Part III.Laboratory methods for detection of CGTPs

Contents - Previous - Next

Enzyme-linked immunosorbent assay (ELISA) for citrus pathogens
Detection of plant viruses and viroids by molecular hybridization
Viroid purification and characterization
Immunosorbent electron microscopy (ISEM) and antibody coating
Isolation and analysis of CTV dsRNA from citrus bark
Isolation and culture of Spiroplasma citri
Detection of citrus tristeza virus inclusion bodies using azure a staining and in situ immunofluorescence


Enzyme-linked immunosorbent assay (ELISA) for citrus pathogens

S M. Garnsey
United States Department of Agriculture

and M. Cambra
IVIA, Valencia, Spain

This is a guide to the use of enzyme-linked immunosorbent assay (ELISA) for detection of citrus pathogens. It describes several common variations of ELISA. Some background information is presented to help the user understand the technique and make modifications to this highly flexible procedure for specific applications. Information on the selection of techniques, on preparation of samples for testing, and on the basic steps of the ELISA protocol is provided. The supplies, reagents and equipment needed are indicated, and some specific examples are shown.

ELISA has become a standard procedure for the detection of several citrus pathogens, especially citrus tristeza virus (CTV). Properly used, ELISA is a sensitive, accurate and rapid detection method. It is especially effective where large numbers of samples must be assayed, where results are needed rapidly, and where suitable indicator plants and/or greenhouse facilities are not available. ELISA has been developed for CTV, satsuma dwarf virus (SDV), citrus variegation virus (CVV), citrus leaf rugose virus (CLRV), citrus mosaic virus (CiMoV), Spiroplasma citri, Xanthomonas campestris pv. citri, and Phoma tracheiphilla. ELISA assays are being developed for the greening organism and for citrus tatterleaf virus.

ELISA is simple and can be carried out by most people after brief training and some practice. As with any indexing procedure, some experience is necessary to use ELISA accurately and confidently. New users should consult several of the excellent general references on ELISA (e.g. Clark and Bar-Joseph,1984; Clark, Lister and Bar-Joseph,1988; Sanchez-Vizcaino and CambraAlvarez,1987) which provide additional details on theory and application. It is very useful to visit a laboratory where ELISA is practiced in order to observe the procedure and to study it under the guidance of an experienced user. Begin with a well-known system, and study the effects of adjusting reactant concentrations and test conditions.

Extensive training and background in serology and immunology are not essential to use ELISA, but it is necessary to understand some basic concepts. ELISA is a serological technique and, in common with other serological procedures, it is based on the concept that many proteins are antigenic when injected into animals and that the immunized animal will form antibodies to them.

These antibodies can be obtained from the serum of the immunized animal and will react specifically with the antigen to which they were formed. A primary requirement to begin ELISA is a useful source of antibody to the pathogen to be detected. This, in turn, means that to obtain such antibodies an antigen specific to the pathogen must be identified and purified sufficiently to produce the needed antibodies. Antigen purification and antibody production are beyond the scope of this section, but information on these topics is contained in some of the references cited (Clark, Lister and Bar Joseph,1988; Van Regenmortel,1982).

Also fundamental to ELISA is the concept that various enzymes can be bound to antibody molecules to form a conjugated molecule that has both enzymatic activity and is also serologically active. Since enzymes are highly active and can be detected at low concentrations, they are effective labels. Enzyme-labelled antibodies can be detected when they are exposed to a substrate which enzymes can change. Normally, a substrate that changes colour as a result of the enzyme action is used. The amount or rate of colour change can then be used to measure the amount of antibody present. Enzyme labels provide a sensitivity similar to that of radioactive labels and have several important advantages: they are stable, the cost is low, they are safe to use and can be used successfully without sophisticated equipment.

The enzyme label may be attached directly to the antibody used to detect the antigen in question (the detecting antibody). This is called a direct assay, of which the highly popular double antibody sandwich technique described below and illustrated in Figure 1 39a is a good example. The label may also be used indirectly. In this case, the detecting antibody is not labelled, but rather the label is attached to a second antibody specific to the detecting antibody. Antibodies of one species are antigenic when injected into an animal of a second species. For example, rabbit immunoglobulins can be injected into another animal such as a goat to create a goat anti-rabbit antiserum. These goat anti-rabbit antibodies are useful to detect antibodies from rabbits which were originally prepared to detect another antigen.

Indirect assays are more sensitive and also avoid the need to prepare a conjugate to each antibody used. Several forms of indirect assay are described in the following section and are also illustrated in Figures 139b and 139d. The relative advantages of direct and indirect systems are discussed in the following section.

Several other molecular interactions are frequently used in conjunction with ELISA, either to purify immunoglobulins or to amplify reactions and increase sensitivity. Protein A is a cell wall component of the bacterium Staphylococcus aureus and has the unique characteristic of binding to the immunoglobulin protein of many mammalian species. The binding site is on the Fc region of the immunoglobulin and not on the antigen binding site. Protein A is frequently used to purify antibodies by affinity chromatography. It can also be conjugated with enzymes and used in assays to detect immunoglobulins.

A second important system is the biotin/avidin system. Biotin, a small vitamin, has a very high affinity for avidin, a 68 000 molecular-weight glycoprotein. Antibodies and enzymes can be conjugated with several molecules of biotin to form a "biotinylated" molecule. Each avidin molecule has four binding sites for biotin. This multiplying interaction has been exploited in several ways to amplify the number of enzyme molecules associated with each antigen-bound antibody and thereby increase sensitivity. One example is illustrated in Figure 139c.

Another fundamental concept for ELISA is that proteins such as antibodies and virus coat proteins will adsorb strongly to the surface of certain plastics such as polystyrene and polyvinyl-chlorides. Protein binding also occurs to some forms of cellulose nitrate. These materials are frequently referred to as "immunosorbents" or the "solid phase" in ELISA protocols. The protein binding to immunosorbent materials is not specific and is not a serological reaction such as occurs between antigen and antibody molecules. If a mixture of antibodies is exposed to an immunosorbent plastic, all will bind. Similarly, when a crude extract from a diseased plant is placed in an ELISA plate, both proteins of the pathogen and proteins of the host present in the extract will be bound.

Binding either the antibody or the antigen component of a serological system to a solid phase is very useful because the bound component can subsequently be used to probe complex mixtures of potential reactants. Only those which are serologically related will be trapped. All non-reactive components can then be removed by washing and do not interfere with subsequent steps. For example, when an extract from a virus-infected plant is placed in the wells of a microtitre plate coated with antibodies to that virus, virus antigens in the extract will be bound to the trapping antibody and all non-related proteins will be removed by the subsequent washing step.

Undesired adsorption of antibody or antigen proteins to the plastic can be avoided by using non-ionic detergents such as Tween 20 in incubating solutions or by adding an excess of a non-specific protein to block all sites not occupied by the desired serological component. For example, the buffer used to coat plates with trapping antibody does not contain Tween 20, but Tween 20 is incorporated in subsequent steps where any non-specific binding of other proteins should be avoided.

Immunoblotting procedures are not specifically discussed in this section. However, much of the information and the general concepts presented are directly applicable to immunoblotting procedures. The main differences are that the solid phase for immunoblotting is usually cellulose nitrate, the substrate used to measure presence of the antigen-antibody-enzyme complex is different, and incubation conditions may be somewhat modified.


Numerous variations of the ELISA procedure can be devised (Clark, Lister and Bar-Joseph,1988; Engvall and Pesce,1978; Jones and Torrance,1986; Koenig and Paul,1982; Maggio,1980). The selection depends on the sensitivity, specificity and convenience required; the presence of interfering factors; and the type and activities of the antisera available. The basic steps for four commonly used variations of ELISA are outlined here and illustrated in Figure 139. In three of the variations, Figures 139a to 139c, the solid-phase (ELISA plate) is coated with antibody to the antigen to be detected. This antibody, identified as trapping antibody (TA), then traps its corresponding antigen (identified as V) from suspension or solution. In the fourth variation, the antigen (V) is trapped directly on the solid phase (Figure 139d) and detected with its specific antibody.

Double antibody sandwich

The inexperienced user should start with the double antibody sandwich (DAS) where possible. This has been the most commonly used form of ELISA for plant virus detection since its description by Clark and Adams (1977). The components of DAS are illustrated in Figure 139a. The immunosorbent surface is the wells in a plastic microtitre plate designed for ELISA as shown in Figure 140. A dilute solution of unlabelled antibody is added to the wells of the plate, and the antibody adsorbed on the plastic becomes the trapping antibody (TA) as illustrated in Figure 139a. After washing to remove any excess antibody (Figure 154), the sample (antigen) is added as shown in Figure 142. Antigens (V in Figure 139) specific to the bound trapping antibody attach themselves to it, but other proteins remain in solution and are removed by washing. The antigen attached to the trapping antibody is detected by adding a labelled antibody (LA in Figure 139a) specific to the antigen (Figure 143). The label is the enzyme (E) previously conjugated to the antibody. When substrate specific to the enzyme is added in the final step (Figure 144). a colour develops as a result of enzyme action (Figures 145 and 146). The amount of colour and rate of development are correlated to the amount of labelled antibody bound to the antigen which had been trapped by the antibody attached to the plate.

DAS can be done with a single good quality polyclonal antiserum. The immunoglobulins present are partially purified, and one portion is saved for use as trapping antibody while another is conjugated to an enzyme. Alkaline phosphatase is commonly used as the enzyme and the conjugation can be done in the presence of dilute glutaraldehyde (Clark, Lister and Bar-Joseph,1988). The antibodies for coating and detection do not have to come from the same source, e.g. monoclonal antibodies could be used for coating, and a polyclonal antiserum used to prepare the enzyme-labelled antibody.

Double antibody sandwich indirect

DAS can be converted to an indirect procedure (DAS-I), which is illustrated in Figure 139b. The first two steps are the same as in DAS. However, the antigen bound to the trapping antibody is detected by an unlabelled intermediate antibody (IA in Figure 139b) which is specific to the same antigen, but is from an animal species different from the one used to prepare the trapping antibody. For example, if the trapping antibody was prepared in rabbits, the detecting or intermediate antibody (IA in Figure 139b) could be from a mouse or a chicken. The unlabelled IA which attaches to the antigen is detected by an enzyme-labelled antibody (LA) specific to the IA. Because the IA is from a species different from the TA, the LA binds only to the IA and no non-specific binding of the LA to the TA occurs. The amount of LA is measured by adding substrate and measuring colour change as in DAS.

DAS-I ELISA involves an additional step (Figure 139b) but is more sensitive and also allows use of a commercially prepared enzymelabelled antibody to the IA. A single LA can also be used for multiple virus detection systems. In addition, the intermediate antibody does not have to be purified and is needed in only a limited quantity. If the intermediate antibody is highly specific, e.g. most monoclonals, then a highly specific antiserum is not required for coating. The major problem is that antibodies to the same antigen must be prepared in two different animals. If the trapping and the intermediate antibodies are from the same species, the labelled antibody used to detect the intermediate antibody will also bind to the trapping antibody and result in a non-specific response.

A system has been devised to carry out DAS-I using a single antiserum (Adams and Barbara,1982; Clark, Lister and Bar-Joseph, 1988). To do this, the antibodies are treated with the enzyme pepsin to remove the Fc portion of the molecule. The remaining F(ab')2 fragment still has the antigen binding sites and will bind to the immunosorbent, but will not bind to protein A. The F(ab')2 fragments are used as trapping "antibody" and the whole antibody is used as the intermediate antibody. Enzyme-conjugated protein A is then used instead of a labelled antibody to detect the intermediate antibody. It does not react to the trapping "antibody" because the Fc region has been removed. The DAS-I procedure can be further modified to amplify the reaction achieved. This is commonly done using a biotin-avidin interaction where the labelled antibody is biotinylated to react with avidin molecules conjugated to multiple enzyme molecules, as illustrated in Figure 139c. Different types of amplification are possible and special kits may be purchased to perform them. Users should be aware of the possibility to increase sensitivity by amplification where the additional sensitivity is needed, but regular procedures should be tested before amplified tests are attempted.

Plate-trapped antigen

Another basic approach to ELISA is the plate trapped antigen procedure (Figure 139d). The approach is to trap the antigen (V) on the plastic surface, then react the trapped antigen with an unlabelled intermediate antibody (IA) specific to it. The IA is then detected as in DAS-I using an enzyme-labelled antibody (LA) specific to the IA. This procedure, called plate-trapped antigen indirect (PTA-I) ELISA, is relatively simple and involves no advance purification of antisera or conjugate preparation if a commercially prepared enzyme-labelled antibody to the unlabelled IA is used. The PTA-I procedure is usually less sensitive than DAS or DAS-I for use with crude plant extracts, and may not be effective when antigen concentration in the sample is low. Since binding to the plate is non-specific, there is competition between the target antigen and other proteins present in the extract for the available binding sites on the plate. Plate-trapped antigen tests can be conducted as a direct assay using an enzymelabelled antibody to the antigen, but sensitivity is even lower than for the indirect method, and the conjugate must still be prepared. Amplification procedures as described for DAS-I can also be used for the PTA-I procedure to increase sensitivity.

The specific steps and schedules for these types of ELISA are described in Schedules 1-3.


Selection of appropriate samples for testing is critical. Although ELISA is a sensitive procedure, reliable results may not be obtained if poor samples are tested. Virus titre in citrus tissue often varies markedly, and thousandfold differences in antigen concentration can occur over a relatively short period. Virus concentrations are generally highest in young, expanding flush tissues. They decrease rapidly as tissues mature under hot-weather conditions and more slowly under cool conditions. Avoid sampling old, mature tissue during the summer months in hot climates unless preliminary testing indicates that reliable samples can be taken. If the virus or pathogen is phloem-limited, such as tristeza, greening, or stubborn, then the tissue sample collected must contain phloem tissue. Older bark tissue can be sampled if the cambium is active, but generally it is less reliable than young shoot flush, bark or young leaf midribs (Figures 148 and 149). The button area tissue and pedicel bark from fruit (Figure 149) are good phloem sources on bearing trees. Young root tips may be useful under some conditions.

A composite sample from several sites on the tree should be collected: normally three-to-five locations per tree are sampled. Increase sampling if the pathogen is irregularly distributed or when trying to monitor a recent infection.

From time to time, it is not convenient to test samples immediately after collection. Fresh tissue can normally be stored for at least seven to ten days at 4C when kept in a plastic bag or sealed container. Tissue can also be dried over a desiccant (or air-dried in dry climates) and then stored over fresh desiccant. Dry samples are convenient for long-distance shipping. One convenient method for field collecting in remote locations is to place 0.25 g samples in gelatin capsules, code them, and place them in a sealed container with a desiccant. Some desiccants having a colour indicator for activity are on the market.

Samples can also be stored frozen at -20C or below for extended periods either as unprocessed fresh tissue, or as diced tissue placed in extraction buffer and frozen in the grinding tube (Figure 150). The sample should not be ground prior to freezing because fresh extracts often lose much activity when frozen. Frozen samples should not be stored in an automatically defrosting freezer. Extracts can be stored for long periods when freeze-dried, which is a good way to store a source of consistent reference (control) samples. Always test a storage method with the specific pathogen under study in order to prove its effectiveness.


Numerous buffers and different additives have been used for extraction of tissue samples with different virus-host systems (Bar-Joseph and Garnsey,1981; Clark,1981; Clark and Bar-Joseph,1984; Clark, Lister and Bar-Joseph,1988; McLaughlin et al,1981). Fortunately, extraction problems have been rare with citrus viruses, and phosphate-buffered saline (PBS) or 0.05 M Tris, pH 7.5 to 8.0 without any additives usually give good results for sandwich assays. Additives such as polyvinylpyrrolidone, EDTA and DIECA are generally unnecessary and may actually reduce reaction efficiency. Test the effect of additives before using them routinely. For plate-trapped antigen procedures, try extraction of the sample in carbonate coating buffer pH 9. 6, or in 0.05M Tris, pH 8.0. Do not use Tween 20 in the extraction buffer for samples to be platetrapped.

Normally, the ratio of buffer to sample tissue should be at least 1: 10. Higher concentrations of tissue may actually reduce reaction rates and make sample preparation more difficult. Good samples of CTV-infected tissue can usually be diluted up to five hundredfold and still give strong reactions.

There are many ways to grind samples. Pestle and mortar are fine for small numbers of tender samples. Addition of an abrasive, such as fine sand or carborundum, to the sample or powdering the tissue in liquid nitrogen makes grinding easier. A dispersion homogenizer (Figure 151) equipped with a 10-25 mm diameter shaft is a good choice when large numbers of samples are to be processed. A 2-10 ml sample can be rapidly ground in a test tube or centrifuge tube of suitable diameter and length with this type of homogenizer. Fibrous tissue, such as bark and leaf midribs, should be cut into short pieces (25 mm) prior to grinding or the shaft will become clogged with fibre and must be cleaned between samples. Two rinses of the grinder shaft in 5001 000 ml clean water are usually adequate (Figure 152). Run the homogenizer briefly in each rinse solution.

Chill samples prior to grinding to offset heating during the grinding process. It is normally not necessary to keep the sample on ice during grinding, unless unusually long grinding is required and the sample becomes warm to the touch. Frequent users of dispersion homogenizers should wear earplugs to protect their hearing, and homogenizers should be isolated.

If necessary, samples can be prepared with very minimal equipment. When virus concentration is high, extensive disruption of the sample tissue is usually unnecessary. One method is simply to place a small piece of tender tissue directly in buffer in the well of an ELISA plate with forceps and then squeeze it to release the cell contents. Tissue can also be crushed in a small plastic bag using a mallet or smooth stone and the extract moved by pipette into the test plate.

Samples containing a lot of debris after extraction can be difficult to extract by pipette. Remedies include centrifugation of the sample to pellet the debris, or filtering the sample through a coarse filter such as cheesecloth or glass wool (Figure 153). Cutting off a portion of the tapered tip of plastic pipettes creates a wider orifice and is often quick and effective. It is frequently quicker to rinse a repeating pipettor between samples than to change tips so only a limited number of tips need to be modified.


Contents - Previous - Next