Expression Analysis of Lrrk1, Lrrk2 and Lrrk2 Splice Variants in Mice

Missense mutations in the leucine-rich repeat kinase 2 gene (LRRK2) are linked to autosomal dominant forms of Parkinson’s disease (PD). In order to get insights into the physiological role of Lrrk2, we examined the distribution of Lrrk2 mRNA and different splice variants in the developing murine embryo and the adult brain of Mus musculus. To analyse if the Lrrk2-paralog, Lrrk1, may have redundant functions in PD-development, we also compared Lrrk1 and Lrrk2 expression in the same tissues. Using radioactive in situ hybridization, we found ubiquitous expression of both genes at low level from embryonic stage E9.5 onward, which progressively increased up until birth. The developing central nervous system (CNS) displayed no prominent Lrrk2 mRNA signals at these time-points. However, in the entire postnatal brain Lrrk2 became detectable, showing strongest level in the striatum and the cortex of adult mice; Lrrk1 was only detectable in the mitral cell layer of the olfactory bulb. Thus, due to the non-overlapping expression patterns, a redundant function of Lrrk2 and Lrrk1 in the pathogenesis of PD seems to be unlikely. Quantification of Lrrk2 mRNA and protein level in several brain regions by real-time PCR and Western blot verified the striatum and cortex as hotspots of postnatal Lrrk2 expression. Strong expression of Lrrk2 is mainly found in neurons, specifically in the dopamine receptor 1 (DRD1a) and 2 (DRD2)-positive subpopulations of the striatal medium spiny neurons. Finally, we identified 2 new splice-variants of Lrrk2 in RNA-samples from various adult brain regions and organs: a variant with a skipped exon 5 and a truncated variant terminating in an alternative exon 42a. In order to identify the origin of these two splice variants, we also analysed primary neural cultures independently and found cell-specific expression patterns for these variants in microglia and astrocytes.


Introduction
Parkinson's Disease (PD), as the second most common neurodegenerative disorder, affects 1-2% of the population over the age of 65 (rising to 4-5% of the population from the age of 85) [1][2][3]. Even though the majority of PD cases are sporadic (.90%), several familial forms of the disease -which could be linked to a single genetic factor -were identified over the last fifteen years [4,5]. Missense mutations in the LRRK2 gene (dardarin), which are linked to changes in the coding region of the PARK8 locus [6][7][8], represent the most common cause of familial (up to 10% overall; up to 40% in certain ethnic populations) and sporadic (1-2%) lateonset forms of PD [7][8][9][10][11][12][13][14]. In contrast to hereditary loss-of-function PD-forms, LRRK2 missense mutations do not interfere with the expression of this large multi-domain protein. Nevertheless, the activity of distinct protein domains is altered. The neuropathology of LRRK2-associated PD is heterogeneous and features the classical nigral degeneration with or without the presence of Lewy bodies and ubiquitin-positive or tau-positive inclusions [8,[15][16][17][18].
The Lrrk2 gene is comprised of 51 exons coding for a large 2527 amino acid multi-domain protein [7]. It shares high homology with its slightly shorter paralog Lrrk1another member of the ROCO protein kinase superfamily [19,20]. Both proteins contain two domains, Roc (Ras in complex proteins, belonging to the Ras/ GTPase family) and COR (C-terminal of Roc) characteristic of ROCO proteins [19], as well as the catalytic kinase domain with close sequence homology to MAPKKKs of the MLK (mixed lineage kinases)-type. In addition, several other domains are found in LRRK1 and/or LRRK2: LRRK2-specific repeats (only in LRRK2), ankyrin repeats (in both) and leucine-rich repeats (LRR, in both) plus a WD40 domain towards the C-terminus (only in LRRK2) [8,20,21]. Although mutations are spread over the whole LRRK2 gene, most of the pathogenic mutations are found either in the kinase (G2019S, I2020T), COR (Y1699C) or Roc-GTPase domain (R1441C/G/H). In vitro data suggests that GTP binding promotes LRRK2 kinase activity and enhances neurotoxicity, indicating a gain-of-function mechanism in the etiology of PD [22][23][24][25][26]. A similar intramolecular, regulatory mechanism has also been reported for LRRK1 [27], but interestingly, no mutation clearly segregating with PD could be identified in this gene so far [28,29]. Nevertheless, due to the highly conserved protein structures of both proteins it has been proposed that there might be a redundancy in their, up to now still unknown, physiological functions [30,31]. Thus, detailed analysis of the organ and brain distribution of Lrrk1 and Lrrk2 could provide valuable hints towards their function. Therefore, in this study a series of detailed and comparative expression analyses for mammalian Lrrk1 and Lrrk2 was performed in mouse. Different analytical methods using samples from early stages of murine embryonic development until adulthood were used for this including: (i) Localization and expression level of Lrrk1 and Lrrk2 were analysed by radioactive in situ hybridization (ISH) in sections of whole embryos, postnatal, adult and aged brains. (ii) Expression level of Lrrk2 mRNA were quantified by real-time PCR and protein level by Western blot in several brain regions. (iii) Lrrk2 co-expression in different neuronal populations of the striatum and cortex were analysed by immunohistological stainings. (iv) In addition, two splice-variants of Lrrk2 were identified in RNA samples from different adult brain regions and organs. Taken together, this is the first time that Lrrk1 and Lrrk2 expression (localization, expression level, cellular distribution and splicing) were analysed using a comparative approach with a number of independent highly sensitive analytical methods (qualitative and quantitative) in parallel. This combination of high-resolution ISH images; region-specific quantification of mRNA and protein level by real-time PCR and Western blot respectively; immunohistochemical analysis of the distribution in neuronal populations and the identification of cell-type specific splice variants is the basis for new insights into potential Lrrk2 function.

Animals Housing and Tissue Preparation
All animal work was carried out in accordance with the European Communities' Council Directive 2010/63/EU. All efforts were made to minimize animal suffering during the work. All protocols involving animal handling were approved by the committee for the Care and Use of Laboratory animals of the Government of Upper Bavaria, Germany. C57BL/6J mice (obtained from Charles River Germany) were group housed in individually ventilated type IIL cages (four mice per cage) and maintained on a 12 h/12 h light/dark cycle with food and water available ad libitum.
Mice were sacrificed and immediately thereafter intracardially perfused with 200 ml of ice-cold paraformaldehyde (PFA, 4%, pH 7.5). For post-fixation, the brains were incubated in PFA for 1-2 hours at RT. For embryo preparation, mothers were sacrificed via cervical dislocation. Embryos were dissected and immediately decapitated. Brains were dissected and immersion fixed in PFA overnight at 4uC. For embedding, brains were dehydrated through an ascending ethanol series (30%, 50%, 75%, 85%, 95% and 26100%; 90 min per step), clarified for 2660 min in 100% RotiHistolH (Carl Roth GmbH, #6640) at RT, equilibrated for 60 min in a 1:1 mixture of RotiHistolH/paraffin at 65uC, transferred for 60 min into 100% paraffin at 65uC and incubated for 8 h in 100% paraffin at 65uC. At the end, samples were cooled down to room temperature in 100% paraffin (embedding) and stored at 4uC until cutting. Paraffin embedded brain tissue were cut on a Cryostat (MICROM, Heidelberg, Germany) in three different planes (coronal, sagittal and horizontal), put into a water bath (37-42uC) for flattening and mounted on SuperFrostH Plus slides (Menzel-Glä ser, #J1800AMNZ), dried overnight on a heating plate and/or in an incubator at 37uC and stored at 4uC until use.

In situ Hybridization
ISH has been performed on paraffin sections using either [ 35 S]-UTP-radiolabeled (Perkin Elmer) or digoxigenin (DIG) labeled riboprobes according to manufacturer's protocol (Roche). The applied protocol is based on the published protocol of Dagerlind and colleagues [37]. In brief, 8 mm thick paraffin sections were rehydrated, partially digested with proteinase K, acetylated, dehydrated again and prehybridized in 100 ml hybridisation mix (containing 50% deionized formamide, 46 SSC, 0.56 Denhardt's solution, 1% N-lauroylsarcosine, 20 mM Na-phosphate buffer and 10% Dextran sulphate) per slide for 1 h at 58uC. Following that, the slides were incubated with either [ 35 S]-UTP-radiolabeled (1?10 6 counts per million/slide) or DIG-labeled riboprobes (15-30 ng/slides) in 100 ml hybridisation mix per slide overnight at 58uC. Washing was performed in decreasing concentrations (46 to 0.16) of saline sodium citrate (SSC) at RT, except the final washing step in 0.16 SSC for 2630 minutes at 65uC). For the radioactive ISH, the hybridized slides were dehydrated and dipped in autoradiographic emulsion (Kodak NTB2), developed after 4-6 weeks and counterstained with cresyl violet according to standard protocols. For detecting DIG-labeled riboprobes, an anti-DIG antibody coupled to the alkaline phosphatase (AP) enzyme has been used; subsequent either nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Boehringer Mannheim, Germany) or a tyramide-cyanine 3 TSA kit (Perkin Elmer Life Sciences, Boston, USA) have been used as chromophore. Riboprobes were obtained from cDNA fragments cloned into the pCR TM II-TOPOH vector (Life Technologies, Carlsbad, USA) by in vitro transcription by RNA polymerase T3, T7 or SP6, either utilizing [ 35 S]-UTP (Perkin Elmer Life Sciences, Boston, USA) or DIG-labeling mix according to manufacturer's protocol (Roche, Mannheim, Germany). To obtain specific ISH-probes for Lrrk1 and Lrrk2, several overlapping and non-overlapping cDNA fragments have been amplified from murine total mRNA by RT-PCRs. Sequences of used primers can be found in ( Figure S9).

Immunhistochemistry on Paraffin Sections
IHC on paraffin sections was performed in coplin jars except for blocking reactions, antibody incubations, incubation with ABCreagent and DAB-stainings. Glass slides were dewaxed for 45 min in xylol, rehydrated in a series of descending alcohol (100%, 96% and 70%) and rinsed with water. To destroy endogenous peroxidases, slides were incubated for 5 min in 0.1% H 2 O 2 (in PBS) before blocking for 1 h in 10% FCS (with 0.05% Triton-X in PBS). Primary antibodies against DARPP-32 (Abcam, ab40801), dopamine D1 receptor (Calbiochem, #324390), dopamine D2 receptor (Chemicon, AB5084P), NeuN (Chemicon, MAB377) were diluted in 10% FCS (with 0.05% Triton-X in PBS) and incubated overnight at 4uC in a humid chamber. Appropriate secondary antibodies were then applied on the next day diluted in 10% FCS (with 0.05% Triton-X in PBS) for 1 h at RT in a humid chamber. Next, the DAB working solution (Sigma) was freshly prepared (1 ml DAB stock solution, 15 ml 30% H 2 O 2 and 19 ml 0.1 M Tris-HCl, pH 7.4) and applied to the sections for 30 min in a humid chamber. Finally, sections were dehydrated in a series of descending alcohol (70%, 96% and 100%) and xylol and embedded in Roti-HistoKit II (Carl Roth GmbH).
Quantitative Lrrk2 expression analysis was performed using total RNA from different brain regions and the ABI PRISM 7900 Sequence Detection System (Applied Biosystems) and TaqManH reaction mixes for endogenous Lrrk2 (Mm00481934_m1, Applied Biosystems) and b-actin (NM_007393.1, Applied Biosystems). All samples were measured in triplicates using the following PCR program: an initial step of 95uC for 10 min and then 40 cycles of 95uC for 15 s and 60uC for 1 min. For quantification of Lrrk2 transcripts, three different qPCR reactions were performed from each cDNA sample (endogenous LRRK2, alternative LRRK2 and b-actin control). A 20 ml reaction contained 10 ml of TaqManH Universal PCR Master Mix (Applied Biosystems), 2 ml cDNA template, 300 nmol/L of each primer and 200 nmol/L of the specific probe. Sequences of used primers and probes can be found in ( Figure S9). For the b-actin control reaction an ACTB TaqManH Gene Expression Assay Mix Mouse (Applied Biosystems) was used. No template controls (NTC) were included in each assay and amplification and detection were performed under standard conditions as recommend by Applied Biosystems. The fluorescent signal intensities were recorded and analysed during PCR amplification using the Sequence Detection Software (SDS) software (Applied Biosystems). To determine the LRRK2 ratios, all values were initially normalized to the b-actin expression. Following, the ratio of variant versus endogenous Lrrk2 was determined (2 2[DDCTvariant2DDCTnormal] ).

Preparation of Primary Neurons, Microglia and Astrocytes
Primary cortical neuronal cultures of C57BL/6J brains were generated as previously described [38]. For preparation of primary astrocyte and microglia cultures, newborn pups (1-3 days old) were decapitated, brains were dissected free of the meninges, and the cerebral cortices were isolated. Then, cortices were mechanically dissociated and cells were resuspended in astrocyte [DMEM (Invitrogen) supplemented with 10% FCS, 1% Pen/Strep, 10 ng/ml human EGF (Invitrogen) and murine bFGF (Peprotech)] or microglia medium [DMEM/F12 (Invitrogen) supplemented with 10% FCS and 1% Pen/Strep], respectively. Culture was performed in 6-well dishes at 5% CO 2 and medium was changed every 3-4 days. To deplete as many microglia as possible, plates were shaken rapidly (15-30 s) prior to each medium change (step done only for astrocyte cultures). After confluency was reached, cultures were trypsinated using a 1:3 dilution of 0.25% trypsin and serum-free medium until the astrocyte containing cell layer slowly detached (while the microglia stayed attached to the plate) [39]. To harvest astrocytes, the floating cell-layer were transferred to a new plate and dissociated by gentle pippeting for further subculture. To harvest microglia, just new medium was added to the old plates after astrocyte-removal and cells were used after 3-6 days. The astrocyte-subculture procedure was repeated 3 times until cultures were taken for experiments and mitogens were removed from the medium 4 days before RNA was isolated.

Results
Comparative Expression Analysis of Lrrk1 and Lrrk2 during Murine Embryonic Development The pattern of expression reflects the physiological role of a protein. Thus, a comparative expression analysis of Lrrk2 and Lrrk1 was performed in C57BL/6J mice by radioactive ISH. This was also to evaluate the possibility of redundant functions between these two highly conserved paralogs. The main emphasis of this expression analysis was put on the brain of adult mice; embryonic development was also taken into account.
Lrrk2 mRNA was first analysed using semi-quantitative RT-PCR in undifferentiated embryonic stem (ES) cells and in early stages of embryonic development (E8.5 to E12.5 embryos) ( Figure  S1A). In virtually all developmental stages and organs (from ES cells until adulthood), expression of Lrrk2 mRNA could be detected ( Figure S1A and data not shown). The earliest embryonic stage, which was analysed by ISH, was E7.5 ( Figure S1B). At this developmental stage only the surrounding maternal tissue (i.e. the parietal yolk sac) exhibited strong Lrrk2 expression, whereas signals in the ectodermal, mesodermal and endodermal embryonic tissues were too weak to be detected even by radioactive ISH ( Figure  S1B). At embryonic stages E9.0 to E11.5, weak expression of Lrrk1 and Lrrk2 was observed widespread over the entire embryo beside the neuroepithelium of the developing CNS. At stage E10, Lrrk1 showed distinct areas of stronger expression in the cephalic mesenchyme tissue and palate ( Figure 1A,B), the inner layer of the optic cup (the future nervous layer of the retina), in the tissue of the first brachial arch and in the developing gut ( Figure 1B), ). Lrrk2 was expressed at higher level besides the head mesenchyme mainly in the urogenital ridge ( Figure 1C,D). At stage E12.5, further Lrrk1 expressing domains became visible in the epithelia of the nose and mouth ( Figure 1E) and significant Lrrk2 expression was observed in Rathke's pouch as part of the developing pituitary gland ( Figure 1F). Again, Lrrk1 and Lrrk2 could not be detected in the developing central nervous system, however, one hot spot of expression was the choroid plexus. Furthermore it is noteworthy that both genes were strongly expressed in the meninges -a nonneuronal tissue directly surrounding the developing central nervous system (CNS) ( Figure 1E,F). At stage E14.5 ( Figure  S1C-E for Lrrk2) and E15.5 ( Figure 1G,H), robust Lrrk1 and Lrrk2 expression was detected in parts of the developing kidney, lung, liver and heart ( Figure 1G,H). Interestingly, the expression of Lrrk1 and Lrrk2 at this developmental stage showed a high degree of overlap, however, the localization within the tissues was diverse in some cases. For example, Lrrk2 mRNA in the developing kidney was more or less restricted to structures like the metanephric vesicles ( Figure S1E), which was not the case for Lrrk1 expression (data not shown). In agreement, both Lrrk1 and Lrrk2 were still strongly expressed in the choroid plexus ( Figure 1G,H and Figure  S1C), while no expression was observed in other parts of the embryonic brain until birth (data not shown). Taken together, the expression pattern of Lrrk1 and Lrrk2 were largely overlapping during embryogenesis in many organs and structures. However, Lrrk1 expression seemed to be stronger and more widespread than expression of Lrrk2.

Comparative Expression Analysis of Lrrk1 and Lrrk2 in the Murine Postnatal, Adult and Aged Brain
Distinct expression of Lrrk2 in neural cells of the CNS were observed around birth (Postnatal day 0, P0), where weak Lrrk2 expression was found in the entire brain, specifically in the developing cortex, cerebellum and brain stem ( Figure S2A). During postnatal development until P7, the overall neuronal ISH signals but also the expression in distinct domains like cortex, striatum and olfactory bulb increased notably ( Figure S2B). Between P7 and P21, the level of Lrrk2 expression in the forebrain and midbrain varied in a highly dynamic manner. As an example, Lrrk2 ISH signals in the striatum increased in the forebrain and surpassing the cortical expression level, which remained rather unchanged ( Figure 2A,C,E). Also, the level of Lrrk2 increased considerably in hippocampus and posterior cortex, while expression in the substantia nigra (SN) was constantly low ( Figure 2B,D,F). In contrast to Lrrk2 ( Figure 3D), Lrrk1 mRNA was barely detectable in the postnatal brain, namely at P21. At this time-point of development, only the non-neuronal meninges and the mitral cell layer of the olfactory bulb showed faint ISH signal for the Lrrk1 specific probe ( Figure 3A and Figure S4I).
In the adult (older than 6 weeks) mouse brain, strong Lrrk2 ISH signals were present all over the brain, with highest level in the forebrain, the olfactory tubercle and the striatum ( Figure 3E and Figure S4B). In addition, strong Lrrk2 expression was also observed in the motor, somatosensory and visual cortex ( Figure S3), as well as in the anterior olfactory nucleus and the hippocampus. Herein, specifically the pyramidal cells of the hippocampus proper (CA1, CA2 and CA3) as well as the dentate gyrus were robustly labeled ( Figure 3E and Figure S4F). Lrrk2 expression was also found in the subventricular zone ( Figure S4D) -the second region of adult neurogenesis besides hippocampus. Interestingly, the dopaminergic system of the ventral midbrain (built up by the SN pars compacta and pars reticulata, as well as the ventral tegmental area) showed only comparable low level of Lrrk2 expression (white arrows in Figure 2D,F; Figure 3E and Figure S4H). In the hindbrain, strongest ISH signals were found in the purkinje cell layer and molecular layer of the cerebellum ( Figure 3E and Figure  S4J). In the adult pituitary gland, Lrrk2 can be observed in the  anterior and especially in the intermediate lobe (data not shown) which constitute the adenohypophysis, but not in the posterior lobe (neurohypophysis). During ageing, no significant differencesneither in expression pattern nor in the expression level -were observed in up to 24 month old mice (data not shown). Lrrk1 retained its postnatal expression pattern also in adulthood ( Figure 3 and Figure S3-4).
Taken together, the expression of Lrrk2 in the murine CNS became detectable around birth and increased during early postnatal development. In adult mice, the level of Lrrk2 mRNA in the forebrain and hindbrain were higher than in the midbrain.
The general Lrrk2 expression patterns observed in six week old brains sustained stable during adulthood. In contrast to Lrrk2, Lrrk1 was barely detectable in the adult mouse brain and only present in the mitral layer cell layer in the olfactory bulb where no prominent Lrrk2 expression can be found.

Quantitative Analysis of Lrrk2 mRNA and Protein Level in the Adult CNS
Since Lrrk1 was barely detectable in the adult mouse brain, all further analyses were focused on Lrrk2. To obtain quantitative data Figure 3. Comparative expression analysis of Lrrk1 and Lrrk2 mRNA in the brain of adult mice. ISH for Lrrk1 (top part) and Lrrk2 (bottom part) mRNA in sections from P21 (A-D) and adult mice (E). Note that Lrrk1 mRNA is barely detectable in the adult mouse brain and only visible in the non-neuronal meninges (white arrow) and the olfactory bulb (white arrowhead) (A). A detailed view onto the adult olfactory bulb depicts the solely neuronal expression of Lrrk1 in the mitral cell layer (B). Specificity of the Lrrk1 signals were verified by using the corresponding sence-probe as negative control (C). In contrast, strong Lrrk2 expression can be detected in various regions throughout the postnatal (D) and adult CNS (E). Abbreviations: An, anterior olfactory nucleus; Bs, brain stem; Cb, cerebellum; Cc, corpus callosum; Cm, motor cortex; Co, cortex; Cp, choroid plexus; Cs, somatosensory cortex; Cv, visual cortex; Gl, glomerular layer; Gr, granual layer; Hi, hippocampus; Me, meninges; Mi, mitral layer; Ob, olfactory bulb; Ot, olfactory tubercle; Pn, parafascicular nucleus; Pl, plexiform layer; Po, pons; SNc, substantia nigra pars compacta; St, striatum; Th, thalamus; Lv, lateral ventricle. Scale bars represent 1 mm (A, D, E) and 500 mm (B). doi:10.1371/journal.pone.0063778.g003 with regard to LRRK2 expression level in the adult brain, we performed quantitative PCR (qPCR) and Western blot analyses on total RNA and protein samples from different brain regions ( Figure 4A). Therefore, mouse brains were separated into their two hemispheres: one half was used for Western blot and the other for qPCR analysis. These were separated into nine different brain regions and utilized to prepare total RNA or protein samples. To allow a direct comparison between Lrrk2 mRNA and protein level, the relative expression level for all brain regions have been summed up and set to 100%. Interestingly, the pattern of protein expression found by Western blot matched the mRNA level observed by ISH (compare Figure 3 and Figure 4B). The highest LRRK2 protein level was found in samples from the posterior cortex (27.2% (63.7%) of total LRRK2 protein in the adult brain), followed by striatum, hippocampus and cerebellum ( Figure 4B). Also by qPCR, the highest Lrrk2 mRNA level were found in samples from cortex and striatum, followed by cerebellum, brain stem and hippocampus ( Figure 4B). Apart from striatum, there were no remarkable discrepancies between the relative protein and mRNA level in the different brain regions. However, the striatal mRNA level was 32.9% (66.2%) of the total Lrrk2 mRNA, while the very same region contained only 17.6% (64.0%) of the total LRRK2 protein. A possible explanation for this phenomenon could be posttranscriptional regulation of LRRK2 in the striatum via specific microRNAs and/or transport of the protein. However, further analyses would be necessary to proof this hypothesis. Taken together, qPCR and quantitative Western blot analyses further verified the ISH-based expression pattern of Lrrk2.
Analysis of Lrrk2-expressing Neuronal Subtypes ISH, Western blot and real-time PCR analyses revealed highest Lrrk2 expression level in samples from cortex and striatum, wherein the last one is known to be the major target area of the dopaminergic neurons from the SN. Based on this observation we studied in detail the cellular expression of Lrrk2 in the striatum by double in situ-hybridization/Immunohistochemical (ISH/IHC) stainings: on paraffin sections of the adult striatum nonradioactive ISH (nonfluorescent dark precipitate as signal indicating strong Lrrk2 expression) were performed, followed by fluorescent IHC for different neuronal markers. High magnifications of radioactive ISH brightfield pictures gave further insights into the cellular expression of Lrrk2. High expression was associated with neurons as identified by their typical morphology in the cresyl violet (CV) counterstaining (black arrows in Figure 5A). Most but not all nonneuronal cells, marked by their more condensed appearance in CV stainings, were devoid of Lrrk2 mRNA (white arrowheads in Figure 5A) We could confirm this finding by a high overlap of Lrrk2 mRNA expressing cells with the neuronal marker NeuN (neuronal nuclei antigen) in striatal sections by ISH/IHC stainings (data not shown), indicating a predominant but not exclusive neuronal expression of Lrrk2. In addition, virtually all striatal cells strongly expressing Lrrk2 mRNA were also positive for the dopamine and cAMP regulated phosphoprotein 32 (DARPP-32) (data not shown); a marker for medium spiny neurons (MSNs) in the striatum [40], suggesting that Lrrk2 plays a functional role in dopaminoceptive neurons. Hence, we studied the cellular expression of Lrrk2 mRNA in the two canonical dopaminergic circuits of the basal ganglia, the direct and indirect pathway, which are marked by the expression of either dopamine receptors D1 (DRD1a) or D2 (DRD2), respectively. The quantification of ISH/ IHC for Lrrk2 and DRD1a depicted in the striatum 36% of the cells being Lrrk2-positive, 25% DRD1a-positive and 39% double positive ( Figure 5B,D). In case of the DRD2 ISH/IHC, 38% of the cells were solely Lrrk2-positive, 24% DRD2-positive and 37% double positive cells could be detected ( Figure 5C,E). This data indicates that neither the direct nor the indirect pathway preferentially expresses Lrrk2. In addition it could be observed that only subpopulations show strong Lrrk2 expression in both pathways.

Analysis of Lrrk2 mRNA Transcripts and Splice Variants
Alternative splicing increases the diversity of mRNAs expressed from the genome, which has profound functional effects [41]. Therefore, we completed the detailed Lrrk1 and Lrrk2 expression analysis by determining Lrrk2 splice variants and their expression in various brain regions and organs from adult mice. Therefore, we first studied the genomic organization of Lrrk2 for the presence of exons which could be spliced out without causing dramatic rearrangements of the residual Lrrk2-transcript (i.e. exons with a number of base-pairs that is divisible by three and thus can be skipped without causing a frame-shift in the residual mRNA coding sequence). Thereafter, we performed six different RT-PCR analyses using primer pairs amplifying the regions of exons 4-9, 16-21, 30-34, 35-39, 41-45 and 47-51 ( Figure S5A). All reactions produced single amplicons representing the canonical Lrrk2 mRNA ( Figure S5B-F) except the primer combination for exons 4 to 9, which revealed an additional, smaller band of unknown origin ( Figure 6B). We analysed RNA samples from the brain regions with the highest Lrrk2 expression (striatum and cortex), but also cerebellum, olfactory bulb and hippocampus. In addition, samples from several organs were analysed using the same primers. Interestingly, all samples analysed showed this additional lower band in a comparable ratio. Sequence analysis of the fragments revealed a novel splice variant in which exon 5 (135bp) is absent ( Figure S6). To test for cell-type-specific expression within the CNS, the same primers were used to analyse RNA samples from primary neuronal, microglia and astrocyte cultures, which all were prepared from cortical tissues of C57Bl/6J mice. As already suggested by the ISH expression analysis, the endogenous Lrrk2 transcript could be detected in all three cell types with several primer pairs (data not shown). Using primers for exons 4-9, we found the same pattern in samples from neurons and microglia as observed in organs and brain tissues ( Figure 6C). However, in astrocytes the splice variant without exon 5 was dominantly expressed whereas the full length transcript was barely detectable ( Figure 6C). In addition, we performed a qPCR expression analysis using a probe located within exon 5 specific for the full-length Lrrk2 transcript and a probe located in the junction between exon 4 and 6 specific for the alternative Lrrk2 mRNA transcript with skipped exon 5 ( Figure 6F). Putative differences in the sensitivity of the primer combinations and probes may not allow a direct quantitative comparison between endogenous and alternative form. Nevertheless, the results confirmed a pronounced expression of alternative Lrrk2 mRNA transcript in primary astrocytes compared to primary neurons and microglia.
Next, we analysed all samples for the presence of an alternative exon 42a (containing several stop codons), which was already annotated in ensembl.org as part of an alternative processed transcript (ensembl.org sequence ENSMUST00000140734, Figure S7). In contrast to the full length Lrrk2 transcript, which consists of 51 exons, 8275 base-pairs and 2527 amino acids (ensembl.org sequence ENSMUST00000060642), the annotated sequence encompasses only 19 of these exons (3452 base-pairs), starts in exon 24 and ends with the alternative exon 42a ( Figure  S5). There is no information available about the sequence which is 59 of the annotated 19 exons. Thus, it could be possible, that this annotation is part of an alternative Lrrk2 transcript, that starts regularly with exon 1, but ends with exon 42 (6456bp and 2152 amino acids long). To confirm the expression of this transcript, we performed RT-PCR analyses using primers amplifying the sequence between exon 39 and the alternative exon 42a ( Figure 6D). Interestingly, expression of this alternative exon 42a could be identified in all brain regions and organs analysed. Sequencing of the fragment confirmed the alternative splicing event from the canonical exon 41 to the alternative exon 42a ( Figure S8). NCBI blast of the corresponding sequence demonstrated that the alternative exon 42a was located between exons 41 and 42 within the published Lrrk2 sequence ( Figure S8). To answer the question whether the expression is -like in case of skipped exon 5 -limited to certain cellular subtypes in the murine CNS, we analysed RNA samples from primary neuronal, microglia and astrocyte cultures and found identical patterns in samples from neurons and astrocytes ( Figure 6E). Interestingly, no band was detectable in samples from microglia, indicating that these cells may not process transcripts containing the alternative exon 42a. For confirmation, we utilized a qPCR analysis using specific primers and probes for the full-length Lrrk2 transcript (i.e. primers located in the endogenous exon 42) and primers and probes located in the alternative exon 42 ( Figure 6G). As expected, in primary microglia we could barely detect the expression of the shortened Lrrk2 mRNA transcript ending in the alternative exon 42, whereas primary neurons and astrocytes show significant higher expression of this variant. Taken together, we were able to identify and describe a noveland so far unknown -splice variant of Lrrk2 (skipped exon 5) and confirm the presence of an already published alternative transcript (alternative exon 42a) in RNA samples from different brain regions, organs and primary neuronal cell cultures. In addition, we . Expression analysis of Lrrk2 mRNA and protein in various brain regions from adult mice. Quantifications of Lrrk2 level in various brain regions by qPCR (mRNA) and Western blot (protein). Before calculation, LRRK2 protein level from Western blot analyses (A, representative Western blot) were normalized to beta-actin (ACTB) as loading control and Lrrk2 mRNA level from qPCR to a beta-actin control accordingly. (B) To allow a direct comparison between Lrrk2 mRNA and protein level, the sum of all individual relative expression level were set to 100% in both cases ({percentage brain region}z. = {relative level brain region}/{sum of all relative level}6100%). The definition of brain regions used for this analysis is indicated in the figure legend (B). Abbreviations: Ac, anterior cortex; Bs, brain stem; Cb, cerebellum; Dc, diencephalon; Hi, hippocampus; Mb, midbrain; Ob, olfactory bulb; Pc, posterior cortex; St, striatum. doi:10.1371/journal.pone.0063778.g004 found differential expression of these new transcripts in a cell-type specific manner.

Discussion
Several studies analyzing mRNA and protein expression of LRRK1 and LRRK2 in mouse, rat, pig and human samples of brain, organs/tissues and cells have been published so far, demonstrating largely ubiquitous expression of Lrrk2 in the CNS and peripheral tissues [30,31,[40][41][42][43][44][45][46][47][48][49][50][51][52][53][54]. Our present study presents novel findings, which gives potentially new insights into the physiological function of these proteins. To our knowledge, this is the first time that Lrrk1 and Lrrk2 expression was comparatively analysed, qualitatively and quantitatively, using several independent and highly sensitive analytical methods. With this we have thus described all important aspects like anatomical localization, expression level, distribution in neuronal cell populations and presence of splice variants in a single study.
We show a detailed analysis of Lrrk1 and Lrrk2 expression during embryonic development (E7.5 to E15.5) as well as in the postnatal (P0 to P21) and adult (.6 weeks) brain of C57BL/6J mice. For this purpose, we used radioactive labeled riboprobes, which are more sensitive than non-radioactive probes and thus are able to identify low expression sites [37]. In addition, we used up to 6 different overlapping and not overlapping riboprobes per gene ( Figure S9), which all showed identical results after hybridization (data not shown). We also used the corresponding sense RNA probes as negative controls to ensure the specificity of the expression pattern. We found moderate ISH signals for Lrrk1 and Lrrk2 during embryogenesis (E10 to E12.5) with their expression patterns showing common characteristics in many organs and structures. Total level of Lrrk1 were slightly stronger and more widespread than Lrrk2. With ongoing organogenesis (E15.5), the overall mRNA level of both genes increased and their non-neuronal expression domains became broader. Increasing Lrrk2 expression during embryonic development was also described by Zechel et al., who performed ISH with DIG-labeled riboprobes on brain sections from C57BL/6 mice and did not observe Lrrk2 expression in the CNS before embryonic stage E12.5 [54]. In addition, similar differences in total brain mRNA level of Lrrk1 and Lrrk2 were shown in two independent studies using radioactive ISH and qPCR [30,31]. First, in rat embryos, Westerlund et al. determined postnatal day P8 to be the earliest time-point in which Lrrk2 expression could be observed in the CNS, while widespread Lrrk1 expression was found already at embryonic stage E 13.5 [31]. Secondly, Biskup et al. showed, in samples from whole mouse brains, that between embryonic stages E11.5 and E17.5 the relative mRNA level of Lrrk1 were constantly higher than those of Lrrk2 [30]. The fact that we and others -in contrast to our ISH studies -detected considerable level of Lrrk1 and Lrrk2 mRNA by RT-PCR or real-time PCR [30,54] in samples of whole embryos and embryonic brain tissues, even at early stages of embryonic development, is not surprising. This can be explained by the observed strong expression of both genes in the choroid plexus and the meninges, which tightly surround the complete CNS and its ventricular system. In line with the non-neuronal expression of Lrrk2 during development is the expression in the Rathke's pouch; the non-neural origin of the adenohypophysis. Strikingly, in adulthood Lrrk2 expression can be found in the anterior and intermediate lobe (adenohypophysis) and not in the posterior lobe (neurohypophysis) (data not shown), which is composed of cells of the ventral diencephalon [55]. Interestingly, there are also two groups presenting stronger ISH signals for Lrrk1 in the developing CNS of rats [31] and for Lrrk2 in the developing CNS of mice [54]. Nevertheless, even though these data may conflict with our findings, they could possibly originate from species or technical differences between their and our studies.
During embryonic development, we observed much stronger ISH signals for Lrrk1 and Lrrk2 expression in kidney, lung, liver and heart when compared to all CNS regions. This is fully in line with other comparative expression studies, most of which  . Qualitative expression analysis of alternative Lrrk2 mRNA transcripts. RNA samples from different brain regions and organs were analysed by RT-PCR using a primer combination that amplifies Lrrk2 transcripts between exon 4 and exon 9 (A). Note that there is -beside the band of the endogenous Lrrk2 (ex.4-9) -an additional lower band visible in all samples analysed representing a mRNA where exon 5 is spliced out (ex.5 skipped) (B). The same primers were used to analyse RNA samples from primary neuronal cultures (i.e. neurons, microglia and astrocytes) (C). Interestingly, the band representing the splice variant without exon 5 (E5 skipped) was dominantly expressed in samples from astrocytes, while the band for the endogenous transcript (E4-E9) was almost completely missing in these cells. (D) RNA samples were analysed by RT-PCR using primers that amplify Lrrk2 transcripts between exon 39 (E39) and an alternative exon 42a (ex.42a). Note that the alternative exon 42a is present in all samples analysed (ex. 39-42a). (E) The same primers were used to analyse RNA samples from primary neuronal cultures (i.e. neurons, microglia and astrocytes). quantified mRNA level of different organs and brain regions by qPCR: (i) Zimprich et al. presented highest Lrrk2 mRNA level in human lung [8]. (ii) Biskup et al. described that both Lrrk1 and Lrrk2 mRNA level were much stronger in murine neonatal lung, heart and kidney than in whole brain samples [30]. (iii) Larsen et al. demonstrated highest Lrrk2 mRNA expression in porcine testis, spleen and lung [45]. (iv) Maekawa et al. showed highest Lrrk2 mRNA level in kidney, spleen, lung and testis even in samples from adult (20, 50 and 98 week old) mice [46]. Taken together, Lrrk1 and Lrrk2 are widely expressed during whole embryogenesis in non-neuronal tissue whereas their expression in neural tissue is barely detectable. On the other hand, Lrrk2 mutant mice (knockout, LRRK2 overexpression or introduction of pointmutations) do not show gross developmental phenotypes; they survive and show normal growth and life span [56][57][58][59]. This suggests that Lrrk2 function during development is not necessary for survival, but is required for non-vital physiological function within non-neuronal tissues.
After birth, Lrrk2 expression became robustly detectable in the murine CNS. It then increased further during early postnatal development in a highly dynamic manner and reached its maximum level around 6 weeks after birth, remaining stable throughout the lifespan. Thus, we confirmed the observations from other studies, namely ascending expression level of Lrrk2 during different stages of postnatal development [31], [25,30]. In the adult mouse brain, strong ubiquitous Lrrk2 ISH signals were present, with highest level in striatum, cortex, hippocampus, cerebellum, anterior olfactory nucleus and brain stem. Interestingly, during these first postnatal weeks synaptogenesis reaches a peak, specifically in the striatum and cortex [60,61]. This raises the questions whether Lrrk2 is directly involved in synaptogenesis or is part of the synaptic machinery. Indeed, a synaptic function of Lrrk2 -as suggested by its appearance in postnatal stages -is also supported by the fact that Lrrk2 has been implicated in neurotransmitter release [36,62]. In addition, LRRK2 protein expression becomes robustly detectable in primary hippocampal and cortical neuron-cultures (prepared from E15-16 embryonic brains) from day 7 onwards: a time point when the neuronal networks in the cultures are established [36]. Surprisingly, the mesencephalic region of the substantia nigra (SN), which contains the most vulnerable neuronal populations affected in PD, exhibited only moderate Lrrk2 expression in the adult brain: the target area of these dopaminergic neurons (i.e. caudate putamen in the striatum) shows the highest level of Lrrk2 mRNA. Most likely due to the rather low expression level of Lrrk2 in the SNc, there have been conflicting statements concerning Lrrk2 expression in this region. Some groups present only low to undetectable level of Lrrk2 in the SNc of mice [47,53] and rats [41,53]. On the other hand, others found clearly detectable Lrrk2 expression in the SNc of humans [40,43,51], mice [44,49,52] and rats [40,42]. Still, the combination of these and the present study clearly indicates that Lrrk2 is indeed expressed in the adult SNc.
In contrast to Lrrk2, Lrrk1 was barely detectable in the adult mouse brain and only present in the mitral cell layer of the olfactory bulb. It is quite surprising that, despite their high degree of homology and their comparable expression patterns during development, the adult brain almost exclusively expresses Lrrk2. Nevetheless, this finding suggests that LRRK1 does not compensate for LRRK2 function in the adult brain and -as a consequence -is not responsible for the absence of neurodegeneration in genetic animal models of Lrrk2 [62]. Furthermore it may also explain why, up to now, no PD-associated mutation has been identified in LRRK1 [28].
We further verified our observed ISH-based expression pattern of Lrrk2 by exact quantification of mRNA and protein level in samples from different brain regions using real-time PCR and Western blot, respectively. The observed patterns match remarkably, not only with the strongest Lrrk2 expression domains identified by our ISH analyses, but also with Northern blot [7], real-time PCR [48,53,63] and Western blot data [49] published by other groups. Interestingly, apart from striatum, no discrepancies between the relative protein and mRNA level in the different brain regions could be detected, suggesting no major posttranscriptional regulation of Lrrk2 outside of the striatum. The discrepancy within the striatum could have manifold explanations including (i) posttranscriptional regulation of Lrrk2 in the striatum via specific microRNAs, leading to blocking of translation [64], (ii) reduced half-life of LRRK2 protein in the striatum due to protein truncations, protein degradation or diminished protein stabilization, (iii) the opposite scenario -increased half-life of the LRRK2 protein in the striatum, so that less mRNA transcripts would be required to maintain the necessary protein level, or (iv) transport of the protein from the striatum to other brain compartments. Thus, further analyses are necessary to distinguish between these theories.
In the next part of the study, Lrrk2 expression was analysed on the cellular level in sections from striatum and cortex with focus on dopaminoceptive neurons. Elucidation of neuronal versus nonneuronal (i.e. the role of Lrrk2 in microglia function as mediator of neuroinflammation) localization of LRRK2 could help to understand its physiological function and/or its role in the pathogenesis of Lrrk2-associated PD [65][66][67]. Indeed, our studies also confirmed expression of Lrrk2 in glial cells; however, the bulk of the expression in the CNS was neuronal. In addition, almost 100% of the Lrrk2+ cells in the striatum were positive for the MSN-marker DARPP-32, indicating that Lrrk2 might play a role in dopaminoceptive neurons. Quantification of double ISH/IHC in the striatum revealed that about 61% (36% Lrrk2+, 25% DRD1a+, 39% Lrrk2+ and DRD1a+) of DRD1a+ MSNs, representing the direct pathway, highly express Lrrk2 mRNA. In this regard, MSNs of the indirect pathway (DRD2+) exhibited almost identical percentages (38% Lrrk2+, 24% DRD2+, 37% Lrrk2+ and DRD2+), indicating that only subpopulations of both pathways express high level of LRRK2. This notion was supported by a partial colocalization of Lrrk2 with DRD1a+ and DRD2+ neurons in the striatum. These data confirm the study of Mandemakers et al. where, via double IHC, LRRK2 protein was shown to colocalize with DRD1a in striatal cells [47] on the mRNA level. Our quantitative data also reveal that only subpopulations of both the direct and the indirect pathway express very high level of LRRK2. Thus, we hypothesize that Lrrk2 could play a role in both dopaminergic pathways, which are directly related to the Interestingly, the alternative exon 42a could not be amplified from primary microglia cDNA. (F) Relative expression level as measured by quantitative RT-PCR with primers and probes specific for either the endogenous Lrrk2 transcript (full-length Lrrk2), or for the alternative Lrrk2 mRNA transcripts with skipped exon 5 (ex.5 skipped). The relative expression is depicted as a ratio between endogenous Lrrk2 transcript and the alternative product on a logarithmic scale. (G) Accordingly for the alternative Lrrk2 mRNA transcript ending in the alternative exon 42a (alt.ex.42), the relative expression level have been determined by quantitative RT-PCR with specific primers and probes for the alternative product and normalized to the expression of the endogenous Lrrk2 transcript (full-length Lrrk2). doi:10.1371/journal.pone.0063778.g006 neuropathology of PD [68]. Nevertheless, differential function(s) of Lrrk2 in DRD1a positive and in DRD2 positive subpopulations of neurons has to be evaluated.
Next, we concentrated on the search for splice variants of the murine Lrrk2 mRNA transcript in various brain regions and organs from adult mice. Zimprich et al. [8] (2004) presented the very first hints for alternative Lrrk2 splicing, when they noticed weak bands at lower sizes in Northern blot analyses of human brain samples; this was also observed by an independent group [7]. When they used RT-PCR to quantify transcripts containing exons 1-3, 12-13 and 32-34 however, they did not find significant differences in the expression level [8]. Later, West et al. described the presence of at least six transcription start sites upstream of the first Kozak sequence and one downstream of the first predicted exon (within intron 1). This would exclude the complete Lrrk2 ORF and produce a protein of unknown biological significance [69]. However, they also found no evidence for alternate Lrrk2 exons or protein coding sequences [69]. Therefore, we analysed the comparably large Lrrk2 transcript for exons, which could be spliced out without causing dramatic rearrangements, and performed RT-PCR analyses using primer pairs amplifying the regions of exons 4-9, 16-21, 30-34, 35-39, 41-45 and 47-51. After screening these six different regions, we were able to identify a novel -and so far unknown -splice variant of Lrrk2, sparing exon 5 (135bp).
Ignacio Marin has published, in 2008, the detailed structure of selected LRRK proteins, including human LRRK2, and described the presence of 14 N-terminal LRRK2-specific repeats in the human and mouse sequence [21]. We calculated that exon 5 covers virtually the complete sequence of the fourth repeat. The lack of exon 5 (and consequently the loss of the fourth repeat) may result in a protein with altered size (45 amino acids smaller) and unknown biological significance. Nevertheless, a possible consequence could be a change in the affinity to different LRRK2 binding partners (i.e. alterations in the LRRK2 interactome). Interestingly, this alternative transcript was present in all samples analysed. When we analyzed RNA samples from primary neuronal, microglia and astrocyte cultures (all prepared from cortical tissues) however, we found identical patterns in neurons and microglia. Surprisingly, there was a dominant expression of the splice variant without exon 5 in astrocytes. Thus, astrocytes may exhibit a different interactome of LRRK2 and therefore LRRK2 function may differ in neurons and astrocytes. Furthermore, we were able to confirm the presence of an alternative transcript that contains an alternative exon 42a (including several stop codons) and was located between the endogenous exons 41 and 42 (annotated in ensembl.org). Interestingly, we also found cell-type specific expression patterns of these transcripts between the different primary cell cultures. While neurons and astrocytes both process the transcript containing the alternative exon 42a, no band was visible in microglia. This suggests a rather low expression of the alternative variant compared to the full-length Lrrk2 transcript. As mentioned above, we have no information about the sequence that is 59 of exon 39 in this transcript. Assuming that the transcript starts regularly with exon 1, it could induce a maximum protein product of 239 kDa (in contrast to the 280 kDa full-length LRRK2 protein). This could somehow explain why we (data not shown) and others [30,46,47,50,65,70,71] observe additional bands in Western blot analyses, which could fit to the calculated size of a truncated LRRK2 isoform. Considering the complex architecture of the LRRK2 protein, implying that Lrrk2 is involved in more than one cellular process [72], we found that the endogenous exons 41, 42 and 43 cover the complete end of the MAPKKK domain [73].
Splicing that introduces an alternative exon 42a will lead to an early truncation of the protein (due to its included stop codons, and may seriously influence its enzymatic activity (due to lack of the c-terminal part of the kinase domain). In addition, the complete WD40 domain would also be absent in such a truncation. Jorgensen et al. demonstrated that removing the WD40 domain prevents complex formation and autophosphorylation but also blocks the neurotoxicity of LRRK2 PD mutations [74]. Moreover, it was shown that fragments of LRRK2 containing the ROCO domain mediate LRRK2 kinase inhibition [75]. It is thus tempting to speculate that the splice variant with alternative exon 42a is acting as an inhibitor of intrinsic LRRK2 function. The absence of this variant in microglia would, in turn, render this cell type more susceptible to LRRK2 pathogenic mutations via further stimulation or through missing the regulation of its potentially toxic kinase activity. It is interesting to note that attenuated inflammatory responses after knockdown of LRRK2 in microglia indicate that it may be a positive regulator of neuroinflammation [67]. Taken together, we observed cell-type specific differences in the expression of full-length, exon 5-skipped and truncated (E1-E42a) transcripts. Thus, we assume that Lrrk2 features different physiological functions in different neural cell populations and cell-type specific differences in its enzymatic functions and its interactome. This is in full agreement with Lewis et al. [72], who postulated that alterations in different glial cell lines could selectively affect susceptible neuronal populations in PD. Furthermore, Lrrk2 may be regulated by different signals; interacting with different upstream regulators or downstream effectors: phosphorylating different substrates, and so formdepending on the presence of appropriate cofactors, interacting proteins or other favorable environmental conditions -tissueselective or non-selective signaling complexes. This results in celltype specific molecular effects and pathology [72]. Further efforts are necessary to figure out if these splicing patterns are conserved between mouse and humans, or if they represent interesting differences between species.
In conclusion, we present here a substantial and detailed analysis of Lrrk1 and Lrrk2 expression, focusing on qualitative and quantitative aspects (anatomical localization, expression level, distribution in neuronal cell populations and presence of splice variants) and covering numerous stages ranging from ES cells, embryonic development until the postnatal and adult brain. To our knowledge, we show for the first time a quantitative analysis of colocalization between Lrrk2 mRNA and both DRD1a and DRD2 expression in MSNs of the striatum. In addition, we were able to identify a novel splice variant of Lrrk2 (skipped exon 5) and confirm an annotated transcript (alternative exon 42a) in RNA samples from different brain regions, organs and primary neuronal cell cultures. Besides that, we found cell-type specific alterations in the transcript expression patterns in the different neuronal lineages.