Langmuir 2007, 23, 11951-11955
11951
Expected and Unexpected Effects of Amino Acid Substitutions on Polypeptide-Directed Crystal Growth Katya Delak, Sebastiano Collino, and John Spencer Evans* Laboratory for Chemical Physics, Center for Biomolecular Materials Spectroscopy, New York UniVersity, 345 East 24th Street, Room 1007, New York, New York 10010 ReceiVed July 14, 2007. In Final Form: October 4, 2007 Nature’s use of biomineralization polypeptides to control and modulate the growth of biogenic minerals is an important process that, if properly understood, could have significant implications for designing and creating new inorganic-based materials. Although the sequences for a number of biomineralization proteins exist, very little is known about the participation of specific amino acids in the mineral modulation process. In this letter, we investigate the impact of global Asp f Asn and Glu f Gln substitutions on two mollusk shell nacre polypeptides, AP7N and n16N. We find that these global substitutions, which remove all anionic Ca(II) binding sites, abolish the expected in vitro mineralization activities associated with each native polypeptide. In addition, the ability of substituted peptides to form complexes with both Ca(II) and Ca(II) metal ion analogs is also abolished. However, some unexpected effects were noted. First, the Asp f Asn, Glu f Gln substituted n16N polypeptide is observed to self-assemble and form biofilms or coatings that appear to mineralize in vitro. Second, both polypeptides are structurally affected by these substitutions, with Asp f Asn substituted AP7N transforming to an R helix and Asp f Asn, Glu f Gln substituted n16N transforming to a more unfolded random-coil-like structure. We find that the participation of Asp and Glu residues is crucial to the inherent mineralization activities and conformations of AP7N and n16N polypeptides. Surprisingly, we find that the replacement of anionic residues within biomineralization polypeptides such as n16N still permits mineral modulation, but in a different form that now involves peptide self-association and biofilm formation.
Introduction One of the more interesting aspects of the biomineralization process is the participation of specific proteins in the nucleation and growth processes of different biominerals.1-4 As an example, in some mollusk shells there exist two layers, nacre and prismatic, each of which possess different polymorphs (prismatic ) calcite, nacre ) aragonite) of calcium carbonate. It is now becoming clear that mineral formation within the nacre and prismatic layers of the mollusk shell involves the participation of distinct proteins with different sequence features.3-6 Recent studies of nacreassociated polypeptides7-10 reveal that these sequences are not exclusively polyanionic but consist of a mixture of anionic, cationic, polar, and hydrophobic sequence regions. Some of these nacre polypeptides appear to behave in a multifunctional capacity in vitro (i.e., they can block and accelerate certain aspects of calcium carbonate crystal growth10,11 and/or induce the formation * To whom correspondance should be addressed. E-mail: jse1@ nyu.edu. Tel: (212)998-9605. Fax: (212)995-4087. (1) Lowenstam, H. A., Weiner, S. On Biomineralization; Oxford University Press: New York, 1989; pp 1-50. (2) Levi-Kalisman, Y.; Falini, G.; Addadi, L.; Weiner, S. J. Struct. Biol. 2001, 10, 4372-4337. (3) Weiner, S.; Addadi, L. J. Mater. Chem. 1997, 7, 689-702. (4) Fritz, M.; Belcher, A. M.; Radmacher, M.; Walters, D. A.; Hansma, P. K.; Stucky, G. D.; Morse, D. E.; Mann, S. Nature 1994, 371, 49-51. (5) Belcher, A. M.; Wu, X. H.; Christensen, R. J.; Hansma, P. K.; Stucky, G. D.; Morse, D. E. Nature 1996, 381, 56-58. (6) Thompson, J. B.; Paloczi, G. T.; Kindt, J. H.; Michenfelder, M.; Smith, B. L.; Stucky, G. D.; Morse, D. E.; Hansma, P. K. Biophys. J. 2000, 79, 33073312. (7) Michenfelder, M.; Fu, G.; Lawrence, C.; Weaver, J. C.; Wustman, B. A.; Taranto, L.; Evans, J. S.; Morse, D. E. Biopolymers 2003, 70, 522-533. Michenfelder, M.; Fu, G.; Lawrence, C.; Weaver, J. C.; Wustman, B. A.; Taranto, L.; Evans, J. S.; Morse, D. E. Biopolymers 2004, 73, 299. (8) Samata, T.; Hayashi, N.; Kono, M.; Hasegawa, K.; Horita, C.; Akera, S. FEBS Lett. 1999, 462, 225-229. (9) Miyashita, T.; Takagi, R.; Okushima, M.; Nakano, S.; Miyamoto, H.; Nishikawa, E.; Matsushiro, A. Mar. Biotechnol. 2000, 2, 409-418. (10) Fu, G. S.; Qiu, R.; Orme, C. A.; Morse, D. E.; De Yoreo, J. J. AdV. Mater. 2005, 17, 2678-2683. (11) Kim, I. W.; Darragh, M.; Orme, C. A.; Evans, J. S. Cryst. Growth Des. 2006, 6, 6-10.
of amorphous mineral deposits11). Clearly, these interesting sequences possess molecular features that, if properly understood and translated, could propel the materials science and nanotechnology fields in new directions. Within the last 10 years, there have been several published studies that have identified the complete or partial amino acid sequences of nacre-specific proteins.7-10,12-14 In some studies, the mineralization functionalities of these proteins have also been identified.7-12 However, the participation of specific amino acid residues in the protein-directed calcium carbonate mineralization process is not fully understood. With this in mind, we embarked on a simple quest to ascertain the impact of amino acid substitution on the mineralization functions of two well-characterized nacrespecific polypeptides, n16N and AP7N.7,11,15 These 30 AA sequences represent the N-terminal mineral modification domains of two nacre-specific proteins, n16 (Japanese pearl oyster, Pinctada fucata)8 and AP7 (Pacific red abalone, Haliotis rufescens).7 These two proteins are linked to the aragonite polymorph selection process with AP7 performing the role of a calcite “blocker” protein7 and n16 acting as a facilitator of aragonite formation.8 Interestingly, both 30-AA N-terminal sequences contain limited quantities of either the Asp or Glu residue (Figure 1), each of which is localized in a specific region within each domain. As a first step toward understanding the participation of these key anionic residues in calcium carbonate mineral formation, we chemically synthesized a 30-AA variant of each domain wherein all Asp and Glu residues were replaced by Asn and Gln, respectively (Figure 1), thus removing the anionic electrostatic charge contributions of the original residues but approximating (12) Weiss, I. M.; Gohring, W.; Fritz, M.; Mann, K. Biochem. Biophys. Res. Commun. 2001, 285, 244-249. (13) Shen, X.; Belcher, A. M.; Hansma, P. K.; Stucky, G. D.; Morse, D. E. J. Biol. Chem. 1997, 272, 32472-32481. (14) Li, S.; Xie, L.; Ma, Z.; Zhang, R. FEBS J. 2005, 272, 4899-4910. (15) Kim, I. W.; DiMasi, E.; Evans, J. S. Cryst. Growth Des. 2004, 4, 11131118.
10.1021/la702113x CCC: $37.00 © 2007 American Chemical Society Published on Web 10/30/2007
11952 Langmuir, Vol. 23, No. 24, 2007
Figure 1. Primary sequences of n16N, AP7N, and their Asn and Gln variants AP7NN and n16NN. Asp f Asn substitutions in AP7NN occur at positions 1, 2, 21, and 22; Asp f Asn substitutions in n16NN occur at positions 17, 21, 24, and 27; and Glu f Gln substitution occurs at position 19. Note that AP7NN and n16NN have net molecular electrostatic charges of +2 and +9, respectively, whereas AP7N and n16N have net molecular charges of -2 and +4, respectively. Color coding: red - carboxylate residues; blue - cationic residues; green - polar, hydrogen-bonding donor/acceptor residues; purple - hydrophobic or nonpolar residues.
the same van der Waals volume and hydrophilicity at these sequence locations. These two globally substituted peptides, termed AP7NN and n16NN, were then tested alongside their original versions in Kevlar-based calcium carbonate mineralization assays15,16 and subjected to biophysical analyses that probed the effect of these substitutions on metal ion binding capabilities and secondary structure characteristics. As detailed in this letter, we find that global Asp f Asn substitutions in AP7NN result in a loss of typical mineralization activity, metal ion binding capabilities, and an intriguing perturbation in secondary structure from an unfolded conformation to a folded R helix. However, the global Asp f Asn and Glu f Gln substitutions in n16NN yielded a surprising result: the loss of typical mineralization activity and structure but the gain of a new activitysthe formation of mineralized biofilms or deposits on crystal surfaces. Materials and Methods The synthesis, purification, and characterization of free amino termini, CR-amide-capped AP7N and n16N, were performed as described in our earlier reports.7,11,15 The two variants, AP7NN and n16NN (Figure 1), were also synthesized at the 100 µmol level and purified at the Wm. Keck Biotechnology Peptide Synthesis Facility, Yale University, by Dr. Janet Crawford and staff using previously published protocols.7,11,15 After resin cleavage and reverse-phase HPLC purification (Waters C-18 column, >95% pure), the experimental molecular masses of n16NN and AP7NN were determined by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) to be 3742.8 and 3221.7 Da, in agreement with the theoretical values of 3742.3 and 3221.5 Da, respectively. In Vitro Kevlar Crystal Growth Assays. We employed a polyimide (Kevlar) assay for the nucleation of calcium carbonate crystals in the presence of AP7N, AP7NN, n16N, and n16NN. The protocol for this assay has been described in detail in earlier reports.15,16 In the current study, total polypeptide assay concentrations were 1 × 10-5, 5 × 10-5, and 1 × 10-4 M. Negative control conditions consisted of no added peptide. Assay conditions and a sample workup for SEM imaging were conducted as described in our previous polypeptide studies.15,16 SEM imaging was conducted using a Hitachi S-3500N SEM microscope at 5 kV after thin Au coatings were (16) Kim, I. W.; Collino, S.; Morse, D. E.; Evans, J. S. Cryst. Growth Des. 2006, 6, 1078-1082.
Letters applied to the samples. The SEM images presented in this report are representative of 10-20 different crystals in each assay sample. The cropping of SEM images and the adjustment of brightness/darkness and contrast levels were performed using Adobe Photoshop. To determine the presence of minerals in structures created in the presence of n16NN within Kevlar assays, parallel samples were coated with carbon and analyzed with EDX using the Hitachi system coupled with a Princeton Gamma Tech IMIX system with a prism light element detector. Analyses were carried out for 100 s at a working distance of 16 mm with a typical count rate of about 1500 counts/s. Backscattered images were obtained at a 16 mm working distance using an ETP Semra Robinson detector. CD Spectrometry Studies. Polypeptides were dissolved in 100 µM Tris-HCl at pH 7.5 to create final concentrations of 2, 4, 6, 8, 12, 20, and 25 µM. CD spectra were obtained using an AVIV 60 CD spectrometer running 60DS software version 4.1t. Samples were scanned from 185 to 260 nm at 20 °C using a 0.5 nm bandwidth and a scan rate of 1 nm/s with the appropriate background buffer subtraction performed. Spectra were obtained as an average of 3 to 5 scans for each peptide sample. The mean residue ellipticity (θ) is expressed in deg cm2 dmol-1 per mole of peptide (cell path length ) 0.1 cm). Ion Trap Mass Spectrometry Studies. Polypeptides were dissolved in deionized distilled water and titrated with microliter volumes of Tris-HCl to make a stock solution (1 mM, pH 7.4). This stock solution was then diluted with deionized distilled water to a final concentration of 10 µM to which appropriate volumes of CaCl2, CdCl2, LaCl3, and EuCl3 stock solutions (each 5 M, 99.99% pure, Sigma-Aldrich, dissolved in deionized distilled water) were then added to create peptide/metal ion ratios of 1:10 for each multivalent metal ion series. Ion trap MS experiments were performed on an Agilent LC/MSD 1100 equipped with an electrospray ionization source using the following procedure: 1000 µL of each sample solution was manually injected directly into the nebulizer at a rate of 10 µL/ min using a N2 nebulizing gas pressure of 14 psi, with a dry N2 gas flow of 6 L/min at a temperature of 330 °C. Positive ionization mode was employed with a skimmer voltage of 40 V and a capillary exit voltage of -200 V.
Results and Discussion As noted elsewhere,15,16 our use of the Kevlar calcium carbonate crystal growth assay is not meant to be a reproduction of the aragonite formation process in the nacre layer but is merely a comparative tool to note the similarities and differences in polypeptide functionality under controlled conditions. As shown in Figure 2, calcium carbonate crystals grown in the presence of the AP7N polypeptide exhibit typical step edge interruption or frustration.16 Here, we observe comparably sized rhombohedral calcite crystals that feature the {104} Miller plane and possess interruptions in crystal growth as well as rounded step edge regions. In contrast, crystal growth assays conducted in the presence of AP7NN also feature the {104} planes but do not feature the same degree of interruption. Instead, we primarily observe surface irregularities and minor interruptions in crystal growth. At higher magnification, these surface irregularities are revealed to be fingerlike extensions, similar to those reported for calcite growth interruption in the presence of nonspecific proteins such as bovine serum albumin.15 From this, we conclude that the global Asp f Asn substitutions within the AP7N sequence at positions 1,2, 21, and 22 have a significant impact on the ability of the polypeptide sequence to interact with calcium carbonates and block the growth of calcite.7,11,16 Hence, Asp residues at these locations are required for full AP7N function. However, we still detect some residual crystal adsorption capability that is retained within the AP7NN sequence despite the loss of all side-chain carboxylate residues. Compared to AP7NN, the situation for n16NN is more complex (Figure 3). For native n16N, we observe that the forming calcite
Letters
Figure 2. Scanning electron microscopy images of in vitro Kevlar calcium carbonate assay systems. (A) Negative control assay, which features typical rhombohedral calcite crystals; (B) AP7N, 100 µM; (C) AP7NN, 100 µM; and (D) high-magnification view of image C revealing interrupted growth. Scale bar dimensions are provided for each image. Note that in these assays the exposed Miller planes of the rhombohedral calcite crystals in the negative control and in the experimental samples are the {104}.
Figure 3. Scanning electron microscopy images of in vitro Kevlar calcium carbonate assay systems. (A) Negative control assay; (B) n16N, 100 µM; (C) n16NN, 100 µM; and (D) the same parameters as for image C but at higher magnification to reveal surface texture and deposits. Scale bar dimensions are provided for each image. Note that in these assays the exposed Miller planes of the rhombohedral calcite crystals in the negative control and in the experimental samples are the {104}. However, in the n16N sample, smaller facets are also noted, and the correct Miller plane designations for these smaller facets are not known at present.
crystals feature the {104} Miller plane but lose their distinct rhombohedral edges and instead feature multiple step edges where
Langmuir, Vol. 23, No. 24, 2007 11953
Figure 4. (A) Backscattering SEM image of a calcite crystal and associated biofilm grown in the presence of 100 µM n16NN. (B) EDX spectrum taken of a typical n16NN-generated free biofilm region (denoted B in image A); and (C) EDX spectrum taken from the surface of a typical crystal-coated biofilm region that exists on top of a grown calcite crystal (denoted C in image A).
both the acute and obtuse surface steps appear to be affected, leading to the formation of “staircase structures” that have small terrace sizes.15 This activity is totally absent from the n16NN assays (Figure 3), where rhombohedral calcite crystals are observed to retain their {104} planes and corner features and appear to have approximately the same dimensions as those observed in the negative controls. Hence, like AP7N, the normal mineralization activity of n16N requires the participation of carboxylate amino acids. However, we also observe n16NN-generated novel effects in our assay systems: the formation of biofilm deposits or coatings over rhombohedral calcite crystals (Figure 3). These coatings or films were noted on the majority of crystals that underwent growth during the assay period, and at higher magnification, we note the presence of rounded deposits or clusters within these coatings (Figure 3). The actual thickness or continuity of these biofilms is not yet known. As shown in Figure 3, we can distinguish where the biofilms overlay on the crystal surfaces and where they are unattached and extend beyond the crystal surfaces. The presence of these coatings on normal-sized rhombohedral crystals suggests that the crystals were allowed to grow and then somehow became coated during the assay period by these biofilms. These results indicate that the Asp f Asn substitutions at positions 17, 21, 24, and 27 and Glu f Gln substitutions at position 19 not only led to the abolishment of normal n16N mineralization activity but also created a new activity, that of polypeptide biofilm or coating formation. We analyzed the composition of the n16NN-generated biofilms using EDX and SEM backscattering techniques (Figure 4). The backscattering image shows that both the free and the crystalcoated biofilms17 are electron-dense, suggesting that the biofilms themselves are calcified either within the film interior and/or at (17) For EDX analysis purposes, we are distinguishing regions of the biofilm in terms of whether a particular region of the film is unattached to the crystal (i.e., “free” biofilm) versus the region that lies on top of an underlying calcite crystal (i.e., “crystal-coated” biofilm). This is shown in Figure 4. When referring to EDX of the free biofilm, we simply mean that to say that EDX of the biofilm was taken by isolating a location where the biofilm did not have crystal underneath or proximal to it (region B). EDX of the crystal-coated biofilm indicates that the spectrum was taken in a region where the crystal had a distinct coating of this film on it (region C).
11954 Langmuir, Vol. 23, No. 24, 2007
Figure 5. CD spectra of 12 µM AP7N, AP7NN, n16N, and n16NN (pH 7.5) apo forms.
the film surface. The presence of biofilm calcium mineralization is also supported by the EDX spectra obtained for both the crystalcoated and free biofilm surface regions (Figure 4). Here, we note that the free biofilm region possesses a very high carbon (C) content, along with above-background levels of calcium (Ca, near 4 keV) and sulfur (S). The C and S peaks most likely originate from the polypeptide (note that n16N and n16NN have 3 Cys residues/30 AA, Figure 1), whereas the calcium peak is most likely calcium carbonate mineral in origin. In contrast, the corresponding EDX spectrum for the crystal-coated biofilm possesses the expected high-intensity Ca peak (near 4 keV), and, atypically, a smaller-intensity S peak not found in negative control crystals (data not shown), which we attribute to crystal-adsorbed n16NN. Thus, we conclude that both the free and crystal-coated regions of the biofilms contain n16NN. Given the tendency of the n16N sequence to self-assemble in solution,15 we believe that the mineralized biofilms or coatings that form in these assays resulted from the self-association of the modified n16NN polypeptide. Because of the fragile nature of the biofilms and their close attachment and proximity to crystal surfaces, we were unable to obtain useful electron diffraction data for these films. Thus, the actual calcium carbonate phase or polymorph that exists within or on the surface of the biofilms is not known at present. The global substitution of Asp f Asn and Glu f Gln in both sequences not only has an impact on polypeptide-mediated mineralization activity but also affects other important features such as secondary structure and ion binding capabilities. As shown in Figure 5, the CD spectrum for AP7N polypeptide at pH 7.5 exhibits a broad negative ellipticity band (π-π* transition, 200 nm), and this spectral feature is consistent with an unfolded structure that consists of random coil structure in equilibrium with turn, extended, loop, and polyproline type II or other labile structures.15,16 However, the spectrum for its counterpart, AP7NN, exhibits three ellipticity bandssa positive band at 190 nm and two negative bands at 208 nm (π-π*) and 216 nm (n-π*) that are signature ellipticities for R-helical structure.18 Similarly, the CD spectra for n16N possesses a broad negative ellipticity band (π-π*) at 203 nm, which indicates that this polypeptide exists in equilibrium between random coil and β-strand conformation,15 whereas the n16NN CD spectrum exhibits a similar band at 198 nm that represents a blue shift toward a more random coil state.15-18 In both cases, the global replacement of anionic residues (18) Kulp, J. L.; Shiba, K.; Evans, J. S. Langmuir 2005, 21, 11907-11914.
Letters
resulted in structural perturbation in both polypeptides, with AP7NN forming a more stable secondary structure (R helix) and n16NN adopting a more unstructured state (random coil). One would expect that the loss of normal mineralization activities in AP7NN and n16NN may be partially attributable to the loss of Ca(II) binding sites (i.e., Asp and Glu) in both sequences. This in turn may affect polypeptide-mineral adsorption and ion clustering activities that are crucial to normal mineralization functionality in AP7N and n16N.11,15,16 To confirm this possibility, we performed ion trap mass spectrometry-based metal ion mass shift experiments19 with AP7N, AP7NN, n16N, and n16NN in the presence of Ca(II) and Ca(II) metal ion analogs, Cd(II), La(III), and Eu(III) (Supporting Information). Here, we found that neither AP7NN nor n16NN forms a peptide/metal ion adduct under the same conditions that AP7N and n16N do. What we note instead is the formation of numerous peptide/water adducts, which we attribute to the presence of additional hydrogen bonding donor/acceptor amino acids (Asn, Gln) in each sequence.19 Thus, the Asp f Asn and Glu f Gln substitutions abolish the ability of both AP7NN and n16NN to form polypeptide-metal ion complexes. The loss of these ion-binding sites has a negative impact on the ability of both substituted polypeptides to modulate mineralization in vitro. More importantly, because n16NN does not possess anionic carboxylate sites for dense Ca(II) attraction and chelation (Figure 1) and is unable to form complexes with Ca(II) ions (Supporting Information), the SEM backscatter and EDX spectra that we see for the n16NN-generated biofilm (Figure 4) cannot arise from peptidebound Ca(II) complexes but must arise from the presence of a calcium carbonate mineral phase that somehow formed within or on this biofilm. In conclusion, we have synthesized and tested global Asp f Asn and Glu f Gln variants of the nacre-associated AP7N and n16N mineral modification polypeptides (Figure 1). We have observed that these global substitutions alter the normal in vitro mineralization activities associated with both sequences (Figures 2 and 3).11,15,16 These results point to the crucial participation of Asp and Glu carboxylate groups in the nucleation and crystal growth processes in which both sequences participate.7,8 More importantly, our studies reveal that these anionic residues also play a crucial role in defining the polypeptide conformations of AP7N and n16N. In the case of AP7N, Asp f Asn substitutions introduced stability into the polypeptide backbone and induced folding (Figure 5), and this indicates two things. First, the result demonstrates that Asp residues and electrostatics play an important role in defining the native unfolded, unstable conformation of this sequence region.7,11,16 Second, it may be that the AP7N sequence possesses inherent helical structure that may become stabilized when confronted with appropriate external influences, such as a mineral surface. In the case of n16N, Asp and Glu residues also play an important role in defining the structure of n16N,15 but with a twist. Here, Asp f Asn and Glu f Gln substitutions induced a destabilization of the n16N backbone structure, as evidenced by a conformational shift toward a more random coil or unfolded state (Figure 5). Clearly, our study demonstrates that the role of carboxylate residues in biomineralization sequences goes beyond that of simply being sites for metal ion or crystal surface interactions. In fact, these residues are prime determinants of structure, and we advise that any future endeavors in this area should take this into consideration. However, the fact that these amino acid substitutions led to unexpected conformational alterations in both AP7N and n16N (19) (a) Collino, S.; Evans, J. S. Biomacromolecules 2007, 8, 1686-1694. (b) Kim, I. W.; Morse, D. E.; Evans, J. S. Langmuir 2004, 20, 11664-11673.
Letters
poses an interesting problem: we do not know if the loss of mineralization activities arises exclusively from the loss of carboxylate sites on each sequence or also arises in part from the structural alterations that occurred within each sequence that, in turn, may have somehow altered polypeptide-mineral recognition and binding and/or polypeptide-mediated ion cluster assembly. The resolution of this problem will be dealt with in subsequent studies. The intruguing phenomenon of n16NN polypeptide biofilm formation within in vitro mineralization assays was an unanticipated result, and there is strong evidence that this biofilm is actually a composite of calcium carbonate mineral and polypeptide components (Figure 4). Because the n16N polypeptide has exhibited some tendency to self-assemble in solution,15 we surmise that the replacement of Asp and Glu residues, along with the resulting conformational shift in polypeptide structure (Figure 5), has somehow altered the self-associative capabilities of the modified n16NN sequence. Unfortunately, we do not know what nonbonding interactions are acting as the driving force for either n16N or n16NN polypeptide-polypeptide self-association, why the replacement of carboxylate groups by amide groups triggered this self-association, or how or why the resultant biofilm
Langmuir, Vol. 23, No. 24, 2007 11955
mineralized in the absence of polypeptide carboxylate groups. Additional studies are now underway to explore the self-assembly characteristics of both n16N and n16NN and to establish the nature of the mineralized phase that is part of this biofilm assembly. It is likely that polypeptide biofilm assemblies that can mineralize may serve as interesting models not only for understanding the biomineralization process but also for developing new strategies in materials technology. Acknowledgment. We thank Dr. Il Won Kim for his assistance in obtaining SEM images. This work was supported by funding from the Department of Energy (DE-FG0203ER46099 to J.S.E.) and represents contribution number 38 from the Laboratory for Chemical Physics, New York University. Supporting Information Available: Ion trap mass spectra (positive ionization mode) of AP7N, AP7NN, n16N, and n16NN polypeptides in the presence and absence of a 10-fold molar excess of CaCl2, CdCl2, LaCl3, and EuCl3, along with tabulated adduct species. This material is available free of charge via the Internet at http: //pubs.acs.org. LA702113X
