Biochemical Society Transactions

LRRK2: Function and Dysfunction

Pharmacological inhibition of LRRK2 cellular phosphorylation sites provides insight into LRRK2 biology

Jing Zhao , Spencer B. Hermanson , Coby B. Carlson , Steven M. Riddle , Kurt W. Vogel , Kun Bi , R. Jeremy Nichols

Abstract

Mutations in LRRK2 (leucine-rich repeat kinase 2) have been linked to inherited forms of PD (Parkinson's disease). Substantial pre-clinical research and drug discovery efforts have focused on LRRK2 with the hope that small-molecule inhibitors of the enzyme may be valuable for the treatment or prevention of the onset of PD. The pathway to develop therapeutic or neuroprotective agents based on LRRK2 function (i.e. kinase activity) has been facilitated by the development of both biochemical and cell-based assays for LRRK2. LRRK2 is phosphorylated on Ser910, Ser935, Ser955 and Ser973 in the N-terminal domain of the enzyme, and these sites of phosphorylation are likely to be regulated by upstream enzymes in an LRRK2 kinase-activity-dependent manner. Knowledge of these phosphorylation sites and their regulation can be adapted to high-throughput-screening-amenable platforms. The present review describes the utilization of LRRK2 phosphorylation as indicators of enzyme inhibition, as well as how such assays can be used to deconvolute the pathways in which LRRK2 plays a role.

  • 14-3-3 protein
  • high-throughput screening
  • leucine-rich repeat kinase 2 (LRRK2)
  • phosphorylation
  • Parkinson's disease

Introduction

LRRK2 (leucine-rich repeat kinase 2) is a large multidomain protein consisting of 2527 amino acids with a molecular mass of nearly 0.3 MDa. The minimal fragment of LRRK2 that exhibits kinase activity encompasses the ROC (Ras of complex proteins)–COR (C-terminal of ROC) domain which contains a GTPase, the kinase domain and the entire C-terminal region which contains a WD40 repeat-like domain [1]. LRRK2 that has been purified from eukaryotic cells exhibits a specific activity that is orders of magnitude lower than other kinases, leading to speculation about the activation state of LRRK2 in mammalian cells. The first commercial mass-produced active kinase preparation of LRRK2 was made available by Invitrogen (now a part of Life Technologies) [2]. This enzyme is expressed in insect cells, encompasses amino acids 970–2527 and is now widely utilized as a source of enzyme for in vitro assays.

The LRRK2 gene is found on chromosome 12 in humans and is located in a mutational ‘hotspot’, with more than 50 mutations being found in the LRRK2 gene and not in control DNA samples [3]. LRRK2 mutations comprise approximately 1% of idiopathic PD (Parkinson's disease) cases and 5% of familial PD cases. The most commonly found mutation in LRRK2 replaces the highly conserved glycine residue within subdomain VII of the kinase domain with serine (G2019S). This mutation has been found to increase kinase activity not by increasing affinity for substrates, but by increasing the catalytic rate of the enzyme [1,2,4]. The prevailing hypothesis is that mutation-induced changes in LRRK2 kinase activity result in deregulation of a downstream substrate, which then leads to degeneration of the substantia nigral neurons that are lost in PD. In this model, development of small-molecule inhibitors of LRRK2 kinase activity could serve a therapeutic function by repressing the overactive LRRK2 regulatory system. Five other pathogenic mutations have been observed in the LRRK2 gene, namely those resulting in R144C, R1441G, Y1699C, I2020T and N1437H [3,5]. These mutations are found within either the ROC–COR or the kinase domain of the enzyme. Most of these mutations have negligible impact on kinase activity, but do induce discrete cytoplasmic accumulations of the enzyme [611]. This allows for a tentative classification of LRRK2 mutations between those that do not increase kinase activity (R144C, R1441G, Y1699C and I2020T) and those that do (G2019S).

Assaying LRRK2 in vitro and in vivo

In vitro assay of LRRK2

Early methods for monitoring LRRK2 kinase activity relied on autophosphorylation of the kinase itself or transphosphorylation of MBP (myelin basic protein). Autophosphorylation rarely goes to completion during kinetic analyses, and LRRK2 has been shown to modify at least 20 residues on itself [1214]. Additionally, MBP has been shown to be modified by LRRK2 at multiple sites at low stochiometry [15]. Given the low catalytic activity of LRRK2 and the promiscuity of MBP as a substrate from many (potentially contaminating) serine/threonine kinases, the use of MBP as a substrate is problematic.

The need for an improved substrate that would enable routine (and high-throughput compatible) monitoring of LRRK2 activity in vitro was met in 2007, when Jaleel et al. [1] discovered that peptide fragments derived from the proteins ezrin, radixin and moesin (ERM proteins) were efficiently modified by LRRK2 in vitro, using the kinase substrate tracking and elucidation method. A peptide comprising the amino acids surrounding Thr558 of moesin, RLGRDKYKT558LRQIRQ is now a commonly used substrate called LRRKtide. This discovery enabled adaptation to high-throughput screening of compound libraries, and has no doubt moved the field forward in terms of understanding the molecular inhibition profile of LRRK2 with established kinase inhibitors. Following this discovery, the LRRKtide substrate was optimized by using a positional scanning peptide library screen to identify a higher-affinity substrate; this substrate (RLGWWRFYTLRRARQGNTKQR) is referred to as Nictide [16].

Both academic and industrial entities have adopted the above assays or variants thereof to define inhibitors of LRRK2. These efforts have isolated multiple compounds that exhibit LRRK2 inhibitory properties [1621]. The first relatively selective and novel tool compound inhibitor of LRRK2, LRRK2-IN1, was disclosed in 2011 [27]. Two other potent and relatively selective inhibitors of LRRK2 have been characterized and disclosed as well: CZC-25146 and TAE684 [22,23]. These tool compounds have allowed for invaluable assessment of the effects of acute LRRK2 inhibition.

LRRK2 phosphorylation sites

Much work has been done to elucidate the regulation of LRRK2. Since many kinases are regulated by upstream kinases, many groups have focused on understanding the sites of LRRK2 phosphorylation and their regulation [6,13,14,24]. Interestingly, LRRK2 has been found to exhibit two classes of phosphorylation sites: autophosphorylation sites that arise after an in vitro kinase assay, and sites of ‘constitutive phosphorylation’ that are observed when active LRRK2 is isolated from mammalian cells. Many groups have characterized LRRK2 autophosphorylation sites with the hope that these can serve as reliable indicators of LRRK2 kinase activity in vivo with immunological detection. However, no sites of LRRK2 autophosphorylation have been identified as being robustly modified in cells and as being detectable without incubation with Mg2+/ATP. Only with time, antibody development and testing can we know whether these autophosphorylation sites are able to be utilized as in vivo readouts of LRRK2 kinase activity.

Independent groups have identified LRRK2 phosphorylation in a cluster of serine residues that precede the leucine-rich repeat domain using MS of mammalian-cell-derived LRRK2 [6,13]. Gloeckner et al. [13] compared the phosphorylation sites of LRRK2 incubated with and without Mg2+/ATP. This comparison confirmed that there are multiple phosphorylation sites present on LRRK2 without in vitro phosphorylation, which were described as constitutive phosphorylation sites. Four of these constitutive sites have been validated with phosphospecific antibodies, namely Ser910, Ser935, Ser955 and Ser973 [6,17,25]. Although the simplest explanation would be that these sites are autophosphorylation sites, two studies have taken efforts to demonstrate directly that these are not direct LRRK2 substrate (autophosphorylation) sites [17,26]. This was accomplished by determining whether dephosphorylated LRRK2 was able to be rephosphorylated at these sites in in vitro kinase assays, and indeed this was not observed. We refer to these four constitutive sites of LRRK2 modification as the ‘cellular phosphorylation sites’ to reflect the demonstrated likelihood that the modification of these sites is by kinases other than LRRK2, and to distinguish them from sites of autophosphorylation. Other potential cellular phosphorylation sites within that cluster have been observed by MS (Ser860 and Ser926), but these have yet to be investigated with immunological reagents.

The phosphorylation status of Ser910, Ser935, Ser955 and Ser973 are linked to pathogenic LRRK2 PD mutations and protein–protein interactions

Certain LRRK2 PD pathogenic mutations disrupt the modification of the four LRRK2 cellular phosphorylation sites. LRRK2 G2019S is fully phosphorylated at these sites, in contrast with the R1441G/C, Y1699C and I2020T mutants, which display a striking decrease in phosphorylation at Ser910, Ser935, Ser955 and Ser973. Interestingly, mutations of either Ser910 or Ser935 induce cytoplasmic accumulations of LRRK2, whereas mutation of Ser955 and Ser973 do not. The similar localization phenotypes associated with Ser910 and Ser935 functionally link phosphorylation with R144G/C, Y1699C and I2020T mutations.

LRRK2 interacts with 14-3-3 proteins via phosphorylated Ser910 and Ser935 [6]. The phosphorylation of these sites was shown to be necessary and sufficient for 14-3-3 protein association. Mutation of Ser955 and Ser973 have no effect on 14-3-3 interactions, but these sites do show disrupted phosphorylation in the context of S910A/S935A mutations, suggesting that phosphorylation at these positions may be linked to phosphorylation at the 910 and 935 positions. LRRK2 R1441G/C, Y1699C and I2020T mutants also show decreased LRRK2–14-3-3 binding, since these mutants show decreased phosphorylation at the 910 and 935 cellular sites.

LRRK2 cellular phosphorylation sites are regulated by LRRK2 kinase activity.

Interestingly, the cellular phosphorylation sites are regulated in a manner that is dependent on LRRK2 kinase activity. Acute inhibition with the aforementioned selective inhibitors of LRRK2 results in a rapid dephosphorylation of the 910, 935, 955 and 973 cellular phosphorylation sites within 30 min of treatment [17,26,27]. LRRK2 interaction with 14-3-3 proteins is disrupted due to this loss of phosphorylation at the 910 and 935 sites [6,25]. Acute inhibition of LRRK2 also forces the relocalization of the enzyme to skein-like structures in the cytoplasm [17,26,27]. It is possible that 14-3-3 binding through association with the 910 and 935 phosphosites is required to maintain LRRK2 in a diffusely cytoplasmic state.

Use of LRRK2 IRM (inhibitor-resistant mutant) to validate direct inhibition of LRRK2

Many of the compounds that target LRRK2 are derivatives of existing kinase inhibitor scaffolds and therefore have the potential for off-target effects. A tool that can be useful to ensure direct engagement of a compound with the kinase active site of LRRK2 is the utilization of an LRRK2 IRM. Previous work had shown that inhibitors of ROCK (Rho-associated kinase) are poor inhibitors of PKA (protein kinase A) owing to a steric clash imparted by a threonine residue preceding the critical and conserved aspartic acid residue of subdomain VII, and that substitution of alanine for this threonine residue was sufficient to render PKA sensitive to ROCK inhibitors [28]. LRRK2 has an alanine at this position preceding the aspartic acid residue (A2016DYG), and after performing the converse substitution, A2016T, LRRK2 was shown to be ‘desensitized’ to some kinase inhibitors [16,22,26,27]. There are, however, differences in the degree of desensitization as exhibited by differences in the shift in IC50 values seen with various inhibitors, regardless of selectivity for LRRK2. However, the LRRK2 IRM is a useful tool for drug development efforts, as this variant has also been shown to exhibit desensitization to inhibitor treatment in the cell-based assays described below, which facilitates an indication of direct target engagement in cells.

Using pharmacological inhibition of LRRK2 phosphorylation to provide insights into LRRK2 biology

Cellular phosphorylation sites as readouts of pharmacological inhibition

Understanding LRRK2 localization, activity and regulation of downstream enzymes can be derived from an understanding of the regulation of Ser910, Ser935, Ser955 and Ser973. These cellular phosphorylation sites can provide an avenue to elucidate not only direct inhibitors of LRRK2, but also the potential to understand the biology of signalling events that feed into and from this important enzyme. Figure 1 illustrates this concept. LRRK2 is hypothesized to activate a downstream kinase(s) that feeds back to modify the cellular phosphosites or to down-regulate a downstream phosphatase that removes the cellular phosphorylation sites (indicated by a broken arrow in Figure 1). These as yet unknown enzymes represent potential novel targets for LRRK2. In addition to using small molecules that may target these enzymes, siRNA (small interfering RNA) libraries targeting the upstream kinase(s) or phosphatase(s) could also be employed to elucidate these relevant LRRK2 regulators. It is plausible that LRRK2 cellular phosphorylation sites are regulated by many kinases and phosphatases and these may be LRRK2-independent. In this case, siRNAs and inhibitors could facilitate the identification of those enzymes.

Figure 1 Model of signalling to LRRK2 cellular phosphorylation sites and points of intervention

LRRK2 is phosphorylated on Ser910, Ser935, Ser955 and Ser973 in a manner dependent on LRRK2 kinase activity. Therefore LRRK2 may be a component of a signal transduction pathway (broken arrows) that directly or indirectly affect the phosphorylation at these sites (+ or −℗). This regulation could be the potentiation of a kinase activity towards the cellular sites or negative regulation of a phosphatase. Alternatively, there could also be kinases that are independent of LRRK2 kinase activity. Points at which one could block these pathways with inhibitors or reverse genetics are indicated.

To date, a robustly modified endogenous substrate of LRRK2 (with appropriately sensitive phosphoantibodies) has yet to be disclosed, leaving the immunological detection of LRRK2 phosphorylation at the 910, 935, 955 or 973 cellular phosphosites as a suitable alternative. In the absence of a method to detect in vivo inhibition of phosphorylation of a direct LRRK2 downstream substrate, immunological tracking of the LRRK2 cellular phosphorylation sites is an appropriate alternative indicator of in vivo inhibition of LRRK2.

Small-molecule screening

Figure 2 shows a scheme for identifying LRRK2 inhibitors starting from a library of small-molecule inhibitors. Figure 2(A) shows a potential high-throughput-screening-amenable adaptation of a cellular phosphosite modification on LRRK2, which we have successfully applied to small-molecule in vivo screening for LRRK2 [29]. Here, the phosphorylated residue of a GFP (green fluorescent protein)-tagged LRRK2 is recognized by a terbium-labelled antibody. Phosphorylation-dependent detection of LRRK2 can be detected by time-resolved fluorescence resonance energy transfer, from the excited terbium label on the antibody to elicit GFP fluorescence. Figure 2(B) shows the workflow if a library is first screened in vitro against the recombinant protein. Compounds found to inhibit LRRK2 in this assay should then be screened for their impact on the cellular phosphosites, which in this context is limited in the potential to reveal LRRK2 activity-independent regulation of those amino acids. If the compound induces LRRK2 dephosphorylation in the wild-type, but not the A2016T mutant, a bona fide LRRK2-specific inhibitor has been identified.

Figure 2 Using LRRK2 cellular phosphorylation sites to identify inhibitors and LRRK2 biology

(A) Example of how immunological detection of LRRK2 phosphorylation at the cellular phosphorylation sites (℗) can be adapted to high-throughput analysis. A terbium (Tb)-labelled antibody can detect phosphorylated GFP-tagged LRRK2. A ratiometric measure of the amount of phosphorylation is achieved by exciting at 340 nm (Tb donor) and taking emission ratios from 495 nm donor (Tb) and 520 nm emission from the GFP acceptor for the time-resolved fluorescence resonance energy transfer pair and plotted as 520/490. (B) and (C) Workflows if a compound library is screened directly against LRRK2 recombinant protein first (B) or in a cellular assay detecting first (C); details are given in the text. HTS, high-throughput screening; WT, wild-type.

Figure 2(C) shows the flow if a library screen is first performed in the cell-based assay. If a loss of LRRK2 phosphosignal is observed, it can be from either direct inhibition of LRRK2 or perturbation of the factors that lead to phosphorylation of LRRK2 (or by alleviating a repressed phosphatase activity). To assess whether LRRK2 is targeted directly, the compounds should be assayed against recombinant enzyme and cells expressing an LRRK2 IRM, which should show a resistance to compound inhibition. If the cellular rescreen of a compound using LRRK2 IRM as the target reveals a similar IC50 value to that seen with the wildtype LRRK2, it is likely that the compound interferes with the activity of an enzyme acting on the cellular phosphosites, and therefore reveals pertinent biology feeding into LRRK2.

The utility of the cellular phosphorylation sites extends to animal models. It has been shown that Ser910 and Ser935 are modified in all tissues where detectable levels of LRRK2 have been studied, e.g. kidney, lung, spleen and brain [26,27]. Intraperitoneal administration of compounds that target LRRK2 can be assessed for efficacy by evaluation of their effects on the phosphorylation of Ser910 and Ser935. Since LRRK2 is highly expressed in peripheral nucleated cells, it is possible that these sites can also be utilized as surrogate peripheral markers for LRRK2 inhibitor efficacy in patients during clinical trials.

Conclusions

The kinase LRRK2 is a disease-relevant protein for which small-molecule inhibitors may serve as a therapeutic or preventative medicine for PD. The exploitation of LRRK2 cellular phosphorylation sites is an avenue to validate inhibitors in a cell-based assay and define new biology around LRRK2 modifications. Certainly, there are many more biological pathways that are affected by LRRK2 in addition to those surrounding the cellular phosphorylation sites and gratifyingly, the field is moving forward rapidly. Studying the regulation of Ser910, Ser935, Ser955 and Ser973 phosphorylation will give insights into important components of that biology.

Funding

This work was supported by funding from the Michael J. Fox Foundation and the benevolence of the Brin/Wojcicki foundation to R.J.N. and internal funding at Life Technologies. R.J.N. is a consultant for Life Technologies.

Acknowledgments

We thank the members of the Nichols Laboratory for helpful discussions.

Footnotes

  • LRRK2: Function and Dysfunction: A Biochemical Society Focused Meeting held at Royal Holloway, University of London, Egham, UK, 28–30 March 2012. Organized and Edited by Patrick Lewis (University College London, U.K.) and Dario Alessi (Dundee, U.K.).

Abbreviations: COR, C-terminal of Ras of complex proteins; GFP, green fluorescent protein; IRM, inhibitor-resistant mutant; LRRK2, leucine-rich repeat kinase 2; MBP, myelin basic protein; PD, Parkinson's disease; PKA, protein kinase A; ROC, Ras of complex proteins; ROCK, Rho-associated kinase; siRNA, small interfering RNA

References

View Abstract