Quantum dots and peptides: A bright future together



Nanocrystalline semi-conductor materials, called quantum dots (QDs), exhibit unique optical and spectroscopic properties, which include, broad absorption, narrow and tunable emission, resistance to photobleaching, strong luminescence, and long luminescent lifetimes. These remarkable properties of QDs have resulted in their use as an alternative to both small-molecule and protein fluorophores in innumerable biological applications. The overlap of QDs with the rich chemistry and biology that is characteristic of the peptide arena is an emerging research area. Peptides engineered with appropriate cysteines or histidines have served as ligands for producing water soluble QDs as well as for tagging protein ligands and biosensors to QD surfaces. Incorporation of cell-penetrating peptides on QD surfaces has allowed for the translocation of functionalized QDs into cells for intracellular imaging applications. QDs containing fluorescently labeled peptide substrates have shown utility in the development of novel protease assays. Moreover, QDs-labeled peptides that serve as ligands for cellular receptors provide an alternative to antibody mediated imaging at the whole-cell and single molecule level to study receptor distribution and trafficking. This review highlights the overlap between QD and peptide chemistry and speculates on future research directions. © 2006 Wiley Periodicals, Inc. Biopolymers (Pept Sci) 88: 325-339, 2007.

This article was originally published online as an accepted preprint. The “Published Online” date corresponds to the preprint version. You can request a copy of the preprint by emailing the Biopolymers editorial office at biopolymers@wiley.com


Semiconductor nanocrystals, also known as quantum dots (QDs), have emerged as a significant new class of materials over the past decade. The ability to synthesize highly monodisperse colloidal QDs, has paved the way for numerous spectroscopic studies assigning the QD electronic states and mapping out their evolution as function of size.1, 2 Fundamental understanding of these semiconductor nanocrystals has further stimulated the development of new materials, such as quantum nanorods,3 magnetic nanocrystals,4 QD-based solar cells, and light-emitting devices.5 These related materials augment the large potential for using luminescent QDs as labeling reagents to empower numerous biotechnological applications.

The unique optical and spectroscopic properties of QDs offer a compelling alternative to traditional fluorophores in almost all fluorescence-based applications.6 The radii of QDs are smaller than the average physical separation between the electrons and holes (Exciton Bohr Radius), which results in a class of materials that dwell between the molecular and bulk forms of matter. Quantum confinement of both the electron and hole in all three dimensions leads to quantized energy levels or band gaps, which increase with decreasing crystalline size.7 Therefore, the emission of QDs can be potentially tuned from 400 to 2000 nm by altering the particle size (Figure 1). QDs can be generated for any emission wavelength with an ∼30 nm of Full-width half-maximum. The typically broad absorption spectra and narrow emission profile allow for excitation of multiple species of QDs utilizing a single light source for multiplexed fluorescence detection without substantial signal cross-coupling.8, 9 Also, the long luminescent lifetime (30–100 ns) of QDs diminishes interference from background autofluorescence in live cell imaging. Furthermore, unlike conventional dyes, QDs are extremely resistant to photobleaching, which makes them useful for continuous monitoring of biological phenomena.10

Figure 1.

Illustration of size-tunable (CdSe)ZnS QDs and their fluorescence spectra.

One of the challenges for using QDs in biological studies is to design hydrophilic, yet luminescent, QDs with a surface that is adaptable to varied biological applications. Monodisperse QDs such as those made of CdSe have been achieved by solution phase synthesis and selective separation methods with a tunable size from 1 to 20 nm.1 However, the luminescent properties of naked CdSe nanocrystals are very sensitive to surface and solvent interactions. The surface defects that lead to nonradiative paths for the electron-hole pair recombination make the quantum yields of bare QDs relative low (10–15% for CdSe).11 Overcoating CdSe nanocrystallites with higher band gap inorganic material, such as ZnS, significantly improves luminescence yields by electronically passifying surface nonradiative recombination sites.12, 13 The quantum yield of these highly luminescent (CdSe)ZnS core-shell QDs can reach >50%.12 Thus, this class of QDs is considered the most promising alternative to organic dyes for fluorescence-based applications and is primarily discussed in this review.


To utilize QDs for imaging in biological environments, numerous efforts have been devoted to coating core-shell QDs with water compatible organic materials. Typically, QDs are synthesized in the presence of hydrophobic inorganic surfactants, such as trioctylphosphine oxide (TOPO). The phosphine oxide functionality chelates the QD surface while the long alkyl chains allow for solubility in nonpolar solvents such as hexane and toluene. The most common method for synthesizing water-soluble QDs is through ligand-exchange. In initial studies, thiol-containing molecules, such as mercaptoacetic acid (MAA) (Figure 2A), were used to replace the hydrophobic TOPO coatings on core-shell QDs,14, 15 where the mercapto group thiolates the nanocrystal surface and the carboxylic acid provides solubility in an aqueous environment. This method is simple, however thiolation to QD is a kinetic chelation process, which results in the labile coating of the QDs16 and long-term instability accompanied by significant decreases in the quantum yield to 7% compared with TOPO-coated QDs.17 To overcome this kinetic instability, dihydrolipoic acid (DHLA, Figure 2B) was subsequently used as a bidentate coating material, which increased the quantum yield to 10–20%.18 However, the desorption of the thiol ligand can still lead to eventual dissolution and aggregation of the nanocrystals, resulting in the release of heavy metal atoms into solution.19 Alivisatos and coworkers developed an elegant coating method, where a silica shell was synthesized to protect the QDs (Figure 2C) and different functional groups, such as primary amines, carboxylic acids, or thiols, could be incorporated on the silica shells for bioconjugation.19 These coated QDs were extremely stable and retained 60–80% of the quantum yield of the original QDs. However, this method, though potentially useful, has not seen significant applications in imaging and related application, possibly because of the time intensive multi-step processing. Phospholipids (Figure 2D),20 polypeptides,21 and bovine serum albumin22 have also been successfully used as capping agents but lack simple and reliable methods for further bioconjugation.

Figure 2.

Common methods for synthesizing biocompatible QDs. (A, B) QDs coated with MAA or DHLA through ligand exchange. (C) QDs coated through ligand exchange followed by silanization. (D) QDs coated with phospholipids. (E) QDs coated with amphiphilic di-block copolymers.

In the past several years, amphiphilic di- and tri-block copolymers, typically containing polyacrylic acids, have been developed for encapsulating QDs via spontaneous self-assembly (Figure 2E).23 This method allows one to circumvent the kinetic chelation based approaches, while maintaining the photoluminescent properties of the QD labels. In most designs of the amphiphilic polymers, a fraction of the carboxylic acid functionalities in the polyacrylic acid are coupled to hydrophobic alkyl chains, which stabilize TOPO capped QDs. The remaining carboxylic acids in the polymer provide solubility in water and can be utilized as chemical handles for conjugation to primary amines in proteins through water soluble cross-linking reagents such as 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDAC). However, almost all proteins and many peptides of interest contain more than one free amino group, many of which have important roles in their specific biological functions. Thus, the EDAC coupling strategy may not always be the best choice for site-specific conjugation of QDs to peptide or protein ligands. Recently, free primary amines have been incorporated in an amphiphilic polymer for QD coating that can be further modified with standard N-hydroxylsuccinimyl ester-malemide containing bifuctional cross-linkers. The maleimide functionalized QDs can be conjugated to any macromolecule selectively through engineered cysteines or thiols on the target ligands24 (Figure 3). Even though the earlier-described methods provide a starting point for synthesizing water miscible QDs for imaging applications, new chemistries for providing stable coatings as well as orthogonal approaches for ligand labeling of the QDs remain of much current interest.

Figure 3.

Maleimide functionalized QDs for conjugating thiol-containing ligands. TOPO stabilized QDs are coated with a primary amine functionalized tri-block amphiphilic copolymer for producing water-soluble QDs, which facilitate further conjugation to ligands with free thiols through bi-functional cross-linkers.

The remainder of this review will focus upon the utility of peptides as an enabling agent in Quantum Dot chemistry.


Peptides have been recently utilized to improve QD solubility in aqueous solution, and in connecting QDs to targeting protein and peptide ligands. Given the success of thiolates in coating QDs, peptide sequences containing cysteine residues have been used to exchange out the TOPO coating on QDs, while appropriate hydrophilic residues facilitate solubility in aqueous solutions. Of note is an example from Pinaud et al.21 where they coated QDs with peptides composed of two domains, a hydrophobic phytochelatin derived adhesive domain rich in cysteines for thiolation to QDs and a hydrophilic domain composed of glycine, serine, and glutamate for increasing the QD solubility in water. While Gomez et al.25 used a mixture of the peptide, CGGGRGDS, and mecaptopropionic acid to simultaneously coat and target QDs in a one step procedure. The peptides helped in exchanging out the TOPO ligand, while at the same time providing the necessary functionality for targeting receptors with the RGD epitope.

In addition to the use of a peptidyl cysteine in constructing water-soluble QDs, histidine can also be engineered into peptides to stabilize QDs in aqueous environments. Similar to His tags, which have been commonly used in protein purification, heterocyclic aromatic amines such as pyridine, imidazole, and histidine can bind to the surface of QDs through Zn coordination.26 Nie and coworkers used histidine (0.5–1M) to coat QDs (Figure 4A), but found a significant decrease in QD fluorescence, possibly because of the low affinity and kinetic instability. By taking advantage of the chelation effect, Mattousi and coworkers engineered histidine tags at the end of proteins such as the maltose binding protein (MBP)27 as well as peptides,28, 29 to significantly increase their affinities30, 31 to the QDs (Figure 4B). Interestingly, when the His-tag fused MBP was bound to the DHLA coated QDs, the quantum yield increased from 15.6% to 39%27 though the mechanism for this increase is unclear. Appropriate design of new peptides or peptide-polymer hybrids32, 33 with appropriate chelating functionalities will certainly help provide new robust coating and QD labeling reagents.

Figure 4.

QD coated with Histidines. (A) A ligand labeled with single histidine binds to QD with low affinity. (B) A peptide containing five histidines increases binding affinity through a chelate effect.


QDs are often targeted to cellular proteins, such as cell-surface receptors, through specific antibodies (Figure 5A). The primary antibody targets the cell-surface receptors, and is then recognized by a biotinylated secondary antibody, which binds a streptavidin coated QD. This sandwich-imaging format leads to very large molecular complexes that may hamper accurate imaging of receptor densities and receptor dynamics. To minimize the size of the QD-label complex, Ting and coworkers34 used an ingenious approach to genetically encode a 15 residue peptide at the terminus of a cell-surface protein of interest. The peptide served as a substrate for biotin ligase, allowing for convenient and selective biotinylation, which allowed for the subsequent recruitment of the QD-labeled streptavidin (Figure 5B). Using this peptide-biotinylation strategy, QD mediated imaging of epidermal growth factor receptors in HeLa cells as well as α-amino-3-hydroxy-5- methyl-4-isoxazolepropionate receptors in neural cells was demonstrated. This selective biotinylation method as well as utilizing peptide motifs35–37 that provide orthogonal binding to streptavidin and avidin will likely provide new methods for targeting QDs and increase the possibility of multiplexing.

Figure 5.

Peptides for noncovalent capture of QDs. (A) General approach for QD mediated labeling of a cell surface protein through secondary and primary antibodies. (B) A peptide encoded at the terminus of a cell surface protein can be selectively biotinylated for subsequent imaging utilizing streptavidin coated QDs.


Though many imaging applications utilizing QDs are targeted to cell-surface proteins, intracellular targeting remains challenging. It has recently been demonstrated that QDs can be delivered into cells by receptor-mediated or nonspecific endocytosis.38 In living cells, many protein transport events are mediated by positively charged peptide transduction domains (PTDs), which likely bind anionic polysaccharides on the cell surface,39 such as heparan sulfate to facilitate entry.39, 40 Thus, several strategies have been designed to deliver QDs into cells by coating QDs with positively charged peptides. One of these PTDs, HIV-1 TAT, first identified in 1988,41, 42 is a cell permeable peptide that delivers heterologous proteins into cells.43 The strongly positively charged TAT peptide sequence GRKKRRQRRR and its analogs are thought to be responsible for the translocation process, and have been fused to other proteins to assist in translocation.44 Since QD sizes are comparable to proteins, de la Fuente et al. coupled the TAT peptide to tiopronin-functionalized QDs and successfully translocated them to the nucleus of the hTERT-BJ1 human fibroblast cells (Figure 6).45 Similar to MAA, tiopronin contains a free thiol and a carboxylic acid. TAT peptides coupled to the tiopronin coated CdS QDs by EDAC, were added to cells for 15 min followed by washing. Fluorescence staining was observed around the cell nucleus, which provided evidence that the QD-TAT complexes were indeed translocated to the nucleus. No fluorescence was found in the cells in which QDs without TAT peptide was incubated. Santra et al. engineered a cysteine in the TAT peptide and conjugated it to an amine-functionalized Cd:Mn/ZnS QD.46, 47 These conjugates penetrated the rat blood-brain-barrier and the labeled brain tissue was visualized in vivo. These experiments indicate that the TAT-QDs migrated to brain parenchyma and reached cell nuclei, which supported the previous report that TAT peptides can rapidly translocate through the plasma membrane and accumulate near the cell nucleus.48 Rozenzhak et al. designed a QD-streptavidin-dual-peptide complex to target HeLa cells.49 One of the biotinylated peptides contained a tryptophan rich region for interacting with the membrane and a lysine rich region for nuclear localization.50 The other peptide, also biotinylated, was derived from GH3 domain of the Grim protein, which is expressed in response to mitochondrial-mediated cell death in mammalian cells.51 Streptavidin coated QDs were mixed with the two peptides and incubated with HeLa cells. Fluorescence micrographs demonstrated nuclear condensation followed by apoptosis. Recently, Delehanty et al.29 designed a cell penetrating peptide (CPP) containing a His-tag, linker, and 8-mer arginine; here, the His-tag could bind QDs with high affinity and the 8-mer arginine assisted in the QD endocytosis. They assembled QDs with the CPP at different ratios, and used different concentrations of QD-CPP to study endocytosis in HEK293T/17 and COS-1 cells. Their results showed that the cellular internalization of QD-CPP was dose-dependent, and a higher fraction of CPP assembled on the QD also had higher intracellular fluorescence. Furthermore, they also investigated the cytotoxity of these QD-CPP assemblies. Both acute and chronic toxicity were measured separately. Importantly, cellular proliferation assays indicated that almost no acute toxicity was found in QDs, with CPP or QD-CPP. However, chronic toxicity was found for QDs and QD-CPP at relatively high concentrations, suggesting that QDs may have long-term toxicity in biological systems, which may preclude their use in human subjects.

Figure 6.

Tat peptide assisted QD endocytosis. Scale bars correspond 20 μm. (A) hTERT-BJ1 human fibroblasts without QDs. (B) hTERT-BJ-1 human fibroblasts incubated with QD/tiopronin complexes without Tat peptide. (C) hTERT-BJ1 human fibroblasts incubated with QD/tiopronin-Tat complexes. (Reproduced from Ref.45, with permission from Wiley).

However, peptide-assisted nanoparticle endocytosis remains attractive because of the ease of use and high quantum efficiency for studies in cell-culture and model organisms. Besides the aforementioned peptides, polyarginine,52, 53 polylysine,54 and other positively charged peptides55, 56 have also recently been shown to be suitable for the delivery of QDs into cells. New methods for intracellular targeting, such as designed receptor/ligand systems57, 58 as well as new PTDs,59 can also be potentially adapted for QD delivery.


Biosensors have widespread applications in all areas of biological research and diagnostics. The main physical phenomena exploited for optical chemical sensing are absorption and fluorescence, although chemical luminescence, Raman scattering, plasmon resonance, and electrical sensing are also of considerable interest. Since QDs exhibit excellent optical properties, they have been utilized as biosensors for detecting maltose,27 aptamers,60 2,4,6-trinitrotoluene,61 toxins,62 and DNA.63 QDs have also been utilized for measuring enzymatic activity for β-lactamase64 and several proteases.28, 65, 66 Fluorescence resonance energy transfer (FRET) is the most commonly utilized technique in these applications because of the high sensitivity, good reproducibility, and real-time monitoring capabilities. As QDs have broad absorption spectra and narrow emission spectra, they are usually used as fluorescence donors while the fluorescence acceptor is often an appropriately labeled peptide. A FRET biosensor based on a QD-protein/peptide conjugate was developed by Mattoussi and coworkers for detecting maltose.27, 67, 68 In initial studies DHLA was used to thiolate the outer shell of the QDs and make them biocompatible, while a poly-histidine tag (5His) at the C-terminus of the maltose binding protein (MBP) allowed for chelating to the QD surface. An organic dye (fluorescence acceptor) was conjugated to β-cyclodextrin (β-CD), which occupied the saccharide-binding pocket of the MBP and resulted in a FRET signal. In more recent work,27 two approaches have been designed and tested as maltose sensors. In the first design, the organic dye QSY-9 (absorption ∼560 nm) labeled β-CD was used to bind to MBP attached to the QD560 (emission ∼560 nm) surface (Figure 7A) such that the fluorescence was quenched. When free maltose was added, it displaced the β-CD-QSY9, resulting in fluorescence recovery from the QD. The optimized sensor contained 10 copies of β-CD-QSY9 bound to the QD complex, where 75% of the QD fluorescence was quenched. Upon complete displacement of β-CD-QSY9 with maltose, an increase in fluorescence of ∼3-fold was observed. Some complications in these applications arise from uncertainty in the distance between QD and acceptor and recent studies indicate that the FRET efficiency may increase through electron transfer between QDs and organic dyes.69 To overcome these limitations, Mattoussi and coworkers designed a maltose sensor (Figure 7B), where 10 copies of Cy3 (absorption 556 nm and emission ∼570 nm) labeled MBP were first incorporated on the QD530 (emission ∼530 nm) surface, followed by binding of the Cy3.5 (absorption ∼575 nm and emission ∼595 nm) labeled β-CD. Fluorescence energy was first transferred from the QD530 to MBP-Cy3, and then the emission energy of Cy3 was transferred to β-CD-Cy3.5. The displacement of β-CD-Cy3.5 from MBP by maltose resulted in concomitant fluorescence decrease from Cy3.5 and an increase from Cy3. This geometry allowed QD530 to transfer sufficient energy to Cy3, but minimized direct energy transfer to Cy3.5. These two methods demonstrate how appropriately designed QD complexes with peptide immobilization tags can be used in determining small molecule concentrations in the 100 nM–10 μM range. Given the large number of peptide domains that have been used in fluorescent sensor architectures it should certainly be possible to couple them to the powerful optical properties of QDs.

Figure 7.

QD based maltose nanosensor. (A) β-CD-QSY-9/MBP complex bound to QD through a peptide His-tag, where QD fluorescence is quenched by QSY-9. The displacement of β-CD-QSY-9 with maltose results in QD fluorescence recovery. (B) β-CD-Cy3.5/MBP-Cy3 complex bound to QD through a peptide His-tag, and Cy3.5 fluorescence emitted through a two-step FRET process. The displacement of β-CD-Cy3.5 with maltose results in Cy3 fluorescence emission.


Most current methods to detect protease activity are based on peptide conjugated organic fluorophores. Auto-quenched probes,70 dual chromophore probes,71 multiphoton FRET,72 peptide-based NIR fluorescence probes,73 etc. have been recently developed to determine protease activities. However, organic dyes often have problems such as photobleaching, lack of tunable wavelengths, sensitivity to environment, and requirement for specific paring between donor and acceptor. Thus, QD labeled peptides are emerging as promising candidates to determine protease activities. In 2005, Chang et al. designed QD probes to target collagenase,65 which is a metalloprotease of therapeutic interest. The CdSe/CdS QDs were coated by a poly (ethylene glycol) (PEG) polymer to increase water solubility and were also functionalized with carboxylic acids. A peptide, GGLGPAGGCG, was chosen as the substrate, which can be cleaved between alanine and glycine in the presence of collagenase. The N-terminal amines of the peptides were coupled to the carboxylic acids on the QDs by EDAC, and the cysteine was conjugated to the maleimide functionalized gold nanoparticles, which quench the QD fluorescence. Incubating this QD-Peptide-Gold substrate with collagenase cleaved the peptides, and subsequently released the gold quencher, leading to recovery of the QD fluorescence. About 30 copies of the peptide were coupled to each QD, while six equivalents of the gold nanoparticles were tethered to each QD. Though a very interesting approach, the signal from these assays was not substantial. The excess peptides on the QD possibly decreased the probe sensitivity to protease hydrolysis, while the gold nanoparticles contain several maleimides possibly resulting in crosslinking between the QDs and the formation of higher order complexes. Additionally, utilizing QDs and gold nanoparticles may also reduce enzyme accessibility to the substrate.

Recently, Medintz et al. developed probes for several different proteases utilizing a similar approach (Figure 8).28 For a caspase-1 assay, they designed a peptide substrate HHHHHHGLAibAAGGWEHDSGC, where the histidine tag was used to bind QD surface, the Aib (α-amino isobutyric acid) was used to break the helical structure of the peptide, the cysteine was used to conjugate to an organic fluorophore or quencher, and the amide bond between glycine and tryptophan was designed to be cleaved by caspase-1. In this case, DHLA coated QD538 (emission at 538 nm) and the Cy3 dye (absorption at 555 nm) were used as the FRET pair. Several advantages are seen in this design. First, each QD was conjugated to about three copies of peptide substrate containing the Cy3 acceptor. Therefore, all peptide substrates on QD likely measure protease activity, which increases the sensitivity of the probe. Second, Cy3 is a relatively small molecule compared with the gold nanoparticles, which possibly increases the substrate accessibility to the protease. Enzyme kinetic analysis revealed that the KM in this geometry was comparable with the KM value reported for normal peptide substrates in solution. Thus, it is likely that the substrates incorporated on the QD complex allow easy access to the desired protease. Moreover, the sensitivity and background of these probes can be easily tuned. Even though eight copies of Cy3 result in almost complete transfer of energy from the QD emission only three copies of Cy3 were required to reduce the QD emission by 60%, which suggests the fluorescence change of the probe is most sensitive with fewer dye substitutions. Here, high Cy3 substitution can completely quench the QD emission and minimize the background, but the FRET change is also minimized at this substitution level. However a remaining problem was that the fluorescence of QDs and Cy3 overlap, thus the change in signal had to be deconvoluted. In subsequent studies, to overcome the fluorescence overlap problem, peptide probes for three other proteases (thrombin, collagenase and chymotrypsin), were conjugated to the quencher QXL-520 instead of Cy3. All of these probes showed good sensitivity to their proteases, and the KM values were comparable with those previously reported in solution.

Figure 8.

QD based protease activity sensors. Quencher conjugated protease substrate peptides bind the QD through a His-tag, where QD fluorescence is quenched through FRET. Upon proteases mediated substrate cleavage the QD fluorescence is recovered.

More recently, Shi et al. synthesized a QD FRET-based protease sensor66 to measure the protease activities in vivo. Approximately 48 copies of a rhodamine labeled collagenase substrate RGDC were coated on each QD through a unique cysteine. The QD had an emission at 545 nm, which matched the absorption wavelength of rhodamine. Peptide cleavage by collagenase caused QD fluorescence at 545 nm to increase and rhodamine fluorescence at 590 nm to decrease. Ratiometric changes in fluorescence were used to quantitatively study protease activity in HTB126 cancer cell lines in which collagenase is aberrantly expressed. These experiments successfully demonstrated that protease activity could be measured in real time by using QD FRET-based ratiometric sensors. Since activities of different proteases can be measured by changing the peptide sequences, it is likely that more and more proteolytic enzyme assays will be performed utilizing QD FRET based sensors with the opportunity for the introduction of multiplexing as well as high throughput formats.


Peptide-QD based nanosensors have also been recently utilized for studying the interactions between a peptide sequence in Rev (an important HIV-1 regulatory protein) and the Rev responsive element (RRE, part of the env gene of the HIV-1 RNA genome).74 This interaction between Rev and RRE plays an important role in HIV replication. Zhang and Johnson used QDs instead of organic dyes to develop a sensor to determine the dissociation constant (Kd) between Rev and RRE based on FRET (Figure 9). The QDs (emission ∼605 nm) were functionalized with streptavidin, which bound biotin-labeled RRE. The RRE functionalized QDs could effectively complex 30 copies of a Cy5 labeled 17-mer peptide containing the arginine-rich region of Rev, thereby reducing the QD fluorescence and increasing Cy5 fluorescence. QD fluorescence could be recovered by titrating unlabeled Rev peptide into the FRET complex. By monitoring the fluorescence change of QD and Cy5 in this titration process, the Kd of the Rev-RRE complex was calculated. The titration curves from the two fluorophores (loss in FRET and increase in nonFRET emission) gave similar results (∼11–14 nM), and were in excellent agreement with literature data. This experiment provides a simple method for monitoring Rev-RRE interaction and possibly any peptide/RNA, peptide/DNA, or peptide/protein interaction.

Figure 9.

QD based nanosensors for detecting Rev-RRE interactions. (Helix represents secondary structure of RRE IIB RNA. Nucleotides important for Rev binding are shown in red.) Cy5 labeled peptide Rev assembled with RRE on the QD complex resulted in QD fluorescence quenching through FRET.


Ligand-receptor interactions are fundamental to most biological processes. These interactions trigger changes in cell metabolism, membrane charge, and transcription. As QDs emerge as useful reagents in imaging, many efforts have been undertaken to investigate ligand-receptor interactions by utilizing QD conjugated ligands to target receptors in cells. In most of the cases, antibodies targeting the receptors of interest75–80 are directly or indirectly incorporated onto the QD surface. However, since many cell-surface receptors bind cognate proteins with known peptide epitopes, or alternatively bind known peptides and their analogs, it should be possible to directly conjugate appropriate peptides to the desired water-soluble QD for imaging applications. Compared with antibodies, peptides have several advantages since they can be covalently conjugated to a QD surface. Antibodies, on the other hand, have to be noncovalently complexed to the QDs, usually through biotin–streptavidin conjugates (Figure 4A). This may ultimately limit multiplexing capabilities, where numerous receptors need to be imaged. Second, the large size of antibodies may increase the likelihood of nonspecific interactions with the antibody (or even streptavidin) and the cell-surface proteome. Third, both the antibody and streptavidin significantly increase the size of the QD complex (∼50 nm), which could potentially affect accessibility to crowded cell-surface locations34, 81 and also interfere with receptor dynamics. A final advantage of using peptide ligands in QD complexes is that one QD can be conjugated to the desired number (10–100s) of peptide ligands (instead of only about five antibodies23, 82). The complexes may therefore exhibit tunable binding affinity and possibly achieve higher targeting efficacy through multivalent interactions. We describe several systems where peptide conjugated QDs have been successfully utilized for imaging cell-surface receptors ranging from growth factor receptors, integrins, and G-protein coupled receptors.

Nerve growth factor (NGF) is the prototype for the neurotrophin family of polypeptides, which are essential in the development and survival of certain sympathetic and sensory neurons in both the central and peripheral nervous systems. Vu et al. chose the β-subunit portion of βNGF to target tyrosine kinase A (TrkA) receptors.83 The peptide was first biotinylated,84 followed by binding to a QD-streptavidin conjugate, and used to target receptors in PC12 cells. The peptide-QD complex was useful not only in imaging receptor distribution but also in effectively activating the targeted TrkA receptor to promote neuronal differentiation.

There is much current interest in integrin targeting. Integrins expressed on endothelial cells modulate cell migration and survival during angiogenesis while integrins (such as αvβ3) expressed on carcinoma cells play a key role in tumor angiogenesis and metastasis. Several groups have used QD-peptide complexes to target integrins to study the distribution of integrins on cell surfaces. Winter et al.80 used both IgG-QD conjugates and RGD peptide labeled QDs to target integrin αv subunit in SK-N-SH cell lines,85 where there is high expression of the αvβ1 complex.86 A short peptide sequence CGGGRGDS was chosen as the recognition module87 in which the RGD tripeptide is known to bind the αvβ1 and αvβ3 integrins and the cysteine was used to bind to the QD surface. The imaging results demonstrated that both the IgG-QD conjugates and RGD peptide labeled QDs can be used to target receptors in living cells, but the peptide coated QDs showed significantly lower aggregation. It is possible that IgG functionalized QDs are more susceptible to aggregation because of multiple secondary antibody binding and cross-linking. Recently, Cai et al.88 reported a cyclic RGD peptide c(RGDyK), a potent integrin αvβ3 antagonist, for targeting tumor vasculature through conjugation to Near IR, CdTe QDs. The imaging of frozen human glioblastoma U87MG tumor tissue staining (high integrin αvβ3 expression) clearly indicated that the QD-RGD peptide complex specifically bound to the αvβ3 receptors (Figure 10A). Interestingly, the dynamics of ligand binding could be measured in mice models through tail vein injection, with imaging at different time points. It was shown that fluorescence at the tumor site reached a maximum at 6 h, and faded after 27 h (Figure 10B). This is the first example of RGD labeled QD utilized for in vivo targeting and imaging of integrin αvβ3-positive tumor vasculature in a murine xenograft model.

Figure 10.

c(RGDyK) functionalized QDs targeted to U87MG tumor tissues in vitro and in vivo. (A) Frozen U87MG tumor tissue staining using 50 nM QD705 (control experiment) and QD705-RGD. (B) In vivo NIR fluorescence imaging of U87MG tumor bearing mice injected with 200 pmol of QD705 (left) and QD705-RGD (right), respectively. Images was taken at the 6th h after injection. (Reproduced from Ref.88, with permission from American Chemical Society).

In 2002, Akerman et al. conjugated several peptide ligands to QDs to target cells of interest in vivo.89 The QD-peptide complex successfully targeted specific tissues and cell types without accumulation in the reticuloendothelial system. The three ligands utilized were: a 13-mer peptide with a sequence CGFECVRQCPERC (GFE) which binds to membrane dipeptidase on the endothelial cells in lung blood vessels90; a 31-mer peptide KDEPQRRSARLSAKPAPPKPEPKPKKAPAKK (F3) which is a fragment of HMGN2 protein and binds to HL-60 and MDA-MB-435 tumor cells91; a shorter peptide CGNKRTRGC (Lyp-1) that recognizes lymphatic cells.92 The peptides were first functionalized with free thiols by reacting with 3-mercaptopropionimidate (iminothiolane). The thiol functionalized peptides were incubated with the thiol-PEG absorbed MAA- coated QDs to form QD-ligand complexes. In vitro experiments demonstrated the imaging of lung endothelial (LE) cells (in which membrane dipeptidase was expressed) with the GFE-QDs. When free GFE ligands or the dipeptidase inhibitors were added, QD fluorescence significantly decreased. The F3-QDs and Lyp-1-QDs also showed binding specificities to their cellular targets. In vivo experiments indicated that GFE-QDs were only found in lungs but not in the brain, kidneys, or other organs, while the F3-QDs and Lyp-1-QDs only accumulated in tumor tissues. Moreover, the QD-ligand complex uptake in the reticuloendothelial system indicated that the PEG coating on QD increases the circulation half-life of the complexes while reducing the accumulation of the complexes in the liver and spleen.

Another class of cell-surface receptors of much current medicinal interest is the G-protein coupled receptors many of which have known peptide and protein ligands. We have recently imaged the human δ-opioid receptor (hDOR) at the cellular and single molecule level. We utilized appropriate polymer coated, maleimide-functionalized QDs (Figure 3) to conjugate a potent peptide ligand of the δ-opioid receptor, deltorphin II(Ile-Ile), through an engineered cysteine (Figure 11). The interactions of the QD labeled ligands with the cognate receptors were visualized on the surface of fixed and living cells24 (Figure 11). The deltorphin II(Ile-Ile) labeled QDs were incubated with hDOR over-expressed in CHO cells and labeling was observed after 1 min. Since deltorphin II(Ile-Ile) is a known agonist, it can activate hDOR and undergo internalization. Interestingly the distribution of hDOR-internalization could be clearly seen after 20 min, suggesting that the agonist-mediated activity is retained in ligand conjugated on QDs (Figure 11, Part I). This QD-ligand complex was also used for probing hDOR at the single molecule level, which can be useful for interrogating receptor localization and trafficking in live cells and in deconvoluting ligand binding kinetics in model bilayers.93 To compare the QD versus traditional organic dyes in single molecule detection, deltorphin II(Ile-Ile) was labeled with QD and Cy3 individually. Both ligand-dye conjugates were incubated with hDOR in a supported lipid bilayer. The single molecule measurements showed that both of the two complexes had similar average residence times (∼50 s). Importantly, the fluorescence from QD-complex was sixfold stronger than that from Cy3 (Figure 11, Part II) even under unoptimized conditions, and was also significantly resistant to photobleaching. Thus, the increased fluorescence, fluorescence lifetimes, resistance to photobleaching, and possibility of multiplexing, may provide an alternative approach to imaging cell surface receptors in single molecule experiments.

Figure 11.

Deltorphin II(Ile-Ile)-QD595 targets hDORs in whole cells and as single molecules. I: Image of deltorphinII(Ile-Ile)-QD595 targeting of human δ-opioid receptor (hDOR) expressed in CHO cells. (A) Cells imaged after 1 min post incubation show only surface labeling. (B) Cells imaged after incubation for 20 min showing significant organization of deltorphinII(Ile-Ile)-QD595 distribution at or near the cell surfaces. (C) Image is a magnification of (B) showing receptor distribution in four cells, at 15 min following the 1-min labeling period. II: Single molecule fluorescence of deltorphin II(Ile-Ile)-QD595 and deltorphin II(Ile-Ile)-Cy3 targeted to hDOR in bilayer membranes. (Reproduced from Ref.24, with permission from American Chemical Society).


Interestingly, though the size-based tunabilty of the fluorescence emission of QDs are possibly their biggest attraction, only a few reports have taken advantage of this in imaging studies. In one such example, Chan et al.94 utilized two different sized QDs (QD605 and QD525) and DAPI to probe three different targets (tyrosine hydroxylase, Vmat RNA, and nuclei) in cells from midbrain sections by excitation with a single blue laser and appropriate emission filters (Figure 12). Multiplexed cellular detection was demonstrated by a combination of a QD-antibody, QD-oligonucleotide, and a nuclear stain. Other QD-based multiplexed imaging has also been recently reported.68 Multiplexed applications are likely to have considerable impact in cellular imaging as well as in biosensor applications.

Figure 12.

Simultaneous multiplex labeling of mRNA and protein in the same cell using QD525 and QD605. (A) Image of QD525 labeled sequence that recognizes Vma2 RNA. (B) Image of QD605 labeled antibody that recognizes tyrosine hydroxylase (TH). (C) Image of nuclei labeled by DAPI. (D) Overlay of (A, B, and C). All chromophores were excited at 405 nm. Scale bar = 15 μm. (Reproduced from Ref.94, with permission from Oxford University Press).


This review highlights the utility of QDs and QD-peptide conjugates in numerous imaging and biosensor applications based on their impressive optical properties, which include broad absorption yet narrow and tunable emission, resistance to photobleaching, strong fluorescence, and long fluorescence lifetimes. QDs are similar to proteins in their dimensions and can be used as fluorescent labels, much like GFP. Multiple copies of small peptide ligands can be easily attached to the QDs by biotin–streptavidin interactions or direct covalent labeling of precoated QDs. Many imaging and biosensor-based applications of QD-peptide conjugates have already been reported. It is quite likely that multiplexed applications utilizing several QD-peptide ligands for targeting multiple cellular receptors will emerge as a powerful imaging method for dissecting complex receptor distribution and trafficking. Moreover, the ability to create CPP fusions with QDs will certainly also allow for multiplexed imaging both in an intra and extracellular context. The control of peptide density on QD surfaces can possibly be utilized to significantly enhance the binding affinity of QDs to their biological targets and facilitate imaging. Moreover, the size and ease of surface modifications may allow QDs to be conjugated to a set of different peptide ligands that can potentially enhance specificity for specific cell types based on their receptor complement and distribution. Given the rapid growth in new QD based materials and the well-established utility of peptides in biology and chemistry, the overlap between these fields will certainly be quite bright.


The authors thank members of the Ghosh group for many helpful comments and Luisa Gronenberg and Eri Nakatani, for contributing to the QD projects. The authors thank Victor Hruby, Ron Lynch, Mary Wirth, and Oliver Monti for numerous discussions. MZ thanks BIO5 for a fellowship.