Introduction

After many years of research, there is still no consensus on the function of the cerebellum. It is accepted today that the cerebellum has a much more broad role than simply that of motor coordination. Purkinje cells (PCs), which are the sole efferent output of the cerebellum, are one of few neuronal populations that express highly specific genes.

This review will focus mainly on the Cre/loxP recombination system, which is a unique tool for Purkinje-cell specific trangenesis that employs the Purkinje cell specific gene coding L7 or pcp2 (Purkinje cells specific protein-2). This protein was described for the first time in 1988 by two independent groups and this was the starting point of Purkinje cell specific transgenesis [2, 3]. In the following text, we will use official nomenclature of pcp2 according to the Mouse Genome Informatics database at www.jax.org.

The gene coding pcp2 is located on the mouse chromosome 8 [4, 5]. Initial studies of the gene structure described four exons [6, 7]. These exons cover approximately 6 % of the gene locus spanning 8 kb. Transcription results in mRNA encoding a protein of approximately 16 kDa [2]. Translation start is at the ATG codon in exon 2 and the reading frame includes the entire exon 3 and part of exon 4. Mature pcp2 proteins incorporate a G protein-regulatory motif (GPR), the GoLoco motif [8, 9]. Experiments on the biochemical properties of pcp2 revealed its modulatory role in GDP release from Gi and Go, and the physical interaction with G subunits [8, 10]. Later experiments confirmed the existence of three further exons present when the gene structure is alternatively spliced [9, 10]. Alternatively spliced pcp2 variants give rise to three proteins of different molecular weight with different biochemical properties and possible differences in function [8, 11, 12] (Fig. 1).

Fig. 1
figure 1

To date identified splice variants of the pcp2 gene

Full size image

Importantly, pcp2 is characterized by its very specific expression exclusively in PCs and retinal bipolar neurons [13, 14]. Expression of pcp2 starts within the first postnatal week and can be easily detected in Purkinje cells with Northern blotting [2, 3], in situ hybridization [2, 3, 10], Western blotting [13, 14], and immunohistochemistry [13, 14].

After discovery of the pcp2 gene a long time has passed until the final construct for Purkinje cell specific recombination has been designed. The overview of experiments and steps which resulted in this construct was presented in an earlier paper [1]. In the first instance, regulatory sequences of the pcp2 gene were used for expression of heterologous genes in Purkinje cell populations [1].

In the following text, we will focus on the use of the pcp2 sequence for targeting of PCs since the last summary published in 2002 [1].

Purkinje Cell Specific Gene Targeting

The possibility of cell or tissue specific deletion of a gene has two big advantages:

  • The ability to study consequences of knockout in a defined cell population. General knockouts almost always bring a question about the origin of an observed phenotype.

  • The ability to study effects of knockouts that are lethal if switched off in the entire organism.

The mostly used system for tissue or cell specific deletion of genes is the Cre/loxP system discovered in 1981 by Sternberg and Hamilton [15]. The system requires two components: a DNA sequence marked with the loxP sites—“floxed”, and expression of the Cre recombinase that recognizes the loxP sites and recombines (removes) the sequence between them. (Fig. 2). If the expression of recombinase is targeted onto a specific cell population, the deletion occurs only in these cells. To prove the specificity of the expression, a so called reporter mouse strain can be utilized. In these transgenic animals, a sequence coding for an easily detectable marker (chromogenic or fluorescent) is inserted into a locus characterized by ubiquitous expression. However, before the translation start, there is a STOP cassette floxed with loxP sites arresting the translation. If we then cross the reporter mouse with a mouse housing a cell-specific Cre expression, we are able to observe the chromogen or the fluorescence in the cells where recombination occurs (Fig. 3). Another recombination system accessible for gene targeting is based on the Flp recombinase [16]. This system can be combined with the Cre/loxP system to achieve recombination restricted to a subpopulation of cells. Both systems can be additionally modified by use of specifically designed inducible promoters. For more details see [1]

Fig. 2
figure 2

Diagram showing the principle of Cre/loxP genomic sequence excision

Full size image
Fig. 3
figure 3

Expression of β-galactosidase in Purkinje cells (black arrows) after crossing of B6.129-Tg(Pcp2-cre)2Mpin/J with the reporter strain—B6.129S4-Gt(ROSA)26Sortm1Sor/J

Full size image

A big step forward in targeting of PCs was the generation of the L7ΔAUG minigene by the John Oberdick group [17]. This construct was designed for easy insertion of DNA or cDNA sequences to express them specifically in Purkinje cells (Fig. 4). Because all transcription starts were removed from the pcp2 coding sequences, it became possible to express the protein of choice only with the use of a co-inserted ATG, using only regulatory capacities of the construct. Accessibility of this very convenient tool enabled us to generate a transgenic mouse strain with Purkinje cell specific expression of Cre recombinase in 2000 [18]. In 2004 and 2005, the generation of another two transgenic strains were reported, which expressed Cre recombinase using regulatory elements of the pcp2 gene. The first of them was generated by means of an artificial BAC chromosome [19], and the other one using the knock-in approach with the help of homologous recombination [20]. The primary strain published in 2000 [18] is accessible at The Jackson Laboratory under the strain name: B6.129-Tg(Pcp2-cre)2Mpin/J. The BAC-based strain can be provided also by JAX under the name B6.Cg-Tg(Pcp2-cre)3555Jdhu/J. The third one, the knock-in strain seems to be available only directly from the lab of origin. The first created strain [18] became the most popular and since its establishment was applied over 60 times for genetic manipulations in murine cerebellum.

Fig. 4
figure 4

Diagram of the L7ΔAUG minigene. Note all mutated ATG and BamHI site. For easy insertion of nucleotide sequences there is the BamHI site in exon 4

Full size image

The use of pcp2Cre based genetic manipulations started in 2003. For the first time the pcp2Cre strain [18] was used by the M. Meyer’s group for Purkinje cells specific knockout of calbindin D-28 k [21, 22]. The strain was created for this particular experiment to confirm the role of calbindin D-28 k in physiology of these neurons. Lack of calbindin D-28 k resulted in a very specific phenotype with clear signs of ataxia, however without signs of LTD disturbances. Detailed analysis of mutants revealed alteration in the early synaptic calcium transients (ESCT) without affecting the delayed synaptic calcium transients (DSCT). It is broadly accepted the DSCT are responsible for LTD, which seems to be independent from the calbindin D-28 k expression. Presence of LTD was additionally confirmed by means of behavioral tests verifying motor-learning.

In the following text, we will briefly review all applications of all three pcp2Cre strains (Table 1).

Table 1 Summary of studies utilizing the pcp2 for Cre based Purkinje cell specific transgenesis
Full size table

Experimental approaches and studies summarized in Table 1 show clearly, how useful is the possibility to specifically manipulate gene expression in the cerebellar PCs. Published data allow for the manipulation of PCs function in many aspects and extend our knowledge on cerebellar physiology. Most of the experiments are typical conditional knockouts where expression of a selected gene is switched off according to the spatial and temporal expression of the pcp2 gene. Some of the cited studies attempted to express conditionally selected genes using a similar approach as in the case of reporter mouse (Tab.1/5,6,15,18,22,30,43–45,56,60–63). In these experiments, 5′ to the coding sequence of nucleotides a floxed STOP cassette was placed. This kind of approach was also applied in the new field of conditional manipulation of neuron activity and namely in optogenetics (Tab.1/18, 44, 45, 56, 61, 63). This sophisticated method uses light to control activity of neurons or an individual neuron, which are sensitive light. The light sensitivity results from transgenic manipulations targeting expression of light-sensitive proteins to neurons. Optogenetics enables researchers to influence the activity of neurons in living animals, allowing for the precise recording of effects in real time.

Another group of experiments aimed to express heterological genes in PCs through the application of regulatory elements of pcp2 (Table 2).

Table 2 Summary of studies utilizing the pcp2 regulatory elements for Purkinje cell specific expression of heterologous genes
Full size table

Among these experiments, we can find four, which have important methodological input in the field of cerebellum research (Tab.2/1,2,4,6). Two of them resulted in stable expression of fluorescent proteins in PCs – GFP, Cerulean and Cherry (Tab.2/1,6). Another two experiments enabled researchers to specifically express a tetracyclin transactivator in these neurons (Tab.2/2,5).

In the following, we will discuss shortly results of selected papers, which are the most interesting from the authors point of view.

Shortly, after M. Meyer’s group report [18] another work concerning motor learning using conditional knockout was published by Feil et al. in 2003. Investigators created a mouse without cGMP-dependent kinase I in PCs to study its role in LTD and cerebellar learning. Researchers emphasized, that cell specific ablation not only gave opportunity to overcome unwanted effect of conventional gene targeting which affects whole organism but also was privileged in that study because pharmaceutical tools had limitations. Firstly, there were no acknowledged specific cGKI inhibitor and secondly there was concern about the exaggeration of the importance of some pathways that might be less important, or even not used under physiological conditions. Tests revealed strongly reduced LTD and impaired adaptation of the vestibulo-ocular reflex in mutants. Results also showed a specific, indispensable role of cGPKI signaling in cerebellar LTD and motor learning. CGKI seems also necessary for motor coordination. Study strongly supports the conception that cerebellar LTD is involved in a specific form of motor learning but not in motor performance in general [23].

A group of publications investigating developmental relationships and refinement of neuronal connections have been initiated by Lorenzetto et al. In 2009, researchers employed a conditional overexpression of muscle chloride channel-YFP fusion protein by means of Cre-loxP system exclusively in PCs and provide a new tool to verify the role electrical activity for elimination of supernumerary climbing fiber inputs. Their new approach had an advantage over previous methods—it allows to verify a single cell type specific perturbation instead of rather global effects of a particular neurotransmitter or signaling system malfunction. The study brought a strong evidence that synapse elimination in the CNS needs appropriate electrical activity in the postsynaptic cells [24].

One of the most sophisticated cre/loxP-based methods has been developed and described by Wall et al. In 2010, they published a work about the new tool for retrograde tracing of presynaptic cells. They used two viruses—one the avian sarcoma leucosis virus glycoprotein EnvA—pseudotyped rabbies virus that was capable to infect only cells expressing an avian receptor protein TVA. The second one was the cre-dependent helper virus that contained genes of the abovementioned TVA and the rabies glycoprotein B19G. This allowed the first one virus to spread retrogradely. Moreover, the EnvA-pseudotyped rabbies virus had no B19G gene so was unable to spread in the absence of another source of that glycoprotein. Combination of these viruses with the concomitant cell specific cre expression gave an ability to identify the direct local and long-distance connections to Cre-expressing neurons in vivo [25].

Another group of studies focused on PCs employed an optogenetic approach. First of the papers we would like to focus on is from 2011. Gutierrez et al. created an optogenetic mouse model for the cell specific expression of vertebrate rhodopsin (vRh) to investigate the functional impact of Gi/o protein mediated modulation on the cerebellar function via the spike modulation in PCs. First, they created mice with a cell-specific expression of vRh-GFP. Secondly, investigators showed that light activation of vRh selectively expressed in PCs, similarly to GABAbR activation by baclofen in PCs, reduces the firing frequency of PCs in vivo. At the end, they conducted series of motor tests confirming that Gi/o-mediated modulation of PC firing alters motor coordination in behaving mice [26].

Comprehensive study of Asrican et al. describes new mouse lines useful for investigation of neural circuits. Researchers compared three strategies for expressing channelrhodopsin-2 (ChR2) in PCs. Pcp-2 promoter was used to target the expression of ChR2 with a H134 mutation and the method was the most selective to Purkinje neurons. The other two lines utilized the parvalbumin promoter and targeted ChR2 expression additionally to other neurons of the cerebellar cortex. Histological analysis revealed significant expression of targeted protein in the cerebellum of all three lines. In contrast to pcp-2-based line where ChR2 expression was present only in Purkinje neurons, other two lines displayed expression of ChR2 also in molecular layer interneurons. The study demonstrated the convergence of a variable number of Purkinje neurons input onto the deep cerebellar nuclei neurons. Authors emphasized that pcp2-based line was ideally suited for PC circuits mapping by means of the optogenetic method [27].

Other application of Purkinje cell specific ChR2 expression was showed by Kruse et al. in 2014. To analyze neuronal networks Kruse with colleagues employed similar method to precisely monitor and modulate PCs firing but their method was improved by a more targeted application of light. Researchers found that PCs switched on and off in a milliseconds time resolution while being light activated. Same effect on PCs was brought by photo stimulation of molecular layer interneurons, whereas granule cells activation only slowly modulated its firing. These results indicate on the functional modulation of PCs firing rate by accompanying cells [28].

Interesting research opportunities brought also the work of Szabo et al. Their method allowed to control and monitor dynamics of microcircuits in freely behaving mice with almost cellular level resolution [29].

Last one of the most interesting optogenetic experiments in this review comes from Hoogland et al. from May 2015. In this study, mice with mutated P/Q-type calcium channels were subjected to the motor coordination test on a disk treadmill to study motor behavior and then to find correlations between the activity of PC ensembles and motor performance parameters. Optogenetics was applied to evoke simple spike activity simultaneously in multiple Purkinje neurons expressing ChR2(H134R)-eYFP driven by pcp2-Cre. Combination of these two methods enabled to characterize walking patterns, and preferred sequences of limb placement after light-induced simple spike activity of PCs groups at rest. During locomotion various gait-inhibition patterns caused by PCs activation could be revealed. Study showed that optogenetically evoked simple spike activity in PC groups can influence the sequence and timing of limb and body movements at rest and during locomotion; output of cerebellar zones in contribution with other downstream regions is involved into coordination of multi-joint and multi-limb-mediated movements [30].

Since the publication of the first two papers describing pcp2, there has been a substantial increase in the understanding of cerebellar physiology [2, 3]. It became easier to study cerebellar function by selectively manipulating its sole output—Purkinje cells. The big step forward however was the possibility to switch off genes with putative importance PC physiology and possibility to manipulate activity of their function without affecting neighboring neuronal elements. Purkinje cell specific gene manipulation makes it possible to analyze even single Purkinje cells with the use one of the very modern techniques—optogenetics.