DREAM (Downstream Regulatory Element Antagonist Modulator)
In 1998, Carrión et al. discovered a 110 kDa protein complex present in nuclear extracts of human neuroblastoma cells that could bind to a small region of DNA (+36 to +44 relative to the transcription start site) within the 5′ untranslated region (5′-UTR) of the human prodynorphin gene. Occupancy of this DNA element, termed the downstream regulatory element (DRE), by the protein complex caused a significant attenuation in prodynorphin gene transcription (Carrión et al. 1998). Utilizing a double-stranded DRE oligonucleotide probe to screen a human caudate expression library, Carrión et al. (1999) subsequently cloned the gene encoding for the DRE-binding protein, which they named downstream regulatory element antagonist modulator (DREAM). Electrophoretic mobility shift assays (EMSAs) confirmed that DREAM can bind to DRE and that the DREAM-DRE interaction is abrogated by increased Ca2+ concentration (Carrión et al. 1999). Numerous studies have since been devoted to identifying additional genes that possess DRE sequences within their 5′-UTRs, as well as the phenotypic abnormalities that are associated with either genetic ablation of dream or expression of Ca2+-insensitive mutants of DREAM. It was recognized early on that DREAM has pleiotropic functions beyond Ca2+-dependent transcriptional repression: these functions are reflected in its alternative names of potassium channel-interacting protein 3 (KChIP3) (An et al. 2000) and calsenilin (Buxbaum et al. 1998), the latter referring to its ability to bind to both Ca2+ and the Alzheimer’s disease (AD) proteins, presenilin 1 (PS1) and presenilin 2 (PS2). DREAM/KChIP3/calsenilin belongs to the neuronal calcium sensor (NCS) superfamily of proteins, which are characterized by the presence of four EF-hand motifs, two or three of which are capable of binding Ca2+, as well as an N-myristoyl group (Burgoyne 2007).
Structure and Expression
In humans, dream/kchip3/calsenilin is located on chromosome 2q11.1. The murine dream/kchip3/calsenilin gene, also located on chromosome 2, consists of nine exons and can potentially encode up to four protein isoforms (Spreafico et al. 2001). Alternative splicing between exons 2 and 3 can result in the inclusion or exclusion of an ACAG nucleotide tetramer (Spreafico et al. 2001). Additionally, translation can be initiated at one of two alternative ATG start codons that are spaced 85 base pairs apart and are out-of-frame relative to each other (Spreafico et al. 2001). Exons 4, 5, 6, and 8 encode the four Ca2+-binding EF-hand domains that are critical to the functionality of DREAM/KChIP3/calsenilin (Spreafico et al. 2001). Two of the variants, which contain the four EF-hand motifs, are identical in protein sequence except for the presence of 29 additional amino acids at the N-terminus of the variant dubbed DREAM by Spreafico et al. (2001). In that study, the variant with the shorter N-terminus was called KChIP3/calsenilin. In reality, it is unclear if the presence of the extended N-terminus distinguishes DREAM from KChIP3 and calsenilin or if one is dealing with the same protein albeit with different names. The other predicted protein isoforms include an N-terminal truncation variant of DREAM that lacks the EF-hand domains and a variant that contains a unique C-terminus lacking EF-hand domains (Spreafico et al. 2001).
In terms of its expression, dream/kchip3/calsenilin mRNA exhibits strong basal expression in the thymus, testis, and thyroid gland (Carrión et al. 1999). DREAM/KChIP3/calsenilin protein is constitutively expressed in the spinal cord, hippocampus, retina, pineal gland, and thyroid gland (Link et al. 2004; Rivas et al. 2004; Cheng et al. 2002). Immunohistochemical analysis revealed that DREAM/KChIP3/calsenilin is very abundant in the cerebellar and retrosplenial granular cortices, as well as moderately present in the optic tract, superior colliculus, olfactory bulb, and various thalamic relay centers, such as the anterior dorsal, medial geniculate, dorsolateral geniculate, ventral posteromedial, and ventral posterolateral nuclei (Hammond et al. 2003). The promoter region of the dream/kchip3/calsenilin gene possesses DRE sequences, suggesting that DREAM can regulate its own expression through an autoinhibitory feedback loop that is Ca2+ dependent (Mellström et al. 2014).
Molecular and Cellular Functions
When intracellular Ca2+ levels are low, Ca2+-free (apo-)DREAM exists as a homotetramer that binds to DRE sequences within target DNA (Osawa et al. 2001). Mg2+ ions bind to the second EF-hand domain, and this interaction is critical in allowing DREAM to bind to DNA (Osawa et al. 2005). High levels of intracellular Ca2+ result in Ca2+ binding to the third and fourth EF-hand domains (Lusin et al. 2008). Ca2+ binding promotes the formation of DREAM homodimers, which are unable to interact with DRE sequences, resulting in transcriptional derepression (Lusin et al. 2008). DREAM recognizes and binds to the consensus DRE sequence 3′-PuNGTCAPuPuG-5′ (Carrión et al. 1998). DRE-containing genes that have been shown to be repressed by DREAM include the opioid precursor prodynorphin; the proto-oncogene c-fos (Carrión et al. 1998); brain-derived neurotrophic factor (bdnf) (Rivera-Arconada et al. 2010); sodium ion (Na+)-Ca2+ exchanger isoform 3 (ncx3) (Gomez-Villafuertes et al. 2005); voltage-dependent L-type calcium channel, alpha 1C subunit (cacna1c) (Ronkainen et al. 2011); the transcription factors neuronal PAS domain 4 (npas4) (Mellström et al. 2014), paired box gene 8 (pax8) (D’Andrea et al. 2005), and forkhead box protein E1 (foxe1) (D’Andrea et al. 2005); the thyroid protein precursor thyroglobulin (tg) (Rivas et al. 2004); calcitonin (Matsuda et al. 2006); the cytokines interleukin-2 (il-2), interleukin-4 (il-4), and interferon-γ (ifn-γ) (Savignac et al. 2005); the deubiquitinase tumor necrosis factor alpha-induced protein 3 (tnfaip3) (Tiruppathi et al. 2014); and the pro-apoptotic protein harakiri (hrk) (Sanz et al. 2001).
In addition to DRE-mediated transcriptional repression, it has been reported that DREAM is capable of activating gene transcription by directly binding to vitamin D and retinoic acid response elements that are located upstream of the transcription start site (Scsucova et al. 2005). Lastly, as discussed in detail in the next section, DREAM can also influence cAMP-dependent gene transcription (Ledo et al. 2002).
Beyond transcriptional regulation, DREAM is involved in other cellular processes through direct protein-protein interaction. As KChIP3, it binds to the cytoplasmic N-terminal domains of Kv4 potassium channels and promotes channel trafficking to the plasma membrane (An et al. 2000). KChIP3 also modulates the biophysical properties of Kv4 channels, reconstituting several features of native A-type currents (An et al. 2000). Ca2+ binding is essential for the ability of KChIP3 to modulate Kv4 channel activity but is not required for the physical association between the two proteins (An et al. 2000). As calsenilin, it interacts with the C-termini of PS1 and PS2 (Buxbaum et al. 1998). Presenilins are components of the γ-secretase protein complex, and the calsenilin-PS interaction has been shown to modulate γ-secretase activity (Jang et al. 2011; Jo et al. 2005).
The molecular functions of DREAM/KChIP3/calsenilin are dependent on specific structural properties of the protein as well as posttranslational modifications. For example, there are numerous positively charged amino acid side chains (K87, K90, K91, R98, K101, R60, and K166) that are clustered on one aspect of the protein surface that may mediate electrostatic interactions with target DNAs (Lusin et al. 2008). Nuclear localization of DREAM appears to depend on its sumoylation by the SUMO-conjugating enzyme UBC9 at two specific residues, lys-26 and lys-90 (Palczewska et al. 2011). G protein-coupled receptor kinase 2 (GRK2)-mediated phosphorylation of DREAM/KChIP3 at ser-95 facilitates KChIP3-dependent trafficking of Kv4.2 channels to the plasma membrane but has no effect on DREAM’s repressor activity (Ruiz-Gomez et al. 2007). Additionally, DREAM/KChIP3 must be palmitoylated at two adjacent residues, ser-45 and ser-46, in order to effectively traffic Kv4 channels to the plasma membrane (Takimoto et al. 2002).
Involvement in Intracellular Signaling
DREAM is an important downstream effector of Ca2+ signaling, as it transduces alterations in intracellular Ca2+ levels into changes in gene expression, which have the potential to dramatically alter the functional capability of the cell. In addition to direct transcriptional regulation through its binding to cis-acting elements, DREAM can also interact with the transcription factor cAMP response element-binding protein (CREB) and prevent its association with CREB-binding protein (CBP) (Ledo et al. 2002). Furthermore, when the α-isoform of cAMP-responsive element modulator (αCREM) becomes phosphorylated by cAMP-dependent protein kinase (PKA), it interacts directly with DREAM via leucine-rich domains, resulting in transcriptional derepression at DRE sites (Ledo et al. 2000). Thus, it appears that DREAM can coordinate the crosstalk between Ca2+ and cAMP signaling to influence gene expression. An example is DREAM-dependent transcriptional activation of the glial fibrillary acidic protein (gfap) gene, which requires both cAMP and Ca2+ signals (Cebolla et al. 2008).
Transcriptional regulation by DREAM can occur by mechanisms other than those described above. DREAM/calsenilin can bind directly to the transcriptional corepressor C-terminal binding protein 2 (CtBP2) in a Ca2+-independent manner (Zaidi et al. 2006). The DREAM/calsenilin-CtBP2 complex is suggested to recruit histone deacetylase (HDAC) enzymes to repress the transcription of numerous genes (Zaidi et al. 2006). DREAM can also directly interact with activating transcription factor-6 (ATF6), preventing its release from the endoplasmic reticulum (ER) membrane and nuclear translocation; the DREAM-ATF6 interaction ultimately blocks the expression of genes involved in the unfolded protein response (UPR) (Naranjo et al. 2016).
Several studies have identified a role of DREAM/calsenilin in programmed cell death (apoptosis). Abrogation of the DREAM-DRE interaction in the 3′-UTR of the hrk gene has been correlated with increased levels of apoptosis of hematopoietic progenitor cells (Sanz et al. 2001). In neuroblastoma cells, Ca2+-bound DREAM interacts with the anti-apoptotic protein hexokinase 1 (HK1), preventing its localization to the surface of mitochondria and ultimately promoting apoptosis (Craig et al. 2013). Another study showed that DREAM/calsenilin enhances apoptosis by promoting release of intracellular Ca2+ stores (Lilliehook et al. 2002). When Ca2+ levels are low, battenin/CLN3 protein suppresses apoptosis by binding to and sequestering DREAM/calsenilin (Chang et al. 2007). DREAM/calsenilin is itself a substrate of caspase-3 and is cleaved at the asp-61 and asp-64 residues (DXXD cleavage motif) to produce two small fragments ofundetermined functions (Choi et al. 2001). Phosphorylation of DREAM/calsenilin at ser-63 inhibits caspase-3-mediated cleavage (Choi et al. 2003).
DREAM can also regulate cellular excitability by various means. For instance, DREAM can bind directly to the NR1 subunit of N-methyl-D-aspartate (NMDA) receptors and attenuate cell surface expression of these receptors (Zhang et al. 2010). Through its Ca2+-dependent modulation of Kv4 channel activity, DREAM/KChIP3 can influence the repolarization properties of excitable cells following an action potential (An et al. 2000).
DREAM has been shown to interact with other proteins, including the antioxidant protein peroxiredoxin 3 (Prdx3) (Rivas et al. 2011) and calmodulin (Ramachandran et al. 2012), although the functional significance of these interactions is less clear. The DREAM-calmodulin interaction is mediated by the N-terminus of DREAM, between residues 29 and 44, and enhances calmodulin-dependent activation of the serine/threonine phosphatase calcineurin (Gonzalez et al. 2015; Ramachandran et al. 2012).
Much of what we know about DREAM’s physiological roles have come from studies using three independently generated knockout ( −/− ) mouse strains and transgenic mice expressing Ca2+-insensitive DREAM (TgDREAM) protein. Dream −/− mice exhibit a global reduction in pain responses that were attributed to increased expression of prodynorphin in the spinal cord (Cheng et al. 2002). The aversive effects of the cannabinoid tetrahydrocannabinol (THC) are also potentiated in these animals, whereas the analgesic effects of THC are reduced (Cheng et al. 2004). Calsenilin −/− mice have reduced levels of Aβ peptide, consistent with an alteration in γ-secretase activity (Lilliehook et al. 2003). An observed enhancement of hippocampal long-term potentiation (LTP) in calsenilin −/− mice corresponded with reduced Kv4 channel current (Lilliehook et al. 2003). Kv4 channel currents in cortical pyramidal neurons are modestly decreased in kchip3 −/− mice (Norris et al. 2010). Dream −/− and kchip3 −/− mice exhibit enhanced hippocampal-dependent learning (Alexander et al. 2009; Fontán-Lozano et al. 2009), potentially through facilitation of CREB-dependent transcription (Fontán-Lozano et al. 2009). Estradiol-enhanced memory formation also appears to be modulated by DREAM (Tunur et al. 2013). Ablation of dream slows age-dependent cognitive decline and hippocampal gliosis (Fontán-Lozano et al. 2009). In neonatal dream −/− mice, GFAP expression and the number of astrocytes are reduced in the cerebral cortex, suggesting a role in astrogliogenesis (Cebolla et a. 2008). TgDREAM mice exhibit impairments in hippocampal-dependent learning, along with a coincident reduction in bdnf expression and NMDA receptor-mediated current in the hippocampus (Mellström et al. 2014; Wu et al. 2010). The hippocampi of TgDREAM mice have reduced dendritic arborisation and spine density of CA1 pyramidal neurons and increased spine density of dentate gyrus granule neurons (Mellström et al. 2016). TgDREAM mice also exhibit reduced viability of cerebellar granule neurons under mild membrane-depolarizing conditions as a result of lower Ncx3 expression (Gomez-Villafuertes et al. 2005).
The physiological functions of DREAM are not restricted to the brain and spinal cord. TgDREAM mice have markedly enlarged thyroid glands, potentially through alteration of thyroid-stimulating hormone receptor (TSHR) activity (Rivas et al. 2009). Dream −/− mice exhibit reduced endotoxin-induced inflammatory injury of the lung and sepsis-induced death as a result of attenuated NF-κB signaling (Tiruppathi et al. 2014). The B cells of TgDREAM transgenic mice are hyperproliferative in vitro but show reduced immunoglobulin (Ig) synthesis (Savignac et al. 2010). siRNA-mediated knockdown of dream expression in human embryonic stem cells (hESCs) reduces hESC pluripotency and promotes cellular differentiation (Fontán-Lozano et al. 2016).
From 1998 to 2000, three independent studies emerged that identified DREAM/KChIP3/calsenilin as a Ca2+-dependent transcriptional repressor, a regulator of Kv4 potassium channels, and a presenilin-interacting protein. The functional pleiotropy of this structurally simple protein, characterized by the presence of four EF-hand motifs, was affirmed in later studies that implicated DREAM in many cellular and physiological processes, including transcriptional regulation by DRE-dependent and DRE-independent mechanisms, protein trafficking, apoptosis, learning and memory, pain modulation, and regulation of thyroid development and immune function. Given the importance of Ca2+ signaling in many aspects of cellular function, we may only be witnessing the tip of the proverbial iceberg in terms of the role of DREAM in mammalian physiology.
- Cheng HY, Pitcher GM, Laviolette SR, Whishaw IQ, Tong KI, Kockeritz LK, Wada T, Joza NA, Crackower M, Goncalves J, Sarosi I, Woodgett JR, Oliveira-dos-Santos AJ, Ikura M, van der Kooy D, Salter MW, Penninger JM. DREAM is a critical transcriptional repressor for pain modulation. Cell. 2002;108:31–43. doi:10.1016/S0092-8674(01)00629-8.PubMedCrossRefGoogle Scholar
- Choi EK, Zaidi NF, Miller JS, Crowley AC, Merriam DE, Lillihook C, Buxbaum JD, Wasco W. Calsenilin is a substrate for caspase-3 that preferentially interacts with the familial Alzheimer’s disease-associated C-terminal fragment of presenilin 2. J Biol Chem. 2001;276:19197–204. doi:10.1074/jbc.M008597200.PubMedCrossRefGoogle Scholar
- Craig TA, Ramachandran PL, Bergen 3rd HR, Podratz JL, Windebank AJ, Kumar R. The regulation of apoptosis by the downstream regulatory element antagonist modulator/potassium channel interacting protein 3 (DREAM/KChIP3) through interactions with hexokinase I. Biochem Biophys Res Commun. 2013;433:508–12. doi:10.1016/j.bbrc.2013.03.016.PubMedPubMedCentralCrossRefGoogle Scholar
- Fontán-Lozano A, Capilla-Gonzalez V, Aguilera Y, Mellado N, Carrión AM, Soria B, Hmadcha A. Impact of transient down-regulation of DREAM in human embryonic stem cell pluripotency: the role of DREAM in the maintenance of hESCs. Stem Cell Res. 2016;16:568–78. doi:10.1016/j.scr.2016.03.001.PubMedCrossRefGoogle Scholar
- Gomez-Villafuertes R, Torres B, Barrio J, Savignac M, Gabellini N, Rizzato F, Pintado B, Gutierrez-Adan A, Mellström B, Carafoli E, Naranjo JR. Downstream regulatory element antagonist modulator regulates Ca2+ homeostasis and viability in cerebellar neurons. J Neurosci. 2005;25:10822–30. doi:10.1523/JNEUROSCI.3912-05.2005.PubMedCrossRefGoogle Scholar
- Link WA, Ledo F, Torres B, Palczewska M, Madsen TM, Savignac M, Albar JP, Mellström B, Naranjo JR. Day-night changes in downstream regulatory element antagonist modulator/potassium channel interacting protein activity contribute to circadian gene expression in pineal gland. J Neurosci. 2004;24:5346–55. doi:10.1523/JNEUROSCI.1460-04.2004.PubMedCrossRefGoogle Scholar
- Mellström B, Sahún I, Ruiz-Nuño A, Murtra P, Gomez-Villafuertes R, Savignac M, Oliveros JC, Gonzalez P, Kastanauskaite A, Knafo S, Zhuo M, Higuera-Matas A, Errington ML, Maldonado R, DeFelipe J, Jefferys JG, Bliss TV, Dierssen M, Naranjo JR. DREAM controls the on/off switch of specific activity-dependent transcription pathways. Mol Cell Biol. 2014;34:877–87. doi:10.1128/MCB.00360-13.PubMedPubMedCentralCrossRefGoogle Scholar
- Mellström B, Kastanauskaite A, Knafo S, Gonzalez P, Dopazo XM, Ruiz-Nuño A, Jefferys JG, Zhuo M, Bliss TV, Naranjo JR, DeFelipe J. Specific cytoarchitectureal changes in hippocampal subareas in daDREAM mice. Mol Brain. 2016;9:22. doi:10.1186/s13041-016-0204-8.PubMedPubMedCentralCrossRefGoogle Scholar
- Naranjo JR, Zhang H, Villar D, González P, Dopazo XM, Morón-Oset J, Higueras E, Oliveros JC, Arrabal MD, Prieto A, Cercós P, González T, De la Cruz A, Casado-Vela J, Rábano A, Valenzuela C, Gutierrez-Rodriguez M, Li JY, Mellström B. Activating transcription factor 6 derepression mediates neuroprotection in Huntington disease. J Clin Invest. 2016;126:627–38. doi:10.1172/JCI82670.PubMedPubMedCentralCrossRefGoogle Scholar
- Ramachandran PL, Craig TA, Atanasova EA, Cui G, Owen BA, Bergen 3rd HR, Mer G, Kumar R. The potassium channel interacting protein 3 (DREAM/KChIP3) heterodimerizes with and regulates calmodulin function. J Biol Chem. 2012;287:39439–48. doi:10.1074/jbc.M112.398495.PubMedPubMedCentralCrossRefGoogle Scholar
- Ronkainen JJ, Hänninen SL, Korhonen T, Koivumäki JT, Skoumal R, Rautio S, Ronkainen VP, Tavi P. Ca2+-calmodulin-dependent protein kinase II represses cardiac transcription of the L-type calcium channel alpha(1C)-subunit gene (Cacna1c) by DREAM translocation. J Physiol. 2011;589:2669–86. doi:10.1113/jphysiol.2010.201400.PubMedPubMedCentralCrossRefGoogle Scholar
- Ruiz-Gomez A, Mellström B, Tornero D, Morato E, Savignac M, Holguín H, Aurrekoetxea K, González P, González-García C, Ceña V, Mayor Jr F, Naranjo JR. G protein-coupled receptor kinase 2-mediated phosphorylation of downstream regulatory element antagonist modulator regulates membrane trafficking of Kv4.2 potassium channel. J Biol Chem. 2007;282:1205–15. doi:10.1074/jbc.M607166200.PubMedCrossRefGoogle Scholar
- Sanz C, Mellström B, Link WA, Naranjo JR, Fernandez-Luna JL. Interleukin 3-dependent activation of DREAM is involved in transcriptional silencing of the apoptotic Hrk gene in hematopoietic progenitor cells. EMBO J. 2001;20:2286–92. doi:10.1093/emboj/20.9.2286.PubMedPubMedCentralCrossRefGoogle Scholar
- Scsucova S, Palacios D, Savignac M, Mellström B, Naranjo JR, Aranda A. The repressor DREAM acts as a transcriptional activator on Vitamin D and retinoic acid response. Elements. 2005;33:2269–79. doi:10.1093/nar/gki503.Google Scholar
- Tiruppathi C, Soni D, Wang DM, Xue J, Singh V, Thippegowda PB, Cheppudira BP, Mishra RK, Debroy A, Qian Z, Bachmaier K, Zhao YY, Christman JW, Vogel SM, Ma A, Malik AB. The transcription factor DREAM represses the deubiquitinase A20 and mediates inflammation. Nat Immunol. 2014;15:239–47. doi:10.1038/ni.2823.PubMedPubMedCentralCrossRefGoogle Scholar
- Zhang Y, Su P, Liang P, Liu T, Liu X, Liu XY, Zhang B, Han T, Zhu YB, Yin DM, Li J, Zhou Z, Wang KW, Wang Y. The DREAM protein negatively regulates the NMDA receptor through interaction with the NR1 subunit. J Neurosci. 2010;30:7575–86. doi:10.1523/JNEUROSCI.1312-10.2010.PubMedCrossRefGoogle Scholar