3724
Langmuir 2004, 20, 3724-3729
Cross-Linking the Protein Precursor of Marine Mussel Adhesives: Bulk Measurements and Reagents for Curing Jennifer Monahan and Jonathan J. Wilker* Department of Chemistry, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907 Received December 3, 2003. In Final Form: February 7, 2004 Marine mussels affix themselves to surfaces by use of a highly cross-linked, protein-based adhesive. Metal levels (e.g., Fe, Zn, Cu, Mn) of the cured glue are significantly concentrated relative to surrounding waters. Specific details on the reagents used by mussels to induce protein cross-linking are not known at this time. To provide insight on the cross-linking agents and reactions taking place while curing mussel glues, we performed a study in which various compounds were tested for the ability to bring about protein curing. A precursor to adhesion, with proteins containing the unusual amino acid 3,4-dihydroxyphenylalanine, was extracted from mussel feet. Potential cross-linking agents were mixed with this gelatinous pellet. The compressibility and shear properties of the resulting material were investigated by use of a penetration test. The reagents examined included simple metal ions (e.g., Na+, Zn2+), oxidizing transition metals (e.g., Fe3+, Cr2O72-), nonmetallic oxidants (e.g., H2O2, IO4-), and oxidizing enzymes (e.g., tyrosinase). We found that protein curing was brought about by simple oxidants and transition metal ions. The results show that optimal curing occurs when the reagent is an oxidizing metal ion (e.g., MnO4-, Fe3+). We conclude that marine mussels are likely to employ Mn3+ and Fe3+ for protein cross-linking and adhesive synthesis.
Introduction Mussel glues, barnacle cements, kelp adhesives, and coral reefs are some examples of the fascinating materials found within ocean waters.1,2 A general theme in marine biomaterial generation appears to be the cross-linking of biological polymers, typically proteins, into hardened matrixes.1-3 Perhaps the most studied of marine adhesives are those produced by mussels.1-6 These mollusks affix themselves to surfaces by synthesizing proteins containing the unusual amino acid 3,4-dihydroxyphenylalanine (DOPA).7,8 The DOPA-containing proteins are deposited and cross-linked into the final matrix of cured mussel glues.5,6,8,9 Still to be discovered are the cross-linking agents used by mussels to cure their adhesive as well as the chemical bonding within the final material. Potential cross-linking agents for DOPA-containing proteins include enzymes, chemical oxidants, and metal ions.2,3,10 To date, no enzymes or simple oxidants have been found in the mussel adhesive assembly. The ability of mussels to concentrate metal ions from seawater into their adhesive has been known for some time.11-13 For example, the mussel adhesive plaques contain iron, zinc, * To whom correspondence should be addressed. E-mail: wilker@ purdue.edu. Telephone: 765-496-3382. (1) Rzepecki, L. M.; Waite, J. H. In Bioorganic Marine Chemistry; Scheuer, P. J., Ed.; Springer-Verlag: New York, 1991; Vol. 4, pp 119148. (2) Waite, J. H. Comp. Biochem. Physiol. 1990, 97B, 19-29. (3) Deming, T. J. Curr. Opin. Chem. Biol. 1999, 3, 100-105. (4) Rzepecki, L. M.; Chin, S.-S.; Waite, J. H.; Lavin, M. F. Mol. Mar. Biol. Biotechnol. 1991, 1, 78-88. (5) Tamarin, A.; Lewis, P.; Askey, J. J. Morphol. 1976, 189, 261270. (6) Benedict, C. V.; Waite, J. H. J. Morphol. 1986, 189, 171-181. (7) Waite, J. H. J. Biol. Chem. 1983, 258, 2911-2915. (8) Waite, J. H.; Housley, T. J.; Tanzer, M. L. Biochemistry 1985, 24, 5010-5014. (9) Vreeland, V.; Waite, J. H.; Epstein, L. J. Phycol. 1998, 34, 1-8. (10) Waite, J. H.; Tanzer, M. L. Science 1981, 212, 1038-1040. (11) George, S. G.; Pirie, B. J. S.; Coombs, T. L. J. Exp. Mar. Biol. 1976, 23, 71-84. (12) Coombs, T. L.; Keller, P. J. Aquat. Toxicol. 1981, 1, 291-300. (13) Donat, J. R.; Bruland, K. W. In Trace Elements in Natural Waters; Salbu, B., Steinnes, E., Eds.; CRC Press: Ann Arbor, MI, 1995; pp 247-281.
copper, and manganese11,12 at levels often 100 000 times greater than those in open ocean waters.13 Consequently we have become interested in exploring functional roles for metal ions in glue synthesis. Iron is known to bind DOPA-containing mussel adhesive proteins14 and hydrolyzed peptides thereof.15 We performed spectroscopic studies on intact adhesive produced by live mussels, protein extracted prior to cross-linking, and synthetic peptide models to reveal Fe(DOPA)3 moieties in this material.16 Electron paramagnetic resonance (EPR) spectroscopy showed an iron-induced oxidation of the DOPA proteins to yield an organic radical.16 Ultravioletvisible absorption (UV-vis) and EPR spectroscopies provided evidence for the existence of Fe(DOPA)3 complexes in mussel glue, an adhesive film prepared by addition of iron to protein, and peptide models.16 The affinity of iron for DOPA,14,15,17 catechol,18 and similar ligands19 is well established. Our spectroscopic experiments led us to the conclusion that mussel glues are formed by addition of Fe3+ to the protein precursors to adhesion.16 An initial cross-link is created by Fe3+ binding three DOPA-containing protein strands.16 Subsequent protein oxidation by this Fe(DOPA)3 moiety generates radical species to enable further protein-protein bonding or a means of creating protein-surface bonds for adhesion.16 We wondered if these spectroscopic findings would correlate to the actual curing of mussel adhesive precursors. Iron has been shown to increase the adhesive energy of isolated mussel protein when observed by atomic force microscopy.20 On a bulk scale, however, little data are (14) Taylor, S. W.; Chase, D. B.; Emptage, M. H.; Nelson, M. J.; Waite, J. H. Inorg. Chem. 1996, 35, 7572-7577. (15) Taylor, S. W.; Luther, G. W., III; Waite, J. H. Inorg. Chem. 1994, 33, 5819-5824. (16) Sever, M. J.; Weisser, J. T.; Monahan, J.; Srinivasan, S.; Wilker, J. J. Angew. Chem., Int. Ed. 2004, 43, 448-450. (17) Taylor, S. W.; Hawkins, C. J.; Winzor, D. J. Inorg. Chem. 1993, 32, 422-427. (18) Avdeef, A.; Sofen, S. R.; Begante, T. L.; Raymond, K. N. J. Am. Chem. Soc. 1978, 100, 5362-5370. (19) Pierpont, C. G.; Lange, C. W. Prog. Inorg. Chem. 1994, 41, 331442. (20) Frank, B. P.; Belfort, G. Biotechnol. Prog. 2002, 18, 580-586.
10.1021/la0362728 CCC: $27.50 © 2004 American Chemical Society Published on Web 03/19/2004
Protein Precursor of Marine Mussel Adhesives
available to indicate which reagents are best at bringing about material formation. Early efforts at measuring adhesive shear forces on purified DOPA-containing proteins did not prove to be practical.21 Consequently, we obtained data on curing of the adhesive precursor by a penetration method.22-26 A gelatinous precursor to the mussel adhesive was extracted and mixed with metal salts and examined for curing. In a preliminary account of these experiments, we added metal salts of the first row transition elements to the adhesive precursor.27 This initial survey of biologically available ions demonstrated preferential curing brought about by Fe3+ ions.27 These materials testing results27 complemented our spectroscopic data16 to indicate the importance of iron in generating biological materials. Beyond simply understanding the processes used by mussels to make their glue, we wish to determine the predominant factors dictating syntheses of biological materials, in general. Consequently, we embarked upon an extensive study in which the precursor to mussel adhesive was reacted with a wide array of inorganic salts and nonmetallic oxidants, as well as enzymes and UV light. These curing experiments were performed with material extracted from excised mussel feet. The foot is the organ in which mussels synthesize their DOPAcontaining proteins.3,9,28 The foot is also the apparatus used for application of the adhesive to a surface.3,9,28 Soluble DOPA-containing proteins can be extracted from excised feet by blending in dilute acid.29 Precipitation of the soluble proteins with acetone, followed by centrifugation, yields a gelatinous pellet containing the proteins Mefp-1 and Mefp-2, known precursors to mussel adhesive plaques.29 In an attempt to mimic the production of mussel glue, we are using this gelatinous mixture of adhesive proteins to determine the factors dictating curing. To obtain data on protein cross-linking, we employed a direct measurement of curing. After mixing the gelatinous precursor with potential cross-linking reagents, the compressibility and shear properties of the resulting materials were recorded. These data were obtained by penetrating the material with a rod at constant velocity. Force data were recorded during this penetration. Low force values indicate a lack of curing.22-26 Conversely, high values represent a hardened material requiring greater force to move the rod through at constant velocity.22-26 These direct measurements of curing uncovered some general principles dictating protein cross-linking. Certain metal ions brought about curing, whereas the majority of those examined did not. Simple oxidants were capable of hardening the extract, but only to a limited degree. Optimal curing was found for reagents that are oxidizing metal ions (e.g., Fe3+, Cr2O72-). Interestingly, no correlation was found between adhesive hardening and reduction potentials. These data are presented in an effort to understand the processes dictating biological material formation. Material and Methods Extraction of Adhesive Precursor. Protein extracts were prepared by use of a published procedure.29 Briefly, excised feet from the common blue mussel, Mytilus edulis (Northeast (21) Monahan, J.; Wilker, J. J. 2003, Unpublished results. (22) Bourne, M. C. J. Food Sci. 1966, 31, 282-291. (23) Bourne, M. C. J. Food Sci. 1967, 32, 601-605. (24) deMan, J. M. J. Texture Stud. 1969, 1, 114-119. (25) Bourne, M. C. J. Texture Stud. 1975, 5, 459-469. (26) Bourne, M. C. Food Texture and Viscosity: Concept and Measurement; Academic Press: New York, 1982. (27) Monahan, J.; Wilker, J. J. Chem. Commun. 2003, 1672-1673. (28) Waite, J. H. Int. J. Adhes. Adhes. 1987, 7, 9-14. (29) Waite, J. H. Methods Enzymol. 1995, 258, 1-20.
Langmuir, Vol. 20, No. 9, 2004 3725 Transport; Waldoboro, ME), were ground in a 0.7% (w/v) perchloric acid solution using an Osterizer blender. The suspension was centrifuged at 31 000g for 30 min at 4 °C. The supernatant, containing soluble adhesive proteins, was retained and acidified with concentrated sulfuric acid (1.68% v/v). The adhesive proteins were then precipitated from solution by dropwise addition of acetone. Centrifuging at 31 000g for 30 min at 4 °C collected the extracted protein as a gelatinous pellet. Drying experiments showed the pellets to be ∼92% water. This pellet was used for the materials tests described below. Pellet Homogenization and Sample Preparation. These experiments required the extraction of a protein-based adhesive precursor from 1340 g of excised mussel feet. The extractions were performed over the course of 3 weeks, and all pellets were stored at 4 °C under a blanket of water. To maintain uniform samples, 833 gelatinous pellets, ∼0.8 g each, were collected and homogenized together after the addition of deionized water (25% w/w). Thus all data reported here are from one batch of extract. Homogenization was accomplished by forcing the pellets through a plastic strainer with 3.5 mm holes, producing a paste. The entire sample was thoroughly mixed with a spatula. A plastic syringe (10 mL volume, 2 mm opening) was used to dispense 1.00-1.05 g of homogenized extract into 9 mm i.d. × 35 mm long plastic microcentrifuge tubes (2 mL volume). Each sample was promptly capped to prevent evaporation and tapped on a laboratory bench to pack the gelatinous material. During preparation, all samples were kept on ice. Once prepared, samples were stored in the dark at 4 °C. Prior to use, no sample was stored longer than 7 days. Under these conditions, no trends were observed to indicate that age influenced the sample. Reagent Solutions. To test the effects of various reagents on cross-linking of the protein extract, a series of metal salts and oxidant solutions were prepared. Each reagent solution was made fresh daily at a concentration of 0.5 M in deionized water (Barnstead Nanopure Infinity). All reagents tested can be found in Table 1. Where commercially available, both the nitrate and chloride salts for a given metal and oxidation state were investigated. In the case of manganese, the acetate salts were included. If the solubility of a given reagent in water would not allow a concentration of 0.5 M, a saturated solution of said metal salt was used instead. The saturated solutions were TiF3, Mn(OOCH3)3, MnO2, CuCl, SnCl2, and NaIO3. For concentration studies, the fresh 500 mM solution of each reagent was diluted to prepare 50 and 5 mM solutions. To manage the large number of samples tested and minimize any sample aging over the course of the experiment, a single test of every reagent at each of the three concentrations was conducted daily. All sample tests were repeated over the course of five consecutive days, and the results were averaged. For example, test results for 50 mM Co(NO3)2 are an average of five runs, each performed on a different day. Penetration Tests. To test the curing ability of each reagent, 100 µL of the reagent solution was added to 1 g of protein extract. The sample was thoroughly mixed using a microspatula, capped, and tapped on the bench to eliminate air bubbles. The sample was then allowed to react at room temperature for 60 min. Control tests were conducted through the addition of 100 µL of deionized water to 1 g of protein extract. When discussing the concentration of a given reagent, concentrations are reported as the final material content. For example, when 100 µL of 500 mM Co(NO3)2 is added to 1 g of protein extract (density ∼ 1 g/mL), the final concentration is reported to be 45 mM (see Table 1). To investigate changes in the materials properties of the protein matrix, a penetration test was conducted on each sample.22-26 An Instron 5544 Materials Testing Machine was used to drive a steel rod into the gelatinous sample at a constant velocity. The rod, a 3.5 mm blank drill bit, penetrated the sample at a rate of 20 mm/min. For each run, the rod was lowered 20 mm from a starting position, typically ∼5 mm above the sample, to yield a total penetration depth of ∼15 mm for each 1 g sample. During penetration, a 5 N load cell was used to monitor the resistive force of the sample against the rod. A sample with extensive cross-linking is expected to generate a higher resistive force against penetration than one with limited cross-linking and thus should require an increased force to lower the rod at a constant velocity.22-26 All penetration forces reported here are measured at a final rod extension of 20 mm. All reagents were
3726
Langmuir, Vol. 20, No. 9, 2004
Monahan and Wilker
Table 1. Average Penetration Forces for Mussel Adhesive Extract Reacted with Various Reagents at Different Concentrationsa reagent LiCl LiNO3 NaCl NaNO3 KCl KNO3 MgCl2 Mg(NO3)2 CaCl2 Ca(NO3)2 TiF3 VCl3 VOSO4 Na3VO4 CrCl2 Cr(NO3)3 Na2Cr2O7 MnCl2 Mn(OAc)2 Mn(OAc)3 MnO2 KMnO4 FeCl2 FeCl3 Fe(NO3)3 CoCl2 Co(NO3)2 NiCl2 Ni(NO3)2 CuCl CuCl2 Cu(NO3)2 ZnCl2 Zn(NO3)2 AlCl3 Al(NO3)3 Ga(NO3)3 SnCl2 SnCl4 AgNO3 NaIO4 NaIO3 Na2S2O8 H2O2 BHP Na2S2O4 DTT HNO3 HCl HClO4 tyrosinase laccase
oxidation state Li+
Li+ Na+ Na+ K+ K+ Mg2+ Mg2+ Ca2+ Ca2+ Ti3+ V3+ V4+ V5+ Cr2+ Cr3+ Cr6+ Mn2+ Mn2+ Mn3+ Mn4+ Mn7+ Fe2+ Fe3+ Fe3+ Co2+ Co2+ Ni2+ Ni2+ Cu+ Cu2+ Cu2+ Zn2+ Zn2+ Al3+ Al3+ Ga3+ Sn2+ Sn4+ Ag+
45 mM force (mN) 15 ( 1 12 ( 3 13 ( 2 14 ( 3 15 ( 5 12 ( 2 13 ( 2 15 ( 2 21 ( 1 21 ( 3 14 ( 2 13 ( 2 12 ( 5 626 ( 219 15 ( 3 13 ( 2 784 ( 114 14 ( 3 15 ( 3 274 ( 75 11 ( 1 409 ( 64 16 ( 9 68 ( 12 66 ( 14 15 ( 2 13 ( 1 13 ( 3 13 ( 2 16 ( 3 21 ( 4 18 ( 4 16 ( 6 13 ( 1 14 ( 2 14 ( 4 13 ( 2 13 ( 2 48 ( 5 27 ( 7 73 ( 6 31 ( 9 35 ( 8 20 ( 4 19 ( 2 14 ( 3 13 ( 1 12 ( 3 13 ( 1 15 ( 4 12 ( 1 14 ( 2
nb 5 5 5 5 5 4 5 5 5 5 5 5 5 4 5 5 5 5 5 5 5 4 4 5 5 4 5 5 5 5 5 5 5 4 5 5 5 4 4 5 5 5 5 5 4 5 5 5 5 5 5 5
4.5 mM force (mN) 13 ( 2 12 ( 3 14 ( 1 14 ( 1 13 ( 2 15 ( 2 13 ( 2 15 ( 2 14 ( 2 16 ( 4 13 ( 2 14 ( 3 14 ( 2 64 ( 12 11 ( 1 12 ( 2 100 ( 8 13 ( 3 14 ( 1 200 ( 5 12 ( 2 27 ( 6 14 ( 2 34 ( 3 33 ( 9 16 ( 4 13 ( 1 13 ( 3 12 ( 2 11 ( 3 13 ( 1 17 ( 4 14 ( 3 19 ( 10 13 ( 2 13 ( 1 13 ( 3 12 ( 2 18 ( 1 18 ( 2 47 ( 5 47 ( 9 14 ( 7 16 ( 6 22 ( 2 13 ( 2 12 ( 3 11 ( 2 15 ( 3 14 ( 4 14 ( 4 14 ( 2
nb 5 5 5 5 5 5 5 5 5 5 4 4 4 4 5 4 5 5 5 5 5 5 5 5 5 5 5 5 5 4 4 4 4 3 4 4 4 4 4 4 5 5 4 4 5 5 4 5 5 5 5 5
0.45 mM force (mN) 16 ( 4 14 ( 2 13 ( 3 15 ( 2 14 ( 4 15 ( 2 12 ( 2 14 ( 2 12 ( 3 15 ( 2 13 ( 1 12 ( 1 12 ( 2 16 ( 5 13 ( 2 12 ( 3 29 ( 3 14 ( 1 12 ( 3 14 ( 1 13 ( 1 13 ( 2 16 ( 3 12 ( 3 17 ( 4 15 ( 3 15 ( 4 14 ( 3 14 ( 4 13 ( 2 13 ( 1 13 ( 2 13 ( 2 12 ( 2 12 ( 2 15 ( 3 13 ( 1 13 ( 1 15 ( 2 15 ( 3 22 ( 4 19 ( 1 12 ( 2 15 ( 4 17 ( 4 15 ( 2 15 ( 3 17 ( 2 14 ( 1 13 ( 2 13 ( 1 14 ( 3
nb 5 5 5 5 5 5 5 5 5 5 5 5 5 5 4 5 4 5 5 5 5 5 5 5 5 5 5 5 5 5 4 5 4 4 5 5 5 5 5 5 4 5 5 4 5 5 5 4 5 5 5 5
E° (V)c Li+/Li
-3.04, -3.04, Li+/Li -2.71, Na+/Na -2.71, Na+/Na -2.93, K+/K -2.93, K+/K -2.37, Mg2+/Mg -2.37, Mg2+/Mg -2.87, Ca2+/Ca -2.87, Ca2+/Ca -0.9, Ti3+/Ti2+ -0.26, V3+/V 0.34, VO2+/V3+ 0.96, HV2O73-/HV2O5- d -0.91, Cr2+/Cr -0.41, Cr3+/Cr2+ 1.23, Cr2O72-/Cr3+ -1.19, Mn2+/Mn -1.19, Mn2+/Mn 1.54, Mn3+/Mn2+ 1.22, MnO2/Mn2+ 1.51, MnO4-/Mn2+ -0.45, Fe2+/Fe 0.77, Fe3+/Fe2+ 0.77, Fe3+/Fe2+ -0.28, Co2+/Co -0.28, Co2+/Co -0.26, Ni2+/Ni -0.26, Ni2+/Ni -0.52, Cu+/Cu 0.15, Cu2+/Cu+ 0.15, Cu2+/Cu+ -0.76, Zn2+/Zn -0.76, Zn2+/Zn -0.17, Al3+/Al -0.17, Al3+/Al -0.55, Ga3+/Ga -0.14, Sn2+/Sn 0.15, Sn4+/Sn2+ 0.80, Ag+/Ag 1.59, IO4-/IO3- e 0.15, IO3-/IO2.01, S2O82-/SO421.78, H2O2/H2O 1.05, t-BuO2•/t-BuO2Hf -1.12, SO32-/S2O42-0.33, RSSR/RSHg
a A water control yielded a force of 14 ( 3 mN averaged from 30 samples. b The number of samples run is indicated by n. c Literature values for common reduction reactions for a given starting material. Potentials (vs NHE) were obtained from ref 34 with the exception of those denoted with d, e, f, or g. d Bard, A. J.; Parsons, R.; Jordan, J. Standard Potentials in Aqueous Solution; Marcel Dekker: New York, 1985; p 523. e Bratsch, S. G. J. Phys. Chem. Ref. Data 1989, 18, 1-21. f Das, T. N.; Dhanasekaran, T.; Alfassi, Z. B.; Netta, P. J. Phys. Chem. A 1998, 102, 280-284. g Cleland, W. W. Biochemistry 1964, 3, 480-482.
tested five times with any outliers examined for removal using a Q-test at 90% confidence. Examples of full rod extension versus penetration force plots are shown below.
Results and Discussion Metal Ions and Oxidants. The average force required to penetrate each protein extract sample after mixing with a specific reagent solution is shown in Table 1. No difference was observed between the chloride salts and the nitrate salts of a given metal oxidation state. Thus Figures 1-4 represent combined averages of both salts for a given ion (e.g., “Mg2+” for MgCl2 and Mg(NO3)2 averaged). The control experiment in which 100 µL of water was mixed with the gelatinous adhesive precursor
protein produced a penetration force of 14 ( 3 mN. This value provides a baseline for comparison of the various reagents. Figure 1 shows the effect of metals in lower oxidation states at 45 mM on the observed penetration force. The most striking feature is that Mn3+ and Fe3+ are pronounced in the abilities to bring about cross-linking of the extract. The effects of Mn3+ and Fe3+ are a respective 20-fold and 5-fold increase in the observed penetration force above the water control. Of the remaining reagents, Cu2+ and Ag+ are the only other metals with penetration forces that are statistically above the water control, at 2× and 3.5×, respectively. Included in this comparison are Cr3+, Al3+, and Ga3+. Despite the fact that these ions have the same
Protein Precursor of Marine Mussel Adhesives
Figure 1. Average penetration forces for mussel adhesive extract in the presence of metal salts with oxidation states of 3+ or lower.
Figure 2. Average penetration forces for mussel adhesive extract in the presence of oxidative reagents. The inset expands select reagents.
Figure 3. Average penetration forces vs standard reduction potential for various reagents reacted with mussel adhesive precursor. Select reagents are labeled.
oxidation state and to some degree coordination chemistry, no increase in penetration force was found. These data imply that curing of the protein precursor in the presence of metal ions is not controlled simply by metal-DOPA coordination chemistry. Prompted by the oxidative properties of both Fe3+ and Mn3+, we examined the ability of various oxidants to cure the DOPA-containing adhesive extract. The hardening effects of several transition metal salts with oxidative properties as well as nonmetallic oxidants are shown in Figure 2. With two exceptions, all oxidants tested induced some degree of cross-linking. Most notable is the curing brought about by Na3VO4 and Na2Cr2O7. These highly oxidized ions yielded a resistance to penetration of 39× and 45× greater than the water control, respectively. Manganese in a high oxidation state (i.e., Mn7+ of KMnO4)
Langmuir, Vol. 20, No. 9, 2004 3727
Figure 4. Average penetration forces for mussel adhesive extract in the presence of group I ions, group II ions, and acids.
yielded an average penetration force almost 30× greater than the water control. The penetration forces of Sn4+ and Ag+ were 48 and 27 mN, respectively. Both the iodate and the periodate ions induced cross-linking, with the IO4- penetration force exceeding that of Fe3+ (IO3- ) 31 mN, IO4- ) 73 mN). Furthermore, nonmetallic oxidants such as Na2S2O8, H2O2, and tert-butyl hydroperoxide (BHP) induced small, but statistically significant, increases in penetration force relative to the water control. Two oxidizing metal ions did not induce cross-linking of the protein matrix, Mn4+ from MnO2 and V4+ from VOSO4. Under these conditions, MnO2 was insoluble, thereby providing a likely explanation for the lack of reactivity. Although the VOSO4 result may appear unexpected, solution speciation of vanadium is quite complex30-33 and makes predictions difficult. Two reductant controls, sodium dithionite (Na2S2O4) and dithiothreitol (DTT), were included for comparison and show no influence on the penetration force. The inset in Figure 2 provides an expanded graph of all oxidants except the vanadium, chromium, and manganese reagents. To provide a context for oxidative curing of the mussel adhesive extract, common reduction potentials34 are presented in Table 1. The literature values provided were obtained under standardized conditions. Our reaction conditions differ greatly given that our system combines reagent solutions with a gelatinous extract. Nevertheless, a simplified view of reduction potentials proves to be useful in considering these results. Scrutiny of the data presented in Table 1 and Figures 1 and 2 reveals that all reagents that induced cross-linking of the protein adhesive precursor have a positive E° versus the normal hydrogen electrode (NHE). This correlation implies that oxidation of the adhesive precursor is critical for curing. Figure 3 shows a plot of standard electrode potential,34 as listed in Table 1, versus penetration force for select reagents examined. Interestingly, the magnitude of the penetration force increases does not appear to correlate with reduction potential. For example, KMnO4 and NaIO4 have nearly identical reduction potentials, but KMnO4 brings about significantly more curing. Thus simple oxidation of protein-bound DOPA residues (-0.80 V)35 is (30) Heath, E.; Howarth, O. W. J. Chem. Soc., Dalton Trans. 1981, 1105-1110. (31) Pettersson, L.; Hedman, B.; Andersson, I.; Ingri, N. Chem. Scr. 1983, 22, 254-264. (32) Crans, D. C.; Rithner, C. D.; Theisen, L. A. J. Am. Chem. Soc. 1990, 112, 2901-2908. (33) Schmidt, H.; Andresson, I.; Rehder, D.; Pettersson, L. Chem.s Eur. J. 2001, 7, 251-257. (34) CRC Handbook of Chemistry and Physics, 73rd ed.; Lide, D. R., Ed.; CRC Press: Boca Raton, FL, 1992. (35) Brun, A.; Rosset, R. Electroanal. Chem. Interfacial Electrochem. 1974, 49, 287-300.
3728
Langmuir, Vol. 20, No. 9, 2004
Monahan and Wilker
Figure 6. Concentration dependence of Fe(NO3)3 and Zn(NO3)2 upon penetration force. (A) Each Fe(NO3)3 trace is the average of five runs. (B) Each Zn(NO3)2 trace is the average of four runs.
Figure 5. Average penetration forces for mussel adhesive extract vs pH of the reagent solution. Select reagents are labeled.
not sufficient for maximal protein cross-linking. Indeed, the oxidative metals generally induced higher penetration forces when compared to nonmetallic oxidants or nonoxidizing metal ions. We conclude that the ideal reagent for cross-linking the adhesive proteins is an oxidizing metal ion. The lack of a direct correlation between curing and reagent reduction potential is particularly interesting in light of recent findings in which adhesive protein aggregation correlated with oxidation potential.36 This earlier work focused upon the solution behavior of a purified protein in the presence of a nonmetallic oxidant (I3-/I-). Our current study differs by exploring the curing behavior of a bulk material with metallic and nonmetallic cross-linkers. The kinetics and products of oxidationreduction reactions will differ greatly between a buffered solution36 and a gelatinous pellet, where protein motion is hindered significantly. To round out our studies of commonly available reagents, we investigated the effect of group I and group II salts. As seen in Figure 4, the alkali and alkaline-earth metal ions had little or no effect on the penetration force when compared to the water control. The one exception was Ca2+, with an average of 21 mN. The calcium ion is neither an oxidant nor a particularly strong chelator of DOPA or other catechol derivatives.37 The observed increase in hardening induced by this ion may be a result of calcium binding to proteins.38 pH Effect. Previous work with purified adhesive precursor protein employed storage in acid (e.g., 5% acetic acid) to inhibit cross-linking.29 As mentioned, the reagent solutions used in the current study were prepared in deionized water. Consequently the pH varied between solutions. For example, a 500 mM Fe(NO3)3 solution was at pH ≈ 1.5 whereas 500 mM CaCl2 was pH ≈ 10. Such pH variations are necessary to maintain solubility of these metal salts. Thus our study also included the use of several acids, shown in Figure 4. No acid demonstrated curing of the extract. Figure 5 plots the pH of reagent solutions versus observed penetration force. No significant correlation was found between solution pH and curing ability. The recent study on solution aggregation of adhesive proteins also showed no influence of pH upon crosslinking.36 Enzymes and Ultraviolet Light. Oxidizing enzymes such as mushroom tyrosinase have been shown to crosslink purified mussel adhesive precursor proteins.39-41 (36) Haemers, S.; Koper, G. J. M.; Ferns, G. Biomacromolecules 2003, 4, 632-640. (37) Tyson, C. A.; Martell, A. E. J. Am. Chem. Soc. 1968, 90, 33793386. (38) Pidcock, E.; Moore, G. R. J. Biol. Inorg. Chem. 2001, 6, 479489.
Thus, we examined the ability of tyrosinase and laccase (Sigma-Aldrich) to bring about curing of the gelatinous protein extract into a hardened material. The penetration experiment was repeated using a stock solution of 1 mg/ mL for each enzyme. Dilution tests were conducted at 0.1 and 0.01 mg/mL, and reaction time was 1 h. No appreciable curing was observed for tyrosinase or laccase under the conditions examined here. However, the gelatinous nature of these samples may not allow for proper enzymatic processing of the proteins contained within. Ultraviolet light may also induce cross-linking via radical-induced oxidative coupling.42 Curing tests with UV light were conducted by exposing 1 g samples of protein extract to a 300 W xenon arc light source (Oriel model 5690) focused through a fiber optic cable. Illumination time was 1 h. Again, no detectable curing was observed under these conditions. Concentration Dependencies. A concentration dependence study was conducted on all reagents (Table 1). For those reagents that showed increased penetration force, the curing generally proceeded in a concentrationdependent manner, with higher concentrations yielding enhanced cross-linking. Figure 6 shows the concentration dependence of penetration force versus rod extension for Fe(NO3)3 and Zn(NO3)2, two metals concentrated in the mussel byssus. As the rod penetrates the sample, it encounters a restive force from the protein sample. For Fe3+, this resistance increased with increasing Fe(NO3)3 concentration, indicating that more Fe3+ induced more cross-linking. Substantial curing was found at 4.5-45 mM Fe(NO3)3. Interestingly, the concentration of iron in mussel plaques,11,12 if converted to a solution value, is in a similar range at ∼30 mM. The degree of concentration dependence varied among the reagents. That found for Fe(NO3)3 (66, 33, and 17 mN at 45, 4.5, and 0.45 mM, respectively) contrasts with that of Mn(OAc)3 (274, 200, and 14 mN); however, the manganese salt was not completely soluble. The oxo compounds Na3VO4, Na2Cr2O7, and KMnO4 showed more pronounced concentration dependencies than Fe(NO3)3. For example, the forces measured for Na2Cr2O7 at 45, 4.5, and 0.45 mM were 784, 100, and 29 mN, respectively. This difference in concentration dependency, relative to Fe(NO3)3, may arise from the greater range of forces found for the oxo species. At 45 mM, the penetration force found for Na2Cr2O7 was more than 10 times greater than that of Fe(NO3)3. The only exception to the trend of higher concentrations bringing about greater curing was NaIO3. For this reagent, 4.5 mM induced more curing than 45 mM. As noted in the Experimental Section, NaIO3 (39) Filpula, D. R.; Lee, S.-M.; Link, R. P.; Strausberg, S. L.; Strausberg, R. L. Biotechnol. Prog. 1990, 6, 171-177. (40) Kitamura, M.; Kawakami, K.; Nakamura, N.; Tsumoto, K.; Uchiyama, H.; Ueda, Y.; Kumagai, I.; Nakaya, T. J. Polym. Sci. 1999, 37A, 729-736. (41) Burzio, L. A.; Burzio, V. A.; Pardo, J.; Burzio, L. O. Comp. Biochem. Physiol. 2000, 126B, 383-389. (42) Pappas, S. P. Radiation Curing: Science and Technology; Plenum Press: New York, 1992.
Protein Precursor of Marine Mussel Adhesives
was not completely soluble at the highest concentration; however, a 10× dilution was soluble. For non-cross-linking samples such as Zn2+, the resistance of the material remained constant, regardless of Zn(NO3)2 concentration. Visual Observations. After each sample reacted for 1 h and was subjected to the penetration test, a visual inspection was conducted. The color and consistency of each sample were noted qualitatively. For most reagents, little or no color change was observed and the protein paste remained a pale beige color. Furthermore, the texture of the majority of samples tested remained the same, flowing easily around the penetration rod or a spatula. For several colored metal salts, light coloration characteristic of the salt solution was detected after mixing. For example, a 500 mM Co(NO3)2 solution is pink; thus mixing the gelatinous protein paste with Co(NO3)3 brought a faint pink hue to the extract. However, distinct color changes were observed for the following reagents: Mn(OOCH3)3, FeCl3, Fe(NO3)3, Na3VO4, NaIO4, NaIO3, Na2Cr2O7, and KMnO4. Each of these solutions resulted in a dark protein sample after 1 h, significantly darker than from simple addition of a colored solution. Furthermore, each of these samples induced a visible thickening verified by the penetration data (Table 1). Specifically, the Fe3+ solutions thickened and darkened to a brown color. Colorless Na3VO4 produced a dark brown paste. Colorless NaIO4 and NaIO3 yielded a dark paste with a reddish hue. Orange Na2Cr2O7 thickened the protein immediately upon contact and produced a greenish paste. After compression, the Na2Cr2O7 solution above the extract maintained a yellow color. Purple KMnO4 generated a dark paste consisting of condensed pellets with dark flecks. Once compressed, the KMnO4 solution above the paste contained only clear liquid, a likely result of Mn7+ reduction. The protein matrix was thickened visibly by addition of Mn(OOCH3)3, but a color change was hard to observe due to the dark nature of the original reagent solution. Several samples did not thicken immediately upon mixing but did become more viscous over the course of an hour. This change in consistency was evidenced by these samples retaining an imprint of the penetration rod after the force measurement. These samples included FeCl3,
Langmuir, Vol. 20, No. 9, 2004 3729
Fe(NO3)3, SnCl4, AgNO3, NaIO4, NaIO3, Na2S2O8, BHP, Na3VO4, Mn(OOCH3)3, KMnO4, and Na2Cr2O7. A comparison of this list with recorded penetration forces quickly revealed that these reagents also yielded high force values. Samples with no thickening, including the water control, flow and thus regenerated a homogeneous surface once the penetration rod was removed. Conclusions The penetration test has proven to be a useful method for probing the ability of various reagents to cure the precursor to marine mussel adhesive. This method allowed quantitative comparison of the cross-linking brought about by different compounds. Cross-linking by metal ions is a selective process with some ions inducing curing and others not. Oxidants, even those not commonly available in ocean waters, are capable of cross-linking this adhesive material. We note that the materials method used here to measure curing does not necessarily implicate DOPA residues in formation of the adhesive. However, given that DOPA is essential to mussel adhesive synthesis,1-3,9 an avid chelator of metal ions,14,15,17 and readily susceptible to oxidation,1-3,9 a role for DOPA in this curing appears likely. Spectroscopic data indicate that chelating of metal ions brings together multiple protein strands by formation of metal-DOPA complexes.16 Subsequently, the metal ion oxidizes the bound proteins to provide a second method of protein cross-linking.16 Results presented here show that, of the metals actively sequestered by marine mussels, Mn3+ and Fe3+ appear to be key for curing. The optimal reagents for cross-linking are oxidizing metal ions. This synergistic combination of both chelation and oxidation may be the very reason mussels concentrate manganese and iron in their adhesive plaques. Acknowledgment. J.J.W. is grateful for support provided by an Arnold and Mabel Beckman Foundation Young Investigator Award, a National Science Foundation Faculty Early Career Development (CAREER) Award, an Alfred P. Sloan Foundation Research Fellowship, and the Lord Corporation. LA0362728
