Surface properties of mussel adhesive protein component films Marion P. Olivieri Roswell Park Cancer Institute,

Biophysics

Department,

Buffalo,

NY 14263, USA

Robert E. Baier Roswell Park Cancer Institute, Biophysics Department, Biomaterials and NSF Center for Biosurfaces, Buffalo,

Buffalo, NY 14263, USA and Department NY 14214, USA

of

Ronald E. Loomis Oral Biology

Department,

State University

of New York at Buffalo,

Buffalo,

NY 14214, USA

Mussel adhesive protein (MAP) is the adhesive agent used by the blue sea mussel (Mytilus e&/is) to attach the animal to various underwater surfaces. It is composed of 75+85 repeating decameric units with the reported primary sequence NH,-A(l)-K(2)-P(3)-S(4)-Y(5)-Hyp(6)-Hyp(7)T(8)-DOPA(K(lO)-COOH. This study identifies and compares the surface properties of the decameric unit, selected fragments and individual amino acid constituents with the complete MAP preparation. These molecular systems were examined: (a) in the solid state as thin films formed on germanium substrata using multiple-attenuated-internal-reflectance infrared (MAIRIR) spectroscopy, ellipsometry and contact angle analysis; and (b) in the solution state using circular dichroism (CD) spectroscopy. Extensive molecular modelling of the decamer was performed making integral use of the experimentally derived data. These cumulative semiempirical and empirical results suggest a conformation for the decamer that closely associates the L-DOPA and tyrosine residues with the solid substratum. This model provides the first representation of MAP derived from a rational integration of theoretical and experimental data. On the basis of this model, a possible explanation for the bioadhesive properties of MAP is suggested. Keywords:

Bioadhesives,

protein

adsorption,

surface

properties,

mussel

adhesive

protein

Received 28 February 1992; revised 8 May 1992; accepted 3 June 1992

Biological adhesive properties have been examined through diverse studies using bacteria, cell contact phenomena, dental and medical implant materials and marine animal exudates. Bioadhesives from marine environments lend themselves to a wide range of surface interaction studies that may allow the underwater adhesion phenomenon to be more fully explained. The common blue sea mussel (Mytilus edulis) is the source of a proteinaceous adhesive called mussel adhesive protein (MAP), that has the ability to adhere its collagen-like fibrous byssus threads non-specifically to virtually any solid surface’. MAP, used as a coating agent, allows optimal cellular attachment to smooth substrata when present in specific film amounts (Ref. 2 and refs therein). The simultaneous addition of MAP and other proteins actually decreases cellular attachment3V4. The present study identifies and compares the surface properties of the decameric repeating unit5 of MAP and the following selected peptide fragments: NH,-DOPA(S)Lys(lO)-COOH, NH,-Thr(8)-DOPA(9)-Lys(lO)-COOH, Correspondence Biomaterials

to Dr R.E. Loomis.

1992, Vol. 13 No. 14

NH,-Hyp(7)-Thr(8)-DOPA(9)-Lys(lO)-COOH, NH,-Hyp(G)Hyp(7)-Thr(8)-DOPA(9)-Lys[lO)-COOH, NH,-Tyr(5)Hyp(6)-Hyp(7)-Thr(8)-DOPA(9)-Lys(lO)-COOH, NH,-Pro(S)Ser(4)-Tyr(5)-Hyp(6)-Hyp(7)-Thr(8)-DOPA(9)-Lys(lO)COOH, NH,-Lys(Z)-Pro(3)-Ser(Q)-Tyr[5)-Hyp(6)-Hyp(7)Thr(8)-DOPA(9)-Lys[lO)-COOH. Using data from the physicochemical characteristics of the adsorbed peptide films in conjunction with solution state circular dichroism (CD) spectroscopy, a first model of the conformation of the decameric unit at an interface has been developed.

EXPERIMENTAL Materials Mussel adhesive protein and its peptide fragments were obtained from BioPolymers (Farmington, CT, USA). The coded primary sequences of the peptide fragments are as follows: [A) NH,-DOPA-K-COOH, (B) NH,-T-DOPA-KCOOH, (C) NH,-Hyp-T-DOPA-K-COOH, (D) NH,-HypHyp-T-DOPA-K-COOH, [E) NH,-Y-Hyp-Hyp-T-DOPAQ 1992 Butterworth-Heinemann Ltd 0142-9612/92/l 41000-09

Surface

DroDerties

of mussel

adhesive

protein: M.P. Olivieri et al.

1001

All samples were analysed non-destructively by both ellipsomet~ and multiple-attenuated-inte~al-~~ectance infrared (MAIR-IR) spectroscopy. This was done before and after each physical challenge was carried out on the coated germanium plates, allowing for a qualitative and semiquantitative assessment of the physical state of the films under each circumstance. Ellipsometry was carried out on a Rudolf Research Thin Film Type 43702-200E Ellipsometer. Three spots per plate were measured and the raw data processed to give film thickness (A) with a National Bureau of Standards computer program7. Contact angles were measured on a Rame-Hart, Inc. Model 100-00 115 NRL C.A. Goniometer. Data were collected using the ‘slowly advancing contact angle method’. Specifically, droplets of 11 pure diagnostic liquids of known surface tensions were separately placed on the films and the contact angles, 8, measured from two droplets, one on top of the other, for each fluid’. The liquids and surface tension (dyne/cm) were as follows: water (72.4) glycerol (64.8) formamide (58.9) thiodiglycol (53.53, methylene iodide (49.0) I-bromonaphthalene (45.0) I-methylnaphthalene (39.3) dicyclohexyl (32.7) n-hexadecane (27.6) n-tridecane (26.0) and n-decane (23.8). The aforementioned surface tensions were obtained from experiments using a secondary standardization method whereby the diagnostic liquids were analysed on a stable Teflon’& surface’. A computer p~gram (THETA)lO was used to compute the critical surface tension (u,) and the slope of the cosine contact angle versus liquid/vapour surface tension plot”. Further, the dispersive (yd) and polar (v,) components of the composite surface tensions (ys) were calculated, where by definition ys = yp + yd (Ref. 11). Circular dichroism (CD) spectra were obtained for each peptide dissolved in 2X distilled water adjusted to pH 7. CD data were collected on a Jasco J-600 Spectropolarimeter interfaced to an IBM Personal System 2 Model 80 computer and Sekonic X-Y SPL-430 plotter. CD cells were cylindrical with 0.1 cm path length. Jasco software version 1,42 was used to calculate molar ellipticity based on the supplied sample molarity and cell path length. Ten computer-averaged transients were obtained to improve the signal-to-noise ratio. All Evans & Sutherland PS390 monitor interfaced to a Digital Microvax II was used to accomplish the molecular modelling. A commercially available program called SYBYL” served to perform energy minimization calculations and construction of a surface for peptide docking experiments. Amino acid starting conformations in vacua consistent with physiological temperatures were extracted directly from the program’s protein dictionary, BIGPRO. L-DOPA was created as a modification of tyrosine. The decamer was initially built with random torsion angles as each amino acid was connected one to another. On the basis of CD data (see Results and discussion), the torsion angles for poly(L-proline) II (Ref. 13) were used to define the spatial arrangements of the two hydroxyproline residues. Bond lengths, bond angles and non-bonded interactions can be used to define the overall geometry and minimum potential energy of a molecular system or aggregate of atoms. The process whereby linear combinations of the respective potential energy functions of the aforementioned forces are altered such that the lowest energy

K-COOH, (G) NH~-P-S-Y-Hi-Hi-T-DOPA-K-COOH, (H) NH~-K-P-S-Y-gyp-Hyp-T-DOPA-K-COOH, and (I) NH,-A-K-P-S-Y-Hyp-Hyp-T-DOPA-K-COOH. Liquids used for contact angle analysis were the highest grade available, while other chemicals used were standard reagent grade. The optically flat, polished, 50 X 20 X 1 mm germanium SPT45 plates were purchased from Harrick Scientific Corporation (Ossining, NY, USA).

Methods High-pressure liquid chromatography (HPLC) and amino acid analysis were performed on each peptide to assure sample purity using previously described method$. The HPLC was performed on a Waters 6OOE HPLC Powerline Multisolvent Delivery System. Amino acid analysis was accomplished by adding -5.0 nmol of peptide to -0.5 ml 6 N HCl in Pyrex test tubes. The samples were evacuated and purged with N, gas several times, sealed in vacua and then heated for 24 h at 11O’C. The tubes were broken and 100~1 of a 0.2 nmol/pl solution of L-a-amino-Pguanidino-propionic acid (AGP; Sigma Chemical Co., St Louis, MO, USA) was added. The contents were removed and 0.5 ml of distilled water was used to rinse the broken tube of remaining contents. Both solutions were dried together in vacua. Sodium citrate buffer was added (150~1 of 0.2 N, pH 2.2) and centrifuged for 2 min. A 100 ~1 aliquot of this solution was loaded into a Beckman Model 6300 High Performance Amino Analyzer sample chamber. The samples were run and compared to standard amino acid solutions. L-DOPA and hydroxyproline are not present in currently available amino acid standard solutions routinely used for amino acid analysis. It was therefore necessary to prepare standards of these two amino acids at the normal standard concentration (5.0 nmol) to determine relative quantities of these residues in the peptide solutions. AGP standards and amino acid destruction factors assisted in the determination of the relative mass of each residue. To examine the peptide layer adsorbed to the substrate, as well as materials not in direct contact with the surface, ‘leaching’ and ‘rinsing’ procedures were followed for the dried samples. Specifically, 1 mg of each peptide was dissolved in 1 ml of distilled water to make stock solutions from which to work. Peptide solution aliquots (200 ~1) were then placed on to three identical detergentcleaned and baseline-characterized germanium internal reflection plates. The solutions were unperturbed and allowed to air dry for 120 min. After drying, one prism was used for contact angle analysis. The remaining two prisms were distilled water leached and again air dried for 10 min. One of these prisms was then used for contact angle analysis and the remaining prism was rinsed with distilled water and given another 10 min air drying. Leaching was accomplished by gently applying just enough distilled water (-0.5 ml) to form a puddle on the plate surface and allowing it to stand for 10 s before discarding the liquid. Rinsing was performed by directing a vigorous stream of distilled water from a wash bottle on to the plate surface for 10 s. In addition, 100 ~1 aliquots of 0.1 M L-DOPA solutions at various pH (2.98, 4.50 and 8.25; using phosphate buffer and trace HCl for the pH 2.98 sample) were dried down on germanium plates, then leached and rinsed as described earlier. -

Biomaterials

1992, Vol. 13 No. 14

1002

Surface properties of mussel adhesive protein: M.P. Olivieri et al.

required to maintain a stable conformation is achieved is referred to as the ‘force field’ or the ‘molecular mechanics’ method. Details of the mathematics and general principles used to define the various energy functions, as well as the algorithms for the molecular mechanics and energy minimization procedures, are well established and can be found in the package available from Sybyl Molecular Modeling Systems (Tripos Assoc. Inc., St Louis, MO, USA). Briefly, potential energy minimization is performed using combinations of first-order derivative and nonderivative methods contained in the MAXIMIN subroutine” of SYBYL. Initially, the non-derivative procedure ‘SIMPLEX’ is used on an atom-by-atom basis until the maximum force on any atom is below some minimum specified value. The major advantage of SIMPLEX is that it handles highly distorted structures with discontinuous derivatives, thereby avoiding energy functions becoming trapped in one or more of the many possible local minima. MAXIMIN then adjusts all the atomic coordinates simultaneously based on the first derivative of the total energy equation with respect to the number of degrees of freedom. Subsequently, fine tuning of the molecular conformations and potential energy functions is performed by the optimization procedure described by Powell14. The resulting peptide can then be used in docking experiments where another molecule, or in this case a surface, is brought into close contact and the intramolecular interactions and total minimum potential energy are evaluated. This feature of our modelling efforts is unique since both the influences of the surface and empirical data were considered when generating the MAP decameric peptide model. Crystallization of the peptide fragments was attempted using both the ‘hanging-drop’ and simple evaporation methods”. Many concentrations and solution types were used to obtain single crystals of selected peptide fragments: for example, ethanol and methanol at 100,80, 50, 40, 33.3, 16.6 and 0% were used as precipitating agents for simple evaporation of fragments A and I; MPD (2-methyl-2,4-pentanediol, PEG (ethylene glycol) and ammonium sulphate at 10,20,30,40 and 50% were used to form hanging droplets of 21.11 each of solvent and peptide (peptide solution was composed of 50~1 of 5% acetic acid added to 1 mg of fragment I). Precipitation was observed in the 20 and 40% ammonium sulphate samples. Additional crystallization attempts used ammonium sulphate at 15,20,25,30, 35,40,45 and 50%.

RESULTS

AND

DISCUSSION

The mussel adhesive protein decamer and fragments all had hygroscopic tendencies. Further, small amounts of the peptide fragments left in ambient laboratory conditions (23°C 45 relative humidity] turned into brownish tar-like masses. All of these peptides contain 3,4-dihydroxyphenylalanine (L-DOPA) that is capable of self-aggregation, quinone and eumelanin formation and photochemical reactions15. When the samples were brought to a high pH using NaOH (pH > 8), they quickly turned dark brown or purple. Although colour has been associated with L-DOPA-hydroxyproline reactions16, it seems unlikely that this was the sole cause of colour Biomaterials

1992, Vol. 13 No. 14

change in these samples, since the smallest fragment (A: L-DOPA-Lys), free from Hyp, also showed these colour changes. Other investigators (Ref. 1 and refs therein) have postulated that in the intact MAP molecule (-800 residues), L-DOPA acts as a tanning agent, first oxidizing to an o-quinone and then participating in a condensation reaction with a lysine residue. Shorter peptide sequences [e.g. fragment A) lend themselves to increased L-DOPALys interactions due to decreased steric hindrance between different peptide chains. When alkaline solutions were brought back to lower pH (e.g. 2-3) using trace HCl, the liquids remained coloured; therefore, some irreversible changes had taken place. The transfer of lyophilized peptide samples was carried out under inert gas (i.e. nitrogen atmosphere) in a drybox. The presence of amino acid constituents, relative peptide purity and molarity of prepared sample solutions were monitored using amino acid analysis. HPLC was the analytical technique used to test for contamination. MAIR-IR (Figure 2) spectroscopy, ellipsometry (Table 2) and contact angle analyses (Table 2) of L-DOPA films formed at various pH show that all samples were capable of spontaneous adsorption. However, only the solutions in the neutral pH region (e.g. 8.25) were capable of uniform film formation with MAIR-IR spectral peaks similar to that of MAP (i.e. strong spectral absorption at 1260 cm-‘, Ref. 2). This absorption has been associated with the hydroxyl groups attached to aromatic compoundsl’. It has been suggested that one requirement for forming a reliable adhesive is the removal of weak boundary layers from the surface (e.g. water)‘. It is likely that the pair of hydroxyl groups from L-DOPA are responsible for it’s unique ability to displace water from an interface. L-DOPA may be capable of this task when occurring both as the free amino acid as well as in the intact MAP molecule, taking up bonding relationships once used by boundary water molecules. This tenacity in adsorption appears to be greater than most other naturally occurring amino acids and their respective homopol~ers, Preliminary experiments have also indicated that the replacement of DOPA with Tyr yields materials with reduced efficiency in surface adsorption. The crystals obtained were coloured and filamentous and therefore inappropriate for further X-ray diffraction studies. To create stable single crystals, it may be necessary to repeat these studies using tyrosine as a substitute for the unstable L-DOPA. However, since the L-DOPA may be key to the activity of this and other bioadhesives, this substitution may compromise the biological significance. When amino acid analysis was performed on the decamer and its fragments under various experimental conditions, there was occasionally a peak at -27.3 min. Since this unidentified peak was present in the smallest fragment (L-DOPA-Lys), it seemed probable that it originated from L-DOPA (no~ally expected at -21.6 min) or one of its derivatives/aggregates. Since it is known that the ability to isolate the intact protein and its subsequent attachment of various cells to surfaces in vitro are both pH dependent, amino acid analysis was performed on L-DOPA at various pH (Figure 2). The unidentified peak was also present in these samples, possibly due to some irreversible conformational change in the L-DOPA molecule. Alternatively, this peak may be

Surface

properties

of mussel adhesive

protein:

M. P. Olivieri

1003

et al.

80

2000 Wave

1500 number

(cm

-1

1000

500

1

Figure 1 Multiple-attenuated-internal-reflectance infrared spectrum of L-DOPA at pt-i 8.25, after water leaching. Most absorbances located below 900 wave numbers are derived from the germanium lattice.

Table 1 Ellipsometry results from 3,4_dihydroxyphenylalanine (L-DOPA) Sample”

Thickness

1 (pH 2.98)

9.0 0.0 0.0 > 1000.0b 0.0 14.0 0.0 542.0b 109.0 275.0 321.0 61 3.0b

2 (pH 4.50)

3 (pH 8.25)

(A)

aAll samples were dried and then subjected to a water leaching procedure before film thickness was measured using three spots per plate. bWhen the sample films contained ‘patches’ that were not representative of the entire film, a fourth spot was measured in the irregular portion.

Table 2 Contact angle results for 3,4-dihydroxyphenylalanine (L-DOPA)a Sample

YC

Yd

YP

YS

Slopeb

1 2 3 4 5

30.5 30.3 30.9 28.7 29.4

27.9 27.3 27.5 24.3 23.3

16.4 15.4 24.6 27.9 35.5

44.3 42.7 52.1 52.2 58.8

-10 -12 -11 -10 -11

(pH (pH (pH (pH (pH

2.98)c 4.50)c 4.50)d 8.25)c 8.25)d

aCalculations’0. ” of the yc, yd, yp and yS values (dyne/cm) which are the critical surface tension, the dispersive and polar components and the composite surface tension values, respectively, from recorded comprehensive contact angle data. bThe ‘slope’ reported (cm/dyne X 10m3) is for the Zisman plotsof the data. ‘Portions of the film were clear and even (i.e. more regular). dPortions of the film were highlycoloured and obviously not thin and uniform.

subject to other unexplored variables such as time and temperature in addition to pH. This could explain the apparent differences in amino acid composition found when examining other proteins containing L-DOPA”. It may be possible to use more extensive analyses to

determine alterations in related aromatic compounds (e.g. those capable of quinone formation) to provide a qualitative and semiquantitative assessment of conformational changes of L-DOPA. Combined infrared and ellipsometric data indicate that intact MAP molecules bind strongly as rinseresistant thin films to clean germanium plates, which serve as surrogate substrata for those surfaces found in many natural situations. These ‘conditioning films’ become thicker with increasing applied concentration, which is typical of most proteins (Ref. 2 and references therein). This effectively masks the physicochemical properties of the original solid surfaces, while expressing new adhesive properties for cell attachment that must reflect the exposed functional groups and surfacelocalized configuration of the protein. Collectively, the contact angle data indicate MAP and its peptide fragments form films with critical surface tensions with the 30+4O dyne/cm range found for most proteins (Ref. 26; see Table 3). MAIR-IR spectra indicate that all of the fragments are capable of forming rinseresistant films (Figure 3) with absorption similarities to MAP’. Also, ellipsometry da .a reveal that the smaller peptides form films that, attough adherent, are very uneven (Table 4). Thus, the inc usion of tyrosine adjacent to the two consecutive hydrr cyproline residues seems crucial to the resultant films becoming more uniform and rinse resistant. Studies done by other investigators examining related compound? revealed similar CD spectra (positive absorption at 230 nm) that were attributed to random coil conformations for the peptides and proteins examined. Furthermore, addition of 6 M guanidine (a known chaotrope capable of hydrogen-bond disruption) produced no apparent changes in the CD spectra. Prior data revealed that the MAP proteins probably do not contain a-helicies (intrachain hydrogen bonds) or /3sheets (interchain hydrogen bonds) in aqueous solution. However, there is no information related to other possible types of MAP secondary structure(s) that are not dependent on hydrogen bonding (Ply-turns involving Biomaterials

1992, Vol. 13 No. 14

1004

Surface properties of mussel adhesive protein: Ml? Otivieri I.

i

Table3 Contact angle results for mussel adhesive protein fragments Peptidea

Film conditionb

ycc

yd

yp

yS

md

A

Dried Leached Rinsed Dried Leached Rinsed Dried Leached Rinsed Dried Leached

31 41 36 30 38

34 24 30 29 27 36 27 25 24 23 28

18 37 20 26 24 17

52 61 50 55 51 53

zs 33 34 25

z 57 57 53

-12 -19 -13 -12 -9 -15 -7 -12 -9 -9 -11

B

a

C D

b E G

C

H

I

d

et al.

23: 41e 41* 33 33

Rinsed

33

28

27

53

-4’

Dried Leached

29 27

27 25

17 27

54 52

-30 -10

Rinsed

33

28

26

Dried

31

27

25

Leached Rinsed Dried Leached Rinsed Dried Leached Rinsed

32 36 30 30 33 31 32 37

29 26 30 26 28 27 30 23

23 29 22 29 21 25 22 31

54 52 52 55 52 54 49 52 52 54

-10 -13 -8 -12 -13 -9 -9 -13 -12 -10

aThe amino acid sequences of the peptide samples are given in the text. Also, the films formed from distilled water solutions. bThe drying, leaching and rinsing procedures are described in the text. CCalculated’o~ ” yc, yd, yP and ys values (dyne/cm). dThe ‘slope’ (m) (cm/dyne X 10e3) from the Zisman plots of the data. eAnomatously high yC values reflect nearly complete elution of the peptide from the surface, exposing high-energy metallic substratum. ‘Note that exceptionally flat Z&man plot slope is characteristicof very clean metal& surfaces, generally having only bound water layers.

Fiiure2 Amino acid analysis: elution times of L-DOPA at various pH: a, 2.90; b, 3.55; c, 6.24, d, 7.48; e, 8.25. The peak at 21.6 min is the expected value for L-DOPA alone; longer times suggest ionic interaction with the column packing leading to slower passage (tighter binding). Peaks found at 36.7 and 38.5 are ammonia and AGP (L-a-amino-fl-guanidino-propiomic acid), respectively.

proline and poly(L-proline) types I and II). Indeed, the positive spectral absorption at 230 nm observed here has been associated with pol~L-proline) II-like structures. This conformation has previously been noted in other small peptide systems, including those with only two consecutive proline residues (Ref. 6 and refs therein). The absorbance demonstrated by MAP fragments including hydroxyproline may thus be related to a poly(L-proline) II-like structure (Figure 4). That there is no hydrogen bonding required for this ~onfo~ation, would also explain why the chaot~pe did not alter the spectra obtained by Williams et al.2’. Although the CD data collected were of peptides in solution state, others have found that the local chain conformations in poly(L-proline) II in solution state are typically the same in the solid state2’. Further studies using these peptides in the adsorbed state need to be completed to discern their molecular orientations at interfaces. Biomaterials

1992, Vol. 13 No. 14

The observed unique adhesiveness of many different cell types to these films probably depends on the outermost expression of polar side-chain arrays of the film’, 3. As the molecular area at the interface changes with concentration, it is plausible that the orientation and/or conformation of the molecules depends on the packing density of the surfacez3. The adsorption of aromatic compounds at low surface concentrations has been attributed to the interaction between n-electrons and the surfacez3. Also, the solid state results obtained in the present study show %&dihydroxyphenylalanine (LDOPA) is independently capable of adsorption and retention to solid surfaces. Our energy-minimized model (Figure 5a) demonstrates how the initial adsorption of the fragments on to the interface may take place. It is likely that at higher concentrations of peptides at an interface, the molecules (L-DOPA] lose their ability to orient parallel to the substrata surface and therefore interact in a more ‘upright’ position (Figure 5b). Although the larger peptides [fragment I) are of lower molecular concentrations than the smaller peptides (fragment A}, the smaller fragments may form ~OPA-DOPA and DOPA-Lys aggregates more readily to give thicker ‘patchier’ films (Table 4). Additionally, in the larger peptides and perhaps MAP, once an L-DOPA residue has been adsorbed, other residues brought close to the interface may decrease the surface sites available for other L-DOPAs to attach. That MAP is most active as a cellular attachment adjacent at specific lower concentrations, and less able to attach cells at higher concentrations,

Surface properties of mussel adhesive protein: MP.

1005

Oliviefi et a/.

A

1500

2000 Wave

number

(cm-’

1

Figure 3 Multiple-a~enuated-internal-reflectance infrared spectra of fragment I: a, before leaching, b, after leaching, c, after water rinsing (with decreasing absorption peak intensities respectively}. In the rinsed fiim (lOOoc 8 >lOOO

37

11

E

>lOOO >lOOO >I000 >lOOO >iooo >lOOO >iooo >lOOO >I000 10 119 >lOOO >lOOO 56 23 >lOOO >lOOO >lOOO 1090 >lOOO 352 582 512 487 420 319 683 838 >lOOO 2008 514 694 >I000 >lOOO 990 351

347 398 430 362 272 291

23

316 266 258 13 12 20

10 9 47

324 209 23 198 32 30

14 16 21

434 408 541 116 196 107

14 16 21

B

c

D

T >1000 436 >lOOO > 1000 20 22 24 0 >lOOO >lOOO 45 >lOOO >lOOO >lOOO 4 >lOOO >lOOO 353 7 >lOOO 927 >lOOO >lOOO >lOOO >I000 >lOOO >lOOO .lOOO >lOOO >lOOO >lOOO

d

i’: 19

16 10

9 9

zi 32

19 18 20

G

11

H

:: 14

480 8 45 0 0 4

706 169 : 12 17

:

I

33;

aThe amino acid sequences of the peptide samples are given in the text. bThe thicknesses are reported in A at a refractive index of 1.5. ‘The actual thickness is greater than the stated amount since values over 1000 A werepresent but werenot further quantified. dlncreased average thicknesses are found in some spots after leaching of distilled water because of redistribution of material and its more uniform adsorption during the leaching period. Note, however, thegeneral increase in film unifarmi~ after excess unadsorbed material(usually in visible patches) was eluted away. eMAIR-lR spectral analyses revealed weak protein absorption bands for these rinsed films. Thus, the films are unevenly distributed on the surfacesor ‘patchy’.

Biomaterials

1992, Vol. 13 No. 14

Surface

1006

-2.0

1

I

properties

I

I

I

Figure 4

protein:

et al.

I

I

I

IMP. Olivieri

230

200

Wavelength

follows:

of mussel adhesive

Circular dichroism spectra of the mussel adhesive 0, A; 0, B; A, C; 0, D; 4, E; A, G; 0, H; +, I.

probably reflects this dependence on substratum surface availability. It has been suggested that such concentrationdependent mechanisms of adhesion are of particular importance at low concentrations, especially when the attaching molecules (L-DOPA) are free to ‘lie down’23. Self-aggregation of the L-DOPA away from the substratum surface allows the films to form thicker layers that also express polar groups outermost (Tables 1 and 2) and are conducive to cellular adhesion. The computer modelling experiments, in conjunction with the ellipsometry data from Tables 3 and4, show that as the peptide size increases from that of fragment A to that of fragment D (Hyp-Hyp-Thr-L-DOPA-Lys), the hydroxyproline residues are critical to presenting the molecule in a conformation that allows the L-DOPA to first dehydrate and then n-bond with the surface. When tyrosine is added [fragment E), the entire molecule then rotates. This brings the Tyr side-chain closer to the interface, thus providing secondary film stabilization with L-DOPA as the primary binding unit as indicated by

fragments

(nm)

(A-E, G-l) in 2X distilled

water,

pH 7. The symbols

are as

the empirical data from Figures 2-3. This results (Tables 1 and 2) in the molecule exposing the polar lysine, serine, hydroxyproline and threonine side-chains outermost in an array and abundance that prompts general bioadhesion’. The proposed models illustrated in Figures 5a,b, where the CD data from Figure 4 are considered, give internal energy values of 27.2 kcal/mol after energy minimization. To test the plausibility of this value, coordinatesz4 for a known poly(L-proline) type II conformation occurring in the N-terminal nine amino acids region of avian pancreatic polypeptidez5 were obtained from the Protein Data Bankz6’ ” at Brookhaven National Laboratory. This model yielded 36.0 kcal/mol of internal energy for the known crystal structure after being subjected to the same energy minimization procedures as the MAP decameric unit. The lower calculated value for MAP attests to the reasonableness of the molecular model presented. CONCLUSIONS Since MAP films are capable of non-specifically attaching a wide variety of cell types (e.g. bovine (cornea11 and canine (vascular) endothelia, hamster kidney fibroblasts

a

b

Figure 5

a, The decameric repeating unit of mussel adhesive protein at an interface. rigid structure around its longitudinal axis also approximates the tyrosine residue stabilization (however, L-DOPA is still the primary binding unit), while maintaining threonine residues as the outermost functional moieties. Biomaterials

1992. Vol. 13 No. 14

b, Note how a simple rotation to the surface providing the polar lysine, serine,

of the relatively for secondary film hydroxyproline and

Surface

properties

of mussel adhesive

protein:

M. P. Olivieri

(BHK-211, human

histocytic lymphoma (U-9371, mouse (thymic) and rabbit [retinal pigmented) epithelia, rat brain neurones, mouse thymocytes, chick embryo chondrocytes, human red blood cells) to a variety of substrata (plastics and glass] (Ref. 2 and refs therein), it is unlikely that the cellular attachment depends on chemically or biologically specific functional moieties. The molecule does contain 20% lysine and it is probable that these positively charged polar groups remain outermost and are responsible for much of the non-specific cellular attachment (e.g. poly(L-lysine) coatings often applied to culture ware for just this purpose). Our molecular model illustrates how MAP combines the dehydration properties of L-DOPA for initial adsorption, with the sequential hydroxyprolines providing conformational rigidity that presents the polar side-chains of the remaining residues as the outermost region. Our continuing experiments have two practical goals which address the functional regulation of protein adhesion. The first goal is to learn how to achieve optimal biological attachment (as required for biotechnology applications such as ‘fixed-film’ processing). The second goal is to develop adhesive agents that can be mass-produced inexpensively. Currently, this process involves either costly and time-consuming extraction of MAP from mussels, or molecular cloning coupled with post-transcriptional modifications”. This work has led to a plausible first structural model for the repeating decameric peptide unit of MAP at an interface. Other studies are underway that will provide additional information that should lead to refinements of the present model. For example, it is possible that the pentapeptide unit Y-HYP-HYP-T-DOPA, rich in waterdisplacing hydroxyls and bond-stabilizing rings, is the interface ‘conditioning’ region of MAP, as suggested by the more surface-localized molecular structure illustrated in Figure 5b. Our future studies will evaluate this possibility, as well as other biophysical, functional and conformational features of MAP. ACKNOWLEDGMENTS We thank Mrs Pamela M. Loomis for the preparation of the manuscript figures, BioPolymers, Inc. for the preparation of the MAP peptide fragments, and Drs Krishna Bhandary and Narayanan Ramasubbu for useful conversations regarding this work. Supported by GSA-Mark Diamond Research Fund [SUNY at Buffalo) and LJSPHS Grants DE07760, DE00229 and DE09580. REFERENCES 1

2

3

Waite, J.H., Nature’s underwater adhesive specialist, Int. I. Adhesion Adhesiv. 1987, 7, 9-14 Olivieri, M.P., Loomis, R.E., Meyer, A.E. and Baier, R.E., Surface characterization of mussel adhesive protein films, I. Adhesion Sci. Technol. 1990, 4, 197-204 Olivieri, M.P., Baier, R.E. and Loomis, R.E., Quenching of surface polarity as a mechanism to reduce cell adhesion, Combined Session of the 22nd International Biomaterials Symposium and the 16th Annual Meeting of the Society for Bioma terials, Charleston, SC, USA, 1990, Abst. No. 78

et al.

1007

4

Olivieri, M.P., Tweden, KS, Rittle, K.H. and Loomis, R.E., A comparative biophysical study of adsorbed calf serum, fetal bovine serum and mussel adhesive protein films, Biomaterials in press Waite, J.H. and Tanzer, M.L., Polyphenolic substances of Mytilus edulis: novel adhesive containing L-Dopa and hydroxyproline, Science 1981, 212, 1038-1040 Loomis, R.E., Gonzalez, M. and Loomis, P.M., Investigation of cis/trans proline isomerism in a multiply occurring peptide fragment from human salivary proline-rich glycoprotein, Int. J. Pept. Protein Res. 1991,38,428-439 McCrackin, F.L. and Colson, J.P., ELIP: a fortran program for analyses of ellipsometer measurements and calculation of reflection coefficients from thin films, National Bureau of Standards (Washington, DC] Misc. Pub. No. 256, 1964 Zisman, W.A., Relation of the equilibrium contact angle to liquid and solid constitution, Adv. Chem. Ser. 1964,43, 1-51 Baier, R.E. and Meyer, A.E., Surface analysis, in Handbook of Biomaterials Evaluation: Scientific, Technical, and Clinical Testing of Implant Materials (Ed. A.F. von Recum), Macmillan, New York, USA, 1986, pp 97-108 Mazurowski, M.J. and Baier, R.E., THETA: a computer program for reducing and plotting contact angle data, Cornell Aeronautical Laboratory (Buffalo, NY, USA] Report No. 185, 1970 Kaelble, D.H., Dispersion-polar surface tension properties of organic solids, 1. Adhesion 1970, 2, 66-81 Sybyl Molecular Modeling System, Version 5.4. Tripos Associates, St Louis, MO, USA, 1991 Cantor, CR. and Schimmel, P.R., Conformational analysis and forces that determine protein structure, in Biophysical Chemistry Part 1: The Conformation of Biological Macromolecules, W.H. Freeman, San Francisco, CA, USA, 1980, p 257 Powell, M.J.D., Restart procedures for conjugate gradient method, Math. Prog. 1977, 12, 241-254 Waite, J.H., The phylogeny and chemical diversity of quinone-tanned glues and varnishes, Comp. Biochem. Physiol. B. 1990, 97, 19-29 Mason, H.S. and Peterson, E.W.. The reaction of quinones with protamine and nucleoprotamine N-terminal proline, 1. Biol. Chem. 1954, 212, 485-493 Cook, B.W. and Jones, K., in A Programmed Introduction to Infrared Spectroscopy Heyden & Son, London, UK, 1970 McPherson Jr, A., Crystallization of proteins by variation of pH or temperature, in Methods in Enzymology: Diffraction Methods for Biological Macromolecules, Part A, Vol. 114, (Eds H.W. Wckoff, C.H.W. Hirs and S.N. Timashell), Academic, New York, USA, 1985, pp 125-127 Waite, J.H., Housley, T.J. and Tanzer, M.L., Peptide repeats in a mussel glue protein: theme and variation, Biochemistry 1985, 24, 5010-5014 Baier, R.E., Meyer, A.E., Natiella, J.R., Natiella, R.R. and Carter, J.M., Surface properties determine bioadhesive outcomes: methods and results, 1. Biomed. Mater. Res. 1984, 18, 337-355 Williams, T.J., Marumo, K., Waite, J.H. and Henkens, R.W., Mussel glue protein has an open conformation, Arch. Biochem. Biophys. 1989, 269, 415-422 Tiffany, M.L. and Krimm, S., Circular dichroism of polyL-proline in an unordered conformation, Biopolymers 1968, 6, 1767-1770 Arwin, H., Lundstrom, I., Arielly, S. and Claeson, G., Orientation of a tripeptide on platinum, Langmuir 1990, 6, 1551-1557

5

6

7

a

9

10

11 12 13

14 15

16

17

18

19

20

21

22

23

Biomaterials

1992. Vol. 13 No. 14

1008 24 25

26

Surface properties of mussel adhesive protein: M.P. O/jvieri et al. Entry lPPT, File 340.DECK, version of October 29, 1987 Blundell, T., Pitts, J.E., Tickle, LJ. and Wood, S.P., X-ray analysis (1.4 A-resolution) of avian pancreatic polypeptide: small globular hormones, Proc, Natl. Acad. Sci. USA 1981, 78, 4175-4179 Bernstein, F.C., Koetzle, T.F., Williams,

G.J.B., Meyer, Jr, E.F., Brice, M.D., Rodgers, J.R., Kennard, O., Shimanouchi, T. and Tasumi, M., The protein data bank: a computerbased archival file for macromolecular structures, 1. Mol. Biol. 1977, 112, 535-542

We offer

a reprints

service

27

28

in respect

Abola,

E.E., Bernstein, F.C., Bryant, S.H., Koetzle, T.F. and Wing, J., in Crystallographic database - Information Content, Software Systems, Scientific Applications (Eds F.H. Allen, G. Bergerhoff and R. Sievers), Data Commission of the International Union of Crystallography, Bonn, 1987, pp 107-132 Filpula, D.R., Lee, S.M., Link, R.P., Strausberg, S.L. and Strausberg, R.L., Structural and functional repetition in a marine mussel adhesive protein, Biotechnol. Proc. 1990, 6,171-177

of all articles

in this journal.

Companies: an article featuring your product or a new application, provides its readers with an independent authoritative comment. Reprints can be used by you to inform and educate your customers and staff about scientific or technical advances in subjects relevant to your business. They are a cost-effective, ethical way of promoting your company and its products. For a quotation

on your reprint requirements,

The Reprints Department, Butterworth-Heinemann Jordan Hill, Oxford, OX2 8Dl? Tel: (0865) 310366.

Biomaterials

1992, Vol. 13 No. 14

please

contact:

Ltd., Linacre House, Fax: (0865) 311270