We performed a series of in vitro protein cross-linking and footprinting experiments to find the domains of the RNAP complex to which SutA binds. To determine the regions of the RNAP complex with which SutA interacts, we undertook two in vitro cross-linking approaches: an unbiased cross-linking approach using bis(sulfosuccinimidyl) suberate (BS3), and a targeted approach that relies on the site-selective approach. uptake of the UV-activated amino acid cross-linker -benzoylphenylalanine (Bpa) (Figure 3.1A). Gel intensity analysis revealed that the addition of SutA to the cross-linking reaction resulted in preferential formation of the smaller product (70% with SutA vs. 35% without) (Figure 3.S2A).
The three cross-links between SutA and RNAP were between Lys62 or Lys69 of SutA and Lys116 of the subunit and between Lys95 of SutA and Lys40 of the ' subunit. No mutants cross-linked to the ↵ subunit (Figure 3.S4B), and the other SutA mutants tested showed minimal evidence of cross-linking to any RNAP subunit. Based on gel analysis, cross-linked products were qualitatively the same in the presence (Figure 3.1C) or absence (Figure 3.S4C) of the nucleic acids.
We identified eight residues whose modification was reduced at least 1.5-fold in the presence of SutA (Figure 3.S6). We also found two residues with higher intensity modified in the presence of SutA: Lys207 of the subunit along the major channel and Lys603 of the subunit ' at the face of the secondary channel. The two BS3 cross-links to lobe 1 coincide with two obscured residues detected in the footpress experiment: Lys116 and Lys116 (Figure 3.2B, blue).
However, Bpa cross-links between positions in the N-terminal part of SutA (Leu6, Leu11 and Leu22) place these residues against the 'clamp', just outside the main channel (Figure 3.2, green).
Discussion
However, our observation that lysine residues within the RNAP main channel are less accessible to chemical modification in the presence of SutA is consistent with clamp closure. Identification of Bpa cross-links between residues within the N-terminal domain of SutA to the. We interpret the partial crosslinking of Leu6, Leu11, Leu22 positions to both and ' subunits to reflect the mobility of the presumably unstructured N-terminal acidic domain.
In addition to rotation of the clamp domain, other movements of large, mobile RNAP domains in the E. Several of the lysines for which the accessibility to chemical modification changes in the presence of SutA (' Lys996 and Lys1231, which show reduced modification in the presence of SutA; and Lys207 and ' Lys603, which show increased modification) are near or part of the ' i6 domain, suggesting that this domain may occupy a different position in the presence of SutA. This domain does not appear in any crystal structures, probably because of its mobility, so alternative conformations are not well characterized.
However, it has been suggested that the position of the 'i6 domain may influence the ability of DksA to act on RNAP [23]. If SutA preferentially interacts with RNAP in a closed clamp conformation, this could contribute to the opposite effects on gene expression and phenotype observed between dksA and sutA mutants. Future efforts to obtain additional structural information of SutA and to define its effects on RNAPin in vitro will help distinguish between these possibilities.
A model of SutA as a protein that can change the clamp position of RNAP is reminiscent of what was observed for the elongation factor NusG [25], although their exact points of interaction are probably different. There are several other distinct differences between the roles and activities of NusG and the potential roles for SutA. NusG is essential and is critical for facilitating interactions between RNAP and other complexes [26], whereas SutA is non-essential, even under the conditions in
Nevertheless, our data suggest that SutA enhances expression of the genes to which it is recruited and one possibility is that, like NusG, it does so primarily by enhancing transcription elongation at protein-coding genes (e.g., rProtein genes). In contrast, for rRNA genes, SutA associates primarily with the promoter region, suggesting that its activity may be different for these genes and, like DksA, may play a primary role in regulating initiation. More detailed structural resolution and measurements of the direct effects of SutA on RNAP activity will also be necessary to adequately test the proposed model.
Future Work
Effects of SutA on RNAP behavior, such as elongation rate or premature termination, can also be examined at these loci.
Experimental Procedures
The Orbitrap was operated in data-dependent acquisition mode to automatically switch between a full scan (m/z=400–1600) in the Orbitrap and subsequent 10 CID MS/MS scans in the linear ion trap. The Orbitrap was operated in data-dependent acquisition mode to automatically alternate between a full scan (m/z=300-1600) in the Orbitrap and subsequent 5 HCD MS/MS scans in the Orbitrap. Raw files were converted to peak lists with Prote-oWizard [ 27 ] and analyzed with Protein Prospector online, version 5.12.4 according to reported protocols with modifications below [ 30 ].
The analysis was performed twice for each set of peak lists to search for both isotopologues of the cross-linker. Raw files were independently screened using MaxQuant for pairs of precursor masses differing by 4.02 Da, representing cross-links made by both linker isotopologues. Crosslinks detected by Protein Prospector were compared to a list of mass pairs to remove crosslinks not present as 4.02 Da aligned mass pairs.
We used this calculated distance as a metric to distinguish "quality" correlations from all others. Based on linker length, the maximum inter↵-carbon distance between lysines linked by BS3is24.6Å, so we considered crosslinks with distances close to or below this value reasonable. A score difference cutoff of 5.6 (similar to the value of 8.5 found by Trnka et al.) separated high- and low-distance crosslinks (Figure 3.S2C) giving an FDR of < 0.05 (see Figure 3. S2B for ROC Curves for this classification model).
The final criteria for assigning quality crosslinks were: (i) found as a precursor mass pair, (ii) Score difference greater than 5.6, and (iii) matched by at least two spectra. These crosslinks were pooled to determine the number of spectra from each replicate and the maximum score difference for each amino acid linkage (Table 3.1). The best spectra used to identify each cross-link between SutA and RNAP are shown in Figure 3.S3.
Raw files were searched using MaxQuant against a protein database containing sequences for SutA, RpoB and RpoC and a. For quantification, the raw files and the list of identified peptides were imported into Skyline version 3.1 and subset for high-quality peak matches between all runs (isotopic product score >. For each replicate, peptide intensity ratios for each peptide ion between SutA and control samples were calculated.
To account for variations in LC-MS/MS loading, all peptide intensity ratios for each experiment were normalized so that the median ratio was 1. When evidence for multiple cross-links between the same peptides was found, all cross-link locations are shown. Lysine residues determined to be obscured (top) or revealed (bottom) in the presence of SutA.
Eight detected intra-RNAP cross-links are shown; the ninth is located on the opposite side of the structure.
