Evaluation of sequencing strategies for whole-genome imputation with hybrid peeling.

BACKGROUND: For assembling large whole-genome sequence datasets for routine use in research and breeding, the sequencing strategy should be adapted to the methods that will be used later for variant discovery and imputation. In this study, we used simulation to explore the impact that the sequencing strategy and level of sequencing investment have on the overall accuracy of imputation using hybrid peeling, a pedigree-based imputation method that is well suited for large livestock populations. METHODS: We simulated marker array and whole-genome sequence data for 15 populations with simulated or real pedigrees that had different structures. In these populations, we evaluated the effect on imputation accuracy of seven methods for selecting which individuals to sequence, the generation of the pedigree to which the sequenced individuals belonged, the use of variable or uniform coverage, and the trade-off between the number of sequenced individuals and their sequencing coverage. For each population, we considered four levels of investment in sequencing that were proportional to the size of the population. RESULTS: Imputation accuracy depended greatly on pedigree depth. The distribution of the sequenced individuals across the generations of the pedigree underlay the performance of the different methods used to select individuals to sequence and it was critical for achieving high imputation accuracy in both early and late generations. Imputation accuracy was highest with a uniform coverage across the sequenced individuals of 2× rather than variable coverage. An investment equivalent to the cost of sequencing 2% of the population at 2× provided high imputation accuracy. The gain in imputation accuracy from additional investment decreased with larger populations and higher levels of investment. However, to achieve the same imputation accuracy, a proportionally greater investment must be used in the smaller populations compared to the larger ones. CONCLUSIONS: Suitable sequencing strategies for subsequent imputation with hybrid peeling involve sequencing ~2% of the population at a uniform coverage 2×, distributed preferably across all generations of the pedigree, except for the few earliest generations that lack genotyped ancestors. Such sequencing strategies are beneficial for generating whole-genome sequence data in populations with deep pedigrees of closely related individuals.

Accuracy of whole-genome sequence imputation using hybrid peeling in large pedigreed livestock populations.

BACKGROUND: The coupling of appropriate sequencing strategies and imputation methods is critical for assembling large whole-genome sequence datasets from livestock populations for research and breeding. In this paper, we describe and validate the coupling of a sequencing strategy with the imputation method hybrid peeling in real animal breeding settings. METHODS: We used data from four pig populations of different size (18,349 to 107,815 individuals) that were widely genotyped at densities between 15,000 and 75,000 markers genome-wide. Around 2% of the individuals in each population were sequenced (most of them at 1× or 2× and 37-92 individuals per population, totalling 284, at 15-30×). We imputed whole-genome sequence data with hybrid peeling. We evaluated the imputation accuracy by removing the sequence data of the 284 individuals with high coverage, using a leave-one-out design. We simulated data that mimicked the sequencing strategy used in the real populations to quantify the factors that affected the individual-wise and variant-wise imputation accuracies using regression trees. RESULTS: Imputation accuracy was high for the majority of individuals in all four populations (median individual-wise dosage correlation: 0.97). Imputation accuracy was lower for individuals in the earliest generations of each population than for the rest, due to the lack of marker array data for themselves and their ancestors. The main factors that determined the individual-wise imputation accuracy were the genotyping status, the availability of marker array data for immediate ancestors, and the degree of connectedness to the rest of the population, but sequencing coverage of the relatives had no effect. The main factors that determined variant-wise imputation accuracy were the minor allele frequency and the number of individuals with sequencing coverage at each variant site. Results were validated with the empirical observations. CONCLUSIONS: We demonstrate that the coupling of an appropriate sequencing strategy and hybrid peeling is a powerful strategy for generating whole-genome sequence data with high accuracy in large pedigreed populations where only a small fraction of individuals (2%) had been sequenced, mostly at low coverage. This is a critical step for the successful implementation of whole-genome sequence data for genomic prediction and fine-mapping of causal variants.

Pigs Lacking the Scavenger Receptor Cysteine-Rich Domain 5 of CD163 Are Resistant to Porcine Reproductive and Respiratory Syndrome Virus 1 Infection.

Porcine reproductive and respiratory syndrome virus (PRRSV) has a narrow host cell tropism, limited to cells of the monocyte/macrophage lineage. CD163 protein is expressed at high levels on the surface of specific macrophage types, and a soluble form is circulating in blood. CD163 has been described as a fusion receptor for PRRSV, with the scavenger receptor cysteine-rich domain 5 (SRCR5) region having been shown to be the interaction site for the virus. As reported previously, we have generated pigs in which exon 7 of the CD163 gene has been deleted using CRISPR/Cas9 editing in pig zygotes. These pigs express CD163 protein lacking SRCR5 (ΔSRCR5 CD163) and show no adverse effects when maintained under standard husbandry conditions. Not only was ΔSRCR5 CD163 detected on the surface of macrophage subsets, but the secreted, soluble protein can also be detected in the serum of the edited pigs, as shown here by a porcine soluble CD163-specific enzyme-linked immunosorbent assay (ELISA). Previous results showed that primary macrophage cells from ΔSRCR5 CD163 animals are resistant to PRRSV-1 subtype 1, 2, and 3 as well as PRRSV-2 infection in vitro Here, ΔSRCR5 pigs were challenged with a highly virulent PRRSV-1 subtype 2 strain. In contrast to the wild-type control group, ΔSRCR5 pigs showed no signs of infection and no viremia or antibody response indicative of a productive infection. Histopathological analysis of lung and lymph node tissue showed no presence of virus-replicating cells in either tissue. This shows that ΔSRCR5 pigs are fully resistant to infection by the virus.IMPORTANCE Porcine reproductive and respiratory syndrome (PRRS) virus (PRRSV) is the etiological agent of PRRS, causing late-term abortions, stillbirths, and respiratory disease in pigs, incurring major economic losses to the worldwide pig industry. The virus is highly mutagenic and can be divided into two species, PRRSV-1 and PRRSV-2, each containing several subtypes. Current control strategies mainly involve biosecurity measures, depopulation, and vaccination. Vaccines are at best only partially protective against infection with heterologous subtypes and sublineages, and modified live vaccines have frequently been reported to revert to virulence. Here, we demonstrate that a genetic-control approach results in complete resistance to PRRSV infection in vivo CD163 is edited so as to remove the viral interaction domain while maintaining protein expression and biological function, averting any potential adverse effect associated with protein knockout. This research demonstrates a genetic-control approach with potential benefits in animal welfare as well as to the pork industry.

Precision engineering for PRRSV resistance in pigs: Macrophages from genome edited pigs lacking CD163 SRCR5 domain are fully resistant to both PRRSV genotypes while maintaining biological function.

Porcine Reproductive and Respiratory Syndrome (PRRS) is a panzootic infectious disease of pigs, causing major economic losses to the world-wide pig industry. PRRS manifests differently in pigs of all ages but primarily causes late-term abortions and stillbirths in sows and respiratory disease in piglets. The causative agent of the disease is the positive-strand RNA PRRS virus (PRRSV). PRRSV has a narrow host cell tropism, limited to cells of the monocyte/macrophage lineage. CD163 has been described as a fusion receptor for PRRSV, whereby the scavenger receptor cysteine-rich domain 5 (SRCR5) region was shown to be an interaction site for the virus in vitro. CD163 is expressed at high levels on the surface of macrophages, particularly in the respiratory system. Here we describe the application of CRISPR/Cas9 to pig zygotes, resulting in the generation of pigs with a deletion of Exon 7 of the CD163 gene, encoding SRCR5. Deletion of SRCR5 showed no adverse effects in pigs maintained under standard husbandry conditions with normal growth rates and complete blood counts observed. Pulmonary alveolar macrophages (PAMs) and peripheral blood monocytes (PBMCs) were isolated from the animals and assessed in vitro. Both PAMs and macrophages obtained from PBMCs by CSF1 stimulation (PMMs) show the characteristic differentiation and cell surface marker expression of macrophages of the respective origin. Expression and correct folding of the SRCR5 deletion CD163 on the surface of macrophages and biological activity of the protein as hemoglobin-haptoglobin scavenger was confirmed. Challenge of both PAMs and PMMs with PRRSV genotype 1, subtypes 1, 2, and 3 and PMMs with PRRSV genotype 2 showed complete resistance to viral infections assessed by replication. Confocal microscopy revealed the absence of replication structures in the SRCR5 CD163 deletion macrophages, indicating an inhibition of infection prior to gene expression, i.e. at entry/fusion or unpacking stages.

Resistance to coronavirus infection in amino peptidase N-deficient pigs.

The alphacoronaviruses, transmissible gastroenteritis virus (TGEV) and Porcine epidemic diarrhea virus (PEDV) are sources of high morbidity and mortality in neonatal pigs, a consequence of dehydration caused by the infection and necrosis of enterocytes. The biological relevance of amino peptidase N (ANPEP) as a putative receptor for TGEV and PEDV in pigs was evaluated by using CRISPR/Cas9 to edit exon 2 of ANPEP resulting in a premature stop codon. Knockout pigs possessing the null ANPEP phenotype and age matched wild type pigs were challenged with either PEDV or TGEV. Fecal swabs were collected daily from each animal beginning 1 day prior to challenge with PEDV until the termination of the study. The presence of virus nucleic acid was determined by PCR. ANPEP null pigs did not support infection with TGEV, but retained susceptibility to infection with PEDV. Immunohistochemistry confirmed the presence of PEDV reactivity and absence of TGEV reactivity in the enterocytes lining the ileum in ANPEP null pigs. The different receptor requirements for TGEV and PEDV have important implications in the development of new genetic tools for the control of enteric disease in pigs.

Replacement of Porcine CD163 Scavenger Receptor Cysteine-Rich Domain 5 with a CD163-Like Homolog Confers Resistance of Pigs to Genotype 1 but Not Genotype 2 Porcine Reproductive and Respiratory Syndrome Virus.

CD163 knockout (KO) pigs are resistant to infection with genotype 2 (type 2) porcine reproductive and respiratory syndrome virus (PRRSV). Furthermore, the substitution of CD163 scavenger receptor cysteine-rich (SRCR) domain 5 with a homolog of human CD163-like (hCD163L1) SRCR 8 domain confers resistance of transfected HEK cells to type 1 PRRSV. As a means to understand the role of domain 5 in PRRSV infection with both type 1 and type 2 viruses, pigs were genetically modified (GM) to possess one of the following genotypes: complete knockout (KO) of CD163, deletions within SRCR domain 5, or replacement (domain swap) of SRCR domain 5 with a synthesized exon encoding a homolog of hCD163L1 SRCR domain 8. Immunophenotyping of porcine alveolar macrophages (PAMs) showed that pigs with the KO or SRCR domain 5 deletion did not express CD163. When placed in culture, PAMs from pigs with the CD163 KO phenotype were completely resistant to a panel consisting of six type 1 and nine type 2 isolates. PAMs from pigs that possessed the hCD163L1 domain 8 homolog expressed CD163 and supported the replication of all type 2 isolates, but no type 1 viruses. Infection of CD163-modified pigs with representative type 1 and type 2 viruses confirmed the in vitro results. The results confirm that CD163 is the likely receptor for all PRRS viruses. Even though type 1 and type 2 viruses are considered phenotypically similar at several levels, there is a distinct difference between the viral genotypes in the recognition of CD163. IMPORTANCE: Genetic modification of the CD163 gene creates the opportunity to develop production animals that are resistant to PRRS, the costliest viral disease to ever face the swine industry. The results create further opportunities to develop refinements in the modification of CD163 with the goal of making pigs refractory to infection while retaining important CD163 functions.

Impact of index hopping and bias towards the reference allele on accuracy of genotype calls from low-coverage sequencing.

BACKGROUND: Inherent sources of error and bias that affect the quality of sequence data include index hopping and bias towards the reference allele. The impact of these artefacts is likely greater for low-coverage data than for high-coverage data because low-coverage data has scant information and many standard tools for processing sequence data were designed for high-coverage data. With the proliferation of cost-effective low-coverage sequencing, there is a need to understand the impact of these errors and bias on resulting genotype calls from low-coverage sequencing. RESULTS: We used a dataset of 26 pigs sequenced both at 2× with multiplexing and at 30× without multiplexing to show that index hopping and bias towards the reference allele due to alignment had little impact on genotype calls. However, pruning of alternative haplotypes supported by a number of reads below a predefined threshold, which is a default and desired step of some variant callers for removing potential sequencing errors in high-coverage data, introduced an unexpected bias towards the reference allele when applied to low-coverage sequence data. This bias reduced best-guess genotype concordance of low-coverage sequence data by 19.0 absolute percentage points. CONCLUSIONS: We propose a simple pipeline to correct the preferential bias towards the reference allele that can occur during variant discovery and we recommend that users of low-coverage sequence data be wary of unexpected biases that may be produced by bioinformatic tools that were designed for high-coverage sequence data.

Potential of gene drives with genome editing to increase genetic gain in livestock breeding programs.

BACKGROUND: This paper uses simulation to explore how gene drives can increase genetic gain in livestock breeding programs. Gene drives are naturally occurring phenomena that cause a mutation on one chromosome to copy itself onto its homologous chromosome. METHODS: We simulated nine different breeding and editing scenarios with a common overall structure. Each scenario began with 21 generations of selection, followed by 20 generations of selection based on true breeding values where the breeder used selection alone, selection in combination with genome editing, or selection with genome editing and gene drives. In the scenarios that used gene drives, we varied the probability of successfully incorporating the gene drive. For each scenario, we evaluated genetic gain, genetic variance [Formula: see text], rate of change in inbreeding ([Formula: see text]), number of distinct quantitative trait nucleotides (QTN) edited, rate of increase in favourable allele frequencies of edited QTN and the time to fix favourable alleles. RESULTS: Gene drives enhanced the benefits of genome editing in seven ways: (1) they amplified the increase in genetic gain brought about by genome editing; (2) they amplified the rate of increase in the frequency of favourable alleles and reduced the time it took to fix them; (3) they enabled more rapid targeting of QTN with lesser effect for genome editing; (4) they distributed fixed editing resources across a larger number of distinct QTN across generations; (5) they focussed editing on a smaller number of QTN within a given generation; (6) they reduced the level of inbreeding when editing a subset of the sires; and (7) they increased the efficiency of converting genetic variation into genetic gain. CONCLUSIONS: Genome editing in livestock breeding results in short-, medium- and long-term increases in genetic gain. The increase in genetic gain occurs because editing increases the frequency of favourable alleles in the population. Gene drives accelerate the increase in allele frequency caused by editing, which results in even higher genetic gain over a shorter period of time with no impact on inbreeding.

Genome edited sheep and cattle.

Genome editing tools enable efficient and accurate genome manipulation. An enhanced ability to modify the genomes of livestock species could be utilized to improve disease resistance, productivity or breeding capability as well as the generation of new biomedical models. To date, with respect to the direct injection of genome editor mRNA into livestock zygotes, this technology has been limited to the generation of pigs with edited genomes. To capture the far-reaching applications of gene-editing, from disease modelling to agricultural improvement, the technology must be easily applied to a number of species using a variety of approaches. In this study, we demonstrate zygote injection of TALEN mRNA can also produce gene-edited cattle and sheep. In both species we have targeted the myostatin (MSTN) gene. In addition, we report a critical innovation for application of gene-editing to the cattle industry whereby gene-edited calves can be produced with specified genetics by ovum pickup, in vitro fertilization and zygote microinjection (OPU-IVF-ZM). This provides a practical alternative to somatic cell nuclear transfer for gene knockout or introgression of desirable alleles into a target breed/genetic line.

Highly efficient targeted chromosome deletions using CRISPR/Cas9.

The CRISPR/Cas9 system has emerged as an intriguing new technology for genome engineering. It utilizes the bacterial endonuclease Cas9 which, when delivered to eukaryotic cells in conjunction with a user-specified small guide RNA (gRNA), cleaves the chromosomal DNA at the target site. Here we show that concurrent delivery of gRNAs designed to target two different sites in a human chromosome introduce DNA double-strand breaks in the chromosome and give rise to targeted deletions of the intervening genomic segment. Predetermined genomic DNA segments ranging from several-hundred base pairs to 1 Mbp can be precisely deleted at frequencies of 1-10%, with no apparent correlation between the size of the deleted fragment and the deletion frequency. The high efficiency of this technique holds promise for large genomic deletions that could be useful in generation of cell and animal models with engineered chromosomes.

Design and development of exome capture sequencing for the domestic pig (Sus scrofa).

BACKGROUND: The domestic pig (Sus scrofa) is both an important livestock species and a model for biomedical research. Exome sequencing has accelerated identification of protein-coding variants underlying phenotypic traits in human and mouse. We aimed to develop and validate a similar resource for the pig. RESULTS: We developed probe sets to capture pig exonic sequences based upon the current Ensembl pig gene annotation supplemented with mapped expressed sequence tags (ESTs) and demonstrated proof-of-principle capture and sequencing of the pig exome in 96 pigs, encompassing 24 capture experiments. For most of the samples at least 10x sequence coverage was achieved for more than 90% of the target bases. Bioinformatic analysis of the data revealed over 236,000 high confidence predicted SNPs and over 28,000 predicted indels. CONCLUSIONS: We have achieved coverage statistics similar to those seen with commercially available human and mouse exome kits. Exome capture in pigs provides a tool to identify coding region variation associated with production traits, including loss of function mutations which may explain embryonic and neonatal losses, and to improve genomic assemblies in the vicinity of protein coding genes in the pig.

Use of the CRISPR/Cas9 system to produce genetically engineered pigs from in vitro-derived oocytes and embryos.

Targeted modification of the pig genome can be challenging. Recent applications of the CRISPR/Cas9 system hold promise for improving the efficacy of genome editing. When a designed CRISPR/Cas9 system targeting CD163 or CD1D was introduced into somatic cells, it was highly efficient in inducing mutations. When these mutated cells were used with somatic cell nuclear transfer, offspring with these modifications were created. When the CRISPR/Cas9 system was delivered into in vitro produced presumptive porcine zygotes, the system was effective in creating mutations in eGFP, CD163, and CD1D (100% targeting efficiency in blastocyst stage embryos); however, it also presented some embryo toxicity. We could also induce deletions in CD163 or CD1D by introducing two types of CRISPRs with Cas9. The system could also disrupt two genes, CD163 and eGFP, simultaneously when two CRISPRs targeting two genes with Cas9 were delivered into zygotes. Direct injection of CRISPR/Cas9 targeting CD163 or CD1D into zygotes resulted in piglets that have mutations on both alleles with only one CD1D pig having a mosaic genotype. We show here that the CRISPR/Cas9 system can be used by two methods. The system can be used to modify somatic cells followed by somatic cell nuclear transfer. System components can also be used in in vitro produced zygotes to generate pigs with specific genetic modifications.

Porcine insulin receptor substrate 4 (IRS4) gene: cloning, polymorphism and association study.

Using PCR and inverse PCR techniques we obtained a 4,498 bp nucleotide sequence FN424076 encompassing the complete coding sequence of the porcine insulin receptor substrate 4 (IRS4) gene and its proximal promoter. The 1,269 amino acid porcine protein deduced from the nucleotide sequence shares 92% identity with the human IRS4 and possesses the same domains and the same number of tyrosine phosphorylation motifs as the human protein. We detected substitution FN424076:g.96C

Substitution of warthog NF-κB motifs into RELA of domestic pigs is not sufficient to confer resilience to African swine fever virus.

African swine fever virus (ASFV) causes a lethal, haemorrhagic disease in domestic swine that threatens pig production across the globe. Unlike domestic pigs, warthogs, which are wildlife hosts of the virus, do not succumb to the lethal effects of infection. There are three amino acid differences between the sequence of the warthog and domestic pig RELA protein; a subunit of the NF-κB transcription factor that plays a key role in regulating the immune response to infections. Domestic pigs with all 3 or 2 of the amino acids from the warthog RELA orthologue have been generated by gene editing. To assess if these variations confer resilience to ASF we established an intranasal challenge model with a moderately virulent ASFV. No difference in clinical, virological or pathological parameters were observed in domestic pigs with the 2 amino acid substitution. Domestic pigs with all 3 amino acids found in warthog RELA were not resilient to ASF but a delay in onset of clinical signs and less viral DNA in blood samples and nasal secretions was observed in some animals. Inclusion of these and additional warthog genetic traits into domestic pigs may be one way to assist in combating the devastating impact of ASFV.

A Strategy To Exploit Surrogate Sire Technology in Livestock Breeding Programs.

In this work, we performed simulations to develop and test a strategy for exploiting surrogate sire technology in animal breeding programs. Surrogate sire technology allows the creation of males that lack their own germline cells, but have transplanted spermatogonial stem cells from donor males. With this technology, a single elite male donor could give rise to huge numbers of progeny, potentially as much as all the production animals in a particular time period. One hundred replicates of various scenarios were performed. Scenarios followed a common overall structure but differed in the strategy used to identify elite donors and how these donors were used in the product development part. The results of this study showed that using surrogate sire technology would significantly increase the genetic merit of commercial sires, by as much as 6.5 to 9.2 years' worth of genetic gain compared to a conventional breeding program. The simulations suggested that a strategy involving three stages (an initial genomic test followed by two subsequent progeny tests) was the most effective of all the strategies tested. The use of one or a handful of elite donors to generate the production animals would be very different to current practice. While the results demonstrate the great potential of surrogate sire technology there are considerable risks but also other opportunities. Practical implementation of surrogate sire technology would need to account for these.

Gene expression profiling in porcine maternal infanticide: a model for puerperal psychosis.

The etiology of mental disorders remains largely unclear. Complex interactions between genetic and environmental factors are key to the development of such disorders. Puerperal psychosis is the most extreme form of postnatal mood disorder in women. Similarly, parturition in the pig can trigger extreme behavioral disturbances, including maternal infanticide. In this study, we have used a targeted cDNA microarray approach using the pig as a model to understand the genes and genetic pathways that are involved in these processes. Two subtracted cDNA libraries from porcine hypothalamus were constructed, which were enriched for genes that were over-expressed and under-expressed in the aberrant behavioral phenotype, compared to the matched control. In addition to this, a normalized library was constructed from hypothalamus and pituitary samples taken from pigs in a variety of reproductive states. The libraries were partially sequenced and combined represented approximately 5,159 different genes. Microarray analysis determined differences in gene expression between hypothalamus samples from nine matched pairs of infanticidal versus control animals, using a common reference design. Microarray analysis of variance (MAANOVA) identified 52 clones as being differentially expressed (P

Targeted gene knock-in by CRISPR/Cas ribonucleoproteins in porcine zygotes.

The domestic pig is an important "dual purpose" animal model for agricultural and biomedical applications. There is an emerging consensus in the biomedical community for the use of large animal models such as pigs to either serve as an alternative, or complement investigations from the mouse. However, the use of pig has not proven popular due to technical difficulties and time required in generating models with desired genetic modifications. In this regard, the ability to directly modify the genome in the zygote and generate edited animals is highly desirable. This report demonstrates for the first time, the generation of gene targeted animals by direct injection of Cas9 ribonucleoprotein complex and short stretches of DNA sequences into porcine zygotes. The Cas9 protein from Streptococcus pyogenes was pre-complexed with a single guide RNA targeting downstream of the ubiquitously expressed COL1A gene, and co-injected with a single-stranded repair template into porcine zygotes. Using this approach a line of pigs that carry pseudo attP sites within the COL1A locus to enable phiC31 integrase mediated introduction of transgenes has been generated. This new route for genome engineering in pigs via zygote injection should greatly enhance applications in both agriculture and biomedicine.

Knockout of maternal CD163 protects fetuses from infection with porcine reproductive and respiratory syndrome virus (PRRSV).

After infection of the porcine dam at about 90 days of gestation, porcine reproductive and respiratory syndrome virus (PRRSV) crosses the placenta and begins to infect fetuses. Outcomes of include abortion, fetal death and respiratory disease in newborn piglets. CD163 is the receptor for the virus. In this study, CD163-positive fetuses, recovered between 109 days of gestation or 20 days after maternal infection, were completely protected from PRRSV in dams possessing a complete knockout of the CD163 receptor. The results demonstrate a practical means to eliminate PRRSV-associated reproductive disease, a major source of economic hardship to agriculture.

Genetically edited pigs lacking CD163 show no resistance following infection with the African swine fever virus isolate, Georgia 2007/1.

African swine fever is a highly contagious, often fatal disease of swine for which there is no vaccine or other curative treatment. The macrophage marker, CD163, is a putative receptor for African swine fever virus (ASFV). Pigs possessing a complete knockout of CD163 on macrophages were inoculated with Georgia 2007/1, a genotype 2 isolate. Knockout and wild type pen mates became infected and showed no differences in clinical signs, mortality, pathology or viremia. There was also no difference following in vitro infection of macrophages. The results do not rule out the possibility that other ASFV strains utilize CD163, but demonstrate that CD163 is not necessary for infection with the Georgia 2007/1 isolate. This work rules out a significant role for CD163 in ASFV infection and creates opportunities to focus on alternative receptors and entry mechanisms.