorescence microscopy. Appl. Envi-ron. Microbiol. 33, 1225^1228.

[17] Nybroe, O., Christo¡ersen, K. and Riemann, B. (1992) Survival ofBacillus licheniformis in seawater model ecosystems. Appl. Environ.Microbiol. 58, 252^259.

[18] Middelboe, M., J�rgensen, N.O.G. and Kroer, N. (1996) E¡ects ofviruses on nutrient turnover and growth e⁄ciency of noninfectedmarine bacterioplankton. Appl. Environ. Microbiol. 62, 1991^1997.

[19] Hansen, M., Kragelund, L., Nybroe, O. and S�rensen, J. (1997) Earlycolonization of barley roots by Pseudomonas £uorescens studied byimmuno£uorescence technique and confocal laser scanning microsco-py. FEMS Microbiol. Ecol. 23, 353^360.

[20] Kroer, N. (1993) Bacterial growth e⁄ciency on natural dissolvedorganic matter. Limnol. Oceanogr. 38, 1282^1290.

[21] Pelz, O., Cifuentes, L.A., Hammer, B.T., Kelley, C.A. and Co⁄n,R.B. (1998) Tracing the assimilation of organic compounds usingdelta C-13 analysis of unique amino acids in the bacterial peptido-glycan cell wall. FEMS Microbiol. Ecol. 25, 229^240.

[22] Borruat, G., Roten, C.A.H., Marchant, R., Fay, L.B. and Karamata,D. (2001) Chromatographic method for diaminopimelic acid detec-tion in calcareous rocks ^ Presence of a bacterial biomarker in stro-matolites. J. Chromatogr. A 922, 219^224.

[23] Horton, H.R., Moran, L.A., Ochs, R.S., Rawn, J.D. and Scrimgeour,K.G. (2002) Principles of Biochemistry, 3rd Edn., pp. 1^862. Pren-tice-Hall, Upper Saddle River, NJ.

[24] Benner, R. and Kaiser, K. (2003) Abundance of amino sugars andpeptidoglycan in marine particulate and dissolved organic matter.Limnol. Oceanogr. 48, 118^128.

[25] Sigleo, A.C., Hare, P.E. and Helz, G.R. (1983) The amino acid com-position of estuarine colloidal material. Estuar. Coast. Shelf Sci. 17,87^96.

[26] Stepanauskas, R., Laudon, H. and J�rgensen, N.O.G. (2000) HighDON bioavailability in boreal streams during a spring £ood. Limnol.Oceanogr. 45, 1298^1307.

[27] Je¡rey, W.H., Pledger, R.J., Aas, P., Hager, S., Co⁄n, R.B., VonHaven, R. and Mitchell, D.L. (1996) Diel and depth pro¢les of DNAphotodamage in bacterioplankton exposed to ambient solar ultravio-let radiation. Mar. Ecol. Prog. Ser. 137, 283^291.

[28] Atrih, A. and Foster, S.J. (2002) Bacterial endospores the ultimatesurvivors. Int. Dairy J. 12, 217^223.

[29] Fuhrman, J.A. and Noble, R.T. (2001) Causative agents of bacterialmortality and the consequences to amine food webs. In: Microbial

Biosystems: New Frontiers. Proceedings of the 8th InternationalSymposium on Microbial Ecology Vol. 1, pp. 145^151.

[30] Hagstro«m, A[ ., Pinhassi, J. and Zweifel, U.L. (2001) Marine bacte-rioplankton show bursts of rapid growth induced by substrate shifts.Aquat. Microb. Ecol. 24, 109^115.

[31] Moriarty, D.J.W. and Hayward, A.C. (1982) Ultrastructure of bac-teria and the proportion of gram-negative bacteria in marine sedi-ments. Microb. Ecol. 8, 1^14.

[32] McKerrow, J., Vagg, S., McKinney, T., Seviour, E.M., Maszenan,A.M., Brooks, P. and Seviour, R.J. (2000) A simple HPLC methodfor analysing diaminopimelic acid diastereomers in cell walls ofGram-positive bacteria. Lett. Appl. Microbiol. 30, 178^182.

[33] Goodfellow, M. and LeChevalier, M.P. (1986) Genus Nocardia Tre-visan 1889. In: Bergey’s Manual of Systematic Bacteriology (Sneath,P.H.A., Ed.), pp. 1459^1471. WilliamspWilkins, Baltimore, MD.

[34] Hagstrom, A., Pinhassi, J. and Zweifel, U.L. (2000) Biogeographicaldiversity among marine bacterioplankton. Aquat. Microb. Ecol. 21,231^244.

[35] de Leeuw, J.W. and Largeau, C. (1993) A review of macromolecularorganic compounds that comprise living organisms and their role inkerogen, coal, and petroleum formation. In: Organic Geochemistry(Engel, M.H. and Macko, S.A., Eds.), pp. 23^72. Plenum Press, NewYork.

[36] Nagata, T., Meon, B. and Kirchman, D.L. (2003) Microbial degra-dation of peptidoglycan in seawater. Limnol. Oceanogr. 48, 745^754.

[37] Bunte, C. and Simon, M. (1999) Bacterioplankton turnover of dis-solved free monosaccharides in a mesotrophic lake. Limnol. Ocean-ogr. 44, 1862^1870.

[38] O’Dowd, R.W. and Hopkins, D.W. (1998) Mineralization of carbonfrom D- and L-amino acids and D- glucose in two contrasting soils.Soil Biol. Biochem. 30, 2009^2016.

[39] Riemann, L. and Azam, F. (2002) Widespread N-acetyl-D-glucosam-ine uptake among pelagic marine bacteria and its ecological implica-tions. Appl. Environ. Microbiol. 68, 5554^5562.

[40] Weinbauer, M.G. and Peduzzi, P. (1995) E¡ect of virus-rich highmolecular weight concentrates of seawater on the dynamics of dis-solved amino acids and carbohydrates. Mar. Ecol. Prog. Ser. 127,245^253.

[41] Cibik, R., Tailliez, P., Langella, P. and Chapot-Chartier, M.P. (2001)Identi¢cation of Mur, an atypical peptidoglycan hydrolase derivedfrom Leuconostoc citreum. Appl. Environ. Microbiol. 67, 858^864.

[42] Smith, T.J., Blackman, S.A. and Foster, S.J. (2000) Autolysins ofBacillus subtilis : multiple enzymes with multiple functions. Microbi-ology 146, 249^262.

[43] Rydman, P.S. and Bamford, D.H. (2002) Phage enzymes digest pep-tidoglycan to deliver DNA. ASM News 68, 330^335.

[44] Karner, M., Fuks, D. and Herndl, G.J. (1992) Bacterial activity alonga trophic gradient. Microb. Ecol. 24, 243^257.

[45] Zubkov, M.V. and Sleigh, M.A. (1998) Heterotrophic nanoplanktonbiomass measured by a glucosaminidase assay. FEMS Microbiol.Ecol. 25, 97^106.

[46] Hahn, M.W., Moore, E.R.B. and Ho«£e, M.G. (2000) Role of micro-colony formation in the protistan grazing defense of the aquaticbacterium Pseudomonas sp. MWH1. Microb. Ecol. 39, 175^185.

[47] Grutters, M., van Raaphorst, W., Epping, E., Helder, W., de Leeuw,J.W., Glavin, D.P. and Bada, J.L. (2002) Preservation of amino acidsfrom in situ-produced bacterial cell wall peptidoglycans in northeast-ern Atlantic continental margin sediments. Limnol. Oceanogr. 47,1521^1524.

FEMSEC 1555 24-11-03 Cyaan Magenta Geel Zwart

N.O.G. J8rgensen et al. / FEMS Microbiology Ecology 46 (2003) 269^280280