Multifunctional compact zwitterionic ligands for preparing robust biocompatible semiconductor quantum dots and gold nanoparticles

ARTICLE

pubs.acs.org/JACS

Multifunctional Compact Zwitterionic Ligands for Preparing Robust

Biocompatible Semiconductor Quantum Dots and Gold Nanoparticles

Kimihiro Susumu,*^,†Eunkeu Oh,^†James B. Delehanty,^‡Juan B. Blanco-Canosa,^§Brandy J. Johnson,^‡

||

Vaibhav Jain,^†William Judson Hervey, IV,^‡W. Russ Algar,^‡,Kelly Boeneman,^‡Philip E. Dawson,^§

and Igor L. Medintz*^,‡

^†Optical Sciences Division, Code 5611, ^‡Center for Bio/Molecular Science and Engineering, Code 6900, U.S. Naval Research

Laboratory, Washington, D.C. 20375, United States

^§Departments of Cell Biology and Chemistry, The Scripps Research Institute, La Jolla, California 92037, United States

College of Science, George Mason University, Fairfax, Virginia 22030, United States

S Supporting Information

b

ABSTRACT: We describe the synthesis of a series of four

diﬀerent ligands which are used to prepare hydrophilic, bio-

compatible luminescent quantum dots (QDs) and gold nano-

particles (AuNPs). Overall, the ligands are designed to be

compact while still imparting a zwitterionic character to the

NPs. Ligands are synthesized appended to a bidentate dihy-

drolipoic acid- (DHLA) anchor group, allowing for high-aﬃnity

NP attachment, and simultaneously incorporate tertiary amines

along with carboxyl and/or hydroxyl groups. These are placed

in close proximity within the ligand structure and their capacity

for joint ionization imparts the requisite zwitterionic nature to

the nanocrystal. QDs functionalized with the four diﬀerent

compact ligands were subjected to extensive physical characterization including surface charge, wettability, hydrodynamic size, and

tolerance to a wide pH range or high salt concentration over time. The utility of the compact ligand coated QDs was further

examined by testing of direct conjugation to polyhistidine-appended protein and peptides, aqueous covalent-coupling chemistry,

and the ability to engage in F€orster resonance energy transfer (FRET). Conjugating cell penetrating peptides to the compact ligand

coated QD series facilitated their rapid and eﬃcient cellular uptake, while subsequent cytotoxicity tests showed no apparent

decreases in cell viability. In vivo biocompatibility was also demonstrated by microinjecting the compact ligand coated QDs into cells

and monitoring their stability over time. Inherent beneﬁts of the ligand design could be extended beyond QDs as AuNPs

functionalized with the same compact ligand series showed similar colloidal properties. The strong potential of these ligands to

expand NP capabilities in many biological applications is highlighted.

’ INTRODUCTION

NP surface; (iv) provide an available ‘handle’ for further chem-

ical modiﬁcation in a relatively simple manner; (v) have minimal

undesired interactions with other molecules native to biological

environments, that is, opsonization; (vi) have minimal to no

toxicity after in vivo administration; and ﬁnally, (vii) minimally

perturb inherent NP properties or those of an attached biomo-

lecule. Although all are important, it has been extremely challen-

ging to design NP ligands that can provide a majority of these

properties, let alone all, in a single molecule. Ligands do vary

widely in structure and, beyond being utilized for direct NP

synthesis, their interactions with NPs can be loosely grouped into

two functional mechanisms. The ﬁrst is encapsulation with

amphiphilic polymeric ligands or similar dendrimeric materials that

interact with and overcoat the native surface while simultaneously

Development of biocompatible nanoparticles (NPs) that

exploit the unique size-dependent properties of their constituent

materials in a biological context continues to be a major priority

in bionanotechnology.^1ꢀ8This goal, in turn, is almost completely

dependent upon the properties of the surface ‘ligands’ utilized to

provide aqueous stability to the NPs and facilitate their integra-

tion within biological systems. Many diﬀerent families of ligand

molecules are currently being utilized for these purposes with

each usually providing its own set of functional beneﬁts and

liabilities.^9,10Besides providing long-term colloidal stability

across a wide pH and buﬀer range to myriad types of mono-

dispersed NPs, the ligands should also have the following

features: (i) easily synthesized in bulk quantities; (ii) have a

small size that minimizes the hydrodynamic radius of the

subsequently functionalized NP; (iii) have high-aﬃnity attach-

ment to the NP while still allowing other molecules access to the

Received: March 2, 2011

Published: May 25, 2011

r

2011 American Chemical Society

9480

dx.doi.org/10.1021/ja201919s J. Am. Chem. Soc. 2011, 133, 9480–9496

|

Journal of the American Chemical Society

ARTICLE

the long chain length of the PEG unit increased the QD’s

hydrodynamic size and prohibited conjugation of His-tagged

proteins directly onto the QD surface.^18,23,31

Here, we report the design, synthesis, and functional charac-

terization of a new family of DHLA-based compact ligands which

can provide QDs with extended pH and salt stability without

increasing their apparent hydrodynamic sizes. In contrast to

utilizing a carboxyl or PEG moiety to mediate colloidal stability,

this ligand design relies on an inherent zwitterionic nature;

chemical structures for the surface compact ligand series (CL1

∼ CL4) are shown in Figure 1. Characterization of QDs coated

with the compact ligands (CL-QDs) included examining surface

charge, wettability, hydrodynamic volumes, pH and salt stability,

testing direct conjugation of His-tagged proteins and peptides,

aqueous covalent-coupling chemistry, cellular microinjection and

peptide-mediated cellular uptake, cellular imaging, and cytotoxi-

city. We also demonstrate that the properties of this compact

ligand series can be extended to AuNPs in a similar fashion.

Figure 1. Chemical structures of DHLA and the DHLA-based ligands

used in this study. DHLA-PEG750-OCH₃was synthesized as previously

described.^25,26

mediating colloidal stability. This, however, usually comes at the

expense of a much larger hydrodynamic size.¹¹In contrast, cap

exchange replaces the original hydrophobic surface ligands with

bifunctional hydrophilic ligands via mass action to render the

NPs water-soluble while keeping relatively small hydrodynamic

size.^9ꢀ11

’ MATERIALS AND METHODS

Chemicals. Thioctic acid, 1,1⁰-carbonyldiimidazole, N-(2-aminoethyl)-

piperazine, and methyl acrylate were purchased from Acros Organics

(Fisher Scientific, Pittsburgh, PA). Lithium hydroxide was purchased

from Sigma-Aldrich (St. Louis, MO). Sodium borohydride was pur-

chased from MP Biomedicals (Solon, OH). N-(2-Hydroxyethyl)-

ethylenediamine and ethylenediamine were purchased from TCI America

(Portland, OR). N,N-Bis(2-hydroxyethyl)ethylenediamine was purchased

from Alfa Aesar (Ward Hill, MA). N-Ethyl-N⁰-(3-dimethylaminopropyl)-

carbodiimide hydrochloride (EDC) and N-hydroxysulfosuccinimide

(sulfo-NHS) were purchased from Pierce Biotechnology (Rockford,

IL). Agarose (low EEO) was purchased from Fisher Scientific. All the

other chemicals including solvents were purchased from Sigma Aldrich

or Acros Organics and used as received. Phosphate buffered saline

(PBS) contains 137 mM NaCl, 10 mM phosphate, and 3 mM KCl

(pH 7.4).

Quantum Dots and Gold Nanoparticles. The 550 nm emitting

CdSe-ZnS coreꢀshell QDs were synthesized as previously described

with some modifications.^32,33The native surface-coating ligands after

the ZnS overcoating step are expected to be a mixture of n-hexylpho-

sphonic acid, trioctylphosphine oxide, and trioctylphosphine. The

615 nm emitting CdSe-CdS-CdZnS-ZnS coreꢀmultishell QDs were

synthesized via modification of published procedures combining suc-

cessive ion layer adsorption and reaction (SILAR) with thermal cycling

techniques.^34ꢀ36The native surface coating ligands of these multishell

QDs following the final ZnS overcoating are expected to be a mixture of

n-dodecylphosphonic acid, diisooctylphosphinic acid, oleic acid, and

trioctylphosphine. Custom Qdot 625 nm emitting ITK organic QDs

used in the microinjection studies were a gift from Life Technologies

(Eugene, OR). Sodium citrate stabilized 5.8 nm diameter gold nano-

particles (AuNPs) were purchased from Ted Pella (Redding, CA).

Ligand Synthesis. Compounds 1 and 2. Thioctic acid (3.00 g,

1.45 ꢁ 10^ꢀ2mol) and 1,1⁰-carbonyldiimidazole (CDI) (2.59 g, 1.60 ꢁ

10^ꢀ2mol) were added to a 100-mL round-bottom flask, and the mixture

was purged with N₂. 30 mL of toluene was added to the mixture by

syringe, and the reaction mixture stirred at room temperature for 1 h

under N₂. The mixture was transferred to an addition funnel and CHCl₃

(20 mL) was further mixed in, and then this solution was added dropwise

over 30 min to a mixture of N-(2-aminoethyl)piperazine (2.10 mL,

1.59 ꢁ 10^ꢀ2mol) and 30 mL of CHCl₃in a 250-mL two-neck round-

bottom flask cooled in an ice-bath under N₂. The reaction mixture was

further stirred at room temperature overnight and transferred to a

separatory funnel. CHCl₃(∼100 mL) and deionized (DI) water

Our research has focused on developing hydrophilic lumines-

cent semiconductor nanocrystals or quantum dots (QDs) for a

variety of biological applications including use in cellular de-

livery, biosensing, functional nanodevices, and as multiplexed

probes.^4,9,12ꢀ15This goal has required a parallel focus on

developing improved ligands for the QDs. The ﬁrst generation of

ligand utilized dihydrolipoic acid (DHLA), see Figure 1.¹⁶In

comparison to monothiol-ligands which suﬀer from a high

dynamic dissociation rate and thus provide poor long-term

colloidal stability to QDs, the bidentate-thiols essentially func-

tion as cooperative anchoring groups that strongly coordinate to

the nanocrystal surface with the charged carboxyl group at the

molecule’s other end mediating colloidal stability.¹⁶This ligand is

easily prepared in one step from commercially available thioctic

acid (TA) by reduction of the terminal disulﬁde with sodium

borohydride (NaBH₄). Importantly, the small, compact size of

DHLA ligands also allows direct self-assembly of a variety of

oligohistidine-tagged (His-tagged) peptides, proteins, and DNA

onto the QD surface.^17ꢀ21However, aqueous solubility of

DHLA-coated QDs relies exclusively on deprotonation of the

terminal carboxyl groups and their colloidally stable pH window

is limited to the basic regime. In contrast, amine-coated QDs are

colloidally stable at moderately acidic pH; however, their stability

drops when moving from neutral to basic conditions.²²The

second generation of ligand mediated solubility by appending

poly(ethylene glycol) (PEG) moieties (molecular weight ran-

ging from 400 up to 1000) to the DHLA anchor group and

provided QDs with a much wider range of pH stability.²³

Improvements to this ligand design were more modular in nature

and replaced an internal acid-labile ester linkage with an amide

bond and provided the option of a wide range of terminal

functional groups ranging from biotin to a primary amine,

carboxyl, or methoxy, the former of which allows for bioconjuga-

tion and covalent modiﬁcations.^24ꢀ27Subsequent work demon-

strated that the beneﬁts of this PEGylated-ligand design could

also be extended to gold nanoparticles (AuNPs).^25,28ꢀ30While

the PEG chain extended pH and intracellular stability of the QDs,

9481

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

(Milli-Q 18 MΩ, ∼100 mL) were added to the mixture and the CHCl₃

layer was collected. The aqueous layer was further washed with CHCl₃

(2 times). The combined organic layers were dried over Na₂SO₄,

filtered, and concentrated to ∼20 mL on a rotary evaporator. The crude

product solution was purified over silica gel column chromatography

with CHCl₃/MeOH (5:1) as the eluent. The product fractions were

concentrated to ∼10 mL and directly used for the next reaction.

Product solution was dissolved in 80 mL of MeOH and purged with

N₂. Methyl acrylate (5.0 mL, 5.6 ꢁ 10^ꢀ2mol) was added dropwise and

the reaction mixture stirred at room temperature overnight under N₂.

The solvent and excess methyl acrylate was subsequently evaporated,

and the residue was separated by chromatography on silica gel with

CHCl₃/MeOH (20:1) as the eluent. Yield = 3.448 g (59% in two steps

based on 3.00 g of thioctic acid). N-[2-(1-Piperazinyl)ethyl]-5-(1,2-

dithiolan-3-yl)pentanamide (1): ¹H NMR (400 MHz, CDCl₃): δ 6.05

(br s, 1H, NH), 3.52ꢀ3.63 (m, 1H), 3.35 (q, 2H, J = 7.6 Hz, ꢀCH₂ꢀ),

3.07ꢀ3.23 (m, 2H), 2.89 (t, 4H, J = 6.4 Hz, ꢀCH₂ꢀ), 2.37ꢀ2.53 (m,

7H), 2.20 (t, 2H, J = 10.0 Hz, ꢀCH₂ꢀCOꢀ), 1.86ꢀ1.97 (m, 1H),

1.55ꢀ1.8 (m, 4H), 1.4ꢀ1.55 (m, 2H). Product was conﬁrmed by

electrospray ionization mass spectrometry (ESI MS) with a mass-to-

charge ratio (m/z) of 318.1743 (M þ H)^þ/theoretical protonated mono-

isotopic mass (M þ H)^þwas 318.167. N-{2-[4-(methoxycarbonylethyl)

piperazin-1-yl]ethyl}-5-(1,2-dithiolan-3-yl)pentanamide (2): ¹H NMR

(400 MHz, CDCl₃): δ 6.08 (br s, 1H, NH), 3.69 (s, 3H, ꢀOCH₃),

3.52ꢀ3.63 (m, 1H), 3.34 (q, 2H, J = 7.6 Hz, ꢀCH₂ꢀ), 3.07ꢀ3.23 (m,

2H), 2.71 (t, 2H, J = 9.6 Hz, ꢀCH₂ꢀ), 2.40ꢀ2.58 (m, 13H, ꢀCH₂ꢀ),

2.20 (t, 2H, J = 9.8 Hz, ꢀCH₂ꢀCOꢀ), 1.86ꢀ1.97 (m, 1H), 1.6ꢀ1.8

(m, 4H), 1.4ꢀ1.55 (m, 2H). ESI MS m/z: 404.2018 (M þ H)^þ/

theoretical protonated monoisotopic mass 404.204.

Hz, ꢀCH₂ꢀ), 3.07ꢀ3.23 (m, 2H), 2.80 (q, 2H, J = 8.2 Hz, ꢀCH₂ꢀ),

2.60 (q, 4H, J = 6.8 Hz, ꢀCH₂ꢀ), 2.49 (t, 2H, J = 8.0 Hz, ꢀCH₂ꢀ),

2.41ꢀ2.53 (m, 1H), 2.21 (t, 2H, J = 10.0 Hz, ꢀCH₂ꢀCOꢀ), 1.86ꢀ1.97

(m, 1H), 1.6ꢀ1.8 (m, 4H), 1.4ꢀ1.55 (m, 2H). ESI MS m/z: 379.1720

(M þ H)^þ/theoretical protonated monoisotopic mass 379.173.

Compound 7. Thioctic acid (2.00 g, 9.69 ꢁ 10^ꢀ3mol) and CDI (1.73 g,

1.07 ꢁ 10^ꢀ2mol) were added to a 100-mL round-bottom flask, and

the mixture was purged with N₂. 20 mL of CHCl₃was added to the

mixture by syringe and the reaction mixture stirred at room temperature

for 1 h under N₂. The mixture above was transferred to an addition

funnel and 5.0 mL of CHCl₃was further mixed in, and the resulting

solution added dropwise over 30 min to a mixture of N,N-bis(2-

hydroxyethyl)ethylenediamine (1.60 g, 1.08 ꢁ 10^ꢀ2mol) and 20 mL

of CHCl₃in a 250-mL two-neck round-bottom flask under N₂. The

reaction mixture was further stirred at room temperature overnight

and poured into DI water (∼100 mL). The CHCl₃layer was collected

and the product was further extracted with CHCl₃(2 times). The

combined organic layers were dried over Na₂SO₄and filtered. The

solvent was evaporated and the crude product was purified by chroma-

tography on silica gel with CHCl₃/MeOH (20:1) as the eluent. Yield =

2.614 g (80% based on 2.00 g of thioctic acid). N-{2-[N⁰,N⁰-Bis(2-

hydroxyethyl)amino]ethyl}-5-(1,2-dithiolan-3-yl)pentanamide (7):

¹H NMR (400 MHz, CDCl₃): δ 6.36 (br s, 1H, NH), 3.63 (t, 4H,

J = 6.8 Hz, ꢀCH₂ꢀ), 3.53ꢀ3.62 (m, 1H), 3.36 (q, 2H, J = 7.6 Hz,

ꢀCH₂ꢀ), 3.07ꢀ3.23 (m, 2H), 2.63ꢀ2.73 (m, 6H, ꢀCH₂ꢀ), 2.41ꢀ2.52

(m, 1H), 2.21 (t, 2H, J = 10.0 Hz, ꢀCH₂ꢀCOꢀ), 1.86ꢀ1.97 (m, 1H),

1.55ꢀ1.8 (m, 4H), 1.4ꢀ1.55 (m, 2H). ESI MS m/z: 337.1788

(M þ H)^þ/theoretical protonated monoisotopic mass 337.162.

Compounds 8 and 9. Thioctic acid (3.00 g, 1.45 ꢁ 10^ꢀ2mol) and

CDI (2.59 g, 1.60 ꢁ 10^ꢀ2mol) were added to a 100-mL round-bottom

flask and the mixture purged with N₂. 30 mL of CHCl₃was added to the

mixture by syringe and the reaction mixture stirred at room temperature

for 1 h under N₂. The reaction solution was then transferredto an addition

funnel and added dropwise over 1 h to a mixture of ethylenediamine

(8.0 mL, 0.12 mol) and 30 mL of CHCl₃in a 250-mL two-neck round-

bottom flask at room temperature under N₂. The reaction mixture was

further stirred at room temperature overnight and transferred to a

separatory funnel. DI water (∼100 mL) was added to the mixture and

the CHCl₃layer was collected. The aqueous layer was further washed with

CHCl₃(3 times). The combined organic layers were concentrated to

∼20 mL and the crude product solution was purified by chromatography

onsilicagelwith CHCl₃/MeOH (5:1) astheeluent. Theproduct fractions

were concentrated to ∼15 mL and used for the next reaction step.

Product solution was dissolved in 70 mL of MeOH and purged with

N₂. Methyl acrylate (10.0 mL, 0.11 mol) was added dropwise and the

reaction mixture was stirred at room temperature for 2 days under N₂.

The solvent and excess methyl acrylate was evaporated. The residue was

separated by chromatography on silica gel with CHCl₃/MeOH (20:1)

as the eluent. Yield = 3.919 g (64% in 2 steps based on 3.00 g of thioctic

acid). N-(2-Aminoethyl)-5-(1,2-dithiolan-3-yl)pentanamide (8): ¹H

NMR (400 MHz, CDCl₃): δ 6.06 (br s, 1H, NH), 3.53ꢀ3.62 (m,

1H), 3.31 (q, 2H, J = 7.7 Hz, ꢀCH₂ꢀ), 3.07ꢀ3.26 (m, 2H), 2.84 (t, 2H,

J = 7.8 Hz, ꢀCH₂ꢀ), 2.41ꢀ2.52 (m, 1H), 2.21 (t, 2H, J = 10.0

Hz, ꢀCH₂ꢀCOꢀ), 1.86ꢀ1.97 (m, 1H), 1.6ꢀ1.8 (m, 4H), 1.4ꢀ1.6

(m, 2H). ESI MS m/z: 249.0941 (M þ H)^þ/theoretical protonated

monoisotopic mass 249.110. N-{2-[N⁰,N⁰-Bis(methoxycarbonylethyl)-

amino]ethyl}-5-(1,2-dithiolan-3-yl)pentanamide (9): ¹H NMR (400

MHz, CDCl₃): δ 6.52 (br s, 1H, NH), 3.69 (s, 6H, -OCH₃), 3.53ꢀ

3.63 (m, 1H), 3.33 (q, 2H, J = 7.2 Hz, ꢀCH₂ꢀ), 3.07ꢀ3.22 (m, 2H),

2.74 (t, 4H, J = 8.6 Hz, ꢀCH₂ꢀ), 2.53 (t, 2H, J = 7.2 Hz, ꢀCH₂ꢀ),

2.47 (m, 1H), 2.43 (t, 4H, J = 8.6 Hz, ꢀCH₂ꢀ), 2.22 (t, 2H, J = 9.8 Hz,

ꢀCH₂ꢀCOꢀ), 1.85ꢀ1.97 (m, 1H), 1.6ꢀ1.8 (m, 4H), 1.4ꢀ1.6

(m, 2H). ESI MS m/z: 421.1647 (M þ H)^þ/theoretical protonated

monoisotopic mass 421.183.

Compounds 4 and 5. Thioctic acid (4.00 g, 1.94 ꢁ 10^ꢀ2mol) and

CDI (3.46 g, 2.13 ꢁ 10^ꢀ2mol) were added to a 100-mL round-bottom

flask, and the mixture was purged with N₂. 50 mL of CHCl₃was added to

the mixture by syringe and the reaction mixture stirred at room

temperature for 1 h under N₂. The mixture was then transferred to an

addition funnel where 20 mL of CHCl₃was further mixed in, and

the resulting solution was added dropwise over 30 min to a mixture of

N-(2-hydroxyethyl)ethylenediamine (2.20 mL, 2.18 ꢁ 10^ꢀ2mol) and

40 mL of CHCl₃in a 250-mL two-neck round-bottom flask cooled in an

ice-bath under N₂. The reaction mixture was further stirred at room

temperature for 5 h and transferred to a separatory funnel. Saturated

sodium bicarbonate solution (∼100 mL) was added to the mixture and

the CHCl₃layer was collected. The aqueous layer was further washed

with CHCl₃(4 times). The combined organic layers were dried over

Na₂SO₄, filtered, and concentrated to ∼20 mL by rotary evaporator.

The crude product solution was purified by chromatography on silica

gel with CHCl₃/MeOH (5:1) as the eluent. The product fractions were

concentrated to ∼10 mL and used for the next reaction.

Product solution was dissolved in 80 mL of MeOH and purged with

N₂. Methyl acrylate (9.0 mL, 0.10 mol) was added dropwise and the

reaction mixture was stirred at room temperature overnight under N₂.

The solvent and excess methyl acrylate were evaporated. The residue

was again puriﬁed by chromatography on silica gel with CHCl₃/MeOH

(10:1) as the eluent. Yield = 3.584 g (49% in two steps based on 4.00 g of

thioctic acid). N-{2-[(2-Hydroxyethyl)amino]ethyl}-5-(1,2-dithiolan-

1

3-yl)pentanamide (4): H NMR (400 MHz, CDCl₃): δ 5.97 (br s,

1H, NH), 3.68 (t, 2H, J = 6.8 Hz, ꢀCH₂ꢀ), 3.53ꢀ3.62 (m, 1H), 3.37 (q,

2H, J = 7.6 Hz, ꢀCH₂ꢀ), 3.07ꢀ3.23 (m, 2H), 2.79 (t, 2H, J =

6.8 Hz, ꢀCH₂ꢀ), 2.78 (t, 2H, J = 7.8 Hz, ꢀCH₂ꢀ), 2.41ꢀ2.52 (m,

1H), 2.20 (t, 2H, J = 9.8 Hz, ꢀCH₂ꢀCOꢀ), 1.86ꢀ1.97 (m, 1H),

1.55ꢀ1.8 (m, 4H), 1.4ꢀ1.55 (m, 2H). ESI MS m/z: 293.1206 (M þ

H)^þ/theoretical protonated monoisotopic mass 293.136. N-{2-[N⁰-(2-

Hydroxyethyl)-N⁰-(methoxycarbonylethyl)amino]ethyl}-5-(1,2-dithio-

lan-3-yl)pentanamide (5): ¹H NMR (400 MHz, CDCl₃): δ 6.39 (br s,

1H, NH), 3.72 (s, 3H, ꢀOCH₃), 3.52ꢀ3.63 (m, 3H), 3.33 (q, 2H, J = 7.6

9482

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

Ester Hydrolysis, Ring-Opening, and Cap Exchange of

Compact Ligands onto Quantum Dots. Typical procedures for

preparing a ligand for cap exchange and the subsequent cap exchange

reaction are as follows: LiOH (20.9 mg, 8.71 ꢁ 10^ꢀ4mol) was added to a

mixture of 2 (0.307 g, 7.61 ꢁ 10^ꢀ4mol), EtOH (2.0 mL), and DI H₂O

(1.0 mL). The reaction mixture was stirred at room temperature for 2 h.

Completion of ester hydrolysis was checked by TLC (thin layer

chromatography) with CHCl₃/MeOH (5:1) as the eluent. For reducing

the dithiolane group, 4 M HCl was then added dropwise to the reaction

mixture to adjust the pH to ∼8, and NaBH₄(60.6 mg, 1.60 ꢁ 10^ꢀ3mol)

was added to the aqueous solution, which was further stirred at room

temperature for 1.5 h under N₂. Then, 4 M HCl was added dropwise to

the reaction mixture to adjust the pH to 7ꢀ8. The EtOH was

evaporated, and the remaining aqueous solution was filtered through a

cotton plug to remove precipitated salts. The filtered aqueous solution

was sealed in a glass vial with a septum and purged with N₂. Cap

exchange utilized a biphasic mixture, see Supplementary Figure S1 for a

schematic. QDs capped with native hydrophobic ligands (∼7.7 nmol in

1.0 mL of CHCl₃) were injected by syringe into the aqueous ligand

solution with vigorous stirring. The biphasic mixture was stirred at room

temperature (or at ∼50 °C) overnight under N₂. The CHCl₃layer was

discarded after collecting by syringe and the residual CHCl₃in the

aqueous layer was removed by evaporation. The aqueous layer was then

filtered through a Millex-LCR membrane filter (pore size 0.45 μm,

Millipore) and transferred to a centrifugal spin dialyzer (Amicon Ultra

50K, Millipore). The mixture was diluted with DI water and centrifuged

at 3400 rpm for 5ꢀ10 min, and the clear, filtered solution was discarded.

To remove excess unbound ligands and other byproducts, the QD

dispersion was subject to a few additional rounds of centrifugation with

DI water.

Cap Exchange of Compact Ligands onto Gold Nanopar-

ticles. For cap exchange with AuNPs, the ligands prepared by ester

hydrolysis step were used without the subsequent ring-opening reaction

of the dithiolane groups due to the much stronger intrinsic AuꢀS

interaction.^28,29Typical procedures for preparing ligand for cap ex-

change with AuNPs and the subsequent cap exchange reaction are as

follows: LiOH (3.4 mg, 1.4 ꢁ 10^ꢀ4mol) was added to a mixture of 2

(38.0 mg, 9.42 ꢁ 10^ꢀ5mol) and DI H₂O (1.0 mL). The reaction

mixture was stirred at room temperature for 2 h. Completion of the ester

hydrolysis was checked by TLC with CHCl₃/MeOH (5:1) as the eluent.

Then, 4 M HCl was added dropwise to the reaction mixture to adjust the

pH to ∼8. Citrate-stabilized AuNPs (5.8 nm diameter, 3.5 mL, 2.8 ꢁ

10^ꢀ10mol) were added to the ligand solution (in case of ligand 7, 0.1 M

NaOH was further added to adjust the pH of the reaction mixture to

∼9), and the reaction mixture was stirred at room temperature over-

night. The aqueous solution was filtered through a Millex-LCR mem-

brane filter (pore size 0.45 μm, Millipore) and transferred to a

centrifugal spin dialyzer (Amicon Ultra 50K, Millipore). The mixture

was diluted with DI water and centrifuged at 3200 rpm for 5ꢀ10 min,

and the clear, filtered solution was discarded. To remove excess unbound

ligands and other byproducts, the AuNP dispersion was subject to a few

additional rounds of centrifugation with DI water.

stability while making overall size significantly smaller, see

Figure 1 and Table 1. For inspiration, we drew directly from

the properties of peptides and proteins. Their excellent aqueous

solubility is partially attributable to the zwitterionic character of

their amino acids; that is, the amines and carboxylic acids

simultaneously present within most constituent residues can

both be ionized.^37,38Tertiary amine, carboxyl, and hydroxyl

groups were chosen as the hydrophilic components to mimic

the zwitterionic properties within the ligands. The acid dissocia-

tion constant (pK_a) of an alkyl carboxylic acid is ∼4.8,³⁹which

means that it will be mostly deprotonated (ionized) at pHs above

the pK_avalue. On the other hand, the pK_aof a tertiary amine

group is ∼10.7⁴⁰and it will be mostly protonated (ionized) at

pHs below the pK_avalue. If these two functional groups are

combined within a single ligand, it is reasonable to expect that the

resulting ligand could hold an ionized state over a wide range of

pH, though the actual pK_avalues in the microenvironments

present on QD surfaces may be different from that of the

individual group. Alternatively, use of hydroxyl groups is also

expected to help aqueous solubility under a wide range of pH

because of their strong inherent tendency to form hydrogen

bonds with water.⁴¹

The basic CL structure and function continues with a modular

design similar to that found in the previous PEGylated ligand

derivatives.²⁴LigandꢀQD interactions are provided by the

DHLA’s bidentate dithiol motif acting as the ﬁrst module. This

high-aﬃnity attachment is still superior to that of QDs coated

with monothiol ligands in aqueous solutions. Rather than

depending upon PEG to mediate aqueous solubility in a second

module, the new ligands incorporate at least two of the functional

groups described above to maintain aqueous solubility in a wide

pH range mainly via a zwitterionic character. These groups are

purposely designed to serve multiple, simultaneous roles. Besides

imparting aqueous solubility under basic pH, the carboxyl groups

can be used for further coupling chemistry with biological

molecules or organic dyes; this provides a third module or utility

to the ligands. The tertiary amino group is expected to enhance

aqueous solubility under acidic pH and remain chemically inert

to common bioconjugation chemistries such as carbodiimide (e.

g., EDC) coupling. Use of the tertiary amine also prevents EDC-

based cross-linking (to carboxyls) between the surface ligands on

QDs and negates potential issues of cellular toxicity previously

noted in the presence of ligands displaying primary amines.^42,43

Four compact ligands were synthesized using these design

criteria (Figure 1). CL1 incorporates two tertiary amines in a

piperazine ring structure and terminates in an alkyl carboxyl

group. CL2 includes a single tertiary amine that displays both an

alkyl carboxyl and a hydroxyl group. CL3 is almost identical to

CL2 but ends in two hydroxyl groups. This ligand essentially

functions as (i) a control to conﬁrm the necessity of having both a

carboxyl and a tertiary amine simultaneously present in the same

ligand, and (ii) to verify if the carboxyl group(s) could be

substituted with a hydroxyl function. DHLA which represents

the non-amine control ligand has already been extensively

characterized.¹⁶CL4 continues with the same CL2 and CL3

motif, but terminates in two alkyl carboxyl groups. The estimated

pK_a’s of the key ionizable groups on each ligand are given in

Table 1. Importantly, CL1, CL2 and CL4 all display values in

both the acidic and basic pH regime matching our key design

criteria. In terms of size and mass, CL1ꢀCL4 are estimated to be

up to twice the ∼1.0 nm extended length of DHLA but less than

twice the mass, see Table 1. This contrasts strongly with the

Detailed descriptions of the analytical instrumentation, and charac-

terization procedures including pH and salt stability tests; contact

angles; protein preparation and labeling; His_n-based self-assembly;

agarose gel electrophoresis; EDC coupling; F€orster resonance energy

transfer (FRET) analysis; along with cell culture procedures including

microinjection, imaging QD uptake with cell penetrating peptide, and

cytotoxicity test can be found in the Supporting Information.

’ RESULTS AND DISCUSSION

Design of the Compact Ligands. The primary design criteria

for the CL series synthesized here was to enhance QD colloidal

9483

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

Table 1. Selected Ligand or Quantum Dot Physical Properties

DHLA

CL1

CL2

CL3

338.54

CL4

394.56

DHLA-PEG750-OCH₃

Molecular weight

208.34

391.60

366.55

926.24 (n = 15)

Synthetic steps

1

4

2

4

5

Extended length^a

pK_a^b

∼1.0 nm

∼2.0 nm

2.87 (amine)

4.55 (carboxyl)

7.58 (amine)

Yes

∼1.8 nm

3.59 (carboxyl)

8.58 (amine)

∼1.6 nm

7.85 (amine)

14.37 (OH)

14.96 (OH)

No

∼1.8 nm

3.29 (carboxyl)

3.93 (carboxyl)

9.06 (amine)

Yes

∼6.8 nm

4.74 (carboxyl)

Dithiols (10.8/10.2)

Central amide (16.3)

Chemical handle

ꢀ

No

Yes

Assembles (His)_npeptide/protein

Contact angles^c

þ/þ

5°

þ/þ

10°

þ/þ

8°

þ/þ

6°

þ/þ

13°

þ/ꢀ

20°

pH stability range over 4 weeks

Zeta potential (mV)^dat pH 8.3

Zeta potential (mV)^dat pH ∼4

DLS (% Intensity)^e

DLS (% Volume)^e

7ꢀ13 (4 days)

ꢀ20.3 ( 3.9

ꢀ

10.8 ( 2.70

9.2 ( 2.36

8.1 ( 1.82

4 (∼5)ꢀ13

ꢀ39.0 ( 1.9

10.8 ( 0.7

8.6 ( 1.76

7.9 ( 1.82

7.2 ( 1.62

4ꢀ13

5ꢀ13

4ꢀ11

ꢀ

ꢀ19.8 ( 1.6

10.3 ( 0.7

9.3 ( 1.66

8.6 ( 1.77

8.0 ( 1.57

ꢀ4.9 ( 0.8

22.4 ( 2.1

9.5 ( 2.10

8.6 ( 2.06

7.8 ( 1.76

ꢀ50.8 ( 2.7

ꢀ19.0 ( 2.2

9.8 ( 2.17

8.6 ( 2.08

7.7 ( 1.76

ꢀ

11.5 ( 2.46

10.2 ( 2.39

9.2 ( 2.01

DLS (% Number)^e

^aEstimated edge-to-edge molecular length in fully extended structure. ^bpK_avalues were calculated by ACD/PhysChem Suite, version 12.01, Advanced

Chemistry Development, Inc., Toronto, ON, Canada. ^cContact angle of water on mica surfaces coated by QDs displaying each ligand. Std. Dev. <10%

for all samples. ^dMeasured in 1ꢁ TBE buﬀer pH 8.3 and acidiﬁed water pH ∼4 adjusted with HCl. DHLA-QDs were not measured due to their colloidal

instability in the acidic environment. ^eBecause of the diverse ways to represent data, the hydrodynamic diameters characterized by diﬀerent proﬁles are

shown here for comparison. Intensity proﬁle is used for the present discussion.

DHLA-PEG750-OCH₃sample, which we use frequently^25,44and

which has an estimated geometric length of up to 6.8 nm (∼6ꢀ7

times longer than that of DHLA) along with a mass ∼4.5 times

larger than that of DHLA.

exchange precursors 2, 5, 7, and 9), clearly fulfilling the desired

criteria of facile synthesis of large quantities (gram scale). Final

CL products were used directly for the cap exchange reaction

with QDs coated with the native hydrophobic ligands. All the

compounds (2, 5, 7, 9) utilized for the subsequent cap exchange

were found to be stable for at least 6 months to a year when

stored in a refrigerator.

Given the high reaction yield of the ligands and their presence

in aqueous solution, we opted to continue working in this media

and utilize a biphasic mixture approach^46ꢀ48for cap exchange

rather than the single-phase method used previously.²⁴The cap

exchange reaction was initiated by injecting the hydrophobic

QDs dissolved in CHCl₃into the aqueous ligand solution under

a N₂atmosphere with vigorous stirring. As the reaction proceeds,

the QDs were gradually transferred from the CHCl₃layer to the

aqueous layer in a few hours. The reaction mixture was left with

vigorous stirring overnight (Supporting Information, Figure S1).

Puriﬁed QDs were found to remain bright and be well dispersed

in water and colloidally stable over 6 months without any sign of

aggregation when stored in a refrigerator (see below).

Characterization of Compact Ligand-Coated Quantum

Dots. Spectroscopic Properties. We next subjected the aqueous-

dispersed CL-QDs to a variety of photophysical analyses to

characterize the properties the ligands impart to the nanocrystals.

Absorption and photoluminescence (PL) spectra were measured

for QDs capped with the native hydrophobic ligands in hexane

prior to cap-exchange and compared to those functionalized with

the CLs in H₂O (Supporting Figure S2). The absorption spectra

measured before and after cap exchange were essentially un-

changed, and occasionally showed only a few nanometers shift of

the lowest absorption maxima from batch to batch; PL spectra

also showed similar trends. These spectroscopic features are quite

similar to our previous QD studies with different DHLA-PEG

ligands and indicate that cap exchange with the new CL series

has no significant optical effect on the inorganic QD coreꢀshell

structure.^24,25It is also quite common to see a decrease in

Synthesis of Compact Ligands and Cap Exchange in a

Biphasic Mixture. An overview of the synthetic procedures

utilized to prepare each ligand is shown in Scheme 1. We carefully

optimized the reaction conditions and minimized the number of

reaction steps. This is an important issue since subsequent QD

cap exchange is a mass-driven reaction and requires a large excess

of ligand per QD to ensure full overcoating. Synthesis of all the

new compact ligands starts from modification of the carboxyl

group of thioctic acid (TA) with a primary amine compound.

Use of 1,1⁰-carbonyldiimidazole (CDI) as a coupling reagent is

key to the chemistry utilized here as it exclusively gives a primary

amine-coupled product in high yield in the presence of a

secondary amine group (1 and 4).⁴⁵This reaction step signifi-

cantly simplifies the entire synthetic process by avoiding the need

for protecting groups. Subsequent Michael addition reactions of

1, 4, and 8 with methyl acrylate formed 2, 5, and 9, respectively.

The last steps of the ligand synthetic procedures are (i) depro-

tection of the methyl ester group to yield the carboxyl group (2 to

3, 5 to 6, and 9 to 10) and (ii) the ring-opening reduction

from a disulfide group to a dithiol group (3 to CL1, 6 to CL2, 7 to

CL3 and 10 to CL4). Since both of these reactions are simple

and quantitative, and each product is highly soluble in water,

the reactions were successively performed in aqueous solution.

Each ligand required four synthetic preparatory steps except

CL3 which only necessitated two steps. This includes the ring-

opening step which is necessary when cap exchanging QDs

with all DHLA-based ligands. Importantly, we should note that

cap exchange with citrate-stabilized AuNPs can be carried out

without the last ring-opening step from TA to DHLA derivatives

due to the strong native AuꢀS interaction (vide infra).^28,29Each

synthetic step for CL1ꢀCL4 was straightforward and had a

reasonably high yield (g50% total yields for synthesis of the cap

9484

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

Scheme 1. Synthetic Schemes for the Compact Ligands Used in This Study: (A) CL1, (B) CL2, (C) CL3, and (D) CL4

ARTICLE

fluorescence quantum yield (QY) of the QD samples following

cap exchange and phase-transfer from organic to aqueous media

due to incomplete surface passivation and charge effects. Again,

similar to our previous reports, we note that the relative QYs after

cap exchange dropped to circa half of that measured for the native

QD in hydrophobic solvents. QYs for the CdSe-ZnS coreꢀshell

QDs used in this study were 25ꢀ35% (in hexane) and 10ꢀ20%

(in water), respectively. As the QD’s inherent QY is also a

function of the size and integrity of the inorganic shell materials,⁴⁶

the CdSe-CdS-CdZnS-ZnS QDs overcoated with several layers

of shell materials displaying a gradient of band gap offsets had

slightly higher QYs. Following cap exchange and phase-transfer,

the QY of these QDs changed from 45ꢀ55% (hexane) to

25ꢀ40% (water). We note that the Qdot 625 nm emitting ITK

9485

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

Figure 2. Physical characterization of a series of compact ligand coated QDs. (A) Gel electrophoretic separation of 550 nm emitting QDs cap-

exchanged with the indicated ligands. Gels were run on 1% agarose gel in 1ꢁ TBE buﬀer (pH 8.3) at ∼7 V/cm for ∼10 min. (B) Hydrodynamic size

distribution of 550 nm emitting QDs cap-exchanged with DHLA (10.8 ( 2.7 nm), CL1 (8.6 ( 1.8 nm), CL2 (9.3 ( 1.7 nm), CL3 (9.5 ( 2.1 nm), CL4

(9.8 ( 2.2 nm), and DHLA-PEG750-OCH₃(11.5 ( 2.5 nm) measured by dynamic light scattering. Data is plotted in arbitrary units of scattering

intensity. (C) PL images for a set of 0.5 μM QDs capped with the indicated compact ligands in diﬀerent buﬀers at pH 2ꢀ13. The 550 nm emitting CdSe-

ZnS QDs were used and excited with a UV lamp at 365 nm. Images were taken <20 min and ∼4 weeks after sample preparation. (D) PL images for a set

of 0.5 μM 550 nm emitting QDs coated with DHLA or the indicated compact ligands in 3 M NaCl solution. Images were taken 1 day after sample

preparation for DHLA and 90 days for CL1∼CL4.

QDs (provided by Life Technologies) maintained relatively high

QYs after cap exchange: QY was ∼78% in hexane and 50ꢀ80%

in water.

only minimal mobility in the electric field. This was expected as

CL3 terminates in hydroxyl groups and not a carboxyl group, and

the surface of a CL3-QD should, therefore, display an incon-

sequential net charge in buffer. In contrast, CL4-QD presented

gel mobility slightly higher than that of the QDs capped with

DHLA or CL1 and CL2. The higher mobility in this case is

directly attributable to the presence of more negative charges per

QD. Since CL4 terminates in two carboxyl groups, one can

expect a higher number of net negative charges present on the

QD surface in basic buffer. We also examined the zeta potentials

of this series of CL-QDs to confirm the gel findings since the

solution electrophoretic mobility of NPs should be proportional

to the zeta potential.³¹Values for the CL-QD samples measured

in 1ꢁ TBE (pH 8.3) are tabulated in Table 1. We note that the

Charge-Based Mobility, Zeta Potential, and Wettability.

Hydrophilic QDs cap-exchanged with the CL series were also

characterized by gel electrophoresis, see Figure 2A. CL-QD

samples were run together with DHLA-coated QDs in a 1%

agarose gel buffered with 1ꢁ TBE buffer (pH 8.3). CL1- and

CL2-QDs essentially showed gel mobilities similar to that of the

DHLA-QDs. These three ligands are all characterized by having

one carboxyl group per ligand which are expected to be exposed

on the QD surface, to be ionized at this pH (based on the pK_a

values in Table 1), and would then act as the major driving force

for electrophoretic mobility in basic buffer. CL3-QDs showed

9486

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

most negative value is seen for CL4 (ꢀ50.8 mV), followed by

CL1 (ꢀ39.0 mV) and CL2 (ꢀ19.8 mV), with CL3 (ꢀ4.9 mV)

representing the lowest value. The latter value is quite close to

that measured for similar QDs capped with DHLA-PEG mol-

ecules which also did not migrate in a gel and displayed an overall

neutral character.³¹The minimal mobility of the CL3-QDs along

with their slight net negative charge in basic buffer is ascribed to

the cumulative effect of all the hydroxyl moieties surrounding the

QD manifesting a very small pseudo-negative character. Lastly,

the higher value for CL4 in basic buffer is more than double that

of CL2 as expected given the presence of twice the number of

carboxyl groups.

ligands may enhance the hydrophilicity of QD surfaces and help

aqueous solubility.

Hydrodynamic Size. We next determined the hydrodynamic

diameters for the CL-QD series by using dynamic light scattering

(DLS) analysis. As shown in Figure 2B and Table 1, the

hydrodynamic diameters of the 550 nm emitting QDs coated

with CL1ꢀCL4 ranged from 8.6 to 9.8 nm. These values are

1.0ꢀ2.2 nm smaller than the 10.8 nm measured for the same

materials coated with DHLA and are 1.7ꢀ2.9 nm smaller than

those functionalized with DHLA-PEG750-OMe (11.5 nm). The

hard diameter of the CdSe-ZnS coreꢀshell QD sample used in

this analysis was directly measured by TEM to be 4.0 ( 0.29 nm

(Supporting Figure S3). The differences in sizes between hydro-

dynamic diameter and the hard coreꢀshell size measured in all

samples arise from hydrodynamic interactions of the ligand layer

on the QD surface.^31,46,47The contribution of the CLs to the

overall hydrodynamic radius is in the ∼2.3ꢀ3.4 nm range, which

is reasonably close to that estimated from just the simple fully

extended geometric length of each CL (1.6ꢀ2 nm), see Table 1.

The hydrodynamic volume of a colloidal NP is far more complex

than this simple extrapolation and is controlled by not only the

size and chemical nature of the surface coating materials, but also

by the size of the NP itself. Nevertheless, the key feature we note

from this analysis is that these new CLs provide the functiona-

lized QDs with hydrodynamic volumes comparable to or even

smaller than those of DHLA, which is one of the smallest ligands

currently utilized for preparing aqueous QDs that are stable over

extended time periods in buffer.¹⁶Moreover, the hydrodynamic

size is smaller than that encountered when utilizing far longer

PEGylated- or far larger polymer/dendrimer/block copolymer

ligands to mediate QD solubility.³¹

Long-Term pH and Salt Stability. We evaluated the long-term

colloidal stability of the CL-QDs across a wide pH range. The

representative images in Figure 2C show the luminescence

images of 550 nm emitting QDs coated with CL1ꢀCL4,

respectively, dispersed in buffer solutions increasing from pH 2

up to 13 at time points of 20 min and 4 weeks following sample

preparation. Images of the intervening time points can be found

in Supporting Figure S4. Strongly acidic conditions (pH 3 or

lower) rapidly led to a partial or complete fluorescence quench-

ing and sedimentation of the QDs, similar to the results observed

for QDs capped with quadradentate thiol ligands at the same

pH.⁴⁹We surmise that this strongly acidic environment results in

reactive ‘acidic’ etching of the nanocrystal or destabilizes the

ligand-QD bonds. The remaining QD dispersions were colloid-

ally stable in weakly acidic to strongly basic pH conditions (pH

4ꢀ5 to 12ꢀ13) without apparent fluorescence quenching for at

least a 2 month period. CL-QDs were also found to be stable in

DI water for at least 6 months. This excellent pH stability directly

contrasts with that observed for DHLA-coated QDs. While

colloidally stable in basic media, DHLA-coated QDs became

unstable even in weakly acidic media in a few days (Supporting

Figure S5). We note that the pH stability of the present CL-QDs

is comparable to that of QDs coated with DHLA-PEG ligands

where the lengthy and neutral ethylene oxide repeat units

mediate aqueous solubility.^24,25We attribute this to their zwit-

terionic nature of CL1, 2 and 4 which display pK_a’s in both the

acidic and basic regime and thus should be ionized across a broad

pH range. Interestingly, this also suggests that the two hydroxyl

groups on CL3 are also capable of at least contributing to

aqueous solubility when presented in conjunction with a tertiary

amine. The colloidal stability of the CL-QD series was also

We also measured the zeta potentials of the CL-QD series in

an acidic media of pH ∼4, see Table 1. Positive values of 10.8 and

10.3 mV were obtained for the same CL1- and CL2-QDs,

respectively, at this pH reﬂecting the balance between proton-

ation of the carboxyl group and ionization of the tertiary amines.

The highest value of 22.4 mV was observed for the CL3-QDs and

reﬂects the high pK_aof the amine in this construct along with the

presence of two hydroxyl groups and no counterbalancing

carboxyl groups. Interestingly, CL4-QDs displayed a value of

ꢀ19.0 mV. The negative value in this case reﬂects the dominating

inﬂuence of the two appended carboxyl groups functioning in

concert and the fact that the surrounding pH is slightly above

their pK_avalues of 3.29 and 3.93. This is further conﬁrmed by

noting that replacing just one of the carboxyl groups with a

hydroxyl group yields CL2 which displays a positive zeta

potential at this pH. As expected, control DHLA-QD samples

were not colloidally stable in this media and did not allow for

reliable measurement. More importantly, the change from a net

negative value in basic buﬀer for CL1-, CL2-, and CL3-QDs to

positive values in acidic environment along with the large change

in net value for CL4-QDs serves as a strong conﬁrmation that the

ligands do indeed display the desired zwitterionic properties and,

more importantly, can impart them to the QDs. We also note

that the microenvironment surrounding the colloidal QD sur-

faces, which accommodate a very high local concentration of

ligands, may have far more complex ionization kinetics than a

simple amino acid residue dissolved in a model buﬀer system.

Some insight into ligand dissolution and interactions within

aqueous environments was provided by probing the relative

wettability of each ligand. This was accomplished by drying

uniform concentrations and volumes of each ligand-coated QD

sample onto ﬂat mica surfaces and measuring the subsequent

contact angles of water drops using sessile drop goniometry, see

Supporting Information. As listed in Table 1, DHLA surfaces

produced the smallest contact angle at 5°, while PEGylated QD

surfaces produce the largest with a value of 20°. The smaller the

observed contact angle, the better the relative interaction of

water with a given surface coating. Data for the CL series are 10°,

8°, 6°, and 13° for CL1, CL2, CL3, and CL4 QD surfaces,

respectively. The smallest contact angle is obtained for the

DHLA-QD and this could be partially attributed to its original

presence in basic buﬀer which was necessary to maintain colloidal

stability; electrolytes would enhance the hydrophilicity of the

sample. All the other samples were in DI water. Importantly, the

contact angles of all the CL-QDs were smaller than that of the

PEG-coated QDs where far bulkier ethylene oxide repeat units

are required for aqueous stability. Indeed we have found that

∼12 ethylene oxide repeats (MW∼600) are usually required for

good solubility and wide pH stability.²³This result suggests that

introduction of multiple ionizable groups onto the surface

9487

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

Figure 3. His-tagged MBP-QD interactions and FRET. (A) Schematic of the FRET assay used to investigate direct assembly of His-tagged MBP labeled

with Texas Red onto QD surfaces. For electrophoresis, QDs were assembled with His-tagged MBP-Texas Red only, while for FRET, a mixture of

His-tagged MBP and His-tagged MBP-Texas Red was used. (B) Gel electrophoresis separation of 550 nm emitting CL1-, CL2-, and CL4-QDs

self-assembled with “n” equivalents of His-tagged MBP-Texas Red. Lane labeled ‘0’ contains only QD while that labeled ‘MBP-TR’ contains only

His-MBP-Texas Red. Bands corresponding to free MBP-TR are indicated with red arrows. Gels were run on 1.5% agarose gel in 1ꢁ TBE buﬀer (pH 8.3)

at ∼10 V/cm for ∼10 min and imaged with 525 nm band-pass or 590 nm long-pass ﬁlters to highlight the QD and Texas Red ﬂuorescence components,

respectively. (C) PL spectral changes from 550 nm emitting DHLA-QD and CL-QDs titrated with a mixture of “n” equivalents of 5ꢁHis-tagged

MBP-Texas Red and “10-n” equivalents of 5ꢁHis-tagged MBP (unlabeled). The total number of His-tagged proteins per QD was ﬁxed as 10 while the

discrete ratio of dye-labeled MBP per QD was varied systematically. (D) FRET eﬃciencies from DHLA-, CL1-, CL2-, and CL4-QDs in (C) as a function

of MBP-TR/QD ratio (n).

examined under high salt concentration (Figure 2D). Each QD

was dispersed in 3 M NaCl solutions (pH ∼6) and monitored

over time. Again DHLA-coated QDs precipitated within a day at

this salt concentration while all the CL-QDs appeared stable and

well dispersed for at least 3 months. The generally accepted

relevant physiological NaCl concentration is ∼150 mM,⁵⁰and

biological environments are also known to be rich in a variety of

other strong electrolytes. This strongly suggests that CL-coated

QDs may be quite tolerant of similar ion-rich biological environ-

ments for long time periods.

€

Assembly of His-Tagged Proteins and Forster Resonance

Energy Transfer. Recombinantly modifying proteins to express

9488

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

oligohistidine tags (His_n) was originally implemented to facilitate

their large-scale, efficient purification with metal-affinity

chromatography.^51,52In recent years, we, along with a growing

number of other groups, have utilized these same metal-affinity

interactions to conjugate CdSe-ZnS QDs to an array of oligo-

histidine-appended proteins, peptides, and even appropriately

modified DNA to assemble a variety of QD bioconjugates for

sensing, imaging, and cellular probing applications.^{17ꢀ21,48,53ꢀ57}

His_n-interactions arise from collective metal-affinity coordina-

tion between the pendant imidazolium side-chains of histidine

residues and the Zn-rich surface of QDs.¹⁸Inherent benefits

provided by this simple but effective QD-bioconjugation ap-

proach include: (i) ease of use as it is based completely on self-

assembly and only requires mixing of the two reagents; (ii) rapid

self-assembly with an intrinsically high affinity (K_d∼ 1 nM); (iii)

control over the valence of His_n-biomolecules attached per QD

through the molar ratios used; (iv) direct use of His_n-QD

bioconjugates in subsequent assays without requiring further

purification; (v) control over biomolecular orientation on the

QD; (vi) application to QDs coated with a number of different

surfaces/ligands; and (vii) due to their small size, His_n-modifica-

tions have only minimal effects on the tertiary structure and

biological activity of the proteins and peptides to which they are

appended.

Given the growing popularity of His_n-QD bioconjugation, it

was important to evaluate whether similar assembly could be

achieved with the CL-QD series. Previous studies revealed that,

although smaller peptides can interact with many types of QDs

via His_n-interaction, conjugation to proteins in the same manner

required the use of QDs capped with short DHLA surface

ligands.^18,20,21,23When attempting to use PEGylated QDs as

His_n-substrates, the larger, globular nature of proteins combined

with the steric hindrance of the long PEG chains on the QD

surface prevents the His-tag from interacting directly with the

metal-rich QD surface. However, as DHLA-capped QDs are

stable only in basic buﬀers, this has severely limited eﬀective use

of DHLA-QD/His_nꢀprotein conjugates. Clearly, signiﬁcant

minimization of the steric hindrance caused by the surface

ligands on the QDs is a prerequisite to ensure a stable conjuga-

tion between QDs and His-tagged proteins. The CL series are

expected to increase the chance that His_n-tags on proteins reach

the QD surface and form stable binding interactions with it.

The capability of direct His_n-tagged protein conjugation onto

CL-QD surfaces was examined in two ways: agarose gel electro-

phoretic separation and FRET analysis. As a test protein, we

utilize the Escherichia coli periplasmic maltose binding protein

(MBP). This medium sized protein (∼44 kDa) has dimensions

of 30 ꢁ 40 ꢁ 65 Å with a volume of ∼78 ꢁ 10³Å³and has found

strong utility as a test platform in biosensing.⁵⁸MBP displaying a

C-terminal pentahistidine sequence was site-speciﬁcally labeled

on residue 95C with Texas Red maleimide dye (5xHis MBP-TR)

as described in the Supporting Information. For gel electro-

phoretic analysis, dye-labeled MBP alone was self-assembled to

550 nm emitting QDs capped with CL1, CL2, and CL4 at ratios

of 0, 2, and 5 per QD (see Figure 3A for schematic). As CL3 is

neutral and is not expected to show sizable gel mobility, it was not

tested in this conﬁguration. QD-MBP-TR conjugates were then

separated side-by-side with free MBP-TR in 1.5% agarose gels

where use of diﬀerent ﬁlters allowed separate visualization of the

QD and sensitized Texas Red emission bands on the gel using

UV excitation.⁵⁹Figure 3B shows that as the ratio of 5ꢁHis

MBP-TR was increased, discrete slower moving bands appeared

along with the higher mobility unconjugated QD bands for each

of the CL-QD samples when observed with the 525 nm band-

pass ﬁlter. The unconjugated QD front-moving band also visibly

decreases in intensity with increasing MBP-TR. The Texas Red

bands visualized with the 590 nm long-pass ﬁlter coincide with

those of the slower moving QD/MBP-TR conjugates. Changes

in mobility of QD/MBP-TR conjugates are due to increases in

relative mass/charge arising from assembly of MBP-TR onto the

QD surface. In comparison, the slowest gel mobility was ob-

served for unconjugated MBP-TR which also appears quite faint

due to a low direct and unsensitized excitation in the UV.

Colocalization of the QD and Texas Red emission, along with

sensitization of the latter for co-assembled samples, conﬁrm that

the QDs and MBP-TR are indeed conjugated and migrating

together.

Since the PL of 550 nm emitting QD and the absorption of

Texas Red have a sizable spectral overlap (I of 2.98 ꢁ 10^ꢀ13

M

^ꢀ1cm³and R₀of 45 Å, see Supporting Figure S6), formation of

the QD-MBP-TR conjugates was expected to facilitate an

eﬃcient level of energy transfer from the QD donor to the

proximal Texas Red dye acceptor. For evaluating CL-QD

intraconjugate FRET, we utilize a slight modiﬁcation during

the self-assembly process. It has been previously demonstrated

that conjugation of His-tagged biomolecules onto QD surfaces

commonly increases the ﬂuorescence intensity or QY of the

QDs; this enhancement is ascribed to protein passivation of

surface trap states on the QD surface.^60,61Assembling unlabeled

5ꢁHis MBP onto the current CL-QD series resulted in similar

changes further conﬁrming His-tagged MBP binding to CL-

functionalized nanocrystals, see Supporting Figure S7 for a

representative example. To eﬀectively address PL enhancement

caused by His-tagged protein binding onto QDs during the

subsequent FRET analysis, the total amount of His-tag MBPs

assembled was ﬁxed at 10 equiv per QD, while the discrete ratio

of Texas Red-labeled (n) and unlabeled His-MBP (10-n) was

systematically varied (n = 0ꢀ10) as shown in Figure 3A. This

approach has been utilized previously,^60,61and mitigates issues of

changes in PL enhancement and the need for a more complex

analysis to account for self-assembly heterogeneity at low

ratios.⁶²

Figure 3C shows a series of representative ﬂuorescence spectra

collected from dispersions of the same CL-QD series as above

incubated with MBP-TR at increasing ratios and excited at

400 nm where the absorption of Texas Red is minimized.

DHLA-coated QDs were also measured as a positive control

sample. As MBP-TR ratios increased, each QD/MBP-TR con-

jugate showed a decrease in QD PL concomitant with the

appearance of a new ﬂuorescence band at ∼615 nm arising from

Texas Red sensitization. The trends in these data sets closely

mirrors that previously collected from a series of DHLA-capped

QDs assembled with MBP labeled with Cy3-acceptor in the same

95C position.⁶¹Increases in FRET eﬃciency within each con-

jugate extracted from the decrease of QD donor PL versus

assembly valence are plotted in Figure 3D. Analysis of this data

using F€orster formalism and eqs 1ꢀ3 in the Supporting Informa-

tion derived QD-donor to dye-acceptor separation distances r of

65 ( 7 Å (DHLA-QDs), 65 ( 5 Å (CL1-QDs), 73 ( 6 Å (CL2-

QDs), 64 ( 3 Å (CL4-QDs). The ﬂuorescence enhancement of

the QDs following assembly with 10 equiv of His-tagged MBPs

(Figure S7) was measured, and the increased QYs of the QD-

MBP conjugates were accounted for within each sample conﬁg-

uration in this analysis. Interestingly, the averaged r values for the

9489

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

Figure 4. EDC coupling chemistry. (A) Schematic representation of EDC coupling of CL-QDs with Lissamine rhodamine B ethylenediamine.

Only one of the two available carboxyl groups on CL4 is shown modiﬁed. Respective absorption and PL spectra of the ﬁrst eluting band following gel

ﬁltration of reaction mixtures of 550 nm emitting QDs and Lissamine rhodamine B ethylenediamine for (B and C) CL1-QDs; (D and E) CL2-QDs; and

(F and G) CL4-QDs versus QD samples in the control reactions lacking EDC. The 15-fold excess dyes per QD were used for each reaction. In panel B,

the native absorption spectrum of Lissamine rhodamine B ethylenediamine is also shown as blue line. PL spectra were measured using either 400 or

440 nm excitation.

CL-QD series are similar to those determined with the DHLA-

QDs, suggesting that CL series are still small enough to provide

close proximity between the QD donor and acceptors; this was

also suggested by the DLS data. Slightly higher FRET eﬃciency

of DHLA-QDs compared with the CL-QD series is due to the

higher ﬂuorescence enhancement of DHLA-QD observed when

assembled with His-tagged MBP (data not shown). We also

surmise that the actual number of His-tagged MBP bound on

9490

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

CL-QDs could be lower than that of DHLA-QDs due to the

slightly bulkier structures of the CL series compared with DHLA.

Diﬀerent ligand structures and packing may also sterically alter

MBP orientation, which may concomitantly aﬀect location of

the acceptor dye and FRET eﬃciency. We note that His-tagged

TR-labeled peptides could also self-assemble on the CL-QD

series and engage in eﬃcient FRET (see representative example

in Supporting Figure S8). This was not surprising given that

His-tagged peptides are usually >10-fold smaller than His-tagged

proteins and can even access the surface of PEGylated QDs.

Overall, it is clear that the present CL series are small enough

to allow direct His-tag mediated conjugation of proteins onto

the QD surface. The ability to directly conjugate QDs with

His-tagged proteins and peptides in a simple manner with good

control over the resulting hybrid architecture will contribute to

wider utility and allow for more sophisticated QD-bioconjugate

design.^14,15

PL spectra of the QD-dye conjugates (Figure 4C,E,G) showed a

signiﬁcant decrease of the original QD ﬂuorescence along with

increase of the dye ﬂuorescence. Since the conjugates were

excited at either 400 or 440 nm where the absorption of the

rhodamine dye is negligible, ﬂuorescence enhancement arises

mostly from FRET sensitization. The FRET eﬃciency of the

QD-dye conjugates was ∼84%, ∼58%, and ∼97% for the

respective CL1, CL2, and CL4 samples as calculated from QD

donor PL changes before and after dye conjugation. These

spectral features indicate that (i) EDC coupling worked to

covalently couple the carboxyl group of the compact ligands on

the QD surface to the amine group of the rhodamine dye, and (ii)

the standard reaction condition used here did not cause any

signiﬁcant nonspeciﬁc adsorption between the QDs and dye. We

also did not observe any visible aggregation or precipitation

during the reactions and the following puriﬁcation step. Though

the coupling conditions were not fully optimized, the present

data set clearly demonstrates that the carboxylic acid groups of

CLs are available for further chemical conjugation. This result is

important in light of several reports indicating that carboxyl

groups may have a high propensity to directly interact with QD

surfaces.^10,63

Cellular Delivery of Compact Ligand-Coated Quantum

Dots. One of the fastest growing areas of QD utility continues

to be that of cellular labeling and imaging. Among the continuing

issues related to this are access to facile methods for delivering

the QDs to the cells and the effects on subsequent cellular

viability.^5,64We examined delivery of the CL-QD series to cells

and monitored their subsequent intracellular stability using both

active microinjection and the more passive technique of peptide-

facilitated uptake.

Microinjection. Adherent COS-1 cells grown in coverslip-

bottom plastic dishes were microinjected with 2 μM of each

CL-QD diluted into 1ꢁ PBS. The cells were then repeatedly

imaged by epifluorescence microscopy over a 6 h period to follow

the intracellular PL distribution of the QDs. This technique

allows us to directly access the cytosol and probe the in vivo

colloidal stability of the CL-QDs. The 550 nm emitting QDs

were used with CL1 and CL2, while 625 nm emitting ITK QDs

were utilized with CL4. As can be seen in the representative time-

lapse image set shown in Figure 5A, PL from the QDs maintained

a strong level of signal inside the cells and an almost uniform

staining pattern. Further, in almost all cells examined, the QDs

were located throughout the cytosol but were clearly excluded

from the nuclei. These results clearly indicate that the CL-QDs

were well dispersed in the cytosol without aggregation. Although

addressing of individual cells by this technique does not allow for

full assessments of toxicity/viability, cells appeared to be intact

and visibly unperturbed for at least 6 h following injection. This

suggests the availability of an experimental window for more

complicated intracellular sensing experiments using this delivery

mechanism with probes assembled around these types of materi-

als. The excellent colloidal stability of these CL-QDs in the

cellular environment is in marked contrast to DHLA-QDs, which

exhibited punctate fluorescence under similar conditions, sug-

gesting rapid aggregate formation and precipitation.²³Repeated

attempts to microinject CL3-QDs failed with the nanocrystal

samples appearing to precipitate at the cell-microinjector tip

interface. The reasons behind this are not readily apparent given

the good pH and salt-stability observed above for QDs capped

with this ligand (vide supra).

Chemical Conjugation to Compact Ligand Functionalized

Quantum Dots. The ubiquitous presence of carboxyl and amine

groups on proteins and peptides along with their easy introduc-

tion onto QDs and other NPs has made carbodiimide (e.g.,

EDC) chemistry perhaps the most popular method for attaching

biological moieties to QDs.^4,9,15,24,41To verify that the terminal

carboxyl groups of CL1, CL2, and CL4, when present on the QD

surface, are available not only for aqueous solubility but also for

chemical conjugation with other molecules, aqueous EDC

coupling reactions were carried out with an amine-terminated

dye. Since the 550 nm emitting QDs used here and Lissamine

rhodamine B ethylenediamine dye (Life Technologies) have

significant spectral overlap (I of 5.3 ꢁ 10^ꢀ13M^ꢀ1cm³and R₀of

49 Å, see Supporting Figure S9), EDC coupling between the

QDs and this dye is expected to place them close enough to allow

efficient FRET from the QD donor to the dye functioning as

acceptor. QDs coated with either CL1, CL2, or CL4 were mixed

with 15-fold excess dye along with EDC and sulfo-NHS in

10 mM phosphate buffer pH 7.0 or PBS buffer as described in

the Supporting Information (Figure 4A). The reaction mixture

was stirred for ∼2 h at room temperature and directly passed

through PD-10 gel filtration columns (GE Healthcare, Inc.) to

remove the reagents and unbound dyes. The same reactions

without EDC were carried out as controls to examine potential

nonspecific adsorption of the dyes onto the QD surfaces. The

first fluorescent fraction eluting from the columns was collected

and used in the subsequent spectroscopic measurements.

Figure 4BꢀG shows the absorption and PL spectra of this first

eluent from the reaction mixtures superimposed over that

collected from control reactions without EDC. For comparison

purposes, Figure 4B also plots the absorption of rhodamine dye.

The absorption spectrum collected from each EDC reaction

(Figure 4B,D,F) is characterized by the presence of a new peak

around 570 nm originating from the rhodamine dye together

with the QDs lowest absorption band around 530 nm. Absorp-

tion spectra of the same fraction collected from control reactions

without EDC were identical to those of the original QD sample.

The average number of conjugated rhodamine dyes per QD

was estimated to be ∼2.0, ∼0.7, and ∼1.9 for the CL1, CL2, and

CL4 samples, respectively, by assuming that the absorption

spectrum of the QD-dye conjugate is the sum of the individual

spectra of QD and dye. The coupling eﬃciency of CL2-QD was

smaller than that of CL1-QD and CL4-QD and we found that

using a higher dye/QD ratio for the EDC coupling increased the

number of the conjugated dye per CL2-QD (data not shown).

9491

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

Figure 5. Cellular delivery. (A) Representative epiﬂuorescent images of COS-1 cells at diﬀerent time periods following microinjection with 550 nm

emitting CL1-QD, 550 nm emitting CL2-QD, and 625 nm emitting CL4-QD. The CL-QDs were injected directly into the cytosol of the cells.

(B) Images corresponding to the diﬀerential interference contrast (DIC), DAPI, 550 nm emitting QD, and AlexaFluor647-transferrin ﬂuorescence for

COS-1 cells exposed to CL3-QD-CPP assemblies. (C) Merged images for cells exposed to assemblies of the CL-QD series with CPP. COS-1 cells were

co-incubated with 550 nm emitting CL-QD-CPP assemblies (CL-QD/CPP ratio 1:25) at QD concentrations of 100 nM and 30 μg/mL AlexaFluor647-

transferrin for 1ꢀ2 h. The nuclei were counterstained with DAPI following ﬁxation.

9492

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

Cellular Uptake of Compact Ligand Coated Quantum Dots

Self-Assembled with Cell Penetrating Peptide. As an alternate to

the microinjection format described above, we also assessed

whether the CL-QD series could undergo peptide-facilitated

delivery to the same cells. The cell penetrating peptide (CPP =

R₉GGLA(Aib)SGWKH₆) used in this study is composed of two

functional sequences separated by a GGLA(Aib)SGWK linker.

The C-terminal oligohistidine sequence allows the peptide to

self-assemble onto the QD surface via metal-affinity interactions,

while the N-terminal oligoarginine sequence allows specific QD

delivery across the cellular membrane.^57,64,65Use of canonical

polyarginine peptidyl motifs, originally derived from the HIV

transcriptional activator or Tat protein, has facilitated endocytic

uptake of a number of different cargos including QDs and other

NPs into cells. Mechanistically, it is believed that their inherently

strongly positive charge interacts with the negatively charged

heparan sulfate proteoglycan receptors present on the cellular

membrane prior to endocytosis and membrane turnover.^64ꢀ66

COS-1 cells were incubated for 1ꢀ2 h with either CL-QD-

CPP assembled at 25 or 50 equiv of CPP per QD or control

nonconjugated CL-QDs. The representative epiﬂuorescent

images in Figure 5B,C show that self-assembly of CPP across

the entire CL-QD series results in eﬃcient cellular uptake of the

conjugates. The complete image set is shown in Supporting

Figure S10. Punctate QD ﬂuorescence was observed throughout

the cytoplasm as well as in close proximity to the cell nuclei

counterstained with DAPI. However, no nuclear staining by the

QDs was apparent. Furthermore, we observed no detectable

ﬂuorescence signal remaining on the plasma membrane, indicat-

ing a negligible degree of nonspeciﬁc binding of CL-QD-CPP

assemblies to the cells after washing. CL-QDs without CPP also

showed only negligible intracellular ﬂuorescence and nonspeciﬁc

binding on the plasma membrane after washing (Supporting

Figure S11). In contrast to the microinjection results, CL3-QD-

CPP assemblies demonstrated very eﬃcient delivery to the cells.

The punctate nature of endosomal QD staining does not,

however, allow us to determine whether they remain dispersed

or aggregated in this acidic cellular compartment. These results

conﬁrm that self-assembly of CPP with CL-QDs allowed for

speciﬁc cellular internalization. The fate of the internalized

CL-QD-CPP assemblies were conﬁrmed by simultaneous

delivery of the endosomal marker AlexaFluor 647-labeled trans-

ferrin (AF647-transferrin), which is commonly used to monitor

the endocytic pathway.⁵⁷The merged ﬂuorescent images in

Figure 5C showed that a signiﬁcant portion of the internalized

QD-CPP assemblies were colocalized with AF647-transferrin

conﬁrming that the CL-QD-CPP assemblies were mostly dis-

tributed within endosomes and intracellular delivery of the

assemblies was driven by endocytotic uptake. We have observed

similar CPP-mediated endocytic internalization for both DHLA

and PEGylated QDs.^57,67The primarily perinuclear localization

of the CL-QD-CPP assemblies is also in keeping with previous

results and is believed to occur as the vesicles transition to late

endosomes within the endolysosomal system.⁵⁷

Figure 6. Cytotoxicity assay results demonstrating the eﬀects of CPP,

CL-QDs, and CL-QD-CPP (QD/CPP ratio = 1:25) assemblies on

cellular proliferation of COS-1 cells. COS-1 cells were incubated with

the materials for 2 h, washed, and subsequently cultured for 72 h prior to

the viability assay. Each data point represents the mean ( standard

deviation of triplicate measurements.

6.2 μM), unconjugated CL-QDs (tested up to 200 nM), or

CL-QD-CPP assemblies (CPP/CL-QD ratio of 25:1, CL-QD

tested up to 200 nM) relative to the untreated control were

observed (Figure 6). Interestingly, CL4-QD-CPP assemblies

appear to have a mildly stimulative effect on the cells in this

assay format, similar to the results reported for exposing mouse

immune cells to mercaptopropanoic acid-capped QDs.⁶⁸We

note that this is the same exposure time utilized to achieve the

CPP-mediated cellular delivery above and that this regime has

also been shown to be nontoxic for delivery of other types of

DHLA and PEGylated QDs to various cell lines.^57,67

Compact Ligand-Functionalization of Gold Nanoparti-

cles. To demonstrate that the CL series could be useful for

solubilizing other NP materials beyond QDs, we also tested their

applicability with AuNPs. While single thiol-appended zwitter-

ionic ligands including cysteine and glutathione have been widely

applied to AuNPs due to the stable AuꢀS interactions,^69ꢀ72use

of these dithiol ligands with zwitterionic character can further

enhance the colloidal stability and thus has some benefits for

long-term use in biological studies under harsh conditions

compared with single thiol analogues.^28,29In contrast to the

coating procedure used with the QDs, surface functionalization

here does not require a reduction step to open the dithiolane

ring, rather the thioctic amide derivatives of CLs (3, 6, 7, 10)

are added in excess to citrate-stabilized AuNPs and the

inherent goldꢀsulfur interactions open the dithiols during cap-

exchange.^28,29,73Citrate-stabilized 5.8 nm AuNPs were cap-

exchanged with the CL series along with DHLA as described

in the Materials and Methods. The CL-AuNPs were then

subjected to a battery of the same characterization tests as the

CL-QD series. Gel electrophoresis was used to verify that the

ligands would impart similar surface charge/mobility character-

istics to the AuNPs. As can be seen in Figure 7A, the CL-AuNPs

have an almost identical gel migration pattern as CL-QDs when

compared to DHLA-AuNPs (compare Figure 7A to Figure 2A)

suggesting that almost the same relative surface-charge proper-

ties are imparted to these materials. Results of the DLS analysis

can be found in Supporting Figure S12. The hydrodynamic

diameters of the 5.8 nm AuNPs coated with CL1ꢀCL4 ranged

from 12.1 to 13.9 nm. These values are comparable within error

to the 13.2 nm measured for the same AuNPs coated with DHLA

and are again much smaller than those functionalized with

Cytotoxicity. Lastly, we examined the potential toxicity of

delivering CL-QDs to cells using a colorimetric tetrazolium-

based cell proliferation assay which monitors the inhibition of

cellular proliferation.^57,67Equivalent concentration ranges of

CL-QD-CPP assemblies, unconjugated CL-QDs, and free CPP

were exposed to the COS-1 cells using the same conditions as

above. Following washing steps and a subsequent 72 h culture,

no detrimental changes in proliferation for the CPP (tested up to

9493

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

overview, these results clearly conﬁrm that CL properties can be

extended to other NPs beyond QDs.

’ CONCLUSIONS

The concept of designing QD ligands to be both compact

and simultaneously zwitterionic has been attempted before.

Bawendi’s group utilized the amino acid cysteine with CdSe-

CdZnS coreꢀshell QDs which allowed it to function as one of

the smallest QD ligands known after mercaptoacetic acid, along

with providing a zwitterionic character.⁴⁶Use of this natural

residue as a ligand, however, resulted in materials that were stable

for ∼24 h without adding the additional presence of a strong

reducing agent; a result attributed to ligand dissociation and

subsequent oxidation which formed cystine dimers that did

not reaccess the nanocrystal surface. The cysteine analogue

penicillamine⁷⁴and the tripeptide glutathione^75,76have also been

used as ligands with the intent of providing similar zwitterionic

properties to QDs.¹⁰Muro et al. reported a QD ligand that is

somewhat closer in design to our CL series.⁴⁷Here, DHLA was

used in a similar manner to provide high aﬃnity with the QD

surface while the zwitterionic nature wasprovided byan appended

sulfobetaine group. Kim recently synthesized a ligand analogous

in concept to that of Muro where DHLA was also linked to a

sulfobetaine group.⁷⁷QDs capped with the latter ligands and

control ligands were utilized for a series of cellular studies.⁷⁷It is

important to note that the quaternary amine and sulfonate groups

in both Kim’s and Muro’s design precluded further chemical

conjugation which could only be provided to those QDs by

creating a mixed nanoparticle surface displaying several diﬀerent

functionalized ligands.

Figure 7. Compact ligand coated gold nanoparticles. (A) Gel electro-

phoresis separation of 5.8 nm AuNPs cap-exchanged with the indicated

ligands. Gels were run on 1.5% agarose gel in 1ꢁ TBE buﬀer (pH 8.3) at

∼10 V/cm for ∼10 min. (B) Images of 200 nM DHLA and CL-AuNPs

suspended in 2 M NaCl. Images were taken 1 day after the sample

preparation for the DHLA sample and 7 days for the CL1∼CL4

samples. (C) Gel electrophoresis separation of CL4-AuNPs EDC-

coupled to a Cy3-labeled peptide displaying a single, internal lysine

amine. The numbers indicate the equivalents of labeled peptide present

in each reaction run in that lane. ‘-EDC’ correspond to control AuNP

samples incubated with 50-fold excess of peptide in the absence of EDC.

Gels were imaged with both white light and UV illumination. Red arrows

(from top to bottom) indicate the presence of unreacted peptide,

AuNP/peptide conjugates and AuNPs.

We report here the design and synthesis of a series of new

DHLA-based compact ligands which allow preparation of robust

and biocompatible QDs for biosensing and cellular imaging

applications. The structures of these new DHLA-based ligands

are simple and they are easily synthesized in large quantities from

available materials in four-steps that include ring-opening. In-

herent beneﬁts that these ligands provide for biological applica-

tions include: (i) high-aﬃnity interactions with the QD surface;

(ii) small hydrodynamic size of QDs, comparable to that of

DHLA-QDs; (iii) enhanced colloidal stability under a wide pH

range and high salt concentration; (iv) direct conjugation of His-

tagged proteins and peptides onto QD surfaces; (v) availability of

carboxyl groups for further chemical modiﬁcation; (vi) excellent

colloidal stability and utility in cellular environments; (vii) low

toxicity in cellular environments; and (viii) providing similar

properties to other NP materials. Cumulative access to all these

properties in a single ligand-coated QD has been extremely hard

to achieve before. While some dendrimers do provide long-term

stability and access to surfaces displaying multiple functional

groups, this is again at the expense of signiﬁcant size.^10,11

Maintaining a small hydrodynamic size is important for any

potential in vivo utility. Equally important, imparting a zwitter-

ionic nature to QD ligands has been repeatedly shown to prevent

adsorption of serum proteins with concomitant increases in

hydrodynamic diameter, a criteria that is deemed critical for

renal clearance.^{46,47,77ꢀ79}

DHLA-PEG750-COOH (17.3 nm). These trends are quite

similar to the results collected from the QDs above. We found

that the CLs also provided the AuNPs with a similar colloidal

stability over time, see Figure 7B.

To verify that the carboxyl groups on the CL series would also

be available for subsequent chemical modiﬁcation, we performed

EDC coupling reactions with a Cy3 dye-labeled peptide that had

a unique lysine-amine within its sequence, see Supporting

Information. Figure 7C shows representative data where CL4-

AuNPs were reacted with increasing concentrations of excess

dye-labeled peptide and then separated on a 1.5% agarose gel. As

can be seen under white light illumination, the conjugated

AuNPs undergo a small decrease in migration relative to control

samples with no EDC present plus 50-fold excess of peptide, see

the corresponding lanes indicated with ‘-EDC’. Mobility shifts

directly track the increased ratios of labeled peptide present in

the reactions along with a concomitant decrease in the intensity

of the fast-moving unconjugated AuNP band. Under UV illumi-

nation, unconjugated peptide is visualized migrating in the

opposite direction coinciding with its very slight positive charge

at the buﬀer pH of 8.3. Fluorescence is also observed between the

fast-moving AuNP band and the loading wells of the gel, and the

relative emission here increases concomitant with the ratio of

peptide present in the reaction. This ﬂuorescence arises from the

peptide that is EDC-conjugated to the AuNPs and its largely

diﬀuse migration corresponds to the heterogeneous valence of

labeled product as expected. Very little emission is seen in lanes

corresponding to the 12.5 ratio of excess peptide used; this was

also expected given both a low predicted peptide attachment and

strong dye quenching from the gold.⁷³Although only an initial

Overall, we note a high degree of functional equivalence and

zwitterionic character among CL1, CL2, and CL4 suggesting

that each is capable of imparting nearly the same properties to

QDs. CL3 served more as a control to verify that the simulta-

neous presence of both a carboxyl and tertiary amine are required

9494

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

for optimal zwitterionic properties. However, its facile synthesis,

small size, and relative in vitro robustness suggest that it is still

useful for selected applications. The use of the high-aﬃnity

dithiol attachment domain should also allow these ligands to

be applied to a number of other NP materials and surfaces

including those derived from (or having shells of) Ag, Pd, and

Pt.⁷In summary, the compact ligand series characterized here

have a strong potential to expand QD and other NP capabilities

in many biological applications.

(16) Mattoussi, H.; Mauro, J. M.; Goldman, E. R.; Anderson, G. P.;

Sundar, V. C.; Mikulec, F. V.; Bawendi, M. G. J. Am. Chem. Soc. 2000,

122, 12142–12150.

(17) Medintz, I. L.; Berti, L.; Pons, T.; Grimes, A. F.; English, D. S.;

Alessandrini, A.; Facci, P.; Mattoussi, H. Nano Lett. 2007, 7, 1741–1748.

(18) Sapsford, K. E.; Pons, T.; Medintz, I. L.; Higashiya, S.; Brunel,

F. M.; Dawson, P. E.; Mattoussi, H. J. Phys. Chem. C 2007,

111, 11528–11538.

(19) Liu, W.; Howarth, M.; Greytak, A. B.; Zheng, Y.; Nocera, D. G.;

Ting, A. Y.; Bawendi, M. G. J. Am. Chem. Soc. 2008, 130, 1274–1284.

(20) Lu, H.; Sch€ops, O.; Woggon, U.; Niemeyer, C. M. J. Am. Chem.

Soc. 2008, 130, 4815–4827.

(21) Dennis, A. M.; Sotto, D. C.; Mei, B. C.; Medintz, I. L.;

Mattoussi, H.; Bao, G. Bioconjugate Chem. 2010, 21, 1160–1170.

(22) Lee, J.; Kim, J.; Park, E.; Jo, S.; Song, R. Phys. Chem. Chem. Phys.

2008, 10, 1739–1742.

(23) Uyeda, H. T.; Medintz, I. L.; Jaiswal, J. K.; Simon, S. M.;

Mattoussi, H. J. Am. Chem. Soc. 2005, 127, 3870–3878.

(24) Susumu, K.; Uyeda, H. T.; Medintz, I. L.; Pons, T.; Delehanty,

J. B.; Mattoussi, H. J. Am. Chem. Soc. 2007, 129, 13987–13996.

(25) Mei, B. C.; Susumu, K.; Medintz, I. L.; Delehanty, J. B.;

Mountziaris, T. J.; Mattoussi, H. J. Mater. Chem. 2008, 18, 4949–4958.

(26) Mei, B. C.; Susumu, K.; Medintz, I. L.; Mattoussi, H. Nat.

Protoc. 2009, 4, 412–423.

’ ASSOCIATED CONTENT

S

Supporting Information. A schematic representation of

b

cap exchange, a series of characterization data of the materials

used, gel images of QD/His-tagged peptides conjugates, images

for the cellular uptake study, DLS data of AuNPs, and detailed

descriptions of the experimental conditions. This material is

available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION

Corresponding Author

susumu@ccs.nrl.navy.mil; Igor.Medintz@nrl.navy.mil

(27) Susumu, K.;Mei, B. C.;Mattoussi, H. Nat. Protoc. 2009, 4, 424–436.

(28) Mei, B. C.; Oh, E.; Susumu, K.; Farrell, D.; Mountziaris, T. J.;

Mattoussi, H. Langmuir 2009, 25, 10604–10611.

(29) Oh, E.; Susumu, K.; Goswami, R.; Mattoussi, H. Langmuir

2010, 26, 7604–7613.

’ ACKNOWLEDGMENT

The authors acknowledge NRL, the NRL NSI, DARPA, ONR,

and ARO/DTRA for ﬁnancial support. J.B.B.-C. acknowledges a

Marie Curie IOF. W.J.H. gratefully acknowledges the NRC RAP

postdoctoral fellowship program. W.R.A. is grateful to the

Natural Sciences and Engineering Research Council of Canada

(NSERC) for a postdoctoral fellowship. We thank Dr. M. Mauro

at Life Technologies for providing a sample of custom Qdot 625

ITK organic QDs and S. MacDonald at ACD for assistance with

pK_adeterminations.

(30) Oh, E.; Susumu, K.; Blanco-Canosa, J. B.; Medintz, I. L.;

Dawson, P. E.; Mattoussi, H. Small 2010, 6, 1273–1278.

(31) Pons, T.; Uyeda, H. T.; Medintz, I. L.; Mattoussi, H. J. Phys.

Chem. B 2006, 110, 20308–20316.

(32) Clapp, A. R.; Goldman, E. R.; Mattoussi, H. Nat. Protoc. 2006,

1, 1258–1266.

(33) Snee, P. T.; Chan, Y.; Nocera, D. G.; Bawendi, M. G. Adv.

Mater. 2005, 17, 1131–1136.

(34) Li, J. J.; Wang, Y. A.; Guo, W.; Keay, J. C.; Mishima, T. D.;

Johnson, M. B.; Peng, X. J. Am. Chem. Soc. 2003, 125, 12567–12575.

(35) Blackman, B.; Battaglia, D. M.; Mishima, T. D.; Johnson, M. B.;

Peng, X. Chem. Mater. 2007, 19, 3815–3821.

’ REFERENCES

(1) Gill, R.; Zayats, M.; Willner, I. Angew. Chem., Int. Ed. 2008,

47, 7602–7625.

(36) Blackman, B.; Battaglia, D. M.; Peng, X. Chem. Mater. 2008,

20, 4847–4853.

(37) Lehninger, A. L.; Nelson, D. L.; Cox, M. M. Principles of

Biochemistry, 2nd ed.; Worth Publishers: New York, 1993.

(38) Berg, J.; Tymoczko, J.; Stryer, L. Biochemistry, 5th ed.; W. H.

Freeman and Company: New York, 2002.

(39) Zhang, S.; Baker, J.; Pulay, P. J. Phys. Chem. A 2010,

114, 432–442.

(40) Hall, H. K., Jr. J. Am. Chem. Soc. 1957, 79, 5441–5444.

(41) Hermanson, G. T. Bioconjugate Techniques, 2nd ed.; Academic

Press, Inc.: San Diego, CA, 2008.

(2) Boisselier, E.; Astruc, D. Chem. Soc. Rev. 2009, 38, 1759–1782.

(3) Yong, K.-T.; Roy, I.; Swihart, M. T.; Prasad, P. N. J. Mater. Chem.

2009, 19, 4655–4672.

(4) Algar, W. R.; Tavares, A. J.; Krull, U. J. Anal. Chim. Acta 2010,

673, 1–25.

(5) Biju, V.; Itoh, T.; Ishikawa, M. Chem. Soc. Rev. 2010,

39, 3031–3056.

(6) Zrazhevskiy, P.; Sena, M.; Gao, X. Chem. Soc. Rev. 2010,

39, 4326–4354.

(42) Duan, H.; Nie, S. J. Am. Chem. Soc. 2007, 129, 3333–3338.

(43) Tan, S. J.; Jana, N. R.; Gao, S.; Patra, P. K.; Ying, J. Y. Chem.

Mater. 2010, 22, 2239–2247.

(7) Thanh, N. T. K.; Green, L. A. W. Nano Today 2010, 5, 213–230.

(8) Jana, N. R. Phys. Chem. Chem. Phys. 2011, 13, 385–396.

(9) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.; Mattoussi, H. Nat.

Mater. 2005, 4, 435–446.

(10) Oh, J. K. J. Mater. Chem. 2010, 20, 8433–8445.

(11) Smith, A. M.; Duan, H.; Rhyner, M. N.; Ruan, G.; Nie, S. Phys.

Chem. Chem. Phys. 2006, 8, 3895–3903.

(12) Medintz, I. L.; Mattoussi, H. Phys. Chem. Chem. Phys. 2009,

11, 17–45.

(13) Algar, W. R.; Massey, M.; Krull, U. J. Trends Anal. Chem. 2009,

28, 292–306.

(14) Boeneman, K.; Prasuhn, D. E.; Blanco-Canosa, J. B.; Dawson,

P. E.; Melinger, J. S.; Ancona, M.; Stewart, M. H.; Susumu, K.; Huston,

A.; Medintz, I. L. J. Am. Chem. Soc. 2010, 132, 18177–18190.

(15) Rosenthal, S. J.; Chang, J. C.; Kovtun, O.; McBride, J. R.;

Thomlinson, I. D. Chem. Biol. 2011, 18, 10–24.

(44) Prasuhn, D. E.; Feltz, A.; Blanco-Canosa, J. B.; Susumu, K.;

Stewart, M. H.; Mei, B. C.; Yakovlev, A. V.; Loukov, C.; Mallet, J. M.;

Oheim, M.; Dawson, P. E.; Medintz, I. L. ACS Nano 2010, 4, 5487–5497.

(45) Rannard, S. P.; Davis, N. J. Org. Lett. 2000, 2, 2117–2120.

(46) Liu, W.; Choi, H. S.; Zimmer, J. P.; Tanaka, E.; Frangioni, J. V.;

Bawendi, M. J. Am. Chem. Soc. 2007, 129, 14530–14531.

(47) Muro, E.; Pons, T.; Lequeux, N.; Fragola, A.; Sanson, N.;

Lenkei, Z.; Dubertret, B. J. Am. Chem. Soc. 2010, 132, 4556–4557.

(48) Zhang, Y.; Schnoes, A. M.; Clapp, A. R. ACS Appl. Mater.

Interfaces 2010, 2, 3384–3395.

(49) Stewart, M. H.; Susumu, K.; Mei, B. C.; Medintz, I. L.;

Delehanty, J. B.; Blanco-Canosa, J. B.; Dawson, P. E.; Mattoussi, H.

J. Am. Chem. Soc. 2010, 132, 9804–9813.

9495

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Journal of the American Chemical Society

ARTICLE

(50) Tietz Textbook of Clinical Chemistry, 3rd ed.; Burtis, C. A.,

Ashwood, E. R., Eds.; WB Saunders: Philadelphia, PA, 1998.

(51) Hochuli, E.; D€obeli, H.; Schacher, A. J. Chromatogr. 1987,

411, 177–184.

(52) Ueda, E. K. M.; Gout, P. W.; Morganti, L. J. Chromatogr., A

2003, 988, 1–23.

(53) Dif, A.; Boulmedais, F.; Pinot, M.; Roullier, V.; Baudy-Floc’h,

M.; Coquelle, F. M.; Clarke, S.; Neveu, P.; Vignaux, F.; Le Borgne, R.;

Dahan, M.; Gueroui, Z.; Marchi-Artzner, V. J. Am. Chem. Soc. 2009,

131, 14738–14746.

(54) Duan, H.; Kuang, M.; Wang, Y. A. Chem. Mater. 2010,

22, 4372–4378.

(55) Ghadiali, J. E.; Cohen, B. E.; Stevens, M. M. ACS Nano 2010,

4, 4915–4919.

(56) Prasuhn, D. E.; Blanco-Canosa, J. B.; Vora, G. J.; Delehanty,

J. B.; Susumu, K.; Mei, B. C.; Dawson, P. E.; Medintz, I. L. ACS Nano

2010, 4, 267–278.

(57) Delehanty, J. B.; Bradburne, C. E.; Boeneman, K.; Susumu, K.;

Farrell, D.; Mei, B. C.; Blanco-Canosa, J. B.; Dawson, G.; Dawson, P. E.;

Mattoussi, H.; Medintz, I. L. Integr. Biol. 2010, 2, 265–277.

(58) Medintz, I. L.; Deschamps, J. R. Curr. Opin. Biotechnol. 2006,

17, 17–27.

(59) Susumu, K.; Medintz, I. L.; Delehanty, J. B.; Boeneman, K.;

Mattoussi, H. J. Phys. Chem. C 2010, 114, 13526–13531.

(60) Medintz, I. L.; Clapp, A. R.; Mattoussi, H.; Goldman, E. R.;

Fisher, B.; Mauro, J. M. Nat. Mater. 2003, 2, 630–638.

(61) Clapp, A. R.; Medintz, I. L.; Mauro, J. M.; Fisher, B. R.;

Bawendi, M. G.; Mattoussi, H. J. Am. Chem. Soc. 2004, 126, 301–310.

(62) Pons, T.; Medintz, I. L.; Wang, X.; English, D. S.; Mattoussi, H.

J. Am. Chem. Soc. 2006, 128, 15324–15331.

(63) Liu, Y.; Kim, M.; Wang, Y.; Wang, Y. A.; Peng, X. Langmuir

2006, 22, 6341–6345.

(64) Delehanty, J. B.; Mattoussi, H.; Medintz, I. L. Anal. Bioanal.

Chem. 2009, 393, 1091–1105.

(65) Vivꢀes, E.; Schmidt, J.; Pꢀelegrin, A. BBA Rev. Cancer 2008,

1786, 126–138.

(66) Fuchs, S. M.; Raines, R. T. Biochemistry 2004, 43, 2438–2444.

(67) Delehanty, J. B.; Medintz, I. L.; Pons, T.; Brunel, F. M.;

Dawson, P. E.; Mattoussi, H. Bioconjugate Chem. 2006, 17, 920–927.

(68) Hoshino, A.; Hanada, S.; Manabe, N.; Nakayama, T.; Yamamoto,

K. IEEE Trans. Nanobiosci. 2009, 8, 51–57.

(69) Rouhana, L. L.; Jaber, J. A.; Schlenoﬀ, J. B. Langmuir 2007,

23, 12799–12801.

(70) Bri~nas, R. P.; Hu, M.; Qian, L.; Lymar, E. S.; Hainfeld, J. F. J. Am.

Chem. Soc. 2008, 130, 975–982.

(71) Majzik, A.; Patakfalvi, R.; Hornok, V.; Dꢁekꢁany, I. Gold Bull.

2009, 42, 113–123.

(72) Cui, R.; Zhang, M.-X.; Tian, Z.-Q.; Zhang, Z.-L.; Pang, D.-W.

Nanoscale 2010, 2, 2120–2125.

(73) Daniel, M. C.; Astruc, D. Chem. Rev. 2004, 104, 293–346.

(74) Breus, V. V.; Heyes, C. D.; Tron, K.; Nienhaus, G. U. ACS Nano

2009, 3, 2573–2580.

(75) Barglik-Chory, C.; Buchold, D.; Schmitt, M.; Kiefer, W.; Heske,

C.; Kumpf, C.; Fuchs, O.; Weinhardt, L.; Stahl, A.; Umbach, E.; Lentze,

M.; Geurts, J.; Muller, G. Chem. Phys. Lett. 2003, 379, 443–451.

(76) Zheng, Y.; Gao, S.; Ying, J. Y. Adv. Mater. 2007, 19, 376–380.

(77) Park, J.; Nam, J.; Won, N.; Jin, H.; Jung, S.; Jung, S.; Cho, S.-H.;

Kim, S. Adv. Funct. Mater. 2011, 21, 1558–1566.

(78) Choi, H. S.; Liu, W.; Misra, P.; Tanaka, E.; Zimmer, J. P.; Ipe,

B. I.; Bawendi, M. G.; Frangioni, J. V. Nat. Biotechnol. 2007,

25, 1165–1170.

(79) Choi, H. S.; Liu, W.; Liu, F.; Nasr, K.; Misra, P.; Bawendi, M. G.;

Frangioni, J. V. Nat. Nanotechnol. 2010, 5, 42–47.

9496

dx.doi.org/10.1021/ja201919s |J. Am. Chem. Soc. 2011, 133, 9480–9496

Article Doi

Multifunctional compact zwitterionic ligands for preparing robust biocompatible semiconductor quantum dots and gold nanoparticles

DOI: 10.1021/ja201919s

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja201919s

Full text of DOI:10.1021/ja201919s

Products guided by the article

R&D Labs maybe for 111-41-1

Relevant to this article

Hot Product