Technical Comments

Comment on "A 3-Hydroxypropionate/4-Hydroxybutyrate Autotrophic Carbon Dioxide Assimilation Pathway in Archaea"

See allHide authors and affiliations

Science  18 Jul 2008:
Vol. 321, Issue 5887, pp. 342
DOI: 10.1126/science.1158766


Berg et al. (Reports, 14 December 2007, p. 1782) reported the discovery of an autotrophic carbon dioxide–fixation pathway in Archaea and implicated a substantial role of this pathway in global carbon cycling based on sequence analysis of Global Ocean Sampling data. We question the validity of the latter claim.

Berg et al. (1) reported the discovery of a CO2-fixation pathway in autotrophic members of the archaeal order Sulfolobales, as well as in members of the Cenarchaeales and Archaeoglobales. This pathway, which the authors refer to as the 3-hydroxypropionate/4-hydroxybutyrate pathway, is shown to comprise a cycle of 16 enzymes, one of which is the proposed key enzyme 4-hydroxybutyryl-CoA dehydratase (4HCD). Based on a comparison of abundances of 4HCD and RuBisCo sequences in the Global Ocean Sampling (GOS) data (2), Berg et al. (1) predicted the abundance of the newly discovered CO2-fixation pathway in the ocean surface waters to be of the same order of magnitude as the Calvin-Bassham-Benson cycle, which is known to be of global importance for carbon cycling. In addition, the authors proposed the existence of an abundant group of mesophilic autotrophic Crenarchaea in the ocean surface waters that uses the proposed 3-hydroxypropionate/4-hydroxybutyrate pathway for CO2 assimilation. We raise several concerns about the validity of these claims.

First, the abundance of 4HCD homologs does not necessarily reflect the abundance of the 3-hydroxypropionate/4-hydroxybutyrate pathway, because this enzyme is also known to participate in 4-aminobutyrate fermentation in a few strict anaerobic bacteria (3). Moreover, homologs of 4HCD are implicated in unrelated metabolic processes, including phenylacetate catabolism (4) and pyoverdine chromophore biosynthesis (5).

Second, solely comparing the “number of identified...gene sequences” (1) in metagenomics data does not provide an accurate estimate of the relative abundances of the enzyme or the species, because these numbers depend on the criteria used for homolog identification, gene copy numbers, presence of paralogs, rates of gene sequence evolution, cloning efficiencies, and lengths of the individual genes, or even the choice of filter-fraction size for sample collection. For example, the relatively large cell sizes of cosmopolitan cyanobacteria such as Synechococcus and Prochlorococcus has likely resulted in a considerable underrepresentation of such sequences in the GOS data set (6), and hence of Calvin-Bassham-Benson cycle sequences. Thus, attempts to infer the abundance of the novel CO2-fixation pathway relative to the (cyanobacterial) Calvin-Bassham-Benson cycle based on the number of sequence reads with gross similarity to 4HCD and RuBisCo, respectively (1), should be considered with caution.

Third, below the euphotic zone (>150 m), pelagic crenarchaeota are known to make up a large fraction of total marine picoplankton (7) and to perform pivotal roles in marine nitrogen and carbon cycles (8). Basedonananalysisof the GOS ocean surface metagenome data, Berg et al. alluded to the existence of a “group of abundant mesophilic autotrophic crenarchaea” in the ocean surface waters, which they propose use the 3-hydroxypropionate/4-hydroxybutyrate pathway for CO2 assimilation (1). This hypothesis is not supported by data from the GOS expedition (2), which indicated that archaeal sequences were less abundant by a factor of ∼33 than bacterial sequences (2.7% and 90.8% of total reads, respectively) (6). Of the 60 most abundant 16S rRNA ribotypes, Prochlorococcus and Synechococcus show 171-fold coverage of the 16S rRNA gene as compared to 8-fold coverage for Archaea (2). Thus, there is at least a 20-fold difference between the most abundant cyanobacterial and archaeal species in the upper surface waters.

These apparent discrepancies prompted us to reexamine the data that brought Berg et al. to their conclusion. We carefully reanalyzed the proposed phylogeny of 4HCD protein sequences (9) and, like Berg et al. (1), observed that the GOS environmental sequences formed a cluster that was more closely related to a clade of anaerobic bacterial species that use 4HCD in fermentation than to the crenarchaea that use this enzyme for CO2 fixation [figure 3 in (1)]. Within the “marine cluster”–1, the GOS1 and GOS2 groups clustered with high support, whereas the proposed 4HCD from the mesophilic crenarchaeaote Cenarchaeum symbiosum was found to cluster within the GOS3 group (Fig. 1). The 4HCD homolog from the strictly aerobic, chemoheterotrophic marine myxobacterium Plesiocystis pacifica is also a member of the “marine cluster”–1, but its exact placement could not be resolved in either analysis [Fig. 1 and figure 3 in (1)].

Fig. 1.

Phylogenetic relationships of 4HCD protein sequences and a kingdom analysis (9) of proteins encoded by the genes flanking the 4HCD sequences in the GOS scaffolds. The GOS1 and GOS2 clades are dominated by bacterial genes (also see fig. S1), indicating that these 4HCD gene products are unlikely to function in the proposed archaeal 3-hydroxypropionate/4-hydroxybutyrate pathway for CO2 assimilation. As a control, the results for the same BLASTP-search-based method are depicted for 50 random proteins from C. symbiosum and P. pacifica, respectively (excluding self-hits). The phylogenetic tree of the 4HCD proteins is based on the same sequence alignment as in Berg et al. (1, 9). The scale bar represents a difference of 1 substitution per site, and the numbers at the nodes indicate the resampling estimated log-likelihood (RELL) support values. Only RELL values above 0.95 are shown. Major groups that gave a similar overall topology as found by Berg et al. [figure 3 in (1)] have been collapsed and depicted as triangles.

Because the phylogenetic analysis fails to distinguish between a bacterial versus an archaeal origin of the environmental sequences in “marine cluster”–1, we analyzed the remainder of the (predicted) genes encoded by the scaffolds on which the 4HCD sequences are located. Sequences that reside on the same scaffold originate from the same species, and an analysis of all genes encoded by the scaffold should in principle distinguish bacterial from archaeal sequences. A search for sequence similarity of proteins encoded by the GOS scaffold using the basic local alignment search tool for proteins (BLASTP) (9) revealed that the scaffolds associated with GOS1 and GOS2 groups are dominated by bacterial genes and are thus likely of bacterial origin (Fig. 1). The scaffolds associated with the GOS3 group of “marine cluster”–1 contain mostly archaeal genes and might thus be of archaeal origin. Because the 3-hydroxypropionate/4-hydroxybutyrate CO2 assimilation pathway is thought to be restricted to archaeal species, the most plausible explanation is that the bacterial 4HCD-like sequences that dominate “marine cluster”–1 are most likely involved in another pathway.

Furthermore, 4HCD is sensitive to oxygen exposure due to inactivation of the [4Fe-4S] iron-sulfur clusters that reside in its active center. The enzyme might be sufficiently stable at the thermophilic, and thus low-oxygen pressure, environments in which members of the Sulfolobales reside. However, the oxygen-rich conditions in the ocean surface waters might require additional measures to prevent inactivation. Interestingly, the amino acid residues that constitute the iron-sulfur cluster pocket are not conserved in the “crenarchaea type-2” 4HCD sequences [figure 3 in (1)], hinting at a possible adaptation toward oxygen exposure. It remains unclear how the seemingly intact iron-sulfur clusters from “marine cluster”–1 4HCDs resist oxygen inactivation.

Finally, Berg et al. predict the 3-hydroxypropionate/4-hydroxybutyrate pathway to be operational in C. symbiosum and Archaeoglobus fulgidus (1). Because we were unable to detect candidate genes for some of the components of the pathway in these species, this conclusion seems premature. For example, candidates for succinyl/malonyl-CoA reductase, succinate-semialdehyde reductase, and the 4-hydroxybutyryl-CoA and 3-hydroxypropionyl-CoA synthetases are missing or cannot be conclusively identified for C. symbiosum (Fig. 2). Thus, one cannot rule out that these organisms rely on yet another variant of the 3-hydroxypropionate pathway for CO2 fixation (10). Future experimental studies of the autotrophic pathways of mesophilic crenarchaeaotes, such as Nitrosopumilus maritimus, should resolve this issue. The anticipated diversity in archaeal CO2 assimilation pathways resembles the diversity observed in other archaeal central carbon metabolic pathways (11).

Fig. 2.

Distribution of the 16 enzymes that make up the 3-hydroxypropionate/4-hydroxybutyrate pathway across archaeal genomes (9). Gene presence was inferred from the archaeal cluster of orthologous groups of proteins (12) (in black shading) or BLASTP searches (in gray shading). The patchy distribution pattern across species hints at the existence of variants of the 3-hydroxypropionate pathway for CO2 assimilation (such as in C. symbiosum and N. maritimus) or at enzymes that connect to or operate in other pathways (13). Species abbreviations: Nitma, Nitrosopumilus maritimus; Censy, Cenarchaeum symbiosum; Metse, Metallospaera sedula; Sulso, Sulfolobus solfataricus P2; Pyrae, Pyrobaculum aerophilum; Thete, Thermoproteus tenax; Calma, Caldivirga maquilingensis IC-167; Thepe, Thermofilum pendens Hrk 5; Stama, Staphylothermus marinus F1; Ighos, Ignicoccus hospitalis; Hypbu, Hyperthermus butylicus; Aerpe, Aeropyrum pernix; Arcfu, Archaeoglobus fulgidus. Enzymes:1, acetyl-CoA carboxylase; 2, malonyl-CoA reductase (NADPH); 3, malonate semialdehyde reductase (NADPH); 4, 3-hydroxypropionyl-CoA synthetase (AMP-forming); 5, 3-hydroxypropionyl-CoA dehydratase; 6, acryloyl-CoA reductase (NADPH); 7, propionyl-CoA carboxylase; 8, methylmalonyl-CoA epimerase; 9, methylmalonyl-CoA mutase; 10, succinyl-CoA reductase (NADPH); 11, succinate semialdehyde reductase (NADPH); 12, 4-hydroxybutyryl-CoA synthetase (AMP-forming); 13, 4-hydroxybutyryl-CoA dehydratase; 14, crotonyl-CoA hydratase [(S)-3-hydroxybutyryl-CoA-forming]; 15, (S)-3-hydroxybutyryl-CoA dehydrogenase (NAD+); 16, acetoacetyl-CoA β-ketothiolase.

Supporting Online Material

Materials and Methods

Fig. S1

Table S1


References and Notes

View Abstract

Navigate This Article