J. gen. ViroL (I969) , 5, 221-236
With 3 plates
22I
Printed in Great Britain
The Structural and Functional Diversity of Adenovirus
Capsid Components
By E. N O R R B Y
Department of Virology, Karolinska Institutet,
School of Medicine Stockholm, Sweden
(Accepted 20 March I969)
Adenoviruses offer a number of advantages for integrated morphological and
functional studies of structural virus components. During the course of multiplication
of these viruses structural components are produced in considerable excess. These
components are soluble, both in the traditional and in the true sense of the word, and
have dimensions which allow their examination by electron microscopy. In addition
most of these soluble components can be easily identified by a number of different
biological tests.
During the last four years there has been a considerable accumulation of information on structural and biological characteristics of adenovirus capsid components
(cf. Norrby, 1968a). This new phase of intensive studies of structural adenovirus components was initiated by the demonstration that vertex capsomeres differed immunologically from non-vertex capsomeres and that they carried projections (Valentine &
Pereira, I965; Norrby, I966a). In the present article an attempt is made to give a brief
summing up of our present state of knowledge in this field. In addition some recent
findings will be presented. Components with subcapsomere position in virus particles
(internal components) will be dealt with only en passant, partly due to the dearth of
information on the nature of these structures. It is hoped that this review will be of
value to non-adenovirus virologists and that it may facilitate a following up of future
developments. Emphasis will be placed on discussions of the correlation between the
biological subgrouping of different human adenoviruses (Rosen, I96o, Table I) and
structural-functional characteristics of their soluble components. In conclusion,
criteria for establishment of serotypes will be proposed.
Table I. Separation of human adenoviruses into subgroups on the
basis of their haemagglutinin (HA) activity
Rosen's subgroup number; H A activity
I. Complete agglutination of monkey
erythrocytes
II. Complete agglutination of rat erythrocytes
IlL Partial agglutination of rat erythrocytes
Serotypo numbers
3, 7, II, 14, I6, 20, 2I, 25, 28
8-to, t3, I5, I7, I9, 22-24, 26, 27, 29, 30
I, 2, 4, 5, 6, I2, I8, 3I*
* Types I2, I8 and 3t were originally described as non-haemagglutinating.
Terminology
The terminology for structural components currently used (Fig. I) was presented by
Ginsberg et al. (I966). They proposed the term hexon for nonvertex capsomeres,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
222
E. NORRBY
since these appear in a position surrounded by six other capsomeres. In one virion
24o out of the total of 252 capsomeres are hexons. Vertex capsomeres plus projections
are called pentons, for analogous reasons. The two parts of a penton, the vertex
capsomere and the projection, are referred to as penton base andfibre, respectively. In
addition the term dodecon (Gelderblom et al. I967) is used to denote aggregates of
12 pentons (see p. 226).
In the following the terms complete and incomplete haemagglutinins (HAs) will be
used to denote components giving a direct agglutination of red cells and components
which require the presence of antibodies against a heterologous serotype in order to
give agglutination, respectively (Fig. 3). The test by which incomplete HAs are demonstrated is called a haemagglutination-enhancement (HE) test.
Monomeric
form
Examples of
polymeric forms
Penton base-vertex capsomere 0
Fibre-vertex
projection
Penton-vertex capsomere
plus projection
:
~
*
~ ~ - e
Virion"
Dodecon
Fig. I. Schematicdescription of monomericand some experimentallyidentifiedpolymeric
forms of adenoviruscapsid components.
Some aspects of biological characteristics of monomeric and polymeric forms of capsid
components.
Figure I gives a schematic description of some different monomeric and polymeric
forms of structural components, which have been identified experimentally.
Hexons (non-vertex capsomeres) represent the dominating soluble component in
harvests from tissue cultures. Most hexons occur in a dispersed form, but there is some
tendency to aggregate into dimers, trimers, tetramers, etc. Disruption of virus particles
by sodium dodecyl sulphate (Smith, Gehle & Trousdale, I965a), heating (Russell,
Valentine & Pereira, I967) or by acetone at room temperature (Laver et al. 1968)
yields a specific polymer comprising nine hexons (Fig. I). This type of polymer might
represent a functional aggregate. Heat-treatment of virus particles under certain conditions can lead to a release of vertex capsomeres plus their neighbouring five hexons
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Diversity o f adenovirus capsid components
223
(Russell et aL I967). The partially degraded capsid which remains is composed of
180 hexons. This kind of structure can easily be constructed from 20 groups of nine
hexons with the configuration depicted in Fig. I. More than 90 ~o of the groups of nine
hexons seen in the electron microscope display a counter-clockwise appearance
suggesting a tendency to a specific orientation (E. C. Follett, personal communication). This might be due to a specific affinity between one side of the groups of nine
hexons and the carbon layer on grids used for electron microscopy.
Rosen's subgroup number
Serotypes analysed
4
III
1, 2, 5, 12 6
Length of fibres (nm.)
P
II
4
9, 15
28-31 23-28 17-19
I
3, 11, 16
'12-13 "10-11
Fig. 2. Diagrammatic presentation of tlae relative position m the elution diagram after
exclusion chromatography on Sephadex G 200 of fibres of various serotypesrepresenting different adenovirus subgroups. The curved line represents extinction values at 280 rim. of
components of normal calf serum, added as a reference in the fractionations. The length (nm.)
of different fibres as determined by electron microscopy of virus particles is also indicated.
Table 2. Some characteristics of fibres of different adenovirus serotypes
Subgroup
number
I
Serotypos Length of Incomstudied fibre (rim.) plete HA
3, tt, i6
Immunological
specificities
9 to It
No
Type-specificHI antigen
Type-specificHI antigen,
intersubgroup-specifie
(II +III) and subgroupspecific HE antigon
Type-specificHI antigen,
intersubgroup-specific
(II+ILO and subgroupspecific HE antigens
II
9, I5
I2 to I3
Yes
III
4
6
x, 2, 5, I2
I7 to 18
23 to 28
28 to 3t
Yes
Yes
Yes
References:
Norrby, I966a,
Norrby, I968b
Norrby et aL I967;
Norrby, I968c.
Wadell et al. I967;
Norrby et at. I969a;
Valentine & Pereira
I965; Pettersson
et aL I968;
Wadell & Norrby,
I969b.
Fibres (vertex projections). There is a considerable variation in length of fibres of
different serotype origin. The most sensitive technique for separating isolated, elongated structures o f the kind represented by fibres appears to be exclusion chromatography. Figure 2 illustrates the relative position of different isolated fibres in the
elution diagram of normal calf serum components (included as a reference) after
fractionation on Sephadex G200 (Pharmacia Fine Chemicals, Uppsala, Sweden) and
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
224
E. NORRBY
Table 2 gives some additional data on characteristics of various fibres. Fibres of all
subgroup I members studied (types 3, II, I6) elute in a position intermediate between
the IgG and albumin-peaks. The length of fibres extending from dodecons or virions
has been estimated as 9 to I i nm. (PI. I, 2; Norrby, I966a, I968b). Fibres of members
of subgroup II (types 9 and 15 have been studied) are somewhat larger, I2 to I3 nm.
(P1. I, 2; Norrby et aL I967; Norrby, I968c) and elute just after the peak o f l g G after
fractionation on Sephadex G2oo. The length of fibres of different members of Rosen's
subgroup III varies somewhat. Type 4, the aberrant member of the group, which with
regard to most biological activities other than haemagglutination belongs to subgroup I
Complete HAs
Incomplete HAs
=Red cellmembrane
Antibody against
=proximal part
of fibres (antigen8)
)--(
Antibody against
[~7/~ =vertex capsomeres
(antigent7)
Fig. 3- Schematic description of the identification of complete and incompleteHAs by
direct agglutinationand heamagglutination-enhancement(HE) tests, respectively.
(cf. Norrby, I968a ), carries fibres with a length of I7 to I8 nm. (PI. I, 2; Wadell,
Norrby & SchOnning, I967). Fibres of most members of subgroup III (types I, 2, 5)
plus type I2 have a length of 28 to 3I nm. (P1. I, 3; Valentine & Pereira, I965;
Pettersson, Philipson & H~Sglund, I968; Norrby & Ankerst, I969; Norrby, Wadell
& Marusyk, I969a ). They elute from a Sephadex G2oo column in a position intermediate between the void volume and the peak of IgG. However, fibres of type 6
elute somewhat later than these fibres and ultrastructural examination of virus
particles has shown that the former are 3 to 4 nm. shorter (PI. I) (Norrby et al. I969a).
Fibres of members of subgroups lI and III can be most readily demonstrated as
incomplete HAs in HE tests. The presumed background to this test is schematically
illustrated in Fig. 3. It is assumed that antibodies present in antisera against a heterologous serotype can aggregate fibres via their proximal parts, antigen 8 (Fig. 3
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Journal of General Virology, Vol. 5, No. 2
Plate I
Ultrastructure of virus particles of different serotype representatives of Rosen's subgroups I, type 3
(a); II, type 15 (b); and III, types 4 (c), 6 (d) and 2 (e). Negative contrasting with sodium tungstosilicate.
E. NORRBY
(Facing p. 224)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Journal of General Virology, Vol. 5, No. 2
Plate 2
Ultrastructure of dodecons of adenovirus types 3 (a), 4 (b), 9 (c) and 11 (d). Negative contrasting with
sodium tungstosilicate. Reproduced with permission of Current Topics Mierobiol. Immunol.
E. NORRBY
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Journal of General Virology, Vol. 5, No. 2
Plate 3
Ultrastructure of 15 to 16S (a and b) and 8 to 9 S (c and d) complete HAs of adenovirus type 6.
Negative contrasting with sodium tungstosilicate.
E. N O R R B Y
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Diversity o f adenovirus capsid components
22 5
Table 5; Pereira & de Figueiredo, I962) into complexes which are polyvalent with
regard to their capacity to interact with receptors on red cells. Antibodies against antigen
cannot interact with this part of a fibre when it occurs combined with vertex capsomeres
into pentons. Fibres of members of subgroup I cannot function as incomplete HA. They
therefore have to be identified by complement fixation tests or by haemagglutinationinhibition antibody consumption (HIC) tests (Norrby & Skaaret, I967).
Pentons (vertex capsomeres plus projections) of all serotypes studied can be identified as an incomplete H A (Norrby, 1966b , I968c, d; Norrby & Skaaret, I968; Norrby
& Ankerst, I969; Norrby et aL I967; Waddell & Norrby, I969a, Wadell, Norrby &
Skaaret, I969). In earlier studies (Pereira & de Figueiredo, I962) it was suggested that
pentons of some members of subgroup III represented a complete HA. This conclusion
was based on the presence of complete H A activity in preparations of pentons isolated
by anion exchange chromatography. In more recent studies (Wadell et aL I969)
preparations of this kind could be further separated by zonal centrifugation into
monomeric and oligomeric forms of pentons. Only the latter were found to carry
complete H A activity (see p. 227).
The penton incomplete H A can be demonstrated in H E tests after aggregation by
antibodies, present in heterotypic antiserum, which can react with vertex capsomeres,
e.g. their antigen specificity fl (Fig. 3, Table 5). Pentons of some serotypes carry a
toxin (cell-detachment) activity which is destroyed by treatment with trypsin (Table 3;
Pereira, I96o; Norrby & Skaaret, I967). This treatment causes a breakdown of the
vertex capsomere part of the penton, which leads to the appearance of isolated fibres.
The effect of trypsin is concentration-dependent. At very low concentrations of the
enzyme only toxin activity is destroyed whereas the capacity of pentons to function as
an incomplete haemagglutinin remains unchanged (Wadell & Norrby, I969b; U.
Pettersson, personal communication).
Table 3. Some characteristics of different kinds of complete and incomplete HAs
Type of
incomplete HA
Type of complete HA
A
Characteristic
features
Rate of sedimentation (S)
Toxin activity*
Degradation by trypsin
Presence of vertex
capsomerest
Dodecons
50 to IOO
Present
Variable
Yes
Dimer of
pentons
I5 to I6
Present
Yes
Yes
Dimer of
fibres
8 to 9
Absent
No
No
~
A
Penton
II
Present
Yes
Yes
Fibres
6
Absent
No
No
Formation of
Effect of heterotypie antiNone
Increase
None
haemagglutinating
sera on agglutinating
aggregates
activity
* Refers to serotypes, which exhibit this activity.
t Determined by the HEC test.
Penton bases (vertex capsomeres) as isolated structures are generally present only in
relatively low concentrations among spontaneously occurring soluble components.
However, among the different serotypes studied, type I2 materials form an exception
to this rule. In preparations of this serotype the relative concentration of free vertex
capsomeres is markedly higher than that of vertex capsomeres combined with fibres
i5
j. Virol. 5
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
226
E. N O R R B Y
into pentons (Norrby & Ankerst, I969). Treatment of type 3 materials with 2.0 Mguanidine+ HC1 caused a dissociation of pentons into vertex capsomeres and projections (Norrby & Skaaret, I967). Other substances such as formamide (Neurath,
Rubin & Stasny, I968) and pyridine (U. Pettersson, personal communication) can also
bring about such a dissociation.
Since isolated vertex capsomeres cannot function as an incomplete HA, a special
technique, the haemagglutination-enhancement antibody consumption (HEC) test,
was designed for their identification (Norrby & Skaaret, I967). A schematic description of this test is given in Fig. 4. The test measures the consumption of HE antibody
reacting with vertex capsomeres, some antigen specificities of which are common to all
Testsample ur homologousantiserum q- indicatorantigen
(heterologouspentons)
Negative --
~
~
%
~--~0-'~"
~
Outcomeof test
Agglutination
Fig. 4. Schematic description of the identification of vertex capsomeres, isolated or combined with fibres into pentons, by the haemagglutination-enhancement antibody consumption
(HEC) test.
human adenoviruses (Table 5)- The test sample, which either should not interact
with the red cells finally added or should be deprived of all complete HA, is incubated
with homologous antiserum. To demonstrate the remaining penton HE antibodyin this
mixture a preparation of heterologous pentons is added. In a situation where the test
sample contains pentons or free vertex capsomeres these will absorb the penton HE
antibody of serum and no agglutination by indicator pentons will occur.
As was mentioned above, the penton base is susceptible to trypsin digestion, which
treatment destroys the toxin activity of penton. There are some findings indicating that
not only pentons, but also isolated vertex capsomeres may carry toxin activity (Wadell
& Norrby, I969b; U. Pettersson, personal communication).
Soluble HAs of three different kinds have been identified (Fig. I, 3). These are (a)
dodecahedral aggregates of Iz pentons plus some extra structure, the type of complex
called dodecons (P1. 2), (b) dimers of pentons (P1. 3a, b) and (c) dimers of fibres
(PI. 3 c, d). Some characteristics of these different HAs as well as their occurrence in
preparations of different serotypes of human adenoviruses have been summarized in
Tables 3 and 4, respectively.
Dodecons vary in their rate of sedimentation, from 50 to Ioo S (Norrby, I966a;
Norrby, I968b; Norrby et aL I967). This is a true variation reflecting structural
differences which occur between dodecons of different serotypes (cf. P1. z). These
differences concern both the size of the central part of the dodecon, i.e. the aggregate of
vertex capsomeres and some extra structure, and the length of fibres. Dodecons and
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Diversity o f adenovirus capsid components
227
isolated pentons share a number of biological characteristics. However, the sensitivity
of dodecons of different serotypes to trypsin digestion varies somewhat. This may be
due to the fact that vertex capsomeres forming a part of a dodecon complex cannot be
as readily attacked by trypsin as free structural components.
Dimers of pentons carry all the properties of isolated pentons (Norrby & Skaaret,
1968, Wadell et aL I969), but sediment with a rate that is about I"3 times greater, I5 to
] 6 S. Dimers of pentons represent the only complete H A which displays an increase in
agglutinating activity in the presence of heterologous antiserum (a true enhancement).
This is caused by antibodies reacting with vertex capsomere antigen. The effect has been
interpreted as being due to a steric reorientation of a considerable fraction of nonagglutinating penton dimers into agglutinating polymers.
Table 4. Occurrence o f different types o f soluble HAs among
various human adenovirus serotypes*
Dimer
Rosen's
subgroup
I
II
II1
Serotype
3
7
II
I6
3-I6
9, 15
4
Dimer
of
of
Dodecon pentons fibres
+
+
+
--+
+
-+
-+
-I+
-
References
Norrby, 1966a
Neurath & Rubin, I968
Norrby, I968b
Norrby & Skaaret, I968
Norrby & Skaaret, I968
Norrby, I968c; Norrby et al. 1967
Norrby & Wadell, I967; Wadell et al.
I967
I, 2, 5, 6
+
+
Wadell & Iqorrby, I969a;
Wadell et al. 1969
12
-+
Norrby & Ankerst, 1969
* Reproduced from Norrby & Skaaret (I968) with permission of
Academic Press: New York & London.
-
-
The third kind of soluble HA, dimers of fibres, shares a number of properties with
isolated fibres (Norrby, 1968 c; Norrby & Skaaret, x968; Norrby et al. x 969 a; Wadell
& Norrby, I969a). However, it is a larger complex sedimenting at a rate of 8 to 9 S
as compared with about 6 S for free fibres. A comparison of sedimentation characteristics and exclusion chromatography behaviour of this kind of complete H A dearly
demonstrates that it is a markedly elongated structure.
The polyvalent and univalent character of complete and incomplete HAs, respectively, with regard to their capacity to interact with receptors on cells should be emphasized (cf. the schematic illustration in Fig. 3). The question why some pentons or
fibres occur in monomeric and others in oligomeric forms of specific configurations
cannot be answered at present. Aggregation of pentons into dodecons seems to occur
around some extra component (Norrby, x966a; Wadell et al. 1967). Possibly the
dimerization of pentons and fibres is also dependent upon some extra component.
In preparations of various human adenovirus serotypes either one or two of the
different kinds of HAs have been identified. As can be seen from Table 4, all theoretically
possible combinations, except a simultaneous occurrence of dodecons and dimers of
pentons, have been found. Possibly the presence of either one of these two penton
polymers may mutually exclude the occurrence of the other.
In the absence of heterotypic antisera, preparations of members of subgroup 1II
15-2
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
228
E. NORRBY
give only a partial agglutination of rat erythrocytes. This has recently been shown
(Wadell, i969) to be due to a competition for receptors on red ceils between complete
and incomplete HAs, the latter of which occur in excess (Wadell & Norrby, I969a,
Wadell et al. I969). In contrast, preparations of members of subgroup II contain
markedly more complete than incomplete H A s (Norrby et aL I967; Norrby, I968c)
and therefore give a pattern of complete agglutination of rat erythrocytes. An additional factor of importance for the haemagglutination pattern obtained is the number of
specific receptors occurring on cells of different kinds. It was recently found (Wadell,
1969) that members of subgroups II and I I I can agglutinate not only rat erythrocytes
but also human O red cells. However, presumably owing to the fact that the latter ceils
carry relatively few receptors, they display a markedly partial agglutination which
readily escapes detection. Certain members of subgroup II, which give complete
agglutiation, are exceptions to this rule.
Table 5. Summary of biological characteristics of
structural adenovirus components*
Antigen
t"
Structure
Non-vertex capsomeres; hexous
Designation
a
?
e
^
- - ' ~
Specificity
Group
Intersubgroup
and intrasubgroup
Type
Remarks
Localized at the inside of the capsid
Common to a restricted number of
serotypes in various combinations
Available at the surface of the capsid
Reacts with neutralizing and virus
particle-HI antibody
Vertex capsomeres; penton
The complex of base and fibre
fl
Group
bases
(= penton) carries (a) toxin activity,
(b) incomplete HA activity
.9
Intersubgroup Common to a restricted number of
and intrasub- serotypes in various combinations
group
8
Intrasubgroup Proximal part of projections. Present
Vertex projections; fibrest
only in members of the same subgroup (not present in members of
subgroup I)
?
Intersubgroup Shared between members of subgroups 11 and III
Type
Distal part of projection. Reacts
with HI and neutralizing antibody
Pentagonal; dodecon?
Unknown
Forms a core of dodecons. Possibly
associated
subcapsomere component of vertices
of virus particles (?)
' Meshwork '-like
?
Unknown
Arginine-rich internal components.
* Reproduced from Norrby 0969) with permission of Springer-Verlag, Berlin, Heidelberg, New York.
t Incomplete HA of members of subgroups II and II1.
Additional structural components, as yet only partly characterized, have been described. A spontaneous disruption of dodecons of type 3 (Norrby, I966a) and type 4
(Wadell et al. i967) caused a release of ring-like or occasionally pentagonal particles.
This type of component has been assumed to form the core of dodecons. It has
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Diversity o f adenovirus capsid components
229
furthermore been speculated that it may occur in a subcapsomere position at the
vertices of virions. It has not as yet been identified by any biological test.
Recently, methods have been described for the isolation of internal components of
virions (Laver et aL I968; Prage, Petterson & Philipson, I968; Russell, Laver &
Sanderson, I968 ). Biochemical and biological characterization of these components
is currently being undertaken.
D
U_o..o° o00 )
(b)
Vertices
of virions
[
]
Dodecon
Pentons
Hexons 0
IgG antibodies !
against hexons
[
Fig. 5. Schematic description of the background to the selective HI of virions by anti-hexon
antibodies. The two mechanisms include (a) aggregation of virions and (b) steric blocking of
fibres by antibody attached to para-vertex hexons.
Immunological specificities carried by different structural components
The present state of knowledge of immunological specificities and other biological
characteristics of structural components has been summarized in Table 5.
Hexons. These structural components carry an antigenic specificity (c0 which is
common to all human and non-human adenoviruses, except the GAL virus of birds
(Pereira et al. I963). The possible presence in hexons of a type-specific antigen
inducing neutralizing antibody is a matter of controversery. In some studies a strong
neutralizing activity of anti-hexon sera was found (Wilcox & Ginsberg, I963; Kjellrn
& Pereira, I968; Gelderblom et al. I968; Norrby, ]969) , whereas in others, highly
purified hexons could not induce the production of neutralizing antibody (Pettersson,
Philipson & HOglund, ] 967). This discrepancy between different results can be explained
in two different ways. Either tile antigen mainly responsible for the induction of neutralizing antibody is a separate component which occurs intimately associated with
hexons and which was removed only by the purification procedure employed by
Pettersson et aL (I967), or this procedure removed or, more likely, altered the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
230
E. NORRBY
specificity of a component which represents an integrated part of the hexons.
Available evidence does not allow a choice between these two possible explanations.
The presence of a type-specific hexon antigen was suggested already from early
studies on CF antigen activity of human adenovirus preparations (Pereira, I956).
These results were recently confirmed in tests including purified hexons (D~Shner,
I968 ). Additional data clearly demonstrating that hexons carry a type-specific
antigen have been accumulated (Norrby, I969; Norrby, Marusyk & Hammarskj~51d.
I969b). One reason for this conclusion was the finding that although antihexon sera
do not contain any H E antibody or HI antibody reactingwith soluble HAs, they do contain a type-specific antibody which can inhibit effectively the haemagglutinating activity
of virions. This effect can be due to two different mechanisms which have been schematically outlined in Fig. 5. These mechanisms are (a) an aggregation of virus particles
(alternative (a), Fig. 5) or (b) a steric interference between IgG antibodies, attached to
hexons with a paravertex position, and projections extending from vertex capsomeres
(alternative (b), Fig. 5). The latter interference might prevent projections from attaching
Table 6. V i r u s p a r t i c l e H I t e s t s with a n t i - h e x o n s e r a a g a i n s t d i f f e r e n t
representative members of Rosen's subgroups
Anti-hexon
serum of
type*
3(I)t
15(II)
4(lid
6 (III)
2(IID
Type 3
virus
particles
Type 15
virus
particles
Type 4
virus
particles
Type 6
virus
particles
Type 2
virus
particles
Homologous
soluble HA
i6oo
< 8
i6oo
<8
< 8
4oo
<8
< 8
< 8
<8
< 8
< 8
<8
< 8
< 8
< 8
< 8
z6:~
< 8
< 8
8:~
< 8
< 8
<
<
<
<
8
8
8
8
< 8
< 8
< 8
* Sera were diluted to contain corresponding quantities of homologous hexon CF activity (64o
units per 0.02 ml).
R o m a n figures within parenthesis denote subgroup belonging.
Traces of agglutination were detectable in tubes counted as negative.
to receptors on cells. As was described above, fibres of different serotypes vary in
length between IO and 3o nm. Since the length of a maximally extended lgG antibody
is I5 to 20 nm. (Valentine & Green, i967), one would anticipate a variation in the
capacity of an antibody of this kind to block sterically the fibres of different serotypes.
In order to analyse this problem, antisera against hexons of different serotypes containing the corresponding amount of homologous CF antibody and lacking HI
activity in tests with soluble components were tested in HI tests with different virions
(Table 6) (Norrby & Wadell, I969). Sera against hexons derived from serotypes
carrying fibres with a length similar to or shorter than that of serotype 4 (I7 to I8 nm.)
inhibited effectively the activity of homologous virions, whereas anti-type 6 and antitype 2 hexon sera only displayed low activities. This might be interpreted as suggesting
that in the case of members of subgroups I and II and in addition type 4, steric interference of the kind described plays a major role, whereas the low virion-Hl activity of
the antihexon sera of subgroup III members other than type 4is due to an aggregation of
the particles. However, recent data obtained in immuno-electron-microscopy studies
suggest that the true explanation may be a more complex one (Norrby, unpublished).
The fact that the virion-HI activity of antihexon sera is demonstrable only with
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Diversity o f adenovirus capsid components
23x
homologous virions implies not only that hexons carry a type-specific component, but
also that this component is available at the surface of virions, whereas the groupspecific part of hexons is turned towards the interior of the particle. The occurrence of
such a specific orientation of hexons in the capsid has been verified by additional
experimental studies, including immuno-electron microscopy (Norrby et al. I969b).
These findings can explain previous observations that only homotypic antibodies
attach to the surface of virions (Smith et aL I965b),
Absorption of antihexon sera with purified heterologous hexons has demonstrated
not only the presence of type and group antigenic specificities but also subgroup
specificities (Norrby, 1969; Norrby & Wadell, 1969). It is of some interest that by use
of this technique it was shown that hexons of the aberrant subgroup III member, type 4,
are more closely related to hexons of members of subgroup I than to those of subgroup
III. In this connexion it might be mentioned that heterologous neutralization of virions
has been demonstrated by incubation of antigen-antibody complexes at low pH
(Kjellrn, I966; Kjellrn & Pereira, 1968). This effect was seen only between virions of
one serotype and antisera against other serotypes belonging to the same subgroup.
Since this effect presumably was due to antibodies reacting with hexons, it suggests
that antigen specificities of a subgroup nature occur in these components.
Penton bases. The antigenic specificities of penton bases are responsible for the
interaction with HE antibody converting the incomplete penton HA into an agglutinating complex (Fig. 4). Since a penton HE activity can be demonstrated by any
heterologous antiserum the penton base must contain one antigenic specificity (/~)
which is common to all human adenoviruses. In addition it contains some intersubgroup and intrasubgroup specificities as demonstrated by determination of the
immunological characteristics of HE antibodies in absorption experiments (Norrby,
I969; Wadell & Norrby, 1969 b). No type-specific component has been found by use of
this technique. This, however, does not exclude the possible presence of a component
of this kind. Immunological specificities of vertex capsomeres of the aberrant subgroup
III member, type 4,like those of non-vertex capsomeres (see above), are more similar to
those of the corresponding components of subgroup I than of subgroup III members.
Fibres. The immunological complexity of fibres increases in parallel with the increase
in their length (Tables 2, 5). The shortest fibres (9 to I I nm.), which are carried by
members of subgroup I, seem to contain only one antigen. This is the antigen y,
which is responsible for the induction of type-specific HI antibodies. Antibodies
against the y antigen in addition carry some neutralizing activity. This is most effectively demonstrated by the rapid technique of fluorescent foci inhibition (U. Pettersson,
personal communication), but less effectively so, or not at all, with other techniques
(Norrby, I969; Kjellrn & Pereira, 1968; Pettersson et al. 1968). The most plausible
explanation for this behaviour seems to be that antibodies against fibres do not completely block but rather cause a delay of the process of virus replication. Fibres of
members of subgroup II in addition to the type-specific antigen ~, carry antigenic
specificities of a subgroup and intersubgroup-specific (II and III) nature (Norrby, i968;
Wadell & Norrby, i969). The longest fibres, i.e. those belonging to members of subgroup IlI, contain at least one more antigenic specificity,which is subgroup-specific
(Wadell & Norrby, 1969). This corresponds to the antigen described by Pereira &
de Figueiredo (1962). Antigenic specificities which reside in the proximal parts of
isolated fibres can interact with HE antibodies (Fig. 4).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
232
E. NORRBY
Other components. The immunological specificity of the internal component of
dodecons and additional components recently isolated from the cores of virions has
not as yet been clearly elucidated.
The occurrence of serotypes with a capsid composed of a mosaic of structural components
related to one or two heterologous serotypes
Typing of adenoviruses is performed by use of neutralization and HI tests. The results
of these two tests do not always agree and some cross-reactions occur between prototype strains in either one or both of the tests. In addition, so-called intermediate strains
have been described which behave like one prototype in the neutralization test but a
completely different one in the HI test (cf. Wigand & Fliedner, 1968 ). The recognition
of two distinct type-specific antigens in virions allowed an analysis, on the level of
structural components, of the background for these cross-reactions. As was mentioned,
these two antigens are the distal parts of fibres (7) which react with HI antibodies,
specifically demonstrable in tests with soluble components, and the e antigen of
hexons. The latter can be readily demonstrated in virion-HI tests with most serotypes
(Table 6). Presumably this antigen is identical with or related to the antigen, which is of
dominating importance for the induction of neutralizing antibody.
Table 7. Antibody activities of antisera from hyperimmunized animals against different
type 4, ~6 and 3-I6 components in H I tests with types 4 and I6 virions and homologous
soluble HA and in CF tests with homologous hexons
HI titres in tests with
Sera against
adenovirus
type
4 hexons
16 hexons
I6 pentons
3-I6 virions
:
Type 4
virions
~
Type 16
virions
Homologous
soluble H A
,
Homologous
hexon CF
activity
512o
640
< 8
< 8
80
2560
320
I6O
< 8
< 8
x28o
128o
2560
2560
< 8
I28O
Ratio of homologous/heterologous neutralization titres
Type 4 16/64
Type i6 2/4
An analysis of cross-reactions in neutralization tests between prototype strains
(see Stevens et al. 1967) reveals some rather close cross-reactions between pairs of
serotypes belonging to the same subgroup. However, an interesting reciprocal crossreaction between types 4 (Rosen's subgroup III) and 16 (Rosen's subgroup I) has also
been demonstrated. This cross-reaction was studied by use of the two kinds of HI tests
mentioned above (Table 7)- It was found that antiserum against type 4 hexons inhibited
mainly homotypic virions. In contrast, antiserum to type I6 hexons inhibited both
homotypic and to a considerable extent also type 4 virions. This one-way crossing
parallels the findings in comparative neutralization tests. The ratio of homotypic to
heterotypic neutralizing antibody titre is much lower for antisera to type I6 hexons
(2 to 4) than for antisera to type 4 hexons (I6 to 64). The fact that antiserum to type I6
pentons did not inhibit type 4 virions demonstrates that the structural similarity
between type 4 and 16 must reside in their hexons. The reason for the one-way
dominance in the reciprocal crossing must be that the antigenic specificity, which is
shared between hexons of types 4 and 16, is available to a greater extent at the surface of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Diversity o f adenovirus capsid components
233
type 4 virions. The occurrence of additional antigenic components of a type-specific
nature has been demonstrated in cross-absorption experiments (Norrby & Wadell,
I969).
The character of intermediate strains has been studied in the author's laboratory
with types 9-I5, described by Cramblett et al. (196o), and type 3-16 (the San Carlos
agent), serologically characterized by Hatch & Siem (I966). The working hypothesis
that the intermediate strains should have a capsid representing a mosaic of hexons and
fibres (or pentons)similar to those of the serotypes with which they can react in neutralization and H I tests, respectively, (cf. Norrby, I968a), has been proved to be correct
(Norrby, in preparation). In Table 7 an antiserum to type 3-16 virions has been ineluded as a biological control. This serum shows a high H I titre with the two adenovirus type I6 HAs, but no reaction with type 4 virions. This implies that hexons of
type 3-16 differ from those of type 4 and consequently also from those of type 16.
This difference is further emphasized by the results of testing antisera to type 3,
3-I6 and 16 hexons and fibres for H I antibodies against virions and soluble H A s
(Table 8). Antisera against type 3 or 3-I6 hexons both inhibited virion agglutination of
types 3 and 3-16, but not 16. In contrast, antisera to fibres of types 3-16 and I6 reacted
similarly, i.e. they inhibited all antigens of these two types, whereas antiserum to
type 3 fibres only contained antibodies reacting with homotypic antigens. These
results indicate the occurrence of close immunological similarities between type 3 and
3-I6 hexons and between type 3-16 and 16 fibres.
Table 8. H I titres of antisera against hexons and fibres of types 3, 3-16 and I6 in tests
with soluble HA and virions of all three types
HI titres in tests with:
A
r
Sera against
adenovirus type
3 hexons
3 fibres
3-I6 hexons
3-16 fibres
16 hexons
I6 fibres
Type 3
~
,
Soluble HA Virions
< 4
640
< 4
< 4
< 4
< 4
I2,8oo
640
64o
< 4
< 4
< 4
Type 3-16
r- ~
Soluble HA
Virions
< 4
< 4
< 4
64o
< 4
16o
i2,8oo
< 4
i6o
64o
< 4
I6O
,
Type 16
,
~--,
Soluble HA Virions
< 4
< 4
< 4
320
< 4
640
< 4
< 4
< 4
16o
x28o
I28o
Some aspects of criteria to be applied for typing of adenoviruses
The importance of the problem concerning criteria to be adopted for typing of
adenoviruses is emphasized by the above-mentioned demonstrations of the occurrence
of a mosaic of multi-related structural components in some adenovirus capsids. Since
there are two different type-specific components, it seems logical to let one of them form
the basis for a separation into different serotypes and the other one for a further
separation into different subtypes. The primary question, then, is which one of the
type-specific components--hexons or fibres--should be given priority? In practice this
means a choice between the use of the neutralization test (or when possible the virionH I tests) or of the H I tests with soluble components. Although the latter test is
advantageous from the point of view of convenience, the neutralization test should
preferably be chosen. One reason for this is the fact that this test gives a more effective
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
234
E. NORRBY
separation into serotypes and that therefore fewer subtypes would be eneountered.
Other, less precise arguments for this choice are (a) that antihexon antibodies are
presumably of relatively greater importance from the point of view of protection in
vivo and (b) that hexons are presumably phylogenetically 'older' than fibres. Taking
this cue the following proposals are made.
(I) The primary separation of adenoviruses into serotypes should be based on an
individualistic behaviour in the neutralization test. Since a number of low-grade crossings occur in this test 'individualistic' might arbitrarily be defined to mean an eightfold or greater reactivity between homologous antiserum and virus than between the
same serum and heterologous serotypes. Special consideration has to be given to
cross-reactions, which only one way is less than 8. Possibly one could indicate such a
relationship by adding the relevant serotype within parenthesis. Thus serotype I6
could be denoted I6(4), whereas the designation of type 4 would not be affected. It
should be pointed out that neutralization tests should be performed with relatively
'slow' systems, since these appear to measure exclusively anti-hexon antibody.
(2) A separation into subtypes (strains) should be made on the basis of an individualistic behaviour (more than eightfold homologous preference of antiserum) in the
HI test employing soluble HA as antigen. By this criterion serotype 29 (Rosen et al.
1962; Stevens et al. 1967) should be regarded as a subtype of type 15.
The situation in which the fibre specificity of one serotype is identical with or
closely related to that of another serotype must be treated separately. Finally, the socalled intermediate strains represent a special case. These could be presented as before
in a hyphenated way, but the serotype with which a cross-reaction occurs in the
neutralization test should be preferably given first. Thus, type 9-I5 should be called
type 15-9, whereas type 3-I6 would retain its designation.
It should be pointed out that the designation of adenovirus strains with different
immunological specificities poses certain problems. Possibly some provisional system
could be adopted. However, it is important that this system allows future modifications.
One can envisage that one day it will be possible to designate different strains by use of
an index system, which describes the combination of different antigenic specificities
carried by various structural components, hexons, penton bases, fibres and possibly
others. Future endeavours will presumably also make it possible to specify the basis
for those variations, concerning for example hexons, by reference to their amino acid
sequences. Possibly the molecular nature of these variations might resemble those
found recently in studies of polypeptides of y-globulins.
Many of the data quoted were obtained in collaborative studies with Dr G~3ran
Wadell. Mrs Halyna Marusyk gave valuable assistance in ultrastructural studies.
The author's work presented in this review was supported by grants from the Swedish
Medical Research Council (projects nos. B 69-I 6 x-548-o5 A and B 6 9-I 6x-744-o4) and
the Swedish Cancer Society (project no. I7I-K68-oi X).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
Diversity of adenovirus capsid components
235
REFERENCES
CRA~a3LEa-r, H.G., KASEL, J.A., LANG~CK, M. & WlLKEN, F.D. (196o). Illnesses in children
infected with an adenovirus antigenically related to types 9 and 15. Pediatrics, N.Y. z5, 822.
D&rNER, L. (I968). Untersuchungen zum Nachweis von serologiscben Unterschieden zwischen den
Gruppenantigenen aus verschiedenen Adenovirustypen. II Demonstration yon Antigendifferenzen mit tier KBR in der isofixationstechnik. Z. Immun.Forsch. x36, 45I.
GELDERBLOM,H., BALrER,H., FRANK,H. &WlGANO,R. (I967)- The structure of group II adenoviruses.
J. gen. Virol. x, 553.
G~LDEP,aLOM,H., MrlSER, W., LmqGYEL,A. & WmAND,R. (1968). The serological relationship of the
soluble antigens of adenovirns type 19. J. gen. ViroL 2, 331.
GrNSBERG,H. S., PEREmA,H. G., VALEICrlNE,R. C. & WILCOX,W. C. (I966). A proposed terminology
for the adertovirus antigens and virion morphological subunits. Virology 28, 782.
HATCH, M. H. & Smi, R. A. (I966). Viruses isolated from children with infectious hepatitis. Am. J.
Epidem. 84, 495.
KXELL~N,L. (I966). Heterologous neutralization of adenovirus type 5 at low pH. Virology 28, 492.
KJELL~N, L. & P~e,EmA, H. G. 0968). Role of adenovirus antigens in the induction of virus neutrafizing antibody. I. gen. Virol. 2, I77.
LAVER, W. G., PEREmA,H. G., RUSSELL,W. C. & VALErCrn,~,R. C. (1968). Isolation of an internal
component from adenovirus type 5. at. molec. Biol. 37, 379.
N~tn~rI~, A. R. & RtmIN, B. A. (1968). Interaction of p-chloromercuribenzoate with adenoviruses.
Inactivation of haemagglutinins and degradation of virions of types 3, 4 and 7. d. gen. ViroL 2,
2I 5.
NEtn~Ta, A.R., RUmN, B.A. & STASNY, J.T. (1968). Cleavage by formamide of intereapsomer
bonds in adenovirus types 4 and 7 virions and hemagglutinins. J. Virol. z, lO86.
Nomu~Y, E. (I966a). The relationship between the soluble antigens and the virion of adenovirus
type 3. I. Morphologic characteristics. Virology 28, 236.
NeaRBY, E. (I966b). The relationship between the soluble antigens and the virion of adenovirus
type 3. II. Identification and characterization of an incomplete hemagglutinin. Virology 3o, 608.
NOgRBV, E. (I968a). Biological significance of structural adenovirus components. Current Topics
MicrobioL lmrnunol. 43, I.
NORgBY, E. (1968b). Comparison of soluble components of adenovirus types 3 and I I. J. gen. ViroL
2, I35.
Nom~sY, E. (I968c). Comparative studies on the soluble components of adenovirns types 9 and 15
and the intermediate strain 9-I5. d. Virol. 2, I2oo.
NORRBY,E. (I968d). Identification of soluble components of adenovirus type I I. J. gen. Virol. 2, 123.
NORRBY,E. (1969). The relationship between the soluble antigens and the virion of adenovirus type 3IV. Immunological complexity of soluble components. Virology 37, 565.
NORRBY, E. & ANKERST, J. (1969). Biological characterization of structural adenovirus type I2
components..1, gen Virol. (In the Press.)
NoggBY, E. & SKA~T, P. (I967). The relationship between the soluble antigens and the virion of
adenovirus type 3- HI. Immunologicalidentification of fiber antigen and isolated vertex capsomer
antigen. Virology 32, 489.
NORRBY, E. & SKA~T, P. (I968). Comparison between soluble components of adenovirus types 3
and I6 and of the intermediate strain 3-16 (the San Carlos agent). Virology 36, 2oi.
Noggmc, E. & W~ELL, G. (I967). Soluble components of adenovirus type 4. Virology 3x, 592.
NORRBV, E. & WADELL, G. (1969). Immunological relationships between hexons of certain human
adenoviruses. Virology. (Submitted for publication.)
NORRBY, E., WAI~ELL, G. & MAgUSYK, H. (I969a). Fiber-associated incomplete and complete
hemagglutinins of adenovirus type 6. Arch. ges. Virusforsch. (In the Press).
NORRBY, E., MARUSYK,H. & HAMMAgSKJ6LD,M.-L. 0969 b). The relationship between the soluble
antigens and the virion of adenovirus type 3. V. Identificationof antigens available at the surface
of virions. Virology. (In the Press).
NOggBY, E., NYBERG, B., SKhA~T, P. & LENGYF-L,A. (I967). Separation and characterization of
soluble adenovirus type 9 components, d. Virol. x, 1lOI.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46
236
E. N O R R B Y
PEREIRA, H. G. (1956). Typing of adenoidal-pharyngoal-conjunctival (APC) viruses by complementfixation. J. Path. Bact. 72, to5.
PEREIRA,H. G. (I960). Antigenic structure of non-infectious adenovirus materials. Nature, Lend. I86,
57t.
PEREIRA, H. G. & DE FIGUEIREDO,M. V. T. (I962). Mechanism of hemagglutination by adenovirus
types I, 2, 4, 5 and 6. Virology x8, I.
PEREIRA,H. G., HUEBNER,g. J., GINSBERG,H. S. & VAN DERVEEN,J. (I963). A short description of
the adenovirus group. Virology 2o, 6I 3.
PETTERSSON,U., PI-IILIPSON,L. & H6GLUND,S. (1967)- Structural proteins of adenoviruses. I. Purification and characterization of the adenovirus type 2 hexon antigen. Virology 33, 575.
PETTERSSON,U., PHILIPSON,L. & H6GLUND,S. (1968). Structural proteins of adonoviruses. II. Purification and characterization of the adonovirus type 2 fiber antigen. Virology 35, 204.
PRAGE, L., PETI'ERSSON,U. & PHILIPSON,L. (1968). Internal basic proteins in adenovirus. Virology
36, 5o8.
ROSEN, L. (1960). A hemagghitination-inhibition technique for typing adenoviruses. Am. J. Hyg. 7x.
I2O.
ROSEN,L., BARON,S. & BELL,J. A. (196t). Four newly recognized adenoviruses. Prec. Soc. exp. Biol
Med. xo7, 434.
ROSEN, L., HOVIS, J. F. & BELL, J.A. (1962). Further observation of typing adenoviruses and a
description of two possible additional serotypes. Prec. Soc. exp. Biol. Med. xio, 7IO.
RUSSELL, W.C., LAVER, W.G. & SANDERSON,P.J. (1968). Internal components of adenovirus.
Nature, Lond. 219, 1I27.
RUSSELL,W. C., VALEN'rlNE,R. C. & PEREmA,H. G. (I967). The effect of heat on the anatomy of the
adenovirus. J. gen. Virol. x, 509.
SMITH, K. O., GEHLE, W. D. & TROUSDALE,M.D. (1965a). Architecture of the adenovirus capsid.
J. Bact. 90, 254.
SMrm, K. O., TROUSDALE,M. & GEHLE, W. (I965b). Adenovirus-antibody agglutination reaction.
J. hnmunol. 95, 8IO.
SVEVENS,D. A., SCnAEErER,R., FOX, J. P., BRANDT, C. D. & ROMANO,M. (1967). Standardization
and certification of reference antigens and antisera for 30 human adenovirus serotypes. Am. J.
Epid. 86, 617.
VALENTINe, R.C. & GREEN, N. M. (1967). Electron microscopy of antibody-hapten complex. J.
molec. Biol. 27, 615.
VALE~rrlNE,R. C. & PEP,EmA, H. G. (1965). Antigens and structure of the adeaovirus. J. molec. Biol.
I3, I3.
WADELL,G. (1969). Hemagglutination with adenovirus sorotypes belonging to Rosen's subgroups II
and III. Proc. Soc. exp. Biol. Med. (In the Press.)
WADELL,G. & NORRaV,E. (I969 a). The soluble hemagglutinins of adenoviruses belonging to Rosen's
subgroup IIL II. The slowly sedimenting complete HA. Arch. ges. Virusforsch. 26, 53.
WADELL, G. & NORRBY,E. (I969b). Immunological and other biological characteristics of pentons
of human adenoviruses. Submitted for publication to J. Virol.
WADELL, G. & SKAARET,P. (I967). Separation of adenovirus type 4 components. Arch. ges. Virusforsch. 2x, 366.
WADELL,G., NORRaY, E. & SCH6NNINO,U. (1967). Ultrastructure of soluble antigens and the virion
of adenovirus type 4. Arch. ges. Virusforsch. 2x, 234.
WADELL, G., NORRBY,E. & SKAARET,P. (1969). The soluble hemagghitinins of adenoviruses belonging to Rosen's subgroup III. I. The rapidly sedimenting complete HA. Arch. ges. Virusforsch. 26,
33WIGAND,R. & FLmD~,X,D. (1968). Serologically intermediate adenovirus strains: A regular feature of
group II adenoviruses. Arch. ges. Virusforsch, 24, 245.
WIGAND, R., BAUER,H., LANG, F. & ADAM,W. (1964). Neutralization of the adenoviruses types I to
28: Specificity and antigenic relationship. Arch. ges. Virusforsch. I5, I88.
WILCOX, W. C. & GINSBER~, H. S. (1963). Production of specific neutralizing antibody with soluble
antigens of type 5 adenovirus. Proc. Soc. exp. Biol. Mecl. xx4, 37.
( R e c e i v e d 5 N o v e m b e r I968 )
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:09:46