A Method for Determining the Sedimentation Behavior - Journal of ...


21 févr. 2014 (il y a 2 années et 9 mois)

85 vue(s)

Robert G. Martin and Bruce N. Ames
Application to Protein Mixtures
Sedimentation Behavior of Enzymes:
A Method for Determining the
1961, 236:1372-1379.J. Biol. Chem. 
Access the most updated version of this article at
JBC AffinityFind articles, minireviews, Reflections and Classics on similar topics on the
When a correction for this article is posted 
When this article is cited 
to choose from all of JBC's e-mail alertsClick here
This article cites 0 references, 0 of which can be accessed free at
by guest on February 21, 2014http://www.jbc.org/Downloaded from
by guest on February 21, 2014http://www.jbc.org/Downloaded from
Vol. 236, No. 5, May 1961
P&&d in U.S.A.
A Method for Determining the Sedimentation Behavior
of Enzymes: Application to Protein Mixtures
From the National Institute of Arthritis and Metabolic Diseases, National Institutes of Health,
United States Public Health Service, Bethesda, Maryland
(Received for publication, December 5, 1960)
During an investigation of the gene-enzyme relationships in
histidine biosynthesis in Xalmonella typhimurium, it became
desirable to determine, in crude extracts, the approximate mo-
lecular weights of several enzymes. We have found sucrose
gradient centrifugation to be a suitable method for determining
sedimentation coefficients of enzymes in protein mixtures. A
variety of enzymes of known properties have been studied in the
development of the method.
Although the separation cell of Yphantis and Waugh (1) has
been demonstrated to be applicable to the determination of
molecular weights when multiprotein solutions are used, the
present method has several advantages over that system and
these advantages will be discussed.
Sucrose gradient centrifugation, using the swinging bucket
head of the preparative ultracentrifuge, has been used exten-
sively in the determination of sedimentation constants of vi-
ruses, mitochondria, microsomes, and ribosomes (2, 3). The
adaptation of this method to relatively low molecular weight
substances is reported here.
In the sucrose gradient technique the sample to be studied is
layered on a gradient and materials of different sedimentation
properties separate from each other during the centrifugation.
A hole is then punched in the bottom of the centrifuge tube and
fractions are collected and analyzed.
With slight modifications in the design of the apparatus for
gradient production and fractionation, we have found the
method to be simple and accurate. Sedimentation coefficients
have been determined for a number of well characterized en-
zymes as well as a sample of ribonucleic acid, and the results
are in good agreement with the values reported by others.
same values have been obtained whether the enzymes were
analyzed as pure proteins or mixed with crude extracts.
Although sedimentation coefficients were directly calculated
for a variety of substances in order to determine the accuracy of
the method, in general use the procedure may be simplified.
The sedimentation coefficient (or approximate molecular weight)
of an unknown enzyme may be determined by a simple ratio of
mobilities when a standard well characterized enzyme has been
added to the protein mixture.
With the use of this technique the sedimentation behavior of
several of the enzymes in the pathway of histidine biosynthesis
in S. typhimurium has been determined.
* This work was begun during service as an officer in the United
States Public Health Service under the Commissioned Officers:
Student Training and Extern Program (CO-STEP).
Materials and AssaysLyophilized yeast alcohol dehydro-
genase obtained from Worthington Biochemical Corporation
was dissolved in 0.05 M Tris-HCl buffer, pH 7.5, to a concentra-
tion of 10 mg per ml and stored at 3”.
Before use on the sucrose
gradient it was diluted to 0.20 mg per ml with the Tris buffer.
The dehydrogenase was assayed in a Cary spectrophotometer
by following the increase in absorption at 340 rnp for 20 seconds
of a l-ml reaction mixture containing 170 pmoles of ethanol, 50
pmoles of Tris, pH 8.5, 15 pmoles of DPN, and 5 to 20 ~1 of
enzyme fraction. Units of activity were expressed in terms of
change in absorbancy per 20 seconds per 10 ~1 of enzyme frac-
Lyophilized egg white lysozyme (Worthington, twice crystal-
lized) was dissolved in 0.05 M Tris buffer, pH 7.5, to a concen-
tration of 100 mg per ml and stored at 3”. Before use it was
diluted to 5 mg per ml in this buffer. Lysozyme was assayed
by following the decrease in turbidity at 650 rnp of a l-ml reac-
tion mixture containing 10 pmoles of Tris, pH 8.0, and enough
Micrococcus Zysodeikticus cell walls (Difco Laboratories, Bacto-
lysozyme substrate) to give an absorbancy of approximately 2.0
absorbancy units in the Cary spectrophotometer. The reaction
was started with 5 to 20 ~1 of enzyme fraction and activity was
expressed in terms of the change in absorbancy per 20 seconds
per 10 ~1 of enzyme fraction.
Beef liver catalase was obtained as an aqueous ammonium
sulfate suspension of approximately 40 mg per ml of protein
(Worthington). Before use, it was diluted to 0.40 mg per ml
in 0.05 M Tris buffer, pH 7.5. Catalase was assayed by follow-
ing the decrease in absorbancy at 240 rnp of a 3-ml reaction
mixture containing 30 amoles of potassium phosphate buffer
at pH 7.5, 18 pmoles of HzOl, and 5 to 20 ~1 of enzyme fraction.
Activities were calculated in terms of change in absorbancy per
20 seconds per 5 ~1 of enzyme fraction.
All enzyme assays were approximately linear in the concen-
tration ranges used.
Soluble RNA from rabbit liver in a concentration of approxi-
mately 6.5 mg per ml (130 absorbancy units per ml at 260 rnp)
was kindly supplied by Dr. Samuel Luborsky. The RNA was
stored at -15” and diluted with 2 volumes of water before use.
It was assayed by its absorption at 260 rnp.
Extracts and Assays of His&line Biosynthetic Enzymes-The
histidine mutants were obtained from Dr. P. E. Hartman. The
medium, growth of strains, preparation of Nossal extracts, and
assays have been previously described (4) with the exception
by guest on February 21, 2014http://www.jbc.org/Downloaded from
May 1961
R. G. Martin and B. N. Ames
that histidinol was substituted for histidine in the growth media
of the histidine-requiring mutants.1 Crude, undialyzed extracts
were usually placed directly on the sucrose gradient. In some
cases the extract was subjected to Sephadex gel filtration (5)
by passing 1 ml of extract through a l- X 5-cm column contain-
ing 1 g of Sephadex G-25 (Pharmacia Company, Uppsala,
Sweden), equilibrated with 0.01 M Tris, pH 7.5. The protein
solution was eluted with the Tris buffer, and the 1.2 ml which
contained the protein free from salt were collected.
In order to obtain a partial purification of the histidine en-
zymes, a DEAE-cellulose (6) column similar to that previously
described was used (7). SaZmonelZa hisG-46 mutant cells were
grown and harvested. The cells were suspended in 10 ml of
TSM buffer2 and subjected to sonic oscillation for 10 minutes
at 0” in a lo-kc Raytheon sonic oscillator. The extract was
spun at 25,000 x 9 for 45 minutes and the supernatant was
passed through a 3- x 21-cm, 20-g Sephadex G-25 column which
had been equilibrated with TSM buffer.
The fractions showing
appreciable activity were combined. (The specific activity of
this sonicated extract was approximately the same as the spe-
cific activities obtained on the Nossal extracts,) The enzyme
was then purified on a 2- x 25-cm, 8.3-g gravity-packed DEAE-
cellulose column, washed according to Peterson and Sober (6),
and equilibrated with TSM buffer. The enzyme extract (20
ml) was washed into the column with 5 ml of TSM buffer and
then eluted with a linear 0.00 to 0.80 M NaCl gradient in TSM
buffer. Fractions of approximately 2.3 ml were collected every
90 seconds. Protein concentrations were determined according
to the method of Lowry et al. (8). As previously reported (7),
complete separation of histidinol dehydrogenase and imidazole
acetol phosphate transaminase from each other and from the
histidinol phosphate phosphatase-imidazole glycerol phosphate
dehydrase peak was obtained.
The ratio of phosphatase activ-
ity to dehydrase activity was constant throughout the eluted
fractions (7). A IO-fold increase in specific activity was ob-
tained for each of the enzymes. The fractions with maximal
activity had a 450-fold higher specific activity than wild-type
and contained less than 2.5% nucleic acid based on their 280 to
260 rnh ratio of 1.1 (9).
Apparatus for Making Sucrose Gradients-A modification of
the simple design of Britten and Roberts (3) was used to pro-
duce linear sucrose gradients.
Their design consists of a block
of Lucite containing two chambers that are connected at the
bottom by a removable screw pin. An outflow tube extends
from one chamber. This basic design was altered only in that
a stopcock was introduced between the two chambers to replace
the screw pin. We employed a polyethylene outflow tube which
was drawn out in a flame so that emptying time with 2.3 ml of
sucrose in each chamber was approximately 10 minutes. The
chamber next to the outflow tube was stirred with a platinum
bacteriological inoculating loop which was mounted on a motor.
The stirring speed was adjusted to give good mixing with mini-
mal disturbance of the meniscus.
The apparatus was filled by turning the stopcock to the
open position so that free flow existed between the two
chambers. The less dense sucrose solution was then added to
1 Specific activities for the histidine enzymes up to 40 times wild
type have been observed when Salmonella his mutants are grown
on 0.05 rnM histidinol .
* TSM buffer: 0.01 M Tris (free base), 0.005 M magnesium ace-
ta,te, and 0.004 M succinic acid, the pH adjusted with NaOH to 7.6,
one chamber and the block rocked back and forth to free air
bubbles that might have been caught in the passage between
the two chambers. The stopcock was then turned to the closed
position and the two chambers were emptied and wiped dry.
The block was placed in a clamp, the stirrer adjusted, the out-
flow tubing bent upward so that its tip was above the top of
the chambers, and each of the two chambers was filled with 2.3
ml of sucrose-buffer solution of the desired concentration. In
all the experiments reported here, 20% (weight per volume) of
cold sucrose (0.584 M) in 0.05 M Tris-HCl buffer at pH 7.5 was
placed in the mixing chamber and 5% (weight per volume) of
cold sucrose (0.146 M) in 0.05 M Tris-HCl buffer at pH 7.5 was
placed in the adjacent chamber. The rotor was started, the
stopcock opened, and the tip of the outflow tube placed at the
top of a Lusteroid centrifuge tube. Occasionally gentle suction
was required to start the flow. To assure linearity of the gra-
dient, care was taken to observe that the fluid levels in the two
chambers were equal during emptying.
In initial experiments the production of gradients was tested
by mixing dichlorophenolindophenol with the 20 y0 sucrose
solution. Perfectly linear plots of absorbancy at 600 mp against
fraction number were obtained when 44 fractions were collected.
Gradients were stable for at least 48 hours.
A 0.05-ml hold-up
volume in the apparatus resulted in the delivery of only 4.55
ml to the centrifuge tube.
Layering of Sample-The gradients were stored in a 3” cold
room for 4 to 18 hours before use.
TO start a run, the substance
to be studied was diluted to the desired concentration and 0.10
ml was layered on the gradient; care was taken to avoid bubble
formation. It is essential that the material to be layered on
the gradient float on 5% sucrose. To insure convection-free
sedimentation, protein solutions of considerably less than 5%
(approximately 2% or less) must be used (3). A sharp inter-
face between the sucrose and protein solution was always ob-
served and this interface remained distinct for the several min-
utes required to load the centrifuge tubes into the rotor buckets.
Centrifugation-The characteristics and dimensions of the
swinging bucket rotor SW-39 designed to fit the model L Spinco
centrifuge (Beckman Instruments, Inc., Spinco Division, Palo
Alto, California) have been described (lo), as have the errors
resulting from the use of nonsector-shaped centrifuge tubes (2).
The total volume used in these experiments was 4.65 ml (4.55-ml
gradient plus O.l-ml sample) and the distance from rotor center
to meniscus, allowing 0.01 cm for rotor stretch and 0.03 cm for
radial shift of the meniscus, was calculated to be 6.02 cm (10).
Further correction is required because the protein layer is not
infinitesimally thin. It was assumed that the protein moves
from the middle of the layer. The calculated distance from the
rotor center to the middle of the protein layer was 6.06 cm.
Because of the stability rendered to the solution by the sucrose
gradient (2, 3) very much less care was needed in the operation
of the centrifuge than described by Hogeboom and Kuff (10).
The rotor was accelerated very slowly for approximately 10
seconds to eliminate the initial lash given to the rotor by the
drive shaft when the two were not fully engaged. After this,
the r.p.m. control knob was immediately turned to full speed, a
setting of 39,000 r.p.m.
The rotor was decelerated by turning
the time knob to zero and allowing the rotor to coast to a halt
with the brake off.
It is worthy of note that accurate calculations of rotor speed
must be based on the odometer readings; the speed setting and
by guest on February 21, 2014http://www.jbc.org/Downloaded from
1374 Sedimentation Behavior of Enzymes
Vol. 236, No. 5
tachometer are unreliable.3 During relatively long runs (over
4 hours), the rotor speed tends to vary somewhat. Conse-
quently, for each centrifugation the odometer was read before
starting and at the time when deceleration was begun so that an
average speed was obtained. The time was then corrected to
the equivalent time of centrifugation at 38,000 r.p.m. by the
equation, (r.p.m.)tti = (r.p.m.)$z. The time of centrifugation
was taken as the period from the start of acceleration of the
rotor to the start of deceleration.
Deceleration took 13 minutes
from full speed, and integration of the plot of time against speed
during deceleration gave an estimated 41 minutes of centrifuga-
tion at 38,000 r.p.m. Acceleration took 4 minutes, which was
equivalent to 24 minutes of centrifugation at 38,000 r.p.m.
Therefore, 3 minutes were added to the time of centrifugation.
Another troublesome aspect of the centrifuge was the main-
tenance of a constant temperature during the centrifugation,
this perhaps because our instrument was not equipped with an
auxiliary ultrahigh vacuum system. Successful runs at 3” in the
swinging bucket rotor were accomplished only by precooling the
rotor chamber (with the vacuum on) to its lowest setting (-18”
on our instrument). The precooled head (3”) was then rapidly
loaded into the centrifuge. With the above precautions, the
temperature increase of the samples during 17 hours of centri-
fugation could be kept to less than O.8o.4
Sampling-In their work with CsCl gradient centrifugation,
Weigle et al. (11) emptied centrifuge tubes progressively and
uniformly by punching a hole in the bottom of each tube with a
needle and collecting the drops. We have devised a simple
fractronator, based on this principle, which yields a constant
number of drops from the 4.65 ml in each tube5 (Fig. 1). In
operation, the bottom end piece with its needle was carefully
cleaned and forced into place.
After the fractionator had been
placed in a clamp, the Lusteroid centrifuge tube was carefully
lowered into the apparatus with a pair of forceps. The upper
end piece was positioned and pressed down, forcing the Lusteroid
tube through the needle and starting the flow of drops. The
rate of flow could be controlled by the syringe and was kept at
approximately 1 drop per second. A soft rubber gasket pre-
vented leakage around the needle. With the system described
above, it was possible to collect 308 =t 5 drops which were usually
divided into 44 fractions (7 drops in each). As the needle rested
approximately 0.7 mm above the bottom of the tube, about 3
drops remained in the tube. To check the efficiency of the
fractionation system, a gradient was made similar to the one de-
scribed for testing the gradient-making apparatus, with dye in
the 20% sucrose. The samples were collected from the frac-
tionator and the absorbancy at 600 rnp was determined. A
linear plot of tube number against absorption was obtained
through tube 43. The last fraction had slightly higher extinc-
tion than expected. Presumably this was due to mixing of the
solutions below and above the level of the needle, resulting from
the turbulence produced when the last frothy drops are forced
through the needle. Although it might be expected that the
sucrose concentration would affect the drop size, the volume
difference between the first and last fractions was less than 3%.
Drops were, therefore, considered to be of constant size, 14.7 f
8 Beckman Instruments, Inc., a personal communication.
4 The temperature immediately before and after centrifugation
was determined in a bucket containing only a sucrose gradient.
6 Subsequent to this investigation, a more elaborate but essen-
tially similar fractionator was reported (12).
FIG. 1. A Nalgene drying tube with two end pieces (No. 1251 B,
Phipps & Bird, Inc., Richmond, Virginia) was cut to 2.5 inches.
A Nalgene centrifuge tube, 100 X 16 mm (No. 1210-1, Phipps &
Bird, Inc.) with an internal diameter which just allowed easy
passage of the Lusteroid centrifuge tube was cut to a 1.5.inch
hollow cylinder. This cylinder was forced into the drying tube
far enough to allow the bottom end piece of the drying tube to
rest against the cylinder when the end niece was inserted in the
drying tube.
The bottom end piece was-then fitted with a No. 00
rubber stopper carefully whittled to fit snugly. On this was
placed a circular piece of soft rubber with a diameter equal to the
internal diameter of the end piece.
A 21.gauge syringe needle
with the adapter end broken off and the sharp end filed down to a
double bevel was forced up through the rubber stopper and soft
rubber gasket so that it protruded 1 mm above the latter. A
cork with a &mm diameter bore through it was placed in the upper
end piece and cut off so that it was even with the plastic ridge of
the end piece. The upper end piece was then fitted with a rubber
hose to a 50-ml syringe.
0.4 ~1. As an added precaution to maintain drop size, the needle
was frequently cleaned with a stylette.
The definition of the sedimentation constant ST,,,, in a medium,
m, at temperature, T, is given by the equation (2, 13):
where w is the angular velocity of the rotor in radians per second,
z is the distance from the rotor center to the boundary, and &/dt
is the velocity of movement of the boundary.
The sedimenta-
tion constant, s, is generally extrapolated to the “standard state”
taken as that of water at 20’:
?m,mbp - Pz0.w)
s2om = STm
~ZO,wbp - PT,m)
where fT,rn
is the viscosity of the medium at the temperature of
centrifugation, r]20, z. is the viscosity of water at 20”, pp is the
density of the protein in solution (i.e. the reciprocal of the partial
specific volume, i), pT,rn is the density of the medium at the
temperature of centrifugation, and ~20,~ is the density of water
at 20”. As the partial specific volume of most proteins varies
little with temperature (14), p, is generally considered constant.
by guest on February 21, 2014http://www.jbc.org/Downloaded from
May 1961
R. G. Martin and B. N. Ames
3.59 -
3.00 -
AT 38,000 RPM
FIG. 2. Theoretical sedimentation behavior of macromolecules
in a 5% (weight per volume) to 20% (weight per volume) sucrose
gradient at 3”.
The partial specific volumes (fi) in cm3 per g and
sedimentation constants at 20” in water (szo.~) in Svedberg units
(S) are indicated.
Equation 2 is applicable to centrifugation in a sucrose gradient,
as it is for centrifugation in uniform media, but in the former case
both the viscosity (~r,~) and the density (PT,~) are functions
of the sucrose concentration and hence of the distance of the
medium from the rotor center. Combining Equations 1 and 2:
ho.wdt = A (pp _ PT,m)
* J dx
(3) x
where the term A = ( pp - p20, ,) /qZo, w is a constant for any given
partial specific volume (1 /p,)
The left-hand side of Equation 3 ran be readily integrated:
[” S20.,rww = sm.&&
The right-hand side of Equation 3 can be numerically inte-
grated using the trapezoidal approximation:
%+I F(x)& = y F(Xi)(Xi+1 - Xi)
0 0
+ ; $2 [F(Xi.l) - F(Zi)l[Xi.l - Zil
This is performed by tabulating arbitrary distances (xi) from
the rotor center starting at 6.10 cm (the start of the sucrose
gradient) and ending at 9.62 cm (the point at which the sucrose
gradient would end if the centrifuge tube were perfectly cylindri-
cal) against sucrose concentration. With standard tables of
sucrose molarity against density and viscosity at given tempera-
tures (2), F(z) = q~,~(zi)/[p, - pr,m(si)]~i can be calculated
for each distance, zi, and hence the right side of Equation 3
determined for any assumed pp. Plotting zi against & (or zi
against t for a given w), one obtains a family of theoretical curves
for substances of different ~20,~ values and an assumed pp.6 Fig.
2 shows the theoretical curves for centrifugation at 3” for sub-
stances of partial specific volume 0.725 cm3 per g, and ~20,~ values
of 11.0, 7.4, and 2.15 S. In the same figure are the curves for
substances of SZO,~ 11.0 S and partial specific volumes 0.500,
0.725, and 0.800 cm3 per g. Fig. 2 indicates that a very nearly
linear relationship should exist between the distance traveled
Large differences in partial specific volume result in signifi-
cantly different values when the SZO,~ is calculated. If the par-
tial specific volume for a protein is assumed to be 0.800 cm3 per
g, an s20,ur can be calculated, whereas a slightly different s20,w
will be arrived at if a partial specific volume of 0.500 cm3 per g
is assumed. For example, one cannot adequately distinguish
between substances with ~20,~ values of 11.0, 12.0, and 12.8 S
with corresponding partial specific volumes of 0.500, 0.725, and
0.800 cm3 per g. However, since most proteins have partial
specific volumes between 0.700 and 0.750 cm3 per g (14), the
assumption in all calculations of a partial specific volume of
0.725 cm3 per g will result in less than 3 $& error in the estimation
of s20,ul for most proteins. Alternatively, one may define an
.s~~,~ as the ~20, w calculated on the assumption of a partial specific
volume of 0.725 cm3 per g. Correction of the s,“,lz so defined to
the true s20,u, cannot be accomplished by a simple mathematical
ratio if the partial specific volume is determined later. In Fig,
3 the effect of partial specific volume on the calculated s~,,,~ has
been plotted for s~,~~
values of 1 to 15 S. Interpolating from
these curves it is possible to obtain the SQO,,,, of a substance from
its s20,zu
0’725 when the partial specific volume is available.
“Xtandard” Enzymes-Three well characterized crystalline en-
zymes of different sedimentation rates (yeast alcohol dehydro-
genase, bovine liver catalase, and egg white lysozyme) were
chosen as standards. In Fig. 4A the sedimentation pattern of
catalase after 4 hours of centrifugation at approximately 20” is
shown. The dotted line represents the sedimentation pattern
obtained from a second centrifuge tube run concomitantly in
which catalase at the same concentration had been mixed with
yeast alcohol dehydrogenase before centrifugation. Fig. 4B
shows the patterns for yeast alcohol dehydrogenase during the
same experiment.’ In other experiments these two enzymes also
7 To obtain the greatest reproducibility it was found necessary
0 This method of analysis is similar to that of de Duve et al.
to count the number of drops in the last fraction and to plot the
(2) except that a different approximation has been used to find the
width of this fraction accordinalv. This has been done in all
numerically determined integral.
sedimentation patterns reported: -
012345 6 7 8 9 IO II 12 13 1415 I6 I7
S20,w’N s
FIG. 3. Theoretical curves demonstrating the effect of assumed
martial soecific volume unon the calculation of SZO.+.. These
curves ar’e applicable only for the particular sucrose-gradient
employed and for a rotor head of the dimensions of the SW-39.
The curves are very steep in the range 0.70 to 0.75 cm3 per g fl.
by the enzyme and the time of centrifugation when substances
of partial specific volume less than 0.80 cm3 per g are examined.
by guest on February 21, 2014http://www.jbc.org/Downloaded from
1376 Sedimentation Behavior of Enzymes Vol. 236, No. 5
appeared to move at their characteristic rates in pure solution or
combined with other proteins.
Fig. 5 shows an example of the centrifugation patterns of
catalase, yeast alcohol dehydrogenase, and lysozyme obtained
by this method. The dotted lines represent the corresponding
enzymes which have been diluted to the same concentration in a
crude extract of S. typhimurium. The distance in centimeters
from the meniscus to the enzyme peak was estimated to two deci-
mal places by the symmetry of the curve.
In Fig. 6 the distances
of the enzyme peaks from the meniscus, as determined from Fig.
5 and many similar centrifugation experiments (all at 3”), are
plotted as a function of the time of centrifugation.
The time of
centrifugation was corrected to the equivalent time at 38,000
r.p.m. (The correction to the time of centrifugation at 38,000
r.p.m. includes both the corrections for the actual rotor speed
and for the time of rotor acceleration and deceleration.) A
similar set of curves was obtained at 15”; steeper slopes were
As predicted, the rate of centrifugation is very nearly constant
for any one of these three enzymes. The .&E” determined for
bovine liver catalase by this method was 11.3 S at 3”. (An
identical value was obtained at 15”.) With the data in Fig. 3
40 36 32
l.3?5 I.9158 26110
26 24 20 16 12
FIG. 4. Sucrose gradients were placed in each of the three
buckets of the SW-39 rotor. Catalase (0.04 mg in 0.10 ml) was
layered on gradient 1, yeast alcohol dehydrogenase (0.02 mg in
0.10 ml) on gradient 2, and a mixture of the two enzymes (0.04
mg of catalase and 0.02 mg of dehydrogenase in 0.10 ml) on gra-
dient 3. The rotor was run at approximately 20” for 4 hours at
38,000 r.p.m. A. Catalase activity was assayed in each fraction
of gradient 1 (solid line) and gradient 3 (dotted line). B. Yeast
alcohol dehydrogenase was assayed in each fraction of gradients
2 (solid line) and 3 (dotted line).
36 32 26 24 20 16 I2 6 4
FIG. 5. Lysozyme (0.5 mg), catalase (0.04 mg), and yeast alco-
hol dehydrogenase (0.02 mg) mixed in 0.10 ml of 0.01 M Tris buffer,
pH 7.5, were layered on a sucrose gradient. After 12.8 hours of
centrifugation at 37,700 r.p.m., 3”, the gradient was fractionated
and analyzed (solid lines). In a second gradient, centrifuged at
the same time, these enzymes were diluted to the same final con-
centrations in a crude extract of Salmonella mutant hisG46 (dotted
linfv) .
FIG. 6. Each point represents the results of a centrifugation
experiment similar to Figs. 5 or 7. All experiments were carried
out at 3”.
The time of each experiment has been corrected to the
equivalent time of centrifugation at 38,000 r.p.m. The solid lines
represent the theoretical sedimentation behavior for macromole-
cules of partial specific value 0.725 cm3 per g and the indicated
~20,. values (in Svedberg units).
and if a partial specific volume of 0.73 cm3 per g (15) is assumed,
the calculated s20,w is 11.4 S. Previous investigators (15) re-
ported a value of 11.3 S (concentration not stated) with optical
Yeast alcohol dehydrogenase has been reported to have an
s20,2u of 7.2 (1 y. solution) (lo), 7.61 (concentration not stated)
(16), and 6.72 S (extrapolated to infinite dilution, centrifugation
at 0’) (17), in the optical centrifuge.
An ~20,~ of 7.6 S (concen-
tration, 0.0005%, 25”) has been reported by Kuff et al. with
their technique (18). In these experiments an a~,~~ of 7.4 S was
found. With the reported (17) partial specific volume for this
enzyme (0.769 cm3 per g) and the data of Fig. 3, an ~20,~ of 7.6
S was calculated.
The .s~~,~ values that have been reported for egg white lyso-
by guest on February 21, 2014http://www.jbc.org/Downloaded from
zyme in the optical centrifuge are 2.11 (1.3% protein), 2.09
(0.9% protein), 2.12 (0.5% protein) (19), 1.9 (1.5 to 0.5% pro-
*F Q zt+
tein) (20), 1.8 (concentration not stated) (21), and 1.94 S (1%
E u .-*
protein) (10). Hogeboom and Kuff (lo), using 1% solutions,
obtained s20,w values of 2.01 and 1.88 S. The .&~~ determined
here was 2.1 S; corrected to a partial specific volume of 0.722
cm3 per g (19), the SPO,,,, is also 2.1 S.
Soluble RNA, a high molecular weight substance of partial
$3 ;
specific volume markedly different from protein, was tested on
a sucrose gradient. A typical centrifugation pattern is shown
in Fig. 7. Fig. 6 indicates that soluble RNA also sediments in
0.65 1.31 1.96 2.61 3.59
this gradient with a constant velocity. The s$lr was calculated
FRACTION 44 40 316 32
:8 24 $0 16 I:
to be 4.6 S.
8 4
The .s~~,~* and partial specific volume of the same
preparation of rabbit liver soluble RNA were determined by
FIG. 7. A centrifugation pattern for rabbit liver soluble RNA
Dr. Samuel Luborsky.9 In a model E Spinco centrifuge
Approximately 0.21 mg of this material in 0.10 ml was layered on a
equipped with ultraviolet optics he obtained a pattern with a
sucrose gradient which was 0.2 N in NaCl (no buffer). The frac-
single major peak at 4.5 f 0.1 S and small amounts of material
tions were assayed by absorbancy (optical density) at 260 rnr
after 9.15 hours of centrifugation at 38,000 r.p.m., 3”.
of lower and higher sedimentation rate. The partial specific
volume determined pycnometrically was 0.48 cm3 per g (22).
The so’726 20,,,, of 4.6 S is equivalent to an SZ~,~ of 4.4 S with this par-
tial specific volume.
w .080 Histidine Biosynthetic Enzyme-We have determined the ap-
proximate molecular weights of several of the enzymes of histi-
= .060
dine biosynthesis by examining partially purified enzyme prep-
arations as well as crude extracts. That crude extracts may be
used for the determination of sedimentation coefficients is dem-
The same sedimentation constants were ob-
onstrated in Fig. 5.
.020 t-
tained with crude and partially purified preparations of the
histidine enzymes.
1 I
Partially purified histidinol dehydrogenase and imidazole
0.75 I.71 ? 0.65
acetol phosphate transaminase were placed on sucrose gradients
NUMBER 44 40 36 32 28 24 44 40 ;6 32 ;8 24
and centrifuged (Fig. 8). Fig. 9 shows the combined results for
FIG. 8. Sedimentation patterns for histidinol dehydrogenase the centrifugation determinations of crude Nossal extracts and
and imidazoleacetol phosphate transaminase. A crude extract of
Salmonella mutant hisEF-135 was centrifuged for 16.50 hours at
of the partially purified enzymes. The calculated s$lf values
33,300 r.p.m., 3”. The dehydrogenase activity is expressed in
for these enzymes are 5.1 S for the dehydrogenase and 4.8 S for
change in absorbancy at 600 rnp per minute per 20 ~1 of enzyme
the transaminase.
fraction. The transaminase activity is expressed in change in Preliminary studies have been carried out on phosphoribosyl-
absorbancy at 295 rns per 20 minutes per 20 ~1 of enzyme fraction.
ATP pyrophosphorylase (22), the first enzyme of the histidine
biosynthetic pathway. This enzyme has a sedimentation con-
stant of about 8.6 S.
May 1961 R. G. Martin and B. N. Ames 1377
The centrifugation results with imidazole glycerol phosphate
dehydrase and histidinol phosphate phosphatase indicate aggre-
gation. The sedimentation patterns are complicated in that
multiple peaks appear. Magnesium ions or mercaptoethanol
alters these peaks. A slow moving component is the major one
in the absence of these substances, and there is, primarily, a heavy
component in the presence of mercaptoethanol or magnesium.
The dehydrase and phosphatase activities appeared to migrate
together in several experiments, although further work on this
complicated system is necessary.
FIG. 9. Sedimentation behavior of histidinol dehydrogenase The ultracentrifugation technique presented here (2, 3) differs
(0) and imidazoleacetol phosphate transaminase (0). Each
point represents a centrifugation experiment similar to the one
from the usual methods of analysis in several aspects. A major
shown in Fig. 8. The enzyme source for each set of experiments
attribute of this system derives from the particular sucrose
was a Salmonella his mutant as indicated in the figure. Both
gradient employed. The viscosity and density of this sucrose
enzymes were assayed from aliquots of the same set of fractions
obtained after centrifugation. The curves represent the theoreti-
* The medium used in Dr. Luborsky’s experiments and our own
cal sedimentation behavior for proteins of partial specific volume was 0.2 M NaCl (no buffer).
0.725 cm3 per g and ~20,~ values of 4.8 S (dotted line) and 5.1 S
Q S. Luborsky and G. L. Cantoni, to be submitted for publica-
(solid line).
tion, Biochim. et Biophys. Acta.
by guest on February 21, 2014http://www.jbc.org/Downloaded from
1378 Sedimentation Behavior of Enzymes Vol. 236, No. 5
gradient are such (at least in the temperature range 3” to 15”),
that essentially linear migration of most biological materials
results. Thus, the ratio of the distances traveled from the
meniscus by any two substances will always be constant.
fore, if an unknown substance is compared with a standard,
careful control of temperature, time, and speed of centrifugation
are unnecessary. The use of such a standard of known sedi-
mentation coefficient highly simplifies the technique. Experi-
mentally, the ratio, R, can be easily determined after any time
of centrifugation:
R = distance travelled from meniscus by unknown
distance travelled from meniscus by standard
And because of the nearly constant rate of movement of any
macromolecule :
~2”;::~ of unknown
s!$,~~ of standard
Or, for macromolecules of the same partial specific volume:
~20.~ of unknown
~~0.~ of standard
For substances of similar partial specific volumes, this last equa-
tion will be very nearly correct. For more accurate determina-
tions of sedimentation constant, the approach outlined in “The-
oretical” may be followed.
A second aspect which distinguishes this technique from con-
ventional optical methods is the form of analysis used. Multiple
fractions of the solution are obtained and these fractions may be
analyzed for any of a variety of properties: radioactivity, en-
zymatic activity, chemical properties, etc. Thus, a particular
biological material in a multicomponent mixture may be localized
by its chemical activity. And hence, the sedimentation coefficient
of a biologically active substance may be determined in a crude
extract. Furthermore, in some cases amounts of material so
small as to be undetectable by the most sensitive optical tech-
niques may be detectable by another parameter. The disadvan-
tage of any technique in which a parameter other than optical
measurement is used is that the centrifugation pattern deter-
mined obtains for a particular time of centrifugation, and
multiple patterns cannot be gotten on the same sample.
Other systems for the determination of sedimentation coeffi-
cients in crude extracts, particularly the separation cell of
Yphantis and Waugh (l), have been demonstrated to be highly
efficient. In the present system, multiple fractions are obtained
and analyzed, whereas only two are obtained in the separation
cell. Several advantages arise from the analysis of multiple
fractions: protein aggregation which can easily go undetected in
the separation cell is readily detected; multiple enzymes of
widely divergent sedimentation coefficients can be analyzed in
the same experiment; and high resolution of the sedimentation
pattern is possible. In addition, the present technique employs
the less expensive “preparative” ultracentrifuge. A disadvan-
tage of this technique relative to the separation cell is that dif-
fusion coefficients cannot be directly determined.
With the use of this technique, a moving zone of material is
analyzed rather than the boundary of an initially uniform solu-
tion. Because materials of different sedimentation properties
are separated from each other during the centrifugation, the
Molecular weight of hi&dine biosynthetic enzymes
Lysozyme . . . . . . 63,000 57,000 140,000
Alcohol dehydrogenase . . . 86,000 78,000 190,000
Catalase . . . . . . . . .
75,000 68,000 170,000
procedure may be used for enzyme purification. Indeed, small
volumes of enzyme have been partially purified with the SW-39
rotor,lO and good results with much larger volumes have been
achieved in preliminary studies with the SW-25 swinging bucket
No protein-protein interactions or protein-nucleic acid inter-
actions were observed in this investigation. Even lysozyme, a
basic protein, showed the same behavior when mixed in a crude
extract and when pure. This lack of interaction may be due to
some effect of the sucrose in minimizing interactions or to the
fact that very dilute solutions were used compared to what is
required in an analytical centrifuge. Nonetheless, protein-pro-
tein interactions are a potential source of error in protein mix-
The greatest disadvantage of sucrose gradient centrifugation
is the necessity of knowing the partial specific volume in order
to determine the true s~~,~. However, as discussed previously,
the error from the assumption of a partial specific volume of
0.725 cm3 per g for any protein will be small.
A crude estimation of the molecular weight (MIV), can be ob-
tained from the sedimentation constant alone (13) :
MW1 *
-= -
( J
and for most proteins the ratio sl/sZ is equal to R. (See Equation
4.) This equation derives from the fact that many proteins are
essentially spherical molecules. Although most globular pro-
teins, i.e. nearly all enzymes, are only roughly spherical, the
relationship between s and MW is approximately correct (13).
With the above approximation and the data presented above,
the molecular weights of the histidine biosynthetic enzymes have
been calculated (Table I). Lysozyme (MW = 17,200 (19)),
alcohol dehydrogenase (MW = 150,000 (17)), and catalase
(MW = 250,000 (15)) were used as standards.
The variation in the estimated molecular weight for any partic-
ular enzyme (Table I) may be due to two factors.
The standard
enzymes vary somewhat in shape, i.e. they are not perfect
spheres. Also, there is some inaccuracy in the reported molecu-
lar weights of the standards.
10 E. Racker and M. Maver, personal communication.
11 One milliliter of a crude extract of Salmonella hisEF 135 was
placed directly on a 29-m& 5 to 40% sucrose gradient and centri-
fuged for 24 hours in the SW-25 swinging bucket rotor.
l-ml fractions were collected. Eighty-seven per cent of the input
activity of phosphoribosyl-ATP pyrophosphorylase (23) was found
in four fractions. The two peak tubes containing over 60% of the
input activity had specific activities close to lo-fold greater than
the starting material. A large portion of the nucleic acid present
in the crude extract was also removed from these fractions, judg-
ing from the absorbancy at 280 and 260 m/L.
by guest on February 21, 2014http://www.jbc.org/Downloaded from
May 1961
R. G. Martin and B. N. Ames
We were interested in seeing whether any correlation could be
made between the molecular weights of the histidine biosynthetic
enzymes and the genetic complementation data of Hartman et al.
(23, 24). These authors showed that the transaminase gene (C
mutants) has one subunit, the dehydrogenase gene (D mutants)
has two subunits, and the pyrophosphorylase gene (G mutants)
has one subunit. These three enzymes have been found to have
molecular weights of roughly68,000,75,000, and 170,000. Thus,
there does not appear to be a correlation between the number of
subunits in a gene and the molecular weight of the corresponding
However, any conclusions based on the molecular
weight of an enzyme could be in error if the enzyme is made up
of a complex of several monomer units, as is the case for glutamic
dehydrogenase (25). Further investigations of genetic map
length and of molecular weight are being undertaken and will be
necessary before any conclusions concerning a correlation be-
tween map length and enzyme size can be made.
We were also interested in trying to determine whether or not
the histidine biosynthetic enzymes are associated intracellularly
in some sort of functional aggregate. In experiments using su-
crose gradient centrifugation, no evidence was found for such an
aggregate, even though S. typhimurium extracts were prepared
in a variety of ways.
of Hogeboom and Kuff is also applicable to the determination of
enzymes in protein mixtures and has been used by Levintoe,
Meister, Hogeboom, and Kuff (J. Am. Chem. Sot., 77,5304,1955).
The advantages of zone over boundary determinations have
been discussed.
YPHANTIS, D. A., AND WAUGH, D. F., J. Phys. Chem., 60, 630
BUTLER AND B. KATZ (Editors), Progress in biphoysics and
biophysical chemistry, Vol. 9, Pergamon Press, New York,
1959, p. 325.
BFXTTEN, R. J., AND ROBERTS, R. B., Science, 131, 32 (1960).
Microbial., 22, 369 (1960).
PORATH, J., Biochim. et Biophys. Acta, 39, 193 (1960).
PETERSON, E. A., AND SOBER, H. A., J. Am. Chem. Sot., 73.
751 (1956).
AMES, B. N., AND GARRY. B.. Proc. Natl. Acad. Sci. U. S..
46, i453 (1959). ’
R. J.. J. Biol. Chem.. 193. 265 (1951).
HOGEBOOM, G. H., AND KUFF, E. L., J. Biol. Chem., 210, 733
1. Sucrose gradient centrifugation with the swinging bucket
rotor of the preparative ultracentrifuge has been used to investi-
gate the sedimentation behavior of macromolecules of relatively
low molecular weight.
WEIGLE, J., MESELSON, M., AND PAIGEN, K., J. Molec. Biol.,
1,379 (1959).
SZYBALSKI, W., Experientia, 16, 164 (1960).
SCHACHMAN, H. K., lJltracentri.fugation in biochemistry, Aca-
2. The theoretical characterization of this system is outlined.
3. The applicability of this technique to the determination of
sedimentation constants of enzymes in multicomponent solutions
is demonstrated.
demic Press, Inc.’ New York,.19i9.
EDSALL. J. T.. in H. NEURATH AND K. BAILEY (Editors). The
Proteins, Vdl. I, Part B, Academic Press, Inc., New ‘York,
SUMNER, J. B., AND GRALI~N, N., J. Biol. Chem., 126,33 (1938).
THEORELL, H., AND BONNICHSEN, R., Acta Chem. Stand.. 6,
1105 (1951).
HAYES. J. E.. JR.. AND VELICK. S. F.. J. Biol. Chem.. 207. 225
4. Sedimentation coefficients of well characterized macromole-
cules were obtained and compared with values determined by
optical techniques. Good agreement was observed.
5. The general applicability of this system to the determina-
tion of sedimentation constants and enzyme purification is dis-
(1954). ’ ’
I , , ,
Chem., 212,439 (1955).
WETTER, L. R.. AND DEUTSCH, H. F., J. Biol. Chem., 192, 237
Chem., 167, 43 (1945).
6. Sedimentation coefficients were determined for three histi-
dine biosynthetic enzymes.
21. ABRAHAM, E. P., Biochem. J., 33, 622 (1939).
22. AMES, B. N., MARTIN, R. G. AND GARRY, B., J. BioZ. Chem.,
in press.
7. Some observations on the correlation of genetic information
with enzyme molecular weight are discussed.
23. HARTMAN, P. E., LOPER, J. C. AND HERMAN, D., J. Gen. Micro-
biol., 22, 323 (1960).
Microbial., 22, 354 (1960).
Addendum-We wish to emphasize that the boundary method
25. FRIEDEN, C., J. BioZ. Chem., 234, 899 (1959).
by guest on February 21, 2014http://www.jbc.org/Downloaded from