Journal of Biochemical and Biophysical Methods, 20 (1990) 143-156

143

Elsevier JBBM 00788

Use of analytical gel chromatography to analyze tertiary and quaternary structural changes in E. coli aspartate transcarbamylase Sarina Bromberg, David S. Burz and Norma M. Allewell Department of Molecular Biology and Biochemistry, Wesleyan University, Middletown, CT 06457, U.S.A. (Received 25 April 1989) (Accepted 17 October 1989)

Summary E. coli aspartate transcarbamylase (ATCase) is a large (310 kDa~ protein that undergoes major changes in quaternary structure when substrates and regulatory nucleotides bind. We have used analytical gel chromatography to detect quaternary structure changes in both the holoenzyme and its catalytic subunit (c3), to characterize the quaternary structure of single site mutant proteins and to monitor urea-induced dissociation and unfolding of c 3. Binding of the bisubstrate analog PALA (N-(phosponacetyi)-L-aspartate) to ATCase and c3 has been shown to alter s20,w by - 3 . 3 ~ and + 1.4~, respectively [Howlett, G.J. and Schachman, H.K. (1977), Biochemistry 23, 5077-5083]. The corresponding changes in the chromatographic partition coefficient (o) are - 2.6 + 0.3~ and 5.5 + i.9r~ on Sephacryl S400HR and $200, respectively. Partition coefficients of mutant ATCases with single site mutations in the c chain differ from those of the wild-type protein by + 0.Sff, in small zone experiments; for example, mutations Arg 269-~ Gly and Glu 239 ~ Gin alter the partition coefficient by 0.4'~ and -0.5~, respectively. The partition coefficient of mutant Glu 50 --* Gin is identical to the wild type enzyme. In the presence of saturating PALA, partition coefficients of Glu 50 ~ Gin and Arg 269 --. Giy, but not Glu 239--* Gin are identical to those of the wild type. Results for Glu 239 ~ Gin are consistent with measurements of activity, small angle X-ray scattering and sedimentation coefficient that indicate that mutations at this site shift the quaternary structure towards the R state [Ladjimi and Kantrowitz (1988), Biochemistry 27, 276-83; Vachette and Herr6, cited by Kantrowitz and Lipscomb (1988), Science 241, 669-674; Newell and Schachman (1988), FASEB J. 2, A551]. Results for Glu 50---* Gin are also consistent with measurements of activity (Ladjimi et al. (1988), Biochemistry 27, 268-276). The changes in tertiary and quaternary structure that result from urea-induced denaturation of c3 result in larger changes in the partition coefficient. Dissociation into folded monomers in 1-1.75 M urea is accompanied by a 4.65 increase in partition coefficient, while denaturation at > 5 M urea gives rise to a 43ff, decrease

Correspondence address: N.M. Allewell, Department of Molecular Biology and Biochemistry, Wesleyan University, Middletown, CT 06457, U.S.A.

Abbreviations: AGC, analytical gel chromatography; HPLC, high performance liquid chromatography; ATCase, E. coli aspartate transcarbamylase; c 3, catalytic subunit; r 2, regulatory subunit; CP, carbamoyl phosphate; PALA, N-(phosphonacetyl)-L-aspartate; OPA, o-phthaldialdehyde. 0165-022X/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

144 on S-300 Sephacryl. The bisubstrate analog PALA suppresses dissociation and increases the cooperativity of the unfolding reaction. Key words: Gel chromatography; Quaternary structure change; Subunit dissociation; Denaturation; Ligand binding; Single site mutant

Introduction

Because conformational changes in proteins are critical to their function, there is a need for simple and direct methods to detect and analyze them in solution. Although hydrodynamic and spectroscopic methods are powerful and widely used, ~'~e instrumentation is sophisticated and expensive. In addition, spectroscopic methods generally detect only local structural changes. In contrast, analytical gel chromatography (AGC) is as sensitive as analytical ultracentrifugation to changes in molecular size and shape, but can be carried out at very modest cost with equipment that is easy to maintain ~,~d operate. The introduction of HPLC methods has dramatically increased the rate at which data can be obtained, although at some sacrifice of precision. AGC has been widely used to study associating systems [for reviews see ref. 1-4] and HPLC-AGC is increasingly being used in protein folding studies [5-9]. The potential of the method for detecting and analyzing conformationai changes of native proteins, although recognized [10,11], has, however, not yet been exploited. We show here that AGC can be used to detect changes in the tertiary and quaternary structure of E. coli aspartate transcarbamylase (ATCase) that result from iigand binding, single site mutations and changes in solution conditions. [For recent reviews of ATCase, see 12-13.] We also show that this method makes the equilibria 3c *-*c2 + c ~ c3 experimentally accessible at low urea concentrations. ATCase is a large (310,000 Da) protein consisting of six catalytic (c) chains and six regulatory (r) chains, organized as two catalytic trimers (c3) and three regulatory dimers (r 2) [14-16]. The molecular weights of the c and r chains are 33 and 17 kDa, respectively [15]. ATCase can be dissociated into c~ and r 2 subunits; the isolated subunits retain the binding properties of the native protein [14]. The enzyme catalyzes the first committed step in pyrimidine biosynthesis, transfer of a carbamoyl group from carbamoyl phosphate (CP) to L-Asp. Activity is ailosterically regulated by CTP (an inhibitor) and ATP (an activator) [17]. Binding of CP, CTP and ATP produces small but detectable changes in tertiary and quaternary structure [18-22], while binding of analogs of L-Asp (in the presence of CP) or the bisubstrate analog N-(phosphonacetyl)-L-Asp (PALA) produces a major change in quaternary structure [19-24]. Several ligands induce changes in the state of association of either the native enzyme or its subunits. CTP increases the self-association constant of r chains [25], ATP increases the rate of dissociation of c 3 [26], while PALA decreases the rate of dissociation of c3 [27] and increases its rate of association [28]. Binding of CP + succinate or PALA to ATCase decreases the sedimentation coefficient by -3.3% and -3.5%, respectively [19]; binding of PALA

145

has also been shown to increase the radius of gyration by 2.5 _+ 1.5 ,A [24]. In contrast, binding of CP + succinate and PALA to c3 increases the sedimentation coefficient by + 1.05~ and 1.4%, respectively [21]. Although these conformational changes were first detected in solution, those associated with binding of CP + succinate, PALA and CTP to ATCase have been confirmed and delineated by X-ray crystallography [29-31]. Both CP + succinate and PALA cause a 12 A (12~) expansion along the three-fold axis of the molecule [29,31], while CTP has no detectable effect on the overall shape of the crystalline enzyme [30]. The crystal structures of the protein with CTP and PALA bound are designated the T and R states, respectively.

,,

Methods

Reagents PALA was obtained from the Drug Synthesis and Chemistry Branch, Division of Cancer Treatment, National Cancer Institute, Bethesda, MD. Purity was determined by ultraviolet difference spectroscopy [32] and verified by 3tp NMR, using an inorganic p h o s p h a t e s t a n d a r d . N e o h y d r i n 1-[3-(chloromercuri-2methoxypropyl)urea] was purchased from I.C.N.-K. and K. Laboratories and twice recrystallized from 95~ ethanol. Ultrapure urea was obtained from Schwarz/Mann Biotech Division of I.C.N. Biomedicals, Inc. Rabbit muscle aldolase, bovine liver catalase, horse spleen ferritin and bovine thyroid thyroglobulin for column calibration were obtained from Pharmacia Fine Chemicals (gel filtration calibration kit). All other calibration proteins were obtained from Sigma Chemical Co. c~ was prepared either from the derepressed, diploid E. coli strain developed by Gerhart and Holoubek [33] as previously described [34-35] or from the transformed overproducing strain EKl104/pEK17 developed by Nowlan and Kantrowitz [36] and kindly provided by Dr. E.R. Kantrowitz [Boston College, Chestnut Hill, MA] by a method that will be described elsewhere. There were no detectable differences in the proteins prepared by the two methods. Wild type and mutant c6r6 were purified from the E. cog strain EKl104 transformed with the appropriate plasmid. Cells were disrupted in 0.1 M Tris-HCI, pH 9.2 by sonication. Protein was precipitated by addition of solid (NH4)2SO4 to 70~ saturation; dissolved in 0.04 M K21-1PO4-KH2PO4, 0.2 mM dithiothreitol, 0.2 mM EDTA, pH 7; treated with RNase A and DNase I; precipitated by dialysis against 0.01 M KH2PO4-K2HPO4, pH 5.9 in the presence of 0.2 mM dithiothreitol, 0.2 mM EDTA; redissolved and passed over a 2.8 x 30 cm. DEAE-Sephadex A-50 column equilibrated with 0.01 M imidazole-HCl, 0.35 M KCI, 0.2 mM EDTA, 0.2 mM dithiothreitol, pH 7. Purity was assessed by nondenaturing PAGE [37] and by pH activity assays at pH 8.3 and was comparable to previous preparations. Protein was stored as a precipitate in 3.6 M (NH4)2SO4, 0.1 M Tris-HCI, 0.2 mM EDTA, 0.2 mM dithiothreitol, pH 8.3 at 4°C and equilibrated with the experimental buffer

146

by passage over a 0.5 x 8 cm Bio-Gel P-4, 100-200 mesh or Sephadex G-50 desalting column. Holoenzyme concentrations were determined spectrophotometrically assuming an extinction coefficient of 0.59 cm- ~ - (mg- ml- ~)- ~ [38]. Concentrations of c 3 were determined either spectrophotometrically assuming an extinction coefficient of 0.72 cm -1 -(rag-ml-1) -1 [38] or, in the presence of high concentrations of flmercaptoethanol, with the Bio-Rad protein assay with bovine serum albumin (Sigma Chemical Co.) as the standard.

Chromatography Flow rate was controlled by a Rainin Rabbit peristaltic pump acting as a brake on the eluant from the column; temperature by the flow of water from a circulating water bath through a water jacket [4]. The eluant was monitored by a fluorometer (Gilson Spectroglo, fitted with OPA filters), ultraviolet spectrophotometer (Varian Techtron 635) or column monitor (LKB Model 2138 Uvicord S Filter Detector or Gilson Holochrome) with a strip-chart recorder attached [4]. Flow rates were 6-11 ml. h - l ; temperatures were 7.5 °C in experiments with c3 and either 5 °C or 25 ° C in experiments with ATCase. Flow rates for small zone experiments were determined at least daily by collecting eluant from the column in a graduated cylinder for 30-60 min and determining the volume. In large zone experiments, removal of the reservoir during loading of the sample was found to change the flow rate by -±0.05 ml. h-!. Accordingly, flow rates were determined in each large zone experiment while the sample was entering the column and while it was eluting. Typically, eluant was collected for 2 h during the 2-3 h load and for 3-7 h during elution. Water-jacketed columns (1.5 x 50 cm) were purchased from Bio-Rad. Resins were Sephacryl $200-400 and S-400HR (Pharmacia L.K.B. Biotechnology Inc.), equilibrated with the appropriate buffer, degassed and brought to the experimental temperature before the column was poured. Sephacryl was chosen because of its stability, particularly in urea, and its relative linear reponse to the Stokes radius [39]. S.400 and S-400HR were used in experiments with ATCase: S-200 and S-300 in experiments with c 3. Columns were packed either at their natural flow rate or, when extreme precision was required, at a flow rate of 70-240 m l - h - t maintained with a peristaltic pump, in order to produce an extremely stable gel bed. Columns for small zone experiments were fitted with a Biorad flow adapter attached to a teflon three-way or Rheodyne injector valve so that samples could be loaded without removing the column head. Loading sample by injection increases the reproducibility of measurements. For the urea denaturation experiments, the buffer with which the resin was equilibrated for pouring the column was made 5 M in urea and the bed volume was fixed with the flow adapter to prevent changes in bed volume as the urea concentration was changed. All column buffers were filtered through Nylon-66 0.2 ~tm filters (Rainin Instrument Co.) and degassed at room temperature before use. When the composition of the column buffer was changed, the column was equilibrated with the new buffer for at least three column volumes or until a stable baseline was achieved. Sample volumes were 0.5 ml for small zone experiments and 18 ml for large zone

147

experiments. Glycine (10/tM) and blue dextran (average molecular weight: 2 x 106 Da; 1 rag- ml-~), both purchased from Sigma Chemical Co, were used for column calibration in the experiments with c3. Glycine was detected by fluorescence, after conjugation with o-phthaldialdehyde (OPA) (see below): blue dextran was detected by visible absorbance at 620 nan. The S-400 and S-400HR columns used in the experiments with ATCase were calibrated either with dextran (average molecular weight 5-40 × 106 Da; Sigma Chemical Co.: 0.6 mg. ml - j ) or tobacco mosaic s~rus (Rs: 60 rim; 25/tg- ml -a) and tyrosine (0.17 mM). All markers were detected at 280 nm. Tyrosine was used simply to eliminate the need for fluorescence detection. Control experiments indicated that markers did not alter the elution volume of the protein. When the buffer contained high concentrations of/3-mercaptoethanol ultraviolet absorbance could not be used for detection. Instead, both protein and glycine (the internal volume marker) were post-column derivatized with OPA by pumping a 0.025~ solution of OPA in 0.2 M K+-borate, pH 10.5 through a T-joint into the column eluant at the column flow rate [40]. Derivatization with OPA allows picomolar concentrations of protein to be detected [40], although the concentrations used in these experiments were considerably higher. o, the fractional included volume or partition coefficient, was calculated with the equation: o = (v,- Vo)l(V

- Vo)

where V~ is the elution volume of the protein, Vo is the void volume, determined with blue dextran, dextran or tobacco mosaic virus, and V~ is the internal volume of the gel, determined with glycine or tyrosine. Elution volumes were calculated from the maximum of the peak in small zone experiments and from the centroid of the leading edge in large zone experiments [1].

Results

The T ~ R quaternary structure change

Binding of PALA is known to produce a major change in the quaternary structure in ATCase [19.22-24.29]. Table 1 shows the values of the partition coefficient obtained with wild type ATCase in small zone experiments in the presence and absence of PALA in two buffers (0.1 M K+-Hepes and 0.04 M NaH2PO4-Na2HPO4) at two pH values (pH 7 and 8.3) and two temperatures (5°C and 25 o C). Under all conditions, the partition coefficient was consistently 1.1-2.0~ smaller in the presence of PALA than in its absence. Calibration of S-400HR columns with standard proteins in 0.04 M KH2PO 4K2HPOo 0.2 mM EDTA, 0.2 mM dithiothreitol, pH 7.0 at 5 °C indicates that the values of the partition coefficient obtained in large zone experiments under these solution conditions in the presence and absence of PALA correspond to Stokes radii of 60.1 A, and 58.§ A respectively (Fig. l y. These values must be interpreted

148 TABLE 1 FRACTIONAL INCLUSION VOLUMES (o) OF WILD TYPE ATCase AND c 3 UNDER VARIOUS CONDITIONS AND PERCENTAGE CHANGE IN o IN THE PRESENCE OF SATURATING PALA Buffer, pH

A TCase a

Hepes 7 b Hepes 8.3 b Phosphate 7 c Phosphate 8.3 ~

5oC

25 o C

o

A o / o (%) d

O

AO/O(~ ) d

0.54304-0.0014 0.5665+0.0012 ~.5366+ 0.0016 (, ~2994-0.0005

--1.58+0.55 -- 1.55-+0.36 -1.40-+0.43 - ~.72 ± 0 ~2

0.5446--+0.0004 0.5677-+0.0008 0.5498-+0.0012 0.54484-0.0017

--1.98-+0.24 -- 1.114-0.26 -1.20+0.27 - 1.174-0.62

7.50C C3 Hepes 8.3 • Phosphate 7 c,f

0.4682 4-0.0016 0.2192 4- 0.0014

2.48 4- 0.80 5.5 + 1.9

a S.400 resin; b 0.1 M K+-Hepes, 0.2 mM dithiothreitol, 0.2 mM EDTA; c 0.04 M NaH2PO4-Na2HPO4, 0.2 mm dithiothreitol, 0.2 mM EDTA; d Ao/o is calculated relative to the unliganded protein under the same conditions; all errors are RMS; e 0.1 M K+-Hepes, 0.2 M /3-mercaptoethanol, 0.2 mM EDTA; S.300 resin; f S-200 resin.

cautiously, because of the asymmetry of the molecule [39,41]. They are, however, in good agreement with the value (60 A) previously determined by analytical gel chromatography [39] and in the same range as the dimensions of the crystalline enzyme (65 A along the two-fold axes; 50 A and 56 A along the three-fold axis in the absence and presence of PALA) [42], but larger than the radius of gyration derived by low-angle X-ray scattering under different solution conditions [24]. The difference between them is comparable to that derived by small-angle X-ray scattering [24], but smaller, as expected, than the 12 A change in length of the molecule along the three-fold axis that results, from the T ~ R transition in the crystal [42]. The dependence of the partition coefficient in l~rge zone experiments in 0.04 M K 2HPO4-KH2PO4 buffer, pH 7.0 on the concentration of PALA is shown in Fig. 2. The partition coefficient decreases with the concentration of PALA up to a molar ratio of PALA" ATCase of approximately 4" 1. This is as expected, since the association constant of PALA is ~pproximately 107 M -~ [43], and, although there are six active sites, the quaternary structure change is complete at a molar ratio of 4 : 1 [19,44]. The maximum value of Ao/o(%) is 2.6%, a value significantly larger than that obtained in small zone experiments. Quaternary structure changes in c~

Partition coefficients of c 3 derived from small zone experiments in the presence and absence of saturating PALA are listed in Table 1. In contrast to the holoenzyme, but as expected from sedimentation studies [19], PALA increases the partition

149

(a)

0.60

_

• catalase

"~~anthine

O"

oxidase

pyruvate kinase

0.50

ferritin ATCase + PALA urease- - ~

0.40

, 1.74

1.70

, 1.78

1.82

log (Rs)

0.61 ' ~ 0.7

~f ~_

RNase

b

¢ytochrome¢

( )

0.5 O"

0.4,

ovalbumin,

0.3.

alkaline p h o s p h a t a s e ~

~

-¢3 + P A L A /c3

serum albumin " transferdn • " ~ . , g-3-p dehydrogenase~"~aldolase alcohol dehydrogenase~ " ~ . . . . . . . ,catalase , .

0.21 0.1

1.3

1.4

1.5 Iog(Rs)

1.6

1.7

1.8

Fig. 1. (a) Calibration curve for S-400HR column (1.5 x 23.5 cm) obtained in small zone experiments at 5oC in 0.04 M KHaPO4-KaHPO4, 0.2 mM dithiothreitol, 0.2 mM EDTA, pH 7; (b) calibration curve for S-200 column (1.5 x 32 cm) obtained in small zone experiments at 7.5°C in 0.04 M NaH2PO 4Na2HPO4, 0.2 mM EDTA, pH 7.0. 3.0 ll

2,5 I

8O

2,0

1,0 O"

E

0

0,5

20

0.0

•

0

2

.

4 6 PALA/ATCase

.

,

8

,

!0

10

Fig. 2. Titration of ATCase with PALA at 5 ° C in 0.04 M KH2 PO4-K 2 HPO4, 0.2 mM dithiothreitol, 0.2 mM EDTA, pH 7.0.; (O) percentage change in partition coefficient in large zone experiments on Sephadex S400HR; (B) percentage change in OD290.

150 0.5700

0.5600 0 0.5500

S

0.5400

'

0

I

10

'

I

20

Temperature (° C)

'

30

Fig, 3. Partition coefficients obtained in small zone experiments over the temperature r~nge 5-25 ° C on S-400 with wild type enzyme in the absence (®) and presence (El) of saturating PALA, Glu 239-* Gin (o), and Arg 269 -* Gly (z~).All experimentswere carried out in 0.1 M K+-Hepes,0.2 mM dithiothreitoi, 0.2 mM EDTA, pH 8.3.

coefficient of c 3 under both sets of buffer conditions, by 2.5~ in Hepes on S-300 and 5.5~ in phosphate on S-200. Partition coefficients obtained in phosphate buffer in the presence and absence of PALA correspond to Stokes radii of 38.1 A and 39.1 ,~, respectively. Single site mutants Several single site mutants are believed to be in either the T or R state on the basis of substrate affinity, Fm~~ and Hill coefficient [cf. 12, 45-46]. Fig. 3 shows the temperature dependence of the partition coefficients of two of these mutants, Glu 239 --, Gin and Arg 269 --~ Gly. Glu 239 --~ Oln is located at the interface between c chains in opposite % subunits while Arg 269 --, Gly lies between c chains within a c3 subunit [29,30,42]. Both activity measurements [45] and low angle X-ray scattering (P. Vachette and O. Herv6, cited in [12]) indicate that Olu 239--, OIn is shifted towards the R state. Sedimentation measurements show that this is also the case for Olu 239 ~ Lys [47]. The partition coefficient of mutant Olu 239 ~ Gin falls between those obtained for the wild type enzyme in the presence and absence of PALA over the entire temperature range examined, while those for mutant Arg 269 ~ Oiy are consistently larger. The Stokes radii of wild type, Arg 269 --, Oly and Olu 239 ~ Oln at 5 ° C derived by calibration in Hepes buffer, pH 8.3 are 58.3 ~,, 57.9 ,~ and 59.6 ,~ respectively. The effects of PALA on partition coefficients derived in small zo~'Leexperiments of these mutants and a third mutant, Olu 50 --, O:n, are given in Table 2. ~'~lu 50 lies between domair ; of the c chain; Glu 50 ~ Gin is a low-activity-low affinity mutant that is, nevertheless, able to undergo the T--, R transition [48]. The partition coefficient of Ofu 50 ~ Gin is identical to the wild type PALA restores the

151 TABLE 2 " ~ , ~. . .•. :~.-NAL ~ I C L U S I O N VOLUMES ( o ) O F W I L D TYPE A N D M U T A N T H O L O E N Z Y M E S IN ABSENCE A N D PRESENCE OF S A T U R A T I N G PALA

Protein

Sigma - PALA

Wild type Giu 50 ~ Gin Glu 239 ~ Gin Arg 269 ~ Giy

a

+ PALA b

0.5532 _+0.0006 0.5538 + 0.0001 : 0.5505 + 0.0014 0.5562 ± 0.0003

0.5446 + 0.0014 0.5452_+ 0.0005 0.5504 + 0.0001 0.5441 + 0.0006

a All experiments were performed in 0.1 M K+-Hepes, 0.2 mM dithiothreitol, 0.2 mM EDTA, pH 8.3 at 5oC.; b 2 5 / t M PALA.

partition coefficients of mutants Glu 50 ~ Gin and Arg 269 --, Gly to that obtained with the wild type enzyme in the presence of PALA; however the partition coefficient of Glu 2 3 9 - , Gin is insensitive to PALA at concentrations that are sufficient to saturate the wild type enzyme.

Urea-induced dissociation and unfolding of c3 The dependence of the partition coefficient upon urea concentration in small zone experiments carried out in the presence and absence of PALA is shown in Fig. 4. In the absence of PALA, the partition coefficient remains nearly constant at urea concentrations of 0-1 M, increases in the range 1-1.75 M, then decreases dramatically over the range 2-5.5 M. In the presence of PALA, the partition coefficient decreases gradually up to 5.5 M urea, then drops abruptly to the same value obtained in the absence of PALA. Column calibration yields a Stokes radius of

_

0..~0

. m

O.4P

0.,00

0,3.¢

O30

O,ZS

I

IL'~EAI t~l)

Fig. 4. Dependence of partition coefficient of c3 on urea concentration in small zone experiments carried out on S-300 at 7.5°C in 0.1 M K+-Hepes, 0.2 M fl-r~rcaptoethanol. 0.2 mM EDTA, pH 8.3. Samples were incubated on ice for at least 2 h before loading on the column. ( a A). control: (B . . . . . . i ) , saturating PALA; (B), the positions of two peaks that were resolved in the presence of PALA at high concentrations of urea. Curves are the best fit to the equation: O,,bs = (0 i + KIo2 + KtK2o.~)/(I + KtK2) where K i -- K ° e x p [ - A/RT(urea)].

152

~"~,....,. ¢:: (9 t--

\

(9 tJ (9 U'J

U.

k

Elution Volume Fig. 5. (a-d) Effect of varying c 3 concentration in 1.25 M urea with S-300 resin. Initial concentrations of c3: (a) 20 FM; (b) 5 ~M; (c) 2 FM; (d) I FM. (e-g) Effect of varying c3 concentration in 1.25 M urea with S.200 resin. Initial protein concentrations: (e) 12.7/~M; (0 5 FM; (g) 1.3/~M. (h-k) Variation in elution profile with urea concentration on S-200 columns with an initial c 3 concentration of 5/~M. Urea concentrations: (h) 1.25 M; (i) 1.5 M; (j) 1.625 M; (k) 1.75 M. All experiments were carried out in 0.1 M K +-Hepes, 0.2 M ~-mercaptoethanol, 0.2 mM EDTA, pH 8.3 at 7.5 o C.

61 ± 3 A at high urea concentrations, whereas Tanford's empirical equation predicts a value of 60 ± 3 A for a polypeptide of 310 residues in the random coil conformation [49]. Sedimentation velocity experiments have established that c3 does in fact exist as unfolded monomers at high urea concentrations [28]. Further experiments, carried out over a range of protein and urea concentrations on a S-200 column, established that the increase in the partition coefficient at 1-1.75 M urea in the absence of PALA is the result of dissociation of the trimer into folded monomers. Whereas only a single, fairly symmetrical peak was observed with the S-300 columns, the S-200 column partially resolved two peaks (Fig. 5). Calibration with standard proteins indicated that the Stokes radius of the more slowly eluting peak lay between that expected for c and c2. In addition, the relative areas of the two peaks varied with protein and urea concentration in the way expected if c, c 2 and c 3 were in equilibrium, with dissociation being driven by increasing concentrations of urea (Fig. 5). in contrast, the results for PALA shown in Fig. 4 indicate that binding of PALA between chains in c 3 suppresses dissociation so that dissociation of PALA, dissociation of c3 into monomers and unfolding of the chains occur at the same urea concentration. Dissociation and unfolding occur at much higher urea concentrations than in the absence of PALA and are much more cooperative. A more detailed analysis of these data will be published elsewhere.

153

Discussion

Because AGC resins are known to interact with proteit~s '~he question of whether the changes in partition coefficient obser,'ed here reflect simply changes in the interaction of the protein with the resin rather than changes in molecular size and shape must be considered. Several features of the results argue against this possibility. First, both ATCase and c a fall below the molecular weight calibration curve in phosphate buffer suggesting that neither is adsorbed to the column (data not shown). (In contrast, c a falls above the molecular weight calibration curve in Hepes buffer at pH 8.3, suggesting it is bound by the resin under these conditions.) Second, binding of PALA has opposite effects on the partition coefficients of ATCase and c a, and the magnitudes of the changes vary little with pH, buffer, ionic strength, resin (in the case of c 3) and temperature in the case of ATCase. Third, the signs of these changes in the partition coefficient are those expected from sedimentation velocity experiments; binding of PALA to ATCase results in decreases of 1.1-2.6~ in o and 3.5~ in s20,w [19], while binding of PALA to c a is accompanied by increases in both o (2.5~ in Hepes, 5.5~ in phosphate) and s20,w (1.4~) [19]. Fourth, there is no, consistent trend in the relationship between the partition coefficient and ionic strength for ATCase; the partition coefficient increases with ionic strength in Hepes, but decreases slightly with ionic strength in phosphate buffer. Similarly, there is no consistent relationship between the change in charge induced by a mutation and the corresponding change in the partition coefficient in fifteen mutants examined to date. The changes in the partition coefficient observed with the mutants are also consistent with conclusions drawn from other types of experiments. Interpretation of the urea denaturation data is more straightforward. The results obtained with the S-200 column indicate clearly that the increase in elution volume from the S-300 column in 1-1.75 M urea in the absence of PALA is the result of protein dissociation. Since PALA binds very tightly between chains in c 3, it is not surprising that this ligand suppresses dissociation. Our results at high urea concentrations confirm the conclusion drawn from sedimentation velocity experiments, that ca exists as unfolded monomers under these conditions [28]. AGC has several advantages over more traditional methods in folding studies of multimeric proteins. In contrast to both spectroscopic and calorimetric methods, the experimental parameter, o, provides quantitative information about the size or shape of molecular species. Dissociation is readily distinguished from unfolding, since the effects on the partition coefficient are opposite. Post-column derivatization with OPA allows picomolar concentrations of protein to be detected even in the presence of strongly ultraviolet-absorbing materials. Extrapolation to 0 M allows free energies in the absence of denaturant to be evaluated. In summary, the results presented here demonstrate that AGC is a powerful method for studying quaternary structL~re change, dissociation and unfolding in subunit proteins. They suggest that detectable changes in the quaternary structure of ATCase can be induced by changes in temperature, pH, and buffer, as well as effector binding. The fact that Ao/o for all of the mutants examined Io date is less

154

than aa/a for PALA binding suggests that none of these mutants is completely frozen in the R form, but, instead, exists in intermediate or different conformations. While small zone experiments allow the effects of mutations, ligands and environmental changes to be screened quickly, to realize fully the potential of the method, large-zone experiments are necessary. With the concentration of protein controlled, it will be possible to derive equilibrium constants for both mutant and wild type protein for the dissociation of c3 and possibly r2 processes which have previously been relatively inaccessible. Analyses of boundary shapes will also provide information on the states of association of mutant proteins and perhaps the existence of intermediates in the T ~ R transition. Although HPLC has the potential to allow experiments to be performed more quickly and preliminary experiments indicate that the T ~ R transition is detectable, HPLC results to date have considerably less precision than those obtained by the low pressure method. Even with low pressure methods, constant monitoring of flow rate is critical to success.

Acknowledgements We thank Dr. Evan Kantrowitz for supplying plasmids and mutant proteins and Himanshu Oberoi for help with Table 1.

References 1 Ackers, G.K. (1970) Analytical gel chromatography of proteins. Adv. Pro. Chem. 23, 343-446. 2 Cann, J. (1970) Interacting Macromolecules. Academic Press, New York, NY. 3 Ackers, O.K. (1975) Molecular Sieve Methods of Analysis. In: Neurath, H. and Hill, R. (Eds.), Proteins, 3rd. Edn., Vol. 1, Academic Press, New York, NY., pp. 1-94. 4 Valdes, R., Jr. and Ackers, G,K. (1979) Study of Protein Subunit Association Equilibria by Elution Gel Chromatography. In: Hirs, C.H.W. and Timasheff, S.N. (Eds.), Methods in Enzymology, Vol. 61, Academic Press, New York, NY., pp. 125-142. 5 Corbett, R.J.T. and Roche, R.S. (1984) Use of high-speed size.exclusion chromatography for the study of protein folding and stability. Biochemistry 23, 1888-1894. 6 Wison, J., Kelley, R.F., Shalongo, W., Lowery, D. and Stellwagen, E. (1986) Equilibrium and kinetic measurement of the cohformational transition of thioredoxin in urea. Biochemistry 25, 7560-7566. 7 Shalongo, W., Ledger, R., Jagannadham, M.V. and Stellwagen, E. (1987) Refolding of denatured thioredoxin observed by size-exclusion chromatography. Biochemistry 26, 3135-3141. 8 Kelley, R.F., Shalongo, W., Jagannadham, M.V. and Stellwagen, E. (1987) Equilibrium and kinetic measurements of the conformation transitions of reduced thioredoain. Biochemistry 26, 1406-1411. 9 AI-Obeidi, A.M. and Light, A. (1988) Size exclusion HPLC of native trypsinogen, the denatured protein and partially refolded molecules. J. Biol. Chem. 263, 8642-8645. 10 Warshaw, H.S. and Ackers, G.K. (1971) Molecular sieve studies of interacting protein systems: VIII critical evaluation of the equilibrium sedimentation technique using stacked gel columns. Anal. Biochem. 42, 405-421. I1 Halvorson, H.R. and Ackers, G.K. (1974) Molecular sieve studies of interacting protein systems. J. Biol. Chem. 249, 967-973. 12 Kantrowitz, E.R. and Lipscomb, W.N. (1988) Escherichia coil aspartate transcarbamylase: the relationship between structure and function. Science 241,669-674. 13 Allewell, N.M. (1989) Escherichia coil aspartate transcarbamoylase: structure, energetics and catalytic and regulatory mechanisms. Annu. Rev. Biophys. Biophys. Chem. 18, 71-92.

155 14 Gerhart, J.C. and Schachman, H.K. (1965) Distinct subunits for the regulation and catalytic activity of aspartate transcarbamyase. Biochemistry 4, 1054-1062. 15 Weber, K. (1968) New structural model of E. coli aspartate transcarbamylase and the amino acid sequence of the regulatory polypeptide chain. Nature 218, 1119. 16 Wiley, D.C. and Lipscomb, W.N. (1968) Crystallographic determination of the symmetry of aspartate transcarbamylase. Nature 218, 1119-1121. 17 Gerhart, J.C. and Pardee, A.B. (1962) The enzymology of control by feedback inhibition. J. Biol. Chem. 237, 891-896. 18 Collins, K.D. and Stark, G.R. (1969) Aspartate transcarbamylase: studies of the catalytic subunit by ultraviolet difference spectroscopy. J. Biol. Chem. 244, 1869-1877. 19 Howlett, G.J. and Schachman, H.K. (1977) Allosteric regulation of aspartate transcarbamoylase: changes in the sedimentation coefficient promoted by the bisubstrate analog N-(phosphonacetyl)-Laspartate. Biochemistry 23, 5077-5083. 20 Lennick, M. and Allewell, N.M. (1981) Changes in the hydrogen exchange kinetics of E. coli aspartate transcarbamylase produced by effector binding and subunit association. Proc. Natl. Acad. Sci. USA 78, 6759-6763. 21 Kirschner, M.W. and Schachman, H.K. (1971) Conformation changes in proteins as measured by difference sedimentation studies of the effect of stereospecific ligands on the catalytic subunit of aspartate transcarbamylase. Biochemistry 10, 1919-1926. 22 Herv6, G., Moody, M.F., Tauc, P., Vachette, P. and Jones, P.T. (1985) Quaternary structure changes in aspartate transcarbamylase studied by X-ray solution scattering. J. Mol. Biol. 185, 189-199. ~ 23 Dubin, S.B. and Cannell, D.S. (1975) The effect of succinate on the translational diffusion coefficient of aspartate transcarbamylase. Biochemistry 14, 192-195. 24 Moody, M.F., Vachette, P. and Foote, A.M. (1979) Changes" in the X-ray solution scattering of aspartate transcarbamylase following the allosteric transition. J. Moi. Biol. 133, 517-532. 25 Cohlberg, J.A., Pigiet, V.P. and Schachman, H.K. (1972) Structure and arrangement of the regulatory subunits in aspartate transcarbamylase. Biochemistry 11, 3396-3411. 26 Yang, Y.R. and Schachman, H.K. (1987) Hybridization as a technique for studying interchain interactions in the catalytic trimer of aspartate transcarbamoylase. Anal. Biochem. 163, 188-195. 27 Burns, D.L. and Schachman, H.K. (1982) Ligand-promoted strengthening of interchain bonding domains in the catalytic subunits of aspartate transcarbamoylase. J. Biol. Chem. 257, 12214-12218. 28 Burns, D.L. and Schachman, H.K. (1982) Assembly of the catalytic trimers of aspartate transcarbamoylase from unfolded polypeptide chains. J. Biol. Chem. 257, 8638-8647. 29 Krause, K.L., Volz, K.W. and Lipscomb, W.N. (1987) 2.5 ,A structure of aspartate carbamoyltransferase complexed with the bisubstrate analog N.(phosphonacetyl)-L-aspartate. !. Mol. Biol. 193, 527-553. 30 Kim, K.H., Pan, Z., Honzatko, R.B., Ke, H.M. and Lipscomb, W.N. (1987) Structural asymmetry in the CTP-liganded form of aspartate carbamoyltransferase from E. coll. J. Mol. Biol. 196, 853-875. 31 Gouaux, J.E., Krause, K.L. and Lipscomb, W.N. (1987) The catalytic mechanism of Escherichia Coli aspartate carbamoyltransferase: a molecular modelling study. Biochem. Biophys. Res. Commun. 142, 893-897. 32 Collins, K.D. and Stark, G.R. (1971) Aspartate transcarbamylase: interaction with the transition state analogue N.(phosphonacetyl)-L-aspartate. J. Biol. Chem. 246, 6599-6605. 33 Gerhart, J.C. and Holoubek, H. (1967) The t~trification of aspartate transcarbamylase of E. co/i and separation of its protein subunits. J. Biol. Chem. 242, 2886-2892. 34 Yang, Y.R., Kirschner, M.W. and Schachman, H.K. (1978) Aspartate transcarbamoylase (Escherichia coli): preparation of subunits. Methods Enzymol. 51, 35-41. 35 Burz, D.S. and Allewell, N.M. (1982) Interactions of ionizable groups in E. coli aspartate transcarbamylase with adenosine and cytidine 5'-triphosphate. Biochemistry 21, 6647-6655. 36 Nowlan, S.F. and Kantrowitz, E.R. (1985) Superproduction and rapid purification of Escherichia coli aspartate transcarbamylase and its catalytic subunit under extreme derepression of the pyrimidine pathway. J. Biol. Chem. 260, 14712-14716. 37 Davis, B.J. (1964) Disc electrophoresis-ll: method and application to human serum proteins. Ann. N.Y. Acad. Sci. 121,404-427.

156 38 Blackburn, M.N. and Schachman, H.K. (1976) Alteration of the allosteric properties of aspartate transcarbamoylase by pyridoxylation of the catalytic and regulatory subunits. Biochemistry 15, 1316-1323. 39 LeMaire, M., Rivas, E. and Moiler, J.V. (1980) Use of gel chromatography for determination of size and molecular weight of proteins: further cautions. Anal. Biochem. 106, 12-21. 40 Benson, J.R. and Hare, P.E. (1975) o-Phthaldehyde: fluorescence detection of primary amines in the picomole range; comparision with fluorescamine and ninhydrin. Proc. Natl. Acad. Sci. USA 72, 619-622. 41 Nozaki, Y., Schechter, N.M., Reynolds, J.A. and Tanford, C. (1976) Use of gel chromatography for the determination of the stokes radii of proteins. Biochemistry 15, 3884-3890. 42 Ladner, £E., Kitchell, J.P., Honzatko, R.B., Ke, H.M., Volz, K.W., Kalb (Gilboa), A.J., Ladner, R.C. and Lipscomb, W.N. (1982) Gross quaternary changes in aspartate carbamoyitransferase are induced by the binding of N.(phosphonacetyl)-L-aspartate. A 3.5 A Resolution study. Proc. Natl. Acad. Sci. USA 79, 3125-3128. 43 Newell, J.O.. Markby, D.W. and Schachman, H.K. (1989) Cooperative binding of the bisubstrate analog N.(phosphonacetyl)-L-aspartate to aspartate transcarbamoylase and the heterotropic effects of ATP and CTP. J. Biol. Chem. 264, 2476-2481. 44 Blackburn, M.N. and Schachman. H.K. (1977) Allosteric regulation of aspartate transcarbamoylase. Effect of active site ligands on the reaction of the sulfhydryl groups of the regulatory subunit. Biochemistry 16. 5084-5091. 45 Ladjimi, M.M. and Kantrowitz, E.R. (1988) A possible model for the concerted allosteric transition in Escherichia coli aspartate transcarbamylase as deduced from site-directed mutagenesis studies. Biochemistry 27, 276-283. 46 Wales, M.E., Hoover, T.A. and Wild, J.R. (1988) Site-specific substitution of the Tyr-165 residue in the catalytic chain of aspartate transcarbamylase promotes a T-state preference in the holoenzyme. J. Biol. Chem. 263, 6109-6114. 47 Newell, J.O. and Schachman, H,K. 0988) Effects of mutational alterations in the catalytic chains of aspartate transcarbamoylase (ATCase} on allosteric regulation of conformational transitions. FASEB Journal 2, A551. 48 Ladjimi, M.M., Middleton, S.A., Kelleher, K.S. and Kantrowitz, E.R. (1988) Relationship between domain closure and binding, catalysis and regulation in £. coli aspartate transcarbamylase. Biochemistry 27, 268-276. 49 Tanford, C., Kawahara, K. and Lapanje, S. (1966) Proteins in 6 M guanidine hydrochloride. Demonstration of random coil behavior. J. Biol. Chem. 241, 1921-1923.