Mol. Biol. Evol. 18(4):530–541. 2001
? 2001 by the Society for Molecular Biology and Evolution. ISSN: 0737-4038
Giardia lamblia Expresses a Proteobacterial-like DnaK Homolog
Hilary G. Morrison,* Andrew J. Roger,† Todd G. Nystul,‡ Frances D. Gillin,‡ and
Mitchell L. Sogin*
*The Josephine Bay Paul Center of Comparative Molecular Biology and Evolution, Marine Biological Laboratory, Woods
Hole, Massachusetts; †Department of Biochemistry and Molecular Biology, Dalhousie University, Halifax, Nova Scotia,
Canada; and ‡Department of Pathology, Division of Infectious Diseases and Center for Molecular Genetics, University of
California at San Diego School of Medicine
We identified a novel gene encoding molecular chaperone HSP70 in the amitochondriate parasite Giardia lamblia.
The predicted protein is similar to bacterial DnaK and mitochondrial HSP70s. The gene is transcribed and translated
at a constant level during trophozoite growth and encystation. Alignment of the sequence with a data set of cytosolic,
endoplasmic reticulum (ER), mitochondrial, and DnaK HSP70 homologs indicated that the sequence was extremely
divergent and contained insertions unique to giardial HSP70s. Phylogenetic analyses demonstrated that this sequence
was distinct from the cytosolic and ER forms and was most similar to proteobacterial and mitochondrial DnaKs.
However, a specific relationship with the alpha proteobacterial and mitochondrial sequences was not strongly sup-
ported by phylogenetic analyses of this data set, in contrast to similar analyses of cpn60. These data neither confirm
nor reject the possibility that this gene is a relic of secondary mitochondrial loss; they leave open the possibility
that it was acquired in a separate endosymbiotic event.
In many phylogenetic studies, Giardia lamblia (Di-
plomonadida) consistently represents one of the most
basal eukaryotic branches (Leipe et al. 1993; Cavalier-
Smith and Chao 1996; Hashimoto and Hasegawa 1996),
although its deep-branching position has been ques-
tioned (Stiller and Hall 1999; Morin 2000). Giardia
lacks mitochondria and peroxisomes, and its energy me-
tabolism appears to be cytosolic. There is no compelling
evidence yet for the presence of introns or functional
splicesomal machinery. Giardia quite successfully
makes its living as a parasite of the mammalian gut,
often using metabolic pathways that are most closely
related to prokaryotes. Giardia’s 12-Mb genome (Adam
2000) is small compared with many eukaryotic ge-
nomes, e.g., those of plants, animals, and many protists.
What is the common denominator that creates a eukary-
otic cell? We undertook sequencing of the Giardia ge-
nome to answer this and other fundamental questions
about this presumptively early diverging eukaryote.
A key event in the evolution of the eukaryotic cell
was the acquisition of the mitochondrion via the endo-
symbiosis of a prokaryote. For many years, this endo-
symbiotic event was assumed to have occurred subse-
quent to the divergence of the most basal lineages. How-
ever, evidence has been accumulating that mitochondria
or other ancestral symbionts occurred in these early-
branching lineages. Genes have recently been discov-
ered in the nuclear genomes of these amitochondriates
that appear to have been introduced by an ancestral en-
dosymbiont. Coding regions for ?-proteobacterial or mi-
tochondrion-targeted molecular chaperones have been
discovered in the nuclear genomes of microsporidia
Key words: Giardia lamblia, HSP70, DnaK, molecular chaper-
Address for correspondence and reprints: Mitchell L. Sogin, The
Josephine Bay Paul Center of Comparative Molecular Biology and
Evolution, Marine Biological Laboratory, Woods Hole, Massachusetts
02543-1015. E-mail: email@example.com.
(Germot, Philippe, and Le Guyader 1997; Hirt et al.
1997; Peyretaillade et al. 1998), trichomonads (Bui,
Bradley, and Johnson 1996; Germot, Philippe, and Le
Guyader 1996; Horner et al. 1996; Roger, Clark, and
Doolittle 1996; Hashimoto et al. 1998), and Giardia
(Roger et al. 1998). The most parsimonious explanation
of this phyletic distribution pattern is that mitochondria
were present in the earliest eukaryotes (the mitochon-
dria-early hypothesis). However, phylogenetic analyses
of cpn60 and HSP70 do not consistently recover topol-
ogies congruent with nuclear gene trees as represented
by rRNA analyses. Either rRNA trees are unreliable, or
the chaperonins lack phylogenetic resolving power, or
the chaperonin genes have been acquired independently
by different lineages.
To address these issues, we characterized a new
HSP70 gene from G. lamblia. Heat shock protein 70s
(HSP70s) are the most prevalent of the molecular chap-
erones and are found in all domains of life. The bacterial
HSP70 homolog is known as DnaK. Most eukaryotes
have at least three HSP70 genes, and the gene products
are compartmentalized to the cytosol, the endoplasmic
reticulum (ER), or the mitochondrion. The cytosolic and
ER forms result from an ancient gene duplication in the
eukaryotic lineage, whereas the mitochondrial form was
acquired from the endosymbiosis of the DnaK-contain-
ing proteobacterium that became the mitochondrion.
Gupta et al. (1994) reported the cloning of the ER and
cytosolic homologs of HSP70 in Giardia, but no mito-
chondrial homolog was discovered. Here, we report on
the cloning, sequencing, expression studies, and phylo-
genetic analysis of a DnaK-like form of HSP70 from
Giardia that is clearly of bacterial origin.
Materials and Methods
Parasites and Culture
Giardia lamblia strain WB was isolated from a pa-
tient with chronic symptomatic giardiasis (Smith et al.
1982) and belongs to the most common group of isolates
from humans (Ey et al. 1996). WB clone C6 (ATCC
by guest on May 30, 2013
Giardia lamblia DnaK Homolog531
number 50803) trophozoites were routinely subcultured
twice weekly in filter-sterilized Diamond’s TYI-S-33
medium with 10% adult bovine serum (Irvine Scientific,
Santa Ana, Calif.) (Diamond, Harlow, and Cunnick
1978) and 0.5 mg/ml of bovine bile (Keister 1983),
without added vitamins, antibiotics, or iron, at pH 7.0–
Cloning and Sequencing of G. lamblia HSP70
The construction of a Giardia genomic DNA li-
brary in the plasmid pBluescript has previously been
described (Henze et al. 1998). A basic local alignment
tool (BLASTX; Altschul et al. 1990) search of sequence
data from one clone returned a highly significant hit to
hydrogenosomal HSP70 from Trichomonas vaginalis
and to various bacterial DnaK proteins. A 558-bp frag-
ment was excised from the clone by EcoRI digestion
and used to probe a Giardia ?-ZapII genomic library
using a digoxigenin detection protocol (Boehringer
Mannheim, Indianapolis, Ind.). A clone containing an
insert of approximately 9 kb was identified, and the
pBluescript phagemid was excised following the man-
ufacturer’s protocols (Stratagene, Valencia, Calif.). A
2,130-bp region containing the HSP70 coding sequence
was sequenced on both strands. Sequencing was carried
out using the Excel II cycle sequencing protocol (Epi-
centre Technologies, Madison, Wis.). Sequencing reac-
tion products were run on LI-COR 4200 automated se-
quencers (Middendorf et al. 1992; Roger et al. 1998).
Expression of total HSP70 proteins during differ-
entiation and heat and oxygen stress was monitored in
preliminary experiments by immunoblotting using poly-
clonal antibodies to yeast ssc-1 (mitochondrial HSP70)
as a probe. (The polyclonal antibodies were a kind gift
from Dr. Elizabeth Craig, University of Chicago.) Since
this reagent was made against a region that was similar
between giardial ER, cytoplasmic, and mitochondrial
HSP70 genes, this gave an estimate of overall changes
in HSP70 levels. Since no significant changes were not-
ed in the initial studies, Northern analyses were carried
out using conditions of the greatest stress tolerated by
the parasites. Protein and RNA isolation were as de-
Parasites were harvested at the indicated stages of
differentiation or after the stress conditions specified by
scraping to release attached cells without chilling, col-
lected by centrifugation, and resuspended in PBS con-
taining the protease inhibitors phenylmethylsulfonyl
fluoride (1 mM, freshly diluted from 80% isopropanol
stock) and E64 (14 ?M) to inhibit most giardial pro-
teinase activity. The proteins were precipitated with cold
6% trichloroacetic acid and collected by centrifugation.
The pellet was resuspended to a cell concentration of
2.5 ? 107cells/ml in SDS/PAGE sample buffer con-
taining 2% SDS with 50 mM dithiothreitol, neutralized
with NaOH, boiled for 6 min, and stored at ?70?C.
Twenty microliters of parasite extract was separated by
SDS-PAGE on 4%–20% gradient gels (Novex, San Di-
ego, Calif.) (Laemmli 1970). The separated antigens
were transferred to nitrocellulose membranes (Towbin,
Staehelin, and Gordon 1979) for 18 h at 30 V and then
for 1 h at 70 V. Membranes were blocked in PBS with
7.5% (w/v) BSA for 2 h and washed thoroughly. The
blots were probed for 2 h with polyclonal antibodies to
yeast ssc1 diluted 1:500 in PBS, washed in Tris-buffered
saline, reacted with Protein A-HRP (Zymed, San Fran-
cisco, Calif.) diluted 1:2,000 in Tris-buffered saline for
1 h, washed in Tris-buffered saline, and developed with
ECL reagents according to the manufacturer’s instruc-
tions (Amersham Life Sciences, Arlington Heights, Ill.).
Total RNA was isolated from G. lamblia at the in-
dicated stages of differentiation or after the stress con-
ditions specified by extraction with RNAzol B according
to the manufacturer’s instructions (Tel-Test Inc., Friends-
wood, Tex.). Samples of total RNA (10 ?g per lane)
were fractionated in 1.5% formaldehyde-agarose gels,
downward capillary blotted in 20 ? SSC, and immo-
bilized onto nylon membranes (Zeta-Probe, Bio-Rad,
Hercules, Calif.) by UV cross-linking. For Northern hy-
bridization, a probe corresponding to the open reading
frame of the giardial DnaK-like HSP70 was purified and
radiolabeled by random priming (Prime It II kit, Stra-
tagene). Blots were prehybridized in 6 ? SSC, 5 ? Den-
hardt’s solution, 0.5% (w/v) SDS, and 20 ?g/ml salmon
sperm DNA for 1 h at 65?C. Hybridization at 65?C was
continued overnight in the presence of the HSP70 probe
(Knodler et al. 1999). The membrane was washed twice
in 2 ? SSC/0.1% (w/v) SDS at room temperature for
15 min, and then once at 60?C for 15 min in 0.2 ? SSC/
0.1% (w/v) SDS. The washed membrane was autoradio-
Induction of Encystation
Pre-encysting cultures were grown to late log phase
(66 h) in TYI-S-33 medium (pH 7.0) without bile but
containing the antibiotics piperacillin (500 ?g/ml, Led-
erle Laboratories, Carolina, Puerto Rico) and amikacin
(125 ?g/ml; Bristol Laboratories, Syracuse, N.Y.),
which do not affect G. lamblia growth or differentiation
(Meng, Hetsko, and Gillin 1996). Encystation was ini-
tiated by removing the spent medium and nonadherent
cells and refeeding the adherent trophozoite monolayer
with encystation medium: TYI-S-33 with antibiotics but
without bovine bile, adjusted to pH 7.8 with NaOH and
supplemented with 0.25 mg/ml porcine bile and 5 mM
lactic acid, (hemicalcium) (Meng, Hetsko, and Gillin
1996), which increases biological activity of cysts (Bou-
cher and Gillin 1990).
For protein isolation, 250 ml of growth medium
was inoculated at 700 cells/ml and grown to confluence
by guest on May 30, 2013
532Morrison et al.
over 3 days at 37?C. Attached cells were isolated by
pouring off medium containing free-swimming cells, re-
fed with fresh media, harvested by scraping, and trans-
ferred to 8-ml glass tubes to facilitate heat transfer. After
a 2-h preincubation period at 37?C, cells were incubated
in a 42?C water bath for 15, 30, or 45 min. The tubes
were periodically inverted during the incubation period
to promote even heat transfer. Following the 42?C in-
cubation, cells were allowed to recover at 37?C for 0,
30, or 60 min and harvested by chilling on ice for 20
min. For RNA isolation, cells were grown and heat
shocked as described above except that the cells were
preincubated at 37?C for 1 h and the 42?C incubation
was carried out in 16-ml tubes for 45 min. Following
the 42?C incubation, cells were allowed to recover at
37?C for 30 or 60 min and then harvested with cell
For protein isolation, 65-ml culture flasks were
filled to 40%, 50%, 60%, 70%, 80%, 90%, and 100%
capacity with growth medium and inoculated with 2 ?
107cells total. Cells were grown overnight at 37?C, and
protein was isolated as above. For RNA isolation, 65-
ml culture flasks were filled to 50% capacity with
growth medium (the maximum oxygen exposure toler-
ated well by the cells in the initial experiments) and
inoculated with 1.5 ? 107cells total. Cells were grown
overnight at 37?C, and total RNA was isolated as above.
Rapid Amplification of cDNA Ends (RACE) Analysis
Rapid amplification of cDNA ends (5? RACE) was
employed to identify the start of transcription of the
DnaK-like HSP70. 5? RACE was performed using the
5? RACE System for Rapid Amplification of cDNA
ends, version 2.0 (Life Technologies, Gaithersburg,
Md.), according to the manufacturer’s instructions. Oli-
go HSP70.3 (TGC-TTC-TTA-GCT-GTT-TCT-GCG-C)
was used as the first-strand primer, and oligo HSP70.4
(TTC-ATG-GAC-TTG-GTG-TCT-GG) was used as the
nested primer (Knodler et al. 1999).
A database containing 56 HSP70 homologs from
the Archaea, Bacteria, and Eukarya was assembled from
GenBank. Protein sequences were aligned using CLUS-
TAL W (Thompson, Higgins, and Gibson 1994). The
alignment was refined manually within the SEQLAB
module of the GCG package (Genetics Computing
Group, Madison, Wis.), taking into account structural
elements identified in bovine HSP70 (bhsc70) and Esch-
erichia coli DnaK (Flaherty et al. 1994; Zhu et al. 1996).
Regions of questionable alignment were removed from
the data sets used in phylogenetic analyses.
Phylogenies were inferred using parsimony, dis-
tance, and maximum-likelihood (ML) methods on the
aligned amino acid sequences. Protein ML trees were
inferred using the quick-add OTU (-q option) tree-
searching procedure implemented in the PROTML pro-
gram, version 2.2, using the JTT-F amino acid substi-
tution model (Adachi and Hasegawa 1996). To avoid
inconsistency problems introduced by the presence of
invariable sites in the alignment, the ML estimate of the
proportion of invariable sites (Pinvar) in the HSP70 data
set was obtained using the PUZZLE program, version
4.0 (Strimmer and von Haeseler 1996), and invariable
positions were selectively removed from the alignment
before PROTML analysis. To study the impact of
among-sites rate variation on the phylogenetic trees, the
quartet puzzling ML method was used to infer trees with
a mixed eight-category discrete gamma and invariable-
sites model of rate heterogeneity and the JTT-F substi-
tution model (JTT-F???Inv). Pinvarand the gamma
shape parameter ? were estimated by ML optimization
on a neighbor-joining topology. ML protein distances
were inferred with PUZZLE using the same model of
evolution, and distance trees were estimated using the
Fitch-Margoliash method with global rearrangements
(using the FITCH program of PHYLIP, version 3.57c;
Felsenstein 1996). Unweighted maximum-parsimony
analysis was carried out by 100 rounds of random step-
wise addition heuristic searches with tree bisection-re-
connection (TBR) branch swapping using the PAUP*
program, version 4b.1 (Swofford 1997). Bootstrap anal-
yses for protein distance trees were carried out using the
PUZZLEBOOT program (http://www.tree-puzzle.de/
puzzleboot.sh). Deviations of amino acid frequencies in
a given sequence from the overall frequencies in the data
set were detected by a chi-square test implemented in
Characterization of the G. lamblia HSP70 Gene
The intronless putative coding region is 1,923 bp,
encoding 640 amino acids (GenBank AF274582). Tran-
scriptional start sites were located 13 and 18 bases up-
stream of the start codon (AAATGTAAATTGCGGAC-
CATG). Both transcript start sites are A/T-rich and are
located downstream from another A/T-rich sequence, as
is typical for Giardia (Holberton and Marshall 1995; Hi-
lario and Gogarten 1998). However, no giardial CAT
box sequence, shown to be important for expression of
the glutamate dehydrogenase gene, was present (Yee et
al. 2000). The overall G?C content of the coding se-
quence was 50.13%. The first of two potential polyad-
enylation signals (CGTAAA) (Adam 1991; Que et al.
1996) overlaps the stop codon. The second (TGTAAAT)
is very similar to the polyadenylation signal of the Giar-
dia protein disulfide isomerase-3 gene (Knodler et al.
Total levels of HSP70 protein were constant during
trophozoite growth and encystation (fig. 1A). Giardia is
sensitive to oxygen; however, we found that HSP70 pro-
tein was not significantly increased by sublethal oxygen
by guest on May 30, 2013
Giardia lamblia DnaK Homolog533
FIG. 1.—HSP70 is expressed constitutively and not upregulated
in response to heat shock or oxygen stress. A, Encystation, Western
analysis. Protein was harvested from trophozoites (T) and 24-h en-
cysting (24) cells. B, Oxygen stress, Western analysis. Cells were
grown overnight in flasks filled to 90%, 80%, 70%, 60%, 50%, or 40%
capacity with normal growth medium. The control flask was filled to
100% with normal growth medium. C and D, Heat shock. Cells were
heat-shocked at 42?C for 15, 30, or 45 min and then allowed to recover
for 0, 30, or 60 min before protein (C) or RNA (D) was harvested.
The control cells were incubated at 37?C, while the rest were heat
shocked and recovered. In D, the ratio of HSP70 message to that of
the GDH control did not change significantly with the most severe heat
stress (fig. 1B) or heat shock (fig. 1C). The constant
expression of the DnaK-like HSP70, in response to ox-
ygen stress or heat shock, was confirmed by Northern
analyses with a specific probe relative to a glutamate
dehydrogenase control (fig. 1D and data not shown).
Alignment of the G. lamblia DnaK Homolog
An alignment of the translated protein sequence
with 56 cytosolic, ER, mitochondrial-like, and bacterial
HSP70s from a variety of taxa (fig. 2) revealed that this
G. lamblia DnaK homolog is distinct from the Giardia
cytosolic and ER HSP70s. The amino terminus of the
Giardia DnaK-like HSP70 is hydrophobic, consistent
with potential translocation across an organelle mem-
brane but not with import into the ER (Nielsen et al.
1997). No C-terminal membrane spanning region or sig-
nature sequences related to endoplasmic reticulum tar-
geting or retrieval are obvious. The protein contains
many of the residues conserved among mitochondrial
and ?-proteobacterial HSP70s (Flaherty et al. 1994; Wil-
banks, DeLuca-Flaherty, and McKay 1994; Kamath-
Loeb et al. 1995; O’Brien, Flaherty, and McKay 1996;
Zhu et al. 1996; Suh et al. 1998; Wang et al. 1998), as
indicated in figure 2. However, other signature sequenc-
es (Germot, Philippe, and Le Guyader 1996, 1997; Hirt
et al. 1997; Peyretaillade et al. 1998) are not well con-
served. For example, the signature sequence GDAW,
present in the mitochondrial-like HSP70 of Trichomonas
vaginalis, is replaced by GEAM in Giardia. Similarly,
the signature sequence YSPAQIG is not well conserved
in Giardia, appearing as VSPIEVG.
The Giardia sequence contained several small in-
sertions not found in any other sequence (fig. 2). The
first occurred at residues 49–51, a site that falls between
two ?-strand sheets in bovine hsc70, and the second
occurred at residues 147–149, between a hydrogen-
bonded turn and a ?-strand sheet in bovine hsc70 (fig.
2). A third insertion occurred at residues 199–202; in-
terestingly, the genes for the cytosolic and ER HSP70s
in Giardia also each contain a distinct insertion at this
location. Finally, a larger insertion occurred at residues
263–276. This site falls between a helical region and a
hydrogen-bonded turn in bovine hsc70. These regions
were excluded from the data set used to infer trees, and
their significance is unknown.
Phylogenetic Analysis of the HSP70 Data Set
Protein ML and maximum parsimony trees were
inferred for 56 taxa using 465 reliably aligned sites (fig.
3 and data not shown). Both methods placed Giardia
within the proteobacteria, but not specifically within the
?-proteobacterial or mitochondrial groups. In figure 3,
Giardia DnaK, the mitochondrial HSP70s, and all of the
proteobacterial sequences branched together with a
bootstrap value of 66%. Within this group, the mito-
chondrial HSP70s plus the ?-proteobacterial DnaKs, ex-
cluding the Ehrlichia and microsporidial sequences,
grouped with a bootstrap value of 56%. Giardia did not
branch within this group, nor was there significant boot-
strap support for its placement as a sister group to the
Ehrlichias. The cytosolic and ER proteins grouped to-
gether, with the Giardia cytosolic and ER proteins
branching most deeply. The Giardia protein branched
with the three microsporidial proteins in both analyses;
however, we suspected that this was a problem of long-
branch attraction, and the group did not hold up on fur-
ther examination. Since it was apparent that the new
Giardia sequence was specifically related to the bacterial
DnaK/mitochondrial HSP70 group, the cytosolic and ER
sequences were not included in the later analyses in or-
der to reduce the number of taxa in the data set. This
permitted the inclusion of a larger number of aligned
sites and the use of more rigorous, computationally de-
manding methods of analysis.
by guest on May 30, 2013
534Morrison et al.
FIG. 2.—Conservation of functional residues in the Giardia DnaK-like gene. Positions 1–320 of the full alignment are shown for selected
DnaK and mitochondrial HSP70s. The new Giardia sequence is designated GiardiaDnaK. Annotations: a ? residues important for ATPase
activity (Flaherty et al. 1994; Wilbanks, DeLuca-Flaherty, and McKay 1994; O’Brien, Flaherty, and McKay 1996); b ? residues important for
substrate binding (Zhu et al. 1996; Wang et al. 1998); j ? residues important for DnaJ binding (Kamath-Loeb et al. 1995; Suh et al. 1998); ^
? mitochondrial/proteobacterial signature sequences (Germot, Philippe, and Le Guyader 1996, 1997; Hirt et al. 1997); INS ? insertions unique
to Giardia lamblia proteins; MITO/PROT ? consensus (80%) of mitochondrial and ?-proteobacterial HSP70s; CONSERVED indicates residues
conserved in at least 80% of the reference HSP70s (table 1). Residues in uppercase are conserved in all reference sequences.
Phylogenetic Analysis of a Restricted Data Set
Based on the preliminary analysis, we created a
smaller data set for more intensive analysis by selecting
taxa that branched near the Giardia sequence. These taxa
included mitochondrial and proteobacterial sequences,
with the Chlamydia and spirochetes as immediate out-
groups. We performed likelihood and distance analyses
on this data set.
In analyses without a correction for among-sites
rate variation (protein ML, quartet puzzling ML, and
ML distance), the Giardia sequence grouped with
HSP70 sequences from the microsporidia, Nosema lo-
custae, and Encephalitozoon cuniculi. This group gained
strong bootstrap support from PROTML, distance, and
quartet puzzling analyses (fig. 4A). However, the exact
placement of this group in the HSP70 tree was poorly
resolved—the ML tree displayed the microsporidia/
Giardia group outside the ?-proteobacterial/mitochon-
drial clade (not shown), while distance analyses placed
it at the base of mitochondria (fig. 4A). Each of these
placements was weakly supported by the respective
methods; bootstrap values and puzzling support values
were below 50% in all cases. Indeed, the quartet puz-
zling consensus tree showed the Giardia, microsporidi-
an, mitochondrial, and ?-proteobacterial sequences in an
unresolved polytomy (not shown).
Not correcting for among-sites rate variation can
lead to problems whereby sequences with higher rates
of evolution artifactually group together (the ‘‘long-
branch attraction’’ problem; Lockhart et al. 1996; Yang
1996). Since the microsporidian and Giardia sequences
formed extremely long branches in our analyses (fig.
4A), we suspected that their strong association could be
due to this kind of artifact. To test this hypothesis, we
performed ML distance and quartet puzzling ML anal-
yses using the gamma model of among-sites rate vari-
ation implemented in PUZZLE, version 4.0. Consistent
with the artifact explanation, trees recovered from these
by guest on May 30, 2013
Giardia lamblia DnaK Homolog535
FIG. 3.—The phylogenetic placement of the Giardia DnaK-like sequence among the HSP70 cytosolic, ER, and mitochondrial/bacterial
homologs. This tree was constructed from 465 positions of an aligned data set of 56 HSP70 sequences with protein maximum likelihood analysis
using the PROTML program, version 2.2 (Adachi and Hasegawa 1996), with the Jones, Taylor, and Thornton amino acid replacement model
adjusted for amino acid frequencies (JTT-F). Tree-searching employed the quick-add OTU stepwise-addition procedure (-q options) with 2,000
iterations. Bootstrap values were estimated using the resampling estimated log likelihood (RELL) method implemented in this program and
compiled using the mol2con PERL script. Species names and accession numbers for HSP70 sequences used: Agrobacterium tumifaciens DnaK
X87113; Aquifex aeolicus DnaK AE000717; Bacillus stearothermophilus DnaK X90709; Borrelia burgdorferi DnaK AE001154; Bradyrhizobium
japonicum DnaK Y09633; Buchnera sp. DnaK D88673; Caulobacter crescentus DnaK M55224; Chlamydia pneumoniae DnaK M69227; Chla-
mydia trachomatis DnaK L22180; Ehrlichia sennetsu DnaK AF060197; Ehrlichia sp. DnaK AF02931; Escherichia coli DnaK AE000112;
Francisella tularensis DnaK L43367; Haemophilus ducreyi DnaK U25996; Haemophilus influenzae DnaK U32803; Halobacterium cutirubrum
DnaK L35530; Legionella pneumophila DnaK D89498; Leptospira interrogans DnaK AF007813; Methanobacterium thermoautotrophicum
DnaK AE000894; Mycobacterium avium paratuberculosis DnaK X59437; Rhizobium meliloti DnaK L36602; Rhodobacter capsulatus DnaK
U57637; Rhodopseudomonas sp. DnaK D78133; Rickettsia prowazekii DnaK AJ235270; Thermoplasma acidophilum DnaK L35529; Caenorhab-
ditis elegans mitochondrial form U88315; Drosophila melanogaster mitochondrial form L01502; Eimeria tenella mitochondrial form Z46965;
Encephalitozoon cuniculi mitochondrial form AJ012470; Leishmania major mitochondrial form X64147; Mus musculus mitochondrial form
D17556; Nosema locustae mitochondrial form U97520; Phaseolus vulgaris mitochondrial form X66874; Pisum sativum mitochondrial form
X54739; Saccharomyces cerevisiae mitochondrial form M27229; Schizosaccharomyces pombe mitochondrial form M60208; Trichomonas va-
ginalis 1 mitochondrial form U70308; T. vaginalis 2 mitochondrial form U27231; Trypanosoma cruzi mitochondrial form M73627; Vairimorpha
necatrix mitochondrial form AF008215; Aspergillus awamorii ER form Y08867; Drosophila melanogaster ER form L01498; Giardia intestinalis
ER form U04875; Zea mays ER form U58209; Blastocladiella emersonii cytoplasmic form L22497; Bos taurus cytoplasmic form X53827; C.
elegans cytoplasmic form M18540; D. melanogaster cytoplasmic form L01501; G. intestinalis cytoplasmic form U04874; L. major cytoplasmic
form X69825; M. musculus cytoplasmic form M35021; P. sativum cytoplasmic form X99515; S. cerevisiae cytoplasmic form M97225; S. pombe
cytoplasmic form AB012387; T. vaginalis cytoplasmic form U93873.
analyses did not show a specific association between the
microsporidian and the Giardia sequences (fig. 4B).
However, the exact placement of Giardia and microspo-
ridia in these trees was again not well resolved. In the
distance analysis, Giardia fell outside the ?-proteobac-
terial/mitochondrial clade, with microsporidia emerging
at the base of mitochondria. Bootstrap and quartet puz-
zling support values for branches separating Giardia
from the mitochondrial sequences were all below 50%.
One source of instability in these analyses was the two
Ehrlichia sequences. Although Ehrlichia are well known
to be a sister group of the Rickettsia (Weisburg et al.
by guest on May 30, 2013
536Morrison et al.
FIG. 4.—The phylogenetic placement of the Giardia DnaK-like sequence among mitochondrial and proteobacterial homologs. A, The Fitch-
Margoliash tree inferred from maximum-likelihood (ML) distance analysis with an equal-rates model of amino acid replacement (the JTT model).
RELL bootstrap values from PROTML analysis, ML quartet puzzling support values, and distance bootstrap values are shown above the branches.
B, The Fitch-Margoliash tree inferred from gamma-corrected ML distances (using the JTT?? model). The estimated shape parameter for the
gamma distribution, alpha, is 0.59. ML quartet puzzling support values and distance bootstrap values are shown above the branches. Bootstrap
or quartet puzzling support values of ?50% are indicated with hyphens.
1989; Roger et al. 1998), in our various HSP70 analyses
the two Ehrlichia sequences moved around, sometimes
grouping together weakly (53% bootstrap support in
PROTML analysis) and sometimes splitting apart, not
always specifically allied with the Rickettsia sequence
To examine the placement of the Giardia sequence
without the confounding bias and instability introduced
by the microsporidian and Ehrlichia sequences, we re-
moved them from the alignment and applied gamma-
corrected ML distance and ML quartet puzzling meth-
ods to the data. The quartet-puzzling consensus tree
by guest on May 30, 2013
Giardia lamblia DnaK Homolog537
FIG. 5.—The phylogenetic placement of the Giardia DnaK-like sequence with microsporidian and Ehrlichia sequences removed. This tree
was constructed by obtaining the optimal Fitch-Margoliash tree, moving the Giardia sequence around to all possible positions on the constrained
backbone of this tree, and evaluating the likelihood of each using the JTT?? model implemented in the PUZZLE program. The gamma shape
parameter used in both distance and likelihood analyses was ? ? 0.58 (derived by maximum-likelihood optimization). The optimal tree (lnL ?
?10,032.45) obtained by this procedure is shown. The second most likely branching position of the Giardia sequence on the same backbone
topology is indicated by the dashed line (lnL ? ?10,032.86); it had a minor difference from the optimal topology of ?lnL ? 0.4. Kishino-
Hasegawa tests (Kishino and Hasegawa 1989) were performed on each placement in comparison to the optimal tree, and only two branching
positions of the Giardia sequence were significantly excluded (indicated by an asterisk). Maximum-likelihood quartet puzzling support values
and distance bootstrap values are shown above the branches.
showed an unresolved polytomy containing mitochon-
drial plus ?-proteobacterial, Giardia, and ?-proteobac-
terial sequences indicating little resolution in the data.
Although the support for the Giardia/proteobacteria/mi-
tochondria clade was strong (quartet puzzling support ?
90%, ML distance bootstrap ? 90%) both trees showed
Giardia being outside the mitochondrial/?-proteobacter-
ial clade, which was moderately supported (quartet puz-
zling support ? 74%, bootstrap support ? 70%). To
determine whether this placement outside of the mito-
chondrial/?-proteobacterial clade was significant, we
moved the position of the Giardia sequence to all pos-
sible branching positions in the backbone ML distance
Fitch-Margoliash tree and evaluated the likelihood score
for each position under the JTT?? model. The optimal
position in ML analysis, shown in figure 5, displayed
the Giardia sequences as an immediate outgroup to the
?-proteobacteria plus mitochondria. However, the next
most likely position of Giardia in this tree was as a sister
group to mitochondria, with a difference in log likeli-
hood of 0.4 compared with the optimal tree (fig. 5). To
test whether the alternative positions for Giardia were
significantly worse, we used Kishino-Hasegawa tests
(Kishino and Hasegawa 1989) to evaluate the signifi-
cance of differences in likelihood for alternative topol-
ogies. These tests indicated that of all the possible place-
ments of Giardia in the tree, only two could be signifi-
cantly excluded at the 5% level (indicated branches in
Therefore, since the branching point of the Giardia
HSP70 can be moved to almost anywhere in the tree
without a significant decrease in the likelihood, it ap-
pears that the data cannot resolve its true position. This
is at least partly due to the extremely long branch lead-
ing to Giardia HSP70 in trees. Most phylogenetic signal
that would allow a clear placement of this sequence in
the HSP70 tree has been erased by the extreme diver-
gence of this sequence. This is in sharp contrast to the
Giardia chaperonin 60 that shows a strongly supported
relationship to the mitochondrial lineage (Roger et al.
1998). HSP70 and cpn60 work sequentially in refolding
proteins during organellar import (Stuart et al. 1994),
and both are of mitochondrial origin in all eukaryotes
studied so far, including the amitochondriate parabasal-
ids (Bui, Bradley, and Johnson 1996; Germot, Philippe,
and Le Guyader 1996; Horner et al. 1996; Roger et al.
1998). While a mitochondrial origin for the Giardia
DnaK is not specifically supported, it nevertheless re-
mains a possibility.
The Impact of Taxon Deletion on the Gamma Shape
The gamma distribution model of sequence evolu-
tion assumes that sites in a gene evolve at different rates
by guest on May 30, 2013
538Morrison et al.
Impact of Taxon Deletion on the Gamma Shape
Parameter (?) Estimate
None . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Rickettsia, Ehrlichia sp. . . . . . . . . . . . . . . . . .
Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Plants, Eimeria . . . . . . . . . . . . . . . . . . . . . . . .
Ehrlichia . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
?-proteobacteria . . . . . . . . . . . . . . . . . . . . . . .
Rhodobacter, Caulobacter . . . . . . . . . . . . . . .
Spirochetes . . . . . . . . . . . . . . . . . . . . . . . . . . .
Agrobacterium, Bradyrhizobium . . . . . . . . . .
Chlamydia . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fungi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Trichomonas . . . . . . . . . . . . . . . . . . . . . . . . . .
Microsporidia . . . . . . . . . . . . . . . . . . . . . . . . .
Trypanosomes . . . . . . . . . . . . . . . . . . . . . . . . .
Giardia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Giardia, Microsporidia . . . . . . . . . . . . . . . . . .
NOTE.—To ensure that missing data did not affect the parameter estimates,
alignment positions containing missing data were removed from the data set,
leaving 47 sites for this analysis. For each data set, ? was estimated by maxi-
mum-likelihood optimization on neighbor-joining topology.
but that those rates do not change over the tree. The
gamma shape parameter, alpha, ranges from ? ? infinity
(no rate heterogeneity) to ? ? 1 (strong heterogeneity).
The value of alpha estimated for the whole data set was
0.6, indicating strong heterogeneity in rates of evolution
at different sites. The highly divergent nature of the
Giardia sequence (see fig. 2) suggested that the pattern
of rates at sites may have changed in this lineage. If so,
the value of alpha should change if Giardia is excluded
from the data set. To test this hypothesis, we investi-
gated the impact of the presence or absence of sequences
on the estimate of the gamma shape parameter by sys-
tematically deleting proteobacterial or mitochondrial
taxa from our analyses and reestimating the shape pa-
rameter (table 1). The deletion of most of these taxa had
little effect on the estimate of the shape parameter,
which remained close to 0.6, the value estimated for the
whole data set. However, deletion of several groups had
a much larger effect. As expected, removing Giardia
produced the largest effect, yielding an alpha estimate
of 0.55, while deletion of trypanosomes and the mi-
crosporidia had the next largest effects, also yielding
lower alpha estimates of 0.56 and 0.57, respectively.
Since the microsporidian and Giardia HSP70 sequences
appear to artifactually group together in some analyses
(see above), we examined the effect of deleting both.
This caused the largest decrease in the alpha estimate,
with ? ? 0.51.
Combining sets of sequences with different patterns
of among-sites rate variation leads to an ‘‘averaging’’
effect on among-sites rate variation for the whole data
set. Under these conditions, the estimate for the gamma
shape parameter, alpha, is increased, falsely indicating
less extreme rate variation among sites for the whole
data set than is displayed by either set of sequences
alone (Lockhart et al. 1998). Conversely, in our analy-
ses, the decreases in the alpha parameter estimates upon
removal of the Giardia, microsporidia, and trypanosome
HSP70s indicate that these sequences have a different
pattern of rates of sites than the rest of the HSP70 data
set. Consideration of the alignment (fig. 2) suggests that
the Giardia and microsporidia sequences, in particular,
have fewer sequence constraints and are more variable
in general than other proteobacterial and mitochondrial
HSP70s (as might be expected for a gene whose func-
tional constraints have been relaxed). It is possible that
the decrease in constraints in the HSP70s from these
amitochondriate groups, coupled with an elevated rate
of evolution, contributes to their artifactual grouping in
some of our phylogenetic analyses (fig. 4A). Further-
more, the changes in the pattern of substitution in these
taxa violate both the equal-rates and the gamma distri-
bution models we used and may be partly responsible
for the difficulties encountered in discerning the true
phylogenetic placement of the Giardia HSP70 sequence.
Genes encoding HSP70 have been extensively used
as phylogenetic markers. They meet several criteria for
useful markers: they are conserved in all domains of
life, they can be aligned based on highly conserved sites
and functional domains, they contain enough variable
sites to provide phylogenetic signal, and they have un-
dergone gene duplication events, so that gene trees can
be reciprocally rooted. However, it is becoming increas-
ingly evident that HSP70 genes do not have the power
to resolve distant phylogenetic relationships, particularly
in cases where the genes are serving different functions
in different organisms. For example, in recently pub-
lished analyses of HSP70 (Germot and Philippe 1999),
the position of T. vaginalis could not be reliably deter-
mined, appearing in some analyses as a sister group to
Giardia, in others as a sister group to fungi/metazoa. The
DnaK HSP70 homologs, in particular, appear unable to
resolve the relationships between lineages that are either
deep-branching or fast-evolving. Ideally, genomic se-
quence for many of these lineages will one day be avail-
able to clarify these relationships.
An objective look at the data cited as evidence that
Giardia once had mitochondria reveals that they are not
robust. Henze et al. (1995), in a study of glyceralde-
hyde-3-phosphate dehydrogenase, considered the possi-
bility that Giardia was secondarily amitochondriate.
However, they correctly point out that the argument is
based solely on the presence of a eubacterial GAPDH
gene. Keeling and Doolittle (1997) concluded that Giar-
dia’s triosephosphate isomerase is of ?-proteobacterial
origins. While their conclusion may be correct, it is a
strong statement to make based on a problematic anal-
ysis that showed (1) no bootstrap support for the mono-
phyly of the eukaryotes, yet subsequent treatment of the
eukaryotes as a single group; (2) a single representative
of the ?-proteobacteria in the data set used, and (3) non-
significantly different tree topologies when the outgroup
to the eukaryotes was all proteobacteria versus ?-
Other analyses, in which a sister taxon relationship
has been observed between Giardia and Trichomonas,
by guest on May 30, 2013
Giardia lamblia DnaK Homolog539
i.e., analyses of valyl-tRNA synthetase (Hashimoto et
al. 1998), beta-tubulin (Keeling and Doolittle 1996), and
cpn60 (Roger et al. 1998), have been taken as evidence
that if Trichomonas once harbored the mitochondrial en-
dosymbiont, Giardia must have done so as well (Roger
1999; Roger, Morrison, and Sogin 1999). However, a
sister relationship of the diplomonads and parabasalids
is not seen in analyses of other proteins, such as alpha
tubulins (Keeling and Doolittle 1996) or the cytosolic
and ER forms of HSP70 (Germot and Philippe 1999).
The credibility of the sister relationship seen in the anal-
ysis of ValRS is weakened by the limited number of
taxa included and the observed specific relationship of
the Trichomonas/Giardia grouping to the green plants.
This is not consistent with any other published molec-
ular analysis, nor is there any phenotypic basis for de-
scribing diplomonads as being specifically related to
photosynthetic chlorophytes or green plants. The sister
group relationship might be artifactual, an example of
the long-branch attraction problem, in which case this
line of reasoning becomes invalid. The sister relation-
ship hypothesis should be tested using additional pro-
tein-coding sequences as such data become available for
both diplomonads and parabasalids.
The strongest evidence thus far that Giardia once
harbored a mitochondrial symbiont comes from analysis
of cpn60 (Roger et al. 1998). If this is true, the simplest
explanation for the origin of the DnaK-like HSP 70,
given its affinity for the proteobacterial DnaK homologs,
is that it is also a mitochondrial relic. The cpn60 phy-
logeny indicates a specific relationship of Giardia’s
cpn60 to the ?-proteobacterial and mitochondrial cpn60.
The gene tree also shows evidence of a sister group
relationship between Trichomonas and Giardia, although
Entamoeba also falls within this clade. The inclusion of
the Entamoeba sequence was considered an artifact
caused by long-branch attraction and by a statistically
significant amino acid composition bias shared by Giar-
dia and Entamoeba (Roger et al. 1998). If the putative
sister relationship is true, it is interesting that the cpn60
phylogeny apparently can resolve deep relationships
while HSP70 cannot. Neither the present function nor
the cellular localization is known for either cpn60 or this
DnaK-like HSP70. If both are of mitochondrial origin,
presumably they have been evolving for the same
amount of time but not under the same constraints,
which suggests that they function independently.
Genes have been discovered in amitochondriate lin-
eages that strongly suggest the occurrence of multiple
lateral transfer or endosymbiotic events from bacteria
other than the ?-proteobacterial symbiont that gave rise
to the mitochondrion. One example in Giardia is a class
2 3-hydroxy-3-methylglutaryl coenzyme A reductase
(Boucher and Doolittle 2000) which appears to have
been acquired from a bacterial source distinct from the
presumptive ?-proteobacterial symbiont. The discovery
in Giardia of an iron-hydrogenase gene most closely re-
lated to a gene found in Entamoeba argues that a lateral
gene transfer event may have occurred, with the possible
source being an anaerobic bacterium living in animal
digestive tracts (J. Nixon et al., personal communica-
tion). It has been noted by others that Giardia’s residence
in the intestinal environment would make it easy to ac-
quire genes from bacteria (Boucher and Doolittle 2000).
It is likely that many other genes of bacterial origin will
be evident in Giardia’s genome when the genome se-
quence data are complete. Even if Giardia is secondarily
amitochondriate, not all of its bacterial-like genes are
necessarily mitochondrial relics. It will be interesting
and informative to determine what proportion of Giar-
dia’s bacterial relics are specifically related to ?-proteo-
The existing data, including the results reported
here, clearly demonstrate that the genomes of amito-
chondriate organisms contain genes of endosymbiotic
origin. However, phylogenetic inferences do not strong-
ly support the most parsimonious hypothesis that a sin-
gle early endosymbiotic event (the mitochondria-early
hypothesis) accounts for the origins of both cpn60 and
DnaK-like HSP70s in amitochondrial lineages. Only the
cpn60 gene shows a specific affinity with homologs
from the mitochondria. The DnaK-like HSP70 genes
merely suggest an unresolved phylogenetic affinity with
proteobacterial and mitochondrial DnaK forms. Evolu-
tionary history has been obscured by rapid evolution of
the genes in amitochondriate taxa, and the present func-
tions of these genes in these taxa are not known.
The complete amino acid alignment has been sub-
mitted to EMBL under alignment number DS44922.
Polyclonal antibodies to yeast ssc1 were a kind gift
from Dr. Elizabeth Craig, University of Chicago. This
work was supported by the G. Unger Vettlesen Foun-
dation; grants GM32964 and AI43273 from the National
Institutes of Health, awarded to M.L.S; and grant
AI42488 from the NIH, awarded to F.D.G. Additional
support for the Giardia genome project was provided by
LI-COR Biotechnology, Inc.
ADACHI, J., and M. HASEGAWA. 1996. MOLPHY version 2.3:
programs for molecular phylogenetics based on maximum
likelihood. Comput. Sci. Monogr. 28:1–150.
ADAM, R. D. 1991. The biology of Giardia spp. Microbiol.
———. 2000. The Giardia lamblia genome. Int. J. Parasitol.
ALTSCHUL, S. F., W. GISH, W. MILLER, E. W. MYERS, and D.
J. LIPMAN. 1990. Basic local alignment search tool. J. Mol.
BOUCHER, S. E., and F. D. GILLIN. 1990. Excystation of in
vitro-derived Giardia lamblia cysts. Infect. Immun. 58:
BOUCHER, Y., and W. F. DOOLITTLE. 2000. The role of lateral
gene transfer in the evolution of isoprenoid biosynthesis
pathways. Mol. Microbiol. 37:703–716.
BUI, E. T., P. J. BRADLEY, and P. J. JOHNSON. 1996. A common
evolutionary origin for mitochondria and hydrogenosomes.
Proc. Natl. Acad. Sci. USA 93:9651–9656.
by guest on May 30, 2013
540Morrison et al.
CAVALIER-SMITH, T., and E. E. CHAO. 1996. Molecular phy-
logeny of the free-living archezoan Trepomonas agilis and
the nature of the first eukaryote. J. Mol. Evol. 43:551–562.
DIAMOND, L. S., D. R. HARLOW, and C. C. CUNNICK. 1978. A
new medium for the axenic cultivation of Entamoeba his-
tolytica and other Entamoeba. Trans. R. Soc. Trop. Med.
EY, P. L., T. BRUDERER, C. WEHRLI, and P. KOHLER. 1996.
Comparison of genetic groups determined by molecular and
immunological analyses of Giardia isolated from animals
and humans in Switzerland and Australia. Parasitol. Res.
FELSENSTEIN, J. 1996. PHYLIP (phylogeny inference package).
Version 3.57c. Distributed by the author, Department of Ge-
netics, University of Washington, Seattle.
FLAHERTY, K. M., S. M. WILBANKS, C. DELUCA-FLAHERTY,
and D. B. MCKAY. 1994. Structural basis of the 70-kilo-
dalton heat shock cognate protein ATP hydrolytic activity.
II. Structure of the active site with ADP or ATP bound to
wild type and mutant ATPase fragment. J. Biol. Chem. 269:
GERMOT, A., and H. PHILIPPE. 1999. Critical analysis of eu-
karyotic phylogeny: a case study based on the HSP70 fam-
ily. J. Eukaryot. Microbiol. 46:116–124.
GERMOT, A., H. PHILIPPE, and H. LE GUYADER. 1996. Presence
of a mitochondrial-type 70-kDa heat shock protein in Trich-
omonas vaginalis suggests a very early mitochondrial en-
dosymbiosis in eukaryotes. Proc. Natl. Acad. Sci. USA 93:
———. 1997. Evidence for loss of mitochondria in Microspo-
ridia from a mitochondrial-type HSP70 in Nosema locustae.
Mol. Biochem. Parasitol. 87:159–168.
GUPTA, R. S., K. AITKEN, M. FALAH, and B. SINGH. 1994.
Cloning of Giardia lamblia heat shock protein HSP70 ho-
mologs: implications regarding origin of eukaryotic cells
and of endoplasmic reticulum. Proc. Natl. Acad. Sci. USA
HASHIMOTO, T., and M. HASEGAWA. 1996. Origin and early
evolution of eukaryotes inferred from the amino acid se-
quence of translation elongation factors 1?/Tu and 2/G.
Adv. Biophys. 32:73–120.
HASHIMOTO, T., L. B. SA´NCHEZ, T. SHIRAKURA, M. MU¨LLER,
and M. HASEGAWA. 1998. Secondary absence of mitochon-
dria in Giardia lamblia and Trichomonas vaginalis revealed
by valyl-tRNA synthetase phylogeny. Proc. Natl. Acad. Sci.
HENZE, K., A. BADR, M. WETTERN, R. CERFF, and W. MARTIN.
1995. A nuclear gene of eubacterial origin in Euglena grac-
ilis reflects cryptic endosymbioses during protist evolution.
Proc. Natl. Acad. Sci. USA 92:9122–9126.
HENZE, K., H. G. MORRISON, M. L. SOGIN, and M. MU¨LLER.
1998. Sequence and phylogenetic position of a class II al-
dolase gene in the amitochondriate protist, Giardia lamblia.
HILARIO, E., and J. P. GOGARTEN. 1998. The prokaryote-to-
eukaryote transition reflected in the evolution of the V/F/
A-ATPase catalytic and proteolipid subunits. J. Mol. Evol.
HIRT, R. P., B. HEALY, C. R. VOSSBRINCK, E. U. CANNING, and
T. M. EMBLEY. 1997. Identification of a mitochondrial
HSP70 homologue in Vairimorpha necatrix: molecular ev-
idence that microsporidia once contained mitochondria.
Curr. Biol. 7:995–998.
HOLBERTON, D. V., and J. MARSHALL. 1995. Analysis of con-
sensus sequence patterns in Giardia cytoskeleton gene pro-
moters. Nucleic Acids Res. 23:2945–2953.
HORNER, D. S., R. P. HIRT, S. KILVINGTON, D. LLOYD, and T.
M. EMBLEY. 1996. Molecular data suggest an early acqui-
sition of the mitochondrion endosymbiont. Proc. R. Soc.
Lond. B Biol. Sci. 263:1053–1059.
KAMATH-LOEB, A. S., C. Z. LU, W. C. SUH, M. A. LONETTO,
and C. A. GROSS. 1995. Analysis of three DnaK mutant
proteins suggests that progression through the ATPase cycle
requires conformational changes. J. Biol. Chem. 270:
KEELING, P. J., and W. F. DOOLITTLE. 1996. Alpha-tubulin from
early-diverging eukaryotic lineages and the evolution of the
tubulin family. Mol. Biol. Evol. 13:1297–1305.
———. 1997. Evidence that eukaryotic triosephosphate isom-
erase is of alpha-proteobacterial origin. Proc. Natl. Acad.
Sci. USA 94:1270–1275.
KEISTER, D. B. 1983. Axenic culture of Giardia lamblia in
TYI-S-33 medium supplemented with bile. Trans. R. Soc.
Trop. Med. Hyg. 77:487–488.
KISHINO, H., and M. HASEGAWA. 1989. Evaluation of the max-
imum likelihood estimate of the evolutionary tree topolo-
gies from DNA sequence data, and the branching order in
Hominoidea. J. Mol. Evol. 29:170–179.
KNODLER, L. A., R. NOIVA, K. MEHTA, J. M. MCCAFFERY, S.
B. ALEY, S. G. SVARD, T. G. NYSTUL, D. S. REINER, J. D.
SILBERMAN, and F. D. GILLIN. 1999. Novel protein-disulfide
isomerases from the early-diverging protist Giardia lam-
blia. J. Biol. Chem. 274:29805–29811.
LAEMMLI, U. K. 1970. Cleavage of structural proteins during
the assembly of the head of bacteriophage T4. Nature 227:
LEIPE, D. D., J. H. GUNDERSON, T. A. NERAD, and M. L. SOGIN.
1993. Small subunit ribosomal RNA of Hexamita inflata
and the quest for the first branch in the eukaryotic tree. Mol.
Biochem. Parasitol. 59:41–48.
LOCKHART, P. J., A. W. LARKUM, M. STEEL, P. J. WADDELL,
and D. PENNY. 1996. Evolution of chlorophyll and bacte-
riochlorophyll: the problem of invariant sites in sequence
analysis. Proc. Natl. Acad. Sci. USA 93:1930–1934.
LOCKHART, P. J., M. A. STEEL, A. C. BARBROOK, D. H. HUSON,
M. A. CHARLESTON, and C. J. HOWE. 1998. A covariotide
model explains apparent phylogenetic structure of oxygenic
photosynthetic lineages. Mol. Biol. Evol. 15:1183–1188.
MENG, T. C., M. L. HETSKO, and F. D. GILLIN. 1996. Inhibition
of Giardia lamblia excystation by antibodies against cyst
walls and by wheat germ agglutinin. Infect. Immun. 64:
MIDDENDORF, L. R., J. C. BRUCE, R. C. BRUCE et al. (11 co-
authors). 1992. Continuous, on-line DNA sequencing using
a versatile infrared laser scanner/electrophoresis apparatus.
MORIN, L. 2000. Long branch attraction effects and the status
of ‘‘basal eukaryotes’’: phylogeny and structural analysis of
the ribosomal RNA gene cluster of the free-living diplo-
monad Trepomonas agilis. J. Eukaryot. Microbiol. 47:167–
NIELSEN, H., J. ENGELBRECHT, S. BRUNAK, and G. VON HEIJNE.
1997. Identification of prokaryotic and eukaryotic signal
peptides and prediction of their cleavage sites. Protein Eng.
O’BRIEN, M. C., K. M. FLAHERTY, and D. B. MCKAY. 1996.
Lysine 71 of the chaperone protein Hsc70 is essential for
ATP hydrolysis. J. Biol. Chem. 271:15874–15878.
PEYRETAILLADE, E., V. BROUSSOLLE, P. PEYRET, G. METENIER,
M. GOUY, and C. P. VIVARES. 1998. Microsporidia, amito-
chondrial protists, possess a 70-kDa heat shock protein gene
of mitochondrial evolutionary origin. Mol. Biol. Evol. 15:
by guest on May 30, 2013
Giardia lamblia DnaK Homolog 541
QUE, X., S. G. SVARD, T. C. MENG, M. L. HETSKO, S. B. ALEY,
and F. D. GILLIN. 1996. Developmentally regulated tran-
scripts and evidence of differential mRNA processing in
Giardia lamblia. Mol. Biochem. Parasitol. 81:101–110.
ROGER, A. J. 1999. Reconstructing early events in eukaryotic
evolution. Am. Nat. 154:S146–S163.
ROGER, A. J., C. G. CLARK, and W. F. DOOLITTLE. 1996. A
possible mitochondrial gene in the early-branching amito-
chondriate protist Trichomonas vaginalis. Proc. Natl. Acad.
Sci. USA 93:14618–14622.
ROGER, A. J., H. G. MORRISON, and M. L. SOGIN. 1999. Pri-
mary structure and phylogenetic relationships of a malate
dehydrogenase gene from Giardia lamblia. J. Mol. Evol.
ROGER, A. J., S. G. SVARD, J. TOVAR, C. G. CLARK, M. W.
SMITH, F. D. GILLIN, and M. L. SOGIN. 1998. A mitochon-
drial-like chaperonin 60 gene in Giardia lamblia: evidence
that diplomonads once harbored an endosymbiont related to
the progenitor of mitochondria. Proc. Natl. Acad. Sci. USA
SMITH, P. D., F. D. GILLIN, W. M. SPIRA, and T. E. NASH. 1982.
Chronic giardiasis: studies on drug sensitivity, toxin pro-
duction, and host immune response. Gastroenterology 83:
STILLER, J. W., and B. D. HALL. 1999. Long-branch attraction
and the rDNA model of early eukaryotic evolution. Mol.
Biol. Evol. 16:1270–1279.
STRIMMER, K., and A. VON HAESELER. 1996. Quartet puzzling:
a quartet maximum-likelihood method for reconstructing
tree topologies. Mol. Biol. Evol. 13:964–969.
STUART, R. A., D. M. CYR, E. A. CRAIG, and W. NEUPERT.
1994. Mitochondrial molecular chaperones: their role in
protein translocation. Trends Biochem. Sci. 19:87–92.
SUH, W. C., W. F. BURKHOLDER, C. Z. LU, X. ZHAO, M. E.
GOTTESMAN, and C. A. GROSS. 1998. Interaction of the
Hsp70 molecular chaperone, DnaK, with its cochaperone
DnaJ. Proc. Natl. Acad. Sci. USA 95:15223–15228.
SWOFFORD, D. L. 1997. PAUP*: phylogenetic analysis using
parsimony (*and other methods). Prerelease version. Sin-
auer, Sunderland, Mass.
THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON. 1994.
CLUSTAL W: improving the sensitivity of progressive mul-
tiple sequence alignment through sequence weighting, po-
sition-specific gap penalties and weight matrix choice. Nu-
cleic Acids Res. 22:4673–4680.
TOWBIN, H., T. STAEHELIN, and J. GORDON. 1979. Electropho-
retic transfer of proteins from polyacrylamide gels to nitro-
cellulose sheets: procedure and some applications. Proc.
Natl. Acad. Sci. USA 76:4350–4354.
WANG, H., A. V. KUROCHKIN, Y. PANG, W. HU, G. C. FLYNN,
and E. R. ZUIDERWEG. 1998. NMR solution structure of the
21 kDa chaperone protein DnaK substrate binding domain:
a preview of chaperone-protein interaction. Biochemistry
WEISBURG, W. G., M. E. DOBSON, J. E. SAMUEL, G. A. DASCH,
L. P. MALLAVIA, O. BACA, L. MANDELCO, J. E. SECHREST,
E. WEISS, and C. R. WOESE. 1989. Phylogenetic diversity
of the Rickettsiae. J. Bacteriol. 171:4202–4206.
WILBANKS, S. M., C. DELUCA-FLAHERTY, and D. B. MCKAY.
1994. Structural basis of the 70-kilodalton heat shock cog-
nate protein ATP hydrolytic activity. I. Kinetic analyses of
active site mutants. J. Biol. Chem. 269:12893–12898.
YANG, Z. 1996. Among-site rate variation and its impact on
phylogenetic analyses. Trends Ecol. Evol. 11:367–372.
YEE, J., M. R. MOWATT, P. P. DENNIS, and T. E. NASH. 2000.
Transcriptional analysis of the glutamate dehydrogenase
gene in the primitive eukaryote, Giardia lamblia. J. Biol.
ZHU, X., X. ZHAO, W. F. BURKHOLDER, A. GRAGEROV, C. M.
OGATA, M. E. GOTTESMAN, and W. A. HENDRICKSON. 1996.
Structural analysis of substrate binding by the molecular
chaperone DnaK. Science 272:1606–1614.
JULIAN ADAMS, reviewing editor
Accepted November 24, 2000
by guest on May 30, 2013