• biological adhesive;
  • extracellular protein;
  • holdfast protein;
  • protein adsorption;
  • sessile organism


  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

Barnacle attachment to various foreign materials in water is guided by an extracellular multiprotein complex. A 19 kDa cement protein was purified from the Megabalanus rosa cement, and its cDNA was cloned and sequenced. The gene was expressed only in the basal portion of the animal, where the histologically identified cement gland is located. The sequence of the protein showed no homology to other known proteins in the databases, indicating that it is a novel protein. Agreement between the molecular mass determined by MS and the molecular weight estimated from the cDNA indicated that the protein bears no post-translational modifications. The bacterial recombinant was prepared in soluble form under physiologic conditions, and was demonstrated to have underwater irreversible adsorption activity to a variety of surface materials, including positively charged, negatively charged and hydrophobic ones. Thus, the function of the protein was suggested to be coupling to foreign material surfaces during underwater attachment. Homologous genes were isolated from Balanus albicostatus and B. improvisus, and their amino acid compositions showed strong resemblance to that of M. rosa, with six amino acids, Ser, Thr, Ala, Gly, Val and Lys, comprising 66–70% of the total, suggesting that such a biased amino acid composition may be important for the function of this protein.


artificial seawater


Balanus albicostatus 19 kDa cement protein


Balanus improvisus 19 kDa cement protein


cement protein



GSF1 and GSF2

cement fractions separated by their solubility in a guanidine hydrochloride solution


Megabalanus rosa cement protein


recombinant 19 ka Megabalanus rosa cement protein in Escherichia coli


response unit


surface plasmon resonance.

Living on a boundary brings various advantages for organisms; such organisms therefore have developed a variety of molecular systems to hold themselves on the boundary during their evolution. Marine sessile organisms possess underwater attachment capability as an indispensable physiologic function, enabling them to live on a liquid–solid boundary during most of their life cycle. This underwater attachment is closely related to other biological functions such as metamorphosis, molting and biomineralization. Recent advances in underwater holdfast studies on mussel [1–3], and barnacle [4], which represent two typical organisms possessing this kind of activity, have indicated that the biological adhesion is, in general, mediated by an insoluble multiprotein complex. Each constitutive protein of the complex has been suggested to have a special function in a multifunctional process of underwater attachment. These functions [5] include displacement of the bound-water layer on a foreign substratum by the adhesive, as well as spreading, coupling of the adhesive with a variety of material surfaces, self-assembly of the adhesive, curing to make the holdfast stiff and tough, and protection from microbial degradation. This multifunctionality, together with the insoluble/sticky and complex nature of the adhesive, have hindered any detailed analysis of its function, especially the direct evaluation of the adhesive process. Thus, biological underwater attachment remains an unachievable technology, which is considered to be based on a completely different approach from that used in developing artificial adhesives in air.

The barnacle, a unique sessile crustacean, has long been noted for its underwater adhesive capability [6–8]. This underwater adhesive material, called cement, joins two different materials, the animal's own calcareous base and the foreign substratum, together in water as a molecular event. Development of a method to render this barnacle cement soluble [9] has enabled us to identify its components. Four cement proteins, designated as Megabalanus rosa cement protein (Mrcp)-100k [9], Mrcp-52k, Mrcp-68k [10], and Mrcp-20k [11], have so far been identified; these were shown to be novel proteins that are distinct from each other. The cp-100k and cp-52k proteins are characterized by their insoluble nature and remarkable hydrophobicity, and are possibly bulk proteins of the cement complex. Reduction treatment with guanidine hydrochloride solution was indispensable to render these proteins soluble. cp-68k is characterized by its bias toward four-amino acids, Ser, Thr, Ala and Gly which comprise 57% of the total residues. cp-20k is characterized by its abundant charged amino acids, with its primary structure being a repeat of a well-defined segment in which Cys residues are found in designated positions. Although both cp-100k and cp-52k seem to constitute the bulk of the adhesive, no proteins contributing the necessary surface functions such as priming, spreading and coupling have been identified. Nor have any direct measurement of these activities been reported for the cement proteins prepared under physiologic conditions, and such kinds of measurement have never been achieved in any biotic underwater adhesive protein studies.

The holdfast system of the barnacle shows no similarity to that of the mussel, a relatively well-characterized one. There are no sequence similarities among the protein components between the two systems. The mussel holdfast system [1] depends on several protein modifications, including 3,4-dihydroxyphenylalanine (Dopa); however, no involvement of Dopa in the barnacle cement was found [10,12]. Thus, the barnacle system represents a novel biological adhesive system.

The present study identified a novel cement protein, cp-19k, in the barnacle holdfast system, and demonstrated its ability to be adsorbed to a foreign material surface in seawater using a bacterial recombinant protein prepared under physiologic conditions. We also show that the function of the protein is reliant upon common amino acids, with no specific modifications.


  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

Purification and characterization of Mrcp-19k

Mrcp-19k was detected by SDS/PAGE in both guanidine hydrochloride-soluble fractions 1 (GSF1) and 2 (GSF2) of barnacle cement [9] with the same mobility (Fig. 1). The molecular mass was estimated to be 18 500 Da from SDS/PAGE. Mrcp-19k was purified from GSF1 by column chromatography, which gave rise to a molecular mass of 16 992.34 Da as measured by MALDI-TOF MS (Table 1). The protein (Table 2) was rich in Gly (17.3%), Thr (12.3%), Ser (11.3%), Ala (10.6%), Lys (8.5%), and Val (8.7%). The amino acid sequence of the mature N-terminus was determined as VPPPXDLGIASKVKQKGVTGGGASVSTT, where X was most likely to be Cys. The N-terminal sequences of three internal peptide fragments were determined to be GVTGGGASVSTTSATQGSG, GFSEGTAAISQTAGANGGATV, and GTVTSSSSHQGSGAGDSIFE. Specific staining for detection of either glycosylation or phosphorylation gave negative results in both cases.


Figure 1. Mrcp-19k from M. rosa cement and its bacterial recombinant analyzed by SDS/PAGE. Lanes 1 and 2: GSF1 and GSF2 prepared from M. rosa cement, respectively. Lane 3: the bacterial recombinant protein rMrcp-19k. Lane 4: molecular weight markers. The samples were separated by SDS/PAGE (a Tris/Tricine buffer system, 16.5% T/3% C [28]) and stained with Coomassie Brilliant Blue R-250. The numbers on the right-hand side indicate molecular masses (kDa). The arrow indicates Mrcp-19k. The bacterial recombinant, rMrcp-19k, has an additional dipeptide, Met-Ala, at the N-terminus of mature Mrcp-19k, due to the vector construction.

Download figure to PowerPoint

Table 1.   Predicted and observed molecular masses and predicted isoelectric points of cp-19ks. Calculated mass (MasscDNA) is based on a sequence deduced from the cDNA, and m/zobs value corresponds to [M + H]+ observed with MALDI TOF-MS.
  Mrcp-19k Balcp-19k Bicp-19k
m/zobs16 993.34
MasscDNA16 995.5217 336.2716 841.99
Table 2.   Amino acid compositions of various cp-19ks and their deviations from standard compositions. The amino acid compositions of mature cp-19ks are presented as the number of residues per protein in columns 1–4. The ratios of each number of residues to the average contents of the amino acids [13] are shown to indicate the bias in columns 5–7. ND, not determined.
  Mrcp19ka Mrcp19kb Balcp19ka Bicp19ka Mrcp19k/standard Balcp19k/standard Bicp19k/ standard
  • a

     The amino acid composition calculated from the deduced sequence.

  • b

     The amino acid composition analyzed by amino acid analysis.

  • c

     Sum of the numbers of Asp and Asn.

  • d

     Sum of the numbers of Glu and Gln.


Cloning of cp-19k cDNA from M. rosa, Balanus albicostatus and B. improvisus

A 53 bp DNA corresponding to the N-terminal part was first amplified from M. rosa cDNA by PCR. The deduced amino acid sequence of the 53 bp DNA completely matched the N-terminal amino acid sequence of the mature Mrcp-19k. Subsequent 3′-RACE and 5′-RACE resulted in a 750 bp and a 102 bp DNA fragment, respectively. An 852 bp cDNA fragment encoding the Mrcp-19k protein was finally determined. Ten randomly selected clones for the coding region of Mrcp-19k had one nonsynonymous substitution and several synonymous substitutions, presumably due to errors introduced by the PCR amplifications (as each substitution was found only in one randomly selected clone but not in any other clones). Both B. albicostatus (Bal)cp-19k (881 bp) and B. improvisus (Bi)cp-19k (970 bp) cDNAs were also amplified by 3′-RACE with the oligonucleotide primers designed from the N-terminal region of Mrcp-19k.

Structural outline of cp-19ks

The coding region of Mrcp-19k encoded 198 amino acids (supplementary Fig. S1A). The mature N-terminal sequence was found to start at residue number 26; thus the first 25 amino acids function as the signal peptide that has been cleaved off in the mature protein. The amino acid sequences of the N-terminal and three internal peptide fragments of Mrcp-19k determined experimentally were found to be contained in the deduced sequence and are in complete agreement with those of the deduced sequence. The cDNA fragments of 881 bp and 970 bp encoding 173 amino acids each were also determined for Balcp-19k and Bicp-19k, respectively (supplementary Fig. S1A). The molecular masses and isoelectric points of the mature polypeptides were predicted to be 16 995.52 Da (Table 1) and 5.8 for Mrcp-19k, 17 336.27 Da and 10.3 for Balcp-19k, and 16 841.99 Da and 10.3 for Bicp-19k, respectively. The molecular mass of Mrcp-19k estimated by SDS/PAGE was slightly higher than that predicted from the cDNA sequence and that determined by MALDI-TOF MS. This may be due to unusual migration on SDS/PAGE caused by the biased amino acid composition. The amino acid composition of Mrcp-19k deduced from the cDNA (Table 2) agreed well with that of Mrcp-19k determined by the amino acid analysis, with six amino acids, Gly (15.6%), Thr (12.1%), Ser (10.4%), Ala (10.4%), Lys (9.8%) and Val (8.1%), as dominant residues and representing 66.4% of all residues. This ratio is significantly higher than that deduced from the standard amino acid composition [13]. The sequence identity and similarity (Fig. 2) were as follows: Mrcp-19k versus Balcp-19k, 54% identity and 65% similarity; Mrcp-19k versus Bicp-19k, 51% identity and 68% similarity; and Balcp-19k versus Bicp-19k, 61% identity and 75% similarity. All cp-19ks contained two Cys residues, whose positions are conserved. The amino acid compositions among three cp-19ks agreed well with each other, especially in terms of the content of the six dominant residues, Gly, Thr, Ser, Ala, Lys and Val (Table 2).


Figure 2.  Alignment of the amino acid sequences of mature cp-19ks.The deduced amino acid sequences of mature Mrcp-19k, Balcp-19k and Bicp-19k were aligned by clustalw[34]. The three homologous proteins have the same amino acid length, and the two Cys residues are conserved. Identical amino acids are reversed.

Download figure to PowerPoint

A blast search of the nonredundant database and a sequence profile-based fold-recognition method for three-dimensional structural prediction failed to provide any homologous sequences and meaningful structure (supplementary Document S1). In particular, no sequence similarity between cp-19ks and foot proteins in the mussel was evident. The primary structures of cp-19ks also showed no homology with cp-100k and cp-20k. Naldrett & Kaplan [14] have reported the partial amino acid sequences of peptide fragments from B. eburneus cement. Among these fragments, WCD-21, a peptide fragment obtained by cyanogen bromide treatment of B. eburneus cement, showed homology to the N-terminal region of cp-19ks (supplementary Fig. S1B), indicating that the protein homologous to cp-19k should also be present in B. eburneus cement.

Characterization of the recombinant Mrcp-19k protein

Recombinant (r)Mrcp-19k was expressed in Escherichia coli as a soluble cytosolic fraction, and was purified to homogeneity (Fig. 1). rMrcp-19k had a slightly lower mobility than that of the native Mrcp-19k isolated from the cement. This was due to the additional N-terminal dipeptide in the recombinant protein as the result of the vector design. The N-terminal sequence and molecular mass were determined to be AMVPPPXDLG and 17 201 Da (predicted molecular mass from the cDNA, 17 197.60 Da), respectively. Digestion of rMrcp-19k by a specific protease generated a peptide fragment with a molecular mass of 4509.24 Da, which corresponds to two peptides; each contains one Cys residue, and they are linked by a disulfide bond (Ala1-Lys14 and Gly19-Lys51, predicted molecular mass, 4509.17 Da). Treatment with reductants led to the loss of the MS peak, and alternatively gave two MS peaks corresponding to each single peptide with molecular masses of 3112.36 Da (Gly19-Lys51, predicted molecular mass, 3111.47 Da) and 1398.6 Da (Ala1-Lys14, predicted molecular mass, 1397.70 Da). This confirmed that the two Cys residues in rMrcp-19k form an intramolecular disulfide bond.

The properties of adsorption of rMrcp-19k to underwater surfaces of glass, formaldehyde resin, alkylated gold, and bare gold were measured either in artificial seawater (ASW) or in a dilute buffer solution. Figure 3 shows the mass uptake by the adsorption of rMrcp-19k on the gold and alkylated gold surfaces versus time from the surface plasmon resonance (SPR) measurement. The proteins showed rapid adsorption to the sensor surfaces that corresponded to sharp increases in the SPR shift. Upon washing, the response units (RUs) were slightly decreased, probably due to dissociation of loosely attached protein. The final RUs after washing were almost the same after repetitive injections of the protein on each surface. The adsorption kinetics were estimated by nonlinear curve fitting with theoretical models described in the biaevaluation software (supplementary Fig. S2). The adsorption constant ka and desorption constant kd were calculated as 2.17 × 105 m−1·s−1 and 4.94 × 10−4 s−1, respectively, for the formation of the rMrcp-19k–Au complex. Using these data, the equilibrium constant Keq = ka/kd could be estimated as 4.39 × 108 m−1. rMrcp-19k was similarly adsorbed to the hydrophobic alkylated gold surface, although the adsorbed amount was two-thirds of that adsorbed to bare gold (Table 3). The values of ka, kd and Keq were calculated to be 9.76 × 104 m−1·s−1, 6.67 × 10−4 s−1 and 1.46 × 108 m−1, respectively. The amounts adsorbed to the glass and formaldehyde resin surfaces in 5 min at 25 °C were estimated, and the results are shown in Table 3 and supplementary Fig. S3.


Figure 3.  Typical SPR analyses on polycrystalline gold and alkylated gold. The arrows and thick arrows indicate the starts of sample loading (2 µm) and washing by the running buffer, respectively. The processes of sample loading and washing were sequentially repeated three times. Open circular symbols, squares and triangles indicate changes of resonance after protein adsorption on polycrystalline gold in ASW, on the same material in a dilute buffer containing 10 mm Tris (pH 7.4)/25 mm NaCl, and on alkylated gold (HPA) in ASW, respectively. ΔRUs after each washing process were as follows: first loading on Au in ASW, 1174 RU; second loading on Au in ASW, 1177 RU; third loading on Au in ASW, 1182 RU; first loading on Au in dilute buffer, 1278 RU; second loading on Au in dilute buffer, 1318 RU; third loading on Au in dilute buffer, 1345 RU; first loading on alkylated gold in ASW, 768 RU; second loading on alkylated gold in ASW, 827 RU; third loading on alkylated gold in ASW, 858 RU.

Download figure to PowerPoint

Table 3.   Amount of adsorption of rMrcp-19k to several material surfaces. The adsorbed amount in ASW or dilute buffer solution was calculated from the change in RU on SPR [36] for gold and alkylated gold, and from a quantitative amino acid analysis for glass and the formaldehyde resin (see details in supplementary Fig. S3). Surface area per molecule was calculated by a assuming full surface monolayer coverage.
 GoldAlkylated goldGlassFormaldehyde resin
Adsorption amount (ng·mm−2)0.76 (0.83 in dilute buffer)0.52.484.38
Surface area per molecule (nm2 per molecule)37 (35 in dilute buffer)57117

Localization and expression site of Mrcp-19k

M. rosa cement was usually collected by gently scraping the surface of the calcareous base on the side attached to the foreign material surface [10], making the cement proteins vulnerable to contamination by calcified material during the process of collection. We therefore attempted to confirm the identified protein as a cement component. The cement joins the animal's own calcareous base to the foreign substratum. Therefore, the cement should be present on one side of the barnacle's calcareous base, whereas the peripheral shell should be free from cement. If the protein is present in the protein fraction from the calcareous base and not in that from the peripheral shell, this would confirm that the protein is a cement component and not a component involved in calcification. Western blot analysis of the primary cement, and of the protein fractions in the calcareous base and periphery, indicated that Mrcp-19k was present in the primary cement and protein fraction in the calcareous base, but not in the peripheral shell (Fig. 4). A western blot analysis with the polyclonal antibody raised against Mrcp-100k gave a similar result to that for Mrcp-19k.


Figure 4.  Western blotting analysis to identify the location of Mrcp-19k and Mrcp-100k in the cement. (A) Antibody to Mrcp-19k was used for western blotting analysis. Lane 4 shows the primary cement [10] with the dithiothreitol/guanidine hydrochloride treatment [9]. Lanes 5 and 6 show the barnacle peripheral shell and base plate, respectively, which have been decalcified and rendered soluble by the dithiothreitol/guanidine hydrochloride treatment. Lanes 1–3 correspond to GSF1, GSF2 and the recombinant protein rMrcp-19k, respectively, as positive controls. (B) Antibody to Mrcp-100k was used for the analysis. Lane 2 shows the primary cement with the dithiothreitol/guanidine hydrochloride treatment. Lanes 3 and 4 show the barnacle peripheral shell and base plate, respectively, which have been decalcified and rendered soluble by the dithiothreitol/guanidine hydrochloride treatment. Lane 1 corresponds to GSF2 as a positive control.

Download figure to PowerPoint

Northern blot analysis using Mrcp-19k DNA as the probe indicated that the corresponding mRNA was specifically expressed in the basal portion of the barnacle where the cement gland was located (Fig. 5).


Figure 5.  Site specificity of Mrcp-19k gene expression in the basal portion of the adult barnacle, where the histologically identified cement gland is located.Twenty micrograms of total RNA extracted from the basal or upper portion of the adult barnacle was electrophoresed in formaldehyde gel, transferred to a nylon membrane, and hybridized with a probe. The basal portion mainly comprises the mantle, muscle, ovariole, cement gland [20–22], and hemolymph, whereas the upper portion contains the cirri, thorax, prosoma and hemolymph. Left, northern blot; right, 18S rRNA on gel stained by ethidium bromide.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

The present study identified a novel protein, cp-19k, in the cement of the barnacle. Amino acid composition analysis indicated that this protein is heavily biased toward six residues, namely, Gly, Thr, Ser, Ala, Lys and Val, with their total proportion exceeding 66% in M. rosa. MALDI-TOF MS analysis of Mrcp-19k isolated from barnacle cement, as well as scrutiny of the specific staining for glycosylation and phosphorylation, revealed that the protein is a simple one bearing no post-translational modifications. As all mussel foot proteins found so far are subjected to extensive post-translational modifications [1], mussel underwater attachment relies heavily on the functionality of modified amino acids [15]. Among the barnacle cement proteins, at least Mrcp-19k, and another cement protein Mrcp-20k, which has been identified previously [11], were shown to be simple proteins. Thus, the barnacle seems to manage its underwater attachment activity well with common amino acids.

The bacterial recombinant protein of Mrcp-19k, rMrcp-19k, was prepared in soluble form under physiologic conditions, enabling us to directly measure its adsorption to underwater surfaces. Two Cys residues in the protein formed an intramolecular disulfide bond, probably with the help of a thioredoxin-tag in the vector system. rMrcp-19k was adsorbed to various characteristic surfaces, including negatively charged, positively charged and hydrophobic surfaces. The barnacle attaches to various foreign material surfaces, including metal oxide, glass, plastic, wood, and rock. Naturally occurring surfaces such as rock are not microscopically homogeneous, and have a patchwork of different surface characteristics. The cement is therefore required to simultaneously adapt the molecular event to different surfaces. The ability of Mrcp-19k to be adsorbed to various surfaces suggests that this protein may be responsible for the surface functions, at least for the ability of the barnacle cement to adsorb to foreign materials with different surface characteristics.

Polycrystalline gold and hydrophobic alkylated gold were used as the representative surfaces in this study for evaluating the adsorption isotherm. The surface attachment area of a protein molecule on the gold surface was calculated to be 37 nm2 per molecule by assuming full surface coverage. Although no information is available on the three-dimensional structure of Mrcp-19k, this value is higher than the surface contact area of the well-known globular protein lysozyme (bacteriophage lambda; molecular mass 17 700 Da [16] 32 × 32 × 40Å, approximately 8–10 nm2 per molecule), which has a similar molecular mass. Thus, the Mrcp-19k molecule may be flatter to maximize contact with the material surface. The adsorption to alkylated gold was two-thirds of that to bare gold. It is not clear from this study whether this was due to an enlarged contact area of the protein as a result of conformational change on the surface, or imperfect surface coverage at some distance as a result of intermolecular repulsion on the surface. The amounts adsorbed to both glass and formaldehyde resin were two-fold to five-fold the amount adsorbed to bare gold. These data, however, were obtained with a method that involved a different principle of measurement, making a direct comparison difficult at this stage.

The fact that the amino acid compositions have been well conserved in cp-19k from three species, although the similarity of sequences was by no means high, indicates that the function of the protein may be associated with the amino acid bias. The four amino acids Ser, Thr, Lys and Val in the six amino acid-biased protein would be useful for coupling with various foreign material surfaces via hydrogen bonding, electrostatic interactions, hydrophobic interactions, etc. During the initial process of underwater attachment, a cement protein is required to approach the solid substratum to which water molecules are bound, and to displace this water prior to coupling with the substratum surface. Waite [5] has suggested the significance of the hydroxyl group on the Ser residue and Thr residue for the priming process. In a relevant protein, the antifreeze protein, which binds to the ice nucleus to inhibit crystal growth in the cytosolic space of several organisms, including bacteria and fish [17], the Ala and/or methyl group of Thr on the molecular surface of the protein are known to be essential in the process of binding to the ice nucleus [18,19], although the exact roles of these amino acids are not yet clearly understood. The requirements of coupling to various foreign material surfaces and displacing water molecules bound to a solid substratum may result in the bias of six amino acids in the barnacle cement protein.

Although the content of Mrcp-19k in cement was not accurately determined in this study, it was by no means a major component. Cement proteins contributing to surface functions might be minor constituents, whereas the proteins for bulk functions [9] would be present in much higher amounts in the adhesive layer. Northern blot analysis has indicated that the Mrcp-19k gene is specifically expressed in the basal portion of the animal, where the histologically identified cement gland is located [20–22]. This result is consistent with that for Mrcp-100k [9]. The cement proteins are probably biosynthesized together in the cement gland and transported by a duct to the narrow interspace outside, between the animal's base and the foreign substratum.

In conclusion, this study has identified a novel protein, cp-19k, in barnacle cement and demonstrated that it is able to be adsorbed to various underwater surfaces, suggesting that this protein is a surface protein of the cement complex. Our results also revealed that the function of cp-19k is dependent on common amino acid residues on the molecular surface. This is in contrast to the underwater adhesive proteins of mussel and tubeworm studied so far, where modified amino acids have been found to play major roles [23,24]. The barnacle cement protein characterized in this study may therefore represent a new mechanism of biological adhesion, which is likely to be useful in helping the interdisciplinary links between biotechnology and material science, e.g. development of adsorbents for various material surfaces, of support for protein alignment on a solid surface [25–27], and of underwater adhesives for surgical use [6].

Experimental procedures

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References


All chemicals used were of the highest grade available, with most being purchased from Wako Pure Chemical Industries (Osaka, Japan) and Takara Shuzo Co. (Otsu, Japan). Two-fold-concentrated ASW was prepared by dissolving ASW (Senju Seiyaku Co., Osaka, Japan) in ultrapure water, which was ultrafiltered through an Mr 3000-cutoff membrane (YM3; Amicon-Millipore, Billerica, MA, USA).

Purification and characterization of Mrcp-19k

GSF1 and GSF2 were prepared from M. rosa cement basically as described previously [9]. Briefly, the cement was suspended in 10 mm sodium phosphate buffer at pH 6.0 containing 6 m guanidine hydrochloride, and the suspension was centrifuged at 200 000 g for 1 h at 20 °C (CS120 centrifuge with RP100AT rotor, Hitachi Koki, Tokyo, Japan). The protein fraction in the supernatant corresponded to GSF1. The precipitate in the GSF1 preparation was reduced with 0.5 m dithiothreitol/7 m guanidine hydrochloride/0.5 m Tris/HCl (pH 8.5)/20 mm EDTA at 60 °C for 2 h in a nitrogen atmosphere. The resulting supernatant was recovered as GSF2. Both fractions were dialyzed against 1% acetic acid at 4 °C, before being evaporated and stored at − 20 °C until needed. GSF1 and GSF2 were separated by SDS/PAGE (a Tris/Tricine buffer system, 16.5% T/3% C [28]). The band corresponding to Mrcp-19k was transferred to a poly(vinylidene difluoride) membrane (ProBlott; Applied Biosystems, Foster City, CA, USA) using a Tris/borate buffer containing 0.1% SDS [29], and was stained with Coomassie Brilliant Blue R-250. In order to get peptide fragments of Mrcp-19k, the band corresponding to Mrcp-19k on the poly(vinylidene difluoride) membrane before Coomassie Brilliant Blue staining was cut out and subjected to in situ enzymatic digestion [30] using lysylendopeptidase (Wako Pure Chemical Industries). The generated peptide fragments were separated and fractionated by RP-HPLC in a 3.9 mm diameter × 150 mm µ-Bondasphere column (C18, 100 Å; Waters, Milford, MA, USA). The amino acid sequence was determined with a Procise 494 cLC (Applied Biosystems) or PSQ-2 protein sequencer (Shimadzu, Kyoto, Japan). Mrcp-19k was also purified from GSF1 by ion exchange chromatography (SP Sepharose FF; Amersham Biosciences, Uppsala, Sweden). The column was equilibrated with 50 mm acetic acid, and eluted with a linear gradient of NaCl from 0 m to 0.6 m in 80 min. The fractions were monitored with a polyclonal antibody raised against the bacterial recombinant protein corresponding to the C-terminal 10 kDa portion of Mrcp-19k, as described in the latter section of recombinant in E. coli except for using 5′-TGG CCG CAG CCA TGG CAT TGG T-3′ as the 5′-primer. The fraction containing Mrcp-19k was concentrated by ultrafiltration (Microcon YM-3; Amicon-Millipore), and further purified by gel filtration chromatography (G3000SWXL; Tosoh, Tokyo, Japan) with 50 mm acetic acid/20 mm NaCl as the eluent. The purified Mrcp-19k was subjected to MALDI-TOF MS with a Voyager-DE STR instrument (Applied Biosystems) incorporating a 337 nm nitrogen laser operated in the linear mode at an acceleration voltage of 20 kV. For the MALDI matrix, saturated sinapinic acid dissolved in 30% (v/v) acetonitrile containing 0.3% (v/v) trifluoroacetic acid was used, and for calibration of the mass, a Sequazyme peptide mass standard kit (Applied Biosystems) was used. The amino acid composition was determined using a double-distilled constant-boiling HCl hydrolysate at 110 °C for 24 h, with an AccQ-Tag system (Waters). Possible modifications of Mrcp-19k by glycosylation and phosphorylation of Mrcp-19k were also investigated. The glycosylation was detected by periodic acid–Schiff staining [31] with BSA as the positive control. Phosphoryation was detected by staining the SDS/PAGE gel with Pro-Q diamond (Invitrogen, Eugene, OR, USA), and then observed under a UV-transilluminator, with BSA and bovine milk β-casein as positive controls.

Molecular cloning of cDNAs encoding Mrcp-19k, Balcp-19k, and Bicp-19k

M. rosa, B. improvisus and B. albicostatus were collected from Miyako Bay (Iwate), Yodo River (Osaka) and Shimizu Bay (Shizuoka, Japan), respectively. RNA and DNA manipulation was generally performed as described previously [9]. Total RNA was extracted from basal tissue of the barnacle by a Total RNA Separator kit (BD Biosciences Clontech, Mountain View, CA, USA), and poly(A)+ RNA was isolated using Oligo(dT)-Latex Super (Takara Shuzo Co.). cDNA was prepared from mRNA with a Zap-cDNA synthesis kit (Stratagene, La Jolla, CA, USA) according to the instructions of the supplier. DNA fragments of Mrcp-19k were first amplified by PCR (ExTaq; Takara) with fully degenerated PCR primers designed from the N-terminal amino acid sequence of Mrcp-19k: 5′-GTN CCN CCN CCN TGY GA-3′ and 5′-CAN CCY TTY TGY TTN ACY TT-3′. The PCR products were resolved by 3% NuSieve 3 : 1 agarose (Takara) gel electrophoresis, and a 53 bp DNA fragment from M. rosa was purified from the gel. The DNA fragment was subcloned in pT7 Blue T-Vector (Novagen, EMD Biosciences, Madison, WI, USA), and the insert was sequenced using a Prism Dye Deoxy sequencing kit and 3700-DNA analyzer (Applied Biosystems). 3′-RACE was then carried out with a specific 3′-RACE primer designed from the 53 bp DNA and using a 3′-RACE core kit (Takara). The 3′-RACE primer used was 5′-CTG ATC TAG AGG TAC CGG ATC CGT TCC CCC ACC ATG CGA CCT TGG CAT-3′. The PCR product was subcloned and then sequenced. To obtain the full-length cDNA, 5′-RACE was carried out with oligonucleotide primers designed from the sequence of 750 bp DNA and using a 5′-RACE core kit (Takara). The 5′-RACE primers used were as follows: 5′-G#CC GTC CCC GGC CGA C-3′, where G# is phosphorylated, for reverse transcription; 5′-GTG CCG GAG CCC TGC GTG GC-3′ and 5′-AAC TCC GTG GAG AAG AAG AA-3′ for the first PCR amplification; and 5′-TGC TGA CCG ACG CGC CTC CT-3′ and 5′-GGC AAC ACG GGC GTC ACC GC-3′ for the second PCR amplification. The 102 bp DNA amplified by 5′-RACE was purified, subcloned, and sequenced. Finally, 665 bp DNA for the coding region of Mrcp-19k was amplified from M. rosa total cDNA using the primers 5′-ACC AAC GCA GCA GTT ATG GT-3′ and 5′-GCT GCA CAT CTT CGA CCT CA-3′, and then subcloned. KOD-plus DNA polymerase (Toyobo, Osaka, Japan) was used for PCR amplification to achieve high fidelity. Ten randomly selected clones were sequenced.

DNA fragments encoding Balcp-19k and Bicp-19k were amplified by 3′-RACE, respectively, using the degenerated oligonucleotide primer designed for 3′-RACE of Mrcp-19k as already described. The amplified DNA fragments were subcloned and sequenced.

A homology search was performed with the nonredundant GenBank CDS translations + Protein Data Bank + swissprot + PIR + PRF database using the blast program [32]. A sequence profile-based fold-recognition method involving the mgenthreader[33] program was used for further analysis to identify a family of homologous proteins. The clustalw[34] program was used to identify the clustered sequence alignment among cp-19ks.

Characterization of the Mrcp-19k recombinant in E. coli

The Mrcp-19k recombinant in E. coli, designated rMrcp-19k, was prepared as follows. The cDNA was amplified by PCR with primers around the mature N-terminal and C-terminal regions, which, respectively, included the newly created NcoI and BamHI restriction sites. The primers used were 5′-ACCGGCCATGGGCAAGGCCGT-3′ and 5′-ATGGTCACGGGATCCCTCCGGTGGTCTTA, whereby the recombinant was designed to have the N-terminal sequence of AMGKAVTV, in which the mature N-terminal sequence of Mrcp-19k with an additional dipeptide sequence, AM, was created after removing the fused tag by enterokinase cleavage, and with the original C-terminal end. The amplified DNA was subcloned in pT7 Blue T-Vector (Novagen), and the sequence was confirmed. Insert DNA was generated by digestion with the NcoI and BamHI restriction enzymes, and then subcloned into pET32b (Novagen) with the same restriction sites. The pET32 vector system produces fusion proteins with a thioredoxin-tag, which enhances disulfide bond formation of the target protein in the cytoplasm of the host strain. The created vector was transformed into the expression host strain Oligami (DE3) (Novagen). The recombinant protein was purified with a metal-chelating column according to the affinity of the His-tag fused into Mrcp-19k. The cells were inoculated in LB medium [35] containing ampicillin at 37 °C for 3 h, and transferred to freshly prepared medium, and inoculated for another 3 h; protein expression was induced by 0.2% isopropyl-thio-β-d-galactoside for an additional 4 h. The cytosolic fraction was prepared by sonicating on ice in 20 mm Tris (pH 7.4)/500 mm NaCl/40 mm imidazole, and purified with an Ni2+-immobilized column (His-bind kit; Novagen) according to the manufacturer's instructions. The fraction containing rMrcp-19k was recovered and dialyzed at 4 °C against 20 mm Tris/HCl (pH 7.4)/50 mm NaCl/2 mm CaCl2, and then treated with enterokinase (recombinant enterokinase; Novagen) to cleave the fused tag. The cleaved protein was purified by Ni2+-immobilized column chromatography and gel filtration column chromatography (TSK-gel G3000 SWXL; Tosoh), using 20 mm Tris (pH 7.4)/50 mm NaCl as the eluent. For the following quantitative amino acid analysis, the solvent was changed to 20 mm Hepes (pH 7.4)/20 mm NaCl by dialysis. The protein was quickly stored at − 80 °C in small volumes. Freeze–thaw cycles and storage for more than 1 month were avoided, and handling of sample solutions was minimized, because these processes caused loss of the protein in solution.

Inspection of the chemical forms of two Cys residues in the recombinant protein was performed as follows. The protein was digested with lysylendopeptidase [Wako; enzyme/substrate, 1 : 100 (molar ratio)] in 20 mm Tris (pH 7.4)/50 mm NaCl at 30 °C for 3 h, and the molecular masses of the resulting peptide fragments were determined by LC-ESI-MS (LCQ-Advantage instrument; Thermo Electron, Waltham, MA, USA), either with or without pretreatment with dithiothreitol.

Adsorption of the recombinant protein to underwater material surfaces was analyzed by: (a) quantitative amino acid analysis; and (b) SPR.

Protein adsorption to glass and a positively charged polymer were evaluated by quantification of the bound protein and unbound protein, respectively, by amino acid analysis after hydrolysis (see details in supplementary Fig. S3). The substrates to be analyzed were the inner surface of small glass test tube (5 mm in diameter and 29 mm in length, 73.6 mm2 for covering the surface area of a 20 µL solution) and benzoguanamine/formaldehyde resin particles (Epostar L15, 11.6 µm in diameter, 73.6 mm2 for the surface area test; Nippon Shokubai, Osaka, Japan). The amount adsorbed in 5 min at 25 °C in ASW was measured with several protein concentrations and fitted using a Langmuir adsorption isotherm.

The SPR measurements were performed with a BIAcore 3000 system (Biacore AB, Uppsala, Sweden) at 25 °C and with a flow rate of 10 µL·min−1. The sensor chips of polycrystalline gold-coated and octadecanethiol-terminated gold, HPA, were purchased from BIAcore. The running buffer was 10 mm Tris (pH 7.4)/25 mm NaCl with or without ASW. A baseline was first established by pumping the buffer, and the port was then switched to the protein solution. After saturation of the protein, the buffer was pumped once again to monitor the desorption behavior. rMrcp-19k at a concentration of 4 µm was adequately diluted by the buffer, before being mixed with the same volume of buffer with or without 2 × ASW immediately before injection. The mixing process with 2 × ASW was used to minimize the exposure of the protein to any higher salt concentration. The mass uptake of protein, ΔmSPR, was evaluated by the relationship

  • image

where ΔRU is the measured change in response units, and Cspr has been calibrated to be 6.5 × 10−2 ng−1·cm−2 for adsorption to a flat surface [36]. The kinetics for adsorption of rMrcp-19k to gold and alkylated gold were evaluated using biaevaluation version 3.1 software that was supplied with the instrument.

Localization and expression site of Mrcp-19k

To confirm that cp-19k was a cement component, the localization of Mrcp-19k in the primary cement and in the protein fractions of both the base shell and peripheral shell of the animal were investigated by western blotting. Polyclonal antibodies were raised using bacterial recombinants of the respective C-terminal regions of approximately 10 kDa in Mrcp-19k and Mrcp-100k as antigens in rabbits with serial subcutaneous injections. The recombinants were prepared as described earlier. The primers used for amplifying the Mrcp-19k portion were 5′-TGG CCG CAG CCA TGG CAT TGG T-3′ and 5′-ACC TCA GGA TCC AGG TCG AGA AAA-3′. The primers used for amplifying the Mrcp-100k portion were 5′-AGT GCA GCC CAT GGG GGC AGC CAT-3′ and 5′-TTG CCT AGG TGG ATC CTC AGC ATC TGA A-3′. M. rosa primary cement was collected as previously reported [10]. The base and peripheral shell were separately collected from living M. rosa specimens, and physically cleaned to remove all contamination by the animals' soft tissue. Each shell was decalcified by dialyzing against 2% acetic acid at 4 °C, and the precipitate was recovered. Although the supernatant was also analyzed, no signal was detected by western blotting. The precipitate was evaporated to dryness, denatured, separated by SDS/PAGE (a Tris/Tricine buffer system, 16.5% T/3% C for Mrcp-19 k, and 8% T [37] with 6 m urea for Mrcp-100k), and finally subjected to western blotting as described elsewhere [38].

To evaluate the expression site of the Mrcp-19k gene in the animal, RNAs were separately purified from tissues in the upper or lower part of the barnacle in the same manner as described above. The upper part included the cirri, thorax, prosoma and hemolymph, and the lower part included the mantle, muscle, ovariole, cement gland and hemolymph. Twenty micrograms of RNA was electrophoresed and transferred to a Hybond-N+ nylon membrane (Amersham Biosciences). The 540 bp DNA encoding the Mrcp-19k ORF was labeled with [32P]dCTP[α32P] using a Random Primer DNA Labeling kit (Takara Shuzo Co.). The labeled probe thus obtained was used for northern blotting analysis with the prepared membrane.


  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References

We thank Ms Futaba Sasaki and Ms Chikako Kajimoto for their technical assistance. Special thanks are given to Professor J.-.R. Shen of Okayama University for his critical reading of the manuscript. Part of this work was performed as an industrial science and technology project entitled Technological Development for Biomaterials Design Based on Self-organizing Proteins, which is supported by The New Energy and Industrial Technology Development Organization (NEDO).


  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  • 1
    Sagert J, Sun C & Waite JH (2006) Chemical subtleties of mussel and polychaete holdfasts. In Biological Adhesives (Smith AM & Callow JA, eds), pp. 125140. Springer-Verlag, Berlin.
  • 2
    Taylor SW & Waite JH (1997) Marine adhesives: from molecular dissection to application. In Protein-Based Materials (McGrath K & Kaplan D, eds), pp. 217248. Birkhauser, Boston, MA.
  • 3
    Zhao H & Waite JH (2006) Linking adhesive and structural proteins in the attachment plaque of Mytilus californianus. J Biol Chem 281, 2615026158.
  • 4
    Kamino K (2006) Barnacle underwater attachment. In Biological Adhesives (Smith AM & Callow JA, eds), pp. 145166. Springer-Verlag, Berlin.
  • 5
    Waite JH (1987) Nature's underwater adhesive specialist. Int J Adhes 7, 914.
  • 6
    Tay FR & Pashley DH (2002) Dental adhesives of the future. J Adhes Dent 4, 91103.
  • 7
    Walker G (1972) The biochemical composition of the cement of two barnacle species, Balanus hameri and Balanus crenatus. J Mar Biol Assoc UK 52, 429435.
  • 8
    Saroyan JR, Linder E, Dooley CA & Bleile HR (1970) Repair and reattachment in the Balanidae as related to their cementing mechanism. Ind Eng Chem Prod Res Dev 9, 122133.
  • 9
    Kamino K, Inoue K, Maruyama T, Takamatsu N, Harayama S & Shizuri Y (2000) Barnacle cement proteins. J Biol Chem 275, 2736027365.
  • 10
    Kamino K, Odo S & Maruyama T (1996) Cement proteins of the acorn barnacle, Megabalanus rosa. Biol Bull 190, 403409.
  • 11
    Kamino K (2001) Novel barnacle underwater adhesive protein is a charged amino acid-rich protein constituted by a Cys-rich repetitive sequence. Biochem J 356, 503507.
  • 12
    Naldrett MJ (1993) The importance of sulphur cross-links and hydrophobic interactions in the polymerization of barnacle cement. J Mar Biol Assoc UK 73, 689702.
  • 13
    Jones DT, Taylor WR & Thornton JM (1992) The rapid generation of mutation data matrices from protein sequences. J CABIOS 8, 275282.
  • 14
    Naldrett MJ & Kaplan DL (1997) Characterization of barnacle (Balanus eburneus and B. crenatus) adhesive proteins. Mar Biol 127, 629635.
  • 15
    Lee H, Scherer NF & Messersmith PB (2006) Single-molecule mechanics of mussel adhesion. Proc Natl Acad Sci USA 103, 1299913003.
  • 16
    Evrard C, Fastrez J & Declercq J-P (1998) Crystal structure of the lysozyme from bacteriophage lambda and its relationship with V and C-type lysozymes. J Mol Biol 276, 151164.
  • 17
    Fletcher GL, Hew CL & Davies PL (2001) Antifreeze proteins of teleost fishes. Annu Rev Physiol 63, 359390.
  • 18
    Zhang W & Laursen RA (1998) Structure–function relationships in a type I antifreeze polypeptide. The role of threonine methyl and hydroxyl groups in antifreeze activity. J Biol Chem 273, 3480634812.
  • 19
    Jia Z & Davies PL (2002) Antifreeze proteins: an unusual receptor–ligand interaction. Trends Biochem Sci 27, 101106.
  • 20
    Walker G (1970) The histology, histochemistry and ultrastructure of the cement apparatus of three adult sessile barnacles, Elminius modestus, Balanus balanoides and Balanus haemri. Mar Biol 7, 239248.
  • 21
    Lacombe D (1970) A comparative study of the cement glands in some balanid barnacles (cirripedia, balanidae). Biol Bull 139, 164179.
  • 22
    Saroyan JR, Lindner E & Dooley CA (1970) Repair and reattachment in the balanidae as related to their cementing mechanism. Biol Bull 139, 333350.
  • 23
    Statz AR, Meagher RJ, Barron AE & Messersmith PB (2005) New peptidomimetic polymers for antifouling surfaces. J Am Chem Soc 127, 79727973.
  • 24
    Deming TJ (1999) Mussel byssus and biomolecular materials. Curr Opin Chem Biol 3, 100105.
  • 25
    Chen X, Ferrigno R, Yang J & Whitesides GM (2002) Redox properties of cytochrome c adsorbed on self-assembled monolayers: a probe for protein conformation and orientation. Langmuir 18, 70097015.
  • 26
    Hyun J, Lee WK, Nath N, Chilkoti A & Zauscher S (2004) Capture and release of proteins on the nanoscale by stimuli-responsive elastin-like polypeptide ‘switches’. J Am Chem Soc 126, 73307335.
  • 27
    Das R, Kiley PJ, Segal M, Norville J & Yu AA (2004) Integration of photosynthetic protein molecular complexes in solid-state electronic devices. Nano Lett 4, 10791083.
  • 28
    Schäger H & Jagow G (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal Biochem 166, 368379.
  • 29
    Ikeuchi M, Takio K & Inoue Y (1989) N-terminal sequencing of photosystem II low-molecular-mass proteins: 5 and 4.1 kDa components of the O2-evolving core complex from higher plants. EJB Lett 242, 263269.
  • 30
    Iwamatsu A & Yoshida-Kubomura N (1996) Systematic peptide fragmentation of polyvinylidene difluoride (PVDF)-immobilized proteins prior to microsequencing. J Biochem (Tokyo) 120, 2934.
  • 31
    Zacharius RM, Tatiana EZ, Morrison JH & Woodlock JJ (1969) Glycoprotein staining following electrophoresis on acrylamide gels. Anal Biochem 30, 148152.
  • 32
    Altschul SF, Gish W, Miller W, Myers EW & Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215, 403410.
  • 33
    McGuffin LJ & Jones DT (2003) Improvement of the GenTHREADER method for genomic fold recognition. Bioinformatics 19, 874881.
  • 34
    Higgins D, Thompson J, Gibson T, Thompson JD, Higgins DG & Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 46734680.
  • 35
    Luria SE & Burrous JW (1957) Hybridization between Escherichia coli and Shigella. J Bacteriol 74, 461476.
  • 36
    Hook F, Kasemo B, Nylander T, Fant C, Sott K & Elwing H (2001) Variations in coupled water, viscoelastic properties, and film thickness of a Mefp-1 protein film during adsorption and cross-linking: a quartz crystal microbalance with dissipation monitoring, ellipsometry, and surface plasmon resonance study. Anal Biochem 73, 57965804.
  • 37
    Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680685.
  • 38
    Enami I, Murayama H, Ohta H, Kamo M, Nakazato K & Shen J-R (1995) Isolation and characterization of a photosystem II complex from the red alga Cyanidium caldarium. Association of cytochrome c-550 and a 12 kDa protein with the complex. Biochim Biophys Acta 1232, 208216.