Semi-supervised drug-protein interaction prediction from heterogeneous biological spaces

Background

Predicting drug-protein interactions from heterogeneous biological data sources is a key step for in silico drug discovery. The difficulty of this prediction task lies in the rarity of known drug-protein interactions and myriad unknown interactions to be predicted. To meet this challenge, a manifold regularization semi-supervised learning method is presented to tackle this issue by using labeled and unlabeled information which often generates better results than using the labeled data alone. Furthermore, our semi-supervised learning method integrates known drug-protein interaction network information as well as chemical structure and genomic sequence data.

Results

Using the proposed method, we predicted certain drug-protein interactions on the enzyme, ion channel, GPCRs, and nuclear receptor data sets. Some of them are confirmed by the latest publicly available drug targets databases such as KEGG.

Conclusions

We report encouraging results of using our method for drug-protein interaction network reconstruction which may shed light on the molecular interaction inference and new uses of marketed drugs.

     

Nuclear binding of progesterone in hen oviduct. Role of acidic chromatin proteins in high-affinity binding.

The multiple classes of binding sites for the progesterone-receptor complex in hen oviduct muclei were found to be of chromatin origin. The highest-affinity, and presumably most physiologically important class, is localized in oviduct chromatin and contains approx. 6000-10000 sites per nucleus. None of these sites is detected in spleen chromatin. Two new techniques were used for assaying rapidly the binding of steroid-receptor complexes to soluble deoxyribonucleoproteins in vito. The extent of high-affinity binding by the nucleo-acidic protein fraction from spleen chromatin is as great as that by the nucleo-acidic protein from oviduct chromatin. Consequently the tissue-specific nuclear binding of the progesterone receptor is found not to be a consequence of the absence of the nuclear binding sites (acceptors) from chromatin of non-target tissue (spleen), but rather a result of complete masking of these sites. In the target-tissue (oviduct) chromatin, approx. 70% of the high-affinity acceptor sites are also masked. Acidic proteins, and not histones, appear to be responsible for the masking of these acceptor sites. In addition, acidic proteins represent (or at least are an essential part of) these high-affinity sites in the oviduct nucleus. Pure DNA displays a few high-and many low-affinity binding sites. In support of previous work with immature chicks, the acidic protein fraction of the nucleo-acidic results thus support the hypotheis that protein complexed with DNA, and not DNA alone, represent the high-affinity binding sites for the steroid-receptor complexes in nuclear chromatin. The lower-affinity classes of binding sites may represent DNA and/or other nuclear components.     

Partial purification and preparation of polyclonal antibodies against candidate chromatin acceptor proteins for the avian oviduct progesterone receptor

Steroid hormones bind to specific receptors in target cells that in turn bind to chromatin acceptor sites to alter gene expression. These chromatin acceptor sites, for a variety of steroid receptors, appear to be composed of acceptor proteins tightly bound to the DNA. This paper describes the preparation of new polyclonal antibodies against the chromatin acceptor proteins of the avian oviduct progesterone receptor (PR) and their use in monitoring the purification of the acceptor proteins. This laboratory recently reported the preparation of monoclonal antibodies that do recognize the intact chromatin acceptor sites containing DNA-bound acceptor proteins but not the unbound acceptor protein for PR [Goldberger, A., Horton, M., Katzmann, J., & Spelsberg, T. C. (1987) Biochemistry 26, 5811-5816]. In order to obtain antibodies that recognize the unbound acceptor protein, polyclonal antibodies were prepared against a highly purified preparation of the acceptor protein(s). Analyses by ELISA indicate that the polyclonal antibodies recognize both the intact acceptor sites and the unbound (free) acceptor protein(s). Using these antibodies in Western immunoblots, two antigenic species of 10 and 6 kDa were detected in crude fractions of acceptor protein. These two protein species could be separated and further enriched while still retaining acceptor activity, i.e., the capacity to generate specific binding of the PR. Thus, the antigenic activity is closely associated with, if not identical with, the acceptor activity. Whether one or both species are used in vivo or whether the 6-kDa species is a proteolytic product of the 10-kDa species is unknown.(ABSTRACT TRUNCATED AT 250 WORDS)     

Loss of ERE binding activity by estrogen receptor-alpha alters basal and estrogen-stimulated bone-related gene expression by osteoblastic cells

Estrogen receptor (ER)-alpha can signal either via estrogen response element (ERE)-mediated pathways or via alternate pathways involving protein-protein or membrane signaling. We previously demonstrated that, as compared to wild type (WT) controls, mice expressing a mutant ER-alpha lacking the ability to bind EREs (non-classical estrogen receptor knock-in (NERKI)) display significant impairments in the skeletal response to estrogen. To elucidate the mechanism(s) underlying these in vivo deficits, we generated U2OS cells stably expressing either WT ER-alpha or the NERKI receptor. Compared to cells transfected with the control vector, stable expression of ER-alpha, even in the absence of E2, resulted in an increase in mRNA levels for alkaline phosphatase (AP, by 400%, P < 0.01) and a decrease in mRNA levels for insulin growth factor-I (IGF-I) (by 65%, P < 0.001), with no effects on collagen I (col I) or osteocalcin (OCN) mRNA levels. By contrast, stable expression of the NERKI receptor resulted in the suppression of mRNA levels for AP, col I, OCN, and IGF-I (by 62, 89, 60, and 70%, P < 0.001). While E2 increased mRNA levels of AP, OCN, col I, and IGF-I in ER-alpha cells, E2 effects in the NERKI cells on AP and OCN mRNA levels were attenuated, with a trend for E2 to inhibit col I mRNA levels. In addition, E2 had no effects on IGF-I mRNA levels in NERKI cells. Collectively, these findings indicate that ERE signaling plays a significant role in mediating effects of estrogen on osteoblastic differentiation markers and on IGF-I mRNA levels.