Transient I-DIRT is a method for unambiguous identification of stable and transient protein–protein interactions. One culture of isotopically light cells containing an affinity-tagged protein is subjected to in vivo chemical cross-linking with formaldehyde (FA), as is a second culture of isotopically heavy cells that does not contain the tagged protein. These cells are frozen independently, subsequently mixed in a ratio of 1:1 and co-lysed under cryogenic conditions. During the affinity purification, proteins that exist in stable interactions with the tagged protein are isotopically light, while contaminants are a 1:1 mixture of light and heavy. Proteins having a transient interaction with the tagged protein complex show an intermediate level of isotopically light peptides.

Mild formaldehyde cross-linking is optimal for the liberation and affinity purification of chromatin-associated protein complexes. YNG1-TAP cells were subjected to various amounts of in vivo formaldehyde (FA) cross-linking. A: Cellular lysates were sonicated to shear DNA. Purified DNA was resolved by agarose gel electrophoresis and visualized with ethidium bromide. B: Sonicated cell lysates were clarified by centrifugation and TAP tagged Yng1 was affinity purified (AP) on IgG-coated Dynabeads. Aliquots were collected during the indicated steps of the purification and the relative amount of Yng1-TAP was visualized by western blotting.

In vivo chemical cross-linking of NuA3 traps novel transiently interacting proteins. YNG1-TAP cells (±0.05% formaldehyde) were collected for affinity purification. A: Cell lysates were subjected to sonication to shear DNA. Isolated DNA was resolved by agarose gel electrophoresis and visualized with ethidium bromide. B: Yng1-TAP and associated proteins were purified on IgG-coated Dynabeads, resolved by SDS-PAGE, visualized by Coomassie staining and excised for protein identification. Proteins were digested with trypsin for peptide analysis by LC-MS² (Thermo LTQ-XL). Proteins identified (99% confidence) were categorized into a Venn diagram. The core NuA3 components (Sas3, Nto1, Yng1, Taf30, Eaf6) were identified in both purifications, while yFACT/Nap1/Rsc were trapped with chemical cross-linking. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Transient I-DIRT technology differentiates specific, transient and nonspecific protein–protein interactions with the NuA3 histone K-acetyltransferase. A: An arginine auxotrophic strain showed complete incorporation of ¹³C₆-arginine. Proteins from an arg4Δ strain grown either in isotopically light or heavy synthetic yeast medium were resolved by SDS-PAGE. Identical bands containing pyruvate decarboxylase PDC1 were excised and treated with trypsin. Peptides were analyzed by MALDI mass spectrometry. Shown is a representative peptide that demonstrates the complete incorporation of heavy arginine. B: Purification of in vivo cross-linked Yng1-TAP from a 1:1 mixture of YNG1-TAP isotopically light and arg4Δ isotopically heavy cells. Proteins co-purifying with Yng1-TAP were resolved by SDS-PAGE, visualized by Coomassie staining and excised for protein identification. Gel bands were treated with trypsin and peptides were analyzed with a Thermo LTQ Orbitrap mass spectrometer. C: Mass spectra of peptides from proteins that are specific, transient or nonspecific interactors with the Yng1-TAP containing protein complex.

NuA3 transiently interacts with the yFACT and RSC complexes in vivo. A: Arginine-containing peptides were identified for 288 of the proteins identified from the gel shown in Figure 4(B). The fraction of isotopically light peptide relative to light + heavy is plotted for each. B: Condensed version of panel A. Core NuA3 components (Sas3, Nto1, Yng1, Taf14, Eaf6) were identified as ∼100% light. A nonspecific threshold (0.47 ± 0.08) was established with a group of ribosomal proteins, which are common contaminants in affinity purifications. Relative to the nonspecific threshold, the majority (278) of the proteins were classified as nonspecific. Five proteins showed a fraction of isotopically light that were >2 standard deviations from the contaminant threshold and were categorized as transient interactions: Pob3/Spt16 (yFACT components), Nap1 (nucleosome assembly protein 1) and Rsc7/8 (RSC chromatin remodeling complex components).

NuA3 and yFACT have overlapping in vivo functions. A: Indicated strains were 10-fold serially diluted and assayed for temperature, hydroxyurea (HU) or methyl methanesulfonate (MMS) sensitivity. B: Proposed model of NuA3 in transcriptional activation/elongation. Set1-modified H3K4me3 recruits NuA3 for acetylation of H3K14. RSC is localized to the H3K14Ac and serves as a bridge to RNA polymerase II. yFACT is localized via protein-protein interactions. Nap1 shuttles in histones to this macromolecular protein assembly. The chromatin remodeling activity of RSC and nucleosome disassembly activity of yFACT destabilize a nucleosome to provide for RNA polymerase II movement. yFACT replaces the nucleosome with histones delivered by Nap1.