OUP user menu

Enumeration and characterization of cellulolytic bacteria from refuse of a landfill

Anne-Marie Pourcher, Laurent Sutra, Isabelle Hébé, Gérard Moguedet, Claude Bollet, Philippe Simoneau, Louis Gardan
DOI: http://dx.doi.org/10.1111/j.1574-6941.2001.tb00774.x 229-241 First published online: 1 January 2001


Enumeration and phenotypic characterization of aerobic cellulolytic bacteria were performed on fresh, 1 year old and 5 years old refuse samples of a French landfill site. Numbers of cellulolytic bacteria ranged from 1.1×106 to 2.3×108 c.f.u. (g dry wt.)−1 and were lower in 5 years old refuse samples. A numerical analysis of phenotypic data based on 80 biochemical tests and performed on 321 Gram-positive isolates from refuse, revealed a high phenotypic diversity of cellulolytic bacteria which were distributed into 21 clusters. Based on the phenotypic analysis and the sequencing of 16S rDNA of five representative strains of major clusters, the predominant cellulolytic groups could be assigned to the family of Bacillaceae and to the genera Cellulomonas, Microbacterium and Lactobacillus. Furthermore, chemical parameters such as pH, carbohydrates and volatile solid contents influenced the composition of the cellulolytic bacterial groups which were reduced essentially to the family of Bacillaceae in the oldest refuse samples.

  • Cellulolytic bacteria
  • Landfill site
  • Phenotypic analysis
  • 16S rDNA sequence analysis

1 Introduction

The landfilling of domestic refuse which represents about 60% of the municipal solid waste (MSW) disposal in France is the major method of waste disposal operation. Studies on landfills have been mainly devoted to waste composition, gas emission and physical parameters [1][2][3][4][5][6]. Despite the importance of microorganisms in the decomposition of organic matter, knowledge on the bacterial populations in refuse is still fragmentary. Most of the experimental data have been on quantification of trophic groups and have been generally obtained from lysimeters [9][7][8][9], and few previous studies have reported enumeration of heterotrophic bacteria in landfill sites [11][11]. As cellulose accounts for 40% to 50% of MSW [8], it is likely that cellulolytic microorganisms play an essential role in the degradation of refuse. Nevertheless, the cellulolytic bacteria involved in the degradation of MSW are poorly defined and only three genera including cellulose-decomposing species have been reported to be found in MSW: Cellulomonas, Clostridium and Eubacterium[12][13][14][15]. However, no information is available on the characterization of cellulolytic bacterial population in landfills. Therefore, many questions concerning the ecology and the identification of this trophic group remain to be answered. The objectives of this study were to quantify total culturable aerobic heterotrophic and cellulolytic bacteria and to examine and compare the microbial diversity of the cellulolytic flora of 29 samples of three types of waste: fresh refuse, 1 year old refuse and 5 years old refuse. A numerical phenotypic analysis based on 80 enzymatic and fermentation tests was performed on 321 Gram-positive isolates and was completed by 16S rDNA sequence analysis of five representative strains of major phenotypic clusters.

2 Materials and methods

2.1 Collection and preparation of samples

Refuse samples were obtained from Séché-Eco Industry Landfill, Laval, France, over a 2 year period. A total of 27 samples including nineteen 1 year old refuse samples and eight 5 years old refuse samples were collected using a bucket auger at a depth of 1 m. Two additional samples of fresh refuse were collected from a trash and on the landfill site. Seven samples of 1 year old refuse were obtained on the landfill from a large area (area 1) and were separated from each other by at least 100 m. Twelve samples of 1 year old refuse and eight samples of 5 years old refuse were collected on two surface areas of 400 m2 (area 2) and 200 m2 (area 3), respectively; excavations were separated from each other by at least 5 m. At each excavation, samples of 0.5–1 kg were removed from the pile of excavated refuse and were transferred into a plastic bag until a final weight of about 5 kg. Samples were stored at room temperature and analyzed within 24 h after collection. Samples were prepared as previously described by Bichet et al. [16]. Briefly, 125 g of refuse was mixed for 1 min in a Warring Blender with 600 ml of phosphate buffer (KH2PO4, 1.61 g l; K2HPO4, 2.07 g l; pH 7.2). A 10 g sample of the obtained slurry was then sonicated (2 min, 80 W) and hand-squeezed through a sieve of 250 μm mesh.

2.2 Chemical analyses

Five chemical parameters were measured on refuse samples excavated from areas 2 and 3, including moisture content, pH, volatile solids, carbohydrates and cellulose contents.

Moisture content was measured by drying refuse (5×100 g) at 105°C until constant weight. Volatile solid percentages were then determined by igniting dry refuse (3×1 g) for 1 h at 550°C. Carbohydrates and cellulose contents and pH of refuse were determined after preliminary mixing for 1 min in distilled water with a Warring Blender. The obtained slurry was then hand-squeezed through a sieve of 250 μm mesh and the pH of the filtrate was measured in duplicate. Carbohydrates and cellulose contents were determined in triplicate on 1 g of dry slurry. After acid hydrolysis, carbohydrates were measured by the anthrone method described by Gaudy [17]. Cellulose was extracted from the dry slurry with a solution of ethanol–toluene (1:2) in a Soxhlet apparatus through a minimum of 25 cycles and washed thoroughly with ethanol and then with water. Cellulose content was determined enzymatically with a glucose oxidase kit (Sigma, no. 510-A) from the liquid extract as described by Leschine and Canale-Parola [18].

2.3 Enumeration of bacteria

Culturable heterotrophic bacteria and cellulolytic bacteria which were enumerated or identified under aerobic conditions included strictly aerobic and aero-anaerobic bacteria.

Total culturable heterotrophic bacteria were estimated by spreading triplicate serial 10-fold dilutions of hand-squeezed waste extract on the surface of Plate Count Agar (PCA, Biokar). Colonies were counted after incubation at 36°C for 8 days under aerobic conditions. Cellulolytic bacteria were enumerated by an indirect method as follows: a volume of 0.1 ml of each serial dilution was plated on the basal medium (Medium B) described by Bagnara et al. [12], containing 2% (w/v) agar and supplemented with 0.4% (w/v) cellobiose. Plates were inoculated in triplicate and incubated at 36°C for 8 days. For plates containing five to 100 colonies, each colony was picked and inoculated into a well of a 96-well microtitration plate containing Medium B supplemented with 0.01% (w/v) 4-methylumbelliferyl β-d-cellobioside (Sigma M 6018). Utilization of cellobiose was detected by fluorescence of the medium under UV light (365 nm). After incubation for 72 h at 36°C, 50 μl of each fluorescent well was transferred into a tube containing Medium B and a cigarette paper strip (0.5×2 cm; OCB Bolloré, Quimper, France). Tubes were incubated at 36°C for 3 months. Cellulose degradation was detected by visible disappearance of cigarette paper. Paper degrading, i.e. cellulolytic bacteria, were transferred to PCA prior to phenotypic characterization.

Statistical comparisons of mean values of bacterial densities were made with the standard t-test [19].

2.4 Bacterial strains

A total of 350 strains were included in the phenotypic analyses, including 29 reference strains obtained from culture collections and 321 strains, among which 11 were isolated from fresh refuse, 72 were isolated from area 1, 124 from area 2 and 114 from area 3 of the landfill. The reference strains used are listed in Table 1.

View this table:

Reference strains used in the study

SpeciesCollection numbersaIsolated from
Bacillus acidovoransCIP 103552soil
Bacillus amyloliquefaciensCIP 103265Tsoil
Bacillus benzoevoransCIP 103477Tsoil
Bacillus circulansCIP 52.76soil
Bacillus coagulansCIP 66.25Tmilk powder
Bacillus licheniformisCIP 52.71Tnot known
Bacillus subtilisCIP 52.65Tnot known
Bacillus viscosusCIP 103554soil
Brevibacillus laterosporusCIP 52.83Tnot known
Paenibacillus alginolyticusCIP 103122Tsoil
Paenibacillus amylolyticusCIP 103117Tsoil
Paenibacillus chondroitinusCIP 103123Tsoil
Paenibacillus maceransCIP 66.19Tsoil
Paenibacillus pabuliCIP 103119Tfermenting fodder
Paenibacillus polymyxaCIP 66.22Tnot known
Paenibacillus validusCIP 103120Tsoil
Virgibacillus pantothenticusCIP 51.24Tsoil
Cellulomonas biazoteaCIP 82.11Tsoil
Cellulomonas cellaseaCIP 102227Tsoil
Cellulomonas cellulansCIP 103404Tsoil
Cellulomonas fermentansCIP 103003Tmunicipal dumping ground
Cellulomonas fimiCIP 102114Tsoil
Cellulomonas flavigenaCIP 82.10Tnot known
Cellulomonas gelidaCIP 102221Tsoil
Cellulomonas hominisDSM 9581Tcerebrospinal fluid
Cellulomonas turbataCIP 100331Tsoil
Cellulomonas udaCIP 102089Tcompost
Promicromonospora citreaDSM 43110Tsoil
Promicromonospora sukumoeDSM 43121Tsoil
  • a CIP, Collection de l'Institut Pasteur, Paris, France; DSM, Deutsche Sammlung von Mikroorganismen, Braunschweig, Germany.

2.5 Biochemical and physiological tests

Gram staining, presence of catalase and oxidase activities, hydrolysis of starch and DNA and reduction of nitrate were tested as described by Smibert and Krieg [20]. Endospores were stained by using the Schaeffer–Fulton staining method [20]. Three types of commercial strips (bioMérieux, Marcy-l'Etoile, France) were used: API 20E strips (20 enzymatic and carbohydrate fermentation tests), API ZYM strips (19 enzymatic tests) and API 50CH strips (fermentation of 48 carbohydrates and hydrolysis of esculin). Strips were inoculated as recommended by the manufacturer and incubated at 36°C. Results were read after 5 h for API ZYM strips, 24 h for API 20E strips and 96 h for API 50CH strips.

2.6 Numerical taxonomy

All strains were characterized using 80 phenotypic tests: catalase, oxidase, DNA and starch hydrolysis, nitrate reduction, seven enzymatic tests of API 20E strips (β-galactosidase, arginine dihydrolase, lysine decarboxylase, ornithine decarboxylase, citrate utilization, Voges Proskauer test and gelatin hydrolysis), 19 enzymatic activities (API ZYM) and fermentation of 48 carbohydrates and esculin hydrolysis (API 50CH). A distance matrix was calculated using the Jaccard coefficient [21]. Cluster analysis was done by using the unweighted pair group method with averages (UPGMA) [21]. At a given level of phenotypic distance, the amount of information for each test was measured by calculating the diagnostic ability coefficient [22] to determine the discriminating biochemical characteristics between phenons or strains.

2.7 16S rRNA gene sequencing

The 16S rDNA sequences of five strains isolated from refuse (strains P8, DOT13, I33, K49, EE17) were determined. Extraction of DNA was carried out by using a QIAmp kit (Qiagen, Hilden, Germany). PCR-mediated amplification of the 16S rRNA gene and sequence determination were performed as previously described [23]. Sequences determined in this study were compared with 16S rDNA sequences obtained from the GenBank database. Sequences were aligned using the Clustal W program [24] and a few further adjustments were made by sight where necessary. Phylogenetic analyses were performed with programs contained in the package PHYLIP version 3.572 [25]. Phylogenetic trees were constructed by using the neighbor-joining method [26]; bootstrap analysis with 300 bootstrapped data sets was performed to determine the statistical support of the branches of neighbor-joining trees.

3 Results

3.1 Microbiological counts

The total culturable heterotrophic bacteria ranged from 8.9×106 to 1.2×109 c.f.u. (g dry wt.)−1. Cellulolytic bacteria were present in all refuse samples and ranged from 1.1×106 to 2.3×108 c.f.u. (g dry wt.)−1 according to the sample (Table 2). The proportion of cellulolytic bacteria to the total culturable heterotrophic bacteria varied according to samples, ranging from 3.7 to 75%. The lowest numbers of bacteria were found in the oldest refuse (area 3). No significant differences were observed between average numbers of heterotrophic bacteria of area 2 (1 year old refuse) and area 3 (5 years old refuse) but average numbers of cellulolytic bacteria and their proportion among culturable heterotrophic bacteria were significantly lower in samples from area 3 than in samples from area 2 (P<0.05). A marked decrease of cellulose content was also observed in area 3 which showed a significantly lower cellulose level than area 2 (P<0.05). Nevertheless, the variability in the numbers of bacteria was high within samples of the same age, whereas the chemical parameters, including pH, moisture level, volatile solids, carbohydrates and cellulose contents, were stable within each area.

View this table:

Chemical and microbial parameters of fresh refuse and of refuse collected in areas 1, 2 and 3 of the landfill

ParametersFresh refuse (2)a1 Year old refuse area 1 (7)1 Year old refuse area 2 (12)5 Years old refuse area 3 (8)
Total culturable heterotrophic bacteria (THB) (10−7×c.f.u. (g dry wt.)−1)38.53.542.637.
Culturable cellulolytic bacteria (CB) (10−7×c.f.u. (g dry wt.)−1)4.42.612.910.
R c11.57.832.023.444.713.422.08.7
Moisture (% wet wt.)nddndndnd48.34.746.75.0
Volatile solids (% dry wt.)ndndndnd77.64.942.51.4
Carbohydrates (% dry wt.)ndndndnd41.
Cellulose (% dry wt.)ndndndnd32.
  • a Number of samples analyzed.

  • b Standard deviation.

  • c R=100×(CB/THB).

  • d nd, not determined.

3.2 Phenotypic characterization of cellulolytic bacteria

A preliminary phenotypic analysis, performed on 355 cellulolytic bacterial isolates from the 29 refuse samples revealed three groups of bacteria: Gram-positive bacilli (321 strains), Gram-negative bacilli (31 strains) and Gram-positive, catalase negative cocci (three strains) which represented ca. 90%, 9% and 1% of the total isolated strains, respectively [27]. Among the Gram-negative isolates, seven strains were identified as Cytophagaceae-like bacteria and 24 strains were assigned to the family of Enterobacteriaceae among which three strains were identified as Serratia marcescens by their API 20E profile [27]. Three Gram-positive cocci were identified as Enterococcus sp. on the basis of the following characteristics: catalase negative, hydrolysis of esculin and production of the group D antigen (data not shown). They were identified as Enterococcus faecium by electrophoretic analysis of the whole cell proteins (kindly performed by Dr. Maarit Niemi, Finnish Environment Institute, Helsinki, Finland).

The numerical analysis restricted to the dominant group of the 321 Gram-positive bacilli and 29 reference strains was performed on the basis of 80 characteristics. At the similarity level of 67%, the dendrogram of phenotypic distances (Fig. 1) showed 21 clusters of at least two strains. Among these 21 clusters, 15 (I, II, V, IX, X, XI and XIII–XXI) contained endospore-forming bacteria and six (III, IV, VI, VII, VIII and XII) included non-endospore-forming bacteria. Thirty strains formed separate branches, including the type strains of Bacillus benzoevorans, Bacillus coagulans, Brevibacillus laterosporus, Paenibacillus validus, Virgibacillus panthotanticus and Bacillus acidovorans CIP 103552 and 24 bacterial isolates from refuse.


Dendrogram of phenotypic distances of 321 Gram-positive strains isolated from refuse samples and 29 reference strains. *Pr, Promicromonsopora.

3.2.1 Endospore-forming bacteria

The 211 endospore-forming bacteria isolated from refuse were assigned to the family of Bacillaceae according to their morphological and biochemical characteristics. The most important cluster (cluster I) contained 135 strains isolated from refuse and three type strains of Bacillus species (Bacillus licheniformis, Bacillus subtilis and Bacillus amyloliquefaciens). Cluster II included 37 strains isolated from refuse, six type strains of Bacillus and Paenibacillus species (Bacillus pabuli, Paenibacillus amylolyticus, Paenibacillus polymyxa, Paenibacillus alginolyticus, Paenibacillus chondroitinus and Paenibacillus macerans) and Bacillus circulans CIP 52.76. Two type strains of Promicromonospora species (Promicromonospora citrea and Promicromonospora sukumoe) were also included in cluster II, as sub-cluster IIg. This result indicates that these two Promicromospora species which belong to the class of Actinobacteria[28] present phenotypic similarities with the family of Bacillaceae.

Except cluster XVI, which included Bacillus viscosus strain CIP 103554, the other 12 clusters of endospore-forming bacteria did not contain any reference strain (Fig. 1). Clusters I and II which included 81.5% of the endospore-forming bacteria, can be differentiated only by the fermentation of β-methyl-d-xyloside. All strains of cluster II fermented this carbohydrate whereas strains of cluster I did not. At a similarity level of 78%, clusters I and II were subdivided into seven sub-clusters (Ia–Ig) and ten sub-clusters (IIa–IIj), respectively. The phenotypic characteristics that differentiate the major sub-clusters are presented in Table 3. Sub-cluster Ia (72 strains including Bacillus subtilis and Bacillus licheniformis type strains), sub-cluster Ib (34 strains) and sub-cluster Ic (four strains) which accounted for 80% of the strains of cluster I, were closely related and were distinguished from sub-clusters Ie and If by two biochemical characteristics: fermentation of sorbitol and mannitol. The five sub-clusters IIa, IIb, IIc, IIe and IIf which accounted for 65% of strains of cluster II were characterized by a high ability to ferment the carbohydrates tested. Strains of the two major sub-clusters IIa (nine strains) and IIf (ten strains) could be separated by hydrolysis of gelatin and fermentation of gluconate. Sub-cluster IIc which included the type strains of Paenibacillus amylolyticus and Paenibacillus polymyxa and sub-cluster IIe which included Bacillus circulans CIP 52.76, were differentiated by the fermentation of three carbohydrates: sorbitol, xylitol and l-fucose.

View this table:

Characteristics that differentiate major sub-clusters of endospore-forming bacteria defined at the 78% similarity level

Biochemical characteristics% of positive strains in sub-clustera (number of strains)
Iab (72)Ib (34)Ic (4)Ie (12)If (4)IIa (9)IIb (4)IIcc (4)IIed (3)IIf (10)
Acid from
Enzymatic tests
Gelatin hydrolysis8755+25752533+
Nitrate reduction+++162577757533+
  • a +, 100% of strains are positive; −, 100% of strains are negative.

  • b Sub-cluster containing the type strains of Bacillus subtilis and Bacillus licheniformis.

  • c Sub-cluster containing the type strains of Paenibacillus amylolyticus and Paenibacillus polymyxa.

  • d Sub-cluster containing Bacillus circulans reference strain CIP 52.76.

3.2.2 Non-endospore-forming bacteria

Non-endospore-forming bacteria were distributed among six clusters. Except for cluster VII, these clusters contained strains which formed pigmented colonies. Three clusters (III, IV and VIII) were constituted of 79 readily decolorized short bacilli, 95% of which showed circular, convex and yellow pigmented colonies on agar medium. Clusters III and IV contained the ten type strains of species belonging to the genus Cellulomonas. Cellulomonas cellulans, Cellulomonas hominis and Cellulomonas turbata type strains were grouped in cluster III with 32 strains isolated from refuse, whereas the seven other type strains of Cellulomonas were gathered together in cluster IV with four strains of refuse. Strains of clusters III and IV presented closed biochemical profiles and could be differentiated only by the α-galactosidase activity which was produced by all strains of cluster IV. Cluster VIII contained 33 oxidase negative and catalase positive short bacilli isolated from refuse which showed morphological similarity with strains of clusters III and IV. Nevertheless, strains of cluster VIII fermented less carbohydrates than strains of clusters III and IV. They could be differentiated from strains of cluster III by their inability to ferment N-acetyl-glucosamine and from strains of cluster IV by their inability to ferment starch and glycogen.

At a similarity level of 78%, clusters III, IV and VIII were subdivided into sub-clusters IIIa–IIId, IVa and IVb and VIIIa–VIIId, respectively, the biochemical characteristics of which are reported in Table 4. Sub-clusters IIIb and IIIc contained Cellulomonas cellulans and Cellulomonas hominis type strains, respectively. Strains of sub-clusters IIIa–IIId could be differentiated by five biochemical characteristics: fermentation of gluconate, melezitose and β-methyl-d-xyloside, production of catalase and nitrate reduction (Table 4). Sub-cluster IIId contained three strains from refuse which grew weakly on PCA medium and produced white circular colonies with diameter of less than 1 mm after incubation at 36°C for 48 h. Similar characteristics were observed for the Cellulomonas cellulans type strain. Sub-cluster IVa contained Cellulomonas biazotea and Cellulomonas fimi type strains and two strains isolated from refuse, and sub-cluster IVb was constituted of Cellulomonas flavigena and Cellulomonas uda type strains. These two sub-clusters could be distinguished from each other by fermentation of amygdaline and N-acetyl-glucosamine. The more important sub-clusters of cluster VIII, sub-clusters VIIIa (18 strains) and VIIIc (6 strains), could be differentiated by fermentation of d-xylose, melezitose and salicine (Table 4).

View this table:

Characteristics that differentiate ten sub-clusters of Actinobacteria-like bacteria defined at the 78% similarity level

Biochemical characteristics% of positive strains in sub-clustera (number of strains)
IIIa (16)IIIbb (13)IIIcc (2)IIId (3)IVad (4)IVbe (2)VIIIa (18)VIIIb (2)VIIIc (6)VIIId (2)
Acid from
Enzymatic tests
Nitrate reduction++50+1616
  • a +, 100% of strains are positive; −, 100% of strains are negative.

  • b Sub-cluster containing the type strain of Cellulomonas cellulans.

  • c Sub-cluster containing the type strain of Cellulomonas hominis.

  • d Sub-cluster containing the type strains of Cellulomonas biazotea and Cellulomonas fimi.

  • e Sub-cluster containing the type strains of Cellulomonas flavigena and Cellulomonas uda.

The other non-endospore-forming bacteria were distributed among three clusters (clusters VI, VII and XII). Cluster VI contained two strains isolated from refuse which consisted of short bacilli readily decolorized and which formed mucous orange-pink colonies on agar medium. They were catalase negative, oxidase positive and reduce nitrate. The three strains of cluster XII which formed orange (becoming brown) colonies were the only strains with branched septate mycelium isolated from refuse samples. Cluster VII was constituted of 17 short bacilli which formed white small colonies on agar medium after incubation at 36°C for 72 h. All strains of cluster VII were oxidase and catalase negative, did not reduce nitrate and did not hydrolyze gelatin (data not shown).

3.3 16S rDNA sequences

In order to better characterize certain strains or groups of strains, the nucleotide sequences of the 16S rDNA of five representative strains of sub-cluster IIf (strain EE17), cluster VII (strain P8), sub-cluster IIIb (strain DOT13), sub-cluster VIIIa (strain I33) and sub-cluster VIIIc (strain K49) were determined. The tree shown in Fig. 2 revealed that strain EE17 exhibited high level of 16S rDNA similarity (99%) with Paenibacillus macerans and Paenibacillus polymyxa. As strain EE17, the type strains of these two Paenibacillus species were included in cluster II. Strain P8 which fell in cluster VII with 16 strains isolated from refuse was highly related to Lactobacillus paracasei (99.9% sequence similarity). The 16S rDNA of strain DOT13 which belonged to sub-cluster IIIb together with the type strain of Cellulomonas cellulans had 100% similarity with the 16S rDNA of Cellulomonas cellulans. Sequencing of the 16S rRNA genes of strains I33 (sub-cluster VIIIa) and K49 (sub-cluster VIIIc) revealed that strain I33 was related to Microbacterium testaceum (99.7% similarity) and that strain K49 was related to Microbacterium liquefaciens (99.6% similarity). Considering the results of 16S rDNA sequencing and the phenotypic data, strains of clusters III, IV and VIII could be assigned to the genera Cellulomonas and Microbacterium which belong to the class of Actinobacteria.


Neighbor-joining tree based on 16S rDNA sequence analysis. The scale bar indicates 0.01 substitution per nucleotide position. Bold values indicate the level (%) of bootstrap support (300 re-sampled data sets); only values that were >75% are given. In parentheses are indicated the sequence accession numbers.

3.4 Distribution of cellulolytic bacteria

The distribution of the dominant aerobic phenotypic groups obtained from refuse samples is presented in Table 5. The presence of endospore-forming bacteria, assigned to the family of Bacillaceae, was detected in 28 of the 29 samples regardless the age of the refuse whereas non-endospore-forming bacteria were essentially found in fresh refuse and in 1 year old refuse samples. Among endospore-forming bacteria, strains of cluster I were isolated from all types of refuse samples whereas strains of cluster II were only recovered from samples of areas 2 and 3. The endospore-forming bacteria of other clusters were present in the three areas but were not detected in the two fresh refuse samples. The distribution of the Actinobacteria-like bacteria differed according to their phenotypic group: strains assigned to the genus Cellulomonas were recovered from fresh refuse and from the three landfill areas, whereas all strains assigned to the genus Microbacterium were isolated only in area 2. The other phenotypic groups were found either in areas 1 and 2 as strains assigned to the genus Lactobacillus and to the family of Enterobacteriaceae, or in a unique area as strains assigned to the genus Enterococcus and to the family of Cytophagaceae.

View this table:

Occurrence of phenotypic groups in 29 refuse samples

Phenotypic group% of refuse samples that contained a phenotypic group
Fresh refuse (2)a1 Year old refuse area 1 (7)1 Year old refuse area 2 (12)5 Years old refuse area 3 (8)
Endospore-forming bacteria10010083.3100
– Cluster I10062.566.687.5
– Cluster II0041.6100
– Others071.458.362.5
Cellulomonas (clusters III and IV)10042.866.612.5
Microbacterium (cluster VIII)0058.30
Lactobacillus (cluster VII)014.3250
Branched mycelium (cluster XII)00025
  • a Number of samples analyzed.

4 Discussion

Microbial degradation of cellulose appears as an important metabolic activity in landfill where cellulose and hemicellulose are considered to represent 91% of the methane potential of fresh refuse according to Barlaz et al. [29]. Previous studies which concerned cellulolytic bacteria of landfill have focused only on counting bacteria [8][10][11] or on measuring global cellulolytic activity [7][10][30][31]. To our knowledge, phenotypic studies of cellulolytic bacteria of MSW have been restricted to the identification of one strain of Cellulomonas[12] and to the characterization of nine anaerobic cellulolytic bacterial isolates [15]. The present study represents the first report on both quantification and phenotypic characterization of aerobic cellulolytic bacterial isolates from refuse samples collected in a landfill. Our results showed that densities of total culturable heterotrophic and cellulolytic bacteria ranged from ca. 107 to 109 and 106 to 108 c.f.u. (g dry wt.)−1, respectively (Table 2). Densities of cellulolytic bacteria were higher than those reported by Jones et al. [10] who detected 104 to 106 c.f.u. (g dry wt.)−1 of cellulolytic bacteria at different depths of the landfill site of Aveley (Essex, UK) and by Palmisano et al. [11] who did not detect any cellulolytic bacteria in Fresh Kill Landfill (New York, USA). In these previous studies, cellulolytic bacteria were numerated by inoculating the refuse extract directly on cellulose-containing medium and incubating plates under anaerobic condition. In our study, refuse extracts were first inoculated on a solid medium supplemented with cellobiose, a less selective substrate than cellulose. As plates were incubated under aerobic condition, cellobiolytic strains which had grown on this first isolation medium likely included aerobic and facultative aerobic bacteria. The strains were then inoculated in a liquid medium containing a paper strip and cellulose degradation was examined during 12 weeks. The high counts of cellulolytic bacteria that we obtained could be explained by the use of this indirect cultural method, which allowed a better recovery of cellulolytic bacteria from refuse as previously observed by Bichet [27]. However, counting of cellulolytic bacteria was variable among refuse samples of the same age, contrary to chemical parameters (pH, moisture level, volatile solids, carbohydrates and cellulose contents) measured in area 1 (1 year old refuse) and area 2 (5 years old refuse) which were stable within each area. The fact that the density of cellulolytic bacteria was higher in area 1 than in area 2 could be related to the cellulose content, which was about 10-fold inferior in 5 years old refuse (area 1) than in 1 year old refuse (area 2). Hydrolytic processes of MSW are known to occur mainly under anaerobic condition but controversy exists about whether hydrolytic bacteria are strictly or facultative anaerobic [32]. In our study, strictly anaerobic cellulolytic bacteria numerated in samples of areas 2 and 3 were about 10-fold less abundant than facultative aerobic cellulolytic bacteria (data not shown). So, the high counts of aerobic or facultative aerobic cellulolytic bacteria that we observed in refuse samples strongly suggest that these bacteria play an important role in the hydrolysis of cellulose.

The phenotypic analysis of 355 isolates from 29 refuse samples revealed a considerable heterogeneity of the aerobic cellulolytic bacterial groups found in MSW. In our study, strains assigned to the family of Bacillaceae and to the class of Actinobacteria represented the most abundant cellulolytic bacteria isolated from refuse. The abundance of strains of both the Bacillus and the Paenibacillus genera in refuse samples is not surprising since these two ubiquitous genera are known to include cellulolytic species and are commonly found in soil and plant litter and compost [33][34] where they play a major role in the decomposition of organic matter. Eighty-two percent of the strains assigned to Bacillaceae were aggregated with 11 reference strains of Bacillus or Paenibacillus at a similarity level of 67% (Fig. 1). Strains of cluster I aggregated with type strains of Bacillus licheniformis, Bacillus subtilis and Bacillus amyloliquefaciens, three species that have been previously described for their ability to degrade polysaccharides of plant tissues [33]. Strains of cluster II aggregated with Bacillus circulans CIP 52.76 which is known for its cellulolytic activity [33] and with seven type strains of the genus Paenibacillus among which six species were previously described for their ability to hydrolyze carboxymethylcellulose [35]. The 16S rDNA sequence analysis showed that one strain of sub-cluster IIf (strain EE17) was closely related to Paenibacillus macerans and Paenibacillus polymyxa (Fig. 2), the latter species not being known for its cellulolytic activity [35].

The second major cellulolytic group isolated from MSW was constituted of strains assigned to the class of Actinobacteria which were distributed into clusters III, IV and VIII. Two of these clusters (III and IV) contained the ten type strains of Cellulomonas described at the moment, all of them, except Cellulomonas hominis, being cellulolytic [36]. Except for gelatin hydrolysis, the phenotypic characteristics of the type strains of Cellulomonas that we obtained in this study agreed with those previously published [36]. The 16S rDNA of strain DOT13 of sub-cluster IIIb which contained the type strain of Cellulomonas cellulans, had 100% similarity with the 16S rDNA of Cellulomonas cellulans. Taken together these results indicated that sub-cluster IIIb was likely to correspond to Cellulomonas cellulans. The phenotypic position of the type strains of Cellulomonas species in the dendrogram (Fig. 1) was consistent with the phylogenetic relationships of Cellulomonas species reported by Rainey et al. [37]. For example, these authors observed that the type strains of Cellulomonas biazotea and Cellulomonas fimi that are included in sub-cluster IVa in our study, exhibited a high level of 16S rDNA homology (99.7%). Furthermore, they reported that Cellulomonas cellulans was the most divergent species in the genus according to the levels of 16S rRNA similarity with the other Cellulomonas species. In our phenotypic analysis, the type strain of Cellulomonas cellulans was indeed separated from the other type strains of Cellulomonas at a similarity level of 78% (Fig. 1). However, the taxonomic status of sub-cluster IIIa (16 strains) and sub-cluster IIId (three strains), which did not contain any type strain of Cellulomonas species, has to be clarified. The morphology and the general biochemical characteristics of strains of cluster VIII (catalase positive, oxidase negative, indole negative and esculin hydrolysis) were consistent with the previous description of the genus Aureobacterium[38], which was recently combined to the genus Microbacterium[39]. The 16S rDNA sequence analysis of representative strains of sub-clusters VIIIa and VIIIc showed that strains of these sub-clusters were closely related to Microbacterium testaceum and Microbacterium liquefaciens, respectively, and thus could be identified to these two species which have never been reported to be cellulolytic [38]. Nevertheless, we observed a phenotypic diversity among strains aggregated in cluster VIII. Further taxonomic investigation based on 16S rDNA analysis and DNA–DNA hybridization are necessary to identify these strains more precisely.

Among Gram-positive cellulolytic bacteria included in the taxonomic analysis, 17 strains included in cluster VII showed morphological and biochemical characteristics of the genus Lactobacillus, which is frequently found in carbohydrate-rich natural habitats but the cellulolytic activity of which has not yet been reported [40]. The 16S rDNA sequence of a representative strain (strain P8) of this cluster which had 99.9% similarity with the 16S rDNA of ‘Lactobacillus paracasei’ confirmed that strains of cluster VII likely correspond to the genus Lactobacillus. According to their morphological and biochemical characteristics, the three strains of cluster XII which formed branched mycelium could be assigned to the cellulolytic genus Micromonospora[41] recently included in the class of Actinobacteria[28].

The other cellulolytic strains isolated from refuse samples included Gram-positive cocci and Gram-negative bacilli which have been characterized in a previous taxonomic analysis [27]. Three Gram-positive cocci were identified as Enterococcus faecium which is not known as a cellulolytic species. Seven strains were assigned to the family of Cytophagaceae, which include cellulolytic species isolated from soils [42] and 24 strains were assigned to the family of Enterobacteriaceae, the cellulolytic activity of which has only been reported for a few species belonging to the genera Serratia or Erwinia[43][44]. Nevertheless, the presence of Enterobacteriaceae, particularly strains identified to the genus Serratia, has been reported in aerosols of domestic refuse composting facilities by Reinthaler et al. [45]. Among the 24 strains of Enterobacteriaceae, three strains were identified as Serratia marcescens, a species which includes strains generally isolated from water, soil and plants [46]. The cellulolytic activity of this species has been reported by Thayer [43], who observed a moderate production of carboxymethylcellulase by a strain isolated from the hind-gut of a termite.

To date, only one study described cellulolytic bacterial diversity of a landfill site [15]. Nevertheless, the selection of bacterial isolates used by Westlake et al. [15] was not exhaustive and only 37 strains were isolated among which only nine strains were identified to a genus. The present study, performed on 355 isolates, revealed different levels of diversity among cellulolytic bacteria between 1 year old refuse and 5 years old refuse. Strains isolated from 1 year old refuse belonged to various families or genera, mainly Bacillaceae, Cellulomonas, Microbacterium but also Cytophagaceae, Enterobacteriaceae, Lactobacillus and Enterococcus (Fig. 1 and Table 5). A high diversity of the soil aerobic cellulolytic population was also reported by Ulrich and Wirth [47] who identified 311 isolates from soil samples by restriction analysis of PCR-amplified 16S rDNA. Similar to these authors, we identified strains belonging to Bacillus, Paenibacillus and Cellulomonas genera but we did not detect any strain belonging to the strictly aerobic genus Streptomyces, which was predominant in soil samples analyzed by Ulrich and Wirth [47]. Contrary to these authors, who reported a relative stability of the community of cellulolytic bacteria in soil samples from different sites, we observed that environmental parameters related to the age of refuse influenced the composition of cellulolytic bacteria of waste. Most of the cellulolytic isolates from 5 years old refuse were endospore-forming bacteria and presumably belonged to the family of Bacillaceae. The lower contents of volatile solids and carbohydrates and the higher value of pH in 5 years old refuse probably contributed to favor the development of most adaptable bacteria like Bacillaceae.

In this paper, cellulolytic bacteria from landfill refuse have been studied using a two-step experimental approach: isolation by an original protocol that allowed a high recovery of bacterial isolates and then numerical analysis of phenotypic data together with 16S rDNA sequence analysis of some representative isolates. Phenotypic data revealed the complexity of the composition of cellulolytic bacterial flora particularly in recent, i.e. 1 year old refuse. The 16S rDNA sequence analysis allowed the identification of some phenotypic clusters at the genus or species level. Interestingly, some of the dominant cellulolytic bacteria belonged to genera which were not previously reported to hydrolyze cellulose, i.e. Microbacterium, Lactobacillus and Enterococcus. However, the taxonomic status of numerous phenotypic clusters has to be clarified using a polyphasic approach based on both phenotypic and genomic methods. The phenotypic approach described in this paper could be completed by a molecular approach, for example based on 16S rDNA sequences [47] to improve the knowledge of trophic groups involved in the decomposition of cellulolytic material in refuse. Furthermore, degradation potentials under aerobic and anaerobic conditions of strains belonging to the different clusters would be considered in further studies.


This work was supported by the French Environment and Energy Agency (ADEME) and SECHE ECO INDUSTRIES Company. The authors thank Dr. Maarit Niemi (Finnish Environment Institute, Helsinki, Finland) for the identification of Enterococcus sp. strains.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
View Abstract