A re-appraisal of the diversity of the methanogens associated with the rumen ciliates

Authors


*Corresponding author. Fax: +44 1224 716687, E-mail address: nrm@rri.sari.ac.uk

Abstract

The diversity of methanogenic archaea associated with different species of ciliated protozoa in the rumen was analysed. Partial fragments of archaeal SSU rRNA genes were amplified from DNA isolated from single cells from the rumen protozoal species Metadinium medium, Entodinium furca, Ophryoscolex caudatus and Diplodinium dentatum. Sequence analysis of these fragments indicated that although all of the new isolates clustered with sequences previously described for methanogens, there was a difference in the relative distribution of sequences detected here as compared to that of previous work. In addition, many of the novel sequences, although clearly of archaeal origin have relatively low identity to the sequences in database which are most closely related to them.

1Introduction

The rumen is a complex ecosystem comprising an as yet undetermined number of protozoal, fungal, archaeal and eubacterial species that are able to breakdown plant material. These different microbes form a complex community of organisms that interact with one another and play an important role in the digestion of feed and the supply of energy to the host in the form of volatile fatty acids and microbial protein. In keeping with many other digestive systems, the species present, and their relative abundance, can change significantly with time. Recently there has been considerable interest in determining the true biodiversity of each of these types of organisms within this ecosystem, in particular the eubacteria [1–5] and protozoa [6].

The diversity of the methanogenic archaea has taken an increasing amount of interest more recently due to the production of methane by domesticated ruminants. Particularly in countries where a significant proportion of a nation's industry is agriculture based, this can account for a large proportion of the national greenhouse gas emissions - e.g., in New Zealand, enteric emissions are responsible for approximately 60% of that country's total greenhouse gas emissions [7]. As with the other microbial populations in the rumen, the methanogen population can be changed according to dietary regime, e.g., [2,3,5,8–10].

In addition to direct dietary changes, the methanogenic population of the rumen may be affected by the presence or absence of ciliated protozoa, as it is known that many methanogens exist in close association with the rumen ciliates. In general most work on rumen methanogen diversity has concentrated on the rumen as a complete organ, e.g., [9,11] and only two previous investigations have looked at the methanogens associated with ciliates in the rumen. These two investigations made use of different isolation methods; ciliate enrichment by sedimentation [12] or centrifugation following transit from one country to another [13]. Both of these methods go some way towards enriching the biomass towards a greater abundance of protozoa relative to other organisms. However, both methods have the potential to still harbour significant amounts of loosely adhered organisms which are not truly associated with the ciliate, and the latter has the additional risk of some organisms being lost from the study due to factors such as lysis in transit.

A third and alternative method for studying methanogens associated with rumen ciliates is presented here, making use of PCR on single cells which have been hand-picked and cleaned up using a Chelex system.

2Material and methods

2.1Harvesting of ciliates

Rumen ciliates were isolated either directly from the rumen or from cultures which had been maintained in vitro. Details of the origins of the samples are described in Table 1. Individual cells were hand-picked using a micromanipulator and a microscope from either rumen fluid or an in vitro culture. Single cells were then washed three times with sterile water and added to 50 μl of 5% Chelex-100 with 10 μg ml−1 proteinase K, and incubated at 98 °C for 5 min before being placed on ice and centrifuged at 3000g for 10 min. Sedimented DNA was used directly for PCR [14].

Table 1.  Sequences of methanogens in GenBank
CloneEBI No.Host ciliateSource of hostBest hit% Age identityBest hit from cultured organism - where best hit is from non-cultured source% Age identity
  1. Archaea isolates associated with ciliata from rumen of cow, sheep and goat. Best hits from BLASTN searches are shown. It should be noted that in some cases the best hit and the best cultured hit have the same percentage identity (after rounding up or down to the nearest 1%) to the test sequence. In all cases, the best value is shown first.

Oph-12AJ606400Ophryoscolex caudatusSheep's rumenMethanomicrobium mobile99--
Met-01AJ606401Metadinium mediumCow's rumenMethanomicrobium mobile89--
Oph-25AJ606402Ophryoscolex caudatusGoat's rumenMethanomicrobium mobile96--
Oph-22AJ606403Ophryoscolex caudatusGoat's rumenP. nitobei symbiont88Methanimicrococcus blatticola87
Oph-24AJ606404Ophryoscolex caudatusGoat's rumenO. formosanus symbiont90Methanimicrococcus blatticola90
Met-05AJ606405Metadinium mediumCow's rumenMethanomicrobium mobile98--
Met-12AJ606406Metadinium mediumCow's rumenMethanomicrobium mobile99--
Met-04AJ606407Metadinium mediumCow's rumenMethanomicrobium mobile97--
Met-19AJ606408Metadinium mediumCow's rumenMethanomicrobium mobile99--
Met-11AJ606409Metadinium mediumCow's rumenRumen isolate [AB034182]98Methanomicrobium mobile98
Met-14AJ606410Metadinium mediumCow's rumenMethanomicrobium mobile98--
Oph-05AJ606411Metadinium mediumCow's rumenMethanomicrobium mobile99--
Ent-15AJ606412Ent. furca monolobumSheep - in vitroUncultured archaeon99Methanomicrobium mobile99
Ent-01AJ606413Ent. furca monolobumSheep - in vitroUncultured archaeon95Methanomicrobium mobile95
Ent-25AJ606414Ent. furca monolobumSheep - in vitroUncultured archaeon99Methanomicrobium mobile99
Ent-11AJ606415Ent. furca monolobumSheep - in vitroMethanomicrobium mobile99--
Ent-14AJ606416Ent. furca monolobumSheep - in vitroMethanomicrobium mobile99--
Ent-18AJ606417Ent. furca monolobumSheep - in vitroMethanomicrobium mobile99--
Dip-23AJ606418Diplodinium dentatumSheep - in vitroUncultured archaeon97Methanomicrobium mobile97
Dip-16AJ606419Diplodinium dentatumSheep - in vitroMethanimicrococcus sp.87--

2.2PCR analysis

Primers already known to amplify around 1200 bp of archaeal the SSU rRNA gene ArcF 7 (GTTGATCCTGCCAGAGG), ArcR 1326 (TGTGTGCAAGGAGCAGGGAC) [15] were used to amplify fragments of the SSU rRNA gene by PCR. PCR was performed in a 50 μl volume using around 250 ng of template DNA, 0.04 mM of each dNTP, 20 pmol of each primers, and 0.5 Units Taq DNA polymerase (Invitrogen), following the manufacturer's instructions. PCR was performed using the following steps: 1 cycle (94 °C for 4 min, 56 °C for 45 s, 72 °C for 1 min 20 s), 35 cycles (94 °C for 30 s, 56 °C for 45 s, 72 °C for 1 min 20 s), and 1 cycle (94 °C for 30 s, 56 °C for 45 s, 72 °C for 10 min). Successful amplification of DNA was demonstrated by electrophoresis on a 1.5% agarose gel in TAE buffer [16] and visualised following staining with 0.5 μg ml−1 ethidium bromide.

2.3Verification of amplicons

PCR products were ligated into the vector pKRX and cloned into Escherichia coli (strain DH5α MCR). Plasmids were isolated from recombinant colonies and checked for inserts. Plasmids with inserts were sequenced in both directions on a Beckmann DNA Sequencer. Inserts were verified as being archaeal SSU genes by DNA sequence analysis using BLASTN. In addition, all sequences were checked for potential chimeric origin at the Ribosomal Database Project (http://rdp.cme.msu.edu/html/).

2.4Phylogenetical analysis

The 10 best “hits” for each of the new sequences, and previously described rumen methanogen SSU sequences were downloaded and saved in FASTA format for alignment. Sequences were aligned using ClustalW [17] with a PHYLIP output, and included Methanococcus vannielii as an out group, having previously been shown to be a useful outlier sequence for the rumen methanogens [9]. The alignment was analysed using the PHYLIP suite of programmes [18]. In the first instance a tree was constructed to depict the genetic distance between DNA sequences, based on the differences between them, using DNAdist, followed by Neighbor. Bootstrap analysis with 1000 iterations was used to assess the statistical strength of the branch positions (1000 re-sampled trees were generated using Seqboot, prior to DNAdist analysis, and the output trees from Neighbor collated using the Consense program). Bootstrap confidence values of 50% or greater are labelled on the respective branch nodes of the tree shown in Fig. 1(a).

Figure 1.

Figure 1.

(a)The relationship between 16S sequences derived for methanogens from the rumen and their “best hits” identified by BLASTN analysis. The tree is the product of NJ analysis and bootstrap values shown are the percentage of occurrences based on 1000 bootstrap iterations. The new sequences reported here are marked with three asterisks [***], those from Tokaru et al. 2001 are shown with a single asterisk [*] and those from Chagan et al. 1999 are shown with two asterisks [**]. (b) A more detailed view of the area of the tree occupied by 16 of the newly identified sequences from methanogens associated with rumen ciliates.

Figure 1.

Figure 1.

(a)The relationship between 16S sequences derived for methanogens from the rumen and their “best hits” identified by BLASTN analysis. The tree is the product of NJ analysis and bootstrap values shown are the percentage of occurrences based on 1000 bootstrap iterations. The new sequences reported here are marked with three asterisks [***], those from Tokaru et al. 2001 are shown with a single asterisk [*] and those from Chagan et al. 1999 are shown with two asterisks [**]. (b) A more detailed view of the area of the tree occupied by 16 of the newly identified sequences from methanogens associated with rumen ciliates.

Figure 1.

Figure 1.

(a)The relationship between 16S sequences derived for methanogens from the rumen and their “best hits” identified by BLASTN analysis. The tree is the product of NJ analysis and bootstrap values shown are the percentage of occurrences based on 1000 bootstrap iterations. The new sequences reported here are marked with three asterisks [***], those from Tokaru et al. 2001 are shown with a single asterisk [*] and those from Chagan et al. 1999 are shown with two asterisks [**]. (b) A more detailed view of the area of the tree occupied by 16 of the newly identified sequences from methanogens associated with rumen ciliates.

3Results

Twenty novel rumen methanogen SSU sequences were obtained. These sequences have been deposited in the EBI database. Of these sequences a number, although having greatest similarity to other sequences from archaea, showed relatively poor similarity levels to any sequences described in GenBank, with some “best identity” levels being less than 90% (summarised in Table 1). Clearly from this information there are a number of sequences which are unlikely to closely align with any known sequences.

The new sequences reported here are marked with three asterisks (***), the previous methanogen sequences from organisms associated with rumen ciliates are shown with one asterisk [12] or two asterisks [13]. The phylogenetic tree depicted in Fig. 1(a) shows that the four sequences which show 90% or less similarity to known sequences cluster together - possibly as a result of “long branch” clustering. However, it is interesting to note that none of the new sequences clusters with any of the previous sequences described as coming from methanogens associated with rumen ciliates. Also a similar pattern of the novel sequences generally clustering away from the majority of the previously described rumen methanogens was also observed when analysis was performed using parsimony analysis (DNAPars program within the PHYLIP suite). The area of the tree containing the sixteen new sequences which cluster together is shown in greater detail in Fig. 1(b). Within the branch of this tree there are other sequences which have been isolated from rumen samples - AB034182 and X99139 - although these sequences were determined from samples which had not specifically been isolated from rumen ciliates.

4Discussion

Clearly all twenty sequences which have been newly determined are very different from those previously described as being associated with the rumen ciliates. Four of these have very low identity levels with any previously described sequences, not only those from methanogens associated with rumen ciliates, and cluster together on the tree - potentially as a result of long-branch clustering. In order that these sequences be properly assigned to a region of the tree it will be necessary to identify additional sequences which are sufficiently similar to allow proper alignment of these sequences within the context of this tree. Nevertheless, it is obvious that these sequences are significantly different from the sequences which had previously been shown to associate with the rumen ciliates.

However, of greater interest are the additional sixteen sequences which show greater similarity to sequences in the database. They lie as a cluster together with two other sequences which have previously been described for rumen methanogens. However, they do not cluster with the sequences which had previously been described for methanogens associated with rumen ciliates. This indicates that the Chelex method used here detects a population of ciliate-associated methanogens which were not detected by the previous studies - although detected by work performed on total rumen fluid.

Also of interest is the observation that these sixteen new sequences show no obvious bias for a particular methanogen population being associated with any specific species of ciliate. Equally there is no obvious division based on original host ruminant, although such a trend may need a greater dataset to allow any such bias to be observed.

More importantly, this tree demonstrates a difference in positioning associated with the sequences here relative to those previously described as associating with the rumen ciliates. This suggests that the method used here to analyse the methanogens associated with the rumen ciliates is able to detect a different population from previous studies, although no such distinction is obvious for the other two methods, with sequences derived by both previous techniques overlapping on the tree (Fig. 1(a)). Clearly all the sequences described by all three methods are from methanogens and due to the isolation procedure must have originated from the rumen. The fact that the two populations of methanogens isolated from ciliates lie as distinct clusters demonstrates that the results obtained are heavily biased by the method of isolation. This in turn demonstrates that the original biodiversity of methanogens, and possibly other organisms in the rumen, may be underestimated at present due to isolation methods needing to be refined.

Acknowledgement

This work was in part supported by EU infrastructure Grant QLRI-CT-2000-01455 (ERCULE). Martina Regensbogenova was the recipient of a Training Studentship under the Rowett Research Institute's Marie Curie Training award (HPMT-CT-2001-00409). The Rowett Research Institute receives funding from SEERAD.

Ancillary