Adaptation to Glucosamine Starvation in Borrelia burgdorferi is Mediated by recA

Lyme borreliosis is a vector-borne disease confined primarily to the Northern hemisphere and caused by spirochetes belonging to the Borrelia burgdorferi sensu lato (s.l.) genospecies (Kurtenbach et al. 2006). The spirochetes are maintained exclusively within an enzootic cycle, alternating between a tick vector and vertebrate host. B. burgdorferi s.l. includes three pathogenic species and at least eight related species that are non-pathogenic or rarely cause infection in humans. The three pathogenic species include B. burgdorferi sensu stricto (s.s.), the only species found in North America, and B. garinii and B. afzelii the causative agents of borreliosis in Europe and Asia (Steere et al. 2004). These three species cause similar acute and chronic illness in humans, although regional differences do exist (Steere 2001).


Introduction
Lyme borreliosis is a vector-borne disease confined primarily to the Northern hemisphere and caused by spirochetes belonging to the Borrelia burgdorferi sensu lato (s.l.) genospecies (Kurtenbach et al. 2006).The spirochetes are maintained exclusively within an enzootic cycle, alternating between a tick vector and vertebrate host.B. burgdorferi s.l.includes three pathogenic species and at least eight related species that are non-pathogenic or rarely cause infection in humans.The three pathogenic species include B. burgdorferi sensu stricto (s.s.), the only species found in North America, and B. garinii and B. afzelii the causative agents of borreliosis in Europe and Asia (Steere et al. 2004).These three species cause similar acute and chronic illness in humans, although regional differences do exist (Steere 2001).
B. burgdorferi is a limited genome organism and must scavenge most essential nutrients from its tick vector or vertebrate host (Das 2000;Saier MH 2000).The organism requires a complex medium for propagation in the laboratory, and specific addition of free N-acetylglucosamine (GlcNAc), a component of the bacterial cell wall, is necessary for cells to reach optimal cell densities in a single exponential phase (Barbour 1984;Tilly et al. 2001;Rhodes et al. 2009;Rhodes et al. 2010).In the absence of free GlcNAc B. burgdorferi can degrade chitin (a polymer of GlcNAc) to chitobiose and import this dimer through a dedicated chitobiose phosphotransferase system to satisfy its requirement for exogenous GlcNAc (Rhodes et al. 2010).However, in the absence of both free GlcNAc and chitin this spirochete exhibits a unique biphasic growth pattern in which the second exponential phase is dependent on the expression of the chitobiose transporter (Tilly et al. 2001;Rhodes et al. 2009;Rhodes et al. 2010).In the first exponential phase cells grow to approximately 2.0  10 6 cells ml -1 within three days by exhausting the small amount of free GlcNAc and/or GlcNAc oligomers from complex medium components such as rabbit serum and yeastolate.After utilizing these sources of GlcNAc cells exhibit a death phase in which cell density decreases below 1.0  10 5 cells ml -1 between days three and five.Finally, a second exponential phase is initiated between days five and seven in which cells grow to near optimal cell density (~5.0  10 7 cells ml -1 ).
This peculiar growth pattern in the absence of free GlcNAc or chitin suggests there is a sequestered (or bound) source of GlcNAc within the medium that cells can access in the second exponential phase, but are unable to utilize in the first exponential phase.While this source remains unknown, as well as the mechanism(s) used to obtain it, our data demonstrating the utilization of chitin by B. burgdorferi in the absence of free GlcNAc suggests that polymerized GlcNAc is not the source (Rhodes et al. 2010).In addition, it is unclear whether the adaptation to GlcNAc starvation (i.e.second exponential phase growth) is physiologic or genetic in nature.A previous report by Tilly et al. (Tilly et al. 2001) suggested that second exponential phase growth was a physiologic response, as cells from the second exponential phase exhibited biphasic growth after being transferred to fresh medium lacking GlcNAc.However, in that experiment starvation-adapted cells grew almost 10-fold higher in the first exponential phase as compared to cells that were not previously exposed to the starvation condition, leaving open the possibility that second exponential phase growth is due to the outgrowth of a mutant population.
Stress-induced mutagenesis occurring in bacteria under growth-limiting conditions is a phenomenon that has been intensively studied in recent years.Nearly two decades ago Cairns and Foster described adaptive reversion using an Escherichia coli Lac assay (Cairns & Foster 1991).In this system cells carrying a lacI-lacZ fusion gene with a +1 frameshift mutation in lacI were unable to utilize lactose, but when cultured with a non-lactose carbon source and then spread onto lactose plates, Lac+ revertants were obtained.Characterization of these stress-induced mutants revealed they were of two types: i) point mutants carrying a compensatory frameshift mutation in lac (Foster & Trimarchi 1994;Rosenberg et al. 1994) or ii) lac-amplified cells in which the leaky lac frameshift allele was amplified 20-50 times (Hastings et al. 2000;Powell & Wartell 2001;Kugelberg et al. 2006;Slack et al. 2006).Substantial work has been conducted to elucidate the underlying molecular mechanisms responsible for these two distinct types of stress-induced mutants.Formation of the Lac+ point mutants can occur via several stress-response and DNA repair and metabolism pathways, including i) homologous recombination via double-strand-break-repair (DSBR) proteins (Cairns & Foster 1991;Harris et al. 1994), ii) error-prone DNA replication (McKenzie et al. 2001), iii) the SOS response (McKenzie et al. 2000), iv) the RpoS starvationstress response (Lombardo et al. 2004), and v) methyl-directed mismatch repair (MMR) (Harris et al. 1999).In contrast, lac-amplification does not require error-prone DNA replication or the SOS response (McKenzie et al. 2001), but does require the general stressresponse regulator RpoS and homologous recombination via DSBR proteins such as RecA, RecBCD and RuvABC (Slack et al. 2006).While B. burgdorferi has a limited genome and lacks error-prone DNA polymerases and the SOS pathway, it is possible that stress-induced mutagenesis occurs in this organism and could be mediated by mismatch repair, RpoS and/or homologous recombination.
Several studies have been conducted in B. burgdorferi to evaluate the function of RecA both in vitro and in vivo (Liveris et al. 2004;Putteet-Driver et al. 2004;Liveris et al. 2008).Early attempts to generate a recA null mutant were unsuccessful; therefore transgenic expression of B. burgdorferi RecA in E. coli was used to determine its functional properties (Liveris et al. 2004;Putteet-Driver et al. 2004).In E. coli RecA has three major activities: i) it participates in homologous recombination by catalyzing DNA strand exchange, ii) it promotes the autocatalytic cleavage of LexA, a repressor of the SOS response, and iii) it directly facilitates bypass of DNA lesions by DNA polymerase V during the SOS response (Cox 2007).Complementation of an E. coli recA null mutant with B. burgdorferi recA restored the ability of the mutant to promote homologous recombination and activated the E. coli SOS pathway in response to DNA-damaging agents (Liveris et al. 2004;Putteet-Driver et al. 2004).In a more recent study, Liveris et al (Liveris et al. 2008) successfully generated a recA mutant in a low passage infectious B. burgdorferi strain and showed that RecA mediates homologous recombination in this organism.However, this report demonstrated RecA is only minimally involved in DNA repair in response to DNA damaging agents, and does not mediate antigenic variation at the vlsE locus after mammalian infection.
As described above, a previous attempt to characterize B. burgdorferi second exponential phase growth in the absence of free GlcNAc has left unanswered questions as to whether this is a physiologic or genetic adaptation to the starvation condition.In this study we follow the growth of wild-type and recA mutant cells in the absence of free GlcNAc and suggest that growth in the second exponential phase is the result of a genetic adaptation that is, at least in part, mediated by RecA.Additionally, we characterize the recA transcriptional unit and provide evidence suggesting this gene is transcribed from two separate promoters.

Adaptation to GlcNAc starvation
Wild-type cells cultured in the absence of free GlcNAc exhibit biphasic growth (Fig. 1A), with growth in the second exponential phase dependent on the expression of the chitobiose transporter chbC (Tilly et al. 2001;Rhodes et al. 2010).To determine if cells in the second exponential phase were adapted to grow in the absence of free GlcNAc we followed cell growth in BSK-II lacking GlcNAc.Cells were diluted to 1.0  10 5 cells ml -1 in the same medium at various time points along the biphasic growth curve, and growth of the new cultures was followed to determine if biphasic growth occurred (Fig. 1B and C).Diluting cells during the first exponential phase on day 2 (46 h) or at the peak of the first exponential phase on day 3 (72 h) into BSK-II lacking GlcNAc resulted in biphasic growth (Fig. 1B).In contrast, cells diluted from the second exponential phase on days 8-11 (191, 216, 237 and 263 h) into BSK-II lacking GlcNAc resulted in growth to high cell densities in a single exponential phase, with cell densities reaching  2.5  10 7 cells ml -1 (Fig. 1C).This cell density is equivalent to that reached by cells in the second exponential phase when cultured in BSK-II lacking free GlcNAc, though about one-tenth the cell densities reached by cells cultured in complete BSK-II.In some experiments cells diluted from the second exponential phase were followed for extended periods of time after reaching their peak cell density (~2.5  10 7 cells ml -1 ) in a single exponential phase, and while there was a decline in cell density a second exponential phase was not observed as previously described (Tilly et al. 2001).

Adaptation to GlcNAc is the result of a mutational event
Two possible explanations for second exponential phase growth in a GlcNAc depleted medium are that: 1) second exponential phase growth occurs as a result of a physiological adaptation to GlcNAc starvation as cells induce genes to utilize sequestered GlcNAc; or that 2) second exponential phase growth occurs as the result of a mutational event that allows for the outgrowth of a mutant population able to obtain sequestered GlcNAc from a source they were previously unable to utilize.If second exponential phase growth was the result of a mutational event, it would be expected that cells removed from the second exponential phase would retain the ability to grow in a single exponential phase in BSK-II lacking GlcNAc after serial passage in complete BSK-II.On the other hand, if a physiologic mechanism was responsible, then it would be expected that cells removed from the second exponential phase and subjected to serial passage in a complete medium would still exhibit biphasic growth when transferred back to a medium lacking GlcNAc.To test this, cells were diluted from the second exponential phase into complete BSK-II, serially passaged up to twelve times in this complete medium, and then re-evaluated for biphasic growth after being transferred back to BSK-II lacking GlcNAc (Fig. 2).Despite being cultured in the presence of free GlcNAc for up to twelve passages, cells exhibited only one exponential phase upon re-introduction to BSK-II lacking GlcNAc.Interestingly, cells diluted into BSK-II without GlcNAc after 7 and 12 serial passages in complete BSK-II grew to a higher cell density (~5-fold) than cells cultured in complete BSK-II for 1 or 2 passages.
To further demonstrate that adaptation to GlcNAc starvation is the result of a genetic event, cells were plated on BSK-II agar and the growth of wild-type or starvation-adapted clones was followed.Specifically, wild-type cells cultured in complete BSK-II and starvation-adapted cells serially passaged seven times in complete BSK-II (P.7 No GlcNAc from Fig. 2; closed squares) were plated on BSK-II agar containing free GlcNAc to obtain isolated colonies.Ten clones from each were transferred to BSK-II broth lacking GlcNAc and growth was followed daily.All ten clones from wild-type cells not previously adapted to GlcNAc starvation exhibited biphasic growth when cultured in the absence of free GlcNAc (Fig. 3A).In contrast, eight of the ten clones obtained from cells previously adapted to GlcNAc starvation and serially passaged seven times in complete BSK-II grew to optimal cell densities in a single exponential phase when cultured in BSK-II lacking free GlcNAc (Fig. 3B).Fig. 2. Growth of starvation-adapted cells in BSK-II without GlcNAc after repeated subculture in complete BSK-II.B31-A cells grown in BSK-II without added GlcNAc (closed gray circle) were transferred (arrow indicates point at which adapted cells were subcultured) to complete BSK-II.After one passage in complete BSK-II cells were diluted to 1.0  10 5 cells ml -1 in BSK-II without GlcNAc (P.1; closed triangle) and growth was followed daily by darkfield microscopy.Growth was also followed in BSK-II without GlcNAc after two (P.2; open triangle), seven (P. 7; closed square) and twelve (P.12; open square) serial passages in complete BSK-II.This is a representative experiment that was repeated twice.Fig. 3. Growth of GlcNAc starvation adapted and non-adapted clones in BSK-II without GlcNAc.(A) B31-A cells cultured in complete BSK-II (non-adapted) were plated on BSK-II agar (containing GlcNAc) and 10 clones were transferred to BSK-II broth without GlcNAc.Growth was followed daily by darkfield microscopy.(B) B31-A cells adapted to GlcNAc starvation and serially passaged in complete BSK-II were plated on solid medium and 10 clones were transferred to liquid BSK-II without GlcNAc.Growth was followed daily by darkfield microscopy.
It is possible that the two clones (clones #4 and #5 from Fig. 3B) from GlcNAc starvationadapted cells exhibited biphasic growth as the result of a second-site mutation or it may be that a sub-population of non-adapted cells was able to persist in the absence of GlcNAc as the result of the mutant population releasing sequestered GlcNAc from an as yet unknown medium component.In an attempt to clarify this point, cells from starvation-adapted clone #10 (Fig. 3B; designated RR45) were passaged twice in complete BSK-II, and then plated on BSK-II agar with GlcNAc.Ten clones were transferred to BSK-II broth without GlcNAc and growth was followed daily by darkfield microscopy.In addition, wild-type cells cultured in complete BSK-II were plated on BSK-II agar with GlcNAc, and the growth of ten clones was followed in liquid BSK-II without GlcNAc (Fig. 4A).As previously demonstrated, the ten wild-type clones exhibited biphasic growth (Fig. 4A).In contrast, all ten RR45 clones grew to optimal cell densities in a single exponential phase (Fig. 4B).Fig. 4. Growth of GlcNAc starvation adapted RR45 and non-adapted wild-type clones in BSK-II without GlcNAc.(A) B31-A cells cultured in complete BSK-II (non-adapted) were plated on complete BSK-II agar and 10 clones were transferred to liquid BSK-II without GlcNAc.Growth was followed daily by darkfield microscopy.(B) RR45 cells (clone #10 from Fig. 3B) were plated on complete BSK-II agar and 10 clones were transferred to liquid BSK-II without GlcNAc.Growth was followed daily by darkfield microscopy.

Construction and characterization of a B. burgdorferi recA mutant and two complemented mutants
As double-strand-break-repair (DSBR) proteins, such as RecA, have been shown to be involved in stress-induced mutation in E. coli (He et al. 2006), we hypothesized that recA may play a role in the ability of B. burgdorferi to adapt to GlcNAc starvation.Therefore, we generated a recA deletion mutant in the B31-A background by allelic exchange, and designated this strain JA10 (see Methods).To confirm that RecA function was inactivated, we examined survival after exposure to DNA-damaging agents, as well as the ability to promote homologous recombination.Recently, Liveris et al. (Liveris et al. 2008) demonstrated that a low-passage infectious B. burgdorferi recA mutant was only moderately sensitive to DNA-damaging agents (UV irradiation and mitomycin C), but showed that RecA was essential for allelic exchange in this organism.We confirmed those results here using a high-passage non-infectious recA mutant (JA10) which showed no significant difference in sensitivity to DNA-damaging agents when compared to wild-type cells (data not shown), but was unable to promote homologous recombination (Table 1).Table 1.Transformation and recombination frequencies of the wild type (B31-A), recA mutant (JA10) and recA complemented mutants (JA15 and JR1).
(ermC::rpoN), a suicide vector carrying a mutant rpoN allele that confers erythromycin resistance.The wild-type strain exhibited a recombination frequency of 3.3  10 -6 (Table 1), while no recombinants were observed for the recA mutant.It is possible, though unlikely, that the lack of recombination observed in the recA mutant was due to the inability of the plasmid to gain entry into the cell during electroporation.Therefore, we determined the transformation frequency of all strains by modifying the pCE320 B. burgdorferi shuttle vector to confer erythromycin resistance, and transformed cells with this new plasmid designated ErmC/pCE320.Transformation frequency was similar for both the wild-type and recA mutant strains (6.7  10 -5 and 5.4  10 -5 , respectively; Table 1), suggesting that the lack of recombinants observed after transformation with pBB0450.1 was due to the absence of a functional recA gene.
To provide further evidence that the recA gene in JA10 was inactivated, the mutant was complemented with plasmid pBB0131comp.2which consists of the wild-type recA gene with 512 bp of upstream DNA, presumably encompassing its natural promoter, cloned into the B. burgdorferi shuttle vector pBSV2 (see Methods).This strain was designated JA15 and transformation and recombination frequencies were determined as described above.The JA15 complemented mutant exhibited a transformation frequency similar to that observed for the wild-type and recA mutant strains (5.6  10 -5 ; Table 1).However, this strain unexpectedly exhibited no homologous recombination, suggesting the 512 bp of DNA upstream of the recA translational start site may not be sufficient for expression of recA at wild-type levels.Therefore, we generated a second recA complemented mutant using plasmid pRecA.3consisting of the wild-type recA gene fused to the constitutive B. burgdorferi flaB promoter and cloned into the shuttle vector pCE320.This strain was designated JR1 and transformation and recombination frequencies were determined.While this strain exhibited a marked reduction in transformation frequency compared to the wildtype strain, RecA function was restored as JR1 demonstrated a 2-fold greater recombination frequency as compared to B31-A (Table 1).The reduction in transformation frequency observed in JR1 is likely due to the incompatibility of pRecA.3 and ErmC/pCE320 as they are both derivatives of the pCE320 shuttle vector.

Expression of recA determined by quantitative RT-PCR (qRT-PCR)
One possible explanation for the lack of recombination observed in the JA15 strain is that recA transcription was reduced, leading to decreased levels of functional RecA protein.
To test this, we evaluated recA expression in B31-A, JA10, JA15 and JR1 using qRT-PCR (Table 2).RNA was extracted from late-log phase cells (~2.0  10 7 cells ml -1 ) and the amount of recA transcript present in 10 ng of total RNA was determined.As expected, recA transcript levels in the JA15 complemented mutant were 61.5-fold lower than those measured in wild-type cells, possibly accounting for the lack of recombination observed in this strain.In contrast, recA transcript levels in the JR1 complemented mutant, in which recA expression was driven by a constitutive promoter, were 15.7-fold higher than the wild type.No recA transcription was detected in JA10 cells.These results may explain the 2-fold increase observed in recombination frequency in JR1 compared to the wild type.It is important to note that despite the differences in recA levels in these different strains, the growth rate between strains did not vary and the growth curves in complete BSK-II were indistinguishable (data not shown).

Characterization of the recA transcriptional unit
Analysis of the B. burgdorferi genome reveals two genes (BB0132 and BB0133) immediately upstream of recA (BB0131) that are oriented in the same direction (Fig. 5A).The reduced expression of recA in the JA15 complemented mutant may be the result of regulatory elements encoded outside of the 512 bp of upstream DNA amplified to generate the complementation construct pBB0131comp.2.It is also possible, given the organization of the flanking genes, that recA is co-transcribed with one or both of the upstream genes as part of an operon, and that the low level of expression observed in JA15 was due to a weak internal promoter.As previous studies have not evaluated the organization of the recA transcript in B. burgdorferi, we attempted to shed light on these questions by characterizing the recA transcriptional unit and identifying the recA promoter(s).
RT-PCR was used to determine if recA is co-transcribed with one or more of its neighboring genes (Fig. 5B-D).Antisense primers (F1 and F3; see Fig. 5) were used to generate cDNA from points originating in the downstream gene BB0129 and within the recA coding region.
The cDNA was amplified in PCR reactions using the antisense primer and various sense primers to determine if mRNA transcripts spanned the junctions between adjacent genes.To determine if recA is co-transcribed with downstream genes, cDNA generated using primer F1 was PCR amplified using the same antisense primer and primers R1, R1A and R2B (Fig. 5B).All primer pairs produced products of the correct size when compared to positive controls, indicating the presence of an RNA transcript spanning genes BB0129-BB0132 and suggesting recA is co-transcribed with at least two downstream genes.In addition, the positive RT product obtained with primers F1 and R2B (lane 9; Fig. 5B) suggested that recA may be part of a larger operon encompassing one or both upstream genes, as the R2B primer was located 1.2 kb upstream of the recA translational start site.To confirm that recA is cotranscribed with BB0132, cDNA was generated using antisense primer F3 and PCR amplified using the same antisense primer and sense primers R1A, R2, R2A, R2B and R3 (Fig. 5C).An RT product of the correct size was obtained with all primer pairs when compared to the positive controls, confirming the results obtained with antisense primer F1 (Fig. 5B) and strongly suggesting that recA is co-transcribed with BB0132.To determine if BB0133 is also part of this transcriptional unit, cDNA generated with antisense primer F3 was PCR amplified using the same antisense primer and sense primers R4 and R5 (Fig. 5D).
An RT product of the correct size was obtained with both primer pairs when compared to the positive controls, providing evidence that BB0133 is also part of this operon.Taken together, these RT results strongly suggest recA is co-transcribed with the upstream genes BB0132 and BB0133 and at least two downstream genes (BB0130 and BB0129).

5' RACE and promoter analysis
The RT results presented above strongly suggest that recA is co-transcribed with the upstream genes BB0132 and BB0133 (Fig. 5B-D); however, we did observe reduced expression of recA in our mutant complemented with the wild-type gene and 512 bp of upstream DNA (Table 2).Together, these results suggest two recA transcripts may exist, one beginning upstream of BB0133 and the other beginning within the 512 bp immediately upstream of the recA gene.Therefore, we used 5' RACE to identify the two recA transcritptional start sites, and compared the promoter regions upstream of the start sites to previously identified RpoD, RpoS and RpoN-dependent promoter sequences in B. burgdorferi (Studholme & Buck 2000;Caimano et al. 2007;Rhodes et al. 2009).
To identify the transcriptional start site immediately upstream of the recA gene, total RNA was extracted from B31-A cells cultured to late log phase (~2.0  10 7 cells ml -1 ) in complete BSK-II and cDNA was generated using antisense primer F3 (Fig. 5A) located within the recA coding region.The cDNA was column purified and a homopolymeric A-tail was added using terminal transferase.PCR amplification of the A-tailed cDNA with nested g e n e s p e c i f i c p r i m e r F 4 a n d a n o l i g o d T -a n c h o r p r i m e r r e s u l t e d i n a p r o d u c t o f approximately 250 bp (lane 2; Fig. 6A).The DNA band was extracted from the gel and sequenced with primer F4, and the transcriptional start site was determined to be 44 or 45 bp upstream of the recA translational start (Fig. 6C).We attempted to clarify the start site using G-tailed cDNA, but were unable to obtain reliable sequence data for this PCR product due to mispriming of the oligo dC-anchor primer to three guanosines (underlined in Fig. 6C) located 29-31 bp upstream of the recA translational start site (data not shown).
Based on the sequence data obtained from the A-tailed reaction we identified the putative -10 and -35 promoter regions immediately upstream of the recA gene.Comparison of these putative promoter regions with previously described B. burgdorferi promoter consensus sequences revealed similarity to the RpoD-dependent consensus promoter, as homology was observed in five of the six bases within the -35 region and seven of the eleven bases within the extended -10 region (Fig. 7).
To determine the transcriptional start site upstream of BB0133, total RNA was extracted from B31-A cells cultured to late log phase (~2.0  10 7 cells ml -1 ) in complete BSK-II, and cDNA was generated using antisense primer F5 (Fig. 5A) located within BB0132.A homopolymeric A-tail was added and PCR of the A-tailed cDNA with nested gene specific primer F6 and the oligo dT-anchor primer resulted in a 1.2 kb product (Fig. 6B).The DNA band was gel extracted and sequenced with nested primer F8, and the transcriptional start site was determined to be 45 or 46 bp upstream of the BB0133 translational start site (Fig. 6D).Both G-tailed and C-tailed cDNA were used to try and resolve this ambiguity, but PCR  amplification with the nested primer F6 and the complementary-tailed anchor primers (C or G, respectively) did not produce a PCR product (data not shown).Therefore, we used the approximate transcriptional start site determined from the A-tailed cDNA to identify the putative -10 and -35 promoter regions.Comparison to B. burgdorferi consensus promoters suggests this operon is under the control of an RpoD-dependent promoter as homology was observed between five of the six bases within the -35 region and ten of eleven bases within the extended -10 region (Fig. 7). www.intechopen.com Lyme Disease 124

Growth of the recA mutant in BSK-II without GlcNAc
Growth analysis of wild-type B. burgdorferi adapted to GlcNAc starvation suggests this organism undergoes a mutational event that allows growth to high cell densities in a single exponential phase in the absence of free GlcNAc (Figs. 2-4).Previous reports have demonstrated that DSBR proteins play a role in stress-induced mutations observed in other organisms (for reviews see (Galhardo et al. 2007;Hastings 2007)).Therefore, we evaluated the ability of a recA mutant in B. burgdorferi (JA10) to grow and adapt to GlcNAc starvation (Fig. 8).Late log phase recA mutant (JA10) cells cultured in complete BSK-II were diluted to 1.0  10 5 cells ml -1 in BSK-II lacking GlcNAc.Growth of the recA mutant prior to adaptation to GlcNAc starvation resulted in biphasic growth similar to wild-type cells that were not previously adapted to GlcNAc starvation (compare Fig. 8 open circles to Fig. 1A open circles).Similarly, when the recA mutant was diluted from the second exponential phase of a GlcNAc-starved culture (Fig. 8; 300 h) into the same medium biphasic growth was still observed (Fig. 8 open and closed triangles).This is in contrast to the single exponential phase typically observed in starvation-adapted wild-type cells that were cultured in the absence of GlcNAc (Fig. 1C).Multiple growth experiments evaluating adaptation to GlcNAc starvation (i.e.dilution from the second exponential phase) were performed for each strain (Table 3).The growth experiment was repeated 15 times with recA mutant cells, and biphasic growth was observed each time cells were diluted from the second exponential phase of a culture starved for GlcNAc.This is in contrast to wild-type cells which exhibited biphasic growth only 3 times in 33 separate growth experiments, and reached high cell densities in a single exponential phase in 30 separate experiments.Table 3. Adaptation to GlcNAc starvation.

Strain
To provide further evidence that RecA plays a role in the adaptation to GlcNAc starvation, we evaluated growth after GlcNAc starvation in the recA complemented mutants.Growth of the JA15 complemented mutant was followed in 17 separate experiments and single exponential phase growth to high cell densities in the absence of free GlcNAc was observed 5 times.Similarly, the JR1 complemented mutant was evaluated in 13 separate growth experiments and exhibited a single exponential phase in BSK-II lacking GlcNAc 3 times.Taken together, these results suggest RecA plays a role in adaptation to GlcNAc starvation and that the level of RecA expression may be important in the process.

Discussion
B. burgdorferi exhibits a unique biphasic growth pattern when cultured in the absence of free GlcNAc, with cells reaching optimal density in the second exponential phase.However, it had not been conclusively determined whether a physiologic or genetic adaptation mechanism is responsible for second exponential phase growth.This study was conducted to address this question as well as identify genes important to the process.We provide evidence that adaptation to GlcNAc starvation in B. burgdorferi is the result of a genetic event, and suggest a role for the homologous recombination protein RecA.In addition, we characterized the recA transcript and showed that transcription occurs from two different promoter sites.
In a previous report, Tilly et al (Tilly et al. 2001) suggested that the biphasic growth observed during B. burgdorferi cultivation in the absence of free GlcNAc was the result of a physiologic adaptation, as cells diluted from the second exponential phase into fresh medium lacking GlcNAc still exhibited biphasic growth.However, careful inspection of the growth curve from that study revealed that starvation-adapted cells grew nearly 10-fold higher in the first exponential phase after re-introduction into BSK-II lacking GlcNAc, than cells that were not previously exposed to the starvation condition (Tilly et al. 2001).To us, these results suggested that the adaptation that permitted second exponential phase growth was maintained by cells after passage in the same medium.We followed the growth of cells transferred from the second exponential phase into BSK-II lacking GlcNAc and found that wild-type cells grew to a high cell density in a single exponential phase (Fig. 1C) in 30 out of 33 independent growth experiments (Table 3).This indicated that adaptation to the starvation condition was retained by the cells in nearly all of the experiments.However, it was not clear whether the adaptation was physiologic or genetic, as the starvation-stress condition was never removed (i.e. cells were only transferred from the second exponential phase directly into media lacking GlcNAc).Since other microorganisms exhibit stressinduced mutagenesis in response to nutrient deprivation and other stresses (for reviews see (Finkel 2006;Galhardo et al. 2007;Hastings 2007)), we hypothesized that B. burgdorferi may use similar mechanisms to adapt to GlcNAc starvation.
To determine if adaptation to GlcNAc starvation was physiologic or genetic, starvationadapted cells were serially passaged in complete BSK-II to relieve the starvation-stress, then re-introduced to the starvation-stress condition (i.e.BSK-II lacking GlcNAc).If adaptation was physiologic then cells should exhibit biphasic growth; however, we found that after as many as twelve serial passages in a complete medium, cells still reached high cell density in a single exponential phase when re-introduced to a medium lacking free GlcNAc (Fig. 2).In addition, starvation-adapted cells were plated on BSK-II agar after seven serial passages in complete BSK-II and the growth of ten clones was followed in BSK-II broth lacking GlcNAc.Eight of the ten clones grew to optimal cell densities in a single exponential phase (Fig. 3B).It is possible that the two clones exhibiting biphasic growth were never adapted and only able to persist due to the release of GlcNAc from sequestered sources by the cells that had undergone adaptation.It is also possible that these two clones were originally adapted to GlcNAc starvation, but during serial passage in complete media became de-adapted, exhibiting the original wild-type phenotype (discussed below).In any case, re-plating of one of the adapted clones (RR45) resulted in all ten clones growing to a high cell density in a single exponential phase (Fig. 4B).These results strongly suggest that B. burgdorferi undergoes stress-induced mutagenesis that allows cells to adapt to GlcNAc starvation.
As discussed above, stress-induced adaptive mutagenesis in the E. coli Lac assay results in either point mutants or amplification of the "leaky" lac region that relieves the stress and allows cells to utilize lactose (Cairns & Foster 1991).This adaptive mutagenesis occurs by a number of different pathways (Cairns & Foster 1991;Harris et al. 1994;Harris et al. 1999;McKenzie et al. 2000;McKenzie et al. 2001;Lombardo et al. 2004), three of which are present in B. burgdorferi: i) MMR, ii) RpoS starvation-stress-response and iii) homologous recombination by DSBR proteins.The gene or genes directly involved in the utilization of sequestered GlcNAc in the second exponential phase remain unknown; therefore, it is difficult to determine if the genetic adaptation observed in B. burgdorferi is the result of a point mutation, amplification or both.However, data collected under GlcNAc starvation conditions relating to these three pathways in B. burgdorferi may shed some light on this question.
MMR is a highly conserved process that repairs mismatched bases that arise during replication, increasing the fidelity of DNA replication 100 to 1000-fold (reviewed in (Kunkel& Erie 2005)).DNA mismatches are recognized by the MutS and MutL proteins, and the mismatched base is removed from the unmethylated strand by the endonuclease activity of MutH.A reduction in MMR activity during stationary phase has been described and is one mechanism by which mutation frequency is increased, possibly resulting in advantageous point mutations (not amplifications) that allow for adaptation to certain stresses (Feng et al. 1996;Harris et al. 1999).In addition, inactivation of MMR genes results in a mutator phenotype that has been observed in natural bacterial populations (LeClerc et al. 1996;Matic et al. 1997;Oliver et al. 2002).We evaluated our GlcNAc starvation-adapted clones for a mutation in mutS or mutL to determine if adaptation occurred due to inactivation of the MMR system.Genomic DNA was extracted from five starvation-adapted clones (Fig. 4, clones 1-5) and the mutS and mutL genes were sequenced along with 1 kb of flanking DNA on either side of the gene (data not shown).No mutations were observed in any of the clones, suggesting these clones have functional MMR systems and that adaptive mutation did not occur due to the absence of this important repair pathway.It is unknown if MutS and MutL levels were reduced during GlcNAc starvation, but it may be of interest to determine these levels in future experiments.
The point mutants observed in the E. coli Lac assay require the presence of RpoS and the general stress response (Lombardo et al. 2004).It is unclear exactly how RpoS or components of the RpoS regulon regulate the formation of point mutations that confer the Lac+ phenotype, but it is known that RpoS promotes the use of the error-prone DNA polymerase DinB in DSBR (Ponder et al. 2005).As B. burgdorferi does not possess a DinB homologue or other error-prone DNA polymerases, it is unlikely this mechanism plays a role in adaptation of this spirochete to GlcNAc starvation.However, the RpoS regulon has also been shown to be necessary for adaptive amplification in the E. coli Lac assay (Lombardo et al. 2004), and previous studies conducted in our lab have demonstrated the importance of RpoS in GlcNAc starvation and biphasic growth (Rhodes et al. 2009).We reported a role for RpoS in the regulation of chitobiose utilization, and also demonstrated that an rpoS mutant was unable to initiate second exponential phase growth in the absence of GlcNAc until after 400 hours.In contrast, wild-type cells initiate a second exponential phase by 125 hours and reach optimal cell densities by 300 hours.It is possible the long delay observed in the rpoS mutant reflects the inability of these cells to amplify the gene(s) needed to obtain sequestered GlcNAc in the second exponential phase.
The DSBR protein RecA is also required for stress-induced point mutation and amplification in the E. coli Lac assay.With regard to point mutants, RecA most likely plays a role in increased mutagenesis through the up-regulation of the SOS response.As discussed above, B. burgdorferi does not possess the components for a functional SOS pathway, and so adaptive mutation in response to GlcNAc starvation by this mechanism is unlikely.However, homologous recombination and DSBR proteins such as RecA are also required for adaptation via amplification.Therefore, we evaluated the ability of a recA mutant to adapt to GlcNAc starvation by culturing cells in BSK-II without GlcNAc, and then diluting cells from the second exponential phase into the same medium.While the recA mutant exhibited biphasic growth similar to the wild type, these cells were unable to permanently adapt to GlcNAc starvation (Fig. 8) in any of the 15 independent experiments conducted (Table 3).These results suggest a role for RecA in the adaptation of wild-type cells to GlcNAc starvation, possibly through amplification of target gene(s) by homologous recombination.
Complementation of the recA mutant with two different complementation plasmids resulted in recA expression that was either 62.5-fold lower (JA15) or 15.7-fold higher (JR1) than wildtype expression (Table 2), and only partial restoration of the adaptive phenotype (Table 3).The partial restoration of the adaptive phenotype in the JA15 complemented mutant may be due to reduced levels of RecA protein resulting in less amplification of the target genes necessary for obtaining sequestered GlcNAc.While only partial restoration of the adaptive phenotype in the JR1 complemented mutant may be due to a deamplification of the repeated target gene(s) because of abnormally high levels of RecA.It has been observed in the E. coli Lac assay that lacamplified clones (Lac + ) plated on rich medium containing a color indicator for -galactosidase give rise to both white (Lac-) and sectored colonies (blue colonies with white sectors) due to deamplification of the repeated lac region, and that deamplification is mediated by RecA (Tlsty et al. 1984).This deamplification mechanism could also account for the two clones that exhibited biphasic growth in BSK-II lacking GlcNAc after repeated passage in complete BSK-II (Fig. 3B), and the three GlcNAc starvation experiments in wild-type cells that did not result in the adaptive phenotype (Table 3).
As a result of the reduced recA expression in our complemented mutant with the wild-type recA gene and 512 bp of upstream DNA (Table 2; JA15) we characterized the recA transcriptional unit (Fig. 5) and identified two putative promoters for this gene (Fig. 6 and  7).RT-PCR analysis confirmed that recA was co-transcribed with both upstream genes (BB0132 encoding the transcription elongation factor GreA; and BB0133 encoding a protein with a tetratricopeptide repeat domain) and at least two of the downstream genes (BB0130 encoding a conserved hypothetical protein; and BB0129 encoding an RNA-uridine isomerase).The combination of our RT-PCR data with our expression data from JA15 suggested the possibility of two promoters, one just upstream of the recA coding region and the other upstream of BB0133.We confirmed this using 5' RACE (Fig. 6), and identified two putative promoters based on comparison to previously identified B. burgdorferi RpoD, RpoS and RpoN-dependent promoters (Fig. 7) (Studholme & Buck 2000;Caimano et al. 2007;Rhodes et al. 2009).The promoter identified upstream of BB0133 is nearly identical to the RpoD consensus promoter indicating constitutive expression under normal conditions, while the putative promoter immediately upstream of the recA coding region shows the most, but only limited, homology to the RpoD consensus sequence.In light of these results and the role recA plays in adaptation to GlcNAc starvation, it is interesting to speculate that the promoter immediately upstream of recA may be involved in the expression of this gene under specific cellular conditions or stresses.

Bacterial strains and culture conditions
The bacterial strains and plasmids used in this study are described in Table 4. Wild-type and mutant B. burgdorferi strains were routinely cultured in modified BSK-II (Barbour 1984) medium supplemented with 7% rabbit serum (Invitrogen, Corp.; Carlsbad, CA) and any necessary antibiotics.BSK-II was modified from the original recipe by replacing 10x CMRL-1066 with 10x Media 199 (Invitrogen).In addition, free GlcNAc was left out of BSK-II for certain growth experiments.For selection and maintenance of antibiotic resistance in B. burgdorferi, streptomycin, kanamycin and erythromycin were used at concentrations of 100 g ml -1 , 340 g ml -1 and 1 g ml -1 respectively.

www.intechopen.com
Adaptation to Glucosamine Starvation in Borrelia burgdorferi is Mediated by recA 129

Growth curve analysis
For analysis of growth in liquid cultures, late-log phase cells (5.0  10 7 cells ml -1 to 1.0  10 8 cells ml -1 ) were diluted to 1.0  10 5 cells ml -1 in BSK-II with or without GlcNAc.In general, 6-12 l of culture was used to inoculate 6 ml of fresh culture medium; therefore, negligible amounts of free GlcNAc were transferred with the inoculum when analyzing growth in medium lacking GlcNAc.Cells were enumerated daily by darkfield microscopy using a Petroff-Hauser (Hauser Scientific; Horsham, PA) counting chamber.Specifically, 2.5 l of liquid culture was transferred to the counting chamber, and cells in all 25 squares were counted.Cells were diluted 1:10 in BSK-II prior to counting when the culture density was greater than 1.0  10 7 cells ml -1 .All growth experiments were carried out at 33C in the presence of 3% CO 2 .Unless otherwise stated, each growth curve presented in this study is representative of at least three independent trials, as growth data could not be pooled due to the length of experiments and the different times at which bacteria were enumerated.

Construction of a recA mutant in B. burgdorferi
The construct used to generate a recA (BB0131) mutation in the B31-A background was created as follows: i) A 2.1 kb fragment of the 3' end of BB0131 and flanking DNA was PCR amplified with engineered restriction sites using primers (Table 5) 5' BB0131 mutF1 (KpnI) and 5' BB0131 mutR2 (XbaI); ii) the PCR fragment was TA cloned into pCR2.1 (Invitrogen) to generate plasmid pBB0131.1;iii) plasmids pBB0131.1 and pKFSS1 (Frank et al. 2003) (a B. burgdorferi shuttle vector conferring streptomycin resistance; Table 4) were digested with KpnI and XbaI and separated by gel electrophoresis; iv) the 2.1 kb band from pBB0131.1 was gel extracted and cloned into the gel extracted pKFSS1 fragment to generate pBB0131.2;v) the 2.1 kb fragment and flanking streptomycin resistance gene were PCR amplified from pBB0131.2 using primers 5' BB0131 mutF1 (KpnI) and pKFSS1 R1; vi) the 3.6 kb amplicon was TA cloned into pGEM T-Easy (Promega, Corp.; Madison, WI) to generate pBB0131.3Aor B (based on the orientation of the PCR product insertion); vii) plasmid pBB0131.3Bwas identified by restriction digest in which the 3' end of the streptomycin resistance gene was adjacent to the XmaI site in the pGEM T-Easy vector; viii) the 5' end of BB0131 and flanking DNA (1.8 kb) was PCR amplified with engineered restriction sites using primers 3' BB0131 mutF1 (XmaI) and 3' BB0131 mutR1 (SacII) and TA cloned into pCR2.1 to generate pBB0131.4;ix) plasmids pBB0131.3Band pBB0131.4 were digested with XmaI and SacII and separated by gel electrophoresis; x) the 1.8 kb fragment from pBB0131.4 was gel extracted and cloned into the gel extracted fragment from pBB0131.3B to generate the final construct pBB0131.5.All plasmids were confirmed by restriction digest and DNA sequencing.In summary, 877 of the 1098 bp constituting the recA coding region were deleted and replaced with the streptomycin resistance gene transcribed in the opposite orientation and under the control of the constitutive B. burgdorferi flgB promoter.Plasmid pBB0131.5 was confirmed by PCR using primers flanking the antibiotic insertion site (BB0131 mut confirm F1 and 5' recA SalI) that resulted in a 1.7 kb fragment (Fig. 9, lane 3) compared to a 1.2 kb fragment generated with the wild-type genomic DNA as a template (Fig. 9, lane 2).
To generate the recA deletion mutant, competent B31-A cells were prepared and transformed with 15 µg of pBB0131.5 as previously described (Samuels et al. 1994).Briefly, transformed cells were resuspended in 10 ml of complete BSK-II and allowed to recover overnight at 33C.The transformation reaction was plated in solid BSK-II medium containing 100 g ml -1 streptomycin to select for the allelic exchange event.Streptomycin resistant colonies were screened by PCR using primers that flanked the insertion site.One clone (JA10) was chosen for further experimentation and PCR of the genomic DNA from this strain with the primers BB0131 mut confirm F1 and 5' recA SalI demonstrated the presence of the mutant recA allele (Fig. 9, lane 5).In addition, the insertion was confirmed by DNA sequencing of the PCR product.

Complementation of the recA mutant
Two approaches were used to complement the recA mutation in JA10.First, the recA gene and 512 bp of upstream DNA were PCR amplified using primers BB0131 comp F1 (KpnI) and BB0131 comp R1 (XbaI), and the 1.6 kb fragment was TA cloned into pCR2.1 to generate pBB0131comp.1.Plasmids pBB0131comp.1 and pBSV2 (Stewart et al. 2001) (E.coli-B.burgdorferi shuttle conferring kanamycin resistance), were digested with KpnI and XbaI and separated by gel electrophoresis.The 1.6 kb fragment from pBB0131comp.1 was gel extracted and ligated with the gel extracted fragment from pBSV2 to generate pBB0131comp.2.Fifteen micrograms of pBB0131comp.2was transformed into competent JA10 cells and transformants were selected by plating on solid BSK-II medium containing 100 g ml -1 streptomycin and 340 g ml -1 kanamycin.Clones were screened by PCR using primers BB0131 mut confirm F1 and 5' recA SalI.One clone, designated JA15, was selected for further experimentation and the presence of both the mutant and wild-type recA alleles was confirmed by PCR using 50 ng of genomic DNA as template (Fig. 9; lane 6).
We also generated a complementation construct in which we fused the wild-type recA gene to the constitutive B. burgdorferi flaB promoter.The flaB promoter was PCR amplified from B31-A genomic DNA using primers PflaB F1 (SalI) and PflaB R1 (NotI), and cloned into pCR2.1 to generate pPflaB.The recA coding region was PCR amplified from B31-A genomic DNA using primers 5' recA (SalI) and BbrecA F1, and cloned into pCR2.1 to generate pRecA.1.Plasmids pPflaB and pRecA.1 were digested with SalI and NotI and separated by gel electrophoresis.The 360 bp product from pPflaB was gel extracted and ligated into the gel extracted fragment of pRecA.1 to generate pRecA.2.The flaB promoter-recA fusion was removed from pRecA.2 by digestion with KpnI and XbaI (sites in pCR2.1 flanking the TA insertion site), separated by gel electrophoresis, gel extracted and cloned into like sites in pCE320 (Eggers et al. 2002) (a B. burgdorferi shuttle vector with a circular plasmid 32 (cp32) origin for replication that confers kanamycin resistance) to generate pRecA.3.The complementation construct pRecA.3 was concentrated and transformed into JA10 as described above.Transformants were screened by PCR using primers BB0131 mut confirm F1 and 5' recA SalI, and one clone was designated JR1 and used for further experiments.PCR using 50 ng of JR1 genomic DNA resulted in amplification of both wild-type and mutant recA alleles (Fig. 9; lane 7).

Homologous recombination assay
A homologous recombination assay similar to that described previously by Liveris et al (Liveris et al. 2008) was used to determine transformation and recombination frequencies for the wild-type, recA mutant and recA complemented mutant strains.To determine transformation frequency we generated plasmid ErmC/pCE320 in which the pCE320 shuttle vector was modified to confer erythromycin resistance.Specifically, primers ErmC F1 and ErmC R1 were used to amplify the erythromycin resistance gene and its corresponding promoter from plasmid pBB0450.1,which was then TA cloned into pCR2.1 to generate pCR2.1/ErmC.Plasmid pCR2.1/ErmC was digested with BamHI and NotI, separated by gel electrophoresis, and the fragment containing ermC was gel extracted and cloned into like sites in pCE320 to generate ErmC/pCE320.Plasmid ErmC/pCE320 was concentrated to >1 g l -1 and cells were electroporated with 15 g of plasmid as previously described (Samuels et al. 1994).After recovery overnight in BSK-II, transformants were selected on solid BSK-II containing 1 g ml -1 erythromycin and transformation frequency was calculated by dividing the number of transformant colonies obtained by the total number of cells plated (CFU) as determined by dilution and plating without antibiotic selection.The presence of ErmC/pCE320 was confirmed in at least ten erythromycin resistant colonies from each experiment by colony PCR with primers specific for ermC.
The ability of recA expressing and non-expressing strains to mediate allelic exchange was determined by calculating the recombination frequency following electroporation with plasmid pBB0450.1 (rpoN::ermC with flanking DNA) (Rhodes et al. 2009).Transformation and plating with pBB0450.1 was carried out as described above for ErmC/pCE320 and recombinants were selected on solid BSK-II containing 1 g ml -1 erythromycin.Recombination frequency was calculated by dividing the number of erythromycin resistant colonies by the total number of cells plated (CFU) as determined by dilution and plating without selection.Allelic exchange was confirmed in at least ten erythromycin resistant colonies from each experiment by colony PCR with primers flanking the insertion site in rpoN.

RNA extraction
RNA was extracted from wild-type cells (B31-A) for use in RT-PCR and qRT-PCR experiments.Briefly, cells were harvested during mid to late log phase (1.0 to 5.0  10 7 cells ml -1 ) by centrifugation (10,000  g, 12 min, 4C).Pellets were resuspended in 500 l of cold BSK-II and transferred to a 2 ml microcentrifuge tube.One milliliter of Bacteria RNAProtect (Qiagen; Valencia, CA) was added, mixed by vortexing and incubated for 5 min at room temperature.Cells were collected by centrifugation at 5,000  g for 10 min, the supernatant was decanted and pellets were stored at -80C for up to 4 weeks prior to RNA extraction.RNA was extracted using the RNeasy Mini kit (Qiagen) according to the manufacturer's instructions and treated with RQ1 RNase-free DNase (Promega, Corp.) and RNasin (Promega, Corp.) to remove any contaminating genomic DNA.RNA was purified using the RNeasy Mini kit and following the RNA Clean-up protocol supplied by the manufacturer.RNA concentration (OD 260 ) and purity (OD 260/280 ) were determined by UV spectroscopy using the Nanodrop ND-1000 (Nanodrop products of Thermo Fisher Scientific; Wilmington, DE) and RNA integrity was verified by visualizing the intensity of the 16S and 23S rRNA bands on a 1% agarose gel.RNA was concentrated to greater that 1 g l -1 by isopropanol precipitation and stored at -80C.

Quantitative RT-PCR (qRT-PCR) to determine recA expression
qRT-PCR was performed using the Mx3005P Multiplex Quantitative PCR System in conjunction with the Brilliant SYBR Green Single-Step qRT-PCR Master Mix kit (Stratagene; La Jolla, CA) according to the manufacturer's instructions.Briefly, a standard curve (10 1 to 10 7 copies per reaction) was prepared using a purified recA PCR product that was generated from B31-A genomic DNA using primers BB0131 F1 and BB0131 R1.Reactions (25 l) containing 10 ng of total RNA extracted from mid-log phase cells (~1.0  10 7 cells ml -1 ) were run under the following conditions: 1 cycle of 50˚C for 30 min and 95˚C for 15 min, followed by 40 cycles of 95˚C for 30 s and 58˚C for 30 s. Fluorescence was measured at the end of the 58˚C step every cycle.Triplicate samples for each experiment were averaged, and the average and standard deviations were calculated for three independent experiments.All qRT-PCR experiments included both no-reverse transcriptase (RT) and no-template controls.
The recA copy number for each sample was calculated using the MxPro (Stratagene) data analysis software and the recA standard curve described above.Copy number in each sample was normalized based on the total RNA input (10 ng per reaction) and the fold difference in recA copy number was determined by comparing the expression in the recA mutant and complemented mutant strains to the expression in the wild-type strain.

RT-PCR to determine the recA transcriptional unit
RT-PCR was used to determine the organization of the recA transcriptional unit.
Complementary DNA (cDNA) was generated from B31-A RNA using the SuperScript First Strand Synthesis for RT-PCR kit (Invitrogen, Corp.) according to the manufacturer's instructions.Briefly, 0.5 g of RNA was reverse transcribed with SuperScript II RT enzyme at 42C for one hour using gene specific primers recA operon F1 or F3.The RT enzyme was heat inactivated and RNA was degraded from the RNA-cDNA hybrid using RNaseH.To confirm there was no genomic DNA contamination a separate no RT reaction was run in conjunction with each RT reaction.For PCR, antisense primers used to generate the cDNA were used in various combinations with the sense primers recA operon R1, R1A, R2, R2A, R2B, R3, R4 or R5.Three reactions were prepared for each primer set and the following templates were added: i) 50 ng of B31-A genomic DNA, ii) 2 l of the appropriate cDNA reaction, and iii) 2 l of the appropriate no RT control reaction.Templates were amplified using TaKaRa ExTaq (Fisher Scientific; Pittsburgh, PA) and the following thermal cycle: i) initial denaturation for 10 min at 95C, ii) 35 cycles of 95C for 30 sec, 58C for 30 sec and 72C for 1 to 4 min (depending on the size of the expected product), and iii) final extension at 72C for 10 min.PCR products were separated on a 1% agarose gel and visualized using ethidium bromide.
4.9 5' Rapid Amplification of cDNA Ends (RACE) and promoter analysis The approximate recA transcriptional start sites were determined using the 2 nd Generation 5'/3' RACE Kit (Roche Applied Science; Mannheim, Germany) according to the manufacturer's instructions.Briefly, two first-strand cDNA synthesis reactions were performed with 1 g of total RNA and antisense primers recA operon F3 and recA operon F5 to determine the transcritptional start sites immediately upstream of BB0131 (recA) and BB0133.The two cDNA reactions were purified using the High Pure PCR Product Purification Kit (Roche Applied Science).The purified cDNAs were divided into two reaction tubes and a homopolymeric A or G-tail was added to the 3' end using recombinant terminal deoxynucleotidyl transferase and dATP or dGTP.A PCR product was amplified from each tailed cDNA using the appropriate anchor primer (oligo dT-AP or oligo dC-AP) and a nested gene specific primer.Nested primer recA operon F4 was used to amplify tailed cDNA generated with recA operon F3, and nested primer recA operon F6 was used to amplify tailed cDNA generated with recA operon F5.The PCR products were subjected to electrophoresis and the bands were gel extracted using the QIAquick PCR Purification Kit (Qiagen).The sequence of the purified PCR products were determined using either recA operon F4 or recA operon F8 and aligned to the upstream sequence of recA or BB0133 to determine the transcriptional start sites.Promoter analysis was carried out by visual inspection and comparison of the region upstream of the approximate transcriptional start sites with previously described RpoD, RpoS and RpoNdependent promoter sequences in B. burgdorferi.

Nucleotide sequencing and computer analysis
Nucleic acid sequencing was performed by the Rhode Island Genomics and Sequencing Center using a 3130xl Genetic Analyzer (Applied Biosystems; Forest City, CA).Sequencing reactions were prepared using the BigDye Terminator v3.0 Cycle Sequencing Kit.Sequences were analyzed using DNASTAR Lasergene software (DNASTAR, Inc.; Madison, WI) and Vector NTI (Invitrogen, Corp.).

Conclusions
In this study we demonstrate that B. burgdorferi undergoes a genetic adaptation during GlcNAc starvation that allows cells to grow to a high cell density when re-introduced to the starvation condition.We also show that recA mutant cells do not respond to GlcNAc starvation the same as wild-type cells, suggesting a role for RecA in this process.Finally, we characterize the recA transcriptional unit and provide evidence that this gene is controlled transcriptionally from two separate promoters.

Fig. 1 .
Fig. 1.Adaptation of B31-A to GlcNAc starvation.(A) Growth of B31-A in BSK-II with and without 1.5mMGlcNAc.Late-log phase cells were diluted to 1.0  10 5 cells ml -1 in complete BSK-II (open circle) or BSK-II lacking GlcNAc (closed circle), and cells were enumerated daily by darkfield microscopy.(B) Growth of B31-A in BSK-II without GlcNAc after dilution from the first exponential phase of a culture starved for GlcNAc.Cells starved for GlcNAc (closed gray circle) were diluted to 1.0  10 5 cells ml -1 at day 2 (closed triangle) and day 3 (open triangle) into BSK-II without GlcNAc, and growth was followed daily by darkfield microscopy.(C) Growth of B31-A in BSK-II without GlcNAc after dilution from the second exponential phase of a culture starved for GlcNAc.Cells starved for GlcNAc (closed gray circle) were diluted to 1.0  10 5 cells ml -1 at day 8 (closed square), day 9 (open square), day 10 (inverted closed triangle) and day 11 (inverted open triangle) into BSK-II without GlcNAc, and growth was followed daily by darkfield microscopy.These growth curves are representative of four independent growth experiments.

a
Recombination frequency was determined by transforming each strain with pBB0450.Transformation frequency is the number of transformants obtained after electroporation of cells with ErmC/pCE320 divided by the number of CFU obtained without antibiotic selection b Recombination frequency is the number of recombinants obtained after electroporation of cells with pBB0450.1 (rpoN::ermC) divided by the number of CFU obtained without antibiotic selection c Number of transformation experiments performed with ErmC/pCE320 to determine transformation frequency.d Number of transformation experiments performed with pBB0450.1 to determine recombination frequency.e ND, Not determined

123Fig. 6 .
Fig.6.Identification of the transcriptional start sites of the recA transcript.5' RACE was used to determine the approximate transcriptional start sites of the recA transcript.A-tailed PCR products obtained using primer F4 (A; lane 2) or primer F6 (B; lane 2) with the oligo dT-anchor primer were separated by gel electrophoresis on a 1% TAE agarose gel.A 1 kb ladder was used as a size standard (lane 1).(C) DNA sequence of the dA-tailed 5' RACE product obtained with primer F4 showing the approximate transcriptional start site (enlarged font).Underlined guanosines indicate site of mispriming with the oligo dC-anchor primer and dG-tailed 5' RACE product.(D) DNA sequence of the dA-tailed 5' RACE product obtained with primer F6 showing the approximate transcriptional start site (enlarged font).

Fig. 7 .
Fig. 7. Identification of the putative recA promoters.Comparison of the putative recA promoters with the RpoD, RpoS and RpoN-dependent promoters in B. burgdorferi.Nucleotide positions in the predicted recA promoters that match the consensus RpoDdependent promoter sequence are highlighted.

Fig. 8 .
Fig. 8. Growth of JA10 (recA mutant) after adaptation to GlcNAc starvation.Growth of JA10 in BSK-II with and without 1.5 mM GlcNAc, and growth of JA10 cells after adaptation to GlcNAc starvation.Late log phase cells were diluted to 1.0  10 5 cells ml -1 in complete BSK-II (open circle) or BSK-II lacking GlcNAc (closed circle), and cells were enumerated daily by darkfield microscopy.Cells adapted to GlcNAc starvation were diluted (arrows) to 1.0  10 5 cells ml -1 at Day 12 (closed triangle) and Day 13 (open triangle) in BSK-II without GlcNAc, and growth was followed daily by darkfield microscopy.

a
Number of independent growth experiments conducted in which cells were diluted from BSK-II lacking GlcNAc into the same medium b Adapted -the number of times cells reached high cell density (1-2  10 7 cells ml -1 ) in a single exponential phase c Not Adapted -the number of times biphasic growth was observed d Percent Adapted calculated as the number adapted (b) divided by the total number of experiments (a) multiplied by 100

Table 4 .
Strains and plasmids used in this study. www.intechopen.com