CHEMO-IMMUNOLOGICAL STUDIES ON CONJUGATED
        CARBOHYDRATE-PROTEINS

VIII. THE INPLUENCE OP THE ACETYL GROW ON THE SPECIFICITY OF
          HEXOSIDE-PROTEIN ANTIGENS

BY WALTHER F. GOEBEL, PH.D., FRANK H. BAFJERS, AND OSWALD T.
              AVERY, M.D.

(From the Hospital oj The Rockefeller Institute for Medica! Research)

(Received for publication, April 6,1934)

That the acetyl group (CHKO-) can exert a specific influence in
determining the serological characteristics of a bacterial carbohydrate
has recently been shown in this laboratory (1).  Tbe capsular poly-
sac&ride of Pneumococcus Type I occurs in the intact cell as an
acetyl derivative of a complex nitrogenous carbohydrate.  The union
of the acetyl groups with the polysaccharide itself is of a highly labile
nature.  When these groups are removed by mild alkaline hydrolysis,
the resultant deacetylated carbohydrate is found to be deprived of cer-
tain immunological properties possessed only by the parent substance.
Since the only demonstrable chemical change that occurs during
alkaline hydrolysis is the removal of acetyl groups, the additional
serological characteristics exhibited by the fully acetylated carbohy-
drate can be attributed only to the presence of the acetyl radical in the
intact bacterial polysaccharide.

In the attempt to understand the immunological significance of the
acetyl group in the more complex bacterial polysaccharides, a study has
been made of the serological specificity of two related antigens each
containing a simple carbohydrate radical in the acetylated and the
unacetylated form.  For this purpose, therefore, the p-aminophenol
8-glucoside of glucose and its monoacetyl derivative have been synthe-
sized. These two glucosides have been combined with the globulin
of horse serum by means of the diazo reaction, and the resultant


86      CONJUGATED CARBOHYDRATE-PROTEINS. VIII

"synthetic" carbohydrate azoproteins have been employed as antigens
in the production of immune rabbit sera.  The carbohydrate radicals
of these two antigens are identical in their stereochemical relationships,
yet they differ in that one of the derivatives contains an acetyl group.
From the mode of synthesis it is probable that the acetyl radical is
bound on the sixth carbon atom of the glucoside, although its exact
allocation in the molecule is, for the purpose of this study, irrelevant.

It has previously been shown that conjugated carbohydrate-protein
antigens containing either the azophenol CX- or &glucoside give rise
in each instance to antibodies which are predominantly specific, yet
capable of a secondary reaction with test antigens containing the
heterologous hexoside (2). By means of the inhibition test it has
been shown that the primary immune reaction is quite specific, since
the union between homologous antigen and antibody is inhibited
only by homologous glucoside, whereas the cross-reaction with the
heterologous antigen is inbibited by both the homologous and the
heterologous glucosides.  Thus a difference in the stereochemical
configuration of the carbon atom bearing the aglucon sufkes to deter-
mine the immunological specificity of the a! and ~3 derivatives of
glucose irrespective of the protein with which they are combined.
The identity in structure of the remaining five carbon atoms of both
these glucosides may account for the cross-reactions between the anti-
sera and the heterologous hexoside-protein antigen.  If the structural
identity of the terminal portion of the hexoside molecule be altered
through the introduction of a chemical grouping such as the acetyl
radical, then one might anticipate that the immunological specificity
of the altered glucoside might well be different from that of the same
glucoside in its unacetylated form.  That this phenomenon actually
occurs will be seen from the following experimental data.  For com-
parison the immunological properties of an antigen containing azo-
phenol a-glucoside are included.  The structural relationships of
the three hexosides, CY and 6 @minophenol glucosides and the acetyl
derivative of the latter, are illustrated by the following graphic
formulae :


w. F. GOEBEL, F. R. BABERS, AND 0. T. AVERY     87

     CHzOH
fl~#i-I>G&m2 l<H $-
     H        OH                  H        OH
    p-aminophenol ar-glucoside         p-aminophenol @-glucoside
                 CHZOCOCHI

H         OH

+nninophenol &giucoside monoacetate

EXPERIMENTAL

I. CHJIMICAL

1. p-Nilropkenol j%GluGoside MpnoM;et&.-10 gm. of p-nitrophenol &glucoside
were dissolved in 30 cc. of anhydrous pyridine.  The solution was cooled to 0"
and 2.61 cc. (1.1 mols) of acetyl chloride were slowly added.  The mixture was
kept cold for several hours and then allowed to stand at room temperature for 3
or 4 days, The pyridine was removed by distillation in vaczM and the residual
syrup evaporated several times with absolute ethyl alcohol.  The mixture was
allowed to stand for 24 hours at ice box temperature to facilitate crystallixation.
Crystals of p-nitrophenol B-glucoside monoacetate were filtered from the mother
liquors and washed with cold ethyl alcohol.  3.4 gm. were recovered.  The com-
pound was recrystallixed several times from ethyl alcohol.  The substance melted
at 202-205oC. (uncorrected).

Analysis: 4.863 mg. substance gave 8.410 mg. COO and 2.165 mg. &O.
      CrdHt709N  Calculated: C 49.00 per cent; H 5.00 per cent.
              Found:   C 49.19 per cent; H 5.19 per cent.
8.236 mg. substance when analyzed by Pregl's (3) method for the determina-
tion of acetyl groups, utilized 1.65 cc. ~/70 NaOH.
       C&CO- Calculated: 12.53 per cent.
              Found:   12.32 per cent.

14: - - 2.33' X 100
    2 x 10 x 0.1102 = - 103.7' (in methyl alcohol)

3. p-Aminophemd &Glucoside Monoacetate.-2.0 gm. of p-nitrophenol &gluco-


88.      CONJUGATED CARBOHYDRATE-PROTEINS. VIII

                    I

side monoacetate were dissolved in 100 cc. of warm methyl alcohol.  The suh-
stance was reduced catalytically with SO mg. of platinum oxide and hydrogen
according to the method of Vorhees and Adams (4).  From the alcoholic solution
1.2 gm. of paminophenol fl-glucoside monoacetate were isolated.  Thii compound
crystallized as a snow-white product diflicultly soluble in ethyl alcohol, more solu-
ble in methyl alcohol, and much more easily soluble in water.  It melted at 177-
179Y. (uncorrected) and had a specific optical rotation of

,I;* =  - 1.69" X 100
    2 x 10 X 0.1204 = - 702" (in methyl alcohol)

Analysis: 8.806 mg. substance when analyzed for acetyl groups, utilized 1.94
cc. ~/70 NaOH.

     C&CO- Calculated: 13.75 per cent.
            Found:   13.55 per cent.
32.3 mg. sample when analyzed for nitrogen by the micro Rjeldahl method
utilized 6.87 cc. ~/70 HCl.
       N    Calculated: 4.47 per cent.
            Found:   4.26 per cent.    '

The monoacetate of kaminophenol fi-glucoside is relatively stable to alkaline
hydrolysis.  Aqueous solutions must be warmed (40-50") with ~/lo alkali before
the acetyl group begins to hydrolyze.
3. FAmino$henol a- and &Glucosides.-These compounds were prepared by
methods previously described (2, 5).
4. Preparation of Protein Aeophenol a-&code, fl-Glucoside, and b-Gkoside
Monoacetate A digem-

(a) Immunizing Antigens.-The amino glucosides were combined with the glob-
ulin of normal horse serum in the manner previously described (2).  The solu-
tions of sugar-aao protein were sterilized by filtration through a Berkefeld candle
and were used for immunization of rabbits.
(b) Test Antigens.-In order to avoid the possibility of antiprotein precipitins
masking the specificity of the anticarbohydrate reactions, the test antigens used in
the precipitin reactions were prepared by combining the respective glucosides to
the proteins of chicken serum.

II. IMMUNOLOGICAL

Methods

Rabbits were immunized by the intravenous injection of solutions of the con-
jugated bexoside-protein antigens.  The rabbits of one series received the antigen
composed of horse serum globulin coupled to diazophenol a-glucoside, those of
another series were treated with the same protein combined with diazophenol &
glucoside, while those of a third series received the protein in combination with
the acetyl derivative of the &glucoside.  The rabbits of all three series were in-


W. F. GOEBEL, F. H. BABERS, AND 0. T. AVERY      89

jwtd intraveIl0UdY with 1 CC. Of the respective antigen, containing 10 mg. pro-
tein per CC., day for six doa=. The course of injections ~a.5 repeated after a rest
period of 1 week.  8 daya fouqhg the fast injection the rabbits were bled and
ha aera tested for precipitina a@st the homologous and heterologoua teat anti-
gens. The technique of the precipitin and inhibition teats ia the same. as that
described in the previous studies (6).

Carbohydrate Antibodies

1. Homologom Precipitis Reactions. -The sera of rabbits immunized
with the carbohydrate-protein antigens were first tested for the pres-
ence of homologous precipitins against test antigens containing the
homologous glucosides combined with the protein of chicken serum.
The precipitin reactions between the homologous antisera and the
corresponding test antigens are summarized in Table I.  The results

                      TABLE I
%onwLogo2ls Precipitins in Sera of Rabbits fmmzcni&cd with ~-GlucpGlobu&n,
      &Glabco-Globulin, and A&y1 &Ghco-Globu&~

1. a-Glucwglobulin
2. &Gluceglobulin
3. Acetyl ,%gluco-globulin

A&.XlD

                Dilatioa of tat antigen
    Test antipa
                1:5,000  1:1o*cKm 1:2o,ooo
                 ---
a-Ghm-chicken serum  +++A +++A ++&
&Gluco-chidrenserum ++++ ++-t& +++
Acetyl fl-gluco-chicken   +++ ++* ++
serum

+ + + + = Complete precipitation with disk-like precipitate.

of the homologous precipitin tests illustrate the property of the im-
mune serum to react with an antigen containing the homologous glu-
coside, irrespective of the protein with which it is combined.  It is
further evident that the diaxophenol glucosides when combined with
protein function as excellent antigens.

2. Hete?oZogous Precipitin React&m.-In order to ascertain whether
the antisera interact with the different carhohydmte antigens, the
respective sera were tested against each of the heterologous test
antigens.  The results are summarized in Table II.  From the results
given in Table II, it may be seen that the immune serum obtained by
immunization with an antigen containing the czglucoside gives rise to


90      CONJUGATED CARBOHYDRATE-PROTEINS. VIII

antibodies which cross-react with a test antigen containing the tm-
acetylated &glucoside, `but not with one containing this glucoside in
the acetylated form.  On the other hand, an antiserum produced by
immtmization with p-gluco-globulin reacts not only with the homol-
ogous test antigen, but with cr-gluco- and acetyl fl-gluco-antigens as
well.  If one of the hydroxyl groups of the immunizing antigen con-
taining the B-glucoside is covered with an acetyl group, then the
homologous antiserum, although it still reacts with a fl-gluco-antigen,
fails to react with the at homologue.  The chemical basis underlying

TABLE II

Hcterologolcs Precifitins in Sera of Rabbits Immunized with a-Gluco-Globdin,
       &Gluco-Globdin, and Acetyl fi-Gluco-Globulin

                                     Dilution of test antigen
      Antiserum            Test antigen
                                      1:5.000   1: 10,000  1: 20,ooo
                                        ---
              u-Gluco-chicken serum     +++* ++* +*
1. a-Gluco-globulin     #-Gluco-chicken serum     +++ +* +
            Acetyl &gluco-chidren serum    0     0      0
                                       ---
              aGluco-chicken serum           +* +
2. /?-Gluco-globulin     @-Gluco-chicken serum  +E* +++ ++
            Acetyl &gluco-chicken serum ++=t ++ +A
                                        ---
              arGluco-chicken serum        0      0     0
3. Acetyl B-gluco-globulin  p-Gluco-chicken serum      +* + +
            Acetyl fi-gluco-chicken serum ++ZIZ ++ ++

+ + + + = Complete precipitation with disk-like precipitate.
0 = No precipitation.

the differences in the immunological properties of these glucosides
will be discussed later.

3. Spec$c Inhibition Tests.-The selective specificity of each of the
gluco-protein antigens is clearly demonstrated by the results of the
inhibition tests which are given in Tables III, IV, and V.  Analysis
of the data presented in Tables III, IV, and V shows in each instance
that upon addition of the corresponding glucoside to its homologous
immune serum the specific precipitins are completely bound and are
no longer capable of reacting with either the homologous or heterolo-


W. F. GOEBEL, F. H. BABERS, AND 0, T. AVERY      91



              TABLE III

spcc,jic Inhibition of PrecipilinS in AcetyE &GlucoCEobulin Antiserum by Homol-
           ogous and Heterologous Glucosi&~*

=
I

Tot No. CL

1     0.2
2    0.2
3    0.2
4    0.2
5    0.2
6    0.2
7    0.2
8    0.2

jwunhopheno;gl&da

A&y1 6

B

a

cc.

cc.

-

-

-

-
-

-

-

-

-

0.3
0.3

cc.


-
-

0.3
0.3
-
-
-
-

0.3
0.3
-
-
-
-

-
0.3    -           -

0.3    -          -

cc. I

0.5
-
0.5
-
0.5
-
0.5
-

-   ++++
0.5    ++
-      0
0.5     0
-    +++
0.5      0
-   +++*
0.5    +*

* a-Gluco-chicken serum antigen is not included in this table since it fails to re-
act in acetyl #I-gluco-globulin arh~er~m (cJ Table II).

TABLE IV

Sgecific Inhibition of Precipitins in ar-Ghuo-GEobuJin Antiserum by Homologous
           and Heterologous Glwosi&s*

B

cdatco-
chicLen

Test No.

cc. I

        1 !j      +t+
                        -
  -L          -
                                       0.5
                0.3 - B           -
             0.3 - d      +J-+
                                 0.5 ++
          L          L              L
* Acetyl &gluco-chicken serum antigen is not included in this table since it fails
to react in cy-gluco-globulin antiserum (cf. Table II).

tc.


0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2

cc.


-
.-
0.3
0.3
-
-
-
-

cc.


-
-
-
-
0.3
0.3
-
-

cc.


-
-
-
-

                      -   +++-I-
                  0.5    ++*
                      -      0
                  0.5
-

  %;;e
0.3 ,p
0.3 9
- 3
- F1

cc.


0.5
-
0.5
-
0.5
-
0.5
-


92 s      CONJUGATED CARBOHYDRATE-PROTEINS. VIII

gous test antigens.  The specificity of the inhibition reaction is demon-
strated by the fact that a heterologous glucoside does not inhibit the
homologous precipitin reaction. The acetylated &glucoside antigen
induces an immune response which is specifically distinct from that
induced by an antigen containing this glucoside in its unacetylated
form, for it is seen that the unacetylated /3-glucoside fails to inhibit
the reaction between the acetyl j3-glucoside antigen and its homologous
antibody (Test 5, Table III). This fact indicates clearly that the

TABLE V

.!?pe&c Inhibitivn of Precipitins in fl-Gluco-Globdin Antiserum by Hmologous and

Hekrologous Glucosides

U-G~iuctdii

Tat No

1
2
3
4
5
6
7
8
9
10
11
12

cc.


0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2

=

1

-




_-









-

bAminopho1 tducoside
     Y/IO

Q



cc.


-
-
-
-
-
-
0.3
0.3
0.3
-
-
-

B

ACCtyl
B

cc.

-
-
-

0.3
0.3
0.3
-
-
-
-
-
-

cc.

-
-
-
-
-
-
-
-
-
0.3
0.3
0.3

cc.


0.3
0.3
0.3
-
-
-
-
-
-
-
-
-

Test a&ens (1: 5.000)

g$c;
sauttt


I%.

0.5
-
-

0.5
-
-
0.5
-
-
0.5
-
-

-

L%ACCt&

f `&en
E


cc.

-
0.5
-
-
0.5
-
-
0.5
-
-

0.5
-

a-Gluco-
chicken
senlm

cc.

-
-
0.5
-
-

0.5
-
-
0.5
-
-

0.5

++++
+++
-t+
  0
  0
  0
+++i
++
  0
+++*
  0
+*

acetyl group confers a distinct and additional specificity upon a simple
carbohydrate the stereochemical structure of which remains unaltered.
From the results in Table IV, it is seen that the acetyl fl-glucoside
does not inhibit the reaction of the unacetylated P-gluco-test antigen in
cu-gluco-globulin antiserum (Test 8).  In Table V it may also be seen
that the reaction between &gluco-globulin antiserum and the homol-
ogous antigen, although inhibited by the /I-glucoside (Test 4), is
not inhibited by the same glucoside when one of the hydroxyl groups
has been covered by an acetyl radical (Test 10).


\V. F. GOEBEL, F. H. BABERi, AND 0. T. AVERY      93

DISCUSSION '

The results of the present study not only confirm the.view previously
held that the immunological specificity of carbohydrates is determined
by their stereo'khemical configuration, but they lend support to the
further assumption that the introduction of a simple chemical group,
such as the acetyl radical, endows a carbohydrate with a new and dis-
tinct specificity which is determined by the chemical nature of the
group thus introduced. It has been previously pointed out that differ-
ences in the specific behavior of cy- and &glucosides of glucose may
be accounted for by differences in the stereochemical configuration of
the carbon atom bearing the aglucon, and that the basis for the im-
munological crossing may lie in the fact that the spatial configuration
of the polar groups on the remaining five carbon atoms is identical.
This explanation appears to be further supported by the results of the
present study. For it can be seen from Table II that when /3-gluco-
test antigen, which normally reacts in cr-gluco-globulin antiserum,
is so altered that one of the polar groups (OH) of the five terminal
carbon atoms of the carbohydrate radical is replaced by acetyl
(CHJZO), the resulting antigen fails to react in cy-gluco-globulin
antiserum.

Furthermore, a-gluco-test antigen, which normally reacts w'ith /3-
gluco-globulin antiserum, fails to react in the immune serum produced
by immunization with acetyl @-gluco-globulin.  Similarly, the acety-
lated fi-glucoside, due to the alteration in chemical constitution, fails
to bind the antibody in ar-gluco-globulin antiserum, and as a result per-
mits both the cy- and @-gluco-test antigens to react in their normal
course. The 8-glucoside liiewise fails to inhibit the reaction between
the acetyl /3-gluco-test antigen, and homologous immune serum.  This
difference in the serological speciscity of the antibodies induced by
fl-gluco-antigens in the acetyIated and unacetylated form can be at-
tributed only to the k!nown Werences in the chemical structure of these
two glucosides. The latter diEer from one another in that the acetyl
&glucoside is a derivative in which one of the polar groups (OH) of
the carbohydrate has been altered by the introduction of an acetyl
radical (CHgCO).

A dhl analysis of the results of these serological tests again
emphasized the fact that the presence of an acetyl group in a carbohy-


94      CONJUGATED CARBOHYDRATE-PROTEINS. VIII

drate exerts a determining influence on the specificity of an antigen of
which it forms a part.  In conclusion it may be pointed out that the
differences in serological specificity exhibited by the acetylated and
deacetylated polysaccharides of Type I Pneumococcus.are accurately
paralleled by the purely synthetic system described.

SUMMARY

The chemical and immunological properties of the acetylated and
unacetylated forms of the p-aminophenol &glucoside of glucose have
been described. The serological specikity of these /3-glucosides in
combination with protein has been correlated with known changes in
chemical structure and has been compared with the immunological
properties of the cu-glucoside of the same hexose.

BIBLIOGRAPHY

1. Avery, 0. T., and Goebel, W. F., J. I&#. &fed., 1933,68, 731.
2. Goebel, W. F., Babers, F. H., and Avery, 0. T., J. Exp. Med., 1932,66,761,769.
3. Pregl, F., Quantitative organic microanalysis, Philadelphia, P. Blakiston's

Son and Co., 1930, 197.

4. Vorhees, V., and Adams, R., J. Am. Ciaem. Sot., 1922,40,1397.
5. Goebel, W. F., and Avery, 0. T., J. Exp. Med., 1929, 60,521.
6. Avery, 0. T., and Goebel, W. F., J. Exf. Med., 1929, 60, 533.