Meselson described some of his work in progress on density gradient centrifugation, and provided Luria with instructions for
doing it. Meselson's experiments with Franklin Stahl on the replication of DNA that were published later in 1958 helped
cement the concept of the double helix by validating a model that many scientists still saw as speculation and demonstrated
how two strands of a helix could physically code for the material of inheritance.
Number of Image Pages:
2 (187,643 Bytes)
1958-01-18 (January 18, 1958)
California Institute of Technology
Luria, Salvador E.
Original Repository: American Philosophical Society. Library. Salvador Luria Papers
Reproduced with permission of Matthew Meselson.
Genetics Lessons from Bacteriophage, 1938-1944
Letter from Matthew Meselson to Salvador E. Luria (October 28, 1958)
Please pardon this long delay in answering your letter. I've been traveling since before Christmas. Last week I visited
Washington University and was quite impressed with the present performance of Kornberg's system. Howard Schactman has
been there and has found that the synthesized DNA resembles the primer in regard to sedimentation constant and (more significantly)
intrinsic viscosity. When sonicated primer is used, the product DNA seems to be shortened accordingly although this last
result is preliminary according to Howard. The system makes 10x increase in DNA over primer, but if T.P. is used as primer
the transforming activity goes neither up nor down. However, if any one of the nucleotide triphosphates is omitted, the transforming
activity is nearly completely eliminated in the experimental time period. They blame the inability to make net increase of
T.P. on nucleases known to be present and are accordingly setting about the preparation of really pure enzyme from several
hundred pounds of Coli! The system puts in deoxy-UTP as well as thymine TP and deoxy-inosine TP for GTP in accord with WC
pairing expectations. Ribonucleotides won't go in. The base composition of DNA made with Coli primer resembles Coli
DNA while T2 primer makes T2-like product DNA. The system seems to be well enough established to justify using heavy nitrogen
DNA for primer and looking for half-heavy molecules in CsCl gradient. I'll make heavy T4 DNA for them and perhaps centrifuge
it too if they don't prefer to do it there.
We have not done any new centrifuging (not even to repeat the transfer experiment) for two reasons. First, we're all
out of CsCl and our old supplier has vanished. (He does not answer letters and his 'phone in Phila. has been disconnected.)
We have contracted with a local firm to make it for us but they won't have any for perhaps another month. The second
difficulty Frank and I have had is that we have almost no time on the centrifuge schedules. We had tied up a machine for
almost a year and now Dintzis, Vinograd, and Sinsheimer quite understandably want to get on with their own work. Our long
equilibrium runs require a machine of our own. Accordingly Max has very quickly arranged for us to buy one to be kept in
the phage group. It will be here in about a month. When we're going again, we'd be glad and interested to run Pl
and the antigenicity mutants too. I hope we could be ready within two months . . . we will be if there are no snags in setting
up the machine.
To make runs yourself, choose a CsCl solution of density equal to that of the phage under the assumption that the density
of the phage is the arithmetic average of the protein (l.30) and the DNA (l.70) densities weighted according to the per cent
DNA in a phage. Buffer at pH 7 with 0.01 M phosphate. In the first runs put in enough phage so the OD at 260A of the starting
solution is about 0.4. With this much phage you will see which way they move in case the band is off scale. You can tell
how much to shift the density of the starting solution to put the band in the middle with the aid of the approximate relation
density gradient = 8x10 ^-10 w(^2) r with w in radians per second and r, the distance in cm from the center of rotation.
I suggest a speed of 30-40 thousand rpm for exploratory runs. The CsCl equilibrium takes about 7 hours and with molecules
as big as Pl two or three hours more will be enough to get quite near equilibrium at the above speeds. At lower speeds, the
CsCl equilibrium takes still 7 hours but the macromolecules band much more slowly. The rate of band formation goes as log
w^2 over w^4. The ratio of separation between two bands to their half-width is speed independent so resolution is the same
as any speed. Good luck,