Even if it were possible to construct a perfect tree, local topologies could be expected to change as new data accumulated and was used to reconstruct the trees. In other words, phylogenetic trees are dynamic constructs that may change when new sequence information is included. New data confer additional information, new identities, and new differences resulting in branch attraction or repulsion. Formerly stable clusters may be dissected, while separated groups may be unified. The trees are "growing" and changing - very much like real ones. In general, for the definition, verification, and positioning of new phyla, one needs data collections containing a reasonable number of sequences representing different levels of relatedness.
As more 16S rRNA sequences accumulate from both cultured and uncultured archaea, the boundaries of existing phyla are being challenged and need to be revised. In order to evaluate the current (May 2002) state of archaeal phylogeny, the ARB software package was used. An ARB database consisting of more than 2500 complete and partial archaeal 16S rRNA sequences was built. Most of the sequences were imported from the EMBL database and about 80% of sequences were published in scientific journals (Table 2). Sequences were aligned and manually corrected taking into account the secondary structure of the rRNA molecule. This database is available upon request at German Jurgens' Website (ARB page http://honeybee.helsinki.fi/users/gjurgens/Arb/arb_page.htm).
Many entries retrieved from public sequence databases contain only partial sequences and some of them do not overlap. To avoid reconstruction artifacts arising from partial sequences, three separate “backbone” maximum-likelihood trees were constructed (Fig. 5): for Korarchaeota (A) (with novel clones, phylogenetically separate from all three phyla), Crenarchaeota (B), and Euryarchaeota (C). The partial sequences were then added, using the unique ARB program "parsimony" option, to backbone trees without changing the overall tree topology. This method allows the construction of big trees with sequences that do not overlap. The constructed trees give an overview of the "initial grouping" and relationship of sequences. These trees help in choosing groups of sequences with correct references which allow further work on more detailed and accurate tree reconstruction. Major lineages (phyla) on Crenarchaeota and Euryarchaeota trees are shown as wedges with horizontal dimensions reflecting the known degree of divergence within that lineage. Phyla with cultivated representatives are in grey and, where possible, named according to the taxonomic outline of Bergey’s Manual (Garrity, 2001). Phyla known only from environmental sequences are in white; because they are not formally recognized as taxonomic groups. They are usually named after the first clones found from within the group (Hugenholtz, 2002). Another interesting and important utilization of the constructed ARB database and trees are the “Probe design" and "Probe match” features of the ARB phylogenetic package. These features allow rapid searching of closest relatives or specific sequence signatures. Such signatures can be evaluated for taxon specificity against the full database and used as probes for hybridization and/or as PCR primers. This handy feature is unique to the ARB software package and allows quick design and evaluation of new and available probes and primers.
Soils are dynamic, complex systems of inorganic, organic and biotic components that have the capacity to support plant life. Both culture-based and culture-independent approaches support the statement that soil represents one of the most diverse habitats for microorganisms (Hugenholtz et al., 1998). Pioneering studies by Torsvik and co-workers examined the diversity of natural communities in soil by DNA-DNA reannealing experiments. In these analyses, bulk DNA was isolated from a soil bacterial fraction obtained by differential centrifugation (Torsvik et al., 1990; Torsvik et al., 1996). Reannealing measurements revealed that the DNA isolated directly from soil was much more complex than expected and suggested that thousands of independent genomes were present in the sample. Extrapolation of the data suggests that there may be thousands of microbial types in a gram of soil, many of which are assumed to be uncultured.
Research on the diversity of "non-extreme" Archaea in soil (apart from euryarchaeotal methanogens) is a rather new sector in Archaea discipline and started with the discovery of the novel Crenarchaeota in Finnish forest soil (paper I). This was the first soil crenarchaeotal 16S rRNA group of sequences ever submitted to a public data bank (EMBL database AC X96688-X96696, submitted 20 March 1996). One short 287 nucleotide long sequence named "FIE16" was obtained by Ueda et al., 1995. The following submission of sequences and research papers (Bintrim et al., 1997; Kudo et al., 1997; Borneman and Triplett, 1997; Buckley et al., 1998; Bomberg et al., 2002) further demonstrated the diversity of Archaea in different types of soils and emphasized our lack of knowledge about them.
In paper I nine archaeal 16S rRNA gene sequences (named FFSB) were obtained by PCR amplification with Archaea-specific primers and template DNA extracted directly from a forest soil sample. The three used methods of phylogenetic analysis used undoubtedly placed eight of the obtained sequences in a group which was distantly related to the all other known crenarchaeotal sequences. This group was situated on the same lineage as the planktonic clade (low-temperature marine Archaea Group I (Delong, 1992)) together with other environmental sequences between the latter and the rest of the Crenarchaeota (Fig.1, paper I). One of the clones, FFSB6, was placed on an individual branch.
FFSB clones from paper I were divided into two clusters according to their hybridization with oligonucleotide probes FFS-Uni and FFS-6. These two probes were designed based on sequence data, as well as phylogenetic and secondary structure analysis of clear-cut followed by prescribed burning soil clones. Probe FFS-6 was designed to be specific to the phylogenetically distinct clone FFSB6.
The first cluster, including clones FFSB1 to 5, 7, 10, and 11 gave positive hybridization signals only with the FFS-Uni probe. The second cluster, consisting only of the clone FFSB6, gave positive signals only with the FFS-6 probe (probe and hybridization conditions are described in paper II).
In paper II the genetic diversity of Archaea in forest soil treated with clear-cutting and prescribed burning was described. The diversity of soil Archaea in intact (C), clear-cut (A) and clear-cut followed by prescribed burning (B) boreal forests was examined based on 19, 49 and 70 clones, respectively.
All obtained clones were subjected to RFLP analysis with AvaII, MspI and MspI/RsaI restriction enzymes, followed by Southern blotting and hybridization with the probes FFS-Uni and FFS-6.
70 clones from clear-cut followed by prescribed burning soil (type B) were divided into 9 classes according to restriction and hybridization patterns. The results shown include representatives of each of these classes (Fig. 1, paper II).
Altogether, 69 clones from clear-cut followed by prescribed burning soil (type B) hybridized only with the FFS-Uni probe and belong to the FFS-Uni cluster. Only one clone (FFSB6) gave a signal with the FFS-6 probe and thus belonged to the FFS-6 cluster.
49 clones from clear-cut soil (type A) were analyzed and fell into seven classes. Only two clones, FFSA8 and FFSA12 (Fig. 3, paper II), gave new restriction and hybridization patterns, not found among the FFSB clones. Nevertheless, they gave positive hybridization signals with the FFS-Uni probe and therefore were classified within the FFS-Uni cluster. 4 clones from clear-cut soil fell into the FFS-6 cluster (Table 1, paper II).
4 new classes among the 19 analyzed clones from intact soil (type C) were found. Their restriction and hybridization patterns differed from those isolated from clear-cut soil and clear-cut followed by prescribed burning soil. Only 2 clones showed a pattern identical to the existing FFSB classes.Clones from the 4 new classes showed homology with the FFS-Uni probe (FFS-Uni cluster).
Restriction patterns belonging to clones FFSB2 and FFSB3 were present in clone libraries from all three soils (A, B, and C).Clones from clear-cut soil and clear-cut followed by prescribed burning soil shared six common patterns, whereas the majority of clones from intact soil had patterns not found in either clear-cut soil and clear-cut followed by prescribed burning soil clones.
Phylogenetic analysis of six sequenced clones from clear-cut and intact soils (Fig 4., maximum likelihood tree, paper II) supported the paper I results. New sequences were affiliated with both the FFS-Uni and FFS-6 subgroups of soil Crenarchaeota. Results based on DNA hybridization and sequence analysis correlated with each other. Sequences FFSC1, 2, 3, FFSC4, from the FFS-Uni hybridization cluster (Table 1, paper II) were all placed within the FFSB group of sequences. Sequences FFSA1 and FFSA2 from the FFS-6 cluster (Fig.4, paper II) were affiliated with the sequence FFSB6. Thus, by using the RFLP-hybridization technique, clones belonging to different phylogenetic groups in the rDNA clone library were distinguished.
The FFS sequences were different (100% bootstrap value in maximum parsimony analysis) from two other soil Archaea groups of sequences isolated from Wisconsin (Bintrim et al., 1997) and Amazon soils (Borneman and Triplett, 1997). These two groups were similar to each other and were situated more closely to the planktonic clade of Crenarchaeota.
Currently all FFS sequences are placed in GroupI.1C on the latest Crenarchaeota tree (Fig. 5B). This group includes altogether 46 sequences isolated from terrestrial environments.
Estuaries are the interfaces between freshwater and marine environments. Tidal estuaries are extremely dynamic zones, with spatial and temporal gradients, due to the variability of several factors such as freshwater input, geomorphology, winds, tidal heights, as well as anthropogenic inputs. As expected, microorganisms play an important role in the dynamics of estuarine environments, particularly in biogeochemical cycles and food webs (Nickson, 1995). Therefore the biodiversity of estuarine microbial communities is an area of great interest in microbiological and ecological studies. Although reports on Archaea in estuarine environments are available (Munson et al., 1997; Crump and Baross, 2000), none have revealed the presence of Crenarchaeota organisms in brackish sediments.
In order to verify the presence and diversity of Archaea in the estuarine sediments of the river Douro (Portugal), total sediment DNA was isolated and archaeal 16S rDNA was selectively amplified by PCR, cloned, sequenced and phylogenetically characterized.
The 235 recombinant plasmids were screened with AvaII restriction enzyme and 14 were selected according to their original restriction pattern and sequenced (DOURO 1 to 14). Phylogenetic analysis of the sequences revealed the presence of Crenarchaeota 16S rRNA gene sequences in the sediment samples. These sequences were recovered mainly from the surface layer of the cores. However, sequences DOURO10 and 11 were retrieved from a depth of 10 cm, corresponding to a region of thin grain and compact sediment, while DOURO 12 and 13 were found at a depth of 19 cm from the same core. Comparison of 12 of the DOURO sequences (1-6, 8, 9, and 11-14) with existing sequences in public databases (EMBL and GenBank), revealed 95-98% identity to uncultured environmental archaeal sequences retrieved from marine ecosystems in other studies (Delong, 1992; Delong et al., 1994; Fuhrman et al., 1993). Indeed, thephylogenetic analysis placed them within the so-called “marine cluster”(Buckley et al., 1998), which is made up of archaeal 16S rDNA sequences recovered from marine environments and lake sediments (Fig. 1, paper III). In the current Crenarchaeota tree, this group is named GroupI.1a (Fig. 5B).
The DOURO7 sequence presented 88-90% identity with Euryarchaeota 16S rRNA sequences from deep-sea sediments (Vetriani et al., 1999) and 82% with Methanobacterium formicicum, and was placed in the euryarchaeotal group named pMC1 (Hugenholtz, 2002, Fig. 5C). DOURO10 showed 89% identity with Halobacterium sp. and was placed in group VAL2 (Fig. 5C).
Freshwaters and wetlands in boreal environments have been identified as important sources of trace gases (Khalil, 1993) and can contribute up to 34% of the total CH4 fluxes of wetlands globally. Prokaryotes in soil, sediments and water play an essential role in this scenario by their contribution to CO2 and CH4 cycling and fluxes (Conrad, 1996). In addition to their global role in biogeochemical cycles, prokaryotes play a key role in the transformation of natural organic matter, which in many boreal freshwaters is of primarily allochthonous origin and recalcitrant (Münster et al., 1999).
A total of 28 clones (named VAL) recovered from lake Valkea Kotinen were sequenced. We found that (1) VAL sequences contained representatives of both Cren- and Euryarchaeota, (2) no cultivated Archaeawere closely related to the VAL sequences, (3) the VAL clones were most closely related (with a similarity around 96%) to archaeal sequences inhabiting a very wide range of environments - marine and lake sediments, different types of soils, rice roots, and an anaerobic digestor. Two separate trees, one for Crenarchaea and one for Euryarchaea, were generated (Fig. 1 and 2, paper IV).
12 freshwater VAL-sequences fall within a subcluster named “freshwater cluster” (Buckley et al., 1998), Group I.3 (Delong, 1998) or Group I.3a (Fig. 5B) of Group I uncultured non-thermophilic Crenarchaeota. Eight of the sequences form a tight VAL I lineage inside the cluster which also includes sequences previously detected in lake sediments (Hershberger et al., 1996; Schleper et al., 1997a), rice roots and soil (Grosskopf et al., 1998b), and an anaerobic digestor (Godon et al., 1997) (Fig. 1, paper IV).
Analysis of 16 VAL sequences identified four distinct lineages related to different orders of Euryarchaeota (Fig. 2, paper IV). Four clones were affiliated with environmental sequences previously detected in anoxic sediments from Rotsee (Switzerland) (Zepp et al., 1999), post-glacial profundal freshwater sediment (Miskin et al., 1998) and peat bogs (Hales et al., 1996) within the order Methanomicrobiales.Another four clones formed a group among sequences inhabiting marine sediments (Vetriani et al., 1998; Munson et al., 1997), rice roots and soil (Grosskopf et al., 1998b) and the deep-sea (Fuhrman and Davis, 1997) (Group III, Figure 2, paper IV). Clone VAL147 clustered together with clone Eel-TA1c9 which was retrieved from a marine sediment (Hinrichs et al., 1999). Convincing evidence (albeit circumstantial) has suggested that organisms represented by this clone are methanotrophic anaerobes involved in the decomposition of methane hydrate. The remaining 7 clones formed two distinct lines of descent which all three methods of phylogenetic analysis (maximum likelihood, maximum parsimony and neighbor-joining) placed in the kingdom Euryarchaeota. However, the exact placement of these two clusters within this kingdom could vary due to the large phylogenetic distances between members of these clades and the rest of the sequences analyzed. Clade VAL II, consisting of four clones, was only distantly related to the family Methanosaeta (Fig. 2, paper IV). At the time of writing paper IV, all methods of analysis allowed sequence VAL33-1 to place this clade on the same branch as Rice cluster I (Fig. 2, paper IV) on the basis of 85.9% similarity between sequence and cluster (Grosskopf et al., 1998b). Levels of rDNA similarity between the other three VAL II clade sequences and sequences from the closest Rice cluster I were extremely low - 65.8-69.2 %.
A similar situation was observed for three clones from cluster VAL III, which were related to the order Thermoplasmatales, which include the Groups II and III (Delong, 1998) and Rice cluster V (Fig. 2, paper IV). The level of similarity between members of this clade, compared to members of orders Thermoplasmatales and Methanobacteriales, ranged from 71.3% (Thermoplasma acidophilum, Thermoplasmatales) to 76.3% (Methanothermus fervidus, Methanobacteriales). Rice cluster V, which was closest to the three VAL III clade sequences, is believed to comprise of non-methanogenic anaerobic Archaea (Grosskopf et al., 1998b). At present, these three clones are placed in group pMC2 (Hugenholtz, 2002) on latest Euryarchaeota tree (Fig. 5C).
In paper IV, new archaeal oligonucleotides were designed for use as PCR primers and fluorescent in situ hybridization probes (Table 3). This allowed study of the distribution of the indigenous Archaea in the water column and their potential contribution to the microbial food web.
With general Archaea-specific probe ARCH915 (Stahl and Amann, 1991) it was possible to detect 7±2% of the DAPI stained cells (data for the hypolimnium layer). No significant differences were detected between samples taken at different times during the sampling period from spring to autumn.
All positive hybridization signals and cell morphologies visualized with the four Euryarchaeotal probes EURY498, EURY514, EURY496 and EURY499 were identical to the signals obtained with the general Archaea probe ARCH915 (Fig. 3, paper IV). EURY496 was the most specific of all the Euryarchaeota probes tested and covered the order Methanomicrobiales and related environmental sequences. It was therefore assumed that the Euryarchaeota clones detected in the FISH experiments belonged to the order Methanomicrobiales. The size and morphotypes of the cells detected (Fig. 3, paper IV) showed similarities to some methanogens (for example Methanospirillum hungatei). Despite a high background fluorescence, maximum intensity and highest cell counts with the EURY496 probe (7±2% of the DAPI stained cells) were observed in the hypolimnion samples. Cell counts in the epilimnion samples were below the detection limit (1%).
Although positive hybridization signals were obtained with the Euryarchaeotaprobes, none of the three Crenarchaeota probes used (CREN499, CREN512 and CREN569) gave a positive hybridization signal.
Analysis of the diagnostic signatures and features of the cloned sequences (Winker and Woese, 1991; Woese, 1987) revealed that they all belonged to the archaeal domain.
soil sequences: 66 out of the 68 relevant positions that define Archaea from Bacteria and Eucarya (Woese, 1987) show features common to Archaea. Intradomain nucleotide signature comparison (Winker and Woese, 1991) of the FFSB group (shown in Table 1 of paper I) and other soil sequences supported their phylogenetic within the Crenarchaeota
estuarine sediment sequences: the intradomain nucleotide signature analysis of the DOURO sequences (Table 1, paper III) confirmed the affiliation of DOURO7 with the Euryarchaeota, as well as the phylogenetic placement of the DOURO1-6, 8, 9, 11-14 sequences within the Crenarchaeota kingdom. Sequence DOURO10 had both crenarchaeal and euryarchaeal features. The phylogenetic analysis assigned it to the Euryarchaeota, while the intradomain signature analysis revealed a mixture of diagnostic features for both Euryarchaeota and Crenarchaeota kingdoms
water sequences: novel VAL sequences shared major parts of the archaeal domain signatures at homologous and non-homologous positions. However, clones with an unstable phylogenetic placement (VAL90, VAL35-1, VAL31-1, VAL84, VAL125, VAL112) contained non-archaeal signatures at the positions 338, 367, 393 and 867 (E. coli numbering). Intradomain signature analysis confirmed affiliation of twelve VAL sequences with the Crenarchaeota, and sixteen VAL sequences with the Euryarchaeota