Expression of two [Fe]-hydrogenases in
Chlamydomonas reinhardtii
under anaerobic conditions
Marc Forestier
1,
*, Paul King
1
, Liping Zhang
1
, Matthew Posewitz
1
, Sarah Schwarzer
2
, Thomas Happe
2
,
Maria L. Ghirardi
1
and Michael Seibert
1
1
National Renewable Energy Laboratory, Golden, CO USA;
2
Ruhr-Universitaet-Bochum, Lehrstuhl Biochemie der Pflanzen AG
Photobiotechnologie, Bochum, Germany
We have isolated and characterized a second [Fe]-hydro-
genase gene from the green alga, Chlamydomonas reinhardtii.
The HydA2 gene encodes a protein of 505 amino acids that
is 74% similar and 68% identical to the known HydA1
hydrogenase from C. reinhardtii. HydA2 contains all the
conserved residues and motifs found in the catalytic core of

2
have been found in many other green
algae [3,4], including Chlamydomonas reinhardtii.This
particular alga is capable of evolving H
2
gas in the dark
[5,6] or in the light, using H
2
O [7] or starch [8,9] as the source
of reductant. The reaction is catalyzed by a monomeric,
49 kDa reversible [Fe]-hydrogenase enzyme, which has been
isolated to purity by Happe and Naber [10].
Other [Fe]-hydrogenases, identified in a small group of
nonphotosynthetic anaerobic microbes (bacteria and pro-
tists) also catalyze either H
2
production or H
2
uptake in vivo
[11,12]. They play an important role in the anaerobic energy
metabolism of these organisms, mainly by reoxidizing
accumulated reducing equivalents. All [Fe]-hydrogenases
whose X-ray structures have been analyzed to date incor-
porate a [2Fe-2S] center bridged by a cysteine residue to a
[4Fe-4S] center at the catalytic site (the H-cluster). It is also
known that most [Fe]-hydrogenases contain additional
iron-sulfur centers that act as electron relays between carrier
molecules and the H-cluster [13,14]. However, the addi-
tional centers are absent in the green algal enzymes [15,16].
In addition, [Fe]-hydrogenases usually exhibit high specific

Research Institute, Batelle and Bechtel under contract number
DE-AC36–99G010337.
*Present address: Limnological Station, Plant Biology Department,
University of Zurich, Seestrasse 187, 8802 Kilchberg, Switzerland.
(Received 1 April 2003, revised 29 April 2003,
accepted 30 April 2003)
Eur. J. Biochem. 270, 2750–2758 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03656.x
physically, by purging algal cultures with a neutral gas, or
physiologically, by incubating algal cultures in sulfur-free
medium [20,21]. The latter condition prevents the turnover
of the D1 protein of photosystem II (PSII), causes the
partial inhibition of PSII, and leads to a reduction of
photosynthetic O
2
evolution. The partial inhibition of PSII
is sufficient to create anaerobic culture conditions and
induce H
2
-production activity, which lasts for 3–4 days
under continuous illumination.
In this paper, we characterize a second hydrogenase gene,
HydA2,inC. reinhardtii, and use the terminology estab-
lished by the international hydrogenase community [12] to
name both prokaryotic and eukaryotic hydrogenase genes.
The primary amino-acid sequence of HydA2 (a) contains all
the strictly preserved motifs present in the core, catalytic-site
sequences found in algal [Fe]-hydrogenases; (b) shows high
similarity to the [Fe]-hydrogenases from Clostridium pasteu-
rianum [22], Trichomonas vaginalis [23], and Desulfovibrio
vulgaris [24]; and (c) is 68% identical to C. reinhardtii

phyll per mL, and the harvested cells were anaerobically
induced as described previously [27] for 4–5 h. For sulfur-
deprivation experiments, cells grown as above were washed
with sulfur-free TAP medium and resuspended in the same
medium at a final chlorophyll concentration of 20 lgÆmL
)1
[20]. The cultures were then incubated in sealed glass bottles
under continuous fluorescent light for up to 4 days.
cDNA library screening
Two specific primers, BE5P1 (5¢-AACATCTTCAAGGA
GCGTGGCATC-3¢)andBE3P1(5¢-AGACAGCAGGA
GACTCACAATCAC-3¢), were used to amplify a C. rein-
hardtii expressed sequence tag (EST), BE337478, from a
strain 21gr cDNA library kindly provided by J. Davies,
Exelixis Inc., South San Francisco, CA, USA [17]. Clone
EST_26,wasthenusedtogenerateanapproximately
1200 bp NotI/EcoRI restriction fragment that was dig-
oxigenin-labeled and used as a probe for cDNA library
screening. The HydA2 clone was retrieved in pBluescript
SK(–), purified and submitted to the Iowa State University
Sequencing and DNA Synthesis Facility for sequencing.
Sequences were evaluated based on the chromatograms,
and they were assembled using the
WISCONSIN PACKAGE
V.10 software by Genetics Computer Group, Inc (San
Diego, CA, USA). Both strands of DNA were sequenced
independently, and the sequence data for HydA2 cDNA
has been deposited in GenBank under Accession No.
AY055756.
Isolation and sequencing of genomic

USA) on an Applied Biosystems 3730 automated
sequencer. The HydA2 gene sequence has been deposited
in GenBank under Accession no. AY090770.
Northern blot analysis
To obtain transcript hybridization signals that truly reflect
the aerobic state of the algae at t ¼ 0, it proved essential to
lyse noninduced algal cells as quickly as possible to avoid
the establishment of anaerobiosis in the dark by respiratory
O
2
consumption. Total RNA was isolated at different time
points from anaerobically induced samples using the SNAP
RNA Isolation Kit (Invitrogen, San Diego, CA, USA).
DNA was removed by treatment with RNase-free DNaseI
(0.013 units per lL). Ten micrograms of RNA were
separated by electrophoresis on denaturing 1.1% (w/v)
agarose, 0.22
M
formaldehyde gels and then blotted onto a
Nytran N
+
nylon membrane with 10· NaCl/Cit [28] as
the transfer buffer. Radiolabeled probes specific for either
HydA1 or HydA2 were generated using Rediprime random
primer labeling kits (Amersham, Piscataway, NJ, USA).
Denatured probes were hybridized to the membranes in
prehybridization buffer [6· NaCl/Cit buffer, 5 · Den-
hardt’s solution, 0.1% (w/v) SDS] overnight at 65 °C.
Ó FEBS 2003 Expression of a second [Fe]-hydrogenase in C. reinhardtii (Eur. J. Biochem. 270) 2751
Following hybridization, the membranes were washed and

The E. coli strain BL21(DE3)pLysS was transformed with
both constructs pET-HydA1 and pET-HydA2. Expression
was induced with 1 m
M
isopropyl thio-b-
D
-galactoside
(IPTG) at an A
600
¼ 0.3.
Antibody generation and immunoblot analysis
A region of low amino acid sequence homology between
HydA1 and HydA2 (the insert between motifs 2 and 3) was
screened for high antigenicity using two methods, Alpha
Diagnostic (San Antonio, Texas) and the JaMBW online
program ( />page/JaMBW). Both methods identified a 14-residue long
oligopeptide (VAE RLAHKVEEAAA) in HydA2 as a
possible candidate. The oligopeptide was synthesized by
Sigma Genosys (The Woodlands, TX, USA), coupled to the
keyhole limpet hemacyanin (KLH) protein carrier and
injected into rabbits to induce antibody generation. The
resulting serum was immunoaffinity purified and tested for
reaction against HydA1 and HydA2 overexpressed in
E. coli and against HydA2 in anaerobically induced algal
extracts.
Algal cells were harvested during the mid-logarithmic
phase by centrifugation at 2000 g for2min.Pelletswere
resuspended in 50 m
M
Tris/HCl pH 8.5 with 20 m

SWISS
-
MODEL
[31].
CLUSTAL W
alignments of the
predicted processed HydA2 sequence to CpI were used to
manually optimize backbone threading. Final versions of
the models were submitted to
SWISS
-
MODEL
for validation.
The resulting homology structures were further refined by
energy minimization with
GROMOS
.Calculatedrmsdvalues
between the resulting HydA2 model and CpI include all
shared backbone atoms.
Results and discussion
Genetic analysis of
HydA2
The first [Fe]-hydrogenase from C. reinhardtii, HydA1,was
recently cloned (GenBank, accession numbers CRE012098,
AY055755, AF289201), and was shown to encode a
functional enzyme [16]. The deduced amino acid sequence
of the HydA1 catalytic site was further utilized in a
BLAST
search and revealed a close match to an expressed sequence
tag (EST) from C. reinhardtii, BE33478. The EST was

nonalgal [Fe] hydrogenases. The three complete motifs
found in the catalytic H-cluster of [Fe]-hydrogenases, motif
1 (PMFTSCCPxW), motif 2 (MPCxxKxxExxR) and motif
3 (FxExMACxGxCV), have also been found in the algal
sequences and are marked in Fig. 1. Each contains cysteine
residues (*) that ligate the catalytic [4Fe-4S] center. The
cysteine residue in motif 3 (#) bridges the [4Fe-4S] to the
[2Fe-2S] center of the active H-cluster. Comparisons of
C. reinhardtii HydA2 with the [Fe]-hydrogenases from
C. reinhardtii and S. obliquus HydA1 show 68% and 61%
identity, respectively.
Algal [Fe]-hydrogenases also share the following charac-
teristic features (Fig. 1): (a) an amino acid residue insertion
(eight residues in both S. obliquus and C. reinhardtii HydA1
and HydA2) upstream of the H-cluster motif 1 and (b) a
second amino acid insertion (16 residues in C. fusca,
16 residues in S. obliquus HydA1 and HydA2, 45 residues
in C. reinhardtii HydA1, and 54 residues in C. reinhardtii
HydA2) between H-cluster motifs 2 and 3. The biological
implications of these unique features are not known at
present, but, given their ubiquity in all cloned algal
hydrogenases, they might be critical to specific functional
or structural peculiarities of the algal enzymes.
In order to determine that HydA1 and HydA2 are
encoded by distinct nuclear genes, C. reinhardtii genomic
DNA was purified, digested with PstI and probed sepa-
rately with HydA1-andHydA2-specific probes (Southern
not shown). A single PstI site is present within the HydA1
genomic sequence where the probe hybridizes. The same
PstI restriction site, however, is absent from the HydA2

hydrogenases, and those highlighted in grey show similarity between at least four of the sequences. Cr, C. reinhardtii HydA1 [16] and HydA2 (this
work); So, S. obliquus HydA1 [15] and HydA2 [19]; and Cp, C. pasteurianum HydI [22].
Ó FEBS 2003 Expression of a second [Fe]-hydrogenase in C. reinhardtii (Eur. J. Biochem. 270) 2753
cleavage site recognized by chloroplast stromal processing
peptidases is not known [38], but the motif VXA has been
identified near the transit peptide cleavage site in a number
of Chlamydomonas chloroplast-targeted proteins [39]. In
HydA1, the transit peptide is cleaved one amino acid
downstream from the sequence VACAA at position 56 of
the nascent peptide [16]. By analogy, a VXA motif cleavage
site is located at amino acid 61 of the HydA2 ORF (VAA),
suggesting a cleavage site after residue 63. This cleavage site
was also identified using the
CHLOROP
program [40].
Additional studies will be required to determine whether
HydA2 is indeed processed proteolytically, and if so, where
cleavage actually occurs in vivo.
The isolation and sequencing of the complete HydA2
gene was done using a HydA2-specific probe (Materials and
methods) to screen a BAC library of cloned C. reinhardtii
genomic DNA. Following identification of four positive
clones, the HydA2 region of a single BAC clone was
amplified by PCR and directly sequenced (GenBank
Accession number AY090770). The length of the complete
HydA2 gene (from the 5¢-UTR to the 3¢-UTR minus the
poly A tail) is 4.62 kb and consists of eight exons and nine
introns (average intron size, 150 bp). The structural
arrangement of HydA2 is thus more complex than HydA1,
which contains only seven introns [16]. Interestingly, the

occurred more slowly, reaching steady-state after
Fig. 2. Effects of anaerobiosis on hydrogenase transcription, enzyme function, and transcript stability. (A) and (B) HydA1 (d)andHydA2 (s)
transcript accumulation following incubation under dark, anaerobic conditions. Transcript levels were measured in C. reinhardtii cells grown on
either TAP (A) or BS (B) medium and normalized to the amount of 23S rRNA of the respective sample. Simultaneous measurements of H
2
photoproduction activity were performed with the same cultures. (C) Northern blots of cultures grown on either TAP or BS medium after
incubation under anaerobic conditions at 4 °C overnight (O/N), followed by exposure to O
2
for 15 min (+O
2
).TherespectiveratesofH
2
photoproduction (lmoles H
2
per mg chlorphyll per h) are shown above each blot.
2754 M. Forestier et al.(Eur. J. Biochem. 270) Ó FEBS 2003
240–300 min of anaerobic treatment. There were no
observed differences in the rate of accumulation of the
two transcripts under these conditions. Induction for longer
periods of time did not further change the levels of the two
transcripts (not shown). The induced rates of H
2
photopro-
duction vary in different experiments and reach steady-state
at slightly different time points, possibly due to the different
levels of anaerobic induction achieved with different
cultures. However, the onset of H
2
-photoproduction acti-
vity in BS-grown cells was consistently later than for TAP-

shows that H
2
-photoproduc-
tion activity in both cultures was completely lost. However,
the transcripts in the photoautotrophic (BS) culture were
virtually undetectable after the cells were exposed to O
2
,
whereas the transcripts in the photoheterotrophic (TAP)
culture were still present. Together, these data demonstrate
the lower stability of the HydA2 transcript on exposure to
O
2
under photoautotrophic conditions (Fig. 2C) and indi-
cate that the levels of the two transcripts may be modulated
by other factors in addition to O
2
.
Gene expression during sulfur deprivation
The expression of HydA1 and HydA2 transcripts was also
studied under sulfur-deprivation conditions. In the absence
of sulfur, the rates of photosynthetic O
2
evolution drop
below those of O
2
consumption by respiration after about
24 h of incubation. As a consequence, sealed cultures of
green algae become anaerobic in the light [41,42]. The
HydA1 transcript was detected after 24 h (Fig. 3, bottom)

sulfur-deprived conditions. (Top) Hydrogen-production activity of
C. reinhardtii cultures incubated in sulfur-deprived TAP medium for
1–3 days in a sealed photobioreactor. The hydrogenase activity at
t ¼ 0 was 0. (Bottom) RNA was isolated from cells subjected to sulfur-
deprivation conditions as above, and hybridized as described in the
Material and methods. The data, showing photographs of typical
blots, are from one representative experiment.
Fig. 4. Heterologous and homologous expression of the HydA2 protein.
Western blots of IPTG-induced E. coli total protein extracts over-
expressing, HydA1 or HydA2 (lanes 1 and 2, respectively) and partially
purified C. reinhardtii cell extracts, either noninduced (lane 3) or
anaerobically induced (lane 4). All blots were probed with the HydA2-
specific antibody. The presence of a band in lane 1 above 49 kDa
represents a nonspecific response seen only in the overexposed blot. It
is not HydA1, as the band is also present in E. coli that had not been
induced by IPTG (data not shown).
Ó FEBS 2003 Expression of a second [Fe]-hydrogenase in C. reinhardtii (Eur. J. Biochem. 270) 2755
major band that migrated at the predicted mass of 49 kDa
(data not shown). Whereas recombinant HydA2 was detec-
ted by HydA2 antibody on the Western blot (Fig. 4, lane 2),
the same antibody did not detect recombinant HydA1
(Fig. 4, lane 1). Thus, the antibody detects the HydA2
protein specifically. As expected, extracts of aerobically
grown C. reinhardtii cells showed no HydA2 reactive protein
(Fig. 4, lane 3). However, the antibody did recognize a
protein in extracts of anaerobically induced cells, which
comigrated with the recombinant HydA2 protein (Fig. 4,
lane 4). These results confirm that HydA2 is expressed in
C. reinhardtii and accumulates after anaerobic induction.
There is precedence for multiple [Fe]-hydrogenases in

a much higher degree of conservation (75% identity, 90%
similarity). In comparison, the mature forms of the HydA1
and HydA2 peptide sequences share 81% identity and 74%
similarity with each other, and in addition, 91% identity is
evident for the active site and H
2
-channel motifs.
In order to visualize the significance of the primary
sequence homology, a theoretical model of HydA2 was
generated from the solved X-ray structure of CpI. As shown
in Fig. 5, HydA2 (as does HydA1, data not shown) exhibits
a high degree of structural similarity to CpI (rmsd of
1.55 A
˚
, 1480 backbone atoms). The predicted locations of
HydA2 (and HydA1, data not shown) peptide sequence
motifs that represent the active site and H
2
-channel match
closely to the positions of the corresponding residues in CpI.
The rmsd between the HydA2 and CpI core regions is
0.74 A
˚
over 1336 shared atoms (334 Ca atoms). This
information corroborates our assignment of a hydrogenase
function to HydA2, provides us with a model to compare
the catalytic functions of isolated HydA1 and HydA2 in the
future, and serves as a guide to future site-directed
mutagenesis studies of the two algal hydrogenase proteins.
In conclusion, we have cloned and sequenced a second

software (Accelrys).
2756 M. Forestier et al.(Eur. J. Biochem. 270) Ó FEBS 2003
for the HydA2 clone. MF would like to acknowledge a grant to
prospective researchers from the Swiss National Science Foundation.
MS and MLG were supported by the U.S. DOE Hydrogen program.
TH thanks the Japanese New Energy and Industrial Technology
Development Organization (NEDO-project no. 01GB1) for financial
support.
References
1. Gaffron, H. (1939) Reduction of carbon dioxide with molecular
hydrogen in green algae. Nature 143, 204–205.
2. Gaffron, H. & Rubin, J. (1942) Fermentative and photochemical
production of hydrogen in algae. J. General Physiol. 26, 219–240.
3. Weaver, P.F., Lien, S. & Seibert, M. (1980) Photobiological pro-
duction of hydrogen. Solar Energy 24, 3–45.
4. Brand, J.J., Wright, J.N. & Lien, S. (1989) Hydrogen production
by eukaryotic algae. Biotechnol. Bioeng. 33, 1482–1488.
5. Miura, Y., Yagi, K., Nakano, Y. & Miyamoto, K. (1981)
Requirement of oxygen for dark hydrogen evolution by a green
alga, Chlamydomonas reinhardtii. J. Ferment. Technol. 59,441–
446.
6. Miura, Y., Yagi, K., Shoga, M. & Miyamoto, K. (1982) Hydrogen
production by a green alga, Chlamydomonas reinhardtii,inan
alternating light/dark cycle. Biotechnol. Bioeng. 24, 1555–1563.
7. Greenbaum, E., Guillard, R.R.L. & Sunda, W.G. (1983) Hydro-
gen and oxygen photoproduction by marine algae. Photochem.
Photobiol. 37, 649–655.
8. Gfeller, R.P. & Gibbs, M. (1984) Fermentative metabolism of
Chlamydomonas reinhardtii.1.Analysisoffermentativeproducts
from starch in dark and light. Plant Physiol. 75, 212–218.

molecular characterization of the [Fe]-hydrogenase from the uni-
cellular green alga Chlorella fusca. Biochim. Biophys. Acta 1576,
330–334.
19. Wu
¨
nschiers,R.,Stangier,K.,Senger,H.&Schulz,R.(2001)
Molecular evidence for a Fe-hydrogenase in the green alga Sce-
nedesmus obliquus. Current Microbiol. 42, 353–360.
20. Melis, A., Zhang, L., Forestier, M., Ghirardi, M.L. & Seibert, M.
(2000) Sustained photobiological hydrogen gas production upon
reversible inactivation of oxygen evolution in the green alga
Chlamydomonas reinhardtii. Plant Physiol. 122, 127–135.
21. Ghirardi, M.L., Zhang, L., Lee, J.W., Flynn, T., Seibert, M.,
Greenbaum, E. & Melis, A. (2000) Microalgae: a green source of
renewable H
2
. Trends Biotechnol. 18, 506–511.
22. Meyer, J. & Gagnon, J. (1991) Primary structure of hydrogenase I
from Clostridium pasteurianum. Biochemistry 30, 9697–9704.
23. Bui, E.T.N. & Johnson, P.J. (1996) Identification and character-
ization of [Fe]-hydrogenases in the hydrogenosome of Tricho-
monas vaginalis. Mol. Biochem. Parasitol. 76, 305–310.
24. Nicolet, Y., Piras, C., Legrand, P., Hatchikian, C.E. & Fontecilla-
Camps, J.C. (1999) Desulfovibrio desulfuricans iron hydrogenase:
the structure shows unusual coordination to an active site Fe
binuclear center. Structure 7, 13–23.
25. Harris, E.H. (1989) The Chlamydomonas Sourcebook.Academic
Press, Inc, San Diego, CA, USA.
26. Flynn, T., Ghirardi, M.L. & Seibert, M. (2002) Accumulation of
O

34. Winkler, M., Hemschemeier, A., Gotor, C., Melis, A. & Happe, T.
(2002) [Fe]-hydrogenases in green algae: photofermentation and
hydrogen evolution under sulfur deprivation. Int. J. Hydrogen
Energy 27, 1431–1439.
35. Melis, A. & Happe, T. (2001) Hydrogen production. Green algae
as a source of energy. Plant Physiol. 127, 740–748.
36. Keegstra, K., Olsen, J.J. & Theg, S.M. (1989) Chloroplastic pre-
cursors and their transport across the envelope membranes. Ann.
Rev. Plant Physiol. Plant Mol. Biol. 40, 415–439.
37. Bruce, B.D. (2001) The paradox of plastid transit peptides: con-
servation of function despite divergence in primary structure.
Biochim. Biophys. Acta 1541, 2–21.
38. Richter, S. & Lamppa, G.K. (1998) A chloroplast processing
_enzyme functions as the general stromal processing peptidase.
Proc.NatlAcad.Sci.USA95, 7463–7468.
39. Franzen, L.G., Rochaix, J.D. & von Heijne, G. (1990) Chloroplast
transit peptides from the green alga Chlamydomonas reinhardtii
share features with both mitochondrial and higher plant chloro-
plast presequences. FEBS Lett. 260, 165–168.
40. Emanuelsson, O., Nielsen, H. & von Heijne, G. (1999) ChloroP, a
neural network-based method for predicting chloroplast transit
peptides and their cleavage sites. Protein Sci. 8, 978–984.
Ó FEBS 2003 Expression of a second [Fe]-hydrogenase in C. reinhardtii (Eur. J. Biochem. 270) 2757
41. Kosourov, S., Tsygankov, A., Seibert, M. & Ghirardi, M.L. (2002)
Sustained hydrogen photoproduction by Chlamydomonas
reinhardtii: Effects of culture parameters. Biotechnol. Bioeng. 78,
731–740.
42. Zhang, L., Happe, T. & Melis, A. (2002) Biochemical and
morphological characterization of sulfur-deprived and H
2