Does Phosphorylation of Presenilin 1 Drive Microglial Migration, Aβ Clearance?

Best known for its role in churning out Aβ peptides as the catalytic subunit of γ-secretase, presenilin-1 may also promote digestion of those Aβ peptides within microglia. That is the upshot of a study published August 13 in Molecular Psychiatry. Led by the late Paul Greengard of Rockefeller University in New York, the study found that in mice expressing a form of PS1 that cannot be phosphorylated on serine-367, microglia became overloaded with autophagic vacuoles and lysosomes that failed to acidify. These microglia migrated sluggishly in response to brain injury. In a mouse model of amyloidosis, the phosphorylation-deficient PS1-S367A mutant stymied microglial digestion of Aβ and exacerbated its accumulation. Together with previous findings from the lab, the study attempts to build a case that presenilin-1 influences amyloidosis via multiple mechanisms and within different cell types.

  • PS1-S367A knock-in mice cannot phosphorylate PS1 on serine-367.
  • Their microglia migrate slowly and have lysosomal defects.
  • In 5xFAD mice, PS1- S367A exacerbates Aβ accumulation.

“This study reminds us that PS1 harbors multiple functions other than the proteolysis,” wrote Taisuke Tomita of the University of Tokyo. “Further analysis of this regulatory mechanism would provide a novel understanding of PS1 biology.” Tomita added that it remains unclear whether this microglial PS1 function is relevant to the pathogenesis of AD. He wondered how familial AD mutations in PS1 might affect its microglial-specific functions.

Previously, a pair of studies from Greengard’s lab cast PS1 as a driver of autophagy in neurons (Jun 2017 news). The researchers reported that when phosphorylated at serine-367, PS1 teamed up with Annexin 2 and other vesicular proteins to facilitate the fusion of autophagosomes to lysosomes. This promoted the autophagic digestion of APP β-C-terminal fragments (CTF), thus keeping Aβ production in check. When the researchers swapped endogenous mouse PS1 for the S367A mutant, amyloidosis skyrocketed in J20 mice. Thwarting phosphorylation of PS1 at this residue did not appear to affect the processing of APP by γ-secretase, suggesting the modification is not essential for proteolysis.

In the new study, first author Jose Ledo and colleagues expanded their investigation to microglia, since these cells use autophagic digestion to clear extracellular debris, including Aβ. The researchers started by assessing fundamental characteristics of microglia—migration and ramification—in PS1-S367A knock-in mice. Using cranial windows to watch the cells in action, the researchers found that they migrated more slowly to the site of a laser-induced injury than did microglia in wild-type mice. Microglia in the knock-ins also extended fewer processes, with less elaborate branching patterns than those expressing normal PS1. Phospho-deficient PS1 did not appear to change how many microglia were there.

How did it alter microglial function? The researchers gleaned some inkling from gene-expression analyses, which uncovered 121 differentially expressed genes when compared with microglia from wild-type mice. Genes involved in phagosome maturation and autophagy were among those most affected. Microglial expression of ATP6v0a1, a vacuolar-ATPase that acidifies the lysosome, was reduced by half in the knock-ins.

Microglial Indigestion. Electron microscopy indicates an overabundance of autophagic vacuoles (arrows) in microglia sorted from phospho-deficient PS1-S367A mice (right) compared with wild-type mice (left). Vacuoles are quantified on right. [Courtesy of Ledo et al., Molecular Psychiatry, 2020.]

In keeping with the gene-expression data, microglia isolated from PS1-S367A mice had a glut of autophagic vacuoles chock-full of undigested material. Their lysosomes had a higher pH, suggesting acidification was not working properly. In agreement with the defect, primary microglia from the knock-in mice internalized Aβ oligomers normally, but did not go on to digest them. The researchers could correct lysosomal defects by overexpressing TFEB, a master transcriptional regulator of autophagy. Ramping up TFEB in primary microglia not only restored levels of ATP6v0a1, but normalized lysosomal pH.

Ralph Nixon, Nathan Kline Institute, New York, is unconvinced that reductions in vacuolar ATPase explain the lysosomal deficits. He has previously reported that APP CTFs raise lysosomal pH. “It is puzzling why this demonstrated action of APP-CTF on lysosomal pH, given the group’s previous report of APP-βCTF elevations in presumably the same lysosomal compartment, was not considered as a possible mechanism for the phospho-deficient mutant, but instead a highly speculative alternative was favored,” he wrote (see comment below).

Does this microglial indigestion influence amyloidosis? To investigate, the researchers crossed the PS1 phospho-deficient mice with 5xFAD mice. Confocal microscopy of brain sections from 3-month-old offspring revealed microglia stuffed with Aβ. Congo Red staining followed by iDISCO, a technique that renders brain tissue transparent, revealed Aβ plaques throughout the brain. Compared with 5xFAD mice, those expressing the phospho-deficient PS1 had a higher plaque load in 35 of about 180 brain regions analyzed.

Aβ Plaque Disco. Tissue clearing reveals substantially more Aβ plaques (Congo Red, gold) in brain hemispheres from PS1-S367A mice. [Courtesy of Ledo et al., Molecular Psychiatry, 2020.]

By mixing APP with microglial extracts, the researchers again found that γ-secretase processing of its C-terminal fragments were unaffected by the S367A mutation in PS1. However, the same was not true for another γ-secretase substrate, Notch. Its processing was enhanced in microglial extracts from the knock-in mice compared with those from wild-type mice. This suggested that PS1 S367 phosphorylation affects γ-secretase activity toward some substrates and not others.

Hui Zheng of Baylor College of Medicine in Houston noted that it is difficult to comprehend how S367A phosphorylation would mediate distinct effects on APP and Notch, and the phenomenon requires further study. “Further, since the S367A mutation is not found in PSEN1 FAD cohorts, the relevance of the microglial phenotypes observed in the KI mice to AD is not evident,” Zheng added.

Ledo told Alzforum that he sees this study as a start to investigating the function of PS1 in microglia. Most studies have focused on the role of the protein in neurons. Previous studies from Greengard’s lab suggested that in neurons, PS1 drives autophagy by promoting the fusion of autophagosomes to lysosomes. However, in microglia, phosphorylation of the protein appears to influence acidification of lysosomes. It is possible that PS1 function is regulated differently depending on cell type, he said. Investigation of PS1 function in microglia is an ongoing focus in the lab, he said, including the generation of conditional knock-in mice that express phospho-deficient PS1, or PS1 with familial AD mutations, only in microglia.

It remains to be seen if these lysosomal functions of PS1 are relevant for the human brain. Ledo said that he has detected phosphorylation of S367 in human fibroblasts, and phosphorylation of the residue has been documented for numerous other human cell types (see PhosphoSitePlus). Rapid changes in phosphorylation complicate postmortem analyses of PS1 phosphorylation in the human brain, though Ledo plans to work with brain banks to minimize this problem. The researchers are also investigating PS1 function in human iPSC-derived neurons and microglia. One hint that PS1 phosphorylation might be affected in the AD brain comes from a methylation study, which found that the gene encoding CK1g2—the kinase that phosphorylates serine-367—is hypermethylated in the brains of people with sporadic AD and poorly expressed (Semick et al., 2019). Reduced expression of this kinase would theoretically lead to less phosphorylation of PS1 and greater Aβ accumulation, Ledo noted.—Jessica Shugart

Comments

  1. The authors extend their previous finding that PS1 phosphorylation regulates the autophagy-lysosomal pathway. Using knock-in mice carrying S367A mutant PS1, they found that phospho-deficient PS1 disturbed the microglial function, causing accumulation of Aβ as well as PSD95 in the microglia in 5xFAD model mice. They concluded that PS1 regulates Aβ pathology at multiple levels.

    They have investigated the microglial dysfunction in the knock-in model by several approaches. One intriguing point of this study is that the regulation of microglial function by PS1 was independent of γ-secretase activity. This study reminds us that PS1 carries out multiple functions other than the proteolysis. Further analysis of this regulatory mechanism would provide a novel understanding of PS1 biology.

    However, it remains unclear whether PS1-mediated microglial function is involved in the pathogenesis of AD, since the authors did not analyze the effect of FAD mutations on the microglia. Several reports suggest that FAD mutations in PS1 cause the partial loss of function in terms of proteolytic function. However, authors showed that γ-secretase inhibitors do not alter the microglial function in vitro. Nevertheless, it would be intriguing to see whether the phosphorylation of PS1 and/or the microglial function was altered by FAD-linked mutations of PS1.

    Moreover, some have reported that PS2, but not PS1, is a critical factor in the regulation of immune responses (Jayadev et al., 2010; Agrawal et al., 2015; Fung et al., 2020). The PS1 phosphorylation site is PS1 specific, but these data raise the possibility that PS-regulated microglial dysfunction modifies the pathophysiology of AD.

    References:

    Jayadev S, Case A, Eastman AJ, Nguyen H, Pollak J, Wiley JC, Möller T, Morrison RS, Garden GA. Presenilin 2 is the predominant γ-secretase in microglia and modulates cytokine release. PLoS One. 2010;5(12):e15743. PubMed.

    Agrawal V, Sawhney N, Hickey E, McCarthy JV. Loss of Presenilin 2 Function Is Associated with Defective LPS-Mediated Innate Immune Responsiveness. Mol Neurobiol. 2015 Jun 18; PubMed.

    Fung S, Smith CL, Prater KE, Case A, Green K, Osnis L, Winston C, Kinoshita Y, Sopher B, Morrison RS, Garden GA, Jayadev S. Early-Onset Familial Alzheimer Disease Variant PSEN2 N141I Heterozygosity is Associated with Altered Microglia Phenotype. J Alzheimers Dis. 2020;77(2):675-688. PubMed.

    View all comments by Taisuke Tomita
    • User Profile Image Hui Zheng
      Baylor College of Medicine
    • Posted:

    The current work is a continuation of the group’s investigations on PS1 serine 367 phosphorylation. Previously the authors showed that this phosphorylation affects, via autophagy and independently of γ-secretase, the conversion of APP β-CTF to Aβ. Here, the authors studied the role of this phosphorylation in microglia using Psen1 S367A phospho-deficient knock-in (KI) mice. They reported that the KI mice exhibit altered microglia morphology, impaired microglial response to brain injury, and increased Aβ and PSD95 in microglia. Thus, the work reveals a novel role of PS1 in microglia regulation through S367 phosphorylation, by which it mediates Aβ metabolism. Intriguingly, and in contrast to their previous report that the PS1 S367 phosphorylation does not affect γ-secretase processing of APP, here the authors show increased γ-secretase activity on Notch signaling in KI microglia. How the S367 phosphorylation mediates distinct effects on APP and Notch is difficult to comprehend and warrants further study. Further, since the S367 mutation is not found in PSEN1 FAD cohorts, the relevance of the microglial phenotypes observed in the KI mice to AD is not evident.

    View all comments by Hui Zheng

  3. Although mostly studied in neurons, Presenilin1 (encoding for PS1) is broadly expressed in the immune system, including microglia. The authors of this interesting work suggest that PS1 is an important regulator of the lysosomal machinery in microglia. The mechanism is not fully elucidated, but the authors suggest that phosphorylation of the PS1 intracellular domain may play some role in the activation of the transcription-factor TFEB, which induces the expression of lysosomal genes and lysosomal acidification. Of note, this type of signaling occurs in a γ-secretase independent manner, therefore it is uncoupled from the mechanism of APP processing.

    The authors generated a transgenic mouse strain harboring a point mutation in the PS1 phosphorylation site S367. To study the role of PS1 phosphorylation in AD-like pathology, these mice were further crossed with the 5xFAD line (a popular model of accelerated amyloid-β deposition). Double-transgenic mice developed a more severe amyloid pathology and synaptic loss compared with 5xFAD control mice expressing a wild-type form of Psen1. In summary, the authors suggest that PS1 phosphorylation in microglia promotes lysosomal degradation of the engulfed Aβ. This novel aspect of PS1 biology may be important not only in AD, but also in other protein misfolding and storage diseases in which tissue resident macrophages are involved.

    View all comments by Marco Colonna View all comments by Simone Brioschi
    • User Profile Image Ralph Nixon
      New York University School of Medicine/Nathan Kline Institute
    • Posted:

    Familial Alzheimer’s Disease (FAD) presenilin1 (PSEN1) mutations or PSEN1 deletions are well-established to cause a γ-secretase-independent failure of lysosomal acidification in neurons and to involve a premature ERAD degradation of the V0a1 vATPase subunit and impaired vATPase assembly (Lee et al., 2010; Lee et al., 2015; Wolfe et al., 2013; Lee et al., 2020; Coffey et al., 2014). This loss-of-function (LOF) effect of PSEN1 mutation on pH regulation has been shown as far back as yeast (Sharma et al., 2019) and across many cell types (Wolfe et al., 2013; Lee et al., 2020; Coffey et al., 2014). Though not surprising, it is therefore reassuring that effects on lysosomal proteolysis and autophagy similar to those of PSEN1 FAD mutations can be recapitulated in microglia when modeled by a phosphorylation-deficient variant of presenilin having similar loss-of-function properties. If dysregulation of such a variant in microglia can be shown to have Alzheimer’s disease relevance, it is likely that there would be negative consequences for phagocytic and proteolytic clearance functions. Given the cell type and species conservation of PSEN1-dependent acidification mechanisms involving V0a1 trafficking and assembly of the vATPase complex and also the lysosomal chloride ion channel ClC7 (Lee et al., 2020), it would be predicted that microglial clearance functions are also impaired in PSEN1 FAD. If so, this may help explain the greater abundance of amyloid and plaque debris in PSEN1 FAD cases also implying slower clearance rather than greater production of amyloid.

    In a previous report (Bustos et al., 2017), the Rockefeller group proposed that the same phospho-deficient mutant interfered with neuronal autophagosome-lysosome fusion and, in doing so, increased levels of APP-βCTF. The rise in APP-βCTF levels noted in that study and the recent findings in the microglial model of impaired lysosomal function due to a rise in pH accord remarkably well with our recent report demonstrating an inhibitory effect of APP-CTF on lysosomal pH (Jiang et al., 2019) and our preliminary report showing this is due to a direct interfering interaction of APP-βCTF with the vATPase (Im et al., ADPD poster presentation 2019). 

    It is puzzling why this demonstrated action of APP-CTF on lysosomal pH and the group’s previous report of APP-βCTF elevations in presumably the same lysosomal compartment were not considered as a possible mechanism for the phospho-deficient mutant and instead a highly speculative alternative was favored. Ledo et al. propose that the 10 percent decrease in the mRNA level of expression of the V0a1 subunit mRNA (FDR. KI-WT, 0.045) or roughly 50 percent by qPCR (p<0.05) is the underlying explanation for the observed lysosomal pH change. While plausible, this is unlikely, especially in the absence of any substantiation such as measuring vATPase complex subunit levels or vATPase complex function. It is well known that mRNA expression is a risky predictor of steady-state protein levels and, in this regard, our RNA-Seq data on PSEN1 FAD fibroblasts and PSEN1 deletion from neurons and other cell types uniformly show an increased expression of V0a1 subunit mRNA yet decreased lysosomal level of V0a1 protein and lowered vATPase function associated with a pH rise. Restoration of normal pH in this model by strong TFEB overexpression, as Ledo et al. reported, cannot be considered persuasive evidence that the minimally lowered V0a1 mRNA is the cause of pH dysfunction.

    Going forward, it will be interesting to know whether presenilin1 phosphorylation is physiologically regulated in neurons or glial subtypes and to establish whether autophagy/lysosomal functions in microglia are impaired in PSEN1 FAD, as predicted.

    References:

    Lee JH, Yu WH, Kumar A, Lee S, Mohan PS, Peterhoff CM, Wolfe DM, Martinez-Vicente M, Massey AC, Sovak G, Uchiyama Y, Westaway D, Cuervo AM, Nixon RA. Lysosomal proteolysis and autophagy require presenilin 1 and are disrupted by Alzheimer-related PS1 mutations. Cell. 2010 Jun 25;141(7):1146-58. PubMed.

    Lee JH, McBrayer MK, Wolfe DM, Haslett LJ, Kumar A, Sato Y, Lie PP, Mohan P, Coffey EE, Kompella U, Mitchell CH, Lloyd-Evans E, Nixon RA. Presenilin 1 Maintains Lysosomal Ca(2+) Homeostasis via TRPML1 by Regulating vATPase-Mediated Lysosome Acidification. Cell Rep. 2015 Sep 1;12(9):1430-44. Epub 2015 Aug 20 PubMed.

    Wolfe DM, Lee JH, Kumar A, Lee S, Orenstein SJ, Nixon RA. Autophagy failure in Alzheimer's disease and the role of defective lysosomal acidification. Eur J Neurosci. 2013 Jun;37(12):1949-61. PubMed.

    Lee JH, Wolfe DM, Darji S, McBrayer MK, Colacurcio DJ, Kumar A, Stavrides P, Mohan PS, Nixon RA. β2-adrenergic Agonists Rescue Lysosome Acidification and Function in PSEN1 Deficiency by Reversing Defective ER-to-lysosome Delivery of ClC-7. J Mol Biol. 2020 Apr 3;432(8):2633-2650. Epub 2020 Feb 24 PubMed.

    Sharma D, Otto G, Warren EC, Beesley P, King JS, Williams RS. Gamma secretase orthologs are required for lysosomal activity and autophagic degradation in Dictyostelium discoideum, independent of PSEN (presenilin) proteolytic function. Autophagy. 2019 Aug;15(8):1407-1418. Epub 2019 Mar 21 PubMed.

    Coffey EE, Beckel JM, Laties AM, Mitchell CH. Lysosomal alkalization and dysfunction in human fibroblasts with the Alzheimer's disease-linked presenilin 1 A246E mutation can be reversed with cAMP. Neuroscience. 2014 Mar 28;263:111-24. Epub 2014 Jan 10 PubMed.

    Bustos V, Pulina MV, Bispo A, Lam A, Flajolet M, Gorelick FS, Greengard P. Phosphorylated Presenilin 1 decreases β-amyloid by facilitating autophagosome-lysosome fusion. Proc Natl Acad Sci U S A. 2017 Jul 3;114(27):7148-7153. Epub 2017 May 22 PubMed.

    Jiang Y, Sato Y, Im E, Berg M, Bordi M, Darji S, Kumar A, Mohan PS, Bandyopadhyay U, Diaz A, Cuervo AM, Nixon RA. Lysosomal Dysfunction in Down Syndrome Is APP-Dependent and Mediated by APP-βCTF (C99). J Neurosci. 2019 Jul 3;39(27):5255-5268. Epub 2019 May 1 PubMed.

    Im E, Jiang Y, Lee J-H, Nixon RA. APP-βCTF regulates vATPase-mediated lysosomal acidification. Poster session presented at AD/PD The 14th International Conference on Alzheimer’s & Parkinson’s Diseases. Lisbon, Portugal 2019 March 26-31.

    View all comments by Ralph Nixon

  5. This manuscript of Ledo and co-workers follows up on their earlier work on a S367-phosphorylation-deficient mutant of PSEN1. Their previous work highlighted a role for S367 phosphorylation in an alternative degradation of β-C-terminal fragments of amyloid-β precursor protein through autophagy and mediated through interactions with Annexin2 and SNARE proteins (Bustos et al., 2017, and 2017). They now document a novel role specifically in microglia. They did so by creating microglia-selective phosphorylation-deficient KI mice and found multiple defects that could be trailed back to a corrupted phagocytosis-lysosome pathway.

    In general, it is exciting that besides neurons, proper PSEN1 function is required for normal microglial biology as well, and in particular for Aβ clearance, highlighting again that likely much more is yet to be discovered about these proteins and in particular about γ-secretase in the context of AD pathogenesis. This study underscores once more the impact a single post-translational modification can have on not only Aβ production but now on Aβ clearance as well. We also reported that a single phosphorylation site in a sorting motif in PSEN2 is sufficient to dramatically alter its subcellular localization and consequently alter the Aβ42/40 ratio (Sannerud et al., 2016). 

    Another intriguing aspect of this work is the γ-secretase-independent nature of the observation: Is this related to other reports of a moonlighting role of PSENs, for instance related to a functional connection to the cytoskeleton in the moss P. patens (Khandelwal et al., 2007), and lysosomal Ca2+ storage/release and autophagy/lysosome defects in PSEN-deficient cells (Coen et al., 2012; Neely et al., 2011)? Because of the relevance of endolysosomal abnormalities in early stages of AD pathogenesis, this is an aspect that requires more investigation: For instance, are PSENs acting herein in or out of the context of other γ-secretase subunits? This study doesn’t answer that question. The authors only tested Notch but it might be worthwhile to investigate other γ-secretase substrates that are functionally involved in Aβ clearance, including TREM2, LRP, and others. Likewise, if the S367A-PSEN1 affects Aβ degradation, is this mutant actually localized on phagosomes or autophagolysosomes?

    With respect to the observed alterations in gene expression, this experiment was done on the whole pool of isolated microglia; it might be worth investigating to what extent the alterations the authors observe (in phagosome maturation, 14-3-3 and ERK/MAPK pathways) are restricted or maybe more pronounced in the pool of disease-associated microglia.

    References:

    Bustos V, Pulina MV, Kelahmetoglu Y, Sinha SC, Gorelick FS, Flajolet M, Greengard P. Bidirectional regulation of Aβ levels by Presenilin 1. Proc Natl Acad Sci U S A. 2017 Jul 3;114(27):7142-7147. Epub 2017 May 22 PubMed.

    Bustos V, Pulina MV, Bispo A, Lam A, Flajolet M, Gorelick FS, Greengard P. Phosphorylated Presenilin 1 decreases β-amyloid by facilitating autophagosome-lysosome fusion. Proc Natl Acad Sci U S A. 2017 Jul 3;114(27):7148-7153. Epub 2017 May 22 PubMed.

    Sannerud R, Esselens C, Ejsmont P, Mattera R, Rochin L, Tharkeshwar AK, De Baets G, De Wever V, Habets R, Baert V, Vermeire W, Michiels C, Groot AJ, Wouters R, Dillen K, Vints K, Baatsen P, Munck S, Derua R, Waelkens E, Basi GS, Mercken M, Vooijs M, Bollen M, Schymkowitz J, Rousseau F, Bonifacino JS, Van Niel G, De Strooper B, Annaert W. Restricted Location of PSEN2/γ-Secretase Determines Substrate Specificity and Generates an Intracellular Aβ Pool. Cell. 2016 Jun 30;166(1):193-208. Epub 2016 Jun 9 PubMed.

    Khandelwal A, Chandu D, Roe CM, Kopan R, Quatrano RS. Moonlighting activity of presenilin in plants is independent of gamma-secretase and evolutionarily conserved. Proc Natl Acad Sci U S A. 2007 Aug 14;104(33):13337-42. PubMed.

    Coen K, Flannagan RS, Baron S, Carraro-Lacroix LR, Wang D, Vermeire W, Michiels C, Munck S, Baert V, Sugita S, Wuytack F, Hiesinger PR, Grinstein S, Annaert W. Lysosomal calcium homeostasis defects, not proton pump defects, cause endo-lysosomal dysfunction in PSEN-deficient cells. J Cell Biol. 2012 Jul 9;198(1):23-35. PubMed.

    Neely KM, Green KN, LaFerla FM. Presenilin is necessary for efficient proteolysis through the autophagy-lysosome system in a γ-secretase-independent manner. J Neurosci. 2011 Feb 23;31(8):2781-91. PubMed.

    View all comments by Wim Annaert

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Does Presenilin 1 Moonlight as Autophagy Driver, Paradoxically Reducing Aβ?

Research Models Citations

  1. 5xFAD (B6SJL)

Paper Citations

  1. Semick SA, Bharadwaj RA, Collado-Torres L, Tao R, Shin JH, Deep-Soboslay A, Weiss JR, Weinberger DR, Hyde TM, Kleinman JE, Jaffe AE, Mattay VS. Integrated DNA methylation and gene expression profiling across multiple brain regions implicate novel genes in Alzheimer's disease. Acta Neuropathol. 2019 Apr;137(4):557-569. Epub 2019 Feb 2 PubMed.

External Citations

  1. PhosphoSitePlus

Further Reading

Papers

  1. Matz A, Halamoda-Kenzaoui B, Hamelin R, Mosser S, Alattia JR, Dimitrov M, Moniatte M, Fraering PC. Identification of new Presenilin-1 phosphosites: implication for γ-secretase activity and Aβ production. J Neurochem. 2015 May;133(3):409-21. Epub 2015 Feb 24 PubMed.

Primary Papers

  1. Ledo JH, Liebmann T, Zhang R, Chang JC, Azevedo EP, Wong E, Silva HM, Troyanskaya OG, Bustos V, Greengard P. Presenilin 1 phosphorylation regulates amyloid-β degradation by microglia. Mol Psychiatry. 2020 Aug 13; PubMed.

Annotate

To make an annotation you must Login or Register.