Báo cáo khoa học: The hemoglobin genes of Drosophila ppt

The hemoglobin genes of Drosophila
Thorsten Burmester
1
, Joachim Storf
1
, Anne Hasenja
¨
ger
2
, Sabine Klawitter
2
and Thomas Hankeln
2
1 Institute of Zoology, University of Mainz, Germany
2 Institute of Molecular Genetics, University of Mainz, Germany
(Hemo-)globins (Hbs) are small globular proteins usu-
ally consisting of a core of about 150 amino acids,
which typically comprises eight a-helical segments
(named A–H) and displays a characteristic 3-over-3
a-helical sandwich structure [1,2]. Hbs include an iron-
containing heme (Fe
2+
-protoporphyrin IX), by which
they reversibly bind gaseous ligands. Most Hbs are
considered respiratory proteins because of their ability
to transport or to store O
2
, thus enhancing the avail-
ability of O
2
to the respiratory chain of the mitochon-

dria [3]. Hbs are widespread among metazoan taxa
[4–6] and it is not surprising that Hbs are among the
best-investigated proteins in biological and biochemical
sciences [1,2]. Recent studies have demonstrated that
some Hbs may also detoxify NO (and possibly other
noxious reactive nitrogen and oxygen species [RNS
and ROS]) and or serve as O
2
sensing molecule [7–9].
For a long time Hbs were essentially unknown in
most insects. Respiratory proteins had been regarded
unnecessary in this taxon because of the well-devel-
oped tracheal system, which enables an efficient diffu-
sion of O
2
from the atmosphere to the inner organs
[10]. Notable exceptions had been only some basal
insects that possess hemocyanin as O
2
transport pro-
tein in the hemolymph, which they inherited from the
crustacean ancestor, but has been lost in the evolution
Keywords
Gene duplication; globin; insects; oxygen;
phylogeny
Correspondence
T. Burmester, Institute of Zoology,
University of Mainz, D-55099 Mainz,
Germany
Fax: +49 6131 39 24652

Tel: +49 6131 39 24477
E-mail:
Note
The nucleotide sequences reported in this
paper have been submitted to the
GenBankTM ⁄ EMBL databases with the
accession numbers AM086021 (D. virilis
glob1 gene), AM086022 (D. virilis glob1
cDNA), AM086023 (D. pseudoobscura glob1
cDNA) and AM086024 (D. melanogaster
glob2 cDNA).
(Received 13 October 2005, revised 18
November 2005, accepted 23 November
2005)
doi:10.1111/j.1742-4658.2005.05073.x
We recently reported the unprecedented occurrence of a hemoglobin gene
(glob1) in the fruitfly Drosophila melanogaster. Here we investigate the
structure and evolution of the glob1 gene in other Drosophila species. We
cloned and sequenced glob1 genes and cDNA from D. pseudoobscura and
D. virilis, and identified the glob1 gene sequences of D. simulans, D. yak-
uba, D. erecta, D. ananassae, D. mojavensis and D. grimshawi in the data-
bases. Gene structure (introns in helix positions D7.0 and G7.0), gene
synteny and sequence of glob1 are highly conserved, with high ds ⁄ dn ratios
indicating strong purifying selection. The data suggest an important role of
the glob1 protein in Drosophila, which may be the control of oxygen flow
from the tracheal system. Furthermore, we identified two additional globin
genes (glob2 and glob3) in the Drosophilidae. Although the sequences are
highly derived, the amino acids required for heme- and oxygen-binding are
conserved. In contrast to other known insect globin, the glob2 and glob3
genes harbour both globin-typical introns at positions B12.2 and G7.0.

Both genes are conserved in various drosophilid species, but only expres-
sion of glob2 could be demonstrated by western blotting and RT-PCR.
Phylogenetic analyses show that the clade leading to glob2 and glob3,
which are sistergroups, diverged first in the evolution of dipteran globins.
glob1 is closely related to the intracellular hemoglobin of the botfly Gastero-
philus intestinalis, and the extracellular hemoglobins from the chironomid
midges derive from this clade.
Abbreviations
AIC, Akaike information criterion; EPO, erythropoietin; Hb, hemoglobin; HRE, hypoxia responsive sequence elements; PIP, percent identity
plot.
468 FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS
of higher neopteran insects [11]. Respiratory proteins
belonging to the Hb-type were considered being con-
fined to a few insect species that are adapted to hypoxic
environments [6]. According to our present knowledge,
the chironomid midges (Diptera:Nematocera) are the
only insects that possess true extracellular Hbs in their
circulatory system for O
2
transport [12,13]. Some
aquatic Hemiptera and the larvae of the horse botfly
Gasterophilus intestinalis harbour intracellular Hbs,
which probably carry out myoglobin (Mb)-like O
2
storage-functions [6,14,15]. These cellular Hbs reach
concentrations in the millimolar range and thus may
easily be identified due to the red colour of the Hb-
containing organs.
Recently, however, we showed that a true Hb gene
(glob1 or Hb1) is present and expressed in the fruitfly

Drosophila melanogaster [16,17]. This finding was
unprecedented because at the first glance D. melano-
gaster is unlikely to face hypoxic conditions during its
life cycle. The crystal structure of D. melanogaster
glob1 identified a protein that displays a typical globin
fold and shows conservation of the residues important
for O
2
binding [18]. Although exhibiting a hexacoordi-
nated binding scheme at the Fe
2+
ion in the deoxygen-
ated state, the O
2
-binding kinetics of glob1 are similar
to those of other insect intracellular Hbs (P
50
¼ 0.14
Torr) [17]. In larval and adult Drosophila, glob1 is
mainly expressed in tracheal cells and the fat body.
These data indicate that Hbs may be in fact involved
in O
2
metabolism and suggest that Hb genes may be
much more widely distributed in insects than previ-
ously thought. Here we investigate the occurrence and
evolution of glob1 in other Drosophila species (Fig. 1).
Moreover, we identified two additional putative globin
genes of Drosophila, which we named glob2 and glob3.
Results

The Drosophila glob1 genes
The D. melanogaster glob1 nucleotide and amino-acid
sequences were used to search the databases of genom-
ic DNA sequences available from various Drosophili-
dae, as they are available at the EMBL ⁄ GenBank
() or at FlyBase (http://
flybase.net/), by the blast algorithm [19]. We identi-
fied the full length glob1 coding sequences from the
genomes of D. yakuba, D. erecta, D. ananassae,
D. pseudoobscura, D. virilis, D. mojavensis and D. grim-
shawi, as well as a partial sequence from D. simulans
that covers only the first coding exon (see Supplement-
ary material). The D. virilis glob1 sequence was also
independently obtained by screening a genomic DNA
library with the D. melanogaster glob1 coding sequence
as probe. The D. virilis glob1 sequence (Acc. No.
AM086021), which covers the complete glob1 gene,
comprises 6265 bp. It differs from the genomic
sequence at FlyBase (contig 462) at 20 positions and
displays four short indels (1–26 bp). In addition, the
cDNA sequences, including their putative 5¢ ends,
from D. virilis and D. pseudoobscura (Acc. Nos.
AM086022 and AM086023) were obtained by
RT-PCR and subsequent RACE experiments. We
identified a single splice variant for each D. virilis and
D. pseudoobscura glob1, which contains three coding
exons and a single 5¢ noncoding exon (Fig. 2). The
glob1 mRNAs of D. pseudoobscura and D. virilis cor-
respond to transcript ‘A’ of the D. melanogaster glob1
gene, which consists of exons 1, 4, 5 and 6 [17]. In

addition, the in silico analysis of the gene predicts the
presence of a second splice variant in D. pseudoobscura,
which corresponds to transcript ‘C’ of D. melanogaster
glob1 (Exons 1, 3, 4, 5 and 6). However, the corres-
ponding cDNA was not recovered by RT-PCR.
The glob1 gene plus the upstream and downstream
genomic regions were scanned for the presence of puta-
tive hypoxia responsive sequence elements (HREs)
(Fig. 2). HREs in hypoxia-regulated genes are defined
by the binding motif of the hypoxia-inducible transcrip-
tion factor HIF-1 (5¢-RCGTG-3¢) [20]. Typically, an
HRE includes two HIF-1 motifs (in direct or inverted
orientation) or one HIF-1 motif combined with an ery-
thropoietin (EPO) box or a HIF ancillary sequence
[21]. Six putative HREs are present in the D. melano-
gaster and D. virilis glob1 genes, while the D . pseudo-
obscura glob1 contains five such motifs. Interestingly,
Fig. 1. Evolutionary history of Drosophilidae. The phylogenetic rela-
tionships of drosophilid species were taken from Russo et al. [52],
divergence times are indicated as calculated by Tamura et al. [24].
T. Burmester et al. Drosophila hemoglobins
FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS 469
the HREs are located in similar positions, although
motif combinations are not always conserved.
The degree of sequence conservation ± 4 kb
upstream and downstream of the glob1 genes were
evaluated with PipMaker, using the D. melanogaster
glob1 gene region as reference (Fig. 3A). The putative
glob1 coding exons (see below) are conserved and
clearly discernable in all Drosophila species. Outside

the melanogaster subgroup (D. simulans, D. yakuba,
D. erecta), hardly any other region displays sequence
similarities. The only exceptions are the noncoding
exon 1, which is transcribed in D. melanogaster,
D. pseudoobscura and D. virilis, and a region within
the third intron, about 2 kb 5¢ of the translation start
codon, which may harbour regulatory sequences. The
D. melanogaster glob1 gene is located on the right arm
of chromosome 3 at band 89A8 between the genes
CG14877, which codes for a putative natriuretic pep-
tide receptor, and CG31292, which has no putative
ortholog outside the Drosophila species (see Supple-
mentary material). In the other species of the melano-
gaster subgroup (Fig. 1) for which the genome
locations are available (D. simulans and D. yakuba),
the glob1 gene is positioned on the same chromosome.
The D. pseudoobscura glob1 gene is located on chromo-
some 2, which is orthologous to chromosome 3R of
the melanogaster subgroup [22,23]. Gene synteny ana-
lyses of 10 neighbouring genes (five on each side)
reveals the conservation of gene locations and orienta-
tions in this region between D. melanogaster and
D. pseudoobscura, which diverged about 55 million
years ago (Fig. 1). Because of the absence of data from
other genomes, gene synteny analyses was restricted to
the two neighbouring gene in these species. This shows
the presence of both CG14877 and CG31292 adjacent
to the glob1 in all Drosophila species (see Supplement-
ary material). Gene orientations were found to be con-
served in all Drosophila genomes except D. ananassae,

in which the putative CG31292 ortholog was found in
reversed orientation. However, this may also reflect an
assembly error.
The Drosophila glob1 proteins and gene
evolution
Gene predictions and comparisons with the available
cDNA sequences show that Drosophila glob1 genes
consistently harbour two introns, which are located at
positions D7.0 (i.e. between codons 6 and 7 of globin
helix D) and G7.0 (Fig. 4). The introns are short, vary-
ing from 64 to 73 bp for the first intron and between
59 and 198 bp for the second intron. Except of
D. pseudoobscura glob1, the coding sequences of all
Drosophila glob1 genes measures 462 bp (including the
stop codon) and translate into polypeptides of 153
amino acids with estimated molecular masses of 17.1–
17.5 kDa. The D. pseudoobscura glob1 sequence has an
insertion of three bp close to the 3¢ end, resulting in a
protein of 154 amino acids. Otherwise no indel was
observed in the coding region. The interspecies amino-
acid differences sum up from two to 36 positions
(75.2–98.7% identity and 91.5–100% similarity, consid-
ering isofunctional replacements). The heme- and
ligand-associated residues, i.e. the proximal and distal
histidines in helix positions E7 and F8, LeuB10,
PheCD1, ArgE10, and IleE11 [18], are invariant in
the glob1 proteins of all Drosophila species (Fig. 4;
Table 1). Only about a third of the amino-acid posi-
tions in the glob1 proteins were found variable. Most
of the observed variations are located in the loop

regions, in particular in the CD, EF and FG corners,
while the helices are well conserved.
Rates of molecular evolution were estimated on the
basis of a PAM substitution matrix assuming the Dro-
sophila divergence times suggested by Tamura et al. [24].
We found that the glob1 proteins have evolved with an
Fig. 2. Structure of Drosophila glob1 genes. The genomic sequences of glob1 genes of D. melanogaster, D. pseudoobscura and D. virilis are
displayed. Exonic sequences are boxed and numbered 1–6 on top, the coding regions are hatched. Putative TATA boxes are indicated and
the positions of potentially relevant hypoxia-responsive sequence elements in the gene region are shown as open squares.
Drosophila hemoglobins T. Burmester et al.
470 FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS
average rate of 1.61 ± 0.56 · 10
)9
amino-acid replace-
ments per site and per year (Table 2). The lowest rate
was calculated from the D. melanogaster – D. yakuba
comparison (0.50 · 10
)9
), which is likely due to a
mutational slow-down in the D. yakuba lineage, whereas
on the other hand a faster average rate of evolution
was observed in the Drosophila subgenus clade, partic-
ularly in the Hawaiian species D. grimshawi. On the
nucleotide level, 0.90 ± 0.31 · 10
)9
nonsynonymous
(dn) and 10.97 ± 5.61 · 10
)9
synonymous (ds) replace-
ments per site per year were inferred. Levels of selective

constraints were measured by ds ⁄ dn ratios [25].
Strong selective pressure on a coding region favours
synonymous substitutions (ds) over nonsynonymous
substitutions that cause amino-acid replacements (dn).
The high average ds ⁄ dn ratio of 14.7 indicates that
significant purifying selection has been imposed on the
glob1 gene.
Identification of two novel globin genes
of Drosophila
By searching the D. melanogaster genome with the
tblastn algorithm and by employing various inverteb-
rate globin sequences as templates, we identified two
novel putative globin genes, which we tentatively called
glob2 and glob3 (Figs 5 and 6). The glob2 gene is anno-
tated as orphan D. melanogaster gene CG15180, glob3 is
known as CG14675. Reverse blastp searches were then
performed with exclusion of the Drosophila sequences.
Assuming the BLOSUM45 model we obtained the high-
est sequence similarity of D. melanogaster glob2 with
the Hb of the bivalve Barbatia virescens (Acc. No.
BAA09587; e-value: 0.003), and of D. melanogaster
glob3 with the globin D of the sea cucumber Caudina
arenicola (Acc. No. AAB19247; e-value: 5 · 10
)5
). The
relatedness of glob2 and glob3 to other globins was cor-
roborated by a random shuffling approach, as it is
implemented in the prss program [26]. The analyses
yield an e-value of 6.96 · 10
)10

for the D. melanogaster
glob2–B. virescens Hb comparison and an e-value
1.02 · 10
)9
for D. melanogaster glob3 vs. globin D
C. arenicola. These comparisons further confirm the
identity of glob2 and glob3 as true globins.
Analyses of Drosophila glob2 genes and proteins
The predicted D. melanogaster glob2 coding sequence
covers 669 bp (including stop codon), which was con-
firmed by sequencing a cDNA fragment that had been
obtained by RT-PCR experiments on adult D. melano-
gaster total RNA (Acc. No. AM086024). In addition,
six entries in the database of D. melanogaster expressed
sequence tags (ESTs [27]); (BE976463, BE978352,
BE978841, BE976596, CB305306, CO331008) demon-
strate that glob2 is actually expressed. Moreover, an
antiserum that had been raised against synthetic glob2
peptides detects a protein band of about 26 kDa in
extracts from total flies (Fig. 7), which corresponds to
the expected size of glob2 (see below).
Comparison of cDNA and genomic sequences show
that the glob2 gene contains two introns, which are
A
B
C
Fig. 3. Conservation of Drosophila globin gene regions. Percent
identity plot (PIP) showing the comparisons of the Drosophila glo-
bin 1 (Gb1; A), globin 2 (Gb2; B) and globin 3 (Gb3; C) genes and
their ± 4 kb flanking regions. The D. melanogaster genes were

used as reference sequences, in the upper row the exons are
boxed, with the black boxes representing the coding sequences.
Gene extensions and transcriptional orientations are indicated by
arrows. Interspecies sequence identities are shown as horizontal
bars on a 50–100% scale. GC-rich regions are indicted as shown
on the right hand side. The boxes in dark grey indicate missing
data. Species abbreviations are: Dsi, D. simulans; Dya, D. yakuba;
Der, D. erecta; Dan, D. ananassae; Dps, D. pseudoobscura; Dvi,
D. virilis; Dmo, D. mojavensis; Dgr, D. grimshawi.
T. Burmester et al. Drosophila hemoglobins
FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS 471
positioned in B12.2 and G7.0 (Fig. 5). The introns are
short, measuring 52 and 127 bp, respectively. The pre-
dicted D. melanogaster glob2 protein covers 222
amino-acid positions (25.3 kDa), which is longer than
glob1 and most other globins (typically  140–150
amino acids). This is mainly due to the 50 and 30
amino-acid extensions at the N- and C-termini,
respectively. Nevertheless, except of an indel of six
amino acids in the DF corner (positions D6 to E4),
the globin core is well conserved. This also comprises
the residues crucial for heme- and O
2
- binding, e.g. the
PheCD1 and the proximal and distal histidines E7 and
F8 (Table 1). The calculated pI of glob2 is 9.18, indi-
cating a highly basic protein. This coincides with the
analyses according to the Reinhardt’s method [28],
which predicts glob2 to be nuclear protein (reliability
76.7%).

Putative glob2 orthologs were identified from the
genomic sequences of D. simulans, D. yakuba,
D. erecta, D. ananassae and D. pseudoobscura (see
Supplementary material; Fig. 5). They all display a
similar length, varying from 218 to 223 amino acids,
and all have short introns in positions B12.2 and
G7.0. Except of an insertion of five amino acids in
the CD loop in D. ananassae glob2 the different
lengths of these proteins result from variations of
the N- and C-terminal extensions. All glob2 proteins
were predicted to reside in the nucleus. The rate cal-
culations, as estimated as described above (Table 2),
show that the Drosophila glob2 proteins evolve with
7.62 ± 3.89 · 10
)9
replacements per site and per
year. This is about six times faster than in the glob1
proteins. The relaxed selective constraint is also indi-
cated by a ds ⁄ dn ratio of 6.5. The PipMaker analy-
sis showed discernable sequence similarities within
Fig. 4. Comparison of Drosophila glob1 amino-acid sequences. The secondary structure elements of Drosophila glob1 (PDB entry 2BK9 [18])
is shown in the upper row. The alpha-helices are designated A through H and the globin consensus helix numbering is given in the lower
row. Conserved amino acids are shaded in grey. The abbreviations are: DmeGb1, D. melanogaster globin 1; DyaGb1, D. yakuba globin 1;
DerGb1, D. erecta globin 1; DanGb1, D. ananassae globin 1; DpsGb1, D. pseudoobscura globin 1; DviGb1, D. virilis globin 1; DmoGb1,
D. mojavensis globin 1, and DgrGb1, D. grimshawi globin 1.
Table 1. Functionally important residues in selected globins. Amino
acids at key positions in human Hb, myoglobin, Chironomus Hb,
G. intestinalis Hb, and Drosophila glob1–3 are given.
Species A12 B10 CD1 E7 E10 E11 F8
Human Hba Trp Leu Phe His Lys Val His

Human Hbb Trp Leu Phe His Lys Val His
Human Mb Trp Leu Phe His Thr Val His
Human Ngb Trp Phe Phe His Lys Val His
Chironomus Hb Trp ⁄ Phe ⁄
Tyr
Leu Phe His Arg Ile ⁄ Val His
G. intestinalis Hb Trp Leu Phe His Arg Ile His
Drosophila glob1 Trp Leu Phe His Arg Ile His
Drosophila glob2 Trp Phe Phe His Ala Met His
Drosophila glob3 Trp Phe ⁄ Tyr Phe His Arg ⁄ Ala ⁄
Thr
Phe His
Table 2. Substitution rates in Drosophila globins. Amino-acid evolu-
tion was modeled according to the PAM matrix [49]. Corrected
nonsynonymous (dn) and synonymous (ds) nucleotide substitutions
per site were calculated by the method of Ota and Nei [51]. Evolu-
tion rates are given as estimated replacements per site and per
year.
Substitution rates (x 10
)9
)
ds ⁄ dn
Amino acids dn ds
Globin 1 1.61 ± 0.56 0.90 ± 0.31 10.97 ± 5.61 14.7
Globin 2 7.62 ± 3.89 4.21 ± 2.02 14.59 ± 6.88 6.5
Globin 3 7.28 ± 1.53 3.86 ± 0.78 17.64 ± 7.27 4.6
Drosophila hemoglobins T. Burmester et al.
472 FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS
the melanogaster subgroup (Fig. 3B), while in
D. ananassae and D. pseudoobscura only the coding

exons are conserved.
The gene of D. melanogaster glob2 was found
localized on the right arm of chromosome 3 at band
83F4. The flanking genes are CG31482 on the proxi-
mal side, which encodes for chain A of the Golgi-
associated Pr-1 protein, and CG15184 on the distal
side, for which no putative ortholog is available and
which is poorly conserved even within the Drosophili-
dae. We therefore used the next neighbouring gene
CG15178 for synteny analyses (see below), which
encodes a highly conserved EF hand calcium binding
protein. In all species of the melanogaster subgroup,
glob2 is located on chromosome 3R, and on the
orthologous chromosome 2 of D. pseudoobscura.
Gene synteny analyses of the neighbouring genes
showed conservation of gene order within the
melanogaster subgroup. However, there is no syntenic
conservation of the glob2 regions between D. melano-
gaster and D. pseudoobscura or D. ananassae. The
D. melanogaster orthologs of the genes adjacent to
D. pseudoobscura glob2 are located on chromosome
3R in regions 83C, which is the glob3 locus (see
below), and 100D.
Fig. 5. Comparison of Drosophila glob2 amino-acid sequences. The secondary structure of Drosophila glob1 is shown in the upper row,
alpha-helices designations and shading was performed as described in Fig. 4. Abbreviations: DmeGb2, D. melanogaster globin 2; DsiGb2,
D. simulans globin 2; DerGb2, D. erecta globin 2; DyaGb2, D. yakuba globin 2; DpsGb2, D. pseudoobscura globin 2; DmoGb2, D. mojavensis
globin 2.
Fig. 6. Comparison of Drosophila glob3 amino-acid sequences. The secondary structure in the upper row derives from Drosophila glob1,
other decorations were performed as described in the legends to Figs 4 and 5. Abbreviations: DmeGb3, D. melanogaster globin 3; DsiGb3,
D. simulans globin 3; DerGb3, D. erecta globin 3; DyaGb3, D. yakuba globin 3; DanGb3, D. ananassae globin 3; DpsGb3, D. pseudoobscura

globin 3; DviGb3, D. virilis globin 3.
Fig. 7. Western blot analyses. About 50 lg of total adult protein
were applied per lane and the glob1 (lane 1) and glob2 (lane 2) pro-
teins were detected using specific antibodies.
T. Burmester et al. Drosophila hemoglobins
FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS 473
Analyses of Drosophila glob3 genes and proteins
The putative coding sequence of the D. melanogaster
glob3 gene measures 588 bp, which translates into a
protein of 195 amino acids. However, we could not
obtain any glob3 cDNA from larval or adult RNA by
RT-PCR. Moreover, no entry in the Drosophila EST
database was recovered that corresponds to glob3.
Because at that stage we could not exclude that glob3
were a pseudogene, we searched the genomic sequences
of the other Drosophila species. In fact, we obtained
apparently conserved orthologs genes of glob3 from
D. simulans, D. yakuba, D. erecta, D. pseudoobscura
and D. virilis (see Supplementary material; Fig. 6).
Gene prediction and comparison with glob2 suggest
the presence of the two introns in B12.2 and G7.0 also
Fig. 8. Phylogeny of Drosophila and other dipteran globins. An alignment of the Drosophila globins, G. intestinalis Hb and selected Hbs from
chironomid midges was analysed by MrBayes 3.1, assuming a WAG model of evolution with gamma-distribution of rates. The Daphnia Hb
domains represent the outgroup. Bayesian posterior probabilities are given at the nodes. The bar corresponds to 0.1 PAM distance. The
abbreviations are: DmaHb1.1 and DmaHb1.2, first and second domain of the hemoglobin (U67067) of D. magna; DpuHb1.1 and DpuHb1.2,
first and second domain of the hemoglobin (Q9U8H0) of D. pulex; Chironomus thummi thummi hemoglobins CttHbI (P02221), CttHbIIb
(AF001292), CttHbIII (M17691), CttHbIV (P02230), CttHbVIIB6 (A30477), CttHbVIII (P02227), CttHbIX (P02223), CttHbX (P02228), CttHbE
(P11582), CttHbZ (P29245); C. thummi piger hemoglobins: CtpHbV (X56271), CtpHbW (X56271), CtpHbY (X56271); GinHb, Gasterophilus
intestinalis hemoglobin (AF063938); for abbreviations of the Drosophila globins refer to Figs 4, 5 and 6.
Drosophila hemoglobins T. Burmester et al.

474 FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS
in the glob3 genes. As for glob2, the glob3 amino-acid
sequences include extensions at the N- and C-terminal
ends, and are longer than the typical globin proteins.
Another similarity of glob2 and glob3 is the six-amino
acids indel in the DF corner. The residues important
for heme-contact and O
2
-binding globins are conserved
(Fig. 6; Table 1). Reinhardt’s method [28] predicts the
glob3 proteins to reside in the cytoplasm. The evolu-
tion rate of the glob3 proteins was calculated to be
7.28 ± 1.53 · 10
)9
replacements per site and per year,
which is in the same range as that of glob2 (Table 2).
The glob3 genes also show relaxed selective constraints,
as indicated by the ds ⁄ dn ratio of 4.6. pipmaker com-
parisons clearly identify the glob3 coding exons
(Fig. 3C), while otherwise there is little conservation
except of closely related species. The 5¢ noncoding
region was found GC-rich. The D. virilis sequence was
too diverged to allow PipMaker analyses.
Database searches show that the glob3 gene of
D. melanogaster is located on chromosome 3R at band
83C4. On the proximal side of glob3, the gene
CG10981 encodes a putative RING zinc finger protein,
on the distal side resides the gene CG1208, which
codes for a sugar-transporter (see Supplementary
material). Gene orders and orientations in this region

are identical in the species of the melanogaster group
(D. melanogaster, D. simulans, D. yakuba, D. erecta,
D. ananassae). Gene synteny is only partially conserved
outside this taxon: in D. pseudoobscura, D. mojavensis
and D. virilis, the genes located proximal to D. melano-
gaster glob3 are not on the same contig as glob3. Gene
synteny is only conserved on the distal side. While in
D. melanogaster glob2 and glob3 are separated on
chromosome 3R by 800 kb, they are neighbouring
genes in D. pseudoobscura on chromosome 2 in a head
to tail orientation separated by 388 bp.
Drosophila hemoglobin phylogeny
Because preliminary phylogenetic studies had demon-
strated the monophyly of arthropod globins (data
not shown), we constructed an alignment that covers
a total of 33 insect globin amino-acid sequences and
includes all Drosophila globins. The globin domains
of the extracellular Hbs of the branchiopod crusta-
ceans Daphnia magna and D. pulex were used as
outgroup. As expected, each of the Drosophila
glob1–3 clades is monophyletic (Fig. 8). The closest
relative of Drosophila glob1 is the intracellular Hb of
the horse botfly G. intestinalis [15]. This clade is associ-
ated with the extracellular Hbs of the Chironomidae.
Although such grouping is supported by only 0.84
Bayesian posterior probability, it was consistently
recovered in all types of analyses. Glob2 and glob3
are highly supported sistergroups. Within the Dro-
sophila clades, the globins generally follow the accep-
ted phylogeny of the species, with the exception of

D. pseudoobscura glob1, which was found to be asso-
ciated with D. erecta glob1. However, the support
value was low and additional Bayesian analyses with
other evolution models or Neighbor joining trees did
not resolve this clade (data not shown).
Discussion
Conservation of Drosophila glob1 and functional
implications
Until recently, it has been assumed that Drosophila has
no Hb genes and even after the determination of the
full D. melanogaster genome [29] their presence in this
species was explicitly denied [30]. Nevertheless, we
identified a true Hb gene in D. melanogaster [16],
which is closely related to the intracellular Hb of the
horse botfly G. intestinalis and to the extracellular Hbs
of the Chironomidae (Fig. 8). While there is no doubt
that the G. intestinalis and Chironomus Hbs carry out
respiratory functions in transporting and storing O
2
[12,14], the actual physiological role of the Drosophila
Hbs remains to be determined. Similarities in the O
2
affinities and tissue distributions of G. intestinalis Hb
and D. melanogaster glob1 may be considered as sup-
port for a respiratory function of glob1 in Drosophila,
possibly facilitating the flow of O
2
from the trachea
and tracheoles into the tissues [17]. This view may sup-
ported by the high expression rate of glob1 mRNA, as

reflected by  130 D. melanogaster ESTs (which corres-
ponds to  0.03% of all D. melanogaster ESTs). More-
over, a rough estimate of glob1 protein content by
comparing the western blot signal intensities (Fig. 7)
with known concentrations of recombinant glob1 sug-
gests that glob1 corresponds to approximately 0.1% of
total adult proteins. Taking into account that the
expression of glob1 is largely restricted to the larval
and adult tracheal system, as well as some regions of
the fat body, this appears to be a significant amount,
which is compatible with the idea that glob1 enhances
local oxygen diffusion rates or acts as oxygen store.
Alternatively, glob1 may function as a buffer system
that protects the inner organs from an excess of O
2
[31,32]. Moreover, other, nonrespiratory functions of
glob1, such as the detoxification of reactive oxygen
species or NO, are also conceivable.
An important role of glob1 in the flies’ metabolism
is supported by the strong conservation of the gene
and protein. Its evolution rate of 1.61 ± 0.56 · 10
)9
T. Burmester et al. Drosophila hemoglobins
FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS 475
amino-acid substitutions per site and per year is low
compared to glob2 and glob3. The evaluation of
D. melanogaster and D. pseudoobscura glob1 coding
sequences shows a low number of nonsynonymous
substitutions (ds) of 0.06, whereas the median of the
predicted orthologs of these species is 0.14 [23]. This is

also reflected by a higher than average amino-acid
identity of 88.9%, whereas the mean identity of 10 987
orthologous protein pairs of D. melanogaster and
D. pseudoobscura is 77.7%. Together with the high
mean ds ⁄ dn ratio (14.7), these data indicate that glob1
is a slowly evolving gene that is under significant
selective pressure.
Recent studies employing the D. melanogaster SL2
cell line have demonstrated that glob1 expression levels
decreased under low oxygen conditions [33]. As in
other animals, hypoxia-response in Drosophila is medi-
ated by binding of HIF-1, which consists of an
a- (Similar in Drosophila) and a b-subunit (tango in
Drosophila), to HREs [34]. Unexpectedly, HIF-1 was
demonstrated to cause a hypoxia-dependent down-
regulation of glob1 in the SL2 cells [33]. Here we have
shown that the glob1 genes of D. melanogaster, D. viri-
lis and D. pseudoobscura harbour several putative
HREs. Although without experimental evidence we
cannot conclude which HRE is functional, the posi-
tional conservation of the HREs in distantly related
Drosophila species suggest that these elements may
actually mediate a hypoxia-response of glob1.
Two novel globin genes in Drosophila with
uncertain function
While the function of glob1 may be related to oxygen
supply, the physiological roles of glob2 and glob3 are
presently unknown. Moreover, we failed to amplify
any glob3 mRNA from D. melanogaster. Although we
cannot exclude technical problems, we must consider

the possibility that this gene is not expressed under
normal physiological conditions. Latter assumption is
also supported by the fact that no glob3 cDNA could
be found in the > 300.000 EST sequences of D. mel-
anogaster. Nevertheless, the conservation of glob3 in
Drosophila evolution (Fig. 6) suggests that this gene is
not a pseudogene, but is functional. It is conceivable
that glob3 is only expressed under particular physio-
logical conditions or only during a short period of
development.
glob2 is expressed in Drosophila, as confirmed by
RT-PCR experiments, EST database entries and
western blotting. The expression level is much lower
than that of glob1, as reflected by only six EST
sequences in the database and own estimations based
on western blots (Fig. 7). A respiratory function of
glob2 is therefore unlikely, and at present the phy-
siological role of this protein must remain uncertain.
There are, however, two interesting features provide
some hints. First, glob2 was predicted to reside in
the nucleus. A nuclear Hb has been observed, e.g. in
alfalfa (Medicago sativa) and has been related to the
NO metabolism [35]. Second, it is noteworthy that
all ESTs derive from a testis specific D. melanogaster
cDNA library. Three additional D. yakuba glob2
ESTs are also available (CV790354, CV790570,
CV784980), which were obtained from testis cDNA,
too. This might suggest a testis-specific function of
glob2 as transcriptional regulator, which, however,
requires further investigations.

Insect globin phylogeny
Globins are widespread in all kingdoms of life and
display a complex evolutionary pattern [4,6,36]. The
last common ancestor of all insects certainly har-
boured a globin gene, although is must remain
uncertain whether it had respiratory functions. Our
analyses indicate that the insect H bs are monophyletic,
and that a clade comprising of glob2 and glob3 is
the earliest offshoot of the insect globins. Therefore,
the separation of the glob2 ⁄ glob3 clade and other
insect globins likely occurred early in the evolution
of this taxon. On the other hand we cannot exclude
the possibility that the high evolution rate of glob2
and glob3 has masked its actual relatedness and
have led to a long-branch attraction effect. The rela-
tionship of glob2 and glob3 is not only confirmed by
the phylogenetic analysis, but also by the fact that
they are neighbouring genes in D. pseudoobscura .
The most parsimonious explanation is therefore that
glob2 and glob3 arose by gene duplication in the
Drosophila stemline. During the evolution of the
melanogaster group the glob2 gene was moved to a
different region of the 3R chromosome. This is com-
patible with the notion that within the Drosophila
genus there is a strong tendency for genes to remain
on the same chromosome arm [22,23].
The Drosophila glob1 genes share a recent common
ancestry with the G. intestinalis Hb, and it may be
assumed that the genes are orthologs that diverged
upon the separation of the species (about 80–100 mil-

lion years ago [37,38]). As mentioned above, this gene
orthology might be considered as support for a role in
respiration also of Drosophila glob1. The respiratory
function of the Chironomus Hbs has convincingly been
proven (for a review, see [5]). The phylogenetic tree
show that the Chironomus Hb genes most likely derive
Drosophila hemoglobins T. Burmester et al.
476 FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS
from intracellular Hbs, as represented today by
G. intestinalis Hb and Drosophila glob1. Therefore,
with the evolution of the chironomid midges there was
a shift in function from an intracellular O
2
-binding
protein with possible O
2
supply function to an extra-
cellular O
2
-transport protein. This event must have
occurred after the separation of the Chironomidae
(Nematocera) and the brachyceran flies. An ancient
origin of the Chironomus Hbs, as originally proposed,
e.g. by Goodman et al. [39,40], is not supported.
Introns and evolution
It is well established that intron positions may be phy-
logenetically informative, particularly in case of globin
genes. In fact, the relatedness and the basal position of
glob2 and glob3 within the insect globins are supported
by the presence of introns B12.2 and G7.0. In fact,

these two intron positions are conserved in all verte-
brate and most invertebrate globin genes, and almost
certainly reflect the ancient gene structure of the
eukaryote globins [4]. Therefore, the loss and gain of
introns occurred in the clade of glob1 and chironomid
globin genes only after they diverged from a common
ancestor. The G7.0 intron has been retained in the
glob1 and G. intestinalis Hb, while the B12.2 intron
was lost in this clade. Instead, a novel intron had
emerged in the common ancestor of glob1 and
G. intestinalis Hb at position D7.0. None of the known
chironomid globin genes have introns in the ‘classical’
positions B12.2 or G7.0. Taking into account the
phylogenetic tree (Fig. 8), the most parsimonious
explanation for this pattern is independent intron
losses. While the B12.2 intron was lost in the clade
leading to the last common ancestor of Drosophila
glob1, G. intestinalis Hb and the chironomid Hbs, the
G7.0 intron was deleted only before the radiation of
the chironomid globins. While most present-day
chironomid globin genes do not have any introns,
some have acquired during evolution central introns at
positions E9.1 and E15.0 [41]. The complex intron
pattern in dipteran globin genes can only be inter-
preted in terms of a dynamic evolution of gene struc-
ture by intron loss and insertion [42,43].
Experimental procedures
Animals
D. melanogaster Oregon R, D. pseudoobscura and D. virilis
were maintained at 18° or 25 °C on standard yeast-corn-

meal-agar-sucrose medium sprinkled with active dry yeast.
Propionic acid and 4-hydroxymethylbenzoate were added
as mould inhibitors. D. virilis were a kind gift of C. Kra-
emer (Institute of Molecular Genetics, Mainz).
Database searches and sequence analyses
The blast algorithm [19] was employed to search the
databases of genomic DNA sequences available at GenBank
() and FlyBase (http://flybase.
net). The significance of the observed amino-acid sequence
similarities were evaluated using a Monte-Carlo shuffling
approach applying the prss3 program (fasta package [26]);.
Probability scores were estimated using the blosum50 matrix
assuming a gap creation penalty of )12 and a gap length
penalty of )2 with 1000 shuffles. The relevant nucleotide
sequences of the genes were extracted from the databases and
assembled by the aid of genedoc 2.6 [44]. DNA translation
and analyses of primary structures were performed with the
programs of the ExPASy Molecular Biology Server (http://
www.expasy.org). Interspecific comparison of gene sequences
was performed with multipipmaker ( />pipmaker/), which computes and visualizes local alignments
in two or more sequences based on the blastz algorithm
[45,46]. The D. melanogaster genes were used as templates.
The multiple sequence alignments from pipmaker were
visualized as ‘percent identity plots’ (PIP). Promoter predic-
tions were carried out using the server at BDGP (http://
www.fruitfly.org/seq_tools/promoter.html) [47].
Molecular biology
RNA from Drosophila species were extracted either with
the GITC method [48] or by the ENZA total RNA extrac-
tion kit (Peqlab). Various specific or degenerate oligonucleo-

tide primers were used to amplify cDNA fragments by
reverse transcription-PCR experiments, employing the
Qiagen OneStep kit system or the Superscript II RNase H
–
(Invitrogen) according to the manufacturer’s instructions.
The primer sequences are available from the authors upon
request. The sequences of the PCR products were either
obtained by direct sequencing or after cloning of the frag-
ments into the pCR4-TOPO vector (Invitrogen). Sequences
were obtained from both strands using a commercial
sequencing service (GENterprise). Missing 5¢- and 3¢ ends
of the cDNAs were obtained using the RACE system by
Invitrogen.
A kFIX II-library of D. virilis genomic DNA was kindly
provided by H. Kress. Screening was performed with a
D. melanogaster glob1 probe that had been labelled with
32
P-ATP employing the random labelling kit by Roche.
Positive phage clones were grown on E. coli p2392 in
standard l-medium until lysis [48]. Phage were then precipi-
tated over night with 7% polyethylene glycol, 6% NaCl
and suspended in 10 mm MgCl
2
. Contaminations by
bacterial nucleic acids were removed by 30 min digestion
with RNase A and DNase I. Phage DNA was obtained by
T. Burmester et al. Drosophila hemoglobins
FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS 477
lysis in SDS and subsequent extraction according to the
standard phenol ⁄ chloroform method.

Protein biochemistry
Antibodies against recombinant D. melanogaster glob1 pro-
tein had been raised in rabbits [17]; antibodies against
D. melanogaster glob2 were prepared by immunizing a rabbit
with two synthetic peptides that correspond to amino acids
75–89 (H
2
N-FRLALDENLPTHLKN-COOH) and 139–153
(H
2
N-DEARKRAMSTALRTT-COOH). The peptides were
coupled to KLH and used for immunization of rabbits.
Peptide syntheses and immunizations were carried out
by EUROGENTEC. The specific anti-glob2 IgG were
purified from the serum employing the synthetic peptides
coupled to SulfoLink columns (Pierce) according to the
instructions of the manufacturer. After elution, the antibod-
ies were stored at 4 °Cin50mm Tris, 100 mm glycine, pH
 7.4) supplemented with 1% BSA and 0.05% NaN
3
. For
western blotting, protein extracts were separated by gel
electrophoresis on a 14% acrylamide gel in the presence of
SDS. The proteins were transferred to nitrocellulose by the
semidry method. The filters were blocked in 2% nonfat dry
milk in Tris buffered saline ⁄ Tween (TBST; 10 mm
Tris ⁄ HCl, pH 7.4, 140 mm NaCl, 0.3% Tween 20) for sev-
eral hours and subsequently incubated with the antibodies
(1 : 10 000) diluted in TBST ⁄ 2% milk. Goat anti-(rabbit
IgG) Ig conjugated with alkaline phosphatase was used

as secondary antibody. Detection was performed using
Nitro Blue tetrazolium and 5-bromo-4-chloroindol-2-yl
phosphate.
Sequence evolution and phylogenetic studies
The globin DNA and amino-acid sequences were aligned
by hand with GeneDoc [44]. Pairwise protein distances
according to the pam model [49] were calculated with the
program protdist from the phylip 3.64 package [50]. Cor-
rected estimates of dn, the number of nonsynonymous
nucleotide substitutions per site, and ds, the number of syn-
onymous nucleotide substitutions per site, were obtained
for each gene according to the method of Nei and Gojobori
[25], as implemented in the snap program [51]. To estimate
the rates of evolution of globin genes and proteins, we fol-
lowed the standard phylogeny of Drosophilidae as pro-
posed by Russo et al. [52], but assumed the divergence
times of drosophilid taxa suggested by Tamura et al. [24].
These are on average about twice as old as those calculated
by Russo et al. [52], but derive from a much larger data set
and fit better own calculations based on multiple protein
alignments (data not shown).
The appropriate model of amino-acid sequence evolution
was selected by ProtTest [53] using the Akaike Information
Criterion (AIC). Distance matrices were calculated with
tree-puzzle 5.01 [54]. Neighbor-joining trees were inferred
with the program neighbor from phylip [50]. The reliabil-
ity of the trees was tested by bootstrap analysis [55] with
100 replications using puzzleboot (shell script by
M. Holder, University of Houston, USA, and A. Roger,
Dalhousie University, USA). mrbayes 3.1 was employed

for Bayesian phylogenetic analyses [56] assuming the wag
model with gamma distribution of rates (four gamma cate-
gories). The analyses were run for 300 000 generations with
random starting trees. The final average standard deviations
of split frequencies were < 0.02. Trees were sampled every
10th generation and posterior probabilities were estimated
on 5000 trees.
Acknowledgements
We are grateful to C. Kraemer (Mainz, Germany) for
providing D. virilis and H. Kress (Berlin, Germany)
for the D. virilis genomic library. We thank S. Dewilde
and L. Moens (University of Antwerp, Belgium) for
recombinant D. melanogaster glob1 samples and
discussions, and D. De Sanctis and M. Bolognesi
(University of Milano, Italy) for making available
the glob1 structure. This work is supported by the
Deutsche Forschungsgemeinschaft (Bu956 ⁄ 5 to T.B.;
Bu956 ⁄ 6 to T.B and T.H).
References
1 Dickerson RE & Geis I (1983) Hemoglobin: Structure,
Function, Evolution, and Pathology. Benjamin ⁄
Cummings Publishers Co, Menlo Park.
2 Berg JM, Tymoszko JL & Stryer L (2002) Biochemistry.
W.H. Freeman and Co., New York.
3 Wittenberg JB (1992) Functions of cytoplasmatic hemo-
globins and myohemerythrin. Adv Comp Environ Physiol
13, 60–85.
4 Hardison RC (1996) A brief history of hemoglobins:
plant, animal, protist, and bacteria. Proc Natl Acad Sci
USA 93, 5675–5679.

5 Weber RE & Vinogradov SN (2001) Nonvertebrate
hemoglobins: functions and molecular adaptations.
Physiol Rev 81, 569–628.
6 Vinogradov SN, Hoogewijs D, Bailly X, Arredondo-
Peter R, Guertin M, Gough J, Dewilde S, Moens L &
Vanfleteren JR (2005) Three globin lineages belonging
to two structural classes in genomes from the three
kingdoms of life. Proc Natl Acad Sci USA 102, 11385–
11389.
7 Burmester T & Hankeln T (2004) Neuroglobin: a
respiratory protein of the nervous system. News Phys
Sci 19, 110–113.
8 Freitas TA, Saito JA, Hou S & Alam M (2005) Globin-
coupled sensors, protoglobins, and the last universal
common ancestor. J Inorg Biochem 99, 23–33.
Drosophila hemoglobins T. Burmester et al.
478 FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS
9 Hankeln T, Ebner B, Fuchs C, Gerlach F, Haberkamp
M, Laufs TL, Roesner A, Schmidt M, Weich B, Wystub
S, Saaler-Reinhardt S, Reuss S, Bolognesi M, De Sanc-
tis D, Marden MC, Kiger L, Moens L, Dewilde S,
Nevo E, Avivi A, Weber RE, Fago A & Burmester T
(2005) Neuroglobin and cytoglobin in search of their
role in the vertebrate globin family. J Inorg Biochem 99,
110–119.
10 Willmer P, Stone G & Johnston I (2000) Environmental
Physiology of Animals. Blackwell, Oxford, UK.
11 Hagner-Holler S, Schoen A, Erker W, Marden JH,
Rupprecht R, Decker H & Burmester T (2004) A
respiratory hemocyanin from an insect. Proc Natl Acad

Sci USA 101, 871–874.
12 Ewer RF (1942) On the function of hemoglobin in Chir-
onomus. J Exp Biol 18, 197–205.
13 Osmulski PA & Leyko W (1985) Structure, function
and physiological role of Chironomus hemoglobin. Comp
Biochem Physiol 85B, 701–722.
14 Keilin D & Wang YL (1946) Haemoglobin of Gastro-
philus larvae. Purification and properties Biochem J 40,
855–866.
15 Dewilde S, Blaxter M, Van Hauwaert ML, Van Houte
K, Pesce A, Griffon N, Kiger L, Marden MC, Vermeire
S, Vanfleteren J, Esmans E & Moens L (1998) Struc-
tural, functional, and genetic characterization of Gastro-
philus hemoglobin. J Biol Chem 273, 32467–32474.
16 Burmester T & Hankeln T (1999) A globin gene of Dro-
sophila melanogaster. Mol Biol Evol 16, 1809–1811.
17 Hankeln T, Jaenicke V, Kiger L, Dewilde S, Ungerechts
G, Schmidt M, Urban J, Marden MC, Moens L & Bur-
mester T (2002) Characterization of Drosophila hemo-
globin: Evidence for hemoglobin-mediated respiration in
insects. J Biol Chem 277, 29012–29017.
18 de Sanctis D, Dewilde S, Vonrhein C, Pesce A, Moens
L, Ascenzi P, Hankeln T, Burmester T, Ponassi M,
Nardini M & Bolognesi M (2005) Bis-histidyl heme
hexacoordination, a key structural property in Droso-
phila melanogaster hemoglobin. J Biol Chem 280,
27222–27229.
19 Altschul SF, Madden TL, Scha
¨
ffer AA, Zhang J, Zhang

Z, Miller W & Lipman DJ (1997) Gapped BLAST and
PSI-BLAST: a new generation of protein database
search programs. Nucleic Acids Res 25, 3389–3402.
20 Wenger RH & Gassmann M (1997) Oxygen(es) and the
hypoxia-inducible factor-1. Biol Chem 378, 609–616.
21 Semenza GL, Jiang B-H, Leung SW, Passantino R,
Concordet J-P, Maire P & Giallongo A (1996) Hypoxia
response elements in the aldolase A, enolase 1, and
lactate dehydrogenase A gene promoters contain essen-
tial binding sites for hypoxia-inducible factor 1. J Biol
Chem 271, 32529–32537.
22 Segarra C, Lozovskaya ER, Ribo G, Aguade M &
Hartl DL (1995) P1 clones from Drosophila melanoga-
ster as markers to study the chromosomal evolution of
Muller’s A element in two species of the obscura group
of Drosophila. Chromosoma 104, 129–136.
23 Richards S, Liu Y, Bettencourt BR, Hradecky P, Letov-
sky S, Nielsen R, Thornton K, Hubisz MJ et al. (2005)
Comparative genome sequencing of Drosophila pseudo-
obscura: chromosomal, gene, and cis-element evolution.
Genome Res 15, 1–18.
24 Tamura K, Subramanian S & Kumar S (2004) Tem-
poral patterns of fruit fly (Drosophila) evolution
revealed by mutation clocks. Mol Biol Evol 21, 36–44.
25 Nei M & Gojobori T (1986) Simple methods for
estimating the numbers of synonymous and nonsyno-
nymous nucleotide substitutions. Mol Biol Evol 3, 418–
426.
26 Pearson WR (2000) Flexible sequence similarity search-
ing with the FASTA3 program package. Methods Mol

Biol 132, 185–219.
27 Boguski MS, Lowe TM & Tolstoshev CM (1993)
dbEST – database for ‘expressed sequence tags’. Nat
Genet 4, 332–333.
28 Reinhardt A & Hubbard T (1998) Using neural net-
works for prediction of the subcellular location of pro-
teins. Nucleic Acids Res 26, 2230–2236.
29 Adams MD, Celniker SE, Holt RA, Evans CA,
Gocayne JD, Amanatides PG, Scherer SE, Li PW et al.
(2000) The genome sequence of Drosophila melanogas-
ter. Science 287, 2185–2195.
30 Rubin GM, Yandell MD, Wortman JR, Gabor Miklos
GL, Nelson CR, Hariharan IK, Fortini ME, Li PW
et al. (2000) Comparative genomics of the eukaryotes.
Science 287, 2204–2215.
31 Hetz SK & Bradley TJ (2005) Insects breathe discon-
tinuously to avoid oxygen toxicity. Nature 433, 516–519.
32 Burmester T (2005) A welcome shortage of breath.
Nature 433, 471–472.
33 Gorr TA, Tomita T, Wappner P & Bunn HF (2004)
Regulation of Drosophila hypoxia-inducible factor
(HIF) activity in SL2 cells: identification of a hypoxia-
induced variant isoform of the HIFalpha homolog gene
similar. J Biol Chem 279, 36048–36058.
34 Crews S (1998) Control of cell lineage-specific develop-
ment and transcription by bHLH-PAS proteins. Genes
Dev 12, 607–620.
35 Seregelyes C, Mustardy L, Ayaydin F, Sass L, Kovacs
L, Endre G, Lukacs N, Kovacs I, Vass I, Kiss GB,
Horvath GV & Dudits D (2000) Nuclear localization of

a hypoxia-inducible novel non-symbiotic hemoglobin in
cultured alfalfa cells. FEBS Lett 482, 125–130.
36 Roesner A, Fuchs C, Hankeln T & Burmester T (2005)
A globin gene of ancient evolutionary origin in lower
vertebrates: Evidence for two distinct globin families in
animals. Mol Biol Evol 22, 12–20.
37 Burmester T, Massey HC Jr, Zakharkin SO & Benes
ˇ
H
(1998) The evolution of hexamerins and the phylogeny
of insects. J Mol Evol 47, 93–108.
T. Burmester et al. Drosophila hemoglobins
FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS 479
38 Wiegmann BM, Yeates DK, Thorne JL & Kishino H
(2003) Time flies, a new molecular time-scale for bra-
chyceran fly evolution without a clock. Syst Biol 52,
745–756.
39 Goodman M, Braunitzer G, Kleinschmidt T &
Aschauer H (1983) The analysis of a protein-poly-
morphism: evolution of monomeric and homodimeric
haemoglobins (Erythrocruorins) of Chironomus
Thummi Thummi. Hoppe-Seyler Z Physiol Chem 364,
205–217.
40 Goodman M, Pedwaydon J, Czelusniak J, Suzuki T,
Gotoh T, Moens L, Shishikura F, Walz D & Vinogra-
dov S (1988) An evolutionary tree for invertebrate glo-
bin sequences. J Mol Evol 27, 236–249.
41 Hankeln T, Friedl H, Ebersberger I, Martin J &
Schmidt ER (1997) A variable intron distribution in glo-
bin genes of Chironomus: evidence for recent intron

gain. Gene 205, 151–160.
42 Logsdon JM, Stoltzfus A & Doolittle WF (1998) Mole-
cular evolution: Recent cases of spliceosomal intron
gain? Curr Biol 8, R560–R563.
43 Roy SW (2003) Recent evidence for the exon theory of
genes. Genetica 118, 251–266.
44 Nicholas KB & Nicholas HB Jr (1997) GeneDoc: analy-
sis and visualization of genetic variation. http://www.
psc.edu/biomed/genedoc/.
45 Schwartz S, Elnitski L, Li M, Weirauch M, Riemer C,
Smit A, Green ED, Hardison RC & Miller W (2003)
MultiPipMaker and supporting tools: alignments and
analysis of multiple genomic DNA sequences. Nucleic
Acids Res 31, 3518–3524.
46 Schwartz S, Zhang Z, Frazer KA, Smit A, Riemer C,
Bouck J, Gibbs R, Hardison RC & Miller W (2000)
PipMaker – a web server for aligning two genomic
DNA sequences. Genome Res 10, 577–586.
47 Reese MG (2001) Application of a time-delay neural
network to promoter annotation in the Drosophila mel-
anogaster genome. Comput Chem 26, 51–56.
48 Sambrook J & Russell DW (2001) Molecular Cloning:
A Laboratory Manual, 3rd edn., Cold Spring Harbor
Laboratory Press, Cold Spring Harbor, New York.
49 Dayhoff MO, Schwartz RM & Orcutt BC (1978) A
Model of Evolutionary Change in Proteins. In Atlas of
Protein Sequence Structur, Vol. 5 (Dayhoff, M O, ed.)
Suppl. 3, pp. 345–352. National Biomedical Research
Foundation, Washington DC, USA.
50 Felsenstein J (2004) phylip (Phylogeny Inference Pack-

age), Version 3.6. Department of Genetics, University
of Washington, Seattle, WA, USA.
51 Ota T & Nei M (1994) Variance and covariances of the
numbers of synonymous and nonsynonymous substitu-
tions per site. Mol Biol Evol 11, 613–619.
52 Russo CA, Takezaki N & Nei M (1995) Molecular phy-
logeny and divergence times of drosophilid species. Mol
Biol Evol 12, 391–404.
53 Abascal F, Zardoya R & Posada D (2005) ProtTest:
Selection of best-fit models of protein evolution. Bioin-
formatics 21, 2104–2105.
54 Strimmer K & von Haeseler A (1996) Quartet puzzling:
a quartet maximum likelihood method for reconstruct-
ing tree topologies. Mol Biol Evol 13, 964–969.
55 Felsenstein J (1985) Confidence limits on phylogenies:
an approach using the bootstrap. Evolution 39, 783–791.
56 Huelsenbeck JP & Ronquist F (2001) mrbayes: Baye-
sian inference of phylogenetic trees. Bioinformatics 17,
754–755.
Supplementary material
The following supplementary material is available
online:
Table S1. Chromosomal localisation of Drosophila
globin and flanking genes.
This material is available as part of the online article
from
Drosophila hemoglobins T. Burmester et al.
480 FEBS Journal 273 (2006) 468–480 ª 2006 The Authors Journal compilation ª 2006 FEBS