Molecular weights of coliphages and coliphage DNA

J. Mol. Biol. (1970) 54, 537-546

Molecular Weights of Coliphages and Coliphage DNA I. Measurement of the Molecular Weight of Bacteriophage T7 by High-speed Equilibrium Centrifugation FRANK C. BANcBoFrt

AND DAVID FREIFELDER

Graduate Department of Biochemistry Brandeis University, Walt?uzm, Mass. 02154, U.S.A. (Received 23 March 1970, and in revised form 28 July 1970) The range of usefulness of the high-speed equilibrium centrifugation method of Yphantis (1964) has been extended to measure the molecular weight of E&erichia co& phage T7. Values for 41of phages T7, TS and T4 were obtained by pycnometry; the phage concentrations were determined by measuring nitrogen and phosphorus contents, and the amino-acid and nucleotide composition. The fi values are: T7,0639; T&O*658 and T4,0*618. Born thenitrogenandphosphorus measurements, the percentage DNA in each phage type has been calculated. These values are: T7, 51.2; T5, 61.7 and T4, 54.9. The molecular weights of T7 phage and T7 DNA are 49.4 and 25.3 million daltona, respectively. In the following paper (Dubin, Benedek, Bancroft & Freifelder, 1970), values are reported of the molecular weights of T7, T5 and T4, determined by another technique. The values yielded by the two techniques for the molecular weight of T7 are in excellent agreement. Discussion of the general problem of molecular weight measurement is given in Freifelder (1970)$.

1. Introduction For a variety of purposes it is important to have available precise values for the molecular weights of different DNA molecules. However, although the techniques for measuring molecular weights of most macromolecules have reached a high degree of perfection, this has not been the case for DNA (Josse ik Eigner, 1966). The so-called absolute methods, such as light-scattering and sedimentation-diffusion, have so far been inadequate for the giant DNA molecules being isolated today. In the case of light-scattering the theoretical framework fails to encompass the complexities of multiple internal scattering; furthermore, most instruments are incapable of being operated at the required small angles. In the case of sedimentation-diffusion, it has not been possible to measure accurately the extremely small diffusion coefficients of large DNA molecules because of the very long time required to perform the measurements. Furthermore, the latter method requires that the partial specific volume, 8, be determined. This is extremely difficult for very large DNA molecules because solutions having the high concentrations necessary for the density determination are gel-like.

As a result of these problems several other methods have been tried. Probably the most important procedure is the electron microscopic determination of molecular t Present address: Department of Biological Sciences, Columbia University, New York, N.Y. 10027, U.S.A. $ A preliminary report of these findings has been made (Bancroft & Freifelder, 1969). 537

538

F. C. BANCROFT

AND

D. FREIFELDER

length (Kleinschmidt & Zahn, 1959; Lang, Kleinschmidt & Zahn, 1964). This method, when well controlled, is probably capable of yielding fairly good values, but is at present deficient for two reasons. First, it has not been possible until this series of papers to determine rigorously the mass per unit length of the molecule on the grid, so that one usually assumes that of the B form determined by X-ray diffraction (Inman, 1967; Lang, Bujard, Wolff & Russell, 1967) ; second, length varies with ionic strength. The limits of validity of the method have been discussed at length by Lang et al. (1967). A second method which has been tried is to measure the number of 32P atoms per DNA molecule of known specific activity by molecular autoradiography (Rubenstein, Thomas & Hershey, 1961). This technique, which for bacteriophage DNA gives values in good agreement with electron micrographic measurements (Thomas & MaoHattie, 1967), relies heavily on the assumption that the film records each decay with lOOo/oefficiency and on a precise determination of specific activity. This latter measurement is often difficult because of the spontaneous formation of polyphosphate in carrier-free 32P0, and the adsorption of 32P0, to glass during the hydrolysis used to destroy the polyphosphate. We do not mean to imply that the electron micrographic and autoradiographic methods are without value; in fact, in general they give fair agreement. The basic criticism is that there is no real criterion for assessing the reliability of the values so obtained. Another method, terminal labeling with polynucleotide kinase (Richardson, 1966), relies on the untestable (but likely) assumption that terminal labeling is 100% efficient. A recent detailed discussion of the difficulties with existing techniques may be found in Schmid & Hearst (1969). Some years ago Davison & Freifelder (1962) pointed out that most of the problems associated with light-scattering and the hydrodynamic methods are a result not of the high molecular weight of the DNA but rather of the great length of the molecule. For this reason they felt that a more reasonable approach would be to determine the molecular weight of a bacteriophage DNA by measuring the molecular weight of the phage particle, and then determining the percentage of this weight that is DNA. Working with phages has an advantage over working with DNA itself in that the phages are easily purified. They are rarely more than 1000 A in their longest dimension (in contrast with the lOO- to IOOO-fold greater length of the DNA’s), and they do not gel at high concentration. Davison & Freifelder (1962) attempted to measure the molecular weight of Escherichia coli phage T7 using both light-scattering and sedimentation-diffusion, but found both that it was difficult to render the phage suspension dust-free, because of both the high sedimentation coefficient of the phage and its tendency to adsorb to membrane filters, and that the diffusion coefficient was too small to be measured precisely by the methods then available. In recent years two techniques have become available that make it feasible to attempt once again to measure DNA molecular weights by determining phage molecular weights, The first is the high-speed equilibrium centrifugation technique of Yphantis (1964) which appears to be useful in measuring molecular weights of large macromolecules. Thus, Yphantis, using a rotor speed of 2378 rev./min, close to the lower limit then available, determined a value of 6.5 x lo6 for the molecular weight of southern bean mosaic virus. We show in this paper that high-speed equilibrium centrifugation experiments can be performed using the electronic speed control at its lower limit of 800 rev./min to measure reproducibly a molecular weight for T7 phage of 49 x 10s.

MOLECULAR

WEIGHT

OF

PHAGE

T7

549

The second new procedure is the 1-r beat frequency spectroscopic method of Dubin, Lunacek & Benedek (1967) for determining diffusion coefficients. In the second paper of this series the use of this method to measure phage molecular weights by sedimentation-diffusion is reported (Dubin et al., 1970). The determination of molecular weights by either of these two methods requires accurate values for the partial specific volumes (G)of the phages. In order to determine B, the density increment of the solution due to the presence of the solute and the solute concentration must be measured. Determination of macromolecular solute concentrations by dry weight measurements performed according to standard techniques is complicated by difficulties encountered in totally removing bound water without some decomposition of the macromolecule (Hunter, 1966,1967). However, phage concentrations can be determined in a tedious but simple and precise method as follows. First, the nitrogen and phosphorus contents of the phage suspension are measured by standard chemical tests. From the nucleoticle composition of the DNA and the P content of the phage, one obtains both the DNA content of the phage suspension and the N content due to DNA alone. Subtracting the N content of the DNA from the total N content, one obtains the protein N content. From a complete amino acid analysis of purified phage protein one can calculate the o/oN in the protein and therefore the protein content of the suspension. Addition of the protein and DNA contents gives the concentration of phage in the sample. It should be noted that from the DNA content and the phage concentration, the percentage of the phage mass which is DNA can also be calculated ; thus DNA molecular weights can be measured accurately from the phage molecular weights. Another method to appear recently for direct determination of DNA molecular weight is an improvement of the CsCl density-gradient equilibrium centrifugation technique (first used by Meselson, Stahl & Vinograd, 195’7) recently described by Schmid & Hearst (1969). They showed that by use of a double-beam optical system to determine the true baseline, and by accounting in the theory for concentration dependence, meaningful values for DNA molecular weights in the 20 to 100 million range could be attained. However, the values they obtained for the DNA’s of coliphages T4, T5 and T7 are significantly lower (about 20%) than currently accepted numbers, and for this reason there has probably been some skepticism concerning their conclusions. In this and the following paper it will be seen that our results agree well with those of Schmid & Hearst. In this and the following papers we report the molecular weights of the phagc particle and of the DNA of E. coli phages T4, T5 and T7.

2. Materials and Methods (a) Preparation

of purijed

phages

Phages were propagated as described elsewhere (Davison & Freifelder, 1962) and purified by four cycles of high- and low-speed centrifugation and two cycles of isopynic centrifugation in CsCl containing 0.01 M-M~& (and lo-” M-CaCls in the case of phage T5). Phages were then dialyzed four times against 0.01 M-‘Iris (pH 7*8), 0.1 rd-NaCI, 0.01 MMgCl,, and, for T5, low3 M-CaCl,. The final dialysate was always saved for use ss spectrophotometric blanks and as solvent blanks for 0 determination and equilibrium centrifugation. For centrifugation equilibrium experiments, pancreatic DNase was always added to the phage at 6 &ml. to degrade any free DNA in the phage preparations.

640

F.

C. BANCROFT

AND

D.

FREIFELDER

(b) Awaino acid cowuposition of purified

phage protein

This was done using a Beckman (for T4) or Technicon (for T5 and T7) ammo acid analyzer. Tryptophan was not determined. Phage protein was purified as follows. For T4 the phage were lysed by osmotic shock, a small quantity of pancreatic DNaae was added and phage ghosts collected by centrifugation. For T5 and T7, the phage DNA was released by boiling for 5 min in 7 M-guanidinium chloride. Solid CsCl was added to make a density of about 1.5 and the sample was centrifuged for 16 hr at 30000 rev./mm in a Spinco SW65 rotor. Phage protein was collected as a thin film on the surface of the CsCl solution. Since there was no DNA in the preparations, corrections for hydrolytic products of nucleotides were unnecessary. Nitrogen content was calculated from the amino acid composition.

(c) Nitrogen content of DNA Nitrogen content was calculated of the DNA%. For T4 the glucose

from the literature values of the nucleotide composition content of the DNA was taken into account.

(d) Nitrogen A standard Werby Labs.,

content of plunge

micro-Kjeldahl procedure was used in N determinations Boston, and for T5 by Dr Peter Davison.

for T4 and T7 by

(e) Phosphorus content of phcqes A modification of the procedure of Martland & Robinson (1926) was used. One ml. of a phage suspension was placed in a HNO,-washed Pyrex tube, 1 ml. of concentrated HaSOl was added, and the mixture heated at 150 to 160°C in a sand bath for approx. 16 hr until fuming stopped. The sample was decolorized by adding six drops of fresh 30% HaOe while still hot and continuing to heat at 150 to 160°C for several hours to destroy the HzOz. Then the sample was cooled and 10 ml. glass-distilled water was added. After mixing, 1 ml. of freshly made 10% ammonium molybdate was added. After mixing, 1 ml. of O.5o/o (w/v) hydroquinone and 1 ml. loo/” (w/v) Na,S04 were added. The blue color developed slowly and continuously for more than 48 hr at 25°C in the dark. Readings were usually taken after 20 to 24 hr. The optical density (o.D.) was determined at 650 nm and the P content obtained from a standard curve for KHaPO+ The authors are indebted to David Farb, Brandeis University, for developing this procedure. (f) Detemzination

of phuge concentrations

for

ii

detemindiorw

The meesured P content of the phage and the percentage P determined from the nucleotide composition were used to determine the DNA content of the solution. From the measured total N content of the phage suspension and the N content of the DNA, the N content of the protein was obtained by difference. The amino acid composition and the percentage N of the protein were used to calculate the protein content of the solution. By adding protein content and DNA content (Na salt), the phage concentration was obtained. Optical density measurements of the original solution yielded extinction coefficients for each. It should be noted that the use of this method of measuring phage concentrations should, according to the theory of Casassa & Eisenberg (1964), yield correct values of the anhydrous molecular weights of the phage particles (and thus also of NaDNA), regardless of any preferential interactions with components of the solvent.

(g) Determination Apparent

specific

volumes

(4) were calculated

of 8

using the equation:

where p. is the solvent density, ps is the solution density, and c is the solute concentration (g/ml .). Phage suspensions were dialyzed for 3 days against four changes of buffer. Densities of outer dialysates and phage suspensions were each determined at 26.0% by three separate weighings to +0-l mg in a 9=2-ml. calibrated pycnometer. The 0.D.260 of the

MOLECULAR

WEIGHT

OF

PHAGE

T7

541

suspension was then measured, and the dry weight was determined as described above. It is assumed that the values of $ obtained at a single concentration are equal within experimental error to 8. Values of B obtained at 26°C were corrected to the temperature of the equilibrium centrifugation experiments (approx. 16°C) using & value of dfi/dt = 0.00037 (Hunter, 1966). (h) Arudyticd

ultmctmtrifugation

Analytical ultracentrifugation was carried out in a Spinco model E ultracentrifugc, equipped with interference optics and electronic speed control. The camera lens was focused at the 0.59 point in the cell, which contained sapphire windows. Equilibrium ultraoentrifugation WEWcarried out according to the method developed by Yphantis (1964), using 2*4-mm liquid column heights, &cell containing a six-channel, 12-mm Durelumin-filled epoxy centerpiece (Yphantis, 1964), and s, heavy An-J rotor to increase stability at the extremely low rotor speed required in these experiments. The solutions were centrifuged to equilibrium at a nominal rotor speed of 800 rev./min. This corresponds for bacteriophage T7 to a value for the effective reduced molecular weight, 0 =waM (1 - @)/RT (Yphantis, 1964) of 5.3 cmm2, compared to the value of 5 cmv2 recommended by Yphantis. The actual rotor speed was determined from periodic readings of the tachometer and was observed to be constant during tho course of an experiment, and reproducible for a given drive. For the drive used in these experiments, a value of 807 rev./min was measured. The RTIC was not used to control temperature during the equilibrium experiments, in order to avoid the possibility that intermittent heating by the RTIC unit might set up steady-state convection currents. Instead, the refrigeration unit was set for 55 to 60”F, and the chamber temperature allowed to reach equilibrium during the course of the experiment. The rotor temperature was measured at the end of the run and was invariably in the range of 15 to 17°C. An equilibrium time of 80 hr was eatimated from the value of 2.3 (b--a)/w2 (Yphantis, 1964). The experiments were continued for at least 120 hr, and the observed distributions. were ascertained to be independent of time over the last 24 hr of each experiment. At the end of the run, the rotor was removed from the chamber, shaken to destroy the solute concentration gradient, returned to the chamber, and accelerated to 800 rev./min. Pictures were taken for the baseline correction. This procedure was considered preferable to the use of a water blank, since it did not require disassembling the cell, or even removing it from the rotor. No redistribution of the solute was observed during the time interval involved in taking the baaeline pictures. One compartment of the Yphantis cell which had been loaded at a high concentration was used to check the condition of the solute at the end of the experiment. After the baseline pictures had been taken, the rotor was accelerated quickly to a higher speed, and the resulting sedimentation pattern of the solute examined by schlieren and fringe optics. A single sedimenting boundary was observed for T7 in these experiments. The fringe patterns were measured with a Nikon two-co-ordinate microcomparator. The plates were read by measuring the y positions of two fringes at intervals of 50 p on the z axis. The value of y0 corresponding to zero solute concentration was determined for each fringe by averaging baseline-corrected values of multiple measurements of y taken near the meniscus. These values were then used to calculate the true fringe displacement y at each value of 2. Values of M,(r) were calculated from the least squares straight line through five adjacent data points in a plot of In Y versus r2, according to the equationt :

Kv(f9 =

2 RT

d(ln

Y)

U?(l-fip)d(T2).

t Abbreviations used: M,(r), weight-average molecular weight at the radius r; M$ M,(r) extrapolated to zero concentration; Mm(r), number-average molecular weight at the radius r; MO,(r), M,(Y) extrapolated to zero concentration; B,, the value of M, calculated according to equation (2) from a least squares fit to a plot of In c versus pa for data obtained over the whole cell sector; R, the universal gas con&ant; T, the absolute temperature; w, the angular velocity; B, the partial specific volume of the solute; p, the density of the solvent; a, t,he meniscus; g, t’hr acceleration due to gravity.

F. C. BANCROFT

542

AND

D. FREIFELDER

Values of i@w were determined from a least squares 6t to equation (2) of all of the data recorded for a cell sector. Values of i&,(r) were calculated according to the equation:

i&(r) = 2 RTY(r)/wa (1 -0~) j”:::Y(u)d(“a) (Yphantis, 1964). The solvent density, determined pycometrioally at 26”C, was 1.0022. At the low rotor speed (800 rev./min) used in these experiments the average radial acceleration is only 42 g. Since there is an additional acceleration of 1 8 along the z axis (parallel to the light beam), the net force on the solute will not be quite parallel to the z (or T) axis. The effect of this phenomenon on the fringe pattern is difficult to calculate since the magnitude of the force in the r direction increases as r2, while that in the z direction is independent of z. However, since the acceleration along the z axis causes particles to be displaced in a direction parallel to the light beam which traverses the cell, to a first approximation at least this acceleration should not change the observed fringe pattern. We have assumed that, since the acceleration along the .z axis is small compared to that in the radial direction, secondary effects of the former upon the fringe pattern obtained may be neglected.

3. Results and Discussion (a) Determination of v for phuges T4, T5 and T7 The partial specific volume, 8, for each phage was determined as described in Materials and Methods. The measured amino acid compositions of the purified phage ghosts and the calculated percentage N are given in Table 1. The remaining data required to determine phage concentration are presented in Table 2, in which the dry weight of a l-ml. suspension of O.D.ss,, = 1 (uncorrected for scattering) is calculated. In Table 3 the raw data for the pyonometric measurements are given, and the Gvalues are calculated. It should be noticed that there are two determinations for T7. These used different phage lysates separately purified. The two final values of C are within 1.1% TABLE

1

Amino acid composition and nitrogen content of phage T4, T5 and T7 protein

T4

Aspartio acid TllBOIdIl0

SC3rillt3 Glutamio Proline Glyoine Alanine

acid

Valine cysteine Methionine Isoleuoine Leuoine Tyrosine Phenylalanine Lysine Histidine Arginine Tryptophsn w&s not determined 0/0 Nitrogen

Moles/100 residues T5

T7

11.87 7.46 5.37 13.58 4.76 IO.40 12.82 7.3 none 21.7 2.76 6.06 4.37 4.26 7.95 l-46 5.34

9.2 6.8 8.4 11.6 2.6 11.3 9.85 7.2 none 2.4 4-4 8.15 2.15 3.3 7.0 1.6 4.2

9.26 5.83 9-37 lo-32 3.81 15.43 11.40 5.53 0.4 4.8 3.0 7.50 2.25 1.84 4.24 1.89 3.14

15.52

15.6

14.97

MOLECULAR

WEIGHT TABLE

Determination

of phage concentration ,@/O.D.

unit?

T7

pg total N/o.D. T4 T5 T7

PHAGE

@DNA, S&/O.D.

9.23 10.03 10.03 unit

Na unitC

5.87&o+% 6.77+2.3% 6.32&0.80/,

pg protein N/o.D. unitd

pg protein/o.o. unit?

5.24*0+6% 4.26& 1.5% 5.66&0.6X

33*77*0*6x 26.92&1.5% 37.81 &O*S%

% DNA

74.85&0*5% 70.22* 1.5% 77.52-&0.5x

54.9& 1.0% 61*66&2.7% 51.2&1.0x

a 1 O.D. unit = 1 ml. suspension having O.D. a00 -- 1. b Calculated from nucleotide composition-T4, 34% G+ C; T5, 39% c Calculated from nuoleotide composition and pg P/o.D. unit. d Total N - DNA N = protein N. e Caloulated from protein N and 0/0 N of Table 1. f Total weight = pg DNA, Na salt + pg protein. f y0 = standard error of the mean. TABLE of partial

PgDNA N/o.D. unit?

4l~OS~O~S% 43.30&2.3% 39.71&0.8%

Total weightjo.~. unit’

Determination

543

of a suspension with O.D. = 1 at 260 nm

11.11-&0.4y0 11*03*0~4% ll*98*o*4y0

T4 T5 T7

T7

2

%P in DNA”

3.57&0.&J% 4.056*2*3% 3.72&0.8x

T4 T5

OF

G + C; T7, 52% G + C.

3 apeci$c

volumet

Phage T7 b-4 T7 (b)

3*57f 1.4% 1*90*2*6%

T5

4*30*o-9yo

T4

l&?&2*2%

lO.OO&O*S% 5.23-&0*6x 12*63* 1.5% 4.86&0.6%

T7 (a) and (b) refer to two independent determinations t See equation (1) for explanation of symbols. $ p. (26.O’C) = 1.0023 g/ml.

from different

0.642&O-8% 0.635If 1.1% O*658&O-9~o 0*618&1*1% phage samples,

of each other, which is consistent with estimated errors. In molecular weight calculawe use the average of the two, i.e. 0.639. One of the difficulties in evaluating measurements of C is that there are no criteria for assessing the reliability of the values obtained, other than reproducibility. We mention this because the value of 6 for TS is higher than we had expected. It is commonly believed that in a complex substance, 5 is equal to the weight-average value of the 6’s of its components. (This assumption is however not based upon first principles, and there are few, if any, situations in which this has been clearly demonstrated experimentally.) Hence one expects that Gfor T7 should be around O-62, since tions

544

F.

C. BANCROFT I

I

AND

D.

I

FREIFELDER I

,

1

I -

-

I.0

- 0.8

_ o’6 2. F -0.4 ‘: .$ z 5c 6 -02 200 -

- 0.1 IOO-

I 6.48

I I 6.52 Radius (cm)

I 6.56

I 6.60

Fra. 1. T7 at sedimentationequilibriumat 807 rev./min, 15*S°C, in a buffer containing 0.1 M-NaCl, 1O-3 M-Mgcl,, 10-s M-Tris (pH 7.4) and 5 pg DNasejml. Initial phage concentration was 0.44 mg/ml. The net fringe displacements (Y) on a logarithmic scale (left-hand ordinate) are plotted as a function of the radius. The error bars correspond to a reading error of f5 ~1.Concentrations in units of mg/ml. (right-hand ordinate) are based on 280 p per fringe, four fringes per mg/ml. The inset shows a linear plot of values of Y recorded near the meniscus as a function of the radius.

the values of Gfor protein and DNA are normally approx. 0.70 to 0.73 and 0.55 to O-57, respectively. Since the DNA content of T5 is higher than that of T7, it might be expected that fi for T5 would be lower than that of T7. This is however not observed. Since the estimated errors in all of the parts of the T?measurement for T5 are low, we believe that this result implies that d cannot always be predicted simply by addition of values of the 5 of components. (b) Measurement of phuge T7 molecular weight by the Yphuntis procedure The result of a typical experiment on T7, at an initial concentration of 0.44 mg/ml., is shown in Figure 1. The fringe displacements near the meniscus are shown in the inset, where it is seen that for a distance of about O-06 cm near the meniscus the observed fringe displacement is constant to within a reading error of f5 CL.This justifies the procedure described under Materials and Methods for calculating absolute concentrations in units of net fringe displacement elsewhere in the cell. In order to examine the behavior of the data across the cell, and to determine whether T7 exhibits concentration dependence, values were calculated for the quantities M,(r) and M,(r), as shown in Figure 2. It is seen that these quantities show little scatter across the cell, and furthermore that they vary only slightly with concentration. (It should be pointed out that the apparent lack of dependence of molecular weight-averages on concentration might arise from a fortuitous combination of aggregation and non-ideality.) The values of the quantities M,O, MO, and 28, have

MOLECULAR

WEIGHT

OF

Concentration

0

0.7 --

I

60 -

PHAG’E

T7

546

(mg/ml.)

n,4

0.6

0.8

I

I

I

I

501

IO

I

0

40-

30-

20-

10 I 600

I 400

I 200

0

I

I

800

1000

I

Y(p)

FIG. 2. Moleaular weight averages as a function of the concentration. Number-average molecular weights were calculated as described in (-O--O--) end weight-avemge (-O-O-) Materiels and Methods using the data of Fig. 1. The lines represent least-squares fits to the data. Concentrations in mg/ml. (upper abscissa) were calculated as in Fig. 1.

TABLE 4 Molecular

weight averages obtained from sedimentation equilibium experimsnts

Experiment

1 2 3 Mean f 4

Molecular

Phage

M::

T7 T7 T7

50.3 50.0 47.6 49*3f 49.6

8.~.

T3t

t Values of molecular

weights

1.5

calculated

weights WV 52.0 47.8 60.0 49.9f2.1 49.3

( x 10-s) J&v 49.8 47.3 48.9 48.7& 1.3 49.0

assuming 5 (T3) = 5 (T7).

been evaluated in three separate experiments, as shown in Table 4, and are quite similar within an individual experiment, again indicating that there is little concentration dependence. In addition, the reproducibility of the molecular weight determinations should be noted.

This work was supported by grant no. GM14358 from the U.S. Public Health Service, grant no. E509 from the American Cancer Society, and contract no. (AT-30.1)3797 from the Atomia Energy Commission. One of us (F.C.B.) was supported by postdoctoral fellowship no. IF2GM21176 from the U.S. Public Health Service. The other author (D.F.) is supported by a Career Development Award (no. GM7617) from the U.S. Public Health Service. 36

546

F. C. BANCROFT

AND

D. FREIFELDER

We would like to thank Peter Davison, Helen Van Vunakis and David Farb for performing many of the chemical analyses, and we wish to thank John Hearst for the results of his measurements prior to publication. This is publication no. 742 from the Graduate Department of Biochemistry, Brandeis University. REFERENCES Bancroft, P. C. Br.Freifelder, D. (1969). Fed. Proc. 28, 834. Casassa, E. F. & Eisenberg, H. (1964). Adwanc. Protein. Chem. 19, 315. Davison, P. F. & Freifelder, D. (1962). J. Mol. Biol. 5, 635. Dubin, S. B., Benedek, G. B., Bancroft, F. C. & Freifelder, D. (1970). J. Mol. Bio2.54,547. Dubin, S. B., Lunacek, J. H. & Benedek, G. B. (1967). Proc. Nut. Acad. Sci., Wash. 57, 1164. Freifelder, D. (1970). J. Mol. Biol. 54, 567. Hunter, E. (1966). J. Phys. Chem. 79, 3205. Hunter, E. (1967). J. Phys. Chem. 70, 3288. Inman, R. B. (1967). J. Mol. BioZ. 25, 209. Josse, J. & Eigner, J. (1966). Ann. Rev. Biochem. 35, 799. Kleinschmidt, A. K. & Zahn, R. K. (1959). 2. Nadurf. lBb, 730. Lang, D., Bujard, H., Wolff, B. & Russell, D. (1967). J. Mol. BioZ. 23, 163. Lang, D., Kleinschmidt, A. K. & Zahn, R. K. (1964). Biochim. biophye. Actu, 88, 142. Martland, M. & Robinson, R. (1926). B&hem. J. 20, 847. Me&son, M., Stahl, F. W. & Vinograd, J. I. (1957). Proc. Nat. Acd Sci., Wash. 43, 581. Richardson, C. C. (1966). J. MoZ. BioZ. 15, 49. Rubinstein, I., Thomas, C. A. Jr., & Hershey, A. D. (1961). Proc. Nwt. Acad. Sci., Wash. 47, 1113. S&mid, C. W. & Hearst, J. E. (1969). J. Mol. BioZ. 44, 143. Thomas, C. A., Jr. & MaoHattie, L. A. (1967). Ann. Rev. Biochem. 36, 485. Yphantis, D. A. (1964). Biochemistry, 3, 297.