J. &fol. Biol. (1977)
114, 507-526
Determination of the Nuclear DNA Content of Succharomyces cerevisiae and Implications for the Organization of DNA in Yeast Chromosomes GAIL II. LATJER-~,THOMAS M. ROBERTS AKD LYNN C. KLOTZ Department Harvard
of Biochemistryy and dfoleculnr Biology Cambridge, Mrxss, 02138. U.S.A.
Univwsity,
(Received 28 September
1976, and in revisal
*form !) April
1977)
The DNA content of tho nucleus of the yeast Saccharomyces cerevisiae has been &t,ermined by both renaturation kin&es and DN.4 per cell measurements. Kcnaturation kinetics experiments acre performed by following the decrease of optical hyperchromicity at 260 nm and by hydroxyapatite chromatography. DNA per cell measurements were made by the diaminobenzoic a,cid method and by the ethidium bromide method of Klotz & Zimm (1972b). The conclusion from the above experiment,s is that the 8. cerevisiae nucleus contains 9 x log& 2 x 10g daltons of DNA. Pre\iously we (Lauer & Klot,z, 1975) had measured the molecular weight of tllc largest piece of DNA iu tlw yeast nucleus to be 2 x lOghO. x log. Here we oxtend t,lris work b>r using a more highly proteindenaturing buffer system and conclude that tile largest) piece of DNA in the S. cerevisiae nucleus contains 1.5~ 10g to 2.2 x IO9 daltons of DNA in both llaploid and diploid cell lysates. From genet,ics, the largest yeast chromosome sl~ould contain 13% of the grnome, or 0.9x 109 to 1.5 x 10” daltons of DNA (using our DNA per cell range). Thus, the large DNA WC measure cont,ains from on,3 to two times t,hc: a.mount of the DNA predicted from genetics to be ill the lnrpcst chro1nosome. In liglit of tkese now data, viscoelastic mf:aslwements on yeast DNA are now consistent with the idea that wrh chromosome contains one piocc of DNA.
1. Introduction The number of pieces of DNA per chromosome in eukargotes is not well-establishctl Light micrographs (Wilson, 1909,1928 ; Takayama, 1975) and electronmicrographs (Dupraw, 1970; Costello, 1961; Hsu et a.Z., 1967) showing fine connectors between experiments (Hoskins, 1968 ; Diacumakos chromosomes, and micromanipulation eI al., IYSl), in which chromosomes seem to be linked together by DNA, suggest that t.here may be more than one chromosome per piece of DNA. In contrast, sedimentation of mouse chromosomal DNA (Lange. 1975) and cyt,ological evidence fol switches in polarity of chromosomal DNA (Wolff ef nZ.. 1976) suggest that there may bc a number of pieces of DNA per chromosome. The type of relationship most compatible with the existence of genetic linkage groups and independent segregation of chromosomes is one piece of DNA per chromosome. Measurements of the molecular weight of Drosophdu chromosomal DNA (Kavenoff & Zimm, 1973) using the viscoelastic retardat’ion-t’ime technique (Klotz & Zimm? 1972a,h) have yielded this answer. t l’rcsent addrws: Rovenstiel Walthem, Mass., 02154, U.S.A.
Basic Medical
Sciencca &search
Center,
Brandeis
University,
508
U.
I). LAl:EH,
‘I’. 11. ROHE’RTS
:INI>
I,.
c’. KLOTZ
However, because of t)he very large size of this DNA, the possibility of DNA aggregat’ion has not been excluded. (See also Roberts et aZ., 1977.) From the above discussion, it is evident’ t,hat’ the question of how DSA is organized in eukaryotic chromosomes remains unanswered. The yeast Sa~cchurornyces cwevisin~ should be an ideal organism t,o use to answer this question because (If it,s small nuclear DNA content and well-studied gen&ics. WC have tried to approach t’his problem (Lauer & Klotz, 1975) by measuring t’hr molecular weight of the largest piece of DNA in the yeast nucleus, using the viscoelast,ic retardation-time technique (Klotz & Zimm. 1972aJ). If the nuclear DNA content is known. and if it is assumed that, the physical length of DKA in a chromosome is proportional to its genetic map length, the amount of DNA that should be contained in each chromosome can be calculated. Specifically, the amount of DNA in the largest chromosome can be calculated and compared to our measured value for the molecular weight, of the largest piece of DNA in the nucleus. However, values reported for the nuclear DNA content of S. cerevisiae vary widely. From renaturation kinetics experiments, Bicknell & Douglas (1970) found 9 x IO0 daltons per haploid nucleus, while Whitney 6 Hall (personal communication) found 5.3 x lo9 to 5.6 x lo9 daltons for the same quantity. DNA per cell measurements have yielded still more values: 1.2 x 10 lo to 1.3 x lOlo daltons per haploid yeast cell (Ogur et al, 1952; Schweizer & Halvorson, 1969; Ciferri et al., 1969) and 8.4~ lo9 daltons per haploid spore (Darland, 1969). Such a wide range of reported values for nuclear DNA content made interpretation of our molecular weight measurements extremely difficult. The problem is doubly complicated in that not only is the total range of values extremely wide for our purposes (5.3 x lo9 to 13 x lo0 daltons), but the two methods of measurement (renaturation kinetics and DNA per cell determinations) gave ranges of values wit,h a minimum of overlap. Additionally, the report’ed DNA per cell measurements were obtained using methods for which proper controls for positively interfering substances are not possible. We describe here experiments directed at resolving this problem. We have performed a very detailed renaturation kinetics study of S. cerevisiae DNA and have performed DNA per cell measurements by the diaminobenzoic acid method, which is believed to be remarkably specific for DNA, and by a technique utilizing DNA enhancement of ethidium bromide fluorescence, for which it is possible to do a DNAase control for positively interfering substances. WC have also extended our previous molecular weight measurements (Lauer $ Klotz, 1975). The results of the various experiments are discussed with respect to their implications for the organization of DNA in chromosomes.
2. Materials and Methods (a) Cells und culture
methods
The cells used were a self-diploidized S. cerevisiae strain derived from S288C and haploid sogregants from that diploid, Escherichia coli strain K12 C600, and Bacillus subtilis strain 3610. E. coli and B. subtilis cells were grown in media containing 8 g nutrient broth, 3 g brain heart infusion, and l-5 g yeast. oxtract/l (Difco) at 37°C with shaking. S. cerevisiae cells were grown in YEPD medium (1 g yeast extract, 2 g peptone, 2 g dextrose per 100 ml) at 30°C with shaking. For all experiments, except tllosc involving viscoelastic retardationtime measurements on logarithmic phase S. cerevisiae, ~11s were harvested in stationary phase. Cells were counted in a hemacytometer under phase contrast illumination at a magnification of 400 X or 1000 X .
CHROMOSOME
STRUCTURE (b) DNA
509
137 8. CEREVISI.4E isolation
E. coli and B. subtilis DNAs were isolated by the method of Marmur (1961). Wlxx~ necessary, cells were pretreated with lysozyme. S. cerezjisiae cells were lysed by the freeze thaw method of Smith & Halvorson (1967). DNA was then isolated by two methods. In the first method, DNA was isolated by the method of Marmur (1961) with additional RNAase and proteinase treatment,s and final purification accomplished by shearing the DNA and passing it over hydroxyapatite. This final step led to no loss of DNA. For the second yeast DNA isolation, the cell lysate was treated extonsively with RNAase and proteinase, and the DNA was then precipitated from the crude lysat,e with isopropanol. Substquent,ly, the DNA was exposed to the enzymes again, oxtract,ed with chloroform/ isoamyl alcohol (24 : l), and precipitated with isopropanol. This sequence was repeated several t,imes. In each case DNA was .redissolved in O-1 x salin+EDTA (0.15M-NaCI, 0.1 M-Na,-EDTA, pH 8) after the isopropanol procipitatiort step. Then 0.5 mg EthBrt and 1 g CsCl (optical grade) were added per g DNA solution, and t’he mixture was banded in tllc Becklnan L4 preparative ultraccntrifuge in a type 65 rotor at 35,000 revs/min for 60 11. The fluorescent DNA ba,nd was well-separated from the denser polysaccharide band, so t,hat t.he plu-r DNA band was easily collected. The EthBr was extracted from the DNA witl, isopropanol. Purity of DNApreparations wasmonitored by optical spectra (A,,,/A,,, = 2.1 t,o 2.2; A,,,/A,,, = 1.8 to 1.85) and hyperchromicities (3 2% to 400/,). For each DNA, hot,11 t.hr buoyant density in analytical CsCl equilibrium centrifugation and the melting t’emperat’urc agreed with previously reported values (Bak et al., 1969 ; Cornea et al., 1966 ; Moustaochi & Williamson, 1966). (c) Shearing of D6A Purified DNA samples in 0.12 M-sodium pllosphatjo buffer (pH 6.8) were sheared approximately 5-ml portions on ice using a Bronson I’ower Sonifier, model W185. (d) DNA
sizing
Molecular weights of single-stranded, alkaline sedimentation in an analytical the procedure of Studier (1965). (e) C&l
density-gradient
by sedimen,tation
velocity
sheared DNA Beckman model
samples were E ultraccnt’rifuge,
equilibriccm
dctermined according
in
by t,o
centrQ%gation
Analytical C&l equilibrium centrifugations were performed ill a Beckman model E ultracentrifuge at either 44,700 revsjmin for 20 to 24 h or 39,460 revs/min for 44 h. The buffer used was 0.01 M-Tris, 0.001 M-Na,-EDTA (pH 8) plus 1.26 g CsCl (optical grade) per g buffer (p := 1.700 g/cm3). Buoyant densities were calculated relative to a bact,erial DNA standard of p = 1.719 g/cm3 according to the procedure of Schildkraut et al. (1962). The areas of the main band and mitochondrial DNA peaks were mea,sured using a planimeter (Keuffel and Esser, model no. 620015) in order to calculate the proportion of mit’ocllondrial DNA in the isolated S. cerevisiae DNA. (f) Thermal
denaturations
and
renaturations
For all denaturations and renaturations, all DNAs were in 0.12 &r-sodium phosphate buffer (pH 6.8). Thermal denaturat’ions were conducted on a Gilford model 2000 recording spectrophotornet,er flllly automated for melting curv~~s. Both E. coli and B. subtilis DNA controls were rlln simultaneously in all melts and all t, values were calculated relative to an E. coli DNA t, value of 90°C’. For renaturations of DNA monitored by following tllca decrease in absorbance at 260 nm on the Gilford 2000. DNA samples were sealed in cuvett,es, denatured by heating at 100°C for 8 min in a thermostatically controlled block, then transferred to the sample chamber of the Gilford spectrophotometer (preheated to the renaturation temperature) and recording begun. Both E. co& DNA and B. subtilis DNA standards were run simultaneously in all renaturations followed optically. A, was taken as the AZ6,, of sheared native DNA measured at tho temperature of renaturation, while ‘4, was taken as the A,,, of the same DNA at 100°C. t Abbreviations used: EthBr, ethidium bromide; t,, mrlting temperature (“C) ; C,t, the product, of the initial DNA concentration in moles nuoleotides per liter and the time of renaturation in seconds. 32
511)
G.
1).
LdUEK,
‘I’.
IV.
KOHEH’l-S
.4X11
I,.
(:.
lCLO’l’%
For reriaturations monitored hy hydroxyapatite cllro~natograpl~~, t11t3 II~‘lr”“~‘tl’atitr was prepared according to tlit: proccdurc descrihcd tiy Hcriiardi (I 97 I ). ‘l’lic capacity. \vas greater than 500 pg DNA/ml hydroxyapatite. All DNA samples \verc in 0.12 nl-sotlium phosphate, pH 6.8 (0.18 AZ-NH+), and cacli sample contained 100 pg DNA. Samples wcrc sealed in vials, denat,ured in hoiling water for 5 mm, aiid t’hen incubated in a t,ticrinostat,ically controlled block for the desired timo at, thr: appropriate tempcraturc to yield an optimal renaturation rate for each particular DNA, i.e. about t, ~ 25 deg. C (65°C for E. coli DNA, 62°C for R. aubtilis DNA, and 60°C for S. cerevisiae DNA). Reassociated DNA samples were immediat~ely passed over 3ml xvater-jacket.ctl lrydroxyapatite columris at 60°C in 0.12 or-sodium phosphate (pH 6.8). Tlie hound fractions wcrc clutod with 0.5 ox-sodium phosphate (pH 6.8). Hyperchromicities of unbound and bourld fractions for each Cat point wero determined hy melting thcsc fractions and t,lic A,,, ratios at 100°C were used to calculate the percentage hound and unbound. The fraction zero-time binding was subtracted from both t)he fraction hound and t,lic t,otal amount of DNA in tlic calculation of pcrccntage DNA reassociated at each (yet value (Davidson et al., 1973). Data were computer fitted using Rritterr’s computerprogram for fittinghydroxyapatite data (Britten et al., 1974). (g) lIlYA per cell by ethidizrm bromide DNA per cell measurements by t’lle ethidium bromida method were performed on S. cerevisiae using a modification of the procedure of Klot)z & Zimm (1972b). Yeast cells were grown well into stationary phaso and spheroplasted (procedure of Lauer & Klotz, 1975). DNA per cell experiments were t,lien performed according to the procedure of Klotz & Zimm (1972b), with the omission of lysozyme from the lysis mixture. Incubation times were increased to 24 11for the lysates, 4 h for the R,NAast: incubation, and overnight for the DNAase incubation. EthBr fluorescence was read on a Perkin Elmer MPE’-44 fluorescence spectropliotomct~cr (excitation wavo1engt.h 540 rim, emission wavelength =z 590 nm). (h) /XVA per cell by diaminobenzoic acid DNA per cell measurements by diamiuobenzoic cells using a modification (Holm-Hanseri et al., Robins (1958). Diaminobcnzoic acid fluorescence taining appropriate filters to yield ail excitation emission peak wavelength = 505 nm. (i) DNA
molecular
acid were performed on S. cerevisiae 1968) of tlin procedure of Kissane & \\as mad on a filter fluorimeter conpeak vvavrlcngth = 405 mn and an
weight measurements
Molecular weight measurements on yeast DNA were made using the viscoelastic previously described technique of Klotz & Zimm (1972a,b) according to the procedures (Lauer & Klotz, 1975; Robert,s et al., 1977). Measurements were made on both haploid and diploid cells using the buffers described by Kavcnoff & Zimm (1973) and Roberts et al. ( 1977). Heat dcnaturatiou and shearing controls were performed as described previously (Lauer & Klotz, 1975; Roberts et al., 1977).
3. Results (a) Measurement
of mitochondrial
DNA
The proportion of mitochondrial DNA in the total cellular yeast DNA was determined from the CsCl equilibrium banding pattern of the S. cerevisiae DNA and the bacterial maker as measured by the ultraviolet-scanner optics at 265 nm in the Beckman model E ultracentrifuge. The S. cerevisiae DNA consisted of two peaks: the main-band nuclear DNA peak with p = 1.694 g/cm3 and the mitochondtial DNA peak with p = 1.6778 g/cm3 (both relative to p = l-7035 g/cm3 for E. coli DNA). The proportion of mitochondrial DNA in the total cell DNA was calculated from the areas of the peaks and found to be about 8% for both isolations of yeast DNA. Since all DNA renaturation kinetics experiments were performed on whole-cell DNA, the
CHROMOSOME
STRUCTVRE
IS
S.
C!ERET~ISldE
51 I
DXA content per nucleus is calculated from these experiments by multiplying the total DNA complexity obtained by 0.92. The DNA contents determined by DNA per cell experiments are also multiplied by 0.92 t’o obtain the nuclear DNA content from these measurements. (1)) DNA
renatwration
kinetics followed
by h.ydroxyapatite
chromatography
The data from the hydroxyapatite monitored rrnaturation of S. cerevisiae DNA are presented in Figure 1. Both E. coli and B. subtilis DNAs wzre renatured as standards, and these data are also in Figure 1. Results are present’ed as C,t plots in the manner of Britten & Kohne (1968). The data on each DNA were computer fitted using Britten’s program (Britten et al., 1974). Fitted values for the fraction of DNA in each component and its corresponding rat,e constant (k,) are listed in Table 1. Also listed are the computer-calculated errors in determining each paramet’er. The errors for Wing the bacterial DNAs and the unique part of the S. cerevisiav DNA are very low. The errors in determining the parnmeters for the repeated fractions of yeast, DNA are large because insufficient data ww collected in these regions of the C,t curve to accurately det,ermine the complexities and amounts of the repeated fractions. This does not matter for our work.
Equivalent
Cot
(mol
I-’ s)
Fro. 1. The kinetics of renaturation as measured by hydroxyapatitn chromat,ography for S. rerevisiae DNA (O), E. coli DNA (o), and I$. subtilis IINA (A). Rrassociations were carried out in 0.12 M-sodium phosphate (pH 6.8) at 60°C for S. cerevisifze DXA, 62°C for B. aubtik? DNA, and 66°C for E. coli DNA. Following incubation to the desired equivalent C,t value, samples were passed over hydroxyapatite columns at 6O’C. Single-stranded mat,erial was eluted with 0.12 M-sodium phosphate (pH 6.8) and double-stranded with 0.6 Msodium phosphate (pH 6.8). The fraction of DNA binding to hydroxyapatite as double strands has been corrected for zero time binding in the manner described by Davidson et al. (1973). DNA fragment lengths were 626 nucleotides for S. cerevisiae DNA, 444 nucleotides for B. subtilis DNA, and 523 nucleotides for E. coli DNA. The curves through the data points represent least-squares computer analyses of the data utilizing 1 kinetic component for each of the bacterial DNAs and 3 kinetic components for the S. ceTeviaiae DNA. Reassociation of the non-repetitive component of the S. cereviaiae DNA is indicated by the broken line, and reassociation of each of the repeated romponent,s of the S. cerevisicce DNA is indicated by a dotted line.
of renaturation
Parameters were reaction were 0.095 fits of the data are t Values are not mine these.
0.906 0.1022t 0.7755*0.025
0.015t 0.114 0.869
0.017
6.315t 0.09871+0.0088
6.964t
0.3187&0.0203
1.0
Observed rate constant ( 1 /s mole nucleotide) 0.3445&0.0236
of
by hydroxyapatite
1.0
Fraction total DNA
monitored
fragment length (base-pairs)
61.78 0.1273
626
444
0.3549
463.4
523
DEL4
0.3701
Pure rate constant (l/s mole nucleotide)
chromatography
obtained from computer least-squares analyses of the data presented in Fig. 1. Fractions of total optical density unbound at the end of for S. cerevisine DNB, 0.0462 for E. coli DNA, and 0.1081 for B. subtilis DSA4. The very low r.m.s. deviations indicate that the computer very close. given for the errors in these quantities because insufficient data w-err collcrted in the proper region of the C,f curve to accurately deter-
1st repeated component 2nd repeated component Unique component
s. cerevisiae
0.8982
0.904
B. subtilis
JJy0.0125
0~9307+0~0171
Fraction of total optlcal density
kinetics experiments
0.889
deviation for fitted curve (%)
R.m.s.
analysis
E. COG
DNA
Kinetic
TABLE 1
(‘HROMOSOME
STRUCTURE
IN
AS.
CEREiT'ISIAE
513
since we are interested only in determining the total nuclear DNA complexity. We can, however, calculate the total amount of repeated sequences: 87% of the total DE,4 is unique, leaving 13’/, t)o be represented by the two repeated components. Since 8:/, of the total DNA was found to be mitochondrial by analytical C&l equilibrium centrifugation, about 50/: of the total cellular DNA (5.4”; of t’he nuclear DNA) of the DNA should consist of nuclear repeated DNB. Th’ is is about’ the proportion which should code for ribosomal RNA (Schweizer et al., 1969). Complexities of yeast nuclear DRI’A were calculated from the rate constants using various reported values for bhe complexities of E. coli DXL4 (Gillis et al., 1970; Klotz & Zimm, 19726) and B. subtilis DNA (Gillis et al., 1970; Klotz & Zimm, 1972b; Wake. 1973; Rau, 1975) as standards. In each calculation, the total calculated DNA DNA to yield the nuclear contjent’ per yeast cell is corrected for 80/, mitochondrial DNA complexit,y. The calculated nuclear DNA complexities are listed in Table 2. The calculated 8. cerevisiae nuclear DNA complexity is 8.42 x log to 9.1 x log daltons when E. coli DNA is used as a standard and 7.01 x IO9 to 8.42 x 10g daltons when 17. xuhtilie DNA is used as a standard. (c) DNA
renaturation
kinetics followed b?poptical density
For each renaturation of 8. cerevisiae DNA monitored by following optical hyperchromicity at 260 nm, both E. coli DNA and B. subtilis DNA were renatured as standards simultaneously with the yeast DNA. The rationale for using these two -DNA standards of differing (G + C) content (5Oq; for E. coli DNA and 43% for B. subtdlis DNA) was that since S. cerevisiae DNA has a low content of (G + C) (409&), several DNA standards of differing percentage (G + C) might be necessary to interpret the data. Also, Bicknell & Douglas (1970) using E. coli DNA as a standard and Whitney &*Hall (personal communication) using B. suhtilis DNA as a standard obtained different complexities for X. cerevisiae DNA. Because of the differences in (G + C) contents and, therefore, in t, values among the DNAs being renatured, experiments were performed at nine different t)emperatures, ranging from 52OC to 76.7”C. Second-order rate constants were calculated from reciprocal second-order plots of l/fraction unrenatured versus time in seconds. Such a plot for a typical experiment on yeast DNA is presented in Figure 2. One obtains the rate constant (k,) of reaction for the unique portion of the DNA from the plot by drawing a line tangent to the nearly linear portion of the plot, which appears after the processes of intrastrand base-pairing and renaturation of repeated DNA have ceased. This line is designated A in Figure 2. k, for the reaction of t’he unique DNA is obtained by dividing the slope of lint A by the initial DNA concentration and then by t,he intercept to correct for the percentage of the Dn’A which is undergoing intrastrand base-pairing plus the percentage which is made up of repeated sequence elements (for details see Rau, 1975). R a\h k, values obtained in this manner are converted into units of l/s mol base-pair, and arc corrected to a standard length of 500 base-pairs assuming an L1j2 dependence of k, (Wetmur & Davidson, 1968). From the second-order rate constants, the complexity 8. wwr~isine can be calculated using the equation N
S. cereuisiae
--
(N) of the unique
k Nstsndard 2 standard k 2 S. cererisiue
,
DNA
of
(1)
(‘HROMOSOME 2.2.
STRUCTURE
5’. CERE
IN
I’ISIAE
5I5
I
I
I
I
I
I
I
I
I
9
I 5
I IO
I I5
1 20
I 25
I 30
I 35
I 40
I 45
50
2.0 -
I.0 0
Time (s x 103J FIG. 2. =\ t,ypical DNA ronaturat,ion kinetics experiment as measured at PBO nm fur DNA from 5’. cerewisirru. Thv data arc plotted as a reciprocal secontl-order plot of (A?” time (s), Tho renaturation, carried out in 0.12 M-sodium phosphate with an initial DNA concn of 1.34 x lo-“ mol nucleotidejl. Line A reaction of the unique DNA. B and B’ represent 2 pwsiblc lines relwatvtl I)NA.
by optical
hyperchromicit\.
~ *4$“)/(A:“” - A$‘) WYW(S (pH 6.8), was run at) 69.75”(’ represants the init,ial rate of through the reaction of tho
where the standard is either B. subtilis DNA or E. coli DNA. Since S. cer~visiu.r contains repeated DNA (in t’he form of mitochondrial DNA plusz possibly, some repeated nuclear DNA), the total cellular DNA content is calculated by dividing t,hc amount of unique DNA by the fraction of total 8. cerevisiae DNA that is unique. The fraction unique DNA is determined from the various intercepts A. B or B’ in Figure 2 as described by Lauer (1976). Finally, t’he nuclear DNA content is obt,ained by multiplying 0.92 (1.0 minus the fraction mitochondrial DNA in the cell) times the total DNA in the cell. The major problem in calculating DNA complexities from renaturation kinetics followed by optical density is that it is not possible to know where to draw the lines t,hrough the initial reactions of the various DNA fractions (lines A and B in Fig. 2) to obtain rate constants. This difficulty is due to the curvature inherent in reciprocal second-order plots of optical renaturations (Rau, 1975). It is especially difficult t,o kno\v u-here t,o draw line B, the line which goes through the reaction of the repeated DNA and the inverse of whose intercept is the fraction of intrastrand base-pairing. Thus. we have drawn B and B’: and the true line could be anywhere bebween these two. At ea,ch temperature, the calculations of percentage single-strand structure and nuclear DNA content are performed twice. once with each of the intercepts for the two lines B and B’, and the correct value for each calculation is probably between the two listed values. In t,he last) two columns of Table 3 are listed the complexities
0.
61 6
I).
LXUER,
‘I’. M. ROHER’I’S
Ah-I)
1,. C’. KLO’I’Z
of yeast nuclear DNA which one obtains using B. auhtilis DKA at, 2.10” tlaltcms (Klotz & Zimm, 197%: Wake, 1973) a11d E. wli DNA at 2.7 I, IO” daltons (Klotz & Zimm, 1972b) as standards and calculating t*he 5’. cPrevi.siae I)NL1 complexity at each temperature from the ratio of its E, \aluo to the k, value of c~~c:hbacterial DSA obtained at the same temperature. As seen in Table 3, the calculated yeast DNA complexities are anomalously large at temperatures above 70°C. One possible explanatfion for this is that, at, higher tcmperatures some (A + ‘I’)-rich componenb of yeast DNA is not renaturing. (Another possible explanation for this result is D1L’A degradation. However, this uxs ruled out by sizing DNA before and afkr renaturations.) For each DNA there is a plateau at’ t, -20 deg. C to 1, -25 deg. C. at which t’he rate of renaturation is maximal. If t,hese maximal k, values are used for the complexity calculations (and using various reported values for the complexibies of B. subtilis DNA and E. coli DNA as st,andards), one calculates the values listed in Table 4. Though k, values obtained at different temperatures are not directly comparable among themselves because they include the effects of temperature, hyperchromicity, viscosity of medium, percentage of (G + C), and possibly other factors TABLE Data from
renaturatiofa
kinetics
experiments hyperchromicity
Renaturation Organism
temperature (“C)
3
monitored
by following
the decrease of optical
at 260 nm
Initial DNA concn (mol nucleotitlc/l I; 103)
(j& Single-strand structure
B. subtilis
52 60.4 62.3 66 69 69.75 72 73.7 76.7
0.146 0.131 0.136 0.131 0.125 0.134 0.113 ,I 0.134 0.134
19.7 13.6 13.8 IO.3 9.3 x.7 X.5 7.7 6.8
E. coli
52 60.4 (i9.y I< 66 69 69.75 72 73.7 76.7
0.145 0.117 @l””II 0.122 0.108 0.120 0.117 0.116 (,.l””--
22.1 14.9 14.9 11.3 !b5 !I.9 7.6 5.9 7.6
S. cerevisiae
52 60.4 62.3 66 69 69.76 72 73.7 76.7
0.301 0.306 0.313 0.311 0.276 0.318 0.307 0.273 0.314
u 21.6 18.7 24.5 23,l 19.4 20.3 19.7 16.7 17.4
Is’ 16.7 9.1 13.0 9.1 9.1 9.1 11.5 9.1 9.1
CHROMOSOME
IN S. C'ERE 1~ISIAE
STRUCTURF, TABLE
517
%-cW?LtiWLPd
0,246 0.329 0.281 0.317 0.285 0.263 0.224 0.183 0.188
0.228 0.316 0.301 0.355 0.349 0.31:: 0.29 1 0.291 0.269 0.098 0.122 0.123 0.09!) 0.081) 0.064 0.050 0.028 0.026
B
B’
B
B’
5.21 5.69 4.48 6.39 6.87 8.67 9.46 13.7 16.1
5.53 6.31 5.17 7.55 8.13 9.89 10.4 15.0 17.7
7.07 7.84 6.49 IO.5 1 I-3 13.9 16.6 ‘9.3 31.1
6.91 8.18 7.47 11.4 13-4 15.9 18.3 32.0 34.2
t Corrected to a DNA length of 500 base-pairs. 3 The S. ewe&sine DNA complexity at. each tempcrat,are was calculated as described in the text from t,he ratio of its kz value to that of each standard bacterial DNA obt,ained at the same tomperaturc.
((iillis Pt al., 1970), this method of calculation should yield better values than the previous method of comparing k, values ohtained for the different DNAs at the same t,emperatures. The values calculated for the complexity of 8. cerevisiae DNA by comparing maximal k, values are 7.07 x 10’ to 8.80 x 10’ if E. coli DNA is used as the standard and 5.25 x log to 7.26 x log if B. subtilis DNA is used as the standard. The problem posed by curvature in the reciprocal second-order plots of the optical renaturations of yeast DNA is severe enough to limit considerably the usefulness of the data. While we doubt that the kinetic complexities thus obtained are incorrect by more than a factor of two, we do not wish 60 argue further for their accuracy. However,
the results of the optical
used for the hydroxyapatite-monitored optimally.
renaturations
do show that under the conditions
renaturation,
yeast DNA is indeed renaturing
DNA
(daltons
from
complexit>
Y 1W g,
cakdated
kiurfics
7.64 to 8.80
DXA
7.07 to 8.15
coli
mo~~ifored
2.7 (b)
E.
260 n,rn
2.5 (a)
E. coli DSA
renaturation
4
5.25
to 6.05
of optical
5.78 to 6.66
1. subtilis 11X.4 2.2 (d)
the decreaar
I?. .subtilis DSA 2.0 (c)
by fdowizg
G.30 to 7.26
2.4 (e)
R. scrbtilia I)S-1
hyprchron?icity
at
Complexities were calculated as described in the text. using the maximal k, value, which \vas obtained at f, 20 dep. c to t,, 26 deg. c’ E,r ~wcb DSAI. The range in each calculated complexity is rlnr to uncertainty in tlrau iny tbrl liw through thcl mitial wartion of r.ppeatwI DS;l. draxvn a< Iinrs H :l,l~l H’ in Pig. 2 (see discussion in test). Rofcrences for the DSd complexit>standards arc> (a) Gillis et trl. (1970). (b) Klutz Br Zimm (1972b). (r) Kl,,tz & Zilllrll (l~:zb): Wake (1973). (d) Rau (1975). (e) Gillis et nl. (19iO).
nuclear
complexity
DNA content
8. cerevisicte x 10e9)
and
nuclear
standard
Calculated (daltons
DNA
8. cerevisiae
TABLE
CHROMOSOME
STRUCTURE
IX
,C. CERE
T7Zh'ZAE
51II
(d) LINA per cell measurr~nrrnts Data from the diaminobenzoic acid and EthBr determinations of DNA per cell for both haploid and diploid S. cerrvi&ae cells are presented in Table 5. The two methods of measuring DNA per cell are very different, in that in the diaminobenzoic acid method one is measuring DNA in a clear supernatant containing acid-hydrolyzed and extracted DNA, while in the EthBr method one is measuring native DNA duplex ‘q that’ a DNAase control for interfering in a 41 lysate. Another important point’ 1, subst,ances is performed in the EthBr met)hod (Klotz & Zimm. 1972h; Lauer, 1976) Ijut not in the diaminobenzoic acid method. For the EthBr experiment#s, the lysates were not’ clear. However? control experiments showed that the addit,ion of Dh’A t-o a lysate gave the proper fluorescence of EthBr proport,ional t’o the amount of DNA added, so that the enhancement fluorescence in the cloudy lysat’es was directly proportional to DNA present. Thus. t#he signal due to the t’urbidit)y could be subtracted by running a DNAase cont,rol for thach sample tube. Bar a discussion of the procedure used for determining the DNA cells hy Et,hBr fluorescence and the various controls for content of S. wrevisiaP positively and negatively inberfering substances, see Lauer (1976). Because the values for S. ceravisiae DNA content which we obtained by the two methods were within 15 to 20”;, of each other? and because neither method is known to l)c bet’ter than the other, we averaged the values measured to obtain 1.1 x 1O’O daltons of DNA per haploid ccl1 and 2.2 Y 10 l O dalt,ons per diploid cell. When these values are corrected for 804, mitochondriwl DNA. the DNA cont,ent, of a haploid 1n~4~~usis found to be lOlo daltons. (e) J!folecular
weight
meahwemevh
WC have reported that the largest piece of DNA in the 8. cerevisiae nucleus has a molec*ular weight of 2.0 x 10Q to & 0.2 x lo9 (Laucr & Klotz, 1975). In this paper we have extended this work by doing more molecular weight measurements on lysates of’ both haploid and diploid cells. In addition, measurements were made in two buffer syst,ems: the buffers used before (the 2 M-Na + buffers of Kavenoff & Zimm (1973)) and the guanidinium=HCl/urea buffers of Roberts et al. (1977), in which the final lysis mixture is 2.76 M-guanidinium.HCl, 5.53 M-urea. l.SSq:, Sarcosyl (Ciba-Geigy). and 0.077 M-EDTA (pH 9.3). Results of the molecular weight measurements are presented in Table 6. Two basic findings resulted. The first’ is that the largest piece of DNA in the S. cerevisiae nucleus is the same size in both haploid and diploid cells. The second is that in the buffer system of Rob&s ef al. (1977), which denatures protttins much more strongly than do the buffers of Kavenoff & Zimm (1973), the molecular weight measured is slightly smaller than that measured in the buffers of Kavttnoff & Zimm (1973). In the more highly danat’uring buffer system, the measured 7 value corresponds to a DKA molecular weight of about I.5 x lo9 t,o l-6 x 10’ using the previously determined (Lauer $ Klotz. 1975) relationship between 7 and M for high salt solutions, M = (2’45)(7~,50)o~~3. C-9 Whether the 2 x lo9 dalton value measured to 1.6 x lo9 value measured in the second unknown. Because the t, value of DEA huffel~s, we performed some measurements
in the first buffer system or the 1.5 x 10’ system is closer to the “true” value is is lower in the guanidinium.HCl/urra at, 40°C. Calculated molecular weights
of cells
0.95
1.04iO.06
DaBrZ
deviations. acid (DABA)
For both method.
2.3710.05
2.19
1.12
lo-lo)
2.0
1.0
and
2 meas~ll.~~mk-lltS
Average nuclear DSA content (daltons x lo-lo)
made by the EthRr m&hod phase for all experiments.
Average cellular DNA content (daltons x
crlls, 5 measurrmrnts were to grow well into stationary
2.18
1.84
2.0&0.16
DABA
EthBr
1.10
1.20&0.03
EthBr
haploid and diploid Cells were allowed
Nuclear DNA content (daltons x IO-lo)
Cellular DNA4 content (daltons x lo-lo)
per cell in X. cerevisiae
5
X&hod
The errors reported are standard were made by the diaminobenzoic
Diploid
Haploid
Ploidy
DNA
TABLE
Kavenoff Roberts
Haploid
Haploid
et al. (1977)
& Zimm
et al. (1977) (1973)
(1973)
4
4
2
9
Xumber lysates
of
weights
6
I
50.6-L
61.05
49.3-r-12.7
63.7
To (s)
5.0
8.1
y.:s
of S. cerevisiae
(s)
19.0%
1.9
30.8 + 4.1
18.4h4.8
:i2.2,4.7
d,,,,
DNA x 10-g)
1.5650.1
2.12kO.2
1.53hO.2
2.18zO.2
M(
All experiments were performed on log phase cells. Experiments were performed at cell concentrations between 0.3 x lo8 and 2.5 x lOa cells/ml. For each measurement on a lysate, 7 was calculated from the slope of the linear semilog plot of rotor angle wwus time according to Klotz & Zimm (1972a,b). From 4 to 8 measurrmrnts WYCI‘Rmade on each lysek, and the measured values v~re arcraged to yield 7 for the lysate. Since no concentration dependence of 7 was observed, TO values were calculated by averaging the average values of T measured at different cell concent,rations. The error estimates in TO values arc standard deviations of the measured T values for each cell concentration from TO (except for the case for which 9 lysates were made, in which case the error estimate in 7” is the standard deviation of the averago T value for each cell concentration from 7”). The values of T:,;,~) are found by dividing the TO values by the relative viscosity of the buffers to water: 1.98 for the buffer system of Kawnoff & Zimm (1973), and 2.67 for the buffer system of Roberts et al. (1977). Molecular weights are calculated from the formula M = (2.45 x 10*)(7:,,,,) O-W from Lauer & Klotz (1975), assuming linear DSA. If the DNA we are measuring is actually circular, molecular weights are 2.15 times the values listed in this Table for linear DNA (Klotz & Zimm, 1972a,b).
Roberts
Diploid
system
and molecular
85 Zimm
times
Buffer
Kevenoff
of cells
Diploid
Ploidy
Retardation
TABLE
;jz_”
c:. 1). LAl:E:R. ‘I’. .\I. Itol3EK’l’S ;\NIl I.. (‘. lil,O’l’%
were the same at 40°C as at 50°C. rTt~us. the lower rnolecuhr weight, otkniwd guanidinium .HCl buffers is not due to melting out8 of’ regions of’t,hr, DSA.
in t hc
Interpretation of these molecular weight’s depends on whcthcr or not DNA aggregation or breakage is occurring. Therefore, we have performed a number of comrols to test for these possibilities. Summarized below are the results of controls for DNA aggregation. The molecular weight obtained is independent of DIL’A concentration (over a 3-fold range), of salt> concentration (0.195 M and 2 M-Na+)? of temperature of measurement (20 deg. C range), and of types of proteinase in the lysates (Pronase and Proteinase K). Other aggregation tests unique to viscoelasticity and viscosity measurements are solution shear stress and rotor wind-up time (time for which shear stress is applied) controls. In some cases, when solutions of large DNA molecules are subjected to shear stresses for periods of time, extensive DNA aggregation occurs (Thompson et ul.? 1969; Klotz $ Zimm, 19723). The effects of this aggregation are a manyfold increase in intrinsic viscosity, a smaller increase in retardation time, and a manyfotd increase in intensit’y of recoil over the expected values. In our lysates, the molecular weights are independent of solution shear stress (varied over a factor of 3) and rotor wind-up time (varied from 0.25 7 to many times T). Prom the intensity of recoil in a viscoelastic experiment and the measured retardation time, the number of largest DNA molecules per millilitre can be calculated from t’he formula
in which r is the intensity of recoil for the largest DNA molecules with retardation time T, Q and qrel are the solvent and relative viscosities, w is the angular velocity of the rotor, and kT has the usual meaning. For DNAs whose molecular weights are known from reliable independent measurements, whenever the size measured in the viscoelastic retardimeter is correct, the calculated number of largest molecules is less than or equal to the actual number of largest molecules which would have been present in the lysate if all molecules were intact (Klotz, unpublished observation; Bowen, unpublished observation). In the few cases where lysates produced abnormally high 7 values, presumably due to DNA aggregation, the number of molecules calculated from recoil intensity is larger than the actual number present (Roberts, Lauer observations). For all of $ Klotz, unpublished observations ; Bowen, unpublished our experiments on S. cerevisiae DNA, the calculated number of largest molecules is always 5 to 15% of the total possible largest molecules, assuming we are measuring chromosome-sized DNA molecules. This 5 to 1576 number is in accord with that found for unaggregated bacterial cell lysates and phage DNAs under similar shear stress conditions. This is another indication that DNA is not aggregated in our 8. cerevisiae cell lysates. The final test for aggregation is that haploid and diploid cell lysates yield the same molecular weight. This result indicates that intracellular aggregation of DNA is not occurring. Breakage is much more difficult to control against than is aggregation. In any experiment directed at measuring the molecular weight of DNA, it is possible that DNA is being broken (unless, of course, the size measured is equal to the total genome).
CHROMOSOME
STRUCTURE
1X ,Y. (‘EIZE
I-ISIdE
523
We have measured larger DNA than we measure for S. cerevisiae in the viscoelastometer (E. coli DNA, Klotz & Zimm, 19723; and AgmenelluwA puadruplicutum DN14, Roberts et al., 1977), so we know that it is possible t’o do so. However, yeast is not a bacterium or a blue-green alga, and DNA may be broken during lysis of S. cerevisiae cells. We at least know that our incubation times in the two buffers used for spheroplasting yeast have been optimized (Lauer & Klot,z. 1975). It is possible t’hat enzymatic degradat,ion of DNA is occurring. Our only indicabion that this is not so is that the DNA size in the lysates is stable for at least seven days. It is also possible that DNA is being broken by mechanical shear. We have never measured DNA as large as t#he total genome of S. cerevisiae (about 9x 10g daltons), and DNA of this large size ma\. bc broken in the viscoclast)ometer.
4. Discussion Our first objective in this work was to obtain a measure of the DNA content of the S. cerevisiae nucleus. We obtained not one value for this quantity but a range of values. Therefore, it is necessary to make some comment on the various values rcksulting from the different met’hods of determinat)ion employed. (a) Renaturation
kinetics
The measurement of the kinetic complexity of yeast, DNA as monitored by the decrease of optical density at 260 nm was sufficiently tortuous as to preclude, at least for us, reliable measurement of nuclear DNA content. It is simply too difficult to draw t)angent lines to compound curves. Renaturation kinetics monitored by hydroxyapatite chromatography, however, yielded data that were much more easily reduced. Depending on which bacterial DNA standard was used and which value was assumed for the kinetic complexity of that standard, the DNA content of the yeast nucleus was determined to be from 7 x IO9 t,o 9 x log daltons. However, two comments may well be made concerning our hydroxyapatite results. (1) In the first place, it should be pointed out that the measured size of the S. cerevisine genome is dependent on the choice of bacterial DNA used as a calibration standard. Although the total range of values from our work is small (7 x log to 9 x 10g daltons), it is interesting that the results of Bicknell & Douglas (1970), using h’. coli DNA as a standard, and Whitney & Hall (personal communication), using both E’. coli and B. subtilis DNAs as standards, suggest a similar though much more pronounced trend (i.e. t’hat the 8. cerevisiae genomc appears smaller when B. subtilis DNA is used as a standard than when E. coli DNA is the standard). It is possible that at least part of the difference caused by the two standards is due t,o an effect of (G + C) content’ on the intrinsic rate of renaturation of a DNA sample. However, the literature references on the effect of (G + C) content’ on DNA renaturat’ion are contradictory (Wetmur & Davidson, 1968; Bak ef aE.. 1964.1970; Gillis et al.? 1970; Seidler & Mandel, 1971; Gillis & de Ley, 1975) with the latest work suggesting that (G + C) content’ may have no effect (Gillis & de Ley, 1975). In summary, it is probably best in the case of the present work bo not,e that the difference in measured values for the DNA content of the S. cerevisiae nucleus caused tJy the change from one bacterial standard to the other is well within expected experimental error. (2) The second comment on the hydroxyapatitemonitored renat’uration concerns the expected accuracy of the method. While it is possible to obtain precise result’s
621
G. I).
LAUER,
‘I’. M. ROHER’I’S
rZh-1) I,. t’.
KLOTZ
using hydroxyapaCte chromatography. t’here are many instances in the lit,eraturc where excellent hydroxyapatite dat~a on nuclear DNA complexit,?- disagree with reliable independent estimations of nuclear DNA content hy up to 20”/, (e.g. the data of Rudkin (1964,1965) and Rwsch ut nl. (1971) ~ersu,s tha’t of Laird (1971) on the nuclear DNA content of D. meZarmgaster). There is a generally unstated reserve on t,he part of many researchers to trust t,he absolute accuracp of kinetic determinations of DNA content to much more than a factor of two. Therefore, though we do not wish to discard our hydroxyapatite dat,a on the nuclear DNA content of S. cerevisiap, its accuracy should not be blindly trusted.
Both methods of DNA per cell measurement gave roughly the same value for the nuclear DNA content: 1.0 x lOlo 3 0.1 x lOlo daltons. That two so different techniques as the diaminobenzoic acid and EthBr methods give such similar results seems a good indication that the systematic errors due to the actual methods of measurement should be small. Assuming that cell counting (which is relatively simple for yeast) and preparation of DNA standards are not sources of significant error, the remaining possible error sources are either an incorrect determination of the mitochondrial DNA content or a failure to obtain through the use of stationary cells a true population of all Gl stage cells. However, our value of Sq:, for the fraction of the cell DNA which is mitochondrial is strictly reproducible and falls within (though at the low end of) the published range of values for S. cerevisiae (5 to 20%,: Hartwell (1974)). If anything, our USE of !3%, for the fraction of mitochondrial DNA might lead to a small (less than lo:,<) overestimat,ion of t’he nuclear DNA content. Failure to obtain 100% Gl cells. too, would cause an overestimation of the nuclear genome size. However, our cells were in stationary phase as judged by growth curves and the lack of buds, and stationary phase yeast cells have been found t’o accumulate in the Gl phase of the cell cycle (Bostock, 1970; Hartwell, 1974). Tn summary, we feel that, using our data from hydroxyapatite-monitored renaturation kinetics and DNA per cell measurements, the nuclear DNA content of the yeast S. cerevisiae may be placed in the range 9 x 10gf2 x log daltons. Notably, recent DNA per cell studies from two other laboratories have yielded a nuclear DNA content for S. cerevisiae of 8 x log to 9 x log daltons (Whitney & Hall, personal communication; Kaback, personal communication) in close agreement’ with our values. Using our values for nuclear DNA content, we can now calculate the expected size of the largest piece of DNA in the S. cerez’isiae nucleus if each chromosome contains one piece of DNA. Assuming that, the largest chromosome contains 13% of the nuclear DNA (it contains 13 1: o f the t,otal genetic map units (Mortimer & Hawthorne, complex (Byers & Goetsch, 1975)), one 1973) and 1376 of the total synaptonemal calculates that the largest piece of DNA would be in t,he range 0.9 x log to 1.5 x lo8 daltons. Our measured range for the size of the largest DNA molecules in yeast lysates is 1.5 x log t’o 2.2 x lo9 daltons (or 3.1 x log to 4.6 x log daltons if the DNA molecules in our lysates are circular, although we do not believe this is so, because bacterial DNAs under similar conditions are linear (Klotz & Zimm, 19723 ; Kavenoff, 1972)). Thus, it appears that, the largest molecules in our lysates contain an amount of DNA equivalent to that expected for the largest chromosome or, in the extreme, the equivalent of several chromosomes’ worth of DNA. In vie\v of the extensive controls presented earlier, we believe aggregation is not a factor in our measurements. AS.cerevisiae
CHROMOSOME
STRUCTURE
IN
5’. CERET’ISIAE
52.5
Therefore, these molecular weights should represent minimum values for the size of the largest DNA molecules in the yeast nucleus. The possibility that breakage has reduced our measured values from the true size of the largest molecules found i~b viva is still quite real. Thus, our data cannot be used to rule out the existence of one genome-sized piece of DNA in yeast,. We wish t’o emphasize, however, t’hat in light, of t,hc new data here, viscoelastic measurements of t,he molecular weight of yeast chromosomal DNA are now consisbent with t,he concept8 that each piece of nuclear DNA represents one chromosome. WV x+ish to thank Ma1 Smart for operating the llockman rnodcl E for all analytical centrifugations, Rich Horn for help wit.h computer programming. .I im Hs.ber for sporulating atld mating the R. cereuisiae cells, Argiris Efstradiatis for help in computer-fitting th(b hydroxyapat,it,e data, Don Rau for discussions of rcnaturation kinetics. J. Woodland Hast,illps for use of his fluorimctor, and Alfred Loeblich and Paul Dot.y for generously allowing us the use? of their laboratory facilities. Tllis research x~as supported by grallt, GB 34293 from the United States National Science Foundation. Two authors (G. D. 1,. and T. M. 12.) were supported for part of the duration of this \vork by Cnitod Statc>s Natiotial Scicncc: Found&ion prcdoctoral fellowships. REFERENCES Bak,
A. L.. Black, F. T.. Christiansen, C. & Frelultlt. E. A, ( 1064). ;Vattrre ( Lo?lrlon), 224. 1209 1210. 224, 27Om271. Bak, A. L., Christ,iansen. C. & Stenderup, A. (1969). IVatrcre (hJ??dOn), Bak, A. L., Christiansen, C. & Stenderup, A. (1970). .I. Cr’en Microhiol. 64, 377-380. Bernardi, G. (1971). In Methods in Enzymology ((Grossman, L. & Moldarr, K., eds), 1.01. 21, pp. 95-139, Academic Press, New York. Bicknell, J. N. & Douglas, H. C. (1970). J. Bacterial. 101, 5OS512. Hostock, C. J. (1970). Eqd. Cell Res. 60, 16.-26. Written, K., Graham, D. E. & Nellfcld, B. K. (1974). 111Mefhotls in I!nzymo(oyy (Grossman, L. & Moldave, K., ods), vol. 29, pp. 363 418, Academic Press, New York. Brit,ten. II. .J. & Kohne, D. E. (1968). Science, 161. 52!4-540. Bpcrs, B. & Goc%sch, L. (1975). Proc. If’&. Acad. Sci.. [!.9.;2. 72. 5056&5060. (‘iferri. O., Sara. S. & Tihoni, 0. (1969). Genetics, 61, 567 576. C”ornoo, G., Moore, C., Sanadi, D. K., Grossmall, L. 1. & Murmurs, .I. (1!)66). Scien.ce, 151. 687-689. Costello, D. P. (1961). Bid. Bull. 120, 285 -31”. Darland, G. K. (1969). Ph.D. thesis, University of ~‘astlingtotl. Davidson, E. H.. Hough, B. R.. Amenson, C. S. 8r Brittrn, R. J. (1973). .I. Mol. Biol. 77, I -23. Diacrunakos, E. G., Holland, S. & Pecora, P. (1971). Satwe (London), 232, 33-36. Dupraw, E. J. (1970). In DNA and Chromosomes. pp. 194 195, Holt, Reinhart and Winston, 11x., New York. Gillis, M. & de Lcy, J. (1975). J. illor. Biol. 98, 447- 464. (Willis, M.. de Ley, ,J. & de Cleent, M. (1970). Eur. .I. JPiochem. 12. 143.-153. Hartwell, I,. H. (1974). Bacterial. Rev. 38, 164.. 198. Holm-Hansen, O., Sutcliffe, W. H. & Sharp, .J. (1968). Lirnllol. Oceanogr. 13, 507-514. Hoskins, G. C. (1968). Nature (London), 217, 748-750. Hsu, T. C., Brinkley, B. R. $ Arrighi, F. E. (1967). Clrromosoncw, 23, 137-I 53. Ka\,rnoff, R. (1972). J. Mol. Biol. 72, 801. 806. Ka\.rnoff. R. REZimm, B. H. (1973). Chromosoma, 41, l-27. Kissanr. J. M. & Robins, E. (1958). J. Biol. Chem. 233, 184 188. Klotz. L. C. & Zimm, B. H. (1972a). Afacromolecules. 5, 471-~481. Klotz. L. C. & Zimm, B. H. (1972b). .J. Mol. Biob. 72, 779--800. Lairtl, C. D. (1971). Chromosoma, 32, 278-416. Lange, C. 8. (1975). Biophys. J. 15, 205a.
526
G. D. LAUER,
‘I?. hf. ROBERTS
AN11
L.
(I. KLOTZ
Lauer, G. D. (1976). Ph.D. t,hesis, Harvard University. Laucr, G. D. & Klotz, L. C. (1975). ,J. 3201. Biol. 95. 309 ~326. Marmur, J. (1961). J. 3102. Biol. 3, 208.-218. Mortimer, R. K. & Hawthorne, D. C:. (1073). Genetics, 74, 33-54. Moustacchi, E. & Williamson, D. H. (I!jfiG). Riochenr. Riol~hys. Rex. C!o,nmrcn. 23, .i(i (il. Ogur, M., Minckler, S., Lindegroll, G. & J,indogrel~, c. c’. (1!)52). Arc/l. Hiochem. Biophyls. 40, 175-184. Rasch, IX. M., Barr, H. J. & Rasch, Ii. LV. (1971). Chromosoma, 33, l-1 8. Rau, D. C. (1975). Ph.D. thesis, Harvard Ulliversity. Roberts, T. M., Klot,z. L. C. & Loeblich, A. li., III (1977). ,7. Mol. Riol. 110, 341--361. Rudkin, G. T. (1964). In (:enetics Today. I’roc. Slth Intern. Congr. Genet. vol. 2, pp. 359~-374, Pergamon Press, New York. Rudkin, G. T. (1965). In Vitro, 1, 12-20. Schildkraut, C. L., Marmur, J. & Doty, I’. (1962). J. Mol. Biol. 4, 430-443. Schweizer, E. & Halvorson, H. 0. (1969). Ex@. Cell EZes. 56, 239-244. Schweizer, E., MacKechnic, D. & Halvorson, H. (1969). .J. Mol. Biol. 40, 261-277. &idler, R. J. & Mandel, M. (1971). J. Bacterial. 106, BOS614. Smith, D. & Halvorson, H. 0. (1967). In Methods in Enzymology (Grossmann, L. & Moldave, K., eds), vol. 12A, pp. 538.. 541. Academic Press, New York. Studier, I?. W. (1965). J. ,Vol. BioZ. 11, 373--390. Takayama, S. (1975). Ea.@. Cell Res. 91, 408 -41%. Thompson, D. S., Hays, J. 13. & Gill, S. ,J. (1969). &opoZymers, 7, 571-580. Wake, R. G. (1973). J. i2loZ. Riol. 77, 569- 575. Wetmur, J. G. Bs Davidson, N. (1968). J. ~l!loZ. Biol. 31, 349.--370. Wilson, E. B. (1909). J. Ezpt. Zoob. 6, 64 99. Wilson, E. B. (1928). Th,e Cell in Development and Heredity, Macmillan, New York. Wolff, S., Lindsley, D. I,. & Peacock, W. J. (1976). Proc. Nat. Acad. Sci., U.S.A. 73, 877-881.