Physicochemical Characterization of Direct Fluorescent Antibody Reagents G. ANN HEBERT, BERTIE PITTMAN, and WILLIAM B. CHERRY Center for Disease Control, Public Health Service, US Department of Health, Education, and Welfare, Atlanta, Georgia 30333, USA
Several measurable characteristics contribute to the quality of a fluorescent antibody (FA) reagent. Performance characteristics such as titer, specificity, nonspecific staining, and stability are the clharacteristics most frequently measured. The plhysicochemical characteristics, which can alter performance, are often neglected. These statistics include labeled protein concentration and composition, fluorochrome-to-protein (F/P) ratio, the presence of unreacted fluorescent material (UFM), the source, concentration, and composition of unlabeled protein, and the amount of specific antibody protein.' The analysis of data and the relationships between performance and physicochemical characteristics are discussed. Materials and Methods PROTEIN.-Total protein concentrations are measured by the biuret method2 with optical density readings at 560 nm.3 This method compares favorably with the micro-Kjeldahl metlhod and does not require sophisticated instrumentation. It is a simple and reliable procedure, free from interference by small amounts of ammonium and sulfate ions and by most of the popular preservatives. FITC.-Fluorescein isothiocyanate (FITC) is determined as protein-bound FITC by absorbance in 0.1 N NaOH at X max for each conjugate near 495 nm.4 The reference standard for FITC determinations is fluorescein diacetate, a commercially available, stable compound that is easily prepared in pure form. For a given spectrophotometer, the extinction coefficient for fluorescein diacetate divided by the extinction coefficient for proThe use of trade names is for identification only and does not constitute endorsement by the Public Health Service or by the US Department of Health, Education, and Welfare.
tein-bound FITC is the constant 1.066. The use of this factor and a standard curve prepared with fluorescein diacetate eliminates the effect of variation in spectrophotometers and relates all measurements to a primary reference standard.4 F/P RATIOS.-The F/P ratios are calculated from the fluorescein and protein measurements and expressed as micrograms of protein-bound FITC per milligram of protein. CASE.-The Beckman Microzone equipment and procedure for cellulose acetate strip electrophoresis (CASE) of human serum proteins5 are used with slight modifications. The standard conditions are 250 v for 20 minutes with a pH 8.6 barbital buffer. As a membrane is removed from the electrophoresis chamber, it is examined under a Wood's light (366 nm) for fluorescence of the separated proteins; a quick sketch is made, indicating the location and relative brightness of each fluorescent band. This is the fluorescence emission profile. After Ponceau S staining to develop the protein bands, the membranes are rinsed in 5% acetic acid until the background is white. Then they are gently blotted to remove excess moisture and placed between dry blotters on a flat surface beneath a glass plate to dry. After they are dried, the uncleared membranes are read on a Beckman Densitometer, model R-110, at the recommended settings for protein profiles. IMMUNOELECTROPHORESIS.-Immunoelectrophoresis (IE) is performed by using 1% Noble agar with pH 8.6 Veronal buffer (0.1 M) under a constant voltage of 250 for 90 minutes. The gel slides are incubated in a humid chamber at 25 C for 24 hours to standardize the development of precipitin lines. Commercial antiserums produced against the whole serum proteins of various A45
A46
J Dent Res Special Issue A
HEBERT, PITTMAN, AND CHERRY
TABLE 1 REPRESENTATIVE PHYSICOCHEMICAL AND PERFORMANCE DATA ON A GROUP OF CONJUGATES Protein
F/P Ratio
Specific Titer
10 mg Protein Titer Factor
Conjugate
(mg/ml)
1 2 3 4 5 6 7
6.2
10
2
1.61
6.1
15
5.9 6.1 2.0 8.3 3.0
20 25 20 20 20
8 64
1.64 1.69 1.64 5.0 1.2 3.33
animal species are diffused against the electrophoresed conjugates to develop the IE patterns. Normal animal serums and conjugates prepared in our laboratory from fractions of these serums are used as control antigens.
Results and Discussion The protein concentration and F/P ratio of a conjugate are essential data that provide a basis for comparison of reagents.1 The specific titer should not be the only criterion of suitability, because titers alone can be misleading, as can protein levels, without F/P ratios, and vice versa. A convenient way to eliminate confusion is to estimate titers per unit of protein by calculating a 10 mg protein factor and applying it to the specific titer determined for the conjugate. This
emission profiles
5
6 4
"1bh
2
J4Rit'
1
1
nl n Pi
done with the representative data shown in Table 1. Conjugates 1 to 6 were prepared from the same batch of immune globulin and conjugate 7 from a later batch of globulin against the same antigen. Conjugates I to 4 (Table I) have about the same protein concentration, but their F/P ratios range from 10 to 25. The specific titers increased with F/P ratios up to a value of 20, which is optimal for this system, because the F/P ratio of 25 was not accompanied by a further increase in titer. Conjugates 5 and 6 have the optimal F/P ratio of 20, but their titers vary because of differences in protein concentrations. Conjugate 7, from a second batch of globulin, has the optimal F/P ratio and low protein, but a high titer. This set of reagents (Table 1) with variable characteristics can be studied by comparing their was
estimated Y globulin F/P ratio 35
*1^it.| 3
64 20 80 64
1
3.2 13.1 108.2 105.0 100.0 96.0 213.1
proteins y
and#
UFM -
30
a.
25
.
20
*
15
V,Ba
+
t;jBa,Alb
-
|35
+
migration of unlabeled proteins Y UFM ae A FIG 1.-Fluorescence emission profiles of FITC conjugates with varying characteristics. UFM, unreacted fluorescent material.
Vol 55 1976
CHARACTERIZATION OF FA REAGENTS
10 mg titers and F/P ratios. Such analysis shows that conjugates 1 and 2 need higher F/P ratios, conjugates 3 to 6 each have about 10 units of titer per unit of protein, and conjugate 7 is superior because it has the most reactivity per unit of protein. CASE is the most informative of the physicochemical procedures. It is a sensitive index of all proteins present in a conjugate, whethier labeled or unlabeled; it allows one to detect the presence of UFM, and it provides an additional means of estimating the F/P ratio. The first data obtained from CASE analysis is the fluorescence emission profile, in whiclh the number, location, and relative brightness of fluorescent bands is recorded and compared with the known migration positions for unlabeled electrophoresed proteins (Fig 1) . Labeling a protein with FITC changes its electrophoretic mobility; the higher the F/P ratio, the faster the protein migrates toward the anode. Gamma globulin is more displaced than albumin by a given amount of bound FITC. Any UFM remaining in a conjugate, or that released by dissociation, will migrate to an anodic position beyond that of a labeled albumin. Conjugates 3 to 6 in Figure 1 are from the same globulin, but they were labeled witlh various amounts of FITC. The labeled y-
A47
globulin with an F/P ratio of 20 is in an anodic f8-globulin position. The higher F/P ratio conjugates are even more anodic and the y-globulin with an F/P ratio of 35 has migrated to an a-globulin position. Conjugate 2 contains more /B- and a-globulins than do conjugates 3 to 6, and conjugates 2 and 5 have a band of UFM. Conjugate 1 is an FITC-labeled whole serum. The albumin h-as combined with the major portion of the FITC, and the weakly labeled y-globulin has migrated slightly cathodic, as though it has an F/P ratio of about 5. Total protein and FITC measurements for the determination of F/P ratios of conjugates 1, 2, and 5 would not be meaningful. Most of the labeled protein in conjugate I is not y-globulin, and some of the FITC in conjugates 2 and 5 is not bound to the globulin. For these three conjugates, the F/P ratio estimated by observing the displaced electrophoretic migration of y-globulin is a more meaningful value. The protein profiles of conjugates after CASE (Fig 2) supply the relative percentages of major classes of proteins present and again reflect the changes in electrical charge caused by labeling with FITC. In addition, protein bands may appear in a region that did not have corresponding bands of fluorescence. Adding unlabeled proteins to conju-
Unlabeled globulin a
Conjugate, F:P 10 =
20
2
8'
30
*I, *
0 .
.
:1*1* :1
:40 %. \ \ '.
*.~ ~ ~ ~ ~ ~ ~ ~1
:1 \
\\
:%~
%%
*.~ ~ ~ ~ ~ ~ ~ ~ ~1
e"
~~_..~. FIG 2.-Effect of FITC labeling on protein electrophoresis profiles.
ratio.
(D F/P. fluorescein-to-protein
A48
J Dent Res Special Issue A
HEBERT, PITTMAN, AND CHERRY
total protein
labeled protein
total fluorochrome
unlabeled protein
bound
non-gamma
gamma
1i
globulin
gamrnma bound
antibody W
free
antibody bound
inappropriate
specific
specific antibody bound
- -
cross-reacting
FIG 3.-Composition and relationships in conjugate analysis.
unusual, but it invalidates the interpretation of measured F/P ratios, and it emphasizes the values of data obtained from both the fluorescence emission profile and the total protein profile of electrophoretic analysis. Figure 3 is a flow diagram of possible conjugate characteristics. Currently, F/P ratios are calculated from the total protein and fluorochrome measurements. The presence of unlabeled protein and UFM can be detected by CASE analysis. If either is present in more than trace amounts, the F/P ratio calculated as indicated previously is invalid, because insofar as possible, the F/P ratio should measure the degree of labeling of the antibody by the fluorochrome. An F/P ratio can be estimated for y-globulin on the basis of observation of its migration after CASE analysis, but unless the antibody is separated and purified, the absolute value of the specific antibody-to-fluorochrome ratio is unknown. gates is not
Many commercial immunofluorescent reacontain unlabeled protein, and frequently this may be from an animal source different from that of the labeled antibody. This unlabeled protein may be a stabilizer, such as albumin or whole serum, or an antibody containing globulin fraction or whole serum designed to help block unwanted cross-reactions. The animal species source of labeled and unlabeled proteins can be identified by IE. The precipitin lines that form during IE of labeled proteins are displaced toward the anode with increasing F/P ratios, just as in CASE analysis. gents
Summary When the data from performance and
physicochemical studies of conjugates are combined for analysis, the performance data and specific titers show a direct relationship to the physicochemical data (Table 2). These reagents were prepared from the same lot of antiserum. The specific titers are very
TABLE 2 ANALYSIS OF PHYSICOCHEMICAL AND PERFORMANCE DATA ON A GROUP OF CONJUGATES Case
Protein
Conjugate (mg/ml) 1
2 3 4
6 4 10 8 8
F/P Ratio
% 'y
24 18 30
75 80 50
20 15 20
10 30
100 30
10 10
y
F/P
Gamma
4+
10 mg -y
(mg/ml)
Titer
Titer
4.5 3.2
64
142
32
100
5.0 8.0 2.4
64
128
32 8
40
32
Vol 55 1976
CHARACTERIZATION OF FA REAGENTS
misleading without the accompanying data (Table 2). The protein concentrations range from 4 to 10 mg/ml, the F/P ratios from 10 to 30, and CASE shows y-globulin to constitute 30 to 100% of the protein. CASE also shows the y-globulin F/P ratios to be only 10 to 20. Using these data, we calculated the concentrations of the -y-globulins and normalized their titers to 10 mg/ml. The value of good fractionation procedures for recovering y-globulin and the desirability of obtaining optimal F/P ratios are reflected in the adjusted titers. Physicochemical characterization of conjugates identifies superior and deficient reagents and frequently reveals the cause of inadequate performance. In this way it serves as a guide for improving reagent quality.
A49
References 1. HEBERT, G.A.; PITTMAN, B.; MCKINNEY, R.M.; and CHERRY, W.B.: The Preparation and Physicochemical Characterization of Fluorescent Antibody Reagents, US Department of Health, Education, and Welfare, Center for Disease Control, Atlanta, Georgia, 1972. 2. GORNALL, A.G.; BARDAWILL, C.J.; and DAVID, M.M.: Determination of Serum Proteins by Means of the Biuret Reaction, J Biol Chem 177: 751-766, 1949. 3. GOLDWASSER, R.A., and SHEPARD, C.C.: Staining of Complement and Modification of Fluorescent Antibody Procedures, J Immun 80: 122-131, 1958. 4. McKINNEY, R.M.; SPILLANE, J.T.; and PEARCE, G.W.: Fluorescein Diacetate as a Reference Color Standard in Fluorescent Antibody Studies, Anal Biochem 9: 474-476, 1964. 5. BECKMAN INSTRUMENTS, INC.: Instruction Manual for Model R-101 Microzone Electrophoresis Cell, Fullerton, Calif, 1965.
Several measurable characteristics contribute to the quality of a fluorescent antibody (FA) reagent. Performance characteristics such as titer, specificity, nonspecific staining, and stability are the clharacteristics most frequently measured. The plhysicochemical characteristics, which can alter performance, are often neglected. These statistics include labeled protein concentration and composition, fluorochrome-to-protein (F/P) ratio, the presence of unreacted fluorescent material (UFM), the source, concentration, and composition of unlabeled protein, and the amount of specific antibody protein.' The analysis of data and the relationships between performance and physicochemical characteristics are discussed. Materials and Methods PROTEIN.-Total protein concentrations are measured by the biuret method2 with optical density readings at 560 nm.3 This method compares favorably with the micro-Kjeldahl metlhod and does not require sophisticated instrumentation. It is a simple and reliable procedure, free from interference by small amounts of ammonium and sulfate ions and by most of the popular preservatives. FITC.-Fluorescein isothiocyanate (FITC) is determined as protein-bound FITC by absorbance in 0.1 N NaOH at X max for each conjugate near 495 nm.4 The reference standard for FITC determinations is fluorescein diacetate, a commercially available, stable compound that is easily prepared in pure form. For a given spectrophotometer, the extinction coefficient for fluorescein diacetate divided by the extinction coefficient for proThe use of trade names is for identification only and does not constitute endorsement by the Public Health Service or by the US Department of Health, Education, and Welfare.
tein-bound FITC is the constant 1.066. The use of this factor and a standard curve prepared with fluorescein diacetate eliminates the effect of variation in spectrophotometers and relates all measurements to a primary reference standard.4 F/P RATIOS.-The F/P ratios are calculated from the fluorescein and protein measurements and expressed as micrograms of protein-bound FITC per milligram of protein. CASE.-The Beckman Microzone equipment and procedure for cellulose acetate strip electrophoresis (CASE) of human serum proteins5 are used with slight modifications. The standard conditions are 250 v for 20 minutes with a pH 8.6 barbital buffer. As a membrane is removed from the electrophoresis chamber, it is examined under a Wood's light (366 nm) for fluorescence of the separated proteins; a quick sketch is made, indicating the location and relative brightness of each fluorescent band. This is the fluorescence emission profile. After Ponceau S staining to develop the protein bands, the membranes are rinsed in 5% acetic acid until the background is white. Then they are gently blotted to remove excess moisture and placed between dry blotters on a flat surface beneath a glass plate to dry. After they are dried, the uncleared membranes are read on a Beckman Densitometer, model R-110, at the recommended settings for protein profiles. IMMUNOELECTROPHORESIS.-Immunoelectrophoresis (IE) is performed by using 1% Noble agar with pH 8.6 Veronal buffer (0.1 M) under a constant voltage of 250 for 90 minutes. The gel slides are incubated in a humid chamber at 25 C for 24 hours to standardize the development of precipitin lines. Commercial antiserums produced against the whole serum proteins of various A45
A46
J Dent Res Special Issue A
HEBERT, PITTMAN, AND CHERRY
TABLE 1 REPRESENTATIVE PHYSICOCHEMICAL AND PERFORMANCE DATA ON A GROUP OF CONJUGATES Protein
F/P Ratio
Specific Titer
10 mg Protein Titer Factor
Conjugate
(mg/ml)
1 2 3 4 5 6 7
6.2
10
2
1.61
6.1
15
5.9 6.1 2.0 8.3 3.0
20 25 20 20 20
8 64
1.64 1.69 1.64 5.0 1.2 3.33
animal species are diffused against the electrophoresed conjugates to develop the IE patterns. Normal animal serums and conjugates prepared in our laboratory from fractions of these serums are used as control antigens.
Results and Discussion The protein concentration and F/P ratio of a conjugate are essential data that provide a basis for comparison of reagents.1 The specific titer should not be the only criterion of suitability, because titers alone can be misleading, as can protein levels, without F/P ratios, and vice versa. A convenient way to eliminate confusion is to estimate titers per unit of protein by calculating a 10 mg protein factor and applying it to the specific titer determined for the conjugate. This
emission profiles
5
6 4
"1bh
2
J4Rit'
1
1
nl n Pi
done with the representative data shown in Table 1. Conjugates 1 to 6 were prepared from the same batch of immune globulin and conjugate 7 from a later batch of globulin against the same antigen. Conjugates I to 4 (Table I) have about the same protein concentration, but their F/P ratios range from 10 to 25. The specific titers increased with F/P ratios up to a value of 20, which is optimal for this system, because the F/P ratio of 25 was not accompanied by a further increase in titer. Conjugates 5 and 6 have the optimal F/P ratio of 20, but their titers vary because of differences in protein concentrations. Conjugate 7, from a second batch of globulin, has the optimal F/P ratio and low protein, but a high titer. This set of reagents (Table 1) with variable characteristics can be studied by comparing their was
estimated Y globulin F/P ratio 35
*1^it.| 3
64 20 80 64
1
3.2 13.1 108.2 105.0 100.0 96.0 213.1
proteins y
and#
UFM -
30
a.
25
.
20
*
15
V,Ba
+
t;jBa,Alb
-
|35
+
migration of unlabeled proteins Y UFM ae A FIG 1.-Fluorescence emission profiles of FITC conjugates with varying characteristics. UFM, unreacted fluorescent material.
Vol 55 1976
CHARACTERIZATION OF FA REAGENTS
10 mg titers and F/P ratios. Such analysis shows that conjugates 1 and 2 need higher F/P ratios, conjugates 3 to 6 each have about 10 units of titer per unit of protein, and conjugate 7 is superior because it has the most reactivity per unit of protein. CASE is the most informative of the physicochemical procedures. It is a sensitive index of all proteins present in a conjugate, whethier labeled or unlabeled; it allows one to detect the presence of UFM, and it provides an additional means of estimating the F/P ratio. The first data obtained from CASE analysis is the fluorescence emission profile, in whiclh the number, location, and relative brightness of fluorescent bands is recorded and compared with the known migration positions for unlabeled electrophoresed proteins (Fig 1) . Labeling a protein with FITC changes its electrophoretic mobility; the higher the F/P ratio, the faster the protein migrates toward the anode. Gamma globulin is more displaced than albumin by a given amount of bound FITC. Any UFM remaining in a conjugate, or that released by dissociation, will migrate to an anodic position beyond that of a labeled albumin. Conjugates 3 to 6 in Figure 1 are from the same globulin, but they were labeled witlh various amounts of FITC. The labeled y-
A47
globulin with an F/P ratio of 20 is in an anodic f8-globulin position. The higher F/P ratio conjugates are even more anodic and the y-globulin with an F/P ratio of 35 has migrated to an a-globulin position. Conjugate 2 contains more /B- and a-globulins than do conjugates 3 to 6, and conjugates 2 and 5 have a band of UFM. Conjugate 1 is an FITC-labeled whole serum. The albumin h-as combined with the major portion of the FITC, and the weakly labeled y-globulin has migrated slightly cathodic, as though it has an F/P ratio of about 5. Total protein and FITC measurements for the determination of F/P ratios of conjugates 1, 2, and 5 would not be meaningful. Most of the labeled protein in conjugate I is not y-globulin, and some of the FITC in conjugates 2 and 5 is not bound to the globulin. For these three conjugates, the F/P ratio estimated by observing the displaced electrophoretic migration of y-globulin is a more meaningful value. The protein profiles of conjugates after CASE (Fig 2) supply the relative percentages of major classes of proteins present and again reflect the changes in electrical charge caused by labeling with FITC. In addition, protein bands may appear in a region that did not have corresponding bands of fluorescence. Adding unlabeled proteins to conju-
Unlabeled globulin a
Conjugate, F:P 10 =
20
2
8'
30
*I, *
0 .
.
:1*1* :1
:40 %. \ \ '.
*.~ ~ ~ ~ ~ ~ ~ ~1
:1 \
\\
:%~
%%
*.~ ~ ~ ~ ~ ~ ~ ~ ~1
e"
~~_..~. FIG 2.-Effect of FITC labeling on protein electrophoresis profiles.
ratio.
(D F/P. fluorescein-to-protein
A48
J Dent Res Special Issue A
HEBERT, PITTMAN, AND CHERRY
total protein
labeled protein
total fluorochrome
unlabeled protein
bound
non-gamma
gamma
1i
globulin
gamrnma bound
antibody W
free
antibody bound
inappropriate
specific
specific antibody bound
- -
cross-reacting
FIG 3.-Composition and relationships in conjugate analysis.
unusual, but it invalidates the interpretation of measured F/P ratios, and it emphasizes the values of data obtained from both the fluorescence emission profile and the total protein profile of electrophoretic analysis. Figure 3 is a flow diagram of possible conjugate characteristics. Currently, F/P ratios are calculated from the total protein and fluorochrome measurements. The presence of unlabeled protein and UFM can be detected by CASE analysis. If either is present in more than trace amounts, the F/P ratio calculated as indicated previously is invalid, because insofar as possible, the F/P ratio should measure the degree of labeling of the antibody by the fluorochrome. An F/P ratio can be estimated for y-globulin on the basis of observation of its migration after CASE analysis, but unless the antibody is separated and purified, the absolute value of the specific antibody-to-fluorochrome ratio is unknown. gates is not
Many commercial immunofluorescent reacontain unlabeled protein, and frequently this may be from an animal source different from that of the labeled antibody. This unlabeled protein may be a stabilizer, such as albumin or whole serum, or an antibody containing globulin fraction or whole serum designed to help block unwanted cross-reactions. The animal species source of labeled and unlabeled proteins can be identified by IE. The precipitin lines that form during IE of labeled proteins are displaced toward the anode with increasing F/P ratios, just as in CASE analysis. gents
Summary When the data from performance and
physicochemical studies of conjugates are combined for analysis, the performance data and specific titers show a direct relationship to the physicochemical data (Table 2). These reagents were prepared from the same lot of antiserum. The specific titers are very
TABLE 2 ANALYSIS OF PHYSICOCHEMICAL AND PERFORMANCE DATA ON A GROUP OF CONJUGATES Case
Protein
Conjugate (mg/ml) 1
2 3 4
6 4 10 8 8
F/P Ratio
% 'y
24 18 30
75 80 50
20 15 20
10 30
100 30
10 10
y
F/P
Gamma
4+
10 mg -y
(mg/ml)
Titer
Titer
4.5 3.2
64
142
32
100
5.0 8.0 2.4
64
128
32 8
40
32
Vol 55 1976
CHARACTERIZATION OF FA REAGENTS
misleading without the accompanying data (Table 2). The protein concentrations range from 4 to 10 mg/ml, the F/P ratios from 10 to 30, and CASE shows y-globulin to constitute 30 to 100% of the protein. CASE also shows the y-globulin F/P ratios to be only 10 to 20. Using these data, we calculated the concentrations of the -y-globulins and normalized their titers to 10 mg/ml. The value of good fractionation procedures for recovering y-globulin and the desirability of obtaining optimal F/P ratios are reflected in the adjusted titers. Physicochemical characterization of conjugates identifies superior and deficient reagents and frequently reveals the cause of inadequate performance. In this way it serves as a guide for improving reagent quality.
A49
References 1. HEBERT, G.A.; PITTMAN, B.; MCKINNEY, R.M.; and CHERRY, W.B.: The Preparation and Physicochemical Characterization of Fluorescent Antibody Reagents, US Department of Health, Education, and Welfare, Center for Disease Control, Atlanta, Georgia, 1972. 2. GORNALL, A.G.; BARDAWILL, C.J.; and DAVID, M.M.: Determination of Serum Proteins by Means of the Biuret Reaction, J Biol Chem 177: 751-766, 1949. 3. GOLDWASSER, R.A., and SHEPARD, C.C.: Staining of Complement and Modification of Fluorescent Antibody Procedures, J Immun 80: 122-131, 1958. 4. McKINNEY, R.M.; SPILLANE, J.T.; and PEARCE, G.W.: Fluorescein Diacetate as a Reference Color Standard in Fluorescent Antibody Studies, Anal Biochem 9: 474-476, 1964. 5. BECKMAN INSTRUMENTS, INC.: Instruction Manual for Model R-101 Microzone Electrophoresis Cell, Fullerton, Calif, 1965.
