Final Report Summary - SYNAPSE NL (When Neurons Touch-Elucidating the Role of Neuroligins in the Formation, Development, Maturation, and Maintenance of Synapses)
Central to the complex functioning of the brain is the ability of neurons to communicate. This occurs at highly specialised contact points known as synapses, the regulation of which is critical for higher functioning tasks such as learning and memory. Neuroligins (NL) are neuron-specific adhesion proteins that facilitate neuronal communication. NLs or related protein families stabilise the synapse by acting as scaffolds or by bridging the synaptic cleft to contact associated molecules on the apposing neuron and to support receptor recruitment, allowing for efficient neurotransmission. It is widely believed that defects in the protein machinery that facilitates synaptic formation, maintenance and function could underlie neuropsychiatric disorders, including autism spectrum disorders (ASDs) and schizophrenia. Our overall goal is to understand exactly how these molecules function in the brain and to elucidate the consequences of their dysfunction for human disorders.
In order to address these issues with respect to NLs, the Synapse-NL project was designed (1) to systematically search for new NL interaction partners, (2) to analyse compound deletion mutant mice lacking either all excitatory synaptic NLs or all inhibitory synaptic NLs, and (3) to generate a conditional knock-in mouse for NL4 and to study the autistic phenotype and other behavioural features in the mature animal.
Objective 1: Neuroligin interaction partners
We used a crude synaptosome preparation of WT and NL4 deletion mutant (KO) mice to screen for NL4 interactors. Solubilisation and subsequent immunoprecipitation of NL4 from these fractions allowed us to visualise immunoprecipitated proteins on Coomassie stained SDS-PAGE gels. By comparing NL4 WT and KO fractions, we could identify non-specific binders and eliminate these from our experiments. Proteins present in WT but not NL4 KO (-ve control) were excised and analysed by mass spectrometry (MALDI-TOF, nanoESI, Orbitrap) (Fig. 1A). The initial results implicated several synaptic proteins as NL4 binders. However, the confidence in these results (number of identified peptides or mass prediction) were low and near the cut-off point, and subsequent validation of these proteins by Western blotting identified them as false positives (Fig 2A).
In order to decrease non-specific interactors and avoid the presence of heavy and light IgG chains in the immunopurified sample, we successfully incorporated a peptide competition step in our immunopurification protocol to elute NL4 and its binders from the affinity beads by incubating the beads with excess peptide corresponding to the NL4 antibody epitope, thus replacing the protein on the antibody and specifically releasing NL4 and its binders into the supernatant (Fig. 1B). In addition, we explored different detergents (Chaps, Triton X-100 and sodium cholate) for protein solubilisation and varied the salt concentrations in order to tease out specific interactions under more stringent conditions. These extended efforts, however, did not yield new NL4-binding proteins. We concluded that the conditions required to extract NL4 from membranes may disrupt weak interactions so that an alternative approach may be required.
Interestingly, we found that we could reproducibly detect all other NL isoforms (NL1, NL2, NL3) in NL4 immunoprecipitates (Fig. 2B). This led us to speculate that functional NLs may be present as oligomers at synapses. Previous results had also alluded to this (Dean et al, 2003) and we had evidence that other NL isoforms are detectable as dimers following cell surface crosslinking in cultured neurons (Fig. 3A). Using a similar paradigm, we assessed NL4 oligomerisation (Fig 3B). We observed NL4-NL4 dimers. Surprisingly however, we did not find evidence of NL4 heterodimerisaton with NL1, NL2 or NL3 (Poulopoulos et al 2012). Promiscuous NL oligomerisation has been observed in our laboratory following NL overexpression in cell lines. Therefore, it is possible that NLs formed artificial oligomers during incubation periods of our immunoprecipitation experiments that do not reflect the heteroligomerisation status at mature synapses. On the other hand, the crosslinking approach relied on crosslinkers of a specific length (BS3, 11.4 Å) between amines. It can therefore not be ruled out that the lack of observed NL4 heterodimers was due to steric restraints that did not allow for crosslinking between NL4 and the other NL isoforms.
In a new approach, we designed a construct encompassing the extracellular domain of NL4 coupled to the Fc domain of human IgG that is cleavable by thrombin and tagged with hemagglutinin (HA). The advantage of this construct is that the protein can be mass produced by expression in cell lines and coupled to protein G sepharose beads for affinity purification purposes without the use of antibodies. The use of the NL4-Fc protein also eliminates the need to extract native NL4 from membranes while retaining interactors, which may bind transiently or weakly. Experiments are currently in progress to produce sufficient NL4-Fc protein to be used as affinity matrix. Additional options will be explored including crosslinking approaches on cultured neurons.
Objective 2: Analysis of compound neuroligin mutants
NL1 is primarily present at excitatory synapses and NL2 at inhibitory synapses. Depending on the brain region and cell type, NL3 is present at both types of synapses. NL4 shares many characteristics with NL2 and is localized to inhibitory synapses in many brain regions. Compound mutants lacking NL1, NL2 and NL3 are embryonic lethal due to respiratory difficulties. NL1/3/4 and NL2/3/4 compound mutants are viable and represent valuable tools to analyse NL function at excitatory and inhibitory synapses, respectively. Having established a breeding regime for sufficient mutants for experiments, we found that triple NL compound KOs were essentially impossible to breed in sufficient numbers in order to perform the planned experiments in time. We therefore focused subsequent activities on the generation of NL2/4 double KO animals as evidence indicates that NL4, similar to NL2, may play a role at inhibitory synapses. Analyses of inhibitory synapse formation and function in NL2/4 double KOs is currently under way. We will investigate mechanisms of compensation for NL loss on established phenotypes and characterise new synaptic phenotypes following loss of inhibitory-type NLs.
Further, we attempted to generate a range of new anti-NL antibodies for immunohistochemistry with local antibody companies Synaptic Systems and Biogenes. However, efforts to date failed to generate new antibodies that satisfactorily detect NLs for the required applications despite numerous attempts. Furthermore, we determined that our current anti-NL4 antibodies, which can specifically detect NL4 in WT but not in NL4 KO animals, are likely only specific for a subset of the total NL4 in the brain (Fig. 4). Quantification of NL4 from brain region specific lysates by Western blot indicates that NL4 is highly expressed in cortex and hippocampus with lower expression in brain stem. In contrast, specific immunohistochemical staining detects NL4 expression most prominently in the brain stem, with much weaker staining observed in layer 4 of the cortex and the stratum lacunosum moleculare of the hippocampus (Fig. 4A). These results indicated that in a brain region specific manner NL4 might possess different quaternary structure conformations, be differentially post-translationally modified or engage with occluding binding partners or oligomers that result in epitope masking. In order to unmask such an epitope, we tested several fixation methods (PFA, methanol and antigen retrieval protocols such as heating/boiling) as well as chemical/enzymatic pre-treatments (pepsin, N- and O-deglycoslation, dephosphorylation) but failed to improve anti-NL4 staining in the forebrain. Currently, aided by the EU Funded EUROSPIN consortium, we are continuing to target new epitopes to yield additional antibodies.
Objective 3: Generation of “stop and rescue” NL4 mouse
The formation and maturation of synapses and the neuronal networks they create occurs primarily up until early adulthood. Thus the dogma was that disease-related defects in synapse maturation and function due to mutations or the absence of key proteins during development would lead to permanent impairment throughout life. A study by Guy and colleagues illustrated that this might not always be the case as a mouse lacking Mecp2 (a Rett-syndrome associated gene) until adolescence developed severe disabilities reminiscent of the Rett syndrome, but made an almost complete recovery upon re-expression of the gene in early adulthood (Guy et al, 2007). We aimed to emulate this study with the ASD-related behavioural phenotype of NL4 KO mice (Jamain et al, 2008). The genomic sequence of NL4 was not available in any public databases. Having obtained a “stop cassette" from our collaborator Dr. K. Tanaka (Okazaki, Japan) and having acquired a cosmid containing partial NL4 genomic sequence, we strived to sequence NL4 genomic DNA in preparation for the generation of targeting vectors. This, however, proved extremely difficult as genomic NL4 DNA was found to be extremely GC rich and to contain numerous long and short tandem repeats, and single nucleotide repeats, which made sequencing and assembly of DNA sequences extremely difficult. This is the likely reason that genomic NL4 sequences have not been available in public databases, and prevents targeting of the NL4 gene with standard homologous recombination techniques in ES cells. Therefore, and in order to save time, we employed Sigma Advanced Genetic Engineering to exploit the use of Zinc finger (ZFN) technology that requires less than a 1 kb of DNA target region. The first two sets of ZFN modules failed to demonstrate cleavage of the target region, and efforts are continuing to obtain an efficient ZFN pair. In addition we are exploring the use of TALENS, a similar technology that may be more specific in its DNA targeting. Continued efforts for genomic NL4 sequencing have resulted in the near completion of the NL4 genomic DNA sequence which no other lab has achieved to date. However, it is clear that working with NL4 genomic DNA in the future may continue to prove challenging with molecular biology techniques.
In summary, the completion of the project aims has proved extremely challenging especially with regards to NL4 as we encountered difficulties on both the protein and DNA levels. In spite of this, we have made steady progress on all fronts and adapted experimental plans to progress within the range of the tools available to us. Furthermore, additional work outside of the scope of these aims is yielding clues to NL function at the synapse. The strong collaborations built over the last two years will further accelerate the progression of these aims. Deciphering the function of NLs and their partners is of paramount interest to the field. Medicine and the pharmaceutical industry will benefit from these data as elucidating these mechanisms opens the door for therapeutic intervention, which would aid those in society suffering from specific neurodevelopmental or neuropsychiatric disorders that are related to NLs.
Bibliography
Dean C, Scholl FG, Choih J, DeMaria S, Berger J, Isacoff E, Scheiffele P (2003) Neurexin mediates the assembly of presynaptic terminals. Nat Neurosci 6, 708-716.
Guy J, Gan J, Selfridge J, Cobb S, Bird A (2007) Reversal of neurological defects in a mouse model of Rett syndrome. Science 315, 1143-1147.
Jamain S, Radyushkin K, Hammerschmidt K, Granon S, Boretius S, Varoqueaux F, Ramanantsoa N, Gallego J, Ronnenberg A, Winter D, Frahm J, Fischer J, Bourgeron T, Ehrenreich H, Brose N Reduced social interaction and ultrasonic communication in a mouse model of monogenic heritable autism. Proc Natl Acad Sci USA 105, 1710-1715.
Poulopoulos A, Soykan T, Tuffy LP, Hammer M, Varoqueaux F, Brose N (2012) Homodimerization and isoform-specific heterodimerization of neuroligins. Biochem J 446, 321-330.