All chemicals were obtained from Sigma Aldrich (St. Louis, MO), unless otherwise specified. Recovery, activity, and fluorescence assays
Concentrations of all purified proteins were determined by 280 nm absorbance using extinction coefficients predicted by ExPASy (www.expasy.org). All extinction coefficients are provided in Supplementary Table S2. Recovery of MBP-BRT17 by either amylose resin purification or precipitation was determined solely using this method.
MBP-BRT17-GFP recoveries were estimated by comparing fluorescence emission intensity at 509 nm with excitation at 487 nm. 100-fold dilutions of both clarified lysate and purified protein were made for fluorescence measurements. Purified proteins were resuspended in the same volume as the lysate from which they were extracted, so signals were compared directly.
For estimation of MBP-BRT17-βlac recovery, protein was added to a nitrocefin solution and the absorbance at 486 nm was tracked corresponding to the hydrolysis of nitrocefin. 500 µL of nitrocefin solution was prepared by placing three nitrocefin disks (Fluka) in 450 µL 50 mM tris-HCl, pH 7.4 and 50 µL DMSO. In each sample well, 50 µL of this solution was mixed with 90 µL of the same tris buffer and 10 µL of protein sample. For each sample tested, serial dilutions from 1X to 1000X were prepared from lysate and purified protein. Initial rates were determined by measuring the change in absorbance at 486 nm over the first 20% of the change in signal between the starting absorbance and the end absorbance. The same nitrocefin stock solution was used for all samples to account for variations in concentration.
MBP-BRT17-AdhD recovery was also evaluated by enzymatic activity using a protocol previously described (28). Since this AdhD was isolated from the hyperthermofile Pyrococcus furiosus, all samples were heat treated at 80°C for 1 h prior to evaluating activity. All assays were performed at saturated conditions of both cofactor and substrate, 0.5 mM NAD+ and 100 mM 2,3-butanediol, respectively. Reaction mixtures containing 2,3-butanediol and protein sample in 50 mM glycine pH 8.8 were incubated at 45°C in a 96 well UV microplate in a spectrophotometer. Reactions were initiated by the addition of NAD+. Initial rates were calculated by following the production of NADH at 340 nm. Specific activity of cleaved AdhD was calculated using an NADH extinction coefficient (∊ = 6.22 mM−1 cm−1).
All spectroscopic measurements were done on a SpectraMax M2 (Molecular Devices; Sunnyvale, CA). Results and discussion
In order to identify the consensus RTX sequence, a database of RTX containing proteins was constructed by searching the UniProt (www.uniprot.org) database for hemolysin-type calcium binding domains. Individual repeats were identified and the frequency of amino acids at each of the nine repeat positions was determined (Figure 1B). From this result, the repeat sequence GGAGNDTLY was identified as the consensus sequence. For a few of these positions, other amino acids were found with nearly equal frequency. However, as this sequence was found to be effective for purification, further investigation on sequence variation was not performed. A variety of synthetic RTX domains of different lengths (BRT5, BRT9, BRT13, BRT17) were prepared as fusions to the C terminus of MBP, with subscripts denoting the number of repeats. These lengths were chosen as they reflect the variability of naturally occurring RTX domains. Unexpectedly, we observed that upon the addition of calcium to the purified BRT17 construct, there was significant precipitation out of solution, which was reversed upon the addition of the chelating agent EGTA.
In order to more thoroughly characterize the observed precipitation behavior, cells were induced to express the four MBP-BRT constructs. Clarified cell lysates wereprepared from these four cultures and then titrated with calcium to assess precipitation behavior by mixing with CaCl2 solution at the indicated concentrations, followed by by centrifugation, and measurement of the mass of the pellet (Figure 2). Due to possible variations in cell growth rates and densities, all cultures were started from saturated overnight cultures and induced simultaneously. Both BRT13 and BRT17 precipitated when calcium concentrations exceeded 25 mM. Some precipitation was observed from BRT5 and BRT9 lysate, similar to what was observed with control cell lysate. Addition of an equivalent concentration of EGTA allowed the pellets to quickly dissolve again upon gentle pipetting.
While both BRT13 and BRT17 precipitated upon calcium addition, BRT17 formed a pellet that was easier to clarify and was therefore selected for further examination. Three additional constructs were prepared by fusing MBP-BRT17 to the N terminus of GFP, β-lactamase, and AdhD (named MBP-BRT17-GFP and MBP-BRT17-βlac, MBP-BRT17-AdhD respectively). These three proteins were fused to MBP to allow for amylose resin chromatography purification as a comparison technique. GFP was chosen as a reporter protein for initial purification experiments to track the location of the BRT. β-lactamase and AdhD were chosen as they are well characterized enzymes whose activity can be measured with straightforward assays.
The folding of RTX domains into β rolls is highly calcium specific. Therefore, we were interested in whether or not the precipitation behavior observed was also calcium-specific. To test this, MBP-BRT17-GFP was purified on an amylose resin and diafiltered into salt-free tris buffer. Diafiltration was necessary as proteins are purified in high salt buffer for the amylose resin step and it was observed that BRT precipitation was significantly reduced in high salt. This is consistent with previous observations that RTX calcium affinity is reduced with increasing salt concentration (29). Solutions of various salts were added to final concentrations of 100 mM. The samples were then gently mixed by pipetting, allowed to sit for 2 min, and centrifuged at 16,000 g in a microcentrifuge for 2 min. Tubes were then inverted and the presence of a pellet at the top was indicative of precipitation (Figure 3). BRT precipitation was observed to be calcium-specific, with near complete precipitation of MBP-BRT17-GFP (as indicated by the remaining color in solution) in the presence of calcium and no precipitation with other salts. While this behavior does not establish the formation of a β roll structure, it does indicate that at least one property of the β roll is preserved in these constructs.