Shotgun polymerase chain reaction: construction of clone libraries specific to a NotI fragment of flow-sorted human chromosome 22.

We have developed the "shotgun polymerase chain reaction," a method for obtaining a large number of DNA markers specific to a giant DNA fragment, which facilitates analysis of a particular chromosomal region. We applied this method to a giant NotI fragment which carries the immunoglobulin lambda constant region on chromosome 22. NotI digests of chromosome 22 flow-sorted from human B-lymphoblastoid cell line GM130B were size fractionated by pulsed-field gel electrophoresis. Preliminary Southern hybridization analysis revealed that the immunoglobulin lambda constant region was conveyed on 1.4- and 1.3-Mb NotI fragments in this cell line. The agarose gel corresponding to 1.2 to 1.5 Mb in size was excised into slices and subjected to polymerase chain reaction to identify gel slices containing NotI fragments carrying Ke-Oz+, a subtype of the immunoglobulin lambda constant region. From the NotI fragment thus identified, a large number of small DNA segments were amplified through the ligation-mediated random polymerase chain reaction method. The amplified products were cloned and analyzed for chromosomal origin and localization to particular NotI fragments. Seven of eighteen clones originated from the 1.4-Mb NotI fragment of chromosome 22 in GM130B cells, which appears to be exactly the same as detected by a probe for the immunoglobulin lambda constant region.     

Construction and characterization of a NotI linking library of human chromosome 21

Effective procedures have been developed for constructing NotI linking libraries starting from chromosome-specific genomic libraries. Fifteen different single copy and two rDNA NotI linking clones from human chromosome 21 were identified in two libraries. Their chromosomal origin was confirmed, and regional location established using hybrid cell panels. Hybridization experiments with these probes revealed pairs of genomic NotI fragments, each ranging in size from less than 0.05 to 4.0 Mb. Many fragments displayed cell type variation. The total size of the NotI fragments detected in a human fibroblast cell line (GM6167) and mouse hybrid cell containing chromosome 21 as its only human component (WAV17) were approximately 32 and 34 Mb, respectively. If these fragments were all non-overlapping, this would correspond to about 70% of the 50-Mb content estimated for the whole chromosome. The linking clones will be enormously useful in the subsequent construction of a NotI restriction map of this chromosome. Characterization of these clones indicates the presence of numerous additional sites for other enzymes that recognize sequences containing CpG. Thus most NotI linking clones appear to derive from CpG islands and probably identify the 5' end of genes.     

A physical map of the human Y-chromosome short arm

A physical map of the Y-chromosome short arm was constructed using DNA probes p19B, Y-286/la5, pZFY, Y-280, and Y-227. These probes hybridize with four NotI fragments of 400 kb (p19B and Y-286/la5), 350 kb, 1.9 Mb, and 3.0 Mb, respectively. The restriction fragments were shown to be adjacent to each other by analysis of NotI partial digests, overlapping restriction fragments, and/or the detection of rearranged restriction fragments in a 46,XX male. The present map covers approximately 5.6 Mb of contiguous DNA of Yp. Previously, the size of the pseudoautosomal region was estimated to be 2.3 Mb, and a 5.3-Mb NotI fragment containing Y-specific repeated DNA was assigned to proximal Yp. These and the present data account for approximately 13 Mb and thus for most of the DNA content of the Y short arm.     

Determination of genome size of Haemophilus influenzae Sb: analysis of DNA restriction fragments.

Genomic DNA size was measured in clinical isolates of Haemophilus influenzae by Pulsed-Field Gel Electrophoresis of DNA restriction fragments. Because of the high (64%) A+T content of H. influenzae DNA, restriction enzymes that recognize sequences with at least four GC base pairs were expected to be rare cutters. Five enzymes that produced fragments greater than 200 kb in size were used to digest intact chromosomes and fragments resolved by TAFE and/or FIGE: ApaI (GGGCCC), EagI (CGGCCG), NotI (GCGGCCGC), RsrI (CGGA/TCCG), and SmaI (CCCGGG). All five had recognition sequences with at least six GC base pairs. The genomic DNA size of H. influenzae serotype b, estimated with ApaI, EagI, NotI, RsrII, and SmaI, is 1,950 kb.     

Construction of a Not I restriction map of the fission yeast Schizosaccharomyces pombe genome.

Pulsed field gel electrophoresis and large DNA technology were used to construct a Not I restriction map of the entire genome of the fission yeast Schizosaccharomyces pombe. There are 14 detectable Not I sites in S. pombe 972h: 9 sites on chromosome I and 5 sites on chromosome II, while no Not I sites were found on chromosome III. The 17 fragments (including intact chromosome III) generated by Not I digestion were resolved by PFG electrophoresis. These fragments ranged in size from 4.5 kb to approximately 3.5 Mb. Various strategies were applied in determining, efficiently, the order of the fragments on the chromosomes. The genomic size measured by adding all the fragments together is about 14 Mb and the sizes of the three chromosomes are I, 5.7 Mb, II, 4.6 to 4.7 Mb, and III, 3.5 Mb. These are generally somewhat smaller than estimated previously.     

Physical maps of the genomes of three Bacillus cereus strains.

NotI restriction maps of the chromosomes from Bacillus cereus ATCC 10876, ATCC 11778, and the B. cereus type strain ATCC 14579 have been established and compared with the previously established map of B. cereus ATCC 10987. Between 10 and 14 NotI fragments were observed, ranging from 15 to 1,300 kb, in digests of DNA from the various strains. The sizes of the genomes varied between 5.4 and 6.3 Mb. The maps were constructed by hybridization of 42 random probes, prepared from B. cereus ATCC 10987 libraries, to fragments from partial and complete NotI digests, separated by pulsed-field gel electrophoresis. Nine probes were specific for ATCC 10987 only. Probes for five B. subtilis and five B. cereus genes were also used. The NotI restriction fragment patterns of the four strains were strikingly different.     

Physical genome map of the unicellular cyanobacterium Synechococcus sp. strain PCC 7002.

A physical restriction map of the genome of the cyanobacterium Synechococcus sp. strain PCC 7002 was assembled from AscI, NotI, SalI, and SfiI digests of intact genomic DNA separated on a contour-clamped homogeneous electric field pulsed-field gel electrophoresis system. An average genome size of 2.7 x 10(6) bp was calculated from 21 NotI, 37 SalI, or 27 SfiI fragments obtained by the digestions. The genomic map was assembled by using three different strategies: linking clone analysis, pulsed-field fragment hybridization, and individual clone hybridization to singly and doubly restriction-digested large DNA fragments. The relative positions of 21 genes or operons were determined, and these data suggest that the gene order is not highly conserved between Synechococcus sp. strain PCC 7002 and Anabaena sp. strain PCC 7120.     

The genome of Haemophilus influenzae Rd has a unique NotI site

Previous analysis of physical maps of Haemophilus influenzae, which is circular and 1.9 Mb in length [Lee and Smith, J. Bacteriol. 170 (1988) 4402-4405; Kauc et al., J. Bacteriol. 171 (1989) 2474-2479], did not detect any NotI (GCGGCCGC) restriction sites. A transposon, Tn916, was constructed to contain a NotI linker cloned into its NciI site and introduced into the H. influenzae chromosome. NotI digestion of chromosomes containing a Tn916-associated NotI site followed by separation of fragments by field-inversion gel electrophoresis revealed the presence of two fragments obtained by two NotI cuts, one in Tn916 and the other, a unique, 'natural' NotI site in the original chromosomal DNA. The examination of other Haemophilus strains demonstrated the presence of one or more NotI sites in all of those tested.     

Physical and genetic map of the genome of Vibrio parahaemolyticus: presence of two chromosomes in Vibrio species

We constructed a physical map of the genomic DNA (5.1 Mb) for Vibrio parahaemolyticus strain AQ4673 by combining 17 adjacent NotI fragments. This map shows two circular replicons of 3.2 and 1.9 Mb. Pulsed-field gel electrophoresis (PFGE) of undigested genomic DNA revealed two bands of corresponding sizes. Analysis both by NotI digestion and by Southern blot of the two isolated bands confirmed the existence of two replicons. The presence of genes for 16S rRNA on both the replicons indicates that the replicons are chromosomes rather than megaplasmids. The two bands were also seen after PFGE of undigested genomic DNA of V. parahaemolyticus strains other than AQ4673, and of strains belonging to other Vibrio species, such as V. vulnificus, V. fluvialis and various serovars and biovars of V. cholerae. It is noteworthy that V. cholerae O1 strain 569B, a classical biovar, was also shown to have two replicons of 2.9 and 1.2 Mb, which does not agree with a physical map proposed in a previous study. Our results suggest that a two-replicon structure is common throughout Vibrio species.     

Insertion of transformation vector DNA into different chromosomal sites of Dictyostelium discoideum as determined by pulse field electrophoresis

Chromosomes of the cellular slime mold Dictyostelium discoideum were fractionated on three pulse field gel electrophoresis systems (pulse field, orthogonal field and C.H.E.F. (Contour-clamped Homogeneous Electric Fields] into a series of 13 bands ranging from 0.1 Mb to over 2 Mb in size. Since this organism has only seven chromosomes (estimated to be 1-10 Mb), and -90 copies of an 88-kilobase linear ribosomal DNA molecule (14% of genome), it was apparent that not all of these bands were whole chromosomes. However these bands were reproducibly obtained with the cell preparation used. They fell into three categories: i) four large poorly resolved DNA molecules (-2 Mb in size) which represent very large fragments or intact chromosomes, ii) eight faint bands ranging from 0.1 Mb to 2 Mb, iii) a prominent band in the apparent size range of about 0.15 Mb. Cloned Fragment V of an EcoR1 digest of the ribosomal DNA, hybridized to the 0.15 Mb band indicating it contained the linear ribosomal DNA. This chromosomal banding pattern was used to examine the stability and location of vector DNA in 16 transformed strains of D. discoideum. Each transformed strain was initially selected on the basis of G418 resistance with an integrating vector containing pBR322 sequences. Eleven transformants still carried pBR322 sequences after more than 60 generations of growth without selection on G418. All four strains transformed with constructs containing regions of the D. discoideum plasmid Ddp1 had lost their pBR322 insert, indicating that integration of Dictyostelium plasmid DNA into chromosomes leads to instability. Orthogonal field electrophoresis of the eleven strains still carrying pBR322 sequences revealed at least seven different integrating sites for the transforming DNA. We conclude that these vectors have many possible sites of integration in the D. discoideum genome.     

Use of pulsed-field agarose gel electrophoresis to size genomes of Campylobacter species and to construct a SalI map of Campylobacter jejuni UA580.

To determine the physical length of the chromosome of Campylobacter jejuni, the genome was subjected to digestion by a series of restriction endonucleases to produce a small number of large restriction fragments. These fragments were then separated by pulsed-field gel electrophoresis with the contour-clamped homogeneous electric field system. The DNA of C. jejuni, with its low G+C content, was found to have no restriction sites for enzymes NotI and SfiI, which cut a high-G+C regions. Most of the restriction enzymes that were used resulted in DNA fragments that were either too numerous or too small for genome size determination, with the exception of the enzymes SalI (5' ... G decreases TCGAG ... 3'), SmaI (5' .... CCC decreases GGG .... 3'), and KpnI (5' ... GGTAC decreases C .... 3'). With SalI, six restriction fragments with average values of 48.5, 80, 110, 220, 280, and 980 kilobases (kb) were obtained when calibrated with both a lambda DNA ladder and yeast Saccharomyces cerevisiae chromosome markers. The sum of these fragments yielded an average genome size of 1.718 megabases (Mb). With SmaI, nine restriction fragments with average values ranging from 39 to 371 kb, which yielded an average genome size of 1.726 Mb were obtained. With KpnI, 11 restriction fragments with sizes ranging from 35 to 387.5 kb, which yielded an average genome size of 1.717 Mb were obtained. A SalI restriction map was derived by partial digestion of the C. jejuni DNA. The genome sizes of C. laridis, C. coli, and C. fetus were also determined with the contour-clamped homogeneous electric field system by SalI, SmaI, and KpnI digestion. Average genome sizes were found to be 1.714 Mb for C. coli, 1.267 Mb for C. fetus subsp. fetus, and 1.451 Mb for C. laridis.     

Breakage of the human Y-chromosome short arm between two blocks of tandemly repeated DNA sequences

Y-chromosomal rearrangements, a common cause of sex reversal in man, frequently occur between two blocks of repeated DNA. Both blocks are composed of 20-kb tandemly repeated Y-chromosome-specific DNA sequences. They are located in the proximal portion of the Y short arm on a NotI restriction fragment of approximately 5.3 Mb and on an MluI fragment of approximately 5.5 Mb. Chromosome breaks positioned between the two blocks were detected in two of three 46,XY females with deletions of Yp and in five of six 46,XX males positive for the repeat sequences. The rearranged NotI fragments in the 46,XX males were 4.4 Mb and the MluI fragments were 2.0 Mb in length. This indicates that breaks occur within a small region of Yp defined by the two blocks of specific repeated DNA sequences. The region between the two blocks thus appears to be a focus of structural lability in the human Y chromosome.     

The African trypanosome genome

The haploid nuclear genome of the African trypanosome, Trypanosoma brucei, is about 35 Mb and varies in size among different trypanosome isolates by as much as 25%. The nuclear DNA of this diploid organism is distributed among three size classes of chromosomes: the megabase chromosomes of which there are at least 11 pairs ranging from 1 Mb to more than 6 Mb (numbered I-XI from smallest to largest); several intermediate chromosomes of 200-900 kb and uncertain ploidy; and about 100 linear minichromosomes of 50-150 kb. Size differences of as much as four-fold can occur, both between the two homologues of a megabase chromosome pair in a specific trypanosome isolate and among chromosome pairs in different isolates. The genomic DNA sequences determined to date indicated that about 50% of the genome is coding sequence. The chromosomal telomeres possess TTAGGG repeats and many, if not all, of the telomeres of the megabase and intermediate chromosomes are linked to expression sites for genes encoding variant surface glycoproteins (VSGs). The minichromosomes serve as repositories for VSG genes since some but not all of their telomeres are linked to unexpressed VSG genes. A gene discovery program, based on sequencing the ends of cloned genomic DNA fragments, has generated more than 20 Mb of discontinuous single-pass genomic sequence data during the past year, and the complete sequences of chromosomes I and II (about 1 Mb each) in T. brucei GUTat 10.1 are currently being determined. It is anticipated that the entire genomic sequence of this organism will be known in a few years. Analysis of a test microarray of 400 cDNAs and small random genomic DNA fragments probed with RNAs from two developmental stages of T. brucei demonstrates that the microarray technology can be used to identify batteries of genes differentially expressed during the various life cycle stages of this parasite.     

Construction of NotI restriction map of the Streptococcus mutans genome

Streptococcus mutans and Streptococcus sobrinus are the major causative organisms of human dental caries. Pulsed-field gel electrophoresis (PFG) showed that the restriction enzyme NotI produced ten and six DNA fragments from the genomes of S. mutans strain MT8148 and S. sobrinus strain 6715, respectively. The sizes of the chromosomes of S. mutans and S. sobrinus were each estimated to be about 2200 kb. The NotI restriction map of S. mutans MT8148 genome was constructed by Southern blot analysis with probes that overlapped two adjacent NotI fragments. Several virulence-associated genes of S. mutans were placed on the NotI restriction map. In addition, unique 'fingerprints' of S. mutans chromosomal DNA digested with NotI were produced by PFG, and these may be useful for epidemiological studies.     

Physical Mapping of Rice Chromosomes 8 and 9 with YAC Clones

First efforts for physical mapping of rice chromosomes 8 and9 were carried out by ordering YAC clones of a rice genomicDNA library covering six genome equivalents with mapped DNAmarkers. A total of 79 and 74 markers from chromosomes 8 and9, respectively, were analyzed by YAC colony and Southern hybridizationusing RFLP markers of cDNA and genomic clones, and by polymerasechain reaction (PCR) screening using PCR-derived and sequence-taggedsite (STS) markers. As a result, 252 YAC clones were confirmedto contain the mapped DNA fragments on both chromosomes. A contigmap was constructed by ordering these YAC clones and about 53%and 43% genome coverage was obtained for chromosomes 8 and 9,respectively, assuming a YAC clone size of 350 kb and overlapbetween neighboring YACs of 50%. A continuous array of YAC cloneswith minimum overlap gave a total size of 18.9 Mb for chromosome8 and 15.6 Mb for chromosome 9, which are close to previousestimates. These contig maps may provide valuable informationthat can be useful in understanding chromosome structure andisolating specific genes by map-based cloning.     

Direct construction of a chromosome-specific NotI linking library from flow-sorted chromosomes.

A linking library consists of genomic DNA fragments which contain a specific rare restriction enzyme site; such clones are very useful as probes in pulsed field gel electrophoresis and in mapping and cloning large regions of DNA. However, identifying those linking clones which map to a certain chromosomal region can be laborious. Therefore, we have developed a straightforward procedure for constructing a linking library directly from flow-sorted chromosomes. As a test of the approach, a NotI linking library was constructed from the chromosome 17 fraction of a flow-sort of human chromosomes, using only 70 ng of DNA. Thirteen of sixteen linking clones were mapped to chromosome 17, suggesting that the library is highly enriched for this chromosome. This method should be generally applicable to other chromosomes and enzymes as well.     

Improved separation of chromosome-sized DNA from Trypanosoma brucei, stock 427-60.

Separation of chromosome-sized DNA from the parasitic protozoan Trypanosoma brucei had previously resulted in the fractionation of DNA molecules that ranged in size from 50 kb up to roughly 1.5 Mb. The number of larger chromosomes and their size, accounting for 80% of the DNA of T. brucei remained unclear. We have now size separated these larger DNA molecules by pulsed field gel electrophoresis (PFG) and resolve a total of 20 bands, accounting for roughly 120 chromosomes, ranging in size from 50 kb up to the size of the largest, 5.7 Mb chromosome of Schizosaccharomyces pombe. Three different VSG gene expression sites were located to chromosomes of 430 kb, 1.5 Mb and 3 Mb, respectively. We have not been able to identify additional, previously cryptic DNA rearrangements, that could explain the activation or inactivation of the expression sites.     

Construction of a Helicobacter pylori genome map and demonstration of diversity at the genome level.

Genomic DNA from 30 strains of Helicobacter pylori was subjected to pulsed-field gel electrophoresis (PFGE) after digestion with NotI and NruI. The genome sizes of the strains ranged from 1.6 to 1.73 Mb, with an average size of 1.67 Mb. By using NotI and NruI, a circular map of H. pylori UA802 (1.7 Mb) which contained three copies of 16S and 23S rRNA genes was constructed. An unusual feature of the H. pylori genome was the separate location of at least two copies of 16S and 23S rRNA genes. Almost all strains had different PFGE patterns after NotI and NruI digestion, suggesting that the H. pylori genome possesses a considerable degree of genetic variability. However, three strains from different sites (the fundus, antrum, and body of the stomach) within the same patient gave identical PFGE patterns. The genomic pattern of individual isolates remained constant during multiple subcultures in vitro. The reason for the genetic diversity observed among H. pylori strains remains to be explained.     

Construction and characterization of a partial library of yeast artificial chromosomes from human chromosome 21

We report a protocol for cloning large DNA fragments in yeast artificial chromosomes (YAC). A partial library has been constructed from a somatic hybrid containing chromosome 21 as the single source of human DNA. About 4.0 Mb of human DNA was recovered in 17 YAC clones. Three clones were analyzed by in situ hybridization and mapped on chromosome 21. One clone hybridized with the chromosome 21 centromeric region and may provide new insight both on the molecular structure of centromere and on the localization of Alzheimer disease gene.     

A complete physical map of a Bacillus thuringiensis chromosome.

Bacillus thuringiensis is the source of the most widely used biological pesticide, through its production of insecticidal toxins. The toxin genes are often localized on plasmids. We have constructed a physical map of a Bacillus thuringiensis chromosome by aligning 16 fragments obtained by digestion with the restriction enzyme NotI. The fragments ranged from 15 to 1,350 kb. The size of the chromosome was 5.4 Mb. The NotI DNA fingerprint patterns of 12 different B. thuringiensis strains showed marked variation. The cryIA-type toxin gene was present on the chromosome in four strains, was extrachromosomal in four strains, and was both chromosomal and extrachromosomal in two strains. A Tn4430 transposon probe hybridized to 5 of the 10 cryIA-positive chromosomal fragments, while cryIA and the transposon often hybridized to different extrachromosomal bands. Ten of the strains were hemolytic when grown on agar plates containing human erythrocytes. Nine of the strains were positive when assayed for the presence of Bacillus cereus enterotoxin. We conclude that B. thuringiensis is very closely related to B. cereus and that the distinction between B. cereus and B. thuringiensis should be reconsidered.