Borrelia burgdorferi malQ mutants utilize disaccharides and traverse the enzootic cycle

Abstract

Borrelia burgdorferi, the causative agent of Lyme disease, cycles in nature between a vertebrate host and a tick vector. We demonstrate that B. burgdorferi can utilize several sugars that may be available during persistence in the tick, including trehalose, N-acetylglucosamine (GlcNAc) and chitobiose. The spirochete grows to a higher cell density in trehalose, which is found in tick hemolymph, than in maltose; these two disaccharides differ only in the glycosidic linkage between the glucose monomers. Additionally, B. burgdorferi grows to a higher density in GlcNAc than in the GlcNAc dimer chitobiose, both of which may be available during tick molting. We have also investigated the role of malQ (bb0166), which encodes an amylomaltase, in sugar utilization during the enzootic cycle. In other bacteria, MalQ is involved in utilizing maltodextrins and trehalose, but we show that, unexpectedly, it is not needed for B. burgdorferi to grow in vitro on any of the sugars assayed. In addition, infection of mice by needle inoculation or tick bite, as well as acquisition and maintenance of the spirochete in the tick vector, does not require MalQ.

Introduction

Borrelia burgdorferi is the spirochete that causes Lyme disease (Burgdorfer et al., 1982; Benach et al., 1983; Steere et al., 1983; Radolf et al., 2012); its enzootic cycle involves an Ixodes tick vector and a vertebrate host (Lane et al., 1991; Spielman, 1994; Piesman & Schwan, 2010). Following acquisition by a feeding tick, B. burgdorferi persists for several months until transmission to a vertebrate, typically a mammal. Little is known about the physiology of the spirochete and its metabolic requirements in the two distinct environments encountered in the enzootic cycle (Gherardini et al., 2010). Disaccharides and oligosaccharides may serve as carbon and energy sources for B. burgdorferi in vivo. Trehalose, an α(1→1)α glucose disaccharide, is found in tick hemolymph (Barker & Lehner, 1976). Chitobiose, a β(1→4)-linked dimer of N-acetylglucosamine (GlcNAc) monomers, also may be available to the spirochete during the chitin rearrangement that occurs as the tick molts; B. burgdorferi can utilize chitobiose in vitro (Tilly et al., 2001).

Escherichia coli and other bacteria can utilize maltose, an α(1→4) glucose disaccharide, as a carbon source (Boos & Shuman, 1998). Maltose and maltodextrins are degraded by amylomaltase, encoded by the malQ gene, and E. coli malQ mutants are unable to grow on maltose (Monod & Torriani, 1948, 1950; Wiesmeyer & Cohn, 1960a, b; Pugsley & Dubreuil, 1988). B. burgdorferi has a malQ homolog (bb0166) (Fraser et al., 1997) and can utilize maltose as a carbon source (von Lackum & Stevenson, 2005). Sequence analysis suggests that MalQ in B. burgdorferi is unusual: it is missing one of four otherwise completely conserved residues (Lys instead of Arg at position 308) (Godány et al., 2008). Godány et al. (2008) purified recombinant B. burgdorferi amylomaltase (MalQ) and demonstrated the release of glucose in the dextrinyl transferase reaction with maltose as well as other maltodextrins as substrates.

Maltose and proteins involved in maltose transport or catabolism, in addition to their roles in metabolism, appear to function as a signaling system to sense the external environment and regulate virulence factors in some bacterial pathogens. For example, maltose inhibits secretion of cholera toxin and a malQ mutant of Vibrio cholerae has attenuated virulence in an animal model (Lång et al., 1994). Moreover, a maltose transport protein and maltodextrin-binding proteins have been implicated in the virulence of streptococci (Shelburne et al., 2006). Therefore, we hypothesized that B. burgdorferi may detect carbohydrates present in the incoming blood meal during tick feeding and/or during persistence in the tick midgut, especially during the molt, via the maltose system and MalQ. Carbohydrate variation may represent another environmental factor, in addition to temperature (Schwan et al., 1995; Stevenson et al., 1995; Fingerle et al., 2000; Yang et al., 2000; Revel et al., 2002; Alverson et al., 2003; Ojaimi et al., 2003), pH (Carroll et al., 1999; Yang et al., 2000), oxygen (Seshu et al., 2004), carbon dioxide (Hyde et al., 2007), and an unidentified factor in blood (Tokarz et al., 2004), sensed by B. burgdorferi to identify the external milieu and alter gene expression to facilitate transmission to and colonization of the mammalian host (Singh & Girschick, 2004; Samuels, 2011; Radolf et al., 2012). Our results demonstrate that B. burgdorferi can utilize trehalose, maltose, GlcNAc, and chitobiose as the main carbon source. However, malQ was required neither for disaccharide utilization nor animal infection and tick persistence.

Materials and Methods

Bacterial strains and culture conditions

Low-passage B. burgdorferi strains B31-A3 (Elias et al., 2002) and 297 (BbAH130) (Hübner et al., 2001), and genetically manipulated derivatives, were maintained in Barbour-Stoenner-Kelly II (BSK II) liquid medium containing 6% rabbit serum (Barbour, 1984) without gelatin (Samuels, 1995). To examine carbohydrate utilization, BSK II (containing GlcNAc) was also prepared without additional glucose, or with 15 mM maltose (EM Science, Hatfield, PA, USA), trehalose (Sigma), GlcNAc (Sigma), or diacetyl chitobiose (V-Labs, Covington, LA, USA) in place of 15 mM glucose (Sigma). Cell density was assayed as previously described by either measuring the OD₆₀₀ of cultures resuspended in one-tenth volume of Dulbecco’s phosphate-buffered saline (138 mM NaCl, 2.7 mM KCl, 8.1 mM Na₂HPO₄, and 1.5 mM KH₂PO₄; dPBS) (Samuels & Garon, 1993) or enumeration using a Petroff-Hausser counting chamber (Caimano et al., 2004).

Construction of malQ mutants

The malQ gene (bb0166) was disrupted by insertion of either flgBp-aadA (conferring streptomycin and spectinomycin resistance) (Frank et al., 2003) or flgBp-aacC1 (conferring gentamicin resistance) (Elias et al., 2002). Genomic regions flanking malQ were amplified by PCR and assembled using restriction sites introduced in the oligonucleotide primers (Table 1). The two flanking sequences were cloned into pCR™2.1-TOPO and ligated together to generate a 2.2-kb recombination substrate that lacks most of the malQ ORF. The antibiotic resistance cassettes were cloned into a synthetic AatII site; the plasmid was linearized with AhdI and electroporated into competent B. burgdorferi as previously described (Samuels, 1995; Gilbert et al., 2007; Lybecker & Samuels, 2007). Transformants were cloned in liquid BSK II medium in 96-well plates (Yang et al., 2004) containing either 50 μg ml⁻¹ streptomycin or 40 μg ml⁻¹ gentamicin at 34°C in a 1.5% CO₂ atmosphere. Positive clones were screened by PCR and assayed for the presence of plasmids lp28-1, lp28-4, lp25, and lp54 (Purser & Norris, 2000; Labandeira-Rey & Skare, 2001).

Table 1.

Oligonucleotides used in this study.

Name	Sequence (5′-3′)
malQ 385F	GCTTCGGCTGATGAGGTTAG
malQ 630R	TAATACGACTCACTATAGGCAAACTGTG
malQ U165F	AATGGTGTAGTCGCCTTGG
malQ U165F+AatII	GACGTCAATGGTGTAGTCGCCTTG
malQ D1681R	AAAAACCTTGAAAACAGAAAAG
malQ U940F	TTTTATTTTTTTTTCTCGGATG
malQ U25R+NdeI+AatII	CATATGATTAATTGACGTCTTAAATAAATTCCCACTG
malQ 1497F+AatII	GACGTCACAAGGCTTTATGGCAGGGC
malQ 1521R+AatII	GACGTCTTAAGCCCTGCCATAAAG
malQ D1252R	TTTTTATTCCAAAGGGCG
flgB 5′	GCTAGCTAATACCCGAGC
flaA 64F	GCTCAAGAGACTGATGGATTAGC
flaA 284R	TAATACGACTCACTATAGGCGCAGAAGG

The malQ mutants were trans-complemented by amplifying the malQ gene, including 165 bp of upstream sequence, using primers malQ U165F+AatII and malQ 1521R+AatII (Table 1). The PCR product was cloned into pCR™2.1-TOPO and confirmed by DNA sequencing. The malQ gene and the shuttle vector pBSV2 (Stewart et al., 2001) were digested with AatII and ligated together to generate pBSmalQ. Competent malQ mutant strains were electroporated with the pBSmalQ and selected in liquid BSK II medium containing 200 μg ml⁻¹ kanamycin.

RNA isolation and RT-PCR

B. burgdorferi cultures were grown at 35°C to late log phase and RNA isolated using TRIzol™ Reagent (Gibco BRL) as previously described (Lybecker & Samuels, 2007). RNA was treated with DNase I (Invitrogen). cDNA was synthesized using the RETROscript™ kit (Ambion) according to the manufacturer’s instructions. cDNA was analyzed by PCR using primers malQ 385F and malQ 630R or flaA 64F and flaA 284R (Table 1).

B. burgdorferi infection in an experimental tick-mouse cycle

The University of Montana Institutional Animal Care and Use Committee approved all mouse experiments. C3H-HeJ female mice were intraperitoneally needle-inoculated with 1 × 10⁴ cells of wild-type, malQ mutant, or complemented 297 clones (Barthold et al., 1990; Barthold et al., 2010). Ear biopsies were taken three weeks post-inoculation and cultured in BSK II containing 50 μg ml⁻¹ rifampicin, 20 μg ml⁻¹ phosphomycin and 2.5 μg ml⁻¹ amphotericin B. Mice were sacrificed five weeks post-injection and ear biopsies, ankles, and bladders were collected and cultured as described above. Cultures were screened for B. burgdorferi by dark-field microscopy.

To examine B. burgdorferi acquisition by ticks, unfed naive Ixodes scapularis larvae (National Tick Research and Education Resource, Oklahoma State University) were allowed to feed to repletion on infected mice five weeks post-injection. Five to ten days after feeding, ticks were crushed with a pestle in a 1.5-ml tube (Jewett et al., 2009) and DNA was isolated (Samuels & Garon, 1993). PCR using primers to the flaA gene (Table 1) was used to detect B. burgdorferi. To follow transmission by tick bite, five infected nymphs were placed on a naive C3H-HeJ female mouse and allowed to feed to repletion. Mouse ear biopsies, bladder tissue and ankle joints were collected five weeks post-tick feeding, cultured in BSK II and screened for B. burgdorferi as described above.

Immunofluorescence microscopy

To determine if fed nymphs were infected with B. burgdorferi, tick midguts were dissected and processed for immunofluorescence microscopy as previously described (Schwan & Piesman, 2000). Briefly, ticks were placed in 10 μl of dPBS with 5 mM MgCl₂ and the midguts were dissected with forceps on silane-coated slides (LabScientific, Inc.) under a dissecting microscope. Midguts were allowed to air dry at room temperature for 30 min before being fixed in acetone for 10 min at room temperature. Slides were washed for 10 min, three times, in dPBS with 5 mM MgCl₂ and 1% goat serum, and incubated with rabbit polyclonal anti-B. burgdorferi antibodies (a gift from T. Schwan) at 1:50 dilution for 1 h. Slides were then washed for 10 min, three times, in dPBS/5 mM MgCl₂/1% goat serum and incubated in goat anti-rabbit AlexaFluor® 488 antibodies (Molecular Probes) at 1:500 dilution for 1 h. Slides were then washed again for 10 min, three times, in dPBS/5 mM MgCl₂/1% goat serum with the final wash containing wheat germ agglutinin-AlexaFluor® 594 (Molecular Probes) at 1:200 dilution. A coverslip was mounted with ProLong Gold antifade reagent (Molecular Probes) and sealed with Permount (Fisher Scientific). Images are a single optical section collected using a FluoView FV1000 Olympus IX81 confocal microscope with a 60 X, NA 1.42 objective. Images were processed using ImageJ (National Institutes of Health; http://rsbweb.nih.gov/ij/) and Pixelmator (Pixelmator Team, Ltd).

Results and Discussion

Disaccharide utilization

Trehalose is a glucose disaccharide found in tick hemolymph (Barker & Lehner, 1976). We tested if trehalose can serve as a carbon and energy source because B. burgdorferi would have access to the sugar as it moves through the hemolymph during transmission to the mammalian host. We also examined growth on maltose, another glucose disaccharide that differs from trehalose in the glycosidic linkage. B31-A3 wild type was grown in BSK II (containing rabbit serum) either without an additional carbon source or with glucose, maltose or trehalose as the sole carbon source other than GlcNAc, which is required for growth (Tilly et al., 2001). B31-A3 grew on trehalose as well as on glucose (Fig. 1a). To the best of our knowledge, this is the first report of B. burgdorferi utilizing trehalose as an energy source. Maltose also supported growth as previously shown (von Lackum & Stevenson, 2005), but cells reached a lower cell density than during growth with glucose (Fig. 1a). A growth curve (Fig. 1b) demonstrated that the decreased cell density in maltose was not due to an extended lag phase from adaptation to the alternative carbon source, which suggests that B. burgdorferi is attenuated in either maltose transport or catabolism. Although B. burgdorferi can utilize many carbohydrates in vitro (von Lackum & Stevenson, 2005), trehalose may be an important energy and carbon source, along with glycerol (He et al., 2011; Pappas et al., 2011), for persistence in the tick vector.

Fig. 1.

Glucose dimer utilization by B. burgdorferi. (a) B. burgdorferi strain B31-A3 wild-type was inoculated at 1 × 10⁴ cells ml⁻¹ and grown for 7 d at 35°C in modified BSK II with 1.8 mM GlcNAc and either no additional sugar (none) or 15 mM of glucose, maltose, or trehalose. Cell density was determined by measuring the OD₆₀₀ of cultures. Values are the mean ± SEM from two separate experiments. (b) Strain B31-A3 wild-type cultures were grown at 35°C in modified BSK II (described above) with no other sugar (open circles), 15 mM of glucose (filled circles), or 15 mM maltose (filled triangles). Cell densities were determined using a Petroff-Hausser counting chamber on days three through seven after initial inoculation with 1 × 10⁴ cells ml⁻¹. Asterisk (*) indicates P < 0.05 by one-way ANOVA comparing growth on maltose and glucose. Values are the mean ± SEM from three separate experiments.

malQ mutagenesis

Maltose is a substrate for MalQ amylomaltases (Boos & Shuman, 1998) and B. burgdorferi has a malQ gene (Fraser et al., 1997; Godány et al., 2008). We hypothesized that MalQ may use trehalose as a substrate in addition to or instead of maltose because the maltose transport system in Thermococcus litoralis is promiscuous for trehalose transport (Xavier et al., 1996; Horlacher et al., 1998). Furthermore, borrelial proteins acting on different sugars than predicted is not unprecedented: the chb gene products were initially categorized as transporting and modifying cellobiose (Fraser et al., 1997), but later found to recognize chitobiose (Tilly et al., 2001). We took a reverse genetic approach to examine malQ function in B. burgdorferi (Brisson et al., 2012). Almost the entire malQ ORF was deleted in B. burgdorferi strains B31-A3 and 297 by exchanging it with the antibiotic resistance cassettes flgBp-aadA (streptomycin and spectinomycin resistance) or flgBp-aacC1 (gentamicin resistance) (Fig. 2a). PCR analyses of genomic DNA from transformants and parental strains confirmed that the antibiotic resistance cassettes replaced the malQ gene (Fig. 2c). In addition, the malQ gene was not detected by PCR in the malQ::aadA and malQ::aacC1 mutants (Fig. 2c). The malQ gene was cloned into the shuttle vector pBSV2 (Stewart et al., 2001) to generate pBSmalQ (Fig. 2b), which was used to complement the malQ mutants in trans yielding strains malQ::aadA/pBSmalQ and malQ::aacC1/pBSmalQ. The malQ transcript was detected by RT-PCR in both the wild-type B31-A3 (Fig. 2d, lane 1) and the complemented malQ::aadA/pBSmalQ strains (Fig. 2d, lane 7), but not in the malQ::aadA mutant strain (Fig. 2d, lane 4).

Fig. 2.

Mutagenesis of malQ. (a) Schematic of malQ null mutants generated with streptomycin and spectinomycin (aadA) or gentamicin (aacC1) resistance genes that were fused to the B. burgdorferi flgB promoter. Primers a (flgB 5′), b (malQ U165F), c (malQ 385F), and d (malQ 630R) are used to analyze genomic DNA and cDNA. (b) The malQ ORF plus 165 bp of upstream sequence was cloned into the shuttle vector pBSV2 to yield pBSmalQ to trans-complement malQ null mutants. The plasmid carries the gene aphI, which confers kanamycin resistance, fused to the flgB promoter (flgBp), a B. burgdorferi replication origin (cp9 PF 57, 50, 49 +IR) from the 9-kb circular plasmid cp9, an E. coli replication origin (ColE1), and a multiple cloning site (MCS) from pCR^®-XL-TOPO. (c) PCR analyses of genomic DNA from wild-type, malQ mutant (malQ::aadA and malQ::aacC1), and complemented (malQ::aadA/pBSmalQ and malQ::aacC1/pBSmalQ) strains. No template is a negative control. Primers a and b were used to determine the orientation of antibiotic resistance genes. Primers c and d were used to detect the malQ gene by PCR. (d) RT-PCR analysis of RNA isolated from wild-type, malQ mutant and complemented strains grown at 35°C. cDNA was synthesized and analyzed by PCR to detect malQ (lanes 1, 4 and 7) or flaA as a positive control (lanes 2, 5 and 8); control reactions without reverse transcriptase were also run (lanes 3, 6 and 9).

Next we examined if MalQ plays a role in carbohydrate utilization. Unexpectedly, malQ was not required for growth on either maltose or trehalose in vitro (Fig. 3a). These results suggest that B. burgdorferi has an alternative pathway to catabolize these disaccharides; in fact, the genome carries a homolog of treA, encoding a putative trehalase (Fraser et al., 1997), although preliminary efforts to disrupt this gene have not been fruitful. We also tested the ability of B. burgdorferi to grow on GlcNAc and its dimer, diacetyl chitobiose, which are components of the tick exoskeleton and the peritrophic membrane that surrounds the blood meal. Chitobiose has previously been shown capable as serving as a carbon and energy source (Tilly et al., 2001). We found that B31-A3 wild type grew at least as well in GlcNAc as in glucose, while cells grown in chitobiose reached a lower cell density after 7 d (Fig. 3b). Again, growth on GlcNAc or chitobiose did not require malQ in vitro (Fig. 3b). These results do not eliminate the possibility that MalQ may be essential to utilize another, as yet unidentified, carbohydrate. In fact, as noted by Godány et al. (2008), the B. burgdorferi MalQ sequence is unusual with several highly conserved amino acids adjacent to the MalQ catalytic triad differing in Borrelia spp. compared to the glycoside hydrolase family amylomaltases from other bacteria, plants and archaea.

Fig. 3.

Growth of malQ mutants on different sugars. (a) Strains B31-A3 wild type (black bars), malQ::aadA (hatched bars), and malQ::aadA/pBSmalQ (gray bars) were inoculated at 1 × 10⁴ cells ml⁻¹ and grown for 7 d at 35°C in modified BSK II with 1.8 mM GlcNAc and either no additional sugar (none) or 15 mM of glucose, maltose or trehalose. Cell density was determined by measuring the OD₆₀₀. Values are the mean ± SEM from two separate experiments. (b) Strains B31-A3 wild type (black bars) and malQ::aadA (hatched bars) were inoculated at 1 × 10⁴ cells ml⁻¹ and grown for 7 d at 35°C in modified BSK II with 1.8 mM GlcNAc and either no additional sugar (none) or 15 mM of glucose, GlcNAc or chitobiose. Cell density was determined using a Petroff-Hausser counting chamber. Values are the mean ± SEM from two separate experiments.

Since MalQ and maltose transport proteins have been implicated in expression of virulence factors in V. cholera and streptococci, respectively (Lång et al., 1994; Shelburne et al., 2006), presumably to relay information about the environment, we assayed if malQ has a similar role in B. burgdorferi. Neither the malQ mutation nor varying carbohydrates available affected expression of outer surface lipoprotein C (data not shown), which is essential for transmission or mammalian infection (Grimm et al., 2004; Pal et al., 2004).

MalQ in the enzootic cycle

While our data suggest that MalQ does not have an essential role in disaccharide utilization in vitro, we hypothesized that MalQ may be important in the enzootic cycle for metabolism or gene regulation in vivo. Therefore, we assayed the malQ::aadA mutant strain in the experimental tick-mouse model. Wild-type, malQ::aadA and complemented strains were needle-inoculated into mice; ear biopsies were collected three weeks after injection, cultured in BSK II and examined for spirochetes by dark-field microscopy. In addition, ear, ankle and bladder tissues were dissected and cultured for B. burgdorferi at five weeks post-inoculation. The malQ mutant was infectious by needle inoculation and successfully disseminated to the ear, ankle and bladder of the mice (Table 2).

Table 2.

Mouse infectivity of B. burgdorferi strains.

		No. of positive mice by culture
		3 weeks		5 weeks
Strain	Inoculation method	Ear	Ear	Ankle	Bladder
Wild-type 297	needle	10/10	3/3	1/1	1/1
	tick bite	3/3	3/3	1/1	1/1
malQ::aadA	needle	9/10	3/3	3/3	2/2
	tick bite	3/3	3/3	3/3	3/3
malQ::aadA/pBSmalQ	needle	6/6	3/3	2/2	3/3
	tick bite	3/3	3/3	2/2	2/2

To examine the role of MalQ in B. burgdorferi acquisition, naive I. scapularis larvae were allowed to feed to repletion on mice infected with wild-type 297, malQ::aadA or complemented strains. Five to 10 days after feeding to repletion, PCR analysis revealed that larvae acquired B. burgdorferi from infected mice independent of the presence of malQ (seven out of seven ticks were infected with each strain).

Larvae that had fed to repletion on infected mice were allowed to molt into nymphs to examine if MalQ functions in tick persistence. After three to four weeks, five nymphs infected with each strain were then fed to repletion on naive mice. About seven days after feeding to repletion, the midguts were dissected and processed for immunofluorescence microscopy using anti-Borrelia antibodies (green) and wheat germ agglutinin-AlexaFluor® 594 that stains tick cells (red). All midguts examined contained B. burgdorferi at similar densities by immunofluorescence microscopy (Fig. 4), suggesting that survival during molting and persistence in nymphs following the blood meal does not require MalQ.

Fig. 4.

B. burgdorferi persistence in fed nymphs does not require malQ. Nymphs infected with (a) wild-type strain 297, (b) malQ::aadA or (c) malQ::aadA/pBSmalQ strains were fed to repletion on naive mice. Midguts of fed nymphs were dissected and examined for the presence of B. burgdorferi by immunofluorescence microscopy using anti-B. burgdorferi antibodies (green). Wheat germ agglutinin-AlexaFluor® 594 (red) was used to visualize tick midguts. Scale bar equals 10 μm.

Although mouse infection by needle inoculation was malQ-independent, the natural route of transmission is by tick bite. Nymphs infected with wild-type, malQ::aadA or complemented strains were allowed to feed to repletion on naive mice to test if transmission of B. burgdorferi by tick bite requires malQ. Five nymphs infected with each strain were fed on three separate mice. Three weeks after tick feeding, ear biopsies were taken, cultured and screened for B. burgdorferi as described above. malQ mutants were able to transmit from ticks to mice (Table 2). Ear, ankle and bladder tissues were cultured for B. burgdorferi at five weeks post-tick feeding, demonstrating that dissemination following infection by tick bite also did not require MalQ (Table 2).

Concluding remarks

Although MalQ seems to have no apparent role in the experimental enzootic cycle of B. burgdorferi or in the ability of the spirochete to utilize glucose disaccharides, the malQ gene is conserved in all sequenced genomes of Borrelia species, albeit encoding an unusual yet functional amylomaltase (Godány et al., 2008). Therefore, MalQ likely has a function that was not discernible in our tick-mouse model system, perhaps related to survival in the tick in nature. There is precedent for our apparently enigmatic results: ospD, encoding an outer surface lipoprotein, and chbC, encoding the chitobiose transporter, are conserved genes that are not essential in an experimental enzootic cycle (Tilly et al., 2004; Li et al., 2007; Stewart et al., 2008). Interestingly, our data indicate that B. burgdorferi can utilize trehalose, which may be physiologically relevant in the tick since trehalose is present in hemolymph (Barker & Lehner, 1976). This may be an important carbon and energy source as B. burgdorferi moves from the tick midgut via the hemolymph to the salivary glands during feeding and transmission.