Transgenic farm animals — A critical analysis

Theriogenology

38:337-357,

1992

TRANSGENIC FARM ANIMALS - A CRITICAL ANALYSIS R.J. Wall’ and G.E. Seidel, Jr.’ ’ U.S. Department of Agriculture, Agricultural Research Service, Beltsville Agricultural Research Center, Gene Evaluation and Mapping Beltsville, Maryland 20705, USA ’ Colorado State University, Department of Physiology, Animal Reproduction Laboratory, Fort Collins, Colorado 80523, USA

Lab.,

ABSTRACT The notion of directly introducing new genes or otherwise manipulating the genotype of an animal is conceptually straightforward and appealing from the standpoints of both speed and precision with which phenotypic changes can be made. Thus, it is little wonder that the imagination of many animal scientists has been captivated by the success others have achieved in introducing foreign genes into mice. Transgenic mice not only exhibit unique phenotypes, but they also pass those traits on to their progeny. However, before transgenic farm animals become a common component of the livestock industry, In this review we attempt to a number of formidable obstacles must be overcome. identify the critical issues that should be considered by both those currently working in the field and those scientists considering the feasibility of initiating a transgenic livestock project. The inefficiency of producing transgenic animals has been well documented, This does not constrain investigators using laboratory animal models, but it has a major impact on applying transgenic technology to farm animals. The molecular mechanisms of transgene integration have not been elucidated, and as a consequence it is difficult to design strategies to improve the efficiency of the process. In addition to the problems associated with integration of new genes, there are inefficiencies associated with collecting and culturing fertilized eggs as well as embryo transfer in farm animals. Transgenic farm animal studies are major logistical undertakings, Even in the face of these practical hindrances, some may be pressured by administrators to embrace this new technology. As powerful as the transgenic animal model system is, currently there are limits to the kinds of agricultural questions that can be addressed. Some uses are so appealing, however, that several commercial organizations have explored this technology. Within the next decade or two, it is likely that many of the technical hurdles will be overcome. Combining new techniques with a better understanding of the genetic control of physiological systems will make it Possible to improve the characteristics of farm animals in highly imaginative ways. Key words:

transgenic, animal, livestock, genetic engineering INTRODUCTION

A transgenic animal might be defined as one whose genetic make-up has been modified by addition or deletion of a specific DNA sequence. The altered chromosomal An animal resulting from a transgenic DNA is transmitted to future generations.

Copyright

0

1992

Butterworth-Heinemann

Theriogenology

338

manipulation is termed a founder animal; about a third of founders transmit the new gene to less than 50% of their offspring because not all cells of the founder acquired the new gene. Descendants of founders will either have the new DNA modification in all of their somatic and pre-meiotic germ cells or in none of them. Transgenic procedures are extremely useful for some kinds of experiments. During the years 1980-1982, a few dozen scientific papers concerning this procedure with mammals were published; the number of such publications now exceeds 2,500. Some of this activity may be due to fashion and fundability, but this approach also is truly a powerful analytical tool. DISCUSSION Advantages of Transgenic Animals over Transfected Cell Lines A substantial percentage of current research in the biological sciences concerns regulation of transcription of mRNA from DNA. Examples are understanding natural or pathological transcription and modifying regulation for some useful or interesting purpose. Most of these studies are performed by transfection of cells in vitro. Such procedures usually are relatively rapid and simple and are the method of choice for most studies. For some studies, however, cells in vitro lack context. The great advantage of transgenic animals is that gene regulation can be studied or exploited in both a developmental and a whole body context. This adds the complexities of whole body physiology, but if the objective is to understand how things happen in nature (e.g., pathogenesis) the whole animal has great advantages. Another advantage is that transgenic constructs can be studied in various genetic contexts, for example, in males versus females, in hemizygous versus homozygous form, A particularly elegant example was the and bred into various genetic backgrounds. demonstration that cytosine methylation patterns in an imprinted construct varied due to sex of parent (but not sex of animal) (1). The Same Transgenic Construct Can Lead to Very Different Transgenic Animals A particularly interesting aspect of transgenic studies is site of integration effects in different transgenic lines from the same DNA construct. Such effects primarily concern the amount of transcription, from very little or none to normal or excessive amounts. A good example is the “shiverer” mouse model in which a transgene for basic myelin protein (2) led to a graded effect, ranging from no benefit to complete normalcy, depending on transcription rates. Because of the expense in characterizing a strain of transgenic animals, investigators tend to home in on the strains that are performing in interesting ways. Transgenic Mice Lead the Way When the team of Ralph Brinster and Richard Palmiter demonstrated that a fusion gene, consisting of regulatory and structural elements from different genes, could be

Theriogenology

339

controlled by dietary additives, the scientific community began to real&e. the enormous potential of transgenic technology (3). Their work not only demonstrated that it was possible to introduce an inducible gene, but also that it was possible to bypass physiological feedback loops (4). They and others have investigated issues in fields as diverse as molecular biology, cancer, reproduction, growth, globin switching, embryo development, mammary gland function, immunology, disease models, and neurology. The successful exploitation of transgenic mice (current publication rate exceeds 500 per year; 5) has stimulated an interest in applying this technology to large animals. The potential value of applying transgenic animal technology to livestock remains speculative, however. The limited participation by the scientific community and industry is not a reflection of lack of acceptance of the concept, but is due, rather, to the impact of technical and financial constraints. As with any emerging new technology, it is difficult to predict which hurdles will be easily vaulted and which will remain insurmountable for an extended period of time. Efficiency of Producing Transgenic Animals is Low The agricultural appropriate mouse can transgenic (6,7,V).

critical issues that impede application of transgenic animal technology to problems can be grouped into two general categories: availability of genes and inefficiencies in producing transgenic livestock. A transgenic routinely be produced from injection of 100 zygotes, whereas production of cattle, goats, pigs or sheep is at least an order of magnitude less efficient

Most livestock ova survive the microinjection process, as assessed by their morphology prior to transfer to recipients. However, under the best conditions, only about 25% of the transferred ova survive to term, and only about 5% of the neonates contain the introduced gene. Thus, the two factors that reduce efficiency of producing transgenic animals are low embryo survival rate and low incidence of transgene integration. Costs arising from these inefficiencies are acceptable for transgenic laboratory animals but are a serious impediment to production of transgenic livestock. The Mechanism of Transgene Integration is Unknown The frequency at which injected genes integrate into the genome is currently measured as a proportion of newborns in which the gene is detected. No method has yet been devised to ascertain whether that parameter accurately reflects the fate of the injected DNA during embryonic development. It is likely that the viability of at least some embryos may be compromised by integration of transgenes; therefore, the measured integration frequency is an underestimate. Transgenes can be detected in approximately 30% of injected mouse embryos cultured to the blastocyst stage. That value is approximately double the proportion of transgenic neonates produced with the same gene construct (Burdon and Wall, unpublished data). should be noted that the method used to evaluate the blastocysts could not distinguish between integrated and unintegrated

340

Theriogenolog

y

transgenes. Therefore, some of the blastocysts in which transgenes were detected were probably not transgenic. Data from this study suggests that embryonic losses due to integration of transgenes is less than 50%, and integration frequency is 30% or less in mice. Even though improving integration frequency would have a significant impact on the overall efficiency of producing transgenic livestock, this problem is not being actively investigated. The lack of intense interest may be partly because the vast majority of investigators who employ transgenic animal technology use the mouse as their animal model. Efficiency of producing transgenic mice is also low, but costs are within an acceptable range; consequently, there is little motivation to address the efficiency issue. The task of improving integration efficiency is further complicated by the fact that the mechanism by which transgenes insert themselves into the genome is not understood. Without knowledge of the mechanism of integration, researchers interested in improving efficiency must either focus on elucidating the mechanism, in the hopes of identifying means of intervention, or take an empirical approach. When does transeene integration occur? It is widely assumed that, for non-mosaic transgenic animals, integration of exogenous DNA occurs prior to DNA synthesis during the S phase of the first cell cycle and, for transgenic mosaics, between DNA syntheses of the first and second cell cycles (10). Remarkably, there is little direct evidence to support this interpretation. The following analysis of the way murine embryos develop leads to the possibility that integration of constructs for many non-mosaic transgenic animals could occur well after the first DNA synthesis. At the 32-cell stage of mouse embryos, there are about 10-l 1 cells in the inner cell mass and 21-22 cells in the trophectoderm (11). The murine inner cell mass is the precursor to the fetus (11, 12). One study involving the formation of chimeras proves that at least four cells were embryonic precursors, although this finding may be atypical due to the huge blastocysts resulting from aggregating four embryos (13). In any case, it appears that the upper limit is that about 20% of early blastocyst cells contribute to the mouse fetus (some inner cell mass cells develop into non-fetal tissue). A second point is that for most embryos, several inside cells at the morula stage may be the only precursors to the fetus (11). In most cases, it is likely that all cells of the fetus arise from only one cell of two-cell embryos. The blastomere that divides first to form a three-cell embryo has a greater tendency to form inner cell mass cells than the one that divides second, and similarly for divisions of four- and eight-cell embryos (see 11, for review). This concept is demonstrated in Fig. 1, which illustrates the effect of a 15% longer cell cycle (heavy lines) on timing of blastomere formation. The first and second cell cycles are 20 h, and basal cell cycle for subsequent divisions is 10 h Clearly, the two to four cells of the eight-cell embryo forming the fetus could arise exclusively from the cells in the left half of Fig. 1 in some or even most animals. Additional evidence for this theory is that when precompaction embryos are aggregated to form chimeras, many resulting fetuses are not chimeric.

Theriogenolog

y

341

In the hypothetical examples in Fig. 1, transgenes, represented by shading within the blastomeres, are shown integrating at the two-cell stage (diagonal lines) and four-cell stage (solid and stippled). Because the two-cell integration event occurred in a blastomere that experienced delayed cleavage, as did one of its daughter blastomeres, the probability that the resulting mouse will be transgenic is low. In one of the examples of integration at the four-cell stage (solid), however, the resulting individual would likely be transgenic since no delay in cleavage was experienced by that cell lineage. It must be pointed out that this scenario is purely hypothetical but not unreasonable. Until a means is developed to detect integration of transgenes in cleavage-stage embryos, it will remain unclear at what point integration takes place. We suspect that integration rarely takes place prior to DNA synthesis at the one-cell stage and indeed may occur much later than is commonly believed. It is less clear to what extent fetuses of farm animal species arising from one transgenic and one non-transgenic blastomere would be expected to be nonmosaic, but, broadly speaking, the same principles are likely to hold (14). A oronosed model for transeene intearation. Characteristically, introduced genes integrate into a single site in the genome in a head-to-tail tandem array consisting of a few to several hundred gene copies. Based on this empirical data (10) and tissue culture transfection experiments (15), a multi-step integration model has been proposed. In this model, an array of introduced genes is formed first, and then the array becomes incorporated into the genome at a break in a chromosome (Fig. 2). The array is thought to be formed by homologous recombination between copies of the introduced DNA rather than end-to-end joining of introduced DNA. If random end joining were the predominant mechanism of array formation (concatamerization), head-to-head, tail-to-tail and head-to-tail arrangements would be equally common in transgene arrays. It is rare, however, to find transgene arrays with head-to-head or tailto-tail gene configurations. Furthermore, head-to-tail concatamers are the only possible outcome of legitimate homologous recombination, regardless of whether the introduced DNA is in a linear or circular form. The rare head-to-head and tail-to-tail configurations could result from illegitimate recombination events or end-to-end joining. End joining definitely does occur intramolecularly and is responsible for the circularization of the injected linear fragments. Bishop and Smith (15) argued that concatamerization most likely precedes integration of the transgene array into the genome. This does seem plausible, since copies of the introduced DNA are more likely to come into contact with one another if they are free to diffuse through the pronucleus than if ligation could only take place at a single site. As compelling as the proposed model is, it does not account for the fact that injection of linearized DNA molecules results in the production of a higher proportion of transgenic animals than does introduction of circularized DNA fragments (lo).

Theriogenology

342

DNA injsted

M

into pronucleus

lntramokcukr

and

intemwkcukr

end joining

occurs within minutes of injection Hom~kgous

recombination

belween circles, betvvaen linear nwkcuks

and

between linear and circular molecules all likely to occur

O----

+/c

A% c3

Influence of asynchronous Figure 1. cleavage on fate of blastomeres during murine embroynoic development. Shading represents transgene integrations for three scenarios. See text.

--_

Ligation

of gene to

chromosomal break point achiwed DNA

by normal

repair mechanism

Recombination with ligated copy o, gene extends knglh or array

Un,ntegraled genes are eventually degraded

Figure 2. A possible mechanism for transgene integration into chromosomal DNA. The 5’ end of the transgenes are marked with arrowheads. The XS between molecules represent cross-over events at sites of homologous recombination.

Possible Approaches for Improving Integration Frequency Some empirical approaches that might improve integration below; all are purely speculative.

frequency are listed

- Reformulating the DNA buffer to encourage DNA covalent bonding (ligase enzymes, ions). In addition to ligation enzymes, proteins such as integrase, and DNA sequences such as P-elements that promote gene integration have been identified in invertebrates ( 17,18). - Synchronize the time of injection with the cell cycle of the ova. It is thought that DNA replication is required for the integration process. Currently, most injections are performed toward the end of pronuclear DNA replication. Performing injections before or at the start of replication may be beneficial. - Encourage homologous recombination of introduced genes with the genome by adding DNA sequences that hybridize with repetitive elements of the genome. The homologous recombination event referred to here is different than the recombination that is thought to form the transgene concatamer. This recombination would enhance the

343

Theriogenology

probability of already formed array to become integrated into the genome. A variation of this approach has been demonstrated in transgenic plants with the LoxP-Cre sitespecific recombination system. First, a transgenic line is produced carrying the LoxP DNA sequence. When a second gene containing the Cre sequence is introduced into the LoxP line, the integration occurs at the same site where the LoxP gene originally integrated (19). It has been postulated that a copy of the - Introduce nicks into the genome. introduced DNA becomes incorporated into the genome by normal DNA repair mechanisms. Nicking the genome stimulates DNA repair processes and could provide more loose ends to which the transgene can attach. Such nicking might be accomplished with endonucleases, irradiation, or even physical damage by injecting a large volume of buffer rapidly while injecting the transgenic DNA construct. Use of Retrovirus-derived

Vectors to Make Transgenic Farm Animals

Several laboratories have used retroviral vectors to make transgenic mice and chickens, but there seem to be no publications using such procedures with mammalian farm animal species. Comparisons of this approach to pronuclear injection of DNA are summarized in Table 1. With birds, the retroviral approach has been used because of the impracticality of injection of DNA into pronuclei or nuclei, since eggs of poultry already have tens of thousands of cells when laid. Because large breeding experiments can be done in poultry at reasonable cost, mosaicism and multiple sites of insertion are not as great a limitation as with sheep, swine, or cattle. The major problem in using pronuclear DNA injection to make transgenic farm animals is the low percentage of transgenic offspring. The retroviral approach appears to result in a much higher percentage of transgenic bovine embryos (Bowen et al., unpublished data), but for studies in which a single site of DNA insertion in all cells in the body is required, this method as currently practiced is likely to be inappropriate because of the extra generation interval(s) required. There is, however, the advantage of obtaining multiple lines from most founders (Bowen et al., unpublished data). For studies in which mosaic animals with multiple sites of insertion are acceptable, or even advantageous, the retroviral approach may be ideal. For example, having only a few percentages of cells transgenic may be sufficient to confer disease resistance to whole animals with some mechanisms. Similarly, if the endpoint is secretion of some substance by transgenic cells, this approach might be ideal. This method also might be used to infect the mammary gland to produce pharmaceuticals without a need to make the whole animal or its descendent transgenic. Designing Designer Genes The most critical step in a transgenic animal research project is the design of the gene construct. A gene construct usually consists of a regulatory element (promoter,

344

Theriogenology

enhancer, or both) from one gene ligated to the structural sequence (information required to form the gene product) of another gene. By judicious selection of regulatory elements, an investigator can theoretically direct expression of the structural gene to a specific tissue, determine when the gene will be expressed, and regulate the amount of gene product produced. Table 1. Comnaxison of two methods of makine transnenic animaIs 1

1

1

RetroviraIderived vectors

Pronuclear DNA injection

Species specificity

considerable

none

Concatamers

none

usually

sometimes

rarely

Size of construct

< 7-g kb

no practical limit

Multiple sites of insertion

can be frequent

infrequent

Retroviral expertise

needed

not needed

Microinjection

not required

required

Exposure to blastomeres to virus-producing cells

required

not required

Percentage of embryos transgenic

high

low

Multiple lines from one founder

usual

infrequent

Mosaic founders

usual

about 30%

Need for breeding experiments to separate effects

usual

sometimes

Public acceptance for food production

very low

low?

Item

Non-expression

due to methylation

skills

Unfortunately, there are few guidelines to assist in this process. Usually, the project design begins with the selection of the structural gene component of the fusion gene construct. Because our knowledge of genetic regulation of physiological processes is meager, even selecting the appropriate structural gene sequence is not a trivial matter. Does one choose to alter the expression of a gene at the top of the physiological control hierarchy, at an intermediate control branch point, or at the lowest control point (Le., at the level of the tissue to be affected)? There are no ready answers. Furthermore, once

Theriogenology

345

selection of a structural gene has been made, based on what is known about the control of the particular physiological system, it may turn out that the desired gene has not been cloned and is therefore not available. One is faced with either choosing a less desirable gene or cloning and characterizing the desired gene, a time-consuming alternative. Many scientists feel that it is advisable, whenever possible, to use the genomic form rather than the cDNA (DNA complementary to the messenger RNA produced by the genomic form of the gene) form of the structural gene component of the fusion gene. Molecular biologists are continuing to discover examples of important regulatory elements that reside within introns of genes. Thus, by using cDNA, which has no introns, proper expression of the transgene may be compromised. This concept is supported by a study that directly compared expression of structural sequences with and without introns in mice (20,22) and by a summary of data from transgenic livestock experiments (Table 2). Although there are exceptions, both integration frequency and the proportion of expressing transgenic lines are depressed when the structural component of fusion gene constructs contains fewer than the total number of naturally occurring introns. Choi et al. (21) have used this concept to advantage by adding a “generic intron” to DNA constructs to increase expression in transgenic mice. Table 2. Comparison of form of structural component of transgenes and transgene expression in transgenic livestock experiments’

Number of transgenic founders

Number of expressing transgenic lines

Percentage of expressing transgenic lines

None (cDNA)

40

6

15

One (mini gene)*

21

8

38

All (genomic)

67

50

66

Number of introns

’ Data were compiled from all transgenic livestock publications in which adequate information was available (courtesy of V.G. Pursel). ’ Mini gene is defined as cDNA to which an intron is added or portions of genomic and cDNA ligated together. Whereas selection of the structural gene sequence is dictated by general understanding of the mechanisms that control particular physiological events, the selection of the regulatory sequence is governed by where, when, and how much expression is desired. One must consider whether the cell type to which the expression is directed has the appropriate post-transcriptional processing machinery. Such post-transcriptional processing will determine if the gene product (protein) will survive within the cell, be glycosylated properly, and be secreted if that is desired. If large amounts of the gene

346

Theriogenology

product are desired, one has the choice of selecting a strong promoter or expressing the gene in an abundant tissue type, such as hepatic cells or muscle. It should be clear from the above that transgene design is more art than science at this time. Even though the mouse may not be a perfect model for livestock, it is most efficient and conservative of resources to test transgenes in mice before they are introduced into large animals. Transgenic Studies are Major Logistical Undertakings Numbers of embrvos and animals. The first objective in most transgenic studies is to obtain one or several interesting founder animals. Even with mice, this is a considerable logistical step. Equipment such as specialized microscopes, micromanipulators, and incubators must be available. Personnel must have or acquire expertise in a variety of areas, including recombinant DNA technology, micromanipulation, and reproductive biology. One needs donor strains of males and females, recipient strains of females, and vasectomized males, as well as facilities to house them. Once such a system is in place and functioning well, the additional cost and time involved in making one or several transgenic founders is minimal. However, start-up and overhead costs can be staggering, and success rates can be poor for months or even years! For many studies, costs to characterize the transgenic line are higher than costs to make the transgenic founder. To character& a transgenic line, one needs many animals of both sexes in both hemizygous and homozygous configurations. The transgenic insert can be detrimental to reproduction, even in the hemizygous state (7,23), and one can have insertional mutagenesis such that embryonic death occurs in the homozygous condition. The timing and cause of such death may become the major focus of the characterization. Complex situations such as integration of the construct on an X chromosome with attendant random inactivation in females can also occur. The construct may become imprinted differently in males and females may also occur. Of course, sorting out these effects can be very interesting, but it also can be expensive and divert resources from the original objectives of the project. Snecies differences in costs. Animal costs for transgenic studies will vary tremendously from species to species and from laboratory to laboratory within species. Some of the factors affecting these costs are summarized in Table 3. Whether or not a species is litter-bearing, hence decreasing recipient needs for embryo transfer, is nearly as important as the cost per day for feed and care. Similarly, efficacy of superovulation procedures is very important. The animal care costs for transgenic studies with rabbits will be roughly 10 times those for mice and, for sheep and goats, 10 times those for rabbits. Although per diem costs per sheep may be lower than those per rabbit, other factors in Table 3 are responsible for these species differences. In a few cases, it will be necessary or desirable to maintain transgenic animals in rather secure isolation, which would be extremely expensive for large farm animals.

Theriogenology

347

Use of slaughterhouse oocytes for in vitro fertilization will greatly decrease costs eventually, particularly in cattle and horses. A special advantage of this approach is more control of stage of the cell cycle than occurs with superovulation and in vivo fertilization, with the result that a higher percentage of pronuclei are at an injectable stage. A disadvantage of this approach is that procedures beginning with in vitro oocyte maturation and fertilization result in embryos with markedly lower pregnancy rates to term than in vivo ones. In vitro procedures with gametes and embryos of farm animals are, however, improving rapidly. Table 3. Some factors affecting species differences in costs of producing tmnsgenic

Need to culture embrvos. To avoid surgical embryo transfer in cattle and horses, it is necessary to culture embryos for 2-3 days after microinjection. In vitro culture procedures are marginal for accomplishing this, even with co-culture using oviduct epithelial cells, Buffalo Rat liver cells, granulosa cells, or conditioned media from such cells. This is particularly a problem if in vitro matured and fertilized oocytes are used, since total time in vitro will be 4-5 days minimum. Frequently, this paradigm is further complicated by resorting to culturing embryos in oviducts of intermediate recipients, such as sheep. Unfortunately, longer-term in vitro culture is required in precisely those species in which embryos are difficult to culture. These problems can be overcome, largely by using huge numbers of embryos and selecting the few that develop normally. Nevertheless, these problems exacerbate logistics.

348

Theriogenology

. . Logrstms of characterizinP lines. Determining the effects of a transgene is usually expensive. The first question is whether the transgene is expressed; frequently, it is not (Table 2). Depending on developmental and tissue specificity, it may be quite difficult to determine if low levels of expression are occurring. One of the most difficult questions to evaluate is the extent of desirable or undesirable epistatic effects of transgenes on other genes. The only way to evaluate such effects is to study the transgene in both sexes, at various ages, and in a variety of the genetic backgrounds of interest. Ironically, for experimental purposes, less characterization may be needed than if the transgenic line will be used in production agriculture. The biggest theoretical hurdle in characterizing transgenic livestock is the need to study the homozygous transgenic state for production use. About 10% of the time, the transgene insert disrupts a normal gene, which may have only slight detrimental effects in the hemizygous state, but is generally lethal in the homozygous state. Since millions of dollars have been invested in removing detrimental recessive genes from populations (e.g., dwarfism) for most applications it is essential to guarantee that one is not adding an undesirable gene via transgenic technology. Because no two transgenic founders have the same site of transgene insertion, the only way of obtaining homozygous transgenic animals with current technology is to mate relatives; generally, brother-sister or father-daughter matings are done, resulting in 25% homozygous offspring. Of course, such offspring will be very inbred, which causes decreased growth rates, decreased reproductive rates, and increased susceptibility to disease. For some situations, it will be extremely difficult to determine how much of the decreased performance is due to inbreeding and how much is due to a deleterious but non-lethal insertional mutagenesis event. There are ways of sorting this out (e.g., mating cousins [24]) but they require many animals and a long time in species with long generation intervals. Approaches to Circumvent Logistical Problems with Transgenic Livestock A major expense in making transgenic farm animals is maintaining large numbers of embryo transfer recipients. One method of greatly reducing this cost is to pre-screen embryos for transgenic status before embryo transfer. To accomplish this, microinjected embryos are either cultured in vitro or in reproductive tracts of intermediate recipients until the blastocyst stage. Embryos are then biopsied (25) and the tissue subjected to polymerase chain reaction (PCR) procedures (26,27). The PCR-positive embryos are transferred, and the negative ones discarded. The success of this procedure depends on accuracy of PCR and how severely pregnancy rates are compromised. Preliminary data with bovine embryos from Colorado State University indicate fewer than 3% false negatives (determined by subjecting the whole embryo to PCR after a negative PCR biopsy), 85% true negatives, 10% false positives, and less than 1% true positives. The 10% false positives could be due to technical problems, retention of residual, unintegrated DNA constructs, or mosaicism. Even with 10% false positives, not transferring the 85% true negatives results in tremendous savings. Pregnancy rates after nonsurgical transfer

Theriogenology

349

of these embryos have been around 35 % , which seems quite reasonable considering they have been centrifuged, microinjected, cultured, and biopsied. Embryonic stem cell technology would provide one solution to making animals homozygous. By using homologous recombination, one could make an embryonic stem cell line homozygous to a line that was already tmnsgenic before using these cells to make a chimeric embryo. Currently, embryonic stem cell technology is not sufficiently developed for use with farm animal species. There is also the problem of having a mosaic founder animal in which only a small percentage of gametes would transmit the transgenes to the next generation. Embryos generated by those founders could be analyzed by PCR to minimize complications caused by germ line mosaicism. One other technique that could greatly facilitate transgenic technology is nuclear transplantation of embryonic stem cell nuclei into oocytes. In some cases, this would save waiting a generation to get an animal with the construct in all cells in the body. One might also clone hemizygous embryos made incapable of reproducing and sell these for production purposes. This would ensure that homozygous transgenic animals would not show up in the population, thereby circumventing the need to test the homozygous state; this also would have built-in proprietary advantages. Transgenic Techniques are Wagging Animal Scientists Transeenic technolozv can be oversold. Workers in production agricultural research are frequently and sometimes justifiably criticized for using dated technology in their studies and for not being sufficiently innovative in thinking. Thus, there is a propensity to embrace new and exciting technologies, particularly by administrators, who quite understandably are not always in a position to appreciate the strengths and weaknesses of the technology. Frequently, bench scientists exploit this situation by proposing projects thought to be fundable or satisfying to administrative leaders. A certain amount of this is healthy; too much is counterproductive. Irony really sets in when a powerful technology is over-promoted, falls into disfavor, and then is underutilized due to a bad reputation. Transgenic technology fits this paradigm perfectly. In our opinion, transgenic procedures are invaluable for obtaining certain kinds of information, some of which will force scientists to re-think completely old problems. On the other hand, these procedures are so expensive and complicated that there are likely to be only a very few production agriculture applications of transgenic animals themselves for some years. The information from transgenic studies, however, is likely to be quite useful. Constraints on aericultural scientists. A major constraint on animal scientists is availability of funding. Very few universities have the wherewithal to carry out transgenic studies with livestock. Although one could request 2 or 3 million dollars for a transgenic project with cattle from the USDA Competitive Research Grant Program, it is unlikely that any study section would give half of their available funds to one project. The result is that many transgenic projects with farm animals are funded by private venture capital. This is not considered inappropriate but does have limitations.

350

Theriogenology

Another problem with farm animal transgenic research is the paucity of genetic information available about specific genes and mutations. Such information is likely to become available fairly rapidly as molecular biology approaches exploit the homologies in gene sequences amongst mammalian species. Even so, however, the lack of highly inbred lines poses severe constraints relative to working with rodents. With completely inbred lines, one does not have the confusion caused by mating relatives to obtain homozygous transgenic lines. A final constraint on applications to food-producing animals is consumer acceptance of transgenic products. No doubt this will sort itself out reasonably in the long run, but it may result in numerous stumbling blocks along the way. Lost onnortunities. The metallothionein-growth hormone (MT-GH) fusion gene was the first gene introduced into livestock (28) in an effort to improve feed efficiency and growth rates and to reduce carcass fat. Since that report, 10 other laboratories have published their successes with transgenic livestock (6, 9, 29, 30, 31, 32, 33, 34, 35, 36). The MTGH experiment was encouraging in that rate of gain was increased by 15%, feed efficiency was increased by 18%) and carcass fat was reduced 80% in transgenic pigs. However, the pigs and sheep exhibited a variety of undesirable side-effects that made it clear that the initial strategy was flawed. Even though the negative aspects of the MTGH project were reported (28), three other laboratories subsequently performed similar experiments and generally reached the same conclusions. Given the enormous resources that must be invested in transgenic livestock projects, one wonders if that much repetition was prudent. Confirmation of results is, of course, part and parcel of the scientific process; however, given current constraints, it is wise to avoid undue repetition. It may be useful to establish an organization for coordination of future projects among groups actively engaged in transgenic livestock research. Although it may be difficult for groups with commercial interests to participate in such a forum, academic and govemmentfunded institutions may benefit by regularly sharing and coordinating their activities. Unique Opportunities For the foreseeable future, the primary value of the transgenic animal model to society will be as a scientific tool for elucidating the mysteries of gene regulation. Much of that work can be successfully performed with laboratory animals. There is clearly a place for the application of transgenic technology to animal production, although marketable commodities may well be years away. Currently-proposed transgenic livestock projects include enhancement of production traits, increasing disease resistance, creation of new animal products, and creation of human genetic disease models. Enhancement of nroduction traits. The MT-GH studies in pigs and sheep failed to produce production animals because the transgene was not regulated as anticipated. In those lines of animals that exhibited only moderate side-effects, however, the beneficial effects of elevated growth hormone were obvious. To overcome the deregulation problems, regulatory elements from phosphoenolpyruvate (PEPCK; 36) and prolactin (35) have been substituted for the metallothionein regulatory elements. Neither of those studies yielded more positive results, demonstrating the difficulty in designing effective

Theriogenology transgenes. Future efforts will probably continue to focus on finding a means to control growth hormone-based transgenes properly, since experiments with growth hormone releasing factor and insulin-like growth factor I failed to elicit improved growth characteristics. With transgenic animal technology, it may be possible to increase wool production in sheep. Wool production requires the amino acid cysteine, and low availability can limit the rate of wool growth. To increase the cysteine availability, two laboratories in Australia (29,37) are attempting to introduce a bacterial-derived biochemical pathway for de novo synthesis of cysteine. The pathway involves two enzymes, serine transacetylase and 0-acetylserine sulfhydrylase, which, in a coupled reaction, converts serine and acetylCoA into cysteine and acetate. Transgenic mice already have been produced with the genes encoding these enzymes (38). Increase resistance to disease. Several approaches have been explored to improve an animal’s resistance to disease. They include introduction of genes that encode the information necessary to produce specific antibodies in pigs and sheep (39) and introduction of genes that encode viral envelope proteins in chickens (40). A sheep experiment currently underway, which parallels the transgenic chicken experiment, is designed to ‘block’ receptors for the Visna virus. Preliminary results are encouraging, but it remains to be determined whether sheep expressing the visna envelope protein will be resistant to infection (C.E. Rexroad, Jr., R.J. Wall, J. Clements and 0. Narayan, unpublished data). If this strategy is effective in mammals, it might be possible to develop strains of farm animals that are genetically immune to a variety of viruses. Creation of new animal oroducts. Currently, the most active area of transgenic livestock research is focused on producing new animal products. Of the ten organizations that have previously published original data on transgenic livestock, five are still active, and three of those are engaged in bioreactor projects. A variety of blood-born coagulation and anticoagulation factors (tPA, blood clotting Factors VIII and IX, and Protein C) have been successfully produced in the mammary glands of goats, pigs and sheep; in addition, human hemoglobin has been produced in the blood of pigs. The appeal of producing pharmaceuticals in the milk stems from the high cost of alternate production processes and the reduced availability of human blood from which many of these products are currently extracted. Also, at least theoretically, confining expression of biologically active compounds to the lactating mammary gland reduces the chance of causing undesired physiological side-effects. It also may be useful to alter the genetic control of mammary glands as a means of reformulating milk composition. One company is attempting to produce human lactoferrin in bovine milk, which will be isolated and then used as an additive to infant formula. USDA is initiating a project to increase porcine lactoferrin in sow’s milk to reduce the incidence of neonatal diarrhea. Because of lactoferrin’s bacteriostatic properties, it may be useful to increase bovine lactoferrin as a means of preventing some forms of mastitis.

352

Theriogenology

Milk composition could also be modified to improve post-harvest processing and nutritional value (41,42). Increasing casein content in milk would increase yields in cheese production. It has been estimated that increasing c&l-casein by 20% would increase cheese yield by approximately 0.1%. That small gain in efficiency would translate into $190 million per year savings to the U.S. dairy industry, which produces about five billion pounds of cheese annually (43). Modifying the primary sequence of c&l-casein to be more susceptible to chymosin cleavage would reduce the time required for cheese ripening. Increasing x-casein could increase the thermal stability of milk. Decreasing or eliminating g-lactoglobulin, a protein of no known function, could free up precursors for synthesis of other proteins of known value. Decreasing or blocking the function of acetyl CoA carboxylase might provide a means of reducing de novo fat And, finally, reducing or synthesis in the mammary gland secretory epithelium. eliminating lactose from cow’s milk would have a number of advantages, including obviating the need to add lactase to milk used in manufacturing ice cream and providing a milk product for individuals with lactose intolerance. A number of recent experiments suggest that it may Somatic cell eenetic eneineering. be possible to shortcut the process of producing transgenic animals by introducing genes directly into adults. Human gene therapy is based on such an approach. As currently practiced, cells harvested from the patient are transfected with the desired gene and then reintroduced into the patient, but a much simpler method involves ‘shooting’ DNA directly into the tissue to be engineered. Approaches have varied from simple injection with a syringe and needle to bombarding tissues with DNA-coated microprojectiles or propelling DNA solutions into tissues with air pressure (44,45,46,47). Tissues treated in vivo in this way include skin, liver, skeletal and cardiac muscles, and mammary glands (48). In all cases, the injected genes were expressed, in some cases for several months. Does this work foretell the day when bioreactors are created by injecting genes into the mammary glands of cows? Possibly. More recent work adds a twist that may cloud the prospects of producing somatic cell transgenic animals, while suggesting equally exciting uses for direct DNA injection. Tang, DeVit and Johnston (49) bombarded mice with human growth hormone gene-coated microprojectiles and demonstrated that it is possible to elicit an immune response to the gene product. Mice produced anti-hGH titers exceeding 0.5 ng/ul of serum lasting over 100 days and responded to booster bombardments with a typical secondary immune response. Such an immune response to injected genes is problematic from the somatic cell transgenic point of view. However, this may well lead to an entirely new approach to vaccine development. Genetics of Farm Animals in the Next Century Goals for farm animal genetics. A serious limitation to applying scientific principles of animal breeding has been confusion about goals. Some goals are incompatible to varying degrees due to negative genetic correlations; for example, low birth weight and heavy weaning weight in beef cattle or heavy milk production with high fertility in most

Theriogenology

353

mammals, especially when nutrition is suboptimal. Some goals are incompatible due to differing objectives among segments of the industry; for example, some heat-tolerant breeds of cattle have less tender meat. Even when goals are more monolithic, problems have arisen; for example, high milk production is associated with increased mastitis. Over the past 50 years, a frequent course of action has been to maximize rather than optimize. This usually has made sense; for example, high milk production, rapid broiler weight gains, or large amounts of wool per sheep were highly correlated with improved efficiency and profitability. For many situations, this is no longer the case, and, in fact, some sweeping corrections are occurring (e.g., decreased birth weights in beef cattle to minim& dystocia). In fact, it is now becoming fashionable to promote optimized animals rather than extremes. What has any of this to do with transgenic technology? Fundamentally new tools will become available to optimize animals. For example, importing genes from other species becomes possible for traits like disease resistance, changing milk composition, or manipulating homeobox genes to obtain more or fewer ribs or hexters instead of quarters for bovine mammary glands. Genes for partial hibernation might be extremely useful for wintering cattle and sheep by decreasing metabolic rates. Altogether different annroaches. Transgenic mice have already been made by injecting fragments of chromosomes into zygotes (50). As suggested some years ago (51), it should be possible to add artificial chromosomes to animals as has been done with yeast (52). Such chromosomes could serve as cassettes on which to place a number of transgenes so that physical interference with transcription of genes in the original genome would not occur. Such transgenic animals should be extremely useful to explore chromosomal arrangements of genes, nuclear matrix attachment regions, etc. Of course, most such research should be done with mice, but applications would shine in farm animals. Another idea is to use the Y chromosome as an artificial chromosome. The Y chromosome could even be used directly as a vector by adding it at the first metaphasc division of a zygote. XYY males are not infertile, and some of the sons would be XYorigina’and some would be XYne”. One might even wish to clone an XYY embryo by nuclear transplantation. With serial cloning, very large numbers of embryos could be made. Such individuals might well grow rapidly and have leaner carcasses without any need for other genetic changes, and without hormone-like implants. One might use related approaches to produce a third sex; in many respects, there already are four sexes (five including freemartins) marketed: females, neutered males, males, and neutered females (in that order for numbers) for cattle, sheep, and swine. With genetic engineering to prevent testicular development plus serial nuclear transplantation and embryo transfer, one could have a line of steers produced asexually. In addition to not requiring castration, they would automatically be sexed and be selected

354

Theriogenology

for growth, carcass qualities, and for the polled trait; disease resistance genes might also be added. Such animals could well be $lOO-$200 more valuable than animals from the best conventional breeding systems, and that extra value might justify the technology, including embryo transfer. It would also be a more humane system than current practices. Clearly, transgenic technology will have made very important contributions to livestock production by the middle of the next century. In the near term, the transgenic animal model will provide new insights into old biological questions. Much of that newly acquired knowledge will be used in non-transgenic applications to solve agricultural problems. There is little doubt that transgenic farm animals will eventually play a direct role in animal agriculture. REFERENCES 1.

Swain, J.L., Stewart, T.A. and Leder, P. Parental legacy determines methylation and expression of an autosomal transgene: A molecular mechanism for parental imprinting. Cell 3:719-727 (1987).

2.

Readhead, C., Popko, B., Takahashi, N., Shine, H.D., Saavedra, R.A., Sidman, R.L. and Hood, L. Expression of a myelin basic protein gene in transgenic Shiverer mice: Correction of the dysmyelinating phenotype. Cell 48:713-712 (1987).

3.

Palmiter, R.D., Norstedt, G., Gelinas, R.E., Hammer, R.E. and Brinster R.L. Human GH fusion genes stimulate growth of mice. Science m:809-814 (1983).

4.

Brinster, R.L., Chen, H.Y., Trumbauer, M.E., Senear, A. W., Warren, R., Palmiter, R.D. Somatic expression of herpes thymidine kinase in mice following injection of a fusion gene into eggs. Cell z:223-231 (1981).

5.

Saffer, J.D.

6.

Hill, K.G., Curry, J., DeMayo, F.J., Jones-Diller, K., Slapak, J.R. and Bondioli, K.R. Production of transgenic cattle by pronuclear injection. Theriogenology =:222 Abstr. (1992).

I.

Pursel, V.G., Pinkert, C.A., Miller, K.F., Bolt, D.J., R.L., Hammer, R.E. Genetic engineering in livestock.

8.

Rexroad, C.E. Jr., Hammer, R.E., Behringer, R.R., Palmiter, R.D. and Brinster, R.L. Insertion, expression and physiology of growth-regulating genes in ruminants. J. Reprod. Fert. Suppl. &l:119-124 (1990).

9.

Ebert, K.M., Selgrath, J.P., DiTullio, P., Denman, J., Smith, T.E., Memon, M.A., Schindler, J.E., Monastersky, G.M., Vitale, J.A. and Gordon, K. Transgenic production of a variant of human tissue-type plasminogen activator in goat milk: Generation of transgenic goats and analysis of expression. BioTechnology $835-838 (1991).

10. Pahniter, (1986).

Transgenic

mice in biomedical

R.D. and Brinster, R.L.

research.

Lab Anim. a:30-38

Metallothionein-

(1992).

Campbell, R.G., Palmiter, R.D., Science 244: 1281-1287 (1989).

Germ-line transformation

Brinster,

of mice. Ann. Rev. Genet. B:465-499

Theriogenology

11. Pedersen, R.A. Potency, lineage, and allocation in preimplantation mouse embryos. and Pedersen, R.A. (eds) Experimental Approaches to Mammalian Embryonic Cambridge University Press, Cambridge, 1985, pp. 3-33. 12. Fleming, T.P. A quantitative analysis of cell allocation to trophectoderm mouse blastocyst. Develop. Biol. m:520-531 (1987). 13. Petters, R.M. and Markert, C.L. Production o&parental mice. J. Hered. a:70-74 (1980).

and reproductive

14. Seidel, G.E., Jr. Mammalian oocytes and preimplantation Biol. Reprod. B:36-49 (1983). 15. Bishop, J.O. and Smith, P. Mechanism Biol. Med. 6:283-298 (1989).

of chromosomal

and inner cell mass in the

performance

of hexapatental

embryos as methodological

integration

In: Rossant, J. Development.

of microinjected

and

components.

DNA.

Mol.

16. Brinster, R.L., Chen, H.Y., Trumbauer, M.E., Yagle, M.K. and Palmiter, R.D. Factors affecting the efficiency of introducing foreign DNA into mice by microinjection of eggs. Pmt. Natl. Acad. Sci. USA &:4438-4442 (1985). 17. Steller, H. and Pirrotta, V. A transposable P vector that confers Drosophila larvae. EMBO J. 4: 167-171 (1985). 18. Stokes, H.W. and Hall, R.M. A novel family of potentially specific gene-integration function: integrons. Mol. Microbial. 19. Gdell, J., Caimi, P., Saver, B. and Russell, S. transgenic tobacco. Mol. Gen. Genet. =:369-378

selectable

G418 resistance

mobile DNA elements encoding 2: 1669-1683 (1989).

Site-directed (1990).

recombination

in the genome

to

site

of

20.

Palmiter, R.D., Sandgren, E.P., Avarbock, M.R., Allen, D.D. and Brinster, R.L. Heterologous introns can enhance expression of transgenes in mice. Proc. Natl. Acad. Sci. USA a:478482 (1991).

21.

Choi, T., Huang, M., Gormon, C. and Jaenisch, R. A generic intron increases gene expression transgenic mice. Mol. Cell. Biol. u:3070-3074 (1991).

in

22. Whitelaw, C.B., Archibald, A.L., Harris, S., McClenaghan, M., Simons, J.P. and Clark, A.J. Targeting expression to the mammary gland: intronic sequences can enhance the efficiency of gene expression in transgenic mice. Transgenic Res. 1:3-13 (1991). 23.

Palmiter, R.D., Wilkie, T.M., Chen, H.Y. andBrinster, R.L. Transmissiondistortionand in an unusual transgenic mouse pedigree. Cell %:869-877 (1984).

mosaicism

24.

Smith, C., Meuwissen, T.H.E. and Gibson, J.P. Anim. Br. Abstr. s:l-10 (1987).

25.

Bondioli, K.R., Ellis, S.B., Pryor, J.H., Williams. M.W. and Harpoid, M.M. The use of malespecific chmmosomal DNA fragments to determine the sex of bovine preimplantation embryos. Theriogenology x:95-104 (1989).

26.

King, D. and Wail, R.J. Identification of specific gene sequences in preimplantation embryos genomic amplification: Detection of a transgene. Mol. Reprod. Develop. I:5762 (1988).

On the use of transgenes in livestock improvement.

by

356

Theriogenology

27.

Ninomiya, T., Hoshi, M., Mizuno, A., Nagao, M. and Yuki, A. Selection of mouse preimplantation embryos carrying exogenous DNA by polymerase chain reaction. Mol. Reprod. Develop. 1:242-248 (1989).

28.

Hammer, R.E., Pursel, V.G., Rexroad, C.E., Wall, R.J., Bolt, D.J., Ebert, K.M., Pahniter, R.D. and Brinster, R.L. Production of transgenic rabbits, sheep and pigs by microinjection. Nature 315:680-683 (1985).

29. Ward, K.A.,Nancarrow, C.D., Byene, C. R., Sham&an, C.M., Murray, J.D., Leish, Z., To-w, C., Rigby, N.M., Wilson, E.W. and Hunt, C.L. The potential of transgenic animals for improved agricultural productivity. Rev. Sci. Tee. Off. Int. Epiz. 9:847-864 (1990). 30. Simons, J.P., Wilmut, I., Clark, A.J., Archibald, in sheep. BioTechnology 6:179-183 (1988). 31. Roschlau, K., Rommel, Huhn, R. and Gazarjan,

A.L.,

Bishop, J.O. and Lathe, R. Gene transfer

P., Andreewa, L., Zackel, M., Roschlau, D., Zackel, B., Schwerin, M., K.G. Gene transfer in cattle. J. Reprod. Fert. Suppl. 2:153-160 (1989).

32. Brem, G., Brenig, B., Goodman, H.M., Selden, R.C., Graf, F., Kruff, B., Springman, K., Hondele, J., Meyer, J., Winnaker, E.-L. and Kausslick, H. Production of transgenic mice, rabbits and pigs Zuchthygiene 3:251-252 (1985). by microinjection into pronuclei. 33.

Krimpenfort, P., Rademakers, A., Eyestone, W., Van der Schans, A., Van den Broek, S., Koolman, P., Kootwijk, E., Platenburg, G., Pieper, F., Strijker, R. and De Boer, H. Generation of transgenic Bio/Technology 9:844-847 (1991). dairy cattle using in vitro embryo production.

34. Vise, P.D., Michalska, A.E., Ashuman, R., Lloyd, B., Stone, A.B., Quinn, P., Wells, J.R.E. and Seamark, R.F. Introduction of a porcine growth hormone fusion gene into transgenic pigs to promote growth. J. Cell. Sci. @:295-300 (1988). 35. Polge, E.J.C., Barton, S.C., Surani, M.H.A., Miller, J.R., Wagner, T., Elsome, K., Davis, A.J., Goode, J.A., Foxroft, G.R. and Heap, R.B. Induced expression of a bovine growth hormone construct in transgenic pigs. Is: Heap, R.B., Prosser, C.G., Lamming, G.E. (eds) Biotechnology of Growth Regulation. Butterworths, London, UK, 1989, pp. 189-199. 36. Wieghart, M., Hoover, J.L., McGrane, M.M., Hanson, R.W., Roltman, F.M., Holtzman, S.H., Wagner, T.E. and Pinkert, C.A. Production of transgenic pigs harbouring a rat phosphoenolpyruvate carboxykinase-bovine growth hormone fusion gene. J. Reprod. Fert. Suppl. a:89-96 (1990). 37.

Rogers, G.E. 11 (1990).

Improvement

of wool production

38. Ward, K.A. and Nancarrow, C.D. Experientia a:913-922 (1991).

through genetic engineering.

The genetic engineering

of production

Trends Biotechnol.

8:6-

traits in domestic animals.

39.

Lo, D., Pursel, V., Linton, P.J., Sandgren, E., Behringer, R., Rexroad, C.E. Jr., Pahuiter, R.D. and Brinster, R.L. Expression of mouse IgA by transgenic mice, pigs and sheep. Eur. J. Immunol. 21:1001-1006 (1991).

40.

Salter, D.W., Smith, E.J., Hughes, S.H., Wright, S.E. and Crittenden, L.B. insertion of retroviral genes into the chicken germ line. Virology m:236-240

Transgenic (1987).

chickens:

Theriogenology

357

41.

Kang, Y., Jimenez-Flores, R. and Richardson, T. Casein genes and genetic engineering of the caseins. In: Evans, J.W. and Hollaender, A. (eds) Genetic Engineering of Animals: An Agricultural Perspective. Plenum Press, New York, 1985, pp. 9.5-112.

42.

Bremel. R.D., Yom, H.-C. and Bleck, G.T. Alteration of milk composition J. Dairy Sci. n:2826-2833 (1989).

43.

Hetighausen, L., Ruiz, L. and Wall, R.J. Transgenic milk. Cur. Opin. in Biotech. 1:74-78 (1990).

44.

Williams, R.S., Johnston, S.A., Riedy, M., DeVit, M.J., McElligott, S.G. and Sanford, J.C. Introduction of foreign genes into tissues of living mice by DNA-coated microprojectiles. Proc. Natl. Acad. Sci. USA a:2726-2730 (1991).

45.

Kitsis, R.N., Buttrick, P.M., McNally, E.M., Kaplan, M.L. and Leinwand, L.A. Hormonal modulation of a gene injected into rat heart in vivo. Proc. Natl. Acad. Sci. USA @:4138-4142 (1991).

46.

Yang, N.-S., Burkholder, J., Roberts, B., Martinell, B. and McCabe, D. In vivo and in vitro gene transfer to mammalian somatic cells by particle bombardment. Proc. Natl. Acad. Sci. USA fl:95689572 (1990).

47.

Brandsma, J.L., Yan, Z.-H., Barthold, S. W. and Johnson, E. A. Use of a rapid, efficient inoculation method to induce papillomas by cottontail rabbit papillomavirus DNA shows that the E7 gene is required. Proc. Natl. Acad. Sci. USA a:48164820 (1991).

48.

Furth, P., Shamay, A., Wall, R.J. and Hennighausen, injection. Anal. B&hem. (in press, 1992).

49.

Tang, D-c, DeVit, M. and Johnston, S.A. Genetic immunization immune response. Nature =:152-154 (1992).

50.

Richa, J. and Lo, C.W. Introduction of human DNA into mouse eggs by injection chromosome fragments. Science =:175-177 (1989).

51.

Seidel, G.E., Jr. Characteristics of future agricultural animals. In: Evans, J.W. and Hollaender, A. (eds) Genetic Engineering of Animals. Plenum Publishing Corporation, New York, 1986, pp. 299309.

52.

Murray, A.W. and Szostak, J.W. 192 (1983).

Construction

using molecular genetics.

animals - production

L.

of foreign proteins in

Gene transfer into somatic tissue by jet

is a simple method for eliciting an

of artificial chromosomes

in yeast.

of dissected

Nature 305: 189-