THURSDAY, FEBRUARY 19, 1976 12:05:33
         <<< FILE GROUP:: NCI.TXT;4  O4040000006
 ------------------------------------------------------------------------------------------
 **-----------------------------------
  << Reading: NCI.TXT;4 8825O7a
 4,26-JAN-76 10:19:04,26-JAN-76 10:19:04,19-FEB-76 12:05:33


 .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .
           NCI (Mol. Genetics) renewal report

                     CA-16896-
      (report covering the period Jan 1- Dec 31 1975)

 1.     For some time, the principal focus of our work has been the
 development of methods for the purification of specific segments of
 bacterial DNA, and their artificial rejoining to produce new, genetically
 active arrangements. These microbial studies were intended to be a
 prototype for important applications in cell and cancer biology, pending
 the solution of formidable technical problems, and the clearing up of the
 biohazard policy issues that currently entail work with human DNA.
      Our laboratory pioneered in the application of the 2-strand ligase of
 T4, which V. Sgaramella had shown here to be active on  "flush-ended" DNA.
 However, incessant difficulties in producing such flush-cut segments led
 us to concentrate on the alternative systems using stagger-cut restriction
 endonucleases, which have indeed attracted explosive interest in dozens of
 laboratories. These enzymes have the special virtue of cleaving DNA at
 specific sites, and we have Oas have many others nowa used them for the
 purification of specific segments of DNA. As reported in O1a, even as
 complex an ensemble as the total DNA of Bacillus subtilis has lent itself
 to separation by this approach (EcoRI endonuclease, followed by agarose
 gel electrophoresis.)
      Now we are     a) returning to the problems of T4-ligase
           b) pursuing specific applications of the DNA-splicing
                methodologies
           c) looking into the use of our current research as
                    a testbed for computer-mechanized assistance
                    to the planning of experimental work.
+
1;  <NORMAL>-TEXTS.SUMEX;1   Fri 22-Sep-78 9:03PM                            PAGE 33:39


 2.     The work reported in detail in O1a was briefly summarized last
 year, and is the foundation for most of our continuing effort. It enables
 us to make consistent preparations of purified segments with specific
 genetic activity in the transforming system. These segments in turn have
 been spliced into various plasmids for amplification, and then tested for
 biological activity either in E. coli or in Bacillus subtilis.
      Methods of using the plasmid DNA clones as reagent-probe-stains for
 electropherotograms have been developed, and they assure that homologous
 DNA has indeed been amplified.  So far, however, we have no  persuasive
 evidence of the biological activity of these segments -- but must also
 wonder whether we have correctly identified their genetic provenience. Our
 cautions -- perhaps even overstimulated by recent discussions about
 presumed biohazards of certain gene complexes -- have hampered rapid
 progress in this area by requiring the painstaking development of
 selective techniques that avoid the use of certain antibiotic-resistance
 markers, and these have come only slowly to hand (e.g. colicin resistance,
 which is sometimes less clearcut than the literature implies.)
      At present we are systematically cataloguing the library of B.
 subtilis DNA segments so far amplified. Present methods stringently
 discriminate against the larger pieces that are of the most interest from
 a genetic standpoint; hence we are working out alternatives , like the
 deleted-lambda system, for application in this context.
      We have also started to examine other endonucleases, e.g. from B.
 amyloliquefaciens, which segments B. subtilis at different sites, with a
 view to using overlapping segments for comprehensive mapping of the genome.

      In work related to the flush-cut ligation, studies on the
 modification of virus P22 activity by exonucleases have been completed and
 mss. submitted to the Journal of Molecular Biology for publication. These
 have been provisionally accepted O3,4a. The most curious finding is the
 considerable AUGMENTATION of infectious activity of the viral DNA upon
 optimal terminal erosion, presumably by facilitating terminal circular
 sequence-pairing and closure.  This is an important lead with respect to
 the early events of infection with purified DNA, and the decision of the
 race between the cell's defense mechanisms (including exonucleases )and
 the virus's initiation of its replication cycle.
      In addition, we have independent evidence to corroborate the
 incorrectness of the random-permutation model of P22-DNA which has lain
 upon this field for many years.
      Studies on B. subtilis transformation, involving heterogenous strains
 and their DNA hybrids, indicate that 1) sequence homology is the principal
 factor regulating the success of distant 'crosses', and 2), unlike
 transduction and crossing, host-restriction-modification systems play
 little if any role in transformation. The most likely explanation is the
 predominance of SINGLE-stranded DNA in transformation. OA practical
 concern is that singled-stranded VIRAL DNA, which is rarely found in
 nature, may be a particularly hazardous agent for xenobiotic exchanges
 when it does occur or may be artificially produced. As we do not really
 know whey even bacteriophage DNA is not (manifestly) infectious for man,
 no nugget of information of this kind is to be overlooked.a

      Besides the earlier models for flush-cut DNA joining (mainly P22), we
 have also looked into the artificial production of such segments, and have
 found a B. subtilis enzyme that seems to work properly to this end.
 Accordingly, there is persuasive (but still preliminary) evidence that
 such segments can be inserted into plasmids and amplified.
+
1;  <NORMAL>-TEXTS.SUMEX;1   Fri 22-Sep-78 9:03PM                            PAGE 33:40


      As already indicated, the present policy turmoil about hybrid DNA has
 given pause about pursuing experiments with animal and human DNA, despite
 their enormous potential in many research and practical medical
 applications (see e.g., O5a), and we are therefore awaiting the resolution
 of these regulatory policies (and wondering how to pay for any new
 facilities that may be indicated) before moving towards such productive
 applications.

      We have, however, decided to begin with the amplification of
 BACTERIAL T-RNA (in systems that must rank near the bottom of imaginable
 risk) as a prototype of similar studies to be done in due course with
 human material. These may be particularly relevant to cancer in the light
 of many reports on altered distribution of t-rna specificities in cancer
 cells. At the very least, amplified clones with specific t-dna sequences
 will provide histochemical reagents that should lend a new dimension of
 discrimination in the microscopic-pathological diagnosis of different cell
 types. We are just at the point of classifying which of the dna clones in
 our B. subtilis library contain the relevant t-rna information.

      For some time I have been collaborating with Prof. E.A. Feigenbaum of
 the Computer Science Department on the development of computer programs
 that emulate some aspects of human scientific thinking. It is commomplace
 now to rely upon computers for tedious 'numbercrunching', data-banking and
 information retrieval applications; but the exaggerated claims for the
 future of artifical intelligence made about 15-20 years ago have not borne
 fruit, and have tended to put the field in some disrepute. Nevertheless,
 we have been able to make demonstrable progress in emulating some more
 creative aspects of scientific thinking, particularly in the formation of
 structural hypotheses -- until now mostly in the field of analytical
 organic chemistry O6a. This work has gone so well, that during the last
 year we have started a new pilot project to do much the same kind of work
 in the field of molecular genetics, with the mutually reinforcing
 advantages of being able to draw upon the 'real-life' activity of our
 research group in pursuing actual and difficult problems; and conversely
 that we might anticipate some reciprocal help in the solution of our own
 day-to-day problems in the laboratory. (This is just what happened in Dr.
 Djerassi's mass-spectrometry laboratory in connection with our other
 work.)     This effort is getting enthusiatic attention from a
 considerable number of staff and students in Computer Science as well as
 from my own group. We are focussing on the computation of candidate PLANS
 for experiments to test hypotheses about the structure of DNA from a given
 source. The work so far seems also to be amenable to expressing (the more
 difficult) issues of the structural hypotheses themselves.  If our
 previous effort with mass spectrometry is a relevant guide, it should take
 about five years for us to develop the data base and the sophistication in
 manipulating it to expect tangible help in our routine laboratory work.
 (This is of course contingent on funding support being sought from various
 sources.)