Báo cáo khoa học: The male seahorse synthesizes and secretes a novel C-type lectin into the brood pouch during early pregnancy pdf

The male seahorse synthesizes and secretes a novel C-type
lectin into the brood pouch during early pregnancy
Philippa Melamed, Yangkui Xue, Jia Fe David Poon, Qiang Wu, Huangming Xie, Julie Yeo,
Tet Wei John Foo and Hui Kheng Chua
Department of Biological Sciences, National University of Singapore, Singapore
The seahorse (Hippocampus) species, which are highly
sought after for both ornamental and traditional Chi-
nese medicine purposes, are in danger of extinction and
their culture presents unique problems in aquaculture,
particularly in rearing of the young. The seahorse
belongs to the Syngnathidae family of fish, which
includes also the pipefish, pipehorses and seadragons. In
all of these, the males incubate the young on or within
their bodies. In the seahorse, this incubation resembles a
true male pregnancy, as the female deposits her eggs into
an enclosed brood pouch on the ventral side of the
male’s abdomen. This brood pouch comprises epithelial
and stoma-like tissue which lines a thick muscular wall.
The epithelium thickens and becomes more vascularized
as the reproductive season approaches (Fig. 1). After
uptake and fertilization of the eggs, the pouch is sealed
and the developing embryos become embedded in the
epithelium. Each embryo becomes compartmentalized
as the epithelium forms a surrounding pit in which it
remains until after yolk absorption is complete [1]. The
embryos continue to develop and grow for several weeks
(depending on the species) until they are able to with-
stand the external environmental conditions independ-
ently, at which point the juveniles are released.
Although appearing to be a true male pregnancy, in
contrast to mammals but comparable to most other

Keywords
Hippocampus comes; C-type lectin; cDNA
library; male pregnancy
Correspondence
P. Melamed, Department Biological
Sciences, National University of Singapore,
14 Science Drive 4, Singapore 117542
Fax: +65 6872 2013
Tel: +65 6874 1882
E-mail:
(Received 23 November 2004, revised 26
December 2004, accepted 6 January 2005)
doi:10.1111/j.1742-4658.2005.04556.x
The male seahorse incubates its young in a manner resembling that of a
mammalian pregnancy. After the female deposits her eggs into the male’s
brood pouch they are fertilized and the embryos develop and grow for several
weeks until they are able to withstand the external environmental conditions
independently, at which point they are irreversibly released. Although the
precise function of the brood pouch is not clear, it is probably related to pro-
viding a suitable protective and osmotic environment for the young. The aim
of this project was to construct and characterize a cDNA library made from
the tissue lining the pouch, in order to help understand the molecular mecha-
nisms regulating its development and function. The library profile indicates
expression of genes encoding proteins involved in metabolism and transport,
as well as structural proteins, gene regulatory proteins, and other proteins
whose function is unknown. However, a large portion of the library con-
tained genes encoding C-type lectins (CTLs), of which three full-length
proteins were identified and found to contain a signal peptide and a single
C-lectin domain, possessing all the conserved structural elements. We have
produced recombinant protein for one of these and raised antisera; we have

shown, using Western analysis and 2D electrophoresis, that this protein is
secreted in significant quantities into the pouch fluid specifically during early
pregnancy. Preliminary functional studies indicate that this CTL causes
erythrocyte agglutination and may help to repress bacterial growth.
Abbreviations
AP, alkaline phosphatase; CTL, C-type lectin; CRD, carbohydrate recognition domain; 2DE, 2D gel electrophoresis; DIG, digoxygenin; hcCTL,
Hippocampus comes C-type lectin; HRP, horseradish peroxidase; IPG, immobilized pH gradient; LB, Luria–Bertani; MBP, mannose binding
protein; NBT ⁄ BCIP, Nitro Blue tetrazolium 5-bromo-4-chloroindol-2-yl-phosphate.
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1221
teleost fish, these fry appear to obtain most of their
nutrition from the yolk sac [2]. Instead, the father’s role
seems to be related to providing a suitable osmotic envi-
ronment for the young, while also supplying oxygen and
calcium, and presumably removing waste products [3,4].
Histological studies have demonstrated the presence of
mitochondria-rich cells in the epithelia lining the pouch
which are postulated to act as ion transporters, as they
do in the gills; the number of these increases with dur-
ation of the incubation period, after which they undergo
apoptosis [4]. In the gills, these cells contain receptors to
prolactin which is one of the major piscine osmoregula-
tory hormones [4,5], and also has a central role in
governing parental behaviour in most animals. The
presence of prolactin receptors in the brood pouch,
however, has yet to be reported.
The aim of this project was to construct and charac-
terize a cDNA library made from the epithelium and
stroma-like tissue lining the incubation pouch, in order
to help understand the molecular mechanisms regula-
ting the development and function of this unique male

pregnancy.
Results
Identification of cDNA clones from the pouch
tissue
A cDNA library was constructed from the tissue lin-
ing the incubation pouch, and over 250 cloned
inserts were sequenced; of these 151 were found to
match sequences in the nucleotide and ⁄ or protein
databases. Another 80 inserts appeared to encode
novel proteins for which matches could not be
found. As expected, the identified inserts contained
genes for ubiquitous proteins such as actin, globin,
keratin, ribosomal proteins and also for transferrins,
and generally showed closest matches with homolog-
ous sequences from other teleosts, where available.
All sequences have been entered to the NCBI Gen-
Bank data base (Table 1).
Many of the cloned inserts encode metabolic
enzymes, including those involved in oxidative phos-
phorylation, fatty acid oxidation and reductive biosyn-
thesis. The presence of these enzymes presumably
reflects the large number of mitochondria in this tissue.
Genes encoding putative regulatory proteins were also
identified, including those for kinases, transcription
factors and binding proteins, indicating that this tissue
is probably regulated by specific signalling pathways.
Genes encoding proteases and protease inhibitors were
also present and a gene with high homology to the
carp zinc endopeptidase, nephrosin, was identified.
This proteinase, which is stimulated by high concentra-

tions of potassium, is expressed specifically in immune
and hematopoietic tissue in carp and shares some
homology with other members of the astacin or fish
hatching enzyme family [6].
By far the most common inserts, however, were
cDNAs encoding proteins with homology to various
C-type lectins (CTL); these comprised inserts in over
15% of all of the clones sequenced.
Fig. 1. Morphology of the seahorse brood
pouch. (A) The brood pouch consists of a
muscular wall (#) which is lined with an
easily detachable layer of stroma (*) and
epithelium (e) which extends towards the
incubation cavity. (B) By the time the male
is ready to receive the female eggs, the epi-
thelium has thickened and is well vascular-
ized (arrow marks blood vessels). (C) With
uptake and fertilization of the eggs, the epi-
thelium becomes more extensive and enve-
lopes the developing embryos (Em). (D) By
the time the fully developed young seahors-
es are hatched and getting ready to leave
the pouch, this tissue has thinned consider-
ably.
C-type lectins in the male seahorse pregnancy P. Melamed et al.
1222 FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS
Table 1. Identified cDNA clones from male seahorse brood pouch, based on gene and ⁄ or protein comparisons.
Clone Gene Protein Accession number
YK1 Beta globin [Oryzias latipes] (4e-87) Adult beta-type globin [O. latipes] (1e-54) CV863925
YK2 Serum lectin isoform 1 precursor [Salmo salar] (1e-23) CV863926

YK3 C-type lectin [Anguilla japonica] (4e-15) CV863927
YK4 NADH ubiquinone oxidoreductase 49 kDa
subunit [Bos taurus] (3e-32)
NADH2 dehydrogenase 49 kDa subunit [B. taurus] (8e-88) CV863928
YK5 C-type lectin 2 [A. japonica] (4e-16) CV863929
YK7 Myosin regulatory light
chain 2 [Mus musculus] (9e-70)
Myosin regulatory light chain 2 [Gallus gallus] (7e-47) CV863930
YK8 FC-epsilon RII [M. musculus] (7e-18) CV863931
YK10 Zymogen granule protein 16 [Homo sapiens] (6e-16) CV863932
YK13 Polyubiquitin [Arabidopsis lyrata] (4e-92) CV863933
YK14 40S Ribosomal protein S25
[Ictalurus punctatus] (1e-61)
Similar to ribosomal protein S25 [Rattus norvegicus] (2e-32) CV863934
YK16 ATPase subunit 8 (ATPase8)
and ATPase subunit 6 (ATPase6)
[Rhamdia sp.] (7e-30)
ATP synthase F0 subunit 6 [Emmelichthys struhsakeri] (4e-75) CV863935
YK20 Lysyl-tRNA synthetase
[Xenopus laevis] (1e-13)
Lysyl-tRNA synthetase [X. laevis] (1e-60) CV863936
YK23 Serotransferrin precursor [O. latipes] (3e-46) CV863937
YK26 Farnesyl diphosphate farnesyl
transferase 1 [H. sapiens] (3e-12)
Farnesyl diphosphate farnesyl transferase 1 [R. norvegicus] (6e-50) CV863938
YK29 Nephrosin precursor [Cyprinus carpio] (3e-35) CV863939
YK35 Clone MGC:55674 [Danio rerio] (1e-18) Makorin 3 [zinc finger protein 127] [M. musculus] (2e-05) CV863940
YK37 Brevican core protein [M. musculus] (5e-15) CV863941
YK39 Ribosomal L6 [Pargus major] (1e-111) 60S ribosomal protein L6 [R. norvegicus] (6e-61) CV863942
YK40 Galectin-like protein

[Oncorhynchus mykiss] (2e-09)
Galectin like protein [O. mykiss] (3e-56) CV863943
YK41 Adult beta type globin
[O. latipes] (6e-79)
Adult beta type globin [O. latipes] (3e-55) CV863944
YK43 Actin related protein 2
homolog [X. laevis] (2e-13)
CV863945
YK45 Transferrin [Melanogrammus aeglefinus] (1e-25) CV863946
YK46 Mannose receptor precursor [G. gallus] (2e-05) CV863947
YK47 Actin-like protein [G. gallus] (8e-93) CV863948
YK49 Novel protein similar to vertebrate mitochondrial enoyl
Coenzyme A hydratase 1 (ECHS1) [D. rerio] (2e-39)
CV863949
YK50 C-type lectin 2 [A. japonica] (4e-15) CV863950
YK51 Cytochrome c oxidase subunit II [Exocoetus volitans] (1e-110) CV863951
YK52 Serotransferrin precursor [O. latipes] (3e-67) CV863952
YK54 Brevican core protein [M. musculus] (2e-15) CV863953
YK55 Brevican core protein [M. musculus] (2e-15) CV863954
YK56 Neurocan core protein precursor [H. sapiens] (9e-11) CV863955
YK57 Transketolase [P. flesus] (1e-15) CV863956
YK59 Similar to ADP-ribosylation factor 2 [M. musculus] (2e-05) CV863957
YK61 DJ-1 [S. salar] (3e-31) Similar to DJ-1 protein [M. musculus] (5e-69) CV863958
YK62 FC-epsilon RII [H. sapiens] (5e-06) CV863959
YK63 Ribosomal protein L23
[Gillichthys mirabilis] (1e-24)
60S ribosomal protein L23 [H. sapiens] (5e-22) CV863960
YK64 Microsatellite marker
[Poecilia reticulata] (8e-54)
CV863961

YK66 C-type lectin 2 [A. japonica] (4e-15) CV863962
YK67 Beta actin 1
[Takifugu rubripes] (1e-101)
Actin [Strongylocentrotus purpuratus] (2e-20) CV863963
YK68 C-type lectin 2 [A. japonica] (7e-15) CV863964
YK69 Gluthionine S-transferase [H. sapiens] (2e-40) CV863965
YK70 C-type lectin 2 [A. japonica] (1e-16) CV863966
P. Melamed et al. C-type lectins in the male seahorse pregnancy
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1223
Table 1. (Continued).
Clone Gene Protein Accession number
YK72 Cytochrome c sububit 1 [Trachipterus trachypterus] (4e-14) CV863967
YK74 Serotransferrin II precursor [S. salar] (6e-23) CV863968
YK75 C-type mannose-binding lectin [O. mykiss] (6e-08) CV863969
YK78 Flavin reductase (NADPH) H. sapiens (1e-38) CV863970
YK79 C-type lectin 2 [A. japonica] (7e-12) CV863971
YK80 C-type lectin 2 [A. japonica] (4e-10) CV863972
YK81 c-src family protein tyrosine kinase [T. rubripes] (3e-29) CV863973
YK82 Transferrin
[Pagrus major] (8e-40)
Transferrin [D. rerio] (7e-44) CV863974
YK84 Ornithine decarboxylase
antizyme [D. rerio] (2e-43)
Ornithine decarboxylase antizyme [D. rerio] (7e-25) CV863975
YK85 Arachidonate 15-lipoxygenase type II [H. sapiens] (2e-08) CV863976
YK86 Type II keratin
[O. mykiss] (5e-74)
Type II cytokeratin [D. rerio] (4e-62) CV863977
YK87 Transferrin [O. latipes] (3e-65) CV863978
YK91 DNA sequence from

clone XX-184L24
[D. rerio] (1e-12)
Novel protein [D. rerio] (1e-46) CV863979
YK92 Retinoic acid binding
protein 1-cellular
[H. sapiens] (1e-18)
Retinoic acid binding protein 1-cellular [T. rubripes] (4e-60) AY437393
YK95 Metalloproteinase inhibitor 4 precursor [R. norvegicus] (4e-13) CV863980
YK98 NIKs-related kinase
[H. sapiens] (8e-08)
Traf2 and NCK interacting kinase [H. sapiens] (2e-14) CV863981
YK99 Ferritin heavy subunit
[S. salar] (6e-16)
Selenocysteine methyltransferase [Astragalus bisulcatus] (5e-12) CV863982
YK102 Ribosomal protein L21
[I. punctatus] (5e-39)
Ribosomal protein L21 [I. punctatus] (1e-55) AY357070
YK103 EF1alpha [Drosophila
melanogaster] (1e-36)
CV863983
YK104 Ribosomal protein L35
[I. punctatus] (7e-29)
60S ribosomal protein L35 [Sus scrofa] (1e-42) AY357071
WQ4 Cytochrome c oxidase polypeptide subunit VIb
[H. sapiens] (4e-14)
CV863984
WQ5 Cytochrome c oxidase subunit I [Mugil cephalus] (1e-11) CV863985
WQ6 Programmed cell death 6
[M. musculus] (1e-06)
Programmed cell death protein 6 [M. musculus] (2e-37) CV863986

WQ7 Ribosomal protein S19
[Gillichthys mirabilis] (8e-72)
Ribosomal protein S19 (3e-62) CV878464
WQ18 eEF-1 beta [X. laevis] (6e-14) CV863987
WQ19 DEAD (Asp-Glu-Ala-Asp) box
polypeptide (D. rerio) [1e-11]
Similar to Eukaryotic initiation factor 4a [D. rerio] (8e-08) CV863988
WQ22 Ependymin 2 [M. musculus] (7e-15) CV863989
WQ25 Hypothetical protein MGC63929 [D. rerio] (1e-29) CV863990
WQ27 Transferrin
[Gadus morhua] (1e-07)
Serotransferrin I precursor [S. salar] (2e-20) CV863991
WQ29 C-type lectin 2 [A. japonica] (6e-11) CV863992
WQ30 C-type lectin 2 [A. japonica] (2e-11) CV863993
WQ31 Similar to ATP synthase H
+
transporting, mitochondrial
F0 complex, subunit c
(subunit 9) isoform 3
[X. laevis] (5e-40)
Similar to ATP synthase C, subunit C, isoform 3 [D. rerio] (1e-35) CV863994
WQ32 40S ribosomal protein S15A
[Paralichthys olivaceus] (1e-115)
40S ribosomal protein S15A [P. olivaceus] (1e-56) AY319480
WQ33 Similar to CG13623-PA [H. sapiens] (1e-28) CV863995
C-type lectins in the male seahorse pregnancy P. Melamed et al.
1224 FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS
Table 1. (Continued).
Clone Gene Protein Accession number
WQ34 ATP synthase, H

+
transporting, mitochondrial F1 complex,
O subunit [B. taurus] (9e-52)
CV863996
WQ36 C-type lectin 2 [A. japonica] (4e-10) CV863997
WQ39 Ribosomal protein L38
[Branchiostoma belcheri] (2e-24)
Similar to ribosomal protein L38, cytosolic [R. norvegicus] (7e-14) CV863998
WQ40 Cytochrome c oxidase subunit II [E. volitans] (3e-70) CV863999
WQ42 Chromosome 20 ORF 42
(C20orf42) [H. sapiens] (1e-06)
Protein c20orf42 homolog [M. musculus] (1e-69) CV864000
WQ43 Transferrin [O. latipes] (0.59) Transferrin [O. latipes] (3e-13) CV864001
WQ44 C-type mannose-binding lectin [O. mykiss] (3e-09) CV864002
WQ51 FC-epsilon-RII (9e-07) CV864003
WQ52 Ferritin heavy subunit
[Oreochromis mossambicus] (2e-64)
Ferritin H [S. salar] (1e-67) CV864004
WQ56 Ribosomal protein L18
[Oreochromis niloticus] (5e-16)
Ribosomal protein L18 [S. salar] (2e-08) CV864005
WQ59 Ferritin heavy subunit
[S. salar] (2e-60)
Ferritin heavy subunit; ferritin H [S. salar] (4e-56) CV864006
WQ60 Villin 2 [ezrin] (VIL2)
[B. taurus] (6e-09)
Ezrin [G. gallus] (2e-23) CV864007
WQ62 40S ribosomal protein S28
[I. punctatus] (2e-6)
40S ribosomal protein S28 [I. punctatus] (7e-12) AY357067

WQ63 40S ribosomal protein S29
[I. punctatus] (2e-25)
40S ribosomal protein S29 [I. punctatus] (1e-21) AY357068
WQ65 Similar to cystatin B (stefin B) [D. rerio] (7e-23) CV864008
WQ69 TANK-binding kinase 1 [M. musculus] (2e-10) CV864009
WQ70 Lithostathine 1 beta [H. sapiens] (2e-07) CV864010
WQ71 Hypothetical protein XP_148064 [M. musculus] (2e-19) CV864011
WQ72 C-type mannose-binding lectin [O. mykiss] (5e-07) CV864012
WQ73 Haplotype VIB.313 cytochrome b
[Hippocampus comes] (0)
Cytochrome b [Hippocampus comes] (1e-89) AF192657
WQ74 Lectin C-type domain containing protein
[Caenorhabditis elegans] (1e-08)
CV864013
WQ75 Type II keratin E3
[O. mykiss] (2e-58)
Type II keratin E3 [O. mykiss] (5e-24) CV864014
WQ76 Serotransferrin precursor [O. latipes] (3e-30) CV864015
WQ77 Similar to eIF3 subunit 9
[M. musculus] (6e-18)
Eukaryotic translation initiation factor 3 subunit 9 [H. sapiens] (1e-11) CV864016
WQ78 C-type lectin 2 [A. japonica] (2e-11) CV864057
WQ79 (i) Kinesin light chain
[G. gallus] (7e-56);
(ii) 40S ribosomal protein S2
[R. norvegicus] (1e-51)
40S ribosomal protein S2 [I. punctatus] (2e-37) CV864058
WQ81 Similar to lysyl-tRNA synthetase
[M. musculus] (3e-17)
Lysyl-tRNA synthetase [X. laevis] (3e-49) CV864059

WQ82 Hypothetical protein LOC51255
[D. rerio] (9e-11)
Zinc finger protein 364 [M. musculus] (7e-11) CV864060
WQ83 Transferrin [O. latipes] (0.52) Transferrin [Salvelinus namaycush] (5e-10) CV864061
WQ86 Metalloproteinase inhibitor 2 precursor (TIMP-2) [Canis familiaris] (1e-22) CV864062
WQ87 cAMP responsive element binding protein-like 2 [H. sapiens] (1e-21) CV864056
WQ89 Lectin C-type [C. elegans] (1e-06) CV864017
WQ90 Elongation factor 1-alpha
[Sparus aurata] (2e-22)
CV864018
WQ93 C-type lectin 2 [A. japonica] (5e-12) CV864019
WQ95 Transferrin Salmo trutta (4e-31) CV864020
WQ97 Lectin C-type domain containing protein [C. elegans] (3e-15) CV864021
WQ99 Serum lectin isoform 1 precursor [S. salar] (1e-15) CV864022
P. Melamed et al. C-type lectins in the male seahorse pregnancy
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1225
Table 1. (Continued).
Clone Gene Protein Accession number
WQ100 Lectin C-type domain containing protein precursor family member
[C. elegans] (2e-15)
CV864023
WQ101 Similar to opioid receptor, sigma 1 [D. rerio] (1e-42) CV864024
WQ102 Cyclophilin A
[Canis familiaris] (2e-18)
Peptidylprolyl isomerase F (cyclophilin F) [H. sapiens] (6e-47) CV864025
WQ104 40S ribosomal protein S30
[I. punctatus] (1e-42)
40S ribosomal protein S30 [I. punctatus] (7e-51) AY357069
WQ105 C-type lectin 2 [A. japonica] (1e-11) CV864026
WQ106 ADP,ATP translocase

[P. flesus] (8e-19)
ADP,ATP translocase [P. flesus] (1e-14) CV864027
WQ107 Similar to ribosomal
protein L27 [D. rerio] (5e-89)
Similar to ribosomal protein L27[H. sapiens] (3e-53) AY437394
WQ110 Similar to retinoid-inducible
serine caroboxypetidase
[D. rerio] (8e-08)
Similar to retinoid-inducible serine caroboxypetidase [D. rerio] (4e-55) CV864028
WQ111 Heat shock protein 90 beta
[P. flesus] (3e-13)
Heat shock protein 90 beta [P. flesus] (2e-09) CV864029
WQ113 Chitinase3 [P. olivaceus] (4e-13) CV864030
WQ114 ATPase subunit 8 (ATPase8) and
ATPase subunit 6 (ATPase6) – mito-
chondrial [Rhamdia laticauda] (4e-17)
ATP synthase F0 subunit 6 [P. olivaceus] (4e-25) CV864031
WQ115 Microsatellite marker Pret-15
[Poecilia reticulata] (1e-37)
CV864032
WQ116 Lectin C-type domain containing protein [C. elegans] (5e-07) CV864033
WQ118 C-type lectin 2 [A. japonica] (4e-16) CV864034
WQ119 Leucine-rich repeat-containing
protein 8 [R. norvegicus] (5e-63)
Leucine-rich repeat-containing protein 8 [M. musculus] (9e-79) CV864035
WQ124 Lectin C-type domain containing protein [C. elegans] (2e-08) CV864036
WQ127 Fructose-1, 6-bisphosphate
aldolase [Sparus aurata] (6e-55)
Fructose-1, 6-bisphosphate aldolase [S. aurata] (2e-67) CV864037
WQ130 Ribosomal protein L19 mRNA

[I. punctatus] (4e-99)
Ribosomal protein L19 [I. punctatus] (2e-64) CV864038
WQ131 Ribosomal protein L31 mRNA
[P. olivaceus] (1e-126)
60S ribosomal protein L31 [P. olivaceus] (2e-46) CV864039
WQ133 Cisplatin resistance related protein
mRNA Length ¼ 2058
[M. musculus] (3e-60)
CRR9p (Crr9-pending), Crr9-pending protein
[M. musculus] (2e-64)
AY437395
WQ134 Machado-Joseph disease protein 1 (Ataxin-3) [M. musculus] (6e-63) CV864040
WQ135 Cytochrome c oxidase subunit VIII
liver form (COX8L) mRNA
[Trachypithecus cristatus] (0.054)
Cytochrome c oxidase subunit VIII liver form [Eulemur fulvus] (9e-08) CV864041
WQ136 Mannose receptor, C type 2; novel lectin [M. musculus] (4e-07) CV864042
WQ137 Eukaryotic translation initiation
factor gamma 2, subunit 3
[D. rerio] (3e-33)
Eukaryotic translation initiation factor 2G; eukaryotic translation
initiation factor 2, subunit 3 (gamma, 52 kDa) [H. sapiens] (1e-99)
CV864043
WQ138 Fatty acyl-CoA hydrolase precursor, medium chain
(thioesterase B) [Anas platyrhynchos] (4e-48)
CV864044
WQ139 Transferrin [O. latipes] (0.85) Transferrin [Oncorhynchus nerka] (4e-31) CV864045
WQ140 RAB26, member RAS
oncogene family (Rab26),
mRNA [R. norvegicus] (1e-07)

RAB37, member of RAS oncogene family; GTPase Rab37
[M. musculus] (1e-22)
CV864046
WQ142 Translocon-associated protein
alpha mRNA [D. rerio] (3e-05)
Translocon-associated protein alpha [D. rerio] (2e-17) CV864047
WQ147 Cytokeratin mRNA
Stizostedion vitreum vitreum] (4e-15)
Type I cytokeratin, enveloping layer;
type I cytokeratin [D. rerio] (1e-38)
CV864048
C-type lectins in the male seahorse pregnancy P. Melamed et al.
1226 FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS
Three different CTLs are expressed
in the incubation pouch
The inserts encoding CTL-like proteins were aligned
and found to comprise three different sequences. For
each of these, a full-length sequence was found in the
library, and the deduced proteins were aligned. Two
of the Hippocampus comes CTLs (hcCTLs), types I
and III are highly similar, while a third, type II differs.
Alignment with the C-type lectins found in whole body
extracts of H. kuda and in the gills of the Japanese eel
[7,8], reveals similarity with the hcCTL II, but less so
to the other two hcCTLs (Fig. 2A). All three novel
hcCTLs contain a signal peptide and a single C-type
lectin domain without other associated domains
(Fig. 2A), defining them as group VII lectins. They
contain many of the 37 residues of the C-type carbohy-
drate recognition domain (CRD), as defined by Weis

et al. [9], as well as six conserved cysteines (Fig. 2A).
The secondary structure of hcCTL III is predicted to
form two helices at the N-terminal end, eight strands
and three disulphide bridges (Fig. 2B). The five residues
crucial in determining mannose binding specificity [10]
are absent in all of the hcCTLs (Fig. 2B), although the
hcCTL II and most of the other aligned CTLs contain
the QPD motif endowing galactose specificity (Fig. 2A).
However, the highly conserved proline contained within
QPD is found in all the CTLs shown (Figs 2A and B).
In situ hybridization confirmed the specific expres-
sion of the hcCTL III in the tissue lining the brood
pouch. Using a digoxygenin (DIG)-labelled 300-bp
fragment of the cDNA, a particularly strong signal
was seen in the stroma-like pouch lining which exten-
ded in the cavity along the epithelial protrusions that
surround the developing embryos. The negative control
completely lacked this signal (Figs 3A and B).
2D gel electrophoresis reveals that hcCTL III
is secreted into the brood pouch
To verify that the hcCTLs are indeed secreted into the
pouch cavity, and to examine other proteins present in
the fluid surrounding the embryos, the proteome of the
pouch fluid of a single incubating male was examined
using 2D gel electrophoresis (2DE) over a pI range of
3–10. After silver staining, several proteins were vis-
ible, the most prominent of which had a low pI and
an apparent relative molecular mass just over 15 kDa
(Fig. 4); this matches the predicted relative molecular
mass (16 kDa) and pI (4) of the hcCTLs identified in

the cDNA library. This protein spot was cut and tryp-
sin-digested for peptide fingerprinting using MALDI
MS. Comparison of the peptide masses with the
deduced peptides for the three hcCTLs revealed pep-
tides that matched the predicted sizes for the novel
hcCTL III and covered 28% of the mature protein.
Analysis of the levels of lectin proteins in the
pouch fluid during pregnancy
The cDNA encoding the hcCTL III was expressed in
Escherichia coli and the recombinant protein (shown in
Table 1. (Continued).
Clone Gene Protein Accession number
WQ149 Chromosome 20 open reading
frame 52 (C20orf52), mRNA
[H. sapiens] (2e-28)
Chromosome 20 open reading frame 52; homolog of mouse
RIKEN 2010100O12 gene [H. sapiens] (2e-21)
CV864049
WQ150 40S ribosomal protein S15A mRNA,
complete [P. olivaceus]
(1e-119)
40S ribosomal protein S15A [P. olivaceus] (3e-61) AY319480
WQ154 mRNA for embryonic alpha-type
globin [O. latipes] (9e-29)
Embryonic alpha-type globin [O. latipes] (1e-52) CV864050
WQ156 Type I cytokeratin (cki),
mRNA [D. rerio] (3e-06)
Type I cytokeratin, enveloping layer; type I cytokeratin
[D. rerio] (3e-18)
CV864051

WQ158 Lectin C-type domain containing protein precursor family
member [C. elegans] (1e-09)
CV864052
WQ159 C-type lectin 2 [A. japonica] (4e-15) CV864053
WQ162 Alpha tubulin mRNA
[Notothenia coriiceps] (1e-134)
Tubulin alpha chain [Notophthalmus viridescens] (3e-64) CV864054
WQ166 Ribosomal protein L28 mRNA
[I. punctatus] (1e-49)
60S ribosomal protein L28 [H. sapiens] (7e-55) AY437397
WQ168 S6 ribosomal protein mRNA
[O. mykiss] (1e-123)
40S S6 ribosomal protein [O. mykiss] (8e-63) CV864055
P. Melamed et al. C-type lectins in the male seahorse pregnancy
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1227
Fig. 5A, lane 3 after elution from Ni-NTA affinity col-
umn) was used to raise antisera in rabbits. The anti-
sera from one of the rabbits was highly specific,
reacting with only a single sized protein in the pouch
fluid of a pregnant but not a nonpregnant seahorse
(Fig. 5B), this reactive protein was not apparent when
A
B
Fig. 2. Three novel H. comes brood pouch C-lectins are homologous with similar proteins from other species and show conserved structural
constaints. (A) The three CTLs identified from screening of the pouch cDNA library (HcI, HcII and HcIII) are aligned with five CTL protein
sequences found in whole body extracts of H. kuda (H00011, H00359, H00385, H00386, H00395 [8]) and two isolated from the gills of the
Japanese eel (Eel1, Eel2 [7]). All of the H. comes and eel CTLs and one H. kuda CTL (H00386) contain a signal peptide (underlined). Con-
served residues of CTLs, as defined by Weis et al. [9] are shown in bold; the six cysteines are marked with asterisks, and the QPD motif
determining galactose binding, where present, is boxed. (B) The predicted structure of hcCTL III, comprising two helices at the N terminus
(marked in bold), eight strands (S1–S8; underlined: both predicted using

PSIPRED at and the three disulphide
bridges (joined by lines and labelled with boxed numbers) are shown. The five residues comprising the part of the CRD that determines
mannose binding (according to Drickamer [12]) are noted in italics above the sequence.
C-type lectins in the male seahorse pregnancy P. Melamed et al.
1228 FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS
the preimmune rabbit sera was used (data not shown).
Given the similarity of protein sequences between the
three novel seahorse lectins and their close sizes, this
reactive band could, however, represent more than just
the hcCTL III. Samples of the pouch fluid from sea-
horses at various stages of incubation were collected
and run on SDS gels for Western analysis to com-
pare the levels of the immunoreactive (ir)-hcCTL III
proteins. The ir-hcCTL III protein was detected only
during incubation of early embryos, but not the devel-
oped seahorses, and was also undetectable both before
uptake of the eggs and after hatching and release of
the juveniles (Fig. 5C).
Functional analysis of hcCTL III
In order to verify a possible antibacterial role for the
novel hcCTL III, bacteriostatic tests were performed.
These involved incubation of E. coli cells with or with-
out addition of the recombinant hcCTL III for up to
2 h, during which the growth of the bacteria was
assessed by O.D. readings every 30 min. Under these
conditions, hcCTL III at a final concentration of
0.7 lm started to inhibit E. coli growth after 1.5 h, and
reached a 25% reduction after 2 h (Fig. 6).
The ability of the novel hcCTL III to recognize cell-
surface glycoproteins was assessed using a haemagglu-

tination assay. Concentrations of 2.25–18 lm of the
hcCTL III were able to agglutinate mouse red blood
cells after 1–1.5 h of incubation (Fig. 7A). In an
attempt to identify the sugars bound by the lectin, the
same assay was repeated after addition of various
mono-, di- and complex carbohydrates, including
mannose, galactose, glucose, maltose, sucrose, fructose,
raffinose, N-acetyl glucosamine and N-acetyl galactosa-
mine, using hcCTL III at a final concentration of
4.5 lm. However none of these was able to inhibit the
agglutination, even at a concentration of 100 mm
(Fig. 7B and not shown).
Discussion
We have created and partially characterized a cDNA
library comprising genes expressed in the epithelium
and stroma-like tissue lining the male seahorse brood
pouch. The profile indicates a high level of expression
of genes encoding proteins involved in metabolism and
transport, as well as structural proteins, gene regula-
tory proteins, and other proteins whose function is
Fig. 3. Confirmation of expression of hcCTL III in the pouch tissue
by in situ hybridization. (A) H. comes pouch tissue was formalin-
fixed and paraffin-embedded before sectioning at 6–8 l
M. The
cDNA for the novel hcCTL III was labelled with DIG and detected
using AP-conjugated antisera and NBT ⁄ BCIP, to give a dark purple
reaction product (*). (B) The negative control, which lacks the same
intense staining, is also shown.
Fig. 4. 2DE of the brood pouch fluid proteome reveals that
hcCTL III is secreted. The incubation fluid that surrounds the sea-

horse embryos was extracted from the pouch of a pregnant male
H. comes comes for analysis of the proteome. The proteins were
separated using 2DE (over the pI range 3–10), and a prominent pro-
tein spot (circled) corresponding to the approximate mass and pI of
the novel hcCTL proteins ( 16 kDa, pI 4) was cut and digested
with trypsin, for peptide fingerprinting using MALDI MS. Of the
peptides obtained, three matched the predicted sizes for the novel
hcCTL III, covering 28% of the mature protein.
P. Melamed et al. C-type lectins in the male seahorse pregnancy
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1229
unknown. However, an unusually large portion of the
library contained genes encoding CTLs. Three full-
length CTLs were identified, which share some similar-
ity to CTLs expressed H. kuda and to a lesser degree,
those in the gills of the Japanese eel [7,8]. The localiza-
tion of hcCTL III mRNA transcripts specifically in the
stroma-like tissue and epithelium of the pouch tissue
was confirmed by in situ hybridization, while 2DE and
Western analysis revealed that it is secreted into the
incubation fluid that surrounds the embryos during
early pregnancy.
CTLs are found universally in eukaryotes and pro-
karyotes and have diverse functions [10]. Although
often containing several domains, they are character-
ized by their ability to bind carbohydrates in a cal-
cium-dependent manner, through a CRD. The CRD
contains two a helices and several strands separated by
loops [11]. At least three disulphide bridges are com-
mon in the long form (approximately 130 residues),
one of which spans from the end of the first helix to

the end of the CRD, the second is shorter and located
at the C-terminal end of the CRD, and the third is
found towards the N-terminal end and spans the first
strand; the latter is lacking in the short (i.e. 115 resi-
due) form. All of the cysteines forming these bridges
are found in the conserved locations in the novel
hcCTLs, as are the positions of the two a helices.
Fig. 6. The novel hcCTL III inhibits growth of E. coli. E. coli cells
(1 mL at an D
595
of 0.1) were incubated with recombinant hcCTL III
at 0.7 l
M, or vehicle alone, for up to 2 h, and D
595
readings taken
every 30 min to assess the rate of bacterial growth. The D values
were calculated relative to the initial readings in the same samples.
An asterisk denotes mean values statistically different (Welch two-
sample t-test, P < 0.05) in hcCTL-treated and control samples
(mean ± SEM, n ¼ 4).
A
B
Fig. 7. The hcCTL III causes erythrocyte agglutination which is not
inhibited by common sugars. (A) A haemagglutination assay was
carried out to test the ability of the hcCTL III to cause erythrocyte
agglutination. After 1 h incubation of mouse erythrocytes with
hcCTL III at 2.25–18 l
M, plaque formation resulted indicating ability
of the hcCTL III to cause agglutination which was absent in the
control samples. (B) In order to verify the carbohydrates recognized

by the hcCTL III, the same assay was repeated using 4.5 l
M
hcCTL III with the addition of fructose, sucrose, maltose, glucose,
galactose or mannose at 12.5–100 m
M. However, no inhibition of
agglutination was apparent with addition of any of the sugars. +C,
Positive control to which no sugars were added; -C, negative con-
trol in which hcCTL III was lacking.
A
C
B
Fig. 5. The amounts of ir-hcCTL III in the pouch fluid vary with pro-
gression of pregnancy. (A) Recombinant hcCTL III was raised and
purified on a Ni–NTA affinity column; the cell lysate (lane 1), column
flow-through (lane 2) and eluted protein (lane 3) are shown on an
SDS ⁄ PAGE gel (12%) stained with Coomassie blue. (B) The eluted
recombinant hcCTL III was used to raise antisera, which recognized
just a single sized-protein in the pouch fluid of a pregnant male
(third lane); shown also are the rainbow marker (first lane) and fluid
from a nonpregnant male (second lane). The proteins were
resolved on an SDS ⁄ PAGE gel (12%); primary antisera was used at
1 : 1000 dilution, and a goat antirabbit IgG–HRP-conjugated secon-
dary antibody (at 1 : 1000 dilution) was used for detection by
chemiluminescence. (C) This antisera was then used in the same
manner to compare levels of ir-hcCTL III in the same volume of
pouch fluid for individuals at various stages of pregnancy: before
uptake of the eggs, during incubation of the developing embryos or
seahorses, or after their release.
C-type lectins in the male seahorse pregnancy P. Melamed et al.
1230 FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS

Furthermore, a highly conserved proline positioned to
induce a turn just before the second disulphide bridge
[11] is also conserved. In the mannose binding protein
(MBP) this is thought to position the flanking side
chains so allowing the correct fold of the CRD to faci-
litate Ca
2+
binding. The presence of these structural
constraints supports the likelihood that the novel
hcCTLs are indeed functional lectins.
Lectins are classified into seven groups, according to
their structural arrangement, including the number and
position of the CRD in relationship to the other func-
tional domains [12]. The CTLs revealed in the current
study belong to group VII, as they contain just a single
CRD and a signal peptide, and are clearly secreted
from the cell. Many of these CTLs, as well as those in
group III (i.e. the collectins, including the MBP), have
been implicated in innate immunity which provides a
rapid first line of defence to help prevent pathogen
penetration. An essential part of this response is the
lectin’s recognition of the pathogen’s cell-surface car-
bohydrates as nonself, in order to target the pathogens
for destruction (e.g. [13,14]). This innate response
appears particularly important in lower vertebrates in
which the acquired immune response may be less well
developed than in mammals [15]. Indeed, lectins of this
type have been found in abundance in mucous of the
skin, gills and intestine, as well as in the blood of sev-
eral species of fish, and some of these have been impli-

cated specifically in the inhibition of bacterial growth
(e.g. [7,16–21]).
The presence of a developed innate defence mech-
anism in the male seahorse brood pouch organ could
be vital as a large number of fry are contained in
an enclosed environment, requiring efficient mecha-
nisms for preventing bacterial colonization. The lec-
tin-mediated pathway could provide a more adept
mechanism than the classical antibody pathway in
allowing specific recognition of microorganisms,
rather than just distinguishing self from nonself
which would leave the embryos, obviously immuno-
logically distinct from the father, also at risk. This
may, however, rely on quite complex recognition of
pathogen-associated molecular patterns on the sur-
face of the microorganisms [14].
Notably, of the group VII fish CTLs reported to
date, nearly all were found to recognize mannose or
galactose, and their respective related sugars [15]. The
Ca
2+
binding sites of the MBP have been well charac-
terized; one of them requires six key residues including
the EPN motif between loops 3 and 4 [10]. Most of
these residues are not present in the novel hcCTLs,
confirming our own findings that the hcCTL III does
not bind mannose, and also previous observations that
these residues have probably been selected as a group
[11]. The proline of the EPN motif is however, extre-
mely well conserved also in CTLs that do not bind

mannose, and a change of the motif to QPD alters the
binding preference to galactose. Although the latter
motif was apparent in the novel hcCTL II, the other
two novel hcCTLs conform to neither. This is in agree-
ment with our finding that hcCTL III-mediated eryth-
rocyte agglutination was not inhibited by galactose or
mannose, or any of the other sugars that we tested,
and suggests that it may be a novel type of CTL pos-
sibly possessing highly specific carbohydrate recogni-
tion. The fact that these hcCTLs in the brood pouch
lack significant similarity with the other CTL reported
in related teleosts further supports the likelihood that
they contain distinct functions, which may be specific
to their role in this distinctive organ. A general lack of
conservation amongst CTLs across species has previ-
ously been noted, as their evolution is thought to have
paralleled that of complex oligosaccharides on cell sur-
face glycoproteins and glycolipids [10].
A number of publications have reported the expres-
sion of high levels of CTLs in fish gonads. These gen-
erally show binding specificity for rhamnose, but also
recognize related sugars containing hydroxyl groups at
C2 and C4 [22–26]. Some of these have been implica-
ted in innate immunity as they inhibit growth of cer-
tain bacteria, however, they also share homology with
the vitellogenic receptor ligand biding domain and
interact with yolk proteins [27]. These are thus thought
to comprise a distinct family of lectins related to the
low density lipoprotein receptor superfamily. They
generally comprise two or three repeats of the CRD

which is comprised of around 20 conserved residues
and is distinct from other characterized CRDs [24,25]
and bears no resemblance to that found in the novel
hcCTLS reported here, which appear more related
functionally and structurally to the skin mucous CTLs.
Interestingly, a study was published recently descri-
bing a cDNA library made from the Chinese seahorse
H. kuda [8], in which a large number of clones enco-
ding CTL proteins was detected. However of the
H. kuda CTL sequences published, only five contain
partial similarity with those found in our study, and
notably more so with the hcCTL II lectin than with the
other two. Given that the library was obtained from
entire animals, it is difficult to conclude much about
the possible expression patterns of these proteins as
localization studies were not carried out, nor were any
functional studies performed; furthermore the repro-
ductive state of the fish used for that study was not
reported. However the authors did attribute a likely
role of the H. kuda CTLs in innate immunity and
P. Melamed et al. C-type lectins in the male seahorse pregnancy
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1231
possibly also some of the beneficial effects associated
with this animal in traditional Chinese medicine.
The secretory profile of the hcCTL III in the pouch
fluid indicates that it is expressed at highest levels after
uptake of the eggs, but that its levels drop off after the
larvae are hatched and as the juvenile seahorses are
preparing to leave the pouch. This suggests the passage
of a signal between the young and the father which

regulates the levels of synthesis and secretion of these
proteins. To date there is little, if any, information on
the ways in which such communication occurs during
this unique pregnancy. Notable however, is that the
drop in hcCTL III levels in the pouch parallels an
increase in osmolality of the pouch fluid, which reaches
that of the seawater just prior to release of the young.
This correlation was also seen in the levels of Japanese
eel CTLs in the gills which are markedly elevated in
fish held in freshwater conditions [7]. That study also
attributed a likely function of the CTLs to a role in
innate immunity, as microorganisms may be more
abundant in freshwater than in seawater conditions,
and it was suggested their presence in the gill mucous
forms a protective layer between the water and the epi-
thelium. Although the Japanese eel CTLs appear to be
galactose specific, it is possible that their regulation
shares some common elements with the hcCTLs, per-
haps as a direct result of the increased salinity.
Our study has thus revealed a novel CTL that is
produced and secreted in significant quantities into the
male H. comes brood pouch in a regulated manner
during specific stages of pregnancy. Preliminary func-
tional studies indicate that this CTL causes cell agglu-
tination and may act to help repress bacterial growth,
but that this is not via the common lectin-binding
sugars. Further studies will be required to elucidate its
precise mechanisms of action and the ways in which its
expression is regulated.
Experimental procedures

Construction of cDNA library and identification
of clones
RNA was extracted from the pouches of five male seahorses
(Hippocampus comes: purchased from Pacific Marine Aqua-
ria, Singapore, after import from Indonesia) at various stages
of pregnancy. After removal of all embryos, the inner tissue
lining the pouch was pulled away from the muscle wall, and
RNA was extracted using TRIzol (Invitrogen, Carlsbad,
CA). The cDNA library was constructed using MMLV-
reverse transcriptase (Stratagene, La Jolla, CA), 2.8 lg linker
primer and 6 lg of the poly(A)
+
RNA at 37 °C for 1 h. After
synthesis using DNA polymerase I, the second-strand cDNA
was created using Pfu DNA polymerase. The purified blunted
cDNA was ligated with EcoRI adapters and the EcoRI ends
phosphorylated with T4 polynucleotide kinase. The product
was digested with XhoI. The DNA was precipitated and re-
suspended in STE buffer before passing through a sepharose
CL-2B gel filtration column. Twenty 100-lL fractions were
collected, from each of which 10 lL was electrophoresed on
a 5% nondenaturing acrylamide gel and silver stained: the
fractions containing larger cDNAs were combined and puri-
fied before re-suspension in 5 lL sterile water.
The cDNA was ligated into the Uni-ZAP XR vector
(Stratagene) using T4 DNA ligase at 4 °C for 2 days and
then packaged into Gigapack III Gold Packaging Extract
(Stratagene) by incubating at room temperature for 2 h,
before addition of SM buffer (500 lL) and chloroform
(20 lL). After centrifugation, the phage were stored at 4 °C

while serial dilutions of primary cDNA library were incuba-
ted with bacteria at 37 °C for 15 min before plating of the
mixtures onto Luria–Bertani (LB) top agar plates.
After 10
7
pfu of the phage, 10
8
XL1-Blue MRF¢ cells and
10
9
pfu of ExAssist helper phage were incubated at 37 °C for
15 min, 20 mL of LB broth was added and the mixture incu-
bated for 2.5 h at 37 °C with shaking. The mixture was hea-
ted at 65 °C for 20 min, spun (1000 g for 10 min) and 1 lLof
the supernatant was added to 200 lL SOLR cells. After
15 min at 37 °C, 100 lL of the cell mixture was plated onto
LB–ampicillin agar plates. Plasmid DNA was isolated from
individual colonies and inserts larger than  500 bp (con-
firmed by colony PCR using T3 and T7 primers, and selected
in order to provide a reasonable chance for accurate identifi-
cation) were sequenced from the plasmid with T3 or T7 prim-
ers using the ABI prism Dye Terminator Cycle Sequencing
Ready Reaction Kit and an ABI PRISM 3100 DNA Se-
quencer (Perkin Elmer Applied Biosystems, Foster City, CA).
Nucleotide sequences were used to search the GenBank
using blastn, and they were also checked in all reading
frames against PDB and ⁄ or SwissProt protein databases
using blastx and blastp (all at NCBI: http://www.
ncbi.nlm.nih.gov/blast/). Insert sequences were also transla-
ted into possible peptides, using specific reading frames

suggested by the results of blastx. Lastly, the translated
peptide sequences were used for pattern searching at Scan-
Prosit ( for identifi-
cation of conserved protein domains.
In situ hybridization and histology
Probes were prepared using a 300-bp PCR amplified frag-
ment from the hcCTL III protein cDNA which was
labelled with digoxigenin-dUTP, using the DIG High
Prime DNA Labeling and Detection Starter Kit I (Roche
Diagnostics, Basel, Switzerland).
The pouch tissues were fixed in formalin and embedded
in paraffin before sectioning at 6–8 lm. The paraffin was
C-type lectins in the male seahorse pregnancy P. Melamed et al.
1232 FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS
removed and they were rehydrated before additional fix-
ing in 0.4% (v ⁄ v) paraformaldehyde in NaCl ⁄ P
i
for
20 min, after which they were washed in NaCl ⁄ P
i
before
treatment with 0.3% Triton-X for 15 min. The sections
were permeablized with 20 lgÆmL
)1
proteinase K in TE
buffer at 37 °C for 30 min, and then acetylated with
0.5% (v ⁄ v) acetic anhydride in 100 mm Tris ⁄ HCl pH 8.0
for 5 min before washing with NaCl ⁄ P
i
.

The sections were incubated with 300 lL prehybridization
solution [formamide (25%), sodium citrate (NaCl ⁄ Cit · 4),
Denhardt’s solution, calf thymus DNA (0.5 mgÆmL
)1
), yeast
tRNA (0.25 mgÆmL
)1
), dextran sulfate (10%) and dithio-
threitol (0.01 m)] for 1 h at 37 ° C, followed by washing in
2· NaCl ⁄ Cit. Probes were denatured for 5 min at 95 °C and
immediately cooled on ice, and diluted in the prehybridiza-
tion solution to 200 ngÆmL
)1
. Hybridization was carried out
for 16 h at 42 °C.
The sections were washed in 2· NaCl ⁄ Cit buffer (1 h at
room temperature), 1· NaCl ⁄ Cit buffer (1 h at room tem-
perature) and then 0.5· NaCl ⁄ Cit buffer (30 min at 37 °C,
twice), and subsequently with Buffer 1 (100 mm Tris ⁄ HCl,
150 mm NaCl, pH 7.5) before incubation in fresh Blocking
solution (Buffer 1 containing 2% normal sheep serum,
0.3% Tritron X-100) for 30 min. The Anti-DIG-alkaline
phosphatase (AP) conjugate was diluted · 5000 with Buffer
1 and applied to the sections for 4 h. After washing in Buf-
fer 2 (100 mm Tris ⁄ HCl, 100 mm NaCl, 60 mm MgCl
2
,
pH 9.5), colour solution (500 lL 200 lm Nitro Blue tetra-
zolium 5-bromo-4-chloroindol-2-yl-phosphate (NBT ⁄ BCIP)
stock solution in 10 mL Detection Buffer) was applied and

incubated in the dark until a desired intensity was acquired.
The colour reaction was stopped using TE buffer and the
sections were rinsed with water and air dried before mount-
ing using DPX resinous mount.
Histological staining was carried out on sections from
the same serials as above, after rehydration and washing
with running water. They were stained using Mayer’s haem-
atoxylin and eosin. Sections were dehydrated and cleared in
xylene before mounting using DPX resinous mount.
Proteomic analysis of incubation fluid
The pouch fluid was obtained by making an incision
along the abdominal fissure of the pouch and draining
into a syringe. Aliquots of 50 lL were treated with 4 mm
tributyl phosphine (for 1 h) followed by 15 mm iodoacet-
amide (for 1.5 h in the dark) to reduce and alkylate the
proteins. The fluid was diluted 10-fold with rehydration
buffer [6 m urea, 2 m thiourea, 4% CHAPS, 0.2% Bio-
loytes 3-10 (Bio-Rad), and Bromophenol blue], centri-
fuged for 20 min before the supernatant was subjected to
ultrafiltration twice with Microcon YM-3 at 12 400 g.
The final retentate was reconstituted to 300 lL with rehy-
dration buffer for in-gel rehydration of immobilized pH
gradient (IPG) strip.
The 17-cm, pH 3–10 IPG strips (Bio-Rad) were rehyd-
rated actively for 12 h. Isoelectric focusing was performed
(Protean IEF Cell, Bio-Rad): 250 V (1 h), 500 V (for 1 h),
1000 V (for 1 h, 10 000 V (gradient over 3 h), and
10 000 V (for 90 000 Vh) at 20 °C. The IPG strips were
equilibrated for 20 min in equilibration buffer (6 m urea,
2% SDS, 50 mm Tris ⁄ HCl pH 8.8, 30% glycerol) before

placing them onto 12% SDS ⁄ PAGE gels and sealing with
molten low-melting agarose (1% in electrode running buf-
fer). Electrophoresis was at 10 mA constant per gel for 1 h
followed by 24 mA constant for about 5 h, at room tem-
perature. The protein spots were visualized by silver-stain-
ing according to Blumb et al. [27].
Protein spots were excised from the gel and minced, before
washing and dehydrating three times with 50% acetonitrile
containing 50 mm ammonium bicarbonate, and acento-
nitrile, before drying. The gel pieces were re-swollen in
15–50 lL of digestion solution [12.5 ngÆlL
)1
trypsin (Prome-
ga, Sequencing grade modified trypsin) in 50 mm ammonium
bicarbonate] and incubated at 4 °C for 30 min. The excess
trypsin solution was removed and 15–50 lLof50mm
ammonium bicarbonate was added, for 15 h at 37 °C.
The supernatants were collected and the gel pieces were
treated with 20 mm ammonium bicarbonate and the super-
natant saved. The gel pieces were then treated with 15–
50 lL of 5% formic acid in 50% acetonitrile for 10 min
and centrifuged at 3800 g. The extracts were saved and the
extraction repeated twice. Lastly, all three supernatants
were combined and dried.
The digests were redissolved in 5 lL of 0.5% formic acid
in 50% acetonitrile and 1 lL of the peptide solution was
applied onto the MALDI plate with a solution of a-cyano-4-
hydroxycinnamic acid (10 mgÆmL
)1
in 50% acetonitri-

le ⁄ 0.3% trifluoroacetic acid) and left to air dry. MS was car-
ried out using a Voyager Biospectrometry Workstation STR
with Delayed Extraction Technology MALDI-TOF mass
spectrometer (PerSeptive Biosystems, Framingham, MA), at
The Protein and Proteomic Centre, Department Biological
Sciences, National University of Singapore. The MALDI
spectra were acquired in the delayed extraction, reflector
mode. The mass scale was calibrated internally with the tryp-
sin autodigestion products of known amino acid sequence,
[M + H]
+
¼ 842.509 and 2211.104. Typically, 180 scans
were averaged to produce the final spectrum.
Expression of recombinant lectin type III
and raising of antisera
The coding sequence of hcCTL III, without the signal
peptide, was amplified by PCR to incorporate BamHI and
XhoI sites at either end, using the following primers:
forward, 5¢-CGCGGATCCTGGTCTTTCCAAAATATTC
AGGCCA-3¢ and reverse, 5¢-GTCCTCGAGGTACATCA
CATCTCTGAT-3¢. After digestion, the PCR-amplified
fragment was cloned into a modified BamHI ⁄ XhoI digested
P. Melamed et al. C-type lectins in the male seahorse pregnancy
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1233
pET-32a(+) plasmid (Novagen, San Diego, CA) to produce
a His
6
-tagged protein. The construct was verified by sequen-
cing.
BL21(DE3)pLysS Singles E. coli competent cells (Nov-

agen) were transformed by heat-shock and a single colony
was inoculated in 10 mL LB–ampicillin medium at 37 °C,
250 r.p.m., for 16 h. Subsequently, 3 mL of the culture was
used to inoculate 1500 mL medium, and incubated until the
D
600
was 0.4–0.6. Protein expression was induced by IPTG
(0.1 mm) and cells cultivated for 5 h (37 °C, 250 r.p.m),
before harvesting (3840 g, 15 min) and lysis in 40 mL lysis
buffer (50 mm NaH
2
PO
4
, 300 mm NaCl, 10 mm imidazole
pH 8.0, 1 mgÆmL
)1
lysozyme) with sonication (20 · 10-s
bursts at 200 W, with a 10-s cooling intervals). The lysate
was cleared by centrifugation (33 700 g, 40 min at 4 °C)
and the supernatant collected.
The lectin was purified by addition of 2 mL Ni–NTA
agarose (Qiagen, Valencia, CA) overnight at 4 °C before
loading onto a column and washing with modified wash
buffer (50 mm NaH
2
PO
4
, 300 mm NaCl, 50 mm imidazole
pH 8.0). The protein was eluted in fractions with modified
elution buffer (50 m m NaH

2
PO
4
,2m NaCl, 250 mm imi-
dazole pH 8.0). The purity was checked by SDS ⁄ PAGE,
and the protein content determined using a Bradford assay
(Bio-Rad, Hercules, CA).
For raising of antisera, proteins were dialysed against
NaCl ⁄ P
i
overnight and adjusted to 500 lL (0.6 mgÆmL
)1
).
An equivalent volume of Freund’s adjuvant was added and
mixed before injection into rabbits four times at 3-week
intervals, during which blood was extracted and the anti-
body titres monitored.
Western analysis
Brood pouch proteins were analysed by separation by
SDS ⁄ PAGE (12% gels) at 200 V for 45 min before transfer
to a PVDF membrane at 30 V for 4 h at 4 °C. The
membrane was blocked with 10% milk in wash buffer (TBS
with 0.1% Tween 20) for 3 h at room temperature
( 22 °C), and then incubated with hcCTL III antisera at
1 : 1000 dilution with 1% milk in wash buffer for 1 h at
room temperature. After washing three times, the mem-
brane was incubated with goat antirabbit IgG–horseradish
peroxidase (HRP) conjugated secondary antibody (Santa
Cruz Biotechnology, Santa Cruz, CA) diluted 1 : 1000 with
1% milk in wash buffer for 1 h at room temperature. After

washing, the immunoreactive protein was detected using
the Super Signal Pico West chemiluminescent system
(Pierce Chemical Co., Rockford, IL), followed by exposure
to Cl-XPosure film (Pierce Chemical Co.) for 1–5 min.
Bacteriostatic assays
Bacteriostatic assays were carried out essentially as in
Biswas et al. [28]. LB media (1 mL) was inoculated with
DH5a E. coli and incubated at 37 °C, at 250 r.p.m. for
3–4 h, until D
595
reached 0.1. The recombinant hcCTL III
protein dissolved in NaCl ⁄ P
i
(20 lL) was added to half of
the tubes to a final concentration of 0.7 lm, while the vehi-
cle alone was added to the controls. The OD reading was
subsequently taken every 30 min in four samples for each
group at each time point.
Haemagglutination assays
Heparinized mouse blood was washed three times in
NaCl ⁄ P
i
(at 420 g, 10 min). The mouse erythrocytes [50 lL,
2% (v ⁄ v)] in NaCl ⁄ P
i
were added to the recombinant
hcCTL III (50 lL to final concentration of 0.5–18 lm) and
incubated at 25 °C for 60 min in a V-shaped 96-well plate.
For inhibition assays, the same volume of erythrocytes was
added to the hcCTL III protein (25 lL to final concentra-

tion of 4.5 lm) and of carbohydrate (25 lL) and incubated
at 25 °C for 60 min.
Acknowledgements
This research was supported by The Academic
Research Fund, National University of Singapore.
References
1 Wetzel J & Wourms JP (1991) Paternal-embryonic rela-
tionships in pipefishes and seahorse (Syngnathidae). Am
Zool 31, A83–A83.
2 Azzarello MY (1991) Some questions concerning the
syngnathidae brood pouch. B Mar Sci 49 , 741–747.
3 Carcupino M, Baldacci A, Mazzini M & Franzoi P
(1997) Morphological organization of the male brood
pouch epithelium of Syngnathus abaster Risso (Teleos-
tea, Syngnathidae) before, during, and after egg incuba-
tion. Tissue Cell 29, 21–30.
4 Watanabe S, Kaneke T & Watanabe Y (1999) Immuno-
logical detection of mitochondria-rich cells in the brood
pouch epithelium of the pipefish, Syngnathus schlegeli:
structural comparison with mitochondria-rich cells in
the gills and larval epidermis. Cell Tissue Res 195,
141–149.
5 Weng CF, Lee TH & Hwang PP (1997) Immune locali-
zation of prolactin receptor in the mitochondria-rich
cells of the euryhaline teleost (Oreochromis mossambi-
cus) gill. FEBS Letts 405, 91–94.
6 Hung CH, Huang HR, Huang CJ, Huang FL & Chang
GD (1997) Purification and cloning of carp nephrosin, a
secreted zinc endopeptidase of the astacin family. J Biol
Chem 272, 13772–13778.

7 Mistry AC, Honda S & Hirose S (2001) Structure, prop-
erties and enhanced expression of galactose-binding
C-type lectins in mucous cells of gills from freshwater
C-type lectins in the male seahorse pregnancy P. Melamed et al.
1234 FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS
Japanese eels (Anguilla japonica). Biochem J 360,
107–115.
8 Zhang N, Xu B, Mou CY, Yang WL, Wei JW, Lu L,
Zhu JJ, Du JC, Wu XK, Ye LT et al. (2003) Molecular
profile of the unique species of traditional Chinese
medicine, Chinese seahorse (Hippocampus kuda Bleeker).
FEBS Letts 550, 124–134.
9 Weis WI, Kahn R, Fourme R, Drickamer K &
Hendrickson WA (1991) Structure of the calcium-
dependent lectin domain from a rat mannose-binding
protein determined by MAD phasing. Science 254,
1608–1615.
10 Dodd RB & Drickamer K (2001) Lectin-like proteins in
model organisms: implications for evolution of carbohy-
drate-binding activity. Glycobiology 11, 71R–79R.
11 Drickamer K & Dodd RB (1999) C-type lectin domains
in Caenorhabditis elegans: predictions from the complete
genome sequence. Glycobiology 9, 1357–1369.
12 Drickamer K (1993) Ca
2+
-dependent carbohydrate-
recognition domains in animal proteins. Curr Opin
Struct Biol 3, 393–400.
13 Weis WI, Taylor ME & Drickamer K (1998) The
C-type lectin superfamily in the immune system.

Immunol Rev 163, 19–34.
14 Fujita T (2002) Evolution of the lectin-complement path-
way and its role in innate immunity. Nat Rev Immunol
2, 346–353.
15 Ewart KV, Johnson SC & Ross NW (2001) Lectins of
the innate immune system and their relevance to fish
health. ICES J Mar Sci 58, 380–385.
16 Tasumi S, Ohira T, Kawazoe I, Suetake H, Suzuki Y
& Aida K (2002) Primary structure and characteristics
of a lectin from skin mucus of the Japanese eel Anguilla
japonica. J Biol Chem 277, 27305–27311.
17 Mitra S & Das HR (2001) A novel mannose binding
lectin from the plasma of Labeo rohita. Fish Physiol
Biochem 25, 121–129.
18 Savan R, Endo M & Sakai M (2004) Characterization
of a new C-type lectin from common carp Cyprinus car-
pio. Mol Immunol 41, 891–899.
19 Suzuki Y, Tasumi S, Tsutsui S, Okamoto M & Suetake
H (2003) Molecular diversity of skin mucus lectins in
fish. Comp Biochem Physiol B Biochem Mol Biol 136,
723–730.
20 Fujiki K, Bayne CJ, Shin DH, Nakao M & Yano T
(2001) Molecular cloning of carp (Cyprinus carpio) C-type
lectin and pentraxin by use of suppression subtractive
hybridization. Fish Shellfish Immunol 11, 275–279.
21 Stratton L, Wu S, Richards RC & Ewart KV (2004)
Oligomerisation and carbohydrate binding in an Atlan-
tic salmon serum C-type lectin consistent with non-self
recognition. Fish Shellfish Immunol 17, 315–323.
22 Shiina N, Tateno H, Ogawa T, Muramoto K, Saneyoshi

M & Kamiya H (2002) Isolation and characterization of
1-rhamnose-binding lectins from chum salmon (Oncor-
hynchus keta) eggs. Fish Sci 68, 1352–1366.
23 Lam YW & Ng TB (2002) Purification and characteriza-
tion of a rhamnose-binding lectin with immunoenhan-
cing activity from grass carp (Ctenopharyngodon idellus)
ovaries Prot Expr Purif 26, 378–385.
24 Tateno H, Saneyoshi A, Ogawa T, Muramoto K,
Kamiya H & Saneyoshi M (1998) Isolation and charac-
terization of rhamnose-binding lectins from eggs of
steelhead trout (Oncorhynchus mykiss) homologous to
low density lipoprotein receptor superfamily. J Biol
Chem 273, 19190–19197.
25 Tateno H, Ogawa T, Muramoto K, Kamiya H &
Saneyoshi M (2002) Distribution and molecular
evolution of rhamnose-binding lectins in Salmonidae:
Isolation and characterization of two lectins from white-
spotted charr (Salvelinus leucomaenis) eggs. Biosci Bio-
technol Biochem 66, 1356–1365.
26 Hosono M, Ishikawa K, Mineki R, Murayama K,
Numata C, Ogawa Y, Takayanagi Y & Nitta K (1999)
Tandem repeat structure of rhamnose-binding lectin
from catfish (Silurus asotus) eggs. Biochem Biophys Acta
1472, 668–675.
27 Blumb H, Beier H & Gross HJ (1987) Improved silver
staining of plant proteins, RNA and DNA in polyacryl-
amide gels. Electrophoresis 8, 93–99.
28 Biswas C, Sinha D & Mandal C (2000) Investigation of
Achatinin, a 9-O-acetyl sialic acid-binding lectin, with
lipopolysaccharide in the innate immunity of Achatina

fulica snails. Mol Immunol 37, 745–754.
P. Melamed et al. C-type lectins in the male seahorse pregnancy
FEBS Journal 272 (2005) 1221–1235 ª 2005 FEBS 1235