Hum. Genet. 40, 1--72 ((cid:127)977) Review Articles © by Springer-Verlag 1977 Molecular Cloning of DNA An Introduction Into Techniques and Problems Hans-Peter Vosberg* Max-Planck-Institut fiir medizinische Forschung, Abt. Molekulare Biologie, Jahnstrasse 29, D-6900 Heidelberg, Federal Republic of Germany Contents Summary ............................ 1 Introduction ..... ................................ 2 The principles of DNA Cloning Experiments 4 Restriction Endonucleases .................. .......................... 7 The Cloning Vectors: ............................ 13 Bacterial Plasmids ........................... 13 Lambda DNA ............................. 20 Other Vectors ............................. 25 In vitro Construction of Recombinant DNA 27 Transfection and Transformation ....................... 32 Selection of Cloned DNA .......................... 33 The Functional Expression of Cloned DNA 36 Genes and DNA Segments Cloned 42 Biohazard Considerations .................. .... Conclusion .......................... 53 References ................................. 57 ................................. 61 Summary. Biochemical, biophysical and genetic studies of DNA segments of complex genomes are greatly facilitated by a variety of techniques, called molecular cloning of DNA, which permit propagation of single DNA seg- ments of virtually any origin in bacterial cells. Molecular cloning requires in vitro recombination of DNA fragments with a prokaryotic genetic element (a plasmid or a bacteriophage DNA) which serves as replication vehicle (also called 7eector) in the bacterial host. The experimental conditions allow the production of bacterial clones each harboring a single fragment of exogeneous DNA out of an initially highly heterogeneous mixture of fragments. The following steps are involved: foreign DNA is first dissected into small pieces of up to some 17 000 basepairs in length. The fragments are then joined in vitro to the vector molecules by means of well characterized DNA enzymes. The resulting recombinant molecules are introduced into the host cells by a To whom offprint requests should be sent
`
`Page 1 of 72
`
`BLUEBIRD EXHIBIT 1018
`
`

`

`2 H.-P. Vosberg transformation or transfection step. Among the progeny of transformed or transfected cells those clones are selected which carry a fragment of interest. Selection is in most cases accomplished by a combination of genetic and physical methods and is based on properties of the vectors as well as on attributes of the cloned foreign DNA. It is anticipated that bacterial host cells are not only suitable for amplifying DNA but also for the expression of useful functions which originate from other, preferably higher organisms. Two questions cannot be answered, conclusively at present: first, is functional expression of genes generally possible in a heterologous cellular environment, and second, if it is possible, is it always harmless or does it create, at least occasionally, a biological hazard. Besides a detailed description of the techniques developed for molecular cloning, problems connected with func- tional expression and biohazards are discussed. In addition, results are presented which were obtained in the recent past by applying DNA cloning procedures. Introduction New experiences in molecular genetics and in biochemistry of nucleic acids have initiated experiments of possibly far reaching influence on scientific knowledge and technical developments in different areas. It has to a certain degree become feasible to recombine genes or parts of genes in vitro in arrangements which do not naturally exist, and to introduce them into living cells in such a way that the DNA is correctly replicated in the host. These recombination experiments have become known as "molecular cloning of DNA" or "recombinant DNA research". Most of the experiments dealing with recombinant DNA are, at present, performed on the level of prokaryotic organisms. These organisms participate in DNA exchange and recombination events in vivo. Among the most powerful molecular tools in recombinant DNA research are bacterial plasmids and bacteriophages. Both are known for their ability to promote gene transfer and recombination in prokaryotes. DNA recombination in prokaryotic organisms is not yet fully understood. It appears that a variety of mechanisms exist. One of them, designated general recombination, operates on the basis of nucleotide homology and requires in E. coli the function of the recA gene (see Clark, 1973; Hotchiss, 1974). In addition, one or more recA-independent systems are known which are involved in non-homologous or "illegitimate" recombination. The latter category operates with a variety of discrete DNA sequences which can be inserted into a bacterial genome at many sites. Since these sequences are able to move from one position in a genome to another one or even to another genome, they were designated translocatable (or transposable) genetic elements. Host genomes, plasmids and bacteriophage DNAs participate in the translocation of DNA in prokaryotes. At least three kinds of transposable elements have been identified: insertion se- quences (IS), transposons (Tn) and certain bacteriophages (for recent reviews see Starlinger and Saedler, 1976; Kleckner, 1977). The bacteriophage lambda which is in use as cloning vector may itself be considered a complex translocatable
`
`Page 2 of 72
`
`

`

`Molecular Cloning of DNA 3 element (Kleckner, 1977). Transposons enclose genes for antibiotic resistance. They play a major role as selection markers for DNA cloning with plasmids. Insertion sequences are known to cause mutations in bacterial genomes through insertion and excision, events of low specificity with respect to their target sites in the genomes. In this article insertion sequences are not considered further. DNA cloning is an experimental procedure which extends the "host range" for DNA of unrelated species in a very significant way. Exchange of genetic material between cells occurs in vivo, occasionally even between cells of different species. Some well known cases of plasmid mediated gene transfer cover a broad range of distantly related or even unrelated prokaryotic organisms (Datta et al., 1971). Cloning of DNA, however, goes beyond the limits of biological systems. It makes "gene transfer" possible between cells of species which most probably do not exchange genetic material in normal life. The stable maintenance of an eukarYotic gene in a prokaryotic organism is, at least to our knowledge, without precedence in nature. In most cloning experiments bacteria are the hosts for foreign DNA. With respect to their biological characteristics these microorganisms are sometimes referred to as "hybrid organisms" or even as new "biotypes". These qualifica, tions may not be regarded adequate. An E. coli cell which harbors a plasmid to which a piece of foreign DNA is covalently connected adds probably less than one thousandth to its own genetic complexity and in many cases the additional information will be more or less silent. Therefore the E. coli cell will at best acquire one or a few new properties, but this should not affect its identity as a bacterial organism or its forming part of a particular species. It is more appropriate to refer to the recombinant plasmids and phage DNAs as "chimeric" or "hybrid" DNA since the contribution of the foreign DNA segment to the vector genome, which promotes replication in the host cell, is significant at least with respect to the amount of DNA added. Amplification and purification of DNA segments from complex genomes are the most noticeable results obtained so far with molecular cloning of DNA. The benefits of these results are obvious. A more detailed understanding of the structure and organisation of genes and genomes of all living organisms will be possible. Many still open questions related to replication of chromosomes, activa- tion or repression of genes during growth, development and differentiation of cells and organs will probably be answered in the near future. The advantages of molecular cloning will eventually contribute to the understanding of pathological conditions, including hereditary diseases. The prospects go, however, beyond the advancement of theoretical knowl- edge. The "implantation" of favorable hereditable properties into the genomes of higher (or lower) organisms are seriously considered together with its use for medicine or in areas of agriculture and industry. Some possibilities of useful applications will be discussed in the last chapter. The general discussion about the consequences of recombinant DNA research includes the biological risks which may accompany the expected benefits. It should be pointed out and will be discussed in some detail later, that at present it is unknown what the risks are and whether there is a real basis for the assumption that potentially hazardous organisms could be created in vitro.
`
`Page 3 of 72
`
`

`

`4 H.-P. Vosberg Molecular cloning of DNA should be distinguished from two other experimen- tal designs both dealing with gene transfer initiated by in vitro manipulations and bypassing natural transfer barriers: the physical injection of either intact nuclei into enucleated eggs from a frog (Gurdon, 1968) or of DNA into the nuclei of intact oocytes (Mertz and Gurdon, 1977), and the biochemical transformation of animal tissue culture cells by the uptake of defined DNA segments or genes of viral origin (e.g. Bacchetti and Graham, 1977). In the case of injected nuclei no selection with respect to a distinct part of the genome is intended. In the two other cases no vector is involved and the transferred DNA is neither amplified nor recovered. The goal in these experiments is the study of functional expression of extraneous DNA by investigating the development of a whole organism or isolated transcription and translation events, or the transformation of a genetical- ly defect cell line. These approaches point to useful extensions of molecular cloning, but they are not part of it. Recombinant DNA research is a young discipline and moving fast. New and important results are coming up almost every month and the number of investigators in this area is still growing. The emphasis of the author, therefore, is not to present a complete survey of the results obtained with DNA cloning to date, but rather to provide first, a detailed description of the technical and procedural prerequisites of cloning experiments and, secondly, the discussion of some essential aspects related to functional expression of cloned genes and, in addition, biohazard problems. Since the new techniques are interesting not per se but for the results which were not available without them, a separate chapter is enclosed which attempts to give a representative view on existing results. This survey is presented in the form of tables of the genes and DNA fragments reported as cloned. For a variety of results additional information is given about the goals of individual cloning experiments. The purpose of this presentation is to provide a summarizing outline about what kind of experiment can be done with molecular cloning and what type of result may be expected in the near future. The Principles of DNA Cloning Experiments Amplification and cloning of DNA fragments are the result of an experimental procedure in the course of which pieces of DNA of different origin are joined together in vitro, to composite molecules. These "hybrid" DNA molecules are able to infect some host cell and to replicate autonomously within the host after infection. Such an experiment is, in the most general terms, characterized by the transfer of genetic material from an environment of high complexity into one of low complexity. The genome of an E. coli cell is estimated to contain up to 5000 different genes. Most corresponding estimates for higher cells range somewhere between 5000 and 50,000 or even more genes. Thus a DNA segment from the genome of any living cell containing a gene or at most a couple of genes comprises a very small part of the total genetic information. Plasmids or viral DNAs to which a DNA segment may be transferred are small, yet autonomous, genetic elements containing not more than approximately 30 genes, in many cases
`
`Page 4 of 72
`
`

`

`Molecular Cloning of DNA 5 even less. This transfer, depending on the plasmid or viral DNA used, is accompanied by a very significant increase in the number of copies per cell. This increase is the amplification component in the experiments to be described. The second component, the cloning, is concerned with the purification of one DNA segment out of a mixture containing a large number of different segments. Cloning results from the fact that under suitable conditions an infected cell contains only one composite DNA molecule and, hence, amplifies only that DNA segment which is part of the infecting small "hybrid" genome. Since single bacterial clones are easily detected and grown to large quantities purification and amplification may be considered as two results from one experiment. In effect, bacterial cells perform what cannot be achieved in vitro, they produce large numbers of homogeneous copies of DNA segments of virtually any origin. Two different DNA elements are brought together in these experiments, the vector (or vehicle) providing the apparatus for autonomous replication and some genetic marker facilitating detection of amplified DNA clones in a population of infected host cells, and the DNA segment of interest. Small DNA molecules serve as vectors, as for instance bacterial plasmids or the DNA of the bacteriophage 2. The vectors are either covalently closed circular DNA molecules (plasmid) or linear DNA molecules (2 DNA). Insertion of a piece of foreign DNA into a vector genome requires cleavage of the latter at a site where insertion destroys neither the capacity to replicate autonomously nor the function which serves later as selective marker. The foreign DNA segments which are to be cloned are obtained after isolation from their natural sources by controlled mechanical or enzymatical cleavage. Coupling to the vector molecules is in most experiments achieved by the spontaneous formation of hydrogen bonds between complemen- tary single-stranded extensions at the 5' ends of the cleaved vectors and the foreign DNA segments, followed by covalent joining through the action of DNA ligase. In the case of plasmids the hybrid vector (or recombinant DNA) results from the correct connection of two components forming a new circular DNA molecule. In the case of lambda DNA three fragments (two 2 fragments at either side of the foreign DNA segment), constitute the recombinant DNA. Pro- pagation of the composite DNA occurs in most experiments within cells of the same species from which the native vector was isolated. The procedure leading to the uptake of recombinant DNA by the host cells is called transformation for plasmids and transfection for bacteriophage DNA. The general scheme of a cloning experiment is illustrated in Figure 1. Essentially three different possibilities exist for obtaining DNA fragments from complex genomes through cloning. The first approach starts with a DNA fragment which is homogeneous when it is coupled to the vector, i.e. purification is preceding amplification and cloning. The specific advantage of this approach is that only minor problems arise in selecting the amplified fragment in the host cell. It is limited, however, to the rare cases where a sufficiently purified DNA frag- ment is available. Examples are synthetic DNA or DNA fragments obtained after reversed transcription of highly purified mRNAs or DNA fragments which can easily be enriched biochemically due to their high copy numbers in the genome from which they are isolated. In this approach the capacity to amplify DNA frag- ments is emphasized over that to purify them.
`
`Page 5 of 72
`
`

`

`6 H.-P. Vosberg vector foreign DNA 1 restriction endonuciease © ,. • I DNA ligase I transformation Bacterial cell Fig. l. Scheme for a typical DNA cloning experiment. The figure depicts cloning of DNA frag- ments generated with a restriction endonuclease and a plasmid as vector. The enzymes involved are indicated. C: chromosome of the bacterial host cell; P: recombinant plasmids in the bacterial cell. The dots (.) indicate the 5'-ends of the DNA fragments. Details for his type of experiment are described under "In vitro Construction of Recombinant DNA" The second approach starts with cleaving the genome containing the DNA segment looked for either randomly by ultrasonication or other mechanical means or, more specifically, with restriction endonucleases. The numerous un- selected (or partially selected) fragments are subsequently spliced to a suitable vector. A particular fragment is identified after propagation of the whole mixture of recombinant DNA molecules by genetic selection based on the expression of its function in the host. This procedure is applicable only for fragments carrying information which is expressed in vivo, too, and is furthermore restricted to genes which are functionally expressed in the host used for cloning. Genetic com- plementation may help to identify bacterial genes in bacterial hosts, but not necessarily eukaryotic genes in bacterial hosts.
`
`Page 6 of 72
`
`

`

`Molecular Cloning of DNA 7 Experiments with unselected sets of fragments are termed "shotgun" cloning. They are subject to specific restrictions due to their inherent potential biohazard risks. The third method is based on random cleavage of whole genomes, too, but uses a different selection procedure in the end. Specific fragments are selected after cloning by physical screening with radioactively labeled messenger RNA (mRNA) or complementary DNA (cDNA). This method requires some manip- ulation starting with lysis of the cells harboring clones of recombinant DNA, followed by DNA denaturation and hybridization of the labeled probe to the DNA in situ. This approach is limited only by the availability of RNA or DNA complementary to the fragment wanted. It is highly efficient and provides, furthermore, a high degree of biological safety because is is simple and only small amounts of material are required (Grunstein and Hogness, 1975; Jones and Murray, 1975; Benton and Davis, 1977). At present this is the most widely used technique for identifying cloned DNA fragments. The two latter approaches equally emphasize DNA amplification and puri- fication by cloning. In the majority of experiments reported so far, DNA was cloned in bacterial hosts with a prevailing interest in the DNA synthesizing capacity of the cells. An important implication of these experiments is that the nucleotide sequences of the foreign DNA fragments are correctly maintained. Whether this assumption is justified will be discussed. Restriction Endonucleases The major step forward in the development of gene cloning techniques was the discovery of a bacterial DNA endonuclease able to recognize and to cleave DNA at specific nucleotide sequences (Smith and Wilcox, 1970; Kelly and Smith, 1970). This enzyme, later designated HindlI 1 (see below), was isolated from the strain Haemophilus influenzae, serotype d. It belongs to the large group of bacterial enzymes termed DNA restriction endonucleases (or restriction enzymes). The discovery of the Haemophilus enzyme and its immediately following use in cleaving the DNA of the tumor virus SV40 into a set of distinct fragments (Danna and Nathans, 1971) laid the ground for all attempts to obtain specific fragments of genomes by a biochemical procedure. The term restriction endonuclease origi.nated from the genetic analysis of bacteriophage-cell interactions first presented by the group of Luria (Luria, 1953) to explain the phenomenon of host-cell induced variation. These authors observed that bacteriophages grown on one strain of bacterial hosts showed a considerable variation in their capacity to grow on other strains. They postulated a cellular mechanism able to impose a biochemical modification upon the phage which was subject to variation. Later, a more complicated model was proposed according to which the phenomenon of variation was the result of modification and restriction, with the term restriction describing the ability of certain bacterial 1 Restriction endonucleases are abbreviated according to the generally used code (Smith and Nathans, 1973). The bacterial strains from which they are obtained are either listed in Table 1 or mentioned in the text together with the enzymes
`
`Page 7 of 72
`
`

`

`8 H.-P. Vosberg strains to reject infecting phage DNA by degrading it (Arber and Dussoix, 1962; Dussoix and Arber, 1962). Two distinct enzymes should be responsible for the underlying mechanisms, one being a DNA endonuclease (i.e. the restriction enzyme ) able to cleave DNA by recognizing defined nucleotide sequences and a second one being a modification enzyme recognizing the same sequence but modifying it in such a way that it is no longer substrate to cleavage by the restricting enzyme. DNA of a cell harboring a restriction endonuclease would always have to be modified in order to prevent digestion by its own enzyme. The biological role of restriction-modification systems is to protect a bacterial cell against invasion of foreign DNA. A variety of microorganisms were identified as possessing such a protective mechanism (for reviews see Arber, 1974; Nathans and Smith, 1975). The first isolation of a restriction endonuclease (from E. coh) was reported by Meselson and Yuan (1968) and Linn and Arber (1968). This enzyme was characterized by site specific action and produced a limited number of fragments with unmodified DNA. Substrate recognition by the enzyme was identified as being specific for certain nucleotide sequences, cleavage, however, occured in a random fashion (Murray et al., 1973; Linnet al., 1974; Adler and Nathans, 1973). Restriction enzymes characterized by this behavior are classified as -class I restriction enzymes. The best known examples of class I enzymes are the restriction enzymes of the E. coli strains B and K (Meselson and Yuan, 1968; Linn and Arber, 1968). Their corresponding modification enzymes are known (Meselson and Yuan, 1968; Lautenberger and Linn, 1972; Eskin and Linn, 1972; Haberman et al., 1972). These restriction enzymes require ATP and S-adenosylmethionin as cofactors. The Y-terminal nucleotides of the cleavage fragments are in some way modified (Eskin and Linn, 1972) and the enzymes possess an ATPase activity which is effective after cleavage of the DNA (Eskin and Linn, 1972; Yuan et al., 1972). In contrast to class I restriction enzymes a large and still growing number of restriction enzymes is known which both recognize and cut DNA at specific nucleotide sequences. These enzymes do not require cofactors. They are classified as class II restriction enzymes. The Haemophilus enzyme mentioned above belongs to this class. The classification of site specific DNA endonucleases as restriction enzymes should include their identification as part of a genetically characterized restriction-modification system. For the majority of the class II enzymes such a correlation does not exist, however. This is in part due to their occurence in microorganisms with a genetics barely understood. Despite some evidence (Bron et al., 1975) supporting the notion that class II restriction enzymes operate in vivo as part of a restriction-modification system comparable to that of the class I enzymes, the possibility exist that they have some other function in vivo, for instance, in site specific recombination (see Roberts, 1976). Whatever their biological implication is, it can be assumed that cells harboring these enzymes have some mechanism to protect themselves against DNA breakdown by their own enzymes. For practical purposes the most important difference between class I and class II restriction enzymes is the nature of the cleavage products. Class I enzyme generated fragments always show atypical, heterogeneous size distributions when analysed by gelelectrophoresis, whereas cleavage with class II enzymes results in well-defined distributions of DNA fragments with the individual patterns, depending on the enzyme used. For the analytical and preparative needs of cloning experiments, therefore, only class II restriction enzymes are useful.
`
`Page 8 of 72
`
`

`

`Molecular Cloning of DNA 9 In the following the term restriction enzyme is exclusively used for class II restriction enzymes. For more detailed information about restriction enzymes the review articles by Nathans and Smith (1975) and Roberts (1976) are recom- mended. More than 80 enzymes of the class II-type are known to date. They were detected in strains throughout the bacterial kingdom. The most productive genus is that of Haemophilus, 22 enzymes alone were isolated from 29 different Haemophilus strains tested. The closely related genus Moraxella is also rich with restriction enzymes. Two restriction enzymes not related to the class I-restriction- modification system were obtained from E. cold carrying drug resistance plasmids. These enzymes are designated EcoRI (Yoshimori, 1971; Greene et al., 1974) and EcoRII (Yoshimori, 1971; Boyer et al., 1973) (for the nomenclature see below). The Gram-positive organism Bacillus subtilis harbors the restriction enzyme BsuRI which is coded on the chromosome, but may be part of a cryptic prophage (Arwert and Rutberg, 1974). It is not known how many of the enzymes are coded for by plasmids or phages where they could be used in vivo for excising and recombining DNA fragments in the course of recombination events principally resembling the in vitro procedures described in this article (Roberts, 1976). It is open whether restriction enzymes are present in higher organisms, too. They may be expected in those species or organs where the organisational pattern of DNA changes during development, as it is the case with the excision of ribosomal DNA in Tetrahymena (Gall, 1974) and Stylonichia (Prescott and Murti, 1973), or where drastic changes such as that from micronucleus to macronucleus as in ciliates are observed (Wesley, 1975). A common abbreviated nomenclature for restriction enzymes (Smith and Nathans, 1973) has been accepted.. The enzymes are designated by a three-letter code describing the genus and the species from which they were obtained. If necessary a fourth-letter is added to define the strain more closely. To distinguish between multiple enzymes from one source, roman numerals follow the ab- breviation. Some of the most frequently used restriction enzymes are listed in Table 1. The more than 80 specific restriction endonucleases that have been discovered recognize more than 40 different specific sequences as judged from the pattern of cleavage fragments obtained with DNA of the animal virus adenovirus 2 and DNA of the bacteriophage lambda, both of which serve as standard substrates in restriction enzyme studies. Some 22 of these sequences were identified. Surprisingly, many enzymes from unrelated sources recognize the same nucleotide sequences. Roberts (1976) suggested classifying them as "isoschizomers". The sequence GGCC 2 is recognized by the enzymes obtained from six different strains. Another frequently recognized sequence is AAGCTT (five enzymes, for details see Roberts, 1976). Not all of the sources are equivalent with respect to yield and ease of purification. Therefore, only some of the genera and strains are of practical interest. Identity in recognition sequences within one group of isoschizomers is normally ac- companied by identical cleavage positions within the sequence characteristic for that group. It has been shown for the group HaplI (Haemophilus aphrophilus), HpalI and MnoI (Moraxella nonliquefaciens) that the site of cleavage within the sequence CCGG is between the two C. Similarly, HaeIII and BsuRI both cleave within their common recognition sequence GGCC 2 The four deoxyribonucleotides are abbreviated in the usual way as A, C, G,T. Nucleotide sequences are presented in 5'-3' direction
`
`Page 9 of 72
`
`

`

`10 Table 1. Some selected restriction enzymes (class II) H.-P. Vosberg Enzyme Microorganism Sequence a R,M b 2 c Reference AluI Arthrobacter luteus AGCT R >50 Roberts et ah (1976) 1 BamI Bacillus amyloliquefaciens GGATCC R 5 Wilson and Young (1975) 1 BsuRI Bacillus subtilis strain R GGCC R >50 Bron et al. (1975) EcoRI Escherichia coil RY 13 GAATTC R,M 5 Greene et al. (1974) HaeIII Haemophilus aegytius GGCC R >50 Middleton et al. (1972) HindII Haemophilus infiuencae R a GTPyPuAC R,M 34 Smith and Wilcox (1970) 1 HindlII Haemophilus influenzae R a AAGCTT R,M 6 Smith and Wilcox (1970) | HpalI Haemophilus parainfluenzae CCGG R >50 Sharp et al. (1973) SalI Streptomyces albus GTCGAC a R 2 Arrand et al., quoted in Roberts (1976) SstI Streptomyces stanford ? R 2 Goff and Rambach, quoted in Roberts (1976) XmaI Xanthomonas malvacearum CCCGGG R 2 Endow and Roberts (1977) a Sequences are written in 5'-3' direction and only one strand is indicated. The complete sequence of both strands is read as shown in Figure 2 for the enzymes EcoRI and HaelII. The arrows designate the cleavage site within the sequence b Enzymes indicated R have been identified as restriction endonucleases, whereas those designated R, M are in addition involved in host modification of DNA (methylation) c This entry defines the number of cleavage sites in 2 DNA for the individual enzymes d This sequence has been identified by Arrand, J. R., Myers, P. A. and Roberts, R. J. (unpublished) between C and G (for references see Roberts, 1976). So far, one case is known where iso- schizomers produce fragments with different end groups: XmaI recognizes CCCGGG and cleaves between the left two C giving rise to a fragment with a 5'-tetra-nucleotide extension. The enzyme SmaI (Serratia marcescens), however, recognizes the same sequence but cleaves between C and G, thus producing fragments with blunt ends (Endow and Roberts, 1977). In regarding the requirements of cloning experiments the most important questions with respect to restriction enzymes are (i) how frequently does a particular cleavage sequence occur in the DNA or, in other words, what is the average fragment length produced with a specific restriction enzyme, and (ii) what is the composition of the endgroups of restriction fragments? The answer to the first question affects the choice of enzyme for a particular cloning purpose, whereas the answer to the second question determines the experimental setup for the in vitro recombination of DNA fragments and vector DNAs.
`
`Page 10 of 72
`
`

`

`Molecular Cloning of DNA 11 With respect to base composition, in a first approximation it may be assumed that the longer and the more complex a recognition sequence is, the less frequently it occurs in the DNA. A direct reflection of the relative frequencies of cleavage sequences is the number of sites present in lambda and adenovirus 2 DNA: whereas EcoRI (recognition sequence GAATTC), cleaves both DNAs at five different sites, HpalI (CCGG), cleaves both DNAs more than 50 times. For EcoRI an average fragment length of about 4000 basepairs has been estimated (Mertz and Davis, 1972; Hamer and Thomas, 1975). The enzymes BamI (GGATCC) and SalI (GTCGAC) cleave DNA less frequently, releasing frag- ments of 6000 and 8000 basepairs in length respectively (Hamer and Thomas, 1976). If isolation of intact genes or even clusters of genes is intended EcoRI or another endonuclease with infrequent cleavage sites would be used. The detailed physical analysis of genes after successful cloning will then be achieved with enzymes generating shorter fragments. Another area where short fragments and hence restriction enzymes cutting more frequently, e.g. HpalI, are generally required is in DNA sequencing (see below). Because of the large number of different recognition sequences a considerable variation in the kind of endgroup base compositions is expected. Restriction enzymes either cut the DNA within the recognition sequence in such a way that staggered single-stranded extensions are produced, as it is observed with EcoRI, or they cut the DNA in a non-staggered fashion releasing fragments with blunt ends as obtained with HaeIII (see Fig. 2). Examples of fragments with single- stranded 5'-extension of five (EcoRII), four (EcoRI), three (Hinfl) (Haemophilus influenzae Rf) and two (HpalI) nucleotides are observed as well as 3'-extensions of four (PstI) (Providentia stuartii 164) and two (HhaI) (Haemophilus haemolyticus) nucleotides. A unique feature of r~striction