origins and fate of fungi and bacteria in the gut of lumbricus … · 2017-12-03 · origins and...

Origins and fate of fungi and bacteria in the gutof Lumbricus terrestris L. studied by image analysis

Frank Schoënholzer, Dittmar Hahn *, Josef ZeyerSwiss Federal Institute of Technology (ETH), Institute of Terrestrial Ecology, Soil Biology, Grabenstr. 3, CH-8952 Schlieren, Switzerland

Received 2 September 1998; received in revised form 11 November 1998; accepted 14 November 1998

Abstract

The effect of the passage through the gut of the earthworm Lumbricus terrestris L. on fungi and bacteria ingested withdecomposing leaves of Taraxacum officinale and with soil was quantified using image analysis tools. Both leaf and soil materialwere labeled with fluorescent latex microbeads to allow a quantification of the food sources in the fore-, mid-, and hindgut ofthe earthworms. The content of leaf material in the gut varied in a range between 4 and 59% of the total gut content in differentearthworms and the different parts of the intestine of individual animals. Filamentous fungi in the gut compartments werefound to originate mainly from leaf material (7700 þ 1800 Wg (g leaf (dry wt.))31), however, the major part was disrupted beforearriving in the intestine. Remaining hyphae in the foregut with a biomass of up to 900 þ 150 Wg (g gut content (dry wt.))31 werecompletely digested during passage through the earthworm gut. Spores of fungi were not detected in our studies. Bacterial cellnumbers in the gut compartments ranged from 63 þ 5U108 to 327 þ 16U108 (g gut content (dry wt.))31 and were significantlyhigher than the numbers found in the soil (50 þ 1U108 cells (g soil (dry wt.))31). Cell numbers usually increased from fore- tohindgut. This increase was not correlated to contents of organic material and only partially due to a multiplication of bacterialcells. Numbers of dividing cells accounted in total for approximately 12% of all bacteria, increasing significantly from fore- tohindgut, counts were from 10 þ 1U108 to 25 þ 2U108 (g gut content (dry wt.))31, respectively. Average cell volumes of bacteriacalculated from cell size distributions in leaf and soil material differed significantly, being 0.197 and 0.063 Wm3, respectively. Inthe gut compartments, average cell volumes ranged from 0.043 to 0.070 Wm3, which may indicate the disruption of large cellsoriginating from the leaves before arriving in the foregut. z 1999 Federation of European Microbiological Societies.Published by Elsevier Science B.V. All rights reserved.

Keywords: Biomass; Calco£uor; Cell size distribution; Fluorescent microbeads; In situ hybridization; rRNA

1. Introduction

Decomposition of plant material is assumed to bemainly mediated by microorganisms, such as bacte-ria and fungi [1]. The rate and extent of the decom-

position depends on the chemical composition of thematerial, on a number of environmental factors, andon the microbial community [2,3]. The activity of thedecomposing microorganisms is accelerated by theactivity of the soil fauna. Anecic earthworm species,such as Lumbricus terrestris L., for example, removepartially decomposed plant material from the soilsurface and transport it to the subsurface layers.

0168-6496 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 1 6 8 - 6 4 9 6 ( 9 8 ) 0 0 1 1 1 - 1

FEMSEC 995 8-3-99 Cyaan Magenta Geel Zwart

* Corresponding author. Tel. : +41 (1) 633-6123;Fax: +41 (1) 633-1122; E-mail: [email protected]

FEMS Microbiology Ecology 28 (1999) 235^248

Plant material is fragmented by ingestion and even-tually mixed with soil which is passed through theintestine and ¢nally excreted as cast [4]. Earthwormstherefore play a major role in the mixing of soil, inimproving soil aeration, and in increasing water-holding capacity [1,5].

Ingestion and passage through the intestine ofearthworms also a¡ects microorganisms that are as-sociated with plant material and soil. It has beensuggested that microorganisms provide a source ofnutrients for earthworms with fungi as a major andbacteria as a minor source [6]. Filamentous fungi, forexample, have been shown to be digested by earth-worms [7^9], although other studies report increasednumbers of fungi in the cast after passage throughthe intestine [10]. For bacteria, a few studies demon-strate a decrease or no consistent change in bacterialnumbers during gut transit [10,11]. However, moreoften an increase in total number of bacteria is ob-served [6,12] and also higher numbers of bacteria inlinings and casts compared to surrounding bulk soil[6,12^15].

Most of these studies are based on growth-depend-ent detection methods which might be impeded bythe fact that many microorganisms are di¤cult toisolate or even resist cultivation. Growth-dependentdetection protocols can therefore be extremely selec-tive and often underestimate numbers of microor-ganisms and diversity of microbial communities[16^18]. Epi£uorescence microscopy is now generallyacknowledged to be the best method available forquanti¢cation of microorganisms in all habitats[19]. Quanti¢cation of total microbial communitiesinvolves staining the microorganisms in the samplewith £uorochromes, such as Calco£uor M2R, £uo-rescein isothiocyanate (FITC), acridine orange, 4P,6-diamidino-2-phenylindole (DAPI) or europium che-late (see [20] for review). Quanti¢cation of speci¢cpopulations is currently achieved by the £uorescentantibody technique in which antibody/£uorochromeconjugates bind to cells or cell components of thetarget organism (see [21] for review), and by the insitu hybridization technique in which £uorescent oli-gonucleotide probes hybridize to rRNA sequences in¢xed whole cells (see [22] for review).

For the quanti¢cation of stained fungi and bacte-ria in aquatic and terrestrial environments, auto-mated image analysis programs have become more

and more important tools because criteria for thequanti¢cation can be de¢ned and standardized andsize measurements be used to estimate microbial bio-masses [23^25]. Automated image analysis, however,depends on the availability of high-quality images.These may most reliably be obtained from hetero-geneous environmental samples by advanced detec-tion technology, e.g. by confocal laser scanning mi-croscopy [24]. In a previous study, we obtained high-quality images also by the optimization of the sam-ple preparation ensuring an equal dispersion of thetarget cells in thin layers [23], by the use of highcontrast £uorochromes and by the de¢nition andstandardization of speci¢c criteria for the automatedimage analysis program [26].

The aim of this study was to quantify the e¡ect ofthe passage through the intestine of the earthwormLumbricus terrestris L. on ¢lamentous fungi and bac-teria in soil and in decomposing leaves of Taraxacumo¤cinale that were used as food source. Both soiland leaves were labeled with £uorescent latex mi-crobeads of di¡erent sizes before feeding the animalsto allow a quanti¢cation of the food source in theintestine. Filamentous fungi were quanti¢ed by im-age analysis after staining with Calco£uor M2R andafter in situ hybridization with Cy3-labeled oligonu-cleotide probes. For the quanti¢cation of bacteriaand the determination of bacterial biomass, an imageanalysis program was developed based on color dif-ferentiation to enable the di¡erentiation of DAPI-stained bacteria from the heterogeneous backgroundof samples of leaves, soil, and the earthworm intes-tine.

2. Materials and methods

2.1. Labeling of soil and leaves with £uorescent beads

Surface samples down to a depth of 10 cm fromthe pristine forest soil Hau (Birmensdorf, Switzer-land, a silty clay with 5.6% organic material ; vege-tation, Aro-Fagetum, [27]) with a water content ofabout 30% were sieved (6 4 mm) and labeled with£uorescent yellow-green latex microbeads with a di-ameter of 0.3 Wm (surfactant free, carboxyl polystyr-ene latex, Interfacial Dynamics, Portland, OR,USA). An even distribution of beads in soil was


F. Schoënholzer et al. / FEMS Microbiology Ecology 28 (1999) 235^248236

achieved by spraying 25 ml of a 0.04% (w/v) bead inwater solution onto a thin layer of 500 g of soilfollowed by careful mixing.

Green leaves of Taraxacum o¤cinale were har-vested from plants in fall. Leaves were dried undervacuum (25 mbar and 40³C) for 24 h and stored at320³C. To increase microbial biomass, leaf pieces(area between 300 and 700 mm2) without big leafribs (10.5 mm in diameter) were incubated on glass¢ber ¢lters soaked with 3 ml of deionized water inPetri dishes for 6 days at 20³C and 90^100% relativehumidity in the dark [28]. The glass ¢ber ¢lters withthe leaf pieces were subsequently placed in a solutionof 0.0057% (w/v) £uorescent yellow-green latex mi-crobeads with a diameter of 0.2 Wm (surfactant free,carboxyl polystyrene latex, Interfacial Dynamics) for30 min at room temperature. Thereafter, they wereair-dried for 15 min at room temperature. The areaof the leaf pieces was related to the dry weight, bymeasuring leaf areas of 15 leaf pieces of scannedleaves by image analysis (OneScanner, Apple) [23]and subsequently determining leaf dry weight afterdrying at 105³C for 24 h.

2.2. Microcosms

Individuals of the earthworm L. terrestris wereretrieved from the original soil Hau using a solutionof 0.1% formaldehyde [29]. They were rinsed imme-diately in water and kept in aerated (moist air) PVC-boxes (35 cm (l)U27 cm (w)U22.5 cm (h)) at 10³C insoil of the ¢eld site and fed with leaves of T. o¤ci-nale. Experiments were performed at 15³C in thedark in microcosms consisting of Plexiglas tubes(with a diameter of 14 mm and a height of 300mm) which were closed on the bottom and whichcontained one individual earthworm [28]. Tubeswere ¢lled with 50 g of soil labeled with £uorescentmicrobeads. A Petri dish (diameter 8.5 cm, with ahole in the center) was lined with moistened paper(Schleicher and Schuell 595) which was covered witha nylon tissue (mesh size 300 Wm). The Petri dish was¢xed on the top of the tube to facilitate controlledfeeding of earthworms [30]. The animals were feddaily with leaf pieces labeled with £uorescent mi-crobeads (16 þ 4 mg (dry wt.) g31 of earthwormlive weight day31). This amount was about 70% ofthe average consumption of L. terrestris when food

was o¡ered ad libitum to ensure immediate and com-plete consumption of the leaves. The feeding experi-ment lasted 4 days. Earthworm microcosms thatcontained withdrawn leaves in the burrows at theend of the experiment were omitted from the study.

2.3. Sample preparation

One day after the last feeding, the animals werekilled in 4% paraformaldehyde at 4³C, washed inwater and dissected from the ventral side. The intes-tine was divided into three parts, the fore-, mid- andhindgut. Subsamples of the gut sections (approxi-mately 0.5^1.0 mg fresh weight) as well as soil sam-ples and leaf pieces were ¢xed in 4% paraformalde-hyde in phosphate-bu¡ered saline (PBS) [31] for 16 h.Fixed samples were subsequently washed twice inPBS. Afterwards, subsamples of the gut sectionswere shaken in 5 ml of 0.1% pyrophosphate at115 rpm at room temperature for 1 h. After removalof the intestinal wall, the suspensions were stored in50% ethanol in PBS at 320³C [12]. Soil samples werestored in 50% ethanol in PBS at 320³C at a concen-tration of 50 mg of soil (dry wt.) ml31 and leaf piecesin 5 ml of 50% ethanol in PBS [31].

Subsamples of the gut sections were dispersed in0.1% sodium pyrophosphate in PBS at a ¢nal con-centration of approximately 1^1.5 mg (dry wt.) ml31.Soil subsamples of 25 Wl (containing the equivalentof 1 mg of soil (dry wt.)) were dispersed in 975 Wl of0.1% sodium pyrophosphate in PBS. Leaf pieceswere transferred into 0.5 ml of 0.1% sodium pyro-phosphate in PBS, ground in a mortar, and an addi-tional 1.5 ml of 0.1% sodium pyrophosphate in PBSwas added. Subsamples of the gut sections, soil sam-ples and leaf suspensions were homogenized with asonicator with a 2-mm-diameter probe (SonicatorB-12, Branson, Danbury, CT) at 90 W for 1 min[28]. Thereafter, leaf suspensions were centrifugedat 5000Ug for 15 min and the supernatant was par-tially discarded to a residue of approximately 400^800 Wl. All suspensions were ¢nally supplementedwith the detergent NP40 (Fluka, Buchs, Switzerland)to a ¢nal concentration of 0.01%.

2.4. Quanti¢cation of the food source

The amount of leaf material in the samples was


F. Schoënholzer et al. / FEMS Microbiology Ecology 28 (1999) 235^248 237

correlated to the number of microbeads. The corre-lations were evaluated on mixtures of microbead-la-beled, ground leaves and microbead-labeled soil sam-ples which were suspended in ratios of 0/100, 20/80,40/60, 60/40, and 100/0 (dry wt.) in 0.1% sodiumpyrophosphate in PBS and homogenized with a so-ni¢er with a 2 mm diameter probe (Soni¢er B-12,Branson) at 90 W for 1 min. Homogenized subsam-ples of 10 Wl (n = 3) were spotted onto gelatin-coatedslides (0.1% gelatin, 0.01% KCr(SO4)2) and micro-beads were counted with a Zeiss Axiophot micro-scope ¢tted for epi£uorescence with a high-pressuremercury bulb (50 W) and ¢lter set 09 (Zeiss ; G450-490, FT510, LP520) at 400U magni¢cation (ZeissPlan-Neo£uar 40U/1.30 oil). Ratios between num-bers of beads in the fore-, mid-, and hindgut of dif-ferent individuals of L. terrestris were then used tocalculate the contents of leaf material in the intestine.

2.5. Analysis of fungi

Fungi were stained for 60 min with a 2 mg ml31

solution of Calco£uor M2R [23,32] after dehydrationin 10-Wl aliquots of leaf (n = 10), soil (n = 10), and gutsamples (n = 3) spotted onto gelatin-coated slides.Preparations were subsequently washed with distilledwater and air-dried. In situ hybridizations with Cy3-labeled oligonucleotide probe EUK516 targeting allmembers of the Domain Eukarya [33] were per-formed in 9 Wl of hybridization bu¡er (0.9 MNaCl, 20 mM Tris/HCl, 5 mM EDTA, 0.01% SDS,pH 7.2) in the presence of 20% formamide, and 1 Wlof the probe (25 ng Wl31) at 42³C for 2 h [34]. Afterhybridization, the slides were washed in hybridiza-tion bu¡er at 48³C for 15 min, subsequently rinsedwith distilled water, and air-dried. Slides weremounted with Citi£uor solution and the preparationswere examined with a Zeiss Axiophot microscope¢tted for epi£uorescence with a high-pressure mer-cury bulb (50 W) at 400U magni¢cation (Zeiss Plan-Neo£uar 40U/1.30 oil). Fungi stained by Calco£uorM2R were analyzed with ¢lter set 02 (Zeiss; G365,FT395, LP420), whereas binding of Cy3-labeledprobes was detected with ¢lter set HQ-Cy3 (AHFAnalysentechnik, Tuëbingen, Germany; G535/50,FT565, BP610/75).

The images were captured (exposure time 1.44 s)via a cooled CCD 3-chip color video camera C5810

(Hamamatsu Photonics) connected via a SCSI cardto a Macintosh computer (Apple Power PC 9500/200) by the software U6341 (Hamamatsu Photonics)with a resolution of 992U672 pixels (RGB, 24 bit)and stored on the hard disk. Fungal biomass wasdetermined following our image analysis protocoldescribed previously [23], except in the ¢rst step, inwhich RGB-signals were converted into a gray scaleimage. Image analysis involved measurements ofboth lengths and diameters of the fungal hyphaeafter staining with Calco£uor M2R and after insitu hybridization with the Cy3-labeled oligonucleo-tide probe EUK516. The fungal biomass was calcu-lated as B = L (d/2)2 Zr, where L and d were thelength and the diameter of fungal hyphae, respec-tively, and r = 0.32 mg mm33 was considered to bethe density of fungal hyphae [35,36].

2.6. Analysis of bacteria

Ten-microliter aliquots of leaf (n = 10), soil(n = 10), and gut samples (n = 3) were spotted ontogelatin-coated slides. The preparations were allowedto air dry and subsequently dehydrated in 50, 80 and96% of ethanol for 3 min each [33]. Bacteria werestained with a 1 Wg ml31 solution of the DNA spe-ci¢c dye DAPI (Sigma, Buchs, Switzerland) by incu-bation for 20 min at 37³C. Preparations were sub-sequently washed in bu¡er containing 10 mM Tris/HCl, pH 7.2, 5 mM EDTA, 0.01% SDS, and 0.9 MNaCl for 15 min at 48³C, washed with distilledwater, and air-dried [31]. Slides were mounted withCiti£uor solution and the preparations were exam-ined with a Zeiss Axiophot microscope ¢tted for epi-£uorescence with a high-pressure mercury bulb(50 W) and ¢lter set 02 (Zeiss; G365, FT395,LP420) at 400U magni¢cation (Zeiss Plan-Neo£uar40U/1.30 oil).

Bacteria were analyzed in 21 images per samplewith up to 400 cells per image. To enhance the con-trast, the range of the color depth detected by thecamera was expanded to the available 24 bit and theRGB image was separated by color channels (auto-mated batch-function of Adobe Photoshop 4.0). Theanalysis consisted of three pathways, that were: (1)the detection of all objects; (2) the detection of de-bris (for example, soil minerals and organic materi-al) ; and (3) the detection of DAPI-halos (Fig. 1). All



pathways were subdivided into three major steps,that were: (1) the gray scale image processing; (2)the thresholding/binarization; and (3) the binary im-age processing.

As for the recognition of fungi, it was essential forthe quanti¢cation of bacteria to capture RGB-im-ages. Bacteria were analyzed using an image analysisprocedure based on the software Prism [37] and theimage analysis tools described by Russ [26] to proc-ess the green and blue channels. While DAPI-stainedcells in the captured image (Fig. 2a,d) could be rec-ognized as distinct objects in the green channel (Fig.2 b,e), they were indistinct and oversized in the bluechannel (Fig. 2c,f). Since debris present as green oryellow particles generally showed less blue signalthan bacteria, this criterion was used by the imageanalysis procedure to di¡erentiate between bacteriaand debris, when a di¡erentiation based on morpho-logical properties was not possible.

For the detection of all objects, in principle, thesetting of a threshold in the green channel could beused to separate the objects present in the image(brighter than the threshold value) from the back-ground (darker than the threshold value). However,due to e.g. auto£uorescence e¡ects or inhomogene-ous illumination, neither bacteria (and other par-ticles) nor the background had a consistent bright-ness over the whole image area. Therefore, a reliabledi¡erentiation between objects (bacteria and otherparticles) and background required a special grayvalue transformation, the Marr^Hildreth ¢lter (Dif-ference of Gaussians; Oé 1 = 1.0, Oé 2 = 0.625) [26,38]which was used to achieve a homogeneous back-ground and to quench large areas of £uorescence(Fig. 2g). For the calculation of the threshold, nogeneral solution was available, since it was highlydependent on the sort and quality of the imagethat had to be analyzed [39]. For our images, weused a procedure that calculated the threshold rangeautomatically by setting the lower limit to white(gray value 0) and a preliminary upper limit togray, representing the average background whichvaried in the di¡erent images. The de¢nitive upperlimit was calculated by subtracting a gray value o¡-set since the resulting gray value was found to cor-relate to the threshold manually determined in 20images from each source (leaves, soil, and gut mate-rial). This step allowed automatic calculation of

thresholds and to perform binarizations. Two thresh-olds, di¡ering slightly by two gray values were set,followed by binarizations that resulted in two binaryimages. Both images were combined by a Feature-AND operation, con¢rming only those objects in the¢rst image that shared at least one pixel with anobject from the second image [37], enabling the re-moval of small and weakly stained objects. After theomission of small and weakly stained objects, thebinarized image (Fig. 2k, black objects), however,not only contained bacteria, but also a lot of other,often yellow stained particles and debris.

For the di¡erentiation of bacteria and debris, twoadditional procedures were necessary: (1) the detec-tion of debris ; and (2) the detection of DAPI-halos(Fig. 1). Both procedures were based on a processedimage of: (1) the blue channel (Fig. 2c,f) after cut-ting the darkest part of the background; and (2) thegreen channel which had sharper structures and amore homogeneous background than the original(Fig. 2h). The detection of debris focused on theidenti¢cation of large objects with an area of s 2.2Wm2 with small furrowed and regular boundaries(formfactor s 0.4) after preprocessing and binariza-tion of the green channel [37] (Fig. 2k, dark grayobjects).

The most reliable indicator for bacteria was thehalo-ring of the DAPI-£uorescence (Fig. 2a^c).This property was used to distinguish a major partof the bacteria from other particles. First, the grayvalue range of the blue channel was reduced to 0^225to eliminate non-essential and disturbing back-ground. Thereafter, the processed blue channel wassubtracted from the processed green channel (Fig. 2i)which enabled the calculation of thresholds to sepa-rate the DAPI-halos from the background. Thethreshold was found by preliminary setting to 165and performing an iterative decreasing or increasingof the threshold, until the area of the pixels con-tained in the threshold range was between 0.7 and3% of the image area. Due to the big gray valuerange the DAPI-halo occurred in the processed im-age (Fig. 2i), two di¡erent thresholds were set. Ob-jects detected by the larger threshold range (Fig. 2j,black pixels) were only accepted when con¢rmed bythe smaller threshold range (Fig. 2j, gray pixels).This allowed the exclusion of blue signals whichwere too weak to represent DAPI-halos originating




Fig. 1. Flow chart of the image processing procedure for the analysis of bacteria. 2a^2l refer to the images presented in Fig. 2.

CFig. 2. Di¡erent stages during image processing of a foregut sample of L. terrestris. High magni¢cation views of the RGB image (a), thegreen channel of the RGB image (b), and the blue channel of the RGB image (c). Low magni¢cation views (320U400 pixels) of an RGBimage (d), the green channel of the RGB image (e), the blue channel of the RGB image (f), the green channel after Marr^Hildreth proce-dure (g); the green channel after background subtraction (h); subtraction of preprocessed images from blue and green channel (i) and thedetection of DAPI-halos after binarizations (j) ; the combined binary image (k) showing: (1) all objects (black objects, binarization of g);(2) debris (dark gray objects, binarization of h); and (3) DAPI-halos (light gray objects) ; and the detection of bacteria after deletion ofall non-bacterial objects (l). Scale bar: 10 Wm.


from bacteria (Fig. 2j, e.g. upper left part). After-wards, two di¡erent binarizations, set at thresholdsdi¡ering by 12 gray values (each followed by an ero-sion operation with a coe¤cient of 7 to remove pixelnoise) were performed, con¢rming each other in therecognition of clearly visible DAPI-halos. After com-bining both binary images by a Feature-AND oper-ation, objects were enlarged with a dilation opera-tion with a coe¤cient of 2.

Finally, the objects detected by the left pathway(Fig. 2k, black objects) were combined with the de-tected DAPI-halos (Fig. 2k, light gray objects) by aFeature AND operation to identify objects as bacte-ria (Fig. 2l). In soil and in material from the gut ofL. terrestris, non-bacterial particles were present inlarge amounts. When located in the direct vicinity ofbacteria they were erroneously recognized as bacteriawhen touching or covering a DAPI-halo. Therefore,all images were checked and `false' bacteria weredeleted manually by click and drag operation inthe resulting binary image after image processing.

Following this procedure, cell numbers and cellsizes, based on measurements of area and perimeterfor each bacterium or bacterial agglomerate, weredetermined. After subsequent determination of ¢berlength and width [37], biovolumes were calculated[24] and divided up into 24 size classes rangingfrom 0 to 0.48 Wm3.

3. Results and discussion

3.1. Quanti¢cation of the food source

Fluorescent latex beads with very de¢ned sizes andintensive £uorescence have already been used for theevaluation of biovolumes of e.g. bacterioplankton[39], for the estimation of feeding activities of e.g.protozoa and oligochaetes [40^42] or for the model-ing of retention times of bacteria in the gut of wood-lice [43]. In our study microbeads with a diameter of0.2 or 0.3 Wm could reliably be di¡erentiated by size.Numbers of microbeads in the original leaf sampleswere 145 þ 6U107 beads (g leaf (dry wt.))31 and inthe soil 107 þ 4U107 beads (g soil (dry wt.))31. Inmixtures of labeled leaf and soil material, theamount of leaf and soil material also correlated tothe numbers of microbeads (Fig. 3). The high corre-

lation between the amount of leaf or soil materialand microbeads with a diameter of 0.2 or 0.3 Wm isindicated by R-values of 0.992 and 0.981, respec-tively.

In the earthworms, ratios between beads of bothsizes (0.2 and 0.3 Wm) and consequently the amountsof leaf and soil material di¡ered signi¢cantly betweenindividuals and also in di¡erent gut compartments ofthe individuals, the fore-, mid-, and hindgut, respec-tively (Fig. 4a). The number of beads was used todetermine percentages of leaf material in the intes-tine. With approximately 5 and 10% of leaf materialindividuals F and A showed no signi¢cant di¡eren-ces in the contents throughout the whole intestine(Fig. 4b). Earthworms B and D exhibited increasingamounts of leaf material from fore- to hindgut withapproximately 5^30% and 10^30%, while earth-worms C, E and G had decreasing amounts of leafmaterial from fore- to hindgut with approximately60 to 5%, 55 to 5% and 25 to 15% (Fig. 4b). Theconsiderably high variability in contents of the foodsource between individual earthworms was attrib-uted to the di¡erent behavior of the earthworms(time and amounts of feeding), although an accumu-lation of beads by retention due to the internal struc-ture of the gut could not be excluded [43]. Such


Fig. 3. Number of microbeads with a diameter of 0.2 Wm(9R9 ; Y =34621+1.45U106 X ; R = 0.992) and 0.3 Wm (989 ;Y = 1.02831.032U106 X ; R = 0.981) as a function of di¡erentamounts of leaf and soil material.


retention, however, should not in£uence the ratiobetween beads with only slightly di¡erent sizes (0.2and 0.3 Wm diameter, respectively). The correlationbetween bead sizes and the contents of leaf and soilmaterial is therefore still valid.

3.2. Analysis of fungi

For the recognition of fungi and their di¡erentia-tion from leaf and soil material, it was essential tocapture RGB images providing full-color informa-tion of the sample. Filamentous fungi were only de-tected in large amounts in leaf material where theyaccounted for a median length of 5760 þ 1330 m (gleaf (dry wt.))31. In addition to their presence on leafmaterial, ¢lamentous fungi were only detected inlarger amounts in the foregut of individuals B andC accounting for lengths of 190 þ 130 and 670 þ 200m (g gut content (dry wt.))31, respectively. In soiland in all remaining gut compartments of the earth-worms investigated, hyphal lengths were usually lessthan 1% of the hyphal lengths found in leaf material.These results indicated a disruption of ¢lamentousfungi originating from leaf material during passagethrough the crop and gizzard of the earthworm, con-¢rming previous reports of a major digestion of fun-gal hyphae in the anterior part of the gut [44]. Re-maining hyphae in the foregut might then further bedisrupted during passage through the earthworm gutresulting in the complete digestion of ¢lamentousfungi during gut transit in earthworms [45]. Sporesof fungi which were reported to survive gut passage[46,47] were not detected in our studies. This failuremight be due to the absence of spores on leaf mate-rial that was incubated under conditions favoringvegetative growth of fungi or be caused by cell wallcomponents which rendered spores impermeable forprobes used for in situ hybridization.

Hyphal lengths re£ected fungal biomass thoughthe accurate determination of biomass was in£u-enced by the staining protocol. Calco£uor M2Rwas shown to be a more even stain for hyphae com-pared to other £uorescent dyes [48]. However, Cal-co£uor M2R also stained cellulose which hinderedreliable detection of fungi especially in leaf material(Fig. 5). Previous studies had shown that these prob-lems could be circumvented by depolymerization ofcellulose by cellulases before staining [23]. An alter-


Fig. 4. Ratios of microbeads (0.2 and 0.3 Wm) (a), the amount ofleaf material (b), the determined number of bacteria (c), the theo-retically expected numbers of bacteria based on the amount ofleaf material (d), and the number of dividing bacteria (e) in thefore-, mid-, and hindgut of seven individuals of L. terrestris (A^G), 24 h after feeding.


native to depolymerization of cellulose was based onin situ hybridization with Cy3-labeled oligonucleo-tide probe EUK516 (Fig. 5). In situ hybridizationwith probe EUK516 targeting rRNA in the cells re-sulted in accurate measurements of hyphal length;however, width measurements and consequently bio-mass estimates were generally lower than measure-ments after staining with Calco£uor M2R. This wascaused by the failure of in situ hybridization to vis-ualize cell walls which were then consequently notincluded into measurements. The estimated fungalbiomass, which is dependent upon the square ofthe fungal diameter, was therefore 2.2^2.3 timessmaller after in situ hybridization than after a com-bined approach based on the determination of hy-phal length by in situ hybridization and of width bymeasuring selected Calco£uor M2R-stained hyphae.On leaf material, fungal biomass determined after insitu hybridization accounted for 3500 þ 800 Wg (g leaf(dry wt.))31 whereas that determined after stainingwith Calco£uor M2R was 7700 þ 1800 Wg (g leaf(dry wt.))31. Similar ratios were found in the foregutof earthworms B and C with 110 þ 80 and 410 þ 70Wg (g gut content (dry wt.))31 after in situ hybrid-ization and 250 þ 170 and 900 þ 150 Wg (g gut con-tent (dry wt.))31 after staining with Calco£uor, re-spectively.

3.3. Analysis of bacteria

3.3.1. Bacterial numbersBacterial numbers in leaves were 216 þ 12U108

cells (g leaf (dry wt.))31 and in soil 50 þ 1U108 cells(g soil (dry wt.))31. Numbers of bacteria in the gutcompartments ranged from 63 þ 5U108 (g gut con-tent (dry wt.))31 to 327 þ 16 108 (g gut content (drywt.))31 (Fig. 4c). They were comparable to the num-bers found in the leaf material and generally higherthan the numbers found in the soil (Fig. 4c). Bacte-rial numbers generally increased from fore- to hind-gut which is consistent with earlier studies [6,10,12].The increase was independent of the content of leafmaterial in the gut. In earthworm B, for example,where the amount of leaf material increased fromfore- to hindgut, bacterial numbers also increasedfrom 121 þ 10U108 to 186 þ 13U108 (g gut content(dry wt.))31, respectively. In earthworm F that hadonly small amounts of leaf material in all three gutcompartments, a slight increase in bacterial numbersfrom 70 þ 6U108 (g gut content (dry wt.))31 in theforegut to 101 þ 5U108 (g gut content (dry wt.))31 inthe hindgut was determined. A slight increase in bac-terial numbers towards the hindgut was also ob-tained for earthworm C where the amount of leafmaterial decreased from fore- to hindgut.


Fig. 5. Detection of ¢lamentous fungi in leaf material after staining with Calco£uor M2R (a) and after in situ hybridization with Cy3-la-beled oligonucleotide probe EUK516 (b). Scale bar: 10 Wm.


Based on the observed amount of leaf and soilmaterial in the gut compartments, on the bacterialnumbers found in the leaves and soil, and under theassumption that: (1) by ingestion and transport tothe foregut, no bacterial growth or loss occurred;and (2) all bacteria in the foregut were originatingfrom either leaf or soil material, theoretical numbersof bacteria were calculated in fore-, mid-, and hind-gut (Fig. 4d). The comparison between observed andcalculated numbers showed that in the foregut, theobserved numbers were higher than theoretical val-ues when the amount of leaf material was small (in-dividuals A, B, D, F and G), and lower than ex-pected in the foregut with large amounts of leafmaterial (individuals C and E) (Fig. 4d). This resultsuggested di¡erential e¡ects of passage through cropand gizzard of the earthworm on bacteria in leaf andsoil material.

It has been speculated that the increase of bacte-rial numbers towards the end of the earthworm gutcould be due to growth of bacterial cells or to anactivation of dormant cells during gut passage[49,50]. A part of this increase might be attributedto the multiplication of bacteria. Numbers of divid-ing cells accounted in total for approximately 12% ofall bacteria, increasing signi¢cantly from fore- tohindgut with 10 þ 1U108 (g gut content (drywt.))31 (mean of all seven foreguts) to 25 þ 2U108

(g gut content (dry wt.))31 (mean of all seven hind-guts) (Fig. 4e). However, numbers of dividing cellsdid not explain the large discrepancies in bacterialnumbers between observed and calculated values.

3.3.2. Bacterial cell size distributionThe ratio between bacterial numbers in leaf and

soil material was 4.3 and much lower than the ratiosbetween bacterial volumes or biomasses which were18. This ratio was based on the estimated total bac-terial biomasses of 1500 þ 100 Wg C (g leaf (drywt.))31 for leaves and 83 þ 6 Wg C (g soil (drywt.))31 for soil. The di¡erent ratios were caused bylarge di¡erences in cell size distributions of the bac-teria in leaf and soil material (Fig. 6). With a meancell volume of 0.197 Wm3, bacteria on the leaves weremuch larger than those in the soil where the meancell volume was 0.063 Wm3. Soil is normally a poornutritional environment for bacteria compared todecomposing leaves in our experiments where de-

composition was initiated under favorable labora-tory conditions. Bacteria may adapt to the nu-trient-limited soil environment by the formation ofresting or dormant cells, such as dwarf cells, cysts orspores, which are usually much smaller in size thancells growing under optimal conditions [51,52]. Insoil, for example, about 70% of the bacterial com-munity may be represented by dwarf cells with adiameter of less than 0.3 Wm [53].

In our study, the total average cell volume of bac-teria in the fore-, mid-, and hindgut of all earth-worms A^G (all gut compartments of all individuals)was 0.050 Wm3 which was comparable to size rangesof bacteria in soil (0.063 Wm3). It is much smallerthan the mean cell volume of bacteria on the leaves(0.197 Wm3), suggesting a disruption of larger cellsoriginating from leaf material. This suggestion wassupported by size distributions of bacteria in the gutcompartments which re£ected the size distribution of


Fig. 6. Size distribution of DAPI-stained bacteria in soil (a) andleaf material (b). Scales for cell volumes as well as for size classesapply for both soil and leaf material.


bacteria in soil rather than that on leaves. Size dis-tributions of bacteria in the gut compartments didnot increase when the contents of leaf material in-creased. In earthworm B, for example, where theamount of leaf material increased from fore- tohindgut, the total average cell volume for all threegut compartments was 0.050 Wm3. Similar resultswere obtained on individual F that had only smallamounts of leaf material in all three gut compart-ments. With a total average cell volume of 0.043Wm3 in all three gut compartments, this earthwormshowed no signi¢cant shift in the bacterial sizes be-tween fore- and hindgut. Only in individual C, a shiftfrom larger bacteria with a mean volume of 0.070Wm3 in the foregut to smaller bacteria with a meanvolume of 0.044 Wm3 in the hindgut was observed(Fig. 7), which was consistent with the decreasingamount of leaf material from fore- to hindgut (Fig.4b). Based on cell size distributions of bacteria in soiland leaves (Fig. 6), calculated cell size distributionsfor this earthworm (C), however, displayed a muchbroader range of cell sizes with many cells in largersize classes (Fig. 7), again suggesting a disruption oflarger cells after ingestion by the earthworm.

The shift in bacterial cell size distribution fromlarge to smaller cells after ingestion, mixing andcomminution of the leaf and soil material might bethe result of a mechanical disintegration of bacteriaby shearing forces during passage through crop andgizzard [4]. Shearing forces resulting in disintegrationmight be more pronounced on large, vegetative cellsof bacteria than on smaller, dormant stages. Num-bers of bacteria determined in the foregut might,therefore, be smaller than theoretical values in sam-ples with high amounts of leaf material (earthwormsC and E).

Numbers of bacteria in the foregut might be high-er than theoretical values in samples with smallamounts of leaf material (individuals A, B, D, Fand G) because dormant cells protected from diges-tion by additional cell wall constituents that also actas permeation barrier for stains or probes becamedetectable after removal of these constituents byshearing forces. Both assumptions, the digestion ofvegetative cells as well as an activation of dormantcells, such as spores, have been shown to happenduring gut passage [54].

These results showed the importance of consider-


Fig. 7. Size distributions of observed and calculated bacteria in fore-, mid-, and hindgut of individual C of L. terrestris. Scales for cellvolumes as well as for size classes apply for both observed and calculated numbers of bacteria.


ing cell sizes in addition to counts during the studyof natural microbial populations. This was impor-tant when comparing fungi and bacteria in habitatswhere the nutrient situation varied considerably as inthe case of decomposing leaves, soil and the earth-worm gut. The analysis of the di¡erent habitatsshowed that, for example, the same bacterial num-bers could account for completely di¡erent bio-masses.

Acknowledgments

This work was supported by grants from SAN-DOZ AGRO SA, Basle, the Swiss National ScienceFoundation (Priority Program Biotechnology), andthe Swiss Federal O¤ce of Environment, Forestsand Landscape (BUWAL).

References

[1] Anderson, R.V., Coleman, D.C. and Cole, C.V. (1981) E¡ectsof soprophytic grazing on net mineralization. Ecol. Bull.(Stockholm) 33, 201^216.

[2] Berg, B. and Staaf, H. (1981) Leaching, accumulation andrelease of nitrogen in decomposing forest litter. Ecol. Bull.(Stockholm) 33, 163^178.

[3] Witkamp, M. (1966) Decomposition of leaf litter in relation toenvironment, micro£ora and microbial respiration. Ecology47, 194^201.

[4] Edwards, C.A. and Bohlen, P.J. (1996) Biology and Ecologyof Earthworms. Chapman and Hall, London, UK.

[5] Jensen, V. (1974) Decomposition of angiosperm tree leaf litter.In: Biology of Plant Litter Decomposition (Dickinson, C.H.and Pugh, G.J.F., Eds.), pp. 69^104. Academic Press, Lon-don, UK.

[6] Edwards, C.A. and Fletcher, K.E. (1988) Interactions betweenearthworms and microorganisms in organic-matter break-down. Agric. Ecosyst. Environ. 24, 235^247.

[7] Dkhar, M.C. and Mishra, R.R. (1991) Micro£ora in the gutcontents of the earthworm (Amynthas di¡ringens Baird).J. Phytol. Res. 4, 155^159.

[8] Dash, M.C., Mishra, P.C. and Behera, N. (1979) Fungal feed-ing by a tropical earthworm. Trop. Ecol. 20, 9^12.

[9] Dash, M.C., Binapani, S., Behera, N. and Dei, C. (1984) Gutload and turnover of soil, plant and fungal material by Dra-wida calebi, a tropical earthworm. Rev. Ecol. Biol. Sol. 21,387^393.

[10] Kristufek, V., Ravasz, K. and Pizl, V. (1992) Changes in den-sities of bacteria and microfungi during gut transit in Lumbri-cus rubellus and Aporrectodea caliginosa (Oligochaeta: Lum-bricidae). Soil Biol. Biochem. 24, 1499^1500.

[11] Pedersen, J.C. and Hendriksen, N.B. (1993) E¡ect of passagethrough the intestinal tract of detrivore earthworms (Lumbri-cus spp. ) on the number of selected Gram-negative and totalbacteria. Biol. Fertil. Soils 16, 227^232.

[12] Fischer, K., Hahn, D., Amann, R.I., Daniel, O. and Zeyer, J.(1995) In situ analysis of the bacterial community in the gut ofthe earthworm Lumbricus terrestris L. by whole-cell hybrid-ization. Can. J. Microbiol. 41, 666^673.

[13] Ross, D.J. and Cairns, A. (1982) E¡ects of earthworms andryegrass on respiration and enzyme activities of soil. Soil Biol.Biochem. 14, 583^587.

[14] Parle, J.N. (1963) Micro-organisms in the intestine of earth-worms. J. Gen. Microbiol. 31, 1^11.

[15] Parle, J.N. (1963) A microbiological study of earthworm casts.J. Gen. Microbiol. 31, 13^22.

[16] Richaume, A., Steinberg, C., Jocteur-Monrozier, L. and Faur-ie, G. (1993) Di¡erences between direct and indirect enumer-ation of soil bacteria: the in£uence of soil structure and celllocation. Soil Biol. Biochem. 25, 641^643.

[17] Skinner, F.E., Jones, P.C.T. and Mollison, J.E. (1952) A com-parison of a direct- and a plate-counting technique for thequantitative estimation of soil micro-organisms. J. Gen. Mi-crobiol. 32, 261^271.

[18] SÖrheim, R., Torsvik, V.L. and GoksÖyr, J. (1989) Phenotyp-ical divergences between populations of soil bacteria isolatedon di¡erent media. Microb. Ecol. 17, 181^192.

[19] Fry, J.C. (1990) Direct methods and biomass estimation.Methods Microbiol. 22, 41^86.

[20] Kepner, R.J.J. and Pratt, J.R. (1994) Use of £uorochromesfor direct enumeration of total bacteria in environmental sam-ples: past and present. Microbiol. Rev. 58, 603^615.

[21] Herbert, R.A. (1990) Methods for enumerating microorgan-isms and determining biomass in natural environments. Meth-ods Microbiol. 22, 1^40.

[22] Amann, R.I., Ludwig, W. and Schleifer, K.-H. (1995) Phylo-genetic identi¢cation and in situ detection of individual micro-bial cells without cultivation. Microbiol. Rev. 59, 143^169.

[23] Daniel, O., Schoënholzer, F. and Zeyer, J. (1995) Quanti¢ca-tion of fungal hyphae in leaves of deciduous trees by auto-mated image analysis. Appl. Environ. Microbiol. 61, 3910^3918.

[24] Bloem, J., Veninga, M. and Shepherd, J. (1995) Fully auto-matic determination of soil bacterium numbers, cell volumes,and frequencies of dividing cells by confocal laser scanningmicroscopy and image analysis. Appl. Environ. Microbiol.61, 926^936.

[25] Sieracki, M.E., Johnson, P.W. and Sieburth, J.M. (1985) De-tection, enumeration, and sizing of planktonic bacteria byimage-analyzed epi£uorescence microscopy. Appl. Environ.Microbiol. 49, 799^810.

[26] Russ, J.C. (1995) The Image Processing Handbook. CRCPress, Tokyo, Japan.

[27] Richard, F., Luëscher, P. and Strobel, T. (1978) PhysikalischeEigenschaften von Boëden der Schweiz. WSL Birmensdorf,Switzerland.

[28] Schoënholzer, F., Kohli, L., Hahn, D., Daniel, O., Goez, C.and Zeyer, J. (1998) E¡ects of decomposition of leaves on



bacterial biomass and on palatability to Lumbricus terrestrisL. Soil Biol. Biochem. 30, 1805^1813.

[29] Daniel, O. and Anderson, J.M. (1992) Microbial biomass andactivity in contrasting soil materials after passage through thegut of the earthworm Lumbricus rubellus Ho¡meister. SoilBiol. Biochem. 24, 465^470.

[30] Kohli, L. (1997) Impact of environmental conditions and foodquality on earthworms. PhD Thesis No. 12189, ETH Zurich,Switzerland.

[31] Hahn, D., Amann, R.I., Ludwig, W., Akkermans, A.D.L. andSchleifer, K.-H. (1992) Detection of microorganisms in soilafter in situ hybridization with rRNA-targeted, £uorescentlylabelled oligonucleotides. J. Gen. Microbiol. 138, 879^887.

[32] Morgan, P., Cooper, C.J., Battersby, N.S., Lee, S.A., Lewis,S.T., Machin, T.M., Graham, S.C. and Watkinson, R.J.(1991) Automated image analysis method to determine fungalbiomass in soils and on solid matrices. Soil Biol. Biochem. 23,609^616.

[33] Amann, R.I., Krumholz, L. and Stahl, D.A. (1990) Fluores-cent-oligonucleotide probing of whole cells for determinative,phylogenetic, and environmental studies in microbiology.J. Bacteriol. 172, 762^770.

[34] Zarda, B., Hahn, D., Chatzinotas, A., Schoënhuber, W., Neef,A., Amann, R.I. and Zeyer, J. (1997) Analysis of bacterialcommunity structure in bulk soil by in situ hybridization.Arch. Microbiol. 168, 185^192.

[35] Newell, S.Y. (1992) The Fungal Community: Its Organisationand Role in the Ecosystem. Marcel Dekker, New York, USA.

[36] van Veen, J.A. and Paul, E.A. (1979) Conversion of biovo-lume measurements of soil organisms, grown under variousmoisture tensions, to biomass and their nutrient content.Appl. Environ. Microbiol. 37, 686^692.

[37] Anonymous (1992) Prism Image Analysis System. User'sManual (Rev. 3.0). Analytical Vision, Raleigh, 457.

[38] Marr, D. and Hildreth, E. (1980) Theory of edge detection.Proc. Roy. Soc. Lond. 207, 187^217.

[39] Bjornsen, P.K. (1986) Automatic determination of bacterio-plankton biomass by image analysis. Appl. Environ. Micro-biol. 51, 1199^1204.

[40] Avery, S.V., Harwood, J.L. and Lloyd, D. (1995) Quanti¢ca-tion and characterization of phagocytosis in the soil amoebaAcanthamoeba castellanii by £ow cytometry. Appl. Environ.Microbiol. 61, 1124^1132.

[41] Pace, M.L. and Baili¡, M.D. (1987) Evaluation of a £uores-cent microsphere technique for measuring grazing rates ofphagotrophic microorganisms. Mar. Ecol. 40, 185^193.

[42] Takada, K., Kato, K. and Okino, T. (1994) Image analysis offeeding activity of aquatic oligochaetes using £uorescent latexbeads. Verh. Internat. Verein. Limnol. 25, 2336^2340.

[43] Clegg, C.D., Anderson, J.M. and Lappin-Scott, H.M. (1996)Biophysical processes a¡ecting the transit of a genetically-modi¢ed Pseudomonas £uorescens through the gut of thewoodlouse Porcellio scaber. Soil Biol. Biochem. 28, 997^1004.

[44] Tiwari, S.C., Tiwari, B.K. and Mishra, R.R. (1990) Micro-fungal species associated with the gut content and cast ofDrawida assamensis Gates. Proc. Indian Acad. Sci. (PlantSci.) 100, 379^382.

[45] Domsch, K.H. and Banse, H.-J. (1972) Mykologische Unter-suchungen an Regenwurmexkrementen. Soil Biol. Biochem. 4,31^38.

[46] Harinikumar, K.M. and Bagyaraj, D.J. (1994) Potential ofearthworms, ants, millipedes, and termites for disseminationof vesicular-arbuscular mycorrhizal fungi in soil. Biol. Fertil.Soils 18, 115^118.

[47] Ho¡man, J.A. and Purdy, L.H. (1964) Germination of dwarfbunt (Tilletia controversa) teliospores after ingestion by earth-worms. Phytopathology 54, 878^879.

[48] West, A.W. (1988) Specimen preparation, stain type, and ex-traction and observation procedures as factors in the estima-tion of soil mycelial lengths and volumes by light microscopy.Biol. Fertil. Soils 7, 88^94.

[49] Lavelle, P. (1988) Earthworm activities and the soil system.Biol. Fertil. Soils 6, 237^251.

[50] Trigo, D. and Lavelle, P. (1993) Changes in respiration rateand some physicochemical properties of soil during gut transitthrough Allobophora molleri (Lumbricidae, Oligochaeta). Biol.Fertil. Soils 15, 185^188.

[51] Bakken, L. and Olsen, R. (1987) The relationship between cellsize and viability of soil bacteria. Microb. Ecol. 13, 103^114.

[52] Roszak, D.B. and Colwell, R.R. (1987) Survival strategies ofbacteria in the natural environment. Microbiol. Rev. 51, 365^379.

[53] Bae, H.C., Costa-Robles, E.H. and Casida, L.E. (1972) Mi-cro£ora of soil as viewed by transition electron microscopy.Appl. Microbiol. 23, 637^648.

[54] Fischer, K., Hahn, D., Hoënerlage, W. and Zeyer, J. (1997)E¡ect of passage through the gut of the earthworm Lumbricusterrestris L. on Bacillus megaterium studied by whole cell hy-bridization. Soil Biol. Biochem. 29, 1149^1152.



origins and fate of fungi and bacteria in the gut of lumbricus … · 2017-12-03 · origins and...

Documents