Genome-edited livestock to secure sustainability

Sequence variant discovery

A list of potential candidate SVs for genome editing were identified by several different methods. A literature search for causative sequence variants with modest effects on milk yield and composition identified a single missense mutation each in diacylglycerol O-acyltransferase 1 (DGAT1) (Grisart et al. 2002) and growth hormone receptor (GHR) (Viitala et al. 2006). Likewise, pleomorphic adenoma gene 1 (PLAG1) was identified to have three putative causative polymorphisms for antagonistic effects on stature and age of puberty (Utsunomiya et al. 2017). The six mutations for heat tolerance found in the prolactin receptor (PRLR) gene were also identified (Flórez Murillo et al. 2021), and the genotyping results for this locus will be reported elsewhere (T Sonstegard, unpubl. data).

The genes ferredoxin 2 (FDX2) and dehydrogenase/reductase 4 (DHRS4) were detected to be under selection in Muturu and N’Dama cattle from West Africa (ES Kim, T Sonstegard, unpubl. data) based on Fst analyses (Edea et al. 2018). A Tajima’s T-test was performed on the SVs found in these genes, and a missense mutation in each gene was found to be evolving under a non-random process (ES Kim, T Sonstegard, unpubl. data). For causative SVs for tolerance to bovine tuberculosis (bTB), sequences of coding genes and microRNAs differentially expressed during immune response to infection from M. bovis were analyzed for contrasting SVs (Hall et al. 2021), when compared between Bos taurus and Bos indicus breeds of cattle. This method of screening was used because greater resistance to bTB has been reported in Zebu breeds compared to taurine breeds (Lee et al. 2024). Changes to protein structure identified through modeling were also considered as criteria for SV selection. Nine genes (macrophage receptor with collagenous structure (MARCO), interferon gamma inducible protein 16 (IFI16), cyclic GMP-AMP synthase (CGAS), interferon regulatory factor 3 (IRF3), nucleotide binding oligomerization domain containing 2 (NOD2), interleukin 1 alpha (IL1A), CXADR like membrane protein (CLMP), NLR family pyrin domain containing 3 (NLRP3), and Macrophage inhibitory cytokine 1 (MIC1)) encompassing 21 SVs were identified as candidates for bTB tolerance (Supplementary Table S1).

SNP genotyping

In total, 34 SVs corresponding to 15 genes were submitted to Geneseek (Lincoln, NE, USA) for iPlex SNP assay design (Table S2). Forward and reverse assay designs were attempted. Every target gene was represented in the iPlex panel (Table S3), but no assays could be designed for two of the three SVs in PLAG1 and the SLICK6 allele in PRLR. Three SVs also had only a single directional design, which taken together with the missing targets resulted in a total of 59 SNP assays for 15 genes.

A panel of 1063 genomic DNA samples representing 40 different cattle breeds that could be grouped into 15 ecotypes were available for genotyping across the candidate gene editing targets (Table 1). Included in this panel were 26 Holsteins and 245 Gir representing the genetics of the embryo donor breeds used to make bESCs for editing and cloning. Genotype allele scores were generated by iPlex assays done by Geneseek. Marker alleles were scored after comparing genotypes from forward and reverse assay results, when possible. Allele scores were converted to allele frequencies determined for each breed and groups of similar breeds (Table 1).

Bovine embryonic stem cells

Procedures for use of animals were approved by the Trans Ova Genetics (Sioux Center, IA, USA) Institutional Animal Care and Use Committee. In vitro fertilized (IVF) embryos were produced from contract matings of Holsteins (N = 8) and Gir (N = 8) from the breeding programs at Progentus (Sioux Center, IA, USA) and ST Genetics (Navasota, TX, USA), respectively. Eggs were harvested by ovum pick up and fertilized in vitro using conventional semen. Between days 6–8 of development at Trans Ova Genetics, the inner cell mass from selected IVF embryos was used to create independent bESCs lines (N = 11) using modified versions of traditional bovine ES-cell derivation methods (Bogliotti et al. 2018).

Genome editing

During all phases of the genome editing process, bESCs were grown with mouse feeder cells. Five days after initial plating, bESC lines were assessed for cell health by staining and pathogen testing (K Martins, J Bostrom, T Sonstegard, D Carlson, unpubl. data). Healthy bESC lines were transfected with between one to four sgRNAs and HDR repair oligos for SV alterations in the following target genes: (1) PRLR, FDX2, DHRS4, and MARCO in Holstein ESCs; or (2) FDX2, ARHGAP15, and GHR in Gir ESCs (J Bostrom, K Martins, T Sonstegard, D Carlson, unpubl. data). Clonal cell colonies were derived from limiting dilution. On-target edits were confirmed by screening colonies via PCR amplification followed by diagnostic restriction fragment length polymorphism or amplicon sequencing. If a single transfection treatment was used, then a total time of approximately 40 days was needed to reach a bESC population ready for the process of somatic cell nuclear transfer (SCNT). A second round of editing and screening for additional edits to reach the desired number homozygous alterations required another 25 days of processing and growth. The final week of growth was partitioned into cells held for either: (1) cloning (frozen for transport); (2) a second round of pathogen testing; and (3) cytogenetic analysis for chromosomal abnormalities.

Somatic cell nuclear transfer

Vials of frozen bESCs were shipped to Trans Ova Genetics for SCNT (Sung et al. 2007). Cells were plated without mouse feeder cells for recovery and expansion prior to SCNT into enucleated eggs collected via ovum pick up from live animals.

Genotyping results and target SV selection

The genotyping results (Table 1) were used to determine if the potential SV targets for disease tolerance to trypanosomes and bTB, stature/fertility, and milk yield were absent or present in the breeds being used for genome editing. This information was also used to better understand if the target SVs were segregating in the base genetics of the indigenous dairy animals of Sub-Saharan Africa, which could be useful information in developing future crossbreeding schemes for maintenance of gene-edited traits.

Assays for determining SV allele frequencies for 11 of the 15 genes passed quality control with a call rate >90% across the more than 1000 samples run on the iPlex system. Assays for three of the bTB response genes failed to genotype and PRLR genotyping data (six SLICK alleles) was removed for separate publication. For the two SV associated with increased milk yield, the missense mutation in GHR provided a breed profile that was most promising as an editing target for Gir ESCs. The high milk allele was not present in the Gir population from Basa Fazenda (Brazil), which averages 18 L/day and is 6 L/day above the typical dairy Gir animal (F. Garcia, pers. comm.). There was some introgression of this allele found in East African Zebu due to Holstein admixture in this experimental herd (Bahbahani et al. 2017). However, there is a complete absence of this allele in dairy animals of West Africa, providing more support for its selection as an editing target. The results of DGAT1 were similar to GHR, but the presence of the low fat allele in some native breeds combined with the negative aspects of reduced fat content relative to human nutrional value allele removed this potential editing target from further consideration. The PLAG1 results demonstrated fixation in Gir and African breeds for the ancient allele associated with moderate growth and stature, while the rapid growth allele was fixed in Holsteins from the ProGentus breeding program. Due to these contrasting results combined with the continued uncertainty for which SV in PLAG1 is causative (D’Occhio et al. 2024), and our previous inability to alter DNA at these positions, we removed this SV from further consideration.

For the candidate SV for disease tolerance to trypanosome and bTB infection, there were some promising results. First, the favorable (alternative) SV for DHRS4 was fixed in West African breeds tolerant to Nagana disease caused by trypanosome infection. The frequency was also high in other admixed cattle in Burkina Faso and was present in both foreign taurine and indicine cattle at a much lower minor allele frequency (Table 1). Unfortunately, this assay did not work in Zebu cattle from West Africa. However, the results were compelling enough to select this SV for editing in Holstein and Gir ESCs. The SV for FDX2, although not fixed in West African taurine cattle, was also selected based on its relatively high alternative allele frequency and absence in both Gir and Holstein ESC donors.

Because three of the six gene targets for bTB tolerance had multiple SV, the genotypes for these genes (IRF3, NOD2, and MARCO) were combined into haplotypes. As expected, the SVs and haplotypes were predictive of sub-species origin and admixture for the genotyped animals across all six bTB candidate genes. Admixed animals often had recombined haplotypes representing years of interbreeding of taurine-zebu mixed animals, like Sanga (Table 1). However, considering the functional nature for two of the SVs in the Zebu haplotype of MARCO, fixation in Gir, and the absence of these alleles in Holsteins, was enough evidence to select the two MARCO SV in close proximity to each with deleterious SIFT scores for genome editing in Holstein ESCs.

In summary, the target SVs selected for alteration in Holstein ESCs for alternative alleles were an allele of PRLR conferring SLICK for heat tolerance (0% allele frequency, data not shown), FDX2 (0% AF), DHRS4 (15% AF), and two SV in MARCO (0% allele frequency). For Gir, the alternative alleles for GHR (0% AF), FDX2 (0% AF), and Rho GTPase Activating Protein 15 (ARHGAP15 – not genotyped) were selected. The latter substitution of ARHGAP15 for DHRS4 in Gir was made possible after COVID-19 delayed procurement of Gir IVF embryos in the USA. Additionally, the more recent findings of an alternative taurine allele for missense mutation had been demonstrated to express phenotypic evidence of tolerance (Obara 2010).

Multiplex genome editing and cloning success rates

Our goals for the overall breeding project were to produce eight multiplex gene-edited animals from each breed with an equal number of each sex (four each). For the Holsteins, 32 ESC lines were produced from the IVF embryos produced from eight different matings (data not shown), and 15 of these lines were selected for genome editing. A total of 58 gene edits were validated in these lines. For SCNT cloning, 12 were selected for embryo production and nine of these lines produced pregnancies (Table 2). The three lines that did not generate any pregnancies were based on a total of 14 transfers (data not shown). From the nine Holstein ESC lines producing pregnancies, an average of five embryos were transferred for each line. The overall success rate for live calves born was 19%, and the post-partum death rate was approximately 6%. Currently, there are 11 Holsteins (five males and six females) representing eight lines (one line has both sexes) on the ground growing and maturing normally.

For the Gir breeding, IVF embryos derived from eight matings produced 17 ESCs; however, some of these lines failed pathogen screening. It was discovered that semen from one of the sires purchased from a private breeder was contaminated with bovine viral diarrhea virus (data not shown). Thus, only six of the ESCs could be selected for editing, and 18 edits have been validated in these lines to date. The rates of SCNT embyro production (7/cell line), and birth of live calves/transferred embryo (21%) is currently comparable to the results for the Holsteins, but the outcomes of the ongoing Gir pregnancies will determine if there are any potential breed differences for this project. To date, only two lines have produced live calves and there are still several active pregnancies that will reach term at the end of 2024. Adding the current Gir data to the totals generated for the Holsteins did not alter the success statistics for live calves 30 days post-partum (Table 2).