Proc. Nat. Acad. Se& USA
Vol. 71, No. 6, pp. 2530-2533, June 1974

Selection for Neuroblastoma Cells that Synthesize Certain Transmitters

(differentiatiqn/tyrosine hydroxylase/catecholamines/cell culture/genetics)

XANDRA 0. BREAKEFIELD AND MARSHALL W. NIRENBERG

Laboratory of Biochemical Genetics, National Heart and Lung Institute, National Institutes of Health, Beth-da, Maryland 20014

Contributed by Marshall W. Nirenberg, March Z8, 1974

ABSTRACT   A selection procedure was devised for
neurons and relqt+ ce!ls that depends upon the ability
of the cells to synthesize certain amine neurotransmitters.
The rationale for selection f that tyrosine is an essential
amino acid for most marn+aIi+n cells and that three en-
zymes from mammalian sources can catalyze the synthesis
of tyrosine: phenylalanine hydroxylase (EC 1.14.16.1),
tyrosine hydroxylase (EC 1.14.16.2), and tryptophan
hydroxylase (EC 1.14.16.4): Tyrpsipe hydroxylabe is found
predominantly in adrenergic neurons and felated cells
that synthesize dopamine, norepinephrine, and epinepb-
rine, and tryptophan hydroxylase in cells synthesizing
serotonin or melatoqin. Only 1 out of 70,QOO uncloned
mouse neurobl+tqma cells grew well iq the absence of
tyrosine. Appmximatily 50% of the cell lines obtained by
selection l?ad tyrosine hydroxylase actiyity. This selection
procedure thus provides a simple means of obtaining cell
lines of neural origin on the basis of their ability to syn-
thesize putative tranirpitters.

It seems likely that a universal code has evolved which
enables n&rons tq communicate with one another by sending
and receiving information encoded in the form of neurotrans-
mitters. Many steps in the coding of neural information can
be studied with neuroblastoma clones currently available;
however, more'cell lines with additional neural properties are
needed. For this reasov we examined various ways of selecting
cells of neural origin on the basis of their ability to synthe-
size tranqnitters.

In this paper we describe a procedure for selecting adrener-
gic neuroblastoma cells by their ability to grow in medium
lacking tyiosine. The selection procedure depends on the
following phenomena: (a) Tyrosine is an essential amino acid
for most mammalian cell types grown in vitro (1) ; (a) tyrosine
hydroxylase (EC 1.14.16.2; tyrosine 3-monooxygenase) cata-
lyzes the conversion of phenyjalaniqe to tyrosine, as well as
the conversion of tyrosine to 3+dihydroxyphenylalapine
(dopa) (2, 3) (see Fig. 1); and (c) tyrosine hydroxylase
activity is found predominantly in adrenergic neurons and
related chromaf& cells of the adrenal medulia (4). During the
course of these studies, we learned that T. Puck has used a
similar procedure to select for cells with phenylalanine hy-
droxylase (EC 1.14.16.1) (personal commtinication).

MATERIALS AND METHODS

Cell Culttqe. The following cell lines were used: an un-
cloned ljne of mouse neuroblastoma C1300, neuroblastoma
clones N-18, NS-20, and NlE-115 (also referred to as N~l15)
(5), a rat glioma, C-6, derived by Benda el al. (6) (from the
American Type CuJture Collection, no. CCL107), and mouse
L-cells B82 (7). Cells were grown in monolayer culture in
polystyrene petri dishes, 60 and 100 mm (20 and 54 cm*,

respectively) (Falcon Plastics) at 37' in a humidified atmo-
sphere of 10% COS and 90% air. Cells were grown routinely
in either F-14 medium (8) or the Dulbecco-Vogt modification
of Eagle's medium with 4.5 g of dextrose per liter and no Na
pyruvate (GIBCO, Catalogue no. H-21). Both media contain
tyrosine and phenylalani$e and both were supplemented
with 5$& fetal-dalf serum (Colorado Serum). Cells were fed at
l- to 5- day intervals and passaged as described (9).

Cells grown in modified Eagle's medium were adapted to
F-14 by cultjvation in this medium for at least 10 generations
prior to selection. When cloning efficiencies were being de-
termined, cells were trypsinized and triturated to achieve a
suspension of single cells and then plated at varying concen-
trations in either F-14, F-14 minus tyrosine, or F-14 minus
tyrosine ad phenylalanine. When tyrosine alone was omitted,
the pheny!alanine concentration was increased to 0.12 mM.
All media were supplemented with 570 fetal-calf serum dia-
lyzed for 2 days it 3" against 10 volumes of 0.85% (w/v)
NaCl (the NaCl solution w&s changed after 24 hr). Thirty
days after they were plated visible colonies (containing >500
cells) were picked with the aid of porcelain penicylinders
(Fisher). Cloning efficiencies were determined in plates con-
taining <lOO colonies, due to the possibility of cross-feeding.
Cell lines originating from selected colonies were further
grown in F-14 without tyrosine for 1 month and then in either
F-14 or modified Eagle's medium with tyrosine. Cell lines
were free of contamination by pleuropneumopia-like organism
(j'P&O) as determined by culture in plates and broth, and by
autoradiography following incubation with [`Hlthymidine.
For g&&h rate experiments, cells were fed daily, dissociated
by trypsiniz&ion, and counted with the aid of a Coulter
counter.

T~TOS~TW Hydrozylase Assay and Chramatographic Idetiifia-
tiola oj Dopa. Cells were grown in modified Eagle's medium

            COOH    f.cai      COOH
    NH$H   NH2-CH  NH$H
&&%-j
                    HO        HO
           PHE       Tm      DOPA

FIG. 1.  Reactions catalyzed by tyrosine hydroxylase. Oxygen
and tetrahydrobiopt.erin are required for both reactions &s
follows: (1) Ltyrosine + O2 + tetrahydropteridine --* HOH +
c3,4-dihydroxyphenylalanine + dihydropteridine; (6) L-
phenylalanine + 0% + tetrahydropteridine + HOH + G
tyrosine + dihydropteridine.


Proc. Nat. Acad. Sci. USA 71 (1974)                         Selection for Neuroblastoma Cells   2531

with 5yo fetal-calf serum, maintained in stationary phase for
5-10 days, then harvested, homogenized, and assayed for
tyrosine hydroxylase by a modification of the methods of
Nagatsu, Levitt, and Udenfriend (10) and Shiman, Akino,
and Kaufman (3) as described (5,ll).

Product identification was performed as follows. Reactions
contained 0.25 mM N-@-hydroxy-benzyl)-N-methylhydra-
sine (Smith and Nephew Ltd.), an inhibitor of aromatic acid
decarboxylase (EC 4.1.1.28) in addition to the usual compo-
nents. Reactions were deproteinized and neutralized by the
method.of Udenfriend and Zaltzman-Nirenberg (12). Samples
then were passed through alumina oxide columns as described
by Nagatsu, Levitt, and Udenfriend (lo), and the tritiated
reaction product was characterized by thin-layer and paper
chromatography using four solvent systems (13), as listed in
the legend to Table 3.

RESULTS

The Effect of Tyrosine Deprivation on Cell Growth. In this
study we have examined the question "Is the growth of neuro-
blastoma cells with and without tyrosine hydroxylase de-
pendent upon added tyrosine?" The ability of clones N-18
and N-115, without and with tyrosine hydroxylase, respec-
tively, to multiply in complete medium containing both tyro-
sine and phenylalanine, in medium lacking tyrosine, and in
medium lacking both tyrosine and phenylalanine, is shown in
Fig. 2. N-18 cells grew well in complete medium, but not in
medium deficient in one or both amino acids. In the absence
of tyrosine, 50yo of these cells died within 8 days. However,
N-115 cells with high tyrosine hydroxylase activity [980
pmoles of dopa formed per min/mg of protein (5)] did multi-
ply in medium lacking tyrosine after a lag period. Cell survival
was dependent upon phenylalanine, but not upon tyrosine.
An N-115 subline obtained by selection in the absence of
tyrosine grew well both in the presence and absence of tyro-
sine with population generation times of 24 and 54 hr, re-
spectively. Again, cells did not survive in the absence of both
phenylalanine and tyrosine.

The cloning efficiencies of different cell limes in the presence
and absence of tyrosine are shown in Table 1. Neuroblastoma
tumor Cl300 contains at least three cell types with respect to
transmitter synthesis: adrenergic cells, cholinergic cells, and
cells with no known transmitter (5). The adrenergic line,
N-115, formed colonies with almost equal efficiency in media
with or without tyrosine. In contrast, no colonies were ob-
served in medium lacking tyrosine with both a cholinergic
clone, NS-20, with high choline acetyltransferase (EC 2.3.1.6),
but no tyrosine hydroxylaae activity, as well as clone N-18,
which lacks both enzyme activities.

The cloning efficiency of an uncloned line of Cl300 with
little or no tyrosine hydroxylase activity was 16% in the pres-
ence of tyrosine. En the absence of tyrosine approximately 1
in ?`Q,OQO cells gave rise to a colony. Two other clonal lines,
rat glioma C-6 and L-cell clone B82, also did not form colonies
in the absence of tyrosine.

When higher concentrations of Lcells were cultivated
with neuroblastoma N-115 cells with high tyrosine hydroxyl-
ase activity, many colonies were found that contained both
neuroblastoma and L-cells; no colonies were found that con-
tained only L-cells. These results suggest that I.-cells obtain
sufficient tyrosine for growth from the neuroblastoma cells
when the two cell types are in close proximity.

1 A. CLOFjE Ni6  11 b.' CLONE Nll5 1

024  66

Fm. 2. Growth of neuroblastoma calls in the presence and
absence of tyrosine. The growth of neuroblastoma cells with and
without tyrosine hydroxylsse was determined in the following
media: (0) F-14 with tyrosine and phenylalanine (complete);
(A) F-14 minus tyrosine with 0.12 mM phenylalanine; and (0)
F-14 minus tyrosine and phenylalanine. All media contained 5%
dialyzed fetal-calf serum.  In panel A the growth of neuroblss-
toma clone N-18 is shown; in panel B, clone N-115 with high
tyrosine hydroxylsee activity; and in panel C a subline of clone
N-115 (N-T45) derived by selection in medium without tyrosine.
Cells were grown in 60-mm Falcon petri dishes (20 cm*) and
dissociated and counted se described in M&&s.

Charactiation of Neuroblastoma Cells Selected by Tyroeine
Deprivation. Colonies obtained from the uncloned Cl300 cell
line by selection in medium without tyrosine were further
propagated in modified Eagle's medium in the presence of
tyrosine. Homogenates were prepared from cells grown to
stationary phase and assayed for tyrosine hydroxylase activ-
ity. Ten out of the 21 cell lines obtained had tyrosine hy-
droxylase activity (Table 2). The specific activity of the
enzyme varied considerably, as shown.

Homogenates of four of the cell lines and a homogenate
prepared from N-115 were used for product characterization.

TABLE 1.  The e$ect of tyrosine on cell cloning e&kncy

      y. of cells
Tyrosine forming colonies
hydrox-
&se    + -
activitv Twosine Tyrceine

Cell lines

Neuroblastoma C-1300 clones
Adrenergic       N-115    + .i 1.4
Cholinergic       NS-20 -      9   <0.0004
No transmitter      N-18 -      16    <o. 0004
Uncloned neuroblsstoms         -      16     0.0014
Glioma             C-6 -
L-cells           B82 -  2   <o. 0002
                             <o. 0001

The cloning efliciencies'of the following celllines were determined
in F-14 with and without tyrosine plus 5oJ0 dialyzed fetal-calf
serum: neuroblastoma clone N-115 with high tyrosine hydrox-
ylase activity; clone NS-20, with choline acetyltransferase but no
tyrosine hydroxylase activity; and clone N-18 with neither
enzyme; as well as an uncloned line derived from neuroblastoma
Cl.%O, glioma clone C6, and L-cell clone B82. In most cases
lOO-mm oetri dishes were inoculated in quintuplicate with lo(,
106, 104, 102, and 10' cells. Colonies were scored after approxi-
mately 30 days of incubation. Duplicate experiments were per-
formed.


2532  Cell Biology: Breakefield and Nirenberg                 Proc. Nd. Acad. Sci. USA Yl (lOY4)

TABLE 2.  Tyrosine hydrozylase activity of
nevtoblastama lines selected without tyrosine

                   pmol of aH+ released
                   from [*HI tyrosine/min
        Cell lines           per mg of protein

N-TDB                            324
N-T4                            162
N-T7                              101
N-TN10                           74
N-T13                             65
N-TN1                             63
N-T1                              42
N-T12                              34
N-TD4                             33
N-T6                              30
N-T16, N-TDL, N-T8, N-T14, N-T15,
N-T16, N-TN2, N-TNlP, N-TN16,
N-TD7, N-TDlO                   <lO

An uncloned line of neuroblastoma C-1300 with negligible
tyrosine hydroxylase activity (<lO pmoles of IH+ released from
[8H] tyrosine per min/mg of protein) was grown in F-14 medium
without tyrosine for 30 days (see Table 1). Well-isolated colonies
were picked, grown, and assayed for tyrosine hydroxylaae as
described in Methods.

Tyrosine hydroxylase reactions were supplemented with
0.25 mM N-(3-hydroxybenzyl)-N-methylhydrazine, an in-
hibitor of aromatic amino-acid decarboxylase. Reaction prod-
ucts were eluted from alumina oxide columns and character-
ized by thin-layer and paper chromatography with four
solvent systems (Table 3). The major reaction product was
identical in chromatographic mobility to authentic dopa.

  TA~LX 3.  Chromatagraphic churucferizalion of
the SH-labeled pwduct of the tyrosine hydrozylase reaction

Solvent systems

Cell lines

A      B      C      D

`% of the aH-labeled product
  with the mobility of dopa

N-TD6        95      89      95      89
N-T4        96      85      97      85
N-T7         95      79      94      91
N-T1         94      87      95      89
N-115        99      95      95      93

Tyrosine hydroxylase reactions were performed ss described in
Methods. Tritiated catechol products were characterized by thin-
layer chromatography wit.h the following solvent systems: so&al
A, 1-butanol-concentrated acetic acid-water (12: 3: 5); solvent B,
methylet,hylketone-concentrated formic acid-water (24: 1: 6);
solvent C, ethylacetate-concentrated acetic, acid-water (15:-
15:lO); &vent D, I-butanol-1 N acetic acid-absolute ethanol
(35: 10: 10). Solvent A was used with Whatman No. 1 filter paper;
solvents B, C, and D were used with thin-layer cellulose plates
(Eastman 6064). The RF values of authentic dops were similar
to those reported previously. One hundred percent corresponds
to the following dpm recovered after chromatography: 6060,
N-TDG; 7860, N-T4; 7550, N-T7; 9590, N-Tl; and 17,200, N-115.
The numbers in the table correspond to the percent of tritiated
product recovered with the chromatographic mobility of authentic
dopa.

In addition, several unidentified radioactive products were
found, which represented <lOTo of the tritiated reaction
products. In other experiments performed with Tom Lloyd,
3-iodotyrosine and antiserum prepared against purified bovine
adrenal medulla tyrosine hydroxylase (14) were found to
inhibit the tyrosine hydroxylase activity of N-TD6 homog-
enates in a dose-dependent manner. These results, together
with studies reported previously (5), strongly suggest that
dopa synthesis is catalyzed by tyrosine hydroxylase.
Some, but not all, of the cell lines obtained in the absence
of tyrosine showed catecholamine hi&fluorescence induced by
formaldehyde, suggesting that catecholamines are both syn-
thesized and stored (David Jaoobowits, paper in preparation).
A small proportion of cells from the Cl300 neuroblastoma
tumor grown in r&o (15) and in Vitro (16,17) have been shown
to exhibit catecholamine fluorescence.

DISCUSSION

The results show that neuroblastoma cell lines differentiated
with respect to adrenergic neurotransmitter synthesis can be
selected on the basis of their ability to grow in medium lacking
tyrosine. Clones derived from various cell types, including
glis and fibroblssts, were tested, but multiplication in the
absence of tyrosine was found only in neuroblastoma cells
with tyrosine hydroxylase activity. When uncloned popula-
tions of neuroblastoma cells were selected by this procedure,
approximately one out of 7 X IO' cells gave rise to a cell line
that multiplied well without tyrosine. Approximately 5070
of the cell lines thus obtained had tyrosine hydroxylase
activity.

Shiman, Akino, and Kaufman have shown that tyrosine
hydroxylase preparations catalyze the synthesis of tyrosine
and its conversion to dopa at similar rates in the presence of
tetrahydrobiopterin, the in uivo cofactor (3). Studies of brain
metabolism in viva (19, 20) and in synaptosome preparations
(21) suggest that phenylalanine is converted to tyrosine in the
normal nervous system. The ability of tyrosine hydroxylase to
catalyze tyrosine synthesis may be advantageous since adren-
ergic neurons thus can generate the substrate needed for
transmitter synthesis by an independent endogenous route.
The observation that adrenergic neuroblastoma cells support
the growth of neighboring cells in the absence of tyrosine
suggests that tyrosine is released by neuroblastoma ceils and
has a tropic effect upon L cells which determines the relative
positions of these two cell types. The interaction of certain
cells in the normal nervous system may be determined in a
similar manner.

It should be possible to select dividing, as well as nondivid-
ing, cells on the basis of their ability to multiply or survive in
the absence of tyrosine. In addition, the selection procedure
coupled with somatic cell hybridization affords a means of
generating and selecting different neural phenotypes on the
basis of their specificity in encoding neural information.
Ultimately, it may be possible to obtain five classes of cells
with respect to putative neurotransmitters, those that syn-
thesize dopamine, norepinephrine, epinephrine, serotonin,
or melatonin, since tryptophan hydroxylase also catalyzes
the formation of tyrosine from phenylalanine (21). E. Held-
man in our laboratory has shown that one of the cell lines,
N-T16, that grows in the absence of tyrosine and does not
have tyrosine hydroxyiase activity, does synthesize sero-
tonin (unpublished observations).


Proc. Nat. Acad. Sci. USA 71 (1974)                         Selection for Neuroblastoma Cells   2533

We gratefully acknowledge the help of Elliot Richelson, Perola
Nirenberg, John Minna, and Tom Lloyd.

11.

Schrier, B. K., Wilson, S. H. & Nirenberg, M. (1974) in
Methods in Enzymology, eds. Fleischere, S., Packer, L. &
Estabrook, R. W. (Academic Press, New York and London),
in press.
Udenfriend, S. & Zaltzman-Nirenberg, P. (1964) fife Sci. 3,
69.5-702.

10.

Eagle, H. (1955) J. B&Z. Chem. 214,83Q-852.
Ikeda. M.. Levitt. M. & Udenfriend. 5. (1965) Biochem.
Biqohys. &es. C&mun. 18, 482-488. '
Shiman, R., Akino, M. & Kaufman, S. (1971) J. Biol. Calm.
246, 1330-1340.
Kaufman. S. (1962) in Methods in Ensmncloau. eds. Colowick
S. P. t Kaphxn, N. D. (Academic Press,""New York and
London), Vol. 5, pp. 809-816.
Amano. T.. Richelson, E. & Nirenbera. M. (1972) Proc. Nat.
Acad. i&i. bSA 69, 2%263.     -.

Benda, P., Lightbody, J., S&o, G., Levine, L. & Sweet, W.
(1968) Science 161, 370-371.
Littlefield, J. W. (1966) Ezp. Cell Res. 41, 190-196.
Vogel, Z., Sytkowski, A. J. t Nirenberg, M. W. (1972)
Proc. Nat. Acad. Sci. USA 69,3180-3184.
Blume, A., Gilbert, F., Wilson, S., Farber, J., Rosenberg, R.
L Nirenberg, M. (1970) Proc. Nat. Acad. 5%. USA 67,786-
792.
Nagatsu, T., Levitt, M. & Udenfriend, S. (1964) J. Bill.
Chem. 239,2QlO-2917.

12.

13.

14.
15.

16.

17.

18.
19.


20.


21.

Richelson, E. & Niienberg, M. (1974) in Methods in Eney-
nwloey, eds. Fleischere, S.. Packer, L.. & E&brook. R. W.
(Academic Press, New.York and London), in press.'
Lloyd, T. &Kaufman, S. (1973) Mol. Pharmucol. 9,438-444.
Anagnoste, B., Freedman, L. S., Goldstein, M., Broome, J.
& Fuxy, K. (1972) Proc. Nat. Acud. Sci. USA 69,1883-1886.
DeLelhs, R. A., Rabson, A. S. & Albert, D. (1970) J. Hz&-
&m. Cytochem. 18,913-914.
Hermeter, J. C., Ciesieiski-Treska, J. & Mandel, P. (1972) J.
Hietochem. Cytochem. 20, 137-138.
Bag&i, S. P. & Zarvcki. E. P. (1970) Life Sci. 9. 111-119.
Bagchi S. P. & Zarycki; E. P. (1973) B&hem. l$armaccl.
22. 1353-1368.

Karobath, M. & Baidessarini, R. J. (1972) Nc&re New Bill.
236,206-208.
Jequier, E., Robinson, D. S., Lovenberg, W. & Sjoerdsma,
A. (1969) &e&em. Phatmucol. 18, 1071-1081.