Nerve growth factor receptors: analysis of the interaction of ßNGF with membranes of chick embryo dorsal root ganglia

Brain Research, 199 (1980) 63-77 © Elsevier/North-Holland Biomedical Press

63

NERVE G R O W T H F A C T O R RECEPTORS: ANALYSIS OF T H E INTERA C T I O N OF flNGF W I T H M E M B R A N E S OF C H I C K EMBRYO DORSAL ROOT G A N G L I A

RICHARD J. RIOPELLE, M1CKI KLEARMAN and ARNE SUTTER Division of Neurology, Queen's University, Kingston K7L 3J7 (Canada) and ( M.K. and A.S.) Department of Neurobiology, Stanford University, School of Medicine, Stanford, Calif. 94305 (U.S.A.)

(Accepted April 17th, 1980) Key words: negative cooperativity - - Nerve Growth Factor - - dorsal root ganglia - - chick embryo

SUMMARY The binding of the fl subunit of Nerve Growth Factor (flNGF) to membrane preparations of 8-day chick embryo dorsal root ganglia (DRG) has been investigated under conditions similar to those used to study the binding of flNGF to intact single cell dissociates of D R G 2a. The equilibrium binding data reveal heterogeneity of binding that is more complex than that seen with intact cells. Binding is not saturable up to 125IflNGF concentrations of 10-8 M. Steady-state and kinetic binding data show two sites with dissociation constants similar to those found on D R G cells. In addition, displacement data reveal a binding component with lower affinity (Ka = 10-8 M) which is not found on intact cells. As with intact cells, the difference in the affinities of the two high affinity sites has been shown to be due to different rate constants of dissociation. The kinetics of dissociation of N G F are slower with membranes than with cells, and dissociation characteristics of I~5IflNGF change with increasing time of exposure to membranes. Degradation of 125IflNGF during incubation with membranes is minimal and does not complicate the analysis of steady-state binding. Insulin does not bind to either of the two high affinity sites. Heterogeneity of the 125IflNGF preparation and cooperativity of binding as a cause for the heterogeneity of the binding of N G F has been ruled out. Although there was an apparent increase in the rate of dissociation of I~sIflNGF in the presence of unlabelled N G F , a finding previously interpreted as evidence for negative cooperativity 7, this was shown to be independent of receptor site occupancy by N G F , and in part due to isotopic dilution within a diffusion barrier around the membranes.

64 INTRODUCTION N G F is a peptide hormone 3 which promotes morphological and biochemical differentiation of sensory and sympathetic neurons lz. Both sensory and sympathetic cells carry specific cell surface receptors for N G F 1,7,9,z3. However, there are certain inconsistencies in published data that may reflect differences in the tissue preparation used in the assay, the assay procedure, and the method used to prepare lzsIflNGF, possibilities which have been discussed previously 15,23. To circumvent some of these problems, the present study was carried out with a membrane preparation of 8-day chick D R G cells using the same assay conditions and techniques, and the same ~zsIflN G F preparation as was used for the experiments with intact cells23. The data presented here indicate that binding to membranes is more complex than that seen with intact cells. There are at least 3 saturable binding components, only two of which are present on intact cells. This heterogeneity cannot be explained by negative cooperativity within a homogeneous class of receptor sites. Some data supporting a model ofligand-induced non-cooperative conformational changes in receptor molecules as a explanation of the heterogeneity in high affinity binding is provided. The absence of saturable low affinity (Ku = 10 -6 M) binding on intact cells suggests that this component on membranes is a distinct intracellular binding site for NGF. MATERIALS AND METHODS All materials and methods used were as previously described 23 with the following exceptions: (a) human growth hormone, sheep growth hormone, and sheep lactogenic hormone were generous gifts of Dr. Morris Krahl, Stanford University; (b) membrane preparation: 8-day chick embryo spinal ganglia were dissected into ice-cold calciumand magnesium-free phosphate-buffered Gey's salt solution (pH 7.4), taking care to remove connective and meningeal tissue prior to ganglia removal. The ganglia were washed 3 times at 4 °C in the same buffer. The final washed pellet was resuspended in 10 vols. of a hypotonic lysing solution consisting of 2 mM NaHCO3, 0.2 mM CaCI2, and 5 mM MgClz (pH 6.8), and incubated at 4 °C for 7-10 min. Ganglia were then disrupted into membranes by homogenization in a tight-fitting Dounce homogenizer. Completion was determined by the absence of whole cells as monitored by phase contrast microscopy. Plasma membrane isolation of D R G was performed by a modification of the two-phase polyethylene gtycol-dextran partition method of Brunette and Till 2. Stock solutions of the two phases were prepared as follows: 103 g of 30 ~ (w/w) polyethylene glycol 6000 (BDH Chemicals) in distilled water, 200 g of 20 ~o (w/w) dextran T-500 (Pharmacia) in distilled water, 333 ml of 0.2 M sodium phosphate buffer (pH 6.4), 172 ml distilled water and 7.2 ml of 0.5 M MgCI2 solution. The solutions were mixed in a separatory funnel, allowed to separate for 48 h in the cold, collected and refrigerated. The ganglia homogenates were centrifuged at 10,000 × g for 15 min and each pellet suspended in 21 ml of upper phase (dextran saturated polyethylene glycol) by vigorous stirring for 10 rain. Seven millilitre aliquots were pipetted into 15 ml conical centrifuge tubes. To each tube, 7 ml of bottom phase polymer (polyethylene glycol

65 saturated dextran) was added. The polymers were gently mixed and centrifuged at 1100 × g in a swinging bucket rotor (Sorval GLC2) for 15 min at 4 °C. Membranes appeared as thin white sheets at the interface of the two polymers, while cell debris pelleted out. These sheets were removed with a Pasteur pipette and resuspended in equal vols. of top and bottom phase polymer. The procedure was repeated 3 times. The final sheets of membrane were then washed by suspension in 30 ml of calcium-, magnesium-free buffered Gey's salt solution (pH 7.4) and centrifuged at 8000 × g for 10 min. After 3 washes, final pellets were suspended in 1-2 ml of calcium-, magnesium-free phosphate buffered Gey's salt solution (pH 7.4), and protein concentration was determined by the method of Lowry et al. 14 using bovine serum albumin as a standard. Membrane solutions were quick frozen in a mixture of acetone and dry-ice and stored at --80 °C until use. The purity of the isolated membrane fraction was estimated using electronmicrography kindly carried out by Dr. Mary Herman, Department of Neuropathology, Stanford University; (c) for the suspension agitation experiments, 25 ml Erlenmeyer flasks, and a New Brunswick Scientific Gyrotory Water Bath Shaker, Model G76, setting 5.5, were used; and (d) for gamma emission counting carried out at Queen's University, a Model 1185 Automatic Gamma System (Searle Analytic) was used. RESULTS

Characteristics of the 1251flNGFpreparation These have been discussed previously23.

Characteristics of the tissue preparation The procedure used for membrane preparation in this study was originally described for the purification of plasma membranes from L-cells. Although this technique used with 8-day embryo DRG resulted in material that was enriched in plasma membrane, electron micrographs revealed some contamination by nuclei, rough endoplasmic reticulum, lysosomes, and other cellular components. The yield from the preparation was calculated at approximately 0.57 #g prot./ganglion. Serial binding studies performed with the preparation following storage at --80 °C indicated that the material remained stable in the frozen state for 5-6 weeks. For any concentration of I~5IflNGF, binding was found to be linearly related to membrane protein concentrations up to 70 #g/ml (data not shown). Except where noted, binding experiments were carried out at a final membrane protein concentration of 30 #g/ml.

Equilibrium binding data The procedures used for these experiments were the same as previously described 29. Using the Beckman Microfuge B, it was found that at least 20 sec were required to pellet the membrane-NGF complex, and routinely 1 min centrifugations were used. The concentration dependence of binding of I~5IflNGF to DRG membranes is shown in Fig. 1a. This binding was not saturable even at concentrations of 1.13 × 10-8 M 125IflNGF. Fig. lb depicts the analysis of the initial part of the concentration curve

66 according to Scatchard ~8. The Scatchard analysis revealed two saturable binding components where binding in the presence of 3.8 × 10 -8 M or 3.8 × 10-7 M unlabetled flNGF was defined as non-displaceable (non-specific) binding. As will be shown later, this so-called non-displaceable binding was in fact displaceable using high concentrations of unlabelled flNGF. The two saturable binding components have equilibrium dissociation constants (Kg) that are two orders of magnitude apart (Kgl = 2.9 ± 1.4 × 10 -H M, and Ka,~ -1.07 ± 0.4 × 10 -9 M). The apparent site numbers/#g prot. when sucrose gradients were

!

~3000 izooo 11ooo /

J

J

ioooo

~

7000

Z6000 5000 4000

o

2ooo iooo

125~NGF { nq/ml) .I4'

t

I

I

~ _ _

le J2

30

.08

z

06

Z

04-

•

02-

b

I __

°

I

~-.

2bo

45o

65O

800

I000

L25 .bS'NGF SPECIFICALLY BOUND (pg/ml)

Fig. 1. Binding of lzsIflNGF to membrane preparation of 8-day chick embryo sensory ganglion cells. a: binding as a function of concentration of 125IflNGF. b: Scatchard plot of binding of I ~ I f l N G F where non-specific binding was defined as binding in the presence of 3.7 × 10 -7 M unlabelled flNGF. Membranes were incubated at 37 °C for 55-60 min with various concentrations of 12ZlflNGF. Bound 125IflNGF was separated from free by centrifugation through sucrose gradients as previously described 2a. The means and S.D.s of three 100 pl aliquots for each point were determined: (3, total binding; O, binding in the presence of 3.7 × 10 -7 M flNGF; I , 'specific' binding. In this experiment the spec. act. of Z2SlflNGF was approximately 75 cpm/pg.

67 TABLE I Degradation of l~ l flNGF by sensory ganglion membrane

Membrane at a protein concentration of 60 pg/ml was incubated with 125IflNGF (1.1 × 10-11 M) at 37 °C. At the times indicated, triplicate 100/~1 aliquots were removed for measurement by centrifugation through sucrose gradients as previously described23; at the same times 350/~1 samples were withdrawn, centrifuged for 1 min at 10,000 × g, and 100 pl aliquots of supernatant were mixed with an equal volume of 20 ~ w/v trichloroacetic acid (TCA). The samples stood on ice for 60 min and were then centrifuged at 10,000 × g for 2 min; supernatants and pellets were counted separately. In the above experiment, the TCA soluble fraction of the stock lzsIflNGF solution was 0.07. The fraction degraded is the difference between the TCA soluble fraction measured, and the TCA soluble fraction expected in the supernatant as a result of binding of 125IflNGF that is not 100 ~ TCA precipitable. Time (min)

Fractionb o u n d

Measured TCA soluble fraction

Expected TCA solublefraction

Fraction degraded

20 40 80

0.089 ::k 0.002 0.102 ± 0.001 0.094 ± 0.002

0.070 ± 0.001 0.082 ± 0.002 0.085 ± 0.002

0.077 ± 0.002 0.078 :k 0.002 0.078 ± 0.002

0.0 0.004 0.007

used were 2.61 -4- 1.3 pg for site 1, and 20.4 ± 6 pg for site 2. Site ratios when the sucrose gradient system was used were approximately 1 : 6-9. The mean and standard deviations are based u p o n data f r o m 7 experiments. Table I shows that even at high concentrations o f m e m b r a n e protein (60 #g/ml), degradation o f 125IflNGF, after 80 min at 37 °C, measured by 10 70 trichloroacetic acid precipitability o f supernatant, was negligible, and it was therefore not necessary to make corrections for steady-state binding data. Displacement data

Displacement o f 125IflNGF at a fixed concentration with increasing concentrations o f unlabelled f l N G F is shown in Fig. 2. At a concentration o f 1.8 × 10 -9 M lzsIflNGF, the first decrement is half maximal at 3 × 10 -9 M, which correlates well with a dissociation constant o f site 2 of 10 -9 M. At concentrations o f unlabelled f l N G F between 10 -s and 10 -7 M, the displacement curve showed a plateau, and then a further decrement with 50 ~ o f the plateau value reached at 10 -6 M. This third binding site was not reflected in the Scatchard analysis o f the equilibrium binding data in Fig. l b because non-specific binding was considered to be binding in the presence o f 3.8 × 10 -s M or 3.8 × 10 -7 M unlabelled flNGF. Displacement data using intact cells are plotted in Fig. 2 as well and reveal that the saturable low affinity binding c o m p o n e n t seen on membranes is virtually absent on intact cells. Insulin over a concentration range o f 10 -9 M to 10 -6 M did not compete with I~5IflNGF to either site 1 (125IflNGF 2.3 × 10 -11 M) or site 2 (I~5IflNGF 3.9 × 10 -1° M). H u m a n growth h o r m o n e , sheep growth h o r m o n e and sheep lactogenic h o r m o n e over the same concentration range did not compete for binding to site 2 (125IflNGF 3.9 x 10 -l° M) (data not shown).

68 w

___

I

I

LL I 0 0 "

_~ 9o800 -~

70-

60-

z 40"63ol,z

I1~

~

~...__~__.~~

2010-

io-9

io'-O TOTAL

16-T

1~6

,X~NGF CONCENTRATION ( M )

Fig. 2. Comparison of binding of 125IflNGF to 8-day sensory ganglion membranes (©), and cells (1.5 × 106/ml) (Q). Incubations were carried out for 60 min at 37 °C in the presence of 1.8 × 10 9 M 125IflNGF and increasing concentrations of unlabelled flNGF. Means and S.D.s of triplicates of 100 td aliquots processed as described previously 23 are plotted.

Kinetic binding data Association kinetics. To determine the association rate constant for site 1, concentrations ofl25IflNGF of 1.14 × 10 -11 M to 2.43 × 10 -11 M were used. At these concentrations, over 70 ~ of the binding occurred to site 1. The association data shown in 280 260240880200CJ Z

180-

0

160-

h

140-

8D _1 U

120" I00-

n

80" Z

60" 4020" 0 0

,~, ;o ~o go ~'o ~o /o ~o ;o ,~,;o

,~o

MINUTES

Fig. 3. Time course at 37 °C of association of 125IflNGF (1.8 x 10-la) to membrane at a protein concentration of 60 #g/ml. Data are corrected for binding in the presence of 3.7 × 10 -v M unlabelled flNG F, and each point is the mean and S. D. of triplicate 100 HI aliquots centrifuged through sucrose gradients for 1 min as previously described 23.

69 Fig. 3 reveal that the process was completed within 50-60 min at 37 °C. By measuring initial reaction velocities derived from initial slopes TM, the association rate constant was calculated to be 3.65 4- 1.4 × 107/M/sec, this high value suggesting that association is a diffusion controlled process. Difficulties with measurement of an association rate constant for site 2 independent of site 1 have been discussed previously 28. In the case of the membrane preparation, analysis is further complicated by the presence of a third site calculated from Fig. 2 to have a capacity approximating 8000 pg//~g. Even with the use of techniques that allowed centrifugation to begin within one second of mixing membrane and 125IflNGF, the association plot was no longer linear by two seconds, and thus, it was not possible to determine an initial slope when high concentrations of 125I/3NGF (9.7 × 10-10 M) were used. Dissociation kinetics. In these experiments, dissociation was measured following the addition of excess amounts ofunlabelled/3NGF, (3.8 × 10-s M or 3.8 × 10-7 M). At all concentrations of 125I/~NGF, the dissociation plot was biphasic. Fig. 4 shows the dissociation pattern at 1.85 × 10-11 M 125I/3NGF (upper) and 2.3 × 10-l° M lzsI/3NGF (lower). Part of the bound 125I/3NGF was released very rapidly, and with the centrifugation techniques employed, it was not possible to measure any points on this part of the dissociation curve; the remainder of the bound lzsIflNGF was released more slowly. For very low concentrations of 125I/3NGF, binding occurred predominantly to site 1, and release from this site followed first order kinetics with a half-time of 18.8 4- 2.8 min (8 experiments). The proportion of bound lzsI/3NGF which was released rapidly increased with increasing concentrations of 125I/~NGF used to pre-equilibrate the cells, and therefore, with increasing occupancy of site 2 and the third site. I00~

i

i

J

I

I

8oO

"6 70

rn

z~ 40-

~

i

O

Z I.tJ C) tY t.d 0_

20-

IO

o

,b

z'o

MN I UTES

3'o

4b

Fig. 4. Dissociation o f 125IflNGF f r o m m e m b r a n e s at 1.85 × 10 -11 M (C)), a n d at 2.3 × 10 -1° M ( 0 ) . A f t e r a p r e i n c u b a t i o n o f 55 rain at 37 °C, dissociation was initiated by addition o f 3.7 x 10 .7 M u n labelled f l N G F . Specific binding at to was m e a s u r e d in quadruplicate a n d at various times thereafter in triplicate, u s i n g centrifugation techniques as previously described 23. All data are corrected for binding in the presence o f 3.7 × 10 -7 M unlabelled f l N G F .

70 IO01~"

8°f 60

4C 3C o ~6

20

8 z

m~ "6 i,i

i

I

0

~o

t

20

3o

40

50

60

MINUTES

Fig. 5, Time course of dissociation of 12~IfiNGFfrom membranes --effect of preincubation time. Membranes were preincubated with 125IflNGF(2.8 x 10~1° M) for 40 min (©), and 80 rain (O) at 37 °C. Dissociation was initiated at those times by addition of 3.8 × 10 7 M unlabelled flNGF, and binding at various times thereafter measured as described in Fig. 4. The release of 125IflNGF from site I has a half-time of approximately 19 min, and complete release from site 2 occurs in less than 50 sec; hence, the rate constants for dissociation are more than two orders of magnitude apart (K 1~ = 6.28 ! 0.93 × 10-4/ sec and K 12 > 5 × 10-Z/sec). Further analysis of dissociation kinetics reveals that once steady state binding of 125IflNGF (as defined by association kinetics) is achieved, secondary processes take place. Fig. 5 illustrates the time dependence of dissociation kinetics of 12HflNGF on pre-equilibration time. At a concentration of 2.8 × 10 10 M 125IflNGF, binding reaches a steady state within 20-30 min at 37 °C (data not shown). The lower plot shows the dissociation kinetics of lzsIflNGF from membranes pre-equilibrated with the label at 2.8 x 10 -~° M for 40 min, while the upper plot is the dissociation kinetics following a pre-equilibration time of 80 min. Thus, with increasing time, the proportion of bound lzsIflNGF which is released rapidly decreases, while that which is released with a halftime approximating 19 min increases. This indicates that the occupancy ratio of site 1 to site 2 (or the third site) increases with length of exposure to 125IflNGF.

Cooperativity DeMeyts et al.5, 6 used the concept of negative cooperativity to explain site heterogeneity observed in Scatchard analysis of insulin binding to its receptors. It was proposed that receptor site occupancy by hormone determined the affinity

71 of binding of the receptor for the hormone either as a result of ligand-ligand interaction or ligand-induced interaction between receptors or their subunits. In a negatively cooperative system, the affinity of binding would decrease as receptor occupancy increased. To test this theory, DeMeyts et al. suggested that the rate of dissociation of labelled hormone from receptor in the presence of excess unlabelled hormone would be faster than dissociation of label where excess hormone was not present, dissociation being initiated in both cases by dilution. For insulin, this was shown to be the case, and there followed experiments with other hormonesT,S,11 and neurotransmitters 1°,13 demonstrating the same phenomenon. Two basic requirements for proof of negative cooperativity are as follows: (1) if the hormone-receptor interaction displays negative cooperativity, the rate of dissociation should be independent of the receptor occupancy prior to the initiation of dissociation, if dissociation occurs in the presence of a large excess of unlabelled hormone, because, due to the added excess of hormone, all available receptors will be occupied and will be in the low affinity state, (2) if the hormone-receptor interaction displays negative cooperativity, when receptor site occupancy is increased the rate of dissociation should be more rapid than when receptor occupancy is decreased. In other words, where, even in the presence of unlabelled hormone, the actual receptor site occupancy is less than that reached during preincubation with trace amounts of labelled hormone insufficient to cause the cooperative change in receptor conformation, the cooperative state should not be induced and the rate of dissociation should not be affected. With respect to these dissociation studies, no investigators proposing negative cooperativity have directly tested the role of site occupancy in dissociation behaviour. The examination of the time course of dissociation at two different preincubation concentrations as seen in Fig. 4, revealed that in this system the first requirement for a negatively cooperative state could not be met. Dissociation in both cases was initiated by 3.8 × 10 -7 M unlabelled IflNGF, but in spite of this, the pattern of dissociation was dependent upon the concentration of lzsIflNGF used to pre-equilibrate the membranes, and thus, dependent upon the relative occupancy of sites 1 and 2 prior to dissociation. To examine the role of site occupancy at the time of dissociation, membranes prelabelled with 125IflNGF were diluted into medium alone, or into medium containing unlabelled flNGF, and the release of 125flNGF was followed. The dilutions (25 ×-150 ×) were sufficient to prevent measurable rebinding of 125IflNGF at the concentrations used (1.9 × 10 -11 M), (data not shown). The time course of dissociation following dilution into medium and medium plus unlabelled flNGF is seen in Fig. 6. These data showed that there was an apparent increase in the rate of dissociation in the presence of unlabelled flNGF, data suggestive of a negatively cooperative state. However, to directly test the role of site occupancy in inducing this state, the dissociation experiments were carried out using a concentration of unlabelled flNGF (1.1 × 10 -11 M) which was lower than the 12~IflNGF concentration used for prelabelling of the membranes. Fig. 6 also reveals that, even while receptor occupancy in the presence of unlabelled flNGF was decreased relative to the occupancy reached in the preincubation, the dissociation rate was still increased relative to that measured in the absence ofunlabelled flNGF. Hence, the second requirement for a negatively cooperative state, that site occupancy determ-

72 I00-

I

I

I

I

Wo

4b

sb

l

60-

~9 5@ r'~ Z (~)4003 ta_ (..9 Z ~:30-

kL 0 Z klA ~C~ 2 0 klA 13.-

,b

~o

MINUTES

Fig. 6. Dissociation of lzsIflNGF in the presence of no unlabelled flNGF (@), 1.1 × 10 11 M unlabelled flNGF (Ak), and 3.7 × 10-7 M unlabelled flNGF (11). Membranes at protein concentration of 60/~g/ml were incubated for 60 rain at 37 °C with 1.9 × 10- n M 125IflNGF. Dissociation was initiated by 30-fold dilution of aliquots into buffer. All data are corrected for binding in the presence of 3.7 × 10-7 M flNGF; binding at to was measured in quadruplicate, while at all other times binding was measured in triplicate by techniques previously described23. ines the negatively c o o p e r a t i v e state, could n o t be met in this system, because the rate o f dissociation was increased even in situations where site o c c u p a n c y at the onset o f dissociation was n o t increased. But an e x p l a n a t i o n was still required for the finding t h a t unlabelled f l N G F resulted in a n a p p a r e n t increased rate o f dissociation. T o test the possibility t h a t 'retention effects' or 'unstirred layers'4,19,2~ might a c c o u n t for these findings, a n d t h a t the a p p a r e n t increase in dissociation rate was the result o f i s o t o p i c dilution o f I~51flNGF by u n l a b e l l e d f l N G F within a diffusion barrier, experiments were carried o u t in which the m e m b r a n e p r e p a r a t i o n was stirred in an a t t e m p t to disturb the barrier. Fig. 7a shows the dissociation p a t t e r n when the mixtures were n o t agitated, a n d Fig. 7b shows dissociations d u r i n g agitation. The lower curves in Fig. 7a a n d b (with N G F ) show the same half-time o f dissociation, while the u p p e r dissociation p a t t e r n (without N G F ) was m o r e

73 bOO

I

,

t

I00

I

I

I

I

t

I

90-

90 80-

,o \i t

70-

60-

60~

\

50-

5o- ~ t ~

40-

~ 40Z

1

m

Z

~ 2,0-

\

Ld (~

~UA 2 0

20"

U C~

o

,o

,b

2'o

go

MINUTES

.;o

5o

b

%

,b

~o

go

.'o

50

MINUTES

Fig. 7. Dissociation of 125IflNGF in the presence of no unlabelled flNGF ( 0 ) , and in the presence of 3.7 × 10 -7 M unlabelled flNGF (m). Preincubation and dissociation conditions were similar to those described in Fig. 6. a" no agitation, b: agitation in a gyrotory shaker bath (see Methods and Materials).

rapid when the mixture was stirred. These data and data from 4 other similar experiments provide evidence for the presence of an unstirred layer around the membrane that can be disrupted by physical forces, and explain, at least in part, the differences observed in the dilution experiments. DISCUSSION

The earlier observation that binding of NGF to embryonic chick sensory ganglia is heterogeneous has been confirmed by the data presented here. Agreeing with the results of Sutter et al. 2z, these data support the concept of two saturable receptor sites specific for flNGF and in agreement with the work of Frazier et al. 7, where ganglia homogenates were used, binding to membranes is heterogeneous and not saturable even at concentrations of 125IflNGF where binding sites with equilibrium dissociation constants equal to or less than 10-9 M should be completely occupied. In displacement experiments, it was shown that this third binding component was saturable with a dissociation constant of approximately 10-6 M. Differences in the techniques employed to measure binding of NGF by different investigators have been discussed previously by Mobley et al. 15 and by Sutter et al. 23. The present studies can be compared most directly to the studies of Sutter et al. 23,

74 because the only difference was in the state of the 8-day embryonic chick sensory ganglia target tissue used - - single cells in the earlier work by Sutter et al. "-'3,and a membrane preparation in the present study. Nevertheless, significant differences have been observed. The kinetics of dissociation of v~sIflNGF from membranes is significantly slower than from intact cells, approximately 2-fold for site I. This observation has been made previously in studies of insulin binding to a number of tissues 5, but no explanation has been forthcoming, and we have not explored the finding here. The second observation relates to receptor site numbers detected using intact cells and membranes. Assuming a yield of 40,000 cells per ganglion, the receptor numbers of sites 1 and 2, using membranes are approximately 25-30 ~ of those found on cells 2a, and significantly less than that observed by Frazier et al. v. This is perhaps not surprising considering that virtually all ganglion material was assayed in the procedures of Sutter et al. and Frazier et al., whereas the preparative procedure for the DRG membranes described here resulted in loss of membrane material during initial centrifugation, and loss from the interface region during successive washings. We have not addressed this problem of receptor loss or the effect of the polymers on receptor integrity during the preparative procedure. Third, the use of membranes reveals the existence of another saturable site with an equilibrium dissociation constant approximating 10-6 M. It is unlikely that this receptor site is an artefact introduced by the use of polymers in the membrane preparation, since binding with a dissociation constant approximating 10-6 M was detected by Frazier et al. in equilibrium binding studies using ganglion homogenates prepared by homogenization of the tissue v. What is more likely, as suggested by Sutter et al. 23, is that cellular disruption exposes internal binding sites for N G F with different characteristics. Finally, degradation of 1251flNGF in the presence of membranes differs from that observed with cells. At a membrane protein concentration of 60/~g/ml, and 12.51flNGF concentrations of 1.9 × 10-11 M, degradation, measured by TCA precipitability of supernatant as previously described 23, was always less than 1 ~,, after 80 min incubation at 37 °C. Under the same conditions, degradation of a25IflNGF in the presence of an equivalent number of cells (4 x 106/ml) was always greater than 10 ~ of the total 125IfiNGF. Because high affinity receptor numbers on membranes are only 25-30 )o of that observed on cells, the true cell equivalent is only I >:: 106/ml instead of 4 7: 106/ml. At cell concentrations of 1 ;< 10°/ml with 1.9 × 10-11 M ~5IflNGF, it was possible to detect degradation at 60 min at approximately the 5 ~ level (data not shown), so that the discrepancy is still not completely explained when receptor site numbers are compared. Other possibilities that have not been explored include loss of degradative activity as a result of exposure to intraceltular molecules, loss of a lysosomal component of degradation, or loss of degradative activity as a result of exposure to polymer; in addition, disruption of cells may expose high affinity receptors not linked to a degradative system, thus making the membrane preparation relatively deficient in degradative activity. There is no direct evidence to suggest that DRG cells have receptor-linked degradative activity, but the indirect evidence is considerable and compelling: the rate of degradation appears to be directly proportional to fractional binding of a2~IflNGF; I~51flNGF is not degraded in the absence of cells or in the presence of supernatant from cells incubated for 60 min at 37 °C; degradation of 1251flNGF cannot be detected in the presence of cells that do

75 not bear high affinity receptors for N G F ; and finally, preliminary evidence suggests that there is substrate specificity for degradation of 1251flNGF (ref. 23, and R. J. Riopelle, unpublished data). The characteristics of the 125IflNGF preparation used in these experiments have been described previously 23. Of particular note is that the preparation was such that cell-independent sedimentation of radioactive label during centrifugation, as well as true non-displaceable binding to cells was a small fraction of total binding, and cellular binding of 125IflNGF at high concentrations of unlabelled /3NGF was found to be approximately what one would expect with only two saturable sites. Hence, non-specific binding is truly non-displaceable binding when cells are used. The 'non-specific' binding to the membrane preparation, definedasbindinginthepresenceofanexcess(3.8 × 10-8 M to 3.8 x 10-7 M) of unlabelled flNGF was found to be in the range of 40-50 ~ of total binding at 1.8 × 10-9 M 125I/3NGF (Fig. 2), and was thus much higher than one would expect if there were but two saturable sites of affinities similar to those found on cells. However, in displacement experiments, it was possible to further compete for 125IflNGF binding with unlabelled flNGF and evidence for a saturable low-affinity binding site for flNGF on membrane was obtained. Thus, true non-displaceable binding in this system occurs at concentrations greater than 10-6 M flNGF. The question of non-linearity in Scatchard analysis plots of equilibrium binding data with N G F has recently been discussed 23. Frazier et al. 7 interpreted their data to suggest that N G F induced changes in receptor affinities, and that, therefore, negative cooperativity was responsible for non-linearity of Scatchard plots in their system. The present data and that of Sutter et al. z2, Riopelle et al. 17, and Sutter et al. z3, have shown conclusively that site 1 does not participate in a negatively cooperative interaction as a result of binding of flNGF, either where cells or membranes of D R G are used. On the other hand, in agreement with Frazier et al. 7, it was observed that the presence of unlabelled N G F resulted in a more rapid dissociation of 125IflNGF. But, as shown previously with intact cells23, and as shown here with membranes, this phenomenon is independent of receptor occupancy by flNGF, and in the present study, evidence for the existence of an unstirred layer phenomenon is provided. Both dissociation experiments at different preincubation concentrations of 125IflNGF, and experiments measuring rates of dissociation of 125IflNGF at different receptor occupancies controlled by the addition of various amounts of unlabelled flNGF indicated that only the existence of a diffusion barrier could explain the observations. Experiments were designed to test the hypothesis that a diffusion barrier around cells or membranes could influence the dissociation of 125IflNGF from the tissue. In these experiments, it was shown that agitation significantly increased the rate of dissociation of 125IflNGF only when the dissociation was induced by dilution without flNGF. These data suggest that hydrophobic, steric, or polar interactions ofligand with membrane components other than specific receptors can create a diffusion barrier in the microenvironment of the membrane surface, and that physical forces can disrupt this barrier. The effect of a barrier would be that ~25IflNGF, dissociating from its receptor, is not readily released from the microenvironment and can rebind to receptor preferentially irrespective of the dilution factor used to initiate dissociation and the achievement of a new steady state of binding. However,

76 when unlabelled N G F is present, the specific radioactivity of the lz5IflNGF released into the 'unstirred layer' becomes diluted, and thus, the measured rebinding is less. As a result, the rate of dissociation of 125IflNGF appears to be more rapid in the presence of unlabelled N G F , than in its absence. The present data thus support and extend the earlier observations with cells 23. These studies confirm the heterogeneous nature of high affinity binding of flNGF to D R G tissue. The finding with membranes is one argument to suggest that intracellular ligand compartmentalization does not explain the site heterogeneity observed with intact cells. One cannot conclude, however, that sites 1 and 2 are distinct molecules; the possibility exists that sites 1 and 2 are the same receptor site in two different conformational states that are not influenced by cooperative interactions. The N G F receptors might undergo ligand-induced conformational change, and this could be an explanation for the observations illustrated in Fig. 5, where the relative occupancy of site 1, as determined from dissociation studies, increases with time after steady-state has been reached, a finding also present with intact cells 23. Alternatively, the two sites on the cell surface may be distinct molecules subserving different biological functions; whereas site 1 is likely responsible for (regenerative) neurite growth 23, site 2 may be required for neurotransmitter synthetic enzyme induction, or the expression of other markers of the differentiated state of the neuron. The third binding site, not seen on intact cells, may represent one component of a high capacity membrane-linked system for retrograde or centripetal transport of N G F once the ligand has been internalized 20. ACKNOWLEDGEMENTS The support of the Medical Research Council of Canada (R. J. R.) and Deutsche Forschungsgemeinschaft (A. S.) is acknowledged. The authors also acknowledge the keen interest and helpful discussion of E. M. Shooter in whose laboratory some of these experiments were carried out.

REFERENCES 1 Banerjee, S. P., Snyder, S. H., Cuatrecasus, P. and Greene, L. A., Binding of Nerve Growth Factor receptor in sympathetic ganglia, Proc. nat. Acad. Sci. (Wash.), 70 (1973) 2519-2528. 2 Brunette, D. M. and Till, J. E., A rapid method for the isolation of L-cell surface membranes using an aqueous two-phase polymer system, J. memb. Biol., 5 (1971) 215-224. 3 Cohen, S., Purification of a nerve-growth promoting protein from the mouse salivary gland and its neurocytotoxic antiserum, Proc. nat. Acad. Sci. (Wash.), 46 (1960) 302-311. 4 DeMeyts, P., In M. Blecher (Ed.), Methods in Receptor Research, Part I, Marcel Dekker, New York, Basel, 1976, pp. 301-373. 5 DeMeyts, P., Bianca, A. R. and Roth, J., Site-site interactions among insulin receptors, J. bioL Chem., 251 (1976) 1877-1888. 6 DeMeyts, P., Roth, J., Neville, D. M., Gavin, J. R. and Lesniak, M. A., Insulin interactions with its receptors. Experimental evidence for negative cooperativity, Biochem. Biophys. res. Commun., 55 (1973) 154-161. 7 Frazier, W. A., Boyd, L. F. and Bradshaw, R. A., Properties of the specific binding of 1251-Nerve Growth Factor to responsive peripheral neurons, J. biol. Chem., 249 (1974) 5513-5519. 8 Frazier, W. A., Boyd, L. F., Pulliam, M. W., Szutowicz, A. and Bradshaw, R. A., Properties and

77

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

specificity of binding sites for i~SI-Nerve Growth Factor in embryonic heart and brain, J. biol. Chem., 249 (1974) 5918-5923. Herrup, K. and Shooter, E. M., Properties of the flNerve Growth Factor receptor in avian dorsal root ganglia, Proc. nat. Acad. Sci. (Wash.), 70 (1973) 3884-3888. Hess, G. P., Bulger, J. E., Fu, J. L., Hindy, E. F. and Silberstein, R. J., Allosteric interactions of the membrane bound acetylcholine receptor: kinetic studies with alpha-bungarotoxin, Biochem. Biophys. Res. Commun., 64 (1975) 1018-1027. Kohn, L. and Winand, R., In H. R. Keback, G. Radda and R. Schwyzer (Eds.), Molecular Aspects of Membrane Phenomena, Springer Verlag, Heidelberg, 1976, pp. 149-181. Levi-Montalcini, R. and Angeletti, P. U., Nerve Growth Factor, PhysioL Rev., 48 (1968) 534-569. Limbird, L. W. and Lefkowitz, R. J., Negative cooperativity among fl-adrenergic receptors in frog erythrocyte membranes, J. biol. Chem., 251 (1976) 5007-5014. Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J., Protein measurement with the Folin phenol reagent, J. biol. Chem., 193 (1951) 265-275. Mobley, W. C., Server, A. C., Ishii, D. N., Riopelle, R. J. and Shooter, E. M., Nerve Growth Factors, New Engl. J. Med., 297 (1977) 1096-1104; 1149-1158; 1211-1218. Moore, W. J., Physical Chemistry, Prentice Hall, Englewood Cliffs, N.J., 1962. Riopelle, R. J., Harris-Warrick, R. M., Sutter, A. and Shooter, E. M., Absence of negative cooperativity of binding of Nerve Growth Factor to embryonic chick sensory ganglion cells, Neurosci. Abstr., 3 (1977) 460. Scatchard, G., The attraction of proteins for small molecules and ions, Ann. N.Y. Acad. Sci., 51 (1949) 660-675. Silhavy, T. J., Szmelcman, S., Boos, W. and Schwartz, M., On the significance of the retention of ligand by protein, Proc. nat. Acad. Sci. (Wash.), 72 (1975) 2120-2124. St6ckel, K., Schwab, M. and Thoenen, H., Comparison between the retrograde axonal transport of NGF and tetanus toxin in motor, sensory, and adrenergic nerves, Brain Research, 99 (1975) 1-16. Stoker, M. G., Role of diffusion boundary layer - - contact inhibition of growth, Nature (Lond.), 246 (1973) 200-203. Sutter, A., Riopelle, R. J., Harris-Warrick, R. M. and Shooter, E. M., Characterization of two distinct classes of high affinity binding sites for Nerve Growth Factor on sensory ganglia ceils from chick embryo, Neurosci. Abstr., 3 (1977) 461. Sutter, A., Riopelle, R. J., Harris-Warrick, R. M. and Shooter, E. M., Nerve Growth Factor receptors, J. Biol. Chem., 254 (1979) 5972-5982.