Medical Device Link.

Originally Published IVD Technology July/August 2001

Membrane immobilization of nucleic acids: Part 1: Substrates

Kevin D. Jones

A variety of membranes have characteristics that can make them useful immobilization substrates for molecular diagnostic applications.

Kevin D. Jones, PhD, is the manager for diagnostic technology at Whatman International Ltd. (Maidstone, Kent, UK). He can be reached via e-mail at kjones@whatman.co.uk.

With the mapping of the human genome, the future of many areas of diagnostics, including rapid testing, appears to be inextricably linked to nucleic acids. The College of American Pathologists designates nucleic acid diagnostics as belonging to the field of molecular pathology, which is defined as covering the diverse areas of disease predisposition, therapeutic suitability, and organism identification. The current market for nucleic acid diagnostics is perhaps more than $2 billion, with a predicted growth rate of better than 20% over the next three years. By 2003, the projected market for nucleic acid diagnostics will approach 20% of all diagnostic tests performed.

Figure 1. Double-stranded DNA immobilized on a glass-fiber matrix. Photo Courtesy Whatman International Ltd.

A significant challenge facing the fledgling nucleic acid diagnostic industry is the transformation of laboratory tests now performed by researchers into commercial diagnostics that can be used by nonspecialist technicians in doctors' offices, and even by consumers at home. This challenge looms especially large for developers of point-of-care diagnostic assays.

A nucleic acid test requires that a series of processes be completed: sample collection, sample preparation, amplification (if required), and detection. Discussion of all aspects of these processes is beyond the scope of a single article. The limited aim of this three-part series therefore is to highlight the materials and methods for nucleic acid immobilization that are available to developers of rapid assays, with a particular focus on how various target materials can be immobilized on a membrane substrate (see Figure 1).

Figure 2. Structures of DNA (a); RNA (b); and peptide nucleic acid (PNA) (c).

Four types of nucleic acid probes can be immobilized onto a solid phase for a rapid assay: large sections of DNA, small DNA (including cDNA), RNA, and peptide nucleic acid (see Figure 2). The methods for immobilizing these different molecules have some similarities, but differences in molecular structure and size make the methods substantially different. This series is restricted to outlining typical procedures to immobilize to a variety of substrates the shorter, more commonly used DNA probes. (The immobilization of large sections of DNA onto solid phases is no longer routinely used for diagnostic tests.).

The term membrane encompasses a wide range of potential substrates that can be used for the immobilization of nucleic acids. They include traditional cast membranes such as nitrocellulose and nylon, and also innovations such as ceramic or track-etched membranes and the types of substrates used for microarrays. Whatever the substrate, there are only a limited number of ways that nucleic acid can be attached. Most common are physical adsorptive processes or chemical linking processes (including ultraviolet [UV] or covalent methods).¹ In addition, a nucleic acid probe can be assembled on the membrane, a novel technique developed by Affymetrix Inc. (Santa Clara, CA).

Nucleic acid probes can be linked to membranes in several ways. In order to appreciate the potential of the various attachment techniques, it is helpful to be familiar with the characteristics of available membrane materials and their application in nucleic acid testing. Future installments in this series will review the generic linking methods, with a concentration on covalent linkage, and consider their potential manufacturability in process-scale production.

Cast Membranes

The original membrane used for nucleic acid immobilization was nitrocellulose, selected by E. M. Southern for his Southern blotting method.² This technique involves transferring DNA fragments, produced by restriction endonuclease digestion of DNA, from an electrophoresis gel to a nitrocellulose membrane. A great deal of published material describes methods to achieve nucleic acid binding to nitrocellulose, usually by means of physical adsorption. Researchers have also developed several newer, more-specific binding techniques that extend the utility of this substrate, including those using aptamers or a poly-T tail.

The principal advantages of nitrocellulose are its ready availability and familiarity. However, it is fragile (unless cast on a polyester backing) and has a lower binding capacity than some other membranes. The use of nitrocellulose membranes with radioactive methods of signal detection is well established, but other materials may offer advantages in applications where the sensitivity of the detection technique is lower than for radiolabels.

Recently, nylon has been promoted as a substrate for nucleic acid binding owing to its greater physical strength and binding capacity, and the wider range of available surface chemistries offered, which optimizes nucleic acid attachment. Immobilization on nylon membranes can be performed via physical adsorption, UV cross-linking, or chemical activation. Immobilization on nylon has been demonstrated to be more durable during repeated probe stripping than immobilization on nitrocellulose.^3,4 Nylon membranes have also been used in methods to detect DNA by colorimetry, fluorometry, and chemiluminescence.⁵

The high background typically observed with nylon membranes is their principal disadvantage. This may be due to a nonspecific binding of the sample or detection system, or to some natural property of the membrane. Nitrocellulose has a lower binding capacity and is weaker than nylon, but it has far lower background for most detection systems.

Other membrane materials have also been used successfully, though they have never achieved significant market penetration. They offer some advantages for particular applications, but their general performance has not been able to match that of nylon or nitrocellulose. These alternative membranes have been investigated for their nucleic acid binding capability. The most widely used are made of charge-modified polyvinylidenedifloride, which binds nucleic acids through interaction of the positively charged surface groups with the phosphate backbone of the acid.

With any membrane type, nucleic acid binding protocols must be optimized for each particular membrane, because various manufacturers will use different formulations and manufacturing processes. This can result in a variety of surface chemistries, for instance, which can lead to different binding and subsequent detection characteristics. Membrane manufacturers often make available protocols to help develop the optimal system for each membrane. The balance of properties between nitrocellulose and nylon means that both products have found widespread use for nucleic acid immobilization within the IVD industry.

Microparticles

In many areas of molecular biology, microparticles are widely used as attachment substrates for nucleic acids. Such particles offer a range of surface chemistries to suit the linking technique required (e.g., silica, oligo dT, agarose, and latex with amine). Often, as in the case of a direct-incubation or direct-reading assay, the microparticles replace the membrane.

However, in some applications the use of microparticles can complement the membrane; the nucleic acid is attached to a large particle that becomes entrapped in the membrane structure or to smaller particles that are trapped by well-defined membranes (e.g., track-etched or ceramic membranes). This use of microparticles is well known in the rapid immunoassay industry, as indicated by the "boulders in a stream" approach with cast membranes and by particle-capture immunoassays defined-pore products such as track-etched membranes (TEMs).^6,7

The use of microparticles as the adsorption phase offers significant advantages over direct attachment of the nucleic acid to a membrane. Perhaps foremost, methods for linking nucleic acids to microparticle surfaces have been well characterized, and all reactions can be carried out in solution phase under a wide range of conditions. With membranes, on the other hand, the chemistries or reaction conditions that can be employed without damaging the membrane are often limited. Also, once the nucleic acid is immobilized, the particle can be passivated easily and without much increased likelihood of introducing nonspecific interactions. By contrast, the activation and subsequent surface passivation of membranes can introduce undesirable chemistries across the entire membrane surface. Such surface chemistries can interfere with the sample or detection system.

The initial immobilization of a nucleic acid on a microparticle would therefore enable a more-controlled procedure than is possible with immobilization directly on a membrane. Subsequent entrapment of the microparticle in either a traditional cast membrane or a TEM would then allow the assay to be run and detection to occur.

Arrays

As array technologies become increasingly important in diagnostics, their price is falling to levels that could make them competitive with other techniques rather than being limited to research laboratories. A number of different techniques and substrates for manufacturing such arrays now exist. No longer limited to the original flat-film arrays such as those assembled on glass slides or plastic films, the technology has been extended to membrane-based, three-dimensional, and flow-through arrays.

Traditional membranes such as nylon and nitrocellulose have been used in the production of macroarrays, but their use in microarrays has been limited because of the problem of spot resolution. Because these membranes exhibit lateral wicking characteristics, the label tends to spread from the point of application. This has been a limitation in the production of very high density arrays. However, some microarray systems use a membrane or porous substance that has been cast onto the surface of a glass slide. And there are certain membranes that have no lateral wicking characteristics. They are typically TEMs or anodic membranes, such as Anopore, whose use in DNA microarrays has been demonstrated by Pamgene (Den Bosch, The Netherlands). Such membranes are not limited with respect to spot resolution.

The use of novel membranes in microarrays presents interesting opportunities. All membranes offer much greater surface area than flat sheets. In traditional cast membranes, the ratio of the within-pore surface area to the top surface area of the membrane is about 200:1.

The most extreme example of this characteristic is an anodically oxidized membrane, one in which almost the entire membrane structure consists of pores, with very little material forming the walls. Anopore has a surface area ratio of approximately 500:1. This means that the amount of probe that can be immobilized per unit area is 500 times greater than would be possible on a glass surface. The benefit is either high sensitivity (more immobilized capture probe results in a higher level of material captured) or a higher-density array (the same amount of capture probe can be immobilized in a smaller area).⁸

Figure 3. Scanning electron microphotograph of nitrocellulose, a typical cast membrane (a); Cyclopore, a track-etched membrane (b); and Anopore, a ceramic membrane (c).

These membranes could be envisaged as the ultimate multiwell plates, each pore becoming in effect a well in which an individual reaction can be completed. Anopore provides a pore density of 10⁸ pores per square centimeter.

If the application technology and bioinformatics could handle the data, probes for the whole human genome could be laid down in an area covered by a single blood spot from a finger prick. With improved detection technology and the capacity for single-copy detection, the potential of these membranes is enormous (see Figure 3).

Novel Membranes

As mentioned above, some types of membrane have no lateral-wicking capacity. Another key difference between traditional cast membranes and these novel materials is that pore size distribution for the newer-style membranes is extremely well controlled. Without lateral wicking, the material applied can travel only through the transverse pores. Anodically oxidized alumina (e.g., Anopore) will bind nucleic acids predominantly under chaotropic conditions. The process is similar to that widely used in binding nucleic acids to silica in preparation columns. TEMs, by contrast, have no direct binding capacity. However, their utility in the trapping of microspheres enables a wide range of surface chemistries and binding techniques to be used.

While not strictly speaking a membrane, powdered silica or alumina on a support matrix has been used to bind nucleic acids (e.g., coated microarray slides). These materials will bind nucleic acids very strongly under chaotropic conditions. Through silane modification, covalent linkage would be possible. However, the relative fragility of these ceramic materials may limit their potential for use in certain applications.

Conclusion

No single substrate material is best for all applications. Each offers some advantages. A manufacturer's choice of substrate will most likely be determined by the requirements of the application.

Whatever the nature of the substrate, there are only a few attachment techniques, including covalent linkage (for example, through silanation of the substrate), photolinkage, physical adsorption, and assembly of the probe on the substrate. The manner in which the nucleic acid is bound is critical for the performance of the assay. Binding via the bases is potentially detrimental to performance, as the bases are responsible for any hybridization events that occur. If the bases are immobilized on the membrane surface, or constrained in such a way that free movement during hybridization is restricted, the binding is not optimal. The effect is far more significant for shorter probes than for long sequences. Thus, any linking techniques that result in attaching bases to the membrane surface (e.g., UV cross-linking) or large-scale attachment of the probe to a flat surface (e.g., traditional physical adsorption) may be less than ideal because of the reduction in sensitivity and specificity of the hybridization. In extreme cases, the immobilization technique can render the immobilized nucleic acid useless. The optimal form of binding, therefore, is to end-link the nucleic acid to the solid phase, ideally using a controlled-size spacer to ensure that the nucleic acid is free to interact with the sample.

The next installment of this series presents an overview of immobilization techniques available to developers of nucleic acid tests. The advantages and disadvantages of each are reviewed, with reference to various membrane substrates and assay performance potential.

1. R Lutgarde et al., "Critical Evaluation of Membrane Supports for Use in Quantitative Hybridizations," Applied Environmental Microbiology 62 (1996): 300–303.

2. EM Southern, "Detection of Specific Sequences among DNA Fragments Separated by Gel Electrophoresis," Journal of Molecular Biology 98 (1975): 503–517.

3. A Dubitsky and J Defiglia, "Stripping of Digoxigenin-Labeled Probes from Nylon Membranes," BioTechniques 19, no. 2 (1995): 210–212.

4. K Noppinger et al., "Evaluation of DNA Probe Removal from Nylon Membrane," BioTechniques 13, no. 4 (1992): 572–575.

5. A Dubitsky, "DNA Arrays on Nylon Membranes," Lab Automation (abstract), June 1998.

6. LB Bangs, "The Latest Applications of Microspheres," in The Latex Course 2000 (Indianapolis: Bangs Labs, 2000).

7. H Christensen et al., "Three Highly Sensitive 'Bedside' Serum and Urine Tests for Pregnancy Compared," Clinical Chemistry 36 (1990): 1686–1688.

8. KD Jones et al., "Comparative Study of Glass Slides versus Microporous Ceramic Slides for Nucleic Acid Arrays" (poster presented at the 32nd Oak Ridge Conference, Boston, May 5–6, 2000).

Copyright ©2001 IVD Technology