figure[Getty Images]

For decades, a variety of biomedical-optics applications—such as clinical diagnostic assays, biosensors and nanoscopic probes for monitoring cellular and subcelluar activity—have relied on organic fluorescent dyes. These fluorophores can be readily modified to reactive versions that can be tied to specific functional proteins, peptides, and cell membranes and organelles, to drive a range of imaging and diagnostic techniques.

In recent years, however, a new player has emerged to challenge the supremacy of dyes: luminescent colloidal nanocrystals, also known as quantum dots (QDs). Luminescent QDs have a number of characteristics—including excellent photostability, outstanding luminescence quantum yields, broad absorption bands, size-tunable narrowband emissions and very large one- and two-photon absorption cross sections—that make them potentially superior to organic dyes in a range of biomedical applications. And the past decade has seen significant advances to address some key limitations of QDs, by improving their biocompatibility and through new methods for understanding and reducing their cytotoxicity.

Here, we take a broad look at how QDs work, and some techniques and technologies that can make them suitable for biomedical applications. We also walk through a few examples of such applications, in sensing and imaging, drawing partly on the experience of our own labs.

Color-tunable by size

The QDs most commonly used in biomedicine consist of a binary nanocrystalline semiconductor core, such as cadmium selenide (CdSe), surrounded by a thin protective shell composed of a wider-bandgap semiconductor such as zinc sulfide (ZnS). The shell acts both to passivate surface defects and to protect the QD from environmental attack. It thus preserves the QD’s electronic and luminescent properties, and also prevents toxic Cd from leaking out of the dot and into the biological system.

One of the most useful properties of QDs is that their optical properties can be controlled simply by varying the nanocrystal’s size. The reason relates to quantum confinement effects that kick in when the diameter of the nanocrystal becomes smaller than the Bohr exciton radius—essentially, the most probable distance separating electrons and holes in the semiconductor structure.

The energy structure of bulk semiconductors includes a valance band and a conduction band, separated by a small band gap. As the size of the material shrinks below the Bohr exciton radius, however, that diffuse energy-band structure transforms into discrete, quantized energy levels akin to those of single atoms (indeed, QDs are sometimes referred to as “artificial atoms”). Further, as the “degree of confinement” increases (that is, as the diameter of the crystal decreases), the energy gap between the ground and excited states increases as well, which in turn implies a decrease in the emission wavelength when the excited state relaxes to the ground state (see figure below).


The notion of quantum confinement, and the increase in energy gaps with reduced particle size, is usually understood using the classic “particle-in-a-box” model of quantum mechanics, describing a particle moving in a confined space in an infinite potential well. When the size reaches the scale of a QD, the electron and hole are confined within the volume of the dot and the surface represents the confining potential. Mathematically, the energy levels in the system are inversely proportional to the box length—so, as the dots become smaller, the energy gap increases as well.

That means that for CdSe QDs, for example, a 2.8-nm-diameter dot emits in the blue, while a 4.8-nm-diameter dot emits in the red. Hence the absorption and emission wavelengths can be tuned by controlling the size of the QD—and a single synthetic approach can be used to create QDs that emit a broad range of distinct colors, simply by controlling the size of the QDs during the growth process. In addition to CdSe dots, a number of other QD materials with different core and shell compositions also exist that cover the spectral range from ~350 nm to beyond 5000 nm.

The atom-like structure of multiple discrete energy levels in QDs also gives the dots broad absorption profiles. The absorption spectra of monodisperse solutions of CdSe/ZnS QDs—that is, solutions containing only a single dot size—show a lowest-energy excitonic absorption peak that corresponds to the 1Sh → 1Se electron-hole transition; weaker 1Ph → 1Pe and 1Dh →1De transitions; and then a near-continuum absorption that becomes stronger at shorter wavelengths. The emission spectra for QDs, by contrast, consist of a single, relatively sharp feature at a wavelength slightly longer than the first excitonic absorption band (as the emission spectra tend to be dominated by relaxation to the lowest excited state, and emission from higher excited states is generally absent).

The broad absorption spectra mean that a single excitation wavelength can be used to excite multiple, different-sized QDs in a single solution. As a result, it’s possible to perform multiplexed assays using QDs—a capability that is very hard to attain with most organic dye labels.

Making QDs—and making them useful

Synthesizing CdSe QD cores commonly involves using a mixture of coordinating ligands such as tri-n-octylphosphine (TOP), tetradecylphosphonic acid and hexadecylamine, to help control the growth rate and ultimate size of the dots. ZnS shells are then grown on the surface of these CdSe cores, one monolayer at a time, until the shell reaches the targeted thickness—typically one to five monolayers (0.3–1.5 nm) thick, depending on the application. For example, QDs with thicker shells tend to be more stable and have larger luminescent quantum yields, which makes them ideal for cell labeling. On the other hand, a thin shell is desirable for applications using Förster resonance energy transfer (FRET), because the separation distance between the QD center and the molecular acceptor, R, should be minimized, since the FRET efficiency falls off as 1/R6.

Using the dots for actual biological applications requires chemical modifications at the surface, to make the dots water soluble and also stable in highly saline biological environments.

This synthesis process has a disadvantage, however: it leaves the newly created CdSe/ZnS QDs with a coating of the hydrophobic coordinating ligands, which in turn make the dots insoluble in water. Using the QDs for actual biological applications requires further chemical modifications at the surface, to make the dots water soluble and also stable in highly saline biological environments. A variety of QD coatings exist that can be tailored for specific biological conditions that the dots are likely to encounter.

One such family of coating materials consists of a three-part modular design containing dihydrolipoic acid (DHLA), to anchor the ligand to the dot; a polyethylene glycol (PEG) derivative, to impart water solubility; and a terminal bio-reactive functional unit (X), such as an amine or carboxyl group, to form chemical bonds that link to dots to selected dyes or specific biological sites of interest. Together, the DHLA-PEG-X assembly provides QDs with the necessary biocompatibility while retaining the outstanding optical properties, including their strong luminescence yields.

QDs in action: Biosensing

The rest of this feature highlights several areas in which QDs are emerging as a compelling alternative to traditional fluorescent dyes. We begin with biosensing, the quest to identify specific biomolecules, pathogens and toxins in both the lab and the clinic.

One approach to using QDs in biosensing is the traditional passive approach—the dots are attached to DNA fragments, petptides or antibodies that bind to a target, and signal a positive response solely through the presence of fluorescence. In this vein, QDs can prove particularly useful for multiplexed assays, in which the presence and intensity of a given, size-determined QD color reflects the presence of one or more targets in the assay.

QDs can particularly excel, however, in more active, real-time biosensing assays, which track changes in the optical properties of the dots that occur owing to chemical or binding processes within the biological target of interest. This approach commonly begins by labeling the target with some type of energy transfer acceptor or donor. Then, when the labeled target interacts with the QD-driven biosensor (acting, in turn, as a donor or acceptor matched to the material used to label the target), there is a small change in the separation between the QD and the labeled receptor.

Both FRET and electron transfer (ET) processes are very sensitive to such changes in separation distance between donor and acceptor, which can in turn produce significant changes in the luminescence intensity of the QD. The magnitude of the luminescence changes can be correlated with the concentrations of the target molecule. Here are a few examples that our labs have studied:

Botulinum Neurotoxin A.

Clostridial botulinum neurotoxins, and especially the serotype A version (BoNT-A), are among the most potent protein toxins, resulting in neuroparalysis, and even in small amounts can contaminate food, making them a potential biothreat. Screening and detection tests for these agents must reflect not only their presence but also whether they are the active, pathogenic zinc-dependent endopeptidase form, which targets specific proteins in cells and leads to eventual paralysis, or a passive, nontoxic variant.

QDs offer a route toward drawing that distinction. In the test, a modular peptide substrate—optimized to contain a central BoNT-A cleavage region that can be used to label an acceptor dye—is tied to an extended linker/spacer sequence and a terminal oligohistidine sequence that attaches to a peptide-bound QD. In the absence of the toxin, photoexcitation of the QD results in efficient FRET to the organic dye, diminishing the luminescence intensity of the QD and increasing the intensity of the dye emission.

Interaction with BoNT-A, however, results in cleavage of the peptide that holds the fluorescent dye, disrupting the FRET process. In turn, the intensity of the QD luminescence increases and the intensity of the dye luminescence decreases. Moreover, this only occurs when the BoNT-A is active—not when it has been denatured or neutralized—and the amount of active BoNT-A can be determined from the ratio of the QD and dye luminescence intensities, within a 350-picomolar detection limit.

figureA sensor for botulinum toxin: Dye-labeled peptides containing specific cleavage sequences recognized by the botulinum neurotoxin A (BoNT-A) protease are assembled on the QD, which then engages in a high rate of FRET. When present, the neurotoxin cleaves the peptide, disrupting the FRET and altering the signal, which is monitored as part of the assay. These sensors can thus report on both the presence of the toxin and its level of activity.

A multiple-protease sensor.

Proteases are enzymes that break down specific bonds within proteins, and preventing certain protease activity in the body has spawned a new class of drug, the protease inhibitors. But studying protease and DNA hybridization activity often involves tracking the coupled action of multiple biomolecules acting in rapid sequence. Meanwhile, traditional QD-biosensing systems that use FRET can be subject to interferences, such as direct excitation of the acceptor in a donor-acceptor pair and interference from auto-fluorescence in the biological sample being studied.

One way to overcome such interference is through time-gated FRET transfer relays. This approach centers around a nanostructure that consists of a long-lived luminescent molecule such as a terbium cryptate complex (Tb), which acts as an efficient energy transfer donor; a QD that can act as both a donor and an acceptor; and a fluorescent dye, which acts as an acceptor.

Illumination of the complex initially excites all of the fluorescent constituents, and also induces autofluorescence in the biological matrix. The fluorescence signals from the dye, the QDs and the autofluorescence have very short lifetimes (ranging from 1 to 20 ns) compared with Tb (which has an excited-state lifetime of around 1 ms), and rapidly decay to the ground state. This allows the initial Tb donor to donate energy to the QD, which acts as a relay, transferring energy to the terminal dye acceptor. The signal detection is delayed by several microseconds, avoiding interference from other fluorescent signals in the mix, and allowing for a clean assay.

Such an assembly can be used to monitor the activity of two distinct proteases. In such an approach, QDs are assembled with peptides specific for each protease and displaying either the initial long-lifetime Tb or the terminal dye acceptor. The presence of a particular protease results in loss of a distinct FRET pathway, correlated to the presence or absence of sensitized emission from one of the components. Such a multiplexed sensing capability is very hard to implement with conventional dyes, as they tend to have overlapping absorption and emission profiles and lifetimes that overlap with that of background autofluorescence.

figureA multi-protease sensor: A time-gated FRET sensor assembled using QDs can monitor the activity of two different proteases. Peptides specific for each protease are labeled with either an initial long-lifetime dye (Tb) or a terminal dye acceptor, and then assembled, via polyhistidine amino acid (His6) coordination, around the central QD. Upon excitation, the first FRET signal occurs from the long-lifetime Tb to the QD; the second occurs from the QD, acting as a relay to the dye. The data are collected in a time-gated mode, where loss of a FRET pathway is correlated to the activity of a specific protease.

QDs in action: Studying the cell

Our labs (and others), along with university and industry collaborators, have also explored how QDs can cast light on activity in the living cell, through use of QDs for cellular labeling and imaging, sensing, cargo delivery and real-time tracking of cell development—at scales ranging from cultured cells to brain tissue slices to whole living organisms. Making these approaches work relies on being able to easily create stable QD-peptide or QD-protein bioconjugates, and to find robust, controlled methods for delivering these conjugates to the living cell itself. A few examples highlight the increasing complexity and sophistication these labs have achieved in using QDs for cell studies.

figureQuantum dots in cellular imaging: (Left) Multicolor labeling of live cells using disparate QD-peptide conjugates, with QDs targeted to the plasma membrane (magenta), early endosomes (red), late endosomes (green) and the cytosol (yellow). The nucleus has been stained with DAPI (blue). (Right) QD-based labeling of the developing chick brain. QD-peptide conjugates (in red) injected into the spinal column at day 3 of the chick embryogenesis localize to the choroid plexus by day 15 of embryonic development. [J. Am. Chem. Soc. 133, 10482 (2011); ACS Chem. Neurosci. 6, 494 (2015)]

One early implementation of QDs with cells focused on getting the dots within the cell itself. This involved finding ways to firmly attach cell-penetrating peptides to the QD surface, which in turn triggers uptake of the QDs into the cell to allow for visualization and labeling. Subsequent work, building on this expertise, has employed these QD-peptide conjugates as a scaffold for carrying protein cargos within the cell; as a two-photon absorption imaging probe; as a probe for targeting discrete subcellular structures; and for the simultaneous uptake and labeling of multiple, disparate cellular locations in live cells.

Another technical advance has been the design of peptide “motifs” that allow QDs and other nanoparticles to be encapsulated within membrane-enclosed packets, or endosomes, within the cell. That encapsulation allows timed release of QDs within the aqueous cell interior (cytosol), making them accessible to the entire cell interior for sensing, drug delivery and other functions. In one example, this approach enabled real-time imaging of pH changes in living cells.

More recently, QD-peptide conjugates have enabled the preferential, specific labeling of neurons in hippocampal brain slices containing multiple cell types, and the multiday tracking of cell migration and partitioning in the developing brain. The large two-photon absorption cross section of QDs has found use in fluorescence-guided imaging and electrophysiological manipulation of deep-tissue neurons—that is, neurons approximately 800 µm below the surface—in the mouse brain. That’s a clear win over the 150-µm tissue depth penetration typical of one-photon imaging using dyes.

As this small sample shows, QDs are coming into their own as versatile, robust tools for imaging, tracking, sensing and cargo delivery. We believe that they’ve emerged as a key enabling research platform—one that will continue to illuminate the intricacies of in vivo imaging and nanoparticle-mediated drug delivery.

A.L. Huston is with the Optical Sciences Division, U.S. Naval Research Laboratory, Washington, D.C., USA. J. B. Delehanty and I. L. Medintz are with the Center for Biomolecular Science and Engineering, U.S. Naval Research Laboratory.

References and Resources

  • J.B. Delehanty et al. “Spatiotemporal multicolor labeling of individual cells using peptide-functionalized quantum dots and mixed delivery techniques,” J. Am. Chem. Soc. 133, 10482 (2011).

  • K.E. Sapsford et al. “Monitoring botulinum neurotoxin A activity with peptide-functionalized quantum dot resonance energy transfer sensors,” ACS Nano 5, 2687 (2011).

  • W.R. Algar et al. “Multiplexed tracking of coupled protease activity using a single color of quantum dot vector and a time-gated Förster resonance energy transfer relay,” Anal. Chem. 84, 10136 (2012).

  • E. Petryayeva et al. “Quantum dots in bioanalysis: A review of applications across various platforms for fluorescence spectroscopy and imaging,” Appl. Spectrosc. 67, 215 (2013).

  • R. Agarwal et al. “Delivery and tracking of quantum dot peptide bioconjugates in an intact developing avian brain,” ACS Chem. Neurosci. 6, 494 (2015).

  • O. Kovtun et al. “Single-quantum-dot tracking reveals altered membrane dynamics of an attention deficit/hyperactivity disorder-derived dopamine transporter coding variant,” ACS Chem. Neurosci. 6, 526 (2015).

  • M.J. Whitley et al. “A mouse-human phase 1 co-clinical trial of a protease-activated fluorescent probe for imaging cancer,” Sci. Transl. Med. 8, 320 (2016).

  • N. Hildebrandt et al. “Energy transfer with semiconductor quantum dot bioconjugates: A versatile platform for biosensing, energy harvesting, and other developing applications,” Chem. Rev. 117, 536 (2017).

  • Y. Park et al. “Medically translatable quantum dots for biosensing and imaging,” J. Photochem. Photobiol. C: Photochem. Rev., doi: 10.1016/j.jphotochemrev.2017.01.002 (2017).