Functional Characterization of Knock-In Mutants 
We have previously shown that antigen-specific IFN-gamma production of splenocytes is a reliable readout system to evaluate whether or not ESAT-6 was secreted by recombinant strains [23,24,26].
Thus, in order to determine the reason for the observed failure of H37Ra to induce ESAT-6-specific responses, H37Rv and H37Ra strains were tested for their potential to express and secrete ESAT-6 in vitro.
The western blot analysis presented in Figure 5 shows that the cell lysates of H37Ra and H37Ra::rpsL contain large amounts of ESAT-6, indicating that the antigen is properly expressed.
However, hardly any ESAT-6 was present in the culture filtrates of these cultures, indicating that these strains were unable to secrete ESAT-6 under the in vitro conditions employed in spite of proper expression.
Similar results were obtained for H37Ra::fadE5 (unpublished data).
In contrast, from examination of the western blots of M. tuberculosis H37Ra::phoP and H37Rv it is clear that a large portion of their ESAT-6 protein is secreted into the culture filtrates.
Analysis of the M. tuberculosis MT103 phoP ko strain SO2 showed strong expression of ESAT-6, but only very little secreted ESAT-6 (Figure 5).
In a previous report [25], the presence of ESAT-6 in cell free extracts was described for this strain, but no secretion assay was performed.
Together, our findings correlate perfectly with the in vivo data described above and suggest that the lack of ESAT-6 specific T cell recognition for H37Ra, H37Ra::fadE5, H37Ra::rpsL, and SO2 is not caused by a loss of ESAT-6 expression, but rather due to a failure of (phoP dependent) ESAT-6 secretion.
The observation that knocking-in the ESX-1 region of H37Rv into H37Ra (H37Ra::RD1-ppe68-ko), and the resultant diploidy, did not improve the ESAT-6-specific T cell responses (Figure 3B), further argues that PhoP-dependent ESAT-secretion might be regulated via a mechanism that lies outside the RD1 region.
As closer inspection of the available literature on transcriptional analyses of a H37Rv-phoP ko strain [21] and H37Ra [12] suggested that the expression of gene cluster rv3612c-rv3616c is reduced in both strains, we evaluated the expression level of rv3614c in H37Rv, H37Ra, and H37Ra::phoP by quantitative real time PCR (qRT-PCR).
In previous studies rv3614c, rv3615c, rv3616c (espA) were independently shown to be essential for proper ESAT-6 secretion [27,28].
We confirmed the trend that rv3614c was expressed much lower in H37Ra compared to H37Rv by qRT-PCR (Figure 6), whereas expression of rv3614c in H37Ra::phoP was restored to wild-type levels, suggesting that the rv3612c-rv3616c gene cluster is regulated by PhoP.
Interestingly, expression of phoP was higher in H37Ra than in H37Rv (Figure 6).
These findings are consistent with data from a previous transcriptional study [12] and suggest that the proposed ability of PhoP to downregulate its own expression [22] is affected by the S219L phoP mutation [29].
They also indicate that it is not a lack of phoP expression that is causing the phoP-associated effects in H37Ra.
