The discovery of the central regulatory network, the Rrp2-RpoN-RpoS pathway, was a significant advance in B. burgdorferi gene regulation. However, the dearth of knowledge regarding the mechanism underlying the activation of this pathway has been a major gap in our understanding of Borrelia host adaptation. In this study, we showed that temperature- and cell density-induced Rrp2-RpoN-RpoS activation occurs via a histidine kinase-independent mechanism. We further provided evidence suggesting the hypothesis that the high-energy metabolic intermediate acetyl~P plays a key role in Rrp2 phosphorylation and, consequently, the activation of the Rrp2-RpoN-RpoS pathway.
In this study we first extended the recent finding by Burtnick et al. 
that Hk2 was not essential for Rrp2 activation under in vitro
cultivation conditions, by further showing that the hk2
mutant was capable of activating the Rrp2-RpoN-RpoS pathway in a mammalian host-adapted model and establishing infection in mice. The fact that the hk2
mutant remained capable of upregulation of OspC and downregulation of OspA in the DMC model () indicates that this sensor kinase and its PAS sensing domain does not play a major in sensing mammalian host-specific signals for RpoS activation. We next tested the hypothesis that Hk1, the only other B. burgdorferi
histidine kinase with no assigned function, could be responsible for activation of the Rrp2 pathway. We found that the hk1
and hk1 hk2
mutants exhibited normal levels of temperature-induced Rrp2-dependent OspC expression. We further found that spirochetes lacking other histidine kinases identified in the B. burgdorferi
genome, the chemotaxis histidine kinases CheA1 or CheA2, also exhibited normal OspC expression. One caveat is that we have not tested cheA1 hk2
and cheA2 hk2
double mutants and thus cannot formally rule out a possible compensatory effect between Hk2 and CheA1 or CheA2.
Several groups have reported the existence of atypical response regulators in other bacteria whose activities do not require phosphorylation of their receiver domains 
. These atypical response regulators either do not possess the conserved aspartate residue shown to function as the phosphorylation site (e.g., HP1021 and HP1043 in Helicobacter pylori
, or lack conserved residues for Mg++
chelation, which is essential for phosphorylation (e.g., FrzS in Myxococcus
or NblR in Synechococcus
. However, Rrp2 retains all the conserved residues for phosphorylation (D52), Mg++
binding (D8, D9), and signal transduction (T80, F99, K102). Thus, it is unlikely that Rrp2 is an atypical response regulator. Indeed, in this study, we showed that Rrp2 can autophosphorylate using acetyl~P as its sole phosphoryl donor. Furthermore, overexpression of the phosphorylatable receiver domain of Rrp2 (Rrp2-N), but not variants of Rrp2-N that carry the D52A or D52E mutations, interfered with endogenous Rrp2 activity. This result is consistent with the assumption that Rrp2 activation requires phosphorylation of D52. Another evidence supporting phosphorylation-dependent Rrp2 activation is our previous observation that the ATPase activity of Rrp2, an activity that is essential for its transcriptional activation function, also is dependent on phosphorylation of Rrp2 
. Of note, overproduction of a protein from a strong constitutive promoter (e.g., flaB
) could have pleiotropic effects. An ideal approach to study the function of Rrp2 phosphorylation would be to replace the endogenous copy of rrp2
with the D52A mutant allele. Despite multiple efforts, however, we failed to generate the desired strain. This lack of success is consistent with previous reports that inactivation of rrp2
may be lethal 
. We hypothesize that phosphorylated Rrp2 may be important for cell growth. Consistent with this hypothesis, overexpression of Rrp2 exhibited a moderate growth defect (data not shown).
The finding that activation of RpoS and OspC requires phosphorylation of Rrp2 but does not require any of the four histidine kinases led us to hypothesize that the phosphoryl donor might be a high-energy central metabolic intermediate 
. Indeed, bioinformatic analysis of the B. burgdorferi
genome revealed one pathway capable of producing carbamoyl-P (ArcA-ArcB) and one pathway that can synthesize acetyl~P (Ack-Pta). Loss of ArcA, which should result in the inability to synthesize carbamoyl-P, had no effect upon Rrp2-dependent expression, suggesting that carbamoyl-P does not serve as the phosphoryl donor to Rrp2.
Acetyl~P is the intermediate of the Ack-Pta pathway. The Ack-Pta pathway functions in acetogenesis through the conversion of acetyl-CoA obtained from pyruvate into acetate; operation of this pathway in the opposite direction enables other bacteria to use acetate as a carbon source by activating acetate to acetyl-CoA, which subsequently enters the tricarboxylic acid (TCA) cycle. In some organisms, such as E. coli
, the pathway is reversible and thus can function in both acetogenesis and acetate activation 
. The relatively small genome of B. burgdorferi
, an obligate parasite, does not encode any enzyme known to convert pyruvate to acetyl-CoA, nor does it encode the enzymes of the TCA cycle. Instead, B. burgdorferi
performs lactogenesis, converting pyruvate to lactate 
(Xu H. and Yang, X.F., unpublished result). As such, the main function of the Ack-Pta pathway of B. burgdorferi
is likely not for converting acetyl-CoA to acetate, but for generating acetyl-CoA from acetate. This acetyl-CoA could then be used for cell wall synthesis (via
the mevalonate pathway [BB0683-BB0688]) and possibly for other metabolic pathways (). Furthermore, B. burgdorferi
seems to lack other acetyl-CoA synthetic pathways (e.g., the AMP-ACS pathway, β-oxidation of fatty acids, and several amino acid degradation pathways). Thus, the Ack-Pta pathway appears to be the sole pathway for biosynthesis of acetyl-CoA. If so, one would predict that the Ack-Pta pathway is essential for spirochetal growth. This notion is consistent with the fact that we failed to generate either an ack
or a pta
mutant by either targeted mutagenesis or random transposon mutagenesis (data not shown). What's the source of acetate for B. burgdorferi
? Our measurement showed that acetate concentration in mouse blood and the midgut of fed ticks is ~1.0 M and ~1.8 mM, respectively (Xu H. and Yang, XF, unpublished data). One of the ingredients of the BSK-H medium, CMRL, also contains 0.61 mM acetate (other ingredients of this complex medium, such as rabbit serum, also may contribute to the overall levels of acetate). Through diffusion or an unknown transport system, B. burgdorferi
may obtain sufficient acetate from these environments for acetyl-CoA production.
Acetyl~P has drawn attention as a global regulator of gene expression via its ability to donate its phosphoryl group to a subset of response regulators under certain environmental conditions 
. In E. coli
, the intracellular acetyl~P concentration can reach levels sufficient to phosphorylate a subset of response regulators 
and thus influence the biological processes controlled by those proteins 
. Although we have not yet measured the intracellular acetyl~P levels to determine if this is also the case in B. burgdorferi
, we were able to provide three lines of evidence to support the conclusion that acetyl~P plays an important role in Rrp2 activation: (i) the activation of the Rrp2-RpoN-RpoS pathway can be induced by increasing concentration of exogenous acetate (); (ii) overexpression of Pta reduced acetate-induced activation of the Rrp2-RpoN-RpoS pathway (); and (iii) acetyl~P served as a phosphoryl donor to Rrp2 in vitro
(). Note that overexpression of Pta did not completely abolish OspC production, suggesting that a low level of Rrp2 activation still occurs. This might be due to the presence of low levels of acetyl~P, as overexpression of Pta does not abolish the production of acetyl~P. Alternatively, Hk2 may contribute to Rrp2 activation. We are currently in the process of testing this possibility by overexpressing Pta in the hk2
mutant. Nevertheless, this partial inhibition of RpoS and OspC expression by overexpression of Pta is consistent with the in vivo
phenotype that overexpression of Pta resulted in a moderate reduction of spirochetal infectivity in mice ().
It is well established that the Rrp2 pathway can be activated by many environmental cues such as temperature, pH, cell density, oxygen, and CO2
. However, the underlying mechanism for these phenomena has not been elucidated. In this regard, it is striking that virtually all the environmental cues that activate the Rrp2 pathway also have been shown to influence the acetyl~P pool in E. coli 
. This observation is consistent with our hypothesis that acetyl~P serves as a signaling molecule that responds to environmental cues and in response activates the Rrp2 pathway. Indeed, we showed that overexpression of pta
greatly inhibited both temperature- and cell density-induced activation of Rrp2 (), suggesting that elevated temperature and increased cell density activate the Rrp2-RpoN-RpoS pathway in an acetyl~P-dependent manner. Elevated temperature may increase acetyl~P levels by enhancing diffusion of acetate into the cells and/or from increased transport efficiency via an unidentified transport system for acetate. Elevated temperature also increases cell growth rates that likely lead to increased levels of acetyl~P 
. The effect of increased cell density on acetyl~P levels, on the other hand, can result simply by a change in extracellular pH. As cell density increases, the culture pH diminishes from 7.5 to 7.0 or lower 
, which favors the passive diffusion of acetate into the cells 
One caveat of this study is that we used expression of RpoS and OspC as the readout for Rrp2 phosphorylation. An ideal approach for such study would be directly to detect the phosphorylated form of Rrp2. Unfortunately this approach is not technically feasible since most forms of the Asp-phosphorylation are unstable and there is no antibody available for detecting Asp-phosphorylation. Thus, a common approach for studying phosphorylation of response regulators is to monitor the output product as a result of phosphorylation of a response regulator. In the case of Rrp2, the only direct target gene identified thus far is rpoS
and therefore, expression of rpoS
faithfully reflects the activation of Rrp2 modulated by phosphorylation. One concern for this approach is whether the effect on RpoS expression observed in this study is through another transcriptional activator, BB647 (BosR). BB647 is a fur homologue and was recently shown that inactivation of this gene significantly reduced rpoS
. Although it remains unclear how BosR fits into the Rrp2-RpoN-RpoS pathway, we found that neither overexpression of Rrp2-N nor overexpression of Pta affected the level of BosR (data not shown), suggesting that the effects of Rrp2-N or Pta overexpression on RpoS and OspC was not through BosR, rather through Rrp2.
In summary, we have shown that temperature- and cell density-induced the activation of the Rrp2-RpoN-RpoS pathway proceeds independently of histidine kinases and carbamoyl-P. In contrast, biochemical and genetic manipulation of the acetyl~P-producing Ack-Pta pathway dramatically impacts activation of the Rrp2-RpoN-RpoS pathway, providing strong evidence that acetyl~P plays an important role in Rrp2 activation under in vitro
growth conditions. We also provide evidence showing that, during mammalian infection, the Rrp2-RpoN-RpoS pathway is also activated via an Hk2-independent mechanism and that acetyl~P plays an important role in this process. Then, what is the function of Hk2? One possibility is that Hk2 may play a role in sensing host signals and activating Rrp2 during the process of tick feeding. In this regard, we have examined the phenotype of the hk2
mutant in ticks and found that the hk2
mutant indeed has reduced infectivity via the route of tick infestation. Unfortunately, we have not been able to construct an infectious complemented strain and, thus, have been unable to show restoration of this defect, which prevents us from drawing a definitive conclusion on Hk2 function in the enzootic cycle of B. burgdorferi
. Nevertheless, this preliminary finding suggests that Hk2 may contribute to Rrp2 activation during the process of tick feeding. In addition, spirochetes likely have increased levels of intracellular acetyl~P in feeding ticks, as they encounter increased temperature 
, as well as a massive influx of nutrients that leads to a dramatic increase of growth rates during this process 
. Thus, we postulate that while acetyl~P plays an important in activating the Rrp2-RpoN-RpoS pathway during mammalian infection, both acetyl~P and Hk2 are likely involved in integrating complex environmental and host signals to modulate the Rrp2-RpoN-RpoS pathway during the process of spirochetal transmission from ticks to mammals.