lipopeptides
Cavalli, S.
Citation
Cavalli, S. (2007, January 25). Chemistry, structural insight and applications of β-sheet forming lipopeptides. Retrieved from https://hdl.handle.net/1887/9452
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/9452
Note: To cite this publication please use the final published version (if applicable).
Self-Organizing β-Sheet Lipopeptide Monolayers
as Templates for the Mineralization of CaCO
3*Abstract. Amphiphilic peptides comprising alternating hydrophobic (Leu) and hydrophilic (Glu) residues have been shown to form ordered monolayers with a β-pleated sheet structure at the air-water interface in which the carboxylate side groups are pointing into the water (Chapter 2). Here, these assemblies were investigated as well-defined two-dimensional templates for the mineralization of calcium carbonate. Monolayers of DOPE, Ac-(Leu-Glu)4-NH2, tetra-ALP, hexa-ALP and octa-ALP were compared in their ability to act as templates. The conjugation of the phospholipid moiety to the octapeptide (Leu-Glu)4 not only enhanced the amphiphilic behavior of the molecule but also increased the flexibility of the monolayer, without compromising the β-sheet structure. This led to a distinct change in the templating behavior. The difference in the ability of the templates to adapt to the developing mineral phase is discussed.
* Part of this work has been published: S. Cavalli†, D. C. Popescu†, E. E. Tellers, M. R. J. Vos, B. P. Pichon, M.
Overhand, H. Rapaport, N. A. J. M. Sommerdijk, A. Kros Angew. Chem. Int. Ed. 2006, 45, 739-744 († both authors contributed equally).
3.1 Introduction
3.1.1 Calcium Carbonate (CaCO3)
Calcium carbonate (CaCO3) is one of the most abundant biomaterials and has therefore received considerable interest in several areas of material research.1 In particular, CaCO3 has emerged as a promising material for bone replacement applications.2 It exists in a variety of polymorphs, of which calcite and aragonite (with similar thermodynamic stabilities) are the most commonly encountered in biological samples.
Figure 3.1. Typical morphologies of (A) calcite, (B) aragonite and (C) vaterite. (D) Cystolith from the leaves of Ficus Microcarpa, which are entirely composed of ACC.
At room temperature and atmospheric pressure, calcite (Figure 3.1, A) is the thermodynamically most stable form of CaCO3. The conventional morphology of calcite is rhombohedral3 and the crystal structure contains layers of Ca2+ ions arranged in deformed cubic closed packing (see Appendix 2). Aragonite (Figure 3.1, B) exists in an orthorhombic space group (see Appendix 2) and shows needle-shape morphology. At ambient temperature and pressure, aragonite is metastable. However, the presence of specific proteins in the crystallization medium can stabilize this polymorph also at room temperature.4 Vaterite (Figure 3.1, C) crystallizes in a trigonal space group (see Appendix 2) and forms spherical- or floret-shaped aggregates.5 Vaterite is extremely rare in nature because is a metastable kinetic product, which rapidly transforms into calcite or aragonite in aqueous solution. Finally, amorphous calcium carbonate (ACC) is an exceptionally unstable transient precursor of more stable crystalline calcite and aragonite. This amorphous form appears to be stabilized through adsorption of biological macromolecules such as polysaccharide on solid surfaces.6 An example of ACC is given by cystoliths, inorganic concretions constituted by irregular-shaped objects, a few tens of micrometers in length, entirely composed of ACC, which are found in the epidermis of leaves from plants of various families (e.g. cystolith from the leaves of Ficus Microcarpa as shown in Figure 3.1, D).7
A B C D
3.1.2 Mineralization of CaCO3
The large variation in inorganic structures encountered in biological systems, often with exquisite and unique morphologies, have been fascinating and inspiring researchers in all scientific disciplines for ages.8 For materials scientists the understanding of the principles underlying the process of biomineralization holds great promise for the design and synthesis of new inorganic and hybrid structures with yet unrealized properties. The central concept in biomineralization research is that organized biomacromolecules, containing well-defined arrays of functional groups, control polymorph selection and oriented nucleation of crystals through lowering the nucleation energy of specific crystal faces.9 Control over crystal morphology is then exerted by the interaction of biomolecules in solution with specific crystal planes during growth. Recently it has become clear that the classical concept which proposes that the template, containing well-defined arrays of functional groups, controls the nucleation of the inorganic crystals via geometric and stereochemical matching, needs expanding.10 Furthermore, a crucial role for amorphous calcium carbonate as a transient intermediate has been proposed.7,11 In the last years it was demonstrated that specific nucleation of calcite can be achieved without having an epitaxial match with the exposed crystal planes12,13,14 and it has been suggested that electrostatic interactions play an important role in the control of the orientation of crystal growth.15,16 In nature crystal nucleation and growth is often controlled by carboxylate-rich polypeptides (e.g. aspartate and glutamate) present within a macromolecular matrix with organized β-sheet domains. Although(poly)peptides have been successfully employed as additives able to modify the growth of calcium carbonate crystals,17 there have been few reports in which peptides with predefined secondary structures have been studied as templates for mineralization.4b,16a,18,19
Recently it was demonstrated that amphiphilic peptides comprising phenylalanine (Phe) and glutamic acid (Glu) residues can form ordered Langmuir monolayers with a β-pleated sheet structure at the air-water interface.20 A family of amphiphilic lipopeptides (ALPs, 1a-c in Chart 3.1) has been investigated for its ability to assemble into β-pleated sheet structures at the air-water interface (Chapter 2). The ALPs are composed of amphiphilic oligopeptide moieties with alternating hydrophobic and hydrophilic amino acid residues, (Leu-Glu)n, conjugated to the phospholipid 1,2-Dioleoyl-sn-Glycero-3-Phosphoethanolamine (DOPE, 3). Although the peptide domain in tetra-ALP (1a) was too short to organize into a defined secondary structure, hexa- (1b) and octa-ALP (1c) have been shown to form β-sheet monolayers with a well-defined two-dimensional architecture and thereby exposing an ordered array of carboxylic acid groups to the aqueous phase.
In this Chapter, the use of these monolayers as biomimetic mineralization templates is reported. Calcium carbonate was grown in the presence of monolayers of the ALPs and the outcome of this study is discussed in the following sections. Furthermore, to gain a better understanding of the role of each domain in the templating effect, the water soluble octapeptide acetylated at the N-terminus Ac-(Leu-Glu)4-NH2 (2) and the phospholipids DOPE (3) were studied as well (Chart 3.1).
HN N H
NH2 O HO O
O O
O O O
O P O OH O
NH O
O n
1a-c (a: n=2; b: n=3; c: n=4)
O O O O
O P O OH O
NH2
3 AcHN
O HN
O
HO O NH
O HN
O
HO O NH
O HN
O
HO O
NH2 NH
O HN
O
HO O 2
Chart 3.1. Chemical structures of the amphiphilic lipopeptides, ALPs (1a-c), N-acetylated octapeptide Ac-(Leu-Glu)4-NH2 2 and phospholipid DOPE 3.
3.2 Results and Discussions
3.2.1 Monolayer Studies on Milli-Q Water, 10 mM CaCl2 and Saturated CaCO3
In order to investigate the influence of the complexation with Ca2+ ions on the secondary structure of the peptide domains, isotherms were recorded using an aqueous 10 mM CaCl2 solution as the subphase. The surface pressure versus area (π-A) isotherms of momolayers of tetra-ALP (1a), hexa-ALP (1b) and octa-ALP (1c) on milli-Q water and 10 mM CaCl2 are shown in Figure 3.2. For comparison, also the monolayer of DOPE (3) was investigated under the same conditions.
Figure 3.2. Surface pressure-area (π-A) isotherms of (a) tetra-ALP, (b) hexa-ALP, (c) octa-ALP and (d) DOPE on (⎯) milli-Q water and (----) 10 mM CaCl2 subphases. The limiting molecular areas were calculated by extrapolating the slope of each curve in the steepest part to zero pressure (see dotted line as an example in d).
Films were allowed to equilibrate for 10 min before compression.
In the case of the monolayers of tetra-ALP (1a) and DOPE (3), no dramatic effect was observed comparing the isotherms on milli-Q water to the 10 mM CaCl2 subphases. In contrast, for both hexa-ALP (1b) and octa-ALP (1c) isotherms, an expansion in the limiting molecular areas of 6.5-7 % was observed when a 10 mM CaCl2 subphase was used (Table 3.1), suggesting that the monolayers interact with the Ca2+ ions. However, CD and IR spectra of the transferred monolayers of compound 1c still revealed the spectroscopic characteristics typical of β-pleated strands (data not shown). Furthermore, in situ grazing incidence X-ray diffraction (GIXD) measurements on CaCl2 (10 mM) subphase showed that the two-dimensional crystalline architecture was retained (see Table 1 in Appendix 1).
Interestingly, although the slope of the isotherm of hexa-ALP on milli-Q water was less steep compared to the slope of octa-ALP isotherm, when 10 mM CaCl2 was used as subphase, hexa- and octa-ALP monolayers showed similar slopes of the isotherms, implying adaptability
50 100 150 200 250
0 10 20 30 40 50
π (mN/m)
Mma (Ų/molecule)
c
50 100 150 200 250
0 10 20 30 40 50
π (mN/m)
Mma (Ų/molecule)
a
50 100 150 200 250
0 10 20 30 40 50
π (mN/m)
Mma (Ų/molecule)
b
50 100 150 200 250
0 10 20 30 40 50
π (mN/m)
Mma (Ų/molecule)
d
of hexa-ALP upon complexation of Ca2+ ions, which consequently resulted in an increased level of organization in the monolayer of the amphiphile 1b.
Table 3.1. Expansion in the limiting molecular area for both hexa- and octa-ALP monolayers, changing from milli-Q water to 10 mM CaCl2 subphase.
For comparison, the N-acetylated octapeptide Ac-(Leu-Glu)4-NH2 (2) was also investigated. The π-A isotherm of compound 2 revealed a sharp transition from a gas to a solid phase (Figure 3.3). From this curve a molecular area of 61 Å2/molecule was determined, which was lower than expected (130 Å2/molecule21), probably due to the high water solubility of the octapeptide 2.
0 50 100 150 200 250 300
0 10 20 30 40 50 60
f e
d c
b
a
π (mN/m)
MmA (Å2/molecule)
Figure 3.3. π-A isotherms of the monolayer of (a-c) N-acetylated octapeptide 2 and (d-f) octa-ALP (1c). π-A isotherms were recorded on (a,d) milli-Q water, (b,e) 10 mM CaCl2 and (c,f) 9 mM Ca(HCO3)2 subphase at 20°C. The limiting molecular areas for compounds 1c and 2 were determined by extrapolating the slope in the liquid condensed region to zero surface pressure.
Indeed, the spectroscopic data indicated the presence of an antiparallel β-sheet structure at the air-water interface also in the case of the octapeptide 2. The CD spectrum of the transferred monolayer of N-acetylated octapeptide 2 displayed a maximum at 201 nm and
π range (mN/m)
Area on milli-Q water
(Å2/molecule) Area on 10 mM CaCl2
(Å2/molecule)
hexa-ALP (1b) 15-25 186 198 (6.5 % increase)
octa-ALP (1c) 25-35 189 202 (7% increase)
a minimum at 221 nm, while the IR spectrum showed a strong band at 1627 cm-1 and a weak one at 1694 cm-1 (Figure 3.4).22
Figure 3.4. (a) CD and (b) FT-IR spectra of the transferred monolayer of N-acetylated octapeptide 2 from an aqueous sub-phase (π=20 mN/m; T=20°C).
When the isotherm of N-acetylated octapeptide 2 was recorded on a 10 mM CaCl2
subphase, an expansion of the monolayer was imperceptible, in contrast to what was observed for the monolayer of octa-ALP (1c), as shown in Figure 3.3. These opposite behaviors were even more pronounced when solutions of octa-ALP (1c) and octapeptide 2 were spread on 9 mM Ca(HCO3) 23 subphase (solution used in the crystallization experiments described below), suggesting that the monolayer of octapeptide 2 does not significantly adapt its structure upon exposure to Ca2+ ions. These results indicated that, due to the DOPE moiety,
1 90 200 21 0 220 230 240 250 260 -0,5
0,0 0,5 1,0 1,5
Ellipticity (a. u.)
W avelength (nm ) 201
221 a
1800 1750 1700 1650 1600 1550 1500 97,5
98,0 98,5 99,0 99,5 100,0 100,5
% Reflectance
Wavenumber(cm-1) 1694
1627 b
the lipopeptide 1c formed a monolayer that self-assembles without compression, but it is still dynamic and adapts its structure upon complexation of calcium ions to efficiently interact with the nucleating crystals.
3.2.2 CaCO3 Crystallization Experiments
A supersaturated solution of Ca(HCO3)2 was prepared using the Kitano method.23 The solution was made by passive carbon dioxide (CO2) out gassing through a suspension of CaCO3 in milli-Q water. Films of ALPs (1a-c), N-acetylated octapeptide (2) and DOPE (3) were spread at the air-water interface of a freshly prepared supersaturated Ca(HCO3)2
solution poured in crystallization dishes. The proper amount of each compound was calculated so that the entire surface area (100% surface coverage) of the crystallization dish was covered by the monolayer (the limiting molecular area values were found extrapolating to zero surface pressure in the isotherms and were used as reference data for these calculations, for details see the experimental section). Furthermore, to investigate the self-organizing properties of the templates, films were spread in order to cover only 10% of the surface of the crystallization dish. Control experiments were performed under the same conditions in the absence of monolayers to enable comparison of the results.
After 20 hrs, the crystallization dishes were observed by the naked eye and compared to the control experiment, in which no monolayer was present. These observations already revealed that, in some cases, the monolayers exerted an effect on the crystallization of CaCO3. The differences in the behavior of the monolayers of ALPs (1a-c), octapeptide 2 and DOPE 3 observed are depicted in Figure 3.5 and summarized in Table 3.2.
Table 3.2. Observation by the naked eye of the influence of the monolayers on the crystallization of CaCO3
compared to the control experiments in which no monolayer was present.
Surface
coverage Influence visible by the naked eye
Top view of the crystallization dishes as
depicted in Figure 3.4
tetra-ALP (1a) 100 % no a
tetra-ALP (1a) 10 % no a
hexa-ALP (1b) 100 % yes b
hexa-ALP (1b) 10 % no a
octa-ALP (1c) 100 % yes c
octa-ALP (1c) 10 % yes d
N-acetylated octapeptide (2) 100 % yes c
DOPE (3) 100 % yes e
Figure 3.5. Schematic representation of the appearance of the surface of the crystallization dishes (top view).
Observation by the naked eye of the influence of the monolayers on the crystallization of CaCO3 were done using the control experiment as reference. Big dots (•) represent dense packed randomly grown crystals, as observed in the control experiment. Small dots (⋅) represent small isolated crystals.
The surface of the crystallization dishes containing the monolayers of tetra-ALP (1a) appeared covered by densely packed randomly grown crystals, as was observed in the control experiment (Figure 3.5, a). In contrast, when the films of hexa-ALP (1b) and DOPE (3) were spread in such a way that the entire surface of the crystallization dishes was covered by those molecules (100% surface coverage), an effect was already perceived by the naked eye.
Among densely packed randomly grown crystals, some isolated smaller crystals were present (Figure 3.5, b and e). N-acetylated octapeptide (2) and octa-ALP (1c) appeared to exert the strongest effect on the crystallization of CaCO3, as the surface of both crystallization dishes at 100% of coverage was entirely constellated by small isolated crystals (Figure 3.5, c).
Interestingly, already at 10% of the surface coverage, the monolayer appeared to act as template for the formation of few isolated crystals (Figure 3.5, d), implying self-organization.
This result is in agreement with the observations gained on a milli-Q water subphase by Brewster Angle Microscopy (see Chapter 2).
Crystals were isolated for further analysis by Langmuir-Blodgett and Langmuir-Shaefer dipping on glass microscopy slides. Optical microscopy (OM) was used to obtain a general overview of the type of crystals formed and to calculate the percentage of modified crystals.
In the control experiments, in the absence of monolayer, only few vaterite and rhombohedral calcite crystals that appeared extensively aggregated and randomly oriented (unmodified) were isolated at the air-water interface (Figure 3.6).
a b c d e
Figure 3.6. Optical micrographs images of crystals isolated from the control experiments, in the absence of monolayer. Crystals collected (A) at the air-water interface and (B) at the bottom of the crystallization dish.
Scale bar: 90 µm (A) and 60 µm (B).
Also in the case of tetra-ALP (1a) and DOPE (3), mainly unmodified calcite was isolated, as found in the control experiments. Noticeably, this could be related to the fact that neither β-sheet nor self-organization was observed for those templates. However, in the presence of DOPE (3), some vaterite crystals (~ 100 µm) were found, most probably induced by the presence of the free amine moiety (Figure 3.7, A).24
Figure 3.7. (A) Optical micrograph image of vaterite crystals formed underneath a monolayer of compound 3.
Scale bar: 90 µm. (B) Optical micrograph of type I and II calcite crystals formed underneath monolayers of 1b (100% surface coverage) and 1c (100% and 10% surface coverage). Scale bar: 500 µm.
Pyramidal (Type I) and rhombohedral (Type II) shaped calcite crystals were formed predominantly in the presence of hexa-ALP monolayer, when the surface of the crystallization dish was totally covered by molecules of the amphiphile 1b (Figure 3.7, B). The same types of modified calcite were also found in the presence of octa-ALP monolayer (Figure 3.7, B). In addition crystallization experiments in which the amount of lipopeptide 1c covered only
A B
A B
10% of the available surface area were performed. Under these conditions a similar ratio of pyramidal and indented crystals were obtained after 20 hrs, indicating that also at lower surface concentrations the monolayer self-assembled to attain the same template structure. For comparison, crystals were also grown under compressed monolayers of the octa-ALP in a Langmuir trough. These experiments were performed at a surface pressure of 30 mN/m (A = 189 Å2/molecule) and yielded similar results as those gained using self-assembled monolayers spread in crystallization dishes.
Scanning electron microscopy (SEM) was used to determine the morphology and orientation of the modified calcite crystals, which were grown underneath the different monolayers.
Type I calcite crystal Type II calcite crystal
Figure 3.8. SEM images of (A) and (B) crystals grown in the presence of hexa-ALP. Scale bar: 50 µm (A) and 100 µm (B). (C) and (D) octa-ALP. Scale bar: 50 µm (C) and 20 µm (D).
In the case of tetra-ALP (1a) only randomly oriented calcite crystals were found, similar to the unmodified crystals that were grown in the absence of a monolayer (control experiments), in agreement with the observation by the naked eye and the optical microscopy investigation discussed above. The pyramidal (Type I) and rhombohedral (Type II) calcite crystals that nucleated underneath the monolayers of hexa-ALP (1b) and
B A
D C
octa-ALP (1c) were further investigated by SEM as shown in Figure 3.8. The rhombohedral crystals, were found to have an interesting uncommon feature, as a concave central region was observed (Figure 3.8, B and D; Figure 3.10, b). In some cases the indented crystals exhibited a central patch (Figure 3.10, c), which was attributed to the nucleation point. Type I and II calcite crystals were also found in the presence of the monolayer of the N-acetylated octapeptide (2). On the contrary, in the case of DOPE (3), besides some vaterite crystals, also a few modified calcite crystals were observed. However, these crystals exhibited several different morphologies and no special type was formed predominantly.
In summary, the analysis of the grown crystals revealed that monolayers of tetra-ALP (1a) did not influence the crystallization of calcium carbonate. In contrast, two types of modified calcium carbonate calcite crystals (type I and II) were grown underneath monolayers of hexa-ALP (1b), octa-ALP (1c) and N-acetylated octapeptide (2). DOPE (3) showed a different templating effect, favoring the nucleation of another polymorph of calcium carbonate, vaterite.
In one set of crystallization experiments, CaCO3 crystals were isolated after 20 hrs.
The two types of modified calcite crystals, triangular (Type I) and indented rhombohedral (Type II) described above, were formed underneath the monolayers of hexa-ALP (1b), octa-ALP (1c) and N-acetylated octapeptide (2) in different percentages, as summarized in Table 3.3. The highest percentage of the total modified crystals (both type I and II) was found for the N-acetylated octapeptide 2.
Table 3.3. Percentage of modified calcite crystal for samples of hexa-ALP, octa-ALP and N-acetylated octapeptide 2, isolated by vertical dipping after 20 hrs (2 samples each) at 100 % coverage.25
Hexa-ALP (1b) Octa-ALP (1b) Ac-(Leu-Glu)4-NH2 (2) Nucleation density
(crystals/mm2) 12 ± 1 16.5 ± 1 11.5 ± 2
Modified crystals (%) 38 44 69
Indented crystals (%) 89 ± 0.5 79 ± 1 49 ± 5
Triangular crystals (%) 11 ± 0.5 21 ± 1 51 ± 5
A trend in the influence of the monolayers on the total amount of modified crystals could be observed: hexa-ALP ≤ octa-ALP < Ac-(Leu-Glu)4-NH2. This tendency could be
related to the difference in both rigidity and self-organization of the three monolayers.
However, the percentage of the new type of indented crystals (type II) was higher in the case of the hexa-ALP (10 % higher compared to the octa-ALP) possibly due to its increased adaptability in the presence of Ca2+ ions.
In the next set of crystallization experiments, crystals grown under monolayers of octa-ALP (1c) and N-acetylated octapeptide (2) were isolated after 5, 20 and 48 hrs and this population slightly increased in time (Figure 3.9).
In the case of octa-ALP, the indented crystals were dominant over the pyramidal crystals. For Ac-(Leu-Glu)4-NH2 predominantly pyramidal crystals were observed after 5 hrs and only a minor (<5%) amount of indented crystals was present. The latter population increased to approximately 50% of the total number of modified crystals after 20 hrs and only after 48 hrs became predominant (65%, Figure 3.9, b). Importantly, most of the pyramidal crystals increased in size rather than in number, suggesting that their nucleation only occurred in the earlier stages of the experiment. In contrast, the appearance of small indented crystals throughout the whole experiment was observed, indicating that these nucleated also in later stages of the experiment.
Figure 3.9. The distribution of the two populations of crystals (Type I – pyramidal shape, Type II – indented crystals) nucleated under monolayers of (a) octa-ALP (1c) and (b) Ac-(Leu-Glu)4-NH2 (2), in percentage of the number of modified crystals (crystals isolated after 5, 20 and 48 hrs).
In summary, the presence of ordered β-sheet domains and the template adaptability appeared key features for the nucleation of the new type of indented calcite crystals (Type II), as indicated by the observation that the more adaptable monolayer induces the formation of the indented rhombohedral crystals. However, self-organization was also required to have an efficient template, especially at low surface coverage.
0 2 0 4 0 60 8 0 100
5hrs 20hrs 48hrs
20 40 60 80 100
5hrs 20hrs 48hrs
b
a Type I
Type II
% of modified
crystals
0
3.2.2 Crystal Analysis26
The pyramidal shaped calcite crystals (Type I) are characterized by 3 thermodynamically stable {10.4} faces, which had been oriented to the aqueous phase, and one face consisting of 3 facets that had been oriented to the monolayer (Figure 3.10, a).
Figure 3.10. Calcite crystals grown under monolayers of 1b, 1c and 2. (a) left: SEM image of a (01.2) – oriented pyramidal crystal as observed from the side attached to the monolayer; right: model27 and SEM image of a similar crystal viewed from the side exposed to the solution. (b) left: SEM image of rhombohedral crystal with a concave central region from the side attached to the monolayer; bottom right: model of this crystal;
top right: SEM image of a similar crystal viewed from the side exposed to the solution. (c) Left: SEM image of indented crystal showing a central patch; right: TEM image and corresponding electron diffraction pattern of a rhombohedral crystal isolated after 20 min. The pattern corresponds to the [10.0] zone of calcite. Reflections A, (01−.1−) (4.2 Å); B, (01−.4) (3.0 Å); C, (00.5) (3.4 Å). Angles, (01−.1−)⌃(00.5) = 104º; (01−.1−)⌃(01−.4) = 59º. Camera length = 60 cm. Bars represent 20 µm except when indicated otherwise.
b a
104 014_ _ _ 114 __
A B A BCC
c
014 _
114 __
104
200 nm 200 nm 200 nm
Crystals showing a similar elevated feature consisting of three inclined facets have been observed before.13b,16c,24 It has been suggested that the apex represented the initial point of attachment to the monolayer and that the outer edges of the crystal detached in time from the monolayer due to gravity. Computer modeling27 of the crystals based on the scanning electron micrographs indicated that in almost all cases the facetted faces belonged to the class {01.l}, with l=1-2. The formation of crystal faces of the type {01.l} with l=1-2 in solution has been reported frequently for biogenic calcite28 and recently also for the octapeptide (Phe-Asp)4.16a In addition, (01.2) oriented calcite crystals have been obtained using different monolayer systems.12,13a,16b,29,30
Previously, Lahiri et al. reported the formation of calcite crystals with symmetrical indentation defined by three {01.2} faces around a (00.1) face which formed the attachment point to a porphyrin monolayer.31 In the present, case the concave indentation of these crystals is defined by 4 roughened planes (Figure 3.10, b), the outer edges of which form a plane that can be modeled by the (11−.4−) face of calcite. This morphological appearance of indented rhombohedral crystals, however, has not been reported before. Also in the present case, on several crystals a central patch could be located, which was assigned to the area that had been adhered to the monolayer and where nucleation had started (Figure 3.10, c). Selected area electron diffraction performed on young crystals with a rhombohedral shape collected within 20 min after the start of the experiment revealed that these crystals had a {10.0} orientation (Figure 3.10, c). This suggests that for the indented crystals the central patch corresponds to a {10.0} face.
A proposed model for the formation of indented crystals is shown in Figure 3.11. The observed crystal indentation might be the combined result of gravity and a limited Ca2+ and HCO32- ion transport from the mineralization solution.
Figure 3.11. Proposed model for the formation of indented crystals.
Central patch (nucleation face)
Time Monolayer
Indented crystal
Gap
As the crystal grows it becomes heavier and tends to sink, while still being attached to the monolayer. Being flexible, the monolayer presumably bends to follow the sinking crystal.
This will create a gap between the monolayer and the face of the crystal attached to it, towards the edges of the crystal. Therefore, the hindering effect of the monolayer on ion transport will be also reduced in that region. This will allow the outer corners of the crystal facing the monolayer to grow upwards, creating an indentation in the crystal.
3.2.3 Mechanistic Considerations
Even though the monolayer of hexa-ALP (1b) formed two-dimensional β-sheet assembly, it did not show self-assembling properties, as discussed above. Both octa-ALP (1c) and N-acetylated octapeptide (2) did self-assemble without compression to form stable monolayers, which were active in the nucleation of oriented calcite. However, in contrast to 2, octa-ALP (1c) formed a monolayer that was still dynamic and easily adapted its structure upon complexation of calcium ions, as demonstrated by the increased surface area in the isotherms on both CaCl2 (10 mM) and CaHCO3 (9 mM), to efficiently interact with the nucleating crystals. From the change of the composition of the mineral phase in time it becomes apparent that the monolayers of 1c and 2 differ in their ability to nucleate the indented versus the pyramidal morphology at the different time points. One could propose that in the case of 2, the nucleation of the face leading to the indented calcite takes place in later stages of the experiments, probably because the solid-like monolayer needs more time to reorganize and to adapt to the developing crystals. It is evident that the origin of this difference lies in the presence of the DOPE moiety. It is interesting to note that the introduction of the DOPE group increased the flexibility of the monolayer without affecting the ability to form a β-sheet structure. From the molecular area, however, it follows that the phopsholipid head group must be in contact with the subphase. One could imagine that in this system the DOPE moiety acts as a kind of buffer region that separates neighboring β-sheet domains, thereby preventing it from forming a rigid structure as seems to be the case for octa-ALP. As stretching of the peptide segment cannot account for the increased surface area on an aqueous Ca(HCO3)2 subphase, it is evident that also significant changes in the phospholipids group must play an important role. Nevertheless, this apparently does not affect the templating behavior of the monolayer other than shifting the ratio between the two crystal types that were already formed by 1c. This is remarkable, in particular if one considers the strong calcium binding properties of phosphate groups.
The {01.1} and {01.2} crystal planes are defined by the same 4.99 Å distance in one direction, but show a gradual change in the orientation of the carbonate groups which are rotated over approximately 10o going from l=1 to l=2 (Figure 3.12). The 4.99 Å distance most probably relates to the interstrand distance in the template as defined by the hydrogen bonds in the β-sheet (4.7-4.8 Å, GIXD data, Chapter 2), allowing an approximate match of the template carboxylate groups with the carbonate ions in the crystal plane. However, although there is only 10o rotation between the planes with l=1 and l=2, significant differences are found in the inter-ion distances in the different nucleation planes. It is therefore likely that in this direction the alignment of the carboxylate groups with respect to the carbonate ions plays a much more important role than the matching of the distances of the two phases.
Figure 3.12. Model of the (a) (01.2); (b) (01.1) and (c) (10.0) plane of calcite with the viewing direction (left) along and (right) perpendicular to the plane.
The more adaptable monolayers of 1b and 1c clearly favor the formation of a different set of planes, i.e. those belonging to the {10.0} family. Interestingly, these planes are related to the {01.l} planes (with l=1-2) by a further rotation of the carbonate ions (Figure 3.12). The {10.0} faces are again defined by a distance of 4.99 Å in one direction and have a spacing of 8.53 Å in the other direction. This suggests that also in this case the spacing of 4.99 Å relates to the interstrand distance in the β-sheet of 4.7-4.8 Å. The interaction of the carboxylate
a
c b
c a
groups of the template with the {10.0} faces along the other direction is most likely related to the ability of the monolayer to adapt to the growing crystal.
The surface pressure-area (π-A) isotherms recorded on milli-Q water and Ca(HCO3)2
subphases showed an expansion of the mean molecular area from 189 to 250 Å2/molecule, respectively (Figure 3.3). The expansion in the direction of the backbone of the molecule could account for most of the observed increase in mean molecular area.32 If a small expansion in the interstrand distance from 4.7-4.8 Å to 4.99 Å is envisioned, the molecule should then expand from ~ 40 Å to ~ 50 Å (i.e. by ~ 25%) along its long axis. It is important to note, however, that stretching of the peptide backbone alone to an all-trans conformation yields a maximum distance between the Glu side chains of ~ 7.6 Å compared to the 8.53 Å spacing of the {10.0} face. Nevertheless, the flexibility of the template should allow it to organize such that the carboxylates align with the carbonate ions in the nucleation plane while matching the 4.99 Å distance, similar to what has been proposed for amide-containing phospholipid derivatives.13 However, the importance of other factors, such as charge density of the amphiphile and the possible presence of a layer of carbonate ions between the crystal and the monolayer should not be neglected.15c,16 It is proposed that the N-acetylated octapeptide 2 has a limited ability to adapt to the growing crystal and hence is responsible for the preferred nucleation of the {01.l} faces (with l=1-2) and the much slower formation of the indented crystals.
3.2.4 Flexibility and Adaptation in Other Systems
In previous studies, the formation of both {01.2}13a,28,30 and {10.0}13b,31 oriented crystals has been discussed in terms of geometrical lattice matching and stereochemical complementarity between the functional groups of the template and the positions of ions in the nucleation plane. In other studies the oriented nucleation of crystals was attributed to non-specific electrostatic effects rather than to an epitaxial match between the monolayer and the crystal face.15,16 Surprisingly, although the importance of the flexibility of organic surfaces used as templates for CaCO3 crystallization had already been emphasized by Mann et al. in 1988,33 the ability of the template to adapt to the growing crystals was only rarely taken into account. Charych et al.34 demonstrated the template adaptability for the first time in situ.
Polymeric films of 10,12-pentacosadiyonic acid were shown to reorganize upon calcite mineralization, in order to optimize the interactions between the carboxylates of the film and the carbonate ions in the (01.2) calcite face, as indicated by in situ Fourier transform infrared (FTIR) and visualized by color changes in the polymer film. In other studies it was observed
that more efficient control of crystal orientation was obtained under monolayers compressed at a low surface pressure, in which the molecules had a higher degree of mobility compared to the fully compressed monolayers.35 An interesting recent example of this was given by Volkmer et al. who compared the effect of compressed monolayers of two macrocyclic compounds bearing carboxylate groups, a resorc[4]arene36 and a calix[4]arene,16c on the crystallization of calcium carbonate.16b It was found that only the resorc[4]arene led to a modified crystallization behavior, i.e. to the formation of aragonite instead of calcite. In contrast, under the calix[4]arene-based monolayers only randomly oriented calcite crystals were formed. At lower degrees of compression the formation of significant amounts of (11−.0) modified calcite together with vaterite was observed for the recorc[4]arene, while (01.2) modified calcite was observed for the calix[4]arene. These observations were attributed to the different charge densities of the respective monolayers.37 In addition, however, it is of interest to note that the monolayer isotherms show that only the resorc[4]arene increased its molecular area upon calcium complexation, whereas the calix[4]arene showed no changes in the π-A isotherm. These results are in line with the observations described above, which showed that the capability of a monolayer to nucleate a specific crystal form is related to its ability to adapt to the developing mineral phase. The necessity of a templating molecule to adapt to the requirements of developing crystals was demonstrated already 15 years ago.38 In this respect it is important to note that also dentin matrix protein 1 (DMP1) has been shown to adapt its structure to form a β-sheet template during the nucleation process.39
3.3 Conclusions
In conclusion, it was demonstrated that amphiphilic lipopeptides 1b and 1c formed stable monolayers with an antiparallel β-sheet conformation on a milli-Q water subphase (Chapter 2) as well as on Ca2+ aqueous solutions. This enabled not only a study of the effect of these monolayers on the crystallization of calcium carbonate, but also to investigate the way they interact with the developing mineral phase. Indeed the formation of habit-modified calcite was observed. Apart from a small amount of pyramidal {01.l} oriented crystals (l=1-2), the majority of the modifications resulted in a new type of indented {10.0} oriented calcite crystals. The formation of this new morphological form was significantly suppressed when the less adaptable peptide 2 was used. Furthermore, few vaterite crystals and different types of modified calcite or only randomly growing crystals were found when the monolayers of compounds 1a and 3 were used, which was mainly attributed to the lack of β-sheet structures.
With the present system it was demonstrated that the nucleation of different crystal faces can be achieved depending on the ability of the template to adapt to the structure of the inorganic phase. Furthermore, the results indicate that stretching of the template in only one direction allows the reorientation of the template’s functional groups such that the stabilization of different crystal planes can be achieved without the need for an epitaxial relation between the two components.
3.4 Experimental Section
General materials and methods. All crystallization experiments were performed in collaboration with dr. Daniela Popescu (Eindhoven University, The Netherlands). Chloroform (AR) and methanol (AR) were purchased from Biosolve B.V., trifluoroacetic acid from Sigma-Aldrich B.V. and nitric acid 65% (AR) from VWR. Calcium carbonate (precipitated, p.a.) was purchased from Merck. All reagents and solvents were used as received. Ultrapure water (18 MΩ/cm)was generated using a Barnstead Easypure® LF water purification system
Synthesis of ALPs. See Chapter 2 (experimental section).
Synthesis of N-Ac-(Leu-Glu)4-NH2. Fmoc-NH-(Leu-Glu)4-NH-anchored to a Sieber Amide resin was prepared following the protocol already described in Chapter 2 (0.25 mmol scale). After Fmoc removal by treatment of the resin with 20% piperidine in NMP, the free N-terminus was acetylated with a solution of 0.25 M AcO2 and 0.125 M of DiPEA in NMP (shaken for 5 min). After the Kaiser test confirmed completeness of the acetylation, the resin was washed first with NMP (3x) and subsequently with DCM (3x). The peptide was cleaved from the solid support by treatment with 10 ml of 9/1 (v/v) TFA/H2O for 1 hour at room temperature.
After filtration of the resin, the TFA solution was dropped into cold Et2O and centrifuged. Freeze-drying yielded 186 mg of crude peptide (0.18 mmol, 72% yield). The crude product was purified by HPLC. LC-MS: 7.6 min (gradient 30 → 90% buffer B, for details see general material and methods in Chapter 2); ESI-MS: m/z = 1028 [M+H]+; m/z = 1011.6 [M-OH+H]+.
Surface pressure-surface area (π-A) isotherms. The isotherms on 10 mM CaCl2 were recorded applying the same conditions used for the water subphase as described in Chapter 2 (experimental section).
Isotherms were recorded on a supersaturated Ca(HCO3)2 subphase (prepared as described in the section crystallization experiments) in a Teflon® Langmuir trough (KSV Optrel BAM, KSV Instrument Ltd., Helsinki, Finland, for details see also Chapter 2).
Brewster Angle Microscopy (BAM). See Chapter 2 (experimental section). In this case, a 9 mM supersaturated solution of Ca(HCO3)2, prepared as described in the crystallization experiments, was used as the subphase.
Crystallization experiments. All glassware used for the crystallization experiments was thoroughly cleaned with soap, nitric acid solution (∼ 10% overnight) and methanol. Between each cleaning step, the glassware was rinsed extensively (3x) with ultrapure water. For all crystallization experiments a 9 mM supersaturated solution of Ca(HCO3)2 prepared following the Kitano method23 was used. The solution was made by bubbling CO2 through a suspension of CaCO3 (3.5-4 g) in ultrapure water (1.5 l) for a period of 1.5 h, followed by filtration and CO2 bubbling for another 30 min to dissolve any CaCO3 particles present. Films were
spread from 9/1 (v/v) Chloroform/TFA solutions (1 mg/ml) at the air-water interface of freshly prepared supersaturated Ca(HCO3)2 solution poured in crystallization dishes (50 ml/dish, Schott Duran®, Germany) or in a Teflon® Langmuir trough (KSV Optrel BAM, KSV Instrument Ltd., Helsinki, Finland). For the crystallization dishes (φ = 60 mm), the amount of molecules to spread was calculated taking into account the limiting MmA, obtained by extrapolation of the steepest part of the π-A isotherms (110 Å2/molecule for 1a, 186 Å2/molecule for 1b, 189 Å2/molecule for 1c, 61 Å2/molecule for 2 and 71 Å2/molecule for 3) such that the molecules would cover 100% and 10% of the surface of the crystallization dishes, respectively. For the experiments performed in the Langmuir trough, the monolayers were compressed until reaching a surface pressure π = 30 mN/m which was maintained constant throughout the experiment. Crystallization of CaCO3 was governed by the slow loss of CO2
gas from the supersaturated solution according to equation (1).33a
Ca2+(aq) + 2HCO3-(aq) CaCO3(s) + CO2(g) + H2O(l) (1)
Crystals used for optical microscopy and SEM investigation were collected on glass microscopy slides (diameter = 15 mm, Menzel-Glasser, Germany) by vertical (Langmuir-Blodgett) and horizontal (Langmuir-Schaefer) transfer after 5, 24 and 48 hrs, following established procedures.31 The crystals isolated by vertical dipping of the glass slides through the monolayer expose the faces that were attached to the monolayer, while the crystals isolated by horizontal dipping expose the side that was facing the mineralization solution (Figure 6.13).
.
Figure 3.13. Sample isolation from Langmuir monolayers. (a) vertical dipping (Langmuir-Blodgett) and (b) horizontal dipping (Langmuir-Schaefer).
Optical Microscopy (OM). CaCO3 samples were examined using a Jenaval polarization microscope with crossed polarizers.
Glass slide Glass slide (a) Vertical dipping
(b) Horizontaldipping
TEM/SEM/optical microscopy TEM/SEM/optical
microscopy
Scanning Electron Microscopy (SEM). The samples were mounted on aluminum stubs with double-sided carbon tape. The specimens were observed with a Philips XL30 scanning electron microscope. A secondary electron detector was used at an accelerating voltage of 1.0-2.0 kV.
Transmission Electron Microscopy (TEM). TEM was performed on a JEOL 2000-FX electron microscope operating at an accelerating voltage of 80 or 120 kV. Crystals used for selected area diffraction were collected within 20 min after the start of the experiment by Langmuir-Schaefer transfer on a carbon-coated TEM grids. The isolation procedure is based on an established technique, which mitigates any disruption of the crystal alignment relative to the template.31 Therefore, the crystals isolated on the TEM grids should be laying on the face that was attached to the monolayer, i.e. the face they nucleate from.
3.5 References and Notes
1 a) Dujardin, E.; Mann, S. Adv. Eng. Mater. 2002, 4, 461-474. b) Aizenberg, J. Adv. Mater. 2004, 16, 1295- 1302.
2 Green, D.; Walsh, D.; Mann, S.; Oreffo, R. O. C. Bone 2002, 30, 810-815.
3 Chakrabarty, D.; Mahapatra, S. J. Mater. Chem. 1999, 9, 2953-2957.
4 a) Belcher, A. M.; Wu, X. H.; Christensen, R. J.; Hansma, P. K.; Stucky, G. D.; Morse, D. E. Nature 1996, 381, 56-58. b) Levi, Y.; Albeck, S.; Brack, A.; Weiner, S.; Addadi, L. Chem. Eur. J. 1998, 4, 389-396.
5 Davies, P.; Dollimore, D.; Heal, G. R. J. Thermal Anal. 1978, 13, 473-487.
6 Mann, S. Biomineralization: Principles and Concepts in Bioinorganic Materials Chemistry 2001, Oxford University Press, New York.
7 Addadi, L.; Raz, F.; Weiner, S. Adv. Mater. 2003, 15, 959-970.
8 Harting, P. Q. J. Microsc. Sci. 1872, 12, 118-123.
9 Mann, S. Nature 1988, 332, 119-124.
10 Cölfen, H.; Mann, S. Angew. Chem. Int. Ed. 2003, 42, 2350-2365.
11 Politi, Y.; Arad, T.; Klein, E.; Weiner, S.; Addadi, L. Science 2004, 306, 1161-1164.
12 a) Aizenberg, J.; Black, A. J.; Whitesides, G. M. Nature 1999, 398, 495-498; b) Travaille, A. M.; Donners, J.
J. J. M.; Gerritsen, J. W.; Sommerdijk, N. A. J. M.; Nolte, R. J. M.; van Kempen, H. Adv. Mater 2002, 14, 492- 495.
13 a) Champ, S.; Dickinson, J. A.; Fallon, P. S.; Heywood, B. R.; Mascal, M. Angew. Chem. Int. Ed. 2000, 39, 2716-2719. b) Buijnsters, P. J. J. A.; Donners, J. J. J. M.; Hill, S. J.; Heywood, B. R.; Nolte, R. J. M.;
Zwanenburg, B.; Sommerdijk, N. A. J. M. Langmuir 2001, 17, 3623-3628.
14 Donners, J. J. J. M.; Nolte, R. J. M.; Sommerdijk, N. A. J. M. J. Am. Chem Soc. 2003, 124, 9700-9701.
15 a) Duffy, D. M.; Harding, J. H. Langmuir 2004, 20, 7630-7636. b) Duffy, D. M.; Harding, J. H. Langmuir 2004, 20, 7637-7642. c) Duffy, D. M.; Travaille, A. M.; van Kempen, H.; Harding, J. H. J. Phys. Chem. B 2005, 109, 5713-5718.
16 a) Volkmer, D.; Fricke, M.; Huber, T.; Sewald, N. Chem. Commun. 2004, 16, 1872-1873; b) Volkmer, D.;
Fricke, M.; Agena, C.; Mattay, J. J. Mater. Chem. 2004, 14, 2249-2259. c) Volkmer, D.; Fricke, M.; Vollhardt, D.; Siegel, S. J. Chem. Soc., Dalton Trans. 2002, 4574-4554.
17 a) Addadi, L.; Weiner, S. Proc. Natl. Acad. Sci. 1985, 82, 4110-4114; b) Gower, L. A.; Tirrell, D. A. J. Crystal Growth 1998, 191, 153-160; c) Sugawara, T.; Suwa, Y.; Ohkawa, K.; Yamamoto, H. Macromol. Rapid.
Commun. 2003, 24, 847-851.
18 a) Addadi, L.; Moradian, J.; Shay, E.; Maroudas, N. G.; Weiner, S. Proc. Natl. Acad. Sci USA. 1987, 84, 2732- 2736. b) Bekele, H.; Fendler, J. H.; Kelly, J. W. J. Am. Chem. Soc. 1999, 121, 7266-7267. c) DeOliveira, D. B.;
Laursen, R. A. J. Am. Chem. Soc. 1997, 119, 10627-10631. d) Falini, G.; Fermani, S.; Gazzano, M.; Ripamonti, A. J. Chem. Soc., Dalton Trans. 2000, 3983-3987. e) Meegan, J. E.; Aggeli, A.; Boden, N.; Brydson, R.; Brown, A. P.; Carrick, L.; Brough, A. R.; Hussain, A.; Ansell, R. J. Adv. Funct. Mater. 2004, 14, 31-37.
19 Bertrand, M.; Brack, A. Chem. Eur. J. 2000, 6, 3452-3455.
20 Rapaport, H.; Kjaer, K.; Jensen, T. R.; Leiserowitz, L.; Tirrell, D. A. J. Am. Chem. Soc. 2000, 122, 12523- 12529.
21 The repeat distances of 4.7 and 6.9 Å, that have been observed previously in crystalline β-sheet structures can be used to estimate the area per molecule, e. g. 4 x 4.7 x 6.9 = 129.7 Å for an 8 residue peptide (See reference 20).
22 In aqueous solution the N-acetylated octapeptide (2) showed a random coil conformation.
23 Kitano, Y. Bull. Chem. Soc. Jpn 1962, 32, 1980-1985.
24 Mann, S.; Heywood, B. R.; Rajam, S.; Walker, J. B. A. J. Phys. D: Appl. Phys. 1991, 24, 154-164.
25 Observation: nucleation densities were comparable to the one reported by Champ et al. for self-organized melamine-cyanuric acid monolayers with carboxylic acid and phosphate functional groups (see reference 13a).
26 Elliott, S. The Physics and Chemistry of Solids, John Wiley & Sons, Chichester, UK, 1998. For a general introduction about crystal systems and crystallographic notations see also Appendix 2.
27 Calcite crystals were modeled with SHAPE V7.1.2. ©2004 by Shape Software, Kingsport (USA).
28 a) Aizenberg, J.; Hanson, J.; Koetzle, T. F.; Leiserowitz, L.; Weiner, S.; Addadi, L. Chem. Eur. J. 1995, 1, 414-422. b) Aizenberg, J.; Albeck, S.; Weiner, S.; Addadi, L. J. Cryst. Growth 1994, 142, 156-164.
29 Lahiri, J.; Xu, G.; Dabbs, D. M.; Yao, N.; Aksay, I. A.; Groves, J. T. J. Am. Chem. Soc. 1997, 119, 5449-5450.
30 Berman, A.; Ahn, D. J. ; Lio, A.; Salmeron, M.; Reichert, A.; Charych, D. Science 1995, 269, 515-518.
31 Rajam, S.; Heywood, B. R.; Walker, J. B. A.; Mann, S.; Davey, R. J.; Birchall, J. D. J. Chem. Soc., Faraday Trans. 1991, 87, 727-734.
32 In situ GIXD measurements on water and CaCl2 solution subphase (see Appendix 1) showed that the interstrand distance remained constant upon compression. The strand length however, appeared to have an elastic behavior, as upon compression the conformation of the backbone changed.
33 Mann, S.; Heywood, B. R.; Rajam, S.; Birchall, J. D. Nature 1988, 334, 692-695.
34 Ahn, D. J.; Berman, A.; Charych, D. J. Phys. Chem. 1996, 100, 12455-12461.
35 Volkmer, D.; Fricke, M. Z. Anorg. Allg. Chem. 2003, 629, 2381-2390.
36 Volkmer, D.; Fricke, M.; Agena, C.; Mattay, J. CrystEngComm 2002, 4, 288-295.
37 Volkmer, D.; Fricke, M.; Gleiche, M.; Chi L. Mater. Sci. Eng. C 2005, 25, 161-167.
38 Weiner, S.; Addadi, L. Trends Biochem. Sci 1991, 16, 252-256.
39 He, G.; Dahl, T.; Veis, A.; George, A. Connect. Tissue Res. 2003, 44, 240-245.
