U.S. Department of Energy

Human Genome 1993 Program Report: Resource Development

We are collaborating on a project that has focused the efforts of several Russian laboratories on cloning, mapping, and sequencing the coding regions of human chromosome 19. We have constructed a partial yeast artificial chromosome (YAC) library from a hybrid (human/hamster) cell line that retains only human chromosome 19 and are screening the library for human DNA fragments. For now, our laboratory is primarily concerned with optimizing YAC library storage.

Long DNA fragments cloned in YACs are often used as a source of starting material for further restriction analysis and sequencing. However, maintaining cloned human DNA in yeast cells is well known to be a cause of human sequence variation. In some cases this variation, sometimes referred to as structure instability, is quite noticeable. Minimizing variation by proper YAC maintenance is therefore highly desirable.

One way to reduce the frequency of inaccurate sequences in a YAC library is to introduce certain changes in the genotype of recipient cells. Mutations of genes RAD52 and RAD1 have been found to cause reduction in chimerism and recombinational rearrangements in some YACs. We have begun a systematic study to determine the relationship of cloned DNA and host genotype. We isolated and characterized mutations in several nuclear spontaneous rho-mutability (SRM) genes that mediate maintenance of various redundant and optional genetic structures in yeast cells. The first step of this project will be to analyze YAC maintenance (i.e., structure variation and mitotic stability of YACs from our library) in mutant SRM cells.

*Refining the Map Location of 5q31-5q33 Deletion with Known Molecular Markers

A useful approach for constructing physical maps of defined chromosomal regions is to use a hybrid cell line containing human chromosomes with a deletion in the DNA region of interest. DNA from such cells may be used for selection of a restricted number of cosmid clones to enable easier mapping of a particular chromosome.

We prepared a hybrid cell line containing human chromosomes having a deletion that maps to 5q31-5q33 by G-banding. Our goal is to identify molecular markers in the deletion region and use them to screen yeast artificial chromosome (YAC) and cosmid libraries for corresponding clones. These markers and clones will provide useful starting points for developing physical maps of ordered clones.

The work will be performed in two steps: (1) from available sequence data, polymerase chain reaction primers will be designed and used to test for the presence of these markers in the 5q31-5q35 region; and (2) markers mapping positively to this region will be used to screen the YAC and cosmid libraries.

This work is being carried out as part of the Chromosome 5 project of the Russian National Human Genome Program.

A Novel Bacteriophage P1-Derived Electroporation-Based Vector for the Construction of Large-Insert Recombinant DNA Libraries

A number of different cloning systems are currently being used to generate contiguous physical maps of individual chromosomes; these systems include yeast artificial chromosomes (YACs), cosmids, and bacteriophage P1. Each of these traditional cloning vehicles is either limited in the ability to propagate large DNA fragments (e.g., 40 kb for cosmids and 90 kb for P1 phage) or contains a high proportion of chimeric DNA molecules (YACs). To mitigate these limitations, we have modified the bacteriophage P1 vector system for producing recombinant molecules by transformation into host cells through electroporation. We call this cloning system P1-derived artificial chromosomes (PACs).

We are presently constructing a fivefold redundant total human genomic library using the PAC cloning system. The PAC cloning vector offers many of the advantages of bacterial artificial chromosomes (BACs), which are derived from the bacterial F-factor and offer ease of manipulation in bacterial hosts as well as a very low frequency of chimeric clones. In addition, the PAC cloning system offers a selectable marker for recombinant clones (SacB gene product, levansucrase) using medium that contains sucrose for selection and has the ability to amplify clones using an isopropyl beta-D-thiogalactopyranoside (IPTG)-inducible high-copy-number origin of replication. The maximum size of PAC clones is not limited by a headful packaging mechanism (traditional P1 cloning systems) and can accommodate inserts up to 300 kb in length.

To isolate chromosome 19-specific PAC clones for closing the chromosome, we used a Biomek workstation to construct a set of high-density colony filters from a portion of the PAC library. The filters have been screened using a variety of chromosome 19-specific probes, including inter-Alu polymerase chain reaction (PCR) products, "degenerate oligo-primed" PCR (DOP-PCR) products, and a 37-bp repetitive element (pe670). We isolated a number of chromosome 19-specific clones, and six putative pe670-positive PAC clones have been identified. Assuming a random distribution of pe670 repeats on chromosome 19, screening a 0.25-fold redundant filter set should result in localizing 13 pe670-positive PACs. The isolation of six pe670-positive PACs probably results from the nonrandom distribution of the pe670 repeat along chromosome 19. PACs that contained pe670 were used as probes to screen chromosome 19-specific high-density cosmid filters. The hybridization of individual PAC clones to a number of chromosome 19-specific cosmids is being confirmed. These data demonstrate the potential of PAC clones for generating a contiguous chromosome 19 physical map.

*Identification and Mapping of DNA-Binding Proteins Along Genomic DNA by DNA-Protein Crosslinking

Techniques such as chemical protection and enzymatic footprinting together with gel-retardation assays enable the mapping of proteins on genomic DNA but provide no information on bound proteins. We have developed the "protein-image" hybridization method, which allows us to identify the size of proteins associated with specific DNA sequences. This is accomplished either directly in vivo in whole-cell experiments by using uv-induced DNA-protein crosslinks or in isolated nuclei by chemical crosslinking.

A protein can be precisely localized on DNA by digestion of crosslinked DNA with restriction endonuclease and exonuclease III. Many sequences can be tested by repeatedly hybridizing the same blot with a number of probes. Successful development of genome programs will provide huge amounts of new sequences, including those of regulatory and structural regions that may be associated with specific proteins. Without much additional effort, mapping of proteins along genomic DNA can be directly coupled with the program of genome sequencing.

*Sequence-Specific Proteins Binding to the Repetitive Sequences of the Human Genome

Although repetitive sequences occupy a large part of the whole eukaryotic genome, their purpose is not well understood. We will attempt to elucidate their role by examining sequence-specific DNA-binding proteins (SSBPs). Methods and technology for finding SSBP were established by using human Alu retroposon-like short interspersed repeated sequences (SINEs) as a model. Two proteins with molecular weights of 40 and 80 kilodaltons (kDa) were found in HeLa nuclear extracts and partially purified. The 80-kDa protein (Alu-SSBP) was found to bind the sequence within the Alu repeat that is homologous to the T-antigen binding site of SV40; cell cycle control may be mediated via Alu-SSBP.

Repetitive sequences with simple structures might play a role in three-dimensional chromatin organization. The presence of SSBP that specifically binds with human satellite 3 (HS3) was shown using the developed model system and a gel shift assay with constituents from a partially solubilized nuclear matrix. HS3 has multiple binding sites for this specific SSBP, and binding influences the secondary structure of HS3 by bending the DNA. The molecular weight of HS3-SSBP is 80 kDa; in the most stable retarded complex it apparently exists as a dimer, and in solution under physiological conditions it assembles with other matrix proteins into complexes with molecular weights of about 800 kDa.

A human embryo lgt11 expression library was screened with labeled HS3 under suitable conditions for DNA-protein binding. The identified clone produced a fusion protein with strong affinity for HS3. Immunofluorescence with antibodies against this fusion protein demonstrated a characteristic intranuclear pattern in HeLa cells and on the isolated nuclear matrix. Future plans are to obtain the genes of the Alu-binding proteins and look for proteins to the other chromosome-specific satellites in the nuclear matrix.

*Protein-Binding DNA Sequences

A random modification method was developed to determine a target site on DNA that would be recognized by specific DNA binding proteins such as the Oct-2B transcription factor. To this end we have used as a probe a random oligonucleotide of the following structure:

CCGGGAAGCTGnnnnnnnnGTGCTGCCTTCGACnnnnnnnnCACGACGGGCC,

where n is any of the four bases A, G, C, or T.

DNA fragments containing the binding area were cloned and sequenced. We found two groups of sequences interacting with the conservative POU domain of octomer proteins, one group containing a common tetranucleotide T/CAAA and the other having the tetranucleotide TAAT. All tested probes differ in their affinity to the POU domain. The stability of DNA-protein complexes depends on the structure of core and flanking sequences. The affinity of a group to the POU domain depends particularly on the presence of the second half of the binding site (ATGC). Our results indicate that the Oct-2B protein interacts with canonical sequence and degenerated sequences. These data have greatly increased the number of potential targets for octamer proteins on DNA and changed our view on gene-expression regulation by these protein factors.

Strategies for Identification of Evolutionarily Conserved DNA Sequences