«DISSERTATION zur Erlangung des Grades doctor rerum naturalium (Dr. rer. nat.) vorgelegt von Sebastian Haesler im Fachbereich Biologie, Chemie, ...»
Studies on the evolution and function of FoxP2, a gene
implicated in human speech and language, using
songbirds as a model
zur Erlangung des Grades
doctor rerum naturalium
(Dr. rer. nat.)
im Fachbereich Biologie, Chemie, Pharmazie
Freien Universität Berlin
Erstgutachter: Prof. Constance Scharff
Zweitgutachter: Prof. Hans-Joachim Pflüger
Language and Genes can be represented by Letters The letters of the Phoenician script shown on top of the DNA sequence of the FoxP2 gene from Taeniopygia guttata (Zebra finch). The Phoenician script which emanated during the 14th century BC was one of the first truly
alphabets. Modern alphabets thought to have descended from Phoenician include Arabic, Greek, Latin (via the Old Italic alphabet), Cyrillic (via the Greek alphabet) and Hebrew (via Aramaic).
FoxP2 is the first gene known to be involved in human speech and language.
3 Contents Contents 1 Introduction
1.1 FoxP2 and Developmental Verbal Dyspraxia
1.2 FoxP2 Expression in the Brains of Mice and Men
1.3 Molecular Function of FoxP2
1.4 FoxP2 Knockout Mouse
1.5 Human Speech and Birdsong
1.6 Genes for Vocal Learning ?
1.7 Analysis of Gene Function by Genetic Manipulation in Songbirds
1.8 The emergence of Vocal Learning and the Molecular Evolution of FoxP2........ 26 1.9 Aims of this Study
2 Materials and Methods
2.1 Solutions and Buffer
2.4 Molecular Biology
2.4.1 Sex determination of young zebra finches
2.4.2 RNA extraction from zebra finch tissue
2.4.3 Northern blotting
2.4.4 Cloning of FoxP2 and FoxP1 from zebra finch
2.4.5 Cloning of FoxP2 and FoxP1 from other avian species and crocodile...... 35 2.4.6 Cloning of V5-tagged FoxP2 expression constructs
2.4.7 Cloning of constructs encoding short-hairpin RNA
2.4.8 Preparation of Plasmid DNA
2.4.10 Sequence analysis
2.4.11 In situ hybridization
2.4.12 Real-time PCR
Contents 2.5 Protein Biochemistry
2.5.1 Protein extraction from cultured cells
2.5.2 Protein extraction from zebra finch brain tissue
2.5.3 BCA assay for protein quantification
2.5.4 Western blot
2.6 Knockdown Efficiency of Hairpin Constructs in vitro
2.7 Generation of Lentivirus
2.8 Surgery and Stereotactic Injection of Virus
2.9 Behavioral Paradigm and Song Analysis
3.1 Expression Pattern of FoxP2 and FoxP1
3.1.1 Cloning of the zebra finch FoxP2 and FoxP1 genes
3.1.2 Embryonic FoxP2 expression
3.1.3 Subtelencephalic FoxP2 expression in the adult zebra finch
3.1.4 Expression of FoxP2 in the adult telencephalon
3.1.5 FoxP1 expression
3.1.6 FoxP2 expression during times of song plasticity
3.1.7 Zebra finch FoxP2 expression and singing
3.1.8 Cellular identity of FoxP2 expressing cells
3.2 Knockdown of FoxP2 in vivo
3.2.1 Establishing lentivirus-mediated RNAi in the zebra finch
3.2.2 Behavioral consequence of FoxP2 knockdown
3.3 Comparison of FoxP2 Sequences from Birds and the Crocodile
4.1 FoxP2 Expression in Avian Vocal Learners and Non-learners
4.2 Analysis of FoxP2 Function in vivo
4.3 Molecular Evolution of FoxP2 in Avians
1 Introduction Learning to imitate sounds and the rules of grammar endows humans with the unique ability to communicate infinite meaning with a finite vocal repertoire using language.
Although language is learned, a genetic bias towards this learning has already been proposed in Charles Darwin´s “Descent of Man” (Chapter III - Comparison of the Mental
Powers of Man and the Lower Animals):
“ […] language is an art, like brewing or baking; but writing would have been a better simile. It certainly is not a true instinct, for every language has to be learnt. It differs, however, widely from all ordinary arts, for man has an instinctive tendency to speak, as we see in the babble of our young children; whilst no child has an instinctive tendency to brew, bake, or write.” A modern account of the idea that learning language is not solely based upon experience was put forward by the Linguist Noam Chomsky. He developed the concept of a “universal grammar”, which posits the existence of a universal set of rules common to all languages (Chomsky, 1957). This universal grammar shared by all languages suggests that some aspects of how language is learned are determined by intrinsic, genetically defined structural and functional characteristics of the human brain. The first example of a gene possibly contributing to such a genetic predisposition for language was provided by the discovery that disruptions of the FoxP2 gene cause developmental verbal dyspraxia (DVD). Individuals suffering from this speech and language disorder have severe difficulties with articulation and show impaired receptive and cognitive language skills.
Although recent theoretical work also puts forward the idea that the universality of certain syntactic rules might just be the by-product of the scale-free network architecture of languages (i Cancho et al., 2005; Nowak et al., 2001), the case of FoxP2 obviously allows to take a closer look on the development and function of neural circuits associated with language from a molecular and cellular perspective.
1.1 FoxP2 and Developmental Verbal Dyspraxia The causative link between FoxP2 and DVD was established when genomic alterations of FoxP2 were identified in all 16 affected members of the british KE-family and an unrelated
individual with a remarkably similar pathophenotype. Affected KE family members carry a substitution of arginine to histidine (R553H), which most likely renders the protein nonfunctional (Figure 1.1). This mutation is inherited in a dominant fashion and was found in KE DVD patients across three generations. In the unrelated individual FoxP2 is disrupted by a balanced translocation (Lai et al., 2001). The direct search for FoxP2 mutations in DVD patient panels meanwhile revealed more individuals with a disrupted FoxP2 allele (Feuk et al., 2006; MacDermot et al., 2005).
What is the common behavioral phenotype of individuals with DVD? Affected members of the KE family have severe difficulty in correctly articulating speech. In both word and the non-word repetition tests, where subjects have to repeat words (e.g. killer) and non-words (e.g. rillek) after hearing them, DVD patients score significantly worse than their unaffected family members (Watkins et al., 2002). The impairment increases gradually with the complexity of the words to be articulated. The DVD family members also have difficulties in the volitional control of skilled non-speech orofacial movements, a symptome called orofacial dyspraxia. Importantly, these difficulties cannot be attributed to a general impairment of motor control, since the patients´ limb praxis performance is indistinguishable from unaffected individuals (Watkins et al., 2002). The patients are also not impaired in their hearing ability. Interestingly, the DVD phenotype resembles that observed in patients with Broca´s aphasia (reviewed in (Damasio and Geschwind, 1984).
However, there are important behavioral differences between the two pathologies.
Aphasics perform better in the word than the non-word repetition test, whereas affected KE family perform equally bad in both tests. This could indicate that despite their actual problems with articulation, aphasics had learned to associate word articulation patterns with word meanings before the onset of the aphasia, which might help them finding the correct words. In contrast, affected KE family that never learned the correct word articulation patterns would fail in using word meaning to solve the word-repetition task.
In addition to the verbal and orofacial dyspraxia, KE family patients perform significantly worse than their unaffected relatives on tests that assess receptive and grammatical language. The deficit includes the inability to correctly inflect words (i.e. change tense or number) or to match sentences describing subtle relationships between objects with the corresponding pictures. Nevertheless, affected individuals score only slightly, but significantly lower on a non-verbal IQ-test than non-affected individuals and there is
considerable overlap between the groups (Alcock et al., 2000; Vargha-Khadem et al., 1998; Watkins et al., 2002). Taken together, these findings suggest that the primary deficit in the affected KE family members reflects a disruption of the sensorimotor mechanisms mediating the selection, control, and sequencing of learned fine movements of the mouth and face. An open question remains, if the receptive cognitive problems result from the primary articulation problem or if they constitute a second independent core deficit of the disorder. The first possibility would be consistent with the motor theory of speech perception (Liberman and Mattingly, 1985), which posits that decoding speech requires the brain circuitry involved in its production. Although recent human studies support this concept (Fadiga et al., 2002; Watkins et al., 2003), the possibility that aberrations of FoxP2 affect the development of grammatic skills independent of the articulation deficit cannot be ruled out.
First insights into the neural basis of the behavioral abnormalities shown by DVD patients came from the examination of affected and unaffected KE family members with structural and functional brain imaging techniques. Affected KE family members displayed bilateral structural deficits consisting of a reduction in the gray matter density of the caudate nucleus in the basal ganglia (Vargha-Khadem et al., 1998; Watkins et al., 2002) the ventral cerebellum (Belton et al., 2003) and Broca´s area. Abnormally high gray matter density was found in the putamen and Wernicke´s area. Interestingly, the volume of the caudate correlated well with the performance in the test of oral praxis (see above; (Watkins et al., 2002), indicating its involvement in the pathology. Given the well-established role of the basal ganglia in motor planning and sequencing (Graybiel, 1995), the structural abnormalities in the striatal regions of the basal ganglia (caudate and putamen) are generally consistent with an impaired control of orofacial motor function. However, it is less clear how they specifically compromise orofacial movements, without affecting other motor functions.
Functional imaging during the performance in covert (silent) and overt (spoken) tasks revealed lateralized disturbances in language-impaired subjects. In contrast to the typical left-dominant activation pattern involving Broca´s Area that is elicited by a verb generation test in unaffected KE family members, the signal distribution in affected individuals is more bilateral. Extensive bilateralization in the activation pattern was also observed for DVD subjects in the word repetition tasks described above. Consistent with the
morphological findings, an underactivation of Broca´s area and the putamen occurred in the affected family KE members (Liegeois et al., 2003). The observed overactivation of areas normally not involved in language has been interpreted to result from compensatory recruitment of additional brain areas, increased attention or a higher cognitive effort to solve the task. Taken together, the imaging work points to the frontostriatal and frontocerebellar networks as key circuitry affected in impaired KE family members.
1.2 FoxP2 Expression in the Brains of Mice and Men
Mapping the expression of FoxP2 in human and murine brains with in situ hybridization and immunohistochemistry has established where mammalian FoxP2 acts (Ferland et al., 2003; Lai et al., 2003; Takahashi et al., 2003). In adulthood, most prominent FoxP2 expression is found in the basal ganglia, in regions of the thalamus that receive input from the basal ganglia, in midbrain visual processing regions and in the inferior olive of the medulla. Further regions expressing high levels of FoxP2 include the cerebellar Purkinje cells, deep cerebellar nuclei, sensory auditory midbrain structures and layer VI neurons of the cerebral cortex (Ferland et al., 2003; Lai et al., 2003). Fetal FoxP2 expression is consistent with the adult expression pattern. In the rodent telencephalon, initial expression of FoxP2 is largely limited to the lateral ganglionic eminence [LGE (Ferland et al., 2003;
Takahashi et al., 2003)], the mammalian subpallial germinal zone that gives rise to the striatal projection neurons of the basal ganglia and to the majority of cortical interneurons (Brazel et al., 2003). Within the LGE, FoxP2 is expressed in the subventricular zone and mantle region but not in the proliferative ventricular zone, suggesting that expression is initiated in postmitotic neurons. This interpretation is also compatible with the additional expression site in the non-proliferative cortical plate of the developing cortex (Ferland et al., 2003; Takahashi et al., 2003). Taken together, the FoxP2 expression pattern is consistent with the sites of pathology identified in affected KE family members by brain imaging techniques. However, the question whether the reduction of functional FoxP2 protein affects the function of speech-related neural circuits as a consequence of their improper development, or by means of disturbed neural transmission or both remains unanswered, due to the purely descriptive nature of gene expression mapping.