Comparison of two methods of virus detection by immunosorbent electron microscopy (ISEM) using protein A

~o~rna~of

Viro~og~~al~ethods,

4 (1982)

155

155-166

Elsevier Biomedical Press

COMPARISON OF TWO METHODS OF VIRUS DETECTION BY IMMUNOSORBENT ELECTRON MICROSCOPY (ISEM) USING PROTEIN A

A. NICOLAIEFF”, D. KATZ2 and M.H.V. VAN REGENMORTEL’ ‘Institut de Biologic ~o~~~ulaire et Ceilukzire, 15 rue Descartes, S~asbourg,

France; and ‘Department

of Viroiog)i, Israel Institute for 3~o~o~‘~alReseareh, P.O.B. 19, Ness-Ziona, fsrael

(Accepted 3 December 1981)

The efficiency of two methods of immunosorbent electron microscopy has been compared. The first method consists in trapping virus particles by means of Staphylococcus aureus cells coated with a layer of viral antibodies; the second method consists in trapping virus particles on electron microscope grids coated with specific antibody. A suspension containing lo7 antibody-coated bacteria trapped the total number of virions present in 1 ml of a 500 ngjml virus preparation; the ceils were then fully saturated with virions, and approximately 100 virions (of 30 nm diameter) were visible at the periphery of each cell. When 10’ cells/ml were used the minimum virus concentration needed to see one virion at the cell periphery was 5 ng/ml. Antl~ody-bats grids allowed for the detection of approx~ately the same quantity of virus, but the data obtained with this method were more reproducible and suitable for quantitation.

immunosorbent

electron microscopy

protein A

serological trapping

antibody-coated

bacteria

INTRODUCTION

The detection of virus particles by immunoelectron vantages of the high specificity of serological reactions electron

microscope

obse~ations

(Almeida

microscopy combines the adwith the extreme sensitivity of

and Waterson,

1969; Doane and Anderson,

1977; Milne and Luisoni, 1977; Van Regemnortel, 1981,1982). A positive serological reaction can be visualized in the electron microscope by three different phenomena that may occur together or separately, i.e. clumping of virus particles, coating of particles with a layer of antibody, and trapping of particles on antibody-coated grids. The serological trapping phenomenon was first described by Derrick (1973), who showed that virus particles adhered much better to grids coated with specific antibody than to untreated grids. Techniques based on the trapping of virions on grids are referred to as immunosorbent electron microscopy (ISEM), which is a sister term to ELBA (enzyme-linked immunosorbent assay) (Roberts and Harrison, 1979; Roberts et al., 1981). The ISEM technique has been used extensively for the identi~cation and quantitative analysis of plant and human viruses (Mime and Luisoni, 1977; Lesemann et al., 1980; Nicolaieff and 0166-0934/82/0000-0000/$02.75

@Elsevier Biomedical Press

1.56

Van Regenmortel, sensitivity

1980; Nicolaieff

of the technique

A from StaphyZococcus Cough,

et al., 1980; Van Regenmortel,

1981). Recently,

the

was further increased by first coating the grids with protein

aureus before coating them with specific antiserum

1979; Gough and Shukla, 1980; Lesemann

(Shukla and

and Paul, 1980; Obert et al., 1981).

Presumably, when a grid is pretreated with protein A, more antibody molecules are bound to the grid (via their Fc portions) than in the absence of protein A (Langone, 1978). The increase in sensitivity of virus detection due to the intermediate layer of protein A is most apparent when low dilutions of antiserum (10-l to 10m2) and high virus concentrations are used. Not all authors agree regarding the advantage of a layer of protein A for detecting low concentrations of virus (Gough and Shukla, 1980; Lesemann and Paul, 1980; Nicolaieff et al., 1980). It has been suggested that in conditions where particle numbers are very few, the advantage theoretically provided by the layer of protein A is no longer operative, because it is then the number of particles rather than the number of antibody molecules that is limiting (Mime, 1980). Recently, it was shown that protein A-containing S. aureus cells could be coated with a layer of viral antibodies by incubating them with specific viral antiserum and that such antibody-coated bacteria were then able to ‘fish out’ the corresponding virus particles from a suspension (Katz et al., 1980). Virions attached at the periphery of the cells are easily visible by electron microscopy. Several other applications of protein A-containing staphylococci in virus serology have been reviewed by Jonsson and Nordenfelt (1979). In the present report we compare the trapping efficiency of such antibody-coated bacteria with the trapping achieved on electron microscope grids coated with viral antibody via an intermediate layer of protein A. MATEKIALSANDMETHODS Viruses and antisera Tomato bushy stunt (TBSV), turnip yellow mosaic (TYMV) and tobacco mosaic (TMV) viruses were purified as described previously (Nicola~eff and Van Regenmortel, 1980). Antisera to these viruses were those used in earlier work (Nicolaieff Regenmortel, 1980).

and Van

Trapping of virus on antibody-coated S. aweus cells S. agrees cells, obtained as described previously (Katz et al., 1980), were sensitized with viral antibodies by mixing 1 ml of a bacterial suspension containing IO* cells with 1 ml of virus antiserum diluted 1 : 10,000. After 30-60 min incubation at 37°C (with stirring), the cells were centrifuged and resuspended in 1 ml of phosphate-buffered saline {PBS), pH 7.0. These antibody-coated cells were used for trapping the corresponding viruses by three different procedures. method I: The suspension of antibody-coated ceils (1 ml) was mixed with 1 ml of a

157

purified virus preparation

and incubated

for 1 h at 37°C. The cells were centrifuged

resuspended

in 0.1 ml of PBS. Drops of this suspension

of parafilm,

and microscopic

and

(50 ~1) were placed on a sheet

grids coated with a film of formvar-carbon

were floated for

5 min on the drops. The grids were rinsed by placing them on drops of PBS and were stained with 1% phosphotungstate, pH 7.0. Grids were examined in a Siemens 101 electron microscope at a magnification of either X4000 or X8000. Several squares of the grids were photographed on 70 mm film. Method 2: After incubation of the bacteria with the virus preparation, the cells were not centrifuged, but were allowed to settle by gravitation for 4.5 h on the surface of grids coated with formvar-carbon. The grids were then rinsed with drops of PBS before staining with phosphotungstate. Method 3: The antibody-coated cells were first allowed to settle for 4.5 h on the surface of the grid. These grids with bacteria attached to them were then floated on the surface of drops of purified virus suspensions. The grids were then examined as described above. Trapping of virus on antibody-coated grids Grids coated with a film of formvar-carbon were floated on 50 ~1 drops of protein A solution (l-10 pg/ml PBS) for S-10 min, and were then placed on drops of PBS (Nicolai’eff et al., 1980). Thereafter, the grids were floated for 5 min on 50 ~1 drops of antiserum diluted in PBS. After an intermediate rinse on a drop of buffer, the antibodycoated grids were left for 1 h on drops of purified virus preparations. After the trapping reaction, the grids were rinsed by placing them on a series of drops of distilled water. Virions were visualized by positive staining with 1% uranyl acetate in 45% ethanol (Nicola’ieff and Van Regenmortel,

1980; Nicolaieff et al., 1980).

RESULTS

Antibody-coated bacteria: method I S. aweus cells (lo8 cells/ml) were incubated with antiserum to TMV, TBSV and TYMV, and were then tested for their ability to trap the corresponding viruses. When the cells were incubated with a 100 ng/ml TMV preparation, about 20% of them became covered with a few particles. Results obtained with isometric viruses, TBSV and TYMV, are summarized in Tables 1 and 2. Under the conditions used (IO’ cells incubated with 1 ml of virus suspension for 1 h) the lowest limit of detection was 20-50 ng of virus; at this virus concentration the bacteria showed an average of about one virion trapped per cell. Approximately 10 cells were present on each square of the 300-mesh microscope grids. The specificity of the method was tested by using bacteria that were coated with antisera against TYMV and TBSV for trapping the homologous and heterologous viruses

158

TABLE

1

Trapping

of TBSV particles

on the surface

of S. aureus (lo8

cells/ml)

sensitized

with a low4 dilution

of TBSV antiserum

Virus concentration

1000

500

100

4

20

(nglml)

No. of cells with trapped Virions

per cell

TABLE

2

Trapping

40/42

(95%)

18/21

(86%)

11/20

(45%)

4/18

(22%)

2122 (9%)

virions 19.8

of TYMV particles

10.8

1.3

of S. aureus (10’

on the surface

0.8

cells/ml)

sensitized

0.2

with a lo*

dilution

of TYMV antiserum

Virus concentration

500

100

8/8 (100%)

11/12

24

11

4

20

0.8

(ng/ml)

No. of cells with trapped Virions

per cell

TABLE

3

Specificity lo-’

of trapping

dilutions

Bacteria

(91%)

lo/28

(35%)

4/20

(20%)

o/19

(0%)

virions 0.6

of TBSV and TYMV on the surface

of homologous

and heterologous

of S. aureus (lo8

cells/ml)

0

sensitized

with

antisera

Viruses

coated

0.2

(500 ng/nl)

with antiserum against

TBSV

TYMV

19192

5121

TBSV No. of cells with trapped Virions

virions

per cell

(85%)

(24%)

8.3

0.43

2130

24132

TYMV No. of cells with trapped Virions

per cell

virions

(6.6%)

(75%)

0.10

4.3

(Table 3). Both the number of virions per cell and the percentage of bacteria with virions on their surface were significantly higher in the homologous systems.

159

Antibody-coated bacteria: method 2 In

an attempt to increase the number of virions trapped per cell, increasing dilutions

of

a preparation of antibody-coated bacteria were incubated for 60 min with 500 ng/ml of TYMV. The bacteria were then allowed to settle on a microscope grid for 4.5 h. The number of cells visible on one square of the 300-mesh grids decreased with increasing dilutions of the bacterial preparation. In the case of 4,8, 16, 32 and 64 million cells/ml, the number of bacteria visible on each square was 1,3,8, 16 and 35. When 4-8X IO6 cells/ml were used to trap the virus, the surface of the bacteria appeared to be completely covered with virions (Figs. 1 and 2). The number of virions/cell was somewhat less when 16X lo6 cells/ml were used, but some bacteria still showed an uninterrupted layer of virus particles (Fig. 3). When 32-64X106 cells/ml were used, the number of virions visible at the cell periphery was about lo-60/tell (Fig. 4). Since this number varied considerably from cell to cell, no attempt was made to quantify precisely the trapping efficiency of the antibody-coated bacteria by method 2. However, it is clear that, compared to method 1, the second method is much more sensitive. When 2X lo6 cells/ml were used to trap TYMV at

Fig. 1. S. aureus (TYMV)

particles.

cell, treated

with

specific

One ml of the incubation

500 ng virus. The stain was 1% phosphotungstate,

virus antiserum, mixture

and subsequently

contained

pH 7. X 30,000.

4 X 10’

incubated

antibody-coated

with

virus

cells and

160

Fig. 2. S. ~ureu~ cells. One ml of the incubation

mixture

contained

8 X IO” cells and 500 ng TYMV.

x 1 S,OOO.

a concentration of 50 ng/ml by method 2, the number of trapped virions visible at the cell periphery was 50-lOO/cell (Fig. 5). Quantitative measurements were made difficult by the fact that considerable agglutination of the cells took place. Antibody-coated bacteria: method 3 Attempts were also made to use antibody-coated bacteria fixed to a microscope grid as a trapping agent for the corresponding virus. This method of virus detection was found to be less sensitive than method 2. After 30 min incubation with TYMV used at a concentration of 1 pg/ml, the bacteria were covered with 3040 virions/cell. In addition, many virus particles seemed to have become detached from the bacterial surface and were surrounded by a layer of material, presumably made up of protein A and antibody (Fig. 6). napping on antibody-coated grids The trapping

of TYMV particles

on antibody-coated

grids was measured in the pre-

161

4

Fig. 3. Antibody-coated

S. aureus

cell. One ml of the incubation

mixture

contained

16

X

10’

cells

mixture

contained

6.4

X

10’

cells

and 500 ng TYMV. x 30,000. Fig. 4.

Antibody-coated

and 500 ng TYMV.

x

S. aureus cell. One ml of the incubation 30,000.

162

6

163

sence and absence of a primary layer of protein A on the grid. In the absence of protein A, the lowest level of detection

(1-2

particles per micrograph)

was reached when a pre-

paration of 5 &ml virus was deposited on the grid for 60 min (Table 4). When the trapping time was increased to 18 h, the number of particles increased 4-fold. In the case of grids pretreated with 1 fig/ml protein A, the number of trapped particles increased 2-S fold. The trapping of TYMV particles on grids was not affected by the presence of plant sap constituents in the preparation. As shown in Table 5, there was only a small difference in the number of particles trapped when the virus was diluted with buffer or with plant sap. DISCUSSION

It is interesting to compare the trapping efficiency of a suspension of bacteria coated with antibody with that of a microscope grid coated in the same way. From the particle weight of TYMV (5.6X106) it can be calculated that: S.6X106 g/l virus contain 6X1O23 virions/l, and that 1 pg/ml virus contains approximately 1X10” virions/mi.

TABLE 4 Trapping of TYMV particles on electron microscope grids coated with protein A and specificantiserum Dilution of serum

10>+

10:” 10-4 10-s No serum

1 pg/ml protein A

No protein A Virus concentration 5 50 500

(ng/ml) 500

Antiserum

Normal serum

Antiserum

0 0 6

5 11 S

2a 1

100 41 20 65

1100 1000 1100 450 22s

5

so

110 160 90 10 10s

so0

10 &g/ml protein A 500

50

Normal serum

Antiserum

Normal serum

25 25 55 10 0

0 0

1500 0 1600 0 1000 4 50 550

so0

500

800 950 125 20 1

The numbers represent mean counts of particles on 2-20 micrographs (68 X 73 mm) taken at an instrumental magnification of X8000. For numbers lower than 10, the given count represents the average from 20 micrographs.

Fig. 5. Antibody-coated S. a~(ret(s cell. One ml of the incubation and SO ng TYMV. x 30,000.

mixture contained 2

x

lo6 cells

Fig. 6. Antibody-coated S. a~etls cell showing trapped TYMV particles. in this experiment the antibody-coated celts were fixed to the formvar carbon-coated grid prior to incubation with 1 #g/ml virus suspension. X 15,000.

164

TABLE

5

Trapping Protein

of TYMV particles A

W/ml)

Antiserum

in plant sap on grids coated diluted

lo-”

Dilution

of sap from

infected

plants

--

with protein

A and specific

Antiserum

antiserum

diluted

10e3

500 ng/ml purified

virus

in sap from healthy plants

diluted

-__ in buffer

lo-*

lo+

10-s

1O-4

0 0.1

5600 7300

3600 5000

220 800

750 _

1000 ._

1

6500

4800

950

1500

900

10

5900

4000

300

880

950

From

the average diameter

of S. augers cells (1 m)

it can be calculated

that the

surface of one cell is equal to: 4 rr (500 nm)'= 3.14X lo6 nm2 . This surface will be able to trap a number of particles approximately equal to the ratio of cell surface to surface section of one virion, i.e. 3.14X106 nm*/3.14 (1.5 nm)’ = 4..5X103 virions. On the circumference of a fully saturated bacterial cell, the number of virions bound should be approximately 2X3.14X500 nm/30 nm = 103 virions. This corresponds in fact to the maximum number of particles observed at the periphery of S. aureus cells (see Figs. 3 and 5). It also follows that IO7 cells have a maximum trapping capacity of 4.5X103X107 = 4.5X1O’o virions, which is about the number of virions (5X10i”) present in 1 ml of a 500 ng/ml virus preparation. The results obtained with antibody-coated bacteria, using method 2 (see Figs. 2--3) showed that IO7 cells incubated with 500 ng virus became fully covered with virions. It is clear, therefore, that this procedure is an extremely efficient way of trapping all the particles present in the virus preparation. Results obtained with method 1 showed the same efficiency of trapping (see Table l), since IO8 cells incubated with 500 ng virus trapped 11 particles at the periphery of each cell (i.e. 480 virions/ceIl). This means that the lo8 cells trapped 4.8X10’” virions, which again corresponds to the total number of particles present in 1 ml of a 500 ngfml virus preparation. It is possible to evaluate the maximum attainable sensitivity in the following way. It was necessary to use at least lo7 bacteria/ml to obtain a sufficient number of cells on the grid, and with this number of cells it was possible to trap 500 ng of virus. Since the periphery of the cells was then covered with approximately 100 virions, it follows that the minimum amount of virus needed to see one virion at the cell periphery is 5 ng. The results we obtained with method 1 using lo8 cells (Tables 1-2) confirm that about 50 ng virus is needed in order to visualize about one virion at the cell periphery. It is obvious that the sensitivity could be increased if a technique were available by which all the

165

virions on the cell surface could be seen. Although done by using the scanning electron

microscope

it has been shown that this could be

(Katz et al., 19X0), this method

is still

not practical for routine use. As far as antibody-coated grids are concerned, the following parameters can be calculated. A microscope grid of 3 mm diameter has a surface of 3.14X(1.5 nm)2X10’2 = 1X lOlo 7x 1o12 ml?, and when fully saturated with virus, it could accommodate particles of TYMV of 700 nm2 cross section. At a magnification of 8000X, the visible area on the photograph is 68X73 nm X IO”/ 8000’ = 8X IO7 nm2, and the total number of such areas on each grid is 7X1012/8X 10’ = 90,000. This means that, if all available &ions deposited on the grid became trapped, about 10’ particles would have to be deposited for IO3 virions to appear on each micrograph. Since we found that 3X lo9 particles (in 50 ~1 of a 500 ng/ml virus suspension) had to be deposited on the grid for lo3 virions to appear on each micrograph (see Table 4), the trapping efficiency is 30 times less than the maximum theoretical value. However, when we used 5 1.11drops of virus suspension (instead of 50 ~1) we found (results not shown) that approximately the same number of virions became trapped on the grids. On the basis of these results, it may be concluded that the trapping efficiency of antibody-coated bacteria is only slightly superior to that of antibody-coated grids. The antibody-coated grids made it possible to detect approximately the same quantity of virus as the antibody-moated bacteria. When a 50 ~1 drop containing 5 ng/ml of TYMV (5X10’ virions) was deposited on a grid coated with a lo* dilution of antiserum, the average number of virions per micrograph was l-5 (see Table 4). In this case the presence of an intermediate layer of protein A (1 @g/ml) on the grid increased somewhat the number of trapped virions. It is possible that the sensitivity bacteria could be further improved

of virus detection by means of antibody-coated by increasing the number of cells visible on the

micrographs. However, at present, the main shortcoming of the use of S. aureus cells lies in various technical problems that are encountered when handling the bacteria (clumping, possible heterogeneity of the cells, recognition of viruses only at the cell periphery). As a result, the data obtained

with the antibody-coated

grid method

are more reproducible

and suitable for quantitation. REFERENCES Almeida, J.D. and A.P. Waterson, 1969, Adv. Virus Res. 15, 307. Derrick, KS., 1973, Virology 56, 652. Doane, F.W. and N. Anderson, 1977, in: Comparative Diagnosisof Viral Diseases, eds. E. Kurstak and C. Kurstak (Academic Press, New York) Vol. 11, p. 505. Gough, K.H. and D.D. Shukla, 1980, J. Gen. Viral. 51,415. Jonsson, S. and E. Nordenfelt, 1979, in: Diagnosis of Viral Infections, eds. D.A. Lennette, S. Specter and K.D. Thompson (University Park Press, Baltimore) p. 115. Katz, D., Y. Straussman, A. Shahar and A. Kohn, 1980, J. Immunol. Methods 38,171.

166

Langone,

J.J., 1978, J. Immunol.

Methods

24,269.

Lesemann,

D.E. and H.L. Paul, 1980, Acta Hortic.

Lesemann,

D.E., R.F. Bozarth

Milne, R.G., 1980,

and R. Koenig,

Acta Hortic.

Milne, R.G. and E. Luisoni, (Academic

110, 129.

1977,

in: Methods

A. and M.H.V. Van Regenmortel,

Nicola’ieff,

A., G. Obert

G., R. Gloeckler,

Roberts, Roberts, Shukla, Van

in Virology,

1980, Ann. Virol. (Inst. Pasteur)

and M.H.V. Van Regenmortel, J. Burckard

I.M. and B.D. Harrison,

Regenmortel, Wagner (Plenum

Van Regenmortel,

M.H.V.,

and H. Koprowski

M.H.V.,

302 pp.

131E, 95. 12, 101.

1981, J. Virol. Methods

3,99.

1979, Ann. Appl. Biol. 93,289. 1981, J. Gen. Virol. (in press).

1979, J. Gen. Viral. 45,533. 1981,

Publishing

1980, J. Clin. Microbial.

and M.H.V. Van Regenmortel,

I.M., R.G. Milne and M.H.V. Van Regenmortel, D.D. and K.H. Gough,

New York)

eds. K. Maramorosch

Press, New York) Vol. VI, p. 265.

Nicola’ieff, Obert,

110, 119.

1980, J. Gen. Virol. 48, 257.

in: Comprehensive

Corporation,

1982,

Serology

Virology,

eds. H. Fraenkel-Conrat

and

R.R.

New York) Vol. 17, p. 183. and Immunochemistry

of Plant Viruses

(Academic

Press,