Microbial rhodopsins on leaf surfaces of terrestrial
Nof Atamna-Ismaeel1,†, Omri M. Finkel2,†, Fabian Glaser3, Itai Sharon1,4,10, Ron
Schneider2, Anton F. Post5, John L. Spudich6, Christian von Mering7, Julia A.
Vorholt8, David Iluz9, Oded Béjà1,*, and Shimshon Belkin2,*
1Faculty of Biology,
Interdisciplinary Center for Life Sciences and Engineering,
Computer Science, Technion – Israel Institute of Technology, Haifa 32000,
3Bioinformatics Knowledge Unit, Lorry I. Lokey
2Department of Plant and Environmental Sciences, Alexander Silberman
Institute of Life Sciences, Hebrew University of Jerusalem, Jerusalem 91904,
5Josephine Bay Paul Center for Comparative Molecular Biology and Evolution,
Marine Biology Laboratory, Woods Hole, MA 02543, USA.
6Center for Membrane Biology, Department of Biochemistry and Molecular
Biology, The University of Texas Medical School, Houston, TX 77030, USA.
7Faculty of Science, Institute of Molecular Life Sciences and Swiss Institute of
Bioinformatics, University of Zurich, 8057 Zurich, Switzerland.
8Institute of Microbiology, Eidgenössische Technische Hochschule Zurich,
Wolfgang-Pauli-Strasse 10, 8093 Zurich, Switzerland.
9Faculty of Life Sciences, Bar-Ilan University, Ramat Gan 52900, Israel
†N.A.-I. and O.M.F. contributed equally to this work
*To whom correspondance should be addressed. E-mails: (@tx.technion.ac.il)
or S.B. ( @vms.huji.ac.il).
10Present address: Department of Earth and Planetary Science, University of
California, Berkeley, CA 94720, USA
The above-ground surfaces of terrestrial plants, the phyllosphere,
comprise the main interface between the terrestrial biosphere and solar
radiation. It is estimated to host up to 1026 microbial cells that may
intercept part of the photon flux impinging on the leaves. Based on 454-
pyrosequencing generated metagenome data, we report on the existence
of diverse microbial rhodopsins in five distinct phyllospheres from
tamarisk (Tamarix nilotica), soybean (Glycine max), Arabidopsis
(Arabidopsis thaliana), clover (Trifolium repens) and rice (Oryza sativa).
Our findings, for the first time describing microbial rhodopsins from non-
aquatic habitats, point toward the potential coexistence of microbial
rhodopsin-based phototrophy and plant chlorophyll-based
photosynthesis, with the different pigments absorbing non-overlapping
fractions of the light spectrum.
Solar radiation is the main source of energy for both marine and terrestrial
organisms, with terrestrial plants and aquatic phytoplankton performing an
equivalent ecological function as chlorophyll-based photosynthetic primary
producers (Field et al., 1998). Marine surface waters are now known to harbour
an additional type of phototrophy; several lineages of bacteria and archaea
utilize rhodopsins (Béjà et al., 2000; Béjà et al., 2001; de la Torre et al., 2003;
Balashov et al., 2005; Giovannoni et al., 2005; Sabehi et al., 2005; Frigaard et
al., 2006; Gómez-Consarnau et al., 2007; Gómez-Consarnau et al., 2010; Oh et
al., 2010), retinal-containing trans-membrane proteins, as light-driven proton
pumps. The first microbial rhodopsin was discovered nearly four decades ago in
the archaeon Halobacterium salinarum from hypersaline environments
(Oesterhelt and Stoeckenius, 1971). Further studies revealed the existence of
microbial rhodopsins in diverse habitats including freshwater, sea ice,
hypersaline and brackish environments (Rusch et al., 2007; Atamna-Ismaeel et
al., 2008; Sharma et al., 2008; Sharma et al., 2009; Koh et al., 2010). To date,
microbial rhodopsins have been reported exclusively for aquatic habitats.
As light is an abundant resource on land, we tested the hypothesis that
microbial rhodopsins also exist and play an important role in terrestrial niches.
The leaf surface of terrestrial plants covers a surface area of an estimated
6.4X108 km2 and comprises the main interface between terrestrial biomass and
solar photon flux. This habitat harbors an immensely diverse microbial
community of up to 106 - 107 cells per cm2 leaf surface (Lindow and Brandl,
2003). A mode of phototrophy that is compatible with the plant’s photosynthesis
would offer a significant ecological advantage to microbes inhabiting this
Results and Discussion
We have identified 156 microbial rhodopsin sequences in five
phyllosphere metagenomes [files S1, S2, S3, S4, and S5], from different
terrestrial plants, soybean (Glycine max) (Delmotte et al., 2009), tamarisk
(Tamarix nilotica), clover (Trifolium repens), rice (Oryza sativa) as well as from a
wild population of the model plant Arabidopsis thaliana. The size of the different
metagenomes obtained was 261 Mb, 448 Mb, 234 Mb, 831 Mb, and 250 Mb for
soybean, tamarisk, clover, rice and Arabidopsis with an average read length of
235, 328, 235, 357 and 233 bp, respectively.
Phylogenetic analysis revealed that some phyllosphere microbial
rhodopsins have branched away from known rhodopsin families within the
bacterial and eukaryal domains (Fig. 1). Some of these sequences clustered
with fungal rhodopsins, while another group clustered with xanthorhodopsins
(Balashov et al., 2005; Lanyi and Balashov, 2008) and actinorhodopsins
(Sharma et al., 2008; Sharma et al., 2009). However, most phyllosphere
rhodopsins appear on novel branches, with no representatives from either
culture-based or environmental datasets, thus rendering them with an as yet
uncertain phylogenetic affiliation. In most cases, the leaf surface rhodopsins
from tamarisk clustered separately from other phyllosphere rhodopsins (Fig.1)
with a statistically significant phylogenetic signal [calculated using Mesquite
(Maddison and Maddison, 2010)], indicating that they reside in distinct microbial
taxa, probably adapted to the unique hypersaline environment of the tamarisk
phyllosphere (Qvit-Raz et al., 2008).
In contrast with soil metagenomes, which do not contain any rhodopsin reads,
the five phyllosphere datasets were found to contain microbial rhodopsins,
however, at frequencies lower than those found in marine and freshwater
metagenomes (Fig. 2). While some of the phyllosphere rhodopsins lack the
retinylidene Schiff base proton donor carboxylate and are thus likely sensory
rhodopsins, others contain both proton acceptor and donor carboxylates at helix
C (bacteriorhodopsin positions 85 and 96, respectively; see files S1, S2, S3, S4,
and S5) and may be considered as potential proton pumps. Compared to the
marine environment, where they make up only 3% of all microbial rhodopsins
(Spudich, 2006), the contribution of sensory rhodopsins to phyllospheres is
much higher (25-70%; Fig. 3). This suggests that microorganisms in the
phyllosphere are intensively engaged in light sensing, to accommodate the
effects of fluctuations in light quality, intensity and UV radiation at the leaf
surface (Ballaré et al., 1990; Beattie and Lindow, 1999).
Interestingly, all phyllosphere rhodopsins detected carry a leucine
residue at position 105 (Fig. 4; based on sequence reads that contain this
region; not all reads cover the entire gene due to the short nature of the 454-
generated sequences, ~250-300 nt on average), which renders them as
putative green light absorbing pigments (Man et al., 2003), thus avoiding an
overlap with the absorption spectrum of the plant’s leaf and possibly even
masking out the negative role of green light on plant growth (Folta and
Maruhnich, 2007). This is opposed to blue light absorbing rhodopsins (Béjà et
al., 2001; Sabehi et al., 2005), which contain a glutamine instead of leucine at
position 105, and are abundantly found in marine habitats (Béjà et al., 2001;
Rusch et al., 2007; Sabehi et al., 2007).
Another indication that this may indeed be the case in the tamarisk
phyllosphere is presented by the absorption spectra in Fig. 5; it is demonstrated
that the microbes washed off the leaves has an absorption maximum around
545 nm, a region of the spectrum where there is no light absorption by the
tamarisk leaves and where the absorption of microbial rhodopsins is maximal.
This absorption peak however, could also be the result of the presence of pink-
pigmented Methylobacterium spp. containing carotenoids (Kutschera, 2007) in
the leaf wash.
This is the first report on microbial rhodopsins existence in terrestrial
habitats; whether it portrays commensalism or mutualism should be a matter of
further investigations. We show that rhodopsin sequences have been found to
be abundant both in the harsh environment of the tamarisk phyllosphere (Qvit-
Raz et al., 2008) as well as on the leaves of cultivated plants; furthermore, they
are common to diverse leaf shapes and plant growth characteristics, but are
absent from both agricultural and forest soils. This indicates that microbial
rhodopsins may be selected for in the phyllosphere environment, thus
conferring an important adaptive trait onto this microbial niche. We propose that
rhodopsin light interception by phyllosphere bacteria needs to be taken into
account in global energy balance and biomass production by the terrestrial
Leaf samples were collected from a Tamarix nilotica tree in an oasis by the
Dead Sea (31º42’41.06’’N 35º27’19.32’’E), and processed within 1 hour of
sampling (Qvit-Raz et al., 2008). Briefly, 50 grams of leaves were placed inside
a 250 ml sterile glass Erlenmeyer flask, immediately immersed in sterile
phosphate buffered saline (1 g leaf/5 ml PBS, pH 7.4), and cavitated in a
sonication bath (Transistor/ultrasonic T7 [L&R Manufacturing Company]) for two
minutes at medium intensity. The preparations were then vortexed 6 X 10 sec at
5-min intervals, and the leaf wash was separated from the leaf debris by
decanting and kept for analysis. Arabidopsis, clover and rice phyllospheres
were prepared according to the previously reported soybean phyllosphere
preparation (Delmotte et al., 2009).
DNA Extraction and pyrosequencing
The leaf wash was filtered on a 0.22 µm membrane filter (Millipore), which was
subjected to total community DNA extraction, using a Power Soil Microbial DNA
extraction kit (MoBio). Sequencing was performed on the Genome Sequencer
FLX system using 3 µg of DNA at a concentration of 17 ng/µl (as determined by
a nanodrop spectrophotometer). The resulting reads were annotated using the
MG-RAST rapid annotation platform (Meyer et al., 2008). Using this platform,
rhodopsin-containing reads were located within each of the compared
metagenomes using an e-value cutoff of 10-5. For inclusion in the phylogenetic
analysis, hits with higher e-values were included as well. The number of reads
was normalized against the average number of selected single-copy genes
found in the datasets using an e-value cutoff of 10-20.
All non-phyllosphere datasets used are publicly available on the MG-
RAST website. The soybean phyllosphere metagenome can be found in the
genbank SRA database. The rhodopsin-containing reads from the phyllosphere
metagenomes are provided in the online supporting material files S1, S2, S3,
S4, and S5.
Phylogenetic tree analysis
In this work, we tried several methods for multiple sequence alignment
calculation (MUSCLE, ProbCons, MAFFT and PROMALS, see references
within (Kemena and Notredame, 2009)). In an effort to automatically identify the
most reliable multiple sequence alignment for a given protein family, we used
the AQUA protocol for automated quality improvement of multiple sequence
alignments (Muller et al., 2010). We performed several alignments using
MUSCLE, MAFFT, ProbCons, along with one refinement program (RASCAL)
and one assessment program (NORMD). According to this protocol the MAFFT
alignment refined by RASCAL produced the most reliable alignment (highest
NORMD value) and was used to produce the phylogenetic tree. Following the
alignment computation, we used FastTree version 2.1.1 SSE3 (Price et al.,
2009) for the calculation of the phylogenetic tree using settings for high
accuracy [-spr 4 (to increase the number of rounds of minimum-evolution SPR
moves) and -mlacc 2 -slownni
To tested if the phylogenetic signal we observe is statically significant we
used the Mesquite program (Maddison and Maddison, 2010). This was done
using a randomization test (to see if the observed number of changes on the
tree is less than 95% of the null values). The 10,000 reshufflings of the
(to make the maximum-likelihood NNIs search
more exhaustive)]. These parameters can produce slight increases in accuracy.
To estimate the reability of each split in the tree, FastTree uses a Shimodaire-
Hasegawa test on the three alternate topologies (NNIs) around that split
(Guindon et al., 2009). Phyllogenetic protein trees were visualized and edited
using Dendroscope software version 2.7.3 (Huson et al., 2007).
characters (5 different plants and other environments) allowed constructing a
character chart of parsimonious changes between the 6 characters assigned.
Relative abundance of microbial rhodopsins in different metagenomes
Frequency of rhodopsin blast hits with an e-value ≤ 10-5 was determined for 14
metagenomes from phyllosphere (5), marine (5), freshwater (1), hypersaline (1)
and soil (2) environments. Rhodopsin abundance was normalized with numbers
of rplA, rplC, rplD, rpoA, rpoB, and rspJ genes (Frank and Sorensen, 2011)
(blast hits with an e-value ≤ 1e-20) according to (Yutin et al., 2007; Howard et
Metagenomic datasets used for comparison (Fig. 2)
Freshwater: GS020, Lake Gatun, Panama (MG_RAST accession: 4441590.3)
Hypersaline: GS033, Punta Cormorant hypersaline lagoon, Galapagos (MG-
RAST accession: 4441599.3)
Open Sea: GS000a, Sargasso Station 11 (MG-RAST accession: 4441570.3)
and GS000b, Sargasso Station 13 (MG-RAST accession: 4441573.3)
Estuary: Monterey Bay (MG-RAST accession: 4443712.3)
Whale Fall: Whale fall Bone (MG-RAST accession: 4441619.3)
Forest Soil: Luquillo experimental forest soil, Puerto Rico (MG-RAST
accession: 4446153.3) and Waseca farm soil (MG-RAST accession:
Soybean: SRA accession: SRX008324
Absorbance spectra of tamarisk leaves and phyllosphere wash (Fig. 5)
Phyllosphere absorbance was calculated as the difference between two
measurements of reflectance spectra (intact tamarisk leaves and phosphate
buffered saline washed, sonicated leaves), obtained at room temperature with a
Labsphere DRA-CA-30I diffuse reflectance accessory. Leaves were organized
in high density on a slide and covered with another slide. Two empty slides
were used as a blank. Measurements were performed on 4 different leaf
samples from different dates. For chlorophylls absorbance, tamarisk leaves
were grinded with acetone 90% and filtered through GFF filters. The extraction
was measured using a Cary 100 spectrophotometer.
We thank Robert Edgar for his help and discussions over issues of short
sequences alignment and Eli Geffen for his help with statistics. Our gratitude to
the staff and scientists of the Josephine Bay Paul Center in Comparative
Molecular Biology and Evolution at the Marine Biology Laboratory (Woods Hole,
MA, USA) for their help. This work was supported in part by a grant from
Bridging the Rift Foundation (O.B. & S.B.), Israel Science Foundation grant
1203/06 (O.B.), the Gruss-Lipper Family Foundation at MBL (O.M.F., S.B. &
A.F.P.), a US-Israel Binational Science Foundation grant 2006324 (S.B.), and
DOE National Institutes of Health Grant R37GM27750, Department of Energy
Grant DE-FG02-07ER15867, and endowed chair AU-0009 from the Robert A.
Welch Foundation (J.L.S.).
Figure 1. A phylogenetic tree of rhodopsin-deduced amino acid sequences
from the phyllospheres of tamarisk, rice, soybean, Arabidopsis and clover.
Following alignment computation (see methods), a FastTree version 2.1.1 was
used for the calculation of the approximately-maximum-likelihood phylogenetic
tree using settings for high accuracy. Bootstraps above 60% are shown as
black circles at the junctions. PR- proteorhodopsins, HR- halorhodopsins, BR-
bacteriorhodopsins, SRI- sensory rhodopsins-I, SRII- sensory rhodopsins-II.
Figure 2. Relative abundance of microbial rhodopsins in different
metagenomes. MG_RAST (Meyer et al., 2008) accession numbers of the
different datasets can be found in the methods section. Abundance was
normalized relative to the numbers of rplA, rplC, rplD, rpoA, rpoB, and rspJ
genes (Frank and Sorensen, 2011) in each environment.
Figure 3. Sensory rhodopsins and proton pumps in different environments.
Proportions of sensory rhodopsins and rhodopsin proton pumps were calculated
only from reads containing the region surrounding the proton acceptor and
donor carboxylates at helix C (bacteriorhodopsin positions 85 and 96,
respectively); Sargasso Sea (Spudich, 2006) (n=732), tamarisk (n=13), soybean
(n=31), rice (n=8), Arabidopsis (n=4), and clover (n=7).
Figure 4. Protein alignment of phyllosphere rhodopsins. Amino acid position
105 is marked with green or blue backgrounds according to the predicted
absorption spectra of the rhodopsin pigments. Only the vicinity of amino acid
105 is shown. Examples from confirmed green absorbing proteorhodopsins
eBAC31A08 (Béjà et al., 2000), Dokdonia MED134 (Gómez-Consarnau et al.,
2007) and confirmed blue absorbing proteorhodopsins PAL-E6 (Béjà et al.,
2001), eBAC49C08 (Sabehi et al., 2005) are shown for reference at the top left
corner. Names of rhodopsins from the soybean phyllosphere start with SRR and
from the tamarisk start with GDOVJJ. Only a subset of the phyllosphere
rhodopsins is shown. See files S1, S2, S3, S4, and S5 for more variations.
Figure 5. Absorbance spectra of tamarisk leaves and phyllosphere wash.
Absorbance of tamarisk chlorophylls (acetone extract) and of phyllosphere leaf
buffer-wash are shown. Chlorophyll absorbance is shown for illustrative
purposes only; note the different scales used.
Atamna-Ismaeel, N., Sabehi, G., Sharon, I., Witzel, K.-P., Labrenz, M., Jürgens, K. et
al. (2008) Widespread distribution of proteorhodopsins in freshwater and 285
brackish ecosystems. ISME J 2: 656-662.
Balashov, S.P., Imasheva, E.S., Boichenko, V.A., Anton, J., Wang, J.M., and Lanyi,
J.K. (2005) Xanthorhodopsin: a proton pump with a light-harvesting carotenoid
antenna. Science 309: 2061-2064.
Ballaré, C.L., Scopel, A.L., and Sánchez, R.A. (1990) Far-red radiation reflected from 290
adjacent leaves: an early signal of competition in plant canopies. Science 247:
Beattie, G.A., and Lindow, S.E. (1999) Bacterial colonization of leaves: a spectrum of
strategies. Phytopathol 89: 353-359.
Béjà, O., Spudich, E.N., Spudich, J.L., Leclerc, M., and DeLong, E.F. (2001) 295
Proteorhodopsin phototrophy in the ocean. Nature 411: 786-789.
Béjà, O., Aravind, L., Koonin, E.V., Suzuki, M.T., Hadd, A., Nguyen, L.P. et al. (2000)
Bacterial rhodopsin: evidence for a new type of phototrophy in the sea. Science
de la Torre, J.R., Christianson, L., Béjà, O., Suzuki, M.T., Karl, D., Heidelberg, J.F., 300
and DeLong, E.F. (2003) Proteorhodopsin genes are widely distributed among
divergent bacterial taxa. Proc Natl Acad Sci U S A 100: 12830-12835.
Delmotte, N., Knief, C., Chaffron, S., Innerebner, G., Roschitzki, B., Schlapbach, R. et
al. (2009) Community proteogenomics reveals insights into the physiology of
phyllosphere bacteria. Proc Natl Acad Sci U S A 106: 16428-16433. 305
Field, C.B., Behrenfeld, M.J., Randerson, J.T., and Falkowski, P. (1998) Primary
production of the biosphere: integrating terrestrial and oceanic components.
Science 281: 237-240.
Folta, K.M., and Maruhnich, S.A. (2007) Green light: a signal to slow down or stop. J
Exp Bot 58: 3099-3111. 310
Frank, J.A., and Sorensen, S.J. (2011) Quantitative metagenomic analyses based on
average genome size normalization. Appl Environ Microbiol 77: 2513-2521.
Frigaard, N.-U., Martinez, A., Mincer, T.J., and DeLong, E.F. (2006) Proteorhodopsin
lateral gene transfer between marine planktonic Bacteria and Archaea. Nature
439: 847-850. 315
Giovannoni, S.J., Bibbs, L., Cho, J.-C., Stapels, M.D., Desiderio, R., Vergin, K.L. et al.
(2005) Proteorhodopsin in the ubiquitous marine bacterium SAR11. Nature 438:
Gómez-Consarnau, L., González, J.M., Coll-Lladó, M., Gourdon, P., Pascher, T.,
Neutze, R. et al. (2007) Light stimulates growth of proteorhodopsin-containing 320
marine Flavobacteria. Nature 445: 210-213.
Gómez-Consarnau, L., Akram, N., Lindell, K., Pedersen, A., Neutze, R., Milton, D.L. et
al. (2010) Proteorhodopsin phototrophy confers enhanced survival of marine
bacteria during starvation. PLoS Biol 8: e1000358.
Guindon, S., Delsuc, F., Dufayard, J.F., and Gascuel, O. (2009) Estimating maximum 325
likelihood phylogenies with PhyML. Methods Mol Biol 537: 113–137.
Howard, E.C., Sun, S., Biers, E.J., and Moran, M.A. (2008) Abundant and diverse
bacteria involved in DMSP degradation in marine surface waters. Environ
Microbiol 10: 2397-2410.
Huson, D.H., Richter, D.C., Rausch, C., Dezulian, T., Franz, M., and Rupp, R. (2007) 330
Dendroscope: An interactive viewer for large phylogenetic trees. BMC
Bioinformatics 8: 460.
Kemena, C., and Notredame, C. (2009) Upcoming challenges for multiple sequence
alignment methods in the high-throughput era. Bioinformatics 25: 2455-2465.
Koh, E.Y., Atamna-Ismaeel, N., Martin, A., Cowie, R.O., Beja, O., Davy, S.K. et al. 335
(2010) Proteorhodopsin-bearing bacteria in Antarctic sea ice. Appl Environ
Microbiol 76: 5918-5925.
Kutschera, U. (2007) Plant-associated methylobacteria as co-evolved phytosymbionts: a
hypothesis. Plant Signal Behav 2: 74-78.
Lanyi, J.K., and Balashov, S.P. (2008) Xanthorhodopsin: a bacteriorhodopsin-like 340
proton pump with a carotenoid antenna. Biochim Biophys Acta 1777: 684-688.
Lindow, S.E., and Brandl, M.T. (2003) Microbiology of the phyllosphere. Appl Environ
Microbiol 69: 1875-1883.
Maddison, W.P., and Maddison, D.R. (2010) Mesquite: a modular system for
evolutionary analysis. Version 2.73 http://mesquiteproject.org. 345
Man, D., Wang, W., Sabehi, G., Aravind, L., Post, A.F., Massana, R. et al. (2003)
Diversification and spectral tuning in marine proteorhodopsins. EMBO J. 22:
Meyer, F., Paarmann, D., D'Souza, M., Olson, R., Glass, E.M., Kubal, M. et al. (2008)
The metagenomics RAST server - a public resource for the automatic 350
phylogenetic and functional analysis of metagenomes. BMC Bioinformatics 9:
Muller, J., Creevey, C.J., Thompson, J.D., Arendt, D., and Bork, P. (2010) AQUA:
automated quality improvement for multiple sequence alignments.
Bioinformatics 26: 263-265. 355
Oesterhelt, D., and Stoeckenius, W. (1971) Rhodopsin-like protein from the purple
membrane of Halobacterium halobium. Nat New Biol 233: 149-152.