The sterol regulatory element binding protein (SREBP) family of transcription activators are critical regulators of cholesterol and fatty acid homeostasis. We previously demonstrated that human SREBPs bind the CREB-binding protein (CBP)/p300 acetyltransferase KIX domain and recruit activator-recruited co-factor (ARC)/Mediator co-activator complexes through unknown mechanisms. Here we show that SREBPs use the evolutionarily conserved ARC105 (also called MED15) subunit to activate target genes. Structural analysis of the SREBP-binding domain in ARC105 by NMR revealed a three-helix bundle with marked similarity to the CBP/p300 KIX domain. In contrast to SREBPs, the CREB and c-Myb activators do not bind the ARC105 KIX domain, although they interact with the CBP KIX domain, revealing a surprising specificity among structurally related activator-binding domains. The Caenorhabditis elegans SREBP homologue SBP-1 promotes fatty acid homeostasis by regulating the expression of lipogenic enzymes. We found that, like SBP-1, the C. elegans ARC105 homologue MDT-15 is required for fatty acid homeostasis, and show that both SBP-1 and MDT-15 control transcription of genes governing desaturation of stearic acid to oleic acid. Notably, dietary addition of oleic acid significantly rescued various defects of nematodes targeted with RNA interference against sbp-1 and mdt-15, including impaired intestinal fat storage, infertility, decreased size and slow locomotion, suggesting that regulation of oleic acid levels represents a physiologically critical function of SBP-1 and MDT-15. Taken together, our findings demonstrate that ARC105 is a key effector of SREBP-dependent gene regulation and control of lipid homeostasis in metazoans.

SYNOPSISThe nature of the chromophore binding site of light-adapted bacteriorhodopsin is analyzed by using all-valence electron MNDO and MNDO-PSDCI molecular orbital theory to interpret previously reported linear and nonlinear optical spectroscopic measurements. A total of 45 binding site models are investigated. The binding site is simulated by including the chromophore, the lysine residue (LYSZl6), the following nearby amino acids ASP,,, ASPIIS, ASPZIZ, THRgo, TRPa6, T R P I~s , TRPlsz, TYR57, TYRs3, and TYRIs5) and zero, one, or two divalent cations. We conclude that the unique two-photon properties of the chromophore are due in part to the electrostatic field associated with a Ca'+ ion near to the chromophore. Four amino acids and three water molecules contribute significantly to the assigned chromophore adjacent calcium binding site (ASPs5, ASP,,,, TYRS7 and TYR,,,), and two conformational minima are predicted. The higher energy conformation has the calcium ion stabilized primarily by ASP,, and the chromophore imine proton by ASP,,,. The lower energy conformation has the calcium ion stabilized primarily by ASPzl2 and the imine proton by ASPs5. The latter configuration is more stable due to strong hydrogen bonding between TYRIs5 and ASP212 coupled with electrostatic stabilization of the divalent cation by TYRS7. Although both tyrosine residues are predicted to exhibit some "unprotonated' character, models involving full deprotonation of either TYR57 or TYRls5 do not fit the spectroscopic data. We conclude that the cation binding site identified in this study is the second high affinity binding site for calcium, and that the chromophore binding site is, to a first approximation, positively charged.The chromophore "lh*t" and "I$*-" states, despite extensive mixing, exhibit significantly different configurational character. The lowest-lying '"&*+" state is dominated by single excitations (> 80% for all models studied) whereas the second-excited "lAg*-'' state is dominated by double excitations (> 70% for all models studied with extensive participation by spin-coupled triplet-triplet excitations).

No abstract

Two short-wavelength cone opsins, frog (Xenopus laevis) violet and mouse UV, were expressed in mammalian COS1 cells, purified in delipidated form, and studied using cryogenic UV-vis spectrophotometry. At room temperature, the X. laevis violet opsin has an absorption maximum at 426 nm when generated with 11-cis-retinal and an absorption maximum of 415 nm when generated with 9-cis-retinal. The frog short-wavelength opsin has two different batho intermediates, one stable at 30 K (lambda(max) approximately 446 nm) and the other at 70 K (lambda(max) approximately 475 nm). Chloride ions do not affect the absorption maximum of the violet opsin. At room temperature, mouse UV opsin has an absorption maximum of 357 nm, while at 70 K, the pigment exhibits a bathochromic shift to 403 nm with distinct vibronic structure and a strong secondary vibronic band at 380 nm. We have observed linear relationships when analyzing the energy difference between the initial and bathochromic intermediates and the normalized difference spectra of the batho-shifted intermediates of rod and cone opsins. We conclude that the binding sites of these pigments change from red to green to violet via systematic shifts in the position of the primary counterion relative to the protonated Schiff base. The mouse UV cone opsin does not fit this trend, and we conclude that wavelength selection in this pigment must operate via a different molecular mechanism. We discuss the possibility that the mouse UV chromophore is initially unprotonated.