1. Introduction

 

 

The earth's organisms are a vast repository of genetic diversity. Each species (n >10⁶) is distinguished from every other by a unique genomic sequence that is passed on to successive generations with extremely high, but not perfect, fidelity. Imperfections in DNA replication and repair mean that the genome of each member of a species is also unique. Intraspecific differences are one basis for individuality, including individual differences in susceptibility to disease. The most striking example of such differences are genetic diseases.

1.1 The Need For Animal Models And A New Approach To Obtaining Them.

Animal models of genetic diseases have been extremely useful. Models can arise from chance discoveries, such as narcopleptic dogs , or by intentional screening of inbred animals . Most importantly, actual creation of mouse models of diseases has been made possible by stem cell lines and methods for introducing specific mutations into those cells , which has led to an explosion of information . Unfortunately, mouse models are not ideal for some human diseases. For example, in the mouse model of cystic fibrosis, the mice fail to develop the lung and pancreatic pathology that are hallmarks of the human disease, but have a more severe form of intestinal disease . Furthermore, even though mice with improved disease features are being developed through selective breeding , mice are still not ideal for many purposes, especially those related to the evaluation of clinical interventions. Thus, alternative animal models would be a boon for researchers. However, in animals other than the mouse it has so far been extremely difficult to develop embryonic stem cell lines that routinely give rise to viable offspring .

In this paper we describe an alternative strategy for the discovery of natural animal models of recessive genetic diseases. The strategy is based on the hypothesis that disease frequencies across human populations offer some guide to disease frequencies in animals. When disease frequencies are high enough (>10^-6), the method is feasible using existing methods for genetic screening of genomic DNA.

The key to the feasibility of the method is the ability to screen for unaffected heterozygotes. It is not sufficiently appreciated that even rare, recessive genetic diseases have relatively high heterozygous gene frequencies. For example, a recessive disease frequency of 1/1,000,000 arises from a carrier frequency of only 1/500. This enormous disparity explains why carriers can be detected readily in populations for which the associated recessive disease is apparently "non-existent".

In this chapter, we introduce the general concept, outline two attempts to implement the approach , and provide a series of steps that should be followed to allow this approach to become a general and cost-effective alternative to stem cell technology.

1.2. Do Recessive Human Genetic Diseases Have Animal Counterparts?

Some recessive diseases have been documented in both humans and animals , but how likely is it that a specific human genetic disease will occur in a specific animal species? The human genome is estimated to contain between 30,000 and 40,000 genes . However, the Online Mendelian Inheritance in Man lists fewer than 10,000 autosomal entries, and fewer than half of these are recessive. The disparity between gene number and disease number has many explanations, but the contribution of each is unknown. If we consider only recessive mutations, we know from experimental work that some of these cause early embryonic lethality when homozygous while others cause no obvious phenotype. Another consideration is that recessive diseases are usually so rare that the chance of a disease escaping diagnosis is high. That leaves an unknown proportion of genes for which it might be argued that the lack of a known disease state arises simply because the mutation frequency in the associated gene is so low that no human exists who has two copies of the mutated gene. How likely is this?

The human population is estimated to be approaching 6 x 10⁹ individuals. To estimate the number of mutations within this vast gene pool we need to know the mutation rate for human genomic DNA. Unfortunately, estimates of that rate vary widely. Based on extensive experiments with Drosophila, Crow gives an estimated mutation rate per nucleotide per generation, of 1.5 x 10^-8, and predicts that the 3 x 10⁹ nucleotide pairs of the human genome, will therefore acquire ~100 new mutations in each human zygote, with ~2% of these affecting genes. To avoid the accumulation of an enormous mutational load, it is proposed that heterozygotes are mildly but cumulatively disadvantaged, and that their preferential elimination culls numerous mutations simultaneously to counterbalance the accumulation . In contrast, experiments in which 50 independent lines of C elegans were allowed to accumulate spontaneous mutations led to the conclusion that the deleterious mutation rate per haploid genome was 0.0026 .

1.3. Carriers Greatly Exceed Affected Individuals.

As stated above, the enormous disparity between heterozygote and homozygote frequencies (Fig. 1) is not widely appreciated. Cystic fibrosis (CF) illustrates some of the consequences of this disparity. CF has an extraordinarily high frequency in the U.S. and northern European Caucasian populations, where about 1/25 individuals are heterozygous for mutations in the causative gene, CFTR. Cystic fibrosis is comparatively rare in other populations. One estimate of the incidence rate of cystic fibrosis in Japan gave a rate of 3.1 per million live births from 1969-1980 (~1/323,000). The highest rates of CF were in Hokkaido (the most northern Island) and lowest in Okinawa (the most southern island). The mean age at death from CF was 3 years for both sexes during the period 1969-1985 . A similar estimate of 1/350,000 using different methods was made more recently . Inspection of Fig. 1 or simple calculation shows that the lowest estimate still corresponds to a carrier frequency of ~1/295, suggesting that the Japanese population (n » 10⁸), has ~ 339,000 individuals carrying disease-causing CFTR mutations. A similar exercise for China suggests it has >4 million cystic fibrosis carriers.

The high incidence of cystic fibrosis in Caucasian populations results primarily but not exclusively from the frequency of one very common allele (D F508). In the CF population of the U.S. and Canada, the D F508 mutation accounts for ~70% of all alleles. Hence, if D F508 were to be subtracted out, the frequency of cystic fibrosis in this population would drop from about 1/2,500 (1/25 carrier frequency) to ~1/28,000 (~1/83 carrier frequency). That is still much higher than the estimates for CF in Japan, and suggests that factors other than the D F508 mutation are at work.

1.4. The Hypothesis: The Aggregate ‘Background’ Frequency Of Human and Animal Mutations are Similar.

We hypothesize that the aggregate frequency of non-D F508 CF causing mutations in human populations offers a rough guide to the aggregate CFTR mutation frequencies in non-human primates. This hypothesis does not assume that any specific mutations found in human populations will necessarily be found in non-human primates. Correspondingly, of course, the frequency of particular mutations in the human population will not provide information about particular mutation frequencies in non-human primate populations; that is evident from the different pattern of mutations observed in separated human populations. Thus, it does not make sense to search non-human populations for specific mutations. Instead, a method is required that is capable of detecting unknown mutations.

We know of no a priori reasoning and certainly no data that would suggest a much different aggregate CF mutation frequency in non-human primates. CFTR is a large gene and like any gene is susceptible to insertions or deletions that cause frame shifts, as well as stop mutations and splicing mutations. CFTR is also, for unknown reasons, extremely susceptible to missense mutations that cause it to be misprocessed . Even wild type CFTR is inefficiently processed. Approximately 75% of wild type CFTR protein is degraded after core glycosylation and never reaches the plasma membrane--this occurs across a range of cells expressing various levels of CFTR and so is not merely an artifact of high levels of exogenous expression . At least four critical regions in CFTR (the pore and the two NBFs) are susceptible to missense mutations that interfere with CFTR's ability to function as a Cl- channel¾ these also cause cystic fibrosis . Finally, recent evidence indicates that some missense mutations which have little affect on processing or chloride channel function can cause CF by altering HCO₃^- transport.

In sum, the large size of CFTR creates many opportunities for mutations, and a high proportion of all mutations render CFTR non-functional and lead to disease in the homozygous state. In humans and mice, heterozygosity for CF has no detectable disadvantage. Thus, leaving aside the possibility of a heterozygote advantage, and barring some unforeseen feature that powerfully selected against monkey carriers, there is no efficient mechanism to prevent CFTR mutations from accumulating in a population at low frequencies, since such frequencies will give rise to homozygotes too infrequently to alter mutation frequencies in the population.

1.5 Which Species Should Be Studied?

The choice of species to be studied is greatly narrowed by several obvious features. Because mice can be genetically manipulated, there is little reason to study any species further removed from humans than mice. Among remaining species, four main criteria determine suitability for discovery of natural animal models. These criteria are (1) availability, (2) experimental tractability, (3) similarity to humans, and (4) genetic diversity. The first two criteria need not be elaborated. However, the criterion of human similarity can vary depending upon the disease of interest, such that a more closely related species may be less optimal than species that are further removed phylogentically. For example, sheep and pigs have lungs that may be better experimental models of some human lung diseases than monkeys. Genetic diversity within the target population is a crucial feature. Unfortunately, the need for high genetic diversity excludes most domestic populations of animals, but even wild populations may be unsuitable. For example, cheetahs display an extreme degree of genetic homogeneity, presumably as a result of a severe population bottleneck that occurred ~10,000 years ago. .

Old world monkeys, particularly the genus Macaca, rank highly on all 4 criteria. (1) Availability is good. Wild populations are still large (though declining at an alarming rate) and are extensively distributed throughout Africa and Asia. An estimated 40,000 primates are imported annually into the U.S for research purposes. Of greater relevance are the large number of primates (~16,000) maintained at NIH Regional Primate Research Centers. This population is bred exclusively for research, and the monkeys receive excellent care and typically live for longer than a decade. The last point is critical, because it is essential to be able to retrieve an animal after a mutation has been identified in its DNA. (2) Monkeys are good experimental subjects, and for some experiments are virtually the only suitable animal subjects. (3) Monkeys are in general more similar to humans than any other species except the great apes, and apes have become so endangered that their use is virtually precluded except for the most essential studies. (4) Finally, with regard to genetic diversity, the evidence suggests that monkeys may be an unusually rich repository of genetic diversity. Studies of 6 species of macaques and of 23 local populations of Rhesus monkeys spread across Vietnam, Burma, and 10 provinces of China extended previous estimates of genetic heterogeneity among and within species. Our own studies have confirmed a high degree of genetic diversity even within Macaca maintained for many years within Regional Primate Research Centers.

A possible heterozygote advantage. With regard to a possible heterozygote advantage, it may be relevant that monkeys are notoriously susceptible to secretory diarrhea. Human CF heterozygotes are thought to be partially protected against certain diarrheal diseases because CFTR is rate-limiting for Cl--mediated electrolyte and fluid secretion from intestinal crypts . In some secretory pathways, fluid secretion by CF heterozygotes is indeed reduced to 50% of normal . Hence, diarrheal diseases that stimulate the CFTR-dependent pathway should cause less fluid and electrolyte loss in heterozygotes. This hypothesis has been tested directly by administering cholera toxin to heterozygous CF mice, but results were conflicting .

A study of serum electrolyte values in 100 rhesus monkeys with diarrhea observed hyponatremia in 88% and hypochloremia in 80% . This strongly suggests that the putative protective effect of CFTR mutations should also apply to non-human primates, and could result in positive selection pressure and hence some enrichment of CF alleles in non-human primate populations. To give some indication of the magnitude of this potential selection pressure for CFTR mutations, in the California Primate Research Center, 34% of non experimental deaths in macaques one year of age and older were due to gastrointestinal disease .

Possible heterozygote disadvantages. The severity of disease caused by CFTR mutations is closely related to the extent to which CFTR-mediated Cl^- conductance is lost. Mild mutations can arise for each class of CFTR mutation; for example, some trafficking mutations allow a certain proportion of CFTR to be processed , some regulatory mutations do not completely disrupt function , and all conductance mutations to date produced only a partial loss of conductance.

Within a critical range of residual CFTR function, subjects no longer display the classic cystic fibrosis syndrome, but instead suffer, if they are male, from sterility secondary to congenital bilateral absence of the vas deferens (CBAVD) . The extreme susceptibility of the vas deferens to mutations in CFTR is not completely understood. However, part of answer may be that CFTR is spliced differently in the vas. A common mutation that contributes to CBAVD is a reduction in a tract of 8 thymidines within intron 8 to a 5-thymidine variant that leads to missplicing of CFTR. The proportion of misspliced CFTR is greater in the vas deferens than in the lung . Unlike humans, mice homozygous for CFTR mutations remain fertile . Given this species difference, a possibility that must be considered is that in some species male fertility will be lost or compromised even in the heterozygous state.

````````

1.6. Testing the Approach: the Search for a Monkey CF Carrier.

In spite of such arguments, only direct experimental test can provide accurate estimates of the frequency of CFTR mutations and polymorphisms in a given species. With no additional information, the chances of mutations being more frequent in non-human primates than in human populations is equal to the chance that they are less frequent. For a mutation frequency of 1/500 (for a disease frequency of 1/1,000,000); screening 1,500 animals, yields ~95% chance of detecting a mutation if the detection method is perfect. Of course no detection method is perfect. The SSCP method is near perfect for detecting small insertions and deletions, and probably detects >90% of point mutations. However, it does not detect intronic mutations or deletions of entire exons. Based on assays of CF populations, the SSCP/HD method we use is estimated to be able to detect ~95% of CF mutations . Thus, screening of 1,500 primates provides a 95% chance of finding a mutation if mutations occur at a frequency of 1/400 or greater, equal to a disease frequency of ~1/640,000. It is worth emphasizing that even a disease frequency of 1/100,000 would make it unlikely that even a single CF birth would have occurred among the entire primate population in all of the U.S. Primate Research Centers during the last decade. Given the infant mortality rate mentioned above, even if such a rare event occurred, the chance that it would have been detected is remote. This emphasizes the power of heterozygote analysis even among populations in which the disease appears to be "non-existent".

The general significance of this program will be to determine the feasibility of establishing animal models for any disease by screening. If our hypothesis of an approximate correspondence in mutation frequencies among primates is confirmed, a program like the one we propose should be at least as cost-effective as the production of mice by stem-cell recombinant methodology.

2. Materials

 

 

2.1. Whole blood (~ 3 ml) was obtained by venipuncture mainly during routine medical checkups of primates and was shipped on ice in EDTA-containing (purple top) tubes.

2.2. EDTA - Ethylenediaminetetraacetic acid ( Sigma, St. Louis, USA) 1 ml of 10% EDTA solution was used for each blood sample.

2.3. Puregene DNA isolation kit (Gentra systems, Inc., Research Triangle Park, North Carolina, USA).

2.4. CFTR Primers, Operon technologies, Inc (For primer sequences, see Wine, 1998 ).

2.5. AmpliTaq® DNA Polymerase and GeneAmp® 10X PCR Buffer was used for PCR amplification (Applied Biosystems, Foster city, CA, USA).

The GeneAmp 10X PCR Buffer is composed of 500 mM potassium chloride, 100 mM Tris-HCl (pH 8.3 at room temperature), 15 mM magnesium chloride and 0.01% (w/v) gelatin GeneAmp® dNTPs GeneAmp® dNTPs deoxynucleoside triphosphates were used Composition: Each of the 4-vial set contains 320 µL of a 10 mM Solution of either dATP, dCTP, dGTP or dTTP.

Isotope a 32P dCTP (3000 Ci/mM) (Amersham Pharmacia Biotech, Piscataway, NJ, USA) 0.5 mCi of a 32P dCTP (3000 Ci/mM) isotope was included in each 10 m L PCR mixture for labeling.

2.6. Denaturing mixture : 95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol, and 20 mM NaOH

Formamide > 99.5 GC ( Sigma, St. Louis, USA)

EDTA - Ethylenediaminetetraacetic acid ( Sigma, St. Louis, USA)

Bromophenol blue ( 3’,3’’,5’,5’’-tetrabromophenolsulfonephthalein sodium salt)

(Sigma, St. Louis, USA)

Xylene Cyanole FF (dye content: ~75%) (Sigma, St. Louis, USA)

NaOH Sodium hydroxide ( Sigma, St. Louis, USA)

2.7. Standard components for Vertical polyacrylamide Gel Electrophoresis (e.g. MDE gel solution or acrylamide/bis-acrylamide mix, glycerol, TEMED, vertical slab gel KODAK BIOMAX STS 45i apparatus for running 35 x 40 cm, 0.4 mm-thick gels).

Plates (35x40cm ) for vertical slab gel KODAK BIOMAX STS 45I apparatus (Eastman Kodak Company, Rochester, New York, USA).

TEMED (N,N,N’,N’-tetramethylethylenediamine ) >99% ( Sigma, St. Louis, USA).

MDE gel (BioWhittaker Molecular Applications, Rockland, ME , USA)

Glycerol >99% Electrophoresis reagent ( Sigma, St. Louis, USA) 10% added to MDE gel solution.

Polyacrylamide: Bio-Rad Laboratories 40% acrylamide : N,N’-Methylenebisacrylamide solution, 37.5:1 (2.6%C).

TBE buffer (TRIS-BORATE-EDTA Buffer) 5X concentration (Sigma, St. Louis, USA).

1 liter of 1 X TBE buffer was used for electrophoresis.

Electrophoresis power supply EPS 1001 (Amersham Pharmacia Biotech, Piscataway, NJ, USA) was used for electrophoresis.

Autoradiography was done on KODAK Scientific Imaging Film X-OMAT AR (35x43 cm) (Eastman Kodak Company, Rochester, New York, USA).

2.8 Mutations were made with Stratagene's Quick-change site- directed mutagenesis kit. (La Jolla, CA) and verified with restriction enzymes (Life Technologies, Grand Island, NY) or sequencing.

2.9. Expression:

Plasmid purification: Qiagen plasmid maxi kit (Valencia, CA)

Transfection: SuperFect transfection reagent. (Qiagen, Valencia, CA)

Dish coating: fibronectin (Sigma, F2006) (St. Louis, MO)

DME H21, 10% fetal bovine serum, 2 mM glutamine, and Pen/Strep (100 U/m g/ml). (Sigma, St. Louis, MO)

2.10 Functional assay. Efflux buffer was (in mM): 50 N-2-hydroxy ethylpiperazine-N'-2-ethane sulfonic acid (HEPES), 5.4 KCl, 130 NaCl, 1.8 CaCl2, 1.0 Sodium Phosphate (monobasic), 0.8 MgSO4, pH adjusted to 7.4 with NaOH, and glucose 100 mg/100 ml, all from Sigma, and 125I /ml (St. Louis, MO).
 
 

3. Methods.

1. Identify an Animal Population. Our sample consisted of 1500 primates Macaca. mulatta, M. nemestrina,M. fuscata, M. arctoides, M. fasicularis and P. anubis. We used different species as a way of increasing genetic diversity. The similarity among species was such that the primers we used amplified CFTR in all species. The primates are maintained at 5 NIH Regional Primate Research Centers. Pedigrees were used when available to select maximally outbred animals. A program is now underway to use genetic typing to establish pedigrees for all primates in the Research Centers. That will make future use of these animals more efficient.
2. Obtain blood. Blood samples (2 ml in EDTA-containing tubes) are obtained during routine testing of animals and shipped on ice via overnight express.
3. Purify DNA Genomic DNA is purified from the blood samples using a commercially available kit (Puregene, Gentra Inc.) and stored at -20° C.
4. Design Primers. It is necessary to know or to obtain the genomic sequence of the target gene before proceeding. We wanted to use intronic primers so that we did not miss mutations in the splice sites and adjacent regions of exons. We started with intronic primers based on the published genomic sequence for human CFTR . Because of homology between human and monkey introns is about 90%, this strategy amplified about half of the CFTR exons. New primers were designed for 12 exons that did not amplify with human primers . To obtain intronic primers for the primate exons that did not amplify with human primers, we reasoned that the high intronic conservation between monkeys and humans would allow us to obtain effective primers simply by shifting the primer site to different regions of the introns. That strategy was successful for many exons on the first try. For others, we selected yet a different region of each flanking intron and paired it with an exonic primer. By repeating this process as required we were able to amplify flanking intron sequences for all remaining exons. Optimal intronic primers were then selected.
5. Amplify and label Exons We studied 25 of the 27 CFTR exons, omitting the two shortest exons. Genomic DNA was amplified and radioactively labeled with conventional polymerase chain reaction (PCR). PCR was carried out with a 10 ml reaction mixture containing 40 ng DNA, 0.2 unit Taq polymerase, 2.5 pmol of each primer, 50 mM KCl, 2.5 mM MgCl2, 10 mM Tris (pH 8.3), 200 m M each of dATP, dGTP, dTTP dCTP; 0.5 m Ci of a32PdCTP (3000 Ci/mM) and 0.001% gelatin. Amplification parameters were: denature 6 min at 94° C, then 30 cycles of: denature for 30 sec at 94° C, anneal 30 sec at 55° C, extend 1 min at 72° C; followed by 7 min at 72° C. We amplified DNA from 96 different monkeys in parallel.
6. Denature-Renature. After amplification each reaction mixture in the 96 well plate was diluted 1:1 with a mixture of 95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol, and 20 mM NaOH. The plate was then returned to the PCR machine and heated to 95° C for 4 min, followed by snap cooled in an ice/water bath.
7. Run on non-denaturing gel. Exons were screened using single strand conformation polymorphism and heteroduplex (SSCP/HD) analysis . The principle of this method is illustrated in Fig. 3. In brief, after denaturing and renaturing, the single strands refold on to themselves as well as annealing with opposite strands. In the purest case, each single strand will refold into one conformation that will migrate differently from the opposite strand and from the duplex, giving 3 distinct bands. However, variations of this simplest pattern can be observed if a portion of one or the other strand assumes a second stable configuration. Regardless, if a mutation exists on one chromosome, it almost always alters the migration pattern of one or both strands, as well as forming heteroduplexes with the unmutated opposite strands. The method depends simply on being able to detect this different migration.

 

 

We used gel conditions that had previously been optimized for CFTR and that should be capable of detecting >95% of CFTR mutations . For each exon from each animal, 2-4 m l of reaction mixture was loaded in one well of a100 lane Lanes were generated with a shark’s tooth comb), polyacrylamide gel consisting of 0.5X MDE (FMC Bioproducts, Rockland, ME) plus 10% glycerol . Gels were run in a 4° C cold room for 4-8 hr at 50W. Gels were then adsorbed onto filter paper and the paper with adherent gel was peeled from the glass plates, dried, and autoradiographed for 12- 48 hrs.

8. Identify shifts. Each gel contains 96 juxtaposed lanes. The method relies on the majority of exons having identical sequences and hence identical SSCP patterns, against which any shifts are obvious by comparison with the flanking sequences.
9. Sequence exons of interest. Amplified and purified DNA from samples displaying different SSCP patterns is then sequenced. Because sequencing is now so commonly used and is often carried out by a separate, dedicated facility, this step is not described further.
10. Identify candidate disease-causing mutations. The choice of which missense mutations to pursue via functional analysis is subjective. Obvious criteria include the frequency of the mutation, whether it changes a conserved amino acid, and whether the change is conservative or non-conservative. Less obvious criteria are based on comparisons with the large number of human missense mutations; certain regions of CFTR (e.g. NBD1) appear to be more sensitive to the missense mutations.
11. Introduce mutations into cDNA. For functional assays, candidate missense mutations were introduced into human CFTR cDNA using Stratagene's Quick-change site- directed mutagenesis kit. The presence of the mutation was verified with restriction site analysis as appropriate and direct sequencing.
12. Express mutated cDNA Plasmids were purified with the Qiagen plasmid maxi kit and concentrated to ~1 m g/m l). Human embryonic kidney (HEK) 293 cells were then transfected with the plasmids using Qiagen's SuperFect transfection reagent. HEK296 cells (1.4 x 10<sup>5</sup> in 2 ml of medium) are seeded onto 35 mm dishes coated with fibronectin (Sigma, F2006) and grown in DME H21 with 10% fetal bovine serum, 2 mM glutamine, and Pen/Strep (100 U/m g/ml).
13. Assay for Function. For mutations that obviously disrupt the gene, this difficult and expensive step is unnecessary, but for missense mutations a functional assay is essential. Assays require knowledge of what the gene product does. CFTR does many things, but its best-documented role is as a protein kinase A (PKA)-dependent anion channel. A simple assay for this CFTR function is to measure the ability of cells transfected with CFTR to increase their conductance to 125I, (a convenient surrogate for Cl-) after exposure to forskolin, which elevates cAMP via a direct action on adenylate cyclase . HEK cells transfected with plasmid alone do not respond to forskolin, while cells transfected with wtCFTR show rapid increases in 125I efflux. If mutations lead to diminished CFTR function, cells transfected with such mutations are predicted to show efflux responses that are significantlyreduced relative to those from cells expressing wt CFTR.

 

 

Cells were incubated at 37° C for 2 hr in efflux buffer containing ~2 m Ci of 125I /ml, then washed 3X with 1 ml aliquots of 22° C buffer. Efflux samples were collected at 30 sec intervals with total fluid replacement. Remaining counts were removed by lysing cells, scintillation fluid was added, and samples were counted in a Beckman liquid scintillation counter. Efflux rate constants were estimated according to the formula given by Venglarik et al. .

14. Establish breeding groups Animals carrying mutations that lead to diminished CFTR function are selected for breeding to homozygosity. All animals used for this study were already maintained as part of breeding populations in NIH primate facilities. Thus breeding groups were establishment by housing the animals together in a facility separate from other animals.
15. Screen offspring for homozygous individuals. As offspring are born in the breeding groups, they are genotyped and then either maintained within the group or returned to the main colony. Homozygous animals are subjected to physiological and biochemical examinations as appropriate to assess possible phenotypic manifestations of the altered genes.

Supported by the Cystic Fibrosis Foundation, by NIH HL51776, and by RR00169 to the California Regional Primate Research Center. We thank the staffs of the Primate Research Centers in California, Louisiana, Oregon and Washington; especially Jenny Short, Phil Allen, Ron Walgenbach, Margaret Clarke, Mark Murchison, Steve Kelley and Debra Glanister. S. Vuillaumier, INSERM, Paris supplied the sequence of exon 1 and flanking segments from several primate species. We thank Ron Kopito and Cristi Ward for supplying 293 cells, the pRBG4 vector, CFTR-pRBG4, and help with transfection protocols. Numerous individuals assisted with SSCP and functional analysis, especially Gregory Hurlock, Eugene Kuo, Mauri Krouse, Clare Robinson, Margaret Lee, Uros Potocnik, and Metka Ravnik-Glava¹ .

1. Identify Animal Population
2. Obtain blood
3. Purify DNA
4. Design Primers
5. Amplify and label Exons
6. Denature-Renature
7. Run on non-denaturing gel
8. Identify shifts
9. Sequence
10. Identify candidate mutations
11. Introduce mutations into cDNA
12. Express mutated cDNA
13. Assay
14. Identify candidate carriers and establish breeding groups
15. Screen offspring for homozygous individuals