Alpha-amanitin-concanavalin A conjugates as inhibitors of specific cell types


Material Information

Alpha-amanitin-concanavalin A conjugates as inhibitors of specific cell types
Physical Description:
vi, 159 leaves : ill. ; 28 cm.
Hencin, Ronald Stephen, 1949-
Publication Date:


Subjects / Keywords:
Conjugation (Biology)   ( lcsh )
Cytology   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis--University of Florida.
Includes bibliographical references (leaves 151-158).
Statement of Responsibility:
by Ronald Stephen Hencin.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000096798
notis - AAL2231
oclc - 06457781
System ID:

Full Text








The author sincerely wishes to acknowledge the support

of the numerous individuals who contributed to this under-

taking. Without their individual efforts on my behalf, a

difficult task would have been arduous at best.

The faculty of the Department of Microbiology and

Cell Science have been without exception generous of their

time, facilities and expertise, all of which were frequently

sought and received. I am truly grateful for having had

their help and teachings, in particular the members of my

graduate committee: Drs. Ed Hoffmann, Bill Clem, Ken

Noonan and Jim Preston.

Dr. Preston, as my major professor who unsuspectingly

opened his lab to me, contributed beyond the limits defin-

able in this simple statement. In the final analysis, it

is his patience that sustained these efforts through com-


The understanding of my personal friends and family

has been essential to my well being. Their confidence in

me throughout reassured and stabilized my goals.

Thank all of you.



ACKNOWLEDGEMENTS........ ............................. ii

ABSTRACT.... ....... ................ ................. iv

SECTION I INTRODUCTION......... ................... 1

Objectives.................................. 1
Background... ............................... 2
Rationale......................................... 14


Modification of a-Amanitin for
Conjugation to Proteins.................... 17
Synthesis of a-Amanitin-Protein
Conjugates ............................... 21
Biochemical Characterization of
Conjugates ................................ 28
Interaction of Conjugates with
Cultured Cells.............................. 32

SECTION III RESULTS................................. 42

ADH-BSA Conjugates.......................... 42
ADH-Con A Conjugates........................ 68
ADGG-BSA Conjugates......................... 75
Hippuric Acid-Con A Conjugates.............. 85
ADGG-Con A Conjugates....................... 101

SECTION IV DISCUSSION.............................. 132

ADH-BSA Conjugates............................ 132
ADH-Con A Conjugates........................ 135
ADGG-BSA Conjugates......................... 136
Hippuric Acid-Con A Conjugates............. 137
ADGG-Con A Conjugates........................ 140

REFERENCES............................................ 151

BIOGRAPHICAL SKETCH................................... 159


Abstract of Dissertation Presented to the Graduate
Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree
of Doctor of Philosophy




August, 1979

Chairman: James F. Preston
Major Department: Microibology and Cell Science

Macromolecular conjugates of the fungal toxin a-

amanitin and concanavalin A (Con A) were used to evalu-

ate the ability of Con A to impart selectivity with

respect to cellular uptake to a-amanitin. The efficiency

of targeting to specific cellular receptors via Con A

saccharide binding sites, the degree to which these re-

ceptors could facilitate entry of toxin into the cell,

and determination of resultant a-amanitin mediated cyto-

toxicity comprise the major objectives of the study.

Covalent a-amanitin-Con A conjugates prepared with

a free carboxyl group containing derivative of a-amanitin

(ADGG) via reaction with carbodiimides contained an aver-

age of 3.6 moles ADGG per mole Con A. They retained a

high degree of binding specificity and affinity for calf
thymus RNA polymerase II in vitro (K = 27 x 10 M versus
6.9 x 10-9M for free ADGG). The ADGG-Con A conjugates
6.9 x 10 M for free ADGG). The ADGG-Con A conjugates

absorbed to Sephadex G-75 and eluted with specific sac-

charide in volumes identical to native Con A. The conju-

gates agglutinated red blood cells at near equivalent con-

centrations to Con A. Evidence from other Con A conjugates

substituted with hippuric acid as an analog for ADGG indi-

cate that the ADGG-Con A conjugates retain full lectin

associated ligand binding specificities.

Exposure of H-7 CHO cells for short periods of time

to ADGG-Con A resulted in 50% inhibition of cell prolifer-
ation at 2.1 x 10 M with respect to ADGG concentration.

At that concentration, free ADGG or the equivalent amount

of Con A to that present in the conjugate had no effect.

Extrapolation from the maximum doses of ADGG used indicates

that a minimum 50-fold enhancement of the toxicity of free

ADGG is achieved by conjugation to Con A. Cell binding

measurements determined that ADGG-Con A is bound by H-7
CHO cells in identical amounts to 1I-Con A, that it
competes for 1I-Con A binding sites as efficiently as

native Con A, and that its binding is inhibitable by pre-

incubation of cells with Con A or by the presence of the

Con A specific ligand, a-methyl-D-mannopyranoside (MDM).

The potent toxicity of ADGG-Con A for H-7 CHO was abolished

by treatment of the conjugate with MDM prior to exposure

of the cells. These data strongly suggest that ADGG-Con A

conjugates derive their enhanced toxicity for CHO cells

from the binding to specific cell receptors with subsequent

internalization of the bound a-amanitin. Confirmation of

this hypothesis was obtained from the interaction of a-

amanitin resistant rat fibroblasts (LAN-2) with ADGG-Con A.

LAN-2 possess an altered RNA polymerase II with greatly

reduced affinity for a-amanitin. Although LAN-2 bind Con

A as well as their parent cell line, they were refractory

to the toxic effects of ADGG-Con A.

These experiments demonstrate that covalent conjugates

of Con A and a-amanitin retain the biological activity of

both portions of the conjugate. Furthermore, the data

clearly indicate that Con A permits efficient and specific

targeting of amanitin to cells as a function of interaction

with specific saccharide receptors on the cell surface.



The overall objectives of this research were to assess

the potential of selected proteins for targeting of inhi-

bitors to specific cells and to evaluate the ability of

cell membrane receptors to mediate uptake of these protein-

inhibitor conjugates. The specific objectives defined for

accomplishing this were: 1) synthesis and biochemical

characterization of conjugates of a-amanitin and bovine

serum albumin (BSA) and concanavalin A (Con A), 2) deter-

mination of the degree to which Con A conjugates retain the

biological properties of each component of the conjugate,

and 3) the use of a-amanitin-Con A conjugates to investi-

gate the effectiveness of Con A as a targeting vehicle

for a-amanitin based on the numbers and/or affinities of

Con A receptor sites present on the target cell. The

effectiveness of Con A for imparting selectivity with

respect to cellular uptake to a-amanitin and the ability

of cell surface receptors to mediate the entry of bound

toxin and thereby be specifically killed, were used as

criteria for assessing the degree to which these objec-

tives were fulfilled.


Alpha-amanitin is one of a group of fungal peptide

toxins isolated principally from members of the fungal

genus Amanita (Wieland, 1968). The structures of the

amanitins have been rigorously established as bicyclic

octapeptides containing two moles glycine, and one mole

each of hydroxytryptophan, cysteine, hydroxyproline, aspar-

agine (or aspartic acid), isoleucine and a hydroxylated

isolencine, which varies depending on the derivative

(Wieland and Wieland, 1972; Fiume and Wieland, 1970;

Wieland and Faulstich, 1978). The sulfur atom of the cys-

teine residue is connected to the indol moiety of the tryp-

tophan ring via a sulfoxide bridge dividing the molecule

into two rings. Alpha-amanitin, the most common derivative,

contains asparagine and dihydroxyisoleucine and because

of its availability is the most thoroughly investigated of

the amanitins.

The amanitins have been proven to possess extreme

toxicity for mammalian cells, in vitro and in vivo (Wieland

and Wieland, 1959; Fiume and Wieland, 1970; Sekeris and

Schmidt, 1972). Amanitins derive their toxicity from the

ability of the toxins to bind with high affinity to eucary-

otic DNA-directed RNA polymerase II (Stirpe and Fiume, 1967;

Seifart and Sekeris, 1969; Lindell at al., 1970; Jacob et

al., 1970). The specificity and high affinity of the inter-

action of amanitins with polymerase is reflected by the

inhibition constant (K ) which was determined for calf
thymus RNA polymerase II to be 3-10 x 10 M (Cochet-

Meilhac and Chambon, 1974). The amanitin mediated inhi-

bition of RNA polymerase II results in cell death due to

the lack of messenger RNA (mRNA) synthesis. RNA polymer-

ases II are those polymerases identified as being respon-

sible for most of the mRNA synthesis in eucaryotic cells

(Chambon, 1975). Evidence for this is derived from several

observations. The synthesis of poly-adenosine containing

heterogenous nucleoplasmic RNA (hnRNA) by HeLa cells is

inhibited by low concentrations of a-amanitin (Zylber and

Penman, 1971). HnRNA is a precursor in eucaryotic cells

to those species of RNA that have been identified as func-

tional mRNA (Brawerman, 1974). Low concentrations of a-

amanitin have also been shown to inhibit the synthesis of

adenovirus mRNA in human KB cells (Ledinko, 1971) but do

not affect the synthesis of mRNA by viruses which use virus-

coded RNA-dependent RNA polymerases that are amanitin

resistant (Wieland and Faulstich, 1978).

Mammalian cell lines resistant to the effects of a-

amanitin have provided additional verification of the

molecular target of a-amanitin. Isolation and purification

of RNA polymerase II from a-amanitin resistant Chinese

hamster ovary (CHIO) cells yielded an enzyme resistant to

a-amanitin (Chan et al., 1972). Amanitin resistant mutants

of BIIK-T6 hamster cells (Amanti et al., 1975), rat myoblasts

(Somers et al., 1975), mouse myeloma MOPC 104 E cells

(Wulf and Bautz, 1976) and human diploid fibroblasts

(Buchwald and Ingles, 1976) have all demonstrated similarly

resistant polymerase II molecules as being responsible for

the observed a-amanitin resistance. Thus a-amanitin can

be seen to be a structurally well characterized toxin that

possesses an extremely specific mechanism of action. The

properties of high specificity, known mechanism of action

and low concentration at which it is inhibitory make a-

amanitin an ideal inhibitor for examining targeting of

inhibitor conjugates to specific cell receptors.

Conjugates of the amanitins and macromolecules were

first prepared as an attempt to produce antibodies to 8-

amanitin as a haptenic substituent linked to rabbit serum

albumin (RSA) (Cessi and Fiume, 1969). Beta-amanitin,which

contains a free carboxyl group, was directly coupled to RSA

by reaction with water soluble carbodiimides. The conju-

gates not only failed to elicit antibody production, but

were found to be approximatley 10-fold more toxic to mice

than free B-amanitin (Cessi and Fiume, 1969). The conju-

gated B-amanitin retained its specificity of interaction

with RNA polymerase II but with a reduced binding affinity

(Fiume et al., 1971; Derenzini et al., 1973). Furthermore

the enhanced in vivo toxicity of the B-amanitin conjugates

was determined to result from increased uptake by the sinu-

soidal cells of the liver and the proximal tubule cells of

the kidney, presumably because of the protein portion of

the conjugate (Fiume, 1969; Fiume et al., 1969). Cultured


cells with high rates of protein uptake, e.g. macrophages,

were preferentially killed by conjugated 3-amanitin

(Barbanti-Brodano and Fiume, 1973; Barbanti-Brodano and

Fiume, 1974; Fiume and Barbanti-Brodano, 1974). Further

coupling of B-amanitin-albumin conjugates to fluorescein

enhanced the conjugate toxicity for hepatocytes, cells

known to possess fluorescein receptors (Fiume et al., 1971).

Although very qualitative in nature, this work demonstrated

that certain conjugates of B-amanitin were preferentially

taken up by cells on the basis of the macromolecular por-

tion of the conjugate. Subsequent cell death resulted

presumably as a result of transport of the conjugate, or

minimally, the amanitin portion to the nucleus where inhi-

bition of mRNA synthesis occurred (Faulstich et al., 1975).

However, it is likely that the enhanced toxicity of these

8-amanitin conjugates is a relatively non-specific phe-

nomenon based upon pinocytosis by cells with a high rate

of protein uptake. The carbodiimide procedure used for

coupling can result in protein cross-linking and ester

formation, increasing the molecular weight of the conjugate

and decreasing the stability of the linkage (Carraway and

Koshland, 1972; Timkovich, 1977). Increased molecular

weight and decreased stability of binding of the inhibitor

and carrier protein would be apt to favor pinocytic uptake

and degradation of the conjugates.

Other conjugates of p-amanitin and proteins have been

synthesized by formation of the hydroxysuccinimide ester

of (j-amanitin which is then directly linked to amino groups

on proteins (Faulstich et al., 1975). This procedure should

not have the carbodiimide associated side reactions and

would lead to well defined, covalently linked conjugates.

To date, no additional work beyond the initial publication

has been presented to define the characteristics of these

conjugates. Furthermore, B-amanitin is a relatively minor

component of the naturally occurring amatoxins and is not

readily available.

Conjugates of the more widely available a-amanitin

were first prepared by the modification of a-amanitin to a

derivative containing a free amino group which can be

coupled to proteins via carbodiimides (Faulstich and

Trischmann, 1973). These derivatives when conjugated to

BSA were shown to be equally as toxic in vivo as free a-

amanitin. In vitro inhibition of calf thymus RNA polymerase

II by the a-amanitin-BSA conjugates was 20-fold less than

free a-amanitin. These studies as well as the previously

described investigations with 8-amanitin conjugates pointed

out the feasibility of using amanitin-protein conjugates for

exploring the targeting of inhibitors to specific cells.

They also underlined the need for careful, quantitative

evaluation of the interaction of a well characterized con-

jugate with the cellular target.

Investigations of drug targeting with inhibitors other

than amanitins have primarily centered on increasing the

specificity of uptake of an inhibitor by conjugation with

immunoglobulins directed against cell membrane constituents.

The inhibitors used are generally non-specific in their

mode of action and without some means for increasing the

selectivity of uptake they are unable to inhibit a specific

subpopulation of cells within a larger population. Without

selectivity, they are of limited use for differentially

inhibiting the growth of transformed cells, the primary

goal of drug targeting research. Various low molecular

weight inhibitors have been coupled to protein carriers.

Most of these inhibitors have as the basis for their toxi-

city interaction with nucleic acid structure or synthetic

processes. 5-Flurodeoxuridine (FUDR) and albumin were

coupled (Barbanti-Brodano and Fiume, 1974), producing con-

jugates that inhibited transformed 3T3 fibroblasts in vitro

almost as well as free FUDR but did not inhibit non-dividing

macrophages. High molecular weight conjugates of methotrexate

and albumin were effective in prolonging the half life of

the drug in vivo but displayed no selectivity in uptake by

cells (Chu and Whiteley, 1977). Triaziquone conjugated to

y-globulin or albumin was toxic to polyoma transformed baby

hamster kidney (BHK) cells. Inhibition of pinocytosis did

not affect the toxicity of the conjugates, suggesting mem-

brane mediated uptake of the drug. However, normal BHK

cells were 3-fold more susceptible to the conjugates than

the transformed cell (Linford and Froese, 1978). Addition-

ally, no actual binding of the conjugate was detectable by

fluorescein labeling which would imply a distinct lack of

specificity in the uptake of conjugates. Cytosine ara-

binoside-albumin conjugates were effective in inhibiting

viral replication within mouse liver cells, suggesting

retention of toxicity after conjugation. However, the high

molecular weight complexes that resulted from the synthesis,

make non-specific endocytosis a likely route for uptake of

these conjugates (Balboni et al., 1976). Greater specificity

of uptake and apparent toxicity was obtained with conjugates

of p-phenylenediamine mustard and anti-lymphoid cell anti-

bodies (Davies and O'Neil, 1977). Other conjugates of

chlorambucil and antibodies against specific cell types

displayed selective toxicity and some degree of chemothera-

peutic potential, in vivo (Ghose et al., 1972). Anti-

Ehrlich ascites tumor cell antibody-chlorambucil conjugates

also were effective in treatment of the neoplasia (Flechner,

1973). However, in similar studies, Rubens and Dulbecco

(1974) have demonstrated a lack of covalent association of

chlorambucil with the antibodies, making it difficult to

ascertain the exact mechanism of the observed toxicity.

Perhaps the most effective conjugates of immunoglo-

bulin and inhibitors that interact with DNA have been

conjugates of daunomycin and anti-mouse lymphoid cell

tumor antibodies (Levy et al., 1975). These conjugates

were examined for both inhibition of specific tumor cell

growth in vivo and for in vitro inhibition of tumor cell

RNA synthesis. In comparison to daunomycin linked to

non-specific immunoglobulin, the anti-tumor antibody-daunomycin

conjugates were significantly more effective in inhibiting

specific tumor cell processes. These conjugates represent

a considerable improvement over those previously discussed

in that their synthesis is a result of the interaction of

a single reactive group on daunomycin with the immunoglo-

bulin. This results in well defined covalent conjugates

that retain the targeting specificity of the immunoglobulin.

The primary disadvantage to their use as chemotherapeutic

agents is that they interact with DNA rather than a critical

enzyme required for macromolecular replication, transcrip-

tion or translation and would thus require much larger doses

to achieve a given cytotoxic effect. This presents a signi-

ficant problem in view of the fact that the inhibitors will

act on normal as well as neoplastic cells with potentially

detrimental side effects.

Conjugates of diphtheria toxin and immunoglobulins

directed against cell surface antigens have proven to be

selective as well as effectively cytotoxic. When coupled

to immunoglobulin with specificity directed against mumps

virus infected monkey kidney cells, the diphtheria toxin

conjugates were selectively toxic to virally infected

cultures (Moolten and Cooperband, 1970). Other conjugates

of diphtheria toxin and anti-DNP antibodies exhibited selec-

tive toxicity in vivo against hapten (DNP) coated tumor

cells in hamsters (Moolten et al., 1972). Hamster lymphoma

growth was effectively suppressed by conjugates of diph-

theria toxin and antibodies prepared against SV-40

transformed cells (Moolten et al, 1976). IIeLa cells that

had been coated with hapten (2,4,6-trinitrophenyl sulfuric

acid) were killed by diphtheria toxin-antibody conjugate

only when the antibody was hapten specific (Philpott et al.,

1973). Mouse anti-lactate dehydrogenase antibody when

coupled to diphtheria toxin resulted in a conjugate more

toxic for Erhlich ascites cells than for normal mouse

kidney cells due to increased expression of lactate dehy-

drogenase on the ascites cells (Samagh and Gregory, 1972).

Thorpi et al. (1978) coupled purified antilymphocytic

globulin-chlorambucil conjugates to dephtheria toxin via

an activated anhydride reaction that yielded well defined

conjugates which could readily be purified by gel filtration.

The conjugates were significantly more toxic to cultured

lymphoblastoid cells than was the free toxin. The con-

jugates overcame a significant problem with the synthesis

of diphtheria toxin-protein conjugates described above.

Dephtheria toxin is a protein macromolecule which is readily

cross-linked during most of the procedures used to produce

conjugates. This results in ill-defined preparations of

varying compositions and specificities.

The advantage to using dephtheria toxin resides in its

mechanism of toxicity. Diphtheria toxin is a potent inhi-

bitor of eucaryotic protein synthesis by virtue of its

catalytic ADP-ribosylation of elongation factor, EF-2

(Collier, 1975).

Chang and Neville (1977) synthesized conjugates of

diphtheria toxin and human placental lactogen that contained

equimolar ratios of toxin to carrier protein. The conju-

gates bound effectively to lactogen receptors on mammary

gland explants but these receptors were unable to mediate

entry of the toxin to the cell as the conjugate did not

inhibit protein synthesis in the target cells (Chang et al.,

1977; Neville and Chang, 1978).

Although previously described conjugates using albu-

mins and other proteins without defined cell binding activ-

ities resulted in some degree of enhanced cytotoxicity,

only those conjugates that could interact with specific

receptors on the target cell produced significant cell

specific toxicity. One other category of receptor specific

macromolecules besides antibody and hormones that may

mediate the targeting of inhibitors are lectins. Lectins

are proteins that possess specific binding sites for

carbohydrate moieties. They have been isolated from a

wide variety of plants and animals and possess a broad

range of individual binding specificities. Their specific

interaction with cell surface glycoproteins has been thor-

oughly documented and leads to an array of complex cellular

responses including agglutination, mitogenic stimulation

and cell toxicity (Sharon and Lis, 1972; Lis and Sharon,

1973; Nicolson, 1974).

Perhaps the most widely investigated lectin is that

protein isolated from the jack bean, Concanavalin A.

Con A has been extensively studied and characterized with

respect to ligand binding specificity (Goldstein et at.,

1965; So and Goldstein, 1968), chemical structure and

properties (Edelman et al., 1972; Sharon and Lis, 1972)

and interactions with cell surfaces (Nicolson, 1974). The

effects of Con A on cells are widely varied and include

agglutination, induction of mitosis, alterations in cell

permeability and transport phenomena as well as cytotoxi-

city. The mechanisms of these complex interactions of

Con A are by no means clear but they all contain as a cen-

tral feature the binding of Con A to cell surface glycopro-

teins. Since the extent of the effect of Con A on a given

cell is related to the surface architecture, density and

mobility of the cell surface glycoprotein and glycolipid

constituents, it seems plausible that variations on Con A

receptors could lead to differential toxicity of Con A-

inhibitor conjugates.

The observations that Con A induces agglutination of

some virally transformed cells at a concentration of lectin

much lower than that required for agglutination of normal

cells (Inbar and Sachs, 1969; Burger, 1969), implied that an

inherent difference in membrane architecture between normal

and transformed cells may exist. Attempts to quantitatively

detect a difference in the number of Con A binding sites on

normal cells and their virally transformed counterparts met

with limited success (Cline and Livingston, 1971; Ozanne and

Sambrook, 1971). Differences, if any, were extremely small

and intrepretations were complicated by variation in

technique and cell lines. Using low temperature conditions,

Noonan and Burger (1973a) were able to demonstrate that in

the absence of endocytosis and with appropriate corrections

for surface area and volume differences, certain virally

transformed cells possessed 3 to 5 times the number of Con

A binding sites of a normal cell. Similar results were

obtained with normal cells at mitosis (Noonan et al., 1973),

after brief protease treatment (Noonan and Burger, 1973b) or

exposure of certain cell lines to dibutyrylcyclic AMP (Veen

et al., 1976). Under these conditions slight differences

in the numbers of Con A binding sites could be detected.

Although the relationship of the Con A binding to increased

agglutinability or the transformed state still remains

undetermined, the altered surface structure of transformed

cells may allow for differential interaction with Con A-

inhibitor conjugates in comparison to normal cells. Kitao

and Hattori (1977) tested a conjugate of Con A and dauno-

mycin for its ability to suppress the in vivo development

of Ehrlich ascites and L1210 cells. Prolonged survival

of the host in comparison to free daunomycin was obtained

following administration of the conjugate with either cell

type. While characterization of this conjugate was minimal

and the test results qualitative in nature, selective toxi-

city of Con A conjugates was implied. The use of Con A

as a targeting agent would therefore seem feasible based

on the known differences between cells with respect to

quantitative and qualitative variations in cell membrane

receptors for the lectin. Con A would also provide a means

for assessing the effects of conjugation on the biological

activities of the native lectin.


Alpha-amanitin presents several distinct advantages

over most inhibitors used for investigations of drug tar-

geting. The affinity of the interaction of a-amanitin with

eucaryotic RNA polymerase II make it an extremely potent

inhibitor. Moreover, a-amanitin is of low molecular weight

and after modification can be coupled via a single reactive

site to proteins. This will allow for reproducible produc-

tion of conjugates that have minimal structural interference

by the amanitin moiety. Work by other investigators has

clearly demonstrated that amanitin conjugates retain inhi-

bitory potential for RNA polymerase II. The fact that a-

amanitin inhibits an enzyme present in limited quantities,

approximately 104 RNA polymerase II molecules per cell

(Cochet-Meilhac et al., 1974), that is critical to survival

of the cell will allow relatively low doses of inhibitor to

generate a measureable cytotoxic effect. Radioactively la-

beled derivatives of (-amanitin may be synthesized and can be

used to quantitate the coupling to protein as well as the

binding to specific cells. Therefore, the initial phase of

this investigation consists of developing procedures neces-

sary for the synthesis of a-amanitin conjugates of defined

chemical nature that retain inhibitory potential for RNA

polymerase II.

For targeting of a-amanitin to specific cell receptors

Con A will be used. Con A is well characterized chemically

and,unlike specific immunoglobulins, is available in quan-

tity from commercial sources. The conjugation with a-ama-

nitin can be monitored by a number of biochemical and bio-

logical parameters for its effects on lectin activity.

Specific ligands with known binding affinities for Con A

are available, as are defined systems for evaluating the

interaction of Con A conjugates with cells.

Following synthesis and characterization of the a-

amanitin-Con A conjugates, the final aspect of the study

will be to evaluate the potential of Con A for targeting

a-amanitin to cells and the ability of Con A receptors to

mediate uptake of bound conjugate. Although it would

appear that the differences in Con A binding between normal

and virally transformed cells would represent a potential

system for discerning targeting differences of Con A con-

jugates, the conditions necessary to achieve quantitative

differences in the number of cell receptors and the low

magnitude of the observed differences, suggest other

approaches may be more productive. A cell line known

to possess a distinct number of Con A receptors, H-7 CHO,

will be used to examine the cytotoxicity of Con A-a-amanitin

conjugate in comparison to free a-amanitin. The study will

be restricted to evaluating whether specific interaction


of the conjugate with membrane receptors for Con A occurs

and if the interaction leads to endocytosis of the conju-

gate. The mechanism of toxicity (if any) of the conjugate

will be evaluated by the use of a Con A resistant CHO

mutant and an a-amanitin resistant rat fibroblast cell.

These cells will differentiate Con A cytotoxicity, antic-

ipated to be minimal with short exposure times and low

concentrations, from toxicity due to the amanitin portion

of the conjugate.


Modification of a-Amanitin for Conjugation to Proteins

Isolation and Purification

All chemicals used throughout these investigations

were of reagent quality and were used without further

purification unless noted otherwise. The primary commercial

suppliers were Scientific Products (Mallinkrodt) and Sigma

Chemical Co. Water for all procedures was deionized and

glass distilled.

The a-amanitin used for these studies was obtained

primarily from specimens of Amanita suballiacea collected

from the Gainesville, Florida,area. Individual specimens

were collected, identified on the basis of gross morpho-

logical characteristics according to published descriptions

(Miller, 1977; Murrill, 1941; Murrill, 1948) and pooled for

extraction of a-amanitin. After thorough washing with

running tap water, the specimens were coarsely chopped and

combined with methanol to an approximate final concentration

of 50% methanol (v/v). This crude slurry was shaken on a

rotary shaker at 100rpm and 22C for 24 hours, filtered

through Whatman #1 filter paper and the crude methanolic

extract was flash evaporated to a thick syrup. Ten volumes

of ice cold 100% methanol were added to precipitate

polysaccharides and the mixture was filtered through

scintered glass. The resulting filtrate was flash evapora-

ted to near dryness, resolubilized in water and extracted

three times with 3 volumes of anhydrous ether. The final

extract was made 50% with respect to methanol (spectral

grade) and subjected to chromatography on Sephadex LH-20

in 50% methanol. The a-amanitin containing fractions were

identified by TLC on silica gel G with methyl ethyl ketone:

methanol (1:1) and detection after spraying with 2% methanolic

t-cinnamaldehyde and exposing to HC1 vapors (Wieland et al.,

1949; Sullivan et al., 1965). The characteristic Rf and

violet color served to identify amanitin containing fractions.

The peak fractions were combined, flash evaporated to dry-

ness, resolubilized in water and chromatographed on Biogel

P-2 in water. After concentration of pooled a-amanitin

fractions from this step, final purification was obtained by

chromatography on Sephadex LH-20 with water. The resulting

a-amanitin was characterized as to purity on the basis of

TLC mobility with two different solvent systems (methanol:

methyl ethyl ketone (1:1), n-butanol:acetic acid:water

(4:1:1), ultraviolet (uv) and visible absorption spectra,

and nuclear magnetic resonance (NMR) spectroscopy. All

samples prepared proved identical by these criteria to a

crystalline a-amanitin standard obtained from Th. Wieland,

Max Planck Institute for Medical Research, Heidelberg,

Germany. These purification procedures were developed from

procedures previously published (Faulstich, et al., 1973;

Wieland, 1968) by other investigators. Additional details

of the experiments resulting in the final purification

procedure are contained in a manuscript currently in

preparation for publication (Preston et al., in preparation

for submission to Lloydia).


The absence of either a free carboxyl or amino group

on a-amanitin necessitated the development of methods for

modifying the basic amanitin structure to allow direct

coupling of a-amanitin to proteins. The basic concept was

derived from the work of Faulstich and Trischmann (1973)

in which a-amanitin was coupled via diazotization to an

aromatic group linked to a six carbon spacer molecule

containing a free terminal amino group. Their procedure

was used with some modification to produce a-amanitin deri-

vatives with free amino groups. A derivative containing a

free carboxyl group was obtained by using the diazonium

coupled spacer molecule approach with new procedures developed

by J. F. Preston (Preston and Hencin, 1979). Details of both

of these syntheses are presented below.

For production of a-amanitin derivatives containing ter-

minal amino groups, a-amanitin was diazotized to N-(4-amino-

benzoyl)-N'-BOC-hexamethylenediamine. Mono-BOC-1,6-diamino-

hexane was prepared according to procedures described by

Faulstich and Trischmann (1973). N-(4-nitrobenzoyl)-N'-BOC-

hexamethylenediamine was prepared by acylation of mono-BOC-

1,6-diaminohexane with 4-nitrobenzoyl chloride. Catalytic

reduction with hydrogen and palladium/BaSO4 of N-(4-nitro-

benzoyl)-N'-BOC-hexamethylenediamine yielded N-(4-amino-

benzoyl)-N'-BOC-hexamethylenediamine which was subjected to

diazotization with a-amanitin. In a typical reaction 34mg of

NaNO2 were added to 112mg of N-(4-aminobenzoyl)-N'-BOC-

hexamethylenediamine in 7ml of 50% acetic acid (5C) and incu-

bated 10 minutes at 5C. This diazonium cation containing

solution was combined with 65mg of a-amanitin dissolved in

4.5ml of pyridine (5C). After 15 minutes of reaction at 5C,

the mixture was flash evaporated to dryness (40C with vacuum

for rapid removal of pyridine and other solvents), resolu-

bilized in spectral grade methanol and applied to a column

of Sephadex LH-20 in methanol (4.25 x 95cm). The a-amanitin

derivative, a-amanitin-diazobenzoyl-N-N'-BOC-hexamethylene-

diamine (ADBH) was identified as a reddish-purple band

eluting in 0.54 column volumes by its typical mobility

during TLC in methanol: methylethylketone (Faulstich and

Trischmann, 1973). The ADBH was further characterized by

absorption and NMR spectra and stored at room temperature,

dried, in darkness until used (Hencin and Preston, 1979).

Free carboxyl group containing a-amanitin derivatives

were produced by diazotization of a-amanitin with p-amino-

benzoylglycylglycine (Preston and IIencin, manuscript in

preparation). p-Aminobenzoylglycylglycine was prepared

by catalytic hydrogenation of p-nitrobenzoylglycylglycine

with Pd/BaSO4. For coupling to a-amanitin, 30mg of p-amino-

benzoylglycylglycine dissolved in 1.0ml of 50% acetic acid

were cooled to 0.2C and 10mg of NaNO2 were added. Follow-

ing 10 minutes of occasional shaking, 18.5mg of a-amanitin

in 2.0ml cold pyridine were added. After 10 more minutes

at 40C the reaction mixture was dried in vacuo, resolubi-

lized in 4.0ml 80% methanol and chromatographed on Sephadex

LH-20 in 80% methanol. Identification of the resulting

product, a-amanitin-diazobenzoylglycylglycine (ADGG) was

achieved by TLC, uv and NMR spectroscopy.

A tritium labeled derivative of ADGG used in these

studies was prepared and purified by Dr. J. F. Preston. The
3 3
derivative, H-demethyl-ADGG ( H-DM-ADGG) was synthesized

by oxidation of ADGG with sodium periodate and subsequent

reduction with H-NaBH Purification by column chromato-

graphy on Sephadex LH-20 with various buffers to isolate the

product and remove exchangeable tritium resulted in a pro-

duct with a specific activity of 7.4 x 10 dpm/pmole. The

H-DM-ADGG was determined to be essentially free from con-

taminants detectable by TLC and fluorography according to

the methods of Randerath (1970). The absorption spectra

and inhibition obtained for calf thymus RNA polymerase II as

well as TLC mobility indicated that it is identical to ADGG.

Synthesis of a-Amanitin-Protein Conjugates


Conjugation to bovine serum albumin (BSA) was performed

by modifications of the methods presented by Faulstich and

Trischmann (1973). The first step of the conjugation was

removal of the t-BOC group from ADBH by dissolving 6mg of

ADBH in 5ml of dry trifluoroacetic acid at room temperature,

swirling for 1 minute followed by immediate evaporation of

the acid at 400C in vacuo. The resulting compound, a-amani-

tin-diazobenzoylhexamethylenediamine (ADH), contains a free

amino group that may be coupled directly to free protein

carboxyl groups by reaction with water soluble carbodiimides

(Carraway and Koshland, 1972). Twenty milligrams BSA dis-

solved in 2ml water with 200mg of N-ethyl-N'(dimethylamino-

propyl)-carbodiimide HC1 (EDC) were added to the dried ADH

in a round bottom flask. After 24 hours at room temperature

with occasional mixing, the reaction mixture was adjusted to

0.05% with respect to NH4HCO3 by addition of solid NH4HCO3,

applied to a Sephadex G-75 column (2.5 x 30cm) pre-equili-

brated with 0.05% NH4 HCO3, and eluted with the same buffer.

Conjugated ADH-BSA eluted as a single peak followed by

unreacted ADH which was pooled and concentrated for reuse.

The peak of protein conjugate was analyzed for protein con-
tent by uv absorbance at 280nm (E% = 6.61) and by the
Lowry method (Lowry et al., 1951), for a-amanitin content

as indicated from absorption of the diazo linkage at 384nm

(E384 = 14000cm /mMole) and by inhibition of calf thymus

RNA polymerase II. After analysis, individual fractions

were pooled, lyophilized and stored at -200C.


Conjugates of a-amanitin and concanavalin A (Con A)

were prepared in a similar fashion to the ADH-BSA conjugates.

Five milligrams of dried ADBH were dissolved in 4ml dry

trifluoroacetic acid, swirled for 1 minute and evaporated

to dryness at 400C. Twenty milligrams of Con A (Sigma,

grade IV) dissolved in 2.5ml of water with 200mg EDC were

added to the ADH and reaction was carried out at room tem-

perature with intermittent mixing. A fine precipitate was

formed after 2 hours that settled out of the reaction mix-

ture. After 24 hours the reaction mixture was adjusted to

0.05% with respect to NH HCO and applied to a Sephadex G-75

column with reasonable care taken to leave the precipitate in

the reaction vessel. The conjugate ADH-Con A eluted with 0.05%

NH4HCO3 as a single peak followed by the unreacted ADH. Anal-
ysis of the individual fractions by absorption at 280nm (E%
1.14) and 384nm and analysis of the pooled ADH-Con A peak by

inhibition of calf thymus RNA polymerase II was performed.


Conjugation of a-amanitin to free amino groups on BSA

was undertaken with the ADGG derivative by procedures similar

to those used to couple ADH. ADGG (3.34mg) was dissolved in

2.0ml of water along with 16.2mg of BSA. The pH was adjusted

to 7.2 with 0.1N NaOH and 200mg of EDC were added. After

24 hours of reaction at room temperature the product ADGG-BSA

was isolated by gel filtration on Sephadex G-75 with 0.05%

NH HCO3 as eluant. The conjugate peak was analyzed by

absorption at 280nm and by the Lowry assay to determine

protein content. Alpha-amanitin content was determined from

the peak absorption of the diazonium moiety which for ADGG

is at 395nm (E395 = 1400cm/mMole). Inhibition of calf

thymus RNA polymerase II was determined for the peak fractions

and the ADGG-BSA conjugate was lyophilized and stored at -200C.


The optimal reaction conditions for production of ADGG-

Con A conjugates were determined from a series of experiments
utlzn 14 14
utilizing C-hippuric acid ( C-HA) as a free carboxyl group

containing analog for ADGG. The optimal conditions for

conjugation with Con A were verified in studies with 3H-DM-

ADGG. Details of the 1C-HA and H-DM-ADGG reactions are

presented in later sections. The procedure described below

represents a typical synthesis used for the production of

ADGG-Con A conjugate.

The Con A used for reaction with ADGG or hippuric acid

was purified by affinity chromatography on Sephadex G-100

followed by exhaustive dialysis versus water and lyophili-

zation as described by Agrawal and Goldstein (1965).

Purified, lyophilized Con A was stored at -200C until used.

Prior to conjugation, Con A was dissolved to an approximate

concentration of l0mg/ml in 0.01M phosphate buffer, pH 5 and

allowed to remain at room temperature for 1 hour. The turbid

solution was then clarified by centrifugation at 10,000rpm

(Beckman J-21B centrifuge; JA-20 rotor) for 30 minutes. The

supernatant was filtered immediately prior to use through a

0.2p Millipore filter and the concentration determined by

absorbancy at 280nm. Reactants were added in the following

order to achieve final concentrations as listed: 0.01pmoles/ml

Con A, 1.0pmole/ml ADGG and sufficient 0.1M phosphate buffer,

pH 5 to give a final concentration of of 0.01M phosphate. The

reaction was initiated by addition of 50ul of Immole/ml EDC

in water per milliliter of reaction mix final volume. The

reaction was allowed to proceed at room temperature for 12

hours after which time it was adjusted to 0.15M with respect

to NaCI and applied to a small column of Sephadex G-100 in

0.15M phosphate buffered saline, pH 7.4 containing 0.1mM

maganese and calcium (PBS ). The column bed volume was

generally 2 to 2.5 times the reaction mix volume. The

reaction mixture was allowed to enter the column at a rate

of approximately 0.lml/minute. Unreacted ADGG emerged from

the column first as essentially 100% of the Con A as ADGG-

Con A conjugate bound to the gel. After elution with 10 col-

umn volumes of PBS+, 0.1M D-glucose was added to the eluant

causing the displacement of ADGG-Con A in a single well-de-

fined band. The conjugate was characterized by absorption

at 280 and 395nm and dialyzed extensively against 0.15M PBS ,

pH 7.4 at 3-50C. ADGG-Con A conjugates were stored at 30C

and were used within three weeks of their synthesis. During

this period of time essentially no change in protein concen-

tration or absorbance spectra were noted.


Verification of the optimal conditions for conjugation

of ADGG and Con A and production of a tritium labeled Con A

conjugate for cell binding studies were performed with

H-DM-ADGG. Details of the individual experiments are

presented in the figure legends. The general format for

determining the extent of conjugation over a period of time

was to first establish the specific reaction conditions

desired. The different parameters investigated included

buffer concentration and pH, Con A, 3H-DM-ADGG and EDC con-

centrations. All reactions were initiated by the addition

of carbodiimide. At the desired sampling point, duplicate

50 or 1001p samples were removed to small tubes containing

200pg each of RNA and BSA in 0.02M Na2P 207 in .lml and

placed on ice. Cold 10% TCA, 2.0ml, was immediately added

and after 15 minutes on ice the precipitates were collected

on GF/C glass fiber discs and processed for scintillation

counting as described by Preston et al. (1975).

Production of 3H-DM-ADGG-Con A for cell binding was

accomplished by reaction of 0.0032pmoles of Con A with

0.22imoles of 3H-DM-ADGG (specific activity = 7.4 x 10 dpm/

pmole) with 750moles of EDC in 1.5ml of 0.01M phosphate buf-

fer pH 5. After 12 hours of reaction, the preparation was

applied to a column of Sephadex G-75 (0.2ml bed volume) pre-

equilibrated with 0.15M PBS+, pH 7.4 and eluted with the

same buffer. Following elution of 80ml, 0.1M D-glucose was

added to the eluting buffer. Individual 1.0ml fractions were

sampled for determination of total radioactivity in Bray's

solution and for TCA precipitable activity as described above.

The 3H-DM-ADGG-Con A peak was pooled after analysis and

dialyzed versus three two liter volumes of PBS at 50C.

4C-HA-Con A

Evaluation of the optimal reaction conditions for

carbodiimide mediated coupling of free carboxyl containing

derivatives of a-amanitin to Con A was performed with the
use of 1C-hippuric acid as an analog. The reaction of
1C-HA and Con A was examined under a number of different

conditions. Reactions were generally carried out in a final

volume of 1.0ml at room temperature. Procedures for sampling

the reaction at various time points were those previously

described for 3H-DM-ADGG reactions. The reaction buffer

was either 0.1M sodium phosphate, pH 5, 6 or 7, or pH 7.2

NaC1, 0.01 or 0.1M containing either 0.1 or 0.01M CaC12 and

MnCl2, respectively. Con A, (0.001pmole/ml) and l.0pmole/ml
14 5
of 1C-carboxyl hippuric acid (8.5 x 10 dpm/pmole, ICN

Chemical and Radioisotope Division) were reacted with either

10 or 100pmoles of EDC in the buffer systems described above.

HA-Con A
Conjugates of hippuric acid and Con A without 1C label

were prepared for analysis of the retention of lectin charac-

teristics by the HA-Con A conjugates. Two different reaction

conditions were chosen for these studies. Reaction mixtures

containing 0.001mole/ml Con A, 0.lmole/ml HA and lOmole/ml

EDC were prepared in either 0.1M phosphate buffer, pH 5 or

0.1M NaC1, pH 7. The reactions were incubated at room tem-

perature for 2 hours followed by overnight dialysis against

0.15M PBS+ at 50C to remove unreacted components. The

resulting HA-Con A conjugates designated HA-Con A (PO4) and

HA-Con A (NaC1) were used without further purification.

Biochemical Characterization of Conjugates


As previously mentioned all conjugates were analyzed

for protein content on the basis of their uv absorption at

280nm and for their a-amanitin content at 384nm (ADH conju-

gates) or 395nm (ADGG conjugates). A Beckman dual beam

spectrophotometer and recorder were used for spectral studies

with standard Icm pathlength quartz cuvettes. Measurements

were made with aqueous solutions unless noted otherwise.

Protein Determination

Protein determinations were made by minor modification

of the Lowry-Folin assay (Lowry et al., 1951). Crystalline

BSA (Sigma, lyophilized, crystallized) or affinity purified

Con A were used as standards. The assay was performed at

20-220C with 0.lml of protein sample containing between 20

and 300pg of protein per milliliter. Results were obtained

by sample absorption at 600nm, a wavelength at which no

interference from the chromogenic diazo linkage of the con-

jugates was found. Duplicate or triplicate determinations

were made for each assay point and standard.

Analysis of Conjugate Linkages

Conjugate bond formation. The nature of the chemical

bond formed between the protein (BSA or Con A) and the EDC
coupled moiety was examined with radioactively labeled 14C-

HA-Con A and H-DM-ADGG-Con A conjugates. The stability of

the bond to hydrolysis by hydroxylamine was determined by

exposure of the conjugates to 0.5M hydroxylamine at 370C

for 12 hours (Carraway and Koshland, 1972). The remaining

TCA precipitable activity in comparison to a nonhydroylzed

control sample was used as an indicator of covalent bonding.

Protein-protein cross-linking. The extent of protein

to protein cross-linking induced during the carbodiimide
conjugation was evaluated for ADH-BSA and 1C-HA-Con A

conjugates by sodium dodecylsulfate-polyacrylamide gel

electrophoresis (SDS-PAGE) according to the method of Weber

and Osborn (1969). Electrophoresis was performed with 5%

acrylamide disc gels run at 8 milliamp per gel. Visuali-

zation of the bands was achieved by staining with Commassie

brilliant blue (0.1%) in 50% TCA for 1 hour at 370C (Laemmli,

1970). Diffusion destaining with multiple changes of 7%

acetic acid was carried out to remove dye not associated

with protein.

Inhibition of Calf Thymus RNA Polymerase II

Calf thymus RNA polymerase II was purified through DEAE

cellulose chromatography according to the method of Kedinger,

et al. (1975) and stored in liquid N2. A single preparation

of enzyme used for all studies presented here proved to be

extremely stable in liquid N2 over a four year period.

Assay of the inhibition of calf thymus RNA polymerase II

was performed according to the procedures developed by Preston,

et al. (1975). The reaction mixture used was that described

by Cochet-Meilhac and Chambon (1974). The inhibition assays

were performed as follows: To the bottom of a new 10 x 75mm

test tube was added 0.01ml of each inhibitor concentration

being tested followed by 0.03ml of enzyme. The reaction

was initiated by adding 0.06ml of reaction mix containing

5- H-uridine triphosphate ( H-UTP). After 10 minutes at

370C the reaction was stopped by addition of O.lml of 0.02M

Na2P207 with 2.0mg/ml each of RNA and BSA and placed on ice.

Ice cold TCA/Na2P207 (7% TCA, 0.02M Na2P207) was added(2.0ml/

tube) and the tubes allowed to stand on ice for 15 minutes.

Precipitates were collected on GF/C glass fiber discs pre-

washed with TCA/Na2P207 on a vacuum manifold. Collected

precipitates were washed three times with TCA/Na2P207, three

times with ethanol:ether (3:1), two times with ether, air

dried at 800C and counted in a toluene based scintillation


Estimation of the apparent inhibition constant for

each inhibitor was performed determining the inhibition of

3H-UTP incorporation obtained by a series of increasing con-

centrations of inhibitor for a constant quantity of enzyme

and a known concentration of labeled substrate ( H-UTP).

The same assay was repeated with two or three different

concentrations of substrate. These results when plotted

according to the method of Dixon (Dixon and Webb, 1958)

yield a value for the apparent inhibition constant, K1.

Determination of Lectin-Associated Properties

Binding to Sephadex. Ligand binding activity of the

conjugates was examined on the basis of their interaction

with Sephadex G-75. Preparations of conjugates were equili-

brated with 0.15M PBS+, pH 7.0 by dialysis and applied to a

column of Sephadex G-75 (2.4 x 40cm). After elution of one

column volume of PBS 0.1M D-glucose was added to the eluant.

Monitoring of fractions for 280nm absorbancy was used to

determine the degree of binding to the gel matrix in refer-

ence to native Con A.

Ligand affinity constant determination. The apparent

affinity constant for the ligand binding of selected conju-

gates was examined spectrophotometrically by the method of

Bessler et al. (1973). A chromogenic ligand, p-nitrophenyl

a-D-mannopyranoside (PNPM) generates a uv difference spectrum

following interaction with Con A the magnitude of which is

proportional to the affinity constant. A series of dilutions

of PNPM in 0.15M PBS were measured for absorbance at 317nm

as was the solution of Con A or Con A conjugate being examined.

A constant amount of the protein was then mixed with the

PNPM dilutions and 317nm absorbance redetermined. From these

data, the following parameters can be defined: Pt, concen-

tration of protein in protein and PNPM mixtures: Dt, concen-

tration of ligand (PNPM); AA317 =[A317 (protein) + A317 (ligand)]-
1 1
A 317 (protein + ligand). A plot of versus D- for values
317 A [Dt]
of [Dt much greater than [Pt] yields a y-intercept the re-

ciprocal of which equals AAmax or the change in absorbancy

when all ligand binding sites are saturated. From this value

the affinity constant can be determined by first calculating

the concentration of protein-ligand complex, [PD] from the

relationship [PD] AAmx [Pt]. The concentration of free
AAmax t
ligand, [D], is then calculated from [D] = [Dt] AAmax [Pt

[A plot of t] 1
A plot of [PD versus 1 generates a straight line with a

slope equal to the reciprocal of the association constant,


Hemagglutination. The hemagglutinating ability of

various Con A conjugates was examined with human type A red

blood cells. Standard microtiter plates (Cooke Engineering)

and dilutions were used throughout. Agglutinations were

determined for serial two-fold dilutions of protein in 0.15M

PBS pH 6.8 and red blood cells at 1 x 108 cells per milli-

liter. After complete mixing and suspension of cells,

hemagglutination plates were allowed to stand at room tem-

perature until a control well containing only red blood

cells in PBS+ had formed a well defined pellet. Agglutina-

tion titers were read by visual inspection and are expressed

as the reciprocal of the largest dilution yielding positive


Interaction of Conjugates with Cultured Cells

Maintenance of Cell Cultures

Cultured mammalian cells used for these studies were

maintained at 370C in a 5% CO2:95% air humidified atmosphere.

Tissue culture media, sera and antibiotics were obtained

from Flow Laboratories, Rockville, MD or International

Scientific Industries, Avon Park, IL. All media were

supplemented with 10% fetal calf serum (FCS), pennicillin

(100 units/ml) and streptomycin (100pg/ml). Stock cultures

were maintained in 25cm or 75cm2 disposable plastic culture

ware (Corniny). Transfer of cell stocks and experimental

procedures were performed in a laminar flow hood isolated

from other laboratory activities.

Human amnionic cells, AV3, were obtained from Dr.

G. Gifford, Department of Immunology and Medical Microbiology,

University of Florida, Gainesville, Florida. AV3 cells

were grown in Eagles MEM with Earle's balanced salts, 10%

heat inactivated FCS (560C, 30 minutes) and penicillin-

streptomycin. They were routinely passage every two days
4 2
at a density of 1.25 x 10 cells/cm2. For transfer of

cells, the cell monolayer (approximately 80% confluent)

was rinsed two times with 0.15M PBS containing 0.05mM EDTA.

One milliliter of PBS containing 0.02% trypsin and 0.05M

EDTA (trypsin-EDTA) was added and the cells were allowed

to detach from the culture vessel, generally within one

to two minutes. Four milliliters of fresh 370C media was

added and cell counts were made with a hemocytometer.

Cells were transferred to new flasks with fresh media after

dilution and incubated.

Chinese Hamster Ovary cells, CHO, were obtained from

Dr. K. D. Noonan, Department of Biochemistry, University of

Florida, Gainesville, Florida. Three CHO lines were used,

M-7, HI-7 and 11-7Wcr. M-7 and H-7 differ primarily in their

response to dibutryl cyclic adenosine monophosphate (db-cAMP)

whereas H-7Wcr is a Con A resistant line derived from 1-7.

CIIO cells were maintained in McCoy's 5A (modified) medium

with 10% FCS, penicillin and streptomycin. They were

routinely transferred every two days at a plating density of

1 x 104 cells/cm2 by trypsinization as described above.

Mouse lymphocytic leukemia cells, EL4, were obtained

from the Salk Institute for Biological Studies, La Jolla,

CA. EL4 cells were grown as suspension cultures in RPMI

1640 with 10% FCS,penicillin and striptomycin. They were

transferred every three days at a final density of 1 x 104


Rat fibroblast cells, A-9, and an a-amanitin resistant

mutant of the A-9 line, LAN-2, were obtained from Dr. J.

Eisenstadt, Institute of Human Genetics, Yale University,

School of Midicine, New Haven, CT. The LAN-2 amanitin

resistant line contained RNA polymerase II activity several

fold more resistant to inhibition by a-amanitin in vitro

compared to the analogous enzyme from the parent line A-9,

(R. Bryant, personal communication). Both lines were grown

in Eagles MEM with Earle's salts, 10% heat inactivated FCS,

penicillin and streptomycin. Stocks of A-9 and LAN-2 were

transferred every three days at a final density of 1 x 104

cells/cm2 by trypsinization as previously described.

Inhibition of Cell Growth

Inhibition of cellular proliferation by amanitin-

protein conjugates was measured by determining cell numbers

with an electronic particle counter (Celloscope, Particle

Data, Inc.). For inhibition by ADH-BSA conjugates, cells

were grown in sterile glass scintillation vials. Prior to

each experiment AV3 and M-7 CHO cells were harvested by

trypsinization, resuspended at a density of 0.5 to 1.0 x

105 cells/ml in the appropriate media and added in 1.0ml

aliquots to scintillation vials. After 12-16 hours of

growth 100pl of each inhibitor concentration (in culture

media) being tested were added to triplicate cultures to

achieve the desired final concentration of inhibitor.

Growth was allowed to proceed for 48 hours after which time

the cells were harvested for counting. Cultures were gently

rinsed two times with 2.0ml cold balanced salt solution

(Gey's A without Ca2+ or Mg +) and treated with 1.0ml tryp-

sin-EDTA for 5 minutes. After detachment from the glass,

9.0ml of Gey's A solution was added to each vial and cells

were counted directly.

EL4 cells were grown in suspension in scintillation

vials for 12-16 hours, exposed to inhibitor for 48 hours

and counted directly by dilution with Gey's A.

Inhibition of CHO H-7 and H-7Wcr cell growth by ADGG

or a-amanitin was also examined by cell number determination

but the cells were grown in plastic multiwell tissue culture

plates (Corning). CHO cells H-7 and H-7Wcr were plated at

1 x 104 cells/cm2 in 1.0ml 10-12 hours prior to the addi-

tion of inhibitor. Forty-eight hours after addition of

50pl of inhibitor, the cells were rinsed two times with

cold PBS and removed from the vessel surface by two

minutes of treatment with 1.0ml trypsin-EDTA. Aliquots

were removed directly into 10ml PBS in a siliconized

scintillation vial and immediately counted with the cello-

scope. Care was taken to ensure that cell clumps were well

dispersed by pasteur pipetting with a silicone treated

pipette. A single plate containing 18 individual wells was

processed at a time with the entire procedure requiring less

than 30 minutes.

Inhibition of H-7 and H-7Wcr cell growth by Con A,

ADGG or ADGG-Con A conjugates was measured by exposure of

established cell monolayers to the protein for 15 minutes.

Cells were plated at a density of 1 x 104 cells/cm2 and

allowed to establish growth for 12 hours. The media was

carefully aspirated and the cells gently rinsed two times

with 0.15M PBS, pH 7.4 at room temperature. Dilutions of

each inhibitor were added in 1.0ml of PBS+ to triplicate

wells for each concentration used. After 15 minutes of

exposure, the protein solution was aspirated, the cells

rinsed twice with PBS and 1.0ml fresh media added to the

well. After 48 hours of growth, the cells were processed

for counting as described above.

A-9 and LAN-2 cells were treated as described for

CHO H-7 and H-7Wcr cells for determination of their sensi-

tivity to free or conjugated a-amanitin.

Inhibition of 3H-Thymidine Incorporation

Incorporation of [methyl- H]-thymidine ( H-TdR) was

measured for AV3, CHO M-7 and EL4 cells as an additional

method for assessing inhibition of cellular functions by

ADH-BSA conjugates. A procedure similar to that of Ball,

et al. (1973) was used. Labeled precursor (Schwartz-Mann,

5mCi/mmole) was added as a 100yl addition to give a final

concentration of lCi/ml 47 hours after addition of inhi-

bitors to cultures. The assay was performed in scintilla-

tion vials as described for inhibition of cell number by

ADH-BSA conjugates. After 1 hour at 370C, incorporation by

AV3 and CHO M-7 cells was stopped by addition of 10ml of

ice cold Gey's A. Cultures were washed three times with

5.0ml of 1.5% perchloric acid, once with 95% ethanol and

drained. The vials were heated for 40 minutes at 800C

after addition of l.0ml/vial of 5% perchloric acid. This

treatment served to hydrolyze the nucleic acids for scin-

tillation counting which was done following addition of

10ml/vial of a Triton X-100 based scintillation cocktail

with a Beckman L5-133 counter.

Measurement of 3H-TdR uptake by EL4 cells was accom-

plished with a 1 hour exposure to the labeled nucleoside.

Collection of the entire culture on GF/C glass fiber filters

(prewashed with 1.5% perchloric acid) was performed after

stopping incorporation with 1.0ml 3% perchloric acid. The

collected precipitates were washed once with 1.5% perchloric

acid, three times with ether:ethanol (1:3), twice with ether,

dried at 1000C and counted in a toluene based scintillation


Measurement of Pinocytosis
Synthesis of 1I-RSA. Measurement of the pinocytic

activity of AV3, CHO M-7 and EL4 cells by uptake of 125I-

BSA was performed by modification of the method of Steinman,

et al. (1974). 1I-BSA was prepared by iodination with

carrier free 125I (New England Nuclear, carrier free, sodium

salt, 50mCi/ml) and solid state lactoperoxidase. The lacto-

peroxidase (Sigma Chemical Co.) was coupled to Sepharose 4B

activated by cyanogen bromide according to David and Reisfeld

(1974). BSA was iodinated to a specific activity of lpCi/wg

and purified by gel filtration on Sephadex G-25 in 0.01M

phosphate buffer, pH 7.

Determination of pinocytic activity. For determination
of pinocytic uptake of 1I-BSA, AV3 and CHO cultures were

grown in scintillation vials to near confluence. Each culture
125 7
received 50pi of 1I-BSA containing approximately 10 cpm.

After 24 hours the media was removed and the monolayers

washed six times with 20ml of serum free medium. Trypsin-

EDTA (0.5ml/culture) was used to detach the cells after which

the cells were removed to a plastic tube for counting in a

gamma counter (Nuclear Chicago). The culture vessel was

washed an additional two times with 1.0ml medium which were

added to the cells. Prior to counting, the cells were separ-

ated from the media by centrifugation (10 minutes at 1500rpm)

and the cell pellet and supernatant were counted independently.

In all cases, negligible activity was found in the superna-

tant fraction. Duplicate cultures which received 501l of
2I-BSA 5 minutes prior to harvesting were processed as

above. The activity associated with the cell pellet for

these cultures was subtracted from the 24 hour culture values

as a control for nonspecific surface absorption. Pinocytosis

by EL4 cells was determined by similar methods with the

addition of a centrifugation step after each wash to pellet

the suspended cells.

Cell Binding
Synthesis of 125I-Con A. Evaluation of the number of

membrane binding sites present on cultured cells for Con A

was acheived by using Con A labeled to a high specific acti-
vity with 1I. The chloramine T method developed by

Cuatrecasas (1973) was used with minor modification. To

400pg of Con A in 0.05ml of 0.15M PBS+, pH 7.4 was added
ImCi of 125I (New England Nuclear, sodium salt, carrier

free) contained in 0.lml of 0.1M sodium phosphate buffer,

pH 7.4. Freshly prepared chloramine T (100l g in 0.025ml

water) was quickly added and mixed for 50 seconds. Sodium

metabisulfate (200pg in 0.025ml water) was added to stop the

reaction followed by addition of 0.30ml PBS The entire

mixture was applied to a 2.5ml column of Sephadex G-100 pre-

equilibrated with PBS The column was washed with 25 vol-

umes of PBS after which time negligible activity eluted from

the gel. D-glucose (0.3M) was added to the elution buffer
and the bulk of the 1I-labeled Con A was collected in a
single 1.0ml fraction. The 1I-Con A was dialyzed versus

two three liter changes of PBS at 50C over a 48 hour period.

The labeled protein was analyzed for total activity, TCA

precipitable activity and total protein by the fluorometric

assay of Bohlen et al. (1973). Following analysis the

preparation was diluted with native Con A to a final specific

activity of 2 x 10 cpm/pg and stored at -200C in Nunclon

freezing vials (Vanguard International) in 0.5ml aliquots

(210pg/vial). The labeled Con A was thawed one time only

and the preparation was used within 18 days of its synthesis.

It should be noted that successful labeling was acheived

only with 125I batches used within five days after their


Cell binding measurement. Methods for the determina-
tion of 1I-Con A binding to cultured CHO H-7, H-7Wcr and

rat A-9, LAN-2 cells were generally adapted from those pre-

sented by Noonan and Burger (1973a). Cells were grown to

approximately 80% confluency in Linbro multiwell plates with

30mm diameter wells (Linbro Scientific). For binding, cul-

tures were removed to room temperature for 5 minutes, media

were removed by aspiration washed gently two times with

0.15M PBS, pH 7.4 and 1.0ml of the desired concentration of

125I-Con A (2x 104 cpm/pg) in PBS+ was added to duplicate

cultures. After 15 minutes incubation the Con A was re-

moved followed by 5 washes of 2.0ml PBS. PBS containing

0.05M EDTA (l.0ml/well) was added and the cells were

allowed to remain at room temperature for 30 minutes. Tryp-

sin-EDTA (0.5ml) was added to each well and the plate incu-

bated at 370C for 15 minutes. A final addition of 1.0ml

20% Triton X-100 was made and the plates heated at 400C

for 60 minutes. Duplicate 0.5ml aliquots were removed from

each well to 10ml of Brey's solution and the activity deter-

mined by counting in a Beckman LS-133 liquid scintillation

counter set on the narrow tritium window. Each experiment

contained controls for background adsorption to empty wells,

inhibition of 125I-Con A binding by 30 minutes of pre-incu-

bation with Con A (100pg/ml), specific ligand (a-methyl-D-

mannopyranoside (MDM), 2mM or non-specific protein (BSA,


Identical procedures to those described above were

used for determination of binding of 3H-DM-ADGG-Con A con-

jugates. For these experiments the specific activity was

3.46 x 103cpm/pg and additional controls for the interfer-

ence of ADGG with the binding were performed.


ADH-BSA Conjugates


Preliminary investigations of the targeting of a-

amanitin to specific cells by macromolecular carriers were

patterned after the work of Faulstich and Trischmann (1973).

The initial objective was to synthesize a-amanitin-protein

conjugates which would then be evaluated on the basis of

their biochemical properties and in vitro toxicity for

selected cell types. Work by other investigators (Wieland,

1968) and NMR studies by Preston and Gabbay (unpublished

results, 1977) had demonstrated that a-amanitin contains a

hydroxytryptophan moiety (Figure 1) that is available for

chemical cross-linking via diazotization to other aromatic

groups. A procedure was developed (Figure 2) from the syn-

thesis reported by Faulstich and Trischmann (1973) that

allowed for the formation of an a-amanitin derivative, ADH,

containing a free amino group.

Synthesis of compounds leading to the formation of

N-(4-aminobenzoyl)-N'-BOC-hexamethylenediamine was per-

formed by Dr. J. F. Preston. This compound was then

coupled to the hydroxytryptophan moiety of a-amanitin via

a diazo bond and the resulting derivative, ADBH, purified


rd .C
03 oo >4

c --
0 *-

41 0 0
U) r44

-U (5 0
0 0 On '

F a) On

4 r) 0

en U) -
o >S n
S >m- n -.

l 4) 0d r-

0 a

0 o ,0o -0

) 0 *H r

p a) 0 E

U a4 -4 0
Sed 4

r -X rd0
U) 4t)rdJ e
a a) -o O

) M4 r 4) d
> 4 1

a) M -I 4J, -i
1 (0 e-1 0
> U 0 (



O --U 0=0

C) (

Z-X = 3----Z--0=0
I~ I

m~ (9

o o
I '
14 G

J d



tc ) 0C) o
IC) -- -

0 Q


O ^ I j

T :1-:
0 0 U 0~O1

~)---- 0C

c c c
c "c --
a i 0 C --
E E EE c

I I 1<


Generalized scheme for derivation of a-amanitin and
carbodiimide mediated conjugation to protein carboxyl



N-1 C 1 4--CO- NH- (CII2) -6 lJ).l()
2 &4' 2 6


4 (ADBH)


AMANIT1N-N=N-C I4 -CO-NH- (C112) 6 -N12


R' NC=":R"


6 4 2 6


by LH-20 chromatography in methanol. Yields were generally

69% or greater. The purified ADBH chromatographed as a

single spot in TLC in methanol:methyl ethyl ketone (1:1)

with an Rf of 0.67. NMR spectroscopy by Drs. E. Gabbay and

W. Brey, Jr. of the Department of Chemistry, University of

Florida, Gainesville, Florida indicated that the hydrogen

atom resonance spectra of the ADBH were consistent with the

proposed structure. The uv and visible spectra presented

for ADBH in Figure 3 show ratios of extinctions at 384nm

to 304nm of 0.85, identical to the ratio of the extinction

coefficient for the azo dye at 384nm, 14000cm2/mmole to that

of free a-amanitin at 304nm, 16400cm2/mmole. This would

imply a stoichiometric relationship between a-amanitin and

the diazo linked spacer molecule.

Bovine serum albumin was chosen for the initial con-

jugate studies based on reports in the literature of selec-

tive toxicity of BSA-B-amanitin conjugates for macrophages

(Fiume and Barbanti-Brodano, 1974). Following removal of

the BOC group from ADBH and carbodiimide coupling to BSA,

the conjugate was purified from the reaction mixture by

chromatography on Sephadex G-75 (Figure 4).


The conjugate peak eluted in the void volume of the

G-75 column and contained significant absorption at 384nm

as well as 280nm (Figure 3) indicative of covalent associa-

tion of the diazo compound and BSA. The ADH-BSA contained

a molar ratio of a-amanitin (as azo compound) to BSA of 1.2


The absorption spectra of ADBH (solid line) dissolved
in methanol (33 wg/ml), ADH-BSA (dashed line) in 0.05%
NH4HCO3, pH 8.0 (1 mg/ml) and BSA (dotted line) in 0.05%
NH4HCO3, pH 8.0 (1 mg/ml) were determined with a Beckman
model 24/25 spectrophotometer at room temperature and
a 1.0 cm light path.




S! I



\ Ni

0.01-.........""".......'.1 1
200 400 GOO
,. \

1, nm

En nU
m m

U) co >

0 0 Oc -

0 0 r E
0 20004

0 0 0 >c

O t)Q t4 .
0 O O 1 r
w. o
14 0 c 0

I q-- -- --
10 v0 1 I
I C:) U 0 4j
m H -4

Sut) 4J
o Oo" Q II

o 0 0

x Q o ) a )
04 0 0 > r-0

04 HQ
O 0 0 43 -H
U 0C M0
0 0 o a II
ui O U 0 l
C- ()0 4-

0 U j

i -4U -' M

0 Q Cd H
C) o rU rd a

I0 -H, r-l

(u Un > Q

ljWH)i''Jj4 in VAI iO]


0 c~j

as determined from the absorption spectra. Lowry protein

analysis of the conjugate yielded slightly lower values for

the amount of protein present than those derived from the

280nm absorbance. Since no interference with the Lowry

assay was noted from ADH added to known quantities of BSA,

it is presumed that the ADH conjugation induces a slight

increase in the 280nm absorption of the BSA. A factor of

1.18 times the 384nm absorbance was subtracted from the

280nm value to empirically correct for these differences.

Inhibition of calf thymus RNA polymerase II by the individ-

ual fractions obtained for G-75 chromotography gave addition-

al confirmation of the association of a-amanitin with BSA

for the ADH-BSA conjugate.

The extent of protein to protein cross-linking result-

ing as a side reaction from carbodiimide coupling was esti-

mated by SDS-PAGE of the ADH-BSA. The relatively high pro-

tein concentrations (10mg/ml) and high EDC concentration

(100mg/ml) used for conjugate synthesis led to significant

cross-linking, shown in Figure 5. Estimation of the area

under each peak of the scan of gel 1 indicated approximately

50% of the protein was in the monomeric form. Molecular

weight calibration with known protein standards (Figure 6)

demonstrated the cross-linked material to be in multiples

of BSA molecular weights (66,000 daltons) which would imply

that no fragmentation of the protein occurs during conjuga-


U 0
*H -H

a) .0 0 a *H
Un tr 4- E
a0 -Hl 0 .-1 0
S(1) 1Q 0O
0 124
4 0 -
O 4 -4 4 TC

-1 S 00
o Mn 4- r-i u

U r (U (
a a) nt -H 4-

I 1 0 ) O)
W E4 0-4
i 0
a) E- 0
k c- oP -
a)4 0
4 () *4 (d

0 U) a)
4 a a) 0

S 4 0 01 -
S 4- C o 0

cn m -*H
U t) (aa) 4Jrd
E- r- M a)
Sa)0 0 0) 0

Z 9 E ar-l H
0 O L *
aI 0 ) 0

f 0 *HO tr (U
cq < r-q F_
a o -Hl*^ e

0 -Io 0 0
4J 0 4J r-I
dt 44'-1 r-L

O 0 U U- O
a) rd) Uc
co -4 C) 0

in +jq d u) r.

7 t 0 4)-i r-l CH
-4 ) : (r- 0
n C) 0Q g4 E- U

" Cz. L:J-


O. o


0 J4
I4 0 -H


0 4

(I) H Q
E-A r

o 10 04

4J 04
H a~
-r- r 4

U 0 0

-4 Q)

0I (1) V.

Lf (J4JQ.0
l) 1) 0 M

(1 a4

3~~ w- c,
0 -1

0) *j(1 C
q rT C:

r0r 0)
I-4 SrA

0 0drl a)O

U 54

0 0~ t7
W E4 rdA

1 E0
U 0)4J~

0) 4W-) in

rJL4 0 E K4l
~H 0r

z( Z
o 0

A-A o


10 0


0 .0
-V- l

The inhibitory potential of the ADH-BSA conjugate

for calf thymus RNA polymerase II was quantitatively com-

pared to that of free a-amanitin or ADH by determining the

apparent inhibition constant (K ) for each compound. The
value of 1.8 x 10 M obtained for a-amanitin agrees with

published values (Chochet-Meilhac and Chambon, 1974) and

does not differ significantly from that for ADH (K =3.0 x
10 M). Both inhibitions were noncompetitive with respect

to UTP. ADH-BSA inhibition deviated from the strictly non-

competitive inhibition seen for a-amanitin and ADH with an

average of the x-intercepts giving an apparent inhibition
constant of 69 x 10 M. Conjugation to BSA resulted in a

38-fold decrease in the affinity of bound a-amanitin for

calf thymus polymerase.

Interaction with Cells

The toxicity of conjugated ADH-BSA was examined in

three cultured cell lines, CHO M7, AV3 and EL4. Table 1

presents the effects of continuous 48 hour exposure to ADH-

BSA or free a-amanitin on cellular proliferation and 3H-

thymidine incorporation of CHO M-7 cells. Comparing equiva-

lent molar concentrations of free or conjugated a-amanitin,

conjugated a-amanitin was found to be slightly less effective

an inhibitor of cell growth than free a-amanitin for CHO

cells. Greater inhibition is seen when examining 3H-TdR

incorporation, with conjugated a-amanitin inducing 60% of
the inhibition obtained at 1 x 10 M concentration in com-

parison to free a-amanitin.


Results are expressed in a Dixon plot of the reciprocal
of velocity expressed as ppmole H-UMP incorporated
per minute versus concentration of inhibitor for three
different substrate (UTP) concentrations: 0.004 mM (0),
0.008 mM (A) and 0.016 mM ( ). Least squares analysis
was used to obtain the best fit for each line. For ADH
and ADH-BSA, the concentrations of a-amanitin were
determined from the diazo linkage extinction coefficient.

-4 0

l18xl O M1-9

4 8 12
[a .AMANIT ., !.'x 10

K = 3.0xIO ,1

-20 20 60 100
[o AMANlTIN]., M. O')

IQ 0

2 :









OU 0


u m
tf -]



4J 0O

C -


u -H


0 -O


H 0



c in t co
o n 4 o (o

rn l m in

-4 r- (N

com 00 C

Sr01- 1 m o


O I I I a

o (N 1-1 C in

H c( (N ('4 n

r- om r-- -4

m co LOn c

l (n rrU)l


2 4 ri

















The same comparisons are presented for EL4 cells in

Table 2. As was observed for CHO cells, free a-amanitin

is more effective in inhibiting cell growth than is conju-

gated a-amanitin with about three fold greater inhibition
seen at 5 x 106 M a-amanitin for nonconjugated toxin. EL4

cells were approximately five times less sensitive to a-

amanitin in either form than were CHO cells. Although the

H-TdR data presented parallels,the cell number data with

a-amanitin being more inhibitory than ADH-BSA at equivalent

concentrations, EL4 cells did not incorporate exogenous

thymidine to any great extent under the conditions examined.

The data is presented only for comparative purposes.

Table 3 contains similar inhibition data for AV3 cells.

Like EL4 cells, AV3 did not incorporate exogenous thymidine

to a significant extent but the inhibition obtained paral-

leled the cell number data. For AV3 however, inhibition of

cell proliferation by ADH-BSA was strikingly different than

that observed in CHO and EL4 cells. Although AV3 cells are

equally as sensitive to free a-amanitin as are CHO cells,

the AV3 cells were eight times more susceptible to conju-

gated a-amanitin in comparison to an equivalent molar dose

of free a-amanitin. This would imply a preferential uptake

of ADH-BSA by AV3 and/or modification of the conjugate to

a more toxic derivative after uptake by AV3 cells.

Figure 8 depicts cell size distributions of AV3, CHO

and EL4 cells after 48 hours of exposure to ADH-BSA or

a-amanitin. The results clearly indicate that no shift in




E-4 C4



u p




u H


cq :

.-l u

4-1 0


Q) -4


4) X



a o
u '-

I *

Ln in i -T i i re

01 ~l IIC

in r-i rr t n

nm m N r o4

o r-4CM Ln
CD 1


lo Ln


ln r- r-i

ln In ml o

* I *

N C)
r- 0

Ow- n

U) ) o()



m CM n

tf i)



0-, -
4-) 0

Q) 4-)

0 -
() -


E n





4 0

- H






m a


O 0

Kt e



**II I
M Ic r-i I I

C 00 rn I i I co I
c' c i l-

II *i
0 lml I o

in in

00I IC O I

01 1 I **

O C I m l 4
r- a a aC m
z c~fO

in r- CN
I I I ,
II I O H )

0 *I II
cm I l l n(N
(N I I r- r

.o 0 r-~

o -1 in IO I i I

0o r- r'3 m i





I x





Following 48 hours of exposure to either ADH-BSA (dotted
line) or a-amanitin (dashed line) as described in
Materials and Methods, cells were compared to normal
cells (solid line) on the basis of size by a 110 channel
celloscope particle counter.

20 GO 100 140













mean cell size has occurred as a result of exposure to

either inhibitor which might have occurred in a growing

but nondividing population of cells or in one that was

dividing in the absence of DNA synthesis.

Differential uptake of ADH-BSA could possibly occur

by means of receptor mediated uptake or enhanced pinocyto-

tic uptake by susceptible cells. In order to explore the

latter possibility, AV3, CHO and EL4 cells were examined for
their relative rates of pinocytosis. Uptake of 125I labeled

BSA over a 24 hour period was used as a measure of pinocyto-

sis (Steinman et al., 1974). Table 4 presents a summary

of the pinocytosis rates in comparison to the molar concen-

trations of free versus conjugated a-amanitin required for

25% inhibition of cell growth and 3H-TdR incorporation for

all three cell lines. The amount of pinocytosis observed

with each cell line was in direct correlation to the rela-

tive sensitivity of the cell line to conjugated a-amanitin.

AV3 cells were 3.5 times more active in the pinocytic uptake
of I-BSA than were CHO cells, whereas EL4 cells took up

negligible amounts under similar conditions. These data

indicate that the increased sensitivity of AV3 cells to

conjugated a-amanitin is a direct function of increased

pinocytotic uptake of the conjugate relative to the other

cell lines tested.

The results of the ADH-BSA experiments pointed out the

necessity for using a macromolecular carrier for which spe-

cific cellular receptors are known to exist to further

o r0

u0 I 0
:C n ry
-4 N -

E- 0O
1 a

3: H

r-4 Q)

U- c


0 0




o 0

r- ChU '


co 'I n









0 0

-,I -r-

H -

clairfy the role of receptor mediated uptake of a-amanitin

conjugates. For this reason conjugates of a-amanitin and

the plant lectin, Concanavalin A, were synthesized.

ADH-Con A Conjugates


Similar conditions to those used for the synthesis of

ADH-BSA were used to prepare ADH-Con A conjugates. The con-

jugates were purified after carbodiimide coupling by chrom-

atography on Sephadex G-75 (Figure 9). The ADH-Con A eluted

in the void volume of the G-75 column and contained substan-

tial absorbance at 384nm and at 280nm (Figure 10). From the

molar extinction coefficients of Con A and the azo moiety

it was determined that all of the Con A applied to the column

eluted in a single peak and contained a molar ratio of a-

amanitin to Con A of 0.67.


The ADH-Con A conjugates were first characterized with

respect to their inhibition of calf thymus RNA polymerase II.

The results presented in Figure 11 generate an inhibition
constant of 186 x 10 M for calf thymus polymerase. The

high KI observed implies that a substantial reduction in the

affinity of a-amanitin for calf thymus RNA polymerase II

has occurred as a result of conjugation to Con A. Further-

more, as evidenced by the lack of binding of the ADH-Con A

to Sephadex G-75, these conjugates possessed little or none

of the ligand binding specificity of the native lectin.

Figure 9.


The reaction mixture from EDC mediated conjugation of
ADH and Con A was chromatographed in 0.05% NH4 HCO3.
The squares ([0) represent absorbancy at 384nm and the
circles (O) represent absorbancy at 280nm.



2.0- :

S* 4


IO 30

-* I

l/.rf* I _

10 30



The absorption spectra of ADH-Con A conjugate was
determined in 0.05% NH HCO3, pH 8.0.

0.4 -


0 -0


I -~

4-i (1)
nl 4-J 4

0 i4 M


rd *Q, )
Q) rO r-- 4J
rH 0

O (U ^-

,-O M (
a) C)

>1 p 0

-,-1 4-4
(u 44-

0 0 "-X
-i *H *W

> 4 O

U4-1 4J

0 0 0

-4 0 .,-

. E
Q) 4-3 0
4 ) ( )

M 0 C

0 Cd -H
X T) )
0) -4 41 >n
0 0 C

.4 H 0)
C 1 4- n
I) C ) 01

rt &.'- e




Other preparations in which Con A and ADH were coupled in

the presence of saturating amounts of specific ligand and

under varying salt and ion concentrations also failed to

absorb to Sephadex. Since it appeared likely that conju-

gation to carboxyl groups on Con A resulted in a loss of

ligand binding activity, investigations of these ADH-Con A

conjugates were discontinued.

ADGG-BSA Conjugates

The results of the ADH-Con A studies demonstrated the

need for an a-amanitin derivative containing a free car-

boxyl group. Such a derivative could be coupled to free

protein amino groups and possibly result in a derivative of

Con A that retained ligand binding properties as well as the

specificity of interaction of a-amanitin and RNA polymerase



The synthetic pathway for one derivative of this type

is shown in Figure 12 as developed by Dr. J. F. Preston.

The synthesis follows the same general pattern as that for

ADH conjugates. It results in the production by diazotiza-

tion of a-amanitin and an aromatic containing carbon chain

spacer molecule, p-aminobenzoylglycylglycine, of a free

carboxyl containing a-amanitin derivative, a-amanitin-

diazobenzoylglycylglycine (ADGG). The ADGG was purified

from the diazotization reaction by chromatography on Sepha-

dex LH-20 in 80% methanol. The product was judged to be








0 0H


.H r


u 0

ro r

-r-i 4-

0 U








7 :



'I _L
v-(9 f)-().


... -'









r Z









C, -





pure on the basis of its appearance as a single spot with

Rf of 0.57 after TLC in the systems described for ADBH.

Concentrations were determined for its absorption at 395nm

with an extinction of 14000cm /mmole.

ADGG was coupled to BSA for production of a conjugate

that could be directly compared with the previous work on

ADH-BSA conjugates. After EDC mediated conjugation, the

ADGG-BSA conjugate was isolated by chromatography on Sepha-

dex G-75 (Figure 13). The ADGG-BSA eluted as a single peak

followed by unreacted ADGG. The elution profile contains

significant absorbance at both 384 and 280nm indicating co-

valent association of the ADGG and BSA. Individual fractions

were further analyzed by inhibition of calf thymus RNA poly-

merase II. The inhibition profile closely parallels the

absorbance at 384nm demonstrating retention of inhibiting

activity of a-amanitin after conjugation.


Figure 14 shows the absorption spectra for ADGG and

ADGG-BSA. The peak absorbancy at 395nm for the azo moiety

was used to determine a-amanitin concentration whereas the

protein concentration was obtained from Lowry protein analy-

sis. No interference with the Lowry assay was noted by ADGG.

The ADGG-BSA conjugate prepared contained 2.9 moles of ADGG

per mole of BSA.

The inhibition of calf thymus RNA polymerase II by

ADGG and ADGG-BSA is shown in Figure 15. Free ADGG com-

pared favorably to a-amanitin with respect to inhibition of

U o

S m 0() 0

W 4J r-

W 0 .) C H
0 c 4 f c(r

0 C N)
0 -,4 0 H
0 <0 (1)

0m 6
'4-1 Q) C \

0 aM e

0 d nam

rO U) O O M
I 0 C 3>i

0 iA 0a)

0 d -I 0
44 4J r,
w a) 4J-H

U d) P( ra)-Q r

r4 H co Q i4
o U0 W

z3 4 II
4 0 o\ 04

0 4 0 *r 4 C
X a)0 a

) 4 41 ) rl
M E -1 U M 1

r H U 0C 0)
(ne E^ u nj -' )

IlU/b76 SLN3'-iv\inii NIJ.INV'WIV-



- -




--~~---- --

-. .10 C
v -
r~l V







Lyophilized ADGG-BSA from the peak eluting from Sephadex
G-75 (Figure 13) was dissolved in water and diluted to
199 pg/ml of BSA in 0.0001 M Tris HCl pH 7.0.



300 X,nM c00


Results are expressed as the reciprocal of velocity
(pimole 3H-UMP incorporated per minute)-1 versus
concentration of inhibitor for different concentra-
tions of substrate (UTP). For ADGG (Figure 15a) the
squares (U) are 0.008 mM H-UTP and the circles (o)
0.016 mM. For ADGG-BSA (Figure 15b) the concentrations
of 3H-UTP are 0.004 mM (n), 0.008 mM (A) and 0.016 mM
( 0). Lines were drawn as the best fit for least squares
determination of the linear regression.


-1.0 0 1.0

K,- 27 x10- M

-16 -12 -8 -4

4 8 12 16

[(cAMA)x 10' M


polymerase and possessed an apparent K of 6.9 x 10 M.

The inhibition seen with ADGG-BSA clearly deviated from the

strictly noncompetitive, as was the case with ADH-BSA con-

jugates. An average of the three x-intercepts resulted in
an apparent K for ADGG-BSA of 127 x 10-M.

These studies verified the retention by ADGG of the

inhibitory potential of free a-amanitin and demonstrated

that its carbodiimide mediated conjugation to proteins was

feasible. From the previously described attempts at conju-

gation of ADH and Con A, the need for a system to evaluate

the conditions under which successful conjugation of ADGG

to Con A could be obtained became apparent. The optimal

conditions for production of conjugate that retained both

a-amanitin and lectin associated properties required defini-

tion. For this reason and to conserve ADGG, the kinetic

analysis of the EDC mediated binding of 1C-hippuric acid

and Con A was undertaken.

Hippuric Acid-Con A Conjugates


The conditions which would allow for the introduction
of a defined number of 1C-labeled hippuric acid residues

onto Con A with a minimum of protein cross-linking were

determined by measuring the amount of 14C activity preci-

pitated by TCA and collected on GF/C glass fiber filters as

described in Materials and Methods. Figure 16 presents the

results of an experiment in which the effect of varying pH


Reactions were carried out in 1.0 ml volumes of 0.1 M
sodium phosphate buffer containing 1.0 moles of C-
HA (8.5 x 105 dpm/iinole), 10 or 100 moles of EDC and
0.001 moles of Con A. Results are expressed as the
total activity of the TCA precipitable material collected
from duplicate 0.05 ml samples on GF/C glass fiber discs
as described in Materials and Methods.



pH 6
EDC 100
1 pH 5
S-EDC "iO0

0 /


/ .EDC 100
pH 5
iA,...,. /_---\----- --- EDC 10

/.EDC 10
i / .-c-- J-~TM--- o----.- hr 1
ipH 7
S---- EDC 10

1 i 6 18
TIME. hr

and EDC concentrations were examined. Carbodiimide

coupling was performed in the presence of 0.1M phosphate

buffers and in all cases higher concentrations of EDC

(l100mole/ml versus lO0mole/ml) gave greater incorporation.

Both the rate of reaction and maximal amount of labeling

were increased. For a given concentration of EDC, the

lowest pH examined, pH 5, proved optimal with an EDC con-

centration of 10lmole/ml resulting in the incorporation of

10 cpm per 0.001pmole Con A. An EDC concentration of

100pmole/ml yielded a maximum of 3.3 x 10 cpm per 0.001pmole

Con A at the same pH. Increasing the pH to 6 or 7 caused

a decrease in the rate and maximum level of incorporation.

Under all conditions examined here, the reactions were

essentially complete by 2 hours.

The effects of the presence of NaC1, an important

parameter affecting the confirmations of Con A, at two

different EDC concentrations are presented in Figure 17.

No effect on the rate or extent of reaction was observed

in the presence of 0.1M NaC1 in comparison to the aqueous

reaction. Increasing the EDC concentration 10-fold resulted

in an approximate 5-fold increase in the maximum amount

of incorporation as well as an increase in the reaction rate.

The reactions in the absence of phosphate buffer took much

longer to go to completion (18-24 hours versus 2 hours) and

attained 10-fold greater maximal values than for the cor-

responding reaction at pH 5 in phosphate buffer. These data

would argue that phosphate is providing a controlled situation


Reactions were carried out with concentrations of
reactants identical to those described for Figure 16.
The buffer systems for this experiment were 0.1 M NaCl
and 0.01 M NaCl pH 7.2 containing 0.1 mM and 0.01 mM
CaCl2 and MnCl2, respectively.

EDC 100

EDC i00
EDC 100

EDC 10

EDC 10



under which a defined number of acid residues may be intro-

duced onto Con A by EDC.


For examination of the extent to which hippuric acid

conjugates of Con A retain lectin associated properties,

two different conjugates were prepared utilizing the same

concentration of 14C-HA described above. One was prepared

in the presence of 0.1M phosphate buffer, pH 5 and the other

in 0.1M NaCi, pH 7.2. Both were reacted with 10pmole/ml of

EDC for 1 hour. After reaction they were dialyzed and sub-

jected to SDS-PAGE. The results are shown in Figure 18.

The upper gel contains molecular weight standards ranging

from 53,000 to 265,000 daltons. The next gel (2) is a

sample of native Con A and shows the typical 24,000 dalton

subunit with a small amount of the naturally occurring
fragments of Con A. Gel 3 contained 1C-HA-Con A conjugate

prepared in phosphate buffer, pH 5. This conjugate prepar-

ation contained 2.4 residues of hippuric acid per mole of

Con A. The electrophoresis shows a lack of material with

a molecular weight greater than the Con A subunit. This

would indicate that no cross-linking of the protein subunits
has occurred. The bottom gel contained 1C-HA-Con A conju-

gate from the NaCl reaction. The molar ratio of hippuric

acid moieties to Con A for this conjugate was 9.6. This

conjugate contained some degree of cross-linked material,

estimated to be less than 8% of the total by scanning the

gels at 600nm on a recording spectrophotometer. Both

.n --4

r4 --T

k) 4-)
O.) en

-1 -1

,- C)
I O --

0,o: 0 o

0 0 in
S.-, ,-

c- r0

0 0 *H 0

.H* -H

- 4-) r
Sa-n 4-

0 0r(

-4 r- tr0
(U ) (Cd

-24 -4 r O
00 o\rd


0 -H

en-r so

" LY


~ ?.;.



-''' I1C'".
'L.a C.

i ;


i FI

i ~~.;:P

i i

-.-; i


reaction conditions, and pH 5 phosphate buffer in particular,

seem to preserve the native subunit structure of Con A.

The nature of the chemical bond between the hippuric

acid residue and Con A was examined by exposure of the con-

jugates, prepared as described above, to hydroxylamine. The

conditions of exposure were such that hydrolysis of ester

type bonds, that may form as a side reaction during carbo-

diimide coupling (Carraway and Koshland, 1972; Timkovich,

1977), would be complete. The conjugate prepared in

phosphate buffer was 88% stable to hydroxylamine and the

preparation from the NaCI reaction contained 65% stable

bonds. These would presumably represent covalent peptide


The ligand binding activity of the two conjugates as

determined from the ability to bind to Sephadex G-75 and be

eluted with specific saccharide ligand is shown in Figure 19.

Complete retention of saccharide binding activity by this

criterion was demonstrated for both conjugates. They ad-

sorbed to the gel and were eluted by the Con A specific

ligand, 0.1M D-glucose, in volumes identical to native Con A.

A more quantitative determination of the ligand binding

activity of the HA-Con A conjugates was made by determining

the apparent association constant (K ) of each conjugate

for the chromogenic ligand, PNPM. Details of the method

are described in Materials and Methods and the results are

presented in Figure 20. At 220C and under the conditions

of the assay used, native Con A had an affinity constant of