cimp1, A Novel Astacin Family Metalloproteinase Gene from East African
Cichlids, Is Differentially Expressed Between Species During Growth
Teiya Kijimoto,* Masakatsu Watanabe,* Koji Fujimura,* Masumi Nakazawa, Yasunori Murakami, Shigeru Kuratani, Yuji Kohara,à Takashi Gojobori,à and
Norihiro Okada*§
*Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology, 4259-B21, Nagatsuta-cho, Midori-ku,
Yokohama, 226-8501, Japan; Laboratory for Evolutionary Morphology, Center for Developmental Biology, RIKEN Kobe,
2-2-3 Minatojima-minami, Kobe 650-0047, Japan; àNational Institute of Genetics, Yata, 1111, Mishima, Shizuoka 411-8540,
Japan; and §National Institute for Basic Biology, Okazaki, 444-8585, Japan
Lake Victoria cichlid fishes are excellent examples of explosive adaptive radiation. Although Lake Victoria cichlids are
believed to have arisen during a short period (;14,000 years), they have various species-specific phenotypes. One important phenotype that distinguishes each species is the shape of the jaw, which has diverged to adapt to the wide variety of
trophic habitats present in the lake. Here we demonstrate a new approach to investigate the diversification of cichlid jaw
morphology at the genetic level by examining differentially expressed genes. We used a DNA chip to compare gene
expression levels between closely related cichlid fishes. This analysis indicated that the expression of some genes differed
in the larvae of two cichlid species. One such clone encodes a new astacin family metalloproteinase. The expression level of
the isolated gene, named cimp1, was analyzed in more detail by real-time quantitative reverse transcription-polymerase
chain reaction. A significant difference in cimp1 expression was observed between two Haplochromis cichlid species
during development. Using in situ hybridization, we found that this gene is expressed only in head and gill epithelia.
Biochemical analysis showed that cichlid metalloproteinase 1 (CiMP1) has proteolytic activity, a common attribute of
all astacin family proteins. Because some astacin family proteins contribute to morphogenesis in animals, CiMP1 is expected to participate in species-specific head morphogenesis in cichlids. This is the first study to demonstrate that differentially expressed genes among cichlids can be identified using a DNA chip.
Introduction
Lake Victoria, which is situated in the East African
Plateau, harbors more than 500 endemic haplochromine
species (Seehausen 1996, Turner et al. 2001). Geological
evidence indicates that this lake dried up and refilled
15,600–14,700 years ago (Johnson et al. 1996). Molecular
phylogenetic analysis suggests that Lake Victoria cichlids
have extremely low neutral genetic variation (Nagl et al.
2000, Watanabe et al. 2004). These facts indicate that Lake
Victoria cichlids have diverged over a surprisingly short period. Moreover, Lake Victoria cichlids can breed fertile
hybrids in unusual environments such as laboratory tanks
(de Caprona 1984) or the turbid eutrophic region of the lake
(Seehausen, van Alphen, and Witte 1997). These observations also suggest that the Lake Victoria haplochromine
species are genetically close.
On the other hand, Haplochromis in Lake Victoria exhibit a wide variety of trophic specializations, including
feeding on insects, algae, mollusks, and scales or predation
on other Haplochromis (Greenwood 1974). The jaw morphologies of cichlids reflect the adaptation of the species to
their trophic habitats. Indeed, Greenwood (1974) suggested
that the evolutionary success of the Haplochromis species
flock could be attributed to head and dentition diversity.
Thus, the Lake Victoria cichlids have evolved diverse
jaw morphologies despite their close genetic relatedness.
Furthermore, the anatomical differences between Lake
Victoria cichlid species are maintained after two generations
of breeding in the laboratory under standardized conditions
(Bouton et al. 1999). These facts suggest that heritable geKey words: cichlid, metalloproteinases, gene expression level,
morphogenesis.
E-mail: [email protected].
Mol. Biol. Evol. 22(8):1649–1660. 2005
doi:10.1093/molbev/msi159
Advance Access publication April 27, 2005
Ó The Author 2005. Published by Oxford University Press on behalf of
the Society for Molecular Biology and Evolution. All rights reserved.
For permissions, please e-mail: [email protected]
netic elements are responsible for their species-specific
jaw morphology. However, the genetic factor(s) responsible
for the distinct morphological characteristics has not been
identified.
Organ morphogenesis is controlled by a variety of
genes (see Carroll, Grenier, and Weatherbee 2001). It is unlikely that every Lake Victoria Haplochromis species has
species-specific genes that account for their unique morphological features. It is more reasonable to expect that species-specific jaw morphologies result from the differential
expression of genes that are shared among cichlid species.
Hence, we focused on identifying such genes that may be
involved in jaw morphogenesis.
Because the jaw is one of the most divergent of cichlid
features, it is critical that we characterize species-specific
expression profiles toward the goal of elucidating the molecular mechanisms of species-specific jaw morphogenesis.
In this study, we introduce a new approach to investigate this
issue. Because Lake Victoria cichlids are genetically similar, a DNA chip can be used to screen for genes that are
responsible for cichlid jaw morphogenesis. DNA chips
(microarrays) have been used to compare genome-wide
gene expression levels in tissues or cells under different conditions (Gasch et al. 2000). Recent reports also show that
a DNA chip can be used to analyze developmental pathways
(Michaut et al. 2003) and to identify regulatory gene sets and
their regulators (Segal et al. 2003). We report here the first
examples of genes that are differentially expressed between
two cichlid species, as screened using a DNA chip.
Materials and Methods
Cichlid Fishes
The cichlid species we used in this work were as follows:
Haplochromis parvidens (HP) (although this species was
1650 Kijimoto et al.
originally described by Greenwood [1959] and reclassified
as Lipochromis parvidens by Greenwood [1980], here we
use Haplochromis for popularity and convenience), Haplochromis chilotes (HC), H. sp. ‘‘rockkribensis’’ (HK) (these
two species were also reclassified as Paralabidochromis chilotes and P. sp. ‘‘rockkribensis,’’ respectively; see a review in
Seehausen [1996] and Greenwood [1959] for HC), and H. sp.
‘‘red tail sheller’’ (HR); (this species was also reclassified as
Ptyochromis sp. ‘‘red tail sheller,’’ see Greenwood [1980]
and Seehausen [1996] for description of Ptyochromis).
The fishes were imported through a local pet supplier and
maintained under constant conditions (28°C, 12-h dark/
12-h light cycle) in 40-Liter tanks. Adults were fed three
times daily with commercial food. One male and three to four
females were placed in a tank to form pair bonds. Eggs nursed
in female mouths were collected at 3 days postfertilization
(dpf) and were kept in a PET bottle with fresh water supplied
intermittently. Larvae hatched around 4–5 dpf and began
to eat food 12–13 dpf, after which they were released into
tanks specific for each species and fed three times daily with
commercial food.
Isolation of Total RNA from Cichlids
Total RNA was purified from the whole body of HC at
various growth stages using the RNeasy mini kit (Qiagen,
Hilden, Germany) according to the manufacturer’s instructions. Purified RNA was suspended in RNase-free water,
and the RNA quality was assessed by electrophoresis.
RNA was stored at 4°C.
Construction of the complementary DNA Library and
DNA Chips
HC complementary DNA (cDNA) was synthesized by
SuperScriptII reverse transcriptase (Invitrogen, Carlsbad,
Calif.). cDNA was purified and fractionated by a cDNA
size-fractionation column (Invitrogen) to exclude short
cDNA fragments. cDNAs (500–1,000 bp) were cloned into
pGEM-T (Promega, Madison, Wisc.). A total of 1,792 clones
were spotted on poly-L-lysine–coated glass slides according
to Nakazawa et al. (2003).
Relative messenger RNA (mRNA) expression levels
were compared between two cichlid species, namely HP
and HR, using the DNA chip system described above.
The procedure was performed in duplicate by two independent researchers. Amplification of total RNA from cichlid
larvae was performed as described by Luo et al. (1999).
Total RNA was isolated from the lower jaw, premaxilla,
and maxilla of three larvae from each of the two cichlid
species (HP and HR).
Total RNA was amplified ;400-fold. Amplified RNA
(1 lg) from HR and HP was tagged with Cy3- and Cy5-labeled
uridine (Amersham Bioscience, Piscataway, N.J.), respectively, by reverse transcription (RT) using random hexamer
primers. Labeled target cDNA was mixed and hybridized to
a single DNA chip for 14 h at 55°C. Hybridized DNA
chips were scanned using a ScanArray 4000 (Perkin Elmer,
Boston, Mass.). After normalization, scanned data were analyzed using QuantArray (GSI Lumonics, Nepean, Canada).
In Situ Hybridization
Full-length cimp1 cDNA was cloned into pGEM-T
(pGEM-CiMP1). Digoxigenin (DIG)-labeled sense and
antisense RNA probes were generated from linearized
pGEM-CiMP1. The plasmid construct was digested with
NotI to generate the sense probe and NcoI to generate
the antisense probe. Transcription was performed with
DIG RNA-labeling mix (Roche, Basel, Switzerland) according to the manufacturer’s instructions.
Whole-Mount In Situ Hybridization
A series of procedures of whole-mount in situ hybridization was performed as described by Murakami et al.
(2001) with the exception that the washing step (with
50% formamide, 5 3 standard saline citrate [SSC], 1%
sodium dodecyl sulfate [SDS]) was performed once for
30 min at 65°C. Detection of the DIG-labeled RNA probe
and coloration of specimens was performed according to
Murakami et al. (2001).
In Situ Hybridization in Frozen Sections
The reagents for hybridization and immunodetection
of DIG-labeled probes were the same as for whole-mount
in situ hybridization.
Samples stored at ÿ20°C in methanol were rehydrated
with a series of ethanol dilutions and then replaced with
30% sucrose. The samples were embedded in OCT compound (Sakura Finetek, Torrance, Calif.) and frozen at
ÿ80°C until sectioning. Sectioning (10 lm) was performed
using a Leica CM1850. Sections were placed on MAScoated glass slides (Matsunami, Osaka, Japan) and dried
immediately. The glass slides were rinsed in diethylpyrocarbonate-treated water. Samples were fixed in 4% paraformaldehyde–phosphate-buffered saline (PBS) containing 0.2%
glutaraldehyde. After washing the fixative with PBS, the
samples were acetylated in triethanolamine-HCl to reduce
nonspecific hybridization of the probe. Prehybridization
was performed at room temperature (rt), and hybridization
was performed at 60°C. The sections were hybridized
with the same probes used for the whole-mount experiments. The specimens were washed with 2 3 SSC for
5 min at 60°C, 0.2 3 SSC for 1 h at 65°C, and gradually
cooled from 65°C to rt in 0.2 3 SSC. Samples were soaked
in 0.2 3 SSC for 5 min at rt and equilibrated with NT buffer
(0.1 M Tris-HCl, pH 7.5, 0.15 M NaCl) for 5 min at rt. The
sections were incubated with 1% blocking solution (Roche)
in NT buffer for 1 h at rt. For immunodetection of DIGlabeled probes, samples were incubated with alkaline phosphatase–conjugated anti-DIG Fab fragments (diluted
1:4,000; Roche) in blocking solution. Bound antibodies
were washed with NT buffer three times for 5 min at rt. Alkaline phosphatase was activated in NTM buffer (0.1 M
Tris-HCl, pH 9.5, 0.1 M NaCl, 50 mM MgCl2) for 10
min at rt. Color development was accomplished with 5bromo-4-chloro-3-indolyl phosphate (BCIP)/nitrotetrazolium blue chloride (NBT) (Sigma, St. Louis, Mo.) in NTM
buffer according to the manufacturer’s instructions.
Differentially Expressed Genes in Cichlids 1651
Quantitative Real-Time RT–Polymerase
Chain Reaction
Quantitative real-time polymerase chain reaction
(PCR) was performed using the Quantitative SYBR Green
RT-PCR kit (Qiagen) with 40 ng total RNA from three
cichlid larvae according to the manufacturer’s instructions.
Amplification and detection of products was performed
using an iCycler iQ real-time PCR analysis system (BioRad, Hercules, Calif.). Glyceraldehyde-3-phosphate dehydrogenase was used as an internal control. Experiments
were performed three times each.
Expression of cichlid metalloproteinase 1 in
Escherichia coli Cells
The region encoding mature cichlid metalloproteinase
1 (CiMP1) (amino acids 59–256) was cloned into the
pET28a expression vector (EMD Biosciences, San Diego,
Calif.). To obtain recombinant protein with a C-terminal
His-tag, primers were designed to amplify the mature region as follows: MP1-CF01, 5#-TATACCATGGAAAATGCTGTTCCAT-3#; MP1-CR01, 5#-ATATCTCGA
GGCATCCATAAAGCCT-3#. The amplified mature region was digested with NcoI and XhoI, and the resulting
fragment was cloned into pET28a. The construct was
transformed into Escherichia coli BL21(DE3), and the
culture was grown at 18°C. The cells were harvested
by centrifugation for 3 h after induction by 0.4 mM
isopropyl-b-D-thiogalactopyranoside and subsequently
stored at ÿ20°C.
Partial Protein Purification
Proteins were extracted by sonication of the recombinant E. coli in extraction buffer A (20 mM Tris-HCl, pH
8.0, 0.5 M NaCl). Most of the recombinant CiMP1 was
sequestered in insoluble inclusion bodies. The insoluble
fraction was suspended in buffer A and centrifuged at
5,0003g for 20 min at 4°C. The precipitate was resuspended in buffer B (20 mM Tris-HCl, pH 8.0, 0.5 M NaCl,
5% [w/v] Triton X-100) and centrifuged as above. This step
was repeated. The precipitate was then suspended in denaturing buffer (50 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 6 M
guanidine hydrochloride). After clearance by centrifugation
at 40,0003g for 30 min at 4°C, the supernatant was applied to a HiTrap–chelating HP column (5 ml, Amersham
Biosciences) that had been previously charged with Ni21
according to the manufacturer’s instructions. The column
was washed extensively with denaturing buffer (as indicated above) and washing buffer (50 mM Tris-HCl, pH
8.0, 0.5 M NaCl, 6 M guanidine hydrochloride, 30 mM imidazole). Proteins were eluted with elution buffer (50 mM
Tris-HCl, pH 8.0, 0.5 M NaCl, 6 M guanidine hydrochloride, 0.5 M imidazole).
Proteolytic Zymography
For the analysis of proteolytic activity of CiMP1, the
insoluble fraction and partially purified enzyme fraction
were subjected to gelatin zymography according to Hung
et al. (1997). The protein-degrading activity in protein sam-
ples can be detected by the appearance of a coomassie brilliant blue (CBB)-negative band on the acrylamide gel
containing its protein substrate. Part of the insoluble fraction was incubated for 30 min at 37°C in 200 ll of SDS–
polyacrylamide gel electrophoresis (PAGE) sample buffer
lacking b-mercaptoethanol so as to not reduce the functional disulfide bond. A Tris-Tricine SDS-PAGE system
was used for electrophoresis. The gel was 7.5% acrylamide
containing 0.1% gelatin. After electrophoresis, the gel was
incubated twice for 1 h at 28°C with 2% (w/v) Triton X-100
in 20 mM Tris-HCl (pH 8.0) followed by incubation in 20
mM Tris-HCl (pH 8.0) containing 0.1 mM ZnCl2 (renaturation buffer) for 16 h at 28°C. The gel was stained with
Quick-CBB (Wako, Osaka, Japan). When we tested other
proteins such as fibronectin, reduced and carboxymethylated bovine serum albumin (BSA), casein, and collagen
I for zymography, the concentration of proteins in the
gel was adjusted to 0.5%–1% (w/v) depending on the samples. For testing the effect of ethylenediaminetetraacetic
acid (EDTA), an aliquot of 0.5 M EDTA was added to
the renaturation buffer to give a final concentration of 20
mM. The gel was incubated in the buffer for 16 h at
28°C and stained with Quick-CBB.
Results
Screening for Genes that are Differentially Expressed
Between Two Cichlid Species
As mentioned in Materials and Methods, this work
primarily used the cichlid species HP, HK, and HR. We
focused on gene expression levels that may affect jaw morphology and used a DNA chip to screen the genes that are
differentially expressed between the Lake Victoria haplochromine species HP and HR. In preliminary studies, we
examined the morphology of cichlid larvae heads at various
ages by staining the cartilage with Alcian blue and the bone
with Alizarin red. We monitored the head morphology
around 20 dpf, but we could not find obvious differences
in the characteristics between the two species before that
time. However, mineralization of bones in the head is drastically achieved from 10 to 20 dpf, and for all the fishes we
tested for staining at 20 dpf, most bones in the head were
stained with Alizarin red. Head morphology became distinct around 50–60 dpf in each species. We therefore compared gene expression levels between HP and HR using 20
dpf larvae, because these species display a prominent difference in lower jaw length during adulthood.
We screened for clones for which the difference in
fluorescence intensity was more than fourfold between
the two species. This criterion defined ‘‘differentially
expressed genes’’ in this study. We repeated the same procedure three times for the same amplified RNA samples of
jaws. Furthermore, two investigators in this laboratory used
different amplified RNA samples for three comparisons
of expression levels between HP and HR, as described
in Materials and Methods. Three clones were identified
reproducibly in these separate experiments. The isolated
fragments were as follows: clone no 08f04, an astacin
family–like gene, the signal for which was approximately
sixfold stronger in HR than HP; clone no 02a11, similar to
1652 Kijimoto et al.
FIG. 1.—Amino acid sequence alignment of CiMP1 and other astacin family proteins. The amino acid sequence of HC CiMP1 was aligned with that
of carp nephrosin and LCE, a medaka hatching enzyme. Asterisks indicate residues that are identical among all three proteins and dots indicate residues
that are identical in two proteins.
mouse c-type lectin; and clone no 34e11, similar to epithelin-granulin. We used clone no 08f04 in subsequent analyses because some astacin family proteins are responsible
for bone morphogenesis in other animals (Wardle, Welch,
and Dale 1999; see also Discussion)
Sequence Analysis of CiMP1
The sequence of the full-length clone no 08f04 was
obtained from our expressed sequence tag data (Watanabe
et al. 2004). We used RT-PCR to isolate this sequence from
cDNA libraries for HC, HP, HK, and HR. Both interspecies
variation and intraspecies polymorphisms were observed,
as often occur in analyses of Lake Victoria cichlids. The
sequence identity of cimp1 among species is more than
98% (data not shown). The open reading frame (ORF)
for clone no 08f04, which we named cimp1 (DNA Data
Bank of Japan accession number AB192347), was 768
bp encoding 256 amino acids. A BlastX database search
suggested that this gene is highly similar to carp nephrosin,
which was originally purified from the head kidney of carp
and identified as a member of the astacin family (Hung et al.
1997). cimp1 is also similar to medaka hatching enzyme
that involves two genes, high choriolytic enzyme (HCE)
and low choriolytic enzyme (LCE, fig. 1). According to
GenBank, the other known astacin family proteins are as
follows: astacin, a digestive enzyme from crayfish (Stocker,
Sauer, and Zwilling 1991); meprin, a multiple domain
membrane component comprised of homologous alpha
and beta chains (Jiang et al. 1992); bone morphogenetic
protein-1 (BMP1), which is involved in vertebrate morphogenesis (Dumermuth et al. 1991); tolloid from Drosophila
(Shimell et al. 1991); and hydra metalloproteinase 1
(HMP1) and HMP2 from hydra (Yan et al. 1995, 2000).
Members of this protein family have an N-terminal signal
and propeptide for secretion from cells and correct activation in tissues, respectively. The mature proteins require
a zinc ion for their activity and thus contain an astacin signature motif (HEXXHXXGFXHEXXRXDRD) that serves
as a zinc ion–binding site and proteinase catalytic center.
These proteins also contain a met-turn (SXMHY). The
cimp1 sequence encodes all these domains (fig. 1). In some
members of this protein family, the C-terminal end of the
proteinase domain has an additional domain, such as an epidermal growth factor-like motif and/or a CUB (complement
components C1r-C1s, the sea urchin protein Uegf, and
BMP1) motif that is not present in CiMP1.
We identified the mature region from sequence similarity with other genes (fig. 1). The molecular weights for
the full-length and mature protein were predicted to be ;33
kDa and ;22 kDa, respectively. The putative signal peptide
was determined by the hydrophobicity ratio of amino acid
residues. No sequence similarity to the putative signal or
propeptide was found in GenBank.
Expression Pattern Analysis of cimp1
In Situ Hybridization
We performed whole-mount in situ hybridization at the
stage of 20 dpf larvae, which was used for comparison of
gene expression on the DNA chip. Using cimp1 antisense
probes generated from HR, HP, and HK, we detected cimp1
mRNA only around the head in all these species (figs. 2 and
3), whereas the sense probe did not generate a significant
signal (data not shown). The mRNA expression was strongly
detected in HR; however, the signal of HP was so weak. For
Differentially Expressed Genes in Cichlids 1653
FIG. 2.—Whole-mount in situ hybridization of cimp1 in three cichlid species at 20 dpf. Larvae at 20 dpf were used for the experiments. An antisense
probe for cimp1 was hybridized with HR, HP, and HK. A, E, and G are photos of adult of each species. Photos of a head in A and E are specimens that had
their bones stained to show morphological differences. Lines are drawn to show head length and lower jaw length according to Barel et al. (1977) (scale
bars indicate 1 cm). B, F, and H are the lateral view of each species. C and D are ventral and dorsal views, respectively, of HR (scale bar in B 5 1 mm).
example, HR larvae indicate strong hybridization signals on
the lateral line (supraorbital canal) and part of premaxilla,
both of which are membranous bones, and on epithelia of
chin and gills, where HP indicated very weak signals except
for around the premaxilla (fig. 2). Next we examined expression of cimp1 at various time points of growth by using HK
larvae. Results showed that cimp1 appeared just after hatch
and keep their high expression level through the growth
stage that we tested (fig. 3). Although the extent of the difference of the signal intensity of in situ hybridization is not
necessarily consistent with the data of DNA chips, the signal
is apparently stronger in HR than in HP.
1654 Kijimoto et al.
FIG. 3.—cimp1 expression during HK growth. An antisense probe for cimp1 was hybridized with embryo or larvae of HK. Specimens at different
ages are indicated (dpf). The photograph of an adult fish is shown in A. (B) The 15 dpf larva of HK. Signals are only seen in head and gills. (C) The 3 dpf
prehatching embryo indicates no signal. (D) The 4 dpf posthatching larva indicates a strong hybridization signal. The signal continues to be seen during
growth when we tested. E and F, photographs of HK at 15 dpf and 20 dpf, respectively (grid in B 5 1 mm).
We also performed in situ hybridization on frozen
sections of HK at 10 dpf for a more detailed profile of
cimp1 expression. A clear hybridization signal was observed in the epithelial region of the specimens but not
around bones (fig. 4). Hybridization to sections of body
parts, including kidney, did not produce a signal (data
not shown), consistent with data from the whole-mount
in situ hybridization study.
description for fish morphology in Discussion). Thus, HK
larvae were used instead of HR, and expression levels of cimp1
were compared with those of HP in this study. Relative cimp1
expression was analyzed at different growth stages for each
species. Maximum expression was observed at 10 dpf in
HK and at 15 dpf in HP. HK had higher levels of gene expression than HP in most of the time points we tested.
Quantitative RT-PCR
Analysis of the Biological Role of CiMP1
Relationship with Other Astacin Family Proteins
We performed quantitative real-time RT-PCR (fig. 5)
to examine cimp1 expression during larval growth and
compared the levels of cimp1 expression in different cichlid
species. We had difficulty obtaining sufficient HR larvae
after DNA chip experiments due to the aggressive behavior
of the adults, which precluded consistent breeding. Like
HR, HK has a shorter lower jaw than HP (see the detailed
To better understand the biological role of CiMP1, we
searched the GenBank database for astacin family proteins
having conserved amino acid sequences. We aligned the
deduced CiMP1 sequence with astacin family proteins such
as the medaka (Oryzias latipes) hatching enzymes, known
as HCE and LCE, carp (Cyprinus carpio) and zebrafish
(Danio rerio) nephrosin, hydra (Hydra vulgaris) HMP1,
Differentially Expressed Genes in Cichlids 1655
FIG. 4.—In situ hybridization of sectioned HK samples. An antisense probe for cimp1 was hybridized with HK at 10 dpf. (A) A transverse section. (B)
Enlarged area shown boxed in (A). (C) Sense probe yielded no signal. (D) Photograph of HK at 10 dpf, prior to sectioning and hybridization; m, Meckel’s
cartilage.
human BMP1, mouse meprin1a, three unknown Fugu
(Fugu rubripes) sequences, two unknown zebrafish sequences, and crayfish (Astacus astacus) astacin. As performed in
a previous study by Hung et al. (1997), we aligned all of the
astacin domains (;200 amino acids) using the ClustalX
sequence alignment program (Thompson et al. 1997). A
phylogenetic tree was subsequently constructed using
these same sequences. A Neighbor-Joining tree (p-distance
method with complete deletion) generated by MEGA2
(Kumar et al. 2001) showed that cimp1 and the nephrosin
gene are clearly distinguishable but are included in a sister
group (fig. 6). cimp1 also was distinguishable from the
hatching enzymes, suggesting another possible function
for this gene. These results were supported by high bootstrap values (fig. 6).
proteolytic band was observed (data not shown), suggesting
that the activity requires a disulfide bond. Inclusion of
20 mM EDTA in the reaction (see Materials and Methods)
reduced the gelatin-degrading activity of CiMP1 (fig. 7),
supporting the prediction that zinc is required for enzymatic
activity (as with other astacin family proteases). Other potential substrates of CiMP1 were also tested by zymography. Recombinant CiMP1 degraded not only gelatin but
also casein, a general substrate of many proteinases in vitro,
reduced-carboxymethylated BSA, which is degraded by
nephrosin, fibronectin, which is also degraded by hydra
HMP1 (Yan et al. 1995), and denatured collagen I (data
not shown).
Enzymatic Activity of Recombinant CiMP1
Discussion
DNA Chips can be Used to Identify Genes Responsible
for Species-Specific Cichlid Morphogenesis
Because our sequence analysis suggested that cimp1 is
a member of the astacin family of metalloproteinases, the
protein product was predicted to have the ability to degrade
extracellular matrix (ECM) proteins and to require a zinc
ion for activity. Hence, we tested whether recombinant
CiMP1 had proteolytic activity in vitro using a zymography
assay. The putative mature protein region of cimp1 cDNA
was cloned into pET28a, and CiMP1 was expressed in bacteria. A partially purified fraction containing recombinant
CiMP1 was subjected to gelatin zymography. As shown
in figure 7, recombinant CiMP1 had gelatin-degrading
activity. When the enzyme solution was treated with
SDS-PAGE sample buffer including reducing reagent, no
We constructed a DNA chip to identify and to isolate
genes that are differentially expressed between species.
According to our expressed sequence tag (EST) analysis,
cDNA sequences of the expressed genes are almost identical among species (Watanabe et al. 2004). Thus, we consider that HC cDNA used here for the construction of
cDNA library and DNA chip will not affect the results that
compare gene expression levels between two cichlid species that are not HC. Among the species we used, HP
has the longest lower jaw (43.3%–55.5% of head length,
Greenwood 1974). HK, being included in Paralabidochromis species according to a recent reclassification (Seehausen 1996), is recognized by a short head (less than 33% of
1656 Kijimoto et al.
FIG. 5.—Real-time RT-PCR of cimp1 in HP and HK at different larval stages. Total RNA from HP and HK jaws was amplified in a one-step RT-PCR
reaction. Relative expression levels of cimp1 are shown at various dpf. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal
control. The x axis indicates the developmental stage (dpf), and the y axis indicates the expression level relative to GAPDH. 5 dpf is shown as 1. Note that
the relative expression level is different between the two species.
the standard length) and a relatively long lower jaw (between 34% and 39% of the head length; Seehausen
1996). HR, being included in Ptyochromis according to recent reclassification, is mainly identified by dentition, tooth
shape, and a relatively short lower jaw (22%–38% of the
head length; Seehausen 1996). We expected to isolate
the genes that may create such morphological difference
and to testify their biological functions after isolation of
the genes.
Using the DNA chip, we detected three clones that
exhibited differential expression between the two cichlid
species. The expression of clone no 34e11, which is similar
to epithelin-granulin, was ;16-fold higher in HR compared
with HP. The expression of clone no 02a11, which is similar to c-type lectin, was ;10-fold higher in HP compared
with HR. Clone no 08f04, subsequently named cimp1, was
expressed approximately sixfold higher in HR compared
with HP and putatively belongs to a metalloproteinase family that degrades the ECM and/or other protein substrates.
Because metalloproteinases that degrade the ECM often
play important roles in organ development (for review
see Perris and Perissinotto 2000), we first analyzed the
function of CiMP1 in vitro. Nephrosin is a single-locus
gene while other two genes belong to the multigene family.
This was another reason to choose this gene for the analysis,
because we planned to quantify the gene expression level
by using real-time quantitative RT-PCR after the DNA chip
experiment, and multi-loci genes are difficult to analyze
their quantity because of their high similarity among paralogs. The other candidate genes identified in this study will
be described elsewhere.
Analysis of CiMP1 Sequences
We compared cimp1 cDNA sequences among Lake
Victoria cichlids. The signal and propeptide sequences encoded by the cDNAs were perfectly conserved among the
four Haplochromis species HP, HR, HK, and HC. There is
no fixed sequence difference of coding region among HP,
HR, HK, and HC, implying that the enzymatic activity of
CiMP1 is essentially the same among cichlid species. This
suggests that the level of cimp1 expression is important for
cichlid differentiation.
Astacin family proteins can be divided into subgroups
as follows (Geier and Zwilling 1998): proteins having morphogenetic activity, such as BMP1; proteins that degrade
various bioactive peptides, such as meprin; and proteinases
that degrade the egg envelope during hatching, such as medaka LCE. Phylogenetic analysis established that cimp1 is
an astacin family member (fig. 6). However, it is interesting
that nephrosin and cimp1 are not included in any of the subgroups described above, at least among the amino acid sequences we considered. This indicates that nephrosin and
cimp1 comprise a new subgroup in the astacin family.
Hence, further analysis is needed to predict the function
of CiMP1.
Analysis of cimp1 Gene Expression and
CiMP1 Activity In Vitro
The biological role of vertebrate astacin family members, except BMP1, is not well understood. BMP1 is a mammalian homolog of tolloid, which activates decapentaplegic
Differentially Expressed Genes in Cichlids 1657
99
zebra ctg26358
fugu ctg7210
84
zebra nephrosin
98
99
carp nephrosin
fugu ctg321
70
99
CiMP1 red
100 CiMP1 parvidens
medaka LCE
72
fugu ctg1806
88
95
medaka HCE
zebra ctg12392
100
mouse
BMP1
73
hydra
crayfish
0.050
FIG. 6.—Phylogenetic tree of CiMP1 and other astacin family proteins. A Neighbor-Joining tree (p-distance method, missing sequence data
was completely deleted) was constructed using the MEGA2 phylogenetic
tree program. Sequences used here are CiMP1 from HP (parvidens) and
HR (red), two unknown zebrafish sequences, three unknown Fugu sequences, carp nephrosin and its zebrafish homolog, medaka LCE and HCE,
mouse meprin 1a (referred to as ‘‘mouse’’), human BMP1, HMP1 (referred
to as ‘‘hydra’’), and crayfish astacin. Only the astacin domain was used for
the analysis. Scale bar indicates amino acid changes per site.
to determine the Drosophila dorsoventral axis (Shimell
et al. 1991). BMP1 cleaves chordin during bone morphogenesis. Chordin is an antagonist of BMP4, which directs
bone and cartilage formation (Wardle, Welch, and Dale
1999). It is likely that some genes of the astacin family
play an important role in organ morphogenesis, and thus
CiMP1 is expected to function in organ tissue morphogenesis in the head.
Astacin family proteins have various expression patterns. Carp nephrosin, which is closely related to CiMP1
(fig. 1), is expressed in various tissues such as head kidney,
blood, and spleen (Hung et al. 1997). On the other hand,
there is another group of astacin family genes whose expression is strictly temporally and/or spatially controlled.
Hydra HMP1, for example, localizes only in the ECM in
the head region and is involved in development (Yan
et al. 1995). HMP2 expression is localized in the foot
and is involved in foot morphogenesis (Yan et al. 2000).
In this work, we found that cimp1 is expressed around epithelia in the head during all developmental stages tested
(3–20 dpf, figs. 2, 3, and 4). Astacin family metalloproteinase BMP1, when expressed in the skin, is responsible for
assembling dermal-epidermal junctions (Burgeson and
Christiano 1997). By analogy, it is likely that epithelial expression of CiMP1, a member of astacin family like BMP1,
may affect cichlid jaw morphogenesis during their growth
via its protein-degrading function.
The in situ hybridization results (figs. 2 and 3) suggest
that the regions in which cimp1 is expressed are similar to
those of hatching enzymes. Hatching enzymes are secreted
from hatching gland cells, which are distributed in the gills,
lower jaws, and yolk of fish embryos (Inohaya et al. 1995).
Given that cimp1 is expressed in epithelial cells of the head
FIG. 7.—Recombinant CiMP1 possesses proteolytic activity. (A) Partially purified recombinant CiMP1 was subjected to SDS-PAGE through
a gel containing 0.1% gelatin. Partially purified protein (2, 10, or 20 lg)
was loaded on the gel. A clear band at ;23 kDa is observed (arrow). We
performed the same expression and purification procedure using bacterial
cells transformed with empty pET28a vector. No band is observed using
this partially purified protein fraction (nc). (B) Addition of 0.5 M EDTA to
give a final concentration of 20 mM reduces CiMP1 activity.
and gills, we initially suspected that it is an ortholog of the
hatching enzymes. The medaka hatching enzyme is expressed in the hatching gland of the lower jaw and around
gills in prehatched embryos, and its expression drastically
drops after hatching (Inohaya et al. 1995). By contrast,
cimp1 is not expressed in prehatching embryo but expressed even after hatching (fig. 3); moreover, its expression level fluctuates during growth (fig. 5). Thus, we
conclude that CiMP1 is neither an ortholog of nephrosin
nor of hatching enzymes. The in vitro enzymatic activity
(fig. 7) and amino acid sequence of CiMP1 (figs. 1 and
6) suggest that it is a new member of the astacin metalloproteinase family derived from hatching enzymes. Interestingly, nephrosin, which is included in the same subgroup of
the family, is also expressed in gills like cimp1. However,
cimp1 is not expressed in internal organs such as kidney or
spleen. Isolation of cichlid ortholog of nephrosin will help
us to understand the evolution of fish genome and astacin
family.
We used quantitative real-time RT-PCR to examine
cimp1 expression as a function of cichlid development
(fig. 5). In the two species that were analyzed (HP and
HK), we found that expression differed in both timing
and magnitude. cimp1 expression reached a maximum 5
days earlier in HK (10 dpf) than in HP (15 dpf). Because
every species of Lake Victoria cichlids in our lab began to
consume food at 12–13 dpf, foraging-related developmental or morphogenetic events occurring at 10–15 dpf (i.e.,
between the time of maximum expression of cimp1 in
HK and HP) may be critical for determining the unique
head morphology of these two species during growth. In
blowfly, transcriptional heterochrony of genes responsible
for morphogenesis can contribute to morphological differences between closely related species (Skaer, Pistillo, and
1658 Kijimoto et al.
Simpson 2002). Similarly, the transcriptional heterochrony
we observed in cichlids may also contribute to the morphological differences between HK and HP. Our next goal will
be to identify the factor(s) that controls cimp1 transcription.
This factor may be a cis regulatory element and/or signal
transduction molecules that act upstream during development. We recently constructed a bacterial artificial chromosome (BAC) library for HC (Watanabe et al. 2003), which
will help us to efficiently isolate the upstream regions of
genes.
Sequence similarities between CiMP1 and other
proteins of the astacin family suggest that it should have
gelatin-degrading activity. Indeed, recombinant CiMP1
has proteolytic activity that is zinc dependent (fig. 7). In
most cases, enzymes that degrade the ECM contribute to
organ morphogenesis (Chin and Werb 1997). Consistent
with this paradigm, some of the astacin family proteinases
degrade ECM proteins and are involved in organogenesis.
We observed that CiMP1 has broad substrate specificity, as
do other astacin family proteins such as BMP1 or meprin A
and B. BMP1 degrades procollagens (Kessler et al. 1996, Li
et al. 1996), prolysyl oxidase (Uzel et al. 2001), and biglycan (Scott et al. 1999) in vitro. We observed that CiMP1 can
degrade denatured type I collagen and fibronectin, ECM
proteins that function in cell adhesion, migration and/or differentiation. Taken together with the results of the in situ
hybridization and quantitative real-time RT-PCR assays,
this information suggests that the CiMP1 expression level
during cichlid growth affects the total mass of ECM proteins in the head structures, including the jaws. Thus,
our current understanding suggests three possibilities for
CiMP1 function. CiMP1 may degrade cell surface proteins,
such as fibronectin, and thereby prevent cell migration and/
or differentiation. Alternatively, CiMP1 may degrade ECM
proteins and affect cell-cell signaling mediated by ECM
proteins during cichlid growth. (Importantly, ECM proteins
interact with secreted signaling molecules such as transforming growth factor-b proteins and/or growth factors.)
Reconstruction of ECM will take place during the growth
of organs of cichlid larva. Thus, it is likely that changes in
the timing of expression levels of ECM protein-degrading
proteinases such as CiMP1 will affect the jaw shapes
among species. Finally, CiMP1 may function analogously
to BMP1 and degrade other signaling molecules directly.
In each case, bony tissues grow and represent speciesspecific jaw shape gradually but after embryogenesis. Interestingly, it should be noted that cimp1 was expressed in
membranous bones of cichlid larvae. Premaxilla is one such
bone, and the shape of this bone is sometimes used for the
description of species-specific characteristics for cichlids.
cimp1 was also strongly expressed in the tissues that are
not directly related to the jaws such as chin or gills. However, expression of cimp1 in these tissues may affect the
morphological characteristics of contiguous tissue such
as lower jaw during the growth.
We are now preparing antibody to observe CiMP1 localization in cichlids. Immunohistochemical analysis will
help us to understand in detail the region where the protein
is secreted and located. Furthermore, the antibody will also
help us to understand protein quantity of cimp1 during the
growth of cichlid.
A recent study of Lake Malawi cichlids (Albertson,
Streelman, and Kocher 2003) illustrates the ongoing effort
to elucidate the genetic factors responsible for morphological differences between closely related species using quantitative trait loci (QTL). On the other hand, we applied DNA
chip technology to directly identify differentially expressed
genes among closely related species on a genome-wide basis. This procedure also will help us to further observations
of gene expression profiles during cichlid morphogenesis.
Genome-wide gene-searching methods, such as QTL
analysis and our method, will directly identify the genes
required for species-specific morphogenesis among cichlids. To demonstrate the direct involvement of such genes
in morphogenesis, it will be necessary to generate and
analyze transgenic animals, the procedure of which was
recently established in the study of transgenic zebrafish
development (Kawakami et al. 2004). The application
of this method to cichlids will enable us to directly test
the contribution of candidate developmental genes during
organ morphogenesis.
Acknowledgments
The authors thank H. Handa and Y. Yamaguchi of the
Tokyo Institute of Technology for technical advice on realtime quantitative RT-PCR. T. K. thanks M. Sato for maintaining the fishes and M. Matsuura for technical assistance
of in situ hybridization for frozen section. This work was
supported by a Grant-in-Aid for Scientific Research on Priority Areas ‘‘Molecular Mechanism of Speciation’’ and
‘‘Genome Science’’ to N.O. from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
Literature Cited
Albertson, R. C., J. T. Streelman, and T. D. Kocher. 2003.
Directional selection has shaped the oral jaws of Lake Malawi
cichlid fishes. Proc. Natl. Acad. Sci. USA 100:5252–5257.
Barel, C. D. N., M. J. P. van Oijen, F. Witte, and E. L. M. WitteMaas. 1977. An introduction to the taxonomy and morphology
of the haplochromine Cichlidae from Lake Victoria. Neth. J.
Zool. 27:333–389.
Bouton, N., F. Witte, J. J. M. van Alphen, A. Schenk, and O.
Seehausen. 1999. Local adaptations in populations of rockdwelling haplochromines (Pisces: Cichlidae) from southern
Lake Victoria. Proc. R. Soc. Lond. B 266:355–360.
Burgeson, R. E., and A. M. Christiano. 1997. The dermalepidermal junction. Curr. Opin. Cell Biol. 9:651–658.
Carroll, S. B., J. K. Grenier, and S. D. Weatherbee. 2001. From
DNA to diversity. Blackwell Science, Oxford.
Chin, J. R., and Z. Werb. 1997. Matrix metalloproteinases regulate
morphogenesis, migration and remodeling of epithelium,
tongue skeletal muscle and cartilage in the mandibular arch.
Development 124:1519–1530.
de Caprona, M.-D. C. 1984. Interspecific fertile hybrids of Haplochromine Cichlidae (Teleostei) and their possible importance
for speciation. Neth. J. Zool. 34:503–538.
Dumermuth, E., E. E. Sterchi, W. P. Jiang, R. L. Wolz, J. S. Bond,
A. V. Flannery, and R. J. Beynon. 1991. The astacin family of
metalloendopeptidases. J. Biol. Chem. 266:21381–21385.
Gasch, A. P., P. T. Spellman, C. M. Kao, O. Carmel-Harel, M. B.
Eisen, G. Storz, D. Botstein, and P. O. Brown. 2000. Genomic
expression programs in the response of yeast cells to environmental changes. Mol. Biol. Cell 11:4241–4257.
Differentially Expressed Genes in Cichlids 1659
Geier, G., and R. Zwilling. 1998. Cloning and characterization of a cDNA coding for Astacus embryonic astacin,
a member of the astacin family of metalloproteases from
the crayfish Astacus astacus. Eur. J. Biochem. 253:
796–803.
Gorbea, C. M., P. Marchand, W. Jiang, N. G. Copeland, D. J.
Gilbert, N. A. Jenkins, and J. S. Bond. 1993. Cloning, expression, and chromosomal localization of the mouse meprin beta
subunit. J. Biol. Chem. 268:21035–21043.
Greenwood, P. H. 1959. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae), Part III. Bull. Br. Mus.
Nat. Hist. (Zool.). 5:179–218.
Greenwood, P. H. 1974. The cichlid fishes of Lake Victoria, East
Africa: the biology and evolution of a species flock. Bull. Br.
Mus. Nat. His. (Zool.) Suppl. 6:1–134.
Greenwood, P. H. 1980. Towards a phyletic classification of the
ÔgenusÕ Haplochromis (Pisces, Cichlidae) are related taxa.
Part II: the species from Lakes Victoria, Nabugabo, Edward,
George, and Kivu. Bull. Br. Mus. Nat. His. (Zool.) 39:
1–101.
Helfman, G. S., B. B. Collette, and D. E. Facey. 1997. The diversity of fishes. Blackwell Science, Inc., Oxford.
Hung, C. H., H. R. Huang, C. J. Huang, F. L. Huang, and G. D.
Chang. 1997. Purification and cloning of carp nephrosin, a secreted zinc endopeptidase of the astacin family. J. Biol. Chem.
272:13772–13778.
Inohaya, K., S. Yasumasu, M. Ishimaru, A. Ohyama, I. Iuchi, and
K. Yamagami. 1995. Temporal and spatial patterns of gene
expression for the hatching enzyme in the teleost embry,
Oryzias latipes. Dev. Biol. 171:374–385.
Jiang, W., C. M. Gorbea, A. V. Flannery, R. J. Beynon, G. A.
Grant, and J. S. Bond. 1992. The alpha subunit of meprin
A. Molecular cloning and sequencing, differential expression
in inbred mouse strains, and evidence for divergent evolution of the alpha and beta subunits. J. Biol. Chem. 267:
9185–9193.
Johnson, T. C., C. A. Scholz, M. R. Talbot, K. Kelts, R. D.
Ricketts, G. Ngobi, K. Beuning, I. I. Ssemmanda, and J. W.
McGill. 1996. Late Pleistocene desiccation of Lake Victoria
and rapid evolution of cichlid fishes. Science 273:
1091–1093.
Kawakami, K., H. Takeda, N. Kawakami, M. Kobayashi, N.
Matsuda, and M. Mishina. 2004. A transposon-mediated gene
trap approach identifies developmentally regulated genes in
Zebrafish. Dev. Cell 7:133–144.
Kessler E., K. Takahara, L. Biniaminov, M. Brusel, and D. S.
Greenspan. 1996. Bone morphogenetic protein-1: the type I
procollagen C-proteinase. Science 271:360–362.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001.
MEGA2: molecular evolutionary genetics analysis software.
Bioinformatics 17:1244–1245.
Li S. W., A. L. Sieron, A. Fertala, Y. Hojima, W. V. Arnold, and
D. J. Prockop. 1996. The C-proteinase that processes procollagens to fibrillar collagens is identical to the protein previously
identified as bone morphogenic protein-1. Proc Natl Acad Sci
USA 93:5127–5130.
Luo, L., R. C. Salunga, H. Guo et al. (11 co-authors). 1999.
Gene expression profiles of laser-captured adjacent neuronal
subtypes. Nat. Med. 5:117–122.
Michaut, L., S. Flister, M. Neeb, K. P. White, U. Certa, and W. J.
Gehring. 2003. Analysis of the eye developmental pathway in
Drosophila using DNA microarrays. Proc. Natl. Acad. Sci.
USA 100:4024–4029.
Murakami, Y., M. Ogasawara, F. Sugahara, S. Hirano, N. Satoh,
and S. Kuratani. 2001. Identification and expression of the
lamprey Pax6 gene: evolutionary origin of the segmented brain
of vertebrates. Development 128:3521–3531.
Nagl, S., H. Tichy, W. E. Mayer, N. Takezaki, N. Takahata, and J.
Klein. 2000. The origin and age of haplochromine fishes in
Lake Victoria, east Africa. Proc. R. Soc. Lond. B. Biol. Sci.
267:1049–1061.
Nakazawa, M., F. Cebria, K. Mineta, K. Ikeo, K. Agata, and T.
Gojobori. 2003. Search for the evolutionary origin of a brain:
planarian brain characterized by microarray. Mol. Biol. Evol.
20:784–791.
Perris, R., and D. Perissinotto. 2000. Role of the extracellular matrix during neural crest migration. Mech. Dev. 95:
3–21.
Scott, I. C., I. L. Blitz, W. N. Pappano et al. (11 co-authors).
1999. Mammalian BMP-1/Tolloid-related metalloproteinases,
including novel family member mammalian Tolloid-like 2,
have differential enzymatic activities and distributions of
expression relevant to patterning and skeletogenesis. Dev. Biol.
213:283–300.
Seehausen, O. 1996. Lake Victoria Rock Cichlids. Verduyn
Cichlids, Zevenhuizen.
Seehausen, O., J. J. M. van Alphen, and F. Witte. 1997. Cichlid
fish diversity threatened by eutrophication that curbs sexual
selection. Science 277:1808–1811.
Segal, E., M. Shapira, A. Regev, D. Pe’er, D. Botstein, D. Koller,
and N. Friedman. 2003. Module networks: identifying regulatory modules and their condition-specific regulators from gene
expression data. Nat. Genet. 34:166–176.
Shimell, M. J., E. L. Ferguson, S. R. Childs, and M. B. O’Connor.
1991. The Drosophila dorsal-ventral patterning gene tolloid
is related to human bone morphogenetic protein 1. Cell 67:
469–481.
Skaer, N., D. Pistillo, and P. Simpson. 2002. Transcriptional
heterochrony of scute and changes in bristle pattern
between two closely related species of blowfly. Dev. Biol.
252:31–45.
Stocker, W., B. Sauer, and R. Zwilling. 1991. Kinetics of
nitroanilide cleavage by astacin. Biol. Chem. Hoppe-Seyler
372:385–392.
Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin,
and D. G. Higgins. 1997. The CLUSTAL_X windows
interface: flexible strategies for multiple sequence alignment
aided by quality analysis tools. Nucleic Acids Res.
25:4876–4882.
Turner, G. F., O. Seehausen, M. E. Knight, C. J. Allender, and
R. L. Robinson. 2001. How many species of cichlid fishes
are there in African lakes? Mol. Ecol. 10:793–806.
Uzel, M.I., I.C. Scott, H. Babakhanlou-Chase, A. H.Palamakumbura,
W. N. Pappano, H. H. Hong, D. S. Greenspan, and P. C.
Trackman. 2001. Multiple bone morphogenetic protein 1-related
mammalian metalloproteinases process pro-lysyl oxidase at the
correct physiological site and control lysyl oxidase activation
in mouse embryo fibroblast cultures. J. Biol. Chem. 276:
22537–22543.
Wardle, F. C., J. V. Welch, and L. Dale. 1999. Bone morphogenetic protein 1 regulates dorsal-ventral patterning in early Xenopus embryos by degrading chordin, a BMP4 antagonist.
Mech. Dev. 86:75–85.
Watanabe, M., N. Kobayashi, A. Fujiyama, and N. Okada. 2003.
Construction of a BAC library for Haplochromis chilotes,
a cichlid fish from Lake Victoria. Genes Genet. Syst.
78:103–105.
Watanabe, M., N. Kobayashi, T. Shin-i, T. Horiike, Y. Tateno, Y.
Kohara, and N. Okada. 2004. Extensive analysis of ORF
sequences from two different cichlid species in Lake Victoria
provides molecular evidence for a recent radiation event of the
Victoria species flock: identity of EST sequences between
Haplochromis chilotes and Haplochromis sp. ‘‘Redtailsheller’’.
Gene 343:263–269.
1660 Kijimoto et al.
Yan, L., K. Fei, J. Zhang, S. Dexter, and M. P. Sarras Jr. 2000.
Identification and characterization of hydra metalloproteinase
2 (HMP2): a meprin-like astacin metalloproteinase that functions in foot morphogenesis. Development 127:129–141.
Yan, L., G. H. Pollock, H. Nagase, and M. P. Sarras Jr. 1995.
A 25.7 3 10(3) M(r) hydra metalloproteinase (HMP1), a member of the astacin family, localizes to the extracellular matrix
of Hydra vulgaris in a head-specific manner and has a developmental function. Development 121:1591–1602.
Billie Swalla, Associate Editor
Accepted April 4, 2005