Bioresource Technology 100 (2009) 2311–2315

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /b ior tech

Short Communication

Impact of crop species on bacterial community structure during anaerobicco-digestion of crops and cow manure

Hong Wang a,*, Annimari Lehtomäki b, Katariina Tolvanen c, Jaakko Puhakka c, Jukka Rintala a

a Department of Biological and Environmental Science, University of Jyväskylä, P.O. Box 35, FI-40014 Jyväskylä, Finlandb Jyväskylä Innovation Ltd., P.O. Box 27, FI-40203 Jyväskylä, Finlandc Institute of Environmental Engineering and Biotechnology, Tampere University of Technology, P.O. Box 541, FI-33101 Tampere, Finland

a r t i c l e i n f o

Article history:Received 28 March 2008Received in revised form 13 October 2008Accepted 16 October 2008Available online 4 December 2008

Keywords:Anaerobic digestionContinuously stirred tank reactorBacterial community structureTerminal restriction fragment lengthpolymorphismDenature gradient gel electrophoresis

0960-8524/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.biortech.2008.10.040

* Corresponding author. Tel.: +358 14 2604248; faxE-mail address: whong@cc.jyu.fi (H. Wang).

a b s t r a c t

The bacterial communities in three continuously stirred tank reactors co-digesting cow manure withgrass silage, oat straw, and sugar beet tops, respectively, were investigated by 16S rRNA gene-based fin-gerprints and clone libraries. The analyses revealed both clearly distinct and similar phylotypes in thebacterial communities between the reactors. The major groups represented in the three reactors wereClostridia, unclassified Bacteria, and Bacteroidetes. Phylotypes affiliated with Bacilli or Deltaproteobacteriawere unique to the sugar beet and straw reactor, respectively. Unclassified Bacteria dominated in sugarbeet reactor while in the straw and grass reactor Clostridia was the dominant group. An increase inorganic loading rate from 2 to 3 kg volatile solids m�3 d�1 resulted in larger changes in the bacterial com-munity in the straw compared to grass reactor. The study shed more light on the evolution of bacterialcommunity during anaerobic co-digestion of different crops and manure to methane.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Crops and crop residues represent a large unexploited potentialfor methane production. The methane yield mainly depends on theorganic content of crops degradable in anaerobic digestion (Amonet al., 2007). Cellulose together with hemicelluloses and lignin, themost abundant component of crop biomass, are found in the cellwall. Cell walls across crop taxa differ importantly in compositionand the anatomical structure. Crops with poorly lignified cell wallsare easily degraded, whereas crops with highly lignified walls mayconstrain degradation (Wilson and Mertens, 1995).

Anaerobic digestion of cellulosic material to methane is a multi-step process mediated by Bacteria and methanogenic Archaea(Chynoweth and Pullammanappallil, 1996). The polymers arehydrolyzed into soluble compounds under fermentative condition.Acidogens and acetogenic bacteria convert these intermediatesinto acetate and one-carbon compounds which can be converteddirectly by methanogenic Archaea into methane and carbon diox-ide. Generally, hydrolysis is considered to be the rate-limiting stepin the anaerobic digestion of cellulosic material (Veeken and Ham-elers, 1999). Studies on the anaerobic digestion of cellulosic mate-rial have concluded that considerable concentration of cellulolyticcapabilities exists among the order Clostridiales (Lynd et al., 2002).

ll rights reserved.

: +358 14 2602321.

Co-digestion of crops and cow manure for methane productionis an option that has several benefits. The main advantages areimprovement of methane yield, reduction of the emissions ofgreenhouse gas, methane and nitrous oxide, from waste manage-ment, and savings in energy costs (Alatriste-Mondragon et al.,2006). Technical performance during co-digestion of cow manureand wheat straw (Somayaji and Khanna, 1994), and forage beet si-lage (Weiland, 2003) has been investigated in the laboratory, butthe microbial population responsible for the anaerobic digestionof crops and wastes has not been linked to substrate characteristicsor digestion performance.

Here, the bacterial community structures involved in the co-digestion of sugar beet tops, grass silage, and oat straw along withcow manure were investigated by 16S rRNA gene-based terminalrestriction fragment length polymorphism (T-RFLP), denaturinggradient gel electrophoresis (DGGE), and clone library analyses.

2. Methods

2.1. Biomass source

Grass silage (75% timothy Phleum pretense, 25% meadow fescueFestuca pratensis) and straw of oat Avena sativa were from centralFinland, whereas tops of sugar beet Beta vulgaris were from south-ern Finland (Lehtomäki et al., 2007). All three plant materials werechopped to a particle size of approximately 3 cm and immediately

2312 H. Wang et al. / Bioresource Technology 100 (2009) 2311–2315

stored at �20 �C. Cow manure was obtained from central Finlandand stored at 4 �C.

2.2. Reactor operation

Four parallel CSTRs were inoculated with digestate from a mes-ophilic digester in Laukaa, Finland (Lehtomäki et al., 2007). Thereactors were semi-continuously fed with cow manure at an or-ganic loading rate (OLR) of 2 kg VS m�3 d�1 and hydraulic reten-tion time (HRT) of 20 days for 27 days and operated at 35 ± 1 �Cwith a continuous stir of 300 rpm. One reactor was then run foradditional 28 days with manure alone, whereas sugar beet tops,grass silage, and oat straw henceforth were fed each along withmanure in the other three reactors (sugar beet, grass, and strawreactor) by replacing 10% of the feedstock VS with crop whilemaintaining constant OLR and HRT. The ratio of crop in the feed-stock was then gradually increased up to 40%. Finally, the OLRsof the grass and straw reactors were increased first to 3 and then4 kg VS m�3 d�1, respectively (Table 1). The samples for investigat-ing the microbial communities were taken during the period whenreactors were fed with 40% crop at OLR 2 and 3 kg VS m�3 d�1.They were taken soon after the feeding or OLR change (early E,day 9–15), after ca 1.5 HRTs operation (middle M, day 22–29)and after ca 2.5 HRTs operation (late L, day 49–50) and storedimmediately at �80 �C.

2.3. T-RFLP analysis

Genomic DNA was extracted by using a FastPrep� Instrumentand a Fast DNA� SPIN Kit for Soil (Qbiogene, Inc.) according tothe manufacturer’s instructions. Bacterial 16S rRNA genes wereamplified using the primer set 27f-6-carboxyfluorescein (FAM)/1492r (Sait et al., 2003). PCR was performed in a reaction mixturecontaining 100 ng of DNA, 1 � PCR buffer, 200 lM dNTP, 2 U ofDyNAzymeTM II DNA polymerase (Finnzymes), and 0.5 lM of eachprimer and followed the PCR conditions specified by Sait et al.(2003) in MBS PCR system (Thermo Fisher Scientific). The PCRproducts were purified by a GenEluteTM PCR clean-up kit (Sigma)and were quantified by a NanoDrop� ND-1000 spectrophotometer(NanoDrop Technologies, Inc., USA).

Approximate 100 ng DNA was digested with 10 U of MspI for 3 hat 37 �C. Fluorescently labeled terminal restriction fragments (T-RFs) were separated on an ABI Prism� 3100 automated sequencer(Applied Biosystems) using an internal size standard (GeneScanTM

1200 LIZ�, Applied Biosystems). T-RFLP electropherograms wereanalyzed with GeneMapper� software version 3.1 (Applied Biosys-tems). The analysis was performed for each DNA extraction in trip-licate. Profiles from different samples were standardized to theprofile with the smallest total fluorescent peak area intensity andthen a 2% of percentage threshold was applied (Dunbar et al.,2001). The relative abundance of T-RF within a profile was calcu-lated on the basis of the standardized peak area.

Table 1Operational conditions of the reactors.

Duration (day) 0–27 28–55 56–84

Grass reactor Cow manure Grass GrassStraw reactor Cow manure Straw StrawSugar beet reactor Cow manure Sugar beet Sugar beeLoad ratioa (% VS) 0 10 20OLRb (kg VS m�3 d�1) 2 2 2HRTc (day) 20 20 20

a Proportion of crops in the feedstock.b Organic loading rate.c Hydraulic retention time.

2.4. DGGE analysis

Bacterial 16S rRNA genes were amplified by nested PCRs, usingthe primer set 27f-1492r for the first PCR and 533f-907rGC for thesecond one (Bodelier et al., 2005). The second PCR was performedusing a touchdown program specified by Bodelier et al. (2005).

DGGE was performed using the INGENYphorU-2 � 2 system(Ingeny, The Netherlands) with a denaturing gradient ranging from30% to 60% where 100% denaturant contained 7 M urea and 40%formamide. The gels were run at 100 V and 60 �C for 20 h and thenstained for half an hour in Sybr-GoldTM (Invitrogen, USA) (1:10000dilution in TAE buffer) and photographed (Kodak 1D v.3.5.4system).

2.5. Cloning, sequencing and phylogenetic analysis

The gene fragments amplified by the primer set 27f-1492r weresubcloned into the vector pGEM-T Easy according to the manufac-turer’s instructions (Promega). The plasmid was transformed intoEscherichia coli JM109 competent cells (Promega). The cloned frag-ments were amplified from randomly selected clones by primer setT7-SP6 and purified for sequencing with Econuclease I-SAP enzyme(Fermentas) according to the manufacturer’s instructions.Sequencing was performed with the 27f primer on an ABI Prism�

3100 sequencer (Applied Biosystems) using an ABI BigDye� termi-nator v3.1 cycle sequence kit (Applied Biosystems).

The obtained sequences with a range of 800–900 bases werechecked for chimerical artifacts by the CHECK-CHIMERA programand uploaded to the MyRDP space in the Ribosomal Databases Pro-ject II (RDP-II) (Maidak et al., 1997; Cole et al., 2007). Homologysearch of the sequences against the sequences available in RDP II(release 9.58 February 2008) were performed with the SeqMatchprogram in RDP II. Finally, a phylogenetic tree was constructedby the neighbor-joining method and the Jukes-Cantor distancemodel with a PHYLIP 3.65 software package. Bootstrap resamplinganalysis for 100 replicates was performed. Operational TaxonomicUnits (OTUs) were defined for sequences that shared greater than99% similar.

2.6. Nucleotide sequence accession number

The nucleotide sequence reported in this paper was depositedin the NCBI nucleotide sequence databases under Accession Num-bers EU551086 through EU551123.

3. Results and discussion

3.1. The bacterial community structures on the basis of their T-RFLPprofiles and DGGE band patterns

The T-RFLP profiles revealed that there were both clearly dis-tinct and also similar T-RFs in their respective bacterial communi-

85–141 142–203 204–266 267–318

Grass Grass Grass GrassStraw Straw Straw Straw

t Sugar beet Sugar beet30 40 40 402 2 3 420 20 18 16

H. Wang et al. / Bioresource Technology 100 (2009) 2311–2315 2313

ties in the different reactors (Fig. 1). A total of 31 T-RFs were iden-tified in the three reactors. Among them, T-RFs of 157 and 540 bp,which together accounted for 15–31% of the profiles, occurred inall the analyzed samples. In contrast, a T-RF of 90 bp was onlyfound in the sugar beet reactor. T-RFs of 89, 98, 180, and 422 bpwere undetected in this reactor while they were often detectedin the other reactors. The relatively abundant T-RF of 66 bp was ob-served throughout in the straw reactor whereas it was detected inthe sugar beet and grass reactors only after the later operationalperiod at OLR 2 kg VS m�3 d�1.

In addition, the abundance of T-RFs was observed to be relatedto the OLR. The abundance of T-RF 98 bp decreased in grass andstraw reactors while the relative abundance of the T-RFs of 89,180, and 422 bp increased with increasing OLR. However, largerchanges in the abundance of these latter T-RFs were found in thestraw reactor than in the grass reactor. Some T-RFs of 85, 134,162, 176, 186, and 537 bp with a relative abundance of up to 10%of the total profile were present in the reactors only at OLR 2 kgVS m�3 d�1, indicating they were minor groups. Overall, the mostdiverse community was detected during the middle operationalphase of the straw reactor at OLR 2 kg VS m�3 d�1. The reductionin HRT may have resulted in enrichment of the microbial speciesoutcompeting for the essential resources in the reactor, whereasspecies with a less competitive advantage were removed fromthe reactor because of washout.

The DGGE analysis also displayed clearly dissimilar band pat-terns between the reactors and during the change of OLR. Mostbands were recorded in the straw reactor. The DGGE band patternsalso revealed clear changes along with the operation of the grassand straw reactors. The band patterns derived from the grass andstraw reactors after the late operational period at OLR 2 kg VSm�3 d�1 showed great dissimilarity with those from the earlierperiods. In addition, the band patterns between the grass andstraw reactor showed larger difference at OLR 3 kg VS m�3 d�1 than2 kg VS m�3 d�1.

The development of the bacterial community in the reactorsobviously depends on the composition of the feedstock, and onthe process conditions, such as OLR, HRT, temperature, and pH.

Fig. 1. Relative abundance of 16S rRNA gene fragments retrieved from the samples of difchange (day 9–15); M, refer to the middle operational period after ca 1.5 HRT operationoperation (day 49–50) in each change.

When the present reactors were all fed with cow manure theirmethane yields and VS removals were very similar (Lehtomäkiet al., 2007), which suggests good reproducibility of the parallelreactors with identical feedstock and process parameters. In thereactors the similarity was favored by the presence of manure(60% of feed VS) in all three reactor feedstock used as co-sub-strates. However, differences in population were obviously pro-moted by differences in the composition of these three crops.High lignin content was found in grass and straw (19 and 21% ofTS, respectively), whereas in sugar beet tops lignin content waslower (10% of TS) (Lehtomaki and Björnsson, 2006).

3.2. Bacterial clone libraries

Three clone libraries were constructed from each reactor sam-ple: (i) B and (ii) S library represented the sugar beet reactor andstraw reactor at M with the OLR of 2 kg VS m�3 d�1, respectively,and (iii) G library represented grass reactor at M with the OLR of3 kg VS m�3 d�1. In total, 11 OTUs were identified for B library(B1–B11), 19 for G library (G1–G19), and 12 for S library (S1–S12). In the G library, the most of OTUs were found to be closelyrelated to species within class Clostridia and Bacteroidetes, whereas3 OTUs were not affiliated with a specific phylogenetic group andassigned to unclassified Bacteria. The OTUs in B library affiliatedto the class Clostridia, Bacteroidetes, Bacilli, and unclassified Bacte-ria. For S library, the OTUs related to the class Clostridia, Bacteroide-tes, Acidobacteria, Deltaproteobacteria, and unclassified Bacteria.Apparently, the S library exhibited most phylogenetic diversity atclass level, which is in good agreement with the observation de-rived from T-RFLP and DGGE data. The dominant component in Blibrary was unclassified Bacteria, accounting for 47.4% of the total,whereas OTUs affiliated with the class Clostridia dominated in bothG (50% of the clones) and S (53.3% of the clones) library. Theunclassified Bacteria were found to be prevalent, accounting for20% of clones, in both S and G library. In addition, OUTs closely re-lated to class Bacteroidetes were much more in G library than inother libraries. A recent study (Cirne et al., 2007) also showed veryfew similarities between the microbial populations anaerobic

ferent reactors. E, refer to the early operational period soon after the feeding or OLR(day 22–29) in each change; L, refer to the late operational period after ca 2.5 HRTs

Fig. 2. Phylogenetic affiliation of 16S rRNA gene sequences. Bootstrap values from 100 replicates are shown for each node. The scale bar represents a 10% estimated differencein nucleotide sequence. Clones were designed B (sugar beet reactor), S (straw reactor), and G (grass reactor). Numbers in parentheses represent the expected T-RF lengths ofthe dominant clones. The tree is rooted using Escherichia coli as outgroup.

2314 H. Wang et al. / Bioresource Technology 100 (2009) 2311–2315

digestion of sugar beets and grass, and the different dominant bac-terial group in the two communities.

Sequence analyses showed similar phylotypes in the bacterialcommunities between the three reactors as shown in Fig. 2. G1(S1, B1) was commonly present in the three library, which wasthe dominant clone in the B library representing 42% of the clones.The OTU S3 (B3) and S2 (B2) were found in both S and B library. Incontrast, distinct phylotypes in the bacterial communities betweenthe three reactors were also observed. An OTU, S11, affiliated withclass Deltaproteobacteria was detected only in the S library. In addi-tion, a sequence related to class Acidobacteria was also only identi-fied in S library. A sequence B11 clustered into class Bacilli was onlyfound in the B library. Moreover, there were differences in specificmethane yields and VS removals among the reactors. The highestspecific methane yields and VS removals were obtained in thegrass reactor, while the lowest value was found in the straw reac-tor (Lehtomäki et al., 2007).

4. Conclusion

The study showed that both similar phylotypes and distinctones in the bacterial communities between the three reactors co-digestion of different crops along with cow manure. The results

suggested that the development of bacterial community in thesereactors was different.

Acknowledgements

We thank Dr. Anna Kaksonen for kindly arranging the DGGEexperiments. We also thank Leena Siitonen, Eila Korhonen and Eli-na Virtanen for their dedicated technical assistance. This work wassupported by Finnish Maj and Tor Nessling Foundation and EU 6thFramework Programme (Project SES6-CT-2004-502824).

References

Alatriste-Mondragon, F., Samar, P., Cox, H.H., Ahring, B.K., Iranpour, R., 2006.Anaerobic codigestion of municipal, farm, and industrial organic wastes: asurvey of recent literature. Water Environ. Res. 78, 607–636.

Amon, T., Amon, B., Kryvoruchko, V., 2007. Methane production through anaerobicdigestion of various energy crops grown in sustainable crop rotations.Bioresour. Technol. 98, 3204–3212.

Bodelier, P.L., Meima-Franke, M., Zwart, G., Laanbroek, H.J., 2005. New DGGEstrategies for the analyses of methanotrophic microbial communities usingdifferent combinations of existing 16S rRNA-based primers. FEMS Microbiol.Ecol. 52, 163–174.

Chynoweth, D.P., Pullammanappallil, P., 1996. Anaerobic digestion of municipalsolid wastes. In: Palmisano, A.C., Barlaz, M.A. (Eds.), Microbiology of SolidWaste. CRC Press, Boca Raton, Florida, pp. 71–113.

H. Wang et al. / Bioresource Technology 100 (2009) 2311–2315 2315

Cirne, D.G., Lehtomaki, A., Bjornsson, L., Blackall, L.L., 2007. Hydrolysis and microbialcommunity analyses in two-stage anaerobic digestion of energy crops. J. Appl.Microbiol. 103, 516–527.

Cole, J.R., Chai, B., Farris, R.J., Wang, Q., Kulam-Syed-Mohideen, A.S., McGarrell, D.M.,Bandela, A.M., Cardenas, E., Garrity, G.M., Tiedje, J.M., 2007. The ribosomaldatabase project (RDP-II): introducing MyRDP space and quality controlledpublic data. Nucleic Acids Res. 35, 169–172.

Dunbar, J., Ticknor, L.O., Kuske, C.R., 2001. Phylogenetic specificity andreproducibility and new method for analysis of terminal restriction fragmentprofiles of 16S rRNA genes from bacterial communities. Appl. Environ.Microbiol. 67, 190–197.

Lehtomaki, A., Björnsson, L., 2006. Two-stage anaerobic digestion of energy crops:methane production, nitrogen mineralisation and heavy metal mobilisation.Environ. Technol. 27, 209–218.

Lehtomäki, A., Huttunen, S., Rintala, J., 2007. Laboratory investigations on co-digestion of energy crops, crop residues and cow manure: effect of crop tomanure ratio. Resour. Conserv. Recycl. 51, 591–609.

Lynd, L.R., Weimer, P.J., van Zyl, W.H., Pretorius, I.S., 2002. Microbial celluloseutilization: fundamentals and biotechnology. Microbiol. Mol. Biol. Rev. 66, 506–577.

Maidak, B.L., Olsen, G.J., Larsen, N., Overbeek, R., McCaughey, M.J., Woese, C.R., 1997.The RDP (Ribosomal Database Project). Nucleic Acids Res. 25, 109–111.

Sait, L., Galic, M., Strugnell, R.A., Janssen, P.H., 2003. Secretory antibodies do notaffect the composition of the bacterial microbiota in the terminal ileum of 10-week-old mice. Appl. Environ. Microbiol. 69, 2100–2109.

Somayaji, D., Khanna, S., 1994. Biomethanation of rice and wheat straw. W. J.Microbiol. Biotechnol. 10, 521–523.

Veeken, A., Hamelers, B., 1999. Effect of temperature on hydrolysis rates of selectedbiowaste components. Bioresour. Technol. 69, 249–254.

Weiland, P., 2003. Production and energetic use of biogas from energy crops andwastes in Germany. Appl. Biochem. Biotechnol. 109, 263–274.

Wilson, J.R., Mertens, D.R., 1995. Cell wall accessibility and cell structure limitationsto microbial digestion of forage. Crop Sci. 35, 251–259.