Hot Pursuit

Photo: Amy FrancisTerry Weber, research technician at the Children's Hospital of Philadelphia, uses radioisotopes to label probes for chromosomal mapping studies. Edwin M. Southern's 1975 paper, "Detection of specific sequences among DNA fragments separated by gel electrophoresis," described one of the landmark advances in early molecular biology. A new era was dawning as it suddenly became possible to detect specific sequences among the thousands of fragments produced by a restriction digest of genomic DNA.1 The "Southern blot," as it was almost immediately called, involved transferring denatured, gel-separated DNA fragments to a membrane (generally nitrocellulose), probing that membrane with a fragment of labeled DNA (generally 32P), and then identifying where the label stuck (generally by autoradiography). In this way it was possible, for example, to use a cloned mouse gene to identify the restriction fragment in which the human homologue resided. There are a huge number of variables in such a procedure, including the degree of sequence similarity, the length of the probe, the probe's specific activity, and the amount of target DNA on the membrane. A standard reference manual for molecular techniques concludes that under the "best conditions" a 32P probe labeled to >109 cpm/µg can detect less than 0.1 pg of target DNA in an autoradiographic exposure of several days.2 Several days! In today's high-throughput, fluorescent environment, a competitor may have identified, cloned, and sequenced the gene and be well on the way to writing the paper in that time. Despite a plethora of nonradioactive, less hazardous, occasionally more sensitive modern ways to probe for a DNA sequence, many of these techniques are not much faster than isotopic autoradiography, and many researchers continue to do things the old way. When asked if fluorescent or enzyme-linked approaches have taken a big bite out of the radioactivity market, one major supplier told us "no, surprisingly not," adding that sales of isotopes continue to be strong. Even in a field that values innovation, many people simply don't want to change from what they know best. That's not to say the new methods don't work--they do, and often admirably well, as LabConsumer has reported in several articles. But the surprising entrenchment of the radioisotope as a detection system for many hybridization-based analyses makes these methods worth another look. This review describes some of the traditional labels for hybridization probes and compares them to what is shaping up as the toughest modern competitor: enzyme-linked colorimetric tests. Golden Oldies While many in vivo labeling methods employ 3H or 14C, and some indirect approaches involve isotopes such as 125I, most in vitro labeling protocols call for radioisotopes of phosphorus or sulphur. Traditionally 32P has been the isotope of choice for many applications involving nucleic acids. An a32P-dNTP has the a phosphate (the one closest to the sugar) substituted with 32P. This form is required when the labeled nucleotide is directly incorporated into the probe as with nick translation or sequencing reactions, because the ß and a phosphates are eliminated during chain elongation. A a 32P-dNTP has 32P substituted at the a position (the outermost phosphate from the sugar). While this form is not suitable for incorporation into a newly synthesized DNA chain, the a phosphate can be transferred to a terminal-OH group by polynucleotide kinase, yielding end-labeled probes. These probes usually display lower specific activity, but are ideally suited for position-sensitive techniques such as S1-nuclease mapping. 32P is a strong ß emitter, emitting particles at 1.7 MeV--the highest-energy particles that most life scientists are likely to encounter. This makes 32P very easy to detect but leads to a number of significant safety problems. Safety First For most radioisotopes, low-density materials such as Plexiglas are the shielding of choice. Not only can you see through these materials better than you can see through, for example, lead, but they also reduce Bremsstrahlung radiation, which is emitted when a particle strikes an absorbing medium. However, the high energy of 32P ß particles can lead to significant quantities of relatively high-energy Bremsstrahlung radiation coming through even the recommended 1 cm Plexiglas shielding. Other safety measures are therefore also important. For almost any significant quantity of 32P--say, 1 mCi or more--finger, wrist, and badge dosimeters should be consistently worn, as well as goggles, double gloves, and all that other stuff that leaves researchers feeling like Star Wars extras when working with radioisotopes. Some sources even recommend lead-impregnated rubber gloves, leading to dexterity akin to typing while wearing boxing gloves. The Nuclear Regulatory Commission (NRC) sets limits regarding acceptable exposure in the United States. Overriding its specific guidelines is a policy frequently referred to as ALARA--exposures must be As Low As Reasonably Achievable. Calculating radioactive dose limits is complicated, because the "harm" done by an exposure depends on the nature of the exposure. For example, the harm from ingested radioisotope is likely to be higher than that from an equivalent external exposure. "Rads" and "grays" are units of absorbed dose and are expressed per unit weight of tissue (1 rad is 100 ergs per gram). The "rem" is a unit of dose equivalent that attempts to convert dose to "biological damage" by factoring in how different types of radiation damage different types of tissue. The NRC places an upper limit on the "total effective dose equivalent" of 5 rem per year for an individual who works with radioisotopes (allowable exposures for the general public are much lower). Other limits, such as 15 rem to the eye, and 50 rem as a shallow dose equivalent to the skin, are also defined. These units, as well as other radiation terminology, are explained at www.eh.doe.gov/docs/rcm/gloss.html; many university radiation safety office Web sites discuss exposure limits. For knocking out a quick probe and getting a fairly fast result, 32P is very convenient, but it has some additional problems. First, the short half-life (14.3 days) prevents long-term usage of probes. Second, the high-energy particles scatter more than lower-energy counterparts, causing fuzzy bands that can be a problem in some applications. Finally, the high-energy particles can literally smash apart molecules in your sample in a process called radiolysis, necessitating prompt use (no more than one or two days) of a labeled sample to avoid high backgrounds and artifacts. Lovely Low Energy In 1983, Mark Biggin and colleagues introduced 35S for use in DNA sequencing.3 An 35S atom is substituted for the a phosphorus of a nucleotide triphosphate, and all the enzymology and chemistry works acceptably well with this element at the a position. 35S has a much longer half-life, 87.9 days, eliminating many drawbacks encountered with 32P. More recently, 33P (half life = 25 days) has been used, most notably for DNA sequencing.4 35S and 33P both emit considerably weaker ß particles than 32P, at 0.167 and 0.248 MeV respectively. This limits background and artifacts due to sample radiolysis and produces crisper, sharper bands due to less scattering. One downside of the lower energy is that traditional autoradiographic image amplification schemes, such as intensifying screens, do not work with 35S and 33P. Safety is also less of a concern with the lower-energy radioisotopes. Bremsstrahlung radiation is much lower in both intensity and energy, allowing use of less bulky shielding. Indeed, some sources claim that for moderate quantities of these isotopes, shielding is not necessary. Caveat Emptor When purchasing isotopically labeled nucleotides for sequencing or making probes, there are a few key features to keep in mind. The first is the age of the stock. Nowadays tracking this is easy, because some companies publish their synthesis schedules on the Web. Decay tables (how much is left after specific time intervals) are also available, as well as a variety of Web-based calculators (see, for example, www.curvefit.com). These tables and calculators are useful for figuring out when to purchase isotopes and for estimating how long to expose the film when seeking the perfect, publication-quality autoradiogram. Next is the specific activity of the material. This is a measure of how much activity is present in a given chemical quantity; for a32P-dATP intended for sequencing or probe generation, this will typically be somewhere in the 2,000-4,000 Ci/mmol range. Historically, isotopically labeled nucleotides were shipped and stored frozen in lead "pigs." Today, many companies advertise with the "green" hook of "lead-free shipping," generally using Plexiglas. The introduction of stabilizers--basically free-radical scavengers that protect the chemical integrity of the sample--allows shipping at ambient temperature and storage generally at 4°C. Using an isotopically labeled nucleotide is very simple: The label is incorporated via a molecule that is identical or very closely related to the normal nucleotide triphosphate, and autoradiographic (or other) imaging takes place without subsequent steps. But when you notice your toes glowing in the dark or when dealing with the radiation safety bureaucracy becomes more painful than you can bear, perhaps it's time to look into a nonisotopic approach to labeling DNA. However, eliminating the need for radioisotopes comes at the cost of increasing the experimental and reagent complexity. In with the New A parallel approach to isotopic labeling is incorporating a fluorophore directly into the DNA probe. While this technique is often successful, the fact that the incorporated nucleotide has a linker and a large fluorescent group attached can lead to enzymatic limitations with consequent low labeling efficiency. Direct fluorescence works admirably well in highly purified circumstances--such as a sequencing ladder on a gel--but the inherent fluorescence in many biological samples and membranes frequently leads to high backgrounds in hybridization experiments. Often, to achieve sensitivities that are comparable to that seen with radiolabels, some form of sandwich assay must be performed. This involves incorporating a reporter group into the probe and subsequently detecting this reporter after hybridization. The first reporter group incorporated into DNA probes was biotin, which is recognized using antibiotin antibodies or avidin (streptavidin).5 However, as biotin is a relatively ubiquitous molecule, high backgrounds are often obtained. Though the biotin-streptavidin system can suffer from nonspecific reactions, the biotin-avidin binding constant is one of the strongest known in nature, allowing relatively stringent approaches. The digoxygenin system was developed to minimize background problems.6 Because modified digoxygenin only occurs in one place in nature (the Digitalis plant), nonspecific reactions are rare. Digoxygenin-labeled probes are recognized using antidigoxygenin antibodies. Fluorescein has also been incorporated into DNA probes, and rather than directly detecting the fluorescein fluorescence, an antifluorescein antibody is used for detection. In each of these cases, the molecule that specifically binds the reporter group (avidin, antidigoxygenin, etc.) is itself linked to an enzyme such as alkaline phosphatase (AP), horseradish peroxidase (HRP), or glucose-6- phosphate dehydrogenase, allowing the enzyme to react with a colored, chemi- or bioluminescent, or fluorogenic substrate.7,8 In some cases, strategies involving direct coupling of AP to the probe (that is, not via a biotin or digoxygenin intermediate) are possible. One of the most common colorimetric reactions involves an AP-catalyzed reaction between BCIP (5-bromo-4-chloro-3-indolyl phosphate) and NBT (nitro-blue tetrazolium salt). In this reaction, the NBT is reduced to form a blue-brown precipitate that remains immobilized on the membrane. Similarly, chemiluminescence detection strategies can be created. AP-mediated dephosphorylation of dioxetane derivatives generate partially stable intermediates that further decay, producing light at 477 nm. Many other strategies have been described, including bioluminescence, fluorescence, and time-gated approaches. Faster reactions are available, such as HRP-catalyzed oxidation of luminol, which produces a chemically excited state of 3-aminophthalate dianion. However, while return to a ground state emits a photon at 428 nm, the reaction is so short lived that enhancers must be used to allow efficient capture on film. Some researchers are clearly obtaining spectacular results with nonisotopic DNA probes, but there are also good reasons to stick with time-tested methods. Perhaps the easiest way to make the choice is based on your environment. In a lab that has traditionally used radioactive techniques, adoption of new nonisotopic technology--which may require learning some new tricks and unlearning some old ones--might be an uphill battle. Whereas when starting a new lab or just introducing labeling/detection into the repertoire, procuring all the safety equipment and approvals required for isotopic work might inspire adoption of nonisotopic methods. If a lab or organization is set up to do hot work, many would argue that isotopic labeling is still the most straightforward--and often the most sensitive--way to generate a probe or sequence ladder, but for now, at least, there seems to be plenty of room for both. S Bob Sinclair (bobsinclair@tech-write-edit-com) is a freelance writer in Salt Lake City, Utah. References 1. E.M. Southern, "Detection of specific sequences among DNA fragments separated by gel electrophoresis," Journal of Molecular Biology, 98:503-17, 1975. 2. T. Maniatis et al., Molecular Cloning, 2nd ed., New York, Cold Spring Harbor Laboratory Press, 1989. 3. M.D. Biggin et al., "Buffer gradient gels and 35S as an aid to rapid DNA sequence determination," Proceedings of the National Academy of Sciences, 80:3963-5, 1983. 4. J. Fan et al., "DNA sequencing with a-33P-labeled ddNTP terminators: a new approach to DNA sequencing with Thermo Sequenase(tm) DNA polymerase," BioTechniques, 21:1132-7, 1996. 5. P.R. Langer et al., "Enzymatic synthesis of biotin-labeled polynucleotides: novel nucleic acid affinity probes," Proceedings of the National Academy of Sciences, 78:6633-7, 1981. 6. C. Kessler et al., "Non-radioactive labeling and detection of nucleic acids. I. A novel DNA labeling and detection system based on digoxigenin: anti-digoxigenin ELISA principle (digoxigenin system)," Biological Chemistry Hoppe-Seyler, 371:917-27, 1990. 7. J.J. Leary et al., "Rapid and sensitive colorimetric method for visualizing biotin-labeled DNA probes hybridized to DNA or RNA immobilized on nitrocellulose: Bio-blots," Proceedings of the National Academy of Sciences, 80:4045-9, 1983. 8. I. Bronstein et al., "Rapid and sensitive detection of DNA in Southern blots with chemiluminescence," BioTechniques, 8:310-4, 1990. Multiplex Madness In the late 1980s, G.M. Church and S. Kieffer-Higgins suggested that large increases in sequencing throughput could be achieved by performing sequencing reactions on pooled samples, immobilizing the size-separated ladders on a membrane, and deconvoluting the sequences using primer-specific hybridizations.1 One pitfall of this idea is that probing, reading, stripping, and reprobing the membrane is very labor intensive, so researchers at the Utah Center for Human Genome Research, led by Robert Weiss, automated the process by creating the Automated Hybridization and Imaging Instrument (AHII) for Southern, northern, western, and slot blots as well as gridded arrays. The core of the AHII, a large-area, small-volume hybridization chamber, comprises two mounted acrylic drums. A puddle of reagents pools in a 5 mm gap between the cylinders at their lowest points. The outer surface of the inner cylinder can accommodate up to 6000 cm2 of nylon membrane, which is continuously rotated through the pool of reagents. The chamber is temperature controlled, and reagents for washing, probing, and stripping the membranes are introduced and removed using computer-controlled solenoid valves. A series of vectors containing unique probe sequences has been constructed, ultimately allowing multiplex-sequence-specific hybridization of a biotin-labeled probe. This probe captures a streptavidin antidigoxygenin alkaline phosphatase conjugate. The fluorogenic substrate Attophos(r) (JBL Scientific, San Luis Obispo, Calif.) was found to give well-defined bands, presumably because the fluorescent hydrolysis product binds relatively tightly to the membrane.2 The AHII incorporates a xenon arc light source and a cooled CCD camera to excite the fluorophore and collect an image of the membrane. The computer controller automatically begins the next strip-wash-reprobe cycle after the image has been collected. Advances in high-throughput, fluorescence-based DNA sequencing instruments have largely outpaced the capabilities of multiplex DNA sequencing. But researchers in the Department of Human Genetics at the University of Utah, led by Mark Leppert, are using the AHII with great success for genotyping a variety of disease families including cancer, epilepsy, and heart disease. For this application, 48-lane gels are directly blotted onto nylon membranes; each lane contains 36 markers, multiplexed to a depth of 18 size-compatible pairs and one internal size marker consisting of an M13mp18 sequence ladder. The membranes are UV crosslinked, mounted in the AHII, and subjected to 20 cycles of probing, imaging, and stripping. While the entire experiment takes around 10 days, it generates approximately 10,800 automated genotypes. --Bob Sinclair References 1. G.M. Church, S. Kieffer-Higgins, "Multiplex DNA sequencing," Science, 240:185-8, 1988. 2. J.L. Cherry et al., "Enzyme-linked fluorescent detection for automated multiplex DNA sequencing," Genomics, 20:68-74, 1994.

Hot Pursuit

Interested in reading more?

Become a Member of