Feb 4, 2014 - The attractiveness is based on a moderate system cost, high lateral ..... High brightness (defined as product of molar extinction coefficient and ...... HMBA linker (Sigma) using HATU/HOAt coupling, as outlined in Protocol 3.
FACULTY OF SCIENCE UNIVERSITY OF COPENHAGEN
PhD Thesis Simon Gregersen
Fluorescent Peptide-Stabilized Silver-Nanoclusters A Solid-Phase Approach for High-Throughput Ligand Discovery
Academic advisors:
Professor Knud J. Jensen Associate Professor Tom Vosch
Submitted: February 4 t h 2014
Institutnavn:
Kemisk Institut
Name of department:
Department of Chemistry
Author:
Simon Gregersen
Titel og evt. undertitel:
Fluorescerende, peptid-stabiliserede sølv-nanoclusters: En fast-fase tilgang for high-throughput opdagelse af ligander
Title / Subtitle:
Fluorescent, peptide-stabilized silver-nanoclusers: A solid-phase approach for high-throughput ligand discovery
Subject description:
Nanobio Science Design, synthesis and characterization of fluorescent, peptide-stabilized silver-nanoclusters. Methodical development for high-throughput ligand discovery
Academic advisors:
Professor Knud J. Jensen Associate Professor Tom Vosch
Submitted:
February 4th 2014
Number of pages:
171
Number of Appendices:
7
ISBN: 978-87-91963-37-7
PREFACE AND ACKNOWLEDGEMENTS This dissertation is the result of a Ph.D. study under supervision of Knud J. Jensen and Tom Vosch at the Department of Chemistry, Faculty of Science, University of Copenhagen, Denmark. The work was carried out in the period from December 15th 2010 to January 29th 2014, and the dissertation is in partial fulfilment of the requirements from the Doctoral School at the Faculty of Science, University of Copenhagen.
The general theme of the dissertation is nanobio-science with focus on design, synthesis and characterization of peptide-stabilized, fluorescent, silver-nanoclusters. The reader is expected to have basic knowledge on principles of fluorescence, peptide structure and function, basic methodology in optical and physical characterization. Prior knowledge within these fields will substantially increase the output from reading this dissertation A full description of experimental procedures and protocols can be found in Appendix A. Experimental conditions and protocols are only briefly described in the text. A full bibliography can be found at the end of the dissertation. Alphabetically ordered lists of abbreviations, chemicals, and symbols can be found prior to the introduction. “This work” refers to the Ph.D. study in its entirety, while “this/the text” refers to central part of the dissertation (main chapters not including appendices).
I would like use opportunity to thank a number of people for their help and support. Firstly, I would like to acknowledge the guidance from my supervisors, Knud Jørgen Jensen and Tom Vosch. They allowed me great freedom for scientific development and shaping the scope of the project, but without them it would never have been successful. I would also like to thank all group members that I have crossed paths with over the last three years. Especially to Kasper K. Sørensen and Søren L. Pedersen for their priceless guidance and assistance regarding peptide synthesis, as well as the Mikkel B. Thygesen for his valuable inputs for the project. I would like to thank Christian T. Hjuler, Klaus Villadsen and Søren Blok for all the non-work related conversations – primarily regarding football. This made the days seem shorter although they often lasted way into the dark hours of night. I would also like to thank Peter W. Thulstrup and Thomas J. Sørensen for making their CD spectrometer and fluorescence plate reader available.
On a personal level, I would like to thank my family - especially my mother Grethe, my father Jens, and my brother Rasmus - for always supporting my interest in science and life. Additionally, thanks to my mother in-law for providing an alternate office-space in her living room – and a lot of coffee. Thanks to all my friends and acquaintances in my hometown of Aalborg which I hold very dear. Last but not least, a large appreciation goes to my girlfriend Stine for putting up with three years of high stress-levels and commuting between Aalborg and Copenhagen. I could not have done it without you.
Simon Gregersen
TABLE OF CONTENTS Preface and Acknowledgements ............................................................................................................................................... i Table of Contents............................................................................................................................................................................ ii Summary (English) ....................................................................................................................................................................... iv Resumé (Dansk).............................................................................................................................................................................. v List of Abbreviations .................................................................................................................................................................... vi List of Chemicals .......................................................................................................................................................................... vii List of Symbols ............................................................................................................................................................................. vii 1.
Introduction ............................................................................................................................................................................ 2 1.1. Fluorescence Imaging ................................................................................................................................................... 3 1.2. Two-Photon Excitation Fluorescence Microscopy............................................................................................ 4 1.3. Super-Resolution and Single-Molecule Techniques ......................................................................................... 7 1.3.1.
Patterned Excitation ........................................................................................................................................ 8
1.3.2.
Single-Molecule Localization ........................................................................................................................ 8
1.4. Fluorophores in Imaging and Diagnostics......................................................................................................... 10 1.4.1.
Organic Dyes..................................................................................................................................................... 11
1.4.2.
Quantum Dots .................................................................................................................................................. 13
1.5. Noble Metal Nanoclusters ........................................................................................................................................ 17 1.5.1.
Physical Models for NC Photophysical Properties............................................................................ 21
1.5.2.
State of the Art in Ag-NC Synthesis and Application ....................................................................... 29
1.5.3.
Peptides and Proteins as Ligands ............................................................................................................ 32
2.
Problem Statement ........................................................................................................................................................... 36
3.
Solid-Phase Peptide Synthesis: A Novel Approach in Ag-NC Synthesis ...................................................... 38 3.1. The Solid-Phase Principle......................................................................................................................................... 38 3.2. Combinatorial Libraries ............................................................................................................................................ 42 3.2.1.
The “One-Bead-One-Compound” Approach ........................................................................................ 43
3.3. Experimental Design .................................................................................................................................................. 45 4.
Proof of Concept ................................................................................................................................................................. 50
5.
Methodical Development ............................................................................................................................................... 60
6.
Library Design..................................................................................................................................................................... 66
7.
Library Synthesis and Screening ................................................................................................................................. 70 8.1. Sequence and Structure Analysis .......................................................................................................................... 85 8.2. Tentative Design Guidelines.................................................................................................................................... 89 ii
9.
Conclusions .......................................................................................................................................................................... 90
10.
Perspectives .................................................................................................................................................................... 92
Appendix A: Experimental ....................................................................................................................................................... 94 a)
General Peptide Synthesis, Cleavage, and Work-up ...................................................................................... 94
b)
Library Synthesis and Work-up ............................................................................................................................. 95
c)
Peptide Analysis and Purification ......................................................................................................................... 96
d)
Silver Nanocluster Synthesis in Solution ........................................................................................................... 97
e)
Silver Nanocluster Synthesis on Resin................................................................................................................ 97
f)
Purification of peptide-Ag-NCs .............................................................................................................................. 98
g)
MALDI-TOF Mass Spectrometry ............................................................................................................................ 98
h)
UV/Vis Absorption Spectroscopy.......................................................................................................................... 98
i)
Fluorescence Spectroscopy ..................................................................................................................................... 98
j)
Fluorescence Microscopy ......................................................................................................................................... 99
k)
Circular Dichroism Spectrometry .......................................................................................................................100
l)
Analysis and Interpretation of CD Spectra ......................................................................................................100
m) Peptide Structure Prediction and Modelling ..................................................................................................101 Appendix B: Detailed MALDI-TOF MS Analysis of Once-Purified Peptides.......................................................102 General Impurities/Adducts and Alternate Explanations ..............................................................................105 Appendix C: Characterization of P-Ag-NCs .....................................................................................................................107 Appendix D: Detailed MALDI-TOF MS Analysis of Purified P-Ag-NCs.................................................................123 Appendix E: Library Design ..................................................................................................................................................129 Appendix F: Structure Prediction and Peptide Modelling........................................................................................133 Appendix G: Analysis of CD Spectra...................................................................................................................................147 References ....................................................................................................................................................................................150
iii
SUMMARY (ENGLISH) Fluorescent probes are widely used in the fields of imaging, detection, and diagnostics, and in order to achieve methodical progress, the search for new tools is an on-going quest. Within the last few decades, few-atom noble metal nanoclusters (NCs) have gathered increasing attention due to their physical and optoelectronic properties. These include great photostability, low toxicity, small size, and tunable spectral properties. Chemical stability of noble metal NCs is, however, very low, and they only exist transiently without a stabilizing scaffold. This has to date been done in solution using for instance small molecules, DNA oligomers, and proteins.
Peptides are an intriguing class of biomolecular ligands, due to the large combinatorial space these provide. Furthermore, as peptides have a propensity to fold up into well-defined and somewhat rigid secondary structures, they may serve as excellent ligands in the synthesis of fluorescent NCs. To date, this class of ligands remains practically unexplored and consequently, not much is known on composition of good peptide ligands for this application.
For the first time, we show that employing a solid-phase methodology in this context can increase throughput dramatically with regards to discovery of novel ligands. Our approach employs Fmoc solid-phase peptide synthesis on a PEGA resin which allows for on-resin screening of peptide ligands which, in turn, removes the tedious and labor-intensive work-up of synthesized peptides. The method allows for on-resin formation of peptide-stabilized AgNCs in a reversible manner, which makes identification of novel lead compound from combinatorial peptide libraries possible with a few simple steps. This resulted in the discovery of at least one promising candidate (P262) showing brighter emission, spectral homogeneity, and better chemical stability than seen for many Ag-NCs published to date.
By physical and optical characterization, we investigate the composition as well as the mechanism for formation of peptide-stabilized fluorescent Ag-NCs, which indicates that the process includes a dynamic folding/reorganization of the peptide to facilitate NC formation. Following an initial chelation involving the thiol-functionality of cysteine side-chains, the coordination of silver into a defined NC is expected to be the driving force of the folding process. This work also illustrates the shortcomings of MALDI-TOF mass spectrometry for the determination of solution-phase composition of organic-silver complexes.
iv
RESUMÉ (DANSK) Fluorescerende prober er brugt i vid udstrækning indenfor bio-imaging og diagnostik. For at sikre at disse metoder udnyttes til deres fulde potentiale, er søgningen efter og udviklingen af nye værktøjer en proces som er yderst vigtig. Indenfor de sidste par årtier har nanoclusters (NCs) med få atomer af et ædelmetal fået øget opmærksomhed grundet deres fysiske og optoelektroniske egenskaber. Disse inkluderer god fotostabilitet, lav toksicitet og justerbare spektrale egenskaber. Den kemiske stabilitet af ædelmetal NCs er dog meget lav og de eksisterer kun kortvarigt uden stabiliserende ligander. Syntesen af disse er hidtil foretaget i opløsning med for eksempel små molekyler, DNA oligomerer og proteiner som ligander.
Peptider er en spændende klasse af biomolekylære ligander grundet de vide muligheder de besidder. Derudover kan peptider folde sig til veldefinerede sekundære strukturer og kan derfor være fantastiske ligander i syntesen af fluorescerende NCs. Hidtil er denne klasse af ligander nærmest uudforsket og som følge heraf vides der på nuværende tidspunkt ikke meget omkring sammensætningen af gode peptid ligander til lige netop dette formål.
Ved at anvende en fast-fase metodologi, viser vi for første gang hvordan man kan øge effektiviteten i opdagelsen og udviklingen af peptid ligander til dannelsen af fluorescerende NCs. Vores tilgang udnytter Fmoc fast-fase peptid syntese på en PEGA resin, der tillader evaluering af peptiderne på resinen og dermed eliminerer en langsommelig og krævendende oprensning af peptiderne. Denne metode muliggør syntesen af peptid-stabiliserede sølv-NCs (Ag-NCs) på en reversibel måde. Reversibiliteten muliggør videre identifikationen af positive kandidater fra kombinatoriske biblioteker af peptider. Dette har resulteret i opdagelsen af mindst en lovende kandidat (P262) der viser kraftigere fluorescens, ensartethed i spektrene og bedre kemisk stabilitet end set for mange af de AgNCs der forekommer i litteraturen.
Gennem fysisk og optisk karakterisering har vi undersøgt sammensætningen såvel som mekanismen for dannelsen af peptid-stabiliserede Ag-NCs. Dette indikerer at processen inkluderer en dynamisk foldning eller omorganisering af peptiderne for a fremme dannelsen af clusteren. Processen påbegyndes ved at sølv binder til thiol-funktionaliteten i sidekæderne fra cystein, hvorefter koordinationen af sølv driver foldningsprocessen. Ydermere illustrer arbejdet også hvilke mangler MALDI-TOF masse spektroskopi har i forbindelse med bestemmelsen af organiske sølv komplekser i opløsninger.
v
LIST OF ABBREVIATIONS AA: Ag-NC: Ag-NP: Au-NC: BAL: BSA: CD: CT: Cy: DFT: DNA-Ag-NC: DPX: ESI: EXAFS: FLIM: FPALM: FRET: FWHM: GM: HOMO: HPLC: HSA: ICG: ITC: IR: LC/MS: LMCT: LUMO: MALDI-TOF: MPE: MRI: MS:
MT: MWCO: NC: NIR: NM: NMR: NP: OBOC: OD: OPE: P-Ag-NC: PALM:
Amino Acid Silver NanoCluster Silver NanoParticle Gold NanoCluster Backbone Amide Linker Bovine Serum Albumin Circular Dichroism Computed Tomography Cyanine Density Function Theory DNA-Silver NanoCluster complex Model Peptide X (Dickson Group [1]) ElectroSpray Ionization Extended X-ray Absorption Fine Structure Fluorescence Lifetime Imaging Fluorescence PhotoActivation Localization Microscopy Fluorescence Resonance Energy Transfer Full-Width at Half-Maximum Göppert-Meyer (unit) Highest Occupied Molecular Orbital High-Pressure Liquid Chromatography Human Serum Albumin IndoCyanine Green Isothermal Titration Calorimetry InfraRed HPLC/ESI-MS Ligand-to-Metal Charge Transfer Lowest unoccupied Molecular Orbital Matrix-Assisted Light Desoprtion /Ionization Time-Of-Flight Multi-Photon Excitation Magnetic Resonance Imaging Mass Spectrometry
PEG: PET: PS: PSF: QD: QY: RESF: RF: RGB: SAM: SCAL: SIM: SPPS: SPR: STED: STORM: S/N: TD: TPACS: TPE: TPM: UV: Vis:
vi
MetalloThionein Molecular Weight Cut-Off NanoCluster Near-InfraRed Noble Metal Nuclear Magnetic Resonance NanoParticle One-Bead-One-Compound Optical Density One-Photon Excitation Peptide-Silver NanoCluster complex PhotoActivation Localization Microscopy PolyEthylene Glycol Positron Emission Tomography PolyStyrene Point-Spread Function Quantum Dot Quantum Yield Red-Edge Shift Factor Radio Frequency Red-Green-Blue Self-Assembled Monolayer Safety-Catch Amide Linker Structured Illumination Microscopy Solid-Phase Peptide-Synthesis Surface Plasmon Resonance Stimulated Emission Depletion STochastic Optical Reconstruction Microscopy Signal-to-Noise Time-Dependent Two-Photon Absorption CrossSection Two-Photon Excitation Two-Photon Microscopy UltraViolet Visible
LIST OF CHEMICALS Boc: DCC: DCM: DIC: DIPEA: DMAP: DMF: EDT: Fmoc: HATU: HBTU: HF: HMBA: HOAt: HOBt: NMP: tBu: TES: TFA: TIS:
N-(tert-butoxycarbonyl) Dicyclohexylcarbodiimide Dichloromethane N,N’-Diisopropylcarbodiimide N,N-Diisopropylethylamine 4-Dimethylaminopyridine Dimethylformamide 1,2-Ethanedithiol 9-Fluorenylmethoxycarbonyl N-[(dimethylamino)-1H-1,2,3-triazolo-[4,5-b]pyridin-1-yl-methylene]-N-methylmethanaminium hexafluorophosphate N-oxide N-[(1H-benzotriazol-1-yl)(dimethylamino)methylene]-N-methylmethanaminium hexafluorophosphate N-oxide Hydrofluoric acid Hydroxymethylbenzoic acid 1-Hydroxy-7-azabenzotriazole 1-Hydroxybenzotriazole N-methyl-2-pyrrolidone tert-butyl Triethylsilane Trifluoroacetic acid Tri-iso-propylsilane
LIST OF SYMBOLS A: c: EFermi: f: fp: h: l: N: NA: nA: Pawe: Q: ελ: λ: λem:
:
Absorbance Speed of light in vacuum Fermi energy Frequency Pulse frequency Planck’s constant Path length Number of Ag atoms in Ag-NC core Numerical aperture Photons absorbed per pulse Average power Quantum yield Extinction coefficient (at wavelength λ) Wavelength Emission wavelength
λex: : λPA: λSTED: λTPE: S0 : S1 : S2 : σ1 : σ2 : T1 : Tc: τ: τp:
vii
Maximum emission wavelength Excitation wavelength Maximum excitation wavelength Wavelength of photo-activation Wavelength of STED pulse Wavelength of maximum TPE Electronic ground state First excited state Second excited state Absorption cross-section (one-photon) Two-photo absorption cross-section Excited triplet state Chelation time Fluorescence lifetime Pulse length
1
1.
INTRODUCTION
Optical imaging using light microscopy is intuitively a very attractive technique for e.g. medical diagnostics and basic science. The technique has found applications in a wide range of investigations ranging from a macroscopic to a molecular level [2-5] as well evaluation of drug efficiency and toxicology [6-8], just to name a few. The attractiveness is based on a moderate system cost, high lateral resolution (down to ~300nm), as well as ease of handling compared to other modern imaging techniques [9-12]. Although providing non-invasive and non-ionizing visualization both in vitro and in vivo, optical imaging of biological samples is impeded by strong scattering and absorption of tissue (and its constituents; e.g. water, lipids, and hemoglobin) in most of the visible (Vis) range [13, 14]. Light transport in biological tissue has been subject to extensive theoretical and experimental investigation by many groups in order to understand and explain it [15-19]. The extent of absorption and scattering varies greatly between specific tissue types and is furthermore highly wavelength- and power-dependent. Due to the complexity in composition of biological tissues, formulation of a generalized model is impossible. Moving towards the near-infrared (NIR) range of the spectrum, a region appears where biological samples and tissue only absorb and scatter light to a smaller degree. This region is generally referred to as the “biological transparency window” or the ”optical window”. There is, however, some disagreement on the range of the window [20-32], but in is generally placed in the range from 650nm (±50nm) to 1350nm (±50nm). An increased penetration depth of light in this region, as illustrated for human skin and mucous tissue in Figure 1, allows for deeper in vivo imaging and characterization of larger samples, than in the 1000
~85
1,23*105(8)
Si
300-700
400-900
~100
0,2-10*105(9)
Table 1: Optical properties of selected organic dyes and quantum dots for comparison. (1)Measured in ethanol. (2)Measured in dioxane. (3)Measured in methanol. (4)Measured in water. (5)Measured in PBS. (6)In quantum dots, optical properties (first excitonic band, emission maximum, and quantum yield) can be tuned according core diameter/shape and shell thickness/material. Additionally, ε can change up to an order of magnitude with size and solvent. An estimated range has been listed. (7)Measured at 350nm – not at the first excitonic band. (8)Measured in chloroform. (9)Measured in octanol.
1.5. NOBLE METAL NANOCLUSTERS Within the last decade, a novel class of fluorescent dyes has gathered increasing attention. This class of dyes are comprised of few-atom noble metal (NM) nanoclusters (NCs) and have gathered increasing attention due to their promising optoelectronic properties [155-160]. The increased attention is especially caused by the superior properties of the NCs compared to conventional fluorescent probes such as organic dyes and quantum dots. These include great photostability, low toxicity, small size, and tunable, optical properties.
The optoelectronic properties of NMs change dramatically with decreasing size. In the bulk, NMs are good conductors due to the continuous density of states and homogenous distribution of free electrons
17
in the valence band, giving them the character of metals. At sizes comparable to or below the electron mean free path, e.g. ~25 nm in nanoparticles (NPs) [161-163], the motion of conduction electrons is restricted by the particle size. The mean free path describes the (average) distance that a particle travels between successive collisions which influence its direction and/or energy. Size-reduction to this regime gives rise to phenomena such as surface plasmon resonance (SPR), as electrons primarily interact with the particle surface. In plasmonic NPs, optical properties are governed by collective oscillations of the conduction electrons resulting from the interaction with incident light. Plasmonic particles and nanostructure are known to be highly absorbing, but are typically non- or weakly fluorescent [164-167].
Figure 11: Schematic illustration showing size-related effects of optoelectronic properties in noble metals. Whereas bulk metal and metal nanoparticles show continuous bands of energy levels, the reduced size of the clusters results in discrete transitions and ultimately; fluorescence emission. Reprinted from [155]
Decreasing the size further, to a regime comparable to the Fermi wavelength of an electron (~0,7nm [163, 168, 169]), leads to breaking of the continuous band structure. As a result, the metal loose the conduction properties due to energy-level separation, and furthermore, these few-atom NCs have discrete energy levels and become “molecular species” or “pseudo-atoms” displaying bright fluorescence emission [155, 170]. In contrast to QD, the reduced size makes collective oscillations from conduction electrons impossible due to quantum confinement, which in turns removes the
18
plasmonic properties in NCs, as Mie’s theory no longer applies [171-176]. Interactions with incident light are now governed by direct electronic transitions due to quantum confinement. A schematic representation showing the differentiated properties of metals with decreasing size can be seen in Figure 11. In general, many of the same theoretical considerations apply when considering ether goldor silver NCs. However, Ag-NCs are of primary interest, as these are regarded to exhibit brighter fluorescence emission than Au-NCs [177].
Even though fluorescence originating from NMs due to quantum confinement has been known for several decades, it was not until the development of protocols for aqueous synthesis of stable NCs a decade ago [178], that the interest accelerated. Though previously synthesized in e.g. zeolites [179183], silver oxide films [184-187], and noble gas matrices [188-190], the preparation techniques were regarded insufficient due to lack of biocompatibility in the synthesized Ag-NCs. Furthermore, if transferred to a more biocompatible scaffold, low chemical stability and relatively low brightness has limited their use in bioimaging-related applications [191-194]. In addition, chemical stability of AgNCs is generally very low, and they only exist transiently to a few hours without a rigid stabilizing scaffold. This is partly due to an increased reactivity towards oxygen, as a consequence of reduced redox potential, for Ag-NCs, in contrast to larger, colloidal Ag-NPs [192, 193, 195-198]. Furthermore, the very unstable NCs are prone to aggregate into larger, non-fluorescent nanoparticles, as well as exhibiting electron and/or atom loss in solution [194, 199]. Therefore, a proper scaffold is needed for synthesis of Ag-NCs with distinct size and high chemical stability.
Water-soluble and biocompatible, fluorescent Ag-NCs are usually formed by in situ chemical reduction [1, 200], but may also be formed using photoreduction, γ-radiation (radiolysis of water), microwaveassisted synthesis, and sonochemical synthesis [195, 201-208]. Depending on the choice of ligand and synthesis protocol, fluorescent NM-NCs have been shown to have varying optical properties [155, 209220]. The great variations in properties have been further elaborated by density functional theory (DFT) calculations, showing the influence of specific binding and oxidative state of the Ag-NCs [221]. QYs are normally in the rage of 2-40%, which is a bit lower than for QDs, but in some instances, QYs of up to almost 70% have been reported for Ag-NCs [222]. Molar excitation coefficients are in the 105-106 M-1 cm-1 range, making Ag-NCs comparable to QDs. Fluorescence lifetimes are often seen in the range of 0,1-20 ns, however triplet-state phosphorescence emission in the millisecond-range is possible [177]. Fluorescence decays cannot be generalized, as both mono-exponential and multi-exponential decays have been observed, depending on the scaffold. In general, Ag-NCs, depending on the stabilizing ligand, show great photostability (using both OPE and TPE) and chemical stability for months to several years. Ag-NC emission spectra are typically narrow and symmetrical with FWHMs in the range from 25nm to 100nm and show Stokes shifts which are often from 50nm to 150nm. Absorption and excitation spectra can be both discrete and complex as a consequence of several allowed electronic transitions within the NC. Analogous to QDs, NCs in some cases display increasing absorption towards the UV depending on the scaffold.
19
TPACSs of MN-NCs are in good agreement with predicted values from time-dependent DFT calculations [223], often ranging from ~103 GM up to ~10 6 GM for both Au- and Ag-NCs [224-226]. TPACSs are comparable to those of QDs and hence make NM-NCs very attractive candidates tor TPM. As some Ag-NCs furthermore exhibit no intermittency on experimentally relevant timescales above 10 μs [200, 227, 228], this implies their potential for single-molecule experiments [229, 230].
Property
Organic Dyes
QDs
NM-NCs
Excitation
Discrete
Increase towards UV
Discrete, multiple
Excitation FWMH (nm)
35-100
N/A(4)
N/A(4)
Emission Emission FWHM (nm)
Asymmetric Tailing 35-100
Symmetric Gaussian 30-90
Symmetric Gaussian 30-100
Stokes Shift (nm)
< 50
> 50
> 50
104-105
105-106
105-106
σ2 (GM)
10-2-103
103-10-5
103-107
τ (ns)
1-10 (mono-exp.)
10-100 (multi-exp.)
0,1-20 (mixed)
Q Size (nm)
0,5-1 (Vis) 0,05-0,3 (NIR) ~0,5
0,1-0,8 (Vis) 0,2-0,7 (NIR) 2-50*
0,03-0,5 (Vis) 0,01-0,4 (NIR) ~1
Toxicity(1)
Low to high
Potentially(5)
Presumed low(8)
Photostability
Very low to High(6) (2) moderate Low to moderate(3) High
High
High
Moderate
Intermittency
Low to moderate (Vis) Very low to low (NIR) All timescales
All timescales
< 100μs ; > 1s
Spectral Multiplexing
Possible (3 colors)
Possible (5 colors)
Possible
Lifetime Multiplexing
Possible
Unlikely(7)
Possible
ε
(M-1 cm-1)
Chemical Stability Thermal Stability
Moderate to high
Table 2: Comparison of optical and physical properties for organic dyes, quantum dots, and noble metal nanoclusters. Vis: Visible range (λem < 700nm). NIR: Near-infrared emission (λem > 700nm). (1)Toxicity varies substantially in relation to specific dye classes, and can therefore be hard to generalize. (2)Unsuitable for long-time imaging and experiments. (3)Stability can be improved by attachment of dye to certain ligands. (4)FWHM is not applicable due to the shape of QD and NC absorption/excitation spectra (see Figure 10). (5)Little is established regarding QD toxicity, but cadmium toxicity raises the concern of leakage from the core which may result in nanotoxicity. (6) Not only are QDs photostable but photobrightening is commonly observed. (7)Distinguishing QDs by their lifetimes has hitherto been deemed unlikely due to their complex and multi-exponential decays. Requires immense developments in algorithm formulation for data analysis. Distinguishing QDs from organic dyes is, however, possible. (8)Primary investigations show low toxicity, but there is need for more intense studies. Note: The listed properties have been generalized for the probe classes, but examples of properties falling beyond this generalization have been demonstrated.
20
The applicability of Ag-NC fluorescence has already been demonstrated in many instances, although primarily on an ensemble base, for the time being. Applications include amyloid fibril imaging [209], surface and nuclear labelling in both fixed and live-cells [210, 220, 231-240], in vivo tumor imaging [241-244], ion-, amino acid-, protein-, DNA-, and mRNA-sensing with detection limits in the nM-range or lower [92, 196, 202, 212, 232, 245-257], FRET assays [258-260], detection of cancerous cells [261263], and enzyme activity assays [264-266], to name a few. Additionally, Ag-NCs have been shown to have very promising properties as a single-molecule dye [200, 211, 267-269].
NM-NCs are regarded as being more biocompatible and more readily bioconjugated than QDs [158], directly implied by the biocompatibility of the scaffold, owing to their overall potential as molecular probes. Toxicity is an aspect still in dire need of investigations, but initial investigations of Au-NCs show very low toxicity as well as high renal clearance compared to NPs and QDs [270, 271]. Furthermore, Ag-NCs have been shown to exhibit no significant toxicity towards HeLa cells [237]. As a general note, it should be mentioned that in contrast to organic dyes, which have identical molecular structure and uniform spectroscopic properties from molecule to molecule, nanomaterial-based probes such as QDs and NM-NCs have limited reproducibility with properties often varying quantitatively from molecule to molecule as well as batch to batch [272]. Moreover, larger NPs are often created as a byproduct in NC synthesis [201]. In addition, many protocols produce spectrally impure NCs, where emission wavelength shifts as a function of excitation wavelength. This type of characteristic spectral shift is a result of the presence of multiple emitters [207, 220]. Great efforts are currently done to create generalized synthesis protocols resulting in monodisperse dyes with high reproducibility.
1.5.1.
PHYSICAL MODELS FOR NC PHOTOPHYSICAL PROPERTIES
Since the discovery of NM-NC fluorescence, much effort has gone into understanding the basis and origin for the phenomenon. To date, greater insight into the fluorescent properties of Au-NCs has been achieved, than is the case for Ag-NCs. There is a general agreement that the NM-NC fluorescence emission arises from a LUMO-HOMO transition, in accordance classical fluorescence theory [177]. This transition arises due to quantum confinement of electrons (and holes) in small NCs. Compared to the continuous band distribution in bulk NMs and NM-NPs due to large overlaps of the energy bands, quantum confinement splits these into discrete energy levels which, in turn, facilitates intramolecular transitions with an energy corresponding to Vis/NIR light and hence the possibility of fluorescence. Though being in the same periodic group, atomic scale properties of Ag and Au differ significantly. There are many theories as to why Au- and Ag-NCs show different physical and optical properties, but one of these relies on the considerably larger relativistic effects in gold, resulting in a small s−d energy gap. Consequently, the d-electrons contribute substantially to the lower lying excitations and thus influence the spectroscopic patterns. In contrast, in the case of silver, the s−d energy gap is much larger, and s−p type excitations lead to strongly localized, intense absorption [273, 274]. Additionally,
21
Ag-Ag interactions are usually weaker than Au-Au interactions, which might also play a role in both cluster stability and optical properties [216]. Furthermore, the role of ligands, with regards to the fluorescent properties, is not yet fully understood, but ligands play an essential role in determining stability and QY of the formed NCs [159, 201, 220].
Traditionally, the free electron Jellium model has been employed to describe the foundation for NM-NC fluorescence emission. The model proposes that the energy of the emitted photon (Eem) is determined by the Fermi energy (EFermi) of the bulk metal (5,5eV and eV for Ag and Au, respectively) but decreases with increasing number of atoms (N) in the cluster, according to Equation 4. The Jellium model originates from quantum mechanics, and regards the NC as a uniform, positively charged sphere surrounded by an electron gas, and can hence be treated as a simple, three-dimensional harmonic oscillator [275, 276]. Consequently, this assumption directly leads to an electronic shell structure with quantized HOMO-LUMO energy gaps within the Vis/NIR range. The model is a simplified version of liquid metal drop model, assuming an N1/3-dependent cluster radius [274, 276, 277], and directly leads to the assumption that for increased cluster size, red-shifted emission will be observed:
!
=
"# $%& '⁄(
(Equation 4)
Although experimental [278-280] and computational [281] findings have supported this theory to some extent, the model is insufficient in fully describing emission energy due to several influencing forces. In the case of Au-NCs the theory has been disproven on several occasions [282, 283] which is also the general case for Ag-NCs. However, Morozov and Ogawa [218] recently reported of fluorescent Ag-NCs showing size-dependent emission following the Jellium model, as illustrated in Figure 12. This is in sharp contrast to previously obtained result for Ag-NCs, as these do not seem to exhibit strict, size-dependent emission. In this respect, it should also be noted, that the authors only fitted three different-sized Ag-NCs to the model, which does validate the assumption.
Figure 12: Correlation of emission energy and cluster atom number (determined by ESI-MS and UV/Vis absorption-based titration) fitted to the Jellium model for three Ag-NCs stabilized by three different coiled-coil peptide complexes. Reprinted from [218].
22
When not only regarding the emissive energy expected from NM-NCs, the electronic closed-shell and odd–even effects play increasing roles in determining the dissociation energies of NCs, which also plays a role in the stability of the NM-NCs [158]. Closed-shell effects are also regarded the reason why “magic number” NC are readily formed with both Au [170, 279, 280, 284-287] and Ag [240, 288-292]. However, when considering “magic number” NCs, one has to distinguish between intrinsic core stability and ligand-induced stability, as this highly influences the “magic” in NC stability. Intrinsic core stability can be explained by means of electronic shell closing (e.g. Au8, Au18, Au20, and Au34) or geometric shell closing (e.g. Au13 and Au55), while ligand-induced stability naturally relies on the specific ligand used [293]. In the case of Au-NCs, one of the most intensely investigated NCs is the glutathione-stabilized Au25 NC, which has high stability due to complete coverage of the Au surface by 18 glutathione molecules.
In addition to closed-shell effects, odd-even effects have also been detected for both neutral and anionic gas-phase, few-atom Ag-NCs [277, 294, 295], as illustrated in Figure 13. From here, we clearly see that although following the Jellium model to some extent, even-number Ag-NCs seem to, in general, be more stable. This does, however, not directly imply that even-number Ag-NCs are also the most stable species in solution. Albeit, this was actually shown to be the case for fluorescent Ag-NCs stabilized by bovine serum albumin (BSA) [296]. In this study, Agn-S complexes (2≥n≥8) were detected by mass spectrometry (MS). One important thing to consider in this respect is that MS-based characterization only detects species that can be efficiently ionized and transferred to the gas-phase. This means that even though even-number Ag-NCs were detected in higher abundances that their oddnumbered counterparts, this could reflect their gas-phase stability or ionization potential rather than their actual abundances and the size-distribution in solution.
Figure 13: Experimental ionization energies (open circles) of gas-phase Ag-NCs as a function of atom number in the cluster. Prediction of ionization energy based on the liquid metal drop model (solid line) as well as the polycrystalline work function of Ag (4,26eV) have been plotted for comparison. Reprinted from [277]
23
Seeking understanding through the properties of bare, gas-phase Ag-NCs might be one way to gain fundamental insight into size-dependent spectral properties, as these have been well characterized [297-305]. Amongst these, Ag1-3 are the best characterized with regards to emissive properties, and collective, spectral properties of these species are shown in Table 3. From this, it is evident, that the number of atoms in the Ag-NC cannot be determined by means of specific, electronic transitions observed, due to the large overlaps and complexity. Furthermore, the complexity and large variations indicate, that geometric factors also play an important role with regards to spectral properties of AgNCs. Gas-phase emission in the red/NIR has primarily been ascribed to Ag3 [304], and one could speculate that the observed properties could be indicative of size-dependent effects. However, this does not seem to be the case as Ag4 has been reported to exhibit emission at 458nm [298] and Ag8 at 321nm [299], which contradicts the anticipated red-shifted emission with growing cluster size, as predicted by the Jellium model. The substantial deviations from the predicted emission energies can, too some extent, be supported by odd/even effect, but nonetheless, this effect seems insufficient in describing the experimental findings presented. That being said, the spectral properties of the bare, gas-phase Ag-NCs will likely not be transferrable to Ag-NCs in solution [306, 307]. At the time being, no complete, structural information has been determined for Ag-NCs in solution other than by MSbase techniques.
Table 3: Emission and absorption bands of bare, gas-phase Agn-NCs (n=1-3). Reprinted from [220].
24
In a recent paper by Schultz and Gwinn [308], further evidence of non-size-dependent emissive properties in Ag-NCs was presented. In this study, it was possible to separate individual DNAstabilized Ag-NCs from a polydisperse mixture of several species, which led to the discovery of both emissive and dark Ag-NCs. Here it was seen that spectral properties were not only determined by the number of atoms in the core, but formation of Ag-NCs with the same number of atoms could result in both dark and emissive Ag-NCs. However, it was speculated that there still was an overall sizedependency in the spectral properties, as larger NCs might be multiple, loosely bound, smaller clusters retaining their individual spectral properties while exhibiting a combined mass spectrum. Following this, the group continued work on DNA-Ag-NCs to find compelling evidence that the emission arise from a thread-like or rod-shaped Ag-NC rather than globular conformations [230]. This means that the Jellium model does not apply, as this assumes (quasi)-spherical geometries in addition to uniform charge distribution of the Ag-NC core. Furthermore, they found that the rod-shaped Ag-NCs are highly charged with roughly the same amount of neutral and cationic Ag, which also facilitates strong ionic binding to the polyanionic DNA. Differences in in spectral properties are primarily attributed to length-dependent effects as well as structural defects. One rod-shaped Ag-NC may be disrupted into disjoint pieces which, in turn, would result in two independent (at least to some extent) intrastrand emitters. These effects are primarily attributed to ligand-dependent interactions with the Ag-NC.
These findings have been further substantiated by computational studies from Ramazanov and Kononov [309], who saw that planar Ag-NCs bound to DNA had lower emitting ability compared to thread-like Ag-NCs. Furthermore, calculated excitation spectra of rod-shaped Ag-NCs show similarity to experimental excitation spectra, whereas the calculated spectra of planar Ag-NCs do not. The modelling also revealed that clusters of Agn as small as n=3, may provide a large variety of emitters due to geometric effects. Bending the structure from the linear state, can cause a strong shift (>1eV or 200nm) in the excitation band of both cationic and neutral few-atom Ag-NCs (see Figure 14), which might explain the complexity of experimentally obtained excitation spectra – especially if assuming a somewhat flexible/dynamic structure of the thread-like Ag-NC. Interestingly, they found that the excitation spectrum Ag3+ bound in the minor groove of a dC3 fragment, closely resembles the spectrum measured for a green-emitting DNA-Ag-NC [310].
25
Figure 14: Calculated excitation spectra for thread-like Ag3+1, Ag40, and Ag60 with different bending angles. Reprinted from [309].
That interactions with the DNA are of utmost importance with regards to spectral properties, has also been shown using extended X-ray absorption fine structure (EXAFS) analysis, which revealed that there is indeed Ag-N/O interactions as well as Ag-Ag interactions [311]. Strong interactions with the ligand would influence the spectral properties to a substantial extent. This has been experimentally verified by Patel et al. for DNA-stabilized Ag-NCs [312]. As previously mentioned, Ag-NCs exhibit a characteristic >10μs (independent of DNA sequence and spectral properties of the Ag-NC) fluorescence intermittency [200, 227, 228], indicative of a different mechanism than for organic dyes and QDs. As nucleotides can function as efficient electron acceptors [313-315], excited-state electron transfer was proposed as the origin of the trap-sites responsible for intermittency as well as a highly influencing factor the photophysical properties of few-atom Ag-NCs in general. A small dark state population of ~1% was determined, originating from the photoinduced electron transfer, which effectively could be depopulated through photoassisted reverse charge transfer, resulting in increased emission rates. Furthermore, the investigation led to the, to date, most complete model for describing the photophysical properties of few-atom Ag-NCs. The model (Figure 15) proposes that following an initial photoexcitation from the ground state (state 0) to an excited charge separated state (state 1), primarily a non-radiative decay back to the ground state occurs. However, a rapid separation to the emissive state (state 2) or a longer-lived dark trap-state (state 3) is also possible, which can describe the observed intermittency.
26
Figure 15: Jablonski (energy) diagram for photo-induced charge transfer mechanism in DNA-stabilized Ag-NCs. Following initial photoexcitation from the ground state (state 0) to the excited charge seperated state (state 1), a non-radiative decay back to the ground stated can occur. Furthermore, separation into the emissive state (state 2) of a longer-lived dark state (state 3) is possible. From the dark state, a nonradiative decay to the ground state or optical repopulation of the charge separated state is possible. Reprinted from [312].
Some Ag-NCs display a shift of the local emission maxima ( ) as a function of λex. The phenomenon is related to two factors: (i) the presence of multiple emitters in the sample and (ii) a large dipole moment of the ligand-Ag-NC complex, which readily interacts with surrounding dipoles (e.g. solvent molecules) by means of dielectric relaxation [316]. The latter can in some cases be regarded as the ground state having a low dipole moment while the excited state has a dramatically increased (induced) dipole moment, and has been referred to as the so-called “red-edge effect” [317]. The rededge effect is easily seen as inhomogeneous spectral broadening. The spectral broadening can be explained as even though a sample consists of merely one emissive species, many different sub-states of this species are present which can be characterized by the extent of contacts/interactions with the environment. Based on the number of contacts within each sub-state, they have different interaction energies with the surroundings (i.e. different extents of dielectric relaxation). Together, they form an ensemble of sub-sets which can be regarded as the collective spectral properties of that specific emitter in a given solvent. As such, the amount of different sub-states for the emitter as well as distribution in their interaction energies with the surroundings, determine the extent of inhomogeneous broadening for that emitter. Furthermore, the relation between time-scales for the relaxation process and the emission is a limiting factor for the extent of broadening. Usually, this is evaluated by the full-width at half-maximum (FWHM) of the emission spectrum as well as shift in . An example of this excitation-emission correlation is illustrated in Figure 16 for Ag-NCs synthesized using a polystyrene-block-poly(methacrylic acid) block copolymer as a scaffold via photoactivation [316], displaying both multiple species and red-edge effect. A large change in the
27
effective dipole moment between the ground state and the emissive state of the ligand-Ag-NC complex will also be revealed through significant solvatochromic effects, as observed in several instances for Ag-NCs synthesized using various organic scaffolds [160, 200, 213, 312, 318]. These effects describe varying dipole interactions between the ligand-Ag-NC complex and solvent as a function of solvent polarity/dielectric function which, in turn, is indicative of solvent exposure. A large change in dipole moment upon excitation indicates a significant excited state charge separation [220], which is consistent with the model proposed by Patel et al.
Figure 16: Excitation-emission correlations for Ag-NCs synthesized via photoactivation with a polystyrene-block-poly(methacrylic acid) block copolymer as scaffold. a: Emission spectra as a function excitation wavelength. b: Excitation spectra as a function emission wavelength. Both plots show three distinct species all showing extensive dipole interactions in both the ground state and first excited state. Reprinted from [316].
In a recent DFT computational study by Gell et al. on the structure and optical properties of thiolatestabilized Ag-NCs, [273] they showed that the number confined electrons in the Ag-NC is the determining factor in the spectral properties. This was done by varying the number of Ag atoms in the NC, the number of thiolate ligands, as well as their ratio. By doing the latter, the number of confined electrons can be adjusted, as sulphur is considered as an electron acceptor, thereby withdrawing valence electrons from the Ag-NC. In general, a certain number of confined electrons is associated with high UV absorption. Furthermore, the study revealed the presence of directional and rigid S-Ag-S bonds arranged in either S-Ag-S or S-Ag-S-Ag-S staple motifs, indicating a very high ligand-NC interaction. These staple motifs are analogous to those seen in larger thiolate-Au-NCs [319]. Furthermore, they pointed out the importance of Ag-ligand interactions with regards to spectral properties as being related to the excited state geometric relaxation of the complex. In contrast to the findings for DNA-Ag-NCs, the thiolate-Ag-NCs have planar and globular geometries, thereby making the two scenarios distinctively different.
28
As previously stated, the role of the ligand is not yet fully understood, but it seems evident that, especially for Ag-NCs, the influence of the ligand is substantial and in some respects is dominant over the size-dependent properties. Regarding the stability of Ag-NCs, this has been further substantiated by the recent paper from Li et al. [320]. Here, a clear correlation between the binding constant (as determined by isothermal titration Calorimetry (ITC)) and the half-time in solution for a range of DNAAg-NCs. Consequently, one could hypothesize, that three effects could be used to explain the spectral properties of Ag-NCs:
1. Structure-dependent properties determined by the shape of the Ag-NC core 2. Ligand-dependent properties determined by interactions between ligand and Ag-NC core 3. Charge state of the Ag-NC core and the resulting dipole moment of the ligand-Ag-NC complex The ligand is in this respect the critical factor, as this (in combination with ligand-to-Ag ratio) is expected dictate the size and shape of the Ag-NC through specific interactions and binding which, in turn, influences the effective dipole moment of the complex. Consequently, all three factors are highly inter-linked thereby complicating the determination of the contribution of the individual effects. Competing ligands might show to be yet another factor further complicating the understanding of AgNC fluorescence [220]. This means that in principle, it is not possible to precisely predict the spectral properties of Ag-NCs without having solid knowledge on method, ligands, and the Ag-ligand interactions. Consequently, construction of a suitable model taking all factors into account could likely prove to be a nearly impossible task.
1.5.2.
STATE OF THE ART IN AG-NC SYNTHESIS AND APPLICATION
As mentioned previously, the pioneering work by the Dickson group in 2002 [178] increased the interest and research in water-soluble, fluorescent Ag-NCs tremendously. When considering the choice of ligand for Ag-NC synthesis, several factors need to be taken into account. Among these is the synthesis route, where NM-NC synthesis generally can be categorized as being top-down or bottom-up. The two routes clearly exemplifies the general possibilities within nanotechnology [321]. They fundamentally differ in the way the Ag-NCs are produced, where the top-down approach utilizes (interfacial) etching of larger NPs by excess ligands into smaller, fluorescent Ag-NCs [177, 215, 322325]. Although highly applicable for synthesis of Au-NCs, this approach is less favorable for synthesis of Ag-NCs, and will not be treated in detail. The bottom-up approach is far more wide-spread, and utilizes direct interaction between the ligand and Ag+-ions followed by nucleation growth or coordination of bound Ag. The last step is usually facilitated by complete or partial, chemical reduction of Ag+ to Ag0.
29
In general, the bottom-up approach can be achieved using either a capping agent or a stabilizing ligand. Using small-molecule capping agents can (to some extent) be regarded as the formation of a self-assembled monolayer (SAM) on the surface of the formed NC. However, this is based on the assumption that the NM-NC in fact has a crystalline character, which is not possible for few-atom NCs below a certain size and hence only applicable for larger (globular) CNs. This approach has been used to generate highly biocompatible Ag-NCs, but also reflects the need for NCs with inherent stability due to the nature of SAMs. In general, thiol-capped Ag-NCs are much rarer than thiol-capped Au-NCs [158, 216, 326]. This is primarily ascribed to the fact that it is difficult to form SAMs on the Ag(111)-face, which is very different from the Au(111)-face [327] that, in contrast, is a very good substrate for SAM formation. Small-molecule capping of NCs usually produces high yield and monodisperse Ag-NCs, but is often impaired by low QYs and tedious purification [177]. Furthermore, the higher susceptibility of nanoscale Ag towards oxidation [220], may require more complete shielding than a flexible, dynamic SAM in many instances provide. This is further substantiated by the fact that SAMs composed of alkanethiolates [328], thioaromats [329], and other SAM-producing capping agents in general bind tighter to Au than Ag, which would make them more prone to displacement or even dissociation, thereby increasing Ag-NC vulnerability to oxidation and further agglomeration/aggregation [170, 216, 330].
Using ligands, however, is fundamentally different from the use of small-molecule capping-agents. Ligands are larger (bio)molecules which function as a template during NC formation. It is proposed, that ligands bind several Ag+-ions within the same molecule, which subsequently folds into a specific conformation with minimized energy, to form the NC. The folding process is influenced by agglomeration of Ag0 to Ag-NCs as a driving force. In contrast to the case with capping agents, this means that stability is even more dependent on the nature of the ligands and its more restricted conformation. Furthermore, due to the physical size of the ligands compared to capping agents, a reduction in ligand-to-NC ratio is achieved (often 1:1). Further reducing the ligand-to-NC ratio is achieved by synthesizing Ag-NCs in larger, organic matrices. This approach can be regarded as a type of physical entrapment and is analogous to Ag-NC synthesis in e.g. zeolites.
Another important aspect is the fact that, the specific ligand influences the physical and optical properties of the Ag-NCs. In addition to providing physical shielding from collisional quenching as well as oxidation protection [34, 158], the ligand can interact either by charge transfer through a direct bond or direct donation of a delocalized electron [283]. To date, DNA [200, 245, 246, 258, 308, 331334], polymer matrices and films [178, 202, 205-208, 213, 335, 336], thiol-containing small molecules [214, 273, 337-339], and non-thiol containing small molecules [232, 339-341], have been the preferred scaffolds. To cover them all is beyond the scope of this work, but a few general remarks and examples will be presented to illustrate the advantages and disadvantages of the different approaches. To date, few-atom DNA-Ag-NCs have been the most comprehensively characterized and computed [221]. A schematic representation of Ag-NC synthesis using an organic ligand can be seen in Figure 17.
30
Figure 17: Schematic illustration of Ag-NC synthesis using an organic ligand. The ligand (left) is mixed with Ag-ions (grey spheres) and allowed to complex. This is followed by an activation step (photoreduction or chemical reduction) to generate well-defined Ag-NCs consisting of Ag0 (orange spheres), which upon illumination emit fluorescence. Emission spectra (right) of a selection of DNA-AgNCs illustrating the tunable emission properties. Inserted images show the color of the DNA-Ag-NCs under ambient and UV light. Reprinted from [220]
DNA is generally regarded as a very efficient ligand for the preparation of fluorescent Ag-NCs. This property is mainly ascribed to Ag+-ions mediating cross-linking between bases in cytosine-rich DNA. The interaction between DNA and Ag is primarily governed by the carbonyl oxygens and doubly bonded ring nitrogens present in the heterocyclic bases. This conclusion has been proven both with experimental results [342-344] and time-dependent DFT calculations [221], showing stronger interactions of few-atom Ag-NCs with cytosine and guanine than with adenine and thymine. As Ag is also able to stabilize a cytosine-cytosine mismatch in double-stranded DNA [345-347], this further indicates the high affinity of Ag towards this nucleobase. Circular Dichroism (CD) spectrometry has been employed to reveal an apparent induced conformation change in the DNA upon Ag-NC formation [161, 337, 348-351], which also reveals an inherent chirality of template-stabilized Ag-NCs. The conformational change implies that the Ag-NC formation is indeed a dynamic process, where a structure with reduced energy for the DNA-AG complex is obtained. DNA-Ag-NCs can be tuned to have emission in the Vis/NIR range with QYs in the range from 10 to 40% [158]. A wide variety of DNA constructs have been investigated for their ability to function as templates for synthesis of bright and stable Ag-NC and in general, this ability is not only highly sequence dependent, but also depends on length and conformational rigidity of the DNA [210, 211, 217, 219, 220, 308, 320, 331, 333, 334, 352]. Noticeably, the same DNA sequence can produce different Ag-NC species which, in turn, indicates that the energy minimization obtained through the dynamic folding process does not lead to one distinct conformation but rather a series of (related) structures with reduced energy.
NM-NCs can also be used to improve the properties of intrinsic and organic fluorophores, as demonstrated by hybrid NCs composed of Ag and thioflavin T for imaging of amyloid fibrils [209]. In contrast to staining with thioflavin T alone which display the anticipated photobleaching, the hybrid NC display photoactivation (i.e. increased emission intensity over time). Furthermore, as bleaching events are eliminated, the resulting emitter density was substantially increased resulting in brighter and better resolved images. Even with an 80-fold reduction in exposure time, the hybrid NCs produced fluorescence images of much higher intensity and contrast, as illustrated in Figure 18. Unfortunately, the specific composition of the hybrid NC was not determined. A recent paper by Chang et al. [353]
31
demonstrated the use fluorescent BSA-stabilized Ag-NCs for bimodal imaging in combination with magnetic resonance imaging (MRI). This was achieved by doping the BSA-stabilized Ag-NC with the paramagnetic (and hence MRI active) gadolinium. In fact, the Gd-doping resulted in fluorescence enhancement, although the mechanism was not discussed. The use of bimodal imaging facilitates more comprehensive characterization in vivo, as two datasets are produced from the same probe, thereby increasing diagnostic effectiveness. In this experiment, however, it was established that Gd-doped AuNCs displayed brighter emission than their Ag counterparts, but the focus was primarily on the MRIrelated part of the bimodal imaging. Solid-state synthesis routes have also been employed to produce fluorescent Ag-NCs [354-356], but will not be treated further.
Figure 18: Epifluorescence imaging of amyloid fibrils with (1) Ag/thioflavin T hybrid nanoclusters (1μM/10μM, exposure time 0,02s) and (2) thioflavin T alone (10μM, exposure time 1,6s). (3) Photobleaching of thioflavin T (black) versus photoactivation of the hybrid NC (red) over 60s. Emission intensity is normalized to the initial intensity for the respective measurements. Reprinted from [209]
1.5.3.
PEPTIDES AND PROTEINS AS LIGANDS
The use of DNA as ligands in synthesis of Ag-NCs opens a vast space in ligand composition, as combinations of the four nucleobases in e.g. a 20-mer, results in ~1012 distinct oligomers. Using peptides or proteins will, however, opens an even larger space for variation. Combining the 20 natural amino acids (AAs) in a 20-mer will result in ~1026 distinct oligomers. This enormous space of variation and consequently conformation might explain why this field has been subject to very sparse investigation, but can also justify why the use of peptides and proteins is very intriguing. Seeking inspiration from nature, one paradigm employed by biological systems for nanocluster synthesis is the expression of small peptides/proteins which serve as binding templates/nucleation sites for metal ions and then serve to stabilize the nanocluster core against continued aggregation [357]. This property has been known for more than a century and utilized in e.g. silver staining of nucleolar proteins for cell visualization and electrophoresis purposes [358-362]. This procedure is, however, not biocompatible as it requires concentrations of at least 5,8M AgNO3 and high temperature treatment, and is hence only applicable for fixed cells. In 2007, Yu et al. [1] published a new approach, where fluorescent Ag-NCs were formed in situ within cells by treatment with 20mM AgNO3 and subsequent photoactivation. Ag-NCs were primarily located in the nucleus and were determined to be synthesized within the framework of the ~70kDa nucleolar protein, nucleolin. The formed Ag-NCs exhibited broad 32
emission from 500-700nm under blue excitation showing a biexponential fluorescence decay with lifetimes of 0,2ns and 1,8ns.
Though the use of proteins and peptides as stabilizing ligands has increased within the last decade, limited knowledge exist on how the interaction occurs and what constitutes a good ligand. One important factor is the fact the 20 natural AAs contain a diversity in functional groups, which likely can facilitate sufficiently strong Ag-ligand interactions. Heterocyclic nitrogens may serve as binding partners as with DNA, thiols could serve as intrastrand capping agents, and carboxylic acids provide stability through electrostatic interactions, as in polymer and dendrimers [178]. Furthermore, secondary, tertiary, and quaternary structures found in peptides and proteins could serve as a rigid and shielding scaffold for the Ag-NCs. Most work to date uses larger, globular proteins such as human serum albumin (HSA) [363], BSA [249, 251, 296], and bovine pancreatic α-chymotrypsin [364] to (presumably) encapsulate the NCs in the hydrophobic core, thereby shielding and protecting them from oxidation. The distinct binding/chelation site of the Ag-NC or the resulting structure/conformation of the protein-Ag complex was not determined and as such, this provides no substantial information as to rational design of peptide and protein ligands. In general, the synthesis of protein-Ag-NCs by chemical reduction yield probes with emission in the red with quantum yields ranging from 1% to 11% and lifetimes in the range from 1-10ns (multi-exponential) [158]. In the case of HSA-Ag-NCs, it was seen that without chemical reduction, a blue emitter formed spontaneously under alkaline conditions, through tyrosine-mediated, intrinsic reduction. What was really striking was the fact that it was possible to convert the blue emitter (determined as HSA-Ag9, Q=16%) into the red emitter (determined as HSA-Ag14, Q=11%) by chemical reduction with NaBH4 in only 5 minutes. Furthermore, it was possible to convert the red emitter to the blue emitter by chemical oxidation with H2O2 also in just 5 minutes. Both conversions yielded roughly the same optical properties as for the asprepared species. The origin of the interchangeable property was not determined, but can be likely be regarded as a result of charge/oxidative state of Ag and/or ligand in addition pH-dependent electrostatic effects. Unfortunately, there has to date been no determination of the specific Ag-NC location and binding partners within the framework of these proteins. As a consequence, it is not possible to directly use the obtained structural information for rational design of novel ligands, but a more detailed overview will follow. A general note on the use of proteins is that reduction (especially chemical) of Ag likely will reduce cysteines thereby breaking the disulfide bonds within the protein. This would result in reduced conformational stability and ultimately aggregation. Partial reconstitution may be achieved, but this might not be sufficient for biocompatibility. Additionally, the physical size of the globular proteins (>5nm) might deem them unsuitable for certain applications.
Peptides have only been exploited to an even smaller extent than proteins, but have still proven to be a suitable ligand for cluster synthesis and stabilization under physiological conditions [1, 220, 365]. As peptides are able to form secondary structure elements and a superiors tertiary structure of welldefined nature, they can serve as excellent templates/ligands for synthesis of nanomaterials with ordered size and shape, as they are capable of limiting continuous grown and hence agglomeration. This is well-illustrated by the fact that they serve as the major biological scaffold for nano-scale selfassembly [366]. A repeating consensus sequence (AHHAHHAAD) from the histidine-rich protein 11
33
originating from Plasmodium falciparum, has also been reported to facilitate Ag- and Au-NC formation [367], albeit of a non-emissive nature. Incorporation of a Cys-X-X-Cys metal binding site [368] in three helical peptides, that are well-known to self-assemble into coiled coil superstructures, also stabilized Ag-NCs with different spectral properties, based on the number of monomers in the coiled coil architecture and hence, the number of Ag atoms in the cluster [218]. The three coiled-coil-Ag-NCs were characterized and displayed red-shifted emission with increasing cluster size. Interestingly, with increasing number of monomers in the superstructure, an increase in quantum yield was also seen (QTrimer=0,25% ; QTetramer=1% ; QHexamer=4%), which could indicate increased protection. All three coiledcoil-Ag-NCs exhibited biexponential fluorescence decays with fluorescence lifetimes ranging from 2,9ns to 17,4ns. Glutathione (γ-Glu-Cys-Gly) has also been reported to be able to stabilize fluorescent Ag-NCs through capping and size focusing/etching of Ag-NPs in an top-down manner [240, 252, 323]. Interestingly, by extending the etching time (and thereby intuitively decreasing Ag-NC size), a blueshift in emission was achieved. A fusion peptide (CCYRGRKKRRQRRR) comprised of C-terminal nucleolar localization signal sequence from the HIV-1 TAT protein and an N-terminal Cys-Cys-Tyr moiety was able to both bind and reduce the silver cluster due to the reductive potential in the phenol side-chain of Tyr under alkaline pH [369]. At pH=9, a mixture of Ag5, Ag6, and Ag7 displaying cumulative emission at 432nm was achieved, while adjusting the pH=12, yielded Ag28 showing emission at 661nm. Both examples show somewhat size-dependent emission, but likely reflects the fact that the peptides were used as capping agents rather than ligands. This assumption is further substantiated by a ligand-NC ratio >1.
Following the discovery of in situ synthesis of Ag-NCs inside cells, the Dickson group [1] performed an AA analysis of nucleolin which revealed that the most prevalent AAs in the nuclear protein were Glu (15,6 %), Lys (12,7 %) and Asp (8,5 %). Using this knowledge, in addition to the implementation of several Cys residues due to its high affinity for silver, a peptide (P1: KECDKKECDKKECDK) was designed. Due to a relatively short chemical lifetime of the peptide-encapsulated, fluorescent nanocluster, two modified peptides were designed. The designs included knowledge from studies with dendrimer-encapsulated clusters [178] where hydrophobic regions facilitate nanocluster formation, as well as introduction of His. This resulted in the two peptides ((P2: HDCHLHLHDCHLHLHCDH) and (P3: HDCNKDKHDCNKDKHDCN)), which indeed showed an increased stability in solution. All three peptide-Ag-NCs (P-Ag-NCs) display red fluorescence with biexponential decays with lifetimes in the range of 0,1ns to 3ns. Interestingly, the latter peptide-Ag-NC was added to NIH3T3 cells and subsequently internalized – even at 4°C – indicating a different uptake mechanism than endocytosis. This example clearly shows that it is not only specific interactions between AA side-chains and metal that facilitates the formation and stabilization of fluorescent metal clusters, but rather a cumulative effect of several factors. In general, peptide-templated Ag-NC synthesis seems to produce spectrally pure emitters, where the emission maximum does not change as a function of excitation wavelength [218]. This property makes peptides even more appealing as ligands for Ag-NC synthesis.
For more comprehensive reviews of properties and synthesis of NM-NCs, the reader is referred to a number of recent reviews with different foci [155-159, 177, 201, 216, 232, 293], where especially the review by Choi et al. [220] has specific focus on Ag-NCs.
34
35
2.
PROBLEM STATEMENT
As described in the introduction, Ag-NCs are a class of intriguing fluorescent probes with very desirable physical and optical properties. Stabilizing ligands are essential for synthesis and stability at physiological conditions, but is still in need of improvement. For this, peptides are a class of biologically relevant molecules, which have shown great promise for the production of highly stable and spectrally pure Ag-NCs. However, the vast space of distinct oligomers still remains practically unexplored. As a consequence, very little is known with regards to what constitutes a good peptide ligand. In addition to the primary structure (i.e. the specific sequence) of a given peptide, secondary structure elements might also play an important role in this context. Therefore, it is essential to gain further insight into these properties. The aim of this thesis is to do just this.
It seems rational to approach this in a highly systematic manner, due to the complex conformational space. However, obtaining a wide variety of peptides for this purpose is either very expensive or extremely time-consuming, as synthesized peptides need purification before evaluation. Consequently, it would be highly beneficial to establish a high-throughput methodology to rapidly screen a large number of peptides in parallel and thereby evaluate their potential as ligands for Ag-NC synthesis. Once established, promising candidates from the screening can be used for evaluation using a more conventional approach. This may lead to the discovery of specific patterns, motifs, or secondary structures, which are highly beneficial for producing suitable peptide ligands. In addition, to ensure biocompatibility, we wish to synthesize the Ag-NCs in aqueous solution at physiological conditions. Around neutral pH, acidic AAs are deprotonated, and basic AAs are protonated. This is expected to facilitate higher binding of Ag.
The scope of this thesis can be summarized as follows:
•
Development of a methodology allowing high-throughput screening of peptide ligands
•
Verification of methodology by utilization of peptides known to facilitate Ag-NC formation
•
Design of a peptide library through systematic, rational design
•
Screening of the library and selection of promising candidates
•
Characterization of these candidates by means of optical and physical methods
•
Discovery of specific properties, patterns, motifs, and secondary structure elements
•
Establishment of tentative guidelines for rational design of novel, peptide ligands
36
37
3.
SOLID-PHASE PEPTIDE SYNTHESIS: A NOVEL APPROACH IN AG-NC SYNTHESIS
As the initial aim of this work is to develop a methodology for high-throughput evaluation of peptides as ligands for Ag-NC synthesis, it is of utmost importance to choose a both convenient and reliable approach. For this purpose, solid-phase peptide synthesis (SPPS) is an obvious candidate, as the approach fulfills both requirements. However, in order to design an appropriate experimental setup, the basic concepts of the methodology must be understood. This chapter will provide a brief overview of SPPS, relevant considerations, and ultimately, a final design for the experimental setup. For a more detailed description of the methodology, chemistry, and possibilities of SPPS, the reader is referred to the excellent reviews and practical guides of Jensen, Tofteng Shenton and Pedersen [370] or Chan and White [371], where the former is very recent and thus state of the art. Unless otherwise mentioned, these are the principle references used throughout this chapter.
3.1.
THE SOLID-PHASE PRINCIPLE
In SPPS, a peptide is assembled in a step-wise fashion starting from its C-terminal AA, which is covalently attached to a solid support through a suitable linker. The functional groups of the AA sidechains are shielded by semi-permanent protecting groups which are not affected by the reaction conditions utilized during chain assembly. The α-amino group of the AA is also shielded, but this is done using a temporary protection group, which is readily removed (Nα-deprotection) prior to coupling of the following AA. The next Nα-protected AA is added to the N-deprotected, resin-bound peptidyl intermediate and the amide bond of the peptide backbone is formed. After the coupling of the AA, excess reagent is removed by washing as all reactions are carried out in fritted reactors with pore sizes sufficiently small to retain the solid support. The newly coupled AA is Nα-deprotected prior to the addition of the following AA. This deprotection and coupling cycle is repeated until the desired peptide has been assembled. Following complete assembly the peptide undergoes a final Nα-deprotection. Subsequently, the peptide is removed from the solid support and the semi-permanent protecting groups are removed from the side-chains. The solid-phase approach decreases the challenging and labor intensive task of isolating each reaction intermediate for subsequent reactions, as surplus reagents and solvated by-products can readily be removed. In addition, the use of repeated reactions allow for a large degree of automation. The basic steps in SPPS have been schematically illustrated in Figure 19.
38
Figure 19: Schematic overview of the basic steps in conventional solid-phase peptide synthesis. Reprinted from [370].
The SPPS principle was first introduced by Bruce Merrifield in the 1960’s, and included a tertbutoxycarbonyl (Boc) group as the temporary Nα-protecting group [372]. Boc is an acid-labile protection group that can be cleaved using trifluoreacetic acid (TFA). Boc-deprotection results in the undesirable formation of trifluoroacetate which, however, can be neutralized using N,Ndiisopropylethylamine (DIPEA), and hence it does not pose any significant problem in practice. Coupling of sequential AAs can be mediated by carbodiimides, pre-formed symmetrical anhydrides of the AAs, or in situ-generated active esters. For protection of side-chain functional groups, acid-labile benzyl-based groups are often used. For the side-chain deprotection and cleavage of the peptide from the resin, the toxic hydrofluoric acid (HF) is the preferred reagent. HF cleavage is necessary as Bocbased SPPS utilizes graduated acidolysis to achieve selectivity in the removal of temporary and permanent protection groups. However, the use of HF requires specialized equipment which is able to withstand the harsh chemistry employed. A more convenient approach to SPPS was established, when a new temporary protection group, 9Fluorenylmethoxycarbonyl (Fmoc), was introduced [373]. Fmoc SPPS is based on an orthogonal protection scheme, where the α-amino group is shielded using the base-labile Fmoc-group, whereas side-chain protection groups are made acid-labile. This means that both permanent protection groups and the linker, if chosen accordingly, are unaffected by the coupling reaction conditions and can hence
39
be chosen to facilitate cleavage under considerably milder conditions than in Boc SPPS. Furthermore, Nα-deprotection (or simply, Fmoc-deprotection) can be followed quantitatively by measuring absorption at 301 nm, due to the conjugated aromaticity of the dibenzofulvene moiety in the Fmoc group.
Fmoc SPPS is performed in polar, aprotic solvents such as dimethylformamide (DMF) or N-methyl-2pyrrolidone (NMP) as they do not contain any protons that can facilitate hydrogen-bonding with the growing peptidyl chain, as do for instance water. The lack of these solvent effects highly increases the rate of nucleophilic attacks hence speeding up reactions and driving them to completion. As illustrated in Figure 19, Fmoc SPPS requires several appropriate reagents for the individual steps. As the Fmoc group is especially labile to secondary amines [374], Fmoc-deprotection is usually performed by treatment with the heterocyclic base piperidine, which it functions both as an initiator of a βelimination reaction and subsequently as a scavenger for the dibenzofulvene moiety, as illustrated in Figure 20. In short, the fluorene moiety is initially deprotonated through a nucleophilic attack by piperidine, forming a cyclopentadiene-type intermediate. This intermediate rapidly eliminates to dibenzofulvene and a peptidyl carboxylate anion. Dibenzofulvene is finally scavenged by piperidine while Fmoc-deprotected peptidyl chain is obtained through the release of carbondioxide.
Figure 20: Schematic representation of the Fmoc-deprotection, β-elimination reaction initiated by a nucleophilic attack from the heterocyclic base, piperidine. Revised from [370].
40
Efficient coupling to the deprotected peptidyl-resin requires activation of the carboxylic acid in the AA to be coupled. In modern Fmoc SPPS, this is done using active esters of the AAs generated in situ. These active esters are usually formed using aminium-based coupling reagents such as N[(dimethylamino)-1H-1,2,3-triazolo-[4,5-b]pyridin-1-yl-methylene]-N-methylmethanaminium hexafluorophosphate N-oxide (HATU) or N-[(1H-benzotriazol-1-yl)(dimethylamino)methylene]N-methylmethanaminium hexafluorophosphate N-oxide (HBTU). The reaction will, however, only occur in the presence of a tertiary amine such as DIPEA. DIPEA facilitates initial deprotonation of the carboxylic acid, which subsequently forms a tetramethyluronium-intermediate though a nucleophilic attack, with release of the anionic OBt- moeity. The OBt-anion subsequently causes an additional nucleophilic attack on the carbonyl in the AA, thus convert the AA into its actived OAt- of OBt-ester, respectively. Activated ester of an Fmoc-protected AA with HBTU is illustrated in Figure 21.
Figure 21: Schematic illustration of in situ formation of activated esters with HBTU/HOBt/DIPEA, employed for couplings in Fmoc SPPS. The reaction is initiated by deprotonation of the carboxylic acid in the Fmoc-AA by DIPEA followed by two successive nucleophilic attacks. Revised from [370].
The active esters readily react with the Nα-deprotected peptidyl resin, through an acyl-transfer mechanism facilitated by nucleophilic attack, thus forming the amide bond, as illustrated in Figure 22. 3-hydroxy-3H-1,2,3-triazolo[4,5-b]pyridine [1-hydroxy-7-azabenzotriazole] (HOAt) or 1hydroxybenzotriazole (HOBt) is often added to the reaction to ensure sufficient formation of activated esters, thereby enhancing coupling yields. Furthermore, the addition suppresses racemization during activation and has been shown to produce very high coupling yields with practically no racemization.
41
Figure 22: Schematic illustration of the acyl-transfer mechanism responsible for coupling between the Nα-deprotected peptidyl-resin and the activated ester-form of the successive amino acid in Fmoc SPPS. Reprinted from [370]
3.2.
COMBINATORIAL LIBRARIES
Combinatorial chemistry has proven to be an extremely powerful tool within drug discovery, as well as in basic research for receptor binding studies and the like [375, 376]. One of the major advantages of using combinatorial libraries is that no prior knowledge about ligand or interaction is required. Also, the method allows for simultaneous synthesis of a library consisting of millions of compound for subsequent screening and selection. Combinatorial libraries can be applied to a range of different molecules such as peptides, oligonucleotides, small molecules, and even large proteins [377-380]. Peptide libraries can be synthesized using both biological and chemical approaches. However, all combinatorial library methods involve a series of main steps:
• • •
Preparation of the library Screening of the library and subsequent selection of promising candidates Determination of the chemical structure of the selected candidates
42
The most widely used biological approach is library preparation using a phage display setup [381], but this method will not be treated further in this work. Chemical libraries are generally prepared using synthesis on a solid support, but there are a number of different approaches that are applicable depending on prior knowledge and aim of the library. One of these is the positional scanning synthetic combinatorial library method, in which one AA in the sequence is specific (Y) while the remaining peptide is an unspecific mixture of AAs (X). The specific AA is then placed in different positions in the peptide, which results in a number of mixtures to be tested depending of the length of the peptide in question (YXn to XnY). This method reveals which specific AAs are best suited at a specific position and results in one (or more) peptides that can be used as lead compound for further development. The method is very suitable for e.g. protein-peptide interaction studies, but is also highly dependent on a suitable screening method for evaluation of the mixtures. Another strong approach is the combination of an Ala-scan and N- and C-terminal truncations of the peptide. This approach is applicable when the aim of the study is to improve on a known or natural peptide. The Ala-scan substitutes every AA sequentially for an alanine, which in turn reveals the importance of this specific AA in the function of the peptide. If the initial AA in a given position is Ala, this is normally substituted for the more bulky Leu to reveal the importance of the Ala residue in that specific position. Truncations of the peptide (from both termini) reveal the minimal number of AAs needed for the peptide to function. Truncations might show that natural peptides are often larger than required to function properly in a given respect [382].
3.2.1.
THE “ONE-BEAD-ONE-COMPOUND” APPROACH
In cases where very little is known about the peptide (or other oligomer), a “split-and-mix” approach is highly applicable. The resin is split into a number of separate reactors, equal to the number of different AAs one wishes to incorporate. After each coupling, the resin from the separate vessels is mixed and afterwards split into individual vessels again for coupling of the next specific AA. This means that each bead only comes into contact with one AA during each coupling step. Hence, every single bead has, in theory, a homogenous product loaded, and the end result is a so-called “one-beadone-compound” (OBOC) library. The method was first introduced in 1991 by Lam et al. [383], and is illustrated in Figure 23 for the simple case of three different AAs. It is noteworthy that the number of distinct peptides in the library (X) grows exponentially with increasing length (N) according to the relation X=AN, where A is the number of different AAs introduced. This means that for growing peptides, the natural diversity, by far, exceeds the capability of standard laboratory facilities. Especially since, as a rule-of-thumb, one needs to synthesize 10*X beads in order to, theoretically, cover the diversity space of the library [384]. This means that for a completely random library of 20mers consisting of all 20 natural AAs synthesized on a standard loading resin, one would need to use over 7 metric tons to cover the diversity space. Using the OBOC approach does, however, give a number of critical advantages compared to other methods:
43
• • •
The synthesis is fast and on a timescale comparable to normal SPPS All peptides are spatially separated at any time and can hence be tested independently Positive candidates can easily be selected and isolated for chemical structure determination
Figure 23: Schematic overview of the ”split-and-mix” method employed in ”one-bead-one-compound” synthesis of combinatorial libraries for a simplified, three-component, trimer synthesis. Revised from [380, 383-385].
There are, on the other hand, a few aspects that need to be taken into account. As different coupling steps might require different conditions and/or reaction times, the general scheme must allow for the “slowest” reactions to complete, as incomplete coupling would lead to more than one compound on each bead. Furthermore, the screening protocol has to be applicable while the peptide is resin-bound. Otherwise, it would normally require a step of isolating every individual bead and subsequently cleave the peptide for a homogenous product. This further implies that side-chain deprotection must be possible on-resin without simultaneous cleavage of the peptide.
44
3.3.
EXPERIMENTAL DESIGN
With the basic chemistry of Fmoc SPPS and the fundamentals of combinatorial OBOC libraries in place, one needs to consider the practical, experimental design. This includes choosing an appropriate resin, linker, and especially the specific chemistry to be employed. These choices should reflect the specific application of the peptide. For this specific experimental design, it is desirable to have several conditions fulfilled.
• • • • •
Screening of the library can be performed on-resin in aqueous solution Conventional, well-documented, and efficient chemistry should be applied Peptides can be deprotected on-resin without facilitating cleavage Cleavage should be possible subsequent to screening and allow for peptide identification The experimental design should not influence the screening results
Several of these criteria can be fulfilled by careful experimental design. This design starts with choice of an appropriate resin. In general, the resins used for SPPS consist of a gel-type, polymer matrix assembled in small microbeads. Naturally, the resin must be compatible with the solvent used in the synthesis, which is determined by its swelling properties. The swelling allows for the reagents to diffuse into the interior of the beads and react with the immobilized substrate. A good resin has a number of properties, such as: Equal distribution of functional groups on both surface and interior of the beads; high solvation of the functional groups throughout the matrix; an inert matrix with good flexibility to allow for peptide growth; high yield per resin volume.
Many resins are based on polystyrene (PS), as it has good swelling properties in standard solvents and can readily be functionalized. Other commonly used supports include polyethylene glycol (PEG) grafted into a low cross-linked PS matrix, such as the TentaGel resin. Other PEG-based resins, such as ChemMatrix [386, 387] and PEGA [388] have been developed, displaying improved properties for some synthesis applications. Today, a large variety of resins are commercially available, which vary greatly in chemical composition, size of the polymer beads, loading capacity, and degree of crosslinking. One major advantage of the PEG-based resins compared to older PS-based resins is good swelling properties in almost any solvent, including water. Poor swelling in water would predominantly make the surface-exposed peptides available for screening. Furthermore, poor swelling might also have an influence on the surface-exposed peptides, as hydrophobic interactions increase. Other advantages in PEG-based resins include the uniformity in bead size, which makes them excellent for preparation of OBOC libraries with a uniform loading.
Linkers are used to provide a reversible linkage between the resin and the growing peptidyl chain, but also determine the C-terminal functionality of the peptide subsequent to cleavage. Among the most
45
conventional and frequently applied are the Wang-type, para-alkoxybensyl-based, linker [389] (for synthesis C-terminal peptide carboxylic acids) and the Rink, trialkoxy-diphenyl-benzhydryl, linker [390] (for synthesis C-terminal peptide amides), just to name a few. The Wang and Rink amide linkers share the property of being acid-labile. This makes them orthogonal to the Fmoc group, and the linker is hence unaffected be the reaction conditions employed during standards Fmoc SPPS. Additionally, the acidic cleavage reagent used, for instance concentrated TFA, is volatile making it easy to remove from the cleaved peptide, which in turn results in a much easier purification process. Acidolysis is also the most commonly employed method in side-chain deprotection, which means side-chain deprotection and cleavage from the resin can be achieved in a one-pot reaction.
However, one might wish to use a linker that is not acid-labile (i.e. unaffected by standard side-chain deprotection conditions) and for this purpose, a range of linkers have also been developed. These are for instance the hydroxymethylbenzoic acid (HMBA) linker [391] which is labile to nucleophiles (however stable under Fmoc-deprotection conditions with piperidine), photocleavable linkers [392], and safety-catch acid labile (SCAL) linkers. The latter type is stable towards TFA until undergoing activation, after which it is readily cleaved by acidolysis [393, 394]. Additionally, linkers have been developed that are bound to the peptidyl chain through a backbone amide such as the BAL linker [395]. Anchoring to the resin away from the C-terminal carboxylic acid allows preparation of Cterminally modified peptides. If one desires the possibility of peptide cleavage under different condition, this can be achieved using a double orthogonally cleavable linker [396]. These combine different linkage strategies on the same resin, which allows for partial cleavage. As loading of the linker onto the resin can be difficult, many pre-loaded resins are commercially available for the most common and popular linkers. Furthermore, loading of the first AA onto the linker can be troublesome and lead to undesirable side-reactions such as racemization. Therefore, a wide range of resins preloaded with both the desired linker, as well as the C-terminal AA, can be acquired commercially.
As previously stated, it is desirable to employ well-established chemistry in the synthesis. In addition to choosing suitable Fmoc-deprotection and coupling conditions, this also requires appropriate sidechain deprotection and cleavage conditions. As there was no need for on-resin modifications of specific AAs in this work, conventional, acid-labile protecting groups are the obvious choice. However, as unwanted side-reactions tend to occur during side-chain deprotection (resulting in the formation of highly cationic species) scavengers are normally added to TFA. These species can, unless trapped, react with and modify AAs that have electron-rich functional groups, e.g. Trp, Tyr, Met, and Cys. By adding scavengers such as ethanedithiol (EDT), phenol, anisol, trialkylsilanes (TES or TIS), and water, side-reactions can be avoided, as these nucleophiles trap the carbocations and quenches unfavorable side-reactions. The choice and concentration of scavengers should, however, reflect the peptide in question and the utilized protection groups. The chemical structure and short names of TFA-labile protecting groups used in conventional Fmoc SPPS are illustrated in Figure 24.
46
Figure 24: Chemical structure and short names of TFA-labile protecting groups used in conventional Fmoc SPPS. Additionally, the amino acids for which the specific protecting groups are used, have been listed. In the case of tert-butyl (tBu), the type of linkage between protecting group and amino acid sidechain, have been listed. Reprinted from [370].
As the aim is to synthesize a library by Fmoc SPPS, which allows for on-resin side-chain deprotection, on-resin screening for (reversible) peptide-templated Ag-NC synthesis, and subsequent cleavage and identification of promising candidates, the design has to allow for all of this. Consequently, the watercompatible PEGA resin was chosen, as this has previously been used for synthesis of combinatorial libraries and subsequent aqueous screening with great success [384]. As a linker, HMBA was chosen for a number of reasons. HMBA has been shown to be completely stable towards TFA treatment which makes it orthogonal to conventional side-chain protecting groups. Furthermore, the HMBA linker has no need for activation before being full cleavable as does the SCAL-type linkers. This fact removes a reaction step in the post synthesis work-up. In addition, HMBA can release both C-terminal peptideamides and –carboxylic acids depending on the employed cleavage conditions. Lastly, the HMBA linker has previously been shown to be fully compatible with the PEGA resin for synthesis of combinatorial peptide libraries [367, 384, 388, 391, 397-401]. Pre-loaded HMBA-PEGA is not commercially available, but sequential loading of the linker and coupling of the first AA can easily be achieved using standard methods. However, as coupling of the first AA (i.e. the C-terminal AA in the final peptide) onto the linker cannot be achieved using standard coupling conditions due to the hydroxyl-functionality of HMBA, esterification can be performed using e.g. a symmetrical anhydride of the C-terminal AA or specialized reagents. To optimize the yield in the first coupling and to avoid the risk of racemization, Fmoc-Gly-OH is a good choice. As the C-terminal AA has the least conformational freedom in the resinbound peptide and the C-terminal group is not accessible during on-resin screening, Gly as the Cterminal AA seems to be a good choice. This approach has also been employed elsewhere with success
47
[402, 403]. In this respect, incorporation of a C-terminal Gly can be regarded as a spacer that should not affect the ability of the peptide for nanocluster stabilization. This naturally has to be tested to ensure methodical validity in this specific case.
Different cleavage conditions have been reported for the HMBA linker including gas-phase ammoniacleavage in a vacuum desiccator followed by elution [367]. This approach results in a C-terminal peptide amide, as the ammonia adds to the ester bond between HMBA and the peptidyl chain. There are both positive and negative aspects using this approach. Though being a bit unconventional, gaseous cleavage does not leave any residuals or surplus reagent in the resulting product compared to aqueous cleavage. Furthermore, the peptide is not physically separated from the bead and can subsequently be eluted using a solvent of choice. However, the process is somewhat more difficult than conventional solution-phase cleavage and the efficiency of the gaseous cleavage and subsequent elution has been reported to be around 60 %, as determined by amino acid analysis. Since HMBA is base-labile, it is cleavable using common bases such as NaOH. The concentration of NaOH has been reported in the range of 0,1-1M depending of the specific product and experimental setup. It should be noted that, in some cases, NaOH concentrations below 1 M has been reported to yield inefficient and incomplete cleavage [401].
Martinez-Ceron et al. [403] reported HMBA cleavage by 40 % aqueous NH4OH and reports yields in the rage of 85-90% for a simple, overnight cleavage. This method allows for identification of peptide candidates from cleavage solutions as well as, to some extent, directly from the resin. It is recommended that peptides are cleaved prior to identification by MALDI-TOF-MS and sequencing using tandem-MS, as direct MS of the peptidyl-resin yields low signal intensities. Additionally, the method allows for cleavage of the peptide either in an isolated reactor or directly on the MALDI target plate using either vapor- or solution-phase cleavage. One report illustrates how a bead was loaded onto a nanoflow tip, subjected to gaseous cleavage and subsequently sequenced using ESI-MS [404]. This method could be applicable using the experimental setup described here, as HMBA is cleavable by gaseous ammonia. Which conditions are best applicable can vary greatly and as such, it is often beneficial to attempt different approaches to establish the optimal conditions for a specific synthesis.
For bulk synthesis of peptides, a more straight forward approach can be employed, which includes conventional Fmoc SPPS on either TentaGel S-RAM or pre-loaded Wang-resins (for synthesis of Cterminal peptide-amides and –acids, respectively) followed by a one-pot, TFA-mediated side-chain deprotection and cleavage.
48
49
4.
PROOF OF CONCEPT
As proof of concept, a set of peptides known to facilitate Ag-NC formation was chosen as models. As previously mentioned, the Dickson group presented three peptide ligands for aqueous synthesis of AgNCs [1]. These three (from here on referred to as DP1 (KECDKKECDKKECDK), DP2 (HDCHLHLHDCHLHLHDCH), and DP3 (HDCNKDKHDCNKDKHDCN), respectively) were selected as model peptides for methodical development. Initially, the peptides were synthesized in order to recreate the original experiment and verify their experimental findings performed in aqueous solutions.
The peptides were synthesized as C-terminal peptide-amides on a TentaGel S-RAM using conventional Fmoc SPPS. In short, Fmoc deprotection was performed using piperidine in NMP (40% for 3min + 20% for 17min), couplings were performed with 4 eq. Fmoc-AA-OH (relative to resin substitution) and HATU (3,8eq), HOAt (4eq), and DIPEA (7,8eq) employing double coupling (2+2h), according to Protocol 1. Cleavage was performed using a cleavage cocktail consisting of TFA/TES/EDT/H2O (90:2,5:2,5:5) for 2h. Crude products were analyzed by HPLC/ESI-MS (LC/MS) and purified by preparative HPLC on a C18 reversed-phase column, using a linear gradient (2-100%) of MeCN (0,1% TFA) in H2O (0,1% TFA) over 45 min, with a flow-rate of 10mL/min. Purified peptides were subsequently re-analyzed by LC/MS and MALDI-TOF-MS to assess purity, as presented for DP1-NH2 in Figure 25.
In the mass spectrum of DP1, we see a major peak contributed to the mass of DP1+18. Deviations from anticipated masses in MALDI-TOF MS were a common sight throughout this work. These can, however, easily be tolerated; especially as the correct mass and anticipated charge/adduct pattern was determined by LC/MS (data not shown). Deviations were primarily ±18, which are commonly observed in MALDI-TOF MS of peptides and can be explained as an ammonium-adduct or associated water (+18) and dehydration (-18). For simplicity, these deviations/adducts are neglected in the peak assignments presented in the text. Instead, the reader is referred to a detailed analysis of all MALDITOF mass spectra recorded in this work, covering assignment and discussion all impurities and adducts (significant and minute), presented in Appendix B. Significant and highly relevant features detected in the mass spectra are, however, assigned and discussed in the text. Furthermore, very broad peaks and shifting isotope patterns in the mass spectra, made specific and precise peak assignments difficult. As a consequence, average peak masses were assigned in peptide identification, and integer-rounded delta masses (Δm) were assigned for detected impurities and adducts. This is, however, a commonly employed approach using MALDI-TOF MS in proteomics and peptidomics due to the aforementioned reasons [405, 406]. Isotope patterns were, however, used to assign the charge of the detected species. As a consequence, both MALDI-TOF and ESI mass data were used to verify synthesis products.
50
Figure 25: MALDI-TOF mass spectrum of once-purified DP1-NH2 recorded in linear negative mode.
One of the aforementioned relevant and significant spectral features is clearly seen in the mass spectrum of DP1-NH2. It shows a peak at a mass corresponding to DP1-NH2 +74 (i.e. Δm=+56 with respect to the principal peak (Na-adduct of DP1-NH2), and is assigned as a tBu-modification of the peptide (+Na-adduct). Whether this is due to an uncleaved side-chain protecting group or a modified Cys due to insufficient scavenging, is principally unknown. This feature will, however, be dealt with later. Nonetheless, this impurity is regarded too significant and hence, the peptide was purified again using a flatter gradient around the elution time of DP1-NH2, which lead to a cleaner mass spectrum, as shown in Figure 26.
Figure 26: MALDI-TOF mass spectrum of twice purified DP1-NH2 recorded in negative linear mode.
In the mass spectrum of the twice purified DP1-NH2 it is clear that the tBu-modified peptide is in much lower (and tolerable) abundance. This is further substantiated by HPLC, displaying a purity of ~97%, . Two rounds of purification were also performed on DP2-NH2 and DP3-NH3 until a satisfactory purity >95% was achieved (data not shown). 51
In order to synthesize Ag-NCs using the three model peptides as ligands, the protocol published by Yu et al., was employed. This protocol dictates aqueous, room-temperature synthesis by mixing the peptide with AgNO3 followed by chemical reduction with NaBH4 with final concentrations of 0,22mM, 0,37mM and 18mM, respectively. The reactions were monitored by means of fluorescence spectroscopy, measuring the emission spectrum with varying λex, as described in Appendix A. In short, this was done by measuring the emission spectrum from 400nm to 730nm using an excitation wavelength spanning from 220nm to 560nm with 20nm increments. The spectra were then plotted in a continuous contour plot to determine the global emission maximum within the measured range.
Intriguingly, this approach produced little-to-none fluorescence emission in the spectral range described by the Dickson group. This lead to a round of protocol optimization where first the silver-topeptide ratio (Ag:P) was adjusted. In the original protocol, this ratio was ~1,7. By varying Ag:P in a range from 0,5 to 20, it was established, that the optimal ratio was Ag:P~5 (evaluated by emission intensity), as illustrated in Figure 27 for DP3-Ag-NCs. This can be interpreted as below this ratio, insufficient amount of Ag is present for efficient formation of emissive Ag-NCs, while above, excess Ag facilitates the formation of larger NPs/agglomeration, thereby decreasing the efficient yield (turnover in %) of emissive Ag-NCs. The concentration of NaBH4 in the original protocol is in large excess relative to the AgNO3 concentration (BH4-:Ag+~50). Varying the NaBH4 concentration did not seem to have significant effect, as long as a large excess was used (data not shown). Consequently, the concentration of NaBH4 was maintained the same as in the original protocol. Using the modified protocol, fluorescence was readily observed in the red.
Figure 27: Maximum emission intensity as a function of Ag-to-peptide ratio for DP3-Ag-NCs, measured 20 hour after reduction in a triplicate experiment. This titration-like protocol optimization reveal an optimal ratio of Ag:P~5.
Maximum emission intensity for Ag-NCs synthesized with the three model peptides as ligands, are plotted over time in Figure 28. Using time-resolved contour plots, it was possible to follow the reaction and evaluate a number of parameters including maximum emission intensity (Imax), excitation (λex) and
52
emission (λem) maxima, formation time (i.e. the time for reaching Imax (Tmax)) as well as chemical halftime in the reaction mixture (T½). These parameters are listed for DP1-DP3 in Table 4, for a 12-day duplicate experiment. Worth to note is the fact that both excitation and emission maxima shifted during the course of the experiment (up to ±15nm). Consequently, the listed wavelengths were measured at Tmax, based on the fact that spectra seemed to ”stabilize” at this time. Furthermore, T½values are based on graphical estimates in the time-resolved plot with ~10h accuracy.
Figure 28: Maximum emission intensity over time of Ag-NC synthesized using DP2-NH2, DP2-NH2, and DP3-NH2 as ligands. Data is averaged averaged over 2 identical experiments. Left: Full 12-day time profile.. Right: Zoom of the first 50 hours.
Ligand
DP1-NH2
DP2-NH2
DP3-NH2
Imax (AU) λex (nm) λem (nm) Tmax (h) T½ (h)
15 ~380 ~595 ~7 50
10 ~400 ~610 ~22 80
17 ~380 ~605 ~21 120
Table 4: Summary of in-solution screening data obtained by fluorescence spectroscopy, as illustrated in Figure 28. Imax: Maximum fluorescence emission intensity observed during screening. Excitation (λex) and emission (λem) wavelength associated with Imax. It should be noted that both wavelengths shift somewhat during the course of synthesis. Tmax: The time after in-situ reduction, where Imax was observed. T½: Halftime of Fluorescence emission with ~10 hour accuracy (graphical estimate).
These results are somewhat consistent with those in the original paper. The emission wavelengths in the original data (DP1=610nm, DP2=615nm, and DP3=630nm) are slightly red-shifted from those measured and in contrast, the re-most emitter is formed using DP2 and not DP3. In good correlation to their results, we also find stability of the three emitters to be DP1 570nm). Although yielding fluorescent, red beads, the emission intensity was quite low. As solutionphase chemistry is not directly adaptable for a solid-phase methodology, another round of protocol optimization was performed. This also included one of the advantages in solid-phase chemistry; washout of excess reagent. This yielded a protocol for on-resin synthesis of Ag-NCs consisting of an initial washing and swelling step, to remove all residual organic solvent from the peptide synthesis, followed by chelation with AgNO3 for 1h (defined as the chelation time, Tc) using a ratio of Ag:P=10 (peptide concentration was assessed from resin loading assuming full synthesis yield). This was followed by an additional washing step and finally chemical reduction using an excess of NaBH4. It proved crucial to include the washing step between chelation and reduction, to remove unbound Ag. Otherwise, a coating was also observed on the surface of the beads. A selection of images from the protocol optimization is shown in Figure 34.
To ensure comparable reaction kinetics and chemical stability of the Ag-NCs, time-resolved fluorescence microscopy was performed on Ag-NCs synthesized using P-Gly-HMBA-PEGA with all three model peptides, as ligands. Emission from the beads was monitored over a 12-day period. The recorded images were analyzed according to the experimental section (Appendix A). In short, images were divided to into their respective RGB channels to obtain intensity images for the individual
60
channels. Subsequently, the area of 5 individual beads was integrated to obtain average, red pixel intensity, as only red/NIR emitters were of specific interest. Time-resolved, normalized emission intensity for DP3-Ag-NCs in solution and on resin is shown in Figure 35 for comparison.
Figure 34: Protocol optimization for on-resin P-Ag-NC synthesis using various Ag:P ratios, chelation times (Tc), and a washing step prior to reduction. (A) Ag:P=5, Tc=1h, -wash. (B) Ag:P=5, Tc=5min, +wash. (C) Ag:P=5, Tc=1h, +wash. (D) Two treatments of: Ag:P=5, Tc=1h, +wash. (E) Ag:P=10, Tc=1h, +wash. (F) Ag:P=25, Tc=1h, +wash. All images were capture using a CCD camera and a 1s exposure time.
Figure 35: Normalized emission intensity over time for DP3-Ag-NCs both in solution and on-resin.
It is evident from the plot, that Tmax is much higher on-resin than in solution, which was also found to be the case for all three model peptides. This, however, was anticipated, as solid-phase reactions are generally slower, due to reduced mobility and conformational freedom in solid-phase as compared to solution-phase. T½ is roughly comparable to those obtained using C-terminal peptide-amides in solution. This further supports the previous finding that C-terminal functionality does indeed facilitate increased stability. Unfortunately, only kinetic parameters can be directly compared, as the
61
experimental setup used, was not equipped for spectral measurements but only imaging. To further ensure methodical validity, the approach was used for synthesis of (Gly-Ala)n with n= (4, 8, 12) and (Lys-Ala-Glu-Ala)n with n=(2, 4, 6). These peptides have been shown not to facilitate fluorescent Ag-NC formation [1] and were used as negative controls. Furthermore, the naked PEGA resin, the HMBAPEGA resin, unprotected peptides, and unreduced DP-Ag complexes were checked for influence on the fluorescence emission. Only side-chain protected peptides displayed fluorescence emission which, however, disappeared after side-chain deprotection, and was hence ascribed to the side-chain protecting groups. Finally, reversibility of on-resin Ag-NC synthesis was demonstrated by excessive EDT treatment. This was done by treating the resin with 50 eq. of aqueous EDT for 2*1h with intermittent washing. A selection of images is shown in Figure 36, demonstrating methodical validity.
It should be noted, that the fluorescent beads are detectable by eye under UV illumination (e.g. on the microscope stage) facilitating a very efficient and rapid initial screening and selection. However, as evaluation of the emission is performed only within the visible range, NIR/IR-emitters are overlooked in the screening. This is the principle disadvantage of performing screening in this setup. One way to extend the detection range into the NIR range and also facilitate automation could be the use of a NIRsensitive fluorescence plate reader programmed to screen at multiple wavelengths. It should, however, be noted that it might not be possible to perform on-resin screening with this approach. Unfortunately, this equipment was unavailable for this work and was not examined further.
A critical aspect with this approach is that with on-resin screening on a high-swelling resin, large spatial separation of the resin-bound peptides is achieved by pseudo-dilution. This directly implies that peptides that could possible serve as capping agents (i.e. P:NC>1) will not be found, as siteisolation for individual resin-bound peptides is expected. However, as PEGA is in fact a dynamic resin, a higher degree of on-resin peptide mobility is expected, compared to e.g. the much more rigid PSbased resins. However, as the aim of this study to investigate peptide-templated Ag-NC synthesis where the peptide functions as a template (i.e. presumably P:NC=1), this aspect can in fact be regarded as an advantage. However, for it to be possible to synthesis Ag-NCs on-resin for an entire peptide library, it requires the establishment of a general protocol. While this might be fruitful for a number of peptides, this approach dictates that the P-Ag-NCs must be synthesized under the same conditions with the same stoichiometry, regardless of peptide sequence and length. The experimental conditions employed, however, use an excess amount of AgNO3 relative to the conditions employed in solution, which would allow for synthesis of larger clusters. As unbound Ag+ is washed out prior to reduction, this should resolve the issue to some extent.
62
Figure 36: Fluorescence microscopy images from different points in the synthesis scheme, using three different filter cubes. Left: λex = 330-385nm ; λem > 420nm. Middle: λex = 470-495nm ; λem = 510-550nm. Right: λex = 510-550nm ; λem > 570nm. Rows (top to bottom): H2N- PEGA; HMBA-PEGA; side-chain protected P3-Gly-HMBA-PEGA; Ag-P3-Gly-HMBA-PEGA (unreduced); Ag-P3-Gly-HMBA-PEGA (reduced); P3-Gly-HMBA-PEGA (after Ag removal by EDT treatment). All images were captured using a CCD camera and an exposure time of 1s.
63
Screening of peptide libraries with regards to metal NC formation has previously been demonstrated [407]. However, the screening was performed in solution on a very small library and did not focus specifically on synthesis of fluorescent NM-NCs. Furthermore, the reaction conditions employed were very different from the ones presented here. For instance, Slocik and Wright performed their screening with Ag:P=0,65, which was found insufficient in this work. To the best of our knowledge, this is the first report of on-resin screening for synthesis of peptide-templated synthesis of fluorescent AgNCs. The concept should be easily adapted to screen for Au-NCs as well, but this is beyond the scope of this work. For fundamental insight into the composition of good peptide ligands, a parallel library was designed. The background and justification for the design will be presented in the following chapter.
A schematic illustration of the synthesis route from resin-bound and side-chain protected peptide to fluorescent on-resin P-Ag-NCs, is illustrated in Figure 37. In this illustration, the final reduction step causes a conformational change into two helical segments. This conformation is merely speculative, although a conformational change is expected in an analogous fashion to the seen for DNA-Ag-NCs.
Figure 37: Schematic overview of on-resin Ag-NC synthesis, developed in this work. Starting from the side-chain protected peptidyl-resin, TFA-mediated deprotection yields the deprotected peptidyl-resin which subsequently is allowed to chelate/complex with silver ions. Through a final in situ chemical reduction, resin-bound P-Ag-NCs are formed. Pg: Side-chain protecting group.
64
65
6.
LIBRARY DESIGN
As a rationally designed library necessitates careful considerations, the following section contains the background for this design. As previously stated, very little is known regarding peptides as ligands for Ag-NC synthesis and consequently, no general design rules have been established. The three model peptides employed in this work were initially designed using very basic principles [1]. DP1 was designed based on a very tentative analysis of AA composition in the nucleolar protein, nucleolin [408] (UniProt: P19338). The three most abundant AAs (Glu, Lys, and Asp constituting 15,6%, 12,7%, and 8,5%, respectively) were combined with Cys in an apparently random 15-mer (KECDKKECDKKECDK). No justification for the inclusion of Cys was given, as this is of ~0,1% abundance in nucleolin. DP2 was designed by including a hydrophobic AA (Leu) based on the finding, that hydrophobic regions facilitate Ag-NC formation in dendrimers. Furthermore, His was included in an apparently random 18-mer (HDCHLHLHDCHLHLHDCH). DP3 was designed by including Asn in an apparently random 18-mer (HDCNKDKHDCNKDKHDCN). However, these peptides diverge somewhat from the initial principles. For instance, His was included for no apparent reason, as this is of only ~0,1% abundance in nucleolin. The justification for including Leu was only employed for DP2, where DP3 included Asn without any further justification (abundance in nucleolin ~3%). The designs were made, not taking into account the specific binding/chelation sites of Ag in native nucleolin, as this still remains unknown. Consequently, no specific structural information from the native protein can be exploited for rational design of peptide ligands. Furthermore, peptide length was 15-18 AAs with no apparent justification. Also, the sequences seemed to be randomly shuffled using pentad (DP1) and heptad (DP2 and DP3) repeats. As stated by Diez and Ras [155], this clearly reflect the lack of understanding of the underlying mechanisms for Ag-NC stabilization using peptides. To make an initial effort for understanding these, this work includes a rationally designed library, which should shed light on what might constitute a good peptide ligand.
From the handful of publications described in Section 1.5.3., only the work by the Dickson group clearly shows the use of peptides as ligand in synthesis of fluorescent Ag-NCs without use of a rigid framework or by capping. Furthermore, a handful of globular proteins have been described, but without knowledge of the binding site the Ag-NCs, little usable information can be drawn from these. Of interest might be the silver binding protein, SilE, found in the periplasm of Escherichia coli where it is associated with silver tolerance of the bacteria [409]. Within this protein, Ag+ ions are bound by bridging His residues placed far apart in the protein. Furthermore, the binding of metals, and particularly Ag, to His in peptides and proteins have been reported on several occasions [410-412]. This might be the background for the inclusion of His in the designs of the Dickson group.
Metallothioneins (MTs) are in interesting class of proteins from which inspiration might be found. MTs are related to heavy metal stress response in many organisms and have been shown to have high affinity for Ag [413-417]. By definition, this class of proteins have a number of characteristic
66
properties: (i) low molecular weight 570nm). Exposure time was kept constant at 1s.
Peptide
Sequence
Length1 Charge2 #Cys
Imax
Tmax
P109 P126 P135 P142 P145 P148 P157 P160 P166 P172 DP3
ACAHACAAACAHACAAACAHACAA HCADACKAHCADACKA HCDNHCDNHCDNHCDN HCAAAACAHCAAAACAHCAAAACA ACHAAACAACHAAACAACHAAACA ACAAHACAACAAHACAACAAHACA HCDAHACDHCDAHACDHCDAHACD KCDAHACAKCDAHACAKCDAHACA HCDNHNCDHCDNHNCDHCDNHNCD DCNHDHCKDCNHDHCKDCNHDHCK HDCNKDKHDCNKDKHDCN
24 16 16 24 24 24 24 24 24 24 18
161,1* 114,5 153,3 156,7* 161,9 140,4 162,4 167,4 148,7 97,5 131,4
225* 52 125 225* 125 125 80 80 80 52 80
+3 +2 0 +3 +3 +3 0 +3 0 +3 +2
6 4 4 6 6 6 6 6 6 6 3
Table 9: Overview of the top-10 peptide candidates determined from on-resin library screening. #Cys refers to the number of Cys residues in the sequence. Imax refers to the maximum, average pixel intensity at Tmax determined from Figure 40. 1Length of the native peptide without the C-terminal Gly linker. 2Charge at physiological pH. *At 225 hours, an apparent maximum had not been reached.
70
Figure 39: Average red-pixel intensity over time for the top-10 P-Ag-NCs fluorescence, as determined from library screening over a 10-day period. On-resin emission for DP3-Ag-NCs has been plotted for comparison. All images were captured through a U-MWG2 filter cube (λex = 510-550nm – λem > 570nm) with an exposure time of 1s.
As previously stated, the on-resin screening is only indicative of peptide potential as a ligand for Ag-NC synthesis. This primarily owes to the reduced flexibility and conformational freedom on the solidphase, but also the elimination of cooperative effects based on the spatial separation of the peptides. In order to assess their actual potential, solution-based evaluation and screening must subsequently be performed. As the library synthesis did not produce sufficient amounts of peptide for comprehensive bulk characterization, another round of SPPS was performed.
71
8.
CHARACTERIZATION OF NOVEL P-AG-NCS
It was noticed, that spectral properties of P-Ag-NCs synthesized using DP1-DP3 as ligands were consistent regardless of the peptides were synthesized on a TentaGel S-RAM resin or on the HMBAPEGA resin, as long as they were sufficiently pure (>95%) and had the same C-terminal functionality (data not shown). Consequently, the candidates were synthesized as C-terminal peptide-amides on a TentaGel S-RAM. For simplicity, this was also done despite the fact that C-terminal peptide-acids showed increased stability. Based on the optimized protocol for P-Ag-NC synthesis using a ratio of Ag:P=5, one variant of the best scoring peptide (P160) was made. This variant contains a C4A substitution to make the Cys:Ag ratio 1:1 during Ag-NC synthesis. This peptide, P160(C4A), will from hereon be referred to as P262.
Synthesis and cleavage was performed by employing the general synthesis approach (Protocol 1 and 2). Synthesis verification by LC/MS and preparative HPLC purification was performed as previously described. As was the case with the three model peptides, tBu-modified peptides were detected. For some of the peptides, this resulted in very low crude yields (~25%) which were highly unsatisfactory. Consequently, the crude products were re-cleaved (i.e. treated once more with the cleavage cocktail), which increased the crude purity substantially (~50%). An additional treatment with the cocktail did not further improve the crude purity. This is illustrated for P142 in Figure 40.
Figure 40: Chromatograms of P142 from preparative HPLC showing the desired peptide (elution time = 20,5min) and the tBu-modified peptide (elution time = 21,9 min). Elution times correspond to ~48% and ~56% MeCN, respectively. Bottom: Crude product displaying a purity of ~23%. Top: Crude product after one additional treatment with the cleavage cocktail (re-cleaved), showing a purity of ~49%. Samples were run through a C18 reversed phase column, using 0,1% TFA in MeCN and 0,1% aqueous TFA as eluents. Traces were recorded at 215nm.
72
As was the case for the model peptides, two rounds of purification was generally a necessity for achieving a purity >95%, as assessed by HPLC. The MALDI-TOF mass spectrum of the re-cleaved and twice-purified P135 is shown in Figure 41. In addition to the observed feature seen with DP1-DP3 (dehydration, water association/ammonia-adducts, and co-crystallized CHCA-matrix), two interesting features were observed. Based on the calculated mass, the observed mass of the peptide displayed Δm=-2. This is assigned to disulfide-formation and was observed for the majority of peptides, displaying either one or two disulfides. Söptei et al. [437] reported that for proteins and peptides used as ligands for synthesis of Au-NCs, free and disulfide-bonded Cys can be regarded as equivalent. This was also assumed in this work and hence, disulfide-formation in the native peptides, were of no further concern. In addition, a small amount of peptide dimer was observed. The dimer also displayed Δm=-2, but as non-covalent dimerization would result in an anticipated Δm=-4, the dimerization was ascribed to a covalent, disulfide formation between two non-disulfide containing monomers. This could be interpreted as dimerization being a competitive reaction for the disulfide-formation within the monomer. Dimerization in low quantities was observed for P135, P145, P148, P160, and P262. This, however, might reflect the specific experimental conditions in the individual MALDI-TOF measurements. For a full analysis of impurities and adducts observed in MALDI-TOF mass spectra recorded from onece-purified peptides in this work, please refer to Appendix B.
Figure 41: MALDI-TOF mass spectrum of once-purified P135 recorded in linear negative mode. Peptide mass measured: M=1887,6 (calculated 1889,6).
As peptides were obtained in >95% purity, they were tested as ligands for Ag-NC synthesis in solution using the optimized Ag:P=5 ratio (Protocol 5). This was done using the top-10 candidates, P262 and DP3 (all as C-terminal amides) for comparison. It should be noted that spectrophotometer settings were adjusted to achieve higher signal. Consequently, intensities cannot be directly compared to the values presented in Chapter 4. For comparison, DP3-NH2 was re-evaluated in parallel and is used as reference in all experiments. Time-resolved emission intensity, as an average over three identical experiments, is plotted in Figure 42 and summarized in Table 10.
73
Figure 42: Maximum emission intensity over time for P-Ag-NCs synthesized using the top-10 candidates, P262, and DP3 (all C-terminal amides) as ligands. Data has been averaged over 3 identical experiments. Left: Full 10-day time profile. Insert: Label specification. Right: Zoom of the first 50 hours.
In contrast to the on-resin screening, where Ag-NCs formed using 8 of the 10 candidates (P109, P135, P142, P145, P148, P157, P160, and P172) displayed higher emission intensity than DP3-Ag-NCs, solution-based synthesis only showed higher intensity for 2 candidates (P160 and P262). Interestingly, is the best performing peptide on-resin and P262 is closely related. Ag-NCs synthesized using P157, P166, P172, and to some extent P135 showed both a maximum intensity and kinetic parameters comparable to those of DP3-Ag-NCs. Noticeably, Ag-NCs formed with P109, P126, P142, P145, and P148 all show low emission intensities, while these outperformed DP3 on-resin. This might be a consequence of some of the previously described features of using this on-resin screening approach, such as spatial separation and reduced conformational freedom. In this respect, it should be noted that P109, P142, P145, ad P148 showed very low aqueous solubility and displayed a high degree of aggregation over time in the solution based Ag-NC synthesis. Furthermore, it should be noted that higher emission intensity cannot be directly related to a brighter emitter. This might also likely be an indication of a higher yield from the synthesis. A higher binding constant is likely linked to the stability of the emitter, as seen for DNA-Ag-NCs [320]. As such, the time-resolved screening approaches (both in solution and on-resin) should primarily be used to evaluate the stability (and related kinetic parameters) of the P-Ag-NCs. To determine true brightness, it is a necessity to determine ε and Q for the individual emitters. Possible reasons for the observations will be discussed later, where experimental findings in this work will be correlated with sequence and structural analysis. The peptides showing properties comparable to or better than DP3 have been highlighted in bold italics, in Table 10, and for these (with the exception of P135) an increased stability (i.e. higher T½) was also seen. For the remaining peptides, a comparable or lower T½ was observed, with the exception of P145 which displayed a larger half-time. Due to the low emission intensity of P145-Ag-NCs, this value should, however, be interpreted with caution.
74
Ligand
P109
P126
P135
P142
P145
P148
P157
P160
P166
P172
P262
DP3
Imax (AU) λEm (nm) λEx (nm) Tmax (h) T½ (h)
23 ~615 ~400 26 100
23 ~610 ~380 12 40
70 ~610 ~400 16 110
20 ~615 ~400 26 120
17 ~610 ~420 56 200
22 ~620 ~400 26 90
83 ~610 380 26 180
114 ~620 ~400 26 160
92 ~610 ~380 20 220
94 ~610 ~380 20 150
108 ~620 ~400 45 220
97 ~605 380 16 120
Table 10: Summary of in-solution screening data using fluorescence spectroscopy, as illustrated in Figure 30. Imax: Maximum fluorescence emission intensity observed during screening. Excitation (λex) and emission (λem) wavelength associated with Imax. Tmax: The time after in-situ reduction, where Imax was observed. T½: Half-time with ~10 hour accuracy (graphical estimate). Peptides showing comparable or better properties than DP3 have been highlighted in bold italics.
Prior to characterization, all emitters were purified according to the established two-step ultracentrifugation approach at a time ~Tmax for the individual emitters. In this respect, it should be noted that an even higher degree of precipitation/aggregation was observed during the purification of Ag-NCs synthesized using P142 and P145, to some degree for P109, and to a lesser degree for P148. As solubility and aggregation was already an issue during P-Ag-NC synthesis with these peptides, volume reduction facilitated this even furthers. However, using the very simple and fast two-step ultracentrifugation, aggregates are removed in the high MWCO filter. This yielded aggregate-free solutions of the four P-Ag-NCs for characterization. Resultantly, emitter concentration in these samples was naturally somewhat lower than in those where no precipitation was seen. Nonetheless, as the purification yielded volume reduction and hence concentration of the samples to an unspecified volume, as well as the fact that conversion/synthesis yield and hence final concentration for the individual emitters is unknown, this was not, as such, a limitation, as emission intensity of the concentrated P-Ag-NCs are not directly comparable.
With the abovementioned observations in mind, spectral characterization of the purified emitters was performed. This resulted in the determination of the absorption, excitation and emission spectra for all P-Ag-NCs as well as a contour plot of the fluorescence within the visible range, analogous to those measured for the model peptides. In general, the synthesized P-Ag-NCs displayed similar spectral properties. Normalized emission spectra for all P-Ag-NCs are shown in Figure 43 for comparison.
75
Figure 43: Normalized emission spectra from the top-10 P-Ag-NCs. λex=380nm.
The P-Ag-NCs can be grouped according to their aggregation/solubility properties. Normalized absorption, excitation, and emission spectra from the four poorly soluble P-Ag-NCs can be seen in Figure 44, while contour plots can be seen in Figure 45. From the spectra it is evident, that the spectral features of P142-Ag-NCs and P145-Ag-NCs differ the most from remaining P-Ag-NCs. This is likely due to their poor solubility which, in turn, makes only very few monomeric peptides available for Ag-NC formation, as aggregation is a competing reaction in a much larger extent than for the remaining peptides. This is reflected especially in their absorption and emission spectra. With increasing solubility (and hence final emitter concentration), higher UV absorption and more distinct emission spectra are observed. Furthermore, from the contour plots it can be seen, that increasing solubility results in higher spectral purity due to decreased aggregation. For the P-Ag-NCs with very low solubility, red-emissive Ag-NCs were produced with very low efficiency and rather, emission at shorter wavelengths was observed – especially when exciting in the UV. Higher (relative) UV excitation of the red emitters is also seen with decreasing solubility (solubility of P-Ag-NCs: P142 4 in an argon matrix. Chemical Physics Letters, 1999. 313(1): p. 105-109. Félix, C., et al., Ag_ {8} Fluorescence in Argon. Physical Review Letters, 2001. 86(14): p. 2992. Rabin, I., et al., Absorption and fluorescence spectra of Ar-matrix-isolated Ag< sub> 3 clusters. Chemical Physics Letters, 2000. 320(1): p. 59-64. Sieber, C., et al., Isomer-specific spectroscopy of metal clusters trapped in a matrix: Ag_ {9}. Physical Review A, 2004. 70(4): p. 041201. Harb, M., et al., Optical absorption of small silver clusters: Ag,(n= 4–22). The Journal of chemical physics, 2008. 129: p. 194108. Fedrigo, S., W. Harbich, and J. Buttet, Collective dipole oscillations in small silver clusters embedded in rare-gas matrices. Physical Review B, 1993. 47(16): p. 10706. Harbich, W., et al., Deposition of mass selected silver clusters in rare gas matrices. The Journal of chemical physics, 1990. 93: p. 8535. Schulze, W., I. Rabin, and G. Ertl, Formation of Light-Emitting Ag2 and Ag3 Species in the Course of Condensation of Ag Atoms with Ar. ChemPhysChem, 2004. 5(3): p. 403-407. Fedrigo, S., W. Harbich, and J. Buttet, Media Effects on the Optical Absorption Spectra of Silver Clusters Embedded in Rare Gas Matrices. International Journal of Modern Physics B, 1992. 6(23n24): p. 3767-3771. Turro, N.J., V. Ramamurthy, and J. Scaiano, Modern molecular photochemistry of organic molecules. Univ. Sciece Books, 2010. Schultz, D. and E.G. Gwinn, Silver atom and strand numbers in fluorescent and dark Ag:DNAs. Chemical Communications, 2012. 48(46): p. 5748-5750. Ramazanov, R.R. and A.I. Kononov, Excitation Spectra Argue for Threadlike Shape of DNA-Stabilized Silver Fluorescent Clusters. The Journal of Physical Chemistry C, 2013. 117(36): p. 18681-18687. O’Neill, P.R., E.G. Gwinn, and D.K. Fygenson, UV excitation of DNA stabilized Ag cluster fluorescence via the DNA bases. The Journal of Physical Chemistry C, 2011. 115(49): p. 24061-24066. Neidig, M.L., et al., Ag K-edge EXAFS analysis of DNA-templated fluorescent silver nanoclusters: insight into the structural origins of emission tuning by DNA sequence variations. Journal of the American Chemical Society, 2011. 133(31): p. 11837-11839. Patel, S.A., et al., Electron Transfer-Induced Blinking in Ag Nanodot Fluorescence. Journal of Physical Chemistry C, 2009. 113(47): p. 20264-20270. Trifonov, A., et al., Real-time observation of hydrogen bond-assisted electron transfer to a DNA base. Chemical Physics Letters, 2005. 409(4): p. 277-280. Huber, R., T. Fiebig, and H.-A. Wagenknecht, Pyrene as a fluorescent probe for DNA base radicals. Chemical Communications, 2003(15): p. 1878-1879. 163
315.
316. 317. 318. 319. 320. 321. 322. 323.
324. 325. 326. 327. 328.
329. 330. 331. 332. 333. 334. 335. 336. 337.
Raytchev, M., et al., Ultrafast Proton-Coupled Electron-Transfer Dynamics in Pyrene-Modified Pyrimidine Nucleosides: Model Studies towards an Understanding of Reductive Electron Transport in DNA. ChemPhysChem, 2004. 5(5): p. 706-712. Díez, I., et al., On heterogeneity in fluorescent few-atom silver nanoclusters. Physical Chemistry Chemical Physics, 2013. 15(3): p. 979-985. Demchenko, A.P., The red-edge effects: 30 years of exploration. Luminescence, 2002. 17(1): p. 1942. Díez, I., et al., Blue, green and red emissive silver nanoclusters formed in organic solvents. Nanoscale, 2012. 4(15): p. 4434-4437. Jiang, D.-e., et al., The “Staple” Motif: A Key to Stability of Thiolate-Protected Gold Nanoclusters. Journal of the American Chemical Society, 2008. 130(9): p. 2777-2779. Li, W., et al., Effects of polymorphic DNA on the fluorescent properties of silver nanoclusters. Photochemical & Photobiological Sciences, 2013. 12(10): p. 1864-1872. Feynman, R.P., There's plenty of room at the bottom [data storage]. Microelectromechanical Systems, Journal of, 1992. 1(1): p. 60-66. Yuan, X., et al., Synthesis of Highly Fluorescent Metal (Ag, Au, Pt, and Cu) Nanoclusters by Electrostatically-Induced Reversible Phase Transfer. Acs Nano, 2011. Yuan, X., et al., Traveling through the Desalting Column Spontaneously Transforms Thiolated Ag Nanoclusters from Nonluminescent to Highly Luminescent. The Journal of Physical Chemistry Letters, 2013. 4: p. 1811-1815. Yuan, X., et al., Highly luminescent silver nanoclusters with tunable emissions: cyclic reduction– decomposition synthesis and antimicrobial properties. NPG Asia Materials, 2013. 5(2): p. e39. Udayabhaskararao, T. and T. Pradeep, Luminescent Ag7 and Ag8 clusters by interfacial synthesis. Angewandte Chemie International Edition, 2010. 49(23): p. 3925-3929. Muhammed, M.A.H., et al., Growth of In Situ Functionalized Luminescent Silver Nanoclusters by Direct Reduction and Size Focusing. Acs Nano, 2012. 6(10): p. 8950-8961. Schreiber, F., Structure and growth of self-assembling monolayers. Progress in surface science, 2000. 65(5): p. 151-257. Sellers, H., et al., Structure and binding of alkanethiolates on gold and silver surfaces: implications for self-assembled monolayers. Journal of the American Chemical Society, 1993. 115(21): p. 93899401. Frey, S., et al., Structure of thioaromatic self-assembled monolayers on gold and silver. Langmuir, 2001. 17(8): p. 2408-2415. Love, J.C., et al., Self-assembled monolayers of thiolates on metals as a form of nanotechnology. Chemical Reviews, 2005. 105(4): p. 1103-1170. Gwinn, E.G., et al., Sequence-dependent fluorescence of DNA-hosted silver nanoclusters. Advanced Materials, 2008. 20(2): p. 279-+. Koszinowski, K. and K. Ballweg, A Highly Charged Ag-6(4+) Core in a DNA-Encapsulated Silver Nanocluster. Chemistry-a European Journal, 2010. 16(11): p. 3285-3290. O’Neill, P.R., et al., Few-Atom Fluorescent Silver Clusters Assemble at Programmed Sites on DNA Nanotubes. Nano Letters, 2012. 12(11): p. 5464-5469. Gwinn, E. and D. Schultz, Stabilization of fluorescent silver clusters by RNA homopolymers and their DNA analogs: C,G versus A,T(U) dichotomy. Chemical Communications, 2011. 47(16): p. 4715-4717. Diez, I., et al., Controlled growth of silver nanoparticle arrays guided by a self-assembled polymerpeptide conjugate. Soft Matter, 2010. 6(14): p. 3160-3162. Zhang, W., et al., In situ generation of fluorescent silver nanoclusters in layer-by-layer assembled films. Journal of Materials Chemistry C, 2013. 1(10): p. 2036-2043. Cathcart, N., et al., Chiral thiol-stabilized silver nanoclusters with well-resolved optical transitions synthesized by a facile etching procedure in aqueous solutions. Langmuir, 2009. 25(10): p. 58405846. 164
338. 339. 340. 341. 342. 343. 344.
345. 346. 347.
348. 349.
350. 351. 352.
353.
354. 355.
356.
357. 358. 359.
Bootharaju, M. and T. Pradeep, Facile and rapid synthesis of dithiol-protected Ag7 quantum cluster for selective adsorption of cationic dyes. Langmuir, 2013. Zheng, C., et al., Synthesis and Spectroscopic Characterization of Water-Soluble Fluorescent Ag Nanoclusters. Journal of analytical methods in chemistry, 2013. 2013. Adhikari, B. and A. Banerjee, Facile synthesis of water-soluble fluorescent silver nanoclusters and HgII sensing. Chemistry of Materials, 2010. 22(15): p. 4364-4371. Zhou, D.-L., et al., Polyinosinic acid-stabilized fluorescent silver nanoclusters for sensitive detection of biological thiols. Analytical Methods, 2013. 5(21): p. 6076-6080. Sengupta, B., et al., Base-Directed Formation of Fluorescent Silver Clusters. Journal of Physical Chemistry C, 2008. 112(48): p. 18776-18782. Ritchie, C.M., et al., Ag nanocluster formation using a cytosine oligonucleotide template. The Journal of Physical Chemistry C, 2007. 111(1): p. 175-181. Schultz, D. and E. Gwinn, Stabilization of fluorescent silver clusters by RNA homopolymers and their DNA analogs: C, G versus A, T (U) dichotomy. Chemical Communications, 2011. 47(16): p. 47154717. Ono, A., et al., Specific interactions between silver (I) ions and cytosine–cytosine pairs in DNA duplexes. Chemical Communications, 2008(39): p. 4825-4827. Ono, A. and Y. Miyake. Highly selective binding of metal ions to thymine-thymine and cytosinecytosine base pairs in DNA duplexes. in Nucleic acids symposium series. 2003. Oxford Univ Press. Freeman, R., T. Finder, and I. Willner, Multiplexed analysis of Hg2+ and Ag+ ions by nucleic acid functionalized CdSe/ZnS quantum dots and their use for logic gate operations. Angewandte Chemie International Edition, 2009. 48(42): p. 7818-7821. Kypr, J., et al., Circular dichroism and conformational polymorphism of DNA. Nucleic Acids Research, 2009. 37(6): p. 1713-1725. Su, Y.-T., et al., Detection of copper ions through recovery of the fluorescence of DNA-templated copper/silver nanoclusters in the presence of mercaptopropionic acid. Analytical Chemistry, 2010. 82(20): p. 8566-8572. Shemer, G., et al., Chirality of silver nanoparticles synthesized on DNA. Journal of the American Chemical Society, 2006. 128(34): p. 11006-11007. Molotsky, T., et al., Synthesis of Chiral Silver Clusters on a DNA Template†. The Journal of Physical Chemistry C, 2010. 114(38): p. 15951-15954. Driehorst, T., et al., Distinct Conformations of DNA-Stabilized Fluorescent Silver Nanoclusters Revealed by Electrophoretic Mobility and Diffusivity Measurements. Langmuir, 2011. 27(14): p. 8923-8933. Chang, W.H., et al. Synthesis of Gadolinium-doped Fluorescent Au/Ag Nanoclusters as Bimodal MRI Contrast Agents. in XIII Mediterranean Conference on Medical and Biological Engineering and Computing 2013. 2014. Springer. Udayabhaskararao, T., B. Nataraju, and T. Pradeep, Ag9 quantum cluster through a solid-state route. Journal of the American Chemical Society, 2010. 132(46): p. 16304-16307. Udayabhaskararao, T., M. Bootharaju, and T. Pradeep, Thiolate-protected Ag 32 clusters: mass spectral studies of composition and insights into the Ag–thiolate structure from NMR. Nanoscale, 2013. 5(19): p. 9404-9411. Chakraborty, I., et al., The superstable 25 kDa monolayer protected silver nanoparticle: Measurements and interpretation as an icosahedral Ag152 (SCH2CH2Ph) 60 cluster. Nano Letters, 2012. 12(11): p. 5861-5866. Slocik, J.M. and D.W. Wright, Biomimetic mineralization of noble metal nanoclusters. Biomacromolecules, 2003. 4(5): p. 1135-1141. y Cajal, S.R., Comparative study of the sensory areas of the human cortex. 1899. Lafarga, M., et al., Cajal’s contribution to the knowledge of the neuronal cell nucleus. Chromosoma, 2009. 118(4): p. 437-443. 165
360. 361. 362.
363. 364. 365. 366. 367. 368.
369. 370. 371. 372. 373. 374.
375. 376. 377. 378.
379. 380. 381. 382. 383.
Merril, C.R., D. Goldman, and M.L. Van Keuren, Simplified silver protein detection and image enhancement methods in polyacrylamide gels. Electrophoresis, 1982. 3(1): p. 17-23. Goodpasture, C. and S.E. Bloom, Visualization of nucleolar organizer regions in mammalian chromosomes using silver staining. Chromosoma, 1975. 53(1): p. 37-50. Kerenyi, L. and F. Gallyas, A highly sensitive method for demonstrating proteins in electrophoretic, immunoelectrophoretic and immunodiffusion preparations. Clinica Chimica Acta, 1972. 38(2): p. 465-467. Anand, U., S. Ghosh, and S. Mukherjee, Toggling between blue-and red-emitting fluorescent silver nanoclusters. J. Phys. Chem. Lett, 2012. 3(23): p. 3605-3609. Narayanan, S.S. and S.K. Pal, Structural and Functional Characterization of Luminescent SilverProtein Nanobioconjugates. The Journal of Physical Chemistry C, 2008. 112(13): p. 4874-4879. Peyser-Capadona, L., et al., Nanoparticle-free single molecule anti-stokes Raman spectroscopy. Physical Review Letters, 2005. 94(5): p. -. Gazit, E., Self-assembled peptide nanostructures: the design of molecular building blocks and their technological utilization. Chemical Society Reviews, 2007. 36(8): p. 1263-1269. Slocik, J.M., J.T. Moore, and D.W. Wright, Monoclonal antibody recognition of histidine-rich peptide encapsulated nanoclusters. Nano Letters, 2002. 2(3): p. 169-173. Zaytsev, D.V., et al., Metal-binding properties and structural characterization of a self-assembled coiled coil: Formation of a polynuclear Cd–thiolate cluster. Journal of Inorganic Biochemistry, 2013. 119(0): p. 1-9. Cui, Y., et al., Serial Silver Clusters Bio-Mineralized By One Peptide. Acs Nano, 2011. Jensen, K.J., Solid-phase peptide synthesis. Peptide Synthesis and Applications. 2013: Springer. Chan, W.C. and P.D. White, Fmoc solid phase peptide synthesis. 2000, New Yourk: Oxford University Press New York. Merrifield, R.B., Solid phase peptide synthesis. I. The synthesis of a tetrapeptide. Journal of the American Chemical Society, 1963. 85(14): p. 2149-2154. Carpino, L.A. and G.Y. Han, 9-Fluorenylmethoxycarbonyl function, a new base-sensitive aminoprotecting group. Journal of the American Chemical Society, 1970. 92(19): p. 5748-5749. Paquet, A., Introduction of 9-fluorenylmethyloxycarbonyl, trichloroethoxycarbonyl, and benzyloxycarbonyl amine protecting groups into O-unprotected hydroxyamino acids using succinimidyl carbonates. Canadian Journal of Chemistry, 1982. 60(8): p. 976-980. Baum, R., Combinatorial chemistry. Chemical & engineering news, 1996. 74(7). Barbas, C.F., et al., Semisynthetic combinatorial antibody libraries: a chemical solution to the diversity problem. Proceedings of the National Academy of Sciences, 1992. 89(10): p. 4457. Thompson, L.A. and J.A. Ellman, Synthesis and applications of small molecule libraries. Chemical reviews, 1996. 96(1): p. 555-600. Denholt, C.L., et al., Identification of novel peptide ligands for the cancer-specific receptor mutation EFGRvIII using a mixture-based synthetic combinatorial library. Biopolymers, 2009. 91(3): p. 201206. Hansen, C.L., et al., Identification of Amino Acid Residues in PEPHC1 Important for Binding to the Tumor-Specific Receptor EGFRvIII. Chemical biology & drug design, 2008. 72(4): p. 273-278. Lam, K.S., M. Lebl, and V. Krchnák, The “one-bead-one-compound” combinatorial library method. Chemical reviews, 1997. 97(2): p. 411-448. Chua, K.H., et al., IDENTIFICATION OF GOLD BINDING PEPTIDES USING A PHAGE DISPLAY PEPTIDE LIBRARY. PHYSICAL SCIENCES, 2008: p. 497. Zasloff, M., B. Martin, and H.C. Chen, Antimicrobial activity of synthetic magainin peptides and several analogues. Proceedings of the National Academy of Sciences, 1988. 85(3): p. 910. Lam, K.S., et al., A new type of synthetic peptide library for identifying ligand-binding activity. Nature, 1991. 354(6348): p. 82-84.
166
384. 385. 386. 387. 388. 389.
390. 391.
392.
393. 394. 395.
396.
397.
398.
399.
400.
401. 402.
Lowe, G., S. Leon, and R. Quarrell, Combinatorial Libraries for Studying Molecular Recognition. 1999. Lebl, M., et al., One-bead–one-structure combinatorial libraries. Biopolymers, 1995. 37(3): p. 177198. García-Martín, F., et al., ChemMatrix, a poly (ethylene glycol)-based support for the solid-phase synthesis of complex peptides. Journal of Combinatorial Chemistry, 2006. 8(2): p. 213-220. Albericio, F., et al., Solid-Phase Peptide Synthesis Using ChemMatrix®, a Polyethylenglycol (PEG)based Solid. Understanding Biology Using Peptides, 2006: p. 114-115. Meldal, M., Pega: a flow stable polyethylene glycol dimethyl acrylamide copolymer for solid phase synthesis. Tetrahedron letters, 1992. 33(21): p. 3077-3080. Wang, S.S., p-Alkoxybenzyl alcohol resin and p-alkoxybenzyloxycarbonylhydrazide resin for solid phase synthesis of protected peptide fragments. Journal of the American Chemical Society, 1973. 95(4): p. 1328-1333. Rink, H., Solid-phase synthesis of protected peptide fragments using a trialkoxy-diphenylmethylester resin. Tetrahedron letters, 1987. 28(33): p. 3787-3790. Bray, A.M., N.J. Maeji, and A.G. Jhingran, Gas phase cleavage of Peptides from a solid support with ammonia vapour. Application in simultaneous multiple Peptide Synthesis. Tetrahedron letters, 1991. 32(43): p. 6163-6166. Hammer, R.P., et al., Practical approach to solid-phase synthesis of C-terminal peptide amides under mild conditions based on a photolysable anchoring linkage1. International Journal of Peptide and Protein Research, 1990. 36(1): p. 31-45. Kenner, G., J. McDermott, and R. Sheppard, The safety catch principle in solid phase peptide synthesis. Journal of the Chemical Society D: Chemical Communications, 1971(12): p. 636-637. Patek, M. and M. Lebl, Safety-catch anchoring linkage for synthesis of peptide amides by Boc/Fmoc strategy. Tetrahedron Letters, 1991. 32(31): p. 3891-3894. Jensen, K.J., et al., Backbone Amide Linker (BAL) Strategy for Solid-Phase Synthesis of C-TerminalModified and Cyclic Peptides1, 2, 3. Journal of the American Chemical Society, 1998. 120(22): p. 5441-5452. LEBL, M., et al., Multiple release of equimolar amounts of peptides from a polymeric carrier using orthogonal linkage-cleavage chemistry. International Journal of Peptide and Protein Research, 1993. 41(2): p. 201-203. Leon, S., R. Quarrell, and G. Lowe, Evaluation of resins for on-bead screening: a study of papain and chymotrypsin specificity using PEGA-bound combinatorial peptide libraries. Bioorganic & Medicinal Chemistry Letters, 1998. 8(21): p. 2997-3002. Christensen, C.A. and M. Meldal, Efficient Solid-Phase Synthesis of Peptide-Based Phosphine Ligands: Towards Combinatorial Libraries of Selective Transition Metal Catalysts. Chemistry-a European Journal, 2005. 11(14): p. 4121-4131. Christensen, C.A. and M. Meldal, Solid-phase synthesis of a peptide-based P, S-ligand system designed for generation of combinatorial catalyst libraries. Journal of Combinatorial Chemistry, 2007. 9(1): p. 79-85. Miyatake, T., M. Nishihara, and S. Matile, A cost-effective method for the optical transduction of chemical reactions. Application to hyaluronidase inhibitor screening with polyarginine-counteranion complexes in lipid bilayers. Journal of the American Chemical Society, 2006. 128(38): p. 1242012421. Steinhagen, M., et al., Simultaneous “One Pot” Expressed Protein Ligation and CuI-Catalyzed Azide/Alkyne Cycloaddition for Protein Immobilization. ChemBioChem, 2011. 12(16): p. 2426-2430. Nielsen, T.E. and M. Meldal, Solid-Phase Synthesis of Pyrroloisoquinolines via the Intramolecular NAcyliminium Pictet-Spengler Reaction. Journal of Combinatorial Chemistry, 2005. 7(4): p. 599-610.
167
403.
404.
405. 406.
407. 408. 409. 410. 411. 412. 413. 414. 415. 416. 417. 418. 419. 420. 421. 422. 423. 424. 425.
Martínez-Ceron, M.C., et al., Sample preparation for sequencing hits from one-bead-one-peptide combinatorial libraries by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Analytical Biochemistry, 2010. 400(2): p. 295-297. Biederman, K., et al., Combinatorial peptide on-resin analysis: optimization of static nanoelectrospray ionization technique for sequence determination. The Journal of peptide research, 1999. 53(3): p. 234-243. Keller, B.O., et al., Interferences and contaminants encountered in modern mass spectrometry. analytica chimica acta, 2008. 627(1): p. 71-81. ABRF, Delta Mass - A Database of Protein Post Translational Modifications, L. Packman, Editor 2013, The Association of Biomelecular Resource Facilities: http://www.abrf.org/index.cfm/dm.home. Slocik, J.M. and D.W. Wright, Screening of Combinatorial Peptide Libraries for Nanocluster Synthesis, in Nanobiotechnology protocols. 2005, Springer. p. 133-141. Srivastava, M., et al., Cloning and sequencing of the human nucleolin cDNA. Febs Letters, 1989. 250(1): p. 99-105. Sedlak, R.H., et al., Engineered Escherichia coli silver-binding periplasmic protein that promotes silver tolerance. Applied and Environmental Microbiology, 2012. 78(7): p. 2289-2296. Silver, S., Bacterial silver resistance: molecular biology and uses and misuses of silver compounds. Fems Microbiology Reviews, 2003. 27(2-3): p. 341-353. Bao, Y.P., et al., Nanoparticle-free synthesis of fluorescent gold nanoclusters at physiological temperature. Journal of Physical Chemistry C, 2007. 111(33): p. 12194-12198. Petty, J.T., et al., DNA-templated Ag nanocluster formation. Journal of the American Chemical Society, 2004. 126(16): p. 5207-5212. Mejáre, M. and L. Bülow, Metal-binding proteins and peptides in bioremediation and phytoremediation of heavy metals. TRENDS in Biotechnology, 2001. 19(2): p. 67-73. Scheuhammer, A. and M.G. Cherian, Quantification of metallothioneins by a silver-saturation method. Toxicology and Applied Pharmacology, 1986. 82(3): p. 417-425. Pulido, P., J.H. Kägi, and B.L. Vallee, Isolation and Some Properties of Human Metallothionein*. Biochemistry, 1966. 5(5): p. 1768-1777. Peterson, C.W., S.S. Narula, and I.M. Armitage, 3D solution structure of copper and silversubstituted yeast metallothioneins. Febs Letters, 1996. 379(1): p. 85-93. Narula, S.S., et al., Establishment of the metal-to-cysteine connectivities in silver-substituted yeast metallothionein. Journal of the American Chemical Society, 1991. 113(24): p. 9354-9358. Naik, R.R., et al., Peptide templates for nanoparticle synthesis derived from polymerase chain reaction-driven phage display. Advanced Functional Materials, 2004. 14(1): p. 25-30. Gruen, L.C., Interaction of Amino-Acids with Silver(I) Ions. Biochimica Et Biophysica Acta, 1975. 386(1): p. 270-274. Stillman, M.J., C.F. Shaw, and K.T. Suzuki, Metallothionein: synthesis, structure, and properties of metallothioneins, phytochelatins, and metal-thiolate complexes. 1992: Wiley-VCH. Bar-Or, D., et al., Characterization of the Co2+ and Ni2+ binding amino-acid residues of the Nterminus of human albumin. European Journal of Biochemistry, 2001. 268(1): p. 42-48. Li, H.B., et al., Complexes of silver(I) with peptides and proteins as produced in electrospray mass spectrometry. Journal of the American Society for Mass Spectrometry, 1997. 8(8): p. 781-792. Vigneshwaran, N., et al., Biological synthesis of silver nanoparticles using the fungus Aspergillus flavus. Materials Letters, 2007. 61(6): p. 1413-1418. Behrens, S., et al., Nanoscale particle arrays induced by highly ordered protein assemblies. Advanced Materials, 2002. 14(22): p. 1621-1625. Xie, J.P., et al., Silver nanoplates: From biological to biomimetic synthesis. Acs Nano, 2007. 1(5): p. 429-439.
168
426. 427. 428. 429. 430. 431. 432. 433. 434. 435. 436. 437. 438. 439. 440. 441. 442. 443. 444. 445. 446.
447. 448.
449.
Mandal, S., S. Phadtare, and M. Sastry, Interfacing biology with nanoparticles. Current Applied Physics, 2005. 5(2): p. 118-127. Sarikaya, M., et al., Molecular biomimetics: nanotechnology through biology. Nature Materials, 2003. 2(9): p. 577-585. Tan, Y.N., J.Y. Lee, and D.I.C. Wang, Uncovering the Design Rules for Peptide Synthesis of Metal Nanoparticles. Journal of the American Chemical Society, 2010. 132(16): p. 5677-5686. Eckhardt, S., et al., Nanobio Silver: Its Interactions with Peptides and Bacteria, and Its Uses in Medicine. Chemical Reviews, 2013. Jover, J., R. Bosque, and J. Sales, A comparison of the binding affinity of the common amino acids with different metal cations. Dalton Transactions, 2008(45): p. 6441-6453. Reed, J.L., Hard and Soft Acids and Bases: Small Molecules. Inorganic Chemistry, 2009. 48(15): p. 7151-7158. Parr, R.G. and R.G. Pearson, Absolute Hardness - Companion Parameter to Absolute Electronegativity. Journal of the American Chemical Society, 1983. 105(26): p. 7512-7516. Pearson, R.G., Hard and Soft Acids and Bases. Journal of the American Chemical Society, 1963. 85(22): p. 3533-&. Gwinn, E.G., et al., Sequence-Dependent Fluorescence of DNA-Hosted Silver Nanoclusters. Advanced Materials, 2008. 20(2): p. 279-283. Petty, J.T., et al., DNA sensing by amplifying the number of near-infrared emitting, oligonucleotideencapsulated silver clusters. Analytical Chemistry, 2011. 83(15): p. 5957-5964. Ma, K., et al., Base-stacking-determined fluorescence emission of DNA abasic site-templated silver nanoclusters. Langmuir, 2012. 28(43): p. 15313-15322. Söptei, B., et al., On the selection and design of proteins and peptide derivatives for the production of photoluminescent, red-emitting gold quantum clusters. Gold Bulletin, 2013. 46(3): p. 195-203. Zheng, J., J.T. Petty, and R.M. Dickson, High quantum yield blue emission from water-soluble Au-8 nanodots. Journal of the American Chemical Society, 2003. 125(26): p. 7780-7781. Balogh, L., et al., Dendrimer-silver complexes and nanocomposites as antimicrobial agents. Nano Letters, 2001. 1(1): p. 18-21. Esumi, K., et al., Role of poly (amidoamine) dendrimers for preparing nanoparticles of gold, platinum, and silver. Langmuir, 2000. 16(6): p. 2604-2608. Fernandez, E.M., et al., Structural properties of bimetallic clusters from density functional calculations. International Journal of Modern Physics B, 2005. 19(15n17): p. 2339-2344. Fernández, E.M., et al., Trends in the structure and bonding of noble metal clusters. Physical Review B, 2004. 70(16): p. 165403. Aikens, C.M., Electronic structure of ligand-passivated gold and silver nanoclusters. The Journal of Physical Chemistry Letters, 2010. 2(2): p. 99-104. Wu, Z., et al., Sequential Observation of Ag n S4−(1≤ n≤ 7) Gas Phase Clusters in MS/MS and Prediction of Their Structures. The Journal of Physical Chemistry Letters, 2010. 1(9): p. 1423-1427. Wang, Y., et al., Bifunctional peptides that precisely biomineralize Au clusters and specifically stain cell nuclei. Chemical Communications, 2012. 48(6): p. 871-873. Létinois-Halbes, U., P. Pale, and S. Berger, Ag NMR as a tool for mechanistic studies of Ag-catalyzed reactions: Evidence for in situ formation of alkyn-1-yl silver from alkynes and silver salts. The Journal of Organic Chemistry, 2005. 70(23): p. 9185-9190. Shukla, S. and M. Sastry, Probing differential Ag+–nucleobase interactions with isothermal titration calorimetry (ITC): Towards patterned DNA metallization. Nanoscale, 2009. 1(1): p. 122-127. Millipore. ZipTip® Pipette Tips. 2013 [cited 2013; Available from: http://www.millipore.com/catalogue/module.do;jsessionid=CDF35ED85084CAD666E0933896CDB1 85?id=C5737#. Karas, M. and F. Hillenkamp, Laser desorption ionization of proteins with molecular masses exceeding 10,000 daltons. Analytical Chemistry, 1988. 60(20): p. 2299-2301. 169
450. 451. 452. 453. 454. 455.
456.
457. 458. 459. 460. 461. 462. 463. 464. 465.
466.
467. 468.
469. 470. 471. 472.
Pérez, J.M.M. and J. Pascau, Image Processing with ImageJ. 2013: Packt Publishing Ltd. Gorry, P.A., General least-squares smoothing and differentiation by the convolution (Savitzky-Golay) method. Analytical Chemistry, 1990. 62(6): p. 570-573. Savitzky, A. and M.J. Golay, Smoothing and differentiation of data by simplified least squares procedures. Analytical Chemistry, 1964. 36(8): p. 1627-1639. Yang, J.T., Calculation of protein conformation from circular dichroism. Methods Enzymol., 1986. 130: p. 208-269. Reed, J. and T.A. Reed, A set of constructed type spectra for the practical estimation of peptide secondary structure from circular dichroism. Analytical Biochemistry, 1997. 254(1): p. 36-40. Whitmore, L. and B. Wallace, DICHROWEB, an online server for protein secondary structure analyses from circular dichroism spectroscopic data. Nucleic Acids Research, 2004. 32(suppl 2): p. W668-W673. Lobley, A., L. Whitmore, and B. Wallace, DICHROWEB: an interactive website for the analysis of protein secondary structure from circular dichroism spectra. Bioinformatics, 2002. 18(1): p. 211212. Mao, D., E. Wachter, and B. Wallace, Folding of the mitochondrial proton adenosine triphosphatase proteolipid channel in phospholipid vesicles. Biochemistry, 1982. 21(20): p. 4960-4968. Brahms, S. and J. Brahms, Determination of protein secondary structure in solution by vacuum ultraviolet circular dichroism. Journal of Molecular Biology, 1980. 138(2): p. 149-178. Andrade, M., et al., Evaluation of secondary structure of proteins from UV circular dichroism spectra using an unsupervised learning neural network. Protein engineering, 1993. 6(4): p. 383-390. Greenfield, N.J., Using circular dichroism spectra to estimate protein secondary structure. Nature Protocols, 2007. 1(6): p. 2876-2890. Whitmore, L. and B.A. Wallace, Protein secondary structure analyses from circular dichroism spectroscopy: methods and reference databases. Biopolymers, 2008. 89(5): p. 392-400. Toniolo, C., et al., Circular dichroism spectrum of a peptide 310-helix. Journal of the American Chemical Society, 1996. 118(11): p. 2744-2745. Nishida, N., H. Yao, and K. Kimura, Chiral functionalization of optically inactive monolayer-protected silver nanoclusters by chiral ligand-exchange reactions. Langmuir, 2008. 24(6): p. 2759-2766. Cheng, J., et al., SCRATCH: a protein structure and structural feature prediction server. Nucleic Acids Research, 2005. 33(suppl 2): p. W72-W76. Pollastri, G., et al., Improving the prediction of protein secondary structure in three and eight classes using recurrent neural networks and profiles. Proteins: Structure, Function, and Bioinformatics, 2002. 47(2): p. 228-235. Prevelige Jr, P. and G.D. Fasman, Chou-Fasman prediction of the secondary structure of proteins, in Prediction of protein structure and the principles of protein conformation. 1989, Springer. p. 391416. Chou, P.Y. and G.D. Fasman, Prediction of the secondary structure of proteins from their amino acid sequence. 2006: Wiley Online Library. Garnier, J., D. Osguthorpe, and B. Robson, Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. Journal of Molecular Biology, 1978. 120(1): p. 97-120. Kaur, H., A. Garg, and G. Raghava, PEPstr: a de novo method for tertiary structure prediction of small bioactive peptides. Protein and Peptide Letters, 2007. 14(7): p. 626-631. Case, D., et al., AMBER 6. University of California, San Francisco, 1999. Maupetit, J., P. Derreumaux, and P. Tuffery, PEP-FOLD: an online resource for de novo peptide structure prediction. Nucleic Acids Research, 2009. 37(suppl 2): p. W498-W503. Thévenet, P., et al., PEP-FOLD: an updated de novo structure prediction server for both linear and disulfide bonded cyclic peptides. Nucleic Acids Research, 2012. 40(W1): p. W288-W293.
170
473. 474. 475. 476. 477. 478.
Alland, C., et al., RPBS: a web resource for structural bioinformatics. Nucleic Acids Research, 2005. 33(suppl 2): p. W44-W49. Néron, B., et al., Mobyle: a new full web bioinformatics framework. Bioinformatics, 2009. 25(22): p. 3005-3011. Camproux, A., R. Gautier, and P. Tuffery, A hidden markov model derived structural alphabet for proteins. Journal of Molecular Biology, 2004. 339(3): p. 591-605. Camproux, A., et al., Hidden Markov model approach for identifying the modular framework of the protein backbone. Protein engineering, 1999. 12(12): p. 1063-1073. Maupetit, J., P. Tuffery, and P. Derreumaux, A coarse-grained protein force field for folding and structure prediction. Proteins: Structure, Function, and Bioinformatics, 2007. 69(2): p. 394-408. Schrodinger, LLC, The PyMOL Molecular Graphics System, Version 1.3r1, 2010.
171
