User:Eamonn B. Mallon/Notebook/Monoparental expression in worker reproduction genes/2010/04/16

Monoparental expression

Main project page Next entry

Entry title Started up lab notebook. Description of proposed research Genomic imprinting (GI) is the inactivation of one allele in diploid individuals, with inactivation being dependent upon the sex of the parent from which it was derived1. Natural selection is expected to favor expression of both alleles in order to protect against recessive mutations that render a gene ineffective2. What, then, is the benefit of silencing one copy of a gene, making the organism functionally haploid at that locus? Haig’s conflict hypothesis is the current leading explanation. The conflict theory is based on the fact that maternally (matrigene) and paternally (patrigene) inherited genes in the same organism can have different interests1. In a species with multiple paternity, a patrigene has a lower probability of being present in siblings that are progeny of the same mother than does a matrigene. As a result, a patrigene will be selected to value the survival of the organism it is in more highly, compared to the survival of siblings. In mammals and angiosperms, this conflict is played out in the provisioning of offspring with resources taken from the mother. A patrigene will benefit by causing more maternal resources to be allocated to the organism it is in, but a matrigene will benefit from sharing resources among all the siblings. Therefore some genes in mammalian embryos are inactive (imprinted) when maternally inherited (putative growth enhancers) while others are inactive when paternally inherited (putative growth suppressors). Although current observations support the conflict theory, there have been no independent tests. Eusocial Hymenoptera (ants, bees and wasps) are an ideal model systems for making truly independent a priori tests of the theory3. Hymenoptera are haplodiploid, with diploid females (queens and workers) arising from fertilized eggs and haploid males arising from unfertilized eggs. This different genetic system combined with the shared brood rearing and reproductive division of labour characteristic of eusociality results in range of novel predictions for the conflict theory3. One major problem in using eusocial Hymenoptera to test the conflict theory was that they were not known to have the mechanisms required for genomic imprinting. In 2006, it was shown that Apis mellifera has a fully functioning cytosine-phosphate-guanine (CpG) methylation system4. DNA methylation is one of the major mechanisms of genomic imprinting in mammals and angiosperms5. CpG methylation has since been shown to be common in social hymenoptera6and the Mallon lab have found it in bumblebees. The honeybee methylation system has been shown to transmit epigenetic information that determines whether a larva will develop into a worker or a queen7. To test the conflict theory, a bumblebee reproductive trait is required for which the theory predicts conflict between an individual’s matrigenes and patrigenes. Worker reproduction is such a trait. Bumblebee workers all possess functional ovaries and can lay unfertilized eggs that develop into males if a colony becomes queenless8. The matrigenes and patrigenes in workers have different selectional pressures in these queenless colonies. A matrigene in a given worker has a 50% chance of ending up in that worker’s son, but only a 25% chance of being in a different worker’s son. The equivalent patrigene is equally likely to be in the focal worker’s son or her sister’s son (50%). This asymmetry stems from the fact that there is only one patrigene in the whole colony, as the father was haploid, but there are two potential matrigenes as the mother was diploid. For a patrigene, it is of equal value if the worker it is in reproduces or the worker’s siblings reproduce, for a matrigene it is twice as beneficial if the worker herself reproduces. As a result of this asymmetry, the conflict theory predicts that genes whose upregulation is associated with worker reproduction should be expressed if they are matrigenes but not if they are patrigenes, that is, these patrigenes should be imprinted3. The reverse would be the case for genes that are downregulated during worker reproduction. Methylation differences between reproducing and non-reproducing workers: We have preliminary data showing that the methylation pattern differs between different reproductive castes (Figure 1). The arrow highlights a band that is different between the samples. A much larger number are visible on the gel. Notice that the queen and the reproducing worker share a similar banding pattern compared to a non-reproducing worker. We will repeat this experiment (5 replicates across 3 colonies) and sequence the resultant bands to identify genes that differ in methylation between reproducing and non-reproducing workers. Methylation differences will be detected using amplification of inter-methylated sites (AIMS)9. We will take genomic DNA from each of the samples and cleave them with the methylation sensitive restriction enzyme SmaI. This cuts the recognition site CCCGGG into 2 blunt ends. A second digestion by the methylation insensitive isoschizomer XmaI will cut the methylated recognition sites leaving sticky ends. Specific adaptors are ligated to the methylated ends of the digested genomic DNA. The ligated sequences are amplified by PCR using adaptor-specific primers extended at the 3' end with two to four arbitrarily chosen nucleotide residues. The samples will then be run on a 9% Polynat© gel. Fingerprints consist of multiple anonymous bands, representing DNA sequences flanked by two methylated sites. We will isolate bands from the gel and sequence them. We will identify these sequences by means of homology searches of major databases. If required we will use the bumblebee BAC library to obtain the full sequence10. The methylation status of every locus identified by AIMS will be confirmed by bisulphite sequencing that will also identify the position of methylated residues. Selection of candidate genes for worker reproduction: Added to the above genes, we will also look at genes from the literature known to be involved with worker reproduction. Juvenille hormone (JV) affects ovary development11, Major royal jelly proteins (MRJP2, MRJP7) and Neimann-Pick type C2 are differentially expressed in sterile and anarchistic honeybees. Anarchistic honeybee colonies contain a large proportion of reproducing workers12. Vitellogenin is upregulated in ovary activated honeybee workers13.Yellow-g2 is involved in insect oogeneisis14 and has been shown to be expressed in egg laying ants15. Chymotrypsin, Cytochrome oxidase and Peroxiredoxin are up-regulated in B. terrestris non-reproducing workers when compared to reproducing workers16. We will test if these genes (methylated and literature) are involved in bumblebee reproduction by looking at their expression difference using quantitative PCR (qPCR) between ten day old reproducing and non-reproducing workers (3 technical and 2 biological replicates). Identification of matrigenes and patrigenes: We will use single stranded conformation polymorphism (SSCP) to find the variation in our candidate genes. Single stranded DNA may experience intra-strand base pairing, resulting in folds that give the single strand a unique 3D structure depending on its sequence causing it to run on a gel at different speeds. We use GMA gels from Elchrom in their Origins electrophoresis system. We will genotype the queen non-destructively and ten workers from up to thirty colonies. Matrigenes for a given gene are the two alleles found in a heterozygote queen or the single allele found in a homozygote queen. The ten workers will contain one of these matrigenes, the other allele in the workers being the single patrigene from the haploid father. A potential problem is if there is no variation in a given gene within a colony (workers are homozygote), we will be unable to tell the patrigenes and matrigenes apart. This seems unlikely as SSCP is extremely sensitive17. We have examined 15 colonies for variation in chymotrypsin and have found heterozygote workers in 14. If however we do come across this problem, one solution is to collect males and queens from different colonies and mate then in a controlled breeding program. The next section could then be carried out on these lab bred colonies. Expression differences between matrigenes and patrigenes: We now have colonies where we know the patrigenes and matrigenes for worker reproduction candidate genes. We predict that for genes involved in the upregulation of worker reproduction only the matrigene will be expressed in workers from queenless colonies. We will measure expression by reverse transcription PCR SSCP. If the cDNA from a reproducing worker is run on a GMA gel only the matrigene should be present. Matrigene/patrigene expression will be verified in at least 5 workers for each colony. We may also develop allele specific primers and look at gene expression using qPCR. We have allowed for the analysis of up to 20 candidate genes. A gene where only the matrigene is expressed is the first imprinted gene discovered in an insect and will confirm conflict theory. References1Haig, D., Ann. Rev. Ecol. and Syst. 31, 9 (2000).2Hurst, L. D. Genomic imprinting (1997).3Queller, D. C., BMC Evol. Biol. 3,15 (2003).4Wang, Y. et al., Sci. 314 (5799), 645 (2006).5Li, E. et al., Nat. 366 (6453), 362 (1993).6Kronforst, M. R. et al., Curr. Biol. 18 (7), R287 (2008).7Kucharski, R. et al., Sci. 319 (5871), 1827 (2008).8Alaux, C. et al., Behav Ecol Soc. 62 (2), 213 (2007).9Frigola, J. et al., Nuc. Acids Res 30 (7) (2002).10Wilfert, L. et al., Insect. Soc. 56 (1), 44 (2009).11Barchuk, A. R. et al., BMC Dev. Biol.7 (2007).12Thompson, G. J. et al., J.Exp. Zoo. Part a 307A (10), 600 (2007).13Koywiwattrakul, P. et al., J Insect Sci 5 (2006).14Claycomb, J. M. et al., Dev Cell 6 (1), 145 (2004).15Graff, J. et al., Mol. Ecol. 16 (3), 675 (2007).16Pereboom, J. J. M. et al., P R SOC B 272 (1568), 1145 (2005).17Sunnucks, P. et al., Mol. Ecol. 9 (11), 1699 (2000).