SEARCH

SEARCH BY CITATION

Keywords:

  • anoxygenic photosynthetic bacteria;
  • genetic diversity;
  • natural environments;
  • pufM;
  • sequence distance;
  • statistical analysis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Aim:  To assess how completely the diversity of anoxygenic phototrophic bacteria (APB) was sampled in natural environments.

Methods and Results:  All nucleotide sequences of the APB marker gene pufM from cultures and environmental clones were retrieved from the GenBank database. A set of cutoff values (sequence distances 0·06, 0·15 and 0·48 for species, genus, and (sub)phylum levels, respectively) was established using a distance-based grouping program. Analysis of the environmental clones revealed that current efforts on APB isolation and sampling in natural environments are largely inadequate. Analysis of the average distance between each identified genus and an uncultured environmental pufM sequence indicated that the majority of cultured APB genera lack environmental representatives.

Conclusions:  The distance-based grouping method is fast and efficient for bulk functional gene sequences analysis. The results clearly show that we are at a relatively early stage in sampling the global richness of APB species. Periodical assessment will undoubtedly facilitate in-depth analysis of potential biogeographical distribution pattern of APB.

Significance and Impact of the Study:  This is the first attempt to assess the present understanding of APB diversity in natural environments. The method used is also useful for assessing the diversity of other functional genes.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Anoxygenic photosynthetic bacteria (APB) have always attracted attention because of their coloration and ability to perform photosynthesis in the absence of air and without producing oxygen. APB grow favourably in polluted environments created from human activities (Kobayashi et al. 1978) and thus have long been widely used for waste remediation and treatment, such as bioremediation of various organic wastes (Kobayashi and Kobayashi 1995). To date, six APB groups have been described, i.e. Chlorobi, Chloroflexi, purple sulfur bacteria (γ-Proteobacteria), purple nonsulfur bacteria (α-Proteobacteria), low G + C Gram-positive heliobacteria (Firmicutes) and an obligately aerobic APB (a,β-Proteobacteria) (Yurkov and Beatty 1998). Recently, Fuchs et al. (2007) first identified a γ-Proteobacteria aerobic APB from the North Sea surface water and further extended APB genetic diversity.

APB have been isolated from soil, sewage, freshwater and brackish waters, sediments and even hot springs and thermal vents in the sea floor (Overmann and Garcia-Pichel 2003; Beatty et al. 2005), indeed nearly every corner of Earth where light radiation can reach. Despite the great efforts on APB isolation, little was known about the extent of current knowledge on APB diversity in natural environments. A marker gene of anoxygenic photosynthesis, pufM, which encodes the photosynthetic reaction centre small subunit, has been widely used to investigate APB diversity in natural environments (Béjàet al. 2002; Allgaier et al. 2003; Schwalbach and Fuhrman 2005; Yutin et al. 2005; Du et al. 2006; Hu et al. 2006; Okubo et al. 2006; Ranchou-Peyruse et al. 2006). However, these independent studies were confined to various local niches and we lack a global-scale comprehensive analysis of APB overall diversity in natural environments.

For this paper, we retrieved all full-length and partial pufM sequences from the GenBank database. When the pufM sequences from APB cultures were classified into the APB species, genus and phylum, respectively, the minimum distances that make the sequences from the same taxonomical unit group into the same operational taxonomic unit (OTU) could thus be calculated using sequence statistical analysis programs. Based on this rationale, a set of cutoff values for grouping pufM sequences at the species, genus and phylum levels, respectively, was first established. Then, we used them to analyse the pufM sequences cloned directly from various environments with the aim to assess the current efforts on sampling global APB diversity.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Collection of pufM sequences

In this study, we target pufM gene because: (i) the 16S rRNA gene is too conservative to differentiate APB from non-APB (Woese 1987), while anoxygenic photosynthetic genes are exclusively distributed in APB; (ii) the functional genes involved in anoxygenic photosynthesis have received much less evolutionary pressure (Xiong et al. 2000; Raymond et al. 2002) and therefore most likely contain enough nucleotide variations to differentiate different APB species and genera and (iii) a large dataset of pufM sequences is available in the GenBank database.

All full-length and partial pufM sequences and more than 1200 bp 16S rRNA gene sequences of related species were retrieved from the GenBank database (ver. Feb. 2007) using gene name search or BLAST analysis (http://www.ncbi.nih.gov/blast). Taxonomic affiliation of pufM-containing species was confirmed through the 16S rRNA gene sequence classifier tool in the Ribosomal Database Project II website (RDP; Cole et al. 2005). One pufM-containing species (Porphyrobacter sp. MBIC3897) that has no corresponding 16S rRNA gene sequence record was omitted from further analysis. Source environment information for each environmental pufM clone was also obtained from the GenBank records.

Sequence distance-based grouping analysis

The pufM sequences retrieved from the database were divided into two groups: Culture and EnvClone (for Environmental clones). The Culture group only contained the sequences from identified APB strains, while the EnvClone group only contained those directly cloned from environmental DNA samples using culture-independent molecular methods. Sequences of both groups were separately subjected to multi-alignment analysis using the ClustalX program (ver. 1.81) (Thompson et al. 1997). Common sequence region was retained for subsequent distance-based grouping analysis. Distance matrices were constructed using dnadist of the phylip package with the Jukes-Cantor correction for multiple substitutions (http://evolution.genetics.washington.edu/phylip.html). Matrices were then input to the program dotur (Distance-based OTU and Richness; Schloss and Handelsman 2005) for assigning sequences into OTUs using the furthest-neighbour algorithm. For each distance level, curves were constructed: (i) rarefaction curve, calculation of the richness estimators with a 95% confidence interval (CI) as a function of sampling effort; (ii) lineage-through-time plot, description of how many OTUs are present for various evolutionary distances and (iii) actual collector’s curve for each estimate by plotting the CI against the sequencing effort to determine how many sequences are required to obtain a desired level of precision for the estimate.

Grouping analysis of Culture sequences was performed to determine the most appropriate cutoff values for species, genus and (sub)phylum levels using the taxonomical information of all cultured APB species as references. In addition, to assess how well each known APB genus is represented in the environmental sampling effort, Culture and EnvClone sequences were pooled for grouping analysis to determine the mean distance (MD) between each genus and environmental pufM sequences. MD was equal to the average of all the minimum distances between each species within the same genus and an environmental pufM sequence.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Sequences from APB cultures and the determination of cutoff values for different taxonomic units

Eighty-eight APB strains that have both pufM and near full-length 16S rRNA gene sequence records were identified in the GenBank database (for detailed information, see Supplementary Table S1). The corresponding nearly full-length (from 704 to 999 bp, 73 sequences) or partial (from 196 to 575 bp, 15 sequences) pufM genes and source environment information were retrieved. These species were classified into five subphyla or phyla (Table 1), covering four APB groups, i.e. Chloroflexi, purple sulfur bacteria, purple nonsulfur bacteria and aerobic APB. Most genera were affiliated with the α-Proteobacteria, followed by the β-Proteobacteria and γ-Proteobacteria. The multi-alignment analysis identified a 196-bp common region, which corresponded to the most targeted region for amplification of pufM partial sequences from environmental DNA samples (Schwalbach and Fuhrman 2005; Du et al. 2006; Hu et al. 2006; Okubo et al. 2006; Ranchou-Peyruse et al. 2006). These environmental pufM surveys also confirmed the efficiency of this region for phylogenetic analysis. Here we also inferred distance matrices from this region using the newly developed grouping program dotur. Detailed distance data for pufM sequences from cultures are shown in Table 1.

Table 1.   Distance matrix of pufM sequences (196 bp) from cultured anoxygenic photosynthetic bacteria
(Sub)phylum genusSpeciesDistance*
Spe.Gen.Phy.
  1. Values indicate the minimum average distance within the same species, genus or (sub)phylum. Bacteroidetes contains only one sequence and is not shown.

  2. *Minimum distance for grouping the sequences; Spe., species level; Gen., genus level; Phy., (sub)phylum level.

  3. †Numbers in parentheses indicate total number of genera/species within each (sub)phylum.

  4. ‡Mean of nonzero values.

α-Proteobacteria (24/71)†   0·68
Acidiphilium  0·14 
RhodospirillumR. rubrum00·20 
Roseococcus    
Rubritepida    
Jannaschia    
Loktanella    
RhodobacterR. capsulatus00·21 
R. sphaeroides0·03  
Rhodovulum    
RoseobacterR. denitrificans00·18 
Roseovarius    
Blastomonas  0·10 
Citromicrobium    
Erythrobacter  0·22 
Erythromicrobium    
Agrobacterium    
Porphyrobacter  0·16 
Sphingomonas  0·12 
Roseospirillum    
Blastochloris    
BradyrhizobiumB. denitrificans0·110·11 
MethylobacteriumM. extorquens0·020·10 
RhodomicrobiumR. vannielii0·01  
Rhodoplanes    
RhodopseudomonasR. palustris0·15  
β-Proteobacteria (4/5)   0·45
Rhodocyclus  0·21 
Rhodoferax    
Roseateles    
Rubrivivax    
γ-Proteobacteria (6/7)   0·42
AllochromatiumA. vinosum0  
Congregibacter    
Ectothiorhodospira    
Lamprocystis    
Thiocapsa    
Thiocystis    
Chloroflexi (2/4)   0·39
ChloroflexusC. aurantiacus00·11 
Roseiflexus    
Mean 0·06‡0·150·48

Ten species, 12 genera, and four (sub)phyla contained more than one different strains, species and genera, respectively. Five species that contained only the strains with identical pufM sequences could not represent real within-species difference and were excluded from the measure of average within-species distance. We calculated the average distance as 0·06, 0·15 and 0·48 at the species, genus and (sub)phylum level, respectively. The lineage-through-time curve showed these cutoff values closely corresponded to the main inflexion points (Fig.1b, Left), suggesting that they may be efficient for grouping pufM sequences at different taxonomical levels.

image

Figure 1.  Rarefaction curve (a), lineage-through-time plot (b), and actual collector’s curve (c) for pufM sequences from 88 anoxygenic photosynthetic bacterial strains (Left) and 314 pufM sequences cloned directly from various environments (Right). Numbers on the right side of curves or dot lines are cutoff values.

Download figure to PowerPoint

We used these cutoff values to assess the sampling effort and richness of APB cultures. The steep slope of rarefaction curves at the species and genus levels (Fig.1a, Left) demonstrates that current efforts related to APB isolation are inadequate. A great number of new APB species and genera remain to be discovered. However, at the (sub)phylum level, the nearly saturated rarefaction curve indicates little possibility of discovering new APB phyla or subphyla. After 60 sequences for species level and 20 sequences for genus level (Fig.1c, Left), the rate of change for the estimator greatly decreased, indicating that we may be close to obtaining reasonable estimates of richness.

Sequences from environmental clones

Total 314 pufM sequences were retrieved from the database (for detailed information, see Supplementary Table S1). Four types of source environment were identified: seawater, lake, sediment and attachment, containing 222, 35, 49 and 8 sequences, respectively. The lineage-through-time analysis also showed that the cutoff values closely corresponded to the main inflexion points (Fig. 2b, Right), which further confirmed their efficiency.

image

Figure 2.  Average distance between each genus and environmental pufM sequences. Gray lines indicate the average value for each (sub)phylum. White columns indicate the genera with the distance of less than 0·15.

Download figure to PowerPoint

EnvClone sequences were grouped into 113 species and 73 genera. Rarefaction curves (Fig. 2a, Right) indicate that the possibility of discovering new sequences remains high. The curves at the species and genus levels show a steep slope with a lower gradient than that of the Culture’s curve (Fig. 2a, Right), demonstrating that current efforts on the sampling of APB diversity in natural environments is also greatly inadequate. Similarly, the nearly saturated curve at the (sub)phyla level (0·48 distance) indicates the low possibility of discovering new APB phyla or subphyla. It is noted that the Chao1 richness estimate rarefaction curve began to level off at the species and genus levels after about 50 sequences (Fig.1c, Right). The greatly decreased rate of change for the estimator also means a possibly reasonable estimate of APB richness in natural environments.

Distance between sequences from cultures and environmental samples

At the species level, only one genus, Allochromatium of the γ-Proteobacteria, shows the distance 0·01 with a clone from the sediment of a wastewater treatment plant; whereas within the genus level, eight genera (four in the α-Proteobacteria, three in the β-Proteobacteria, and two in the γ-Proteobacteria) show representatives in the environmental clones (Fig. 2). Average distances for the α-Proteobacteria, β-Proteobacteria and γ-Proteobacteria were similar, but slightly higher than the genus level of 0·15 (Fig. 2). Only one and two APB genera that have pufM sequenced were identified within the phyla Bacteroidetes and Chloroflexi, respectively, making it difficult to produce data with statistical significance. However, it is clear that the Chloroflexi has no any relative pufM sequences in the environmental clones.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Assessment of the distance-based grouping method and cutoff values

A commonly used method for grouping sequences is phylogenetic analysis. However, when the dataset is more than 100 sequences, this method is extremely time-consuming, especially for the most likelihood algorithm. Therefore, we used a distance-based grouping method to analyse the more than 400 pufM sequences in this study. Currently, there are few methods available to quickly assign sequences to OTUs, such as FastGroup (Seguritan and Rohwer 2001) and dotur (Schloss and Handelsman 2005). dotur was shown to assign sequences to OTUs more accurately and consistently than had previous methods (Schloss and Handelsman 2005) and therefore was used in this study. dotur assigned the pufM sequences rapidly (ca. 30s for 402 200-bp-length sequences on a P4-level PC) and systematically to OTUs at all possible cutoff values. Furthermore, dotur assisted in assessing the completeness of sequencing effort and the reliability of richness estimates (Schloss and Handelsman 2005). Nonparametric richness estimators by dotur permit a mathematical estimate of richness without requiring that each OTU be sampled (Hughes et al. 2001). When an estimator rarefaction curve levels off, the value to which the curve converges is a reasonable estimate of the true richness (Hughes et al. 2001). These fast and multiple functions make it an ideal sequence grouping tool.

Using the distance analysis of APB culture sequences, we first established a set of cutoff values. For 16S rRNA gene sequences, the mostly used distance values of 0·03, 0·05, 0·10 and 0·20 are thought to differentiate at the species, genus, family/class and phylum levels, respectively (Stackebrandt and Goebel 1994; Hugenholtz et al. 1998; Sait et al. 2002). However, little was reported about cutoff values for functional genes. The few studies available to date on the assessment of functional gene diversity used only empirical cutoff values. For instance, 5% sequence difference was used as the OTU definition for the ammonia monooxygenase gene (amoA) (Beman and Francis 2006). Schloss and Handelsman (2005) predicted a richness of 1040 species in the Sargasso Sea metagenome by using an empirical cutoff value of 6% difference as species definition for rpoB gene, comparable to the previous estimate by Venter et al. (2004). In this study, we calculated an equivalent cutoff value, suggesting the general applicability of the distance 0·06 for grouping protein-coding sequences at the species level. Meanwhile, we calculated cutoff values for the genus and even (sub)phylum levels, providing more parameters for analysing the environmental sequences and therefore a more thorough understanding of natural microbial diversity.

The method of establishing cutoff values from culture sequences would also be useful for other functional genes. However, some limitations should be noted: (i) the number of APB whose pufM have been identified remains insufficient to reach a statistically significant result; (ii) different regions of the pufM gene or different fragment lengths selected may result in different cutoff values and (iii) if other functional genes specific for APB have similar large datasets for grouping analysis, it may give different cutoff values. Therefore, cautions should be taken to perform an analysis similar to that used in this study.

Diversity of APB in natural environments

The establishment of cutoff values made it possible to assess our current understanding of the diversity of APB species and genera in natural environments. Although any distance level that is selected to differentiate species or genera will be controversial due to insufficient data, it will still serve as a useful benchmark for further analyses. Furthermore, with the present both less expensive and largely automated sequencing technology, more sequence data will be available for APB. This study provides some new ideas for the exploitation of sequence resources in public databases.

Because the Chao1 collector’s curves do not level off, it is difficult to determine how many more sequences would need to be sampled in order to obtain an accurate estimate of global APB richness. Considering the steady rise in the richness estimate with sampling, these estimates are clearly minimum values of richness and should rise with increased isolation or sampling effort. It has been proposed that at least 10 000 16S rRNA gene sequences would be necessary to approach an estimate of the true species richness in soil (Schloss and Handelsman 2005). For the widely distributed APB, current limited data are insufficient to give an accurate estimate of natural APB diversity.

For improving the assessment’s accuracy, we make four proposals based on our results: (i) increase sampling effort of pufM in nonmarine environments, such as species-rich surface soil, light-reachable sediment or other attachments, lakes or rivers, and other anoxic but light-reachable niches, etc., because marine pufM sequences account for the largest part of environmental sequences in the database; (ii) improve the coverage of current pufM primers as no sequence closely related to Chloroflexi were found in the environmental clones and (iii) establish a secondary database for APB, specially designed for storage of APB-related information, such as library size, primers, gene sequences and source information, to facilitate more in-depth analysis in the future, such as the potential biogeographical distribution of APB in natural environments. Clearly, these endeavours will undoubtedly increase to better understand the APB diversity pattern and their potential distribution mechanisms.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work was supported by NSFC Projects: No. 40576063, 40521003 and 40632013. Professor I. J. Hodgkiss (The University of Hong Kong) is thanked for his assistance with English.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Allgaier, M., Uphoff, H., Felske, A. and Wagner-Dolber, I. (2003) Aerobic anoxygenic photosynthesis in Roseobacter clade bacteria from diverse marine habitats. Appl Environ Microbiol 69, 50515059.
  • Beatty, J.T., Overmann, J., Lince, M.T., Manske, A.K., Lang, A.S., Blankenship, R.E., Van Dover, C.L., Martinson, T.A. et al. (2005) An obligately photosynthetic bacterial anaerobe from a deep-sea hydrothermal vent. Proc Natl Acad Sci USA 102, 93069310.
  • Béjà, O., Suzuki, M.T., Heidelberg, J.F., Nelson, W.C., Preston, C.M., Hamada, T., Eisen, J.A., Fraser, C.M. et al. (2002) Unsuspected diversity among marine aerobic anoxygenic phototrophs. Nature 415, 630633.
  • Beman, J.M. and Francis, C.A. (2006) Diversity of ammonia-oxidizing archaea and bacteria in the sediments of a hypernutrified subtropical estuary: Bahia del Tobari, Mexico. Appl Environ Microbiol 72, 77677777.
  • Cole, J.R., Chai, B., Farris, R.J., Wang, Q., Kulam, S.A., McGarrell, D.M., Garrity, G.M. and Tiedje, J.M. (2005) The Ribosomal Database Project (RDP-II): sequences and tools for high-throughput rRNA analysis. Nucleic Acids Res 33, D294D296.
  • Du, H.L., Jiao, N.Z., Hu, Y.H. and Zeng, Y.H. (2006) Real-time PCR for quantification of aerobic anoxygenic phototrophic bacteria based on pufM gene in marine environment. J Exp Mar Biol Ecol 329, 113121.
  • Fuchs, B.M., Spring, S., Teeling, H., Quast, C., Wulf, J., Schattenhofer, M., Yan, S., Ferriera, S. et al. (2007) Characterization of a marine gamma- proteobacterium capable of aerobic anoxygenic photosynthesis. Proc Natl Acad Sci USA 104, 28912896.
  • Hu, Y.H., Du, H.L., Jiao, N.Z. and Zeng, Y.H. (2006) Abundant presence of the gamma-like Proteobacterial pufM gene in oxic seawater. FEMS Microbiol Lett 263, 200206.
  • Hugenholtz, P., Goebel, B.M. and Pace, N.R. (1998) Impact of culture independent studies on the emerging phylogenetic view of bacterial diversity. J Bacteriol 180, 47654774.
  • Hughes, J.B., Hellmann, J.J., Ricketts, T.H. and Bohannan, B.J.M. (2001) Counting the uncountable: statistical approaches to estimating microbial diversity. Appl Environ Microbiol 67, 43994406.
  • Kobayashi, M. and Kobayashi, M. (1995) Waste remediation and treatment using anoxygenic phototrophic bacteria. In Anoxygenic Photosynthetic Bacteria ed. Blankenship, R.E., Madigan, M.T. and Baucer, C.E. pp. 12691282 Dordrecht, the Netherlands: Kluwer Academic.
  • Kobayashi, M., Fujii, K., Shimamoto, I. and Maki, T. (1978) Treatment and re-use of industrial waste water by phototrophic bacteria. Prog Water Technol 11, 279284.
  • Okubo, Y., Futamata, H. and Hiraishi, A. (2006) Characterization of phototrophic purple nonsulfur bacteria forming colored microbial mats in a swine wastewater ditch. Appl Environ Microbiol 72, 62256233.
  • Overmann, J. and Garcia-Pichel, F. (2003) The phototrophic way of life. In The Prokaryotes, an Evolving Electronic Resource for the Microbiological Community ed. Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K. and Stackebrandt, E. Available at: http://141.150.157.117:8080/prokPUB/index.htm.
  • Ranchou-Peyruse, A., Herbert, R., Caumette, P. and Guyoneaud, R. (2006) Comparison of cultivation-dependent and molecular methods for studying the diversity of anoxygenic purple phototrophs in sediments of a eutrophic brackish lagoon. Environ Microbiol 8, 15901599.
  • Raymond, J., Zhaxybayeva, O., Gogarten, J.P., Gerdes, S.Y. and Blankenship, R.E. (2002) Whole-genome analysis of photosynthetic prokaryotes. Science 298, 16161620.
  • Sait, M., Hugenholtz, P. and Janssen, P.H. (2002) Cultivation of globally distributed soil bacteria from phylogenetic lineages previously only detected in cultivation- independent surveys. Environ Microbiol 4, 654666.
  • Schloss, P.D. and Handelsman, J. (2005) Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl Environ Microbiol 71, 15011506.
  • Schwalbach, M.S. and Fuhrman, J.A. (2005) Wide-ranging abundances of aerobic anoxygenic phototrophic bacteria in the world ocean revealed by epifluorescence microscopy and quantitative PCR. Limnol Oceanogr 50, 620628.
  • Seguritan, V. and Rohwer, F. (2001) FastGroup: a program to dereplicate libraries of 16S rDNA sequences. BMC Bioinformatics 2, 9.
  • Stackebrandt, E. and Goebel, B.M. (1994) A place for DNA–DNA reassociation and 16S rRNA sequence-analysis in the present species definition in bacteriology. Int J Syst Bacteriol 44, 846849.
  • Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. and Higgins, D.G. (1997) The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 24, 48764882.
  • Venter, J.C., Remington, K., Heidelberg, J.F., Halpern, A.L., Rusch, D., Eisen, J.A., Wu, D., Paulsen, I. et al. (2004) Environmental genome shotgun sequencing of the Sargasso Sea. Science 304, 6674.
  • Woese, C.R. (1987) Bacterial evolution. Microbiol Mol Biol Rev 51, 221271.
  • Xiong, J., Fischer, W.M., Inoue, K., Nakahara, M. and Bauer, C.E. (2000) Molecular evidence for the early evolution of photosynthesis. Science 289, 17241730.
  • Yurkov, V.V. and Beatty, J.T. (1998) Aerobic anoxygenic phototrophic bacteria. Microbiol Mol Biol Rev 62, 695724.
  • Yutin, N., Suzuki, M.T. and Béjà, O. (2005) Novel primers reveal wider diversity among marine aerobic anoxygenic phototrophs. Appl Environ Microbiol 71, 89588962.