Histochemistry

Abstract

Histochemistry—chemistry in the context of biological tissue—is an invaluable set of techniques used to visualize biological structures. This field lies at the interface of organic chemistry, biochemistry, and biology. Integration of these disciplines over the past century has permitted the imaging of cells and tissues using microscopy. Today, by exploiting the unique chemical environments within cells, heterologous expression techniques, and enzymatic activity, histochemical methods can be used to visualize structures in living matter. This review focuses on the labeling techniques and organic fluorophores used in live cells.

The goal of histochemistry is to provide color and contrast to microscopic images. The field uses disparate techniques to accomplish the specific labeling of biological structures. Histochemists pioneered the use of small-molecule cellular stains, labeled molecules such as antibodies, and enzyme-mediated detection and signal amplification (Kiernan 2008). Historically, however, histochemistry has been synonymous with the imaging of fixed (i.e., dead) cells and tissues. The advent of genetic manipulation techniques has greatly expanded histochemical methods to living cells. This review examines the current collection of labeling strategies and discusses the correlating fluorescent dyes that allow biologists to add color to live systems.

Fluorescence and Fluorophores

Fluorescence occurs when a dye absorbs one or more photons, yielding a molecule in an excited state (Valeur 2002; Lakowicz 2006). This excited molecule can relax through a variety of processes, including the emission of a photon of another, typically longer, wavelength. It is this property—emission of a photon at a different wavelength—that allows detection of fluorophores in the presence of billions of other molecules in a biological sample. Thus, fluorescence-based techniques find widespread utility in bioresearch as they provide exquisite sensitivity and contrast in imaging experiments (Lichtman and Conchello 2005).

Fluorescent molecules come in various forms, including small-molecule dyes accessed by organic chemistry (Haugland et al. 2005; Lavis and Raines 2008), the genetically encoded fluorescent proteins (Giepmans et al. 2006; Chudakov et al. 2010), or the inorganic “quantum dots” (Smith A and Nie 2009). Here, we focus on the organic fluorophore classes, which encompass both synthetic small-molecule dyes and the fluorescent proteins. A key principle to understanding organic fluorophores is modularity. Building different substituents in and around the molecular structure of the dye allows tuning of the excitation maxima (λ_ex) and emission maxima (λ_em) across the entire visible spectrum and into both the ultraviolet and infrared (Lavis and Raines 2008). Figure 1 summarizes the strategies used to label cells with organic fluorophores and shows specific examples of different fluorescent dye structures (compounds 1–17).

Figure 1.

Labeling strategies and representative fluorescent labels 1–17. The fluorophoric portions of ligands are colored according to emission maxima of the dye. (A) Molecules with intrinsic binding affinity for native cellular biomolecules and subcellular regions. (B) Genetically encoded fluorescent proteins. (C) Affinity-based tag–ligand systems with a genetically encoded binding element. (D) Enzyme-based “self-labeling” tag–ligand system. (E) Enzyme-based tag–ligand system based on posttranslational modification.

Intrinsic Labeling

Small fluorescent molecules with intrinsic affinity for the target can be used to stain regions of interest (Fig. 1A). This concept harkens back to the genesis of the histochemical discipline. Painstaking application of large collections of colored molecules allowed the discovery of compounds with affinity for different biological structures. The idea of the “magic bullet” put forth by Paul Ehrlich, a noted histochemist and Nobel laureate, was inspired by the affinity of dyes such as methylene blue and trypan red for particular types of cells (Strebhardt and Ullrich 2008). Today, this concept is being revisited as large libraries of fluorescent dyes are generated via diversity-oriented synthesis and screened for utility in live-cell imaging experiments (Vendrell et al. 2010).

An important subset of fluorescent stains includes molecules with intrinsic affinity for nucleic acids. For example, the dibenzimidazole Hoechst 33342 (1) binds the minor groove of DNA with high affinity and exhibits a large increase in fluorescence intensity upon binding with λ_ex/λ_em = 350/461 nm. The increased membrane permeability of Hoechst 33342 makes it more useful for live-cell applications than other nucleic acid stains (Crissman and Hirons 1994). To allow DNA staining of live cells with a fluorophore emitting in the red region of the spectrum, a heavily modified anthraquinone derivative, 1,5-bis[[2-(dimethylamino)ethyl]amino]-4,8-dihydroxyanthracene-9,10-dione (“DRAQ5”, 2; λ_ex/λ_em = 646/681 nm), can be used to stain live cells (Smith PJ et al. 2000).

In addition to the screening approach, rational design of small-molecule fluorophores can furnish stains for native cellular structures. The labeling of organelles can be accomplished by exploiting different chemical environments, including electric potential, pH, and lipophilicity within a specific subcellular region (Haugland et al. 2005). For example, positively charged molecules can label mitochondria; this selective concentration is due to the potential across the mitochondrial inner membrane. Thus, dyes bearing a fixed positive charge such as rhodamine 123 (3; λ_ex/λ_em = 507/529 nm) have found use as fluorescent markers for these important organelles (Scaduto and Grotyohann 1999). Another example of a subcellular labeling probe is the boron dipyrromethene (BODIPY) derivative 4 (i.e., LysoTracker Green), which exibits λ_ex = 504 nm and λ_em = 511 nm (Haugland et al. 2005). By appending a weakly basic tertiary amino group to the fluorophore structure, the compound accumulates in acidic vesicles such as lysosomes. BODIPY dyes are particularly useful for intrinsic staining, as they exhibit sharp excitation and emission peaks, are relatively insensitive to the chemical environment, and can be modified to extend the λ_ex into the red region of the spectrum (Ulrich et al. 2008).

Lipophilic regions inside cells can be stained with non-polar dyes. Compounds that exhibit increased fluorescence in hydrophobic environments are particularly advantageous for this application. One example of an environmentally sensitive dye is the oxazine-based Nile Red (5; λ_ex/λ_em = 532/636 nm). This molecule is used to image lipid droplets in tissue (Fowler and Greenspan 1985) and in super-resolution localization microscopy to visualize membranes (Sharonov and Hochstrasser 2006). Another class of useful lipophilic dyes are the cyanine derivatives such as 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine (DiI; 6; λ_ex/λ_em = 549/565 nm). These compounds bear long alkyl chains that embed in lipid bilayers. This cyanine dye platform can be modified in several ways to fine-tune the spectral and chemical properties of these dyes and generate numerous imaging probes for membranes.

In addition to exploiting environmental differences to label subcellular locations, appending binding motifs for native proteins allows specific staining of cellular structures such as the endoplasmic reticulum (Zünkler et al. 2004). An interesting and useful extension of this strategy involves fashioning a suicide substrate for an enzyme where the enzymatic reaction elicits a large change in fluorescence. This fluorogenic reaction yields a labeled protein. Although used extensively for proteome investigations (Sadaghiani et al. 2007), these “activity-based probes” also can enable the imaging of native enzymes inside live cells (Blum et al. 2007).

Genetically Encoded Fluorophores

Perhaps the most straightforward method to label specific proteins in living cells is to use the genetically encoded fluorescent proteins (FPs; Fig. 1B) (Chudakov et al. 2010). These compounds are essentially small-molecule fluorophores encased in a protein shell. The chromophores form spontaneously, requiring only molecular oxygen. The ability to express many proteins as fusions with a fluorescent protein tag allows a myriad of biological imaging experiments. In addition to protein labeling, these fluorophores can be directed to various parts of the cell such as the nucleus, endoplasmic reticulum, cellular membrane, or mitochondria by expression of variants with different localization motifs (Chudakov et al. 2010).

The relative ease of labeling with fluorescent proteins has led to the development of numerous variants displaying a broad range of colors and properties. Encased in the protein structure, the green fluorescent protein (GFP) chromophore (7) exhibits λ_ex = 490 nm and λ_em = 509 nm (Tsien 1998). Mutagenesis can change the makeup of the fluorophore or surrounding pocket to modify the spectral characteristics of the dye. For example, replacement of Tyr66 residue in GFP with a His residue results in the blue fluorescent protein chromophore (8; λ_ex/λ_em = 383/448 nm) (Ai et al. 2007).

Genetically encoded fluorophores with larger chromophore structures have been discovered and further engineered to open up imaging in the red region of the visible spectrum using fluorescent proteins (Piatkevich and Verkhusha 2009). Such imaging agents are known collectively (and imprecisely) as red fluorescent proteins (RFPs). The chromophore of DsRed (9), a common RFP, is extended by further oxidation of adjoining residues, giving rise to the longer λ_ex = 558 nm and λ_em = 583 nm (Gross et al. 2000).

Despite the unquestionable utility of fluorescent proteins, this labeling strategy has some limitations. The relatively large size of the FP can disrupt the activity and location of the resulting fusion protein. These systems can use only the 20 proteogenic amino acids and thus suffer from a finite (albeit still growing) number of accessible fluorophore structures. Most FP systems express a protein variant with a single color. Given the larger diversity of small-molecule fluorescent dyes, it would be useful to combine the precise genetic targeting available with FPs and the structural flexibility of small-molecule fluorophores. Incorporation of non-natural amino acids is one possible route to this goal (Young and Schultz 2010), although the ultimate utility in mammalian systems might be limited (Uttamapinant et al. 2010). Alternatively, intersectional strategies that use a genetically expressed “tag” that interacts specifically with a small-molecule “ligand” can be advantageous. Appending the ligand motif to a fluorescent dye can allow the labeling of a single tagged protein with fluorophores of different colors or properties, depending on experimental requirements.

Affinity-Based Tag–Ligand Systems

Numerous genetically encoded tags have been discovered or engineered (O’Hare et al. 2007). The first type involves expression of a protein bearing an affinity tag that can bind tightly and specifically to a small-molecule ligand (Fig. 1C). Examples of tags include small-molecule binding proteins (Marks et al. 2004; Krusemark and Belshaw 2007; Wombacher et al. 2010), peptides (Kelly et al. 2007), antibody fragments (Szent-Gyorgyi et al. 2007), streptavidin for biotinylated probes (Wu et al. 2000), tetra-aspartate to bind Zn²⁺ complexes (Nonaka et al. 2010), and polyhistidine to bind nitrilotriacetic acid (NTA) motifs via a Ni²⁺ bridge (Goldsmith et al. 2006). Further engineering can introduce a nucleophilic residue on the tag near the ligand-binding site that can react with a corresponding electrophilic motif installed in the small-molecule ligand to form a covalent conjugate (Krusemark and Belshaw 2007; Gallagher et al. 2009; Nonaka et al. 2010).

A classic small-molecule binding tag is the tetracysteine-containing helix described by Griffin and coworkers (1998). The fluorescein arsenical helix binder (FlAsH, 10), based on the well-known fluorophore fluorescein, binds preferentially to this small tetracysteine motif installed on a protein. Upon binding, the compound displays increased fluorescence with the conjugate exhibiting λ_ex = 508 nm and λ_em = 530 nm. A more recent tag–ligand system involves the triarylmethane dye malachite green (11) and a cognate single-chain antibody fragment (scFv). This pair has been used in live-cell imaging experiments (Szent-Gyorgyi et al. 2007) where the bound fluorophore exhibits λ_ex = 635 nm and λ_em = 656 nm. Importantly, the binding of 11 to the scFv gives a ~10⁴ enhancement in fluorescence, thereby diminishing background signal from the population of free dye.

Endogenous molecules can also serve as ligands, negating the need for delivery of exogenous compounds. The engineering of bacteriophytochrome proteins that bind the naturally occurring biliverdin 12 has yielded a fluorophore with λ_ex = 684 nm and λ_em = 708 (Shu et al. 2009). Because of the obstinate chromophore structure of fluorescent proteins, variants exhibiting spectral properties in the infrared have been lacking. This affinity strategy overcomes this limitation of existing FPs.

Self-Labeling Tag–Ligand Systems

In addition to labeling based on expressed affinity tags, exogenous enzymes can be used to label structures inside cells. Exploiting enzymatic activity can increase the specificity and speed of labeling in bioconjugation experiments (Kalia and Raines 2010). Enzymes in which the mechanism uses a covalent intermediate can be engineered to stop at that step of the reaction. Small proteins have been modified to serve as tags with substrate moiety–fluorophore conjugates as ligands (Fig. 1D). Examples of these “selflabeling tags” include a halogenase (HaloTag) (Los et al. 2008; Watkins et al. 2009) and variants of the O ⁶alkylguanine-DNA alkyltransferase (SNAPTag and CLIPTag) (Maurel et al. 2010). In addition, substrates for β-galactosidase have been constructed that label the protein with a concomitant change in the spectral properties of the dye (Komatsu et al. 2006). Such fluorogenic systems eliminate the need for washing unbound ligand out of the cell.

These self-labeling tags have been engineered to accommodate disparate dye structures. Derivatives of the pH-insensitive 7-aminocoumarin have established utility as the basis for fluorescent ligands for many labeling strategies, including the HaloTag (e.g., compound 13; λ_ex/λ_em = 351/430 nm). The photostable rhodamines (Beija et al. 2009) are also used extensively in these systems. The tetramethylrhodamine CLIPTag 14 (λ_ex/λ_em = 554/580) can label proteins inside cells (Gautier et al. 2008). Cyanine derivatives, such as the SNAPTag ligand 15 (λ_ex/λ_em = 660/673 nm), are also employed. To use cyanines as labels, however, the lipophilic structure of these dyes demands addition of sulfonate groups to improve solubility (Mujumdar et al. 1996). Currently, these polar functional groups render the molecule cell impermeant, limiting the utility of these long-wavelength ligands to cell surface labeling experiments.

Ligase-Based Tag–Ligand Systems

Enzymatic activity can be applied another way by exploiting ligases that facilitate posttranslational modifications on particular protein motifs (Fig. 1E). Here the expressed peptide tag can be significantly smaller than fluorescent proteins, affinity tags, and enzyme variants while facilitating the formation of a specific covalent bond. Examples include a formyl glycine-generating enzyme to install aldehydes (Carrico et al. 2007), the acyltransferase protein (ACPTag) (George et al. 2004), and biotin ligase from Escherichia coli that can attach biotin and biotin mimics to a short peptide sequence on the N-terminus of proteins (Chen et al. 2005).

It must be noted that these ligase-based strategies require two proteins and efficacious delivery of the small-molecule label. Therefore, the utility of these labeling systems is often confined to the exterior of cells. For example, coenzyme A derivative 16 serves as a substrate for phosphopantetheinyl transferase and allows the attachment of fluorescein (λ_ex/λ_em = 490/514 nm) to a serine residue on a short peptide sequence (PCPTag) installed on cell surface proteins (Yin et al. 2005). In a recent development that allows facile labeling inside living cells, lipoic acid ligase can be modified to attach small-molecule probes such as 7-hydroxycoumarin derivative 17 (λ_ex/λ_em = 360/449 nm) directly to a 13-mer peptide tag (Uttamapinant et al. 2010).

Future Directions

Numerous systems are available to allow the extension of histochemical techniques inside living cells. The modularity of organic fluorophores allows multiple colors to be used to label structures for imaging experiments. Each labeling strategy has distinct advantages and weaknesses. Molecules with intrinsic affinity do not require genetic manipulation and could be useful for imaging in living organisms where genetic access is limited or unethical (e.g., humans). Fluorescent proteins will likely remain the dominant imaging modality in cell biology and model organisms, as addition of exogenous compounds is not required. Limitations of fluorescent protein labels can be circumvented by intersectional strategies employing both small-molecule fluorophore ligands and genetically encoded tags. These strategies do, however, add complexity to the experiment as they require expression of one (or more) proteins, delivery of the small-molecule ligand, and removal of any excess dye.

The established histochemical methods were developed and refined through collaborative efforts between chemists, biochemists, and biologists. Further improvements in live-cell labeling strategies will also require interdisciplinary research. Development of efficient synthetic methods for constructing fluorescent molecules along with improved screens is required to discover new stains with intrinsic affinity for biological structures. New fluorescent proteins with lower toxicity and enhanced optical properties are needed to perform long-term imaging experiments in vivo. Novel strategies for the efficient delivery and targeting of small molecules are also essential to expand the utility of labeling in complex cellular environments. Collectively, these exciting and useful tools will allow increasingly intricate histochemical experiments to illuminate biological processes in living cells.